5-HT2B receptor-mediated serotonin morphogenetic functions in mouse cranial neural crest and myocardiac cells

Serotonin (5-hydroxytryptamine, 5-HT) is one of the well known monoamine neurotransmitters, mitogens, and hormones, which mediate a wide variety of physiological effect, including peripheral and central actions. The large variety of 5-HT functions is paralleled by the pharmacological complexity of 5-HT receptors which can be classified into different families depending on their signalling pathways. The family including 5-HT1 and 5-HT5 receptors interacts negatively with adenylyl cyclase; the 5-HT2 receptor family is coupled to the activation of phospholipase C (PLC); the family, including 5-HT4, 5-HT6 and 5-HT7 receptors, activates adenylyl cyclase, whereas the 5-HT3 receptor is a ligand gated ion channel.

Recently, we have cloned the 5-HT2B receptor cDNA from the mouse (Loric et al., 1992) and human (Choi et al., 1994). These receptors are functionally coupled to Gq and to the ras signalling pathway, and can be considered as a ligand dependent oncogene acting on protein kinase C (PKC) and MAPKinase activation (Launay et al., 1996). The development of a subtype-specific antiserum allowed a precise mapping of the distribution of the mouse 5-HT2B receptors. In adult mice, the major sites of expression of 5-HT2B receptors are the gut and heart, and there is also detectable levels of expression in the brain and kidney (Choi and Maroteaux, 1996). Interestingly, the expression of 5-HT2B receptor mRNA is observed as early as 8 days postcoitum (d.p.c.) of mouse embryonic development (Loric et al., 1992). Additionally, the 5-HT2 receptor homologue in Drosophila is expressed during gastrulation (Colas et al., 1995), and the 5-HT2B receptors are also expressed during the serotonergic differentiation of the mouse teratocarcinoma derived 1C11* cell lines (Loric et al., 1995; Kellermann et al., 1996).

Several line of evidence indicate that 5-HT possess developmental functions (Lauder, 1988; Lauder, 1993). For instance, it has been shown that 5-HT is present early in mammalian embryonic development and is probably maternally derived (Yavarone et al., 1993a). Moreover in the mouse, the ability to take up 5-HT is detected at early embryonic stages (8.5 d.p.c., 10 somite pairs) in the heart myocardium (Shuey et al., 1993) and in the rhombencephalic neuroepithelium where it is restricted to rhombomeres 3 and 5 (Shuey et al., 1993). The physiological relevance of these findings is stressed by the data showing that embryos grown in the presence of high levels of 5-HT or 5-HT uptake blockers develop deficient head mesenchyme, hypoplastic mandibular arches and forebrain, open cranial neural folds and abnormal eyes. These malformations overlap those observed in whole embryo culture following exposure of mouse embryos to 13-cis-retinoic acid (Lauder, 1988; Lauder et al., 1988). 5-HT has also been shown to participate in rat craniofacial development (Van Cauteren et al., 1986) and to be involved in the formation and migration of the cranial neural crest (NC) cells (Moiseiwitsch and Lauder, 1995) as well as in chicken and mouse cardiovascular morphogenesis (Huether et al., 1992; Yavarone et al., 1993b).

Here we used molecular biology, pharmacology and whole embryo culture techniques to present evidence for involvement of 5-HT2B receptors in mouse embryonic development at the neurulation stage.

Total RNA from 8-13 d.p.c. mouse embryos was isolated using the guanidinium-thiocyanate method. Reverse transcriptase-PCR (RT-PCR) experiments were performed using specific oligonucleotides for mouse 5HT2A, 2B, 2C, receptors and for the serotonin transporter (SERT) as follows:

5-HT2A-AAGCCTCGAACTGGACAATTGATG/AAGATTTCA-GGAAGGCTTTGGTT

5-HT2B-CAGAAGACATGTGATCACCTGATC/TGTAATCTTG-ATGAATGCAGTAGCC

5-HT2C-GTTTGGCAGTTCGATATTCCATA/GGCAATCTTCA-TGATGTAGTC

SERT-ACATCTGGCGTTTTCCCTACAT/TTTTGACTCCTTTC-CAGATG CT

To avoid the possible amplification of genomic DNA, oligonucleotides were designed in different exons. RT-PCR reaction was tested with different amounts of RNA (3, 4, 5 μg/50 μl reaction) to keep the reaction in the exponential phase for quantitative analysis of mRNA, and its specificity was verified by Southern blotting and hybridisation to a third primer.

