BRD4 Inhibition Is Synthetic Lethal with PARP Inhibitors through the Induction of Homologous Recombination Deficiency

Abstract

Poly(ADP-ribose) polymerase inhibitors (PARPi) are selectively active in cells with homologous recombination (HR) deficiency (HRD) caused by mutations in BRCA1, BRCA2, and other pathway members. We sought small molecules that induce HRD in HR-competent cells to induce synthetic lethality with PARPi and extend the utility of PARPi. We demonstrated that inhibition of bromodomain containing 4 (BRD4) induced HRD and sensitized cells across multiple tumor lineages to PARPi regardless of BRCA1/2, TP53, RAS, or BRAF mutation status through depletion of the DNA double-stand break resection protein CtIP (C-terminal binding protein interacting protein). Importantly, BRD4 inhibitor (BRD4i) treatment reversed multiple mechanisms of resistance to PARPi. Furthermore, PARPi and BRD4i are synergistic in multiple in vivo models.

Keywords: BRD4 inhibitor; CtBP-interacting protein; CtIP; PARP inhibitor; homologous recombination.

Conflict of interest statement

Declaration of interests: G.B.M. has licensed an HRD assay to Myriad Genetics; is a SAB (Scientific Advisory Board) member/consultant with AstraZeneca, Catena Pharmaceuticals, Critical Outcome Technologies, ImmunoMET, Ionis, MedImmune, Nuevolution, Pfizer, SignalChem Lifesciences, Symphogen, Takeda/Millennium Pharmaceuticals, and Tarveda; and has stock options with Catena Pharmaceuticals, ImmunoMet, Spindle Top Ventures, and Tarveda. G.B.M. receives research support from AbbVie, AstraZeneca, Critical Outcomes Technology, Illumina, Ionis, Karus Therapeutics, NanoString, Pfizer, Takeda/Millennium Pharmaceuticals, and Tesaro. M.H., M.O’.C., S.F. and H.C. are employees of AstraZeneca.

Copyright © 2018 Elsevier Inc. All rights reserved.

Figures

Figure 1. Effect of BRD4 inhibition on HR

(A) Heat map (left) and HRD scores (right) from unsupervised clustering of HRD gene signatures using GSE29799 dataset. Higher scores represent defective HR. Statistical significances were determined using Student’s t test. (B) Relative HRD score represents change (treated minus control) in HRD scores in indicated GEO datasets after BRD4 inhibition. The top symbol indicates method of BRD4 inhibition used. Green circle size indicates change in HRD scores, while color indicates −log (p) by Student’s t test. (C) U2OS DR-GFP cells were treated with 100 nM JQ1 or 100 nM AZD5153 for 24 hr (upper), or transfected with control or BRD4 siRNA for 24 hr (lower) and then transfected with I-Sce1 endonuclease for 48 hours. HR efficiency of treated cells was compared to DMSO or control siRNA respectively based on percentage of GFP+ cells detected by flow cytometry. Data represent mean±SEM of three independent experiments. Student’s t test: *** p<0.001. (D) Results of Ingenuity Pathway Analysis (IPA) of pathways significantly altered by BRD4 inhibition in the indicated GEO datasets. Symbol of intervention is as in B. Numbers in the box correspond to −log(p) calculated by Benjamini-Hochberg (B-H) method. See also Figure S1.

Figure 2. Effect of BRD4 inhibition on CtIP expression

(A) Heat map of RPPA data representing “rank-ordered” changes induced by BRD4i treatment (see STAR methods). Proteins with consistent decreases (green) are on the left and increases (red) are on the right of the heat map. Statistically significant changes (z scores) indicated in boxes. (B) Western blot of indicated proteins in HOC1 cells treated with indicated dose of JQ1 for 48 hr (left) or treated with 200 nM JQ1 for the indicated length (right). (C) Western blot of indicated proteins in HOC1 cells after BRD4 silencing for 48 hr. (D) Western blot of indicated proteins in HOC1 cells treated with 200 nM JQ1 or 200 nM GSK1324726A for 48 hr. (E) Correlation between BRD4 and CtIP protein expression in MCLP database. (F) Representative image of IHC with BRD4 or CtIP antibody (left) and correlation between BRD4 and CtIP expression by IHC (right) in ovarian cancer tissues. Scale bar, 25 μm. (G) Correlation between BRD4 and CtIP protein expression in NCI60 dataset. (H) IPA with genes in CtIP coexpression signature. (I) GSEA plot of the CtIP coexpression signature in GSE29799 after BRD4 inhibition (see STAR methods). (J) GSEA plot of Enrichment Score (ES) of CtIP coexpression signature in indicated GEO datasets after BRD4 inhibition. Symbol of intervention is as in Figure 1B. See also Figure S2.

