Integrative multi-omics analysis identifies a stage-immune prognostic signature and therapeutic targets in lung adenocarcinoma

To understanding the molecular mechanisms and tumor microenvironment characteristics that contribute to LUAD progression. Multi-omics data, including miRNA-seq, RNA-seq, DNA methylation, whole-exome sequencing (WES), clinical, and survival data of Lung adenocarcinoma (LUAD) samples, were acquired from The Cancer Genome Atlas (TCGA). Differentially expressed genes (DEGs) were identified using the “limma” algorithm, with Gene Ontology (GO) annotations performed. The ESTIMATE algorithm was applied to determine immune and stromal scores, CIBERSORT was used for immune cell infiltration analysis. Patients were stratified based on LUAD stages and ESTIMATE scores for survival analysis and prognostic model development using univariate and multivariate Cox regression, followed by lasso regression. LUAD cell line experiments validated key findings using siRNA transfection, qRT-PCR, Western blot, and immunofluorescence. Patients with Stage I LUAD exhibited higher overall survival compared to those with stages II-IV (P < 0.0001). DEGs were identified between stage I and stages II-IV, revealing 144 upregulated and 331 downregulated genes. In the context of LUAD occurrence, 544 genes were upregulated and 877 downregulated, with 67 upregulated and 133 downregulated genes overlapping between stages and LUAD occurrence. GO analysis highlighted pathways involving mitotic progression and cytokinesis. Higher ESTIMATE scores correlated with better prognosis and increased immune cell infiltration, indicating higher immunotherapy response. The prognostic model incorporated five alterations (ANGPTL1, WNT7A, chr11:102207640–102208464, chr5:82876126-82878122, SLC38A3) and demonstrated robust predictive performance. In NCI-H1395 LUAD cells, siRNA knockdown confirmed ANGPTL1 and WNT7A as pro-tumorigenic, while SLC38A3 acted protectively. Our findings provide a clinically actionable framework for risk stratification and nominate ANGPTL1, WNT7A, and SLC38A3 as potential targets for combination therapies integrating immune modulation and molecular intervention in LUAD. The final prognostic signature integrates mRNA expression and two somatic alteration regions from WES, selected through LASSO-Cox regression. This study addresses a critical gap in stage-aware, immune-informed LUAD risk modeling.