Endurance Training in Humans Leads to Fiber Type-Specific Increases in Levels of Peroxisome Proliferator-Activated Receptor-γ Coactivator-1 and Peroxisome Proliferator-Activated Receptor-α in Skeletal Muscle

Seven healthy men, age (mean ± SD) 34 ± 5 years, mass 78 ± 8 kg, maximal oxygen consumption (V˙o_2max) 54 ± 4 ml · kg⁻¹ · min⁻¹, participated in the study, which was approved by the local medical society ethical committee. All participants gave their informed consent and agreed to a muscle biopsy followed by a test to measure aerobic capacity.

V˙o_2max was measured using a Quark B2 metabolic cart (Cosmed, Rome, Italy) while the participants ran on a treadmill (Technogym Runrace, Gambettola, Italy). The participants began running at 7.2 km/h with a 1% gradient. The speed was increased by 1.8 km/h every 3 min until the participant could not maintain the speed. The duration of the tests lasted between 10 and 15 min. Heart rate (Polar) and oxygen consumption were measured continually throughout the test. V˙o_2max was calculated as the highest value averaged over a 30-s period.

Skeletal muscle samples were obtained under local anesthesia (Xylocaine, 1% plain) from the belly of the vastus lateralis muscle using a percutaneous needle biopsy technique (22) modified to include suction. A single incision was made in the skin, and two muscle samples were taken from a single insertion of the biopsy needle. After the first biopsy cut, the needle was rotated and a second cut was performed. One muscle sample was mounted in embedding medium and frozen in isopentane, previously cooled to its freezing point. The other muscle sample was immediately frozen in liquid nitrogen and used for RNA extraction.

The endurance training program comprised three supervised sessions per week, which included two interval sessions separated by a constant intensity session. Each session commenced with a warm-up consisting of ∼10 min of light running followed by stretching. The interval sessions consisted of five to six intervals at an individual intensity corresponding to 70–80% of V˙o_2max. A 1-min recovery was performed between each test at an intensity of 50% of V˙o_2max. The duration of these intervals ranged between 1 and 3 min for each participant. The number of repetitions and their intensity were gradually increased over the training period so that during week 6, the participants were completing 6 × 80% repetitions. For example, during the first week, 5 × 70% repetitions were performed; during the second week, it was 6 × 70%; and during the third week, it was 5 × 75%. This pattern was followed so that during week 6, the participants were completing 6 × 80% repetitions. A 1-min recovery was performed between each test at an intensity of 50% of V˙o_2max. The duration of these intervals ranged between 1 and 3 min for each participant. All intensities were controlled by individual heart rates that were measured during the duration of each session using a Polar heart rate monitor. The constant intensity session consisted of running for 40 min at an intensity of 60% of V˙o_2,max .

Total RNA was extracted, and oligo-dT primed first-strand cDNA was synthesized as reported previously (23). Real-time PCR was performed using a Lightcycler rapid thermal cycler system; a Lightcycler-DNA Fast-Start Master SYBR Green I mix for the quantification of PGC-1, PPAR-β/δ, Tn1-slow, COX4, GLUT4, and CPT-1; and a Lightcycler-DNA Master SYBR Green I mix for the quantification of PPAR-α, PPAR-γ, and β-actin (Roche Diagnostics, Rotkreuz, Switzerland), according to the manufacturers’ instructions. β-Actin was used as a control to account for any variations due to efficiencies of the reverse transcription and PCR and has previously been shown to remain stable after training (24). The primer sequences are presented in Table 1. For confirming the amplification specificity, the PCR product was subjected to a melting curve analysis and following this, agarose gel electrophoresis. PCR products were purified using a QIAGEN purification kit (QIAGEN AG, Basel, Switzerland), and the DNA sequence was confirmed on an ABI Prism 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA).

