Enteric neural crest-derived cells: Origin, identification, migration, and differentiation
2001; Wiley; Volume: 262; Issue: 1 Linguagem: Inglês
10.1002/1097-0185(20010101)262
ISSN1097-0185
AutoresHeather M. Young, Don Newgreen,
Tópico(s)Intestinal Malrotation and Obstruction Disorders
ResumoNeurons and glial cells forming the enteric nervous system (ENS) arise from neural crest cells that migrate away from two different rostrocaudal levels of the neural axis, the vagal and sacral regions. The vagal region is defined as the post-otic hindbrain level with somites 1–7 (Le Douarin, 1982), and the sacral region is caudal to somite 28 in chick embryos and caudal to somite 24 in embryonic mice. Vagal level neural crest cells give rise to enteric neurons and glial cells throughout the entire gastrointestinal tract. Yntema and Hammond (1954) removed the neural crest and dorsal portion of the neural tube from defined rostrocaudal levels of the neural axis of chick embryos, and then approximately a week later, the embryos were fixed, sectioned, and silver stained to determine if enteric ganglia were present in the gastrointestinal tract. They found that enteric ganglia were absent from the esophagus, stomach, and small and large intestines if the lesions removed neural crest cells from a region commencing 4 somites width rostral to the first somite and extended caudally to the predicted level of the 10th somite. ENS deficits were not observed when any other levels of the neural crest were ablated (Yntema and Hammond, 1954, 1955; Hammond and Yntema, 1947). Thus, vagal level neural crest cells were deduced to be the only, or at least the major, source of the ENS throughout the entire gastrointestinal tract. These results were confirmed by Le Douarin and Teillet (1973, 1974) and Burns and Le Douarin (1998), who found that following replacement of the vagal neural axis, between somites 1–7, of chick embryos with the equivalent region from quail embryos, quail neurons and glial cells were found in ganglia throughout the gastrointestinal tract of the host embryo. More exact studies have shown that most neural cells in the mid- and hindgut of chick embryos are derived from the levels of somites 3–6 (Peters-van der Sanden et al., 1993; Epstein et al., 1994). There may be some species differences, since in mice neural crest cells adjacent to somites 1–4 appear to give rise to enteric neurons throughout the gut, whereas neural crest adjacent to somites 6–7 gives rise to a sub-population of neurons in the foregut, but no other gut region (Durbec et al., 1996; note that Durbec and colleagues define neural crest adjacent to somites 6–7 as "trunk" level neural crest since it also gives rise to dorsal root ganglia). Since Yntema and Hammond (1954) observed no neurons within the gut following ablation of the vagal-level neural crest, it appeared that the vagal neural crest was the sole source of enteric neurons. However, using chick-quail grafts, Le Douarin and Teillet (1973) showed that the sacral-level neural crest contributes some cells to enteric ganglia of the hindgut, although they did not ascertain with molecular markers if the sacral-derived cells were neurons and/or glial cells. In the following 25 years, different studies reached different conclusions as to the contribution of the sacral-level neural crest to the enteric nervous system. In both chick and mouse embryos in which pre-migratory sacral level neural crest cells were labelled with DiI or retroviruses, labelled cells were reported in the hindgut (Pomeranz et al., 1991; Serbedzija et al., 1991), at least 2 days prior to the arrival of vagal neural crest cells. However, variable results were obtained when segments of hindgut were removed prior to the arrival of vagal level cells and grown in culture, or explanted to the chorioallantoic membrane or kidney capsule; most studies reported a complete, or almost complete, absence of enteric neurons in the hindgut (Smith et al., 1977; Allan and Newgreen, 1980; Nishijima et al., 1990; Kapur et al., 1992; Lecoin et al., 1996; Young et al., 1996, 1998a), whereas other studies reported that neurons were present (Rothman and Gershon, 1982; Pomeranz and Gershon, 1990). Following transection of the midgut of chick embryos prior to the arrival of vagal neural crest cells, no