Systematic characterization of pyroptosis-related gene patterns identifies potential prognostic inflammatory...

Sepsis is a life-threatening syndrome driven by a dysregulated host response to infection, yet reliable prognostic biomarkers remain limited. Pyroptosis has emerged as an important contributor to immune dysregulation in sepsis. Here, we systematically characterized pyroptosis-related gene (PRG) patterns using public transcriptomic cohorts (GSE65682). Consensus clustering identified two distinct molecular subtypes, with the poorer-prognosis subtype enriched in inflammatory pathways. Based on differentially expressed genes, we constructed a 7-gene PRG-score (TUBG2, TNFAIP3, CXCL8, WFDC1, DEFA4, CX3CR1, and ZBP1) via LASSO and Cox regression. This model demonstrated good predictive performance for short-term mortality (AUCs > 0.75), which was further evaluated in an independent septic shock cohort (GSE95233, AUC = 0.984). Single-cell RNA sequencing analysis (GSE167363) showed that PRG-scores tended to remain elevated in non-survivors but decreased over time in survivors. Immune infiltration analysis indicated that higher PRG-scores were associated with features of immune remodeling, including increased M0/M2 macrophage proportions, reduced CD8 + T cells, and higher expression of immune checkpoints (e.g., CTLA4, TIGIT) and IL-10. In addition, a prognostic nomogram integrating the PRG-score and age improved individualized survival estimation. Overall, these findings suggest that the 7-gene PRG-score may reflect immune status and is associated with prognosis, providing insights into molecular subtyping and potential immunomodulatory strategies in sepsis.