Active ingredients of Astragalus-ginseng against COPD-associated sarcopenia: network pharmacology and molecular...

Chronic obstructive pulmonary disease (COPD)-associated sarcopenia is a critical comorbidity associated with high morbidity and mortality, driven by chronic systemic inflammation, oxidative stress, and proteometabolic dysregulation. The Astragalus membranaceus-Ginseng (HQ-RS) herbal pair is a traditional remedy for consumptive diseases; however, its molecular mechanisms against this comorbidity remain undefined.

We integrated transcriptomic data from four Gene Expression Omnibus (GEO) datasets (two discovery cohorts and two validation cohorts) for COPD and sarcopenia. Differentially expressed genes (DEGs) were identified using the thresholds of P < 0.05 and -log₂FC- ≥ 1. Active compounds of HQ-RS were screened using the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP) database (oral bioavailability (OB) ≥ 30%, drug-likeness (DL) ≥ 0.18). Candidate targets were identified by intersecting drug targets with shared disease targets (GeneCards relevance score > 20). A protein-protein interaction (PPI) network was constructed to identify hub genes using multi-centrality metrics. Molecular docking was performed with celecoxib as a positive control. Three independent 100 ns molecular dynamics (MD) simulations of the FMNT-PTGS2 complex were then performed, and the trajectories were analyzed using the Simulation Interaction Diagram (SID) tool.

We identified 1,267 COPD-specific and 115 sarcopenia-specific DEGs, yielding nine shared DEGs. Intersection of drug targets with 279 shared disease targets identified 36 candidate targets. PPI network analysis highlighted PTGS2, CDKN1A, NFE2L2, NOS3, and MMP2 as key hub genes. Molecular docking revealed that formononetin (FMNT) exhibited the most favorable binding affinity to PTGS2 (-8.94 kcal/mol), comparable to the positive control celecoxib (-7.97 kcal/mol). Three independent 100 ns MD simulations of the FMNT-PTGS2 complex demonstrated that the system reached dynamic equilibrium after approximately 20 ns in all replicates, with the ligand maintaining a stable binding pose throughout the simulations. Molecular Mechanics Generalized Born Surface Area (MM-GBSA) calculations across three replicate simulations yielded a mean binding free energy of -47.82 ± 1.13 kcal/mol, supporting the thermodynamic favorability of the interaction. Protein-ligand interaction analysis identified sustained hydrogen bonds and hydrophobic contacts with key active-site residues.

This study computationally predicts that the FMNT-PTGS2 interaction may represent a potential mechanism underlying the efficacy of the HQ-RS herbal pair in treating COPD-associated sarcopenia. This study is entirely computational and limited by small discovery cohort sample sizes (COPD n = 6, sarcopenia n = 51). These findings provide a multi-component, multi-target mechanistic hypothesis that requires experimental validation.