Division of Stem Cell Research
Department of Cell Differentiation

Development of the hematopoietic system in the mouse

We have been trying to elucidate molecular mechanisms responsible for the hematopoietic commitment of mesodermal cells and endothelial cells. We have developed a culture system in which endothelial cells and hematopoietic cells differentiate from murine ES cells via lateral mesoderm as an intermediate precursor. We identified three distinct developmental pathways of hematopoietic cell lineages, i.e., differentiation of primitive erythrocytes from a subset of mesodermal cell population, hematopoietic differentiation of CD41-expressing precursors derived from lateral mesoderm, and hematopoietic differentiation of VE-cadherin-expressing endothelial cells. Among them, hemogenic differentiation of endothelium serves as a major source of the definitive hematopoietic cell lineages in the ES cell culture system.


Fig. 1. Differentiation of lateral mesodermal derivatives from mouse ES cells


We are interested in the molecular basis of the hematopoietic program inherent in the endothelial cells. The c-myb proto-oncogene is essential for the development of definitive hematopoiesis and is activated in hemogenic endothelial cells. We conditionally induced c-myb over-expression in the endothelial cells during in vitro differentiation of ES cells. Over-expression of c-myb augmented hemogenic potentials of endothelial cells, suggesting that c-myb is able to exert a function in endothelial cells fostering the establishment of their hemogenic potential. We also established an experimental system based on c-myb-null ES cells in which expression of c-myb can be induced at any stage and at any level during differentiation. Our data indicate that c-myb is a major gene that controls differentiation as well as proliferation of hematopoietic progenitors derived from hemogenic endothelial cells, and that appropriate levels of c-Myb protein are strictly defined at distinct differentiation steps of each hematopoietic cell lineage.


Fig. 2. Developmental programs inherent in the endothelial cells


Hematopoietic stem cells have been thought to develop from the vascular endothelium located in the aorta-gonad-mesonephros (AGM) region of the mouse embryo. However, several reports have suggested that most hematopoietic progenitors are derived from non-endothelial precursor cells expressing CD41. Our observations indicate that hemogenic endothelial cells and CD41-expressing progenitors are both able to serve as a developmental source of definitive hematopoietic cells, yet possess distinct hemogenic potentials depending on the embryonic stages and tissues where they reside. It is not clear whether the hematopoietic stem cells in the AGM region are derived from hemogenic endothelial cells or CD41-expressing progenitors. It has not also been proven whether the hematopoietic stem cells residing in the adult bone marrow originate from the stem cell population which developed in the AGM region. Several findings including ours have suggested that neither VE-cadherin-expressing cells nor CD41-expressing cells may contribute to the hematopoietic stem cell population in the adult mouse. These remain critical issues which we keep trying to resolve.

Development of the vascular system in the mouse

We are also interested in cellular mechanisms responsible for angiogenesis and vascular remodeling. ES cell-derived endothelial cell colonies generated on stromal cell layer have been used as an in vitro model to examine the behavior of individual endothelial cells in response to various angiogenic stimuli such as VEGF-A/VEGFR-2 and VEGF-C/VEGFR-3 signals. The VEGFR-2 signal was shown to induce elongation and dispersion of endothelial cells while the VEGFR-3 signal maintains integrity of cell-cell adhesion by modulating the VEGFR-2 signal. We also identified a forkhead-type transcription factor Foxo1 as a regulatory molecule in the elongation reaction of endothelial cells in response to the VEGFR-2 signal. Endothelial cells derived from foxo1-null ES cells showed a markedly different morphological response to exogenous VEGF-A compared with wild-type endothelial cells. This phenotype might account for the severe abnormality of angiogenesis found in foxo1-deficient mouse embryos. Therefore, elucidation of the function of Foxo1 in the morphological response of endothelial cells should provide a clue to how angiogenic growth factors regulate angiogenesis at the cellular level.


Fig. 3. Endothelial cell colonies stained with anti-VE-cadherin antibody


Confluent endothelial cells in culture are generally regarded as a model of resting endothelium in blood vessels (i.e., forming junctions at points of cell-cell contact, losing ability to proliferate in response to growth factors, and remaining stationary). However, incompatibility between junctional integrity and endothelial cell motility remains uncertain. Time-lapse analyses of endothelial cell colonies showed that though cells were connected to each other through adherens junctions and tight junctions, they were moving continuously within the colonies. Endothelial cell-specific expression of fluorescent protein-tagged VE-cadherin and claudin-5 revealed that adherens junctions and tight junctions persisted during endothelial cell migration. Our observations suggest that endothelial cells can remain highly motile without losing intercellular junctions. Therefore, as in the case of epithelial cell behaviors in morphogenesis, the compatibility of cell motility with junctional integrity should be an important characteristic of endothelial cells participating in angiogenesis. We are focusing on the mechanisms underlying the regulation of endothelial cell motility and junctional integrity during vascular development.