Reversal of developmental restrictions in neural crest lineages: Transition from Schwann cells to glial-melanocytic precursors in vitro

Transcription

1 Reversal of developmental restrictions in neural crest lineages: Transition from Schwann cells to glial-melanocytic precursors in vitro Elisabeth Dupin, Carla Real, Corinne Glavieux-Pardanaud, Pierre Vaigot, and Nicole M. Le Douarin* Laboratoire d Embryologie Cellulaire et Moléculaire, Centre National de la Recherche Scientifique, Unité Mixte de Recherche 7128, 49 Bis Avenue Belle Gabrielle, Nogent-sur-Marne Cedex, France Contributed by Nicole M. Le Douarin, March 3, 2003 In vertebrate embryos, diversification of the lineages arising from the neural crest (NC) is controlled to a large extent by environmental factors. In previous work, we showed that endothelin 3 (ET3) peptide favors the development of glial and melanocytic NC precursors in vitro. This factor is also capable of inducing proliferation of cultured epidermal pigment cells and their conversion to glia. ET3 therefore strongly promotes the emergence of melanocytic and glial phenotypes from precursors and acts on the maintenance of these phenotypes. In the present work, we explored the capacity of ET3 to reprogram glial cells into melanocytes. Schwann cells expressing glial-specific markers [such as the Schwann cell myelin protein (SMP)] were isolated from sciatic nerves of quail embryos and cultured in vitro. We found that ET3 promotes cell growth and sequential expression of melanocyte differentiation markers in cultures of purified SMP-expressing cells, whereas it had no significant effect on SMP-negative cells from the same nerves. Moreover, we provide evidence for the transition of differentiated Schwann cells to melanocytes in clonal cultures. This transition involves the production of a mixed progeny of melanoblasts melanocytes, glia, and cells bearing differentiation markers of both phenotypes. Therefore, Schwann cells exposed to ET3 transdifferentiate to melanocytes through reversion to the stage of bipotent glial-melanocytic NC precursors. These findings show that NC-derived pigment and glial cells are phenotypically unstable in vitro and may undergo reversal of precursor hierarchy to function as bipotent stem cells. melanocytes quail embryo clonal cultures endothelin 3 During vertebrate embryogenesis, neural crest (NC) cells arise from the dorsal aspect of the closing neural tube and undergo extensive migration along distinct pathways as they proliferate. The NC gives rise to a wide range of cell types, including neurons and glia of the peripheral nervous system (PNS), melanocytes, craniofacial bones and cartilage, fibroblasts, smooth muscle cells, and endocrine cells (1, 2). Migrating NC cells are a heterogeneous collection of progenitors, including multipotent stem cells, cells with limited developmental potential and fate-restricted precursors (3 11). The influence of environmental signals is therefore crucial to ensure that postmigratory NC-derived cells express a defined phenotype corresponding to their position in the body (for reviews, see refs. 1, 2, and 12). Among such signals, endothelin 3 (ET3) signaling through endothelin receptor (ETR) B was shown in mouse and human to be required in vivo for proper development of melanocytes and posterior enteric nerve cells (13 16). In vitro culture studies have revealed that ET3 exerts trophic and mitogenic activities on early avian and mammalian NC cells (17 20). The main target cells of ET3 have been identified (18) and include faterestricted glial and melanocytic cells as well as a bipotent glialmelanocytic (GM) precursor previously evidenced in quail NC clonal cultures (7). Enhanced melanogenesis by quail NC cells treated with ET3 is coupled to switching from expression of ETRB to that of ETRB2 (18), which was characterized as the melanocytic lineage-specific form of ETR in avians (21). At later stages of embryogenesis, NC-derived pigment cells in the epidermis express ETRB2. If transferred in culture, these differentiated melanocytes respond to ET3 by increased proliferation rate and change in their developmental fate. Pigment cells in clonal cultures treated with ET3 acquire the ability to generate both melanocytes and glial cells, thus recovering the GM