Hyperglycemia Abrogates Ischemic Postconditioning Cardioprotection by Impairing AdipoR1/Caveolin-3/STAT3 Signaling in Diabetic Rats

Myocardial infarction is a major perioperative complication in patients with diabetes (1). Reperfusion therapies restore coronary flow but may cause lethal tissue injury, called “reperfusion injury.” Ischemic postconditioning (IPo), the phenomenon that brief repetitive episodes of ischemia and reperfusion (IR) at the immediate onset of reperfusion, can protect the hearts against myocardial IR injury (MIRI) (2), which needs to activate signal transducer and activator of transcription 3 (STAT3) (3). However, how IPo activates STAT3 remains unclear. In addition, hearts from subjects with diabetes are less or not sensitive to IPo, and the mechanism is unknown (4). The reduction of myocardial STAT3 activation in diabetes and the subsequent reduction of myocardial nitric oxide (NO) bioavailability (5) may be primarily responsible for the loss of myocardial sensitivity to IPo (4).

Adiponectin (APN), an adipocyte-derived plasma protein with antidiabetic properties, protects against MIRI by reducing myocardial oxidative/nitrative stress and activating endothelial nitric oxide synthase (eNOS) and increasing NO bioavailability (6), which requires the interaction of caveolin-3 (Cav3), the structural protein for caveolae, which are flask-shaped plasma membrane invaginations, and APN receptor 1 (AdipoR1), the predominant APN receptor expressed in cardiomyocytes (7). In cultured adult mouse cardiac fibroblasts, APN activates STAT3 (8). However, whether APN can activate or facilitate the activation of STAT3 in cardiomyocytes in the context of MIRI in nondiabetic and especially in diabetic conditions, a pathological condition that is accompanied with increased oxidative stress, is unknown. Nevertheless, the properties of APN (i.e., activating STAT3 and enhancing NO bioavailability) suggest that APN may play significant roles in the cardioprotective actions of IPo. Furthermore, the finding that antioxidant treatment enhanced STAT3 activation in the heart of diabetic rats through APN signaling (9) suggests that malfunction of APN might be a key mechanism that rendered the diabetic heart less or not responsive to IPo cardioprotection. In addition, cardioprotective effects of APN were diminished in high-fat diet–induced diabetes (10,11), in which Cav3 is downregulated (12). This, together with our recent finding that Cav3 is disrupted (13) and STAT3 inactivated (9), which deranged eNOS signaling in diabetic myocardium with concomitantly reduced APN (9), indicates that APN signaling (e.g., AdipoR1/Cav3) is impaired in diabetes. Although AdipoR1 (14,15) and STAT3 (16) have both been shown to be important procell survival factors in multiple cells, their potential interplay in affecting postischemic cell survival in general and specifically in affecting the myocardial responsiveness to IPo cardioprotection in diabetes is unknown. We hypothesized that the impairment of cardiac APN signaling is responsible for the inactivation of STAT3 in diabetes that disables the ability of the diabetic heart to respond to IPo and that an intact or adequate AdipoR1/Cav3 interaction is critical for IPo to activate STAT3 via APN.

Male Sprague-Dawley rats (250 ± 8 g, 6–8 weeks) were obtained from the Laboratory Animal Unit (The University of Hong Kong). Male APN knockout (Adipo^−/−) mice (6–8 weeks) with a C57BL/6J background and age-matched wild-type (WT) control mice with the same genetic background were generated as previously described (17). All of the experiments were conducted in adherence with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee.

After anesthesia, animals were randomized to receive sham operation, myocardial IR, or IPo. Myocardial IR was induced by temporarily exteriorizing the heart via a left thoracic incision, and the left anterior descending coronary artery was occluded for 30 min, followed by reperfusion for 2 h (to assess postischemic mRNA and protein expressions of cardiac APN and other signaling molecules) or 24 h (to assess postischemic cardiac function, myocardial injury, and infarct size) (6). IPo was produced by three cycles of 10 s of reocclusion and 10 s of reperfusion immediately after ischemia (18).

All assays were performed using tissue from IR area or an area at risk identified with Evans blue negative staining. Myocardial infarct size was determined by the Evans blue/2,3,5-triphenyl-2H-tetrazolium chloride (TTC) staining, as previously described (9), and the operators were blinded to the information of study design and intervention. Global cardiac functions were monitored by using a pressure-volume conductance catheter and analyzed using LabChart 8 software (ADInstruments, Colorado Springs, CO), as we described previously (19).

