In Silico discovery of novel androgen receptor inhibitors for prostate cancer therapy using virtual screening, molecular docking, and molecular dynamics simulations.

This study employs an integrated computational approach to identify novel androgen receptor (AR) inhibitors, a key target in prostate cancer (PC) therapy. The full-length AR structure was modeled using MODELLER v10 (template: 1GS4) and validated via Ramachandran analysis, DOPE scoring, and normal mode analysis. A ligand-based pharmacophore derived from 20 known AR inhibitors guided high-throughput virtual screening and molecular docking with AutoDock Vina. ADMET profiling assessed pharmacokinetics, while in silico target prediction, STRING-based PPI network analysis, and Gene Ontology enrichment elucidated the functional role of AR. The stability of the AR-ligand complex was evaluated through a 100-ns molecular dynamics simulation using GROMACS, with RMSD analysis. MODELLER achieved 92.5% sequence identity, 99% query coverage, and a 2.30 Å resolution, yielding the optimal model (DOPE score: - 29,412.36), validated by Ramachandran analysis (98.33% favored residues) and normal mode analysis (eigenvalue: 5.28563e-04). The pharmacophore model (AUC: 0.92, EF: 8.5, MCC: 0.78) facilitated virtual screening and docking, identifying Estrone (ZINC000013509425) as the lead inhibitor (docking score: - 10.9 kcal/mol). Key interactions included hydrogen bonding with Asn705(A) and hydrophobic contacts with Trp741(A), Leu704(A), Met742(A), and Met780(A). ADMET analysis confirmed favorable pharmacokinetics, while network analysis reinforced AR's role in oncogenic pathways. Molecular dynamics simulations indicated complex stability, with protein RMSD stabilizing at 1.5-2.0 Å and ligand RMSD at 3.5-4.0 Å. Estrone was identified as a potent AR inhibitor with strong binding, stable dynamics, and favorable pharmacokinetics for PC therapy.