Melanocytes arise from the neural crest

Vertebrate skin pigment cells or melanocytes that synthesize and store melanins are found in epidermal and dermal layers of the skin. Melanocytic progenitors are the melanoblasts which themselves are unpigmented, but have the potential to produce melanin. The neural crest (NC) origin of pigment cells during vertebrate development was first deduced from pioneer extirpation and transplantation experiments in amphibian and birds (reviewed by Lecoin et al., 1998a). However, the early dispersion of NC cells and the migration routes of unpigmented melanoblasts were largely unknown until cell markers were devised to follow NC cells in birds. The construction of quail-chick neural primordium chimeras allowed to identify and trace the multiple NC-derived cell types including melanocytes. The NC forms on the mediodorsal aspect of the closing neural folds at late neurula stages, whereas the neural tube itself gives rise to the CNS. The anterior limit of the NC is located in the diencephalon at the level of origin of the epiphysis (Couly et al., 1993). NC cells undergo an epithelio-mesenchymal transition to yield migratory cells dispersing along definite pathways throughout the developing embryo. They stop at different locations and give rise to a large array of derivatives. NC-derived cells include neurons and glial cells of the PNS, endocrine cells and melanocytes. At the cephalic level, the NC also yields mesenchymal cells that differentiate into the musculo-connective wall of large arteries derived from aortic arches (except endothelial cells), facial and visceral arch skeleton and dermis, the skull vault (for references, Le Douarin, 1982; Le Douarin and Kalcheim, 1999) and meninges in the anterior brain (Etchevers et al., 2001).

Pigment cells found in epidermal and dermal skin and in the choroid layer of the eye are therefore derived from the NC, in contrast pigment cells in the retinal pigmented epithelium originate from the neural epithelium of the optic cup. The migration paths, final location and terminal differentiation of donor quail NC cells were followed in quail-chick neural primordium chimera. At trunk level, NC cell migration proceeds in two different waves: the first one, dorsoventrally near the neural tube within the anterior part of the somites, leads to forming peripheral ganglia and nerves. The second one, mediolaterally between the superficial ectoderm and the dermomyotome, is used by melanocyte precursors that invade the subectodermal mesenchyme (Teillet and Le Douarin, 1970; Teillet, 1971). The latter mediolateral path 'opens' 1 day later than the dorsoventral one, so that NC presumptive melanoblasts accumulate near the dorsal neural tube in the 'migration staging area' (Erickson et al., 1992) before colonizing the skin. Prospective melanocytes migrate essentially through the mesenchyme and then invade the epidermis, to differentiate into pigment cells in the feather germs of chicken at embryonic day 9 (E9) (Teillet, 1971).

Labelling melanoblast precursors with monoclonal antibodies (Mabs) has been instrumental to further describe their migration and differentiation steps. The HNK1/NC1 Mab (Abo and Balch, 1981; Vincent et al., 1983) that reacts with chick (but not quail) melanoblasts at early stages, allowed the migration of presumptive melanocytes to be followed up to embryonic day 4.5 (E4.5) (Erickson et al., 1992). In toto labelling showed that emigration of chick melanoblasts from the dorsal neuroepithelium to the so-called 'staging area' is not segmented. The HNK1-positive cells do not enter the lateral space between ectoderm and somites before E3.5. Antigenic markers have been discovered in the past decade that are expressed by pigment cell precursors at early stages of differentiation (reviewed by Lecoin et al., 1998a). In avian embryos, MEBL-1 is the most precocious marker since it is expressed from E3 in chick NC cells emigrating from the neural tube along the dorsal path (Kitamura et al., 1992). Nataf et al. (1993) have produced several Mabs against melanocytes from quail NC cultures; Mel1 and Mel2 recognize melanosomal antigens in pigment cells from E8. The melanoblast/cyte early marker (MelEM) Mab produced by Nataf et al. (1993) is specific for NC-derived melanocytes (it does not react with the retinal pigmented epithelium) and labels unpigmented melanoblasts in NC cultures and in vivo from the stage of mediolateral migration.

