We randomly collected patient samples from the Cancer Hospital of Shantou University Medical College. 128 formalin-fixed paraffin-embedded (FFPE) specimens were obtained from patients with gastric cancer undergoing curative surgery between 2000 and 2005 (median age, 61 years; range, 35–84 years). The median follow-up period was 37 months (range, 1–80 months) from the date of surgery. Additionally, cancer tissues (n=20) and their matching noncancerous tissues (n=20), for immunoblot assays, were immediately snap frozen in liquid nitrogen and kept at −80°C until further use. Of note, they were harvested from another independent cohort of patients with gastric cancer who underwent surgery at the same institution between May 2010 and July 2012.

Tumor grade and stage were classified according to the International Union against Cancer (UICC)/American Joint Committee on Cancer (AJCC) pathologic tumor-node-metastasis (TNM) classification, 7th edition (2010). Table I demonstrates the clinicopathologic characteristics for these patients. No patient was found to have undergone preoperative radiotherapy or chemotherapy. The current study conformed to the ethical guidelines of the Declaration of Helsinki and was approved by the Institution Review Board (no. 04-470) of the Cancer Hospital of Shantou University Medical College. Written informed consent was obtained from all patients before sample collection. All samples were coded and data was stored anonymously.

CDK10 protein levels were examined by immunoblot analysis as described previously (6–8,16,17). 60 µg proteins in RIPA lysis buffer were separated by SDS-PAGE and transferred to a PVDF membrane. After one hour incubation in blocking buffer (Tris-buffered saline with 0.1% Tween and 5% nonfat dry milk), the membrane was then incubated with a rabbit polyclonal antibody against CDK10 (catalog no. ab72710; dilution, 1:500; Abcam, Cambridge, MA, USA), followed by a horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG; catalog no. sc-2004; dilution, 1:2,000; Santa Cruz Biotechnology, Inc., Dallas, TX, USA). Signals were visualized using an enhanced chemiluminescence kit (GE Healthcare Life Sciences, Little Chalfont, UK) as described by the manufacturer. Anti-GAPDH monoclonal antibody (catalog no. ab9485; dilution, 1:2,500; Abcam) was used to assure equal loading of protein. Quantification of the intensity of CDK10 in the immunoblot was conducted using the Bio-Rad Quantity One quantitation software (20), with the ratio between the tumors and the paired noncancerous tissues identified as being less than two folds, suggesting decreased CDK10 expression.

We performed immunohistochemical staining to detect CDK10 expression using a standard EnVision complex method as described previous (7,14,17). 4-µm sections were cut from FFPE specimens and then processed with deparaffinization and rehydration. The endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide for 30 min at room temperature. After antigen retrieval, tissue sections were incubated with a rabbit polyclonal anti-CDK10 antibody (catalog no. ab72710; dilution, 1:50; Abcam), after which immunohistochemical staining was conducted by an EnVision antibody complex (anti-Mouse/Rabbit) method using an EnVision^™ Detection kit (ZSGB-BIO, Beijing, China) and 3,3′-diaminobenzidine as the chromogen substrate. As a negative control, the primary antibody was replaced by a normal rabbit IgG.

The staining evaluation was carried out as follows (7): Ten 400× microscopic fields per slide were random selected and evaluated by two independent pathologists. CDK10 immunostaining was determined using a semi-quantitative approach combining the percentage of positive cells and the staining intensity. The mean percentage of stained cells was scored as follows: 0, 0%; 1, 1–25%; 2, 26–50%; 3, 51–75%; and 4, 76–100%. The staining intensity was classified as follows: 0, Absent; 1, weak staining; 2, moderate staining; and 3, strong staining. We summed up the intensity and extent scores as the final staining score. For the purpose of statistical evaluation, we grouped the tumor samples with a final staining score of <3 into negative CDK10 expression and those with scores ≥3 into positive CDK10 expression.

Two established gastric cancer cell lines (HGC-27 and SGC-7901) were cultured as described previously (6,16). A recombinant CDK10-expressing plasmid (pcDNA3.1-CDK10) and a pre-designed validated siRNA targeting CDK10 (5′-GUCCCAGUAAAGCCAAUGATT-3′ and 5′-UCAUUGGCUUUACUGGGACTT3′) were prepared as described previously (6). Cell transfection was conducted using Lipofectamine 2000 reagent (GE Healthcare Life Sciences, Little Chalfont, UK) according to the manufacturer's instructions. Cells were harvested 48 h post-transfection and used for subsequent experiments described below. The expression levels of CDK10 after transfection was confirmed by immunoblot analysis.

The cell proliferative and cologenic capacities were measured using a 3-(4,5-dimethylthiazol-2-yl)-2 5-diphenyl-2H-tetrazolium bromide (MTT) colorimetric assay and a colony formation (soft agar culture) assay, respectively, according to standard methods described before (6,16,18). Each experiment was conducted three times in replicates of six wells.

The cell motility and invasive capacities were measured using a wound healing assay and a Matrigel invasion chamber assay, respectively, according to standard methods described before (6,16,18). Each experiment was conducted in triplicate wells and repeated three times.

