Peroxisome proliferator-activated receptor-γ coactivator-1α and -1β (PGC-1α and PGC-1β) were overexpressed by adenovirus-mediated gene transfer in cultures of primary rat skeletal muscle cells derived from neonatal myoblasts. Effects on muscle fiber type transition and metabolism were studied from days 5 to 22 of culture. PGC-1α and PGC-1β overexpression caused a three- to fourfold increase in mRNA level, a doubling of enzymatic activity of citrate synthase, a slight increase in short-chain acyl-CoA dehydrogenase mRNA, a doubling of the mRNA level, and a 30–50% increase in enzymatic activity of glyceraldehyde-3-phosphate dehydrogenase. Lactate dehydrogenase or creatine kinase activity was unchanged. PGC-1α enhanced glycogen buildup twofold at 5 or 25 mM glucose, whereas PGC-1β caused a decrease. Both PGC-1α and PGC-1β overexpression caused a faster maturation of myotubes, as seen by mRNA downregulation of the immature embryonal and perinatal myosin heavy-chain (MHC) isoforms. PGC-1α or PGC-1β overexpression enhanced mRNA of the slow oxidative-associated MHC isoform MHCIb and downregulated mRNA levels of the fast glycolytic-associated MHC isoforms MHCIIX and MHCIIB. Only PGC-1β overexpression caused an increase in mRNA of the intermediary fast oxidative-associated MHC isoform MHCIIA. PGC-1α or PGC-1β overexpression upregulated GLUT4 mRNA and downregulated myocyte enhancer factor 2C transcription factor mRNA; only PGC-1α overexpression caused an increase in the mRNA expression of TRB3, a negative regulator of insulin signaling. These results show that both PGC-1α and PGC-1β are involved in the regulation of skeletal muscle fiber transition and metabolism and that they have both overlapping and differing effects.
- skeletal muscle cell culture
- peroxisome proliferator-activated receptor γ coactivator-1
- myosin heavy chain
- enzyme activities
the beneficial effect of exercise on various diseases is widely acknowledged (7), and exercise has proven particularly successful as a treatment for the metabolic syndrome, including type 2 diabetes (48); however, the exact molecular pathway responsible for the effect is largely unknown. Recently, genes involved in oxidative phosphorylation in skeletal muscle were found to be downregulated in patients with type 2 diabetes (37, 43), and speculations about type 2 diabetes being caused by mitochondrial dysfunction, primarily in the type I skeletal muscle fibers, have surfaced (13, 31, 51). Regular exercise has been shown to induce changes in both skeletal muscle metabolism and muscle fiber type over time, most notably an increase in mitochondrial content and oxidative metabolism as well as a shift toward a more slow oxidative fiber type (8, 44). Single bouts of exercise transiently change the mRNA expression levels of a vast amount of genes both during and immediately after exercise (3, 10, 45). It has been speculated that it is the repeated transient induction of these genes, leading to an average higher gene expression over time that is ultimately responsible for the changes seen in skeletal muscle phenotype after regular exercise (57).
One of the genes induced in skeletal muscle after exercise is the peroxisome proliferator-activated receptor (PPAR)-γ coactivator-1α (PGC-1α; see Refs. 1, 2, and 10). PGC-1α, a coactivator of several nuclear receptors and other transcription factors, has been shown to be involved in the regulation of mitochondrial biogenesis, adaptive thermogenesis, and enzymes involved in oxidative metabolism as well as glucose transport in skeletal muscle (27, 46). Furthermore, PGC-1α was recently shown to be involved in muscle fiber-type switching during development (29). Peroxisome proliferator-activated receptor-γ coactivator-1β (PGC-1β), another family member of the PGC-1 family of transcriptional coactivators, has been less extensively studied than PGC-1α, but like PGC-1α it seems to be involved in the regulation of energy metabolism and energy expenditure, although with subtle differences compared with PGC-1α (21, 27–29). Even though PGC-1β is highly expressed in skeletal muscle, it does not seem to be regulated by endurance exercise (34). The involvement of PGC-1β in skeletal muscle fiber plasticity has not been studied, although overexpression of PGC-1α and PGC-1β in skeletal muscle cells suggests a different role of these coactivators in controlling mitochondrial function (52). Both PGC-1α and PGC-1β have been shown to be downregulated in skeletal muscle in type 2 diabetes (43), and a correlation to the downregulation of genes involved in oxidative phosphorylation has been established (37).
