Distinct transcriptional regulation of long-chain acyl-CoA synthetase isoforms and cytosolic thioesterase 1 in the rodent heart by fatty acids and insulin

David J. Durgan, Justin K. Smith, Margaret A. Hotze, Oluwaseun Egbejimi, Karalyn D. Cuthbert, Vlad G. Zaha, Jason R. B. Dyck, E. Dale Abel, Martin E. Young

Abstract

The molecular mechanism(s) responsible for channeling long-chain fatty acids (LCFAs) into oxidative versus nonoxidative pathways is (are) poorly understood in the heart. Intracellular LCFAs are converted to long-chain fatty acyl-CoAs (LCFA-CoAs) by a family of long-chain acyl-CoA synthetases (ACSLs). Cytosolic thioesterase 1 (CTE1) hydrolyzes cytosolic LCFA-CoAs to LCFAs, generating a potential futile cycle at the expense of ATP utilization. We hypothesized that ACSL isoforms and CTE1 are differentially regulated in the heart during physiological and pathophysiological conditions. Using quantitative RT-PCR, we report that the five known acsl isoforms (acsl1, acsl3, acsl4, acsl5, and acsl6) and cte1 are expressed in whole rat and mouse hearts, as well as adult rat cardiomyocytes (ARCs). Streptozotocin-induced insulin-dependent diabetes (4 wk) and fasting (≤24 h) both dramatically induced cte1 and repressed acsl6 mRNA, with no significant effects on the other acsl isoforms. In contrast, high-fat feeding (4 wk) induced cte1 without affecting expression of the acsl isoforms in the heart. Investigation into the mechanism(s) responsible for these transcriptional changes uncovered roles for peroxisome proliferator-activated receptor-α (PPARα) and insulin as regulators of specific acsl isoforms and cte1 in the heart. Culturing ARCs with oleate (0.1–0.4 mM) or the PPARα agonists WY-14643 (1 μM) and fenofibrate (10 μM) consistently induced acsl1 and cte1. Conversely, PPARα null mouse hearts exhibited decreased acsl1 and cte1 expression. Culturing ARCs with insulin (10 nM) induced acsl6, whereas specific loss of insulin signaling within the heart (cardiac-specific insulin receptor knockout mice) caused decreased acsl6 expression. Our data expose differential regulation of acsl isoforms and cte1 in the heart, where acsl1 and cte1 are PPARα-regulated genes, whereas acsl6 is an insulin-regulated gene.

  • gene expression
  • metabolism
  • peroxisome proliferator-activated receptor-α

diabetes is a primary risk factor for the development of cardiovascular disease. People with diabetes exhibit a two- to fourfold greater risk for ischemic heart disease and heart failure (30). In addition, cardiovascular disease risk factors, such as hypertension and obesity, are often associated with diabetes, thereby exacerbating the likelihood of cardiovascular morbidity and mortality (27). It has become increasingly clear that alterations in myocardial metabolism during diabetes likely play a role in the development of diabetic cardiomyopathy (35). Dyslipidemia during diabetes is associated with increased rates of fatty acid oxidative and nonoxidative metabolism, with decreased reliance on carbohydrate as a fuel (32, 35, 41). Mechanical efficiency of the left ventricle (LV) [defined as the external cardiac power for a given rate of myocardial oxygen consumption (MV̇o2)] is reduced when the heart is exposed to high-plasma nonesterified fatty acid (NEFA) concentrations, and it is improved by suppression of NEFA levels or mitochondrial fatty acid oxidation both in vivo (humans, dogs, and pigs) and in perfused rodent hearts (4, 17, 19, 24, 25, 29). In addition to reduced myocardial efficiency, dyslipidemia promotes channeling of fatty acyl derivatives into so-called lipotoxic pathways, which have been suggested to contribute to contractile dysfunction (36, 41). Despite growing appreciation for the vast therapeutic potential of targeting fatty acid metabolism for the treatment of ischemic heart disease and heart failure, relatively little is known regarding the mechanism(s) by which fatty acids cause O2 wasting and toxicity to the cardiomyocyte.

Differences in the theoretical O2 requirements for fatty acid and carbohydrate metabolism do not account fully for the depression in myocardial efficiency by high-plasma NEFA levels (31). This suggests that energy wastage during fatty acid metabolism occurs through one or more futile (substrate) cycles. Potential fatty acid-augmented futile cycles include mitochondrial uncoupling, fatty acid ↔ triacylglyceride, and fatty acid ↔ fatty acid-CoA. Himms-Hagen and Harper (15) recently suggested that uncoupling protein 3 (UCP3), mitochondrial thioesterase 1 (MTE1), long-chain acyl-CoA synthetase (ACSL), and the carnitine shuttle may together generate a futile cycle, wherein the cycling of fatty acyl groups between the mitochondrial matrix and the cytosol occurs at the expense of ATP utilization. Consistent with the idea that this futile cycle potentially contributes to reduced efficiency of the heart, we have recently reported that myocardial expression of both ucp3 and mte1 is markedly induced in an animal model of uncontrolled insulin-dependent, streptozotocin (STZ)-induced diabetes (33).

