D-2-hydroxyglutarate impairs DNA repair through epigenetic reprogramming

Abstract

Cancer-associated mutations in IDH are associated with multiple types of human malignancies, which exhibit distinctive metabolic reprogramming, production of oncometabolite D-2-HG, and shifted epigenetic landscape. IDH mutated malignancies are signatured with "BRCAness", highlighted with the sensitivity to DNA repair inhibitors and genotoxic agents, although the underlying molecular mechanism remains elusive. In the present study, we demonstrate that D-2-HG impacts the chromatin conformation adjustments, which are associated with DNA repair process. Mechanistically, D-2-HG diminishes the chromatin interactions in the DNA damage regions via revoking CTCF binding. The hypermethylation of cytosine, resulting from the suppression of TET1 and TET2 activities by D-2-HG, contributes to the dissociation of CTCF from DNA damage regions. CTCF depletion leads to the disruption of chromatin organization around the DNA damage sites, which abolishes the recruitment of essential DNA damage repair proteins BRCA2 and RAD51, as well as impairs homologous repair in the IDH mutant cancer cells. These findings provide evidence that CTCF-mediated chromatin interactions play a key role in DNA damage repair proceedings. Oncometabolites jeopardize genome stability and DNA repair by affecting high-order chromatin structure.

Fig. 1. D-2-HG compromises HR DNA repair via inhibition of TET1/2.

a γH2A.X foci number. U251 cells were treated with DMSO, 150 μM TMZ or 150 μM BCNU for 24 h. Group differences were tested with one-way ANOVA. **p < 0.01. All p < 0.0001; data are mean ± SEM from three independent experiments. b–e Activation of DDR. IDH1^WT, IDH1^R132H U251 or U251 cells were treated with 150 μM TMZ ± 5 μM AG-120 (b); 150 μM TMZ/150 μM BCNU, 0.5 mM D-2-HG, 1 mM αKG (+), 10 mM αKG (++) for 24 h (c–e). f, g Comet assay. f U251 were treated with 150 μM TMZ, 0.5 mM D-2-HG, 10 mM αKG for 24 h. **p < 0.01. All the indicated p < 0.0001; g U251 with TET1/TET2 knockout were treated with DMSO or 150 μM TMZ for 24 h. **p < 0.01. All the indicated p < 0.0001; group differences were tested with one-way ANOVA. Data are mean ± SEM from three independent experiments. h, i Western blotting. U251 with control sgRNA, TET1/TET2 knockout, TET1/TET2 knockout with TET1 or TET2 overexpression were treated with 150 μM TMZ for 24 h. j Flow cytometry analysis. DD-Sce-I-GR DRGFP U2OS were treated with 0.5 μM Shield1 and 0.2 mM TA, 0.5 mM D-2-HG, 10 mM αKG, TET1 or TET2 overexpression for 24 h. k Quantification for (j). *p < 0.05, **p < 0.01. D-2-HG vs D-2-HG + 1 μg TET1, p = 0.01. Other p < 0.0001. Group differences were tested with one-way ANOVA. Data are mean ± SEM from three independent experiments. l Activation of DDR. U251 IDH1^WT/IDH1^R132H with TET1 or TET2 overexpression were treated with 150 μM TMZ. m, n γH2A.X foci number. U251 with TET1/TET2 knockout, TET1/TET2 overexpression was treated with 150 μM TMZ for 24 h. *p < 0.05, **p < 0.01. TET2 KO+ vs TET2 OE, p = 0.0393. All other p < 0.0001. Group differences were tested with one-way ANOVA. Data are mean ± SEM from three independent experiments. Source Data are provided as a Source Data file.

