YY1 induced USP13 transcriptional activation drives the malignant progression of hepatocellular carcinoma by deubiquitinating WWP1

Abstract

Background

Hepatocellular carcinoma (HCC) is the sixth most prevalent cancer globally and the third leading cause of cancer-related mortality. Protein ubiquitination and deubiquitination play vital roles in human cancers. Ubiquitin-specific protease 13 (USP13) is a deubiquitinating enzyme (DUB) that is involved in many cellular processes. However, the mechanism by which USP13 regulates deubiquitination remains largely unknown.

Methods

Clinical data were analyzed via online databases. USP13 expression in HCC cell lines and tissues was analyzed via western blotting and immunohistochemistry. A lentivirus was used to established stable USP13-knockdown and USP13-overexpression cells. Cell Counting Kit-8, colony formation, wound healing, Transwell, and sphere formation assays were used to detect the malignant behaviors of HCC cells in vitro. A subcutaneous mouse model was used to investigate the function of USP13 in vivo. Co-immunoprecipitation, chromatin immunoprecipitation and dual-luciferase reporter assays were conducted to explore the molecular regulation.

Results

USP13 was upregulated in HCC cell lines and tissues, which predicted a poor prognosis in patients with HCC. Functional experiments in which USP13 was overexpressed or depleted revealed the oncogenic role of USP13 in driving HCC progression both in vitro and in vivo. Mechanistically, WW domain–containing ubiquitin E3 ligase 1 (WWP1) was identified as a binding protein of USP13. Furthermore, USP13 can interact with WWP1 and then remove the K29- and K48-linked polyubiquitination chains from WWP1 to stabilize the WWP1 protein via the ubiquitin–proteasome pathway. Moreover, Yin Yang 1 (YY1) was explored as a new transcription factor of USP13, and YY1 could also upregulate WWP1 expression through USP13. Moreover, YY1 and WWP1 were shown to participate in the oncogenic role of USP13.

Conclusions

Our findings revealed the functional YY1/USP13/WWP1 signaling axis in HCC, identifying a promising therapeutic target for anti-HCC treatment.

Graphical Abstract

Introduction

Hepatocellular carcinoma (HCC) ranks as the sixth most prevalent cancer globally and the third leading cause of cancer-related mortality [1]. In China, it is the second most common cancer and the fourth leading cause of cancer-related death, with comparable rates of mortality and morbidity [2]. Owing to its inconspicuous early symptoms and lack of specific early biomarkers, only 30% of patients with HCC are eligible for curative treatment upon initial diagnosis. The remaining 70% of patients with HCC still require intervention treatments such as targeted and immune therapies to eventually render the tumor amenable to surgical resection [3]. Despite advancements in HCC treatment in recent years, the prognosis for patients with HCC remains suboptimal. Therefore, investigating the molecular mechanism of HCC malignant progression and revealing the potential molecular targets of HCC are highly important, as identifying these targets will eventually contribute to the development of targeted therapies for HCC treatment.

Ubiquitination, a post-translational modification (PTM), plays a crucial role in various cellular processes, such as cell cycle regulation, DNA damage repair, and cell signaling pathways [4]. Ubiquitination involves three types of enzymes: ubiquitin-activating enzymes (E1), ubiquitin-conjugating enzymes (E2), and ubiquitin ligases (E3) [5]. In brief, E1s activate and transfer ubiquitin to E2, and then E3 cooperates with E2s to mediate the ubiquitination of substrates [6]. During E1–E2–E3 cascade enzymatic reactions, ubiquitin, which is linked to substrates, can be removed by deubiquitinating enzymes (DUBs) to control protein fate [7]. To date, 2 E1s, 42 E2s, more than 600 types of E3s, and hundreds of DUBs have been discovered [4, 8, 9], forming a complex molecular network. Ubiquitination, which occurs towards the end of a protein’s lifespan, primarily triggers the degradation of its substrate proteins [10], which also regulates the phosphorylation [11] and other modifications of substrates, as reported in previous studies [12, 13].

Ubiquitin-specific protease 13 (USP13), which serves as a DUB, controls multiple cellular functions, including mitochondrial energy metabolism, autophagy, DNA repair, endoplasmic reticulum-associated degradation (ERAD), and other activities [14]. Many findings confirm that USP13 promotes the progression of various human cancers. USP13 influences cell cycle progression in gastric cancer by stabilizing cyclin D1 [15] and enhances the deubiquitination and stabilization of ATP citrate lyase (ACLY) and oxoglutarate dehydrogenase (OGDH) in ovarian tumors [16]. In gastrointestinal stromal tumors, USP13 stabilizes autophagy-related protein 5 (ATG5) to induce autophagy and imatinib resistance [17]. Conversely, tumor suppressive roles of USP13 have also been identified. By acting as a DUB, USP13 stabilizes PTEN to suppress breast cancer [18] and bladder cancer development [19]. Therefore, the function of USP13 is bidirectional.

Ubiquitin E3 ligase WW domain–containing ubiquitin E3 ligase 1 (WWP1) is a member of the homologous to the E6-associated protein carboxyl terminus (HECT) E3 family [20]. According to previous reports, WWP1 is involved in several diseases including lung cancer, breast cancer, prostate cancer, etc. [20]. Mechanistically, WWP1 ubiquitinates and degrades the proteins KLF5, Smad2, CXCR4, and LAST1 by acting as an E3 ligase in different tumors [21]. Interestingly, WWP1 uniquely modified PTEN through nondegradative K27-linked polyubiquitination, and thus, inhibited PTEN dimerization and membrane recruitment, subsequently activating the Akt pathway indirectly to increase cancer progression [12, 13]. However, there is no evidence that the regulation of WWP1 deubiquitination serves as a substrate.