Binding assays were performed with embryo membranes using [125I]2,5-dimethoxy-4-iodophenyl-2-aminopropane ([125I]DOI), alone or in presence of ketanserin, mesulergine or ritanserin (10 −8 M) as described by Loric et al. (1992). Membrane extracts were also used for SERT quantification by binding experiments using [3H]paroxetine and for tryptophan hydroxylase (TPH) activity dosage by enzymatic assay according to published procedures (Buc-Caron et al., 1990).

For whole-mount in situ hybridisation, embryos were fixed with 4% paraformaldehyde followed by dehydration in serial dilutions of methanol and stored at −20°C in methanol until use. Hybridisation experiments were performed as described by Wilkinson and Dourish (1991), using whole 5-HT2B cDNA sense or antisense probes. Embryos were fixed with 3% paraformaldehyde and then treated for whole-mount immunohistochemistry with an affinity-purified antiserum specific for 5-HT2B receptors as described by Mark et al. (1993) and Choi and Maroteaux (1996).

CD1 strain mouse embryos were cultured as described by MorrissKay (1993). Briefly embryos were dissected in cold Tyrode’s solution, and then placed in 30 ml sterile glass bottles with 2.5 ml rat serum, 2.5 ml Tyrode’s buffer, 0.1 ml penicillin/streptomycin. The bottles were filled with a mixture of 90% nitrogen, 5% oxygen, 5% carbon dioxide. Embryos of 8 d.p.c. (around 5 somite pairs) were incubated for 24 hours on a rotator wheel (30 revolutions per minute) at 37.5°C (Bavik et al., 1996). Drugs were dissolved in water or ethanol and then diluted at least 1000× (5 μl/5 ml) in culture medium. The morphological analysis is based on that described by New (1990). Experiments were performed at least 4 times with 4 or 5 treated embryos each.

Embryos were fixed by immersion in 2.5% glutaraldehyde in cacodylate buffer (0.1 M, pH 7.2) overnight at room temperature and washed in cacodylate buffer for further 30 minutes followed by postfixation with 1% osmium tetroxide in 0.1 M cacodylate buffer for 1 hour at 4°C. Embryos were dehydrated through graded alcohol (50, 70, 90, 100%) for 30 minutes each and embedded in Epon 812. Semithin sections were cut at 1.5 μm and stained with toluidine blue, and histologically analysed by light microscopy. Ultrathin sections were cut at 70 nm and contrasted with uranyl acetate and lead citrate, and examined with a Philips 208 electron microscope. For scanning electron microscopy, embryos were fixed, dehydrated as above, dried with critical point-drying apparatus, and then mounted on aluminium stubs coated with palladium-gold using a cold sputter-coater and observed with a Philips XL-20 microscope.

Nile Blue sulphate (NBS) dye staining of apoptotic cells was performed as described by Jeffs et al. (1992) except that phosphate-buffered saline was used as the incubation medium.

The staging of the mouse embryos follows that described by Hogan et al. (1994) with the following criteria: 8 d.p.c. corresponds to 1-7 somite pairs; 8.5 d.p.c. to 8-12 somite pairs (embryo turning); 9 d.p.c. to 13-20 somite pairs; 9.5 d.p.c. to 20-30 somite pairs and 10 d.p.c. to 30-35 somite pairs.

Previous studies have indicated that different 5-HT2 receptors are expressed early during rodent embryonic devel-opment (Loric et al., 1992; Hellendall et al., 1993). We have refined these data by performing semi-quantitative RT-PCR on RNA extracts from staged embryos using a set of amplimers specific for the 5-HT2A, 2B and 2C receptors. Representative results of such experiments are shown in Fig. 1A,B. We observed a strong expression of 5-HT2B mRNA at 8 d.p.c. followed by a sharp decreased by 10 d.p.c. In contrast, low levels of 5-HT2A mRNA were detected at 8 d.p.c. and rose slowly. The expression of 5-HT2C mRNA appeared only at 13 d.p.c. in parallel with a second wave of 5-HT2B mRNA expression. Interestingly, elevated levels of SERT mRNA and protein (as [³H]paroxetine binding sites) were detected from 8 to 10 d.p.c. (Table 1; Fig. 1A). Binding experiments performed on protein extracts from similarly staged embryos, using [¹²⁵I]DOI, a specific 5-HT2 radio-ligand, revealed a peak of specific binding at 9 d.p.c. (Fig. 1C). The specific binding of [¹²⁵I]DOI was completely blocked by ritanserin, partially reduced by mesulergine and totally insensitive to ketanserin. Given the respective affinity of these drugs for the different 5-HT2 receptors (Table 2), these results indicate that the peak of [¹²⁵I]DOI binding detected at 9 d.p.c. is mainly due to the expression of 5-HT2B receptors.

