SEARCH

SEARCH BY CITATION

Keywords:

  • Xenopus;
  • tail;
  • fin;
  • epidermis;
  • neural crest;
  • fate map;
  • ablation

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

It has been known since the 1930s that the dorsal fin is induced by the underlying neural crest. The inducer of the ventral fin, however, has remained elusive. We have investigated the source of the inducer of the ventral fin in Xenopus and show that it is the ventral mesoderm and not the neural crest. This induction takes place during mid-neurula stages and is completed by late neurulation. In terms of cell composition, the dorsal fin mesenchyme core arises from neural crest cells, while the mesenchyme of the ventral fin has a dual origin. The ventral fin contains neural crest cells that migrate in from the dorsal side of the embryo, but a contribution is also made by cells from the ventral mesoderm. Developmental Dynamics 230:461–467, 2004. © 2004 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The Xenopus fin starts to develop along the trunk and tail of the embryo at the tail bud stage and continues to grow until the start of metamorphosis when it begins to regress. The fin has an important role in tadpole locomotion; large defects causing the tadpole to swim in a lop-sided manner. The fin consists of flattened epidermal cells elevated into a keel-like structure by a supporting core of mesenchyme and extracellular matrix. The mesenchyme separates the epidermis on either side of the fin, except at the crest where the two sides meet and fuse to form a single layered epithelial structure. This structure gives the fin its characteristic two-part appearance. The dorsal fin mesenchyme has been shown to be derived from the neural crest (Raven, 1931; Smith et al., 1994). Neural crest migrates out of the neural tube in a wave along the body starting at the head at stage 25. Migration finishes rostrally at stage 41 and more caudally around stage 46. The formation of the dorsal fin is summarized in Figure 1. As the fin expands, it becomes filled with a thick extracellular matrix (ECM) in which the neural crest-derived cells sit (Tucker, 1986). This ECM consists of collagen, fibronectin, tenascin, and glycosaminoglycans, such as hyaluronan and sulfated proteoglycans (Tucker and Erickson, 1986; Epperlein et al., 1988). According to these authors, the melanophores rarely migrate into the fin mesenchyme, due to the high levels of glycosaminoglycans, but sit below the dorsal fin in a stripe where there is a high concentration of fibronectin.

thumbnail image

Figure 1. Diagram representing the stages of fin induction and formation. A: Stage 25. B: Stage 28. C: Stage 31. D: Stage 32.

Download figure to PowerPoint

The ventral fin is also composed of neural crest derived mesenchyme. The neural crest cells migrate into the ventral fin by means of two routes (Collazo et al., 1993). The tail tip pathway extends along the dorsal fin around the tip of the tail and into the ventral fin at stage 27–32. The enteric pathway extends in a direct ventral direction from the trunk toward the proctodaeum where the neural crest enters the ventral fin. This pathway develops at a much later stage than the tail tip pathway, the cells populating the ventral fin once it is fully formed between stages 41 and 46, by which time the tadpole is already a free-swimming larva. The migration of neural crest cells in the fin appears to be an active process, passive transport of the cells as the tail grows being only of secondary importance (Collazo et al., 1993).

The neural crest not only provides the mesenchyme of the fin but is also involved in the induction of the epidermal part of the dorsal fin. When the trunk neural folds are extirpated in axolotls or lamprey embryos the dorsal fin does not form at the appropriate body level (DuShane, 1935; Newth, 1956). Induction of the dorsal epidermis appears to occur soon after neural crest closure at the late neurula stage, while fin formation does not occur until mid-late tail bud stages. Once the dorsal epidermis has been induced to form a fin, but before the crest has started to move up into the developing fin bud, the epidermis can be transplanted to a lateral site over the somites of a host embryo and will recruit mesenchyme from the underlying tissue to form an ectopic fin (Bodenstein, 1952). Dorsal epidermis from earlier stages, before induction by the neural crest, was unable to recruit mesenchymal cells.

