Abstract
Fatty liver disease is a serious health problem worldwide and is the most common cause for chronic liver disease and metabolic disorders. The major challenge in the prevention and intervention of this disease is the incomplete understanding of the underlying mechanism and thus lack of potent therapeutic targets due to multifaceted and interdependent disease factors. In this study, we investigated the role of a signaling adaptor protein, GRB2-associated-binding protein 2 (Gab2), in fatty liver using an animal disease model. Gab2 expression in hepatocytes responded to various disease factor stimulations, and Gab2 knockout mice exhibited resistance to fat-induced obesity, fat- or alcohol-stimulated hepatic steatosis, as well as methionine and choline deficiency-induced steatohepatitis. Concordantly, the forced expression or knockdown of Gab2 enhanced or diminished oleic acid (OA)- or ethanol-induced lipid production in hepatocytes in vitro, respectively. During lipid accumulation in hepatocytes, both fat and alcohol induced the recruitment of PI3K or Socs3 by Gab2 and the activation of their downstream signaling proteins AKT, ERK, and Stat3. Therefore, Gab2 may be a disease-associated protein that is induced by pathogenic factors to amplify and coordinate multifactor-induced signals to govern disease development in the liver. Our research provides a novel potential target for the prevention and intervention of fatty liver disease.
Introduction
Alcoholic fatty liver disease (AFLD) and non-alcoholic fatty liver disease (NAFLD) are major global health problems and the most common causes for chronic liver disease and metabolic disorders (Purohit et al., 2009; Birkenfeld and Shulman, 2014). Alcoholic or non-alcoholic pathogenic factors cause hepatic impairment and lipid metabolism disruption, first inducing hepatic steatosis (Malaguarnera et al., 2009; Purohit et al., 2009). If the accumulation of excessive hepatic fat is not eliminated, continual disease factors and excessive free fatty acids (FFA) cause steatohepatitis, followed by cirrhosis and hepatocellular carcinoma (Malaguarnera et al., 2009; Purohit et al., 2009). In general, 8%–20% of heavy drinkers with steatosis will develop steatohepatitis, and 10%–30% of steatosis in NAFLD patients can progress to non-alcoholic steatohepatitis (NASH) (Vuppalanchi and Chalasani, 2009; Birkenfeld and Shulman, 2014; Duan et al., 2014). Steatohepatitis can easily result in cirrhosis and increase the risk of metabolic complications due to insulin resistance (IR), hepatic damage, and inflammation (Vuppalanchi and Chalasani, 2009). Therefore, hepatic steatosis and steatohepatitis are key steps in the development of fatty liver disease (Malaguarnera et al., 2009; Purohit et al., 2009; Birkenfeld and Shulman, 2014; Suk et al., 2014).
Fat accumulation is the result of an imbalance between fatty acid intake/lipid synthesis and lipid oxidation/transport in the liver (Malaguarnera et al., 2009; Purohit et al., 2009). Disease factors and excessive FFA stimulate or suppress many transcription factors and enzymes to govern these processes (Purohit et al., 2009; Birkenfeld and Shulman, 2014). Continual disease factors also induce IR, oxidative stress, abnormalities in mitochondrial oxidation, detrimental inflammation, and impairment of glucose homeostasis, thus aggravating the phenotypes of hepatocellular damage and lipid accumulation and facilitating pathological development (Tilg and Moschen, 2008; Malaguarnera et al., 2009; Purohit et al., 2009). In the development of fatty liver disease, various factors, including insulin, FFA, reactive oxygen species, and inflammatory cytokines, are either positive or negative mediators that trigger several cytoplasmic signals, such as Ras/MAPK, PI3K/AKT, AMPK, JNK-STAT, NF-κB, mTOR, and Socs3 (Schwabe and Brenner, 2007; Aghazadeh and Yazdanparast, 2010; Sengupta et al., 2010; Matsuda et al., 2013; Gautheron et al., 2014; Hsu et al., 2014), which coordinate to produce an integrated biological effect (Malaguarnera et al., 2009; Purohit et al., 2009). However, little is known how these signals are integrated, which is crucial for understanding the disease progression and developing effective therapeutic strategies (Farrell et al., 2013; Birkenfeld and Shulman, 2014; Suk et al., 2014).
The GRB2-associated binder (Gab) family comprises scaffolding/adaptor proteins that recruit cytoplasmic signaling proteins to amplify and integrate the bioeffects of various extracellular signals (Goldstein et al., 2006). Gab1 and Gab2 collect several SH2- and SH3-containing proteins, such as PI3K, GRB2, Shp2, and PLCγ, to transduce and integrate signals emanating from growth factor receptors including EGF, insulin, IGF-1, PDGF, and leptin (Nishida and Hirano, 2003; Nakaoka and Komuro, 2013). These growth factors and signaling proteins regulate glucose and lipid metabolism (Aghazadeh and Yazdanparast, 2010; Krycer et al., 2010; Matsuda et al., 2013). In the liver, Gab1 balances hepatic insulin action, and deletion of Gab1 enhances hepatic insulin sensitivity (Bard-Chapeau et al., 2005). This evidence indicates that Gab proteins may act as signaling scaffolding proteins to integrate various signals in hepatic tissue. However, the regulation of Gab proteins in the liver and their roles in liver disease remain unclear (Nakaoka and Komuro, 2013).
