HBV X protein regulates cancer stemness and tumor invasiveness through SENP1 in hepatocellular carcinoma.

To evaluate the clinical relevance of SENP1 in HCC, SENP1 mRNA levels were assessed in paired tumor (T) and adjacent peritumor (PT) tissues from 211 patients. Levels of SENP1 mRNA were classified as high (T/PT ≥2) or low (T/PT < 2) ( Table 1 ). Patients with high SENP1 levels exhibited a poorer OS (median survival: 69.2 vs. 172.8 months, p < 0.001) and shorter DFS (median survival: 15.8 vs. 39.7 months, p = 0.004) ( Fig. 1 A), and high SENP1 expression was significantly associated with advanced TNM stages ( Fig. 1 B, Table 1 ). Elevated SENP1 expression was further validated in the GSE76427 cohort, 17 showing significantly higher SENP1 mRNA levels in tumor tissues vs . adjacent non-tumor tissues ( Fig. 1 C). Additionally, a significant and positive correlation was observed between SENP1 expression and cancer stemness, which was also indicated by results of a gene set-enrichment analysis targeting HCC recurrence and embryonic stem cell (ESC) signatures ( Fig. 1 D). Consistently, SENP1 mRNA levels were correlated with stemness/EMT markers ( OCT4 , SNAIL , TWIST ) in tissues from 211 patients with HCC we further examined ( Fig. 1 E). These clinical observations highlight the potential role of SENP1 in HCC tumor recurrence and metastasis. Table 1 Variables associated with high and low SENP1 expression in HCC. Table 1 SENP1 High a (n = 40) SENP1 Low a (n = 171) p value b Gender 0.940  Male 33 (73.3%) 139 (72.8%)  Female 12 (26.7%) 52 (27.2%) Age (mean ± SD) 57.6 ± 13.2 61.9 ± 10.7 0.236 c HBV 24 (60.0%) 59 (34.5%) 0.003∗ HCV 10 (25.0%) 69 (40.4%) 0.071 Bilirubin ≥1.2 mg/dl 8 (20.0%) 38 (22.2%) 0.759 Albumin <3.5 g/dl 9 (22.5%) 30 (17.5%) 0.467 ALT ≥35 U/L 22 (55.5%) 103 (60.23%) 0.606 PT, INR ≥1.2 5 (12.5%) 6 (3.5%) 0.021∗ AFP ≥400 ng/ml 12 (30.0%) 33 (19.3%) 0.137 ICG, retention rate ≥15% 7 (17.5%) 28 (16.4%) 0.863 TNM stage <0.001∗  Stage 1/2 28 (70.0%) 158 (92.4%)  Stage 3/4 12 (30.0%) 13 (7.6%) Multiple tumors 5 (12.5%) 16 (9.4%) 0.550 Child-Pugh class 0.193  Class A 40 (100%) 164 (95.8%)  Class B 0 (0%) 7 (4.2%) Complete tumor capsule 9 (22.5%) 47 (27.5%) 0.520 Microvascular invasion 12 (30.0%) 22 (12.87%) 0.008∗ Macrovascular invasion 2 (5.0%) 4 (2.3%) 0.362 Cut margin free 36 (90.0%) 165 (96.5%) 0.082 Differentiation 0.149  Grade 1/2 11 (27.5%) 68 (39.8%)  Grade 3/4 29 (72.5%) 103 (60.2%) Tumor size ≥3 cm 27 (67.5%) 101 (59.1%) 0.326 Satellite nodules 10 (25.0%) 24 (14.0%) 0.090 OCT4 expression (≥2x) 38 (95.0%) 11 (6.4%) <0.0001∗ SNAIL expression (≥2x) 34 (85.0%) 12 (7.0%) <0.0001∗ TWIST expression (≥2x) 30 (75.0%) 12 (7.0%) <0.0001∗ All three genes (≥2x) 28 (70.0%) 2 (1.2%) <0.0001∗ ∗ p < 0.05. n = 211. AFP, alpha-fetoprotein; ALT, alanine aminotransferase; ICG, indocyanine green; PT-INR, prothrombin time-international normalized ratio. Bold font indicates that the item is statistically significant. a SENP1 High: SENP1 expression level of tumor/peritumor tissue ≥2; SENP1 Low: SENP1 expression level of tumor/peritumor tissue <2. b Chi-squared test unless specified. c Student’s t test. Fig. 1 SENP1 expression is associated with tumor recurrence, a pluripotent embryonic stem cell signature, and poor prognosis in HCC. (A) Kaplan-Meier curve for OS and DFS after HCC resection for assessment of the transcriptional expression of SENP1 in HCC tissues (n = 211, 40 SENP1_High and 171 SENP1_Low). (B) Statistics on TNM stages of patients with HCC by SENP1 expression levels (n = 211). p < 0.0001, Chi-squared test. (C) SENP1 expression in ANTT vs. T tissues from the GSE76427 GEO dataset. ∗∗ p < 0.01, Mann-Whitney test. (D) GSEA showing the correlation of SENP1 with HCC recurrence (HCC_RECURRENCE_UP and HCC_RECURRENCE_DN) and embryonic stem cell signature (ESC_CORE) gene sets in HCC from the GSE76427 GEO dataset. (E) qPCR analysis of SENP1 , OCT4 , SNAIL , and TWIST in T and PT tissues; correlations presented as (−ΔΔCt of T/PT) ( n = 211). ∗∗∗ p < 0.001, Spearman’s test. ANTT, adjacent non-tumorous tissue; DFS, disease-free survival; ESC, embryonic stem cell; GEO, Gene Expression Omnibus; GSEA, gene set-enrichment analysis; HCC, hepatocellular carcinoma; NES, normalized enrichment score; OS, overall survival; PT, peritumor; T, tumor. Fig. 1 Variables associated with high and low SENP1 expression in HCC. ∗ p < 0.05. n = 211. AFP, alpha-fetoprotein; ALT, alanine aminotransferase; ICG, indocyanine green; PT-INR, prothrombin time-international normalized ratio. Bold font indicates that the item is statistically significant. SENP1 High: SENP1 expression level of tumor/peritumor tissue ≥2; SENP1 Low: SENP1 expression level of tumor/peritumor tissue <2. Chi-squared test unless specified. Student’s t test. SENP1 expression is associated with tumor recurrence, a pluripotent embryonic stem cell signature, and poor prognosis in HCC. (A) Kaplan-Meier curve for OS and DFS after HCC resection for assessment of the transcriptional expression of SENP1 in HCC tissues (n = 211, 40 SENP1_High and 171 SENP1_Low). (B) Statistics on TNM stages of patients with HCC by SENP1 expression levels (n = 211). p < 0.0001, Chi-squared test. (C) SENP1 expression in ANTT vs. T tissues from the GSE76427 GEO dataset. ∗∗ p < 0.01, Mann-Whitney test. (D) GSEA showing the correlation of SENP1 with HCC recurrence (HCC_RECURRENCE_UP and HCC_RECURRENCE_DN) and embryonic stem cell signature (ESC_CORE) gene sets in HCC from the GSE76427 GEO dataset. (E) qPCR analysis of SENP1 , OCT4 , SNAIL , and TWIST in T and PT tissues; correlations presented as (−ΔΔCt of T/PT) ( n = 211). ∗∗∗ p < 0.001, Spearman’s test. ANTT, adjacent non-tumorous tissue; DFS, disease-free survival; ESC, embryonic stem cell; GEO, Gene Expression Omnibus; GSEA, gene set-enrichment analysis; HCC, hepatocellular carcinoma; NES, normalized enrichment score; OS, overall survival; PT, peritumor; T, tumor. We previously reported elevated OCT4 expression in tumors, especially in HBV-related HCC. 7 Given the link between SENP1 and stemness/EMT markers, we analyzed expression profiles of