The induction of the ventral fin has been neglected in comparison to the dorsal fin. The ventral fin runs from the tail tip to the proctodaeum and is morphologically continuous with the dorsal fin. In some experiments, ablation of the most caudal neural crest has led to defects in the ventral fin, leading to the conclusion that the neural crest induces both the dorsal and ventral fins (Ford, 1949). In other experiments, ventral fin defects were only seen after ablation of the non-neuronal tissue ventral to the residual slit-like blastopore at the neurula stages (Bytinski-Salz, 1936; Smithberg, 1954). However, interpretation of this result is complicated by the fact that excision of the tissue ventral to the blastopore will ablate part of the epidermis fated to develop into the ventral fin (Tucker and Slack, 1995). The neural crest appears to provide a small proportion of the ventral fin mesenchyme. The source of the remaining mesenchyme has been predicted to be both the postanal endoderm (Elsdale and Davidson, 1983; Bytinski-Salz, 1936) and the ventral mesoderm (van Aufsess, 1941). We therefore decided to investigate whether the neural crest was responsible for induction of both the dorsal and ventral fins in Xenopus, and to determine from where the bulk of the ventral fin mesenchyme originated. To do this, we have adopted a strategy of explant culture, ablation, and fate mapping.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Ablation of the Posterior Neural Crest

To investigate the role of the neural crest in induction of the ventral fin, the most caudal third of the neural folds were ablated in stage 13 neurulae (red regions in Fig. 2A). The position of the neural crest cells at this stage was inferred from the published expression pattern of the neural crest specific markers Xsnail and Xslug (Essex et al., 1993; Mayor et al., 1995). The operated embryos were grown to stage 41, by which time a full scale fin normally forms (Fig. 2B). In all cases, no dorsal fin developed above the regions where the neural crest had been ablated (Fig. 2C), agreeing with previous studies using axolotols (DuShane, 1935). No melanocytes were found within the tail from the point where development of the dorsal fin ceased (Fig. 2C, arrow), confirming the loss of neural crest from this region. The caudal tip of the dorsal fin and the whole ventral fin, however, was unaffected by this operation (Fig. 2C). So the induction of the ventral fin appears not to involve the neural crest. When the neural folds are ablated at stage 13, the presumptive fin epidermis (green areas in Fig. 2A) should not be affected, as this lies lateral to the folds at early neurula stages (Tucker and Slack, 1995). If a portion of this prospective fin epidermis is excised, the wound area heals over remarkably well and normal ventral and dorsal fins form (Fig. 2D).

thumbnail image

Figure 2. Role of the neural crest. A: Schematic of neurula embryo stage 13. Dorsal and side view. Red areas represent the posterior neural crest ablated in C. Green areas represent the posterior dorsal fin epidermis ablated in D. Blue areas represent the ventral region. Pink areas represent the lateral region. B: Control embryo stage 41. C: Ablation of the posterior neural crest at stage 13. Embryo cultured to stage 41. Embryo shows loss of the dorsal fin from arrow, except at the very tip. The ventral fin remains unaffected. D: Ablation of the dorsal fin epidermis at stage 13. Embryos cultured to stage 41. No fin defect is observed. n = 35 for C, n = 30 for D.

