A Role for the Transcriptional Coactivator PGC-1α in Muscle Refueling*

The transcriptional coactivator peroxisome proliferator-activated receptor γ coactivator-1α (PGC-1α) has been identified as an inducible regulator of mitochondrial function. Skeletal muscle PGC-1α expression is induced post-exercise. Therefore, we sought to determine its role in the regulation of muscle fuel metabolism. Studies were performed using conditional, muscle-specific, PGC-1α gain-of-function and constitutive, generalized, loss-of-function mice. Forced expression of PGC-1α increased muscle glucose uptake concomitant with augmentation of glycogen stores, a metabolic response similar to post-exercise recovery. Induction of muscle PGC-1α expression prevented muscle glycogen depletion during exercise. Conversely, PGC-1α-deficient animals exhibited reduced rates of muscle glycogen repletion post-exercise. PGC-1α was shown to increase muscle glycogen stores via several mechanisms including stimulation of glucose import, suppression of glycolytic flux, and by down-regulation of the expression of glycogen phosphorylase and its activating kinase, phosphorylase kinase α. These findings identify PGC-1α as a critical regulator of skeletal muscle fuel stores.

Glucose and fatty acids are the chief fuel sources for skeletal muscle. During prolonged bouts of low intensity exercise, muscle energy needs are met through utilization of both substrates with mitochondrial fatty acid oxidation serving a "glucose sparing" function (1,2). During acute high intensity exercise, glucose derived from hepatic and muscle glycogen stores serves as the chief energy source (reviewed in Refs. [3][4][5]. Rapid glycogen repletion following a bout of exhausting intense exercise is an important adaptive response, preparing the muscle for subsequent bouts of activity. With endurance exercise training, the capacity for mitochondrial oxidation of fatty acids is augmented and muscle glycogen reserves increase (2). In disease states such as diabetes and heart failure, the capacity for muscle energy substrate utilization is reduced due to alterations in glucose metabolism and derangements in mitochondrial function (6, 7) (reviewed in Ref. 8).
Glucose, stored as glycogen, serves as a key muscle energy substrate, especially during periods of high intensity activity. Whereas the role of PGC-1␣ in the control of muscle fatty acid oxidation and mitochondrial respiratory capacity is well established, its contribution to the regulation of skeletal muscle glucose metabolism has not been well defined. Several recent lines of evidence suggest that PGC-1␣ exerts control on muscle glucose metabolism. First, PGC-1␣ serves as a key regulator of hepatic gluconeogenesis, an important source of substrate for muscle (32)(33)(34). Second, PGC-1␣ has been shown to activate transcription of the GLUT4 gene in myogenic cells in culture (35). Third, PGC-1␣ activates expression of the gene encoding pyruvate dehydrogenase kinase 4 (PDK4), a negative regulator of glucose oxidation (22).
In the current study, we sought to clarify the role of PGC-1␣ in the regulation of muscle fuel metabolism in vivo. To this end, we developed an inducible, muscle-specific PGC-1␣ transgenic system to be used in comparison with normal and PGC-1␣deficient mice (31). We found that muscle-specific activation of PGC-1␣ increases muscle glucose uptake and glycogen levels at baseline and prevented depletion of glycogen stores following exercise. This glycogen "sparing" effect involved several mechanisms including increased capacity for fatty acid oxidation, induction of glucose transporter (GLUT1 and GLUT4) expression leading to augmentation of muscle glucose import, inhibition of glycolysis, and suppression of pathways involved in glycogen degradation.

EXPERIMENTAL PROCEDURES
Animal Studies-All animal experiments and euthanasia protocols were conducted in accordance with the National Institutes of Health guidelines for humane treatment of animals and were reviewed and approved by the Animal Studies Committee of Washington University. Mice doubly transgenic for PGC-1␣ under control of the tetracycline response element promoter (TRE-PGC-1␣) (30), and for the tetracycline transactivator under control of the muscle-specific creatine kinase promoter (MCK-tTA) (36) were generated. Breeding pairs and offspring were maintained on chow containing doxycycline (200 mg/kg; Research Diet Inc., Brunswick, NJ). To activate the TRE-PGC-1␣ transgene, doxy-chow was removed and replaced with standard chow (Lab Diet 5052; Purina Mills Inc., St. Louis, MO). Unless otherwise stated, transgene induction was initiated at ϳ5-6 weeks of age and experiments were performed 3-4 weeks later. For all studies, age-and sex-matched littermates were used. The generation and initial characterization of PGC-1␣-deficient (PGC-1␣ Ϫ/Ϫ ) mice has been described (31). Approximately 2-3-month-old PGC-1␣ Ϫ/Ϫ mice were used for all studies with age-and sex-matched wild-type controls unless specified otherwise.
