AMP-activated protein kinase (AMPK) is emerging as a key signaling pathway that modulates cellular metabolic processes. In skeletal muscle, AMPK is activated during exercise. Increased myocardial substrate metabolism during exercise could be explained by AMPK activation. Although AMPK is known to be activated during myocardial ischemia, it remains uncertain whether AMPK is activated in response to the physiological increases in cardiac work associated with exercise. Therefore, we evaluated cardiac AMPK activity in rats at rest and after 10 min of treadmill running at moderate (15% grade, 16 m/min) or high (15% grade, 32 m/min) intensity. Total AMPK activity in the heart increased in proportion to exercise intensity (P < 0.05). AMPK activity associated with the α2-catalytic subunit increased 2.8 ± 0.4-fold (P < 0.02 vs. rest) and 4.5 ± 0.6-fold (P < 0.001 vs. rest) with moderate- and high-intensity exercise, respectively. AMPK activity associated with the α1-subunit increased to a lesser extent. Phosphorylation of the Thr172-regulatory site on AMPK α-catalytic subunits increased during exercise (P < 0.001). There was no increase in Akt phosphorylation during exercise. The changes in AMPK activity during exercise were associated with physiological AMPK effects (GLUT4 translocation to the sarcolemma and ACC phosphorylation). Thus cardiac AMPK activity increases progressively with exercise intensity, supporting the hypothesis that AMPK has a physiological role in the heart.
- energy metabolism
- signal transduction
amp-activated protein kinase (AMPK) is a serine-threonine kinase that is activated by energetic stress and has an important role in the regulation of cellular metabolism (22, 30). AMPK is activated in the ischemic heart (31) and under hypoxic conditions in skeletal muscle (26). AMPK is also known to undergo physiological activation during contraction and exercise in skeletal muscle (51, 53, 54). However, it is not known whether heart AMPK is activated during exercise or the extent to which AMPK might have physiological importance in the normal heart.
Exercise is known to increase the uptake and utilization of free fatty acids and glucose in the heart (18, 52), as it does in skeletal muscle (54). During exercise, heart contractile function and metabolism are regulated in a complex fashion by a number of signaling pathways, including adrenergic receptor and calcium-activated mechanisms. In skeletal muscle, activation of AMPK appears to have a role in mediating the enhanced glucose uptake (36) and fatty acid oxidation (48) that occur with increased contractile activity, although additional pathways may be involved. However, in the heart, it remains uncertain whether AMPK has a role in mediating the cardiac response to exercise.
When activated pharmacologically with the compound 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside, increased free fatty acid and glucose uptake occur in skeletal muscle (35) and heart (4). One of the best characterized downstream targets of AMPK is acetyl-CoA carboxylase (ACC; see Refs. 8, 12, and 53). The ACC-1 and ACC-2 isoforms both contain serine residues that are phosphorylated by AMPK (2, 14). ACC phosphorylation inhibits malonyl-CoA synthesis, enhancing carnitine palmitoyltransferase I activity and free fatty acid oxidation (31, 32, 54). AMPK may also modulate free fatty acid oxidation through the activation of malonyl-CoA decarboxylase (43). In addition, AMPK increases cellular glucose uptake through translocation of the GLUT4 transporter to the sarcolemma in skeletal muscle (33) and heart (42). Although the downstream targets of AMPK responsible for GLUT4 translocation are unknown, AMPK signaling (7, 27, 42) is distinct from the phosphatidylinositol 3-kinase (PI 3-kinase) pathway that mediates insulin-activated glucose transport. AMPK also increases glucose utilization through activation of 6-phosphofructo-2-kinase, which leads to the production of fructose 2,6-bisphosphate, an activator of glycolysis (34).
AMPK is a heterotrimeric complex comprised of a catalytic α-subunit as well as regulatory β- and γ-subunits (23, 28, 46). In most tissues, including the heart and skeletal muscle, there are two isoforms of the catalytic subunit, α1 and α2. In skeletal muscle, there is evidence that the two α-isoforms may be differentially activated during contraction and exercise, with a greater degree of activation of α2 (9, 17, 48, 58). These findings are consistent with the observation that the α2-isoform may have a greater dependence on the cellular AMP concentration (45). Although both α-isoforms are activated during ischemia (10), the response of the heart AMPK α-isoforms to exercise is unknown.