enteric neurons were observed in the hindgut (Meijers et al., 1989). Thus, most of these experiments in which the vagal source was absent suggested that sacral neural crest cells do not colonize the hindgut, or that they require the presence of vagal cells in order to colonize the hindgut and form a recognizable ENS, or that they arrive simultaneous to or later than the vagal cells (Allan and Newgreen, 1980). The contribution of sacral level neural crest cells to the ENS in birds was resolved by a detailed study utilizing chick-quail chimeric grafting by Burns and Le Douarin (1998), in which they showed definitively that sacral cells contribute to the ENS in the hindgut. Importantly, they examined the time of arrival of sacral cells into the hindgut, and the phenotypes of the derivatives of sacral neural crest cells. Sacral cells were observed to migrate ventrally by E4.5 to form the pelvic plexus and nerve of Remak. The latter is an extension of the pelvic plexus in the form of a ganglionated chain, which lies in the mesentery adjacent to the dorsal border of the post-umbilical intestine; it is unique to birds. However, the sacral neural crest cells did not immediately continue their migration further ventrally from the nerve of Remak to enter the hindgut; instead, they entered the hindgut approximately 3 days later, around the time that the vagal cells arrived in the hindgut, at E7.5. Sacral cells were found to give rise to neurons and glial cells. Cells derived from the sacral neural crest gave rise to 17% of enteric neurons in the distal hindgut, with their contribution decreasing rostrally to 0.3% in the rostral hindgut. Several important questions remain unresolved about the contribution of sacral neural crest cells to the ENS of the hindgut. The first question is why the sacral cells pause for approximately 3 days in the nerve of Remak prior to entering the hindgut. The gut has been postulated to synthesize chemoattractive substances that attract migrating neural crest cells (Le Douarin and Teillet, 1974), and it is possible that the hindgut produces these substances at a later stage than the fore- and midgut. The possibility has been raised that the presence of vagal level neural crest cells within the hindgut may be necessary to induce the sacral level neural crest cells to immigrate into the hindgut (Allan and Newgreen, 1980; Gershon et al., 1993). However, a recent study has shown that following ablation of the vagal level neural tube, the migration of the sacral cells into the hindgut is unaffected (Burns et al., 2000), and combination grafts of aneural hindgut with sacral cells with and without vagal cells have shown the same effect (Hearn and Newgreen, 2000). Thus, if there is a chemoattractive substance produced in the hindgut, it appears to be produced by the mesenchyme, and not vagal level neural crest cells. Rather than an absence of chemoattractive substances to attract sacral cells into the hindgut, it is also possible that the delayed entry of sacral cells into the hindgut is due to a transient repellent effect of the hindgut. A recent study has shown that the mesenchyme of the colo-rectum of chick embryos expresses collapsin-1, a secreted glycoprotein that is a member of the semaphorin family (Shepherd and Raper, 1999). Studies in other parts of the nervous system have established that collapsin-1 acts as axon repellent and neural crest guidance cue (Behar et al., 1996; Taniguchi et al., 1997; Eickholt et al., 1999). Shepherd and Raper (1999) showed that, in vitro, extension of neurites from neurons in the nerve of Remak is inhibited by explants of colo-rectum. Collapsin-1 is initially (at E6) expressed throughout the mesenchyme of the wall of the colo-rectum; at a later stage (E8), which coincides with the time at which the axons of the nerve of Remak first enter the outer (muscle) layers of the hindgut, the expression of collapsin recedes to the inner submucosal and mucosal layers and is completely absent from the muscle layers (Shepherd and Raper, 1999). Since the sacral neural crest cells migrate into the hindgut along nerve fibres from the nerve of Remak that are projecting into the hindgut (Burns and Le Douarin, 1998), it is possible that the expression of collapsin-1, either directly or indirectly, regulates the time of