Both physiological and pathophysiological elevations in fatty acid availability result in an upregulation of enzymes involved in LCFA oxidation (41). This feed-forward mechanism is mediated in a large part by the nuclear transcription factor peroxisome proliferator-activated receptor-α (PPARα) (1). However, when fatty acid availability exceeds the capacity for fatty acid oxidation, fatty acyl derivatives accumulate within the cell (e.g., as observed during diabetes). These fatty acyl derivatives have been implicated in the pathogenesis of various aspects of Type 2 diabetes, including insulin resistance, pancreatic dysfunction, and diabetic cardiomyopathy (23, 28, 43). Although the exact mechanism(s) by which fatty acyl derivatives mediate their detrimental effects is (are) currently unknown, increased flux into pathways for ceramide biosynthesis, PKC activation, and reactive oxygen species generation have been implicated (22, 41, 43). What is even less clear is whether passage of fatty acids into oxidative versus nonoxidative pathways is due to a passive “spill over” or an active channeling. Recent studies focusing on the liver show that fatty acids entering a cell can be partitioned into oxidative and nonoxidative pathways by five distinct ACSL isoforms (ACSL1, ACSL3, ACSL4, ACSL5, and ACSL6). These studies suggest that localization of ACSL isoforms at distinct subcellular compartments channels fatty acids into specific pathways. For example, association of ACSL4 with peroxisomes infers a partitioning of fatty acyl-CoAs into peroxisomal metabolism by this isoform (21). In contrast, localization of ACSL5 at the mitochondrion suggests a role of this isoform in facilitating entry of fatty acids into β-oxidation (7). Localization of acyl-CoA synthetase activity at the cell surface would facilitate uptake of fatty acids into the cell; indeed, fatty acid transport (FAT) proteins (FATPs) exhibit endogenous acyl-CoA synthetase activity (13). Once within the bona fide cytosol, fatty acyl-CoAs may become a substrate for cytosolic thioesterase 1 (CTE1) (16). The liberated NEFA would then be channeled into distinct metabolic pathways by specific ACSL isoforms. However, such a mechanism of fatty acid channeling would be at the expense of ATP utilization (two ATP equivalents for each ACSL reaction).

Very little is known regarding either CTE1 or the distinct ACSL isoforms in the heart. In an attempt to gain insight into the roles of these enzymes in the heart, we tested the hypothesis that cte1 and the acsl isoforms are differentially regulated in the heart at the transcriptional level. After characterization of cte1 and acsl isoforms in normal rodent hearts, we utilized various in vivo and in vitro models to investigate candidate molecular mechanisms regulating myocardial expression of these genes.

MATERIALS AND METHODS

Animals.

Male Wistar rats (200 g initial weight, Charles River) were housed either at the Animal Care Center of the University of Texas Health Science Center at Houston (for diabetes, PPARα agonist, and cardiomyocyte studies) or at the Animal Care Center of the Children's Nutrition Research Center, Baylor College of Medicine (for high-fat feeding and fasting studies). PPARα null mice (20 g initial weight, Jackson) and strain- and age-matched controls (Svlmj) were housed at the Animal Care Center of the University of Texas Health Science Center at Houston. Cardiac-specific insulin receptor knockout (CIRKO) mice (8–10 wk old) and age-matched littermate wild-type mice (C57BL6/FVB/129SvJ) were housed at the Animal Resources Center of the University of Utah. All animal experiments were approved by respective Institutional Animal Care and Use Committees. All animals were housed under controlled conditions (23 ± 1°C; 12-h:12-h light-dark cycle) and received standard laboratory chow and water ad libitum, unless otherwise stated. Ten days before they were killed, animals were housed in a separate environment-controlled room, within which a strict 12-h:12-h light-dark cycle regime was enforced [lights on at 7 AM; zeitgeber time (ZT) 0]. On the day of the experiment, hearts were isolated from anesthetized rodents (pentobarbital sodium, 100 mg/kg ip), freeze clamped in liquid nitrogen, and stored at −80°C before RNA extraction. We have previously shown that the responsiveness of the rat heart to fatty acids is greatest during the middle of the dark phase, when the rodent is most active (34). Therefore, with the exception of the circadian and fasting studies (wherein time courses were performed), all animals in the present study were euthanized between 10 PM (ZT 15) and 4 AM (ZT 21) to optimize the validity of the data obtained.

Dietary manipulations.

Rodents were fed either standard laboratory chow (Purina Mills), a low-fat diet (LFD; Research Diets), or a high-fat diet (HFD; Research Diets). The LFD and HFD were isocaloric and varied only in the proportion of energy obtained from carbohydrate and fat. The contribution of carbohydrate, fat, and protein to total energy available was 70%, 10%, and 20% for the LFD and 20%, 60%, and 20% for the HFD, respectively. The source of carbohydrate was a combination of cornstarch, maltodexrin, and sucrose, whereas the fat source was a combination of soybean oil and lard. Standard laboratory chow is very similar to the LFD in terms of contribution of carbohydrate, fat, and protein to total energy available (60%, 12%, and 28%, respectively).

Induction of diabetes, fasting, high-fat feeding, and specific PPARα activation.

Experiments were performed in which rats were subjected to STZ-induced diabetes (65 mg/kg iv; euthanized after 4 wk), fasting (maximum duration of 24 h), high-fat feeding (4 wk duration), or administered with a specific PPARα agonist (WY-14643, 50 mg/kg ip; euthanized after 4 h), as described previously (34). We have previously shown that STZ-induced diabetes, fasting, and high-fat feeding all significantly increase plasma NEFA levels in these rats (33, 34).

Genetically modified mouse models.

The potential role of PPARα in regulating myocardial gene expression was investigated with the use of the previously established PPARα null mouse (20). Additionally, previously established CIRKO mice were utilized to investigate the potential influence of insulin on myocardial gene expression (2).

Isolated adult rat cardiomyocytes.

