Faculty Bibliography

In this paper we examine the expression of several cell adhesion molecules and muscle differentiation markers during larval development and growth of the axial musculature of Xenopus laevis. We have identified a small group of cells, located on the dorsal extremity of the myotome, which express very high levels of N-CAM and early muscle markers such as desmin and muscle myosin, but do not express later-stage markers of muscle development such as sarcomeric actinin. These cells are most probably muscle precursor cells; they may participate in the generation of new fibres in the myotomal muscles of Xenopus. On the basis of these observations we propose a model of myotomal development in Xenopus characterized by a spatial segregation between regions of the myotome where new fibres are generated and regions where myotubes mature. As during muscle differentiation in vivo we observe a sharp change in the profile of expression of cell adhesion molecules, we suggest that different adhesion receptors are involved in the generation of new muscle fibres and in the growth and differentiation of already existing ones.

In the amphibian embryo, the ectoderm becomes a neural structure during gastrulation as a result of an interaction with the dorsal mesoderm. From that time onward, neurectodermal cells have the ability to express in vitro a large variety of mature phenotypes without further interaction with the inducing tissue. The neuralization of the ectoderm can be reproduced in a variety of experimental situations that do not involve the dorsal mesoderm. In this study we have analyzed biochemically the extent to which artificial inductions mimic the natural inducing process. Making use of antibodies specific to different neurotransmitter pathways, we have shown that the repertoire of the phenotypes expressed by experimentally induced neurons is always restricted compared to those obtained after induction in vivo. Only a limited number of generic neuronal characteristics are expressed. These results suggest that the expression of a complete neuronal phenotype normally involves a sequence of inductive events that can be experimentally uncoupled.

Mesodermal patterning in the amphibian embryo has been extensively studied in its dorsal aspects, whereas little is known regarding its ventrolateral regionalization due to a lack of specific molecular markers for derivatives of this type of mesoderm. Since smooth muscles (SM) are thought to arise from lateral plate mesoderm, we have analyzed the expression of an alpha-actin isoform specific for SM with regard to mesoderm patterning. Using an antibody directed against alpha-SM actin that recognized specifically this actin isoform in Xenopus, we have found that the expression of alpha-SM actin is restricted to visceral and vascular SM with a transient expression in the heart. The overall expression of the alpha-SM actin appears restricted to the ventral aspects of the differentiating embryo. alpha-SM actin expression appears to be activated following mesoderm induction in animal cap derivatives. Moreover, at the gastrula stage, SM precursor cells are regionalized since they will only differentiate from ventrolateral marginal zone explants. Using the animal cap assay, we have found that alpha-SM actin expression is specifically induced in treated animal cap with bFGF or a low concentration of XTC-MIF, which induce ventral structures, but not with a high concentration of XTC-MIF, which induces dorsal structures. Altogether, these results establish that alpha-SM actin is a reliable marker for ventrolateral mesoderm. We discuss the importance of this novel marker in studying mesoderm regionalization.

The nucleotide-binding protein Go is a transducing molecule closely associated with neural structures in vertebrates. Because of the potential importance of molecules of this type during the first step of neurogenesis, we have investigated the kinetics of expression of Go in the amphibian (Pleurodeles waltl) embryo, focusing our attention on the stages corresponding to the acquisition of neural competence by presumptive ectoderm and to the process of neural induction. Using affinity-purified IgGs directed against the alpha subunit of Go, Go-like immunoreaction (GoLI) is first detected at the midblastula stage in some animal cap (future ectodermal) cells just before they have attained competence to be neuralized. At the early gastrula stage, GoLI is almost exclusively expressed by neural-competent tissue as a whole, with no obvious difference between the dorsal (prospective neural) and the ventral (prospective epidermal) ectoderm. The expression of GoLI is therefore related to the state of competence of the tissue rather than to its fate. At the early neurula stage, immediately following neural induction, the expression of GoLI persists essentially in that part of ectoderm that has been diverted from epidermal differentiation towards the neural pathway; in the ventral ectoderm, as neural competence is lost GoLI disappears. Furthermore, in the neurectoderm, only approximately 70% of the cells conserve GoLI, demonstrating that immediately following neural induction the population of neurectodermal cells is not homogeneous.

In amphibian development, neural structures arise from the presumptive ectoderm at the gastrula stage by an inductive interaction with the chordamesoderm. It has been previously reported that early gastrula presumptive ectoderm can be neuralized when it is dissociated into single cells. A similar result is reported here with regard to Pleurodeles waltl presumptive ectoderm. Using this experimental model system we demonstrate: first, that neuronal and glial lineages can be specified from the presumptive ectoderm without any intervention of the natural inducing tissue; and second, that whereas rupture of cell-cell contacts evoked neural induction, dissociation immediately followed by reaggregation reduces the neuralizing response, pointing toward an active role played by cell-cell contacts of presumptive ectodermal cells in the modulation of neural commitment.

After neural induction certain cells in the neuroepithelium immediately acquire the property to express certain neural phenotypes (Duprat et al., 1984, 1987). However, the activity of almost all the specific enzymes involved in the biosynthesis of neurotransmitters is considerably higher when neurectodermal cells are cultured with chordamesodermal cells than when they are cultured alone. The stimulating effects of chordamesoderm do not appear to be due to diffusible factors (Duprat et al., 1985b). The present study was designed to investigate the role of extracellular matrix components in neuronal cell differentiation. We showed that the extracellular matrix cannot replace chordamesoderm in stimulating the biochemical differentiation of neuroblasts, although fibronectin and especially laminin stimulate morphological differentiation. We suggest that interaction between neuronal and non-neuronal cells plays an important part in functional biochemical differentiation, whereas the molecules of extracellular matrix are important for morphological differentiation.

As an immediate consequence of neural induction during gastrulation, some neuroectodermal cells acquire the ability to develop a number of specific neuronal and astroglial features, without requiring subsequent chordamesodermal cues. Thus, cholinergic, dopaminergic, noradrenergic, gabaergic, somatostatinergic, enkephalinergic, etc. traits are expressed in cultures of neural plate and neural fold isolated from amphibian late gastrulae immediately after induction and cultured in a defined medium. These results strongly suggest that at the late gastrula stage, the neural precursor population does not yet constitute a homogeneous set of cells. It was of interest to know the origin of this heterogeneity. Is it a direct result of the process of neural induction itself, stochastic phenomena being involved or not at the cellular level, or does it reflect a pre-existing heterogeneity in the presumptive ectoderm? At the early gastrula state, presumptive ectoderm can be neuralized consecutively to its dissociation into single cells. Using this experimental model, we have demonstrated by means of immunological probes that neuralized presumptive ectodermal cells, without any intervention of the chordamesoderm (natural inducing tissue), can develop autonomously into glial and neuronal lineages. These data suggest the existence of diverse predispositions of presumptive ectodermal cells. Competent ectoderm seems to be a heterogeneous structure with cells presenting distinct neural predispositions that can emerge as a consequence of a permissive inductive signal without real specificity (such as a target tissue dissociation). Moreover, such a differentiated neuronal population includes neurons of the GABAergic and enkephalinergic phenotypes but not of the cholinergic, catecholaminergic, somatostatinergic, etc. phenotypes. These data show that the developmental program of ectodermal cells induced without interaction with the chordamesoderm appears restricted compared to the naturally induced ectoderm. Experiments are now under way to analyze such sequential neural events.

The appearance and localization of N-CAM during neural induction were studied in Pleurodeles waltl embryos and compared with recent contradictory results reported in Xenopus laevis. A monoclonal antibody raised against mouse N-CAM was used. In the nervous system of Pleurodeles, it recognized two glycoproteins of 180 and 140x10(3) M(r) which are the Pleurodeles equivalent of N-CAM-180 and -140. Using this probe for immunohistochemistry and immunocytochemistry, we showed that N-CAM was already expressed in presumptive ectoderm at the early gastrula stage. In late gastrula embryos, a slight increase in staining was observed in the neurectoderm, whereas the labelling persisted in the noninduced ectoderm. When induced ectodermal cells were isolated at the late gastrula stage and cultured in vitro up to 14 days, a faint polarized labelling of cells was observed initially. During differentiation, the staining increased and became progressively restricted to differentiating neurons.

The molecular mechanism of neural induction of embryonic cells is an important but poorly understood problem in neuroembryology. Glycoconjugates in the target cell plasma membrane and/or its structural organization play a key role in the reception of the inductive signal. It is the competent target tissue itself which probably contains the capacity and specificity for neuralization. However, the mechanism of transmission of the signal which leads to activation of the intracellular machinery involved in the process of neural determination remains to be elucidated. With respect to the information acquired by the target cells during neural induction, and the early events in differentiation, neuronal precursor cells have been shown to have acquired the potential to display a high degree of biochemical and phenotypic differentiation, even in the absence of further embryonic influences.