Empagliflozin Ameliorates Obesity-Related Cardiac Dysfunction by Regulating Sestrin2-Mediated AMPK-mTOR Signaling and Redox Homeostasis in High-Fat Diet–Induced Obese Mice

Obesity is a growing health problem worldwide that is strongly linked to an increased risk of cardiovascular disease, cardiac dysfunction, and chronic heart failure (1,2). Thus, developing new strategies to prevent and treat obesity-related cardiac dysfunction is highly desirable. Numerous mechanisms such as insulin resistance, ectopic fat accumulation, lipotoxicity, and imbalanced adipocyte secretion have been shown to play a role in these pathophysiological processes. Moreover, these alterations can lead to progression of the pathological cardiac remodeling that manifests in cardiac hypertrophy and fibrosis (3). The process of cardiac hypertrophy is an adaptive response to multiple pathological stresses and is accompanied by excessive oxidative stress and chronic activation of mammalian target of rapamycin (mTOR) (4,5). AMPK is a major cellular sensor of energy availability. In obesity, AMPK phosphorylation in the heart is often downregulated due to overnutrition (6). Acting as a cardioprotective molecule, AMPK balances oxidative stress and inhibits mTOR signaling by directly phosphorylating the TSC1/2 complex (7,8). This AMPK-mediated mTOR regulation has been shown to have pleiotropic cardioprotective effects that restore cardiac dysfunction, including improving the energy supply and regulating other physiological processes (7).

Sestrin2 (Sesn2) is a novel stress-inducible protein that lacks kinase activity but has been shown to have two important functions: inhibition of mTOR by its COOH-terminal domain and suppression of oxidative stress through its N-terminal domain (9). Both functions are vital for preventing and treating cardiovascular disease, including ischemia-reperfusion injury and cardiomyocytes hypertrophy (10). For example, Sesn2 has been shown to attenuate phenylephrine-induced cardiac hypertrophy in neonatal rats (10). Our recent findings revealed that Sesn2 prevents age-related intolerance to myocardial infarction by activating AMPK (11). Sesn2 also promotes phosphorylation of AMPK through its upstream kinase liver kinase B1 (LKB1)–dependent way and mediates genotoxic stress-induced AMPK activation and mTOR inhibition (12,13). However, whether Sesn2 is also involved in obesity-related cardiac dysfunction remains unknown.

Sodium–glucose cotransporter 2 inhibitors (SGLT2i) are a novel class of hypoglycemic agents that selectively inhibit sodium and glucose reabsorption from the proximal renal tubule, increase the excretion of urine glucose, and thus decrease blood glucose levels (14). Several large randomized, controlled cardiovascular outcomes trials (EMPA-REG OUTCOME trial, CANVAS Program, and DECLARE‐TIMI 58) and a meta-analysis have demonstrated unprecedented favorable cardiovascular outcomes in patients with diabetes treated with empagliflozin (EMPA), canagliflozin, or dapagliflozin (15–18). There are several mechanisms to explain these beneficial effects of SGLT2i on cardiovascular outcomes, such as weight loss, and improved glucose metabolism, blood pressure (BP), and sodium overload (19). However, in subjects without obvious diabetes and hypertension, SGLT2i also seemed to show cardiac protection (20). Additionally, the direct impact of SGLT2i on the heart remains controversial, as only SGLT1, and not SGLT2, is expressed in the myocardium (21). Hence, understanding the underlying pathophysiological mechanisms responsible for these positive effects requires further investigation.

While the consensus is that SGLT2i improves cardiac function in diabetes, no study has focused on whether SGLT2i could directly improve obesity-related cardiac dysfunction. In this study, we used the high-fat diet (HFD)–induced obesity model and Sesn2 knockout (KO) mice to test the hypothesis that EMPA improves obesity-related cardiac dysfunction via regulating Sesn2-mediated AMPK-mTOR signaling and the cellular antioxidant system.

