Hydrogen peroxide (H2O2) controls axon pathfinding during zebrafish development 

Get Complete Project Material File(s) Now! »

Regeneration and asexual reproduction

But why did regeneration originate in evolution? And did it originate as an independent property? Or has it originate together with life, as one of its intrinsic property? If we think at the regeneration as a restoration of body components, it becomes the equivalent of physiological maintenance, which is required to all living organism to survive. That is why Morgan and Korschelt considered the regeneration process as a primordial attribute of life. This idea was reinforced by Needham, who defined regeneration as “identical to life itself” (Needham, 1952) since in the organisms who reproduce themselves by asexual reproduction, this process is indistinguishable from their mode of regeneration.
If we take as example the protozoan Bursaria, the only difference that we can find between asexual reproduction and regeneration is the stimulus triggering the two events: in the first case, the favourable environmental conditions and, in the second case, an injury. Therefore, asexual reproduction and regeneration in protozoans appear to be different manifestations of the same ability: reverse their morphogenetic processes (Lund, 1917).
Could, then, regeneration have its origin in asexual reproduction? If the only difference between the two processes is the stimulus needed to trigger the process, it is sufficient that the stimulus necessary to induce regeneration activate the same messenger molecules that mediate asexual reproduction. And hydra provides a supporting example: during asexual reproduction it has been observed a localised increase of the secreted peptide head activator (HA), necessary for the budding process (Schaller, 1973); similarly, the amputation of this organism induce the release of high levels of HA, necessary for the formation of the regeneration bud that will differentiate into the missing body part (Galliot, 1997). Then, in both case we have the activation of the same genetic cascade.
Clearly, the explanation cannot be so easy. In fact, there are animals able to regenerate the entire organism from body parts in which asexual reproduction and regenerative abilities are not associated. This is the case of the planarians Dendrocoelum lacteum and Bdelocephalla punctata. Contrary to other planarian species, these two organisms are unable to regenerate if the amputation occurs at the level of the pharynx or below; this is also accompanied by a lack of asexual reproduction (Brønsted, 1969). Why do they reproduce exclusively by sexual means even if they have the potential for asexual reproduction, evidenced by the ability of pharyngeal fragments to regenerate complete animals? Cases such as these need to be further study to identify the relationship between regeneration and asexual reproduction, and the factors necessary either to allow or prevent the reversion of the morphogenetic processes.

Regeneration and cellular pluripotentiality

Important differences exist between the regeneration processes of simple and histologically more complex organisms. These differences rely on the different ability of the animals to reverse their morphogenetic processes, which depends on the regulation of cellular pluripotentiality. If we then consider two very different examples of regeneration, as they can be the amputation of the salamander hand or of the planarian tail, it is however clear that in both cases the regenerating fragments have in common the property to be pluripotent; the main difference among them lies in how this potential is controlled in the adult organism (Figure 3). Then, if the ability of animals to reproduce missing body parts depends on the availability of a source of pluripotent cells, it is not a coincidence that in more complex animals, in which it is harder to induce cell differentiation and reveal the hidden pluripotentiality of the cells, we observe less regenerative abilities (Sanchez-Alvarado, 2000).
Figure 3: Increasing regulation of cellular pluripotentiality with increasing morphological complexity. The addition of several regulatory steps (wound healing, dedifferentiation, blastema formation and differentiation) as evolutionary complexity increases is here illustrated taking as example three representative organisms: Hydra, planarian and salamander. The regeneration process starts always with the wound healing and culminates in in the fully regenerated structure. Adapted from Sanchez-Alvarado, 2000.
In the simple diploblastic hydra, we observe the presence of undifferentiated and constantly dividing cells in the body column (David and Campbell, 1972; Campbell and David, 1974); moreover, the pluripotent cells of the gastric column (Gierer et al., 1972) continuously change their position in an apical or basal direction, reaching the whole organism body. The gastric column can then be functionally linked to a blastema because it serves as a sources of pluripotent cells (Galliot, 1997). Therefore, in this animal the cellular proliferation and the creation of the blastema are not needed thanks to a constant source of pluripotent cells in the body column and, in this way, the regeneration process can directly start with the post-blastemal stages observed in higher animals, as determination and differentiation.
Tripoblasts, instead, possess a more tight regulation of cell pluripotentiality, which is severely restricted even in animals with extensive regeneration power. In planarians, for example, we find only one population of undifferentiated and mitotically active cells, the neoblasts (Brønsted, 1969); after amputation, these cells have first to proliferate to form the blastema, which can after direct the determination and differentiation processes. In the more complex salamanders, the pluripotentiality is restricted even further: in these animals there are no reserves of pluripotent cells in the body; on the contrary, they are produced de-novo by pre-existing terminally differentiated cells after amputation. Therefore, in these organism there is the addition of a new step in the regeneration process: the de-differentiation of differentiated cells that then undergo mitosis and proliferate to form the blastema (Brockes, 1997).
In conclusion of this paragraph, it is evident that with the increase of animal complexity, more and more regulatory checkpoints have been added to control the inherent pluripotentiality of cells; this augmented control of pluripotent cells might be a reflection of selective forces that may have been acting on the evolutionary origins of the regeneration blastema itself (Sanchez-Alvarado, 2000).