In this study, we confirmed that USP13 was elevated in HCC and acted as an oncogene to enhance the malignant characteristics of HCC cells both in vitro and in vivo. WWP1 was subsequently revealed as a new binding protein of USP13. Moreover, USP13 stabilized WWP1 by removing the K29- and K48-linked polyubiquitination chains from WWP1 and then activated the Akt/mTOR pathway in HCC. In addition, via bioinformatic analyses and molecular biology experiments, we revealed that Yin Yang 1 (YY1) is a new transcription factor of USP13 that promotes USP13 transcription through binding to its promoter region. In conclusion, we demonstrated that YY1-USP13-WWP1 forms an oncogene axis to accelerate HCC development via the Akt/mTOR pathway.

Materials and methods

Cell culture

HCC cell lines including Li-7 (SCSP-5062), MHCC97H (SCSP-5092), Huh7 (SCSP-526), Hep3B (SCSP-5045), PLC/PRF/5 (SCSP-5095), HCCLM3 (SCSP-5093), and SNU387 (SCSP-5046); the hepatoblastoma cell line, HepG2 (SCSP-510); and human embryonic kidney 293 T (HEK293T) (SCSP-502) were kindly provided by Cell Bank, Chinese Academy of Sciences. The human hepatic cell line, THLE-2, was obtained from Jinyuan Biological (JY658, Shanghai, China). Short tandem repeat (STR) profiling was performed to authenticate all cell lines. Li-7 and SNU387 were cultured in RPMI-1640 (VivaCell, C3010–0500); Hep3B, PLC/PRF/5, and HepG2 were cultured with minimum essential medium (MEM, VivaCell, C3060–0500); and the other cells were grown in Dulbecco’s modified Eagle medium (DMEM, VivaCell, C3113–0500). Each culture medium was supplemented with 10% fetal bovine serum (SERANA, FBS-AS500) and 1% penicillin and streptomycin (Senrui biological, CR15140), then incubated at 37 ℃ with 5% CO₂.

siRNAs, plasmids, and lentivirus infection

siRNAs targeting and plasmid sequences of WWP1 and YY1, and lentiviruses of USP13 were all purchased from GenePharma (Shanghai, China). The USP13 truncation mutant and luciferase reporter vectors were purchased from Sangon Biotech (Shanghai, China). The HA-ubiquitin plasmid and its mutant plasmids came from our lab. The sequences of siRNA and shRNA are listed in Supplementary Table S1. Plasmids and siRNAs were transfected into cells using Hieff Trans™ Liposomal Transfection Reagent (Yeasen, 40802ES03) and Lipofectamine 3000 (Invitrogen, L3000015). The experiment followed the manufacturer’s instructions.

For lentivirus infection, 1 × 10⁵ cells were plated in a six-well dish and lentivirus was added to the cells with 1 mL fresh medium the next day. After 24 h, the medium with lentivirus was replaced with new medium. After 48 h, puromycin was used to screen stably infected cells. Finally, total RNA and protein were extracted to evaluate the knockdown or overexpression efficiency of the target gene.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR)

Total mRNA was isolated from HCC cells by RNA-Quick Purification Kit (ES science, RN001). cDNA was synthesized by PrimeScript™ RT Reagent Kit (Takara, RR037A). qRT-PCR was carried out with SYBR Green (Yeasen, 11184ES08) on an ABI 7500. The primers used in this study are listed in Supplementary Table S2.

Western blotting (WB)

Cells were lysed with cell lysis buffer (Beyotime, P0013) and supplemented with a proteinase inhibitor and a phosphatase inhibitor. Protein samples were quantified using the Pierce™ BCA Protein Assay Kit (ThermoFisher, 23225). Proteins were denatured, separated in SDS-PAGE under an electric field, and then transferred to a polyvinylidene difluoride (PVDF) membrane (Merck Millipore, IPFL00005). The membrane was then blocked in 5% nonfat milk for 2 h, and incubated with primary antibodies at 4 °C overnight. After washing 10 min in 1× Tris-buffered saline with Tween 20 (TBST) three times, the membrane was then incubated with a secondary antibody conjugated with horseradish peroxidase (ThermoFisher, 31460, 31430) for 1 h at room temperature, and finally visualized using ECL reagent (BIO-RAD, 1705061). The primary antibodies used are listed in Supplementary Table S3.

Cell Counting Kit-8 (CCK-8) and colony formation

Cells were seeded into 96-well plates at a density of 2000 cells per well and cultured for 24, 48, and 72 h, followed by the addition of 10 μl CCK-8 detection solution (Yeasen, 40203ES80) and absorbance measurements at 450 nm within 1–2 h.

For colony formation, 1000 cells were seeded in six-well plates and cultured for 10–14 days. The former colonies were fixed for 15 min with 4% paraformaldehyde, then dyed with 2% crystal violet for 30 min. Lastly, colonies were imaged and counted.

Wound healing and Transwell assays

Cells were placed in six-well plates. Then, cells were gently scraped using a fresh pipette tip, rinsed multiple times with phosphate-buffered saline (PBS), and incubated in serum-free DMEM for 24 h. The cells’ migrated area was photographed under a microscope and analyzed by ImageJ software.

Cell Transwell assays were performed using 24-well Transwell chambers (Corning, 3422). First, 200μl serum-free DMEM with 5 × 10⁴ cells were added into the upper chambers and 600 μl of DMEM with 10% serum was added into lower chambers. After 24 h, the chambers were soaked with 4% paraformaldehyde for 20 min and stained with 2% crystal violet for 1 h. The cells on the upper surface were cleaned. The fixed cells were imaged by microscope and analyzed by ImageJ.

Sphere formation assay

A total of 5 × 10⁴ HCC cells were seeded in ultra-low binding six-well dishes (Corning, 3471) and maintained in DMEM/F12 medium (VivaCell, C3130) supplemented with recombinant human FGF-basic (Peprotech, 100-18B), animal-free recombinant human EGF (Peprotech, AF-100–15), and N-2 supplement (ThermoFisher, 17502048). The spheres were imaged after 10 days.

Co-immunoprecipitation (Co-IP)

Cells treated were collected and lysed, followed by incubating with primary antibody at 4 °C overnight. The next day, Protein A/G Magnetic Beads (Selleck, B23201) were added to the antigen–antibody mixture and gently mixed at room temperature for 15 min. Then the supernatant was removed and the remaining beads were heated at 95 °C for 10 min with 1× loading buffer. Afterwards, the beads were removed through magnetic separation and the IP protein samples were examined by WB.

