AJ adherens junctions AS amnioserosa ASCs adult stem cells DC dorsal closure DME dorsal-most epidermal ECM extracellular matrix GBR germ band retraction GSCs germinal stem cells HI head involution ISCs intestinal stem cells ICD intracellular domain JNK Jun-N-terminal kinases LE leading edge miRNAs microRNAs MMP matrix metalloproteinase PcG Polycomb group PCP planar cell polarity SRF serum response factor SSCs somatic stem cells 20E 20-hydroxyecdysone
Understanding the molecular basis of wound healing and regeneration in vertebrates is one of the main challenges in biology and medicine. This understanding will lead to medical advances allowing accelerated tissue repair after wounding, rebuilding new tissues/organs and restoring homeostasis. Drosophila has emerged as a valuable model for studying these processes because the genetic networks and cytoskeletal machinery involved in epithelial movements occurring during embryonic dorsal closure, larval imaginal disc fusion/regeneration, and epithelial repair are similar to those acting during wound healing and regeneration in vertebrates. Recent studies have also focused on the use of Drosophila adult stem cells to maintain tissue homeostasis. Here, we review how Drosophila has contributed to our understanding of these processes, primarily through live-imaging and genetic tools that are impractical in mammals. Furthermore, we highlight future research areas where this insect may provide novel insights and potential therapeutic strategies for wound healing and regeneration. Developmental Dynamics 240:2739–2404, 2011. © 2011 Wiley Periodicals, Inc.
Organisms maintain their structural and functional integrity in response to external injuries and physiological dysfunction through a variety of mechanisms, including wound healing/tissue repair, tissue regeneration and cell turnover (Fig. 1). The first two processes are similar in that they are both triggered by exogenous damage and can affect multiple cell types, requiring cell replacement over a long time scale time (months–years); however, there is a clear distinction between them. The purpose of wound healing is to restore tissue continuity without precise replacement of lost/damaged tissue to prevent infections and protect the organism from the external environment. This is achieved by formation of a scar, a connective tissue matrix without other apparent functional properties. On the contrary, tissue regeneration refers to the replacement of lost/damaged tissue so that both morphology and functionality are completely restored. This can occur in combination with wound healing (Nguyen et al.,2009). The third process is cell turnover, also called homeostatic regeneration, a class of regeneration that is initiated endogenously and occurs as a rapid process (days–months). It is restricted to specific tissues and organs that exhibit rapid turnover, such as the mammalian epidermis, intestinal mucosa, lung epithelium, blood cells, bone marrow, thymus, testis, uterine lining (endometrium), and mammary glands and is mediated by adult stem cells (ASCs; Wagers and Weissman,2004; Ohlstein and Sprading, 2006; Blanpain et al.,2007, for review).
Improving wound healing is particularly important for victims of burns or blasts and people with slow healing skin lesions, such as patients suffering from diabetes, which usually present foot ulcers that can lead to amputation and a significant impact on patient quality of life (Greenhalgh,2003, for review). There are also cases of excessive healing (as occurs in hypertrophic and keloidal scars), which can lead to abnormal scars that grow beyond the boundary of the original site of a skin injury. Furthermore, wound closure is the most important step after surgery because it contributes to the success or failure of surgical interventions. Indeed, many studies on scar formation and their minimization after myocardial infarctions or cosmetic/plastic surgery are being performed. Another motivation to study wound healing is the fact that DNA microarray analyses have shown that the gene expression pattern of healing skin resembles that of malignant tumors (Iyer et al.,1999; Chang et al.,2004); nonhealing wounds could be a risk factor for malignant transformation (Schäfer and Werner,2008, for review). Furthermore, the regeneration of organs such as liver, heart, or pancreas is being extensively studied in humans and other animal models (Lien et al.,2006; Poss,2007; Dor and Stanger,2007; Zaret and Grompe,2008). It has been suggested that the loss of regenerative homeostasis can cause cancer, syndromes, and diseases (Freeman,2008, Casali and Batlle,2009). Thus, regenerative medicine is not only focused on creating functional tissues/organs to replace, with embryonic or ASCs, those lost or damaged, but also on understanding stem cell biology for diagnosis, prevention, and treatment of human diseases. It will be essential to identify those signals/instructions that are required to specify cell types from undifferentiated stem cells, which will help to restore the function of damaged tissues (Blanpain et al.,2007, for review).
All animals have some capacity to heal wounds, but regeneration varies according to the species, developmental stage, and tissue involved. Several invertebrate organisms, such as echinoderms, planarians, hydras, and ascidians, have good regeneration potential and are able to regenerate a complete organism from small body fragments or to regenerate amputated limbs or tails (Saló,2006; Galliot et al.,2006; San Miguel-Ruiz and García-Arrarás,2007; Tseng and Levin,2008; Yokoyama,2008). Among vertebrates, only amphibians and teleosts can regenerate limbs during adult stages (Poss et al.,2003; Stocum,2006; Kragl et al.,2009). Several observations have led to the conclusion that wound healing and tissue regeneration recapitulate several embryonic morphogenetic processes such as gastrulation, neural tube closure, palate formation, and eyelid fusion in vertebrates, but also dorsal closure (DC) in Drosophila and ventral closure in Caenorhabditis elegans (Agnès and Noselli,1999; Sabapathy et al.,1999; Simsje and Hardin,2001, for review; Wei et al.,2001; Harden,2002; Martin and Parkhurst,2004). Indeed, it seems that the molecular mechanisms underlying these processes are conserved across the animal kingdom (Martin and Lewis,1992; McCluskey et al.,1993; Brock et al.,1996; Davidson et al.,2002; Wood et al.,2002). Thus, investigations not only in vertebrates but also in invertebrate model organisms could help to understand how wound healing and regeneration of tissues/organs occur. Among insects, the fruit fly Drosophila melanogaster is the most commonly used model organism, and genetic, cellular and molecular studies in this system have provided important contributions in those fields. The first studies revealing the mechanisms of wound healing and regeneration in Drosophila showed that injured larval imaginal discs transplanted into female abdomens were able to regenerate and differentiate (Hadorn and Buck,1962; Hadorn,1968). Moreover, embryonic DC, imaginal disc fusion/regeneration, and regeneration of wounded epithelia in embryos, larvae, and adult organisms have been studied to highlight the genetic and cellular events underlying tissue repair (Fig. 2; Agnès and Noselli,1999; Zeitlinger and Bohmann,1999; Martín-Blanco et al., 2000; Jacinto et al.,2001; Martin and Wood,2002; Rämet et al.,2002; Wood et al.,2002; Galko and Krasnow,2004; Martin and Parkhurst,2004; Bergantiños et al.,2010b). Fly organs such as the midgut, ovaries, or testes are used as models in which to study homeostatic regeneration and to determine the molecular mechanisms that control stem cell behavior (Fig. 2; Ohlstein and Sprading, 2006; Singh and Hou,2008,2009). Drosophila presents many advantages when compared with other animal models of wound healing and regeneration. It is a comparatively simple organism with less genetic redundancy than vertebrates, and fundamental cellular processes as well as many genes and signaling pathways are conserved between insects and vertebrates. Furthermore, the availability of powerful genetic tools and the deep knowledge of many Drosophila developmental processes at the genetic and cellular levels are features that make this organism an ideal model system to address novel biological questions including those relevant to human health. It is also possible to perform genetic screens in Drosophila, which allows genome-wide analyses of genetic interactions based on the dominant modification of a given phenotype obtained by loss or gain of function of the gene of interest. These screens have led to the identification of genes and signaling pathways involved in DC and imaginal disc regeneration that also regulate wound healing and regeneration in vertebrates (see Table 1 in Martin and Parkhurst,2004). Importantly, the application of the yeast GAL4/UAS system in Drosophila allows the study of the effects of overexpression or inactivation of any gene in a selected tissue and developmental stage (Brand and Perrimon,1993). Furthermore, the use of advanced microscopy techniques such as time-lapse confocal microscopy for live imaging and laser ablation experiments on transgenic embryos expressing fluorescent proteins has helped to describe the exact sequence of cell movements occurring during epithelial repair and regeneration in Drosophila (Wood et al.,2002).
Numerous works and reviews have been previously published describing the various morphogenetic Drosophila processes that are good systems for the study of wound healing and regeneration in vertebrates. In this review, we provide an overall and comprehensive view of how Drosophila has contributed to the understanding of such processes. First, we present a brief description of the cellular and molecular mechanisms underlying wound healing and regeneration in vertebrates to introduce the reader in the field. Next, we review the Drosophila morphogenetic events that can be studied as models to understand those processes, including embryonic DC, wound healing in embryos and larvae, imaginal disc fusion and regeneration, also reinforcing the potential of germline and midgut stem cells (Fig. 2). We also present the common genetic pathways that trigger these processes and discuss the main contributions of Drosophila research to our understanding of wound healing and regeneration, highlighting the most important advances and techniques developed from fly research (Box 1), as well as the similarities and evident differences that exist between the Drosophila and vertebrate healing/regeneration processes (Box 2). Finally, we present future applications of Drosophila as a model of wound healing and regeneration, and discuss how this organism could help to advance this field.
CELLULAR AND MOLECULAR BASES OF WOUND HEALING AND REGENERATION IN VERTEBRATES
As mentioned above, tissues and organs can be structurally or functionally injured by external or internal causes, and recovery can be achieved by two different processes: wound healing and scar formation or wound healing followed by regeneration (Fig. 1). In this section we describe the cellular and molecular bases of both processes.
