To address the functions of Rac1 in keratinocytes of the basal epidermal layer and in the outer root sheath of hair follicles, we generated transgenic mice expressing a dominant inhibitory mutant of Rac, N17Rac1, under the control of the keratin 14 promoter. These mice do not exhibit an overt skin phenotype but show protracted skin wound re-epithelialization. Investigation into the underlying mechanisms revealed that in vivo both proliferation of wound-edge keratinocytes and centripetal migration of the neo-epidermis were impaired. Similar results were obtained in mice with an epidermis-specific deletion of Rac1. Primary epidermal keratinocytes that expressed the N17Rac1 transgene were less proliferative than control cells and showed reduced ERK1/2 phosphorylation upon growth factor stimulation. Adhesion, spreading, random migration and closure of scratch wounds in vitro were significantly inhibited on collagen I and, to a lesser extent, on fibronectin. Stroboscopic analysis of cell dynamics (SACED) of N17Rac1 transgenic and control keratinocytes identified decreased lamella-protrusion persistence in connection with increased ruffle frequency as a probable mechanism for the observed impairment of keratinocyte adhesion and migration. We conclude that Rac1 is functionally required for normal epidermal wound healing and, in this context, exerts a dual function – namely the regulation of keratinocyte proliferation and migration.
Wound re-epithelialization relies on two basic functions of epidermal keratinocytes: first, the ability to adapt their proliferative potential to the needs of fast tissue expansion and, second, to cover a denuded tissue surface by migration. It is generally believed that in skin wounds these functions are regulated by the extracellular wound milieu, i.e. the presence of growth factors, cytokines and extracellular matrix molecules released on the spot or transported to the wound site by components of the blood clot or by inflammatory cells. Although a number of growth factors and cytokines have been shown to contribute to wound re-epithelialization in skin in vivo (for a review, see Werner and Grose, 2003), the signaling mechanisms by which these factors regulate this function remain poorly defined. One of the few intracellular signaling molecules for which a non-redundant function in wound re-epithelialization has been established is signal transducer and activator of transcription 3 (STAT3) (Sano et al., 1999). STAT3 is directly regulated by the Rho family GTP binding protein Rac1 (Simon et al., 2000).
The small GTPase Rac1 is expressed ubiquitously and exerts important functions within the epidermis. Although in vitro studies have suggested an essential function for Rac in the establishment of cell-cell contacts (Betson et al., 2002; Braga et al., 2000; Braga et al., 1997; Mertens et al., 2005), targeted deletion of Rac1 in epidermal keratinocytes does not necessarily lead to their disruption (Chrostek et al., 2006). Inducible elimination of the Rac1 gene in the epidermis of adult mice has been reported to result in loss of epidermal stem cells (Benitah et al., 2005). Mice with constitutive epidermis-specific deletion of Rac1 show hair-follicle disruption but no detectable changes of the interfollicular epidermis (Chrostek et al., 2006).
In Drosophila melanogaster development the dorsal closure of the epidermis, which partly resembles wound healing in mammals, depends on the presence of at least one of the three Drosophila genes encoding Rac (Rac1, Rac2, Mtl) and on Rac activity (Harden et al., 1995; Harden et al., 1999; Woolner et al., 2005). In the absence of all three Rac genes dorsal closure is severely impaired (Hakeda-Suzuki et al., 2002). Studies in rat embryonic fibroblasts have suggested an essential function for Rac in the process of wound closure by fibroblasts in vitro (Nobes and Hall, 1999). A similar requirement for Rac activity in the closure of a wounded epithelial monolayer has been shown for Madin-Darby canine kidney (MDCK) cells (Fenteany et al., 2000). In these studies, inhibition of Rac function resulted in the failure of wound-edge cells to form lamellipodia and migrate, leading to impaired wound closure.
Although these results suggest an important function for Rac in epidermal wound healing in mammals, this issue has not been addressed so far in vivo. We therefore analysed the wound healing phenotypes of two different in vivo models of epidermal Rac impairment, mice expressing a dominant-negative version of Rac1 as a transgene in epidermal keratinocytes and mice with epidermis-specific deletion of Rac1.
To investigate functions of Rac in the epidermis in vivo we inhibited its activation in basal epidermal keratinocytes. This was achieved by expressing a dominant-negative version of Rac1, N17Rac1, as a transgene under the control of the keratin 14 promoter (Hafner et al., 2004) in mice (Fig. 1a). This mutant has been shown previously to efficiently inhibit Rac-mediated cellular responses (Kjoller and Hall, 2001). Three independent transgenic founder lines were generated. Polymerase chain reaction and Southern blot analysis with genomic DNA from mice of the F1 generation showed stable integration of the transgene (Fig. 1b,c). Semi-quantitative analysis using a 600-bp fragment of the IL-1 receptor as reference showed a low transgene copy number (approx. 1-3 copies) for each of the three founder lines (data not shown). In situ expression of the N17Rac1 transgene as analysed by immunostaining with an antibody against the Myc-tag was, as expected, exclusively detectable in the basal epidermal layer and the outer root sheath of transgenic animals (Fig. 1d). Western blot analysis of extracts from primary adult and newborn mouse keratinocytes with monoclonal antibodies against the Myc-tag and against full-length Rac1 demonstrated correct expression of the transgene in cells isolated from N17Rac1 transgenic mice but not from wild-type littermates (Fig. 1e,f). Considerable variability with respect to expression levels of N17Rac1 protein was observed between primary keratinocytes isolated from different transgenic mice (Fig. 1g). In order to investigate whether these different expression levels had any functional consequences, we determined the proliferative potential of keratinocytes with strong versus weak N17Rac1 expression (Fig. 1h,i). Analysis of cell number per colony and colony area demonstrated that expression of N17Rac1 significantly decreases colony-forming efficiency (CFE) of keratinocytes in culture and, furthermore, suggests that inhibition of CFE is a function of the expression level of the transgene.