By using adenoviral-mediated overexpression of either PGC-1α or PGC-1β, we examined the different effects of increased levels of these coactivators on the metabolic and contractile phenotype of skeletal muscle cells. PGC-1α and PGC-1β induced similar changes in both the metabolic and contractile phenotype of primary rat skeletal muscle cells, with subtle differences between PGC-1α and PGC-1β overexpression.
MATERIALS AND METHODS
FCS and horse serum were from In Vitro (Fredensborg, Denmark). Matrigel was from BD Biosciences (Bedford, MA), accutase was from Cytotech (Hellebaek, Denmark), and all plasticware was from Nunc (Roskilde, Denmark). ADP, NADH, creatine phosphate, glucose-6-phosphate dehydrogenase, and hexokinase were from Roche (Basel, Switzerland). AMP was from Calbiochem (La Jolla, CA); gentamicin, acetyl-CoA, BSA, DTT, and NADP+ were from Gerbu (Gaiberg, Germany); and bovine insulin was from Novo Nordisk (Gentofte, Denmark). Adeno-X Virus Purification Kit and Adeno-X Rapid Titer Kit were from Clontech (Mountain View, CA). Omniscript RT, Quantitect SYBR Green Master Mix, and Quantitect SYBR Green RT-PCR kit were from Qiagen (Valencia, CA), and the QuantumRNA 18S internal standards kit was from Ambion (Cambridgeshire, UK). All other chemicals and enzymes were from Sigma-Aldrich (St. Louis, MO).
Primary skeletal muscle cell culture preparation.
Primary cultures of rat myotubes were prepared from the hindleg muscles from <24-h-old Wistar rats by a modification of a previously described method for rabbits (24). On day 0, 25 newborn rats were decapitated, and the bodies were carefully rinsed in 70% ethanol and allowed to dry for 15 min, after which the hindlegs were cut off. With the use of a stereomicroscope, the legs were skinned, washed two times in ice-cold Dulbecco's PBS without Ca2+ and Mg2+ (CMf-DPBS), and carefully trimmed for muscle tissue in a few drops of CMf-DPBS. The muscle tissue was cut with a pair of iris scissors for ∼10 min in a few drops of CMf-DPBS, transferred to an extraction beaker, and incubated for 5 min at 37°C with magnetic stirring at 300 revolutions/min (30 s left, 30 s right) in 40 ml of a trypsin solution (1.25 mg/ml). The tissue was allowed to sediment, the supernatant was discarded, and 40 ml of fresh trypsin solution (as above) were added together with 1 ml of collagenase I solution (32 mg/ml); the suspension was incubated at 37°C for 25 min with stirring as above. The suspension was triturated every 10 min at least 15 times with a wide mouth pipette. After 25 min of digestion, the suspension was again triturated at least 15 times, and the undigested tissue was allowed to sediment for a few minutes. The supernatant was filtered through a stainless steel sieve (0.2-mm pore size) into a 50-ml tube containing 1 ml plating medium (DMEM containing 5 mM glucose, 2 × 105 IU/l penicillin, 50 mg/l gentamicin, 1.25 mg/l amphotericin B, and 20% FCS). The volume of filtrate should not exceed 15 ml/tube. CMf-DPBS (35 ml) was added to each tube, and the tubes were gently mixed and centrifuged at 778 g for 8 min at room temperature. The supernatant was discarded, and the pellets were resuspended in 10 ml plating medium/tube. Meanwhile, 40 ml fresh trypsin solution plus 1 ml of collagenase I solution (as above) were added to the remaining tissue in the extraction beaker, and the tissue was incubated for a further 20 min, after which the cells were collected as described above. The combined cell suspensions were filtered through a Millipore filter (20 μm pore size) in a total volume of 180 ml plating media, after which the cell suspension was transferred to three 75-cm2 culture flasks (60 ml/flask) and placed in the incubator for 50 min to allow fibroblasts to attach to the surface. The myoblasts still in suspension were carefully poured into a sterile bottle and counted, and plating medium was added to a final cell concentration of ∼0.5 × 106 cells/ml. Cell suspension (2 ml) was plated on 35-mm culture dishes precoated overnight in the incubator with a 1:100 dilution of Matrigel in DMEM. The culture medium was changed to fusion medium on day 3 (DMEM containing 5 mM glucose, 2 × 105 IU/l penicillin, 50 mg/l gentamicin, 1.25 mg/l amphotericin B, and 10% horse serum) and transduced with adenovirus on day 5, as described below. On day 7, the culture medium was changed to differentiation medium (DMEM containing 5 mM glucose, 2 × 105 IU/l penicillin, 50 mg/l gentamicin, 1.25 mg/l amphotericin B, 10% horse serum, and 1 μM dexamethasone) in which the cells were kept for the remainder of the experiment. The medium contained 10 μM cytosine β-d-arabinofuranoside from day 5 to day 9. All cultures had their medium changed every 2nd or 3rd day.
