- Email: [email protected]
Intravenously Injected Human Fibroblasts Home to Skin Wounds, Deliver Type VII Collagen, and Promote Wound Healing
original article
& The American Society of Gene Therapy
Intravenously Injected Human Fibroblasts Home to Skin Wounds, Deliver Type VII Collagen, and Promote Wound Healing David T Woodley1, Jennifer Remington1, Yi Huang1, Yingping Hou1, Wei Li1, Douglas R Keene2 and Mei Chen1 1 Department of Dermatology, The Keck School of Medicine, University of Southern California, Los Angeles, California, USA; 2Department of Biochemistry and Molecular Biology, Shriners Hospital for Children, Portland, Oregon, USA
Patients with dystrophic epidermolysis bullosa (DEB) have incurable skin fragility, blistering, and multiple skin wounds because of mutations in the gene that encodes for type VII collagen (C7), which holds together the epidermal and dermal layers of human skin. The intradermal injection of gene-corrected DEB fibroblasts, recombinant C7 protein, or lentiviral vectors expressing C7 is a potential therapy for DEB. Nevertheless, severe DEB causes widespread wounds and treatment would require multiple injections. An alternative strategy might be to inject genetically engineered cells into the patient’s circulation that home to the skin wounds and deposit the transgene product. In this study, we demonstrated that intravenously (IV) injected, molecularly engineered DEB fibroblasts (overexpressing human C7) homed to murine skin wounds and continuously delivered C7 at the wound site, where it incorporated into the skin’s basement membrane zone and formed anchoring fibril structures. Wounds made on murine or grafted human skin demonstrated accelerated healing when the animals were IV injected with gene-corrected DEB fibroblasts. Our data demonstrate that abundant C7 promotes wound healing. This is also the first evidence that IV injected, molecularly engineered skin fibroblasts can deliver C7 to skin wounds. This strategy could be useful for treating DEB patients. Received 19 May 2006; accepted 2 October 2006; published online 23 January 2007. doi:10.1038/sj.mt.6300041
INTRODUCTION Healing human skin wounds is a major medical problem worldwide particularly in the elderly patient population. According to the Wound Healing Society, about 15% of older adults suffer from chronic, hard-to-heal wounds.1 It is estimated that 18% of diabetic patients over the age of 65 years will have chronic, non-healing skin ulcers.2 Although the topical application of growth factors has been envisioned to improve skin
wound healing,3–5 overall this strategy has been disappointing. The topical application of epidermal growth factor has been shown to accelerate wound closure of acute human wounds.3 Nevertheless, topical epidermal growth factor never became a commercially viable therapy because of cost and practical considerations. Only platelet-derived growth factor has been approved by the US Food and Drug Administration for nonhealing diabetic foot ulcers, but practitioners find its application limited and not always successful. One problem with the topical application of growth factors is that the wound bed is often laden with proteolytic enzymes that degrade and nullify the applied agent. A more sustained presence of the therapeutic agent may be desirable. In this regard, a molecular approach in which genetically engineered cells home to the wound and deliver a wound-healing agent continuously at that site may overcome the limitations associated with topically applied growth factors. Children who have dystrophic forms of epidermolysis bullosa (DEB) are born with skin fragility, blistering, and repeated wounding and healing of their skin.6 Their wounds typically heal with fibrosis, scarring, and small epidermal inclusion cysts called milia. Their skin poorly adheres to the underlying dermal connective tissue layer and the slightest trauma causes epidermal–dermal disadherence. The disadherence is due to a defect in the gene that encodes for type VII (anchoring fibril) collagen (C7), which serves to anchor the epidermis onto the dermis.7,8 C7 at the epidermal–dermal junction (EDJ) forms anchoring fibrils, 200–700 nm structures that emanate down perpendicularly from the EDJ into the papillary dermis. The EDJ of patients with DEB is characterized by a paucity of normal anchoring fibrils. C7 is composed of three identical alpha chains, each consisting of a 145-kDa central collagenous triple-helical segment, flanked by a large 145-kDa amino-terminal, noncollagenous domain (NC1), and a small 34-kDa carboxylterminal non-collagenous domain (NC2).9,10 Within the extracellular space, C7 molecules form antiparallel dimers which aggregate laterally to form anchoring fibrils.
Correspondence: Mei Chen, Department of Dermatology, The USC Laboratories for Investigative Dermatology, Room 204, Cancer Research Laboratory, 1303 N Mission Road, Los Angeles, California 90033, USA. E-mail: [email protected]
628
www.moleculartherapy.org vol. 15 no. 3, 628–635 march. 2007
& The American Society of Gene Therapy
There are several recent studies related to ex vivo gene therapy for DEB. Oritz-Urda et al. recently used a phi C31 integrasebased non-viral gene transfer approach to integrate stably COL7A1 cDNA into C7-null keratinocytes from recessive DEB (RDEB) patients. Using gene-corrected cells in a human skin equivalent, which was transplanted onto severe combined immunodeficient mice, they showed that many of the RDEB features were corrected after gene transfer.11 We developed a minimal lentiviral vector to express C7 in RDEB keratinocytes and fibroblasts (in which C7 was absent) and demonstrated the reversion of the RDEB cellular phenotype.12 We then used these gene-corrected RDEB cells to create a composite human skin equivalent, which was transplanted onto severe combined immunodeficient mice. The transplanted human skin exhibited C7 at the EDJ, and the transgene-derived C7 created anchoring fibril structures that were correctly organized into the basement membrane zone (BMZ). An ex vivo approach requires transplantation of genecorrected cells onto surgically prepared sites of the patient’s skin. The experience of using cultured keratinocyte autografts for transplantation onto human wounds has shown that this technology is often fraught with technical difficulties and poor graft take. We and others have also developed more straightforward direct in vivo gene therapy approaches to correct DEB. We showed that the intradermal injection of gene-corrected RDEB fibroblasts (so-called ‘‘cell therapy’’), recombinant human C7 (so-called ‘‘protein therapy’’), or lentiviral vectors expressing human C7 (so-called ‘‘vector therapy’’) into mouse skin or a DEB human skin equivalent engrafted onto a mouse achieves long-term expression of C7 that incorporates into the BMZ and reverses RDEB disease features, including dermal–epidermal separation and anchoring fibril defects.13–16 Severe RDEB causes widespread lesions and multiple wounds spanning large areas of trauma prone sites such as the sacrum, hips, feet, mouth, scalp, lower back, and hands. Therefore, treatment via intradermal injection would require numerous injections into multiple wound sites. An alternative strategy might be to inject genetically engineered cells into the patient’s circulation that home to the skin wounds and deposit the transgene product. In the study reported here, we demonstrated that IV injected normal human fibroblasts and gene-corrected RDEB fibroblasts (overexpressing human C7) home to a mouse’s wound site, continuously deliver C7 to the BMZ of the skin, and form anchoring fibril structures. Surprisingly, we found that C7 delivered to the wound sites significantly enhanced wound healing. Our data identified a previously unrecognized use of C7 to promote wound healing. This study provides the first demonstration of the potential use of IV injected gene-corrected DEB fibroblasts to restore C7 in DEB patients who have multiple open wounds.
RESULTS Restoration of C7 expression in RDEB fibroblasts We previously developed a minimal lentiviral transfer vector to express full-length human C7 cDNA and used it to transduce fibroblasts from RDEB patients.12 As shown in the Western blot Molecular Therapy vol. 15 no. 3, march. 2007
Intravenously Injected Human Fibroblasts
experiment depicted in Figure 1a, C7 is secreted into the conditioned media of normal human dermal fibroblasts, but not in cells from a severe RDEB patient. However, the gene-corrected RDEB cells expressed higher levels of C7 than normal fibroblasts (compare lane 2 with lane 3). Figure 1b shows immunofluorescence staining of these cells using an affinity-purified antibody to the NC1 domain of C7. Consistent with the immunoblot data, normal human fibroblasts displayed significantly lower levels of C7 than gene-corrected RDEB cells. As expected, there was no immunostaining of C7 in parental RDEB fibroblasts.
