Glucagon‐like peptide‐1 ameliorates cardiac lipotoxicity in diabetic cardiomyopathy via the PPARα pathway

Summary Lipotoxicity cardiomyopathy is the result of excessive accumulation and oxidation of toxic lipids in the heart. It is a major threat to patients with diabetes. Glucagon‐like peptide‐1 (GLP‐1) has aroused considerable interest as a novel therapeutic target for diabetes mellitus because it stimulates insulin secretion. Here, we investigated the effects and mechanisms of the GLP‐1 analog exendin‐4 and the dipeptidyl peptidase‐4 inhibitor saxagliptin on cardiac lipid metabolism in diabetic mice (DM). The increased myocardial lipid accumulation, oxidative stress, apoptosis, and cardiac remodeling and dysfunction induced in DM by low streptozotocin doses and high‐fat diets were significantly reversed by exendin‐4 and saxagliptin treatments for 8 weeks. We found that exendin‐4 inhibited abnormal activation of the (PPARα)‐CD36 pathway by stimulating protein kinase A (PKA) but suppressing the Rho‐associated protein kinase (ROCK) pathway in DM hearts, palmitic acid (PA)‐treated rat h9c2 cardiomyocytes (CMs), and isolated adult mouse CMs. Cardioprotection in DM mediated by exendin‐4 was abolished by combination therapy with the PPARα agonist wy‐14643 but mimicked by PPARα gene deficiency. Therefore, the PPARα pathway accounted for the effects of exendin‐4. This conclusion was confirmed in cardiac‐restricted overexpression of PPARα mediated by adeno‐associated virus serotype‐9 containing a cardiac troponin T promoter. Our results provide the first direct evidence that GLP‐1 protects cardiac function by inhibiting the ROCK/PPARα pathway, thereby ameliorating lipotoxicity in diabetic cardiomyopathy.

properties (Liu et al., 2001). Considerable progress has been made in DCM management. Nevertheless, the molecular etiologies of DCM remain poorly understood, and currently available therapies are far from ideal. Therefore, further research in this area is urgently required.
It has been confirmed that disturbances in cardiac substrate metabolism and energetics are the key contributors to DCM (Anderson et al., 2009;Lopaschuk, Folmes & Stanley, 2007). In diabetes, cardiac palmitate oxidation doubles and glucose oxidation decreases by 30%-40% relative to the levels observed in nondiabetic patients (Anderson et al., 2009;Rijzewijk et al., 2009).
Although the switching of substrate utilization may meet the energy demand for heart function maintenance, it also brings many deleterious consequences (Rodrigues, Cam & McNeill, 1995;Stanley, Lopaschuk & McCormack, 1997). Increased fatty acid (FA) oxidation along with reduced ATP/O ratios decreases cardiac efficiency and contributes to ventricular dysfunction by increasing the generation of reactive oxygen species (ROS) and toxic lipid intermediates (Battiprolu et al., 2013;Houstis, Rosen & Lander, 2006). ROS damage DNA, mitochondria, and other cellular components by oxidizing proteins, converting lipids into reactive lipid peroxides, and increasing protein tyrosine nitration (Boudina et al., 2005). Lipid metabolite accumulation in cardiomyocytes (CMs) results in lipotoxicity and apoptosis (Drosatos & Schulze, 2013;van de Weijer, Schrauwen-Hinderling & Schrauwen, 2011). Therefore, inhibition of FA accumulation and oxidation has become important therapeutic strategies in DCM management.
Peroxisome proliferator-activated receptor alpha (PPARa) plays an important role in myocardial substrate metabolism by regulating the transcription of genes involved in FA transport, esterification, and oxidation (Banke et al., 2010;Gilde et al., 2003). Increases in FA oxidation and uptake in diabetic hearts were significantly reduced in PPARa À/À mice (Campbell et al., 2002). Cardiacrestricted PPARa overexpression (MHC-PPARa) in mice mimicked the DCM phenotype. These animals were relatively more susceptible to serious cardiomyopathy in response to high-fat diets (HFD) or streptozotocin (STZ) stimulation and presented with significant increases in lipids accumulation (Finck et al., 2002(Finck et al., , 2003Yang et al., 2007). Therefore, PPARa activation-induced metabolic abnormalities in diabetic hearts may be promising as therapeutic DCM targets.