SENP1 , OCT4 , SNAIL , and TWIST across 211 HCC samples of various etiologies. As shown in Fig. 2 A, patients with HBV-HCC (n = 83) exhibited significantly higher mRNA levels of these genes compared to NBNC-HCC (n = 23), HCV-HCC (n = 79), and BC-HCC (n = 26), indicating a strong correlation between SENP1 and OCT4 in HBV-HCC. To evaluate the prognostic significance, patients were stratified by gene expression and their DFS was analyzed within 24 months. A Kaplan-Meier analysis demonstrated that elevated mRNA levels of SENP1 (7.8 vs. 15.7 months, p = 0.003), OCT4 (7.8 vs. 16.7 months, p < 0.001), SNAIL (8.6 vs. 15.7 months, p = 0.012), and TWIST (8.2 vs. 15.5 months, p = 0.028) were significantly associated with a shorter DFS ( Fig. 2 B). Immunohistochemical (IHC) staining confirmed significantly higher expression of SENP1 and the OCT4 and CD133 stemness-related proteins in HBV-HCC tumor tissues compared to NBNC-HCC tumor tissues ( Fig. 2 C, D). A positive correlation was observed between SENP1, OCT4, and CD133 protein levels ( Fig. 2 E). Fig. 2 Positive correlations of SENP1 , OCT4 , CD133 , SNAIL , and TWIST expression levels with early tumor recurrence in HBV-related HCC. (A) qPCR analysis of SENP1 , OCT4 , SNAIL , and TWIST in paired T and PT tissues from patients with HCC of different etiologies: NBNC (n = 23), HBV (n = 83), HCV (n = 79), and BC (n = 26). ∗p < 0.05, ∗∗ p < 0.01, Mann-Whitney test. (B) Kaplan-Meier analysis of DFS (early tumor recurrence, 24 months) in HBV-related HCC (n = 83) based on transcriptional expression of the indicated genes. (C) Representative immunohistochemical staining images of SENP1, OCT4, and CD133 in tumor sections from patients with NBNC- and HBV-related HCC. Scale bars = 100 μm. (D) Quantitative data of (c) are shown as intensity/pixel. ∗ p < 0.05, ∗∗ p < 0.01, Mann-Whitney test. (E) Correlations of protein levels (intensity/pixel) of SENP1 with OCT4 (n = 27) and CD133 (n = 24) in HBV-HCC tumor sections. Spearman’s test. BC, HBV and HCV; DFS, disease-free survival; HCC, hepatocellular carcinoma; NBNC, non-HBV and non-HCV; PT, peritumor; T, tumor. Fig. 2 Positive correlations of SENP1 , OCT4 , CD133 , SNAIL , and TWIST expression levels with early tumor recurrence in HBV-related HCC. (A) qPCR analysis of SENP1 , OCT4 , SNAIL , and TWIST in paired T and PT tissues from patients with HCC of different etiologies: NBNC (n = 23), HBV (n = 83), HCV (n = 79), and BC (n = 26). ∗p < 0.05, ∗∗ p < 0.01, Mann-Whitney test. (B) Kaplan-Meier analysis of DFS (early tumor recurrence, 24 months) in HBV-related HCC (n = 83) based on transcriptional expression of the indicated genes. (C) Representative immunohistochemical staining images of SENP1, OCT4, and CD133 in tumor sections from patients with NBNC- and HBV-related HCC. Scale bars = 100 μm. (D) Quantitative data of (c) are shown as intensity/pixel. ∗ p < 0.05, ∗∗ p < 0.01, Mann-Whitney test. (E) Correlations of protein levels (intensity/pixel) of SENP1 with OCT4 (n = 27) and CD133 (n = 24) in HBV-HCC tumor sections. Spearman’s test. BC, HBV and HCV; DFS, disease-free survival; HCC, hepatocellular carcinoma; NBNC, non-HBV and non-HCV; PT, peritumor; T, tumor. CSC-associated properties have been well reported to involve expression of stemness-related markers such as OCT4 and CD133, and EMT-related N-cadherin, SNAIL, and TWIST, and with the ability to form tumor secondary spheres. 7 , 18 To investigate the effect of SENP1 on CSC-associated properties in HCC, overexpression and silencing approaches were respectively applied to HCC cell lines with low (Huh7 and PLC5) and high (HepG2 and Hep3B) endogenous SENP1 expression ( Fig. S1 ). SENP1 overexpression in Huh7 and PLC5 cells significantly upregulated OCT4-and EMT-associated proteins such as N-cadherin, SNAIL, and TWIST ( Fig. 3 A), whereas the silencing of SENP1 expression in HepG2 and Hep3B cells reduced levels of these markers ( Fig. 3 B). Fig. 3 SENP1 regulates OCT4 expression, the EMT, tumor sphere formation, and the CD133 + cell population in HCC. (A,B) Western blot analysis of SENP1, OCT4, N-cadherin, SNAIL, and TWIST in Huh7 and PLC5 cells overexpressing SENP1 (A) and HepG2 and Hep3B cells with SENP1-knockdown (shSENP1#1 and shSENP1#2) (B). (C,D) Secondary tumor sphere-formation assay in HepG2 cells with OCT4 (C) or SENP1 (D) overexpression or silencing. Spheres of >100 μm were quantified (right panels). Scale bars = 100 μm. (E) Wound-closure assay to assess the effect of SENP1 overexpression on the migration of Hep3B and PLC5 cells at 0 and 24 h. Quantified data are shown (right panel). (F) Western blot of CD133, SENP1, OCT4, N-cadherin, SNAIL, and TWIST protein levels in sorted CD133 high and CD133 −/low HepG2 and PLC5 cells. (G) Flow cytometric analysis of CD133 + cell populations in HepG2 cells. (H) The CD133 + cell population was analyzed by flow cytometry in sorted CD133 high cells from HepG2 cells after being cultured for the indicated time. Quantification is shown (right panels). β-ACTIN served as the loading control; quantified values are shown below. ∗ p < 0.05, ∗∗∗ p < 0.001, Student’s t test. Ctrl, control vector; EMT, epithelial-mesenchymal transition; HCC, hepatocellular carcinoma; shCtrl, shRNA of LacZ gene; shSENP1, small hairpin RNA targeting SENP1. Fig. 3 SENP1 regulates OCT4 expression, the EMT, tumor sphere formation, and the CD133 + cell population in HCC. (A,B) Western blot analysis of SENP1, OCT4, N-cadherin, SNAIL, and TWIST in Huh7 and PLC5 cells overexpressing SENP1 (A) and HepG2 and Hep3B cells with SENP1-knockdown (shSENP1#1 and shSENP1#2) (B). (C,D) Secondary tumor sphere-formation assay in HepG2 cells with OCT4 (C) or SENP1 (D) overexpression or silencing. Spheres of >100 μm were quantified (right panels). Scale bars = 100 μm. (E) Wound-closure assay to assess the effect of SENP1 overexpression on the migration of Hep3B and PLC5 cells at 0 and 24 h. Quantified data are shown (right panel). (F) Western blot of CD133, SENP1, OCT4, N-cadherin, SNAIL, and TWIST protein levels in sorted CD133 high and CD133 −/low HepG2 and PLC5 cells. (G) Flow