Search

The definitive long-term repopulating human hematopoietic stem cell, which seeds the adult blood system, was previously thought to derive from the extra-embryonic yolk sac. However, there is now considerable evidence that in both avian and murine systems, yolk sac hematopoietic cells are largely a transient, embryonic population and the definitive stem cell, in fact, derives from a distinct region within the embryonic mesoderm, the aorta-gonad-mesonephros region. In the human embryo, an analogous region has been found to contain a cluster of cells distinct from, but closely associated with, the ventral endothelium of the dorsal aorta, the appearance of which is restricted both spatially and temporally. We have used antibodies recognising hematopoietic regulatory factors to further characterise this region in the human embryo. These studies indicate that all factors examined, including vascular endothelial growth factor and its receptor FLK-1, Flt-3 ligand and its receptor STK-1, and stem cell leukemia transcription factor, are expressed by both hematopoietic cells in the cluster and endothelial cells. However, there is some discontinuity in cells directly underlying the cluster. Furthermore, we have identified a morphologically distinct region of densely-packed, rounded cells in the mesenchyme directly beneath the ventral wall of the dorsal aorta, and running along its entire length. In the preumbilical AGM region, directly underlying the hematopoietic cluster, but not at more rostral and caudal levels, this region of mesenchyme expresses tenascin-C, an extracellular matrix glycoprotein known to facilitate cell-cell interactions and migration. This region of cells may therefore provide the microenvironmental support for the intraembryonic development of definitive hematopoietic stem cells, a process in which tenascin-C may play a pivotal role.

Vitamin A requirement for early embryonic development is clearly evident in the gross cardiovascular and central nervous system abnormalities and an early death of the vitamin A-deficient quail embryo. This retinoid knockout model system was used to examine the biological activity of various natural retinoids in early cardiovascular development. We demonstrate that all-trans-, 9-cis-, 4-oxo-, and didehydroretinoic acids, and didehydroretinol and all-trans-retinol induce and maintain normal cardiovascular development as well as induce expression of the retinoic acid receptor beta2 in the vitamin A-deficient quail embryo. The expression of RARbeta2 is at the same level and at the same sites where it is expressed in the normal embryo. Retinoids provided to the vitamin A-deficient embryo up to the 5-somite stage of development, but not later, completely rescue embryonic development, suggesting the 5-somite stage as a critical retinoid-sensitive time point during early avian embryogenesis. Retinoid receptors RARalpha, RARgamma, and RXRalpha are expressed in both the precardiac endoderm and mesoderm in the normal and the vitamin A-deficient quail embryo, while the expression of RXRgamma is restricted to precardiac endoderm. Vitamin A deficiency downregulates the expression of RARalpha and RARbeta. Our studies provide strong evidence for a narrow retinoid-requiring developmental window during early embryogenesis, in which the presence of bioactive retinoids and their receptors is essential for a subsequent normal embryonic development.

Even though previous studies have shown that transcripts encoding the murine retinoid X receptor gamma (RXRgamma) are present in skeletal muscle of mouse embryos and that cultured myoblasts are induced to differentiate upon retinoid treatment, a function for RXRgamma and retinoids in mammalian myogenesis has not yet been identified. To begin to understand the possible role of RXRgamma during mammalian myogenesis we isolated novel rat RXRgamma cDNA sequences and examined in detail the spatio-temporal expression pattern of RXRgamma transcripts in relation to skeletal muscle differentiation in rat embryos and cultured myoblasts. We show that the onset of RXRgamma expression coincides with the differentiation of limb myoblasts in vivo. In vitro, RXRgamma is expressed in differentiating myoblasts, but not in proliferating myoblasts. In the myotome, however, RXRgamma is first expressed after myoblast differentiation, with RXRgamma transcripts being confined initially to its ventral region. Subsequently, RXRgamma becomes expressed throughout limb and myotome-derived muscle masses, and by the end of the primary myogenic wave, RXRgamma transcripts are mainly confined to their periphery. This dynamic expression pattern of RXRgamma during myogenesis suggests its possible involvement in the differentiation of limb myoblasts but excludes a role in the differentiation of early myotomal myoblasts.

The development of the vertebrate skeleton is under complex genetic control, and good progress is being made towards identifying the genes responsible. A recent paper contributes to this progress by describing transgenic mice in which the homeobox-containing MHox gene has been disrupted. MHox(-/-) mice have a range of skeletal defects, involving loss or shortening of structures in the skull, face and limb. Puzzling features of the MHox(-/-) mutation, which has similar effects on bones with very different embryological origins and yet spares other bones completely, may hold clues to the mechanisms that shape the skeleton. MHox(-/-) mice, used in conjunction with other skeletal mutants, will be important tools for exploring these mechanisms further.

Bone morphogenetic proteins BMP-4 and BMP-2 are closely-related members of the transforming growth factor-beta superfamily that have been implicated in signalling in a number of developmental systems. To determine whether they could be involved in the epithelial-mesenchymal interactions that control face development, we mapped the distribution of Bmp-4 and Bmp-2 gene transcripts in the developing chick facial primordia. At stages when primordia were becoming established, Bmp-4 transcripts were present in specific regions of epithelium in all facial primordia, but were undetectable in the mesenchyme. Bmp-4 transcripts appeared subsequently in specific regions of mesenchyme at the distal tips of the primordia. This mesenchymal expression first appeared in the frontonasal mass and then, in turn, in the lateral nasal processes, the maxillary primordia and the mandibular primordia. There was a complex relationship between domains of epithelial and mesenchymal Bmp-4 expression, and at many sites there was an inverse correlation between epithelial and mesenchymal Bmp-4 expression. Bmp-2 transcripts were found in the epithelium and mesenchyme of the maxillary and mandibular primordia at early stages in facial development. Bmp-2 transcripts appeared in the frontonasal mass and lateral nasal processes at later stages, with epithelial expression preceding mesenchymal expression. In general, mesenchymal Bmp-2 expression was associated with overlying epithelial Bmp-2 expression. The domains of Bmp-4 expression overlapped with those of Bmp-2, but detailed examination showed that there was no precise correlation between the expression patterns of the two genes. Indeed, in some places the Bmp-4 and Bmp-2 expression domains were complementary. The expression of the Bmp-4 and Bmp-2 genes in the epithelium and distal mesenchyme of the facial primordia suggests that BMP-4 and BMP-2 may be involved in the epithelial-mesenchymal interactions that control outgrowth of these primordia.

Retinoic acid is a good candidate for a morphogen in chick limb bud development. The challenge now is to determine how retinoic acid interacts with limb bud cells and how the retinoic acid signal is integrated with other signals to mould and pattern the developing limb.