Therefore, the objectives of the present study were 1) to determine whether cardiac AMPK activity increases during exercise; 2) to examine whether in vivo activation of AMPK in the heart depends on the exercise intensity; 3) to evaluate the degree of activation of the two α-isoforms of AMPK in the heart in response to exercise; and 4) to relate activation of AMPK during exercise with potential downstream actions, including phosphorylation of ACC and GLUT4 translocation.
Male Sprague-Dawley rats weighing 250-350 g were housed in a pathogen-free facility on a 12:12-h light-dark cycle and were allowed access to standard chow and water ad libitum before experiments. Rats were habituated to a treadmill for 5 days before placement of an indwelling jugular venous polyethylene catheter (29) and then were allowed to recover for 72 h. On the day of the experiment, rats were randomly assigned to rest or treadmill exercise with either moderate-(16 m/min, 15% grade, ∼40% of V̇O2 max) or high-intensity (32 m/min, 15% grade, ∼80% of V̇O2 max) running. Rats were anesthetized, either while resting or during active exercise, by the intravenous injection of pentobarbital sodium, the hearts were rapidly excised, and the ventricles were frozen while beating in aluminum clamps cooled in liquid nitrogen. All procedures were approved by the Yale University Animal Care and Use Committee.
Tissue homogenization. Myocardial tissue samples were homogenized in Tris buffer (125 mM Tris, 10 mM EDTA, 10 mM EGTA, 250 mM mannitol, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 1 mM DTT, 1 mM benzamidine, 0.004% trypsin inhibitor, and 3 mM sodium azide, pH 7.5) with a Polytron as previously described (31, 42). The homogenate was centrifuged, and the supernatant was subjected to polyethylene glycol (PEG) precipitation. The 2.5-6% PEG fraction was suspended in homogenization buffer without mannitol for measurement of AMPK activity. All procedures were performed at 4°C. Protein concentration was determined spectrophotometrically using the Bio-Rad reagent.
AMPK immunoprecipitation. The AMPK activity associated with specific α-subunit isoforms was examined after immunoprecipitation with polyclonal sheep antibodies raised against synthetic peptides derived from the α1 (TSPPDSFLDDHHLTR)- or α2 (MDDSAMHIPPGLKPH)-isoforms (45, 59). Antibodies were prebound in excess to protein G-Sepharose beads before incubation with 50 μg of PEG-precipitated homogenate overnight at 4°C. The immunoprecipitates were washed extensively with homogenization buffer containing 0.1% Igepal, and then with assay buffer (40 mM HEPES, 80 mM NaCl, 8% glycerol, and 0.8 mM EDTA, pH 7.0) before kinase assay.
AMPK assay. AMPK activity was measured using the SAMS peptide HMRSAMSGLHLVKRR phosphorylation assay. PEG extracts or α-subunit immunoprecipitates were resuspended in homogenization buffer containing 0.8 mM DTT and 0.2 mM AMP, with or without 0.2 mM SAMS peptide (see Refs. 31, 42, and 45). The kinase assay was performed in the presence of 5 mM MgCl2, 0.2 mM ATP, and [32P]ATP (New England Nuclear, Boston, MA) for 10 min at 37°C. Aliquots of the reaction mixture supernatant were spotted on Whatman filter paper (P81). The filters were washed with cold 150 mM phosphoric acid for 40 min and with acetone for 20 min and then were allowed to dry before scintillation counting. AMPK activity was calculated as picomoles per milligram PEG-precipitated protein per minute, and results were expressed as the degree of increase compared with resting rats.
