Insulin Stimulates Long-Chain Fatty Acid Utilization by Rat Cardiac Myocytes Through Cellular Redistribution of FAT/CD36

All animal studies were approved by the Experimental Animal Committee of Maastricht University. Rats were subjected to overnight fasting so as to lower systemic insulin concentrations. Cardiac myocytes were isolated from male Lewis rats (200–250 g) using a Langendorff perfusion system and a Krebs-Henseleit bicarbonate medium supplemented with 11 mmol/l glucose and equilibrated with 95% O₂ and 5% CO₂ (medium A) at 37°C as previously described (7). The isolated cells were allowed to recover for ∼2 h at room temperature in a Krebs-Henseleit bicarbonate medium supplemented with 1.0 mmol/l CaCl₂ and 2% (wt/vol) BSA and equilibrated with 95% O₂ and 5% CO₂ (medium B) at 37°C. Only when >80% of the cells had a rod-shaped appearance and excluded trypan blue were they used for subsequent tracer uptake studies.

Cell suspensions were subjected to an electric field via two platinum electrodes (distance 1.4 cm), which were connected to a pulse generator, capable of generating biphasic pulses up to 250 V (16). The monophasic components of the pulses exhibit a block profile. The duration of a monophasic pulse was set at 100 μs, and the time interval between the monophasic components before reversal of the voltage at 10 μs. The voltage was set at 200 V (140 V/cm) and the stimulation frequency at 4 Hz.

Cells (1.8 ml; 5–8 mg wet mass/ml), suspended in medium B, were preincubated in capped 20-ml incubation vials for 15 min at 37°C under continuous shaking. At the start of the incubations, 0.6 ml of the [1-¹⁴C]palmitate/BSA complex was added so that the final concentration of palmitate amounted to 100 μmol/l with a corresponding palmitate-to-BSA ratio of 0.3. This palmitate/BSA complex was prepared as previously described (7). Palmitate uptake (3 min incubation), oxidation (measured as production of ¹⁴CO₂ after 20 min of incubation), and esterification (measured as incorporation of radiolabel into phospholipids and triacylglycerols after 20 min of incubation) were determined as previously described (7). Uptake of 100 μmol/l 2-deoxy-d-[1-³H]glucose by cardiac myocytes after 3 min of incubation was also measured as previously described (7).

Giant vesicles were isolated from heart muscle, and palmitate uptake rates by these vesicles were determined exactly as previously described (11).

Cardiac myocytes (2.25 ml; 20–25 mg wet mass/ml) were incubated in medium B in the absence and presence of 10 nmol/l insulin for 15 min. At the end of the incubation, the total cell suspension was diluted with 1.0 ml, and NaN₃ was added to a final concentration of 5 mmol/l to stop ATP-dependent vesicular trafficking events, such as GLUT4 translocation (17). Immediately thereafter, cell suspensions were homogenized in a tightly fitting 10-ml Potter-Elvejhem glass homogenizer with 10 strokes and frozen in liquid nitrogen. Subsequently, fractionation was carried out according to Fischer et al. (18). Briefly, the thawed homogenates were centrifuged for 15 min at 17,000g. The pellet was washed once with 5 ml TES-buffer (20 mmol/l Tris, pH 7.4, 1 mmol/l EDTA, and 250 mmol/l sucrose, supplemented with 100 μmol/l phenylmethylsulfonyl fluoride), executing a 17,000g spin for 20 min, and resuspended in 1.5 ml TES-buffer. This total volume was layered on top of a sucrose cushion (38% wt/vol) in 20 mmol/l Tris, pH 7.4, 1 mmol/l EDTA, and ultracentrifuged for 65 min at 65,000g using a Beckmann SW41 rotor. The interface was collected, amply diluted with TES-buffer, and pelleted at 48,000g during 30 min. The pellet was resuspended in 100 μl TES-buffer and referred to as the plasma membrane (PM) fraction, because it is 13.5-fold enriched with ouabain-sensitive p-nitrophenyl-phosphatase, whereas the specific activity of the sarcoplasmatic EGTA-sensitive Ca²⁺-ATPase was 3.6-fold decreased. For collection of intracellular membrane pools containing recycling proteins, the supernatant of the first 17,000g spin was centrifuged for 30 min at 48,000g, resulting in the separation of a high-density microsomal fraction (pellet) and a low-density microsomal (LDM) fraction (supernatant). Pelleting of LDM occurred with a 250,000g spin for 65 min, after which this fraction was resuspended in 100 μl TES-buffer. The high-density microsomes were contaminated with plasma membranes. In the LDM fraction, no activity of p-nitrophenyl-phosphatase or of Ca²⁺-ATPase could be detected, indicating that this fraction was devoid of plasma membrane and of sarcoplasmic reticulum.

