Rosiglitazone (RSG) is an insulin-sensitizing thiazolidinedione (TZD) that exerts peroxisome proliferator-activated receptor-γ (PPARγ)-dependent and -independent effects. We tested the hypothesis that part of the insulin-sensitizing effect of RSG is mediated through the action of AMP-activated protein kinase (AMPK). First, we determined the effect of acute (30–60 min) incubation of L6 myotubes with RSG on AMPK regulation and palmitate oxidation. Compared with control (DMSO), 200 μM RSG increased (P < 0.05) AMPKα1 activity and phosphorylation of AMPK (Thr172). In addition, acetyl-CoA carboxylase (Ser218) phosphorylation and palmitate oxidation were increased (P < 0.05) in these cells. To investigate the effects of chronic RSG treatment on AMPK regulation in skeletal muscle in vivo, obese Zucker rats were randomly allocated into two experimental groups: control and RSG. Lean Zucker rats were treated with vehicle and acted as a control group for obese Zucker rats. Rats were dosed daily for 6 wk with either vehicle (0.5% carboxymethylcellulose, 100 μl/100 g body mass), or 3 mg/kg RSG. AMPKα1 activity was similar in muscle from lean and obese animals and was unaffected by RSG treatment. AMPKα2 activity was ∼25% lower in obese vs. lean animals (P < 0.05) but was normalized to control values after RSG treatment. ACC phosphorylation was decreased with obesity (P < 0.05) but restored to the level of lean controls with RSG treatment. Our data demonstrate that RSG restores AMPK signaling in skeletal muscle of insulin-resistant obese Zucker rats.
- peroxisome proliferator-activated receptor-γ
- Zucker rats
- adenosine monophosphate-activated protein kinase-α2
skeletal muscle insulin resistance is a common state associated with inactivity, aging, genetic predisposition, and environmental factors and is a hallmark feature of a variety of disease states, including obesity, hyperlipidemia, hypertension, and type 2 diabetes (31). Rosiglitazone (RSG) is a member of the thiazolidinedione (TZD) class of oral antidiabetic agents that improve insulin sensitivity in a range of insulin-resistant states (15, 18). TZDs are peroxisome proliferator-activated receptor-γ (PPARγ) agonists that, upon activation by fatty acids (FA) or FA-derived compounds, bind to responsive elements located in the promoter regions of many genes and modulate their transcriptive activities (1, 32). Although PPARγ plays a role in the TZD-induced insulin sensitization, metabolic responses to TZDs can be dissociated from PPARγ-induced gene transcription (4, 14, 29). Furthermore, we (23) have recently reported that RSG improves glucose tolerance by mechanisms other than a reduction of FA accumulation within skeletal muscle. Hence, it is likely that there exists PPARγ-independent mechanisms by which TZDs improve insulin sensitivity.
The AMP-activated protein kinase (AMPK) is a cellular energy sensor that regulates glucose and lipid metabolism by phosphorylating key regulatory enzymes (21, 39). AMPK activation causes many metabolic changes that would be beneficial to individuals with type 2 diabetes and the metabolic syndrome, including increased glucose uptake and metabolism by muscle, decreased glucose production by the liver and decreased synthesis, and increased oxidation of FA (16, 39). As TZDs increase insulin-stimulated glucose uptake and utilization and decrease circulating levels of free FA and triglycerides (for reviews, see Refs. 26 and 32), it is reasonable that a component of the insulin-sensitizing effect of these drugs could be mediated through the action of the AMPK pathway. In this regard, Fryer et al. (13) reported that acute incubation of H-2Kb muscle cells with RSG increased the AMP/ATP ratio and activated both α1- and α2-containing AMPK isoforms, which lead to a marked increase in the phosphorylation of acetyl-CoA carboxylase (ACC). Furthermore, AMPK has been reported to phosphorylate insulin receptor substrate-1 at Ser789 and enhance p85 docking, which is consistent with improving insulin signaling (19).
In the present study, we determined the effects of acute RSG administration in L6 muscle cells and chronic RSG treatment in skeletal muscle from obese, glucose-intolerant Zucker rats on AMPK regulation and its downstream phosphorylation target, ACC. We found that RSG treatment normalizes AMPKα2 activity in the skeletal muscle of obese Zucker rats compared with their lean littermates.
MATERIALS AND METHODS
Female obese Zucker and age-matched lean rats were obtained from Monash Animal Services (Victoria, Australia) at 19 wk of age. Rats were housed under controlled light (12:12-h light-dark cycle) at an ambient temperature of 21°C. Animals had free access to standard rat chow and water, except when overnight fasting was required for blood measurements. All procedures were approved by the Royal Melbourne Institute of Technology University Animal Ethics Committee. Obese rats were randomly divided into two experimental groups: control (OB CON, n = 9) and RSG treated (OB RSG, n = 9). The lean rats (LN CON, n = 9) were treated with vehicle and acted as a control group for the OB CON rats. The rats were dosed daily, for 6 wk by oral gavage with either vehicle, which consisted of 0.5% carboxymethylcellulose (CMC; 100 μl/100 g body mass), or 3 mg/kg RSG (GlaxoSmithKline, Stevenage, UK) suspended in an equal volume of CMC. After the 6-wk experimental period, rats were fasted overnight and euthanized by CO2 asphyxia, followed by exsanguination. The time interval between CO2 administration and muscle harvesting was ∼1–2 min, and the interval between the last RSG treatment and tissue collection was ∼24 h. The red gastrocnemius muscles were rapidly dissected from the rat hindlimbs, snap-frozen in liquid nitrogen, and stored at −80°C for later analysis.
Muscle metabolite and adenine nucleotide contents.
Red gastrocnemius muscle was freeze-dried, powdered, and cleaned of all visible connective tissue and blood under magnification. An ∼2-mg aliquot was then taken and extracted in 0.5 M HClO4 (1 mM EDTA) and neutralized with 2.2 M KHCO3. Muscle ATP, phosphocreatine (PCr), creatine, and lactate were measured, and muscle ADP and AMP were calculated as described for cells. Muscle for glycogen analysis was extracted in 2 M HCl and neutralized with 0.67 M NaOH, and glycogen content was determined by fluorometric techniques (27).
AMPK activity and AMPK subunit protein expression.
Approximately 80 mg of wet muscle were homogenized in buffer A (50 mM Tris·HCl, pH 7.5, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 50 mM NaF, 5 mM Na pyrophosphate, 10% glycerol, 1% Triton X-100, 10 mg/ml trypsin inhibitor, 2 mg/ml aprotinin, 1 mM benzamidine, and 1 mM phenylmethylsulfonyl fluoride). The muscle homogenate was then centrifuged at 20,000 g for 25 min, and the supernatant was collected. The supernatant was aliquoted for determination of 1) protein concentration (Pierce, Rockford, IL), 2) immunoprecipitation (AMPKα1 and AMPKα2), 3) ACC affinity purification, and 4) protein expression by Western blotting [uncoupling protein-3 (UCP3)]. Aliquots were stored at −80°C until use, and immunoprecipitates were used to determine protein content and enzyme activity (AMPKα1 and AMPKα2, described below).
Approximately 5 mg of protein from the supernatants were incubated with AMPKα1 or AMPKα2 antibody-bound protein A-Sepharose beads (6 MB; Amersham Biosciences, Uppsala, Sweden) for 2 h at 4°C. The polyclonal antipeptide antibodies to AMPKα1 and AMPKα2 were raised to nonconserved regions of the AMPKα1 (rat 231–251) and AMPKα2 isoforms (rat 351–366). Immunocomplexes were washed with PBS and suspended in 50 mM Tris (pH 7.5) buffer for the AMPK activity assay (9). The AMPK activities in the immune complexes were measured in the presence of 200 μM AMP. Activities were calculated as picomoles of phosphate incorporated into the serine-alanine-methionine-serine (SAMS) peptide per minute per milligram of protein subjected to immunoprecipitation (pmol·min−1·mg−1). ACC was affinity purified by incubating the post-AMPK immunoprecipitation supernatants in streptavidin-Sepharose high-performance beads (Amersham Biosciences) for 1 h at 4°C.
The affinity-purified ACC fraction was electrophoresed on 7.5% SDS-PAGE and detected by immunoblotting with anti-phospho-Ser218-ACC polyclonal antibody (8). The blots were then stripped (50 mM Tris, 2% SDS, 115 mM β-mercaptoethanol) at 55°C for 20 min, blocked, and incubated with horseradish peroxidase-conjugated streptavidin (Amersham Pharmacia Biotech UK, Little Chalfont, UK) to determine total ACC. Aliquots of immunoprecipitated AMPKα1 and AMPKα2 proteins were electrophoresed on 10% SDS-PAGE and detected by immunoblotting with anti-phospho-Thr172 antibody and raised against AMPKα peptide (KDGEFLRpTSCGSPNY) as described previously (10). The membranes were then stripped (as described above), washed in PBS, blocked for 60 min in 5% skim milk powder-PBS, and reprobed with antibodies specific for the AMPKα1 and AMPKα2 isoforms. Aliquots containing 80 μg of total protein were electrophoresed on 15% SDS-PAGE and detected by immunoblotting with anti-UCP3 antibody (Alpha Diagnostic, San Antonio, TX).
L6 myoblasts were maintained at 37°C (95% O2-5% CO2) on 60-mm collagen-coated plastic dishes in α-modified Eagle's medium (α-MEM) containing 10% fetal bovine serum (FBS) culture medium and 1% penicillin-streptomycin and 5 mM glucose. Differentiation was induced by switching to medium containing 2% horse serum when the myoblasts were ∼90% confluent. Experimental treatments were started after 4 days, by which time nearly all of the myoblasts had fused to form myotubes.
To determine the effect of RSG treatment on AMPK activity, AMPK Thr172 phosphorylation, and ACC Ser218 phophorylation, L6 myotubes were incubated with 0 (DMSO), 5, 50, or 200 μM RSG for 30 min at 37°C. After incubation, cells were rinsed three times with PBS and lysed with 200 μl of buffer A. Lysates were immediately frozen in liquid N2 and kept at −80°C until further analysis. AMPKα1 and AMPKα2 subunits and ACCβ protein were isolated by immunoprecipitation and affinity chromatography, respectively, as described above. AMPK activity was determined using the SAMS peptide assay, and protein expression and phosphorylation were determined as described above.
To examine the effects of RSG on fat oxidation, cells were grown in 60-mm plates and incubated for 1 h in the absence (DMSO, n = 6) or presence of 5 μM (n = 6) or 200 μM (n = 6) RSG in 3 ml of medium consisting of α-MEM, 2% BSA, and 0.5 mM [14C]palmitate (Amersham Biosciences, Castle Hill, Australia). After the incubation period, the reaction was stopped, and 2 ml of medium were added to an equal volume of 1 M acetic acid, and the released 14CO2 was collected in benzothonium hydroxide. The cells were rinsed twice with PBS, methanol was added to the plate, and cells were scraped for subsequent analysis of the acid-soluble metabolite (ASM) fraction. The 14C content of both fractions was determined using scinitillation counting, and the oxidation rates (CO2 + ASM) were calculated.
L6 myotubes were incubated with DMSO (vehicle), 200 μM RSG, 500 μM dinitrophenol (DNP), or 200 μM 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR) for 30 min. The medium was removed, and the cells were rinsed three times with PBS, lysed with 100 μl of 0.5 M HClO4 (1 mM EDTA), and neutralized with 2.2 M KHCO3. This extract was used for the determination of adenosine triphosphate (ATP), PCr, creatine, and lactate by enzymatic fluorometric assays (2). Free ADP and AMP concentrations were calculated with the assumption of equilibrium of the adenylate kinase and creatine kinase reactions (12). Free ADP was calculated using the measured ATP, creatine, and PCr values, an estimated H+ concentration (30), and the creatine kinase equilibrium constant of 1.66 × 109. Free AMP concentration was calculated from the estimated free ADP and the measured ATP with the adenylate kinase constant of 1.05.
Results are presented as means ± SE. Statistical analysis of skeletal muscle measurements was performed using an unpaired Student's t-test. Analysis of cell culture experiments was undertaken using one-way ANOVA. A P value of <0.05 was considered significant. When ANOVA revealed significant differences, a Neuman-Keuls post hoc test was used to locate such differences.
Effects of acute incubation of muscle cells with RSG.
Table 1 displays the AMP/ATP ratio of L6 myotubes after incubation with DMSO (control), 200 μM RSG, 500 μM DNP, or 200 μM AICAR for 30 min. Incubation with DNP caused a 2.4-fold increase in the AMP/ATP ratio (P = 0.02 vs. control). No change in AMP/ATP was observed after RSG or AICAR treatment.
Figure 1 shows the effect of varying concentrations of RSG on the activity of AMPKα1 (Fig 1A) and Thr172 phosphorylation of AMPKα1 (Fig. 1B), whereas Fig. 2 displays the ACC Ser218 phosphorylation (Fig. 2A) and palmitate oxidation (Fig. 2B) in L6 myotubes after acute incubation. The dose-response trend of increasing AMPKα1 activity was significant at a dose of 200 μM RSG (P < 0.05) and corresponded with similar increases in Thr172 phosphorylation of the AMPKα1 isoform (Fig. 1B). Increased AMPKα1 activity in L6 myotubes at the highest dose of RSG was also associated with a threefold increase in the ACC Ser218 phosphorylation compared with control (0 vs. 200 μM RSG, P < 0.05; Fig. 2A) and a significant increase in palmitate oxidation (0 vs. 200 μM RSG, 391 ± 20 vs. 455 ± 19 pmol/h, respectively; Fig. 2B). AMPKα2 activity in L6 myotubes was undetectable.
Effects of chronic RSG treatment in skeletal muscle.
Table 2 displays the resting nucleotide and metabolite concentrations for the three groups under investigation. There were no differences in ATP, AMP, or ADP concentrations between muscle from LN CON or OB CON animals or OB RSG muscle. Consequently, there were no changes in the AMP/ATP ratio for any group (Table 2). Resting concentrations of PCr, creatine, and lactate were also similar between groups. However, glycogen content was elevated in muscle from obese compared with lean animals (P < 0.05), and this remained high in RSG-treated rats.
The activities of the AMPKα1 and AMPKα2 isoforms, along with the ACC Ser218 phosphorylation from skeletal muscle of animals treated for 6 wk with RSG, are shown in Fig. 3. The activity of AMPKα1 was similar in lean and obese animals and was unaffected by 6 wk of RSG treatment (Fig. 3A). By contrast, the activity of AMPKα2 was twofold lower in obese than in lean animals (7.45 ± 1.0 vs. 15.7 ± 2.2 pmol·mg−1·min−1, P < 0.05) but was restored to LN CON values after RSG treatment (OB RSG, 14.2 ± 1.5 pmol·mg−1·min−1; Fig. 3B). The observed changes in AMPKα2 activity were most likely due to corresponding changes in protein expression rather than to changes in phosphorylation of the protein at the Thr172 residue, which were similar in all groups (data not shown). Altered AMPKα2 activity with obesity and after RSG treatment paralleled changes in the phosphorylation state of its target protein, ACC. The ratio of phosphorylated (p-)ACC to total ACC protein (p-ACC/total ACC) was significantly decreased in muscle from obese animals (P < 0.05) but was restored to LN CON levels after RSG treatment (Fig. 3C). AMPKα1 protein expression tended to be higher (P = 0.06) in muscle from OB CON than from LN CON animals but was significantly decreased after RSG treatment (P < 0.05, OB CON vs. OB RSG; Fig. 4A). There were reciprocal changes in the protein expression of the AMPKα2 isoform, which was lower in muscle from obese animals but restored to LN CON values after chronic RSG treatment (Fig. 4B).
Figure 5 displays the protein levels of UCP3 in skeletal muscle of lean and obese animals after RSG treatment. UCP3 protein abundance was not different between lean and obese animals. However, RSG treatment resulted in a significant increase in UCP3 protein levels (P < 0.05).
TZDs are known to improve insulin sensitivity by activating PPARγ and inducing adipogenesis in adipose tissue (11) but also via PPARγ-independent mechanisms. Here, we report that RSG increases AMPK, leading to phosphorylation of ACC in vitro and in vivo.
In the present investigation, we observed maximal activation of AMPK in L6 myotubes when treated with 200 μM RSG, due to an increase in the activity of the AMPKα1 isoform. Moreover, the increased AMPKα1 activity at the highest dose of RSG was associated with a marked increase in the phosphorylation of ACC and a concomitant increase in palmitate oxidation. Recently, Cha et al. (6) reported that cultured muscle cells from type 2 diabetic patients had lower rates of basal β-oxidation but that this defect was normalized after chronic (4 days) TZD treatment. Accordingly, an AMPK-stimulated increase in lipid oxidation might be one mechanism by which RSG exerts its insulin-sensitizing effect.
Only one previous study has examined the effect of RSG on AMPK activity in skeletal muscle myotubes. Fryer et al. (13) observed that treating muscle cells derived from heterozygous H-2Kb-tsA58-transgenic mice with RSG leads to increases in the AMP/ATP ratio and a concomitant increase in AMPK activity. We, too, observed an increase in AMPK activity in RSG-treated cells. However, we did not detect any difference in the energy charge (AMP/ATP) of the cell, suggesting that RSG's action was exerting an energy charge-independent effect on AMPK. The reasons for the differences in results between the present and previous studies (13), in terms of the AMP/ATP ratio, are not readily apparent. The H-2Kb cell line contains both AMPKα1 and AMPKα2 isoforms (13), and we cannot rule out the possibility that this is a more sensitive cell line to treatments that affect energy charge. However, it should be noted that the H-2Kb lines are immortalized cells derived from the simian virus 40 large tumor antigen transgenic mouse (20) that might have some limitations with respect to their physiological significance. To be consistent in the present investigation, we chose to use L6 myotubes derived from rat skeletal muscle so that the acute effects of RSG might be compared with the chronic effects of this treatment regimen in obese Zucker rats.
Previous studies (17, 24, 34, 40) have suggested that altered skeletal muscle AMPK activity may not play a role in obesity and diabetes. However, a recent study (7) found that, although basal AMPK activity was not altered, AMPK activation was impaired in cultured human skeletal muscle from obese type 2 diabetics. There is also evidence that basal AMPK phosphorylation is reduced in the cardiac muscle of both Zucker diabetic fatty rats and ob/ob mice (37) and that this defect is normalized by TZD treatment. Further study is clearly needed to draw general conclusions on the role of skeletal muscle AMPK defects in obesity and diabetes.
To date, no studies have determined the effects of chronic RSG treatment on the responses of AMPK in skeletal muscle or, more importantly, the effect of RSG on skeletal muscle from insulin-resistant animals. Our finding that chronic RSG treatment increased both AMPKα2 and ACC phosphorylation in obese Zucker rats is of major clinical significance, given the widespread use of TZDs as an antidiabetic drug. The altered AMPKα2 activity, with obesity and RSG treatment, corresponded to changes in the phosphorylation state of its target protein, ACC, which was decreased with obesity but restored to the level of lean controls after RSG administration. Of interest was the reciprocal response of the AMPK protein levels to the chronic RSG administration. AMPKα2 protein was reduced in obesity but restored to lean control values after RSG treatment, whereas AMPKα1 protein abundance was suppressed with treatment. Consistent with our in vitro cellular data, the muscle nucleotide concentrations were unaltered by chronic RSG treatment in vivo (Table 2).
Because hypoxia stimulates AMPK activity and a catecholamine response (3, 15), the combination of euthanasia by CO2 asphyxiation and the elapsed time between tissue collection and freezing might have resulted in an altered intracellular environment that could have contributed to changes in AMPK activity. These factors are potential limitations of the present study. However, we are confident that the observed AMPK activation in the present study was not due to hypoxia for the following reasons. First, muscle metabolites were within the expected ranges for all treatment groups and were similar to those obtained by other groups studying obese Zucker rats (22, 29). Second, the observed changes in AMPK activation were paralleled by changes in AMPK protein levels (Fig. 4) and were not associated with changes in the Thr172 phosphorylation of AMPK. This suggests that the increase in activity was due to chronically altered protein levels rather than to acute changes in cellular energy status. Finally, the time taken to extract muscle was not different between groups; therefore, any differences in AMPK activity are indicative of the treatment.
The improved metabolic action, as a consequence of upregulating AMPKα2 by RSG, might be due to an increase in fat oxidation, because increased β-oxidation in muscle cells has been shown (28) to enhance insulin-stimulated glucose metabolism and protect against insulin resistance in the face of elevated intramyocellular lipid content. Future studies should determine whether the observed increase in skeletal muscle AMPK activity with chronic RSG treatment is associated with a concomitant increase in fat oxidation in vivo. However, it is also possible that RSG-mediated upregulation of AMPKα2 might improve metabolic action via alternative mechanisms. Treatment with either RSG (5) or AICAR (35, 36) increases UCP expression in skeletal muscle. In addition, UCP3 overexpressing transgenic mice display enhanced glucose tolerance and palmitate oxidation (38), a phenotype that is compatible with chronic TZD treatment. In the present investigation, we found that UCP3 protein expression was upregulated by chronic RSG treatment, suggesting that this protein may account for part of its metabolic action in skeletal muscle.
Previously, Musi et al. (25) reported that the antidiabetic drug metformin (biguanide) increased AMPKα2 activity in type 2 diabetic patients after a 10-wk treatment regimen. A significant increase in Thr172 phosphorylation was observed, but there was no significant increase in AMPKα2 levels in the biopsy samples. The increase in AMPKα2 activity seen in their study was paralleled by decreases in ACC activity consistent with AMPK mediated phosphorylation. Previously, it has been suggested that metformin and RSG activate AMPK by different pathways (13), with RSG causing an increase in AMP levels. Irrespective of the mechanism of activation, both drugs (metformin, human type 2 diabetics; RSG, obese Zucker rats) enhance AMPK signaling via the α2-isoform and inactivate ACC.
In conclusion, the results from the present study demonstrate that chronic RSG administration stimulated isoform-specific regulation of the AMPK signaling cascade in skeletal muscle from obese Zucker rats and in myotubes in vitro. The increased basal activity of AMPKα2 paralleled changes in the phosphorylation state of its target protein, ACCβ, and occurred without detectable changes in cellular energy charge. These changes in AMPKα2 seem likely to be an important component of the insulin-sensitizing effect of RSG that contributes to the improved metabolic action associated with chronic TZD drug administration. Future studies should determine the mechanisms underlying altered transcriptional regulation and/or turnover of AMPKα2 protein in skeletal muscle from obese Zucker rats and how this is reversed by TZD treatment.
This study was supported by grants from the National Health and Medical Research Council of Australia (NHMRC), Australian Research Center (ARC), and National Heart Foundation. M. J. Watt is an NHMRC Peter Doherty postdoctoral fellow, M. A. Febbraio is an NHMRC Senior Research Fellow, and B. E. Kemp is an ARC Federation Fellow.
We acknowledge Winnie Lau for technical assistance. We also thank GlaxoSmithKline (Stevenage, UK) for the gift of rosiglitazone.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by American Physiological Society