2.1 Clinical specimens

All human primary HCCs, paired peritumoral liver tissues and nontumor liver tissues were collected from the Department of Hepatobiliary Surgery, the Affiliated Drum Tower Hospital of Nanjing University Medical School. Among them, 49 pairs of tumor and paired peritumoral liver tissues were collected to detect the mRNA levels of UBE2S, and 8 of 49 were collected to detect protein levels of UBE2S. None of the patients received any other antitumor treatment before surgery such as chemotherapy, radiotherapy, targeted therapy. The use of human tumor samples was approved by the Ethics Committee of the Affiliated Drum Tower Hospital of Nanjing University Medical School.

2.2 Cell lines and cell culture

The LO₂ (human embryonic liver) cell line and the Huh7 (human HCC) cell line were purchased from the Cell Bank of Xiangya Central Laboratory, Central South University. Three liver cancer cell lines, HCCLM3, MHCC-97 L, and MHCC-97H, were gifts from the Liver Cancer Institute, Zhongshan Hospital, Fudan University. SMMC-7721, HepG2, Bel-7402, Hep3B, PLC, and Li-7 cell lines were obtained from the cell bank of the Chinese Academy of Sciences. HCCLM3, MHCC-97 L and MHCC-97H, HepG2, Hep3B, SMMC-7721, PLC, Huh7, and Li-7 cells were cultured in complete DMEM (Wisent, Inc.). LO2 and Bel-7402 cells were cultured in complete RPMI-1640 medium (Wisent, Inc.). All cells were supplemented with 10% fetal bovine serum (Wisent, Inc.) at 37°C in a humidified incubator with 5% CO₂.

2.3 Western blot analysis

Cells were lysed with NP40 solution (Beyotime Biotechnology) containing 1% protease inhibitors (Thermo Fisher Scientific). A BCA protein assay kit (Thermo Fisher Scientific) was used to quantify proteins. Subsequently, the proteins were boiled in loading buffer, fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and then transferred onto polyvinylidene difluoride (PVDF) membranes. After that, they were stained with primary antibodies overnight at 4°C, and the protein bands were incubated with horseradish peroxidase (HRP)-conjugated secondary antibody and detected using enhanced chemiluminescence (EMD Millipore, Billerica, MA, USA) with a Tanon 5200 Chemiluminescent Imaging System (Tanon Science and Technology Co., Ltd.). Antibodies against UBE2S, cyclin B, P-CDC20, CDC20, E-cadherin, N-cadherin, vimentin, Snail, JAK2, p-JAK2, STAT3, p-STAT3, and p21 were purchased from Cell Signaling Technology. Antibodies against VHL and HIF-1α were purchased from Proteintech (Proteintech Group, Inc.). Antibodies against cyclin A and cyclin E2 were purchased from Santa Cruz Biotechnology (Santa Cruz Biotechnology). The dilution of the primary antibodies was 1:1000. The dilution of anti-GAPDH (Bioworld Technology, Inc.) was 1:10,000.

2.4 Quantitative real-time polymerase chain reaction analysis (qRT–PCR)

The primer sequences of human UBE2S and GAPDH were as follows: UBE2S: (5′-3′) CGACACGTACTGCTGACCAT, (5′-3′) GCCGCATACTCCTCGTAGTT, and GAPDH: (5′-3′) CCATGTTCGTCATGGGTGTGAACCA, (5′-3′) GCCAGTAGAGGCAGGGATGATGTTC. Total RNA was extracted through TRIzol® reagent (Takara Bio, Inc.) and converted to cDNA with PrimeScript RT Master Mix (Takara Bio, Inc.). SYBR Premix Ex Taq II (Takara Bio, Inc.) was used to detect the relative mRNA expression levels, which was calculated by the 2^−ΔΔCq method with GAPDH.

