The hemangioblast revisited

In this issue of Blood, Padrón-Barthe et al explore the role of the hemangioblast as the cell of origin for yolk sac blood and endothelium.¹ 

Although recent years have seen a significant increase in our understanding of the complex onset of hematopoiesis in the embryo,²  the precise cellular origin of blood cells remains associated with recurrent controversies: is the adult mammalian hematopoietic system ultimately derived from the dorsal aorta or the yolk sac? Is blood derived from a hemangioblast, hemogenic endothelium, or a combination of both? Clarifying such controversies will be important for informing studies aimed at de novo generation of blood cells for clinical purposes. Through in vivo clonal analysis of early mouse hematopoiesis, Padrón-Barthe and colleagues find little evidence for the hemangioblast as the common cell of origin for both blood and endothelium.¹  Their results provide important in vivo support for previous ex vivo fate mapping studies in the mouse^3,4  and prompt us to revisit the concept of the hemangioblast.

A common origin of blood and endothelium was first proposed at the beginning of the last century. Building on the work of Sabin, Maximov, and others, it was Murray who in 1932 coined the term “hemangioblast” to indicate the thickenings of the mesoderm in the chick yolk sac, the mesodermal “masses” located at the sites where later the blood islands emerge.⁵  It is worth noting that Murray defines the hemangioblast as a population of cells, which as a whole gives rise to the blood and endothelium of the blood islands. At the time, he could not say whether extrinsic cues determine the fate of individual hemangioblast cells or whether inherently different precursors for the 2 lineages coexist within his hemangioblast. Fast forward to the 1990s when a clonal mesodermal precursor for blood and endothelium was identified in embryonic stem (ES) cell differentiation cultures,⁶  the blast colony-forming cell (BL-CFC). This cell was thought to be the in vitro equivalent of the in vivo hemangioblast, implying a conceptual change, where “hemangioblast” no longer represents a population of cells, but a clonal bipotent progenitor to blood and endothelium. However, in contrast to the original hemangioblast that had its physical presence in the mesodermal masses of the prospective blood islands, the in vivo site of residence of this clonal hemangioblast remained uncertain. It was not until the identification of low numbers of BL-CFC in the primitive streak of the E7.5 mouse embryo⁷  that this newly defined hemangioblast found its place. The BL-CFC was reported to give rise to primitive erythrocytes, as well as other hematopoietic progenitors in culture. More recently, a closer examination of ES cultures placed the BL-CFC and a putative in vitro equivalent of hemogenic endothelium in 1 linear pathway,⁸  lending support for a model of developmental hematopoiesis in which the hemangioblast gives rise to blood cells through a hemogenic endothelial intermediate.⁹  As, so far, no hematopoietic stem cells have been derived from ES cell cultures, this model has been restricted to yolk sac hematopoiesis.

In their current work, Padrón-Barthe and colleagues assess the in vivo relevance of the clonal hemangioblast and the hemogenic endothelium to yolk sac hematopoiesis. They use an inducible Cre mouse model where Cre is expressed from the ubiquitous polII locus. Injection of a low dose of 4-hydroxy-tamoxifen early in gestation results in the random genetic labeling of single cells of (pre)gastrulation stage embryos. These cells grow out to clones, and labeled clones contributing to the yolk sac blood islands are analyzed for the presence of blood and/or endothelial cells. The strength of this genetic labeling approach over previous mouse studies^3,4  is that it is clonal, does not require embryo manipulation, and, importantly, labeling and tracing occur entirely in vivo. Remarkably, the majority of the labeled clones contain either blood or endothelial cells, whereas only a small percentage harbor both lineages. These bilineage clones appear to result from labeling events in the pregastrulation embryo that preceded the separation of blood and endothelium-fated territories. To increase the likelihood of labeling the rare hemangioblast, an inducible Tie2-Cre mouse model was used in a directed genetic approach. Also in this model, the vast majority of yolk sac clones contained either blood or endothelial cells. Interestingly, Tie2-Cre–mediated labeling also generates a small proportion of clones with both endothelial and hematopoietic cells. In this instance, however, it appears to be the yolk sac hemogenic endothelium that is labeled, suggestive of labeling cells of the second yolk sac wave.

Altogether, the high frequency of unilineage clones does not favor a model where yolk sac blood and endothelium is derived from bipotent hemangioblasts. Instead, the data support a model (depicted in Figure 6 of Padrón-Barthe et al¹ ) where the precursors for primitive blood and yolk sac endothelium are already specified before gastrulation,^3,4  and colonize the yolk sac sequentially with the endothelial-fated lineage encircling the blood-fated cells.¹⁰  The authors propose that part of the labeled yolk sac endothelium represents the hemogenic endothelium that contributes to blood through an endothelial to hematopoietic transition. It remains to be determined, however, whether the choice for endothelium to be hemogenic or not is one dictated by extrinsic cues in the yolk sac, or due to the presence of a subset of intrinsically different cells specified earlier. Another question that remains is how to reconcile the BL-CFC with this model? The BL-CFC was shown to generate vascular smooth muscle as well as blood and endothelium, and may represent a mesodermal cell with an even broader differentiation potential. However, in vitro potential does not equal in vivo fate, which can be more restricted. Thus, could the BL-CFC be the in vitro readout of a primitive streak precursor that in vivo gives rise to the blood-fated cells of the prospective blood islands that Murray reports⁵ ? Or the precursor to the hemogenic endothelial lineage? Future lineage tracing studies, tracking specific blood progenitors and hemogenic endothelium back to their origins, will no doubt shed light on this.