For further in vitro experiments we used keratinocytes with strong expression of N17Rac1. Transgenic mice expressing the N17Rac1 transgene were indistinguishable from wild-type littermates. There were no overt signs of skin or hair abnormalities. The mice had a normal life span and showed no signs of stem cell depletion or hair follicle disintegration – even in mice older than 12 months. Comparative histopathological and immunohistological analysis of skin samples from transgenic and wild-type mice showed no detectable phenotypic changes. Expression of keratin 14, keratin 10, loricrin, filaggrin, the adhesion molecules E-cadherin, α1 integrin and α2 integrin and also the deposition of laminin 5 and collagen IV in the basement-membrane zone appeared normal (data not shown).
Delayed wound re-epithelialization in N17Rac1 transgenic mice
In vitro studies and analyses of Drosophila mutants have suggested an important role of Rac isoforms, particularly Rac1, in epithelial closure during development (Farooqui and Fenteany, 2005; Fenteany et al., 2000; Hakeda-Suzuki et al., 2002; Harden et al., 1995; Harden et al., 1999; Woolner et al., 2005), a process that closely resembles epithelial wound healing in vertebrates (Jacinto et al., 2001). To investigate the effects of RAC inhibition in the context of regenerating epidermis we carried out wound healing experiments in N17Rac1 transgenic mice and wild-type controls. The back skin of 17 female transgenic mice and 15 female wild-type mice was lifted and the skin fold was punched with a 4-mm biopsy punch, thus creating two wounds per mouse. The degree of re-epithelialization was analysed after 2 days (seven transgenics, seven controls) and after 5 days (ten transgenics, eight controls). Wounds were fixed in 4% paraformaldehyde, cut in half and embedded in paraffin. Starting from the intersecting plane, 4-μm sections were cut. The first complete section from each half was used for morphometric analysis after hematoxylin and eosin (H-E) staining. For quantification of wound closure we used two parameters: (1) the length of the epithelial tongue starting from the wound edge to the tip of the tongue (denominated `length') as readout for keratinocyte migratory activity (red line in Fig. 2a) and, (2) the area covered by the neo-epidermis (denominated `area') as readout for its volume (green area in Fig. 2a). At day 2 after wounding, cohesive formations of epidermal keratinocytes had started to migrate over the wound bed in wounds of wild-type controls and were visible as epithelial tongues in cross sections of such wounds (Fig. 2b). By contrast, cross sections through wounds of N17Rac1 transgenic mice showed no, or very short, epithelial tongues (Fig. 2c). By day 5 after wounding epithelial tongues had elongated both in wild-type controls and in N17Rac1 transgenics. Epithelial tongues in wounds of transgenic mice were, however, consistently shorter than those in wounds of controls (Fig. 2d,e). Quantification of `length' demonstrated a dramatic delay of wound re-epithelialization in the transgenic mice at days 2 and 5 after wounding (Fig. 2f).
Upon histopathological examination, we noticed another difference between N17Rac1 transgenic and wild-type mice: at day 2 the wounds already exhibited epidermal hyperplasia, which was strongly increased by day 5 (Fig. 2b,d). This was markedly less pronounced in wounds of transgenic mice: at day 2 after wounding no hyperplasia of the wound-edge epithelium was seen. At day 5 after wounding hyperplastic wound-edge epithelium and neo-epidermis were present in both N17Rac1 transgenic mice and in wild-type controls; epithelial hyperplasia was, however, much less pronounced in transgenic mice (Fig. 2e). Quantification of `area' revealed a dramatic decrease in the efficiency of neo-epidermis formation of N17Rac1 transgenic mice compared with controls (Fig. 2g). From these experiments we conclude that wound re-epithelialization is profoundly disturbed upon inhibition of Rac activity in basal epidermal keratinocytes.
Delayed wound re-epithelialization in mice with epidermis specific deletion of Rac1
To confirm that inhibition of wound re-epithelialization in N17Rac1 transgenic mice was not due to unspecific inhibitory effects of N17Rac1 on other Rho family GTPases, or signaling molecules other than Rac (Vidali et al., 2006), we carried out wound healing experiments in mice with an epidermis-specific deletion of Rac1 (Rac1E-KO mice) and in mice with epidermis-specific deletion of only one Rac1 allele (Rac1E-HET mice) as controls. Experimental conditions were identical to those used in N17Rac1 mice. Rac1E-KO mice and Rac1E-HET mice were generated by means of conditional gene targeting, using the Cre/lox system with a Cre recombinase transgene driven by the keratin 5 promoter as described elsewhere (Chrostek et al., 2006). These mice are viable although Rac1E-KO mice suffer from progressive hair loss starting 1 week after birth, whereas Rac1E-HET mice are phenotypically normal.
Analysis of wound re-epithelialization in eight Rac1E-KO mice and six Rac1E-HET mice (mixed sex groups) 5 days after wounding showed dramatic shortening of epithelial tongues, and reduced hyperplasia of wound-edge epithelium and neo-epidermis in Rac1E-KO mice compared with controls. The decrease of `length'and `area' in Rac1E-KO mice compared with Rac1E-HET mice was even more pronounced than in N17Rac1 transgenic mice compared with wild-type controls. We conclude that inhibition of Rac in basal epidermal keratinocytes, either by expression of a dominant-negative mutant or through the targeted deletion of Rac1 from the epidermis leads to severe disturbances in wound re-epithelialization.