Isolated adult rat cardiomyocytes were prepared with protocols as described previously (8). Freshly isolated cardiomyocytes were cultured overnight in serum-free, DMEM-containing, laminin-coated plates. The cells were then challenged with oleate (0.1, 0.2, or 0.4 mM), the PPARα agonist WY-14643 (1 μM), the PPARα agonist fenofibrate (10 μM), or insulin (10 nM). Oleate was conjugated to fatty acid-free BSA at a final BSA concentration in the culture medium of 1% wt/vol. WY-14643 was prepared in DMSO, with a final DMSO concentration of 0.1% in the culture medium. Fenofibrate and insulin were prepared in sterile water. During the overnight culture before challenge, cardiomyocytes were cultured in DMEM containing the anticipated vehicle (i.e., 1% BSA for oleate, 0.1% DMSO for WY-14643, and sterile water for fenofibrate and insulin). After 0, 12, 24, or 48 h of challenge, cardiomyocytes were harvested in TriReagent and stored at −80°C before RNA isolation.

RNA extraction and quantitative RT-PCR.

RNA extraction and quantitative RT-PCR of samples were performed by using previously described methods (6, 10, 14). Specific quantitative assays were designed from rat and mouse sequences available in GenBank. Taqman assays for rat and mouse acsl1, acsl3, acsl4, acsl5, acsl6, cte1, cyclophilin, and gapdh are presented in Table 1. Standard RNA was made for all assays by the T7 polymerase method (Ambion, Austin, TX), using total RNA isolated from either rat or mouse hearts; the use of standard RNA allows absolute quantification of gene expression. The correlation between the number of PCR cycles required for the fluorescent signal to reach a detection threshold (Ct) and the amount of standard was linear over at least a 5-log range of RNA for all assays (data not shown). Gene expression data are represented as mRNA molecules (per ∼20 ng total RNA in the case of cultured cardiomyocytes or per ng total RNA in the case of intact rat and mouse hearts). Because of the lower yield of total RNA for isolated cardiomyocytes versus intact hearts, the accuracy in spectrophotometric determination of RNA concentration is reduced; typically, 20 ng total RNA are used for each cardiomyocyte gene expression measurement.

View this table:
Table 1.

Primer and probe sequences used in real-time quantitative RT-PCR

Plasma insulin level determination.

Immediately before heart isolation, blood was withdrawn from all rodents. The sample was placed on ice before centrifugation for 10 min at full speed using a desktop microfuge. The plasma was retained and stored at −80°C until insulin levels were measured spectrophotometrically, with a commercially available ELISA kit (LINCO).

Statistical analysis.

Two-way ANOVA was conducted to investigate the main effects of group (tissue or experimental condition) and time. The general linear model procedure in SAS software, version 8.2, was used for this analysis (SAS Institute). A full model, including second-order interactions, was conducted for each experiment. Significant differences were determined by using type III sums of squares. Tukey's post hoc test was conducted for all significant two-way ANOVAs. The null hypothesis of no model effects was rejected at P < 0.05. Repeated-measures analysis was not utilized due to the fact that samples in each group were from different animals.

RESULTS

Characterization of acs1 isoforms and cte1 at transcriptional level in rodent heart.

A comparison of the absolute level of expression between the different acsl isoforms was performed for intact rat and mouse hearts, as well as isolated adult rat cardiomyocytes. Similar expression profiles were observed in all three sample types. The predominant isoform was acsl1, followed by acsl5, acsl4, acsl3, and, lastly, acsl6 (Fig. 1). A similar type of analysis was performed for the absolute level of expression of cte1 and mte1; here, previously published data from Young's laboratory (33) for mte1 were utilized. We report that mte1 gene expression is at least 33-fold greater than cte1 expression in intact rat and mouse hearts, as well as isolated adult rat cardiomyocytes (data not shown). We next investigated diurnal variations in the expression of the acsl isoforms, cte1, and cyclophilin in the rat heart. Figure 2 shows circadian rhythms in expression for acsl3 and cte1, with peak levels of expression at or near the dark-to-light phase transition. In contrast, no significant diurnal variations in mRNAs encoding for acsl1, acsl4, acsl5, acsl6, and the housekeeping gene cyclophilin were observed (Fig. 2).

Fig. 1.

Relative abundance of long-chain acyl-CoA synthetases (acsl) isoform gene expression in intact rat (A) and mouse hearts (B), as well as isolated adult rat cardiomyocytes (C). Values are shown as means for ∼200 (A), 50 (B), and 10 (C) observations for each acsl isoform. Data are represented as percentage of total acsl expression.

Fig. 2.

Diurnal variations in acsl1 (A), acsl3 (B), acsl4 (C), acsl5 (D), acsl6 (E), cytosolic thioesterase 1 (cte1; F), and cyclophilin (G) gene expression in hearts isolated from rats fed ad libitum. Values are shown as means ± SE for between 24 and 30 observations at each time point. Data are normalized to nanograms total RNA.

Differential regulation of acsl isoforms and cte1 gene expression in heart during STZ-induced diabetes.

Four weeks of uncontrolled STZ-induced diabetes were associated with a significant decrease in myocardial acsl6 expression, with no significant effects on acsl1, acsl3, acsl4, or acsl5 (Fig. 3). In contrast, induction of diabetes dramatically increased cte1 expression in the rat heart (26-fold; Fig. 3). Diabetes had no effect on myocardial cyclophilin expression (Fig. 3). Young's laboratory (33) has previously reported that plasma NEFA levels are 2.4-fold increased in these diabetic rats, compared with their controls.

Fig. 3.