entry of the sacral cells into the hindgut. The second unresolved question about the contribution of the sacral neural crest to the ENS in the hindgut is the identity of labelled cells observed in the hindgut following introduction of DiI or a retrovirus into the sacral neural tube prior to migration of neural crest cells (Pomeranz et al., 1991; Serbedzija et al., 1991), as these cells were observed within the hindgut 2–3 days prior to the arrival of sacral neural crest cells as reported by Burns and Le Douarin (1998). It also remains to be determined whether cells derived from the sacral neural crest give rise to specific functional classes of enteric neurons, and whether sacral cells have the potential to colonize other regions of gut apart from the distal hindgut. Although a contribution from the sacral neural crest to the ENS in the hindgut has been shown conclusively in chick/quails (Burns and Le Douarin, 1998), the situation is still unclear in mammals. Mice in which the genes encoding either Ret or GFRα1 are inactivated lack enteric neurons in most of the gastrointestinal tract caudal to the stomach, but some neurons are present in the distal rectum (Durbec et al., 1996; Cacalano et al., 1998). The most parsimonious explanation for the existence of neurons in the distal rectum, but not in the small intestine or colon, is that they arise from the sacral-level neural crest. However, as in birds, it appears that if sacral level neural crest cells do give rise to enteric neurons in the hindgut, they do not enter the hindgut until after it has been fully colonized by vagal level neural crest cells. A range of markers of neural crest cells have been used to examine the colonization of the embryonic mouse gut (Kapur et al., 1992; Young et al., 1998a, 1999). These studies have shown a single, unidirectional, rostral-to-caudal colonization of the gut by neural crest cells; labelled cells are not observed in the hindgut prior to the arrival of the cells, which are deduced to be vagal neural crest cells (Fig. 1). Furthermore, if the hindgut is removed prior to the arrival of vagal crest cells and explanted or cultured, no neurons develop in the vast majority of explants (Nishijima et al., 1990; Kapur et al., 1992; Young et al., 1996, 1998a). Mammals do not have a nerve of Remak, but the pelvic plexus is closely apposed to the developing distal hindgut in mouse embryos (see Young et al., 1998a), and it is possible that sacral neural crest cells migrate into the hindgut, but only after pausing in the pelvic plexus for 3–4 days, as they do in the nerve of Remak in birds. This is supported by the observation that cells with a similar phenotype to vagal enteric neural crest cells are present in the pelvic plexus adjacent to the distal hindgut of embryonic mice from E10.5 (see below; Young et al., 1996). The wavefront of vagal-level neural crest cells starts to colonize the rostral hindgut of mice at E11.5 and reaches the distal hindgut at around E14 (Kapur et al., 1992; Young et al., 1998a). Even at late E13.5, when the wavefront of the vagal cells is within about 500 μm of the anal end, cells within the pelvic plexus that are derived from sacral-level neural crest do not appear to have entered the hindgut (Fig. 2). Although most studies have reported an absence of neurons in the hindgut if it is removed and cultured prior to the arrival of vagal neural crest cells (see above), Rothman and Gershon (1982) reported neurons were present in cultured explants of hindgut from E9-E10 mice. Since separation of the distal hindgut from adjacent tissue is particularly difficult at these early stages, it is possible that the gut explants also had part of the pelvic plexus attached, from which neuronal precursors migrated into the explants. Location of neural crest–derived cells (shaded area) in the gut of E10.5–E14.5 mice. The uni-directional colonisation of the gut represents the migration of cells derived from vagal level neural crest. Note that there are no cells in the hindgut prior to the arrival of vagal neural crest cells. Wholemount preparation of the hindgut of a late E13.5 mouse showing Phox2b-immunoreactive cells. The most caudal cells (arrows) of the wave of Phox2b-positive cells derived from the vagal neural crest is approximately 500 μm from the anal