Identification and validation of hypoxia and energy metabolism-related gene signatures in lung adenocarcinoma and their potential implications for immunotherapy.

e20587 Background: Lung cancer is the leading cause of cancer-related deaths globally, with non-small cell lung cancer (NSCLC) constituting a majority of cases. Immune checkpoint inhibitors targeting PD-1/PD-L1 have improved 5-year survival rates, yet nearly half of NSCLC patients fail to respond, underscoring critical need for effective predictive biomarkers to identify responsive individuals. Hypoxia and energy metabolism in the tumor microenvironment (TME) have emerged as pivotal factors influencing ICI efficacy. Methods: Transcriptomic data from TCGA and GEO datasets were analyzed to identify hypoxia and energy metabolism-related subtypes in lung adenocarcinoma (LUAD) by ssGSEA scoring. The prognostic risk model was constructed using LASSO-Cox regression on TCGA-LUAD cohort and validated by two GEO cohorts. Immune cell infiltration, immune checkpoint gene (ICGs) expression, tumor immune dysfunction and rejection (TIDE) scores were further assessed between high-risk and low-risk groups. Single-cell RNA data was used to assess gene expression in T cell subtypes. The expression of hub genes was validated through qRT-PCR. Results: We found that patients with high hypoxia and low energy metabolism scores (HL) had a significantly worse prognosis compared to the ones with low hypoxia and high energy metabolism scores (LH). A total of 138 core differential genes were identified between the HL and LH groups. LASSO-Cox regression identified nine genes ( RHOV , CYP4B1 , KRT6A , TNS4 , HMGA2 , FIGF , SFTPB , TMPRSS11E and GRIA1 ) to construct a prognostic risk model. Patients with high risk scores exhibit significantly worse overall survival, and the model was validated using two independent cohorts. We found increased infiltration of Tregs in the high-risk group, and expressions of RHOV and TNS4 correlated with this infiltration. Additionally, A high-risk score was associated with lower expression of immune cell-dominated ICGs, a higher TIDE score and high partial response rates. At the single cell level, RHOV and TNS4 were mainly expressed in TOX+ CD4-Tfh and FOXP3+ CD4-Tregs, respectively, and both showed significantly higher expression in the non-major pathological response patients. Finally, the expression of nine hub genes in tumor samples of progressive disease (PD, 19) compared to stable disease (SD, 20) were validated by RT-qPCR, showing consistent pattern with the risk stratification. Specifically, TNS4 and RHOV showed significantly higher expression in PD. Conclusions: The identified hypoxia and energy metabolism-based prognostic signature offers a robust tool for stratifying LUAD patients, predicting immunotherapy outcomes, and guiding personalized treatment. Insights into the interplay between hypoxia, metabolism, and immune modulation in the TME may guide therapeutic advancements in NSCLC.