RESEARCH DESIGN AND METHODS

Five-week-old male Lewis rats were randomly divided into five groups: control, untreated diabetes (STZ), and diabetes treated with ICT or insulin implants (insulin), PKCβ selective inhibitor ruboxistaurin, or ACE inhibitor captopril. Diabetes was induced with STZ (70 mg/kg body wt i.p.) (Sigma, St. Louis, MO). Blood pressure was measured weekly with tail plethysmography. Four weeks after STZ injection, treatments for an additional 4 weeks was initiated in the ICT/insulin, ruboxistaurin, and ACE inhibitor groups. ICT in rats was performed as previously described (15). For some experiments, rats with insulin implants (2 units/day) (Linshin, Toronto, Ontario, Canada) were used in place of ICT rats. Ruboxistaurin and ACE inhibitor groups received 0.01% wt/wt ruboxistaurin (Eli Lilly, Indianapolis, IN) in diet or 1 mg/ml captopril (Sigma) in drinking water, respectively. All procedures were conducted according to Joslin Diabetes Center's Animal Care and Use Committee guidelines. On the day of termination, food was removed, and all rats were provided water only 5 h before being killed.

Preparation of RNA for microarray study.

Total RNA was isolated from ventricles using TRI Reagent (Molecular Research Center, Cincinnati, OH) and purified with RNeasy columns (Qiagen, Valencia, CA). Biotinylated cRNA was prepared and hybridized to the rat Affymetrix RG-U34A GeneChip using standard protocols provided by Affymetrix (Santa Clara, CA). Arrays were then washed, scanned, and analyzed with MAS 5.0 software (Affymetrix) as previously described (16). cRNA prepared from one animal was used for each array, and three microarrays formed the data set for each treatment group.

Analysis of microarray data.

Gene expression data were globally scaled in MAS 5.0. Subsequent data analysis was performed using GeneSpring 5.0 software (Silicon Genetics, Redwood City, CA). Annotation of genes was done using UNCHIP, ONTO-EXPRESS, and GENMAPP softwares (16). Differences in gene expression across all groups were analyzed by ANOVA with multiple comparison correction using the Benjamini-Hochberg method (16). Differences between individual groups were evaluated by t test with a significance threshold of <0.05. In addition, gene expression in control and STZ rats was compared with gene expression in rats from ICT, ruboxistaurin, and ACE inhibitor groups. The resulting list of differentially expressed genes was also compared with the genes identified using the twofold change filter. K-means clustering and correlation analysis were performed to identify coregulated genes and genes correlating with metabolic variables.

Cardiac expression of several selected genes (PDK4, UCP3, PFK-2, L-CPT1, and CD36) were also assessed by quantitative real-time RT-PCR using TaqMan PCR Master Mix Reagents kits or SYBR Green PCR Master Mix system (Applied Biosystems). Primer and probe sequences are listed as follows: PDK4 (forward 5′-TTCACACCTTCACCACATGC-3′; reverse 5′-AAAGGGCGGTTTTCTTGATG-3′; probe 5′-CGTGGCCCTCATGGCATTCTTG-3′), UCP3 (forward 5′-GTGACCTATGACATCATCAAGGA-3′; reverse 5′-GCTCCAAAGGCAGAGACAAAG-3′; probe 5′-CTGGACTCTCACCTGTTCACTGACAACTTCC-3′), PFK-2 (forward 5′-cctatgcactagccaacttc-3′; reverse 5′-cacccgcatcaatctcattc-3′; probe 5′-atcagctccctgaaagtatggactagccac-3′), L-CPT1 (forward 5′-AGAGGTGCTGGCTGAACAGT-3′; reverse 5′-GTCACAGCAGCGAGAGTCAG-3′), and CD36 (forward 5′-CTCTGACATTTGCAGGTCCA-3′; reverse 5′-CACAGGCTTTCCTTCTTTGC-3′). To assess changes in protein expression, immunoblot analyses of candidate genes were performed as described (17).

Electrophoretic mobility shift assay.

