We identified frequently methylated genes in ESGN and we suggest they could be used to screen for gastric cancer at early stages. The methylation markers generally displayed three distinct patterns during multistage gastric tumorigenesis: type 1, cons ...
We identified frequently methylated genes in ESGN and we suggest they could be used to screen for gastric cancer at early stages. The methylation markers generally displayed three distinct patterns during multistage gastric tumorigenesis: type 1, consistently high methylation throughout tumor progression; type 2, high methylation in early stages but low in advanced cancer; and type 3, low methylation in early-stage neoplasia but high methylation in advanced cancer. Also, the distribution of methylation levels of these genes according to tumor progression was gradual with no evidence of bimodality or drastic discontinuity. Therefore, methylation changes of these markers during gastric carcinogenesis is gene- and tumor stage-dependent and that they can serve as useful biomarkers for the different stages of gastric neoplastic progression. In particular, type 1 genes could be useful for screening while type 3 genes should be examined for prognostic value. Of the genes we studied, GDNF was the most stable biomarker, as it was methylated in more than 70% of gastric tumor patients irrespective of tumor stage. Among type 2 markers, RORAa was the most frequently methylated in ESGN. Based on this observation, we developed a combined panel of type 1 and type 2 markers because we expected this combination to cover all stages of gastric neoplastic progression. As a result, use of a panel consisting of GDNF and RORAa differentiated ESGN from NGM with a sensitivity and specificity of 98% and 95%, respectively. Also, use of this panel increased the detection sensitivity for all stages of gastric neoplasia from 78% to 91%. These two genes could, therefore, be useful to molecularly detect ESGN in biopsies, gastric washing fluids or possibly serum studies. Of interest, we found that gastric precancerous lesions and EGCs showed higher methylation levels of type 2 marker genes than did AGCs. This was consistent with previous studies showing the early occurrence of DNA hypermethylation in a subset of genes in gastric precursor lesions and is also consistent with studies in colorectal neoplasia. In contrast, we saw gradual increases in the level and frequency of methylation of type 3 marker genes reaching their peak in AGCs. Another important concern of the study analysis was to examine the role of these markers as an epigenetic field defect marker. Our previous studies demonstrated the involvement of aberrant methylation in an epigenetic field defect for colon cancer37, 38 and gastric cancer.8 In the present study, three (GDNF, RORAa, and ADAM23) out of 5 type 1 and 2 markers showed significant higher methylation in ADJ samples obtained from cancer patients than those in NGM. In particular, the mean level of RORAa methylation in ADJ samples from EGC patients was significantly higher than in those from AGC patients (12.1% vs. 7.2%, P = .032), indicating a potential field defect marker for gastric cancer risk prediction. Besides local promoter CGI hypermethylation, genomic global hypomethylation may promote cancer development through genome instability and proto-oncogene hypomethylation. However, researchers have yet to explore in detail the role of DNA hypomethylation in gastric tumorigenesis and the relationships between global hypomethylation and gastric cancer phenotypes. Of interest, LINE-1 methylation was significantly lower in AGC samples showing well to moderate differentiation or intestinal-type Lauren classification. However, this association was not clear in EGC samples, suggesting the probable presence of a distinct cancer subtype, marked by the demethylation phenotype in AGC. In summary, we identified seven methylation markers for the early detection of gastric neoplasia. Of these, combined methylation analysis using GDNF and RORA was highly sensitive and specific, suggesting that this analysis is a rational approach for EGC screening that should be tested in future studies.