Rats were injected intravenously with insulin (2 units/kg body mass) or an equal volume of saline; 15 min later they were killed, and their hearts were removed and immediately freeze-clamped in liquid nitrogen. Upon thawing, hearts were diced and incubated for 30 min in a high-salt solution (2 mol/l NaCl, 20 mmol/l HEPES pH 7.4, and 5 mmol/l NaN₃) at 4°C as recommended by Fuller et al. (19). Thereafter, the suspension was centrifuged for 5 min at 1,000g and the pellet homogenized in 6.0 ml TES-buffer using a tightly fitting 10-ml Potter-Elvejhem glass homogenizer with 10 strokes. The resulting homogenate was centrifuged for 5 min at 1,000g, after which the pellet was rehomogenized in 4.0 ml TES-buffer with 10 strokes and then recombined with the 1,000g supernatant. Subsequently, fractionation was carried out as described by Fuller et al. (19). In brief, the homogenate was centrifuged for 10 min at 100g. The pellet (P1) was resuspended in 300 μl TES-buffer and saved. The supernatant was centrifuged for 10 min at 5,000g. The pellet (P2) was resuspended in 300 μl TES buffer and saved. The supernatant was centrifuged for 20 min at 20,000g. The pellet (P3) was resuspended in 300 μl TES-buffer and saved. The supernatant was centrifuged for 30 min at 48,000g. The pellet (P4) was resuspended in 150 μl TES-buffer and saved. The supernatant was centrifuged for 65 min at 250,000g. The pellet (P5) was resuspended in 150 μl TES-buffer and saved. Upon analysis of P1 to P5 with ouabain-sensitive p-nitrophenyl-phosphatase and with EGTA-sensitive Ca²⁺-ATPase, we decided to refer to P2 as PM-fraction and to P5 as LDM fraction.

For determination of the GLUT4 and FAT/CD36 content in PM and LDM, aliquots of the membrane fractions were separated with SDS-PAGE and Western blotting, as we have described recently (15). To detect FAT/CD36, we used MO25, and for detection of GLUT4, a polyclonal antiserum was applied. Signals obtained by Western blotting were quantified by densitometry.

Cellular wet mass was obtained from cell samples taken during the incubation period and determined after centrifugation for 2–3 s in a microcentrifuge and subsequent removal of the supernatant. Protein was quantified according to the bicinchoninic acid (BCA) method. Determination of blood glucose was carried out with Euroflash teststrips and a blood glucose meter (Lifescan Diagnostics, Beerse, Belgium) according to the manufacturer’s instructions. Ouabain-sensitive p-nitrophenyl-phosphatase was determined according to Bers (20) and EGTA-sensitive Ca²⁺-ATPase according to Jones et al. (21).

[1-¹⁴C]palmitic acid and 2-deoxy-d-[1-³H]glucose were obtained from Amersham Life Science, Little Chalfont, U.K. BSA (fraction V, essentially FA free), wortmannin, and phloretin were obtained from Sigma (St. Louis, MO). Collagenase type 2 was purchased from Worthington (Lakewood, NJ). BCA protein assay reagent kit was from Pierce (Rockford, IL). Antibodies against GLUT4 were obtained from Sanver Tech (Heerhugowaard, the Netherlands). SSP was routinely synthesized in our laboratory, as has been previously described (22). Its purity was confirmed with infrared spectroscopy (kindly performed by Dr. van Genderen, Eindhoven Technical University).