IV injected normal or gene-corrected RDEB fibroblasts delivered human C7 to the mouse’s regenerated BMZ We used an animal model to evaluate the feasibility of IV injecting gene-corrected RDEB fibroblasts for DEB treatment. We made a standard 1 1 cm full-thickness wound on the back of nu/nu athymic hairless mice. After 8–10 h, we IV injected the mice with normal human fibroblasts or gene-corrected human RDEB fibroblasts capable of synthesizing and secreting C7. The controls for this experiment were no injection of cells and the injection of RDEB fibroblasts that did not express C7. Skin biopsies from wounded and unwounded areas of the mouse’s skin were obtained and subjected to indirect immunofluorescence using an antibody that recognizes only human C7, but not rodent. As shown in Figure 2A, in wounded animals, the injected gene-corrected RDEB fibroblasts homed to the wound site and deposited human C7 into the EDJ of the mouse’s skin as early as 2 weeks after injection (Figure 2A, panel a). Identical results occurred when normal dermal fibroblasts were injected (Figure 2A, panel b), although with less deposition of human C7 at the healed BMZ (compare panel a with panel b). In contrast, there was no detectable human C7 in unwounded skin (Figure 2A, panel c) or in mice that were not wounded before injection
Figure 1 Expression of C7 in the cultured human RDEB fibroblasts used for IV injection. (a) Conditioned media from RDEB fibroblasts (lane 1), gene-corrected RDEB fibroblasts (lane 2), and normal human fibroblasts (lane 3) were concentrated and subjected to 6% sodium dodecyl sulfate–polyacrylamide gel electrophoresis followed by immunoblot analysis using a polyclonal antibody to NC1. The positions of fulllength 290 kDa C7 and molecular weight markers are indicated. (b) Immunohistochemistry with an affinity-purified polyclonal antibody to the NC1 domain of C7. Note the very high efficiency of C7 gene transfer to RDEB cells which, in their native state, cannot express C7. RDEB, parental RDEB fibroblasts; RDEB/C7, RDEB fibroblasts transduced with lentiviral expression vector for C7; NFB, normal human fibroblasts.
629
Intravenously Injected Human Fibroblasts
& The American Society of Gene Therapy
Figure 2 IV injected gene-corrected RDEB fibroblasts delivered human C7 to the mouse’s regenerated BMZ. A: (a–h) Immunofluorescence staining of the mouse’s skin after IV injection of various types of human fibroblasts was performed with antibodies specific for human C7 at 2 weeks after the injection. Note that the gene-corrected RDEB fibroblasts (a) and normal human fibroblasts (b) synthesized and delivered human C7 to the mouse’s regenerated BMZ in wounded sites, but not in unwounded sites (c). Further, there was no human C7 expression in mice (n ¼ 10 mice) that were not wounded before injection with either gene-corrected RDEB fibroblasts (d) or normal human fibroblasts (e). Healed wounds of mice injected with parental RDEB cells entirely lacked human C7 (f). (g and h) show the stable expression of human C7 at the mouse’s BMZ at 8 weeks after the injection with gene-corrected RDEB fibroblasts and normal human fibroblasts respectively. e, epidermis; d, dermis. B: (i–k) Dose-dependent deposition of human C7 at the mouse’s BMZ after IV injection with gene-corrected RDEB fibroblasts. Immunofluorescence staining with an antibody specific for human C7 was performed on healed mouse skin wounds after the animals were injected with (i) 0.75, (j) 1.5, and (k) 3 106 genecorrected RDEB fibroblasts, respectively. Note the dose-dependent increase in the deposition of human C7 at the mouse BMZ. C: Immunofluorescence staining of mouse skin was performed 3 weeks after the animals were IV injected with gene-corrected RDEB fibroblasts (RDEB/C7) or normal human fibroblasts (NFB). The skin was labeled with either a monoclonal antibody specific for human C7 (green, panel a-H) or a rabbit polyclonal antibody that recognizes both mouse and human C7 (red, panel a-H þ M). Merged images demonstrate co-localization of human C7 with mouse C7 in the mouse’s BMZ. RDEB/C7, gene-corrected RDEB fibroblasts (n ¼ 40 mice); NFB, normal fibroblast (n ¼ 20 mice); RDEB/FB, parent RDEB fibroblasts (n ¼ 20 mice).
(Figure 2A, panels d and e). Also, as expected, if uncorrected RDEB cells were injected, no human C7 was detected in the mouse’s wounded skin (Figure 2A, panel f). Further, the human fibroblast-derived C7 incorporated into the mouse’s BMZ and was sustained for at least 8 weeks, the longest duration tested in this study (Figure 2A, panels g and h). After demonstrating the ability of the injected fibroblasts to home to skin wounds and deliver human C7 to the mouse’s regenerated BMZ, we wished to determine the minimal concentration of fibroblasts required for C7 deposition at the BMZ. We injected 0.75, 1.5, and 3 106 gene-corrected DEB fibroblasts IV into the tail of nude mice. As shown in Figure 2B, 630
the injected fibroblasts mediated a dose-dependent increase in the deposition of human C7 into the mouse’s BMZ. To confirm that the human C7 derived from the injected cells was truly localized at the BMZ of the healing mouse skin, we colabeled the same vertical sections with a polyclonal antibody that recognizes both mouse and human C7 (Figure 2C, (a-H þ M) and our human-specific C7 antibody (a-H). As shown in Figure 2C, these two antibodies showed perfect co-localization in the mouse’s BMZ when the images were merged. Therefore, the IV injected normal and gene-corrected RDEB fibroblasts synthesized and delivered human C7 to the same location in the BMZ as the mouse’s endogenous C7. www.moleculartherapy.org vol. 15 no. 3, march. 2007
& The American Society of Gene Therapy
Tissue distribution of IV injected gene-corrected RDEB fibroblasts To determine whether the injected human fibroblasts became resident cells within the mouse’s wound, we stained healing murine skin with an antibody that recognizes human, but not murine, fibroblasts. As shown in Figure 3a, double staining for both human dermal fibroblasts and human C7 demonstrated that these IV injected human cells, residing now in the neodermis of the mouse’s healing wound, expressed human C7 that incorporated into the mouse’s BMZ. Further, IV injected gene-corrected RDEB fibroblasts were stably retained in the dermis of the wound tissues for 6 weeks after injection, where they continuously expressed C7. To determine whether these IV injected human fibroblasts also trafficked to tissues other than skin, we performed immunostaining on tissue sections from brain, kidney, liver, lung, spleen, heart, and small intestine with our antibody specific for human fibroblasts. As shown in Figure 3b, the cells were readily observed in healing skin, but not in brain, kidney, liver, spleen, heart, or small intestine. However, in approximately half of the IV injected mice, we did detect some injected human fibroblasts in small foci of the animals’ lungs. Nevertheless, we noted no untoward effects of the fibroblast injections in any of the mice. Daily weight, activity, and feeding habits were identical
Intravenously Injected Human Fibroblasts
between experimental and uninjected control groups of mice. Finally, in mice that were not wounded before injection, IV injected human fibroblasts were detected only in some of the animal’s lung and not in any other internal organs or the skin (data not shown). These results indicate that skin wounding is a prerequisite for the injected cells to home to the wound site.
Formation of anchoring fibrils by IV injected genecorrected RDEB fibroblasts Having shown that IV injected human fibroblasts into mice would take up residence in the healing wound and deliver human C7 that incorporated into the mouse’s BMZ, we next sought to determine whether the human collagen would form anchoring fibril structures. We performed immunoelectron microscopy on vertical sections of healing mouse skin using our human-specific antibody to C7. As shown in Figure 4, the human-specific antibody labeled the ends of wheat-stack-shaped structures emanating perpendicularly from the mouse’s BMZ. The immunogold labeling occurred at the ends of anchoring fibrils because the antibody used in this experiment is specifically directed against the NC1 domain, which is located at the ends of C7 antiparallel dimers. These data demonstrate that IV injected gene-corrected RDEB fibroblasts can synthesize and deliver
Figure 3 Tissue distribution of IV injected gene-corrected RDEB fibroblasts. (a) IV injected gene-corrected RDEB fibroblasts were detected in wounded dermal tissue. Three and 6 weeks after IV injection of gene-corrected RDEB fibroblasts, mouse skin was subjected to doubleimmunofluorescence labeling with a monoclonal antibody specific for human C7 (a-HC7) and a fluorescein isothiocyanate -conjugated secondary antibody (green) and an antibody specific for human fibroblasts (a-HFB) and Cy3-conjugated secondary antibodies (red). Shown are images from the same field at different wavelengths of excitation and merged. Representative fields of random sections from one of 40 IV injected mice are shown. Merged images (yellow) showed that injected cells were detected within the dermis of the wounded sites. (b) Tissue distribution of IV injected genecorrected RDEB fibroblasts. Four weeks after IV injection of gene-corrected RDEB fibroblasts, necropsies were performed on the mice (n ¼ 20 mice) and tissue sections obtained from brain, kidney, liver, lung, spleen, heart, small intestine, and healed skin were subjected to immunostaining using an antibody specific for human fibroblasts. Note that the injected fibroblasts were readily detected in the wounded skin and in small foci of lung, but not in any other organs.
Molecular Therapy vol. 15 no. 3, march. 2007
631
Intravenously Injected Human Fibroblasts
& The American Society of Gene Therapy
Figure 4 Immunoelectron microscopy of mouse skin injected with gene-corrected RDEB fibroblasts. Immunogold labeling of mouse skin after IV injection with gene-corrected RDEB fibroblasts was performed using a human-specific anti-C7 antibody and revealed that human C7 delivered via the injected cells incorporated into the mouse’s BMZ and formed anchoring fibrils. Note numerous gold particles decorating the ends of anchoring fibrils. Arrows denote gold particle-labeled NC1 domains of human C7 derived from injected fibroblasts. D, dermis; E, epidermis; HD, hemidesmosome. Bar ¼ 200 nm.
human C7 that forms anchoring fibril structures at the BMZ of healed mouse skin in vivo.
