Microarray dataset GSE14520 (17) was downloaded from GEO. A total of 183 HCC and 179 corresponding para-carcinoma tissues that were assayed with Affymetrix U133A GeneChips from cohort 2 Chinese patients were used in this study. Gene expression profiling data was re-summarized using the RMA method (18) and Entrez gene-centric CDF files (19) (instead of original Affymetrix CDF files), which filtered non-specific probes on the GeneChips and merged multiple probe sets representing the same Entrez gene into one probe set. Significance analysis of microarray (SAM) (20) was performed to identify differently expressed genes between HCC and corresponding para-carcinoma tissues. Delta was set to 2.25, and the threshold of FDR was set to 0.001. Genes overexpressed in HCC that were expressed in more than 50 HCC tissues but less than 10 corresponding para-carcinoma tissues were studied. The genes were further analyzed with GenCLiP software (21) (http://ci.smu.edu.cn) to annotate gene functions and construct molecular interaction networks.

The clinical samples were used for research purposes only. Approval was obtained from the ethics committee of Chinese PLA General Hospital (LREC 2012/40). All samples were collected under the condition of a prior written informed consent from the patients.

Sixteen pairs of fresh HCC and adjacent non-tumor tissues used for real-time PCR and western blot analyses were collected during surgery from the Chinese PLA General Hospital (Beijing, China) from November, 2012 to February, 2013. Tissues were snap-frozen in liquid nitrogen until use. Paraffin-embedded, archived HCC and adjacent non-tumor tissues used for immunohistochemistry (IHC) were obtained from 132 HCC patients who underwent partial liver resection at the same hospital between January, 2006 and December, 2007. The median age of the patients was 52 years (range 24–80 years).

One normal liver cell line (LO2) and three HCC cell lines (HepG2, SMMC7721, Huh7) were purchased from the American Type Culture Collection (ATCC, Manassas, VA) or Chinese Academy of Science Cell Bank and were maintained in the Institute of Biotechnology, Academy of Military Medical Sciences. All cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen, CA) supplemented with 10% fetal bovine serum (FBS, Gibco, Carlsbad, CA) and 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin at 37°C with 5% CO₂.

Total RNA was extracted from cell lines and tissue samples using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. A total of 2 μg RNA was reverse transcribed using the TransScript and cDNA Synthesis Kit from TransGen Biotech (Beijing, China). Real-time PCR was performed to examine CDC20 mRNA level in cell lines and tissue samples by using a Bio-Rad iQ5 Multicolor Real-Time PCR Detection System (Bio-Rad, Hercules, CA). β-actin was used as an internal control for normalization. PCR primers were designed using the Primer Premier 5 software and the sequences were: CDC20 forward, 5′-TCGCATCTGGAATGTGTGCT-3′; and reverse, 5′-CCCGGGATGTGTGACCTTTG-3′; β-actin forward, 5′-TGACGTGGACATCCGCAAAG-3′; and reverse, 5′-CTGG AAGGTGGACAGCGAGG-3′. Expression data were normalized to the geometric mean of the housekeeping gene β-actin and calculated with the ΔΔCt (22) and results were expressed with 2^−ΔΔCt.

Western blot analysis was performed under the standard protocol. SDS-polyacrylamide gel electrophoresis (10%) was used to separate the protein which was then electrotransferred from the gel to a polyvinylidene fluoride (PVDF) membrane. After blocking with 5% dried skim milk for 1 h, the membrane was incubated with rabbit anti-human CDC20 polyclonal antibody (1:1,000, Bioworld Technology, St. Louis Park, MN) for 1 h at room temperature. The mouse anti-human α-tubulin monoclonal antibody (1:5,000; Santa Cruz Biotechnology, Santa Cruz, CA) was used as an internal control. After washing with Tris-buffered saline with Tween-20 (TBST) three times, the membrane was incubated with secondary horseradish peroxidase-conjugated antibody against rabbit or mouse (dilution 1:5,000). Chemiluminescent detection was performed with the Immobilon Western Chemiluminescent HRP Substrate kit (Millipore Corporation, Billerica, MA).

Immunohistochemistry for CDC20 expression in HCC and adjacent non-tumor samples was performed using standard methods. Briefly, tissue sections were incubated with rabbit anti-CDC20 diluted 1:200 (Bioworld Technology) overnight at 4°C. Bovine serum albumin (1%; BSA) without primary antibody was used as the negative control. The secondary poly-horseradish peroxidase (HRP) anti-rabbit IgG antibody (ZSGB-Bio, Beijing, China) was incubated in room temperature for 20 min.

All of the immunostained sections were reviewed and scored independently by two pathologists in a blinded manner without knowledge of the clinicopathological information, based on the H-score method, which considers the staining intensity together with the percentage of cells staining positively (23,24). For H-score method, 10 fields were chosen randomly at ×400 magnification. The staining intensity in the cells was scored as 0, 1, 2 and 3 corresponding to the negative, weak, intermediate and strong staining, respectively. In each field the total number of cells and cells stained at each intensity were counted. The H-score was calculated following the formula: (% of cells stained at intensity category 1×1) + (% of cells stained at intensity category 2×2) + (% of cells stained at intensity category 3×3). H-scores varied from 0 to 300 where 300 represented 100% of cells strongly stained (3+) (24). High CDC20 expression was defined as staining H-scores of cells ≥200.

Double-stranded, small interfering RNA (siRNA) was synthesized and purified by GenePharma (Shanghai, China). The subsequences corresponding to the coding region of human CDC20 were: sense, 5′-GGAGCUCAUCUCAGGCCA UUU-3′; antisense, 5′-AUGGCCUGAGAUGAGCUCCUU-3′. A scrambled non-targeting siRNA was used for the negative control. These siRNAs were dissolved in diethyl pyrocarbonate (DEPC) water to reach a concentration of 20 μM. Liver cancer HepG2 cells were treated with CDC20 siRNA or negative control siRNA in 20 nM by using the INTERFERin in vitro siRNA transfection reagent (Polyplus Transfection, New York, NY).

Quantitative real-time PCR and western blot analyses were used for detecting the interference effect of siCDC20. Cells that were untreated or treated with negative control siRNA oligonucleotides were the control groups.

Two 25-cm² plastic flasks were each inoculated with 2×10⁵ liver cancer cells (HepG2), which were then transfected with 20 nM negative control siRNA and CDC20 siRNA, respectively, for 48-h incubation. Then, cells were digested and seeded into 6-well plates containing 2 ml medium per well. Afterwards, the cells from a well were digested and counted by the automated cell counter (Invitrogen) every 24 h until day five.

siRNA for CDC20 and negative control siRNA oligonucleotides were transfected into HepG2 cells with the INTERFERin in vitro siRNA transfection reagent for 48 h. After treatment with 2 μg/ml thymidine (Sigma-Aldrich, St. Louis, MO) for 24 h, cells were harvested at 0, 3, 6, 9, 12-h after removing thymidine from the medium and fixed with 70% alcohol. Prior to analyses, cells were washed with PBS, treated with 1 mg/ml RNase A at 37°C for 30 min and then stained with propidium iodide (PI). Analyses were performed by BD FACSCalibur (Becton-Dickinson, Franklin Lakes, NJ) and results were analyzed by WinMDI Version 2.9 software.

All statistical analyses were made using the IBM SPSS 20.0 statistical software package. One-way analysis of variance (ANOVA) was used to compare the expression of CDC20 mRNA normalized to β-actin in cell lines. The same method was also performed in comparing the histological differentiation in tumor tissues normalized to normal liver samples. Data comparisons in two groups were performed by Student’s t-test. The χ² test was used to analyze the relationship between CDC20 expression and clinicopathological features. Bivariate correlations between variables were calculated by Spearman’s correlation coefficients. The level of statistic significance was set at P<0.05 for all tests.

SAM analysis identified 4,358 genes overexpressed in HCC, in which 137 genes were expressed in more than 50 HCC tissues, but in less than 10 para-carcinoma tissues. GenCLiP analysis showed that among the 137 genes, 72 genes were related to the cell cycle (P=1.238e-15, by χ² test), 19 genes were related to spindle assembly (P=2.172e-55, by χ² test), and 26 genes were related to chromosome segregation (P=5.472e-55, by χ² test). Among the molecular networks constructed with the 137 genes, CDC20 was the major node that has not been previously reported to be related with HCC. Nine genes (NEK2, E2F1, BUB1, BUB1B, AURKB, CCNB1, CCNB2, UBE2C and CDKN2A) are known to interact with CDC20, in which 7 genes were related with spindle assembly checkpoint (SAC) (Fig. 1). The target of the SAC is CDC20 (25). Thus, we inferred that in HCC, overexpression of CDC20 leads to absence of SAC, and cells rapidly become aneuploid.

