SEARCH

SEARCH BY CITATION

Keywords:

  • blastema;
  • epigenesis;
  • limb;
  • morphology;
  • positional memory;
  • regeneration

Abstract

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References

The limb blastema cell, which is a major source of mesenchymal components in the limb regenerate, serves as a stem cell that possesses an undifferentiated state and multipotency. A remarkable property of the limb blastema cell can be seen in its capability for morphogenesis. Elucidation of the molecular basis for morphological regeneration is essential for success in organ regeneration in humans, and characterization of limb blastema cells will provide many insights into how to create three-dimensional morphology during the regeneration process. In this review, we deal with positional memory, a key trait of the limb blastema cell in regard to morphological regeneration, making reference to classic surgical experiments, comparative descriptions of limb and fin blastemas, and genetic/epigenetic regulation of gene transcription. Urodele amphibians, anuran amphibians, and teleosts are likely to share fundamental mechanisms for morphological regeneration, but there are several differences in the process of regeneration, including the epigenetic conditions. Accumulation of knowledge of the molecular mechanisms and epigenetic modifications of gene activation in morphological regeneration of the model organisms for which an overview is provided in this review will lead to successful stimulation of regenerative capacity in amniotes, which only have a limited capability for morphological regeneration.

Recent progress in stem cell biology has resulted in the creation of induced pluripotent stem (iPS) cells, and further progress will hopefully enable the production of all kinds of cells in the human body in the near future (Yamanaka 2007). The development of new technology, together with rapid progress in bioengineering, may make organ regeneration in mammals possible. There are at least three steps for successful organ regeneration: preparation of every kind of cells composing the organ, tissue organization, and establishment of three-dimensional morphology of the organ (morphological regeneration). With the recent remarkable progress in stem cell biology, it is expected that the first two steps will become possible in the near future, but the third step, morphological regeneration, might be impossible. The concept of morphological regeneration appears not to be included in current stem cell biology, mainly because there are few good model systems for organ regeneration, especially morphological regeneration, in mammals.

The development of technology for iPS cell preparation is based on results of extensive studies on molecular characterization of ES cells (Takahashi & Yamanaka 2006). We should target a good model of stem cells for morphological regeneration of an organ as has been done in iPS cell research, which targeted ES cells for preparation of totipotent cells. If such stem cells do not exist in mammals, we should search for them in other vertebrates, as we know that many species of vertebrates can regenerate various organs. In order to approach organ regeneration in humans in a logical rather than an alchemical manner, it is necessary to understand the regeneration of organ morphology in non-mammalians. For example, in caudal fin regeneration in zebrafish, an amputated fin can regenerate a simple formation of cell/tissue types, but stem cells for fin regeneration can restore the original morphology of the caudal fin, including the M-shape configuration seen on a lateral view, indicating that fin regeneration serves as a typical morphological regeneration (Fig. 1A, for reviews, see Akimenko & Smith 2007; Yin & Poss 2008 and references therein). We can see another example in jaw regeneration in the newt, in which an amputated jaw reconstructs a very complex arrangement of tissues, resulting in a complete replica of the original jaw morphology (Kurosaka et al. 2008). Limb regeneration in amphibians also serves as a good model of a system of morphological regeneration. If a forelimb and hindlimb of a urodele amphibian is amputated, a four-digit regenerate of the forelimb and a five-digit regenerate of the hindlimb develop. If a urodele forelimb is amputated at the wrist/ankle, elbow/knee, or upper arm/thigh levels, the regenerates which develop are an autopod, a zeugopod plus autopod, or a stylopod plus zeugopod plus autopod respectively (Fig. 1B, see Tamura et al. 2008 for terminology of limb structure).

image

Figure 1.  Morphological regeneration of appendages in vertebrates. (A) Zebrafish. (B) Axolotl. (C) Xenopus froglet. Pink lines with scissors indicate amputation levels. The pink area in each picture corresponds to the regenerated part: note that zebrafish (A) and axolotl (B) regenerate the same morphology as that of the original organ shown on the left, while the froglet (C) regenerates a spike-like structure regardless of amputation level.

Download figure to PowerPoint

Limb regeneration as a model of morphological regeneration

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References

The limb regeneration process in amphibians can be dissected into several successive but overlapping steps: (i) wound epidermis formation, (ii) blastema formation and (iii) repatterning and redifferentiation (for reviews, see Bryant et al. 2002; Gardiner et al. 2002; Suzuki et al. 2006; Yokoyama 2008; and references therein). In this process, the blastema cells (undifferentiated mesenchymal cells) are the main source of a regenerate. The blastema cell is believed to be a type of stem cell that has multipotency (Stocum 2004; Brockes & Kumar 2005), whereas the potency for plasticity of cell differentiation which blastema cells possess appears not to be exerted during the normal limb regeneration process in urodele amphibians (Kragl et al. 2009). Regardless of the origin and final differentiation of blastema cells, an interesting issue is that blastema cells memorize the original position in which they were placed before amputation. This is a fundamental but unresolved characteristic of blastema cells in regards to morphological regeneration. A blastema (a group of blastema cells with an epidermal jacket) can form a normal regenerate autonomously even when transplanted into the eye chamber or the dorsal fin (Pietsch & Webber 1965 ; Stocum 1968). When a blastema derived from the wrist level of an axolotl forelimb is heterografted into a host stylopod-level blastema which is regenerating from mid-thigh level, the donor autopod-level blastema does not start regenerating until the host regeneration reaches the ankle level (Fig. 2B, Crawford & Stocum 1988). In this process, the autopod-level blastema is ejected (sorted out) from the stylopod level and displaced to the ankle level. Figure 2B shows a schematic representation of this experiment. A simple explanation for this phenomenon, known as distal displacement, is as follows: Limb cells have their own values that differ from neighboring cells at different levels, as indicated by 10-to-1 numbers (codes) along the proximal–distal axis. Each number corresponds to a positional value which each cell memorizes. Blastema cells with a fixed positional value in their memories always regenerate only a more distal structure than the memory, at a location corresponding to the value. Therefore, for example, when blastema cells with the value “4” are implanted into the more proximal “8” region, the blastema cells are displaced to a position next to the value “5” and regenerate a structure corresponding to “4-3-2-1”. This phenomenon of distal displacement suggests several important characteristics of blastema cells in regards to morphological regeneration: (i) the blastema cells memorize the original positional value along the proximal–distal axis, (ii) the positional memory involves a cell surface property, giving rise to displacement, (iii) the memory is not provided/controlled by the stump tissue but installed in the blastema cells themselves, and (iv) the memory is not erased or modified, even when the cell is in a different positional environment.

image

Figure 2.  Schematic representations of surgical experiments on morphological regeneration. In all diagrams, stump structures are shown in grey and regenerate skeletons are shown in pink and blue. (A) Jagged amputation of the caudal fin of zebrafish (this illustration incorporates data and figures from Akimenko & Smith 2007). After this jagged cut, the dorsal-most edge (upper) and ventral-most edge (lower) of the amputation plane were at the same proximal–distal level, indicated by a broken line. The same-level blastema cells showed different growth rates (blue double-headed lines). (B) Heterotopic transplantation of blastema in the axolotl limb (this illustration incorporates data and figures from Crawford & Stocum 1988). Positional values are indicated by numbers (“10”-“1” from proximal to distal). In this experiment, a wrist level blastema with the value “4” (in pink), was grafted to the blastema-stump junction of a limb, where the value is “7” (in blue). The wrist blastema displaced to the level of the host limb regenerate that corresponded to its own level of origin (the value “4”). (C) Limb regeneration with reversed proximal–distal axis (this illustration incorporates data and figures from Butler 1955). A limb that had been amputated at the wrist level was sewn into a pocket on the caudal flank. This bridged limb, which was innervated and vascularized from both proximal and distal ends, now had the positional value “10-9-8-7-6-5”. Then the bridged limb was amputated between the values “7” and “8”. The normal proximal stump (upper side) regenerated a limb with a distal value “7-6-5-4-3-2-1”. The reversed distal stump (lower side) regenerated a limb with a distal value “6-5-4-3-2-1”.