IV injected gene-corrected RDEB fibroblasts promote wound healing We and others have shown that dermal fibroblasts participate in wound healing by synthesizing extracellular matrices and orchestrating a cascade of events to heal cutaneous wounds.17,18 We next assessed whether human C7 delivered to the skin wound by IV injected fibroblasts could promote wound healing. We measured the area of each animal’s open wound at days 0, 7, 9, 11, 14, and 17 after injection. Figure 5 shows representative control and experimental wounds on days 0, 7, 9, 11, 14, and 17 (Figure 5a) and wound size measurements for each time point in 10 duplicate experiments (Figure 5b). The wounds of mice that were IV injected with gene-corrected RDEB fibroblasts overexpressing C7 demonstrated accelerated healing compared with those of mice that were injected with uncorrected RDEB fibroblasts (lacking C7 expression) when observed between days 7 and 11. By day 14, all wounds had closed and no significant differences between control and experimental wounds were observed. The wound-healing rates of mice injected with uncorrected RDEB fibroblasts and uninjected control mice were identical (data not shown). This suggests that uncorrected RDEB fibroblasts do not have an inhibitory effect on wound healing. We also assessed the effect of IV injected fibroblasts on the healing of human skin wounds. We grafted 1.5 1.5 cm human skin onto the back of nude mice using a previously described xenografting system.19 Eight weeks after grafting, the human skin was wounded with an 8 mm punch biopsy tool. The animals were then IV injected with gene-corrected RDEB fibroblasts or controls. As shown in Figure 6, the human skin wounds healed significantly faster in animals that received IV injected genecorrected RDEB fibroblasts compared with the control animals between days 5 and 11. By day 20, all wounds had healed and no detectable differences between control and experimental wounds were observed. With the use of species-specific antibodies (human laminin 5, C7, and human fibroblast antibodies), we confirmed that the cells re-epithelializing the wound and 632
Figure 5 IV injected gene-corrected RDEB fibroblasts promoted wound healing in mouse skin. A 1.0-cm2 (1 1 cm) full-thickness excision wound was made on the mid-back of 8- to 10-week-old athymic nude mice, which were then injected 8 h after wounding with gene-corrected RDEB fibroblasts expressing human C7 or uncorrected RDEB fibroblasts (n ¼ 10 mice per group). (a) Representative days 0, 7, 9, 11, 14, and 17 wounds are shown. Wound sizes were significantly reduced in mice injected with gene-corrected RDEB fibroblasts expressing C7 (RDEB/C7) compared with the mice that were injected with uncorrected RDEB fibroblasts (Control) at 7, 9, and 11 days after wounding. (b) Mean7SD wound size measurements at day 0, 7, 9, 11, 14, and 17 post-wounding (n ¼ 10 mice for each group).
re-populating the wound’s neodermis were of human origin (data not shown). Taken together, these data indicate that C7 delivered to the wound bed via the IV injection of human dermal fibroblasts promotes the closure of skin wounds in both murine and human skin.
DISCUSSION In this study, we used a murine wound-healing model to evaluate the feasibility of IV injecting gene-corrected human RDEB fibroblasts for therapy of DEB. Our studies showed that human RDEB fibroblasts could be molecularly engineered to secrete into the extracellular space of a host’s wounded skin a very large connective tissue molecule, namely, human C7 (MrE900,000). This macromolecule can incorporate correctly into the regenerated BMZ of the mouse’s skin and form anchoring fibril structures. Our data also demonstrated that in order for IV injected human dermal fibroblasts to home to the animal’s skin, the skin must first be wounded. In addition, it appears that IV injected gene-corrected RDEB fibroblasts, but not uncorrected RDEB fibroblasts, dramatically enhanced wound healing in these animals, suggesting the potential use of C7 to promote wound healing. We and others have shown that gene-corrected RDEB keratinocytes alone can be used to restore C7 expression and anchoring fibril formation in vivo.11,12 Although grafting genecorrected epidermal autografts onto RDEB patients may be www.moleculartherapy.org vol. 15 no. 3, march. 2007
& The American Society of Gene Therapy
Figure 6 IV injected gene-corrected RDEB fibroblasts promoted wound healing in human skin grafted onto mice. A 0.5-cm2 (8 mm diameter punch biopsy) full-thickness excision wound was made in human skin grafted onto the mid-back of 8- to 10-week-old athymic nude mice, which were then injected 8 h after wounding with genecorrected RDEB fibroblasts expressing human C7 or without injection (n ¼ 5 mice per group). (a) Representative days 0, 5, 8, and 11 wounds are shown. Wound sizes were significantly reduced in mice injected with gene-corrected RDEB fibroblasts expressing C7 (RDEB/C7) compared with control at 5, 8, and 11 days after wounding. (b) Mean7SD wound size measurements at days 0, 5, 8, 11, 15, and 20 post-wounding (n ¼ 5 mice for each group).
feasible, the procedure may have many technical problems similar to those encountered when transplanting epidermal autografts onto burn wounds. Among these, the degree of autograft take has been a major concern given that approximately 40% of cultured epidermal autografts transplanted onto burn wounds are lost. Also, the grafting procedure requires surgical excision of the patient’s skin down to fascia for optimal autograft take.20–22 Pain, immobilization of the patient, scarring, extensive wound care, and other potential morbidity issues limit the success of this strategy. We and others have recently also developed strategies involving the direct intradermal injection of gene-corrected RDEB fibroblasts, recombinant C7, or lentiviral vectors expressing C7 for the correction of DEB which offers many advantages over grafting thin-cultured epidermal autografts onto wounds.13–16 We and others showed that the intradermal injection of various therapeutic agents including cells, proteins, or lentivectors into intact RDEB skin stably restored C7 expression and anchoring fibrils at the BMZ and corrected the RDEB disease features. However, the small diffusion radius of intradermally injected agents and the inaccessibility of some sites such as the esophagus limit intradermal injection therapies. Patients with generalized, extensive blistering would require Molecular Therapy vol. 15 no. 3, march. 2007
Intravenously Injected Human Fibroblasts
many intradermal injections at multiple sites. These issues pose practical and logistical concerns for the clinical use of intradermal injections for RDEB patients. In this study, we found that IV injected fibroblasts specifically trafficked to wounded skin and synthesized and deposited human C7 at the newly formed BMZ of the host’s skin. It appears that this human fibroblast homing to wound sites is specific to skin wounding because we did not detect human fibroblasts in the liver when mouse liver was injured with carbon tetrachloride (CCl4) and then IV injected with either genecorrected RDEB fibroblasts or normal human fibroblasts (data not shown). We are not sure of the mechanisms underlying these observations. It is possible that wounded skin tissue sends signals to the injected fibroblasts. The cells then respond to the signals by migrating to the damaged sites and helping repair the wound. It has been shown recently that bone marrow-derived stem cells IV injected into wounded mice will home to the wound and participate in skin wound healing.23,24 In a recent study,25 murine green fluorescent protein-labeled bone marrow cells were transplanted into non-green fluorescent protein-positive mice. The green fluorescent protein-labeled cells were observed trafficking through both wounded and unwounded skin of the green fluorescent protein-negative recipients. Likewise, using a chimeric mouse wound model, it was shown that a percentage of cells in the healed dermis of skin wounds came from the bone marrow.26 It is unclear if bone marrow stem cells participate in the healing of naturally occurring human skin wounds. It is possible that local stem cells from nearby unwounded skin participate in the healing of the wound. Using an animal model, a recent study found that stem cells in hair follicles are enlisted to help heal skin wounds.26 Most surprisingly, we observed that the wounds of mice IV injected with gene-corrected RDEB fibroblasts overexpressing C7 demonstrated accelerated wound healing compared with control mice injected with uncorrected RDEB cells. Further, there was no significant difference in the wound-healing rate between the mice injected with uncorrected RDEB cells and those without injection (data not shown). Similarly, C7 delivered to the wound bed via intradermal or IV injection of the recombinant protein also promoted the closure of skin wounds (unpublished observation). In addition, when wounds were made on human skin grafted onto mice, IV injected gene-corrected RDEB fibroblasts also significantly enhanced healing of the human skin. How exogenously delivered C7 promotes healing is unknown. The presence of RGD motifs, fibronectin type III, cartilage matrix protein, and von Willebrand factor homology domains in the C7 molecules would allow C7 to modulate cells and matrix–matrix interactions at the wound site. We have shown previously that C7 promotes keratinocyte and fibroblast matrix attachment and cell motility.12,27 In acute wound healing, C7 staining is detected in the wound bed and the neodermis.28,29 It is possible that C7 may recruit or maintain fibroblasts at this location and promote cellular remodeling of the wound bed. C7 also has specific domains with affinity for other BMZ and dermal molecules such as type IV collagen, laminin-5, fibronectin, and type I collagen.30–32 This may promote the organization and maturation of the connective tissue macromolecules within the 633
Intravenously Injected Human Fibroblasts
BMZ and the high papillary dermis which would be useful for reepithelialization, building granulation tissue, and wound closure. Interestingly, it has been shown that C7 expression was undetectable in two patients’ wounds—one a lower-limb ulcer and the other a non-healing leg ulcer. These patients did not show any clinical evidence of healing.33 Therefore, C7 synthesis and incorporation into anchoring fibrils is probably a prerequisite for the synthesis of a stable basement membrane in healing wounds and is required for healing to proceed. The small number of IV injected fibroblasts detected in foci of lung tissue raise a potential safety issue. Nevertheless, IV injected mice did not exhibit any abnormal behavior. Their activity, eating habits, and growth rate were identical to those of uninjected mice. There was no histological evidence of fibrosis in the lungs of those mice with small foci of injected human fibroblasts (data not shown). Furthermore, when identical experiments were conducted with immunocompetent mice, we did not observe any injected fibroblasts in the lung (data not shown). It is possible that fibroblasts trafficking to the lung will be eliminated by the host’s immune system, which is missing in the immunodeficient mice. Our hope is that fibroblasts IV injected into immune competent DEB patients would be rapidly cleared as well. In summary, our studies demonstrate that IV injected genecorrected RDEB fibroblasts were capable of homing to skin wounds and synthesizing and delivering human C7 that incorporated into the newly regenerated BMZ and promoted wound healing. Defining the potential advantages and drawbacks of IV injected cell-based therapy will be of significant interest for future studies. As the exogenous application of biological products to skin wounds has been very disappointing in terms of promoting wound healing owing to the excess proteolytic enzymes in the wound bed that degrade and nullify the applied agent, we believe that IV delivery of pro-wound healing agents by the method detailed in this study could be used to deliver extracellular matrix molecules and growth factors continuously to healing wounds. For example, physicians could IV inject molecularly engineered fibroblasts that overexpress products that promote various steps in wound healing such as re-epithelialization, fibroplasia, or angiogenesis. In addition to being a therapy for healing chronic wounds, this strategy may be particularly useful for patients with DEB who have multiple open wounds and skin fragility due to a paucity of C7 and anchoring fibrils in their skin.