All data are presented as means ± SD for the indicated number of myocyte preparations. Statistical difference between groups of observations was tested with a paired Student’s t test. P ≤ 0.05 were considered significant.

Insulin induced deoxyglucose uptake by more than twofold. This effect was already near-maximal at 0.1 nmol/l, i.e., a low physiological concentration (Fig. 1). This establishes that the cardiac myocytes were insulin sensitive toward hexose uptake under physiological circumstances. Wortmannin, an inhibitor of PI-3 kinase, had no effect on basal glucose uptake, but completely inhibited insulin-sensitive glucose uptake (Fig. 2). These findings are in close agreement with studies of Ramrath et al. (23) in cardiac myocytes. Addition of SSP, a specific inhibitor of FAT/CD36 (11,24), had no significant effect on basal nor insulin-sensitive deoxyglucose uptake (Fig. 2).

Insulin stimulated FA uptake by cardiac myocytes up to 1.5-fold (Figs. 1, 3, and 4). This effect was already apparent within 30 s after addition of radiolabeled palmitate (Fig. 3) and, similar to deoxyglucose, near-maximal at 0.1 nmol/l of insulin (Fig. 1). Whereas phloretin inhibited basal FA uptake by 87%, in agreement with our previous observations (7), insulin was not able to increase palmitate uptake in the presence of this general inhibitor of carrier-mediated membrane transport processes. FA uptake in fasted cardiac myocytes is inhibited by SSP by 25%. In the presence of this inhibitor, cellular FA uptake is not altered upon insulin addition (Fig. 4). Hence, inhibition of FAT/CD36 action abolishes insulin-inducible FA uptake completely. To investigate the role of PI-3 kinase on cardiomyocytic FA uptake, wortmannin was used at a concentration range from 50 to 200 nmol/l. The effects of wortmannin on aspects of vesicular trafficking are routinely carried out at 200 nmol/l and higher (e.g., see refs. 23,25). However, side effects on PI-4 kinase, which is also inhibited by wortmannin, are negligible at 50 nmol/l (26). Wortmannin had no effect on basal FA uptake, and it completely inhibited insulin-sensitive FA uptake at both concentrations tested (Fig. 4).

Preincubation of cardiac myocytes for 15 min in the presence of 10 nmol/l insulin resulted in an increase in the amount of GLUT4 and FAT/CD36 in the PM by 1.91-fold and 1.45-fold, respectively, and a simultaneous decrease in the LDM of 61 and 62%, respectively (Fig. 5). To verify whether these in vitro effects of insulin were also occurring in vivo, rats were injected with insulin (2 units/kg rat) and killed 15 min later. At the time of death, blood glucose levels were 4.55 ± 0.30 mmol/l in control (saline-injected) rats (n = 4) and 2.52 ± 0.26 in insulin-treated rats (n = 5), confirming that insulin administration resulted in a rapid physiological response. Compared with the in vitro results with cardiac myocytes, in vivo administration of insulin elicited similar effects on distribution of both GLUT4 and FAT/CD36 between PM and LDM in the heart (Fig. 5).

The ability of vanadate, an insulin-mimetic agent, to alter the intrinsic activity of GLUT4 upon preincubation of giant vesicles from skeletal muscle with this compound (27) indicates the suitability of these vesicles to investigate aspects of insulin signaling at the sarcolemmal level. Basal FA uptake by these vesicles amounted to 1.14 ± 0.08 pmol/mg protein per second (n = 6). Preincubation of giant vesicles for 15 min with insulin resulted in an uptake rate of 1.06 ± 0.14 pmol/mg protein per second (n = 3). Hence, insulin had no direct effect on FA uptake by giant vesicles.

