Integrating prior knowledge inference with computational multi-omics analysis to reveal host antiviral networks of natural compounds against influenza A virus.

Introduction: Influenza A virus (IAV) remains one of the major global health threats, and a rapid emerging resistance to direct acting antivirals underscores the need for host directed therapies. Methods: Here, we integrated prior knowledge inference with a multilevel phenotypic screening and computational multi-omics analysis to identify natural candidate chemical compounds from Traditional Chinese Medicine (TCM)-derived libraries. Twenty compounds enriched for links to antiviral signaling were tested via a GFP-IAV reporter assay, followed by experimental validation using replication kinetics of wild type IAV in MDCK and A549 cells, cytotoxicity measurements, and a viral polymerase minigenome assay. Results: Four structurally distinct compounds, Aloe emodin, Cryptotanshinone, Emodin, and Andrographolide, consistently inhibited IAV replication with low cytotoxicity. The polymerase assay results indicated no substantial direct inhibition of viral polymerase, except for modest effects of Aloe emodin at high concentration. Transcriptomic and proteomic profiling of compound-treated and virus-infected A549 cells showed that all four compounds reprogrammed host antiviral and inflammatory networks, including innate immune and stress response pathways, and virus-host interaction modules. Discussion: These findings nominate the four natural chemical compounds as promising antiviral scaffolds against IAV.