Low Intensity Exercise-To assess endurance exercise capacity and glycogen depletion, male and female mice were run to exhaustion on an Exer4-OxyMax motorized treadmill (Columbus Instruments, Columbus, OH). Briefly, mice were acclimated to the treadmill for 2 days prior to the experimental protocol by running for 9 min at 10 m/min followed by 1 min at 20 m/min. On the day of the experiment, mice were run for 1 h at 10 m/min followed by an increase in speed of 2 m/min each additional 15 min until failure. Mice were defined as exhausted if they remained on the shock grid for five continuous seconds.
High Intensity Exercise-Mice were acclimated to the treadmill as above. On the day of the experiment, mice were run alternating 1 min of running with 2 min of rest. The running intervals started at 10 m/min and increased 5 m/min each interval until a speed of 50 m/min was reached. Thereafter, speed was further increased 5 m/min every 6th interval until failure. Exhaustion was defined as above.
For either exercise protocol, tail blood was taken prior to exercise and immediately following failure for glucose (HemoCue AB, Á ngelholm, Sweden) and lactate (Lactate Pro Arkray, Kyoto, Japan) measurements, as per the manufacturers' instructions. To prevent blood loss, wounds were cauterized following sample collection. For glycogen replenishment studies, mice were gavaged with a glucose solution (0.56 M, 1 mg of glucose/g of mouse) immediately following completion of the exercise protocol and euthanized by CO 2 inhalation at varying time points after cessation of exercise. The vastus muscles were harvested and immediately clamp-frozen in liquid nitrogen for later glycogen content, RNA, and protein analyses.
Histology and Microscopy-Gastrocnemius was analyzed histologically. For light microscopy, frozen tissue was used for succinate dehydrogenase staining. Formalin-fixed samples were embedded in paraffin, sectioned, and stained with hematoxylin/eosin (H & E) or periodic acid Schiff. For electron microscopy, tissue was fixed in 2% glutaraldehyde and 1% paraformaldehyde, post-fixed in 1% osmium tetroxide, embedded in plastic epoxy resin (poly bed 812). Sections on copper grids were then stained with uranyl acetate and lead citrate for visualization.
RNA and Protein Expression Analyses-Total cellular RNA was isolated from gastrocnemius using the RNAzol method (Tel-Test, Friendswood, TX). Northern blot hybridizations with random-primed 32 P-labeled cDNA probes were performed using QuikHyb (Stratagene). Real-time quantitative reverse transcription-PCR (RT-PCR) was performed as previously described (37) and results were normalized to the expression of 36B4. Mouse-specific primers used for RNA analysis may be found in supplemental materials Table S1.
Isolated Mitochondrial Respiration-Mitochondria were isolated from whole hindlimb using a trypsin digestion procedure as previously described (44). Briefly, tissue was minced, washed, and suspended in isolation medium (300 mM sucrose, 10 mM NaHepes, 0.2 mM EDTA, 1 mg/ml bovine serum albumin, pH 7.4). Following digestion, samples were gently homogenized with a glass Teflon homogenizer. Following centrifugation and two washes, isolated mitochondria were incubated with palmitate-containing media and respiration was determined as previously described (31). Oxygen consumption was measured at 25°C using an optical probe (Oxygen FOXY Probe; Ocean Optics, Dunedin, FL). Following basal state measurements, maximal (ADP-stimu-lated) state 3 respiration was determined by adding ADP to a 1 mM final concentration. Thereafter, rates of uncoupled respiration was determined indirectly following addition of oligomycin (1 g/ml). The solubility of oxygen at 25°C was taken as 246.87 mmol of O 2 ml Ϫ1 . Respiration rates are expressed as nanomole of O 2 min Ϫ1 mg protein Ϫ1 as determined by the Pierce Micro BCA protein assay.