Regeneration and development

For long time the researcher wondered if the process of limb regeneration could recapitulate the events observed during limb development. Effectively, as already mentioned, the two processes deploy similar sets of regulatory genes to carry out their morphogenetic purposes (Gardiner et al., 1995; Gardiner and Bryant, 1996; Imokawa and Yoshizato, 1997; Simon et al., 1997; Logan et al., 1998). However, in the last years new data suggest the existence of important differences between the two processes (Nacu and Tanaka, 2011).
First of all it is important to notice that the spatial and temporal scale in which the two processes take place is completely different (Figure 4): even if a molecule is necessary for both processes (and there are several examples: Hh, Wnt, Fgf, …), the levels of expression and the time required to the formation of a signal gradient have to be adapted to really different dimensions (Brockes and Kumar, 2005). On the other hand, the sequential expression of hox genes seems to be conserved, at least in the axolotl, between the developing and regenerating limb (Roensch et al., 2013).
In this context, also the codification of the positional information has to adapt to different spatial and temporal scales, even if the same patterning molecules are used. Some decades ago, it has been proposed that, contrary to the development process in which the cells acquire position identity in a proximal-to-distal sequence, during regeneration occurs an intercalation process: the blastema cells acquire distal identity, creating a gap with the more proximal identity stump cells that triggers regeneration of the intervening regions; then, when a salamander hand blastema is grafted onto an upper limb stump, intercalation occurs, whereby the upper arm stump regenerates the missing lower arm segment and the grafted hand blastema regenerates the hand (Stocum, 1975; Iten and Bryant, 1975; Pescitelli and Stocum, 1980). The intercalation is unidirectional, because the hand blastema cells do not contribute to restoration of the missing lower arm. Unidirectional intercalation reflects a cellular determination state, called “the rule of distal transformation,” in which connective tissue-derived blastema cells can only form limb segments more distal to their original identity (Pescitelli and Stocum, 1980). However, in the last years, the already mentioned studies on hox genes expression during axolotl regeneration, coupled with transplantation assay, showed that blastema cells acquire positional identity in a proximal-to-distal sequence, suggesting a progressive specification rather than an intercalation process during limb regeneration of this animal (Roensch et al., 2013). Then, the intervention of one or both the intercalation and progressive specification mechanisms during regeneration has still to be clarified.
In any case, it is important to understand how the cells can determine their position along the proximal-distal axis and use this information to regenerate the missing body part. The positional information has to possess two characteristics: first, it has to be present in the adult before the amputation and, second, its perturbation has to be transmitted to regulate the regeneration process. Nowadays, only few factors controlling the positional information during limb regeneration have been identified. The best example is Prod1 (Proximo-distal 1), the newt ortholog of CD59; it is a glycoprotein located at the cell membrane with a GPI (glycosyl-phosphatidylinositol) anchor and its expression forms a gradient in adult limb, conferring the positional information to the cells (Kumar et al., 2007a; da Silva et al., 2002). The role of Prod1 during development, if it exists, is not yet clear (Nacu and Tanaka, 2011). Looking for potential ligands of Prod1, the team of J. Brockes discovered the factor nAG (newt Anterior Gradient), which is responsible for the nerve dependence of limb regeneration and for the interdependence between the wound epithelium and the nervous fibres during regeneration (Kumar et al., 2007b). After amputation, injured axons firstly retract in the stump, before to regrow, surrounded by Schwann cells. At the same time nAG starts to be expressed: firstly, in the Schwann cells of the stump, where we find the retracted axons, to confer the positional information corresponding to the level of amputation and, secondly, in the wound epithelium, to give the positional information corresponding to the most distal extremity of the structure that has to be regenerated. Brockes team also found that nAG expression in both locations is abrogated by denervation and that its the ectopic expression is sufficient to rescue a denervated blastema and regenerate the distal structures, linking together for the first time the positional identity of the blastema and the classical nerve dependence of limb regeneration necessity (Kumar et al., 2007b). nAG is very conserved during evolution; on the contrary Prod1 is identified for the moment only in the newt (Kumar et al., 2007a).
Another important difference between development and regeneration processes is the proliferation of stem cells, which is sustained during development, whereas in many cases stem cells are not present in mature tissues. As already said, the ability of animals to regenerate body parts depends on the availability of pluripotent cells or on the ability to reprogram terminally differentiated cells (Brockes and Kumar, 2002; Odelberg, 2005; Poss, 2007; Birnbaum and Sanchez Alvarado, 2008); these cells, which have different origins but at this point are all committed in the regeneration process, proliferate to form the blastema with a behaviour that has been studied in several species, as amphibians, zebrafish and planarian (Wagner et al., 2011). In this context, the comparison between adult regeneration and regeneration of the embryo – or its development – is difficult, and maybe also devoid of significance, because in organisms that are still in development the process of pluripotent cells mobilisation is not necessary (Galliot, 2011). Finally, a last difference between development and regeneration is the nerve dependance: if in embryos limbs are able to develop even in absence of nerves, in adult the innervation is essential to launch the regenerative program. I will come back later in this manuscript on the relationship between nerves and regeneration.

