OBJECTIVE—Heart disease is a leading cause of death in diabetes and could occur because of excessive use of fatty acid for energy generation. Our objective was to determine the mechanisms by which AMP-activated protein kinase (AMPK) augments cardiac lipoprotein lipase (LPL), the enzyme that provides the heart with the majority of its fatty acid.
RESEARCH DESIGN AND METHODS—We used diazoxide in rats to induce hyperglycemia or used 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR) and thrombin to directly stimulate AMPK and p38 mitogen-activated protein kinase (MAPK), respectively, in cardiomyocytes.
RESULTS—There was a substantial increase in LPL at the coronary lumen following 4 h of diazoxide. In these diabetic animals, phosphorylation of AMPK, p38 MAPK, and heat shock protein (Hsp)25 produced actin cytoskeleton rearrangement to facilitate LPL translocation to the myocyte surface and, eventually, the vascular lumen. AICAR activated AMPK, p38 MAPK, and Hsp25 in a pattern similar to that seen with diabetes. AICAR also appreciably enhanced LPL, an effect reduced by preincubation with the p38 MAPK inhibitor SB202190 or by cytochalasin D, which inhibits actin polymerization. Thrombin activated p38 MAPK in the absence of AMPK phosphorylation. Comparable with diabetes, activation of p38 MAPK and, subsequently, Hsp25 phosphorylation and F-actin polymerization corresponded with an enhanced LPL activity. SB202190 and silencing of p38 MAPK also prevented these effects induced by thrombin and AICAR, respectively.
CONCLUSIONS—We propose that AMPK recruitment of LPL to the cardiomyocyte surface (which embraces p38 MAPK activation and actin cytoskeleton polymerization) represents an immediate compensatory response by the heart to guarantee fatty acid supply when glucose utilization is compromised.
Heart disease is a leading cause of death in diabetic patients (1), and coronary vessel disease and atherosclerosis play important roles in the increased incidence of cardiovascular dysfunction (2). However, a predisposition to heart failure in patients with both type 1 and type 2 diabetes might also reflect the effects of underlying abnormalities in diastolic function that can be detected in asymptomatic patients with diabetes alone (3–5). These observations suggest a specific impairment of heart muscle, termed diabetic cardiomyopathy. Because rodent models of chronic diabetes also display abnormalities in diastolic left ventricular function, with or without systolic left ventricular dysfunction (6), it can be proposed that the diabetic state can directly induce abnormalities in cardiac tissue independent of vascular defects. Several etiological factors have been proposed to explain the development of diabetic cardiomyopathy, including an increased stiffness of the left ventricular wall associated with accumulation of connective tissue and insoluble collagen (7) and abnormalities of various proteins that regulate ion flux, specifically intracellular calcium (8). More recently, the view that diabetic cardiomyopathy could also occur as a consequence of metabolic alterations has been proposed (9).
During insulin resistance or diabetes, glucose utilization is compromised. This alteration, together with increased fatty acid supply, switches cardiac energy generation to use of fatty acid. High fatty acid uptake and metabolism not only augment accumulation of fatty acid intermediates and triglycerides but also increase oxygen demand and generation of reactive oxygen species, leading to cardiac damage. Interestingly, increasing fatty acid uptake through overexpression of cardiac human lipoprotein lipase (LPL) (10) or fatty acid transport protein (11) or augmenting fatty acid oxidation through overexpression of cardiac peroxisome proliferator–activated receptor α (12) or long-chain acyl CoA synthase (13), results in a cardiac phenotype that resembles diabetic cardiomyopathy. Conversely, normalizing cardiac metabolism in diabetic animals reverses the development of cardiomyopathy (14). Taken together, these studies strongly support the role of altered metabolism in the development of diabetic cardiomyopathy.
