Epigenetic Reprogramming with Antisense Oligonucleotides Enhances the Effectiveness of Androgen Receptor Inhibition in Castration-Resistant Prostate Cancer Free

Prostate cancer remains the most commonly diagnosed cancer and the second leading cause of cancer-related deaths in American men, with most mortalities resulting from advanced metastatic tumors (1). Since its development and progression depend on signals through the androgen receptor (AR), physical and pharmacologic castration became therapeutic mainstays for advanced disease (2). Although initially responsive to androgen deprivation therapy (ADT), these tumors almost invariably recur in an incurable androgen-independent form, called castration-resistant prostate cancer (CRPC). The pathophysiologic underpinnings of CRPC are many, but frequently involve the continued activation of AR despite low levels of androgens (2). To address this phenomenon, contemporary antiandrogen therapy, typified by the second-generation antagonist enzalutamide, seeks to directly bind and target AR, blocking its intracellular activity. Although clinical outcomes are superior to conventional castration, enzalutamide-treated CRPC tumors still uniformly recur (3, 4).

The role of epigenetic modifications in CRPC has become an important area of inquiry. For example, we and others have shown that inhibition of BET bromodomain chromatin reader proteins attenuates growth of AR- wild-type- and AR-mutant CRPC (5, 6). Our group has also shown that the histone-lysine N-methyltransferase EZH2 is overexpressed and associated with poor prognosis in prostate cancer (7). Although EZH2 has been described as a transcriptional repressor that methylates histone H3 lysine 27 (H3K27Me3) to mediate epigenetic silencing of multiple tumor suppressors, studies also suggest that EZH2 possesses additional activities in prostate cancer (8). Recently, EZH2 has been found to be highly expressed in neuroendocrine prostate cancer (NEPC); progression to NEPC includes loss of lineage-specific AR-mediated gene expression and gain-of-function in EZH2 (9, 10). Another study in a NEPC model showed that EZH2 inhibition could reactivate AR signaling and sensitivity to drugs (11), whereas an additional study showed that NEPC may represent a late-stage with loss of plasticity, and AR expression/signaling was not amenable to EZH2 targeting (12). Whether inhibition of EZH2 can impact AR signaling during resistance to current AR-targeting therapies (e.g., enzalutamide) outside of NEPC has not been defined.

In this study, we identified EZH2 as a key regulator of sensitivity to AR-targeted therapies in CRPC adenocarcinoma models through an unbiased epigenetic inhibitor screen. EZH2 inhibition had profound effects on the AR cistrome and signaling, suggesting an enhanced dependence on this pathway. Furthermore, we tested antisense oligonucleotides (ASO) targeting EZH2 and AR as an alternative strategy to target oncogenic factors, which may be beneficial in settings where ablation of these proteins is desired compared with inhibition of select enzymatic activities or domains. ASOs have demonstrated clinical activity for the treatment of multiple diseases, including diabetes, hyperlipidemias, cardiovascular diseases, neurodegenerative diseases, and cancer; those currently approved for therapeutic use include Kynamro, Spinraza, and Exondys 51 (13). Herein, a novel ASO strategy targeting EZH2 and AR was shown to be effective in inhibiting CRPC growth in vivo, suggesting that targeting this axis may provide clinical benefit in patients that have developed resistance to established therapies.

LNCaP, C4-2B, and CWR-R1 prostate cancer cell lines were grown in RPMI1640 (Invitrogen) supplemented with 10% FBS(Invitrogen) in a 37°C incubator with 5% CO₂. The LNCaP-abl (14) and LNCaP-EnzR (15) cell lines were propagated as described previously. LNCaP cells were obtained from ATCC. C4-2B cells were generously provided by Evan Keller, Ph.D., at the University of Michigan (Ann Arbor, MI), CWR-R1 and LNCaP-EnzR cell lines were kindly provided by Donald Vander Griend, Ph.D., at the University of Chicago, (Chicago, IL) and the LNCaP-abl cell line was kindly provided by Myles Brown, Ph.D., at Harvard University (Cambridge, Massachusetts). Cells were genotyped to confirm identity at the University of Michigan Sequencing Core and tested routinely for Mycoplasma contamination.