Immunoblotting. Western blot analyses of α-isoforms of AMPK, pThr172 AMPK, and phosphorylated ACC (pACC) were performed on 40 μg of PEG-precipitated heart protein (2.5-6% fraction) after SDS-PAGE on 8% polyacrylamide for AMPK and 5% gels for pACC. AMPK immunoprecipitates were suspended in sample buffer, boiled for 10 min, and then used for SDS-PAGE. Proteins were subjected to electrophoresis and transferred to PVDF membranes. For immunoblotting, antibodies were diluted as follows: anti-α1 (45) at 1:2,000, anti-α2-AMPK (45) at 1:3,000, anti-pan-α-AMPK (Cell Signaling, Beverly, MA) at 1:5,000, anti-pThr172 AMPK (Cell Signaling) at 1:5,000, anti-pACC that recognizes both the Ser79 of ACC-1 and the equivalent Ser218 of ACC-2 (Upstate, Waltham, MA) at 1:5,000, anti-pThr308 Akt (Upstate) at 1:1,000, and anti-pSer473 Akt and anti-pan Akt1/2 (Cell Signaling) at 1:1,000.
Membrane fractionation. Membrane fractions were prepared from hearts as described previously (42, 61). In brief, crude homogenates were prepared with a Polytron. The supernatant containing the crude membrane fraction was pelleted by ultracentrifugation, and the membrane fractions were separated on a discontinuous sucrose gradient (25, 30, and 35% wt/vol) at 150,000 g for 20 h. The sarcolemma and the intracellular membranes were harvested and stored at -80°C. GLUT4 immunoblots were performed on 40-μg membrane protein in low-ionic-strength Laemli buffer on 8% polyacrylamide gels.
Measurement of high-energy phosphates and glycogen. Heart samples were extracted with 6% perchloric acid, and the supernatants were neutralized with 3 M K2CO3. Myocardial nucleotide contents were measured by reverse-phase HPLC, whereas creatine phosphate was measured using spectrophotometric methods, as previously described (6, 7). Glycogen was measured after KOH extraction and ethanol precipitation as previously described (6, 7). Results are expressed as micromoles per gram wet weight.
Statistical analysis. All data are reported as means ± SE. The number of rats in each group is presented in Figs. 1, 2, 3, 4, 5, 6, 7. Data were analyzed by ANOVA, and contrasts were used for planned comparisons between groups using Statistical Analysis Software (SAS Institute, Cary, NC). Differences were considered significant at P < 0.05.
Total AMPK activity in heart during exercise. We first measured total AMPK activity in PEG-precipitated heart homogenates from rats that remained resting and those that ran for 10 min at either moderate or high intensity. AMPK activity increased 1.6 ± 0.1-fold (P < 0.02) in the moderate-intensity group and 2.1 ± 0.2-fold (P < 0.001) in the high-intensity group compared with resting rats (Fig. 1). This AMPK activation was significantly greater at the faster treadmill speed (P < 0.03), indicating that there was a graded response of cardiac AMPK to increasing exercise intensity.
Heart α-isoform AMPK activities during exercise. To examine whether there was differential activation of the catalytic subunits of AMPK, SAMS kinase activity was also measured in α1- or α2-immunoprecipitates from hearts after exercise. Immunoprecipitation procedures were isoform specific with no demonstrable cross-immunoreactivity of precipitated proteins when subjected to immunoblotting (Fig. 2). AMPK activity in α1-immunoprecipitates tended to increase after both moderate (2.1 ± 0.6-fold)- and high (1.9 ± 0.4-fold)-intensity exercise compared with that in resting rats (Fig. 3). However, more pronounced increases in AMPK activity were observed in cardiac α2-immunoprecipitates: 2.8 ± 0.4-fold after moderate-intensity exercise (P < 0.02) and 4.5 ± 0.6-fold after high-intensity exercise (P < 0.001; Fig. 3). The increase in heart α2 activity was significantly greater (P < 0.02) in rats exercising at the faster treadmill speed, indicating that there was a graded activation of the α2-isoform in the heart with exercise. In addition, high-intensity exercise increased the activity of α2-AMPK to a significantly greater extent than α1 activity (P < 0.05), suggesting that this isoform is more responsive to the physiological stress of exercise in the heart, as it is in skeletal muscle (17, 48).