LPL hydrolyzes triglyceride-rich lipoproteins, thus regulating the supply of fatty acid to meet the metabolic demands of different tissues. It is synthesized in myocytes and subsequently transported onto heparan sulfate proteoglycan (HSPG)-binding sites on the myocyte cell surface (15). Through mechanisms that are not completely understood, LPL is then transported onto HSPG-binding sites on the luminal surface of the capillary endothelium (16). At this location, the enzyme plays a crucial role in hydrolysis of triglyceride-rich lipoproteins to fatty acids, which are transported to the heart and used either for energy production or for resynthesis of triglycerides. Recently, LPL-mediated hydrolysis of circulating triglycerides was suggested as the principal source of fatty acids for cardiac use (17,18). In addition to its role as a lipolytic enzyme, LPL also mediates a noncatalytic bridging function that allows it to bind simultaneously to both lipoproteins and specific cell-surface proteins, facilitating cellular uptake of lipoproteins (19).
AMP-activated protein kinase (AMPK) plays a key role in the regulation of cardiac metabolism. Once activated, AMPK switches off energy-consuming processes such as protein synthesis, whereas ATP-generating mechanisms, such as fatty acid oxidation and glycolysis, are turned on (20). Additionally, results from our laboratory have demonstrated a strong correlation between activation of AMPK and increases in LPL activity (21). The objective of the present study was to determine the mechanisms by which AMPK augments cardiac LPL. Our data demonstrate that stress kinases like AMPK and p38 mitogen-activated protein kinase (MAPK), through their control of heat shock protein (Hsp) and the actin cytoskeleton, act in unison to facilitate LPL translocation to the myocyte cell surface and, eventually, to the coronary lumen. The ensuing alteration in cardiac fatty acid metabolism could be translated into increased cardiovascular risk following diabetes.
RESEARCH DESIGN AND METHODS
The current study adheres to the guide for the care and use of laboratory animals published by the U.S. National Institutes of Health and the University of British Columbia. Adult male Wistar rats (260–300 g) were obtained from the University of British Columbia Animal Care Unit and supplied with a standard laboratory diet (PMI Feeds, Richmond, VA) and water ad libitum. Diazoxide, a selective ATP-sensitive K+ channel opener, decreases insulin secretion and causes hyperglycemia (22,23). Although doses of 25 and 50 mg/kg increased plasma glucose, the extent and duration of hyperglycemia were not as substantial as with 100 mg/kg diazoxide, which caused a rapid decline in serum insulin within 1 h (23). Diazoxide (100 mg/kg) was administered intraperitoneally, and animals were killed at various times. Subsequently, hearts were removed for measurement of coronary luminal LPL activity and Western blotting.
Plasma measurements.
Control rats were injected with diazoxide at 10:00 a.m. (fed state). Following diazoxide, blood samples from the tail vein were collected over a period of 4 h and blood glucose determined using a glucometer (AccuSoft) and glucose test strips (Accu-Chek Advantage; Roche). At varying intervals, blood was also acquired in heparinized glass capillary tubes. Blood samples were immediately centrifuged and plasma collected and assayed. A diagnostic kit was used to measure nonesterified fatty acid (NEFA) (Wako).
Isolated heart perfusion.
Rats were anesthetized with 65 mg/kg sodium pentobarbital i.p., the thoracic cavity opened, and the heart carefully excised. After cannulation of the aorta, hearts were secured by tying below the innominate artery and perfused retrogradely by the nonrecirculating Langendorff technique with Krebs-Henseleit buffer containing 10 mmol/l glucose (pH 7.4). Perfusion fluid was continuously gassed with 95% O2/5% CO2. The rate of coronary flow (7–8 ml/min) was controlled by a peristaltic pump (24).
LPL activity and gene and protein expression.
To measure coronary endothelium–bound LPL, the perfusion solution was changed to buffer containing fatty acid–free BSA (1%) and heparin (5 units/ml). This concentration of heparin can maximally release cardiac LPL from its HSPG-binding sites. The coronary effluent (perfusate that drips down to the apex of the heart) was collected in timed fractions (10 s) over 5 min and assayed for LPL activity by measuring the hydrolysis of a [3H]triolein substrate emulsion (25). Retrograde perfusion of whole hearts with heparin results in a discharge of LPL that is rapid (within 0.5–1 min; suggested to represent LPL located at or near the endothelial cell surface), followed by a prolonged slow release (that is considered to originate from the myocyte cell surface) (24). As we were primarily concerned with examining regulation of LPL at the coronary lumen, only peak LPL activities are illustrated. LPL activity is expressed as nanomoles oleate released per hour per milliliter. LPL gene expression was measured using RT-PCR (26), and protein expression was determined using the 5D2 monoclonal mouse anti-bovine LPL (generously provided by Dr. J. Brunzell, University of Washington, Seattle, WA) (27,28).
