G9a inhibition potentiates the anti-tumour activity of DNA double-strand break inducing agents by impairing DNA repair independent of p53 status

Abstract

Cancer cells often exhibit altered epigenetic signatures that can misregulate genes involved in processes such as transcription, proliferation, apoptosis and DNA repair. As regulation of chromatin structure is crucial for DNA repair processes, and both DNA repair and epigenetic controls are deregulated in many cancers, we speculated that simultaneously targeting both might provide new opportunities for cancer therapy. Here, we describe a focused screen that profiled small-molecule inhibitors targeting epigenetic regulators in combination with DNA double-strand break (DSB) inducing agents. We identify UNC0638, a catalytic inhibitor of histone lysine N-methyl-transferase G9a, as hypersensitising tumour cells to low doses of DSB-inducing agents without affecting the growth of the non-tumorigenic cells tested. Similar effects are also observed with another, structurally distinct, G9a inhibitor A-366. We also show that small-molecule inhibition of G9a or siRNA-mediated G9a depletion induces tumour cell death under low DNA damage conditions by impairing DSB repair in a p53 independent manner. Furthermore, we establish that G9a promotes DNA non-homologous end-joining in response to DSB-inducing genotoxic stress. This study thus highlights the potential for using G9a inhibitors as anti-cancer therapeutic agents in combination with DSB-inducing chemotherapeutic drugs such as etoposide.

Keywords: Cancer epigenetics; Chemical probes; Chemotherapeutics; Non-homologous end joining; UNC0638.

Fig. 1

Screening chemical probe inhibitors identifies UNC0638 as potentiating with phleomycin to impede cell proliferation. (A) Schematic elaborating the screening strategy. (B) U2OS cells were seeded in 96 well plates at ~10% confluence and treated with vehicle (1% DMSO), small-molecule inhibitors targeting chromatin modulators (Chr. Regulators; PFI-1, GSK2801, UNC1215, PFI-3), histone acetyl-transferases (HATs; I-CBP112, SGC-CBP30) and histone methyl-transferases (UNC0638, SGC0946, GSK343, PFI-2, UNC1999). Details of inhibitors used are provided in Supplementary Table S1. One-hour post treatment, cells were damaged with low doses of phleomycin (0.5 µM or 1 µM) and were allowed to proliferate for 6 days. The area occupied by cells (% confluence) over time is a surrogate marker for proliferation. Data presented here provide percentage growth rates of cells over a period of 6 days. Error bars correspond to SDs of three independent experiments. Small-molecule Inhibitor UNC0638 inhibiting G9a methyl-transferase hypersensitised U2OS cells to DSB-inducing agent phleomycin at both phleomycin concentrations tested.

Fig. 2

G9a inhibition potentiates the anti-tumour activity of DSB-inducing agents. (A) Growth curves for U2OS cells show that co-treatment of UNC0638 (1 µM) with phleomycin (1 µM) significantly inhibits tumour cell growth compared to cells treated with phleomycin alone. (B) Representative images of DAPI stained cell nuclei at the end-point of the proliferation assay are shown. (C) Western blot analysis reveals a reduction in levels of histone H3K9 dimethylation (H3K9me2) upon UNC0638 treatment. Total levels of histone H3 and tubulin served as loading controls. (D and E) Growth curves of U2OS cells demonstrate that co-treatment of UNC0638 (1 µM) hypersensitised tumour cells to etoposide (100 nM) but not carboplatin (15 µM). At the same concentrations of phleomycin and etoposide, UNC0638 did not adversely affect the cell growth of non-tumorigenic RPE1 and MRC5 cells (see Supplementary Fig. S3). (F) Structurally distinct chemotype of G9a small-molecule inhibitor, A-366, similarly hypersensitises U2OS cells to phleomycin. (G) Western blots confirm G9a depletion and corresponding reductions in H3K9 dimethylation (H3K9me2) upon treatment of U2OS cells with three independent siRNAs (siG9a-1, 2, 3) compared to control siRNA-transfected cells (siContl). Tubulin and total histone H3 were used as loading controls. (H) Depletion of G9a, the target of UNC0638 and A-366, with three independent siRNAs (siG9a-1, 2, 3) hypersensitises U2OS cells to phleomycin (1 µM) and etoposide (100 nM) compared to control siRNA-transfected cells (siContl). Histograms depict % confluence of cells at the endpoint of the proliferation assays. (I) Histograms provide the % confluence of cells at the endpoint of the proliferation assays with indicated treatments of U2OS cells showing that UNC0638 does not further increase the hypersensitivity of G9a depleted cells (siG9a-1, 2, 3) to phleomycin (1 µM) supporting the effects of UNC0638 being via G9a inhibition. Error bars correspond to standard deviations (SDs) of three independent experiments. Non-significant p-values are represented as ns.

