Nonalcoholic steatohepatitis (NASH), characterized

Transcription

1 AMERICAN ASSOCIATION FOR THE STUDY OFLIVERD I S E ASES HEPATOLOGY, VOL. 65, NO. 5, 2017 STEATOHEPATITIS/METABOLIC LIVER DISEASE The E3 Ligase Tripartite Motif 8 Targets TAK1 to Promote Insulin Resistance and Steatohepatitis Feng-Juan Yan, 1,2* Xiao-Jing Zhang, 2-4* Wen-Xin Wang, 2-4* Yan-Xiao Ji, 2-4 Pi-Xiao Wang, 2-4 Yang Yang, 2 Jun Gong, 2-4 Li-Jun Shen, 2-4 Xue-Yong Zhu, 2-4 Zan Huang, 1,2 and Hongliang Li 2-4 Tripartite motif 8 (TRIM8), an E3 ligase ubiquitously expressed in various cells, is closely involved in innate immunity. However, its role in nonalcoholic steatohepatitis is largely unknown. Here, we report evidence that TRIM8 is a robust enhancer of steatohepatitis and its complications induced by a high-fat diet or a genetic deficiency (ob/ob). Using gain-offunction and loss-of-function approaches, we observed dramatic exacerbation of insulin resistance, hepatic steatosis, inflammation, and fibrosis by hepatocyte-specific TRIM8 overexpression, whereas deletion or down-regulation of TRIM8 in hepatocytes led to a completely opposite phenotype. Furthermore, investigations of the underlying mechanisms revealed that TRIM8 directly binds to and ubiquitinates transforming growth factor-beta activated kinase 1, thus promoting its phosphorylation and the activation of downstream c-jun N-terminal kinase/p38 and nuclear factor jb signaling. Importantly, the participation of TRIM8 in human nonalcoholic fatty liver disease and nonalcoholic steatohepatitis was verified on the basis of its dramatically increased expression in the livers of these patients, suggesting a promising development of TRIM8 disturbance for the treatment of nonalcoholic steatohepatitis related metabolic disorders. Conclusion: The E3 ligase TRIM8 is a potent regulator that exacerbates steatohepatitis and metabolic disorders dependent on its binding and ubiquitinating capacity on transforming growth factor-beta activated kinase 1. (HEPATOLOGY 2017;65: ) Nonalcoholic steatohepatitis (NASH), characterized by lipid accumulation and inflammation in the liver, contributes to a broad spectrum of severe pathologies, ranging from metabolic syndrome to hepatocellular carcinoma. (1,2) The everincreasing incidence of NASH has placed it as the predicted leading cause of liver transplantation in the next decade. However, no pharmacological approach is available for its clinical treatment due to our poor understanding of its complex pathogenesis. (3-5) Tripartite motif 8 (TRIM8) is a ubiquitously expressed factor in the TRIM family that orchestrates Abbreviations: Ad, adenoviral; Adsh, adenoviral short hairpin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; cdna, complementary DNA; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GFP, green fluorescent protein; G6Pase, glucose 6-phosphatase; GTT, glucose tolerance test; HA, hemagglutinin; HFD, high-fat diet; HOMA-IR, homeostatic model assessment of the insulin resistance; IKKb, inhibitor of NF-jB kinase subunit b; IL, interleukin; IP, immunoprecipitation; IRS, insulin receptor substrate; ITT, insulin tolerance test; JNK, c-jun N-terminal kinase; MAPK, mitogen-activated protein kinase; NASH, nonalcoholic steatohepatitis; NC, normal chow; NEFA, nonesterified fatty acid; NF-jB, nuclear factor jb; NTG, nontransgenic; 5Z-7-OX, 5Z-7-Oxozeaenol; PEPCK, phosphoenolpyruvate carboxykinase; TAK1, transforming growth factor-beta activated kinase 1; TC, total cholesterol; TG, triglyceride; TRIM8, tripartite motif 8; TRIM8-HKO, hepatocyte-specific TRIM8-knockout; TRIM8-HTG, hepatocyte-specific TRIM8 transgenic. Received August 30, 2016; accepted November 22, Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/ /hep.28971/suppinfo. *These authors contributed equally to this work. Supported by grants from the National Science Fund for Distinguished Young Scholars ( ), the National Natural Science Foundation of China ( , ), the National Science and Technology Support Project (2011BAI15B02, 2012BAI39B05, 2013YQ , 2014BAI02B01, 2015BAI08B01, and 2016YFF ), the Key Project of the National Natural Science Foundation ( , ), the National Basic Research Program China (2011CB503902), and the Natural Science Foundation of Hubei Province (2013CFB259). Copyright VC 2016 by the American Association for the Study of Liver Diseases. View this article online at wileyonlinelibrary.com. DOI /hep Potential conflict of interest: Nothing to report. 1492

2 HEPATOLOGY, Vol. 65, No. 5, 2017 YAN, ZHANG, ET AL. multiple biological processes, including cell survival, differentiation, apoptosis, and, in particular, the innate immune response. (6-8) Harboring a conserved RBCC motif composed of a RING domain followed by 2 B- box and a coiled-coil domain at the N terminus, TRIM8 possesses E3 ligase activity. (9) Previous studies of this factor and its family members have mainly focused on their participation in cancer and immune diseases deriving from their potent regulation of inflammation. (8,10-12) Given the close involvement of inflammation in NASH, (13) we hypothesized that TRIM8 might be critical for the pathogenesis of steatohepatitis and related metabolic disorders; however, this hypothesis and associated underlying mechanisms remain to be tested. In the present study, we performed in vivo and in vitro experiments to test our hypothesis regarding the function of TRIM8 in steatohepatitis and its complications. Employing gain-of-function and loss-offunction approaches, we found that TRIM8 exacerbated high-fat diet (HFD) induced or gene deficiency (ob/ob) induced insulin resistance, hepatic lipid accumulation, inflammation, and fibrosis through binding to and ubiquitinating transforming growth factorbeta activated kinase 1 (TAK1). Materials and Methods ANIMALS AND TREATMENTS (TRIM8-flox) mice were created using the CRISPR/Cas9 system, (14) which were then crossed with Albumin-Cre mice (Jackson Laboratory, Bar Harbor, ME) to obtain hepatocyte-specific TRIM8-knockout mice (TRIM8-HKO). Constructs containing full-length TRIM8 complementary DNA TRIM8 flox/flox (cdna) were inserted after CAG-loxp-CAT-loxp cassettes, which were microinjected into fertilized C57BL/6J embryos to generate conditional hepatocyte-specific TRIM8 overexpressing mice. The generated mice were crossed with Albumin-Cre mice to obtain hepatocyte-specific TRIM8 transgenic (TRIM8-HTG) mice. To knock down the expression of TRIM8 in ob/ob mice, adenoviral short hairpin (Adsh) TRIM8 ( pfu) was injected at the age of 8 weeks after normal chow (NC) diet administration. The ob/ob mice injected with AdshRNA were used as controls for the AdshTRIM8-treated mice. AdTRIM8, AdTRIM8-M (an adenovirus-encoding mutant TRIM8 lacking the TAK1-binding domain [59-182]), and AdTRIM8(C15A; C18A) were injected into C57BL/6J mice to overexpress TRIM8 or its indicated mutants, respectively. Mice injected with adenoviral green fluorescent protein (AdGFP) were used as controls. 5Z-7-Oxozeaenol (5 mg/kg; 5Z-7-OX, O MG; Sigma, St. Louis, MO) was intraperitoneally administered into nontransgenic (NTG) and TRIM8-overexpressing mice in parallel to specifically inhibit TAK1 activation. A mouse NASH model was established through feeding mice an HFD (60% kcal from fat; D-12492; Research Diets, NJ) for 12 weeks starting at the age of 8-10 weeks, whereas NC-fed mice were maintained on a rodent diet with 10% kcal from fat (D-12450B). Food and water were provided ad libitum. Mice were housed in a standard environment with a 12-hour light/12-hour dark cycle. Body weights, fasting blood glucose levels, and fasting insulin levels were recorded every 2 weeks. All of the animal experiments were approved by the Animal Care and Use Committee of Renmin Hospital at Wuhan University, and care was provided according to the criteria outlined in the Guide ARTICLE INFORMATION: From the 1 College of Life Sciences, 2 School of Basic Medical Sciences, 3 Institute of Model Animals, and 4 Medical Research Institute, School of Medicine, Wuhan University, Wuhan, China. ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO: Zan Huang, Ph.D. College of Life Sciences, Wuhan University 16 Luo-Jia-Shan Road, Wuhan , China z-huang@whu.edu.cn Tel: or Hongliang Li, M.D., Ph.D. Renmin Hospital and Institute of Model Animals Collaborative Innovation Center of Model Animal and Cardiovascular Research Institute Wuhan University Luojia Mount Wuchang Wuhan , China lihl@whu.edu.cn Tel:

