Both Primitive and Definitive Hematopoiesis Arise From Precursors with Endothelial Potential in the Zebrafish.

Abstract 697

Development of the blood system is complex, with multiple waves of hematopoietic precursors arising in different embryonic locations. Monopotent, or primitive, precursors first give rise to embryonic macrophages or erythrocytes. Multipotent, or definitive, precursors are subsequently generated to produce the adult-type lineages. In both the zebrafish and mouse, the first definitive precursors are committed erythromyeloid progenitors (EMPs) that lack lymphoid potential. We have previously shown that zebrafish EMPs arise in the posterior blood island independently from hematopoietic stem cells (HSCs) that arise along the ventral side of the aorta. By using the CRE/LOX technology, recent reports have nicely showed that vascular endothelial cells on the ventral side of the aorta were giving rise to HSCs in the mouse embryo.

Here we show similar results in the zebrafish, that HSCs are also born from hemogenic endothelial cells (ECs). We used the kdrl/flk1:CRE transgenic line. Unlike in mammals, the zebrafish flk1 promoter used here is not pan-mesodermic but restricted to ECs and their precursors, so we could specifically trace the progeny of ECs in the zebrafish by crossing this CRE line to a switch-reporter line. This latter consists of the b-actin promoter driving the expression of DsRED after a STOP cassette is excised by the CRE recombinase. Double transgenic adults were sacrificed at 3 months, and in all animals (n=30), 100% of adult leukocytes were switched (DsRED+). Thus, zebrafish HSCs originate from ECs in the embryo. To corroborate these findings, we could isolate cells during their transition from EC to hematopoietic cells using the flk1:RFP and cmyb:GFP reporter lines, and show that, during the transition, the endothelial genetic program is shut down, while hematopoietic genes are turned on. Moreover, by using the transparency of the zebrafish embryo, we could directly image the birth of HSCs from the ventral floor of the aorta.

As stated above, three other hematopoietic lineages arise during embryogenesis: primitive macrophages/erythrocytes and definitive EMPs. Whether or not these other lineages arise from hemogenic ECs is still unknown. In order to address this question, we generated triple transgenic animals (flk1:CRE; b-actin:switch; lineage:GFP), where the lineage specific promoter was either lysozymeC or cd41, in order to mark primitive macrophages (lysC+ at 24hpf) and EMPs (cd41+ at 26-30hpf), respectively. In all combinations studied, GFP-positive cells were always also expressing DsRED from the switch reporter, revealing that primitive macrophages and EMPs were also derived from flk1+ cells in the embryo. Primitive erythrocytes could not be traced as well, since the b-actin promoter driving the switch reporter is silenced during erythroid commitment. However, flk1:eGFP+ cells could be sorted as early as 8-10 somites during zebrafish development, and we could show that only this fraction in the embryo was expressing the transcripts for gata1 and hbae3/hbae1 (embryonic hemoglobins). Thus we conclude that primitive erythroblasts originate from a flk1-expressing subset of cells in the embryo.

As previously described, HSCs do arise from hemogenic ECs in the aorta, which we can now directly visualize. Primitive macrophages and erythrocytes also derive from flk1-expressing cells in the early embryo, which are likely to be hemangioblasts since vasculogenesis has not occurred yet. Finally, our data do not allow to precisely define the origin of EMPs, which could be attributed to either a caudal hemangioblast or a hemogenic endothelium. Altogether, these results should give new insights on the generation of blood cells in the vertebrate embryo.