Effects of streptozotocin (STZ)-induced diabetes (STZ-diabetes) on acsl1 (A), acsl3 (B), acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and cyclophilin (G) gene expression in rat heart. Rats were injected intravenously with either vehicle (saline; control) or STZ (65 mg/kg) 4 wk before heart isolation. Values are shown as means ± SE for between 20 and 21 observations in each group. Data are normalized to nanograms total RNA. ***P < 0.001 vs. control.

Differential regulation of acsl isoforms and cte1 gene expression in heart during fasting.

Major humoral factor alterations during diabetes include hypoinsulinemia, hyperlipidemia, and hyperglycemia. In contrast, fasting is associated with hypoinsulinemia and hyperlipidemia, in the absence of hyperglycemia. We therefore investigated the effects of fasting on myocardial expression of acsl isoforms and cte1. Fasting for >21 h was associated with a significant decrease in the expression of acsl6 in the rat heart, with no significant effects on acsl1, acsl3, acsl4, or acsl5 (Fig. 4). In contrast, fasting caused a rapid (within 12 h) induction of cte1 in the rat heart (Fig. 4). Prolonged fasting (i.e., 24 h) resulted in a slight, but significant, decrease in myocardial cyclophilin expression (Fig. 4). Young's laboratory (34) has previously reported that plasma NEFA levels increase 9.0-fold in these fasted rats, within 12 h of food withdrawal.

Fig. 4.

Diurnal variations in acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and cyclophilin (G) gene expression in hearts isolated from standard chow fed (⧫) and fasted (▪) rats. For fasted rats, food was withdrawn at zeitgeber time (ZT) 6. Values are shown as means ± SE for between 5 and 7 observations at each time point. Data are normalized to nanograms total RNA. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. fed control at same time point.

High-fat feeding induces cte1 gene expression in rat heart.

Whether increased fatty acid availability contributes toward the changes in myocardial gene expression observed during diabetes and fasting was investigated further by feeding rats a HFD for 4 wk. High-fat feeding had no significant effects on expression of the acsl isoforms investigated or of cyclophilin in the rat heart (Fig. 5). In contrast, high-fat feeding dramatically induced myocardial cte1 expression (Fig. 5). Young's laboratory (33) has previously reported that plasma NEFA levels are 3.9-fold increased in these HFD rats, compared with their controls. Plasma insulin levels were not significantly different between HFD and LFD rats (5.92 ± 0.69 vs. 6.41 ± 0.52 ng/ml, respectively).

Fig. 5.

Effects of high-fat feeding on acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and cyclophilin (G) gene expression in rat heart. Rats were fed low- or high-fat diets (LFD and HFD, respectively) for 4 wk before heart isolation. Values are shown as means ± SE for 17 observations in each group. Data are normalized to nanograms total RNA. ***P < 0.001 vs. LFD control.

Concentration- and time-dependent effects of oleate on acsl isoforms and cte1 gene expression in isolated adult rat cardiomyocytes.

To investigate further the potential direct effects of fatty acids on myocardial acsl isoforms and cte1 gene expression, we next utilized isolated adult rat cardiomyocytes. Prolonged culturing of cardiomyocytes in the absence of fatty acids (i.e., control cells) was associated with a time-dependent decrease in all genes investigated, with the exception of acsl3 (Fig. 6). In the latter case, acsl3 expression exhibited a rhythmic pattern (Fig. 6). Supplementation of the culture medium with oleate resulted in a concentration-dependent increase in acsl1, acsl3, acsl4, and cte1 expression (Fig. 6). In contrast, prolonged exposure of adult rat cardiomyocytes to oleate caused a slight, but significant, decrease in acsl5 expression (Fig. 6). Oleate had no significant effects on either acsl6 or cyclophilin expression (Fig. 6).

Fig. 6.

Concentration- and time-dependent effects of oleate on acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and cyclophilin (G) gene expression in isolated adult rat cardiomyocytes. Cardiomyocytes were incubated either in the absence [control (⧫)] or presence of oleate [0.1 (▪), 0.2 (▴), and 0.4 (★) mM] for 12, 24, or 48 h. Values are shown as means ± SE for 5 observations in each group. Data are represented as mRNA molecules/∼20 ng total RNA. *P < 0.05, **P < 0.01, and ***P < 0.001 for 0.1 mM vs. control at same time point. #P < 0.05, ##P < 0.01, and ###P < 0.001 for 0.2 mM vs. control at same time point. $$P < 0.01 and $$$P < 0.001 for 0.4 mM vs. control at same time point.

Effects of PPARα-agonists on acsl isoforms and cte1 gene expression in isolated adult rat cardiomyocytes.

We next investigated whether direct activation of PPARα, through the use of the PPARα agonists WY-14643 and fenofibrate, mimicked the effects of oleate on myocardial expression of acsl isoforms and cte1. Figure 7 shows that WY-14643 (1 μM for 24 h) causes a significant induction of acsl1 and cte1 in isolated adult rat cardiomyocytes, with no significant effects on acsl3, acsl4, acsl5, acsl6, or cyclophilin expression. Similarly, fenofibrate (10 μM for 24 h) caused an induction of acsl1, acsl5, and cte1 with no effects on acsl3, acsl4, acsl6, or cyclophilin expression (Fig. 8).

Fig. 7.

Effects of WY-14643 on acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and cyclophilin (G) gene expression in isolated adult rat cardiomyocytes. Cardiomyocytes were incubated either in the absence (control) or presence of WY-14643 (1 μM) for 24 h. Values are shown as means ± SE for 5 observations in each group. Data are represented as mRNA molecules/∼20 ng total RNA. *P < 0.01 and ***P < 0.001 vs. control.

Fig. 8.

