Abstract 3830: Employing reduced representation bisulfite sequencing for biomarker discovery in colorectal cancer

Using reduced representation bisulfite sequencing (RRBS), we have identified 3 candidate loci with aberrant methylation in colon cancer that in combination have 92% sensitivity for CRC tumor detection with 100% specificity. Aberrant DNA methylation is an important epigenetic event occurring early in the pathogenesis of colorectal cancer (CRC). Detection of aberrantly methylated DNA in CRC can be exploited for clinical utility as a minimally invasive screening and/or monitoring tool by assaying for methylated tumor DNA in either stool or plasma. In this study we employed RRBS to identify novel methylated loci in CRC for use as early detection markers. We first performed RRBS on a discovery set of 21 normal colon epithelia samples and 43 CRC tumor samples seeking loci unmethylated in normal colon yet methylated in CRC, referred to as UnUp loci. We identified 50 such UnUp loci using the following criteria: individual CpGs demonstrating less than 5% DNA methylation level in each normal sample (with > = 20 read coverage), having > = 20 percentage point increased methylation above the normal range in at least 80% CRC tumors (with > = 10 read coverage), and encompassing more than one such informative CpG within a 200 bp interval. We validated these 50 UnUp loci by PCR amplification of each locus followed by next generation bisulfite sequencing in an expanded validation cohort consisting of 40 CRC tumors, their matched normal colon tissue, plus 8 CRC cell lines and their matching parental tumors. The best 12 candidate UnUp biomarkers were then identified that had a combination of highest sensitivity for tumor and highest specificity for normal colon, based on total CpGs methylated in each sequencing read. Individually, these 12 best candidate loci were 6%-90% sensitive in detecting colon cancers, and 92%-100% specific in not detecting normal colon. The 12 best candidate loci were further scrutinized for the feasibility of designing an assay for methylation detection in body fluids. To this end, we restricted window size to ∼100 bp and performed a bioinformatic scan to identify the best small windows containing the greatest number of CpGs simultaneously having the lowest per-CpG methylation in normal samples, and the highest methylation in cancer. A panel of the 3 most robust small windows showed 92% sensitivity and 100% specificity of distinguishing colon cancers from normal colon tissue. The performance of these small-window loci is being further tested in an independent set of 96 CRC tumors with matched normal colon tissue. Other ongoing studies include examining expression of neighboring genes in these loci as well as testing the performance of our top candidates in detecting CRC from patient blood samples.Citation Format: Ryan E. Fecteau, Helen Moinova, Joseph E. Willis, Omar De la Cruz, Thomas LaFramboise, Sanford D. Markowitz. Employing reduced representation bisulfite sequencing for biomarker discovery in colorectal cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3830. doi:10.1158/1538-7445.AM2015-3830

Duke Scholars

Author Ryan Everett Fecteau Radiation Oncology