The DNMT1 inhibitor GSK-3484862 mediates global demethylation in murine embryonic stem cells

Abstract

Background: DNA methylation plays an important role in regulating gene expression in mammals. The covalent DNMT1 inhibitors 5-azacytidine and decitabine are widely used in research to reduce DNA methylation levels, but they impart severe cytotoxicity which limits their demethylation capability and confounds interpretation of experiments. Recently, a non-covalent inhibitor of DNMT1 called GSK-3484862 was developed by GlaxoSmithKline. We sought to determine whether GSK-3484862 can induce demethylation more effectively than 5-azanucleosides. Murine embryonic stem cells (mESCs) are an ideal cell type in which to conduct such experiments, as they have a high degree of DNA methylation but tolerate dramatic methylation loss.

Results: We determined the cytotoxicity and optimal concentration of GSK-3484862 by treating wild-type (WT) or Dnmt1/3a/3b triple knockout (TKO) mESC with different concentrations of the compound, which was obtained from two commercial sources. Concentrations of 10 µM or below were readily tolerated for 14 days of culture. Known DNA methylation targets such as germline genes and GLN-family transposons were upregulated within 2 days of the start of GSK-3484862 treatment. By contrast, 5-azacytidine and decitabine induced weaker upregulation of methylated genes and extensive cell death. Whole-genome bisulfite sequencing showed that treatment with GSK-3484862 induced dramatic DNA methylation loss, with global CpG methylation levels falling from near 70% in WT mESC to less than 18% after 6 days of treatment with GSK-3484862. The treated cells showed a methylation level and pattern similar to that observed in Dnmt1-deficient mESCs.

Conclusions: GSK-3484862 mediates striking demethylation in mESCs with minimal non-specific toxicity.

Keywords: 5-Azacytidine; DNA methylation; DNMT1; Decitabine; Demethylation; GSK-3484862; GSK-3685032; Whole-genome bisulfite sequencing.

Fig. 1

GSK-3484862 inhibitor treatment in mESC. A Brightfield images showing morphology of WT and TKO cells after six days of treatment with 0.1% DMSO or GSK-3484862 (2 µM, 20 µM and 200 µM) obtained from ChemieTek. Scale bar = 500 μm B Cell numbers after 30,000 WT or TKO mESC were treated with indicated culture conditions for six days. C Brightfield images of WT cells after 14 days of treatment with 0.1% DMSO, 2 µM or 10 µM of GSK-3484862 from ChemieTek or MedChemExpress Scale bar= 500 μm. D Graphs from two independent experiments show the cell growth of WT or TKO cells passaged and counted every 2–3 days

Fig. 2

Expression of methylation-regulated transcripts in GSK-3484862 treated mESC. RT-qPCR, normalized to Actb, of known methylation-regulated germline genes/transposons in cells treated with GSK-3484862 over the number of days indicated. N = 2–6 biological replicates per sample. Mean and standard error are indicated

Fig. 3

Treatment of mESCs with 5-azanucleosides. A Brightfield images of WT and Dnmt1/3a/3b TKO mESCs after 48 h of exposure to 0.1% DMSO, 0.1 µM or 0.3 µM 5-azacytidine (left columns) and WT cells 46 h after the end of treatment with 5-azacytidine (right column). Scale bar = 500 μm B Brightfield images taken of WT and TKO mESCs after 48 h of exposure to 0.1% DMSO, 0.1 or 0.3 µM decitabine. To enhance survival, these cells were cultured on murine embryonic feeders (MEFs) rather than gelatin, hence the distinct morphology. Scale bar = 500 μm C, D Cell numbers after treatment and recovery for 5-azacytidine (C) and decitabine (D). E, F Expression of methylated genes upon 5-azacytidine (E) or decitabine (F) treatment and recovery, normalized to Gapdh and Actb, respectively. Mean and standard error of two technical replicates are indicated. G V6.5 mESCs after exposure to decitabine. These cells showed similar lethality to J1 WT mESCs

Fig. 4

CpG methylation reduction after treatment with GSK-3484862 or 5-azacytidine. A Global CpG methylation levels of WT cells treated with 0.1% DMSO, Dnmt1/3a/3b TKO cells treated with 0.1% DMSO, and WT cells after six or 14 days of treatment with 2 µM or 10 µM GSK-3484862. Methylation levels of published Dnmt1^−/− and Dnmt3a^−/−3b^−/− knockout cells are shown in comparison [31]. B Methylation levels in WT mESCs after treatment with 0.1% DMSO, 0.1 µM or 0.3 µM 5-azacytidine. C DNA methylation over a 4-MB region over chromosome 6 in samples indicated. The first three samples are from published data [31]. Height of bar indicates extent of CpG methylation in a region, from 0 to 100%. D DNA methylation level of 10 kb regions of the genome are indicated, with each region plotted as a single point. E, F DNA methylation level over the Tuba3b and Dazl genes. G, I DNA methylation level of 10 kb regions of the genome are plotted for the samples shown, with Pearson correlation coefficient indicated. The Dnmt1^−/− and Dnmt3a^−/−3b^−/− data are from published data [31]