Reduced keratinocyte proliferation in vivo and in vitro upon inhibition of Rac activity or targeted deletion of Rac1
Reduced hyperplasia of wound-edge epithelium and neo-epidermis in N17Rac1 transgenic mice and Rac1E-KO mice suggested that the growth rate of mutant epidermal keratinocytes was decreased in both. We therefore quantified cell proliferation by analysing incorporation of bromodeoxyuridine (BrdU) into keratinocytes of the wound edge and the neo-epidermis. Per mouse, 3.2 μg BrdU were injected exactly 2 hours before the mice were killed. BrdU-positive nuclei within the epidermis were counted following immunostaining of sections of paraffin-embedded samples of wound tissue. Analysis of wound sections from 14 female N17Rac1 transgenic mice, 11 Rac1E-KO mice (mixed sex groups), eight Rac1E-HET mice (mixed sex groups) and 12 female wild-type controls revealed decreased BrdU incorporation into the wound epithelium of N17Rac1 transgenic mice and Rac1E-KO mice compared with the respective controls (Fig. 4a-c).
To investigate whether reduced keratinocyte proliferation and ERK phosphorylation in wound epithelium of N17Rac1 transgenic mice was a cell-autonomous phenomenon or due to an altered wound environment, we isolated primary keratinocytes from N17Rac1 transgenic mice and controls and analysed BrdU incorporation after stimulation with epidermal growth factor (EGF) and insulin-like growth factor 1 (IGF1) in vitro. Analysis of three transgenic cell lines and three wild-type lines stimulated with either EGF or IGF1 revealed that incorporation of BrdU was consistently decreased in keratinocytes of N17Rac1 transgenic mice (Fig. 4e), indicating that inhibition of cell proliferation upon RAC inhibition is a cell autonomous phenomenon.
Since Rac is able to regulate STAT3 and ERK activity (Frost et al., 1997; Simon et al., 2000) and these signaling molecules have been implicated in the regulation of keratinocyte proliferation and wound re-epithelialization (Haase et al., 2003; Haase et al., 2001; Li et al., 2004b; Sano et al., 1999), we carried out immunostainings for phosphorylated STAT3 and ERK1/2 proteins to determine their activation status in wounded skin of Rac1E-KO and Rac1E-HET mice. Epidermis of Rac1E-KO mice at wound edges still showed phosphorylated STAT3, although the number of STAT3-positive keratinocytes seemed to be reduced (data not shown). This indicates that Rac1 is not required for the activation of STAT3 signaling in skin wound healing. Immunostaining with a phosphorylation specific ERK1/2 antibody indicated decreased ERK phosphorylation in basal epidermal keratinocytes of skin wounds of Rac1E-KO mice (Fig. 4e, upper panel). We, therefore, quantified ERK phosphorylation in situ by counting 20 basal keratinocytes from the wound margin towards the tip of the neo-epidermis in wound sections of five different Rac1E-KO mice and three different Rac1E-HET mice. In Rac1-deficient epidermis we found 13±4.4 and in Rac1-expressing epidermis we found 19±1 nuclei staining positive for phosphorylated ERK. This shows that Rac1 deficiency in epidermal keratinocytes leads to reduced ERK phosphorylation.
To compare ERK phosphorylation between wild-type and transgenic keratinocytes in vitro, adherent keratinocytes were stimulated with EGF (Fig. 4f) or, alternatively, suspended keratinocytes were re-plated onto collagen- and fibronectin-coated dishes (Fig. 4e, lower panel) and subjected to western blot analysis using a phosphorylation-specific ERK1/2 antibody. Consistently, this analysis revealed weaker ERK phosphorylation in transgenic keratinocytes. Representative results of one out of four experiments (Fig. 4f) and one out of three experiments (Fig. 4e) with three different transgenic and three different wild-type lines are shown. We conclude that inhibition of RAC activity in basal epidermal keratinocytes results in an intrinsic defect of ERK phosphorylation and cell proliferation.
Reduced adhesion, spreading and lamella-protrusion persistence in N17Rac1 transgenic keratinocytes
Keratinocyte migration is achieved through temporally and spatially controlled cell attachment, membrane protrusion, cell contraction and cell detachment (Li et al., 2004a). Rac activity is important for these cellular functions. We therefore analysed attachment, cell spreading and lamella protrusion as individual functions of the migration process. For analysis of cell adhesion we used two substrates that are abundant in skin wounds: collagen I and fibronectin. To test whether Rac activation was dependent on the substratum, we carried out Rac pull-down experiments. We used a biotinylated peptide containing the Rac-interacting (CRIB) motif of the effector kinase PAK (Price et al., 2003). GTP bound Rac1 interacts specificially with this peptide and can be co-precipitated. The results of the pull-down experiments show that Rac activation is comparable in keratinocytes plated onto dishes coated with collagen I or fibronectin (Fig. 5a). Adhesion of N17Rac1 transgenic keratinocytes to collagen I was consistently less efficient compared with controls (Fig. 5b). In addition, N17Rac1 transgenic keratinocytes showed reduced spreading on collagen I (Fig. 5c,d). Cell spreading was less pronounced on fibronectin, and inhibition of attachment and spreading were not detectable when keratinocytes were plated onto fibronectin-coated surfaces (Fig. 5b,e). This suggests that Rac activity is required for attachment and spreading of keratinocytes on collagen I. To investigate whether reduced adhesion and spreading on collagen were due to reduced expression of matrix receptors on the keratinocyte surface, we carried out fluorescence-activated cell scanning (FACS) analysis of the surface levels of α2 integrin, α5 integrin and β1 integrin. We analysed primary keratinocytes isolated from seven transgenic and four control mice. There were no consistent changes in the surface levels of either integrin subunit (data not shown), suggesting that inhibited attachment and spreading of N17Rac1 keratinocytes on collagen I was not the consequence of reduced integrin expression.
