Migratory Patterns and Developmental Potential of Trunk Neural Crest Cells in the Axolotl Embryo

Home , Neural fold, Neurula, Trunk neural crest

DEVELOPMENTAL DYNAMICS 236:389–403, 2007

RESEARCH ARTICLE

Hans-Henning Epperlein,1* Mark A.J. Selleck,2 Daniel Meulemans,3 Levan Mchedlishvili,4 Robert Cerny,5 Lidia Sobkow,4 and Marianne Bronner-Fraser3

Using cell markers and grafting, we examined the timing of migration and developmental potential of trunk neural crest cells in axolotl. No obvious differences in pathway choice were noted for DiI-labeling at different lateral or medial positions of the trunk neural folds in neurulae, which contributed not only to neural crest but also to Rohon-Beard neurons. Labeling wild-type dorsal trunks at pre- and early-migratory stages revealed that individual neural crest cells migrate away from the neural tube along two main routes: first, dorsolaterally between the epidermis and somites and, later, ventromedially between the somites and neural tube/notochord. Dorsolaterally migrating crest primarily forms pigment cells, with those from anterior (but not mid or posterior) trunk neural folds also contributing glia and neurons to the lateral line. White mutants have impaired dorsolateral but normal ventromedial migration. At late migratory stages, most labeled cells move along the ventromedial pathway or into the dorsal fin. Contrasting with other anamniotes, axolotl has a minor neural crest contribution to the dorsal fin, most of which arises from the dermomyotome. Taken together, the results reveal stereotypic migration and differentiation of neural crest cells in axolotl that differ from other vertebrates in timing of entry onto the dorsolateral pathway and extent of contribution to some derivatives. Developmental Dynamics 236:389–403, 2007. © 2006 Wiley-Liss, Inc.

Key words: DiI; ﬂuorescent dextrans; GFP transgenic embryos; trunk neural crest; migration; dorsal ﬁn; Rohon-Beard cells; lateral line; axolotl

Accepted 31 October 2006

INTRODUCTION of the peripheral nervous system, pig- border between prospective neural The neural crest is a transient migra- ment and adrenomedullary cells, the plate and epidermis (Moury and Ja- tory population of cells found in all craniofacial skeleton and, in lower cobson, 1989; Selleck and Bronner- vertebrate embryos (Knecht and vertebrates, mesenchyme of the dor- Fraser, 1995). Because the neural Bronner-Fraser, 2002). These cells mi- sal ﬁn (Hall and Ho¨rstadius, 1988; Le folds are contiguous with neural plate grate extensively along deﬁned path- Douarin and Kalcheim, 1999). Neural and epidermis (Moury and Jacobson ways and give rise to diverse deriva- crest cells arise in a prospective neu- (1990), it is not clear where they “end” tives, including neurons and glial cells ral fold area of the ectoderm at the and neural tissue or epidermis begin.

1Department of Anatomy, TU Dresden, Dresden, Germany, 2Department of Cell and Neurobiology, University of Southern California Keck School of Medicine, Los Angeles, California 3Division of Biology and Beckman Institute, California Institute of Technology, Pasadena, California 4Max-Planck-Institute of Molecular Cell Biology and Genetics, Dresden, Germany 5Department of Zoology, Charles University Prague, Prague, Czech Republic Grant sponsor: DFG; Grant number: EP8/7-1; Grant sponsor: NIH; Grant number: NS051051; Grant sponsor: James H. Zumberge Research and Innovation Fund; Grant sponsor: Donald E. and Delia B. Baxter Foundation. *Correspondence to: Hans H. Epperlein, Department of Anatomy, TU Dresden, Fetscherstr. 74, 01307 Dresden, Germany. E-mail: [email protected] DOI 10.1002/dvdy.21039 Published online 19 December 2006 in Wiley InterScience (www.interscience.wiley.com).

Furthermore, single cell lineage anal- al., 1988; Collazo et al., 1993). In the lyze the migratory patterns and deriv- ysis in several species suggests that a mouse, no timing differences exist be- atives of trunk neural crest cells single neural fold precursor can give tween entry onto the dorsolateral and emerging from the neural folds/tube rise not only to neural crest but also to ventromedial pathways (Serbedzija et (Fig. 1A–E). DiI injections were per- neural tube and epidermal derivatives al., 1990). Thus, species-specific differ- formed at different positions along the (Selleck and Bronner-Fraser, 1995). ences exist in some aspects of neural mediolateral axis (Fig. 1A,F) prior to Accordingly, Chibon (1966) concluded crest migration. There is reason to the onset of neural crest migration. that Rohon-Beard neurons, missing think that migratory behavior may Along the rostrocaudal axis, anterior, after neural fold ablation, were neural differ between axolotl (Ambystoma middle, or posterior levels of the dor- crest derived. mexicanum) and chick as well. In the somedial neural folds or neural tube A major unresolved question in neu- axolotl, some premigratory neural were injected (Fig. 1B–E;F). In the ral crest development concerns the re- crest cells, such as presumptive pig- trunk of early tailbud stage embryos, ment cells, appear to be committed to lationship between neural crest and the neural crest can be morphologi- neural tube with respect to the alloca- their fate prior to the onset of migra- cally identified as a wedge of cells that tion of different lineages and path- tion. Therefore, it is possible that sub- forms on the dorsal aspect of the neu- ways of migration. This question has sequent migration may follow differ- ral tube and expresses snail tran- been extensively studied in the chick ent rules. It is unknown whether the scripts (Fig. 1G). Later, an elevated embryo, which is amenable to culture premigratory crest is entirely “mo- neural crest string is present on top of and microsurgical manipulation in saic” and contains only unipotent cells combination with numerous cell- or a mixture of cells with different the neural tube (Fig. 1H). Once neural marking techniques for following neu- ranges of developmental potential. crest cells migrate, they move along ral crest migration (Le Douarin, 1969; Here, we address this issue by two primary pathways: a dorsolateral Le Douarin and Kalcheim, 1999; Ser- studying the migratory patterns and pathway underneath the epidermis bedzija et al., 1989; Elena de Bellard developmental potential of early dif- that is primarily thought to give rise and Bronner-Fraser, 2005). The mi- ferentiating trunk neural crest cells in to pigmented derivatives (melano- gratory behavior of avian neural crest the axolotl such as pigment cells, Ro- phores and xanthophores) and a ven- appears to be highly organized, such hon-Beard cells, or components of the tromedial pathway followed by pre- that the first cells to exit the neural lateral line. As a first step, we have cursors to sensory, sympathetic, tube migrate along a ventromedial performed a series of cell-marking and adrenomedullary, and glial cells. It is pathway through the somites, while grafting studies using dark wild-type sometimes difficult to discriminate be- later migrating cells are restricted to a and white mutant embryos, the latter tween laterally migrating neural dorsolateral pathway between the epi- having defective pigment cell migra- crest–derived pigment precursors, dermis and somites where they form tion (Dalton, 1953; Keller et al., 1982; epidermis, or neural tube on the one pigment cells (Le Douarin and Teillet, Lo¨fberg et al., 1985). The results show hand and between pigment precursors 1974; Guillory and Bronner-Fraser, that neural crest cells follow both a and components of the lateral line on 1986; Erickson et al., 1992; Serbedzija dorsolateral and a ventromedial mi- the other. To best illustrate this point, et al., 1989). In the chick, it has been gratory route similar to other verte- we examined dark instead of white proposed that melanocytes are speci- brates. However, the time course is mutant embryos since they have fied before entering the dorsolateral different than in other anamniotes or many more laterally migrating cells migratory route (Reedy et al., 1998) amniotes, since the dorsolateral route (Figs. 1I–L). DiI-labeled cells can be is taken first. We confirm that neural and that distinct neurogenic and mel- distinguished in transverse sections crest and Rohon Beard cells share a anogenic sublineages may diverge be- after one anterior (Fig. 1I) or three common lineage within the neural fore or soon after neural crest cell em- midtrunk injections (Fig. 1J) into the igration from the neural tube (Luo et folds and find that only few axolotl trunk neural fold from epidermal or al., 2003). neural crest cells contribute to the neural tube cells by means of their Despite the wealth of information dorsal fin, which also has a contribu- position (Fig. 1K). Pigment precursors on avian neural crest migration, it re- tion from the dermomyotome (Sobkow and components of the lateral line mains unknown whether ordered mi- et al., 2006). Furthermore, we show such as neurons (ganglionic cells), gratory behavior is common to all ver- that anterior but not mid- or posterior tebrate embryos, or whether it is trunk neural folds contribute to the glial cells, and lateral line nerves can avian specific. Some differences in mi- lateral line system. These results re- be distinguished through the use of gratory behavior from those observed veal interesting similarities and dif- specific markers. We found that glial in chick have been noted in amphibi- ferences in the migratory pathways cells are present only in the vicinity of ans like Xenopus as well as in mam- and derivatives between axolotl neu- the lateral line primordium or nerve mals like the mouse. In Xenopus, neural crest and that of other vertebrates. (Fig. 1L). They follow these structures ral crest cells including presumptive caudally from the posterior cranial pigment cells migrate preferentially neural crest situated at a level caudal RESULTS on the ventromedial route, between to the otic placode (Northcutt et al., the somite and neural tube/notochord DiI labeling was performed at premi- 1994; Schlosser, 2002). They do not and not laterally or between the gratory stages (15–19; 25), early (33), migrate laterally from mid- or caudal somites (Macmillan, 1976; Krotoski et and late migratory (35) stages to ana- trunk neural crest regions. TRUNK NEURAL CREST CELLS IN THE AXOLOTL EMBRYO 391