Regeneration and homeostasis

The tissue homeostasis is the result of three different processes – cell death, proliferation and differentiation – and the same processes are found during regeneration.
In mammals, the cell renewal has been studied in different tissues and the time to obtain a complete renewal is very variable, from few day for the intestinal epithelium to some months for the epithelium of the lungs (Blanpain et al., 2007); in organs as the brain and the heart the renewal is very limited and very slow and then also very difficult to study. Also the regeneration process requires the proliferation of progenitor cells but with some important differences compared to homeostasis. First, the regeneration needs to start a “massive renewal process” to allow the formation of the blastema, later responsible for the orchestration of the morphogenesis. Second, pluripotent mitotically-active cells are not always present in the tissue and then during regeneration could be necessary an additional step of dedifferentiation; different studies of cell lineage in transgenic animals have demonstrated that all differentiated cells can turn into progenitors cells, maintaining the memory of their origin and finally re-differentiating according to it (Kragl et al., 2009; Knopf et al., 2011; Tu and Johnson, 2011). Third, participating cells must be precisely guided to needed areas, and once regeneration is complete specific cues are required to report regenerative success and signal termination, otherwise the initial response would continue indefinitely (Oviedo and Beane, 2009).
The studies to better understand the link between homeostasis and regeneration are very important for regenerative medicine. For example, recent results in planarian show that the signal involved in stem cells renewal are also involved in stem cells activation after amputation (Forsthoefel et al., 2012; Newmark, 2005).

Regeneration and scarring

Before to conclude this first part of the introduction on the regeneration process and go on with the manuscript, I think that is also important to clarify the difference between the terms regeneration and wound healing/scarring, being both commonly used to indicate the process that leads to recover the tissue integrity after an injury caused by the environment or diseases. The process of wound healing is included in the process of regeneration and consists in very rapid mechanisms that allow to cover the wound with a layer of epithelial cells, leading to the protection of the injured tissue as soon as possible. In mammals, this process is more and more inefficient with aging: the embryo and the foetus have healing capacities definitely more important than the adult (Redd et al., 2004). On the contrary, the cell layer that is formed during the healing phase of the regeneration process is very characteristic and has similarities with the apical epidermal layer that is formed in embryos during limb development (Christensen and Tassava, 2000; Nye et al., 2003); moreover, it produces signals necessary for the formation and the maintenance of the blastema (Thornton, 1957; Nye et al., 2003). Contrary to normal wound healing/scarring, the regeneration process allows the complete reconstitution of a functional tissue without the formation of the scar.
RNAi experiments on planarians and other experiments on other organisms allowed the identification of different genes involved in the regeneration process, as for example some genes important for the signalisation occurring just after healing (Reddien et al., 2005). These findings allow the analysis of the loss of these genes or problems in their functionality in organisms that are not able to regenerate. In fact, it has been proposed that some genes necessary for regeneration to take place would be present only in the specie effectively able to regenerate (Brockes and Kumar, 2008). An example is the already mentioned Prod1, responsible for the positional identity and probably for the nerve dependence of the regeneration process, that is found only in the pleurodeles (Kumar et al., 2007b).