Figure 3. BRD4 binding to CtIP promoter and enhancer and effect on CtIP transcription

(A) qRT-PCR analysis of cMYC and RBBP8 in cells treated with 200 nM JQ1 (upper), or 200 nM AZD5153 (middle) for 24 hr, or after silencing of BRD4 or CtIP by siRNA for 48 hr (lower). (B) Schematic diagram of BRD4 binding regions in CtIP promoter and enhancer in ENCODE. Primers for ChIP-qPCR validation are indicated (upper). ChIP-seq of anti-BRD4 at the RBBP8 locus in HCC1395 or T47D cells treated with JQ1 in GSE63581 dataset (lower). (C) HOC1 treated with vehicle or 200 nM JQ1 for 24 hr and subjected to ChIP with normal rabbit IgG, BRD4, H3K27Ac, H3K4Me1 or Pol-II antibody as indicated. ChIP samples were analyzed by qPCR using primers indicated in B. Data across panels represent mean±SEM of three independent experiments. Student’s t test: * p

Figure 4. Effect of down regulation of CtIP on DNA end resection, generation of ssDNA, and HR function

(A) Representative images of BrdU and γH2AX staining under non-denaturing conditions at 4 hr after 10 Gy IR in HOC1 cells cultured with or without 200 nM JQ1 (see STAR methods). BrdU positive cells were quantified below. Scale bar, 20 m. (B) Representative images of RPA foci in HOC1 cells after 24 hr BRD4 inhibition (200 nM JQ1 or siRNA) or CtIP downregulation (siRNA) and then treated with BMN673 (200 nM) for 48 hr. RPA foci positive cells were quantified below. Scale bar, 20 μm. (C) Western blotting of indicated proteins in HOC1 cells 24 hr after transfection with control, CtIP or BRD4 siRNA and then treated with 200 nM BMN673 for 48 hr. (D) Western blotting of indicated proteins in HOC1 cells treated with BMN673 (200 nM), JQ1 (200 nM), GSK1324726A (200 nM) or the indicated combination for 48 hr. (E) Western blot of indicated proteins in chromatin-bound fractions from HOC1 cell treated with BMN673 (200 nM), JQ1 (200 nM), or combination for 48 hr. Histone H3 was used as marker for chromatin-bound fraction. (F) Representative images of RAD51 and γH2AX foci in HOC1 cells after 24 hr BRD4 inhibition (200 nM JQ1 or siRNA) or CtIP downregulation (siRNA) and then treated with BMN673 (200 nM) for 48 hr. Scale bar, 5 μm. (G) Comet assay in HOC1 cells treated with BMN673 (200 nM), JQ1 (200 nM), or combination for 48 hr. DNA damage quantified via % DNA in tails. Each data point represents at least 50 cells counted. Scale bar, 10 μm. (H) Comet assay in HOC1 cells 24 hr after transfection with control, CtIP, or BRD4 siRNA and then treated with 200 nM BMN673 for 48 hr. Each data point represents at least 50 cells counted. Scale bar, 10 μm. (I) 24 hours after transfection with control or CtIP siRNA, U2OS DR-GFP cells were transfected with the I-Sce1 endonuclease for 48 hours. HR efficiency of CtIP siRNA treated cells was compared to control siRNA based on percentage of GFP+ cells detected by flow cytometry. (J) 24 hr after transfection with control, BRD4, or CtIP siRNA, clonogenic assay was performed with indicated dose of BMN673 for 7 days. Representative pictures are shown. (K) Cells were transfected with CtIP siRNA (50 nM) or treated with 200 nM JQ1 with or without 50 nM RAD51 siRNA for 24 hr. Western blots of indicated proteins are in left panel. Cells were then treated for 96 hr with indicated doses of BMN673 and viability assessed (right). Short: short time exposure, long: long time exposure. Data across panels represents mean±SEM of three independent experiments, Student’s t test: **p

Figure 5. Effect of CtIP expression on BRD4 inhibition induced DNA end resection and HR defects

(A) Representative images (upper) and quantification (lower) of native BrdU foci staining in Dox inducible GFP-CtIP or GFP-CtIP (T847A) HOC1 cells at 4 hr after 10 Gy IR plus 200 nM JQ1 treatment with or without Dox induction. Scale bar, 20 μm. (B) Representative images (left) and quantification of positive cells (right) of RPA (upper) and RAD51 foci (lower) in Dox inducible GFP-CtIP or GFP-CtIP (T847A) HOC1 cells treated with combination of 200 nM BMN673 and 200 nM JQ1 for 48 hr with or without Dox induction. Scale bar, 20 μm. (C) Western blotting of indicated proteins in Dox inducible GFP-CtIP and GFP-CtIP (T847A) HOC1 cells treated with BMN673 (200 nM), JQ1 (200 nM), or combination for 48 hr with or without Dox induction. (D) Representative pictures of clonogenic assay in Dox inducible GFP-CtIP and GFP-CtIP (T847A) cells treated with BMN673 (200 nM), JQ1 (200 nM), or combination for 10 days with or without Dox induction. (E) Relative colony formation rates of cell in D are presented as percent relative to DMSO. Data across studies represent mean±SEM of three independent experiments, Student’s t-test: *p