MATERIALS AND METHODS Cell culture. Normal human fibroblasts were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. Primary RDEB fibroblasts (gift of Dr Scott Herron, Stanford University) and RDEB fibroblasts genetically modified to re-express C7 by lentiviral transduction, as described previously, were cultured in Dulbecco’s modified Eagle’s medium /Ham’s F12 (1:1) supplemented with 10% fetal bovine serum.12 Protein analysis. For immunoblot analysis, normal human dermal
fibroblasts, RDEB fibroblasts, and lentivirus-transduced RDEB fibroblasts were grown to confluence, the medium was changed to serum-free medium containing 150 mm ascorbic acid, and the cultures were maintained for an
634
& The American Society of Gene Therapy
additional 24 h. The media were collected, equilibrated to 5 mM EDTA, 50 mM N-ethylmaleimide, and 50 mM phenylmethylsulfonyl fluoride, concentrated 10- to 15-fold (Centricon-100, Amicon, Beverly, MA) and subjected to 6% sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Proteins were electrotransferred onto a nitrocellulose membrane. The presence of C7 was detected by Western blot analysis using a polyclonal antibody to the NC1 domain of C7.34 Immunofluorescence staining of cells. RDEB or normal human cells
were plated in TissueTek chamber slides (Nunc Inc., Naperville, IL) on polylysine at 371C for 18 h. Cells were immersed in periodate-lysineparaformaldehyde fixative for 10 min at room temperature, washed several times with phosphate buffered saline (PBS) to remove fixative, and then permeabilized and blocked by incubating in PBS with 3% bovine serum albumin, 1% saponin, and 10% normal goat serum for 15 min at room temperature. The cells were incubated with an affinitypurified polyclonal antibody to the NC1 domain of C7 (ref. 34) at a dilution of 1:200 in a humidified chamber for 2 h, washed three times with PBS and 1% saponin, counterstained with a fluorescein isothiocyanate-conjugated goat antibody to rabbit immunoglobulin IgG (1:200 dilution) for 1 h (Organon Teknika-Cappel, MP Biomedicals Inc, Aurora, Ohio) and washed again. The cells were examined and photographed on a Zeiss microscope equipped with an epiluminating fluorescent light source. IV injection of fibroblasts into mice. To prepare mice for injection, we first made a 1.0 1.0-cm full-thickness excision wound by lifting the skin with forceps and removing full-thickness skin using a curved scissors on the mid-back of 8- to 10-week-old athymic nude mice (The Jackson Laboratory, Bar Harbor, ME) and then covered the wound with a Band-Aid and a Coban, self-adherent wrap, to prevent desiccation. Within 8 h after wounding, 1.5 106 gene-corrected RDEB fibroblasts, normal human fibroblasts, and parental uncorrected RDEB fibroblasts in PBS (100 ml PBS with 50 U heparin/ml) were injected into the tail vein of wounded mice using a 30 G needle. We then re-injected mice one more time with 1.5 106 cells 8 h later after the initial injection to increase the number of injected cells. We injected 40 mice with genecorrected RDEB fibroblasts, 20 mice with normal human fibroblasts, and 20 mice with uncorrected parental RDEB fibroblasts. Two to 8 weeks after injection, mouse skin biopsies from wounded or unwounded areas were obtained and subjected to immunostaining using an antibody specific for human C7 (clone LH 7.2; Sigma, St Louis, MO) or a rabbit polyclonal antibody recognizing both human and mouse C7 as described below. To measure the wound size, standardized digital photographs were taken of the wounds at 0, 7, 9, 11, 14, and 17 days after wounding and open wound areas determined with an image analyzer (AlphaEase FC version 4.1.0, Alpha Innotech Corporation, San Leandro, CA) on a Macintosh computer. For evaluating wound healing of human skin, a 1.5 1.5-cm of fullthickness human skin was grafted onto nude mice as described previously.19 Eight weeks after grafting, the human skin was wounded using an 8 mm punch biopsy tool, bandaged, and then injected with gene-corrected RDEB fibroblasts as described above. The assessment of wound healing was performed by area planimetry in the same way as in the assessment of the murine wounds described above. All animal studies were conducted using protocols approved by the University of Southern California Institutional Animal Use Committee. Immunofluorescence staining and ultrastructural analysis of tissue.
Five-micrometer thick sections of the optimal cutting temperature–embedded tissues were cut on a cryostat, fixed for 5 min in cold acetone, and air-dried. Immunolabeling of the tissue was performed using standard immunofluorescence methods as www.moleculartherapy.org vol. 15 no. 3, march. 2007
Intravenously Injected Human Fibroblasts
& The American Society of Gene Therapy
described previously.35 Briefly, for single- and double-immunofluorescence staining, sections were blocked with MOM. Mouse IgG Blocking Reagent (Vector Laboratories, Burlingame, CA) for 1 h at room temperature. Primary antibodies were diluted in PBS with 1% bovine serum albumin. For type C7 staining, we used monoclonal antibodies against human C7, clone LH 7.2 (Sigma) or a rabbit polyclonal antibody recognizing both mouse and human C7.34 For double-immunofluorescence, we incubated the mouse monoclonal anti-human C7 together with the following rabbit or mouse anti-human antibodies: a rabbit polyclonal antibody to both human and mouse C7; a mouse monoclonal antibody to human skin fibroblasts (Clone D7-FIB, Acris Antibodies GmbH, Hiddenhausen, Germany). All primary antibody dilutions were 1:100. After incubation for 1 h at room temperature, sections were washed in PBS three times and stained for 1 h with fluorescein isothiocyanate -conjugated goat antimouse IgG1 with or without Cy3-conjugated goat anti-rabbit IgG (Sigma) or Alex fluor 568 goat anti-mouse-Ig G2a (Molecular Probe, Eugene, OR) diluted 1:500 in PBS with 1% bovine serum albumin. Slides were mounted with 40% glycerol. Photographs of stained sections were taken using a Zeiss Axioplan fluorescence microscope equipped with a Zeiss Axiocam MRM digital camera system.. Immunogold electron microscopy was performed on the mouse skin using a standardized method as described previously.36 To assess human anchoring fibril formation and ultrastructure, 40 mm sections were fixed in 0.1% glutaraldehyde, rinsed in 0.15 M Tris (pH 7.5), then incubated in an antibody specific for human C7 (NP185, a gift of Dr Lynn Sakai, Shriners Hospital for Children, Portland, OR) followed by 1 nm gold secondary antibody and enhancement as described previously.37
11. 12. 13.
