Hyperglycemia-Induced Protein Kinase C β₂ Activation Induces Diastolic Cardiac Dysfunction in Diabetic Rats by Impairing Caveolin-3 Expression and Akt/eNOS Signaling

Male Sprague-Dawley rats (aged 8 weeks) weighing 260 ± 10 g equilibrated to surroundings for 3 days before experiments. Diabetes was induced via single tail vein injection of STZ (60 mg/kg; Sigma, St. Louis, MO) dissolved in citrate buffer (0.1 mol/L, pH 4.5), while control rats were injected with an equal volume of citrate buffer alone. One week after STZ injection, rats exhibiting hyperglycemia (blood glucose ≥16.7 mmol/L) were considered diabetic and were subjected to outlined experiments. One week after diabetes induction, rats were treated with vehicle or PKCβ inhibitor LY333531 (also named ruboxistaurin, a drug that has been approved by the U.S. Food and Drug Administration for the prevention of vision loss in patients with diabetic retinopathy [22]) by oral gavage for 4 weeks at a dose of 1 mg/kg/day (demonstrated to adequately inhibit PKCβ activation in rat heart and vasculature [23,24]). This model was chosen based on our most recent study (25) and a study of others (26) showing that STZ diabetic rats developed cardiac dysfunction 35 days after STZ injection, with concomitant cardiomyocyte hypertrophy and cardiac fibrosis formation (25)—two major features of diabetic cardiomyopathy. After 4 weeks’ treatment, cardiac functions were determined; the rats were then deeply anesthetized with sodium pentobarbital (65 mg/kg), and hearts were either rapidly excised for cardiomyocyte isolation or frozen in liquid nitrogen for later analysis. Subgroups of control and untreated diabetic rats were terminated at 8 weeks of STZ-induced diabetes, and heart tissue samples were processed to analyze changes of cardiac PKCβ₂ and Cav-3 at a relatively later phase of the disease. All experiments performed conformed to the Guide for the Care and Use of Laboratory Animals, published by the National Institutes of Health (NIH publication no. 86–23, revised 1996), and were approved by the Institutional Animal Care and Use Committee of Hong Kong University.

At the conclusion of 4 weeks’ treatments, transthoracic echocardiography was performed at experiment termination via a 17.5-MHz linear array transducer system (Vevo 770, High Resolution Imaging System; VisualSonics, Toronto, Ontario, Canada), and left ventricular (LV) dimensions and LV diastolic and systolic function were assessed by M-mode and Doppler echocardiography as we previously described (25). LV internal dimensions at end systole (LVIDs) and diastole (LVIDd) were used to calculate fractional shorting (FS) by the following formula: FS (%) = (LVIDd − LVIDs)/LVIDd × 100%. LV posterior wall dimensions at end diastole (LVPWd) and systole (LVPWs) were used to calculate fractional LV posterior wall thickening (FLVPW) by the following formula: LVPW(%) = LVPWs − LVPWd/LVPWd × 100%. The peak velocity of early (E) and late (A) diastolic filling was used to calculate the ratio of E and A (E/A). LV end diastolic volume (LVVd) and end systolic volume (LVVs) were used to calculate ejection fraction (EF) by the following formula: EF (%) = LVVd − LVVs/LVVd × 100%. The heart rate (HR), systolic interventricular septal thickness, diastolic interventricular septal thickness, LV isovolumic relaxation time (IVRT), and stroke volume (SV) were also monitored. All echocardiographically derived measures were obtained by averaging the readings of three consecutive beats.

After the completion of 4 weeks’ treatment, cardiomyocyte cross-sectional diameters were assessed by hematoxylin-eosin–stained paraffin-embedded sections of left ventricles (1–2 μm) longitudinally orientated to the muscle fibers in the subendocardium and subepicardium as previously described (6). Cross-sectional areas were randomly selected in five fields that visualized capillary profiles and nuclei. Images of the left ventricle sections were captured by an Axisoplus image-capturing system (Zeiss) and analyzed by Axiovision Rel. 4.5 image-analyzing software. A minimum of 150 cells per animal was chosen for analysis.

