INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE AMP-ACTIVATED PROTEIN kinase (AMPK) system has been described as a protective system for individual cells during metabolic stress by acting as a "fuel gauge," being activated by depletion of ATP and the resultant increase in the AMP:ATP ratio (17). Once activated, AMPK initiates energy-saving and energy-generating systems. In this regard, AMPK has been shown to phosphorylate and inhibit 3-hydroxy-3-methylglutaryl (HMG)-CoA reductase (12, 43) and acetyl-CoA carboxylase (ACC) in isolated hepatocytes (43) and in intact rat liver (12). By inhibiting HMG-CoA reductase and ACC, activated AMPK promotes energy saving in these tissues by reducing sterol and fatty acid synthesis, respectively.

Interestingly, AMPK mRNA expression is substantial in cardiac (47) and skeletal (14) muscle, even though neither cardiac nor skeletal muscle possesses the biosynthetic capabilities of fat and liver. In these tissues, activation of AMPK has been shown to be associated with increased energy production. For example, studies of skeletal muscle in vivo (49) and in vitro (23) show that AMPK is activated and fatty acid oxidation elevated during prolonged exercise or increased electrical stimulation. In isolated hearts exposed to global ischemia, activation of AMPK is accompanied by phosphorylation and inhibition of ACC as well as a reduction in its end product, malonyl-CoA (26, 27, 30). During subsequent reperfusion of the hearts, AMPK activity remains high, and the accompanying reduction in malonyl-CoA levels, which relieves the inhibition of carnitine palmitoyltransferase-1 and stimulates transport of long-chain fatty acids into the mitochondria (26, 27, 30), leads to accelerated rates of fatty acid oxidation (26, 27). Glucose uptake in skeletal muscle (7, 18, 34, 50) and in cardiac muscle (41) has also been shown to be increased on activation of AMPK by increases in workload or by exposure to 5-aminoimidazole-4-carboxamide 1--D-ribofuranoside (AICAR), an adenosine analog and cell-permeable activator of AMPK (11). In addition, AMPK has recently been shown to phosphorylate and activate 6-phosphofructo-2-kinase (PFK-2) in vitro and in isolated cells (33), suggesting that AMPK-induced activation of PFK-2 is involved in the stimulation of glycolysis during times of metabolic stress. Taken together, the results of these studies support the concept that activation of AMPK increases myocardial energy production by stimulating both fatty acid oxidation and glucose utilization.

There is evidence to suggest that AMPK may participate in the control of glycogen metabolism. AMPK has been shown to phosphorylate glycogen synthase purified from rabbit skeletal muscle at Ser 7 (10), a site known to be phosphorylated in vivo (37, 52). This phosphorylation promotes subsequent phosphorylation at Ser 10 by casein kinase-1, which ultimately causes inactivation of glycogen synthase (37, 44, 52). More recently, activation of AMPK by AICAR was shown to lead to the phosphorylation and inactivation of glycogen synthase in rat skeletal muscle (50). AMPK has also been shown to phosphorylate purified rabbit skeletal muscle phosphorylase kinase, the kinase responsible for phosphorylating and activating glycogen phosphorylase (10), although another report suggests that this may not be the case (9). If such AMPK-dependent phosphorylations occur in the heart and are physiologically relevant, they should lead to inactivation of glycogen synthase and/or activation of glycogen phosphorylase and, thereby, promote the breakdown of glycogen in the intact heart to increase energy production from endogenous glucose stores. This scenario makes great intuitive sense when the importance of glycogen as a fuel, especially during times of metabolic stress, is considered (1). Importantly, the effects of AMPK-induced changes in phosphorylation on enzyme activity and on glycogen metabolism in the intact heart are not yet fully known.

In the present study, we tested the hypothesis that increased AMPK activity promotes glycogenolysis in the intact heart by increasing glycogen phosphorylase activity and decreasing glycogen synthase activity. The effects of AMPK activation on myocardial glycogen metabolism were determined in isolated working rat hearts exposed to the cell-permeable AMPK activator AICAR. We specifically chose to perform this investigation in vitro, where conditions could be accurately controlled, to avoid confounding effects of changes induced by AICAR administration in vivo.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Isolated Heart Preparation and Perfusion Protocol

