Mechanistic multiscale modeling identifies putative natural tri-target candidates of MAO-B, LRRK2 and A₂A for Parkinson’s disease

Abstract

Parkinson’s disease (PD) is a rapidly growing neurodegenerative disorder for which current dopaminergic and device-based therapies remain purely symptomatic and fail to modify disease progression. Addressing the multifactorial biology of PD under stringent blood-brain barrier constraints requires CNS-penetrant small molecules that can simultaneously engage several validated targets. Here, we combined ADMET-AI profiling, structure-based docking, 100 ns molecular dynamics, MM/GBSA ensemble free energies, and principal component analysis to evaluate falcarinol, 20-hydroxyecdysone, and arnicolide D as putative tri-target candidates of MAO-B, LRRK2 and the A₂A receptor. Falcarinol and arnicolide D occupy a CNS drug-like ADMET space with high predicted BBB penetration and acceptable safety, and show stable, hydrophobically driven binding across all three proteins. In contrast, 20-hydroxyecdysone achieves the most favorable MM/GBSA binding free energies in MAO-B and A₂A (ΔGbind down to -53.4 and − 56.6 kcal·mol⁻¹, respectively) through persistent multi-point hydrogen bonding. Still, it lies outside the optimal CNS physicochemical window. Integrating these orthogonal readouts, we propose an ADMET-constrained multi-target pharmacophore framework for simultaneous MAO-B/LRRK2/A₂A modulation and nominate falcarinol- and arnicolide D-based chemotypes, alongside polarity-masked 20-hydroxyecdysone analogues, as prioritized starting points for experimental validation and next-generation multi-target-directed ligands in Parkinson’s disease.