AMPK Re-Activation Suppresses Hepatic Steatosis but its Downregulation Does Not Promote Fatty Liver Development

Abstract

Nonalcoholic fatty liver disease is a highly prevalent component of disorders associated with disrupted energy homeostasis. Although dysregulation of the energy sensor AMP-activated protein kinase (AMPK) is viewed as a pathogenic factor in the development of fatty liver its role has not been directly demonstrated. Unexpectedly, we show here that liver-specific AMPK KO mice display normal hepatic lipid homeostasis and are not prone to fatty liver development, indicating that the decreases in AMPK activity associated with hepatic steatosis may be a consequence, rather than a cause, of changes in hepatic metabolism. In contrast, we found that pharmacological re-activation of downregulated AMPK in fatty liver is sufficient to normalize hepatic lipid content. Mechanistically, AMPK activation reduces hepatic triglyceride content both by inhibiting lipid synthesis and by stimulating fatty acid oxidation in an LKB1-dependent manner, through a transcription-independent mechanism. Furthermore, the effect of the antidiabetic drug metformin on lipogenesis inhibition and fatty acid oxidation stimulation was enhanced by combination treatment with small-molecule AMPK activators in primary hepatocytes from mice and humans. Overall, these results demonstrate that AMPK downregulation is not a triggering factor in fatty liver development but in contrast, establish the therapeutic impact of pharmacological AMPK re-activation in the treatment of fatty liver disease.

Keywords: AMPK; Fatty liver treatment; Lipid metabolism; Metformin; Nonalcoholic fatty liver disease (NAFLD); Small-molecule AMPK activators.

Graphical abstract

Fig. 1

Metabolic characteristics of liver AMPK-deficient mice fasted, fasted-refed a high-carbohydrate diet and fed a high-fat diet. (A) Western-blot analysis of various tissus from WT, AMPKα1^lox/lox, α2^lox/lox (Lox) and AMPKα1^lox/lox, α2^lox/lox-Alfp-CRE (CRE) mice, with the indicated antibodies. (B‐G) Control AMPKα1α2 floxed (WT) and liver-specific AMPKα1α2 KO (Liver AMPK KO) mice were fasted for 24 h (Fasted) or fasted for 24 h then refed a high-carbohydrate diet (Refed) for 12 h. (B) Blood glucose levels, (C) plasma insulin levels, (D) hepatic triglyceride content and (E) hepatic cholesterol content, and (F) plasma triglyceride levels and (G) plasma β-hydroxybutyrate levels. Data are means ± SEM. n = 6-8 per group. ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001 versus fasted WT or liver AMPK KO mice by ANOVA. (H) In vivo synthesis rates of lipids in the livers of WT and liver AMPK KO mice. These data are basal values (vehicle) of experiments from the Fig. 7H. (I‐N) Control AMPKα1α2 floxed (WT) and liver-specific AMPKα1α2 KO (Liver AMPK KO) mice were fed either a standard diet (SD) or a high-fat diet (HFD) for 5 months. (I) Blood glucose levels, (J) Blood glucose levels during oral glucose tolerance test (OGTT), (K) Plasma insulin levels at time 0 and at 20 min in the OGTT, (L) Hepatic triglyceride content and (M) hepatic cholesterol content and (N) plasma β-hydroxybutyrate levels. Data are means ± SEM. n = 6 per group. ^⁎P < 0.05, ^⁎⁎P < 0.01 versus SD-fed WT or liver AMPK KO mice; ^#P < 0.05 versus time T0 in WT or liver AMPK KO mice; by ANOVA.

Fig. 2

Effect of various AMPK activators on triglyceride content and lipogenic gene expression in control and AMPK-deficient hepatocytes. control AMPKα1α2 floxed (WT) and AMPKα1α2 KO (AMPK KO) primary hepatocytes were cultured for 16 h in medium containing 5 or 25 mM glucose (G5 or G25), with or without 100 nM insulin, in the presence or absence of various concentrations of AICAR, metformin or A-769662. (A) Intracellular triglyceride content. (B) Immunoblots were performed with the antibodies indicated. (C) Lipogenic gene expression. Data are means ± SEM from 5 independent experiments. ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001 versus WT or AMPK KO hepatocytes incubated with G25 + insulin; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus WT hepatocytes incubated under the same conditions; by ANOVA.

Fig. 3

The forced expression of lipogenic genes does not prevent the AMPK-induced inhibition of hepatic lipogenesis. WT primary hepatocytes were infected with Ad-GFP or Ad-mSREBP-1c adenovirus and cultured for 16 h in the presence 25 mM glucose. Hepatocytes were then incubated in fresh medium containing 25 mM glucose with or without 100 μM AICAR or 30 μM A-769662 for 3 h. (A) Lipogenic gene expression. (B) Immunoblots performed with the antibodies indicated and quantification of immunoblot images for the ACC and FAS proteins normalized to GAPDH. (C) The rate of lipid synthesis was estimated from the incorporation of [1-¹⁴C]-acetate into total lipids. Data are means ± SEM from 3 independent experiments. ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001, versus Ad-GFP-infected hepatocytes; ^###P < 0.001 versus Ad-mSREBP-1c-infected hepatocytes; by ANOVA.

Fig. 4

Effects of various AMPK activators on lipid synthesis and fatty acid oxidation rates in control and AMPK-deficient hepatocytes. control AMPKα1α2 floxed (WT) and AMPKα1α2 KO (AMPK KO) primary hepatocytes were incubated in the presence or absence of various concentrations of AICAR, metformin, A-769662, C13, 991 or TOFA for 3 h. (A) Immunoblots were performed with the indicated antibodies. (B, C) Fatty acid and sterol synthesis were assessed from the incorporation of [1-¹⁴C]-acetate into saponifiable and non-saponifiable lipids, respectively. (D) Palmitate oxidation rates were determined by measuring the production of ¹⁴C–labeled acid-soluble metabolites from [1-¹⁴C]-palmitic acid. Data are means ± SEM from 3 to 5 independent experiments. ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001 versus WT or AMPK KO hepatocytes incubated in the absence of compounds; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 WT versus AMPK KO hepatocytes incubated under the same conditions; by ANOVA.

Fig. 5

Small-molecule AMPK activators enhance the action of metformin and AICAR on AMPK activation, lipid synthesis and fatty acid oxidation in primary hepatocytes. Control AMPKα1α2 floxed (WT) and AMPKα1α2 KO (AMPK KO) primary hepatocytes were incubated with or without various concentrations of AICAR or metformin in the absence or presence of A-769662 (1 or 10 μM) or C13 (1 μM) for 3 h. (A) Immunoblots were performed with the antibodies indicated. (B) P-AMPKα/AMPKα and P-ACC/ACC ratios from the quantification of immunoblot images. Data are means ± SEM. ^⁎P < 0.05, ^⁎⁎P < 0.01 versus WT hepatocytes incubated in the absence of A-769662 by ANOVA. (C, E, G) Fatty acid and sterol synthesis were assessed from the incorporation of [1-¹⁴C]-acetate into saponifiable and non-saponifiable lipids, respectively. Results are presented as a percentage of acetate incorporated into WT or AMPK KO hepatocytes incubated in the absence of compounds. (D, F, H) Palmitate oxidation rates were determined by measuring the production of ¹⁴C–labeled acid-soluble metabolites from [1-¹⁴C]-palmitic acid. Data are means ± SEM from 3 independent experiments. ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎P < 0.001 versus WT hepatocytes incubated in the absence of compounds; ^##P < 0.01, ^###P < 0.001 for comparisons between the conditions indicated; by ANOVA.

Fig. 6

Indirect and small-molecule AMPK activators inhibit lipid synthesis in primary human hepatocytes. (A) Assessment of AMPK subunit levels in primary human and mouse hepatocytes by western blotting with the indicated antibodies. (B-J) Human primary hepatocytes were cultured in the absence or presence of various concentrations of A-769662, C13, AICAR or metformin, alone or in combination, at the indicated concentrations for 3 h. (B) Immunoblots were performed with the indicated antibodies. (C-J) Fatty acid and sterol synthesis were assessed from the incorporation of [1-¹⁴C]-acetate into saponifiable and non-saponifiable lipids, respectively. Data are means ± SEM from 3 independent experiments. ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001 versus hepatocytes incubated in the absence of activators (C—F); for comparisons between the conditions indicated (G-J); by ANOVA.

Fig. 7

Small-molecule-mediated activation of AMPK in the liver restores fatty acid oxidation, abolishes hepatic steatosis and improves insulin sensitivity in aP2-nSREBP-1c transgenic mice. (A-E) aP2-nSREBP-1c (aP2) and non-transgenic (WT) littermate mice were treated for 8 days with vehicle or A-769662 (30 mg/kg, i.p., b.i.d.). Mice were sacrificed after a 24-h fast. (A) Western-blot analysis of liver lysates with the antibodies indicated. Each lane represents an individual mouse. The lower panels show P-AMPKα/AMPKα and P-ACC/ACC ratios from the quantification of immunoblot images (n = 5 per group). (B) Oil Red O staining of representative liver sections. Scale bars: 50 μm. (C) Liver triglyceride content, (D) liver cholesterol content and (E) plasma β-hydroxybutyrate levels (n = 5-6 per group). (F) Insulin tolerance test in aP2-nSREBP-1c transgenic mice (n = 5-6 per group) treated for 7 days with vehicle or A-769662 (30 mg/kg, i.p., b.i.d.). Data are means ± SEM. ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001 versus WT + Veh; ^##P < 0.01, ^###P < 0.001 versus aP2 + Veh; by ANOVA. (G) Control AMPKα1α2 floxed (WT) and liver-specific AMPKα1α2 KO (Liver AMPK KO) mice fed either a standard diet (SD) or a high-fat diet (HFD) for 5 months were treated with vehicle or A-769662 (30 mg/kg, i.p., b.i.d.). After 5 days of treatment, mice were sacrificed in the fed state, and their livers were collected for hepatic triglyceride content determination. Data are means ± SEM. (n = 6 per group). ^⁎P < 0.05 HFD versus SD, ^#P < 0.05 WT versus liver AMPK KO by ANOVA. (H) In vivo synthesis rate of lipids in the livers of AMPKα1α2 floxed (WT) and liver-specific AMPKα1α2 KO mice (Liver AMPK KO) mice treated with vehicle or A-769662 were assessed by determining the incorporation of ³H-labeled water into lipids. Data are means ± SEM. n = 4 per group. ^⁎P < 0.05 versus vehicle; ^#P < 0.05 versus WT by ANOVA.