Exogenous H2S Protects Against Diabetic Cardiomyopathy by Activating Autophagy via the AMPK/mTOR Pathway Open Access

In diabetic patients, a wide range of structural reconfigurations have been observed, such as cardiomyocyte hypertrophy, ventricular dilation, prominent interstitial fibrosis [1, 2], diastolic dysfunction, systolic dysfunction, and left ventricular hypertrophy [3, 4] _ ENREF_4.This disease process is known as diabetic cardiomyopathy (DC), which is a cardiac muscle-specific disease without other vascular pathology_ENREF_6.

Macroautophagy (henceforth, ‘autophagy’) at a low level under physiological conditions exerts its cellular protective and quality control process, which can be activated in response to stress conditions including starvation and oxidative stress [5]. Autophagy is an evolutionarily conserved catabolic process to degrade and recycle long-lived proteins and damaged organelles [6]. The mechanisms of the autophagy induction are not entirely clear. Much of the research has established the role of autophagy in the development of diabetic cardiomyopathy [7]. It has been reported that both apoptosis and autophagy are essential for cell survival [8]. Well-established type 1 diabetic animal models, namely STZ-induced diabetic mice [9] and OVE26 mice [10], exhibit inhibited of cardiac autophagy along with cardiomyocyte apoptosis and cardiac dysfunction. The crosstalk between autophagy and apoptosis is complex, and Hu et al. revealed that autophagy or autophagy-related proteins can lead to apoptosis or autophagic cell death under certain conditions [11].

AMP-activated protein kinase (AMPK), a key regulator of energy metabolism in the heart, is involved in the regulation of many cellular processes. As both anti-apoptotic and pro-apoptotic actions of AMPK have been reported, cardiomyocytes exposed to hyperglycaemia and oxidative stress under diabetic conditions can trigger both autophagy and apoptosis [8]. Despite of an increasing interest in the potential role of AMPK in the pathogenesis of diabetes and the possible use of AMPK-stimulating agents as treatment for this disease, this is the first study to demonstrate that exogenous H₂S leads to a significant increase in AMPK activity in the cardiac tissue of rats with increasing autophagosomes in type I diabetes. This finding supports the idea that increasing autophagy in cardiomyocytes through pharmacological activation of AMPK could be a novel strategy for the treatment of type I diabetes.

Hydrogen sulphide (H₂S) is the third endogenous gas to be identified as an important cell signalling molecule [12, 13]. H₂S is produced naturally in mammalian tissues and exhibits various biological and physiological effects [14-17], such as modulated metabolic state [18], mitochondrial function, cellular redox status, and cell apoptosis [19].

High glucose induces formation of reactive oxygen species (ROS), whereas H₂S represents a remarkable capacity to scavenge ROS. As a gasotransmitter with exclusive biologic effects, H₂S can be endogenously generated by cystathionine-β-synthetase (CBS) and cystathionine-γ-lyase (CSE) in mammalian tissues [13, 20, 21].

In the diabetic cardiovascular system, little information is available for the relationship between H₂S and autophagy. Therefore, we utilized in vivo and in vitro studies to identify the involvement of autophagy with H₂S treatment during cardiac dysfunction. We also investigated the potential mechanism(s) of action in H₂S -facilitated autophagy with a focus on the AMPK–mTOR pathway.

One hundred and twenty male Wistar rats were purchased by Harbin Medical University (8 weeks of age weighing 200-250 g). The rats were fed with a normal rat diet and had free access to water under conditions of standard temperature (20–22˚C) and humidity (50-60%) in standard lighting (alternating 12 h light/dark cycle). After one week of acclimatization, the rats were fasted overnight before the experiments. The animal protocol was reviewed and adheres to the Declaration of Helsinki and International Ethical Guidelines for Biomedical Research. This study protocol was approved by the Animal Ethics Committee of Harbin Medical University.

Random, morning blood glucose, non-fasting, concentration obtained from tail veins was determined at 4- to 8-week intervals in all groups by Roche AccuChek ® Inform (Roche Diagnostics, Indianapolis, IN).