Effects of fenofibrate on acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and cyclophilin (G) gene expression in isolated adult rat cardiomyocytes. Cardiomyocytes were incubated either in the absence (control) or presence of fenofibrate (10 μM) for 24 h. Values are shown as means ± SE for 5 observations in each group. Data are represented as mRNA molecules/∼20 ng total RNA. *P < 0.05 and **P < 0.01 vs. control.

Influence of acute PPARα activation on myocardial acsl isoforms and cte1 gene expression in vivo.

Whether specific activation of PPARα influenced myocardial gene expression in vivo, to a similar extent as observed in vitro, was investigated by injecting rats with WY-14643 (50 mg/kg) and subsequently isolating hearts 4 h later. WY-14643 had no significant acute effects on the expression of acsl isoforms or cyclophilin in the intact rat heart (Fig. 9). In contrast, WY-14643 dramatically induced myocardial cte1 expression (Fig. 9).

Fig. 9.

Effects of peroxisome proliferator-activated receptor (PPARα) agonist WY-14643 administration on acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and cyclophilin (G) gene expression in rat heart. Rats were injected intraperitoneally with either vehicle (1:1 DMSO/saline) or WY-14643 (50 mg/kg) 4 h before heart isolation. Values are shown as means ± SE for 18 observations in each group. Data are normalized to nanograms total RNA. ***P < 0.001 vs. vehicle control.

Myocardial expression of acsl isoforms and cte1 in PPARα null mouse hearts.

To investigate further the potential influence of the transcription factor PPARα on basal myocardial expression of the acsl isoforms, as well as cte1, the PPARα null mouse was utilized. Hearts isolated from PPARα null mice exhibited significantly lower levels of acsl1 and cte1 expression, compared with hearts isolated from wild-type mice (Fig. 10). In contrast, acsl6 expression was significantly higher in hearts isolated from PPARα null mice compared with those isolated from wild-type hearts (Fig. 10). Expression of acsl3, acsl4, acsl5, and cyclophilin did not differ between hearts isolated from wild-type and PPARα null mice (Fig. 10). Furthermore, consistent with previously published observations (11), plasma insulin levels are elevated in PPARα null mice compared with wild-type controls (0.71 ± 0.09 vs. 0.44 ± 0.03 ng/ml, respectively; P < 0.05).

Fig. 10.

Expression of acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and cyclophilin (G) in hearts isolated from age-matched wild-type and PPARα null mice. Values are shown as means ± SE for 18 observations in each group. Data are normalized to nanograms total RNA. ***P < 0.001 vs. wild-type mice.

Insulin modulates acsl isoform gene expression in isolated adult rat cardiomyocytes.

In an attempt to investigate the role of insulin as a potential regulator of acsl isoforms and/or cte1 expression, adult rat cardiomyocytes were challenged with insulin (10 nM for 24 h). Insulin significantly increased the expression of acsl1, acsl3, and acsl6 expression, with no effect on acsl4 or acsl5 expression (Fig. 11). In addition, insulin had no effects on cte1 expression in adult rat cardiomyocytes (Fig. 11). In contrast, insulin significantly increased cyclophilin expression in adult rat cardiomyocytes (71,192 ± 1,221 vs. 105,463 ± 2,353 cyclophilin mRNA molecules/20 ng total RNA, for control vs. insulin-treated cardiomyocytes, respectively). However, insulin did not affect the expression of a second “housekeeping” gene, namely gapdh (Fig. 11), suggesting total RNA levels were not different between the two groups.

Fig. 11.

Effects of insulin on acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and gapdh (G) gene expression in isolated adult rat cardiomyocytes. Cardiomyocytes were incubated either in the absence (control) or presence of insulin (10 nM) for 24 h. Values are shown as means ± SE for 5 observations in each group. Data are represented as mRNA molecules/∼20 ng total RNA. **P < 0.01 and ***P < 0.001 vs. control.

Myocardial expression of acsl isoforms and cte1 in CIRKO mouse hearts.

To investigate further the potential influence of insulin on myocardial expression of acsl isoforms and cte1, we utilized the CIRKO mouse. Myocardial-specific loss of the insulin receptor was associated with decreased expression of acsl1, acsl6, and cte1, compared with wild-type hearts (Fig. 12). No significant differences were observed for acsl3, acsl4, or acsl5 between hearts isolated from wild-type and CIRKO mice (Fig. 12). In contrast, cyclophilin expression was slightly, but significantly, higher in hearts isolated from CIRKO mice (4,053 ± 250 cyclophilin mRNA molecules/ng total RNA), compared with wild-type hearts (3,220 ± 152 cyclophilin mRNA molecules/ng total RNA). However, the expression of a second “housekeeping” gene, namely gapdh, was not significantly different between hearts isolated from wild-type and CIRKO mice (Fig. 12), suggesting that total RNA levels were not different between the two groups.

Fig. 12.

Expression of acsl1 (A), acsl3 (B) acsl4 (C), acsl5 (D), acsl6 (E), cte1 (F), and gapdh (G) in hearts isolated from age-matched wild-type and cardiac-specific insulin receptor knockout (CIRKO) mice. Values are shown as means ± SE for 6 observations in each group. Data are normalized to nanograms total RNA. **P < 0.01 and ***P < 0.001 vs. wild-type mice.

