Mammals, and especially humans, have paid a high price for climbing up the evolutionary ladder. They have lost much of the regenerative power found in lower animals. Lower animals show amazing regenerative abilities and develop three principal strategies to regenerate organs. First, cells that normally do not divide can multiply and grow to replenish lost tissue, as occurs in injured salamander hearts. Second, specialized cells can undergo a process known as dedifferentiation, replicate and later respecialize to reconstruct a missing part. Thirdly, pools of stem cells can step in to perform required renovations. On decapitation, planaria regenerates a new head within five days, using this approach (Davenport, 2004). 

Throughout the course of time we have witnessed many animals, such as tritons and salamanders, with the ability to regenerate their shed or torn tails and broken jaws. Moreover, some animals also exhibit the ability to regenerate their damaged hearts, eye tissues, spinal cords and even skin. The skin of vertebrates serves as a protective barrier against the external world which highlights the need for a fast and efficient repair system. Of note, a temporary repair can be achieved by the formation of a blood clot to serve as a ‘plug’ at the site of the wound. In addition to providing this temporary shield and protection against invading microorganisms, the blood clot also serves as a provisional matrix for invading cells and importantly, as a reservoir of growth factors that are required during the later stages of wound healing. It is well established that within a few hours after injury, inflammatory cells are recruited to invade the wounded area. Neutrophils appear first at the site of inflammation, followed by monocytes then lymphocytes. It is the infiltrating neutrophils that mop-up the area of foreign particles and contaminating bacteria to clean the wound which is proceeded by a process known as phagocytosis, performed by the macrophages. Wounding of skin can cause damage to both epidermal and dermal structures. In order to restore the damaged dermis, fibroblasts invade the wound area to form a contractile granulation tissue. The new stroma has granular appearance owing to massive angiogenic invasion

by a network of capillary blood vessels, which supply the metabolically active wound tissue with nutrients and oxygen. Some of the fibroblasts within this granulation tissue transform into specialist contractile myofibroblasts, which has been speculated to contribute to the wound contractive force. During reestablishment of the epithelial barrier, keratinocytes, originating from outside the wound, migrate over the injured dermis and the granulation tissue (Fig.1).

Figure 1. Scheme of different stages of wound repair in mammals A: 12–24 h after injury the wounded area is filled with a blood clot. B: at days 3–7 after injury, endothelial cells migrate into the clot; they proliferate and form new blood vessels. Fibroblasts migrate into the wound tissue, where they proliferate and form extracellular matrix. The new tissue is called granulation tissue. Keratinocytes proliferate at the wound edge and migrate down the injured dermis and above the provisional matrix. C: 1–2 wk after injury the wound is completely filled with granulation tissue. Fibroblasts have transformed into myofibroblasts, leading to wound contraction and collagen deposition. The wound is completely covered with a neoepidermis. Modified from Werner and Grose, 2003.

At the wound edges, these keratinocytes form the socalled hyperproliferative epithelium, which strongly proliferates and migrates to replenish the wounded area with new tissue. Cells from the hyperproliferative epithelium overtime displace the fibrin clot. The hyperproliferative epithelium is characterized by the expression of keratins 6 and 16, which are normally expressed in the unwounded epidermis (Martin, 1997; Werner and Grose, 2003).

Under certain circumstances, a wound may fail to heal and develop into a chronic wound. Incidences of chronic wounds are higher among the elderly and diabetic, as well as among people with vasculature problems (Harding et al., 2002; Falanga, 2005). The epidermis of a chronic wound has a typical appearance (Fig.2). It is thick and hyperproliferative, with mitotically active cells located in the upper, differentiated layers. Furthermore, the cornified layer is hyperkeratotic (thick cornified layer) and parakeratotic (presence of nuclei in the cornified layer). Keratinocytes on a chronic wound edge are capable of proliferating, but are unable to migrate properly (Morasso and Tomic-Canic, 2005). Particularly, these types of wounds or lifethreatening skin burns may require special treatments of wound mediators to accelerate healing. However, at this point in time there is not yet enough clinical data to support the routine use of such factors.

Figure 2. Chronic wound. Keratinocytes at the edge of the wound (purple) are hyperproliferative (indicated by mitotically active cells present throughout the suprabasal layers), hyperkeratotic (indicated by thick cornified layer) and parakeratotic (indicated by presence of nuclei in the cornified layer). BM=basement membrane.