Male 6-week-old wild-type (WT) C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and randomly assigned to the normal control (NC), NC-EMPA (NC-E), HFD, and HFD-EMPA (HFD-E) groups. The NC and NC-E groups were fed a normal chow diet (17% energy from fat, 29% from protein, and 54% from carbohydrate; 300 kcal/100 g) (catalog number 8640; Envigo, Somerset, NJ), while the other two groups were fed an HFD (45% energy from fat, 19% from protein, and 36% from carbohydrate; 460 kcal/100 g) (catalog number TD.06415; Envigo) to induce obesity. After 12 weeks, the NC-E and HFD-E groups received EMPA (10 mg/kg) (Sigma-Aldrich, St. Louis, MO) by oral gavage once daily for another 8 weeks, whereas the other two control groups received saline. Sesn2 KO mice (C57BL/6J background) were generated as previously described (12,13). The KO mice and WT littermates were randomly assigned to normal diet and HFD with EMPA or saline treatment. All groups were fed an HFD for 12 weeks and then treated with EMPA (10 mg/kg) or without EMPA for another 8 weeks. All mice were caged in a temperature-controlled environment on a 12-h light/dark cycle. Body weight (BW) and food intake were measured weekly. After 8 weeks of EMPA treatment, body composition and metabolic profiles, including glucose homeostasis, plasma insulin, lipids, and adipokine, were performed. These are described in the Supplementary Materials and Methods. The study protocol was approved by the Institutional Animal Care and Use Committees of the University of Mississippi Medical Center, Affiliated Hospital of Weifang Medical University, and University of South Florida.

Cardiomyocytes were isolated enzymatically from NC and HFD mice as previously described by our group (22). Briefly, after mice were anesthetized, hearts were excised and fixed onto a cardiomyocyte’s perfusion apparatus (Radnoti Ltd., Monrovia, CA). Hearts were first perfused with Krebs-Henseleit buffer oxygenated with 95% O₂/5% CO₂ and then digested with a solution containing liberase blendzyme. After successful digestion, cardiomyocytes were incubated with EMPA (1 μmol/L) at 37°C for 2 h, and then total protein was extracted for Western blot analysis.

BP was measured by a noninvasive tail-cuff device (BP2000; VisiTech International, Sunderland, U.K.) at the end of the study period. Cardiac function was determined by high-frequency ultrasonography using a Vevo 3100 imaging system (VisualSonics Inc., Toronto, Ontario, Canada) as previously described (23). The following main variables were assessed: left ventricular (LV) ejection fraction (EF), LV fractional shortening (FS), LV mass, the ratio of early diastolic to late diastolic mitral inflow velocities (E/A), isovolumic relaxation time (IVRT), and myocardial performance index (MPI).

To assess morphological changes, LV tissues were fixed immediately in 10% neutral buffered formalin and embedded in paraffin after anesthesia. Then, 5-μm sections were cut from the paraffin blocks and stained with hematoxylin and eosin for histopathological examination. To evaluate perivascular/interstitial fibrosis, sections were stained with both Masson trichrome and picrosirius red (Sigma-Aldrich). To evaluate myocardial inflammation, sections were stained with CD68 antibody. The heart sections were also used to evaluate lipid accumulation by Oil Red O staining. Micrographs were acquired using light microscopy (Nikon, Tokyo, Japan).

The cardiac ultrastructure was observed using transmission electron microscopy. Heart tissues were prepared as previously described by our group (23). Digital images were taken using an XR611 digital camera (Advanced Microscopy Techniques, Corp., Woburn, MA).

Mitochondrial reactive oxygen species (ROS) production was evaluated by MitoSOX Red (Invitrogen, Carlsbad, CA) as previously described (24). Briefly, fresh-frozen sections of LV were washed with PBS and then incubated in PBS containing 1 μmol/L MitoSOX Red mitochondrial superoxide indicator (Invitrogen) for 10 min at 37°C, protected from light. Images were taken using a fluorescence microscope (excitation: 510 nm; and emission: 580 nm).

Equal amounts of total protein from LV myocardial tissues were electrophoresed on 7–15% SDS-PAGE gels and transferred to polyvinylidene difluoride membranes. The membranes were incubated with primary antibodies (Supplementary Materials and Methods) overnight at 4°C, washed, and then incubated with horseradish peroxidase–conjugated secondary antibodies for 1 h. Immunoreactive bands were detected with SuperSignal West Femto (Thermo Fisher Scientific, Waltham, MA) in a Western blotting detection system (Bio-Rad Laboratories, Hercules, CA). Results are expressed as density values normalized to β-tubulin or GAPDH.