A general scheme for the regeneration process

The regeneration of missing body parts in response to tissue injury, degeneration or amputation requires the succession of three modules (Figure 5):
1. the immediate response to the injury/amputation with the activation of repair genes;
2. the regeneration-induction module, with the formation of the blastema;
3. the regeneration re-development module, consisting in morphogenesis, differentiation and growth processes.
The second module is composed by a variety of cellular processes that trigger the structure-specific developmental process; these cellular processes arise as a response to injury, in some contexts via the wound epidermis, and can be combined or not, providing multiple ways to bridge injury to regeneration (Galliot and Ghila, 2010). As consequence, this module is very variable: if we only think at how progenitor cells can be recruited, we have already a relatively long list of possible options (stem cell proliferation, dedifferentiation or the not mentioned above apoptosis-induced compensatory proliferation, direct proliferation of differentiated cells and transdifferentiation). This variability is reflected by the facts that different organisms could follow different ways to regenerate homologous structures or that the same organism could use different mechanisms to regenerate different body structures or even the same structure. An example of this large variability is given, for example, by the evolutionary divergent mechanisms of muscle regeneration observed during limb regeneration in two closely related salamander species, Notophthalmus viridescens (newt) and Ambystoma mexicanum (axolotl). In newt, the new skeletal muscle is formed by proliferation of mononuclear cells originated from myofibers fragmentation and dedifferentiation; on the contrary, myofibers in axolotl do not dedifferentiate to generate proliferating cells and do not contribute to newly regenerated muscle, which is formed by contribution of resident satellite cells (Sandoval-Guzmán et al., 2014).
The variability of the second module is due to both the developmental status of the organism and the homeostatic conditions of the tissue before the injury (Galliot and Chera, 2010). Indeed, the length of the module is directly influenced by the developmental and aging processes ongoing in the organism (Carlson and Conboy, 2007; Auger et al., 2010): for the embryos there is no need to have a bridge between wound healing and re-development, whereas in the adult the bridge is required and it becomes longer with the closing of the developmental processes and the appearance of aging. The homeostatic status at the time and place of injury, on its side, influences the cellular tools that can be activated upon injury (Galliot and Chera, 2010). The recruitment of these tools leads to the activation of the third module, which is less plastic because the re-development of the missing structure often relies on the same signaling pathways used during organogenesis.
Figure 5: Scheme of the tri-modular composition of the regeneration process. To regenerate a given structure, a variety of cellular processes, which form the “induction module”, induce a structure-specific developmental process; these processes can be combined or not, providing multiple routes to bridge injury to regeneration. The developmental stage of the organism and the homeostatic context at the time of injury largely influence this second module. Adapted from Galliot and Chera, 2010 and Galliot and Ghila, 2010.

Zebrafish and its caudal fin regeneration ability

Danio rerio, or zebrafish, is a teleost fish first described by Francis Hamilton, a surgeon with the British East India Company, stationed principally in West Bengal at the beginning of the 19th Century. He published An Account of the Fishes Found in the River Ganges and its Branches in 1822 that included ten Danio species. It is characterized by the presence of a distinctive colour pattern based on alternating dark and light horizontal stripes.
Zebrafish is one of the most important vertebrate model organisms in genetics, developmental biology, neurophysiology and biomedicine, thanks to some of its characteristics that make it particularly suitable to experimental manipulation: it is small, then easy and cheap to maintain in the laboratory; it breeds all year round and females can spawn every 2-3 day hundreds of eggs; generation time is short, around 3-4 months, making it suitable for selection experiments; the eggs are large relative to other fish and transparent; fertilisation is external and live embryos are accessible to manipulation and can be monitored through all developmental stages; finally, the development is rapid, with major organs precursors developing within 36 hours post fertilisation and larvae displaying food seeking and active behaviour within 5 days post fecundation (Kimmel et al., 1995; Spence et al., 2008).
The zebrafish is increasingly important in biomedical research (Dooley and Zon, 2000; Shin and Fishman, 2002), particularly as a model of human disease (Berghmans et al., 2005; Guyon et al., 2007) and for screening of therapeutic drugs (Rubinstein, 2003, 2006); its strengthen as a model organism is that as vertebrate it is more comparable to humans than invertebrate model species as Drosophila melanogaster, while being more tractable to genetic and embryological manipulation than mammalian model species as mice, in which such procedures are both more complicated and costly (Spence et al., 2008).
Zebrafish is particularly interesting because, as other teleost fish, is able to regenerate many organs and structures at the adult age: the retina, the spinal cord, the liver, the pancreas, the heart and all its fins. Even if it is a relatively “new” model organism, it is more and more present in the research laboratories and its spread utilisation in developmental biology has allowed the development of many genetic tools and the in vivo imaging.
During my thesis, I worked on the process of epimorphic regeneration of the caudal fin in the adult zebrafish. It is a recent model to study the regeneration of body appendages, compared to the regeneration of urodelans limb regeneration, and they share many similitudes, in particular regarding the blastema formation. On the contrary, the morphogenesis of the new-formed zebrafish caudal fin is faster than the re-formation of the urodelans limb (Figure 6).
Figure 6: Different time scales in the regeneration process of different organisms. The regeneration of the zebrafish caudal fin is completed in 20 days, whereas salamander limb takes more than 70 days to completely regenerate. Adapted from Pfefferli and Jaźwińska, 2015 (zebrafish) and Brockes, 1997 (salamander).

