Abstract

There are three isoforms of peroxisome proliferator-activated receptor-γ coactivator 1α (PGC-1α) mRNA, which promotes mitochondrial biogenesis in skeletal muscles. Compared with PGC-1α-a mRNA, PGC-1α-b or PGC-1α-c mRNA is transcribed by a different exon 1 of the PGC-1α gene. In this study, effects of exercise intensity and 5-aminoimidazole-4-carboxamide-1β-d-ribofuranoside (AICAR) on isoform-specific expressions of PGC-1α were investigated. All isoforms were increased in proportion to exercise intensity of treadmill running (10–30 m/min for 30 min). Preinjection of β2-adrenergic receptor (AR) antagonist (ICI 118551) inhibited the increase in PGC-1α-b and PGC-1α-c mRNAs, but not the increase in PGC-1α-a mRNA, in response to high-intensity exercise. Although high-intensity exercise activated α2-AMP-activated protein kinase (α2-AMPK) in skeletal muscles, inactivation of α2-AMPK activity did not affect high-intensity exercise-induced mRNA expression of all PGC-1α isoforms, suggesting that activation of α2-AMPK is not mandatory for an increase in PGC-1α mRNA by high-intensity exercise. A single injection in mice of AICAR, an AMPK activator, increased mRNAs of all PGC-1α isoforms. AICAR increased blood catecholamine concentrations, and preinjection of β2-AR antagonist inhibited the increase in PGC-1α-b and PGC-1α-c mRNAs but not the increase in PGC-1α-a mRNA. Direct exposure of epitrochlearis muscle to AICAR increased PGC-1α-a but not the -b isoform. These data indicate that exercise-induced PGC-1α expression was dependent on the intensity of exercise. Exercise or AICAR injection increased PGC-1α-b and PGC-1α-c mRNAs via β2-AR activation, whereas high-intensity exercise increased PGC-1α-a expression by a multiple mechanism in which α2-AMPK is one of the signaling pathways.

peroxisome proliferator-activated receptor-γ coactivator 1α (PGC-1α), which was originally identified as a nuclear receptor coactivator, is expressed in brown adipose tissue (BAT), skeletal muscle, heart, kidney, liver, and brain, is markedly upregulated in BAT and skeletal muscle after acute exposure to cold stress, and promotes mitochondrial biogenesis (30). There are three isoforms of PGC-1α mRNA, which promotes mitochondrial biogenesis and angiogenesis in skeletal muscles (6, 25, 46). Compared with PGC-1α-a mRNA (a previously described isoform), PGC-1α-b and PGC-1α-c mRNAs are transcribed by a different exon 1 of the PGC-1α gene and produce slightly smaller-sized proteins. Transgenic mice overexpressing PGC-1α-b or PGC-1α-c in skeletal muscles showed increased gene expression related to mitochondrial biogenesis and fatty acid oxidation (25). In mice, low-intensity exercise (15 m/min for 45 min) or injection of the β2-adrenergic receptor (AR) agonist clenbuterol increased PGC-1α-b and PGC-1α-c mRNA expression more than 40-fold, but not that of PGC-1α-a, in skeletal muscle (25). The increase in PGC-1α mRNA expression was specific to β2-AR; injection of α-, β1-, or β3-AR agonist did not increase PGC-1α mRNA expression (26). Increased expressions in skeletal muscles in response to exercise were inhibited by pretreatment with the β2-AR-specific antagonist ICI 118551 or nonspecific β-AR antagonist propranolol (26). These data indicate that β2-AR activation is a major mechanism in the increase in PGC-1α expression in skeletal muscles and that the increase in PGC-1α mRNAs is isoform specific. However, it is not known which isoform of PGC-1α was increased in response to the high-intensity exercise, by which α2-AMPK was markedly activated (5, 42). Therefore, in the present study, we examined the effects of exercise intensity on increases and contributions of AMPK to isoform-specific increase of PGC-1α mRNA.

5-Aminoimidazole-4-carboxamide-1β-d-ribofuranoside (AICAR) is a pharmacological activator of AMPK demonstrating similar effects of exercise for increasing glucose uptake (23) and fatty acid oxidation (27) in isolated skeletal muscles. In vivo, AICAR injection in rodents resulted in an increase in fatty acid oxidation (15), mitochondrial biogenesis (41), PGC-1α mRNA (17), and protein (36). Therefore, we also investigated isoform-specific increases in PGC-1α mRNA in response to AICAR.

METHODS

Experimental animals.

Eight-week-old male C57BL/6J mice were obtained from Japan SLC (Hamamatsu, Japan). Five-week-old male Wistar rats were obtained from Japan CLEA (Tokyo, Japan). Mice and rats were cared for in accordance with the NIH Guide for the Care and Use of Laboratory Animals and our own institutional guidelines. All animal experiments were conducted with the approval of the National Institute of Health and Nutrition Ethics Committee on Animal Research (No. 1008).