DISCUSSION

The purpose of the present study was to investigate whether distinct acsl isoforms, as well as cte1, were differentially regulated at the transcriptional level in the heart. We hypothesized that such a differential regulation may play an important role in fatty acid channeling and decreased myocardial efficiency, both of which are observed during conditions associated with increased fatty acid availability, such as diabetes. Consistent with this hypothesis, we report that STZ-induced diabetes causes a dramatic induction of cte1 in the rat heart, decreases myocardial acsl6 expression, with no significant effects on acsl1, acsl3, acsl4, or acsl5 expression. Identical observations were made during fasting, but not high-fat feeding, of rats. Through the use of multiple in vivo and in vitro manipulations, we report that fatty acids, likely through PPARα, induce the expression of cte1 in the heart. Similarly, acsl1 is induced by fatty acids and PPARα-activation in vitro, although increased fatty acid availability in vivo cannot increase expression of this acsl isoform further (because it may be maximally expressed). In contrast, the repression of myocardial acsl6 during diabetes (and fasting) is likely due to hypoinsulinemia. We therefore conclude that acsl isoforms and cte1 are differentially regulated in the heart, in which acsl1 and cte1 are PPARα-regulated genes, whereas acsl6 is an insulin-regulated gene (see Table 2). Whether these changes in gene expression contribute toward fatty acid channeling and/or decreased myocardial efficiency during diabetes requires further investigation.

View this table:
Table 2.

Relationship between PPARα, insulin, and myocardial expression of acsl isoforms and cte1

Regulation of acsl isoform gene expression in heart.

Significant evidence exists to suggest differential regulation of acsl isoforms in tissues, such as the liver and adipose (7, 38). For example, acsl1 and acsl4 are specifically induced in the liver by PPARα agonists (21). In contrast, PPARγ agonists induce acsl1 and acsl5 in white and brown adipose tissue, respectively (7). Far less is known regarding the regulation of acsl isoforms in the heart. To date, studies investigating the heart have focused primarily on acsl1 expression. Indeed, several reports (9) suggest that specific acsl isoforms are not expressed in the heart (e.g., acsl3). The present study clearly shows, through the use of real-time quantitative RT-PCR, that the mRNA of all five known isoforms of acsl are expressed in the heart (Fig. 1). To ensure that the acsl isoforms are expressed within the cardiomyocytes, as opposed to nonmyocyte cell types found within the intact heart (e.g., endothelial cells and fibroblasts), we utilized isolated adult rat cardiomyocytes. Relative abundance of the acsl isoforms was identical between intact mouse and rat hearts, as well as isolated cardiomyocytes; this relative level of expression was acsl1 > acsl5 > acsl4 > acsl3 > acsl6 (Fig. 1). Interestingly, of the five acsl isoforms investigated, only acsl3 exhibited marked circadian rhythms in expression (Fig. 2). The latter circadian rhythm is also observed for the mouse heart (Durgan and Young, unpublished observations). Specific circadian oscillations in acsl3 provide the first evidence that distinct acsl isoforms are differentially regulated in the rodent heart. Intriguingly, oscillations in acsl3 mRNA appear to persist in cultured cardiomyocytes (Fig. 6), indicative of regulation by the intramyocellular circadian clock.(8)

Chronic (4 wk) STZ-induced diabetes was associated with decreased myocardial acsl6 expression, with no significant effects on acsl1, acsl3, acsl4, or acsl5 (Fig. 3). Type 1 diabetes mellitus is associated with a plethora of neurohumoral changes. In the perspective of myocardial metabolism, a selection of the most notable humoral changes occurring during diabetes includes hyperglycemia, hyperlipidemia, and hypoinsulinemia. Similar to diabetes, prolonged fasting results in hypoinsulinemia and elevation of plasma NEFA but is not associated with hyperglycemia. We therefore investigated whether prolonged fasting (up to 24 h) mimicked the effects of chronic diabetes on myocardial acsl isoform gene expression. Consistent with STZ-induced diabetes, prolonged fasting selectively repressed acsl6 expression, with no effects on the other acsl isoforms (Fig. 4). These observations suggest that basal expression of myocardial acsl6 is repressed by fatty acids and/or dependent on insulin. In an attempt to investigate the former possibility, rats were fed a HFD for 4 wk to increase circulating fatty acid levels (without affecting circulating insulin levels) (33). However, unlike diabetes and fasting, high-fat feeding had no significant effects on the expression of any acsl isoforms in the heart (Fig. 5). These data suggest that fatty acids may not repress acsl6 expression in the rat heart.

We investigated further the potential influence of fatty acids on the regulation of myocardial acsl isoform gene expression through the use of isolated adult rat cardiomyocytes. Consistent with the high-fat feeding studies (Fig. 5), oleate (a highly abundant, monounsaturated, 18-carbon fatty acid) did not affect acsl6 expression in isolated adult rat cardiomyocytes (Fig. 6). In contrast, oleate induced acsl1, acsl3, and acsl4 in both a time- and concentration-dependent manner and caused a slight, but significant, repression of acsl5 expression. Fatty acids exert their effects on gene expression through a number of molecular mechanisms (26). One such mechanism involves the direct binding of fatty acids (or fatty acid derivatives) to, and subsequent activation of, the nuclear receptor PPARα (1). On activation, PPARα, in combination with its heterodimerization partner retinoid X receptor (RXR), binds to fatty acid responsive elements in the promoter of various target genes (18). The latter includes key regulators of both fatty acid and carbohydrate metabolism (3, 5, 12, 33, 37, 39, 40, 42). We therefore investigated whether specific activation of PPARα, through use of the specific PPARα agonists WY14643 and fenofibrate, affected acsl isoform gene expression in isolated adult rat cardiomyocytes. WY-14643 (1 μM) and fenofibrate (10 μM) both induced acsl1 gene expression without affecting acsl3, acsl4, acsl5, and acsl6 expression (Figs. 7 and 8). These observations are consistent with the hypothesis that oleate induced acsl1 gene expression through PPARα. As such, focus was turned toward in vivo models in which PPARα activity was either specifically increased (through acute administration of rats with WY-14643) or decreased (through the use of PPARα null mice). In the latter case, hearts isolated from PPARα null mice exhibit decreased acsl1 expression, compared with wild-type hearts (Fig. 10). In contrast, acute activation of PPARα in vivo (4 h; thereby reducing potential systemic effects associated with chronic pharmacological interventions) had no significant effects on myocardial acsl1 gene expression (Fig. 9). Collectively, both our in vitro and in vivo observations are consistent with previous suggestions that acsl1 is a PPARα-regulated gene in the heart (Table 2) (37). It is likely that the high basal expression of acsl1 in the heart in vivo attenuates further induction during diabetes, fasting, high-fat feeding, and specific PPARα activation.

