Heat Shock Prevents Insulin Resistance–Induced Vascular Complications by Augmenting Angiotensin-(1-7) Signaling

Insulin resistance is a prominent component of metabolic syndrome, which is characterized by obesity, hypertension, and atherosclerosis. A reciprocal relationship has been established between insulin resistance and endothelial dysfunction, which may lead to cardiovascular disorders (1). Apart from the metabolic actions, insulin performs various important hemodynamic functions such as peripheral vasodilation and increased regional blood flow via stimulation of nitric oxide (NO). Therefore, anything that impairs insulin action is expected to result in endothelial dysfunction and vice versa (2). Insulin has been shown to induce endothelial-dependent vasodilation without affecting endothelium-independent vasodilation. We and others (3,4) have reported that insulin resistance impairs endothelium-dependent vasodilation.

Several studies have pointed out active involvement of renin-angiotensin system (RAS) in cardiovascular disorders and insulin resistance (5,6). RAS further splits in two, as follows: angiotensin (ANG) II/ACE1 and ANG-(1-7)/ACE2 axis. ANG-(1-7), acting through receptor Mas (G-protein–coupled), counter-regulates the actions of ANG II. Mas receptor dimerizes with AT₁ receptor and inhibits ANG II signaling, suggesting direct interaction of ANG II and the ANG-(1-7) axis (7). ANG-(1-7) promotes vasodilation and inhibits cell proliferation thrombosis, hence opposing the effects of ANG II (8,9). In endothelial cells, ANG-(1-7) inhibited ANG II–induced c-Src, extracellular signal–related kinase 1/2, and NADPH oxidase by activating SHP-2 phosphatase (10). ANG-(1-7) also inhibits the negative regulation of ANG II on insulin-induced AKT and NO production, suggesting the crucial role of ANG-(1-7) in maintaining insulin sensitivity (11). ANG-(1-7) and Mas agonist AVE0991 have shown beneficial effects in atherosclerosis and hypertension (12,13). Long-term treatment with ANG-(1-7) improved glucose uptake and insulin resistance in rats. ANG-(1-7) has also been found to be effective in mitigating diabetes complications like nephropathy and memory impairment (14,15). ANG-(1-7) is produced mostly from ANG II, and this reaction is catalyzed by ACE2. Overexpression of ACE2 or ACE1 inhibition with captopril is an effective way to increase ANG-(1-7) expression in tissues and circulation. ACE2 overexpression ameliorated the formation of atherosclerotic plaques, cardiac hypertrophy, and glomerulosclerosis, and this effect, in part, is mediated through ANG-(1-7) (16). Therefore, an imbalance between the ANG II/ACE1 axis and the ANG-(1-7)/ACE2 axis in the RAS has emerged as a common denominator in cardiovascular disorders (17).

It has been reported that endothelial NO synthase (eNOS) protein expression and NO release are decreased in clinically relevant human atherosclerosis (18). It has also been reported that loss in the expression of eNOS accelerated atheroma formation (19,20). Recently, Sansbury et al. (21) provided striking evidence that eNOS overexpression reduces adiposity in rodents fed with a high-fat diet (HFD). They have reported that eNOS activation in adipose and conduit or resistant vessels has profound clinical significance. As far as the stress response is concerned, calorie restriction or exercise training has also been shown to increase eNOS expression and NO bioavailability, which are implicated in mitochondrial biogenesis (22). Therefore, activated eNOS not only improves endothelial dysfunction but also alters processes such as adipocyte phenotype and mitochondrial biogenesis. Hence, eNOS appears to be much wiser than we thought. Hence, we hypothesized that in our model of insulin resistance eNOS expression might be altered.

Over the years, from the discovery of heat shock (HS) response, several studies pointed out the protective role of HS protein (HSP) 72 in various disorders like cardiac ischemia, Huntington disease, hypertension, and diabetes (23–25). In the study by Bruce et al. (26), correlation between decreased HSP72 expression and the degree of insulin resistance has been demonstrated. Further, diabetes has been shown to limit HSP72 expression in various tissues (25). These studies suggest that alteration in HSP expression is involved in the pathogenesis of insulin resistance. Induction of HSP72 by HS, transfection, or mild electrical stimulation prevented the insulin resistance in skeletal muscle and recently in monocytes (27–29). Several hypotheses for the beneficial effects of HSPs, such as inhibition of stress-activated kinases (Jun NH₂-terminal kinase), reduction in oxidative stress, and nuclear factor-κβ (NF-κβ) activation have been proposed, but the mechanism is still poorly understood (29).