Gab1 is stably expressed in the body and is a required functional protein for normal life (Bard-Chapeau et al., 2005). Deletion of the Gab1 gene induces embryonic death, and abnormalities in Gab1 cause defects and diseases of several organs in mouse models (Bard-Chapeau et al., 2005; Nakaoka and Komuro, 2013). Gab2 is selectively expressed in some organs, and a deficiency of Gab2 has no effect on the lifespan of mice (Gu et al., 2001; Nakaoka and Komuro, 2013). It is highly expressed in many cancers and promotes cancer development (Bentires-Alj et al., 2006; Ke et al., 2007; Adams et al., 2012; Nakaoka and Komuro, 2013). Abnormal Gab2 protein signaling is linked to Alzheimer's disease (Zhong et al., 2011). In both cell lines and animal tissues of the liver, we found that fat or alcohol markedly increased Gab2 but not Gab1 protein expression. Thus, we infer that Gab2 may be a disease-associated protein that coordinates various disease signals to promote liver disease.
To test this hypothesis, we treated Gab2 gene knockout mice with alcohol or the high-fat diet (HFD). Deletion of Gab2 ameliorated fat- or alcohol-induced fatty liver. Furthermore, PI3K, Socs3, and Shp2 were recruited by Gab2 to activate their downstream pathways AKT, Stat3, and ERK, respectively. Therefore, our study suggests that Gab2 may be a novel key disease-associated protein in fatty liver and an effective therapeutic target for fatty liver disease.
Results
Deletion of Gab2 diminishes hepatic pathology in mouse models
To explore the role of Gab2 in fatty liver in vivo, we generated several disease models with Gab2 gene knockout mice (Ke et al., 2007) and assessed the effect of Gab2 deletion on disease development. First, we constructed a non-alcoholic fatty liver disease (NAFLD) mouse model by feeding mice the HFD for 24 weeks. Mice fed the HFD were fat (Figure 1A, left-hand images), and the average body weight of HFD-fed mice was obviously higher than that of mice fed a standard diet (SD) (P < 0.05) (Figure 1A, right graph). However, HFD-induced obesity was distinctly reduced in Gab2−/− mice (P < 0.05) (Figure 1A), and the body weight of HFD-fed Gab2−/− mice increased more slowly than that of heterozygous mice (Gab2+/−) fed the HFD (Supplementary Figure S1A). Liver weight and hepatic index were increased in the HFD-fed Gab2+/− mice, and the liver became lighter in color (Figure 1B and Supplementary Figure S1B). Liver tissue sections were stained with H&E, which revealed the accumulation of many lipid droplets in the liver of HFD-fed Gab2+/− mice (Figure 1C, fatty vacuoles). However, deletion of Gab2 protected against hepatic lipid accumulation. Liver weight, hepatic index, and the area of fatty vacuoles were all lower in Gab2−/− mice than Gab2+/− mice (Figure 1B and C). Furthermore, HFD-fed Gab2+/− mice exhibited higher levels of hepatic triglycerides (TG) as well as serum total cholesterol (T-CHO) and glutamate pyruvate transaminase ALT, which are important markers of NAFLD. However, the increases in hepatic TG, serum T-CHO, and ALT were clearly reduced in HFD-fed Gab2−/− mice (Figure 1D). These results demonstrate that deletion of Gab2 alleviated lipid accumulation in NAFLD mice.
Deletion of Gab2 protects against hepatic pathological development induced by fat in different NAFLD models. (A−C) Phenotypes of mice fed the standard diet (SD) or high-fat diet (HFD) for 24 weeks. Left panels show the representative images of the body (A), liver (B), and fat accumulation (vacuoles, H&E staining; C), respectively. Right panels show the quantifications for body weight (A), liver weight (B), and the number of fat vacuoles (C), respectively. (D) Concentrations of hepatic triglycerides (TG), total cholesterol (T-CHO), and ALT in the sera of mice after 12 h of fasting. (E) Impaired function and inflammation in the liver tissue from NASH mice fed the methionine- and choline-deficient (MCD) diet was revealed by H&E staining. Arrows indicate inflammatory foci, and asterisks designate lipid droplets. (F) Concentrations of TG, T-CHO, ALT, and TNFα in the sera of mice after 12 h of fasting. All quantified data are the average values for all mice in more than three replicate experiments and are expressed as mean ± SEM. *P < 0.05; ^^, **P < 0.01. ^ indicates treatment difference; * indicates group difference. Scale bar, 25 μm.
To further determine the effects of Gab2 deletion on steatohepatitis, we generated a NASH mouse model fed a methionine- and choline-deficient (MCD) diet for 8 weeks (Vuppalanchi and Chalasani, 2009). Mice fed the MCD diet exhibited a serious steatohepatitis phenotype. Compared with mice fed methionine and choline supplements (MCS), the body weights of Gab2+/− mice fed the MCD diet were markedly decreased, but the hepatic index was increased (Supplementary Figure S1C and D). H&E staining clearly revealed lipid accumulation and inflammatory cell infiltration in the liver of Gab2+/− mice fed the MCD diet (Figure 1E, right upper image), and the concentrations of hepatic TG, serum T-CHO, serum ALT, and the inflammatory cytokine TNFα were significantly increased in these mice (Figure 1F, black column). The pathological phenotypes were strongly improved in Gab2−/− mice fed the MCD diet. Hepatic color was nearly recovered (Supplementary Figure S1D), and very few lipid droplets and inflammatory cell infiltrations were observed in their liver tissues (Figure 1E, right lower image). The quantity of lipid droplets was markedly smaller than that in Gab2+/− mice fed the MCD diet. The levels of hepatic TG, serum T-CHO, serum ALT, and TNFα nearly decreased to the same levels in mice fed the MCS diet (Figure 1F, gray column). In order to clarify the inflammation in liver tissue, we also checked the activity of NF-κB with IHC staining. NF-κB protein entered into the nucleus in Gab2+/− mice fed the MCD diet, while nuclear NF-κB was little in Gab2−/− mice fed the MCD diet (Supplementary Figure S1E), indicating that deletion of Gab2 diminished MCD-induced inflammatory reaction in the liver. These results suggest that the absence of Gab2 restrained the development of NASH caused by MCD.