DiI Labeling at Different Lateral or Medial Positions of the Trunk Neural Folds At neurula stages, the exact position of the prospective neural crest within the neural fold is not known. To address whether different aspects of the neural folds of neurulae differ in developmental potential, neural folds were labeled laterally or medially along their anterior/posterior axis (Fig. 1F) prior to the onset of neural crest migration (stages 15–19; Figs. 1A, 2A,D; Table 1). Embryos were ﬁxed 2 or 4 days post-injection and the

Fig. 1. DiI labeling and identification of neural crest cells in the axolotl trunk. DiI labeling was carried out at stages 15–19, 25, 33, and 35 to analyze the migratory potential of trunk neural crest. A: At stage 16 (mean stage used in this schematic to represent stages 15–19), three focal DiI injections were made into the lateral or medial aspects of the trunk neural folds (F). B: At stage 16, the left trunk neural fold was injected anteriorly or posteriorly in a middle position (F). At stages 25 (C), 33 (D), and 35 (E), three positions along the dorsal midtrunk were labeled. F: Red dots indicate positions of DiI labeling in the left trunk neural fold at stage 16. G: In transverse sections through the trunk of early tailbuds (stage 23), the neural crest (arrow) can be morphologically identified as wedge on Fig. 1. top of the neural tube where it stains with the snail riboprobe. H: In transverse sections through the trunk of later tailbuds (stage 34), an elevated neural crest string is present on top of the neural tube (DiI labeling of the neural crest, dapi staining, anti-fibronectin staining). I–L: DiI labeling of neural fold in dark (D/D) neurulae and identification of neural crest cells in larvae in relation to epidermis and neural tube. I, J show left trunks in larvae (stage 37); in I, one left anterior trunk neural fold, and in J, three left midtrunk neural fold positions were injected with DiI. In I, labeled cells are scattered in the upper anterior trunk and the middle lateral line primordium is stained. The arrow marks the injection site. In J, labeled cells are found throughout the lateral trunk, on top of the dorsal fin, and sparsely in the middle lateral line primordium. Injection sites are not more visible. K: Transverse section through the midtrunk of the larva in I counterstained with anti-fibronectin; in contrast to DiI-labeled epidermal cells (arrow), labeled neural crest cells (arrowheads) are found more internal to the epidermal basement membrane. L: Transverse section through the midtrunk of the larva in J counterstained with anti-fibronectin and dapi. Arrowheads point to DiI-labeled neural crest cells on the lateral and ventromedial pathway. The arrow indicates the Fig. 2. A–F: Lateral or medial fold injections (average: stage 16). Prior to the onset of neural crest migration middle lateral line primordium; the two labeled in dark (D/D) neurulae (stages 15–19), lateral (A) or medial (D) positions of trunk neural folds were labeled with cells in its vicinity are very likely glial cells be- three focal injections of DiI. Embryos were fixed 2 days (B,E) and 4 days (C,F) after the injection. In either coming associated with the middle lateral line case, after 4 days many more labeled cells were observed on the lateral than on the medial route whereas nerve. m, middle lateral line primordium; nc, the distance traversed by neural crest cells was similar on both routes (see also Table 1). The pink numbers neural crest; nt, neural tube; not, notochord. indicate DiI-labeled cells in the dorsal fin (df), the green numbers indicate sessile neural crest (nc) cells/ derivatives on the neural tube, and the red and blue numbers indicate neural cress cells/derivatives on the lateral (lat) and medial (med) route, respectively. 392 EPPERLEIN ET AL.

TABLE 1. Trunk Neural Crest Cell Migration In Axolotl Based on DiI Infectionsa

Migrated Time/stage distance Stage at Site/number of after Neural crest cells (␮m) Groups Figure Genotype injection N injections injection df nc lat med lat med 1 2B D/D 18/19 5 Lateral fold/3 2d/35 01 02 09 10 450 450 2 2C D/D 15 5 Lateral fold/3 4d/38 06 04 51 35 1100 1050 3 2E D/D 17/18 5 Medial fold/3 2d/34–35 02 00 16 12 600 650 4 2F D/D 15 5 Medial fold/3 4d/38 09 02 58 24 1150 1050 14 3B d/d 16 5 Anterior fold/1 4d/38 02 04 91 22 800 1200 15 3C d/d 16 5 Posterior fold/1 4d/38 15 15 23 26 800 950 5 6B D/D 25 5 Dorsal mid trunk/3 1d/32–33 00 14 05 00 300 0 6 6C D/D 25 5 Dorsal mid trunk/3 2d/36 00 14 55 07 700 400 7 6E d/d 25 5 Dorsal mid trunk/3 1d/33 00 00 01 09 350 200 8 6F d/d 25 5 Dorsal mid trunk/3 2-3d/36–37 05 03 14 28 450 650 9 7B D/D 33/34 5 Dorsal mid trunk/3 1d/36 02 07 24 07 750 600 10 7C D/D 33/34 5 Dorsal mid trunk/3 2d/37 03 17 40 44 750 650 11 7E d/d 33/34 5 Dorsal mid trunk/3 1d/36 00 13 03 02 200 200 12 7F d/d 33/34 5 Dorsal mid trunk/3 2d/37 02 22 07 31 450 650 13 8B D/D 35/36 5 Dorsal mid trunk/3 4d/39–40 24 17 22 79 700 1100 16 8C D/D 35/36 5 Dorsal mid trunk/3 4d/39–40 16 09 32 62 800 1050 4A D/D 16 5 Anterior fold/1 4d/38 4A D/D 16 10 Anterior fold/1 5d/39 5A d/d 16 10 Anterior fold/1 2d/32 5A d/d 16 5 Anterior fold/1 4d/38 5A d/d 16 5 Anterior fold/1 5d/39 4B d/d 16 12 Posterior fold/1 2d/32 4B d/d 16 15 Posterior fold/1 4d/38 4B d/d 16 15 Posterior fold/1 5d/39 – D/D 16 5 Posterior fold/1 4d/38 – D/D 16 10 Posterior fold/1 5d/39

aFigures 2,3,6–8 show a quantitative evaluation of DiI-labeled cells. Cell counting in each of 16 groups is based on approximately 20 transverse sections (100-␮m thick) through the trunk of embryos until stage 34 and 30 sections until stage 39/40. Five specimens were sectioned per group. Labeled cells in sections through all specimens of a group were counted and mirror-imaged to the left side of a summary section that represented the group (Figs. 2,3,6–8; Table 1). Figures 4A,B and 5A contain data on anterior or posterior trunk neural fold injections with Dil, which were qualitatively evaluated. D/D, wild-type (dark); d/d, white mutant; df, dorsal ﬁn; nc, neural crest; lat, lateral route; med, med at route; n, number of specimens investigated.