end. Phox2b-positive cells are also present in a group in the primordium of the pelvic plexus, but there do not appear to be any cells within the most caudal part of the hindgut adjacent to the pelvic plexus. Scale bar = 100 μm. The neural crest population contains multipotent progenitors whose choice of fate is determined by environmental signals (Sieber-Blum and Cohen, 1980; Fraser and Bronner-Fraser, 1991). However, there is also increasing evidence that subsets of neural crest cells are committed to a particular fate either prior to, or shortly after, emigrating from the neural tube. For example, both lineage analysis (Frank and Sanes, 1991) and experiments in which neural crest cells are grown in vitro (Greenwood et al., 1999), suggest that the fate of a sub-population of sensory neurons is determined early, perhaps soon after they emigrate from the neural tube. In addition, neural crest cells destined to become melanocytes appear to be specified at the time that they emigrate from the neural tube (Henion and Weston, 1997; Reedy et al., 1998). There is also some evidence for pre-specification of neural crest cells destined to colonize the gut. For example, "adrenomedullary" level neural crest cells, corresponding to somites 18–24, do not normally penetrate the dorsal mesentery, and hence do not give rise to enteric neurons (Le Douarin and Teillet, 1974). However, when quail vagal-level neural crest cells were grafted in to replace the host chick embryo's adrenomedullary level neural crest cells, quail cells were present within the gut of the host, principally in the post-umbilical small intestine (Le Douarin and Teillet, 1974; Smith et al., 1977). Le Douarin and her colleagues suggested that vagal neural crest cells are pre-specified to form the ENS, and follow chemotactic cues to reach the gut. Endothelin (ET) (B) receptor (see Yanagisawa et al., 1998), Ret (Roberston and Mason, 1995), and CCK-LacZ (Lay et al., 1999) are expressed by pre-migratory cells in the neural tube. Unlike ET(B) receptor, which is expressed by all neural crest lineages including pigment cells, Ret appears to be expressed only by crest-derived neural precursors including sensory, sympathetic and enteric ganglia, and CCK-LacZ is expressed only by sympathetic and enteric ganglia. It is, therefore, possible that the Ret-positive and the CCK-LacZ-positive cells in the neural tube represent cells pre-specified to restricted lineages. While they are migrating, both from the neural tube to the gut, and then through the gut mesenchyme, neural crest cells are morphologically indistinguishable from the mesenchymal cells through which they migrate. Hence, experiments examining the origin and migration of neural crest cells initially relied on observing the effect of ablations, or the use of chick-quail chimeras (see above). However, a rapidly expanding number of markers has been discovered that identify undifferentiated neural crest cells prior to their differentiation into enteric neurons and glial cells. These markers are shown in Table 1. Some of the markers are expressed by neural crest cells prior to their entry into the gut mesenchyme, whereas other markers are not detectable until after the cells have arrived within the gut mesenchyme. One of the first markers of undifferentiated neural crest-derived cells in the chick embryo was the E/C8 antibody, which recognizes an avian-specific, 73-kD neurofilament-associated protein (Ciment and Weston, 1985; Ciment et al., 1986). Around the same time that the E/C8 antibody was first utilized to label neural crest cells, the equivalent monoclonal antibodies, HNK-1 and NC-1, were discovered to recognise a carbohydrate epitope on the cell surface of neural crest cells (Tucker et al., 1984). HNK-1 and NC-1 stain virtually the same population of cells that is stained with E/C8 (Tucker et al., 1986). Since then, HNK-1 and NC-1 have been frequently used to identify neural crest-derived cells within the embryonic gut, primarily of chick and quail embryos (Figs. 3A,4; Tucker et al., 1986; Pomeranz and Gershon, 1990; Epstein et al., 1991; Newgreen et al., 1996, 1997). HNK-1 and NC-1 also bind to neural crest-derived cells in a range of species including rat (Fig. 5A), but not mice or amphibians (Tucker et al., 1988; Erickson et al., 1989; Newgreen