Activity of peroxisome proliferator–activated receptor (PPAR)α was measured by an electrophoretic mobility shift assay probe (5′-AAAAACTGGGCCAAAGGTCT-3′) using a kit commercially available from Panomics (Redwood City, CA) according to the manufacturer's instructions. This assay has been previously validated and described (18).

Heart membrane PKC activity assay.

Hearts isolated from three to five rats of each experimental group were used to perform PKC activity assays in the cytosolic and membranous fractions as previously described in detail (6,10).

Rat heart perfusion experiments.

An isolated isovolumetric rat heart preparation (Langendorff model) was perfused with modified Krebs-Henseleit buffer (composition in mmol/l: 118 NaCl, 5.3 KCl, 2 CaCl₂, 1.2 MgSO₄, 25 NaHCO₃, 0.5 EDTA, and 11 mmol/l d-glucose equilibrated with 95%O₂/5%CO₂ at pH 7.4) at a constant pressure of 80 mmHg and paced at a rate of 300 bpm as previously described (19,20). After a stabilization period of 30 min, the perfusate was changed to Krebs-Henseleit buffer containing 11 mmol/l [1-¹³C]glucose instead of unlabeled glucose. Perfusion was continued for another 40 min after which hearts were freeze clamped in liquid N₂ and stored at −80°C until assayed. Heart tissues were extracted, lyophilized, and dissolved in 700 μl deuterium oxide for ¹³C nuclear magnetic resonance (NMR) spectroscopy.

¹³C NMR spectroscopy.

Proton-decoupled ¹³C NMR spectra of the heart extracts were collected as previously described (21). The spectrum peak areas of the C3 and C4 carbons of glutamate were quantified using the McNuts software (Acorn NMR). The contributions of labeled glucose and unlabeled endogenous substrates to the oxidative metabolism were determined by modeling the tricarboxylic acid (TCA) cycle fluxes using the relative peak areas of C3 and C4 resonances of glutamate according to the methods of Malloy et al. (21).

Cardiac function assessment.

Hearts were perfused with Krebs-Henseleit buffer in Langendorff mode and paced at 300 bpm as described above. After 10–15 min of stabilization, coronary flow rate was recorded. Next, the water-filled balloon inserted in the left ventricle was emptied, and a left ventricular pressure–volume curve was performed by stepwise inflation of the balloon by 0.05 ml until maximal left ventricular developed pressure was obtained. The heart was then arrested in diastole by introducing high KCl (25 mmol/l) into the perfusate to determine the passive pressure volume characteristics of the left ventricle.

Cardiomyocyte culture experiment.

Neonatal rat cardiomyocytes (NRCM) were harvested from 1- to 2-day-old rats and plated to ∼100% confluency in Dulbecco's modified Eagle's medium (DMEM)/F12 medium with 5.5 mmol/l glucose and 5% horse serum on to gelatin-coated dishes for gene expression studies or Nunclon Δ polystyrene culture tubes (NalgeNunc, Roskilde, Denmark) for glucose oxidation assays (22,23). Sixteen hours after plating, culture media were replaced with DMEM/F12 plus 5.5 mmol/l glucose without serum. Three days after plating, NRCM were exposed to 0.05 mmol/l oleic acid (OA) in the presence or absence of 1 μmol/l GF109203X (GFX; general PKC inhibitor; LC Laboratories, Woburn, MA) and 50 nmol/l ruboxistaurin in DMEM/F12 media containing 0.33% FFA BSA (Sigma) for 24 h. Cell viability by the trypan blue method showed >85% viability under all conditions.

Glucose oxidation assay.