cytometric analysis of CD133 + cell populations in HepG2 cells. (H) The CD133 + cell population was analyzed by flow cytometry in sorted CD133 high cells from HepG2 cells after being cultured for the indicated time. Quantification is shown (right panels). β-ACTIN served as the loading control; quantified values are shown below. ∗ p < 0.05, ∗∗∗ p < 0.001, Student’s t test. Ctrl, control vector; EMT, epithelial-mesenchymal transition; HCC, hepatocellular carcinoma; shCtrl, shRNA of LacZ gene; shSENP1, small hairpin RNA targeting SENP1. In the secondary sphere-formation assay, the effect of OCT4 was utilized as a positive control. As depicted in Fig. 3 C, OCT4 overexpression in HepG2 cells led to increases in both the size and number of secondary tumor spheres, whereas suppression of OCT4 expression inhibited sphere formation. Similar to the effects of OCT4, SENP1 overexpression notably enhanced the formation of secondary tumor spheres, and silencing of SENP1 expression mitigated the enhanced effect ( Fig. 3 D). Consistently, wound-closure assays revealed an enhanced migratory ability upon overexpression of SENP1 ( Fig. 3 E), supporting its involvement in the EMT and cell motility. These findings strongly support the role of SENP1 in CSC-associated properties and highlight the close relationship between SENP1 and OCT4/EMT in HCC. To investigate the correlation between SENP1 and the CSC-related factor, CD133, magnetic cell sorting was used to isolate CD133 high and CD133 -/low fractions from HepG2 and PLC5 cells. Compared to CD133 −/low cells, CD133 high cells exhibited high expression levels of SENP1 and CSC-related proteins, including OCT4, N-cadherin, SNAIL, and TWIST ( Fig. 3 F). To investigate the impact of SENP1 on the CD133 + cell population, knockdown of SENP1 expression was performed in HepG2 cells. SENP1 silencing significantly reduced the ratio of the CD133 + cell population ( Fig. 3 G). To further explore the role of SENP1 in CD133 + cell self-renewal, CD133 high HepG2 cells were transduced with short hairpin (sh)SENP1 and cultured for 7 days. Results demonstrated that the silencing of SENP1 expression considerably decreased the CD133 + cell population in HepG2 cells compared to the vector control group ( Fig. 3 H). Results indicated that SENP1 plays an important role in the self-renewal of CD133 + HepG2 cells, thus maintaining the CSC population. The HBx protein was demonstrated to increase OCT4 expression and drive the pathogenesis of HBV-related HCC. 10 Given elevated SENP1/OCT4 expression in HBV-related HCC, we investigated the HBx protein’s effect on SENP1 and its downstream impact on stemness features in HCC cells, including OCT4, EMT markers, secondary sphere formation, and migration/invasion. HBx was overexpressed in HBV-negative (Huh7 and HepG2) and HBV-positive (Hep3B and PLC5) HCC cell lines. HBx significantly increased SENP1, OCT4, N-cadherin, SNAIL, and TWIST expression, while slightly decreasing E-cadherin expression ( Figs. 4 A, and S2A). Immunofluorescence staining results showed that cells with high levels of HBx-GFP expression had higher SENP1 expression compared to cells without HBx-GFP expression ( Fig. 4 B, SENP1 in red, and HBx-GFP in green, as indicated by arrowheads). Additionally, HBx overexpression induced a morphological transition in HepG2 cells from an epithelial-like to a mesenchymal-like phenotype, which was reversed upon SENP1 silencing ( Fig. S2B ). Fig. 4 HBx increases CSC-associated properties through SENP1 regulation in HCC. (A) Western blot analysis of indicated proteins in HepG2 cells transduced with GFP or HBx-GFP expression (a); quantified protein levels are shown in (b). (B) Co-localization of the HBx protein-GFP (green), SENP1 (red), and DAPI (blue) in HepG2 cells. An arrowhead indicates a cell with HBx-GFP expression. (C) Western blot of indicated proteins in GFP-HepG2 and HBx-HepG2 cells with or without SENP1-knockdown (shSENP1#1 and shSENP1#2) (a); quantified protein levels are shown in (b). (D) Secondary tumor sphere formation assay in GFP-HepG2 and HBx-HepG2 cells with or without silencing SENP1 or OCT4 expression (a); quantification of spheres >100 μm is shown in (b). (E) Flow cytometric analysis of GFP + CD133 + populations in HepG2 cells (a); quantified in (b). (F) Wound-closure assay to assess the effect of SENP1 silencing on HBx-induced cell migration of HepG2 cells at 0, 12, and 24 h (a); quantified in (b). (G) Matrigel-coated transwell assay evaluating cell invasion under indicated conditions (a); quantified in (b). ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001, Student’s t test. Scale bars = 100 μm. Western blot analyses were conducted across all conditions, with β-ACTIN as the loading control. CSC, cancer stem cell; HBx, HBV X protein; HCC, hepatocellular carcinoma; shSENP1, small hairpin RNA targeting SENP1. Fig. 4 HBx increases CSC-associated properties through SENP1 regulation in HCC. (A) Western blot analysis of indicated proteins in HepG2 cells transduced with GFP or HBx-GFP expression (a); quantified protein levels are shown in (b). (B) Co-localization of the HBx protein-GFP (green), SENP1 (red), and DAPI (blue) in HepG2 cells. An arrowhead indicates a cell with HBx-GFP expression. (C) Western blot of indicated proteins in GFP-HepG2 and HBx-HepG2 cells with or without SENP1-knockdown (shSENP1#1 and shSENP1#2) (a); quantified protein levels are shown in (b). (D) Secondary tumor sphere formation assay in GFP-HepG2 and HBx-HepG2 cells with or without silencing SENP1 or OCT4 expression (a); quantification of spheres >100 μm is shown in (b). (E) Flow cytometric analysis of GFP + CD133 + populations in HepG2 cells (a); quantified in (b). (F) Wound-closure assay to assess the effect of SENP1 silencing on HBx-induced cell migration of HepG2 cells at 0, 12, and 24 h (a); quantified in (b). (G) Matrigel-coated transwell assay evaluating cell invasion under indicated conditions (a); quantified in (b). ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001, Student’s t test. Scale bars = 100 μm. Western blot analyses were conducted across all conditions, with β-ACTIN as the loading control. CSC, cancer stem cell; HBx, HBV X protein; HCC, hepatocellular carcinoma; shSENP1, small hairpin RNA targeting SENP1. To further investigate whether HBx regulates stemness and the EMT via SENP1, HepG2 cells were transduced with HBx-GFP or control-GFP vectors, with or without SENP1-knockdown (shSENP1#1 or shSENP1#2). SENP1 silencing markedly reduced HBx-induced expression of OCT4-and EMT-associated markers, including N-cadherin, SNAIL, and TWIST ( Fig. 4 C). The impact of HBx on CSC-associated properties was further assessed by analyzing secondary tumor sphere formation and the CD133 + cell population in HepG2 cells. HBx overexpression significantly enhanced sphere formation ( Fig. 4 D) and increased the CD133 + cell population ( Fig. 4 E), whereas SENP1-knockdown markedly inhibited these HBx-induced effects. Moreover, HBx promoted EMT-associated phenotypes, including enhanced cell migration and invasion, as shown by wound-closure and transwell assays. These phenotypes were effectively suppressed by SENP1 silencing ( Fig. 4 F, G). PIN1 inhibition enhances sorafenib sensitivity and suppresses HCC growth, 15 while its overexpression is frequently observed in HBV-related HCC. 14 To examine the clinical relationship between PIN1 and SENP1, IHC staining was performed on tumor tissues from patients with NBNC and HBV-related HCC. As shown in Fig. 5 A, SENP1 and PIN1 protein levels were elevated in HBV-related HCC, with a strong positive correlation observed between the two ( Fig. 5 B). Fig. 5 HBx increases PIN1 and cyclin D1 expressions through SENP1 in HBV-related HCC. (A) Immunohistochemical staining of SENP1 and PIN1 in representative tumor sections from patients with NBNC-HCC (n = 8) and HBV-HCC (n = 25). Quantification of protein levels (intensity/pixel) is shown on the right. Scale bars = 100 μm ∗∗ p < 0.01, Mann-Whitney test. (B) Correlation of protein levels of SENP1 and PIN1 in HBV-HCC tumor sections (n = 25). p < 0.001, Spearman’s test. (C) Western blot analysis of SENP1, PIN1, cyclin D1, and HBx in HepG2 cells overexpressing HBx-GFP. (D) Co-localization of the HBx protein-GFP (green), PIN1 (red), and DAPI (blue) in HepG2 cells. The arrowhead indicates a non-transduced cell. (E) Western blot analysis of SENP1, PIN1, and cyclin D1 in SENP1-overexpressing HepG2 cells. (F) Western blot analysis of the indicated proteins in GFP- and HBx-GFP HepG2 cells with or without SENP1-knockdown (shSENP1#1 and shSENP1#2). β-ACTIN served as the loading control; quantified values are shown below. HBx, HBV X protein; HCC, hepatocellular carcinoma; NBNC, non-B non-C; shSENP1, small hairpin RNA targeting SENP1. Fig. 5 HBx increases PIN1 and cyclin D1 expressions through SENP1 in HBV-related HCC. (A) Immunohistochemical staining of SENP1 and PIN1 in representative tumor sections from patients with NBNC-HCC (n = 8) and HBV-HCC (n = 25). Quantification of protein levels (intensity/pixel) is shown on the right. Scale bars = 100 μm ∗∗ p < 0.01, Mann-Whitney test. (B) Correlation of protein levels of SENP1 and PIN1 in HBV-HCC tumor sections (n = 25). p < 0.001, Spearman’s test. (C) Western blot analysis of SENP1, PIN1, cyclin D1, and HBx in HepG2 cells overexpressing HBx-GFP. (D) Co-localization of the HBx protein-GFP (green), PIN1 (red), and DAPI (blue) in HepG2 cells. The arrowhead indicates a non-transduced cell. (E) Western blot analysis of SENP1, PIN1, and cyclin D1 in SENP1-overexpressing HepG2 cells. (F) Western blot analysis of the indicated proteins in GFP- and HBx-GFP HepG2 cells with or without SENP1-knockdown (shSENP1#1 and shSENP1#2). β-ACTIN served as the loading control; quantified values are shown below. HBx, HBV X protein; HCC, hepatocellular carcinoma; NBNC, non-B non-C; shSENP1, small hairpin RNA targeting SENP1. The role of HBx/SENP1 in regulating PIN1 expression was further investigated. Consistent with data described earlier in this text, we found that HBx protein overexpression significantly increased SENP1 protein levels ( Fig. 4 , Fig. 5 C), and importantly, it effectively induced expression of the PIN1 and cyclin D1 proteins ( Fig. 5 C). The effect of the HBx protein on PIN1 expression was confirmed by immunofluorescence staining targeting PIN1, which revealed the clear co-localization of HBx-GFP (indicated by GFP) and PIN1 (indicated by red fluorescence) in HCC cells. Higher HBx expression was correlated with more-intense PIN1 immunostaining. In Fig. 5 D, an arrowhead points to an HCC cell showing low levels of both the HBx and PIN1 proteins. Additionally, to investigate the upstream regulatory role of SENP1 in PIN1 protein expression, HA-SENP1 was overexpressed in HepG2 cells to assess the induction of PIN1 and cyclin D1. As shown in Fig. 5 E, SENP1 increased PIN1 and cyclin D1 expression in HepG2 cells. Furthermore, the role of SENP1 in HBx-PIN1/cyclin D1 regulation was investigated in HepG2 cells transfected with HBx-GFP alone or in combination with SENP1 shRNA. As illustrated in Fig. 5 F, HBx-GFP overexpression significantly increased protein levels of SENP1, PIN1, and cyclin D1, and shSENP1, effectively attenuating the effect of HBx-GFP. These results strongly support the regulatory role of HBx/SENP1 in PIN1 expression in HCC cells. To validate the in vitro findings of HBx/SENP1-mediated regulation of OCT4 and EMT-related properties ( Table 1 , Fig. 1 , Fig. 2 , Fig. 3 , Fig. 4 ), an orthotopic liver xenograft model was established using HepG2 cells stably expressing the GFP or HBx protein, with or without SENP1 shRNA (shSENP1). Control shRNA (shCtrl) was used as the control group. HCC tumor-bearing mice were divided into four groups: GFP-shCtrl (control group, n = 5), GFP-shSENP1 (control group with shSENP1, n = 6), HBx-shCtrl (HBx group, n = 8), and HBx-shSENP1 (HBx group with shSENP1, n = 12). Notably, mice in the GFP-shCtrl and HBx-shCtrl groups exhibited significant reductions in body weight compared to those in the SENP1-silenced groups (GFP-shSENP1 and HBx-shSENP1) ( Fig. 6 A). Tumor sizes in each group were monitored by