Disclosure(s): No relevant financial relationship(s) to disclose.
First Author: Xue Zhang, N/A Co-Author: Haiqing Shi, N/A – Doctor, Department of Critical Care Medicine, West China School of Medicine, West China Hospital, Sichuan University Co-Author: Xuelian Liao, N/A – Doctor, Department of Critical Care Medicine, West China School of Medicine, West China Hospital, Sichuan University
Introduction: Sepsis remains a critical clinical challenge, driven by immune dysregulation and high mortality. Macrophage metabolism profoundly shapes immune responses, yet reliable metabolic biomarkers for diagnosis and targeted intervention are lacking.
Methods: Transcriptomic data from three PBMC cohorts (GSE236713, GSE134347, GSE100150) were integrated for discovery analysis. GSE154918 served as an independent public validation dataset, and a clinical PBMC cohort with mortality data (Huaxi cohort) was prospectively collected from our center. Candidate genes were screened by intersecting differentially expressed genes, WGCNA modules, and macrophage metabolism-related databases. Hub genes were identified using LASSO, random forest, and SVM-RFE algorithms. A diagnostic model and nomogram were constructed and validated via ROC analysis. Immune cell infiltration was assessed using CIBERSORT and ssGSEA. Consensus clustering defined immune subtypes, and their mortality differences were evaluated. Expression of hub genes was validated in external datasets and single-cell profiles, and correlated with clinical indicators. Molecular docking explored interaction between IL2RB and glucomannan. A CLP mouse model was used for qPCR validation.
Results: IL2RB, FLT3LG, and SLC38A1 were significantly downregulated in sepsis, especially in non-survivors. The diagnostic model based on these genes showed high accuracy (AUCs: 0.960–0.982). Subtype C1, with higher hub gene expression, had a lower mortality trend than C2 (44.6% vs. 71.4%, p = 0.0714). Gene expression correlated with lactate, procalcitonin, IL-6, and immune cell subsets (CD4⁺, CD8⁺ T cells, memory T cells, monocytes), supported by both computational deconvolution and clinical immune profiling. Docking analysis showed strong affinity between IL2RB and glucomannan (−8.5 kcal/mol). In septic mice, qPCR confirmed increased HIF1A, STAT5, and LDHA expression in lung tissues, indicating activation of glycolysis, HIF-1α, and JAK-STAT pathways.
Conclusions: IL2RB, FLT3LG, and SLC38A1 may serve as metabolic–immune biomarkers for sepsis diagnosis and stratification. Their association with clinical severity and mortality, and the regulatory potential of glucomannan via IL2RB, highlight new avenues for precision immunometabolic therapy.