Wound Healing: Travelling Through a Wound
When we have a wound after injury or surgery, our epithelium, dermis, and vascular system as well as other tissues may be disrupted. An orchestrated process begins involving various cell types and extracellular matrix (ECM) components, which need to collaborate and coordinate to protect the organism and heal the wound. In general, the wound healing process is initiated by mechanical and chemical signals and can be summarized in three steps that overlap in time: inflammation, new tissue formation and maturation/remodeling (Fig. 1).
First, bleeding and inflammation appear to prevent infections and accelerate wound healing. Subsequently, platelet aggregation leads to the formation of a fibrin-rich clot, which controls active bleeding (hemostasis), and the release of growth factors that promote angiogenesis, inflammation and migration of keratinocytes, fibroblasts, neutrophils and macrophages (Martin,1997; Werner and Grose,2003; Dovi et al.,2004). In addition, cytokines, interleukins and interferons activate the serum response factor (SRF), which induces the expression of several genes (including fos, jun, and the early growth response genes egr-1 and egr-2) involved in cell growth, ECM remodeling and growth factor action (Chai and Tarnawski,2002; Grose et al.,2002; Fu et al.,2003). As soon as the mechanisms of defense have fulfilled their mission, the organism has to face the task of repairing the wound. This second stage occurs within two days to three weeks in distinct phases that partially overlap in time: angiogenesis, fibroplasia and granulation tissue formation, collagen deposition, epithelialization/proliferation and wound contraction (Midwood et al.,2004, for review). Endothelial ASCs migrate from uninjured blood vessels and develop pseudopodia by establishing new blood vessels (a process known as angiogenesis), which supply oxygen and nutrients to cells that are rebuilding the affected tissue (Greenhalgh,1998). Simultaneously, fibroblasts migrate from uninjured tissues or arise from blood-borne ASCs (Song et al.,2010) and proliferate into the wound (a process called fibroplasia). Then granulation tissue formation occurs, in which fibroblast form a new, provisional ECM by excreting collagen and fibronectin. Keratinocytes, fibroblasts, inflammatory cells, and endothelial cells migrate through this new ECM to the damaged area for reepithelialization (Midwood et al.,2004, for review). Later, cell proliferation starts, and a series of events takes place: (1) epithelial cells at the edges of the wound start to divide and grow toward each other and eventually join to form an intact layer of skin, (2) cell migration forms a scar by the loss of inhibitory contact, (3) mitosis starts in the basal cell layer of the adjacent epithelium, and (4) differentiation with vertical stratification occurs. Subsequently, deep cells undergo contraction and ingression, and fibroblasts are stimulated by growth factors to differentiate into myofibroblasts, which are responsible for wound contraction (Mirastschijski et al.,2004). Finally, and for the next three weeks to years, collagen is remodeled and realigned along tension lines, and unnecessary cells are eliminated by apoptosis (Haslett,1992; Hinz, 2007). A scar is the memory of the wound in adult organisms.
It should be noted that several differences exist between adult and embryonic wound healing. First, tissue movements during embryonic wound healing use cell machinery fundamentally different from that driving the analogous movements in adults. In embryonic wounds, in some adult tissues and in small wounds where cell–cell contacts exist, an actomyosin cable assembles at the apical edges of the cells at the wound margin to close it (McCluskey and Martin,1995; Danjo and Gipson,1998). This supramolecular structure constricts the limits of the wound acting like a purse string (Martin and Lewis,1992; McCluskey and Martin,1995; Brock et al.,1996; Danjo and Gipson,1998), and actin-rich protrusions form filopodia and lamellipodia from the wound margins helping them to fuse together (Martin and Lewis,1992; Bement et al.,1993; Brock et al.,1996; Wood et al.,2002). Second, mammalian fetal wounds heal rapidly either without or with minimal inflammation and scarring, showing signs of regeneration (Hess,1954a,b; Dixon,1960; Burrington,1971; Wilgus,2007). It seems that the inflammatory response in adults is produced by different combinations and levels of growth factors and cytokines that enhance scarring (Whitby and Ferguson,1991; Ferguson and O'Kane,2004). However, there are adult tissues, such as the oral mucosa, that retain the abilities observed in embryonic tissues (Szpaderska et al.,2003; Wong et al.,2009), thus suggesting the existence of different environmental and tissue properties for wound healing.
Regeneration: Blastema Formation and the Role of Adult Stem Cells
Regeneration involves a combination of cell rearrangements and compensatory proliferation to restore damaged tissues. Multicellular organisms use three mechanisms of regeneration: (1) proliferation of the residual differentiated healthy cells in the wound (compensatory hyperplasia or real regeneration), (2) dedifferentiation of cells that adopt early patterns of gene activity and proliferate, or (3) activation and proliferation of ASCs, undifferentiated cells located in regenerative tissues (Fig. 1). Compensatory hyperplasia is the proliferation of residual cells in the wound directed to restore tissue mass and integrity while maintaining most of their differentiated functions. In some amphibians and certain species of fish, dedifferentiation of cells during adult limb regeneration involves the formation of a blastema, which classically was considered a mass of morphologically undifferentiated, pluripotent proliferating cells that are covered by an epithelium and differentiate to replace the missing structures. Several key questions about blastema cells are still unknown, such as if they derive from differentiated or undifferentiated cells, if they are pluripotent, and whether and how the cells committed to a specific cell lineage can differentiate to form another cell type. Indeed, several reports have demonstrated the existence of a set of transcription factors that can generate pluripotent cells from fibroblasts and other cell types (Okita et al.,2008; Hanna et al.,2008). However, recent research indicates that in some organisms blastema cells may retain a memory of their tissue or embryonic origin (Kragl et al.,2009). By tracking limb tissues marked by a GFP transgene in axolotls, it has been shown that limb regeneration can be achieved without complete dedifferentiation to a pluripotent state (Kragl et al.,2009), suggesting that the blastema is not formed by a unique cell type and that regeneration does not require cells to undergo acute reprogramming. Finally, activation of quiescent lineage-restricted ASCs by specific signals is a common mechanism by which adult tissues in amphibians, reptiles, mammals and Drosophila regenerate (Gurley and Alvarado,2008, for review). For example, after skin tissue injury, ASCs mobilize from the bone marrow into the pool of circulating cells (Song et al.,2010). These cells migrate to the site of injury and regulate proliferation and migration of epithelial cells/dermal mesenchymal cells during the early inflammatory phase (Fathke et al.,2004). Indeed, simpler animals such as planarians have an enhanced capacity to regenerate because adults retain clusters of ASCs within their bodies that migrate to wound sites, where they divide and differentiate to reform the missing tissue and organs. Recently, Wagner and co-workers provided evidence that neoblasts, which are adult pluripotent stem cells, underlie the remarkable regenerative abilities of planarians (Wagner et al.,2011). ASCs are also present in most adult tissues/organs and serve as a reservoir for tissue renewal and repair by homeostatic regeneration like adult intestinal stem cells (ISCs), which control intestinal regeneration in mammals and flies (see below). During homeostatic regeneration, older or damaged differentiated cells are eliminated by apoptosis and replaced by cells derived from ASC division. Subsequently, newly generated cells differentiate and become functionally integrated within the preexisting tissue. Interestingly, deregulation of ASC behavior results in cancer formation, tissue degeneration and premature aging (Casali and Batlle,2009). Little is known about the intrinsic programs that trigger cell death and how they are coordinated with ASC division and differentiation in adult organs, but one critical parameter appears to be how ASCs and their local microenvironments, or niches, interact. Both neighboring cells and the ECM can provide cues that regulate ASC behavior. This can entail signaling pathways, physical interactions, metabolic cues, and neural inputs (Molofsky et al.,2004, for review; Scadden,2006).
Regeneration mechanisms vary according to species and tissue (see Table 1 in Stocum and Zupanc,2008). Although different, all these processes share signaling pathways and molecular mechanisms to restore normal function of the affected tissue (Bosch et al.,2005; Mattila et al.,2005; Gurtner et al.,2008). Determining how common signaling pathways are used in different processes is one of the main questions in developmental biology and regenerative medicine.
Signaling Pathways and Epigenetic Factors Involved in Vertebrate Wound Healing/Regeneration
It has been shown that wound repair and regeneration are mainly regulated by the highly evolutionary conserved Jun N-terminal kinase (JNK) and Wnt/Wg pathways. However, several studies have identified a variety of different genes, other pathways and factors involved in these processes. More recently, the involvement of epigenetics in wound healing and regeneration has been demonstrated. A detailed description of these signaling pathways and epigenetic mechanisms is provided in the following sections.
The JNK pathway.
