Abstract

METHODS

RESULTS

Pgc-1α MKO mice have normal IIb-to-IIa fiber-type transformation in response to voluntary running.

To determine if PGC-1α is essential for endurance exercise-induced skeletal muscle adaptation, we obtained MKO and WT mice by crossbreeding (Fig. 1A). The efficacy of the gene deletion was confirmed by real-time PCR analysis of total RNA from plantaris muscles, showing that MKO mice only had trace amount of Pgc-1α mRNA expression (Fig. 1B). Similar to the findings from previous studies (23, 24, 32), MKO mice showed a trend of decreased expression of Esrra, Ndufs1, and Vegf mRNA with no appreciable changes in Pgc-1β mRNA in skeletal muscle (Fig. 1B). When we subjected MKO and WT mice (male, 8 wk of age) to 4 wk of voluntary running, the daily running distance for each group increased gradually, reaching a steady-state level after 2 wk (Fig. 1C). No significant differences in running distance were observed between MKO and WT mice (Fig. 1D). Both MKO and WT mice had moderate but significant increases in the heart weight (normalized by body weight) following endurance exercise training (Fig. 1E), indicating similar degrees of endurance exercise-induced cardiac hypertrophy. Neither the body weight [24.8 ± 0.5, 23.7 ± 0.9, 24.6 ± 1.4, and 23.2 ± 0.4 g for the WT sedentary (WT-Sed), WT endurance exercise-trained (WT-Ex), MKO-Sed, and MKO-Ex groups, respectively, P > 0.05] nor plantaris muscle weight (0.62 ± 0.01, 0.59 ± 0.02, 0.59 ± 0.02, and 0.61 ± 0.01 mg/g body wt for WT-Sed, WT-Ex, MKO-Sed and MKO-Ex groups, respectively, P > 0.05) showed any significant difference among different groups.

Open in a separate window

Fig. 1.

Voluntary exercise training in muscle-specific peroxisome proliferator-activated receptor γ coactivator-1α gene (Pgc-1α) knockout (MKO) and wild-type (WT) mice. A: gel images of PCR products amplified from genomic DNA from WT (Pgc-1α +/+:Myog-Cre +/+ or Pgc-1α fl/+:Myog-Cre +/+) and MKO (Pgc-1α fl/fl:Myog-Cre Cre/+) mice. B: total RNA from plantaris muscles of WT and MKO mice were assayed for mRNA expression of Pgc-1α, Pgc-1β, estrogen-related receptor α (Esrra), NADH dehydrogenase (ubiquinone) Fe-S protein 1 (Ndufs1), and vascular endothelial growth factor (Vegf) by real-time PCR. 18S ribosomal RNA was used to normalize the expression (n = 6). *P < 0.05, ***P < 0.001. C: representative running activity recording. Six continuous voluntary running activity recordings are shown for WT and MKO mice during the 4 wk of voluntary running. Each of the clusters of running activities (spikes) that correspond to a duration of half-day represents the nocturnal running activity at night. D: daily voluntary running distance in WT and MKO mice during the 4 wk of voluntary running (P = 0.99). Values are means ± SE (n = 7). E: heart weight (normalized by body weight) in sedentary (Sed) and 4-wk endurance exercise-trained (Ex) WT and MKO mice (n = 6–7). *P < 0.05.