Fig. 3

Under low damage condition UNC0638 induces tumour cell death independent of p53 status. (A) U2OS cells were treated as indicated for 4 days and stained with Annexin V and PI for flow cytometry analyses. Combined treatment of UNC0638 (1 µM) and low dose phleomycin (1 µM) significantly increases U2OS cell death as measured by percentage of Annexin V/propidium iodide double-positive cells compared to cells treated with phleomycin alone. Error bars correspond to SDs of three independent experiments. (B) Growth curves for HCT116 p53+/+ (WT) and p53−/− (KO) cells with indicated treatments show that co-treatment of UNC0638 (1 µM) with phleomycin (0.5 µM) inhibits growth of HCT116 cells independent of p53 status. Representative phase-contrast images of cells at the end-point of the confluence assay are shown. (C) HCT116 p53 WT and KO cells were treated with indicated treatments for 4 days and stained with Annexin V and PI for flow cytometry analyses. UNC0638 treatment in the presence of low dose phleomycin increases the cell death (% Annexin V/propidium iodide double-positive cells) for both p53 WT and KO cells compared to phleomycin treatment alone. Statistical analyses were performed as in (A). (D) No increase in PARP1 cleavage was observed upon co-treatment of UNC0638/phleomycin compared to phleomycin treatment only. See also Fig. S4.

Fig. 4

G9a inhibition impairs DNA DSB repair via NHEJ. (A) Representative immuno-fluorescent images of U2OS cells stained with antibodies recognising 53BP1, γH2AX and nuclear stain DAPI (all in grey) after indicated treatments for 4 days are shown. Dotted lines mark nuclear peripheries and the scale bar represents 10 µm. (B) Quantification of average numbers of γH2AX and 53BP1 foci per cell upon the treatments indicated in (A). Error bars correspond to SDs of three independent experiments (>100 cells were analysed per condition per experiment). Combined treatment of UNC0638 with phleomycin significantly increases the average number of γH2AX and 53BP1 foci per cell compared to phleomycin treatment alone. (C and D) DNA repair efficiencies were assayed by neutral comet assay. After the indicated treatments, U2OS and HCT116 p53 WT and KO cells were damaged with phleomycin (26 µM) for two hours (damaged), and were allowed to repair for 2 hours (recovery) after washing off the phleomycin, in the presence of indicated treatments. DSB repair efficiency was measured as the ratio of comet tail moments in recovery by damaged condition. UNC0638 treatment impaired DSB repair both in U2OS and HCT116 cells. (E) Depletion of G9a using three independent siRNAs (siG9a-1, 2, 3) impaired DSB repair upon phleomycin treatment. Depletion of ATM Kinase (siATM) served as a positive control. Comet assays were conducted as in (C). (F) Percentage efficiency of NHEJ upon depletion of G9a with three independent siRNA (siG9a-1, 2, 3) measured by random plasmid integration. Depletion of XRCC4 (siXRCC4) and control siRNA (siContl) served as positive and negative controls, respectively. Error bars correspond to SDs of three independent experiments. See also Fig. S6.