To evaluate the effect of Gab2 on lipid metabolic disorders caused by alcohol, we established acute and chronic AFLD mouse models by feeding with ethanol for 12 h or 6 weeks, respectively. Both acute and chronic ethanol feeding increased liver lipid production (Figure 2A and B, Oil Red O staining, right upper image). Serum ALT and TG, and hepatic TG levels were increased in chronic AFLD mouse model (Figure 2C, black column, Gab2+/− mice). However, no differences in food intake, serum AST, and hepatic T-CHO levels were observed among the groups in chronic AFLD model (Supplementary Figure S1F and G). As expected, the hepatic disorder was limited in ethanol-fed Gab2−/− mice. The production of lipid droplets in the liver was markedly decreased (Figure 2A and B, right lower image compared with the right upper image), and the concentrations of ALT and TG in both serum and hepatic tissues were nearly identical to those of normal control mice (Figure 2C, gray column). In addition, the effects of deletion of Gab2 on alcohol-induced hepatic steatosis were confirmed by comparing the phenotypes of Gab2−/− mice with that of wild-type mice in the chronic AFLD model (Supplementary Figure S2A and B). Therefore, the absence of Gab2 reduced the liver impairment induced by alcohol.
Deletion of Gab2 blocks hepatic steatosis induced by alcohol in vivo. (A and B) Lipid production was indicated by Oil Red O staining of the liver tissue from mice subjected to acute (A) or chronic (B) ethanol feeding. (C) Concentrations of serum ALT and TG, hepatic TG, and hepatic T-CHO in the chronic AFLD model. All quantified data are the average values for all mice in more than three replicate experiments and are expressed as mean ± SEM. Eth, ethanol; ^, *P < 0.05; ^^, **P < 0.01. ^ indicates treatment difference; * indicates group difference. Scale bar, 25 μm.
Taken together, these results indicate that Gab2 deficiency restricted in vivo fatty liver pathologies, i.e. hepatic lipid accumulation and steatohepatitis.
Gab2 is crucial for lipid production induced by oleic acid and ethanol in HepG2 cells
To further investigate the effects of Gab2 on the pathological phenotype in hepatic cells, we constructed a stable HepG2 cell line that highly expresses EGFP-Gab2 (Figure 3E) and evaluate oleic acid (OA)- or ethanol-induced lipid accumulation, which is the main phenotype in fatty liver, in these cells. EGFP- or EGFP-Gab2-expressing cells were treated with 1.0 mM OA for 12 h and stained with Oil Red O for lipid droplets (Figure 3A, indicated by arrows). The quantity of lipids was notably larger in EGFP-Gab2 cells than EGFP cells (P < 0.01) (Figure 3A, right-hand images; Figure 3B, left graph). When cells were stimulated with 100 mM ethanol for 12 h, more lipid droplets were observed in EGFP-Gab2 cells than EGFP cells (Figure 3C, right-hand images; Figure 3D, left graph). However, no lipids were observed in cells cultured in normal medium (Figure 3A and C, left-hand images), indicating that Gab2 overexpression did not affect lipid synthesis. Furthermore, TG secretion was significantly higher in EGFP-Gab2 cells than EGFP cells when cultured with 1.0 mM OA or 100 mM ethanol (Figure 3B and D, right graphs). These results suggest that Gab2 overexpression accelerated the generation of lipid droplets induced by OA or ethanol.
Gab2 promotes lipid production induced by OA or ethanol in HepG2 cells in vitro. (A−D) Cells were treated with 1.0 mM OA (A and B) or 100 mM ethanol (C and D) for 12 h. Lipid production was assessed by Oil Red O staining. (A and C) Representative images of each assay are shown. Arrowheads designate lipid droplets. (B and D) The OD value of Oil Red O was measured in an automatic microplate reader at 450 nm, and the TG concentration in the culture medium was assayed by ELISA. (E) Gab2 was overexpressed in the stable HepG2 cell line expressing EGFP-Gab2. α-Tubulin was used as a total protein control. All quantified data are the average values from more than three replicate experiments and are expressed as mean ± SEM. Eth, ethanol; Con, normal medium; *P < 0.05; ^^, **P < 0.01. ^ indicates treatment difference; * indicates group difference. Scale bar, 25 μm.
When small interfering RNA (siRNA) was used to decrease Gab2 protein levels in HepG2 cells (Figure 4E), lipid droplets and TG in siRNA-transfected cells treated with 1.0 mM OA or 100 mM ethanol for 12 h were examined (Figure 4A–D). Knockdown of Gab2 still had no apparent effect on basal lipid production (Figure 4A and C, left-hand images) but clearly suppressed OA- or ethanol-induced lipid production in HepG2 cells (Figure 4A and C, right-hand images). TG secretion was also decreased by Gab2 protein deficiency in HepG2 cells (Figure 4B and D, right graphs). Notably, restoration of Gab2 protein in primary hepatocytes from Gab2−/− mice, using a lentivirus containing Gab2 mRNA, reactivated TG formation induced by OA or ethanol (Supplementary Figure S3).
