Akt Activity Negatively Regulates Phosphorylation of AMP-activated Protein Kinase in the Heart*

In the heart, insulin stimulates a variety of kinase cascades and controls glucose utilization. Because insulin is able to activate Akt and inactivate AMP-activated protein kinase (AMPK) in the heart, we hypothesized that Akt can regulate the activity of AMPK. To address the potential existence of this novel signaling pathway, we used a number of experimental protocols to activate Akt in cardiac myocytes and monitored the activation status of AMPK. Mouse hearts perfused in the presence of insulin demonstrated accelerated glycolysis and glucose oxidation rates as compared with non-insulin-perfused hearts. In addition, insulin caused an increase in Akt phosphorylation and a decrease in AMPK phosphorylation at its major regulatory site (threonine 172 of the α catalytic subunit). Transgenic mice overexpressing a constitutively active mutant form of Akt1 displayed decreased phosphorylation of cardiac α-AMPK. Isolated neonatal cardiac myocytes infected with an adenovirus expressing constitutively active mutant forms of either Akt1 or Akt2 also suppressed AMPK phosphorylation. However, Akt-dependent depression of α-AMPK phosphorylation could be overcome in the presence of the AMPK activator, metformin, suggesting that an override mechanism exists that can restore AMPK activity. Taken together, this study suggests that there is cross-talk between the AMPK and Akt pathways and that Akt activation can lead to decreased AMPK activity. In addition, our data suggest that the ability of insulin to inhibit AMPK may be controlled via an Akt-mediated mechanism.

The insulin signaling cascade is an important signaling pathway responsible for controlling substrate preference in the heart. Indeed, insulin has been shown to stimulate myocardial glucose uptake (1,2) and accelerate glycolysis (2, 3) and glucose oxidation rates (2,4) as well as promote glycogen synthesis (5,6). These effects of insulin on glucose metabolism are particularly important because a significant amount of glucose-derived ATP is used for maintaining proper cardiac function. In addition, the stimulation of myocardial glucose utilization and subsequent decrease in fatty acid oxidation has been proven to be efficacious in reducing ischemic injury (7)(8)(9)(10)(11). Moreover, insulin-stimulated glucose utilization may be a central component of the beneficial effects of glucose-insulin-potassium therapy on the ischemic heart (12). Two kinases that can be regulated by insulin are Akt and AMPK. 1 Akt is a serine/threonine protein kinase that can be activated by insulin via a multistep pathway involving a phosphatidylinositol 3-kinase-dependent mechanism (see Ref. 13 for review). Once Akt is phosphorylated and activated, it can promote glucose uptake and subsequent metabolism via translocation of glucose transporter (GLUT) 4 to the plasma membrane (14 -18). In addition, the phosphorylation and inhibition of glycogen synthase kinase (GSK) 3 by Akt activates glycogen synthase and thereby promotes glycogen synthesis (19). Recent evidence in heart indicates that the presence of elevated exogenous palmitate inhibits insulin-induced Akt phosphorylation (2), suggesting that Akt may be more central to the control of cardiac energy substrate preference than previously thought.
AMPK is a major regulator of cardiac energy substrate use. It is a heterotrimeric protein consisting of a catalytic subunit (␣) and two non-catalytic subunits (␤ and ␥) (20,21). AMPK activity is regulated via protein-protein interactions, allosteric interactions, and by phosphorylation/dephosphorylation of a major activating site (Thr-172) located on the ␣ catalytic subunit (see Ref. 22 for review). This phosphorylation site is a major controller of AMPK activity and is thought to be the target of at least one upstream kinase (23)(24)(25).