Experimental protocols.

For exercise experiments, mice were subjected to treadmill running at 10 m/min (low-intensity exercise), 20 m/min (medium-intensity exercise), or 30 m/min (high-intensity exercise) for 30 min. For AICAR experiments, mice were injected subcutaneously with 1 g/kg body wt AICAR or the same volume of saline. At 0 h (just after exercise) and at 1, 3, 6, and 24 h after exercise or AICAR injection, skeletal muscles (gastrocnemius) were isolated. For β2-AR antagonist experiments, mice were injected subcutaneously with 10 mg/kg body wt ICI 118551 or the same volume of saline 1 h before exercise or AICAR injection. In all experiments, skeletal muscle was rapidly (within 30–60 s) removed from mice killed by decapitation, frozen immediately in liquid nitrogen, and kept at −80°C.

Measurement of AMPK activity.

Muscles were homogenized in ice-cold lysis buffer containing 50 mM Tris·HCl, pH 7.4, 1 mM EGTA, 1 mM EDTA, 1% Triton X-100, 1 mM sodium orthovanadate, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 0.27 M sucrose, 0.1% 2-mercaptoethanol, and “Complete Mini EDTA-free” protease inhibitor cocktail (1 tablet per 10 ml; Roche Diagnostics, Mannheim, Germany). Homogenates were centrifuged at 13,000 g for 10 min at 4°C, supernatants were removed, and aliquots were snap-frozen in liquid nitrogen. Muscle lysate protein (0.1 mg) was incubated at 4°C for 30 min with the corresponding antibody conjugated to protein G-Sepharose (GE Healthcare Bio-Science, Uppsala, Sweden). The immunoprecipitates were washed twice with lysis buffer containing 0.5 mM NaCl and twice with buffer A (50 mM Tris·HCl, pH 7.4, 0.1 mM EGTA, and 0.1% 2-mercaptoethanol). Phosphotransferase activity of the immunoprecipitates toward the SAMS peptide was measured as described previously (32). The antibodies against the α1 or α2 catalytic subunits of AMPK (cat. no. 07-350 and 07-363; Upstate Biotechnology, Lake Placid, NY) were purchased.

Transgenic mice.

Transgenic mice were produced expressing a dominant-negative (DN) mutant of α1-AMPK in skeletal muscle (α1-AMPK-DN mice) as previously described (24). Briefly, human skeletal muscle α-actin promoter provided by Drs. E. D. Hardeman and K. Guven (Children’s Medical Research Institute, Westmead, Australia) was used to express α1-AMPK-DN in skeletal muscle. α1-AMPK-DN mice exhibited a 50% reduction in α1-AMPK activity and almost complete loss of α2-AMPK activity in skeletal muscle compared with wild-type littermates (24). Male chimeras (BDF1 background) harboring the α1-AMPK-DN transgene were mated with pure C57BL/6J females to obtain F1 offspring. The heterozygous F1 male offspring from this breeding were then back-crossed with purebred C57BL/6J females to obtain F2 offspring, and this process was continued until the F7 generation of mice was obtained. Heterozygous F7 offspring and wild type, born at the same period were used for the experiments.

Western blot.

Frozen skeletal muscle was powdered under liquid nitrogen and then homogenized in RIPA lysis buffer (Upstate Biotechnology) containing 1 mM sodium orthovanadate, 1 mM sodium fluoride, and “Complete Mini, EDTA-free” protease inhibitor cocktail (1 tablet per 10 ml). After three cycles of freezing and thawing, the supernatant was separated by centrifugation at 20,400 g for 15 min at 4°C. Fifteen micrograms of protein of the supernatant was applied onto SDS-PAGE. Phosphorylated and total acetyl-CoA carboxylase (ACC) protein was measured by Western blotting with anti-phospho-ACC (S79) antibody (cat. no. 3661; Cell Signaling Technology, Beverly, MA) and anti-ACC antibody (cat. no. 3662; Cell Signaling Technology), respectively.

Epitrochlearis preparation.

Male Wistar rats, whose body weights were ∼110 g, were anesthetized with an intraperitoneal injection of pentobarbital sodium, 5 mg/100 g body wt, and the epitrochlearis muscles were dissected out and incubated at 35°C for 5 h in 3 ml of oxygenated Krebs-Henseleit bicarbonate buffer containing 8 mM glucose and 32 mM mannitol (18). The gas phase in the flasks was 95% O2-5% CO2. Epitrochlearis muscles were exposed to 1 μM β2-AR agonist clenbuterol for 5 h or 0.5 mM AICAR for 6 h. After the incubations, muscles were blotted and frozen at −80°C for RNA preparation.

Quantitative real-time RT-PCR.

