Abstract
BACKGROUND: Guanylate-binding proteins (GBPs) induced by interferons play a crucial role in inflammatory signaling. This study aimed to investigate their anti-cancer properties in lung adenocarcinoma (LUAD). METHOD: Seven GBPs genes were obtained from the TCGA-LUAD and GSE31210 datasets. The GBPs score was computed for each patient using ssGSEA, followed by conducting WGCNA. Differentially expressed genes (DEGs) were identified with the Limma package, and further refined by univariate and multivariate Cox regression analyses to develop a risk score model. Kaplan-Meier survival and time-dependent receiver operating characteristic (ROC) analyses were performed with the survminer and timeROC packages, respectively. Immune microenvironment was profiled using ESTIMATE and ssGSEA algorithms, while the pRRophetic package was applied to predict chemotherapeutic sensitivity. Relevant signaling pathways were identified by GSEA. Finally, the functional roles of key genes were experimentally validated using CCK-8, wound healing, and Transwell assays. RESULTS: The GBPs score was significantly higher in para-carcinoma tissues. The green-yellow module (β = 7), which contained 422 genes, was closely associated with the GBPs score. Intersection of this module with the DEGs yielded 92 overlapping candidates. A 4-gene risk model constructed from Cox regression demonstrated high predictive accuracy for 1-, 3-, and 5 year Overall Survival. Patients were assigned by the median risk score into low- and high-risk groups, with low-risk patients showing better prognostic outcomes and higher immune infiltration scores, particularly of CD4(+) T cells and activated B cells. A nomogram combining multiple clinical factors exhibited robust prognostic performance. We also further identified potential drugs for treating different risk groups of LUAD. Pathways such as reactive oxygen species (ROS) and Myc target V2 were enriched in high-risk groups, suggesting potential targets for immunotherapy. In vitro assays confirmed that knockdown of the model gene CBFA2T3 significantly suppressed the migratory and invasive capacities of A549 and NCI-H838 cells. CONCLUSION: We analyzed the GBPs phenotype associated with LUAD progression and developed a risk score model for patient prognosis. This finding provided a reference model for clinically assessing patient prognosis, contributing to personalized treatment for LUAD patients.