Fibroblast cultures were prepared from the cells, which adhered to plain plastic in the purification step described above. After removal of the medium containing myoblasts, the fibroblasts were washed one time with CMf-DPBS and loosened from the culture flask by addition of 10 ml accutase. Cultures from this fibroblast suspension were prepared exactly as described for myoblast cultures.
Transduction of primary skeletal muscle cell cultures with adenovirus.
Immediately before transduction, all dishes were changed to serum-free plating medium, and adenovirus, diluted in serum-free plating medium, was added to a final virus concentration of 108 infectious units/35-mm dish (resulting in a multiplicity of infection of ∼25). The dishes were gently shaken and incubated at 37°C in the incubator for 3 h, shaking the dishes gently one time per hour, after which the medium was changed to fusion medium as described above.
Preparation of cell lysates for enzyme assays.
The dishes were placed on ice, and the media were removed. Five hundred microliters of ice-cold buffer (25 mM glycyl-glycin, 150 mM KCl, 5 mM MgSO4, 5 mM EDTA, 1 mM DTT, 0.02% BSA, and 0.1% Triton X-100, pH 7.5) were added to each dish. The cell lysate was scraped off, and the resulting lysate was transferred to an Eppendorf tube, whirl mixed, frozen in liquid nitrogen, thawed on ice, and whirl mixed once again. The lysate was then centrifuged at 22,000 g for 2 min at 4°C. The supernatant was stored at −80°C for later determination of enzyme activities. Two separate 35-mm dishes were harvested for each time point.
Preparation of cell lysates from adult muscle tissue.
Male Wistar rats were sedated using thiomebumal and killed by neck dislocation. Extensor digitorum longus (EDL) or soleus muscles were dissected and freeze-clamped in liquid nitrogen. Frozen muscle tissue was stored at −80°C until time of analysis. A 5% (wt/vol) homogenate was prepared in a 2-ml Potter-Elvehjem homogenization glass on ice in 25 mM glycyl-glycin, 150 mM KCl, 5 mM MgSO4, 5 mM EDTA, 1 mM DTT, 0.02% BSA, and 0.1% Triton X-100, pH 7.5, using a Teflon piston for 1 min. The homogenate was frozen in liquid nitrogen, thawed on ice, whirl mixed, and centrifuged at 22,000 g for 2 min at 4°C. The supernatant was stored at −80°C for determination of enzyme activities.
Protein content was measured using BSA (fraction V) as standard (32).
After the medium was removed, the cells were extracted with 500 μl of 0.5 M KOH at room temperature for 10–15 min. The lysate was boiled for 10 min and frozen at −20°C for later determination of protein and glycogen content. Glycogen was measured as glucose after hydrolysis in 1 M HCl at 100°C for 2 h. Glucose was determined enzymatically in the neutralized hydrolysate using glucose oxidase as described previously (25).
Citrate synthase (CS, EC 18.104.22.168) activity was determined spectrophotometrically at 409.5 nm and 30°C by measuring the reduction rate of DTNB in a reaction mixture with 100 mM glycyl-glycin, 0.5 mM EGTA and 2 mM MgCl2, 100 μM DTNB, and 50 μM acetyl-CoA, pH 8.2, as described previously (42). The reaction was initiated by adding oxaloacetate to 50 μM.
Creatine kinase (CK, EC 22.214.171.124) activity was determined spectrophotometrically at 340 nm and 37°C in a reaction mixture with 100 mM glycyl-glycin, 10 mM MgCl2, 0.02% BSA, 5 mM DTT, 1 mM ADP, 5 mM AMP, 2 mM glucose, 1.5 mM NADP+, 3 U/ml hexokinase, and 0.175 U/ml glucose-6-phosphate dehydrogenase, pH 7.0, as described previously (42). The reaction was initiated by adding phosphocreatine to 25 mM.
Lactate dehydrogenase (LDH, EC 126.96.36.199) activity was determined spectrophotometrically at 340 nm and 37°C in 100 mM triethanolamine, 2.5 mM DTT, and 0.154 mM NADH, pH 7.4, as described previously (42). The reaction was initiated by addition of pyruvate to 5 mM.
