The present study used paraffin-embedded HCC samples from 110 patients who underwent curative hepatectomy. Patients should fulfill the following criteria: No evidence of extrahepatic metastasis, main portal vein infiltration/thrombosis, and eligible for curative hepatectomy. Exclusion criteria for this study included: Rehepatectomy, preoperative combination with other tumors or prior history of other tumors. The samples had been clinically and histologically diagnosed at the Third Affiliated Hospital of Sun Yat-sen University (Guangzhou, China) between January 2011 and December 2012. A total of 10 HCC and 3 normal tissue samples were frozen and stored in liquid nitrogen and these tissues to investigate the mRNA and protein levels of SKA3. A total of three normal liver specimens obtained from the edge of the liver were used to isolate normal primary human hepatocytes. The HCC samples were obtained from patients: 99 (90%) men and 11 (10%) women. The median age of the patients was 54 years (range, 31–76 years). The patients were followed up from 1–77 months, with a median of 23 months. Clinicopathological information is summarized in Table I. All clinical specimens were obtained with informed consent and the study was approved by the Clinical Research Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University. Informed written consent was obtained from all patients. Tumor stages were determined according to the HCC TNM staging system of the 8th American Joint Committee on Cancer (AJCC) in January 2018 (15).

Hep3B, Huh7, SK-Hep1, HepG2, MHCC97-H, MHCC97-L, SNU-387 and SNU-475 liver cancer cells were obtained from the Cell Bank of Shanghai Institute of Biology, Chinese Academy of Science (Shanghai, China). Normal primary human hepatocytes were isolated from liver specimens using a standardized two-step collagenase perfusion technique (16,17). High yield and high activity normal liver cells were obtained. All cells were cultured in DMEM, supplemented with 1% penicillin-streptomycin (both Invitrogen; Thermo Fisher Scientific, Inc.) and 10% FBS (HyClone; GE Healthcare Life Sciences) at 37°C with 5% CO₂. Cells were sub-cultured every 1–2 days to maintain logarithmic growth. Huh7 and HepG2 were transfected with short hairpin (sh)RNA against SKA3 to knock down SKA3, and lentivirus overexpressing vector to increase the expression of SKA3. Vector and non-specific scramble shRNA were used as negative controls. The plasmids for expression of SKA3, SKA3-specific shRNA, and their relevant lentiviruses were obtained from Shanghai GeneChem, Co., Ltd. The cells were incubated in a humidified 5% CO₂ atmosphere for 24 h at 37°C. Finally, the diluted virus was replaced with fresh medium and incubated for another 48 h under the same conditions. Subsequently, the transfection efficiency was examined by western blotting. Liver cancer cells were transduced with individual types of lentivirus at a multiplicity of infection (MOI) of 10 in the presence of 5 µg/ml puromycin [Hanheng Biotechnology (Shanghai) Co., Ltd.]. The two SKA3 sequence were: shSKA3#1, GATCTGTCTGATCCTCCTGTT; and shSKA3#2, CCACAGGCAGTGAACAACTAT. According to the manufacturer's instructions, liver cancer cells were transfected using Block Lentiviral shRNA Expression system (Invitrogen; Thermo Fisher Scientific, Inc.). Subsequent experiments were performed 48 h after transfection.

Total RNA from cell lines, 11 fresh tissue samples and tissue from athymic nude mice were extracted using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instruction. The cDNA synthesis was performed according to the manufacturer's instructions (Invitrogen; Thermo Fisher Scientific, Inc.). PCR reaction conditions for all assays were 95°C for 2 min, followed by 40 cycles of amplification (95°C for 10 sec, 60°C for 20 sec and 72°C for 20 sec). RT-qPCR was performed using the following primers: SKA3, forward, 5′-TACACGAGCAAGAAGCCATTAAC-3′ and reverse, 5′-GGATACGATGTACCGCTCAAGT-3′; p21, forward, 5′-CGATGCCAACCTCCTCAACGA-3′ and reverse, 5′-TCGCAGACCTCCAGCATCCA-3′; and Cyclin D1, forward, 5′-AACTACCTGGACCGCTTCCT-3′ and reverse, 5′-CCACTTGAGCTTGTTCACCA-3′. GAPDH was used as the endogenous control with the following primers: Forward, 5′-ACAACTTTGGTATCGTGGAAGG-3′ and reverse, 5′-GCCATCACGCCACAGTTTC-3′. Expression levels were normalized to those of GAPDH and calculated as 2^{−[(Cq of genes)-(Cq of GAPDH)]}, where Cq represents the threshold cycle for each transcript (18). All experiments were repeated three times.

Western blotting was performed as described previously (18). The following antibodies (all 1:1,000) were used: SKA3 (cat. no. ab118560; Abcam), anti-p-Rb (cat. no. D20B12; Cell Signaling Technology, Inc.), anti-Rb (cat. no. D20; Cell Signaling Technology, Inc.) and α-tubulin (cat. no. ab7291; Abcam) as an internal loading control. The expression levels of SKA3 were detected using enhanced chemiluminescence detection system (Amersham; Cytiva) according to the manufacturer's instructions.

IHC and SKA3 expression scoring were performed as previously reported (19). Image-Pro Plus version 6.0 software (Media Cybernetics, Inc.) was used to judge the area and density of the dyed region, and the integrated optical density (IOD) value of the IHC section. The intensity of staining was graded as follows: 0 (no staining), 1 (weak staining, light yellow), 2 (moderate staining, yellow brown), and 3 (strong staining, brown). The proportion of tumor cells was scored as follows: 0 (no positive tumor cells), 1 (<10% positive tumor cells), 2 (10–50% positive tumor cells), 3 (50–75% positive tumor cells), and 4 (>75% positive tumor cells). The staining index (SI) was calculated by the product of the staining intensity score and the proportion of positive tumor cell score. The optimal cut-off value was determined: SI score ≥6 was considered to indicate high SKA3 expression, while a score <6 indicated low expression.

Cell proliferation was measured via MTT assay (Sigma-Aldrich; Merck KGaA). In brief, treated and untreated Huh7 and HepG2 cells were inoculated into 96-well plates. The cells were incubated with 20 µl 5 mg/ml MTT solution in a humid atmosphere containing 5% CO₂ at 37°C for 4 h. The OD was measured at 490 nm using a spectrophotometer following cell lysis and dissolution of methoxypyrine in 150 µl DMSO. The calculation formula of proliferation inhibition rate was as follows: Proliferation inhibition rate=OD_sample/OD_control. All experiments were repeated three times, and the mean values were plotted to construct cell proliferation inhibition curves.

Both non-transfected and transfected Huh7 and HepG2 cells were seeded in 6-well plates at a density of 1,000 cells/well. After 12 days, cells were fixed with 4% paraformaldehyde for 30 min and stained with 0.1% crystal violet for 30 min at 37°C. Next, the plates were washed gently with PBS, then air dried and the stained colonies were photographed under an inverted light microscope (×200 magnification; Leica Biosystems). The colonies were counted in 10 randomly chosen microscope fields. The experiments were repeated three times.

Propidium iodide (PI) staining was used to analyze DNA content. Treated and untreated Huh7 and HepG2 cells were washed with PBS and fixed in ice-cold ethanol at −20°C overnight. and stained with staining solution containing 20 µg/ml PI and 100 µg/ml RNase (both Sigma-Aldrich; Merck KGaA) for 20 min at 4°C. A FACScan flow cytometer (BD Biosciences) was used to analyze DNA content and FlowJo version 7.6 software (FlowJo LLC) was used to determine the percentage of cells in G₁/G₀, S and G₂/M phase. The experiments were repeated three times.

Cells were seeded in 24-well plates in triplicate and allowed to stand for 24 h. E2F 3′-untranslated region (UTR) was cloned into the downstream region of the luciferase gene in the pGME2F-Lu luciferase vector. For the reporter assay, the E2F luciferase reporter plasmid pGME2F-Lu Genomeditech Inc.), 1 ng pRL-TK Renilla plasmid (Promega Corporation), and a SKA3 mimic, SKA3 inhibitor or SKA3 control vector were transfected into cells using Lipofectamine 2000 Reagent (Invitrogen; Thermo Fisher Scientific, Inc.). The primer sequences of E2F 5′-3′ UTR were 5′-GGCCCGAUCGAUGUUUUCCTT-3′ (forward) and 5′-GGAAAACAUCGAUCGGGCCTT-3′ (reverse). Following 48 h transfection, cells were lyzed and luciferase activity was measured using Dual-Luciferase Reporter Assay kit (Promega Corporation) according to the manufacturer's instructions. Data were standardized to Renilla luciferase activity.

In order to investigate the expression levels of SKA3 in liver tissue, data from TCGA (tcgadata.nci.nih.gov/tcga/) was analyzed. Gene expression levels and clinical data of patients were integrated according to their barcode ID. In order to determine the biological pathways underlying the effect of SKA3 in liver cancer, GSEA, a method of using biological knowledge to analyze and interpret microarray and data (20), was conducted using GSEA version 2.0 of the Broad Institute (21). GSEA firstly generated an ordered gene list according to correlation with SKA3 expression, then a pre-defined gene set enrichment score (ES) was obtained to refute the hypothesis that its members are randomly distributed in the ordered list. The BENPORATH_PROLIFERATION gene set (C2. BENPORATH. V3.0) biological process database from the Molecular Signatures Database was used for enrichment analysis (22). P-value was calculated using an arrangement of 1,000 random samples. A false discovery rate (FDR) <25% and nominal P-value <0.05 were considered to indicate a statistically significant difference.

Complete medium agar (1%; Sigma-Aldrich; Merck KGaA) mixture was poured into the wells of a 6-well plate. Following solidification, cells (1×10³) were digested with trypsin and suspended in 2 ml medium supplemented with 0.3% agar, and then spread on the bottom layer. Following incubating at 37°C, 5% CO₂ for 10 days, colony size was measured with an ocular micrometer and colonies with a diameter >0.1 mm were counted. The experiment was performed independently three times.

For in vivo tumor proliferation assays, a subcutaneous HCC model was established. Twenty six-week-old male BALB/c (nu/nu) mice (average weight, 15 g) were supplied by Beijing Weitong Lihua Laboratory Animal Technology Co., Ltd. The mice were raised in the animal facility at Sun Yat-sen University (Guangzhou, China) under specific pathogen-free laboratory with a 12 h light/dark cycle at 22±2°C and 60±5% humidity, provided with free access to food and water. The mice were randomly divided into four groups (n=5/group) and injected with 2×10⁶ Huh7 cells transfected with empty vector, SKA3, scramble or shSKA3#2. Tumor volume was monitored by measuring the length (L) and width (W) at 3-day intervals for 4 weeks and calculated using the equation (LxW²)/2. At 4 weeks following liver cancer cell inoculation, all animals were euthanized with 100% CO₂ infused at 30% volume/minute displacement and then subjected to cervical dislocation in accordance with AVMA institutional guidelines (23). Respiratory and cardiac arrest and pupil dilation were considered to indicate animal death. The tumor tissues were dissected, weighed and subjected to pathological examination. The mean animal weight before tumor excision was 21.3±2.2 g. Animal experiments were approved by the Institutional Animal Ethics Committee at Sun Yat-sen University (no. 2017-047).

Statistical analysis was performed using SPSS 21.0 (IBM Corp.) statistical software. Each experiment was performed at least in triplicate. All the data are generally expressed as the mean ± SEM or counts and and percentages. A chi-square test and Fisher's exact test were used to assess the association between SKA3 expression and clinicopathological factors. Pearson's correlation analysis was used to evaluate the relationship between SKA3 and clinical pathologic factors. The differences between groups were analyzed by two-tailed Student's t-test or one-way ANOVA followed by Tukey's honestly significant difference post hoc test. Survival curves were constructed via the Kaplan-Meier method and compared with log-rank test. Univariate and multivariate Cox regression analysis was used to evaluate the significance of variables for survival. P<0.05 was considered to indicate a statistically significant difference.

SKA3 expression levels were extracted from TCGA liver cancer cohorts and were revealed to be significantly increased in HCC tissue compared with normal liver tissue (Fig. 1A). Furthermore, RT-qPCR was performed to evaluate mRNA expression levels of SKA3 in 10 solid tumor and three normal liver tissue samples from patients. SKA3 mRNA expression levels were significantly elevated in HCC tumors compared with normal tissue (Fig. 1B). SKA3 mRNA levels in all cultured liver cancer cell lines were also significantly higher than those in the normal hepatocyte lines (Fig. 1C). In order to investigate whether SKA3 was also upregulated at the protein level, western blotting was performed. The protein levels of SKA3 in clinical HCC tissue and cultured liver cancer cell lines were upregulated compared with their normal counterparts (Fig. 1D and E). These results indicated that SKA3 is overexpressed in HCC at both the mRNA and protein level.

In order to determine the clinical significance of SKA3 in liver cancer, the expression levels of SKA3 were studied in 110 paraffin-embedded liver cancer tissue samples using IHC analysis. Samples including 66 cases of stage I, 26 cases of stage II, 12 cases of stage III and 6 cases of stage IV tumor (Table I). As shown in Table II, SKA3 was positively expressed in 105 samples (95.5%), with high expression in 59 samples (53.6%) and low expression in 51 samples (46.4%). Correlations between SKA3 expression levels and clinicopathological characteristics of HCC were evaluated using Chi-square test and Pearson's correlation analysis (Tables III and IV). Kaplan-Meier survival analysis and log-rank test were used to evaluate the association between SKA3 expression levels and prognosis. Survival status of patients with high and low expression levels of SKA3 was analyzed (Fig. 2A). The overall survival time of patients with HCC with high levels of SKA3 was shorter than those of patients with low levels of SKA3 (Fig. 2B; log-rank P-value=0.002). Univariate and multivariate Cox regression analysis demonstrated that SKA3 expression and clinical stage were independent prognostic factors for patients with HCC (Table V).

In order to study the potential signaling pathways that contribute to SKA3-mediated proliferation of liver cancer cells, GSEA was used to identify associated signaling pathways. Gene expression level profiles of patients with low and high expression of SKA3 in the TCGA HCC cohort (371 patients, divided by the median expression of SKA3 mRNA) were compared. In the BENPORATH_PROLIFERATION gene set, the proliferation pathway was significantly enriched in the SKA3_High group (Fig. 3A). In order to evaluate the function of SKA3 in vitro, Huh7 and HepG2 cells were transfected with SKA3-expressing lentivirus to overexpress SKA3, or shRNA to silence SKA3. The efficiency of the overexpression and knockdown of SKA3 expression was examined by western blotting (Fig. 3B). In addition, in order to explore the association between the expression of SKA3 and the proliferation ability of HCC cells, MTT assay was performed. Knockdown of SKA3 significantly decreased the proliferation rate, and overexpression of SKA3 resulted in increased proliferation compared with the corresponding control liver cancer cell lines, both in Huh7 and HepG2 cells (Fig. 3C). The results demonstrated that SKA3 affected the proliferation of liver cancer cells. Colony formation assays were performed to evaluate the colony formation function of SKA3 in vitro, which demonstrated that silencing SKA3 colony formation, whereas overexpression of SKA3 increased colony formation (Fig. 3D).

Flow cytometry analysis was performed to determine whether the effect of SKA3 on the proliferation of liver cancer cells was affected by cell cycle arrest. The results showed that shRNA-transfected liver cancer cells were arrested at the G₁/G₀ phase and the percentage of cells in S-phase decreased, whereas in cells transfected with SKA3-expressing lentivirus, the percentage of cells in S-phase increased (Fig. 4A). GSEA analysis showed that high expression of SKA3 was associated with E2F1 and its target genes (Fig. 4B). Luciferase reporter activity assay revealed that the luciferase activity of the wild-type E2F-regulated 3′UTR was significantly upregulated by 6.5-fold, whereas knockdown of SKA3 significantly inhibited luciferase activity by 80% in Huh7 and HepG2 cells (Fig. 4C). Protein p21 inhibits the activity of the cyclin D1 complex, which results in decreased phosphorylation of Rb, in turn inhibiting the release of E2F1 from the Rb-E2F1 inhibitory complex, followed by enhanced expression of E2F1 target genes (24). In order to investigate whether mRNA and protein levels of associated cell cycle regulatory oncogenes are also upregulated in liver cancer cell lines with overexpressed and silenced SKA3, RT-qPCR and western blot analysis were performed. RT-qPCR demonstrated significant upregulation of p21 and downregulation of cyclin D1 in Huh7 and HepG2 cells overexpressing SKA3 (Fig. 4D). Western blotting revealed that the levels of cyclin D1 and p-Rb were elevated in liver cancer cells transfected with SKA3 and downregulated in liver cancer cells transfected with shSKA3, whereas levels of p21 were downregulated in liver cancer cells transfected with SKA3 and elevated in liver cancer cells transfected with shSKA3 (Fig. 4E). These results demonstrated that expression of SKA3 facilitated cell cycle progression in liver cancer cells.

In order to examine the effects of SKA3 on the growth of liver cancer cells, the anchorage-independent growth assay was performed. Quantitation of the number of spheres formed over serial passages showed that overexpression of SKA3 significantly increased sphere formation, whereas knockdown of SKA3 by transfection with shSKA3 significantly inhibited the sphere formation ability of Huh7 and HepG2 cell lines (Fig. 5A). In order to examine the effects of SKA3 on liver cancer cell growth in vivo, a subcutaneous tumor xenograft model was established. In athymic nude mice injected subcutaneously with Huh7 cells transfected with SKA3, the tumor volume and weight increased significantly, whereas in those injected with cells transfected with shSKA3, they decreased significantly (Fig. 5B-D). Differences in SKA3 expression levels in tumor tissue of each group were investigated by IHC, which demonstrated that SKA3 expression levels significantly increased in the group injected subcutaneously with Huh7 cells transfected with SKA3 (Fig. 5E). Furthermore, to explore whether associated cell cycle regulatory oncogenes were also upregulated in liver cancer tissue from these mice, mRNA and protein were extracted and RT-qPCR and western blotting were performed. RT-qPCR revealed significant downregulation of p21 and upregulation of cyclin D1 in the SKA3-overexpression HCC tissue group (Fig. S1). Western blotting revealed that the levels of cyclin D1 and p-Rb were elevated in the SKA3-overexpression and downregulated in shSKA3 HCC tissue group, while the levels of p21 were downregulated in liver cancer cells transfected with SKA3 and elevated in liver cancer cells transfected with shSKA3 (Fig. 5F). These results demonstrated that the expression of SKA3 increased liver tumor cell growth both in vivo and in vitro.