USP26 as a hepatitis B virus-induced deubiquitinase primes hepatocellular carcinogenesis by epigenetic remodeling

Abstract

Despite recent advances in systemic therapy for hepatocellular carcinoma (HCC), the prognosis of hepatitis B virus (HBV)-induced HCC patients remains poor. By screening a sgRNA library targeting human deubiquitinases, we find that ubiquitin-specific peptidase 26 (USP26) deficiency impairs HBV-positive HCC cell proliferation. Genetically engineered murine models with Usp26 knockout confirm that Usp26 drives HCC tumorigenesis. Mechanistically, we find that the HBV-encoded protein HBx binds to the promoter and induces the production of USP26, which is an X-linked gene exclusively expressed in the testis. HBx consequently promotes the association of USP26 with SIRT1 to synergistically stabilize SIRT1 by deubiquitination, which promotes cell proliferation and impedes cell apoptosis to accelerate HCC tumorigenesis. In patients with HBV-positive HCC, USP26 is robustly induced, and its levels correlate with SIRT1 levels and poor prognosis. Collectively, our study highlights a causative link between HBV infection, deubiquitinase induction and development of HCC, identifying a druggable target, USP26.

Fig. 1. Identification of USP26 as a potential regulator in HCC cells through CRISPR-based DUB library screening.

a Schematic strategy for CRISPR-based DUB library screening. A total of 348 sgRNAs targeting 87 DUBs (4 sgRNAs/DUB) were transfected into the HCCLM3 cell line, which was subsequently screened. Left: Dot plots showing the rank of DUBs from 1 (lowest) to 87 (highest) based on colony formation assay (b) and EdU staining assay (c) of sg-DUBs HCCLM3 cells, which were measured and calculated by log₂^{fold change}. The top DUBs with inhibitory effects on cell growth were listed as indicated. Right: The percentage of significantly changed DUBs in each group (p < 0.01, log₂ ^{fold change} < −0.58, down group; p < 0.01, log₂^{fold change} > 0.26, up group; p > 0.05, −0.58 <log₂ ^{fold change} < 0.26, no difference group (No-diff)). d Venn diagram showing the overlapping DUBs that inhibited cell survival in the two screening assays described in b and c. e, f Western blot analysis and colony formation assay in HCCLM3 cells with or without CRISPR/Cas9-mediated USP26 knockout. g Flow cytometry analysis showing the percentages of cells in S phase in the sg-Control and USP26 knockout groups. h. Representative immunoblots showing USP26 protein levels in tumors and paired normal tissues from HBV-positive (HBV⁺) HCC patients. N, normal tissue; T, tumor tissue. Quantification of relative USP26 protein levels in tumors and paired normal tissues from HBV⁺ HCC patients (i) and HBV-negative (HBV⁻) HCC patients (j). k Kaplan-Meier survival analysis of HCC patients with high or low USP26 expression. The data were obtained from the Kaplan-Meier plotter database. Each graph is presented as the mean ± SEM. b, c n =  3 biologically independent experiments; (e–h) n =  3 biologically independent experiments; n = 42 (i) and n = 8 (j) paired tumor and normal samples. Statistical significance was calculated by (b–d) one-way ANOVA; (e, g) two-tailed unpaired t-test; (i, j) two-tailed paired t-test; (k) log-rank test. Source data are provided as a Source Data file.

Fig. 2. Loss of USP26 inhibits cell proliferation and promotes cell apoptosis in vitro.

Western blot analysis (a), MTT (b), colony formation (c) and EdU staining assays (d, e) of USP26-knockdown HCCLM3 cells with or without USP26 (shRNA-resistant USP26) restoration. Scale bar: 100 μm. Western blot analysis (f), MTT (g), colony formation (h) and EdU staining assays (i, j) of USP26-knockdown MHCC97H cells with or without USP26 (shRNA-resistant USP26) restoration. Scale bar: 100 μm. k–m. The apoptosis rate of MHCC97H cells as described in f was measured by flow cytometry after Annexin/V and PI dual staining. n Cleaved PARP (C-PARP) expression in MHCC97H cells as described in f was measured by western blotting. The red stars (*) in a, f and n indicate the positions of USP26. Each graph is presented as the mean ± SEM. Each blot data is representative of three independent experiments. (b, g) n = 4 independent experiments; (c, d, h, i, k) n = 3 independent experiments. Statistical significance was calculated by (b, g) two-way ANOVA; (c, e, h, j, l, m) one-way ANOVA. Source data are provided as a Source Data file.

