Definitive EMP and Pre-HSC Emerge in Myb-Null Murine Embryos and Retain Macrophage Potential

Myb-null murine embryos lack definitive erythropoiesis but can produce primitive erythroid cells, allowing their survival to embryonic day 15 (E15) (Mucenski et al., Cell, 1991). Myb expression has been used to detect emerging HSC and is required for HSC maintenance (Lieu et al., PNAS 2009; North et al., Cell 2009). These data have led to the model that myb function is required for definitive hematopoiesis. Interestingly, macrophages and megakaryocytes are still detected in myb-null embryos and, as both of these lineages are also components of primitive hematopoiesis, it is proposed that these cells are not definitive in origin. Myb-independent macrophages infiltrate fetal tissues and have been implicated as a self-renewing source for several adult tissue-resident macrophage populations (Schulz et al., Nature 2012; Gomez Perdiguero et al., Glia 2013; Hoeffel et al., Immunity 2015). We tested the hypothesis that definitive hematopoiesis is entirely myb-dependent by examining two distinct sources of definitive erythroid/myeloid potential: 1) HSCs that emerge from hemogenic endothelium, including the AGM region at E10.5 in mice and 2) HSC-independent definitive EMP that emerge after primitive hematopoiesis from yolk sac hemogenic endothelium beginning at E8.25 (Frame et al., Stem Cells 2016). By E9.5, EMP can be prospectively isolated based on immunophenotype and contain all the erythroid/myeloid progenitor activity present at this time (McGrath et al., Cell Reports 2015). Surprisingly, we found normal numbers of immunophenotypic EMP in E9.5 Myb-null yolk sacs, and immunohistochemical analysis confirmed their emergence from Runx1+ hemogenic endothelium. At E10.5, reduced numbers of Myb-null EMP were found not only in the yolk sac, but also in the bloodstream and the liver. This decrease correlated with fewer hemogenic clusters in the yolk sacs of Myb-null embryos, as well as alterations in their cell-cycle status. The presence of significant numbers of immunophenotypic EMP suggests they could serve as an alternate source of Myb-null macrophages. Clonal analysis of sorted EMP confirmed that Myb function is necessary for erythroid and granulocyte progenitors, but Myb-null EMP retain normal plating efficiencies for macrophage progenitors. Indeed, Myb-null EMP generate only macrophages in liquid culture, lacking not only erythroid and granulocyte cells, but also Ly6C+ monocytes. Consistent with these results, RNA sequencing analysis of Myb-null versus wildtype EMP demonstrated decreased expression of genes in pathways associated with cellular growth, as well as erythroid and granulocyte fates. We further determined that Myb is not required for the emergence of immunophenotypic pre-HSC in the AGM region of E10.5 embryos. In addition, there were normal numbers of clusters in the aorta of E10.5 myb-null embryos. We were also able to detect lineage-, Kit+ (LK) cells in E14.5 livers as previously reported (Sumner et al., Oncogene 2000). While LK numbers were reduced, the Sca1+ (LSK) subset was present in normal numbers. Like EMP, sorted Myb-null E10.5 pre-HSC, as well as E14.5 liver LK and LSK, lacked erythroid or granulocyte CFC activity, but retained normal CFC-M plating efficiencies. In addition, Ly6C+ monocytes were not observed in liquid cultures of sorted Myb-null E10.5 pre-HSC, which produced only macrophages in vitro, or in Myb-null E14.5 livers. Together these data indicate that Myb is not required for hematopoietic emergence of definitive EMP or HSC, but does facilitate the expansion of these definitive stem/progenitor cells and is required for erythroid and granulocyte differentiation. Additionally, EMP and HSC contain Myb-independent macrophage potential, which does not appear to differentiate from a monocyte intermediate.