Abstract
BACKGROUND: Bloodstream infections (BSIs) caused by gram-positive cocci (GPC) and gram-negative bacilli (GNB) are major causes of sepsis. However, their distinct effects on host responses remain poorly characterized at the single-cell level. This study used single-cell transcriptomics to define pathogen-specific monocyte heterogeneity in BSIs to identify the mechanisms underlying clinical differences. METHODS: Single-cell RNA sequencing (scRNA-seq) was performed on peripheral blood mononuclear cells obtained from healthy volunteers, two patients with GNB-BSI sepsis, and two patients with GPC-BSI sepsis. Differential gene expression, particularly in monocytes, was analyzed. The key findings were validated with clinical characteristics and outcomes of 45 patients with GNB-BSI sepsis and 40 patients with GPC-BSI sepsis. The distinguishing performances of identified biomarkers were evaluated via receiver operating characteristic (ROC) curve. RESULTS: In pathogen-specific transcriptomes, 54 identified genes were significantly associated with GNB-BSI (upregulated genes enriched in inflammatory pathways and downregulated genes enriched in oxidative phosphorylation). Twenty-one identified genes were associated with GPC-BSI (downregulated genes associated with cell adhesion molecules and upregulated genes involved in PI3K-Akt signaling). Nineteen genes were common to both groups, with distinct pathogen sensitivities. Patients with GNB-BSI presented with significantly greater disease severity, systemic inflammation and lymphopenia than patients with GPC-BSI. Conversely, patients with GPC-BSI had higher S100A12 and globulin levels and platelet counts. The combination of S100A12(high) and procalcitonin (PCT)(low) discriminated GPC-BSI from GNB-BSI (area under the curve=0.882, sensitivity 75%, specificity 91%; cutoff value 0.56). CONCLUSION: ScRNA-seq reveals the heterogeneity of GPC-BSI and GNB-BSI. Compared with GPC-BSI, GNB-BSI causes severe inflammation and metabolic suppression, which are associated with poor outcomes. The S100A12(high)+PCT(low) combination may have potential to discriminate among the major causes of BSI.