and Hartley, 1995). However, the HNK-1 epitope is also found on a number of cells not derived from the neural crest, and also on extracellular matrix molecules (Fig. 3A; Newgreen et al., 1990), and so results must be interpreted with caution. E/C8, HNK-1, and NC-1-positive cells are observed within the somites adjacent to the vagal level neural tube, and thus the antigens to which the antibodies bind appear to be first expressed shortly after the neural crest cells leave the neural tube (Tucker et al., 1986). A: Transverse slice through the level of somites 5–6 of an E2.75 quail embryo. HNK-1-positive cells are present in the position of the dorsal root ganglia (DRG) and sympathetic chain (asterisk), and in some cells (arrows) that have migrated further ventrally and are entering the developing gut. Note that HNK-1 staining is also observed in extracellular matrix around the notochord. Scale bar = 50 μm. B: Confocal microscope image of neural crest derived cells in the gut of an E10.5 mouse after processing for Phox2b (red) and p75NTR (green) immunoreactivity. Phox2b-immunoreactivity is confined to the nuclei and p75NTR immunoreactivity is predominantly on the cell membranes. All of the Phox2b-positive nuclei are surrounded by p75NTR-positive membranes. Nucleated blood cells (asterisks) show non-specific fluorescence. Scale bar = 10 μm. C: Transverse section of resin-embedded small intestine from an E15 mouse stained with toluidine blue. Phox2b-positive cells (arrows) are only present in the outer part of the intestine, just under the serosa. There are no cells at the inner side of the circular muscle, where the submucosal plexus will later form. Scale bar = 10 μm. D: Wholemount preparation of a myenteric ganglion from an adult mosaic mouse in which 50% of the cells express the reporter gene, lacZ. The ganglion is composed of both β-gal-positive cells (asterisks) and β-gal-negative cells (arrows), and hence does not arise from a single progenitor cell. Scale bar = 20 μm. E: Confocal microscope image of a wholemount preparation of the hindgut from an E14.5 mouse. A sub-population (arrows) of the Phox2b cells (stained red) has differentiated into NOS neurons (stained green), but most Phox2b-positive cells (asterisks) do not express NOS. Scale bar = 10 μm. F: Wholemount preparation of gut from an E12.5 mouse stained with an antibody to p75NTR (red) and the nucleic acid stain, SYTO 16. The two p75NTR-positive neural crest cells in this field of view (arrows) are both undergoing mitosis, as is evident from their chromosomal arrangement. The SYTO 16-stained cells that are not neural crest–derived cells are mesenchymal cells. Scale bar = 10 μm. G: High-magnification image of a neural crest–derived cell undergoing mitosis (telophase) that was immunostained for p75NTR (green) and Phox2b (red). At the late stages of mitosis, Phox2b immunostaining is associated with the chromosomes. Scale bar = 5 μm. HNK-1 labeled whole-mount of E5 (HH27-) quail mid and hindgut. Vagal neural crest–derived HNK-1-labelled enteric neural cells extend through the pre-umbilical (Pre-U) and post-umbilical (Po-U) intestine, to the rostral border of the caecum (arrow), but no HNK-1-stained cells are seen distally at this stage, for example in the colo-rectum. Sacral neural crest–derived HNK-1-labelled cells form the nerve of Remak in the dorsal mesentery directly adjacent to the intestine but do not at this stage extend into the intestine at any level. Scale bar = 200 μm. A: High-magnification image of an HNK-1-positive cell in the midgut of an E16 rat. Scale bar = 5 μm. B: p75NTR-positive cells in the midgut of an E11 mouse. Scale bar = 20 μm. C: Phox2b-positive nuclei in the esophagus of an E13.5 mouse. Scale bar = 50 μm. ET(B) receptor is expressed by the precursors of melanocytes, the adrenal medulla and virtually all components of the peripheral nervous system, including dorsal root ganglia and autonomic (including enteric) ganglia. Neural crest cells at all levels of the neural axis of chick and quail embryos express ET(B) receptor both before, and during, their emigration from the neural tube (Nataf et al., 1996). Ret is also expressed in the neural tube prior to the emigration of neural crest cells, but is restricted to vagal level neural tube, between rhombomeres 6 and somite 7 (Robertson and Mason, 1995). Following migration, Ret is expressed by some cranial ganglia, dorsal root ganglia, sympathetic ganglia, the ENS, and some cranial ganglia of the chick (Robertson and Mason, 1995). The only detailed study that examined the migration of sacral-level neural crest cells into the hindgut used quail-chick chimeric grafting to detect sacral-derived neural crest cells (Burns and Le Douarin, 1998, see above), rather than any of the markers shown in Table 1. Hence, little is known about the phenotype of undifferentiated sacral-derived neural crest cells in the chick, either prior to, or following, their entry into the hindgut. Although Ret is expressed in the vagal neural tube of the chick prior to emigration of the neural crest cells, Ret is not expressed by the sacral level neural tube (Robertson and Mason, 1995). Sacral cells destined to populate the chick hindgut appear to pause in the nerve of Remak for about 3 days prior to entering the hindgut (Burns and Le Douarin, 1998; see above). HNK-1 staining, and Ret- and GFRα1-immunoreactivity is shown by many cells in the nerve of Remak (Tucker et al., 1986; Epstein et al., 1991; Schiltz et al., 1999), but it is unknown if any of these cells emigrate into the hindgut. ET(B) receptor, tyrosine hydroxylase (TH), p75NTR, Phox2b, Ret, MASH1, CCK-lacZ, and SOX10 are all expressed by vagal neural crest cells prior to their entry into the gut mesenchyme (for references see Table 1). ET(B) receptor (see Yanagisawa et al., 1998) and CCK-lacZ (Lay et al., 1999) are expressed in the neural tube before the cells commence migration, but the other molecules are not detectable until after the cells have left the neural tube. Although Ret is expressed by pre-migratory crest cells of chick embryos (see above), it is first detected in migratory phase neural crest cells in mice embryos (Pachnis et al., 1993). Some molecules or genes, such as TH (Cochard et al., 1978; Teitelman et al., 1979), Phox2b (Pattyn et al., 1997, 1999), and CCK-lacZ (Lay et al., 1999), are expressed by enteric and sympathetic neuron precursors, but not dorsal root ganglion precursors, whereas other molecules such as ET(B) receptor (Southard-Smith et al., 1998), Ret (Pachnis et al., 1993; Tsuzuki et al., 1995; Nosrat et al., 1997), p75NTR (Britsch et al., 1998), ErbB3 receptor (Britsch et al., 1998), and SOX10 (Herbarth et al., 1998; Southard-Smith et al., 1998), are expressed by the precursors of enteric, sympathetic, and sensory neurons. It is unknown whether all neural crest cells destined to enter the gut express all of the molecules, or whether different phenotypic classes of cells exist, based on the expression of different markers. However, after entering the mouse gut, undifferentiated neural crest cells are of only two phenotypes (Young et al., 1999; see phenotype of neural crest-derived cells within the gut wall below). It, therefore, seems likely that prior to entering the gut, the neural crest cells will show a similar, limited range of phenotypes. Little is known about the phenotype of sacral-level neural crest cells in mammals. Migrating sacral neural crest cells in the embryonic mouse express SOX10 (Southard-Smith et al., 1998). In addition, cells in the pelvic plexus primordium express Phox2b and p75NTR (Young et al., 1998a), and thus if sacral-derived neural crest cells migrate into the mammalian hindgut via the pelvic plexus (see above), then, like vagal cells, sacral-derived cells are also likely to express both of these molecules prior to entering the gut. All of the markers expressed by neural crest-derived cells en route to the gut are also expressed by the crest-derived cells within the gut wall, for varying periods of time. Vagally-derived HNK-1/NC-1-positive cells colonize the embryonic chick and quail gut in a rostral-to-caudal wave (Fig. 4; Tucker et al., 1986; Epstein et al., 1991; Newgreen et al., 1997; Schiltz et al., 1999), and it is assumed that all undifferentiated neural crest-derived cells in the chick gut are HNK-1-postive. Little is known about the overlap of expression of other neural crest markers by the HNK-1-positive cells. However, the relationship between HNK-1 staining and the expression of the GDNF receptors, Ret and GFRα1, by neural crest cells as they colonize the gut has recently