Total RNA from LV myocardial tissues was extracted using TRIzol (Invitrogen) and then synthesized into cDNA using the TaqMan High-Capacity RNA-to-cDNA Kit (#4387406; Applied Biosystems, Foster City, CA). Quantitative PCR was performed using iTaq Universal SYBR Green Supermix (#1725124; Bio-Rad Laboratories). Primer sequences are listed in the Supplementary Materials and Methods.

All data are expressed as the mean ± SD. Differences among more than two groups were compared by one- or two-way ANOVA with multiple-comparisons and post hoc tests for selected comparisons. P < 0.05 was accepted as statistically significant. Statistical analyses were performed with SPSS 22.0 (IBM Corporation, Armonk, NY) or Prism 8 (GraphPad Software, San Diego, CA).

The data sets analyzed during the current study are available from the corresponding author on reasonable request.

After 12 weeks, mice fed the HFD had significantly increased BW (32.6%) and fat mass (2.92-fold; P < 0.05). EMPA treatment of HFD mice for 8 weeks prevented further BW and fat mass gain (P < 0.05) (Fig. 1A–D). However, EMPA had no effect on lean mass (P > 0.05) (Supplementary Fig. 1A). Although HFD mice ingested fewer grams of food (2.8 vs. 4.3 g; P < 0.05) (Supplementary Fig. 1D), there was no significant difference in kilocalories of food intake between control and HFD mice (12.87 vs. 13.07 kcal; P > 0.05) (Supplementary Fig. 1E). EMPA had no effect on food intake (both by grams and kilocalories; P > 0.05) (Supplementary Fig. 1B–E). As expected, compared with control mice, HFD mice showed impaired glucose/insulin tolerance (higher glucose area under the curve), both of which were improved by EMPA treatment (P < 0.05) (Fig. 1E and F and Supplementary Fig. 1F and G). There was no significant difference in fasting blood glucose among the groups (P > 0.05) (Fig. 1G). Fasting insulin levels were higher in HFD mice compared with control mice (P < 0.05) (Fig. 1H). EMPA significantly decreased insulin levels by 58.27% (P < 0.05) but had no beneficial effects on free fatty acid (FFA), triglyceride, or cholesterol levels (P > 0.05) (Fig. 1I and J). To determine whether EMPA could improve adipokine profiles in HFD mice, we measured circulating adiponectin and leptin levels. EMPA treatment enhanced circulating adiponectin by 32.7% and reduced leptin levels by 75.36% in HFD mice (P < 0.05) (Fig. 1K and L). No significant differences in either systolic or diastolic BP were found among the groups (P > 0.05) (Supplementary Fig. 1H).

There were significant increases in LV and heart mass in HFD mice compared with control mice (increased by 40.18% and 39.7%, respectively; P < 0.05). EMPA treatment significantly reduced LV/heart mass (reduced by 20.66% and 23.61%; P < 0.05) (Fig. 2A and B). Echocardiographic measurements of heart rates, LV volume, interventricular septal dimensions, LV internal dimensions, and LV posterior wall thickness were similar among the four groups (P > 0.05) (Supplementary Fig. 2). No significant differences were observed among groups regarding systolic function parameters (LVEF and LVFS, P > 0.05) (Fig. 2C and D). However, compared with NC mice, HFD mice exhibited impaired diastolic function, including E/A, IVRT, and MPI (P < 0.05) (Fig. 2E–H). EMPA treatment significantly improved diastolic function, which manifested as enhanced E/A and reduced changes in IVRT and MPI (P < 0.05) (Fig. 2E–H). We did not observe differences between the NC and NC-E groups (P > 0.05).

In accordance with the observed morphometric changes, HFD mice exhibited obvious cardiac perivascular/interstitial fibrosis and increased fibrosis-related genes expression (FSP-1 and Postn) (P < 0.05) (Fig. 3A–C, E, and F and Supplementary Fig. 3). Additionally, Oil Red O staining revealed increased relative lipid contents in myocardial tissue from HFD mice (Fig. 3D). Interestingly, all of these abnormalities were ameliorated by EMPA treatment (Fig. 3). No significant differences were found between NC and NC-E mice (P > 0.05). To explore how EMPA markedly decreased lipid content, we detected acetyl CoA carboxylase (ACC) phosphorylation and found that EMPA could increase ACC phosphorylation (P < 0.05) (Supplementary Fig. 4).

