Reprimo as a Potential Biomarker for Early Detection in Gastric Cancer

Materials and Methods

Clinical samples and cell line. To identified the DNA methylation patterns of 24 genes, we used a testing-validation approach (12). As a testing set, we selected 32 surgically resected gastrectomy specimens with diagnoses of gastric cancer according to the WHO (13). Testing set was composed by 21 males (65.6%) with an age average of 65 years old, 11 (34.3%) of these tumors were located in the cardia, 7 (21.8%) in the middle third, and 14 (43.7%) in the antrum. Four (12.5%) tumors were at early stage and 22 (68.8%) were lymph node positive. Thirteen (40.6%) tumors were signet-ring cell histology according to the WHO (14). As a validation set, 43 prospectively collected endoscopic biopsies and plasma samples from gastric cancer patients with similar characteristics to that of the testing set were obtained together with plasma samples from 31 asymptomatic age- and gender-matched controls. All cases were immediately snap-frozen at the time of surgery or endoscopic procedure. None of cases had family history of gastric cancer. This study was approved by the Institutional Review Board at Hospital San Borja-Arriaran and Pontificia Universidad Catolica de Chile. The cell line MKN45 was used to confirm the association between CpG island hypermethylation and RNA expression of selected genes.

DNA extraction. Five, 15-μm cryostat sections with representative areas of gastric cancer were cut and placed into a 0.5 mL tube for DNA extraction. DNA extraction was done in 100 μL extraction solution [1 mol/L Tris (pH 8.0), 50 mmol/L EDTA, 0.5% and Tween 20] with 1 mg/mL Proteinase K (Sigma) for 12 h at 55°C. Proteinase K was inactivated by boiling at 100°C for 10 min and DNA was purified by phenol-chloroform extraction and ethanol precipitation according to standard protocols. DNA concentration was determined by spectroscopy using 1 OD260 for 50 μg/mL. We chose plasma as a source for DNA analysis based on previous reports indicating that plasma is superior than serum for DNA methylation detection (15).

Methylation assays. DNA was treated with sodium bisulfite as described previously (16). Briefly, 1 μg of genomic DNA was denatured by incubation with 0.2 mol/L NaOH for 10 min at 37°C. Hydroquinone (10 mmol/L; 30 μL; Sigma) and 3 mol/L sodium bisulfite (pH 5.0; 520 μL; Sigma) were added, and the solution was incubated at 50°C for 16 h. Treated DNA was purified by use of Wizard DNA Purification System (Promega Corp.), desulfonated with 0.3 mol/L NaOH, precipitated with ethanol, and resuspended in water. Modified DNA was stored at −80°C until used. The methylation status of 24 genes was determined by Methylation-specific PCR (16). These genes were chosen because they are tumor suppressors genes, methylation at these CpG sites is associated with gene silencing, they cover six essential alterations in cell physiology that collectively dictate malignant growth (self-sufficiency in growth signals, insensitivity to growth-inhibitory signals, evasion of programmed cell death, limitless replicative potential, sustained angiogenesis, and tissue invasion and metastasis), and they had previously been described as undergoing hypermethylation in other tumor types (16–33). Gene names, gene location, function, and reference for methodology for each gene selected in this study is summarized in Table 1 . Only cases with positive unmethylated bands were considered informative for methylation status in this study. Therefore, our rate of informative cases ranges from 6 to 32. Negative control samples without DNA were included for each set of PCR.

Reverse transcription-PCR and demethylating agent 5-aza-2′-deoxycytidine. For all 24 genes selected for analysis, an association between hypermethylation and gene silencing has been previously established (16–33). However, we confirmed this association for SHP1, APC, Reprimo, FHIT, E-cadherin, and SEMA3B by using reverse transcription-PCR in the cell line MKN45, a poorly differentiated gastric cancer cell line from mongoloid female origin. Total RNA was extracted using the Qiagen RNeasy System (Qiagen), according to the manufacturer's recommendations. RNA concentration was determined by measuring absorbance at 260 nm, and quality was verified by the integrity of 28S and 18S rRNA after ethidium bromide staining of total RNA samples subjected to 0.8% agarose gel electrophoresis. Total cDNA was synthesized with Moloney Murine Leukemia Virus reverse transcriptase (ThermoScript RT; Invitrogen). Reverse transcription-PCR was done using 1 μg of total cellular RNA to generate cDNA. Primers sequences have been previously established (29, 31, 32, 34–36). Reverse transcription-PCR for the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene served as a control for RNA loading. Ten microliters of each PCR reaction were directly loaded onto 2% agarose gels and separated by electrophoresis; the gels were stained with ethidium bromide and visualized under UV illumination. To restore silenced gene expression of methylated genes, we treated a cell line for 72 h with the demethylating agent 5-aza-2′-deoxycytidine (Sigma) at a concentration of 2 μmol/L, as described previously (29).

