Identification of the Origin of Eosinophils

Eosinophil granulocytes participate in parasite protection and also play a part in asthma and allergy. Histology staining shows a characteristic red granular stain, which separates the eosinophils from neutrophil and basophil granulocytes. Eosinophils can be characterized by SiglecF and CCL3 surface expression, and differentiate from bone-marrow CD34+ Granulocyte Monocyte Progenitors (GMP).  Isolating GMP and stimulating with IL-3, GM-CSF and IL-5 leads to eosinophil colony forming units (CFU) in methylcellulose. Eosinophil precursors are currently are isolated by IL-5ra expression in GMP and IL-5 is critical for eosinophil differentiation and expansion both in vivo and in vitro. Gata1 is a transcription factor necessary for erythrocyte and megakaryocyte formation. Previously, a Gata1-GFP transgenic reporter mouse was used to show that all eosinophil CFU capacity within the GMP came from those cells which induce Gata1 expression. Apart from that, little is known about the required transcriptional activity for eosinophil development. We performed single cell RNA seq using the Fluidigm C1 system on murine hematopoietic stem and progenitor cells to characterize myeloid differentiation and development. While primarily studying neutrophil granulocyte differentiation programs within the GMP, we noticed progenitors that expressed eosinophil genes. Gfi1 is a known granulocyte transcription factor, which is critical for neutrophil granulocytes in mice and man. We utilized a Gfi1-GFP knock in reporter mouse to exclude Gfi1-GFP^high GMP (which we find are neutrophil precursors). Unbiased hierarchical clustering of single-cell RNA Seq from Gfi1-GFP^low GMP identified four clusters. To identify uniquely expressed transcripts within each cluster, we utilized the MarkerFinder algorithm in a bioinformatics analysis platform called AltAnalyze (p<0.05 and fold >2). One cluster expressed transcripts like Fos, Jun, Gata1 and Gata2, indicating a stem-like progenitor. A second cluster, lost the stem-progenitor signature but expressed Elane, Mpo and Csf3r, suggesting a neutrophil progenitor signature. A third cluster lost Fos and Jun expression while gaining Il5ra, Epx and Prg2, suggesting that this cluster contained eosinophil-primed progenitors. Interestingly, although part of the stem-progenitor gene expression was lost, this cluster retained Gata1 expression.  Thus, we have identified rare Gata1-expressing GMP populations with either a stem-progenitor or an eosinophil gene expression program.  To biologically test whether Gfi1-GFP^low GMP have eosinophil potential, we performed CFU assays with or without IL-5 and then analyzed the colonies for eosinophil markers.  We found that IL-5 stimulation of Gfi1-GFP^low GMP led to a significant CFU-G eosinophil-like colony formation, whereas IL-5 stimulation of Gfi1-GFP^high GMP did not result in any eosinophil-like colonies. Flow cytometric analysis showed SiglecF and CCL3 surface expression in IL-5 treated Gfi1-GFP^low GMP. Furthermore, we observed increased IL-5ra, Epx and Prg2 transcript levels in IL-5 treated Gfi1-GFP^low GMP. Histology stain showed characteristic eosinophilic granules in the IL-5 treated condition. While earlier studies posited the induction of Gata1 expression in GMP as part of the program leading to eosinophil development, our data reveal a rare Gata1-expressing stem-progenitor population which gives rise to eosinophil progenitors as these cells differentiate along the granulocyte-monocyte pathway.  Thus, single-cell RNA Seq, combined with conventional methodology can be used to define previously rare and intermediary progenitor populations to understand hematopoietic development.