Compared with NC mice, HFD mice exhibited an altered mitochondrial cristae architecture, including swollen, distorted mitochondria, fission, and inner mitochondrial membrane disruption (Fig. 4A). However, EMPA treatment reduced myocardial mitochondria injury (Fig. 4A). To further determine whether EMPA could reduce mitochondrial injury, we measured mtDNA contents by RT-PCR and evaluated mitochondrial ROS through the MitoSOX Red probe, which specifically targets mitochondria. EMPA treatment of HFD mice significantly reduced mitochondrial ROS (P < 0.05) (Fig. 4B and C) and enhanced mtDNA contents (P < 0.05) (Fig. 4D).

As the AMPK-mTOR signaling pathway is essential for cardiac hypertrophy (25), we investigated whether EMPA treatment can ameliorate myocardial abnormalities via the AMPK-mTOR signaling cascade. The results demonstrated that EMPA treatment significantly enhanced AMPK phosphorylation and reduced phosphorylation of mTOR and phospho-S6 (P < 0.05) (Fig. 5A and B). We have previously reported that Sesn2 serves as a scaffold protein to modulate AMPK-mTOR signaling in the heart (11,26); thus, we next determined whether these beneficial effects of EMPA treatment involved Sesn2-mediated regulation. Intriguingly, we found that EMPA treatment augmented both Sesn2 protein and Sesn2 mRNA levels (P < 0.05) (Fig. 5A and B). To further explore whether EMPA could directly activate cardiac Sesn2-AMPK, we incubated isolated cardiomyocytes from normal diet and HFD mice with or without EMPA for 2 h. Interestingly, we found that EMPA treatment not only augmented Sesn2 protein but also triggered AMPK phosphorylation, indicating that EMPA could activate cardiomyocytes Sesn2-AMPK independent of beneficial whole-body metabolic changes (P < 0.05) (Fig. 5C).

In addition to the important role of mTOR signaling in cardiac hypertrophy, phosphatidylinositol 3-kinase (PI3K)–Akt signaling is also critical for cardiac hypertrophy (25); thus, we investigated whether EMPA treatment could modulate the PI3K-Akt pathway. The results demonstrated that EMPA treatment inhibited phosphorylation of both PI3K and Akt triggered by HFD (P < 0.05) (Fig. 5D and E). Additionally, endothelial nitric oxide synthase (eNOS) phosphorylation was attenuated in HFD mice compared with NC mice (P < 0.05) (Fig. 5D and E). EMPA treatment increased phosphorylation of eNOS at Ser¹¹⁷⁷ with HFD, indicating the endothelial-protective effects of EMPA (P < 0.05) (Fig. 5D and E).

There is evidence that inflammation and oxidative stress are involved in HFD-induced cardiac dysfunction (27); thus, we examined the effects of EMPA treatment on HFD-induced cardiac functions. The results demonstrated that EMPA treatment significantly reduced myocardial nuclear factor-κB (NF-κB) p65 activity and MCP-1 protein levels caused by HFD (P < 0.05) (Fig. 6A). Moreover, EMPA treatment reduced macrophage marker (CD68) expression in the heart (Fig. 6B). Interestingly, EMPA treatment significantly upregulated the expression of oxidative signaling proteins and downstream antioxidant genes, including Nrf2/HO-1, catalase, and GCLM (P < 0.05) (Fig. 6A and C). The isolated cardiomyocytes from HFD mice were incubated with EMPA, augmented Nrf2/HO-1, and reduced NF-κB p65 activity (P < 0.05) (Supplementary Fig. 5).

