Characterization of androgen receptor properties in mediating transcriptional biomolecular condensates in prostate cancer

One in eight men in the United States are estimated to develop prostate cancer (PCa) during their lifetime. The main oncogenic driver of PCa is the androgen receptor (AR), a ligand activated transcription factor that contains three domains: a large intrinsically disordered N-terminal domain (NTD), a DNA-binding domain (DBD), and a C-terminal ligand-binding domain (LBD). We recently showed that the full-length AR forms nuclear condensates upon androgen stimulation in PCa models. These condensates colocalize with players of the transcriptional machinery and promote the expression of oncogenic programs. We also discovered that in androgen-dependent models, the treatment-resistant AR splice variant V7 (AR-V7), which lacks the LBD, does not form condensates. Nevertheless, in castration-resistant models, AR-V7 condensates form independently of androgens. In this study, we explore the impact of mutating or truncating critical residues or regions associated with AR's transcriptional activity on its capacity to form nuclear condensates in LNCaP cells. We also employ various bioinformatics tools to predict additional AR residues and sequences with high likelihood of undergoing phase separation. Consequently, we generate these predicted truncations and mutations, assessing their potential to form condensates in LNCaP cells. Additionally, we examined the effects of MED1 phosphorylation and treatment with various inhibitors targeting distinct AR domains on the ability of the full-length and AR-V7 proteins to form droplets in vitro and in PCa cells. Our objective is to deepen our comprehension of the factors driving AR condensate formation in PCa. By doing so, we aim to elucidate this novel mechanism of transcriptional regulation and identify novel therapeutic options for patients with advanced stages of PCa disease.