Calcium-tolerant cardiomyocytes were prepared from rat ventricles via a modified method as previously described (27). Cells isolated from a single rat heart were plated on Matrigel-coated culture dishes and allowed to recover for 3 h. Cultured ventricular cardiomyocytes were incubated in low glucose (LG) (5.5 mmol/L), high glucose (HG) (25 mmol/L), or mannitol/glucose (19.5 mmol/L mannitol plus 5.5 mmol/L glucose) at 37°C in Medium 199 (Gibco, Grand Island, NY) containing various treatments and then snap-frozen in liquid nitrogen for future analysis. Lactate dehydrogenase (LDH) release (a measure of cell injury) in culture medium was detected via a commercial LDH kit (Roche, Mannheim, Germany).

Isolated cardiomyocytes were homogenized in lysis buffer. A total of 500 μg cell extracts were subjected to immunoprecipitation with 2 μg Cav-3 primary antibody in the presence of 20 μL protein A/G plus-agarose. After extensive PBS washes, the immunoprecipitates were denatured with 1× sodium dodecyl sulfate loading buffer and subjected to analysis for PKCβ₂ expression by Western blot as described below.

Isolated cardiomyocytes were plated on Matrigel precoated glass coverslips, incubated either in LG or HG in Medium 199 for 36 h, and fixed in ice-cold acetone for 5 min. The fixed cells were blocked in PBS with Tween 20 (PBST) with 10% goat serum and 1% BSA for 30 min and further incubated with a mixture of mouse against rat Cav-3 antibody (1:50; Santa Cruz Biotechnology, Santa Cruz, CA) and rabbit against rat PKCβ₂ antibody (1:100; Santa Cruz Biotechnology) in 1% BSA in PBST in a humidified chamber for 1 h at room temperature. After three PBST washings, the cells were incubated for 1 h with a mixture of Alexa Fluor 488 goat anti-mouse IgG and Alexa Fluor 594 goat anti-rabbit IgG (1:2,000; Invitrogen, Carlsbad, CA). Cells were washed three times and prepared for confocal laser scanning microscopic imaging with mounting medium with DAPI (Vector Laboratories, Burlingame, CA).

Embryonic rat cardiac H9C2 cells were maintained in Dulbecco’s modified Eagle’s medium containing 10% FBS in a humidified atmosphere (5% CO₂) at 37°C. Commercial PKCβ₂ siRNA and Cav-3 siRNA (Santa Cruz Biotechnology) were used for inhibition of both PKCβ₂ and Cav-3 expression per the manufacturer’s protocol. After transfection with control, PKCβ₂, or Cav-3 siRNA, cells were incubated in either LG or HG in Dulbecco’s modified Eagle’s medium for 36 h and snap-frozen in liquid nitrogen.

Frozen heart tissues were pulverized separately with mortar and pestle in liquid nitrogen, homogenized in ice-cold PBS, and centrifuged at 3,000g for 15 min at 4°C for supernatant collection. The supernatant protein concentration was determined via a Lowry assay kit (Bio-Rad, Hercules, CA). Concentrations of nitrites (NO₂⁻) and nitrates (NO₃⁻), the stable end products of NO, were determined by the Griess reaction as previously described (28). NO levels were expressed as nanomoles per microgram of protein. Myocardial O₂⁻ production was determined via lucigenin chemiluminescence method (29,30). The supernatant samples were loaded with dark-adapted lucigenin (5 μmol/L) and read in 96-well microplates by luminometer (GloMax; Promega) with and without pretreatment with the NOS inhibitor L-N^G-nitroarginine methyl ester (L-NAME) (100 μmol/L [15]) for 30 min at room temperature. Light emission, expressed as mean light units/min/100 μg protein, was recorded for 5 min. Myocardial nitrotyrosine levels (micrograms per milligram protein) in the collected supernatant were determined by chemiluminescence detection via the Nitrotyrosine Assay kit per the manufacturer's protocol (Millipore).