Glucose oxidation was determined by measuring ¹⁴CO₂ released from the metabolism of [U-¹⁴C]glucose as modified from the methods previously described (23). After incubation with the media containing the experimental conditions for 24 h, the media were replaced with DMEM/F12 containing 0.2 mmol/l glucose, and each culture tube of NRCM was gassed with 95%O₂/5%CO₂ for 15 s. [U-¹⁴C]glucose was then added and the tube sealed with a rubber stopper with an overhanging centerwell containing a piece of Whatman Grade 3 filter paper. For insulin stimulation, 100 nmol/l insulin was added 15 min before the addition of radio-labeled glucose. The tubes were incubated at 37°C for 4 h. Hyamine hydroxide (200 μl) (ICN Biomedicals, Irvine, CA) was then injected onto the filter paper and 0.5 ml 6M HCl injected into the media to release CO₂ and to trap ¹⁴CO₂, after which the filter paper was measured in a liquid scintillation counter. For each condition of each experiment, the assay was performed in duplicates. We confirmed in the initial studies that the protein content was similar between the culture tubes within each experiment but varied between different experiments. Hence, to correct for interexperiment variability, we compared the glucose oxidation of NRCM in each condition to that of the basal control condition, which was set at 100% for every experiment.

Expression of metabolic genes in NRCM.

The direct effects of OA on mRNA expressions of pyruvate dehydrogenase kinase (PDK)4 and uncoupling protein (UCP)3 were examined in NRCM as previously described (24) with a few modifications. Three days after plating, cells were exposed to 0.05 mmol/l OA with 0.33% FFA BSA and with or without 100 nmol/l insulin for 12 h or 1 μmol/l GFX for 24 h. Cardiomyocytes were washed twice with cold PBS, and total RNA was isolated using TRI Reagent. Expression of PDK4 and UCP3 mRNA was determined by quantitative RT-PCR as described above.

Measurements of metabolites.

Glucose levels from tail blood samples were measured with an Elite XL Glucometer (Bayer, Elkhart, IN). Serum insulin levels were measured with a radioimmunoassay kit (Linco Research, St. Louis, MO). Triglyceride and nonesterified fatty acid levels were determined by enzymatic colorimetric kits from Sigma and Wako Chemicals (Richmond, VA). Measurements of myocardial contents of glycogen and triglyceride were performed as previously described (25,26).

Statistical analyses.

Results are expressed as means ± SD unless otherwise noted. The data in the line graphs of Fig. 3 are presented as means ± SE to allow for easy viewing. Statistical analyses for comparisons among three or more groups were performed using one-way ANOVA or general linear models ANOVA, as appropriate, followed by a Newman-Keuls test. Comparisons of values non–normally distributed or with unequal variances between groups were made with the Kruskal Wallis one-way ANOVA on ranks followed by Dunn's method. Comparisons between two groups were analyzed by the parametric unpaired Student's t test or nonparametric Wilcoxon's rank-sum test as appropriate. The level of significance was set at <0.05.

RESULTS

Physiological and biochemical parameters of rats.

Induction of STZ-induced diabetes in rats resulted in hyperglycemia, hypoinsulinemia, and decreased body weight gain when compared with age-matched controls (Table 1). At week 4 of STZ-induced diabetes, diabetic rats randomized to the different treatment groups were not different in terms of body weight, blood glucose, or blood pressure. Four weeks after the interventional treatments, all diabetic rats except those treated with ICT or insulin implants had significantly higher blood glucose, lower serum insulin levels, and smaller body weight gain than control rats. Furthermore, captopril-treated rats had lower blood pressure than other groups (Table 1). Fasted plasma triglyceride and FFA levels were higher in diabetic rats than in control rats (Table 1) and were corrected by ICT but not by ruboxistaurin or captopril treatment (Table 1).

Diabetic rats in STZ, ruboxistaurin, and ACE inhibitor groups had a lower absolute ventricular weight than control and ICT rats (Table 1). However, diabetic rats in the STZ groups exhibited a greater ventricular–to–body weight ratio than the control rats. This ratio was partially normalized by ICT and totally reversed by captopril treatment but not affected by ruboxistaurin treatment (Table 1).