C4-2B cells were seeded in 96-well plates at a density of 2 × 10³ cells per well in a total volume of 50-μL media. Each epigenetic inhibitor (see Supplementary Table S1) in the customized panel (Selleck) was added at 1 μmol/L concentration to an individual well containing either enzalutamide or DMSO vehicle (final concentration 2 μmol/L). Cells were incubated for 5 days before quantification with CellTiter-Glo reagent (Promega), as per the manufacturer's instructions. CellTiter-Glo analysis was performed after transferring 100 μL of CellTiter-Glo solution from each well into a Costar 96-well clear flat bottom plate for reading with a Tecan Infinite M1000 PRO reader. CellTiter-Glo luminescent signal was normalized to negative control wells (without epigenetic inhibitor); raw values are included in Supplementary Table S2.

Six-week-old male CB17 SCID mice were used for experiments. C4-2B prostate cancer cells (1 × 10⁶ cells) in 50% Matrigel (BD Biosciences) were injected subcutaneously into bilateral dorsal flanks. When xenografts reached an average size of 50 to 100 mm³, animals were randomized into groups (each containing 10 animals) subsequently treated with one of the following: PBS, vehicle, enzalutamide (30 mg/kg), ASO-Ctrl (50 mg/kg), ASO-AR (25 mg/kg), ASO-EZH2-65 (50 mg/kg), or ASO combo (25 mg/kg ASO-AR, 50 mg/kg ASO-EZH2-65). For experiment in Supplementary Data, treatment cohorts were as follows: ASO-Ctrl (75 mg/kg), ASO-AR (25 mg/kg) + ASO-Ctrl (50 mg/kg), ASO-EZH2-65 (50 mg/kg) + ASO-Ctrl (25 mg/kg), or ASO combo (25 mg/kg ASO-AR, 50 mg/kg ASO-EZH2-65). All reagents were dosed 5 days per week for 3 weeks. Enzalutamide was delivered by gavage and ASOs by subcutaneous injection. Tumor volume was calculated from digital caliper measurements made at the study outset, then twice weekly. Tumors were resected and weighed at the end of the time course. Animal experiments were approved by the University of Michigan Institution Animal Care and Use Committee.

Loewe additivity is a dose-effect model, which states that additivity occurs in a two-drug combination if the sum of the ratios of the dose versus the median effect for each individual drug is 1. In this model, combination index (CI) scores estimate the interaction between the two drugs. If CI < 1, the drugs have a synergistic effect and if CI > 1, the drugs have an antagonistic effect. CI = 1 means the drugs have additive effect. The CI coefficients were computed on the basis of the Chou–Talalay median-effect model as implemented in CalcuSynv2.11 (http://www.biosoft.com/w/calcusyn.htm).

Antibodies used in the immunoblotting assays are as follows: AR (Millipore, catalog no. 06-680), PSA (DAKO, catalog no. A0562), EZH2 (Cell Signaling Technology, catalog no. 5246S), H3K27Me3 (Diagenode, catalog no. C15410069), Histone H3 (Cell Signaling Technology, catalog no. 9715S), and GAPDH (Cell Signaling Technology, catalog no. 3683S). All antibodies were employed at dilutions suggested by the manufacturers. For Western blot analysis, 200 μg total protein extract was boiled in sample buffer and 10 to 20 μg aliquots were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membrane (GE Healthcare). The membrane was incubated for 1 hour in blocking buffer [Tris-buffered saline, 0.1% Tween (TBS-T), 5% nonfat dry milk] followed by incubation overnight at 4°C with primary antibody. Following a wash with TBS-T, the blot was incubated with horseradish peroxidase–conjugated secondary antibody, and signals were visualized by enhanced chemiluminescence system as per the manufacturer's protocol (GE Healthcare).