For characterization of subcellular distributions of targeted proteins, cytosol and membrane fractions of cardiac tissue lysate were separated by ultracentrifugation as previously described (5). Cytosol and membrane fractions were denatured by 5× sodium dodecyl sulfate loading buffer and subjected to analysis for PKCβ₁ and PKCβ₂ expression by Western blot as described below.

Caveolae were isolated by discontinuous sucrose gradient centrifugation as previously described (21). Each heart sample gradient was separated into 12 fractions. Fractions 4–6 were considered the lipid raft fractions, and fractions 8–12 were considered the heavier fractions. Equal protein amounts were loaded for Western blot analysis.

Equal protein amounts from isolated cardiomyocytes, H9C2 cells, and rat heart homogenate were resolved by 7.5–12.5% SDS-PAGE and subsequently transferred to polyvinylidene fluoride membrane for immunoblot analysis as previously described (31).

Densitometry was obtained by image analysis software (Bio-Rad). All values are presented as means ± SEM. Comparisons between multiple groups were made by one-way ANOVA followed by Tukey test for multiple comparisons. Statistical analysis was performed by GraphPad Prism (GraphPad Software, San Diego, CA). P values <0.05 were considered significant.

We previously reported activation of the PKCβ₂, but not PKCβ₁, isoform in the diabetic heart (6). In the current study, we examined whether PKCβ₂ activation was associated with abnormal Cav-3 expression, a muscle-specific marker of caveolae (16). Diabetes moderately increased PKCβ₂ phosphorylation on Thr⁶⁴² residue (data not shown), but the increase in phosphorylation of PKCβ₂ was most profound at Ser⁶⁶⁰, without influencing total PKCβ₂, resulting in a markedly increased ratio of phosphorylated PKCβ₂ to total PKCβ₂ (Fig. 1A). Decreased Cav-3 expression was observed in cardiomyocytes isolated from 8-week diabetic rat hearts compared with age-matched controls (Fig. 1B). We next examined the relationship between Cav-3 and PKCβ₂ by immunoprecipitation experiments in isolated cardiomyocytes. While a small amount of PKCβ₂ remained constitutively associated with Cav-3 during basal conditions, the diabetic condition increased its association with Cav-3 (Fig. 1C). To confirm our findings, we used confocal immunofluorescence staining. Limited PKCβ₂ was present during basal conditions in association with Cav-3 in the cell membrane (indicated by scant yellow punctate staining of the cell periphery); 36 h of HG stimulation significantly increased regions of colocalization between PKCβ₂ and Cav-3 compared with LG stimulation (Fig. 1D).

HG conditions significantly increased the ratio of p-PKCβ₂ to total PKCβ₂ (indicating PKCβ₂ activation) in cardiomyocytes within 1 h for up to 48 h (Fig. 2A). Peak increase in the ratio of p-PKCβ₂ to total PKCβ₂ occurred after 12 h HG exposure. The osmotic control mannitol exerted no effects upon p-PKCβ₂ to total PKCβ₂ and Cav-3 expression (data not shown). In contrast to the quick increase of phosphorylated (p)-PKCβ₂/total PKCβ₂ as early as 1 h after HG exposure, Cav-3 expression did not significantly increase until 6–12 h after HG exposure, reduced to basal levels within 24 h, and significantly decreased 36–48 h after initial HG exposure (Fig. 2B). Cardiomyocyte LDH release significantly increased 24 h after HG exposure, with a rising tendency continuing 36–48 h after initial HG exposure (Fig. 2C).

