Exploring CRISPR/Cas9 as a tool to investigate the role of Androgen Receptor in basal to luminal cell differentiation using human prostate cell lines

The prostate epithelium is defined in terms of a hierarchical structure, where each cell type can differentiate into the next cell in the lineage until terminally differentiated luminal cells are formed. Androgen receptor (AR) appears to be essential for differentiation of basal to luminal cells. However previous research was predominantly conducted on mouse models, which have a different epithelial structure and therefore this project focussed on 3D models of human prostate cell lines. This project aims to knockout (KO) AR in BPH-1 PPmO cells (benign cell line with fluorescent differentiation indicator) and P4E6 cells (localised cancer cell line) using CRISPR/Cas9, then to generate tools with which to investigate the effects of AR KO on prostate differentiation. The hypothesis was that AR expression is essential for cell differentiation during prostate cancer (PCa) development. Copy number analysis determined that BPH-1 PPmO contains three copies of AR, whilst P4E6 contains one copy. This allowed for the generation of P4E6 homozygous and BPH-1 PPmO heterozygous clones. The CRISPR/Cas9 lentivirus successfully created edits in both cell lines, with the highest efficiency occurring in P4E6 cells. Analysis via qRT-PCR confirmed that AR was successfully knocked out in both the BPH-1 PPmO and P4E6 clones. Growing prostate basal cells in 3D culture induces differentiation and the expression of AR, by mimicking the prostate in vivo. The differentiation indicator only fluoresced in wild-type (WT) BPH-1 PPmO spheroids, indicating AR may indeed be required for differentiation. However the survival of ARKO clones demonstrated that AR is not essential for basal cell survival. Future work should focus on further optimising differentiation via the addition of retinoic acid (RA) to spheroids and engrafting the edited cell lines into immunocompromised mouse models. This project has generated a wide selection of ARKO models which are suitable for the investigation of the role of AR during PCa development