Epigenetics includes the stable alteration in DNA and histone proteins responsible for cell- and tissue-specific gene expression. DNA methylation is perhaps one of the most widely studied epigenetic phenomenon (30). DNA methylation, at CpG dinucleotides, has been implicated in a number of processes, such as transcriptional regulation, chromatin structure, and cancer pathogenesis. A growing body of evidence shows an inverse correlation between transcription and DNA methylation, and is marked as a gene-silencing signal (31). DNA methyltransferase (DNMT) 1 is an abundantly expressed DNMT that is essential for the maintenance of methylation patterns and the silencing of tumor suppressor genes (32). DNMT1 has been shown to be activated in proatherogenic conditions and to be responsible for the repression of genes like estrogen receptor alpha (ER-α) (33). But there are no reports depicting an association between HS and DNMT1 expression and its role in regulating the key molecules responsible for endothelial dysfunction in insulin resistance.

Chen et al. (23,34) for the first time reported the presence of cross talk between HSPs and RAS. Mild heat treatment (41°C) was able to decrease the ANG II–induced inflammation and hypertension. On the contrary, heat treatment at 45°C augmented ANG II–induced contraction and oxidative stress, suggesting the hormetic effect of HS (35). These studies determined the effect of HS on the vasoconstrictor arm, ANG II, of RAS. However, the vasoprotective arm of RAS, ANG-(1-7), has not been investigated. Therefore, we hypothesized that in vivo hormetic mild heat treatment may enhance the ANG-(1-7)/Mas/ACE2 axis and prevent endothelial dysfunction in insulin resistance.

The 6- to 8-week-old male Sprague-Dawley rats (160–180 g) were procured from the Central Animal Facility of the National Institute of Pharmaceutical Education and Research. They were maintained under standard environmental conditions (temperature 20 ± 10°C; humidity 50 ± 10%; and 12 h light/dark cycle) with food and water available ad libitum. Animals were fed for 12 weeks with a normal pellet diet or an HFD (58 kcal% fat) (4). Experiments were conducted 48 h after the last heat or sham treatment, and rats were fasted 12 h before undergoing experimental procedures. All protocols were approved by the Institutional Animal Ethics Committee of the National Institute of Pharmaceutical Education and Research (IAEC13/14).

ANG-(1-7), nicotinamide (NAM), PD123319, and l-N^G-nitro-l-arginine methyl ester (L-NAME) were procured from Sigma-Aldrich. A-779 was procured from Bachem. Antibodies against HSP72 (1:3,000), phosphorylated (p) AMP-activated protein kinase (AMPK) (1:3,000), ACE2 (1:1,000), eNOS (1:2,000), p-eNOS (1:2,000), DNMT1 (1:1,000), and tubulin (1:3,000) were obtained from Santa Cruz Biotechnology; and HSP27 (1:3,000) and sirtuin 1 (SIRT1) (1:1,000) were obtained from Upstate Chemicals. Losartan was a generous gift from Novartis India. Unless specified, all other reagents were from Sigma-Aldrich.

Animals were anesthetized with an intraperitoneal injection of ketamine/xylazine. Animals were wrapped in a homeothermic blanket (Harvard Apparatus). The blanket temperature was set at 45°C and the core body temperature was monitored using a rectal probe. The animal core temperature was maintained at 41°C for 20 min by opening and closing the blanket, as described before by Gupte et al. (29). For the sham treatment, rats were anesthetized, and their core temperature was maintained at 37°C. After the heat or sham treatment, the rats were administered 5 ml 0.9% saline i.p. to prevent dehydration.

Aortic ring preparation and isometric tension measurement studies were performed with Panlab (Harvard Apparatus) and PowerLab data acquisition system (AD Instruments), as described previously (4). In brief, ANG-(1-7)–induced relaxation was studied by precontracting aortic rings with 100 nmol/L phenylephrine (PE) or 40 mmol/L KCl, after which the concentration-response curve (CRC) to ANG-(1-7) (1 nmol/L to 1 μmol/L) was recorded. To study the mechanism of ANG-(1-7)–induced relaxation, the second CRC to ANG-(1-7) was constructed. A total of two CRCs to ANG-(1-7) were constructed from each preparation with an interval of 2 h and with washing once every 15 min. The tissue was incubated with the vehicle, L-NAME (100 µmol/L), A-779 (10 µmol/L), PD123319 (10 µmol/L), and losartan (10 µmol/L) for 30 min, after which a second cumulative relaxation-response curve to ANG-(1-7) was elicited. Preliminary studies were performed to ascertain that the first and second cumulative relaxation-response curves were identical in the absence of any modulators.

