Background/Purpose: Neutrophils are key first responders of innate immunity, participating in both defense and inflammatory disease. Phenotypic heterogeneity is recognized but the existence of bone fide subsets remains unknown. We applied single-cell transcriptomics to define the relationships among individual neutrophils in heathy mice.

Methods: We profiled neutrophils from bone marrow, peripheral blood and spleen from 6-8 week male C57Bl/6 mice. Ly6g+ neutrophils were sorted for single-cell RNA sequencing using the 10X platform. We applied a multinomial model to the raw unique molecular identifier (UMI) counts to assign cells to one of 265 known cell populations and thereby ensure restriction of the analysis to neutrophils. Following exclusion of low-quality cells and transcripts expressed in fewer than 10 cells, 12,011 cells and 2,709 highly-expressed transcripts were retained for downstream analysis. We compared methods for dimensionality reduction, including PCA/tSNE and diffusion maps, to optimally reflect gene expression patterns.

Results: An average of 1,439 UMI and 482 genes were detected per cell. Linear dimensionality reduction could identify clusters of neutrophils, but underlying patterns were in fact best captured through diffusion maps, a nonlinear method. Using diffusion pseudotime analysis, we reconstructed a continuum of neutrophil maturation that defined a sequential gene expression program. The genes anchoring the two poles of this “neutrotime” spectrum were Ly6g, Camp, Ltf, Lcn2 in less mature cells and Il1b, Ccl6, Csf3r in more differentiated cells. Cells all along the neutrotime continuum were found in all profiled organs, but regional transcriptomic differences were also observed. For example, Il1b was highly expressed in blood and spleen but weakly in bone marrow. Conversely, Ly6g transcript abundance was highest in bone marrow and to a lesser extent in spleen, but only weakly detectable in blood. Interestingly, we identified a small cluster of cells defined by high expression of interferon response genes including Ifit1, Ifit3, Isg15, Ifitm3. This interferon-responding population was present in all profiled organs but less common in bone marrow than in blood and spleen, suggesting a more mature neutrophil population with a function yet to be defined.

Conclusion: Neutrophils have traditionally been characterized by nuclear morphology, cell density, and surface markers. Here we expanded this perspective via unbiased single-cell approach transcriptomics, demonstrating technical feasibility despite low transcript density. We identify a single developmental continuum shared by neutrophils across biological compartments, termed here ‘neutrotime’, that appears to represent much but potentially not all of their phenotypic heterogeneity. Our data define the baseline gene expression landscape of murine neutrophils and will facilitate further transcriptomic and functional investigations, as well as rational gene perturbations, potentially highlighting opportunities for therapeutic exploitation of the diversity within this important immune population.

« Back to 2019 ACR/ARP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/single-cell-rna-sequencing-of-murine-neutrophils-identifies-a-transcriptional-continuum-neutrotime-across-biological-compartments/