Fig. 3. USP26 deficiency blocks hepatocellular carcinogenesis in mice.

a Schematic illustration of hepatoblastoma model by hydrodynamic injection (HDI) of ΔN90-β-catenin and YAP-S127A plasmids. IR/DR refers to inverted repeats and direct repeats. Images of liver tumors (b), liver weight (c) and the ratio of liver weight to body weight (d) from Usp26^+/Y and Usp26^−/Y mice generated in a. Scale bar: 10 mm. Representative images and quantitative results of H&E staining (e, f), Ki67 staining (g, h) and TUNEL-positive cells (i, j) in tumor nodules from Usp26^+/Y and Usp26^−/Y mouse liver sections. The areas circled by black dashed lines in e point to tumor nodules. N normal tissue, T tumor. Scale bar: 200 μm (e), 50 μm (g, i). k Kaplan‒Meier curves of the HDI hepatoblastoma model. l Schematic illustration of the HCC model with HDI of oncogenic β-catenin-S33/37 A and MYC-luciferase plasmids. IRES indicates the internal ribosome entry site. m, n Representative images and quantitative bioluminescence intensity of liver tumors in l. Representative liver images (o), liver-to-body ratios (p), H&E staining (q, r) and Ki67 staining (s, t) of the tumor nodules in Usp26^+/Y and Usp26^−/Y mouse liver sections. Scale bar: 10 mm (o), 200 μm (q), 50 μm (s). u. DNA sequencing was used to validate the gene editing of USP26 in single cell-derived knockout clones. Schematic representation with various insertions (red) and guide sequences (blue) are marked. v–x Representative images, volumes, and weights of xenograft tumors derived from USP26 knockout (KO) HCCLM3 cells with or without USP26 restoration. Scale bar: 10 mm. Each graph is presented as the mean ± SEM. c, d n = 10 in Usp26^+/Y group, 9 in Usp26^−/Y group; (f, h) n = 6 per group; (k) n = 11 in Usp26^+/Y group, 7 in Usp26^−/Y group; (n) n = 9 in Usp26^+/Y group, 6 in Usp26^−/Y group; (j, p, r, t, w, x) n = 5 per group. Statistical significance was calculated by (c, d, f, h, j, n, p, r, t, x) two-tailed unpaired t-test; (k) log-rank test; (w) two-way ANOVA. Source data are provided as a Source Data file.

Fig. 4. USP26 directly interacts with SIRT1.

a A Co-IP assay combined with mass spectrometry analysis was performed to screen for USP26-interacting proteins. A partial list of accession numbers and the matched peptides of candidate proteins is shown. b HEK293T cells were co-transfected with SFB-USP26 and Myc-SIRT1, and then Co-IP was performed using S-protein beads. c The interaction between endogenous USP26 and SIRT1 was measured by Co-IP. Lysates of HCCLM3 cells were immunoprecipitated with an anti-USP26 antibody or nonspecific IgG, and the associated SIRT1 was detected by immunoblotting. d. A GST pull-down assay was carried out to detect the interaction between USP26-His and GST-SIRT1. e His pull-down assay was carried out to detect the interaction between USP26-His and GST-SIRT1. f U2OS cells were transfected with the GFP-USP26 plasmid. The intracellular localization of USP26 (green) and endogenous SIRT1 (red) was visualized by using a laser confocal microscope. Scale bar: 10 μm. g Mapping the domain structure of SIRT1 that interacts with USP26. Full-length SIRT1 and its various deletion mutants together with USP26 were co-expressed in HEK293T cells. The binding regions between USP26 and SIRT1 were detected by Co-IP. A schematic diagram is shown above the panel. h The domain structure of USP26, which interacts with SIRT1, was measured by Co-IP. The binding regions were detected by Co-IP. A schematic diagram is shown above the panel. The red star (*) in c indicates the position of SIRT1. Data in (b–h) are representative from three independent experiments. Source data are provided as a Source Data file.

Fig. 5. USP26 deubiquitinates and stabilizes SIRT1.

a The relative protein levels of SIRT1 and USP26 in different HCC cell lines. b The relative protein levels of SIRT1 in Huh7 cells transfected with the indicated doses of USP26. c. U2OS cells stably expressing GFP-USP26 were immunostained for endogenous SIRT1. Scale bar: 10 μm. The relative fluorescence intensity of SIRT1 in cells with or without GFP-USP26 expression was quantified in the same visual fields. d The protein levels of SIRT1 in USP26 knockout (KO) HCCLM3 cells. e The SIRT1 protein levels in liver tumors from Usp26^+/Y and Usp26^−/Y mice generated through the strategy described in Fig. 3a. f The protein levels of SIRT1 in USP26-knockdown and re-expressed shRNA-resistant USP26 MHCC97H cells. g The half-life of SIRT1 was measured by adding CHX to HEK293T cells transfected with wild-type (WT) USP26 or mutant USP26 (C304S). h The half-life of SIRT1 in USP26-knockdown and re-expressed shRNA-resistant USP26 MHCC97H cells was measured after treatment with CHX. i. HCCLM3 cells stably expressing SIRT1 were transfected with USP26 and then treated with MG132 (10 μM). Endogenous polyubiquitinated SIRT1 was detected by Western blotting. j, k USP26 KO HEK293T cells were co-transfected with HA-Ub, Myc-SIRT1, USP26 or mutant USP26 as indicated and then treated with MG132. Finally, the cells were lysed and subjected to a ubiquitination assay. l, m Ubiquitinated Myc-SIRT1 was purified from HEK293T cells, followed by incubation with bacteria-purified WT or mutant USP26 for an exogenous ubiquitination assay. The reaction buffer was subjected to immunoblotting or stained with Coomassie blue. n–q Similar to j, K29 and K48 ubiquitination assays were carried out. The red stars (*) in f indicate the positions of USP26. Each graph presents the mean ± SEM. Each blot data is representative of three independent experiments. Data in (c) n = 42 or 48 cells for WT or USP26 overexpressing group from three independent experiments. Data in (e) n = 7 mice per group. Statistical significance was calculated by (b, d, k, m, o, q) one-way ANOVA; (c) two-tailed unpaired t-test; (g, h) two-way ANOVA. Source data are provided as a Source Data file.

