Moringa oleifera phytochemicals were predicted to target insulin resistance proteins using a modified network pharmacology and molecular docking approach. Two hundred ninety M. oleifera phytochemicals with their aglycones, acetylase and myrosinase degradation products were compiled from literature and phytochemical databases. Nine protein targets were identified from the intersection of gene lists with high relevance to insulin resistance from GeneCards and DisGeNET and the target genes predicted by reverse screening using Swiss Target Prediction: protein-tyrosine phosphatase 1B (PTPN1), 11-beta-hydroxysteroid dehydrogenase 1 (HSD11B1), peroxisome proliferator-activated receptor α (PPARα), peroxisome proliferator-activated receptor γ (PPARγ), PI3-kinase p85-alpha subunit (PIK3R1), insulin receptor (INSR), tumor necrosis factor α (TNF-α), endothelial nitric oxide synthase (eNOS) and hepatic lipase (LIPC). Binding affinities of phytochemicals with the targets were predicted using Autodock Vina. The predicted binding affinities were classified according to calculated thresholds using receiver operating characteristic (ROC) calculations of binding affinities of: (a) binders (annotated drugs and other molecules with known interaction with each target), and (b) decoys (molecules not expected to bind to a specific target). In addition, ubiquitous phytochemicals were filtered out to differentiate the effect on insulin resistance of M. oleifera from that of other plants. The resulting phytochemical-protein interaction network was visualized using Cytoscape. All mentioned targets, except hepatic lipase, were key targets based on various network centrality measures. Previous studies on murine models have shown that isothiocyanate-rich M. oleifera extracts ameliorate insulin resistance. Using our approach, the following phytochemicals, with predicted moderate bioavailability, high GI absorption, and probable binding with insulin resistance targets, are recommended for further in vivo or in vitro validation for insulin resistance activity: boldione (a steroid); aurantiamide acetate and aurantiamide (peptide derivatives); O-ethyl-(3,4-dihydroxyphenyl)methyl carbamothioate and O-methyl-N-(4-hydroxyphenyl)methyl carbamothioate (thiocarbamates); 4α,6α-dihydroxyeudesman-8β,12-olide (a sesquiterpenoid); sanleng acid and tianshic acid (fatty acid derivatives); 2′,5,5′,7-tetrahydroxyflavone; 2′,3,5,7-tetrahydroxyflavone; and 6-hydroxykaempferol (flavonoids). By combining network centrality measures of targets, using ROC-derived thresholds for docking energies, and considering ubiquity of phytochemicals, our refined network pharmacology approach may aid in discovering key bioactive phytochemicals as potential chemical markers for standardization and differentiation of an herbal drug.
Santos-Enriquez et al. (Tue,) studied this question.