Untreated diabetes resulted in higher contents of glycogen (in μmol glucose/g tissue wt: STZ 46.1 ± 5.0 [n = 6] vs. control animals 12.3 ± 4.0 [n = 6], P < 0.05) and triglyceride (in μg glycerol/mg tissue wt: STZ 5.8 ± 1.7 [n = 6] vs. control 3.6 ± 0.6 [n = 6], P < 0.05) in the heart. Insulin treatment normalized diabetes-induced changes in myocardial content of glycogen (16.5 ± 6.1 μmol glucose/g tissue wt [n = 6], P < 0.05 vs. STZ) and triglyceride (3.7 ± 0.8 μg glycerol/mg tissue wt [n = 6], P < 0.05 vs. STZ). Interestingly, ruboxistaurin treatment also lowered both glycogen and triglyceride contents in diabetic rat hearts (glycogen in μmol glucose/g tissue wt: 36.1 ± 8.9 [n = 7], triglyceride in μg glycerol/mg tissue wt: 4.0 ± 1.2 [n = 7], P < 0.05 vs. STZ). Captopril treatment decreased glycogen (34.9 ± 11.3 μmol glucose/g tissue wt [n = 6], P < 0.05 vs. STZ) but not triglyceride content (5.4 ± 1.6 μg glycerol/mg tissue wt [n = 6], P > 0.05 vs. STZ) in diabetic rat hearts. Diabetes significantly increased PKC activity in the heart membranous fractions (Table 1), which was normalized in rats treated with ICT, ruboxistaurin, or captopril.

Changes in gene expression in the diabetic heart.

Of the 8,799 genes and expressed sequence tags present on the microarray, 1,985 genes were found to have differential expression (P < 0.05) among the five different groups of rats by ANOVA analysis. No single gene remained differentially expressed after controlling for multiple comparison false discovery using the Benjamini-Hochberg method. To identify genes that are differentially expressed between the diabetic and control animals, we performed unpaired t tests comparing the two groups using P < 0.05 as the cutoff value. This identified 376 genes with differential expression between the diabetic and control states (see supplementary Table 4 [found in an online appendix at http://dx.doi.org/10.2337/db06-0655]). Using MAPP FINDER (16), many of these genes were identified as those involved in fatty acid and carbohydrate metabolism as listed in Table 2.

Expression of genes involved in glucose and fatty acid metabolism.

In the STZ-induced diabetic heart, several key enzymes and transporter proteins involved in glucose and fatty acid metabolism were increased versus controls, including fatty acid translocase/CD36 (2.1-fold), fatty acid transport protein (FATP)1 (1.6-fold), liver carnitine palmitoyltransferase (L-CPT)1 (2.7-fold), UCP3 (2.9-fold), mitochondrial 2,4-dienoyl CoA reductase 1 (2.3-fold), Δ2-enoyl-CoA isomerase (1.9-fold), and PDK4 (18.7-fold) (Table 2). A decrease in 6-phoshofructo-2-kinase (PFK-2) was also noted in STZ hearts (3.4-fold vs. control) (Table 2). Unexpectedly, the expressions of all the genes mentioned above were either completely or partially normalized by treatment with ICT, ruboxistaurin, or ACE inhibitor (Fig. 1A and Table 2). Results from the unpaired t test (P < 0.05) on these genes between the STZ rats and each of the other three treatment groups showed most of these metabolism genes as being differentially expressed. Similarly, K-means clustering identified several of these genes to show a similar, coordinated expression pattern across the different treatment groups (data not shown). Real-time PCR of the identified genes (CD36, L-CPT1, UCP3, and PDK4) concurred with the microarray data on control, STZ, and ICT rat groups (Table 3). PCR analyses also demonstrated normalization in the diabetic rat myocardium of PDK4, UCP3, and PFK-2 by ruboxistaurin treatment (P < 0.05 vs. STZ) and concordant changes with microarray results in L-CPT1 expression by ruboxistaurin and captopril treatments (P > 0.05 vs. control and STZ groups). Immunoblot analyses of PDK4 and UCP3 protein expression demonstrated concordance with array data with marked increases in STZ diabetic rat hearts, which were ameliorated by insulin, ruboxistaurin, and captopril (Fig. 1B and C). Moreover, the change in L-CPT1 protein expression, which was not statistically significant, agrees with the array data in the direction of change (fold change from control: STZ, 2.3 ± 1.4; insulin, 1.2 ± 0.7; ACE inhibitor, 2.1 ± 1.2; and ruboxistaurin, 1.4 ± 0.7). Interestingly, expression and activation of PPARα, as assessed by microarray, PCR, and electrophoretic mobility shift assay activity, did not significantly differ among the five groups (Fig. 2A; E.A., R.C.W.M., G.L.K., unpublished data).