distribution of DiI-labeled cells deter- major differences were noted for me- rostrocaudal axis such that by stage mined (stages 35 or 38; Fig. 2B,C,E,F; dial versus lateral neural fold injec- 35, migrating pigment cells have Table 1). Because DiI labeling of the tions with respect to the distance tra- spread throughout the flank. It was neural folds marks epidermis and versed by neural crest cells away from previously assumed that only pigment neural tube in addition to neural the injection site. However, more mi- cells migrate along the dorsolateral crest, neural crest cells were carefully grating cells were observed along the pathway (Lo¨fberg et al. 1980; Epper- distinguished from epidermal and dorsolateral pathway after 4 days lein and Lo¨fberg, 1990). neural tube using morphological crite- (Fig. 2C,F). Here, we investigated whether ria (Fig. 1K). other neural crest populations mi- Two days following lateral injec- DiI Labeling of Different grated along the dorsolateral pathway tions into neural folds of wild type em- using DiI labeling to mark the neural Rostrocaudal Axial Levels bryos, the leading edge of neural crest folds of white mutant embryos, which cells had migrated ϳ450 ␮m away Using scanning electron microscopy have a defect in pigment cell migra- from the injection site and ϳ1,100 ␮m and dopa-histochemistry, we have tion and thus have only a few melano- after 4 days and was found equally far previously shown that delamination of phores that could obscure recognition along the dorsomedial and ventrolat- neural crest cells starts around stage of migrating non-pigmented neural eral pathways (Fig. 2B,C; Table 1). 28 in the anterior trunk of wild-type crest derivatives. A focal DiI injection Following medial injections, neural embryos, while dorsolateral migration was made into one dorsal middle area crest cells had migrated 600–650 ␮m of prospective pigment cells com- of either the anterior or posterior after 2 days and 1,150/1,050 ␮m after mences by stage 31. Initiation of mi- trunk neural fold on the left side of a 4 days (Fig. 2E,F; Table 1). Thus, no gration proceeds gradually along the neurula (Fig. 3AB). Previously, the TRUNK NEURAL CREST CELLS IN THE AXOLOTL EMBRYO 393 postotic posterior lateral line placode rived from the lateral line placode pri- fragments or the entire left trunk neu- and associated neural crest have been mordium situated posterior to the otic ral fold, from post-otic region to the proposed to contribute to the lateral placode (Northcutt et al., 1994), re- tail, of white mutant embryos. By line system in the trunk (Northcutt et ceive afferent and efferent innerva- stage 35/36, the primordium of the al. 1994; Schlosser, 2002; Collazo et tion, and grow along invariant path- middle lateral line was labeled inter- al., 1994). ways (Schlosser, 2002). mittently (Fig. 4C). By stage 41, GFP- Following anterior injections into The primordium contains prospec- positive glia cells and neuromasts white neurulae (Fig. 3A,B; stage 16, tive lateral line receptors (neuromast were observed intermittently along axial level of prospective 2nd somite), and ampullary primordia) and sen- the middle nerve (not visible) on both the majority of labeled cells migrated sory neurons of lateral line ganglia the ispilateral side bearing the graft along the dorsolateral rather than the and glia (Schlosser, 2002; Gilmour et (Fig. 4D) as well as the contralateral ventromedial pathway (e.g., 91 vs. 22 al., 2002). By stage 37, the middle lat- side (Fig. 4E). In addition, a few GFP- cells for summing 5 representative an- eral line nerve extends to the tail re- positive pigment and epidermal cells terior injections; Table 1). After poste- gion, and by stage 41, the dorsal and were randomly distributed in the dor- rior injections (Fig. 3A,C, stage 16, at middle lateral lines fuse above the clo- solateral trunk (not shown). In some the most posterior trunk neural fold aca (Northcutt et al., 1994). cases, only the right, unilateral neural level), dorsolateral migration was To address the source of neural fold graft was used to replace the en- similar to ventromedial migration (23 crest cells that appear to cease migra- tire host since this often aided in vi- vs. 26 labeled cells summing 5 repre- tion in the vicinity of the dorsal or sualizing the GFP-labeled cells on the sentative posterior injections; Table middle lateral line nerve, focal injec- left side better (no left epidermis la- 1), but lateral migration was fourfold tions of DiI were applied into anterior beled). Using this approach, the left less than after anterior injections. (Fig. 4A) or posterior (Fig. 4B) neural dorsal, middle, and ventral lateral Interestingly, the front of dorsolat- folds of dark and white neurulae line nerves become clearly labeled in erally migrating neural crest cells ap- (stage 16; for exact level of injections, larvae at stage 40/41 (Fig. 4F). GFP- pears to stagnate around the middle see Fig. 3A). This results in random positive neural crest cells were ob- lateral line nerve regardless of the site distribution of labeled cells under- served associated with all three lat- of injection. Following anterior injec- neath the epidermis of developing lar- eral line nerves (Fig. 4G). However, tions, there appears to be an addi- vae. For anterior injections, labeled only surrounding the ventral nerve tional dorsolateral group of labeled cells were occasionally found along the was there a swarm of GFP-positive cells at stage 38 (Fig. 3B) correspond- middle lateral line (dark larva at cells (Fig. 4G) whose origin from the ing to the position of the dorsal lateral stage 37; Fig. 4A) probably depositing mid-trunk neural crest was question- line, which by this stage is present in the neuromast primordia (Northcutt, able. In transverse sections, they the anterior but not yet in the poste- personal communication). In contrast, proved to be localized more deeply, in rior trunk (see fig. 16 in Northcutt et DiI-labeled cells were not found along the lateral plate mesoderm (Fig. 4H), al., 1994). Because the number of dor- the dorsal or middle lateral line and, thus, represent neural crest–de- solateral melanophores in white lar- nerves after posterior injections rived enteric neurons/glial cells mi- vae is ϳ30 (see “Quantification of DiI- (white larva at stage 41; Fig. 4B). grating on the ventromedial route and Labeling and Control” in the These results suggest that only ante- turning to the surface from under- Experimental Procedures section), the rior trunk neural fold/neural crest neath the myotomes. These findings much higher number of labeled cells cells contribute to the lateral line. To confirm our DiI labeling and suggest observed in white larvae for those in- gain better cellular resolution of neu- that GFP-positive glial cells associ- jected anteriorly at stage 38 can only ral crest contributions to the lateral ated with lateral line nerves at trun- be ascribed to the presence of many line, we used a second labeling ap- cal levels arise from anterior neural more non-melanophores. This raises proach of grafting trunk neural folds folds from behind the otic placode. the intriguing question if mid and pos- from GFP-positive transgenic donor Immunohistochemistry was used to terior trunk neural fold cells might embryos into white hosts. Because examine the phenotype of labeled neu- contribute glia or neurons to the lat- white mutants have few melano- ral crest cells traveling along the eral line. phores compared with wild-type, pig- nerves. Five days after an anterior in- ment cells tend not to obscure the GFP jection when neuromasts are present Anterior But Not Mid or signal. For grafts of the posterior along the dorsal and middle lateral Posterior Trunk Neural trunk neural fold region, we observed line nerves, the nerves themselves are Folds Contribute to Neurons only GFPϩ pigment cells, leaving the only lightly labeled with DiI, appear- graft and no fluorescent cells migrating as a fine continuous line (white and Glia of the Lateral Line ing toward or associating with lateral larva at stage 39; Fig. 5A,B). In con- In fish and aquatic amphibians, the line nerves (data not shown), suggest- trast, Schwann cells scattered along lateral line system forms part of the ing that the posterior trunk fails to the nerves were brightly stained with VIIIth cranial nerve complex. In the contribute to the lateral line. DiI. Anti-tubulin staining of trans- trunk of axolotl larvae, three lines of Next, small mid trunk fragments or verse sections through this larva re- neuromasts develop, one dorsally, one entire trunk neural folds from GFPϩ vealed the nerves as two spots on ei- medially, and one ventrolaterally transgenic neurulae (stage 17) were ther side of the dorsal trunk (Fig. 5C). (Smith et al., 1990). All three are de- grafted in place of small left midtrunk Anti-tubulin staining was observed 394 EPPERLEIN ET AL.

Fig. 3. A–C: Anterior or posterior trunk neural fold injections (stage 16). One focal DiI injection was made into the left trunk neural fold in a middle position either at an anterior or posterior position of a white (d/d) neurula (stage 16) in order to investigate the contribution of neural crest cells to the lateral line system. Following anterior injections, labeled cells greatly favored Fig. 4. the lateral pathway over the medial pathway (B). In white embryos injected posteriorly, many fewer labeled cells migrated laterally (C). In either type of injection, the front of lateral migration stagnated in the ﬂank at the level of the middle lateral line nerve (mll); following anterior injections, labeled cells seemed to be grouped also around the dorsal lateral line nerve (dll). intermixed with DiI staining of puta- tive glial cells on the left side (Fig. 5C). The glia character of these DiI- positive cells (Fig. 5A–C) was shown by anti-GFAP staining of microwave- treated parafﬁn sections through the midtrunk of stage 41 larvae (Fig. 5D, blue circle) adjacent to anti-tubulin staining of the lateral line nerve (Fig. 5E, condensed brown structure).