Download figure to PowerPoint

Induction of Fin Development in Explant Culture

Explants were dissected from stage 13 neurulae and cultured in isolation or in coculture until control embryos reached stage 40. In isolation, presumptive dorsal or ventral fin epidermis (green and blue areas, respectively, in Fig. 2A) rounded up into semitransparent vesicles (Fig. 3A). A similar rounding up of epidermis was observed after epidermis from lateral regions of the embryo, not fated to form fin, was cultured (pink areas in Fig. 2A; data not shown). However, cocultures of neural crest cells plus epidermis from lateral, ventral, or dorsal regions of the embryo developed thin ruffles of flattened cells that projected out from a central mass of tissue (Fig. 3B). These transparent fringes took on the characteristics of fin tissue (compare with fins seen in controls, Fig. 2B). In contrast, when lateral, ventral, or dorsal epidermis was cocultured with mesoderm, or mesoderm and endoderm, from lateral regions of the neurula embryo the explanted tissue remained dense with no projections (Fig. 3C). To further test whether neural crest cells had the ability to induce fin formation in early epidermis from other positions, fluorescein dextran amine (FDA) -labeled epidermis from regions of the neurula embryo, which did not normally contribute to fin development (pink regions in Fig. 2A, see fate map Tucker and Slack, 1995), were grafted above the dorsal neural tube of host embryos where the host epidermis had been ablated. These grafts were carried out at stage 20, once the neural tube has closed. In each case, dorsal fins developed normally with a substantial contribution from the FDA-labeled donor (Fig. 3D). Neural crest cells, thus, appear to have the ability to induce fin development in all early epidermis, not just in that normally fated to become fin.

thumbnail image

Figure 3. Explant cocultures. A: Uninduced epidermis from alongside the neural folds and under the blastopore, extirpated at stage 13 and cultured until controls reached stage 40. See green and blue areas in Figure 2A. B: Lateral, ventral, and dorsal epidermis cocultured as a sandwich with neural folds. See green, blue, and pink areas, plus red area, in Figure 2A. Note the ruffles of fin induced around the explant. C: Lateral, ventral, and dorsal epidermis cocultured with lateral mesoderm and endoderm. See green, blue, and pink areas, plus underlying pink area in Figure 2A. No fin induction is observed. D: Fluorescein dextran amine–labeled non-fin epidermis grafted at stage 20 above the posterior neural tube and incorporated into the development of a normal dorsal fin by stage 40. Fin is outlined by dashed white line. E,F: Ventral tissue explanted at stage 13 and cultured until controls reached stage 40. See blue area in Figure 2A. Arrow in E points to fin-like structures. n = 20 for A and B, n = 30 for C and E.

Download figure to PowerPoint

Having set up positive and negative controls for fin induction in explant culture, we then sort to identify the location of the endogenous ventral inducer. To do this, we cocultured ventral epidermis in combination with the underlying ventral mesoderm and endoderm from stage 13 neurulae (blue areas in Fig. 2A). In this case, ventral explants developed thin fringes of tissue (Fig. 3E, arrow), indicative of fins, although the fin development was not as pronounced as that seen after coculture with neural crest (Fig. 3E). In other cases of ventral tissue cocultures, two fin-like projections were seen, on either side of a central mass, the explants looking like butterflies (Fig. 3F). The ventral mesoderm and endoderm thus appeared to be able to induce a degree of fin development in isolated epidermis, something that lateral mesoderm and endoderm was unable to do. To verify the presence of a fin inducer in this ventral tissue, ablation experiments were then performed.

Ablation of Candidate Ventral Fin Inducers

A 600 μm by 600 μm square of tissue ventral to the blastopore was ablated from stage 13–15 neurulae (see blue areas in Fig. 2A). This region encompasses most of the epidermis fated to form the ventral fin (Tucker and Slack, 1995) along with the underlying mesoderm and endoderm. The resulting embryos when grown to stage 41 had a range of defects in the ventral fin (Fig. 4A–D). The fin at the tip of the tail and the dorsal fin were unaffected. The ventral fin defect was more extensive in those embryos where the tissue was removed at stage 15, compared with stage 13. In approximately 50% of cases, where the ventral tissue was ablated at stage 13, only a slight defect in ventral fin development was observed, affecting the length of the ventral fin rather than its presence (Fig. 4A,B). However, at stage 15, all the embryos showed a substantial defect in the ventral fin (Fig. 4C,D). This range in defects may represent a difference in timing of when the ventral fin is induced. If a similar sized region of epidermis, mesoderm, and endoderm is ablated from the lateral region of the neurula embryo at these stages, no defect in fin development is seen (Fig. 4E,F), the fin defect, thus, is specific to loss of ventral tissue.