The process of wound healing involves a complex interaction between epidermal and dermal cells. It is coordinated by many signals that trigger relatively sedentary cell lineages at the wound margin to proliferate, to become invasive, and then to lay down a new matrix. After injury keratinocytes become activated and secrete various cytokines and growth factors and, at the same time, respond to them. Keratinocytes release interleukin-1 (IL-1), which is the first signal upon wounding and has a dual function: to activate keratinocytes and to signal-alert the surrounding tissues. In the last decade in vivo and in vitro studies have provided the identification of a list of growth factors and cytokines that are important during wound repair. For example, plateletderived growth factors (PDGFs), fibroblast growth factors (FGFs), and granulocytemacrophage colony stimulating factor (GMCSF) (Scheid et al., 2000; Werner and Grose, 2003).

The development of genetically modified mouse technologies gives new insights into the role of genes during skin repair processes. For instance, genetic evidence obtained in mice indicates that signaling via the epidermal growth factor (EGF) receptor and the keratinocyte growth factor (KGF/FGF7) receptor are important for reepithelialization (Werner et al., 1994; Repertinger et al., 2004; Shirakata et al., 2005). Furthermore, downregulation of the transforming growth factor β(TGFβ) receptor in keratinocytes reduce the rate of reepithelialization (Amendt et al., 2002). Smad3 is a downstream component of TGFβ signaling and, in contrast, Smad3 mutant mice show an increased rate of reepithelialization and reduced monocyte infiltration during wound healing (Ashcroft et al., 1999). It has also been demonstrated that cJun and STAT3 may signaling downstream of growth factors, such as interleukins and integrins. Specifically a conditional mutation of cJun and STAT3 in the epidermis delayed wound closure (Sano et al., 1999; Li et al., 2003).

The skin is the largest organ in the body, which protects against environmental insults as well as from dehydration. The mammalian skin is composed of several layers, including an underlying dermis, separated by a basement membrane from the epidermis and its appendages, including the hair follicles, sebaceous glands and sweat glands (Fig. 3). The basement membrane is composed of extracellular matrix proteins, such as collagen IV, fibronectin and laminin 5. Both the epidermis and the dermis contribute to the synthesis of basement membrane components (DiPersio et al., 1997; Raghavan et al., 2000). The epidermis is a thin multilayer of stratified squamous epithelium that is mainly comprised of keratinocytes. The undifferentiated basal layer stratifies to give rise to differentiated cell layers of the spinous layer, granular layer, and the outer most stratum corneum. As cells withdraw from the basal layer, they stop dividing and induce a programme of terminal differentiation that will ultimately allow them to function as barrier of the skin. The epidermis originates from the outer layer of the embryo, the surface ectoderm. BMPs activate the epidermal differentiation program and induce the expression of keratin proteins via several known transcription factors (Meulemans and Bronner-Fraser, 2004; Byrne et al., 1994). The surface ectoderm proliferates and migrates from the dorsal midline to cover the embryo, and persists as a simple epithelium until approximately embryonic day 9.5 of mouse embryogenesis. At this stage basal cells begin to express keratin 5 and 14, presaging epidermal stratification, which requires the activity of a key epidermal transcription factor that also regulates epidermal fate, proliferation, and adhesion (Yang et al., 1999; Bakkers et al., 2002; Koster et al., 2004; Lechler and Fuchs, 2005). By the birth, the epidermis consists of a proliferative basal layer that differentiates to form outer layers.

To date there is considerable amount of data on the profiles of structural gene expression in the epidermis and its appendages, however much less is known about how these are established during development and what programmes are orchestrated to terminate differentiation at the transcriptional level (Fuchs and Raghavan, 2002). In normal conditions, growth and proliferation are precisely balanced and regulated processes in the epidermis. Tyrosine kinase receptors and their ligands have important role in regulating this balance. For instance, the dermal fibroblasts secrete GM-CSF and FGF7 to promote the proliferation and differentiation of overlying epidermal keratinocytes (Szabowski et al., 2000). Keratinocytes themselves are a source of autocrine growth factors that stimulate tyrosine kinase receptors, such as transforming growth factor α (TGF α), a ligand for EGFR (Luetteke et al., 1994; Sibilia and Wagner, 1995). EGFR and its downstream Ras–MAPK pathway have been implicated in epithelial proliferation (Hansen et al., 2000). Tyrosine kinase receptors can activate phosphoinositol 3 kinase (PI3K) and the Akt cell-survival pathway, which both control epidermal homeostasis. Aberrations in these signaling pathways may result in hyperproliferative disorders of the skin, such a psorasis and basal or squamouscell carcinomas.

