Identification, Isolation and Transcriptome Analyses of Mouse, Rat and Man Erythroblastic Island Central Macrophages

Erythroblastic island (EBI), composed of a central macrophage and surrounding erythroid cells, is the first hematopoietic niche discovered for erythropoiesis. Yet, the identity of the central macrophage has so far remained elusive. Based on the previous findings that F4/80, VCAM1 and CD169 are potential mouse central macrophage markers, we first calculated the number of F4/80⁺VCAM1⁺CD169⁺ mouse macrophages in the mouse bone marrow and compared it to the number of Ter119⁺ erythroblasts. We found that the ratio of F4/80⁺VCAM1⁺CD169⁺ macrophage and erythroblasts is about 1:2. Given the fact that one central macrophage is surrounded by multiple erythroblasts, the above finding suggests that it is unlikely that all the F4/80⁺VCAM1⁺CD169⁺ macrophages are central macrophages.

Erythropoietin (Epo) is essential for erythropoiesis. It has been reported that the Epo receptor (Epor) is expressed in peritoneal macrophages. These findings promoted us to speculate that EBI central macrophages may express Epor so that Epo acts on both erythroid cells and the central macrophages simultaneously in the niche to ensure efficient and optimal red cell production. To test this notion, we first examined whether mouse bone marrow and fetal liver macrophages express Epor using the Epor-GFPcre knockin mouse model. We found that ~5% of bone marrow F4/80⁺ macrophages and ~35% of fetal liver F4/80⁺ macrophages express Epor-GFP. As negative control, no Epor-GFP macrophages are noted in wild type F4/80⁺ macrophages. Importantly, ImageStream analyses revealed the native EBIs in bone marrow and fetal liver are formed by Epor⁺ but not Epor^- macrophages. Bioinformatics analyses of RNA-seq data on the sorted Epor⁺ and Epor^- macrophage populations revealed that molecules involved in central macrophage-erythroblast association such as VCAM1, CD169, and molecules known to be important for central macrophage function such as Dnase2a, ferroportin, are highly expressed in Epor⁺ macrophages. In marked contrast, highly expressed pathways in Epor^- macrophages are associated with immune responses including antigen process and presentation. Intriguingly, the immune related pathways are dramatically downregulated in the Epor⁺ macrophages, suggesting that the Epor⁺ macrophages in bone marrow and fetal liver have evolved a specialized function in supporting erythropoiesis.

To examine whether expression of Epor in EBI central macrophages is a conserved feature across species, we generated Epor-GFPcre knockin rat using the CRISP/Cas9 technology. Using CD163 as rat macrophage marker, we found that a subpopulation of rat bone marrow CD163⁺ macrophages expresses Epor-GFP. As a negative control, no Epor-GFP macrophages are noted in wild type CD163⁺ macrophages. To examine whether EPOR is expressed in human EBI central macrophages, antibody specificity for human EPOR is critical. To this end, we employed CRISP/Cas9 approach to knock out EPOR in K562 and Hela cell lines and validated the specificity of a commercially available anti-human EPOR antibody. Using CD163, CD169 as human macrophage markers, we found that EPOR is also expressed in a subpopulation of human macrophages. Moreover, in vitro EBI formation assay revealed that human EPOR⁺ but not EPOR^- macrophages form EBIs with erythroid cells and that the EBI formation is enhanced by EPO.

In summary, we for the first time, after discovery of the EBIs 60 years ago, have identified Epor⁺ macrophages in mouse bone marrow and fetal liver as EBI central macrophages. Our findings provide solid foundation for studying the mechanisms by which erythropoieis is supported EBI central macrophages. A better understanding of such mechanisms will provide extensive new knowledge on basic biology of erythropoiesis. It is also important to understand the pathology of erythropoietic disorders as well as to improve ex vivo erythrocyte production.