Measurement of glucose utilization in the isolated perfused rat heart.

The abnormalities of gene expression in glucose and fatty acid metabolism in the diabetic heart were normalized by ICT or insulin replacement, ruboxistaurin, and captopril. This suggests that glucose utilization in the diabetic heart could be increased by these treatments. Hence, glucose utilization was assessed in control and diabetic rats from different treatment groups using ¹³C NMR spectroscopy and Langendorff perfused hearts. In hearts from control animals, glucose contributed 83.0 ± 11.6% of the acetyl-CoA entering the TCA cycle when perfused with buffer containing ¹³C-labeled glucose as the only exogenous substrate (Fig. 2C). In STZ-induced diabetic rats, the contribution of glucose to the TCA cycle was significantly decreased to 53.0 ± 14.8% (a relative decrease of 36% vs. control, P < 0.05). This decrease in glucose utilization in the diabetic heart was normalized by ICT, as expected (85.3 ± 9.2%, P < 0.05 vs. STZ). Treatments with captopril and ruboxistaurin also significantly increased glucose utilization to 79.5 ± 10.7% and 65.4 ± 15.4%, respectively (P < 0.05 vs. STZ) (Fig. 2C).

Cardiac function in diabetic rats treated with ICT, ruboxistaurin, or captopril.

To investigate whether the alterations in cardiac glucose metabolism are associated with changes in performance, cardiac function in animals from different treatment groups was characterized using Langendorff perfusion. Untreated diabetic animals exhibited left ventricular diastolic dysfunction that was manifested by a leftward shift in the end diastolic pressure (EDP) volume curve when compared with control animals (EDP at maximum left ventricular volume tested in mmHg: control 18.6 ± 3.7 vs. STZ 27.9 ± 3.3; P < 0.05) (Fig. 3A). ICT, captopril, and ruboxistaurin significantly improved diastolic function versus STZ rats (EDP at maximum left ventricular volume tested in mmHg: ICT 15.8 ± 6.3, ACE inhibitor 20.4 ± 1.9, ruboxistaurin 20.0 ± 2.7; P < 0.05 vs. STZ) (Fig. 3A). Left ventricular systolic function was not changed in STZ, ICT, or ruboxistaurin rats when compared with control rats (Fig. 3B). There was a downward shift in the left ventricular systolic pressure volume relation of the hearts from animals treated with captopril, which was possibly due to the lower blood pressure and ventricular mass in these animals. When the passive pressure volume characteristics of the left ventricle were assessed in KCl-arrested hearts, there was a trend for a leftward shift in the pressure-volume relation of the STZ rat hearts versus the control and ICT hearts (left ventricular pressure at maximum left ventricular volume tested in mmHg: STZ 39.0 ± 6.4 vs. control 30.8 ± 5.4 and ICT 26.0 ± 5.1) (Fig. 3C). The curves of both ruboxistaurin and ACE inhibitor rat hearts shifted significantly to the right when compared with the STZ hearts (left ventricular pressure at maximum left ventricular volume tested in mmHg: ruboxistaurin 24.1 ± 2.1 and ACE inhibitor 21.0 ± 2.5; P < 0.05 vs. STZ), suggesting that the structural components of the myocardium might be normalized by ruboxistaurin and captopril treatments. Coronary blood flow in paced hearts was similar among different rat groups (in ml/min per g ventricular weight: control 22.8 ± 5.1, STZ 22.4 ± 5.5, ICT 21.5 ± 4.3, ACE inhibitor 27.1 ± 4.8, ruboxistaurin 23.0 ± 4.6; P > 0.05).

Measurement of basal and insulin-stimulated glucose oxidation in isolated cultured rat cardiomyocytes.