We have previously identified Sesn2 as a critical scaffold protein that mediates cardiac AMPK signaling in response to stress conditions (13). To determine whether the beneficial effects of EMPA treatment on cardiac dysfunction induced by HFD are Sesn2 dependent, we compared the response of both WT littermates and Sesn2 KO mice to HFD with or without EMPA treatment. The results demonstrated that Sesn2 KO mice showed more obesity and metabolic disorders when exposed to HFD compared with WT littermates, as well as greater BW gain, fat mass, and glucose tolerance (P < 0.05) (Fig. 7A–F). Compared with Sesn2 KO mice with normal chow diet, Sesn2 KO mice with HFD (KO-HFD) mice showed both impaired systolic (reduced LVEF and LVFS) and diastolic function (P < 0.05) (Fig. 7G). However, no obvious improvement in cardiac function was observed in Sesn2 KO-HFD mice with EMPA (KO-HFD-E) treatment (P > 0.05) (Fig. 7G). Moreover, Masson staining and transmission electron microscopy data demonstrated that there was no improvement in cardiac interstitial fibrosis or myocardial mitochondria injury in KO-HFD-E treatment (Fig. 7H).

The immunoblotting results demonstrated that EMPA treatment led to the upregulation of AMPK phosphorylation and reduction of mTOR phosphorylation in WT littermates with HFD (WT-HFD) (Fig. 8A). However, these effects of EMPA treatment were not observed in Sesn2 KO-HFD mice (Fig. 8A). Additionally, compared with WT-HFD mice, EMPA treatment had no beneficial effects on HO-1 protein (Fig. 8B), Nrf2 mRNA levels (Fig. 8C), or NF-κB activation in Sesn2 KO-HFD mice (P > 0.05) (Fig. 8B). Similarly, no enhancement was observed in myocardial mtDNA in KO-HFD-E treatment (P > 0.05) (Fig. 8C).

The experiments described above elucidated the effect of EMPA treatment on obesity-related cardiac dysfunction in obese mice. We found for the first time that EMPA treatment attenuated myocardial hypertrophy and improved cardiac function via regulating cardiac Sesn2-mediated AMPK-mTOR signaling and maintaining cardiac redox homeostasis. Further studies found that Sesn2 is a stress-inducible protein that is required for the effects of EMPA in obesity-related cardiac dysfunction. This study demonstrated that the protective role of EMPA in obesity-related cardiac dysfunction is via regulating Sesn2.

The incidence of cardiac dysfunction has risen in parallel with increasing obesity-related metabolic disorders. Obese subjects are more likely to display the phenotype of heart failure with preserved ejection fraction, without significant reduction of LVEF or LVFS (28). The worldwide obesity-related cardiac dysfunction epidemic has heightened the need to find an effective therapy that can improve cardiac dysfunction and reduce the risk of cardiovascular disease. Although some drugs such as ACE receptor and angiotensin receptor blockers as well as mineralocorticoid receptor antagonists have been used to treat cardiac dysfunction, few effective therapeutic strategies have been found to prevent or attenuate the progression of cardiac dysfunction.

Recently, SGLT2i has been shown to have beneficial effects on cardiovascular disease outcomes by improving cardiometabolic profiles. The Dapagliflozin and Prevention of Adverse-Outcomes in Heart Failure (DAPA-HF) study demonstrated the incremental efficacy and safety of SGLT2i in patients with heart failure and reduced ejection fraction without diabetes (29). Our study found that HFD mice exhibited impairments in diastolic function, including reduced E/A, prolonged IVRT, and increased MPI, without significant alterations in LVEF and LVFS. However, 8 weeks of EMPA treatment improved LV hypertrophy and diastolic function. Other studies have also demonstrated that EMPA can improve cardiac dysfunction in other models, such as leptin-deficient and nondiabetic mice (29–31). In addition to improving cardiac function, EMPA also reduced cardiac fibrosis and lipid accumulation and attenuated myocardial mitochondrial injury. Our results are the first to demonstrate that EMPA can improve obesity-related cardiac structure and function in HFD mice.

To explore the potential mechanism of these cardioprotective effects of EMPA, we assessed metabolic profiles. The results demonstrated that EMPA prevented BW gain, reduced whole-body fat, and improved glucose homeostasis. We did not observe significant changes in BP or lipid levels. Thus, the cardiac function improvements of EMPA-treated HFD mice are independent of changes in BP and lipid levels. It is well known that altered adipokine secretion is a link between obesity and the increased prevalence of obesity-associated diseases. The high ratio of leptin/adiponectin appears to be associated with high risk of cardiovascular diseases in obesity (32). Our data found that EMPA treatment significantly ameliorated these disorders along with reducing leptin and enhancing adiponectin, which were also confirmed by a recent meta-analysis (33). These alterations are possibly due to the reduction of BW and whole fat mass.