AMPK phosphorylation. AMPK activity measured with the SAMS kinase assay largely reflects the extent to which the α-subunit was phosphorylated in vivo, since the assay conditions do not replicate the in vivo concentrations of AMP that were present in the heart during exercise. The primary phosphorylation site responsible for regulating AMPK activity is the Thr172 residue of both the α1- and α2-catalytic subunits (22, 23, 25). Thus we assessed the degree of phosphorylation of Thr172, using a phosphopeptide-specific antibody, and expressed the amount of pThr172 relative to the total amount of AMPK in the samples. There was a twofold increase (P < 0.001) in the pThr172 content in PEG precipitates of hearts from rats running at high intensity (Fig. 4); no significant increase was apparent after moderate-intensity exercise.
Recent evidence suggests that cellular stress may activate noninsulin receptor-linked PI 3-kinase (13) and its downstream kinase Akt (44). To examine whether AMPK activation might be associated with Akt activation, we evaluated the effects of exercise on Akt phosphorylation by immunoblotting heart homogenates with antibodies against the pThr308 and pSer473 residues of Akt. In contrast to the increased phosphorylation of AMPK observed during exercise, there was no apparent increase in either Thr308 or Ser473 Akt phosphorylation (Fig. 5).
Total heart nucleotide and glycogen content. The activity of AMPK is modulated by several factors in muscle tissues. Increases in the ratio of the cellular contents of AMP/ATP are known to activate AMPK through enhanced phosphorylation, decreased dephosphorylation, and allosteric activation of the α-subunit (22, 23). The total contents of adenine nucleotides were measured in neutralized acid extracts of freeze-clamped hearts, but no significant changes in either total ATP or AMP were apparent after exercise (Table 1). These measurements do not exclude the possibility that increases in the free concentration of AMP might have occurred. Free AMP concentration typically increases when creatine phosphate decreases (15), as it did after both moderate- and high-activity exercise (Table 1). In addition, the cardiac glycogen content fell in proportion to exercise activity, with 31 and 62% reductions after 10 min of moderate- and high-intensity exercise, respectively (Table 1).
Downstream effects of AMPK activation during exercise. Free fatty acid oxidation is an important source of ATP generation in the heart during exercise (52). We assessed whether the increase in AMPK during exercise is associated with downstream phosphorylation of ACC, a key mediator of free fatty acid oxidation in the heart (31, 32). Although both AMPK and protein kinase A (PKA) phosphorylate multiple sites on ACC (11, 20), both Ser79 on ACC-1 and the equivalent Ser218 on ACC-2 are phosphorylated by AMPK but not by PKA. Immunoblots of PEG heart precipitates with an antibody that recognizes these specific phosphorylated serine residues showed a twofold increase in phosphorylation of both ACC-1 (265 kDa) and ACC-2 (280 kDa) after exercise (Fig. 6).
AMPK is also known to increase glucose uptake in both heart (42) and skeletal (7, 16, 27, 35) muscle by translocating GLUT4 transport proteins to the sarcolemma (33, 42). To evaluate whether AMPK activation was associated with GLUT4 translocation, we immunoblotted GLUT4 in sarcolemma and intracellular membranes from hearts after rest or high-intensity exercise. After high-intensity exercise, there was an increase in the sarcolemma GLUT4 content and a reduction in the intracellular membrane GLUT4 content, indicating the translocation of transporters during exercise (Fig. 7).
These experiments examined whether AMPK is activated during the acute physiological stress of exercise in normal rats during treadmill running. Exercise increased the total AMPK (SAMS kinase) activity in the heart, as well as that associated with both of the α-isoforms of the catalytic subunit of AMPK. The α2-isoform showed more pronounced activation than the α1-isoform and also had a graded response to exercise intensity, increasing progressively with higher treadmill speed. Exercise also increased the degree of phosphorylation of Thr172, a key site on the α-subunits that determines their activity and is phosphorylated by the upstream AMPK kinase. We also observed increases in AMPK-modulated downstream pathways important in the regulation of heart substrate metabolism, specifi-cally ACC phosphorylation and GLUT4 transporter translocation to the sarcolemma. Thus these results demonstrate for the first time that activation of AMPK occurs during acute exercise in the heart and suggest that AMPK may have a physiological role in the normal heart.
