AR coactivators, CBP/p300, are critical mediators of DNA repair in prostate cancer

Abstract

Castration resistant prostate cancer (CRPC) remains an incurable disease stage with ineffective treatments options. Here, the androgen receptor (AR) coactivators CBP/p300, which are histone acetyltransferases, were identified as critical mediators of DNA damage repair (DDR) to potentially enhance therapeutic targeting of CRPC. Key findings demonstrate that CBP/p300 expression increases with disease progression and selects for poor prognosis in metastatic disease. CBP/p300 bromodomain inhibition enhances response to standard of care therapeutics. Functional studies, CBP/p300 cistrome mapping, and transcriptome in CRPC revealed that CBP/p300 regulates DDR. Further mechanistic investigation showed that CBP/p300 attenuation via therapeutic targeting and genomic knockdown decreases homologous recombination (HR) factors in vitro, in vivo, and in human prostate cancer (PCa) tumors ex vivo. Similarly, CBP/p300 expression in human prostate tissue correlates with HR factors. Lastly, targeting CBP/p300 impacts HR-mediate repair and patient outcome. Collectively, these studies identify CBP/p300 as drivers of PCa tumorigenesis and lay the groundwork to optimize therapeutic strategies for advanced PCa via CBP/p300 inhibition, potentially in combination with AR-directed and DDR therapies.

Fig. 1. Altered CBP/p300 expression correlates with aggressive disease.

A Frequency of AR, CBP, and p300 gene alterations (i.e., amplifications, mutations, and/or deletions) in primary and metastatic PCa datasets from cBioPortal—TCGA Firehose Legacy and SU2C/PCF Dream Team (PNAS 2019) datasets. Hormone sensitive PCa total alterations AR n = 9, CBP n = 12, and p300 n = 7, and castration-resistant PCa AR n = 252, CBP n = 36, and p300 n = 8. B Correlation of CBP and p300 mRNA expression in metastatic PCa patient tumors from SU2C/PCF Dream Team (PNAS 2019) data set from cBioportal. C CBP and p300 are highly expressed in primary PCa (top 10% expressed genes) and mCRPC (top 25% expressed genes) and are associated with poor outcome. D cBioportal analyses for disease-free survival from prostate adenocarcinoma data set (SKCC, Cancer Cell 2010).

Fig. 2. CCS1477 sensitizes CRPC cells to AR-dependent DNA repair.

A, B 22Rv1 and C4-2 cells treated with CBP/p300 inhibitor (CCS1477—1 mM) for increasing time (4–72 h). A Cells were harvested and analyzed for protein expression of AR, AR-V7, c-MYC, CBP, p300, and Vinculin. B Cells were harvested and analyzed for cell cycle analyses with flow cytometry. C, D 22Rv1 (top) and C4-2 (bottom) cells were treated with increasing dosages of indicated DDR agents (IR, Cisplatin, Olaparib, or Doxorubicin) with or without CCS1477, and drug sensitivity assays were performed on Day 5 using Picogreen. D Non-linear regression analyses were performed to determine the IC₅₀ values. E WB analyses of AR, CBP, p300, c-MYC, and Vinculin in response to IR, Cisplatin, Olaparib, and Doxorubicin in 22Rv1 and C4-2 cells. n = 3, *p < 0.05, **p < 0.01, and ***p < 0.001.

Fig. 3. CBP/p300 inhibition drastically reduces the AR, CBP, p300 cistromic landscape.

A ChIP-Seq of intensity plots of 22Rv1 control and CCS1477 treated cells. AR, CBP, and p300 were immunoprecipitated in CCS1477 exclusive, common binding, and VEH exclusive in a 6-kb region. B Percent of total binding of overlap of common bindings sites of AR, CBP, and p300 in CCS treated 22Rv1 cells. C Known motif of AR, CBP, and p300 of common binding in 22Rv1 CCS1477 treated cells. D Motifs of binding regions of AR, CBP, and p300 visualized for top 6 highest p-values. E Hallmarks GSEA pathways regulated by CBP and p300 in CCS1477 treated 22Rv1 cells.

Fig. 4. CBP/p300 acts in concert to alter the global transcriptome in PCa cells and regulate DDR.