3 YAN, ZHANG, ET AL. HEPATOLOGY, May 2017 for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences and published by the National Institutes of Health. HUMAN LIVER SAMPLES Human liver samples including nonsteatosis and steatohepatitis were collected from Renmin Hospital of Wuhan University. Written informed consent was signed by the subjects or by the families of the liver donors. Procedures involving human samples were approved by the Renmin Hospital of Wuhan University Review Board and were consistent with the principles outlined in the Declaration of Helsinki. METABOLIC PARAMETER ANALYSIS Fasting blood glucose levels, fasting serum insulin levels, homeostatic model assessment of insulin resistance (HOMA-IR) values, glucose tolerance tests (GTTs), and insulin tolerance tests (ITTs) were examined and performed as described. (15) HISTOLOGICAL ANALYSES AND IMMUNOHISTOCHEMISTRY Oil red O staining and hematoxylin and eosin staining were performed on frozen and paraffin-embedded liver sections, respectively. Periodic acid Schiff staining was performed to detect glycogen as reported. (15) Histological fibrosis was examined by sirius red and alpha-smooth muscle actin staining. The expression and localization of TRIM8 were investigated by immunohistochemistry. LIVER LIPID AND FUNCTION ASSAYS Triglyceride (TG), total cholesterol (TC), and nonesterified fatty acid (NEFA) levels were detected in liver samples using commercial kits (Wako, Osaka, Japan), while serum levels of alanine amino transferase (ALT), aspartate amino transferase (AST), and alkaline phosphatase (ALP) were measured using an ADVIA 2400 Chemistry System analyzer (Siemens, Tarrytown, NY) according to the manufacturer s instructions to evaluate the liver function of the mice. PRIMARY HEPATOCYTE ISOLATION, CULTURE, AND TREATMENT Primary hepatocytes were isolated from 6-week-old to 8-week-old male mice using a liver perfusion approach, as described. (16) Hepatocytes were cultured in Dulbecco s modified Eagle s medium supplemented with 10% fetal bovine serum and 1% penicillin streptomycin in a 5% CO 2 /water-saturated incubator at 378C. QUANTITATIVE REAL-TIME PCR After total mrna isolation and cdna synthesis, PCR amplification was performed with the SYBR Green PCR Master Mix ( ; Roche). The mrna expression levels of the target genes were normalized to the expression of glyceraldehyde 3- phosphate dehydrogenase (GAPDH). WESTERN BLOT ANALYSIS For western blotting, total protein samples were extracted from tissue or cell samples and 30 lg of protein samples were separated on a 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis gel (and transferred to a polyvinylidene fluoride membrane). Protein expression was visualized by incubating primary antibodies (Supporting Table S2) overnight at 48C, followed by the corresponding secondary antibodies. PLASMID CONSTRUCTS The entire mouse TRIM8 coding region was cloned into pcdna5-ha-c1 and psi-flag-c1 plasmids to generate the pcdna5-ha-trim8 and psi-flag- TRIM8 recombinant plasmids. TRIM8 truncations were obtained by PCR amplification of psi-flag- TRIM8 using corresponding primer pairs. The pcdna5-ha-tak1, pcdna5-flag-tak1, and Flag-labeled TAK1 truncations were constructed as described. (17) IMMUNOPRECIPITATION AND IN VIVO UBIQUITINATION ASSAYS Immunoprecipitation (IP) assays were performed as described in previous studies to identify the binding capacity and domain of TRIM8 that binds to TAK1. (18) Ubiquitination assays were performed as described. (19) 1494