Data analysis. Frequencies of methylation were compared using χ² test or Fisher exact tests. For all tests, probability values of P < 0.05 were regarded as statistically significant.

Results

DNA methylation patterns of the CpG Island of multiple genes. The complete spectrum of DNA methylation patterns of 24 genes observed in 32 retrospectively collected gastric cancer cases is shown in Fig. 1A . Twenty-three (96%) genes showed promoter methylation in at least 3 or more of the tested cases, whereas 11 genes were hypermethylated in at least 50% of cases (APC, SHP1, E-cadherin, ER, Reprimo, SEMA3B, 3OST2, p14, p15, DAPK, and p16). Two of these 11 genes (SHP1 and SEMA3B) have not been reported previously methylated in gastric cancer. No promoter methylation was evident for GSTp1 gene in any of the tested cases. Representative examples of the DNA methylation analysis are shown in Fig. 1B.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Methylation-specific PCR analysis of 32 retrospectivelly collected gastric cancer cases. A, histogram representing the percentage of tumors showing methylation for the 24 genes as indicated. B, illustrative methylation-specific PCR results for seven of the genes studied in 6 cases (lanes 1-6). WM, weight marker; M, PCR product with primers specific for methylated DNA; U, PCR product with primers for unmethylated DNA; CM, MKN-45 in vitro methylated MKN-45 cell line (positive control for methylated genes). CU, peripheral blood lymphocytes DNA (used as a control of unmethylated genes).

DNA methylation patterns and loss of RNA expression in MKN-45 cell line. To establish the association between DNA methylation patterns and gene silencing in gastric cancer, we determined mRNA expression of 5 genes that were methylated in at least 50% of cases (APC, SHP1, E-cadherin, Reprimo, and SEMA3B). To this purpose, we treated the cell line MKN45 with the demethylating agent 5-aza-2-deoxycitidine. As shown in Fig. 2 , the gene expression reactivation was associated with the addition of the demethylating drug.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Representative examples of RT-PCR for (A) SHP1 and (B) Reprimo expression in gastric cancer MKN45 cell line before (0 h) and after (48 h) treatment with 5-Aza-CdR. Expression of the housekeeping gene GAPDH was run as a control for RNA integrity. cDNA was synthesized with (+) or without (−) reverse transcriptase. CM, MKN-45 in vitro methylated MKN-45 cell line (positives control for methylated genes); NC, negative control (water blank).

Clinicopathologic associations. We explore the relationship among clinical and pathologic characteristics of cases and DNA methylation patterns in the testing group cases. Methylation of eight genes (BRCA1, p73, RARβ, hMLH1, RIZI, RUNX3, MGMT, and TIMP3) was statistically associated with a particular variant of gastric cancer, the signet-ring cell type (P = 0.03 by Fisher's exact test). These data suggest that at molecular level, the signet-ring cell type should be considered a distinct subtype of gastric cancer. No other clinical or pathologic associations were found either by multiple- or single-gene analysis.

DNA methylation patterns of the CpG island in prospectively collected gastric cancer cases and asymptomatic controls. Next, to evaluate the significance of the DNA methylation patterns of the CpG island for clinical biomarkers for early diagnosis of gastric cancer, the seven most frequently hypermethylated genes in the testing group (APC, SHP1, E-cadherin, ER, Reprimo, SEMA3B, and 3OST2) were examined in 43 prospectively collected gastric cancer cases from which pair tumor gastric biopsies and plasma samples were available. In addition, plasma from 31 asymptomatic age- and gender-matched controls were also prospectively collected and examined for these genes. As shown in Fig. 3A , only methylation of Reprimo was identified in 97.7% (42 of 43) and 95.3% (41 of 43) of tumor and plasma from gastric cancer patients, respectively. However, among asymptomatic controls, methylation of Reprimo was identified in only 9.7% (3 of 31) of cases tested. These differences were statistically significant (P < 0.00001 by Fisher's exact test). Although methylation of APC was frequently methylated in tumor and pair plasma samples from gastric cancer patients, no differences were observed with pair plasma from asymptomatic controls. Representative examples of these analyses are shown in Fig. 3B-D. The other five genes (SHP1, E-cadherin, ER, SEMA3B, and 3OST2), although frequently methylated in tumor samples were less frequently methylated in pair plasma samples.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Methylation-specific PCR analysis of 43 prospectively collected gastric cancer cases and 31 asymptomatic age- and gender-matched controls. A, histogram representing the percentage of tumors and pair samples showing methylation for the 7 genes as indicated. B to D, illustrative methylation-specific PCR results for two (APC and Reprimo) of the seven genes in tubular/papillary and poorly differentiated types (B), mucinous/signet-ring cell type (C) and asymptomatic controls. D, WM, weight marker, M, PCR product with primers specific for methylated DNA; U, PCR product with primers for unmethylated DNA; CM, MKN-45 in vitro methylated MKN-45 cell line (positive control for methylated genes). CU, peripheral blood lymphocytes DNA (used as a control of unmethylated genes).