thumbnail image

Figure 4. Location of the ventral fin inducer. A–P: Stage 41 embryos. A,C,E,G,I,K,M,O: Side views of fins. B,D,F,H,J,L,N,P: Frontal sections through developing tails. A,B: Operations carried out at stage 13. C–J,M–P: Operations carried out at stage 15. K,L: Operations carried out at stage 20. A,B: Square of tissue ventral to blastopore was ablated, resulting in reduction of the ventral fin. C,D: Similar ablation at stage 15, resulting in complete loss of the ventral fin. E,F: Similar-sized piece of lateral tissue ablated. No effect on fin development. G,H: Ventral epidermis only ablated, resulting in no fin defect. I,J: Ventral mesoderm and endoderm ablated, resulting in defects in the ventral fin. K,L: Ventral mesoderm and endoderm ablated at late neurula stage, showing only limited defect in the ventral fin. M,N: Ventral endoderm alone ablated, resulting in no ventral fin defect. O,P: Ventral mesoderm alone ablated, resulting in ventral fin defects. N = 30–40 for each extirpation.

Download figure to PowerPoint

Removal of the ventral tissue is a less-informative experiment than removal of the neural folds, as carried out in Figure 2C, as the presumptive fin epidermis is ablated along with the potential inducer. However, if only the epidermis is removed at the same stages, the wound heals and a normal ventral fin forms (Fig. 4G,H). The ablations were then repeated at stage 15 but instead of ablating the whole square of tissue under the blastopore, the epidermis was folded back to enable removal of the underlying tissue and was then replaced. When such embryos were cultured to stage 41, ventral fin development could still be seen to be disrupted (Fig. 4I,J). When the same operation was carried out at stage 20, late neurula, the resulting tail had a fairly well-formed ventral fin (Fig. 4K,L), suggesting that the signal for fin induction had already occurred. In these cases, the ventral fin formed but was often ruffled and folded over. The fin contained mesenchyme, suggesting that the induced ventral epidermis was able to recruit mesenchymal cells into its core, in a similar way to that observed after late dorsal epidermis was grafted to a lateral site (Bodenstein, 1952). The ventral fin inducing signal, therefore, would appear to emanate from the underlying ventral tissues. To identify which tissue acts as the inducer of the ventral fin, either the endoderm or the mesoderm was removed separately. This strategy showed clearly that removal of endoderm alone had no effect on fin development (Fig. 4M,N), while removal of the mesoderm alone suppressed ventral fin development (Fig. 4O,P).

Contribution to the Ventral Fin Core

The neural crest has been shown to migrate into the ventral fin providing a source of mesenchyme in the fin core (Collazo et al., 1993). However, not all of the fin core appears to be neural crest derived as, when the neural crest is ablated, the ventral fin core does not appear depleted of cells and remains as large as controls (Fig. 2C). To identify other possible sources for the ventral fin core, we decided to carry out a fate map of the ventral tissue. Orthotopic grafting of the 600 μm × 600 μm square lying ventral to the blastopore was carried out using FDA-labeled donor tissue at stage 13 (see blue areas in Fig. 2A). First, only the epidermis was labeled, then the endoderm, and finally the mesoderm and epidermis together. The mesoderm consists of a very narrow sheet on the ventral side, and as such, it was difficult to graft alone. When the endoderm was labeled, no fluorescent cells were found within the fin core, the tissue remaining in a compact patch near to the proctodaeum (Fig. 5B). When the epidermis was labeled, the majority of the ventral fin epidermis was labeled as expected (Fig. 5C). When both the mesoderm and epidermis were labeled, fluorescent cells could be seen both in the epidermis and in the underlying fin mesenchyme and the adjacent somites (Fig. 5D). It would appear, therefore, that the mesoderm does indeed contribute cells to the ventral fin core.