The dermis of the skin consists mostly of loosely packed fibroblasts. Mature dermis is also composed of collagen, elastin fibers and interfibrillar glycosaminoglycans (GAG)/proteoglycan gel. The dermis has a remarkable variety of embryonic origins, i.e. the ventral dermis arises from the lateral plate mesoderm, while a part of the head dermis arises from neural crest (Candille et al., 2004; Fernandes et al., 2004). However, fate mapping of dorsal dermis in mammalian embryos has not yet been described (Millar, 2005).

Figure 3. Mammalian skin and its appendages. Skin consists of the epidermis and dermis, separated by a basement membrane (BL). The epidermis is composed of is the basal layer (BL), differentiated spinous layer (SL), granular layer (GL) and the stratum corneum (SC). Also shown is a cross-section of a hair follicle, which consists of an outer root sheath that is contiguous with the basal epidermal layer, the hair bulb, made from proliferating matrix cells, and the bulge, which is part of the outer root sheath and is where epidermal stem cells reside. Modified from Fuchs and Raghavan, 2002.

The Met receptor tyrosine kinase was first identified as an activated oncogene (Park et al., 1986). Subsequently, the cDNA of this protooncogene was isolated and found to encode a transmembrane receptor tyrosine kinase. The Met receptor is synthesized as a single polypeptide chain of 1436 amino acids that undergoes intracellular proteolytic cleavage into a twochain heterodimer. This encompasses an Nterminal α chain located outside the membrane, and a C-terminal β chain that contains an extramembrane sequence, a single transmembrane domain and a cytoplasmic protein kinase domain.

Met belongs to a family of receptors that also includes mammalian Ron and the avian Sea receptors. They share the heterodimeric structural motif of an extracellular α chain and transmembrane β chain harboring the tyrosine kinase activity. The similarities are observed not only among receptors but also between ligands since Macrophage-stimulating protein (MSP), the ligand for Ron, resembles the Met receptor ligand, HGF/SF, in many aspects (Leonard and Danilkovitch, 2000). In evolutionary terms, Met is a young molecule, which appear during evolution of vertebrates for the first time (Birchmeier et al., 2003).

The ligand for Met was first identified as a factor that induces proliferation of hepatocytes and was subsequently named hepatocyte growth factor (HGF) (Miyazawa et al., 1989; Nakamura et al., 1989; Zarnegar and Michalopoulos, 1989). The activity of HGF was observed in pairs of rats with a surgically connected circulation system, which one rat had an injured liver. In the blood stream of these animals, the presence of circulating factors was having a dramatic affect on the growth of both damaged and normal liver. As a result, one of these circulating growth factors was purified form the media of primary cultured rat hepatocytes, and identified as a novel, very potent mitogen, named as HGF (Matsumoto and Nakamura, 1993). HGF was later shown to be identical to scatter factor, SF, discovered independently due to its ability to induce motility of epithelial cells (Stoker et al., 1987). The identity between HGF and SF was demonstrated by amino acid sequencing, by immunological methods, by comparison of the biological activities (Gherardi and Stoker, 1990; Weidner et al., 1990; Furlong et al., 1991), by cDNA cloning, and by receptor binding studies (Weidner et al., 1991; Naldini et al., 1991). HGF/SF is a unique growth factor that elicits multiple cellular responses including mitogenesis, cell motility and morphogenesis. The structure of HGF/SF contains a domain that resembles that of plasminogen as well as other complex serine proteinases involved in blood coagulation and fibrinolysis in vertebrate organisms (Donate et al., 1994). They represent a family with related biological activities, termed plasminogen-related growth factors.