Cell spreading is the net result of extension and retraction of lamellipodia. We have measured the persistence of lamellipodia extension and ruffle frequency in N17Rac1 transgenic and wild-type keratinocytes by using computer assisted stroboscopic analysis of cell dynamics (SACED) (Hinz et al., 1999). This analysis revealed that the presence of the N17Rac1 transgene in keratinocytes resulted in reduced persistence of lamellipodia protrusion and in increased ruffle frequency, suggesting that protruded membrane domains were retracted before real lamellipodia could be assembled (Fig. 6a-d). In addition to its function in the regulation of lamella protrusion, Rac1 plays a crucial role in the formation of focal adhesion complexes at attachment sites of the cell membrane to the underlying matrix (Guo et al., 2006). To analyse whether reduced lamella protrusion persistence is due to impaired membrane attachment, we stained N17Rac1 transgenic keratinocytes and wild-type cells adherent to collagen-coated glass slides for focal contact complexes using vinculin as a marker. N17Rac1-expressing keratinocytes contained less focal contacts compared with controls (Fig. 6e,f). We conclude that N17Rac1-expressing keratinocytes have a decreased ability to form focal contacts through which protruded lamellipodia are attached to the matrix.
Reduced migration and wound closure of N17Rac1 transgenic keratinocytes in vitro
Adhesion and lamella protrusion are functions that determine the efficiency of cell migration. Apart from proliferation, keratinocyte migration is a prerequisite for a normal repair function of the epidermis. Although Rac activity has been suggested to be of general importance for cell migration (Nobes and Hall, 1999), previous studies have shown that inhibition of Rac activity or loss of Rac1 does not necessarily result in impairment of cell migration (Ahram et al., 2000; O'Connor et al., 2000; Vidali et al., 2006; Wells et al., 2004). To study whether Rac contributes to the migration required for wound re-epithelialization in keratinocytes, we analysed keratinocyte movements by time-lapse video microscopy, again using collagen I or fibronectin as a matrix. Keratinocytes were plated onto dishes coated with these substrates and cell migration was monitored for 20 hours with a time-lapse video microscope in the absence and presence of growth factors. Under all conditions tested, transgenic keratinocytes showed a reduction in the distance and speed of migration (Fig. 7a,b). Notably, inhibition of motility in transgenic keratinocytes was more pronounced when cells were migrating on collagen I compared with fibronectin (see Fig. 7a). To investigate whether this deficiency in migration could account for the delay in wound re-epithelialization, we tested the capacity of growth-arrested cultured N17Rac1 transgenic and control keratinocytes to close a scratch wound in vitro. In analogy to measurements of keratinocyte motility, assays of wound closure in vitro were carried out with cells plated on either collagen I or fibronectin. Analysis of primary keratinocytes from three transgenic and three control mice revealed a delay in wound re-epithelialization similar to that observed in N17Rac1 transgenic mice in vivo only when cells were plated on collagen I, but not when plated on fibronectin (Fig. 7c,d; see Fig. 2 for comparison). We, therefore, conclude that inhibition of RAC activity in epidermal keratinocytes by expression of the N17Rac1 transgene results in a cell-autonomous migration defect on collagen I that contributes to the observed delay in wound re-epithelialization in vivo.
Rac1 activity in primary human keratinocytes in vitro is required for the maintenance of cell-cell contacts (Braga et al., 2000; Braga et al., 1997) and the inducible deletion of Rac1 from the epidermis of adult mice in vivo results in epidermal stem cell depletion (Benitah et al., 2005). Constitutive, epidermis-specific deletion of Rac1 during murine embryonic development results in destruction of hair follicles but does not affect the interfollicular epidermis (Chrostek et al., 2006). In view of such dramatic changes observed after loss of RAC function in epidermal keratinocytes, we were surprised that expression of N17Rac1 in the epidermis of mice did not result in an overt skin phenotype. The cause for this discrepancy remains currently open. Possibly, inhibition of Rac function by expressing a dominant-negative mutant is incomplete and remaining Rac activity in N17Rac1 transgenic mice may result in a normal phenotype. Alternatively, the restriction of Rac inhibition to basal epidermal and outer root sheath keratinocytes only could be, in contrast to deletion of Rac1 from all epidermal layers, insufficient to produce skin pathology. Nevertheless, the absence of a spontaneous phenotype makes N17Rac1 transgenic mice most suitable to directly investigate functions of Rac in physiological stress situations, such as wound healing.
Although a potential function for Rac1 in wound healing has been predicted from previous in vitro studies (Farooqui and Fenteany, 2005; Fenteany et al., 2000; Nobes and Hall, 1999) and from loss-of-function or inhibition experiments in Drosophila (Hakeda-Suzuki et al., 2002; Harden et al., 1999; Woolner et al., 2005) the complexity of the wound healing process in mammals cannot truly be represented by these models. We show here that N17Rac1 transgenic mice and Rac1E-KO mice have a defect in wound re-epithelialization in vivo. Our results demonstrate for the first time that epidermal keratinocytes require normal Rac activity for efficient wound re-epithelialization.