Segregation of melanocyte precursors from pluripotent NC cells

The multiple cell types yielded by the NC raised the question of how and when subsets of NC-derived cells enter the pigment cell lineage. Experiments investigating the developmental potentials of individual cells in vivo and in vitro were developed to address the question as to whether melanocytes derive from multipotent or from lineage-restricted cells. The first evidence for a dual origin of melanocytes came from the first in vitro colony assay of NC cells devised by Cohen and Konigsberg (1975). They found that colonies formed by quail NC cells either contained only pigment cells or only unpigmented cells, or both (Cohen and Konigsberg, 1975). Unpigmented cells in mixed colonies were later shown to include several types of neurons, thus revealing a common precursor for neural cells and melanocytes in trunk NC (Sieber-Blum and Cohen, 1980; Sieber-Blum, 1989).

The developmental potentials of NC cells in vitro was further studied by using single-cell plating under microscopic control and culture on a feeder layer of growth-arrested 3T3 fibroblasts (Baroffio et al., 1988). The analysis with a panel of lineage markers of the clones from cephalic migratory NC cells provided evidence for heterogenous cell proliferation and differentiation potentials. Most progenitors gave rise to two to five different cell types with various combinations of phenotypes, whereas others generated only glial cells or neurons. Some clones contained all the main phenotypes derived from the cephalic NC including cartilage, a mesectodermal derivative, and were therefore considered to derive from a totipotent precursor (Baroffio et al., 1988, 1991; Dupin et al., 1990). In these experiments, pigment cells were thus generated from multipotent neural-melanocytic progenitors or bipotent glial-melanocytic precursors (Figure 1). Labelling with the MelEM Mab identified unpigmented melanoblasts in multiphenotypic colonies derived from cephalic and trunk NC cells (Dupin and Le Douarin, 1995). Taken together, these results suggest that NC precursors are lineally related, thus supporting the hypothesis that, as cells divide during migration, progressive restriction takes place in the developmental options of progenitors generated from an initial pool of multipotent stem cells.

Figure 1.

Model for the segregation of cell lineages derived from the cephalic NC. This diagram illustrates the different progenitors identified from quail mesencephalic NC cells grown in clonal cultures. The presence of cartilage (C), neurons (N), glial cells (G), adrenergic cells (A) and melanocytes (M) was recorded in the colonies. Progenitors have been classified according to the number of cell phenotypes in their progeny. Those endowed with a melanocytic developmental potential are shown in grey. The results are consistent with the generation of unipotent progenitors from a 'totipotent' stem-like cell through several intermediate oligopotent precursors. Filiations between precursors are only hypothetical (adapted from Baroffio et al., 1991)

Full figure and legend (48K)

Pluripotency of NC cells was confirmed for mammalian cells by in vitro clonal analysis. In the rat, trunk NC cells give rise to neurons, glial cells and fibroblasts and are able to self-maintain (Stemple and Anderson, 1992). Common precursors for melanocytes and neurons developed in colonies from NC cells isolated from mouse embryos at day 9.5 of gestation (Ito et al., 1993).

Consistent with the results of in vitro clonal cultures, in vivo experiments tracing the fate of single NC cells in the chick embryo provided evidence that at premigratory and early migratory stages, the NC cell population includes pluripotent precursors (Bronner-Fraser and Fraser, 1988,1989). The progeny of NC cells labelled after injection with a fluorescent dye populated distinct derivatives in many cases and include cell types as diverse as sensory neurons, adrenomedullary cells and presumptive melanocytic cells located in the skin.

These results raised the possibility that multipotent cells may be present at later stages of NC development. The differentiation capacities of cells en route or present in NC derivatives were therefore investigated in single-cell cultures. These experiments consistently showed loss of differentiation options as development proceeds. However, they also revealed that pluripotent precursors still exist at advanced stages of migration and even in NC derivatives. Thus, NC cells in the epidermis, which are normally fated to become melanocytes, retain multiple developmental potentials at early stages and are able to give rise to neurons in vitro (Richardson and Sieber-Blum, 1993). However, in these experiments, quail epidermal NC-derived cells expressing nonmelanogenic options were no longer evidenced from E6 onward.

Taken together, these results show that NC cells are heterogeneous in their developmental potentialities, including multipotent stem cell-like progenitors as well as committed precursors. Therefore, specific environmental cues are required to promote the expression of the phenotypes adopted in each NC derivative. Knowing the factors that influence the expression of developmental potentialities of NC cells locally and are responsible for final fate decisions is of great importance for understanding the molecular mechanisms underlying NC lineage segregation.

