The E3 ligase tripartite motif 7 drives the progression of non-alcoholic fatty liver disease by targeting DUSP10 degradation in male mice

Abstract

Non-alcoholic fatty liver disease (NAFLD) and its more severe form, nonalcoholic steatohepatitis (NASH), have emerged as a burgeoning global epidemic and impose an enormous socioeconomic burden. However, the lack of effective pharmacotherapies is due to incomplete understanding of the molecular mechanisms of NASH. In the present study, we observe that E3 ligase TRIM7 expression is significantly increased in both liver tissues and hepatocytes from NAFLD models. In vivo gain- and loss-of-function experiments reveal that hepatic-specific TRIM7 deletion significantly alleviates hepatic steatosis, inflammation and insulin resistance in diet-induced male mouse models, whereas overexpression of wild-type TRIM7 (but not its E3-deficient mutant) shows diametrically opposite effects. Mechanistic studies reveal that TRIM7 interacts with and catalyzes DUSP10 ubiquitination and proteasomal degradation, thus leading to hyperactivation of IKKβ-NF-κB and JNK/p38 MAPK signaling pathways. Importantly, silencing DUSP10 expression abrogates the protective effects of hepatic TRIM7 deficiency on NAFLD-related pathological phenotypes. Collectively, our findings identify TRIM7 as a key regulator of the pathogenesis of NAFLD/NASH and provide a promising therapeutic strategy for NAFLD by targeting the TRIM7-DUSP10 axis.

Metabolic dysfunction-associated steatohepatitis (MASH) has an incompletely understood mechanistic basis. Here the authors use gain- and loss-of-function mouse models to show that that E3 ligase TRIM7 in hepatocyte promotes MASH pathogenesis via DUSP10 ubiquitination.

Introduction

With the rising prevalence of obesity in both the adult and pediatric populations, nonalcoholic fatty liver disease (NAFLD) has become a burgeoning global epidemic¹. Given its strong association with obesity, type 2 diabetes (T2D) and metabolic abnormalities, NAFLD has been redefined as metabolic dysfunction-associated steatotic liver disease (MASLD)^2,3. NAFLD is a progressive disease that begins with excessive hepatic lipid accumulation (simple steatosis), which can progress into a more severe form, NASH, and even NASH-related cirrhosis and hepatocellular carcinoma if left untreated^4,5. In addition, NAFLD increases the risk of extrahepatic chronic diseases, such as cardiovascular diseases, and chronic kidney disease⁶. Despite the recent FDA approval of Resmetirom—the only drug currently available for NAFLD—this condition still affects over a third of the global population, with prevalence expected to surge to 55.7% by 2040, due to the limited therapeutic efficacy^1,7,8. The enormous socioeconomic burden imposed by NAFLD has prompted an urgent need to further elucidate the molecular mechanisms of NAFLD progression and develop potential therapeutic targets.

Increasing evidence suggests that hepatic inflammatory response and insulin resistance (IR) are the two key drivers of NAFLD development and progression^9,10. At the molecular level, the dysregulation of critical signaling molecules leads to aberrant activation of inflammatory and insulin-related pathways, triggering hepatic inflammation and IR^11–13. Ubiquitination, one common post-translational modification (PTM), orchestrates diverse cellular functions, including signaling cascade activation, protein turnover, and subcellular localization¹⁴. Emerging evidence demonstrated that the aberrant ubiquitination/deubiquitination modifications are intricately linked to metabolic dysfunction, such as insulin resistance and hepatic inflammation^15–17. Therefore, targeting critical ubiquitination regulators involved in inflammation and insulin signaling pathway has emerged as potential therapeutic strategies for NAFLD. In order to dig out key ubiquitination-related molecules (E3 ligase and deubiquitinating enzymes) involved in NAFLD progression, we performed phenotype-based RNA sequencing using liver tissues from two independent NAFLD mice models. We then intersected the differentially expressed genes (DEGs) with ubiquitination-related genes (UbRGs) from the iUUCD 2.0 database (http://iuucd.biocuckoo.org/). The E3 ubiquitin ligase TRIM7 was identified as a top candidate in NAFLD development and progression.

TRIM7, a novel member of the tripartite motif (TRIM) family of E3 ubiquitin ligases, contains a canonical N-terminal RBCC motif (comprising one RING finger, one B-box and one coiled-coil domain) and a C-terminal PRY/SPRY domain¹⁸. Initially identified as glycogenin-interacting protein (GNIP), TRIM7 was first implicated in glycogen metabolism^19,20. Subsequent studies have demonstrated that TRIM7 plays critical roles in tumorigenesis and metastasis, antiviral response and TLR4-mediated innate immunity, with these functions largely dependent on its E3 ligase activity^21–25. Notably, two independent studies have reported opposing effects of TRIM7 on hepatocellular carcinoma progression^26,27, highlighting the context-dependent functional in different physiological and pathological conditions. However, the role of TRIM7 and its underlying molecular mechanisms in the development and progression of NAFLD remain entirely uncharacterized.

In the present study, we observed dramatic upregulation of TRIM7 expression in both in vivo and in vitro NAFLD models. Hepatocyte-specific Trim7 knockout significantly alleviated diet-induced hepatic steatosis, inflammation, insulin resistance, and fibrosis in mouse models, whereas hepatic Trim7 overexpression exacerbated NAFLD phenotypes. Mechanistically, we identified a direct interaction between TRIM7 and DUSP10 (dual-specificity phosphatase 10, also called MKP5), a known negative regulator of JNK and p38 MAPK signaling by dephosphorylating their serine/threonine and tyrosine residues^28,29. Dysregulation of DUSP10 has been implicated in inflammatory diseases, immune disorders²⁹, and obesity-related metabolic dysfunction-where its deficiency promotes brown adipocyte differentiation but inducing insulin resistance and inflammation in white adipose tissue^30,31. Given the critical roles of insulin resistance and inflammation in NAFLD progression, we investigated the regulatory axis of TRIM7 and DUSP10. Our results demonstrate that TRIM7 catalyzes K63-linked polyubiquitination of DUSP10, triggering its proteasomal degradation and consequent hyperactivation of the IKKβ-NF-κB and JNK/p38 signaling pathways under metabolic stress. Importantly, hepatic Dusp10 silencing in Trim7 knockout mice abolished the protective effects of TRIM7 deficiency on NAFLD progression. Collectively, these findings establish the hepatic TRIM7 -DUSP10 axis as a novel and promising therapeutic target for NAFLD.

Results

TRIM7 expression is upregulated in the development and progression of NAFLD

To comprehensively characterize the role of ubiquitination modifications in NAFLD pathophysiology, we performed unbiased RNA-seq on the liver tissues from two independent NAFLD mice models to screen for key regulatory proteins (Supplementary Fig. 1a). Among differentially expressed genes (DEGs) encoding E3 ligases and deubiquitinating enzymes, Trim7 exhibited robust upregulation in both models (Fig. 1a, Supplementary Fig. 1b). RT-qPCR confirmed significant Trim7 mRNA upregulation in livers of ob/ob mice as well as HFD or HFFC (high-fat, high-fructose, high-cholesterol diet)-fed mice compared to NCD controls (Fig. 1b, c). Consistent with transcriptomic data, immunoblotting and immunohistochemistry revealed a substantial increase in TRIM7 protein expression in NAFLD mouse livers versus NCD controls (Fig. 1d–f). Time-course analysis showed progressive TRIM7 protein upregulation with prolonged HFD or HFFC feeding (Fig. 1g, h, Supplementary Fig. 1c, d). Additionally, TRIM7 expression was significantly elevated in liver biopsies from histologically diagnosed NAFLD patients compared to non-steatotic controls (Fig. 1i–k, Supplementary Fig. 1e). To determine the cellular source of TRIM7 upregulation, primary mouse liver cells were isolated and TRIM7 was found predominantly expressed in hepatocytes (Supplementary Fig. 1f). The immunofluorescent staining further revealed that TRIM7 protein was mainly localized to the hepatocyte cytoplasm (Fig. 1l). In vitro, TRIM7 mRNA and protein levels were also markedly induced in primary hepatocytes and hepatic cell lines treated with palmitic and oleic acids (POA) mixture (Fig. 1m, n, Supplementary Fig. 1g–j). Collectively, these results demonstrate the upregulation of TRIM7 in hepatocytes during NAFLD development and progression.

