AI-Driven Discovery and Optimization of Positive Allosteric Modulators for NMDA Receptors: Potential Applications in Depression

N-Methyl-d-aspartate receptors (NMDARs)
are extensively distributed throughout the central nervous system
(CNS), and their dysfunction is implicated in depressive disorder.
Positive allosteric modulators (PAMs) enhance the receptor’s
sensitivity and activity to agonists without direct activation. In
this study, using structure-based virtual screening and artificial
intelligence (AI)-assisted optimization, we identified Y36, a benzene-substituted piperidinol derivative and potent GluN2A-selective
PAM. Y36 showed higher efficacy (Emax = 397.7%) than GNE-3419 (Emax = 196.4%), reducing the EC50 values and increasing the Emax values for glutamate/glycine at GluN2A receptor.
In chronic restraint stress (CRS) mice, Y36 significantly
alleviated depression-related behaviors in multiple behavioral assessments,
highlighting its superior antidepressant effects. Preliminary studies
also confirmed favorable pharmacokinetic (PK) profiles and blood-brain
barrier (BBB) penetration for Y36, with no signs of addiction,
weight gain, or organ and tissue damage in mice. These results suggest
that Y36 offers promising potential as a novel antidepressant
with multiple antidepressant-like properties.

N-甲基-D-天冬氨酸受体（N-Methyl-d-aspartate receptors, NMDARs）广泛分布于中枢神经系统（central nervous system, CNS），其功能异常与抑郁症相关。正变构调节剂（Positive allosteric modulators, PAMs）无需直接激活受体，即可增强其对激动剂的敏感性与活性。本研究通过基于结构的虚拟筛选结合人工智能（artificial intelligence, AI）辅助优化，成功鉴定出化合物Y36——一种苯取代哌啶醇衍生物，且为强效GluN2A选择性正变构调节剂。相较于GNE-3419（最大效能Emax=196.4%），Y36的最大效能更高（Emax=397.7%），可降低谷氨酸/甘氨酸作用于GluN2A受体时的半最大效应浓度（EC50）并提升其最大效能。在慢性束缚应激（chronic restraint stress, CRS）小鼠模型中，Y36在多项行为学评价中显著改善了抑郁样行为，凸显其优异的抗抑郁效果。初步研究同时证实，Y36具备良好的药代动力学（pharmacokinetic, PK）特征与血脑屏障（blood-brain barrier, BBB）穿透能力，且未在小鼠体内观察到成瘾、体重增加或器官与组织损伤等不良现象。上述结果表明，Y36作为一种兼具多重抗抑郁样活性的新型抗抑郁候选化合物，具有可观的开发潜力。