DiI-Labeling After Neural Tube Closure Reveals Differences in Timing of Neural Crest Entry Onto Lateral Versus Ventromedial Routes We next investigated the choice of pathway and timing of migration of Fig. 5. TRUNK NEURAL CREST CELLS IN THE AXOLOTL EMBRYO 395 neural crest cells by labeling the dor- pared to the wild type. Labeled cells in 36), two days (stage 37), or four days sal neural tube after its closure but the mutant appeared stagnant in (stage 40) post-injection. prior to the onset of neural crest emi- their migration ϳ450 ␮m distant from In wild-type embryos labeled during gration. Results were compared be- the injection site. Again, their site of early migratory stages (Fig. 7A), tween wild-type and white mutant arrest corresponded to the position of many more labeled cells were ob- embryos. DiI was injected at three the middle lateral line nerve. served on the dorsolateral than the sites along the middle of the trunk Taken together, these results sug- ventromedial pathway one day after neural tube of stage 25 wild-type and gest that neural crest cells in wild injection (Fig. 7B; Table 1) but num- white mutant embryos (Fig. 6A,D) type embryos labeled at stage 25 mi- bers were approximately equal by two and the results examined 1 or 2–3 grate first on the lateral route where days after injection (Fig. 7C; Table 1). days later in fixed embryos (stages they are more numerous than those The front of dorsolaterally migrating 32–33 or 36–37). As in DiI labeling of migrating ventromedially. In white cells was about 750 ␮m away from the lateral and medial trunk neural folds, mutants, in contrast, lateral migra- injection site at both 1 and 2 days. we also carefully distinguished be- tion is reduced whereas ventromedial Ventromedial migration was about tween neural crest and epidermal migration (stage 36–37, Fig. 6F) re- 600 ␮m away from the injection site cells when counting neural crest cells sembles that in the wild type (stage after 1 day and 650 ␮m after 2 days. on the dorsolateral and ventromedial 36, Fig. 6C; stage 37, Fig, 7C). In white mutant embryos (injection pathways. sites Fig. 7D), few labeled cells were In dark embryos, the front of neural present on either route after 1 day crest cells on the lateral route had DiI-Labeling at Early and (Fig. 7E; Table 1). After 2 days, a few moved about 300 ␮m after 1 day (5 Late Migration Stages more labeled cells were observed on cells) and 700 ␮m after 2–3 days (55 the dorsolateral pathway but ap- cells; Figs. 6B,C; Table 1). On the ven- To determine how long neural crest peared to stagnate in their migration tromedial pathway, labeled cells were migration persists on different migra- (Fig. 7F; Table 1). Those on the ven- absent after 1 day but had moved tory pathways, cell marking was car- tromedial pathway, however, had in- maximally 400 ␮m away from the in- ried out at progressively later stages creased considerably in number and jection site after 2 days (7 cells), sug- ranging from early (stage 33; Fig. traveled 650 ␮m. Thus, in dark em- gesting that this pathway opens sec- 7A–F) to late migration (stage 35; Fig. bryos dorsolateral cells migrate first ondarily (Fig. 2A,B; Table 1). In the 8A–C). Each embryo received focal in- and are more frequent than those on white embryo, dorsolateral migration jections of DiI at three rostrocaudal the ventromedial pathway similar to was reduced (1 cell after one day, 14 locations on the dorsum of the neural embryos injected at stage 25. In white cells after 2–3 days; Fig. 6E,F) com- tube and was examined one day (stage mutant embryos, dorsolateral migra-

Fig. 4. Contribution of neural crest cells to the lateral line. A: Left side of a dark axolotl larva at stage 37, 4 days after injection of DiI into an anterior location of the left trunk neural fold. DiI likely labels the main trunk lateral line placode depositing the neuromast primordia for the middle lateral line (m). B: Left tail region of a white axolotl larva at stage 41, 5–6 days after injection of DiI into a posterior location of the left trunk neural fold. Neuromasts of the middle lateral line nerve are encircled. A few labeled cells spread laterally from the injection site but do not associate with the middle or dorsal lateral line nerves. C: Left side of a white larva (stage 35/36) bearing a trunk neural fold graft from a GFPϩ transgenic neurula on its left side. The dorsal neural tube and covering epidermis are heavily labeled. The primordium of the middle lateral line nerve (m) migrates posteriorly and has reached a level above the cloaca. The intermittent labeling along the nerve is very likely due to glial cells migrating posteriorly. From the labeled neural tube, no labeled neural crest cells were observed to migrate laterally and associate with the middle lateral line nerve. D: Lateral aspect of left midtrunk of white larva (stage 41) whose entire left trunk neural fold was replaced with a GFPϩ fold. Elongated ﬂuorescent glial cells (arrows) were present intermittently along the middle lateral line nerves of the ipsi- (D) and contralateral side (E). Arrowheads in E point to GFPϩ centers of neuromasts where possibly supportive cells are stained. From the GFPϩ grafted neural fold (dorsal, outside the image), no lateral migration of labeled neural crest cells was observed towards lateral line nerves. F: Left side of a white larva (stage 40/41) following removal of both trunk neural folds and replacement of the right fold with a GFPϩ fold. The dorsal and middle lateral line nerve and the advancing ventral nerve are labeled. G: Enlarged inset from F showing the enlarged middle (m) and ventral (v) nerve. The advancing tip of the ventral nerve is surrounded by a swarm of GFPϩ cells. Only those at the nerve or in close vicinity to it seem to be glial cells. GFPϩ cells further apart are very likely enteric neurons/glial cells that migrated into the lateral plate on the ventromedial route. H: Transverse section through larva in F,G in the plane indicated by a white vertical line in G; GFPϩ cells in the epidermis were observed only in the position of the ventral lateral line (v). All other GFPϩ cells were lying in the lateral plate mesoderm (arrowheads) and had reached this position after migrating on the ventromedial pathway below the myotomes (my). Immunostaining with primary anti-ﬁbronectin and secondary Cy3-conjugated antibody to reveal basement membranes; dapi staining.

Fig. 5. Identity of lateral line nerves and glia. A: Left side of a white larva at stage 39, 4 days after injection of DiI into an anterior location of the left trunk neural fold. Irregular labeling of the dorsal (d) and middle (m) lateral line nerve with DiI. B: Enlarged inset from A showing neuromasts (arrows) dorsal to the middle lateral line nerve; irregular DiI-labeling along the nerve stains possibly neuromast primordia and glial cells. C: Immunostaining of a transverse vibratome section through the midtrunk of the larva in A with a primary anti-tubulin antibody and a FITC-conjugated secondary antibody reveals two green fluorescent subepidermal spots on either side of the dorsal trunk that label the dorsal and midde lateral line nerve. Green spots on the left side colocalize with DiI-labeled glial cells. D: Immunostaining of transverse microwave-treated paraffin sections through the trunk of a larva (stage 41) with a primary anti-GFAP-antibody, a secondary biotinylated antibody, and a tertiary avidin-peroxidase complex (Vectastain ABC-Kit). Glia (arrow; blue color) around the middle lateral line “nerve” and in the neural tube (nt) reacted positive and (E) colocalized with the middle lateral line nerve (arrow, brown color) dorsal to a blood vessel (v) in a similar transverse microwave-treated paraffin section (stage 41 larva) that reacted positive with a primary anti-tubulin antibody, a secondary biotinylated antibody, and a tertiary avidin-peroxidase complex (Vectastain ABC-Kit). 396 EPPERLEIN ET AL.

Fig. 6. A–F: Dorsal mid trunk injections (stage 25, premigratory stage). Three separate focal injections of DiI were applied along the mid trunk neural tube/premigratory neural crest of wild type(D/D;6A) and white mutant (d/d) embryos (6D) at stage 25. Neural crest cells in the wild type migrate first on the lateral route where they are more numerous than on the medial pathway. In the white mutant, lateral migration remains strongly reduced. The front of cells coincides with the positon of the middle lateral line nerve. Medial migration dominates and resembles that in the wild Fig. 8. A–C: Dorsal mid trunk injections (stage type (compare F with C). 35, late migration stage). Wild-type (D/D) embryos (two groups, B and C, with 5 individuals in each group) were injected with DiI into the dorsal neural tube/neural crest at late migration (stage 35) and fixed 4 days post-injection (stage 40). At stage 40, ventromedial migration and migration into the dorsal fin dominate.

into the dorsal fin with some migration observed along the lateral pathway. It is difficult to perform DiI labeling of the trunk neural tube much later than stage 35 due to the impenetrable nature of the epidermis. Based on larvae injected at stage 35 (Fig. 8A) and investigated 4 days later (stage 40), it is likely that neural crest cells are still emigrating from the neural tube after stage 35 (Fig. 8B,C; Table 1). Pigment cells (melanophores, xanthophores) start to visibly differenti- Fig. 7. A–F: Dorsal mid trunk injections (stage 33, early migration stage). Three separate focal ate under the epidermis by stage 35. injections of DiI were carried out along the mid trunk neural tube/early migratory neural crest of wild type (D/D; A) and white mutant (d/d; D) embryos at stage 33. In wild type embryos one day after the By stages 38–40 neurons and glial injections, many more labeled cells were observed on the lateral than on the medial pathway (B). cells are condensing into dorsal root By 2 days (C), numbers were approximately equal. In white embryos one day after the injection, few ganglia at the ventrolateral margin of labeled cells were found on both routes (E). By 2 days, many more labeled cells were found on the the neural tube as revealed in trans- ventromedial route but their number stagnated laterally (F). verse sections stained with anti-tubulin and anti-GFAP (not shown), re- tion appears to be compromised com- (Fig. 8A) and fixed 4 days post-injec- spectively. In 14-mm-long larvae, pared with that in wild-type counter- tion (stage 40). As late as stage 40, DiI tyrosine hydroxylase (TH)-positive parts. labeling resulted in many labeled mi- chromaffin cells start to appear, preced- To examine late stages of neural gratory neural crest cells (Fig. 8B,C). ing THϩ sympathetic neurons that dif- crest migration, wild type embryos The majority of labeled cells moved ferentiate in larvae Ͼ24 mm (not were injected as above, but at stage 35 along the ventromedial pathway or shown; see also Vogel and Model, 1977). TRUNK NEURAL CREST CELLS IN THE AXOLOTL EMBRYO 397

