Unraveling potent Glycyrrhiza glabra flavonoids as AKT1 inhibitors using network pharmacology and machine learning-assisted QSAR

Abstract

Glycyrrhiza glabra (G. glabra) phytocompounds have been reported to interact with neurological targets, including those implicated in epilepsy, and may modulate epilepsy-related targets. While substantial evidence supports their potential antiepileptic effects, the underlying molecular mechanisms remain unclear. This study aims to elucidate the molecular mechanism of G. glabra phytocompounds by integrating network pharmacology and machine learning (ML)-based quantitative structure-activity relationship (QSAR) techniques. Network pharmacology analysis identified AKT1 as a key epilepsy-related target, and four ML-based 2D-QSAR models were developed using AKT1 inhibitors. The developed models underwent comprehensive validation, including internal and external validation, Y-randomization, statistical analysis, and applicability domain assessment to ensure robustness and reliability. Among them, the Multilayer Perceptron (MLP) model excelled as the most robust and demonstrated superior predictive ability with a correlation coefficient r²_training = 0.95, r²_test = 0.84, and cross-validation coefficient q² = 0.72. The MLP model accurately predicted pIC₅₀ values of phytoflavonoids, leading to the identification of 19 active molecules through the activity atlas model. ADME and drug-likeliness screening narrowed the selection to eleven promising compounds for further docking analysis. Molecular docking highlighted glabranin and 3'-hydroxy-4'-O-methylglabridin as top lead compounds with a binding energy of - 5.75 and - 5.37 kcal/mol, respectively. Additionally, 400 ns molecular dynamics simulation confirmed the structural and conformational stability of the glabranin-AKT1 complex, further reinforced by Principal Component Analysis, free energy landscapes, and Molecular Mechanics Poisson-Boltzmann/Generalized Born Surface Area. Collectively, these findings underscore the potential of G. glabra phytocompounds as promising antiepileptic candidates, paving the way for future advancements in this field.

Keywords: Glycyrrhiza glabra; AKT1; Machine learning QSAR; Molecular docking; Molecular dynamics simulation; Network pharmacology.