Activation of ERK1/2 Ameliorates Liver Steatosis in Leptin Receptor–Deficient (db/db) Mice via Stimulating ATG7-Dependent Autophagy

Nonalcoholic fatty liver disease involves a serious pathological change in liver (1). The initial stage of nonalcoholic fatty liver disease is liver steatosis, characterized by the excess deposition of triglyceride (TG) and/or cholesterol (TC) in liver (2). If uncontrolled, liver steatosis will progress to life-threatening diseases, such as liver cirrhosis and dysfunction (3). Abnormal hepatic lipid accumulation results from increased uptake of fatty acid/augmented de novo lipogenesis and/or decreased β-oxidation/impaired TG export (4).

Autophagy, a cellular process that degrades intracellular organelles and proteins (5), has recently been demonstrated to regulate lipid metabolism (6,7). Lipid droplets are sequestered by autophagosome with the coordination of autophagy-related genes (ATGs). Autophagosome is then fused with lysosome (8) for the degradation of lipid droplets into free fatty acids (FFAs). FFAs are then degraded by mitochondrial β-oxidation to produce ATP or are reesterified into TG for storage (9). Impaired autophagy decreases hepatic fatty acid β-oxidation (FAO) and TG export and results in liver steatosis in mice (7,10), and fatty liver is ameliorated when hepatic autophagy is stimulated by certain compounds (11,12) or some signaling pathways (13) in various animal models.

The mitogen-activated protein kinase–extracellular signal–regulated kinase (MEK-ERK) signaling pathway is involved in a wide variety of cellular processes (14–16). Several lines of evidence, however, have implied a link between ERK1/2 and lipid metabolism (17–20). A direct effect of ERK1/2 on hepatic lipid metabolism, however, has not been reported. Based on the knowledge detailed above and the fact that ERK1/2 is involved in autophagy (21–23), we hypothesized that ERK1/2 may play a role in liver steatosis via affecting autophagy. The aim of current study was to test this hypothesis and elucidate underlying mechanisms.

Male C57/B6J wild-type (WT) mice were purchased from Shanghai Laboratory Animal Co. Ltd. (Shanghai, China). Male 10-week-old leptin receptor–deficient (db/db) mice were kindly provided by Xiang Gao, Nanjing University, Nanjing, China. Mice were maintained as previously described (24). A high-fat diet (HFD) was purchased from Research Diets (60% fat, D12492; Research Diets). For autophagy inhibitor treatment, chloroquine (C6628; Sigma-Aldrich) was dissolved in PBS and injected at the dose of 60 mg/kg i.p. daily in mice for 14 days (12,25). For VLDL secretion assay, mice were injected with poloxamer 407 (16758; Sigma-Aldrich) at the dose of 1 g/kg i.p. and TG levels were measured in serum of tail vein blood taken at different time points (26). These experiments were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee of the Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences.

Mouse primary hepatocytes were prepared by collagenase perfusion and treated with SB203580 (S8307; Sigma-Aldrich) as previously described (24,27).

Recombinant adenoviral vector expressing rat constitutively active MEK1 (Ad-CA MEK1) was kindly provided by Haiyan Xu from Brown University. Adenoviruses expressing (Ad-) scrambled (Ad-scrambled) and short-hairpin (sh)RNAs specific for ERK1/2 (Ad-shERK1/2) or ATG7 (Ad-shATG7) were generated using the BLOCK-iT Adenoviral RNAi Expression System (K4941-00; Invitrogen) according to the manufacturer’s instructions. The scrambled sequence is 5′-TTCTCCGAACGTGTCACGT-3′, and the shRNA sequence for mouse ERK1/2 is 5′-CACCGCAATGACCACATCTGCTACTCGAAAGTAGCAGATGTGGTCATTGC-3′ and for mouse ATG7 is 5′-CACCATGAGATCTGGGAAGCCATCGAAATGGCTTCCCAGATCTCAT-3′. Double-stranded small interfering (si)RNA targeting mouse p38α was from Cell Signaling Technology (catalog no. 6417). The Ad-ATG7 was constructed using a human ATG7 expression vector (Neuronbiotech Company, Shanghai, China). Recombinant adenoviruses were purified (24) and diluted in PBS and administrated at a dose of 1 × 10⁷ plaque-forming units (pfu)/well in 12-well plates or via tail vein injection using 5 × 10⁸ pfu/mouse (Ad-scrambled, Ad-shERK1/2, and Ad-shATG7) or 1× 10⁹ pfu/mouse (Ad–green fluorescent protein [Ad-GFP], Ad-ATG7, and Ad-CA MEK1).