The major role of AMPK is to respond to alterations in the AMP/ATP or PCr/Cr ratios and control the energy homeostasis of the cell. In times of low energy supply, AMPK responds by either switching off energy-consuming pathways or by promoting energy producing processes. In the heart, one physiologically relevant condition that results in diminished ATP supply is ischemia (26). During myocardial ischemia, AMPK is activated and contributes to accelerated fatty acid oxidation rates upon reperfusion (26). In addition, AMPK activation has been shown to promote the translocation of GLUT 4 to the plasma membrane, thus stimulating glucose uptake (27). However, the role of AMPK in promoting glucose uptake and utilization appears to occur via a different signaling pathway than the insulin signaling pathway (27). Indeed, AMPK has been shown to be inactivated by insulin during normoxic conditions (28,29). In addition, insulin can antagonize ischemia-induced activation of AMPK, presumably by preventing phosphorylation at the activation site (Thr-172) of the catalytic subunit of AMPK by the upstream AMPK kinase (29). * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  Although Akt and AMPK have different cellular roles, common downstream targets do exist. For example Akt and AMPK both can phosphorylate endothelial nitric-oxide synthase (30). In addition, both enzymes appear to be responsible for the regulation of GLUT 4 translocation to the plasma membrane to promote glucose utilization (18,27). Despite this commonality, there appear to be distinct mechanisms by which Akt and AMPK signal GLUT 4 translocation. Therefore, although Akt and AMPK signaling may have at least one common end point in the regulation of cardiac substrate utilization, the mechanisms controlling each kinase cascade may be distinct. Indeed, in many instances Akt and AMPK activity are inversely correlated. For example, myocardial ischemia causes an activation of AMPK and inactivation of Akt, whereas the addition of insulin in the aerobic period activates Akt phosphorylation and blunts the AMPK response to a subsequent period of ischemia (29,32). Recently, it has been suggested that AMPK activation can regulate Akt in skeletal muscle (33,34), although it is not known if the inverse is true. In addition, whether Akt can regulate AMPK activity in the heart is currently unknown. Therefore, understanding how these two kinases interact at a molecular level in response to insulin may provide insights into how glucose-insulin-potassium therapy is cardioprotective.

MATERIALS AND METHODS
Animal Care-The University of Alberta adheres to the principles developed by the Council for International Organizations of Medical Sciences for biomedical research involving animals and complies with National Institutes of Health animal care guidelines.
Mouse Heart Perfusions-Hearts from male C57/BL mice (10 -12 weeks of age) were perfused in working mode essentially as described (35). Briefly, hearts were excised and immediately immersed in ice-cold Krebs-Henseleit bicarbonate buffer containing 118.5 mM NaCl, 25 mM NaHCO 3 , 4.7 mM KCl, 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 2.5 mM CaCl 2 , 5 mM dual-labeled 5-3 H and 14 C glucose with or without the addition of 100 microunits/ml insulin, and 1.2 mM palmitate prebound to 3% fatty acid-free bovine serum albumin. The aorta was then cannulated, and the heart was subjected to a retrograde Langendorff perfusion, at 60 mm Hg for 10 min. Upon cannulation of the left atria, the heart was switched from Langendorff to working mode. The left atrium was perfused at a preload pressure of 15 mm Hg, and afterload was set at 50 mm Hg. A 5-ml sample of perfusate and hyamine was taken every 10 min. At the end of the 60-min aerobic perfusion period, hearts were frozen with liquid nitrogen and stored at Ϫ80°C.
Akt1 Transgenic Mice-Transgenic mice overexpressing a constitutively active form of Akt1 (myrAkt1) in the heart were generated by crossing ␣MHC-tTA mice (36) with Tet-myrAkt1 mice, which express myrAkt1 under the control of tetracycline-responsive promoter. 2 Mice were maintained on doxycycline until 12 weeks of age, at which time doxycycline was withdrawn from the drinking water for 2 weeks to induce the myrAkt1 transgene expression in the heart. Control hearts were excised from uninduced littermates.