Total RNA was isolated from cells using Direct-zol Kit (Zymo), and cDNA was synthesized from 1,000 ng total RNA using Maxima First Strand cDNA Synthesis Kit for qRT-PCR (Thermo Fisher Scientific). qRT-PCR was performed in triplicate using standard SYBR Green reagents and protocols on a StepOnePlus RT-PCR system (Applied Biosystems). The target mRNA expression was quantified using the ΔΔC_t method and normalized to HMBS expression. All primers were designed using Primer 3 (http://frodo.wi.mit.edu/primer3/) and synthesized by Integrated DNA Technologies (Coralville). Primer sequences are listed in Supplementary Table S3. RNA-seq was performed using the Illumina HiSeq 2000 in paired-end mode, as described previously (16). Detailed description of gene set enrichment analysis (GSEA) is provided in the Supplementary Materials and Methods.

Chromatin immunoprecipitation (ChIP) assays for AR were performed using HighCell ChIP Kit (Diagenode) according to the manufacturer's protocol, using AR antibody (Millipore, catalog no. 06-680). For AR ChIP-seq experiments with EPZ-6438 and EZH2 ASOs, C4-2B cells were treated with 1 μmol/L EPZ-6438 or ASO-EZH2-65 for 5 days. Next, cells were cross-linked for 10 minutes with 1% formaldehyde. Cross-linking was terminated by the addition of 1/10 volume 1.25 mol/L glycine for 5 minutes at room temperature followed by cell lysis and sonication (Bioruptor, Diagenode). This resulted in an average chromatin fragment size of 200 base pairs. A chromatin equivalent of 5 × 10⁶ cells was used for each assay. ChIP DNA was isolated (IPure Kit, Diagenode) from samples by incubation with antibody at 4°C overnight followed by wash and reversal of cross-linking. The ChIP-seq sample preparation for sequencing was performed according to the manufacturer's instructions (Illumina). ChIP-enriched DNA samples (1–10 ng) were converted to blunt-ended fragments using T4 DNA polymerase, E. coli DNA polymerase I large fragment (Klenow polymerase), and T4 polynucleotide kinase (New England BioLabs). A single A-base was added to fragment ends by Klenow fragment (3′ to 5′ exo minus; New England BioLabs) followed by ligation of Illumina adapters (Quick ligase, New England BioLabs). The adapter-modified DNA fragments were enriched by PCR using the Illumina Barcode primers and Phusion DNA polymerase (New England BioLabs). PCR products were size selected using 3% NuSieve agarose gels (Lonza) followed by gel extraction using QIAEX II reagents (Qiagen). Libraries were quantified with the Bioanalyzer 2100 (Agilent) and sequenced on the Illumina HiSeq 2000 Sequencer (100 nucleotide read length). Detailed description of ChIP-seq data analysis is provided in the Supplementary Materials and Methods.

Assay for transposase-accessible chromatin using sequencing (ATAC-seq) was performed as described previously (17). Briefly, C4-2B cells were treated with 1 μmol/L EPZ-6438 or DMSO control for 5 days (or ASO-Ctrl and ASO-EZH2), then trypsinized and pelleted. A total of 2.5 × 10⁴ cells were washed in cold PBS and resuspended in lysis buffer (CER-I of NE-PER Kit, Invitrogen, catalog no. 78833) on ice for 10 minutes with occasional pipetting. The lysate was centrifuged at 1,300 × g for 5 minutes at 4°C. Nuclei were resuspended in 2× TD buffer, then incubated with Tn5 enzyme for 30 minutes at 37°C(Nextera DNA Library Preparation Kit, catalog no. FC-121-1031). Samples were immediately purified by Qiagen minElute column and PCR amplified with the NEBNext High-Fidelity 2X PCR Master Mix (New England BioLabs, catalog no. M0541L). qPCR was utilized to determine the optimal PCR cycles to prevent overamplification. The amplified library was further purified by Qiagen minElute column and SPRI beads (Beckman Coulter, catalog no. A63881). ATAC-seq libraries were sequenced on the Illumina HiSeq 2500, paired-end, 50 cycles. Detailed description of ATAC-seq data analysis is provided in the Supplementary Materials and Methods.