Download figure to PowerPoint

Limb cells in urodeles memorize their position. What then would regenerate from the amputated distal part of a limb if the limb could be kept alive? Possible predictions would include no regeneration, proximal regeneration or distal regeneration; the third is the correct answer. The amputated distal piece that, for example, has “-5-6-7” as its positional value, regenerates “6-5-4-3-2-1”, resulting in “-5-6-7-6-5-4-3-2-1” codes from the proximal to distal direction (Fig. 2D, Butler 1955). This “rule of distal transformation” suggests that the positional value at the amputation plane only carries the memory to regenerate the distal part and that the memory at levels deeper than the amputation plane has nothing to do with distal transformation. Thus, each blastema cell inherits level-specific positional memory from its mother limb cell, and each cell hoards this memory until amputation occurs at that level. The molecular nature of positional memory remains unresolved thus far but, as we will discuss later, one possible key for this memory is epigenetic regulation of gene expression in the genome.

Morphological regeneration in the teleost fish fin

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References

Fins, appendages in fish (osteichthyes) which are homologous to limbs in tetrapods, are composed of two skeletal parts developed through distinct ontogenic processes. The outer layer of the skeleton (the lepidotrichia and actinotrichia composed of dermal bones), can be regenerated in many different teleost fishes (for a review, see Akimenko & Smith 2007). While teleost fish do not regenerate the inner layer of the skeleton (the endoskeleton composed of endochondral bones), the lungfish, a sacropterygian fish, can regenerate this structure (Conant 1970, 1973). The fact that the zebrafish regenerates the caudal fin with its original M-shape morphology (Fig. 1A) suggests that cells in the fin also memorize a positional value. While there have not been many studies focusing on morphogenesis in fin regeneration, fin regeneration in the zebrafish has fascinated many researchers, particularly in regard to genetic analysis aimed at elucidation of the molecular mechanisms involved in organ regeneration (for reviews see Akimenko et al. 2003; Poss et al. 2003; Akimenko & Smith 2007; Nakatani et al. 2007).

It is thought that blastema-like mesenchymal cells participate in fin regeneration. In caudal fin regeneration, the more proximally the fin is amputated, the faster the fin regenerates (Tassava & Goss 1966; Akimenko et al. 1995; Lee et al. 2005), suggesting that fin blastema cells organize the regeneration process differently along the proximal–distal axis. When the caudal fin is amputated in the shape of two teeth of a saw (Fig. 2A), the dorsal-most and ventral-most edges of the amputated plane are located at the same level of the proximal–distal axis. Interestingly, the growth rates on the two sides differ, the ventral lateral-most region regenerating faster than the dorsal lateral-most region, suggesting that growth of the regenerating blastema is influenced by the neighboring blastema and/or stump tissue (Akimenko et al. 2003; Akimenko & Smith 2007; in this review, the authors stated that a similar experiment was first performed by Morgan, who is well known for his prominent work on Drosophila genetics). The memory that the fin blastema cells may possess appears to be rewritable. Plasticity of memory has also been shown by experiments involving heterotopical grafting of ray fragments (Murciano et al. 2002). When, for example, a fragment of a lateral ray is transplanted into an intermediate region, during regeneration the implanted fragment develops intermediate-ray traits with bifurcations, according to its new position in the fin, suggesting that the rays have a morphogenetic plasticity which depends on the environment. Plasticity for the proximal–distal axis has also been shown (Murciano et al. 2007). Since such plasticity is not seen in amphibian blastemas, the mechanisms for morphological regeneration in fish fins do not completely correspond with those in tetrapod limbs. The memory for positional value may be stored differently in fin blastema than in limb blastema cells.

Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References

In contrast to the great capacity for morphological regeneration in urodele limbs, this capacity is limited in anuran amphibians. Anuran tadpoles can generally regenerate their developing limb buds, but regenerative capacity declines before/during metamorphosis (for reviews, see Stocum 1995). Some species have no capacity for limb regeneration in young adult froglets and frogs after metamorphosis, but in other species, including Xenopus laevis, this capacity is partially conserved in froglets and adults. In Xenopus froglets the capacity is incomplete and, regardless of the amputation level, they regenerate only a hypomorphic protrusion consisting of a single shaft of spike-like cartilage (Fig. 1C).