RNA preparation methods and quantitative real-time RT-PCR were performed as described previously (25). PCR primers to detect all isoforms of PGC-1α (total PGC-1α) were selected to correspond to sequences in the second and third exons of the PGC-1α gene (25). The mouse-specific primer pairs used were as follows: PGC-1α-a forward 5′-GGGACATGTGCAGCCAAGA-3′, reverse, 5′-AAGAGGCTGGTCCTCACCAA-3′; PGC-1α-b forward 5′-GACATGGATGTTGGGATTGTCA-3′, reverse 5′-ACCAACCAGAGCAGCACATTT-3′; PGC-1α-c forward 5′-TGAGTAACCGGAGGCATTCTCT-3′, reverse 5′-TGAGGACCGCTAGCAAGTTTG-3′; 36B4 forward, 5′-GGCCCTGCACTCTCGCTTTC-3′; 36B4 reverse 5′-TGCCAGGACGCGCTTGT-3′. The rat-specific primer pairs used were as follows: PGC-1α-a forward 5′-GGGACATGTGCAGCCAAGA-3′, reverse 5′-AAGAGGCTGGTCCTCACCAA-3′; PGC-1α-b forward 5′-GATATGGATGTCGGGTTTGTCA-3′, reverse 5′-ACCAACCAGAGCAGCACATTT-3′; 36B4 forward 5′-CCTTCCCACTGGCTGAAAAGG-3′, reverse 5′-AGCCGCAGCCGCAAATGCAG-3′. For comparison of the amount of mRNA in each isoform and total PGC-1α, cDNA was amplified by the same primer sets in quantitative real-time RT-PCR. After purification and quantification of the amount, using cDNA from each isoform or total PGC-1α as a standard, we made a standard titration curve of each isoform or total PGC-1α mRNA by quantitative real-time RT-PCR. We then quantified the amount of mRNA in each isoform and total PGC-1α relative to amount of each cDNA. By these means, we were able to compare the amounts of mRNAs derived from each isoform PGC-1α gene (or total PGC-1α). For comparison, relative values of the mean of PGC1α-a mRNA from the control mice are shown in Figs. 15.

Measurements of ZMP contents.

Gastrocnemius was homogenized with 1.0 N perchloric acid and centrifuged at 10,000 g for 15 min. After neutralization of the supernatant with calcium carbonate, ZMP [5-aminoimidazole-4-carboxamide-1β-d-ribofuranosyl 5′-monophosphate (AICAR monophosphate)] concentrations were determined by high-performance liquid chromatogram (Phenomenex Luna 5 μ NH2; mobile phase, sodium phosphate buffer; detector, 260 nm) (33, 39).

Blood glucose and plasma catecholamine.

Blood samples were obtained intravenously under anesthesia, and blood glucose concentration was measured with a glucose analyzer (Glucometer DEx; Bayer Medical, Tokyo, Japan). For catecholamine analysis, plasma samples were separated by centrifugation in the presence of EDTA and snap-frozen at −80°C until analysis. Plasma catecholamine was measured with an automated catecholamine analyzer/combined post-column reactor system HLC-725CA II (Tosoh, Tokyo, Japan).

Statistical analysis.

All values are means ± SE. Data were analyzed by one-way or two-way ANOVA. Where differences were significant, each group was compared with the other by Student’s t-test (JMP 5.1.2; SAS, Cary, NC). Statistical significance was defined as P < 0.05.

RESULTS

Exercise intensity-dependent expression of PGC-1α mRNA isoform.

Total PGC-1α mRNA expression was determined in gastrocnemius obtained at time 0 (just after exercise) and at 1, 3, 6, and 24 h after treadmill running at 10, 20, or 30 m/min for 30 min. PGC-1α expression was not increased just after exercise (0 h), and for each exercise intensity, PGC-1α expression was highest at 1 h after exercise (Fig. 1A). The amounts of these increases were dependent on the intensity of exercise; at 1 h after the exercise, the amount of PGC-1α mRNA was elevated 2.0-, 5.0-, and 7.1-fold after 10, 20, and 30 m/min running, respectively. In the following experiments, PGC-1α mRNA expression was analyzed in gastrocnemius obtained at 1 h after a 30-min exercise, the time showing highest expression. Expression of the three PGC-1α mRNA isoforms was measured using isoform-specific primers by quantitative RT-PCR (Fig. 1B). Although the magnitudes of the response to exercise were different between isoforms, the expressions of each PGC-1α isoform were increased in accordance with the level of exercise intensity. PGC-1α-a was not increased after 10 m/min running but was significantly increased 1.4- and 1.8-fold after 20 and 30 m/min exercise, respectively. In contrast, PGC-1α-b expressions were increased ∼6.0-, 20.0-, and 33.3-fold, and PGC-1α-c expressions were increased 5.5-, 17.7-, and 16.0-fold after 10, 20, and 30 m/min running, respectively.