Since plantaris muscles are actively recruited during voluntary running (51), and fiber-type analysis could be done precisely for the cross section in the middle portion of the muscle, we performed immunofluorescence-based fiber-type analysis in plantaris muscles following 4 wk of voluntary running as described (4). Our results showed that endurance exercise induced a significant increase in the percentage of type IIa myofibers in WT mice with a concurrent decrease in the percentage of type IIb fibers (Fig. 2A and Table 1). The percentage of type I fiber is extremely low in plantaris muscles, and endurance exercise training with the duration had no effect on the percentage of type I fibers in either genotypes. The increases in the percentage of type IIa myofibers in plantaris muscles were similar between MKO and WT mice, indicating this fiber-type transformation was independent of the genotypes (Fig. 2A and Table 1). To further confirm this result, we performed immunoblot analysis to quantify the expression of different isoforms of MHC proteins. As shown in Fig. 2, B and C, voluntary running induced significant increases in MHC IIa protein expression in plantaris muscles with a trend of reduction of MHC IIb protein; these changes were not affected by muscle-specific disruption of the Pgc-1α gene. Consistent with the immunofluorescence-based fiber type analysis, we could only detect trace amount of MHC I protein expression, and it was not affected by voluntary running. We observed the same MHC IIa protein induction by immunoblot analysis in two other predominantly glycolytic muscles, gastrocnemius (GA) and extensor digitorum longus (EDL), independent of the genotype (see Supplementary Figure S1, available with the online version of this article).

Open in a separate window

Fig. 2.

Immunofluorescence and immunoblot analyses for fiber-type composition in plantaris muscles of sedentary (Sed) and 4-wk endurance exercise-trained (Ex) WT and MKO mice. A: cross sections of plantaris muscles were immunostained for myosin heavy chains (MHC) I (red), IIa (blue), and IIb (green). Bars, 500 μm. All the fibers were counted and measured for cross-sectional area for calculation of fiber-type composition (please see Table 1). B: immunoblot analysis of MHC I, MHC IIa, and MHC IIb in plantaris muscles from sedentary (S) and endurance exercise-trained (E) mice with β-actin as loading control. Soleus muscle (Sol) was used as a positive control for MHC I. C: quantifications of MHC I, MHC IIa, and MHC IIb protein expression in plantaris muscles. Values are means ± SE (n = 6–7). **P < 0.01.

Table 1.

Comparison of fiber type composition in plantaris muscles in response to voluntary running in MKO and WT mice

	WT		MKO
	Sed (n = 6)	Ex (n = 5)	Sed (n = 6)	Ex (n = 7)
Fiber-type composition
MHC I
%Total fibers	1.42 ± 0.49	1.49 ± 0.59	3.32 ± 0.73‡	3.51 ± 0.65‡
%Area	0.46 ± 0.17	0.92 ± 0.54	1.50 ± 0.35‡	1.39 ± 0.27
MHC IIa
%Total fibers	21.1 ± 1.1	29.5 ± 2.5†	22.5 ± 1.1	29.6 ± 1.3†
%Area	15.1 ± 1.6	26.3 ± 5.0*	20.1 ± 1.0	32.4 ± 2.5†
MHC IId/x
%Total fibers	29.5 ± 1.3	30.0 ± 2.6	27.7 ± 1.1	26.4 ± 1.7
%Area	20.7 ± 1.2	26.8 ± 5.2*	24.9 ± 1.3	28.6 ± 2.2
MHC IIb
%Total fibers	48.0 ± 1.6	39.0 ± 4.4*	46.5 ± 1.5	40.5 ± 2.2
%Area	63.7 ± 2.7	46.0 ± 9.6*	53.5 ± 2.0	37.6 ± 3.9*

Open in a separate window

Values are means ± SE. MKO, muscle-specific Pgc-1α gene knockout mice; WT, wild-type littermates; Sed, sedentary; Ex, endurance exercise trained; MHC, myosin heavy chain; %Total fibers, fiber type as percentage of total fibers counted; %Area, fiber type area as percentage of total area.

^*P < 0.05,

^†P < 0.01 vs. Sed group of the same genotype.

^‡P < 0.05 vs. WT group of the same treatment.

Muscle-specific deletion of the Pgc-1α gene attenuates endurance exercise-induced mitochondrial biogenesis.