bioluminescence intensity. As shown in Fig. 6 B, bioluminescence imaging demonstrated a progressive increase in signal intensity in tumors in both the GFP-shCtrl and HBx-shCtrl groups, with the HBx-shCtrl group exhibiting a markedly stronger signal. In contrast, SENP1 silencing effectively suppressed tumor progression, as evidenced by the bioluminescence intensity of liver tissues in the GFP-shSENP1 and HBx-shSENP1 groups at 8 weeks post-implantation of HCC cells ( Fig. 6 B). Quantitative data are shown in Fig. 6 C. Further gross examination of liver tissues at 8 weeks post-implantation revealed discernible HCC tumors in the experimental group without SENP1 silencing ( Figs. 6 D and S3, GFP-shCtrl and HBx-shCtrl groups). Tumor sizes and liver masses were notably greater in the HBx-shCtrl group compared to the GFP-shCtrl group, indicating enhanced tumorigenicity ( Fig. 6 D). Consistent with the bioluminescence results in Fig. 6 B, SENP1-silenced groups (GFP-shSENP1 and HBx-shSENP1) exhibited either markedly smaller liver tumors or no visible tumor formation, demonstrating that SENP1 silencing substantially impaired tumor outgrowth in vivo ( Figs. 6 D and S3). These observations were further verified by H&E staining of representative histological sections corresponding to HCC tumors in Fig. 6 D ( Fig. 6 E). Additionally, H&E staining revealed lung metastases in tumor-bearing mice ( Fig. 6 F, Table S3 ). The HBx-shCtrl group showed more-extensive pulmonary lesions compared to the other groups, based on a quantitative analysis of the lung metastatic area ( Fig. 6 G). Importantly, liver tumors exhibited invasion into the fibrous capsule in mice with lung metastasis ( Fig. 6 H). In contrast, no such invasion was observed in the SENP1-silenced group ( Fig. 6 H). Immunostaining results revealed elevated levels of SENP1, OCT4, CD133, N-cadherin, and PIN1 proteins in the HBx-shCtrl group, whereas these proteins were markedly reduced in both the liver and lung metastatic tumors in the SENP1-silenced groups ( Fig. 6 H, I). Fig. 6 HBx-induced SENP1 expression promotes tumor growth, metastasis, and sorafenib refractoriness in an orthotopic HCC xenograft model. (A) Body weights of mice in each group were monitored throughout the experiment: GFP-shCtrl (n = 5), GFP-shSENP1 (n = 6), HBx-shCtrl (n = 8), and HBx-shSENP1 (n = 12). Data from week 8 were analyzed using the Mann-Whitney U test. ∗ p < 0.05, ∗∗ p < 0.01. (B) Representative bioluminescence images of orthotopic HCC xenografts from the GFP-shCtrl (n = 5), GFP-shSENP1 (n = 6), HBx-shCtrl (n = 8), and HBx-shSENP1 (n = 12) groups. Data from week 8 were analyzed using the Mann-Whitney U test. ∗∗ p < 0.01, ∗∗∗ p < 0.001. (C) Bioluminescence quantification of the tumor burden, expressed as total flux (photons/s). (D) Liver morphology of mice of the GFP-shCtrl, GFP-shSENP1, HBx-shCtrl, and HBx-shSENP1 groups at week 8 post orthotopic HCC cell injection. (E) Representative H&E staining of liver tumor xenograft tissues. (F) Representative lung images highlighting metastatic lesions in the GFP-shCtrl and HBx-shCtrl groups. (G) Quantification of metastatic tumor areas in lung tissue sections (10 4 μm 2 /mouse/slide), corresponding to panel F. (H,I) Histological evaluation of liver tumors (H) and lung metastases (I) in the xenograft mouse model. H&E staining of representative liver tumor (H) and lung metastasis (I) tissue sections. The yellow dotted line in panel H highlights the fibrous capsule at the tumor margin; the red boxed region is shown at higher magnification below. Immunohistochemical staining of SENP1, OCT4, CD133, E-cadherin, N-cadherin, and PIN1 was performed on representative samples from liver tumors and lung metastases. Scale bars = 50 μm. HBx, HBV X protein; HCC, hepatocellular carcinoma; shCtrl, control small hairpin RNA; shSENP1, small hairpin RNA targeting SENP1. Fig. 6 HBx-induced SENP1 expression promotes tumor growth, metastasis, and sorafenib refractoriness in an orthotopic HCC xenograft model. (A) Body weights of mice in each group were monitored throughout the experiment: GFP-shCtrl (n = 5), GFP-shSENP1 (n = 6), HBx-shCtrl (n = 8), and HBx-shSENP1 (n = 12). Data from week 8 were analyzed using the Mann-Whitney U test. ∗ p < 0.05, ∗∗ p < 0.01. (B) Representative bioluminescence images of orthotopic HCC xenografts from the GFP-shCtrl (n = 5), GFP-shSENP1 (n = 6), HBx-shCtrl (n = 8), and HBx-shSENP1 (n = 12) groups. Data from week 8 were analyzed using the Mann-Whitney U test. ∗∗ p < 0.01, ∗∗∗ p < 0.001. (C) Bioluminescence quantification of the tumor burden, expressed as total flux (photons/s). (D) Liver morphology of mice of the GFP-shCtrl, GFP-shSENP1, HBx-shCtrl, and HBx-shSENP1 groups at week 8 post orthotopic HCC cell injection. (E) Representative H&E staining of liver tumor xenograft tissues. (F) Representative lung images highlighting metastatic lesions in the GFP-shCtrl and HBx-shCtrl groups. (G) Quantification of metastatic tumor areas in lung tissue sections (10 4 μm 2 /mouse/slide), corresponding to panel F. (H,I) Histological evaluation of liver tumors (H) and lung metastases (I) in the xenograft mouse model. H&E staining of representative liver tumor (H) and lung metastasis (I) tissue sections. The yellow dotted line in panel H highlights the fibrous capsule at the tumor margin; the red boxed region is shown at higher magnification below. Immunohistochemical staining of SENP1, OCT4, CD133, E-cadherin, N-cadherin, and PIN1 was performed on representative samples from liver tumors and lung metastases. Scale bars = 50 μm. HBx, HBV X protein; HCC, hepatocellular carcinoma; shCtrl, control small hairpin RNA; shSENP1, small hairpin RNA targeting SENP1. Sorafenib refractoriness remains a key therapeutic concern in HCC. To evaluate whether SENP1 modulates sorafenib responsiveness, we treated Hep3B, PLC5, HepG2215, and sorafenib-resistant HepG2215_R and Mahlavu_R cells with increasing concentrations of sorafenib in the presence of various doses of a SENP1 inhibitor (SENP1-IN-3). The individual IC 50 (50% inhibitory concentration) values for sorafenib