Native glucagon-like peptide-1 (GLP-1) is a hormone produced by the L-cells of the distal ileum and colon in response to the entry of nutrients and destroyed by the circulating dipeptidyl peptidase-4 (DPP-4) (Ban et al., 2008). In the past decade, GLP-1 and its analogs have been introduced as a new class of antidiabetic medications for their pleiotropic effects, including increasing glucose-dependent insulin secretion, suppressing glucagon secretion, decreasing appetite, and reducing body weight (Park, Lim, Lee & Na, 2016). A functional GLP-1 receptor (GLP-1R) is highly expressed in the heart. GLP-1R agonists and DPP-4 inhibitors have beneficial effects on the cardiovascular system (Ban et al., 2008;Noyan-Ashraf et al., 2009;Timmers et al., 2009). Previous studies have shown that GLP-1 and its analogs protected the heart against ischemia-reperfusion injury and diabetes mellitus (Tate, Robinson, Green, McDermott & Grieve, 2016;Wang et al., 2013). They also protected isolated CMs from oxidative damage (Chang et al., 2013) and high-glucose stress (Younce, Burmeister & Ayala, 2013). However, the protective effects of GLP-1 exendin-4 (Ex-4) on lipid metabolism in diabetic hearts and the relationship between GLP-1R activation and the PPARa pathway have not yet been elucidated.
In this study, we investigated the effects of the GLP-1R agonist exendin-4 and the DPP-4 inhibitor saxagliptin on DCM induced by a HFD and STZ injections. We demonstrated that the cardioprotective effects of GLP-1, including reductions in lipid accumulation and potentiation of antioxidant and anti-apoptosis properties, may be driven by a PPARa-mediated mechanism.
2 | RESULTS 2.1 | Exendin-4 and saxagliptin reversed symptoms in diabetic mice (DM) Type 2 diabetes was induced in mice with low-dose STZ injections and continuous HFD, exhibiting hyperglycemia, body weight gain, and glucose intolerance compared with normal diet (ND) group . Administration of the GLP-1R analog Ex-4 and the DPP-4 inhibitor saxagliptin for 8 weeks significantly decreased plasma glucose levels and improved glucose tolerance compared with the diabetic mice (DM) group (Figure 1b,c). Saxagliptin provided slower glycemic control than Ex-4 and insulin but all three groups eventually reached similar glucose levels. Ex-4 reduced body weight compared with the untreated DM group, but saxagliptin did not ( Figure 1d). Both hyperglycemia and dyslipidemia impair heart function in diabetes. To exclude the hypoglycemic capacity when we assessed the cardioprotective effects of Ex-4 and saxagliptin, we used a 1.5 U/day insulin treatment as a control, as previously reported . As expected, insulin treatment had the same effects on blood glucose, glucose tolerance, and body weight as exendin-4 and saxagliptin in DM (Figure 1b-d).
2.2 | Exendin-4 and saxagliptin treatments attenuated cardiac remodeling and improved cardiac function in diabetic mice As shown in Figure 1e, the hearts in the DM were significantly larger than those in the ND mice. In addition, the heart weight to tibial length (HW/TL) ratios ( Figure 1e) and atrial natriuretic polypeptide expression ( Figure 1i) were higher in the DM than the ND mice.