Sodium hydrosulfide (NaHS), Streptozotocin (STZ; Sigma Chemical), GYY4137 (morpholin-4-ium-4- methoxyphenyl morpholino phosphinodithioate), bafilomycin A1 (02911643), DL-proparglycine (PPG) were purchased from Sigma Chemical (St. Louis, MO). Hoechst 33342, Propidium iodide (PI) and 2, 7- dichlorofluorescein diacetate (DCFH-DA) were obtained from Beyotime (Shanghai, China). N-acetylcysteine (NAC), siRNA AMPK were purchased from Cell Signalling Technology (Beverly, MA). Lipofectamine 2000 and JC-1 were both obtained from Invitrogen (Carlsbad, CA). CSE, Beclin-1, AMP-activated protein kinase (AMPK), p-AMPK (Thr172 ), p62 , LC3 , mammalian target of rapamycin (mTOR ), p-mTOR p70-S6 Kinase，phospho-p70 S6K，4EBP, phospho-4EBP antibodies are provided by Cell Signalling Technology (Beverly, MA, USA). Cytochrome c, pro-caspase-3/9 and cleavage-caspase 3/9 antibodies were provided by Proteintech Group (Chicago, IL). DMEM-F12 medium and fetal bovine serums (FBS) were supplied by Hyclone (Logan, UT).

Streptozotocin (STZ; Sigma Chemical) dissolved in citrate buffer (pH 4.5) and diluted in 0.3 ml of normal saline (0.9% NaCl injection USP, Baxter) was administered intraperitoneally at a dose of 50 mg/kg body weight. The rats were randomly divided into three groups as follows: control group (n=40; neither injected with STZ nor treated with NaHS), diabetic group (Dia group; n=40; injected with STZ), diabetes + NaHS group (Dia+NaHS group; n=40; injected with STZ and administered with NaHS). Successful generation of a type I diabetic model occurred when a blood glucose level of 16.7 mM was obtained 72 hrs after STZ injection [22], and blood glucose levels ranged from 5 to 10 mM in the control group injected with saline only [23]. Rats with blood glucose levels <6 mM were non-diabetic and used as controls.

In the NaHS treatment group, rats were given the intraperitoneal injection of NaHS (100 µM) with dosages based on a previous study [24] and our preliminary experiments [25]. Briefly, age-matched diabetic rats and control rats were sacrificed after treatment for 4 and 8 weeks. For every blood sample, the serum was obtained after centrifugation, and the serum was stored at -80°C. Each left ventricular was subjected to 3 tests as follows: western blotting for autophagic and apoptotic-related proteins; haematoxylin and eosin staining; and transmission electron microscopy (TEM).

The mixed cells and left ventricular myocardial tissues were collected and fixed in 2.5% glutaraldehyde under transmission electron microscopy. After dehydrated through an ascending ethanol series and embedded in epoxy resin, ultrathin sections were cut and stained with uranyl acetate and lead citrate. All ultrastructural analyses were performed in a blinded and nonbiased manner from photomicrographs captured using the Philips CM120 electron microscope.

After 4 and 8 weeks, the plasma and myocardial contents of H₂S were measured by methylene blue as described by Chunyu [26] with modifications [27]. Briefly, tissues were homogenized in ice-cold 100 mM potassium phosphate buffer (pH 7.4). The assay mixture (500 µl) contained tissue homogenate (450 µl), l-cysteine (10 mM; 20 µl), pyridoxal 5′-phosphate (2 mM; 20 µl), and saline (30 µl). Incubation was carried out in tightly sealed eppendorf vials. After incubation (37 °C, 20min), 1% zinc acetate (250 µl) was injected to trap evolved H₂S. This was followed followed by 10% trichloroacetic acid (250 µl) to precipitate protein and stop the reaction. Thereafter, N,N-dimethyl-p-phenylenediamine sulfate (20 mM;133 µl) in 7.2 M HCl was added followed by FeCl₃ (30 mM; 133 µl) in 1.2 M HCl and absorbance (670 nm) of aliquots of the resulting solution (300 µl) was determined after 10 mins using a 96-well microplate reader (Costar, USA). The same procedure was applied to measure the plasma H₂S level. All chemicals and reagents were obtained from Sigma (USA).

Protein homogenates were prepared from left ventricular tissue, and the protein concentration of the cell lysates was then determined using a BCA protein assay reagent kit (Beyotime，China). Equal amounts of protein (50 µg) were loaded on 10 % SDS-PAGE gels, transferred onto nitrocellulose membranes and blocked with 5 % non-fat milk. The membranes were incubated with primary antibodies overnight at 4 °C. The membranes were then incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse secondary antibodies at room temperature for one hour. In this study, β-actin and VDAC were served as loading control for the cytosolic and mitochondrial fraction, respectively.