The normalization of PKC activation in the diabetic heart by ruboxistaurin and ACE inhibitors suggests that opposing PKC activation may underlie the actions of both inhibitors. To determine that these inhibitors mediated their effects directly on cardiomyocytes, we studied the effects of OA, one of the main plasma FFAs commonly elevated in diabetes, on glucose oxidation in vitro using NRCM. Insulin (100 nmol/l) increased oxidation of glucose to CO₂ to 213 ± 39% of the basal rate in NRCM (P < 0.001) (Fig. 4). Cell viability was similar among the different treatment groups. OA decreased both basal and insulin-stimulated glucose oxidation to 25 ± 15 and 54 ± 26% of basal values, respectively (Fig. 4). Incubation with ruboxistaurin partially restored basal and insulin-stimulated glucose oxidation from the effects of OA to 47 ± 22% (P < 0.05 vs. OA) and 100 ± 58% (P = 0.1161 vs. OA, rank-sum test) of basal values, respectively (Fig. 4). General PKC inhibitor, GFX, significantly increased basal- and insulin-mediated CO₂ production in OA-treated NRCM to 73 ± 38 and 136 ± 70% of basal values, respectively (Fig. 4).

Effects of OA on mRNA expression of metabolic genes.

In parallel with its inhibitory effect on glucose oxidation, OA also increased the mRNA expression of PDK4 by 2.2 ± 0.9–fold (P < 0.05 vs. control) and UCP3 by 12.6 ± 17.6–fold (P < 0.05 vs. control) in NRCM (Fig. 5). These increases in the expression of the two genes were attenuated by insulin and PKC inhibition with GFX (Fig. 5). However, OA did not affect PPARα activity in NRCM (Fig. 2B).

DISCUSSION

In this study, we showed that normalizing circulating insulin levels by ICT or insulin implants with the correction of metabolites could reinstate cardiac metabolic genes for glucose and fatty acid utilization and cardiac function in the insulin-deficient model of diabetes. Surprisingly, these results also revealed that PKCβ inhibitor ruboxistaurin and ACE inhibitor captopril normalized the expression of many metabolic genes, fuel metabolism, and function in the diabetic myocardium without normalizing hyperglycemia and hyperlipidemia. These unexpected findings suggest the interesting idea that PKC and ACE inhibitors may have cardioprotective effects by directly altering the gene expression in the metabolic pathways in the myocardium, in addition to their peripheral metabolic or hemodynamic actions.

Results from this study showed that not only could PKC activation be inhibited by ruboxistaurin, but it could also be secondarily normalized by ICT and captopril in the diabetic heart. PKC activation, especially the β-isoform, has been associated with multiple diabetic microvascular and cardiovascular pathologies (6,10–14). Previously, we have reported that ruboxistaurin can prevent PKC activation and cardiac pathologies in mice overexpressing PKCβ in the heart and normalize PKC activity in membranous fractions of failing human hearts (10,11,27). Angiotensin II type 1 receptor blockade has been reported to prevent the activation of PKCε in diabetic hearts (28) and block glucose-induced activation of several PKC isoforms including PKCβ1/2, -δ, and -ε in cultured rat cardiomyocytes (29). Our results suggest that PKC activation induced by diabetes is partially due to increased actions of angiotensin II on the myocardium, likely via significant activation of PKCβ isoform.

Changes of gene expressions associated with glucose and fatty acid metabolism in the diabetic heart, such as increases in UCP3, CD36, FATP (involved in fatty acid uptake), and PDK4, as shown by our microarray analysis and confirmed by RT-PCR and immunoblot studies, are expected (3,4). Normalization of these gene changes by ICT or insulin implants is likewise expected, since elevations in glucose and FFA levels that were inducing these gene changes in the diabetic heart were reversed by these treatments. In contrast, the novel effects of ruboxistaurin and captopril treatments on these metabolic genes were unexpected because abnormalities of the metabolic milieu were still present. Interestingly, the quantitative changes in these genes as measured by real-time PCR, microarray, and immunoblot tests were very different, although they changed in the same direction. This suggests that many different processes are regulating the mRNA and protein levels of these genes.