Fig. 1. TRIM7 expression is increased in the development and progression of NAFLD.

a TRIM7 gene expression data from RNA-Seq analysis of liver tissues from the indicated mouse models. b, c RT-qPCR analysis of relative TRIM7 mRNA expression in liver tissues of ob/ob mice (12 weeks) and C57BL6/J mice fed with HFD for 12 weeks (NCD, HFD group n = 8; ob/ob group n = 6; P values determined by one-way ANOVA) (b) or HFFC fed for 19 weeks (n = 10 per group; P values determined by 2 tailed t test) (c). d, e Western Blot (d) and quantitative analysis (e) of TRIM7 protein expression in liver tissues (ob/ob group n = 3; NCD, HFD, HFFC group n = 4; P < 0.001 versus NCD group by 2 tailed t test). f Representative immunohistochemistry (IHC) staining images of TRIM7 in liver tissues from mice fed NCD, HFD, or HFFC diet (n = 3). g, h Analysis of TRIM7 protein expression in liver tissues of mice fed a HFD and HFFC diet for different time (n = 3/group, P values determined by one-way ANOVA). i, j TRIM7 protein expression analysis in liver tissues from non-steatosis and NAFLD patients (non-steatosis group n = 6; NAFLD group n = 8; P < 0.001 by 2 tailed t test). k Representative immunohistochemistry staining images of TRIM7 in liver tissues from non-steatosis and NAFLD patients (n = 3). l Representative immunofluorescence staining images of TRIM7 and HNF4 in liver tissues from HFD-fed mice for 12 weeks (n = 3). m, n Western Blot (m) and quantitative analysis (n) of TRIM7 protein expression in murine primary hepatocytes treated with POA or BSA (n = 4; P < 0.001 versus BSA group by 2 tailed t test). Data are presented as mean ± SD. Source data are provided as a Source Data file.

Hepatocyte-specific TRIM7 deficiency ameliorates HFD-induced metabolic disturbance and liver injury

To investigate the biological effects of TRIM7 on NAFLD in vivo, we generated hepatocyte-specific Trim7 knockout mice (Trim7-HKO), and confirmed Trim7 deficiency in the livers of Trim7-HKO mice (Supplementary Fig. 2a–d). Both Trim7-HKO mice and their littermate controls (Trim7-Flox) were then fed an HFD or NCD feeding for 14 weeks. Under NCD conditions, no significant differences in body weights were observed between Trim7-HKO and Trim7-Flox mice. However, Trim7-HKO mice exhibited lower body weights than Trim7-Flox mice when fed an HFD (Fig. 2a, Supplementary Fig. 2e). Of note, food intake was comparable between the two mouse genotypes in both dietary groups (Supplementary Fig. 2f). After 14 weeks of HFD feeding, Trim7-HKO mice showed significantly decreased fasting blood glucose, insulin levels, and homeostatic model assessment of insulin resistance (HOMA-IR) (Fig. 2b–d). Additionally, Trim7 deficiency effectively improved HFD-induced glucose metabolism disturbance, as indicated by intraperitoneal glucose tolerance tests (IPGTTs) and intraperitoneal insulin tolerance tests (IPITTs) (Fig. 2e, f). Western Blot analyses revealed that Trim7 knockout increased the phosphorylation of AKT and GSK3β while dramatically reducing the expression of gluconeogenesis-related genes (PEPCK and G6Pase) under HFD conditions, suggesting restoration of HFD-disrupted hepatic insulin signaling in Trim7-HKO mice (Fig. 2g, Supplementary Fig. 2g–i). Correspondingly, liver glycogen content was markedly higher in Trim7-HKO mice than in Trim7-Flox mice after HFD challenge as shown in Fig. 2h. Moreover, Trim7 deficiency also influenced liver lipid metabolism. On one hand, Trim7-HKO mice exhibited smaller and lighter livers than their littermate controls under HFD feeding, whereas no obvious difference in liver appearance or weight were observed under NCD conditions (Fig. 2i, j, Supplementary Fig. 2j). On the other hand, histological analysis (H&E and Oil Red O staining) and NAFLD Activity Score (NAS) assessment confirmed that Trim7 knockout substantially attenuated hepatic steatosis compared to the control group after HFD treatment, while Trim7 deficiency had no significant effect on liver histological structure in NCD-fed mice (Fig. 2k, l, Supplementary Fig. 2k). Further analysis of hepatic TG, TC, NEFA levels, and lipid metabolism-related gene expression supported the protective role of Trim7 deletion against hepatic steatosis (Supplementary Fig. 2l, m). Furthermore, Trim7 knockout displayed reduced liver injury under HFD conditions, as evidenced by lower ALT and AST levels in the livers of Trim7-HKO mice (Supplementary Fig. 2n). Finally, the effect of Trim7 deficiency on hepatic inflammation was evaluated. The expression of pro-inflammatory cytokines (TNF-α, IL-6 and Ccl2) were significantly decreased in the serum and livers of HFD-fed Trim7-HKO mice, while the anti-inflammatory cytokine IL-10 was increased compared to controls (Fig. 2m–o, Supplementary Fig. 2o). Immunofluorescence staining for F4/80-positive inflammatory cells in liver tissue revealed fewer infiltrating macrophages in the livers of Trim7-HKO mice under HFD conditions (Fig. 2p, Supplementary Fig. 2p). Collectively, these data indicated that Trim7 deficiency ameliorates HFD-induced metabolic disturbance and liver injury.

Fig. 2. Hepatocyte-specific TRIM7 deficiency ameliorates HFD-induced metabolic disturbance and liver injury.

a–d Body weight (a), fasting blood glucose levels (b), fasting insulin levels (c), and HOMA-IR (d) in Trim7-HKO and Trim7-Flox mice at the indicated time points during feeding with NCD or HFD (n = 8; P values determined by one-way ANOVA). e, f The IPGTT (e) and IPITT (f) were performed in Trim7-HKO and Trim7-Flox mice at 12 weeks and 13 weeks of NCD or HFD feeding, respectively, and the areas under the curve (AUC) were calculated for both tests (n = 8; P values determined by one-way ANOVA); *P < 0.05, **P < 0.01, ***P < 0.001 vs. NCD-Trim7-Flox group; #P < 0.05, ##P < 0.01 vs. HFD-Trim7-Flox group. g Immunoblotting analysis of the phosphorylated and total expression levels of the indicated proteins in the insulin signaling pathway within liver tissues (n = 3). h Representative PAS staining images of liver tissues from Trim7-HKO and Trim7-Flox mice fed a NCD or HFD for 14 weeks (n = 4). i, j The liver weight (i), liver weight to body weight ratio (j) in Trim7-HKO and Trim7-Flox mice fed a NCD or HFD for 14 weeks (n = 8; P values determined by one-way ANOVA). k, l H&E and Oil red O staining (k), quantitative analysis of Oil red O positive area (n = 5; P < 0.001 by 2 tailed t test) (l) in liver tissues from Trim7-HKO and Trim7-Flox mice fed a NCD or HFD for 14 weeks. m–o Serum levels of TNFα (m), IL-6 (n) and IL-10 (o) in Trim7-HKO and Trim7-Flox mice fed a NCD or HFD for 14 weeks (n = 8; P values determined by one-way ANOVA). p Immunofluorescence staining for F4/80 (macrophage marker) in liver sections from HFD-fed Trim7-HKO and Trim7-Flox mice (n = 6). Data are presented as mean ± SD. Source data are provided as a Source Data file.