The structure of the adult caudal fin

Zebrafish possesses five sets of fins divided into two types: the paired fins (pectoral and pelvic) and the median fins (dorsal, anal and caudal), structurally comparable (Figure 7). These fins are traversed by dermal bone rays, which are segmented along the whole fin length: the lepidotrichia. These lepidotrichia arise at the base of the fin and extend to the distal extremity, where they are terminated by elastoidin non-mineralised fibrils, the actinotrichia. Each segment is constituted by two concave hemi-segments that form in this way a tubular structure traversed by arteries, nerves and connective tissue; each hemi-segment is embedded in a monolayer of bone secretory cells, the osteoblasts. Between the different rays there are the connective tissue, the veins and the sensory neurons. All the structure is keep together by ligaments, which confer cohesion and flexibility at the fin (Becerra et al., 1983). In the fin there are no muscles, but at the base of the fin we find some tendons that connect the most proximal segment of each ray with the adductor and abductor muscles of the fish body (Akimenko et al., 2003).
The adult caudal fin is composed by 18 rays; all of them, except the most dorsal and the most ventral, have a bifurcation at the half of their length (Montes et al., 1982; Becerra et al., 1983). After amputation, at the standard temperature of 28°C, the regeneration of the caudal fin is completely finished in less than three weeks.
Figure 7: Zebrafish caudal fin skeleton structure. (A) Alizarin-red/alcian-blue staining of lepidotrichia and cartilage of the adult zebrafish caudal fin. The fin is constituted of segmented bony lepidotrichia (l), which form dichotomously branching points along the proximodistal axis of the fin. Scale bar 1mm. (B) SEM of lepidotrichia showing that each lepidotrichium (l) is made of two parallel segmented hemi-rays. Two adjacent segments (se) are separated by a narrow space and connected o each other by collagenous ligaments making a joint (j). Scale bar 100µm. Adapted from Borday et al., 2001. (C) Scheme of the fin and magnifications to better show the actinotrichia at the tip of each ray (top, right) and the segmented lepidotrichia formed by two hemi-rays and connected by a joint (bottom). Adapted from Rolland-Lagan et al., 2012.