We next investigate the interplay between PKCβ₂ activation and caveolae (and Cav-3) under hyperglycemic conditions. Given that PKCβ₁ activation induced by HG requires caveolae in primary mesangial cells (32), we determined whether caveolae are crucial in HG-induced PKCβ₂ activation in isolated cardiomyocytes from nondiabetic rats. As shown in Fig. 3A, phosphorylation of PKCβ₂ induced by HG was prevented by either the selective PKCβ₂ inhibitor CGP-53353 (1 μmol/L; Sigma-Aldrich [IC₅₀ values are 0.41 and 3.8 μmol/L, respectively, for PKCβ₂ and PKCβ₁]) or methyl-β-cyclodextrin (CD) (50 μmol/L), a disrupter of cholesterol-rich caveolae (33). To determine whether Cav-3 is required for PKCβ₂ activation, we subjected H9C2 cells treated with rat-specific Cav3 small interfering (si)RNA to both LG and HG conditions. siRNA-mediated reduction of Cav-3 expression by ∼60% (Fig. 3B) prevented augmented phosphorylation of PKCβ₂ in HG conditions. No effects upon PKCβ₂ phosphorylation were observed in cells exposed to LG (Fig. 3B). We also determined whether excessive PKCβ₂ activation induced by HG is associated with reduced Cav-3 expression. Selective inhibition of PKCβ₂ activation by CGP reversed the reduction of Cav-3 expression in primary cardiomyocytes exposed to HG (Fig. 3C). Similarly, in H9C2 cells, knockdown of PKCβ₂ by siRNA reduced PKCβ₂ phosphorylation in cells incubated in LG and HG conditions (Fig. 3D) and attenuated decreased Cav-3 expression in cells exposed to HG, with no impact upon Cav-3 expression in cells exposed to LG (Fig. 3E).

Next, we investigated the impact of PKCβ₂ activation by HG in the downstream signaling molecules Akt and eNOS, both of which are modulated by Cavs (34). Cardiomyocytes incubated in HG exhibited decreased phosphorylation of Akt at Ser⁴⁷³ and eNOS at Ser¹¹⁷⁷, and these decreases were reversed by CGP treatment (Fig. 4A and B). Caveolar disruption by CD further exaggerated HG-mediated reduction of Akt phosphorylation (Fig. 4A) but did not further exacerbate HG-induced reduction of p-eNOS expression (Fig. 4B). However, CGP-mediated restoration of eNOS phosphorylation in HG-treated cardiomyocytes was abolished during concomitant CD treatment (Fig. 4B). For confirmation of the relative effects of PKCβ₂ and Cav-3 upon HG-mediated changes in p-AKT and p-eNOS, H9C2 cells were subject to both PKCβ₂ and Cav-3 knockdown by siRNA. PKCβ₂ knockdown significantly increased the phosphorylation of Akt and eNOS in HG-treated cells—effects that were not observed in LG-treated cells (Fig. 4C). Knockdown of Cav-3 resulted in further reduced expression of both p-Akt and p-eNOS in both LG- and HG-treated cells (Fig. 4D).

To further investigate the role of PKCβ₂ activation in diabetes-induced abnormalities, we treated STZ-induced diabetic rats with the PKCβ inhibitor LY333531 for 4 weeks. PKCβ_2, but not PKCβ₁, isoform was excessively activated in the diabetic heart, as demonstrated by increased membrane translocation of PKCβ₂ but not PKCβ₁—a phenomenon inhibited by LY333531 (Fig. 5A). PKCβ2 inhibitor LY333531 administration suppressed augmented whole-heart Cav-1 expression (Fig. 5B) and prevented whole-heart decreased Cav-3 expression (Fig. 5C). Caveolae fractions were isolated via discontinuous sucrose gradient centrifugation of whole cell lysates. Cav-1 was found predominantly in fractions 8–10, whereas Cav-3 was located within both lipid fractions (4–6) and heavier fractions (8–12). PKCβ₂ was predominantly coexpressed within Cav-3–rich fractions (Fig. 5D). Densitometric analysis of all fractions (1–12) demonstrated that PKCβ2-inhibitor LY333531 significantly reduced augmented Cav-1 and PKCβ₂ expression and suppressed decreased Cav-3 expression in diabetes (Fig. 5E).