and the SENP1 inhibitor in each HCC cell line are presented in Fig. S4 . Dose-response profiles for combined treatments with sorafenib and the SENP1 inhibitor are plotted in Fig. 7 A, and corresponding IC 50 values were calculated and are summarized in Fig. 7 B. These results demonstrate that SENP1 inhibition sensitized both naïve and sorafenib-resistant HCC cells to sorafenib treatment. To further assess the impact of SENP1 on the sorafenib response, HCC xenograft-bearing mice were treated with sorafenib and monitored by bioluminescence imaging. As shown in Fig. 7 C,D, the GFP-shCtrl group (n = 4) exhibited a significant reduction in the tumor signal after 1 week of treatment, whereas the HBx-shCtrl group (n = 7) showed a marked increase in the tumor burden over the same period. After 4 weeks of sorafenib treatment, no therapeutic response was observed in the HBx-shCtrl group, as evidenced by continued tumor progression. In contrast, two mice in the GFP-shCtrl group demonstrated sustained suppression of tumor growth throughout the treatment period ( Fig. 7 C,D). Importantly, no apparent tumor progression was observed in the SENP1-silenced groups throughout the treatment period, indicating effective suppression of tumor growth ( Fig. 7 C,D). Fig. 7 HBx/SENP1 signaling decreases cellular responsiveness to sorafenib in vitro and reduces tumor sensitivity in an in vivo xenograft model. (A) Dose-response curves of sorafenib in Hep3B, PLC5, HepG2215, HepG2215_R, and Mahlavu_R HCC cell lines treated with increasing concentrations of SENP1-IN-3. Combination treatment was administered for 48 h, and cell viability was measured to assess the effect of SENP1-IN-3 on sorafenib sensitivity. (B) IC 50 values of sorafenib in five HCC cell lines following combination treatment with SENP1-IN-3, as calculated from dose-response curves in panel A. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001, Student’s t test. (C) In vivo assessment of sorafenib response in four xenograft mouse groups: GFP-shCtrl (n = 4), GFP-shSENP1 (n = 5), HBx-shCtrl (n = 7), and HBx-shSENP1 (n = 10). Mice received intraperitoneal sorafenib treatment (30 mg/kg body weight, twice weekly) beginning at week 8 post-cell injection (pretreatment). (D) The tumor burden was monitored weekly by bioluminescence imaging. Flux (photons/s) at week 8 was set to 100% as the baseline for normalization. The panel shows the tumor burden in individual mice. HCC, hepatocellular carcinoma; shCtrl, shRNA of LacZ gene; shSENP1, small hairpin RNA targeting SENP1. Fig. 7 HBx/SENP1 signaling decreases cellular responsiveness to sorafenib in vitro and reduces tumor sensitivity in an in vivo xenograft model. (A) Dose-response curves of sorafenib in Hep3B, PLC5, HepG2215, HepG2215_R, and Mahlavu_R HCC cell lines treated with increasing concentrations of SENP1-IN-3. Combination treatment was administered for 48 h, and cell viability was measured to assess the effect of SENP1-IN-3 on sorafenib sensitivity. (B) IC 50 values of sorafenib in five HCC cell lines following combination treatment with SENP1-IN-3, as calculated from dose-response curves in panel A. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001, Student’s t test. (C) In vivo assessment of sorafenib response in four xenograft mouse groups: GFP-shCtrl (n = 4), GFP-shSENP1 (n = 5), HBx-shCtrl (n = 7), and HBx-shSENP1 (n = 10). Mice received intraperitoneal sorafenib treatment (30 mg/kg body weight, twice weekly) beginning at week 8 post-cell injection (pretreatment). (D) The tumor burden was monitored weekly by bioluminescence imaging. Flux (photons/s) at week 8 was set to 100% as the baseline for normalization. The panel shows the tumor burden in individual mice. HCC, hepatocellular carcinoma; shCtrl, shRNA of LacZ gene; shSENP1, small hairpin RNA targeting SENP1.

This work was supported by research grants from the National Science and Technology Council, Taiwan (grant nos.: MOST 106-2314-B-038-057 and MOST 107-2314-B-038-087-MY3 [to YC Wu]; NSC 102-2628-B-038-008-MY3, NSTC 113-2314-B-038-136 and NSTC 114-2314-B-038-012 [to YH Huang]; MOST 104-2314-B-182-073-MY2 and MOST 105-2628-B-182-013-MY3 [to TS Chang]), and by grants from the Chen Wei-Tien Research Found for Thoracic Medicine, Taipei Medical University (TMU 111-5431-001-400 [to YC Wu]).

HepG2 (HBV-negative; BCRC, Taiwan), Huh7 (HBV-negative; JCRB, Japan), Hep3B (HBV-positive; BCRC), PLC5 (HBV-positive; BCRC), HepG2215 (derived from HepG2 by stable transfection with the HBV genome; kindly provided by Dr. Jun-Jen Liu, Institute of Medical Biotechnology, Taipei Medical University, Taipei, Taiwan), and Mahlavu (HBV-negative; kindly provided by Dr. Muh-Hwa Yang, Institute of Clinical Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan) HCC cell lines were used in this study. Sorafenib-resistant cell lines (HepG2215_R and Mahlavu_R) were established by gradually exposing parental cells (HepG2215 and Mahlavu) to increasing concentrations of sorafenib (Cell Signaling Technology, Boston, MA, USA). The resistant cell lines were subsequently maintained in medium containing 10 μM sorafenib. Patients’ frozen HCC tissues and corresponding formalin-fixed, paraffin-embedded HCC tissues were obtained from Chang Gung Memorial Hospital (Chiayi, Taiwan). This study was approved by the Institutional Review Board of Chang Gung Medical Foundation (IRB approval no. 101-5232B). A detailed description of the materials and methods is provided in Supplementary Materials.

YH Huang and TS Chang initiated and supervised the research; YC Wu and YH Huang conceived and designed the project; YC Wu, YC Huang, YC Kuo, MHT Ngo, and YT Sung performed the in vitro experiments and data analysis; KF Lee and LM Kuo handled pathological tissue processing and tissue microarray fabrication; SL Doong provided the HBx expression vector materials and offered valuable insights on the HBx-related research; YC Wu, YC Kuo, MHT Ngo, and HF Wang conducted the animal experiments; YC Wu, YH Huang, and TS Chang wrote the paper. All authors discussed the results and commented on the manuscript.