Fig. 6. USP26 promotes HCC tumorigenesis by stabilizing SIRT1.

Western blot analysis (a), MTT (b) and colony formation (c) assays of USP26-knockdown MHCC97H cells with or without shRNA-resistant USP26 or SIRT1 overexpression. Western blot analysis (d), MTT (e) and colony formation (f) assays of USP26-knockdown HCCLM3 cells with or without USP26 (shRNA-resistant USP26) or SIRT1 overexpression. g. An EdU staining assay was carried out on USP26-knockdown and SIRT1 restored Huh7 and HCCLM3 cells. Flow cytometry was carried out to screen the active DNA replicated cells. h The apoptosis rates of USP26-knockdown and SIRT1 restored MHCC97H cells were measured by flow cytometry after Annexin/V/PI dual staining. i Immunoblotting of cleaved PARP in USP26-knockdown SIRT1restored MHCC97H cells. j Immunoblotting of the indicated proteins in USP26-knockdown MHCC97H cells with or without shRNA-resistant USP26 or SIRT1 overexpression. k The relative p66Shc mRNA levels in USP26-knockdown MHCC97H cells with or without shRNA-resistant USP26 or SIRT1 were detected by RT‒qPCR. l Representative images of xenograft tumors derived from USP26-knockout (KO) HCCLM3 cells with or without USP26 or SIRT1 overexpression. Scale bar: 10 mm. m The volumes of tumors burdened in nude mice receiving the HCCLM3 cells were measured on different days after implantation. n Weights of the mouse xenograft tumors. Each graph is presented as the mean ± SEM. Each blot data is representative of three independent experiments. b, e n = 4 independent experiments; c, f, g, h, k n = 3 independent experiments; (m, n) n = 4 mice per group. Statistical significance was calculated by (b, e, m) two-way ANOVA; (c, f, g, h, k) one-way ANOVA; (n) two-tailed unpaired t-test. Source data are provided as a Source Data file.

Fig. 7. HBx transcriptionally promotes USP26 expression and enhances USP26-mediated deubiquitination of SIRT1.

a Left: Representative immunoblots of USP26 and SIRT1 protein levels in tumor and paired normal tissues of HBV-positive (HBV⁺) HCC patients. N normal tissue; T tumor. Right: Spearman correlation analysis between USP26 and SIRT1 protein expression in HBV⁺ HCC samples. b Quantification of relative USP26 and SIRT1 expression in HBV-negative (HBV⁻) and HBV⁺ HCC cell lines as indicated in Fig. 5a. c The endogenous protein levels of USP26 in Huh7 and HepG2 cells transfected with HBx. The relative mRNA levels of SIRT1 and USP26 in Huh7 (d) and L02 (e) cells transfected with HBx. f The protein levels of USP26 and SIRT1 in USP26-knockdown Huh7 cells transfected with HBx. g USP26 promoter activity was measured in Huh7 cells transfected with HBx, HBc or HBs. h USP26 promoter activity was measured in Huh7 cells transfected with HBx. i, j A series of USP26 promoter truncations were constructed to identify the regulatory region mediated by HBx. k, l In vivo and in vitro interaction between biotinylated USP26 promoter and HBx were detected by DNA pull-down assay. m EMSA assay was used to examine binding between HBx and the −450 to +100 region of the USP26 promoter. When the band of the free probe became weaker, it denoted strong binding. n The interaction between Myc-USP26 and HBx-Flag was detected in HEK293T cells. o The interaction between SFB-USP26 and Myc-SIRT1 was detected in the presence of HBx-Flag in HEK293T cells. p USP26-KO HEK293T cells were co-transfected with HA-Ub and Myc-SIRT1 with or without SFB-USP26 or SFB-HBx. The ubiquitination of SIRT1 was detected by Western blotting. Each graph presents the mean ± SEM. Data in (a) n = 41 human HBV⁺ HCC samples. Each blot data is representative of three independent experiments. Data in (d, e, g–j) n = 3 independent experiments. Statistical significance was calculated by (a) Spearman’s rank correlation coefficient, analysis of two-tailed t-test; (d, e, h) two-tailed unpaired t-test; (f, g, p) one-way ANOVA; (i, j) two-way ANOVA. Source data are provided as a Source Data file.

Fig. 8. Schematic of HBx-induced elevation of USP26 involved in hepatocellular carcinoma by stabilizing SIRT1.

HBx acts as a transcriptional coactivator to recruit β-catenin and subsequently promote the transcriptional activation of USP26. Then, HBx promotes the interaction between induced USP26 and SIRT1, enhancing the stability of the SIRT1 protein and promoting the progression of HCC.