Deciphering the role of pyroptosis-related genes and natural killer T cells in sepsis pathogenesis: a comprehensive bioinformatics and Mendelian randomization analysis.

Pyroptosis is increasingly recognized as crucial in sepsis development, but the specific roles of pyroptosis-related genes (PRGs) in sepsis remain underexplored. Gene expression profiles of sepsis and control samples were retrieved from the Gene Expression Omnibus (GEO) database for analysis (GSE57065, GSE95233). Differentially expressed genes (DEGs) were identified, followed by functional enrichment analysis. Weighted gene co-expression network analysis (WGCNA) was employed to identify genes associated with sepsis, with intersecting DEGs and PRGs highlighted via Venn diagrams. Hub genes were further analyzed across both the training and validation datasets (GSE65682) for differential expression, receiver operating characteristic (ROC) analysis, correlation analysis, and Kaplan-Meier (KM) survival analysis. Immune cell infiltration was evaluated in both datasets using the single-sample gene set enrichment analysis (ssGSEA) algorithm. Machine learning approaches were applied to identify critical immune cell types involved in sepsis regulation, which were subsequently correlated with the hub genes. Single-cell RNA sequencing (scRNA-seq) analysis of sepsis samples was conducted using the GSE167363 dataset. Finally, Mendelian randomization (MR) was utilized to investigate causal relationships between exposures and outcomes. In results eight hub PRGs were identified, including NLRC4, PLCG1, TP53, AIM2, GZMB, GZMA, ELANE, and CASP5. Functional enrichment analysis implicated dysregulated immune responses in sepsis progression, aligning with established pathophysiological mechanisms. These eight key genes exhibited consistent expression patterns. Several genes (NLRC4, PLCG1, AIM2, GZMB, and ELANE) emerged as promising diagnostic biomarkers (AUC>0.85). Machine learning revealed that 15 immune cell types may play important roles in sepsis. Correlation analysis indicated a positive relationship between granzyme B (GZMB) and natural killer T (NKT) cells, a finding further corroborated by scRNA-seq analysis. In the validation cohort, GZMB and ELANE were linked to patient prognosis (p<0.05). MR analysis using the inverse variance weighting (IVW) method demonstrated a positive causal relationship between GZMB and NKT cells (OR=1.063, 95% CI=1.013-1.115, p=0.013). Moreover, elevated NKT cell levels were associated with a reduced risk of sepsis (OR=0.977, 95% CI=0.955-1.000, p=0.046), and NKT cells served as protective factors for 28-day mortality in sepsis (OR=0.938, 95% CI=0.881-0.997, p=0.040). This study provides a comprehensive analysis of the roles of PRGs and NKT cells in sepsis, offering valuable insights for diagnostic and therapeutic approaches in sepsis immunotherapy.