The decline in capacity for morphological regeneration in Xenopus tadpoles is fascinating in light of the contrast between the small capacity in mammals and the complete capacity in urodeles. Xenopus tadpoles generate a complete morphology of the limb after amputation at early stages of limb development. The appendages that a tadpole has at these stages are developing limb buds, and limb regeneration at the embryonic stage, strictly speaking, should be discriminated from that in froglets and frogs. In spite of the immaturity of tadpole limb tissues, regeneration of the tadpole limb bud is based on a great capability for morphological regeneration. A limb equipped with fewer digits is regenerated when the limb is amputated at intermediate stages of limb development and finally, when amputated at a later stage before/during metamorphosis, by which time tissues in the limbs have matured, the tadpole regenerates at best only a stunted protrusion, sometimes nothing is regenerated (Dent 1962; Muneoka et al. 1986). This decline in regenerative capacity is accompanied by defects in morphological regeneration, including abnormal expression of key genes (ex. Shh and Lmx1) for limb morphogenesis (Endo et al. 2000; Matsuda et al. 2001; Yakushiji et al. 2007). The mesenchyme composing the limb bud is responsible for the decline in regenerative capacity; a combination of regenerative mesenchyme and non-regenerative epidermis can regenerate the complete limb, but the reverse combination cannot (Yokoyama et al. 2000). Application of FGF10 at the amputated plane of later stage limbs results in multi-digit regenerates, suggesting that this protein stimulates morphological regeneration in the later-stage tadpole, although under normal conditions it has almost no capacity to regenerate (Yokoyama et al. 2001).

There has been debate as to whether hypomorphic limb regeneration of spike-like cartilage in froglets and frogs should be categorized as epimorphic regeneration (a type of regeneration accompanied by outgrowing blastema formation; Agata et al. 2007) or simply as hypertrophic tissue repair. Results of cellular and molecular studies (Endo et al. 2000; Suzuki et al. 2005, 2007; Satoh et al. 2006; Yakushiji et al. 2007), plus discovery of nerve-dependency of the spike formation found in urodele limb regeneration (Endo et al. 2000; Suzuki et al. 2005), has provided evidence that limb regeneration in Xenopus froglets is a type of epimorphic regeneration (for review, see Suzuki et al. 2006). The fact that there is no pattern (no segments, no bifurcation, and no digit-like structures) in the spike regenerates indicates that epimorphic regeneration in Xenopus froglet limbs lacks some aspects of morphological regeneration. Determination of the cause of the pattern-less regeneration is expected to reveal the essence of morphological regeneration.

In the past decade, studies of gene expression have revealed information about the froglet limb blastema. Genes related to an undifferentiated state and to cell proliferation, including Msx1, fgf8 and fgf10, are activated in the froglet limb blastema (Endo et al. 2000). Genes involved in epimorphic regeneration, such as Tbx5 (Simon et al. 1997), are also expressed in the froglet blastema (Suzuki et al. 2005). These findings strongly suggest that froglet limb regeneration has epimorphic aspects. In contrast, some defects in gene expression related to morphogenesis are observed. Lmx1, which is a key transcription factor for dorsal–ventral axis formation in the developing limb bud (Riddle et al. 1995; Vogel et al. 1995; Cygan et al. 1997; Chen et al. 1998), is absent from the froglet limb blastema (Matsuda et al. 2001), suggesting that dorsal–ventral axis formation is disrupted. Hoxa11 and hoxa13, which contribute to proximal–distal axis formation in developing limbs with dynamic change in each expression domain (Yokouchi et al. 1991, 1995), are also expressed in the froglet blastema, but hoxa11 and hoxa13 are uniformly expressed in the blastema throughout the following regeneration processes (Endo et al. 2000; and our unpublished observations). Shh and target genes (such as Ptc1 and Gli1) downstream of Shh signaling are not expressed in the froglet blastema (Endo et al. 2000; Yakushiji et al. 2007, 2009a). Transcriptional regulator genes of Shh expression, dHand and Gli3, are expressed in the froglet blastema, suggesting that a certain point of Shh transcription regarding anterior–posterior axis formation is defective. Shh plays a pivotal role in anterior–posterior axis formation for digit morphogenesis in developing limbs (Riddle et al. 1993). As with the developing limbs of amniotes, in both urodeles and larval anurans Shh is expressed in the posterior margin of the blastema (Endo et al. 1997; Imokawa & Yoshizato 1997; Torok et al. 1999). Furthermore, in froglet limb regeneration excess administration of a Shh agonist gives rise to branching of spike cartilage (Yakushiji et al. 2009b). Therefore, spike formation in the Xenopus froglet seems to include multiple deficiencies in patterning along all three axes for three-dimensional morphological regeneration.

Epigenetic gene regulation as positional memory

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References

Once morphological regeneration has been initiated after limb amputation, key molecules for the patterning that occurs during limb development are upregulated. The basic molecular mechanisms of limb regeneration recapitulate the mechanisms used in developing limbs, but some regeneration-specific modes of gene expression have been pointed out (for a review, see Gardiner & Bryant 2007). By using their “positional memory”, urodele amphibians correctly educe the mechanisms that they once used for limb development, but Xenopus may lack the memory or fail to educe the required mechanisms during the process of morphological regeneration. The central difference responsible for the variation in capacity for morphology regeneration between species, a difference which is closely related to positional memory, has not yet been identified. However, accumulating knowledge concerning chromatin remodeling and its modification for gene transcription, so-called epigenetic gene regulation, will provide clues for clarifying the nature of positional memory.

Control of gene expression is not only mediated through genetic regulation but is also modified by ‘epigenetic’ alterations, including histone modifications (usually deacetylation) and DNA methylation. Such alterations condense nucleosomes more tightly and generally prevent transcription. This additional mechanism for regulation of gene expression is now considered to be very important, because epigenetics includes multiple mechanisms by which DNA transcription is altered in various tissues and at different times without changing the underlying gene sequence. Epigenetic marking by modification of DNA and histones creates molecular landmarks that distinguish between active and inactive chromatin, and such epigenetic marks are heritable through mitosis. During development, two families of proteins have been shown to be involved in epigenetic changes: the Trithorax (TrX) and Polycomb (PcG) groups of proteins. TrX proteins function as transcriptional activators, while PcG genes are involved as silencers of gene expression. Response elements for these proteins can remember and maintain an active or inactive state of gene expression over many cell generations, long after the activators and repressors have disappeared, indicating that the TrX/PcG system can be utilized as epigenetic memory of cell identity (Ringrose & Paro 2007). The response elements (targets of the TrX/PcG system) include regulatory elements of genes for homeobox-containing transcription factors (hox genes). Both TrX and PcG were discovered in Drosophila as activators and repressors of homeotic gene expression.

Although little is known about the relationship between the TrX/PcG system and morphological regeneration in amphibian limbs, it is probable that epigenetic regulation through the TrX/PcG system of homeobox genes plays an important role in establishing and implementing positional memory. We here discuss a hypothetical model of epigenetic landmarks as positional memory (Figs 3, 4). For simplification, we have selected three homeobox genes, Meis, hoxa11 and hoxa13, all of which are known to be involved in proximal–distal axis formation during vertebrate limb development (for a review of proximal–distal axis formation in the developing limb, see Tabin & Wolpert 2007). Meis determines the proximal-most part, the stylopod (Capdevila et al. 1999; Mercader et al. 1999). Although Meis is initially expressed broadly in the early limb bud (in green in Fig. 3), Meis expression is restricted to the proximal-most region of the more mature limb bud. The distal region of the Meis domain begins to express hoxa11 (in yellow in Fig. 3), and Meis is restricted to the more proximal region where hoxa11 is not expressed. The Meis-positive region gives rise to the stylopod. Subsequently, the distal region of the expanding hoxa11 domain begins expressing hoxa13 (in orange in Fig. 3), and negative regulation of hoxa11 by hoxa13 represses hoxa11 expression in the distal-most region. This negative regulation causes separation of hoxa11 and hoxa13 domains along the proximal–distal axis, resulting in determination of the hoxa11-positive zeugopod and hoxa13-positive autopod. Finally, the limb establishes three distinct compartments, stylopod, zeugopod, and autopod. The cells in the zeugopod region have encountered expression of Meis and hoxa11 but not of hoxa13. In this region, Meis is initially turned on but then turned off. The final status of epigenetic landmarks in the zeugopod is Meis = OFF, hoxa11 = ON, and hoxa13 = never experienced (lower center in Fig. 3). In the same way, the stylopod is recognized as Meis = ON, hoxa11 = never experienced, and hoxa13 = never experienced (lower left in Fig. 3), and the autopod is recognized as Meis = OFF, hoxa11 = OFF, and hoxa13 = ON (lower right in Fig. 3). The final state of these homeobox genes could be marked by the epigenetic mechanism, and a cell could memorize the state.

image

Figure 3.  A hypothetical model of imprinting of the positional value during the limb development process. As the limb bud develops distally (from left to right), the expression domains of M and A11 are first overlapped and then separated along the proximal–distal axis. As the limb bud further elongates distally, the A11 domain first overlaps with the A13 domain but is later separated. The lower column shows the final situation with regard to gene activation in each region along the proximal–distal axis. “Never experienced” indicates that the cells in the region have never expressed the gene. A, autopod (upper panels); A11, hoxa11 (in yellow); A13, hoxa13 (in orange); M, Meis (in green); OFF, gene inactivated; ON, gene activated; S, stylopod; Z, zeugopod.

Download figure to PowerPoint

image

Figure 4.  A hypothetical model of emergence of the memorized positional value in limb regeneration. The final situation with regard to gene activation is surrounded by a blue line, and this memorized situation leads to regeneration of a part corresponding to the combination of genes expressed (blue arrow). More distal positional values (pink box and pink arrow) are newly provided by a de novo process. Lines with scissors indicate amputation levels. Ne, never experienced; OFF, inactivation; ON, activation.

Download figure to PowerPoint

When the limb is amputated, the genomic conditions should be initialized into an undifferentiated state, but the above positional memory is not erased, resulting in emergence of memory (as if a carved letter on a relief sculpture surfaced after the sand on the relief was blown away). If the limb is amputated in the middle of the zeugopod, for example, the cells which possess memory of the zeugopod (Meis = OFF, hoxa11 = ON, and hoxa13 = never experienced) contribute to making the blastema cells (in yellow in Fig. 4). These cells recognize that they should not express Meis. They turn Meis expression off, but retain hoxa11 expression as dictated by their memory for hoxa11. They have no memory of hoxa13, but the re-activated hoxa11 recaptures the halfway process of development (indicated in the upper middle part in Fig. 3) and newly induces hoxa13 and the resultant autopod. The combination of the epigenetic pathway of Meis repression/hoxa11 activation (memorized, blue lines and arrows in Fig. 4) and the genetic pathway for hoxa13 induction (renewed, red lines and arrows in Fig. 4) enables regeneration of the morphology distal to the site of amputation of the zeugopod. We can apply this model for both proximal and distal amputations (left and right in Fig. 4). If the limb is amputated inside the autopod (orange region to the right in Fig. 4), the blastema should activate hoxa13, but neither hoxa11 nor Meis would be re-expressed in the autopod blastema. Alternatively, the autopod blastema immediately turns hoxa11/Meis expression off. Our observations of the expression pattern of hoxa11 and hoxa13 in the regenerating limb bud of the Xenopus tadpole, which can regenerate a complete structure along the proximal–distal axis, support this idea (Fig. 5). In the early (Fig. 5A), middle (Fig. 5B), and later (Fig. 5C) stages of autopod blastemas, expression of hoxa11 is not found, suggesting either that this gene is not reactivated in the distal (autopod) blastema or that the gene is turned off at an early stage. In contrast, hoxa13 transcripts were detected in the autopod blastemas at all stages that we examined (Fig. 5D–F).

image

Figure 5. hoxa11 and hoxa13 expression in tadpole blastemas. Stage 53 (Nieuwkoop & Faber 1956) hindlimb buds were amputated inside the autopod region, and in situ hybridization was performed for hoxa11 and hoxa13 as described by Endo et al. (2000). (A–C) hoxa11 expression in (A) 2 dpa, (B) 3 dpa, and (C) 5 dpa blastemas. (D–F) hoxa13 expression in (D) 2 dpa, (E) 3 dpa, and (F) 5 dpa blastemas. Note that expression of hoxa11 cannot be seen in the blastema whereas its expression can be detected in the stump zeugopod region (arrowheads in A–C). The extent of expression of hoxa13 during tadpole limb regeneration is similar to that of developing limb buds. Lines indicate the estimated amputation planes. Distal is to the right in all figures. Scale bar = 200 μm. Dpa, days post amputation.

