Lactate-associated H3K18la upregulates RNF166 to promote immune evasion in hepatocellular carcinoma by remodeling c-Jun ubiquitination Authors: Bingyi Ren, Yichen Yang, Kai Sheng, Qiuyu Mu, Zhiyuan Bai, Min Tian, Yang Ren, Yi Lv, Kang Liu Published: 2026 ABSTRACT Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality, and PD-L1-mediated immune evasion is a major mechanism of tumor immune suppression and an important target of immune checkpoint blockade. Here, we identified RNF166 as an E3 ubiquitin ligase that promotes immune evasion in HCC. RNF166 expression was upregulated in HCC tissues and correlated with PD-L1 levels, whereas RNF166 did not directly affect the proliferation or apoptosis of HCC cells. In immunocompetent mouse models, RNF166 overexpression accelerated tumor growth and suppressed CD8 + T-cell infiltration and cytotoxic function, while these effects were dependent on PD-L1 signaling and were attenuated in immunodeficient settings. Mechanistically, RNF166 physically interacted with c-Jun and promoted K63-linked ubiquitination of c-Jun at K50/K56, which was associated with reduced K48-linked ubiquitination and enhanced c-Jun stability. Stabilized c-Jun bound to the PD-L1 promoter and promoted PD-L1 transcription, thereby linking RNF166-mediated c-Jun stabilization to tumor immune suppression. Upstream, lactate accumulation increased H3K18la enrichment at the RNF166 promoter and contributed to RNF166 transcriptional upregulation. Therapeutically, pharmacological inhibition of c-Jun with T-5224 enhanced the antitumor effect of anti-PD-L1 blockade. Collectively, our findings delineate a lactate-H3K18la-RNF166-c-Jun-PD-L1 axis that promotes immune evasion in HCC and suggest that targeting RNF166 or the RNF166-c-Jun axis may improve PD-1/PD-L1-based immunotherapy.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s10495-026-02435-7. Introduction Hepatocellular carcinoma (HCC) is a highly malignant tumor with poor prognosis. The current therapeutic strategies for HCC include surgical resection, radiofrequency ablation, transarterial chemoembolization, and systemic therapies [1–3]. With the advances in cancer immunology, immunotherapy has emerged as an important treatment modality for advanced HCC [4, 5]. Immune checkpoint inhibitors (ICIs) restore antitumor immunity by blocking the inhibitory immune signaling pathways, thereby reactivating cytotoxic immune cells and suppressing tumor growth. Among them, PD-1/PD-L1 inhibitors disrupt the interaction between PD-1 and PD-L1, thereby alleviating the immunosuppressive tumor microenvironment [6]. Notably, the IMbrave150 trial demonstrated that the combination of atezolizumab and bevacizumab significantly improved overall survival in patients with unresectable HCC [7], establishing immunotherapy as a frontline treatment option. PD-L1 expression is regulated at multiple levels, including genomic, transcriptional, post-transcriptional, and post-translational mechanisms [8, 9]. Elucidating the regulatory network governing PD-L1 expression may provide novel therapeutic targets and combinatorial strategies to enhance immunotherapy responsiveness, particularly in patients who exhibit primary or acquired resistance to immune checkpoint blockade. RNF166 belongs to the RING-UIM family of E3 ubiquitin ligases and contains the characteristic C3H4-RING, C2HC, C2H2, and UIM domains [10]. Although other members of the RING-UIM family, such as RNF114, RNF125, and RNF138, have been extensively studied in cancer biology, the role of RNF166 in tumor progression and immunity remains poorly characterized. Previous studies have shown that RNF166 mediates K29- and K33-linked ubiquitination to regulate antibacterial autophagy [11]. RNF166 has also been reported to promote XIAP degradation and neuronal apoptosis [12] and to stabilize Ku80 through K63-linked ubiquitination [13]. In cancer research, RNF166 has been shown to decrease angiomotin stability, leading to YAP activation and promotion of colorectal cancer progression [14]. However, the biological functions of RNF166 in HCC and tumor immune regulation remain largely unknown. c-Jun is a classic transcription factor that functions as a component of the AP-1 complex and regulates the transcription of numerous downstream target genes [15]. Accumulating evidence indicates that c-Jun can directly transactivate PD-L1 expression [16, 17], thereby contributing to tumor immune evasion. Nevertheless, the upstream regulatory mechanisms controlling c-Jun protein stability and activity, particularly the role of E3 ubiquitin ligases in modulating c-Jun in the context of tumor immunity, remain insufficiently understood. Protein lactylation is a recently identified post-translational modification in which lactyl groups are added to lysine residues, thereby influencing protein function [18]. Lactylation has been implicated in transcriptional regulation and signal transduction [19], particularly in histones, where it alters chromatin accessibility and gene transcription. Histone H3 lactylation represents one of the most extensively characterized forms of this modification. Emerging evidence suggests that lactylation participates in tumor immune regulation [20, 21] and may influence the proteins involved in PD-L1 expression and stability [22, 23]. However, the precise role of lactylation in tumor immunity, particularly in HCC, remains largely unexplored. Results RNF166 is upregulated in HCC but does not affect intrinsic tumor growth We first analyzed RNF166 expression across multiple cancer types using the TIMER2.0 Gene_DE module. RNF166 expression was significantly higher in HCC tissues than in normal tissues (Fig. 1a). To further validate this observation, we analyzed RNF166 expression in the ICGC LIRI-JP cohort. Paired analysis revealed that RNF166 mRNA expression was markedly higher in HCC tissues than in matched adjacent non-tumor tissues (Fig. 1b). Consistently, comparison between normal liver tissues and HCC tissues in the ICGC LIRI-JP cohort demonstrated significantly increased RNF166 expression in HCC tissues (Fig. 1c). Data from the GSE190174 dataset further confirmed the upregulation of RNF166 in HCC samples (Fig. 1d). In addition, immunohistochemical analysis of a clinical tissue microarray showed that RNF166 expression was markedly elevated in HCC tissues compared with adjacent non-tumor tissues (Fig. 1e). Subsequently, RNF166-overexpressing and RNF166-knockdown HCC cell lines were established, and the efficiency of overexpression and knockdown was confirmed using Western blotting and qPCR analyses (Fig. 1f, g). To evaluate the effect of RNF166 on tumor cell-intrinsic biological behavior, colony formation assays were performed. Neither overexpression nor knockdown of RNF166 significantly altered colony-forming ability (Fig. 1h). Consistently, CCK-8 assays (Fig. 1i) and EdU incorporation assays (Fig. 1j–l) showed no significant effects of RNF166 manipulation on HCC cell proliferation. Western blotting analysis of proliferation- and apoptosis-related markers further revealed no marked changes after RNF166 overexpression or knockdown (Fig. 1m). To further validate these findings in vivo, RNF166-overexpressing Hepa1-6 cells were generated (Supplementary Fig. S1a, b) and subjected to subcutaneous tumorigenesis assays in immunodeficient BALB/c nude mice. No significant differences in tumor volume, tumor weight, or tumor growth curves were observed between the RNF166-overexpressing and vector-control groups (Supplementary Fig. S1c–e). Immunohistochemical staining for Ki67 in tumor tissues also showed no significant difference between the two groups (Supplementary Fig. S1f). Moreover, in NOD-SCID mice, RNF166 overexpression did not significantly alter tumor volume, tumor weight, or tumor growth kinetics compared with vector control (Supplementary Fig. S1g–i). These results indicate that RNF166 is upregulated in HCC but does not directly promote tumor-cell-intrinsic proliferation or tumor growth in immunodeficient settings. Fig. 1RNF166 is upregulated in HCC but does not directly promote HCC cell proliferation. a Pan-cancer analysis of RNF166 expression in tumor and normal tissues using the TIMER2.0 Gene_DE module. Differences between tumor and normal tissues were assessed using a two-sided Wilcoxon rank-sum test. Sample sizes were based on the available TCGA samples for each cancer type; *P < 0.05, **P < 0.01, and ***P < 0.001. b Paired analysis of RNF166 mRNA expression in HCC tissues and matched adjacent non-tumor tissues from the ICGC LIRI-JP cohort (n = 179 matched pairs; two-tailed paired Student’s t-test, P = 0.02). c RNF166 mRNA expression in normal liver tissues (n = 179) and HCC tissues (n = 391) from the ICGC LIRI-JP cohort (two-tailed unpaired Student’s t-test, P = 0.036). d RNF166 mRNA expression in normal liver tissues (n = 5) and HCC tissues (n = 5) from the GSE190174 dataset (two-tailed unpaired Student’s t-test, P = 0.02). e Representative immunohistochemical staining of RNF166 in HCC and adjacent non-tumor tissues, with quantification of RNF166 IHC scores (n = 76 per group; two-tailed paired Student’s t-test, P < 0.0001). (f, g) Western blotting and qPCR analyses confirming RNF166 knockdown in Huh7 and HCCLM3 cells and RNF166 overexpression in Hep3B cells. qPCR data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P < 0.0001). HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (shNC vs. shRNF166-01, P < 0.0001; shNC vs. shRNF166-02, P = 0.0002). h Colony formation assays evaluating the effect of RNF166 knockdown or overexpression on HCC cell proliferation. Colony areas were normalized to the mean value of the corresponding control group, which was set to 100%. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.6131). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-01, P = 0.6098; shNC vs. shRNF166-02, P = 0.9115; HCCLM3: shNC vs. shRNF166-01, P = 0.7867; shNC vs. shRNF166-02, P = 0.5541). i CCK-8 assays evaluating HCC cell viability after RNF166 manipulation. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-way ANOVA followed by Šídák’s multiple comparisons test. No significant differences were observed between the Hep3B Vec and RNF166 groups at any time point; the adjusted P value at day 4 was 0.8598. For knockdown cells, day-4 adjusted P values were as follows: Huh7 shNC vs. shRNF166-01, P = 0.8496; Huh7 shNC vs. shRNF166-02, P = 0.9467; HCCLM3 shNC vs. shRNF166-01, P = 0.9406; HCCLM3 shNC vs. shRNF166-02, P = 0.9998. j–l EdU incorporation assays evaluating HCC cell proliferation after RNF166 knockdown or overexpression. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.7027). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-01, P = 0.9921; shNC vs. shRNF166-02, P = 0.9016; HCCLM3: shNC vs. shRNF166-01, P = 0.9980; shNC vs. shRNF166-02, P = 0.8511). m Western blotting analysis of proliferation- and apoptosis-related markers after RNF166 knockdown or overexpression. Data are presented as mean ± SD RNF166 is upregulated in HCC but does not directly promote HCC cell proliferation. a Pan-cancer analysis of RNF166 expression in tumor and normal tissues using the TIMER2.0 Gene_DE module. Differences between tumor and normal tissues were assessed using a two-sided Wilcoxon rank-sum test. Sample sizes were based on the available TCGA samples for each cancer type; *P < 0.05, **P < 0.01, and ***P < 0.001. b Paired analysis of RNF166 mRNA expression in HCC tissues and matched adjacent non-tumor tissues from the ICGC LIRI-JP cohort (n = 179 matched pairs; two-tailed paired Student’s t-test, P = 0.02). c RNF166 mRNA expression in normal liver tissues (n = 179) and HCC tissues (n = 391) from the ICGC LIRI-JP cohort (two-tailed unpaired Student’s t-test, P = 0.036). d RNF166 mRNA expression in normal liver tissues (n = 5) and HCC tissues (n = 5) from the GSE190174 dataset (two-tailed unpaired Student’s t-test, P = 0.02). e Representative immunohistochemical staining of RNF166 in HCC and adjacent non-tumor tissues, with quantification of RNF166 IHC scores (n = 76 per group; two-tailed paired Student’s t-test, P < 0.0001). (f, g) Western blotting and qPCR analyses confirming RNF166 knockdown in Huh7 and HCCLM3 cells and RNF166 overexpression in Hep3B cells. qPCR data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P < 0.0001). HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (shNC vs. shRNF166-01, P < 0.0001; shNC vs. shRNF166-02, P = 0.0002). h Colony formation assays evaluating the effect of RNF166 knockdown or overexpression on HCC cell proliferation. Colony areas were normalized to the mean value of the corresponding control group, which was set to 100%. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.6131). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-01, P = 0.6098; shNC vs. shRNF166-02, P = 0.9115; HCCLM3: shNC vs. shRNF166-01, P = 0.7867; shNC vs. shRNF166-02, P = 0.5541). i CCK-8 assays evaluating HCC cell viability after RNF166 manipulation. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-way ANOVA followed by Šídák’s multiple comparisons test. No significant differences were observed between the Hep3B Vec and RNF166 groups at any time point; the adjusted P value at day 4 was 0.8598. For knockdown cells, day-4 adjusted P values were as follows: Huh7 shNC vs. shRNF166-01, P = 0.8496; Huh7 shNC vs. shRNF166-02, P = 0.9467; HCCLM3 shNC vs. shRNF166-01, P = 0.9406; HCCLM3 shNC vs. shRNF166-02, P = 0.9998. j–l EdU incorporation assays evaluating HCC cell proliferation after RNF166 knockdown or overexpression. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.7027). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-01, P = 0.9921; shNC vs. shRNF166-02, P = 0.9016; HCCLM3: shNC vs. shRNF166-01, P = 0.9980; shNC vs. shRNF166-02, P = 0.8511). m Western blotting analysis of proliferation- and apoptosis-related markers after RNF166 knockdown or overexpression. Data are presented as mean ± SD RNF166 upregulates PD-L1 and suppresses antitumor T-cell responses GO enrichment analysis using TCGA data showed that RNF166-associated genes were enriched in immune-related pathways (Fig. 2a). Analysis of the GSE202069 cohort showed that RNF166 expression was higher in HCC patients who responded to anti-PD-1 therapy than in non-responders (Fig. 2b). We then assessed the correlation between RNF166 and CD274 expression in LIHC tumor samples using the GEPIA2 database, with only tumor samples included in the analysis (Fig. 2c). Because exploratory survival analyses did not show a significant association between RNF166 expression and overall survival in HCC cohorts, we focused on the potential relationship between RNF166 and immune regulation rather than presenting RNF166 as a prognostic biomarker (Supplementary Fig. S2). In vitro experiments using human HCC cell lines showed that RNF166 overexpression or knockdown altered PD-L1 levels, as detected by Western blotting (Fig. 2d), qPCR (Fig. 2e), and flow cytometry (Fig. 2f). RNF166 expression was positively associated with PD-L1 protein expression, PD-L1 mRNA levels, and cell-surface PD-L1 expression. Similarly, in the mouse liver cancer cell line Hepa1-6, RNF166 overexpression significantly increased PD-L1 expression (Fig. 2g, h). To evaluate whether RNF166 broadly regulated immune-related genes, we performed a PCR array in RNF166-overexpressing Hep3B cells. Among the altered immune-related genes, CD274/PD-L1 was prominently upregulated and was reduced by c-Jun knockdown, whereas several other candidates, including VEGFA, IL10, IDO1, CSF1, and TGFB1, were not consistently dependent on c-Jun under the same conditions (Supplementary Fig. S3). These data support PD-L1 as an important RNF166-regulated immune checkpoint target. In the subcutaneous C57BL/6 mouse model, tumors formed by RNF166-overexpressing Hepa1-6 cells grew significantly faster than vector-control tumors (Fig. 2i–k). Flow cytometry analysis of the tumor microenvironment revealed that RNF166 overexpression was associated with reduced CD8 + T-cell infiltration and impaired cytotoxic function, as indicated by decreased GZMB, TNF-α, and IFN-γ expression (Fig. 2l). Immunofluorescence analysis yielded consistent results, showing reduced CD8 and GZMB signals in RNF166-overexpressing tumors (Fig. 2m). In an orthotopic C57BL/6 liver tumor model, RNF166 overexpression similarly promoted tumor growth, increased PD-L1 expression, and decreased CD8α and GZMB levels within tumors (Supplementary Fig. S4), consistent with the subcutaneous model. Because the tumor immune microenvironment contains multiple immune-cell subsets, we further examined additional immune markers in RNF166-overexpressing tumors. Immunofluorescence analysis of NK1.1, FOXP3, and F4/80 did not show significant differences between RNF166-overexpressing and vector-control tumors (Supplementary Fig. S5). Together, these data indicate that RNF166 upregulates PD-L1 and is associated with impaired antitumor CD8 + T-cell responses in immunocompetent HCC models. Fig. 2RNF166 enhances PD-L1 expression in HCC cells and suppresses antitumor T-cell responses. a Gene ontology enrichment analysis of RNF166-associated genes using TCGA data. b RNF166 mRNA expression in anti-PD-1-treated HCC patients from the GSE202069 cohort stratified according to treatment response. Each dot represents one patient, and horizontal lines indicate the mean (non-responders, n = 9; responders, n = 8; two-tailed unpaired Student’s t-test, P = 0.0130). c Correlation analysis between RNF166 and CD274 expression in LIHC tumor samples using the GEPIA2 database; the correlation coefficient and P value are shown in the panel. d Western blotting analysis of PD-L1 expression after RNF166 knockdown in Huh7 and HCCLM3 cells and RNF166 overexpression in Hep3B cells. Representative Western blots are shown from three independent experiments. e qPCR analysis of PD-L1 mRNA expression after RNF166 manipulation. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.0006). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-1, P = 0.0002; shNC vs. shRNF166-2, P = 0.0002; HCCLM3: shNC vs. shRNF166-1, P < 0.0001; shNC vs. shRNF166-2, P < 0.0001). f Flow cytometric analysis of cell-surface PD-L1 expression after RNF166 knockdown or overexpression. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.0003). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-1, P < 0.0001; shNC vs. shRNF166-2, P < 0.0001; HCCLM3: shNC vs. shRNF166-1, P < 0.0001; shNC vs. shRNF166-2, P < 0.0001). g, h Western blotting and qPCR analyses of PD-L1 expression in Hepa1-6 cells after RNF166 overexpression. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using a two-tailed unpaired Student’s t-test (P < 0.0001). Representative Western blots are shown from three independent experiments. i Representative images of subcutaneous tumors formed by Hepa1-6 cells expressing vector control or RNF166 in immunocompetent C57BL/6 mice. j Tumor volume and tumor weight measurements in the subcutaneous C57BL/6 model. Each data point represents one mouse (n = 5 mice per group; two-tailed unpaired Student’s t-test; tumor volume, P = 0.0081; tumor weight, P = 0.0089). k Tumor growth curves in the subcutaneous C57BL/6 model (n = 5 mice per group; two-way ANOVA followed by Šídák’s multiple comparisons test; adjusted P < 0.0001 at day 18). l Flow cytometric analysis of CD8⁺ T-cell infiltration and cytotoxic markers in tumor tissues. Each data point represents one mouse (n = 5 mice per group; two-tailed unpaired Student’s t-test; CD8⁺ T cells among CD45⁺ TILs, P = 0.0005; GZMB⁺ cells among CD8⁺ T cells, P < 0.0001; IFN-γ⁺ cells among CD8⁺ T cells, P = 0.0184; TNF-α⁺ cells among CD8⁺ T cells, P = 0.0177). m Immunofluorescence analysis of CD8 and GZMB expression in tumor tissues, with fluorescence intensity normalized to the Vec group (n = 5 mice per group; two-tailed unpaired Student’s t-test; CD8, P < 0.0001; GZMB, P = 0.0005). Quantitative data are presented as mean ± SD RNF166 enhances PD-L1 expression in HCC cells and suppresses antitumor T-cell responses. a Gene ontology enrichment analysis of RNF166-associated genes using TCGA data. b RNF166 mRNA expression in anti-PD-1-treated HCC patients from the GSE202069 cohort stratified according to treatment response. Each dot represents one patient, and horizontal lines indicate the mean (non-responders, n = 9; responders, n = 8; two-tailed unpaired Student’s t-test, P = 0.0130). c Correlation analysis between RNF166 and CD274 expression in LIHC tumor samples using the GEPIA2 database; the correlation coefficient and P value are shown in the panel. d Western blotting analysis of PD-L1 expression after RNF166 knockdown in Huh7 and HCCLM3 cells and RNF166 overexpression in Hep3B cells. Representative Western blots are shown from three independent experiments. e qPCR analysis of PD-L1 mRNA expression after RNF166 manipulation. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.0006). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-1, P = 0.0002; shNC vs. shRNF166-2, P = 0.0002; HCCLM3: shNC vs. shRNF166-1, P < 0.0001; shNC vs. shRNF166-2, P < 0.0001). f Flow cytometric analysis of cell-surface PD-L1 expression after RNF166 knockdown or overexpression. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.0003). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-1, P < 0.0001; shNC vs. shRNF166-2, P < 0.0001; HCCLM3: shNC vs. shRNF166-1, P < 0.0001; shNC vs. shRNF166-2, P < 0.0001). g, h Western blotting and qPCR analyses of PD-L1 expression in Hepa1-6 cells after RNF166 overexpression. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using a two-tailed unpaired Student’s t-test (P < 0.0001). Representative Western blots are shown from three independent experiments. i Representative images of subcutaneous tumors formed by Hepa1-6 cells expressing vector control or RNF166 in immunocompetent C57BL/6 mice. j Tumor volume and tumor weight measurements in the subcutaneous C57BL/6 model. Each data point represents one mouse (n = 5 mice per group; two-tailed unpaired Student’s t-test; tumor volume, P = 0.0081; tumor weight, P = 0.0089). k Tumor growth curves in the subcutaneous C57BL/6 model (n = 5 mice per group; two-way ANOVA followed by Šídák’s multiple comparisons test; adjusted P < 0.0001 at day 18). l Flow cytometric analysis of CD8⁺ T-cell infiltration and cytotoxic markers in tumor tissues. Each data point represents one mouse (n = 5 mice per group; two-tailed unpaired Student’s t-test; CD8⁺ T cells among CD45⁺ TILs, P = 0.0005; GZMB⁺ cells among CD8⁺ T cells, P < 0.0001; IFN-γ⁺ cells among CD8⁺ T cells, P = 0.0184; TNF-α⁺ cells among CD8⁺ T cells, P = 0.0177). m Immunofluorescence analysis of CD8 and GZMB expression in tumor tissues, with fluorescence intensity normalized to the Vec group (n = 5 mice per group; two-tailed unpaired Student’s t-test; CD8, P < 0.0001; GZMB, P = 0.0005). Quantitative data are presented as mean ± SD PD-L1 knockdown attenuates RNF166-mediated immunosuppression We knocked down PD-L1 in RNF166-overexpressing Hepa1-6 cells (Fig. 3a) and established subcutaneous tumor models in immunocompetent C57BL/6 mice. PD-L1 knockdown markedly slowed tumor growth (Fig. 3b), with corresponding reductions in tumor volume and tumor weight (Fig. 3c, d). Flow cytometry analysis of the tumor microenvironment revealed that PD-L1 knockdown restored CD8 + T-cell infiltration and enhanced cytotoxic function, as indicated by increased GZMB, TNF-α, and IFN-γ expression (Fig. 3e). Immunofluorescence analysis confirmed these findings, showing increased CD8 and GZMB signals in PD-L1-knockdown tumors (Fig. 3f). Similar results were observed in the orthotopic liver tumor model. PD-L1 knockdown attenuated RNF166-induced tumor growth, as shown by reduced liver tumor burden and liver-to-body weight ratio (Supplementary Fig. S6a, b). Immunohistochemical analysis showed that PD-L1 knockdown decreased PD-L1 expression while restoring CD8α and GZMB levels in orthotopic tumor tissues (Supplementary Fig. S6c), consistent with the subcutaneous model. Together, these data indicate that PD-L1 contributes to RNF166-mediated tumor growth and immune suppression by limiting CD8 + T-cell infiltration and cytotoxic function in immunocompetent HCC models. Fig. 3PD-L1 knockdown attenuates RNF166-induced T-cell suppression. a Western blotting analysis of RNF166 and PD-L1 expression in RNF166-overexpressing Hepa1-6 cells after PD-L1 knockdown. Representative Western blots are shown from three independent experiments. b Representative images of subcutaneous tumors formed by RNF166-overexpressing Hepa1-6 cells expressing shNC or shPD-L1 in immunocompetent C57BL/6 mice. c Tumor volume and tumor weight measurements in the RNF166-shNC and RNF166-shPD-L1 groups. Each data point represents one mouse (n = 5 mice per group; two-tailed unpaired Student’s t-test; tumor volume, P = 0.0022; tumor weight, P = 0.0009). d Tumor growth curves in the RNF166-shNC and RNF166-shPD-L1 groups (n = 5 mice per group; two-way ANOVA followed by Šídák’s multiple comparisons test; adjusted P < 0.0001 at day 18). e Flow cytometric analysis of CD8⁺ T-cell infiltration and cytotoxic markers in tumor tissues. Each data point represents one mouse (n = 5 mice per group; two-tailed unpaired Student’s t-test; CD8⁺ T cells among CD45⁺ tumor-infiltrating lymphocytes, P < 0.0001; GZMB⁺ cells among CD8⁺ T cells, P = 0.0066; IFN-γ⁺ cells among CD8⁺ T cells, P = 0.0170; TNF-α⁺ cells among CD8⁺ T cells, P = 0.0013). f Immunofluorescence analysis of CD8 and GZMB expression in tumor tissues, with fluorescence intensity normalized to the RNF166-shNC group. (n = 5; two-tailed unpaired Student’s t-test; CD8, P < 0.0001; GZMB, P = 0.0002). Quantitative data are presented as mean ± SD. Scale bars, 50 μm PD-L1 knockdown attenuates RNF166-induced T-cell suppression. a Western blotting analysis of RNF166 and PD-L1 expression in RNF166-overexpressing Hepa1-6 cells after PD-L1 knockdown. Representative Western blots are shown from three independent experiments. b Representative images of subcutaneous tumors formed by RNF166-overexpressing Hepa1-6 cells expressing shNC or shPD-L1 in immunocompetent C57BL/6 mice. c Tumor volume and tumor weight measurements in the RNF166-shNC and RNF166-shPD-L1 groups. Each data point represents one mouse (n = 5 mice per group; two-tailed unpaired Student’s t-test; tumor volume, P = 0.0022; tumor weight, P = 0.0009). d