The regeneration process of the adult caudal fin

The tail of teleost fish has been very used in the last decades to study the regeneration process and the histological mechanisms intervening in it have been carefully described.
When the majority or even the whole fin are amputated, the regeneration process allows the re-formation of a well-shaped and functional fin (Shao et al., 2009). Histological studies have suggested that zebrafish caudal fin regeneration is similar to the ephimorphic regeneration of urodelans limbs; as the latter, the process can be subdivided into four steps:
• wound closure;
• wound epithelium formation;
• blastema formation;
• cell differentiation and morphogenesis, leading to the missing tissue restoration.
After the injury, the connective tissue is rapidly sealed by a clot, formed by blood cells, while the entire wound is covered by lateral epithelial cells rearrangement and migration; this process occurs without cellular proliferation to be as rapid as possible, in order to avoid excessive bleeding and the risk of infection. This first step is concluded about 6 hours post amputation and the new-formed epithelium, the wound epithelium, is first a thin layer but it becomes faster a thick multilayer tissue, distinguishable by the later epithelium (Santos-Ruiz et al., 2002). The wound epithelium is also characterised by the expression of some genes, as fgf24 (Poss et al., 2000a; Kawakami et al., 2004) and dlx5a (Miyama et al., 1999).
After the wound closure and the formation of the wound epithelium, the proliferation starts first in the lateral epithelial cells; successively, also the mesenchymal cells of the stump de-organise and these quiescent cells re-enter the cell cycle. Nowadays there are not specific markers for these cells, which could derive from differentiated cells, as osteoblasts and endothelial cells, or fibroblasts, whose differentiation state is still debated. These cycling cells are then visualised with proliferation markers as phophorylated Histone 3 (pH3) and Proliferating Cell Nuclear Antigen (PCNA). These progenitors cells successively start their migration toward the wound, where they accumulate as a mass of mesenchymal proliferating cells covered by the wound epithelium: the blastema.
The proliferation of the blastema cells is maintained until a sufficient number of cells are produced to re-form the missing fin. It has been suggested that in urodelans the blastema would be formed by dedifferentiated cells coming from the surrounding tissues (Lo et al., 1993; Kumar et al., 2000; Velloso et al., 2000; Brockes and Kumar, 2002; McHedlishvili et al., 2007); moreover, studies conducted with axolotl suggested that the dedifferentiated blastema cells are able to re-differentiate in a different type of cell compared to their origin (Echeverri and Tanaka, 2002). On the contrary, more recent studies propose that these cells lose temporarily their markers of differentiation, but then re-differentiate in the same cell type from where they originated (Kragl et al., 2009). Similar observations have been done during Xenopus tadpole tail regeneration, in which cells derived from muscles, spinal cord and notochord re-differentiate only in their respective cellular types in the regenerated tissue (Ryffel et al., 2003; Gargioli and Slack, 2004; Slack et al., 2004), demonstrating that the regeneration process can occur without a complete dedifferentiation to the pluripotent state.
In zebrafish, cell lineage studies have demonstrated that each cell of the stump is able to dedifferentiate to take part at the blastema formation; these cells maintain the memory of their origin and then re-differentiate according to the cell type they were at the beginning of the process (Knopf et al., 2011; Tu and Johnson, 2011; Singh et al., 2012b; Sousa et al., 2011, 2012; Stewart and Stankunas, 2012), suggesting also a conservation in the process of ephimorphic regeneration between fish and amphibians.
Finally, there are still some opens questions. For example, even if it has been demonstrated that differentiated cells can participate in the blastema formation, their proportion compared to fibroblasts or stem cells is not known; moreover, cell lineage experiments with cellular differentiation marker have shown that cell differentiation occurs in the first two segments adjacent the amputation plane (Knopf et al., 2011), whereas other histological experiments and of cell lineage with colorants proposed that cell recruitment happens in the first six segments (Poleo et al., 2001; Santos-Ruiz et al., 2002; Gauron et al., 2013; Rampon et al., 2014).
The first part of the fin regeneration process, from fin amputation to blastema formation, is schematized in Figure 8.
Figure 8: Schematic representation of the events occurring from fin amputation to blastema formation. After amputation, the connective tissue is sealed by a clot and the lateral epidermal cells migrate to form the wound epithelium. Successively, proliferation starts first in the lateral epidermis and then also in the mesenchyme. These mesenchymal cycling cells continue to proliferate and then start to migrate toward the tip of the fin, where they form the blastema. Blastema cells proliferation and differentiation finally lead to the restoration of the amputated fin. Adapted from Santos-Ruiz et al., 2002.