Citrate Synthase Activity Assay-Citrate synthase activity was determined in tibialis anterior muscle as described (45).
Measurement of Glucose Transport Activity-Animals were anesthetized with an intraperitoneal injection of sodium phenobarbital (5 mg/100 g body weight). Epitrochlearis muscle (and EDL, data not shown) was isolated following an overnight fast. Muscles were immediately washed with Krebs-Henseleit bicarbonate (KHB) and incubated for 20 min in 1.0 ml of KHB buffer containing 1 M 2-deoxy-D-[1,2-3 H]glucose and 0.1% bovine serum albumin and the rate of accumulation of 2-deoxyglucose (2-DG) in intracellular water was determined as described (46). Values were corrected to tissue weight.
Glucose 6-Phosphate Assay-The concentration of Glu-6-P was determined as described (47) by assaying gastrocnemius tissue extract with Glu-6-P dehydrogenase (G8289, Sigma) and 0.2 mM NADP in 50 mM Tris (pH 7.4) and 3 mM MgCl 2 . Resulting changes in absorption at 340 nm were compared with a standard of 0 -33 nmol of Glu-6-P and corrected to protein concentration.
Isolated Muscle Glycolysis Studies-Glycolysis was assayed in epitrochlearis muscle isolated from mice following an overnight fast using a modified protocol from isolated working hearts (48). Briefly, isolated muscle was immediately placed in KHB buffer with 8 mM glucose and 32 mM mannitol for 1 h at 37°C for recovery. Muscle was then transferred to fresh KHB as above, supplemented with 125 M oleate for substrate selection and preincubated for 30 min at 37°C. Finally, muscle was transferred to fresh KHB as above, spiked with 1 Ci/ml [5-3 H]glucose (15 Ci/mmol), and incubated for 1 h at 37°C. The release of 3 H 2 O from glucose into incubation buffer was measured by evaporative 3 H 2 O transfer from the incubation buffer to water in a sealed scintillation vial. Tritium was measured by scintillation counting and disintegrations/min were normalized to control values corrected for tissue weight.
Glycogen Measurements-Muscle extract from either vastus or gastrocnemius (data not shown) was analyzed as previously described (49). Briefly, snap frozen tissue was powdered under liquid nitrogen and homogenized in a 0.3 M perchloric acid solution. This extract was assayed with and without amyloglucosidase digestion (A7420, Sigma) in 50 mM Na acetate (pH 5.5) and 0.02% bovine serum albumin. Resulting changes in absorption at 340 nM were compared with a standard of 0 to 80 mol of glucose. Results are presented as glucose released from glycogen corrected to tissue weight.
Statistical Analysis-Statistical comparisons were made using unpaired t test or analysis of variance coupled to Tukey's post hoc test when appropriate. All data are presented as mean Ϯ S.E., with a statistically significant difference defined as a value of p Ͻ 0.05.

Short-term, Muscle-specific Expression of PGC-1␣ Activates Mitochondrial Biogenesis and Oxidative Gene Regulatory
Programs-Double transgenic mice (PGC-1␣-TRE(ϩ) mice) were generated by breeding MCK-tTA ("tet-off" transactivator) and TRE-PGC-1␣ (transresponder) lines so that the short-term effects of muscle-specific overexpression of PGC-1␣ could be investigated. Single transgenic transresponder (PGC-1␣-TRE(Ϫ)) littermates receiving doxycycline were used as controls. PGC-1␣-TRE(ϩ) mice were maintained on doxycyclinecontaining chow until 5-6 weeks of age, after which doxycycline was removed. The transgene was not expressed in the presence of doxycycline (Fig. 1A). Transgene expression was detectable in muscle after a 2-week doxy "wash-out" period and continued to increase until 3-4 weeks after doxycycline removal (Fig. 1A). Examination of various tissues revealed that transgene expression was specific to skeletal muscle ( Fig. 1B and data not shown). As described previously for constitutive skeletal muscle PGC-1␣ transgenic mice (28), forced expression of PGC-1␣ increased red coloration of the muscle within 3 weeks of the induction of transgene expression (Fig. 1C).
