Androgen receptor mediated gene regulatory networks in prostate cancer

Abstract

Prostate cancer (PCa) is the most common cancer diagnosed in men. The androgen receptor (AR) is a well-studied master regulator of PCa. Patients with metastatic PCa develop resistance to AR-targeted therapies, which can arise in part from tumour microenvironment signaling. AR activity is often maintained in treatment-resistant metastatic PCa alongside other important changes including epithelial-to-mesenchymal transition and hypoxia. It is challenging to define the causative/driver events as treatment resistance emerges due to the cellular and genomic heterogeneity in prostate tissue.In this study, I have undertaken a meta-analysis of datasets mapping AR recruitment to chromatin to identify conserved features of AR binding to the genome across clinical datasets. To enhance this approach, a conventional peak calling and a novel machine learning approach, LanceOtron was employed, to identify binding sites that would otherwise be missed by applying rigid thresholding to peak calling. Chromatin immunoprecipitation (ChIP), is used to map AR recruitment to the genome, which conventionally has sensitivity issues and creates challenges when working with low cell numbers/heterogenous samples. To address this, I have used a “low cell ChIP” method known as MNase ChIP for the first time to assess AR recruitment to chromatin.Finally, to provide insights into the transcriptional impact of stromal and immune cells on PCa cells, a pre-clinical co-culture model was used. Using a combination of phenotypic assays, ChIP, transcript profiling and metabolomics I show that co-culture with either stromal or immune cells induces HIF1α signaling and TGF beta responses, leading to a more glycolytic state with partial EMT in cells and reduced recruitment of AR to many canonical ARBs. This recapitulates some of the key changes that occur in tissue during the emergence of treatment resistance disease and provides a framework for future translational studies to further address the crosstalk between this signaling and changes in AR activity.

Date of Award	Dec 2022
Original language	English
Awarding Institution	Queen's University Belfast
Sponsors	Marie Sklodowska Curie COFUND
Supervisor	Ian Mills (Supervisor), Simon McDade (Supervisor) & Nuala McCabe (Supervisor)