Statistical analyses were performed using the SPSS statistical software package (version 17.0; SPSS, Inc., Chicago, IL, USA). A Student's t test was used to evaluate the statistical significance of differences in numerical data. The Pearson χ² test or Fisher's exact test was used to analyze the correlations between CDK10 expression and clinicopathologic parameters, and the Spearman's rank method was used to obtain the correlation coefficient between variables. Overall survival (OS) rates were generated using the Kaplan-Meier method with log rank test. The prognostic significance of clinicopathological variables was determined by univariate and multivariate regression analysis with the Cox hazards model. P<0.05 (two-tailed) was considered to indicate a statistically significant difference.

Giving that evidenced function of CDK10 as a potential tumor suppressor frequent silenced during tumorigenesis (5–7), we first determined whether CDK10 expression was downregulated in tumors vs. noncancerous tissues. Fig. 1 shows the representative results in a cohort of gastric cancer specimen (n=20) by immunoblot analysis. We found that 85% (17/20) of cancer tissues had lower CDK10 protein levels compared to their matching adjacent noncancerous tissues (P<0.001).

To verify the above observation, we next determined the expression profile and localization of CDK10 protein in 168 gastric cancer specimens by utilizing immunohistochemistry. Positive CDK10 immunostaining was observed primarily in the cytoplasm of neoplastic cells in 50.8% (65/128) of primary gastric tumors tested (Fig. 2). To better understand the significance of CDK10 expression in gastric cancer, we further assessed the association between CDK10 expression and the clinicopathological parameters. We found that negative CDK10 expression was associated with advanced tumor stage (stage III/IV vs. stage I/II; P<0.001), frequent lymph node metastasis (N2/N3 vs. N0/N1; P<0.001), distant metastasis (P=0.013) and tumor differentiation (poor vs. well/moderate; P=0.004) (Table I). No significant relationship was observed between CDK10 expression and the other clinicopathologic factors such as age, gender, tumor size and invasive depth. Spearman correlation analysis revealed CDK10 expression levels were correlated with clinical tumor staging (r=−0.375; P<0.001), lymph node metastasis (r=−0.349; P<0.001), distant metastasis (r=−0.239; P=0.002) and tumor differentiation (r=−0.288; P<0.001). Together, these data demonstrate that a lack of CDK10 expression may correlate with malignant properties mainly relevant to lymph node metastasis and tumor advancement. These observations provide solid evidence that dysregulation of CDK10 expression may contribute to gastric cancer progression.

Kaplan-Meier survival analyses were performed to investigate the prognostic impacts of CDK10 expression on the outcome of patients with gastric cancer. We found that the OS rate of patients with tumors that expressed CDK10 was higher than that of patients with tumors did not express CDK10 (P<0.001) (Fig. 3). Further, we examined whether CDK10 expression may be served as a prognostic factor for gastric cancer patients by utilizing univariate and multivariate analyses. On univariate analysis, CDK10 expression (P<0.001), patient age (P=0.039), tumor size (P<0.001), histological type (P=0.001), stage of tumors (P=0.001) and lymph node metastasis (P<0.001) were significantly correlated with an unfavorable OS (Table II). After adjusting the prognostic factors that achieved significance in univariate analysis, CDK10 expression (P=0.045), histological type (P=0.004) and tumor size (P=0.001) maintained independent significance for predicting the prognosis of gastric cancer patients.

To confirm that CDK10 functions as a tumor suppressor in gastric cancer pathogenesis and progression, we examined the effect of CDK10 on cell proliferation by transfecting a CDK10 expression construct into gastric cancer cell lines HGC-27 and SGC-7901. Overexpression of CDK10 was achieved in these two cell lines (Fig. 4A). As evidenced by MTT assays, we found that proliferation rates of these cell lines with CDK10 overexpression were significantly reduced in comparison to the empty vector-transfected control cells (Fig. 4B). Moreover, HGC-27 and SGC-7901 cells exogenously expressing CDK10 produced fewer colonies than their respective control cells (Fig. 4C).

We subsequently investigated whether CDK10 could regulate the metastasis capacities of gastric cancer cells. Confluent HGC-27 and SGC-7901 cells with or without ectopic CDK10 expression were scratched and cell motility was determined. By doing so, we found that ectopic CDK10 expression, in both gastric cell lines, significantly delayed closure of wound area compared to the empty vector controls (Fig. 4D). Next, we examined the invasive ability of these gastric cell lines using a Matrigel assay. As shown in Fig. 4E, CDK10 overexpression let to a significant decrease in invasiveness of both cell lines, as compared with their own controls. All these observations clearly demonstrated that CDK10 expression restoration inhibits the migration and invasion of gastric cancer cells.

We further assessed the biological effects of silencing CDK10 expression on gastric cancer cells. The efficiency of CDK10 knockdown using a pre-design and validated siRNA was tested by immunoblot analysis (Fig. 5A). We observed that HGC-27 and SGC-7901 cells with CDK10 knockdown exhibited a consistent and significant increase in cell proliferation and colony formation, as compared to those cells transfected with non-specific control siRNA (Fig. 5B and C). Further, the inhibitory effect of CDK10 on cell invasion was confirmed using a Matrigel assay. The results showed that HGC-27 and SGC-7901 cells with CDK10 knockdown showed a significant increase in cell invasive capacity compare to their respective control cells (Fig. 5D). Together, these findings clearly demonstrate that knockdown of CDK10 promotes the proliferation and the migration-invasion of gastric cancer cells.