Sestrins are a novel family of stress-inducible proteins that are involved in AMPK-mTOR signaling, redox, and autophagy. Our laboratory and others (13) have highlighted that Sestrins are vital for cardiac homeostasis. Sesn2 protects cardiomyocytes from ischemia-reperfusion–induced injury by initiating AMPK activation (11). However, whether Sesn2 is also involved in obesity-related cardiac function has not been studied. Our data showed that both Sesn2 protein and Sesn2 mRNA levels were lower in HFD cardiac tissue and that EMPA treatment augmented both Sesn2 protein and mRNA levels. To further explore whether these alterations were dependent on whole-body metabolic changes, we incubated the cardiomyocytes with EMPA for 2 h and found that EMPA treatment triggered Sesn2 protein levels, indicating that EMPA can induce cardiomyocytes Sesn2 upregulation independent of beneficial whole-body metabolic changes. Next, we examined whether the cardioprotective effect of EMPA on cardiac dysfunction was Sesn2 dependent. Sesn2 KO mice were fed an HFD for 12 weeks and then treated with EMPA for another 8 weeks. Although Sesn2 KO mice were more prone to developing obesity and metabolic disorders compared with WT littermates, EMPA could not attenuate these adverse alterations in the mutants. Sesn2 KO mice displayed not only diastolic dysfunction but also adverse systolic function, with reduced LVEF and LVFS. The results indicated that Sesn2 is a critical protein that mediates the beneficial effects of EMPA on obesity-related cardiac dysfunction. HFD decreases Sesn2 levels in the C57BL/6J mouse hearts and consequently decreases downstream AMPK phosphorylation that modulates mTOR activation (11,25). Although the isolated cardiomyocytes from HFD hearts did not show significant alterations in Sesn2 levels, the downstream AMPK phosphorylation was downregulated in the HFD cardiomyocytes. These observations indicate two potential underlying mechanisms of observed cardiac dysfunction in HFD mice. First, the Sesn2-AMPK signaling cascade, rather than the levels of Sesn2 protein alone, plays a role in the HFD-induced diastolic function. Second, a reduced expression of Sesn2 in noncardiomyocytes, such as endothelial cells, contributes to the observed reduction of Sesn2 in HFD hearts. A reduced Sesn2 in endothelial cells could affect cardiac diastolic function by regulating the nitric oxide pathway (34,35). In agreement, we found that EMPA treatment modulates the phosphorylation of eNOS at Ser¹¹⁷⁷ in the mouse hearts with HFD, indicating the endothelial-protective effects of EMPA. Future studies are needed to explore the difference of Sesn2 signaling response between cardiomyocytes and noncardiomyocytes in HFD heart.

mTOR is a central regulator of biogenesis that responds to the nutritional status of the cell. In obesity, activated mTOR can phosphorylate p70 S6 kinase, which is involved in myocardial proliferation and hypertrophy (36,37). AMPK inhibits protein synthesis and myocardial proliferation by inhibiting the mTOR-p70S6 kinase pathway (36). Sesn2 has two main functions: modulating AMPK-mTOR and antioxidative capability. We first detected whether the upregulation of Sesn2 by EMPA regulated the AMPK-mTOR pathway and found that EMPA treatment significantly enhanced AMPK phosphorylation and reduced phosphorylation of mTOR and p70 S6 kinase, indicating that EMPA could alleviate myocardial hypertrophy through the Sesn2-AMPK-mTOR pathway in HFD mice.