The JNK pathway is a conserved class of Mitogen-Activated Protein Kinase (MAPK) signaling pathway composed of a series of serine-threonine kinases whose activity results in phosphorylation of c-Jun and other transcription factors in mammals and invertebrates (Adler et al.,1992; Hibi et al.,1993; Dérijard et al.,1994; Kyriakis et al.,1994; Agnès et al.,1999; Rämet et al.,2002; Huang et al.,2003; Chang et al.,2003). Numerous upstream activators and downstream effectors of this pathway have been described (Fig. 3A). Indeed, several small GTPases of the Rho family are upstream activators of JNK signaling during wound healing in mammals and Drosophila (Brock et al.,1996; Su et al.,1998; Harden,2002; Stronach and Perrimon,2002; Yin et al.,2008; Baek et al.,2010), regulating the actin cytoskeleton, cell migration and proliferation (Hall,1998; Kaibuchi et al.,1999). However, the specific signal(s) required to trigger JNK activation is(are) still unknown in many contexts. Growth factors, mechanical stretching of the wound edge cells, loss of cell polarity or necrotic signals are all candidates for upstream activators of this signaling pathway (Rämet et al.,2002; Nelson et al.,2005; Igaki et al.,2006). Alternatively, release of intracellular ATP or Ca2+ by damage or signals secreted by inflammatory cells recruited to the wound may participate in JNK activation (Kushida et al.,2001). The JNK pathway has been implicated in epithelial migration and multiple developmental/homeostatic processes across the animal kingdom, including cellular stress responses, the innate immune response, planar cell polarity (PCP) establishment, apoptosis, and compensatory cell proliferation following injury, among others (Sluss et al.,1996; Boutros et al.,1998; Paricio et al.,1999; Weber et al.,2000; Boutros et al.,2002; Ryoo et al.,2004; McEwen and Peifer,2005). In mammals, inhibition of JNK signaling results in proliferation impairment and cell/tissue movement defects. Indeed, knockout mice for Jun present eyelid closure defects and defects in fibroblasts migration during wound healing (Grose,2003; Javelaud et al.,2003; Li et al.,2003; Zenz et al.,2003; Weston et al.,2004). Furthermore, mammalian AP-1 transcription factors (Fig. 3A) regulate the expression of several genes involved in wound healing, such as those encoding matrix metalloproteinases (MMPs), integrins, cytokines, and growth factors (Angel et al.,2001, Yates and Rayner,2002; Florin et al.,2004; Zenz et al.,2008) and that act during the inflammatory/proliferative phases and in keratinocyte migration (Fig. 3B) (Kondo and Yonezawa,2000; Li et al.,2003). It has also been shown that the Scribble (Scrib) complex, an important determinant of apicobasal polarity, contributes to epithelial cell migration and fusion during both development and wound healing in mammals by recruiting the small GTPases Cdc42, Rac and the serine-threonine kinase Pak to the leading edge (LE) of migrating cells (Murdoch et al.,2003; Zarbalis et al.,2004; Humbert et al.,2006; Dow et al.,2007; Bahri et al.,2010). JNK signaling is required for regeneration after wounding in zebrafish (Kawakami,2010, for review), and it plays a crucial role in the differentiation of blastema neoblasts by regulating the G2/M cell cycle transition during planarian regeneration (Tasaki et al.,2011). Finally, it was shown that expression of Wnt target genes was stimulated by JNK activation, accelerating tumorigenesis (Sancho et al.,2009).
In Wnt/β-catenin signaling, secreted Wnt proteins bind to cell-surface receptors of the Frizzled (Fz) family and activate Dishevelled (Dsh) family proteins, which in turn inhibit a complex of proteins that includes axin, GSK-3 and APC. This leads to the accumulation of cytoplasmic β-catenin protein, which translocates to the nucleus and interacts with transcription factors of the TCF/LEF family to promote specific gene expression, which ultimately leads to regulation of cell proliferation and cell fate as well as cell transformation (Logan and Nusse,2004; Widelitz,2005; Clevers,2006; Cheng et al.,2008). In absence of the Wnt/Wg signal, the axin/GSK-3/APC complex phosphorylates β-catenin and triggers its cytoplasmic degradation, and Wnt-responsive genes are not transcribed (Harden,2002, for review). Moreover, there are several noncanonical Wnt/Wg signaling pathways in Drosophila and vertebrates that appear to function in a β-catenin-independent manner (Cheng et al.,2008). The two best studied noncanonical Wnt pathways are the PCP/Wnt and Wnt/Ca2+ pathways (Kühl et al.,2000). Numerous studies have shown that Wnt signaling is involved in cell shape changes and polarization during wound healing in insects and vertebrates (Labus et al.,1998; Harden,2002; Baek et al.,2010; Zhang et al.,2009, for review). However, it seems that canonical and noncanonical Wnt signaling pathways may have opposing effects during this process (Kühl et al.,2001; Ishitani et al.,2003; Weidinger and Moon,2003). β-catenin is also a component of adherens junctions (AJ), in association with cadherins, which are essential for cell adhesion and have to be disassembled to allow cell migration, an important process during wound healing. During regeneration, Wnt/β-catenin signaling is necessary for blastema formation and regeneration of the zebrafish tail fin and adult heart, as well as the hindlimbs of Xenopus tadpoles (Lepilina et al.,2006). It is established that a β-catenin-independent, noncanonical Wnt network antagonizes β-catenin signaling during tail fin regeneration (Stoick-Cooper et al.,2007). Moreover, Wnt/β-catenin signaling seems to play a role in mammalian liver and intestinal regeneration. First, conditional β-catenin knockout in the mouse liver causes decreased hepatic cell proliferation and suboptimal regeneration (Sodhi et al.,2005; Tan et al.,2006). Wnt signaling is responsible for maintaining ISCs in an undifferentiated state, and its activation increases ISCs mitosis (Crosnier et al.,2006). Indeed, conditional mutations in the APC gene, a negative regulator of the Wnt cascade, lead to ISC mitosis and epithelial multilayering, which promotes gastrointestinal cancer (Cordero et al.,2009, for review). Importantly, JNK and insulin receptor (InR) pathways activate Wnt signaling and regulate ISC proliferation (Nateri et al.,2005). Thus, manipulation of Wnt signaling could be potentially beneficial for regenerative therapies.
Notch (N) is a membrane receptor involved in lateral inhibition (Heitzler and Simpson,1991; Artavanis-Tsakonas et al.,1995). Interaction between N and the Delta (Dl)/Jagged ligand leads to the cleavage of the N intracellular domain (ICD) (Heitzler and Simpson,1991), which is translocated to the nucleus to bind to the CSL transcription factor (also known as RBP-J in mice, CBF-1 in humans and Suppressor of Hairless (Su(H)) in Drosophila) and activates the transcription of N target genes (Bray,2006; Fiúza and Arias,2007). N signaling has been implicated in wound healing and tissue repair, regeneration and stem cell turnover. Indeed, both activation and inhibition of N signaling alter the behavior of cultured vascular endothelial cells, keratinocytes and fibroblast in a scratch wound healing model in mice (Chigurupati et al.,2007). The N cascade has been shown to regulate endothelial cell proliferation and migration during angiogenesis in normal tissues and tumors (Nakajima et al.,2003; Sainson et al.,2005; Takeshita et al.,2007; Paris et al.,2005). N signaling is also required for normal tail regeneration in Xenopus, as its inhibition completely abolishes this process while its overexpression has the opposite effect (Beck et al.,2003). Furthermore, N signaling regulates intestinal regeneration in zebrafish and humans, induces ISCs proliferation and promotes intestinal cell determination of ISC progeny (Crosnier et al.,2006; Fre et al.,2005; Stanger et al.,2005; van Es et al.,2005; van der Flier and Clevers,2009). Consistently, ablation of N target genes impairs intestinal epithelial homeostasis (Jensen et al.,2000; Yang et al.,2001; Lee et al.,2002).
The cytokine-activated Janus kinase (JAK)/signal transducer and activator of transcription (STAT) is a conserved signaling pathway composed of a large family of extracellular ligands that bind to transmembrane receptors that are associated with JAK kinases. Following stimulation, ligand/receptor/JAK complexes are activated by tyrosine phosphorylation to recruit STATs, which are then phosphorylated and translocated to the nucleus to activate target genes. This pathway plays an important role promoting cell proliferation in different cellular contexts and in response to cytokines and growth factors (Zeidler et al.,2000). It has been shown that JAK/STAT signaling is activated during the wound healing process by IL-6, leading to increased inflammation, ECM degradation, and cancer (Yoshimura et al.,2007; Atreya and Neurath,2008; Iliopoulos et al.,2009). The JAK/STAT pathway affects the expression of numerous genes, including those involved in cellular proliferation and migration (Hirano et al.,2000). Additionally, it increases collagen expression at both mRNA and protein levels, supporting the idea that this pathway is critical in modulating fibrosis (Lakner et al.,2010).
Other activating signals.
Additional signaling pathways and transcription factors are required during wound healing after epithelial injury (Schäfer and Werner,2007, for review). It has been shown that an evolutionary conserved pathway that includes the transcription factors Fos and Grainy head (Grh) is required for mammalian skin and Drosophila cuticle repair (Yates and Rayner,2002; Li et al.,2003; Mace et al.,2005; Ting et al.,2005; Mehic et al.,2005; Zenz and Wagner,2006; Gazel et al.,2008; Pearson et al.,2009). Of interest, Ting et al. (2005) found that knockout mice for one of the Drosophila grh homologs, grainy head-like3 (grhl3) show defects in epidermal protein barrier cross-linking and in epidermal embryonic wound repair. Thus, the Grh transcription factor is important for constructing and healing the animal body surface barrier both in the insect cuticle and mammalian skin.
Epigenetics in wounds.