TRIM7-HKO alleviates HFFC-induced NASH

To further explore the role of TRIM7 in NASH, one advanced stage of NAFLD, we established a mouse NASH model by subjecting Trim7-HKO mice and their littermate controls to a HFFC diet for 19 weeks. After HFFC feeding, Trim7 knockout mice exhibited significantly lower body weights, liver weights and liver to body weight ratio (LW/BW) than Trim7-Flox mice, with no significant intergroup differences in food intake (Fig. 3a–c, Supplementary Fig. 3a–c). Notably, Trim7 deletion markedly attenuated HFFC-induced abnormalities in fasting blood glucose, fasting insulin, HOMA-IR, and glucose intolerance (Fig. 3d–h). Consistent with these findings, hepatic insulin signaling activity was restored, while gluconeogenesis-related gene expression was suppressed in liver tissues of HFFC-fed Trim7-HKO mice compared to Trim7-Flox controls (Fig. 3i, Supplementary Fig. 3d–f). Periodic acid-Schiff (PAS) staining further revealed higher glycogen content in Trim7-HKO livers than the controls under HFFC conditions, with no significant changes under NCD status (Fig. 3j). In addition, Trim7 knockout dramatically improved HFFC-induced hepatic steatosis, as evidenced by reduced lipid accumulation in histological analyses (H&E and Oil Red O staining), decreased hepatic lipid content (TG, TC, NEFA), and altered lipid metabolism-related gene expression (Fig. 3k–l and Supplementary Fig. 3g–i). The hepatic dysfunction and inflammation response were also ameliorated in Trim7-HKO mice compared to control mice under HFFC feeding (Fig. 3m, n, Supplementary Fig. 3j–l). What’s more, Trim7 deficiency strongly reduced liver collagen deposition, as indicated by Masson staining and downregulation of profibrotic markers (α-SMA, Col1) (Fig. 3o, Supplementary Fig. 3m–o). In summary, these findings demonstrate that Trim7 knockout alleviates HFFC-induced NASH development by improving metabolic dysfunction, reducing steatosis, inflammation, and fibrosis.

Fig. 3. TRIM7-HKO alleviates HFFC-induced NASH.

a–f Body weight (a), liver weight (b), liver weight-to-body weight ratio (c), fasting blood glucose levels (d), fasting insulin levels (e) and HOMA-IR (f) in Trim7-HKO and Trim7-Flox mice at the indicated time points during feeding with NCD or HFFC (n = 8; a performed by 2 tailed t test, b-f performed by one-way ANOVA). g, h The IPGTT (g) and IPITT (h) were performed at 17 weeks and 18 weeks of HFFC feeding, respectively, and the areas under the curve (AUC) were calculated for both tests (n = 8; P values determined by 2 tailed t test). i Immunoblotting analysis of phosphorylated and total expression levels of the indicated proteins in the insulin signaling pathway in liver tissues (n = 3). j–m Representative PAS staining images of the liver (j), liver H&E and Oil red O staining images (k), hepatic contents of TG, TC and NEFA (l), and serum levels of TNFα, IL-6 and IL-10 (m) in Trim7-HKO and Trim7-Flox mice fed a NCD or HFFC diet for 19 weeks (n = 8; P values determined by one-way ANOVA). n, o Representative immunofluorescence staining images for F4/80 (n) and Sirius Red staining images (o) in liver sections from Trim7-HKO and Trim7-Flox mice fed a HFFC diet (n = 4). Data are presented as mean ± SD. Source data are provided as a Source Data file.

Hepatic TRIM7 overexpression exacerbates HFD/HFFC-induced steatohepatitis

To further validate the role of hepatic Trim7 in NASH development, we generated hepatocyte-specific Trim7 overexpression mice via tail vein injection of liver-targeting recombinant adeno-associated virus serotype 8 (AAV8) encoding Flag-tagged Trim7 wildtype (AAV8-Trim7) or its E3 dead ligase mutant (Trim7-C29A, C32A; AAV8-Trim7-DL) into mice fed NCD or HFD. The AAV8 vector also co-expressed enhanced green fluorescent protein (EGFP) for viral transduction tracking, so AAV8-EGFP was used as the control group. All mice were monitored throughout the treatment period and sacrificed 10 weeks after AAV8 administration (Fig. 4a). EGFP fluorescent imaging confirmed the virus successful enrichment in liver (Supplementary Fig. 4a), and immunohistochemistry/ immunoblotting verified effective hepatic overexpression of Trim7 wildtype and Trim7-DL proteins in their respective groups compared to controls (Fig. 4b, Supplementary Fig. 4b). Under NCD conditions, no differences in body weight, liver weight, liver-to-body weight ratio (LW/BW), or glucose metabolism were observed among groups. However, HFD-fed mice overexpressing Trim7 wildtype exhibited significantly higher body weights, liver weights, LW/BW, fasting blood glucose levels, and HOMA-IR values than EGFP controls (Fig. 4c–h, Supplementary Fig. 4c–e). Trim7 wildtype overexpression further exacerbated glucose intolerance, impaired insulin sensitivity, and suppressed both insulin signaling pathway activity and hepatic glycogen accumulation (Fig. 4i–k, Supplementary Fig. 4f, g). Furthermore, hepatic lipid accumulation induced by HFD was significantly aggravated by Trim7 wildtype overexpression, as evidenced by H&E and Oil Red O staining, elevated TG, TC, and NEFA levels, and dysregulated lipid metabolism-related genes expression (Fig. 4l, m, Supplementary Fig. 4h–k). Additionally, Trim7 wildtype overexpression further worsened HFD-induced liver injury and inflammation, characterized by increased serum markers of liver dysfunction, elevated pro-inflammatory cytokines, decreased anti-inflammatory cytokines, and enhanced hepatic macrophage infiltration (Fig. 4n–p, Supplementary Fig. 4l–n). Notably, overexpression of the E3 dead ligase Trim7-DL mutant did not affect these NAFLD-related phenotypes under HFD conditions (Fig. 4). Similar trends were observed in HFFC-fed mice, where Trim7 wildtype overexpression exacerbated weight gain, metabolic dysfunction, hepatic steatosis, liver injury, and inflammation (Fig. 5a–p, Supplementary Fig. 5a–l). Moreover, Trim7 wildtype but not Trim7-DL overexpression promoted liver collagen deposition (Sirius Red staining) and upregulated profibrotic markers (α-SMA, Col1) (Fig. 5q, Supplementary Fig. 5m, n). Taken together, these results demonstrate that hepatic TRIM7 promotes NAFLD/NASH progression in an E3 ligase activity-dependent manner.

Fig. 4. Hepatic TRIM7 overexpression exacerbates HFD-induced steatohepatitis.

a Schematic of the AAV8-mediated Trim7 overexpression model in NCD- or HFD-fed mice. b Representative immunohistochemistry staining images of Trim7 in the liver of AAV8-injected mice (n = 5). c–h Body weight (c), liver weight (d), liver weight-to-body weight ratio (e), fasting blood glucose levels (f), fasting insulin levels (g), and HOMA-IR (h) in AAV8-injected mice fed a NCD or HFD for 14 weeks (n = 5; P values determined by one-way ANOVA). i–j The IPGTT (i) and IPITT (j) were performed at 12 weeks and 13 weeks of HFD feeding, respectively, and the areas under the curve (AUC) were calculated for both tests (n = 5; P values determined by one-way ANOVA). k–m Immunoblotting analysis of the indicated protein expression (k), H&E and Oil red O staining images (l), and quantitative analysis of Oil red O positive area (m) in liver tissues of AAV8-injected mice fed a HFD for 14 weeks (n = 5; P values determined by one-way ANOVA). n, o Serum levels of TNFα, IL-6 (n) and IL-10 (o) in AAV8-injected mice fed a NCD or HFD for 14 weeks (n = 10; P values determined by one-way ANOVA). p Immunofluorescence staining for F4/80 in liver sections of AAV8-injected mice fed a HFD for 14 weeks (n = 5). Data are presented as mean ± SD. Source data are provided as a Source Data file.

Fig. 5. Hepatic TRIM7 overexpression exacerbates HFFC-induced steatohepatitis.

a Schematic illustration of the AAV8-mediated Trim7 overexpression model in HFFC-fed mice. b Representative immunohistochemistry staining images of Trim7 in the liver of AAV8-injected mice (n = 5). c–h Body weight (c), liver weight (d), liver weight-to-body weight ratio (e), fasting blood glucose levels (f), fasting insulin levels (g) and HOMA-IR (h) in AAV8-injected mice fed a HFFC diet for the indicated time (n = 5; P values determined by one-way ANOVA). i, j The IPGTT (i) and IPITT (j) were performed after 17 weeks and 18 weeks of HFFC diet feeding, respectively, and the areas under the curve (AUC) were calculated for both assays (n = 5; P values determined by one-way ANOVA). k–m Immunoblotting analysis of the indicated protein expression in liver tissues (k), H&E staining and Oil red O staining images of liver sections (l), and quantitative analysis of Oil Red O positive areas in liver tissues (m) in AAV8-injected mice fed a HFFC diet for 19 weeks (n = 5; P values determined by one-way ANOVA). n–q Serum levels of IL-10 (n) and TNFα, IL-6 (o) (n = 10; P values determined by one-way ANOVA); liver immunofluorescence staining images for F4/80 (p), and Sirius Red staining images (q) in AAV8-injected mice fed a HFFC diet for 19 weeks. Data are presented as mean ± SD. Source data are provided as a Source Data file.

TRIM7 regulates steatohepatitis through the TAK1-JNK/p38 signaling pathway

To systemically elucidate the molecular mechanisms underlying TRIM7-mediated regulation of NAFLD/NASH, we performed RNA-seq on liver tissues from HFD-fed Trim7-HKO mice and their littermate controls. RNA-seq analysis identified 229 differentially expressed genes (DEGs), including 110 upregulated and 119 downregulated genes (Fig. 6a). KEGG pathway analysis revealed that Trim7 deficiency significantly modulated genes involved in fatty acid metabolism, inflammation and insulin signaling, with the MAPK signaling pathway and inflammatory responses ranked among the most enriched categories (Fig. 6b). We subsequently validated the activation of the MAPK pathway and the canonical inflammatory pathway (IKKβ-NF-κB) in liver tissues of Trim7-knockout and Trim7-overexpression mice under HFD or HFFC feeding. The Immunoblotting results showed that both HFD and HFFC feeding dramatically activated MAPK signaling, as evidenced by increased phosphorylation of ERK, JNK, and p38. Notably, Trim7 overexpression further exacerbated the phosphorylation of JNK and p38 (but not ERK), while Trim7 deficiency attenuated HFD/HFFC-induced JNK and p38 activation (Fig. 6c, d, Supplementary Fig. 6a). Consistent with the observed inflammatory responses, Trim7 overexpression obviously enhanced IKKβ and NF-κB phosphorylation in HFD/HFFC-fed mice, whereas Trim7 knockout restrained these events (Fig. 6c, d, Supplementary Fig. 6a). Given the crucial role of TAK1 as an upstream regulator of both MAPK and IKKβ-NF-κB pathways in NASH pathogenesis¹¹, we assessed TAK1 activation in relation to Trim7 expression. The results indicated that HFD/HFFC treatment markedly activated TAK1, and this activation was further enhanced by wildtype Trim7 (but not E3 ligase mutant Trim7-DL) overexpression, while Trim7 knockout inhibited TAK1 activation (Fig. 6c, d, Supplementary Fig. 6a). Altogether, these data demonstrate that TRIM7, via its E3 ligase activity, promotes NAFLD progression by activating the TAK1-JNK/p38 signaling axis.

Fig. 6. TRIM7 regulates steatohepatitis through the TAK1-JNK/p38 signaling pathway.

a Volcano plot showing the distribution of all detected genes from RNA sequencing in liver tissues of Trim7-HKO and Trim7-Flox mice after 14 weeks of HFD feeding (n = 3). Differentially expressed genes (DEGs) analysis was performed using the two-sided Wald test, which was selected to account for both potential upregulation and downregulation of genes between the two groups. Genes with a P-value < 0.05 and |log2 (fold change)|≥1.0 were defined as DEGs; red dots represent significantly upregulated DEGs, blue dots represent significantly downregulated DEGs, and black dots represent non-significant genes. b KEGG pathway enrichment analysis of the aforementioned DEGs. c, d Immunoblotting (c) and quantitative (d) analysis of the expression levels of the indicated proteins in liver tissues from NCD- or HFFC-fed Trim7-Flox, Trim7-HKO mice (left panel) and AAV8-injected mice (right panel) (n = 3; P values determined by one-way ANOVA). Data are presented as mean ± SD. Source data are provided as a Source Data file.

TRIM7 interacts with DUSP10 and promotes its degradation via K63-linked ubiquitination

To identify the key molecule through which TRIM7 regulates the TAK1-JNK/p38 pathway, we performed IP-MS analysis (Fig. 7a) and identified DUSP10 as a potential TRIM7-interacting partner. Subcellular localization studies showed that TRIM7 and DUSP10 exhibit cytoplasmic colocalization under basal conditions, with TRIM7 predominantly in the cytoplasm and DUSP10 distributed in both the cytoplasm and nucleus (Supplementary Fig. 7a–c). However, DUSP10 nuclear localization significantly decreased, with enhanced cytoplasmic colocalization with TRIM7 in liver tissues under HFD conditions (Supplementary Fig. 7c). Co-immunoprecipitation (co-IP) assays confirmed that both exogenous and endogenous TRIM7 interact with DUSP10 (Fig. 7b, Supplementary Fig. 7d). To map the interaction domains, we constructed a series of Flag-tagged TRIM7 and HA-tagged DUSP10 truncated mutants (Fig. 7c). Co-IP results identified the PRY/SPRY domain of TRIM7 and the N-terminal domain of DUSP10 as essential for their binding (Fig. 7d, e, Supplementary Fig. 7e, f). In vivo ubiquitination assays showed that wildtype TRIM7 (but not its E3 dead ligase mutant) dramatically enhanced DUSP10 ubiquitination level (Fig. 7f, Supplementary Fig. 7g, h). Conversely, TRIM7 knockdown significantly reduced DUSP10 ubiquitination (Fig. 7g, Supplementary Fig. 7i). Polyubiquitin chain type analysis revealed that TRIM7 specifically catalyzes K63-linked polyubiquitination of DUSP10 (Fig. 7h, Supplementary Fig. 7j). Mutation of lysine 425 to arginine in DUSP10 (DUSP10-K425R) abrogated TRIM7-mediated ubiquitination, identifying K425 as the critical ubiquitination site (Fig. 7i, Supplementary Fig. 7k). Further study showed no significant changes in DUSP10 mRNA levels upon TRIM7 overexpression or knockout (Supplementary Fig. 7l–p), whereas wildtype TRIM7 (but not the E3 mutant) reduced DUSP10 protein expression in a dose-dependent manner (Fig. 7j). Cycloheximide (CHX) chase assays revealed that wildtype TRIM7 shortened the half-life of DUSP10 protein (whereas its E3 ligase-dead mutant did not), while TRIM7 knockout prolonged the DUSP10 protein stability (Fig. 7k, Supplementary Fig. 7q, r). Treatment with the proteasome inhibitor MG132 (but not lysosome inhibitors chloroquine or NH4Cl) rescued TRIM7-induced DUSP10 degradation, indicating that TRIM7 promotes DUSP10 degradation via the ubiquitin-proteasome pathway (Fig. 7l). Finally, analysis of NAFLD model samples confirmed significant downregulation of DUSP10 protein, with a negative correlation between TRIM7 and DUSP10 expression levels (Fig.7m, Supplementary Fig. 7s–v). In summary, all above results suggested that TRIM7 interacts with DUSP10 and promotes its degradation via K63-linked ubiquitination at lysine 425.

Fig. 7. TRIM7 interacts with DUSP10 and facilitates degradation via K63-linked ubiquitination.

a Schematic of the design for IP Flag-TRIM7 and subsequent LC-MS/MS analysis. b Co-IP assay to examine the interaction between endogenic TRIM7 and DUSP10 proteins. c Schematic representation of full-length (FL) TRIM7 and DUSP10, as well as their truncated mutant constructs. d, e IP mapping analysis of the interaction domains between TRIM7 and DUSP10 in HEK293T cells co-transfected with plasmids encoding HA-DUSP10 and Flag-TRIM7 or its truncated mutants (d); or plasmids encoding Flag-TRIM7 and HA-DUSP10 or its truncated mutants (e). f, g Ubiquitination of DUSP10 in HEK293T cells co-transfected with Myc-Ub, HA-DUSP10, and Flag-tagged TRIM7 wildtype or its E3 dead ligase mutant plasmids (f) or TRIM7 knockdown plasmids (g). h Ubiquitination screening to identify the ubiquitin type of DUSP10 catalyzed by TRIM7 in HEK293T cells co-transfected with HA-DUSP10, Flag-TRIM7, and the indicated Myc-Ub constructs. i Ubiquitination analysis of DUSP10-WT and DUSP10-K425R in 293T cells transfected with or without Flag-TRIM7 plasmids. j Immunoblotting analysis of DUSP10 expression in 293T cells co-transfected with HA-DUSP10 and different dose of Flag-tagged TRIM7 wildtype or TRIM7-DL plasmids. k Immunoblotting (left panel) and quantification (right panel) of DUSP10 protein levels to assess the effect of TRIM7-WT or TRIM7-DL on DUSP10 half-life via CHX chase assay (n = 3; P values determined by one-way ANOVA). Data are presented as mean ± SD. l Western blot analysis of DUSP10 expression in 293T cells transfected with the indicated plasmids, followed by treatment with CHX alone or in combination with MG132, CHQ or NH₄Cl prior to cell harvest. m Western Blot analysis of DUSP10 and TRIM7 protein expression in liver tissues from mice fed a HFD (left panel) or HFFC (right panel) diet (n = 4). These data were obtained from at least three independent experiments using independent samples. Source data are provided as a Source Data file.

DUSP10 is essential for TRIM7-mediated function in steatohepatitis

To determine whether DUSP10 is indispensable for TRIM7’s function in NAFLD, we first investigated its role in the TRIM7-regulated TAK1-JNK/p38 signaling pathway. In POA-treated HepG2 cells stably overexpressing TRIM7, wildtype DUSP10 overexpression significantly inhibited POA-induced activation of the TAK1-JNK/p38 pathway, whereas the phosphatase-dead mutant (DUSP10-C408S) had no effect (Supplementary Fig. 8a). This suggests DUSP10 inhibits this pathway via dephosphorylation. To further confirm the physiological relevance of DUSP10 in TRIM7-mediated NAFLD progression in vivo, we injected hepatocyte-specific DUSP10-knockdown adeno-associated virus (AAV8-shDUSP10) into HFD-fed Trim7-HKO mice via tail vein (Fig. 8a). Ten weeks post-injection, strong EGFP fluorescence and western blot analysis confirmed efficient DUSP10 knockdown in the liver (Fig. 8b, c, Supplementary Fig. 8d). Notably, DUSP10 knockdown reversed the protective effects of Trim7 deficiency, as evidenced by increased liver weight and LW/BW ratio, impaired glucose tolerance and insulin sensitivity, and reduced hepatic glycogen storage (Fig. 8d–l, Supplementary Fig. 8e). Additionally, DUSP10 downregulation abolished Trim7-HKO-mediated improvements in HFD-induced lipid accumulation, liver injury and inflammatory response (Fig. 8m–p, Supplementary Fig. 8f–k). Importantly, silencing DUSP10 restored TAK1-JNK/p38 pathway activation in livers of HFD-fed Trim7-HKO mice (Fig. 8q, Supplementary Fig. 8l). Consistently, DUSP10 knockdown also diminished the protective effects of Trim7 knockout against NASH progression in HFFC-fed Trim7-HKO mice (Supplementary Figs. 9, 10). All in all, these results demonstrated that DUSP10 is an essential downstream effector of TRIM7 in NAFLD/NASH pathogenesis through regulation of the TAK1-JNK/p38 pathway.

Fig. 8. DUSP10 is essential for TRIM7-mediated function in steatohepatitis.

a Schematic illustration of the AAV8-mediated DUSP10 downregulation in the liver of HFD-fed Trim7-HKO mice. b, c Immunoblotting analysis (b) and quantitative (c) analysis of DUSP10 expression in liver tissues from the above AAV8-injected mice (n = 4; P values determined by 2 tailed t test). d–h Liver weight (d), liver weight/body weight ratio (e), fasting blood glucose levels (f), fasting insulin levels (g), and HOMA-IR (h) in the above AAV8-injected mice after 14 weeks of HFD feeding (n = 8; P values determined by 2 tailed t test). i, j The IPGTT (i) and IPITT (j) were conducted on the above AAV8-injected mice at 12 weeks and 13 weeks of HFD feeding, respectively. Subsequently, the areas under the curve (AUC) for each test were calculated (n = 8; P values determined by 2 tailed t test). k Immunoblotting analysis of the indicated proteins expression in liver tissues from the above AAV8-injected mice after 14 weeks of HFD feeding (n = 4). l–n Liver PAS staining images (l), liver H&E and Oil red O staining images (m), and quantitative analysis of Oil Red O-positive areas in liver tissues (n = 5; P values determined by 2 tailed t test). o ELISA analysis of the serum levels of TNFα, IL-6, and IL-10 in the above AAV8-injected mice after 14 weeks of HFD feeding (n = 8; P values determined by 2 tailed t test). p Immunofluorescence staining images for F4/80 in liver tissues from the above AAV8-injected mice (n = 6). q Immunoblotting analysis of the indicated protein expression in liver tissues from the above AAV8-injected mice (n = 4). Data are presented as mean ± SD. Source data are provided as a Source Data file.

Discussion

The escalating prevalence of nonalcoholic fatty liver disease (NAFLD) and its advanced form, nonalcoholic steatohepatitis (NASH), has become a critical public health concern. However, current effective pharmacotherapies remains insufficient due to the incomplete elucidation of their underlying molecular mechanisms¹. Therefore, in-depth exploration of the pathogenic mechanisms underlying NAFLD/NASH has emerged as an urgent priority for the development of promising therapeutic targets. In this study, we identified E3 ubiquitin ligase TRIM7 as a pivotal driver of NAFLD/NASH pathogenesis, thereby providing novel mechanistic insights into the disease.

TRIM7, as an E3 ubiquitin ligase, is implicated in diverse physiological and pathological processes through the ubiquitination of its specific substrates. Emerging evidence has demonstrated critical roles for TRIM7 in regulating innate immunity and antiviral response^23,25,32,33. Additionally, existing literature highlights that aberrant TRIM7 expression is associated with tumorigenesis and metastasis in multiple cancers, such as hepatocellular carcinoma (HCC), lung cancer, and gastric cancer^22,26. Notably, two independently studies have reported opposing effects of TRIM7 on HCC progression^26,27, underscoring the complexity and functional heterogeneity of TRIM7 in different tumor microenvironments. Intriguingly, despite its roles in diverse cellular processes, hepatic-specific TRIM7 knockout mice exhibit normal fertility and survival to adulthood, suggesting a limited role in liver tissue development. In the present study, we observed a dramatic upregulation of TRIM7 in both NAFLD liver tissues and hepatocytes. In vivo studies demonstrated that hepatic-specific TRIM7 deficiency significantly alleviated diet-induced hepatic steatosis, insulin resistance, and inflammation. Conversely, overexpression of wild-type TRIM7 in hepatocytes markedly exacerbated these pathological phenotypes, while the E3-deficient TRIM7 mutant failed to exert such effects, indicating that the E3 ligase activity of TRIM7 is essential for its detrimental role in NAFLD/NASH. Considering the correlation between TRIM7 and immune responses, and that the liver serves as the most critical metabolism-immune organ, our findings uncover a novel role for TRIM7 in metabolic dysregulation, bridging immune regulation and metabolic diseases.

Hepatic steatosis, inflammation and insulin resistance form a vicious circle that collectively drives the progression of NASH. At the molecular level, metabolic stress triggers hyperactivation of two canonical inflammatory pathways (IKKβ-NF-κB and JNK), thereby inducing hepatocellular injury and inflammation. Hyperactivated IKKβ/JNK promotes serine 307 phosphorylation of IRS1, which disrupts IRS1-AKT signaling and induces insulin resistance³⁴. Concurrently, JNK directly regulates lipid metabolism by binding to downstream c-Jun/c-Fos transcription factors and targeting PPARs^11,35. Insulin resistance resulting from impaired IRS1-AKT signaling not only alleviates FOXO1-mediated inhibition of gluconeogenic genes but also selectively preserves its lipogenesis effects, further exacerbating hepatic steatosis^10,36. Consistent with these reports, MAPK signaling pathway and inflammatory responses significantly enriched in our KEGG pathway analysis of liver tissues from HFD-fed TRIM7-HKO mice. Further study revealed that TRIM7 deletion attenuated activation of the IKKβ-NF-κB, JNK/p38 MAPK pathway and their common upstream kinase TAK1 in HFD- or HFFC-challenged mice. In contrast, wildtype TRIM7 overexpression (but not its E3-deficient mutant) robustly hyperactivated the TAK1-NF-κB/JNK/p38 axis. In addition, wildtype TRIM7 overexpression blunted insulin signaling and enhanced gluconeogenesis. Altogether, these findings establish TRIM7 as a central regulator linking inflammatory pathway hyperactivation, insulin resistance, and lipid dysregulation in NASH, with its E3 ligase-dependent modulation of the TAK1-NF-κB/JNK/p38 axis serving as a pivotal mechanistic driver.