bipotentiality of their NC ancestor (22). Taken together, these data point to a key role for ET3 in regulating cell cycling and differentiation at different commitment levels in NC-derived glial and melanocytic lineages. The finding that developmental restrictions can be reversed by ET3 raises the possibility that PNS glial cells, which express ETRB (23), might be induced by ET3 to adopt the alternative melanocytic phenotype. Further support for this notion came from previous studies that revealed a melanogenic potential in early embryonic avian dorsal root ganglia and peripheral nerve (PN) explant cultures treated with phorbol ester and basic fibroblast growth factor (24 26). More recently, it was shown that ET3 triggers pigment cell differentiation in dissociated cell cultures of embryonic day 10 (E10) to E14 quail sciatic PN, which consist mostly of differentiated Schwann cells (27). These data suggest that Schwann cells can transdifferentiate to melanocytes in vitro, although the possibility remains that undifferentiated precursors present in PN could be at the origin of de novo produced melanocytes. In the present work, we have examined the effects of ET3 on purified Schwann and non-schwann PN cell subpopulations cultured independently and as single cells to determine from which PN cell type melanocytes can emerge. We show that melanogenesis in PN cultures results from the transdifferentiation of Schwann cells coupled to proliferation and reversion to the bipotent GM stage. Materials and Methods Isolation of Embryonic PN and Schwann Cells. The sciatic PN were dissected from E10.5 quails and the perineural sheets were carefully removed before dissociation to single cells as described (27). PN cells were collected in DMEM containing 3% FCS (Dutscher, Brumath, France) and labeled with mab against Schwann cell myelin protein (SMP), a surface glycoprotein expressed by avian PN Schwann cells (28). The cell suspension was incubated with anti-smp ascitic fluid (1:250 for 30 min) followed by phycoerythrin-conjugated goat anti-mouse IgG1 (1:400; Southern Biotechnology Associates). Cells (10 6 per ml in PBS) were filtered through a nylon membrane (25- m pore diameter; VWR International) before cell sorting by using a fluorescence-activated cell sorter (FACStar Plus; Becton Dickinson). Control samples included cells treated as described above except that anti-smp was omitted. Purified fractions of SMPpositive (SMP ) and SMP-negative (SMP ) cells were isolated according to fluorescence intensity. Abbreviations: ET3, endothelin 3; ETR, endothelin receptor; NC, neural crest; dn, day n of culture; SMP, Schwann cell myelin protein; MelEM, melanoblast melanocyte early marker; PN, peripheral nerve; PNS, PN system; GM, glial-melanocytic; En, embryonic day n. *To whom correspondence should be addressed. nicole.le-douarin@collegede-france.fr. cgi doi pnas PNAS April 29, 2003 vol. 100 no

2 Mass and Clonal Cultures of Purified Subpopulations of PN Cells. For mass cultures, sorted SMP and SMP PN cells ( per 20 l of culture medium) were plated at the center of 35-mm culture dishes or four-well culture plates (Nunc). Clonal cultures of SMP PN cells were performed by micromanipulation as described for NC cells (3, 4, 7, 18, 29). Single cells were seeded in 70- l drops of culture medium on 60-mm-diameter dishes (Nunc) previously coated with rat-tail collagen (Biomedical Technologies, Stoughton, MA). After overnight incubation, culture medium was added to the cultures and changed every 3 days thereafter. Culture medium included 10% FCS, 2% chicken embryo extract, and various hormones and growth factors as described (18, 22). When indicated, it was supplemented with ET3 (Sigma) at a 100-nM concentration, which triggered pigment cell differentiation in NC (17, 18) and PN cultures (27). Cultures were maintained at 37 C in an atmosphere of 5% CO 2 and 95% air. Analysis of the Cultures. Cell phenotypes were analyzed from day 1 of culture (d1) to d19. Pigment cells were recorded microscopically, and unpigmented cells were characterized by immunocytochemistry with several differentiation markers. Schwann cells were identified by using the anti-smp (7, 28) and antiprotein Po (30) mabs. Unpigmented melanocytes were detected with the melanoblast melanocyte early