The glyceraldehyde-3-phosphate dehydrogenase (GAPDH, EC 188.8.131.52) activity was determined spectrophotometrically at 340 nm and 37°C by measuring the rate of NAD+ reduction. To take into account possible interference from LDH, the activity was measured as (GAPDH + LDH) activity − LDH activity. LDH activity was measured in a reaction mixture with 100 mM HEPES, 2 mM MgSO4, 1 mM EDTA, 1 mM ATP, 20 μM NADH, 0.15 U/ml 3-phosphoglycerate kinase, 0.15 U/ml α-glycerophosphate dehydrogenase, and 1.5 U/ml triosephosphate isomerase, pH 7.5. The reaction was initiated by addition of pyruvate to 5 mM. GAPDH + LDH activity was measured in the same reaction mixture, and the reaction was initiated by addition of pyruvate and 3-phosphoglycerate to 5 mM each.
Total RNA was isolated from the cell cultures with TriReagent as previously described (18). Intact RNA was confirmed by denaturing agarose gel electrophoresis.
mRNA levels were determined by Northern blotting using cloned PCR products as described previously (20). The PCR primers are shown in Table 1. Total RNA (500 ng) was used per lane, and 28S rRNA was used for normalization.
Quantitative real-time RT-PCR (excluding tribbles 3).
Total RNA (500 ng) was converted to cDNA in 20 μl using the OmniScript RT (Qiagen) and poly(dT) according to the manufacturer's protocol. For each target mRNA, 0.25 μl cDNA was amplified in a 25-μl SYBR Green PCR reaction containing 1× Quantitect SYBR Green Master Mix (Qiagen) and 100 nM of each primer (Table 1). The amplification was monitored real-time using the MX3000P Real-time PCR machine (Stratagene). The threshold cycle values were related to a standard curve made with the cloned PCR products. The quantities were normalized to mRNA for the large ribosomal protein P0 (RPLP0; see Ref. 11). No statistically significant difference was found between RPLP0 mRNA levels of all samples when RPLP0 were normalized to 18S RNA levels (data not shown).
Quantitative real-time RT-PCR of tribbles 3.
Tribbles 3 (TRB3) mRNA was measured with 18S rRNA as internal standard on a Rotor-Gene 3000 (Corbett Research) employing quantitative real-time RT-PCR using the QuantiTect SYBR Green RT-PCR kit according to the manufacturer's instructions. The primers used to measure 18S rRNA were part of the QuantumRN 18S Internal Standards kit (Ambion). The primers used to measure TRB3 mRNA are shown in Table 1. All quantitative RT-PCR determinations were done in duplicate, and TRB3 mRNA levels were normalized to 18S rRNA levels.
Ad-CMV-GFP, Ad-CMV-mPGC-1α, and Ad-CMV-mPGC-1β are all recombinant adenovirus vectors (6) capable of expressing either green fluorescent protein (GFP), mouse PGC-1α, or mouse PGC-1β (52). The preparation of purified adenovirus was carried out using the Adeno-X Virus Purification Kit according to the manufacturer's instructions. Recombinant adenoviruses were titered using the Adeno-X Rapid Titer Kit (Clontech).
mRNA and glycogen data were log-transformed before statistical analysis. All graphs depict the mean (geometric mean for mRNA and glycogen) of three independent experiments done in duplicate ± SE (backtransformed for mRNA and glycogen) unless otherwise noted. All data were analyzed using mixed-model statistics on each day, with post hoc tests being Bonferroni adjusted (SAS 9.1.2; The SAS Institute). The experiments were considered to be a randomized complete block design. Statistically significant differences between GFP and control, PGC-1α, or PGC-1β or between PGC-1α and PGC-1β are indicated in the legends for Figs. 1–5.
Primary rat skeletal muscle cell cultures.
The cultures matured from single dispensed myoblasts on day 0 into myotubes exhibiting a morphology resembling striated muscle and showing spontaneous contractile activity from day 6. By visual assessment, it was estimated that >90% of the single myoblasts fused into myotubes. The cultures were readily transduced with adenoviruses on day 5 postisolation when infected with adenoviruses at a multiplicity of infection of ∼25. The level of infection could be observed visually by the expression of GFP (Fig. 1), and, judging from GFP expression, >90% of the cells were successfully transduced. The cultures showed a significant increase in PGC-1α mRNA levels in the PGC-1α-transduced cultures and a significant increase in PGC-1β mRNA levels in the PGC-1β-transduced cultures (see Fig. 4), whereas no increase in PGC-1α nor PGC-1β expression was seen in the GFP-transduced cultures. No direct visual difference between nontransduced cultures and cultures transduced with adenoviruses capable of driving the expression of either GFP, PGC-1α, or PGC-1β was observed (Fig. 1). However, a marked increase in the spontaneous contractile activity of the PGC-1α-overexpressing cultures was observed visually.
