Androgen receptor signaling regulates DNA repair in prostate cancers

Abstract

We demonstrate that the androgen receptor (AR) regulates a transcriptional program of DNA repair genes that promotes prostate cancer radioresistance, providing a potential mechanism by which androgen deprivation therapy synergizes with ionizing radiation. Using a model of castration-resistant prostate cancer, we show that second-generation antiandrogen therapy results in downregulation of DNA repair genes. Next, we demonstrate that primary prostate cancers display a significant spectrum of AR transcriptional output, which correlates with expression of a set of DNA repair genes. Using RNA-seq and ChIP-seq, we define which of these DNA repair genes are both induced by androgen and represent direct AR targets. We establish that prostate cancer cells treated with ionizing radiation plus androgen demonstrate enhanced DNA repair and decreased DNA damage and furthermore that antiandrogen treatment causes increased DNA damage and decreased clonogenic survival. Finally, we demonstrate that antiandrogen treatment results in decreased classical nonhomologous end-joining.

Significance: We demonstrate that the AR regulates a network of DNA repair genes, providing a potential mechanism by which androgen deprivation synergizes with radiotherapy for prostate cancer.

Figure 1. AR transcriptional output is associated with expression of DNA repair genes

A. Gene set enrichment analysis of second-generation anti-androgen (ARN-509) treatment of a castration-resistant xenograft model (LNCaP-AR). Xenografts were treated with 4 days of either control (red) or ARN-509 (blue). B. Heatmap demonstrating wide spectrum of AR transcriptional output (“canonical AR output”), calculated using Hieronymus et al. (18), across cohort of 131 primary prostate cancer tumors. C. Correlation between enriched DNA repair genes from the xenograft experiment (Fig. 1A) and canonical AR output in primary prostate cancer cohort. D. Union of top six enriched DNA repair gene sets from xenograft experiment filtered for an association with canonical AR output (p<0.01 and r>0) in same human cohort. Primary prostate tumors ranked from low to high canonical AR output, left to right, with corresponding heatmap of associated 144 DNA repair genes (“AR-associated DNA repair” signature).

Figure 2. A network of DNA repair genes are both induced by androgen and represent AR target genes

A. Gene set enrichment analysis of LNCaP cell line grown in charcoal-stripped serum (CSS) in the presence of exogenous androgen (R1881), red, or control, blue. Previously defined AR-associated DNA repair signature enriched for in addition to other well-established AR signatures, such as Nelson et al (20). B. Of the 144 gene AR-associated DNA repair signature, 72 genes are significantly induced by R1881 and of these genes 32 represent AR target genes by ChiP-seq. C. Examples of AR peaks on representative DNA repair genes. D. Motif analysis of the AR binding peaks of these 32 genes revealed the classic consensus AR-binding site.

Figure 3. Enhanced DNA repair in prostate cancer cells treated with androgen plus IR, decreased DNA repair and survival of cells treated with anti-androgen plus IR

A. LNCaP cells, pretreated with either synthetic androgen (R1881) or mock, exposed to 2 Gy of IR, with DNA damage measured by gamma-H2AX foci. Under androgen-depleted conditions the foci peaked later and higher and resolved more slowly (* = p<0.001). B. Neutral Comet assay of LNCaP cell line, showing increased double-strand breaks when cells irradiated under androgen-deprived conditions (* = p<0.001). C. Neutral Comet assay of LNCaP, LNCaP-AR, and VCaP cells after 48 hours of treatment with second-generation anti-androgen (ARN-509) versus mock demonstrates increased double-strand breaks (* = p<0.001). D. Clonogenic survival assay of LNCaP cells demonstrates decreased survival when irradiated in the presence of ARN-509 versus mock (* = p≤0.02).

Figure 4. AR regulates a network of DNA repair genes

A. The AR-regulated transcriptome includes a number of DNA repair genes, one of which is the known AR-cofactor, PARP1. Other AR-regulated DNA repair genes play important roles in DNA damage sensing, non-homologous end joining (NHEJ), homologous recombination (HR), mismatch repair (MMR), and the Fanconi pathway. B. Schematic outlining the transient V(D)J recombination assay in which substrate along with RAG1 and RAG2 expression vectors are transfected into cells, after which bacteria are transformed with the rescued substrate and plated. Rescued substrate plasmids that fail to undergo recombination by classical NHEJ (C-NHEJ) only express the ampicillin (Amp) resistant gene, while those that successfully undergo recombination express both Amp and chloramphenicol (Cam) resistant genes. Recombination frequency can be calculated by the ratio of colonies grown on the Amp + Cam plates compared to the Amp plates. C. Compared with control-treated LNCaP cells, treatment with ARN-509 significantly (p<0.01) decreases C-NHEJ as measured using transient V(D)J recombination assay. The magnitude of effect is comparable to that observed when Lig4, a known mediator of C-NHEJ, has been stably knocked down in a LNCaP cell line (p<0.01). D. Schematic outlining transient DR-GFP assay, in which DR-GFP contains direct repeats of two defective GFP genes with the upstream repeat containing a recognition site for the I-SceI endonuclease. When a double-strand break (DSB) is introduced by I-SceI followed by successful homologous recombination (HR) repair using the downstream copy as a template, it produces a functional GFP gene; GFP+ cells are scored by flow cytometry. E. Compared with control-treated LNCaP cells, treatment with ARN-509 has no measurable effect upon HR repair as measured by the transient DR-GFP assay (p=0.38; n=4 transfections), while expression of the dominant negative BRC3 peptide (n=3) results in an approximately 2.8 fold reduction in HR compared to vector control (n=2) (p=0.015).