Abstract
Protocols for the isolation of neutrophils from whole blood often result in neutrophil preparations containing low numbers (~5%) of contaminating leukocytes, and it is possible that these contaminating cells contribute to highly sensitive assays that measure neutrophil gene expression (e.g. qPCR). We investigated the contribution of contaminating leukocytes on the transcriptome profile of human neutrophils following stimulation with inflammatory cytokines (GM-CSF, TNFα), using RNA-Seq. Neutrophils were isolated using Polymorphprep or the StemCell untouched neutrophil isolation kit (negative selection of "highly pure" neutrophils). The level of contamination was assessed by morphology and flow cytometry. The major source of contamination in Polymorphprep neutrophil preparations was from eosinophils and was highly donor dependent. Contaminating cells were largely, but not completely, absent in neutrophil suspensions prepared using negative selection, but the overall yield of neutrophils was decreased by around 50%. RNA-seq analysis identified only 25 genes that were significantly differentially-expressed between Polymorphprep and negatively-selected neutrophils across all three treatment groups (untreated, GM-CSF, TNFα). The expression levels of 34 cytokines/chemokines both before and after GM-CSF or TNFα treatment were not significantly different between neutrophil isolation methods and therefore not affected by contributions from non-neutrophil cell types. This work demonstrates that low numbers (<5%) of contaminating leukocytes in neutrophil preparations contribute very little to the overall gene expression profile of cytokine-stimulated neutrophils, and that protocols for the isolation of highly pure neutrophils result in significantly lower yields of cells which may hinder investigations where large numbers of cells are required or where volumes of blood are limited.