4 HEPATOLOGY, Vol. 65, No. 5, 2017 YAN, ZHANG, ET AL. STATISTICAL ANALYSIS Statistical analyses were performed using SPSS software (version 13.0), and all data are expressed as means 6 standard deviations. Statistical differences among more than 2 groups were compared using a one-way analysis of variance, followed by Bonferroni s post hoc tests (assuming equal variances) or Tamhane s T2 post hoc tests. Student t tests were performed to compare the differences between two groups. P < 0.05 was considered significant. ADDITIONAL METHODS Detailed methods are provided in the Supporting Information. Results HEPATIC TRIM8 EXPRESSION IS INCREASED IN STEATOTIC LIVERS Our hypothesis on the potential regulation of TRIM8 in hepatic steatosis was first tested by examining variations in the expression of the TRIM8 protein in fatty livers. The TRIM8 protein level was dramatically increased in the livers from patients with steatosis and steatohepatitis compared with nonsteatotic controls, with the highest expression level in the steatohepatitis group (steatohepatitis > steatosis > nonsteatosis) (Fig. 1A). A mouse model of hepatic steatosis was established through feeding an HFD for 12 weeks to investigate the relationship between TRIM8 and steatohepatitis, and TRIM8 expression was measured in liver samples from these mice. Consistent with our observations in human samples, the expression of TRIM8 in the livers of HFD-fed mice was significantly increased compared with its expression in NC-fed controls (Fig. 1B). Increased TRIM8 expression was also observed in the livers of ob/ob mice (Fig. 1C). In primary hepatocytes, expression of TRIM8 protein was significantly increased after palmitate challenge compared with the control group (Fig. 1D). Immunohistochemistry on liver sections further confirmed the significant elevation of TRIM8 in the livers of mice upon HFD treatment (Fig. 1E). However, the mrna level of TRIM8 was not significantly changed in fatty livers or in palmitate-stimulated primary hepatocytes, suggesting that the up-regulated TRIM8 protein level in fatty liver was not due to the changed transcriptional level (Supporting Fig. S1). The consistent increase in TRIM8 expression in livers of nonalcoholic fatty liver disease/nash patients and in the mouse NASH model induced by different stimuli suggests a potential regulatory effect of TRIM8 on steatohepatitis. HEPATIC TRIM8 OVEREXPRESSION EXACERBATES HFD-INDUCED INSULIN RESISTANCE Considering the dramatic increase in TRIM8 expression in fatty livers, we generated hepatocyte-specific TRIM8-overexpressing mice (Fig. 2A) to evaluate the role of increased TRIM8 expression in the pathogenesis of NASH and related metabolic disorders. With HFD consumption, the body weights, fasting blood glucose levels, and insulin levels of NTG and TRIM8-HTG mice were rapidly increased compared to the NCtreated controls, and the fasting blood glucose and insulin levels in TRIM8-HTG mice were much higher than those in the NTG littermates from 2 to 12 weeks of HFD administration (Fig. 2B-D). Correspondingly, HOMA-IR values were markedly increased by TRIM8 overexpression (Fig. 2E). TRIM8-HTG mice exhibited significantly higher glucose concentrations and areas under the curve in GTT and ITT assays, indicating that TRIM8 compromised insulin sensitivity (Fig. 2F,G). Consistently, insulin signaling was impaired by TRIM8 as evidenced by the significantly reduced levels of phosphorylated insulin receptor substrate on Tyr 608 (IRS Tyr608 ), AKT Ser473, and glycogen synthase kinase 3b in the TRIM8-HTG group (Fig. 2H). Additionally, the glycogen content in the TRIM8-HTG group was much lower than in NTG control mice after HFD treatment (Fig. 2I), accompanied by marked increases in phosphoenolpyruvate carboxykinase (PEPCK) and the catalytic subunit of glucose 6-phosphatase (G6Pase) mrna and protein levels (Fig. 2J,K). Together, these results reveal that overexpression of TRIM8 in the liver exacerbates insulin resistance and glucose metabolism dysregulation induced by an HFD. TRIM8 DEFICIENCY IN THE LIVER INHIBITS INSULIN RESISTANCE INDUCED BY AN HFD To further validate the functional role of TRIM8 in HFD-induced glucose metabolism and insulin 1495

5 YAN, ZHANG, ET AL. HEPATOLOGY, May 2017 FIG. 1. TRIM8 expression is up-regulated in fatty livers. (A) TRIM8 protein expression in livers of nonsteatotic donors (n 5 6), patients with simple steatosis (n 5 8), and steatohepatitis subjects (n 5 12). *P < 0.05 versus nonsteatotic controls. (B) Expression level of TRIM8 protein in mouse livers after HFD or NC feeding for 12 weeks (n 5 4). *P < 0.05 versus the NC group. (C) TRIM8 expression level measured in livers of ob/ob mice at the indicated time points (n 5 4). *P < 0.05 versus the expression of TRIM8 at 2 weeks. (D) Expression of TRIM8 in primary hepatocytes that were treated with palmitate or vehicle controls. Three independent experiments were performed. *P < 0.05 versus control group. For (A-D), results are presented as means 6 standard deviation; b-actin level was used for normalization. (E) Representative immunohistochemical images of the TRIM8 expression profile in the mouse liver sections (n 5 4/group). 1496

6 HEPATOLOGY, Vol. 65, No. 5, 2017 YAN, ZHANG, ET AL. FIG. 2

7 YAN, ZHANG, ET AL. HEPATOLOGY, May 2017 resistance, TRIM8-HKO mice were generated (Fig. 3A). As anticipated, mice with a TRIM8 deficiency showed reduced fasting blood glucose levels, fasting serum insulin levels, and HOMA-IR values compared with TRIM8-flox littermate controls after a 12-week HFD challenge (Fig. 3B-E). Furthermore, the insulin sensitivity of HFD-fed mice was clearly improved by TRIM8 deletion, as indicated by the GTT and ITT results (Fig. 3F,G). Examinations of IRS AKT signaling suggested that insulin signaling activity was enhanced by TRIM8 ablation (Fig. 3H). Moreover, glycogen content and gluconeogenesis were increased and suppressed, respectively, in the livers of TRIM8- HKO mice compared with control mice (Fig. 3I-K). Interestingly, the influence of TRIM8 on the physiological phenotype of mice treated with NC was negligible (Figs. 2 and 3). The combined observations of gain-of-function and loss-of-function approaches clearly demonstrate that TRIM8 promotes HFDinduced insulin resistance and glucose metabolic impairment. HEPATIC TRIM8 EXACERBATES HEPATIC STEATOSIS, INFLAMMATION, AND FIBROSIS Lipid accumulation in the liver is the most prominent characteristic of NASH and can be promoted by insulin resistance and hepatic inflammation. (20) Sustained HFD elicited dramatic increases in liver weights and the ratio of liver weight to body weight, which were further augmented by TRIM8 up-regulation but blunted in TRIM8-HKO mice compared with controls (Fig. 4A,B). Hematoxylin and eosin and oil red O staining showed increased lipid accumulation in TRIM8-HTG mice, whereas artificial TRIM8 deficiency led to a decrease in hepatic steatosis after HFD administration (Fig. 4C). Consistent with these findings, TG, TC, and NEFA contents were significantly increased by TRIM8 (Fig. 4D,E). Furthermore, the liver dysfunction evaluated by serum ALT, AST, and ALP levels was effectively ameliorated by hepatic TRIM8 deficiency but aggravated by TRIM8 overexpression (Supporting Fig. S2A,B). Real-time PCR analyses indicated that TRIM8 significantly increased cholesterol synthesis, fatty acid uptake, and fatty acid synthesis but reduced mrna levels of genes related to cholesterol efflux and fatty acid b-oxidation (Supporting Fig. S2C-J). The aggravating effect of TRIM8 on lipid accumulation was further confirmed in L02 hepatocytes challenged with palmitate using Bodipy-C16 staining (Supporting Fig. S2K). TRIM8-mediated inflammation during hepatic steatosis was demonstrated by the production of cytokines and chemokines in the livers of mice in the indicated groups. The mrna expression of interleukin-6 (IL-6), tumor necrosis factor-a, and monocyte chemoattractant protein 1 was dramatically elevated; and IL- 10 mrna levels were significantly reduced in TRIM8-HTG mice compared with the levels in NTG controls (Fig. 4F). In contrast, TRIM8 deletion in the liver alleviated the production of proinflammatory cytokines and chemokines (Fig. 4G). Consistently, the infiltration of F4/80-positive cells in the liver was remarkably increased by TRIM8 overexpression but decreased by TRIM8 deficiency (Fig. 4H). Western blotting showed that the classic nuclear factor jb (NF-jB) proinflammatory signaling pathway was dramatically activated by TRIM8 up-regulation, which was demonstrated by the significant increase in the expression of phosphorylated inhibitor of NF-jB kinase subunit b (IKKb) and p-ijba proteins, accompanied by a decreased IjBa expression (Fig. 4I). Conversely, the activation of NF-jB signaling was blocked in TRIM8-HKO mice compared with that in the TRIM8-flox controls (Fig. 4I). In terms of fibrosis, real-time PCR analysis showed that a 12-week HFD administration triggered a much higher mrna expression of profibrotic genes, FIG. 2. TRIM8 overexpression exacerbates HFD-induced insulin resistance and gluconeogenesis. (A) Expression of TRIM8 in the livers of NTG and TRIM8-HTG mice. b-actin served as a loading control. *P < 0.05 versus the NTG group. (B) Body weights of TRIM8-HTG mice or their NTG controls after 0, 2, 4, 6, 8, 10 and 12 weeks of NC or HFD feeding (n mice/group). (C- E) Fasting blood glucose levels (C), fasting serum insulin levels (D), and HOMA-IR values (E) in NC-fed or HFD-fed NTG or TRIM8-HTG mice at the indicated time points (n mice/group). (F,G) Blood glucose levels after NC or HFD feeding for 12 continuous weeks in NTG or TRIM8-HTG mice during IP GTTs (F) and IP ITTs (G). The corresponding areas under the curve are indicated on the right (n mice/group). (H) Expression levels of phosphorylated and total IRS1, AKT, and glycogen synthase kinase 3b in the livers of NTG and TRIM8-HTG mice fed HFD. Protein expression was normalized to b-actin levels. (I) Representative images of periodic acid Schiff staining in liver sections from each indicated group (n 5 4/group). (J,K) Expression of PEPCK and G6Pase mrnas (J) and proteins (K) in liver samples from TRIM8-HTG mice and their littermate controls after HFD treatment for 12 weeks (n 5 4/group). GAPDH and b-actin served as controls for the PCR and western blot analyses, respectively. (B-K) *P < 0.05 versus the NTG/NC group; #P<0.05 versus NTG/HFD group. Abbreviations: GSK, glycogen synthase kinase; PAS, periodic acid Schiff. 1498