Myocardial apoptosis was determined by TUNEL staining using an In Situ Cell Death Detection Kit (Roche Diagnostics GmbH, Mannheim, Germany), as previously described (20). The sections were observed in the light microscope by an investigator who was initially blinded to treatment groups. Five randomly selected fields of each slide were analyzed, and the apoptotic index was calculated as a percentage of apoptotic nuclei to total nuclei.

Heart tissues were fixed with 2.5% glutaraldehyde in 0.1 mol/L phosphate buffer (pH 7.4), followed by 1% OsO₄. After dehydration, thin sections were stained with uranyl acetate and lead citrate for observation under a JEM 1011CX electron microscope. Images were acquired digitally.

Type 1 diabetes was induced via signal tail-vein injection of streptozocin (STZ) (65 mg/kg; Sigma-Aldrich, St. Louis, MO), as previously described (21). One week after STZ injection, rats exhibiting hyperglycemia (blood glucose ≥16.7 mmol/L) were considered diabetic and subjected to outlined experiments. Recombinant APN adenovirus (1*10⁹ plaque-forming units) was used to overexpress APN in vivo, and luciferase was used as the control, which were injected by tail vein into rats 1 week before inducing IR, as previously described (19).

At the end of experiments, heart tissues were immediately collected for cardiac mitochondria isolation according to the manufacturer’s protocol per the mitochondria extraction kit (Thermo Fisher Scientific, Chicago, IL). The resulting mitochondrial pellets were resuspended in Tris-hydroxymethyl aminomethane solution (pH 8.0) and kept at −80°C for subsequent measurements of mitochondrial respiratory chain complex activities (complex I-V), as previously described (22).

Cardiac tissue (area at risk) was rinsed and homogenized. Plasma and heart tissue free 15-F_2t-isoprostane (15-F_2t-IsoP) was measured using immunoassay kits (Cayman Chemical, Ann Arbor, MI), as previously described (23). Concentrations of nitrites (NO₂⁻) and nitrates (NO₃⁻) were determined by the Griess reaction, as previously described (13). Myocardial nitrotyrosine levels were determined using the Nitrotyrosine Assay Kit (Millipore) according to the manufacturer’s protocol. Myocardial ATP content was measured using an ATP ELISA kit (Cloud-Clone, Houston, TX). Superoxide generation in cultured cardiomyocytes was estimated by dihydroethidium (DHE) staining, as previously described (24).

Cardiac DNA was extracted from frozen heart tissues with TaKaRa Genomic DNA Extraction Kit (TaKaRa Bio, Inc., Shiga, Japan). Quantitative real-time PCR was performed using SYBR Green QPCR system (TaKaRa Bio, Inc.) with specific primers, mouse nuclear DNA (nDNA) contamination nDNA forward: ATGGAAAGCCTGCCATCATG, reverse: TCCTTGTTGTTCAGCATCAC; mitochondrial DNA (mtDNA) forward: CCTATCACCCTTGCCATCAT, reverse: GAGGCTGTTGCTTGTGTGAC. The PCR reactions were performed using an Applied Biosystems Prism 7000 Sequence Detection System. Relative quantification of the copy number of mtDNA was analyzed using nDNA as the standard. Total RNA was extracted using TRIzol (Invitrogen Life Technologies, Carlsbad, CA), and quantitative real-time PCR was performed with a SYBR Green PCT Master Mix (TaKaRa Bio, Inc.) on a Applied Biosystems Prism 7000 Sequence Detection System. Gene-specific primers were as follows:

• mouse Tfam forward: 5′-TCAGGAGCAGCAGGCACTACA-3′, reverse: 5′-CTGAGCTCC GAGTCCTTGAACAC-3′;

• mouse Nrf1 forward: 5′-GATGCTTCAGAACTGCCAACCA-3′, reverse: 5′GGTCATTTCACCGCCCTGTAAC-3′;

• mouse Ppargc1a forward: 5′-CACTGACAGATGGAGCCGTGA-3′, reverse: 5′-TGTTGGCTGGTGCCAGTAAGAG-3′;

• mouse β-actin forward: 5′-CATCCGTAAAGACCTCTATGCCAAC-3′, reverse: 5′-ATGGAGCCACCGATCCACA-3′.