In the 12th week, animals were anesthetized and submitted to surgical procedure as soon as anesthesia was assured by loss of pedal and corneal reflexes. The jugular vein was cannulated with a PE catheter (AD Instruments), and in vivo stimulation of the aorta was performed by the injection of 200 μL of either saline or ANG-(1-7) (1,000 pmol/kg) into rats from each group; after 10 min, the aorta was isolated and frozen immediately for Western blotting. For selective antagonism of Mas receptor, A-779 (80 pmol/kg) was coadministered with ANG-(1-7) intravenously. In another independent study, NAM (200 mg/kg i.p.) was administered to HFD-fed rats before heat treatment to discover the role of SIRT1 (36). NAM was administered 2 h prior to heat treatment. Intra-arterial blood pressure was measured by a fluid-filled pressure transducer connected to a data acquisition system (AD Instruments) in controls, HFD-fed rats, and HFD-fed rats subjected to a single bout of HS.

After the last heat treatment, animals were anesthetized and given a glucose load of 2 g/kg i.p., and blood sampling was performed. The blood samples were taken at 0 (just before glucose administration), and 15, 30, 60, and 120 min after administration of the glucose load from all animals for estimation of glucose levels.

Protein isolation and Western blotting were performed as previously described (4). Briefly, frozen aortic tissues were homogenized in liquid nitrogen and transferred to lysis buffer. Protein samples were resolved using 7.5%, 10%, and 14% SDS-PAGE, depending on the molecular weight of the desired proteins. These were then transferred to polyvinylidene fluoride membranes (PALL) and were analyzed with primary antibodies. The antibodies used were as follows: HSP72, p-AMPK, ACE2, eNOS, p-eNOS, HSP27, and SIRT1. Tubulin was used as the loading control. The antigen-primary antibody complexes were incubated with horseradish peroxidase–conjugated secondary antibodies and visualized using an enhanced chemiluminescence system and Hyperfilm ECL (Amersham). Blots were scanned and analyzed using ImageJ software.

RNA was isolated from aorta using RNA extraction kit (Ambion). After reverse transcription with Superscript II (Invitrogen), real-time RT-PCR was performed on a Light Cycler 2 (Roche) using SYBR Master mix (Invitrogen) and the specific forward and reverse primers (Eurofins MWG Operon) (MAS: forward-ACGTCCCCAGACCAGTCA; reverse-TGAGGAGTTCTTGTGCTG). After amplification, a melting-curve analysis was performed to verify the specificity of the reaction. The 18S gene was used as an internal control, and results were determined by 2^-ΔΔCt expressed as the fold change over control rats.

Rats were anesthetized and aorta, liver, muscle, white adipose tissue (WAT), and brown adipose tissue (BAT) were removed and stored in 10% formal saline. Paraffin blocks were prepared after completing the routine processing. Sections (3–5 mm) were prepared from paraffin blocks and stained with hematoxylin-eosin. Sections of liver, muscle, WAT, and BAT were deparaffinized with xylene, followed by antigen retrieval by heating in citrate buffer (10 mmol/L). The following rabbit polyclonal primary antibodies were used: anti-HSP72 (1:125). Polyvalent biotinylated goat anti-rabbit secondary antibody and a streptavidin-horseradish peroxidase system (Vectastain) were used to amplify the signals followed by detection with diaminobenzidine as a chromogen. Slides were counterstained with hematoxylin, dehydrated with alcohols and xylene, and mounted in DPX (Sigma-Aldrich). Histological images were captured by a charged-coupled device camera attached to a microscope (Model BX 51; Olympus, Tokyo, Japan). The intensity of the spot was graded from 1 to 4 (1, slight or no color; 2, very low color; 3, moderate brown color; and 4, very intense brown color). The immunohistochemistry score was expressed as the mean ± SEM for each experimental group.

Results were expressed as the mean ± SEM, and n refers to the number of samples studied. Cumulative CRCs were analyzed by nonlinear curve fitting, and statistical differences between the means (maximal relaxation) were determined by one-way ANOVA using Prism software (version 5.0; GraphPad, San Diego, CA) for Windows. P < 0.05 was considered to be statistically significant.

To evaluate whether HSP72 deficiency is involved in endothelial dysfunction associated with insulin resistance, we determined the expression profile of HSP72 in arteries of control and HFD-fed rats. We show that HSP72 is predominantly expressed in the aorta of healthy control rats. Aortic segments from HFD-fed, insulin-resistant rats showed markedly reduced expression of HSP72, as confirmed by immunohistochemical analysis (Fig. 1A and B) (P < 0.01). Western blotting analysis revealed that the expression of HSP72 was higher in arteries of control rats than in those of HFD-fed rats (Fig. 1C and D) (P < 0.05). These observations suggest that HSP72 may have a beneficial role in insulin resistance and its complications.

Insulin resistance was developed in rats by feeding them an HFD for 12 weeks. From the first week of feeding, a subgroup of control and HFD-fed rats received mild heat treatment (i.e., HS; at 41°C once a week for 20 min), which is an established method for the induction of HSP72 (29). After the last HS treatment, animals were fasted overnight for the measurement of fasting glucose and insulin levels, and to perform an intraperitoneal glucose tolerance test (IPGTT). As expected, an HFD resulted in an elevation in body weight (Fig. 2A), fasting glucose (Fig. 2B) and insulin (Fig. 2C) levels, and insulin resistance, as analyzed by homeostatic model assessment for insulin resistance (HOMA-IR), compared with control rats (Fig. 2D). On the other hand, HFD-fed rats subjected to HS were protected from hyperglycemia, hyperinsulinemia, and elevated HOMA-IR. However, control rats subjected to HS showed no significant changes in fasting glucose and insulin levels, and HOMA-IR compared with control rats. Further, to check whether rats were protected against glucose load, we performed an IPGTT. The HFD-fed rats showed impaired glucose disposal, as observed by an increased area under the curve (AUC) than control rats (P < 0.01). However, HFD-fed rats subjected to HS showed improved glucose disposal and reduced AUC than HFD-fed rats. There was no change in the AUC in control rats and control rats subjected to heat treatment (Fig. 2E and F) (P < 0.05).

Previous studies have shown that mRNA and protein level of HSP72 are reduced in the skeletal muscle of patients with type 2 diabetes and insulin resistance (26,29). However, there are no reports regarding the inhibition of vascular HSP72 expression in insulin resistance. Here, we report that HSP72 expression is reduced significantly in the aortas of HFD-fed rats than in those of control rats (Fig. 2G–I) (P < 0.05). HS markedly restored the expression of HSP72 in HFD-fed rats (P < 0.05). We also determined the expression of HSP27. An HFD also decreased HSP27 expression to a certain extent, and HS slightly elevated it. In order to determine a correlation between HSP72 protein expression and fasting insulin levels, we performed regression analysis. Interestingly, we observed an inverse correlation between fasting insulin level and aortic HSP72 expression (Fig. 2J) (r² = 0.65, P < 0.01). This observation was matched with earlier reports (26,37). Further, HFD-fed rats subjected to heat treatment showed decreased plasma cholesterol and triglyceride levels than rats fed an HFD alone (Supplementary Fig. 1A and B). HFD feeding also promoted adiposity because the weights of the visceral, mesentery, epididymal, and retroperitoneal fat pads were increased. HFD-fed rats subjected to HS were protected from any increase in adiposity index (Supplementary Fig. 1C–G). Also, HS redirected the fat from nonadipose tissue to adipose tissue. The liver weight and liver mass index were significantly increased by HFD feeding, and HS suppressed these changes (Supplementary Fig. 1H–M). Histopathological analysis revealed excessive fat deposition in hepatocytes and skeletal muscles by HFD feeding, and its reversal by heat treatment (Supplementary Fig. 2A). HFD feeding increased the adipocyte size and cholesterol deposition in WAT and BAT. HS significantly reduced the adipocyte size of WAT and BAT in HFD-fed rats (Supplementary Fig. 2B).

It is generally accepted that the tonic release of NO from the endothelium exerts vasculoprotective and cardioprotective effects in response to exercise or calorie restriction (38,39). We and others (1,4) have reported that the activation of ANG II is responsible for endothelial dysfunction in insulin resistance. Earlier, HS has been shown to suppress ANG II–induced hypertension and inflammation, but there are no reports suggesting a beneficial role of HS on the ANG-(1-7) pathway of RAS (23). Therefore, we determined vasodilator responses to ANG-(1-7) in insulin resistance condition. ANG-(1-7) induced concentration-dependent vasodilation in thoracic aorta. ANG-(1-7)–induced 40 ± 6% of maximum relaxation in control aortic rings. HFD feeding reduced the ANG-(1-7)–induced maximal relaxation to 22 ± 5% (Fig. 3A) (P < 0.01). This indicates that the protective action of ANG-(1-7) is masked by HFD-induced insulin resistance. Contrary to that, ANG-(1-7)–induced vasodilation was augmented to 39 ± 3% in aortic rings isolated from rats subjected to HS than HFD-fed rats (Fig. 3A) (P < 0.01). This indicates that heat treatment protected HFD-fed rats against reduced ANG-(1-7)–induced vasodilatory effect. The vasodilatory responses to ANG-(1-7) were comparable among control rats and control rats subjected to HS rat aortic rings (Fig. 3A). We also compared the sensitivity (negative logarithm value [pD₂]) to ANG-(1-7) in control and HFD-fed rats subjected to either sham or heat treatment. The sensitivity was similar for all groups (Fig. 3B).

Further, we investigated the direct effect of HS on NO bioavailability and endothelial-dependent vasodilatation in insulin resistance. HS enhanced the vasodilation in response to carbachol (CCh) (Fig. 3C) (P < 0.05 vs. HFD), which was suppressed under HFD feeding (P < 0.001 vs. control). Vasodilation in response to CCh was endothelium-dependent, and therefore we assessed endothelium-independent vasodilation by using NO donor sodium nitroprusside (SNP). The relaxation responses to SNP were unaltered by either an HFD or HS and were comparable with control rings (Fig. 3D). This indicates an active involvement of endothelium in mediating the effects of HS. In order to determine NO bioavailability, we performed a Griess assay and a functional study with PE in the presence of L-NAME. HFD feeding resulted in decreased NO levels in the aorta compared with control rats, which corroborates our previous report (4). HS significantly elevated the NO levels obtained from the Griess method (Fig. 3E) (P < 0.05). We observed that the vascular reactivity to PE was increased in aortic rings obtained from rats subjected to heat treatment when preincubated with eNOS inhibitor L-NAME. This indicates that NO levels were elevated in aortas of rats subjected to HS (Fig. 3F) (P < 0.05). In the overall samples, the AUC for ANG-(1-7)–induced vasodilation was positively correlated to aortic HSP72 expression (Fig. 3G) (r² = 0.71, P < 0.01). Gross histopathology of the aorta was performed by staining with hematoxylin-eosin. We did not observe any appreciable changes in microscopic characters of the aorta among all the groups (Supplementary Fig. 3). We have also checked the expression of inflammatory molecules such as p-NF-κβ and inhibitor of κB in the aorta. Interestingly we found that HFD increases p-NF-κβ and decreases inhibitor of κB expression. HS treatment was able to attenuate this effect and hence HFD-induced inflammation (Supplementary Fig. 4C and D). These observations matched with those of previous studies (23,29).

The study by Sampaio et al. (40) shows that ANG-(1-7) binds to Mas receptor, activates eNOS (S1177) phosphorylation, and releases NO. The AKT/protein kinase B pathway is also shown to be involved in eNOS activation and NO release (40). We reported, earlier in RESULTS, on ex vivo ANG-(1-7) signaling in isolated aortic rings, and we also assessed the effect of HFD and HS on the in vivo ANG-(1-7) signaling.

First, we checked ANG-(1-7)–induced S1177 phosphorylation of eNOS as it has been correlated with NO release. ANG-(1-7) significantly increased the p-eNOS expression within 10 min in control rats (P < 0.05). However, HFD feeding prevented the ANG-(1-7)–induced p-eNOS expression (P < 0.01).The p-eNOS expression was fully rescued by HS treatment (Fig. 4A, panels a and b). Further, we have assessed both the basal and ANG-(1-7)–stimulated AKT activation. The basal expression of p-AKT S473 was comparable in control and HFD-fed rats but increased significantly in HS rats (P < 0.05). ANG-(1-7) induced p-AKT expression in control rats (P < 0.001), whereas this effect was inhibited in HFD-fed rats. Interestingly, HS treatment enhanced the ANG-(1-7)–induced p-AKT expression significantly (P < 0.01) (Fig. 4A, panels c and d). Next, we determined the effect of ANG-(1-7) on p-AMPK expression. We found that ANG-(1-7) acute stimulation increased p-AMPK expression and that HFD feeding ablated this effect. HS restored the ANG-(1-7)–induced AMPK phosphorylation (Fig. 4A, panels e and f) (P < 0.01). Further, we analyzed HSP72 and SIRT1 expression after the acute stimulation of ANG-(1-7). It was found that neither HSP72 nor SIRT1 expression was altered after short stimulation of ANG-(1-7) (Fig. 4A, panels g and h). This indicates that acute stimulation of ANG-(1-7) was not able to increase the expression of SIRT1 or HSP72. Next, we checked the change in blood pressure by one-time HS exposure for 20 min in 3-month-old HFD-fed rats. Three months of HFD feeding significantly increased the systolic blood pressure compared with control rats on a normal pellet diet (P < 0.01). Heat treatment of HFD-fed rats showed reduced systolic blood pressure (P < 0.01 vs. HFD). The heart rate was comparable among the groups (Fig. 4H and I).

ANG-(1-7) is produced mostly from ANG II by the enzyme ACE2, and its overexpression has been shown to increase the circulating and local levels of ANG-(1-7) (41). Accumulating evidence indicates that beneficial actions of ANG-(1-7) are mediated through the Mas receptor activation (40,41). Checking ANG-(1-7) receptor Mas expression by real-time PCR, we found that HFD feeding decreased the mRNA levels of Mas receptor gene (Fig. 5A) (P < 0.01). The rats subjected to HS treatment had higher expression of Mas receptor than HFD-fed rats (P < 0.01). We also determined the protein expression of ACE2 in all the groups. HFD feeding promoted a decrease in ACE2 expression, and HS treatment increased it (Fig. 5B and C) (P < 0.01).

There exists an inverse correlation between transcription and DNA methylation, which is marked as a gene-silencing signal. The expression of eNOS, SIRT1, and HSP72 was reduced in insulin resistance, which indicates that activation of silencing marks enzymes such as DNMT1. Therefore, we assessed the expression of the enzyme responsible for altering DNA methylation, DNMT1, in the insulin resistance state and whether HS has any effect on it.

We found that DNMT1 expression was markedly increased in the aortas of HFD-fed rats compared with those of controls (P < 0.01) (Fig. 5D and E). Interestingly, heat treatment ameliorated this increase in the expression of DNMT1 in HFD-fed rats to a significant extent (P < 0.01). This interesting observation suggests that insulin resistance is a state of enriched DNA methylation, which might be playing a role in the regulation of many vasoprotective molecules like eNOS, SIRT1, and HSP72.

We further sought to understand the mechanism of ANG-(1-7)–induced vasodilation in HFD-fed rats subjected to HS. To trace out the type of receptor involved, we incubated aortic rings with series of angiotensin receptor blockers for 30 min prior to visualizing ANG-(1-7) responses. Losartan (AT₁ receptor blocker) and PD123319 (AT₂ receptor blocker) did not affect ANG-(1-7)–induced relaxation responses in aortic rings isolated from rats subjected to HS. But A-779, a Mas receptor blocker, completely abolished the ANG-(1-7)–induced vasodilation (Fig. 6A) (P < 0.001). Further, we assessed ANG-(1-7)–induced eNOS and AKT phosphorylation in vivo, in the presence of A-779. ANG-(1-7) (1,000 pmol/kg)–induced eNOS (P < 0.01) and AKT phosphorylation (S473) (P < 0.01) were significantly inhibited by the administration of A-779 (80 pmol/kg) (Fig. 6B–D). This provides us with a plausible mechanism for ANG-(1-7)–induced vasodilation via Mas receptor.

Aortic rings obtained from HFD-fed rats subjected to HS were incubated either with or without L-NAME, an eNOS inhibitor. We observed that L-NAME promoted a decrease in ANG-(1-7)–induced relaxation responses in HFD+HS rats (Fig. 6E) (P < 0.01). ANG-(1-7) elicits some relaxation in the presence of L-NAME, which indicates that other factors, such as prostanoids or endothelium-dependent hyperpolarizing factor, are also involved in mediating the responses to ANG-(1-7) in the rat aorta.

Recent findings pointed out the role of SIRT1 in mediating HS response. SIRT1 deacetylates HS factor-1 (HSF-1) and promotes its occupancy of the HSP72 gene (36). The expression of SIRT1 was increased in rats subjected to HS. Hence, we became more interested in examining its role in mediating the effects of HS. We used NAM, which is an endogenous inhibitor of SIRT1 and acts through a salvage pathway (36).

NAM treatment indeed suppresses the induction of HSP72 in HFD-fed rats subjected to HS (Fig. 7A). It also prevented the increase in p-eNOS (S1177) level in HFD+HS rats (P < 0.01), which was very much comparable with the level in HFD-fed rats. Furthermore, the phosphorylation levels of AMPK and SIRT1 expression were also checked. As expected, HFD feeding reduced p-AMPK and SIRT1 expression compared with that in control rats, whereas HS ablated this effect. Treatment with NAM significantly suppressed the elevated p-AMPK expression in HFD+HS rats (P < 0.01). However, the expression of SIRT1 was not affected by NAM treatment in HFD+HS rats. This can be explained if we assume that NAM only inhibited the catalytic activity of SIRT1. Overall, all of these results show that p-AMPK, SIRT1, and p-eNOS are negatively correlated with insulin resistance (Fig. 7A–E).

At the functional level, NAM promoted a decrease in ANG-(1-7)–induced vasodilation in HS rats (Fig. 7F) (HFD+HS+NAM −21.6 ± 2.2; HFD+HS −35.1 ± 1.7; P < 0.05). This shows the direct involvement of SIRT1 in augmenting ANG-(1-7)–induced vasodilation. Likewise, NAM also suppressed the augmented CCh-induced vasodilation in aortic rings of HS rats (Fig. 7G) (HFD+HS+NAM −31.2 ± 2.2; HFD+HS −54.7 ± 2.4). Overall, NAM prevents the beneficial effects of HSP72 induction in HFD-fed rats, which highlights the presence of cross talk between SIRT1 and HSP72.