8 HEPATOLOGY, Vol. 65, No. 5, 2017 YAN, ZHANG, ET AL. FIG

9 YAN, ZHANG, ET AL. HEPATOLOGY, May 2017 including Col1a1, Tgfb1, Ctgf, and a-sma, than NC treatment (Fig. 4J). Histological fibrosis was also developed as evidenced by sirius red and a-smooth muscle actin staining (Fig. 4K). The increased profibrotic gene expression and histological fibrosis were further enhanced by TRIM8 overexpression but greatly reversed by TRIM8 deficiency (Fig. 4J,K). Interestingly, TRIM8 showed no significant influence on fatty acid and cholesterol metabolism, inflammation, and fibrosis in NC groups (Fig. 4; Supporting Fig. S2). Together, these results suggest that TRIM8 functions as a potent and positive regulator of hepatic steatosis, inflammation, and fibrosis in response to HFD administration. TRIM8 DOWN-REGULATION ALLEVIATES INSULIN RESISTANCE AND HEPATIC STEATOSIS IN ob/ob MICE Given that gene mutation represents another important NASH etiological factor, beyond HFD challenge, (21) we examined the possible roles of TRIM8 in steatohepatitis and metabolic disorders in ob/ob mice by down-regulating TRIM8 expression by infecting mice with AdshTRIM8 (Fig. 5A). AdshRNA-infected mice served as controls. Similar to the observations in TRIM8-HKO mice, no significant regulation on body weight was found by TRIM8 down-regulation, while TRIM8 knockdown contributed to the significant reductions of fasting blood glucose levels, fasting serum insulin levels, and HOMA-IR index values (Fig. 5B- E). Additionally, the GTT and ITT results suggested an improved insulin resistance, accompanied by higher glycogen levels in the liver sections from the AdshTRIM8-infected ob/ob mice compared with that in the control group (Fig. 5F-I). In terms of hepatic steatosis, AdshTRIM8 administration attenuated the increase in liver weight and hepatic lipid accumulation (Fig. 5J-L). The data obtained from ob/ob mice indicated that the TRIM8 down-regulation may have a potential therapeutic capacity for treating gene deficiency-induced insulin resistance and hepatic steatosis. TAK1 C-JUN N-TERMINAL KINASE/p38 SIGNALING IS INVOLVED IN TRIM8- REGULATED STEATOHEPATITIS The potent function of TRIM8 in steatohepatitis and related metabolic disorders prompted us to explore the mechanisms underlying its effects. Accumulating evidence suggests the close participation of mitogenactivated protein kinase (MAPK) signaling in the initiation and progression of NASH and its complications, (22) and thus, our efforts to investigate the mechanisms underlying TRIM8-regulated hepatic steatosis were first focused on MAPK signaling. Figure 6A indicates that all of these 3 subunits of MAPK, i.e., extracellular signal regulated kinase, c-jun N-terminal kinase (JNK) and p38, were activated after HFD feeding; however, the expression of phosphorylated JNK and p38, but not phosphorylated extracellular signal regulated kinase, was significantly influenced by the alterations in TRIM8 expression. Among the classic JNK and p38 members, including apoptosis signal regulating kinase 1, TAK1, and MAPK/extracellular signal regulated kinase kinase 1, (23) we found that only the activation of TAK1 responsive to TRIM8 expression changes (Fig. 6B). Effects of TRIM8 on TAK1 JNK/p38 were substantiated by an in vitro experiment using palmitate-challenged primary hepatocytes transfected with AdTRIM8 or Adsh- TRIM8 (Supporting Fig. S3A). We then investigated whether the activation of TAK1 is required for TRIM8-regulated steatohepatitis and metabolic disorders, and thus, a specific inhibitor of TAK1, 5Z-7-OX, was introduced and intraperitoneally injected into TRIM8-overexpressing FIG. 3. TRIM8 deficiency significantly ameliorates HFD-induced insulin resistance. (A) Expression of TRIM8 in liver, fat, muscle, and heart samples of TRIM8-flox and TRIM8-HKO mice (n 5 4/group). Expression level of TRIM8 was normalized to b-actin levels. *P < 0.05 versus TRIM8-flox mice. (B-E) Body weights (B), fasting blood glucose levels (C), fasting insulin levels (D), and HOMA-IR values (E) in TRIM8-HKO mice and their littermate controls at the indicated time points in response to HFD feeding (n /group). (F,G) IP GTTs (F) and IP ITTs (G) were performed in TRIM8-flox and TRIM8-HKO mice after HFD feeding for 12 weeks, and the corresponding areas under the curves were calculated (n mice/group). (H) Expression levels of phosphorylated and total IRS1, AKT, and glycogen synthase kinase 3b in the livers of TRIM8-HKO mice and their littermate controls after 12 weeks of HFD feeding. Expression levels were normalized to b-actin levels. (I) Representative images of periodic acid Schiff stained liver sections of TRIM8-flox and TRIM8-HKO mice after 12 weeks of HFD feeding (n 5 4/group). (J,K) PEPCK and G6Pase mrna (J) and protein (K) expression levels in livers of TRIM8-HKO and control mice (n 5 4/group). GAPDH and b-actin served as controls for the PCR and western blot analyses, respectively. (B-L) *P < 0.05 versus the TRIM8-flox/NC group; #P<0.05 versus the TRIM8-flox/HFD group. Abbreviations: GSK, glycogen synthase kinase; PAS, periodic acid Schiff. 1500