Calcium-tolerant cardiomyocytes were prepared from rat/mouse ventricles via a modified method, as previously described (13). Mice cardiomyocytes were treated with stattic (a specific STAT3 inhibitor; 100 µmol/L, 1 h before hypoxia/reoxygenation [HR]), methyl-β-cyclodextrin (CD, a disrupter of cholesterol-rich caveolae; 10 mmol/L, 30 min before HR), or recombinant globular APN (gAd) (2 µg/mL, for 24 h), before being subjected to HR and hypoxic postconditioning (HPo). Rat cardiomyocytes were incubated in normal glucose (5.5 mmol/L) or high glucose (HG) (25 mmol/L) for 18 h or 38 h. Some of the subgroups were subjected to HR and HPo. HR was achieved by hypoxia for 45 min, followed by 2 h of reoxygenation (25). HPo was achieved by three cycles of 5 min of hypoxia and 5 min of reoxygenation. Hypoxia conditions were obtained by equilibrating a humidified Plexiglas chamber containing myocytes with 95% N₂ and 5% CO₂ and confirmed by measuring the chamber O₂ concentration falling to 0.1%. Reoxygenation was achieved by exposing cells to room air (9).

At the end of treatments, cells were fixed for immunofluorescence staining, as described below, or collected and snap frozen in liquid N₂ for future analysis. Lactate dehydrogenase (LDH) release in culture medium was detected via a commercial LDH kit (Roche, Mannheim, Germany) (13). After the experiments were completed, cultured cardiomyocytes were homogenized in lysis buffer. The protein concentration was determined via a Lowry assay kit (Bio-Rad, Hercules, CA). Concentrations of APN were determined by a commercially available APN EIA kit (Antibody and Immunoassay Services, Hong Kong) (22). APN levels were expressed as nanogram per microgram of protein.

Isolated cardiomyocytes or heart tissue were homogenized in lysis buffer. A total of 500 mg extracts were subjected to immunoprecipitation with 2 mg Cav3 primary antibody in the presence of 20 mL protein A/G PLUS-agarose. After extensive PBS washes, the immunoprecipitates were denatured with SDS loading buffer and analyzed for AdipoR1 or APN receptor 2 (AdipoR2) expression by Western blot, as described below.

Isolated cardiomyocytes were incubated in Medium 199 and subjected to various treatments. Thereafter, cells were fixed and blocked with 10% goat serum and further incubated with a mixture of mouse against rat Cav3 antibody and rabbit against rat AdipoR1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Then, the cells were incubated with a mixture of Alexa Fluor 488 goat anti-mouse IgG and Alexa Fluor 594 goat anti-rabbit IgG (Invitrogen, Carlsbad, CA) and were imaged with a confocal laser scanning microscopic, with mounting medium containing DAPI (Vector Laboratories, Burlingame, CA) (13).

Embryonic rat cardiac H9C2 cells were maintained in DMEM containing 10% FBS in a humidified atmosphere (5% CO₂) at 37°C. Commercial small interfering RNAs (siRNAs) (Santa Cruz Biotechnology) were used to respectively knockdown the gene expression of APN, AdipoR1, AdipoR2, T-cadherin, and STAT3, according to the manufacturer’s protocol. After transfection with control or specific siRNA, cells were incubated in DMEM for 36 h. Some of the subgroups were treated with gAd (2 µg/mL) for 24 h and snap frozen in liquid N₂.

Caveolae were isolated by discontinuous sucrose gradient centrifugation, as previously described (13). Each heart sample gradient was separated into 12 fractions. Fractions 4–6 were considered the lipid raft fractions (buoyant membrane), and fractions 8–12 were considered the heavier fractions (nonbuoyant membrane).

Equal protein amounts from isolated cardiomyocytes, H9C2 cells, rat heart, and isolated mitochondria or caveolae fractions were resolved by 7.5–12.5% SDS-PAGE and transferred to polyvinylidene fluoride membrane for immunoblot analysis, as previously described (13).

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 the Tukey test for multiple comparisons. Statistical analysis was performed by GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA). P values of <0.05 were considered statistically significant.

APN deficiency exacerbated MIRI (6,26). To determine the role of APN in IPo-mediated cardioprotection, WT and Adipo^−/− mice were subjected to myocardial IR with or without IPo. After 24 h of reperfusion, Adipo^−/− mice exhibited more severe MIRI, manifested as larger myocardial infarct size (Fig. 1A-C), higher plasma creatine kinase (CK)-MB (Fig. 1D), and increased cardiomyocyte apoptosis (Fig. 1E and F), concomitant with reductions in the maximal slope of systolic pressure (dP/dt_Max), cardiac diastolic decrement (dP/dt_Min) (Fig. 1H and I), stroke work, cardiac output, ejection fraction, and elevations in cardiac end-diastolic pressure (Fig. 1G) and Tau (Supplementary Table 1). IPo significantly attenuated all these changes in WT but not in Adipo^−/− mice.