Isolation of mitochondrial protein from left ventricular tissue was performed according to a mitochondria isolation kit (Beyotime Inc., Shanghai, China). The cardiac tissues (n=4-5, per group) were washed twice with ice-cold PBS, resuspended in lysis buffer, and then homogenized by an homogenizer in ice water. After removing the nuclei and cell debris by centrifugation at 1, 000 g for 10 min at 4°C, the supernatants were further centrifuged at 10, 000 g for 20 min at 4°C. Th protein concentration was determined using the Bicinchoninic acid (BCA) assay kit (Beyotime Inc., Shanghai, China).The resulting mitochondrial pellets were resuspended in lysis buffer,which was stored on ice and used for experiments within 4 h.

A Clark oxygen electrode was used for examining the oxygen consumption of mitochondria [28]. Mitochondria were incubated in respiration buffer (100 mM KCl, 25 mM sucrose, 5 mM KH₂PO₄, 1 mM MgCl₂, 1 mM EGTA, 10 mM HEPES, 10 mM glutamate, and 2.5 mM malate) [29], and state 3 of respiration (consumption of oxygen in the presence of substrate and ADP) was initiated with ADP (75 nmol/mg protein). States 3 and 4 (consumption of oxygen after ADP phosphorylation) respiration, respiratory control ratio (RCR = state 3/state 4), and ADP/O index (a marker of the mitochondrial ability to couple oxygen consumption to ADP phosphorylation during state 3 respiration) were determined according to Chance and Williams (1956) [30].

H9C2 cells have been shown to retain several electrical and hormonal characteristics similar with adults’ cardiomyocytes. In current study, we used H9C2 cells to establish high glucose model. H9C2 cells (Chinese Academy of Medical Sciences, Shanghai, China) were cultured in Dulbecco’s Modified Eagle Media supplemented with 10% FBS and 100 µg/mL penicillin&streptomycin (Beyotime , China) at 37°C under an atmosphere of 5% CO₂ and 95% O₂. H9C2 Cells were cultured to 70–80% confluence and treated with either high glucose (40 mM) or H₂S (dissolved in phosphate buffered saline) in combination.For inhibitor experiments, H9C2 cells were pre-incubated with a selective autophagy inhibitor, 3-MA (10 µM), or an AMPK inhibitor, compound C (2.5 µM), PPG (0.3 mM) and then treated without or with H₂S together with high glucose. The stimulant concentration of high glucose (40 mM) for H9C2 cells was determined from our pilot studies.

Cell viability was determined using a short-term microculture tetrazolium (MTT) assay. Cells were seeded into 96-well plates at a density of 2.5 x 10⁴ cells/well and incubated in 100 µl of culture media then exposed to different concentrations of glucose for varying time periods. The viability of H9C2 cells was determined by adding MTT to the cell cultures to reach a final concentration of 1 mg/ml. After the MTT-containing DMEM was removed, the remaining dark crystals formed were dissolved in 75 µl HCl/isopropyl alcohol solution. Optical densities at 570 nm were measured by a plate reader with an appropriate filter.

Morphological changes of apoptosis including cellular morphological change, chromatin condensation chromosomal condensation of H9C2 cells, were observed by Hoechst 33342 staining followed by fluorescent microscope. In brief, after different treatments, H9C2 cells were washed after three-time washes with ice cold PBS, cells were resuspended in 1 ml of PBS solution with Hoechst 33342 /PI and the cells were visualized under a fluorescent microscope.

After different treatments, cells were incubated with 10 µM Probes 2′, 7′- dichlorofluorescin diacetate (DCFH-DA; Beyotime, Shanghai, China) in a 37°C humidified incubator for 30 mins and washed twice with phosphate-buffered saline, pH 7.4 [31]. The slides were washed three times with PBS and DCF fluorescence was measured over the entire field of vision by using a fluorescence microscope connected to an imaging system (BX50-FLA, Olympus, Tokyo, Japan). Mean fluorescence intensity (MFI) from three random fields was analyzed by Image pro plus，and MFI was used to represent the amount of ROS. The experiment was carried out three times.