Data from our studies using perfused hearts and ¹³C NMR spectroscopy supported the idea that ruboxistaurin and captopril could improve glucose utilization and cardiac functions in the myocardium of diabetic rats but without normalization of plasma metabolites. Specifically, diastolic dysfunction, an early cardiac manifestation (1) observed in stressed diabetic rats, was normalized by ICT and was significantly improved by treatment with ruboxistaurin or captopril. Previous studies in mice with targeted overexpression or disruption of glucose (GLUT4) and fatty acid transport proteins suggested that subtle changes in substrate uptake and utilization may cause cardiac dysfunction (30,31), potentially by altering the coupling of glycolysis to glucose oxidation with increasing proton production, tissue acidosis, and decreasing cardiac efficiency. The results from the isolated hearts (Fig. 2C) reflected the changes in glucose utilization in vivo, with decreases exhibited in diabetes and normalized by insulin replacement, as previously reported by others (3,4). However, these results should be extrapolated cautiously to in vivo conditions, since the isolated heart studies were performed in the absence of fatty acids, which are the primary and preferred substrate of the myocardium. These new results suggest that PKC activation and angiotensin's actions may partly impair cardiac function via coordinated gene expression changes in glucose and fatty acid metabolic pathways, in conjunction with the effects of circulating metabolites.

The increase in triglyceride in the diabetic heart was expected since the genes involved in transporting fatty acids into cells across the sarcolemma (i.e., FATP and CD36) were increased. Moreover, the decrease in myocardial triglyceride content by insulin and ruboxistaurin treatments in the diabetic rats was likely a consequence of the normalization of FATP and CD36 genes. The marked increase in glycogen content in the hearts from untreated STZ-induced diabetic rats is thought to be the result of reduced glycolysis (3) and glucose oxidation (32), leading to increases in glucose-6-phosphate (G6P) (33). Thus, a decrease in PFK-2 action or expression in the heart of diabetic rats will decrease fructose 2,6-bisphosphate level, increase G6P levels, and reduce glycolysis (34). Insulin replacements and ruboxistaurin and captopril treatments all attenuated the glycogen accumulation in diabetic hearts, which was in concordance with the normalization of the gene expression of PFK-2 and PDK4, the enzymes involved in glucose metabolism. Accordingly, the effects of treatments on myocardial glycogen content were likely due to improvements in glucose utilization via increased glycolysis and glucose oxidation, thereby shuttling G6P away from glycogen synthesis (34).

One potential mechanism of ruboxistaurin and captopril's actions is to affect the expressions or actions of PPARα. Many metabolism genes that are changed by diabetes are reported to be regulated at the transcriptional level by PPARα (35,36). However, in this study, cardiac mRNA expression of PPARα, which some have reported to be decreased in hearts from chronically diabetic rats (36), were not changed by diabetes. It is also possible that PPARα activity might be increased in the diabetic heart due to the increased availability of fatty acids (36) or alterations in coactivator or corepressor function or expression (16). Unexpectedly, we did not observe any alterations in PPARα activity in the diabetic heart or in NRCM after incubation with OA. The lack of detectable changes in PPARα expression or its activity could be due to the duration of diabetes in these rats. Indeed, recent studies have reported that PPARα expression or activities may not be changed with short duration of STZ-induced diabetes even though glucose and fatty acid oxidation have changed. With longer duration of diabetes, PPARα expression was increased (37,38). Thus, it is possible that regulation of cardiac fuel metabolism in diabetes are mediated by PPARα- and non-PPARα–dependent pathways (37–39).

