SEARCH

SEARCH BY CITATION

Keywords:

  • axolotl;
  • regeneration;
  • tail;
  • Xenopus

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Overview of Xenopus tail regeneration
  5. Conclusions
  6. Acknowledgements
  7. References

The tail of the Xenopus tadpole contains major axial structures, including a spinal cord, notochord and myotomes, and regenerates within 2 weeks following amputation. The tail regeneration in Xenopus can provide insights into the molecular basis of the regeneration mechanism. The regenerated tail has some differences from the normal tail, including an immature spinal cord and incomplete segmentation of the muscle masses. Lineage analyses have suggested that the tail tissues are reconstructed with lineage-restricted stem cells derived from their own tissues in clear contrast to urodele regeneration, in which multipotent blastema cells derived from differentiated cells play a major role. Comprehensive gene expression analyses resulted in the identification of a panel of genes involved in sequential steps of the regeneration. Manipulation of genes’ activities suggested that the tail regeneration is regulated through several major signaling pathways.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Overview of Xenopus tail regeneration
  5. Conclusions
  6. Acknowledgements
  7. References

Amphibians have long been used in the study of tail regeneration (Goss 1969; Ferretti 2001). Urodeles regenerate their tails throughout their lives, but anurans do so only in a limited period, the tadpole stage, as the anuran tail is a transient appendage and is not present throughout an anuran's life. Much research has been performed using newts and salamanders, but the tail regeneration of Xenopus larva is now considered to be a useful model system for analyzing the molecular mechanism underlying the appendage regeneration. The tail of the Xenopus tadpole is transparent and suitable for whole mount observation at the cellular level. It regenerates within 2 weeks following amputation. Molecular studies using Xenopus have been facilitated by recent advances in genetic resources and techniques, including expressed sequence tags (ESTs), microarray and transgenesis. Analyses of cell lineage and molecular pathways using transgenic tadpoles have been reviewed in detail (Slack et al. 2004). This review briefly summarizes the results of the recent research on Xenopus tail regeneration in comparison to tail generation in axolotls (Ambystoma mexicanum), another amphibian model used in recent molecular analyses of tail regeneration.

Overview of Xenopus tail regeneration

  1. Top of page
  2. Abstract
  3. Introduction
  4. Overview of Xenopus tail regeneration
  5. Conclusions
  6. Acknowledgements
  7. References

A notochord is a major axial structure located in the center of the larval tail (Fig. 1A). It consists of vacuolated cells surrounded by a collagenous sheath. A spinal cord and a pair of sensory ganglions are located at the dorsal side of the notochord. Bilateral muscle masses cover the central notochord and the spinal cord. Dorsal and ventral fins are located at the midline of the tail. Major veins and arteries are in the mesenchymal space. The Xenopus larva regenerates most of its tail after amputation during its larval life until metamorphosis, with an exception in the refractory period between stages 45 and 47, at which the ability to regenerate the tail is lost in some populations of tadpoles (Beck et al. 2003). Reproducible tail regenerations were obtained within 2 weeks using tadpoles at stages 48–50 (Sugiura et al. 2004).

image

Figure 1. Morphology of normal and regenerating tail of Xenopus tadpole. (A) Transverse section of a stage 50 tadpole tail was stained with hematoxylin and eosin. (B, C) Sagittal sections of the regenerating tail at days 2 and 3 after amputation. The pair of arrowheads indicates the amputation plane. Arrows in (B) indicate cell masses of the notochord precursor. The dorsal side is up and the anterior side is to the left in (B) and (C). N, notochord; SC, spinal cord; M, muscle mass. Scale bar, 50 µm.

Download figure to PowerPoint

The surface of the amputated stump is covered with an epidermis within 24 h after half of the tail is removed. The most distal epidermis thickens to form a multicell layer (Sugiura et al. 2004), which may be equivalent to the apical ectodermal cap reported in urodele limb regeneration. Notochord precursor cells accumulate to create a compact cell mass adjacent to the edge of the amputated notochord sheath (Fig. 1B). The cells proliferate and align along the proximal-to-distal direction to make an immature notochord (Fig. 1C), which continues to elongate in the posterior direction. The regenerating notochord increases its diameter in the proximal region where the cells finally differentiate into vacuolated cells. The axolotl larva also contains a notochord in its tail but differs greatly from Xenopus after tail amputation, as the regenerated axolotl tail contains cartilage rather than notochord (Echeverri et al. 2001). The amputated spinal cord creates a terminal vesicle or bulb at its proximal end and then elongates posteriorly along the growing notochord. The regenerating spinal cord shows a simple tube without an apparent dorso-ventral pattern and is sometimes known as an ependymal tube in contrast to the complete regeneration of the spinal cord in urodeles. Myogenic cells accumulate in a mesenchymal region of the regenerating tail and differentiate into myofibers, which are not aligned with the regular segments observed in the normal tail. The regenerated Xenopus tail therefore is not the same as the normal tail, at least in terms of the patterns of the spinal cord and the muscle mass.

A typical feature of urodele regeneration is the formation of the blastema, which contains undifferentiated uniform cells with the potential to differentiate into multiple types of cells. The regenerating Xenopus tail shows no obvious structural equivalent to the urodele blastema, and there is no evidence to show the presence of multipotent cells in the regenerating Xenopus tail, as described below. This is the most critical difference between Xenopus and axolotl tail regeneration.

Cell lineage is not changed during Xenopus tail regeneration

Cell lineage analyses during Xenopus tail regeneration showed a clear contrast to the results obtained for the axolotl. The blastema cells in urodeles are derived from differentiated cells and redifferentiate into a variety of cells, including muscle, cartilage and dermis (Lo et al. 1993; Kumar et al. 2000). Multinucleate myofibers in the axolotl tail become mononucleate cells through fragmentation, proliferate rapidly and contribute a significant population of the blastema cells during the regeneration (Echeverri et al. 2001). In contrast to the case of the axolotl, Ryffel et al. (2003) showed that non-muscle cells were never derived from differentiated muscle cells in the regenerating Xenopus tail. The researchers labeled differentiated muscle cells and their descendants by making a transgenic animal in which Cre recombinase driven by a muscle-specific promoter induced a recombination of loxP sites and the inheritable expression of a downstream green fluorescent protein (GFP) gene. If labeled cells had been found in differentiated tissues other than muscles after amputation, it would have meant that differentiated muscle cells could trans-differentiate into other types of cells. However, no labeled cells were found in such differentiated tissues. Gargioli and Slack (2004) labeled myotome by grafting a piece of presomitic mesoderm from a transgenic embryo expressing GFP into a non-transgenic embryo, and showed that myofibers in the larval tail degenerated after amputation and did not contribute to any tissues in the regenerated tail. The same result was obtained after our experiment in which myofibers were labeled by electroporation-mediated gene transfer (Fig. 2).

image

Figure 2. Lineage analyses during the tail regeneration. Arrows in (A) indicate green fluorescent protein (GFP)-labeled notochord cells before amputation. (B) The same tadpole shown in (A) at day 4 after amputation. All of the GFP-labeled cells were found in the regenerating notochord. Arrows in (C) indicate three muscle fibers labeled with GFP. (D) The same tadpole as shown in (C) at day 5 after amputation. Two labeled cells were not changed while the other cell degenerated and was not found. (E, F) Spinal cord cells were electroporated after amputation and observed at day 2 (C) and day 7 (D). All the labeled cells were found in the regenerating spinal cord. The pair of arrowheads indicates the amputation plane. The dorsal side is up and the anterior side is to the left. Scale bar, 250 µm. The solution containing the expression plasmid pCAX-AFP (1 µg/µL) that coded for a mutant form of GFP (Inouye et al. 1997) was injected into the muscle, notochord or spinal cord region of a stage 48 tadpole anesthetized in MS222. The injected tadpole was covered with 0.1 × MMR-saturated papers and electroporated with an electroporator (ECM 830, BTX) by applying 10 pulses (30 V, 5 ms) according to the method of Momose et al. (1999). After confirmation of the GFP expression using a fluorescence microscope the tail was amputated at the plane indicated by a white line.

Download figure to PowerPoint

A mononucleate muscle satellite cell expresses PAX7 and is a myogenic stem cell in mammals. After stimulation by injury or exercise the satellite cells proliferate and differentiate into myofibers (Collins 2006). Xenopus satellite cells also express PAX7, and PAX7-expressing cells increase in number after tail amputation (Chen et al. 2006; Fig. 3). Gargioli and Slack (2004) suggested from the grafting experiment that the satellite cell is the source of regenerated myofibers after tail amputation. Forced expression of a dominant-negative form of Pax7 resulted in incomplete regenerations with few or no muscle cells, suggesting that the Pax7 plays an essential role in the muscle regeneration of Xenopus larva (Chen et al. 2006). These results showed that Xenopus larvae regenerate tail muscles through the proliferation and differentiation of the lineage-restricted stem cells, the muscle satellite cells, but not through the dedifferentiation of myofibers.

image

Figure 3. PAX7 expression during tail regeneration. A frozen cross-section of the tail muscle region was stained with anti-laminin antibody (A), Hoechst 33342 (B) and anti-PAX7 antibody (C). (D) Merged image of (A) and (C). The subpopulations of muscle nuclei were labeled with ani-PAX7. (E) Higher magnification view of a section immunostained for laminin (green) and PAX7 (magenta). A PAX7-labeled cell is situated underneath the myofiber basement membrane visualized with anti-laminin antibody, suggesting that it is the muscle satellite cell. (F, G) Horizontal sections of amputated tail at days 1 (F) and 3 (G) were stained with anti-PAX7 antibody. The PAX7-positive cells increase in number during the regeneration. The dorsal side is up in (A, B, C, D, E) and the anterior side is to the left in (F, G). A white line in (F, G) indicates the amputation plane. Scale bar, 50 µm in (D, F, G) and 25 µm in (E). A stage 48 tadpole was fixed in 4% paraformaldehyde, infiltrated in 18% sucrose, embedded in Tissue-Tek OCT compound (Sakura) and frozen in n-hexane cooled with liquid nitrogen. A cryosection was re-fixed in 4% paraformaldehyde and treated with a mixture of antibodies, containing a monoclonal anti-PAX7 antibody (DSHB, Iowa; Kawakami et al. 1997) and a rabbit anti-laminin antibody (Sigma). The signals were detected with Cy3- and Alexa488-conjugated secondary antibodies (molecular probes).

Download figure to PowerPoint

A lineage switch from neural cells to other cell types was reported in axolotl tail regeneration (Echeverri & Tanaka 2002). When radial glia cells in the spinal cord were labeled with GFP by electroporation and traced after tail amputation, GFP fluorescence was found in a variety of cell types, including neurons, glias, melanocytes and muscle and cartilage cells. It is plausible that cells in the Xenopus spinal cord neither change their lineage nor exit from the spinal cord during tail regeneration, because GFP-labeled cells were only detected in the regenerated spinal cord when spinal cord cells were labeled by grafting or electroporation (Slack et al. 2004; Fig. 2).

Tracing the notochord cells labeled with GFP by grafting or electroporation showed that notochord cells in the regenerated tail are derived from the pre-existing notochord (Slack et al. 2004; Fig. 2). The above results show that muscle, spinal cord and notochord are reconstructed in the regenerating tail with cells derived from their own type of tissues, suggesting that Xenopus tail regeneration is performed through the proliferation and differentiation of lineage-restricted stem cells rather than through trans-differentiation from differentiated cells.

Gene expression analyses in Xenopus tail regeneration

Gene expression analyses were performed in several laboratories to elucidate the molecular basis of Xenopus tail regeneration (Beck et al. 2003; Christen et al. 2003; Ishino et al. 2003; Sugiura et al. 2004; Tazaki et al. 2005). Most genes analyzed are expressed both in the developing tail bud of the embryo and in the regenerating larval tail, suggesting the presence of common genetic programs regulating both types of tail formations. Conversely, it is obvious that the mechanism underlying tail regeneration is not identical to that of tail development. Inflammation response and wound healing are specific events in tail regeneration. The regenerated tail is morphologically different from the intact tail, as described above. Neural and mesodermal cells are continuously produced from undifferentiated cells in the embryonic tailbud, as shown by morphological observation and the expression of organizer-specific bone morphogenetic protein (BMP)-antagonists, which induce neural cells (Gont et al. 1993; Beck & Slack 1998), while neither neural nor mesodermal cells are produced from undifferentiated cells in the tail regeneration process (Sugiura et al. 2004). The gene expression profile in the regenerating tail is similar to that in the normal tail tip region of the larva rather than that in the embryonic tailbud, suggesting that Xenopus tail regeneration occurs through the reconstruction and growth of the larval tail tip rather than through reconstruction of the embryonic tailbud. The above idea is also supported by the following observation. The notochord, spinal cord and muscle tissues continue to grow posteriorly in the most distal region of the normal tail with morphological similarity to the tissues in the regenerating tail.

Comprehensive analysis of gene expression has been performed using a cDNA array technique to identify genes whose expression was up- or downregulated during tail regeneration (Tazaki et al. 2005). The identified genes were categorized into three groups, namely; early responding, late responding and downregulated gene groups, according to their expression timings during the regeneration. The early responding group includes a large number of genes related to the inflammation response and wound healing, while the late-responding group includes genes related to cell proliferation and differentiation. The downregulated genes are reactivated at a much later stage of the regeneration and represent the final differentiation of cells, especially muscle differentiation. The temporal expression profile is therefore well correlated to morphologically identified events during the tail regeneration. Additionally, the array analysis resulted in the identification of 20 unclassified or novel genes whose expression was upregulated during the tail regeneration. It is necessary to reveal the functions of the genes in both embryogenesis and regeneration in order to obtain a complete understanding of the molecular mechanism underlying the tail regeneration.

Molecular signals required for the tail regeneration

Nerve-dependent regeneration is well documented in the limbs of both urodeles and anurans, and in the tails of urodeles (Goldfarb 1909; Dinsmore & Mescher 1998). Studies of the nerve-derived soluble factors have resulted in the identification of neurotrophic factors or growth factors. Ablation of the spinal cord from the urodele tail inhibits the regeneration completely (Holtzer 1956). Conversely, the tail regeneration of the anuran larva is believed to be independent of the spinal cord. It was reported that the destruction of the Xenopus spinal cord caused little or no effect on the tail regeneration (Rogusuki 1954), while the exact role of the spinal cord in Xenopus tail regeneration remained to be elucidated by ablating the spinal cord completely. We recently developed a simple operation method to remove the spinal cord completely from a part of the tadpole tail and showed that the spinal cord is required for the proper growth and differentiation of notochord during tail regeneration (Taniguchi et al. in preparation). This method should be useful for identifying and characterizing the soluble factors derived from the Xenopus spinal cord.

Fibroblast growth factors (FGFs) were expressed in the urodele spinal cord and were suggested to be required for the tail regeneration (Zhang et al. 2000). But SU5402, an inhibitor of FGF signaling, does not have significant effects on either axolotl tail regeneration (Schnapp et al. 2005) or Xenopus tail regeneration (data not shown), suggesting that FGFs are not essential for tail regeneration in either species. Sonic hedgehog (shh) is expressed in the ventral spinal cord of the axolotl. Treatment with cyclopamin, an inhibitor of hedgehog signaling, suppressed the axolotl tail regeneration completely, while treatment with a hedgehog agonist could not rescue the defect caused by the spinal cord-ablation (Schnapp et al. 2005). Shh therefore seems to be a necessary but not sufficient factor secreted from the spinal cord in the axolotl. The expression pattern of Xenopus shh (Xshh) in regenerating tail is quite different from that in axolotl. Xshh is not expressed in the regenerating spinal cord but is, rather, expressed abundantly in the regenerating notochord (Sugiura et al. 2004). Xshh is therefore not the spinal cord-derived factor discussed above, while its role in Xenopus tail regeneration should be determined.

Beck et al. (2003) reported that Xenopus tail regeneration requires signals of BMP and Notch. These researchers made transgenic tadpoles harboring hsp 70 promoter-regulated constructs in order to activate BMP or Notch signaling. When the expression of the transgene was induced by heat shock at the regeneration-deficient refractory stage, tail regeneration was restored. Conversely, the forced expression of inhibitors of BMP or Notch signaling suppressed the tail regeneration at the regenerative stage. A temporal analysis using a stable transgenic line harboring hsp promoter-driven noggin showed that BMP signaling is not required for the early phase of the tail regeneration, including wound healing and the early morphological changes of the notochord and spinal cord, but is required for the growth of the regenerating bud through the activation of a target gene, Msx1.