14.
15.
16.
17. 18. 19.
20.
21.
22.
23.
24. 25. 26. 27.
28. 29.
ACKNOWLEDGMENTS This work was supported by grants RO1 AR47981 (Dr Chen) and RO1 AR33625 (Dr Woodley) from the National Institutes of Health. We thank Sara Tufa for technical support of immuno-EM.
30.
31.
REFERENCES 1. 2. 3. 4. 5.
6. 7. 8.
9. 10.
Stein, J (2000). Chronic wounds on the rise. Internal Medicine World Report. /http:// 207.21.218.87/news1a.asp?cid=27S. Hess, CT (1999). Care for a diabetic ulcer. Nursing 29: 70–71. Brown, GL et al. (1989). Enhancement of wound healing by topical treatment with epidermal growth factor. N Engl J Med 321: 76–79. Goldman, R (2004). Adv Skin Wound Care 17: 24. Review. Fu, X, Li, X, Cheng, B, Chen, W and Sheng, Z (2005). Engineered growth factors and cutaneous wound healing: success and possible questions in the past 10 years. Wound Repair Regen 13: 122–130. Lin, AN and Carter, DM (eds) (1992). Epidermolysis Bullosa: Basic and Clinical Aspects. Springer-Verlag Inc.: NY. Briggaman, RA and Wheeler Jr, CE (1975). The epidermal–dermal junction. J Invest Dermatol 65: 71–84. Uitto, J and Christiano, AM (1994). Molecular basis for the dystrophic forms of epidermolysis bullosa: mutations in the type VII collagen gene. Arch Dermatol Res 287: 16–22. Burgeson, RE (1993). Type VII collagen, anchoring fibrils, and epidermolysis bullosa. J Invest Dermatol 101: 252–255. Sakai, LY, Keene, DR, Morris, NP and Burgeson, RE (1986). Type VII collagen is a major structural component of anchoring fibrils. J Cell Biol 103: 1577–1586.
Molecular Therapy vol. 15 no. 3, march. 2007
32.
33. 34.
35.
36.
37.
Ortiz-Urda, S et al. (2002). Stable nonviral genetic correction of inherited human skin disease. Nat Med 8: 1166–1170. Chen, M et al. (2002). Restoration of type VII collagen expression and function in dystrophic epidermolysis bullosa. Nat Genet 32: 670–675. Woodley, DT et al. (2003). Normal and gene-corrected dystrophic epidermolysis bullosa fibroblasts alone can produce type VII collagen at the basement membrane zone. J Invest Dermatol 121: 1021–1028. Ortiz-Urda, S, Lin, Q, Green, CL, Keene, DR, Marinkovich, MP and Khavari, PA (2003). Injection of genetically engineered fibroblasts corrects regenerated human epidermolysis bullosa skin tissue. J Clin Invest 111: 251–255. Woodley, DT et al. (2004). Intradermal injection of lentiviral vectors corrects regenerated human dystrophic epidermolysis bullosa skin tissue in vivo. Mol Ther 10: 318–326. Woodley, DT et al. (2004). Injection of recombinant human type VII collagen restores collagen function in dystrophic epidermolysis bullosa. Nat Med 10: 693–695. Singer, AJ and Clark, RAF (1999). Cutaneous wound healing. Rev N Engl J Med 341: 738–746. Li, W, Fan, J, Chen, M and Woodley, DT (2004). Mechanisms of human skin cell motility. Histol Histopathol 19: 1311–1324. Kim, YH, Woodley, DT, Wynn, KC, Giomi, W and Bauer, EA (1992). Recessive dystrohpic bullosa phenotype is preserved in xenografts using SCID mice: development of an experimental in vivo model. J Invest Dermatol 92: 191–197. Gallico III, GG, O’Connor, NE, Compton, CC, Kehinde, O and Green, H (1984). Permanent coverage of large burn wounds with autologous cultured human epithelium. N Engl J Med 311: 448–451. Herzog, SR, Meyer, A, Woodley, DT and Peterson, HD (1988). Wound coverage with cultured autologous keratinocytes: use after burn wound excision, including biopsy followup. J Trauma 28: 195–198. Woodley, DT, Peterson, HD and Herzog, SR (1988). Burn wounds resurfaced by cultured epidermal autografts show abnormal reconstitution of anchoring fibrils. JAMA 259: 2566–2571. Badiavas, EV, Abedi, M, Butmaarc, J, Falanga, JV and Quesenberry, P (2003). Participation of bone marrow derived cells in cutaneous wounding healing. J Cell Physiol 196: 245–250. Badiavas, EV and Falanga, V (2003). Treatment of chronic wounds with bone marrow-derived cells. Arch Dermatol 139: 510–516. Fathke, C et al. (2004). Contribution of bone marrow-derived cells to skin: collagen deposition and wound repair. Stem Cells 22: 812–822. Ito, M et al. (2005). Stem cells in the hair follicle bulge contribute to wound repair but not to homeostasis of the epidermis. Nat Med 11: 1351–1354. Chen, M, O’Toole, EA, Li, YY and Woodley, DT (1999). a2b1 integrin mediates dermal fibroblast attachment to type VII collagen via a 158 amino acid segment of the NC1 domain. Exp Cell Res 249: 231–239. Haapasalmi, K et al. (1995). Expression of epithelial adhesion proteins and integrins in chronic inflammation. Am J Pathol 147: 193–206. Woodley, DT, Peterson, HD and Herzog, SR (1988). Burn wounds resurfaced by cultured epidermal autografts show abnormal reconstitution of anchoring fibrils. JAMA 259: 2566–2571. Chen, M et al. (1997). Interactions of the amino-terminal noncollagenous (NC1) domain of type VII collagen with extracellular matrix components. J Biol Chem 272: 14516–14522. Chen, M, Marinkovich, MP, Jones, JJ, O’Toole, EA and Woodley, DT (1999). NC1 domain of type VII collagen binds to the beta 3 chain of laminin 5 via a unique subdomain within the fibronectin-like repeats. J Invest Dermatol 112: 177–183. Rousselle, P, Keene, DR, Ruggiero, F, Champliaud, MF, Rest, M and Burgeson, RE (1997). Laminin 5 binds the NC-1 domain of type VII collagen. J Cell Biol 138: 719–728. Hopkinson, I, Anglin, IE, Evans, DL and Harding, KG (1997). Collagen VII expression in human chronic wounds and scars. J Pathol 182: 192–196. Chen, M, Petersen, MJ, Li, H-L, Cai, X-Y, O’Toole, EA and Woodley, DT (1997). Ultraviolet A irradiation upregulates type VII collagen expression in human dermal fibroblasts. J Invest Dermatol 108: 125–128. Gammon, WR, Briggaman, RA, Inman III, AQ, Queen, LL and Wheeler, CE (1984). Differentiating anti-lamina lucida and anti-sublamina densa anti-BMZ antibodies by indirect immunofluorescence on 1.0 M sodium chloride-separated skin. J Invest Dermatol 82: 139–144. Keene, DR, Sakai, LY, Lunstrum, GP, Morris, NP and Burgeson, RE (1987). Type VII collagen forms an extended network of anchoring fibrils. J Cell Bio 104: 611–621. Champliaud, MF, Lunstrum, GP, Rousselle, P, Nishiyama, T, Keene, DR and Burgeson, RE (1996). Human amnion contains a novel laminin variant, laminin 7, which like laminin 6, covalently associates with laminin 5 to promote stable epithelial-stromal attachment. J Cell Biol 132: 1189–1198.