Relative quantification RT-PCR and Western blot analysis were performed as previously described (24). Primary antibodies (anti–phosphorylated [p]-ERK1/2 [Thr²⁰²/Tyr²⁰⁴] [9106], anti–total (t)-ERK1/2 [9102], anti–p-p38 [Thr¹⁸⁰/Tyr¹⁸²] [9211], anti–t-p38 [8690], anti-MEK1 [9124], anti-LC3 [2775], anti-CPT1α [12252], anti-proliferating cell nuclear antigen (anti-PCNA) [13110], anti–cyclin D1 [2926], anti-PARP [9542], anti-caspase3 [9662], anti–t-AKT [9272], anti–p-AKT [Ser⁴⁷³] [9271], anti–t-GSK3β [9315], anti–p-GSK3β [Ser⁹] [9336], anti–t-HSL [4107], anti–p-HSL [Ser⁶⁶⁰] [4126], and anti-ATG7 [2631]) (all from Cell Signaling Technology); anti-actin (A4700; Sigma-Aldrich); anti-tubulin (P8203; Sigma-Aldrich); and anti-SQSTM1 (ab56416; Abcam) were incubated overnight at 4°C, and specific proteins were visualized using ECL Plus (9589151; Amersham Biosciences).

Hepatic and cellular lipids were extracted with chloroform methanol (2:1) as previously described (28). TG, TC, FFAs, alanine transaminase (ALT), aspartate aminotransferase (AST), and β-hydroxybutyrate (3-HB) were measured with a TG kit (290-63701; Wako), TC kit (294-65801; Wako), FFA kit (436-91995; Wako), ALT kit (700260; Cayman Chemical), AST kit (K735-100; Bio Vision), and 3-HB colorimetric assay kit (700190; Cayman Chemical), respectively, according to the manufacturers’ instructions.

FAO assays were conducted as previously described (29).

Immunofluorescence assay was performed as previously described (30). Images were obtained using a Zeiss LSM 510 confocal microscope (Carl Zeiss Imaging, Oberkochen, Germany).

Frozen sections of liver were stained with Oil red O. Paraformaldehyde-fixed, paraffin-embedded sections of liver were stained with hematoxylin-eosin (H-E) for histology.

Levels of blood glucose and serum insulin were measured using a Glucometer Elite monitor and a Mercodia Ultrasensitive Rat Insulin ELISA kit (03113-1; ALPCO Diagnostic), respectively. The HOMA of insulin resistance (HOMA-IR) index was calculated according to the following formula: [fasting glucose levels (mmol/L)] × [fasting serum insulin (μU/mL)]/22.5. The insulin tolerance test was performed by injection of 0.75 units/kg i.p. insulin after 4 h of fasting. The in vivo insulin signaling assay was performed as previously described (24).

All data are expressed as means ± SEM. Significant differences were assessed either by two-tailed Student t test or by one-way ANOVA followed by the Student-Newman-Keuls (SNK) test. P < 0.05 was considered statistically significant.

To explore a role of ERK1/2 in liver steatosis, we examined ERK1/2 activity in liver of db/db mice, a classic animal model for liver steatosis (31). We found that ERK1/2 activity was compromised in the livers of db/db mice compared with control mice (Fig. 1A). We then injected db/db mice with Ad-CA MEK1, the specific upstream activator of ERK1/2 (17), or control Ad-GFP. As predicted, overexpression of MEK1 increased hepatic ERK1/2 phosphorylation compared with control mice but had no effect on body weight, liver weight, fat mass, or food intake (Fig. 1B–D). In contrast, the extensive lipid deposition manifested as macro- and microvesicular steatosis (as examined by Oil red O and H-E staining) in the livers of db/db mice disappeared after overexpression of ERK1/2 compared with control mice (Fig. 1B and E). Consistently, liver TG content was also decreased by ERK1/2 activation (Fig. 1F). However, serum TG levels were increased in Ad-CA MEK1 mice (Fig. 1G), suggesting that VLDL secretion may be largely enhanced in these mice, as confirmed by our VLDL secretion assay (Supplementary Fig. 1). Liver TC levels were not changed, but serum TC levels were increased by ERK1/2 activation (Fig. 1F and G). Neither serum nor liver FFA levels were affected by ERK1/2 activation (Fig. 1F and G).

