Differential cancer predisposition in Lynch syndrome: insights from molecular analysis of brain and urinary tract tumors

Abstract

Hereditary non-polyposis colorectal carcinoma (Lynch syndrome) is among the most common hereditary cancers in man and a model of cancers arising through deficient DNA mismatch repair (MMR). Lynch syndrome patients are predisposed to different cancers in a non-random fashion, the basis of which is poorly understood. We addressed this issue by determining the molecular profiles for different tumors from a nationwide cohort of Lynch syndrome families (approximately 150 tumors in total). We focused on some less prevalent cancers, affecting the brain (n = 7) and urinary tract (five bladder and five ureter uroepithelial cancers and four kidney adenocarcinomas), and compared their molecular characteristics to those of the most common cancers, colorectal, gastric and endometrial adenocarcinomas, from the same families. Despite origin from verified MMR gene mutation carriers, the frequency of high-level microsatellite instability in tumors varied between high (100-96% for ureter, stomach and colon), intermediate (63-60% for endometrium and bladder) and low (25-0% for kidney and brain). In contrast to gastrointestinal and endometrial carcinomas, active (nuclear) beta-catenin was rare and KRAS mutations were absent in brain and urological tumors. Compared with other tumors, frequent stabilization of p53 protein characterized urinary tract cancers. Promoter methylation of tumor suppressor genes discriminated the tumors in an organ-specific manner. Our findings suggest that different Lynch syndrome tumors develop along different routes. Uroepithelial cancers of the ureter (and bladder to lesser extent) share many characteristics of MMR deficiency-driven tumorigenesis, whereas brain tumors and kidney adenocarcinomas follow separate pathways.