Figure 6. Effect of PARPi and BRD4i on survival of different cell lineages

(A) Dose response curves of BMN673 or JQ1 alone or combined in 55 cancer cell lines treated with varying concentrations of the JQ1 and BMN673 for 96 hr. Combination index (CI) was calculated using CalcuSyn software with the Chou-Talalay equation. (B) BMN673 IC50 of (top) and selected mutations (middle) in cell lines. Red indicates a mutation in the respective gene, white indicates no mutation; red text indicates significant differences in frequency of mutations between PARPi sensitive and resistant cells (Pearson’s chi square test: p<0.05). The plot (bottom) shows the CtIP protein level in PARPi sensitive and resistant cells (Student’s t-test. p<0.001). (C) CI values of (top) and selected mutations (middle) in cell lines. Red indicates a mutation within the respective gene, white indicates no mutation; red text indicates significant differences in frequency of mutation between cells with or without synergism between BRD4i and PARPi (Pearson’s chi square test: p

Figure 7. Effect of BRD4i on acquired PARPi resistance

(A) Dose response curves of parental or PARPi-resistant OAW42 and A2780CP cells treated with BMN673 or JQ1 alone and combined for 96 hr. (B) Dose response curves of parental or six individual monoclonal populations of PARPi-resistant OC316 treated with BMN673 (upper left) for 96 hr. Remaining graphs show dose response curves of six individual monoclonal populations of PARPi-resistant OC316 treated with various concentrations of BMN673 alone (blue) or combined with 200 nM JQ1 (red) for 96 hr. (C) Dose response curves of UWB1.289 and UWB1.289-BRCA1 treated with BMN673 or JQ1 alone or combined for 96 hr. (D) Western blot of BRCA1/53BP1 knockdown efficiency in MCF10A stable lines or of 53BP1 knockdown efficiency in UWB1.289 and COV326 cell lines by siRNA for 48 hr. (E) Dose response curves of BMN673 in parental, shBRCA1, or shBRCA1/53BP1 stable MCF10A cells with or without 200 nM JQ1 for 96 hr. (F) Dose response curves of BMN673 in COV326 cells transfected with control or 53BP1 siRNA with or without 200 nM JQ1 for 96 hr. (G) Representative images of clonogenic assay in parental, shBRCA1, or shBRCA1/53BP1 stable MCF10A cells in presence of the indicated inhibitor for 10 days. (H) 24 hr after transfection with control or 53BP1 siRNA in UWB1.289 or COV326 cells, clonogenic assays were performed with indicated dose of BMN673 for 10 days. Representative pictures are shown. (I) Western blot of PARP1 in A2780CP cells after PARP1 silencing by siRNA for 48 hr (left). Dose response curves in control or PARP1 knockdown cells treated with BMN673 or JQ1 alone or combined for 96 hr (right). (J) Western blot of PARP1 in parental or shPARP1 stable MDA-MB-231 cells (left). Dose response curves in parental or shPARP1 stable MDA-MB-231 cells treated with BMN673 or JQ1 alone or combined for 96 hr (right). Short: short time exposure, long: long time exposure. Data represent mean±SEM of three independent experiments. See also Figure S7.

Figure 8. Efficacy of BRD4i and PARPi in vivo

(A, B) Tumor volume curves (upper) or waterfall plot of tumor burden changes (lower) of OVCAR8 xenografts (A) and WU-BC3 PDX (B) mice treated with vehicle (0.5% hydroxypropylmethylcellulose and 0.2% Tween 80), BMN673 (0.333 mg/kg, oral gavage, per day), JQ1 (40 mg/kg, I.P., per day), or combination of BMN673 and JQ1. (C–E) Tumor volume curves (upper) or waterfall plot of tumor burden changes (lower) of OVCAR3 (C) or PATX53 (D) xenografts or LPA1-T127 allograft (E) mice treated with vehicle (0.5% hydroxypropylmethylcellulose and 0.2% Tween 80), Olaparib (100 mg/kg, oral gavage, per day), AZD5153 (2.5 mg/kg, oral gavage, per day), or combination of Olaparib and AZD5153. (F, G) Representative images of IHC with indicated antibodies in tumor tissues from OVCAR8 xenografts (F) or WU-BC3 PDX (G). Scale bar, 50 μm. Data represent mean±SEM. Analysis of variance (ANOVA) was used to compare differences among multiple groups. **p