At the end of the treatment period, untreated diabetic rats had significantly elevated blood glucose and reduced body weight and heart weight compared with control rats, which were not altered by LY333531 treatment (Table 1). However, the ratio of heart weight to body weight, an indirect index of myocardial hypertrophy, in the untreated diabetic rats was significantly higher than that in the control rats, which was significantly attenuated by LY333531 treatment (Table 1). Further, the LV cardiomyocyte cross-sectional areas as assessed in hematoxylin-eosin–stained cardiac sections in the untreated diabetic rats (478.8 ± 58.9 µm²) were significantly bigger than in the control rats (256.7 ± 37.9 µm²) and significantly attenuated by LY333531 treatment (329.5 ± 47.7 µm²), showing that LY333531 can attenuate cardiomyocyte hypertrophy in diabetes. We further determined rat cardiac function via echocardiography. As shown in Table 1, no significant change in FS, FLVPW, or EF was observed among all experimental groups (although nonsignificant decreased values were recorded in the untreated diabetic group). Such data suggest preserved systolic function at 5 weeks in diabetic rats used in our model. However, diastolic dysfunction was manifested. The HR and the E/A ratio were significantly decreased in diabetic rats as a consequence of significant reduction of E velocity and enhancement of A velocity that was concomitant with significant increase in LV IVRT and reductions in LVVd and SV. Four weeks’ treatment with LY333531 restored the values of E/A, IVRT, LVVd, and SV to levels that were comparable with those in the control rats but without significant effect upon HR.

Diabetes is associated with decreased NO levels and increased O₂⁻ and nitrotyrosine production (14,35), agents of oxidative and nitrative stress. For determination of whether LY333531 conferred cardioprotection in part by reducing oxidative and nitrative stress in diabetes, the levels of NO, O₂⁻, and nitrotyrosine in diabetic heart tissues were assessed. The diabetic condition significantly decreased NO levels (Fig. 6A) and increased O₂⁻ (Fig. 6B) and nitrotyrosine production (Fig. 6C) in cardiac tissues. LY333531 suppressed all of these derangements. Further studies revealed that the diabetes-induced augmented O₂⁻ levels could be blocked by the NOS inhibitor L-NAME (100 μmol/L) (Fig. 6B), suggesting an NOS-dependent mechanism for O₂⁻ accumulation. We next investigated related signaling molecules, including Akt, eNOS, and inducible NOS (iNOS). The diabetic condition did not affect total cardiac Akt and eNOS expression but significantly decreased p-Akt (Ser⁴⁷³) and p-eNOS (Ser¹¹⁷⁷) expression, both of which were reversed by LY333531 (Fig. 6D and E). Consistent with a recent study (36), our results demonstrate that diabetes increased myocardial iNOS (an adverse marker mediating nitrative stress [23]), which was reversed by LY333531 (Fig. 6F).

In the current study, we have demonstrated that hyperglycemia-induced cardiac PKCβ₂ activation requires caveolae. We provide evidence that excessive PKCβ₂ activation is associated with reduced Cav-3 expression, contributing to abnormal Akt/eNOS signaling during hyperglycemia. Inhibition of excessive activation of PKCβ₂ by compound LY333531 improves cardiac diastolic function, possibly via attenuation of caveolar dysfunction and rescuing Akt/eNOS/NO function in the diabetic heart. To our best knowledge, this is the first study examining the relationship between PKCβ₂ and Cav-3 in cardiomyocytes subjected to hyperglycemic conditions.