As described above, acsl6 expression is decreased in the heart during diabetes and fasting (Figs. 3 and 4). Given that activation of PPARα within the cardiomyocyte, either in vivo (high-fat feeding and WY-14643 administration) or in vitro (challenging adult rat cardiomyocytes with oleate, WY-14643, or fenofibrate), did not affect acsl6 expression, it is unlikely that activation of PPARα during diabetes and fasting is responsible for repression of myocardial acsl6. We therefore hypothesized that insulin influenced expression of acsl6 (and potentially other acsl isoforms) in the heart. Challenging isolated adult rat cardiomyocytes with insulin (10 nM for 24 h) resulted in induction of acsl6 (as well as acsl1 and acsl3; Fig. 11). Conversely, loss of insulin signaling specifically within murine cardiomyocytes in vivo (i.e., CIRKO hearts) results in a dramatic decrease in acsl6 expression (as well as decreased expression of acsl1; Fig. 12). Taken together, these data strongly support the hypothesis that acsl6 is an insulin-regulated gene in the rodent heart and that decreased acsl6 expression during diabetes and fasting is likely due to decreased myocardial insulin stimulation. These data also suggest that acsl1 may be an insulin-regulated gene in the heart.

It should be noted that acsl6 expression was elevated in the PPARα null mouse heart compared with wild-type hearts (Fig. 10). The PPARα null mouse is a model in which both a chronic and ubiquitous loss of PPARα has occurred. As such, multiple secondary adaptive mechanisms will be activated in the adult mouse heart, one of which might have resulted in the induction of acsl6. Interestingly, PPARα null mice on a 129Sv background have been reported to have a slight elevation in circulating insulin levels compared with wild-type mice (11). Consistent with the latter observations, we find that plasma insulin levels are 1.6-fold higher in PPARα null mice compared with wild-type littermates. Whether increased myocardial insulin signaling in the PPARα null mouse heart contributes to increased acsl6 expression is an attractive possibility.

Regulation of cte1 gene expression in rodent heart.

We have recently reported that mte1, like ucp3, is a PPARα-regulated gene in the heart and have speculated that concomitant induction of mte1 and ucp3 during periods of increased fatty acid availability likely promotes increased rates of fatty acid oxidation (33). Here, we report that cte1, like mte1, is expressed in the rodent heart. Basal levels of cte1 expression are relatively low in the heart, compared with mte1 (see results). However, during periods of increased fatty acid availability, cte1 expression is dramatically induced. For example, during STZ-induced diabetes, cte1 expression is induced 26-fold (Fig. 3), whereas mte1 is induced only 2-fold (33). As such, cte1 can be considered a gene the activity of which is relatively low in the basal state but can be activated dramatically in response to increased fatty acid availability.

It is clear from the present study that cte1, like mte1, is PPARα-regulated in the heart (Table 2). Increased fatty acid availability induces cte1 expression both in vivo (diabetes, fasting, and high-fat feeding; Figs. 35) and in vitro (oleate; Fig. 6). These observations are mimicked by the specific PPARα agonists WY-14643 and fenofibrate (Figs. 79). Furthermore, cte1 expression is decreased by 80% in hearts isolated from PPARα null mice, as opposed to hearts isolated from wild-type mice (Fig. 10). These observations provide strong evidence that cte1 is a PPARα-regulated gene in the heart. This is consistent with previously published reports (16) that cte1 is PPARα regulated in the liver. Furthermore, decreased expression of cte1 (and acsl1) in hearts isolated from CIRKO mice is consistent with previous reports (26) that PPARα-regulated genes are repressed in this model. Indeed, we observe decreased expression of known PPARα-regulated genes (ucp3 and mte1) in the CIRKO hearts utilized in the present study compared with wild-type hearts (Durgan and Young, unpublished observations).

Hypothetical fatty acid channeling at expense of myocardial efficiency.

The mechanisms responsible for channeling fatty acids into oxidative versus nonoxidative pathways are unknown in the heart. Similarly, our current understanding of the molecular mechanisms responsible for decreased myocardial efficiency in the face of increased fatty acid availability remains in its infancy. Figure 13 proposes a potential model in which fatty acid channeling would contribute toward decreased myocardial efficiency during conditions of increased fatty acid availability, such as diabetes. Consistent with data reported for adipose, ACSL1 may reside at the cell surface, in close proximity to the fatty acid transporter CD36/FAT (38). In doing so, ACSL1 would promote fatty acid uptake into the cell, by maintaining a diffusion gradient such that the fatty acyl-CoA is generated immediately after import of the fatty acid. The high level of myocardial expression of acsl1 observed in the present study is consistent with high rates of fatty acid uptake by the heart. A second class of FATPs has been shown to possess an intrinsic acyl-CoA synthetase activity, which promotes fatty acid uptake (13). According to our proposed model (Fig. 13), once the fatty acyl-CoA generated at the cell surface moves deeper into the cytosol, it would become an available substrate for CTE1, resulting in generation of the NEFA. The latter would then become available for channeling into specific metabolic pathways, by distinct ACSL isoforms located within different compartments of the cardiomyocyte. Studies in the liver suggest that ACSL4 and ACSL5 channel fatty acids into peroxisomal and mitochondrial metabolism, respectively, through association with these organelles (7). Clearly, further studies are required in the heart to allow designation of a specific ACSL isoform with a particular metabolic pathway.