The data presented in this study provide compelling evidence that HSP72 (heat treatment) can be a potential target for treating endothelial dysfunction. We, for the first time, show that insulin resistance is a state of vascular HSP72 deficiency. HSP72, an inducible form of HSP, functions at the cellular level to protect cells against many stress conditions, such as temperature, hypoxia, and ischemia, and stabilizes the normal functions of proteins (42). In our study we show that vascular HSP72 expression is reduced in insulin resistance and is inversely correlated with circulating insulin levels. Our data support earlier findings demonstrating that there exists an inverse relationship between mRNA levels of HSP72 and insulin sensitivity in skeletal muscles of type 2 diabetic and obese patients (26,28,29). Previous studies have failed to address the mechanism of loss of HSP72 expression in insulin resistance (26,28). HSF-1 regulates the transcription of a large number of genes regulating protein homeostasis, including HSP72. Transient activation of HSF-1 by stress leads to conversion of the inactive monomer to DNA binding trimer (42). Westerheide et al. (36) have reported that SIRT1 deacetylates HSF-1 and enhances its affinity to DNA. This suggests that SIRTs are essential to induce HS response. In our study, we report that SIRT1 expression is reduced in insulin resistance, which correlates well with our observation of the reduced expression of HSP72 in insulin resistance. Apart from SIRT1, our data also highlight the role of AMPK in regulating HSF-1/HSP72 activity, either directly or indirectly.

Genome-wide promoter analysis of DNA methylation in patients with type 2 diabetes has suggested that hypermethylation of mitochondrial genes such as peroxisome proliferator–activated receptor γ coactivator-1-α promotes its silencing (43). Recently, a decrease in the expression of ER-α has been positively correlated with enhanced DNMT1 expression in the aorta of HFD-fed rats (29). Here we also report that DNMT1 expression is enhanced in the aortas of HFD-fed rats, and this increase was suppressed by HS. This interesting observation can explain the decreased expression of HSP72, SIRT1, and eNOS in insulin resistance, and its epigenetic regulation. However, additional studies are required to confirm that these genes are regulated by DNA methylation.

HS has been shown to inhibit ANG II–induced hypertension and inflammation in the heart and aorta (23,34). This suggests the presence of cross talk between HS and RAS. We now report for the first time that insulin resistance impairs the ANG-(1-7) vasodilator effect, and the induction of HSP72 attenuates this impairment. Previously, we have reported that HFD upregulates ANG II/AT₁ signaling, which leads to enhanced vasoconstriction in the aorta (4). It has been reported that the initial response in the ANG II/AT₁ axis downregulates the ANG-(1-7)/Mas axis (44,45). We now report that HFD impairs ANG-(1-7)/Mas signaling the in aorta, and HS prevents this loss. This suggests that enhanced ANG II/AT₁ signaling in the aortas of HFD-fed rats might be responsible for the downregulation of ANG-(1-7)/MAS/ACE2 axis and NO bioavailability. Overexpression of ACE2 or chronic ANG-(1-7) infusion has been shown to reduce blood pressure, and this effect may be due to the degrading of local ANG II and the improving of endothelial dysfunction (46,47). Previously, HS has been shown to suppress ANG II–induced hypertension in aorta (23). Together, the plausible explanation for the antihypertensive effect of HS may be the activation of ANG-(1-7) and the suppression of ANG II signaling in HFD-fed rats. The mechanism for the increase in vasodilator responses to ANG-(1-7) is probably due to enhanced downstream signaling of ANG-(1-7) as the sensitivity (pD₂ values) did not change. The molecular mechanisms of ANG-(1-7) were first demonstrated by Sampaio et al. (40) in endothelial cells. Previously, Santos et al. (48) discovered that the Mas receptor was a binding site of ANG-(1-7). Activation by ANG-(1-7) causes dimerization of Mas receptors and the release of NO via activation of the AKT/eNOS pathway (40). In our study, we report that HS augments ANG-(1-7) signaling (ex vivo and in vivo) through activation of the Mas/AKT/eNOS pathway as both Mas receptor inhibitor (A-779) and eNOS inhibitor (L-NAME) prevent the effects of HS.

We also observed reduced eNOS expression in HFD-fed rats, and HS prevented this effect. Increased DNA methylation marks on the promoter of eNOS have been shown in vascular smooth muscle cells, which do not express eNOS (31). Interestingly, we provide evidence that HFD-fed rats had increased levels of DNMT1 in the aorta and that HS could prevent this increase. This definitely points out that eNOS expression might also be downregulated by DNA methylation. However, further studies are required to warrant any conclusion.

The role of ANG-(1-7) has been studied in various disorders such as hypertension, insulin resistance, and atherosclerosis (16). Like ACE inhibitors or AT₁ receptor blockers, ANG-(1-7) attenuates the characteristics of metabolic syndrome (49). Oral formulation of ANG-(1-7) improves metabolism, and decreases the proinflammatory profile and fat deposition in HFD-fed mice as well as hypertensive rats (50,51). Prolonged ANG-(1-7) treatment has also been shown to improve endothelial dysfunction and oxidative stress in a mouse model of obesity (52). Recently, the therapeutic potential of ANG-(1-7) axis was highlighted by the development of an orally active nonpeptide Mas agonist, such as AVE 0991 and CGEN-865S. AVE 0991 showed beneficial effects in heart dysfunction, atherosclerosis, blood pressure, and pulmonary remodeling (53–56). These observations further confirm the protective role of ANG-(1-7) in different models of cardiovascular disorders.