Masson's trichrome stain displayed higher fibrotic areas (Figure 1f Table S1. Relative to the ND mice, the DM were characterized by significant decreases in E/A ratio, ejection fraction F I G U R E 1 Exendin-4 and saxagliptin improved metabolic characteristics and cardiac dysfunction in diabetic mice induced by high-fat diet and low-dose STZ injection. (a) Schematic of animal experimentation in vivo. Briefly, mice were fed with 60% high-fat diet (HFD) for 12 weeks, followed by 50 mg/kg STZ treatment for 5 days. At 14 week, experimental diabetes mice divided into different groups were treated with exendin-4, saxagliptin, or insulin, respectively, for 8 weeks. (b) After an 8-hr fast, serial tail blood glucose was measured before and after glucose administration (1 g/kg, intraperitoneal injection). *p < .05 vs. ND, # p < .05 DM vs. DM + treatments (exendin-4, saxagliptin or insulin).  (EF), fractional shortening (FS), maximal slope of systolic pressure increment (dP/dt max ), minimal slope of diastolic pressure decrement (dP/dt min ), and an increase in the left ventricular posterior wall thickness at diastole (LVIDd). All the aforementioned parameters were improved after the Ex-4 and saxagliptin treatments. However, insulin failed to produce therapeutic effects on diabetic myocardial remodeling and dysfunction (Figure 1e-i). Therefore, the cardioprotection conferred by GLP-1 may not depend on the ability of this agent to induce hypoglycemia. Taken together, these data showed that it was the activation of the GLP-1 receptor with Ex-4 or the elevation of endogenous GLP-1 with saxagliptin rather than insulin that alleviated cardiac injury in DM.

| GLP-1 suppressed the diabetes-related activation of Rho kinase and PPARa in vivo and in vitro
The Rho/Rho-associated kinase (ROCK) pathway may play important roles in oxidative stress and apoptosis and could be associated with complications of diabetes (Liu, Tan, Lai, Li & Wang, 2016;Zhou & Li, 2012). We examined the effects of GLP-1 on ROCK activation. As Peroxisome proliferator-activated receptor alpha (PPARa) is a key regulator of cardiac lipid metabolism. It alone drives the pathologic changes and functional abnormalities in diabetic hearts (Finck et al., 2002(Finck et al., , 2003. Therefore, the effects of GLP-1 on PPARa expression in diabetes were explored. The results showed that GLP-  2.10 | Exendin-4 reduced myocardial lipid accumulation, oxidative stress, and apoptosis via a PPARa-mediated mechanism in vivo As shown in Figure 5a, the combination of Ex-4 and wy-14643 elevated myocardial lipid accumulation more than the Ex-4 treatment alone in WT DM. PPARa deficiency mimicked the effects of Ex-4. In addition, Ex-4 treatment of PPARa KO in DM did not inhibit lipid accumulation relative to the WT DM + Ex-4 group. Therefore, Ex-4 attenuated diabetes-induced lipid metabolic disorder mainly in a PPARa-dependent manner (Figure 5a). Expression of the long-chain FA transporter CD36 significantly decreased when the PPARa protein was absent in mice (Figure 5d). Myocardial oxidative stress and apoptosis displayed the same trend upon evaluation by staining and Western blotting (Figure 5b,c,e,f). These data suggest that exendin-4 decreases myocardial lipid accumulation, oxidative stress, and apoptosis via a PPARa-mediated mechanism.
2.11 | Cardiac-specific PPARa overexpression induced by the adeno-associated virus serotype-9 (rAAV-9) reversed the salutary effects of exendin-4 on diabetic cardiomyopathy Previous studies reported the beneficial effects of Ex-4 on cardiac microvascular endothelium and inflammatory cells. Therefore, we induced cardiac-specific PPARa overexpression in mice via rAAV-cTNT-PPARa virus treatment to determine whether our conclusions were based on CMs alone. As expected, the rAAV-cTNT-GFP and rAAV-cTNT-PPARa viruses significantly increased the expression of the genes they bore in mouse hearts ( Figure S5A). Both rAAV-GFP and rAAV-PPARa mice developed stable hyperglycemia, glucose intolerance (Figures 6b and S5B). These were corrected by Ex-4 treatment. However, no difference was observed between the Ex-4 treatment and control groups in terms of weight gain ( Figure S5C).