Immunofluorescence (IF) staining

HCC cells were treated with 4% paraformaldehyde for 15 min and rinsed with PBS three times. Then, QuickBlock™ Blocking Buffer for Immunol Staining (Beyotime, P0260) was used for blocking. After that, the dishes were processed with primary antibodies at 4 °C overnight, then stained with Alexa Fluor 488- and Cy3-conjugated secondary antibodies (Beyotime, A0423 and A0521). Finally, nuclei were stained with DAPI (Beyotime, P0131). Images were acquired by confocal microscope.

MG132, cycloheximide (CHX), and hydroxychloroquine (HCQ) treatment

HCC cells with stable knockdown of USP13 were treated with CHX (100 μg/ml, Selleck, S7418) and proteasome inhibitor MG132 (10 μM, Selleck, S2619) or lysosome inhibitor HCQ (20 μM, Selleck, E4824) for different lengths of time. The protein of the cells was extracted and subjected to WB.

Immunohistochemistry (IHC) analysis

Samples of HCC were obtained from the Department of Pathology of Zhejiang Provincial Hospital, selected from patients with untreated primary HCC. Paraffin-embedded tissue sections were dewaxed, rehydrated, and boiled. Then tissues were processed by hydrogen peroxide for 15 min and washed with PBS. Primary antibodies were applied to tissues and incubated at room temperature for 1.5 h. Immunodetection was carried out with the DAB kit (ZSGB-BIO, PV8000) according to the provided guidelines. Two independent observers determined the staining scores according to previously published literature [22].

Chromatin immunoprecipitation (ChIP)

ChIP assays were performed on MHCC97H cells with a ChIP Assay Kit (Beyotime, P2078) following the manufacturer’s instructions. A YY1 antibody was used to bind the promoter region of USP13. qRT-PCR was then conducted to verify the enrichment of the USP13 promoter region with targeted primers.

Dual luciferase reporter assay

For 24 h, HEK293T cells were simultaneously transfected with USP13-luc reporter, YY1, and pRL-TK plasmid. Luciferase assays were conducted with a dual luciferase reporter assay kit (Beyotime, RG027) following the provided guidelines.

In vivo tumorigenesis assay

Female BALB/C nude mice (5–6 weeks old) were purchased from Hangzhou Medical College. MHCC97H cells (4 × 10⁶) with a stable knockdown or overexpression of USP13, or control, were resuspended in DMEM and injected into the mice subcutaneously. The tumor volume (V) was determined by measuring the tumor’s length (L) and width (W) every 4 days with the formula V = 0.52 × L × W². For mice in given groups, indole-3-carbinol (I3C, MCE, HY-N0170) was intraperitoneally injected every 2 days in two different dosages (5% I3C, 95% DMSO, 20 mg/kg or 40 mg/kg) to suppress WWP1 activation [13]. The mice were euthanized over 20 days post-injection, and the tumors were removed.

Statistical analysis

The data from public databases was collected and analyzed by R 4.2.2. Statistical analysis was conducted using GraphPad Prism 8 software. All data are shown as mean ± SD from at least three independent experiments. Student’s t test (two-tailed) was used to compare the statistical significance between the two groups. One-way analysis of variance (ANOVA) tests were used to compare the statistical significance when there were three or more groups. The Fisher’s exact test is used for comparisons between qualitative data. A P-value less than 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001) was determined as statistically significant. “ns” indicated no significance.

Results

USP13 is highly expressed in HCC and predicts poor prognosis

To explore the expression of USP13 in tumor tissues, we analyzed USP13 expression through the Gene Expression Omnibus (GEO) and The Cancer Genomics Atlas (TCGA) databases, which revealed that USP13 mRNA was elevated in many human tumor tissues compared with normal tissues (Fig. 1A), including HCC (Fig. 1B, C; Supplementary Fig. S1A). To further verify the expression of USP13 in HCC, we tested USP13 expression in HCC and adjacent normal tissues obtained from Zhejiang Provincial People’s Hospital. The IHC results revealed upregulation of USP13 expression in HCC tissues compared with adjacent normal tissues (Fig. 1D, E), which was consistent with the above bioinformatics findings. The clinicopathological data of these 22 patients revealed that the expression level of USP13 was correlated with tumor size (P = 0.032, Supplementary Table S4). In the TCGA database, we also found that higher expression of USP13 was associated with worse overall survival (OS, Fig. 1F), disease-specific survival (DSS, Fig. 1G), and progression-free interval (PFI, Fig. 1H) in patients with HCC. In addition, we conducted receiver operating characteristic (ROC) analysis to assess the predictive precision of USP13 in predicting the outcome of patients with HCC, and the area under the curve (AUC) values exceeded 0.6 at both 1 year and 3 years (Supplementary Fig. S1B). A forest plot revealed that USP13 was a risk factor (Supplementary Fig. S1C). Furthermore, a nomogram was created to predict the OS of patients with HCC by incorporating the USP13 expression level, sex, grade, age, and stage for accurate and convenient quantification (Supplementary Fig. S1D). This nomogram will be useful for accurately assessing patient risk and enables us to make well-informed choices about future treatment plans. To validate the accuracy of the nomogram’s predictions, we utilized a calibration curve to compare the actual results with the predicted results (Supplementary Fig. S1E). In addition, we examined USP13 expression in different HCC cell lines. Compared with THLE-2, USP13 was highly expressed in the majority of HCC cell lines at both the mRNA (Fig. 1I) and protein levels (Fig. 1J, K). Therefore, USP13 is upregulated in HCC tissues and is an unfavorable risk factor for HCC.