Role of endothelin 3 in the onset and maintenance of the melanocyte phenotype

Characterizing environmental signals provided by embryonic structures, extracellular matrix molecules and/or growth factors on the phenotypic choice of NC cells remains a major issue in understanding NC cell differentiation (for reviews, see Anderson, 1997; Le Douarin and Kalcheim, 1999). Mouse genetics and particularly the study of white spotting mutations proved to be very useful for the knowledge of the factors that influence the development of melanocytic precursors (Jackson, 1994).

Two mouse spotting mutants, white dominant spotting (w) and Steel (sl), have been analysed in detail. They are embryonic lethals as homozygotes, whereas heterozygote embryos show clear defects in melanocytes, hematopoietic cells, germ cells and pacemaker cells of the intestine. The w and sl loci encode the receptor tyrosine kinase c-kit and its cognate ligand, steel factor or stem cell factor (SCF), respectively (for references, Reith and Bernstein, 1991; Williams et al., 1992; Galli et al., 1993). A number of in vivo and in vitro approaches have been used to pinpoint the critical period when SCF/c-kit signalling is required for survival and migration of pigment precursors (reviewed by Yoshida et al., 2001). In utero injection of c-kit blocking antibody has revealed that c-kit is needed during a restricted period when mouse melanocyte precursors invade the epidermis from the dermis (Nishikawa et al., 1991; Yoshida et al., 1993). Analysis of different mutations of Sl indicated that soluble SCF is required for melanoblast early survival or dispersal, whereas membrane-bound SCF promotes melanoblast survival in the dermis (Wehrle-Haller and Weston, 1995). Consistent with these inferences, studies of mouse and avian NC melanogenesis in vitro suggest that SCF promotes survival and moderate proliferation of melanocytes and their precursors (Murphy et al., 1992; Morrison-Graham and Weston, 1993; Lahav et al., 1994; Reid et al., 1995).

Endothelins have been identified recently as crucial factors in the development of subsets of NC cell types, including melanocytes. Endothelins 1, 2 and 3 (EDN1, EDN2 and EDN3) are a family of 21 aa vasoactive peptides, which are synthesized by proteolytic cleavage from larger precursors and which bind to G-protein-coupled heptahelical receptors (Yanagisawa et al., 1988). They are of two types in mammals, endothelin receptors A and B (EDNRA and EDNRB). EDNRA displays a preferential affinity for EDN1, whereas EDNRB accepts all the three peptides equally (Sakurai et al., 1992 for a review).

The function of EDN3 and EDNRB in pigment cell development was discovered when knockout mice have been generated. Targeted mutation of the EDNRB gene as well as the spontaneous piebald-lethal (s^l) mouse mutant are characterized by coat colour spots and megacolon. Similar defects result from targeted disruption or spontaneous mutation (lethal spotting, ls) of the EDN3 gene (Hosoda et al., 1994; Greenstein-Baynash et al., 1994). The EDN3/EDNRB ligand/receptor is therefore crucial for the development of two NC-derived lineages: melanocytes and enteric nerve cells. Analysis of mutant embryos further indicates that migration of early melanoblasts in the dermis requires both c-kit and EDN3 (Yoshida et al., 1996). Conditional mutation in mouse revealed that requirement for functional EDNRB corresponds to a restricted period of melanoblast early migration (Shin et al., 1999), which precedes the requirement for c-kit/SCF function (Pavan and Tilghman, 1994).

To address the issues of the timing and mechanisms of EDN3 function during melanogenesis, we have studied the expression patterns of the different genes encoding EDNRs and their ligands in the avian embryo. The results are summarized in Figure 2. EDNRB starts to be expressed by NC cells still in the neural folds and may therefore be acting before cells enter the migration path to the skin (Nataf et al., 1996). At a later stage, only cells that migrate dorsoventrally and later on, their neural derivatives, express EDNRB whereas melanoblasts and melanocytes do not. We have cloned a third type of EDNR receptor in the quail, referred to as EDNRB2, since it binds the three types of endothelins with equal affinity (Lecoin et al., 1998b). EDNRB2 is expressed by early melanoblasts, as soon as they enter the medio-lateral migration pathway and later on, by differentiated melanocytes and is not present in NC cells migrating dorsoventrally. Thus, EDNRB2 and EDNRB display complementary expression patterns in avian NC cells and their derivatives (Figure 2). The finding that EDN3 is expressed in the environment in which melanoblasts migrate, that is, by the ectoderm from E3 and later by the epidermis (Nataf et al., 1998) argues for a paracrine action exerted by EDN3 via EDNRB2, on NC cells during migration in the skin and melanogenic differentiation.