Further refinement of the 5-HT2B receptor expression was obtained by studying its mRNA distribution by whole mount in situ hybridisation. Strong expression of the 5-HT2B receptor mRNA was detected at 8 d.p.c. in the neural fold, neural tube and heart primordia (Fig. 2A). This expression persisted in the cephalic region, neural tube and heart at 9 d.p.c. (Fig. 2B). This mRNA expression was confirmed by immunostaining whole-mount embryos using a 5-HT2B C-terminal-specific antiserum (Choi and Maroteaux, 1996): 5-HT2B receptor protein appears to be expressed in the somites, neural tube and heart region at 8.5 d.p.c. (Fig. 3B,D), and persisted in the same regions with addition of the otic and optic vesicles, and pharyngeal arches at 9.5 d.p.c. (Fig. 3F,H).

Early embryonic expression of 5-HT2B receptors raised the possibility that these receptors are involved in 5-HT-dependent morphogenetic processes. Therefore, we initiated a series of experiments to test this hypothesis by performing whole embryo culture. Embryos at about the 5 somite-pairs stage were incubated for 24 hours in medium containing 1:1 rat serum/Tyrodes salt solution and various serotonergic drugs. Normal rat serum contains nearly micromolar amount of 5-HT (0.5-1 μM) and dialysed serum cannot support embryonic development in our experimental conditions, even supplemented by 1 μM 5-HT. Therefore, in order to block the action of 5-HT2B receptors, we selected specific antagonists of the 5-HT2 subtypes of receptors, ritanserin, methysergide, mesulergine, ketanserin and mianserin, for their differential affinity for these receptors (Table 2). No defects were detected in embryos incubated in normal serum, whereas treatment with antagonist induced several reproducible embryonic defects which were characterised by optical microscopy (Fig. 4A-C), and by scanning electron microscopy (Fig. 4D-F). As shown in (Fig. 4B,E) ritanserin-treated embryos show a strong growth retardation as compared to controls (Fig. 4A,D). The yolk sac circulation was also impaired as indicated by the formation of few blood islands. Moreover, the cephalic region showed apparent defects in flexure, and the forebrain, hindbrain and the first pharyngeal arch developed abnormally. In addition, the epicardial layer of the ventricular wall was swollen. Defects in embryonic turning, in somite number and shape, and in neural tube shape and closure were also frequently observed after ritanserin treatment (Fig. 4B,E; Table 3). Ritanserin induced these defects in all treated embryos at 1 μM con-centration (Table 3), whereas at 100 nM ritanserin nearly 20% of the embryos were already affected (not shown). Defects could be visually detected 6 hours after the beginning of treatment. Simultaneous addition of agonist with ritanserin, 5-HT (Fig. 5; Table 3) or the highly specific 5-HT2B agonist N-acetyl-5-HT (NAS) (Colas et al., 1997), prevented the onset of these defects in almost 50% of the treated embryos (Table 3). Interestingly, among the other 5-HT antagonists tested (Table 3) only those having a high affinity for 5-HT2B receptors (Table 2), gave a similar phenotype to that of ritanserin. In contrast, the antagonists mianserin (not shown) and ketanserin which have a lower affinity for 5-HT2B than for 5-HT2A or 5-HT2C receptors (Table 2), gave a milder phenotype even at a concentration of 10 μM (Table 3; Fig. 4C,F), pointing out a 5-HT2B-mediated action.