There is a growing body of evidences supporting the idea that cells reprogram their gene expression profiles during wound healing and regeneration. Reprogramming or modification of genetic programs without modification of gene sequences involves the concept of epigenetics. Different epigenetic mechanisms have been described, including chromatin modification and microRNAs (miRNAs)-based mechanisms (Solter et al.,2004). miRNAs are 21- to 22-nucleotide-long RNA molecules that, among other functions, inhibit gene expression by binding to the 3′UTRs of target mRNAs, thereby blocking their translation or destabilizing them (Carrington and Ambros,2003; Bartel,2004; Lai,2004). Of interest, miRNAs act at several phases of wound healing, including inflammation, angiogenesis, fibroblast and keratinocyte function and apoptosis (Bavan et al.,2011, for review). They are critical factors in immune cell function and control of cytokine networks, and they regulate fibroblast proliferation and the differentiation of wound-specific cells (Gu and Iyer,2006; Asirvatham et al.,2009; Lu and Liston,2009). miRNAs also regulate the expression of MMPs and collagen, determining ECM composition and fibroblast activity (Chen et al.,2009; Roy et al.,2009; Stanczyk et al.,2011), thus affecting scar formation (Cheng et al.,2010). Moreover, miRNAs are involved in zebrafish fin regeneration (Yin et al.,2008; Yin and Poss,2008) and have been identified in regenerative tissues in planaria (González-Estévez et al., 2009; Tian et al.,2011), chicken (Frucht et al.,2011), newts (Nakamura et al.,2010), and rats (Castro et al.,2010). On the other hand, Polycomb group (PcG) proteins are a family of conserved epigenetic modifiers that direct cellular fates during embryogenesis, repressing the expression of homeotic genes by means of histone modifications (Sparmann and van Lohuizen,2006). PcG proteins are downregulated during wound healing, thus inducing the expression of genes involved in repair (Shaw and Martin,2009). Furthermore, two opposing functions have been described for PcG proteins in embryonic cells and ASCs: they are required to maintain stem cell multipotency by repressing developmental genes but also antagonize stem cell self-renewal, thus facilitating cell lineage differentiation (Su et al.,2011, for review).
DROSOPHILA MODELS OF WOUND HEALING AND REGENERATION
Although pioneering studies of embryonic and adult wound healing were performed in vertebrates, the genetic and cellular events underlying these processes were only described in detail after analyzing Drosophila developmental morphogenetic events and Drosophila wound healing at several developmental stages (see Box 1). In this section we describe, at the cellular and molecular levels, the Drosophila morphogenetic processes that are used as in vivo models of wound healing and regeneration in vertebrates (Fig. 2). Similarities and differences existing between these processes in flies and vertebrates are summarized in Box 2. In addition, we present the different approaches used to dissect the cellular and molecular mechanisms involved in those processes. Moreover, we also highlight the potential of Drosophila germline and midgut stem cells to study homeostatic regeneration.
Embryonic Dorsal Closure: A Developmental Model of Wound Healing
DC is a morphogenetic process that takes place at the end of Drosophila embryogenesis (Martinez-Arias, 1993) after germ band retraction (GBR) and coincident with head involution (HI) (VanHook and Letsou,2008, for review). At this developmental stage, embryos exhibit a large dorsal hole covered by an extra-embryonic epithelial tissue, the amnioserosa (AS), which lays over the yolk sac and is surrounded by two epithelial sheets (Fig. 4A). During DC, the hole closes by migration of the lateral epithelial sheets toward the dorsal midline and subsequent fusion or zippering (Fig. 4A). Simultaneously, the AS constricts apically and then disappears through basal extrusion of cells and activation of a cell death program, similar to mesenchymal contraction in a wound (Kiehart et al.,2000; Jacinto et al.,2002; Fernández et al.,2007; Gorfinkiel and Arias,2007; Toyama et al.,2008). Defects during DC cause clear phenotypes in embryos, which present a large hole in the dorsal epidermis (Fig. 4B). The use of time-lapse confocal microcopy techniques in transgenic embryos expressing a GFP-Moesin (GFP-Moe) fusion protein (Edwards et al.,1997) allowed the examination of cell and tissue dynamics during DC and revealed it to be a powerful model of epithelial morphogenesis, migration and resealing that recapitulates the process of wound healing (Harden,2002; Jacinto et al.,2002). The resemblance between DC and wound healing was also emphasized with live-imaging techniques to analyze DC and cell dynamics after mechanical perturbation in embryos (laser ablations or incisions) (Kiehart et al.,2000; Hutson et al.,2003; Stramer et al.,2005; Peralta et al.,2008; Rauzi et al.,2008; Rodriguez-Diaz et al.,2008; Toyama et al.,2008; Solon et al.,2009; Gorfinkiel et al.,2009; Gettings et al.,2010; Blanchard et al.,2010). These techniques have provided quantitative information on tissue movements during DC and have led to the development of biophysical models that use mathematical language to represent physical and kinematic properties of the tissues involved in DC (Gorfinkiel et al., 2011, for review).
DC requires coordinated cell shape changes within the AS and the lateral epidermis. At the LE of the two lateral epidermal sheets, both tissues contribute to the formation of an actomyosin cable (Young et al.,1993; Mizuno et al.,2002), similar to that previously observed in wounded chick embryos (Martin and Lewis,1992). It was proposed that the LE generates a contractile force oriented along the length of the cable and perpendicular to the direction of the closure (Hutson et al.,2003; Peralta et al.,2007) to coordinate the dorsal migration of the epithelial sheets toward the dorsal midline and to close the hole (Fig. 4C). This is known as the purse string model (Young et al.,1993). However, laser and genetic ablation experiments followed by physical modeling have suggested that other forces contribute to this process (Kiehart et al.,2000; Hutson et al.,2003; Peralta et al.,2007). Indeed, live-imaging studies using actin-GFP expressing embryos showed the formation of filopodia and lamellipodia structures beyond the LE of the epithelial sheets during DC (Jacinto et al.,2000; Wood et al.,2002; Jankovics and Brunner,2006; Gates et al.,2007; Liu et al.,2008). These structures recognize cells of the same anterior/posterior segment identity on each side of the hole, interdigitate and fuse by AJ formation (Fig. 4D) (Millard and Martin,2008). It has been shown that the fusion of epithelial sheets by actin projections starts first in the posterior and later in the anterior canthi through the formation of contacts modulated by E-Cadherin (E-Cad) and integrins (Danjo and Gipson,1998; Jacinto et al.,2000; Wood et al.,2002; Narasimha and Brown,2004; Fernández et al.,2007; Gorfinkiel and Arias,2007; Gorfinkiel et al.,2009). On the other hand, some evidence suggests that the AS plays an essential role during DC. First, it was demonstrated that reduction of the AS surface area is required to generate tension in epithelial sheets (Kiehart et al.,2000). Second, it was shown that AS cells in contact with the lateral epidermis also contain purse string-like structures (Wada et al.,2007). Finally, other experiments showed that removal of the entire AS by genetic ablation produced DC defects caused by disruption of mechanical and molecular signals (Scuderi and Letsou,2005). Taken together, these results suggested that the AS plays structural and signaling roles during DC. Subsequent studies have demonstrated that apical constriction of AS cells is an active process driven by Myosin II (Myo II) (Franke et al.,2005) and that a secreted signal which diffuses from the dorsal-most epidermal (DME) cells to the AS (the morphogen Decapentaplegic (Dpp), a member of the transforming growth factor-β (TGF-β) family) coordinates AS cell contraction and adhesion between AS and epidermal cells (Fernández et al.,2007). Moreover, it has been recently shown that AS cells have the intrinsic capability to pulse during DC through constant contracting and relaxing forces along their surfaces (Fig. 4E) (Solon et al.,2009; Gorfinkiel et al.,2009; David et al.,2010; Blanchard et al.,2010). The pulsed forces created by AS cells are sufficient to transiently displace the flanking epidermal cells dorsally (Solon et al.,2009). Then the epidermal actomyosin cable acts like a cellular ratchet that prevents the relaxation of AS cells during their periodic cycles of contraction and expansion (Solon et al.,2009). However, in the absence of a fully functional actomyosin cable the AS can contract, suggesting that AS cells have the intrinsic capability to contract (Laplante and Nilson,2006; Gorfinkiel et al.,2009). Although the AS is an important tissue during DC, it is not yet clear whether there is an equivalent structure in wound healing. However, it has been suggested that granulation tissue, which is composed of fibroblasts and myofibroblasts, produces contractile movements in wounds to contribute to zippering (Martin and Parkhurst,2004).