marker (MelEM) mab (31). When indicated, cultures were analyzed by staining with human natural killer 1 (HNK1) mab (32) and anti-p75 lowaffinity nerve growth factor receptor antibody (Chemicon). The HNK1 epitope and p75 protein are expressed by most NC cells and their derivatives in the PNS (8, 33). Antibody to 200-kDa neurofilament protein (Sigma) and mab to quail tyrosine hydroxylase (34) were used to identify neurons and adrenergic cells, respectively. Secondary antibodies were purchased from Southern Biotechnology Associates. Detailed immunofluorescence procedures were described elsewhere (18, 22). Fluorescence was observed with an Olympus (Melville, NY) 70 microscope. Expression of the ETRB2 gene (21) was studied by RNA in situ hybridization of the cultures according to Lahav et al. (18). The frequencies of the colonies were analyzed by 2 test (STATISTICA for Macintosh; Statsoft, Tulsa, OK), and differences between ET3-treated and control cultures were considered to be statistically significant when P Fig. 1. Cultures of PN cell subpopulations at d1. (A and B) Unsorted PN cells include elongated SMP and flat SMP cells (arrows). (C and D) The subpopulation of SMP Schwann cells was separated by fluorescence-activated cell sorting from SMP fibroblast-like cells (E and F). (A, C, and E) Phase contrast. (B, D, and F) Corresponding SMP fluorescence. (Bar 100 m.) Results Characterization of PN Cell Subpopulations. Cultures of PN cells from E10 14 quails undergo melanogenesis when treated with phorbol 12-myristate 13-acetate (PMA) or ET3 (27), suggesting that NC-derived glial cells can change their differentiation program in vitro. Previous experiments, however, did not address directly the issue of which cell type is able to generate melanocytes because the population of PN cells included, in addition to Schwann cells, fibroblasts and putative undifferentiated NC cells. To address this question, we have isolated Schwann cells by means of expression of the avian glial-specific membrane glycoprotein SMP (28). Dissociated cells from E10.5 quail sciatic nerves were labeled with anti-smp mab and analyzed by fluorescence-activated cell sorting (FACS) (Fig. 1). The fraction of sorted SMP cells accounted for % of total gated cells (mean SEM from 19 experiments) and was estimated by FACS reanalysis to be 98% pure. Unsorted and sorted PN cells were plated in control medium and, after 18 h of incubation, the cultures were analyzed for SMP expression (Fig. 1). Purified SMP cells tended to group as small cords or clusters of thin elongated cells (Fig. 1 C and D), whereas SMP cells formed a homogenous layer of fibroblast-like cells (Fig. 1 E and F). SMP cells expressed myelin protein Po and p75 low-affinity nerve growth factor receptor (data not shown) as expected for NCderived glial cells. None of these markers were detected in SMP cells, which hence were likely fibroblasts of non-nc origin. Effect of ET3 on Cultured SMP Schwann Cells. To identify ET3- responsive cells in embryonic PN, the isolated fractions of SMP and SMP cells were cultured in control medium with or without ET3 at a 100-nM concentration, a dose previously shown to highly promote melanogenesis by quail NC (18) and sciatic PN (27) cells. After d13 15, in the presence of ET3, pigment cells differentiated in cultures of SMP cells (Fig. 2 B, G, and H), whereas none were detected in cultures of SMP cells (data not shown). As compared with controls, addition of ET3 to SMP cell cultures also increased the overall cell growth (Fig. 2 A and B), which was not observed in SMP cell cultures. These data indicate that the SMP PN glial cell subpopulation responds to ET3 and is the source of cells with a melanogenic potential. However, as overt pigmentation occurs at the terminal stage of melanocyte differentiation, the possibility remained that unpigmented melanoblasts were present in cultures of both SMP and SMP PN cells. We therefore investigated whether the isolated PN cell subpopulations express in vitro early markers of avian melanocyte differentiation, that is, ETRB2 and MelEM. Such proteins are induced in melanocyte precursors from early stages of development in vivo and in vitro (18, 21, 31) and are not present in PNS glial cells, which instead express ETRB (23) cgi doi pnas Dupin et al.

