Single-cell genomic profiling to identify immune signatures of bacterial sepsis in humans

Sepsis, life-threatening organ dysfunction due to a dysregulated host response to infection, is prevalent and highly lethal. Early detection is a major challenge, largely because, as we and others showed, sepsis is a heterogeneous syndrome, with many patients presenting with vague symptoms and signs; improved biomarkers

would enable earlier diagnosis, especially of these patients. A second major challenge is the shortage of treatments, as mechanisms of immune dysfunction and vascular leakage in sepsis are poorly understood, limiting therapeutic development. We hypothesize that unbiased single-cell transcriptional profiling (scRNA-seq)

of circulating immune cells will identify transcriptional signatures that address both challenges. In our proof-of- principle study (Reyes et al. Nat Med 2020), we discovered a unique monocyte cell state (MS1) that is expanded in patients with urosepsis and absent in patients with milder urinary tract infection or healthy controls. MS1

discriminates septic patients from patients with other diseases in public transcriptomic data, in a second cohort of bacterial sepsis, and in preliminary data from our new cohort of patients with sepsis from any body site. Our goals in this proposal are to test the hypothesis that when extended to patients with sepsis from all body sites,

unbiased single-cell approaches will generate immune signatures that distinguish sepsis from non-infectious organ dysfunction, define new clinically-relevant sepsis endotypes, identify markers that enhance sepsis diagnostics and subtyping, and reveal critical mechanistic biology underlying the immune dysfunction and

vascular leak present in sepsis, thus facilitating future development of endotype-specific therapeutics. In this proposal, we will test our hypothesis that among patients presenting to the ED with acute organ dysfunction, transcriptional signatures can be identified that distinguish those with sepsis from those with non-infectious

etiologies. We propose to: (1) discover blood single-cell transcriptional signatures that discriminate sepsis from non-infectious organ dysfunction, define unbiased molecular endotypes, and associate with clinical outcomes (Aim 1); (2) identify cell surface markers associated with scRNA-seq-defined cell states, including MS1, that are

significantly expanded in sepsis (Aim 2); (3) define alterations in cellular functions in patients and mice with sepsis and in response to sepsis-induced MS1 cells (Aim 3). The proposed studies are highly likely to lead to substantially improved cellular and molecular signatures for sepsis that could be translated into clinical use and

to new insights into the nature of immune dysregulation in sepsis. Investigations into the function of genes and pathways identified in our studies will impact mechanistic understanding of sepsis and may lead to new therapeutic concepts, especially for subsets of sepsis patients stratified based on single cell-derived molecular

endotypes. Whereas sepsis remains among of the most challenging problems in human health, our use of a high-resolution unbiased approaches to address the major questions of heterogeneity and mechanism will provide highly useful datasets for the field and promote the development of new targeted therapeutics.