Histological examination revealed enhanced succinate dehydrogenase staining in gastrocnemius, consistent with an increase in proportion of oxidative fibers in this muscle ( Fig.  2A). To further examine the cellular and structural changes accompanying PGC-1␣ overexpression in skeletal muscle, electron microscopic analysis was performed. Forced expression of PGC-1␣ resulted in a dramatic increase in the cellular volume density of mitochondria ( Fig. 2A). Importantly, muscle from PGC-1␣-TRE(ϩ) animals exhibited lipid droplets associated with mitochondria and granular staining resembling glycogen ␤-particles suggesting an increase in cellular fuel stores ( Fig. 2A).
Respiratory function studies were performed on mitochondria isolated from whole hindlimb of PGC-1␣-TRE(ϩ) and PGC-1␣-TRE(Ϫ) mice. Using palmitoylcarnitine as substrate, state 3 (ADP-stimulated) respiration rates were significantly higher in mitochondria isolated from the PGC-1␣-TRE(ϩ) mice (Fig. 2B). In the presence of oligomycin, respiration was largely abolished in both PGC-1␣-TRE(ϩ) and PGC-1␣-TRE(Ϫ) mitochondria, suggesting that the majority of the PGC-1␣-stimulated respiration was coupled to ATP production. In contrast to the results obtained with palmitoylcarnitine, respiration rates with pyruvate were not different between the groups (data not shown). The activity of the citric acid cycle enzyme, citrate synthase (Fig. 2B), and the expression of nuclear and mitochondrial genes encoding enzymes involved in mitochondrial fatty acid ␤-oxidation, electron transport, and oxidative phosphorylation were coordinately increased in the muscle of the PGC-1␣-TRE(ϩ) mice (Fig. 2C). Expression of the cellular fatty acid transporter, CD36, was also induced in the muscle of PGC-1␣-TRE(ϩ) mice (Fig. 2C). Taken together, these results demonstrate that PGC-1␣ augments the capacity for muscle mitochondrial fatty acid oxidation.
To determine the fate of the glucose transported into PGC-1␣-TRE(ϩ) muscle, glycolysis rates were measured in isolated skeletal muscle by following the release of 3 H 2 O from [5-3 H]glucose. Surprisingly, PGC-1␣ overexpression led to a marked reduction (over 50%) in glycolytic flux (Fig. 3A). The decreased glycolytic rate was associated with a modest but significant reduction in the expression of the gene encoding phosphofructokinase, which catalyzes a tightly regulated, rate-limiting step in the glycolytic pathway (Fig. 3B). Recently we demonstrated that PGC-1␣ exerts repression on glucose oxidation by increasing expression of the gene encoding PDK4, a negative regulator of the pyruvate dehydrogenase complex (22). Taken together with the data shown here, these results indicate that PGC-1␣ suppresses glycolysis and glucose oxidation.
The observed increased rates of muscle glucose uptake combined with decreased glycolytic flux in the PGC-1␣-TRE(ϩ) mice suggested that the imported glucose was diverted to storage. Consistent with this possibility, levels of Glu-6-P were increased (Fig. 3A), which is known to increase glycogen synthesis rates. Muscle glycogen levels were assessed by periodic acid Schiff staining. Periodic acid Schiff staining was markedly increased in PGC-1␣-TRE(ϩ) muscle compared with controls (Fig. 3C). Enzymatic quantification of muscle extracts revealed that glycogen levels were 2.5-fold higher in PGC-1␣-TRE(ϩ) muscle compared with control muscle (Fig. 3C). Collectively, these results support that forced expression of PGC-1␣ increases and maintains muscle glycogen stores through the combined effects of increased glucose import and down-regulation of glycolytic flux.   DECEMBER 14, 2007 • VOLUME 282 • NUMBER 50

Forced Expression of PGC-1␣ Prevents Depletion of Muscle
Glycogen during Exercise-Glucose derived from muscle and hepatic glycogen is a critical source of fuel for exercising muscle. Following exercise, glycogen stores are rapidly replenished (2,5). Recently, we have shown that PGC-1␣ levels are induced in skeletal muscle immediately post-exercise (22). To assess the physiological relevance of the PGC-1␣-mediated glycogenic response, muscle glycogen levels were measured before and after exhausting exercise. For the first series of experiments, muscle PGC-1␣ expression was induced in PGC-1␣-TRE(ϩ) mice 3-4 weeks prior to low intensity exhausting treadmill running. Baseline muscle glycogen levels were significantly increased following induction of PGC-1␣ expression in the PGC-1␣-TRE(ϩ) mice (Fig. 4A). Exercise performance was not significantly different in the two groups (Fig. 4A). Following exercise, muscle glycogen levels were markedly reduced in PGC-1␣-TRE(Ϫ) mice, but did not fall below baseline levels in the PGC-1␣-TRE(ϩ) mice (Fig. 4A). Liver glycogen was significantly depleted in both groups but to a lesser extent in transgenic animals (data not shown).