Tumor growth curves in the RNF166-shNC and RNF166-shPD-L1 groups (n = 5 mice per group; two-way ANOVA followed by Šídák’s multiple comparisons test; adjusted P < 0.0001 at day 18). e Flow cytometric analysis of CD8⁺ T-cell infiltration and cytotoxic markers in tumor tissues. Each data point represents one mouse (n = 5 mice per group; two-tailed unpaired Student’s t-test; CD8⁺ T cells among CD45⁺ tumor-infiltrating lymphocytes, P < 0.0001; GZMB⁺ cells among CD8⁺ T cells, P = 0.0066; IFN-γ⁺ cells among CD8⁺ T cells, P = 0.0170; TNF-α⁺ cells among CD8⁺ T cells, P = 0.0013). f Immunofluorescence analysis of CD8 and GZMB expression in tumor tissues, with fluorescence intensity normalized to the RNF166-shNC group. (n = 5; two-tailed unpaired Student’s t-test; CD8, P < 0.0001; GZMB, P = 0.0002). Quantitative data are presented as mean ± SD. Scale bars, 50 μm RNF166 interacts with c-Jun We performed immunoprecipitation followed by mass spectrometry (IP-MS) to identify potential RNF166-interacting proteins and conducted Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. RNF166-interacting proteins were significantly enriched in immune-related pathways, including the PD-L1 expression and PD-1 checkpoint pathway in cancer and the type I interferon signaling pathway (Fig. 4a), suggesting a potential link between RNF166 and tumor immune-related signaling. Among the candidate proteins enriched in the PD-L1 expression and PD-1 checkpoint pathway, we focused on c-Jun because c-Jun is a well-established transcriptional regulator of PD-L1 (Fig. 4b). Mass spectrometry spectra revealed characteristic peptides of c-Jun in RNF166 immunoprecipitates (Fig. 4c), and structural docking predicted potential binding between RNF166 and c-Jun (Fig. 4d). We further validated this interaction using endogenous co-immunoprecipitation assays in HCC cells (Fig. 4e) and exogenous reciprocal co-immunoprecipitation assays in HEK293T cells (Fig. 4f), confirming that RNF166 associates with c-Jun. Immunofluorescence staining showed partial co-localization of RNF166 and c-Jun in the cytoplasm of Huh7 and HCCLM3 cells (Fig. 4g). Moreover, a GST pull-down assay showed that GST-tagged RNF166, but not the GST control, pulled down V5-tagged c-Jun (Fig. 4h). Collectively, these findings support a physical association and spatial overlap between RNF166 and c-Jun. Fig. 4RNF166 interacts with c-Jun. a KEGG pathway enrichment analysis of RNF166-interacting proteins identified by immunoprecipitation followed by mass spectrometry. Immune-related pathways, including the PD-L1 expression and PD-1 checkpoint pathway in cancer and type I interferon signaling, are highlighted. b Candidate RNF166-interacting proteins enriched in the PD-L1 expression and PD-1 checkpoint pathway in cancer. c Representative mass spectrometry spectra showing c-Jun peptides identified in RNF166 immunoprecipitates. d Structural docking analysis predicting the interaction between RNF166 and c-Jun. e Endogenous co-immunoprecipitation assays validating the interaction between RNF166 and c-Jun in HCCLM3 and Huh7 cells. f Exogenous reciprocal co-immunoprecipitation assays validating the interaction between Flag-RNF166 and V5-c-Jun in HEK293T cells. g Immunofluorescence staining showing the subcellular localization and co-localization of RNF166 and c-Jun in Huh7 and HCCLM3 cells. Scale bars, 10 μm. h GST pull-down assay assessing the physical association between GST-RNF166 and V5-c-Jun RNF166 interacts with c-Jun. a KEGG pathway enrichment analysis of RNF166-interacting proteins identified by immunoprecipitation followed by mass spectrometry. Immune-related pathways, including the PD-L1 expression and PD-1 checkpoint pathway in cancer and type I interferon signaling, are highlighted. b Candidate RNF166-interacting proteins enriched in the PD-L1 expression and PD-1 checkpoint pathway in cancer. c Representative mass spectrometry spectra showing c-Jun peptides identified in RNF166 immunoprecipitates. d Structural docking analysis predicting the interaction between RNF166 and c-Jun. e Endogenous co-immunoprecipitation assays validating the interaction between RNF166 and c-Jun in HCCLM3 and Huh7 cells. f Exogenous reciprocal co-immunoprecipitation assays validating the interaction between Flag-RNF166 and V5-c-Jun in HEK293T cells. g Immunofluorescence staining showing the subcellular localization and co-localization of RNF166 and c-Jun in Huh7 and HCCLM3 cells. Scale bars, 10 μm. h GST pull-down assay assessing the physical association between GST-RNF166 and V5-c-Jun RNF166 stabilizes c-Jun protein We examined the relationship between RNF166 and c-Jun expression in liver cancer cells. At the protein level (Fig. 5a), RNF166 expression was positively correlated with c-Jun expression, whereas at the mRNA level (Fig. 5b), neither knockdown nor overexpression of RNF166 significantly affected c-Jun expression. These results suggest that RNF166 regulates c-Jun primarily at the protein level. To investigate the underlying mechanism, cells were treated with the proteasome inhibitor MG132 or autophagy inhibitors (CQ, BafA1, and 3-MA). The results showed that MG132 markedly rescued the reduction in c-Jun protein levels caused by RNF166 knockdown (Fig. 5c, d). Further, protein degradation assays revealed that RNF166 knockdown accelerated c-Jun degradation (Fig. 5e and f), whereas RNF166 overexpression slowed V5-c-Jun degradation (Fig. 5g). These findings indicate that RNF166 stabilizes c-Jun protein by modulating proteasomal degradation. Fig. 5RNF166 stabilizes c-Jun protein. a Western blotting analysis of c-Jun protein levels after RNF166 knockdown in Huh7 and HCCLM3 cells or RNF166 overexpression in Hep3B cells. b qPCR analysis of c-Jun mRNA expression after RNF166 knockdown or overexpression. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.6831). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-1, P = 0.9449; shNC vs. shRNF166-2, P = 0.9283; HCCLM3: shNC vs. shRNF166-1, P = 0.6669; shNC vs. shRNF166-2, P = 0.7811). c, d Western blotting analysis of c-Jun protein levels after treatment with the proteasome inhibitor MG132 or the autophagy inhibitors CQ, BafA1, and 3-MA in RNF166-knockdown Huh7 and HCCLM3 cells. e, f Cycloheximide chase assays evaluating endogenous c-Jun protein degradation after RNF166 knockdown in Huh7 and HCCLM3 cells. g Cycloheximide chase assay evaluating V5-c-Jun protein degradation after RNF166 overexpression in Hep3B cells. Quantitative qPCR data are presented as mean ± SD. Representative Western blots are shown from three independent experiments RNF166 stabilizes c-Jun protein. a Western blotting analysis of c-Jun protein levels after RNF166 knockdown in Huh7 and HCCLM3 cells or RNF166 overexpression in Hep3B cells. b qPCR analysis of c-Jun mRNA expression after RNF166 knockdown or overexpression. Data were obtained from three independent biological experiments (n = 3). Hep3B Vec vs. RNF166 was analyzed using a two-tailed unpaired Student’s t-test (P = 0.6831). Huh7 and HCCLM3 shNC vs. shRNF166 groups were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (Huh7: shNC vs. shRNF166-1, P = 0.9449; shNC vs. shRNF166-2, P = 0.9283; HCCLM3: shNC vs. shRNF166-1, P = 0.6669; shNC vs. shRNF166-2, P = 0.7811). c, d Western blotting analysis of c-Jun protein levels after treatment with the proteasome inhibitor MG132 or the autophagy inhibitors CQ, BafA1, and 3-MA in RNF166-knockdown Huh7 and HCCLM3 cells. e, f Cycloheximide chase assays evaluating endogenous c-Jun protein degradation after RNF166 knockdown in Huh7 and HCCLM3 cells. g Cycloheximide chase assay evaluating V5-c-Jun protein degradation after RNF166 overexpression in Hep3B cells. Quantitative qPCR data are presented as mean ± SD. Representative Western blots are shown from three independent experiments RNF166 remodels the ubiquitin linkage pattern of c-Jun Because RNF166 is an E3 ubiquitin ligase, we hypothesized that it regulates c-Jun stability through ubiquitination. Ubiquitination assays showed that RNF166 did not markedly alter the overall ubiquitination level of c-Jun (Fig. 6a). To determine the ubiquitin linkage type regulated by RNF166, we performed ubiquitination assays using HA-Ub K-only and K-to-R mutants. These analyses indicated that RNF166 preferentially promoted K63-linked ubiquitination of c-Jun (Fig. 6b, c). We next examined K48- and K63-linked ubiquitination of c-Jun. In cells transfected with linkage-specific HA-Ub constructs, RNF166 decreased K48-linked ubiquitination of V5-c-Jun and increased K63-linked ubiquitination of V5-c-Jun (Fig. 6d, e). Dose-response experiments further showed that increasing Flag-RNF166 expression was associated with a progressive reduction in K48-linked ubiquitination and an increase in K63-linked ubiquitination of c-Jun, and these effects were confirmed by densitometric quantification from three independent experiments (Fig. 6e and Supplementary Fig. S7f, g). Together, these results indicate that RNF166 regulates the balance between K48- and K63-linked ubiquitination of c-Jun. To determine whether the E3 ligase activity of RNF166 is required for this effect, we generated a catalytically inactive RNF166-C33A/C36A mutant within the RING domain. Compared with wild-type RNF166, the RNF166-C33A/C36A mutant failed to efficiently reduce K48-linked ubiquitination or promote K63-linked ubiquitination of c-Jun (Fig. 6f and Supplementary Fig. S7h, i). Consistently, wild-type RNF166, but not the C33A/C36A mutant, increased c-Jun and PD-L1 protein levels (Supplementary Fig. S7a). In endogenous ubiquitination assays, RNF166 knockdown in Huh7 and HCCLM3 cells increased K48-linked ubiquitination and reduced K63-linked ubiquitination of endogenous c-Jun after c-Jun immunoprecipitation (Supplementary Fig. S7b–e). These findings indicate that RNF166 catalytic activity is required for RNF166-mediated regulation of c-Jun ubiquitination. To identify the major lysine residues involved in RNF166-associated K63-linked ubiquitination of c-Jun, we generated K-to-R point mutants of c-Jun. Mutational analysis showed that K50 and K56 were the major residues required for RNF166-associated K63-linked ubiquitination of c-Jun (Fig. 6g, h). Collectively, these results indicate that RNF166 stabilizes c-Jun by promoting K63-linked ubiquitination at K50/K56 and regulating the K48/K63 ubiquitination balance of c-Jun. Fig. 6RNF166 remodels the ubiquitin linkage pattern of c-Jun. a Ubiquitination assay evaluating the effect of RNF166 on the overall ubiquitination of V5-c-Jun in HEK293T cells. Cells were transfected with HA-Ub, Flag-RNF166, and V5-c-Jun as indicated and treated with MG132 before immunoprecipitation. b, c Ubiquitination assays using HA-Ub K-only and K-to-R mutants to determine the ubiquitin linkage type regulated by RNF166. d Ubiquitination assay assessing the effect of RNF166 on K63-linked ubiquitination of V5-c-Jun. e Ubiquitination assay examining the dose-dependent effects of RNF166 on K48-linked and K63-linked ubiquitination of V5-c-Jun. f Ubiquitination assay comparing the effects of wild-type RNF166 and the catalytically inactive RNF166-C33A/C36A mutant on K48-linked and K63-linked ubiquitination of V5-c-Jun. g, h K-to-R mutational analysis identifying K50 and K56 as major lysine residues required for RNF166-associated K63-linked ubiquitination of c-Jun. V5-c-Jun was immunoprecipitated, and ubiquitinated c-Jun was detected using anti-HA antibody. Representative blots are shown from three independent experiments RNF166 remodels the ubiquitin linkage pattern of c-Jun. a Ubiquitination assay evaluating the effect of RNF166 on the overall ubiquitination of V5-c-Jun in HEK293T cells. Cells were transfected with HA-Ub, Flag-RNF166, and V5-c-Jun as indicated and treated with MG132 before immunoprecipitation. b, c Ubiquitination assays using HA-Ub K-only and K-to-R mutants to determine the ubiquitin linkage type regulated by RNF166. d Ubiquitination assay assessing the effect of RNF166 on K63-linked ubiquitination of V5-c-Jun. e Ubiquitination assay examining the dose-dependent effects of RNF166 on K48-linked and K63-linked ubiquitination of V5-c-Jun. f Ubiquitination assay comparing the effects of wild-type RNF166 and the catalytically inactive RNF166-C33A/C36A mutant on K48-linked and K63-linked ubiquitination of V5-c-Jun. g, h K-to-R mutational analysis identifying K50 and K56 as major lysine residues required for RNF166-associated K63-linked ubiquitination of c-Jun. V5-c-Jun was immunoprecipitated, and ubiquitinated c-Jun was detected using anti-HA antibody. Representative blots are shown from three independent experiments RNF166 upregulates PD-L1 through c-Jun stabilization and promoter binding To determine whether RNF166 regulates PD-L1 through c-Jun, we overexpressed c-Jun in RNF166-knockdown cells and knocked down c-Jun in RNF166-overexpressing cells. Western blotting, qPCR, and flow cytometry showed that c-Jun overexpression restored PD-L1 expression in RNF166-knockdown cells, whereas c-Jun knockdown reduced PD-L1 levels in RNF166-overexpressing cells (Fig. 7a–c). Treatment with the c-Jun inhibitor T-5224 also decreased PD-L1 expression in RNF166-overexpressing cells (Fig. 7d–f), supporting