Cell Culture-Hearts from 2-day-old neonatal rat pups were isolated and placed in ice-cold 1 ϫ phosphate-buffered saline solution. After repeated rinsing, the atria were removed and the ventricles were minced with scissors. The minced tissue was washed three times in ice-cold 1 ϫ phosphate-buffered saline solution and then placed in a T-25-cm 2 tissue culture flask containing 19.5 ml of ice-cold 1 ϫ phosphate-buffered saline, (0.025%) DNase (w/v), (0.10%) collagenase (w/v), and (0.05%) trypsin (w/v). The tissue was digested on rotary shaker at 37°C for 20 min. After digestion the tissue was centrifuged at 114 ϫ g for 1 min at 4°C in 20 ml of DF20 media, 20% fetal bovine serum, and 50 g/ml gentamicin. The supernatant was discarded, and the pellet was subsequently added to DNase/collagenase/trypsin buffer for further digestion at 37°C for 20 min. After a second digestion, the tissue was again transferred into a 50-ml falcon tube with 20 ml of DF20 media and centrifuged at 114 ϫ g for 1 min at 4°C. This step was repeated 2 times. After final digestion, all the supernatant fractions were pooled and centrifuged at 300 ϫ g for 7 min at 4°C. The resulting pellet was resuspended in 10 ml of plating media (DF20 media, 5% fetal bovine serum, 10% horse serum, 50 g/ml gentamicin) and incubated at 37°C in a T-25-cm 2 tissue culture flask for 60 min. After 60 min the supernatant was removed and placed in another T-25-cm 2 tissue culture flask for an additional 60 min. This step was repeated twice. After serial plating, the resulting pellet was resuspended in plating media. The cells were plated on primeria dishes (Falcon) at a density of 1.8 -2.0 ϫ 10 6 cells/plate. Cell Treatment-After 18 h of culture, neonatal rat cardiac myocytes were rinsed two times with serum-free Dulbecco's modified Eagle's/ HAM'S F12 media and infected with Ad⅐GFP, Ad⅐myrAkt1 (37), or Ad⅐myrAkt2 (38) adenovirus at the multiplicity of infection of 25. After infection, the media were changed to Dulbecco's modified Eagle's/ HAM'S F12 media supplemented with insulin/transferrin/sodium selenite and cultured for an additional 48 h at 37°C in 5% CO 2 . Appropriate dishes of cells were also treated for 18 h with 5 mM of 1,1-dimethylbiguanide hydrochloride (Metformin) 30 h postinfection. After the appropriate treatment time, cells were rinsed two times with ice-cold 1 ϫ phosphate-buffered saline and lysed in 200 l of lysis buffer (20 mM Tris-HCl (pH 7.4), 50 mM NaCl, 50 mM NaF, 5 mM Na pyrophosphate, 0.25 M sucrose, 1% Triton X-100, mammalian protease inhibitor mixture, phosphatase inhibitor mixture I), and 1 mM dithiothreitol was added to each plate. Cells were scraped and lysed for 15 min on ice and then centrifuged at 1000 ϫ g for 10 min at 4°C. The protein concentration of the supernatant was then determined, and samples were subjected to SDS-PAGE and immunoblot analysis.
Statistical Analysis-All data are presented as means Ϯ S.E. of the mean (S.E.). The unpaired t test was used for determination of statistical differences between two groups. For comparison of three groups, analysis of variance followed by the Neuman-Keuls post hoc test was used. A value of p Ͻ0.05 was considered significant.

Cardiac Function and Glucose Metabolism in Perfused
Mouse Hearts-Mouse hearts perfused with or without insulin demonstrated no significant alterations in left ventricular work or coronary flow and were functionally very similar (Table I). Hearts perfused without insulin demonstrated glycolytic rates of 1964 Ϯ 422 mol⅐g dry wt Ϫ1 ⅐min Ϫ1 (Fig. 1A, open bars). With the addition of insulin these rates increased by 4-fold (Fig. 1A,  filled bars). A similar profile is observed in these hearts with respect to glucose oxidation rates, with glucose oxidation rates increasing ϳ8-fold in the insulin-perfused group as compared with controls (Fig. 1B).
Effects of Insulin on Akt and AMPK Phosphorylation-In an attempt to better understand the mechanisms that control the metabolic effects observed in the perfused hearts, extracts from hearts perfused with or without insulin were subjected to immunoblot analysis using anti-phospho-Akt (Ser-473 or Thr-308) or anti-phospho-␣ subunit of AMPK (Thr-172) antibodies (Fig. 2, A and B). Hearts perfused without insulin had relatively little phosphorylated Akt protein. Upon treatment with insulin, the level of phosphorylated Akt increased dramatically at both phosphorylation sites ( Fig. 2A, Ser-473 and Thr-308). Although phosphorylation at Ser-473 is indicative of the activation state of Akt, we also immunoblotted with anti-phospho-GSK3␣/␤ (Ser-21/9) antibody to ascertain whether altered phosphorylation of Akt had any effects on downstream target proteins such as GSK3␤, a known Akt target. In accordance with Akt activation, GSK3␤ phosphorylation was markedly elevated by insulin relative to hearts perfused in the absence of insulin ( Fig. 2A). In contrast to these changes in Akt phosphorylation, insulin treatment resulted in reduced phosphorylation of ␣-AMPK (Thr-172) and was inversely correlated with Akt phosphorylation status (Fig. 2B). During the relatively short perfusion period, the levels of total Akt, GSK3␤, or ␣ subunit of AMPK were not altered (Fig. 2, A and B).
Effect of Constitutively Active Akt1 on AMPK Phosphorylation Status in Transgenic Mice-Transgenic mice expressing constitutively active Akt1 (myrAkt1) specifically in the heart were used to determine the effects of Akt1 activation on the ␣-AMPK phosphorylation status. Expression of myrAkt1 in the heart induced an increase in phospho-Akt levels comparable with those observed in insulin-perfused hearts (Fig. 2) and dramatically reduced the phosphorylation of ␣-AMPK at Thr-172 (Fig. 3). Although phosphorylation of the ␣-AMPK was reduced, no changes in total ␣ subunit of AMPK protein levels were observed.