Figure 2.

Complementary expression of EDNRs in avian migratory NC cells. Expression patterns of EDNRB (a, c, e, g) and EDNRB2 (b, d, f, h) in quail embryos sectioned transversally and in situ hybridized with digoxygenin-labelled riboprobes. The results are summarized on the left. At E2, NC that migrate along the dorsoventral pathway (1) express EDNRB (a, c). At E3 and E4, there is no EDNRB labelling in cells located in the skin (e, g). NC cells along the mediolateral pathway (2) express EDNRB2 from E3 (f) and become more numerous at E4 (h). D, dermomyotome; Sc, sclerotome; NC, notochord; Ao, dorsal aorta; Mel, melanocytes; DRG, dorsal root ganglia; Sy.G, sympathetic ganglia; Ag, adrenal gland (from Le Douarin and Kalcheim, 1999)

Full figure and legend (120K)

The action of EDN3 on quail NC cells was investigated using in vitro cultures. Addition of EDN3 promoted a dose-dependent increase of NC cell division and delayed the onset of melanogenesis. Numerous melanocytes differentiated after 10 days of culture in the presence of EDN3 and distributed in a characteristic pattern of pigmented and unpigmented cells (Lahav et al., 1996) (Figure 3a–c). In mouse NC cultures, EDN3 acts in conjunction with SCF in regulating the survival and proliferation of melanocyte progenitors (Reid et al., 1996).

Figure 3.

Effects of EDN3 on the behavior of quail trunk NC cells in vitro. (a) Quantification of the NC total cell number at different time points of a 16-day culture period in control medium and in the presence of 100 nM EDN3. (b, c) Effect of EDN3 on melanogenesis: after 11 days, control cultures show dispersed melanocytes (b), whereas EDN3-treated cultures display reproducible pattern of pigment cells and unpigmented premelanocytes (c) (from Lahav et al., 1996). (d) Diagram illustrating the results of in vitro clonal analysis of NC cells (Lahav et al., 1998): addition of EDN3 increases the survival of unipotent and bipotent precursors for glial and melanocytic cells

Full figure and legend (91K)

In an attempt to identify the potential target of EDN3, we examined the developmental potentials of trunk NC cells that respond to EDN3 in clonal cultures. Six different types of progenitors were defined by the various combinations of cell phenotypes in the clones. These progenitors included multipotent, bipotent and committed cells as previously described (Dupin and Le Douarin, 1995). However, EDN3 was found to highly enhance the survival of only three types of clonogenic cells, those generating both melanocytes and glial cells, and those yielding melanocytes only and glial cells only (Figure 3d). EDN3 also promoted a large increase of the total cell number in the colonies. The increase in cell survival and proliferation by EDN3 was stronger on melanogenic than on glial precursors, which explains that pigment cell expansion is the most prominent EDN3 effect. Therefore, prolonged treatment of NC cells with EDN3 exerts a selective positive effect on melanogenic and glial progenitors, including bipotent precursors (Lahav et al., 1998). However, if quail NC cells are treated with EDN3 only transiently, another target of EDN3 can be evidenced, that is a multipotent progenitor giving rise to cells of the melanocytic, adrenergic and sensory-like neuronal lineages (Stone et al., 1997)

The multiple effects of EDN3 on avian NC cells are mediated through EDNRB and EDNRB2. EDN3 increases the number of EDNRB-expressing cells in NC cultures only transiently, whereas it expands the population of EDNRB2-expressing melanocytic cells during the whole culture period, resulting in decline of EDNRB and induction of EDNRB2 (Lahav et al., 1998). This is in agreement with the dynamic expression pattern of EDNRB and EDNRB2 in vivo, which shows first, expression of EDNRB by premigratory and early migratory NC cells and later, switch off of EDNRB and onset of EDNRB2 in cells that migrate mediolaterally to the skin. As ectodermal and epidermal tissues express EDN3 transcripts, developing EDNRB2-positive NC cells remain in close contact with a source of ligand, therefore suggesting that EDN3 in the skin promotes survival and proliferation of melanoblasts and melanocytes via activation of EDNRB2.