Some Neural Crest Cells tralateral side of the embryo and then The same injections also gave rise to Migrate Contralaterally follow analogous migratory pathways. neural crest derivatives and epidermis. This result suggests that the In birds and frogs, some neural crest Both Neural Crest and Rohon-Beard cells, epidermis, and cells cross the midline to migrate on Rohon-Beard Neurons neural crest cells share a common the contralateral side (Hall and Ho¨rst- origin from the dorsal neural folds Originate From Neural Folds adius, 1988; Couly et al., 1996; Le- (Fig. 10E,F). Douarin and Kalcheim, 1999). In axo- Rohon-Beard cells are transient sen- lotl, this possibility was first sory neurons in the dorsal trunk neu- investigated by Ho¨rstadius and Sell- ral tube of fish and amphibian em- DISCUSSION man (1946) using vital dyes. To extend bryos (Hall and Ho¨rstadius, 1988; Le Neural Crest Cells From this work with higher resolution label- Douarin and Kalcheim, 1999). They ing methods, we performed unilateral form a primary sensory system that Different Sites Within the grafts of left neural folds that had degenerates when dorsal root ganglia Neural Fold Have been injected with fluorescent dex- develop (Hughes, 1957; Roberts and Equivalent Developmental trans at the two-cell stage, thus label- Patton, 1985) and persist in axolotl at and Migratory Potential ing all neural fold cells in the graft. least through larval stages (DuShane, In amphibian embryos, it has been Prior to the onset of trunk neural 1938). proposed that different neural crest crest migration (stage 30/31), fluores- Rohon-Beard cells were identified derivatives may arise from different cent cells derived from a rhodamine using a combination of anti-tubulin staining and DiI injections in sections lateral or medial positions within the dextran–labeled neural fold were con- through the trunk of embryonic (4 neural folds. Testing this proposal in fined to the epidermis and neural tube mm) to larval (17 mm) stage axolotls axolotl, Brun (1985) had shown that on the grafted side and to a premigra- (Fig. 10A). In larvae, Rohon-Beard melanophores differentiate from the tory neural crest on top of the neural cells appeared as large neurons at the medial cranial neural fold (“L.N.F.” in tube, which also contained unlabeled dorsal-most aspect of the spinal cord Brun, 1985) but not from lateral epi- cells presumably from the contralat- (Fig. 10B), in contrast to interneurons dermis adjacent to lateral cranial neueral neural fold (Fig. 9A). The neural that are localized more laterally (Rob- ral fold. Furthermore Northcutt et al. crest was surrounded by basement erts, 2000). Both cell types were about (1996) had demonstrated in axolotl membranes as revealed by anti-fi- 2–3 times larger than neighboring that the medial wall of (cranial) neu- bronectin staining (Fig. 9A). At early cells of the spinal cord. In younger ral folds forms neural crest and tube migratory stages (stages 32–34), fluo- larvae (9 mm; stage 39/40), several derivatives, whereas the lateral wall rescent neural crest cells from a left, large cells with a rectangular or differentiates into dorsal and dorso- FITC-labeled neural fold entered the rounded shape were observed; these lateral ectoderm including neurogenic ventromedial pathways on both the can be distinguished from Rohon- placodes. The latter authors empha- ipsilateral and contralateral sides Beard cells because they reach from sized that “the neural folds . . . .must (Figs. 9B,C). When the left trunk neu- the ependymal layer to the outer sur- be qualified by division into medial ral fold contains a FITC- and the right face of the tube and have nuclei at and lateral fields which each give rise different levels from the center (Fig. a rhodamine dextran graft (Fig. 9D), to multiple lineages” (Northcutt et al., 10C,D; see Sauer, 1935). Thus, they contralateral migration of neural 1996). It also has been reported that are likely to be mitotic neuronal pre- crest cells from each fold was ob- neural plate adjacent to neural fold in cursors. served. We show rhodamine dextran– axolotl gives rise to melanophores To examine the origin of Rohon- labeled neural crest cells on the lat- whereas epidermis adjacent to fold Beard cells in the axolotl, DiI was eral and ventromedial pathway of the contributes to spinal and cranial gan- injected into the left trunk neural contralateral side (Fig. 9D). glia (Moury and Jacobson, 1990). fold of neurulae (stage 16), similar to In horizontal sections (Fig. 9F) those injections performed to exam- Because previous studies did not de- through the dorsal trunk of an embryo ine neural crest migratory patterns. finitively address which neural crest at stage 35 bearing a FITC-dextran Embryos were subsequently fixed at populations differentiate from posi- graft on its left side (Fig. 9E), single embryonic stage 33–34. In horizon- tions within medial or lateral neural labeled cells were clearly identified on tal sections through the dorsal neu- fold positions, we re-examined this is- the dorsolateral and ventromedial ral tube, several giant cell bodies at sue in the axolotl trunk. However, we pathway on both sides of the embryo. the left margin of the neural tube failed to find any difference in the Interestingly, those cells moving ven- were labeled with DiI and appeared pathways chosen, distances traversed, tromedially were observed within the several times larger than adjacent or derivatives produced by neural narrow sclerotome (Fig. 9B), although neurons. Furthermore, they ex- crest cells derived from either medial no labeled neural crest cells were ob- tended nerve fibers laterally into the or lateral trunk neural folds. Thus, served migrating through the myo- intersomitic clefts (Fig. 10E,F), char- there was no evidence for obvious me- tome as proposed by Vogel and Model acteristic of Rohon-Beard neurons diolateral differences or of sublin- (1977). These results show that neural (for details see Hughes, 1957; Taylor eages within the premigratory neural crest cells can migrate to the con- and Roberts, 1986; Roberts, 2000). crest in the trunk neural folds. 398 EPPERLEIN ET AL.

Fig. 9. Migration of trunk neural crest to the contralateral side. The left trunk neural fold of wild-type (D/-) hosts (stages 15–17) was replaced orthotopically by a neural fold from dark donors (same age) labeled with FITC or rhodamine dextran. A: Dorsal part of transverse section through the trunk of a stage 30/31 embryo to demonstrate the distribution of labeled tis- sues that were derived from the left, grafted rhodamine labeled neural fold before the onset of neural crest migration. Labeled cells from the left neural fold graft were observed in the left epidermis (epi) and left neural tube (nt). The neural crest (nc, arrow) cannot be divided into left or right portions; it is partially labeled and contains also unlabeled cells from the right fold. B–D: Transverse sections through different embryonic trunks at early migratory stages of the neural crest (stages 30/31–34). Embryos contain either one FITC-dextran labeled neural fold on their left side (B,C) or one FITC-dextran- on the left and one rhodamine-dextran labeled fold on their right side (D). Sections were immuno- stained with a primary polyclonal anti-fibronectin antibody visualized by a FITC-conjugated (A) or a Cy3-conjugated secondary antibody (B,C). Fluorescent neural crest cells entered ventromedial (arrowheads) and lateral pathways (arrows) on the ipsi- and contralateral side. In B, fluorescent neural crest cells on the left medial pathway pass through the sclerotome. E: Left side of a larva (stage 35) bearing a FITC-dextran-labeled trunk neural fold graft on the left side. F: Horizontal section through larva shown in E at a middle neural tube level. Single labeled cells are visible on the medial (arrowheads) and lateral pathway (arrows) of both the ipsi- and contralateral side. B,C,F: counterstaining with dapi. Fig. 9. Fig. 10. Origin and distribution of Rohon-Beard neurons. Reinvestigation of Rohon-Beard cells on sections through the axolotl trunk from embryos (stage 33) to larvae (17 mm length). A: Schematic showing planes of sectioning in an axolotl embryo at stage 34. B: In transverse sections through the trunk of larvae (12 mm), large neurons (arrowheads) were observed at the dorsal-most aspect of the spinal cord following immunostaining with anti-tubulin, which fulfills the criteria of Rohon-Beard cells. C,D: Horizontal section through the dorsal neural tube of a younger larva (stage 39/40). C: Immunostaining with primary anti-tubulin and secondary Cy3-conjugated antibody followed by dapi staining. D: Same section uncolorized. Large cells with a rectangular (arrow) or rounded shape (arrowhead) are mitotic precursors and not Rohon-Beard cells because they are continuous from the ependymal layer to the outer surface of the neural tube and have their nuclei at different distances from the central lumen. E,F: Horizontal section (E, DiI, and dapi; F, laser scanning micrograph) through the dorsal neural tube of an embryo (stage 33–34), which had received DiI injections into the left trunk neural fold. Three giant labeled cell bodies at the left margin of the neural tube extend nerve fibres laterally into the intersomitic clefts. They are likely Rohon Beard cells. nt, neural tube; som, somite; nf, nerve fibre. Fig. 10. TRUNK NEURAL CREST CELLS IN THE AXOLOTL EMBRYO 399