OCT4 is widely expressed in various tumor types and strongly linked to CSC-related properties. 19 A meta-analysis also revealed that OCT4 expression was correlated with tumor sizes, tumor numbers, cell differentiation, and the TNM stage, and that OCT4 expression was associated with poor 3- and 5-year OS and DFS rates in HCC. 20 However, the molecular mechanism through which OCT4, which is closely related to the pluripotency of ESCs, is re-expressed in somatic cancer cells remains unknown. In this study, we demonstrated that the HBx protein promotes CSC-associated properties, such as OCT4 upregulation, through regulation by SENP1. We also found that SENP1 and OCT4/EMT expression levels were positively correlated with early tumor recurrence in patients with HCC. The co-expression of HBx and SENP1 may contribute to the upregulation of OCT4/EMT expression in HCC. SENP1-knockdown effectively suppressed HBx-induced OCT4/EMT expression and inhibited intrahepatic and pulmonary metastases in vivo , suggesting that SENP1 could serve as a prognostic marker and a potential therapeutic target in HBV-related HCC. Niche and epigenetic regulation are regarded as pivotal factors in OCT4 functionality. 19 DNA methylation at CpG sites in promoter and exonic regions modulates OCT4 transcription in trophoblast cells and ESCs. 21 , 22 In somatic cancers such as glioblastomas, aberrant regulation by DNA methyltransferase (DNMT) can trigger CSC development through OCT4 reprogramming. 23 Our previous studies demonstrated that patients with HBV-HCC have higher serum IL-6 concentrations, where IL-6 promotes CSC-associated properties by activating the IGF-1/IGF-1R signaling pathway. 7 IL-6/IGF-1R signaling-mediated OCT4 expression and sorafenib resistance may involve DNMT3b in HBV + HCC cells. 24 In this study, we found that the IL-6/STAT3 signaling pathway not only mediated OCT4 expression but also partially increased SENP1 expression in HCC cells with HBx expression ( Fig. S5 ). Our previous data demonstrated that SENP1 stabilizes OCT4 protein levels, 11 and that OCT4 expression is modulated by IL-6/IGF-1R signaling. 7 Herein, we further validated that HBx-induced SENP1 contributes to IGF-1R upregulation in HepG2 cells ( Fig. S6 ). Collectively, these findings indicate that IL-6 enhances OCT4 expression and CSC-associated properties in HBV-related HCC via a complex regulatory network involving IGF-1/IGF-1R, DNMT3b, and SENP1. The HBx protein was shown to regulate epigenetic mechanisms, including micro (mi)RNA, DNA methylation, and histone modifications, which contribute to CSC features. 25 Previous studies indicated that miR-145 is downregulated in HBV-related HCC, which is potentially influenced by HBx. 26 MiR-145 was reported to target the 3′-untranslated region (UTR) of OCT4 and SENP1 mRNAs, 27 , 28 thereby suppressing their expression. Interestingly, OCT4 also binds to the miR-145 promoter and represses its transcription in human ESCs, 27 suggesting a potential positive feedback loop that may contribute to an enhancement of CSC properties. In this study, using our HBx-HepG2 cell model, we observed that HBx partially enhanced SENP1 promoter activity and significantly suppressed miR-145-5p expression ( Fig. S7A and B ). Sequence alignment identified a putative miR-145-5p-binding site within the SENP1 3′-UTR ( Fig. S7C ), and luciferase assays confirmed that miR-145-5p overexpression significantly reduced SENP1 3′-UTR activity ( Fig. S7D ). Collectively, these findings demonstrated that HBx increases SENP1 expression by partially enhancing SENP1 promoter activity and suppressing miR-145-5p expression, which counteracts its inhibitory effect on SENP1 mRNA ( Fig. S7E ). HBx influences several critical signaling pathways relevant to CSC-associated properties, including the Wnt/β-catenin, TGF-β1, NF-κB, Notch, Hedgehog, STAT3, and PI3K/AKT signaling pathways. 29 A study demonstrated that HBx upregulates CSC-related gene expression, such as CD133 and OCT4 , in OV6 + CSCs in HCC through the β-catenin signaling pathway. 30 Another intriguing study found that HBx increases levels of OCT4 and MYC, facilitating the cellular reprogramming of induced pluripotent stem cells and potentially contributing to the formation of liver CSCs. 31 A recent study employed the CRISPR/Cas9 system to target the HBx gene; HBx-knockdown effectively reduced signatures of EMT and CSC genes in HBV-HCC cells. 32 In this study, in addition to demonstrating that SENP1 regulates OCT4 to enhance CSC properties, we also found that SENP1 modulated the self-renewal capacity of CD133 high HepG2 cells in vitro ( Fig. 3 ). Furthermore, in vivo animal experiments revealed that silencing SENP1 reduced HBx-enhanced CD133 expression ( Fig. 6 ). To analyze the correlation between CD1 33 mRNA ( PROM1 ) and SENP1 mRNA in HCC, we utilized the GSE76427 public dataset. As shown in Fig. S8 , a significant positive correlation was observed between PRMO1 and SENP1 mRNA expression (n = 115). To further investigate, we examined protein levels of CD133 and SENP1 through IHC staining in HCC tissues, comprising eight NBNC-HCC and 24 HBV-HCC samples. Expression levels of SENP1 and CD133 were positively correlated in patients with HBV-HCC ( Fig. 2 C-E). Overall, our results further clarify the relationship between HBx and OCT4/CSC properties through regulation by SENP1. Upregulation of SENP1 was recently reported in various cancer types, including breast cancer, lung cancer, prostate cancer, HCC, and colorectal cancer. 33 SENP1 overexpression disrupts the balance of SUMOylation by targeting specific proteins, significantly contributing to tumor progression and poor prognoses. In this study, we demonstrated for the first time that SENP1 is overexpressed in HBV-related HCC and is correlated with poor OS, poor DFS and extrahepatic metastasis, a critical feature of malignant tumors that considerably affects patient prognosis. The EMT is a potential mechanism of tumor cell metastasis, and it recently emerged as an important regulator of CSC-associated properties in HCC. 34 Numerous studies identified the significant role of SENP1 in the EMT across various tumors. In HCC, SENP1 was demonstrated to promote hypoxia-induced cancer stemness by deSUMOylating HIF-1α, thereby establishing a positive feedback loop. 35 Additionally, SENP1 has been shown to regulate the hepatocyte growth factor-induced invasion and migration of HCC cells. 36 In our in vitro experiments, we demonstrated that SENP1 was associated with an increased EMT phenotype in HCC cells. Furthermore, in vivo experiments using an orthotopic xenograft model revealed a significant reduction in tumor metastasis following SENP1-knockdown. In addition, OCT4 is also involved in regulating the EMT. Our previous studies demonstrated that OCT4 increases levels of EMT-related factors, including SNAIL, TWIST, and SLUG, in HCC and human endometriosis. 