Western blotting.
Western blot was performed as described previously (21). Briefly, ventricles (50 mg) or plated myocytes (0.4 × 106) were homogenized in ice-cold lysis buffer. After centrifugation at 5,000g for 20 min, the protein content of the supernatant was quantified using a Bradford protein assay. Samples were diluted, boiled with sample loading dye, and 50 μg used in SDS-PAGE. After blotting, membranes were blocked in 5% skim milk in Tris-buffered saline containing 0.1% Tween-20. Membranes were incubated with rabbit AMPK-α, phospho-AMPK (Thr-172), p38 MAPK, phospho-p38 MAPK (Thr180/Tyr182), Hsp25, and phospho-Hsp25 (S86) antibodies and subsequently with secondary goat anti-rabbit horseradish peroxidase–conjugated antibody. Reaction products were visualized using an enhanced chemiluminescence detection kit and quantified by densitometry.
Isolated cardiac myocytes.
Ventricular calcium–tolerant myocytes were prepared by a previously described procedure (23,28). Briefly, myocytes were made calcium tolerant by successive exposure to increasing concentrations of calcium. Our method of isolation yields a highly enriched population of calcium-tolerant myocardial cells that are rod shaped in the presence of 1 mmol/l Ca2+ with clear cross striations. Intolerant cells are intact but hypercontract into vesiculated spheres. Yield of myocytes (cell number, ∼4.8 × 106) was determined microscopically using an improved Neubauer hemocytometer. Myocyte viability (generally between 75 and 85%) was assessed as the percentage of elongated cells with clear cross striations that excluded 0.2% trypan blue. To examine the influence of 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR) (an AMPK activator) and the serine protease thrombin (to activate p38 MAPK) on LPL activity, cardiomyocytes were plated on laminin-coated six-well culture plates (to a density of 200,000 cells/well). Cells were maintained using Media-199 and incubated at 37°C under an atmosphere of 95% O2/5% CO2 for 16 h. Subsequently, and where indicated, AICAR (2 mmol/l) or thrombin (0.05 units/ml) was added to the culture medium. Following the indicated times, myocyte basal LPL activity released into the medium was measured. To release surface-bound LPL activity, heparin (8 units/ml; 1 min) was added to the culture plates and aliquots of cell medium were removed, separated by centrifugation, and assayed for LPL activity. In separate experiments, following incubation of plated myocytes (0.4 × 106 cells in a 60 × 15 mm tissue culture dish) with AICAR or thrombin, myocyte cell lysates were also used for Western blotting. In addition, myocytes were 1) preincubated with a p38 MAPK inhibitor (SB202190, 20 μmol/l) for 60 min before addition of either AICAR or thrombin (at the indicated times) and LPL activity and phospho-Hsp25 determined or 2) incubated with 0.5–1.5 mmol/l albumin-bound palmitic acid (1:2) for 15 min before Western blotting for AMPK.
Nuclear localization of p38 MAPK.
Following diazoxide, heart tissue was homogenized, whereas after incubation of myocytes with thrombin, cells were scraped and washed twice with 0.5 ml PBS. Subsequently, samples were lysed in ice-cold buffer A (10 mmol/l HEPES, pH 7.9, 10 mmol/l KCl, 0.1 mmol/l EDTA, 0.1 mmol/l EGTA, 1 mmol/l dithiothreitol, 0.5 mmol/l phenylmethylsulphonyl fluoride, and 0.5% NP40) for 15 min. After centrifugation (13,000 rpm, 3 min, 4°C), the supernatant (cytosolic fraction) was separated and the pellet vigorously vortexed with buffer B (20 mmol/l HEPES, pH 7.9, 0.42 mol/l NaCl, 1 mmol/l EDTA, 1 mmol/l EGTA, 1 mmol/l dithiothreitol, and 1 mmol/l phenylmethylsulphonyl fluoride) for 10 min. Following centrifugation (13,000 rpm, 10 min, 4°C), the supernatant (nuclear fraction) was quantified using a Bradford protein assay and used for Western blotting to determine nuclear localization of p38 MAPK. Using an antibody against Histone H3 as a nuclear marker, we show good purity of nuclear fractions (data not shown).