Binding of active HGF/SF to Met results in phosphorylation of the receptor. These phosphorylation events lead to the activation of the receptor and create recruitment sites for many signaling mediators. Two tyrosine residues (Tyr1349 and Tyr1356, Fig.1), together with a short aminoacid sequence motif located near the Cterminus of the intracellular domain, constitute a multidocking site (Ponzetto et al., 1994). Studies using chimeric receptors containing the extracellular and membrane domains of other receptor fused to the intracellular portion of Met demonstrate that this docking site is both necessary and sufficient to mediate Met signal transduction and biological function (Weidner et al., 1993). Recruitment of adaptor proteins and signaling molecules to the docking site of Met enables amplification of the signal, the activation of multiple downstream signaling pathways and thus induces various cellular responses (Fig.5). Specificity at the receptor level is achieved by binding of various cytoplasmic signaling proteins to phosphotyrosines and surrounding amino acid residues of the activated receptor (Fig.4). These proteins that are recruited to activated Met include the adapter proteins Gab1, SHC, Grb2 and Crk/CRKL, along with other signal transducers, like phosphoinositol3 kinase, PI3K and Shp2 (Pelicci et al., 1995; Ponzetto et al., 1994; Garcia-Guzman et al., 1999; Graziani et al., 1991; Weidner et al., 1996; Fixman et al., 1996).

Figure 4. Docking sites of Met. Shown are the phosphotyrosines binding sites of Met, as well as their direct interaction partners. α and β refer to the subunits of the receptor. Gab1, growth-factor-receptor-bound protein 2 (Grb2)-associated binder 1; HGF/SF, hepatocyte growth factor/scatter factor; PI3K, phosphatidylinositol 3-kinase; PLCγ, phospholipase Cγ; Shc, Src-homology-2 (SH2)-domain-containing; Shp2, SH2-domain containing protein tyrosine phosphatase 2; Grb2 and Grb10, growth-factor-receptor bound-protein 2 and 10; Ship, SH2-domain-containing inositol-5-phosphatase; p85 refers to the regulatory subunit of PI3K. Modified from Birchmeier et al., 2003.

Most of these proteins recognize specific tyrosine-phosphorylated regions of a receptor via their src-homology region 2 (SH2) domains. Thereby, this domain has a key role in relaying cascades of signal transduction (Koch et al., 1991). In contrast, the docking protein Gab1 can be recruited to the receptor directly via a unique 13amino acids sequence, the Metbinding site (Schaeper et al., 2000). In addition to the direct association of Gab1 to Met, binding may be also enhanced indirectly by coupling Gab1 to Met via Grb2. Phosphorylated Gab1 binds several downstream signaling molecules, like PI3K, phospholipase Cγ, PLCγ, the phosphatase Shp2, and the adaptor proteins Crk/CRKL (for recent review see Birchmeier et al., 2003). In vitro studies have shown, that binding of Shp2 by Gab1 is important for HGF/SF and Met dependent branching morphogenesis of epithelial cells (Schaeper et al., 2000; Maroun et al., 2000). Recruitment of Shp2 is critical for activation of the ERK/MAPK pathway, which plays an important role in cell proliferation, differentiation and migration (Fig.5). Met activates also other signaling branches that regulate cell motility and invasion by the phosphorylation and activation of paxillin and focal adhesion kinase (Liu et al., 2002). The driving forces for cell motility and polarity are derived from the cytoskeletal reorganization of actin, which is controled by Cdc42, Rac and Rho small GTPases (Ridley, 2001). Cdc42 promotes filopodia and microspike formation while Rac induces lamellipodia and membrane ruffling. Several effectors for Cdc42 and/or Rac have also been found to be involved in HGF/SF induced cellcell dissociation and migration, such as Cdc42/Rac-regulated p21activated kinase (PAK) (Royal et al., 2000). Moreover Met can also contribute to cell survival via activation of the PI3K/Akt pathway (Fan et al., 2001). In addition, other molecules such as βcatenin, integrins, and cjun amino terminal kinase (JNK) have been reported to participate in HGF/SF/Met signaling (Monga et al., 2002; Muller et al., 2002; Chiu et al., 2002; Lamorte et al., 2000).

Figure 5. Signaling by the receptor tyrosine kinase, Met. Upon binding of HGF/SF, Met recruits various adapter proteins like Gab1 and Grb2 and activate Shp2, Ras, Erk and PI3K pathway. These pathways regulate cell adhesion, cytoskeleton, motility, cell cycle and apoptosis (Birchmeier et al., 2003).

Genetic and cell biological evidence have shown that Gab1 is the most crucial substrate for Met signaling (Maroun et al., 2000; Sachs et al., 2000). Targeted mutagenesis of Gab1 in mice revealed similar phenotypes of Gab1 and Met null mutants, which proofs the essential role of Gab1 in Met signaling. Gab1 null mutation embryos die between embryonic day 13.5 and 17.5 and are characterized by a placenta defect, a small liver and lack of muscle of limbs and diaphragm (Sachs et al., 2000). Furthermore, Gab1 null mutation mice display skin and heart defects, which could be related to defective signaling downstream of EGFR (epidermal growth factor receptor), PDGFR (platelet derived growth factor receptor) or gp130 signaling cascades (Itoh et al., 2000; Cai et al., 2002), since Gab1 is also adaptor protein for these receptors. Gab1 contains a pleckstrin homology domain (PH) at its Nterminus, a unique Met receptor tyrosine kinase binding site (MBS) and two Grb2binding sites. The central region of Gab1 protein is rich in prolines and contains multiple tyrosine residues, which when are phosphorylated, they bind the SH2 domains of many downstream signaling proteins. Pleckstrin homology domains can recognize membrane components, and therefore contribute to the membrane targeting of Gab1 (Fig.6). Interaction of Gab1 with the adaptor protein Grb2 is important for coupling Gab1 to activated EGFR (Lock et al., 2000). Gab1 binds constitutively the SH3 domains of Grb2 via proline-rich sequences, while phosphotyrosine residues of activated EGFR bind the SH2 domain of Grb2. EGFR mutants deficient in binding Grb2, are unable to recruit and activate Gab1 (Rodrigues et al., 2000). Binding of Grb2 is also important for coupling Gab1 to other tyrosine kinases, like the FGFRs (fibroblast growth factor receptors) (Ong et al., 2001; Lamothe et al., 2004). In the FGF receptor pathway, Gab1 phosphorylation occurs via an additional scaffolding adaptor, FRS2. Upon receptor activation, FRS2 becomes tyrosine phosphrylated and binds to GRB2, which in turn recruits Gab1 (Hadari et al., 2001). Many extracellular stimuli like insulin, IL3 (interleukin 3), IL6, Epo1 (Erythropoetin1) as well as Bcell receptor activation induce Gab1 phosphorylation and association with Shc, PI3K, PLCγ and Shp2 (Lecoq-Lafon et al., 1999; Ingham et al., 1998). Gab1 binding with PI3K has been shown to be important for prevention of apoptosis in response to NGF (nerve growth factor) stimulation (Holgado-Madruga et al., 1997).

Figure 6. Gab family proteins. Schematic domain structures of mammalian Gab1 proteins, zebrafish Gab1 (zfGab1) and their invertebrate orthologs in Drosophila (DOS) are shown. All Gab family proteins consist of N-terminal pleckstrin-homology (PH) domain, prolinerich domains (P) and multiple tyrosines. The unique Met-binding site (MBS) that allows for direct interaction with Met is also indicated.

The Gab1 docking proteins are evolutionary conserved from worms to mammals and homologues have been identified in vertebrates, like zebrafish (zfGab1), but also invertebrates, like Drosophila (Dos, Daughter of Sevenless) and Caenorabditis elegans (Soc1, Suppressor of Clear, Fig.6) (Gu and Neel, 2003).

The importance of HGF/SF and Met signaling system in development has been assessed by genetic analyses in the mouse. Mice that carry a null mutation of either HGF/SF or Met die in uterus between embryonic day 12.5 and 16.5 due to placenta defect. The placental labyrinth layer formed by epithelial trophoblast is significantly reduced, which leads to an impairment of oxygen exchange and nutrients between the maternal and embryonic bloodstream (Schmidt et al., 1995; Uehara et al., 1995; Bladt et al., 1995). Ablation of Met or HGF/SF results in complete absence of the muscle groups in the mouse embryo that derive from migrating precursor cells, whereas other muscle groups form normally. The migrating progenitors delaminate by epithelialmesenchymal transition from the dermomyotome and migrate to the limb, tongue and diaphragm, where they differentiate into skeletal muscle. In HGF/SF and Met mutant mice, these migrating cells do not detach and do not emigrate from the dermomyotome (Fig.7).