Expression of several Wnts and a wnt inhibitor is increased during Xenopus tail regeneration (Sugiura et al. 2004; Tazaki et al. 2005; Sugiura et al., unpubl. data). Wnts activate the canonical beta-catenin and/or the non-canonical signals according to a combination of ligands and receptors. The role of the canonical Wnt signal in cell proliferation and cell differentiation may be essential in tail regeneration, as its role has been well documented in other experimental systems. Xwnt-5a is expressed in the distal part of the regenerating tail and probably activates non-canonical signals. We recently found that forced expression of Xwnt-5a affects the proximo-distal pattern in the regenerating tissue (Sugiura et al., unpubl. data).

Gene manipulation in tail regeneration research

Genetic manipulation is a powerful and direct approach to analyzing the functions of genes in the regeneration process. Transgenesis using a sperm nucleus is now a popular method of manipulating the Xenopus genome (Kroll & Amaya 1996), while an inducible promoter is sometimes required in regeneration research because ectopic expression of a transgene in embryonic stages often results in severe developmental defects. The promoter of hsp 70 is very useful, as the expression of a downstream gene can be induced by a simple heat shock treatment (Beck et al. 2003; Beck et al. 2006; Chen et al. 2006), while repeated heat shock treatment, which may seriously damage amputated tadpoles, is sometimes required to obtain the desired effect. The use of other types of inducible promoters and/or cell type-specific promoters may overcome this problem.

Gene transfer by electroporation is another way of expressing exogenous genes at larval stages and is available to manipulate cells in the spinal cord, notochord and epidermis as mentioned above (see Fig. 2). The condition of electroporation, however, should be modified to increase the efficiency of the gene transfer, because only a small population of cells is labeled using a standard method. The injection of morpholino antisense-oligo into blastomeres is a very popular and effective method to knock down endogenous genes during early development, but could not be applied at the larval stage by a simple injection method. Delivering morpholino oligo by electroporation was reported to reduce the expression of Msx1 and Pax7 during tail regeneration in the axolotl (Schnapp & Tanaka 2005). The same strategy may be effective in Xenopus. Improving the manipulation methods and genetic resources, including promoters and transgenic strains, will facilitate regeneration studies using Xenopus.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Overview of Xenopus tail regeneration
  5. Conclusions
  6. Acknowledgements
  7. References

The most significant difference between Xenopus and axolotl tail regeneration is in the plasticity of the cells that contribute to regenerates. Mononucleate cells derived from differentiated myofibers contribute the blastema cells, and cells derived from the spinal cord differentiate into non-neural cells in the axolotl. In Xenopus, however, the major components of the tail are reconstructed with cells derived from the same types of tissues, probably through the proliferation and differentiation of lineage-restricted stem cells. Repair or regeneration through the activation of stem cells is well known in mammals. Proliferation and differentiation of the muscle satellite cells expressing PAX7 are observed in both Xenopus tail regeneration and in mammalian muscle regeneration. The Xenopus regeneration is still essentially different from mammalian tissue repair, even if they share a common cellular mechanism, as Xenopus larvae regenerate an entire appendage but mammals do not. Comparing the molecular aspects of regeneration in anurans, urodeles and mammals will elucidate the common and species-specific molecular mechanisms in these animals, thus facilitating our understanding of vertebrate regeneration.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Overview of Xenopus tail regeneration
  5. Conclusions
  6. Acknowledgements
  7. References

This work was supported in part by a Grant-in-Aid for Scientific Research from JSPS to M. M. We thank Drs N. Ueno, K. Agata, C. Kobayashi, Y. Momose, A. Tazaki, T. Sugiura, K. Watanabe, H. Orii and other members of our laboratory for their technical advice and helpful discussions.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Overview of Xenopus tail regeneration
  5. Conclusions
  6. Acknowledgements
  7. References