In the absence of electrostimulation and in agreement with experiments shown in Fig. 4, insulin at 10 nmol/l stimulated FA uptake by 1.51-fold, and this effect was completely abolished in the presence of wortmannin or SSP (Fig. 6). Electrical stimulation of cardiac mycytes at 4 Hz and 200 V caused FA uptake to increase by 1.64-fold, an effect that is remarkably similar in magnitude compared with insulin’s maximal effect. In addition, SSP was able to fully antagonize FA uptake into electrically stimulated cardiac myocytes (Fig. 6), in agreement with our earlier observations (16). However, contraction-inducible FA uptake was not reversed by 200 nmol/l wortmannin, indicating that insulin and contractions use different signaling pathways to recruit FAT/CD36. Indeed, when electrically stimulated cells were also stimulated with insulin, FA uptake was increased by 2.29-fold (Fig. 6). This effect is similar to the sum of the individual effects of contractions and of insulin on FA uptake. Furthermore, in the presence of both insulin and cellular contractions, SSP reduced FA uptake to the level observed when SSP was added to nonstimulated (control) cardiac myocytes, whereas the inhibitory effect of wortmannin was only partial (Fig. 6). Thus, despite the fact that the effects of insulin and cellular contractions on cellular FA uptake are additive, indicating their mutual independence, the effects of both stimuli are dependent on FAT/CD36.

Palmitate oxidation rates were not influenced by wortmannin or insulin (Fig. 7). In contrast, palmitate oxidation was markedly (2.77-fold) enhanced in contracting cells, which is in agreement with previous observations (16). Taking oxidation rates in these contracting myocytes as reference point, wortmannin and insulin were again without significant effect (Fig. 7). When examining intracellular storage of FA, the rate of palmitate esterification into cellular lipid pools was increased by insulin (1.36-fold), whereas contractions were without effect (Fig. 7).

This study demonstrates for the first time that insulin directly stimulates FA uptake by cardiac myocytes, i.e., by ∼1.5-fold. The abilities of wortmannin (inhibition of PI-3 kinase) and SSP (specific inhibitor of FAT/CD36) (16) to fully antagonize the effect of insulin on FA uptake indicate that both FAT/CD36 and PI-3 kinase are involved in this insulin effect.

Using subcellular fractionation of cardiac myocytes, we discovered that insulin is able to increase surface amounts of FAT/CD36 at the expense of intracellularly stored FAT/CD36, both in vitro and in vivo. In vivo insulin administration undoubtedly lowered circulating glucose, but in vitro insulin-inducible FA uptake and FAT/CD36 translocation were observed in the absence of glucose in the medium. Yet, despite these differences in the glucose milieu in these models, the FAT/CD36 translocation was remarkably similar in magnitude. Hence, it is very unlikely that the observed translocation should be attributed to an indirect effect of glucose. Furthermore, the insulin-induced increase in surface content of FAT/CD36 in cardiac myocytes (1.45-fold in vitro and 1.49-fold in vivo) closely parallels the effect of insulin on FA uptake (1.51-fold), indicating that the increase in surface amount of FAT/CD36 can fully account for the enhanced FA fluxes in the presence of insulin. In agreement with this notion, the fact that insulin did not influence FA uptake by heart giant vesicles, which lack endosomal storage compartments (11), argues against an insulin-induced change in intrinsic activity of FAT/CD36, such as through altered phosphorylation, as a likely mechanism to explain the insulin-induced increase in FA uptake.

It is obvious that the effects of insulin on surface accumulation of GLUT4 are somewhat more pronounced than its effects on surface content of FAT/CD36 (Fig. 5). In contrast, the relative insulin-induced alteration in intracellular contents of both GLUT4 (−61%) and FAT/CD36 (−62%) are very similar. This implies that the portion of FAT/CD36 stored intracellularly under basal conditions is substantially smaller than that of GLUT4, of which > 80% is stored in endosomal compartments in quiescent myocytes (1,28,29). These very similar relative decreases in intracellular GLUT4 and FAT/CD36 upon insulin addition may suggest that both transporters are localized in the same endosomal compartment and share the cellular machinery involved in their recruitment.

