BRD4 Promotes DNA Repair and Mediates the Formation of TMPRSS2-ERG Gene Rearrangements in Prostate Cancer

Abstract

BRD4 belongs to the bromodomain and extraterminal (BET) family of chromatin reader proteins that bind acetylated histones and regulate gene expression. Pharmacological inhibition of BRD4 by BET inhibitors (BETi) has indicated antitumor activity against multiple cancer types. We show that BRD4 is essential for the repair of DNA double-strand breaks (DSBs) and mediates the formation of oncogenic gene rearrangements by engaging the non-homologous end joining (NHEJ) pathway. Mechanistically, genome-wide DNA breaks are associated with enhanced acetylation of histone H4, leading to BRD4 recruitment, and stable establishment of the DNA repair complex. In support of this, we also show that, in clinical tumor samples, BRD4 protein levels are negatively associated with outcome after prostate cancer (PCa) radiation therapy. Thus, in addition to regulating gene expression, BRD4 is also a central player in the repair of DNA DSBs, with significant implications for cancer therapy.

Keywords: BRD2; BRD4; CRPC; DNA repair; NHEJ; TMPRSS2-ERG; gene fusion; genomic rearrangements; non-homologous end joining; prostate cancer.

Figure 1. Regulation of DNA Repair Genes and NHEJ DNA Repair by BET Inhibitors

(A) Histogram representation of beta values indicating gene expression (from RNA-seq experiment) upon treatment of LNCaP cells with I-BET151 for 8 hr. Red colored bars represent 10 NHEJ genes. (B) Heatmap representation of the expression of DNA repair genes (from RNA-seq experiment) upon treatment of LNCaP cells with the indicated doses of I-BET151. (C) Gene set enrichment analysis (GSEA) of BRD4 expression level from castration-resistant PCa (CRPC) specimens (n = 122) against the 10 NHEJ DNA repair genes. (D) Schematic of the NHEJ DNA repair assay. (E) Representative flow cytometry profiles to describe the effects JQ1 on the repair of I-SceI-induced DNA DSBs by the NHEJ pathway. (F and G) Quantitative analysis of the effects of JQ1 (F) or siRNA against BRD4 (G) on the repair of I-SceI-induced DNA DSBs by the NHEJ pathway (*p < 0.05; **p < 0.01; ***p < 0.001 by two-tailed Student's t test; error bars, SD of 3 technicalreplicates). (H) BRD4 RNA expression in normal prostate, ERG fusion positive and SPOP mutant primary prostate adenocarcinoma samples from TCGA dataset.

Figure 2. BRD4 Promotes TMPRSS2-ERG Genomic Rearrangements

(A) Schematic representation of the CRISPR-Cas9 assay to engineer TMPRSS2-ERG genomic rearrangements. The single-guide RNA (sgRNA) target sites in the introns of TMPRSS2 and ERG genes are indicated by red arrows. (B) A TaqMan qRT-PCR assay to detect TMPRSS2-ERG fusion RNA transcripts. T1, T2, and T3 represent sgRNAs that target the TMPRSS2 locus; E1, E2, and E3 represent sgRNAs that target the ERG locus. (C) A TaqMan quantitative qPCR assay to detect the specific TMPRSS2-ERG fusion genomic DNA junction induced by combination treatment with T3 and E2 sgRNAs. (D) Sequence analysis of TMPRSS2-ERG fusion genomic DNA junction in LNCaP cells obtained by combination treatment with T3 and E2 sgRNAs. Gene structures for TMPRSS2 and ERG are shown using the GenBank reference sequences NM_005656 and NM_004449, respectively. (E) Sequence analysis of TMPRSS2-ERG fusion transcript junction in LNCaP cells obtained by combination treatment with T3 and E2 sgRNAs. (F and G) TaqMan qPCR assay to detect the specific TMPRSS2-ERG fusion genomic DNA junction (F) or TaqMan-qRT-PCR assay for RNA transcript junction (G) in LNCaP cells treated with siRNA against BRD4, in combination with sgRNAs targeting TMPRSS2 and ERG genes. siBRD4 (pool) represents combination of the four-individual siRNAs against BRD4 (*p < 0.001 by two-tailed Student's t test; error bars, SD of 3 technical replicates). (H and I) TaqMan qPCR assay to detect the specific TMPRSS2-ERG fusion genomic DNA junction (H) or TaqMan-qRT-PCR assay for RNA transcript junction (I) in LNCaP cells treated with various doses of JQ1, in combination with sgRNAs targeting TMPRSS2 and ERG genes. (J) Quantitative analysis of the effects of siRNA against BRD2 on the repair of I-SceI-induced DNA DSBs by the NHEJ pathway (***p < 0.001 by two-tailed Student's t test; error bars, SD of 3 technical replicates). (K and L) TaqMan qPCR assay to detect the specific TMPRSS2-ERG fusion genomic DNA junction (K) or TaqMan-qRT-PCR assay for RNA transcript junction (L) in LNCaP cells treated with siRNA against BRD2, in combination with sgRNAs targeting TMPRSS2 and ERG genes. siBRD2 (pool) represents combination of the four individual siRNAs against BRD2 (*p < 0.001 by two-tailed Student's t test; error bars, SD of 3 technical replicates).