Decreased expression of Gab2 suppresses lipid generation induced by OA or ethanol in HepG2 cells. Gab2 expression was knocked down with a specific siRNA in HepG2 cells, followed by stimulation of the cells with 1.0 mM OA or 100 mM ethanol for 12 h. (A and C) Oil Red O staining of lipid droplets induced by OA (A) or ethanol (C). Arrowheads designate lipid droplets. (B and D) The OD value of Oil Red O and TG concentration in the culture medium were measured. (E) The Gab2 protein was detected with a specific antibody in HepG2 cells. Eth, ethanol; Con, normal medium; siCon, siRNA of control protein; siGab2, siRNA of Gab2; *P < 0.05; ^^, **P < 0.01. ^ indicates treatment difference; * indicates group difference. Scale bar, 25 μm.
In summary, these data suggest that Gab2 is fundamental to lipid generation in hepatic cells exposed to disease factors, i.e. fat and alcohol.
Gab2 but not Gab1 or Shp2 is highly expressed in pathological hepatocytes
In the human hepatic carcinoma cell line HepG2, the expression of Gab2, but not Gab1, was markedly increased when cells were exposed to 1.0 mM OA or 100 mM ethanol (Figure 5A and B). In primary hepatocytes from Gab2+/− mice, OA or ethanol also increased Gab2 protein levels (Figure 5C). Furthermore, the protein level of Gab2 but not Gab1 was increased in the mouse liver from fatty liver models (Figure 5D). IHC staining results demonstrated that Gab2 expression was increased in whole liver tissues from all mouse fatty liver models in this study (Supplementary Figure S4). Interestingly, the expression of Gab1 was not increased in primary cells isolated from Gab2−/− mice and treated with OA or ethanol or hepatic tissues from HFD-fed Gab2−/− mice (Figure 5C and D). Previous studies have demonstrated that Gab family proteins are regulated by phosphorylation (Gu and Neel, 2003). Tyr452 phosphorylation site of Gab2 is a potential binding site of p85, a regulatory subunit of PI3 kinase. In addition, AKT phosphorylates Gab2 at Ser159. The phosphorylation on these sites of Gab2 inhibits its activity and downstream signal amplification (Lynch and Daly, 2002). We observed that pY452-Gab2 rather than pS159-Gab2 was activated by OA or ethanol in hepatocytes (Figure 5A−C). In addition, the expression of Shp2, a partner protein of Gab1 and Gab2, was not affected by either OA or ethanol treatment or deletion of Gab2 (Figure 5).
Expression of Gab2 but not Gab1 or Shp2 is increased in pathological hepatic cells. (A−C) Protein expression and phosphorylation of Gab2 in hepatic cells cultured in vitro. HepG2 cells were treated with 1.0 mM OA (A) or 100 mM ethanol (B) for 12 h, and primary liver cells from Gab2 knockout (Gab2−/−) and heterozygous control (Gab2+/−) mice were exposed to 1.0 mM OA or 100 mM ethanol for 24 h (C), respectively. (D) Protein levels of Gab2, Gab1, and Shp2 in the liver tissue from NAFLD (HFD for 24 weeks) and AFLD (ethanol for six weeks) mouse models. β-actin or Gapdh was used as a total protein control. The protein levels of Gab2, Gab1, Shp2, and the background protein β-actin or Gapdh were analyzed by western blot using specific antibodies. Representative blots are shown in the panels. The average relative grayscale values normalized with the total control protein were determined for all experiments and expressed as mean ± SEM. Eth, ethanol. ^^, **P < 0.01. ^ indicates treatment difference; * indicates group difference.
These data suggest that Gab2 expression in hepatocytes is specifically stimulated by pathogenic factors and Gab2 may be an accessory protein for disease development in the liver.
Gab2 is triggered by fat or alcohol to activate the cytoplasmic signaling proteins AKT, ERK, and Stat3 and the transcription factor SREBP-1c in hepatocytes
Lipid synthesis and accumulation are the initial steps and primary phenotypes of fatty liver (Malaguarnera et al., 2009; Purohit et al., 2009). To elucidate the mechanism of Gab2 regulation in fatty liver disease, we examined the activities of several cytoplasmic signaling proteins and transcription factors involved in lipid generation in NAFLD and AFLD mouse models and OA- or ethanol-treated HepG2 cells.
HFD clearly increased the phosphorylation of AKT and ERK in liver tissue, but deletion of Gab2 suppressed both basal and fat-induced activation of AKT and ERK, with the total protein levels of AKT and ERK not changed (Figure 6A). Similar findings were obtained in primary hepatocytes isolated from Gab2−/− mice and treated with 1.0 mM OA for 24 h (Figure 6B). In contrast, overexpression of Gab2 in HepG2 cells enhanced basal and OA-induced activation of AKT and ERK (Figure 6C).