In obesity, there is increased FFA-induced myocardial lipid accumulation, which can increase the risk of heart failure in obesity by the imbalance of oxidative/antioxidant and inflammatory cell infiltration (38). Nrf2/HO-1 is the primary endogenous antioxidant defense pathway and counteracts oxidative stress-related cardiac injury, contributes to myocardia hypertrophy, and participates in protecting cardiac function (39,40). Our results showed that EMPA treatment inactivated NF-κB p65 activity and MCP-1 protein, as well as effectively enhancing oxidative signaling proteins, including the Nrf2/HO-1 system and its downstream antioxidant genes catalase and GCLM. Bae et al. (41) reported that Sesn2 induces Nrf2 activation by enhancing the autophagic degradation of Keap1. We also found that EMPA activated Nrf2/HO-1 by Sesn2, because treatment of KO-HFD-E did not upregulate Nrf2 mRNA or HO-1 protein, nor did it promote NF-κB activation. These findings indicated that Sesn2 is an essential factor for the EMPA-induced Nrf2/HO-1 antioxidative system. EMPA treatment activates the AMPK-signaling pathway, which can phosphorylate downstream ACC. The ACC phosphorylation leads to an augmented fatty acid oxidation that occurred in the HFD hearts. It could be a reason that EMPA treatment decreased lipid content in the HFD hearts.

The endothelial dysfunction induced by eNOS uncoupling initiates atherosclerosis, which is the leading cause of cardiovascular disease and often observed in obesity. We found that EMPA treatment triggered eNOS phosphorylation at Ser¹¹⁷⁷, indicating endothelial-protective effects of EMPA, which is consistent with a previous study (42). The PI3K/Akt pathway is also involved in cell proliferation, and activation of this pathway stimulates cell cycle progression and cardiac remodeling via the mTOR pathway. Akt activation in myocardial cells leads to hypertrophy and cardiac dysfunction in HFD mice with heart failure (43). We found that EMPA treatment inhibited phosphorylation of PI3K/Akt, indicating that EMPA could also inhibit the cardiac hypertrophy caused by the PI3K/Akt/mTOR pathway. We reported that EMPA triggers AMPK phosphorylation through AMPK upstream LKB1 activation (44). We also revealed that Sesn2 serves as a scaffold protein to modulate AMPK activation via recruiting AMPK upstream LKB1 to the Sesn2-AMPK complex (13). Therefore, Sesn2 KO limits LKB1 access to AMPK to phosphorylate AMPK by EMPA treatment.

The absence of data on the inhibition of Sesn2 by siRNA or overexpress Sesn2 in cardiomyocytes is a limitation of our study. In this regard, we used Sesn2 KO-HFD mice and treated them with EMPA, which better reflects the beneficial effects of EMPA for obesity-related cardiac dysfunction in vivo. Therefore, it is possible that the notable improvements in cardiac structure and function observed in our study could be due to Sesn2-mediated metabolic alterations. Additionally, some studies found that Sesn2 activation is dependent on hematopoietic progenitor kinase-germinal cell kinase-like kinase or protein kinase-like ER kinase (45,46). Further studies are needed to determine whether other SGLT2i can protect the heart via Sesn2 and whether EMPA activates Sesn2 through hematopoietic progenitor kinase-germinal cell kinase-like kinase or other signaling pathways.

This work clarified the critical functions of Sesn2 in obesity-related cardiac dysfunction and revealed that the molecular mechanism underlying the cardioprotective effects of EMPA on obesity-induced cardiac dysfunction was through regulating Sesn2-mediated AMPK-mTOR signaling and maintaining redox homeostasis. These promising findings will have an important positive impact in the field of heart failure research, as they provide potential therapeutic approaches for obesity-related cardiac dysfunction.

Acknowledgments. The authors thank Haiyan Zhang for measurement of the ELISA kit, Glenn Hoskins for transmission electron microscopy analysis, and Joshua Jefferson and Jesus Monico for tissue preparation and specific staining (all from University of Mississippi Medical Center).

Funding. The current study was supported by the National Natural Science Foundation of China (81870593), the American Diabetes Association (1-17-IBS-296), and the Overseas Study Program from the Government in Shandong Province.

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

Author Contributions. J.L. conceived and designed the study and participated in the revision and final approval of the manuscript. X.S. and F.H. participated in the study design, data collection, and writing of the draft manuscript. Q.L., X.L., D.R., J.Z., Y.H., and Y.K.X. participated in the partial data collection. All authors read and approved the final manuscript. J.L. is the guarantor of this work and, as such, had full access to all of 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 this study were presented in oral form at the 55th Annual Meeting of the European Association for the Study of Diabetes, Barcelona, Spain, 16–20 September 2019.