Construction and Validation of Prognostic Features Related to Mitochondrial Function and Programmed Cell Death in Lung Adenocarcinoma

Background : In the process of programmed cell death (PCD), mitochondria can release proteins stored within the cell, triggering apoptosis and thereby playing a critical role in cancer progression. This study aimed to develop a prognostic signature related to mitochondrial function and PCD in lung adenocarcinoma (LUAD). Methods : Candidate genes were identified by selecting the overlapping differentially expressed genes from the TCGA-LUAD dataset and genes associated with mitochondria and PCD, which were then subjected to enrichment analysis. Mitochondria- and PCD-linked prognostic genes were further screened using univariate Cox and LASSO regression analyses to establish a risk model. A nomogram integrating risk scores and clinical features was constructed for more accurate prediction of survival in LUAD patients. Immune characteristics across risk groups were assessed via immune cell infiltration, ESTIMATE scoring, and immune checkpoint inhibitor (ICI) profiling. Additionally, cell clusters and subpopulations were identified in the GSE131907 dataset and analyzed through pseudo-time trajectory analysis. Finally, qRT-PCR was employed to validate differential expression of the prognostic genes. Results : We initially identified 50 candidate genes associated with mitochondria and PCD. A risk model based on five prognostic genes (ACSL1, VDAC1, CYCS, GLS2, and MPV17L) was established and proved effective in predicting LUAD patient outcomes. Using the median risk score, patients were divided into high- and low-risk subgroups, with the high-risk group exhibiting significantly poorer survival. The nomogram, incorporating independent prognostic factors (risk score, T-stage, and overall stage), showed improved accuracy in predicting patient survival. Significant differences were observed between the two risk subgroups regarding the infiltration levels of 29 immune cell types (e.g., Macrophages and Th1 cells) and expression of 20 immune checkpoints (including CD70 and CTLA4). Single-cell and pseudo-time analyses indicated that the differentiation states of fibroblasts and epithelial cells were linked to LUAD progression, and active intercellular communication was detected among fibroblasts, epithelial cells, and other cell subpopulations. Conclusion: We constructed a prognostic signature linked to mitochondria and PCD based on five genes, which proved to be a significant reference in the prognosis and treatment of LUAD.