The similarity of regulation of FAT/CD36 translocation with that of GLUT4 translocation by contractions and by insulin is striking and allows the speculation that the analogy goes even further. In this respect, the generally held view is that the effects of insulin and contractions on GLUT4 translocation are additive and that both factors recruit GLUT4 from distinct intracellular storage compartments involving different signal transduction pathways (25,30,31). Using wortmannin at low concentrations so as to specifically inhibit PI-3 kinase, we were able to show that insulin induces FAT/CD36 translocation through PI-3 kinase-dependent signaling and that contraction-induced FAT/CD36 translocation occurs independent of PI-3 kinase action, again in analogy with GLUT4 translocation (25,30). The observation that the effects of contractions and of insulin on FA uptake are additive also points toward two functionally distinct intracellular FAT/CD36 stores and shows that cardiac myocytes are able to mobilize these stores simultaneously when both types of stimuli are present.

Both insulin and contractions stimulate cellular FA uptake to a similar extent, i.e., 1.5- to 1.6-fold, by means of FAT/CD36 translocation, despite the fact that different signaling pathways are being used. Importantly, the metabolic fate of the extra palmitate taken up in the presence of insulin is different than that taken up with cellular contractions. Notably, in the presence of insulin, the extra palmitate taken up is solely directed toward lipid esterification with the oxidation rate remaining unchanged. In contrast, in the presence of cellular contractions, the extra palmitate is oxidized, and esterification remains unchanged. It is tempting to speculate that the differences in recruitment of FAT/CD36 between stimulation by insulin and by contractions underlie the differences in metabolic fate of incoming FA. However, it is questionable whether contraction-recruitable FAT/CD36, once present at the sarcolemma, is functionally and spatially distinguishable from its insulin-recruitable counterpart. Besides uptake of FA by FAT/CD36, another important and possibly rate-limiting step in FA utilization is the conversion of fatty acyl-coenzyme-A into fatty acyl-l-carnitine by carnitine palmitoyltransferase-1 before mitochondrial entry (32). CPT1 is sensitive to inhibition of malonyl-CoA, a “by-product” from FA metabolism, and is formed from the linkage of β-oxidation- derived acetyl-CoA with CO₂ by acetyl-CoA carboxylase. Insulin is known to increase malonyl-CoA levels through dephosphorylation of acetyl-CoA carboxylase (2,3), whereas increased workload of the heart is accompanied by decreased malonyl-CoA (33). Thus, it is conceivable that intracellular levels of malonyl-CoA could regulate the destination of the extra incoming FA in the presence of insulin or contractions. In addition, regulation of FA metabolism at other sites is plausible. For instance, insulin is known to rapidly increase glycerol-3-phosphate acyltransferase, at least in adipocytes (34), which could explain accelerated FA esterification in the presence of insulin.

In summary, we have demonstrated a novel role of insulin in FA utilization by the heart. Insulin induces a translocation of FAT/CD36 from an intracellular pool to the sarcolemma of cardiac myocytes, thereby permitting an increased rate of cellular uptake of FA. Preliminary studies have shown that insulin also induces FAT/CD36 translocation in rat skeletal muscle (35).

The inducement of FAT/CD36 translocation by insulin allows the speculation that this translocation could be malfunctioning in type 2 diabetes-related syndromes. There is ample evidence showing disturbances in cardiac and skeletal muscle lipid metabolism in insulin-resistant conditions (36–38). For instance, the type 2 diabetic heart is characterized by increased triacylglycerol stores, which most likely are the result of a mismatch between uptake and utilization of FA. In recent studies with obese Zucker rats, we observed an increased abundance of FAT/CD36 at the sarcolemma, whereas in comparison with lean nondiabetic Zucker rats the total cardiac expression of this transporter was not altered (39). This more permanent relocalization of FAT/CD36 to the sarcolemma suggests that there is a chronic elevation of FAT/CD36-mediated FA influx, which then leads to the progressive build-up of intracellular lipid stores. The greater size of these triacylglycerol pools could be one of the critical facors leading to insulin resistance and the development of type 2 diabetic cardiomyopathies (37,38).

This study was supported by the Netherlands Heart Foundation (Grant D98.012) and by the Heart & Stroke Foundation of Ontario. J.J.F.P.L. is a Dekker postdoctoral fellow of the Netherlands Heart Foundation.

The authors would like to thank Ing. W.A. Coumans for his expert technical assistence with the synthesis of SSP and S.L.M. Coort for her participation in the subcellular fractionation procedures.