The reversal effects of ERK1/2 activation on fatty liver in db/db mice are likely to reflect an effect on hepatic TG synthesis, β-oxidation, uptake, and/or secretion of fatty acids. We first investigated whether genes underlying the synthesis of triglycerides and uptake of fatty acids are differentially regulated in Ad-CA MEK1 or Ad-GFP db/db mice. These genes included fatty acid synthase (Fas), acetyl CoA carboxylase 1 (Acc1), stearoyl CoA desaturase (Scd1), malic enzyme (Me), glycerol-3-phosphate acyltransferase (Gpat), and transcription factors including peroxisome proliferator–activated receptor γ (Pparγ), sterol regulatory element–binding protein 1c (Srebp1c), and carbohydrate-responsive element-binding protein (Chrebp) (32). Genes related to fatty acid uptake include cluster of differentiation 36 (Cd36), fatty acid–binding protein (Fabp), and fatty acid–transporting protein (Fatp) (32). Interestingly, genes related to lipogenesis were not affected or some even increased and fatty acid uptake genes showed inconsistent changes in livers by Ad-CA MEK1 injection in db/db mice (Fig. 1H and I). Consistent with reversal effects of ERK1/2 activation on liver steatosis, genes involved in TG secretion, including apolipoprotein B (ApoB) and apolipoprotein E (ApoE), and fatty acid oxidation, including peroxisome proliferator–activated receptor α (Pparα) and carnitine palmitoyltransferase 1α (Cpt1α), were significantly increased in the livers of Ad-CA MEK1 db/db mice (Fig. 1J and K). Consistently, levels of hepatic CPT1α protein and serum 3-HB were increased in Ad-CA MEK1 db/db mice (Fig. 1L and M). In addition, levels of serum ALT and AST were decreased by Ad-CA MEK1 (Fig. 1N and O).

To further investigate the impact of ERK1/2 on liver steatosis in vivo, we injected WT mice with Ad-shERK1/2 or Ad-scrambled. Ad-shERK1/2 significantly decreased hepatic ERK1/2 protein levels, with no effects on body weight, liver weight, fat mass, or food intake, but caused liver steatosis as confirmed by Oil red O and H-E staining (Fig. 2A–D). Consistently, levels of hepatic TG, TC, and FFAs were also significantly increased in Ad-shERK1/2 mice compared with control mice (Fig. 2E). In contrast, serum TC and FFA levels were decreased, but serum TG levels were not affected, in Ad-shERK1/2 mice (Fig. 2F). Most genes related to lipogenesis (except for Pparγ), FAO, and TG secretion were significantly decreased, but genes related to fatty acid uptake showed inconsistent changes in liver of Ad-shERK1/2 mice (Fig. 2G–J). Consistently, levels of hepatic CPT1α protein and serum 3-HB were also decreased in Ad-shERK1/2 mice (Fig. 2K and L). In addition, serum levels of ALT and AST were increased by Ad-shERK1/2 injection (Fig. 2M and N).

Because autophagy accelerates FAO and TG export in mouse livers (7) and ERK1/2 is shown to regulate autophagy in HepG2 cells (22), we hypothesized that ERK1/2 might regulate liver steatosis via autophagy. To test this possibility, we examined expression of hepatic LC3-II (a positive autophagy marker) and SQSTM1/p62 (a negative autophagy marker) (33) in db/db and WT mice and found that autophagy was attenuated (as demonstrated by decreased LC3-II and increased SQSTM1 expression) in db/db mice, as reported previously (33), which was reversed by Ad-CA MEK1 (Fig. 3A and B). Opposite effects were observed in WT mice with ERK1/2 knockdown (Fig. 3C). Similar effects were observed in primary hepatocytes (Fig. 3D and E). Consistently, autophagic flux was enhanced by ERK1/2 activation or inhibited by ERK1/2 knockdown, respectively, in primary hepatocytes transfected or infected with RFP-LC3 plasmid or Ad-CA MEK1 or Ad-shERK1/2 (Fig. 3F). Because autophagy is a dynamic process, we further treated cells with chloroquine (CQ), a lysosomotropic weak base that blocks the fusion of autophagosome with lysosome (12,25,33). In the presence of CQ, a greater amount of autophagasome was seen in control cells, but the increase was much bigger in Ad-CA MEK1 cells or smaller in Ad-shERK1/2 cells (Fig. 3F).

To confirm a role for autophagy in mediating ERK1/2-ameliorated liver steatosis in db/db mice, we injected Ad-CA MEK1 db/db mice with PBS or CQ (12,25) for 14 days. p-ERK1/2 and MEK1 were increased by Ad-CA MEK1, in the presence or absence of CQ treatment, in db/db mice (Supplementary Fig. 2). Autophagic flux inhibited by CQ (as confirmed by Western blot) had no effects on body weight, liver weight, fat mass, or food intake but reversed the effects of Ad-CA MEK1 on ameliorated liver steatosis (as confirmed by Oil red O and H-E staining and hepatic TG measurement) in db/db mice (Fig. 4A–E). CQ treatment had no effects on levels of liver TC and FFAs and serum FFAs but decreased serum TG and TC levels in Ad-CA MEK1 db/db mice (Fig. 4E and F). In addition, CQ treatment reversed the effects of Ad-CA MEK1 on hepatic genes expression related to TG secretion and FAO, but had no effects on other genes, in db/db mice (Fig. 4G–J). Levels of hepatic CPT1α protein and serum 3-HB were also reversed in Ad-CA MEK1 db/db mice by CQ treatment (Fig. 4K and L). Furthermore, the decreased serum levels of ALT and AST of Ad-CA MEK1 db/db mice were also reversed by CQ treatment (Fig. 4M and N). In addition, CQ treatment had similar effects in db/db mice injected with control Ad-GFP (Supplementary Fig. 3).

To investigate molecular mechanisms underlying ERK1/2 control of autophagy, we examined mRNA expression of autophagy regulators Atg4a, Atg5, Atg6 (Beclin-1 [Becn1]), and Atg7, all important for the production of autophagosome (34), in primary hepatocytes infected with Ad-CA MEK1 or control Ad-GFP. We found that mRNA levels of Atg4a, Atg5, and Atg7 were increased by ERK1/2 activation (Fig. 5A). Opposite effects were observed when ERK1/2 was knocked down by Ad-shERK1/2 (Fig. 5B). As observed in vitro, Ad-CA MEK1 also induced Atg4a, Atg5, and Atg7 expression in the livers of db/db mice, whereas Ad-shERK1/2 only inhibited Atg7, but not Atg4a and Atg5, expression in the livers of WT mice (Fig. 5C and D). Hepatic Becn1 expression, however, was not affected by Ad-CA MEK1 or Ad-shERK1/2 (Fig. 5A–D). Consistent with changes in Atg7 mRNA, ATG7 protein levels were also increased or decreased by ERK1/2 activation or knockdown, respectively, in primary hepatocytes (Fig. 5E and F). Compared with WT mice, hepatic ATG7 protein levels were decreased in db/db mice, which was further stimulated by ERK1/2 activation (Fig. 5G and H). Consistently, ERK1/2 knockdown decreased hepatic ATG7 protein levels in WT mice (Fig. 5I).

To investigate whether ATG7-dependent autophagy is responsible for the ameliorated liver steatosis in Ad-CA MEK1 db/db mice, we injected db/db mice with Ad-CA MEK1 in combination with Ad-shATG7 or Ad-scrambled. ATG7 knockdown (as confirmed by Western blot) had no effects on body weight, liver weight, fat mass, or food intake but reversed the effects of Ad-CA MEK1 on improved autophagy (as measured by autophagy-related markers) and ameliorated liver steatosis (as confirmed by Oil red O and H-E staining and hepatic TG measurement) in db/db mice (Fig. 6A–D). ATG7 knockdown had no effects on levels of liver TC and FFAs and serum FFAs but decreased serum TG and TC levels in Ad-CA MEK1 db/db mice (Fig. 6E and F). In addition, ATG7 knockdown reversed the effects of Ad-CA MEK1 on hepatic gene expression related to TG secretion and FAO, but not other genes, in db/db mice (Fig. 6G–J). Consistently, levels of hepatic CPT1α protein and serum 3-HB were reversed in Ad-CA MEK1 db/db mice injected with Ad-shATG7 (Fig. 6K and L). Furthermore, the decreased serum levels of ALT and AST of Ad-CA MEK1 db/db mice were also reversed by Ad-shATG7 injection (Fig. 6M and N). In addition, Ad-shATG7 had similar effects in db/db mice injected with control Ad-GFP (Supplementary Fig. 4).

Atg7 triggers autophagy and ameliorates liver lipid storage in leptin-deficient (ob/ob) mice (33). So, we injected Ad-shERK1/2 mice with Ad-ATG7 or Ad-GFP and examined whether ATG7 overexpression could ameliorate fatty liver in these mice. Ad-ATG7 overexpression (as confirmed by Western blot analysis) had no effects on body weight, liver weight, fat mass, or food intake but reversed the suppressive effects of Ad-shERK1/2 on hepatic autophagy as demonstrated by the changes of autophagy-related markers and ameliorated ERK1/2 inhibition–induced liver steatosis as confirmed by H-E and Oil red O staining (Fig. 7A–D). In addition, Ad-ATG7 reversed the increasing effects of Ad-shERK1/2 on liver TG and FFAs, but not liver TC, compared with control mice (Fig. 7E). Ad-shERK1/2 decreased serum TC but not TGs, and FFAs were increased by Ad-ATG7 (Fig. 7F). Furthermore, Ad-ATG7 reversed the suppressive effects of Ad-shERK1/2 on some of genes related to lipogenesis and fatty acid uptake in liver (Fig. 7G and H). Genes related to TG secretion and FAO, which were decreased by Ad-shERK1/2 in liver, were also largely reversed by Ad-ATG7 (Fig. 7I–J). Consistently, levels of hepatic CPT1α protein and serum 3-HB were reversed by Ad-ATG7 in Ad-shERK1/2 mice (Fig. 7K and L). Furthermore, the increased serum levels of ALT and AST of Ad-shERK1/2 mice were also reversed by Ad-ATG7 (Fig. 7M and N).

Previous work has shown that ERK1/2 inhibits phosphorylation of p38 (35), which is a negative regulator of autophagy (23,36). We found that p38 phosphorylation was increased in the livers of db/db mice, and the increased p38 phosphorylation was significantly decreased by Ad-CA MEK1 in the livers of db/db mice or primary hepatocytes (Fig. 8A and B). Consistent with these changes, inhibition of ERK1/2 increased p38 phosphorylation in the livers of WT mice and primary hepatocytes (Fig. 8C). To confirm a role of p38 in ERK1/2 regulated autophagy, we infected or transfected primary hepatocytes with Ad-shERK1/2 or p38α siRNA for 72 h. As predicted, p38 knockdown significantly increased expression of ATG7 and LC3-II and decreased expression of SQSTM1 in primary hepatocytes infected with Ad-shERK1/2 (Fig. 8D). Similar results were obtained with p38-specific inhibitor SB203580 (Supplementary Fig. 5).

The research concerning the metabolic functions of ERK1/2 is primarily focused on nonhepatic tissues, such as hypothalamus (37) and adipose tissue (14). In this study, we found that 1) ERK1/2 activity is compromised in the livers of db/db mice and activation of ERK1/2 upstream regulator MEK1 is sufficient to reduce lipid accumulation in these mice, 2) knockdown of ERK1/2 in C57/B6J WT mice results in liver steatosis, and 3) manipulation of ERK1/2 activity controls lipid accumulation in HepG2 cells and primary hepatocytes (Supplementary Fig. 6). These results strongly suggest that ERK1/2 serves as a potent regulator of lipid accumulation both in vivo and in vitro. However, hepatic ERK1/2 activity is increased in ob/ob and HFD mice, as demonstrated by previous work (17). We speculate that the different ERK1/2 activity might be caused by the complicated mechanisms involved in the control of ERK1/2 activity. In addition to well-known upstream cellular surface receptors and their downstream effectors, various hormone levels (including insulin, glucagon, and leptin) and nutritional status (including HFD and high-carbohydrate diet) and even self-signaling can influence ERK1/2 activity (17,38,39).

Our results show that hepatic expression of genes related to FAO (Pparα, Cpt1α) and TG export (ApoB, ApoE) was increased by activation of ERK1/2 in db/db mice or decreased by ERK1/2 knockdown in WT mice. The change in β-oxidation genes was accompanied by corresponding changes in CPT1α protein expression, serum 3-HB levels, and FAO rates measured under each treatment (Supplementary Fig. 7). In contrast, Jiao et al. (17) showed that the expression of these genes is decreased when hepatic ERK1/2 is enhanced in WT mice, which might be explained by the difference in fat content between WT and db/db mice. Because genes related to TG export were also stimulated by ERK1/2 activation, the possible contribution of increased TG export in ERK1/2 ameliorating liver steatosis should not be ignored. Liver lipid levels can also be largely controlled by insulin signaling in liver and lipolysis of fat tissue (40). Consistent with previous reports (16,17), we found that insulin signaling is enhanced in Ad-shERK1/2 mouse livers, and mRNAs and proteins related to lipolysis in the fat tissue of Ad-shERK1/2 mice are inhibited (Supplementary Fig. 8). These results suggest that insulin signaling and lipolysis in fat tissue may contribute to ERK1/2-regulated hepatic lipid storage in mice.

Autophagy is shown to stimulate lipid clearance (6,7,9), and manipulation of autophagic activity affects fat accumulation in liver (11–13). In this study, we found that ERK1/2 positively regulates hepatic autophagy both in vivo and in vitro. Furthermore, expression of ATG7, a key protein for autophagosome formation (34), is also regulated by ERK1/2. Although ATG7 protein can be posttranslationally regulated via altered calpain activity (33), we did not observe any effects of ERK1/2 on calpain activity (data not shown), suggesting that ERK1/2 regulation of ATG7 expression most likely occurs at transcriptional levels. The involvement of ATG7 in ERK1/2 amelioration of liver steatosis was then confirmed by the reversal effects of CQ treatment, and ATG7 knockdown on Ad-CA MEK1 ameliorated liver steatosis in db/db mice. We also found that ATG7 overexpression significantly reversed liver steatosis in Ad-shERK1/2 mice, as shown previously (33). Again, ERK1/2-regulated expression of hepatic genes related to FAO and TG export, as well as serum 3-HB levels, was affected by CQ and ATG7 under each different treatment.

Previous study has shown that ERK1/2 enhances autophagy by positively regulating BECN1 (21). We did not, however, observe any significant effects of ERK1/2 on Becn1 expression in vitro or in vivo. Interestingly, we found that p38 knockdown reversed the suppressive effects of Ad-shERK1/2 on autophagy and ATG7 expression in mouse primary hepatocytes. Consistent with our results, a previous study has shown that activation of p38 by osmotic stress inhibits autophagy in rat hepatocytes (36). A possible role for p38 in ERK1/2 regulation of ATG7-dependent autophagy in vivo and underlying mechanisms, however, remains to be further explored in the future.

A change in liver fat content is normally associated with a change in liver weight, and ERK1/2 has been shown to promote cell proliferation (15). Unexpectedly, we found that liver weight remains unchanged, though liver steatosis is ameliorated or induced by manipulation of hepatic ERK1/2 activity. We assume that the weight-lowering effects of the decreased liver TG on liver might be compensated by the enhanced cell proliferation as demonstrated by the increased expression of proliferation markers PCNA and cyclin D1 after ERK1/2 activation (Supplementary Fig. 9). Reciprocally, a balanced liver weight is reached between the increased liver TG and the decreased cell proliferation by ERK1/2 knockdown (Supplementary Fig. 9).

ERK1/2 is possibly activated in liver of young db/db and HFD mice, as insulin activates ERK1/2 (41) and serum insulin increases during this period (42). However, ERK1/2 may become inactivated in elder mice and result in decreasing hepatocyte proliferation. Given the close relationship between cell proliferation and lipid metabolism (43,44), changes in proliferation resulting from manipulation of ERK1/2 activity may also influence the lipid accumulation in db/db or HFD mice.

In addition to db/db mice, we also investigated the relevance of the role of ERK1/2 in the regulation of autophagy and liver steatosis in mice maintained on an HFD for 12 weeks. Because hepatic ERK1/2 activity is increased in HFD mice (17), we injected HFD mice with Ad-shERK1/2 and found that ERK1/2 knockdown significantly decreased hepatic autophagy and exacerbated liver steatosis in HFD mice (Supplementary Fig. 10). However, another study shows that ERK1/2 knockdown had no effect on hepatic TG in mice fed an HFD for 38 weeks (17). The lack of the response to ERK1/2 knockdown in the latter case could be due to the much longer period for HFD feeding: mice maintained on an HFD for 38 weeks already have much severe liver steatosis, which makes it difficult to further increase liver steatosis by ERK1/2 knockdown. A role for ERK1/2 in liver steatosis of ob/ob mice is implied by a previous study showing that global knockout of ERK1 reduces liver steatosis in ob/ob mice (45), but it is unclear whether it was due to a direct or indirect effect of ERK1 on liver steatosis. Therefore, a role for ERK1/2 in HFD and ob/ob mice remains to be investigated in the future.

We also explored the possible influence of ERK1/2 on insulin resistance, as liver steatosis is shown to be linked to insulin resistance. For example, it is shown that activation of ERK1/2 promotes insulin resistance and inhibition of ERK1/2 improves insulin sensitivity in WT mice and mice maintained on an HFD for 38 weeks (16,17), the effect of which might be mediated by ERK1/2 stimulation of serine phosphorylation of insulin receptor substrate 1 (46). On the other hand, enhancement of autophagy is reported to ameliorate insulin resistance and liver steatosis (33). Possibly due to the combined effects of activation of insulin receptor substrate 1 serine phosphorylation and stimulation of autophagy, we did not observe any significant effects of ERK1/2 activation on fed/fasting blood glucose, fasting serum insulin, and HOMA-IR index in db/db mice, while the insulin tolerance test results indicated that the insulin resistance is exacerbated in Ad-CA MEK1 db/db mice (Supplementary Fig. 11).

Amelioration of liver steatosis by inhibiting TG synthesis exacerbates liver damage (47). On the other hand, enhanced autophagy ameliorates liver injury and inhibited autophagy exacerbates liver damage (12,25). Therefore, the increased autophagy by ERK1/2 activation may account for the protective effects for liver damage in db/db mice, as demonstrated by the effects of ERK1/2 on serum ALT and AST under different treatment. Furthermore, lipid storage in hepatocytes, autophagy, and ERK1/2 itself are linked with apoptosis in a complex manner (15,48,49). However, we found that apoptosis is enhanced by ERK1/2 knockdown or inhibited by ERK1/2 activation in primary hepatocytes (Supplementary Fig. 12).

In summary, we have discovered a novel function of ERK1/2 in regulating hepatic lipid metabolism that is mediated by ATG7-dependent autophagy (Fig. 8E). These results provide novel insights into a physiological role of ERK1/2 in liver and theoretical basis for activating ERK1/2 as a potential treatment target for liver steatosis and its associated diseases. However, in evaluation of the beneficial effects of ERK1/2 activation on liver steatosis, its potential deleterious effects on insulin sensitivity and elevated serum TC, a risk factor for the development of heart attacks and strokes (50), should not be ignored.

Acknowledgments. The authors thank Haiyan Xu from Brown University for kindly providing Ad-CA MEK1. The authors thank Zhixue Liu from the Institute for Nutritional Sciences, Shanghai Institute for Biological Sciences, The Chinese Academy of Sciences, for helpful discussion.

Funding. This work was supported by grants from the National Natural Science Foundation (81130076, 81325005, 31271269, 81100615, and 81390350), the Ministry of Science and Technology of China (973 Program 2010CB912502), the Basic Research Project of Shanghai Science and Technology Commission (13JC1409000), and the International S&T Cooperation Program of China (Singapore 2014DFG32470) and by research supported by the The Chinese Academy of Sciences/State Administration of Foreign Experts Affairs international partnership program for creative research teams. F.G. was also supported by the One Hundred Talents Program of the Chinese Academy of Sciences.

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

Author Contributions. Y.X. and H.L. researched data and wrote, reviewed, and edited the manuscript. J.Y., Z.Z., T.X., C.W., K.L., J.D., and Y.G. researched data. F.X. and Y.C. contributed to writing and helpful discussion. S.C. provided research material. F.G. directed the project, contributed to discussion, and wrote, reviewed, and edited the manuscript. F.G. 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.