The RNAscope 2.5 HD BROWN assay (Advanced Cell Diagnostics, catalog no. 322300) was performed using target probes to EZH2 or AR according to the manufacturer's instructions. EZH2 RNA probes (catalog no. 405491, accession #NM_001203248.1, targeting 197–1337) and AR RNA probes (catalog no. 40049, accession # NM_000044.3, targeting 5604–6660) are complementary to the target mRNA. Probes Hs-PPIB (human peptidylprolyl isomerase B) and DapB (bacterial dihydrodipicolinate reductase) were used as positive and negative controls, respectively. Paraffin sections from formalin-fixed tumor samples were incubated for 1 hour in a 60°C drying oven, before deparaffinization in xylene, brief incubation in 100% ethanol, and air drying. Samples were permeabilized with Protease Plus, then subjected to probe hybridization with 2-hour incubation in HybEZ at 40°C. After washing, slides were processed by standard signal amplification steps. Chromogenic detection was performed using DAB, followed by counterstaining with 50% Gill hematoxylin I (Thermo Fisher Scientific, catalog no. 26801-01).

All slides were examined for EZH2 or AR ISH signals in morphologically intact cells and scored by a blinded independent pathologist. ISH signal was defined as brown, punctate dots, and expression level scored as described in ref. 18: 0, no staining or less than one dot per 10 cells; 1, one to three dots per cell; 2, four to nine dots per cell (few or no dot clusters); 3, 10 to 15 dots per cell (less than 10% in dot clusters); and 4, greater than 15 dots per cell (more than 10% in dot clusters). Cumulative ISH product score was calculated for each evaluable slide as the sum of the individual products of the expression level (0–4) and percentage of cells [0–100; i.e., (A% × 0) + (B% × 1) + (C% × 2) + (D% × 3) + (E% × 4); total range = 0 to 400].

The raw RNA-seq, ChIP-seq AR, and ATAC-seq data have been deposited at SRA(SRP157942).

Because past studies associated epigenetic changes with CRPC development, we postulated that inhibition of key epigenetic regulators may sensitize CRPC cells to antiandrogen therapy, such as enzalutamide (5–7). An unbiased epigenetic inhibitor screening experiment was conducted, subjecting CRPC-derived C4-2B cells to treatment with a library of 92 epigenetic regulator inhibitors (Supplementary Table S1; Fig. 1A). This screen found that EZH2 inhibitors markedly enhanced the inhibitory effect of enzalutamide on cell proliferation and ranked the highest among the compounds assessed (Fig. 1A). The experiment was repeated with the 10 most efficacious epigenetic inhibitors and found 6 targeted EZH2 inhibitors (Fig. 1B). Immunoblot analysis demonstrated that these six reduced H3K27Me3 protein content, indicating diminution of EZH2 function (Fig. 1B).

These findings were validated by employing EPZ-6438, a small-molecule EZH2 inhibitor that is effective in treating hematologic malignancies and solid tumors (19). CI analysis indicated that cotreatment with EPZ-6438 and enzalutamide produced synergistic antiproliferative effects in C4-2B and androgen-dependent LNCaP cells (Fig. 1C). Cotreatment yielded the slowest growth curve compared with treatment by DMSO or single agent and resulted in near-complete inhibition of cell proliferation (Fig. 1D). We further confirmed that EPZ-6438 was able to sensitize cells to enzalutamide using an enzalutamide-resistant LNCaP model, LNCap-EnzR (Supplementary Fig. S1; ref. 15). Selective disruption of EZH2 with CRISPR/Cas9 further established that loss of EZH2 enhanced the growth-inhibitory effect of enzalutamide in prostate cancer cells and in an additional model, CWR-R1 cells (Supplementary Fig. S2A–D; Supplementary Fig. S3A–C).