The signaling pathways involved in the tail regeneration process

In the last years the signaling pathways involved in the blastema growth have been the objects of several studies; on the contrary, the early signals that could been involved in the activation of the regeneration module following the amputation/injury are not well known, even if there are several candidate signals.
The studies of these early signals have being conducted in different tissues and in different organisms, vertebrates and invertebrates. Since several years, it is known that amphibian limb and teleost fish fins regeneration processes start with the formation of the wound epithelium; before its maturation in a thick multilayer tissue, it has to secrete several factors that would be essential for the organization of the forming blastema. It has been shown that the rapid and continuous elimination of the wound epithelium as soon as it forms impairs the progression of the regeneration process, in particularly affecting the blastema formation (Stocum and Cameron, 2011).
As already mentioned, the process of blastema formation shows similarities with the development of the embryonic bud. A number of factors involved in the development process have been identified also in the wound epithelium, both in amphibians and zebrafish (Suzuki et al., 2006; Yoshinari et al., 2009); these factors are expressed very early after injury and their role in blastema growth is well documented, but there are also some evidences suggesting their implication in the wound healing and the blastema formation processes. In the fish, for example, fgf20a and wnt10a are expressed within the first six hours after amputation (Whitehead et al., 2005; Stoick-Cooper et al., 2007a); loss-of-function studies for both FGF and Wnt/β-catenin pathways showed that cells migrate to cover the wound but the formed wound epithelium is abnormal and incorrectly specified (Whitehead et al., 2005; Kawakami et al., 2006; Stoick-Cooper et al., 2007a); they are also involved in the process of blastema formation because their inhibition after wound healing leads to the non expression of msx gene and absence of blastema (Poss et al., 2000a; Stoick-Cooper et al., 2007a). The Wnt/β-catenin pathway seems to act upstream the FGF signaling because its inhibition induces the loss of fgf20a expression (Stoick-Cooper et al., 2007a); however, it is also possible that FGF signaling also induces Wnt/β-catenin pathway in a positive feed-back loop (Stoick-Cooper et al., 2007b). However, there are also other factors that inhibit blastema formation when overexpressed, as the chemokine stromal cell-derived factor (SDF-1) and Wnt5b (Dufourcq and Vriz, 2006; Stoick-Cooper et al., 2007a).
Other potential signals for the induction of the regeneration process are the bioelectric signals. In planarians, the polarity of regeneration can be manipulated thanks to an electric field, because it has been demonstrated that the head is always formed at the cathode side (Carlson, 2007); moreover, different studies have shown that the diminution of β-catenin expression or the dysfunction of a Wnt agonist interfere on head regeneration (Gurley and Sanchez Alvarado, 2008; Petersen and Reddien, 2008), leading for the first time to find a relationship between an electric field and the signal transduction. Other studies conducted on the regeneration of the Xenopus tadpole tail have demonstrated the implication of the V-ATPase H+ pump in the regeneration process but not in the wound healing (Adams et al., 2007; Levin, 2007): the pump inhibition by concanamycin inhibits regeneration, whereas the expression of a pump insensible to the drug can rescue the inhibition. More recently, similar studies on zebrafish adult fin have shown that this H+ pump is necessary for the expression of the gene aldh1a2, responsible for retinoic acid synthesis, and MAP Kinase Phosphatase 3 (MKP3): its inhibition blocks the blastema formation in the adult but, interestingly, has no effect on the larvae tail regeneration, suggesting the existence of different mechanisms intervening in the larvae or adult regeneration processes (Monteiro et al., 2014). Altogether, these studies suggest that the V-ATPase H+ pump has a central role in the regulation of the membrane potential of the stump cells and then it would be interesting to understand how voltage gradients are detected and transduces during the regeneration process. In this context, the implication of nerves seems to be determinant (Stewart and Stankunas, 2012) and, indeed, the innervation is essential for the regeneration process (Brockes and Kumar, 2008). I will speak more precisely about the role of nerves during the regeneration process in the next chapter.
Finally, another signal that could be involved in the induction of the blastema formation could be the activation of the thrombin, necessary to seal the wound just after the injury. It has been shown in the urodelans that the local activation of the thrombin is a signal for the regeneration process and that thrombin is necessary for muscle cell dedifferentiation in vitro (Tanaka et al., 1999; Lööf et al., 2007; Stocum and Cameron, 2011). Moreover, the thrombin-dependent activation of platelets plays a role in serotonin secretion after hepatectomy, acquiring an essential role for hepatocytes proliferation process (Lesurtel et al., 2006).
After the formation of the blastema, this structure starts to grow to allow the regeneration process to proceed and the involved signaling pathways have been carefully studied in different species, as reviewed for example in Stoick-Cooper et al., 2007b. At this point, the cell cycle speeds up (Nechiporuk and Keating, 2002) and the fin begins to rapidly growth back. Blastema cells differentiate into sceroblasts that start to secrete the matrix that will form the new lepidotrichia (Geraudie and Singer, 1992; Santamaria et al., 1992; Becerra et al., 1996; Mari-Beffa et al., 1996). Both FGF and Wnt/β-catenin pathways are still active and now required for expression of shh, which becomes an important player during this phase (Laforest et al., 1998; Poss et al., 2000a; Stoick-Cooper et al., 2007a), in the cells of the basal layer of the epidermis, adjacent to the newly forming bone structures (Laforest et al., 1998). The ectopic expression of shh or bmp2 in the blastema induces excess bone deposition and mispatterning of the regenerate (Quint et al., 2002); interestingly, both the inhibitions of BMP and Shh pathways through chordin overexpression or cyclopamine treatment, respectively, block the fin outgrowth, suggesting that both BMP and Shh are involved in sceroblasts proliferation and/or differentiation (Quint et al., 2002). Also the retinoic acid receptors are expressed during fin regeneration and exogenous retinoic acid application causes bifurcating bones to fuse together, suggesting that retinoic acid is involved in the process of bone patterning (White et al., 1994).
The regenerating fins display positional memory: they grow back to the correct length, independently where on the proximodistal axis they were amputated, with fins amputated more proximally growing back faster than fins amputated more distally (Akimenko et al., 1995; Lee et al., 2005b). It has been shown that the FGF signaling determines regenerative growth rates, being then involved in the positional memory of the fin (Lee et al., 2005b). Because it is known that FGF signaling regulates msxb expression (Poss et al., 2000a) and that the levels of expression depend on where along the proximodistal axis the fin is amputated, with the rapidly proliferating cells of the proximal blastema expressing higher levels than the less rapidly proliferating cells of the distal blastema (Akimenko et al., 1995; Lee et al., 2005b), the positional memory could be conferred precisely through Msxb.
Finally, at the end of the regeneration process, the fin has to stop to grow. To achieve this, there are two possible mechanisms: the gradual extinction of positive regulators or the activation of negative regulators. The latter seems to be the case of Wnt5b, which activates the β-catenin independent Wnt pathway (Stoick-Cooper et al., 2007a): the overexpression of Wnt5b inhibits β-catenin target genes expression and cell proliferation and leads to the failure of the regeneration process; moreover, the fish that are homozygous mutants for wnt5b regenerate faster than the wild type. Wnt5b is a negative regulator of the β-catenin signaling, and it has been demonstrated that β-catenin induce wnt5b expression to regulate itself, as the inhibition of β-catenin results also in the suppression of wnt5b expression (Stoick-Cooper et al., 2007a). Altogether these results lead to the conclusion that Wnt5b acts in a negative feedback loop to regulate Wnt/β-catenin signaling during the regrowth of proper sized fin. However, it is also interesting to notice that wnt5b is expressed throughout the regeneration and then it may also play other roles that are still to be uncovered. Another factor that has been recently involved in the termination process of fin outgrowth is calcineurin. Indeed, calcineurin activity gradually increases as the fin final size and shape are recovered, indicating that its activity could slow or stop the regenerative process. Moreover, calcineurin inhibition results in continued disproportionate fin outgrowth beyond its original dimensions because this inhibition shifts fin regeneration from a distal growth program to a proximal program, probably with the promotion of retinoic acid signaling. Calcineurin inhibition has a similar effect in juvenile developing fins and induces fins outgrowth in adult (Kujawski et al., 2014). Therefore, also calcineurin can be considered as a negative factor that regulates the termination of the regeneration process in order to obtain a proper sized regenerated fin (Tornini and Poss, 2014).