A 22Rv1- shCON, shCBP, and shP300 cells were treated with doxycycline to induce knockdown, and RNA Seq analysis was performed in quadruplet samples. MA plot depicts gene expression modulation with the number of significant upregulated and downregulated transcripts with p < 0.001 and fold change (FC) > 2.0. B GSEA of RNA-Seq (KEGG Pathways) identified deenriched pathways with CBP and p300 knockdown treatment in 22Rv1 cells using FDR < 0.25. C–E. 22Rv1-shCON, -shCBP, and –shp300 cells were treated with doxycycline and 5 Gy IR. C Flow analysis was performed to determine changes in cell cycle progression. Cells were harvested, and mRNA (D) was isolated. Changes in c-MYC, TP53, and CDKN1A (p21) and 18S mRNA expression. E 22Rv1-shCON, -shCBP, and –shp300 cells were treated with doxycycline to knockdown CBP and p300. Cells were also treated with IR (5 Gy) at Day 1. Cells were harvested for growth assays analyses on Days 1, 3, and 5 using Picogreen. F 22Rv1 (top) and C4-2 (bottom) cells were treated with increasing dosages of indicated DDR agents (Doxorubicin, Cisplatin, or Olaparib), and drug sensitivity assays were performed on Day 5 using Picogreen. G Non-linear regression analyses were performed to determine the IC₅₀ values. n = 3, *p < 0.05, **p < 0.01, and ***p < 0.001.

Fig. 5. CBP/p300 attenuation impacts double-strand DNA break repair and decreases HR efficiency.

A Schematic describing the comparison of RNA-Seq data sets. Briefly, common transcriptions with p < 0.001 and fold change (FC) > 2.0 were identified and then organized into specific DDR pathways. B Validation of HR gene targets in CRPC cells with shCON, shCBP, and shp300 CRPC models (22Rv1 and C4-2). n = 3, *p < 0.05, **p < 0.01, and ***p < 0.001.

Fig. 6. CBP/p300 impacts double-strand DNA break repair and decreases HR efficiency.

A U20S-DR-GFP cells were treated with increasing doses of CCS1477. CBP and p300 was knocked down in U20S-DR-GFP cells for 72 h via siRNA. Cells were treated with ATM inhibitor for 24 h. Cells were harvested for flow cytometry. B–D. Changes in HR factors (ATM, CHEK2, and RAD50) and 18S mRNA and protein expression were analyzed with CBP/p300 attenuation (via shRNA or CCS1477 treatment) and 5 Gy IR for 24 h. B 22Rv1 and C4-2-shCON, -shCBP, and –shp300 cells were treated with doxycycline to knockdown CBP and p300. Then cells were treated with 5 Gy IR. Cells were harvested and mRNA was isolated. C Changes in ATM, CHECK2, RAD50, and 18S mRNA expression were analyzed in CBP/p300 inhibitor treated CRPC cells. D Changes in HR factors protein expression were analyzed with CBP/p300 inhibitor with or without genotoxic insult of 5 Gy IR or 10 nM Doxorubicin treatment for 24 h. n = 3, *p < 0.05, **p < 0.01, and ***p < 0.001.

Fig. 7. CBP/p300 inhibition mediates HR repair and impacts patient outcome.

A HR gene profiling in 22Rv1 and C4-2 cells with CBP/p300 inhibition (CCS1477) or CBP/p300 knockdown in inducible cell models (22Rv1). B Correlation of HR genes (ATM, ATR, and RAD50 with CBP and p300 mRNA expression in patient cohorts from SU2C/PCF Dream Team (PNAS 2019) data set from cBioportal. C HR expression in 22Rv1 xenografts with CCS1477 (20 mg/kg) and harvested to examine HR gene expression. D Patient-Derived Xenografts (PDX) prostate cancer tumor tissue treated with CCS1477 (20 mg/kg) for 8 days and then longer (see text for details). Harvested, RNA isolated, and analyzed for HR gene expression. E Patient-Derived Explants (PDE) prostate cancer tumor tissue treated with CCS1477 (1 mM and 5 mM) for 48 h and analyzed for Ki67 positivity. F HR expression in PDEs with CCS. G Module summarizing findings. n = 3, *p < 0.05, **p < 0.01, and ***p < 0.001. n > 3, *p < 0.05, **p < 0.01, and ***p < 0.001.