The defects induced by 5-HT antagonists were further characterised by light and transmission electron microscopy (TEM) on thin sections of these embryos. Toluidine blue stained sections showed that ritanserin-treated embryos have hypoplastic pharyngeal arches and an irregular neural tube with dilated blood vessels (Fig. 6B-H). It is important to point out that the reduced size of pharyngeal arches is indicative of impairment of NC cell migration and/or proliferation. The NC cells in the ritanserin-treated embryos, seemed not to migrate properly and remained in a more dorsal aspect of the cephalic region (Fig. 6F) than in the control embryos (Fig. 6E). Fig. 7 shows higher magnification of the neural crest region of these sections. Although dividing cells can be seen in the epithelium, densely stained pyknotic nuclei, which are indicative of dying cells, were specifically observed in NC cells of the first pharyngeal arch and in the neuroepithelium near the optic stalk. Several characteristics of apoptosis were identified, cells presenting electron dense nuclei with condensed chromatin, dark cytoplasm and fragmented cells, whereas no necrotic cells could be seen after ritanserin treatment (Fig. 7B-D). Counting the number of apoptotic NC cells on these sections gave an average of 52% apoptotic cells in ritanserin-treated embryos whereas only 6% were observed in vehicle-treated embryos. In order to confirm the apoptotic nature of these cells we stained embryos with the NBS dye which has been frequently used to visualise patterns of cell death during embryogenesis (Lumsden et al., 1991; Jeffs et al., 1992). The cephalic region of the ritanserin-treated embryos show extensive NBS staining over the pharyngeal arch 1, forebrain region and optic vesicle (Fig. 8B,D) in addition to the staining observed in control embryos over the rhombomeres 3 and 5 (Fig. 8A,C) already described (Graham et al., 1993, 1994; Lumsden and Graham, 1996). Further-more, in treated embryos, the kinked neural tube and somites have more NBS-labelled cells than control embryos (Fig. 8C,D).

Similar studies of the cardiac region were performed (Fig. 9). In ritanserin-treated embryos the heart structure was disorganised (Fig. 9B,D). In particular, the bulboventricular groove was not well defined and the swollen atrioventricular canal showed reticulocyte accumulation indicating an inefficient circulation (Fig. 9B). The subepicardial layer was thin and the cardiac trabecular cells were absent in the ventricle. Elongated cells, however, were present in this layer which normally contains myocardial trabecular stem cells (Fig. 9D). TEM analysis of thin sections of ritanserin-treated embryos (Fig. 9E,F) indicated that this compound induces abnormal differentiation of myofilament sarcomeres in the subepicardial layer. These anomalies were not detected in vehicle-treated embryos at the same stage. This results suggest either a modification of the differentiation program of the myocardial stem cells and/or a deficient migration of the precursors of the trabecular cells.

The presence of 5-HT has been detected during early mammalian embryonic development (Lauder et al., 1988). We observed that the limiting enzyme in 5-HT biosynthesis, tryptophan hydroxylase (TPH) cannot be detected in embryos between 8 and 10 d.p.c. (Table 1), substantiating its suspected maternal origin (Yavarone et al., 1993a). Similarly, we detected the presence of binding sites at 8-10 d.p.c. for the SERT-specific ligand [³H]paroxetine (Table 1), corroborating the 5-HT uptake activity previously described at the 10 somite-pairs stage (8.5 d.p.c.; Lauder et al., 1988) in the neuroepithelium and myocardium (Shuey et al., 1993). Furthermore, 5-HT has been shown to affect rat and mouse craniofacial (Van Cauteren et al., 1986; Shuey et al., 1993) and chicken and mouse cardiovascular morphogenesis (Huether et al., 1992; Yavarone et al., 1993b). Some of these effects have been reported to be mediated by 5-HT2-like receptors (Van Cauteren et al., 1986; Huether et al., 1992).

We describe here, an extensive identification of 5-HT receptors present at the neurulation stage of early mouse embryogenesis. We have focused on the 5-HT2 receptor family since the mitogenic activity of 5-HT has been linked to 5-HT2 receptor-dependent stimulation of PLC/PKC. (i) 5-HT has mitogenic effects on NIH3T3 fibroblasts expressing a high density of 5-HT2A or 5-HT2C receptors (Julius et al., 1989), as well as on LMTKcells expressing 5-HT2B receptors, which in response form foci and induce tumours in nude mice (Launay et al., 1996). (ii) We have shown that, in these later cells, 5-HT2B receptors are functionally coupled to IP3 stimulation via the heterotrimeric alpha subunit of Gq and to the ras signalling pathway via the beta-gamma subunit of the same protein. These trigger the effects of 5-HT2B selective agonists on PKC and MAPKinase activity in these cells (Launay et al., 1996). (iii) 5-HT acts on the serotonergic differentiation of the teratocarcinoma-derived cell line 1C11* which expresses 5-HT2B receptors before 5-HT2A receptors and regulates the final phenotype of these cells (unpublished data; Loric et al., 1995; Kellermann et al., 1996). These overall properties strongly suggest that 5-HT2B receptors mediate some trophic functions of 5-HT.