635
& The American Society of Gene Therapy
Intravenously Injected Human Fibroblasts Home to Skin Wounds, Deliver Type VII Collagen, and Promote Wound Healing David T Woodley1, Jennifer Remington1, Yi Huang1, Yingping Hou1, Wei Li1, Douglas R Keene2 and Mei Chen1 1 Department of Dermatology, The Keck School of Medicine, University of Southern California, Los Angeles, California, USA; 2Department of Biochemistry and Molecular Biology, Shriners Hospital for Children, Portland, Oregon, USA
Patients with dystrophic epidermolysis bullosa (DEB) have incurable skin fragility, blistering, and multiple skin wounds because of mutations in the gene that encodes for type VII collagen (C7), which holds together the epidermal and dermal layers of human skin. The intradermal injection of gene-corrected DEB fibroblasts, recombinant C7 protein, or lentiviral vectors expressing C7 is a potential therapy for DEB. Nevertheless, severe DEB causes widespread wounds and treatment would require multiple injections. An alternative strategy might be to inject genetically engineered cells into the patient’s circulation that home to the skin wounds and deposit the transgene product. In this study, we demonstrated that intravenously (IV) injected, molecularly engineered DEB fibroblasts (overexpressing human C7) homed to murine skin wounds and continuously delivered C7 at the wound site, where it incorporated into the skin’s basement membrane zone and formed anchoring fibril structures. Wounds made on murine or grafted human skin demonstrated accelerated healing when the animals were IV injected with gene-corrected DEB fibroblasts. Our data demonstrate that abundant C7 promotes wound healing. This is also the first evidence that IV injected, molecularly engineered skin fibroblasts can deliver C7 to skin wounds. This strategy could be useful for treating DEB patients. Received 19 May 2006; accepted 2 October 2006; published online 23 January 2007. doi:10.1038/sj.mt.6300041
INTRODUCTION Healing human skin wounds is a major medical problem worldwide particularly in the elderly patient population. According to the Wound Healing Society, about 15% of older adults suffer from chronic, hard-to-heal wounds.1 It is estimated that 18% of diabetic patients over the age of 65 years will have chronic, non-healing skin ulcers.2 Although the topical application of growth factors has been envisioned to improve skin
wound healing,3–5 overall this strategy has been disappointing. The topical application of epidermal growth factor has been shown to accelerate wound closure of acute human wounds.3 Nevertheless, topical epidermal growth factor never became a commercially viable therapy because of cost and practical considerations. Only platelet-derived growth factor has been approved by the US Food and Drug Administration for nonhealing diabetic foot ulcers, but practitioners find its application limited and not always successful. One problem with the topical application of growth factors is that the wound bed is often laden with proteolytic enzymes that degrade and nullify the applied agent. A more sustained presence of the therapeutic agent may be desirable. In this regard, a molecular approach in which genetically engineered cells home to the wound and deliver a wound-healing agent continuously at that site may overcome the limitations associated with topically applied growth factors. Children who have dystrophic forms of epidermolysis bullosa (DEB) are born with skin fragility, blistering, and repeated wounding and healing of their skin.6 Their wounds typically heal with fibrosis, scarring, and small epidermal inclusion cysts called milia. Their skin poorly adheres to the underlying dermal connective tissue layer and the slightest trauma causes epidermal–dermal disadherence. The disadherence is due to a defect in the gene that encodes for type VII (anchoring fibril) collagen (C7), which serves to anchor the epidermis onto the dermis.7,8 C7 at the epidermal–dermal junction (EDJ) forms anchoring fibrils, 200–700 nm structures that emanate down perpendicularly from the EDJ into the papillary dermis. The EDJ of patients with DEB is characterized by a paucity of normal anchoring fibrils. C7 is composed of three identical alpha chains, each consisting of a 145-kDa central collagenous triple-helical segment, flanked by a large 145-kDa amino-terminal, noncollagenous domain (NC1), and a small 34-kDa carboxylterminal non-collagenous domain (NC2).9,10 Within the extracellular space, C7 molecules form antiparallel dimers which aggregate laterally to form anchoring fibrils.
Correspondence: Mei Chen, Department of Dermatology, The USC Laboratories for Investigative Dermatology, Room 204, Cancer Research Laboratory, 1303 N Mission Road, Los Angeles, California 90033, USA. E-mail: [email protected]
628
www.moleculartherapy.org vol. 15 no. 3, 628–635 march. 2007
& The American Society of Gene Therapy
There are several recent studies related to ex vivo gene therapy for DEB. Oritz-Urda et al. recently used a phi C31 integrasebased non-viral gene transfer approach to integrate stably COL7A1 cDNA into C7-null keratinocytes from recessive DEB (RDEB) patients. Using gene-corrected cells in a human skin equivalent, which was transplanted onto severe combined immunodeficient mice, they showed that many of the RDEB features were corrected after gene transfer.11 We developed a minimal lentiviral vector to express C7 in RDEB keratinocytes and fibroblasts (in which C7 was absent) and demonstrated the reversion of the RDEB cellular phenotype.12 We then used these gene-corrected RDEB cells to create a composite human skin equivalent, which was transplanted onto severe combined immunodeficient mice. The transplanted human skin exhibited C7 at the EDJ, and the transgene-derived C7 created anchoring fibril structures that were correctly organized into the basement membrane zone (BMZ). An ex vivo approach requires transplantation of genecorrected cells onto surgically prepared sites of the patient’s skin. The experience of using cultured keratinocyte autografts for transplantation onto human wounds has shown that this technology is often fraught with technical difficulties and poor graft take. We and others have also developed more straightforward direct in vivo gene therapy approaches to correct DEB. We showed that the intradermal injection of gene-corrected RDEB fibroblasts (so-called ‘‘cell therapy’’), recombinant human C7 (so-called ‘‘protein therapy’’), or lentiviral vectors expressing human C7 (so-called ‘‘vector therapy’’) into mouse skin or a DEB human skin equivalent engrafted onto a mouse achieves long-term expression of C7 that incorporates into the BMZ and reverses RDEB disease features, including dermal–epidermal separation and anchoring fibril defects.13–16 Severe RDEB causes widespread lesions and multiple wounds spanning large areas of trauma prone sites such as the sacrum, hips, feet, mouth, scalp, lower back, and hands. Therefore, treatment via intradermal injection would require numerous injections into multiple wound sites. An alternative strategy might be to inject genetically engineered cells into the patient’s circulation that home to the skin wounds and deposit the transgene product. In the study reported here, we demonstrated that IV injected normal human fibroblasts and gene-corrected RDEB fibroblasts (overexpressing human C7) home to a mouse’s wound site, continuously deliver C7 to the BMZ of the skin, and form anchoring fibril structures. Surprisingly, we found that C7 delivered to the wound sites significantly enhanced wound healing. Our data identified a previously unrecognized use of C7 to promote wound healing. This study provides the first demonstration of the potential use of IV injected gene-corrected DEB fibroblasts to restore C7 in DEB patients who have multiple open wounds.
RESULTS Restoration of C7 expression in RDEB fibroblasts We previously developed a minimal lentiviral transfer vector to express full-length human C7 cDNA and used it to transduce fibroblasts from RDEB patients.12 As shown in the Western blot Molecular Therapy vol. 15 no. 3, march. 2007
Intravenously Injected Human Fibroblasts
experiment depicted in Figure 1a, C7 is secreted into the conditioned media of normal human dermal fibroblasts, but not in cells from a severe RDEB patient. However, the gene-corrected RDEB cells expressed higher levels of C7 than normal fibroblasts (compare lane 2 with lane 3). Figure 1b shows immunofluorescence staining of these cells using an affinity-purified antibody to the NC1 domain of C7. Consistent with the immunoblot data, normal human fibroblasts displayed significantly lower levels of C7 than gene-corrected RDEB cells. As expected, there was no immunostaining of C7 in parental RDEB fibroblasts.
IV injected normal or gene-corrected RDEB fibroblasts delivered human C7 to the mouse’s regenerated BMZ We used an animal model to evaluate the feasibility of IV injecting gene-corrected RDEB fibroblasts for DEB treatment. We made a standard 1 1 cm full-thickness wound on the back of nu/nu athymic hairless mice. After 8–10 h, we IV injected the mice with normal human fibroblasts or gene-corrected human RDEB fibroblasts capable of synthesizing and secreting C7. The controls for this experiment were no injection of cells and the injection of RDEB fibroblasts that did not express C7. Skin biopsies from wounded and unwounded areas of the mouse’s skin were obtained and subjected to indirect immunofluorescence using an antibody that recognizes only human C7, but not rodent. As shown in Figure 2A, in wounded animals, the injected gene-corrected RDEB fibroblasts homed to the wound site and deposited human C7 into the EDJ of the mouse’s skin as early as 2 weeks after injection (Figure 2A, panel a). Identical results occurred when normal dermal fibroblasts were injected (Figure 2A, panel b), although with less deposition of human C7 at the healed BMZ (compare panel a with panel b). In contrast, there was no detectable human C7 in unwounded skin (Figure 2A, panel c) or in mice that were not wounded before injection
Figure 1 Expression of C7 in the cultured human RDEB fibroblasts used for IV injection. (a) Conditioned media from RDEB fibroblasts (lane 1), gene-corrected RDEB fibroblasts (lane 2), and normal human fibroblasts (lane 3) were concentrated and subjected to 6% sodium dodecyl sulfate–polyacrylamide gel electrophoresis followed by immunoblot analysis using a polyclonal antibody to NC1. The positions of fulllength 290 kDa C7 and molecular weight markers are indicated. (b) Immunohistochemistry with an affinity-purified polyclonal antibody to the NC1 domain of C7. Note the very high efficiency of C7 gene transfer to RDEB cells which, in their native state, cannot express C7. RDEB, parental RDEB fibroblasts; RDEB/C7, RDEB fibroblasts transduced with lentiviral expression vector for C7; NFB, normal human fibroblasts.
