Article Text

Download PDFPDF

Germline testing of mismatch repair genes is not aided by prescreening tumours for allelic loss
  1. R L Ward1,
  2. E Mokany2
  1. 1Department of Medical Oncology, St Vincent’s Hospital Sydney, Sydney, Australia
  2. 2School of Medicine, University of NSW, NSW, Australia
  1. Correspondence to:
    Dr R L Ward
    Department of Medical Oncology, St Vincent’s Hospital, Darlinghurst, Sydney 2010, Australia;

Statistics from

Immunostaining and microsatellite testing of tumours is increasingly being used to guide germline testing in individuals with suspected hereditary non-polyposis colorectal cancer (HNPCC).1 While the aim of these prescreening tests is to minimise the cost and maximise the chance of identifying a pathogenic germline change, it is clear that neither alone is ideal. In clinical practice, germline testing can often only be justified where an individual has developed a tumour which is microsatellite unstable, and which fails to express a mismatch repair protein. Clearly, this approach is imperfect as not all pathogenic germline mutations are associated with failure of expression of the mismatch repair proteins. The aim of this pilot study was to retrospectively assess the utility of loss of heterozygosity studies in predicting the mutated mismatch repair gene.

Seven individuals with germline mutations in hMSH2 were identified from the family cancer clinic at St Vincent’s Hospital, Sydney. The tumours from each of these individuals were microsatellite unstable and failed to express hMSH2, but demonstrated normal expression of hMLH1. For loss of heterozygosity (LOH) analysis, we used microsatellite markers D1S180 and D1S235 (for Exo1), PMS1.1, D2S118, and D2S155 (for PMS1), D2S21153, D2S2156, D2S2292, D2S2369, and D2S378 (for hMSH2 and hMSH6) and D3S1447 and D3S3685 (for hMLH1). Only heterozygous loci were regarded as informative and LOH was scored when there was a major reduction (at least 50%) or total loss of one allele in the tumour compared to normal tissues.

Of the seven tumours examined in this study, six showed allelic loss of hMSH2, suggesting that the residual normal allele was silenced by LOH (fig 1). In five tumours, allelic loss of hMSH2 occurred in association with LOH in at least one other mismatch repair gene. Only one tumour had retained heterozygosity at all assessable loci, possibly indicating that a mutation had caused the second hit in this tumour.

Allelic loss of hMSH2 often occurs in association with germline mutations but it is clear that loss of the other mismatch repair genes is also a frequent finding. Screening tumours for LOH should not be employed to select patients for mutation analysis of mismatch repair genes. The use of immunohistochemistry and microsatellite testing remain the best available prescreening tools.

Figure 1

Loss of heterozygosity analysis of four mismatch repair genes tumours from seven individuals with germline mutations in hMSH2. White areas, retention of heterozygosity; grey areas, not informative; black areas, loss of heterozygosity.


Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.