AR is central to prostate cancer development, progression, and drug resistance (2). Among the many AR signaling perturbations in CRPC, AR-mutant forms or ligand-independent variants (AR-Vs) are not effectively inhibited by currently approved therapies, such as enzalutamide (20, 21). Similarly, most inhibitors of EZH2, including EPZ-6438, target the SET domain of EZH2 responsible for methyltransferase activity. Studies have, however, shown that the oncogenic activity of EZH2 in CRPC can occur independently of its Polycomb-repressive function, and targeting the non-PRC2 (polycomb repressive complex-2) function of EZH2 may have therapeutic efficacy (8). ASOs optimized for in vivo delivery present a novel solution by targeting all forms of EZH2 and AR at the mRNA level, leading to ablation of the protein itself rather than inhibition of enzymatic function.

In collaboration with Ionis Pharmaceuticals, we have tested clinical-grade ASOs targeting AR and EZH2. ASOs were delivered to the cells via free uptake, as described previously (22). As shown in Fig. 2A and Supplementary Fig. S4A and B, ASOs inhibited their targets and downstream activities efficiently, as indicated by respective dose-dependent reductions in EZH2/H3K27Me3 and AR/PSA levels. We compared the growth-inhibitory effects of EZH2 ASOs with EPZ-6438 in LNCaP-abl cells, an androgen-independent cell line that relies on the PRC2-independent functions of EZH2 (8). As shown in Supplementary Fig. S5A, EPZ-6438 and EZH2 ASOs both inhibited LNCaP-abl cell growth, but EZH2 ASOs exhibited enhanced efficacy. Furthermore, ASO-AR treatment was able to decrease growth of enzalutamide-resistant LNCaP-EnzR cells (Supplementary Fig. S5B), suggesting that ablation of EZH2 and AR by ASOs may indeed be advantageous to enzymatic inhibition in certain settings.

Consequently, we determined whether the results generated in Fig. 1 could be recapitulated with ASOs to ablate EZH2 and AR. In corroboration with data in Fig. 1, EZH2 ASOs sensitized LNCaP and C4-2B cells to ASO-AR, lowering the IC₅₀ of ASO-AR and shifting the growth-inhibitory curve to the left (Fig. 2B). Furthermore, both CellTiter-Glo assays (Fig. 2C) and soft agar colony experiments (Supplementary Fig. S6A and S6B) demonstrated that ASOs cotargeting EZH2 and AR yielded the most potent cellular growth inhibition. CI analysis indicated that EZH2 and AR ASOs synergized in both LNCaP and C4-2B cells (Fig. 2D). Similar results were obtained using CRISPR/Cas9 to target EZH2 in the presence of ASO-AR (Supplementary Fig. S6C and D). Combination ASO-AR and ASO-EZH2 treatment was also tested in LNCaP-EnzR cells and shown to be more effective at inhibiting growth of this model than either agent alone (Supplementary Fig. S5B).

We sought to evaluate how EZH2 inhibition may sensitize CRPC cells to AR-targeted therapies. RNA-seq was employed to profile transcriptomic changes in C4-2B cells treated with EPZ-6438 or ASO-EZH2. Interestingly, GSEA revealed AR signaling to be the set most significantly activated by either agent (Fig. 3A and B). RT-PCR confirmed that multiple AR target genes were significantly upregulated upon treatment with EPZ-6438 or EZH2 ASOs in LNCaP and C4-2B cells (Fig. 3C).