Filamentous and globular actin.
The ratio of filamentous to globular actin (F-actin/G-actin ratio) in the whole heart was determined using an in vivo assay kit. Briefly, hearts from control and diazoxide animals were isolated and lysed. Lysates were homogenized and centrifuged at 2,000 rpm for 5 min. Total actin content of the supernatant was centrifuged at 100,000g for 1 h at 37°C to isolate F-actin (pellet) and G-actin (supernatant). The pellets were resuspended to the same volume as the G-actin fraction using ice-cold Milli-Q water plus 10 μmol/l cytochalasin D (CTD) and left on ice for 1 h to dissociate F-actin. The ratio of F-actin to G-actin was determined using Western blotting and densitometry.
F-actin and G-actin were also determined in isolated cardiomyocytes using immunofluorescence. Briefly, myocytes were plated on laminin-coated cover glass slides and rinsed with PBS. Following incubation with thrombin or SB202190 at the indicated times, myocytes were fixed for 10 min with 4% paraformaldehyde in PBS, permeabilized with 0.1% Triton X-100 in PBS for 3 min, treated with PBS containing 1% BSA for 20 min, and finally rinsed with PBS. Cells were double stained with DNAaseI AlexaFluor594 and Rhodamine488 Phalloidin to colocalize monomeric globular actin (red, G-actin), and polymerized filamentous actin (green, F-actin) (29). The unbound fluorescent probe was rinsed with PBS buffer and slides visualized and photographed with a Leica fluorescent microscope (Wetzlar, Postfach, Germany). The effects of AICAR in myocytes were also determined in the presence or absence of 1 μmol/l CTD (an actin polymerization inhibitor) (30).
Silencing of p38 MAPK by small interfering RNA.
Small interfering RNA (siRNA) transfection of p38 MAPK in cardiomyocytes was performed using a kit from Santa Cruz. Briefly, in six-well culture plates, 0.1 × 106 cells were plated and subsequently exposed to the siRNA (or scrambled, Scr) solution for 8 h at 37°C in a CO2 incubator. Then, the media was changed to Media-199 and the cells incubated for another 18 h. Subsequently, AICAR (2 mmol/l) was added to the culture medium for 2 h, and LPL (released by heparin), p38 MAPK, and Hsp25 (using Western blotting) were determined.
Materials.
[3H]triolein was purchased from Amersham Canada. Heparin sodium injection (Hapalean; 1,000 USP units/ml) was obtained from Organon Teknika. The F-actin/G-actin in vivo assay kit was obtained from Cytoskeleton (Denver, CO). Total AMPK-α, phospho–AMPK-α, p38 MAPK, phospho–p38 MAPK, glyceraldehyde 3-phosphate dehydrogenase, and histone H3 antibodies were obtained from Cell Signaling (Danvers, MA). Hsp25 and phospho-Hsp25 antibodies were obtained from GeneTex (San Antonio, TX). SB202190 was purchased from Sigma-Aldrich. The enhanced chemiluminescence detection kit was obtained from Amersham. A diagnostic kit was used to measure NEFA (Wako). All other chemicals were from Sigma Chemical.
Statistical analysis.
Values are means ± SE. Wherever appropriate, one-way ANOVA followed by the Bonferroni test was used to determine differences between group mean values. The level of statistical significance was set at P < 0.05.
RESULTS
Acute diabetes increases LPL at the vascular lumen.
Following diazoxide, blood glucose levels increased and were significantly higher at 1 and 4 h after administration (Fig. 1A). Another characteristic feature associated with hyperglycemia, such as polydipsia, was also observed in diazoxide-treated animals. Retrograde perfusion of hearts with heparin resulted in release of LPL into the coronary perfusate. Compared with control rat hearts, there was a substantial increase in LPL activity (∼400%, Fig. 1B) at the vascular lumen following 4 h of diazoxide. This change in LPL activity was independent of shifts in mRNA (data not shown), suggesting a posttranscriptional increase in myocyte LPL. In addition, because no change in whole-heart LPL protein was observed (data not shown), it is likely that the increase in LPL activity at the coronary lumen is simply a result of transport of enzyme from myocytes to the endothelial cell. We have previously reported that control of LPL by diazoxide is dependent on its lowering of insulin rather than its direct effects on the heart or blood pressure (23).
