New Clues About the Origin of Hematopoietic Stem Cells

To date, scientists have yet to succeed in differentiating pluripotent human stem cells (either embryonic stem cells or induced pluripotent stem cells) into functional hematopoietic stem cells (HSCs). Optimal in vitro differentiation of pluripotent cells into HSCs will likely require that we recapitulate the key stages of embryonic development in which these cells develop. Recent findings from the laboratory of Dr. Nancy Speck at the University of Pennsylvania have brought us one step closer in our understanding of the critical stages of hematopoietic development during embryogenesis.

Embryonic hematopoietic development is complex, with distinct hematopoietic populations appearing at different times and in different anatomical sites. One of the early waves of blood cell production is initiated in the yolk sac from erythroid-myeloid progenitor (EMP) cells, which give rise to erythrocytes, platelets, and myeloid cells but not lymphocytes. While EMP cells and their progeny are critical for the viability of the developing fetus, they provide only transient blood cell production. The final definitive stage of blood cell production occurs later in development, and, in contrast to EMP cells, is derived from HSCs, which supply all blood cell types — erythroid, megakaryocytic, myeloid, and lymphoid — for the entire lifespan of the organism. A key feature that sets HSCs apart from EMP cells is that only HSCs are capable of long-term engraftment in a myeloablated (usually by irradiation) host animal. Prior to this study by Chen et al., it was predicted that both EMP cells and HSCs develop from special endothelial cells, termed hemogenic endothelial cells, but it was not clear which endothelial subpopulations give rise to EMP cells and whether these were the same endothelial cells that also give rise to HSCs later in development.

The critical new finding in this paper is that EMP cells and HSCs are derived from two distinct hemogenic endothelial cell precursors. Formation of both EMP cells and HSCs requires a transcription complex containing a heterodimer of Runx1 (also known as AML1) and Core Binding Factor-beta (CBFβ). In order to test their hypothesis, the investigators used mice lacking the CBFβ gene, which make neither EMP nor HSC, and die early in gestation. In this study, the investigators restored CBFβ expression using either the TEK or the Ly6a promoters, which are activated in endothelial cells at different developmental stages. In TEK:Cbfβ mice, the TEK promoter drives CBFβ expression in endothelial cells of the early yolk sac as well as endothelial cells of the aorto-gonad-mesonephros (AGM) region. In contrast, CBFβ expression is restored in Ly6a:Cfbβ mice only at a later time in the AGM region. In TEK:Cbfβ mice, normal EMP cells developed, but the HSCs did not survive to provide long-term hematopoiesis, which caused perinatal death. In contrast, in the Ly6a-Cfbβ mice, the early transient EMP population was completely absent but fully functional HSCs developed. These data strongly suggest that these two major waves of hematopoietic cell development are derived from two different hemogenic endothelial populations, which are separate biological entities differing in the expression of key genes.