Immunfluorescence experiments were performed as reported previously (25). The PGC-1 (sc-5816; Santa Cruz Biotechnologies, San Diego, CA) and PPAR-α (Dr. David Bell, The University of Nottingham, U.K.) antibodies, respectively, were diluted 1:200 and 1:500 in PBS containing 0.5% BSA and 0.1% NaN₃ and were visualized separately, using a donkey anti-goat (IgG) and a goat anti-mouse (IgG) antibody, both conjugated with Alexa Fluor 594 (Molecular Probes Europe, Leiden, the Netherlands) at a concentration of 1:500, for 60 min at room temperature. Muscle fiber composition was determined by counting the number of positive fibers stained with antibodies obtained from the Development Studies Hybridoma Bank, University of Iowa, raised against the human slow type I MHC (A4.84) and the fast IIa MHC (N2.261), respectively (25). The muscle fibers that were not stained were deemed to be MHC type IIx fibers. To determine the saturation point of the fluorescence signals, we performed preliminary experiments using serial dilutions of the primary and secondary antibodies. For the PGC-1 antibody, which has previously been used for Western blot analysis (15), SAOS2-GR(+) cells were infected with an adenovirus-expressing human PGC-1 (26) (Fig. 5) and were used to determine incubation conditions and the specificity of the antibody. A negative control for the PPAR-α antibody was performed by incubating the sections with the preimmune serum, which resulted in no signal. Negative controls, using only the secondary antibody, were also performed in parallel and on the same glass slide as the experimental samples. In addition, pre- and posttraining sections were placed on the same slide and analyzed in duplicate. Sections were viewed and photographed using a Zeiss Axiophot I microscope mounted with an Axiocam color CCD camera. The specific fluorescence within each fiber was quantified using the Zeiss KS400 V3.0 program (25). Approximately 310 ± 45 fibers were scanned for each participant.

Group data were presented as mean ± SE. Paired t tests were used to determine any changes in V˙o_2max and gene expression changes after training. A 2 (training status, between factor) × 3 (fiber types, within factor) factor ANOVA was used to compare the influence of training status and fiber type on PGC-1 and PPAR-α protein expression, respectively. When a significant interaction between training status and fiber type was observed, stratified analysis was used to locate the significant differences. To be specific, one-way repeated measures ANOVA followed by contrasts were used to compare the percentage changes in PGC-1 and PPAR-α, respectively, between type I, IIa, and IIx muscle fibers, for the groups pre- and posttraining. The α level of significance for the paired t tests was set at 0.05. The overall α level for the ANOVA was set at 0.05, and the Sharpened Bonferroni method was used to adjust the individual α level to significance of P < 0.0167 when multiple testings were performed in the stratified analysis.

The participant data are presented in Table 2. Endurance training increased V˙o_2max by 10% and type I fibers by 18% (P < 0.05). The increase in type I fibers was accompanied by a decrease in both type IIa and IIx fibers by 12 and 7%, respectively (P < 0.05).

Analysis of mRNA expression indicated that PGC-1 and PPAR-α mRNA levels increased by 2.7- and 2.2-fold (P < 0.01), respectively, whereas PPAR-β/δ and PPAR-γ mRNA did not change after endurance training (P > 0.05; Fig. 1). Tn1-slow, CPT-1, COX4, GLUT4, and mRNA increased by 40, 75, 72, and 59%, respectively (P < 0.01; Fig. 2). β-Actin mRNA did not change with training (pre- versus post-training values; 5.9 ± 0.9 vs. 6.3 ± 0.7 AU; P > 0.05).

Immunofluorescence staining was performed using antibodies that distinguish the different fiber types and with antibodies against PGC-1 and PPAR-α. This enabled the identification of PGC-1 and PPAR-α in the different fiber types. Figure 3 shows a representative immunofluorescence stain of PGC-1 and PPAR-α before and after endurance training. It can be observed that the level of expression of PGC-1, both before and after endurance training, is the highest in type IIa fibers, lower in the type I fibers, and the lowest in the type IIx fibers, whereas the level of PPAR-α was higher in type I than in both the type II fibers (Figs. 3 and 4).

Quantification of the images indicated that before endurance training, PGC-1 protein levels in the type IIa fibers were 2.2- and 6-fold greater than in the I and IIx fibers, respectively (P < 0.001), whereas PGC-1 was 2.5-fold greater in the type I than in the type IIx fibers (P < 0.001). After endurance training, PGC-1 protein content increased by 2.8-fold in the type IIa fibers, an effect that was significantly greater than the 1.5-fold increase measured in both the type I and IIx fibers (P < 0.015; Fig. 4). In contrast, before endurance training, PPAR-α levels were 1.9-fold greater in the type I than in the IIa and IIx fibers (P < 0.001). No differences in the PPAR-α protein contents were observed when comparing the type IIa with the type IIx fibers (P > 0.05). After endurance training, PPAR-α protein content increased by 3.0-fold in the type I fibers, an effect that was significantly greater (P < 0.015) than the 1.5-fold increase in both the type IIa and IIx fibers (P < 0.01).

In the present study, we examined the effect of 6 weeks of endurance training on the skeletal muscle gene expression of PGC-1 and of PPAR-α, -β/δ, and -γ in humans. We also studied, in the different muscle fiber types, the distribution of PGC-1 and PPAR-α proteins before and after the 6-week endurance training period. Two findings of our study are novel. First, we showed that several weeks of whole-body endurance training in humans induces a pronounced increase in skeletal muscle PGC-1 and PPAR-α mRNA and their encoded proteins. Second, we showed that before and after endurance training, PGC-1 protein content is greatest in type IIa, lower in type I, and lowest in type IIx fibers, whereas PPAR-α protein content is greater in the type I than in the type IIa and IIx fibers of the vastus lateralis muscle.