Fig. 1

USP13 is highly expressed in HCC and predicts poor prognosis. A USP13 mRNA expression levels in different tumor tissues and normal tissues from the GEO and the TCGA databases were analyzed with Xena (http://xena.ucsc.edu/). The blue bars represent normal tissues and red bars represent tumor tissues. B USP13 mRNA levels in HCC and normal liver tissue was analyzed based on TCGA. C USP13 mRNA levels in HCC and normal liver tissue was analyzed based on the GEO (GSE54236) database. D Representative IHC staining images of USP13 in HCC and adjacent tissues. E Quantitative analysis of IHC staining results by ImageJ. F–H Prognostic information including overall survival (OS), disease-specific survival (DSS), and progression-free interval (PFI) were analyzed in the TCGA database. I USP13 mRNA levels in different HCC cell lines. J, K USP13 protein levels in different HCC cell lines, and quantified by ImageJ. Scale bars: 50 µm

USP13 is associated with HCC cell proliferation, migration, and stemness

To examine the function of USP13 in HCC, we introduced two shRNAs to silence USP13 and generated USP13 stable knockdown Huh7 and MHCC97H cells via infection with a lentivirus harboring USP13 shRNA (shUSP13). Both qRT-PCR and WB analysis confirmed the successful silencing of USP13 at both the mRNA and protein levels (Fig. 2A, B). CCK-8 (Fig. 2C, D) and colony formation (Fig. 2E) assays revealed that USP13 knockdown significantly suppressed HCC cell proliferation. Then, we used Transwell (Fig. 2F) and wound healing experiments (Fig. 2G) to determine the influence of USP13 on the migration ability of HCC cells. These results indicated that depletion of USP13 weakened the capacity of HCC cells to migrate. We also detected stemness-related protein expression and performed sphere formation assays to determine the effect of USP13 on the stemness of HCC cells. As shown in Fig. 2H, the sphere formation capacity of HCC cells drastically decreased after USP13 knockdown, as the sphere size decreased in USP13-knockdown group. Similarly, the expression of stemness-related proteins, including c-Myc, Nanog and Lin28B, was reduced in USP13-knockdown cells (Fig. 2I). Thus, USP13 deficiency suppressed HCC cell proliferation, migration, and stemness in vitro.

Fig. 2

USP13 was associated with HCC cell proliferation, migration, and stemness. A, B Huh7 and MHCC97H were infected with shNC, shUSP13-1, and shUSP13-2 lentivirus. qRT-PCR and WB were used to determine the mRNA and protein level of USP13. C, D CCK-8 and (E) colony formation assays were used to detect cell proliferation ability. F Transwell and (G) wound healing assays were used to detect cell migrate ability. H Spheroid formation assay was used to detect cell stemness. I Several stemness-related proteins expression was detected. Scale bars: 275 µm

To further validate the function of USP13 in HCC, we established USP13-overexpressing HCC cells via lentivirus (Supplementary Fig. S2A, B). As revealed by the CCK-8 (Supplementary Fig. S2C, D) and colony formation (Supplementary Fig. S2E) assays, USP13 overexpression enhanced HCC cell proliferation. According to the results of the Transwell assay (Supplementary Fig. S2F) and wound healing assay (Supplementary Fig. S2G), USP13 promoted HCC cell migration. Furthermore, USP13 overexpression enhanced the sphere formation ability (Supplementary Fig. S2H) and stemness-related protein expression (Supplementary Fig. S2I) of HCC cells. Taken together, these results indicate that USP13 serves as an oncogene to accelerate the malignant behaviors of HCC cells.

USP13 interacted with WWP1 in HCC cells

Our findings described above revealed the tumor-promoting role of USP13 in HCC. However, how USP13 accelerates HCC development remains to be elucidated. Through immunoprecipitation with an anti-USP13 antibody, WWP1 was identified as a new binding protein of USP13 (Fig. 3A). Immunoprecipitation with WWP1 antibody further validated the interaction between WWP1 and USP13 (Fig. 3B). Next, we transfected Myc-tagged WWP1 (Myc-WWP1), along with Flag-tagged USP13 (Flag-USP13) or an empty vector, into HCC and HEK293T cells, followed by a Co-IP assay to validate the binding between USP13 and WWP1. As shown in Fig. 3C, Myc-WWP1 was pulled down by the anti-Flag antibody. Flag-USP13 was also detected in the Myc-WWP1-binding complex with the Myc antibody (Fig. 3D).

Fig. 3

USP13 interacted with WWP1 in HCC cells. A Co-IP assay detected the endogenous interaction between USP13 and WWP1 with a USP13 antibody in HCC cells and HEK293T cells. B Co-IP results with a WWP1 antibody in HCC cells and HEK293T cells. C Co-IP assays showed the exogenous interaction between USP13 and WWP1 with a Flag antibody. D Co-IP results of the binding between USP13 and WWP1 with a Myc antibody. E The interaction between USP13 truncation mutants and WWP1. F Immunofluorescence indicated the colocalization of USP13 and WWP1 in Huh7 and MHCC97H cells. Scale bars: 50 µm

As previously reported, USP13 consists of a UBP domain, a USP domain, and two UBA domains [23]. To determine the specific region of USP13 that interacts with WWP1, we constructed five Flag-tagged USP13 truncation plasmids. These truncated plasmids were then co-transfected with Myc-WWP1 into HEK293T cells for Co-IP experiments. On the basis of our results, the UBP domain or two UBA domains of USP13, but not the USP domain, were essential for its binding to WWP1 (Fig. 3E). Moreover, the immunofluorescence staining results revealed that USP13 colocalized with WWP1, mainly in the nucleus of HCC cells (Fig. 3F). Overall, there was an interaction between USP13 and WWP1.

USP13 deubiquitinated and regulated WWP1 in a proteasomal-dependent way.