Download figure to PowerPoint

In this model, we have divided the positional values of the limb cells into only three regions along the proximal–distal axis, but each limb cell should have an individual property, such as cell adhesiveness, that gradually changes along the axis (Yajima et al. 2002). If the condition of gene expression that a cell experiences at the final determination of its position is fixed and carved into the genome as an epigenetic landmark, each cell will return to the same condition of gene expression when it is de-differentiated into blastema stem cells during limb regeneration. This model could also be applied to the formation of other axes along the anterior–posterior and dorsal–ventral directions, each of which provides the positional values for three-dimensional morphogenesis.

Epigenetic gene regulation and diversity of capacity for morphological regeneration

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References

The model we presented in the previous section is hypothetical, but epigenetic gene regulation is a likely candidate for explication of positional memory and of precise reconstruction of the original structure in morphological regeneration. Evidence for the importance of epigenetic gene regulation is accumulating for many systems, including stem cells, cancer, the immune system and development, but we have much less knowledge in regard to organ regeneration.

One interesting study has shown a correlation between DNA methylation status and fin regeneration in the zebrafish (Thummel et al. 2006). DNA methylation status serves as an epigenetic landmark for active chromatin (poorly methylated) or inactive chromatin (highly methylated). In that study, the authors analyzed the DNA methylation status of the promoter sequences in some transgene constructs. Although the ef1-alpha promoter drives ubiquitous EGFP expression in early zebrafish development, the signal for EGFP disappears in several adult tissues, including the caudal fin. In the adult fin, where expression of the EGFP transgene is inactivated, the promoter sequence is highly methylated, showing a good correlation between gene silencing and high degree of DNA methylation. The EGFP signal can be re-detected in the fin blastema after caudal fin amputation, and the fin blastema contains a non-methylated pattern, indicating that the blastema cells demethylate the DNA sequence of the transgene. It appears that zebrafish can control the epigenetic condition for gene silencing during morphological regeneration (uppermost in Fig. 6). They may have a key to open a locked (silenced) condition.

image

Figure 6.  A phylogenetic comparison of the capability for morphological regeneration (a model from epigenetic aspects). “Locked” and “Unlocked” indicate inactive and active conditions of gene transcription, respectively. See the text for details.

Download figure to PowerPoint

As we have mentioned, in limb regeneration in the Xenopus froglet, the blastema cells that eventually regenerate a non-patterned spike do not activate Shh expression, suggesting that there are defects in expression of the Shh gene. The defects are not sequence defects, such as deletions and mutations of a genome sequence, because Shh is expressed in developing and regenerating limb buds of the Xenopus tadpole and organizes the digit pattern. Rather, it seems that epigenetic silencing is involved in the failure of Shh activation in the froglet limb blastema cells. The degree of DNA methylation in, not the promoter, but the limb-specific enhancer of the Shh gene, MFCS1 enhancer (Maas & Fallon 2004; Sagai et al. 2004, 2005), is very high in the froglet limb and the limb blastema, whereas there is little methylation in the developing tadpole limb bud and regenerating tadpole blastema (Yakushiji et al. 2007). These results strongly suggest that epigenetic gene regulation, including DNA methylation, plays a role in morphological regeneration, and that failure of demethylation of an appropriate sequence for gene expression may be a cause for failure of morphological regeneration in metamorphosed anural amphibians. Indeed, Shh expression is strongly activated in froglet blastema cells treated in vitro with a combination of an inhibitor of DNA methyltransferase and a histone deacetylase inhibitor (Yakushiji et al. 2009a), suggesting that epigenetic regulation is definitely involved in gene expression in the froglet limb blastema. Epigenetic regulation does not occur at random but shows a kind of target-specificity because the Shh promoter is not highly methylated in the froglet limb. In addition, regulation seems to be independent of aging because the limb-specific Shh enhancer is highly methylated in tadpole tissues other than the limb bud even before metamorphosis. It is likely that, after metamorphosis, Xenopus adults cannot control the epigenetic condition and cannot be released from negative landmarks for gene silencing during morphological regeneration (lower middle in Fig. 6).