37 , 38 These findings suggest that HBx-induced SENP1 and OCT4 may be a promising target for inhibiting the EMT and tumor metastasis in HBV-related HCC. PIN1 is overexpressed in HCC, particularly in HBV-related HCC, and is associated with adverse features such as increased tumor sizes, intrahepatic metastasis, and poor prognoses. 12 , 14 , 39 Functionally, PIN1 interacts with cyclin D1 and β-catenin to modulate oncogenic pathways such as PI3K/Akt/mTOR and the EMT, 40 and was shown to stabilize the HBx protein and enhance its transactivation potential in HCC. 14 Recent evidence implicates PIN1 in therapeutic resistance, with its knockdown enhancing sorafenib sensitivity in HCC models. 15 SENP1-mediated deSUMOylation enhances PIN1 activity in breast cancer. 16 Consistent with previous findings, our study demonstrated increased expression and a significant correlation between SENP1 and PIN1 in HBV-related HCC. Notably, HBx promotes expression of PIN1 and cyclin D1 through SENP1, suggesting that SENP1 facilitates HBx/PIN1-driven transcriptional activity. Collectively, these findings provide compelling evidence that HBx may regulate PIN1 expression and HBx/PIN1 transactivation through SENP1-mediated deSUMOylation in HBV-related HCC; thus, HBx promotes tumorigenesis. Drug refractoriness is a major issue impeding the successful chemotherapeutic treatment of HCC. Many factors contribute to the development of drug resistance in HCC; for example, a hypoxic tumor microenvironment is known to play a pivotal role in promoting refractoriness to therapy. 41 Sorafenib, a first-line multikinase inhibitor approved for advanced HCC, was shown to modestly prolong median survival by 3–5 months. [42] , [43] , [44] SENP1 was implicated in cancer progression and may contribute to sorafenib resistance via its regulation of oncogenic proteins such as PIN1. A study revealed that PIN1-knockdown effectively increased sensitivity to sorafenib in HCC. 15 Our findings demonstrated that HBx enhances PIN1 and cyclin D1 expression through SENP1, supporting its role in HBx-driven drug refractoriness. To assess the therapeutic relevance, we evaluated combination treatment using sorafenib and a SENP1 inhibitor (SENP1-IN-3) across five HCC cell lines, including resistant models. This approach significantly reduced the effective dose of sorafenib ( Fig. 7 A,B). SENP1-knockdown restored sorafenib sensitivity in HBx-expressing tumors in vivo ( Fig. 7 C), indicating that SENP1 is a key modulator of drug responsiveness in HBV-related HCC. In this study, we identified significant clinical associations of the SENP1-associated CSC- and EMT-related factors with poor OS and poor DFS in HBV-related HCC. We showed that HBV upregulates SENP1 expression through HBx, possibly by suppressing miR145 levels. This upregulation subsequently leads to increased expressions of CSC-associated factors (OCT4, CD133, and IGF-1R), EMT-associated factors (SNAIL and TWIST), and cell proliferation factors (PIN1 and cyclin D1). This promotes the self-renewal, migration, invasion, and sorafenib resistance of CSCs in HBV-related HCC. These findings highlight the crucial role of SENP1 in the progression of HBV-related HCC and suggest that SENP1 could serve as a prognostic marker and therapeutic target in this context.

Hepatocellular carcinoma (HCC), the predominant form of primary liver cancer, ranks as the third leading cause of cancer-related mortality worldwide and accounts for approximately 80% of liver cancer cases. 1 Chronic infection with the HBV remains the principal etiological factor, especially in the Asian-Pacific and sub-Saharan African regions, although the incidence of metabolic dysfunction-associated steatotic liver disease (MASLD)-related HCC is rising in Western populations. 2 Despite curative strategies such as surgical resection and local ablation, recurrence rates remain high, with intrahepatic recurrence being most common and pulmonary metastases representing the predominant extrahepatic manifestation. [3] , [4] , [5] Intrahepatic recurrences can arise from occult metastases within 2 years or as de novo tumors >2 years after surgery. 6 Preventing postoperative HCC recurrence is crucial for improving prognoses, but effective strategies remain limited despite numerous clinical trials. A recent study demonstrated that cancer stem cells (CSCs) play an important role in early HCC recurrence, especially in HBV-related HCC. 7 However, the role of HBV in HCC recurrence and CSC-associated properties remains unclear. Numerous studies have shown that the HBV X (HBx) protein plays an important role in HCC development. 8 , 9 The HBx protein was reported to induce CSC-associated properties, such as upregulation of the OCT4 and NANOG proteins in HCC, although the underlying mechanism remains unclear. 10 Our previous research indicated that patients with HBV-related HCC have shorter disease-free survival (DFS) compared to those with non-HBV/non-HCV, HCV-related, or HBV/HCV-related HCC. 7 We also demonstrated that SUMO/sentrin-specific protease 1 (SENP1) regulates OCT4 stability and activity in embryonal carcinoma cells. 11 Nonetheless, how OCT4 is regulated in HCC remains to be elucidated. Additionally , PIN1, a phosphorylation-dependent prolyl isomerase, is overexpressed in HCC and promotes tumorigenesis by modulating cell cycle regulators, apoptosis, and oncogenic signaling pathways. 12 , 13 PIN1 interacts with the HBx protein and enhances its transcriptional activity. 14 Notably, PIN1-knockdown was shown to sensitize HCC cells to sorafenib treatment, underscoring its role in therapeutic resistance. 15 Furthermore, SENP1 was reported to enhance PIN1 activity via deSUMOylation in breast cancer, 16 suggesting a potential regulatory axis involving SENP1, PIN1, and the HBx protein in HCC that remains to be elucidated. In this study, we demonstrate that HBx-induced SENP1 expression regulates levels of OCT4 and PIN1, promoting CSC-associated properties in vitro and in vivo . There properties include tumor sphere formation, the epithelial-mesenchymal transition (EMT), tumor growth, metastasis, and drug refractoriness. Our findings demonstrate that SENP1 is a key factor promoting cancer stemness in HBV-related HCC and may contribute to HCC recurrence and metastasis after surgery.

BC, HBV and HCV; CSC, cancer stem cell; DFS, disease-free survival; EMT, epithelial-mesenchymal transition; ESC, embryonic stem cell; HBx, HBV X protein; HCC, hepatocellular carcinoma; NBNC, non-HBV and non-HCV; OS, overall survival; PIN1, peptidylprolyl isomerase 1; SENP1, SUMO-specific peptidase 1; SNAIL, snail family transcriptional repressor 1; TWIST, twist family bHLH transcription factor 1.

The authors have no conflicts of interest to declare. Please refer to the accompanying ICMJE disclosure forms for further details.

The datasets analyzed during the current study are publicly available in the Gene Expression Omnibus (GEO) under accession no. GSE76427 .