Hypoxia-induced genome-wide DNA demethylation by DNMT3A and EMT of cancer cells.

Background: Despite the comprehensive advancement in the field of cancer therapeutics, there remains an urgent need to identify new pathophysiological mechanisms that can be targeted in isolation or in combination with existing therapeutic regimens. The epithelial-to-mesenchymal transitions (EMT) induced by hypoxia, cytokines, and growth factors involves acquisition of invasive and migratory properties by cancer cells. Epigenetic alterations of DNA methylations and/or histone modifications cause substantial transcriptomic reprogramming in cancer cells during EMT and metastasis, which can be therapeutically targeted by a thorough understanding of the mutual interactions among the epigenetic processes. Previously, the mammalian DNA methyltransferases (DNMTs) have been shown to possess redox- and Ca⁺⁺- dependent active DNA 5mC demethylation activities in addition to the cytosine methylation activity.

Methods: In this study, we have carried out experiments using a range of molecular, cellular, and genome editing approaches including cell culturing, CRISPR/Cas9-editing, si- or sh-RNA-mediated knockdown, quantitative RT-PCR, western blotting, ChIP-qPCR, Na-bisulfite sequencing, EMT and lung colonization assays in conjunction with DNA methylome and DNMT3A ChIP-Seq analyses, RESULTS: We found that active DNA demethylation activity of DNMT3A is essential for hypoxia-induced EMT of the SW480 colon cancer cells, its global genomic DNA demethylation, and promoter DNA demethylation/transcriptional activation of EMT-associated genes including TWIST1 and SNAIL1. DNMT3A also regulates hypoxia-induced HIF-1α binding to and transcriptional activation of the TWIST1 promoter as well as genome-wide DNA demethylation and EMT of breast cancer and liver cancer cells. Mechanistic analysis supports a regulatory model where hypoxia-induced H3K36me3 mark recruits DNMT3A to demethylate CpG in the hypoxia-responsive element (HRE), thereby facilitating HIF-1α binding and activation of the promoters of EMT genes.

Conclusions: Altogether, this study has provided the first demonstration of a physiological function of the active DNA demethylation activity of the DNMTs. Equally important, our findings have revealed a missing link between the HIF-1α pathway and the O₂-sensing KDM pathway both of which are known to be essential for a wide set of normal and disease-associated cellular processes. Finally, the active DNA demethylation activity of DNMT3A has now emerged as a new potential target for therapeutic development to prevent EMT and metastasis of cancer cells.

Clinical trial number: Not applicable.