Genetic progression in sporadic endometrial and gastrointestinal cancers with high microsatellite instability

Abstract

This study selected a series of 136 MSI-H (microsatellite instability at high frequency) gastric, colorectal, and endometrial carcinomas combining immunohistochemical analysis for hMLH1 or hMSH2 gene products and microsatellite study. The clinico-pathological profile of all tumours was correlated with the overall instability rates at coding and non-coding repeats, in order to clarify the role and the mutation timing of seven target genes (TGFbetaRII, IGFIIR, BAX, hMSH3, hMSH6, CHK1, and BRCA2) in the progression of an MSI-H neoplasm. Regardless of the primary site, the results confirm a model of oncogenesis in which inactivation of hMLH1, or less frequently hMSH2, may initiate a pathway culminating in a progressive accumulation of frameshifts in coding region (CDR) microsatellites. Comparing gastrointestinal and endometrial tumours, significantly lower levels of microsatellite instability at both coding and non-coding repeats were observed. Among gastric and colorectal tumours, the detection of small shortening within Bat-26 and Bat-25 markers defines a subgroup of MSI-H gastrointestinal tumours invariably characterized by early stage at diagnosis. In these tumours, mutations of TGFbetaRII or BAX genes precede frameshifts in the other tested genes. The analysis of correlations between the mutational and clinico-pathological profiles of advanced gastrointestinal tumours revealed that the higher levels of microsatellite instability at both coding and non-coding repeats were not associated with a more advanced clinico-pathological stage or a less favourable outcome. A significant association was observed between a low number of CDR frameshifts and the presence of lymph-node metastasis in advanced gastrointestinal tumours. The existence of advanced MSI-H tumours with more aggressive behaviour and a 'mild mutator phenotype' could be explained by hypothesizing an overlapping of different mechanisms of tumourigenesis, including both the mutator and the suppressor pathways; this should be tested by further studies.