Determination of the replication error phenotype in human tumors without the requirement for matching normal DNA by analysis of mononucleotide repeat microsatellites

Abstract

Microsatellite instability (MI) characterizing tumors with replication errors (RER+ tumors) was first described in colorectal tumors from hereditary non-polyposis colorectal cancer (HNPCC) patients as well as in sporadic cases. It has also been observed in subgroups of extracolonic sporadic tumors, but there is no consensus as to the number of microsatellite loci to examine, and the threshold percentage of unstable loci required to classify a tumor as RER+. We have recently shown that BAT-26, a mononucleotide repeat microsatellite, was quasi-monomorphic in DNA from normal individuals and from colorectal RER- samples, and showed important size variations in RER+ samples. In the present work, we analyzed BAT-26 allelic profiles in tumors of the breast (n = 107), brain (n = 78), stomach (n = 59), prostate (n = 49), esophagus (n = 36), thyroid (n = 31), endometrium (n = 12), and cervix (n = 10) whose RER status was already known, thus extending BAT-26 analysis to a total of 542 human solid tumors. BAT-26 alleles were quasi-monomorphic in RER- samples (475/481) and shortened in RER+ tumors (57/61), including four tumors shown to have been misclassified on the basis of dinucleotide repeat microsatellite analysis. In 3/481 RER- and 4/61 RER+ cases, BAT-26 size variation was important enough to attract attention, but not sufficient to establish the RER status of the corresponding tumors. In these cases, the analysis of BAT-25 and BAT-34C4, two other mononucleotide repeat microsatellites, was necessary to resolve the ambiguity. There were only 3 false positive cases. In conclusion, BAT-26 was able to identify the RER status of 539 out of 542 tumors from various origins (99.5% efficiency) in a single-step experiment without the requirement for matching normal DNA.