629
Intravenously Injected Human Fibroblasts
& The American Society of Gene Therapy
Figure 2 IV injected gene-corrected RDEB fibroblasts delivered human C7 to the mouse’s regenerated BMZ. A: (a–h) Immunofluorescence staining of the mouse’s skin after IV injection of various types of human fibroblasts was performed with antibodies specific for human C7 at 2 weeks after the injection. Note that the gene-corrected RDEB fibroblasts (a) and normal human fibroblasts (b) synthesized and delivered human C7 to the mouse’s regenerated BMZ in wounded sites, but not in unwounded sites (c). Further, there was no human C7 expression in mice (n ¼ 10 mice) that were not wounded before injection with either gene-corrected RDEB fibroblasts (d) or normal human fibroblasts (e). Healed wounds of mice injected with parental RDEB cells entirely lacked human C7 (f). (g and h) show the stable expression of human C7 at the mouse’s BMZ at 8 weeks after the injection with gene-corrected RDEB fibroblasts and normal human fibroblasts respectively. e, epidermis; d, dermis. B: (i–k) Dose-dependent deposition of human C7 at the mouse’s BMZ after IV injection with gene-corrected RDEB fibroblasts. Immunofluorescence staining with an antibody specific for human C7 was performed on healed mouse skin wounds after the animals were injected with (i) 0.75, (j) 1.5, and (k) 3 106 genecorrected RDEB fibroblasts, respectively. Note the dose-dependent increase in the deposition of human C7 at the mouse BMZ. C: Immunofluorescence staining of mouse skin was performed 3 weeks after the animals were IV injected with gene-corrected RDEB fibroblasts (RDEB/C7) or normal human fibroblasts (NFB). The skin was labeled with either a monoclonal antibody specific for human C7 (green, panel a-H) or a rabbit polyclonal antibody that recognizes both mouse and human C7 (red, panel a-H þ M). Merged images demonstrate co-localization of human C7 with mouse C7 in the mouse’s BMZ. RDEB/C7, gene-corrected RDEB fibroblasts (n ¼ 40 mice); NFB, normal fibroblast (n ¼ 20 mice); RDEB/FB, parent RDEB fibroblasts (n ¼ 20 mice).
(Figure 2A, panels d and e). Also, as expected, if uncorrected RDEB cells were injected, no human C7 was detected in the mouse’s wounded skin (Figure 2A, panel f). Further, the human fibroblast-derived C7 incorporated into the mouse’s BMZ and was sustained for at least 8 weeks, the longest duration tested in this study (Figure 2A, panels g and h). After demonstrating the ability of the injected fibroblasts to home to skin wounds and deliver human C7 to the mouse’s regenerated BMZ, we wished to determine the minimal concentration of fibroblasts required for C7 deposition at the BMZ. We injected 0.75, 1.5, and 3 106 gene-corrected DEB fibroblasts IV into the tail of nude mice. As shown in Figure 2B, 630
the injected fibroblasts mediated a dose-dependent increase in the deposition of human C7 into the mouse’s BMZ. To confirm that the human C7 derived from the injected cells was truly localized at the BMZ of the healing mouse skin, we colabeled the same vertical sections with a polyclonal antibody that recognizes both mouse and human C7 (Figure 2C, (a-H þ M) and our human-specific C7 antibody (a-H). As shown in Figure 2C, these two antibodies showed perfect co-localization in the mouse’s BMZ when the images were merged. Therefore, the IV injected normal and gene-corrected RDEB fibroblasts synthesized and delivered human C7 to the same location in the BMZ as the mouse’s endogenous C7. www.moleculartherapy.org vol. 15 no. 3, march. 2007
& The American Society of Gene Therapy
Tissue distribution of IV injected gene-corrected RDEB fibroblasts To determine whether the injected human fibroblasts became resident cells within the mouse’s wound, we stained healing murine skin with an antibody that recognizes human, but not murine, fibroblasts. As shown in Figure 3a, double staining for both human dermal fibroblasts and human C7 demonstrated that these IV injected human cells, residing now in the neodermis of the mouse’s healing wound, expressed human C7 that incorporated into the mouse’s BMZ. Further, IV injected gene-corrected RDEB fibroblasts were stably retained in the dermis of the wound tissues for 6 weeks after injection, where they continuously expressed C7. To determine whether these IV injected human fibroblasts also trafficked to tissues other than skin, we performed immunostaining on tissue sections from brain, kidney, liver, lung, spleen, heart, and small intestine with our antibody specific for human fibroblasts. As shown in Figure 3b, the cells were readily observed in healing skin, but not in brain, kidney, liver, spleen, heart, or small intestine. However, in approximately half of the IV injected mice, we did detect some injected human fibroblasts in small foci of the animals’ lungs. Nevertheless, we noted no untoward effects of the fibroblast injections in any of the mice. Daily weight, activity, and feeding habits were identical
Intravenously Injected Human Fibroblasts
between experimental and uninjected control groups of mice. Finally, in mice that were not wounded before injection, IV injected human fibroblasts were detected only in some of the animal’s lung and not in any other internal organs or the skin (data not shown). These results indicate that skin wounding is a prerequisite for the injected cells to home to the wound site.
Formation of anchoring fibrils by IV injected genecorrected RDEB fibroblasts Having shown that IV injected human fibroblasts into mice would take up residence in the healing wound and deliver human C7 that incorporated into the mouse’s BMZ, we next sought to determine whether the human collagen would form anchoring fibril structures. We performed immunoelectron microscopy on vertical sections of healing mouse skin using our human-specific antibody to C7. As shown in Figure 4, the human-specific antibody labeled the ends of wheat-stack-shaped structures emanating perpendicularly from the mouse’s BMZ. The immunogold labeling occurred at the ends of anchoring fibrils because the antibody used in this experiment is specifically directed against the NC1 domain, which is located at the ends of C7 antiparallel dimers. These data demonstrate that IV injected gene-corrected RDEB fibroblasts can synthesize and deliver
Figure 3 Tissue distribution of IV injected gene-corrected RDEB fibroblasts. (a) IV injected gene-corrected RDEB fibroblasts were detected in wounded dermal tissue. Three and 6 weeks after IV injection of gene-corrected RDEB fibroblasts, mouse skin was subjected to doubleimmunofluorescence labeling with a monoclonal antibody specific for human C7 (a-HC7) and a fluorescein isothiocyanate -conjugated secondary antibody (green) and an antibody specific for human fibroblasts (a-HFB) and Cy3-conjugated secondary antibodies (red). Shown are images from the same field at different wavelengths of excitation and merged. Representative fields of random sections from one of 40 IV injected mice are shown. Merged images (yellow) showed that injected cells were detected within the dermis of the wounded sites. (b) Tissue distribution of IV injected genecorrected RDEB fibroblasts. Four weeks after IV injection of gene-corrected RDEB fibroblasts, necropsies were performed on the mice (n ¼ 20 mice) and tissue sections obtained from brain, kidney, liver, lung, spleen, heart, small intestine, and healed skin were subjected to immunostaining using an antibody specific for human fibroblasts. Note that the injected fibroblasts were readily detected in the wounded skin and in small foci of lung, but not in any other organs.
Molecular Therapy vol. 15 no. 3, march. 2007
631
Intravenously Injected Human Fibroblasts
& The American Society of Gene Therapy
Figure 4 Immunoelectron microscopy of mouse skin injected with gene-corrected RDEB fibroblasts. Immunogold labeling of mouse skin after IV injection with gene-corrected RDEB fibroblasts was performed using a human-specific anti-C7 antibody and revealed that human C7 delivered via the injected cells incorporated into the mouse’s BMZ and formed anchoring fibrils. Note numerous gold particles decorating the ends of anchoring fibrils. Arrows denote gold particle-labeled NC1 domains of human C7 derived from injected fibroblasts. D, dermis; E, epidermis; HD, hemidesmosome. Bar ¼ 200 nm.
human C7 that forms anchoring fibril structures at the BMZ of healed mouse skin in vivo.