10 FIG. 4. TRIM8 aggravates hepatic steatosis, inflammation, and fibrosis in response to HFD. (A,B) Liver weights and ratios of liver weight to body weight of TRIM8-HTG (A) and TRIM8-HKO (B) mice and their corresponding littermate controls after NC or HFD feeding for 12 weeks (n mice/group). *P < 0.05 versus the NTG/NC or TRIM8-flox/NC group; #P<0.05 versus the NTG/ HFD or TRIM8-flox/HFD group. (C) Representative images of hematoxylin and eosin (upper panel) and oil red O (bottom panel) staining in liver sections of HFD-fed TRIM8-HTG, TRIM8-HKO, and littermate control mice (n 5 4/group). (D,E) TG, TC, and NEFA contents in livers of mice in the indicated groups after NC or HFD feeding for 12 weeks (n /group). (F,G) Relative mrna expression levels of IL-6, tumor necrosis factor-a, monocyte chemoattractant protein 1, and IL-10 in liver samples from the indicated groups after NC or HFD feeding for 12 weeks. Expression levels of the genes of interest were normalized to GAPDH levels. (H) Representative images of F4/80-positive inflammatory cell infiltration in liver sections of indicated mice (n 5 4/group). (I) Expression levels of total and phosphorylated IKKb, IjBa livers of NTG, TRIM8-HTG, TRIM8-Flox, and TRIM8-HKO mice after NC or HFD feeding (n 5 4). b-actin served as a loading control. (J) Relative mrna expression levels of Col1a1, Tgfb1, Ctgf, and a-sma in liver samples from the indicated groups. Expression levels of genes were normalized to GAPDH levels. (K) Representative images of sirius red (upper panel) and a-smooth muscle actin (bottom panel) staining in liver sections of TRIM8-HTG, TRIM8-HKO, and littermate control mice after HFD administration (n 5 4/group). (D-J) *P < 0.05 versus the NTG/NC or TRIM8-flox/NC group; #P<0.05 versus the NTG/HFD or TRIM8-flox/HFD group. Abbreviations: Col1a1, collagen type 1a1; Ctgf, connective tissue growth factor; H&E, hematoxylin and eosin; LW/BW, liver weight to body weight ratio; MCP1, monocyte chemoattractant protein 1; a-sma, a- smooth muscle actin; Tgfb1, transforming growth factor b1; TNF-a, tumor necrosis factor-a.

11 YAN, ZHANG, ET AL. HEPATOLOGY, May 2017 FIG. 4. Continued

12 HEPATOLOGY, Vol. 65, No. 5, 2017 YAN, ZHANG, ET AL. FIG

13 YAN, ZHANG, ET AL. HEPATOLOGY, May 2017 mice. The administration of 5Z-7-OX effectively reduced the level of phosphorylated TAK1 in mouse livers (Supporting Fig. S3B). Phenotypic studies indicated that the specific reduction of TAK1 activation significantly blocked HFD-induced insulin resistance and gluconeogenesis (Fig. 6C-G). More importantly, the enhanced increases in fasting blood glucose levels, fasting serum insulin levels, and HOMA-IR values (Fig. 6C-E), as well as the decreases in insulin sensitivity and glycogen content (Supporting Fig. S3C-G), by TRIM8 overexpression were largely reversed by TAK1 inhibition. In accord with its function in insulin resistance, 5Z-7-OX almost completely abolished the positive regulation of TRIM8 on liver weight gain, hepatic lipid accumulation, and liver dysfunction (Fig. 6F-I). A significant suppression of the activation of JNK and IKK as well as increased expression of total IjBa and phosphorylated AKT might account for the neutralizing effect of TAK1 inhibition on TRIM8 s role in these pathologies (Fig. 6J). As it stands, the activation of TAK1 is essential for the function of TRIM8 in regulating insulin resistance and hepatic steatosis. TRIM8 INTERACTS WITH AND UBIQUITINATES TAK1 Next, we examined how TRIM8 and TAK1 communicate during the development of steatohepatitis, and thus, the binding of TRIM8 to TAK1 was measured. IP assays indicated that TRIM8 physically interacted with TAK1 (Fig. 7A). Interestingly, TRIM8 overexpression induced enhancement of its binding to TAK1 only occurred in cytoplasm and requires palmitate stimulation. We found that TRIM8 mainly located at the nucleus at baseline but was translocated into the cytoplasm of hepatocytes where it colocalizes with and binds to TAK1 after palmitate administration (Supporting Fig. S4). This provided an explanation for the stress-dependent manner of TRIM8-regulated nonalcoholic fatty liver disease phenotypes. A mapping experiment was performed to further identify domains that were responsible for TRIM8 TAK1 interaction. TRIM8 and TAK1 were both truncated into small fragments according to their respective functional domain distributions (Fig. 7B,C). Plasmids encoding TRIM8 and TAK1 fragments were transfected into HEK293T cells with hemagglutinin (HA)-TAK1 and HA-TRIM8, respectively. After IP analysis, we observed that the amino acid domain had an indispensable role in the binding of TRIM8 to TAK1, while the amino acid domain of TAK1 was sufficient for its interaction with TRIM8 (Fig. 7D,E). Because TRIM8 has an E3 ligase activity and TAK1 phosphorylation and activation are largely dependent on its ubiquitination, (24,25) we wondered whether TRIM8 can ubiquitinate TAK1 and facilitate its activation. Therefore, the ubiquitination level of TAK1 in livers of TRIM8-flox, TRIM8-HKO, NTG, and TRIM8-HTG mice were examined after HFD treatment. We found that TRIM8 deficiency dramatically decreased the TAK1 ubiquitination level, which, however, was largely increased by TRIM8 overexpression compared with their corresponding controls (Fig. 7F). In HEK293T cells cotransfected with TRIM8 and TAK1, the ubiquitination of TAK1 was also dramatically increased compared with cells that were transfected with TAK1 alone (Fig. 7G). It has been established that the RING domain of TRIM8 is responsible for its E3 ligase activity. (24) After cys15 and cys18 in the RING domain were mutated to alanine (C15A; C18A), TRIM8 lost its ubiquitination activity on TAK1 (Fig. 7H). The failure of TRIM8 on TAK1 ubiquitination also occurred when the amino acid domain of TRIM8 was mutated (Fig. 7I). In light of these findings, we deduce that the binding capacity and E3 ligase activity of TRIM8 to TAK1 are required for the activation of TAK1 and its downstream signaling. FIG. 5. TRIM8 down-regulation reduces spontaneous insulin resistance and hepatic steatosis in ob/ob mice. (A) Expression of TRIM8 in livers of ob/ob mice after AdshRNA or AdshTRIM8 infection (n 5 4). AdshRNA-infected mice served as controls. (B-E) Body weights (B), fasting blood glucose levels (C), fasting insulin levels (D), and HOMA-IR values (E) in AdshRNA-infected or AdshTRIM8-infected ob/ob mice (n mice/group). (F,G) IP GTT (F) and IP ITT (G) results and corresponding areas under the curve calculated for AdshRNA-infected or Adsh-TRIM8-infected ob/ob mice (n /group). (A-G) *P < 0.05 versus AdshRNA-infected ob/ob mice. (H) Representative periodic acid Schiff stained images of liver sections from ob/ob mice after AdshRNA or AdshTRIM8 infection (n 5 4). (I) mrna expression of PEPCK and G6Pase in liver samples from AdshRNAinfected or AdshTRIM8-infected ob/ob mice (n 5 4 mice/group). (J) Liver weights and ratios of liver weight to body weight in AdshRNA-infected or Adsh-TRIM8-infected ob/ob mice (n /group). *P < 0.05 versus AdshRNA group. (K) Representative images of hematoxylin and eosin and oil red O staining in liver sections of ob/ob mice after AdshRNA or AdTRIM8 infection (n 5 4). (L) TG, TC, and NEFA contents in livers of AdshRNA ob/ob and AdTRIM8 ob/ob mice. *P < 0.05 versus AdshRNA ob/ob group. Abbreviations: AUC, area under the curve; H&E, hematoxylin and eosin; LW/BW, liver weight to body weight ratio; PAS, periodic acid Schiff. 1504