been examined (Schiltz et al., 1999). Unlike the situation in the mouse embryo (Young et al., 1999; see mouse and rat below), not all of the undifferentiated neural crest–derived cells in the embryonic chick gut appear to express Ret or GFRα1 as they are colonizing the gut, since HNK-1-positive cells at the migratory wavefront do not show Ret- or GFRα1-immunoreactivity. Although TH is expressed by vagal neural crest cells both prior to their entry into the gut, and for several days after they have entered the gut mesenchyme of embryonic mouse and rat (see below), it is not expressed by neural crest cells in chick embryos either prior to or after they have entered the gut mesenchyme (Smith et al., 1977). However, when neural crest cells are immunoselected from dissociated segments of embryonic chick gut using HNK-1, and then grown in culture, some cells do express TH (Pomeranz et al., 1993; Sextier-Sainte-Claire Deville et al., 1994). Sacral neural crest-derived cells differentiate into neurons (including nitric oxide synthase-containing neurons) and glial cells in the hindgut (Burns and Le Douarin, 1998). However, to our knowledge, nothing is known about the phenotype of sacral-derived neural crest cells within the chick hindgut prior to their differentiation into neurons and glial cells. A large number of molecules are expressed by vagal neural crest–derived cells in the embryonic mouse and rat gut (see Table 1; Fig. 5B,C). The colonization of the gut by neural crest–derived cells in mice was first examined in detail in transgenic mice in which the promotor of the gene encoding the catecholamine synthetic enzyme, dopamine-β-hydroxylase (DbH), was used to direct expression of the reporter gene, lacZ (Kapur et al., 1992). DβH-lacZ appears to be expressed by many, if not all, vagally-derived neural crest cells within the gut, and the expression of the transgene by enteric neurons persists in adult mice, even though there are no catecholaminergic neurons in the adult ENS. Although DβH-lacZ-expressing cells almost certainly include the cells that transiently express TH (see below), DβH-lacZ is expressed by many more cells, and for a far longer period of time, than TH. By breeding DβH-lacZ mice with mice carrying mutations that result in ENS defects, the role of particular genes in the migration of neural crest–derived cells into the gut has been examined (see Kapur et al., 1993, 1996). Vagal crest cells colonise the mouse and rat gut in a rostral-to-caudal wave (Fig. 1; Kapur et al., 1992; Newgreen and Hartley, 1995; Young et al., 1998a), and thus during the colonization, the most caudal cells are undifferentiated, whereas cells rostral to the wavefront are at various stages of differentiation. The overlap in the expression of some of these molecules was examined recently in the embryonic mouse gut (Young et al., 1999). Between E10.5 and E13.5, undifferentiated (the most caudal) neural crest cells were all found to co-express Ret, p75NTR and Phox2b (Fig. 3B), but none was Phox2a-positive. TH is also expressed by neural crest–derived cells within the early embryonic rat and mouse gut (Cochard et al., 1978; Teitelman et al., 1979). At E10.5, TH is expressed by about 20% of the Ret/p75NTR/Phox2b cells at the wavefront (Young et al., 1999). In the E10.5 mouse gut, although we have never observed the most caudal cell to be TH-positive, we have observed preparations where the second and third most caudal cells are TH-positive (Fig. 6A,A'). However, at subsequent stages, TH-positive cells become progressively further (more rostral) from the wavefront, and, in fact, TH cells are never observed in the hindgut. It is unclear whether the absence of TH cells from the hindgut is due to the environment of the hindgut, or to the poor migratory ability of TH cells. Since TH cells show many features of neurons, including the expression of neuron-specific proteins such as peripherin, neurofilament, and PGP9.5 (Fig. 6B,B',C,C'; Baetge et al., 1990), it is possible that TH cells fail to reach the hindgut because they are early differentiated neurons and hence have poor migratory abilities. Although all undifferentiated neural crest–derived cells within the gut appear to express Ret, p75NTR and Phox2b, it remains to be determined for most of the other markers
Referência(s)