Hearts were perfused in the absence or presence of AICAR (0.8 or 1.2 mM, as indicated) for 15 min. Concentrations of AICAR in the range chosen have been shown to activate AMPK in skeletal muscle (7) and heart (32, 41). Additionally, we have previously shown that lower concentrations of AICAR stimulate long-chain fatty acid oxidation in isolated working rat hearts perfused under similar conditions (29). Higher concentrations of AICAR were not used because they adversely affected heart function in preliminary experiments. AICAR is the riboside derivative of the Z-base compound 5-amino-4-imidazole carboxamide (42). After entry into the cell, AICAR is monophosphorylated by adenosine kinase to form 5-aminoimidazole-4-carboxamide 1--D-ribofuranotide (ZMP), which is a normal constituent of the de novo purine synthesis pathway (42). As a consequence, AICAR has been used experimentally to restore purine nucleotide pools in postischemic myocardium (45) and in cerebral tissue of patients with Lesch-Nyan Syndrome (31). ZMP mimics the multiple effects of AMP on AMPK, causing allosteric activation and also promoting phosphorylation and activation by the upstream kinase, AMPK kinase (AMPKK) (17, 19). The time course for these experiments was chosen because 15 min is sufficient for AICAR-induced activation of AMPK in skeletal muscle (34) and because it minimizes the decrease in glycogen content observed in hearts perfused with octanoate for longer time periods (2). Selected octanoate-perfused hearts were exposed to either isoproterenol (100 nM) or insulin (100 mU/l). These latter experiments were used to validate the methods to measure glycogen synthesis and degradation under conditions in which glycogen is known to decrease (isoproterenol) or increase (insulin).

During the perfusion period, heart rate and peak systolic pressure were recorded at 5-min intervals with a DIREC physiological recording system (Fine Science Tools) using a pressure transducer (Viggo-Spectramed) in the aortic afterload line. The rate-pressure product (heart rate multiplied by peak systolic pressure) was calculated to determine the work performed by the heart (3). At the end of the 15-min perfusion period, hearts were quickly frozen using aluminum tongs cooled to the temperature of liquid nitrogen. Frozen heart tissue was powdered using a mortar and pestle and then stored in cryovials at 70°C until use.

Analytical Procedures

Myocardial glycogen. Myocardial glycogen content was determined after extraction of frozen ventricular tissue with 30% KOH, ethanol precipitation, and acid hydrolysis of glycogen (20). As described, a portion of the glycogen extract was used to determine net incorporation of radiolabeled glucose to measure net glycogen synthesis (4, 20). The extent of glycogen degradation was calculated as the difference between myocardial glycogen content at the start of the perfusion and the amount of unlabeled glycogen in the myocardium at the end of the perfusion (4). The former value was obtained from hearts frozen at the end of the Langendorff perfusion (114.9 ± 4.3 µmol/g dry wt, n = 8), while the latter value represents the difference between total glycogen content and net glycogen synthesis in myocardium at the end of the perfusion.

Activity of glycogen synthase, glycogen phosphorylase, and AMPK. Glycogen synthase, calculated from the ratio of activities determined with low or high glucose-6-phosphate and expressed as a percentage of total activity (%a), was determined, essentially as previously described (46). This method is based on the measurement of the incorporation of radioactivity into glycogen from UDP-[U-¹⁴C]glucose. To approximate in vivo conditions, synthase activity was measured in the presence of glucose-6-phosphate at a concentration of either 0.25 mM (a form) or 15 mM (total activity) (16).

Phosphorylation of AMPK and ACC. Phosphorylation of AMPK was determined to complement AMPK activity measurements in myocardial tissue extracts. The extent of phosphorylation of ACC, an important downstream target of AMPK, was also quantified as a means to assess AMPK activity in the intact heart (48). Phosphorylation of ACC has been shown to be a useful reporter of AMPK activity in skeletal muscle stimulated at low frequencies where AMPK activity is not measurably altered (36). The extent of phosphorylation of AMPK and ACC proteins in myocardium was determined by Western analysis by a previously described method (4). Briefly, samples of frozen ventricular tissue homogenate (containing 20 µg total protein) were solubilized by boiling in reducing sample buffer, separated by electrophoresis on 10% SDS-polyacrylamide gels, and transferred by electroblotting to a nitrocellulose membrane. After nonspecific blocking, the blots were probed overnight with primary rabbit antibodies against rat phospho-AMPK (1:1,000 dilution, Cell Signaling Technology, Mississauga, Ontario, Canada) or phospho-ACC (1:100 dilution, Upstate Biochemicals, Lake Placid, NY). After incubation with goat anti-rabbit secondary antibody, the signal was detected by an ECL-based detection system. Bands were quantified by densitometry. Equivalence of protein loading was confirmed by detection of total AMPK and ACC.