Since melanocytes continue to strongly express EDNRB2 while differentiating in the epidermis, the question was raised as to whether EDN3 influences the terminal differentiation and maintenance of differentiated pigment cells. We have thus tested the effect of EDN3 on the behaviour of pigment cells isolated from the quail embryonic skin. Addition of EDN3 triggered high proliferative activity in melanocytes grown in clonal cultures. Dividing pigment cells progressively loose melanogenic traits and, under prolonged treatment with EDN3, generated large colonies of mixed phenotypes including melanocytes and cells expressing the glial-specific Po and Schwann cell myelin proteins (Dupin et al., 2000) (Figure 4).These results show that EDN3 is able to induce transdifferentiation of embryonic pigment cells to glia. Epidermal pigment cells in vitro thus reverse to the bipotent stage of glial-melanocytic NC progenitors, which suggests that strong proliferative signals such as EDN3 are able to alter the stability of the pigment cell phenotype.

Figure 4.

Transdifferentiation in vitro of epidermal pigment cells to glia. (a, bright field) Individual pigment cells isolated from E7 quail epidermis after 2 h of plating. (b, bright field) After 13 days of culture in the presence of 50 nM EDN3, single pigment cells have generated large colonies showing a dense core of pigment cells. (c, fluorescence microscopy) The colonies include both MelEM+ melanoblasts (red) and cells expressing the glial marker SMP (green), indicating a mixed progeny. The distinct steps of this transdifferentiation process are summarized on the right: proliferation of pigment cells by EDN3 led to progressive melanocyte dedifferentiation followed by induction of glial proteins Po and SMP. Bars, 100 m in a and c, 1 mm in b

Full figure and legend (70K)

These data also suggest that peripheral glial cells, which are derived from bipotent glial-melanocytic NC progenitors and which express EDNRB, could also respond to EDN3 by phenotypic alteration. This possibility is supported by previous data showing that melanogenesis can be induced in avian peripheral nerve cultures, although the precise origin of melanocytes (from glial cell conversion or from undifferentiated cells) was not determined (Nataf and Le Douarin, 2000). To address the question as to whether Schwann cells are capable to convert to melanocytes, we recently submitted Schwann cells purified from embryonic sciatic nerves to EDN3 in vitro. A mixed progeny of glial cells and melanocytes could be obtained in clonal cultures treated with EDN3 (Dupin et al., 2003), therefore arguing that glial cells can recapitulate in vitro the bipotency of their immediate ancestor upstream in the NC lineage tree, as do embryonic pigment cells.

Concluding remarks

Molecular and functional analysis of the genes in conjunction with in vitro cellular approaches led to define some of the regulatory pathways that control the early stages of pigment cell development from the NC. Cell diversification in the NC appears to be generated by progressive restrictions in the developmental potencies of an initial pool of stem cells, which yield more restricted intermediate progenitors and diversified committed cells. This process is largely governed by the action of external cytokines on particular target progenitors.

The selection and expansion of migrating melanogenic precursors depend on the local action of EDN3 on bipotent glial-melanocytic cells and melanocytic precursors. In addition, the immature bipotent stage of this common ancestor can be recapitulated in vitro by differentiated pigment and glial cells when they are induced to proliferate in the presence of high doses of EDN3. This suggests that reversion to more immature stages upstream in the NC lineage tree might be a general property of NC-derived cells. Reciprocal conversion between pigment and glial phenotypes in vitro thus provides further evidence for the plasticity of NC cell fate and suggests that EDN3, owing to its strong proliferative activity on NC precursors and NC-derived cells, could play a role in vivo in several pathologies implicating both glia and melanocytes.

References

1. Abo T and Balch CM. (1981). J Immunol., 127, 1024–1029. | PubMed | ISI | ChemPort |
2. Anderson DJ. (1997). Trends Genet., 13, 276–280. | Article | PubMed | ISI | ChemPort |
3. Baroffio A, Dupin E and Le Douarin NM. (1988). Proc. Natl. Acad. Sci. USA, 85, 5325–5329. | Article | PubMed | ChemPort |
4. Baroffio A, Dupin E and Le Douarin NM. (1991). Development, 112, 301–305. | PubMed | ChemPort |
5. Bronner-Fraser M and Fraser SE. (1988). Nature, 335, 161–164. | Article | PubMed | ChemPort |
6. Bronner-Fraser M and Fraser SE. (1989). Neuron, 3, 755–766. | Article | PubMed | ChemPort |
7. Cohen AM and Konigsberg IR. (1975). Dev. Biol., 46, 262–280. | Article | PubMed | ChemPort |
8. Couly GF, Coltey PM and Le Douarin NM. (1993). Development, 117, 409–429. | PubMed | ISI | ChemPort |
9. Dupin E, Glavieux C, Vaigot P and Le Douarin NM. (2000). Proc. Natl. Acad. Sci. USA, 97, 7882–7887. | Article | PubMed | ChemPort |
10. Dupin E, Real C, Glavieux-Pardanaud C, Vaigot P and Le Douarin NM. (2003). Proc. Natl. Acad. Sci. USA, Apr. 17; [epub ahead of print].
11. Dupin E and Le Douarin NM. (1995). Dev. Biol., 168, 529–548.
12. Erickson CA, Duong TD and Tosney KW. (1992). Dev. Biol., 151, 251–272. | Article | PubMed | ISI | ChemPort |
13. Etchevers HC, Vincent C, Le Douarin NM and Couly GF. (2001). Development, 128, 1059–1068. | PubMed | ISI | ChemPort |
14. Galli SJ, Zsebo KM and Geissler EN. (1993). Adv. Immunol., 55, 1–96. | ISI |
15. Greenstein-Baynash A, Hosoda K, Giaid A, Richardson JA, Emoto N, Hammer RE and Yanagisawa M. (1994). Cell, 79, 1277–1285. | Article | PubMed | ISI | ChemPort |
16. Hosoda K, Hammer RE, Richardson JA, Greenstein-Baynash A, Cheung JC, Giaid A and Yanagisawa M. (1994). Cell, 79, 1267–1276. | Article | PubMed | ISI | ChemPort |
17. Ito K, Morita T and Sieber-Blum M. (1993). Dev. Biol., 157, 517–525. | Article | PubMed | ISI | ChemPort |
18. Jackson J. (1994). Ann. Rev. Genet., 28, 189–217.
19. Kitamura K, Takiguchi-Hayashi K, Sezaki M, Yamamoto H and Takeuchi T. (1992). Development, 114, 367–378.
20. Lahav R, Dupin E, Lecoin L, Glavieux C, Champeval D, Ziller C and Le Douarin NM. (1998). Proc. Natl. Acad. Sci. USA, 95, 14214–14219. | Article | PubMed | ChemPort |
21. Lahav R, Lecoin L, Ziller C, Nataf V, Carnahan JF, Martin FH and Le Douarin NM. (1994). Differentiation, 58, 133–139. | PubMed |
22. Lahav R, Ziller C, Dupin E and Le Douarin NM. (1996). Proc. Natl. Acad. Sci. USA, 93, 3892–3897. | Article | PubMed | ChemPort |
23. Lecoin L, Lahav R, Dupin E and Le Douarin NM. (1998a). Molecular Basis of Epithelial Appendage Morphogenesis. CM Chuong (ed). RG Landes Company: Georgetown, USA, pp. 131–154.
24. Lecoin L, Sakurai T, Ngo MT, Abe Y, Yanagisawa M and Le Douarin NM. (1998b). Proc. Natl. Acad. Sci. USA, 95, 3024–3029. | PubMed |
25. Le Douarin N. (1982). The Neural Crest. Cambridge University Press: Cambridge, 259pp.
26. Le Douarin NM and Kalcheim C. (1999). The Neural Crest. 2nd ed. Cambridge University Press: New York, 445pp.
27. Morrison-Graham K and Weston JA. (1993). Dev. Biol., 159, 346–352. | Article | PubMed | ChemPort |