3 Table 1. Effect of ET3 on SMP PN cells in clonal cultures Control ET3 P d1 clones 279 (100%) 265 (100%) d7 clones 212 (76%) 238 (90%) ns d13 clones 150 (56%) 201 (76%) 0.01 d17 clones 112 (40%) 184 (69%) Small Medium Large The total number of clones derived from SMP cells at d1, d7, d13, and d17 in control and ET3-supplemented medium is expressed as percent of d1 clonogenic cells. Counting the total number of cells per clone (n) at d17 allowed determination of the proportions of three categories of small (1 n 100), medium (100 n 1,000), and large (n 1,000) clones. P values are from statistical analysis of control and ET3-treated cultures; ns, not significant. Fig. 2. ET3 promotes melanocytic differentiation in SMP PN cultures. (A and B) At d15, control cultures form a sparse network of unpigmented cells (A), whereas dense ET3-treated cultures contain pigment cells (B). (C and D) Atd6, ET3-treated (C) but not untreated (D) cultures express ETRB2 mrna. (E and F) At d11, cultures exposed to ET3 include both SMP glial (E) and MelEM melanocytic (F) cells and some cells expressing the two markers (arrows). (G and H) At d17, in the presence of ET3, cultures show a high density of melanocytes (G) and glial cells (H). (A and B) Phase contrast. (C, D, and G) Bright field. MelEM (F) and SMP (E and H) fluorescence. (Bars 200 m ina and B, 100 m inc F, and 400 m ing and H.) We found that, at d1 and d4, neither ETRB2 nor MelEM markers were present in cultured SMP and SMP PN cells (data not shown). At d6, cultures were still negative for MelEM whatever the medium used; however, ETRB2 transcripts were detected in a subset of cells in SMP cultures treated with ET3 (Fig. 2 C and D). At d9, such cultures showed enhanced expression of ETRB2 and gave rise to MelEM-positive cells, the number of which increased at d11 (Fig. 2 E and F). At d17, SMP cell cultures exposed to ET3 yielded dense areas of pigment cells and glial cells (Fig. 2 G and H). SMP PN cell cultures analyzed from d9 to d19 did not express ETRB2 nor MelEM markers (data not shown), which argues that, in embryonic PN, the potential to yield melanocyte precursors exclusively belongs to the glial SMP cell compartment. Neurons and adrenergic cells were not detected in the cultures (data not shown) indicating that, at least in these culture conditions, PN cells are not able to adopt a neurogenic fate. Action of ET3 on the Clonal Progeny of SMP Schwann Cells. We further characterized the effects of ET3 on Schwann cells by clonal cultures. Eighteen hours after plating single SMP cells in control medium, the cloning efficiency was 65% (given by the number of d1 clonogenic cells from a total of 840 plated cells). Cultures were then fed with control or ET3-supplemented medium, and the clones were analyzed after various time periods for cell survival and growth. Table 1 shows that the total number of clones decreased between d1 and d17 in both media. However, when exposed to ET3, the clones exhibited a higher survival rate than in control medium, and, at d17, 69% of ET3-treated versus 40% of control clones were still alive. ET3 therefore significantly promotes long-term survival of colonies derived from SMP Schwann cells. We also investigated the effect of ET3 on clonal cell growth. d17 colonies were classified according to the total number of cells per clone (n) into small (1 n 100), medium (100 n 1,000), or large (n 1,000) colonies. The results (Table 1) show a highly significant increase in the number of medium and large colonies in ET3-treated compared with control cultures (61.4% versus 12.5% with n 100, respectively). To analyze the influence of ET3 on the developmental potential of glial PN cells, we examined d17 colonies for the presence of SMP glial cells, MelEM-expressing melanoblasts, and pigment cells. Five different types of clones were thus defined according to the marker(s) expressed (Table 2). In control medium, the large majority of the clones contained glial cells and neither melanoblasts nor pigment cells (Glia clones, Table 2), whereas only 4% of the cultures were unpigmented GM clones (containing both glial cells and melanoblasts). In the presence of ET3, pigmentation was induced in 10% of total clones, and the frequency of GM clones was raised to 52% (Table 2). GM clones contained a heterogeneous population of glial and melanocytic cells in variable proportions (Fig. 3A) and occasionally cells carrying both lineage markers (SMP MelEM cells) (Fig. 3 B D). A small subset of ET3-treated cultures was Table 2. Phenotypic analysis of clones from SMP PN cells Clone type Control ET3 P Glia 73 (67%) 60 (33%) GM pigm 4 (4%) 78 (42%) GM pigm 0 (0%) 19 (10%) Mel 0 (0%) 7 (4%) 0.04 Unidentified 32 (29%) 20 (11%) Cell phenotypes were analyzed at d17 in 119 control and 184 ET3-treated colonies. The total number (and percent of total clones) was determined for five different types of clones as defined by the presence of SMP glial cells, MelEM melanoblasts, and pigment cells: Glia clones (SMP, MelEM, pigm ), unpigmented and pigmented GM clones (SMP, MelEM ), Mel clones (MelEM, pigm, SMP ), and phenotypically unidentified clones (SMP, MelEM, pigm ). P values are from statistical analysis of control and ET3- treated cultures. Dupin et al. PNAS April 29, 2003 vol. 100 no

4 Fig. 3. Cell phenotypes in d17 clonal cultures of SMP PN cells. (A) Part of a GM clone after treatment with ET3, including SMP (green) and MelEM (red) cells. (B D) Detailed views of one cell immunoreactive for both SMP and MelEM markers. (B) SMP. (C) MelEM. (D) Merge of B and C fluorescence. (Bars 100 m ina and 50 m inb D.) Fig. 4. Model for the reversal of developmental restrictions by ET3. During ontogeny, differentiation of glia and melanocytes from pluripotent NC stem cells results from progressive developmental restrictions (black arrows) and involves an intermediate bipotent GM precursor that was identified by its clonal progeny but that has not yet been fully characterized. GM precursors (18), skin melanocytes (22), and PN glia (the present work) are the main targets of ET3 in vitro (orange circles). In response to ET3, melanocytes and glial cells proliferate and reverse to GM precursors (orange arrows). The reciprocal conversion between glia and melanocytes also produces a transient cell type expressing genes specific for both phenotypes (i.e., SMP MelEM cells). Whether GM precursors can be driven to reverse to precursors upstream in lineage hierarchy is still uncertain. pure melanocytic clones (Mel clones, Table 2). In both media, cells that express neither of the above markers formed the last type of colonies recorded (Unidentified clones, Table 2). Such cells, which were also negative for HNK1 epitope and protein Po (data not shown), were detected in most Glia, Mel, and GM clones, indicating that a subset of SMP PN cells tends to lose glial traits and becomes undifferentiated in long-term culture. Taken together, these data show that addition of ET3 to clonal cultures of Schwann cells promotes cell survival and growth and induces the production of a heterogeneous progeny in which unpigmented melanoblasts and eventually pigment cells differentiate. Discussion The NC is the source of a variety of cell types that are derived during ontogeny from precursors endowed with multiple potentialities (3 10). In previous studies, we have characterized in the quail NC a bipotent glial-melanocytic precursor, called GM, that responds to the cytokine ET3 (18). In this work on NC clonal cultures, a spectacular increase in cloning efficiency, growth, and differentiation was observed in GM precursors as well as in monopotent committed melanocytic progenitors. In contrast, ET3 had a modest activity on committed glial cells and no effect on the other NC precursors (18). ET3 appears, therefore, as a potent agent for inducing proliferation and differentiation of at least two NC-derived cell types: melanocytic and GM precursors. This effect is mediated in melanocytic cells through a receptor, ETRB2, discovered in birds by Lecoin et al. (21). When NC cells are in the neural fold and as they start migrating, they express a receptor showing 92% homology to human ETRB (23). This gene remains activated in the cells taking the dorsoventral pathway of migration. In contrast, NC cells that enter the developing skin switch from ETRB to ETRB2 (21, 23). These cells are subjected to ET3, which is produced by the avian ectoderm (35). The switch of the expression of ETRB to ETRB2 results from an induction by ET3, as shown by in vitro cultures where native ETRB-expressing NC cells rapidly switch to ETRB2 when treated with ET3 (18). In addition to its action on the emergence of melanocytes and glia from NC precursors, ET3 was shown to be capable of altering the fate of fully differentiated melanocytes. If they are treated in vitro with appropriate concentrations of ET3, pigment cells isolated from the quail epidermis can reverse to the bipotential GM precursor and give rise to a large clonal progeny containing glial and pigment cells together with cells carrying markers of both lineages (22). Taken together, these results identify GM NC precursors and NC-derived melanocytes as ET3-responsive cells. They also suggest that PNS glial cells, which like melanocytes derive from GM precursors, could be the target of ET3 and respond to this factor by changing their differentiation program. Previous studies have shown that explant cultures of dorsal root ganglia and PN from early avian embryos produce pigment cells if exposed to phorbol ester phorbol 12-myristate 13-acetate (PMA) and basic fibroblast growth factor (24 26, 36). Later, melanocyte differentiation was obtained from E10 14 quail PN whose Schwann cells expressing SMP were dissociated and cgi doi pnas Dupin et al.