Contrary to our expectation, the limb blastemas in urodele amphibians do not demethylate the Shh limb-specific enhancer region during limb regeneration. The limb blastema of the axolotl and newt includes Shh-expressing cells (Imokawa & Yoshizato 1997; Torok et al. 1999), and the level of methylation of the Shh limb-specific enhancer region is sufficiently small (Yakushiji et al. 2007). However, before limb amputation, the cells in mature limb tissues maintain their low methylation status in the Shh limb-specific enhancer region, and therefore do not have to demethylate the DNA sequence in that region during limb regeneration (Yakushiji et al. 2007). Surprisingly, mature tissues in the eyes and heart also have a low level of DNA methylation in this region, although these tissues can have never used this limb-specific enhancer throughout their life.

Urodele amphibians presumably retain their epigenetic status in a non-silenced condition, and it is possible that they do not have to release the limb cells from an epigenetically silenced condition for gene transcription during morphological regeneration (upper middle in Fig. 6). Taken together, the findings suggest that epigenetic gene regulation has pivotal roles in organ regeneration, and that teleosts and urodele amphibians, both of which show great capacity for morphological regeneration, may possess their own distinct mechanisms for epigenetic gene regulation. From a scientific point of view the non-silenced condition in mature tissue of urodele amphibians is very interesting. However, maybe what we should focus on is the regeneration mechanism that teleosts have, because the human body must have a highly silenced condition of epigenetic gene regulation in its mature tissues, as do zebrafish and Xenopus froglets (lowermost in Fig. 6).

Conclusion

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References

The blastema cell is a less differentiated and multipotent mesenchymal stem cell that retains its original position as positional memory. Thus, blastema cells must have originated from cells that have once experienced positional identification. Accordingly, the origin of blastema cells cannot be naïve undifferentiated cells such as ES and iPS cells. It might be possible to create blastema cells in humans if we could cancel the differentiated state of limb mesenchymal cells without erasing their positional memory. Alternatively, it might be possible to create blastema cells from undifferentiated totipotent stem cells by introducing accurate positional values. In both cases, a key to blastema cell creation is epigenetic gene regulation which makes memory for the positional value possible. Although we are still far from a full understanding of positional memory, we are optimistic because of the day-by-day accumulation of information on genetic and epigenetic mechanisms and technical progress in teleosts and amphibians. The key to success in creation of blastema cells is objective assessment of various models for morphological regeneration and complete elucidation of the organ regeneration process. Epimorphic regeneration in teleosts and amphibians may involve different epigenetic controls on key gene expressions and, besides these two, spike formation in the Xenopus froglet may provide another example of epigenetic control of epimorphic regeneration. Studies using model animals of morphological regeneration are essential if we are to progress toward successful organ regeneration in humans.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References

We thank Dr Yonei-Tamura for her drawing/paintings of all illustrations in the figures. We are grateful to Dr Marie-Andree Akimenko for invaluable comments on morphological regeneration in the zebrafish fin. Our laboratory’s work on Xenopus limb regeneration presented here was supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan, KAKENHI (Grant-in-Aid for Scientific Research) on Priority Areas “Comparative Genomics”, a grant from Graduate School of Life Sciences, Tohoku University, and the Toray Science Foundation.

References

  1. Top of page
  2. Abstract
  3. Limb regeneration as a model of morphological regeneration
  4. Morphological regeneration in the teleost fish fin
  5. Xenopus limb regeneration: a clue for morphological regeneration in non-regenerative animals
  6. Epigenetic gene regulation as positional memory
  7. Epigenetic gene regulation and diversity of capacity for morphological regeneration
  8. Conclusion
  9. Acknowledgments
  10. References