Abstract

Since the discovery of altered DNA methylation in cancer 25 years ago, the functional roles of epigenetic changes in cancer have been debated. Two recent converging lines of investigation from our group help to shed light on the key role of epigenetic plasticity in cancer causation. Both approaches involve comprehensive genome-scale analysis of cancer epigenetics. In the first approach, we identified cancer-specific differentially DNA-methylated regions (cDMRs), and found that they also distinguish normal tissue types from each other (tDMRs). In a comprehensive examination of most of the common cancer types, we found a high degree of stochastic variation in methylation within each tumour type, but involving the same loci across tumour types, with intermediate values of variation for premalignant tumours. Furthermore, whole genome bisulfite sequencing of colon cancer shows relative hypomethylation of large (5 kb - 10 Mb) blocks encompassing half the genome and one-third of annotated genes, associated with extreme heterogeneity of gene expression, as well as loss of sharply delimited methylation boundaries at CpG islands. Thus, cancer appears to involve increased stochastic epigenetic variation at loci that play a critical role in normal development. In the second approach, we have investigated the mechanism of epithelial-to-mesenchymal transition (EMT), an extreme example of cell plasticity that is important for normal development, injury repair and malignant progression. We identified a mechanistic basis for EMT that involves similar domains that show epigenetic instability in cancer. Together, these data support a model we have proposed for epigenetic plasticity in normal development, under genetic and environmental control, that would provide a selective advantage for species adapting to a fluctuating environment, but also provide a strong selective advantage to tumour development.