Key points are not available for this paper at this time.
Background Glioblastoma (GBM) remains a highly aggressive malignancy with limited effective therapeutic options. Integrating traditional medicine resources with artificial intelligence–based analytical strategies may accelerate the identification of novel biomarkers and drug targets. Methods We applied a network pharmacology framework to screen bioactive compounds of Acorus tatarinowii and identify intersecting targets with GBM-related genes. LASSO Cox regression was performed in TCGA-GBM to construct a prognostic model. Multi-omics analyses, including transcriptomic validation, CPTAC proteomics, copy-number alteration profiling, mutation landscape characterization, immune infiltration assessment, and pharmacogenomic correlation analysis, were conducted to prioritize key targets. Functional experiments based on genetic perturbation of KCNH2, including colony formation, CCK-8 proliferation, wound-healing, apoptosis assays, mitochondrial membrane potential measurement, and xenograft models, were performed to evaluate its biological effects. Results Twenty-five intersecting targets were identified, and an eight-gene LASSO signature demonstrated favorable prognostic performance. Among these genes, KCNH2 showed consistent transcriptomic upregulation, proteomic validation, association with genomic instability, immune modulation patterns, and drug sensitivity correlations. Functional assays confirmed that KCNH2 silencing suppressed proliferation and migration-associated behavior while inducing mitochondria-dependent apoptosis in GBM cells. In vivo , KCNH2 knockdown significantly inhibited tumor growth. Conclusion This integrative AI-driven and multi-omics strategy identifies KCNH2 as a biologically relevant and translationally informative candidate target potentially linked to the anti-glioma effects of Acorus tatarinowii. These findings highlight the value of combining traditional medicine resources with computational prioritization and experimental validation to uncover biologically actionable mechanisms in GBM.
Song et al. (Thu,) studied this question.