Studies showed that APN accumulated in the injured area of the heart after IR (26) and that mitochondrial (mito)STAT3 activation is critical for IPo to confer cardioprotection (27). To determine the role of APN in IPo-mediated cardioprotection and mitoSTAT3 activation, we determined the levels of plasma and cardiac APN as well as mitoSTAT3 in WT and Adipo^−/− mice subjected to IR and IPo. Postischemic plasma and cardiac APN were significantly reduced and IPo prevented IR-induced reduction of cardiac APN protein expression at the early stage of reperfusion (i.e., 1 and 2 h of reperfusion) (Supplementary Fig. 1A) and increased the cardiac mitoSTAT3 phosphorylation at site Ser727 in WT but not in Adipo^−/− mice (Fig. 2A–C), suggesting that IPo required cardiac APN to induce/activate STAT3, whereas IPo had no effect on plasma APN (Fig. 2A).

Further, we determined the effect of IPo on mitochondrial function. Compared with WT mice, Adipo^−/− mice exhibited more severely impaired postischemic myocardial mitochondrial function, evidenced as lower complex I/II and III/IV/V activities (Fig. 2D) and reduced cardiac ATP production (Fig. 2E), as well as more severe mitochondrial damage manifested as significant morphological defects (mitochondrial swelling and dissolving) and a reduction of the mtDNA-to-nDNA ratio (Fig. 2F and G) concomitant with decreases of mRNA expression of Tfam, Nrf1, and Ppargc1a (Supplementary Fig. 1B–D). IPo significantly increased mitochondrial complex I/IV/V (but not complex II and III) activities, increased myocardial ATP production, attenuated mitochondrial morphological changes, increased the mtDNA-to- nDNA ratio (Fig. 2D–G), and upregulated mRNA expression of Tfam, Nrf1, and Ppargc1a in WT but not in Adipo^−/− mice (Supplementary Fig. 1B–D).

Postischemic cardiac NO was significantly reduced, which was associated with decreased eNOS phosphorylation at site Ser1177 and increased nitrotyrosine formation in WT and Adipo^−/− mice, and IPo attenuated all of these changes in WT but not in Adipo^−/− mice (Supplementary Fig. 2A–C). Similarly, IR-induced myocardial oxidative stress was more pronounced in Adipo^−/− mice, evidenced as elevated cardiac and plasma levels of 15-F_2t-IsoP (P < 0.05, WT-IR vs. WT-Sham; Adipo^−/−-IR vs. WT-IR), a specific index of oxidative stress induced by reactive oxygen species (ROS), and IPo significantly attenuated all these changes in WT but not in Adipo^−/− mice (Supplementary Fig. 2D and E).

To confirm the role of APN in IPo cardioprotection and to see whether this exactly happens in the cardiomyocytes, cardiomyocytes from WT and Adipo^−/− mice were isolated and subjected to HR and HPo in the presence or absence of gAd and/or the STAT3-specific inhibitor stattic. In cardiomyocytes isolated from WT mice, posthypoxic cellular injury was significantly increased, manifested as increased LDH and cleaved caspase 3 expression, which was concomitant with reduced APN content, and all of these changes were significantly attenuated by HPo (Fig. 3A–C and Supplementary Fig. 3A). In cardiomyocytes isolated from Adipo^−/− mice, HR significantly increased cell death (elevated release of LDH, Fig. 3D), increased apoptosis (enhanced cleaved caspase 3 expression, Fig. 3E), impaired mitochondrial function (increased release of cytochrome C, Fig. 3F), and elevated ROS production (increased the number of DHE-positive cells, Fig. 3G) that was associated with reduced mitoSTAT3 activation (Fig. 3H). Similar to our in vivo data, cellular protection of HPo was diminished in cardiomyocytes isolated from Adipo^−/− mice but was restored by APN administration in a dose-dependent manner manifested as progressively reduced posthypoxic LDH release and cytochrome C expression (Supplementary Fig. 3C and D). However, stattic abolished APN administration–mediated restoration of HPo cellular protection. These results from isolated cardiomyocytes were confirmed in H9C2 cells with APN or STAT3 gene knockdown by using respective siRNAs, which showed that APN supplementation restored HPo cellular protection in H9C2 with APN gene knockdown. However, this beneficial effect of APN supplementation was abolished by STAT3 gene knockdown (Supplementary Fig. 3E and F).