In the present study we have observed both by mRNA analysis (Fig. 1A,B) and receptor protein binding (Fig. 1C), that the 5-HT2B subtype is the major 5-HT2 receptor expressed between 8 and 9 d.p.c.. Whole-mount in situ hybridisation (Fig. 2) and immunohistochemistry experiments (Fig. 3) reveal expression of this receptor on the neural tube, cranial neural crest cells, on somites, and in the heart tube at 8.5 d.p.c. (Fig. 3B,D). At 9.5 d.p.c., the forebrain region, the pharyngeal arch 1 and the otic and optic vesicles are also positively labelled (Fig. 3F,H). At a later stage of development, 11.5 d.p.c., the cardiac staining is restricted to the trabecular cells of the myocardium (not shown). Interestingly, the 5-HT2B staining at 8.5 d.p.c. is observed in the region of migratory NC cells (Fig 3D), and the staining observed at 9.5 d.p.c. over the pharyngeal arch is suggestive of that of the cranial NC cells after migration. The 5-HT2B receptor staining over the somites, neural tube and cranial neural crest is overlapping with that of the product of the paired box gene Pax-3 (Goulding et al., 1993; Pourquié et al., 1995). This transcription factor is responsible for the splotch phenotype in mice and the Waardenburg syndrome in human (Wehr and Gruss, 1996) which primarily affect neural crest cell derivatives and its gene has a similar localisation to the 5-HT2B receptor gene on chromosome 2q36 (LeConiat et al., 1996).

The treatment of embryos with high affinity 5-HT2B receptor antagonist induce a highly reproducible phenotype which includes strong growth retardation (Fig. 4B,E), abnormal flexure of the cephalic region, underdeveloped forebrain and hindbrain, small pharyngeal arches and distended epicardial layer (Figs 6, 8). Defects in embryo turning, in somite number and shape, and in neural tube shape and closure are also frequently observed after these treatments (Table 3). This phenotype, including small brain and pericardial oedema, is very similar to that observed in the mouse mutated for the fused gene (Perry III et al., 1995), and in the knock-out mice phenotype of Ras GTPase-acti-vating protein (Henkemeyer et al., 1995) and of the neurofi-bromatosis type-1 gene which is a regulator of the ras signal transduction pathway (Brannan et al., 1994). Finally, it is worth noting that a human pathological condition showing similar embryonic defects to those we observed, has been described in the embryos of phenylketonuria mothers (Lenke and Levy, 1980), and seems to be associated with low levels of blood 5-HT (Roux et al., 1995).

All the affected areas of the embryos express high levels of 5-HT2B receptors (Figs 2, 3). In addition, the effects induced by ritanserin or mesulergine seem to be specific. First, morphological modifications induced by antagonist treatment are dose-(Table 3) and time-dependent, and can be partially prevented by simultaneous treatment with NAS, a highly selective 5-HT2B receptor agonist or with 5-HT (Fig. 5; Table 3). Second, the yolk sac circulation impairment is not directly involved since nearly normal embryos were asso-ciated with blood island formation in ritanserin-plus-5-HT treatment (Table 3). Third, a general toxic effects of these compounds can be excluded since the defects are tissue-specific and, despite the presence on sections of apoptotic cells, dividing cells can still be observed (see Fig. 7B). Finally, the non-migrating NC cells undergo cell death with typical apoptotic but not necrotic morphology (Fig. 7D). The possibility that ritanserin acts at other 5-HT receptor subtypes cannot be completely excluded but is very unlikely since these compounds are typical 5-HT2-specific ligands (Hoyer et al., 1994); the only known serotonin receptors having fairly high affinity for ritanserin are the rat 5-HT6 and 5-HT7 subtypes but their affinity for ritanserin is at least 10 times lower than that of mouse 5-HT2B receptors (Boess and Martin, 1994).

The 5-HT action on early embryonic development includes head mesenchyme and pharyngeal arch formation, neural tube closure, eye and heart development, and is partially over-lapping with retinoic acid teratology (Lauder, 1988). The ritanserin treatment of embryos induced a modification of the differentiation program of the myocardial stem cells and/or a deficient migration of the precursors of the trabecular cells (Fig. 9). The absence of trabeculation in the myocardium is similar to that reported in neuregulins knock-out mice (for review see Lemke, 1996), and in RXRα mutated animals (Kastner et al., 1994). In both cases, the thickness of the ven-tricular wall is markedly decreased. The retinoic acid-induced transcription factor AP-2, when knocked-out, also gives a similar cardiac phenotype (Shorle et al., 1996; Zhang et al., 1996). Since the promoter region of the 5-HT2B receptor contains consensus binding sequence for retinoic acid receptor and AP-2 (unpublished), we propose that the 5-HT2B receptors are interacting with these transcription factors.

