SIRT1 regulates hepatocyte lipid metabolism through activating AMP-activated protein kinase

Abstract

Resveratrol may protect against metabolic disease through activating SIRT1 deacetylase. Because we have recently defined AMPK activation as a key mechanism for the beneficial effects of polyphenols on hepatic lipid accumulation, hyperlipidemia, and atherosclerosis in type 1 diabetic mice, we hypothesize that polyphenol-activated SIRT1 acts upstream of AMPK signaling and hepatocellular lipid metabolism. Here we show that polyphenols, including resveratrol and the synthetic polyphenol S17834, increase SIRT1 deacetylase activity, LKB1 phosphorylation at Ser(428), and AMPK activity. Polyphenols substantially prevent the impairment in phosphorylation of AMPK and its downstream target, ACC (acetyl-CoA carboxylase), elevation in expression of FAS (fatty acid synthase), and lipid accumulation in human HepG2 hepatocytes exposed to high glucose. These effects of polyphenols are largely abolished by pharmacological and genetic inhibition of SIRT1, suggesting that the stimulation of AMPK and lipid-lowering effect of polyphenols depend on SIRT1 activity. Furthermore, adenoviral overexpression of SIRT1 stimulates the basal AMPK signaling in HepG2 cells and in the mouse liver. AMPK activation by SIRT1 also protects against FAS induction and lipid accumulation caused by high glucose. Moreover, LKB1, but not CaMKKbeta, is required for activation of AMPK by polyphenols and SIRT1. These findings suggest that SIRT1 functions as a novel upstream regulator for LKB1/AMPK signaling and plays an essential role in the regulation of hepatocyte lipid metabolism. Targeting SIRT1/LKB1/AMPK signaling by polyphenols may have potential therapeutic implications for dyslipidemia and accelerated atherosclerosis in diabetes and age-related diseases.

FIGURE 1.

Pharmacological activation of SIRT1 by resveratrol prevents the decrease in AMPK activity and elevation in FAS expression and triglyceride level in human HepG2 hepatocytes exposed to high glucose. A, the dose-response effect of resveratrol on in vitro SIRT1 deacetylase activity. Fluor de Lys fluorescence deacetylase assays were preformed with human recombinant SIRT1, using a synthetic acetylated Lys³⁸² p53 peptide and NAD as substrates in the absence or presence of increasing concentrations (10–100 μm) of resveratrol as described under “Experimental Procedures.” SIRT1 activity was expressed as arbitrary fluorescence units relative to the control (mean ± S.E., n = 4). *, p < 0.05 versus control. B, resveratrol prevents the inhibition of AMPK caused by high glucose in a dose-dependent manner. HepG2 cells were maintained in serum-free DMEM containing normal glucose overnight and incubated for 24 h without or with increasing concentrations (1–50 μm) of resveratrol in the absence or presence of 30 mm d-glucose (high glucose). Representative immunoblotting analysis with antibodies against AMPKα phosphorylated at Thr¹⁷² (pAMPK) and total AMPKα1 or -α2 for loading controls, respectively, is shown. Immunoblots with anti-SIRT1 antibody show the existence of ∼120 kDa of endogenous SIRT1 in HepG2 cells and no detectable change in SIRT1 throughout treatment. C and D, high glucose increases and resveratrol suppresses expression of FAS. Expression of FAS was analyzed by immunoblots with anti-FAS and normalized to β-actin level and presented as the -fold change (mean ± S.E., n = 3). E, resveratrol protects against high glucose-induced triglyceride accumulation. Intracellular triglyceride contents were measured and expressed as μg of lipid/mg of protein (the mean ± S.E., n = 4) as described under “Experimental Procedures.” *, p < 0.05 versus normal glucose; #, p < 0.05 versus high glucose alone.

FIGURE 2.

Pharmacological inhibition of SIRT1 attenuates polyphenol-induced AMPK activation and lipid reduction in human HepG2 cells or HEK293 cells. A, SIRT1 deacetylase activity is largely inhibited by nicotinamide. The SIRT1 activity assay was carried out in the absence or presence of nicotinamide or suramin, an SIRT1 inhibitor included in the SIRT1 activity assay kit. *, p < 0.05 versus control. B–D, inhibition of SIRT1 activity by nicotinamide or splitomicin diminishes enhanced phosphorylation of AMPK and ACC in response to the polyphenol in HepG2 cells. Cells were pretreated without or with either nicotinamide (10 mm) or splitomicin (100 μm) in the serum-free medium for 24 h and then incubated without or with S17834 (10 μm) or resveratrol (50 μm) for an another 1 h. Densitometric quantification of the phosphorylation of AMPK and ACC is shown. *, p < 0.05 versus control; #, p < 0.05 versus polyphenol alone (mean ± S.E., n = 3). E, nicotinamide prevents the lipid-lowering effect of polyphenols in HepG2 cells. Cells were pretreated for 24 h without or with nicotinamide (10 mm) in serum-free medium and incubated for 24 h without or with polyphenols in the presence of high glucose. *, p < 0.05 versus normal glucose alone; #, p < 0.05 versus high glucose alone; ##, p < 0.05 versus high glucose plus polyphenols (mean ± S.E., n = 4). F–H, the polyphenols increase and splitomicin decreases phosphorylation of AMPK and ACC in HEK293 cells. Phosphorylation of AMPK and ACC was assessed in cells pretreated with splitomicin (100μm) for 24 h and incubated with resveratrol (50 μm) or S17834 (10 μm) for an additional 1 h as indicated. *, p < 0.05 versus control; #, p < 0.05 versus polyphenol alone (mean ± S.E., n = 3). No detectable change in the expression of endogenous SIRT1 was observed throughout treatments in both human cell lines.

FIGURE 3.

Lentivirus-mediated knockdown of SIRT1 diminishes the basal and polyphenol-induced AMPK activation in HepG2 cells. HepG2 cells were infected without or with lentivirus expressing either an shRNA control or SIRT1 shRNA and then selected in 0.6 μg/ml puromycin. Cells were allowed to recover from the selection for 1 week prior to the experiments. A, co-expression of GFP in both control shRNA cells and SIRT1 shRNA cells was observed under fluorescence microscopy. B, endogenous SIRT1 expression was largely suppressed by lentivirus expressing SIRT1 shRNA. Representative immunoblots for the expression of SIRT1 and β-actin are shown in duplicates under the identical condition. C–E, knockdown of SIRT1 by lentivirus-mediated SIRT1 shRNA down-regulates the basal and polyphenol-stimulated AMPK and ACC phosphorylation. HepG2 cells expressing either control or SIRT1 shRNA were quiesced in serum-free medium overnight and treated with S17834 (10 μm, 1 h). *, p < 0.05 versus untreatment in cells expressing control shRNA; #, p < 0.05 versus S17834 treatment in cells expressing control shRNA (mean ± S.E., n = 3).

FIGURE 4.

Polyphenol-induced AMPK activation and lipid reduction is abolished by overexpression of a catalytically inactive mutant of SIRT1 (SIRT1H355A) in HepG2 cells. A, a representative immunoblot of overexpression of an adenovirus vector encoding a catalytically inactive SIRT1 mutant in HepG2 cells is shown. B, polyphenol-stimulated AMPK phosphorylation is abolished by the SIRT1H355A mutant under normal glucose conditions. HepG2 cells were infected for 48 h with Ad-GFP or Ad-SIRT1H355A, followed by treatment with S17834 (10 μm) or resveratrol (50 μm) for 1 h. C and D, polyphenol-stimulated AMPK signaling is diminished by the SIRT1H355A mutant under high glucose conditions. HepG2 cells infected with Ad-GFP or Ad-SIRT1H355A were quiesced in serum-free medium overnight and treated for 24 h without or with 10 μm of resveratrol or S17834 in the presence of high glucose. E, the lipid-lowering effect of resveratrol is attenuated by the SIRT1H355A mutant. *, p < 0.05 versus normal glucose in cells expressing GFP; #, p < 0.05 versus high glucose alone in cells expressing GFP; ##, p < 0.05 versus polyphenol treatment in cells expressing GFP (mean ± S.E., n = 4).

FIGURE 5.

Hepatic overexpression of SIRT1 is sufficient to enhance phosphorylation of AMPK and ACC over the basal levels in HepG2 cells and in mouse liver in vivo. A, immunoblots with SIRT1 antibody confirm overexpression of recombinant SIRT1 protein (∼120 kDa) in HepG2 cells infected with adenovirus-mediated vector encoding wild type SIRT1 (Ad-SIRT1). B and C, overexpression of SIRT1 is sufficient to increase the baseline phosphorylation of AMPK and ACC in HepG2 cells. D, in vivo expression of adenovirus-mediated vector encoding FLAG-tagged wild type SIRT1 (Ad-FLAG-SIRT1) is visualized by Western blots with anti-FLAG antibody in the livers from different C57BL/6 mice that were sacrificed 7 days postinjection as indicated. Ad-GFP-infected mice were used as a control. E and F, phosphorylation of AMPK and ACC is increased in the livers of mice injected with Ad-FLAG-SIRT1 in vivo. Representative immunoblots of phosphorylation of AMPK and ACC in the livers from two mice each group as indicated and densitometric analysis are shown. *, p < 0.05 versus Ad-GFP-infected HepG2 cells or Ad-GFP-injected mice (mean ± S.E., n = 3).

FIGURE 6.

AMPK is required for SIRT1 to suppress FAS induction and lipid accumulation in HepG2 cells exposed to high glucose. HepG2 cells were infected with adenoviral vectors encoding GFP or a Myc-tagged dominant-negative AMPK mutant (Ad-DN-AMPK, AMPKαK45R) or co-infected with Ad-SIRT1 and subsequently incubated for 24 h without or with resveratrol (10 μm) in the absence or presence of high glucose. A, overexpression of the DN-AMPK (∼64 kDa) and SIRT1 (∼120 kDa) was confirmed by immunoblots with anti-Myc and anti-AMPKα2 and with anti-SIRT1 antibodies, respectively. B and C, the ability of SIRT1 to prevent the decrease in ACC phosphorylation caused by high glucose is attenuated by the DN-AMPK. D and E, SIRT1 suppression of high glucose-enhanced FAS expression is abrogated by the DN-AMPK. F, DN-AMPK blocks the effect of SIRT1 and resveratrol on lipid accumulation. *, p < 0.05 versus normal glucose in cells expressing GFP; #, p < 0.05 versus high glucose alone in cells expressing GFP; ##, p < 0.05 versus polyphenol treatment in cells expressing GFP (mean ± S.E., n = 4).

FIGURE 7.

Polyphenols and SIRT1 stimulate AMPK signaling in an LKB1-dependent manner. A–C, phosphorylation of LKB1 at Ser⁴²⁸ and AMPK at Thr¹⁷² in response to polyphenols is increased in HepG2 cells expressing LKB1 but not in HeLa cells lacking LKB1. Both human HepG2 hepatocytes and HeLa cells were quiesced in serum-free DMEM overnight and incubated with AICAR (1 mm), S17834 (10 μm), or resveratrol (50 μm) for an additional 1 h. The levels of phosphorylated LKB1 (pLKB1) or AMPK (pAMPK) were normalized to those of total LKB1 or AMPKα2, respectively, and expressed as relative phosphorylation (mean ± S.E., n = 3). *, p < 0.05, versus control in the identical cell lines. D and E, overexpression of LKB1 restores AMPK activation by polyphenols in LKB1-deficient HeLa cells. Adenovirus-mediated vectors encoding either GFP (Ad-GFP) or FLAG-tagged wild type LKB1 (Ad-FLAG-LKB1) were infected into HeLa cells, followed by treatment with S17834 (10 μm) or resveratrol (50 μm) for 1 h as indicated. F–H, SIRT1-dependent AMPK signaling is restored by overexpression of LKB1 in HeLa cells. Ad-GFP, Ad-FLAG-LKB1, or Ad-SIRT1 was infected into HeLa cells as indicated. Immunoblots with anti-FLAG, anti-LKB1, or anti-SIRT1 antibodies confirmed overexpression of recombinant FLAG-LKB1 (∼60 kDa) or SIRT1 (∼120 kDa). *, p < 0.05 versus Ad-GFP alone; #, p < 0.05 versus Ad-FLAG-LKB1 alone (mean ± S.E., n = 3).

FIGURE 8.

Proposed scheme for the role of SIRT1-activating polyphenols in the regulation of AMPK signaling and hepatocyte lipid metabolism. SIRT1-activating polyphenols, such as resveratrol and S17834, stimulate LKB1 phosphorylation as well as AMPK phosphorylation and activation. Similarly, overexpression of wild type SIRT1 also increases AMPK phosphorylation, which in turn increases ACC phosphorylation and inhibits ACC activity. As a consequence, decreased production of malonyl-CoA results in up-regulation of fatty acid oxidation as well as down-regulation of fatty acid synthesis, thereby leading to hepatocyte lipid reduction. On the other hand, activation of AMPK by polyphenols inhibits glucose-induced expression of FAS, which contributes to the reduction in triglycerides through inhibition of fatty acid biosynthesis. Moreover, SIRT1 is required for the effects of polyphenols on AMPK and lipids, since the beneficial effects of polyphenols are mimicked by overexpression of wild type SIRT1 and abrogated by inhibition of SIRT1, such as the inactive SIRT1 mutant (SIRT1H355A), shRNA SIRT1, or inhibitors of SIRT1. Importantly, the stimulation of AMPK and the lipid-lowering effect of SIRT1 are abolished by DN-AMPK, suggesting that AMPK acts as a novel functional downstream regulator for the metabolic effect of SIRT1. In addition, polyphenols and SIRT1 stimulate AMPK via an LKB1-dependent mechanism. Therefore, SIRT1/LKB1/AMPK signaling may be considered a novel molecular mechanism for potential therapeutic effects of polyphenols on hepatic lipid accumulation in diabetes and age-related metabolic disorders.