In-depth molecular mechanistic investigations of TRIM7 have demonstrated the colocalization and direct interaction between TRIM7 and DUSP10. Notably, DUSP10 predominantly localizes to the nucleus in liver tissue of NCD-fed mice but redistributes to the cytoplasm where it colocalizes with TRIM7 under HFD conditions. We hypothesized that the cytoplasmic translocation of DUSP10 is induced by metabolic stress, a hypothesis further supported by experiments showing that treatment with the nuclear export inhibitor leptomycin B (LMB) reduces DUSP10 cytoplasmic expression in palmitic acid (POA)-treated hepatocytes (Supplementary Fig. 11a). Such subcellular redistribution likely represents a critical regulatory paradigm in the TRIM7-DUSP10 axis, partially explaining why TRIM7-mediated regulation of NAFLD and metabolic signaling pathways occurs in a stress-dependent manner. Additionally, we found that DUSP10 protein expression was dramatically decreased in NAFLD model samples, which was contrary to the expression trend of TRIM7 protein. More detailed in vitro studies showed that TRIM7 catalyzes K63-linked polyubiquitination of DUSP10 at the K425 site and the subsequent proteasomal degradation. Of note, overexpression of wildtype TRIM7 (but not its E3 dead ligase mutant) enhanced both K63- and K48-linked ubiquitination of DUSP10, while TRIM7 knockout reduced these modifications compared to controls (Supplementary Fig. 11b, c). These seemingly contradictory with the results in Fig. 7g may be due to the use of ubiquitin mutant plasmids. We therefore hypothesize that the K63 ubiquitination of DUSP10 catalyzed by TRIM7 may act as a seed to recruit other E3 enzyme, which in turn mediate K48-linked polyubiquitination and proteasomal degradation of DUSP10. The detailed molecular mechanisms underlying this sequential ubiquitination cascade warrant further investigation in future studies.

In this study, the change pattern of DUSP10 protein expression showed an inverse trend with that of TRIM7 protein in NAFLD model samples, which further confirmed the correlation between TRIM7 and DUSP10. Restoration of wildtype DUSP10 expression (but not its phosphatase-inactive mutant) suppressed POA-induced activation of the TAK1-NF-κB/JNK/p38 pathway in TRIM7-overexpressing hepatocytes. In contrast, silencing DUSP10 in Trim7-HKO mice reversed both the TRIM7 deficiency-mediated suppression of this pathway and the protective effect against NAFLD. DUSP10 is known to be a negative regulator of MAPK pathway, with a preferential ability to inactivate the JNK and p38 MAPK subtypes^28,37. In this preliminary study, we identified an interaction between DUSP10 and TAK1 (Supplementary Fig. 11d, e). In vitro phosphatase assays showed that wildtype DUSP10 significantly dephosphorylated TAK1, whereas its phosphatase-inactive mutant exhibited minimal activity (Supplementary Fig. 11f), suggesting that DUSP10 negatively regulates TAK1. However, the detailed regulatory mechanism requires further investigation. While our study establishes a mechanistic framework, several questions remain to be further investigated. First, the upstream regulators driving TRIM7 overexpression in NAFLD require detailed characterization, and the role of TRIM7 in NAFLD has only been studied in male mice here, whether TRIM7 exhibits the same function in female mice needs to be explored in future research. Additionally, preclinical studies using more complex animal models and human primary cells are essential to translate these findings into clinical applications. Third, it is necessary to explore the association between genetic variations in TRIM7 and human diseases, as well as the generalizability of the TRIM7-DUSP10 axis across diverse liver pathologies, such as alcoholic-related liver injuries, NAFLD-related hepatocellular carcinoma, and viral hepatitis.

In conclusion, our study reveals that stress-induced upregulation of TRIM7 promotes the ubiquitination and degradation of DUSP10, leading to hyperactivation of TAK1-NF-κB/JNK/p38 signaling pathways, thus triggering inflammation, insulin resistance, and hepatic steatosis, facilitating the progression of NAFL to NASH (Fig. 9). These findings not only enhance our understanding of NAFLD pathogenesis but also provide a promising therapeutic strategy targeting the TRIM7-DUSP10 axis. Potential approaches include the development of small-molecule inhibitors to disrupt the interaction between TRIM7 and DUSP10 or inhibit the E3 ligase activity of TRIM7, as well as genetic knockdown of TRIM7.

Fig. 9. TRIM7 promotes NAFLD/NASH progression by targeting DUSP10 for degradation.

Schematic diagram illustrating the underlying molecular mechanism by which TRIM7 mediates DUSP10 degradation and promotes the progression of NAFLD to NASH.

Methods

Animal experiments

To mitigate the impact of hormonal fluctuations on metabolic processes, only male animals were used in this study. Six to eight-week-old C57BL/6J male mice and age-matched ob/ob male mice used in this study were purchased from Jinan Pengyue Laboratory Animal Breeding Co., LTD (Jinan, China). All mice were housed in a specific pathogen-free environment with a constant temperature (22–24 °C), humidity (50 ± 5%), and a 12-h light/dark cycle, and had ad libitum access to food and water. After one week of acclimatization to the new environment, all mice were randomly assigned to experimental groups. Mice were fed either a high-fat diet (HFD; protein 20%, carbohydrate 20%, fat 60%; XTM04; Xietong Bio-engineering, Nanjing, China) or a high-fat, high-fructose and high-cholesterol diet (HFFC; protein 20%, fructose 20%, fat 40%, cholesterol 2%; XTM05; Xietong Bio-engineering, Nanjing, China) for 14 or 19 weeks, respectively, to establish mouse models of NAFL or NASH. Mice in the control group were fed a normal chow diet (NCD, protein 21.5%, carbohydrate 67.4%, fat 11.1%; XT101WC; Xietong Bio-engineering, Nanjing, China). The body weights and food intake of each mouse were measured weekly throughout the experiment period.

Generation of hepatocytes-specific Trim7 knockout mice

Trim7-flox/flox mice were generated using CRISPR/Cas9 by Cyagen (Suzhou, China). Briefly, exon 4–6 of Trim7 allele were targeted for conditional knockout, using 4 guide RNAs against introns 3 and 6 (Supplementary Fig. 2A). The targeting vector containing loxp-flanked exon 4–6 and 2 homology arms was co-injected with Cas9 and gRNA into fertilized eggs. The positive founder mice were mated with C57BL/6J mice to obtain Trim7-flox heterozygous mice, with germ-line transmission confirmed by PCR and sequencing. Hepatocyte-specific Trim7-knockout mice (Trim7-HKO, Trim7-flox/flox; Ale-Cre^+/−) and control mice (Trim7-Flox, genotype: Trim7-flox/flox) were generated by crossing Trim7-flox/flox mice with Albumin-Cre mice. All mice were of the C57BL/6J background. Knockout efficiency was verified by locus-specific PCR and western blot. The primers used for genotyping Trim7-HKO mice were synthesized by Beijing Tsingke Biotechnology Co., Ltd. (Nanjing, China), and their sequences are listed in Supplementary Table 1.

Mouse adeno-associated virus serotype 8 (AAV8) injection

The AAV8 delivery system was used to manipulate the overexpression or knockdown of specific genes in mice livers. For hepatocyte-specific Dusp10 knockdown, shRNA targeting Dusp10 was inserted into the vector pAAV8-TBG-GdGreen-WPRE by ObiO Technology (Shanghai, China) and generated the pAAV8-TBG-GdGreen-miR30shRNA (Dusp10) vector (AAV8-shDUSP10). The AAV8 vector was under the TBG promoter to achieve hepatocyte-specific expression, and the corresponding empty vector (AAV8-shNC) was used as control. To specifically overexpress Trim7 in hepatocytes in mice, AAV8 vectors encoding 3×Flag-tagged Trim7 wildtype (AAV8-Trim7) and its E3 dead ligase mutant (C29A, C32A, AAV8-Trim7-DL) were constructed by ObiO Technology. The empty vector (AAV8-EGFP) was used as control. Mice fed with a HFD for 4 weeks or a HFFC for 9 weeks were injected with 150 μL of AAV8 virus (3 × 10¹¹ virus genomes) via the tail vein, then continued HFD or HFFC for 10 weeks. The equal amount of empty vector virus were injected into mice as the control.

Intraperitoneal glucose tolerance test (IPGTT) and insulin tolerance tests (IPITT)

IPGTT tests were performed before 2 weeks of HFD or HFFC-fed ending, and IPITT tests were conducted before one week of fed ending. After fasting for 8 h, mice were weighted and injected with 2 g/kg glucose or 0.75 U/kg insulin intraperitoneally. Subsequently, the concentration of blood glucose of tail venous blood at the indicated time after intraperitoneal injection was measured using commercial glucometer. The date of IPGTT and IPITT were analyzed and calculated the areas under the curve (AUC) using graphpad prism9 software. Homeostasis model assessment of insulin resistance (HOMA-IR) index was calculated based on the concentration of fasting blood glucose and insulin in serum of mice.

Serum insulin and cytokine parameters detection

Blood samples were obtained from the orbital vein of mice after an overnight fast and then collected the supernatant serum through centrifugation. The concentration of serum insulin and cytokines of mice were detected using commercial enzyme linked immunosorbent assay (ELISA) kit according to the manufacturer’s instructions. The insulin (KE10089), TNFα (KE10002), IL-6 (KE10007) and IL-10 (KE10008) ELISA kits were purchased from Proteintech (Wuhan, China).

Biochemical indicators analysis

The contents of serum alanine transaminase (ALT), aspartate aminotransferase (AST) and liver triglyceride (TG), total cholesterol (TC) and non-esterified free fatty acids (NEFA) were measured using commercial kits according to the manufacturers’ protocols. These commercial kits were purchased from Nanjing Jiancheng Bio-engineering Institute and the detailed information are listed in Supplementary Table 2.

Histopathologic staining

Liver tissues were fixed with 4% paraformaldehyde (BL539A, biosharp) overnight. For histomorphology, paraffin-embedded sections were stained with hematoxylin and eosin (H&E) and evaluated by two blinded pathologists. The NAS is the sum of separate scores for steatosis (0 = < 5%; 1 = 5–33%; 2 = 34–66%; 3 = > 66%), hepatocellular ballooning (0 = none; 1 = mild; 2 = moderate/severe), and lobular inflammation (0 = none; 1 = 1–2 foci/200×; 2 = 3–4 foci/200×; 3 = ≥ 5 foci/200× field). NASH was diagnosed based on the presence of steatosis (≥5%), lobular inflammation, and hepatocyte ballooning on liver histology according to the NASH Clinical Research Network (CRN) criteria with NAS ≥ 5 or NAS 3–4 with accompanying fibrosis. To evaluate the lipid deposition in the liver, OCT-embedded frozen sections were stained with Oil Red O (G1015, Servicebio, Wuhan, China) and then re-stained with hematoxylin after rinsed with 60% isopropyl alcohol. To measure the content of glycogen in liver tissues, paraffin-embedded liver tissue sections were stained with Periodic Acid Schiff stain (PAS) (G1008, Servicebio, Wuhan, China). Fibrosis was evaluated by picrosirius red or Masson trichrome staining (G1006, Servicebio). For Trim7 immunohistochemistry, paraffin sections were deparaffinized, rehydrated, and subjected to antigen retrieval (citrate buffer pH 6.0). Endogenous peroxidase was quenched with 3% hydrogen peroxide, and sections were blocked with 3% BSA before overnight incubation with anti-Trim7 primary antibody at 4 °C. After HRP-conjugated secondary antibody incubation, signals were visualized with DAB, and nuclei were counterstained with hematoxylin. All images were captured using a Leica light microscope (Leica; Germany).

Immunofluorescence staining

For immunofluorescence (IF) of liver tissues: after deparaffinization, rehydration, antigen retrieval, peroxidase quenching, and blocking with BSA, the liver tissue sections were incubated with primary antibodies overnight at 4 °C, followed by fluorescein-conjugated secondary antibodies. Nuclei were counterstained with DAPI. For cellular IF on coverslip: cells were fixed with 4% paraformaldehyde, permeabilized, blocked, and incubated with primary antibodies overnight at 4 °C. After incubation with fluorescein-conjugated secondary antibodies, nuclei were stained with DAPI. Images were acquired using a confocal laser scanning microscope (TCS SP8; Leica; Germany).

Human liver tissue samples

Paraffin-embedded human liver tissue samples (NAFLD and non-steatosis controls) were collected from Hubei Cancer Hospital (sex was not considered as a selection criterion). Steatotic samples were acquired from NAFLD/NASH patients who underwent liver biopsy or surgical resection; patients with hepatic steatosis secondary to excessive alcohol consumption (>140 g for men or >70 g for women per week) or viral hepatitis were excluded from the study. Non-steatotic controls were collected from histologically normal liver regions of NAFLD-free donors who received surgical resection because of hepatic benign tumor. A total of 8 non-steatosis samples and 8 NAFLD/NASH phenotypic liver samples were included in the current study. Detailed clinicopathological information (sex, BMI and liver function indicators) is provided in Supplementary Table 3.

Mice primary hepatocytes isolation

Primary hepatocytes were isolated from 6–8-weeks-old C57BL/6J male mice as previously described⁵. Briefly, anesthetized mice were subjected to in situ liver perfusion via the inferior vena cava using perfusion buffer, followed by liver digestion with perfusion buffer containing 0.05% collagenase IV (V900893, Sigma) using a peristaltic pump. The excised liver was washed with DMEM (KGL1206, KeyGen) and minced. The cell suspension was filtered and then centrifuged at 50 × g for 1 min. The harvested hepatocytes were resuspended in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin and then seeded on rat tail collagen-precoated dishes. After 3 h of incubation for cell attachment, the medium was replaced with fresh complete medium. These cultured hepatocytes were used for subsequent experiments.

Cell culture and treatments

HEK293T and HepG2 cell lines were obtained from American Type Culture Collection (ATCC, USA), and human normal hepatocyte L02 were previously kindly provided by Professor Jianguo Wu laboratory (Wuhan University). All cells were grown in DMEM medium (KGL1206, KeyGen) supplemented with 10% FBS and 1% penicillin/streptomycin (Gbico BRL) at 37 °C, 5% CO₂. To establish fatty liver cell models, primary hepatocytes or hepatocyte cell lines were treated with a free fatty acid (FFA) mixture containing palmitic acid (PA, P0500, Sigma-Aldrich) and oleic acid (OA, O1383, Sigma-Aldrich) for the indicated time periods. 50 mM PA/OA stock solutions were prepared in 50% ethanol and filtered through a 0.22 μm sterile membrane. A 2 mM FFA working solution was prepared by adding PA/OA (1:2) and 1% fatty acid-free BSA to complete DMEM. Control cells were treated with complete DMEM supplemented with 1% BSA and 0.5% ethanol. In the protein half-life assay, transfected HEK293T cells were treated with 50 μg/mL cycloheximide (CHX, #239764, Sigma-Aldrich) for indicated time. In the protein stability assays, transfected HEK293T cells were pretreated with 15 μM MG132 (M7449, Sigma-Aldrich), 100 μM chloroquine (CHQ, C6628, Sigma-Aldrich) or 20 μM NH₄Cl for 1 h prior to CHX treated.

Plasmid construction and transfection

Human full-length cDNA sequences of TRIM7 and DUSP10 were generated by PCR-based amplification of human cDNA library, and then inserted into the 3 × Flag-tagged phage vector or HA-tagged pCAGGS vector, respectively. Truncated TRIM7 and DUSP10 fragments as indicated in Fig. 7 were also generated by PCR methods and cloned into the corresponding vectors. The fragments of TRIM7 and DUSP10 site-specific mutant were constructed using overlap PCR, and then cloned into the corresponding vectors. All recombinant plasmids constructed in this study were verified by DNA sequencing in Beijing Tsingke Biotechnology Co., Ltd. (Nanjing, China). The indicated plasmids were transfected into 293T cells or hepatocyte cell lines with transfection reagent according to the manufacturer’s instructions.