Fig. 2. D-2-HG inhibition on TET1/2 alters DNA damage protein recruitment in DNA damage regions.

a, b Laser micro-irradiation assay. U251 cells were treated with 0.5 mM D-2-HG, 10 mM αKG or TET1 and TET2 overexpression for 24 h. Scale bar: 5 μm. c, d Quantification of the cells with BRCA2 or RAD51 recruitment to DSB sites. A total of 40 cells from 5–9 views (c DMSO, n = 7; D-2-HG, n = 9, D-2-HG + αKG, n = 7, D-2-HG + TET1, n = 7, D-2-HG + TET2, n = 7; d DMSO, n = 7; D-2-HG, n = 8, D-2-HG + αKG, n = 5, D-2-HG + TET1, n = 6, D-2-HG + TET2, n = 7) were obtained for quantifications. **p < 0.01. c D-2-HG vs D-2-HG + αKG, p < 0.0001; D-2-HG vs D-2-HG + TET1, p < 0.0001; D-2-HG vs D-2-HG + TET2, p < 0.0001; d D-2-HG vs D-2-HG + αKG, p < 0.0001; D-2-HG vs D-2-HG + TET1, p = 0.0006; D-2-HG vs D-2-HG + TET2, p < 0.0001; Group differences were tested with one-way ANOVA. Data are mean ± SEM from three independent experiments. e Immunofluorescence staining shows the induction of γH2A.X and RAD51 puncta in the nucleus. DD-Sce-I-GR DRGFP U2OS cells were treated with 0.5 μM Shield1 and 0.2 mM TA in the presence of 0.5 mM D-2-HG with 10 mM αKG, TET1 or TET2 overexpression for 24 h. Scale bar: 5 μm. f Immunofluorescence staining shows the recruitment of RAD51 DSB sites. U251 cells were treated with 150 μM TMZ for 24 h. Scale bar: 5 μm. g Line profile analysis for data shown in (f). Red, RAD51; cyan, γH2A.X. h ChIP-qPCR shows the recruitment of BRCA2, RAD51, and γH2A.X modification around the Sce-I site. DD-Sce-I-GR DRGFP U2OS cells were treated with 0.5 μM Shield1 and 0.2 mM TA in the presence of 0.5 mM D-2-HG with 10 mM αKG. ACTB loci were used as the negative control. Data are presented as mean values ± SEM from three independent experiments. Source Data are provided as a Source Data file.

Fig. 3. Cytosine hypermethylation hampers CTCF binding and subsequent recruitment of HR elements.

a 5hmC Dot blot assay. Cells were treated with 150 μM TMZ, 0.5 mM D-2-HG, 10 mM αKG, TET1 or TET2 overexpression for 24 h, 400, 200, 100, 50, 25, 12.5 ng gDNA was used for the assay. 5hmC was monitored with the antibodies. b MeDIP assay. DRGFP U2OS cells were treated with 0.5 μM Shield1 and 0.2 mM TA for 2 h. *p < 0.05, **p < 0.01. ns: non-significant. For −1500 to 1000 locus, Sce-I+ vs D-2-HG, D-2-HG vs D-2-HG + αKG, all p < 0.0001; −1500, D-2-HG vs D-2-HG + TET1/TET2, p < 0.0001; −500, D-2-HG vs D-2-HG + TET1, p = 0.0423; D-2-HG vs D-2-HG + TET2, p = 0.0230; −180, D-2-HG vs D-2-HG + TET1, p = 0.0014; D-2-HG vs D-2-HG + TET2, p = 0.0006; 150, D-2-HG vs D-2-HG + TET1, p = 0.0010; D-2-HG vs D-2-HG + TET2, p = 0.0034; 500, D-2-HG vs D-2-HG + TET1, p = 0.0060; D-2-HG vs D-2-HG + TET2, p = 0.0156; 1000, D-2-HG vs D-2-HG + TET1, p = 0.0024; D-2-HG vs D-2-HG + TET2, p = 0.0096; group differences were tested with one-way ANOVA. Data are mean ± SEM from three independent experiments. c, d Laser micro-irradiation. Scale bar: 5 μm. e, f Quantification of data shown in (c) and (d). Over 60 cells were obtained from 5–7 imaging views (e, DMSO, n = 5; D-2-HG, n = 5, D-2-HG + αKG, n = 6, D-2-HG + TET1, n = 8, D-2-HG + TET2, n = 7; f DMSO, n = 6; D-2-HG, n = 5, D-2-HG + αKG, n = 5, D-2-HG + TET1, n = 7, D-2-HG + TET2, n = 6). *p < 0.05, **p < 0.01. e All indicated p < 0.0001; f DMSO vs D-2-HG, p < 0.0001; D-2-HG vs D-2-HG + αKG, p < 0.0001; D-2-HG vs D-2-HG + TET1, p = 0.0069; D-2-HG vs D-2-HG + TET2, p = 0.0277; group differences were tested with one-way ANOVA. Data are mean ± SEM from three independent experiments. g ChIP-qPCR assay. DRGFP U2OS cells were treated with 0.5 μM Shield1 and 0.2 mM TA for 0.5, 1, 2, 4, 6, 12, or 24 h. Two genomic regions −180 bp and +150 bp adjacent of DSB sites were monitored. h ChIP-qPCR assay targeting FRA7D, FRA7-2, and FRA21. Data are presented as mean ± SEM from three independent experiments. Source Data are provided as a Source Data file.

Fig. 4. D-2-HG affects the assembly of HR repair complex through TET1 and TET2 inhibition.

a–d PLA assay shows the interactions among CTCF and BRCA2 or RAD51 in the presence of 150 μM TMZ with/without 0.5 mM D-2-HG, 10 mM αKG, TET1 or TET2 overexpression for 24 h. Scale bar: 15 μm. e–h Quantification of PLA foci shown in (a–d). Over 40 cells among three biological replicates were obtained for quantifications. The statistical significance of differences among groups was tested using the one-way analysis of variance (ANOVA). **p < 0.01. e, g TMZ− vs TMZ+, DMSO vs D-2-HG, D-2-HG vs D-2-HG + αKG, D-2-HG vs D-2-HG + TET1, D-2-HG vs D-2-HG + TET2, p < 0.0001; f, h TMZ− vs TMZ+, TET1 KO− vs TET1 KO+, TET1 KO+ vs TET1 OE, TET2 KO− vs TET2 KO+, TET2 KO+ vs TET2 OE, p < 0.0001. i Co-IP experiments show the interactions among CTCF, BRCA2, and RAD51 in the U251 IDH1^WT, IDH1^R132H, and IDH1^R132C cells with/without 10 mM αKG, TET1 or TET2 overexpression. j Co-IP experiments show the interactions among CTCF, BRCA2, and RAD51 in U251 cells in the presence of 150 μM TMZ with/without 0.5 mM D-2-HG, 10 mM αKG, TET1 or TET2 overexpression. k, l Co-IP experiment shows the interactions between BRCA2 and RAD51 in the absence of TET1 or TET2. TET1 and TET2 overexpression were utilized to evaluate their rescue effects on the knockout consequence. Cells were treated with TMZ (150 μM) for 24 h. Source Data are provided as a Source Data file.

Fig. 5. IDH mutation disrupts chromatin structure around DSB through DNA hypermethylation.

a ChIP-seq assay shows alteration of CTCF coverage around AsiSI DSB sites in IDH1^WT and IDH1^R132H DIvA cells. Two biological replicates were performed. b Top, genomic tracks show differential 4C-seq signal (log2[+DSB/−DSB]) in IDH1^WT and IDH1^R132H DIvA cells treated with 4-OHT (300 nM) for 4 h. Bottom, CTCF enrichment at the same genomics loci. Two biological replicates were performed. c Bisulfite PCR shows the DNA methylation status of the CpG islands around AsiSI sites shown in (b). d ChIP-qPCR assay shows the CTCF, BRCA2, RAD51 binding, and γH2A.X modification around the DSB sites. DD-Sce-I-GR DRGFP U2OS cells were treated with 0.5 μM Shield1 and 0.2 mM TA for 0.5, 1, 2, 4, 6, 12, or 24 h before collection for ChIP-qPCR. Two genomic regions (−180 bp and +150 bp) adjacent of DSB sites were monitored. Cells were transfected with TET1 CD/dCD. The recruitments were quantified relative to the IgG control. e MeDIP/hMeDIP assay shows epigenetic status around the Sce-I site. DD-Sce-I-GR DRGFP U2OS cells were treated with D-2-HG (0.5 mM), αKG (10 mM), or overexpression of TET CD/dCD, or TET2. DSB was induced with 0.5 μM Shield1 and 0.2 mM TA. Data are presented as mean ± SEM from three independent experiments. Source Data are provided as a Source Data file.