We assessed the mitochondrial membrane potential using JC-1 as described below. Mitochondrial membrane potential (MMP, Ψm) is the main indicator of mitochondrial metabolism [32]. MMP was assessed using a fluorescent dye, JC-1, a cell-permeable cationic dye that preferentially enters the mitochondria based on the highly negative MMP. Depolarization of MMP results in the loss of JC-1 from the mitochondria and a decrease in intracellular fluorescence. The cells were seeded on a slide with DMEM. After the indicated treatments, the slides were washed three times in PBS. Cells were stained with 5 µM JC-1 dye by incubating at 37 °C for 15 min. The dye changed reversibly from red to green as the mitochondrial membrane became depolarized. Fluorescence was measured using a fluorescence microscope and the MFI of JC-1 from 5 random fields was analysed using ImagePro Plus software.

Autophagosomes in the cells were detected by MDC staining using the method as previously described [33]. Briefly, the cells were incubated with 0.05 mM MDC (Sigma) in Hanks’ buffered salt solution at 37 °C for 10 mins. After being washed three times with 0.1 M PBS, MDC-labelled autophagosomes were examined by a fluorescence microscope and counted in x200 fields (five sequential fields were counted and averaged per coverslip) for three coverslips in each experiment as a percentage of total cell number.

Intracellular free H₂S levels were determined using 7-Azido-4-methylcoumarin (C-7Az). This probe is used both to monitor enzymatic production of H₂S in vitro [34] and to visualize H₂S in cells [35]. After the indicated treatments, H9C2 cells were washed with PBS and incubated with 50 µM AzMc combined with a surfactant at 37 °C for 20 mins in the dark. The MFI of 4 random fields was measured over the entire field of vision and analysed by Image pro plus software.

AMPK siRNA (a pool of 3 target-specific19–25 nucleotide siRNAs designed to knockdown the gene expression of mitofusin-2 in rats) and scrambled control were purchased from Cell Signalling Technology, Inc. (Beverly, MA). Transient transfection was initially standardized to improve knockdown efficiency. Subsequently, all the transient transfections were performed with 150 nM AMPK siRNAs using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol. After 6 hours, the transfection mixtures were replaced with regular medium. After 48 hours’ transfection, cells were incubated with medium containing 5 or 40 mM glucose for 24 hours. Cells were processed for assessment of mitochondrial networks and protein expression.

All values were given as the means ± the standard error of the mean (S.E.M.) for at least three replicates. Statistical data was analyzed with Newman t –test or ANOVA as appropriate using Graph Pad Prism 5 for Windows platform (Graph Pad Software, San Diego, CA). P＜0.05 was considered statistically significant.

A well-established type I diabetes mellitus model of STZ injection was used in this study. As expected, levels of blood glucose significantly increased in long-term STZ-induced diabetic rats. Meanwhile, an obvious decrease in body weight was found when compared with control rats (Table 1), confirming their diabetic state. Heart weight and blood glucose levels in 8 weeks diabetic rats were similar to those of 4 weeks diabetic rats, and these levels were higher than those of the control group (Table 1). Echocardiography was used to assess cardiac structure and function (Table 2). LVEDD was significantly increased in the 4 wks and 8wks Dia groups compared with control group. However, LVESD was significantly increased in the 4 wks Dia group but decreased in the 4 wks Dia+ NaHS group (P< 0.05). Left ventricular Ejection Fraction (LVEF) and Left Ventricular Fractional Shortening (LVFS) were both decreased in the Dia and Dia+NaHS group. Administration of NaHS modestly increased the LVEF in the 8 wks Dia+NaHS group compared to the 8 wks Dia group (P>0.05). Furthermore, the E/A ratio was decreased in the 4 wks Dia groups compared to the control, and the indexes increased in the 4 wks Dia+NaHS group compared to the diabetic groups (P< 0.05).

Hydrogen sulphide content was determined by methylene blue and sulphide electrode methods. Our results showed that H₂S contents in plasma and myocardial tissue were significantly lower in the Dia group than in the control group. However, in the Dia+ NaHS group, H₂S levels were markedly higher than those in the Dia group (Fig. 1A-D).

The expression of CSE proteins in cardiac tissues decreased in the 4 and 8 wks Dia groups compared to the control group (P< 0.05). The expression of myocardial CSE increased significantly in the Dia+NaHS groups compared to the diabetes groups (P<0.05) after 4 and 8 weeks of exogenous H₂S treatment (Fig. 1E).

