Proteomic Characterization of 1000 Human and Murine Neutrophils Freshly Isolated From Blood and Sites of Sterile Inflammation

Abstract

Neutrophils are indispensable for defense against pathogens. Injured tissue-infiltrated neutrophils can establish a niche of chronic inflammation and promote degeneration. Studies investigated transcriptome of single-infiltrated neutrophils which could misinterpret molecular states of these post mitotic cells. However, neutrophil proteome characterization has been challenging due to low harvests from affected tissues. Here, we present a workflow to obtain proteome of 1000 murine and human tissue-infiltrated neutrophils. We generated spectral libraries containing ∼6200 mouse and ∼5300 human proteins from circulating neutrophils. 4800 mouse and 3400 human proteins were recovered from 1000 cells with 10²-10⁸ copies/cell. Neutrophils from stroke-affected mouse brains adapted to the glucose-deprived environment with increased mitochondrial activity and ROS-production, while cells invading inflamed human oral cavities increased phagocytosis and granule release. We provide an extensive protein repository for resting human and mouse neutrophils, identify proteins lost in low input samples, thus enabling the proteomic characterization of limited tissue-infiltrated neutrophils.

Keywords: circulation; inflammation; low input proteomics; neutrophils; stroke; transmigration.

Graphical abstract

Fig. 1

Workflow for analyzing the proteome of murine and human neutrophils isolated from blood and sterile inflammatory sites. The sample cohort included healthy human and naive mouse blood neutrophils with five biological replicates that were utilized for the establishment of a 1000-cell proteome method. Neutrophils isolated from mouse ischemic brain and the inflamed human oral cavity were used for the validation of established technologies.

Fig. 2

Protein profiling of 1000 neutrophils isolated from human and mouse blood (n = 5). A and B, comparison of proteins and peptides identified from a 1000 cell digest, 1000 cell equivalent (eq.), and 4000 cell eq. 1000 and 4000 cell equivalents were derived from standard digests (100,000 cells). Data are represented as mean ± SD. Statistical significance was determined by paired t test, where ∗∗p < 0.01, ∗∗∗p < 0.001. C and D, the correlation among biological replicates of 1000 cells based on protein quantity. E and F, proteome correlation between 1000 cells and 1000 cell eq. This includes the proteins present in both conditions. G and H, Venn diagram showing the overlap of proteins identified from different conditions.

Fig. 3

Gene Ontology analysis of neutrophil proteomes.A, biological Processes and (B) cellular components that passed the Benjamini–Hochberg (32) adjusted p-value cut-off (<0.05) were ranked by gene count. The top pathways in each or both organisms are shown as a function of gene count.

Fig. 4

Comparison of the protein repertoire between human and mouse neutrophils.A, protein and peptide identifications obtained from different cell quantities in both organisms. Data represent the mean ± s.d. of biological replicates (n = 5). ∗∗∗∗p < 0.0001 (two-tailed t test). B, Venn diagram illustrating shared proteins between two species, identified from the proteomics data in (A). C, reactome pathways associated with proteins identified in both organisms and uniquely detected in one organism (from panel B). Only pathways with Benjamini–Hochberg (32) adjusted p-values <0.05 are shown. D, expression pattern of neutrophil-lineage proteins identified in both human and mouse neutrophils compared with previously reported transcriptomics data (7). The upper half shows the proteins and transcripts detected in both organisms by both methods. The lower half highlights the proteins identified in both organisms but transcripts only in one organism. E, neutrophil-lineage proteins uniquely identified in human and mouse in accordance with transcriptomics data.

Fig. 5

Estimation of protein copy numbers in human and mouse neutrophils.A and B, dynamic range of protein copy numbers detected in 4000 and 1000 neutrophils. Highlighted proteins are identified in both quantities (green area) or not identified in 1000 cells (blue area). C and D, consistency of quantified protein copies across various sample quantities as indicated. Only proteins common to all three conditions were considered for the analysis. E, resemblance between the current study and a published repository (36) in protein copy number estimation of human neutrophils. F, correlation between human and mouse neutrophils based on estimated copy numbers of orthologous proteins. G and H, concordance between the quantitative flow cytometry (qFlow) and proteomics approaches in quantifying copies of surface proteins. The average of all measurements is plotted. Repository data (36, 37) were included to validate the trend.

Fig. 6

Differential expression analysis between neutrophils isolated from murine ischemic brain and the circulation (n = 3) 24 h after tMCAO.A, experimental approach of stroke induction and cell isolation. B, volcano plot indicating proteins with significantly differential expression (Log₂ fold change ≥ ±1, paired t test, p-value <0.05) Red: proteins upregulated in brain neutrophils; blue: proteins downregulated in brain neutrophils. C, expression pattern of specific proteins involved in maturation in the different neutrophil populations. D, overexpression of mitochondrial proteins in brain-infiltrated neutrophils 24 h after stroke. E, biological processes (Benjamini–Hochberg (32) adjusted p-value <0.05) of significantly dysregulated proteins in two populations obtained from (B).

Fig. 7

Proteome comparison between circulatory neutrophils and neutrophils transmigrated to the oral cavity (lavage) after tabasco treatment.A, experimental approach of tabasco treatment and cell isolation from healthy human individuals (n = 7). B, volcano plot highlighting upregulated (red) and downregulated (blue) proteins in transmigrated neutrophils isolated from lavage. Statistical significance was determined by Paired t test with p-value <0.05, Log₂ fold change ≥ ±1. C, paired analysis in surface markers expression in neutrophils obtained from blood and lavage samples. Analyses were made by our proteomics workflow (top) and spectral flow cytometry (bottom). Paired t test was used for the expression comparison (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001). D, biological processes (Benjamini–Hochberg (32) adjusted p-value <0.05) associated with proteins upregulated either in transmigrated neutrophils or in circulatory neutrophils.