Fig. 13.

Proposed model for ACSL-mediated fatty acid (FA) channeling and futile cycling in heart. After entry of FAs into cardiomyocyte [mediated by CD36/FA transport (FAT) or 1 of 6 FAT proteins (FATP) family members; FATP1 is shown], the fatty acyl-CoA is rapidly generated. The latter is catalyzed by either an ACSL isoform or endogenous acyl-CoA synthetase activity possessed by FATP. Once within the bona fide cytosol, fatty acyl-CoA is hydrolyzed back to a free FA by CTE1. Distinct ACSL isoforms channel FAs into oxidative and nonoxidative pathways. Additional hydrolysis of fatty acyl-CoAs by CTE1 or mitochondrial thioesterase 1 (MTE1) reactions will generate futile cycles at the expense of ATP utilization; each ACSL reaction utilizes 2 ATPs. Numbers in parenthesis denote fold induction of key enzymes in rat heart during uncontrolled STZ-diabetes (at the transcriptional level): for origins of fold change data, see Fig. 3 (cte1), Ref. 33 (ucp3 and mte1), and unpublished observations (Durgan and Young; cd36/fat and fatp1). Induction of CTE1 and MTE1, as well as uncoupling protein 3 (UCP3), during periods of increased FA availability (e.g., diabetes) may contribute to decreased myocardial efficiency via ATP utilization at the ACSL reactions.

Classically, it is assumed that two ATP equivalents are required to “activate” a fatty acid for subsequent entry into a specific metabolic pathway (as the ACSL reaction cleaves the phosphate bond between the α- and β-phosphate groups of ATP during acyl-CoA synthesis). However, in the proposed pathway illustrated in Fig. 13, at least four ATPs are required for each fatty acid to be channeled into a specific pathway. This number of ATP would be increased further if futile cycling occurs between the CTE1 and ACSL reactions. The dramatic induction of cte1 in the heart during diabetes (Fig. 3) would promote such a futile cycling. Furthermore, futile cycling of fatty acyl groups between the cytosol and mitochondrial matrix (via the carnitine shuttle and MTE1/UCP3 system) would decrease myocardial efficiency further through ATP hydrolysis at the ACSL reaction (Fig. 13).

Study limitations.

The purpose of the proposed study has been to investigate the transcriptional regulation of acsl isoforms and cte1 in the rodent heart. Consistent with the hypothesis that different ACSL isoforms channel fatty acids into distinct metabolic pathways, we have found that these isoforms are differentially regulated at the transcriptional level in the heart. Furthermore, consistent with the hypothesis that increased futile cycling contributes toward reduced myocardial efficiency in the presence of increased fatty acid availability, we have found that myocardial cte1 expression is dramatically increased by fatty acids, both in vitro and in vivo. However, the present study has not investigated whether the observed changes in gene expression are mirrored by changes in protein expression and/or activity of the distinct ACSL isoforms or CTE1. Furthermore, we have not investigated the subcellular localization of the distinct ACSL isoforms within the cardiomyocyte or identified the metabolic pathway into which each isoform channels fatty acyl groups. Such studies are hindered by the lack of availability of commercial antibodies specific for these different enzymes. Future studies are required to elucidate these important questions. In an attempt to initiate such studies, we have recently generated an antibody against ACSL1; through use of this antibody, we found that total ACSL1 protein levels are not different between hearts isolated from control and STZ-diabetic rats (Cuthbert and Dyck, unpublished observations), consistent with the acsl1 gene expression data reported for these samples.

In conclusion, the present study reports differential regulation of distinct acsl isoforms, as well as cte1, in the rodent heart at the transcriptional level. We have found that during STZ-induced diabetes and fasting, myocardial expression of cte1 is dramatically induced, whereas acsl6 expression is repressed. Our data suggest that induction of cte1 during diabetes is likely through fatty acid activation of PPARα, whereas repression of acsl6 is likely due to hypoinsulinemia. Whether these changes in gene expression contribute toward fatty acid channeling and/or decreased myocardial efficiency during diabetes requires further investigation.

GRANTS

This work was supported by the American Heart Association Texas Affiliate Beginning Grant-In-Aid Award No. 0365028Y and by the National Heart, Lung, and Blood Institute Grants HL-074259-01 (to M. E. Young) and HL-070070 (to E. D. Abel).

Acknowledgments

We thank Drs. Molly S. Bray and William C. Stanley for aiding with statistical analysis of the data presented and constructive criticisms before submission of this manuscript. E. D. Abel is an established investigator of the American Heart Association. J. R. B. Dyck is a Scholar of the Alberta Heritage Foundation for Medical Research and a Canadian Institutes of Health Research New Investigator. K. D. Cuthbert holds a Natural Science and Engineering Research Council postgraduate scholarship. V. G. Zaha is supported by a postdoctoral fellowship from the American Heart Association (Western Affiliates).

Footnotes

  • * D. J. Durgan and J. K. Smith contributed equally to this work.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract