¹Turku Centre for Biotechnology, University of Turku and �bo Akademi University, Finland

²Departments of Medical Biochemistry, and Dermatology, University of Turku, Turku, Finland

Correspondence to: V-M Kähäri, Turku Centre for Biotechnology, University of Turku, Tykistökatu 6B, FIN-20520 Turku, Finland

^aCurrent address: Wellcome Trust Centre, University of Oxford, Roosevelt Drive, Headington, Oxford OX3 7BN, UK

Impaired cutaneous wound healing is a common complication in diabetes, ischemia and venous insufficiency of lower extremities, and in long-term treatment with corticosteroids or other immunosuppressive agents. In development of gene therapy for wound repair, expression of therapeutic transgenes should be precisely targeted and controlled. Here, we describe a recombinant adenovirus RAdFiRE-EGFP, in which a growth factor inducible element (FiRE) of the murine syndecan-1 gene controls the expression of enhanced green fluorescent protein (EGFP) reporter gene. Treatment of RAdFiRE-EGFP-transduced murine epidermal keratinocytes in culture with FiRE-activating growth factor markedly enhanced the expression of EGFP. In ex vivo organ culture of wounded murine skin transduced with RAdFiRE-EGFP, the EGFP expression was specifically detected in wound margin keratinocytes, but not in intact skin. Activity of EGFP was first detected 2 days after a single application of RAdFiRE-EGFP and persisted up to 10 days. Similarly, FiRE-driven EGFP expression was detected specifically in epidermal keratinocytes in the edge of incisional wounds in murine skin transduced with RAdFiRE-EGFP. In contrast, adenovirus-mediated lacZ expression driven by CMV promoter was detected scattered in epidermal, dermal and subcutaneous layers in ex vivo and in vivo wounds, as well as in intact skin. These data demonstrate the feasibility of FiRE as a tool for transcriptional targeting of adenovirus-mediated transgene expression to cutaneous wound edge keratinocytes. Gene Therapy (2000) 7, 1640-1647.

adenovirus; EGF; FiRE; syndecan-1; transcription; wound repair

Introduction

Wound healing is a dynamic process, which includes inflammation, tissue formation, and tissue remodeling.¹ Impaired cutaneous wound healing is an important medical problem and a common complication, eg in diabetes, ischemia and venous insufficiency of lower extremities, as well as in long-term treatment with corticosteroids or other immunosuppressive agents. Chronic dermal ulcers are characterized by impaired re-epithelialization and granulation tissue formation, as well as insufficient neovascularization.¹ Recently, the possibility of correcting these defects in wound repair by introducing polypeptide growth factors or cytokines to chronic wounds has been under intensive investigation. Accordingly, topical application of epidermal growth factor (EGF), fibroblast growth factors (FGFs) and platelet-derived growth factor (PDGF) have been shown to stimulate wound healing.^1,2,3,4,5 In addition, the feasibility of other therapeutic proteins, such as inducible nitric oxide synthase (iNOS) in therapy of chronic ulcers has been suggested.⁶

Although easily performed, topical administration of therapeutic polypeptides to wounds has several disadvantages. First, the effect of topically applied growth factors can not be targeted to specific cells in the wounds and their effects are therefore difficult to control. In addition, production of recombinant growth factors is expensive, and due to their instability in the wound environment they should be administered frequently to obtain a sufficient effect. It is expected that controlled local production of therapeutic polypeptides in the wound as a result of effective gene delivery, would overcome these problems. The feasibility of viral gene delivery using adenovirus, herpes simplex virus and retrovirus vectors, as well as nonviral gene transfer by naked DNA injection, liposomes or particle-mediated gene delivery to skin cells either in vivo or ex vivo has been demonstrated.^7,8,9,10,11 One main goal in development of gene therapy aimed at improving cutaneous wound repair is targeting of therapeutic genes to a defined cell population in the wound to avoid undesired effects of growth factors. This could be achieved by transcriptional targeting of transgene expression to a specific cell population using cell specific regulatory elements, as has been demonstrated in gene delivery to melanoma, liver, mammary gland, and smooth muscle cells.^12,13,14,15

We have previously characterized an FGF-inducible response element (FiRE), a 280 bp gene fragment located 10 kb upstream of the promoter of the murine gene for heparan sulfate proteoglycan syndecan-1.¹⁶ The activity of FiRE is induced by FGF-2 in NIH-3T3 cells, whereas in epidermal keratinocytes FiRE is activated by EGF, transforming growth factor- (TGF-), and keratinocyte growth factor (KGF, FGF-7).¹⁷ In murine keratinocytes FiRE binds two inducible AP-1 complexes, a constitutively expressed upstream stimulatory factor-1 (USF-1), and an uncharacterized 46 kDa protein. In vivo studies with transgenic mice harboring a FiRE-syndecan-1 promoter cassette in front of lacZ reporter gene, have revealed that FiRE is activated in migrating keratinocytes during re-epithelialization of cutaneous wounds, but not in fibroblasts or endothelial cells, or in intact skin or other adult tissues.¹⁸ Activation of FiRE-driven transcription is first detected in the epithelial sheet adjacent to the incision site 24 h after wounding, when the keratinocytes start to migrate towards the wound base. The activation of FiRE is marked at the leading edge of the migrating keratinocytes and in the merging epithelium where the re-epithelializing keratinocyte sheets fuse, but at the end of re-epithelialization the activity declines. In addition, FiRE activation can be detected in the remnants of the hair follicle keratinocytes adjacent to the wound site.¹⁸

To examine the feasibility of FiRE in transcriptional targeting of gene delivery to healing wounds we have constructed a recombinant adenovirus RAdFiRE-EGFP, in which FiRE drives the expression of enhanced green fluorescent protein (EGFP) reporter gene. We show that transduction with this adenoviral construct results in growth factor-inducible EGFP expression in cultured epidermal keratinocytes, as well as in EGFP expression specifically in wound edge keratinocytes in ex vivo and in vivo skin wound models. This study is the first demonstration of adenoviral gene delivery transcriptionally targeted to migrating epidermal keratinocytes, and supports the feasibility of FiRE as a tool in targeted gene therapy for improving cutaneous wound repair.

Results

EGF activates FiRE-syndecan promoter in adenovirally transduced keratinocytes

To study the feasibility of FiRE in gene delivery targeted to wound epidermal keratinocytes, we constructed a recombinant replication-deficient adenovirus RAdFiRE-EGFP, in which the FiRE sequence together with a 2.2 kb proximal promoter of the murine syndecan-1 gene directs the expression of enhanced green fluorescent protein (EGFP) reporter gene (Figure 1a). We have previously shown that FiRE is activated by EGF, in transiently and stably transfected epidermal keratinocytes in culture.¹⁷ Therefore, we initially tested whether EGF similarly activates the RAdFiRE-EGFP in transduced epidermal keratinocytes in culture. Murine epidermal keratinocytes, (MCA3D cells) were transduced with RAdFiRE-EGFP (1000 p.f.u. per cell) and incubated without or with EGF (10 ng/ml), added 24 h after transduction. Expression of EGFP was detected in untreated RAdFiRE-EGFP- transduced cells, and it was markedly enhanced by 48-h EGF stimulation with virtually all cells becoming EGFP positive and expressing higher quantities of EGFP (Figure 1b). Likewise, quantification of the cell population expressing EGFP (rather than EGFP intensity) without (control) or with EGF treatment demonstrates marked enhancement by EGF stimulus (Figure 1c). The EGFP activity remained high over the following days and persisted at least up to 7 days of incubation (not shown). This observation demonstrates that the FiRE-syndecan-1 promoter cassette embedded within adenovirus functions similarly as in transiently transfected episomal plasmid vectors, or when integrated into the genome of stably transfected cells.

FiRE targets adenovirus-mediated EGFP expression to dermal wound margins ex vivo

We have recently noted, that in transgenic mice the FiRE-syndecan-1 promoter cassette directs the expression of lacZ reporter gene to migrating epidermal keratinocytes in the wound edges during the re-epithelialization of cutaneous wounds.¹⁸ In contrast, no activation of FiRE is seen in the intact skin or other organs. To examine, whether FiRE can also be used for transcriptional targeting of gene expression to migrating keratinocytes when embedded in adenoviral vector, we initially tested RAdFiRE-EGFP in an ex vivo cutaneous wound model.

Mice were killed, the hair was removed from their back, and linear full thickness wounds were incised in their backs and tails. Pieces of skin containing the wound with a 1 cm margin, as well as 1 cm pieces of wounded tail were transferred to organ culture, RAdFiRE-EGFP (3 ´ 10⁸ p.f.u.) was added to the culture medium (2 ml) and the tissues were examined daily under fluorescence microscope for up to 2 weeks. The EGFP activity was first detected 1 day after transduction and marked expression was noted at the wound edge 2 days after transduction (Figure 2a). EGFP expression was even more potent from day 3 up to 10 days after infection, after which it gradually declined (Figure 2b). The expression of EGFP was restricted to the wound site and it was not detected in adjacent intact skin (Figure 2a and b). Notably, no EGFP was detected in the intact skin even after tape stripping known to activate expression of EGF receptors on keratinocytes¹⁹ (not shown). We also wanted to determine the viral dose required for maximal transgene expression after ex vivo adenoviral transduction of wound tissue. With dose 7.5 ´ 10⁷ p.f.u., the expression of EGFP was barely detectable 3 days after infection, whereas with dose 1.5 and 3 ´ 10⁸ p.f.u. the expression was marked (not shown). With doses of 6 ´ 10⁸ p.f.u. or higher the expression was less prominent (not shown).

FiRE targets adenovirally delivered EGFP gene expression to migrating keratinocytes

To determine the exact localization of FiRE-driven EGFP expression in RAdFiRE-EGFP-infected cutaneous wounds, we examined EGFP expression in native frozen tissue sections by UV microscopy 3 days after transduction with RAdFiRE-EGFP (3 ´ 10⁸ p.f.u.). In UV microscopy EGFP expression was noted in epidermal keratinocytes at the wound edge, but not in adjacent intact skin (Figure 3). As opposed to migrating keratinocytes, proliferating keratinocytes reside farther apart from the wound edge. Our previous results with transgenic mice showed that FiRE-driven gene expression is located in migrating, not in proliferating, keratinocytes as illustrated by PCNA staining.¹⁸ The expression pattern of EGFP after transduction with RAdFiRE-EGFP was undistinguishable from the FiRE-driven transgene expression in cutaneous wounds in transgenic mice and was restricted to the migrating keratinocytes. Expression was not detected in adjacent proliferating or differentiating keratinocytes (Figure 3). These results provide evidence that FiRE can be used to target genes into wound edge epidermal keratinocytes using adenoviral gene transfer in ex vivo wound model.

To exclude the possibility that selective expression of EGFP in migrating keratinocytes in RAdFiRE-EGFP-infected wounds would be due to selective adenoviral infection of these cells, wounded skin pieces in organ culture were transduced with RAdlacZ (3 ´ 10⁸ p.f.u.), an adenovirus bearing CMV IE promoter in front of E. coli -galactosidase (lacZ) reporter gene, incubated for 3 days, fixed and stained for -galactosidase activity. In contrast to RAdFiRE-EGFP infected wounds, the reporter gene expression was detected scattered in epidermal and dermal layers of RAdlacZ-infected wounded skin, as well as in the adjacent intact skin (Figure 4). These data imply that selective expression of EGFP in wound edge keratinocytes in RAdFiRE-EGFP infected wounds tissue cannot be attributed to selective adenoviral infection of these cells. In addition, these results show, that in contrast to RAdFiRE-EGFP, the CMV promoter-driven gene expression is not cell specific and is therefore unlikely to prove useful for targeted gene delivery to wound cells.

FiRE targets adenoviral EGFP gene expression to migrating keratinocytes in vivo

Next, we examined the efficiency and specificity of RAdFiRE-EGFP and RAdlacZ in gene delivery to linear cutaneous wounds in vivo. Full thickness linear wounds were incised in the back of mice after hair removal under general anesthesia, RAdFiRE-EGFP or RAdlacZ (1.1 ´ 10⁹ p.f.u. each) were administered into the site of wound before wounding, immediately after wounding, or 24 h later, and the wounds were either left uncovered or were covered with plastic. Mice were killed 1, 3 or 7 days after transduction, native frozen sections from RAdFiRE-EGFP transduced mice were examined under fluorescence microscope for EGFP activity, and sections from the RAdlacZ-transduced wounds were stained with X-gal for -galactosidase activity.

The most prominent expression of EGFP was found 3 days after transduction with RAdFiRE-EGFP, whereas the level of expression detected 1 or 7 days after transduction was low or undetectable (not shown). As seen in the ex vivo wound model, FiRE targeted the adenovirus-mediated EGFP expression to migrating keratinocytes in cutaneous wounds in vivo (Figure 5a), In contrast, no expression of EGFP was detected in adjacent epidermal, dermal, or subcutaneous layer of RAdFiRE-EGFP transduced wounds, or in intact skin. The expression of EGFP detected in RAdFiRE-EGFP transduced mice was detectably lower than in the ex vivo transduced wounds, possibly due to lower infection efficiency or inflammation at the wound site in vivo.

As in the ex vivo wounds, infection with RAdlacZ resulted in -galactosidase expression scattered in keratinocytes, dermal fibroblasts, and subcutaneous adipocytes at the wound site (Figure 5b) and in adjacent intact skin. The expression of -galactosidase was most prominent in wound tissue 3 days after transduction with RAdlacZ, but the overall -galactosidase expression was clearly lower than in the ex vivo experiments. No marked difference in transgene expression in RAdFiRE-EGFP or RAdlacZ-infected wounds was detected, whether the wound was covered or left uncovered after adenovirus application (not shown). The most abundant expression was detected when transduction was performed directly after wounding of the skin (not shown).

Discussion

Cutaneous wound repair is a tightly controlled dynamic process, which includes three overlapping phases: inflammation, tissue formation, and tissue remodeling.¹ Impaired wound healing due to ischemia, pressure, or immunosuppression is an important medical problem, which is characterized by defective re-epithelialization, granulation tissue formation, and neovascularization. Several different approaches have been suggested to improve the repair of poorly healing cutaneous wounds, including pressure ulcers and burn wounds. These include topical application of therapeutic polypeptides, transplantation of either autologous grafts or xenografts, and transfer of cultured keratinocytes to the wound.¹ A number of proteins that may enhance re-epithelialization, granulation tissue formation, and angiogenesis have also been proposed for therapy of poorly healing wounds. These include adhesion molecules, eg integrins^20,21 and growth factors such as EGF, KGF, and platelet-derived growth factor (PDGF).^2,3,4 However, topical application of therapeutic proteins is expensive, inefficient, and difficult to control. In contrast, delivery of therapeutic genes directly to wound cells is expected to overcome these problems. In this context, the efficiency of both viral and nonviral delivery of therapeutic genes to wound cells ex vivo or in vivo has been demonstrated, the former being more efficient.^{6,22,23,24,25,26} It is conceivable, that in gene delivery to the wound site, it would be highly desirable to restrict the expression to a defined cell population, as growth factors might induce inappropriate proliferation of normal wound tissue or adjacent intact skin. Furthermore, prolonged exposure to these factors might predispose normal skin cells to malignant transformation. Although transient gene expression would be achieved by using adenoviral vectors, cell-specific promoter elements could add more strict transcriptional control on the duration and location of the transgene expression. Finally, targeting migrating keratinocytes at the edge of the wound, as opposed to proliferating epidermal keratinocytes further away from the wound site, would have obvious advantages if cell migration assisting genes, eg matrix degrading proteinases, were to be delivered.

In the present study, we have used a growth factor inducible promoter element (FiRE) of the murine syndecan-1 gene to construct an adenovirus, RAdFiRE-EGFP, for transcriptional targeting of transgene expression to wound edge keratinocytes. We have previously shown, that this element directs growth factor inducible gene expression in cultured keratinocytes and targets reporter gene expression selectively into migrating keratinocytes in healing cutaneous wounds of transgenic mice both ex vivo and in vivo.^16,17,18 The results of the present study show that infection of cultured murine epidermal keratinocytes with RAdFiRE-EGFP results in low basal EGFP reporter gene expression, but the expression is markedly stimulated by treatment of transduced cells with EGF. These results show, that the activity of FiRE is growth factor-inducible even when embedded in the adenoviral genome. Further experiments showed that transduction of wounded murine skin in ex vivo organ culture with RAdFiRE-EGFP resulted in EGFP expression, which was restricted to the migrating keratinocytes at the wound margins, but was not detected in the intact skin. Interestingly, expression of EGFP was first noted 2 days after a single application of RAdFiRE-EGFP and was detectable up to 10 days. Similarly, FiRE-driven EGFP expression was detected specifically in migrating keratinocytes of incisional wounds in murine skin in vivo. Although the half-life of EGFP in keratinocytes is not known, studies in other cell types have indicated a half-life of 1 to 1.5 days.²⁷ In the present study EGFP expression persisted up to 14 days in ex vivo, but was not detectable after 7 days in vivo. Therefore, it is unlikely that the longer expression in ex vivo could be merely caused by extended EGFP half-life. In contrast to FiRE-driven EGFP expression, lacZ expression driven by CMV promoter was detected widely within the epidermal, dermal and subcutaneous layers besides migrating keratinocytes. In addition, CMV-driven reporter gene expression was prominent in transduced intact skin. Together these results show that the function and specificity of the FiRE element is retained when embedded in a viral vector, in contrast to many other tissue or cell specific regulatory elements.²⁸

Based on the results of the present study, it can be proposed, that the FiRE element can be used to target gene expression to migrating keratinocytes in different approaches to gene therapy of wound repair. For example, the genes encoding for EGF or KGF, both known to enhance keratinocyte proliferation and migration,^29,30 could be delivered to cultured cells ex vivo, followed by grafting of cells to the wound site. Thereby FiRE could create a positive autoregulatory loop to produce large quantities of the particular growth factor. Also PDGF, which is down-regulated during impaired wound healing,³¹ could be used to stimulate granulation tissue formation and enhance the survival and growth of keratinocytes in skin grafts.²⁴ In addition, FiRE could be used to transcriptionally target therapeutic transgene expression to the edges of autologous transplants and xenotransplants, to provide additional stimuli for keratinocyte migration and adhesion. Obviously, transient expression of, for example EGF or KGF, persisting until the transplant has attached, would be desirable, and should be achieved by using FiRE-driven adenoviral transgene expression.

Finally, FiRE could be used to deliver therapeutic polypeptides into poorly healing wounds of, for example diabetic patients or patients receiving immunosuppressive treatment. A number of candidate genes that would be beneficial for wound healing, depending on the primary defect in the healing process, have been proposed. Besides the epidermally acting growth factors, these could include, for example the inducible nitric oxide synthase, which has been shown to reverse impaired wound healing when delivered within an adenovirus.⁶ It remains to be studied whether the RAdFiRE would function as an efficient gene delivery vector in the therapy of chronic wounds. However, it needs to be mentioned that in chemically induced chronic ulcerations in the FIRE-syndecan promoter transgenic mice, FiRE is activated and remains active for long periods (up to 6 months) at the wound edge (P Jaakkola, unpublished observations).

As discussed above, transient expression of therapeutic genes until the closure of the wounds, would be desirable in gene therapy of wound repair. In keeping with this, we chose to test the function of FiRE in adenoviral vector, since these are known to provide an efficient and transient transgene expression in skin in vivo.⁷ However, FiRE could also be used as a means for transcriptional targeting using other methods for gene delivery, for example nonviral and retroviral gene transfer. These systems should offer the advantage of producing less inflammatory reaction compared with adenoviruses. However, nonviral gene delivery has been demonstrated to be a less efficient way of gene transfer to keratinocytes compared with adenoviruses.⁷ Retroviral vectors on the other hand would be predicted to target merely the proliferating keratinocytes, especially the epidermal stem cells, but not migrating keratinocytes.³² Nonviral methods may still provide a useful way for transfecting cultured keratinocytes followed by transplantation.²⁵

The results of the present study show, that FiRE offers a powerful tool for targeting and controlling gene expression in cultured keratinocytes, as well as in the ex vivo maintained skin. However, when transduced to the wound sites in vivo, the expression is clearly lower than in transduction of ex vivo wounds. This is equally true for the CMV promoter driven expression noted in RAdlacZ infected wounds. This can be due to lower transduction efficiency of wound cells in vivo and could be circumvented by using higher viral dose or multiple applications of the virus. Both these adenoviruses cause an inflammatory reaction, which may also be a reason for attenuation of transgene expression in the wound in vivo. It is possible that adenoviral transduction of wound cells enhances the inflammatory phase of wound repair, as adenoviral transduction of healthy intact human skin in vivo results only in mild or moderate inflammation.³³ Next generation adenoviral vectors are predicted to cause much less inflammation and therefore, might also increase the level of cell specific transgene expression in vivo.³⁴ Obviously, when available, the ability of these vectors to carry functional FiRE driven genes should be tested.

In conclusion, the results of this study demonstrate that transcriptional targeting of adenovirus-mediated genes to migrating epidermal keratinocytes is possible, efficient and can be tightly controlled using FiRE. Therefore, the results of this study support our hypothesis of using the FiRE as a tool in developing targeted gene therapy for cutaneous wounds.

Materials and methods

Construction of RAdFiRE-EGFP

For construction of recombinant replication deficient (E1-/E3-) adenovirus RAdFiRE-EGFP the XhoI and EcoRI EGFP cDNA fragment from pIRES-EGFP (Clontech, Palo Alto, CA, USA) was subcloned into shuttle vector pE1sp1a (Microbix, Biosystems, Toronto, ON, Canada) under the control of CMV promoter creating plasmid pE1-(CMV)IRES-EGFP. The CMV promoter together with IVS and IRES sequences from pE1-(CMV)IRES-EGFP plasmid was removed by SalI-HindIII digestion. A SalI-XbaI/Blunt fragment from pFiRESynProm plasmid used for generating transgenic mice¹⁸ was ligated to SalI-HindIII/Blunt site of pE1-(CMV)IRES-EGFP. The XbaI/Blunt fragment from pFiRESynProm contains the 350 bp FiRE element next to 2.2 kb of syndecan-1 proximal promoter which ranges from -115 bp to -2163 bp from syndecan-1 translation initiation site.³⁵

Adenoviral genomic plasmid pHBG10 (Microbix Biosystems) and the shuttle vector containing the FiRE syndecan promoter-EGFP expression cassette were cotransfected into 293 cells (Microbix Biosystems) with CalPhosMaximizer kit (Clontech) according to the manufacturer's instructions. After 2 weeks plaques were visible and cell layer was subsequently harvested in PBS containing 10% glycerol and viruses were released from cells with freon extraction and subjected to plaque purification in 96-well plates. Positive recombinants were identified with PCR using recombinant clone viral DNA as template with oligonucleotide primers 5'-AGAGGAGACAGAGCCTAAC-3' from the syndecan-1 gene and 5'-TCTTTGCTCAGGGCGGAC TG-3' from the EGFP gene. The expression of EGFP reporter gene by the PCR positive recombinant adenovirus clones was verified by examination of infected 293 cells by fluorescence microscopy. One EGFP positive clone of recombinant adenovirus was named RAdFire-EGFP and was chosen to generate high titer preparation by freon extraction, cesium chloride banding and dialysis.³⁶ Determination of viral titer was conducted as described previously.³⁷ Recombinant replication-deficient adenovirus RAdlacZ (RAd35),³⁸ which contains the Escherichia coli - galactosidase (lacZ) gene under the control of CMV IE promoter was kindly provided by Dr Gavin WG Wilkinson (University of Cardiff, Wales).

Adenovirus infections of cultured keratinocytes and ex vivo skin wounds

MCA3D cells were routinely cultured in Ham's F-12 medium (Gibco BRL, Paisley, UK) supplemented with 10% fetal calf serum (FCS). For adenoviral transduction, 1.5 ´ 10⁵ cells were plated and incubated with RAdFiRE-EGFP (1000 p.f.u. per cell) for 16 h in serum-free culture medium, as described previously.³⁹ The media were then changed, EGF (10 ng/ml; PeproTech, London, UK) was added and the incubations continued for 48 h. For quantitation of EGFP expression in MCA3D keratinocytes the cells were photographed under UV microscope, the photographs were scanned at 150 dots per inch and imported into an image analyzer system of the Microcomputer Image Device system (MCID, Imaging Research, St Catherines, ON, Canada). The area covered by green color in a given field was quantified.

For organ culture studies, mice were killed and after hair removal, full thickness linear wounds were cut to their backs or tails. Wounded tissues were removed with a margin of approximately 1 cm and allowed to float in cell culture conditions in DMEM supplemented with 10% FCS in a 24-well or six-well plate. Immediately following transfer of tissues to organ culture adenoviruses (7.5 ´ 10⁷ p.f.u. to 6 ´ 10⁸ p.f.u.) were added to the culture medium, and the tissues were harvested at indicated time-points, frozen or fixed in formalin. For EGFP detection from sections native frozen sections were examined under UV microscope. For -galactosidase staining, tissues were fixed in 4% formalin and dehydrated in ascending concentrations of ethanol and embedded in paraffin. Tail pieces were treated with 4% formic acid for 2-4 days before embedment. Microtome sections (5 M) were mounted on glass slides treated with poly-L-lysine. Standard techniques were used for hematoxylin-eosin staining. X-gal staining was performed overnight with 1 mg/ml concentration, as described previously.¹⁶

Adenovirus infections of cutaneous wounds in vivo

For in vivo wound healing studies 3- to 9-month-old mice from (C57Bl/6 ´ DBA/2) strain of both sexes were anesthetized by 2.5% avertin, the hair was removed and full thickness incisional wounds were made on the back or on the tails of the mice with a scalpel. Adenoviruses RAdFiRE-EGFP and RAdlacZ (1.1 ´ 10⁹ p.f.u. each) in 100-300 l of PBS were inoculated on top of the wounds. In certain experiments adenoviruses were inoculated to the site of wound before the incision. The wounds were either left uncovered or covered with plastic. At the indicated time-points the mice were killed, the wounded skin was removed followed by production of histological sections. The experimental procedures were approved by the Animal Test Review Committee of the University of Turku. The production of histological sections was performed as described above.

Acknowledgements

The authors are grateful to Mrs Hanna Haavisto, Mrs Taina Kalevo-Mattila and Mrs Anni Kieksi for expert technical help. Professor Seppo Ylä-Herttuala is acknowledged for fruitful discussions. This work was financially supported by the Academy of Finland, the Technical Research Center of Finland (TEKES), the Sigrid Jusélius Foundation, the Finnish Cancer Union, Turku University Central Hospital, Diabetes Research Foundation of Finland, Turku Graduate School of Biomedical Sciences, the Maud Kuistila Memorial Foundation (to PJ). PJ is Junior Research Fellow of the Academy of Finland.

1 Singer AJ, Clark RAF. Cutaneous wound healing. New Engl J Med 1999; 341: 738-746, MEDLINE

2 Brown GL et al. Enhancement of wound healing by topical treatment with epidermal growth factor. New Engl J Med 1989; 321: 76-79,

3 Davidson JM, Broadley KN. Manipulation of the wound healing process with basic fibroblast growth factor. Ann NY Acad Sci 1991; 638: 306-315,

4 Staiano-Coico L et al. Human keratinocyte growth factor effects in a porcine model of epidermal wound healing. J Exp Med 1993; 178: 865-878, MEDLINE

5 Danilienko DM et al. Growth factors in porcine full and partial thickness burn repair. Differing targets and effects of keratinocyte growth factor, platelet-derived growth factor, and neu differentiation factor. Am J Pathol 1995; 147: 1261-1277,

6 Yamasaki K et al. Reversal of impaired wound repair in iNOS-deficient mice by topical adenoviral-mediated iNOS gene transfer. J Clin Invest 1998; 101: 967-971,

7 Lu B et al. Topical application of viral vectors for epidermal gene transfer. J Invest Dermatol 1997; 108: 803-808, MEDLINE

8 Hengge UR, Walker PS, Vogel JC. Expression of naked DNA in human, pig, and mouse skin. J Clin Invest 1996; 97: 2911-2916, MEDLINE

9 Yu WH et al. Topical gene delivery to murine skin. J Invest Dermatol 1999; 112: 370-375, Article MEDLINE

10 Sawamura D et al. In vivo gene introduction into keratinocytes using jet injection. Gene Therapy 1999; 6: 1785-1787, MEDLINE

11 Ghazizadeh S, Harrington R, Taichman L. In vivo transduction of mouse epidermis with recombinant retroviral vectors: implications for cutaneous gene therapy. Gene Therapy 1999; 6: 1267-1275,

12 Siders WM, Halloran PJ, Fenton RG. Transcriptional targeting of recombinant adenoviruses to human and murine melanoma cells. Cancer Res 1996; 56: 5638-5646, MEDLINE

13 Sandig V et al. HBV-derived promoters direct liver-specific expression of an adenovirally transduced LDL receptor gene. Gene Therapy 1996; 3: 1002-1009, MEDLINE

14 Jeng MH et al. Reconstitution of estrogen-dependent transcriptional activation of an adenoviral target gene in select regions of the rat mammary gland. Endocrinology 1998; 139: 2916-2925,

15 Kim S et al. Transcriptional targeting of replication-defective adenovirus transgene expression to smooth muscle cells in vivo. J Clin Invest 1997; 100: 1006-1014, MEDLINE

16 Jaakkola P, Vihinen T, Määttä A, Jalkanen M. Activation of an enhancer on the syndecan-1 gene is restricted to fibroblast growth factor family members in mesenchymal cells. Mol Cell Biol 1997; 17: 3210-3219,

17 Jaakkola P, Määttä A, Jalkanen M. The activation and composition of FiRE (an FGF-inducible response element) differ in a cell-type and growth factor-specific manner. Oncogene 1998; 17: 1279-1286,

18 Jaakkola P et al. Wound re-epithelialization activates a growth factor-responsive enhancer in migrating keratinocytes. FASEB J 1998; 12: 959-969,

19 Stoscheck CM, Nanney LB, King LEJ. Quantitative determination of EGF-R during epidermal wound healing. J Invest Dermatol 1992; 99: 645-649,

20 Galiano RD et al. Interaction between the insulin-like growth factor family and the integrin receptor family in the tissue repair processes. Evidence in a rabbit ear dermal ulcer model. J Clin Invest 1996; 98: 2462-2468,

21 Goldfinger LE et al. The 3 laminin subunit, 64 and 31 integrin coordinately regulate wound healing in cultured epithelial cells and in the skin. J Cell Sci 1999; 112: 2615-2629,

22 Liechty KW et al. Adenoviral-mediated overexpression of platelet-derived growth factor-B corrects ischemic impaired wound healing. J Invest Dermatol 1999; 113: 375-383,

23 Eming SA et al. Genetically modified human keratinocytes overexpressing PDGF-A enhance the performance of a composite skin graft. Hum Gene Ther 1998; 9: 529-539, MEDLINE

24 Eming SA et al. Particle-mediated gene transfer of PDGF isoforms promotes wound repair. J Invest Dermatol 1999; 112: 297-302,

25 Rio MD et al. Nonviral transfer of genes to pig primary keratinocytes. Induction of angiogenesis by composite grafts of modified keratinocytes overexpressing VEGF driven by a keratin promoter. Gene Therapy 1999; 6: 1734-1741, MEDLINE

26 Benn SI et al. Particle-mediated gene transfer with transforming growth factor-1 cDNAs enhances wound repair in rat skin. J Clin Invest 1996; 98: 2894-2902,

27 Fukumura D et al. Tumor induction of VEGF promoter activity in stromal cells. Cell 1998; 94: 715-725, MEDLINE

28 Miller N, Whelan J. Progress in transcriptionally targeted regulatable vectors for genetic therapy. Hum Gene Ther 1997; 8: 803-815, MEDLINE

29 Bennet N, Schultz G. Growth factors and wound healing: biochemical properties of growth factors and their receptors. Am J Surg 1993; 165: 728-737,

30 Tsuboi R et al. Keratinocyte growth factor (FGF-7) stimulates migration and plasminogen activator activity of normal human keratinocytes. J Invest Dermatol 1993; 101: 49-53,

31 Beer HD, Longaker MT, Werner S. Reduced expression of PDGF and PDGF receptors during impaired wound healing. J Invest Dermatol 1997; 109: 132-138,

32 Kolodka TM, Garlick JA, Taichman LB. Evidence for keratinocyte stem cells in vitro: long-term engraftment and persistence of transgene expression in retrovirus-transduced keratinocytes. Proc Natl Acad Sci USA 1998; 95: 4356-4361, MEDLINE

33 Harvey BG et al. Cellular immune response of healthy individuals to intradermal administration of an E1-E3- adenovirus gene transfer vector. Hum Gene Ther 1999; 10: 2823-2837, Article MEDLINE

34 Amalfitano A. Next-generation adenoviruses: new and improved. Gene Therapy 1999; 6: 1643-1645, MEDLINE

35 Vihinen T et al. Functional characterization of mouse syndecan-1 promoter. J Biol Chem 1996; 271: 12532-12541,

36 Graham FL, Prevec L. Methods for construction of adenovirus vectors. Mol Biotechnol 1995; 3: 207-220, MEDLINE

37 Lu Y, Zhang Y, Steiner MS. Efficient identification of recombinant adenoviruses by direct plaque screening. DNA Cell Biol 1998; 17: 643-645,

38 Wilkinson GW, Akrigg A. Constitutive and enhanced expression from the CMV major IE promoter in a defective adenovirus vector. Nucleic Acids Res 1992; 20: 2233-2239, MEDLINE

39 Ahonen M, Baker AH, Kähäri V-M. Adenovirus-mediated gene delivery of tissue inhibitor of metalloproteinases-3 inhibits invasion and induces apoptosis in melanoma cells. Cancer Res 1998; 58: 2310-2315, MEDLINE

Figure 1 EGF activates the FiRE-driven EGFP expression in adenovirally transduced keratinocytes. (a) Schematic model of the FiRE-syndecan core promoter-EGFP reporter gene expression cassette embedded in adenovirus RAdFiRE-EGFP. (b) Murine epidermal keratinocytes (MCA3D) were transduced with RAdFiRE-EGFP (1000 p.f.u. per cell) and subsequently incubated without (control) or with epidermal growth factor (EGF; 10 ng/ml) for 48 h. The expression of EGFP is detectable as green color under fluorescence microscope. Magnification ´ 100. (c) Quantitation of the area of EGFP expression in MCA3D keratinocyte monolayers incubated without (control) or with epidermal growth factor (EGF; 10 ng/ml) for 48 h. The results indicate the total EGFP-positive cell population in a given field of nearly confluent cells. Notably, the results do not take into account the stronger intensity of EGFP expression in EGF stimulated cells. Means and standard deviations of three independent experiments are shown.

Figure 2 Expression of EGFP in the wound margins of ex vivo maintained skin transduced with RAdFiRE-EGFP. Linear wound incised to mouse skin (a) or tail (b) was excised immediately, maintained ex vivo in organ culture and transduced with RAdFiRE-EGFP (3 ´ 10⁸ p.f.u.). (a) The expression of EGFP is detected in the margin of the cutaneous wound as green color under fluorescence microscope 2 days after transduction (left panel). The same wound, photographed under white light, is shown in the right side panel. Magnification ´ 30. (b) In tail wounds the expression of EGFP is detectable under fluorescence microscope 3 and 10 days after transduction. Only the other wound margin is shown. Magnification ´ 100. The asterisks indicate the site of incisional wound.

Figure 3 FiRE-driven expression of EGFP in migrating keratinocytes of cutaneous wounds transduced ex vivo with RAdFiRE-EGFP. Linear incisional wound in mouse skin was excised immediately, maintained ex vivo in organ culture, and transduced with RAdFiRE-EGFP (3 ´ 10⁸ p.f.u.). EGFP expression was examined by fluorescence microscopy in frozen sections (left-hand panels) at the day of transduction (0 d) and 3 days after transduction (3 d). White light photographs of the corresponding fields are shown in the right-hand panels. The dashed line marks the epidermal (E)-dermal (D) junction. The arrows indicate the site of the migrating wound tip.

Figure 4 CMV-driven lacZ expression in ex vivo cutaneous wounds transduced with RAdlacZ is non-specific. Linear wounds were incised to mouse skin maintained in organ culture and transduced with CMV-lacZ adenovirus (RAdlacZ) (3 ´ 10⁸ p.f.u.). Expression of lacZ was visualized 3 days after transduction by X-gal staining of formalin fixed sections and is indicated as blue color in keratinocytes and fibroblasts widely in wound area and in transduced intact skin, whereas non-transduced wound (control) is negative.

Figure 5 FiRE-driven EGFP expression in migrating keratinocytes in cutaneous wounds transduced with RadFiRE-EGFP in vivo. Linear wounds were incised to murine skin and transduced with (a) RAdFiRE-EGFP or (b) RAdlacZ (1.1 ´ 10⁹ p.f.u. each). Three days after transduction the expression of EGFP was detected in native frozen sections by UV microscopy and the lacZ expression by X-gal staining in formalin-fixed sections. (a) In RadFiRE-EGFP infected wounds EGFP expression is restricted to wound edge keratinocytes. White light photograph of the native frozen section is shown in the right-hand panel. The line marks the epidermal (E)-dermal (D) junction. The arrows indicate the site of the migrating wound tip. (b) Two representative areas of RAdlacZ-infected wounds with -galactosidase expression are shown. The arrows point to the migrating wound tip and the asterisk marks the wound clot.