We next sought to assess the effects of a high intensity exercise regimen that is known to markedly increase reliance of skeletal muscle on glycogenolysis and glycolysis (reviewed in Ref. 50). Given that glycolytic flux is constrained in PGC-1␣-TRE(ϩ) mice, it was predicted that exercise performance would be impaired in the PGC-1␣-TRE(ϩ) animals. As with the low intensity regimen, high intensity exhausting exercise depleted muscle glycogen in PGC-1␣-TRE(Ϫ) mice but not PGC-1␣-TRE(ϩ) mice (Fig. 4B). In addition, blood lactate levels were significantly lower at baseline and following exercise in the PGC-1␣-overexpressing mice compared with controls, consistent with reduced glycolytic flux (Fig. 4B). Importantly, the high intensity exercise protocol unveiled a performance deficit for PGC-1␣-TRE(ϩ) mice compared with control animals (Fig. 4B). These results provide further evidence that PGC-1␣ simultaneously inhibits glycolysis and maintains glycogen stores. Whereas this function is likely adaptive during the post-exercise period, forced expression of PGC-1␣ preceding and during intense exercise likely diminishes performance by imposing constraints on muscle glycogen degradation and glucose utilization.
PGC-1␣ Is Required for Normal Muscle Glycogen Replenishment Post-exercise-To determine whether PGC-1␣ is required for normal glycogen homeostasis in skeletal muscle, we utilized PGC-1␣-deficient (PGC-1␣ Ϫ/Ϫ ) mice (31). For these experiments, muscle glycogen levels were assessed at baseline and following low intensity exhausting exercise in PGC-1␣ Ϫ/Ϫ and   wild-type control animals. Basal glycogen levels were not different in the two groups indicating that PGC-1␣ is not required for basal muscle glycogen synthesis (Fig. 5B). We next sought to explore the effects of exercise. As we have shown previously (31), PGC-1␣ Ϫ/Ϫ mice exhibited reduced exercise capacity (Fig. 5A). Despite the decreased performance, muscle glycogen levels were depleted to undetectable levels in both control and PGC-1␣ Ϫ/Ϫ animals (Fig. 5B). Importantly, the rate of postexercise replenishment of muscle glycogen in the PGC-1␣ Ϫ/Ϫ animals lagged significantly behind wild-type animals (Fig. 5B). These findings are consistent with a requirement for PGC-1␣ in the rapid re-accumulation of glycogen post-exercise.
The observed reduction in skeletal muscle glycogen replenishment rates in the PGC-1␣ Ϫ/Ϫ mice could be due to impaired glucose delivery related to a reduction in circulating glucose levels. To investigate this possibility, blood glucose levels were measured before and at the end of exercise in the PGC-1␣ Ϫ/Ϫ mice and wild-type controls. There was no significant difference in mean baseline blood glucose levels between the groups (Fig. 5C). Post-exercise, blood glucose levels were actually higher in the PGC-1␣ Ϫ/Ϫ mice. Muscle GLUT4 levels were also assessed in the PGC-1␣ Ϫ/Ϫ mice. As predicted by the studies in the PGC-1␣ Ϫ/Ϫ -TRE(ϩ) mice, GLUT4 protein levels were decreased in the muscle of PGC-1␣ Ϫ/Ϫ mice (Fig. 5D). These data suggest that altered capacity for muscle glucose uptake contributes to the defect in post-exercise glycogen replenishment in the PGC-1␣ Ϫ/Ϫ mice.