PGC-1α is a pivotal regulator of mitochondrial biogenesis by interacting with and stimulating the expression of transcription factors that function in control of nuclear-mitochondrial interactions, such as nuclear respiratory factor 1 (NRF1) and NRF2 (52). Despite the fact that overexpression of PGC-1α in cultured muscle cells and adult skeletal muscle induces mitochondrial biogenesis (33, 52), it was not known whether PGC-1α is required for endurance exercise-induced mitochondrial biogenesis, particularly in light of the recent finding that some exercise-responsive genes, including genes that function in mitochondrial oxidative phosphorylation, were not affected in mice with global knockout of the Pgc-1α gene (31). To address this important question, we performed immunoblot analysis for COX IV and Cyt c in skeletal muscles from MKO and WT mice following 4 wk of endurance exercise. These two proteins are critical components of the mitochondrial electron transport chain, and their expression levels are indicative of mitochondrial biogenesis (15, 26, 45). Following 4 wk of voluntary running, WT mice showed a 1.8-fold increase in COX IV expression (P < 0.001) while MKO mice showed a significantly less increase (1.3-fold, P < 0.05) (Fig. 3, A and B). The posttraining level of COX IV protein in MKO was significantly lower than that in WT mice (P < 0.001). Similar results were obtained for Cyt c protein; endurance exercise training induced a significant (1.6-fold) increase of Cyt c protein in WT mice (P < 0.001), but not in MKO mice (1.3-fold; P = 0.10) (Fig. 3, A and B). Interestingly, endurance exercise-induced expression of myoglobin, which is known to be upregulated in cultured muscle cells and in skeletal muscle in transgenic mice where the Pgc-1α gene is overexpressed (25, 33), was not affected by muscle-specific deletion of the Pgc-1α gene (P < 0.05 and P < 0.01 for WT and MKO mice, respectively). Similar observations for these proteins were made in three other glycolytic muscles, GA, EDL, and white vastus lateralis (WV) muscles (see Supplementary Figure S1, available with the online version of this article).

Open in a separate window

Fig. 3.

Immunoblot analysis for myoglobin and mitochondrial enzymes and immunofluorescence analysis for capillary density in skeletal muscles of sedentary (Sed) and 4-wk endurance exercise-trained (Ex) WT and MKO mice. A: immunoblot analysis of cytochrome oxidase IV (COX IV), cytochrome c (Cyt c), and myoglobin (Mb) protein expression in plantaris muscles from sedentary (S) and endurance exercise trained mice (E) with β-actin as loading control. Each lane was loaded with 40 μg of muscle protein from whole muscle homogenates. B: quantifications of COX IV, Cyt c and Mb protein expression in plantaris muscle (n = 6–7). *P < 0.05, **P < 0.01, ***P < 0.001. C: immunofluorescence staining of endothelial cells with anti-CD31 antibody in cross sections of plantaris muscle. Bars = 100 μm. D: quantifications of capillary density in plantaris muscle (n = 6–7). *P < 0.05, ***P < 0.001, respectively.

Muscle-specific deletion of the Pgc-1α gene attenuates endurance exercise-induced angiogenesis.

Angiogenesis is one of the important adaptations of skeletal muscle in response to endurance exercise training, and PGC-1α has recently been shown to play an important role in hypoxia-induced VEGF expression and angiogenesis independent of hypoxia-inducible factor-1 (HIF-1) in skeletal muscles (7). Considering this and the fact that MKO mice present a trend for reduced Vegf mRNA expression (Fig. 1B), we hypothesized that PGC-1α plays an essential role in endurance exercise-induced angiogenesis and performed immunofluorescence by using specific antibodies against platelet endothelial cell adhesion molecule-1 (PECAM-1, CD31) to evaluate capillary density in skeletal muscle following voluntary running. Consistent with our hypothesis, endurance exercise induced a ~60% increase in the number of CD31-positive endothelial cells per unit cross-sectional area (capillary density) in the WT mice. This adaptation was abolished in plantaris muscles from mice with muscle-specific deletion of the Pgc-1α gene (P = 0.17) (Fig. 3, C and D).

Voluntary running corrects abnormal mitochondrial morphology and restores SOD2 expression in skeletal muscles in MKO mice.