Currently, USP13 has been reported as a DUB that removes ubiquitin from substrates, thereby stabilizing its substrate protein [14]. Since our findings described above demonstrated the interaction between USP13 and WWP1, we wondered whether USP13 regulated WWP1 through deubiquitination. First, we analyzed the effect of USP13 on the expression of WWP1. An obvious decrease in WWP1 protein levels was observed in USP13-knockdown cells (Fig. 4A), while there was no change in the RNA level of WWP1 (Supplementary Fig. S3A and S3C). Moreover, USP13 increased the protein level (Fig. 4B) but not the RNA level of WWP1 (Supplementary Fig. S3B and S3D). Second, we co-transfected Myc-WWP1 and HA-ubiquitin (HA-Ub) plasmids into USP13-knockdown or -overexpressing cells to detect the ubiquitination of WWP1. The results indicated that USP13 knockdown increased, while USP13 overexpression reduced, the level of ubiquitinated WWP1 protein (Fig. 4C, D). Currently, two protein degradation methods have been identified, including proteasomal-dependent and lysosomal-dependent degradation [24]. To identify the manner of WWP1 degradation by USP13, we treated HCC cells with MG132 (a proteasome inhibitor) or HCQ (a lysosome inhibitor) along with USP13 knockdown. Our results indicated that MG132 could restrain the degradation of WWP1 induced by USP13 knockdown, whereas HCQ could not (Fig. 4E). Loss of USP13 markedly decreased the protein stability of WWP1 (Fig. 4F, G; Supplementary Fig. S3E, F). Together, these findings revealed that USP13 maintained WWP1 stabilization by restraining its proteasome degradation rather than the lysosome pathway.

Fig. 4

USP13 removed the K29- and K48-linked polyubiquitination chain of WWP1. A WB results of WWP1 expression upon USP13 knockdown in HCC cells. B WB results of WWP1 expression upon USP13 overexpression in HCC cells. C, D The ubiquitination of WWP1 was analyzed in HCC cells with stable USP13 knockdown or overexpression. E WWP1 protein level was detected in HCC cells treated with CHX and MG132 or HCQ. F CHX assays were conducted to detect the half-life of WWP1 in Huh7 shUSP13 and shNC cells. G Quantitative analysis of WWP1 protein expression by ImageJ. H, I WB results of the ubiquitination of WWP1 in HEK293T cells with co-transfection of Myc-WWP1, HA-ubiquitin, or mutations, and either the empty vectors or Flag-USP13

USP13 removed the K29- and K48-linked polyubiquitination chains of WWP1.

There are seven lysine sites in the ubiquitin protein (K6, K11, K27, K29, K33, K48, and K63). K48- and K63-linked ubiquitination are the most general polyubiquitination types that regulate the stability of substrates [25]. To determine the ubiquitin linkage patterns of WWP1 by USP13, we co-transfected Myc-WWP1 with HA-Ub mutants with all lysine mutations (K0), or all lysine mutations but one (K6, K11, K27, K29, K33, K48, or K63), along with or without Flag-USP13, into HEK293T cells (Fig. 4H). Our findings indicated that USP13 remarkably removed K29- and K48-linked polyubiquitination of WWP1. To further verify the involvement of K29- and K48-linked polyubiquitination of WWP1 by USP13, we co-transfected Myc-WWP1 with wild-type (WT) HA-Ub or HA-Ub with a single lysine mutation (K6R, K11R, K27R, K29R, K33R, K48R, or K63R), along with or without Flag-USP13, into HEK293T cells (Fig. 4I). As shown in Fig. 4I, the HA-Ub mutants K27R, K48R, and K63R attenuated the deubiquitination of WWP1 by USP13. Taken together, our findings demonstrated that USP13 effectively reversed the K29- and K48-linked ubiquitination of WWP1.

USP13 regulates HCC cell behaviors via WWP1

As an E3 ubiquitin ligase, WWP1 has been reported to interact with and then ubiquitinate the PTEN protein to suppress its dimerization and membrane recruitment without affecting PTEN expression, thus activating the Akt pathway [13]. Here, WWP1 was knocked down or overexpressed in HCC cells. The downregulation or upregulation of phosphorylated Akt (p-Akt) and phosphorylated mTOR (p-mTOR) levels was observed upon WWP1 knockdown or overexpression, whereas there was no change in PTEN expression (Supplementary Fig. S4A-B), which was consistent with previous reports [12, 13]. To explore whether WWP1 is necessary for the oncogenic function of USP13, we transfected WWP1 plasmids or vectors into cells with stable knockdown of USP13. The results revealed that USP13 knockdown significantly reduced the protein levels of WWP1, p-Akt, and p-mTOR, which was reversed by WWP1 overexpression (Fig. 5A). Conversely, we also transfected WWP1 siRNA or negative control siRNA into HCC cells overexpressing USP13 and found that WWP1 knockdown attenuated USP13-induced WWP1 expression and Akt/mTOR pathway activation (Fig. 5B). In addition, we also observed no change in PTEN expression after USP13 and WWP1 were knocked down or overexpressed (Fig. 5A, B). These results demonstrate that USP13 indirectly activates the Akt/mTOR pathway through WWP1.

Fig. 5

USP13 regulates HCC cell behaviors via WWP1. A WB assay detected the Akt/mTOR pathway-related protein expression levels with WWP1 overexpression in USP13-knockdown HCC cells. B WB assay detected the Akt/mTOR pathway-related protein expression levels with WWP1-knockdown in USP13-overexpression HCC cells. C CCK-8 and (D) colony formation assays were used to detect cell proliferation ability. E Spheroid formation assays were used to detect cell stemness ability. F, G Wound healing and (H, I) Transwell assays were used to detect cell migration ability. Scale bars: 275 µm

CCK-8 and colony formation assays revealed that WWP1 rescued the suppression of HCC cell proliferation caused by USP13 silencing (Fig. 5C, D). Furthermore, sphere formation (Fig. 5E), wound healing (Fig. 5F, G), and Transwell assays (Fig. 5H, I) revealed that the USP13-silencing mediated cell migration and stemness suppression were reversed by WWP1 overexpression. Conversely, WWP1 knockdown reduced the increase in cell proliferation, migration, and stemness induced by USP13 overexpression (Supplementary Fig. S4C–I). Overall, USP13 facilitates the malignant progression of HCC cells through the upregulation of WWP1.

Thus, we concluded that USP13 stabilized WWP1, and then activated the Akt/mTOR pathway. To explore the role of the Akt/mTOR pathway in the function of USP13, we employed pharmacological modulators of mTOR signaling, namely, rapamycin (Rapa, an inhibitor) and MHY1485 (MHY, an activator), both of which are well characterized agents with established specificity in preclinical models [26, 27]. Dose–response experiments in Huh7 and MHCC97H cell lines revealed that Rapa or MHY treatment obviously inhibited or activated the mTOR pathway (Supplementary Fig. S5A). Then, we treated USP13 or NC cells with Rapa. CCK-8 (Supplementary Fig. S5B) and colony formation assays (Supplementary Fig. S5C, D) revealed that Rapa significantly weakened the increase in cell proliferation induced by USP13. In contrast, MHY effectively increased HCC cell proliferation and rescued the cell growth inhibition induced by USP13 knockdown (Supplementary Fig. S5B–D). Taken together, our findings suggest that USP13 contributes to HCC progression via the WWP1/Akt/mTOR pathway.

USP13 regulates the tumorigenesis of HCC cells via WWP1 in vivo

To validate the role of USP13 in tumor growth in vivo, MHCC97H shUSP13 or shNC cells were harvested and then subcutaneously injected into nude mice (Fig. 6A). Our results revealed that the sizes of the tumors in the shUSP13 groups were notably reduced compared with those in the shNC group (Fig. 6B, C). In addition, NC and USP13 HCC cells were injected into nude mice. I3C is a naturally occurring compound with minimal toxicity that is widely found in cruciferous vegetables, and has previously been identified as an effective inhibitor of WWP1 both in vivo and in vitro [13, 28]. Therefore, we treated the USP13 group mice with I3C to suppress WWP1 activity (Fig. 6D). USP13 promoted the growth of HCC tumors, since tumors in the USP13 group were larger than those in the NC group. Moreover, I3C effectively attenuated the tumor-promoting ability of USP13, while no significant differences were observed between the two doses of I3C (Fig. 6E, F). The body weight changes of the nude mice in the different groups are shown in Supplementary Fig. S6A-B. Hematoxylin and eosin (HE) stainings of the tumors in Fig. 6B and E are shown in Supplementary Fig. S6C. IHC analysis revealed that the expression of the proliferative marker Ki67 was significantly decreased in the shUSP13 group but increased in the USP13-overexpressing group (Fig. 6G–J). In addition, I3C treatment also reduced Ki67 expression (Fig. 6I, J). Western blotting analysis further revealed the downregulation of WWP1 in the tumors of the shUSP13 group (Fig. 6K). In contrast, WWP1 was elevated in the USP13 group, and I3C suppressed the upregulation of WWP1 mediated by USP13 overexpression (Fig. 6L). In summary, we concluded that USP13 facilitates HCC tumorigenesis by upregulating WWP1 in vivo.

Fig. 6

USP13 regulated tumorigenesis of HCC cells via WWP1 in vivo. A The nude mice bearing tumors were euthanized and photographed. B Images of tumors from the nude mice in A. A ruler was used to indicate the size of tumors. C Tumor growth curves from the nude mice in A . D The nude mice bearing tumors were euthanized and photographed. E Images of tumors from the nude mice in D. A ruler was used to indicate the size of tumors. F Tumor growth curves from the nude mice in D . G, I IHC analysis of Ki67 expression in tumors shown in B and E. H, J Quantification and statistical analysis of the Ki67 IHC scores. K, L USP13 and WWP1 protein level in tumors from nude mice were detected by WB. Scale bars: 50 µm

YY1 transcriptionally activated USP13 and formed a carcinogenic YY1/USP13/WWP1 axis in HCC

To better understand the regulatory mechanism of USP13 in HCC, ALGGEN-PROMO, AliBaba2.1, Human TFDB, and GeneCards were used to investigate the transcription factors (TFs) that regulate USP13. Two TFs (YY1 and TBP) were revealed to bind to the promoter region of USP13 (Fig. 7A). Next, we analyzed the correlation between USP13 expression and YY1 or TBP expression in the TCGA cohort. The correlation between YY1 and USP13 (R = 0.656, P < 2.2 × 10⁻¹⁶) was greater than the correlation between TBP and USP13 (R = 0.376, P < 9.11 × 10⁻¹⁴) (Supplementary Fig. S7A, B). YY1 was selected for further study. Our results revealed that YY1 depletion in HCC cells led to a decrease in both the mRNA and protein levels of USP13 (Fig. 7B, C) and that YY1 overexpression upregulated USP13 expression (Supplementary Fig. S7C, D). JASPAR was used to predict the binding region of the USP13 promoter with YY1 [29], and three YY1 binding motifs located within the −2900 to 100 bp range of the USP13 promoter region were identified (Supplementary Fig. S7E, F). To verify the binding sites of the USP13 promoter to YY1, we designed three primers that cover the three predicted binding regions (Supplementary Table S2) and conducted qRT-PCR after the ChIP assay. We found an obvious enrichment of motif 3 (−143 to 81 bp) in the YY1 group (Fig. 7D). To verify the binding site located in the USP13 promoter region with YY1, we developed two luciferase reporter vectors with truncated promoters. The luciferase reporter pGL3 motif 3 (−900 to 100 bp), which contains motif 3 but not motif 1 or motif 2, was strongly activated by YY1 expression, which was significantly decreased when motif 3 was mutated in the pGL3-motif 3-mutant group (Fig. 7E). Moreover, YY1 inhibition could suppress WWP1 expression, which was reversed by USP13 expression and vice versa (Fig. 7F, G), leading to molecular regulation of the YY1/USP13/WWP1 axis. To further confirm the expression patterns of YY1, USP13, and WWP1, IHC staining was performed on HCC samples from our hospital, and representative IHC images are shown in Fig. 7H. The statistical results confirmed that the expression of YY1, USP13, and WWP1 was positively correlated with each other (Fig. 7I–K). We multiplied the IHC scores of three proteins for each sample and divided the 28 samples into high- and low-expression groups. Statistical analysis of the clinical data revealed that patients in the high-expression group had larger tumor volumes (P = 0.029) and a greater incidence of microvascular invasion (P = 0.026, Supplementary Table S5).

Fig. 7

YY1 transcriptionally activates USP13 and forms a carcinogenic YY1/USP13 axis in HCC. A The transcriptional factors that regulate USP13 were predicted by PROMO, AliBaba2.1, Animal TFDB, and GeneCards. B, C WB and qRT-PCR results of USP13 expression after knockdown or overexpression of YY1. D ChIP-qPCR analyzed the binding motifs of USP13 promoter with YY1 protein in MHCC97H cells. E The binding motif 3 of the USP13 promoter was mutated, and the relative luciferase activity from wild-type and mutant-type-transfected cells was measured by a dual luciferase reporter system. F WWP1 proteins were detected by WB with overexpression of USP13 in YY1-knockdown HCC cells. G WWP1 proteins were detected by WB with knockdown of USP13 in YY1-overexpression HCC cells. H Representative IHC staining image of YY1, USP13 and WWP1 in HCC tissue were displayed. I The correlation between YY1 and USP13 were analyzed based on IHC score. J The correlation between USP13 and WWP1 was analyzed on the basis of IHC score. K The correlation of between YY1 and WWP1 was analyzed according to IHC score. Scale bars: 50 µm

Furthermore, YY1 has been reported to be upregulated in many cancers, which leads to poor prognosis [30]; here, we wondered whether YY1 promotes HCC progression via USP13. The results of the CCK-8 assay confirmed that YY1 depletion weakened the proliferation of HCC cells and that USP13 could eliminate or even reverse this effect (Fig. 8A). A colony formation assay confirmed the necessity of USP13 for the oncogenic role of YY1 (Fig. 8B, C). We subsequently used wound healing experiments to explore whether the migration of HCC cells was affected by YY1, and the results revealed that YY1 accelerated HCC cell migration via USP13 (Fig. 8D, E).

Fig. 8

YY1 regulates HCC cell proliferation and migration via USP13. A CCK-8 assay of YY1-knockdown in USP13-overexpression and YY1-overexpression in USP13-knockdown HCC cells were used to detect cell proliferation ability. B, C Colony formation assays confirmed the difference in cell proliferation. D, E Wound healing assays were used to detect cell migration ability of the above cells. Scale bars: 275 µm

Together, our findings revealed that YY1 transcriptionally upregulated USP13 expression, indirectly increased WWP1 protein expression and promoted HCC progression.

Discussion

HCC is the predominant histologic subtype of liver cancer and accounts for 90% of primary liver cancers, which is known for its high rates of both mortality and morbidity [1, 2]. Therefore, identification of the regulatory mechanism of HCC progression and exploration of potential therapeutic targets for patients with HCC are crucial. Our current study revealed that USP13 expression was increased in HCC samples, indicating a poorer prognosis and a larger tumor size in patients with HCC. Mechanistically, WWP1 was further demonstrated to be a substrate of USP13. By acting as a DUB, USP13 can bind to and then deubiquitinate and stabilize the WWP1 protein, thus activating the Akt/mTOR pathway. Finally, our functional experiments in HCC cells and a subcutaneous xenograft model further revealed the oncogenic role of the USP13-WWP1 axis both in vitro and in vivo.

At present, many studies have demonstrated the oncogenic role of USP13 as a DUB in human cancers. For example, USP13 removes ubiquitin from the c-Myc protein and then stabilizes and upregulates c-Myc expression, ultimately facilitating the malignant progression of both lung squamous cell carcinoma and glioblastoma [31, 32]. Xie et al. revealed that USP13 can deubiquitinate and stabilize ZHX2 to promote the tumorigenesis of clear cell renal cell carcinoma [33]. Morgan et al. further revealed that USP13 enhances cervical cancer cell growth via the deubiquitination of Mcl-1 [34]. USP13, which acts as an oncogene, was shown to promote breast cancer metastasis by deubiquitinating Twist 1 [35]. However, the tumor suppressive role of USP13 in breast cancer has also been explored. Zhang et al. reported that USP13 suppresses tumorigenesis and glycolysis in PTEN-positive breast cancer through the deubiquitination of PTEN [18]. In addition to its role in breast cancer, USP13 has been demonstrated to suppress tumor development via PTEN in oral squamous cell cancer [36] and bladder cancer [19]. Therefore, USP13 plays a dual role in human cancers that is largely dependent on the functional properties of its substrates. In our current study, an oncogenic role of USP13 in HCC cell growth, migration, and stemness was identified.

WWP1, a ubiquitin E3 ligase, has been shown to be involved in several human cancers, including breast cancer, intrahepatic cholangiocarcinoma, non-small cell lung cancer, prostate cancer and acute myeloid leukemia, through ubiquitination of its protein substrates [20]. However, the above studies focused mainly on the function of WWP1 as an E3 ligase, and the regulation of WWP1 protein deubiquitination has not been elucidated. Here, we revealed that USP13, which acts as a DUB, deubiquitinates WWP1, which was identified as a new substrate of USP13. According to our results, USP13 truncations containing either the N-terminal region (aa1–300 and aa1–624) or the C-terminal region (aa301–863 and aa625–863) could bind to WWP1. Similar to our results, a previous study revealed that the complete N-terminal (aa1–624) or C-terminal (aa625–863) region of USP13 is necessary for the interaction between USP13 and STING [37]. Gao et al. and Shin et al. reported that USP13 with a C-terminal region (aa301–863 and aa337–863) has the ability to interact with ATG5 and HMGB1 [17, 38]. Thus, it seems that the two-terminal regions of USP13 mainly mediate the binding of USP13 to its substrates.

At present, there are seven lysine sites in ubiquitin proteins (K6, K11, K27, K29, K33, K48, and K63), indicating different ubiquitin-linked regulation of substrates [39]. Among the above polyubiquitination types, K48- and K63-linked polyubiquitination of substrates were the most common and has been extensively reported. Here, through site mutation of the lysines of ubiquitin, we further revealed that USP13 specifically removed the K29- and K48-linked polyubiquitination of WWP1. To date, two protein degradation methods have been identified: proteasomal-dependent and lysosomal-dependent degradation. K48-linked polyubiquitination of proteins is usually subjected to proteasome degradation, whereas K29-linked polyubiquitination of proteins is reportedly under lysosomal degradation [40,41,42]. In HCC, emerging evidence has demonstrated that the deubiquitinating enzyme zinc finger, RAN-binding domain containing 1 (ZRANB1) selectively removes K29- and K33-linked polyubiquitin chains on UV radiation resistance-associated gene (UVRAG), increases the binding of UVRAG to the rubicon autophagy regulator (RUBCN), and ultimately suppressing autophagic flux and promoting HCC growth [43]. Recently, K48-linked polyubiquitination of proteins was also associated with lysosomal-mediated degradation [44, 45]. Our results revealed that USP13 suppression induced degradation of WWP1 could be rescued by co-treatment with MG132 but not with HCQ, indicating that USP13-mediated polyubiquitination of WWP1 was proteasome dependent. Next, the role of K29-linked polyubiquitination of WWP1 needs to be further determined. Moreover, the critical lysine residues on the WWP1 protein affected by USP13 should also be identified.

In addition, previous studies revealed that WWP1 can activate Akt by promoting PTEN K27 ubiquitination without influencing PTEN expression [12, 13]. However, Zhang et al. revealed that USP13 can deubiquitinate and stabilize the PTEN protein, thus inhibiting the Akt pathway and suppressing breast cancer progression [18]. Here, we found that USP13 did not affect PTEN expression, but could activate the Akt/mTOR pathway via WWP1 in HCC cells, which was consistent with the findings of Lee et al. [12, 13]. The AKT/mTOR pathway is a classic pathway that governs essential biological functions under normal conditions and plays a pivotal role in various cancer-related processes, such as proliferation, metabolism, apoptosis resistance, and migration [46, 47]. Next, we demonstrated the molecular regulation of the USP13/WWP1/Akt/mTOR axis. Our further functional rescue results suggested that USP13 facilitated the malignant behaviors of HCC cells through upregulating WWP1 and activating the Akt/mTOR pathway.

Previous studies have identified various regulatory mechanisms of USP13, including METTL3 and IGF2BP2-dependent N6-methyladenosine modifications of USP13 [17] and CK2-induced phosphorylation of USP13 [48]. In addition, Guo et al. reported that RREB1 is a TF of USP13, that promotes USP13 expression [49]. However, there is still little research on how USP13 is transcriptionally modified. To identify other transcription factors of USP13, we amalgamated four databases and identified YY1 and TBP as new TFs of USP13. Moreover, the correlation of YY1 with USP13 was greater than that of TBP with USP13 in the TCGA cohort (R = 0.656 versus R = 0.376). In addition, studies on the function of YY1 in cancers are comprehensive and thorough, but the relationship between YY1 and USP13 has not yet been elucidated. YY1, a member of the GLI-Krüppel family [50], is a zinc-finger protein highly expressed in various tumors including HCC. YY1 enhances aggressive behaviors, including cell angiogenesis, by transcriptionally activating vascular endothelial growth factor A (VEGFA) [51]. YY1 can also cooperate with centromere protein A (CENPA) to transcriptionally activate cyclin D1 and neuropilin 2 (NRP2), ultimately driving HCC tumorigenesis [52]. Moreover, YY1 enhances HCC cell resistance to sorafenib by upregulating epidermal growth factor receptor (EGFR) [53]. In addition, YY1 can bind to and activate Linc01608, thus facilitating HCC progression by activating the EGFR/ERK pathway [54]. Conversely, YY1 reportedly serves as a suppressor to inhibit PGC-1β expression, ultimately promoting fatty acid oxidation and tumorigenesis in HCC [55]. Here, we showed that YY1 expression was positively correlated with USP13 expression in HCC tissues, as revealed by IHC. Moreover, ChIP analysis revealed that YY1 bound to the motif 3 region in the USP13 promoter. Next, a dual luciferase reporter assay demonstrated that YY1 bound to the AAAAATAGCGTC sequence in motif 3 and activated USP13 transcription. We then confirmed the regulation and correlation of the YY1/USP13/WWP1 signaling axis via western blotting and IHC. In addition, functional rescue assays revealed that YY1 facilitated HCC malignant behaviors through USP13. Taken together, our findings indicate that YY1 is a new TF of USP13 and supports the oncogenic YY1/USP13/WWP1 signaling axis in HCC.

Notably, USP13 has been demonstrated to possess anti-inflammatory properties [56] and to exert dual regulatory effects on host antiviral responses [37, 57]. However, its pathophysiological function across the hepatitis B (HBV) infection–chronic hepatitis–HCC carcinogenic continuum remains poorly characterized. Emerging evidence from multiple studies links USP13 to tumor immune microenvironment modulation through immune cell infiltration [58] and metabolic reprogramming [16, 59], highlighting the imperative for comprehensive mechanistic investigations to elucidate its precise contributions to HCC pathogenesis.

There were limitations in this study, including a lack of prognostic survival information for patients with HCC, a small HCC cohort (n = 22, n = 28), and an imbalance in tumor stage distribution, since most of the HCC tissues used in our study were defined as early-stage HCC, as shown in Supplementary Tables S4 and S5. Furthermore, the high prevalence of HBV-positive cases (90.9% and 89.3%) in Chinese patients with HCC also affected the statistical results (Supplementary Tables S4 and S5). To address these limitations, prospective multicenter validation with expanded cohorts encompassing diverse tumor stages and comprehensive clinical information and long-term prognostic follow-up will be valuable for further translational studies.

Conclusions

In this study, we first identified WWP1 as a novel protein substrate of USP13 that can remove K29- and K48-linked polyubiquitination chains and is stabilized by USP13. By acting as an oncogene, USP13 affects HCC malignant progression through the WWP1-mediated Akt/mTOR pathway. Moreover, YY1 was revealed to act as a transcription factor of USP13 through binding to the motif 3 region of the USP13 promoter. Finally, we validated the regulation of the YY1/USP13/WWP1 signaling axis in HCC cells and tissues (Fig. 9). Our findings described above indicate that the oncogenic YY1/USP13/WWP1 axis might serve as a potential therapeutic target for HCC treatment.

Fig. 9

The mechanistic scheme of this study