Cardiac-restricted PPARa overexpression failed to increase heart size and HW/TL relative to those of rAAV-GFP DM (Figure 6a). Nevertheless, there were substantial increases in myocardial fibrotic area, cardiac lipid accumulation, and triglyceride (TG) content (Figure 6c,d).
Cardiac cell apoptosis is the most frequently proposed mechanism of DCM progress (Ouyang, You & Xie, 2014). Apoptotic cell death is also regarded as a terminal junction of various molecular mechanisms.
It contributes to cardiac remodeling by destroying contractile units and inducing compensatory myocardial cell hypertrophy and reparative fibrosis (Kusminski, Shetty, Orci, Unger & Scherer, 2009). The rate of CM apoptosis in patients with diabetes is 85-fold greater than that in nondiabetics (Ho, Liu, Liau, Huang & Lin-Shiau, 2000). Diabetesinduced CM apoptosis has been associated with excessive generation of reactive free radicals even though other inductive pathways exist as well (Dorn, 2009;Robertson et al., 2004). Increased ROS production and reduced antioxidant levels in diabetes have been widely documented in previous reports (Fiordaliso et al., 2004;Houstis et al., 2006). Earlier studies showed that CMs incubated with GLP-1 or its analogs remained viable and lowered ROS levels and apoptosis rates in both diabetic and nondiabetic models (Inoue et al., 2015;Raab, Vuguin, Stoffers & Simmons, 2009;XiaoTian et al., 2016). Nevertheless, these reports failed to address the possible mechanisms responsible for these effects. In the present study, we observed decreases in CM apoptosis and oxidative stress in the presence of exendin-4 or saxagliptin and elucidated their modes of action.
Some studies have attributed the benefits of Ex-4 on DCM to its effects on infiltrating macrophages, cardiac microvascular injury, and mitochondrial dysfunction (Tate et al., 2016;Wang et al., 2013;Wassef et al., 2017). However, the present study mainly focused on the effects of GLP-1 on lipid regulation because, along with hyperglycemia, lipid accumulation and toxicity play key roles in DCM (Kusminski et al., 2009;Yang et al., 2007). The lack of glycemic control in cardiovascular disease progress in obese and T2DM patients underscores the importance of lowering cardiac steatosis in them.
Excessive epicardial fat accumulation closely linked to cardiometabolic disruptions and mortality in T2DM patients through secretion of lipids, adipokines, and pro inflammatory and oxidative factors (Fitzgibbons & Czech, 2014;Gonzalez, Moreno-Villegas, Gonzalez-Bris, Egido & Lorenzo, 2017). Unoxidized FA accumulation in cardiac myocytes impairs energy metabolism and aggravates mitochondrial dysfunction, ROS overproduction, and lipoapoptosis (Drosatos & Schulze, 2013;Rodrigues et al., 1995). Several recent studies concluded that strategies to minimize ectopic fat accumulation and lipotoxicity have direct cardioprotective effects (Mori et al., 2014;Yang et al., 2007). To the best of our knowledge, the present study is the first to revealed that both exendin-4 and saxagliptin significantly reduced lipid content in CMs both in vivo and in vitro by controlling the PPARa-CD36 pathway, which is a major regulatory signal in cardiac FA metabolism. The long-chain FA transporter CD36 is responsible for >60% of the cardiac FA uptake (Angin et al., 2012). Heart-specific CD36 deficiency prevents myocardial lipid accumulation and rescues cardiac dysfunction. Therefore, CD36 may be a key therapeutic target for DCM (Yang et al., 2007). In the present study, we report for the first time that the restricted expression of myocardial CD36 was associated with the cardiac benefits of exendin-4 treatment.