thumbnail image

Figure 5. Orthotopic grafting of fluorescein dextran amine–labeled ventral tissue at stage 13. Embryos grown to stage 41. A: Schematic of stage 41 tail. B: Ventral endoderm only labeled. Label is restricted to the ventral site near the proctodaeum. C: Ventral epidermis only labeled. Label found throughout ventral fin epithelium. D: Ventral mesoderm and epidermis labeled. Label is seen in the fin epithelium (as in C). In addition, label was observed in the ventral somites and in cells within the fin core, see arrows. Fin is outlined by dashed white line. n = 15 for each.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Previous work on other species has shown that the neural crest plays a double role in dorsal fin development, eliciting fin induction and then becoming incorporated as part of the fin structure (Raven, 1931; DuShane, 1935; Newth, 1956; Smith et al., 1994). We show that the neural crest is also vital for dorsal fin induction in Xenopus and that the ability to respond to the neural crest signal is not restricted to the prospective fin epidermis. The ventral fin, however, remains largely unaffected after ablation of the neural crest suggesting the existence of another inducer, and arguing against the results of Ford that the neural crest controls both dorsal and ventral fin induction (Ford, 1949). From extirpation experiments, it would appear that the ventral fin inducer resides within the tissue that lies just under the blastopore, agreeing with the preliminary results of Bytinski-Salz (1936) and Smithberg (1954). The induction signal appears not to be from the endoderm as ablation of this tissue has no effect on ventral fin formation; however, ablation of the mesoderm underlying the epidermis does lead to defects in ventral fin formation. Of interest, the ventral fin defect seen is dependent on stage. At stage 13, 50% of embryos show relatively weak defects in the ventral fin, while those ablations carried out at stage 15 result in much more severe defects. At even later neurula stages (stage 20), ablation of the mesoderm and endoderm tissue ventral to the blastopore has again only limited affect on ventral fin development. These results suggest that the ventral inducer acts during mid-neurula stages and is completed by late neurulation. A similar change in status of the presumptive fin epidermis is seen on the dorsal side of the embryo.

Although the ventral fin core contains some neural crest-derived mesenchyme, it is not entirely composed of cells derived from the neural crest, as is evident from neural crest ablation. In his fate map of Rana, Smithberg indicated that the underlying material a short distance ventral and lateral to the blastopore contributed to the ventral fin. In the zebrafish, ventral tail fin tissues were fate mapped to the ventral most regions of the epiblast at 6 hr of development (Woo and Fraser, 1998). From our fate map of the area under the blastopore at the early neurula stage, the mesoderm layer appears to provide cells that migrate into the ventral fin core. Thus, the mesoderm layer ventral to the blastopore both induces and physically contributes to the ventral fin, in a manner similar to the role of the neural crest in the dorsal fin. The endoderm, by comparison, did not migrate away from the graft site and remains incorporated in the ventral endodermal mass. A mesenchymal transformation and migration into the fin of postanal endoderm, as proposed by Elsdale and Davidson (1983), can thus be ruled out.

Not all the ventral fin is lost when the 600 μm × 600 μm square of tissue was excised in our experiments, as the tip remains unaffected. This same tip region is also unlabeled when the epidermis and mesoderm under the blastopore is fate mapped. The actual domain of the ventral inducer probably extends further laterally around the blastopore than the 600 μm square. The lateral region was indicated by Smithberg (1954) as having a role in ventral fin development. It seems likely that, as well as the posterior neural crest, Ford also extirpated lateral tissue, and this likelihood may explain why he saw ventral fin defects (Ford, 1949). In keeping with this idea, it is interesting to note that Ford's ventral fin defects occurred near to the fin tip.

The independence of the dorsal and ventral fins is supported by the phenotypes of a variety of zebrafish mutants. The tadpole's fin is homologous to the unpaired (medial) fin fold of larval fish. This fin fold disappears during development and is replaced by three distinct adult fins, the unpaired anal, tail, and dorsal fins, during the first 4 weeks of postembryonic development. Defects in this embryonic fin fold are observed in many zebrafish mutants, such as those that show bubbly fin folds, or undulations in the fin epithelium, but of relevance to this study, many mutants have defects that affect specifically the ventral fin fold (van Eeden et al., 1996). Dorsalising mutations such as mini fin and lost-a-fin show deletions of the ventral tail fin. mini fin encodes tolloid, a metalloprotease that can cleave chordin and, thus, increase BMP activity, while lost-a-fin encodes Alk8, which acts as a Bmp receptor. In both mutants, Bmp activity is reduced and the whole embryo becomes dorsalized at the end of gastrulation (Connors et al., 1999; Bauer et al., 2001). Ventral markers are reduced in these mutants, and it possible that the loss of ventral tissue and, therefore, the ventral inducer has resulted in loss of the ventral fin. We have shown previously that BMP signaling can drive tail outgrowth in Xenopus embryos (Beck et al., 2001). In these zebrafish mutants, the dorsal inducer, the neural crest, is still present but is unable to induce ventral fin development. In contrast, in mutants that are ventralized, such as mercedes and dino, the ventral fin is duplicated (van Eeden et al., 1996). In embryos that are severely ventralized and posteriorized, such as ichabod, the embryos fail to form a head and notochord but instead have an abnormally large ventral tail fin (Kelly et al., 2000). An expansion of the ventral domain thus leads to the formation of larger or multiple ventral fins. The phenotype of these mutants agrees with our conclusion that the ventral fin is both induced by and populated with cells from the ventral mesoderm and is initially independent of the neural crest.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Embryos

Xenopus laevis embryos were obtained by standard procedures and staged according to Nieuwkoop and Faber (1967). Embryos were incubated in NAM/10 at 24°C after dejellying with 2% cysteine, pH 7.9 in Petri dishes coated with 1% noble agar. Embryos were fixed in 10% formalin in phosphate buffered saline overnight at 4°C and stored in methanol.

Fate Mapping

Embryos were injected shortly after fertilization with 9.6 nl of FDA at 50 mg/ml in water. Injections were carried out in NAM/2plus 5% Ficoll. The label had been previously dialyzed to remove any low molecular mass impurities. Pieces of labeled tissue were carefully excised by using tungsten needles and then implanted into unlabeled host embryos where similar-sized pieces of tissue had been ablated. Grafts were held in place by using glass coverslip fragments until the tissue had adhered, the host embryo having been embedded in an agar hole to prevent movement and squashing. Embryos were left to develop until the fins had fully formed (stage 40/41). The graft was visualized by using a Zeiss fluorescence microscope.

Explant Culture

Operations were carried out in full-strength NAM salts, after which the explanted tissue was cultured in NAM/2 at 14–18°C. Explants were incubated until control embryos had reached stage 40/41. For cocultures, the prospective inducing tissue (neural crest, lateral or ventral mesoderm, or mesoderm and endoderm) was sandwiched between two pieces of epidermis and held together by a small piece of glass coverslip in a hole cut in a plate coated with agarose until healing around the tissue had occurred.

Extirpation Experiments

Operations were carried out in full-strength NAM salts, after which the embryos were cultured in NAM/2 at 14–18°C. Embryos were embedded in agarose holes to prevent movement during the dissections. For the experiments where the epidermis was replaced, tissue was cut on three sides and the epidermis (with or without mesoderm) folded back to reveal the underlying mesoderm/endoderm. After ablation, the epidermis was then folded back and held in place with a glass coverslip until the wound healed.

After photographing, tails were embedded in wax and sectioned to show the extent of the ventral fin defects. Slides were stained with hematoxylin/eosin.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

This work was supported by Cancer Research UK, the Wellcome Trust, and the Medical Research Council. For the experiments at King's College London, Xenopus embryos were kindly provided by Alison Snape and Georgina Fletcher.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES