Epigenetic stochasticity, nuclear structure and cancer: the implications for medicine.

The aim of this review is to summarize an evolution of thinking about the epigenetic basis of human cancer, from the earliest studies of altered DNA methylation in cancer to the modern comprehensive epigenomic era. Converging data from epigenetic studies of primary cancers and from experimental studies of chromatin in development and epithelial-mesenchymal transition suggest a role for epigenetic stochasticity as a driving force of cancer, with Darwinian selection of tumour cells at the expense of the host. This increased epigenetic stochasticity appears to be mediated by large-scale changes in DNA methylation and chromatin in domains associated with the nuclear lamina. The implications for diagnosis include the potential to identify stochastically disrupted progenitor cells years before cancer develops, and to target drugs to epigenetic drivers of gene expression instability rather than to mean effects per se.