It is well established that chronic hyperglycemia induces abnormal activation of PKC, which contributes to diabetic cardiovascular complications (37,38). However, the PKC-signaling pathway is complicated by numerous isoforms, each with varying cellular distribution and opposing function at times (39). The PKCβ₂ isoform is most frequently implicated in diabetic cardiovascular complications (5–8). Our current study further confirmed that PKCβ₂, but not PKCβ₁, is excessively activated in the diabetic heart. Although the precise mechanisms by which hyperglycemia induces PKCβ₂ activation in cardiomyocytes are not fully understood, evidence supports the vital role of caveolae (the specialized plasma membrane microdomains modulating signaling transduction pathways of molecules docked within them [16]) in hyperglycemia-induced PKCβ₂ activation. This is well supported by our immunoprecipitation and immunofluorescence studies demonstrating that hyperglycemia increased the association and colocalization of PKCβ₂ and Cav-3. Caveolar disruption by methyl-β-cyclodextrin (33) suppressed hyperglycemia-induced PKCβ₂ activation in isolated cardiomyocytes. Cav-3 knockdown by siRNA prevented augmented PKCβ₂ phosphorylation in H9C2 cells exposed to HG, suggesting that Cav-3 is required specifically for hyperglycemia-induced PKCβ₂ activation in cardiomyocytes.

Cav-3 is the predominant cardiomyocyte Cav isoform essential for caveolar function. In the current study, we demonstrated that cardiomyocyte Cav-3 expression increased 6–12 h after HG exposure but reduced to basal levels within 24 h and progressively further reduced to lower than basal levels after 36–48 h of HG exposure. The initial increase in Cav-3 expression after HG exposure observed in our study is an acute response to the calorie surplus similar to that reported by other researchers (40). The significant reduction of Cav-3 expression in cardiomyocytes after prolonged HG exposure is consistent with our results from the intact rats, which showed that Cav-3 expression was decreased in isolated cardiomyocytes from rats of 8- and 5-week duration of diabetes. Loss of Cav-3 expression results in cardiomyopathy (41), and reduction in cardiac Cav3 protein expression is highly correlated with reduction in LV FS in mice with constitutive overexpression of A1-adenosine receptor–induced cardiac dilatation and dysfunction and with makers of heart failure phenotype in humans (42). Our study results suggest that excess PKCβ₂ activation contributes to attenuated Cav-3 expression in the diabetic heart, as inhibition of PKCβ₂ activation by CGP53353 or siRNA-mediated PKCβ₂ expression knockdown prevented the decline of Cav-3 expression in cells exposed to HG. LY333531 treatment ameliorated diabetic heart caveolae dysfunction.

PKCβ₂ activation likely exerts adverse effects in the diabetic heart via alteration of the Akt/eNOS signaling pathway, which is modulated by Cavs (34). Although Cav-1 negatively regulates eNOS in cardiovascular tissues (43,44), the colocalization of Cav-3 and eNOS may facilitate eNOS activation by both cell surface receptors and cellular surface NO release for intercellular signaling in cardiomyocytes (17). This is supported by our findings that disruption of caveolae function by methyl-β-cyclodextrin or Cav-3 siRNA prevented Akt phosphorylation and suppressed eNOS phosphorylation in cardiomyocytes. Additionally, hyperglycemia decreased phosphorylated Akt and eNOS both in isolated cardiomyocytes and in diabetic heart tissues, leading to decreased myocardial NO levels, reduced Cav-3 levels, and increased Cav-1 cardiac levels. Inhibition of PKCβ₂ activation suppressed or abrogated these alterations. Therefore, inhibition of PKCβ₂ activation may rescue proper Akt/eNOS/NO signaling in the diabetic heart via Cav regulation.