Transmission electron microscopy revealed that there were some autophagic vacuoles (two-layer membrane structures that wrap around partially degraded cargo) in the sections from the experimental groups. In contrast, there was no evident abnormality in the control group. The percentage of autophagic vacuoles per cytoplasm area was significantly increased in the Dia+NaHS groups compared to the diabetic group (Fig. 2A). The expression of Beclin-1 was significantly increased in vivo in the Dia+NaHS groups compared to the Dia group (Fig. 2B). Moreover, the expression of p-AMPK/AMPK was decreased in the diabetic groups, whereas administration of NaHS significantly increased the expression of p-AMPK/AMPK (Fig. 2C).

In a first approach to investigate the impact of H₂S on mitochondrial bioenergetic function, mitochondrial respiration and oxidative phosphorylation were evaluated in freshly isolated cardiac mitochondria energized with succinate, a complex II substrate. Mitochondrial coupling and oxidative phosphorylation were measured and indicated by the RCR and ADP/O indexes, respectively. ∆Ψm results in the conversion of ADP to ATP via ATP synthase, which is essential for oxidative phosphorylation occurrence. Mitochondrial respiratory chain pumps H⁺ out of the mitochondrial matrix across the inner mitochondrial membrane. Our results showed that state 3 respiration was decreased significantly in cardiac mitochondria in long-term hyperglycaemic rats and that state 4 respiration was increased compared to Dia+NaHS groups (Fig. 3A&B). The RCR and ADP/O index decreased in the diabetes groups, which indicated that the respiratory chain was impaired (Fig. 3C&D).

To explore the effect of HG on cell viability, H9C2 cells were treated with normal (5.5 mM), 25 mM and 40 mM glucose for 24 and 48 hrs. Cell viability decreased by 21% and 35% in the 40 mM HG treatment groups after 24 and 48 hrs compared to the LG groups (Fig. 4A). Therefore, we selected two time points, namely 24 and 48 hrs, for autophagic observation (Fig. 4D) by MDC. Based on the data in Fig. 4A, 40 mM HG treatment for 48 hrs caused significant cell injury (measured by cell viability). We selected 40 mM HG treatment for 48 hrs as our model group for further research as these conditions were suitable to mimic the hyperglycaemia in DC.

It has been reported that GYY4137 releases H₂S slowly in aqueous solution in vitro [36]. Release of H₂S from GYY4137 in aqueous solution is pH-dependent with considerably greater release at pH 3.0 than at neutral or alkaline pH [37]. GYY4137 is more expensive but causes long-lasting relaxation of cells in vitro compared to NaHS. Thus, in this study, we used GYY4137 to further examine the role of H₂S in high glucose-treated H9C2 cells, and we used NaHS to examine the role of H₂S in rats. When H9C2 cells were incubated with 50, 100, and 200 µM GYY4137 for 48 hrs, cell viability (Fig. 4B) and MMP collapse were improved in the HG-treated group in a concentration-dependent manner (Fig. 4E&F). However, treatment with 100 µM GYY4137 alone had no such effects. As shown in Fig. 4E, mitochondria were considerably damaged after 40 mM HG treatment for 48 hrs, and MMP collapse was prevented by different concentrations of GYY4137 treatment. Our results suggested that 100 µM GYY4137 protects mitochondrial function and significantly increases cell viability of cells treated with high glucose. As shown in Fig. 4C, GYY4137 increased autophagosomes compared to the HG group. GYY4137 treatment after 48 hrs further increased autophagy under HG, indicating that H₂S at least in part increased autophagy production under HG. Therefore, we selected 100 µM GYY4137 for further research. Autophagic vacuoles detected by MDC were increased after treatment with GYY4137 100 µM for 48 hrs (Fig. 4D). Thus, 48 hrs was used as the endpoint for observation in vitro.

As shown in Fig. 5 A, the mean fluorescence intensity was significantly increased in the high glucose-cultured group but was decreased after 48 h treatment with GYY4137. In addition, the exogenous supply of GYY4137 can be converted into H₂S through enzymatic H₂S biosynthesis in cells. Our results revealed that the H₂S level in cells treated by HG was significantly decreased, and these alterations were attenuated after treatment with GYY4137 (Fig. 5B).