It is unknown whether the N17Rac1 mutant shows specificity for the Rac1 isoform with respect to its effects in cells and tissues. Rac2 is specifically expressed in the hematopoietic system (Roberts et al., 1999); however, we cannot exclude a possible effect of Rac3 inhibition by the N17Rac1 transgene, even though Rac3-null mice have not been reported to show an overt skin phenotype (Cho et al., 2005; Corbetta et al., 2005). Dominant-negative mutants of Rac1 have also been suggested to inhibit non-Rac dependent cellular functions (Vidali et al., 2006). A specific role for Rac1 in epidermal wound healing is, however, suggested by our finding that wound re-epithelialization is severely disturbed in mice carrying an epidermis-specific deletion of Rac1. In contrast to N17Rac1 transgenic mice, Rac1E-KO mice show progressive hair-follicle destruction after birth (Chrostek et al., 2006). Since in mice the bulge region of the hair follicle has been identified as the location of hair follicle stem cells (Cotsarelis et al., 1990; Tumbar et al., 2004), which are known to temporarily contribute to the re-population of the epidermis after wounding (Ito et al., 2005; Tumbar et al., 2004), it is perceivable that in Rac1E-KO mice disturbed wound re-epithelialization is due to the loss of this cell population. Although this could account for the observed delay in wound re-epithelialization, our analysis of keratinocytes isolated from N17Rac1 transgenic mice shows that inhibition of Rac has also direct effects on keratinocyte proliferation and migration, thus, making it probable that Rac exerts multiple functions in epidermal wound healing in mice.
A mechanism by which depletion of bulge stem cells leads to impaired wound healing would be analogous to the described depletion of interfollicular epidermal stem cells following inducible deletion of Rac1 in the epidermis. Recently, a Rac1 dependent pathway involving PAK2 and Myc was shown to be able to regulate the proliferative potential of keratinocytes (Benitah et al., 2005). Our results confirm a function of Rac in the regulation of keratinocyte proliferative potential. In addition, we have now found an influence of Rac inhibition on the activity of the ERK signaling pathway which we have shown previously to be relevant for the regulation of keratinocyte proliferation and epidermal stem cell properties (Haase et al., 2001; Zhu et al., 1999). It is therefore conceivable that wound re-epithelialization in N17Rac1 mice is impaired because the necessary expansion of a pool of keratinocytes with high proliferative potential is inhibited by expression of the transgene. Since repression of Rac1 function in transgene bearing basal keratinocytes is incomplete the consequences do not become obvious under steady state conditions of the epidermis but only in a situation that requires increased production of tissue mass. Alternatively, it cannot be excluded that enhanced activity of the keratin 14 promoter in wounded skin facilitates transgene expression under these conditions. An explanation for diminished ERK phosphorylation upon Rac inhibition could be reduced keratinocyte adhesion, because integrin engagement is required for growth-factor-stimulated ERK activation (Renshaw et al., 1997), and interference with integrin binding to extracellular matrix has been shown to inhibit ERK phosphorylation in human keratinocytes (Zhu et al., 1999).
However, isolated keratinocytes from N17Rac1 transgenic mice show intrinsic proliferation and migration defects, arguing that regulation of proliferative potential is not the only possible mechanism that could contribute to delayed wound re-epithelialization. Keratinocyte proliferation is an essential component of the wound healing response because it serves the replacement of lost tissue mass. An involvement of Rac in the control of cell proliferation, possibly by the control of cyclin D expression, has been shown in different cell lines (e.g. Joneson et al., 1996; Joyce et al., 1999) and Rac1 deficient mouse embryonic fibroblasts have a severe proliferation defect (Vidali et al., 2006). Our finding of decreased BrdU incorporation into N17Rac1 transgenic keratinocytes compared with wild-type keratinocytes in the presence of two growth factors with an established role in wound healing could either point to a protraction of the cell cycle or to a decreased number of cycling keratinocytes in the N17Rac1 expressing populations.
Rac is also known as a cardinal regulator of cell migration. Inhibition of Rac results in disturbed migration of keratinocytes both in vitro and in vivo. This is in contrast to macrophages and mouse embryonic fibroblasts, in which deficiency of Rac1 resulted in defects in lamellipodia formation and cell spreading, but not in cell migration in vitro (Guo et al., 2006; Vidali et al., 2006; Wells et al., 2004). Although it cannot be excluded that, in macrophages, Rac2 can compensate for the loss of Rac1, these results suggest that Rac is not necessarily required for cell migration in vitro. In contrast to this view, Rac is required for normal recruitment of hemocytes to wounds in Drosophila embryos (Stramer et al., 2005).
In fibroblasts, migration speed is directly correlated with persistence of lamellipodia protrusion, which is negatively regulated by ENA/VASP proteins (Bear et al., 2000; Bear et al., 2002). SACED analysis of N17Rac1-expressing and control keratinocytes revealed a reduction in persistence of lamellipodia protrusion upon inhibition of Rac function that may represent a possible mechanism for impaired migration. Interestingly, we have observed an inverse correlation between persistence of lamella protrusion and ruffle frequency. Ruffles are protruded membrane domains that do not attach to the underlying matrix but fold towards the dorsal cell side. Together with the reduced number of focal contacts in N17Rac1 transgenic keratinocytes, which has also been found in mouse embryonic fibroblasts upon deletion of Rac1 (Guo et al., 2006; Vidali et al., 2006) the observed increase in ruffle frequency in N17Rac1 transgenic cells may suggest that the primary defect responsible for decreased protrusion persistence does not necessarily lie in the assembly and disassembly of the actin-filament network at the leading edge that drives lamella protrusion, but rather in an impairment to establish firm contacts between the cell membrane and the extracellular matrix.