28. Murphy M, Reid K, Williams DE, Lyman SD and Bartlett PF. (1992). Dev. Biol., 153, 396–401. | Article | PubMed | ISI | ChemPort |
29. Nataf V, Amemiya A, Yanagisawa M and Le Douarin NM. (1998). Mech. Dev., 73, 217–220.
30. Nataf V, Lecoin L, Eichmann A and Le Douarin NM. (1996). Proc. Natl. Acad. Sci. USA, 93, 9645–9650. | Article | PubMed | ChemPort |
31. Nataf V and Le Douarin NM. (2000). Pigment Cell Res., 13, 172–178.
32. Nataf V, Mercier P, Ziller C and Le Douarin NM. (1993). Exp. Cell Res., 207, 171–182. | PubMed |
33. Nishikawa S, Kusakabe M, Yoshinaga K, Ogawa M, Hayashi S, Kunisada T, Era T, Sakakura T and Nishikawa SI. (1991). EMBO J., 10, 2111–2118. | PubMed | ISI | ChemPort |
34. Pavan WJ and Tilghman SM. (1994). Proc. Natl. Acad. Sci. USA, 91, 7159–7163. | Article | PubMed | ChemPort |
35. Reid K, Nishikawa SI, Bartlett PF and Murphy M. (1995). Dev. Biol., 169, 568–579. | Article | PubMed | ISI | ChemPort |
36. Reid K, Turnley AM, Maxwell GD, Kurihara Y, Kurihara H, Bartlett PF and Murphy M. (1996). Development, 122, 3911–3919. | PubMed | ChemPort |
37. Reith AD and Bernstein A. (1991). Genome Analysis. Vol. 3: Genes and Phenotypes. Cold Spring Harbor Laboratory Press: New York, pp. 10–133.
38. Richardson MK and Sieber-Blum M. (1993). Dev. Biol., 157, 348–358.
39. Sakurai T, Yanagisawa M and Masaki T. (1992). Trends Pharmacol. Sci., 13, 103–108. | Article | PubMed | ISI | ChemPort |
40. Shin MK, Levorse JM, Ingram RS and Tilghman SM. (1999). Nature, 402, 496–501. | Article | PubMed | ISI | ChemPort |
41. Sieber-Blum M. (1989). Science, 243, 1608–1611. | PubMed | ChemPort |
42. Sieber-Blum M and Cohen AM. (1980). Dev. Biol., 80, 96–106.
43. Stemple DL and Anderson DJ. (1992). Cell, 71, 973–985. | Article | PubMed | ISI | ChemPort |
44. Stone JG, Spirling LI and Richardson MK. (1997). J. Cell Sci., 110, 1673–1682.
45. Teillet MA. (1971). Ann. Embryol. Morphogenèse, 4, 95–109.
46. Teillet MA and Le Douarin NM. (1970). C. R. Acad. Sci., Ser. III (Paris), 270, 3095–3098.
47. Vincent M, Duband JL and Thiery JP. (1983). Brain Res., 285, 235–238.
48. Wehrle-Haller B and Weston JA. (1995). Development, 121, 731–742. | PubMed | ChemPort |
49. Williams DE, de Vries P, Namen AE, Widmer MB and Lyman SD. (1992). Dev. Biol., 151, 368–376. | Article | PubMed | ISI | ChemPort |
50. Yanagisawa M, Kurihara H, Kimura S, Tomobe Y, Kobayashi M, Mitsui Y, Yazaki Y, Goto K and Masaki T. (1988). Nature, 332, 411–415. | Article | PubMed | ISI | ChemPort |
51. Yoshida H, Kunisada T, Grimm T, Nishimura EK, Nishioka E and Nishikawa SI. (2001). J. Investig. Dermatol. Symp. Proc., 6, 1–5. | Article | PubMed | ISI | ChemPort |
52. Yoshida H, Kunisada T, Kusakabe M, Nishikawa S and Nishikawa SI. (1996). Development, 122, 1207–1214. | PubMed | ISI | ChemPort |
53. Yoshida H, Nishikawa S, Okamura H, Sakakura T and Kusakabe M. (1993). Dev. Growth Differ., 35, 209–220. | Article | ISI | ChemPort |

Acknowledgements

This work was supported by the Centre National pour la Recherche Scientifique and by a grant from the Association pour la Recherche contre le Cancer (no. 5578).