As a subunit of PRC2, EZH2 associates with chromatin, inducing its compaction and limiting access of transcription factors and ATP-dependent remodeling machinery; therefore, EZH2 inhibition may yield global redistribution of transcription factor cistromes. Genome-wide assay for ATAC-seq was performed to profile chromatin accessibility changes resulting from EZH2 inhibition in C4-2B cells with either EPZ-6438 or ASO-EZH2. As shown in Fig. 3D and Supplementary Fig. S7A, both treatments dramatically increased chromatin accessibility. De novo transcription factor binding site motif analysis was performed on differential accessible chromatin regions and revealed binding sites for AR and AR-associated pioneer factors to be the most enriched motifs upon EZH2 inhibition or ablation (Fig. 3D; Supplementary Fig. S7B). Genome-wide AR ChIP-seq was employed and integrated with ATAC-seq data to directly profile EPZ-6438 and ASO-EZH2–mediated AR occupancy changes. Consistent with motif analysis, these data showed that differentially accessible regions (DAR) were highly associated with AR binding (Fig. 3D; Supplementary Fig. S7A and S7C). Combined, these data propose that inhibition of EZH2 may lead to reprogramming of the AR cistrome and upregulation of AR signaling. These alterations could increase dependence of prostate cancer cells on this pathway and their sensitivity to AR-targeted therapies.

ASOs engineered for in vivo delivery are an emerging class of oncology therapeutics (13). An ASO-based strategy was employed to determine the effect of inhibiting AR and EZH2 in prostate cancer xenografts. The efficacy of AR-targeting ASO as an antitumor agent in C4-2B cells was first evaluated. Following a 3-week time course, xenografts in both active agent groups had significantly lower volumes than those of controls (Fig. 4A). Xenograft volumes were significantly reduced in animals treated with ASO-AR versus enzalutamide, highlighting advantages of the ASO technology; xenograft weights confirmed volume analysis (Fig. 4A).

The same experimental protocol was used to determine the impact of combining ASOs targeting AR and EZH2 on xenograft growth. As observed above, ASO-AR significantly inhibited xenograft growth, whereas ASO-EZH2 (50 mg/kg) had no effect (Fig. 4B). However, combination of ASO-AR and ASO-EZH2 yielded significant reduction in tumor volume versus ASO-AR alone. Measurements of xenograft weight paralleled those of volume (Fig. 4B), and a clear size decrement was appreciable in the combination treatment (Supplementary Fig. S8A). Quantitative RNA in situ hybridization (RNA-ISH) verified significant reductions in mRNA encoding EZH2 and AR (Fig. 4C; Supplementary Fig. S8B and S8C). Tumor growth and weights in the combination group were also significantly reduced when compared with single-agent arms employing equal amounts of total ASOs (i.e., ASO-AR + ASO-Ctrl and ASO-EZH2 + ASO-Ctrl). This ensured that antitumor effects of combination treatment did not result from higher total ASO load (Supplementary Fig. S8D). These data demonstrate the ability of EZH2 inhibition to augment the growth inhibitory effect of AR inhibition in CRPC. In addition, the findings highlight ASO therapy as a novel method of targeting both AR and EZH2 in vivo.

Although advanced prostate cancer often responds initially to therapies that suppress androgen-axis signaling, resistance inevitably develops, leading to the emergence of CRPC. The clinical efficacies of therapies targeting AR, such as abiraterone and enzalutamide, have confirmed that AR signaling remains an important driver of CRPC (3, 4, 23). Approaches to improve the duration of response and address key pathways of resistance are needed. The data presented here suggest that administration of therapies that inhibit EZH2 signaling may sensitize tumor cells to AR-targeting therapies in CRPC. EZH2 inhibitors lead to increased open chromatin and AR binding throughout the genome, upregulating AR signaling. Interestingly, this appears to increase dependence of CRPC cells on this pathway, as sensitivity to AR-targeting therapies is enhanced (Fig. 4D).