Previous studies have demonstrated involvement of increased iNOS expression with cardiovascular abnormalities in STZ-induced diabetic rats (23,45). Our study confirms increased iNOS cardiac content in diabetic rats, an adverse mediator of nitrative stress (23), and increased nitrotyrosine was also demonstrated in diabetic heart tissue. The NOS inhibitor L-NAME blocked diabetes-induced augmentation of O₂⁻ levels, indicating eNOS uncoupling, which is in line with a recent study in STZ-induced diabetic mice (46). Furthermore, treatment of diabetic rats with LY333531 inhibited cardiac iNOS expression and reduced both nitrotyrosine and O₂⁻ production. Diabetes is associated with decreased NO levels and increased O₂⁻ and nitrotyrosine production (14,35), which are implicated with oxidative/nitrative stress and eNOS uncoupling. Our study provides direct evidence showing that inhibition of PKCβ₂ activation can mitigate oxidative/nitrative stress and eNOS uncoupling.

Initial LV diastolic dysfunction, reduced contractility, and prolonged diastole are the hallmarks of diabetic cardiomyopathy (47,48). In the current study, diabetes significantly reduced E/A ratio, which was concomitant with significantly increased IVRT and decreased LVVd and SV but did not alter FS, FLVPW, or EF. Such data indicate that myocardial diastolic (but not systolic) dysfunction occurs in 5-week STZ-induced diabetes rats, which can be ameliorated by LY333531. Our findings are in general agreement with the findings of Mihm et al. (26), who conducted a series study in similar STZ diabetic rats and showed that diastolic dysfunction occurred early during the course of the disease, which progressed to LV dilation reflected as increases in LVIDd and LVIDs with concomitant increase in LV luminal area and systolic dysfunction reflected as reduction in LV FS 35 days after STZ injection and onward. It should be noted that the E/A ratio derived from the conventional Doppler echocardiography as used in our current study is not a load-independent parameter and may have inherent limitations, e.g., the peak E wave velocity can be highly dependent upon HR (26), while the HR in the diabetic rats was lower than that in the control group (Table 1). The newly developed Doppler tissue echocardiography (DTE) can acquire myocardial wall and mitral annular velocity online, and the early diastolic annular velocity measured using DTE has been reported to be a preload independent index for evaluating LV diastolic function. A combination of DTE-derived E′/A′ ratio and the mitral inflow patterns (E/A ratio) obtained by conventional Doppler echocardiography in future studies should help to provide better estimations of diastolic dysfunction. However, the significant reduction of E/A ratio as a consequence of significant reduction of E velocity and significant enhancement of A velocity in combination with concomitant increase in IVRT and reductions in LVVD and SV in the diabetic group could jointly suggest diastolic dysfunction in the current study. LY333531 treatment in diabetic rats did not affect HR but corrected all of the above changes, suggesting that LY333531 treatment prevented the development of diastolic dysfunction. Future functional study with cardiomyocytes isolated from rodents in various stages of diabetes will help establish the relative contribution of the slowing in cardiomyocyte relaxation time and LV stiffness (due to fibrosis) to the development of diastolic dysfunction.

In summary, our study demonstrates that hyperglycemia-induced PKCβ₂ activation is associated with caveolar dysfunction and, consequently, deranged Akt/eNOS signaling (Fig. 7). Inhibition of PKCβ₂ activation attenuated cardiac diastolic dysfunction by restoring Cav-3 expression and subsequently rescuing Akt/eNOS/NO signaling. PKCβ₂ blockade may therefore represent a novel therapeutic avenue in the treatment of diabetic cardiomyopathy.

This study was supported by a grant from the National Natural Science Foundation of China (NSFC 81270899) and by General Research Fund grants (784011M and 782910M) from the Research Grants Council of Hong Kong.

No potential conflicts of interest relevant to this article were reported.

S.L. performed the study and wrote the manuscript. H.L., J.X., Y.L., and X.G. performed the study. J.W., K.F.J.N., W.B.L., X.-l.M., and B.R. contributed to data analysis and interpretation. M.G.I. reviewed and approved the research protocol. Z.X. reviewed and approved the research protocol and wrote the manuscript. Z.X. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.