DC is primarily regulated by the JNK and Wnt/Wg pathways, similar to wound healing and regeneration in vertebrates, although other factors are involved. Therefore, DC has contributed to the discovery of molecular mechanisms underlying key processes in vertebrates. The JNK pathway is activated at the LE and in AS cells at the end of GBR and downregulated in the AS and maintained at the LE during DC, where it induces the expression of at least two genes: dpp and puckered (puc) (Fig. 3A) (Noselli,1998, for review; Noselli and Agnès,1999; Glise and Noselli,1997; Knust,1997; Martin-Blanco, 1997; Reed et al.,2001; Stronach and Perrimon,2001). It is required for AS cell contraction (Stronach and Perrimon,2001; Fernández et al.,2007) and during actin stress fiber formation, actomyosin cable assembly and filopodia polymerization (Kiehart et al.,2000; Jacinto et al.,2000,2002), formation of focal adhesion complexes in DME cells to regulate adhesion and possibly signal transduction (Ricos et al.,1999; Reed et al.,2001; Kaltschmidt et al.,2002; Homsy et al.,2006), cell-shape changes and ventral epithelial elongation (Fig. 3C) (Glise et al.,1995; Riesgo-Escovar et al.,1996; Sluss et al.,1996; Hou et al.,1997; Kockel et al.,1997; Riesgo-Escovar and Hafen,1997a,b; Zeitleinger et al., 1997; Ricos et al.,1999). Moreover, the JNK cascade has been recently shown to be involved in reprogramming and intercalation of cells at the LE, which helps to relax tissue tension (Gettings et al.,2010). Different assays have been performed to identify new JNK targets during Drosophila DC (Jasper et al.,2001; Thomas et al.,2009). Some of these genes encode cytoskeleton regulators such as Chickadee (Chic) involved in actin polymerization in the LE, cell adhesion molecules (like Myospheroid and RalA), proteins involved in ecdysone response (such as ImpE1 and ImpL1), the trypsin-like protease Scarface (Scaf) that negatively regulates JNK activity in the epidermis and the small GTPase Rab 30, which is involved in DC, HI and thorax fusion (see below) through the regulation of intracellular transport of vesicles in the LE (Thomas et al.,2009; Rousset et al.,2010). Other genes encoding JNK regulators or effectors involved in DC are: raw, which is expressed in the AS and influences JNK activity in the LE (Byars et al.,1999), and cabut (cbt) that encodes a transcription factor expressed in the yolk sac, AS, and lateral epidermis, and functions downstream of JNK signaling (Muñoz-Descalzo et al.,2005; Belacortu et al.,2011). Regarding Wnt/Wg signaling, it is required for normal elongation of DME cells, actomyosin cable formation (Kaltschmidt et al.,2002; Morel and Arias,2004), and promotes expression of target genes in the DME cells in collaboration with JNK signaling (McEwen et al.,2000). Arm also contributes to the DC process in a Wg-independent manner through its role in AJ formation (see below) (McEwen et al.,2000). In addition, it has been shown that N is involved in DC through its ICD, specifically regulating patterning of the dorsal epidermis during embryogenesis in a Su(H)-independent manner. Indeed, reduction of N function overactivates the JNK pathway during DC and can rescue bsk and hep mutant phenotypes, thus suggesting that N activity can repress JNK signaling (Zecchini et al.,1999). Cell junctions also contribute to cell shape changes and to transmit the mechanical forces generated during DC. Therefore, different AJ components, such as Canoe (Cno), ZO-1, E-Cad and Arm have been shown to be involved in DC (Brock et al.,1996; Takahashi et al.,1998; McEwen et al.,2000; Tepass et al.,2001; Harden,2002; Gorfinkiel and Arias,2007). Similarly, components of septate junctions (SJ), such as Coracle (Cora), Neurexin IV (NrxIV), Disc large (Dlg), Scrib and Lethal giant larvae (Lgl), are also involved in this process (Lamb et al.,1998; Perrimon,1988; Bilder and Perrimon,2000; Arquier et al.,2001). Proteins involved in intracellular trafficking by vesicles and recycling of AJ components, such as Rab 5, Rab 11 and Rab 30 are also involved in DC, HI and thorax fusion (see below) (Roeth et al.,2009; Sasikumar and Roy,2009; Thomas et al.,2009; Mateus et al.,2011). Other factors such as miRNAs of the 310 family (310/311/12/313/92) have a role in DC, as their silencing causes DC and HI defects in Drosophila embryos (Leaman et al.,2005). Moreover, several lines of evidence support a relationship between ecdysteroid hormones and DC (Fernández et al., 1995; Lamka and Lipshitz,1999; Chávez et al.,2000; Kozlova and Thummel,2003; Petryk et al.,2003; Chavoshi et al.,2010). First, one of the 20-hydroxy-ecdysone (20E) hormone peaks in the Drosophila life cycle occurs during DC. Second, embryos expressing a dominant negative form of the 20E receptor (EcR-DN) and embryos homozygous for the thermosensitive ecdysoneless1 (ecd1) allele, which synthesize low quantities of the hormone, present GBR, CD, and HI defects (Fernández et al., 1995; Lamka and Lipshitz,1999; Kozlova and Thummel,2003; Chavoshi et al.,2010). Moreover, loss of function of genes such as disembodied (dib) and shade (shd) that encode P450 enzymes involved in embryonic ecdysone biosynthesis present DC defects (Chávez et al.,2000; Petryk et al.,2003). In addition, early hormone response genes such as ImpE1 and ImpL1 have been identified as JNK targets during DC (see above) (Jasper et al.,2001). Finally, we have demonstrated in our laboratory that the Cbt transcription factor is probably regulated by ecdysone signaling during DC (Y.B. and N.P., unpublished results), as it occurs during metamorphosis (Beckstead et al.,2005). However, more experiments are necessary to unravel the exact role of ecdysteroids during DC.
Wounding Drosophila Embryos, Larvae, and Adults
Other studies in Drosophila have been performed after inducing mechanical wounds in embryos, larvae and adult organisms (Fig. 2), which are used as models of wound healing in vertebrates. However, differences between this process in flies and vertebrates exist (Box 2). These studies have revealed that during late embryogenesis and larval development, wounding caused an immediate barrier response by forming a plug in the wound gap, which later melanized to form a scab (Galko and Krasnow,2004). Time-lapse studies of wound healing in Drosophila embryos have shown that cells move by coordinately changing cell–cell relationships and that only in small wounds is an actomyosin cable assembled to close the wound (Bement et al.,1993; Wood et al.,2002). It seems that not only the wound LE but also cells lying distal to that LE actively participate in epithelial cell sheet migration during wound hole closure (Rämet et al.,2002; Kwon et al.,2010). Recently, the mechanisms of single-cell wound repair in syncytial embryos have been also studied (Abreu-Blanco et al.,2011). It seems that plasma membrane mobilization and assembly of a contractile actomyosin ring are required for this process (Abreu-Blanco et al.,2011). In larval wounds, epidermal cells surrounding the wound site orient toward it and fuse to form a syncytium (without proliferation), and subsequently cells spread along and through the wound to reestablish epithelial continuity using lamellipodia but not an actomyosin cable (Galko and Krasnow,2004). Filopodial matching is also observed during healing of embryonic wounds in the ventral epithelium, suggesting that molecules mediating recognition are found not only at the LE of the epidermal sheets but throughout the embryonic epithelium (Millard and Martin,2008). Analyses of wounds induced in puc-LacZ adult flies (a JNK signaling reporter) revealed the formation of filopodia and lamellipodia at the wound LE as well as activation of the JNK pathway in epidermal cells close to the wound (Fig. 3D) (Rämet et al.,2002). The requirement of JNK signaling during wound healing in Drosophila embryos and larvae was also evident in mutants of the JNK downstream effector fos gene, which present defects in cell spreading and wound closure (Rämet et al.,2002; Galko and Krasnow,2004). It appears that the JNK pathway is activated in a gradient emanating from the wound limited by the scab and diffuses to the syncytium to promote cell spreading and reepithelialization. It has also been shown that the JNK pathway and its upstream regulators, the Rho family of small GTPases, are involved in localization of Nonmuscle Myo II to larval wounds (Rämet et al.,2002; Jacinto et al.,2002; Galko and Krasnow,2004; Mattila et al.,2005; Baek et al.,2010; Kwon et al.,2010). The JNK pathway and Fos activate the expression of MMP-1, which degrades matricellular proteins during wound healing, as well as during disc eversion/regeneration and tumor invasion in Drosophila (Uhlirova and Bohmann,2006; Srivastava et al.,2007; McClure et al.,2008). Moreover, the unpaired (upd) genes, which encode the JAK/STAT-activating cytokines, are also upregulated through JNK activation during Drosophila wound healing (Pastor-Pareja et al.,2008). Genetic screens have been recently performed to identify new genes involved in wound healing and/or regeneration downstream of JNK signaling. These studies revealed that different JNK pathway components are acting in larval wound healing and DC (Campos et al.,2010; Lesch et al.,2010), similar to other processes (Boutros et al.,2002; Silverman et al.,2003; Geuking et al.,2009; Igaki,2009). Several studies have also suggested that the transcription factors Fos and Grh activate the Dopa decarboxylase (Ddc) and Tyrosine hydroxylase (Ple) enzymes, which are responsible for producing reactive quinones that crosslink chitin and cuticle proteins in Drosophila (Mace et al.,2005). As mentioned above, both are required in mammalian skin repair (Yates and Rayner,2002; Li et al.,2003; Ting et al.,2005; Mehic et al.,2005; Zenz and Wagner,2006; Gazel et al.,2008; Pearson et al.,2009). It seems that Grh and Fos are directly regulated by Extracellular signal-Regulated Kinase (ERK) phosphorylation (Uv et al.,1997; Ciapponi et al.,2001) and that activated ERK is required for a robust wound response in Drosophila (Mace et al.,2005). Moreover, it has been recently reported that the Grh target stitcher (stit) gene, which encodes a Ret-family receptor tyrosine kinase, is required for efficient epidermal wound healing (Wang et al.,2009). The transcriptional stimulation of stit after injury triggers a positive feedback loop increasing the magnitude of epithelial responses by ERK phosphorylation (Wang et al.,2009).