Attachment, spreading and wound re-epithelialization in vitro of N17Rac1 transgenic keratinocytes showed a greater impairment on collagen I than on fibronectin. Obviously, other pathways depending on the engagement of different adhesion receptors must be involved. Since both α subunits of the relevant matrix receptors (α2β1 integrin for collagen and α5β1 integrin for fibronectin) couple to the β1 integrin chain, the information that confers specificity for these functions must reside in the α2 and the α5 integrin subunits. At present, little is known about integrin α-subunit-specific signaling events. Available evidence points, however, to differential signaling – even between different α subunits of the collagen receptors α1β1 and α2β1. The α2-integrin subunit, but not the α1-integrin subunit, was able to activate MAP kinase, which resulted in collagen synthesis and cell migration (Ivaska et al., 1999; Klekotka et al., 2001). Interestingly, α2-integrin-dependent cell migration could be augmented or reduced by modulation of Rac activity (Klekotka et al., 2001), suggesting specific crosstalk between Rac and an α2-integrin-dependent and/or MAP-kinase-dependent signaling pathway. This would provide a plausible explanation for the selective effects of the N17Rac1 transgene on keratinocyte adhesion, spreading and wound re-epithelialization in vitro when keratinocytes are grown on collagen I but not fibronectin.
In summary, we provide the first evidence for an important function of Rac1 in normal skin wound healing in vivo by regulating keratinocyte proliferation and migration. This contributes to our understanding of basic mechanisms regarding wound healing and may have future therapeutic implications for the treatment of non-healing or chronic skin wounds.
Materials and Methods
Generation of mutant mice
Rac1 c-DNA containing a point mutation C17 to A and a C-terminal Myc-tag (Kjoller and Hall, 2001) was inserted via SacI-XbaI sites into a keratin 14 expression cassette harboring an SV40 intron and an SV40 polyA signal sequence at the N-terminus (Hafner et al., 2004). Correct insertion was confirmed by direct DNA sequencing. After digestion of the plasmid with BssH2 the transgene was separated by agarose gel electrophoresis and isolated from the gel using the MinElute gel extraction kit (Qiagen, Hilden, Germany). Purification was performed using an Elutip mini column (Schleicher & Schüll, Dassel, Germany) and subsequent precipitation with ethanol. For pronucleus injection, the DNA was dissolved in microinjection buffer and adjusted to a final concentration of 10 ng/μl.
Transgenic mice were generated by injection of the DNA construct into the pronucleus of fertilized oocytes. For screening of transgene insertion, genomic DNA was isolated from mouse tails and analysed bymeans of polymerase chain reaction (PCR) using the primers SF3-25 5′-TTGGTTGTGTAACTGATCAGTAGGC-3′ and SF5-23 5′-TGGAGAGCTAGCAGGAAACTAGG-3′. Insertion was confirmed by Southern blot analysis using a 600-bp fragment as probe. Copy numbers of the inserted transgene were estimated using the IL-1 receptor gene as a single copy gene for comparison (Carroll et al., 1995). The presence of the N17 mutation in the transgenic Rac1 cDNA was verified by sequencing of PCR fragments obtained from the screening of transgenic mice. Founder mice were backcrossed into a wild-type C57Bl6 background for five generations.
Isolation and culture of epidermal keratinocytes
Keratinocytes were isolated from transgenic and wild-type newborn and adult mice. For the isolation of neonatal murine keratinocytes mice were killed between 1 and 3 days after birth and skinned. The epidermis was removed en bloc after incubation of total skin in Trypsin (0.1%)-EDTA (0.06%) solution overnight. Pieces of epidermis after mincing were suspended in low Ca2+ FAD medium and incubated on a shaker at 1000 rpm for 1 hour at room temperature. Cells were resuspended in low-Ca2+ FAD medium and plated onto collagen-coated dishes (BioCoat, BD Biosciences, Heidelberg, Germany) on which a feeder layer of mitomycin-treated 3T3 fibroblasts had been prepared (Watt, 1998). Low-Ca2+ FAD medium had essentially the same composition as FAD medium (Watt, 1998), except that the Ca2+ concentration was 50 μM, and contained the following supplements: 10% fetal calf serum (FCS) treated with Chelex 100 resin (BioRad, Munich, Germany) and sterile-filtered prior to use and a cocktail of 1.8×10–10 M adenine, hydrocortisone (0.5 μg/ml), insulin (5 μg/ml), cholera toxin (10–10 M) and epidermal growth factor (EGF) (10 ng/ml). Primary neonatal keratinocytes were cultured on a 3T3 cell feeder layer up to passage four and thereafter without feeder cells.
Primary adult murine keratinocytes were isolated from the back skin of transgenic and wild-type mice as described (Morris et al., 1987). After shaving mice were immersed in povidone iodine surgigal solution followed by washes in sterile water, 70% ethanol and again sterile water. The mice were skinned and the skins were placed fur-side down in a Petri dish with PBS-2× antibiotic solution (PBS + 1/5 [v/v] penicillin-streptomycin; 104 IU penicillin/ml, 104 μg streptomycin/ml + 1/2000 [v/v] amphothericin B 5 mg/ml). Subcutaneous fat was removed and skin was added (dermis-side down) to a solution containing 0.25% trypsin (Gibco BRL, Grand Island, NY) in PBS (Ca2+-Mg2+-free) for 2 hours at 32°C. Skins were washed in 10 ml harvesting medium (FAD media + FCS + 2× antibiotics). The epidermis was scraped off with a scalpel into 10 ml harvesting medium. The tissue suspension was stirred for 20 minutes at room temperature and filtered through a cell strainer with 70 μm pore size. After a centrifugation step the supernatant was removed, the cell pellets resuspended in 2 ml FAD low-Ca2+ and plated onto collagen coated 6-cm Petri dishes (BioCoat, Beckton Dickinson). After overnight incubation at 32°C and 5% CO2 attached cells were lysed and used for western blot analysis.
