Down-regulated expression of OPCML predicts an unfavorable prognosis and promotes disease progression in human gastric cancer

Tissue samples were obtained from 30 gastric cancer patients who underwent surgical resection at the First Affiliated Hospital of Sun Yat-sen University between 2010 and 2014. These samples contained both the tumor and tumor-free locations and used to assess the differential expression of OPCML using immunohistochemical analysis. In addition, we included 133 patients undergoing gastrectomy in the First Affiliated Hospital of Sun Yat-sen University between 1995 and 2004. Survival data were available from these 133 patients, with a median follow-up time of 24.4 months. None of the patients was treated with radiotherapy or chemotherapy preoperatively. Tumors were classified according to guidelines of the International Union against Cancer [12]. All patients provided informed consent for collection of the tissue samples and publication of the related data. This study was approved by the Ethics Committee of the Sun Yat-sen University.

Immunohistochemical analysis of OPCML protein was conducted as previously described [13]. The degree of OPCML immunostaining was evaluated according to both the proportion of positively stained cancer cells and intensity of staining. The proportion of cancer cells was scored according to the following criteria: 0 (no positive cancer cells), 1 (<10% positive cancer cells), 2 (10–50% positive cancer cells), and 3 (>50% positive cancer cells). 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). Staining index was calculated as the proportion score × the staining intensity score. We assessed the expression of OPCML in gastric cancer specimens according to the staining index, which scores as 0, 1, 2, 3, 4, 6, and 9. The staining index score of 4 was applied as the cut-off to differentiate between low and high expression of OPCML. OPCML antibody was purchased from Abcam (Cambridge, UK). The positive and negative controls for immunostaining are applied to assess the specificity of OPCML antibody [Additional file 1]. Assessment of the pathological slides was performed by two experienced pathologists who were blinded to clinicopathological characteristics and clinical outcome of the patients.

Total RNA was extracted using NucleoSpin RNA Kit (Macherey-Nagel GmbH, Germany) according to the manufacturer’s instructions. Reverse transcription PCR (RT-PCR) and quantitative PCR was carried out as previously described [14]. PCR primers for the human OPCML gene were as follows: sense: 5′-CCTAGGTCCTCTGAGCAACG-3′, antisense: 5′-GGTCAAGGTAGCAGGAGCAG-3′. Primer sequences for S12 were as follows, sense: 5′- GCATTGCTGCTGGAGGTGTAAT-3′, antisense: 5′- CTGCAACCAACCACTTTACGG-3′.

Total protein of cells was extracted by lysis buffer and concentration determination was conducted using the DC protein assay method of Bradford (Bio-Rad, Hercules, CA). Twenty mg of protein was loaded and separated in 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred to polyvinylidine difluoride membranes (Millipore, Bedford, MA). The antibodies used for the analysis are as follows, OPCML (Abcam, Cambridge, UK), cyclin-dependent kinase inhibitor 1B (p27), cleaved caspase 3, cleaved caspase 9, poly ADP-ribose polymerase (PARP), phospho-Protein Kinase B (AKT) (Ser473) and phosphor-glycogen synthesis kinase 3β (GSK3β) (Cell Signaling Technology, Inc., Danvers, MA), and GAPDH (Good Here Biotechnology, Hangzhou, P.R. China).

Gastric cancer cell lines (SGC7901, BGC823, AGS, MKN28, NCI-N87, SNU1 and MKN45) were obtained from American Type Culture Collection (Manassas, VA) in 2011. The cell lines were last tested and authenticated via cell line STR genotyping assay in 6 months before the experiment ended. The cell lines were cultured in RPMI-1640 media. Treatment of cell lines (SGC7901, BGC823 and AGS) with 5-aza-2′-deoxycytidine was performed as previously described [15].

The cDNA of human OPCML was inserted into pcDNA3.1 (Invitrogen, Carlsbad, CA), as previously described [16]. SGC7901 and BGC823 cells were transfected with OPCML-pcDNA3.1 and the empty vector pcDNA3.1 using Lipofectamine 2000 according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA), and stable clones were then generated and confirmed as previously described [16].

Genomic DNA was extracted using NucleoSpin DNA Kit (Macherey-Nagel GmbH, Germany) according to the manufacturer’s instructions, and was analyzed by the methylation-specific PCR (MSP) posterior to bisulfite conversion, as previously reported [17]. MSP primers used for OPCML promoter was as follows, M, sense: 5′-CGTTTAGTTTTTCGTGCGTTC-3′, antisense: 5′-CGAAAACGCGCAACCGACG-3′; U, sense: 5′-TTTGTTTAGTTTTTTGTGTGTTTG-3′, antisense: 5′-CAAAACAAAAACACACAACAACA-3′.

Cell viability was examined using Cell Counting Assay Kit-8 (CCK-8) (YiYuan Biotech, Guangzhou, China), according to the manufacturer’s instructions. Anchorage-dependent and -independent growth was respectively assessed by colony formation assay and colony formation in soft agar, as previously described [16]. By flow cytometry, cell cycle assay was performed using propidium iodide staining, and apoptosis assay was conducted using annexin V-FITC/PI staining (BD Biosciences, Erembodegem, Belgium), as previously reported [16, 17].

All the experiments on mice were approved by the Sun Yat-Sen University Committee for Animal Research and conducted in accordance with the highest international standards of humane care in animal experimentation. Four-week-old male athymic BALB/c nude mice were purchased from Vital River Laboratories (Beijing, China). All mice had free access to sterilized food and autoclaved water. In vivo experiments were initiated posterior to 1 week of acclimatization. 1 × 10⁷ suspended SGC-7901 cells (in 0.1 ml phosphate-buffered saline) transfected with OPCML or empty vector were injected subcutaneously into the dorsal left flank of each mouse. Tumor size was measured at each third day for 4 weeks. Tumor volume (mm³) and tumor weight were determined as previously described [17].

Student’s t test, One-way ANOVA, Mann-Whitney U test and X² test were used to determine the statistical difference. The survival curves were plotted by the Kaplan–Meier method, and comparison of survival times was conducted using the log-rank test. All tests were two-sided, and a P < 0.05 was considered to be statistically significant. Statistical analysis was carried out using SPSS 16.0 and Graphpad Prism V4.0.

In the first place, we performed immunohistochemistry to assess OPCML expression in 30 paired gastric cancer tissues and their adjacent normal gastric tissues. Abundant OPCML protein was easily detectable in non-cancerous gastric tissues and was shown to be majorly expressed in cell membrane and cytoplasm. Twenty-four of thirty patients exhibited the significantly decreased or lost OPCML expression in malignant tissues, as compared to the matched normal stomach tissues (Fig. 1 a1, b; P < 0.001).

Subsequently, we obtained tumor samples from 133 patients with adenocarcinoma of stomach and evaluated the differential expression of OPCML in gastric cancer, using the normal stomach tissues as control (Fig. 1 a2–4). Low expression of OPCML protein was exhibited in tumor tissues from 96/133 (72.2%) patients with gastric cancer (Table 1). Moreover, OPCML expression was found to be completely lost in samples from 45/133 (33.8%) gastric cancers. We subsequently analyzed the association between OPCML expression and clinicopathological characteristics of gastric cancer patients. Of note, tumors with more advanced tumor stages T3 and T4 tended to exhibit higher rates of low expression of OPCML compared with tumor stages T1 and T2 (Table 1, P = 0.007). Furthermore, OPCML expression was significantly associated with tumor grading, with poorly differentiated tumors possessing higher probability to demonstrate low expression of OPCML as compared to highly and moderately differentiated ones (Fig. 1 a3–4, Table 1, P < 0.001). Whereas, the results did not reveal the significant correlation between OPCML expression and other clinicopathological characteristics including age, gender, nodal status or state of metastasis (Table 1).

We next assessed the correlation between OPCML expression and clinical outcome of patients with gastric cancer. As indicated in Fig. 1c and Table 2, OPCML expression exhibited a significant association with overall survival time of gastric cancer patients (P = 0.031). The patients with tumors of low OPCML expression had a significantly shorter mean overall survival time in comparison to patients with high OPCML expression (26.1 vs 39.7 months; Table 2). To assess whether OPCML expression is an independent prognostic factor for gastric cancer, we conducted a Cox multivariate regression analysis including age, tumor stage, nodal status, state of metastasis, grading and OPCML expression (Table 3). The results point to the independent prognostic significance of reduced OPCML expression in gastric cancer (HR = 2.34, 1.38–3.97; P = 0.002).

We detected the expression of OPCML in seven gastric cancer cell lines. The results showed that expressions of both OPCML mRNA and protein were markedly down-regulated or lost in all seven gastric cancer cell lines, while it is readily detectable in the normal stomach tissues (Fig. 2a, b).

Using MSP, we determined the promoter methylation of OPCML in these seven gastric cancer cell lines. Interestingly, OPCML promoter was methylated in all seven cell lines, while not in the normal stomach tissues (Fig. 2c). Next, the expression changes of OPCML mRNA were analyzed in three cell lines (SGC7901, BGC823, AGS) after incubation with the methylation inhibitor 5-aza-2′-deoxycytidine. As shown in Fig. 2d, all these three cell lines exhibited restored mRNA levels of OPCML posterior to the treatment with 5-aza-2′-deoxycytidine. These data suggested the significant correlation of OPCML mRNA expression and promoter methylation.