ROS releases in high glucose states are related to various cardiac diseases. Reducing oxidative stress leads to significant reduction of heart pathology in degenerative diseases [12]. The levels intracellular ROS in H9C2 cells cultured with high glucose for 24 hrs and 48 hrs were determined (Fig. 5C). As expected, exogenous H₂S treatment significantly reduced high glucose-induced intracellular ROS levels (P<0.05). H9C2 cells cultured in high glucose were also treated with the ROS scavenger NAC (100 µM). Intracellular ROS levels increased significantly 48 hrs after exposure to 40 mM DMEM. The NAC treatment group also reduced hyperglycaemia-induced ROS accumulation (Fig. 5D).

Moreover, our results showed the protective effect of H₂S on high glucose-induced cell death in H9C2 cells. H9C2 cells were exposed to high glucose and then treated with H₂S for the indicated time period. Cell viability was assayed by a MTT assay. As shown in Fig. 4B, treatment of H9C2 cells with H₂S markedly reduced high glucose-induced cell injury. H₂S also protected against high glucose-induced cellular apoptosis (Fig. 5E). DL-proparglycine (PPG), an irreversible competitive CSE inhibitor, was used to further investigate the role of H₂S in inhibiting apoptosis and inducing autophagy. For the PPG group, cells were pre-treated with PPG (0.3 mM) for 60 mins before high glucose induction was administered. The induction of apoptosis is a major event in the response of H9C2 cells to hyperglycaemic stress. Thus, we investigated the role of apoptosis in PPG-treated cells by Hoechst/PI staining. Necrotic cell death was as detected by propidium iodide (PI). Fig. 5E shows that the apoptotic cells increased in HG+PPG cells compared to GYY4137-treated cells after 24 or 48 hrs. However, GYY4137+PPG treatment did not significantly affect cell apoptosis compared to the PPG group (Fig. 5F). Viable cells displayed normal nuclear size and uniform fluorescence in the control group, whereas apoptotic cells showed condensed, fractured, or distorted nuclei in the high glucose group. White arrows indicate the fragmented and condensed nuclei by Hoechst staining. Our results showed that the H₂S-treated group showed low levels of cytochrome C release from the mitochondria into the cytoplasm compared to the HG group (Fig. 4A). Our data showed that high glucose significantly induced pro-caspase 3 and pro-caspase 9 cleavages in H9C2 cells (Fig. 6B&C). Our data confirmed that high glucose activated caspase-3 and caspase-9 in H9C2 cells, and H₂S suppressed this activation. These results indicated that H₂S inhibits oxidative stress, which attenuates apoptosis induced by high glucose.

AMPK, a major regulator of cellular energy homeostasis, is known to activate autophagy, and inhibition of mTOR is associated with autophagy induction [38]. We examined the phosphorylations levels of AMPK and mTOR to elucidate the molecular autophagy mechanism induced by H₂S during high glucose in vitro. The ratio of p-AMPK/AMPK was decreased after HG induction, while the ratio of p-mTOR/mTOR was significantly increased in high glucose group. To determine the involvement of mTOR signalling in H₂S-mediated activation of autophagy during high glucose-induced pathogenesis, Compound C (CC), an inhibitor of AMPK, was used to determine the involvement of mTOR signaling in H₂S-induced activation of autophagy during high glucose treatment (Fig. 6D).

Interestingly, the activity of the mTOR pathway was dramatically increased by high glucose, as indicated by decreased phosphorylation of P70S6K (a downstream substrate of mTOR Complex 1) and 4EBP (a downstream target of mTOR Complex 1), compared to low glucose. After treatment with GYY4137 or pre-treatment with CC, these indexes decreased compared to those in the high glucose treatment alone group (Fig. 6E). Therefore, mTOR signalling acted as a downstream effector of AMPK signalling in regulation of autophagy mediated by the administration of H₂S during high glucose treatment.