Gab2 mediates the activation of the cytoplasmic signaling proteins AKT and ERK and the transcription factor SREBP-1 in hepatic lipid formation induced by fat. (A−C) The protein levels and phosphorylation of AKT and ERK and expression of mature SREBP-1c in the mouse liver tissue from NAFLD model (A), primary liver cells isolated from Gab2−/− mice and treated with 1.0 mM OA for 24 h (B), or EGFP-Gab2 HepG2 cells stimulated by 0.5 mM OA for 12 h (C) were analyzed by western blot using specific antibodies. β-actin or α-Tubulin was used as a total protein control. (D and E) Transcriptional activity of SREBP-1c in HepG2 cells transfected with siRNA against Gab2 (D) or EGFP-Gab2 HepG2 cells (E) treated with 1.0 mM OA for 12 h was analyzed by the luciferase reporter assay. Representative blots are shown in the panels. The average relative grayscale values normalized with the total control protein were obtained from all experiments and expressed as mean ± SEM. mSREBP-1c, mature SREBP-1c; pSREBP-1c, precursor protein of SREBP-1c; Con, normal medium; siCon, siRNA of control protein; siGab2, siRNA of Gab2. ^, *P < 0.05; ^^, **P < 0.01. ^ indicates treatment difference; * indicates group difference.
SREBP-1 isoform c is a key transcription factor in lipid synthesis that regulates cholesterol and TG synthesis by activating de novo fatty acid biosynthesis genes ACC1 and FAS (Xiao and Song, 2013). Mature SREBP-1c and ACC were increased in the liver tissue of NAFLD mice, but deletion of Gab2 decreased both basal and fat-induced protein levels of mature SREBP-1c and ACC (Figure 6A). This was further confirmed in primary hepatocytes treated with 1.0 mM OA for 24 h (Figure 6B). As expected, overexpression of Gab2 in HepG2 cells increased basal (Figure 6C) and OA-induced both mature SREBP-1c and ACC protein levels (Figure 6C). In addition, we determined the levels of mature SREBP-1c in the nuclear extract of HepG2 cells with high or low levels of Gab2 protein, which further confirmed that Gab2 mediated the activation of SREBP-1c (Supplementary Figure S5A and B). The SREBP-1c precursor protein level did not change in all experiments, indicating that SREBP-1c gene expression was not affected by Gab2 (Figure 6A–C). To further elucidate the effect of Gab2 on the transcriptional activity of SREBP-1c, we employed a pSyn SRE-luciferase reporter construct containing an SREBP-1-responsive element (SRE). Overexpression of Gab2 in EGFP-Gab2 HepG2 cells markedly increased (Figure 6E), whereas knockdown of Gab2 suppressed (Figure 6D) the OA-induced transcriptional activity of SREBP-1c.
Similarly, deletion of Gab2 eliminated ethanol-induced phosphorylation of AKT and the maturation of SREBP-1c in the liver tissue of chronic ALFD mice (Figure 7A) and primary hepatocytes treated with 100 mM ethanol for 24 h (Figure 7B). Using the pSyn SRE-luciferase reporter, Gab2 was also observed to mediate the ethanol-induced transcriptional activity of SREBP-1c (Figure 7C). Although Gab2 deficiency did not inhibit the activation of ERK in the liver tissue from chronic ALFD mice (Figure 7A) or hepatocytes treated with ethanol for 24 h (Figure 7B; Supplementary Figure S5C), knockdown of Gab2 clearly reduced ERK activation induced by acute ethanol treatment (12 h) (Supplementary Figure S5C).
Gab2 is involved in the activation of AKT, Stat3, and SREBP-1 in hepatic lipid formation induced by alcohol. (A, B, D, and E) Phosphorylation of AKT, ERK, Stat3, Socs3 and expression of mature SREBP-1c in the liver tissue from chronic AFLD mice fed ethanol for six weeks (A), primary liver cells isolated from Gab2−/− mice and treated with 100 mM ethanol for 24 h (B and D), or EGFP-Gab2 HepG2 cells stimulated with 100 mM ethanol for 12 h (E) were measured by western blot with specific antibodies. (C) The transcriptional activity of SREBP-1c in HepG2 cells transfected with siRNA against Gab2 or EGFP-Gab2 HepG2 cells treated with 100 mM ethanol for 12 h was analyzed using the luciferase reporter assay. The background protein Gapdh or β-actin was used as a total protein control. Representative blots are shown in the panels. The average relative grayscale values normalized with the total control protein were obtained from all experiments and are expressed as mean ± SEM. mSREBP-1c, mature SREBP-1c; Eth, ethanol; Con, normal medium; siCon, siRNA of control protein; siGab2, siRNA of Gab2. ^, *P < 0.05; ^^, **P < 0.01. ^ indicates treatment difference; * indicates group difference.
Notably, we also observed that Stat3 was associated with Gab2 regulation of lipid production. In liver tissue and primary hepatocytes, ethanol reduced the phosphorylation of Stat3, which was rescued by deletion of Gab2 (Figure 7A and D). Suppressor of cytokine signaling 3 (Socs3) is a suppressor of Stat3 activation (Mori et al., 2004). Ethanol stimulated the expression of Socs3 in primary hepatocytes (Figure 7D), which was inhibited by deletion of Gab2 (Figure 7D). These results suggest that the Socs3−Stat3 signaling pathway is recruited by Gab2 to transduce the pathogenic stimulation.
The suppressive effects of Gab2 deficiency on ethanol-activated cytoplasmic signaling proteins and transcription factors were further confirmed by decreasing Gab2 protein level with specific siRNA in HepG2 cells (Supplementary Figure S5C–E). In contrast, overexpression of Gab2 in HepG2 cells markedly increased the basal phosphorylation of these cytoplasmic signaling proteins and transcription factors (Figures 6C and 7E) and moderately enhanced their activation induced by OA or ethanol (Figures 6C and 7E).