2.5 Immunohistochemistry (IHC)

Protein levels were detected by IHC as previously described.¹⁷ Human HCC specimens and HCCLM3 tumor tissues were collected from nude mice and fixed in 10% formalin. The tumor tissues were dehydrated in an ascending ethanol series, then cleared with xylene and embedded in paraffin. They were incubated with the primary antibody at 4°C overnight, after tumor tissue sections were prepared. The tumor tissue sections were washed with PBST three times. Then, they were inoculated with HRP-conjugated secondary antibody and developed using the 3,3′-diaminobenzidine (DAB) substrate. The tissue sections were further counterstained with haematoxylin. Stained sections were examined and photographed via an Olympus BX51 microscope. The percentage of positively stained cells was scored as 0: 0%; 1: 1%–25%; 2: 26%–50%; 3: 51%–75%; 4: 76%–100%. The intensity was scored as 0: negative staining; 1: weak staining; 2: moderate staining; 3: strong staining. The results were performed manually by two experienced pathologists. The product of the staining intensity score and the percentage score was scored as 0: 0, 1–4: +, 5–8: ++, 9–12: +++. According to the UBE2S levels in HCC tissue detected by IHC, 103 patients with HCC were divided into two groups: the high UBE2S expression group (score: ++, +++, n = 57) and the low UBE2S expression group (score: 0, +, n = 46). The primary antibodies used for IHC were anti-UBE2S (ab197945, 1:200, Abcam), anti-p-JAK2 (ab32101, 1:200, Abcam), anti-VHL (16538-1-AP, 1:400, Proteintech), anti-HIF-1α (20960-1-AP, 1:200, Proteintech), anti-p21 (2947, 1:200, Cell Signaling Technology), anti-Ki67 (9027, 1:400, Cell Signaling Technology), and anti-p-STAT3 (9145, 1:100, Cell Signaling Technology). The secondary antibody for IHC was Super Sensitive TM IHC Detection System Kit (BD5001, Bioworld). The integral optical density (IOD) from IHC staining of anti-VHL, anti-HIF-1α, anti-p21, anti-Ki67, anti-p-STAT3, and anti-p-JAK2 was calculated using Image-Pro Plus software, and three samples were used for statistics.

2.6 Cell proliferation assay

Cells were plated into 96-well plates at the same concentration of 5 × 10³ cells/well for equal amounts of cell seeding. After culturing for 48, 72, and 96 h, MTT solution (5 mg/mL, 20 μL/well) was added to each well for 4 h. Then, 150 μL of DMSO was added to each well to dissolve the insoluble crystals. A spectrophotometer was used to measure the absorbance at 570 nm. Each assay was carried out at least three times.

2.7 Colony formation assay

Cells were seeded into 6-well plates (500 cells/well) for 2 weeks. After colony formation, the medium was discarded. After washing with PBS three times, the cells were fixed with methanol and stained with crystal violet for 15 min. The cell colonies were photographed and counted.

2.8 Wound healing assay

Wound healing culture inserts (Ibidi) were used to count migration capacity. First, the cells were seeded into each well of a culture insert (40,000 cells/well). After incubation overnight, the culture insert was removed and were carefully washed with PBS. Then, they were cultured with DMEM or RPMI-1640 medium without FBS. The cells migrated into the cell-free area, and the migration was monitored and photographed with an Olympus BX51 microscope at 24, 48, and 72 h. ImageJ software was used to calculate the wound closure rate. All experiments were repeated three times.

2.9 Transwell assay

Transwell inserts (Merck Millipore) were used to detect the migration capacity. The upper chamber with or without Matrigel (354230) coating (BD) for 30 min at 37°C for invasion and migration tests, respectively, was seeded with 20,000 HCC cells resuspended in DMEM without FBS. Then, the lower chamber was added with 0.8 mL DMEM supplemented with 10% FBS. At the end of the experiment, the medium was removed. The upper chamber membrane was fixed with methanol. The cells on the upper side were removed via a cotton swab. Then, the transwell inserts were stained with crystal violet. The cells on the opposite side of the membrane were photographed and counted in six random fields with an Olympus BX51 microscope.