Effects of PGC-1α and PGC-1β overexpression on the activities and mRNA levels of metabolic enzymes.
The activities of four different metabolic enzymes (CS, LDH, CK, and GAPDH) were measured in the skeletal muscle cell cultures at the indicated days (Fig. 2). For comparison, the activities of LDH, CS, and CK were measured in two different rat hindleg muscles (EDL and soleus) in 3-wk-, 8-wk-, and 4-mo-old rats as well as in rat primary fibroblast cultures (Table 2).
Both PGC-1α and PGC-1β overexpression enhanced CS activity approximately twofold from days 14 to 22 (Fig. 2), indicating a shift toward a more oxidative metabolism. This corresponds well with previous observations that both PGC-1α and PGC-1β stimulate mitochondrial biogenesis and oxidative metabolism (28, 29, 37, 43, 47). As expected, the mRNA level of CS was increased in both the PGC-1α- and PGC-1β-overexpressing cultures from days 9 to 22 (see Fig. 4). The highest increase in CS mRNA levels was observed on day 9, with a three- to fourfold upregulation compared with both controls. The level of CS activity in the skeletal muscle cell cultures was comparable to the CS activity observed in adult muscle (Table 2). Fibroblast cultures, however, had about fourfold lower CS activity than the skeletal muscle cell cultures.
Transduction of the skeletal muscle cell cultures with an adenovirus expressing GFP increased the CK activity compared with the nontransduced control on days 14 and 22 (P < 0.05 and P < 0.01, respectively). This effect was, however, negated to some extent by the overexpression of PGC-1α and PGC-1β, suggesting a negative regulation of CK activity by PGC-1α and PGC-1β (Fig. 2). The levels of CK activity measured in the skeletal muscle cell cultures were ∼35-fold higher than in fibroblasts, but approximately sixfold lower than the lowest activity measured for adult muscle tissue (Table 2).
No significant change in LDH activity was observed upon overexpression of either PGC-1α or PGC-1β in the skeletal muscle cell cultures (Fig. 2). The level of LDH activity measured in the skeletal muscle cell cultures was comparable with the level of LDH activity in adult rat muscle and about twofold higher than the activity measured in fibroblasts (Table 2).
The activity of GAPDH was found to be slightly elevated by PGC-1α overexpression at the end of the culture period (Fig. 2), whereas the relative differences in the mRNA levels of GAPDH varied a lot more (Figs. 3 and 4). Most notably, a significant upregulation of GAPDH mRNA of about fourfold by PGC-1α was detected by both quantitative RT-PCR and Northern blot, but also PGC-1β overexpression upregulated GAPDH mRNA expression. The levels of GAPDH activity measured in the skeletal muscle cell cultures were comparable with the GAPDH activity observed in adult muscle (30).
The mRNA levels of short-chain acyl-CoA dehydrogenase were increased by PGC-1α and PGC-1β overexpression, although statistically significant on day 9 only (Fig. 4).
Both PGC-1α and PGC-1β cause an increase in GLUT4 mRNA levels, but only PGC-1α causes an increase in glycogen accumulation.
PGC-1α overexpression caused an increase in GLUT4 mRNA levels (Fig. 4), as has previously been reported (35). Interestingly, we observed a similar increase in GLUT4 mRNA levels in the cultures overexpressing PGC-1β (Fig. 4). PGC-1α overexpression significantly increased the glycogen content by ∼100% at both 5 and 25 mM glucose on day 9 (Fig. 5), whereas PGC-1β seemed to have the opposite effect, decreasing glycogen content, on days 14 and 22 (Fig. 5). The cultures were found not to be insulin sensitive, since treatment with or without 10 nM insulin at both 5 and 25 mM glucose resulted in no difference in glycogen accumulation (data not shown).
Both PGC-1α and PGC-1β overexpression quicken the maturation of muscle myotubes and induce a more oxidative muscle fiber type.
To determine the fiber type composition of the primary skeletal muscle cell cultures, the mRNA levels of different myosin heavy-chain (MHC) isoforms (44) were measured in the skeletal muscle cell cultures (Fig. 3). The skeletal muscle cell cultures contained up to 13% of the amount of adult MHC mRNA found in adult rat skeletal muscle (Schjerling, unpublished observation) and expressed all adult MHC isoforms normally found in adult skeletal muscle tissue. The mRNA levels of embryonal MHC (MHCemb) and perinatal MHC (MHCperi), both developmental and immature MHC isoforms, were significantly faster downregulated in the cultures overexpressing PGC-1α and PGC-1β, indicating that these cultures matured faster than the control cultures. Both PGC-1α and PGC-1β overexpression increased the mRNA levels of the slow oxidative associated MHCIb isoform, whereas only PGC-1β significantly increased the mRNA level of the intermediate fast oxidative-associated MHCIIA isoform. PGC-1α and PGC-1β overexpression both decreased the mRNA levels of the fast glycolytic-associated MHC isoforms, MHCIIX and MHCIIB, throughout the culture period. All in all, the overexpression of either PGC-1α or PGC-1β leads to a more mature and oxidative MHC isoform mRNA expression pattern in the skeletal muscle cell cultures.
PGC-1α and PGC-1β overexpression alter myocyte enhancer factor 2C and TRB3 mRNA levels.
In this study, PGC-1α and PGC-1β overexpression caused an ∼50% downregulation of myocyte enhancer factor 2C (MEF2C) mRNA levels throughout the culture period (Fig. 4). The transcription factor MEF2C has been implicated in myofiber plasticity (41) and in PGC-1α regulation of GLUT4 expression in skeletal muscle (29, 33).
PGC-1α overexpression was recently shown to enhance the expression of TRB3, a negative regulator of insulin signaling, in hepatocytes (12, 22), although some controversy exists (19). In the present experiment, TRB3 mRNA levels increased ∼25-fold from day 5 to day 9, suggesting a regulation of TRB3 expression by mechanisms involved in myotube maturation (Fig. 3). Notably, only PGC-1α overexpression was found to significantly increase the TRB3 mRNA levels on days 9 and 14, whereas no significant effect on TRB3 mRNA levels was observed in the PGC-1β-overexpressing cultures. The TRB3 mRNA levels observed on days 9, 14, and 22 were ∼10% of the level present in primary hepatocytes (data not shown).
The PGC-1 family of transcription factors comprises three members [PGC-1α, PGC-1β, and PGC-1-related coactivator (PRC)], with both PGC-1α and PRC being upregulated in skeletal muscle by acute bouts of exercise (2, 45, 49) and PGC-1β being unaffected (34). PGC-1α and PGC-1β have furthermore been shown to be differentially regulated in oxidative and glycolytic skeletal muscle tissue, with PGC-1α having a high expression level in oxidative type 1 muscle and a low expression level in type 2 glycolytic muscle, whereas the opposite holds true for PGC-1β (23, 29). PGC-1α transgenic mice expressing the coactivator specifically in skeletal muscle tissue show a distinct muscular phenotype, with all skeletal muscle becoming more oxidative (29). The effect of enhanced PGC-1β expression in vivo on skeletal muscle phenotype has not been examined (21). Overexpression of PGC-1α or PGC-1β in vitro has shown that both coactivators are able to alter mitochondrial metabolism by increasing the level of proton leak rates in mitochondrial respiration, with PGC-1α being the most potent (52). Even though PGC-1α and PGC-1β share many target genes (27), PGC-1β is unable to compensate for the loss of PGC-1α as seen in PGC-1α knockout mice, who exhibit deficits in oxidative metabolism in multiple tissues and exercise intolerance (26). Here we show that PGC-1β also contributes to the regulation of the skeletal muscle contractile and metabolic phenotype.
Effect of PGC-1α and PGC-1β overexpression on metabolic phenotype.
The present study confirms the stimulating effect of both PGC-1α and PGC-1β expression on oxidative metabolism (21, 29, 52). This is most obviously seen by the increase in CS activity and mRNA expression and is corroborated by the enhanced expression of short-chain acyl-CoA dehydrogenase on day 9 of culture (Figs. 2 and 4). The decrease in CK activity in PGC-1α- and PGC-1β-transduced cultures may also suggest a change toward a slow oxidative phenotype, since the CK activity was found to be lower in slow oxidative soleus muscle than in fast glycolytic EDL muscle of young rats (Table 2). Although this difference diminished with age, the properties of the muscle cell cultures prepared from neonatal myoblasts probably reflect those of younger rather than older rats. CK activity has also been reported to be significantly higher in fast glycolytic than in slow oxidative muscle fibers (59), although others find no difference (9). This discrepancy might be explained by our observation that CK activities in rat skeletal muscles change with age in soleus muscle (Table 2).
Neither PGC-1α nor PGC-1β appears to downregulate glycolytic metabolism as one might have anticipated. Thus only a slight effect was found on the activity of GAPDH, and both coactivators caused an unexpected increase in GAPDH mRNA levels (Figs. 2–4). In adult rats, the activity of GAPDH has been reported to be fourfold higher in glycolytic than in oxidative muscles (30). Similarly, the activity of LDH, which is often taken as a measure of glycolytic metabolic activity (53), was not affected by PGC-1α or PGC-1β overexpression (Fig. 2), although the enzyme activity was higher in EDL than in soleus muscle of young rats (Table 2). Furthermore, the expression of GLUT4 was enhanced by PGC-1α or PGC-1β overexpression (Fig. 4). Interestingly, glycogen accumulation was increased, and the effect appeared specific for PGC-1α overexpression (Fig. 5). This corresponds with the observation that endurance training, which is known to upregulate PGC-1α (45), increases the maximum glycogen content of skeletal muscle (16, 38, 50). The observation that PGC-1β seems to cause a decrease in glycogen accumulation has not been shown before, and the possible involvement of PGC-1β in skeletal muscle glucose metabolism warrants further investigation, since it has only been thoroughly examined in the liver (28). It may be suggested that PGC-1α causes a diversion of glucose metabolism from oxidation to deposition in accordance with the lowered pyruvate oxidation, as reflected by enhanced expression of pyruvate dehydrogenase kinase-4 in C2C12 myotubes overexpressing PGC-1α (56).
GLUT4 expression is known to be enhanced by PGC-1α via coactivation of MEF2C (35); however, regulation of MEF2C by PGC-1α and PGC-1β as seen in this study (Fig. 4) has previously only been reported for PGC-1α (36) and not for PGC-1β. Despite the downregulation of MEF2C, we still observed an increase in GLUT4 expression, in contrast to a previous in vivo finding where overexpression of PGC-1α in skeletal muscle of mice prompted both a downregulation of MEF2C and GLUT4 (36). This apparent discrepancy in the results of PGC-1α overexpression in vivo and in vitro is not easily explained; however, others have found a similar increase in GLUT4 mRNA expression after PGC-1α overexpression in myotubes in vitro (35). Interestingly, PGC-1α expression is regulated by MEF2C as well in a feedforward autoregulatory loop (15). One might speculate that the downregulation of MEF2C expression seen in both PGC-1α and PGC-1β overexpression constitutes a failsafe mechanism to control their runaway expression.
The upregulation of TRB3 mRNA after myotube formation from day 5 to 9 of culture (Fig. 3) suggests a possible role for TRB3 in myogenesis. TRB3 binds to and inhibits the activation of Akt1 and Akt2 (12), which are both implicated in signal transduction events by insulin and insulin-like growth factor I (IGF-I; see Ref. 14). IGF-I has been shown to be involved in muscle fiber growth (39), and the marked upregulation of TRB3 after myofiber formation might be one mechanism by which myogenesis prevents excessive skeletal muscle hypertrophy. PGC-1α and to some extent PGC-1β overexpression in the skeletal muscle cell cultures caused an increase in TRB3 mRNA (Fig. 3) in accordance with the observation that TRB3 upregulation in liver is caused by PGC-1α (22). The observed increase is comparable with the relative increase observed in fasted liver (12).
In several cases, an effect of the GFP-expressing adenovirus alone was noted. This was most pronounced in the case of CK activity but was also observed in LDH activity and GAPDH measurements as well as in MHCemb, MHCperi, MHCIb, and MHCIIA mRNA levels; however, the majority of parameters measured showed no significant difference between nontransduced controls and GFP-transduced controls. Similar effects were observed in skeletal muscle cell cultures transduced with an adenovirus expressing β-galactosidase (data not shown), which has led us to conclude that the effect is not caused by the overexpression of GFP but is rather an effect inherent in the adenovirus expression vector. Most of the changes caused by transduction with the GFP-expressing adenovirus perturbs the skeletal muscle cell culture toward a more glycolytic phenotype, something that has previously been observed to happen upon infection of cells with adenovirus (4, 17).
Effect of PGC-1α and PGC-1β overexpression on contractile phenotype.
Both PGC-1α and PGC-1β seem to be involved in myofiber maturation, since overexpression of either coactivator causes a downregulation of MHCemb and MHCperi (Fig. 3).
