Triaptosis-related prognostic model for lung adenocarcinoma based on transcriptomic analysis and experimental validation

Abstract Background The role of triaptosis, as a novel form of regulated cell death (RCD), remains poorly understood in lung adenocarcinoma (LUAD). This study scout to identify triaptosis-related prognostic genes in LUAD and construct a prognostic model. Methods Triaptosis-related gene (TRG) scores were calculated using ssGSEA. Prognostic genes were identified through differential expression analysis, Cox regression, and expression validation, and a prognostic model was constructed using machine learning algorithms for risk stratification. We further explored functional enrichment, the tumor microenvironment, immune infiltration, and drug sensitivity. Single-cell analysis was integrated with Scissor and TRG sets to identify key cell types and investigated the expression of prognostic genes. Finally, reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and Western-blot were performed to validate the expression of genes in LUAD cells. Results A prognostic model based on 14 TRGs (AURKB, NUF2, RAB3B, S100P, TROAP, ADAMTS8, C1QTNF7, CHRDL1, GRIA1, HLF, MS4A2, SCN7A, SFTPC, and SLC15A2) was constructed using StepCox[forward] + RSF, stratifying patients into high- and low-risk groups. Different risk groups exhibited distinct pathway enrichment, immunological profiles, and drug responses. Single-cell analysis revealed that AURKB, HLF, S100P, SFTPC, and TROAP were differentially expressed between LUAD and controls in epithelial, T, and myeloid cells. Myeloid cells exhibited the highest TRG activity and were closely associated with LUAD prognosis. Experimental validation verified the bioinformatics predictions, confirming the upregulation of AURKB and the downregulation of HLF and SLC15A2 in LUAD cells at both mRNA and protein levels. Conclusion The TRG-related risk model captures distinct biological and immunological features associated with LUAD prognosis, offering transcriptomic-level insights into triaptosis-associated tumor heterogeneity. However, its predictive performance requires further optimization before clinical translation, and the functional links to triaptosis warrant experimental validation.