Effect of Constitutively Active Akt on AMPK Phosphorylation Status in Cultured Neonatal Rat Cardiac
Myocytes-To determine whether both Akt1 and -2 had similar abilities to reduce phosphorylation of ␣-AMPK at Thr-172, rat neonatal cardiac myocytes were cultured and infected with adenoviruses expressing either constitutively active Akt1 or Akt2 (Ad⅐myrAkt1 or Ad⅐myrAkt2). Cardiac myocytes expressing myrAkt2 reduced phosphorylation of the ␣ subunit of AMPK to similar levels as in transgenic mice expressing constitutively active Akt1 (Fig. 4). Almost identical results were obtained using Ad⅐myrAkt1 adenovirus (Fig. 5, lanes 1, 3, and 5). This reduction in the phosphorylation of ␣-AMPK was observed in cardiac myocytes cultured in the presence or absence of insulin (data not shown), suggesting that insulin signaling mechanisms distinct from Akt activation are not responsible for reduced AMPK phosphorylation.
Although activated Akt can reduce the basal phosphorylation status of ␣-AMPK at Thr-172, we determined whether activated Akt could maintain this effect in the presence of a known AMPK activator, metformin. Metformin has been shown to promote phosphorylation of the ␣-AMPK at Thr-172 without altering the cellular AMP/ATP ratio (39). Thus metformin was used in these experiments to avoid confounding variables such as altered energy status of the cell. With the addition of metformin, the inhibitory effect of Ad⅐myrAkt1 or -2 on ␣-AMPK phosphorylation was overcome (Fig. 5A), which corresponded to a dramatic increase in phosphorylation of an AMPK target protein, ACC, at Ser-79 (Fig. 5B). Metformin treatment did not result in increased expression of total ␣ subunit of AMPK protein or either of the two ACC isoforms.

DISCUSSION
Using insulin-perfused mouse hearts, hearts from transgenic mice expressing constitutively active Akt, and isolated cardiac myocytes expressing constitutively active Akt, our data suggest that the ability of insulin to inhibit AMPK can be controlled via Akt. The ability of insulin to lead to the phosphorylation and activation of Akt presumably causes the acceleration of glucose uptake via GLUT 4 translocation to the plasma membrane (18). This increased availability of glucose accelerates glycolysis and subsequently increases rates of glucose oxidation (40). This was observed in our studies. Because both glycolysis and glucose oxidation rates are dramatically accelerated in the mouse heart perfused with insulin, we investigated whether insulin modified phosphorylation control of AMPK. Although the activation of AMPK by insulin would contradict previous observations (28,29), AMPK activation is known to accelerate glucose uptake and utilization in the heart (27). Therefore, the phosphorylation status of the ␣ subunit of AMPK was measured in mouse hearts perfused in the presence or absence of insulin.
Similar to what has been reported in the perfused rat heart (29), insulin resulted in reduced phosphorylation of ␣-AMPK (Thr-172) in the perfused mouse heart. Therefore, the dramatic acceleration of glucose utilization in mouse hearts perfused with insulin may not be directly attributed to AMPK activation and increased GLUT 4 translocation to the plasma membrane. However, another major metabolic action of AMPK in the heart is the acceleration of fatty acid oxidation rates (26). Indeed, insulin has been shown to decrease fatty acid oxidation rates, presumably by decreasing AMPK activity (28). Therefore, it is possible that insulin promotes glucose utilization in the heart by 1) activating Akt and 2) inhibiting fatty acid oxidation by inhibiting AMPK activity. Although the mechanism by which AMPK is inactivated by insulin is unknown, it is possible that activated Akt may play a role.
Because cardiac Akt and AMPK activation seem to be inversely correlated during insulin stimulation, we investigated whether increased Akt activity could lead to inactivation of AMPK. Hearts from transgenic mice expressing constitutively active Akt1 were used to determine the effects of Akt activation on the phosphorylation of ␣-AMPK. Hearts from transgenic mice demonstrated a dramatic reduction in ␣-AMPK phosphorylation when compared with control littermates that do not express the transgene. Although the phosphorylation of the ␣-AMPK at Thr-172 was reduced, no changes in total ␣-AMPK protein levels between transgenic and control hearts were observed, demonstrating that decreased signal with anti-phospho-␣-AMPK (Thr-172) antibody was not due to a decrease in ␣-AMPK protein levels. Furthermore, these data indicate that activation of Akt in the heart in vivo is sufficient to downregulate ␣-AMPK phosphorylation during conditions of normal insulin levels, suggesting that insulin-induced down-regulation of ␣-AMPK phosphorylation is mediated by Akt-dependent pathways.