Influence of fatty acids on cardiac AMPK phosphorylation.
Previous studies from our lab have reported significantly higher AMPK phosphorylation in hearts from moderately diabetic streptozotocin-induced diabetic animals (31). In the present study, following diazoxide, an approximately sixfold increase of cardiac AMPK phosphorylation was observed after 15 min (Fig. 2A). With time, despite persistence of hyperglycemia, AMPK phosphorylation in hearts from diazoxide-treated animals declined and reached basal levels within 4 h (Fig. 2A). Because AMPK activation is prevented in severe streptozotocin-induced diabetes with its attendant enlargement of plasma and heart lipids (31), we measured changes in plasma fatty acid to resolve whether a relationship exists between AMPK and fatty acid in diazoxide-treated animals. Interestingly, with time, the reduction in AMPK activation corresponded to a significant and rapid increase of fatty acid that peaked at 60 min and remained high until 4 h (Fig. 2B). When control myocytes were incubated with appropriate concentrations of palmitic acid, concentrations that varied from 0.5 to 0.8 mmol/l activated cardiac AMPK (Fig. 2C). However, with high concentrations of palmitic acid (that resembled the peak circulating concentrations seen with diazoxide), activation of AMPK was absent (Fig. 2C), suggesting that in vivo and in vitro, fatty acid has dual effects on AMPK activation.
Diazoxide stimulates cardiac actin polymerization.
p38 MAPK, a downstream target of AMPK (32), is suggested to regulate actin polymerization through its phosphorylation of Hsp25. In turn, the actin cytoskeleton has been implicated in management of myocyte LPL secretion (33). Estimation of cardiac cytosolic p38 MAPK phosphorylation showed a similar pattern to that seen with activation of AMPK: rapid activation followed by a decline to control levels within 4 h (Fig. 3A). Once phosphorylated, p38 MAPK relocates to the nucleus (34). Separation of the nuclear fraction revealed that concurrent to the decline in cytosolic p38 MAPK phosphorylation, nuclear p38 MAPK phosphorylation increased (Fig. 3A, inset). Phosphorylation of Hsp25 also intensified (Fig. 3B). However, this increase was time dependent, reaching a maximum at ∼60 min and remaining elevated for the next 3 h. To determine whether Hsp25 phosphorylation elicits F-actin polymerization, we quantitated F-actin and G-actin cellular fractions using Western blot. In the resting cardiomyocyte, the proportion of polymerized F-actin is consistently higher than G-actin (90:10) and is predominantly localized along the cell periphery. Additionally, an increase in the ratio of F-actin to G-actin indicates actin polymerization. Diazoxide increased the ratio of F-actin to G-actin (Fig. 3C). Interestingly, the increase in F-actin polymerization closely mirrored the enlargement of LPL activity at 4 h after diazoxide (Fig. 1B).
Promotion of AMPK phosphorylation in isolated control myocytes activates p38 MAPK and Hsp25 and recruits LPL to the cardiomyocyte cell surface.
To directly turn on AMPK, control myocytes were incubated for different times and with varying concentrations of AICAR. AICAR, up to 1 mmol/l, was incapable of phosphorylating cardiac AMPK (data not shown). Interestingly, increasing the concentration to 2 mmol/l activated AMPK phosphorylation in a pattern similar to that seen with acute diabetes: rapid activation followed by a decline to control levels with time (Fig. 4A). Comparable with acute diabetes induced with diazoxide, the activation of AMPK was temporally related to phosphorylation of p38 MAPK (Fig. 4B) and, subsequently, Hsp25 (Fig. 4C). We evaluated whether AICAR can augment LPL in myocytes, and Table 1 illustrates both basal and heparin-releasable LPL activity. Incubation of myocytes with AICAR had no effect on basal LPL activity. Interestingly, 2 mmol/l AICAR appreciably enhanced heparin-releasable activity in the medium. This increase occurred in the absence of any change in LPL mRNA or protein in cardiomyocyte lysates (data not shown).
We hypothesized that inhibition of Hsp25 phosphorylation should decrease cardiomyocyte LPL activity. In the absence of specific inhibitors of Hsp25, we used SB202190, an inhibitor of p38 MAPK. Incubation of control myocytes for 1 h with SB202190 decreased Hsp25 phosphorylation that is produced by AICAR (Fig. 5A). More importantly, the robust increase in heparin-releasable LPL activity induced by AICAR was also reduced by preincubation of myocytes with SB202190 (Fig. 5B). To investigate the involvement of the actin cytoskeleton in AICAR-mediated augmentation of myocyte LPL, myocytes were pretreated with an actin polymerization inhibitor, CTD, before incubation with AICAR. CTD reduced the effect of AICAR to increase myocyte heparin-releasable LPL without any effect on basal activity (control 3,417 ± 181, AICAR 5,305 ± 223, and AICAR plus CTD 2,163 ± 178 nmol · h−1 · 106 cells; P < 0.05).
Directly increasing p38 MAPK activity also enlarges the cardiomyocyte cell-surface LPL pool.
To activate p38 MAPK in the absence of AMPK phosphorylation, we used the serine protease thrombin. As predicted, control myocytes in the presence of thrombin did not display any change in AMPK phosphorylation (Fig. 6A, inset). Nevertheless, thrombin rapidly (within 5 min) phosphorylated cytosolic p38 MAPK, which was followed by a decline to control levels within 30 min. Comparable with diabetes, the decline in cytosolic p38 MAPK phosphorylation corresponded to an activation of nuclear p38 MAPK (Fig. 6B), Hsp25 phosphorylation (Fig. 6C), F-actin polymerization (Fig. 8), and enhanced heparin-releasable activity (Fig. 7B). Preincubation of control myocytes for 1 h with SB202190 prevented all of these effects induced by thrombin (Figs. 7 and 8).
Silencing of p38 MAPK prevents cardiomyocyte LPL recruitment observed with AICAR.
To confirm the relationship between p38 MAPK and LPL, we used siRNA to silence p38 MAPK expression in isolated cardiomyocytes. We first validated successful p38 MAPK inhibition using Western blotting (Fig. 9, inset). Interestingly, in myocytes in which p38 MAPK was silenced, heparin-releasable LPL activity was reduced (control plus heparin 2,308 ± 150, p38 MAPK–silenced control plus heparin 1,142 ± 167 nmol · h−1 · 106 cells; P < 0.05). Cardiomyocytes were next exposed to AICAR and Hsp25 and LPL activity determined. In myocytes in which p38 MAPK was silenced, AICAR had no influence on total p38 MAPK, which remained low (Fig. 9A) and was unable to phosphorylate Hsp25 (Fig. 9B) or increase LPL activity (Fig. 9C).
DISCUSSION
The major source of fatty acid for myocardial energy use is LPL-mediated hydrolysis of triglyceride-rich lipoproteins at the vascular endothelium (17). Despite this essential role of LPL at the coronary luminal surface, endothelial cells do not manufacture LPL. In the heart, the enzyme is synthesized in the underlying myocytes (35) before it is translocated to the luminal side of the coronary vessel wall with the help of heparan sulfate oligosaccharides acting as extracellular chaperones (16,36). Within the myocyte, we (30) and others (33) have reported actin cytoskeleton reorganization as an important means by which LPL is secreted onto plasma membrane HSPG-binding sites. In this study, for the first time, we demonstrate that following diabetes, it is the phosphorylation of AMPK, p38 MAPK, and Hsp25 that causes actin cytoskeleton rearrangement to facilitate LPL translocation to the myocyte cell surface and, eventually, to the coronary lumen.
AMPK is the switch that regulates cellular energy metabolism (37). Changes in intracellular AMP/ATP levels promote Threonine (Thr-172) phosphorylation and activation of AMPK, an important regulator of both carbohydrate and lipid metabolism (38,39). Thus, in heart and skeletal muscle, phosphorylated AMPK stimulates glucose uptake by inducing GLUT4 recruitment to the plasma membrane (40,41) and subsequent glycolysis through activation of 6-phosphofructo-2-kinase (42). AMPK control of fatty acid use includes its effect on fatty acid delivery to cardiomyocytes through its regulation of CD36 (43) and its role in facilitating fatty acid oxidation through its effect on acetyl-CoA carboxylase (44). Recently, we have also demonstrated that following AMPK activation after overnight fasting (with its attendant hypoinsulinemia), heparin-releasable LPL activity is amplified, providing an additional mechanism whereby cellular energy is regulated (21). In the present study, we report that acute diabetes increases cardiac LPL activity within 4 h. This augmentation in LPL activity was preceded by a rapid and intense phosphorylation in AMPK that was not sustainable and declined to control levels at 4 h. The early increase in AMPK may be a product of either metabolic stress associated with a decrease in insulin or a direct activation by circulating free fatty acids. Interestingly, in studies using L6 skeletal muscle, fatty acids have been shown to allosterically activate AMPK without changing energy charge (45). Despite the prevailing hyperglycemia following increasing durations of diazoxide, the decrease of AMPK activation to control levels is likely a consequence of the excessive amount of both circulating and LPL-derived fatty acids. This idea was strengthened by our experiment using isolated myocytes incubated with high concentrations of palmitate. In addition, moderate diabetes significantly increases cardiac AMPK and acetyl-CoA carboxylase phosphorylation, whereas in severe diabetes, with the addition of augmented plasma and heart lipids, AMPK activation is prevented (31). Recently, high fatty acid or triglycerides, through their formation of ceramide, have been shown to activate protein phosphatase 2A, leading to dephosphorylation of AMPK (46).
Given the observation that when LPL activity was the highest, AMPK phosphorylation returned to normal, we considered the possibility that the early activation of AMPK may have turned on other downstream signals. One downstream target of AMPK is p38 MAPK, and there was coincident activation of both AMPK and p38 MAPK following injection of diazoxide. Other studies have demonstrated that AMPK activates p38 MAPK through its interaction with transforming growth factor β–activated protein kinase 1–binding protein 1 (32). Cytosolic activation of p38 MAPK results in its transfer to the nucleus and in gene activation through a number of transcription factors (47). In the nucleus, p38 MAPK can also activate MAPK-activated protein kinase 2, which is then exported to phosphorylate Hsp25. Our studies in the heart confirmed that cytosolic activation of p38 MAPK was followed by its nuclear translocation. More importantly, with increasing duration of hyperglycemia, phosphorylation of Hsp25 progressively increased. Hsp25 is known to inhibit actin polymerization, and its phosphorylation results in a decline of this inhibitory function (48). In this setting, actin monomers are released from the phosphorylated Hsp25 to self-associate to form fibrillar actin. Because the increase in myocyte LPL activity at 4 h corresponded to an enlargement in the ratio of F-actin to G-actin, our data suggest that AMPK and p38 MAPK, through their control of Hsp25 and the actin cytoskeleton, act in unison to facilitate LPL translocation to the myocyte cell surface.
To eliminate the possibility that the above changes are 1) an outcome of a direct cardiotoxic effect of diazoxide and 2) a result of the myriad metabolic and hormonal changes that arise during diabetes, we used compounds to directly stimulate AMPK and p38 MAPK. AICAR is a cell-permeable activator of AMPK (49). In neonatal myocytes, 0.5–1 mmol/l AICAR is required to stimulate AMPK activity (50), whereas a concentration of up to 2 mmol/l is essential for AMPK phosphorylation in adult myocytes (51). In the present study, 2 mmol/l AICAR activated AMPK in a manner comparable with that seen with diazoxide. Prompt activation was followed by a reduction to control levels. At present, the mechanism for this decrease in cardiomyocyte AMPK phosphorylation with time is unknown (given the absence of fatty acid in the myocyte incubation medium). An increase in energy charge resulting from amplification in glucose uptake and the limited demand for energy in these nonbeating quiescent myocytes are potential explanations. Similar to diazoxide, there was a close relationship among AMPK activation, p38 MAPK and Hsp25 phosphorylation, and the increase in cardiomyocyte heparin-releasable LPL activity after exposure to AICAR. A different approach used thrombin to activate p38 MAPK. It should be noted that the more traditional methods to activate p38 MAPK include sorbitol and anisomycin; however, sorbitol is also known to activate AMPK (52), whereas anisomycin is shown to cause insulin resistance (53). Thrombin, without affecting AMPK phosphorylation, had a robust effect to provoke cardiomyocyte p38 MAPK, likely through proteinase-activated receptor 4 and c-Src tyrosine kinase activation (54). More importantly, p38 MAPK phosphorylation was followed by nuclear translocation, phosphorylation of Hsp25, actin cytoskeleton reorganization, and an increase in cell-surface LPL activity. Because the p38 MAPK inhibitor SB202190 and siRNA-mediated inhibition of p38 MAPK blocked the effects of thrombin and AICAR, respectively, on Hsp25 and LPL activity, our data suggest that F-actin polymerization produced by activation of p38 MAPK is an important means by which vesicle transport of LPL is made possible.
In summary, we propose that in addition to its direct role in promoting fatty acid oxidation, AMPK recruitment of LPL to the cardiomyocyte cell surface could represent an immediate compensatory response by the heart to guarantee fatty acid supply when glucose utilization is compromised. The mechanism underlying this process embraces p38 MAPK activation and an increase in actin cytoskeleton polymerization (Fig. 10). Interestingly, the actin cytoskeleton also plays a key role in promoting insulin-induced GLUT4 translocation. At present, the process of LPL vesicular movement along the actin filament network is unknown and merits further investigation. Understanding this mechanism could lead to strategies that overcome contractile dysfunction following diabetes. This is because changes in cardiac LPL activity may predispose people with diabetes to premature death from cardiac disease. In mice, both cardiac-specific overexpression (with its attendant lipid deposition, muscle fiber degeneration, and proliferation of mitochondria and peroxisomes [55]) and knockout (associated with cardiac interstitial and perivascular fibrosis [56]) of LPL have been implicated in cardiac dysfunction.
Limitations.
One limitation of this study is the lack of mouse models supporting the role of AMPK/p38 MAPK in regulating LPL recruitment to the cardiomyocyte cell surface. Transgenic mice overexpressing MAPK kinase 6 and MAPK phosphatase 1 are available and could potentially be used in future studies. However, it should be noted that endothelium-bound heparin-releasable LPL activity was unchanged in both type 1 and type 2 diabetic mouse hearts (57). This could be a consequence of genetic adaptation or the excessive heart rate in control animals (∼600 bpm), permitting prior translocation of LPL from the cardiomyocyte to the coronary lumen to saturate all of the LPL HSPG-binding sites.
. | AICAR (min) . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|---|
. | 0 . | 30 . | 60 . | 90 . | 120 . | ||||
LPL activity (nmol · h−1 · 106 cells) | |||||||||
Without heparin | 1,678 ± 94 | 1,724 ± 32 | 1,825 ± 164 | 1,832 ± 73 | 1,798 ± 92 | ||||
With heparin | 2,215 ± 85 | 2,305 ± 48 | 3,725 ± 256 | 5,624 ± 75* | 5,782 ± 115* |
. | AICAR (min) . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|---|
. | 0 . | 30 . | 60 . | 90 . | 120 . | ||||
LPL activity (nmol · h−1 · 106 cells) | |||||||||
Without heparin | 1,678 ± 94 | 1,724 ± 32 | 1,825 ± 164 | 1,832 ± 73 | 1,798 ± 92 | ||||
With heparin | 2,215 ± 85 | 2,305 ± 48 | 3,725 ± 256 | 5,624 ± 75* | 5,782 ± 115* |
Myocytes were prepared as described in research design and methods. AICAR (2 mmol/l) was added to the culture medium and myocytes kept for 30–120 min. Following the indicated times, myocyte basal LPL activity released into the medium was measured. To release surface-bound LPL activity, heparin (8 units/ml; 1 min) was added to the culture plates and aliquots of cell medium were removed, separated by centrifugation, and assayed for LPL activity. Results are means ± SE of three rats in each group.
Significantly different (P < 0.05) from control.
Published ahead of print at http://diabetes.diabetesjournals.org on 17 October 2007. DOI: 10.2337/db07-0832.
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 U.S.C. Section 1734 solely to indicate this fact.
Article Information
This study was supported by an operating grant from the Canadian Diabetes Association.