2.10 Immunofluorescence

Round coverslips were placed into 24-well plates, and cells were seeded onto the coverslips. After 24 h, the cells were fixed with 4% formaldehyde and permeabilized for 30 min. Then, they were washed with PBS three times, and 500 μL of 0.2% Triton X-100 was added to the 24-well plate for 15 min. After the cells were blocked with 1% BSA for 1 h, they were incubated with primary antibody overnight at 4°C. Then, they were washed with PBST three times, and the cells were incubated with FITC-conjugated goat anti-rabbit IgG (Zenon® Alexa Fluor® 546 Rabbit IgG Labelling Kit, Thermo Fisher Scientific) for 2 h. After washing with PBS three times, the cells were incubated with DAPI (Thermo Fisher Scientific) for 15 min before being mounted with ProLong Gold Antifade Reagents. Finally, the cells were photographed through a fluorescence confocal microscope (500 IX71, Olympus). The primary antibodies used for immunofluorescence were anti-VHL (16538-1-AP, 1:400, Proteintech), anti-HIF-1α (20960-1-AP, 1:200, Proteintech), anti-vimentin (5741, 1:200, Cell Signaling Technology), and anti-E-cadherin (14472, 1:400, Cell Signaling Technology).

2.11 Small interfering RNA (siRNA) and plasmid transfection

The sequences of human UBE2S siRNA (siUBE2S) and VHL siRNA (siVHL) were as follows: siUBE2S sense (5′-3′): GACACGUACUGCUGACCAUTT and siVHL sense (5′-3′): CCAAUGGAUUCAUGGAGUA, which were purchased from GenePharma. Lipofectamine® RNAiMAX (Thermo Fisher Scientific) was used to transfect cells according to the manufacturer's instructions.

The pcDNA3.1-UBE2S, -VHL, and -p21 plasmids were synthesized by GeneChem (Shanghai GeneChem Co., Ltd.). The pcDNA3.1 empty vector was used as a negative control. The plasmids were delivered into cells via Lipofectamine 3000 transfection reagent (Thermo Fisher Scientific). The cells were transfected at a density of 80%–90%.

2.12 Cell cycle analysis

Cells (5 × 10⁵) were seeded into 6-well plates. The cells were harvested and fixed with cold 70% ethanol overnight at −20°C after 24 h. Then, after washing with PBS, they were stained in a solution with propidium iodide (PI) (0.5 mg/mL) and RNase A (10 mg/mL). Flow cytometry (FACSCalibur, BD) was used to confirm the cell cycle qualitatively.

2.13 Apoptosis assay

Cells (1 × 10⁶) were seeded into 6-well plates. After 24 h, the cells were collected including apoptotic, dead, and adherent cells. Then, the cells were washed with cold PBS and resuspended in annexin V-FITC binding buffer. Apoptosis was detected by flow cytometry (FACSCalibur, BD) after incubation with annexin V-FITC/PI. A TUNEL Apoptosis Assay Kit (Roche) was used for TUNEL staining according to the manufacturer's instructions.

2.14 HCC cells with stable UBE2S knockdown and overexpression

Lentivirus vector targeting UBE2S was synthesized by Shanghai GeneChem Company, Ltd. (GeneChem). Transfection efficiency was detected via GFP in the lentiviral vectors. Cells were transfected with constructs of the UBE2S firefly luciferase reporter for 48 h in order to stably knockdown or overexpress corresponding genes. After approximately 80% of the cells were transfected, puromycin (1 μg/mL) was added to the cells for approximately 2 weeks to generate stably transfected cells.

2.15 Coimmunoprecipitation (co-IP) assay

For Co-IP, the cells were lysed with NP40 (Beyotime Biotechnology) solution containing 1% protease inhibitors. Then, the cell lysates were incubated with antibodies against UBE2S, VHL, or IgG at 4°C overnight, and then incubated with protein A + G Sepharose beads (Beyotime Biotechnology) for another 1 h. Eluted proteins were resolved via 12% SDS–PAGE, and detected by western blotting with antibodies of UBE2S, VHL, or IgG.