Although only myrAkt1 overexpressing mice were available for this study, we also investigated whether Akt-dependent inactivation of ␣-AMPK was specific to Akt1 or Akt2. In addition, because Akt2 has been suggested to be a more important regulator of glucose homeostasis than Akt1 (41,42), it was possible that we would observe different effects on ␣-AMPK phosphorylation with the two Akt isoforms. Rat cardiac myocytes expressing either myrAkt1 or myrAkt2 resulted in similar levels of decreased ␣-AMPK phosphorylation. These data suggest that Akt1 and Akt2 have at least one common downstream target that results in similar decreased phosphorylation of ␣-AMPK. These data also suggest that the ability of insulin in the intact heart to decrease phosphorylation of ␣-AMPK may be dependent on Akt activation and not on other pathways involved in the insulin signaling cascade.
Although the exact mechanism involved in decreased phosphorylation of ␣-AMPK by Akt is currently unknown, it is possible that either inactivation of the upstream AMPK kinase or stimulation of the AMPK phosphatase is involved. Alternatively, activated Akt may directly phosphorylate AMPK on a separate site other than Thr-172, which may prevent subsequent phosphorylation by AMPK kinase under steady-state conditions. Although Akt activation can reduce ␣-AMPK phosphorylation at Thr-172, the mechanism by which this occurs is reversible even in the presence of continued Akt activity. This may explain why insulin addition to ischemic hearts does not reduce AMPK activity (29). Although increased phosphorylation of ␣-AMPK may occur when the ATP/AMP ratio falls, it also may occur when nucleotide ratios are not altered, as is the case with metformin treatment (39). This suggests that independent of nucleotide changes, activation of the upstream AMPK kinase or inhibition of the AMPK phosphatase can increase phosphorylation of ␣-AMPK, regardless of Akt activity. In addition, because of the studies by Beauloye et al. (29) involving the ischemic heart, it is likely that ability of Akt to inhibit/reduce AMPK phosphorylation only occurs in hearts that are not metabolically stressed.
This study has provided us with information suggesting that there is cross-talk between the AMPK and Akt pathways. Although other studies have suggested that AMPK can regulate Akt expression and/or activity (33,43), this is the first report suggesting that Akt activation can lead to decreased ␣-AMPK phosphorylation. In addition, our data suggest that the ability of insulin to inhibit ␣-AMPK phosphorylation may be controlled via Akt. This opens the door to the possibility that the cardioprotective mechanism of glucose-insulin-potassium therapy to prevent ischemia/reperfusion injury occurs via Akt in- hibition of AMPK. Recently, Beauloye et al. (29,32) have shown that insulin administered prior to an anaerobic episode was able to lessen AMPK activation during ischemia. Because reduced AMPK activity during ischemia would presumably decrease fatty acid oxidation rates, it may explain why insulin is cardioprotective against ischemia/reperfusion injury.
The ability of activated Akt to decrease the phosphorylation of AMPK at its primary activation site of the catalytic subunit is a novel pathway that adds to the already complex mecha-nisms involved in AMPK regulation. Although the reasons for this inverse correlation of Akt and AMPK activity are as yet unknown, this study provides insight into the potential mechanism by which insulin inhibits AMPK activity in the heart. Although the metabolic effects of AMPK and Akt have been studied here, the ability of Akt to inhibit AMPK has implications beyond cardiac energy metabolism. Recently, AMPK activation has been implicated in reducing mTOR, p70S6 kinase, and eEF2 activity and seems to be very much involved in the regulation of protein translation and possibly the hypertrophic Each experiment was performed in duplicate, and four independent experiments were performed. Representative immunoblots obtained from the cell extracts of these groups using anti-phospho-␣AMPK (Thr-172) antibody (A) or anti-phospho-acetyl-CoA carboxylase (Ser-79) antibody (B) are shown. Anti-␣AMPK antibody (A) and peroxidase-labeled streptavidin (B) served as loading controls for total ␣AMPK and ACC protein; anti-␤-actin antibody served as a loading control for total protein (A). Extracts from cells that were treated with metformin are indicated by the ϩ symbol. response (44,45). In addition, insulin and IGF-1 have been shown to induce protein synthesis and cardiac hypertrophy, possibly via Akt activation (31,46,47). Therefore, it is possible that a contributing factor to Akt-induced hypertrophy may be the reduction of a protein synthesis inhibitor such as AMPK. Studies are ongoing to investigate this relationship.