APN confers cardioprotective effect mainly through its three receptors: AdipoR1, AdipoR2, and T-cadherin (28). To determine which receptor is involved in the APN-mediated STAT3 activation in the context of IPo, cultured H9C2 cells were treated with specific AdipoR1, AdipoR2, and T-cadherin siRNA, respectively, and then subjected to HR and HPo. AdipoR1 siRNA, but not AdipoR2 or T-cadherin siRNA, abolished HPo cellular protection manifested as increased LDH release, elevated cleaved caspase 3 expression, and decreased Bcl2-to-Bax ratio (Fig. 4A–C). However, supplementation of APN (by gAd) cannot restore HPo cellular protective effects in H9C2 cells treated with AdipoR1 gene knockdown (Supplementary Fig. 4E), suggesting that APN-mediated HPo cellular protection is dependent on AdipoR1. AdipoR1 siRNA, but not AdipoR2 or T-cadherin siRNA, diminished HPo-induced upregulation of eNOS and phosphorylated STAT3 protein expression (Fig. 4D–F).

To determine whether Cav3 contributes to APN-mediated IPo cardioprotection, the expression and distribution of Cav3 and AdipoR1 were examined. Postischemic myocardial Cav3 and AdipoR1 were significantly decreased (Fig. 5A and B) concomitant with reduced Cav3 and AdipoR1 accumulation in the buoyant fraction (Fig. 5C). Although cardiac Cav3 protein expression was significantly lower in Adipo^−/− mice than that in WT mice, IPo significantly increased Cav3 protein expression both in WT and Adipo^−/− mice. However, IPo significantly increased AdipoR1 protein expression, Cav3, and AdipoR1 buoyant fraction accumulation (Fig. 5A-C), with concomitantly increased colocalization of Cav3 and AdipoR1 in WT but not in Adipo^−/− mice (Fig. 5D). Furthermore, cardiomyocytes were isolated from WT and Adipo^−/− and subjected to HR and HPo, and HPo significantly upregulated Cav3 and AdipoR1 colocalization (Fig. 5E) and attenuated HR-induced cellular injury and reduction of mitoSTAT3 phosphorylation in cardiomyocytes from WT but not Adipo^−/− mice. All of these cellular protective effects of HPo in WT cardiomyocytes were abolished by CD (a disrupter of cholesterol-rich caveolae) (Fig. 5F and G). Whereas in cardiomyocytes isolated from Adipo^−/− mice, APN (gAd) supplementation restored HPo cellular protection and HPo-induced mitoSTAT3 activation, but these beneficial effects were also abolished by CD (Fig. 5H and I).

Postischemic myocardial injury was significantly increased, evidenced as increased infarct size and plasma CK-MB release in both 4- and 8-week diabetic rats (Supplementary Fig. 5A–C), and these were associated with increased cardiac 15-F_2t-IsoP, reduced cardiac NO production, and elevated cardiac nitrotyrosine formation (Supplementary Fig. 5D–F). IPo significantly attenuated/prevented all of these changes in nondiabetic but not in 4- or 8-week diabetic rats.

In nondiabetic control rats, IPo significantly reduced postischemic myocardial injury manifested as a reduction of infarct size (Fig. 6A and B). APN supplementation further enhanced IPo cardioprotective effects, whereas these protective effects of APN were abolished by STAT3 inhibition (by AG490, a Jak inhibitor that inhibits STAT3) (Fig. 6A and B).

In diabetic rats, blood glucose was increased (28.9 ± 0.4 mmol/L in 4-week diabetic rats vs. 6.2 ± 0.1 in age-matched control rats; 29.1 ± 0.5 mmol/L in 8-week diabetic rats vs. 5.9 ± 0.2 in age-matched control rats), which was associated with reduced plasma and cardiac APN (Supplementary Fig. 6A and B). APN supplementation alone, at the dose used, did not reduce postischemic infarct size in 4- or 8-week diabetic rats (Fig. 6A and B), but reduced postischemic CK-MB release in 4-week but not in 8-week diabetic rats (Fig. 6C). By contrast, APN supplementation in combination with IPo significantly reduced postischemic infarct size and plasma CK-MB release and improved cardiac functional recovery in 4-week but not in 8-week diabetic rats (Fig. 6A–D and Supplementary Tables 2 and 3). Plasma APN was significantly elevated after APN injection, but no significant changes in cardiac APN protein expression were observed in 4- or 8-week diabetic rats before or after IR and IPo (Supplementary Fig. 6C and D). IPo had no significant effect on postischemic cardiac APN protein expression in diabetic rats irrespective of APN supplementation (Supplementary Fig. 6D), which was in keeping with the result gained in primarily cultured cardiomyocytes in which IPo did not affect HR-induced reduction of APN in cardiomyocytes incubated with HG (Supplementary Fig. 3B). Further, AG490 significantly reduced mitoSTAT3 phosphorylation and canceled the cardioprotection conferred by the combined use of APN and IPo in 4-week diabetic rats (Fig. 6E). Also, mitochondrial complex I/II and III/IV/V activities were significantly increased after APN supplementation and were further enhanced by IPo, but these beneficial effects of IPo were abolished by AG490 (Fig. 6F). APN supplementation alone or combined with IPo significantly attenuated postischemic mitochondrial damage (reduction of cytochrome C release) and reduced cardiac free 15-F_2t-IsoP formation, whereas these beneficial effects of APN were abolished by AG490 (Fig. 6G and H).