the role of c-Jun in RNF166-mediated PD-L1 upregulation. Because K50 and K56 were identified as major residues involved in RNF166-associated K63-linked ubiquitination of c-Jun, we next examined whether these sites were required for c-Jun stability and PD-L1 regulation. In RNF166-overexpressing Hep3B cells, endogenous c-Jun was knocked down and then rescued with either wild-type c-Jun or the K50R/K56R double mutant c-Jun-2KR. Wild-type c-Jun restored PD-L1 protein and mRNA expression more effectively than c-Jun-2KR (Fig. 7g, h). Cycloheximide chase assays further showed that c-Jun-2KR was less stable than wild-type c-Jun in RNF166-overexpressing cells (Fig. 7i). These results indicate that K50/K56-dependent regulation of c-Jun stability contributes to RNF166-mediated PD-L1 expression. We then examined whether c-Jun directly regulates PD-L1 transcription. JASPAR analysis identified predicted c-Jun-binding motifs in the PD-L1 promoter (Fig. 7j). Dual-luciferase reporter assays showed that c-Jun increased PD-L1 promoter activity, whereas c-Jun knockdown attenuated RNF166-induced promoter activation (Fig. 7k, l). Mutational analysis showed that binding sites 2 and 3 were required for c-Jun-mediated PD-L1 promoter activation (Fig. 7m). Consistently, ChIP-qPCR confirmed c-Jun enrichment at sites 2 and 3 of the PD-L1 promoter (Fig. 7n). To determine whether RNF166 promotes PD-L1 transcription through these c-Jun-binding sites, we performed promoter-mutant luciferase assays under RNF166 overexpression. Mutation of site 2, site 3, or both sites markedly reduced RNF166-induced PD-L1 promoter activation (Fig. 7o). In addition, ChIP-qPCR showed that RNF166 overexpression increased c-Jun enrichment at PD-L1 promoter sites 2 and 3 (Fig. 7p). Together, these findings indicate that RNF166 stabilizes c-Jun and promotes c-Jun binding to the PD-L1 promoter, thereby upregulating PD-L1 transcription. To further assess the tissue relevance of this axis, we analyzed HCC tissue microarrays and mouse tumor tissues. In clinical HCC samples, RNF166 expression was positively correlated with c-Jun and PD-L1 expression and negatively correlated with CD8 and GZMB expression (Supplementary Fig. S8a). Multiplex immunofluorescence in mouse tumor tissues further supported increased RNF166, c-Jun, and PD-L1 signals in RNF166-overexpressing tumors (Supplementary Fig. S8b). These tissue-level findings are consistent with the RNF166-c-Jun-PD-L1 axis identified in cellular assays. Fig. 7RNF166 promotes PD-L1 transcription through c-Jun stabilization and c-Jun binding to the PD-L1 promoter. a Western blotting analysis of RNF166, c-Jun, and PD-L1 expression after c-Jun rescue in RNF166-knockdown Huh7 and HCCLM3 cells or c-Jun knockdown in RNF166-overexpressing Hep3B cells. Representative Western blots are shown from three independent experiments. b qPCR analysis of PD-L1 mRNA expression after c-Jun overexpression in RNF166-knockdown Huh7 and HCCLM3 cells or c-Jun knockdown in RNF166-overexpressing Hep3B cells. Data were obtained from three independent biological experiments (n = 3). Huh7 and HCCLM3 rescue experiments were analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0003; HCCLM3, P < 0.0001). Hep3B c-Jun knockdown experiments were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (shNC vs. sh-c-Jun-1, P < 0.0001; shNC vs. sh-c-Jun-2, P < 0.0001). c Flow cytometric analysis of cell-surface PD-L1 expression after c-Jun rescue in RNF166-knockdown Huh7 and HCCLM3 cells or c-Jun knockdown in RNF166-overexpressing Hep3B cells. Data were obtained from three independent biological experiments (n = 3). Huh7 and HCCLM3 rescue experiments were analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0015; HCCLM3, P = 0.0010). Hep3B c-Jun knockdown experiments were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (RNF166-shNC vs. RNF166-sh-c-Jun-1, P < 0.0001; RNF166-shNC vs. RNF166-sh-c-Jun-2, P < 0.0001). d Western blotting analysis of c-Jun and PD-L1 expression in RNF166-overexpressing Hep3B cells treated with the c-Jun inhibitor T-5224. e, f qPCR and flow cytometric analyses of PD-L1 expression in RNF166-overexpressing Hep3B cells after T-5224 treatment. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-tailed unpaired Student’s t-tests (qPCR, P = 0.0089; flow cytometry, P = 0.0065). g Western blotting analysis of RNF166, c-Jun, and PD-L1 expression in RNF166-overexpressing Hep3B cells after endogenous c-Jun knockdown followed by rescue with V5-c-Jun-WT or V5-c-Jun-2KR. h qPCR analysis of PD-L1 mRNA expression in RNF166-overexpressing Hep3B cells after c-Jun knockdown and rescue with V5-c-Jun-WT or V5-c-Jun-2KR. Data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test (shNC vs. sh-c-Jun, P < 0.0001; sh-c-Jun vs. sh-c-Jun + V5-c-Jun-WT, P < 0.0001; sh-c-Jun + V5-c-Jun-WT vs. sh-c-Jun + V5-c-Jun-2KR, P = 0.0001). i Cycloheximide chase assay comparing the stability of V5-c-Jun-WT and V5-c-Jun-2KR in RNF166-overexpressing Hep3B cells after endogenous c-Jun knockdown. Representative Western blots are shown from three independent experiments. j Predicted c-Jun-binding motif in the PD-L1 promoter. k Dual-luciferase reporter assays assessing the effect of c-Jun overexpression on PD-L1 promoter activity in Huh7 and Hep3B cells. Firefly luciferase activity was normalized to Renilla luciferase activity and expressed relative to the pCMV-Tag control group. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0020; Hep3B, P = 0.0005). l Dual-luciferase reporter assay assessing PD-L1 promoter activity after RNF166 overexpression and c-Jun knockdown in Hep3B cells. Data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test (control vs. RNF166, P = 0.0002; RNF166 + si-Control vs. RNF166 + si-c-Jun, P = 0.0012). m Dual-luciferase reporter assay assessing the effects of mutations in predicted c-Jun-binding sites on PD-L1 promoter activity. Data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the WT promoter construct serving as the control (WT vs. Mut1, P = 0.6083; WT vs. Mut2, P = 0.0002; WT vs. Mut3, P = 0.0001; WT vs. Mut4, P = 0.9967). n ChIP-qPCR analysis of c-Jun enrichment at the indicated regions of the PD-L1 promoter in Hep3B-Vec and Hep3B-c-Jun cells. Data were obtained from three independent biological experiments (n = 3) and analyzed on log2-transformed values using two-way repeated-measures ANOVA followed by Šídák’s multiple comparisons test (distant region, P = 0.9965; binding site 1, P = 0.3787; binding site 2, P < 0.0001; binding site 3, P < 0.0001; binding site 4, P = 0.9025). o Dual-luciferase reporter assay assessing the effects of mutations in c-Jun-binding sites 2 and/or 3 on RNF166-induced PD-L1 promoter activity in Hep3B cells. Data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test (Vector + WT vs. RNF166 + WT, P < 0.0001; RNF166 + WT vs. RNF166 + Site 2 mutant, P < 0.0001; RNF166 + WT vs. RNF166 + Site 3 mutant, P < 0.0001; RNF166 + WT vs. RNF166 + Site 2/3 double mutant, P < 0.0001). p ChIP-qPCR analysis of c-Jun enrichment at PD-L1 promoter binding sites 2 and 3 after RNF166 overexpression in Hep3B cells. Data were obtained from three independent biological experiments (n = 3) and analyzed on log2-transformed values using two-way repeated-measures ANOVA followed by Šídák’s multiple comparisons test; adjusted P values for binding sites 2 and 3 were both P < 0.0001. Quantitative data are presented as mean ± SD RNF166 promotes PD-L1 transcription through c-Jun stabilization and c-Jun binding to the PD-L1 promoter. a Western blotting analysis of RNF166, c-Jun, and PD-L1 expression after c-Jun rescue in RNF166-knockdown Huh7 and HCCLM3 cells or c-Jun knockdown in RNF166-overexpressing Hep3B cells. Representative Western blots are shown from three independent experiments. b qPCR analysis of PD-L1 mRNA expression after c-Jun overexpression in RNF166-knockdown Huh7 and HCCLM3 cells or c-Jun knockdown in RNF166-overexpressing Hep3B cells. Data were obtained from three independent biological experiments (n = 3). Huh7 and HCCLM3 rescue experiments were analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0003; HCCLM3, P < 0.0001). Hep3B c-Jun knockdown experiments were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (shNC vs. sh-c-Jun-1, P < 0.0001; shNC vs. sh-c-Jun-2, P < 0.0001). c Flow cytometric analysis of cell-surface PD-L1 expression after c-Jun rescue in RNF166-knockdown Huh7 and HCCLM3 cells or c-Jun knockdown in RNF166-overexpressing Hep3B cells. Data were obtained from three independent biological experiments (n = 3). Huh7 and HCCLM3 rescue experiments were analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0015; HCCLM3, P = 0.0010). Hep3B c-Jun knockdown experiments were analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (RNF166-shNC vs. RNF166-sh-c-Jun-1, P < 0.0001; RNF166-shNC vs. RNF166-sh-c-Jun-2, P < 0.0001). d Western blotting analysis of c-Jun and PD-L1 expression in RNF166-overexpressing Hep3B cells treated with the c-Jun inhibitor T-5224. e, f qPCR and flow cytometric analyses of PD-L1 expression in RNF166-overexpressing Hep3B cells after T-5224 treatment. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-tailed unpaired Student’s t-tests (qPCR, P = 0.0089; flow cytometry, P = 0.0065). g Western blotting analysis of RNF166, c-Jun, and PD-L1 expression in RNF166-overexpressing Hep3B cells after endogenous c-Jun knockdown followed by rescue with V5-c-Jun-WT or V5-c-Jun-2KR. h qPCR analysis of PD-L1 mRNA expression in RNF166-overexpressing Hep3B cells after c-Jun knockdown and rescue with V5-c-Jun-WT or V5-c-Jun-2KR. Data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test (shNC vs. sh-c-Jun, P < 0.0001; sh-c-Jun vs. sh-c-Jun + V5-c-Jun-WT, P < 0.0001; sh-c-Jun + V5-c-Jun-WT vs. sh-c-Jun + V5-c-Jun-2KR, P = 0.0001). i Cycloheximide chase assay comparing the stability of V5-c-Jun-WT and V5-c-Jun-2KR in RNF166-overexpressing Hep3B cells after endogenous c-Jun knockdown. Representative Western blots are shown from three independent experiments. j Predicted c-Jun-binding motif in the PD-L1 promoter. k Dual-luciferase reporter assays assessing the effect of c-Jun overexpression on PD-L1 promoter activity in Huh7 and Hep3B cells. Firefly luciferase activity was normalized to Renilla luciferase activity and expressed relative to the pCMV-Tag control group. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0020; Hep3B, P = 0.0005). l Dual-luciferase reporter assay assessing PD-L1 promoter activity after RNF166 overexpression and c-Jun knockdown in Hep3B cells. Data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test (control vs. RNF166, P = 0.0002; RNF166 + si-Control vs. RNF166 + si-c-Jun, P = 0.0012). m Dual-luciferase reporter assay assessing the effects of mutations in predicted c-Jun-binding sites on PD-L1 promoter activity. Data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the WT promoter construct serving as the control (WT vs. Mut1, P = 0.6083; WT vs. Mut2, P = 0.0002; WT vs. Mut3, P = 0.0001; WT vs. Mut4, P = 0.9967). n ChIP-qPCR analysis of c-Jun enrichment at the indicated regions of the PD-L1 promoter in Hep3B-Vec and Hep3B-c-Jun cells. Data were obtained from three independent biological experiments (n = 3) and analyzed on log2-transformed values using two-way repeated-measures ANOVA followed by Šídák’s multiple comparisons test (distant region, P = 0.9965; binding site 1, P = 0.3787; binding site 2, P < 0.0001; binding site 3, P < 0.0001; binding site 4, P = 0.9025). o Dual-luciferase reporter assay assessing the effects of mutations in c-Jun-binding sites 2 and/or 3 on RNF166-induced PD-L1 promoter activity in Hep3B cells. Data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test (Vector + WT vs. RNF166 + WT, P < 0.0001; RNF166 + WT vs. RNF166 + Site 2 mutant, P < 0.0001; RNF166 + WT vs. RNF166 + Site 3 mutant, P < 0.0001; RNF166 + WT vs. RNF166 + Site 2/3 double mutant, P < 0.0001). p ChIP-qPCR analysis of c-Jun enrichment at PD-L1 promoter binding sites 2 and 3 after RNF166 overexpression in Hep3B cells. Data were obtained from three independent biological experiments (n = 3) and analyzed on log2-transformed values using two-way repeated-measures ANOVA followed by Šídák’s multiple comparisons test; adjusted P values for binding sites 2 and 3 were both P < 0.0001. Quantitative data are presented as mean ± SD Targeting the RNF166–c-Jun axis enhances the response to PD-L1 blockade To explore potential therapeutic strategies based on the RNF166-c-Jun-PD-L1 axis, we first evaluated the antitumor effect of anti-PD-L1 antibody treatment in immunocompetent mice. Hepa1-6 cells overexpressing RNF166 or vector control cells were subcutaneously implanted, followed by treatment with anti-PD-L1 mAb or control IgG (Fig. 8a). Anti-PD-L1 blockade produced a more pronounced tumor-suppressive effect in RNF166-overexpressing tumors, as shown by gross tumor images, endpoint tumor volume and weight, and tumor growth curves (Fig. 8b–d). Flow cytometry analysis showed that anti-PD-L1 treatment increased CD8 + T-cell infiltration in RNF166-overexpressing tumors, accompanied by elevated GZMB, TNF-α, and IFN-γ expression (Fig. 8e). Immunofluorescence staining also showed increased