Mechanisms of DNA demethylation

Abstract

While every cell in an organism is genetically identical, there are marked phenotypic differences between tissues and organs that are controlled by epigenetic modifications. These epigenetic modifications provide an important role in controlling the machinery of gene expression. DNA methylation is the most stable epigenetic modification and is necessary for cellular activities. Many cancers show dysregulation of DNA methylation, with significant global loss of methylation. While the phenomenon of DNA methylation is well described in mammals, the mechanisms of reversal of methylation are incompletely understood. In previous studies, our laboratory observed rapid DNA demethylation in Jurkat cells (T cell leukaemia) after exposure to oxidative stress. This rapid effect indicated active demethylation and suggested an activation of TET (Ten Eleven Translocation) proteins may be occurring. My research investigates the mechanism of changes in DNA methylation using a barcoded hairpin-bisulfite sequencing technique. This high throughput hairpin-bisulfite assay provides a direct assessment of the methylation status of the DNA strands by linking the complementary strands together with a barcoded DNA hairpin linker. We established and developed a bioinformatic workflow to analyse hairpin-bisulfite sequencing data, initially using the Galaxy online platform and later using UNIX command line tools. We demonstrated hemi-methylation of densely methylated promoter regions, as early as 2 h after DNA replication in cells treated with decitabine, and this hemi-methylation increased with decitabine treatment. With the power of the hairpin-bisulfite sequencing assay, we were able to describe the kinetics and pattern of DNMT1 inhibitor treatment. Even though these results were predicted by previous studies, the kinetics of decitabine-induced hemi-methylation has not been demonstrated before in such detail and clarity. We observed an increase (40%) in unmethylated hairpin reads of the PCDHGA12 promoter after treatment with ascorbate/decitabine compared to untreated controls. This observed rapid demethylation implicates a novel mechanism of active demethylation that has not yet been recognised by researchers in the field. Importantly, we demonstrated that decitabine is capable of erasing the epigenetic memory of somatic cells by full removal of methyl groups from both complementary DNA strands. This study is the first to report active demethylation in somatic cells treated with decitabine. Furthermore, in both single-gene and global investigations, we found loss of DNA methylation in different leukaemia cell lines (Molt4, Nalm6 and HL60) following ascorbate and DNMT inhibitor treatment, as well as an increase in 5-hydroxymethyl cytosine in Jurkat cells treated with decitabine and combined ascorbate/decitabine. We performed a low-coverage methylation sequencing assay (PBAT) to detect changes in global DNA methylation. We observed a gradual loss in global DNA methylation in synchronised Jurkat cells treated with decitabine and combined ascorbate/decitabine, confirming the results from locus-specific demethylation. The clinical relevance of our study is that it supports the premise that ascorbate is necessary to enhance the efficacy of decitabine by promoting the function of TET. It is likely that the demethylation pathways that we are studying operate during the onset of cancer and the existence of molecules that alter TET activity may have implications for modification of this process. Cancer patients are often markedly ascorbate (vitamin C) deficient, so the addition of ascorbate to treatment protocols may increase the clinical efficacy (or toxicity) of such drugs in patients with leukaemia.