Identifying androgen receptor antagonists using a metabolically competent high-throughput screening assay

Androgen receptor (AR) is a nuclear receptor with a well-established role in sexual function and development. Modifications in AR can lead to endocrine disruption, cancer, and other diseases, making it imperative to identify compounds that influence these changes. AR modulators have been identified using immortalized cell lines in a high-throughput screening assay. However, most of these methods do not incorporate metabolism, leading to misclassification of compounds that normally require it to become AR modulators. Metabolism transforms exogenous parent compounds into metabolites that are easier to excrete, and normally less active than the parent. However, some metabolites modulate AR more effectively than the parent compound. Incorporating metabolism into a large compound screen can identify active metabolites as potential AR modulators. In this study, we optimized a high-throughput screening assay that included rat liver microsomes (RLM) as an exogenous metabolic system to detect AR antagonists. A robotic screen of the LOPAC library + 88 Tox21 compounds (a total of 1365 unique compounds) was then performed to validate the assay and identify any bioactivated AR modulators within the test library. Fifty-five compounds were identified as potential AR antagonists; 9 compounds out of these 55 compounds were found to have significant potency shifts between RLM free and RLM assays, suggesting the necessity of metabolism for their AR activity. A concurrent assay using heat-inactivated RLM was conducted to discern the true activity of each compound. Metabolic stability assays were also performed on the top compounds to clarify their ability to transition from parent to metabolite using RLM. Four compounds were identified as novel parent compounds requiring metabolism to become more potent AR antagonists. However, only 4,5-dianilinophthalimide (DAPH) displayed a clear concentration-response curve with a more potent IC₅₀ when RLM was included compared to its parallel screens, identifying it as a true AR antagonist requiring metabolism.