It has been shown previously that muscle-specific deletion of the Pgc-1α gene leads to enhanced expression of IL-6 and oxidative genes and increased susceptibility to muscle injury (24). We performed electron microscopy for plantaris muscle sections and found that MKO mice had significantly more vacuolated mitochondria (18.6 ± 3.2%) in both glycolytic and oxidative myofibers than WT mice (10.2 ± 3.2%). The vacuolated mitochondria showed reduced density of the matrix, disarray of the cristae, and significant enlargement (Fig. 4, A and B). Importantly, 4 wk of voluntary running resulted in significant reduction in the number and severity of mitochondrial vacuolation in both WT (4.8 ± 0.7%, P < 0.05 vs. WT-Sed mice) and MKO mice (3.9 ± 0.9%, P < 0.01 vs. MKO-Sed mice) (Fig. 4, A and B). Immunoblot analysis revealed that MKO mice had significantly less SOD2 protein expression in plantaris muscles compared with the wild-type littermates, which could be completely corrected by voluntary running (Fig. 4, C and D). These findings suggest that deletion of the Pgc-1α gene resulted in significantly compromised mitochondria integrity and impaired mitochondrial SOD2 expression in skeletal muscle, which could be restored by endurance exercise training in the absence of functional PGC-1α.

Open in a separate window

Fig. 4.

Transmission electron microscopy (TEM) analysis for mitochondrial morphology and immunoblot analysis for antioxidant enzyme expression in plantaris muscles of sedentary (Sed) and 4-wk endurance exercise-trained (Ex) WT and MKO mice. A: TEM images of longitudinal plantaris muscle sections of glycolytic and oxidative myofibers. Arrows show degenerative mitochondrion. To show detailed structure of some degenerative mitochondria, cropped images are shown with 2.5× magnification of the original images on one corner. Bars = 1 μm. B: quantification of abnormal mitochondria in ≥14 randomly acquired images as a percentage of total mitochondria. *P < 0.05, ***P < 0.001, respectively. C: immunoblot analysis of superoxide dismutase 2 (SOD2) protein expression in plantaris muscles from sedentary (S) and endurance exercise trained mice (E) with β-actin as loading control. Each lane was loaded with 40 μg of muscle protein from whole muscle homogenates. D: quantifications of SOD2 protein expression in plantaris muscle (n = 5–7). *P < 0.05.

DISCUSSION

One of the important questions in exercise physiology is about the regulatory factor(s) that is functionally responsible for endurance exercise-induced skeletal muscle adaptation. Since the cloning of the Pgc-1α gene, there have been increasing interests in its function in skeletal muscle plasticity. The most intriguing finding is the formation of significantly more oxidative myofibers when the Pgc-1α gene is overexpressed in skeletal muscle in transgenic mice (33). However, it was not known whether PGC-1α was required for endurance exercise-induced skeletal muscle adaptation and what aspects, if not all, of muscle adaptation would be affected when the Pgc-1α gene was deleted. This study confirmed the essential function of PGC-1α in endurance exercise-induced adaptation based on the findings that MKO mice had significant attenuation of 1) the induction of mitochondrial electron transport protein expression, and 2) new capillary formation in skeletal muscles. These two adaptations constitute the major adaptive processes that confer the oxidative metabolic phenotype leading to improved oxidative phosphorylation and microcirculation during exercise.

The voluntary running model has been used to induce skeletal muscle adaptation in rodents (6). While the periodic bursts of running activity are not the same as the endurance exercise in humans, the phenotypic adaptations are quite similar (4, 6, 47). The MKO mice were able to run similar distance as WT mice with a nearly twofold increase in daily running distance after 2 wk of endurance exercise. Therefore, the attenuated muscle adaptation in mitochondrial biogenesis and angiogenesis in MKO mice is clearly not due to reduced running activity in these mice. MKO had a similar degree of physiological cardiac hypertrophy (heart weight normalized by body weight) compared with WT mice, further supporting this notion. Furthermore, induced expression of other exercise responsive genes, such as myoglobin and MHC IIa, was not affected by the Pgc-1α gene deletion in skeletal muscle, providing biochemical evidence for the exercise perturbation.

The similar daily running distance between MKO and WT mice seems to be contradictory to the substantial attenuation of endurance exercise-induced COX IV and Cyt c expression as these proteins play important role in mitochondrial oxidative phosphorylation and hence endurance exercise capacity. There are at least three possible reasons for this apparently paradoxical finding. First and most likely, mitochondrial function and/or efficiency may be enhanced independent of mitochondrial biogenesis. In other words, in the absence of increased mitochondrial electron transport protein expression, mitochondrial function may still be improved through other processes, such as posttranslational modification and/or improved physical structure of the electron transport chain complexes. Our finding that MKO mice had a significant amount of abnormal mitochondria and reduced SOD2 protein expression that could be corrected by voluntary running support to a certain degree the independence of the mitochondrial remodeling process on PGC-1α function. Consistent with this notion, Adhihetty et al. (1) have recently shown that Pgc-1α gene disruption impaired mitochondrial function (state 3 and state 4 respiration) in isolated mitochondria from skeletal muscle of sedentary animals, but that endurance exercise training led to enhanced mitochondrial function regardless of the genetic background. Second, other muscle adaptive processes, such as those related to calcium handling and the contractile apparatus, may result in compensation for the Pgc-1α gene deletion-induced loss of oxidative function in skeletal muscles. Finally, although voluntary running is a great physiological exercise model to induce skeletal muscle adaptation, the daily running distance on the voluntary wheel is not an index of endurance capacity. It would be very interesting to determine if MKO mice have lower endurance capacity compared with WT mice and if endurance exercise training improves exercise capacity in these mice.

Mitochondrial biogenesis, the process by which new mitochondria are formed in the cell along with increased mitochondrial enzyme expression/activity, is not only a hallmark of skeletal muscle adaptation (20, 27, 28) but is also pivotal for the enhanced exercise capacity (10, 27). Much of the recent attentions have been paid to PGC-1α function in endurance exercise-induced mitochondrial biogenesis since activation of Pgc-1α transcription (9, 42) and activity (13) occurs upon a single bout of endurance exercise, and transgenic overexpression of the Pgc-1α gene or upstream signaling molecules in skeletal muscle resulted in enhanced mitochondrial biogenesis (3, 19, 33). While we were conducting this study in muscle-specific Pgc-1α knockout mice, Leick et al. (31) employed a mouse model of global gene disruption of the Pgc-1α gene and obtained surprising results indicating that Pgc-1α is not mandatory for exercise- and training-induced adaptive gene responses. Since the mice with the global disruption of the Pgc-1α gene had hyperactivity due to lesions in the central nervous system, an approach of skeletal muscle-specific gene disruption is desirable to minimize the potential confounding factors. Another note to consider is that two of the protein markers, hexokinase II (HK II) and aminolevulinate synthase 1 (ALAS1) used in the aforementioned study are not mitochondrial enzymes, and the other one, COX I, is a protein that has been shown not to be controlled by PGC-1α (18). Our finding that endurance exercise-induced COX IV and Cyt c protein expression were significantly attenuated in MKO mice is consistent with an essential functional role of PGC-1α in mediating endurance exercise-induced mitochondrial biogenesis. A final note on mitochondrial adaptations in MKO is the observation that mitochondrial morphology and SOD2 protein expression were improved with endurance exercise training in these animals. Thus, whereas normal PGC-1α expression seems to be required for mitochondrial maintenance in sedentary animals, the mechanisms involved in endurance exercise-induced improvement in mitochondrial morphology remain to be determined.