2.16 Animal models

All animal experimentation were approved by the Animal Care Committee of Nanjing University in accordance with the guidelines of the Institutional Animal Care and Use Committee. The subcutaneous xenograft model in nude mice was performed as previously described.^{24, 25} HCCLM3 cells (control-shRNA group and UBE2S-shRNA group) and Bel-7402 cells (control group and UBE2S overexpression group) were resuspended with PBS and then injected into the right axillary region of 4- to 5-week-old male nude mice (2 × 10⁶ cells in 200 μL of PBS per mouse). The longest (L) and shortest (W) tumor diameters and mouse body weights were measured. Tumor volume (mm³) = 1/2 × L × W2. All mice were sacrificed 25 days after the subcutaneous HCCLM3 cell injection and 16 days after the subcutaneous Bel-7402 cell injection, and the tumor weights were measured. The tumors were fixed in 10% phosphate-buffered formalin for IHC staining. For sorafenib treatment, there were three groups (control-shRNA, control-shRNA+sorafenib, and UBE2S-shRNA+sorafenib). Seven days after the subcutaneous cell injection, the control-shRNA+sorafenib and UBE2S-shRNA+sorafenib groups were intraperitoneally injected with sorafenib (50 mg/kg) daily for 4 weeks. 0.4% dimethyl sulfoxide (DMSO) in PBS was used as negative control. The volumes were measured every 7 days, and after 35 days, all mice were sacrificed.

Orthotopic xenograft tumor models were created as previously described.^{26, 27} The HCCLM3 tumor tissues, which were obtained from the subcutaneously transplanted tumor model in the nude mice, were cut into small pieces (1 × 1 × 1 mm). Then, they were implanted into the livers of nude mice to develop the transplanted tumor model (control-shRNA group and UBE2S-shRNA group). All mice were sacrificed 8 weeks after the orthotopic xenograft tumor model was formed, and the tumor weight and volume were measured. The lungs and livers were harvested for histopathological examination. For sorafenib treatment, there were four groups (control-shRNA, UBE2S-shRNA, control-shRNA+sorafenib, and UBE2S-shRNA+sorafenib). Seven days after the transplanted tumor model developed, the control-shRNA+sorafenib and UBE2S-shRNA+sorafenib groups were intraperitoneally injected with sorafenib (50 mg/kg) daily for 25 days, and 0.4% dimethyl sulfoxide (DMSO) in PBS was intraperitoneally injected to the control group. The volumes and weights of the tumors were measured. All mice were sacrificed through cervical dislocation at the indicated time points. Inhalation anaesthesia with isoflurane was used in all experimental procedures using animals.

2.17 Haematoxylin and eosin (H&E) staining

Lungs obtained from the subcutaneously transplanted tumor model were fixed in 4% paraformaldehyde. The tissue sections were embedded in paraffin and stained with H&E as previously described.²⁸ The metastatic nodules in the lungs were photographed and counted using an Olympus BX51 microscope.

2.18 Bioinformatics analysis

The RNA-seq and clinical data of different cancer categories were obtained from the TCGA database (https://gdc.cancer.gov, up to January 28, 2016). The microarray datasets (GSE14520, GSE54236) were downloaded from GEO (http://www.ncbi.nlm.nih.gov/geo/). The level of UBE2S was analyzed by a log2-fold-change. Average gene expression was calculated for all involved genes for one sample to analyse the expression level of the pathway gene. The average value represents the sample's average gene expression of the pathway as previously reported.

2.19 Statistical analysis

All experimental data were presented as the mean ± standard deviation and analyzed using Student's t-test. Patient survival curves were generated via the Kaplan–Meier method. The correlation between different protein levels was determined by Spearman's rank analysis. When p < 0.05, the data were considered significantly different.

3.1 UBE2S was overexpressed in human HCCs and associated with poor prognosis

The mRNA levels of UBE2S in 42 pairs of HCC tissues and matched peritumoral tissues were first detected by RT–PCR. The results indicated that upregulation of UBE2S mRNA, log2 (the concentration of UBE2S mRNA in tumor tissue/that in matched peritumoral tissue) > 1, was detected in 29 of 42 (69%) HCCs (Figure 1A). The upregulation of UBE2S mRNA was also confirmed by analysis of TCGA data, GSE14520 and GSE54236 in HCC (Figure 1B). In addition, we also analyzed 49 pairs of HCC tumor tissues and their matched peritumoral tissues from the TCGA database (Figure 1C), which also had similar results to those illustrated in Figure 1A. Moreover, we randomly selected eight pairs of tumors to detect the protein level of UBE2S in tumor tissues, measured by western blotting. The results indicated that the protein levels of UBE2S were significantly upregulated in HCC tissues (Figure 1D). Spearman correlation analysis showed that there was a positive correlation between the UBE2S and Ki-67 mRNA expression levels, which was used as a proliferation marker for human tumor cells (Figure 1E). It indicated that upregulation of UBE2S may promote cell proliferation in HCC. Then, we measured the protein levels of UBE2S in 103 pairs of HCC tissues by IHC. The protein levels of UBE2S were higher in HCC tumor tissues than that in peritumoral liver tissues (Figure 1F). In addition, the results showed that HCC patients with high UBE2S expression had shorter overall survival and disease-free survival (Figure 1G). We also analyzed the associations between clinicopathological parameters and UBE2S expression levels. The results showed that the high-expression group had higher clinical stages (p = 0.021, Table S1) and larger tumor size (p = 0.032, Table S1), compared with the low UBE2S expression group. However, there was no significant difference in other clinical features between the two groups, including sex, age, tumor number, HBV and AFP level. Moreover, we further measured the mRNA and protein levels of UBE2S in LO₂ and HCC cell lines. We found that the UBE2S levels in the HCC cell lines were significantly higher than those in LO2 cells (Figure 1H,I). In conclusion, these results showed that UBE2S is overexpressed in HCC, which suggested that UBE2S is associated with HCC development. UBE2S may be a potential prognostic biomarker for HCC. These results were consistent with previous studies.¹⁷

3.2 UBE2S promoted HCC cell proliferation, migration, and invasion in vitro

To further investigate the role of UBE2S in HCC cell proliferation, migration, and invasion, we downregulated the expression level of UBE2S through siRNA transfection (siUBE2S and sicontrol) and overexpressed UBE2S via plasmid transfection (pUBE2S and MOCK). The choice of those specific cells to perform UBE2S downregulation or upregulation depended on the UBE2S levels in the HCC cell lines (Figure 1H). HCCLM3 and Huh7 cells with high expression levels of UBE2S were used to downregulate the expression of UBE2S, while Hep3B and Bel-7402 cells with low expression levels of UBE2S were used to upregulate the expression of UBE2S. First, we detected the interference and overexpression efficiency of UBE2S in HCC cell lines through western blotting and RT–PCR (Figure 2A,B). The reason for the two UBE2S western blot bands may be that UBE2S degradation is stimulated in HCC cells after UBE2S overexpression (partial degradation), and the protein that has not been completely degraded could still bind to the antibody. Then, the effects of UBE2S on cell proliferation were analyzed by an MTT assay. The results indicated that at 48 h, no significant difference in cell proliferation was observed between the siUBE2S and sicontrol groups. However, the cell proliferation rate was significantly inhibited after downregulation of UBE2S at 72 and 96 h. In addition, overexpression of UBE2S in Hep3B and Bel-7402 cells induced faster cell growth than that in control cells at 48, 72, and 96 h. Over time, the effect of UBE2S on cell proliferation was increased. The results showed that the cell proliferation rate was significantly inhibited after downregulation of UBE2S (Figure 2C) and increased after overexpression of UBE2S (Figure 2D). A colony formation assay also indicated that after downregulation of UBE2S, the number of clones was decreased (Figure 2E) and that after upregulation of UBE2S, the number of clones was increased (Figure 2F). Furthermore, wound healing assays and transwell assays were used to explore the potential role of UBE2S in the regulation of HCC cell migration. We found that knockdown of UBE2S decreased the wound healing rate (Figure 3A), while overexpression of UBE2S increased the wound healing rate (Figure 3B). Similarly, the transwell assays showed that downregulation of UBE2S caused fewer cells to pass through the membranes (siUBE2S) than did those of the control group (sicontrol) (Figure 3C), while overexpression of UBE2S (pUBE2S) increased the number of invasive cells compared with the control group (MOCK) (Figure 3D). In addition, knockdown of UBE2S decreased HCC cell invasion, while overexpression of UBE2S increased HCC cell invasion (Figure 3E). Western blotting (Figure 3F,G) and immunofluorescence assays (Figure 3H,I) demonstrated that downregulation of UBE2S upregulated the expression levels of E-cadherin (the epithelial cell marker) and downregulated the expression levels of vimentin and N-cadherin (the mesenchymal cell markers), while overexpression of UBE2S had the opposite results. In addition, UBE2S also regulated the EMT transcription factor snail (Figure 3F,G). Taken together, these results showed that in HCC, UBE2S could promote the malignant properties by increasing cell proliferation, migration, and invasion in vitro.

3.3 Downregulation of UBE2S inhibited HCC growth and metastasis in vivo

To evaluate the effect of UBE2S on HCC progression in vivo, HCCLM3 cells with stable knockdown of UBE2S expression (named UBE2S-shRNA HCCLM3) and Bel-7402 cells with stable overexpression of UBE2S (named UBE2S- overexpression Bel-7402) were established via lentivirus transfection. Then, we used UBE2S-shRNA HCCLM3 cells and control-shRNA HCCLM3 cells to establish subcutaneous xenograft tumor models and orthotopic xenograft tumor models. The results showed that downregulation of UBE2S significantly inhibited tumor size (Figure 4A). The tumor sizes and weights of the UBE2S-shRNA group were markedly decreased (Figure 4B,C). However, there was no obvious distinction in the weights of mice from the UBE2S-shRNA and control-shRNA groups in subcutaneous xenograft tumor models (Figure 4D). Moreover, upregulation of UBE2S significantly improved tumor size (Figure 4E). The tumor sizes and weights of the UBE2S overexpression group were markedly increased (Figure 4F,G). There was also no obvious distinction in the weights of mice from the control and UBE2S overexpression groups (Figure 4H).

In addition, in liver orthotopic xenograft tumor models, the results were consistent with those of subcutaneous xenograft tumor models, which indicated that knockdown of UBE2S could inhibit tumor growth in vivo (Figure 4I,J). In addition, H&E staining of lungs confirmed that downregulation of UBE2S in HCC decreased the number of pulmonary metastatic nodules in liver orthotopic xenograft tumor models (Figure 4K). We also determined the ratio of lung metastatic nodule number per tumor weight for each mouse, and there were still significant differences between the two groups, which indicated that UBE2S knockdown reduced metastasis independently of tumor growth (Figure 4L). In summary, the results indicated that UBE2S plays an important role in HCC growth and metastasis in vivo.

3.4 Downregulation of UBE2S arrested HCC cells in the G2/M phase by decreasing the degradation of p21

To explore the molecular mechanism of downregulation of UBE2S-mediated inhibition of HCC growth, gene set enrichment analysis (GSEA) was conducted based on TCGA data. The results showed that the G2/M checkpoint signaling pathway was upregulated in HCC patients with high expression of UBE2S mRNA (Figure 5A). Then, the cell cycle of HCCLM3 and HUH7 cell lines after knockdown of UBE2S was detected by flow cytometry. The percentage of cells in G2/M was increased compared with that of control cells (Figure 5B). In addition, we further measured the G2/M-arrest-related proteins in HCCLM3 and HUH7 cell lines after UBE2S knockdown. The results indicated that cyclin A, cyclin B1, and p-CDC2 were upregulated and cyclin E2 was downregulated in the HCCLM3 and HUH7 cell lines after UBE2S knockdown (Figure 5C). These results indicated that downregulation of UBE2S arrested HCC cells in the G2/M phase. More importantly, apoptosis assays were conducted, and there was no difference between the siUBE2S group and the sicontrol group, which indicated that downregulation of UBE2S suppressed HCC cell growth by arresting the cell cycle rather than inducing cell apoptosis (Figure 5D).

Many studies have shown that the p21 protein is a regulator of the cell cycle at the G2/M checkpoint.^29-31 In addition, knockdown of UBE2S can arrest OSCC cells in the G2/M phase by decreasing the degradation of p21.¹⁶ Inspired by the abovementioned findings, we detected the expression levels of p21 protein. The results indicated that knockdown of UBE2S increased p21 protein levels (Figure 5C). In addition, p21 was downregulated by UBE2S overexpression in Hep3B cells (Figure 5E). To examine the role of p21 in UBE2S-mediated cell proliferation, p21 was upregulated after overexpression of UBE2S in Hep3B cells. The results incated that overexpression of p21 reversed the upregulation effects of UBE2S on HCC cell proliferation (Figure 5F). Moreover, downregulation of p21 reversed the effect of arresting HCC cells in the G2/M phase after inhibition of UBE2S in HCCLM3 cells (Figure 5G). In subcutaneous xenograft tumor models, the p21 protein levels in the UBE2S knockdown group (UBE2S-shRNA) were significantly increased compared with those in the control-shRNA group, while there were fewer Ki67-positive cells in the UBE2S knockdown group than in the control-shRNA group (Figure 6A,B). However, TUNEL staining revealed no differences between the two groups (Figure 6A,B). All these results demonstrated that downregulation of UBE2S arrested HCC cells in the G2/M phase to inhibit cell proliferation by decreasing the degradation of p21.

3.5 UBE2S promoted HCC growth and metastasis by inactivating not only VHL/HIF-1α signaling but also the VHL/JAK2/STAT3 signaling pathway

Previous studies have shown that UBE2S can regulate tumor development through the VHL/HIF-1α signaling pathway.^{21, 22} In addition, UBE2S can target VHL and stabilize HIF-α in the mouse liver.¹⁵ However, the key mechanisms of UBE2S in promoting HCC growth and metastasis still needs in-depth study. Based on these findings, we first verify whether UBE2S promotes HCC growth and metastasis by inactivating VHL/HIF-1α signaling and their moderating relationship. We detected the protein levels of VHL and HIF-1α in HCC cells with downregulated or overexpressed UBE2S by western blotting. The results indicated that the expression of VHL protein in HCCLM3 cells was increased significantly after downregulation of UBE2S, and the expression level of HIF-1α protein was decreased significantly (Figure 7A). In contrast, after the overexpression of UBE2S in Bel-7402 cells, the expression of VHL in Bel-7402 cells was significantly inhibited, and the level of HIF-1α protein was increased (Figure 7A). We then further measured the expression of VHL and HIF-1α by immunofluorescence. The results were similar to the western blotting results (Figure 7B). Moreover, the protein expression levels of VHL and HIF-1α in tumor tissues collected from xenograft tumor models were detected by IHC, and the results were similar to those from western blotting (Figure 7C).

In addition, we tried to explore the direct interaction between UBE2S and VHL in HCC cells by co-IP. First, HCCLM3 cells were transfected with the UBE2S plasmid and then incubated for 12 h in the presence of 10 mM MG132, a proteasome inhibitor, and the cell lysates were immunoprecipitated with anti-UBE2S antibodies. Then, western blotting was used to detect the expression level of VHL. The results showed that VHL was contained in the cell lysates immunoprecipitated with anti-UBE2S antibodies (Figure 7D). Similarly, UBE2S was also detected in cell lysates immunoprecipitated with anti-VHL antibodies in Bel-7402 cells (Figure 7D). We also examined the levels of VHL in Bel-7402 cells after transfection with the UBE2S plasmid and incubation for 12 h in the presence or absence of 10 mM MG132. This result showed that VHL was decreased in a manner dependent on the expression level of UBE2S in the absence of MG132 and this decrease was eliminated in the presence of MG132 (Figure 7E). In addition, Jung et al. and Lim et al. also showed that UBE2S targeted VHL for ubiquitin-mediated proteolysis to degrade VHL (15, 21). Taken together, UBE2S modulates VHL/HIF-1α signal transduction by binding to VHL and degrading VHL.

It has been reported that VHL can regulate the transduction of JAK/STAT signaling.^{32, 33} Therefore, we speculated that in HCC cells, UBE2S may regulate the transduction of JAK/STAT signaling by directly interacting with VHL. To verify the role of UBE2S in the regulation of the JAK/STAT signaling pathway in HCC, we detected the protein levels of JAK2, p-JAK2, STAT3, and p-STAT3. The results indicated that the protein levels of p-JAK2 and p-STAT3 in HCC cells were decreased after downregulating the expression of UBE2S (Figure 7F). In contrast, the upregulated expression of UBE2S in Bel-7402 cells increased the protein levels of p-JAK2 and p-STAT3 (Figure 7F). Furthermore, we detected the protein levels of p-JAK2 and p-STAT3 in tumor tissues collected from HCC xenograft tumor models by IHC. The results showed that the expression of p-JAK2 and p-STAT3 proteins in the UBE2S-shRNA group was significantly lower than that in the control-shRNA group (Figure 7G).

In conclusion, UBE2S may regulate HCC growth and metastasis by inactivating not only VHL/HIF-1α signaling but also the VHL/JAK2/STAT3 pathway.

3.6 Downregulation or upregulation of VHL reversed the effect of UBE2S on cell proliferation and migration

Because downregulation of UBE2S increases VHL protein levels, we questioned whether the effect of downregulation of UBE2S in HCC is attributable to the enhancement in VHL stability. We cotransfected a lentivirus to generate stable UBE2S knockdown and siVHL in HCCLM3 cells and examined cell proliferation and migration. The influence of VHL on cell proliferation at 96 h was detected by MTT assays (Figure 8A). In Figure 8B,C, the migration abilities were also determined. Indeed, siVHL reversed the inhibition of cell proliferation and migration mediated by UBE2S knockdown (Figure 8A–C). We found that knockdown of VHL reversed the inhibition of cell proliferation and migration caused by the downregulation of UBE2S in HCCLM3 cells (Figure 8A–C). We further measured the expression levels of Snail, HIF-1α, JAK2, p-JAK2, STAT3, and p-STAT3. The results indicated that knockdown of VHL in UBE2S-shRNA HCCLM3 cells increased the expression levels of Snail, HIF-1α, JAK2, p-JAK2, and p-STAT3 (Figure 8D).

In addition, we also checked whether overexpression of VHL could reverse the accelerated cell proliferation, migration, and invasion mediated by UBE2S overexpression. UBE2S-overexpressing Bel-7402 cells were cotransfected with pVHL, and cell proliferation and migration were examined. pVHL reversed the enhancement of cell proliferation and migration mediated by UBE2S overexpression in Bel-7402 cells (Figure 8E–G). We further measured the protein levels of Snail, HIF-1α, JAK2, p-JAK2, STAT3, and p-STAT3. The results showed that VHL overexpression in UBE2S cells reduced the protein levels of Snail, HIF-1α, JAK2, p-JAK2, and p-STAT3 (Figure 8H).

The data suggested that VHL plays a key role in UBE2S-mediated HCC development, in which UBE2S could bind VHL and promote VHL degradation to activate VHL/HIF-1α signaling and the VHL/JAK2/STAT3 pathway.

3.7 Knockdown of UBE2S increases the sensitivity of HCC cells to sorafenib

To measure the biological significance of UBE2S in the sensitivity of HCC cells to sorafenib, the viability of HCC cells treated with sorafenib was detected. We found that after downregulation of UBE2S in HCCLM3 cells, the cell death induced by sorafenib was markedly increased (Figure 9A). The half maximal inhibitory concentrations (IC₅₀) of sorafenib in HCCLM3-control-shRNA cells and HCCLM3-UBE2S-shRNA cells were 10.57 μM and 8.68 μM, respectively. In addition, upregulation of UBE2S enhanced sorafenib resistance in Bel-7402 cells (Figure 9B). The IC50 of sorafenib was 15.87 and 24.57 μM before and after overexpression of UBE2S in Bel-7402 cells, respectively. To further explore the role of UBE2S in sorafenib sensitivity in vivo, we established subcutaneous xenograft models with UBE2S-shRNA HCCLM3 and their negative control cells, and then the nude mice were treated with sorafenib daily by intraperitoneal injection. The results indicated that tumors in the UBE2S-shRNA group treated with sorafenib were smaller than those in their negative control group treated with sorafenib (Figure 9C–E). In addition, in liver orthotopic xenograft tumor models, the results also indicated that tumors in the UBE2S-shRNA group treated with sorafenib were smaller than those in their negative control group treated with sorafenib, suggesting that knockdown of UBE2S made HCC cells more sensitive to sorafenib in vivo (Figure 9F,G).