CD8 and GZMB signals in tumor tissues after anti-PD-L1 treatment (Fig. 8f). To determine whether the antitumor efficacy of anti-PD-L1 blockade depended on CD8 + T cells, RNF166-overexpressing tumor-bearing mice were treated with anti-PD-L1 mAb together with either control IgG or CD8α-depleting antibody. CD8α depletion markedly weakened the antitumor effect of anti-PD-L1 blockade, as shown by gross tumor images, tumor volume, tumor weight, and tumor growth curves (Supplementary Fig. S9a–d). These results indicate that anti-PD-L1 blockade suppresses RNF166-overexpressing tumors largely through CD8 + T-cell-mediated antitumor immunity. Because RNF166 promoted PD-L1 expression through c-Jun stabilization, we next tested whether simultaneous inhibition of c-Jun and PD-L1 could enhance antitumor efficacy in the RNF166-overexpressing model. In RNF166-overexpressing Hepa1-6 tumors, both T-5224 and anti-PD-L1 mAb monotherapy reduced tumor growth, whereas the combination treatment produced stronger tumor suppression than either monotherapy, as shown by gross tumor images, endpoint tumor volume and weight, and tumor growth curves (Fig. 8g–j). Kaplan-Meier analysis further showed that the combination treatment prolonged survival compared with vehicle or either monotherapy (Fig. 8k). We also performed the same treatment experiment in vector-control Hepa1-6 tumors to assess the general activity of this regimen. T-5224 and anti-PD-L1 mAb also suppressed vector-control tumor growth in the absence of RNF166 overexpression, indicating that the treatment effect was not exclusively RNF166-dependent (Supplementary Fig. S9e–g). Nevertheless, the RNF166-overexpressing model provides a mechanistically relevant setting in which activation of the RNF166-c-Jun-PD-L1 axis can be counteracted by combined c-Jun inhibition and anti-PD-L1 blockade. Together, these findings indicate that combined c-Jun inhibition and anti-PD-L1 blockade can effectively suppress RNF166-driven tumor growth while also showing broader antitumor activity in this model. Fig. 8PD-L1 blockade and c-Jun inhibition counteract RNF166-driven tumor growth and improve survival in vivo. a Schematic illustration of the PD-L1 blockade treatment schedule in the subcutaneous Hepa1-6 tumor model. Mice were inoculated with Hepa1-6 cells expressing vector control or RNF166 and treated with IgG control or αPD-L1 mAb as indicated. Tumor tissues were collected on day 18. b Representative images of subcutaneous tumors from the Vec + IgG, RNF166 + IgG, Vec + αPD-L1 mAb, and RNF166 + αPD-L1 mAb groups. c Tumor volume and tumor weight measurements at endpoint (n= 5 mice per group). Tumor volume: Vec + IgG vs. RNF166 + IgG, P = 0.0016; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0029; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P = 0.0002. Tumor weight: Vec + IgG vs. RNF166 + IgG, P = 0.0011; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0012; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P = 0.0001. d Tumor growth curves in the indicated groups (n = 5 mice per group). The indicated comparisons were assessed on day 18 (Vec + IgG vs. RNF166 + IgG, P < 0.0001; Vec + IgG vs. Vec + αPD-L1 mAb, P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001). e Flow cytometric analysis of CD8⁺ T-cell infiltration and cytotoxic markers in tumor tissues (n = 5 mice per group). CD8⁺ T cells among CD45⁺ tumor-infiltrating lymphocytes: Vec + IgG vs. RNF166 + IgG, P = 0.0037; Vec + IgG vs. Vec + αPD-L1 mAb, P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001. GZMB⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, P = 0.0054; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001. IFN-γ⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, P = 0.0234; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001. TNF-α⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, P = 0.0401; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0453; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P = 0.0006. f Immunofluorescence analysis of CD8 and GZMB expression in tumor tissues from the indicated groups. CD8: Vec + IgG vs. RNF166 + IgG, P = 0.0350; Vec + IgG vs. Vec + αPD-L1 mAb, P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001. GZMB: Vec + IgG vs. RNF166 + IgG, P = 0.0362; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0002; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P = 0.0003. Scale bars, 50 μm. g Schematic illustration of the treatment schedule for RNF166-overexpressing Hepa1-6 tumors treated with IgG+Vehicle, IgG + T-5224, αPD-L1 mAb+Vehicle, or αPD-L1 mAb + T-5224. h Representative images of subcutaneous tumors from the indicated treatment groups. i Tumor volume and tumor weight measurements at endpoint (n = 5 mice per group). Tumor volume: IgG+Vehicle vs. IgG + T-5224, P = 0.0017; IgG+Vehicle vs. αPD-L1 mAb+Vehicle, P = 0.0019; IgG+Vehicle vs. αPD-L1 mAb + T-5224, P = 0.0005; IgG + T-5224 vs. αPD-L1 mAb + T-5224, P = 0.0037; αPD-L1 mAb+Vehicle vs. αPD-L1 mAb + T-5224, P = 0.0027. Tumor weight: IgG+Vehicle vs. IgG + T-5224, P = 0.0028; IgG+Vehicle vs. αPD-L1 mAb+Vehicle, P = 0.0012; IgG+Vehicle vs. αPD-L1 mAb + T-5224, P = 0.0005; IgG + T-5224 vs. αPD-L1 mAb + T-5224, P = 0.0021; αPD-L1 mAb+Vehicle vs. αPD-L1 mAb + T-5224, P = 0.0058. j Tumor growth curves in the indicated treatment groups (n = 5 mice per group). The indicated comparisons were assessed on day 23 (IgG+Vehicle vs. IgG + T-5224, P < 0.0001; IgG+Vehicle vs. αPD-L1 mAb+Vehicle, P < 0.0001; IgG+Vehicle vs. αPD-L1 mAb + T-5224, P < 0.0001). k Kaplan–Meier survival analysis of mice treated with IgG+Vehicle, IgG + T-5224, αPD-L1 mAb+Vehicle, or αPD-L1 mAb + T-5224. Each group included 10 mice. Median survival times were 29.5, 41.0, 44.0, and 59.5 days, respectively. Overall survival curves were compared using the log-rank Mantel-Cox test (χ2 = 69.58, df = 3, P < 0.0001). Pairwise log-rank analyses showed significant differences between IgG+Vehicle and IgG + T-5224 (χ2 = 20.06, P < 0.0001), IgG+Vehicle and αPD-L1 mAb+Vehicle (χ2 = 19.32, P < 0.0001), IgG + T-5224 and αPD-L1 mAb + T-5224 (χ2 = 19.47, P < 0.0001), and αPD-L1 mAb+Vehicle and αPD-L1 mAb + T-5224 (χ2 = 18.37, PPD-L1 blockade and c-Jun inhibition counteract RNF166-driven tumor growth and improve survival in vivo. All experiments shown in this figure were performed using immunocompetent C57BL/6 mice. a Schematic illustration of the PD-L1 blockade treatment schedule in the subcutaneous Hepa1-6 tumor model. Mice were inoculated with Hepa1-6 cells expressing vector control or RNF166 and treated with IgG control or αPD-L1 mAb as indicated. Tumor tissues were collected on day 18. b Representative images of subcutaneous tumors from the Vec + IgG, RNF166 + IgG, Vec + αPD-L1 mAb, and RNF166 + αPD-L1 mAb groups. c Tumor volume and tumor weight measurements at endpoint (n = 5 mice per group). Data were analyzed using Brown–Forsythe and Welch ANOVA followed by Dunnett’s T3 multiple comparisons test. Tumor volume: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0016; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0029; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P = 0.0002. Tumor weight: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0011; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0012; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P = 0.0001. d Tumor growth curves in the indicated groups (n = 5 mice per group). Data were analyzed using ordinary two-way ANOVA followed by Šídák’s multiple comparisons test. The indicated comparisons were assessed on day 18 (Vec + IgG vs. RNF166 + IgG, adjusted P < 0.0001; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001). e Flow cytometric analysis of CD8⁺ T-cell infiltration and cytotoxic markers in tumor tissues (n = 5 mice per group). Data were analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test. CD8⁺ T cells among CD45⁺ tumor-infiltrating lymphocytes: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0037; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001. GZMB⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0054; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001. IFN-γ⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0234; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001. TNF-α⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0401; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0453; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P = 0.0006. f Immunofluorescence analysis of CD8 and GZMB expression in tumor tissues from the indicated groups (n = 5 mice per group). Data were analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test. CD8: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0350; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001. GZMB: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0362; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0002; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P = 0.0003. Scale bars, 50 μm. g Schematic illustration of the treatment schedule for RNF166-overexpressing Hepa1-6 tumors treated with IgG + Vehicle, IgG + T-5224, αPD-L1 mAb + Vehicle, or αPD-L1 mAb + T-5224. h Representative images of subcutaneous tumors from the indicated treatment groups. i Tumor volume and tumor weight measurements at endpoint (n = 5 mice per group). Data were analyzed using Brown–Forsythe and Welch ANOVA followed by Dunnett’s T3 multiple comparisons test. Tumor volume: IgG + Vehicle vs. IgG + T-5224, adjusted P = 0.0017; IgG + Vehicle vs. αPD-L1 mAb + Vehicle, adjusted P = 0.0019; IgG + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P = 0.0005; IgG + T-5224 vs. αPD-L1 mAb + T-5224, adjusted P = 0.0037; αPD-L1 mAb + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P = 0.0027. Tumor weight: IgG + Vehicle vs. IgG + T-5224, adjusted P = 0.0028; IgG + Vehicle vs. αPD-L1 mAb + Vehicle, adjusted P = 0.0012; IgG + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P = 0.0005; IgG + T-5224 vs. αPD-L1 mAb + T-5224, adjusted P = 0.0021; αPD-L1 mAb + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P = 0.0058. j Tumor growth curves in the indicated treatment groups (n = 5 mice per group). Data were analyzed using ordinary two-way ANOVA followed by Šídák’s multiple comparisons test. The indicated comparisons were assessed on day 23 (IgG + Vehicle vs. IgG + T-5224, adjusted P < 0.0001; IgG + Vehicle vs. αPD-L1 mAb + Vehicle, adjusted P < 0.0001; IgG + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P < 0.0001). k Kaplan–Meier survival analysis of mice treated with IgG + Vehicle, IgG + T-5224, αPD-L1 mAb + Vehicle, or αPD-L1 mAb + T-5224. Each group included 10 mice. Median survival times were 29.5, 41.0, 44.0, and 59.5 days, respectively. Overall survival curves were compared using the log-rank Mantel–Cox test (χ² = 69.58, df = 3, P < 0.0001). Pairwise log-rank analyses showed significant differences between IgG + Vehicle and IgG + T-5224 (χ² = 20.06, P < 0.0001), IgG + Vehicle and αPD-L1 mAb + Vehicle (χ² = 19.32, P < 0.0001), IgG + T-5224 and αPD-L1 mAb + T-5224 (χ² = 19.47, P < 0.0001), and αPD-L1 mAb + Vehicle and αPD-L1 mAb + T-5224 (χ² = 18.37, P < 0.0001). Tick marks indicate censored observations. Quantitative data are presented as mean ± SD. PD-L1 blockade and c-Jun inhibition counteract RNF166-driven tumor growth and improve survival in vivo. a Schematic illustration of the PD-L1 blockade treatment schedule in the subcutaneous Hepa1-6 tumor model. Mice were inoculated with Hepa1-6 cells expressing vector control or RNF166 and treated with IgG control or αPD-L1 mAb as indicated. Tumor tissues were collected on day 18. b Representative images of subcutaneous tumors from the Vec + IgG, RNF166 + IgG, Vec + αPD-L1 mAb, and RNF166 + αPD-L1 mAb groups. c Tumor volume and tumor weight measurements at endpoint (n= 5 mice per group). Tumor volume: Vec + IgG vs. RNF166 + IgG, P = 0.0016; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0029; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P = 0.0002. Tumor weight: Vec + IgG vs. RNF166 + IgG, P = 0.0011; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0012; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P = 0.0001. d Tumor growth curves in the indicated groups (n = 5 mice per group). The indicated comparisons were assessed on day 18 (Vec + IgG vs. RNF166 + IgG, P < 0.0001; Vec + IgG vs. Vec + αPD-L1 mAb, P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001). e Flow cytometric analysis of CD8⁺ T-cell infiltration and cytotoxic markers in tumor tissues (n = 5 mice per group). CD8⁺ T cells among CD45⁺ tumor-infiltrating lymphocytes: Vec + IgG vs. RNF166 + IgG, P = 0.0037; Vec + IgG vs. Vec + αPD-L1 mAb, P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001. GZMB⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, P = 0.0054; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001. IFN-γ⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, P = 0.0234; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001. TNF-α⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, P = 0.0401; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0453; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P = 0.0006. f Immunofluorescence analysis of CD8 and GZMB expression in tumor tissues from the indicated groups. CD8: Vec + IgG vs. RNF166 + IgG, P = 0.0350; Vec + IgG vs. Vec + αPD-L1 mAb, P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P < 0.0001. GZMB: Vec + IgG vs. RNF166 + IgG, P = 0.0362; Vec + IgG vs. Vec + αPD-L1 mAb, P = 0.0002; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, P = 0.0003. Scale bars, 50 μm. g Schematic illustration of the treatment schedule for RNF166-overexpressing Hepa1-6 tumors treated with IgG+Vehicle, IgG + T-5224, αPD-L1 mAb+Vehicle, or αPD-L1 mAb + T-5224. h Representative images of subcutaneous tumors from the indicated treatment groups. i Tumor volume and tumor weight measurements at endpoint (n = 5 mice per group). Tumor volume: IgG+Vehicle vs. IgG + T-5224, P = 0.0017; IgG+Vehicle vs. αPD-L1 mAb+Vehicle, P = 0.0019; IgG+Vehicle vs. αPD-L1 mAb + T-5224, P = 0.0005; IgG + T-5224 vs. αPD-L1 mAb + T-5224, P = 0.0037; αPD-L1 mAb+Vehicle vs. αPD-L1 mAb + T-5224, P = 0.0027. Tumor weight: IgG+Vehicle vs. IgG + T-5224, P = 0.0028; IgG+Vehicle vs. αPD-L1 mAb+Vehicle, P = 0.0012; IgG+Vehicle vs. αPD-L1 mAb + T-5224, P = 0.0005; IgG + T-5224 vs. αPD-L1 mAb + T-5224, P = 0.0021; αPD-L1 mAb+Vehicle vs. αPD-L1 mAb + T-5224, P = 0.0058. j Tumor growth curves in the indicated treatment groups (n = 5 mice per group). The indicated comparisons were assessed on day 23 (IgG+Vehicle vs. IgG + T-5224, P < 0.0001; IgG+Vehicle vs. αPD-L1 mAb+Vehicle, P < 0.0001; IgG+Vehicle vs. αPD-L1 mAb + T-5224, P < 0.0001). k Kaplan–Meier survival analysis of mice treated with IgG+Vehicle, IgG + T-5224, αPD-L1 mAb+Vehicle, or αPD-L1 mAb + T-5224. Each group included 10 mice. Median survival times were 29.5, 41.0, 44.0, and 59.5 days, respectively. Overall survival curves were compared using the log-rank Mantel-Cox test (χ2 = 69.58, df = 3, P < 0.0001). Pairwise log-rank analyses showed significant differences between IgG+Vehicle and IgG + T-5224 (χ2 = 20.06, P < 0.0001), IgG+Vehicle and αPD-L1 mAb+Vehicle (χ2 = 19.32, P < 0.0001), IgG + T-5224 and αPD-L1 mAb + T-5224 (χ2 = 19.47, P < 0.0001), and αPD-L1 mAb+Vehicle and αPD-L1 mAb + T-5224 (χ2 = 18.37, PPD-L1 blockade and c-Jun inhibition counteract RNF166-driven tumor growth and improve survival in vivo. All experiments shown in this figure were performed using immunocompetent C57BL/6 mice. a Schematic illustration of the PD-L1 blockade treatment schedule in the subcutaneous Hepa1-6 tumor model. Mice were inoculated with Hepa1-6 cells expressing vector control or RNF166 and treated with IgG control or αPD-L1 mAb as indicated. Tumor tissues were collected on day 18. b Representative images of subcutaneous tumors from the Vec + IgG, RNF166 + IgG, Vec + αPD-L1 mAb, and RNF166 + αPD-L1 mAb groups. c Tumor volume and tumor weight measurements at endpoint (n = 5 mice per group). Data were analyzed using Brown–Forsythe and Welch ANOVA followed by Dunnett’s T3 multiple comparisons test. Tumor volume: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0016; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0029; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P = 0.0002. Tumor weight: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0011; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0012; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P = 0.0001. d Tumor growth curves in the indicated groups (n = 5 mice per group). Data were analyzed using ordinary two-way ANOVA followed by Šídák’s multiple comparisons test. The indicated comparisons were assessed on day 18 (Vec + IgG vs. RNF166 + IgG, adjusted P < 0.0001; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001). e Flow cytometric analysis of CD8⁺ T-cell infiltration and cytotoxic markers in tumor tissues (n = 5 mice per group). Data were analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test. CD8⁺ T cells among CD45⁺ tumor-infiltrating lymphocytes: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0037; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001. GZMB⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0054; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001. IFN-γ⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0234; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001. TNF-α⁺ cells among CD8⁺ T cells: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0401; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0453; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P = 0.0006. f Immunofluorescence analysis of CD8 and GZMB expression in tumor tissues from the indicated groups (n = 5 mice per group). Data were analyzed using ordinary one-way ANOVA followed by Tukey’s multiple comparisons test. CD8: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0350; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P < 0.0001; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P < 0.0001. GZMB: Vec + IgG vs. RNF166 + IgG, adjusted P = 0.0362; Vec + IgG vs. Vec + αPD-L1 mAb, adjusted P = 0.0002; RNF166 + IgG vs. RNF166 + αPD-L1 mAb, adjusted P = 0.0003. Scale bars, 50 μm. g Schematic illustration of the treatment schedule for RNF166-overexpressing Hepa1-6 tumors treated with IgG + Vehicle, IgG + T-5224, αPD-L1 mAb + Vehicle, or αPD-L1 mAb + T-5224. h Representative images of subcutaneous tumors from the indicated treatment groups. i Tumor volume and tumor weight measurements at endpoint (n = 5 mice per group). Data were analyzed using Brown–Forsythe and Welch ANOVA followed by Dunnett’s T3 multiple comparisons test. Tumor volume: IgG + Vehicle vs. IgG + T-5224, adjusted P = 0.0017; IgG + Vehicle vs. αPD-L1 mAb + Vehicle, adjusted P = 0.0019; IgG + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P = 0.0005; IgG + T-5224 vs. αPD-L1 mAb + T-5224, adjusted P = 0.0037; αPD-L1 mAb + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P = 0.0027. Tumor weight: IgG + Vehicle vs. IgG + T-5224, adjusted P = 0.0028; IgG + Vehicle vs. αPD-L1 mAb + Vehicle, adjusted P = 0.0012; IgG + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P = 0.0005; IgG + T-5224 vs. αPD-L1 mAb + T-5224, adjusted P = 0.0021; αPD-L1 mAb + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P = 0.0058. j Tumor growth curves in the indicated treatment groups (n = 5 mice per group). Data were analyzed using ordinary two-way ANOVA followed by Šídák’s multiple comparisons test. The indicated comparisons were assessed on day 23 (IgG + Vehicle vs. IgG + T-5224, adjusted P < 0.0001; IgG + Vehicle vs. αPD-L1 mAb + Vehicle, adjusted P < 0.0001; IgG + Vehicle vs. αPD-L1 mAb + T-5224, adjusted P < 0.0001). k Kaplan–Meier survival analysis of mice treated with IgG + Vehicle, IgG + T-5224, αPD-L1 mAb + Vehicle, or αPD-L1 mAb + T-5224. Each group included 10 mice. Median survival times were 29.5, 41.0, 44.0, and 59.5 days, respectively. Overall survival curves were compared using the log-rank Mantel–Cox test (χ² = 69.58, df = 3, P < 0.0001). Pairwise log-rank analyses showed significant differences between IgG + Vehicle and IgG + T-5224 (χ² = 20.06, P < 0.0001), IgG + Vehicle and αPD-L1 mAb + Vehicle (χ² = 19.32, P < 0.0001), IgG + T-5224 and αPD-L1 mAb + T-5224 (χ² = 19.47, P < 0.0001), and αPD-L1 mAb + Vehicle and αPD-L1 mAb + T-5224 (χ² = 18.37, P < 0.0001). Tick marks indicate censored observations. Quantitative data are presented as mean ± SD. Lactate-associated H3K18la promotes RNF166 transcription in HCC cells To investigate the upstream mechanism underlying RNF166 upregulation, we first analyzed publicly available datasets. Lactate treatment increased RNF166 expression in the GSE191252 dataset (Fig. 9a). Analysis of H3K18la ChIP-seq datasets further showed H3K18la enrichment around the RNF166 promoter region, suggesting that RNF166 may be regulated by lactate-associated histone lactylation (Fig. 9b). We next examined this relationship in HCC cells. Treatment of Huh7 and HCCLM3 cells with lactate increased RNF166 protein and mRNA levels over time, accompanied by increased H3K18la and pan-lysine lactylation levels (Fig. 9c, d). Conversely, treatment with oxamate or GSK2837808A reduced RNF166 expression in a dose-dependent manner (Fig. 9e–h). In addition, oxamate, lactate, and rotenone treatment further supported the association between lactate metabolism, histone lactylation, and RNF166 expression (Fig. 9i, j). To further test whether endogenous lactate production contributes to this axis, we knocked down LDHA in Huh7 and HCCLM3 cells. LDHA knockdown reduced H3K18la and pan-lysine lactylation levels, accompanied by decreased RNF166, c-Jun, and PD-L1 expression (Supplementary Fig. S10a, b). In rescue experiments, lactate supplementation restored RNF166 expression under oxamate treatment (Supplementary Fig. S10c, d). Consistently, pharmacological inhibition of LDHA with FX-11 reduced RNF166 expression and downstream c-Jun and PD-L1 levels (Supplementary Fig. S10e, f). Treatment with C646, a histone acetyltransferase/lactylation-related inhibitor, also decreased H3K18la, RNF166, c-Jun, and PD-L1 expression (Supplementary Fig. S10g, h). Finally, ChIP-qPCR was performed to examine H3K18la enrichment at the RNF166 promoter. H3K18la enrichment was detected at the RNF166 promoter in both Huh7 and HCCLM3 cells and was quantified as fold enrichment over the IgG control (Fig. 9k). Lactate treatment further increased H3K18la enrichment at this promoter region when quantified as the percentage of input DNA (Fig. 9l). Together, these data support that lactate-LDHA-dependent H3K18la contributes to RNF166 upregulation and suggest that this mechanism contributes to activation of the RNF166-c-Jun-PD-L1 axis in HCC cells. Fig. 9Lactate-associated H3K18la promotes RNF166 transcription in HCC cells. a RNF166 mRNA expression in control and lactate-treated samples from the GSE191252 dataset. Expression levels are presented as TPM values. Each point represents one biological sample (n = 2 per group; two-tailed unpaired t-test with Welch’s correction, P = 0.0087). b Genome-browser tracks showing H3K18la signals at the RNF166 locus. The RNF166 promoter region analyzed by ChIP-qPCR is highlighted. c, d Western blotting and RT-qPCR analyses of RNF166 expression in Huh7 and HCCLM3 cells after lactate treatment for the indicated durations. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the 0-h group serving as the control. Huh7: all indicated comparisons, P < 0.0001. HCCLM3: 0 vs. 3 h, P = 0.0959; 0 vs. 6 h, P < 0.0001; 0 vs. 9 h, P < 0.0001; 0 vs. 12 h, P < 0.0001. e, f Western blotting and RT-qPCR analyses of RNF166 expression in Huh7 and HCCLM3 cells treated with the indicated concentrations of oxamate. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the 0 group serving as the control. Huh7: 0 vs. 5, P = 0.0464; 0 vs. 10, P = 0.0004; 0 vs. 20, P < 0.0001. HCCLM3: 0 vs. 5, P = 0.0004; 0 vs. 10, P < 0.0001; 0 vs. 20, P < 0.0001. g, h Western blotting and RT-qPCR analyses of RNF166 expression in Huh7 and HCCLM3 cells treated with the indicated concentrations of GSK. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the 0 group serving as the control. Huh7: 0 vs. 5, P < 0.0001; 0 vs. 10, P < 0.0001; 0 vs. 20, P < 0.0001. HCCLM3: 0 vs. 5, P = 0.0118; 0 vs. 10, P = 0.0003; 0 vs. 20, P < 0.0001. i, j Western blotting and RT-qPCR analyses of RNF166 expression in Huh7 and HCCLM3 cells treated with DMSO, oxamate, lactate, or rotenone. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the DMSO group serving as the reference. Huh7: DMSO vs. oxamate, P = 0.0456; DMSO vs. lactate, P < 0.0001; DMSO vs. rotenone, P < 0.0001. HCCLM3: DMSO vs. oxamate, P = 0.0157; DMSO vs. lactate, P < 0.0001; DMSO vs. rotenone, P < 0.0001. k ChIP-qPCR analysis of H3K18la enrichment at the RNF166 promoter in Huh7 and HCCLM3 cells. Enrichment is expressed as fold over the IgG control. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0001; HCCLM3, P = 0.0002). l ChIP-qPCR analysis of H3K18la enrichment at the RNF166 promoter after lactate treatment in Huh7 and HCCLM3 cells. Enrichment is expressed as the percentage of input DNA. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0023; HCCLM3, P = 0.0004). Quantitative data are presented as mean ± SD. Representative Western blots are shown from three independent experiments Lactate-associated H3K18la promotes RNF166 transcription in HCC cells. a RNF166 mRNA expression in control and lactate-treated samples from the GSE191252 dataset. Expression levels are presented as TPM values. Each point represents one biological sample (n = 2 per group; two-tailed unpaired t-test with Welch’s correction, P = 0.0087). b Genome-browser tracks showing H3K18la signals at the RNF166 locus. The RNF166 promoter region analyzed by ChIP-qPCR is highlighted. c, d Western blotting and RT-qPCR analyses of RNF166 expression in Huh7 and HCCLM3 cells after lactate treatment for the indicated durations. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the 0-h group serving as the control. Huh7: all indicated comparisons, P < 0.0001. HCCLM3: 0 vs. 3 h, P = 0.0959; 0 vs. 6 h, P < 0.0001; 0 vs. 9 h, P < 0.0001; 0 vs. 12 h, P < 0.0001. e, f Western blotting and RT-qPCR analyses of RNF166 expression in Huh7 and HCCLM3 cells treated with the indicated concentrations of oxamate. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the 0 group serving as the control. Huh7: 0 vs. 5, P = 0.0464; 0 vs. 10, P = 0.0004; 0 vs. 20, P < 0.0001. HCCLM3: 0 vs. 5, P = 0.0004; 0 vs. 10, P < 0.0001; 0 vs. 20, P < 0.0001. g, h Western blotting and RT-qPCR analyses of RNF166 expression in Huh7 and HCCLM3 cells treated with the indicated concentrations of GSK. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the 0 group serving as the control. Huh7: 0 vs. 5, P < 0.0001; 0 vs. 10, P < 0.0001; 0 vs. 20, P < 0.0001. HCCLM3: 0 vs. 5, P = 0.0118; 0 vs. 10, P = 0.0003; 0 vs. 20, P < 0.0001. i, j Western blotting and RT-qPCR analyses of RNF166 expression in Huh7 and HCCLM3 cells treated with DMSO, oxamate, lactate, or rotenone. qPCR data were obtained from three independent biological experiments (n = 3) and analyzed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test, with the DMSO group serving as the reference. Huh7: DMSO vs. oxamate, P = 0.0456; DMSO vs. lactate, P < 0.0001; DMSO vs. rotenone, P < 0.0001. HCCLM3: DMSO vs. oxamate, P = 0.0157; DMSO vs. lactate, P < 0.0001; DMSO vs. rotenone, P < 0.0001. k ChIP-qPCR analysis of H3K18la enrichment at the RNF166 promoter in Huh7 and HCCLM3 cells. Enrichment is expressed as fold over the IgG control. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0001; HCCLM3, P = 0.0002). l ChIP-qPCR analysis of H3K18la enrichment at the RNF166 promoter after lactate treatment in Huh7 and HCCLM3 cells. Enrichment is expressed as the percentage of input DNA. Data were obtained from three independent biological experiments (n = 3) and analyzed using two-tailed unpaired Student’s t-tests (Huh7, P = 0.0023; HCCLM3, P = 0.0004). Quantitative data are presented as mean ± SD. Representative Western blots are shown from three independent experiments Fig. 10Schematic representation of the lactate–H3K18la–RNF166–c-Jun–PD-L1 signaling axis in HCC immune evasion (by Figdraw) Schematic representation of the lactate–H3K18la–RNF166–c-Jun–PD-L1 signaling axis in HCC immune evasion (by Figdraw) Discussion PD-L1 and PD-1 are the classical tumor immune checkpoints. Although numerous potential new immune checkpoints have been identified in recent years [24, 25], clinical studies have indicated that these novel targets have not yet surpassed the efficacy of PD-L1/PD-1. Moreover, PD-L1/PD-1 checkpoint inhibitors continue to play a major role in clinical practice. Therefore, elucidating the regulatory mechanisms of PD-L1 and developing rational combination strategies remain important. Previous studies have shown that PD-L1 can be directly regulated through ubiquitin-mediated modifications; for example, CSN5 [26] and USP22 [27] deubiquitinate and stabilize PD-L1, whereas SPOP mediates its degradation through ubiquitination [28]. Beyond direct regulation, ubiquitination can also modulate PD-L1 indirectly, such as SPOP regulating PD-L1 expression through the degradation of IRF1 [29], or TRAF6 acting through the YAP1/TFCP2 pathway [30]. In this study, we found that RNF166 indirectly regulates PD-L1 transcription by stabilizing c-Jun through ubiquitination, further expanding our understanding of indirect ubiquitin-mediated regulation of PD-L1. Although RNF166 has been reported to promote tumor cell growth in colorectal cancer, our study demonstrated that RNF166 does not significantly affect tumor cell proliferation or apoptosis. Instead, public database analyses indicated that RNF166 is closely associated with immune-related pathways. Similarly, previous studies have shown that PRDM1 suppresses tumor-cell-intrinsic growth yet promotes immune evasion through PD-L1 upregulation [31]. In vivo experiments revealed that RNF166 alters the distribution and cytotoxic function of tumor-infiltrating T cells by upregulating PD-L1. In immunodeficient BALB/c nude and NOD-SCID mice, RNF166 overexpression did not markedly promote tumor growth, supporting an immune-dependent component of its pro-tumor effect. Notably, PD-L1 knockdown restored T-cell infiltration and effector function, indicating that RNF166 exerts its pro-tumorigenic effects, at least in part, through PD-L1. Clinical TMA analyses further showed that RNF166 expression was positively correlated with c-Jun and PD-L1 expression and negatively correlated with CD8 and GZMB expression. These findings support the clinical relevance of the RNF166-c-Jun-PD-L1 axis, although exploratory survival analyses did not support RNF166 as a standalone prognostic biomarker. Future studies are warranted to determine whether RNF166 regulates additional tumor-associated signaling pathways, thereby influencing other immune cell populations, or whether it functions in other cell types to modulate the tumor immune microenvironment. Mechanistic studies revealed that RNF166 stabilizes c-Jun by promoting K63-linked ubiquitination at K50/K56, which was associated with reduced K48-linked ubiquitination. The catalytically inactive RNF166-C33A/C36A mutant failed to efficiently regulate K48- and K63-linked ubiquitination of c-Jun or restore c-Jun and PD-L1 expression, supporting a requirement for RNF166 E3 ligase activity. Endogenous c-Jun ubiquitination assays further confirmed that RNF166 knockdown increased K48-linked ubiquitination and reduced K63-linked ubiquitination of c-Jun. These results support a model in which RNF166 remodels the ubiquitin linkage pattern of c-Jun, rather than implying direct enzymatic removal of K48-linked chains. The current data do not distinguish whether the reduction in K48-linked ubiquitination results from competition at c-Jun lysine residues, steric interference with K48-linked chain formation, altered recruitment or activity of deubiquitinases or other ubiquitin-regulatory enzymes, or additional indirect mechanisms. K63-linked ubiquitination is the second most prevalent type of ubiquitination after K48-linked ubiquitination; however, its biological functions remain incompletely understood. K63-linked ubiquitination can mediate substrate degradation through autophagy. For example, TRIM32 catalyzes K63-linked ubiquitination at GPX4 K107, promoting p62-mediated autophagic degradation [32]. Concomitantly, K63-linked ubiquitination can also stabilize protein expression; for instance, TRIM26 stabilizes GPX4 by competing with K48-linked ubiquitination [33], and FBXL8 stabilizes IκBα through K63-linked ubiquitination [34]. These studies highlight the diverse roles of K63-linked ubiquitination in protein regulation. The therapeutic experiments further support the relevance of this pathway. In RNF166-overexpressing tumors, either T-5224 or anti-PD-L1 mAb monotherapy reduced tumor growth, whereas combined c-Jun inhibition and anti-PD-L1 blockade produced stronger tumor suppression and prolonged survival. The vector-control treatment experiments further showed that this regimen also exerted antitumor activity in tumors without RNF166 overexpression, suggesting that c-Jun inhibition and anti-PD-L1 blockade may have broader therapeutic value. Together, these findings support the therapeutic relevance of the RNF166-c-Jun-PD-L1 axis and provide a rationale for combined targeting of c-Jun and PD-L1 in HCC. From a translational perspective, these findings suggest several potential strategies. First, RNF166 expression, together with c-Jun and PD-L1 status, may help identify tumors in which the RNF166-c-Jun-PD-L1 axis is activated and may therefore inform patient stratification for combination immunotherapy. Second, pharmacological inhibition of c-Jun may be explored as a strategy to enhance anti-PD-1/PD-L1 therapy in HCC, particularly in tumors with high RNF166 expression and active PD-L1 signaling. Third, because RNF166 is an E3 ubiquitin ligase, future development of RNF166-targeted approaches, such as small-molecule inhibitors or targeted protein-degradation strategies, may provide a more upstream intervention. However, these translational applications will require prospective validation in larger HCC cohorts, clinically relevant treatment models, and safety studies before clinical implementation. The role of lactylation in tumor immunity remains largely unexplored. Previous studies have shown that lactate can induce GCN5 expression, thereby promoting PD-L1 lactylation and stabilizing its protein levels [23]. In addition, LDHA-mediated H4K5 lactylation in bladder cancer upregulates PD-L1 expression, thereby facilitating tumor immune evasion [35]. In our study, we found that lactate promoted H3K18 lactylation, which increased RNF166 expression, and subsequently upregulated PD-L1 expression through an indirect mechanism. LDHA knockdown, LDHA inhibition, oxamate treatment, and C646 treatment reduced RNF166 expression and downstream c-Jun and PD-L1 levels. Lactate supplementation restored RNF166 expression under oxamate treatment, and ChIP-qPCR confirmed increased H3K18la enrichment at the RNF166 promoter. These findings support that lactate-LDHA-dependent H3K18la contributes to RNF166 upregulation, while acknowledging that broader metabolic and chromatin-regulatory effects may also participate. In summary, our study revealed that RNF166 promotes tumor immune evasion in HCC. RNF166 upregulates PD-L1 expression, thereby enhancing the ability of liver cancer cells to escape CD8 + T cell–mediated cytotoxicity. Moreover, we found that RNF166 remodels c-Jun ubiquitination in an E3 ligase activity-dependent manner. Finally, lactate-associated H3K18la contributes to RNF166 transcriptional upregulation, providing a mechanistic explanation for its aberrant elevation in tumors (Fig. 10). These findings delineate a lactate-H3K18la-RNF166-c-Jun-PD-L1 axis and provide a mechanistic rationale for combining c-Jun inhibition with anti-PD-L1 blockade in HCC. Methods Ethics statement This study complied with all the relevant ethical regulations. All animal experiments were approved by the Institutional Animal Care and Use Committee of Xi’an Jiaotong University. The use of patient samples was approved by the Institutional Research Ethics Committee of First Affiliated Hospital of Xi’an Jiaotong University, which was in accordance with the Declaration of Helsinki. These samples were obtained from the Biobank of the First Affiliated Hospital of Xi’an Jiaotong University. Clinical samples A total of 76 paired HCC tissues and adjacent non-tumor tissues were obtained from the Biobank of the First Affiliated Hospital of Xi’an Jiaotong University. All patients were pathologically diagnosed with HCC, had not received chemotherapy, radiotherapy, or immunotherapy prior to surgery, and had complete clinicopathological information. These paired samples were used for immunohistochemical analysis of RNF166 expression. Tissue microarray sections from the same cohort were further used to evaluate c-Jun, PD-L1, CD8, and GZMB expression for correlation analyses. This study was approved by the Ethics Committee of the First Affiliated Hospital of Xi’an Jiaotong University, and written informed consent was obtained from all patients prior to surgery. Animal experiments Four-week-old male BALB/c nude mice and NOD-SCID mice were used as immunodeficient tumor models. Four-week-old male C57BL/6 mice were used as immunocompetent tumor models. RNF166-overexpressing Hepa1-6 cells and the corresponding vector-control cells were generated through lentiviral transduction and used for subsequent in vivo experiments. For subcutaneous tumor models, mice were randomly assigned to groups (n = 5 per group unless otherwise indicated). A total of 2.5 × 106 Hepa1-6 cells suspended in 100 µL phosphate-buffered saline (PBS) were subcutaneously injected into the left flank of each mouse. Tumor size was measured every three days using calipers, and tumor volume was calculated using the following formula: (length × width2) / 2. At the indicated experimental endpoints, mice were sacrificed, and tumors were excised, weighed, photographed, and processed for subsequent analyses. For orthotopic liver tumor models, mice were randomly divided into groups (n = 5 per group). A total of 1 × 106 Hepa1-6 cells suspended in 50 µL PBS were injected into the liver of each mouse. Two weeks after tumor implantation, mice were sacrificed, and tumor tissues were collected, weighed, and processed for immunohistochemical analysis. For anti-PD-L1 treatment experiments, mice were treated with anti-mouse PD-L1 mAb (B7-H1; Selleck, A2115) or the corresponding rat IgG2b isotype control antibody (Selleck, A2116) by intraperitoneal injection at 100 µg per mouse every three days, according to the indicated schedules. For CD8 + T-cell depletion experiments, anti-mouse CD8α mAb (Selleck, A2102) or control IgG was administered by intraperitoneal injection at 100 µg per mouse every three days. For combination therapy experiments, tumor-bearing mice were treated with T-5224, anti-PD-L1 mAb, or their combination. T-5224 was administered by oral gavage at 100 mg/kg every three days. Anti-PD-L1 mAb or control IgG was administered by intraperitoneal injection at 100 µg per mouse every three days. The vehicle for T-5224 was 10% DMSO and 90% corn oil. Treatment schedules are shown in the