Article Text


Microsatellite instability in colitis associated colorectal cancer
  1. S H ITZKOWITZ, The Dr Burrill B Crohn Professor of Medicine Director
  1. The Dr Henry D Janowitz Division of Gastroenterology
  2. Box 1069, GI Division
  3. Mount Sinai School of Medicine
  4. One Gustave Levy Place
  5. New York, NY 10029, USA
  6. email: sitzk{at}

    Statistics from

    See article on page 367

    As molecular pathways of colon carcinogenesis continue to be defined for high risk hereditary colon cancer syndromes and for average risk sporadic colon cancers, it seems that colon carcinogenesis in the setting of chronic idiopathic inflammatory bowel disease (IBD) may be unique.1 Although to some extent this notion seems intuitive because of the substrate of chronic inflammation from which these cancers arise, it is nevertheless curious that despite the setting of chronic inflammation, colon cancers occurring in patients with colitis share several features in common with those that arise in patients with hereditary non-polyposis colorectal cancer (HNPCC). For example, in both conditions there is a tendency for the colorectal cancers to affect young individuals, be multifocal, show a proximal colonic distribution, and display mucinous, signet-ring cell or undifferentiated histology. In addition, HNPCC colorectal cancers often have a rim of surrounding inflammation, referred to as a Crohn's-like reaction. These clinicopathological similarities have prompted some investigators to explore whether colitis associated cancers might in fact share a common molecular pathogenesis with HNPCC colorectal cancers.

    Colorectal cancers in patients with HNPCC occur by virtue of a defective ability to repair DNA base pair mismatches. This results in DNA replication errors that have the net effect of altering genes that are critical for maintaining normal growth and behaviour of colonic epithelial cells. Such replication errors can be identified using markers that detect short repetitive sequences, or microsatellites, located throughout the genome. Errors in replicating these sequences result in a phenotype termed microsatellite instability (MI). Several genes act in concert to orchestrate normal repair of DNA base mismatches, including hMLH1, hMSH2, hMSH3, hMSH6, hPMS1, and hPMS2.2 Except for hMSH3, germline mutations of any one of these genes are known to give rise to tumours of the colorectum and other organs in patients with HNPCC, with mutations of hMLH1 and hMSH2 accounting for the vast majority of known HNPCC families.

    In this issue (see page 367), Cawkwell and colleagues investigated the frequency of MI in colitis associated colorectal cancers to determine whether these tumours might share a similar molecular pathogenesis with HNPCC. They analysed 46 colitis associated colonic adenocarcinomas for MI using a panel of four markers in a fluorescence PCR-based MI assay, and also stained tumours by using immunohistochemistry for loss of hMLH1 and hMSH2. Six (15%) of 41 cases demonstrated MI at one or more marker, whereas only one (2.4%) case had MI at two or more of the four markers. The latter case was also the only one that demonstrated loss of immunohistochemical staining of hMSH2, a useful surrogate marker of higher rates of MI in tumours. The authors suggest that MI is uncommon in colitis associated colon cancers, thus distinguishing these cancers from HNPCC where the frequency of MI positivity is about 85%, but perhaps somewhat similar to sporadic colon cancers which have a 15% MI positivity rate.

    Before accepting that MI is uncommon in colitis associated neoplasms, it is worth considering that other investigators have reported a somewhat higher frequency of MI in colitis associated dysplasias and cancers. Using a panel of five microsatellite markers, Suzukiet al noted MI of at least one marker in 17/63 (27%) patients, and 15/120 (8.3%) dysplastic and cancerous lesions manifested MI of two or more markers.3 These investigators also reported that some colitis associated neoplasms manifested mutations of transforming growth factor β1 type II receptor, a target gene that is often affected by MI in HNPCC.4 Brentnall and coworkers used a panel of 13 markers and observed MI of at least one marker in 40% of cancers and 85% of high grade dysplasias, and MI of at least two markers in 40% of cancers and 46% of high grade dysplasias.5 In that study, MI was quite frequent even in colitic mucosa that was non-dysplastic, regardless of whether the patient had colitis of short (less than three years) or long (more than eight years) duration, and MI was not found in acute, self limited forms of colitis. Furthermore, Heinen and colleagues used a panel of six microsatellite markers and noted MI at one locus in 13% non-dysplastic, 5% dysplastic, and 17% cancerous lesions in colitis, and curiously, only non-dysplastic mucosa manifested MI at two loci.6

    How can we reconcile the differences among these studies? To be sure, the studies differ with regard to the number and types of neoplastic tissues analysed and to some extent the method of analysing MI. However, the crucial difference relates to the choice and number of markers used in the various studies. Because MI testing has been so variable between laboratories, a National Cancer Institute workshop established consensus recommendations in 1998 suggesting that a reference panel of five markers be used to classify a tumour as having either a high (MI-H), low (MI-L), or stable (MS) microsatellite phenotype.7 If the recommended five loci are analysed, a tumour is classified as MI-L if one locus exhibits instability, or MI-H if two or more loci are unstable. If more than the five markers are used, MI-L status would apply to tumours manifesting instability at less than 30–40% of loci, whereas tumours with instability of more than 30–40% of loci would be considered MI-H. Importantly, included among the five reference markers are BAT25 and BAT26, which recognise mononucleotide repeats and are therefore considered more sensitive markers. It is possible that if Cawkwell and coworkers had used these more sensitive markers, the frequency of MI in their colitis associated cancers might in fact be higher than reported.

    As we await additional studies using standard markers, a common theme seems to already be emerging from the existing literature: microsatellite instability, especially the MI-L phenotype, may in fact be fairly common in ulcerative colitis, both in neoplastic and non-neoplastic tissue. Some hypothesise that if the DNA repair machinery is defective in ulcerative colitis mucosa, the oxidative stress that accompanies chronic inflammation can damage DNA and overwhelm the ability of the colonocytes to repair DNA, resulting in the MI phenotype of even non-dysplastic mucosa. Although the biological and clinical significance of low level MI is not yet clear, this issue is receiving increasing attention. For example, germline hMSH6 mutations have recently been reported to give rise to an attenuated HNPCC phenotype.8 Moreover, based upon observations made in yeast models, inactivation or inhibition of hMSH6 might result in errors at mononucleotide tracts, but not dinucleotide or longer repeat elements.9 This raises the possibility that mutations of hMSH6 or perhaps silencing of the gene by methylation might contribute to the neoplastic process in ulcerative colitis, a prospect worthy of further investigation.

    See article on page 367


    View Abstract

    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.

    Linked Articles