Aberrant DNA Methylation: Have We Entered the Era of More Than One Type of Colorectal Cancer?

Abstract

It has been known for over two decades that DNA from colorectal cancer (CRC) is hypomethylated compared to that from normal colonic epithelia. 1,2 More recently it has also been recognized that specific CpG dinucleotide islands are often hypermethylated in this tumor type, 3 giving rise to the term CpG island methylator phenotype, or CIMP+. 4 Hypomethylation may contribute to the activation of oncogenes, whereas CpG-island specific hypermethylation has been linked to the transcriptional silencing of tumor suppressor genes. The relationship between these two epigenetic phenomena has not been examined, however, in CRC. Intuitively one might expect that tumors with severe global DNA hypomethylation might also show a lower frequency of CpG island hypermethylation due to a lower overall content of methyl groups. The meticulous study by Bariol et al 5 in this issue of The American Journal of Pathology reveals this not to be the case. No association was apparent between global 5-methyl cytosine content and the number of methylated CpG sites in a relatively large and representative series of 97 premalignant and malignant lesions. With only one exception, all of the 22 CRCs investigated showed a lower 5-methyl cytosine content compared to normal mucosa from the same patient. Moreover, approximately 30 to 40% of hyperplastic polyps, adenomas, and CRCs showed no methylation at any of the five CpG sites examined. This most recent study by the Ward laboratory follows on from their previous work in characterizing CIMP+ CRCs. 6 It is also relevant to the notion put forward by these and other workers that CIMP+ and the closely related tumor subgroup with microsatellite instability (MSI+) originate in large hyperplastic polyps and serrated adenomas. 7,8 Increasingly, it is becoming recognized that tumors with widespread DNA hypermethylation have distinctive clinicopathological and molecular features. 4,6,9 Although the study by Bariol et al 5 emphasizes and probably serves to justify the current interest in hypermethylation as opposed to hypomethylation, there is clearly much more to be learned about the general role of aberrant DNA methylation in the development and progression of CRC.