Another possibility is that captopril and ruboxistaurin could still have systemic actions thereby causing secondary effects on the myocardium. However, the lack of changes in circulating metabolites and the finding that PKC inhibition can directly attenuate fatty acid–mediated induction of the metabolic genes PDK4 and UCP3 and restore OA-induced inhibition of glucose oxidation in cultured cardiomyocytes support a direct action of PKC inhibitors on the myocardium. More importantly, we demonstrated that OA-mediated suppression of both basal- and insulin-induced glucose oxidation could be reversed by PKC inhibition, suggesting that PKC activation can antagonize glucose utilization through inhibition of basal metabolism as well as insulin action in cardiomyocytes. We have previously reported that PKC activation can inhibit insulin-induced Akt activation and endothelial nitric oxide synthase expression in various vascular cell types (40,41). FFA in skeletal muscle has been reported to change glucose metabolism by inhibiting insulin signaling via PKC activation (42). These effects of FFA are linked with increases in diacylglycerol mass and membrane-associated PKC-β1, -β2, and -δ (43), which are similar to the effects of hyperglycemia on diacylglycerol synthesis (6) and the impairments of cardiomyocyte function (44).

Our in vitro experiments using NRCM demonstrated reduced glucose oxidation when cardiomycoytes were cultured in media containing OA. There are several limitations in these studies. First, NRCM and adult cardiomyocytes are different. However, multiple studies have shown that glucose utilization of the two cells are comparable and have been used for the study of cardiac metabolism and substrate utilization (24). Second, the inhibitory effects of OA on glucose oxidation were achieved at levels below those observed in pathophysiological conditions; thus, it is not clear that the findings from NRCM can be directly translated to in vitro conditions. On the other hand, it has been reported that OA would not induce apoptosis or significantly affect cell viability at a concentration of 0.5 mmol/l up to 48 h (45). Indeed, we did not observe any difference in cell viability between the control and the NRCM exposed to OA at 0.05 mmol/l for 24 h. Hence, the low concentration used for OA in this study allows one to exclude the possibility that OA-induced effects on glucose oxidation were primarily due to changes in cell viability. Another potential limitation of our studies is that no data concerning contractility or the actions of ACE inhibitor in cardiomyocytes were obtained. These studies will need to be done in the future, since they will greatly increase the scope of the present goal, which is to characterize the role of PKC activation in diabetic cardiac metabolism and function.

Besides metabolic dysfunction, ruboxistaurin and captopril also caused a rightward shift in the passive pressure-volume relation in diabetic hearts, which suggests that they decreased the stiffness of the ventricular wall in diabetes. This finding supports our previous report that mice with cardiac-specific overexpression of PKCβ isoform developed cardiac hypertrophy and fibrosis similar to diabetic cardiomyopathy, which were reversed by ruboxistaurin treatment (10,11). Similarly, inhibition of ACE has been shown to reduce interstitial and perivascular fibrosis in hearts of diabetic rats (46). These new findings suggest that PKC activation could contribute to the increase in stiffness of the myocardium induced by diabetes or angiotensin II, possibly by increasing the expressions of transforming growth factor-β or connective tissue growth factor as previously reported (11,22,46).

Our studies did not address the roles of PKC isoforms, specifically in substrate metabolism of the heart. PKCβ knockout mice have been reported to have an enhanced glucose transport in adipocytes and soleus muscle (47). A recent proteomic study by Mayr et al. (48) showed that certain glycolytic enzyme isoforms were decreased, while enzymes involved in lipid metabolism were more abundant in the hearts of PKCδ knockout mice versus wild-type mice. This suggests a shift from glucose to lipid metabolism in the absence of PKCδ in the heart. In our study, PKCβ inhibitor treatment alone improved cardiac glucose utilization in diabetic rats, which is consistent with previous reports of PKCβ isoform being activated in the myocardium of diabetic rodents and heart failure patients (6,11,27). However, in vitro studies using NRCM showed only a partial restoration by ruboxistaurin of basal and insulin-stimulated glucose oxidation in isolated cardiomyocytes exposed to OA. These findings suggest that PKCβ as well as other isoforms may inhibit insulin's actions on glucose utilization through a direct effect on the expression of metabolic genes in the myocardium.

In conclusion, these results indicated a direct mechanism by which PKCβ and ACE inhibitors can partly ameliorate cardiac metabolism, structure, and function in diabetes by normalizing the gene expression involved in fuel metabolism in the myocardium. This was shown to be true even with the continuous presence of elevated plasma FFA and hyperglycemia.