AMPK is activated by exercise in skeletal muscle, where it has an important role in regulating substrate metabolism (30, 54). Although AMPK is known to be activated in the ischemic (31, 32) and hypertrophied heart (49), it has been unclear whether AMPK serves as a physiological regulator in the normal heart. The current findings indicate that heart AMPK is activated during normal exercise. In previous experimental models, AMPK activation was not evident when cardiac workload was manipulated by dobutamine infusion in pigs (21) or by increasing afterload in isolated working rat hearts (5). However, confounding comparisons with the current results were the effects of surgery in the former study (21) and the relatively lower workloads examined in the latter report (5). During exercise, several determinants of cardiac work increase, including blood pressure, heart rate, and cardiac contractility, although it is possible that additional factors may be involved in the exercise activation of AMPK in vivo.
These results also indicate a greater degree of activation of the α2-compared with the α1-isoform in the heart during exercise. The α2-associated activity increased 2.8-fold during moderate-intensity and 4.5-fold during high-intensity exercise. Although treadmill running also appeared to be associated with a trend to an increase in α1-AMPK activity, it was less consistent than that in α2-AMPK activity. These observations parallel those in skeletal muscle with regard to greater α2 activation during moderate acute exercise (17, 48, 58). However, activation of the α1-isoform has been shown only after sprinting in humans (9) and during electrically stimulated contraction of isolated rat skeletal muscles (26). Thus, although α1 may be somewhat more readily activated in heart than in skeletal muscle during exercise, heart α2 is significantly more responsive than α1. The role of each of these isoforms in modulating the cardiac exercise response remains to be determined.
AMPK activity is modulated by the phosphorylation state of the Thr172 residue of the α-subunits, which is determined by the activities of upstream AMPK kinase(s) and protein phosphatases (22-24). The Thr172 site lies within the critical kinase activation domain of the α-catalytic subunit (23). We observed a significant increase in phosphorylation of Thr172 in heart homogenates from rats run at high intensity, using a phosphopeptide-specific AMPK antibody directed against this domain. Similarly, exercise appears to increase Thr172 phosphorylation in skeletal muscle after exercise (48). We did not detect a significant increase in Thr172 phosphorylation in hearts from rats run at moderate-intensity exercise, despite a modest increase in AMPK activity. This most likely reflects the lesser sensitivity of the phosphopeptide immunoblots compared with the enzymatic AMPK assay, rather than activation of the kinase through alternate phosphorylation sites.
Several mechanisms may contribute to the activation of heart AMPK during exercise. A rise in the intracellular concentration of AMP (or the AMP-to-ATP ratio) increases the activity of upstream AMPK kinase(s) (25, 47), enhances the sensitivity of AMPK to phosphorylation, and decreases its susceptibility to dephosphorylation (24). The total cardiac AMP concentrations measured by HPLC in freeze-clamped hearts are two orders of magnitude higher than the intracellular concentration of free AMP (as estimated by NMR spectroscopy in vitro), which regulates AMPK activity (15). Although it is not feasible to make such measurements in the heart during exercise, it is likely that the free AMP concentration increases during exercise, given the decrease observed in the creatine phosphate concentration (15). In addition, the decrease in the creatine phosphate concentration may also regulate AMPK through allosteric mechanisms that would operate in vivo (38) but may not be reflected by the in vitro measurements of kinase activity. An additional mechanism regulating the activity of AMPK is the concentration of glycogen (48, 56), through glycogen binding to the β-subunit of AMPK (28, 37). Thus the finding that cardiac glycogen content was reduced significantly in rats, as in previous reports (19), is of interest in terms of AMPK activation in the heart during exercise.
In these studies, we found that exercise was associated with phosphorylation of ACC in the heart. ACC is a well-recognized downstream target of AMPK (22) and is an important regulator of malonyl-CoA concentrations that modulate carnitine palmitoyltransferase-1 activity in both heart (31, 32) and skeletal muscle (40). In skeletal muscle, AMPK phosphorylates ACC during acute exercise (54, 55), although the role of AMPK in maintaining high levels of free fatty acid oxidation during more prolonged exercise remains somewhat uncertain (57). We observed exercise-induced phosphorylation of both ACC-1 (265 kDa) and ACC-2 (280 kDa). ACC-2 is associated with the mitochondria (1) and has an important role in modulating fatty acid oxidation (3), which is an important metabolic pathway in the heart during exercise. The pACC antibody that was utilized detects phosphorylation of Ser79 on ACC-1 and the equivalent Ser218 site on ACC-2, neither of which is thought to be a target for PKA (11, 14, 20). Catecholamines have a well-recognized role in modulating both the cardiac metabolic and contractile responses to exercise by increasing circulating free fatty acid supply to the heart as well as through direct PKA-mediated effects. These observations suggest that the AMPK pathway and catecholamines may have distinct but complementary effects on the heart during exercise.
These results also provide evidence that GLUT4 translocation to the sarcolemma occurs in association with AMPK activation in the heart during high-intensity treadmill running. Low levels of GLUT4 translocation are difficult to detect, and these measurements were not performed during moderate-intensity exercise. Pharmacological activation of AMPK is also known to increase GLUT4 translocation and glucose uptake in heart (4, 42) and skeletal (27, 33, 35, 36) muscle. In addition, AMPK appears to have a critical role in modulating glucose transport in both hypoxic skeletal muscle (36) and ischemic heart (41). However, the exact role of AMPK in mediating glucose utilization during exercise remains uncertain. In skeletal muscle, AMPK appears to be only partially responsible for contraction-mediated GLUT4 translocation, based on evidence from transgenic mice expressing a kinasedeficient α-isoform of AMPK (36). Catecholamines and increased intracellular calcium concentrations (39) also cause GLUT4 translocation in the heart, and it is possible that these additional mechanisms play a role in enhancing cardiac glucose uptake during exercise.
Although the downstream targets of AMPK that mediate GLUT4 vesicular trafficking have not been identified, AMPK stimulation of glucose transport does not require activation of PI 3-kinase, a key lipid kinase in the pathway of insulin-stimulated glucose transporter translocation (27, 42). Although PI 3-kinase does not mediate glucose transport during contraction in skeletal muscle (60), there is some evidence that it may be involved in glucose transport in isolated cardiac myocytes during electrical stimulation-induced contraction (50). In the current studies, we examined the effect of exercise on the phosphorylation state of the serine-threonine kinase Akt, which is distal to PI 3-kinase and appears to be activated by contraction in isolated skeletal muscle (44). However, we found no evidence that exercise increased the phosphorylation of either of the two key regulatory sites (Thr308 and Ser473) that mediate Akt activity in the heart. Thus these results suggest that, to the extent that AMPK stimulates glucose transporter translocation during exercise, it does so through a mechanism that does not involve downstream activation of the Akt pathway in the heart.
In conclusion, this study is the first to demonstrate that AMPK is activated by exercise in the normal heart, consistent with the hypothesis that AMPK may have a role in the cardiac response to physiological stress. Further studies will help to elucidate the extent to which AMPK activation is required for regulation of key physiological pathways in the normal heart.
This work was supported by National Institutes of Health Grants R01 HL-63811 (L. H. Young) and R01 DK-40936 (G. I. Shulman), and the Robert Leet and Clara Guthrie Patterson Trust (L. H. Young). D. L. Coven was the recipient of a postdoctoral research fellowship (National Research Service Award HL-10301). D. G. Hardie was supported by a Programme Grant from the Wellcome Trust and by a Research and Technological Development Contract (QLG1-CT-2001-01488) from the European Commission.
We thank Syed Hasan for expert technical assistance and Dr. Raymond R. Russell III for reviewing the manuscript.
Current address for R. Bergeron: Merck Research Laboratories, Rahway, NJ 07065.
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 © 2003 by American Physiological Society