Extensively Self-Renewing Erythroid Precursors Emerge from the Embryo but Not from the Adult

The only cell in the hematopoietic hierarchy thought to be capable of long-term selfrenewal is the stem cell. An erythroid progenitor derived from mammalian hematopoietic tissue, fetal or adult, is capable of limited proliferation (103–106 fold expansion; Bauer, 1999; Panzenböck, 1998; von Lindern, 1999). Here we report that an erythroid precursor derived from the mouse embryo is capable not only of limited, but also of extensive proliferation (~1030 fold expansion). These cells resemble proerythroblasts and basophilic erythroblasts based on their morphology, their globin gene expression profile, and their immunophenotype. While aneuploidy is not necessary for extensive proliferation, it sporadically begins to accumulate after prolonged culture. These cells are capable of massive (>100 days) daily proliferation in vitro in the presence of Epo, SCF, IGF-1, and dexamethasone. Examination of cultures lacking each of these factors support that glucocorticoids play an important role in this expansion by uncoupling erythroid precursor proliferation from maturation. Despite prolonged in vitro culture, these cells preserve their potential to fully differentiate into enucleated red blood cells with the removal of dexamethasone. Differentiation occurs over 2–3 days and is characterized by the accumulation of adult (α, β1, and β2), but not embryonic (ζ, εy, and βH1), globins. The retention of full differentiation potential despite >1030 fold expansion indicates that this proliferation represents self-renewal. To determine the developmental origin of these extensively self-renewing erythroblasts (ESREs), we initiated in vitro cultures from staged mouse embryos as well as adult tissues. E7.5 embryos, that contain primitive but not definitive erythroid progenitors, failed to generate ESREs. In contrast, ESREs can be derived from E8.5–E10.5 yolk sac and E11.5–E14.5 fetal liver. These findings along with the globin expression pattern indicate that erythroblast self-renewal is associated with definitive, but not primitive, erythropoiesis. Surprisingly, marrow from adult steady-state hematopoiesis failed to yield ESREs. Furthermore, despite the characteristics shared by stress erythropoiesis (adult spleen) and fetal erythropoiesis (liver), stress erythropoiesis only yielded erythroblasts with limited, and not extensive, self-renewal capacity. This result suggests that extensive self-renewal potential is linked either to the transient yolk sac-derived definitive erythroid lineage or to the fetal hematopoietic microenvironment. We are currently investigating the mechanisms responsible for the extensive self-renewal capacity of such lineage-restricted and mature hematopoietic precursors. Our findings raise the possibility that the expansive cellular output of the erythron within the midgestation mammalian embryo may be regulated, in part, at the level of late stage erythroid precursors.