Our observation of an increase in PGC-1 mRNA after 6 weeks of endurance training supports previous observations made in rats (15,17) but is in contrast with studies performed in humans (18,19). These conflicting results are most likely explained by the differences in the training duration and intensity. Our results suggest that 6 weeks of moderate- to high-intensity running endurance training is required to stimulate an increase in PGC-1 mRNA. More important, we showed for the first time that this same training program can induce increases in the PGC-1 protein content in human skeletal muscle. The increase in PGC-1, in the vastus lateralis muscle of our participants after endurance training, was paralleled by a shift in fiber composition toward a more oxidative type I muscle with an increase in the percentage of type I fibers and a decrease in the percentage of type IIa and IIx fibers. This was supported by an increase in the mRNA expression of Tn1-slow, a gene expressed mostly in slow oxidative muscle fibers (27) and suggested to be regulated by PGC-1 (6). The greater percentage of type I fibers measured in the vastus lateralis muscle after endurance training is in support of previous observations made in endurance-trained humans (28,29). Our observation of an increase in PGC-1, in association with an increase in the percentage of type I muscle fibers and Tn1-slow mRNA, supports previous findings of Lin et al. (6), who showed that overexpression of PGC-1 in the plantaris, a muscle consisting of predominantly type II fibers, up to the level found in the soleus, a muscle consisting of predominantly type I fibers, induced a type II to type I fiber conversion as well as an induction of Tn1-slow mRNA. These findings support the idea that PGC-1 possibly controls the slow oxidative type I muscle phenotype (6). Our findings also suggest that a naturally caused induction of endogenous PGC-1 activity may lead to similar phenotypes as experimentally raised levels of PGC-1, which may be important not only for increasing the content of type I fibers but also for improving insulin sensitivity.

The increase in PPAR-α mRNA and protein content after the 6 weeks of endurance training performed in the present study is in contrast with the results of Tunstall et al. (19), who observed no change in PPAR-α mRNA in humans after 9 days of endurance training, but in agreement with those of Horowitz et al. (20), who observed a twofold increase in PPAR-α protein content after 12 weeks of endurance training. Again, these differences may be due to the duration and intensity of the endurance training performed. The metabolic stress caused by a short training duration of 9 days at an intensity of 63% of V˙o_2max may not be sufficient to evoke an increase in PPAR-α gene transcription (19) and, one can speculate, on PPAR-α protein translation. The pretraining PPAR-α levels may be adequate to increase the transcription of genes involved in fatty acid oxidation, such as CPT-1 and FA translocase (FAT/CD36), and therefore assist, in part, with the increased fat oxidation that was observed after 9 days of training (19). Our results suggest that if a longer training period of at least 6 weeks at intensities between 60 and 80% of V˙o_2max is performed, then an increase in PPAR-α mRNA and protein levels is necessary to meet these requirements. Indeed, this is supported by the 73% increase in CPT-1 mRNA observed in the present study.

Unlike PPAR-α mRNA, PPAR-β/δ and PPAR-γ mRNA were not changed after endurance training. This is the first time that PPAR-β/δ gene expression has been measured in human skeletal muscle after endurance training. In mouse skeletal muscle, PPAR-β/δ gene expression is severalfold higher than PPAR-α and PPAR-γ (9). In both human and rodent myocytes (9) and in mouse white adipose tissue (12), the activation of PPAR-β/δ increases fatty acid oxidation and induces the expression of lipid regulatory genes. It is possible that the basal level of PPAR-β/δ in those who participated in the present study may have been sufficient to regulate the genes that assist in the increase in fat oxidation commonly observed after endurance training (20,30,31). In addition, even though an increase in PPAR-β/δ mRNA was not observed, its activity would presumably be higher as a consequence, of among others, the increase in PGC-1. In contrast with our results, a small but significant 20% decrease in PPAR-γ mRNA was reported in the study by Tunstall et al. (19) after 9 days of endurance training. However, whether the 20% decrease in PPAR-γ mRNA was paralleled by a decrease in the PPAR-γ protein content was not determined. Again, the discrepancy between these studies may be due to the different training protocols used. The downregulation of PPAR-γ mRNA may be an early but transient response to training. It is possible that after prolonged endurance training, PPAR-γ mRNA returns to its pretraining levels.

