Glycosylation-related gene risk model and functional validation of OSTC and TUBA1C in lung adenocarcinoma.

INTRODUCTION: OSTC and TUBA1C drive the malignant progression of lung adenocarcinoma (LUAD) through the modulation of N-glycosylation and the PI3K/AKT signaling pathway; Silencing these two genes markedly suppresses the malignant phenotypes of tumor cells, thereby identifying them as promising novel candidate targets for LUAD diagnosis and therapeutic intervention. METHODS: This study integrated multi-omics data from TCGA and GEO, constructed a risk model (GPRS) for LUAD using 10 machine learning algorithms based on a glycosylation gene set (GRGs). Prognosis was evaluated using Cox regression and survival plots, and analyses of gene function, the microenvironment, and drug sensitivity were conducted. Gene distribution was characterized using single-cell and spatial transcriptomics, and in vitro experiments validated the effects of OSTC and TUBA1C on the function of LUAD cells. RESULTS: This study screened 134 GRGs to construct an 18-gene risk model, which showed high predictive precision in the TCGA-LUAD dataset (1-, 3-, and 5-year AUCs ranging from 0.966 to 0.982). The high-risk group displayed significant enrichment of the PI3K/AKT pathway, had a low immune score, and was sensitive to EGFR inhibitors; Bcl-2 inhibitors showed significant efficacy in the low-risk group. Single-cell profiling revealed high expression of OSTC and TUBA1C in the course of tumor progression. In vitro experiments confirmed that the knockdown of these two genes impairs LUAD cell proliferation and migration and triggers apoptosis, a mechanism correlated with decreased PI3K/AKT phosphorylation levels. CONCLUSION: This study identifies OSTC and TUBA1C as glycosylation-related oncogenes that drive LUAD progression via the PI3K/AKT pathway. The GPRS, derived from a multi-omics machine learning pipeline, serves as a discovery tool that captures the prognostic significance of these genes and reflects LUAD metabolic features, immune status, and drug responsiveness. Together, these findings provide mechanistic insights and candidate targets for precision therapy in LUAD.