AbstractBackground:Current prostate cancer (PC) therapies for both mCRPC and nmCRPC patients are limited to targeting the ligand-binding domain (LBD). But due to the emergence of LBD lacking variants, and mutations in the LBD of androgen receptor (AR), these therapies become ineffective over time. Thus, there is an unmet need to develop treatments targeting N-terminal Domain (NTD) of the AR to potentially inactivate AR. Although the AR-NTD is a promising target but the structural aspects such as conformational flexibility and lack of a stable secondary structure without defined binding pocket make it technically challenging for designing NTD specific inhibitors. Thus, our efforts are being focused on generating consensus structure and developing small molecules antagonists that bind to AR-NTD.Methods:For NTD targeting we generated a consensus AR-NTD structure by proprietary and integrated drug discovery AI (Artificial Intelligence)-based platform. Multiple MDS (T= 100 ns) together with deep learning neural networks identified two druggable sub-pockets in the consensus NTD structure model (M5). Fragment and structure based virtual screening against one of the deep sub pockets of AR-NTD combined with novel protein motion evaluation methods generated novel chemical matters by cell based assay that resulted nano-molar inhibitors over several SAR rounds.Results:To validate the binding of NTD binding small molecules to the AR-NTD protein in real time, surface plasmon resonance (SPR) analysis was performed. This allowed for the measurement of binding kinetics and affinity, providing robust evidence of the direct interaction between the lead CAP-molecules and AR-NTD with Kd 5-10 micromolar/sec. Research Lead molecules, selected based on their favorable in vitro potency, ADME, pharmacokinetic, and tolerability profiles, were advanced to 28-day AR/ARV7 driven VCaP xenograft model to test for in vivo efficacy. One molecule showed outstanding efficacy and was well tolerated and moved to an ARV7 driven model, 22RV1, a model known for its expression of AR variants & resistance to standard AR-targeted therapies, making it ideal for evaluating the potential of NTD inhibitors. The study results showed significant tumor growth inhibition and impressive synergistic effect when combined with enzalutamide while maintaining stable animal weight. This promising effect of an NTD targeted inhibitor towards 22RV1 xenograft model validates the computer modelling, AI driven structure prediction and virtual screening for generating IP protected allosteric small molecule NTD inhibitor with novel druggable pockets in AR. The novel AR-NTD inhibitor has been further incorporated into two chemistry strategies: AR-NTD targeting PROTACs, AR-NTD & cellular essential protein targeting heterobifunctional molecules; pharmacological activity evaluation is being pursued for these two approaches.Citation Format:Jeffrey N. Lindquist, Sumana Ghosh, Naveen Naganaboina, Nishan Shettigar, Sanjita Banerjee, Michael Green, Ravi K. Muttineni, Hirdesh Uppal. Generation of a best-in-class AR-NTD inhibitor targeting a unique binding pocket identified by computational modeling. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1740.
