Computational Evaluation of Chromen-2-one-Oxadiazole Derivatives as Next-Generation Non-Steroidal Androgen Receptor Antagonists

Castration-resistant prostate cancer (CRPC) develops mainly because mutation occurs in the androgen receptor (AR) which reduce the effectiveness of existing anti-androgen drugs. This creates a need for new AR antagonists that can remain active even in mutated receptors. In this study, a complete computational approach combining machine learning (ML)-based virtual screening, molecular docking, Density function theory (DFT) calculations, drug likeness prediction, molecular dynamic simulations and MM/PBSA analysis was used to identify chromen-2-one-oxadiazole derivatives as potential non-steroidal AR antagonists. A random forest model trained on BindingDB data predicted the activity of designed compounds after screening. Among them, VP1-071 and VP1-095 showed best results, with strong docking scores against wild type-AR (-10.8 and -10.4 kcal/mol). Both compounds also maintained good binding with resistance mutants T877A and W741L. DFT calculations indicated stable electronic properties, while 200ns molecular dynamic simulations confirmed stable protein-ligand complexes. MM/PBSA analysis further supported strong binding interactions.去势抵抗性前列腺癌（Castration-resistant prostate cancer, CRPC）的发生主要源于雄激素受体（androgen receptor, AR）发生突变，导致现有抗雄激素药物的疗效下降，因此亟需开发即便在AR突变体中仍能保持活性的新型AR拮抗剂。本研究采用整套计算分析方法，结合基于机器学习（machine learning, ML）的虚拟筛选、分子对接、密度泛函理论（Density function theory, DFT）计算、药物相似性预测、分子动力学模拟以及MM/PBSA分析，旨在鉴定色原烷-2-酮-噁二唑衍生物作为潜在的非甾体类AR拮抗剂。本研究基于BindingDB数据集训练得到随机森林模型，用于在虚拟筛选后预测所设计化合物的活性。筛选结果显示，VP1-071与VP1-095表现最优，针对野生型AR的对接得分分别为-10.8和-10.4 kcal/mol，结合作用强劲。两种化合物对耐药突变体T877A与W741L亦保持了良好的结合能力。密度泛函理论计算结果表明这类衍生物具备稳定的电子特性，而200 ns分子动力学模拟则验证了蛋白-配体复合物的结构稳定性；MM/PBSA分析进一步证实了其较强的结合相互作用。