Abstract / Summary
Abstract Objective To systematically dissect multi-level molecular perturbations driven by serine/threonine kinase 11 ( STK11 ) mutations in lung adenocarcinoma (LUAD) using integrative multi-omics. Methods TCGA-LUAD (65 mutant, 492 wild-type) and CPTAC-LUAD (21 mutant, 89 wild-type) data were analyzed; samples were stratified by mutation status and protein expression; differential expression, kinase activity inference, immune infiltration, survival, AI-driven drug repurposing, molecular docking, and immunotherapy cohort validation were performed. Results STK11 inactivation mainly arose from truncating mutations and copy-number loss, with weak mRNA-protein correlation (r = 0.39). Mutant tumors exhibited upregulated oxidative phosphorylation and MYC targets, downregulated interferon-γ response and PD-1/PD-L1 pathways, and extensive phosphoproteome remodeling (2,030 phosphoproteome-specific molecules versus 563 proteome-specific DEPs). Kinase activities of CDK1/2, ERK1/2, and mTOR were enhanced, whereas AMPKA1, PKACA, and Akt1 were suppressed. The microenvironment showed an immune-desert phenotype with reduced effector T/NK cells and PD-L1. Low STK11 protein independently predicted poor prognosis (HR = 5.90, 95% CI 1.33–26.22), with the Mut-Low subgroup having shortest survival; STK11 mutation independently predicted worse immunotherapy outcome (HR = 1.686, 95% CI 1.187–2.396). Intersection of multi-omics layers yielded 35 hub genes; AI screening and docking suggested ruxolitinib stably binds STK11. Conclusions STK11 mutations drive a “triad” pro-tumorigenic circuit through kinome remodeling, metabolic reprogramming, and immune desertification. The stratification strategy combining mutation status with protein expression improves prognostic stratification. The multi-omics perturbation landscape and identified candidate compounds provide a new basis for precision therapy targeting this subgroup.