Should we treat ductal prostate cancers with androgen receptor targeted therapies?.

Abstract

Introduction & Objectives: Ductal prostate cancer (DAC) is an aggressive, histological variant of prostate cancer. Despite poor outcomes, DACs are currently treated with androgen receptor (AR) targeted therapies, similar to prostatic acinar adenocarcinoma (PAC). We aimed to assess the DAC-response to AR therapies and identify a biologically targeted therapy for these tumours. Method(s): All DAC patients treated with adjuvant or salvage ADT post radical prostatectomy (RP) from 2005 to 2018 were compared with cohorts of high-risk PAC patients using Kaplan-Meier and Cox regression models. A separate group of patients who received neoadjuvant therapies prior to RP were evaluated to assess pathological responses to AR directed therapies. 20 DAC and 10 PAC tumours, matched for stage, grade and volume, were micro-dissected from radical prostatectomy samples. Whole exome sequencing, RNA sequencing and Hallmark pathway analyses were performed. To model DAC, organoids were developed in Matrigel from patientderived xenografts (PDXs) originating from DDR-proficient radical prostatectomy (287R, 275R) and BRCA2 heterozygous mutant metastatic (201.1) tumours, retaining key histologic and genomic features. These organoids were exposed to different PARP inhibitors (PARPi) and androgen pathway inhibitors (ARPi) and cell viability and growth responses were assessed. Results were further validated in vivo using DDRproficient 287R PDXs. Result(s): 26 DAC patients and 50 PAC patients received ADT post RP. DAC patients receiving ADT had worse 5-year MFS and 5-year OS (both, P < 0.05) compared with PAC patients. Further, none of the DAC patients (0/15) who received only neoadjuvant ADT prior to RP had any pathologic downgrading. On genomic analyses, DACs had higher rates of mutations compared to matched PAC tumours. Unsupervised hierarchical clustering of RNA expression identified a distinct DAC cluster, with downregulation of AR pathways and enrichment of HR pathways, compared to PAC. Overall, PARPi/ARPi combination treatment significantly reduced organoid viability compared to PARPi alone or ARPi in DDR-proficient and heterozygous BRCA2-mutant DAC tumours. The levels of synergy were comparable regardless of which PARPi and ARPi agents were combined. In vivo results using DDRproficient PDX 287R confirmed the efficacy of PARPi + ARPi combination by reducing tumour volume by 58% compared to control (P = 0.0198) and by 40% versus PARPi alone (P = 0.0326). Conclusion(s): In contrast to PACs, primary DACs have a unique genomic and transcriptomic landscape with high rates of mutations associated with intrinsic androgen resistance. PARPi+ARPi were more effective at decreasing DAC organoid viability than either treatment alone, regardless of the HR status. Our results show that PARPi increases the efficacy of ARPI in DAC, providing the rationale for a pre-planned phase 1/2 study.