5 cultured, provided that ET3 or PMA was added to the cultures. PMA and ET3 synergistically increased the yield of melanocyte differentiation (27). These studies therefore have revealed that exogenous factors such as ET3 can drive PN cells to express a melanogenic potential. However, they did not address the issue of whether melanogenesis by PN cells resulted from the transdifferentiation of already committed glial cells or from the differentiation of silent precursors present in the PN. To address this question directly and further document the action of ET3 on glial cells, we have, in the present work, explored the capacity of cloned Schwann cells from E10.5 quail PN to adopt the alternative melanocytic phenotype when exposed to ET3. The glial identity of the cells that we put in culture was attested to by the fact that they are selected on the basis of their expression of the surface antigen SMP present in the embryo exclusively on Schwann cells of the PN and oligodendrocytes in the central nervous system (28). We show that purified SMP glial cells respond to ET3 in vitro by enhanced survival and proliferation and they transition toward a melanocytic differentiation program. They gave rise to cells that sequentially express ETRB2 and MelEM and synthesize pigment, thus recapitulating the temporal sequence of melanocytic differentiation markers in the avian NC (21, 31). By contrast, cultured SMP PN cells, presumably fibroblasts, did not respond to ET3. Moreover, in single-cell cultures, SMP PN cells generated a mixed progeny of glial cells and pigment cells and or melanoblasts in a large proportion of the clones (52%). As in the case of epidermal pigment cell cultures (22), clonal efficiency, as well as the growth of the cells clonally generating glia and melanocytes, was promoted by ET3 in these cultures. Taken together, these results demonstrate that the terminal phenotypes resulting from the differentiation of GM precursors, i.e., glia and melanocytes, can be reversed to the bipotent stage through ET3-induced cell proliferation. A rare population of cells carrying the two protein markers MelEM and SMP was found in GM colonies derived from Schwann cells and melanocytes (22), which suggests that the reciprocal transition of glia and melanocytes involves a transient step where genes that characterize both phenotypes are simultaneously activated. Whether such a transient cell 1. Le Douarin, N. (1982) The Neural Crest (Cambridge Univ. Press, Cambridge, U.K.). 2. Le Douarin, N. M. & Kalcheim, C. (1999) The Neural Crest (Cambridge Univ. Press, New York), 2nd Ed. 3. Baroffio, A., Dupin, E. & Le Douarin, N. M. (1988) Proc. Natl. Acad. Sci. USA 85, Baroffio, A., Dupin, E. & Le Douarin, N. M. (1991) Development (Cambridge, U.K.) 112, Bronner-Fraser, M. & Fraser, S. E. (1988) Nature 335, Bronner-Fraser, M. & Fraser, S. E. (1989) Neuron 3, Dupin E., Baroffio, A., Dulac, C., Cameron-Curry, P. & Le Douarin, N. M. (1990) Proc. Natl. Acad. Sci. USA 87, Stemple, D. L. & Anderson, D. J. (1992) Cell 71, Henion, P. D. & Weston, J. A. (1997) Development (Cambridge, U.K.) 124, Morrison, S. J., White, P. M., Zock, C. & Anderson, D. J. (1999) Cell 96, Greenwood, A. L., Turner, E. E. & Anderson, D. J. (1999) Development (Cambridge, U.K.) 126, Anderson, D. J. (1997) Trends Genet. 13, Baynash, A. G., Hosoda, K., Giaid, A., Richardson, J. A., Emoto, N., Hammer, R. E. & Yanagisawa, M. (1994) Cell 79, Hosoda, K., Hammer, R. E., Richardson, J. A., Greenstein-Baynash, A., Cheung, J. C., Giaid, A. & Yanagisawa, M. (1994) Cell 79, Puffenberger, E. G., Hosoda, K., Washington, S. S., Nakao, K., Dewit, D., Yanagisawa, M. & Chakravarti, A. (1994) Cell 79, Shin, M. K., Levorse, J. M., Ingram, R. S. & Tilghman, S. M. (1999) Nature 402, Lahav, R., Ziller, C., Dupin, E. & Le Douarin, N. M. (1996) Proc. Natl. Acad. Sci. USA 93, Lahav, R., Dupin, E., Lecoin, L., Glavieux, C., Champeval, D., Ziller, C. & Le Douarin, N. M. (1998) Proc. Natl. Acad. Sci. USA 95, type is included in the