As shown in Fig. 7A and B, AdipoR1, but not AdipoR2, protein was progressively reduced from 1 week to 8 weeks of diabetic induction (Fig. 7A and B). The colocalization of AdipoR1 and Cav3 was significantly and progressively reduced after diabetic induction (Fig. 7C). In cardiomyocytes isolated from nondiabetic rats exposed to normal glucose or HG for different durations, the colocalization of AdipoR1 and Cav3 was increased after 18 h of HG exposure but was significantly decreased after 38 h of HG exposure (Fig. 7D). These results were further confirmed by coimmunoprecipitation showing that the colocalization of AdipoR1 but not AdipoR2 with Cav3 was significantly reduced after 38 h of HG treatment (Fig. 7E and F). Similarly, AdipoR1 and Cav3 colocalization was significantly reduced in both 4- and 8-week diabetic rats but was reversed by IPo in 4-week but not in 8-week diabetic rats (Fig. 7G).

In the current study, we demonstrated that APN is required for IPo to activate mitoSTAT3 to confer cardioprotection and that APN mediates mitoSTAT3 activation after IPo via AdipoR1/Cav3 signaling. Reduction in APN and impairment of AdipoR1/Cav3 signaling are responsible for the loss of IPo cardioprotection in diabetes. In the early stage of diabetes, loss of IPo cardioprotection is mainly due to the reduced APN, whereas with the progression of diabetes, it is mainly due to the impairment of AdipoR1/Cav3 signaling. Supplementation of APN by preserving mitoSTAT3 activation can restore heart sensitivity to IPo in the early stage of diabetes, where AdipoR1 and Cav3 are still functionally interactive, but not in the late stage of diabetes when AdipoR1/Cav3 signaling is severely impaired.

APN is an abundant circulating adipocytokine secreted from adipose tissue (29) and cardiomyocytes (30). Our results suggest that IPo confers cardioprotection through increasing endogenous APN, which is generated by and may be subsequently secreted from cardiomyocytes. Conversely, Shibata et al. (26) reported that 24 h after myocardial IR, APN was upregulated and accumulated in the damaged myocardium, which was concomitant with the dramatic decrease of postischemic circulating APN. The inconsistencies of these results may be due to the different time points of sample collection but provide further information that there is a delay for APN to be shuttled from other compartments to the injured heart and that cardioprotection of IPo seen in our present study may be mainly through acutely/rapidly upregulating cardiomyocyte-driven APN. This was confirmed in the in vitro study in isolated cardiomyocytes (Fig. 3), which showed that IPo increased APN protein expression and reduced cellular injury in cardiomyocytes isolated from WT but not from Adipo^−/− mice. The loss of IPo protection in Adipo^−/− cardiomyocytes was reversed by APN supplementation, which further indicates the essential role of cardiac APN in IPo cardioprotection.

IPo exerts its cardioprotection via three major intracellular pathways, the NO synthase/protein kinase G pathway (31), the reperfusion injury salvage kinase pathway (i.e., PI3K/Akt) (32), and the survival activating factor enhancement pathway (Jak/STAT3, which is essential for IPo cardioprotection [33]). These pathways converge at the mitochondria as an integration point that is decisive for survival of cardiomyocytes (34). However, how IPo improves postischemic mitochondrial function subsequent to acute STAT3 activation is unknown. By using loss-of-function approach in Adipo^−/− mice and gain-of-function approach in isolated cardiomyocytes, we demonstrated that IPo, by increasing cardiac APN, activated STAT3, which subsequently shuttled into mitochondria, reversed the activities of mitochondrial complex I/IV/V, leading to increased ATP content and an increased mtDNA-to-nDNA ratio (Fig. 2), and reduced myocardial oxidative stress and attenuated postischemic myocardial injury. We previously showed that APN attenuated postischemic myocardial dysfunction and apoptosis by decreasing NADPH oxidase expression and blocking peroxynitrite formation (6,35). In addition, we recently showed that administration of APN increased heme oxygenase-1 induction by concomitantly increasing Brg1 and Nrf2, which reduced myocardial oxidative stress and cardiac dysfunction in STZ-induced diabetes (19). Similarly, in the current study, we found that IPo attenuated postischemic myocardial oxidative stress in WT but not in Adipo^−/− mice (Supplementary Fig. 2), suggesting that APN is required for IPo to confer cardioprotection through decreasing myocardial oxidative stress.

