基于线粒体基因的宫颈癌新型程序性细胞死亡相关预后风险模型

Despite established links between mitochondrial (mt) function, programmed cell death (PCD) dysregulation, and cervical cancer progression, the prognostic potential of integrating PCD- with mitochondria-related genes remains elusive. To address this gap, this study aims to develop a reliable RiskScore model leveraging PCD and mt characteristics for cervical cancer prognostic prediction.

mRNA expression profiles with clinical data of 309 cervical cancer patients from The Cancer Genome Atlas (TCGA), complemented by datasets from the Gene Expression Omnibus (GEO) and the European Molecular Biology Laboratory's European Bioinformatics Institute (EMBL-EBI) databases were utilized. We identified differentially expressed genes (DEGs) using the R "limma" package. Least absolute shrinkage and selection operator (LASSO) Cox regression analysis refined gene selection for RiskScore construction. The "rms" package was employed to develop a nomogram model. Additionally, we leveraged generalized structured component analysis (GSCA) for mutation analysis and Cell-type Identification By Estimating Relative Subsets Of RNA Transcripts (CIBERSORT) for immune infiltration assessment. Single-cell sequencing data were processed with the "Seurat" package.

Our study identified 119 PCD-related mt DEGs. A prognostic model was developed based on IL1B, NAMPT, PRKAB2, SLC2A1, ACAA2, E2F1, GZMB, and BNIP3 , yielding a RiskScore that stratified patients into low- and high-risk groups with significant survival differences. The model's predictive accuracy was validated with high area under the curve (AUC) values and confirmed as an independent prognostic factor. A nomogram model was constructed for predicting patient survival rates, showing good agreement with actual outcomes. Mutation analysis indicated a low mutation rate in our model genes, with a significant negative correlation between RiskScore and tumor mutation burden. Single-cell analysis uncovered GZMB's high expression in T/natural killer (NK) cells, with dynamic changes along pseudotime trajectories.

This study introduces a RiskScore model by using PCD-related mt genes for cervical cancer prognosis, offering a personalized therapeutic strategy and insights into immunotherapy potential, potentially advancing future clinical applications.