Validation of Data-Mining in Identifying Novel Genes in the Colonic Polyp-Cancer Sequence

Background: We previously described identification of genes associated with colon cancer using a data-mining approach. We sought to validate this approach at a biological level, and determine their role in the polyp-cancer sequence.

Methods: Data-mining was used to select transcripts preferentially expressed in colon cancer libraries. A literature search identified a cohort of these genes that were not previously implicated in this disease process. Primers were designed for these transcripts.

RNA was extracted from ex vivo samples and an in vitro model of colonic neoplasia. Matched normals and an in vitro cell line were used as controls. RT-PCR and Taqman PCR were used to confirm expression of genes of interest.

Results: Of 15 genes highly expressed in colon cancer libraries, seven had not previously been implicated in colonic polyp-cancer sequence. RT-PCR confirmed expression of four of these in colon carcinomas. Of 8 unknown genes highly expressed in colon cancer libraries, two were demonstrated in colon cancer samples. Expression of these genes was also shown in cell line models.

Two genes were selected for further analysis, MYEOV and ETV4. In order to examine the role of these genes in the polyp-cancer sequence, Taqman PCR demonstrated expression of MYEOV in advanced adenomas and carcinomas, and expression of ETV4 in carcinomas only.

Conclusions: We have confirmed the utility of library mining for selection of disease-associated genes. We have demonstrated the novel finding of increased expression of MYEOV in advanced adenomas and carcinomas.