Figure 7. Migrating muscle precursor cells in HGF/SF/ and Met/ embryo. During normal development muscle precursor cells (red) delaminate from the epithelial dermomyotome (Dm, blue) and migrate to the limb bud, where they differentiate into myoblasts. In HGF/SF–/– or Met–/– embryos, the progenitor cells are not released, but remain in the dermomyotome. HGF/SF, hepatocyte growth factor/scatter factor; My, myotome; Sc, sclerotome. Modified from Birchmeier and Brohmann, 2000.

During delamination and migration of muscle precursor cells, HGF/SF is expressed in a highly dynamic pattern, first in the mesenchyme close to the epithelial dermomyotome, and then along the ways and at the targets. HGF/SF and Met null mutation mice are also characterized by small livers due to decreased proliferation and increased apoptosis of hepatocytes.

The genetic analyses of HGF/SF and Met in mice revealed that the phenotypes are identical, demonstrating that HGF/SF is the only Met ligand and Met is the only functional receptor for HGF/SF in vivo. HGF/SF and Met arose late in evolution. This is consistent with the function carried out by this signaling system in the embryo, related to processes such as placentation and liver development, which arose also late in evolution.

The Met tyrosine kinase receptor and its ligand, HGF/SF are expressed not only during embryogenesis, but also in the adult. Several experimental approaches have shown that deregulation of this pathway is implicated in many human malignancies. Transgenic mice that overexpress Met or HGF/SF have been shown to develop different types of tumors. Furthermore, the receptor or the ligand are frequently expressed in human carcinomas and other types of solid tumors, as well as in their metastases. Overexpression of Met and/or HGF/SF often correlates with poor prognosis, for example, Met-activating mutations have been found in human sporadic and inherited renal papillary carcinomas (Rong et al., 1994; Takayama et al., 1997; Abounader et al., 2002; Danilkovitch-Miagkova and Zbar, 2002). It is highly probable that HGF/SF and Met signaling might participate at different stages of tumor progression, since it is implicated during several stages of tumourigenesis such as proliferation, invasion, angiogenesis and antiapoptosis.

Under normal physiological conditions HGF/SF acts as a paracrine factor, i.e. mesenchymal cells produce HGF, which acts on epithelial and other cells. Met is predominantly expressed in the cells derived from the epithelial or endothelial origins (Birchmeier and Gherardi, 1998). In pathological situations, such as cancer, activation of Met occurs most often through autocrine, although it is possible that Met can act through paracrine mechanisms. For instance, osteosarcomas and glioblastoma multiforme express both Met and HGF/SF (Birchmeier et al., 2003). HGF/SF and Met have also been implicated in various physiological processes in the adult. For instance, during liver regeneration, HGF/SF levels in the blood stream raise, and conditional mutagenesis in mice has shown that Met signals are essential during liver regeneration and repair (Michalopoulos and DeFrances, 1997; Taub, 2004; Borowiak et al., 2004; Huh et al., 2004). Moreover, upregulated HGF/SF and Met expression is observed after tissue injury, for instance in the lung, kidney and heart (Ohmichi et al., 1996; Kawaida et al., 1994; Nakamura et al., 2000). Interestingly, application of exogenous HGF/SF to a skin wound promoted the formation of new tissue and fast healing (Bevan et al., 2004). Thus, it suggests that upregulated HGF/SF expression might be part of a general, defensive response to tissue injury.

It has been demonstrated that HGF/SF/Met signaling has a potential role during regeneration and tissue remodelling. These data were achieved through studies carried out by biochemical and cell culture experiments in vitro, and by genetic liver regeneration studies. HGF/SF/Met signaling has been implicated in skin regeneration, since HGF/SF levels increase rapidly after skin injury in the serum. However, the role of endogenous HGF/SF and its receptor, Met, in skin development and wound healing has as yet not been elucidated. The targeted mutation of HGF/SF or Met causes embryonic death due to defect in placental development, which had precluded a genetic analysis of Met or HGF/SF function in the adult. To overcome embryonic lethality, the creloxP technology was here employed to clarify the role of Met specifically in the adult skin. Keratin14cre knockin mice were used to introduce a Met null mutation into keratinocytes. The present study was aimed to understand further the role of Met in the skin and during skin wound healing. For the first time, a genetic method for the analysis of the Met gene function in the skin has been developed.