Skeletal myofibrils differ greatly with respect to their contractile and metabolic capabilities. In humans, the type I fibers are highly oxidative with a slow contraction speed, the type IIa fibers are considered an oxidative-glycolytic fiber with intermediate contraction speed, and the IIx fibers demonstrate a highly glycolytic metabolism and are the most rapidly contracting fibers (21). This study is the first to report the distribution of PGC-1 and PPAR-α proteins in human muscle fiber types. PGC-1 protein levels were the greatest in the type IIa fibers, lower in the type I fibers, and the lowest in the type IIx fibers. As type I fibers have the greatest mitochondrial content, it could have been expected that the content of PGC-1, a coactivator of mitochondria biogenesis, would be the highest in these fibers. It has been suggested that the responsiveness to stimuli of muscle fibers to alter their phenotype would follow as such, type I ↔ type IIa ↔ type IIx (32). However, in light of the ability of PGC-1 to induce mitochondrial biogenesis (4), the greater PGC-1 content in the type IIa fibers observed in the present study may entail part of the mechanism, that when properly activated, promotes mitochondrial biogenesis and the switch of muscles to a type I phenotype. This notion would be in agreement with the findings of Howald et al. (33), who observed, after 6 weeks of endurance training in humans, an increase in mitochondrial density in all three fiber types, with the greatest increase occurring in the type IIa fibers, as compared with the type I and IIx muscle fibers. Our observation of an increase in the gene expression of the mitochondrial marker COX4 supports previous observations made after endurance training (34,35) and suggests that the increase in PGC-1 in the present study resulted in an increased mitochondrial biogenesis. In addition, the high levels of PGC-1 in the type IIa fibers may be indicative that PGC-1 functions in other aspects of muscle metabolism. PGC-1 overexpression in C₂C₁₂ myoblasts increases both basal and insulin-stimulated glucose uptake (5) and has been shown to regulate gene expression of GLUT4, a glucose transporter (5,15). Because obesity and diabetes are associated with a decrease in the population of muscle type I fibers and this change of phenotype has been postulated to play a role in obesity/diabetes-related insulin resistance (36), targeting PGC-1 in obese and diabetic patients may increase their type I fiber phenotype and GLUT4 levels, with a concomitant improvement in insulin sensitivity. Indeed, endurance training has previously been shown to increase the mRNA levels of GLUT4 in both healthy (35,37) and diabetic patients (37), an observation supported by the 59% increase in GLUT4 mRNA in the present study.

PPAR-α content was greater in the type I as compared with the IIa and IIx fibers. Because PGC-1 acts as a coactivator of PPAR-α, the simultaneous increase in PGC-1 and PPAR-α levels in type I fibers would be expected to lead to a robust increase in PPAR-α activity per se. As PPAR-α is associated with the control of genes involved in fatty acid metabolism, this would be, among others, a mechanism contributing to the greater fatty acid oxidative capacity of these fibers (38). In addition, the increase in PPAR-α content in all fiber types would play a role in the improved oxidative capacity after endurance training (39). An increase in PPAR-α activity has been shown to improve muscle insulin responsiveness (40). Therefore, our observed increase in PPAR-α in response to endurance training may be another factor contributing to the improved skeletal muscle insulin sensitivity observed after endurance training (41,42).

In summary, the present results show that in human skeletal muscle, endurance training induces an increase in both PGC-1 and PPAR-α mRNA as well as in PGC-1 and PPAR-α protein contents in the type I, IIa, and IIx fiber types. PGC-1 is expressed the most in type IIa as compared with type I and IIx fibers, whereas PPAR-α is expressed the most in type I as compared with type IIa and IIx fibers. The fiber type differences in PGC-1 and PPAR-α remain after endurance training. The increase in PGC-1 was also associated with a shift toward an increased type I muscle fiber phenotype, which supports the role, previously observed in transgenic mice, that PGC-1 acts as a co-activator of oxidative type I muscle fibers in humans. In addition, endurance training increased the gene expression of target genes of PGC-1, including troponin-1, COX4, and GLUT4, and of PPAR-α, such as CPT-1. Combined, our results suggest that the increases in muscle mitochondria content, type I muscle fiber phenotype, oxidative capacity, and insulin sensitivity, induced by endurance training, might be mediated by increases in PGC-1 and PPAR-α and their nuclear encoded target genes.

A.P.R. was supported by grants from the Office Féderal du Sport Macolin, Fonds Eugène Rapin, the Fondation du Centenaire de la Société Suisse d’Assurances générales sur la vie humaine pour la santé publique et les recherches médicales. This work was also supported by the Swiss National Science Foundation Grant 31-54306.98 and the Fondation Suisse de Recherche sur les Maladies Musculaires.

We thank the volunteers for commitment and enthusiasm, Ingrid Tofor and Chantal Ansiaux for assistance with the muscle biopsies, and Patrick Crettenand for expertise in supervising all of the training sessions.