14 FIG. 6. TRIM8 regulates TAK1 JNK signaling during hepatic steatosis. (A) Expression levels of phosphorylated and total extracellular signal regulated kinase, JNK, and p38 proteins in the livers of the NC-fed or HFD-fed mice in the NTG, TRIM8-HTG, TRIM8-flox, and TRIM8-HKO groups (n 5 4/group). *P < 0.05 versus the NTG/NC or TRIM8-flox/NC group; # P < 0.05 versus the NTG/HFD or TRIM8-flox/HFD group. (B) Expression levels of total and phosphorylated apoptosis signal regulating kinase 1, TAK1, and MAPK/extracellular signal regulated kinase kinase 1 in livers of mice in the indicated groups (n 5 4/group). *P < 0.05 versus the NTG/NC or TRIM8- flox/nc group; # P < 0.05 versus the NTG/HFD or TRIM8-flox/HFD group. (C-E) Fasting blood glucose levels (C), fasting insulin levels (D), and HOMA-IR values (E) in mice from the indicated groups after HFD feeding for 12 continuous weeks (n /group). (F) Liver weights and ratios of liver weight to body weight of mice in the indicated groups (n /group). *P < 0.05 versus NTG/DMSO mice; # P < 0.05 versus TRIM8-HTG/dimethyl sulfoxide mice. (G) Representative hematoxylin and eosin stained and oil red O stained images of liver sections from NTG or TRIM8-HTG mice treated with 5Z-7-OX or the dimethyl sulfoxide vehicle control. (H) TG, TC, and NEFA contents in the livers of mice in the indicated groups. (I) Serum levels of ALT, AST, and ALP in mice with or without TRIM8 overexpression that were treated with 5Z-7-OX or the dimethyl sulfoxide vehicle control. (J) Expression levels of phosphorylated JNK, IKKb, IjBa, and AKT proteins in the livers of the indicated mice. Expression level of b-actin served as a loading control. *P < 0.05 versus the NTG/DMSO group; # P < 0.05 versus the TRIM8-HTG/DMSO group. Abbreviations: ASK1, apoptosis signal regulating kinase 1; DMSO, dimethyl sulfoxide; ERK, extracellular signal regulated kinase; H&E, hematoxylin and eosin; LW/BW, liver weight to body weight ratio; MEKK1, MAPK/extracellular signal regulated kinase kinase 1; NS, nonsignificant.

15 YAN, ZHANG, ET AL. HEPATOLOGY, May 2017 FIG. 6. Continued TRIM8-REGULATED TAK1 UBIQUITINATION IS ESSENTIAL FOR ITS ROLE IN STEATOHEPATITIS We next attempted to gain insight into whether the TRIM8 TAK1 interaction and the E3 ligase activity of TRIM8 are required for the regulatory effect of TRIM8 on metabolic disorders. To evaluate the role of the binding capacity of TRIM8 to TAK1 in its functional role, we introduced AdTRIM8-M into mice in parallel with AdTRIM8. AdGFP-infected mice served as controls. Administration of AdTRIM8-M failed to significantly enhance TAK1 s phosphorylation compared with that observed in AdTRIM8-injected group (Fig. 8A). Furthermore, the exacerbated insulin resistance and steatohepatitis by TRIM8 overexpression all disappeared in AdTRIM8-M treated mice (Fig. 8B-M). 1506

16 FIG. 7. TRIM8 directly interacts with and ubiquitinates TAK1. (A) A coimmunoprecipitation assay was performed in HEK293T cells cotransfected with Flag-TAK1 and HA-TRIM8 to examine the interaction between TRIM8 and TAK1. (B,C) Schematics of the TRIM8 (B) and TAK1 (C) constructs. (D,E) The interaction domains of TRIM8 and TAK1 were explored using full-length and truncated TRIM8 and TAK1 expression constructs and coimmunoprecipitation assays followed by western blotting. (F) An immunoprecipitation assay was used to examine TAK1 ubiquitination in livers from the indicated groups after HFD feeding for 12 weeks. Ubiquitination of TAK1 was measured using an anti-ub antibody. (G) HEK293T cells were transfected with HA-TAK1 with or without Myc-Ub and Flag-TRIM8. TAK1 ubiquitination was examined by immunoprecipitation and western blot analyses. (H) An immunoprecipitation assay was used to examine TAK1 ubiquitination in HEK293T cells that were transfected with HA-TAK1, Myc-Ub, Flag-TRIM8, and/or a Flag-TRIM8 mutant (C15A, C18A). (I) HEK293T cells were transfected with the HA-TAK1, Myc-Ub and Flag-TRIM8 constructs and subjected to immunoprecipitation and western blotting assays. Abbreviations: aa, amino acid; IB, immunoblot; IgG, immunoglobulin G.

17 FIG. 8. The TRIM8 TAK1 interaction is required for TRIM8-regulated insulin resistance and hepatic steatosis. (A) Expression of total and phosphorylated TAK1 in livers of mice infected with AdGFP, AdTRIM8, or AdTRIM8-M (plasmid encoding TRIM8 that lacks the TAK1 binding domain). (B-E) Body weights (B), fasting blood glucose levels (C), fasting insulin levels (D), and HOMA-IR values (E) of mice treated with AdGFP, AdTRIM8, or AdTRIM8-M. (F,G) Blood glucose levels in IP GTTs (F) and IP ITTs (G) and the corresponding calculated areas under the curve. (A-G) n /group; *P < 0.05 versus the AdGFP group; # P < 0.05 versus the AdTRIM8 group. (H) Representative images of periodic acid Schiff stained liver sections from AdGFPinfected, AdTRIM8-infected, or AdTRIM8-M-infected mice after HFD feeding for 12 weeks (n 5 4/group). (I) Expression of PEPCK and G6Pase mrnas in liver samples from indicated groups (n 5 4/group). (J) Liver weights and ratios of liver weight to body weight of mice in indicated groups (n /group). *P < 0.05 versus the AdGFP group; # P < 0.05 versus the AdTRIM8 group. (K) Representative images of hematoxylin and eosin and oil red O staining in liver sections from mice treated with AdGFP, AdTRIM8, or AdTRIM8-M after HFD feeding. (L) TG, TC, and NEFA contents in liver samples from indicated mice after 12- week HFD challenge (n mice/group). (M) Serum ALT, AST, and ALP levels in mice of indicated groups after HFD feeding for 12 continuous weeks (n /group). (K,L) *P < 0.05 versus the AdGFP group; # P < 0.05 versus the AdTRIM8 group. Abbreviations: AUC, area under the curve; H&E, hematoxylin and eosin; LW/BW, liver weight to body weight ratio.

18 HEPATOLOGY, Vol. 65, No. 5, 2017 YAN, ZHANG, ET AL. To further investigate the requirement of TRIM8 E3 ligase activity in its regulation in steatohepatitis, we injected AdTRIM8(C15A; C18A) mutant to mice that had been treated with HFD for 8 weeks. AdGFPinjected and AdTRIM8-injected mice served as negative and positive controls, respectively. Unlike the AdTRIM8 group, AdTRIM8(C15A; C18A) failed to activate TAK1 and did not show aggravated effect on glucose metabolic disorder, hepatic steatosis, and inflammation compared with the AdGFP group (Supporting Fig. S5). Interestingly, AdTRIM8(C15A; C18A) treatment led to a mild attenuation on NASH pathologies (Supporting Fig. S5). The improved steatohepatitis by TRIM8 E3 inactivated mutants might be explained by its properties of lacking ubiquitinating capacity but holding the ability to bind to TAK1, which might comprise the binding, ubiquitination, and activation of TAK1 by wild-type TRIM8. Collectively, the interaction with TAK1 and the E3 ligase activity of TRIM8 are indispensable for its exacerbated effects on steatohepatitis and its complications. Discussion TRIM8, a ubiquitously expressed E3 ligase, has been shown to be tightly involved in the innate immune response and cell cycle. (11,12,26) However, the regulatory effect of TRIM8 on steatohepatitis remains largely unknown. Here, utilizing in vivo and in vitro gain-offunction and loss-of-function approaches, we characterized TRIM8 as a pivotal positive regulator of HFDinduced and gene deficiency-induced steatohepatitis and its related metabolic disorders. Further investigations into the underlying molecular events revealed that TRIM8 directly interacts with and ubiquitinates TAK1, thereby eliciting its autophosphorylation and the hyperactivation of downstream IRS1-AKT, IKKb- NF-jB, and JNK/p38 signaling, which promotes the progression of NASH and its complications. Hepatic steatosis, insulin resistance, and inflammation are characteristic features/complications of NASH and can mutually enhance each other to form a vicious circle. (27,28) At the cellular level, insulin signaling, MAPK cascades, and IKK NF-jB pathway are major links among these pathologies. (29) After long-term HFD feeding, compromised activation of IRS1 AKT induces a sustained activation of FOXO1-PEPCK/the catalytic subunit of G6Pase to promote gluconeogenesis that in turn facilitates hepatic lipogenesis. (30-32) JNK increases IRS1 phosphorylation at ser307 and inhibits its tyrosine phosphorylation, (33) while p38 targets phosphatase and tensin homolog to inhibit phosphoinositide 3-kinase AKT cascades. (34) JNK also directly mediates lipid metabolism by targeting peroxisome proliferator activated receptor target genes and exacerbates inflammation through downstream c-jun and c-fos that bind to promoters of proinflammatory mediators. (35) During steatohepatitis, the inflammatory response was mainly driven by NF-jB signaling, which also activates RHOC to inhibit insulin receptor signaling and suppresses cytochrome P450 7A1 to reduce bile acid synthesis. (36) In our study, the capacity of TRIM8 to simultaneously activate all of these molecular events makes targeting TRIM8 a promising therapeutic strategy for NASH treatment. TAK1, a MAPK kinase kinase protein, has been recognized as a common upstream molecule of MAPK kinase JNK/p38 and IKK NF-jB cascades. (25) Using a specific TAK1 inhibitor (5Z-7-OX) that can block both the kinase activity and adenosine triphosphatase activity of TAK1, (37) we demonstrated a beneficial effect of TAK1 inhibition on steatohepatitis improvement and an indispensable role of its activation in TRIM8-exacerbated NASH pathologies. Because TAK1 kinase activity directly contributes to the phosphorylation of downstream factors (25,38) and adenosine triphosphatase activity is intrinsic for a kinase, (39,40) both kinase activity and adenosine triphosphatase activity of TAK1 might be involved in its role in steatohepatitis. Interestingly, Inokuchi-Shimizu et al. (41) and Morioka et al. (42) recently reported that the liver-specific knockout of TAK1 exacerbated hepatic steatosis. This appears to present a paradox concerning the role of TAK1 in steatohepatitis. However, it is important to contrast the TAK1 inhibition in our study with the deletion in these two previous studies. Remarkably, TAK1 maintains critically important functions in various vital pathophysiological processes, including cell survival, apoptosis, oxidative processes, and immune response. (43,44) Therefore, total deletion of TAK1 would be expected to impair its fundamental role in body homeostasis and lead to detrimental outcomes. However, it does not mean that TAK1 is a good guy with respect to steatohepatitis. Notably, the TAK1 constitutive activation also resulted in spontaneous inflammation and liver cancer. (45) Further, TAK1 can induce hyperactivation of downstream JNK and NFjB signaling, (15,17,45,46) which have been firmly confirmed to promote steatohepatitis development. Thus, an aberrant increase or decrease of TAK1 activity is 1509

19 YAN, ZHANG, ET AL. HEPATOLOGY, May 2017 detrimental to hepatocyte function, and TAK1 should be kept in balance in order to maintain hepatic homeostasis and physiological functions. A key question that remains to be answered when exploring the mechanisms underlying TRIM8- regulated hepatic steatosis is how TRIM8 activates TAK1. Lys63-linked polyubiquitination is required for the autophosphorylation and activation of TAK1. (47) Coincidentally, TRIM8 has a RING domain that endows it with E3 ligase activity. (48) Here, we report that TRIM8 directly binds to TAK1 and that the RING domain of TRIM8 is mainly responsible for TAK1 s ubiquitination. The capacity of TRIM8 to ubiquitinate TAK1 demonstrated in our present study is consistent with the results reported by Li et al. (26) The evidence that TRIM8 failed to aggravate steatohepatitis and its complications when its binding domain was missing or the RING domain was mutated suggests that both the TRIM8 TAK1 interaction and TRIM8 s E3 ligase activity are required for its regulatory role. In conclusion, our present study reports a novel function for TRIM8 in diet-induced and geneinduced NASH. Based on in vivo and in vitro experiments, we clearly identified that TRIM8 significantly exacerbates insulin resistance, hepatic steatosis, inflammation, and fibrosis by directly binding to and ubiquitinating TAK1 that activates downstream JNK/p38 and NF-jB signaling pathways. Treatments that interfere with the TRIM8 TAK1 interaction or TRIM8 s E3 ligase activity might prove to be a promising strategy for NASH and metabolic syndrome treatment. REFERENCES 1) Tilg H, Moschen AR. Insulin resistance, inflammation, and non-alcoholic fatty liver disease. Trends Endocrinol Metab 2008; 19: ) Birkenfeld AL, Shulman GI. Nonalcoholic fatty liver disease, hepatic insulin resistance, and type 2 diabetes. HEPATOLOGY 2014;59: ) Zhang HJ, He J, Pan LL, Ma ZM, Han CK, Chen CS, et al. Effects of moderate and vigorous exercise on nonalcoholic fatty liver disease: a randomized clinical trial. JAMA Intern Med 2016;176: ) Nascimbeni F, Pais R, Bellentani S, Day CP, Ratziu V, Loria P, et al. From NAFLD in clinical practice to answers from guidelines. J Hepatol 2013;59: ) Review T, LaBrecque DR, Abbas Z, Anania F, Ferenci P, Khan AG, et al. World Gastroenterology Organisation global guidelines: nonalcoholic fatty liver disease and nonalcoholic steatohepatitis. J Clin Gastroenterol 2014;48: ) Versteeg GA, Rajsbaum R, Sanchez-Aparicio MT, Maestre AM, Valdiviezo J, Shi M, et al. The E3-ligase TRIM family of proteins regulates signaling pathways triggered by innate immune pattern-recognition receptors. Immunity 2013;38: ) Rajsbaum R, Garcia-Sastre A, Versteeg GA. TRIMmunity: the roles of the TRIM E3-ubiquitin ligase family in innate antiviral immunity. J Mol Biol 2014;426: ) McNab FW, Rajsbaum R, Stoye JP, O Garra A. Tripartite-motif proteins and innate immune regulation. Curr Opin Immunol 2011;23: ) Vincent SR, Kwasnicka DA, Fretier P. A novel RING finger-b box-coiled-coil protein, GERP. Biochem Biophys Res Commun 2000;279: ) Caratozzolo MF, Micale L, Turturo MG, Cornacchia S, Fusco C, Marzano F, et al. TRIM8 modulates p53 activity to dictate cell cycle arrest. Cell Cycle 2012;11: ) Tomar D, Sripada L, Prajapati P, Singh R, Singh AK, Singh R. Nucleo-cytoplasmic trafficking of TRIM8, a novel oncogene, is involved in positive regulation of TNF induced NF-kappaB pathway. PLoS One 2012;7:e ) Okumura F, Matsunaga Y, Katayama Y, Nakayama KI, Hatakeyama S. TRIM8 modulates STAT3 activity through negative regulation of PIAS3. J Cell Sci 2010;123: ) Zhang XJ, Zhang P, Li H. Interferon regulatory factor signalings in cardiometabolic diseases. Hypertension 2015;66: ) Yang H, Wang H, Shivalila CS, Cheng AW, Shi L, Jaenisch R. One-step generation of mice carrying reporter and conditional alleles by CRISPR/Cas-mediated genome engineering. Cell 2013;154: ) Wang PX, Zhang XJ, Luo P, Jiang X, Zhang P, Guo J, et al. Hepatocyte TRAF3 promotes liver steatosis and systemic insulin resistance through targeting TAK1-dependent signalling. Nat Commun 2016;7: ) Luo P, Wang PX, Li ZZ, Zhang XJ, Jiang X, Gong J, et al. Hepatic oncostatin M receptor beta regulates obesity-induced steatosis and insulin resistance. Am J Pathol 2016;186: ) Ji YX, Zhang P, Zhang XJ, Zhao YC, Deng KQ, Jiang X, et al. The ubiquitin E3 ligase TRAF6 exacerbates pathological cardiac hypertrophy via TAK1-dependent signalling. Nat Commun 2016;7: ) Zhu LH, Wang A, Luo P, Wang X, Jiang DS, Deng W, et al. Mindin/Spondin 2 inhibits hepatic steatosis, insulin resistance, and obesity via interaction with peroxisome proliferator-activated receptor alpha in mice. J Hepatol 2014;60: ) Lee SW, Li CF, Jin G, Cai Z, Han F, Chan CH, et al. Skp2- dependent ubiquitination and activation of LKB1 is essential for cancer cell survival under energy stress. Mol Cell 2015;57: ) Pawlak M, Lefebvre P, Staels B. Molecular mechanism of PPARalpha action and its impact on lipid metabolism, inflammation and fibrosis in non-alcoholic fatty liver disease. J Hepatol 2015;62: ) Anstee QM, Goldin RD. Mouse models in non-alcoholic fatty liver disease and steatohepatitis research. Int J Exp Pathol 2006; 87: ) Nelson JE, Roth CL, Wilson LA, Yates KP, Aouizerat B, Morgan-Stevenson V, et al. Vitamin D deficiency is associated with increased risk of non-alcoholic steatohepatitis in adults with non-alcoholic fatty liver disease: possible role for MAPK and NF-kappaB? Am J Gastroenterol 2016;111: ) Chang L, Karin M. Mammalian MAP kinase signalling cascades. Nature 2001;410: