SIRT1/HSF1/HSP pathway is essential for exenatide-alleviated, lipid-induced hepatic endoplasmic reticulum stress

ANIMAL MODELS

Seven-week-old male C57BL/6J mice were purchased from the Model Animal Research Center of Nanjing University (Nanjing, China). After a 1-week acclimatization period, mice were randomly distributed into two initial groups fed with a chow diet (4% fat [wt/wt]; Guangdong Medical Laboratory Animal Center, Guangzhou, China) or a high-fat diet (HFD) (35.8% fat [wt/wt], D12331; Research Diets, New Brunswick, NJ). After 12-week diet intervention, each group of mice was then randomly subdivided to receive either exenatide (24 nmol/kg/d; Eli Lilly and Company, Indianapolis, IN) or normal saline as control by intraperitoneal injection during the light cycle with the respective diet feeding for 4 weeks. Body weight and daily food intake were monitored every 2 weeks. Glucose tolerance tests and insulin tolerance tests were carried out at the end of intervention as described in the Supporting Information.

The whole-body Sirt1 heterozygous knockout (Sirt1^+/–) mice in a C57BL/6J genetic background were from J. Ye's laboratory at the Pennington Biomedical Research Center (Louisiana State University, Baton Rouge, LA)23 and bred in the animal experimental center of our hospital. Eight-week-old male Sirt1^+/– mice and their wild-type (WT) littermates were both challenged with an HFD for 12 weeks and then randomly subjected to either exenatide or normal saline for 8 weeks in the same way as described above.

A Sirt1 knockdown mouse model was constructed in C57BL/6J mice by tail vein injection of lentivirus (Lv) expressing short hairpin RNA (shRNA) targeting Sirt1, which was designed and chemically synthesized by Shanghai GeneChem Co. Ltd. (Shanghai, China). The Lv vector expressing green fluorescence protein (GFP) only was used as the RNA inference (RNAi) control. Eight-week-old male C57BL/6J mice were randomly divided into two groups fed a chow diet or an HFD for 8 weeks. The HFD-fed mice were further divided into four groups: a normal saline–treated group (HFD), an exenatide-treated group (HFD+Exe), an Lv-GFP with exenatide–treated group (HFD+Exe+GFP), and an Lv-Sirt1 RNAi with exenatide–treated group (HFD+Exe+Sirt1 RNAi). One week after the injection of Lv (5 × 10⁷ transfection units) or control solvent, mice were subjected to either exenatide or normal saline for 2 weeks as described above.

All animals received humane care, and all study procedures were approved by the Institutional Animal Care and Use Committee of Sun Yat-Sen University. Serum sampling and tissue collection, liver histology and immunohistochemistry, assays of serum and liver lipids, as well as serum adiponectin are detailed in the Supporting Information.

CELL CULTURE AND TREATMENT

Human HepG2 hepatocytes were cultured in minimum essential medium supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate, 100 U/mL penicillin, and 100 μg/mL streptomycin. Cells at 70%-80% confluence were incubated in serum-free medium for 4 hours before treatment. Primary murine hepatocytes were isolated from 8-week-old male C57BL/6J mice using a two-step collagenase perfusion technique as described24 and cultured on type I collagen–coated plates in Williams E medium supplemented with 2 mM L-glutamine, 100 nM dexamethasone, and antibiotics described above.

Thapsigargin and palmitate (PA) (Sigma-Aldrich, St. Louis, MO) were used as ER stress inducers. To examine the effect of exendin-4 (Sigma-Aldrich) on hepatic ER stress in vitro, hepatocytes were incubated in medium containing ER stress inducers with or without 100 nM exendin-4 for 24 hours. A SIRT1 inhibitor, EX-527 (Sigma-Aldrich), as well as two SIRT1 agonists, resveratrol (Sigma-Aldrich) and SRT1720 (Selleck, Houston, TX), were also included to investigate the role of SIRT1 in the effect of exendin-4 on PA-induced ER stress. For Sirt1 knockdown in vitro, HepG2 cells were transfected with Lv Sirt1 shRNA or scrambled shRNA at a multiplicity of infection (=10/20) according to the manufacturer's instructions. Transfected cells were then incubated in medium containing PA with or without exendin-4 for 24 hours.

IMMUNOPRECIPITATION

Total protein (500 μg) was incubated with 10 μL specific antibody against HSF1 (Millipore, Billerica, MA) overnight at 4 °C with continuous mixing. Then Protein A/G Mix Magnetic Beads (Millipore) were added, followed by incubation for 2 hours at 4 °C. After washing three times, the immunocomplexes were resuspended with denaturing elution buffer, heated at 100 °C for 5 minutes, and separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, followed by immunoblotting with primary antibody against HSF1. For acetylation western blotting, 50 mM trishydroxymethylaminomethane (pH 7.5) with 10% (vol/vol) Tween-20 and 1% peptone was used for blocking, and 50 mM trishydroxymethylaminomethane (pH 7.5) with 0.1% peptone was used to prepare primary antibody against acetylated-lysine (Cell Signaling Technology, Danvers, MA) and secondary antibody.25

CHROMATIN IMMUNOPRECIPITATION ASSAY

Chromatin immunoprecipitation (ChIP) reactions were performed with an EZ-Magna ChIP HiSens kit (Millipore) according to the manufacturer's instructions. Briefly, chromatin samples isolated from hepatocytes were sonicated, followed by incubation with 5 μL anti-HSF1 antibody (Millipore) or with control immunoglobulin G and ChIP A/G magnetic beads overnight at 4 °C with continuous mixing. Immunoprecipitated DNA and input genomic DNA were amplified with primers targeting the Hsp70.1 promoter region containing heat shock element by real-time PCR as described above. Primers used for the Hsp70.1 promoter were (forward) 5′-TGTCCCCTCCAGTGAATCCCA-3′ and (reverse) 5′-TATTCCAGGGTTTTCGCCTCC-3′. Results were normalized to reactions performed with 100% input samples.

STATISTICAL ANALYSIS

Data are presented as mean ± standard error of the mean (SEM). The unpaired two-tailed Student t test was used to determine significant differences between groups. P < 0.05 was considered statistically significant.

Additional materials and methods are described in the Supporting Information.

EXENATIDE REDUCES BODY WEIGHT AND RESTORES SYSTEMIC AND HEPATIC GLUCOSE HOMEOSTASIS IN HFD-FED MICE

From the fourth week of intervention, the mice in the HFD group began to show a significant increase in body weight compared to the chow diet group, which lasted until the end of intervention (Fig. 1A). Interestingly, the HFD-induced mice subjected to 4-week exenatide treatment exhibited significant weight loss compared to the HFD group, especially in the first 2 weeks, which was accompanied by a decrease in daily food intake (Fig. 1A,B). Meanwhile, exenatide treatment showed beneficial effects on glucose tolerance and insulin sensitivity in HFD-challenged mice as indicated by glucose tolerance tests and insulin tolerance tests (Fig. 1C,D). To investigate the physiological mechanisms responsible for exenatide-restored systemic glucose homeostasis, the expression of hepatic glycometabolic genes was measured. As a result, the mRNA levels of gluconeogenic genes including Pepck and Pck1 were increased in the liver by HFD challenge, and these changes were significantly reduced by exenatide. In contrast, the expression of a glycolytic gene, Gck, which was obviously reduced by HFD feeding, was restored by exenatide treatment (Fig. 1E). These findings demonstrate that exenatide reduces body weight and restores systemic and hepatic glucose homeostasis in HFD-fed mice.

EXENATIDE ALLEVIATES ER STRESS AND HEPATIC STEATOSIS IN HFD-FED MICE

To investigate whether exenatide alleviates lipid-induced hepatic ER stress in an HFD-fed mouse model, the protein levels of the ER stress marker GRP78 and the PERK/eIF2α arm of the UPR pathway were measured in the liver tissues. We found that HFD challenge strongly induced hepatic ER stress, as indicated by increased protein levels of GRP78, phosphorylated PERK (p-PERK) and p-eIF2α in the livers of the HFD group compared with the chow diet group. However, these increases were significantly inhibited by exenatide treatment (Fig. 2A). Similarly, both the liver weight and the ratio of liver to body weights in the HFD group were significantly decreased by exenatide (Fig. 2B). Consistently, histological analysis including hematoxylin and eosin (H&E) staining and oil red O staining revealed that exenatide greatly alleviated the lipid accumulation in the livers of HFD-fed mice, which was also indicated by the reduced liver triglyceride and cholesterol levels (Fig. 2C,D). The same trends were observed in the serum levels of triglyceride and cholesterol (Fig. 2E). Furthermore, the serum levels of adiponectin, an adipose-derived cytokine, were decreased by HFD challenge but reversed after exenatide treatment (Fig. 2E). The gene expression of two adiponectin receptors, AdipoR1 and AdipoR2, was reduced in the livers of the HFD group. However, only the expression of AdipoR2 was significantly increased by exenatide (Fig. 2F). Taken together, these results indicate that exenatide alleviates lipid-induced ER stress and hepatic steatosis in HFD-fed mice.

EXENDIN-4 ALLEVIATES PA-INDUCED ER STRESS AND LIPID ACCUMULATION IN HEPATOCYTES

To confirm whether exendin-4 alleviates lipid-induced hepatic ER stress in vitro, human HepG2 cells and primary murine hepatocytes were introduced in our study. Thapsigargin, modifying Ca²⁺ concentration in the ER lumen by inhibiting Ca²⁺ adenosine triphosphatase, was served as a positive ER stress inducer. Here, thapsigargin exposure was observed to increase the protein levels of GRP78, p-PERK, and p-eIF2α in parallel with the lipid accumulation in HepG2 cells (Fig. 3B; Supporting Fig. S1), suggesting that ER stress could directly lead to lipid accumulation in hepatocytes. Both immunofluorescence staining and western blotting revealed that PA exposure also induced the protein levels of GRP78, p-PERK, and p-eIF2α in HepG2 cells, to a lesser extent than thapsigargin (Fig. 3A,B). Notably, the increased expression of these ER stress markers induced by thapsigargin and PA was significantly inhibited by exendin-4 treatment (Fig. 3A,B). Meanwhile, exendin-4 significantly alleviated the lipid deposition induced by thapsigargin and PA in HepG2 cells (Fig. 3C; Supporting Fig. S1). We also confirmed that exendin-4 reduced PA-induced lipid accumulation in primary murine hepatocytes (Fig. 3D). Immunofluorescence staining revealed that the enhanced positive staining of GRP78 located in the cytoplasm of primary murine hepatocytes induced by PA was substantially reduced by exendin-4 treatment (Fig. 3E). Likewise, the protein levels of GRP78 and p-eIF2α were increased by PA exposure alone and significantly decreased by exendin-4 (Fig. 3F). Collectively, these results suggest that exendin-4 alleviates PA-induced ER stress, which may contribute to its amelioration of lipid accumulation in hepatocytes.

EXENDIN-4 ALLEVIATES PA-INDUCED HEPATIC ER STRESS IN A SIRT1-DEPENDENT MANNER

In HepG2 cells transfected with scrambled shRNA, PA exposure was observed to increase the protein levels of GRP78, p-PERK, and p-eIF2α in parallel with the reduced SIRT1 expression; and these changes were significantly reversed by exendin-4 treatment. By contrast, no obvious change in the expression of GRP78, p-PERK, and p-eIF2α was observed after exendin-4 treatment when Sirt1 was knocked down by Sirt1 shRNA (Fig. 4A). We noted that PA exposure induced a higher expression of GRP78 and p-eIF2α in HepG2 cells transfected with Sirt1 shRNA than that with scrambled shRNA (Fig. 4A). We also observed that exendin-4 significantly decreased the mRNA expression of Grp78 in PA-exposed HepG2 cells, but this was greatly blocked by pretreatment with a SIRT1 inhibitor, EX-527 (Fig. 4B). In addition, the ability of exendin-4 to restore the reduced SIRT1 expression and the increased expression of ER stress markers induced by PA exposure was close to two SIRT1 activators, resveratrol and SRT1720 (Fig. 4C,D). These data demonstrate that exendin-4 alleviates PA-induced hepatic ER stress in a SIRT1-dependent manner.

EXENDIN-4-AMELIORATED PA-INDUCED HEPATIC ER STRESS IS ASSOCIATED WITH INCREASED HSP EXPRESSION THROUGH SIRT1-MEDIATED HSF1 DEACETYLATION

HSP, a large family of molecular chaperones that regulate protein homeostasis, prevent protein aggregation and participate in refolding or elimination of misfolded proteins in their capacity.26, 27 To investigate whether HSP were involved in exendin-4-ameliorated, PA-induced hepatic ER stress, we screened the expression of the HSP family in this system. As a result, exendin-4 treatment significantly increased the protein levels of HSP70 and HSP40 in PA-induced HepG2 cells. However, this effect disappeared when Sirt1 expression was inhibited by Sirt1 shRNA (Fig. 5A). The same trend was confirmed in the mRNA levels by quantitative real-time PCR (Fig. 5B; Supporting Fig. S2). Moreover, the effect of exendin-4 on gene expression induction of HSP was greatly blocked by the pretreatment with EX-527 (Fig. 5C). By contrast, exendin-4 was observed to up-regulate the gene expression of HSP70 and HSP40 comparable to resveratrol and SRT1720 (Fig. 5C). Taken together, these findings show that exendin-4-ameliorated, PA-induced hepatic ER stress is associated with increased HSP expression in a SIRT1-dependent manner.

Because HSF1 is a major transcription factor regulating the expression of HSP genes,28 we performed a ChIP assay to investigate whether exendin-4 influenced the binding of HSF1 to the promoter of HSP genes in a SIRT1-dependent manner. As a result, exendin-4 significantly increased the binding of HSF1 to the Hsp70 promoter in PA-induced HepG2 cells. However, no change was observed after exendin-4 treatment in Sirt1 shRNA-transfected cells (Fig. 5D). Because SIRT1 functions as a protein deacetylase, the acetylation status of HSF1 was determined. In scrambled shRNA-transfected HepG2 cells, PA exposure increased the acetylation of HSF1, and this was significantly reversed by exendin-4 treatment. In Sirt1 shRNA-transfected HepG2 cells, the acetylation of HSF1 was significantly higher and had no change regardless of exendin-4 treatment (Fig. 5E). We also confirmed that the acetylation of HSF1 was increased in the liver after HFD challenge, while this was significantly inhibited by exenatide (Fig. 5F). Altogether, these data indicate that exendin-4-ameliorated, PA-induced hepatic ER stress is associated with increased HSP expression through SIRT1-mediated HSF1 deacetylation.

EXENATIDE-INCREASED HSP EXPRESSION TO ATTENUATE ER STRESS AND HEPATIC STEATOSIS IS DIMINISHED IN HFD-FED Sirt1^+/– MICE

To validate in vitro findings in animal studies, an HFD-fed Sirt1^+/– mouse model was introduced in our study. We observed decreased SIRT1 expression in combination with an exaggerated ER stress response indicated by increased protein levels of GRP78 and p-PERK in the livers of Sirt1^+/– mice compared to WT controls under HFD feeding (Fig. 6A). Not surprisingly, increased SIRT1 expression in parallel with decreased protein levels of GRP78, p-PERK, and p-eIF2α were observed after exenatide treatment in HFD-fed WT mice. However, no significant change was found in HFD-fed Sirt1^+/– mice regardless of exenatide treatment (Fig. 6A). In addition, exenatide notably increased the protein levels of HSP70 and HSP40 in HFD-fed WT mice, while this effect was greatly diminished in HFD-fed Sirt1^+/– mice (Fig. 6B). Histological analysis including H&E staining and oil red O staining showed more severe lipid accumulation in the livers of Sirt1^+/– mice compared to WT controls under HFD feeding. The hepatic lipid accumulation induced by HFD challenge was dramatically improved by exenatide in WT controls. However, this effect was greatly weakened in Sirt1^+/– mice (Fig. 6C). Consistently, the same tendency was observed in the triglyceride and cholesterol contents in both the liver and serum (Fig. 6D,E). These results suggest that exenatide-increased HSP expression to attenuate ER stress and hepatic steatosis is diminished in HFD-fed Sirt1^+/– mice.

Lv Sirt1 KNOCKDOWN ABOLISHES THE EFFECT OF EXENATIDE ON HSF1 DEACETYLATION, HSP INDUCTION, AND ALLEVIATION OF ER STRESS AND HEPATIC STEATOSIS

A Sirt1 knockdown mouse model was also used to further confirm our conclusion. As a result, neither SIRT1 and p-eIF2α nor HSP70 and HSP40 displayed any change in protein expression after Lv-GFP injection compared to the HFD with the exenatide-treated group, while Lv-Sirt1 RNAi injection dramatically decreased SIRT1 expression and abolished the effect of exenatide on increasing HSP expression and reducing p-eIF2α expression in the livers of HFD-fed mice (Fig. 7A,B). Consistent with the in vitro finding, we observed that exenatide-inhibited acetylation of HSF1 was greatly blocked by Lv-Sirt1 RNAi injection (Fig. 7C). As shown by H&E staining and oil red O staining, 2-week exenatide treatment ameliorated the hepatic steatosis induced by 8-week HFD challenge, which was consistent with the results described above. Unlike Lv-GFP injection, Lv-Sirt1 RNAi injection greatly diminished exenatide-ameliorated hepatic steatosis (Fig. 7D). Triglyceride and cholesterol quantification in the liver and serum displayed a similar tendency with the histological analysis of liver (Fig. 7E,F). By contrast, the increased serum adiponectin levels and AdipoR2 gene expression in the liver induced by exenatide treatment were not changed after Lv-Sirt1 RNAi injection (Fig. 7F; Supporting Fig. S3A). In addition, exenatide-restored hepatic glucose homeostasis in HFD-fed mice was significantly weakened by Lv-Sirt1 RNAi injection (Supporting Fig. S3B). Taken together, these findings verify again that Sirt1 knockdown abolishes the effect of exenatide on HSF1 deacetylation, HSP induction, and alleviation of ER stress and hepatic steatosis.