Anterior Trunk Neural Fold (stage 33) of the trunk neural crest in sis. Similarly, we observed glial cells Gives Rise to More Laterally axolotl revealed that neural crest cells and neurons migrating dorsolaterally Migrating Cells Than in wild type embryos migrate first dor- at least from anterior trunk neural solaterally and then ventromedially folds in the axolotl. Posterior Fold and are more numerous on the former The migration pathways followed In dorsal trunks of white mutant em- route. In contrast, dorsolateral migra- by neural crest cells in axolotl are sur- bryos injected with DiI at stage 25, a tion is restricted in the white mutant prisingly similar to those previously quantitative investigation showed whereas ventromedial migration (Fig. described in the chick. For example, that labeled cells were distributed on 6F) resembles that in the wild type axolotl neural crest cells migrate lat- the lateral migratory route at stage (Figs. 6C, 7C). At late-migratory erally between the epidermis and a 36–37 and stopped in an area of the stages (stage 35), there was still mod- narrow dermomyotome (referred to as middle lateral line nerve (Fig. 6F). erate migration of neural crest cells “dermatome” in Sobkow et al., 2006; These DiI-labeled cells (14 per 5 indi- along the dorsolateral pathway, but Epperlein et al., unpublished data), viduals) formed only a fraction (about most labeled cells moved along the comparable to the dorsolateral migra- 10%) of all lateral melanophores ventromedial pathway or into the dor- tion between the epidermis and der- counted in white controls (ϳ30/indi- sal fin. momyotome observed in chick (Erick- vidual); however, it was unclear Our results contrast with previous son et al., 1992). Also as in chick whether their arrest at the lateral line findings by Vogel and Model (1977), (Bronner-Fraser, 1986), medial neural nerve was coincidental or reflected a who, based on (3H)thymidine-labeling crest cells in axolotl migrate through a contribution to derivatives of the lat- of neural folds, suggested that ventro- sclerotome (Fig. 9B), albeit very nar- eral line system. To distinguish be- medial migration of presumptive neu- row. tween these possibilities, we injected rons and glial cells at stage 34 had It is uncertain whether an earlier DiI focally into either anterior or pos- advanced similar to pigment cell mi- lateral migration of pigment cells is terior positions of the trunk neural gration along the dorsolateral path- an evolutionary advantage for axolotl fold (stage 16). The number of labeled way (staging according to Schrecken- embryos. Pigment cells might provide cells on the lateral route was fourfold berg and Jacobson, 1975; stage 34 is the larvae with disguise. On the other higher for anterior than posterior in- equivalent to stage 32 in the table of hand, Rohon-Beard cells constitute a jections. While anterior injections con- Bordzilovskaya et al., 1989; see Ep- primary sensory system in axolotl, tributed labeled cells to the lateral perlein and Junginger, 1982, for com- perhaps making formation of deriva- line primordium (Northcutt et al., parison). Similarly, we found no evi- tives along the ventromedial pathway 1994), posterior injections failed to do dence for migration of neural crest less crucial at early times in develop- so. Thus, labeled cells migrating from cells through the somite (stage 30) as ment. an injected anterior neural fold may Vogel and Model (1977). contain neuronal (ganglionic) and The early migration of axolotl neu- Dual Neural Crest and glial precursors for the lateral line in ral crest along the dorsolateral path- Somite Origin for the addition to few pigment cells. way contrasts with that in several Mesenchyme of the Dorsal Further evidence for this assump- other amniotes/anamniotes. Dorsolat- Fin tion was corroborated by a qualitative erally, neural crest migration is about characterization of the labeled cells 300–400 ␮m advanced compared to Classically, the mesenchyme of the distributed, either after injecting DiI medial migration. Dorsolateral cells dorsal fin in lower vertebrates has into anterior or posterior trunk neural need about 1 day to cover such a dis- been assumed to be neural crest de- fold as before (Figs. 4A, 5A) or by tance (Fig. 6B,C). One possible expla- rived. The dorsal fin epidermis is in- grafting GFP transgenic neural folds nation for this earlier lateral migra- duced by the neural crest and removal into white hosts (Fig. 4C–H). The lat- tion is that presumptive pigment cells of the neural folds leads to the absence ter experiments, in particular, left no may be committed to their fate prior to of the dorsal fin (Raven, 1931; DuSh- doubt that midtrunk neural crest does initiation of migration (Epperlein and ane, 1935; Bodenstein, 1952). Follow- not supply glial and neuronal cells for Lo¨fberg, 1984) and that this prespeci- ing injections of DiI into dorsal mid the lateral line. The results confirmed fication is essential for migrating dor- trunks of axolotl tailbuds at stage 25 the view that these cells appear to solaterally. Thus, pigment cells in ax- (Fig. 6A–F) and 33 (Fig. 7A–F), we arise from anterior trunk neural fold olotl might follow the “phenotype- observed only a few labeled cells in the levels behind the otic placode (North- directed model” for neural crest developing dorsal fin after 2 days cutt et al., 1994), which migrate pos- migration as proposed for pigment cell (stages 36–37). When DiI was injected teriorly together with or in the vicinity precursors in the chick (Erickson and into larvae at stage 35 (Fig. 8A–C), the of lateral line primordia/lateral line Reedy, 1998). However, results by dorsal fin mesenchyme was abun- nerves. Wakamatsu et al. (1998) argue dantly labeled after 4 days (stage 40). against such a general prespecifica- Although DiI injections hit only a frac- Neural Crest Migrates First tion of the neural crest. Chicken neu- tion of all neural crest cells present at ral crest cells migrating dorsolaterally a certain stage, we found the number Laterally and Then Medially appear initially to include neuronal of labeled mesenchymal cells 2 days DiI labeling of premigratory stages cells in addition to melanocytes and after an injection (stages 36–37) sur- (stage 25) and early migratory stages these are later eliminated by apopto- prisingly small. 400 EPPERLEIN ET AL.

By grafting single GFPϩ labeled Trunk neural fold/neural tube. and mirror-imaged to the left side of a somites into unlabeled hosts, we re- summary section that represents the DiI injections were made into three cently found that the dermomyotome group (Figs. 2,3,6–8, Table 1). different rostrocaudal locations of the participates in early stages of mesen- left lateral or medial trunk neural fold chyme formation of the dorsal fin of dark neurulae (stages 15–19) to de- Controls. The distribution of mela- (Sobkow et al., 2006). Thus, in axolotl termine whether regional differences nophores in whole white larvae was neural crest appears to make only a exist in migratory potential and cell determined between gills and cloaca limited contribution to the mesen- fate at different mediolateral posi- at stages 36 and 42 in order to obtain chyme of the developing dorsal fin at tions within the neural folds (Figs. 1, a measure that would allow compari- early larval stages; later its contribu- 2A–F; Table 1). Injected embryos were son with the migration of Di-labeleld tion appears to increase as shown by fixed 2 or 4 days post-injection and cells in sectioned white embryos. At DiI labeling of dorsal trunks at stage vibratomesectioned (transverse sec- stage 36, six larvae were used and 35 (Fig. 8B,C). tions, see below). subjected to Dopa treatment (Epper- DiI was injected into one middle an- lein and Lo¨fberg, 1984) in order to en- terior or posterior location of the left hance the visibility of melanin pig- EXPERIMENTAL trunk neural fold of dark and white ment in melanophores. At stage 42 PROCEDURES neurulae (stage 16) to examine when melanophores are clearly visi- whether and in which way trunk neu- ble, we used 10 larvae. In both stages, Embryos ral crest in these sites contributes to we found a mean value of ϳ30 mela- Wild-type (dark, D/-) and white mu- lateral line nerves (Figs. 1, 3A–C, 4, 5; nophores in the dorsolateral trunk. tant (dd) embryos of the Mexican ax- Table 1). The anterior injection site olotl (Ambystoma mexicanum) were was localized at the level of the pro- FITC- and Rhodamine obtained from the axolotl colony in spective 2nd somite, the posterior one Dextran Injections Bloomington, Indiana, or from our in the most posterior positon of the left Dark axolotl embryos at the two- to own colonies in Dresden. GFP em- fold at the level of still unsegmented mesoderm. White injected embryos four-cell stage were kept at room tem- bryos were spawned from a ␤-actin perature in a deepening of an agar promoter-driven GFP germline trans- were fixed 2, 4, and 5 days post-injection, dark ones after 4 and 5 days and dish filled with 5% ficoll in 1/10 Stein- genic animal that had been produced berg solution containing antibiotics. by plasmid injection (Sobkow et al., mostly investigated as whole mounts. Some embryos fixed 4 days after the One blastomere was injected with 2006). Embryos were kept in tap wa- FITC- or rhodamine dextran (Molecu- ter at room temperature or at 7–8°C injection were vibratomesectioned (transverse sections, see below). If polar Probes; 50 mg/ml Steinberg solu- and were staged according to the nor- sitions or identity of labeled cells were tion; 50 nl/injection) through the egg mal table of Bordzilovskaya et al. unclear, sections were counterstained membrane using an IM 300 injector (1989). Before being used for grafting, with anti-fibronectin to visualize tis- (Narshige). After one day, the em- embryos were washed thoroughly sue borders. bryos were transferred into normal with tap water and sterile Steinberg DiI injections were made into three strength saline. At the neurula stage solution (Steinberg, 1957) containing different rostrocaudal locations of the parts of fluorescent trunk neural folds antibiotics (Antibiotic-Antimycotic, dorsal midtrunk neural tube of dark were used for grafting. Gibco) and then decapsulated me- and white embryos at stages 25, 33, chanically. and 35 to examine whether there were Axolotl Embryos Expressing different migratory potentials of Green Fluorescent Protein trunk neural crest cells (Figs. 6A–F, DiI-Injections (GFP) 7A–F, 8A–C; Table 1). Embryos were CellTracker CM-DiI (C-7000; Molecu- fixed at different times post-injection White axolotl embryos transgenic for lar Probes, Eugene, OR) was dissolved and vibratomesectioned (transverse GFP were produced by mating a male GFP transgenic with a normal white in absolute ethanol to a concentration sections, see below). female (Sobkow et al., 2006). ␤-actin of 1 mg/ml and further diluted in 4 or promoter-driven GFP expression first 9 parts of 10% sucrose in water just Quantification of DiI-labeling became clearly visible in the neurula before injection. For injection, either and controls. stage. Trunk neural fold material was glass micropipettes were used that Quantification. For a quantitative grafted into white donors in order to were backfilled with DiI solution and evaluation of the distribution of DiI- investigate a possible contribution of attached to a Parker Hannifin Corpo- labeled neural crest cells, approxi- neural crest derivatives to the lateral ration Picospritzer II assembly or mately 20 transverse sections (100 ␮m line system (see below). glass capillaries that were connected thick) were cut through the trunk of to a IM 300 injector (Narshige). DiI DiI-labeled embryos until stage 34 Neural Fold Grafting was injected into embryos as indicated and 30 until stage 39/40. Five speci- Experiments below and monitored with different mens were sectioned per group. The stereo and epifluorescence micro- labeled cells in sections through all Head and trunk neural fold material scopes. specimens of a group were counted pre- or postlabeled in different ways TRUNK NEURAL CREST CELLS IN THE AXOLOTL EMBRYO 401

