Making Bad Cells Go Good: The Promise of Epigenetic Therapy

Abstract

Listen

Translate article

Among the prevailing explanations for the development of cancer over the past 50 years has risen the notion that clinically apparent cancers represent the failure of the tumor cells to differentiate along a pathway that would be expected for their tissues of origin. This point of view springs from and leads to the familiar characterization of epithelial neoplasms as well-, moderately, and poorly differentiated. This obviously orients our thinking about hematologic cancers, where the neoplastic cells frequently seem to be parodies of normal maturing marrow elements. Indeed, laboratory studies have demonstrated that exposure to various substances causes differentiation of certain hematopoietic tumor cells with their acquisition of more normal characteristics. Historically, the field received a tremendous boost—and a challenge—when Leder and Leder 1 demonstrated that high concentrations of sodium butyrate could cause marked differentiation of murine erythroleukemia cells. The implications that profound regulation of gene expression could be manipulated by so simple a molecule had tantalizing mechanistic and therapeutic implications. The subsequent demonstration that high concentrations of sodium butyrate caused marked change in the acetylation state of histones pointed toward modification of DNA or chromatin-associated proteins as correlating with the basis for this effect. 2 In parallel with these developments, longstanding observations that the methylation status of certain genes could influence their expression had arisen from the observations that there are actually five bases that could be detected in DNA. In addition to the familiar adenine, guanine, thymine, and cytosine, a variable proportion of cytosine existed as a methylated form. Moreover, the pattern of methylated cytosines was conserved from cycle to cycle of replication by the action of methyltransferases so that only one of the two chromosomes bearing certain genes was methylated. The methylated allele tended to be preferentially underexpressed. This nongenetic (in the sense that it did not arise from the sequence of the gene) but heritable feature was described as the mechanism ‘‘imprinting’’ certain maternally, as opposed to paternally, derived gene sets. 3 Methylation