PGC-1␣ Inhibits the Glycogen Degradation Pathway-Glycogen levels are determined by the input rate of glucose substrate and the balance between glycogen biosynthesis and degradation. The activities of the glycogen synthesis and degradation pathways are precisely controlled by insulin and other upstream signaling pathways (reviewed in Ref. 51). We hypothesized that, in addition to increasing substrate delivery, the effect of PGC-1␣ on maintenance of steady-state glycogen levels involved regulation of synthesis or degradation pathways. We focused on the two enzymes controlling glycogen turnover, glycogen synthase (GS; synthesis) and glycogen phosphorylase (GPh; degradation). Regulation of the activity of these enzymes occurs at the level of expression and through post-translational control via inhibitory (GS) or activating (GPh) phosphorylation, respectively. Phosphorylation of GS in the C-terminal region by glycogen synthase kinases (GSKs) is known to decrease its activity (52). Total GS levels and phosphorylation of its C-terminal serine residues were not significantly different in PGC-1␣-TRE(ϩ) and PGC-1␣-TRE(Ϫ) mice (Fig. 6A). The expression and phosphorylation status of GSK-3␣ and -␤ were also unchanged (data not shown). In contrast, total GPh levels were modestly decreased and phospho-GPh (pGPh) levels were  DECEMBER 14, 2007 • VOLUME 282 • NUMBER 50 markedly decreased such that the pGPh/total GPh ratio was reduced to ϳ50% of control values in the PGC-1␣-TRE(ϩ) muscle (Fig. 6A). These changes would be predicted to reduce steady-state activity of glycogen phosphorylase, particularly during exercise when phosphorylation of GPh by phosphorylase kinase (PhK) is induced. Consistent with the decreased ratio of pGPh/GPh, protein levels of PhK␣ were markedly reduced in the PGC-1␣-TRE(ϩ) muscle (Fig. 6A). PhK-mediated phosphorylation of GPh is triggered by ␤-adrenergic stimulation consistent with the observation that muscle glycogen levels are maintained in PGC-1␣-TRE(ϩ) mice during exercise. Finally, levels of mRNA encoding both GPh and PhK␣ were significantly reduced indicating that the down-regulation of this pathway by PGC-1␣ occurs, at least in part, at the level of gene expression (Fig. 6B). These results identify an additional mechanism whereby PGC-1␣ serves to maintain muscle glycogen levels, via inhibition of glycogen degradation.

DISCUSSION
To characterize the role of PGC-1␣ in the dynamic control of muscle energy metabolism in vivo, we generated an inducible, skeletal muscle-specific murine PGC-1␣ transgenic system. As expected, we found that short-term induction of PGC-1␣ expression drives muscle toward a more oxidative phenotype. Importantly, our gain-of-function and loss-of-function studies revealed that PGC-1␣ exerts dramatic effects on muscle glucose metabolism. Previous studies have demonstrated that PGC-1␣ regulates hepatic glucose metabolism by inducing expression of gluconeogenic genes (32)(33)(34)53). Gluconeogenesis is an important source of glucose for skeletal muscle, especially during periods of exertional activity (54). Several lines of evidence shown here support the conclusion that PGC-1␣ also exerts regulatory action directly on muscle to increase glucose availability by augmenting glycogen stores post-exercise through stimulation of glucose import and via "glycogen sparing" effects ( Fig. 7). First, we found that muscle glucose uptake is robustly increased in PGC-1␣-TRE(ϩ) muscle. These results are consistent with the observation that PGC-1␣ activates expression of the GLUT4 gene in cultured muscle cells (35). Forced expression of PGC-1␣ triggers a significant increase in GLUT4, GLUT1, and hexokinase levels. Second, in the context of increased glucose uptake, PGC-1␣ overexpression results in suppression of muscle glycolytic rates as