Rac GTP loading assay
Rac1 activity was assayed with a biotinylated peptide corresponding to the CRIB motive of the Rac1 effector Pak, essentially as described previously (Price et al., 2003). Briefly, keratinocytes were plated for 40 minutes on collagen-I-coated or fibronectin-coated 6-cm Petri dishes. The cells were then scraped off the plates into 400 μl lysis buffer containing 2 μl of the CRIB protein (2 mg/ml) and protease inhibitor mix. Cell lysates were cleared by centrifugation and the Rac1-GTP complexes were precipitated with streptavidin-agarose and solubilized in SDS sample buffer. Detection of Rac1 was by western blotting using an anti-Rac1 antibody (clone 23A8; Upstate Biotechnology).
Keratinocytes were lysed in situ in modified RIPA buffer (containing 5 mM EDTA, 1% Triton X-100, 1% NP40, 0,1% SDS, 0,5% deoxycholate, 20 μM leupeptin, 1 mM PMSF, 0.5 mg/ml soybean trypsin inhibitor, 0.5 mM NaVO3 and 10 mg/ml p-nitrophenylphosphate), scraped from the dishes and sonicated for 30 seconds at full power. Lysates were centrifuged at 14,000 g for 10 minutes and the supernatant was used for western blot analysis. Equal amounts of protein were separated by SDS-PAGE and blotted onto nitrocellulose LC 2000 membranes (Invitrogen, Carlsbad, CA) probed with antibodies against the Myc-epitope (9B11, Cell Signalling Technology, Frankfurt, Germany; dilution 1:500), Rac1 (clone 23A8; Upstate, Charlottesville, VA), phosphorylated ERK1/2 (rabbit monoclonal, cat. no. 4376, Cell Signalling Technology) and total ERK (Santa Cruz Biotechnology, Santa Cruz, CA). Immunoreactive proteins were visualized with horseradish peroxidase (HRP)-coupled secondary antibodies on Hyperfilm by chemiluminescence (Western Lightning, Perkin Elmer, Rodgau, Germany). For quantification of the densitiy of protein bands a Fluro-S Multimager (Bio-Rad) was used. Protein quantification was carried out using the Pierce protein quantification kit (Pierce).
Wound healing in vivo
Wounding experiments were carried out as described previously (Eckes et al., 2000) and after obtaining approval by the local authorities. Wounds were created on the back skin of mice using a 4-mm biopsy punch. Two or five days after wounding mice were killed and wounds were excised, bisected and either fixed in 4% paraformaldehyde overnight or embedded in OCT compound (Tissue-Tek, Sakura, Zoeterwoude, The Netherlands) and frozen immediately. Morphometric measurements of `length' and `area' (for illustration, see Results section and Fig. 2) were performed on hematoxylin-eosin (H-E) stained sections using a Leica DM 4000 B microscope (Leica, Wetzlar, Germany) equipped with a KY-F75U digital camera (JVC, Wayne, NJ) and Diskus software (Carl H. Hilgers, Technisches Büro, Koenigswinter, Germany).
For visualization of S-phase keratinocytes in sections of wound tissue, mice were injected with 3.2 mg BrdU i.p. 2 hours prior to killing. Detection of BrdU was performed on sections of paraffin embedded tissue which were de-paraffinized in xylene, rehydrated in graded alcohols and rinsed in automation buffer (Biomedia, Foster City, CA). Sections were incubated in 2 M HCl at 37°C, washed in borate buffer and digested with 0.01% trypsin in 0.05 M Tris for 3 minutes at 37°C. After blocking with 10% goat serum for 20 minutes, sections were incubated for 1 hour at room temperature with mouse BrdU antiserum (Becton Dickinson; dilution 1:25). Secondary antibodies were conjugated with Alexa Fluor-488 (Molecular Probes).
For determination of BrdU incorporation into cultured keratinocytes cells were seeded onto chamber slides (Invitrogen) at equal densities, starved for 2 hours in low-Ca2+ FAD medium without supplements and then stimulated with 10 ng/ml EGF or 100 ng/ml IGF1. 1 μM BrdU were added simultaneously with the respective growth factor. After 30 minutes incubation at 32°C and 5% CO2 the medium was removed and cells were fixed in 70% ethanol for 30 minutes at room temperature. Cells were allowed to dry and then incubated with 0.07 M NaOH for 2 minutes. After washing cells were incubated with anti-BrdU antibody (see above) diluted 1:5 in PBS-Tween 20 (0.5%) for 45 minutes at room temperature. For detection an Alexa Fluor-488-labeled goat anti-mouse IgG was used (Molecular Probes).