PPARa had been widely accepted as a transcriptional switch for various genes involved in cardiac FA uptake and oxidation. PPARa may hasten the progress of DCM (Finck et al., 2002(Finck et al., , 2003. In the present study, PPARa KO mice failed to develop DCM. Cardiac-specific PPARa overexpression showed more severe DCM. WT DM mice receiving wy-14643 treatment alone presented with higher myocardial lipid levels and fibrosis severity than WT DM. Nevertheless, no significant differences were observed between the two groups in terms of cardiac function, oxidative stress, or apoptosis. It is possible that wy-14643 alone does not induce cardiac damage as severe as that caused by cardiac-specific PPARa overexpression. On the other hand, wy-14643 could reverse the PPARa inhibition promoted by exendin-4. The present study demonstrates that exendin-4 suppresses the PPARa expression and nuclear translocation induced by diabetes mellitus. These are the key mechanisms explaining the lipid-lowering property and cardioprotective effect of Ex-4. In contrast, contradictory conclusions about the role of PPARa on DM have also been documented (Baraka & AbdelGawad, 2010;Young et al., 2001). CMs chronically exposed to FA showed relatively lower PPARa expression and treated with the PPAR agonist fenofibrate showed suppression of PA-induced apoptosis (Young et al., 2001).
We propose that the model and treatment style differences among these studies account for these discrepancies.
Our data indicate that Ex-4 inhibited the ROCK/PPARa/CD36 pathway by PKA activation. The RhoA/ROCK pathway is a key mediator of oxidative stress-mediated cell injury (Liu et al., 2016;Zhou & Li, 2012). Our previous study demonstrated that the RhoA/ Rock pathway is strongly activated in patients with diabetes (Liu et al., 2016). Moreover, there is evidence that the RhoA/ROCK pathway contributes to DM pathogenesis both in vitro and in vivo (Furukawa et al., 2005;Wang et al., 2013). GLP-1 may attenuate the oxidative stress induced in cardiac microvascular endothelial cells by high glucose via the activation of cAMP/PKA and the inhibition of downstream ROCK activity . However, these results do not explain the beneficial effects of exendin-4 on cardiac lipotoxicity. In the current study, our data clearly linked the GLP-1/ PKA/ROCK regulatory axis and the PPARa/CD36 lipid metabolic signal.
In conclusion, we demonstrate that the GLP-1 analog exendin-4 improved the structural and functional abnormalities of diabetic hearts at least in part by inhibiting the PPARa-mediated lipid accu- Institutes of Health (NIH Publication No. 85-23, revised 1985 (Finck et al., 2002). The DM were randomly divided into four groups: (i) DM; (ii) DM subcutaneously injected with exendin-4 (Ex-4) at 100 lg kg À1 day À1 ; (iii) DM orally treated with saxagliptin (Saxa) at 10 mg kg À1 day À1 ; (iv) DM treated with 1.5 U insulin.

| Experiment 2
The PPARa KO on a C57Bl/6J background was purchased from Jackson Labs (Bar Harbor, ME, USA). Purebred wild-type littermate mice were used. Diabetes was induced as described in Experiment 1.

| Statistical analysis
All data are presented as means AE standard error of the means unless otherwise stated. Blood glucose and body weight at the 22nd week, OGTT at 120 min, heart weight:tibia length, fibrotic area, cardiac function parameters, Western blot densitometry, real-time polymerase chain reaction data, and fluorescence intensity in the first animal and most cell experiments were analyzed by one-way ANOVA. The Student-Newman-Keuls post hoc test was used to evaluate differences between groups. Data from the second and third animal experiments and from the adult mouse CM experiments shown in Figure S3E-H were analyzed by two-way ANOVA. The Tukey's post hoc test was used to evaluate differences between groups. p < .05 was considered statistically significant. All statistical tests were performed using GraphPad Prism v. 5.0 (GraphPad Software, San Diego, CA, USA) and SPSS v. 18.0 (IBM Corp., Armonk, NY, USA).
Other details of the experimental procedures are available in the Supporting information.

ACKNOWLEDG MENTS
We thank Dr. Wang's group for their technical assistance and constructive feedback during the course of this investigation. This work was supported by grants from National Nature Science Foundation

CONFLI CT OF INTEREST
The authors declare no potential conflict of interests relevant to this article.