Quantitative real-time PCR was used to examine transcript level of CDC20 in three HCC cell lines (HepG2, SMMC-7721 and Huh7), and one normal liver cell line (LO2). CDC20 mRNA level was higher in all the three HCC cell lines than that in the normal cell line (Fig. 2).

In order to determine whether the upregulation of CDC20 in HCC cell lines is similar in HCC patients, we performed quantitative real-time PCR on 16 pairs of matched HCC and adjacent normal liver tissues. As shown in Fig. 3A, CDC20 was found to be differentially overexpressed in 14 of all examined human primary HCC samples compared with non-cancerous samples from the same patients. Additionally, the tumor/non-tumor (T/NT) ratio of CDC20 mRNA expression was at least >2-fold in all these 14 upregulated samples, and the highest ratio was even up to about 40-fold. The protein level of CDC20 was confirmed in all the 16 paired tissue samples by western blot analysis. Image J 1.47v software was used to quantize the grey level of each band and calculate the CDC20 T/NT ratio of each patient normalizing with α-tubulin. Results revealed that all examined HCC samples showed a higher expression level of CDC20 protein except sample 1 and sample 4, which was consistent with the mRNA level (Fig. 3B). These findings indicated that CDC20 was commonly upregulated in either HCC tissues or cell lines.

Immunohistochemistry (IHC) was performed to analyze the protein expression and localization of CDC20 in 132 paraffin-embedded archived HCC tissue samples, including 14 cases of well differentiated, 78 cases of moderate differentiated and 40 cases of poor differentiated tumors. CDC20 protein was upregulated (H-score ≥200) in 90 (68.18%) of 132 HCC tissues. As shown in Fig. 4A, high levels of CDC20 were present in primary HCC lesions and the positive staining mainly localized in cytoplasm and nuclei. In contrast, CDC20 was negatively or weakly stained in corresponding non-tumor tissues. By quantitatively comparing the scores of CDC20 staining between 132 archived HCC and adjacent non-tumor tissues (mainly including hepatocirrhosis and hepatitis), the scores of CDC20 staining was significantly increased in HCC samples compared to peritumoral samples (Fig. 4B). In addition, the scores of CDC20 staining was significantly increased along with the progression of tumor histological differentiation from well to poorly differentiated tissues (Fig. 5).

The relationship between CDC20 protein expression and the clinicopatholigical features of 132 HCC patients was further analyzed. As summarized in Table I, CDC20 expression was significantly associated with gender (P=0.013), tumor differentiation (P=0.000), TNM stage (P=0.012), P53 and Ki-67 expression (P=0.023 and P=0.007, respectively). However, no statistically significant association was found between high CDC20 expression and other clinicopathological characteristics including age, tumor size, HBV infection, serum AFP level, hepatic cirrhosis, vascular invasion and intra/extra hepatic metastasis. Spearman correlation analysis revealed that high expression of CDC20 was closely related with gender (R=0.215, P=0.013), poorer tumor differentiation (R=0.421, P<0.001), advanced TNM staging (R=0.218, P=0.012), higher expression level of p53 (R=0.212, P=0.014) and Ki-67 (R=0.235, P=0.007) (Table II). Taken together, these results indicated that CDC20 was highly expressed in HCC tissues and its expression closely correlated with the differentiation and progression of HCC.

The quantitative real-time PCR revealed 90% transcriptional level suppression of CDC20 at 48 h after transfection with siCDC20, comparing with untreated cells or cells transfected with negative control siRNA (Fig. 6A) (P<0.0001). The western blot analysis also showed an obvious suppression of CDC20 protein level at 48 h after transfection with siCDC20 (Fig. 6B). The altered expression of cyclin-dependent kinase inhibitor 1A (p21) protein was also examined by western blot analysis, and the results showed that P21 protein level was remarkably upregulated owing to the suppression of CDC20 (Fig. 6B).

Cell proliferation assay revealed that the cell growth of CDC20 siRNA transfected cells was remarkably inhibited compared to the cells transfected with negative control siRNA (Fig. 6C). The inhibition of cell proliferation by siCDC20 was expected to be accompanied by the arrest at metaphase of the cell cycle. By flow cytometry test, an increased number of cells in the G2/M phase was observed in siCDC20 transfected cells compared to cells transfected with negative control siRNA (Fig. 6D). These data indicated that suppression of CDC20 inhibited cell growth by inducing G2/M phase arrest.