Figure 3. Ionizing Radiation Induces Acetylation of Histone H4 in the Chromatin

(A) Heatmap representation of ChIP-seq signals ±4 kb around Acetyl histone H4 peaks in LNCaP cells treated with or without 5 Gy IR. Samples were processed 4 hr post-treatment. The heatmaps are paired and sorted by the 0-Gy treatment. (B) Venn diagram representing acetyl histone H4 peaks in untreated or 5-Gy IR-treated LNCaP cells. (C–E) Average coverage plots showing enrichment of Acetyl histone H4 genome-wide (C), at transcription start sites (TSS) (D) and DNase I hypersensitivity sites (DHSs) (E). (F) Genome browser representation of Acetyl histone H4 peaks in TMPRSS2 and PTEN genes. Black triangles represent common peaks and red triangles represent IR-induced peaks.

Figure 4. BRD4 Is Recruited to the Chromatin upon DNA Damage and Functionally Interacts with DNA Repair Proteins

(A) Histone H4 acetylation and BRD4 recruitment to the chromatin upon ionizing radiation (IR)-induced DNA damage (20 Gy) in LNCaP cells (top). γ-H2AX is the positive control for IR treatment; H2AX serves as positive control for the chromatin fraction and negative control for cytosolic fraction; β-tubulin serves as positive control for cytosolic fraction and negative control for chromatin fraction. BRD4 was immunoprecipitated from the same lysates and analyzed by immunoblot using anti-acetyl histone H4 antibody (bottom). (B) The role of JQ1 in recruitment of BRD4 to the chromatin upon IR-induced DNA damage (20 Gy) in LNCaP cells. (C) Co-immunoprecipitation (Co-IP) experiments with nuclear extracts from untreated or IR-treated LNCaP cells (20 Gy). 1 hr post-IR treatment, the immunoprecipitation (IP) was performed using anti-BRD4 antibody and analyzed by immunoblot with the indicated antibodies. (D) The role of JQ1 (10 μM) in the recruitment of BRD4, 53BP1, Artemis, and Ku80 to the chromatin upon IR-induced DNA damage (20 Gy) in LNCaP cells. (E and F) The role of dBET1 (1 μM) in the recruitment of BRD4 and DNA repair proteins to the chromatin upon IR-induced DNA damage in LNCaP (E) and 22Rv1 cells (F).

Figure 5. Loss of BRD4 Function Promotes H2AX Phosphorylation after IR Treatment

(A–D) The effect of BRD4 knockdown (A and B) or incubation with 1 μM JQ1 (C and D) in the phos-phorylation of histone H2A.X (Ser139) upon treatment of LNCaP cells with IR (5 Gy). The cells were analyzed at 30 and 120 min post-IR treatment; scale bar, 10 μm. The number of γ-H2AX foci, above threshold intensity per nucleus (n = 165) was quantified using the ImageJ software (***p < 0.0001 by two-tailed Mann-Whitney U test).

Figure 6. BRD4 Inhibition Synergizes with IR to Promote DNA Damage

(A) Alkaline comet assay of LNCaP cells with non-targeting siRNA orBRD4 siRNA, followed by treatment with the indicated doses of IR. Cells were irradiated 72-hr post-siRNA treatment, followed by recovery after 30 min. (B) Quantification of alkaline comet assay Olive tail moment (top) and validation of BRD4 knockdown by immunoblotting (bottom). (C) Alkaline comet assay of LNCaP cells treated with 1 μM JQ1 for 24 hr, followed by IR treatment, and recovery after 30 min. (D) Quantification of alkaline comet assay Olive tail moment. (E) Alkaline comet assay quantification of single-agent or combination treatment of LNCaP (left) and 22Rv1 cells (right) with 1 μM JQ1 and/or 1 μM enzalutamide for 24 hr followed by IR treatment and recovery after 30 min (**p < 0.01; ***p < 0.001; ****p < 0.0001 by unpaired Student's t test; error bars, SD of n > 50 cells fromeach sample).

Figure 7. Nuclear BRD4 Expression Correlates with Progression to Castration Resistance PCa

(A) Representative IHC images and HS for BRD4 expression in diagnostic biopsies of prostate adenocarcinoma. (B) Median HS and interquartile range for nuclear BRD4 expression in normal (39 patients), prostatic intraepithelial neoplasia (PIN; 37 patients), and adenocarcinoma (28 patients). Patients with adenocarcinoma were divided by BRD4 low (HS <100; 11 patients; gray) and BRD4 high (HS ≥100; 17 patients; red) for further analysis. BRD4 expression between groups was not significantly different using Kruskal-Wallis equality-of-populations rank test (p = 0.92). (C and D) Kaplan-Meier curves of time to CRPC (C) (y axis represents percent hormone-sensitive PCa [HSPC]) and overall survival (D) from diagnosis after radical primary therapy are shown for low BRD4 (gray) and high BRD4 (red) groups. Hazard ratios (HRs) with 95% confidence intervals and p values for univariate cox survival model are shown.