Feasibility of a DNA methylation assay for noninvasive CRC screening

B16 Introduction: Colorectal cancer (CRC) is the fourth leading cause of cancer-related death in the Western world. Early-stage CRC often has no symptoms, highlighting the need for screening age-appropriate, asymptomatic individuals. Less than 40% of CRCs are diagnosed at a localized stage when CRC is the most treatable. Colonoscopy is the gold standard screening method but this procedure is invasive, costly and not readily accessible or acceptable to a large percentage of age-eligible individuals. Fecal occult blood testing (FOBT) is non-invasive and inexpensive but suffers from lack of sensitivity and specificity. As data demonstrating the clinical and cost-effectiveness of CRC screening has emerged, the impetus for improving non-invasive testing options has intensified.
 Stool-based screening has the advantage of the test substrate coming in direct contact with cancerous lesions. Epigenetic changes in colonic mucosal cells are early events in CRC development. A blood-based screening test would overcome many practical disadvantages of current options and, when coupled with good performance, would become an excellent means for identifying individuals for colonoscopy.
 Materials and Methods: Marker identification: Using re-expression profiles of colon cancerous cell lines, candidate genes were identified and the most promising markers were tested on tissue using the Base5 methylation profiling platform 1 . Promoter sequences were linked with gene expression to identify epigenetically silenced genes. An established pharmacologic unmasking strategy (5-aza-2’-deoxycytidine (DAC) and trichostatin A (TSA)) for re-expression analysis of epigenetically targeted genes was combined with proprietary advanced bioinformatics tools to identify genes prone to promoter methylation. Marker validation: Differentially methylated genes were validated in tissue using real-time methylation specific PCR (real-time MSP). The best performing tissue markers were selected for testing DNA from stool and blood samples. Sample collection, preparation and processing: Prospectively collected stool and blood samples from multiple centers in Germany and The Netherlands are used in the present study. Individuals with no suspicious findings, adenomas or carcinomas based on colonoscopy findings are enrolled. DNA is isolated using standard DNA isolation methods and bisulphite modified using a commercially available kit. Analyte quantitations are done using real-time MSP.
 Results : Based on re-expression, 145 markers were further tested on the Base5 methylation profiling platform. The 30 most differentially methylated genes were validated in tissue using real-time MSP. Of those, a combination of markers reliably detected CRC in an independent tissue test set. A multicenter screening trial has been initiated in 2006. In this trial subjects of age 50 or above are screened with colonoscopy, FOBT and real-time MSP using DNA from stool and blood. The trial is expected to enroll 1600 individuals within 3 years. The best performing tissue markers were assessed on available stool and blood samples from this ongoing trial. Initial data analysis confirmed that a combination of methylation markers has strong performance characteristics to correctly identify patients with a positive colonoscopy result.
 Reference: 1 Straub, J. et al., AB-104-AACRMD (2007)