These results indicate that Gab2 activates AKT, ERK, and Stat3 to exert the pathogenic effects of fat or alcohol on liver lipid synthesis.
Gab2 activates the cytoplasmic signaling pathways PI3K/P85α−AKT and Socs3−Stat3
Gab2 usually recruits proteins containing the SH2 or SH3 domain to activate cytoplasmic signals (Nishida and Hirano, 2003). Thus, we investigated the interaction between such proteins and Gab2 in HepG2 cells stimulated with 1.0 mM OA or 100 mM ethanol for 12 h. PI3K is the first identified Gab2-binding cytoplasmic signaling protein (Bouscary et al., 2001; Gu et al., 2001). In this study, protein complexes in cell lysates were pulled down using the Gab2 antibody, and the P85α subunit of PI3K was identified in the protein complexes containing Gab2 in HepG2 cells (Figure 8A, third band in upper first panel; Figure 8F, first band in upper second panel). The detected P85α signal was increased when cells were exposed to OA or ethanol (Figure 8A, last band in upper first panel; Figure 8F, third band in upper second panel), indicating that OA or ethanol promotes more signal protein PI3K-P85α binding to Gab2. Concordantly, the interaction between PI3K-P85α and Gab2 was enhanced in cells overexpressing Gab2 (Figure 8B) but diminished when Gab2 was knocked down in HepG2 cells (Figure 8C). Furthermore, the AKT inhibitor GSK690693 inhibited the transcriptional activity of SREBP-1c induced by OA or ethanol in both EGFP and EGFP-Gab2 cells (Figure 8D). The PI3K inhibitor LY294002 also suppressed the transcriptional activity of SREBP-1c induced by OA or ethanol (Figure 8E). These findings suggest that the PI3K−AKT signaling pathway may be activated by Gab2 to regulate the effects of fat or alcohol on lipid synthesis through activating SREBP-1c.
OA- or ethanol-induced interaction between Gab2 and PI3K/P85α or Socs3 in HepG2 cells. (A−C) Interactions between Gab2 and the PI3K-P85α, Shp2, or Socs3 in HepG2 cells with normal expression (A), overexpression (B), or knockdown of Gab2 (C) treated with OA for 12 h. (D) The activity of the SRE reporter was measured in EGFP and EGFP-Gab2 cells treated with OA or ethanol alone or in combination with the AKT inhibitor GSK690693. (E) The effects of the PI3K inhibitor LY294002 on the SRE reporter activity stimulated by OA or ethanol. (F) The interaction between Gab2 and PI3K-P85α or Socs3 in HepG2 cells stimulated by 100 mM ethanol for 12 h. (G) A schematic representation of the effects of Gab2 on the signal transduction induced by disease factors in fatty liver disease. The dotted line indicates the pathway described in the literature. The interactions between proteins were analyzed by immunoprecipitation (IP) followed by western blot using specific antibodies. The background protein α-Tubulin was used as a total protein control. Representative blots are shown in the panels. The average relative grayscale values normalized with the total control protein were obtained from all experiments and expressed as mean ± SEM. Eth, ethanol; Con, normal medium; siCon, siRNA of control protein; siGab2, siRNA of Gab2. **P < 0.01.
Interestingly, we demonstrated that Socs3 might be a novel interacting protein of Gab2. In HepG2 cells, ethanol stimulated the expression of Socs3 (Figure 7E) and induced Socs3 binding to Gab2 (Figure 8F, third band in upper first panel). we also found that OA enhanced the interaction between Gab2 and Socs3(Figure 8A), although OA did not stimulate Socs3 expression in HepG2 cells. Socs3 specifically activates AKT or suppresses Stat3 and contributes to IR and lipid production in fatty liver pathology (Mori et al., 2004; Malaguarnera et al., 2009). Therefore, our results identified a new pathway employed by Gab2 to activate AKT or suppress Stat3 via SCOS3.
Shp2 is a partner protein of Gab2, and these proteins bind together to activate ERK, AKT, or Stat3 (Huang et al., 2014). Although Shp2 expression did not respond to pathological factors (Figure 5), the interaction between Gab2 and Shp2 was enhanced by both OA and ethanol (Figure 8A and F). The Shp2 inhibitor PHPS1 suppressed OA- or ethanol-induced phosphorylation of ERK (Supplementary Figure S6B). However, neither the ERK inhibitor PD98059 nor the Shp2 inhibitor PHPS1 blocked the SRE-luciferase signal induced by the stimulus (Supplementary Figure S6A and C), which indicates that the Shp2/ERK pathway is not involved in the transcriptional activity of SREBP-1c induced by OA and ethanol. We found that OA reduced the phosphorylation of insulin receptor substrate 1 (IRS1) in HepG2 cells (Supplementary Figure S7A), which was restored when Gab2 was knocked down by siRNA (Supplementary Figure S7B). Conversely, overexpression of Gab2 suppressed the phosphorylation of IRS1 and aggravated the inhibitory effect of OA (Supplementary Figure S7C). These results suggest that deletion of Gab2 may improve IR.
In summary, pathogenic factors such as fat and alcohol induce the upregulation of Gab2 that recruits the cytoplasmic signaling proteins PI3K, Socs3, and Shp2 to activate downstream AKT, Stat3 and ERK signalings, which coordinate to affect lipid metabolism or other pathological effects in fatty liver disease (Figure 8G).
Discussion
Although most hepatic steatosis does not impair liver function and can be reversed, it is usually the beginning of fatty liver disease (Birkenfeld and Shulman, 2014; Duan et al., 2014; Suk et al., 2014). Continual disease factors and excessive FFA can cause steatohepatitis, which is a pathological phenotype of the liver that causes hepatic damage and induces the development of fatty liver disease (Malaguarnera et al., 2009; Purohit et al., 2009; Vuppalanchi and Chalasani, 2009; Birkenfeld and Shulman, 2014; Duan et al., 2014; Suk et al., 2014). The present study demonstrated that deletion of Gab2 protected against fat- or alcohol-induced hepatic steatosis and MCD-induced steatohepatitis in mice. This discovery provides a novel target for the prevention and intervention of fatty liver disease.
Accumulation of hepatic fat is the primary phenotype in ALD and NAFLD (Malaguarnera et al., 2009; Birkenfeld and Shulman, 2014; Suk et al., 2014). Subsequent pathologies depend on the coordination of different signaling pathways in the fat-loaded liver (Malaguarnera et al., 2009; Purohit et al., 2009). Elucidating this process will enable a complete understanding of signaling regulation in fatty liver disease (Vuppalanchi and Chalasani, 2009; Birkenfeld and Shulman, 2014; Suk et al., 2014). Therefore, we concentrated on the mechanism of Gab2 regulation of hepatic steatosis. Although fat accumulation involves fatty acid intake, lipid synthesis, and lipid oxidation and transport, abnormal lipid synthesis is most important in fatty liver disease (Malaguarnera et al., 2009; Purohit et al., 2009). SREBP-1 is a key transcription factor that regulates the product of hepatocyte cholesterol and TG by increasing enzymes involved in de novo fatty acid biosynthesis, such as ACC1 and FAS (Krycer et al., 2010; Zeng et al., 2012; Matsuda et al., 2013). Therefore, SREBP-1 and ACC1 were considered the molecular markers of fatty acid biosynthesis in our study. In both liver tissue and hepatocytes, deletion of Gab2 reduced the activity of SREBP-1 and the protein expression of ACC1 stimulated by fat or alcohol, thus confirming the regulation of fat accumulation by Gab2 in fatty liver disease.
Gab2 is a signaling adaptor protein and usually integrates various signaling pathways to govern disease progress in cancer and AD (Adams et al., 2012; Jin et al., 2013; Nakaoka and Komuro, 2013). Here, using SREBP-1 as a marker of lipid synthesis, we demonstrated that Gab2 affects several signaling pathways to orchestrate fat production in the liver in response to fat or alcohol stimulation. First, our results confirmed the interaction between Gab2 and PI3K in hepatocytes. PI3K usually triggers the AKT signaling pathway, and PI3K−AKT is the key signaling pathway in lipid and glucose metabolism in the liver (Malaguarnera et al., 2009; Matsuda et al., 2013; Suk et al., 2014). This pathway mediates insulin-, fat-, and alcohol-induced lipid synthesis by activating SREBP-1. PI3K/AKT is recruited by Gab2 in numerous types of non-liver cells (Gu et al., 2001; Hunzicker-Dunn et al., 2012). Here, we demonstrated that Gab2 also employs the PI3K−AKT pathway to transduce pathogenic signals in hepatic steatosis. Furthermore, we identified Socs3 as a new binding protein of Gab2. In the development of NAFLD, Socs3 is activated by FFA to decrease insulin sensitivity and promote SREBP-1 activity in the liver (Mori et al., 2004; Malaguarnera et al., 2009). Socs3 can activate AKT or suppress Stat3 (Mori et al., 2004). Therefore, in addition to the partner protein PI3K, Socs3 represents a previously unidentified signaling protein employed by Gab2 to trigger AKT or Stat3 in fatty liver disease.
Shp2 is a partner protein of Gab1 or Gab2 in a number of tissues and usually triggers the MAPK−ERK pathway in vivo (He et al., 2012; Huang et al., 2014; Kaneshiro et al., 2014). The Shp2−ERK pathway mediates the regulation of hepatic insulin activity by Gab1, and deletion of Shp2 in the liver blocks lipid accumulation in HFD-fed mice (Nagata et al., 2012). However, in the present study, we observed that Gab2 did not employ the Shp2−ERK pathway to regulate lipid synthesis. Although an inhibitor of Shp2 suppressed the activation of ERK in HepG2 cells treated with OA or ethanol, Shp2 expression was not increased in hepatocytes exposed to a pathological factor. Furthermore, inhibitors of ERK or Shp2 did not suppress the transcriptional activity of SREBP-1c stimulated by OA or ethanol. The Shp2−ERK pathway regulates cell proliferation in several cell lines and mediates hepatocyte injury and growth (Bard-Chapeau et al., 2005, 2006). In addition, Gab2 was demonstrated to regulate the phosphorylation of IRS1 in HepG2 cells treated with OA. Therefore, Gab2 may recruit Shp2/ERK to induce IR or hepatocyte impairment without disturbing lipid deposition.
In addition, the extrahepatic effects of global knockout of Gab2 may also contribute to the improvement of liver impairments caused by disease factors. Deletion of Gab2 may also improve the lipid production in liver tissue by regulation of the whole body lipid metabolism (Malaguarnera et al., 2009). In our study, the obesity was improved in HFD-fed Gab2−/− mice, which could decrease lipid accumulation in the liver. Furthermore, deletion of Gab2 may regulate the lipid metabolism in other tissues through the same mechanisms as in the liver, such as regulating the phosphorylation of IRS1, IR, and the expression of SREBP-1 (Malaguarnera et al., 2009; Xiao and Song, 2013). The details of extrahepatic effects of global knockout of Gab2 on metabolism will be investigated in the future.
Our results support that Gab2 activates several cytoplasmic signaling pathways (PI3K−AKT, Socs3−Stat3, and Shp2−ERK) to mediate pathological progress in fatty liver, a complicated disease caused by multifaceted factors. Because each pathogenic factor employs a different signaling pathway, it is difficult to fully elucidate the mechanism of fatty liver and identify an effective therapeutic target through a single pathway (Birkenfeld and Shulman, 2014; Suk et al., 2014). Gab2 is a central platform to govern the integrated regulatory signals in disease development and may represent an efficient therapeutic target. In addition, Gab2 is weakly expressed in the liver of mice, and mice without Gab2 behave normally (Gu et al., 2001; Nakaoka and Komuro, 2013). Thus, treatment by reducing the Gab2 protein may have very minor side effects on normal physiological functions. Based on these advantages, Gab2, as a novel key modulator in fatty liver disease, may be a powerful target for the prevention and therapy of fatty liver disease.
Materials and methods
Animal models and treatments
C57BL/6 mice were obtained from the Laboratory Animal Center of Xiamen University, China. Gab2-transgene mice were the gift from the Burnham Institute for Medical Research, California, USA (Ke et al., 2007). Mice were housed under standard conditions with free access to food and water. All experimental procedures were approved by the Animal Welfare Committee of Research Organization (X200810), Xiamen University. The details of mouse models (NAFLD, NASH, and AFLD) and animal treatments are described in Supplementary Materials and methods.
Lipid droplet assays
Liver tissue from NAFLD mice was fixed in formalin for at least 24 h and then embedded in paraffin. Tissue sections were stained with hematoxylin-eosin (H&E). Fat vacuole areas were quantified using the color segmentation function of Scn Image software. Frozen sections of liver tissue in AFLD mice were stained with Oil Red O following the previous description. HepG2 cells were fixed in 10% formalin for 30 min and stained in Oil Red O. Lipid droplets in cells were eluted with isopropanol, and the absorbance of the solution was measured using an ELISA reader (Bio-Rad) at a wavelength of 450 nm.
Biochemical indicator detection assays
In brief, liver tissue (50 mg for TG, 30 mg for cholesterol) or HepG2 cells (1 × 107) were lysed in 1 ml of lysis buffer. The supernatants of the lysate were analyzed using an ELISA kit. After standing for 2 h at room temperature, fresh blood was centrifuged at 9000 rpm for 3 min. TG, cholesterol, ALT, and AST in the supernatant were detected using ELISA kits according to the manufacturers’ protocols. For TG and cholesterol, the kit was obtained from Nanjing Jiancheng Bioengineering Institute (China, E1003 and E1005). For ALT and AST, the kit was obtained from Applygen Technologies Inc. (China, C009-2 and C0010-2).
Extraction of liver primary cells, cell culture and treatment
Extraction of liver primary cells was performed using a two-step perfusion method (Kreamer et al., 1986). The human hepatoblastoma HepG2 cell line was purchased from American Type Culture Collection (ATCC). The detailed protocols are described in Supplementary Materials and methods.
Transfection and luciferase assays
The transfection of plasmids and siRNA into HepG2 cells and transcriptional activity reporter assays for SREBP-1c were performed as previously described (Lu et al., 2006). Briefly, HepG2 cells were seeded in 24-well plates and co-transfected with Gab2-siRNA (5′-GGAGTGCCAGCTTCTCTCA-3′) and pSyn SRE (a kind gift from Dr Timothy F. Osborne, Professor of University of California, Irvine, USA) into different cells using Lipofectamine 2000. After treatment with OA or ethanol, luciferase activity was measured in a Victor 3 V multilabel plate reader (Perkin Elmer).
Nuclear and cytoplasmic protein preparation, co-immunoprecipitation, and western blot analyses
Nuclear and cytoplasmic protein extracts of HepG2 cells were prepared using the NE-PER Nuclear and Cytoplasmic Extraction Kit (Cat. 78833, Thermo) as previously described (Li et al., 2014). Immunoblotting and co-immunoprecipitation analyses of whole-cell lysates or cytoplasmic/nuclear protein solutions were performed as previously described (Li et al., 2014). Grayscale western blotting bands were quantified using Quantity One software (Bio-Rad). The antibodies are described in Supplementary Materials and methods.
Statistical analyses
All statistics were analyzed using the single-tailed Student's t-test or two-way ANOVA using GraphPad Prism software.
Supplementary material
Supplementary material is available at Journal of Molecular Cell Biology online.
Acknowledgements
We thank Dr Timothy F. Osborne (University of California, Irvine, USA) for the gift of pSyn SRE plasmid.
Funding
This work was supported by the National Basic Research Program of China (grant no. 2010CB945004 and 2013CB945503) and the National Natural Science Foundation of China (grant no. 30772546).
Conflict of interest: none declared.
References
Author notes
These authors contributed equally to this work.