Table of contents :

General context and aims of the thesis project
1. INTRODUCTION
1.1. The regeneration process
1.1.1. The history of regeneration research
1.1.2. Regeneration among metazoans
1.1.3. Regeneration and asexual reproduction
1.1.4. Regeneration and cellular pluripotentiality
1.1.5. Regeneration and development
1.1.6. Regeneration and homeostasis
1.1.7. Regeneration and scarring
1.1.8. A general scheme for the regeneration process
1.2. Zebrafish and its caudal fin regeneration ability
1.2.1. The structure of the adult caudal fin
1.2.2. The regeneration process of the adult caudal fin
1.2.3. The signaling pathways involved in the tail regeneration process
1.3. The role of nerves during regeneration
1.3.1. The neurotrophic hypothesis and the factor X
1.4. Reactive Oxygen Species
1.4.1. ROS signaling: the mechanisms
1.4.2. ROS signaling targets: molecules, signaling pathways and cellular processes
1.4.3. ROS and regeneration
1.4.4. ROS as possible mediator between nerves and regeneration process
1.4.5. How to visualise ROS (and in particular H2O2): HyPer
1.5. Hedgehog (Hh) proteins and signaling pathway
1.5.1. Expression, processing and secretion of the Hh active peptide
1.5.2. Hh signaling pathway
1.5.3. Hh signaling during zebrafish fin regeneration
1.6. Palmitoylation
1.6.1. S-palmitoylation: features and targets
1.6.2. S-palmitoylation: regulatory effects
1.6.3. S-palmitoylation: the mechanisms
1.6.4. N-palmitoylation of Shh: mechanism and functions
1.6.5. Protein palmitoylation: regulation of axonal growth and synaptic plasticity
2. RESULTS
2.1. Nerves control redox levels in mature tissues through Schwann cells and Hedgehog signaling
2.2. Hydrogen peroxide (H2O2) controls axon pathfinding during zebrafish development
2.3. H2O2 gradients control protein S-acylation during morphogenesis
3. DISCUSSION
3.1. Similarities and differences among different reactive oxygen species
3.2. Gradient formation: a different process between development and regeneration
3.3. The relationship between nerves and H2O2 (and Sonic Hedgehog)
3.3.1. nAG is a thioredoxin
3.4. Perspectives
4. BIBLIOGRAPHY

GET THE COMPLETE PROJECT