Other procedures. Adenine nucleotides as well as glucose-6-phosphate were determined in perchloric acid extracts of frozen ventricular tissue. Adenine nucleotides were measured in the tissue extracts by high-performance liquid chromatography according to a previously described method (5). Glucose-6-phosphate was measured by a standard spectrophotometric technique (8).

Data Analysis

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Heart Function

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Mechanical function of isolated working hearts perfused with octanoate (open symbols) or palmitate (filled symbols) for 15 min. Circles, no 5-aminoimidazole-4-carboxamide 1--D-ribofuranoside (AICAR); squares, 0.8 mM AICAR; triangles, 1.2 mM AICAR. bpm, Beats/min.

Glycogen Content

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Glycogen degraded (A), newly synthesized glycogen (B), and total glycogen content (C) in octanoate-perfused (open bars) and palmitate-perfused (filled bars) hearts. Values are means ± SE. The amount of glycogen degraded was calculated as the difference between glycogen at the outset of the perfusion (114.9 ± 4.3 µmol/g dry wt) and the amount of unlabeled glycogen at the end of the perfusion. The latter represents the difference between total glycogen content and net glycogen synthesized at the end of perfusion. * Significantly different from control value, P < 0.05; significantly different from octanoate-perfused value, P < 0.05; n = 3-5/group.

Incorporation of radiolabeled glucose, which is a measure of net glycogen synthesis during the perfusion period (4, 20), was not significantly altered by the presence of AICAR in either octanoate- or palmitate-perfused hearts (Fig. 2B). Total myocardial glycogen content tended to decrease in a stepwise manner with increasing AICAR concentration in both octanoate- and palmitate-perfused hearts (Fig. 2C). However, compared with corresponding hearts without AICAR, total glycogen content was reduced significantly only in octanoate-perfused hearts exposed to 1.2 mM AICAR.

As anticipated, insulin increased net glycogen synthesis (6.7 ± 0.3 vs. 3.1 ± 0.9 µmol/g dry wt, P < 0.05) and isoproterenol increased glycogen degradation (77.4 ± 6.7 vs. 3.6 ± 2.8 µmol/g dry wt, P < 0.05) compared with values in hearts perfused in their absence. Correspondingly, isoproterenol decreased (45.3 ± 4.4 µmol/g dry wt, n = 4, P < 0.05) and insulin increased (132 ± 6 µmol/g dry wt, n = 3, P < 0.05) total myocardial glycogen content compared with hearts perfused in the absence of either agent (114.0 ± 3.7 µmol/g dry wt).

Activities of AMPK, Glycogen Synthase, and Glycogen Phosphorylase

1

1

Significant concentration-dependent effects of ZMP on glycogen phosphorylase activity in myocardial homogenates were observed (Fig. 3A), yielding a K_m of glycogen phosphorylase for ZMP of 708 ± 37 µM. As shown previously by Corton et al. (11), AMPK is dose dependently activated by relevant concentrations of ZMP (Fig. 3B). We also found that ZMP inhibits glycogen synthase but does so only at concentrations >5 mM (Fig. 3C). This latter finding is in keeping with that of other investigators who showed no alteration in glycogen synthase by 3 mM ZMP (50).

View larger version (9K): [in this window] [in a new window]   Fig. 3.   Representative plots of the concentration-dependent effects of 5-aminoimidazole-4-carboxamide 1--D-ribofuranotide (ZMP) on glycogen phosphorylase (A), AMP-activated protein kinase (AMPK; B), and glycogen synthase (C) activity in homogenates of rat myocardium. U, µmol/min; prt, Protein.

Phosphorylation of AMPK and ACC

View larger version (46K): [in this window] [in a new window]   Fig. 4.   Representative immunoblots of the phosphorylated forms of AMPK (phosphoAMPK; A, top) and acetyl-CoA carboxylase (phosphoACC; B, top) as well as total AMPK (AMPK 1/2; A, bottom) and ACC (B, bottom) in isolated working rat hearts perfused in the absence (Control) or presence of AICAR. Homogenate from hearts exposed to 30 min no-flow ischemia (Ischemia) serves as a control for changes in phosphorylation of AMPK and ACC. Protein (20 µg) from the myocardial homogenate was loaded on each lane. Each lane represents a single heart.

Myocardial Metabolites

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Myocardial glycogen content reflects the balance between the flux through glycogen synthase and glycogen phosphorylase, both enzymes being controlled in a complex manner via phosphorylation and key allosteric mediators that include AMP, ATP, inorganic phosphate, and glucose-6-phosphate (35, 38). In the present experiments, measurable changes in glycogen phosphorylase or glycogen synthase activity were not observed in tissue extracts prepared from hearts after perfusion in the presence of AICAR (Table 2), even though glycogen degradation was stimulated (Fig. 2). This finding indicates that changes in phosphorylation of these key regulatory enzymes did not occur under the conditions of study and cannot explain the stimulation of glycogen degradation observed in response to AICAR. Other investigators have observed that AICAR activates glycogen phosphorylase and stimulates glycogenolysis, assessed indirectly as lactate production, in isolated skeletal muscle (51), presumably through phosphorylation and activation of phosphorylase kinase. The discrepant findings between cardiac and skeletal muscle may be a reflection of differing sensitivities of adult rat heart and skeletal muscle to the effects of AICAR. Alternatively, they may result from inherent differences in the relative importance of phosphorylation as opposed to allosteric control of glycogen metabolism between cardiac and skeletal muscle. The constant work load and dynamics of the heart might demand a greater reliance on the rapid control features afforded by allosteric mechanisms.

Changes in the cellular concentrations of key allosteric mediators, including glucose-6-phosphate and adenine nucleotides, were not observed in the presence of AICAR, indicating that these factors are not responsible for the stimulation of glycogen degradation observed. In concordance with our results, others have found that AICAR does not alter adenine nucleotide levels in well-perfused hearts (42). Inorganic phosphate was not measured in the current experiments. However, given the absence of changes in adenine nucleotides, it is unlikely that inorganic phosphate increased in the present experiments and, thereby, was presumably not a contributing factor to the changes in glycogen observed. We did, however, detect a substantial increase in ZMP in the presence of AICAR (Table 2). Importantly, ZMP was found to mimic the effects of AMP on glycogen phosphorylase in extracts of myocardium (Fig. 3). These effects on glycogen phosphorylase activity were observed at concentrations of ZMP in the range of those detected in the hearts in our study. On the basis of the ZMP content in hearts exposed to 1.2 mM AICAR, and assuming an intracellular water content of 2 ml/g dry wt (25), we calculate that the intracellular concentration of ZMP was ~600 µM in these hearts. This concentration of ZMP is very similar to the K_m value of ZMP for glycogen phosphorylase determined in myocardial extracts. Others have observed similar concentration-dependent effects of ZMP on glycogen phosphorylase activity in skeletal muscle extracts (51). We also found that glycogen synthase activity in myocardial extracts is unaffected by concentrations of ZMP achieved in these experiments (Fig. 3). Thus we speculate that, in the absence of changes in the phosphorylation state of regulatory enzymes or in content of key allosteric mediators, the AICAR-induced elevation of ZMP is responsible for the increased glycogen degradation observed in the present experiments by way of allosteric activation of glycogen phosphorylase. This viewpoint is entirely consistent with the well-recognized importance of allosteric modifiers in the control of glycogen metabolism in muscle (21, 39).

Our finding that AICAR stimulates glycogen degradation stands in contrast to those of other investigators. The majority have found that AICAR does not significantly alter total glycogen content in skeletal muscle (7, 28, 50) or heart muscle (41). However, Aschenbach et al. (6) recently found that total glycogen content of skeletal muscle was increased by AICAR administration in vivo. This discrepancy with respect to the effects of AICAR on muscle glycogen may be related to the different methods and/or preparations used to assess the effect of AICAR on glycogen metabolism. Determination of total glycogen content alone does not provide an accurate assessment of glycogen metabolism because it fails to account for simultaneous synthesis and degradation of glycogen (i.e., glycogen turnover), a process that actively occurs in rat hearts perfused under similar conditions (15, 20). By accounting for newly synthesized glycogen as well as the amount of glycogen at the outset of the perfusion (4), a more accurate determination of the extent of glycogen degradation was obtained in this study than could be obtained by measuring changes in total glycogen content alone. Failure to account for newly synthesized glycogen may thus explain why AICAR-induced changes in the extent of glycogen degradation were not observed in previous studies. The effect of AICAR on glycogen metabolism in vivo is complicated by accompanying effects on the concentration of relevant substrates in the circulation (6), which are likely a contributing factor to the increased skeletal muscle glycogen content observed. Of relevance, the authors of this investigation concluded that AICAR did not promote glycogen synthesis by changing glycogen synthase or phosphorylase activity. Rather, the increased synthesis of glycogen was attributed to alterations in intracellular concentrations of key allosteric modulators, specifically glucose-6-phosphate.

The AMPK activities observed in this study correspond well with previously reported values measured in isolated, perfused rat hearts (13, 27). Activity of AMPK measured in myocardial extracts in the absence of AMP, which is a reflection of the phosphorylation state of the kinase (34), was not altered after perfusion of hearts in the presence of AICAR in the current study (Table 2). This finding is supported by the observation that phosphorylation of AMPK, assessed by immunoblot analysis with anti-phospho-AMPK antibodies, was not altered by AICAR (Fig. 4). This failure to detect measurable changes in AMPK activity in myocardial extracts occurred even though the concentrations of AICAR used were comparable to or higher than those previously shown to activate AMPK in rat hearts in vivo (41) and in isolated hearts from newborn rabbits (32). One potential factor that may account for the difference between the present and previous studies is the duration of exposure of the heart to AICAR, even though studies in skeletal muscle have shown it to be sufficient (34). In the current study, AMPK was not measurably activated by exposure to AICAR for 15 min, a shorter duration of exposure than in investigations in which measurable AMPK activity was significantly elevated in the heart (32, 41).

It is important to recognize that absence of measurable changes in AMPK activity in tissue extracts does not exclude activation of AMPK allosterically, the effects of which are not readily detectable by in vitro assay. In fact, finding that phosphorylation of ACC, which serves as a reporter of AMPK activity in the intact organ or tissue (48), is increased in hearts exposed to AICAR (Fig. 4) indicates that AMPK is activated by AICAR in the intact heart independent of changes in phosphorylation, presumably by allosteric mechanisms. Support for this viewpoint comes from consideration of the concentration of ZMP achieved in these experiments, which is sufficient to increase AMPK activity (Ref. 11 and Fig. 3). Further support for this viewpoint comes from the finding that palmitate oxidation, an important AMPK-sensitive metabolic pathway in the heart, is stimulated by lower concentrations of AICAR in rat hearts perfused under similar conditions (29). Moreover, fatty acid oxidation is stimulated by AICAR in skeletal muscle in the absence of measurable changes in AMPK activity (24). An important corollary of this observation is that it suggests that AMPK is more sensitive to the allosteric effects of ZMP than its upstream kinase, AMPKK.

The likelihood of allosteric activation of AMPK (by AICAR) notwithstanding, our data do not unequivocally rule out the possibility that AMPK plays a role in the control of myocardial glycogen metabolism. In fact, recent data from skeletal muscle, where AMPK activity was increased by AICAR, suggest that AMPK phosphorylates and inactivates glycogen synthase (50), in keeping with phosphorylation of isolated glycogen synthase protein (10). Coupled with the fact that activation of AMPK is reported to increase intracellular glucose-6-phosphate concentration (7, 18, 34, 50), an important allosteric modulator, our data and that of others indicate that glycogen metabolism may be influenced by AMPK activation and changes in ZMP (or AMP) in several ways (Fig. 5). Further studies, including those in which AMPK is substantially and measurably activated, will be required to clarify this issue in the heart.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Diagram illustrating the potential control of myocardial glycogen metabolism by ZMP and AMPK. After its formation inside the cardiac myocyte, ZMP allosterically activates glycogen phosphorylase (GP), leading to enhanced degradation of glycogen. ZMP also causes activation of AMPK, which may inhibit glycogen synthase (GS). At the same time, ZMP-induced activation of AMPK results in increased uptake of glucose and glucose-6-phosphate (G-6-P) formation and activation of GS. In the current experiments, allosteric activation of GP by ZMP was presumably the predominant effect, accounting for the acceleration of glycogenolysis observed.

The results of these experiments have important ramifications with respect to the interpretation of experiments using AICAR and to the assessment of myocardial glycogen metabolism. First, it cannot be assumed that effects observed in response to AICAR are necessarily attributable to persistent activation of AMPK by phosphorylation, at least in the isolated working adult rat heart. Second, assessment of downstream targets, such as phosphorylation of relevant protein(s) or fatty acid oxidation rates, is required to account for the fact that AMPK can be activated (by AICAR) independent of changes in phosphorylation and in a manner not detectable in tissue extracts. Third, measurement of glycogen content alone fails to provide an accurate assessment of glycogen metabolism, primarily because it does not account for simultaneous synthesis and degradation of glycogen. Conclusions based on glycogen content alone should, therefore, be made with caution.