Discussion

We and others are currently attempting to define the DNA methylation patterns of each type of human cancer (16, 21, 37–42). In gastric cancer, the first documented epigenetic alteration was the promoter hypermethylation of the DNA mistmatch repair genes (hMSH2 and hMLH1; refs. 8, 43). Although several other genes have been described inactivated by epigenetics (5, 9–11, 18, 44–46), a comprehensive methylation profiling in gastric cancer has not been done up to date. Our approach using 24 genes led us to identify specific genes (BRCA1, p73, RARβ, hMLH1, RIZI, RUNX3, MGMT, and TIMP3) associated with a particular variant of gastric cancer, the signet-ring cell type. It has been suggested that signet-ring cell type is epidemiologically, clinico-pathologic, and molecularly a distinct subtype of gastric cancer (47, 48). Thus, our finding not only identified the methylation profile of this emerging variant of gastric cancer but also gives support to the hypothesis that the hypermethylation of the CpG island does not occur randomly but by specific selection process targeting key tumor suppressor genes (41). Our candidate gene approach also led us to identify a two genes (SHP1 and SEMA3B) significative methylated in gastric cancer. SHP1 (Hematopoietic cell-specific protein-tyrosinephosphatase) is a member of JAK-STAT pathway and located in 12p13. This gene has been described frequently inactivated by methylation in leukemia and lymphomas (49) and more recently in gallbladder carcinoma (21). SEMA3B, a member of the 3p21.3 tumor suppressor cluster, has been also described inactivated by methylation in several neoplasms such as liver, gallbladder, lung, and ovary (16). To our knowledge, this is the first report on SHP1 and SEMA3B methylation in gastric cancer.

Several studies have addressed the diagnostic utility of epigenetic biomarkers in detection of human cancer (7). Methylation abnormalities have been detected in blood or sputum of patients with lung cancer, in serum or plasma of head and neck cancers, in ductal lavage fluid of breast cancer, and in urine from patients with prostate and bladder cancer (7). To explore the diagnostic utility of epigenetic biomarkers in the detection of gastric cancer, we evaluated the most frequently hypermethylated genes from the testing set (APC, SHP1, E-cadherin, ER, Reprimo, SEMA3B, and 3OST2) in an independent prospectively collected validation set. In this validation set, we confirm the high frequency of methylation among primary tissues. However, only two (APC and Reprimo) were frequently methylated (>70%) in pair plasma samples. When these genes were evaluated among plasma samples from asymptomatic controls, only Reprimo was significantly less frequently methylated than the others. Taken together, our data are consistent with previous reports in which Reprimo has been found frequently methylated in several cancers but rarely in nonmalignant tissues (29). However, our results in plasma samples are the first to indicate Reprimo as a potential biomarker for early detection of gastric cancer. Reprimo is a downstream mediator of p53-induced G₂ cell cycle arrest (50). When overexpressed, Reprimo induces cell cycle arrest at the G₂ phase, suggesting that has tumor suppressor function (50). Because functional abrogation of the p53 tumor suppressor gene and its downstream mediators, such as 14-3-3-σ, Reprimo is central to the development of human cancers.

Our findings of APC cannot substantiate the results of Leung et al. (51) who found only 17% of methylation of APC in serum of patients with gastric cancer. This discrepancy may partly be due to differences in the pathogenesis of gastric cancer between South America and Hong Kong, or methodologies used (Methylight versus Methylation-specific PCR). On the other hand, our findings of methylation of CDH1 in plasma of gastric cancer cases were similar to Lee et al. (52), the first to study the feasibility of detecting aberrant methylation in serum of gastric cancer patients.

The prognosis of gastric cancer in prognosis is correlated with tumor invasion (2, 3). Prospective studies using photofluorographic methods has shown 2-fold decrease in gastric cancer mortality (relative risk, 0.52; 95% confidence interval, 0.36-0.74) between screened and nonscreened subjects (53). However, the implementation of such massive screening program in asymptomatic population is expensive. This scenario indicates the necessity of develop methodologies for early detection of gastric cancer. The term “serologic biopsy” is a noninvasive alternative method for massive detection of gastric cancer. The serologic biopsy includes the detection of pepsinogen and gastrin 17 for the identification of gastric atrophy, the precursor lesion of gastric cancer (54). Here, we describe a novel gene, Reprimo, which displayed high frequency of methylation in primary tumors and pair plasma samples in gastric cancer cases but low frequency in plasma samples from asymptomatic controls. Thus, Reprimo should be considered to be included in the serologic biopsy to directly identified gastric cancer.

In summary, our results provide additional information on the significance of epigenetic modification in gastric cancer and delineate distinct methylation profiles of histologic variants of gastric cancer. Our findings are relevant from clinical perspective because we found that promoter hypermethylation of Reprimo might be a potential candidate for early detection of gastric cancer. Further research will be necessary to validated the potential clinical effect of these results.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.