A rapid bioinformatic method identifies novel genes with direct clinical relevance to colon cancer

Abstract

Identifying genes whose differential expression affect the survival of patients after primary tumor surgery is a major aim of clinical cancer research. To address this issue we combined rapid bioinformatic search algorithms with quantitative RT–PCR in a panel of clearly defined cases of colorectal carcinomas with detailed patient histories. Search algorithms were written that identified Expressed Sequence Tags (ESTs) from the Unigene EST collection of putative open reading frames (ORFs). Expression ratios of healthy to cancerous tissue of each Unigene ORF were calculated. The first 35 candidates arising from bioinformatic searches were examined for mRNA expression in a panel of 20 well documented cases of colon cancer. Four of these 35 genes showed significant correlations with histopathological parameters. Therefore, their expression was further analysed by quantitative RT–PCR in a larger patient cohort. Kaplan-Meier/log rank statistical tests of up to 49 patients in three of the four genes demonstrated significant association of gene expression with poor survival. All four genes demonstrated a strong association with metastatic tumor progression. Expression of the genes was localized to epithelial cells by in-situ hybridization.