Loci for efficient detection of microsatellite instability in hereditary non-polyposis colorectal cancer.

Most hereditary non-polyposis colorectal cancer (HNPCC) is due to germline mutations in DNA mismatch repair genes. Tumors arising as a result of these mutations display instability in microsatellites, which are short tandem repeats of DNA that are distributed throughout the genome. Although a subset of sporadic colorectal carcinomas also have microsatellite instability (MSI), the phenotype is a useful screening test in identifying patients with HNPCC caused by mutations in mismatch repair (MMR) genes. Studies have shown that some microsatellite markers are more efficient than others in identifying tumors with MSI. Furthermore, the frequency of instability can be assessed by categorizing patients into high (MSI-H, >/= 30-40% positive markers), low (MSI-L), and microsatellite stable (MSS) groups. Using a panel of 28 microsatellite markers, tumor and normal DNA from 10 HNPCC patients was used to identify the five most efficient markers for detecting MSI (BAT26, D2S123, FGA, D18S35, and TP53-DI). Each of the five markers detected MSI in 80-100% of the cases examined. We then expanded the sample size to 17 tumors from HNPCC patients. Each case had evidence for a mutation in either hMSH2 or hMLH1. We compared the efficiency of our panel of five best markers with another panel of five markers (BAT25, BAT26, D2S123, D17S250, and D5S346) identified as being efficient markers for detection of MSI at a recent NCI workshop. Our five selected markers were more efficient (85% vs. 79%) in detecting MSI. However, using either panel, 100% of the cases fell into the MSI-H category and the probability of misclassifying an MSI-H case as MSI-L is very low (0.002-0.008). We also examined four cases meeting the Amsterdam criteria for HNPCC, but with no evidence for mutation in either the hMSH2 or hMLH1 gene. With our panel, three were classified as MSI-H, while only two were classified as such with the NCI reference panel. The probability of misclassifying an MSI-L case as an MSI-H, using a panel of five markers is high (0.263).