Fig. 1.

Effects of exercise intensity on peroxisome proliferator-activated receptor-γ coactivator 1α (PGC-1α) isoform mRNAs. A: changes in total PGC-1α mRNA expression in skeletal muscles (gastrocnemius) at time 0 (just after exercise) and at 1, 3, 6, and 24 h after treadmill running at 10, 20, or 30 m/min for 30 min. Control mice were kept sedentary (Sed). Graph shows percentage of total PGC-1α mRNA level in sedentary mice. Values are means ± SE (n = 3–10). *P < 0.05, **P < 0.01, ***P < 0.001 vs. Sed. B: increases in PGC-1α isoform mRNA expression in skeletal muscles (gastrocnemius) 1 h after treadmill running at 10, 20, or 30 m/min for 30 min. Control mice were kept sedentary. Graph shows percentage in PGC-1α isoform mRNA level relative to the PGC-1α-a mRNA level of sedentary mice. Values are means ± SE (n = 3–8). *P < 0.05, **P < 0.01 vs. Sed.

Exercise-activated β2-AR for the expression of PGC-1α mRNA isoform.

In our previous study, an injection of clenbuterol (β2-AR agonist) in mice markedly increased PGC-1α-b and PGC-1α-c mRNAs in skeletal muscles but not PGC-1α-a mRNA (25). Pretreatment with the β2-AR-specific antagonist ICI 118551 or the nonspecific β-AR antagonist propranolol inhibited increases in PGC-1α-b and PGC-1α-c isoforms in response to low-intensity exercise (15 m/min, 45 min) (25). The effect of β-AR blockade on levels of PGC-1α isoform mRNA in response to high-intensity exercise (30 m/min), by which PGC-1α-a mRNA was increased, was examined using a β2-AR-specific antagonist. Increased expression of PGC-1α-b and PGC-1α-c isoforms in response to high-intensity exercise was markedly inhibited by pretreatment with ICI 118551 at 1 h before the exercise, whereas increased PGC-1α-a expression was not inhibited (Fig. 2). The same results were observed by pretreatment with the nonspecific β-AR antagonist propranolol (data not shown).

Fig. 2.

Effects of β2-adrenergic receptor (AR) antagonist injection on PGC-1α isoform mRNA expressions in response to high-intensity exercise. Mice were injected with 10 mg/kg body wt sc ICI 118551 (Pre-β2-AR antagonist). An identical volume of saline was injected as a control (Pre-Saline). At 1 h after injection, 30 m/min treadmill exercise was performed for 30 min (Run). Control mice were kept sedentary (Sed). Skeletal muscles (gastrocnemius) at 1 h after a 30-min exercise were taken from mice. Graph shows percentage of PGC-1α isoform mRNA level relative to PGC-1α-a mRNA level of sedentary mice. Values are means ± SE (n = 5–6). **P < 0.01, ***P < 0.001 vs. Sed., †††P < 0.001 vs. Pre-Saline.

α2-AMPK is not essential for increased expression of PGC-1α isoforms in response to exercise.

Activation of AMPK in skeletal muscles increases fatty acid oxidation and mitochondrial biogenesis (40). AMPK phosphorylates PGC-1α protein directly and activates the promoter activity of PGC-1α (17). Because activation of AMPK was observed after exercise in skeletal muscle (5, 42), AMPK might be involved in increased expression of PGC-1α mRNAs after exercise. Gastrocnemius was obtained just after treadmill running at 10, 20, or 30 m/min for 30 min, and AMPK activity was measured. Although α1-AMPK activity was not altered, α2-AMPK activity was increased 1.6- and 2.0-fold just after 20 and 30 m/min running for 30 min, respectively (Fig. 3A). To examine whether activation of AMPK in response to a single bout of exercise might lead to increased expression of PGC-1α mRNA isoforms, transgenic mice expressing α1-AMPK-DN in skeletal muscle were subjected to treadmill running. Because α1-AMPK-DN mice cannot maintain a 30 m/min running speed for 30 min (24), the intensity of exercise was set at 20 m/min for 30 min. α2-AMPK activity was severely inhibited in the skeletal muscle of the α1-AMPK-DN mice when sedentary and after running at 10 m/min (24). α2-AMPK activity was also inhibited in the skeletal muscle of the α1-AMPK-DN mice just after running at 20 m/min for 30min (Fig. 3B). In contrast, α2-AMPK activity was increased 2.7-fold in wild-type skeletal muscle after running. To determine AMPK activity in vivo, phosphorylation of ACC, one of the substrates for AMPK in vivo, was measured. In α1-AMPK-DN mice after running, although α2-AMPK activity was almost completely inhibited, phosphorylation of ACC was significantly increased. However, the amount of phosphorylated ACC in α1-AMPK-DN mice was significantly low compared with that in wild-type mice not only when the mice were sedentary but also after running. Although α2-AMPK activity was severely inhibited in the skeletal muscle of the α1-AMPK-DN mice, 30 min of exercise at 20 m/min resulted in increased expression of each isoform of PGC-1α mRNA (Fig. 3C).