To discover new genes involved in wound healing and inflammatory response in Drosophila, microarray analyses in wild type and macrophage-deficient serpent (srp) mutant embryos (Stramer et al.,2008) as well as genetic screens in injured embryos and larvae (Campos et al.,2010; Lesch et al.,2010) have been performed. Comparison of the RNA profiles of laser wounded vs. unwounded wild type and srp mutant embryos led to the identification of wound-activated hemocyte genes such as secreted phospholipase A2 (PLA2; CG14507), which is conserved in mammals and controls eicosanoids biosynthesis, or the antimicrobial peptide drosomycin (Drs); hemocyte-specific genes such as the cell-cycle regulator cdc25 (Milchanowski et al.,2004) or the Drosophila procollagen-lysine dioxygenase 3 (PLOD3), which is involved in posttranslational modification of collagens (Stramer et al.,2008); and the inflammation-associated wound response GADD45 (growth arrest and DNA damage-inducible gene 45) gene, which could have a role in epigenetic regulation of wound target genes (Stramer et al.,2008). Interestingly, a murine homologue of Drosophila GADD45 was shown to be upregulated rapidly after wounding and this response was much reduced in PU.1 null mice (Stramer et al.,2008), in which inflammatory cells were missing (Cooper et al.,2005). This finding provides further evidence for an evolutionarily conserved repair response in flies and vertebrates. In a different screen, Campos et al. (2010) demonstrated that genes previously described to have a role in DC are also involved in embryonic epithelial repair, although they identified other genes such as βHeavy-spectrin. Similarly, Lesch et al. (2010) performed a UAS-RNAi screen in Drosophila larvae to discover genes involved in control and coordination of epithelial migration during postembryonic wounds and identified genes involved in JNK and Stress-Activated Protein Kinases (SAPK) signaling and actin cytoskeleton remodeling genes (gγ1, ced-12, arp14D, mbc, rac1, SCAR, and arp11). Their results suggest that different processes are essential for normal wound closure such as the directionality of the wound edge and the regulation of syncytium formation, among others. They also proposed a model of epidermal cell behavior in normal and perturbed wound healing to infer some of the biological functions that occur within the epidermal sheet and are required for proper closure (see Fig. 7 in Lesch et al.,2010).
Injury and Regeneration of Drosophila Imaginal Discs: A Model of Regenerative Biology
Imaginal discs are larval epithelial sacs composed of groups of differentiated cells that subsequently will give rise to the adult insect limbs (Cohen, 1993). Imaginal disc precursor cells appear in embryos as clusters of cells that invaginate from the embryonic epithelium to form those structures during larval development. During metamorphosis, they undergo morphogenetic changes to form the adult legs, wing, eyes, antennae, head capsule, halters, and genital organs. In a pioneering study, Hardon and co-workers found that fragments of Drosophila larval imaginal discs were able to heal and proliferate (Fig. 5A), but they did not differentiate, when implanted into the abdomen of adult females (Hadorn and Buck,1962). Differentiation could only take place when introduced in larvae that will undergo metamorphosis (Fig. 5A) (Hadorn et al.,1968). It was also shown that when 50% of imaginal disc cells were killed by X-ray irradiation an adult normal appendage could still develop (Haynie and Bryant,1976). This result suggested that compensatory proliferation of imaginal disc tissue was taking place, in which the remaining living tissue could rescue the appropriate size of the disc. This ability of imaginal discs to regenerate after injury has been confirmed by other researches either in vivo and in vitro (Marsh and Theisen,1999; McClure and Schubiger,2007). Subsequent studies support the activation of this compensatory proliferation activity after apoptosis or injury triggering either tissue repair or normal growth processes and regulating organ size during development (Wells et al.,2006; Fan and Bergmann,2008; Pérez-Garijo et al.,2009). Taken together, these results established Drosophila imaginal discs as an amenable experimental system in which to study tissue regeneration and to identify the genetic program underlying this process (Hadorn et al., 1968; Schubiger,1971; Bryant,1975; Bergantiños et al.,2010a,b).
Drosophila imaginal disc regeneration depends on the type of injury, the cellular context and the topology of the cut in the fragmented discs. Thus, imaginal discs regenerate by reconstructing the missing central tissue (regeneration by intercalation) if the two opposite peripheral pieces are cultured together, but they duplicate if they are cultured separately (Haynie and Bryant,1976; Bryant et al.,1978). Meinhardt established in 1983 that the compartment boundaries of imaginal discs are the signaling centers for regeneration and that the decision to regenerate or duplicate depends on the compartmental border confrontation after the injury (Meinhardt,1983). Understanding of this model came after the description of the patterns of morphogens that are released from compartment boundaries: BMP/Dpp and Wnt/Wg morphogens diffuse from the D–V and the P–D boundary, respectively (Campbell et al.,1993; Diaz-Benjumea et al.,1994; Campbell and Tomlinson,1995). Therefore, if the injury eliminates a complete compartment, cells expressing the morphogen required to specify one of the disc axis and for regeneration are not present, and the fragment duplicates. In contrast, if some cells located within the A–P and D–V boundaries remain in the fragment after the injury, the disc regenerates. As mentioned above, the cellular response after irradiation is a compensatory proliferation to replace the tissue (Haynie and Bryant,1976). However, if the damage is produced by surgical removal of tissue, an epimorphic process starts in which cytoskeletal rearrangements produce cell-shape and polarity changes. Indeed columnar and peripodial epithelial cells in the disc form a contractile actin cable and filopodia in the wound edges, thus closing the wound as in DC (Fig. 5B,C) (Bosch et al.,2005; Mattila et al.,2005). Later a blastema is formed (Abbott et al.,1981; Kiehle and Schubiger,1985; Bosch et al.,2005; Bergantiños et al.,2010a). The blastema cells in Drosophila imaginal discs are not pluripotent and retain their compartment origin identity, contributing only to the reconstruction of the damaged compartment (Bosch et al.,2008; Bergantiños et al.,2010a). For that reason, the main role of blastema cells is to maintain their disc-specific identity. For example, if the injured disc is a leg disc, after regenerative cell divisions, the new cells will be leg cells. However, some new cells exhibit stem cell-like properties and can switch to a different fate, specifying cells in an injured leg disc in wing cells, for example (Fig. 5A). This phenomenon, which is characterized by the change of cell identity by a reprogrammable mechanism, is known as transdetermination (Hadorn,1968; McClure and Schubiger,2007; McClure et al.,2008). Transdetermination has been ascribed to the action of ectopic morphogens, which induce cells to activate incorrect signaling cascades (Maves and Schubiger,1998,2003), chromatin modifications by PcG proteins or other factors (Lee et al.,2005; Klebes et al.,2005).
Imaginal disc regeneration is regulated by the JNK pathway, as inactivation of JNK signaling compromises regeneration of cut/injured discs (Bosch et al.,2005). Moreover, cell-lineage experiments using the JNK-responsive gene puc in the wound edge of microsurgically or genetically ablated discs revealed that a substantial number of blastema cells arise from cells in which JNK is activated, which promote proliferation and rebuilding of the lost tissue (Bosch et al.,2005,2008; Smith-Bolton et al.,2009; Bergantiños et al.,2010a; Wu et al.,2010). It has also been observed that the JNK pathway is activated in the LE of the wound at the onset of regeneration, but not in apoptotic cells, and contributes to actin cable and filopodia formation (Fig. 3E) (Bergatiños et al., 2010b). Moreover, it has been shown that in genetically ablated wing imaginal discs, JNK signaling activates the Hippo pathway, which controls growth during normal development (Zhao et al., 2010, for review), resulting in compensatory cell proliferation and regeneration (Sun and Irvine,2011). To identify new genes involved in regeneration, microarray analyses were performed in regenerating wing disc after being manually fragmented and implanted into adult females at several time frames. These assays allowed the identification of early- and late-expressing genes involved in the regeneration process, such as members of the JNK and N pathways and chromatin regulators (Blanco et al.,2010). One of the JNK effectors identified during this process was cbt, also involved in DC downstream of JNK signaling (see above). Additionally, regeneration and transdetermination events can be induced by missexpression of the Wg signaling molecule (Maves and Schubiger,1995; Klebes et al.,2005). Notably, hyperactive Wnt signaling in mammals can induce a similar switch in cell lineage, indicating that is a potent inducer of cell fate changes in many organisms (Okubo and Hogan,2004). Genetic screens to identify the regulators of regenerative growth have been performed by activating the apoptosis program specifically in wing imaginal discs with the UAS/GAL4 system (Smith-Bolton et al.,2009). These analyses revealed that wg is expressed in the cells surrounding the ablated zone and is upregulated during regeneration. Smith-Bolton and co-workers also proposed a model in which Wg promotes growth by N repression, allowing the expression of the dmyc gene and the bantam miRNA (Herranz et al.,2008). Interestingly, wound healing and tissue repair in mammals is also associated with the loss of PcG-mediated silencing, leading to the derepression of wound-induced genes such as dmyc gene (Shaw and Martin,2009). Indeed, some PcG genes are downregulated in the proliferating cells at the wound site in injured discs upon activation of the JNK pathway (Lee et al.,2005). It has been suggested that another possible function of PcG proteins in regeneration and transdetermination is the regulation of cell-cycle genes in blastema cells, which present a distinct cell-cycle profile compared with the surrounding normal disc cells (Ringrose et al.,2003; Valk-Lingbeek et al.,2004; Sustar and Schubiger,2005). Further studies using Drosophila injured imaginal discs as model may help to decipher the mechanisms of reprogramming during regeneration and transdetermination.