evidenced by direct measurements of glycolytic flux and determination of lactate levels. In addition, we have recently shown that PGC-1␣ inhibits glucose oxidation by increasing PDK4 gene expression (22). Suppression of glycolysis and glucose oxidation set the stage for diversion of imported glucose to storage (Fig. 7). Third, our results indicate that PGC-1␣ exerts regulatory effects on glycogen synthesis. Glu-6-P, an important precursor for glycogen synthesis accumulates in PGC-1␣-TRE(ϩ) muscle. Through allosteric effects, increased levels of Glu-6-P activate GS, a key enzyme in the glycogen synthesis pathway (51,55). Last, PGC-1␣ was shown to reduce levels and phosphorylation of GPh, the key enzyme responsible for glycogen degradation. During the preparation of this article, Mortensen et al. (56) showed that adenoviral overexpression of PGC-1␣ increased GLUT4 expression and glycogen content in cell culture (primary rat myotubes), consistent with the results of the in vivo studies presented here. In addition, recent work by Al-Khalili et al. (57) supports a role for PGC-1␣-mediated GLUT4 induction in enhancing insulin-stimulated glycogen synthesis in primary human myotubes.  . PGC-1␣ regulates muscle glucose metabolism at multiple levels to increase glycogen stores. Schematic representation of the regulatory nodal points involved in the effects of PGC-1␣ on muscle glucose and glycogen metabolism. Following rigorous activity (e.g. exercise) the expression of PGC-1␣ is rapidly induced in skeletal muscle. As denoted by the arrows, PGC-1␣ increases (green) or decreases (red) expression of genes involved in glucose metabolism. PGC-1␣ augments steady-state glycogen levels by increasing muscle glucose uptake and Glu-6-P levels, suppressing glycolysis and glucose oxidation, increasing substrate delivery for glycogenesis, and inhibiting glycogenolysis during exercise via down-regulation of PhK-mediated phosphorylation of GPh. Hx, hexokinase; TCA, tricarboxylic acid cycle; PDC, pyruvate dehydrogenase complex; FAO, fatty acid oxidation (other abbreviations defined in text).
It should be noted that hepatic glucose production is an important source of glucose for exercising skeletal muscle. The results of the PGC-1␣-TRE(ϩ) mouse studies demonstrate that PGC-1␣ is capable of increasing muscle glucose uptake and glycogen stores. The results of the PGC-1␣ Ϫ/Ϫ mouse experiments are consistent with this conclusion. However, given that the PGC-1␣ Ϫ/Ϫ mice have a generalized deficiency of PGC-1␣, a contribution of altered hepatic glucose production to the observed reduced rate of muscle glycogen replenishment postexercise cannot be excluded. Indeed, hepatic glucose output from gluconeogenesis has been shown to be reduced in the PGC-1␣ Ϫ/Ϫ mice (58). However, several lines of evidence support the conclusion that muscle PGC-1␣ deficiency contributes to the observed phenotype. First, we found that, as predicted by the overexpression studies, muscle GLUT4 levels are reduced in the PGC-1␣ Ϫ/Ϫ mice. Second, it is known that glycogenolysis serves as the chief source of glucose export from liver during exercise (59,60). Whereas hepatic gluconeogenic capacity is reduced in PGC-1␣ Ϫ/Ϫ mice, glycogen stores and glycogenolysis rates are increased (58). Third, circulating blood glucose levels were found to be increased rather than decreased in PGC-1␣ Ϫ/Ϫ post-exercise. Taken together, we conclude that the muscle glycogen replenishment defect in PGC-1␣ Ϫ/Ϫ mice involves reduced capacity for muscle glucose uptake and glycogen storage, consistent with the gain-of-function results. A potential contribution to this phenotype by defective hepatic glucose production cannot, however, be excluded by our results. Future studies using tissue-specific PGC-1␣ gene deletion mice could shed light on this issue.
It is likely that the effects of PGC-1␣ on muscle glucose metabolism are mediated by both direct and indirect gene regulatory mechanisms. We found that forced expression of PGC-1␣ robustly increased the expression of GLUT1, GLUT4, and hexokinase. Previous studies have shown that PGC-1␣ directly activates GLUT4 gene transcription via the transcription factor MEF-2 (35). Whether this same mechanism accounts for the activation of GLUT1 and hexokinase expression remains unknown. Recently, we have shown that PGC-1␣ increases PDK4 gene transcription by coactivating the orphan nuclear receptor estrogen-related receptor ␣ (22). Whereas the induction of glucose uptake and inhibition of glucose oxidation involve direct effects of PGC-1␣ on known transcription factor targets, the mechanisms involved in suppressing glycolysis and glycogenolysis are less evident. The observed repression of phosphofructokinase is likely involved in the PGC-1␣-mediated inhibition of glycolytic flux, given that this enzyme catalyzes a rate-limiting step in this pathway. We have shown previously that muscle-specific overexpression of PPAR␣, a known partner of PGC-1␣, also leads to down-regulation of phosphofructokinase gene expression (61). However, given that neither PGC-1␣ nor PPAR␣ are known transcriptional repressors, it is difficult to envision how the observed effect on phosphofructokinase is direct. Similarly, the mechanisms involved in repression of GPh expression and its activating kinase PhK␣ may also be indirect. Future studies aimed at the mechanisms involved in these repressive effects will be informative.
Endurance training has been shown to increase muscle fiber mitochondrial content and boost capacity for muscle fatty acid oxidation (1,2,62), adaptations that are mimicked by forced expression of PGC-1␣ (9,10,28,63). Endurance exercise training also affects muscle glucose metabolism in the post-exercise state, including an increase in muscle glycogen stores and a shift in resting energy substrate utilization from glycolysis to fatty acid oxidation. The results described here indicate that PGC-1␣ is necessary and sufficient for rapid re-fueling of muscle post-exercise. PGC-1␣ increases expression of GLUT1, GLUT4, and hexokinase, each of which has been shown to be activated by exercise (64 -68). Increased levels of PGC-1␣ leads to reduced phosphorylation of GPh, a critical determinant of glycogen breakdown during periods of increased adrenergic drive, such as occurs with exercise. Consistent with this finding, glycogen levels were maintained during exercise in PGC-1␣-TRE(ϩ) mice. Increased glucose uptake coupled with suppression of glycolysis and diminished glycogen breakdown mediated by PGC-1␣ leads to an increase in the glycogen fuel depot in a manner similar to that of endurance training. Additionally, we also observed increased intracellular lipid droplets in the PGC-1␣-TRE(ϩ) mice, another signature of increased fuel storage related to exercise training. Interestingly, the results of our experiments with PGC-1␣-TRE(ϩ) mice subjected to intense exercise demonstrated that the temporal pattern of the induction of PGC-1␣ post-exercise is critical. PGC-1␣-TRE(ϩ) mice exhibited poor exercise performance associated with reduced lactate production and no change in glycogen levels, consistent with inhibition of glycolysis, the main source of energy during intense exercise. The rapid inducibility of PGC-1␣ allows for precise temporal regulation of this response (14,15,22,27,69).
Skeletal muscle is a major site of whole body glucose utilization. It has been suggested that molecular regulatory events triggered by exercise impart a protective effect against the effects of diabetes, heart failure, and aging (reviewed in Ref. 70). Indeed the development of insulin resistance and diabetes is associated with decreased muscle substrate flexibility (71) and an association between insulin resistance and reduced levels of muscle PGC-1␣ has been shown by several groups (6,72). Whereas the contribution of reduced muscle PGC-1␣ to insulin resistance could be related to altered mitochondrial function or effects on insulin signaling, our results suggest that derangements in glucose and glycogen metabolism are important. Insulin increases muscle glucose uptake and glycogen synthesis, metabolic effects also mediated by PGC-1␣. However, PGC-1␣ suppresses glycolysis and glucose oxidation, pathways that are stimulated by insulin. Thus, it would appear that the effects of PGC-1␣ on muscle glucose disposal occurs, at least in part, independent of insulin signaling by increased expression of glucose transporters and hexokinase. This notion is consistent with the results of previous studies demonstrating that exercise improves glucose tolerance in an insulin-independent manner (65).
In summary, our results unveil an expanded role for the inducible transcriptional coactivator PGC-1␣. In addition to exerting effects on oxidative metabolism in muscle, PGC-1␣ is capable of orchestrating metabolic programs involved in the replenishment of muscle glycogen, a key source of fuel for muscle during periods of intense activity.