(see below) from sterile neurulae Histochemistry/Histology/ In situ–hybridized wild-type embryos (stages 15–17) was grafted isochroni- Immunostaining were washed for 2 hours in several cally into orthotopic locations of dark changes of PBS and bleached for 1–2 White larvae at stage 35/36 were sub- or white hosts. Before the operations, hr in 1% H202/5% formamide/.5ϫSSC host neurulae were placed into a deep- jected to the Dopa reaction (Epperlein under intense illumination to reveal ening of an agar dish filled with cold and Lo¨fberg, 1984) in order to enhance hybridization Steinberg saline. Grafting was per- the visibilitiy of melanophores and fa- cilitate their exact counting. The num- formed under sterile conditions using Laser Scanning Microscopy tungsten needles. bers of melanophores were needed for Whole left trunk neural folds were assessing the numbers of DiI labeled Horizontal vibratome sections (100 transplanted orthotopically from dark cells in white embryos. ␮m) through the dorsal trunk of DiI- FITC- dextran (nϭ5) or rhodamine Some vibratome sections (100 ␮m) injected embryos (stage 33–34) were dextran (nϭ2) labeled neurulae into were cut through PFA-fixed DiI-, also investigated in the laser-scan- dark unlabeled hosts or from white FITC- or rhodamine dextran-, or GFP- ning microscope (Zeiss; ϫ10 objective) GFPϩ transgenic neurulae (nϭ6) into labeled embryos and counterstained to gain a better resolution of Rohon white unlabeled hosts. In addition with an anti-fibronectin antibody to Beard cells and their nerve fibres. small left midtrunk (nϭ5) or posterior visualize tissue borders. Sections were trunk neural fold fragments (nϭ5) then incubated with a polyclonal anti- Image Analysis were grafted orthotopically from fibronectin antibody (Dako, Hamburg, Whole embryos and sections were an- white GFPϩ donors into white hosts. Germany) followed by a FITC- or Cy3- alyzed with an epifluorescence micro- In some cases, both trunk neural folds conjugated goat-anti-rabbit secondary scope and images were captured with of dark embryos were removed and antibody (Dianova, Hamburg, Ger- a Spot digital camera (Visitron). Us- the left fold replaced with a FITC- and many). Some sections were stained ing Adobe Photoshop software, the the right one with a rhodamine dex- with dapi (0.1–1 ␮g/ml in PBS) to contrast and brightness of captured tran–labeleld neural fold fragment. mark cell nuclei. Vibratome sections images were optimized, and bright Finally, both trunk neural folds of through larvae that had received one field images of whole embryos were white hosts were removed and only focal DiI injection into the left anterior combined with corresponding fluores- the right fold replaced with a GFPϩ trunk neural fold were also stained for cence images. Separate images were fold (nϭ10). Donor and hosts were at anti-tubulin or anti-GFAP to reveal captured from sectioned material la- stages 15–17. After 2–4 days, migra- lateral line nerves or glia, respec- beled with DiI, fluorescent-dextrans tion of labeled trunk neural crest cells tively. For anti-tubulin staining, a pri- or GFP (neural crest cells), anti-fi- through somites (FITC-label) or con- mary monoclonal antibody (W. Half- bronectin (tissue borders), anti-tubu- tribution to the lateral line (GFPϩ ter, Pittsburgh) was used followed by lin (lateral line nerve), anti-GFAP transgenic) was analyzed. a FITC-conjugated goat-anti-mouse secondary antibody (Dianova). Stain- (glial cells), and dapi (cell nuclei), being of glia was carried out with a fore combinations of two or three of Fixation, Embedding, and primary polyclonal GFAP-antibody them were superimposed into a single Sectioning (Dako) followed by a FITC-conjugated image. Embryos were fixed in 4% paraformal- goat-anti-rabbit secondary antibody dehyde (PFA) in PBS overnight and (Dako). Because staining with fluores- ACKNOWLEDGMENTS kept either in fixative for a later use in cence techniques was sometimes weak This research was supported by DFG histology or were transferred directly (anti-tubulin) or had high background (EP8/7-1) to H.H.E., by NS051051 to into 100% methanol and kept there at (anti-GFAP), some unlabeled control M.B.F., and a James H. Zumberge Re- Ϫ20°C before use in in situ hybridiza- larvae of a similar age as DiI-injected search and Innovation Fund and a tion. Some individuals were examined specimens were embedded into paraf- Donald E. and Delia B. Baxter Foun- in an epifluorescence microscope for fin. Microwaved transverse sections (5 dation award to M.A.J.S. H.H.E. the distribution of DiI- or GFP-labeled ␮m) were stained with the same pri- thanks G. Northcutt (San Diego) and cells. For histology, specimens were mary antibodies but with biotinylated G. Schlosser (Bremen) for valuable in- washed in PBS for 1 h, sectioned with secondary antibodies and tertiary avi- formation on the lateral line, I. Beck, a vibratome (vibratome series 1000 din-peroxidase complexes (Vectastain P. Mirtschink, F. Heidrich, and T. sectioning system, Ted Pella, Inc.) in ABC-Kit). Schwalm for analysis of cell migration transverse or horizontal planes at 100 or computer work, and S. Bramke for ␮m thickness, stained as indicated be- histology. In Situ Hybridization low, and examined with an epifluorescence microscope. Some PFA-fixed Axolotl Snail riboprobes were pre- REFERENCES embryos were embedded into paraffin pared as previously described (Epper- for microwave-treatment of trans- lein et al., 2000). In situ hybridization Bodenstein D. 1952. Studies on the devel- verse sections (5 ␮m), which were was performed on albino and wild- opment of the dorsal fin in amphibians. J Exp Zool 120:213–245. used for immunohistochemical dem- type axolotl embryos as described by Bordzilovskaya NP, Dettlaff TA, Duhon onstration of lateral line nerves and Henrique et al. (1995), with the addi- ST, Malacinski GM. 1989. Developmen- glia. tion of an overnight wash in MAB-T. tal-stage series of axolotl embryos. In: 402 EPPERLEIN ET AL.

Armstrong JB, Malacinski GM, editors. Erickson CA, Duong TD, Tosney KW. 1992. neurogenic and melanogenic neural crest The developmental biology of the axolotl. Descriptive and experimental analysis of sublineages. Development 130:321–330. New York: Oxford University Press. p the dispersion of neural crest cells along Macmillan GJ. 1976. Melanoblast-tissue 201–219. the dorsolateral path and their entry interactions and the development of pig- Bronner-Fraser M. 1986. Analysis of the into ectoderm in the chick embryo. Dev ment pattern in Xenopus larvae. J Em- early stages of trunk neueral crest mi- Biol 151:251–272. bryol Exp Morphol 35:463–484. graion in avian embryos using nmono- Gilmour, DT, Maischein, HM, Nu¨ sslein- Moury JD, Jacobson AG. 1989. Neural fold clonal antibody HNK-1 Dev Biol 115:44– Volhard, 2002. Migration and function of formation at newly created boundaries 55. a glial subtype in the vertebrate periph- between neural plate and epidermis in Brun RB. 1985. Neural fold and neural eral nervous system. Neuron 34:577–588 the axolotl. Dev Biol 133:44–57. crest movement in the Mexican Guillory G, Bronner-Fraser M. 1986. An in Moury JD, Jacobson AG. 1990. The origins salamander Ambystoma mexicanum. J vitro assay for neural crest cell migra- of neural crest cells in the axolotl. Dev Exp Zool 234:57–61. tion through the somites. J Embryol Exp Biol 141:243–253. Chibon P. 1966. Analyse expe´rimentale de Morphol 98:85–97. Northcutt RG, Catania KC, Criley BB. la re´gionalisation et des capacite´s mor- Hall BK, Ho¨rstadius S. 1988. The neural 1994. Development of lateral line organs phogeńe´tiques de la crete neurale chez crest. London: Oxford University Press. in the axolotl. J Comp Neurol 340:480– l‘Amphibien Urode`le Pleurodeles waltlii. Henrique D, Adam J, Myat A, Chitnis A, 514. Michah Mem Soc Fr Zool 36:1–107. Lewis J, Ish-Horowicz D. 1995. Expres- Northcutt RG, Barlow LA, Catania KC, Collazo A, Bronner-Fraser M, Fraser SE. sion of a Delta homologue in prospective Braun CB. 1996. Developmental fate of 1993. Vital dye labelling of Xenopus lae- neurons in the chick. Nature 375:787– the lateral and medial walls of the neu- vis trunk neural crest reveals multipo- 790. ral folds in axolotls. Am Zool 36:74A. tency and novel pathways of migration. Horstadius S, Sellman S. 1946. Experi- Raven CP. 1931. Zur Entwicklung der Development 118:363–376. mentelle untersuchungen u¨ ber die deter- Ganglienleiste. I. Die Kinematik der Collazo A, Fraser SE, Mabee PM. 1994. A mination des knorpeligen kopfskelettes Ganglienleistenentwicklung bei den dual embryonic origin for vertebrate bei Urodelen. Nova Acta Regiae Soc Sci Urodelen. Wilhelm Roux Arch Entwick- mechanoreceptors. Science 264:426– Ups Ser 13:1–170. lungsmech Org 125:210–293. 430. Hughes A. 1957. The development of the Reedy MV, Faraco CD, Erickson CA. 1998. Couly G, Grapin-Botton A, Coltey P, Le primary sensory system in Xenopus lae- The delayed entry of thoracic neural Douarin NM. 1996. The regeneration of vis (Daudin). J Anat 91:323–338. crest cells into the dorsolateral path is a the cephalic neural crest, a problem re- Keller RE, Lofberg J, Spieth J. 1982. Neu- consequence of the late emigration of visited: the regenerating cells originate ral crest cell behavior in white and dark melanogenic neural crest cells from the from the contralateral or from the ante- embryos of Ambystoma mexicanum: epi- neural tube. Dev Biol 200:234–246. rior and posterior neural fold. Develop- dermal inhibition of pigment cell migra- Roberts A. 2000. Early functional organi- ment 122:3393–3407. tion in the white axolotl. Dev Biol 89: zation of spinal neurons in developing Dalton HC. 1953. Relations between devel- 179–195. lower vertebrates. Brain Res Bull 53:585– oping melanophores and embryonic tis- Knecht AK, Bronner-Fraser M. 2002. In- 593. sues in the Mexican axolotl. Gordon M duction of the neural crest: a multigene Roberts A, Patton DT. 1985. Growth cones (ed) Pigment cell growth. New York: Ac- process. Nat Rev Genet 3:453–461. and the formation of central and periph- ademic Press. P 17–27. Krotoski DM, Fraser SE, Bronner-Fraser eral neurites by sensory neurones in am- DuShane GP. 1935. An experimental study M. 1988. Mapping of neural crest path- phibian embryos. J Neurosci Res 13:23– of the origin of pigment cells in am- ways in Xenopus laevis using inter- and 38. phibia. J Exp Zool 72:1–31. intra-specific cell markers. Dev Biol 127: DuShane GP. 1938. Neural fold derivatives 119–132. Sauer FC. 1935. Mitosis in the neural tube. J Comp Neurol 62: 377-405. in the amphibia: pigment cells, spinal Le Douarin NM. 1969. Particularite´s du ganglia and Rohon-Beard cells. J Exp noyau interphasique chez la Caille Schlosser G. 2002. Development and evo- Zool 82:193–215. japonaise (Coturnix coturnix japonica). lution of lateral line placodes in amphib- Elena de Bellard M, Bronner-Fraser M. Utilisation de ces particularite´s comme ians I. Development. Zoology (Jena) 105: 2005. Neural crest migration methods in ’marquage biologique’ dans les recher- 119–146. the chicken embryo. Methods Mol Biol ches sur les interactions tissulaires et les Schreckenberg GM, Jacobson AG. 1975. 294:247–267. migrations cellulaires au cour de l‘onto- Normal stages of development of the ax- Epperlein HH, Junginger M. 1982. The gene`se. Bull Biol Fr Belg 103:435–452. olotl Ambystoma mexicanum. Dev Biol normal development of the newt, Tritu- Le Douarin NM, Kalcheim C. 1999. The 42:391–400. rus alpestris (Daudin). Amphibia Rep- neural crest. Cambridge: Cambridge Selleck MA, Bronner-Fraser M. 1995. Ori- tilia 2:295–308. University Press. gins of the avian neural crest: the role of Epperlein HH, Lo¨fberg J. 1984. Xan- Le Douarin NM, Teillet MA. 1974. Experi- neural plate-epidermal interactions. De- thophores in chromatophore groups of mental analysis of the migration and dif- velopment 121:525–538. the premigratory neural crest initiate ferentiation of neuroblasts of the auto- Selleck MA, Bronner-Fraser M. 1996. The the pigment pattern of the axolotl larva. nomic nervous system and of genesis of avian neural crest cells: a clas- Roux Arch Dev Biol 193:357–369. neurectodermal mesenchymal deriva- sic embryonic induction. Proc Natl Acad Epperlein HH, Lo¨fberg J. 1990. The devel- tives, using a biological cell marking Sci USA 93:9352–9357. opment of the larval pigment patterns in technique. Dev Biol 41:162–184. Serbedzija GN, Bronner-Fraser M, Fraser Triturus alpestris and Ambystoma mexi- Lo¨fberg J, Ahlfors K, Fallstrom C. 1980. SE. 1989. A vital dye analysis of the tim- canum. Adv Anat Embryol Cell Biol 118: Neural crest cell migration in relation to ing and pathways of avian trunk neural 1–99. extracellular matrix organization in the crest cell migration. Development 106: Epperlein HH, Meulemans D, Bronner- embryonicaxolotltrunk.DevBiol75:148– 809–816. Fraser M, Steinbeisser H, Selleck MA. 167. Serbedzija GN, Fraser SE, Bronner-Fraser 2000. Analysis of cranial neural crest mi- Lo¨fberg J, Nynas-McCoy A, Olsson C, Jon- M. 1990. Pathways of trunk neural crest gratory pathways in axolotl using cell sson L, Perris R. 1985. Stimulation of cell migration in the mouse embryo as markers and transplantation. Develop- initial neural crest cell migration in the revealed by vital dye labelling. Develop- ment 127:2751–2761. axolotl embryo by tissue grafts and ex- ment 108:605–612. Erickson CA, Reedy MV. 1998. Neural tracellular matrix transplanted on mi- Smith SC, Lannoo MJ, Armstrong JB. crest development: the interplay be- crocarriers. Dev Biol 107:442–459. 1990. Development of the mechanorecep- tween morphogenesis and cell differenti- Luo R, Gao J, Wehrle-Haller B, Henion PD. tive lateral-line system in the axolotl: ation. Curr Top Dev Biol 40:177–209. 2003. Molecular identification of distinct placode specification, guidance of migra- TRUNK NEURAL CREST CELLS IN THE AXOLOTL EMBRYO 403

tion, and the origin of neuromast polar- Steinberg MS. 1957. A nonnutrient culture Vogel DL, Model PG. 1977. Development of ity. Anat Embryol (Berl) 182:171–180. medium for amphibian embryonic tissue. the sympathetic system in the Mexican Sobkow L, Epperlein HH, Herklotz S, Carnegie Inst Washington Year b56:347– axolotl, Ambystoma mexicanum. Dev Straube WL, Tanaka EM. 2006. A germ- 348. Biol 56:76–96. line GFP transgenic axolotl and its use to Taylor JS, Roberts A. 1983. The early de- Wakamatsu Y, Mochii M, Vogel KS, track cell fate: dual origin of the ﬁn mes- velopment of the primary sensory neu- Weston JA. 1998. Avian neural crest-de- enchyme during development and the rones in an amphibian embryo: a scan- rived neurogenic precursors undergo ap- fate of blood cells during regeneration. ning electron microscope study. J optosis on the lateral migration path- Dev Biol 290:386–397. Embryol Exp Morphol 75:49–66. way. Development 125:4205–4213.