Angiogenesis, the formation of new capillary from the existing vasculature, is equally important to improved functional capacity after exercise training (11, 22). It has even been postulated that angiogenesis precedes and play a permissive role for the other adaptive processes, such as fiber-type transformation (47). The recent study by Arany et al. (7) showed that PGC-1α is necessary and sufficient for hypoxia-induced angiogenesis in skeletal muscle. This raised the possibility of a functional role of PGC-1α in endurance exercise-induced angiogenesis although it is still debatable whether hypoxia constitutes an endurance exercise signal for phenotypic adaptations. Interestingly, mice with global gene disruption of the Pgc-1α gene showed impaired VEGF and CD31 expression in skeletal muscle in response to endurance exercise training (30) despite the findings that these mice had relatively normal exercise-induced gene responses (31). Our findings that MKO mice had abolished increase of capillary density in skeletal muscle following 4 wk of voluntary running provide direct evidence that PGC-1α is essential for endurance exercise-induced angiogenesis. In light of the findings that hypoxia-induced angiogenesis in skeletal muscle is mediated by PGC-1α independent of HIF-1, it would be interesting to determine if endurance exercise-induced angiogenesis in skeletal muscle is achieved through the same mechanism.

In addition to the phenotypic changes that confer characteristics in metabolism, endurance exercise also promotes IIb-to-IIa fiber-type transformation (16, 47). While overexpression of Pgc-1α in skeletal muscle promotes more oxidative myofiber formation (33), we showed here that voluntary exercise-induced IIb-to-IIa transformation was not affected by the deletion of the Pgc-1α gene in skeletal muscle, suggesting that endurance exercise-induced fiber-type transformation in skeletal muscle could occur independent of the function of the Pgc-1α gene. These findings have several implications. First, considering the impaired angiogenesis in the muscle-specific knockout mice, one might argue that angiogenesis is not obligatory for endurance exercise-induced fiber-type switching. In other words, these findings are not consistent with the hypothesis that angiogenesis plays an instrumental role for fiber-type transformation. Second, it has previously been shown that overexpression of active upstream signaling molecules in skeletal muscle that upregulate PGC-1α expression can promote mitochondrial biogenesis without inducing shifts in fiber-type composition (3, 19). We have also obtained new evidence in skeletal muscle-specific p38 knockout mice that endurance exercise-induced mitochondrial biogenesis and angiogenesis in skeletal muscle, but not fiber-type transformation, are dependent on the upstream p38γ MAPK signaling pathway (43). The findings in this study once again in a tissue-specific loss-of-function mouse model genetically separate the endurance exercise-induced mitochondrial biogenesis and angiogenesis from the contractile adaptation, specifically IIb-to-IIa fiber-type transformation. All these findings collectively support the view that PGC-1α and its regulatory control axis play a functional role in complete skeletal muscle adaptations that lead to improved microcirculatory function and oxidative metabolism, but their function seems not absolutely required for endurance exercise-induced contractile adaptation in mice.

It should be pointed out that although we have employed skeletal muscle-specific Pgc-1α knockout mice to minimize the possibility of developmental and interorgan influences from the disruption of the gene globally, we could not rule out the possibility of other regulatory factors that function in a compensatory manner. For example, Pgc-1β has recently been shown to promote formation of type IIx fibers (8), although the decreased Pgc-1β expression in human skeletal muscle by endurance exercise training (38) appears to be contradictory to this potential function in promoting fiber-type transformation from IIb to IIa. The possibility that a redundant functional role of a different isoform of PGC-1α as a result of alternative promoter (37) in MKO mice could be ruled out as we deleted exons 3, 4, and 5 of the Pgc-1α gene and prevented the formation of this PGC-1α isoform. The fact that endurance exercise-induced mitochondrial biogenesis and angiogenesis are significantly attenuated in Pgc-1α MKO mice suggests that the essential function of PGC-1α in these fundamental adaptive processes could not be compensated adequately by redundant systems if they existed.

In summary, our studies revealed that Pgc-1α plays an important functional role for endurance exercise-induced mitochondrial biogenesis and angiogenesis but not for IIb-to-IIa fiber-type transformation in mouse skeletal muscle. Our findings provide evidence in a loss-of-function mode that some of the metabolic adaptations could be genetically separated from the contractile adaptation. The improved understanding of the molecular mechanisms by which endurance exercise training promotes skeletal muscle adaptation will be instrumental for future development of novel interventions for many chronic diseases that have been shown to be benefited from regular exercise.

DISCLOSURES

No conflicts of interest are declared by the authors.

Supplementary Material

REFERENCES