[Identification of immune key genes in sepsis and the regulatory effects of dexmedetomidine based on transcriptome analysis and machine learning].

Objective This study aims to identify immune-related key genes in sepsis using transcriptomic analysis and machine learning, and to explore the role of dexmedetomidine (DEX) in sepsis regulation through these genes. Methods The transcriptomic dataset GSE9960 related to sepsis was obtained from the gene expression omnibus (GEO) database. Differential expression analysis and visualization were performed using the limma and ggplot2 packages in R 4.5.1. DEX targets were collected from the comparative toxicogenomics database (CTD), SwissTargetPrediction, and GeneCards databases, followed by integration and removal of duplicates to obtain the overall DEX target set. Immune-related targets were also retrieved from the GeneCards database. The DEX targets, sepsis-related targets, and immune-related targets were intersected to identify DEX-targeted immune-related sepsis targets. Key targets were screened using least absolute shrinkage and selection operator (LASSO) regression and random forest (RF) algorithms, and the overlapping targets were further identified as core DEX-targeted immune targets in sepsis. Immune infiltration was evaluated using the CIBERSORT algorithm combined with Spearman correlation analysis, and the correlations between core targets and immune cell populations were further investigated. In vitro, a sepsis cell model was established by lipopolysaccharide (LPS) stimulation of the human monocyte cell line THP-1. Target gene knockdown or overexpression was achieved through gene transfection. Western blot analysis was performed to evaluate the effects of DEX treatment and potassium calcium-activated channel subfamily M α1 (KCNMA1) regulation on sepsis-related proteins and inflammatory cytokine levels. Results Sepsis-related dataset GSE9960 exhibited significant differential expression. After integrating and removing duplicates among sepsis-, immune-, and DEX-related targets, 25 key immune targets of DEX against sepsis were identified. LASSO and RF analyses further screened three core targets. Immune infiltration analysis showed that these core targets were significantly associated with plasma cells, neutrophils, and resting memory CD4+ T cells. Since KCNMA1 exhibited the most prominent upregulation and the greatest differential expression, it was selected for further experimental validation. The results demonstrated that DEX treatment reduced the expression of sepsis-related proteins, including KCNMA1, cyclin B2 (CCNB2), and NEDD4 E3 ubiquitin protein ligase (NEDD4), and suppressed inflammatory cytokine levels by downregulating KCNMA1. Conclusion DEX suppresses the development of sepsis by downregulating KCNMA1 expression.