Imaginal Disc Spreading, Fusion and Closure: A Model of Epithelial Migration and Fusion
Imaginal disc spreading and fusion is a morphogenetic process that occurs during Drosophila prepupal stages. It involves epithelial sheet movements and fusion, thus sharing morphological and molecular similarities with embryonic DC and with wound healing in vertebrates. Imaginal discs are formed by a monolayer of cells coated with a sheet of squamous peripodial epithelium that folds and adheres to the basal lamina. Just before eversion takes place, the cells detach from the basal lamina; the epithelial cells then columnarize and the accompanying contraction forces the discs to evert through the peripodial stalks. Stalk widening and disc eversion appear to result from microfilament contraction, which leads to dramatic changes in cell shape, rather than from changes in membrane adhesiveness (Martin-Blanco et al.,2000). Then imaginal cells spread over the larval tissues and fuse, similar to DC. However, a clear difference has been observed between DC and imaginal disc spreading and fusion. During embryonic DC, epithelial cells are pulled by the AS, which contracts and invaginates into the yolk sac and disappears by apoptosis. In contrast, during pupariation, imaginal discs crawl over the larval epidermis (see Fig. 6 in Martin-Blanco et al.,2000). During this process, cells are left below and behind and eventually delaminate from the edges (Martin-Blanco et al.,2000). Regarding imaginal disc fusion, one of the best known processes is thorax closure, in which the dorsal sides of both wing discs approach and fuse to midline 6–8 hr post pupariation (Fristrom and Fristrom, 1993; see Fig. 1 in Zeitlinger and Bohmann,1999). Furthermore, the male genital imaginal disc also presents epithelial spreading and fusion events during prepupal stages, a process known as “dorsal closure” (Macías et al.,2004; Rousset et al.,2010). Three groups of cells originally coming from the embryonic A8, A9, and A10 abdominal segments assemble to form the complete larval disc (Chen and Baker,1997; Freeland and Kuhn,1996). During pupariation, the male genital disc undergoes partial eversion, cell spreading and fusion, and subsequently it twists dextrally accomplishing a full 360° rotation. Defects during cell spreading and disc rotation result in externally misrotated male genitalia and are called “left–right” defects (Spéder and Noselli,2007; Suzanne et al.,2010). Recently, live-imaging assays of male pupal discs expressing GFP with the AbdB-GAL4 driver have led to a detailed description of male genitalia DC and rotation (Rousset et al.,2010).
Martín-Blanco et al. (2000) showed that loss of JNK activity alters the adhesive properties of larval cells and leads to the detachment of the imaginal and larval tissues. Moreover, the absence of dpp signaling affects the actin cytoskeleton, blocks the emission of filopodia, and promotes the collapse of the LE in imaginal tissues, defects similar to those occurring during DC (Martín-Blanco et al., 2000). Thus, both cytoskeletal dynamics and the JNK pathway play a central role during imaginal disc fusion (Fig. 3F). Indeed semi-lethal mutants of JNK pathway components present a thoracic cleft phenotype (Zeitlinger et al.,1997; Agnès et al.,1999; Pastor-Pareja et al., 2004). Furthermore, it has been shown that ectopic puc expression in male genitalia (Macías et al.,2004; McEwen and Peifer,2005), as well as mutations in JNK pathway components such as hep, slp, and scaf, affect male genital plate morphogenesis (Holland et al.,1997; Polaski et al.,2006; Rousset et al.,2010), suggesting that JNK signaling is necessary for that process. Recently, Rousset et al. (2010) established that this pathway is required in the A8 segment for correct male imaginal disc DC and rotation. Interestingly, we have found in our laboratory that cbt, another JNK target, appears to also be involved in thorax and male genital disc fusion (Muñoz-Descalzo et al.,2005; Y.B. and N.P., unpublished results). Furthermore, mutant flies for the pvf1/pvr, hid, and MyoD genes also present male genitalia misrotation (Macías et al.,2004; Spéder et al.,2006).
Drosophila Adult Stem Cells: A Model of Homeostatic Regeneration
ASCs have been identified in several Drosophila tissues, including ovaries and testes (Singh and Hou,2008), midgut, hindgut, foregut/midgut junction (Micchelli and Perrimon,2006; Ohlstein and Spradling,2006, 2007; Takashima et al.,2008; Singh et al.,2011) and in the Malpighian tubule system, an excretory organ similar to the kidney (Affolter and Barder, 2007; Singh et al.,2007; Singh and Hou,2008, 2009) (Fig. 2 and data not shown). Stem cells can be identified because they are actively dividing but do not differentiate (Margolis and Spradling,1995). These Drosophila cells present unique advantages for studying the molecular and genetic networks controlling stem cell regulation because they can be distinguished from their neighboring cells and can be easily marked and manipulated, facilitating qualitative and quantitative analyses (Xie and Spradling, 2001; Kiger and Fuller, 2001; Decotta and Spradling,2005). Furthermore, the potent genetic and molecular tools available in Drosophila and the existence of large collections of mutants have made this organism one of the best animal models to study ASC biology. The molecular mechanisms used by ASCs to regulate their division patterns remain poorly understood, but research into this process will increase our understanding of the choice between self-renewal and differentiation. Indeed, the identification of signaling pathways and molecules that regulate ASC homeostasis is critical for developing tools to manipulate stem cells for therapeutic purposes. It is also important to study the microenvironments, or cell niches, where insect and vertebrate ASCs reside. Interestingly, the first experimental evidence of the existence of a stem cell niche was obtained in the Drosophila ovary (Xie and Spradling,2000).
Intestinal Stem Cells (ISCs) are the most thoroughly studied ASCs in Drosophila and are required to maintain normal gut structure and function, likely by responding to variations in cell number and damage (Amcheslavsky et al.,2009; Buchon et al.,2009; Jiang et al.,2009). ISCs self-renew and produce the two main differentiated cell types of the intestinal epithelium: the nutrient absorptive enterocytes (ECs) and secretory enteroendocrine cells (EEs) (Micchelli and Perrimon,2006; Ohlstein and Spradling,2006). The Drosophila midgut has structural and functional similarities with its mammalian counterpart, thus mechanisms that regulate their self-maintenance could be similar (Ohlstein and Spradling,2006; Michelli and Perrimon, 2006). Several signaling pathways are required in the ISC niche. Indeed, Wg/Wnt signaling regulates ISC self-renewal both in Drosophila and mammals, promoting ISC proliferation but not maintenance of their identity (Lin et al.,2008; Lee et al.,2009; van der Flier and Clevers,2009). Thus, reduction of wg function causes ISC quiescence and differentiation, whereas wg overexpression produces ISC overproliferation (Lin et al.,2008). Epistasis analyses suggest that the N pathway acts downstream of the Wg pathway and that this hierarchy controls the balance between ISC self-renewal and differentiation (Lin et al.,2008). N is specifically expressed in the ISCs and is required to produce an appropriate fraction of EE cells (Schonhoff et al.,2004; Micchelli and Perrimon,2006; Ohlstein and Spradling,2006,2007). Of interest, the N cascade also controls the absorptive/secretory fate choice in mammals (van der Flier and Clevers,2009), suggesting that genetic control of intestinal lineage differentiation is evolutionary conserved. Recently, it has been proposed that ISCs maintenance requires inhibition of the Enhancer of split complex (E(spl)-C), which contain N target genes, as well as activation of ISC-specific genes like daughterless (da), which encodes a bHLH transcription factor (Bardin et al.,2010). This model supports the observation that transcriptional repression of differentiation genes may be a central hallmark of stem cells in general (Jepsen et al.,2007; Maines et al.,2007; Dejosez et al.,2008; Liang et al.,2008; Pietersen and van Lohuizen,2008). The JAK-STAT pathway cooperates with Wg, EGFR and N signaling in midgut ISC self-renewal, and it plays a critical role in promoting terminal differentiation from epithelial progenitors, preferentially toward the EE cell fate (Lin et al.,2010; Xu et al.,2011). Finally, it has been shown that EGFR signaling (Biteau et al.,2008; Jiang et al.,2011), the JNK pathway (Karpowicz et al.,2010), JAK-STAT signaling (Jiang et al.,2009), the Hippo pathway (Staley and Irvine,2010; Shaw et al.,2010) and apoptosis (Bergmann and Steller,2010) are induced in response to damage or stress in the Drosophila midgut epithelium, promoting ISC division and midgut epithelium regeneration.
Germinal stem cells (GSCs) have been identified in both testes and ovaries of adult flies (Fig. 2). In males they are located at the tip of the testes, where there is a germinal proliferation center in which 12 quiescent somatic cells form the germ stem cell niche (known as the hub). Anchored around the hub, there are five or nine GSCs and many somatic stem cells (SSCs), called cyst progenitor cells, which maintain spermatogenesis (Fig. 2). The Drosophila ovary harbors three different types of stem cell populations, GSCs, SSCs, and escort stem cells (ESCs), located in the germarium (Fig. 2) and whose activities have been confirmed by lineage tracing and laser ablation experiments (Wieschaus and Szabad,1979; Lin and Spradling,1993; Margolis and Spradling,1995). Expanded N activation causes the formation of more cap cells (CCs) and bigger niches, thus increasing GSC number. In contrast, reduction of N function results in rapid loss of the GSC niche, including CCs and thus GSCs (Song et al.,2007). It has been shown that N signaling is also important for GSC niche formation and maintenance in the Drosophila ovary. Furthermore, JAK/STAT signaling is necessary for the maintenance of GSCs and SSCs (Kiger et al.,2001; Tulina and Matunis,2001; Deccota and Spradling, 2005; Beebe et al.,2010). Different classes of intrinsic factors, translational regulators, chromatin remodeling factors, cell cycle regulators and miRNAs are also required for controlling GSC or SSC self-renewal (Park et al.,2007; Hatfield and Ruohola-Baker,2008; Chan and Ruohola-Baker,2010, for review). It has been recently shown that ecdysone functions in GSC niche formation and the control of female GSCs proliferation through interactions with chromatin remodeling factors (Ables and Drummond-Barbosa,2010; König et al.,2011). These results suggest that steroid hormones and the intrinsic chromatin remodeling machinery could act as potential mechanisms to promote broad transcriptional programs not only involved in ASC self-renewal but also in other processes such as DC or imaginal disc regeneration in Drosophila.