corresponding figures. For survival experiments, survival time was calculated from the day of treatment initiation. Mice were followed for up to 60 days after treatment initiation or until they reached predefined humane endpoints. Humane endpoints included tumor volume reaching 1500 mm3, maximum tumor diameter reaching 20 mm, tumor ulceration, bleeding, infection, necrosis, obvious limitation of movement, inability to eat or drink normally, body weight loss of 20% or more, body condition score of 2 or lower, or other signs requiring euthanasia according to institutional animal welfare guidelines. Mice that did not reach humane endpoints by the end of follow-up were censored. Immunohistochemistry (IHC) Formalin-fixed tissue samples were embedded in paraffin and sectioned for staining. After deparaffinization, the sections were rehydrated through a graded ethanol series, followed by antigen retrieval using citrate buffer (pH 6.0). Endogenous peroxidase activity was blocked with 3% hydrogen peroxide and the sections were blocked with 5% BSA. The sections were incubated overnight with primary antibodies against RNF166 (Abclonal, A8276, 1:100), Ki67 (Abclonal, A20018, 1:200), PD-L1 (Abclonal, A27937, 1:200), CD8α (Abclonal, A23305PM, 1:300), c-Jun (Proteintech, 24909-1-AP, 1:500), and GZMB (Abclonal, A2557, 1:200). The following day, sections were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies. Immunoreactivity was visualized using DAB, and nuclei were counterstained with hematoxylin. Images were captured under a microscope and evaluated. For tissue microarray analysis, RNF166, c-Jun, PD-L1, CD8, and GZMB staining were evaluated using semi-quantitative IHC scores. Correlations among RNF166, c-Jun, PD-L1, CD8, and GZMB IHC scores were analyzed using Spearman correlation analysis. Flow cytometry Tumor tissues from mice were minced and digested at 37 °C in a digestion buffer containing collagenase IV and DNase I to prepare single-cell suspensions. After washing with PBS, the cells were stimulated for 4 h in culture medium containing PMA, Ionomycin, and Brefeldin A. Cells were then incubated with CD45-APC-Cy7, CD8α-FITC, and CD4-Pacific Blue at 4 °C in the dark for 30 min for surface staining. After staining, the cells were fixed and permeabilized using a fixation/permeabilization buffer, followed by washing with permeabilization/wash buffer. Subsequently, cells were incubated with APC-IFN-γ, PE-TNF-α, and GZMB-PE-Cy7 for intracellular staining at 4 °C in the dark for 30 min. Data acquisition was performed using a CytoFLEX flow cytometer (Beckman Coulter) and data were analyzed using FlowJo software. Immunoprecipitation (IP) assay Cells were lysed on ice for 15 min using IP lysis buffer, followed by centrifugation at 13,000 rpm for 10 min. 10% of the supernatant was collected as the input control for protein quantification. The remaining lysates were incubated overnight at 4 °C with the indicated primary antibodies or an equivalent amount of the corresponding IgG with gentle rotation. The next day, 20 µL of Protein A/G agarose beads were added to the lysates and incubated at 4 °C for an additional 1 h with rotation. The beads were then washed three times with IP lysis buffer. Finally, an equal volume of 2× loading buffer was added, and the samples were boiled at 100 °C for 10 min before western blot analysis. Ubiquitination assay Cells transfected with the indicated plasmids were treated with 10 µM MG132 for 6 h before lysis. Cells were then lysed on ice using IP lysis buffer. After centrifugation, the supernatants were collected and incubated overnight at 4 °C with the indicated antibody with gentle rotation. For exogenous ubiquitination assays, V5-c-Jun was immunoprecipitated, and ubiquitinated c-Jun was detected using anti-HA antibody after co-transfection with HA-Ub or the indicated HA-Ub mutants. For linkage-specific ubiquitination assays, K48-linked and K63-linked ubiquitination of c-Jun was examined using HA-Ub K48/K63 constructs or linkage-specific ubiquitin antibodies, as indicated. For endogenous ubiquitination assays, endogenous c-Jun was immunoprecipitated from Huh7 and HCCLM3 cells after RNF166 knockdown, and K48-linked or K63-linked ubiquitinated c-Jun was detected using linkage-specific ubiquitin antibodies. The RNF166-C33A/C36A mutant was used to assess the requirement for RNF166 catalytic activity. The following day, Protein A/G agarose beads were added to the lysates and incubated at 4 °C for 1 h with rotation. The beads were washed three times with IP lysis buffer. Finally, an equal volume of 2× loading buffer was added, and the samples were boiled at 100 °C for 10 min before Western blotting analysis. For densitometric quantification, ubiquitin smear signals were quantified over the same molecular-weight range across groups and normalized to the corresponding immunoprecipitated c-Jun signal. ChIP-qPCR Cells were cross-linked with formaldehyde at a final concentration of 1% at room temperature, and the reaction was terminated by adding glycine. The cells were resuspended in ChIP lysis buffer containing SDS and lysed on ice, followed by sonication to shear chromatin. After centrifugation, the supernatant was collected, and 1% of the lysate was reserved as the input control. The remaining samples were diluted and incubated overnight at 4 °C with 2 µg of the indicated primary antibody or IgG control. Anti-H3K18la antibody was used to examine H3K18la enrichment at the RNF166 promoter, and anti-c-Jun antibody was used to examine c-Jun enrichment at predicted binding sites in the PD-L1 promoter. The next day, Protein A/G magnetic beads were added and incubated. The beads were sequentially washed with low-salt, high-salt, LiCl, and TE wash buffers. Chromatin was eluted and incubated at 65 °C overnight to reverse crosslinking. After treatment with RNase A and Proteinase K, DNA was purified using a DNA purification kit and subjected to qPCR. ChIP-qPCR enrichment was calculated as fold enrichment over IgG or as the percentage of input DNA, as indicated in the corresponding figure legends. For the RNF166 promoter ChIP-qPCR assay, the primer sequences were as follows: forward, 5’-CAGTGCTGTCCTTATTCGGAGAGA-3’; reverse, 5’-CATAAAGGCCGCTTTTCAGTCTCC-3’. The amplicon was located at chr16:88706830–88,706,919 (GRCh38/hg38), with a size of 90 bp and a position of -511 to -422 bp relative to the transcription start site. Public H3K18la ChIP-seq peak annotation at the RNF166 promoter is provided in Supplementary Table 1. Dual-luciferase reporter assay When cell confluence reached approximately 70%, cells were co-transfected using PEI with the firefly luciferase reporter plasmid containing the indicated PD-L1 promoter sequence and the Renilla luciferase plasmid as an internal control. Wild-type PD-L1 promoter reporter constructs and promoter mutant constructs containing mutations in predicted c-Jun-binding sites were used as indicated. For c-Jun gain-of-function assays, cells were co-transfected with pCMV-c-Jun or the corresponding control vector. For RNF166-dependent promoter assays, cells were co-transfected with the RNF166 expression plasmid together with si-c-Jun or si-Control, as indicated. After 48 h of transfection, cells were washed with PBS and lysed in passive lysis buffer at room temperature for 15 min. Firefly and Renilla luciferase activities were measured using a dual-luciferase reporter assay kit. Relative luciferase activity was calculated as the ratio of firefly luciferase activity to Renilla luciferase activity and normalized to the indicated control group. Plasmids and Lentiviral Production The RNF166 expression construct was generated with an N-terminal FLAG tag, and the c-Jun expression construct was generated with a V5 tag. These plasmids were cloned into the pcDNA3.1 backbone and used for ubiquitination assays. The catalytically inactive RNF166-C33A/C36A mutant and the c-Jun-K50R/K56R double mutant (c-Jun-2KR) were generated by site-directed mutagenesis and verified by DNA sequencing. For gene knockdown experiments, shRNA sequences targeting RNF166, PD-L1, and JUN were inserted into the pLKO.1 vector. For gene overexpression studies, the full-length coding sequences of RNF166 and c-Jun were cloned into the pCDH lentiviral expression vector. For luciferase reporter assays, siRNAs targeting JUN and the corresponding siRNA control were used as indicated. All constructs were verified by DNA sequencing. For lentiviral production, transfer plasmids containing the target sequences were co-transfected into 293T cells together with the packaging plasmids psPAX2 and pMD2.G. Viral supernatants were collected 48 h after transfection and used to infect target cells in the presence of polybrene. To establish stable cell lines, appropriate concentrations of antibiotics were added to the culture medium for selection 48 h after infection. Western blotting and RT-qPCR Cells or tumor tissues were lysed using RIPA lysis buffer supplemented with protease and phosphatase inhibitors. Protein concentrations were determined using a BCA protein assay kit. Equal amounts of protein were separated by SDS-PAGE and transferred onto PVDF membranes. After blocking with 5% non-fat milk, the membranes were incubated overnight at 4 °C with the indicated primary antibodies, followed by incubation with HRP-conjugated secondary antibodies. Protein bands were visualized using enhanced chemiluminescence reagents. GAPDH or β-actin was used as the loading control unless otherwise indicated. For ubiquitination assays, ubiquitin smear signals were normalized to the corresponding immunoprecipitated substrate signal. Total RNA was extracted from cells using TRIzol reagent according to the manufacturer’s instructions. Complementary DNA was synthesized using a reverse transcription kit, and RT-qPCR was performed using SYBR Green master mix. Relative mRNA expression was calculated using the 2^−ΔΔCt method and normalized to the indicated internal control gene. PCR array analysis Total RNA was extracted from RNF166-overexpressing Hep3B cells and the corresponding vector-control cells according to the manufacturer’s instructions. RNA concentration and purity were assessed using a spectrophotometer, and equal amounts of total RNA were reverse-transcribed into cDNA. The expression profiles of inflammation- and immunity-related genes were examined using the Human Cancer Inflammation and Immunity Crosstalk qPCR Array (Wcgene Biotechnology, WC-MRNA0021-H) according to the manufacturer’s protocol. Quantitative PCR was performed using a real-time PCR system. Three independent biological replicates were analyzed for each group. Gene-expression levels were normalized to the housekeeping gene(s) included in the array and calculated using the 2^-ΔΔCt method, with the vector-control group used as the reference. For heatmap visualization, normalized expression values were transformed to row Z-scores for each gene across all samples. For ranked fold-change analysis, log2 fold changes were calculated by comparing the RNF166-overexpressing group with the vector-control group. Selected immune-related genes were further validated by RT-qPCR in vector-control, RNF166-overexpressing, and RNF166-overexpressing plus c-Jun-knockdown cells. Immunofluorescence and multiplex immunofluorescence For immunofluorescence staining, cells or tumor tissue sections were fixed, permeabilized, blocked, and incubated with the indicated primary antibodies overnight at 4 °C, followed by fluorescence-conjugated secondary antibodies. Nuclei were counterstained with DAPI. Images were acquired using a fluorescence or confocal microscope. For multiplex immunofluorescence, tumor tissue sections were stained with the indicated antibodies according to the manufacturer’s protocol. Fluorescence intensity or positive areas were quantified as indicated in the figure legends. Scale bars are shown in the corresponding images. Statistical analysis All quantitative data are presented as mean ± standard deviation (SD) or standard error of the mean (SEM), as indicated in the figure legends. Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). Comparisons between two groups were performed using two-tailed unpaired Student’s t-test, paired Student’s t-test, or Welch’s t-test, as appropriate. Comparisons among multiple groups were performed using one-way ANOVA followed by Dunnett’s or Tukey’s multiple comparisons test, or Brown-Forsythe and Welch ANOVA followed by Dunnett’s T3 multiple comparisons test, as indicated. Tumor growth curves and grouped ChIP-qPCR analyses were analyzed using two-way ANOVA or repeated-measures ANOVA followed by Šídák’s multiple comparisons test, as indicated. Correlations between IHC scores were evaluated using Spearman correlation analysis. Survival curves were analyzed using the Kaplan-Meier method and compared using the log-rank (Mantel-Cox) test. A P value < 0.05 was considered statistically significant. Statistical significance is indicated as *P < 0.05, **P < 0.01, and ***P < 0.001; ns indicates no significant difference. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1 Supplementary Material 2 Supplementary Material 2 Supplementary Material 3 Supplementary Material 3