normal ontogenetic process of NC development in vivo is doubtful. According to our own observations, it was never detected in the embryo. In fact, its presence may be due to the culture conditions, which turn out to be permissive for both glial and melanocytic phenotypes. In contrast, in vivo, the microenvironments in which NC cells are led to develop at the term of migration are different and strictly selective. Thus, NC cells are led to occupy specific niches where they can express definite sets of genes at the exclusion of others. When transferred to in vitro culture, some of their differentiation capacities that were repressed can then be disclosed. This implies that, in vivo, the NC cells that take the mediolateral skin migration pathway and are exposed to ET3 are at the same time inhibited to differentiate into glia. Similarly, the cells that follow the dorsoventral pathway or migrate into the gut where they are subjected to ET3 (35) are repressed to differentiate into melanocytes. Such an inhibition has been elegantly demonstrated by Rizvi et al. (37) in a transgenic mouse system where Schwann cells show the capacity to produce pigment cells if they migrate from a nerve that has been severely injured. This remarkable plasticity holds also for central nervous system glial cells: Committed rat oligodendrocyte precursors treated in culture with a definite set of factors can reverse to the pluripotent phenotype of bona fide neural stem cells (38). In conclusion, melanocytes and glial cells that are generated from NC pluripotent precursors exhibit a state of differentiation that turns out to be unstable if the cells are exposed in vitro to the potent mitogen ET3 (Fig. 4). This factor induces reciprocal conversion of these two phenotypes through reversion to GM precursors. Because no other NC-derived cell types have been observed so far in the descendants of melanocytes (22) or glial cells (the present work), we conclude that in these culture conditions, ET3 is not able to reverse the bipotent GM precursor to a more pluripotent NC progenitor. We thank F. Beaujean and S. Gournet for illustrations, and M. Scaglia for preparing the bibliography. This work was supported by the Centre National de la Recherche Scientifique, the Collège de France, and the Institut Curie. C.R. is recipient of a doctoral fellowship from Fundaçao para Ciencia e Tecnologia, Portugal (SFRH BD ). 19. Reid, K., Turnley, A. M., Maxwell, G. D., Kurihara, Y., Kurihara, H., Bartlett, P. F. & Murphy, M. (1996) Development (Cambridge, U.K.) 122, Stone, J. G., Spiraling, L. I. & Richardson, M. K. (1997) J. Cell Sci. 110, Lecoin, L., Sakurai, T., Ngo, M. T., Abe, Y., Yanagisawa, M. & Le Douarin, N. M. (1998) Proc. Natl. Acad. Sci. USA 95, Dupin, E., Glavieux, C., Vaigot, P. & Le Douarin, N. M. (2000) Proc. Natl. Acad. Sci. USA 97, Nataf, V., Lecoin, L., Eichmann, A. & Le Douarin, N. M. (1996) Proc. Natl. Acad. Sci. USA 93, Ciment, G. (1990) Comments Dev. Neurobiol. 1, Stocker, K. M., Sherman, L., Rees, S. & Ciment, G. (1991) Development (Cambridge, U.K.) 111, Sherman, L., Stocker, K. M., Morrison, R. & Ciment, G. (1993) Development (Cambridge, U.K.) 118, Nataf, V. & Le Douarin, N. M. (2000) Pigm. Cell Res. 13, Dulac, C., Cameron-Curry, P., Ziller, C. & Le Douarin, N. M. (1988) Neuron 1, Dupin, E. & Le Douarin, N. M. (1995) Dev. Biol. 168, Bhattacharyya, A., Frank, E., Ratner, N. & Brackenbury, R. (1991) Neuron 7, Nataf, V., Mercier, P., Ziller, C. & Le Douarin, N. M. (1993) Exp. Cell Res. 207, Abo, T. & Balch, C. M. (1981) J. Immunol. 127, Vincent, M. & Thiery, J. P. (1984) Dev. Biol. 103, Fauquet, M. & Ziller, C. (1989) J. Histochem. Cytochem. 37, Nataf, V., Amemiya, A., Yanagisawa, M. & Le Douarin, N. M. (1998) Mech. Dev. 73, Nichols, D. H. & Weston, J. A. (1977) Dev. Biol. 60, Rizvi, T. A., Huang, Y., Sidani, A., Atit, R., Largaespada, D. A., Boissy, R. E. & Ratner, N. (2002) J. Neurosci. 22, Kondo, T. & Raff, M. (2000) Science 289, Dupin et al. PNAS April 29, 2003 vol. 100 no