CONTRIBUTIONS OF THE DROSOPHILA MODELS TO WOUND HEALING AND REGENERATION IN VERTEBRATES
In this review, we have shown how studies performed in Drosophila have contributed to a better understanding of wound healing and regeneration in vertebrates (Box 1). Indeed, different morphogenetic processes occurring during Drosophila development, such as DC and imaginal discs eversion and fusion, involve movements and subsequent sealing of epithelial sheets that resemble those occurring during wound healing. Furthermore, epithelial repair after wounding at different Drosophila developmental stages as well as imaginal disc regeneration and cell turnover mediated by ASCs have also been investigated. First, studies of Drosophila DC have led to the development of potent live-imaging techniques that have been subsequently applied to describe in detail cell and tissue movements occurring during wound repair. This has been mainly achieved by the generation of transgenic lines expressing fluorescent proteins allowing embryos, larvae or adult flies to be thoroughly analyzed by time-lapse confocal microscopy. These techniques also allowed analyses of actin-based structures and cell contractions during epithelial resealing, both in DC and in wound closure. Furthermore, thorough genetic analyses have led to the identification of genes and signaling pathways involved in DC. Subsequent studies have confirmed that some of these mechanisms also function in wound healing in Drosophila and vertebrates, demonstrating that the genetic programs underlying epithelial repair are evolutionary conserved between flies and vertebrates. More recent studies based on genetic or laser ablation experiments followed by physical modeling have demonstrated the involvement of different tissues and forces in the DC process. Of interest, some of these tissues or forces have equivalents in vertebrate wound healing. Also imaginal disc fusion and closure during Drosophila metamorphosis has been analyzed, which involves epithelial sheet movements and fusion and shares morphological and molecular similarities with wound healing in vertebrates. In addition, imaginal disc regeneration after injury in Drosophila has been studied as a model of regeneration in vertebrates. The analysis of imaginal disc regeneration has mainly contributed to understand the genetic basis of cell proliferation and reprogramming, two important processes occurring during tissue regeneration. In addition, the recent discovery of ASCs in Drosophila has been relevant to understanding the mechanisms underlying tissue homeostasis. It is important to note that the first in vivo characterization of a cell niche was performed in Drosophila. Moreover, studies in Drosophila have led to the identification of signaling pathways and molecules that regulate ASC differentiation and proliferation in vertebrates, an important step toward their manipulation for therapeutic purposes. Finally, wound healing after injury has also been studied in Drosophila, which has contributed to a better knowledge of wound healing in vertebrates. Although inherent differences do exist between fly and vertebrate wound healing, there is substantial conservation of important cellular mechanisms such as cytoskeletal dynamics and required signaling pathways (Box 2). For all these reasons, all the Drosophila processes discussed in this review can be considered good models for either wound healing or regeneration in vertebrates, and therefore their study is potentially relevant to human health.
In this review, we have provided an overview of how Drosophila has been used as a model organism to study wound healing and regeneration in vertebrates. Although the biomedical applications of Drosophila studies of wound healing/regeneration are limited, flies have contributed to the accumulation of basic knowledge about those processes and the discovery of novel genes and epigenetic mechanisms that could aid in the development of therapies, mainly because the cellular mechanisms and genetic pathways underlying these processes are conserved between flies and vertebrates. However, the development of innovative strategies to promote tissue repair is an important task that requires a more thorough analysis of these processes. In this regard, Drosophila could help in the identification of therapeutic compounds able to stimulate wound healing by performing large-scale compound screens, such as those performed in fly models of human diseases (Chang et al.,2008; García-Lopez et al.,2008). This fast and high-throughput methodology could be adapted to screen chemical compound libraries to identify drugs able to accelerate wound healing in wild type or injured embryos/larvae that could be useful in vertebrates.
It is clear that Drosophila will continue to be a useful tool for analyzing the cellular and molecular aspects of wound healing and regeneration. Indeed, future perspectives and directions for each individual morphogenetic process described here will be and are already being developed. Future works on DC will include new imaging techniques based on 3D cell shape dynamics and quantitative analyses at the subcellular level with nanodissection tools (Cavey et al.,2008) or photoactivatable reporters (Wang et al.,2010). Moreover, many aspects of the DC process are still unknown, such as the nature of the signal that induces formation and contraction of the actin ring, the contribution of other tissues, such as the yolk, or molecular mechanisms like ecdysone signaling and cell reprogramming. Regarding wound healing in Drosophila, the relationship between JNK signaling activation and actin cytoskeletal dynamics that drive epidermal cell migrations across the wound remain unclear. Similarly, the precise machinery and nature of the signal(s) required to initiate and stop wound closure and how activation of the migratory and proliferative machinery of reepithelialization is achieved are unknown. Research designed to answer these questions may help to better understand similar processes in vertebrates. The study of the epigenetic contributions to DC regulation, imaginal disc fusion and repair and stem cell biology could open new perspectives on wound healing and regeneration in vertebrates. In this section, we focus on future prospects of wound healing and regeneration in vertebrates based on the Drosophila morphogenetic events described in this review.
Signals that are sensed by cells surrounding wounds are as yet unknown in Drosophila. Ongoing genetic screens are likely to clarify the relationships between reepithelialization and epidermal barrier-repair pathways as well as identify diffusible signals and receptors that cause the response to epidermal wounding. It would be interesting to perform large-scale genomic and proteomic analyses to search for potential biomarkers of chronic wounds or aberrant wound healing that are able to promote malignant transformation. This will give rise to a better understanding and monitoring of chronic wound progression. Currently, few reports on the functional characterization of proteins involved in wound healing and regeneration in vertebrates are available, and large-scale proteomic analyses have not yet been performed in fly wounds or morphogenetic processes involving epithelial fusions.
Wound healing and cancer progression show several morphological similarities, including angiogenesis and ECM/cellular rearrangements. For that reason, it is not surprising that genetic studies confirm a close relationship between both processes. Chang et al. (2004) discovered the activation of wound healing genes in some human malignant tumors, including prostate, liver, colon and breast cancers. These findings provide a molecular basis for understanding cancer as a deregulation of normal wound healing processes (Dauer et al.,2005). Additional assays are necessary to discover more similarities between these processes and to identify drugs to prevent cancer progression. Moreover, chronic wounds can also result in malignant tumors. More studies are necessary to identify the mechanisms responsible for such transformations, which will allow the development of diagnostic biomarkers. Similarly, it will be interesting to determine the different gene expression profiles that distinguish regular wounds that heal correctly and chronic wounds that can result in cancer. These results could also aid in the design of medical treatments to prevent such malignant transformations.
Currently, biological wound treatments are based on skin substitutes, stem cell therapy, growth factors and gene therapy. Indeed, there are four major research fields whose development could contribute to the healing of complex wounds: (1) use of artificial extracellular matrices for regeneration, (2) use of biological tools such as growth factors and stem cells, (3) analysis of the effects of mechanical forces on wound healing and (4) use of gene therapy. Research in Drosophila could contribute in at least the second and fourth topics by determining the molecular mechanisms of stem cell differentiation and by performing large-scale genetic screens to identify crucial genes involved in chronic wounds or cancer transformation. Of interest, several patents have been filed describing beneficial effects of ecdysteroids in mammalian wound healing (Syrov and Khushbatkova, 1996; Darmograi et al.,1998; Meybeck and Bonté,1990; Detmar et al.,1994). Moreover, 20E shortens the duration of skin repair after wounding and stimulates keratinocyte differentiation in vitro (Detmar et al.,1994). It has also been shown that 20E accelerates the healing process after experimental bone fracture when administered orally to rats (Syrov et al.,1986) and that the same molecule can stimulate in vitro proliferation of rat osteosarcoma (Gao and Wang, 2000). Although numerous reports have been published describing different and contradictory effects of ecdysteroids in mammals, more than 50 companies commercialize ecdysteroid plant extracts with different applications (Lafont and Dinan,2003, for review). In this scenario, it would be of particular interest to understand more precisely their mode(s) of action in mammals. Studies about the role of ecdysteroids in Drosophila DC and homeostatic regeneration could shed light on these mechanisms.
Stem Cell Biology and Homeostasis
Over the past few years, Drosophila testes, ovaries, midgut and Malpighian tubules have emerged as reference systems to study tissue homeostasis due to the identification of ASCs in such organs. Genetic analysis of the Drosophila ISC niche may also provide insights into ISC self-renewal, homeostasis and tumorigenesis in mammals and humans. One of the most exciting discoveries in cancer research is the existence of cancer stem cells and the potential relation between cancer formation and normal ASCs (Clarke et al.,2006). Cancer stem cells can also self-renew, but unlike normal ASCs, they cannot differentiate normally to generate differentiated and functional cells able to integrate into tissues, instead generating abnormally differentiated cells that drive tumor formation and growth. Therefore, understanding the molecular mechanisms and triggering signals governing homeostasis and ASC self-renewal and differentiation is an important aim of developmental biology, regenerative medicine, and has implications for cancer biology. Indeed, the identification of reliable ASCs biomarkers would accelerate our understanding of stem cell roles in tissue homeostasis and cancer, and would help to find new strategies for cancer treatment. These issues can now be investigated in several Drosophila tissues by studying ASC biology in this organism.
We thank Joaquín de Navascués for critical reading of the manuscript and to Salvador Herrero for technical help during artwork preparation. Y.B. was funded by a predoctoral fellowship from the Gobierno de La Rioja and N.P. was funded by grants from the Generalitat Valenciana and Ministerio de Educación y Ciencia.