RNA extraction and quantitative PCR

Total RNA was extracted from tissues (30 mg) or cells using Trizol reagent (#191012, Invitrogen) according to the manufacturer’s protocol. The quality and concentration of RNA were evaluated using NanoDrop 2000. Extracted RNA with A260/A280 ratio 1.9–2.1 were reverse transcribed into cDNA using commercial kit with gDNA eraser (R212V21.1, Vazyme) according to the instructions. RT-qPCR was conducted using the SYBR Green Mix (DBI-2044, Germany) on an ABI 7500 StepOne Real-Time PCR Detection System. The specific primers for detected genes were synthesized by Beijing Tsingke Biotechnology Co., Ltd. (Nanjing, China), and their sequences are provided in Supplementary Table S1. The relative mRNA expression of target genes was calculated using the 2^−ΔΔCT method, and β-actin was used as an internal control.

Protein extraction and western blot

Mice tissue and cells were homogenized and lysed with RIPA lysis buffer containing phosphatase and protease inhibitors cocktail. After incubation on ice for 30 min, the cell lysates were centrifuged at 12,000 × g and 4 °C for 15 min. The collected protein supernatants were then denatured with SDS-loading buffer containing DTT at 95 °C for 10 min. Protein extraction from paraffin-embedded human liver sections was performed using a commercial kit (BB-3164, BestBio, Shanghai, China) according to the manufacturer’s protocol. Briefly, four 50 μm-thick sections were deparaffinized in xylene, rehydrated in graded ethanol, rinsed with ddH₂O, and air-dried. Then, protein were incubated with extraction buffer containing DTT under the following conditions: 5 min on ice, 20 min at 100 °C, 2 h at 80 °C, and 1 min at 4 °C. Protein supernatants were collected by centrifugation at 4 °C and denatured with SDS-loading buffer containing DTT at 95 °C for 10 min. All protein samples were stored at −80 °C for further experiments. To examine the expression of proteins, equal amounts of protein samples were subjected to SDS-PAGE and transferred to PVDF membranes (Millipore, USA). After blocking with 5% skim milk, the membrane were incubated with primary antibodies (diluted with primary antibody diluent, VP6022, VICMED) overnight at 4 °C, followed by incubation with HRP-conjugated secondary antibodies. The proteins were detected with ECL and visualized in a chemiluminescence imaging system (GE Amersham Imager 600, USA).

Coimmunoprecipitation

Protein-protein interaction was detected by co-immunoprecipitation assay as previously described^5,35. Briefly, the corresponding protein expression plasmids were co-transfected into HEK293T cells using transfection reagent for exogenous protein immunoprecipitation. After transfected for 30 h, 293T cells were collected and lysed with lysis buffer containing protease inhibitor for 30 min on ice, followed by centrifugation at 12,000 × g and 4 °C for 15 min. The collected cell lysates were incubated with magnetic beads conjugated to anti-Flag (B26102, Bimake) or anti-HA (B26202, Bimake) at 4 °C overnight. For endogenous protein immunoprecipitation, the samples of cell or tissue were lysed as above described. The protein lysates were incubated with protein A/G sepharose (17-0963-03, 17-0618-01; GE Healthcare, USA) and the indicated primary antibodies at 4 °C overnight. The corresponding IgG antibody was used as negative control. The immunocomplexes were washed with lysis buffer for six times and then eluted with 2 × SDS-loading buffer containing DTT at 95 °C. The eluted protein were subjected to western blot detection as aforementioned.

Mass spectrometry analysis

Hepatocyte line (L02) stably overexpressing Flag-tagged TRIM7 and its corresponding control cells were subjected to immunoprecipitation (co-IP) as described above. The supernatants were incubated with pre-washed anti-Flag M2 magnetic beads (M8823, Sigma) in lysis buffer at 4 °C overnight with rotation. Immunocomplexes were washed with lysis buffer and eluted with 3 × Flag peptide (F4799, Sigma). The eluted protein samples were separated by SDS-PAGE, and the target gel bands were excised, trypsin-digested, and peptides separated via an Ultimate 3000 UHPLC system (Thermo Fisher Scientific). Peptides were analyzed using a Q Exactive HF X tandem mass spectrometer (Thermo Fisher Scientific) in data-dependent acquisition (DDA) mode by BGI Co., Ltd. (Shenzhen, China) (no biological replicates). Raw data (.mgf files) were converted and searched against a protein database using Mascot (v2.3.02, Matrix Science). Results underwent quality control (peptide spectrum match (PSM)-level FDR ≤ 0.01) to retain high-confidence peptides. Candidates were selected if detected exclusively in the TRIM7-overexpressing group (not controls). Interacting proteins were further validated by gene ontology (GO) enrichment analysis and literature mining.

RNA-seq and data processing

For profiling the gene expression differences, liver total RNA was used for cDNA library construction and sequenced on Illumina NovaSeq X Plus (Tsingke Biotech, Beijing). RNA-seq data were preprocessed with fastp (v0.20.1) to generate CleanData (removing adaptors/low-quality reads), mapped to Ensembl mouse genome (mm10/GRCm38) using HISAT2 (v2.2.1). FPKM values were calculated via StringTie (v2.0.4). DESeq2 (v1.26.0) identified differential expression; DEGs were defined as |log₂(fold change)| ≥ 1.0 and P < 0.05 (two-tailed Wald test).

KEGG pathway enrichment analysis

KEGG pathway enrichment analysis was performed using Fisher’s exact test with R script. The KEGG pathway annotations were downloaded from the KEGG database. Pathways with a P value < 0.05 were considered as significantly enriched pathways.

Ubiquitination assay

HEK293T cells that cotransfected with HA-DUSP10, myc-Ub and with or without Flag-TRIM7 plasmids as well as liver tissues, hepatocyte cell line were used to detect the ubiquitination level of exogenous and endogenous DUSP10 proteins. Liver tissues or cells were homogenized and lysed with IP buffer T (20 mM Tris-HCl, pH 7.4; 1 mM EDTA; 150 mM NaCl and 1% Triton X-100) containing protease inhibitors, and then denatured by adding 1% SDS as well as heating at 95 °C for 10 min. After dilution 10 fold with IP buffer T, the lysates were sonicated, centrifuged. The collected supernatants were immunoprecipitated with magnetic beads coupled to anti-HA (B26202, Bimake) or protein A/G sepharose with indicated target protein antibody at 4 °C overnight. The immunocomplexes were rinsed six times with IP buffer T containing 0.5 M NaCl, eluted in 2×SDS-loading buffer with DTT at 95 °C, and analyzed by western blot as described.

In ubiquitination assays: “O” denotes ubiquitin with all lysines (except its own) mutated; “R” indicates mutation of the specified ubiquitin lysine to arginine.

Statistics and reproducibility

All relevant experiments data are presented as mean ± SD. Except for some animal studies, each experiment was performed independently at least three times with similar results. Statistical analyses were conducted using GraphPad Prism 9. Two-tailed Student’s t test was used for two-group comparisons; one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was used for multiple-group comparisons. A P value less than 0.05 was considered to be statistically significant.

Ethical statement

Animal experiments were approved by the Ethical Review Committee on Laboratory Animal Research of Jiangsu Normal University (JSNU-IACUC-2024016), and conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals and China’s Regulations on the Administration of Experimental Animals. All procedures involving human samples in this study followed the principles outlined in the Declaration of Helsinki and were approved by the Ethics Committees of Hubei Cancer Hospital (LLHBCH2024YN-089). All participants volunteered to participate in this study on an unpaid basis and signed a written informed consent form prior to participating in this study. Appropriate security measures were implemented to protect the privacy and confidentiality of the subjects.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

F.-J.Y., R.-P.L. and Q.W. designed and supervised this study, analyzed data and wrote the manuscript. F.-J.Y., H.D., N.Z., M.-X.H. performed experiments and data analysis. Q.W. collected the clinical liver tissue samples. Q.W., R.-S.Y. and F.-J.Y. performed pathological analysis and provided useful advice regarding the manuscript. F.-J.Y., H.D., N.Z., J.-W.L. and Y.-J.Y. constructed the Trim7-HKO mice and established mouse NAFLD/NASH models. R.-P.L. and Y.-J.W. helped design the project and edit the manuscript. All authors reviewed and approved the final manuscript.

Peer review

Peer review information

Nature Communications thanks Minxuan Xu and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The data that support the findings of this study are available in the main text, the supplementary information, and Source Data file. The RNA-Seq data generated have been deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database under the accession code PRJNA1268184. The mass spectrometry proteomics data have been deposited in the ProteomeXchange Consortium via the iProX partner repository with the dataset identifier PXD069230.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-65415-6.