To understand the function of OPCML gene in gastric cancer, we assessed the effect of ectopic expression of OPCML on the growth of SGC-7901 and BGC-823 cells in which OPCML gene was silenced. As shown by RT-PCR and western blot, OPCML was significantly over-expressed in SGC-7901 and BGC-823 cells after stably transfected with OPCML-pcDNA3.1 plasmid (Fig. 3a, b). Representative clones of SGC-7901 and BGC-823 transfectants were cultivated and assayed for their viability using CCK-8 assay. Ectopic expression of OPCML resulted in an approximately 36% (P < 0.01) and 42% (P < 0.01) decrease of cell viability in SGC-7901 and BGC-823 cells at the 5th day of experiment, respectively (Fig. 3c, d). Colony formation assay revealed that the number of colonies formed by OPCML transfectants were significantly fewer compared with empty vector transfectants. OPCML SGC-7901 and BGC-823 transfectants exhibited an approximately 54% (P < 0.001) and 65% (P < 0.001) reduction in colony formation respectively, in comparison to empty vector–transfected control cells (Fig. 3e, f). Soft agar assay further confirmed the significant suppressive effect of OPCML on the growth of SGC-7901 (P < 0.0001) and BGC-823 cells (P < 0.001) (Fig. 3g, h).

We further examined whether OPCML inhibited the growth of gastric cancer cells in vivo. As indicated in Fig. 3i, OPCML-pcDNA 3.1 transfection led to the significant reduction of the tumor volume formed by SGC-7901 cells in nude mice (P < 0.01). When the in vivo experiment was concluded till the 27th day, mice with OPCML transfectants had a 66.7% decrease in tumor weight as compared to the empty vector (Fig. 3j, P <0.001).

The following cell cycle analysis by flow cytometry indicated that, after trasnfected with OPCML-pcDNA3.1 plasmid, an elevated percentage of both SGC-7901(from 35.5% to 60.5%, P < 0.01) and BGC-823 (from 45.3% to 68.8%, P < 0.01) cells accumulated in the G0/G1 phase, as compared to cells transfected with empty vector (Fig. 4a, b). While ectopic expression of OPCML led to a decreased proportion of cell population of both SGC-7901 and BGC-823 cells at S and G2/M phase (all P < 0.05) (Fig. 4a, b). These results revealed that OPCML suppressed proliferation of gastric cancer cells by arresting cell cycle in the G0/G1 phase.

Because apoptosis was also frequently associated with cell growth inhibition by tumor suppressor, Annexin V-FITC/PI flow cytometric analysis was used to determine the effect of ectopic OPCML expression on apoptosis of SGC-7901 and BGC-823 cells. The analysis demonstrated a significant increase of cell population of both early apoptosis (P < 0.01) and late apoptosis (P < 0.01) in SGC-7901 cells transfected with OPCML-pcDNA 3.1, as compared to empty vector transfectants (Fig. 4c1, c2). Similar results were indicated in OPCML transfected BGC-823 cells, with a significant elevation of the percentage of both early apoptotic cell population (P < 0.01) and late apoptotic population (P < 0.01), compared with cells transfected with empty vector (Fig. 4d1, d2).

We further analyzed the expression of genes implicated in cell cycle arrest and apoptosis induction. Western blot was used to assess the expression of p27, an important regulator involved in transition checkpoint of G1 to S phase, and the expressions of the pro-apoptotic regulators, encompassing the active form of caspase-3, caspase-9 and nuclear enzyme poly (ADPribose) polymerase (PARP). As shown in Fig. 4e, expression of p27 protein was significantly up-regulated in both SGC-7901 and BGC-823 cells by ectopic OPCML expression. Moreover, expressions of activated form of caspase-3 and caspase-9, and PARP were markedly elevated in SGC-7901 and BGC-823 cells by OPCML (Fig. 4e). To investigate whether the activity of AKT was associated with the effects of OPCML on proliferation and apoptosis, we determined the phosphorylation status of AKT posterior to OPCML plasmid transfection. The results showed that AKT was constitutively activated in the two gastric cancer cell lines, and the phosphorylation level of AKT was significantly decreased by ectopic expression of OPCML. We next determined the effect of OPCML transfection on the phosphorylation status of the AKT downstream target, GSK3β. The constitutive phosphorylation of GSK3β was present in gastric cancer cells and its phosphorylation level was shown to be markedly reduced by ectopic OPCML expression (Fig. 4e).