Distinct functional neutrophil phenotypes in sepsis patients correlate with disease severity

Abstract

Purpose: Sepsis is a clinical syndrome defined as life-threatening organ dysfunction caused by a dysregulated host response to infection. Sepsis is a highly heterogeneous syndrome with distinct phenotypes that impact immune function and response to infection. To develop targeted therapeutics, immunophenotyping is needed to identify distinct functional phenotypes of immune cells. In this study, we utilized our Organ-on-Chip assay to categorize sepsis patients into distinct phenotypes using patient data, neutrophil functional analysis, and proteomics.

Methods: Following informed consent, neutrophils and plasma were isolated from sepsis patients in the Temple University Hospital ICU (n=45) and healthy control donors (n=7). Human lung microvascular endothelial cells (HLMVEC) were cultured in the Organ-on-Chip and treated with buffer or cytomix ((TNF/IL-1β/IFNγ). Neutrophil adhesion and migration across HLMVEC in the Organ-on-Chip were used to categorize functional neutrophil phenotypes. Quantitative label-free global proteomics was performed on neutrophils to identify differentially expressed proteins. Plasma levels of sepsis biomarkers and neutrophil extracellular traps (NETs) were determined by ELISA.

Results: We identified three functional phenotypes in critically ill ICU sepsis patients based on ex vivo neutrophil adhesion and migration patterns. The phenotypes were classified as: Hyperimmune characterized by enhanced neutrophil adhesion and migration, Hypoimmune that was unresponsive to stimulation, and Hybrid with increased adhesion but blunted migration. These functional phenotypes were associated with distinct proteomic signatures and differentiated sepsis patients by important clinical parameters related to disease severity. The Hyperimmune group demonstrated higher oxygen requirements, increased mechanical ventilation, and longer ICU length of stay compared to the Hypoimmune and Hybrid groups. Patients with the Hyperimmune neutrophil phenotype had significantly increased circulating neutrophils and elevated plasma levels NETs.

Conclusion: Neutrophils and NETs play a critical role in vascular barrier dysfunction in sepsis and elevated NETs may be a key biomarker identifying the Hyperimmune group. Our results establish significant associations between specific neutrophil functional phenotypes and disease severity and identify important functional parameters in sepsis pathophysiology that may provide a new approach to classify sepsis patients for specific therapeutic interventions.

Keywords: Organ-on-Chip; neutrophil extracellular traps; neutrophil heterogeneity; proteomics; sepsis.

Figure 1

Organ-on-Chip Design (A) The image shows the vascular channel network and inlets and outlets with tubing inserted. (B) Bright field image shows the vascular channels and tissue compartment of the organ-on-chip. The vascular channels and tissue compartment are connected by 3 µm pores. (C) HLMVEC grow to confluent to cover the vascular channels. F-actin is labeled green with phalloidin and nuclei labeled blue with Hoechst 33342. (Scale bar = 100 µm). (D) Spatial variations in flow conditions in the vascular networks and at bifurcations showing shear rates in different vessel in the network in the organ-on-chip. Blue indicates a low shear rate and red a high shear rate. The effects of shear flow and vessel geometry on neutrophil adhesion and migration can be determined in this system.

Figure 2

Neutrophil adhesion and migration patterns identify distinct neutrophil functional phenotypes. (A) Discriminant analysis of neutrophil adhesion and migration patterns to identify different neutrophil functional phenotypes. Based on distinct adhesion and migration responses to cytomix activation, three different phenotypes were identified, and labeled as Hyperimmune, Hypoimmune and Hybrid. Continuous variables used for Discriminant Analysis include neutrophil adhesion ± cytomix at different shear rates (<15, 15-30, 30-60, 60-150 s^-1) and at bifurcations, and neutrophil migration ± cytomix over 60 min (10,15, 30, 60 min). (B) Hyperimmune: Representative functional response of Hyperimmune neutrophils to cytomix in the organ-on-chip demonstrating increased neutrophil adhesion in the vascular channels and increased migration across human endothelial cells into the Tissue Compartment, (C) Hypoimmune: Representative functional response of Hypoimmune neutrophils to cytomix in the organ-on-chip demonstrating decreased neutrophil adhesion in the vascular channels and decreased migration across human endothelial cells into the Tissue Compartment and (D) Hybrid: Representative functional response of Hybrid neutrophils to cytomix in the organ-on-chip demonstrating increased neutrophil adhesion in the vascular channels and decreased migration across human endothelial cells into the Tissue Compartment.

Figure 3

Neutrophil adhesion and migration in response to cytomix (A) Total neutrophil adhesion in organ-on-chip in the different sepsis neutrophil phenotypes and healthy controls in response to buffer or cytomix activation. (B) Adhesion of different neutrophil phenotypes at various shear rates and at bifurcations in the organ-on-chip. (C) Total neutrophil migration across HLMVEC into the tissue compartment following cytomix activation in response to fMLP. (D) The kinetics of sepsis patient neutrophil migration at different time points over a 60 min observation period. (A-D) Values are Mean ± SEM, *P<0.05, **P<0.01, ***P<0.001 by ANOVA. (E) Boxplot of time from sepsis diagnosis to blood draw in the Hyperimmune, Hypoimmune and Hybrid patient groups. P=0.29 by one way ANOVA.

Figure 4

The three sepsis patient phenotypes have unique proteomic signatures. (A) Heatmap of neutrophil proteins (B) Volcano Plot of Hypoimmune phenotype as compared to healthy controls, (C) Volcano Plot of Hyperimmune phenotype as compared to healthy controls, (D) Volcano Plot of Hybrid phenotype as compared to healthy controls. For all three phenotypes (3B-3D), the red dots represent upregulated proteins, the green dots represent downregulated proteins, and the gray dots are proteins that are not significantly different from healthy control samples.

Figure 5

Venn Diagrams of protein expression changes in the neutrophils from the three sepsis patient phenotypes as compared to controls. (A) Common and unique upregulated proteins in the neutrophils from the three sepsis patient phenotypes. (B) Common and unique downregulated proteins in the neutrophils from the three sepsis patient phenotypes.

Figure 6

Differential expression of neutrophil adherence proteins in sepsis phenotypes (A) A clustered heatmap showing expression of neutrophil adherence associated proteins in control, Hypoimmune, Hyperimmune and Hybrid groups. The color key at the top left indicates whether protein expression was greater or less than the mean. (B) Quantitation of the individual adherence associated proteins. (C) A clustered heatmap showing expression of neutrophil cytoskeleton associated proteins. (D) Quantitation of the individual cytoskeleton associated proteins. (E) A clustered heatmap showing expression of neutrophil defense associated proteins. (F) Quantitation of the individual defense associated proteins. Values are Mean ± SEM, *P<0.05, **P<0.01, ***P<0.001, n=4/group.

Figure 7

(A) Mosaic plot comparing the causes of respiratory failure in the Hypoimmune, Hyperimmune and Hybrid patient groups. (B) Plasma levels of MPO-DNA in healthy controls, Hypoimmune, Hyperimmune and Hybrid groups. Values are Mean ± SEM, (N= 8-23) **p<0.01.

Figure 8

Schematic Illustration: Three distinct neutrophil functional phenotypes were identified in ICU sepsis patients by organ-on-chip analysis of neutrophil ex vivo adhesion and migration patterns. These functional neutrophil phenotypes had distinct clinical correlates.