Autophagy usually improves cell survival, and recent studies have reported that autophagy protects cells from apoptosis [39, 40]. LC3 is initially synthesized in its unprocessed form, and cardiac autophagy is suppressed as evidenced by a decrease in conversion of LC3-I to LC3-II [41]. LC3-II is localized in the autophagosome membrane and serves as a reliable protein marker of autophagy. LC3-II can be degraded in autolysosomes during autophagy [42]. Therefore, treatment with the autophagic inhibitor, 3-MA, or CC will enhance the level of LC3-II if H₂S increases autophagic flux. H9C2 cells were treated with GYY4137 (100 µM) and AICAR (AMPK activator, 20 µM) for 48 hrs. Cells showed an increase in the Beclin-1/β-actin ratio and LC3II/I ratio compared to cells treated with high glucose alone (Fig. 7A). In addition, a significantly decreased p62/β-actin ratio was observed after high glucose treatment. The p62/β-actin ratio indicated that GYY4137-induced autophagy was significantly decreased by 3-MA or CC pre-treatment. There was no significant autophagy after treatment with 3-MA or CC alone for 24 hrs compared to control cells (both P > 0.05). Our results indicated that H₂S induces autophagy in H9C2 cells in an AMPK-dependent manner (Fig. 7B).

To determine autophagic flux after exogenous H₂S addition, we cultured cells in the absence or presence of BAF (100 nM) DMEM with low glucose (5.5 mM) and high glucose (40 mM) for 48 hrs. BAF markedly increased LC3 II levels in cells treated with GYY4137, whereas BAF had only a small effect on LC3 II content when cells were exposed to high glucose, suggesting that H₂S-mediated induction of LC3 II levels was attributed mainly to increased AV formation rather than reduction of lysosome degradation (Fig. 7C).

To determine if AMPK is responsible for the autophagic induction by H₂S, H9C2 cells were transfected with AMPK siRNA. As shown in Fig. 8A, cells exhibited a profound decrease of AMPK. These results strongly suggested that AMPK activation is essential for a protective effect of H₂S in H9C2 cells under oxidative stress. In addition to H₂S, knockdown of AMPK increased the ratio of p-mTOR/mTOR, while the ratio of LC3-II/LC3-I was significantly reduced (Fig. 8B&C). These results demonstrated that H₂S-induced autophagy is linked to the AMPK pathway. Knockdown of AMPK in cells led to an augmentation of apoptosis (Fig. 8D), thereby suggests the involvement of AMPK-dependent autophagy in cardioprotection.

Diabetic cardiomyopathy is a critical complication of diabetes [43]. A comprehensive understanding of mechanisms underlying the pathogenesis of diabetic cardiomyopathy is urgently needed. Our study presented three major findings. Firstly, NaHS administered after high glucose treatment elevated serum levels of H₂S and protected against hyperglycaemia injury. Secondly, cardioprotection conferred by the H₂S donor, NaHS and GYY4137, was dependent on AMPK activation in vivo and in vitro. Finally, H₂S suppressed high glucose injury by activating autophagy via the AMPK/mTOR pathway, which is likely the mechanism underlying cardioprotection conferred by H₂S.

H₂S exerts a wide influence on the cardiovascular system [44, 45]. There is ample evidence that cardiovascular disease is related to decreased expression of CSE [46, 47]. Our previous studies have found that H₂S regulates the interactions and cause a switch among cell death pathways during hyperglycaemia [25, 32]. In the present study, we established a classic type I diabetic rat model to investigate the protective effects of H₂S against hyperglycaemia-induced cardiac dysfunction. GYY4137 is a well-known H₂S donor that has vasodilatory and antihypertensive activities [37]. We selected GYY4137 as our tool to explore the cardiovascular biology of H₂S in vitro as it reflected endogenous physiological release.

Our findings indicated that administration of NaHS increased H₂S contents in plasma and myocardial tissue in diabetic rats (Fig. 1A-D) and that H₂S could protect against left ventricular systolic dysfunction in the diabetic group according to echocardiographic data (Table 2).

Previous studies have shown that mitochondria are involved in the process of high glucose injury through enhancing the production of mitochondrial ROS and opening of the mPT pore [25]. The formation of ROS from cardiomyocytes is a key factor in the pathogenesis of diabetic complications [43, 48, 49]. In this study, we surprisingly found that high glucose decreased mitochondrial oxidative phosphorylation efficiency, induced mitochondrial swelling, promoted mitochondrial Cyt C release and promoted ROS generation (Fig. 5B&Fig. 6A). In contrast, the administration of exogenous H₂S to isolated mitochondria significantly improved mitochondrial state 3 respiration in a concentration-dependent manner (Fig. 3). These results strongly indicated that there may be a causal relationship between mitochondrial dysfunction and anomalous generation of ROS in high glucose injury. Elevated levels of mitochondrial ROS could damage the mitochondrial membrane and associate with increased mitochondrial membrane permeabilization. In response to high glucose, autophagosomes chelate ROS-generating mitochondria to protect cells, while 3-MA increases the production of ROS and advances apoptosis [50]. Cellular apoptosis associated with oxidative stress in multiple organs of diabetes mellitus has been documented [51, 52]. Autophagy inhibition in isolated cardiomyocytes has been shown to increase apoptosis and necrosis, while autophagy inhibition has also been shown to be protective in some studies [53, 54]_ENREF_50. Although H₂S has been demonstrated to inhibit apoptosis and necrosis, it remained unclear if autophagy is involved in H₂S-triggered protection [55]. In our study, we demonstrated that H₂S activated cytoprotective autophagy in vivo and in vitro. Autophagy was evaluated by the following methods: TEM and MDC fluorescence evaluation of autophagosomes provided a valid insight into what events were happening inside in the cell; and western blot analysis was used to interpret the molecular changes as indicated by LC3 induction and p62 degradation. Our results further supported that autophagy is a homeostatic mechanism for maintaining cardiac function by managing cellular injury.

To date, several cellular signalling pathways are speculated to turn on autophagy under high glucose [54]. AMP-activated protein kinase (AMPK) exerts its biological effects in cardiovascular functional regulation. Activation of AMPK inhibits diabetes-induced cardiomyocyte apoptosis and prevents cardiac dysfunction in diabetic animals after treatment with metformin [9], suggesting that further investigations are needed to explain the role of AMPK in the regulation of autophagy and apoptosis. Interestingly, our previous study demonstrated that H₂S inhibits induction of protein synthesis in the endothelial cells after high glucose treatment by activation of AMPK [56]. Indeed, activation of AMPK inhibits mTOR, the negative regulator of autophagy, and subsequently stimulates autophagy. Moreover, synergistic effects are observed when H₂S treatment is combined with an AMPK inhibitor. Our current study demonstrated that inhibition of AMPK phosphorylation and obverted phosphorylation of mTOR in response to high glucose injury (Fig. 6 D&E).

Compound C and 3-MA were applied in the high glucose model to confirm the role of autophagy in the cardioprotection-induced by H₂S via the AMPK-mTOR pathway. These compounds potently reduced the cytoprotective effects of H₂S. Cardiac autophagy was suppressed as evidenced by a decrease in conversion of LC3-I to LC3-II [5]. The decreased LC3 conversion and expression of Beclin-1 after treatment with GYY4137 or AICAR in the present of high glucose was attributed to alterations in the formation of autophagosomes (Fig. 7A) .The upstream activity of autophagy processes was determined by the significant alterations of p62 (Fig. 7B). Our results agreed with previously published studies as they showed that compound C and 3-MA resulted in cardiac dysfunction by blocking autophagy activation [57].

Based on the aforementioned results, we postulated that the GYY4137-elicited cytoprotective effects against high glucose might be due to autophagy activation via the AMPK-mTOR pathway. Thus, we further investigated the role of AMPK in the protection of diabetic cardiomyopathy. Finally, AMPK siRNA increased mTOR phosphorylation and successfully inhibited autophagy, leading to increased apoptosis and cell injury (Fig. 8A-B). Thus, these results suggested that the AMPK-mTOR pathway participates in H₂S -induced autophagy regulation (Fig. 9). However, the role of autophagy is still controversial in cardiovascular diseases. Autophagy as a repair mechanism is activated with low intensity stress. With increasing stress levels, apoptosis begins to occur. Under extreme stress, ATP depletion results in necrosis and neither autophagy nor apoptosis can progress [58].

In conclusion, autophagy may be essential in H₂S-induced cardioprotection against high glucose injury. H₂S-activated AMPK-mTOR-dependent pathway may attribute to the regulatory mechanism for HG-induced autophagy. Therapeutic H₂S or autophagy activation may offer a promising new strategy in diabetic heart disease treatment. In addition, further studies are needed to examine the association between H₂S-induced autophagy and the AMPK-mTOR pathway in both type I and type II diabetes.

The project is supported by the grants from National Science Foundation of China (81370330, 81370421, and 81670344). Graduate innovation Foundation of Harbin Medical University (YJSCX2014-05HYD).

No conflicts of interest, financial or otherwise, are declared by the author(s).