NO represents one of the most important defense mechanisms against MIRI and is also one of the major mediators of IPo cardioprotection (36). In the current study, IPo conferred cardioprotective effects through APN that was associated with reduced postischemic nitrotyrosine formation and increased cardiac NO production in WT but not in Adipo^−/− mice. All of these suggest that IPo confers cardioprotection by reducing myocardial oxidative stress and increasing NO bioavailability through APN (Supplementary Fig. 2A and B). Thus, it is reasonable to postulate that inhibition of NO may abrogate APN-mediated IPo cardioprotection. This hypothesis deserves to be further investigated. Further, in the current study, we found that the above-mentioned APN-mediated cardioprotective effects in IPo were abolished by STAT3 inhibition, which was associated with increased ROS formation (Fig. 3). This indicates that APN-mediated mitoSTAT3 activation plays essential roles in IPo cardioprotection, especially in IPo-induced reduction of postischemic myocardial oxidative stress.

We previously showed that APN-mediated protection against MIRI depends on the activation of AdipoR1, the predominate form of AdipoRs in the heart (37). Similarly, in the current study, AdipoR1 siRNA, but not AdipoR2 or T-cadherin siRNA, abolished HPo-induced STAT3 activation, eNOS expression, and cellular protection, indicating that cardiac APN-induced STAT3 activation during IPo is AdipoR1 dependent. We further found that IPo significantly increased myocardial Cav3 protein expression and increased the colocalization of AdipoR1 and Cav3. The critical role of effective AdipoR1/Cav3 interaction in IPo-mediated cardioprotection was confirmed by our findings that the application of a Cav3 disrupter, CD, not only canceled HPo-induced cellular protection but also reduced HPo-induced STAT3 activation in cardiomyocytes isolated from WT mice despite not affecting the cardiomyocyte content of APN and AdipoR1 (data not shown). These provide additional evidence that Cav3 plays an essential role in APN transmembrane signaling and in APN anti-IR actions, as we reported (37). Our finding that APN supplementation mediated restoration of HPo-induced STAT3 activation and HPo cardioprotection in Adipo^−/− cardiomyocytes was canceled by Cav3 disruptor indicates that APN-induced STAT3 activation in IPo cardioprotection is mediated by or needs the participation Cav3.

Increased oxidative stress in the heart of diabetic animals is a major mechanism that renders the diabetic heart less or not responsive to cardioprotective interventions that are otherwise effective in nondiabetic hearts (4,38), but the underlying mechanism is unclear. Plasma and cardiac APN levels are reduced in patients with diabetes (39) and in STZ-induced diabetic rats (9,40). Moreover, oxidative stress can downregulate APN level (9,41). Thus, reduced APN, in part as a consequence of increased oxidative stress in diabetes, should be responsible for the loss of IPo cardioprotection in diabetes. To this end, we used an adenovirus to overexpress APN in diabetic rats and subjected them to IR and IPo. Interestingly, supplementation of APN restored IPo cardioprotection by improving mitochondrial function and ameliorating oxidative stress subsequent to mitoSTAT3 activation (Fig. 8), but these effects of APN were only seen in 4-week but not in 8-week diabetes. This promoted us postulate that the different responses to APN at 4- and 8-week diabetes may be due to the impaired APN signaling in diabetes. After diabetes induction, AdipoR1 expression decreased progressively with the significant reduction at 4 and 8 weeks, whereas AdipoR2 expression did not change. More importantly, the colocalization of cardiac AdipoR1 and Cav3 reduced progressively in diabetes, with the most significant reduction seen at 8-week diabetes. Further, our in vitro study results showed that AdipoR1 and Cav3 colocalization increased after HG exposure for 18 h but significantly decreased when cardiomyocytes were exposed to HG for a longer period (38 h), whereas the in vivo study showed that IPo significantly increased AdipoR1 and Cav3 colocalization in 4-week but not in 8-week diabetes (Fig. 7). These findings collectively indicate that in the early stage of diabetes, the loss of IPo was mainly due to the reduced APN, whereas at the late stage of the disease, it was mainly due to the impaired APN signaling (AdipoR1/Cav3), which consequently diminished IPo-induced STAT3 activation.

Our results may provide an explanation for the discrepancy showing that although clinical and animal studies all indicated that APN is cardioprotective, those with type 1 diabetes are more susceptible to myocardial IR and not sensitive to IPo irrespective of increased or decreased APN seen in patients and animals with type 1 diabetes (42,43). Our results showed that loss of the APN cardioprotective effect seen in the late stage of diabetes is largely due to the malfunction of APN, especially the impaired AdipoR1/Cav3 signaling, but not merely the reduction of APN content. That the protective effects of IPo may be abolished by diabetes mainly through APN-independent mechanisms is unlikely, because we have clearly demonstrated that APN is required for IPo cardioprotection by using Adipo^−/− mice. It should be noted, however, that reduction of APN is not the main reason why diabetic hearts lose responsiveness to IPo. This explains why APN supplementation combined with IPo cannot restore diabetic heart responsiveness to IPo at the late state of the disease when cardiac Cav3/AdipoR1 signaling is impaired. In addition to type 1 diabetes, IPo cardioprotective effects were also reduced in type 2 diabetes (44), and we have shown that APN cardioprotective effects also reduced in type 2 diabetes associated with reduced APN content (10,11). However, whether IPo can increase the interaction of AdipoR1 and Cav3 in type 2 diabetes is not known. If so, then the mechanism identified in the type 1 diabetic model in the current study may also operate in type 2 diabetes. This needs to be studied further.

Of note, previous studies showed that APN exerts its metabolic regulative effects largely through the AMPK-dependent pathway. However, our recent study (6) shows that APN reduced postischemic myocardial injury via its antioxidative effects rather than through metabolic regulation. In addition, exogenous APN supplementation could attenuate myocardial apoptosis in cardiomyocytes subjected to HR by decreasing NADPH oxidase expression and blocking peroxynitrite formation in an AMPK-independent fashion (35). These results strongly suggest that the degree of AMPK involvement in the biological functions of APN is determined by the intracellular environment, particularly AMP concentration. Under pathological conditions (such as myocardial IR) in which intracellular AMP concentrations are elevated (45,46) and AMPK has already been significantly activated, APN may exert its biological actions largely through signaling molecules other than AMPK. In the current study, we found APN activates STAT3 that during IPo and subsequently reduces myocardial oxidative stress mainly through AdipoR1/Cav3. Because we did not determine the change of AMPK in this process, we cannot exclude the possibility that AMPK may also play a role in this APN-induced STAT3 activation in IPo, and thus, further study regarding the role of AMPK in APN-mediated STAT3 activation in IPo is needed, and its resolution will advance the understanding of IPo cardioprotection.

To our knowledge, our study demonstrated for the first time that IPo, by activating APN, activates STAT3 and leads to its mitochondrial shuttling and consequently improves postischemic cardiac mitochondrial function and attenuates myocardial injury and that intact AdipoR1/Cav3 signaling is critical for IPo to activate STAT3 through APN. Diabetes, by reducing APN and disarranging the AdipoR1/Cav3 interaction, diminishes APN-induced STAT3 activation and eventually reduces IPo cardioprotection in the hearts from diabetes (Fig. 8).

Although most experimental animal studies have shown consistent results of effective cardioprotection of IPo, attempts to translate IPo into clinical use are not successful. The reason might be that most patients typically have comorbidities, such as diabetes, aging, dyslipidemia, and hypertension, which have been shown to influence the protective effects of IPo. By contrast, most of the animals used in experimental studies are healthy, juvenile, adult subjects, which cannot mimic the clinical situation. Findings from our current study indicate that effective means that may concomitantly activate APN and repair APN signaling (i.e., AdipoR1/Cav3) in diabetes may represent promising avenues to restore IPo cardioprotection in the treatment of MIRI in diabetes. Recently, APN receptor(s) agonist, which can be given orally, has been successfully generated (47), therefore providing optimism that the application of an APN receptor agonist in combination with IPo could be promising therapies in combating MIRI in diabetes.

Acknowledgments. The authors thank Dr. Yan Chen and Dr. Jiao Peng, Department of Surgery, The University of Hong Kong, Hong Kong, China, for excellent technical assistance.

Funding. Z.X. has received General Research Fund grants 784011, 17124614, and 17123915 from the Research Grants Council of Hong Kong and grant 81270899 from the National Natural Science Foundation of China.

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

Author Contributions. H.L. and Z.X. designed the study. H.L., W.Y., and Z.L. performed experiments and analyzed data. H.L., W.Y., Z.L., A.X., Y.H., X.-l.M., M.G.I., and Z.X. interpreted results of experiments. H.L. drafted the manuscript. M.G.I. and Z.X. edited and revised the manuscript. Z.X. approved the final version of 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.

Prior Presentation. Parts of these data were presented at the American Heart Association Scientific Sessions 2014, Chicago, IL, 15–19 November 2014, and published as an abstract in Circulation 2014;130:A13879.