Past studies have associated epigenetic aberrations with prostate cancer progression and drug resistance and support targeting of epigenetic mediators as a potential therapeutic strategy. Our group previously found EZH2 to be overexpressed in prostate cancer and associated with progression and poor prognosis (7). Recent studies showed that EZH2 plays a critical role in NEPC development and suggested that inhibition of EZH2 is a promising therapeutic avenue for NEPC that could reactivate AR signaling and restore sensitivity to AR-targeting drugs (9–11). Here, we defined the role of EZH2 inhibition in CRPC adenocarcinoma models and demonstrated that EZH2 inhibition is a viable treatment strategy to restore sensitivity; results were similar in LNCaP cells, proposing EZH2 inhibition may also augment responses to AR-targeted drugs in androgen-dependent stages of prostate cancer. Interestingly, inhibition of EZH2 caused redistribution of AR binding and upregulated AR signaling, suggesting a greater dependence on this pathway underlies ADT sensitization. Recently, a phase II study showed that bipolar androgen therapy resulted in resensitization to enzalutamide in patients undergoing rechallenge (24). Although a different avenue of intervention, these data support our hypothesis that further activation of AR signaling may restore response to ADT.

Through exploration of alternative strategies to inhibit all functions of EZH2 and AR, we found ASOs, a rapidly expanding class of oncology therapeutic agents, as a novel treatment strategy for combinatorial targeting of these factors in CRPC (13). In reports involving other cancer lineages, AZD4785 (ASO targeting KRAS) has been shown to be an attractive therapeutic for treatment of KRAS-driven human cancers, and AZD9150 (ASO targeting STAT3) exhibited promising antitumor activity in patients with treatment-refractory lymphoma and non–small-cell lung cancer (22, 25). Previously, AR ASOs targeting both full-length AR and AR variants have been shown as a rational approach for treatment of AR-dependent CRPC (26). Our data demonstrate that combining AR-ASO with EZH2 ASOs enhances the antitumor effect in CRPC. Furthermore, tumors treated with ASO-AR alone exhibited enhanced growth inhibition compared with enzalutamide, another advantage of ASO technology in this setting. Finally, ASOs have the power to target previously undruggable proteins or disease-associated noncoding RNAs, such as AR variants or Polycomb-independent functions of EZH2 (8, 20). Altogether, data from this study suggest that clinical studies investigating ASO-AR and ASO-EZH2 therapy in patients with CRPC are warranted.

A.R. MacLeod reports receiving a commercial research grant from Ionis Pharmaceuticals and has ownership interest (including stock and patents) in Ionis Pharmaceuticals. Y. Kim has ownership interest (including stock and patents) in Ionis Pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.

Conception and design: L. Xiao, J.C. Tien, A.M. Chinnaiyan

Development of methodology: L. Xiao, J.C. Tien, J. Vo, M. Tan, Y. Zhang, L. Wang, Y. Qiao, S. Shukla, X. Wang, F. Su, A.R. MacLeod, Y. Kim

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Parolia, H. Zheng, X. Jing, E. Luo, A. Delekta, K.M. Juckette, A. Xu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L. Xiao, J.C. Tien, J. Vo, A. Parolia, Y. Zhang, L. Wang, Y. Qiao, S. Shukla, X. Wang, A.R. MacLeod

Writing, review, and/or revision of the manuscript: L. Xiao, J.C. Tien, A.R. MacLeod, Y. Kim, A.M. Chinnaiyan

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Cao, A.S Alva, Y. Kim

Study supervision: A.M. Chinnaiyan

We thank S.M. Dhanasekaran, C. Kumar, M. Cieślik, D. Robinson, Y-M. Wu, S. Shankar, and S. Kregel for helpful discussions. We also thank S. Ellison for assistance in writing and editing of the manuscript. This work was supported in part by a NCI Prostate SPORE (P50CA186786). A.M. Chinnaiyan is a Howard Hughes Medical Institute Investigator, A. Alfred Taubman Scholar, and American Cancer Society Professor as well as is supported by the Prostate Cancer Foundation. L. Xiao is supported by a Department of Defense (DoD) Postdoctoral Award (W81XWH-16-1-0195). J.C. Tien and Y. Qiao are supported by the PCF Young Investigator Award. A. Parolia is supported by the DoD Early Investigator Research Award (W81XWH-17-1-0130). Y. Zhang is also supported by the DoD Early Investigator Research Award (W81XWH-17-1-0134).