A Novel Single Cell RNA-seq Analysis of Non-Myeloid Circulating Cells in Late Sepsis

Abstract

Background: With the successful implementation of the Surviving Sepsis Campaign guidelines, post-sepsis in-hospital mortality to sepsis continues to decrease. Those who acutely survive surgical sepsis will either rapidly recover or develop a chronic critical illness (CCI). CCI is associated with adverse long-term outcomes and 1-year mortality. Although the pathobiology of CCI remains undefined, emerging evidence suggests a post-sepsis state of pathologic myeloid activation, inducing suboptimal lymphopoiesis and erythropoiesis, as well as downstream leukocyte dysfunction. Our goal was to use single-cell RNA sequencing (scRNA-seq) to perform a detailed transcriptomic analysis of lymphoid-derived leukocytes to better understand the pathology of late sepsis.

Methods: A mixture of whole blood myeloid-enriched and Ficoll-enriched peripheral blood mononuclear cells from four late septic patients (post-sepsis day 14-21) and five healthy subjects underwent Cellular Indexing of Transcriptomes and Epitopes by Sequencing (CITE-seq).

Results: We identified unique transcriptomic patterns for multiple circulating immune cell subtypes, including B- and CD4⁺, CD8⁺, activated CD4⁺ and activated CD8⁺ T-lymphocytes, as well as natural killer (NK), NKT, and plasmacytoid dendritic cells in late sepsis patients. Analysis demonstrated that the circulating lymphoid cells maintained a transcriptome reflecting immunosuppression and low-grade inflammation. We also identified transcriptomic differences between patients with bacterial versus fungal sepsis, such as greater expression of cytotoxic genes among CD8⁺ T-lymphocytes in late bacterial sepsis.

Conclusion: Circulating non-myeloid cells display a unique transcriptomic pattern late after sepsis. Non-myeloid leukocytes in particular reveal a host endotype of inflammation, immunosuppression, and dysfunction, suggesting a role for precision medicine-guided immunomodulatory therapy.

Keywords: chronic critical illness; immune cells; lymphocytes; scRNA-seq; sepsis; transcriptome.

Figure 1

scRNA-seq analysis at 14-21 days post-sepsis versus healthy control. Cells depicted are from all subjects in the study, in each corresponding group (sepsis n=4, healthy n=5). Using Seurat’s method of integrating data across conditions/batches, the integration allows for joint clustering and to identify shared (or possibly unshared) cell clusters. Cells are visualized on uniform manifold approximation and projection (UMAP) plots colored by cell types. (A) UMAP representation of cell clusters identified in healthy patients versus late sepsis. (B) Annotation of T-lymphocyte subsets was performed manually using expression of CD3D, CD4, CD8A, CCL5, CCR7, FOXP3, GNLY, IL2RA, IL7R, NCAM1 and NKG7. (C) UMAP representation of T-cell subset clusters from manual annotation identified in healthy patients versus late sepsis in three dimensions. (E-MDSC, early myeloid derived suppressor cell; G-MDSC, granulocytic myeloid derived suppressor cell; M-MDSC, monocytic myeloid derived suppressor cell; pDC, plasmacytoid dendritic cells).