Overexpression of both PGC-1α and PGC-1β causes a downregulation of the fast glycolytic-associated MHC isoforms, MHCIIX and MHCIIB, and an upregulation of the slow oxidative-associated MHC isoforms, MHCIb and MHCIIA (Fig. 3). PGC-1β, which has not previously been shown to be involved in skeletal muscle plasticity, increases the levels of the fast glycolytic oxidative-associated MHCIIA mRNA both sooner and to a higher degree than PGC-1α, indicating a difference in MHCIIA regulation by the two coactivators. Interestingly, transgenic mice overexpressing either PPARδ or calcineurin in skeletal muscle have a similar phenotype to mice overexpressing PGC-1α in skeletal muscle (40, 54). We have previously shown that calcineurin signaling is involved in regulation of an oxidative muscle fiber phenotype (18), whereas others have shown involvement of MEF2D and PGC-1α in calcineurin regulation of skeletal muscle phenotype (29). With the regulation of MEF2C expression by PGC-1α and PGC-1β shown here and the knowledge that MEF2 transcription factors act downstream of calcineurin signaling (5), one might speculate that a shift toward more oxidative phenotype in skeletal muscle is carried out by the combined transcriptional regulation of PGC-1α, PGC-1β, PPARδ, and calcineurin signaling, e.g., MEF2 transcription factors.
PGC-1α has been shown to cause a dramatic shift toward a slow oxidative muscle fiber type in transgenic mice when expressed from about embryonic day 14 in skeletal muscle (29, 55). Because PGC-1α in these mice was expressed throughout both primary and secondary myofiber formation, it cannot be ruled out that the observed effects were the result of a role of PGC-1α in myogenesis. The effects seen by PGC-1α and PGC-1β overexpression in the present experiments cannot be purely developmental and may indeed be the result of true adult myofiber plasticity, since the PGC-1α and PGC-1β overexpression in the skeletal muscle cell cultures starts at ∼days 6–7. This corresponds to a time when myotubes have already been formed and some maturation toward myofibers has begun. On the other hand, the gradual downregulation of MHCemb mRNA during culture suggests that the cultures are not yet fully adult in nature.
The myotubes in the skeletal muscle cell cultures exhibited plenty of spontaneous contractions. PGC-1α has been shown to be induced in skeletal muscle by exercise (1, 2), and exercise is known to cause a shift in skeletal muscle fiber phenotype from fast glycolytic to slow oxidative. It can therefore not be excluded that the spontaneous contractions of the muscle cell cultures are required for the myofiber-type switch caused by PGC-1α or PGC-1β overexpression and that the effect might be more pronounced in the PGC-1α-overexpressing cultures, since these qualitatively seemed to spontaneously contract more frequently. However, if the differences in spontaneous contractile activity had a major impact on the metabolic properties of the myotubes, one might also expect the more active myotubes to store less glycogen, since the energy would be required for contraction. This does, however, not seem to be the case, since the more active myotubes, the ones overexpressing PGC-1α, stored more glycogen than the cultures that were less active, indicating that the contribution from spontaneous contractions most likely can be dismissed.
In conclusion, both PGC-1α and PGC-1β were shown to have an effect on skeletal muscle fiber metabolism and fiber type. Both upregulated oxidative metabolism, with no apparent effect on glycolytic metabolism, and both conferred a switch toward a more slow myofiber phenotype. The switch in myofiber phenotype was, however, differently regulated by PGC-1α and PGC-1β, since both caused a decrease in fast glycolytic-associated MHCIIB and MHCIIX, with PGC-1β having a more profound effect on the glycolytic oxidative MHCIIA. The differences between PGC-1α and PGC-1β became more apparent when the glycogen content of the myofibers was measured, since only PGC-1α upregulated cell glycogen content and only PGC-1α was able to upregulate TRB3. All in all, PGC-1α and PGC-1β seem to have both overlapping yet distinctively different effects on the phenotype of skeletal muscle fibers.
The adenoviruses used were a generous gift from Dr. Christopher Newgard (Duke University). This work was supported by the Danish Academy of Technical Sciences, Novo Nordisk, Novo Nordisk Foundation, Rigshospitalet, Hovedstadens Sygehusfællesskab, University of Copenhagen, and Savværksejer Jeppe og Hustru Ovita Juhls Mindefond.
We are grateful for the excellent technical assistance by Tina Kisbye Rasmussen and Christina Bøg Sørensen.
The current address for O. Hartvig Mortensen is Centre of Inflammation and Metabolism, Rigshospitalet (Copenhagen, Denmark).
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.
- Copyright © 2006 by American Physiological Society