Either Mas deficiency or Mas inhibition has been shown to ablate ANG-(1-7)–induced vasodilator responses or NO availability (40,48). Here, we report that mRNA levels of the Mas receptor gene decreases insulin resistance, and this may explain the impaired vasodilator responses to ANG-(1-7). Interestingly, HS treatment increases the expression of Mas receptor, and hence ANG-(1-7) signaling. This raises an interesting question of whether HSP72 interacts directly or indirectly with Mas receptor to regulate vascular tone. Additional studies are needed to elucidate this interaction. Enhancement in ANG-(1-7) signaling can also be explained by the observed increase in the expression of ACE2 by HS. ACE2 is a major enzyme for ANG-(1-7) formation. Studies involving the activation of ACE2 further confirm the role of ANG-(1-7) axis cardiovascular dysfunction. Either pharmacological or genetic activation of ACE2 improved endothelial migration, and impaired tube formation in renin transgenic mice and endothelial dysfunction in SHR rats (57,58). ACE2 activation has been reported to enhance the reparative function of endothelial progenitor cells in the diabetic condition (59). Further, ACE2 overexpression has been shown to prevent atherosclerosis, cardiac fibrosis, and hypertrophy, and these effects were in part mediated through upregulation of ANG-(1-7) (60,61).

Endothelial dysfunction is a hallmark of cardiovascular risk factors and has been linked to pathophysiology of hypertension. We and others have shown (1,4) an opposite correlation between NO bioavailability and endothelial dysfunction. Hypertensive subjects generated higher levels of free radicals, which quench NO and thereby reduce its bioavailability (62). Exercise or other nonpharmacological approaches such as endurance training is a systemic way to increase NO levels and treat hypertension (63). Like exercise, we also provide evidence that HS significantly increases NO levels and reduces systolic blood pressure in insulin-resistant rats.

SIRT1 (a silent mating type information regulator 2 homolog) is known to extend the life span and insulin sensitivity in calorie restriction (64). SIRT1 is also considered to be beneficial in endothelial biology. Previously, it was shown that overexpression of SIRT1 prevented oxidative stress–induced endothelial senescence and enhanced endothelial-dependent vasodilation by deacetylating eNOS (65,66). SIRT1 closely works with AMPK in the regulation of cell stress and energy homeostasis. The lipid metabolism in hepatocytes has been shown to be regulated by SIRT1 through AMPK (67). However, in skeletal muscles AMPK modulates SIRT1 by regulating the activity of enzymes (Nampt) in a salvage pathway (68). A recent report suggests that shear stress primes AMPK and SIRT1 to act synergistically to increase NO bioavailability in endothelial cells (69). In light of this evidence, we report that insulin resistance represents a state of vascular SIRT1 and p-AMPK deficiency, which may contribute to the endothelial dysfunction. The induction of HSP72 restores the expression of SIRT1 and p-AMPK, which is another novel finding of our study. Assuming that effects mediated by HSP72 induction are, in part, mediated by SIRT1, we used NAM to inhibit SIRT1. SIRT1 inhibition suppressed the ANG-(1-7)–induced relaxant responses and endothelial-dependent vasodilation. This was also associated with reduced p-eNOS, p-AMPK, and SIRT1 expression, indicating a direct role of SIRT1 in mediating the effects of HSP72.

In conclusion, we show that the elevation of HSP72 levels by HS treatment augments ANG-(1-7) signaling and protects against endothelial dysfunction in insulin resistance. This protection was tightly associated with enhanced ANG-(1-7)–induced eNOS phosphorylation. The increased DNMT1 expression in insulin resistance suggests an epigenetic regulation of eNOS and SIRT1 expression. We propose that the induction of HSP72 can be a realistic approach to amplify ANG-(1-7)/Mas/ACE2 signaling and to prevent hypertension (Fig. 8).

Acknowledgments. The authors thank Sandeep Kumar and Jasmine Kaur for insightful discussion of this manuscript.

Funding. This research was supported by the Department of Biotechnolgy, Government of India (via. sanction no. BT/ PR13221/MED/12/448/2009), National Institute of Pharmaceutical Education and Research, which provided necessary facilities and funding.

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

Author Contributions. P.A.K. designed and performed all the experiments, and wrote the manuscript. K.T. designed and supervised the experiments, and approved the final version of the manuscript. K.T. 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.