Histopathological analysis and immunostaining
After excision, tissue samples were fixed in 4% paraformaldehyde or embedded in OCT compound and frozen immediately. Further processing, paraffin embedding of paraformaldehyde fixed tissue and H-E staining was according to standard histopathological procedures. Immunostaining of keratin 14, keratin 10 and loricrin was performed as described in (Stratis et al., 2006). For staining of E-cadherin, integrins, matrix proteins, Rac and the Myc-tag the following antibodies and procedures were applied on frozen sections: E-cadherin, antibody GP84 (kindly provided by Rolf Kemler, Max-Planck-Institute of Immunobiology, Freiburg, Germany) at 1:400, 4% paraformaldehyde fixed tissue; α2 integrin, antibody M075-0 (emfret, Würzburg, Germany) at 1:200, acetone fixed tissue; β1 integrin, MAB1997 (Chemicon, California, CA) at 1:200, acetone fixed tissue; collagen IV, antibody CL50451AP (Cedarlane, Ontario, Canada) at 1:1500, acetone fixed tissue; laminin 5, antibody 8LN5 (kindly provided by Manuel Koch, Institute of Biochemistry, University of Cologne, Germany) at 1:100, acetone fixed tissue; Rac, antibody 23A8 (Sigma) at 1:100, 4% paraformaldehyde fixed tissue; Myc-tag, antibody 9B11 (Cell Signalling Technology, Frankfurt, Germany) at 1:1000, 4% paraformaldehyde fixed tissue. Phospho-specific monoclonal rabbit anti-STAT3 and anti-ERK1/2 antibodies were obtained from Cell Signaling Technology and used according to the manufacturer's protocols.
Cultured keratinocytes were fixed in 4% paraformaldehyde for 10 minutes at room temperature, followed by permeabilization with 0.5% Triton X-100 for 5 minutes at room temperature. Immunostaining was carried out using antibodies against Myc (9B11, New England Biolabs; dilution 1:500) and vinculin (Chemicon, Temecula, CA,USA). Polymerized actin was stained using TRITC-labeled phalloidin. Secondary antibodies coupled to Alexa Fluor-488 (Molecular Probes, Eugene, OR) at a dilution of 1:500 or 1:1000 were used for visualization.
Adhesion and spreading assay
Adhesion assays were carried out as described in (Zhu et al., 1999). In brief, keratinocytes were plated onto collagen-I-coated (20 μg/ml), fibronectin-coated (10 μg/ml) and poly-L-lysine-coated (20 μg/ml) bacteriological 96-well plates (Invitrogen, Karlsruhe, Germany) in low-Ca2+ FAD medium without supplements. Cells were allowed to adhere for 1 hour, washed twice, and lysed with 0.9% Triton X-100 in PBS. Release of lactate dehydrogenase was determined using the CytoTox 96 kit (Promega, Mannheim, Germany). Spreading assays were carried out as described by Haase et al. (Haase et al., 2003). Cells were plated in DMEM and stained with 9B11 for the Myc-tag and with TRITC-phalloidin for polymerized actin.
Wound epithelialization in vitro
Primary murine keratinocytes (passages 1-4) were plated at equal densities onto collagen-I-coated six-well plates (BioCoat; BD Biosciences) or six-well plates (Falcon, BD Biosciences) coated overnight with 10 μg/ml human fibronectin (Sigma, Taufkirchen, Germany) in low-Ca2+ medium. Cultures were treated with 4 μg/ml mitomycin C in low-Ca2+ FAD medium without FCS for 2 hours, then washed in PBS and the monolayer was wounded with a tip of a glass pipette. Cells were incubated in low-Ca2+ FAD medium for up to 24 hours. Phase-contrast pictures were taken every 5 minutes with a time-lapse video microscope (see below).
Time-lapse video microscopy and SACED analysis
Sixteen hours after plating keratinocytes onto collagen-I- and fibronectin-coated dishes in low-Ca2+ FAD medium, the medium was changed to keratinocyte SFM (Invitrogen) in which cells were kept throughout the experiment. The surface of the medium was overlaid with mineral oil to avoid evaporation. Cells were observed for up to 24 hours in an incubator chamber attached to an IX81 microscope (Olympus Europa GmbH, Hamburg, Germany) at 32°C, 5% CO2 and 60% humidity. Frames were taken every 5 minutes using an OBS CCD FV2T camera (Olympus). For visualization and quantification of cell movements, the computer programs OBS Cell R (Olympus) and DIAS (Solltech, Oakdale, IA) were used.
Lamella dynamics were quantified using a computer-assisted method of stroboscopic analysis of cell dynamics (SACED) that has been described previously (Hinz et al., 1999). Phase-contrast image series of motile keratinocytes on collagen-coated glass chamber slides (Invitrogen, Carlsbad, CA) were collected using an inverted confocal laser-scanning microscope LSM 510 (Zeiss, Oberkochem, Germany) equipped with a 63× 1.4 NA Ph3 plan apochromat objective and an airconditioned incubation chamber. At least 80 individual keratinocytes per line between passages three and five obtained from three different N17Rac1 transgenic mice and three different wild-type mice were monitored for 5 minutes. Pictures were captured every 2 seconds resulting in time-space plots that represent lamella dynamics. From these pictures lamella protrusion persistence and ruffle frequency were calculated as described (Hinz et al., 1999). Statistical evaluation of SACED-derived motility data was performed using SPSS software (SPSS Inc., Chicago, IL Statistical significance between data groups was determined using the Whitney U-Test and considered to be significant at P<0.05 and highly significant at P<0.01.
Determination of significance between samples were made using the Student's t test (GrapPad Prism; GraphPad Software), except for the SACED analysis for which SPSS software (SPSS Inc., Chicago, Il) was used. For all statistical tests, the 0.05 level of the confidence interval was accepted as statistically significant.
This work was supported by a grant from Deutsche Forschungsgemeinschaft (SFB 589). We are grateful to Beate Eckes and Manuel Koch for providing antibodies.
- Accepted February 8, 2007.
- © The Company of Biologists Limited 2007