Fig. 3.

Effects of AMPK on increases in PGC-1α isoform mRNAs in response to a bout of exercise. A: changes in α1- and α2-AMPK subunit activities in skeletal muscle (gastrocnemius) after exercise. Skeletal muscles were obtained just after treadmill running at 10, 20, or 30 m/min for 30 min. Control mice were kept sedentary. AMPK activity was measured in immunoprecipitates. Each value (pmol/min/mg protein), is the mean ± SE of 4–6 mice. ***P < 0.001 vs. Sed. B: changes in α1- and α2-AMPK subunit activities, and phosphorylation of ACC in skeletal muscle (gastrocnemius) from α1-AMPK-DN (dominant-negative) mice (AMPK-DN) and wild-type littermates (Wt) after running at 20 m/min for 30 min. Skeletal muscles were obtained just after treadmill exercise (Run) or during the resting state (Sed). AMPK activity was measured in immunoprecipitates. Each value is the mean ± SE of 5–6 mice. ***P < 0.001 vs. Sed., †P < 0.05 and †††P < 0.001 vs. Wt. Phospho-ACC/total ACC was measured by Western blotting. Each value, expressed as a ratio of the value for sedentary Wt littermates, is mean ± SE (n = 5–6). Typical blots are also shown. **P < 0.01 vs. Sed., †P < 0.05 vs. Wt. C: AMPK-DN and Wt mice were run for 30 min on a treadmill at 20 m/min side by side (Run). At 1 h after a 30-min exercise, skeletal muscles (gastrocnemius) were taken, and the level of PGC-1α mRNA was analyzed. Control mice were kept sedentary. Graph shows percentage in PGC-1α isoform mRNA level relative to PGC-1α-a mRNA level of sedentary mice. Values are means ± SE (n = 6–7). ***P < 0.001 vs. Sed.

AICAR-induced expression of PGC-1α mRNA isoform.

AICAR, an AMPK activator, increased the expression of PGC-1α mRNA in skeletal muscle in in vivo and ex vivo experiments (17, 38). Total PGC-1α mRNA expression was measured in gastrocnemius obtained at 0, 1, 3, 6, and 24 h after injection of 1 g/kg body wt AICAR under feeding conditions. This dose of AICAR decreased blood glucose concentrations but did not cause overt hypoglycemia; blood glucose concentrations at 0, 60, and 180 min after AICAR injection were 219, 87, and 149 mg/dl, respectively. Total PGC-1α expression was not increased at 1 h but did increase 8.7- and 6.9-fold at 3 and 6 h after the injection, respectively (Fig. 4A). In the following experiments of the effects of AICAR on PGC-1α expression, expression of PGC-1α mRNA was measured in gastrocnemius obtained at 3 h, the time showing highest expression, after AICAR injection. Expression of all three PGC-1α isoforms was increased significantly, similar to that after a bout of exercise (Fig. 4B). Pretreatment with β2-AR antagonist at 1 h before AICAR injection inhibited the increased expression of PGC-1α-b and PGC-1α-c isoforms but not that of the PGC-1α-a isoform (Fig. 4B). The increased expression of PGC-1α-b and PGC-1α-c isoforms in response to AICAR injection was also inhibited by pretreatment with the nonspecific β-AR antagonist propranolol (data not shown).

Fig. 4.

Effect of AICAR injection on increases in PGC-1α isoform mRNAs and blood catecholamine concentrations. A: changes in total PGC-1α mRNA expression in skeletal muscles (gastrocnemius) at 1, 3, 6, and 24 h after 1 g/kg body wt AICAR injected sc (AICAR). An identical volume of saline was injected as a control (Saline). Graph shows percentage of total PGC-1α mRNA level vs. that before injection (Pre). Values are means ± SE (n = 3–4). ***P < 0.001 vs. Saline. B: effects of β2-AR antagonist on increased expression of PGC-1α isoforms in response to AICAR. Mice were preinjected sc with 10 mg/kg body wt of ICI 118551 (Pre-β2-AR antagonist). An identical volume of saline was injected as a control (Pre-Saline). At 1 h after injection, 1 g/kg body wt of AICAR (AICAR) or an identical volume of saline (Saline) was injected sc. Skeletal muscles (gastrocnemius) at 3 h after AICAR or saline injections were taken from mice. Graph shows percentage in PGC-1α isoform mRNA level relative to PGC-1α-a mRNA level of sedentary mice. Values are means ± SE (n = 5). ***P < 0.001 vs. Saline, †P < 0.05 and †††P < 0.001 vs. Pre-Saline. C: skeletal muscle ZMP concentration and phosphorylation of ACC after AICAR injection with or without preinjection of β2-AR antagonist. Mice were preinjected sc with 10 mg/kg body wt of ICI 118551 (Pre-β2-AR antagonist). An identical volume of saline was injected as a control (Pre-Saline). At 1 h after injection, 1 g/kg body wt of AICAR or an identical volume of saline was injected sc. Skeletal muscles (gastrocnemius) at 1 h after AICAR or saline injections were taken from mice. Values are means ± SE (n = 5–6). N.D., not detected. **P < 0.01 vs. Saline, †††P < 0.001 vs. Pre-Saline. D: plasma catecholamine concentrations after AICAR injection. Epinephrine and norepinephrine concentrations were measured in plasma obtained before (Pre) and at 1 and 3 h after injection of 1 g/kg body wt of AICAR (AICAR) or an identical volume of saline (Saline). Values are means ± SE (n = 6). *P < 0.05, ***P < 0.001 vs. Saline.

AICAR is an adenosine analog taken up by skeletal muscle and phosphorylated to form ZMP, which stimulates AMPK activity (8). To examine whether the concentration of ZMP in gastrocnemius was affected by preinjection of β2-AR antagonist, measurement of ZMP in gastrocnemius was performed at 1 h after AICAR injection (2 h before a maximum increase in PGC-1α mRNA was noted) (Fig. 4C). The concentration of ZMP in the mice preinjected with β2-AR antagonist was 44% lower than that in the mice preinjected with saline, suggesting that pretreatment of mice with β2-AR antagonist might decrease delivery of AICAR or phosphorylation of AICAR into ZMP in the skeletal muscle. However, phosphorylation of ACC in gastrocnemius was increased at 1 h after AICAR injection up to 3.0- and 3.4-fold with or without preinjection of β2-AR antagonist (Fig. 4C). These data suggested that the amount of ZMP in the skeletal muscle would be enough to activate AMPK in vivo (shown as phosphorylation of ACC) and upregulate PGC-1α-a mRNA expression irrespectively of preinjection of β2-AR antagonist.

The increased expression of PGC-1α-b and PGC-1α-c isoforms in response to AICAR injection might be due to the increased plasma catecholamine concentrations. Plasma epinephrine concentrations measured at 1 and 3 h after AICAR injection were increased 9.2- and 2.5-fold, respectively (Fig. 4D). However, plasma norepinephrine concentration was only slightly increased at 1 h after AICAR injection.

Basal plasma catecholamine concentrations might be higher due to the stress induced by handling of the mice. Indeed, the pretreatment with β2-AR antagonist reduced the expression of mRNAs of PGC-1α-b and PGC-1α-c isoforms by saline injection (Fig. 4B), in which catecholamines may induce. Even though basal catecholamine concentration might be elevated, injection of AICAR markedly increased epinephrine concentration in the plasma much more than in the control sample collected under the same condition.

Direct effects of β2-AR agonist and AICAR on the expression of PGC-1α mRNA isoform.

Incubation of isolated muscle with the β2-AR agonist clenbuterol increased the expression of total PGC-1α mRNA (26). To determine which isoform of PGC-1α mRNA was increased, epitrochlearis muscle isolated from rat was incubated with 1 μM clenbuterol for 5 h at 35°C and then frozen, and expression levels of PGC-1α mRNA isoforms in epitrochlearis were measured later (Fig. 5A). We used rat epitrochlearis because of its larger size and easier manipulation. PGC-1α-c was not expressed in rat skeletal muscle, because the sequence required for splicing to produce PGC-1α-c does not exist in rat DNA. A significant increase in the expression of PGC-1α-b mRNA was observed by the incubation with clenbuterol, whereas clenbuterol treatment did not significantly increase expression of PGC-1α-a mRNA. In contrast, when epitrochlearis muscle was incubated with 0.5 mM AICAR for 6 h at 35°C, a significant increase in the expression of PGC-1α-a mRNA was observed, whereas expression of PGC-1α-b mRNA was unchanged (Fig. 5B). These data indicated that an increase of PGC-1α-a mRNA in response to AICAR injection might be due, at least in part, to a direct effect of AICAR on skeletal muscle, and an increase of PGC-1α-b mRNA in response to clenbuterol injection might be due to a direct effect of β2-AR stimulation. However, in vivo, AICAR injection in mice activates β2-AR by an increase in plasma epinephrine concentration, which leads to increases in PGC-1α-b and PGC-1α-c mRNAs.

Fig. 5.

Ex vivo effects of β2-AR agonist (A) and AICAR (B) treatment on PGC-1α isoform mRNAs. Isolated rat epitrochlearis was exposed for 5 h to 1 μM clenbuterol or for 6 h to 0.5 mM AICAR and frozen for mRNA analysis. An identical volume of saline was used as a control. Graph shows percentage in PGC-1α isoform mRNA level relative to PGC-1α-a mRNA level of saline-treated epitrochlearis. Values are means ± SE (n = 9). *P < 0.05, **P < 0.01 vs. Saline.

DISCUSSION

It has been reported that there are several molecular pathways to increase PGC-1α expression after exercise. The activation of AMPK, calcium/calmodulin-dependent protein kinase (CaMK), calcineurin A (CnA), and p38 mitogen-activated protein kinase (MAPK) signaling cascades are well-known upstream regulators of PGC-1α expression in skeletal muscle (2, 17, 43). In addition, we have previously shown that β2-AR activation is required for low-intensity exercise-induced PGC-1α expression (26), especially for that of PGC-1α-b and PGC-1α-c (25). In the present study, we found that low-, medium-, and high-intensity exercise and AICAR injection increased PGC-1α-b and PGC-1α-c transcripts via β2-AR activation, whereas high-intensity exercise or AICAR injection increased PGC-1α-a transcripts independently of β2-AR activation (Fig. 6, A and B). PGC-1α-a is transcripted from exon 1a; PGC-1α-b and PGC-1α-c are transcripted from an alternate, exon 1b (25).

Fig. 6.

Proposed model of PGC-1α gene expression in response to different intensity of exercise (A) or AICAR injection (B). Exercise activates the promoter for the expression of exon 1b via the sympathetic nervous system (SNS), and that magnitude is dependent on the intensity of the exercise. Pretreatment with β2-AR antagonist markedly inhibited activation of this promoter. Exon 1a is activated only by high-intensity exercise. α2-AMPK activation was not essential for either signaling. AICAR injection increases epinephrine secretion. Epinephrine activates β2-AR and increases PGC-1α-b and PGC-1α-c mRNA expressions. Also, AICAR increased exon 1a expression via β2-AR-independent pathways.

We show here that high-intensity exercise was required for an increase in PGC-1α-a expression by exercise and that such expression was not inhibited by β2-AR blockade (Fig. 2). During high-intensity exercise, α2-AMPK activity was increased in the skeletal muscle (Fig. 3A). Furthermore, treatment with AICAR, an activator of AMPK, in vivo and ex vivo induced PGC-1α-a expression (Figs. 4B and 5B). Although from these results it appeared that the high-intensity exercise-induced activation of α2-AMPK was required for PGC-1α-a expression in skeletal muscle, the high-intensity exercise-induced expression of PGC-1α-a was observed in α1-AMPK-DN mice (Fig. 3C), which showed a marked decrease in α2-AMPK activity (Fig. 3B), suggesting that other mechanisms rather than activation of α2-AMPK were involved in the exercise-induced expression of PGC-1α-a. Existence of other mechanisms was also suggested by an observation that medium-intensity exercise could induce phosphorylation of ACC substantially in α1-AMPK-DN mice, up to a level similar to that in wild-type mice in the sedentary state (Fig. 3B). Activation of sucrose nonfermenting AMPK-related kinase (SNARK) by exercise is a candidate for other mechanisms. Muscle contraction- and exercise-induced activation of SNARK, one of the AMPK-related kinases, was observed in mouse skeletal muscle, and muscle contraction-induced phosphorylation of AS160 or TBC1D1, which might be necessary for increase in glucose transport in skeletal muscle, was mediated by SNARK (19). AICAR treatment also could phosphorylate AS160 or TBC1D1 (37). Therefore, phosphorylation of ACC and expression of PGC-1α-a might also be increased by not only AMPK but also AMPK-related kinases such as SNARK. Another mechanism is also possible: CaMK might be involved in the expression of PGC-1α-a, because high-intensity exercise-induced phosphorylation was observed not only in AMPK but also in CaMK II, and these phosphorylations did not occur after low-intensity exercise (11). These results suggest that the high-intensity exercise-induced PGC-1α-a expressions are regulated by a redundant mechanism in which α2-AMPK is one of multiple signaling pathways.

In contrast, exercise-induced expression of PGC-1α-b or PGC-1α-c is mediated by β2-AR irrespective of intensity. β-AR activation increases not only in cAMP but also in Ca2+ concentration (28), which was observed in sympathetic nerve activation. In addition to β-AR activation, muscle contraction also increases in Ca2+ concentration during exercise (4, 31). However, only β2-AR-mediated change might be involved in exercise-induced PGC-1α-b and PGC-1a-c expression, because β2-AR blockade remarkably inhibited such expression in the present study (Fig. 2). AICAR-induced expression of PGC-1α-b or PGC-1α-c in vivo was probably due to an increase in the concentration of plasma epinephrine by direct stimulation of the adrenal glands by AICAR, as shown in PC12 cells (12) and/or by indirect stimulation via a reduction of blood glucose.

It is unlikely that inhibition of the exercise-induced expression of PGC-1α-b or PGC-1α-c by β2-AR antagonist injection before the exercise was due to whole body effects including change in sympathetic tone and the cardiovascular system, because incubation of isolated rat skeletal muscle with β2-AR agonist increased in PGC-1α-b mRNA (Fig. 5A). In addition, exercise performance was not impaired in the presence of β2-AR antagonist in our experimental conditions (data not shown). Contracting muscles are known to be exposed to much higher concentrations of catecholamines from the nerve endings than are noncontracting muscles; stretch (0.5 kg tension) increased muscle interstitial norepinephrine concentration 45% in the stretched leg and 7% in the control leg in rat (20). Therefore, it is expected that during exercise contracting mouse muscle is exposed to a higher concentration of norepinephrine and increased PGC-1α-b or PGC-1α-c expression compared with that in noncontracting muscle. Inhibition of exercise-induced expression of PGC-1α-b or PGC-1α-c by β2-AR antagonist was likely due to the direct effect on contracting muscle.

Extensive promoter analysis of PGC-1α-a has been conducted (2, 9, 10, 13, 45). There are NFAT (nuclear factor of activated T cell), myocyte enhancer factor (MEF)2, forkhead box O (FOXO)1, cAMP-responsive element-binding protein (CREB), and activating transcription factor (ATF)2 binding sites in the PGC-1α-a promoter (13). Among these sites, conserved MEF2 binding site and cAMP-responsive element (CRE) sequences regulate PGC-1α-a transcription in response to muscle contraction (1, 3). It was suggested that transcriptional activation of PGC-1α-a is induced by the interaction on the MEF2 binding site and CRE sequences on the PGC-1α-a promoter with regulatory factors such as MEF2, class IIa histone deacetylases (HDACs), ATF2, and CREB (1). AMPK, CaMK, CnA, and p38 MAPK have been reported to modulate these regulatory factors (2, 7, 9, 22, 29, 34, 35, 38, 44). Modulation of these transcription factors and the regulators by these protein kinases and phosphatase might be involved in high-intensity exercise-induced PGC-1α-a expression. Recently, it was reported that the effect of AICAR on transcription activation is mediated by the overlapping GATA/E-box binding site at −495 in the human PGC-1α-a promoter (16). Furthermore, newly synthesized PGC-1α protein can coactivate MEF2 and thus positively regulates its own transcription (13). AMPK phosphorylates PGC-1α protein at Thr177 and Ser538, which is required for the PGC-1α-dependent induction of the PGC-1α promoter (17). These results suggest that high-intensity exercise or AICAR-induced expression of PGC-1α-a might also be regulated by the GATA/E-box binding site and phosphorylation of PGC-1α protein by AMPK.

For exercise-induced expression of PGC-1α-b or PGC-1α-c, β2-AR-mediated regulation of CREB and its binding to the CRE sequence are important. In analysis of the PGC-1α-b promoter, CRE and E-box sequences are required for activation of human PGC-1α-b transcription (6, 46). The E-box sequence may contribute to the muscle-enriched expression, whereas the CRE sequence likely serves as a key target of exercise-initiated signaling pathways and mediates the induction of PGC-1α-b and PGC-1α-c expression in skeletal muscles. Expression of constitutively activated forms of CaMKIV and CnA or forskolin treatment, which increases the cytosolic concentrations of cAMP, increases PGC-1α-b promoter activity depending on the CRE sequence via activation of CREB (46). There is another putative CRE sequence in the PGC-1α-b promoter that is also essential for β2-AR-induced expression (6). In agreement with our in vivo finding, PGC-1α-b promoter is activated by CREB activation, but not by AICAR treatment, in the luciferase assay of C2C12 myoblasts (46).

AICAR stimulates PGC-1α-a expression via exon 1a, whereas β2-AR activation stimulates PGC-1α-b and PGC-1α-c expression via exon 1b. Exercise uses both pathways, but only high-intensity exercise uses exon 1a. It is unclear why PGC-1α-b and PGC-1α-c expression is much more sensitive to β2-AR activation than is PGC-1α-a expression, although both promoters possess a functional CRE sequence. There are two types of CRE sequences, the palindromic and variant CRE sequences. In the promoter of the human α-subunit gene, the variant CRE sequence binds the same proteins as does the palindromic CRE sequence but with lower affinity (14). A palindromic CRE sequence was found in the promoter for PGC-1α-a, and both the variant and palindromic (antisense) CRE sequences were found in the promoters for PGC-1α-b and PGC-1α-c (Fig. 7). These CRE sequences might cause the difference in sensitivity to the β2-AR stimuli. These findings suggest that the molecular mechanism by which exercise induces expression of PGC-1α might be more complex. Further studies of promoter analysis including the physiological role of the PGC-1α isoform are warranted.

Advertisements