In mouse, the NC cells emerge by epitheliomesenchymal transformation (Newgreen and Minichiello, 1995), from epithelial cells on both sides of the neural tube at the neurulation stage (Le Douarin, 1982; Hall and Hörstadius, 1988). There is strong evidence that the cranial NC cells are actively migrating toward the developing pharyngeal arches and the frontonasal process (Morriss-Kay and Tan, 1987; Hall and Hörstadius, 1988; Selleck and Bronner-Fraser, 1996). Here, they differentiate into mesenchymal derivatives such as bone, cartilage and muscles (Le Douarin, 1982; Hall and Hörstadius, 1988). Several factors have been implicated in stimulation and guidance of NC migration including cell adhesion molecules (tenascin, fibronectin, laminin) and their receptors (integrins) (Bronner-Fraser, 1993). Humoral factors such as growth factors (Le Douarin et al., 1993; Selleck and Bronner-Fraser, 1996) or neurotransmitters are also implicated in NC migration. Evidence that 5-HT and some of its receptors participate in these processes have also been reported (Moisei-witsch and Lauder, 1995).

In our experiments of embryos treatment with 5-HT2B antagonist, widespread cell death was detected in the hindbrain, in cranial NC and in disorganised somites (Figs 7, 8). The site of expression of the 5-HT2B receptor in cranial NC cells and the phenotype observed after antagonist treatment, where NC cells present induced-cell death with typical apoptotic nuclei (Fig. 7D,F), suggest a role for 5-HT2B receptors in NC migration, cellular proliferation and/or survival. One possibility is the involvement of 5-HT in the formation of cranial NC cells by epitheliomesenchymal transformation which has been shown to be dependent on PKC activity (Newgreen and Minichiello, 1995), and is under the control of members of the transforming growth factor β superfamily (Selleck and Bronner-Fraser, 1996). Conversely, 5-HT is involved in cranial NC cell migration (Moiseiwitsch and Lauder, 1995), and previous experiments have suggested the involvement of a 5-HT2-like receptor since the migration of mesenchymal cells can be blocked by methysergide (Lauder and Zimmerman, 1988). The contribution of 5-HT2C receptors in these effects can nevertheless be ruled out, since 5-HT2C knock-out mice do not show any developmental deficits (Tecott et al., 1995). Interestingly, we have detected SERT mRNA (Fig. 1) and paroxetine-binding (Table 1) in extracts from embryos at 8-10 d.p.c., and SERT uptake activity has been localised in the neuroepithelium at the 12 somite-pair stage in rhombomeres r3 and r5 (Shuey et al., 1993). These data along with the observation that apoptotic NC cells are present in the same rhombomeres of chicken embryos at the same stage (Graham et al., 1993; Lumsden and Graham, 1996) indicate that rhombomeres r3 and r5 are associated with apoptotic NC cells and that local uptake activity reduces the levels of 5-HT. It is therefore, tempting to speculate that treatment by 5-HT2B-antagonist mimics the absence of 5-HT stimulation resulting normally from the local uptake of 5-HT and therefore broadens the apoptosis of NC cells (Fig. 8). This would predict that 5-HT acting on the 5-HT2B receptors exerts a trophic action on cranial NC cells to prevent the induction of programmed cell death.

In conclusion, our data indicate that 5-HT2B receptors participate in the regulation of cranial NC cells, by affecting their migration, cellular proliferation and/or survival, as well as migration and/or inhibition of cellular differentiation of the heart myocardium during early mouse embryogenesis. Investigations with gene targeted (knockout) mice should further validate these findings about the functions of 5-HT2B receptors during early embryonic development.

We wish to acknowledge P. Hickel, T. Ding, B. Schuhbaur, C. Dennfeld, for excellent technical assistance, F. Perrin-Schmidt for the early in situ hybridisation experiments, V. Dupé for immunohistochemistry advise, and B. Boulay and J.-M. Lafontaine for help in preparing the artwork of this manuscript. We thank Dr F. Bola ños, for critical reading of the manuscript, Drs E. Borrelli, V. Mutell and G. M. Morriss-Kay for helpful discussions. This work was supported by funds from the Centre National de la Recherche Scientifique, the Institut National de la Santé et de la Recherche Médicale, the Centre Hospitalier Universitaire Régional, and by grants from the European Community, from the Ministère de la Recherche, from the Fondation pour la Recherche Médicale, and from the Association pour la Recherche contre le Cancer # 6800.