FOXA1 regulates androgen receptor variant activity in models of castrate-resistant prostate cancer

Abstract

Retention of androgen receptor (AR) signalling in castrate-resistant prostate cancer (CRPC) highlights the requirement for the development of more effective AR targeting therapies. A key mechanism of resistance to anti-androgens is through expression of constitutively active AR variants (AR-Vs) that are refractory to next-generation therapies, including Enzalutamide and Abiraterone. By maintaining an androgenic gene signature, AR-Vs drive tumour survival and progression in castrate conditions. Critically, however, our understanding of the mechanics of AR-V-driven transcription is limited, particularly with respect to dependency on pioneer factor function. Here we show that depletion of FOXA1 in the CWR22Rv1 CRPC cell line abrogates the oncogenic potential of AR-Vs. Gene expression profiling reveals that approximately 41% of the AR-V transcriptome requires FOXA1 and that depletion of FOXA1 attenuates AR-V binding at a sub-set of analysed co-regulated genes. Interestingly, AR-V levels are elevated in cells depleted of FOXA1 as a consequence of attenuated negative feedback on the AR gene, but is insufficient to maintain cell growth as evidenced by marked anti-proliferative effects in FOXA1 knockdown cells. In all, our data suggests that AR-Vs are dependent on FOXA1 for sustaining a pro-proliferative gene signature and agents targeting FOXA1 may represent novel therapeutic options for CRPC patients.

Keywords: FOXA1; androgen receptor variants; prostate cancer; transcriptional regulation.

Figure 1. AR-Vs are constitutively chromatin bound

Chromatin immunoprecipitation (ChIP) was performed using AR (C-19) A. or AR (N-20) B. antibodies in VCaP and CWR22Rv1 cells treated with 10 nM DHT or 1 μM Enzalutamide for 4 hours and AR recruitment to the PSA enhancer element was analysed by quantitative PCR. C. Chromatin fractionation was performed in CWR22Rv1 cells treated as above and resultant samples were subject to anti-AR, PARP1 (chromatin) and α-tubulin (cytoplasmic) antibodies. Formaldehyde-assisted isolation of response elements (FAIRE) was performed in LNCaP D. and CWR22Rv1 E. cells after 4 hours 10 nM DHT treatment and resultant DNA was analysed by quantitative PCR using primers specific to the indicated genes. All data represents the mean of at least three independent experiments +/− SE (*denotes p-value < 0.05)

Figure 2. FOXA1 depletion up-regulates PSA and KLK2 expression in CWR22Rv1 cells

A. FOXA1 levels were reduced by siRNA knockdown for 48 hours in CWR22Rv1 cells and expression of PSA and KLK2 was measured after 24 hours 10 nM DHT stimulation by quantitative PCR. Representative FOXA1 protein levels are shown on the right. B. As in (A), but with the inclusion of a 1 μM Enzalutamide treatment arm for 24 hours prior to quantitative analysis. C. AR and FOXA1 knockdown is confirmed by immunoblotting using anti-AR and –FOXA1 antibodies. All data represents the mean of at least three independent experiments +/− SE (*denotes p-value < 0.05)

Figure 3. FOXA1 knockdown elevates AR-V expression in CWR22Rv1 cells

A. FOXA1 was depleted by siRNA for 48 hours in CWR22Rv1 cells and levels of AR-V7, AR-1/2/3/2b were measured after 24 hours 10 nM DHT stimulation by quantitative PCR. B. As in (A), but cells were additionally depleted of AR by siRNA and treated for 24 hours with Enzalutamide prior to analysis of FL-AR mRNA. C. Western analysis of FOXA1 knockdown samples using AR (N-20), AR-V7, FOXA1 and α-Tubulin antibodies. D. As in (A) but with the inclusion of a 1 μM Enzalutamide treatment arm. All data represents the mean of at least three independent experiments +/− SE (*denotes p-value < 0.05)

Figure 4. A downstream repressive element in the AR gene controls AR-V expression

Chromatin immunoprecipitation (ChIP) experiments in CWR22Rv1 cells using AR (N-20) A. and FOXA1 B. antibodies to assess recruitment at intron 2 of the AR gene in response to 4 hour 10 nM DHT treatment. CWR22Rv1 cells depleted of FOXA1 were treated with Enzalutamide for 4 hours and subject to ChIP using FOXA1 C. and AR D. antibodies to assess recruitment to intron 2 of the AR gene. Fold enrichment is calculated in (D) to demonstrate loss of AR recruitment upon FOXA1 depletion (*denotes p = < 0.05 significance between siScr control and FOXA1 depletion). All data represents the mean of at least three independent experiments +/− SE (*denotes p-value < 0.05)

Figure 5. AR-V- and FOXA1-target genes show considerable overlap

A. Venn diagrams of genes demonstrating 1.5-fold up- and down-regulation in response to AR and FOXA1 depletion in CWR22Rv1 cells grown in steroid-depleted conditions + 1 μM Enzalutamide. B. Validation of UBE2C, NKx3.1 and ATAD2 as AR-V- and FOXA1-co-regulated genes using quantitative PCR analysis of CWR22Rv1 cells depleted of AR or FOXA1 treated with and without Enzalutamide for 24 hours. C. Chromatin immunoprecipitation (ChIP) in CWR22Rv1 cells depleted of FOXA1 using an AR (N-20) antibody to assess AR-V binding to the UBE2C promoter (*denotes p = < 0.05 significance between siScr control and FOXA1 depletion). All data represents the mean of at least three independent experiments +/− SE (*denotes p-value < 0.05)

Figure 6. AR-Vs and FOXA1 co-regulate pro-proliferative genes

A. Gene ontology analysis of co-regulated AR- and FOXA1-target genes shows significant enrichment of cell cycle and cell division associated genes. B. Impact of FOXA1 and AR knockdown on Cyclin A2 (CCNA2) expression in CWR22Rv1 cells grown in steroid-depleted media treated with 10 nM DHT or 1 μM Enzalutamide as measured by quantitative PCR. C. Chromatin immunoprecipitation (ChIP) was performed in CWR22Rv1 cells depleted of FOXA1 to assess AR-V recruitment to the CCNA2 promoter (*denotes p = < 0.05 significance between siScr control and FOXA1 depletion). D. Proliferation assays to assess effect of AR or FOXA1 knockdown in CWR22Rv1 cells treated with 10 nM DHT or 1 μM Enzalutamide for 96 hours (* denotes p = < 0.05; ** denotes p = < 0.01 significance between siScr control and AR or FOXA1 knockdown, respectively). Representative western blots indicate AR and FOXA1 levels in response to protein knockdown. E. Cell cycle analysis using propidium iodide (PI) flow cytometry in CWR22Rv1 cells depleted of FOXA1 grown in the presence and absence of 1 μM Enzalutamide for 72 hours. All data represents the mean of at least three independent experiments +/− SE (*denotes p–value < 0.05). F. Diagrammatic representation of FOXA1-AR-V interplay in CRPC. Expression of AR-Vs are regulated by FOXA1 at the downstream repressive element (DRE) of the AR gene. FOXA1 also facilitates respective androgen-independent activation and repression of genes, such as CCNA2 and SRSF2 by AR-Vs.