IV injected gene-corrected RDEB fibroblasts promote wound healing We and others have shown that dermal fibroblasts participate in wound healing by synthesizing extracellular matrices and orchestrating a cascade of events to heal cutaneous wounds.17,18 We next assessed whether human C7 delivered to the skin wound by IV injected fibroblasts could promote wound healing. We measured the area of each animal’s open wound at days 0, 7, 9, 11, 14, and 17 after injection. Figure 5 shows representative control and experimental wounds on days 0, 7, 9, 11, 14, and 17 (Figure 5a) and wound size measurements for each time point in 10 duplicate experiments (Figure 5b). The wounds of mice that were IV injected with gene-corrected RDEB fibroblasts overexpressing C7 demonstrated accelerated healing compared with those of mice that were injected with uncorrected RDEB fibroblasts (lacking C7 expression) when observed between days 7 and 11. By day 14, all wounds had closed and no significant differences between control and experimental wounds were observed. The wound-healing rates of mice injected with uncorrected RDEB fibroblasts and uninjected control mice were identical (data not shown). This suggests that uncorrected RDEB fibroblasts do not have an inhibitory effect on wound healing. We also assessed the effect of IV injected fibroblasts on the healing of human skin wounds. We grafted 1.5 1.5 cm human skin onto the back of nude mice using a previously described xenografting system.19 Eight weeks after grafting, the human skin was wounded with an 8 mm punch biopsy tool. The animals were then IV injected with gene-corrected RDEB fibroblasts or controls. As shown in Figure 6, the human skin wounds healed significantly faster in animals that received IV injected genecorrected RDEB fibroblasts compared with the control animals between days 5 and 11. By day 20, all wounds had healed and no detectable differences between control and experimental wounds were observed. With the use of species-specific antibodies (human laminin 5, C7, and human fibroblast antibodies), we confirmed that the cells re-epithelializing the wound and 632
Figure 5 IV injected gene-corrected RDEB fibroblasts promoted wound healing in mouse skin. A 1.0-cm2 (1 1 cm) full-thickness excision wound was made on the mid-back of 8- to 10-week-old athymic nude mice, which were then injected 8 h after wounding with gene-corrected RDEB fibroblasts expressing human C7 or uncorrected RDEB fibroblasts (n ¼ 10 mice per group). (a) Representative days 0, 7, 9, 11, 14, and 17 wounds are shown. Wound sizes were significantly reduced in mice injected with gene-corrected RDEB fibroblasts expressing C7 (RDEB/C7) compared with the mice that were injected with uncorrected RDEB fibroblasts (Control) at 7, 9, and 11 days after wounding. (b) Mean7SD wound size measurements at day 0, 7, 9, 11, 14, and 17 post-wounding (n ¼ 10 mice for each group).
re-populating the wound’s neodermis were of human origin (data not shown). Taken together, these data indicate that C7 delivered to the wound bed via the IV injection of human dermal fibroblasts promotes the closure of skin wounds in both murine and human skin.
DISCUSSION In this study, we used a murine wound-healing model to evaluate the feasibility of IV injecting gene-corrected human RDEB fibroblasts for therapy of DEB. Our studies showed that human RDEB fibroblasts could be molecularly engineered to secrete into the extracellular space of a host’s wounded skin a very large connective tissue molecule, namely, human C7 (MrE900,000). This macromolecule can incorporate correctly into the regenerated BMZ of the mouse’s skin and form anchoring fibril structures. Our data also demonstrated that in order for IV injected human dermal fibroblasts to home to the animal’s skin, the skin must first be wounded. In addition, it appears that IV injected gene-corrected RDEB fibroblasts, but not uncorrected RDEB fibroblasts, dramatically enhanced wound healing in these animals, suggesting the potential use of C7 to promote wound healing. We and others have shown that gene-corrected RDEB keratinocytes alone can be used to restore C7 expression and anchoring fibril formation in vivo.11,12 Although grafting genecorrected epidermal autografts onto RDEB patients may be www.moleculartherapy.org vol. 15 no. 3, march. 2007
& The American Society of Gene Therapy
Figure 6 IV injected gene-corrected RDEB fibroblasts promoted wound healing in human skin grafted onto mice. A 0.5-cm2 (8 mm diameter punch biopsy) full-thickness excision wound was made in human skin grafted onto the mid-back of 8- to 10-week-old athymic nude mice, which were then injected 8 h after wounding with genecorrected RDEB fibroblasts expressing human C7 or without injection (n ¼ 5 mice per group). (a) Representative days 0, 5, 8, and 11 wounds are shown. Wound sizes were significantly reduced in mice injected with gene-corrected RDEB fibroblasts expressing C7 (RDEB/C7) compared with control at 5, 8, and 11 days after wounding. (b) Mean7SD wound size measurements at days 0, 5, 8, 11, 15, and 20 post-wounding (n ¼ 5 mice for each group).
feasible, the procedure may have many technical problems similar to those encountered when transplanting epidermal autografts onto burn wounds. Among these, the degree of autograft take has been a major concern given that approximately 40% of cultured epidermal autografts transplanted onto burn wounds are lost. Also, the grafting procedure requires surgical excision of the patient’s skin down to fascia for optimal autograft take.20–22 Pain, immobilization of the patient, scarring, extensive wound care, and other potential morbidity issues limit the success of this strategy. We and others have recently also developed strategies involving the direct intradermal injection of gene-corrected RDEB fibroblasts, recombinant C7, or lentiviral vectors expressing C7 for the correction of DEB which offers many advantages over grafting thin-cultured epidermal autografts onto wounds.13–16 We and others showed that the intradermal injection of various therapeutic agents including cells, proteins, or lentivectors into intact RDEB skin stably restored C7 expression and anchoring fibrils at the BMZ and corrected the RDEB disease features. However, the small diffusion radius of intradermally injected agents and the inaccessibility of some sites such as the esophagus limit intradermal injection therapies. Patients with generalized, extensive blistering would require Molecular Therapy vol. 15 no. 3, march. 2007
Intravenously Injected Human Fibroblasts
many intradermal injections at multiple sites. These issues pose practical and logistical concerns for the clinical use of intradermal injections for RDEB patients. In this study, we found that IV injected fibroblasts specifically trafficked to wounded skin and synthesized and deposited human C7 at the newly formed BMZ of the host’s skin. It appears that this human fibroblast homing to wound sites is specific to skin wounding because we did not detect human fibroblasts in the liver when mouse liver was injured with carbon tetrachloride (CCl4) and then IV injected with either genecorrected RDEB fibroblasts or normal human fibroblasts (data not shown). We are not sure of the mechanisms underlying these observations. It is possible that wounded skin tissue sends signals to the injected fibroblasts. The cells then respond to the signals by migrating to the damaged sites and helping repair the wound. It has been shown recently that bone marrow-derived stem cells IV injected into wounded mice will home to the wound and participate in skin wound healing.23,24 In a recent study,25 murine green fluorescent protein-labeled bone marrow cells were transplanted into non-green fluorescent protein-positive mice. The green fluorescent protein-labeled cells were observed trafficking through both wounded and unwounded skin of the green fluorescent protein-negative recipients. Likewise, using a chimeric mouse wound model, it was shown that a percentage of cells in the healed dermis of skin wounds came from the bone marrow.26 It is unclear if bone marrow stem cells participate in the healing of naturally occurring human skin wounds. It is possible that local stem cells from nearby unwounded skin participate in the healing of the wound. Using an animal model, a recent study found that stem cells in hair follicles are enlisted to help heal skin wounds.26 Most surprisingly, we observed that the wounds of mice IV injected with gene-corrected RDEB fibroblasts overexpressing C7 demonstrated accelerated wound healing compared with control mice injected with uncorrected RDEB cells. Further, there was no significant difference in the wound-healing rate between the mice injected with uncorrected RDEB cells and those without injection (data not shown). Similarly, C7 delivered to the wound bed via intradermal or IV injection of the recombinant protein also promoted the closure of skin wounds (unpublished observation). In addition, when wounds were made on human skin grafted onto mice, IV injected gene-corrected RDEB fibroblasts also significantly enhanced healing of the human skin. How exogenously delivered C7 promotes healing is unknown. The presence of RGD motifs, fibronectin type III, cartilage matrix protein, and von Willebrand factor homology domains in the C7 molecules would allow C7 to modulate cells and matrix–matrix interactions at the wound site. We have shown previously that C7 promotes keratinocyte and fibroblast matrix attachment and cell motility.12,27 In acute wound healing, C7 staining is detected in the wound bed and the neodermis.28,29 It is possible that C7 may recruit or maintain fibroblasts at this location and promote cellular remodeling of the wound bed. C7 also has specific domains with affinity for other BMZ and dermal molecules such as type IV collagen, laminin-5, fibronectin, and type I collagen.30–32 This may promote the organization and maturation of the connective tissue macromolecules within the 633
Intravenously Injected Human Fibroblasts
BMZ and the high papillary dermis which would be useful for reepithelialization, building granulation tissue, and wound closure. Interestingly, it has been shown that C7 expression was undetectable in two patients’ wounds—one a lower-limb ulcer and the other a non-healing leg ulcer. These patients did not show any clinical evidence of healing.33 Therefore, C7 synthesis and incorporation into anchoring fibrils is probably a prerequisite for the synthesis of a stable basement membrane in healing wounds and is required for healing to proceed. The small number of IV injected fibroblasts detected in foci of lung tissue raise a potential safety issue. Nevertheless, IV injected mice did not exhibit any abnormal behavior. Their activity, eating habits, and growth rate were identical to those of uninjected mice. There was no histological evidence of fibrosis in the lungs of those mice with small foci of injected human fibroblasts (data not shown). Furthermore, when identical experiments were conducted with immunocompetent mice, we did not observe any injected fibroblasts in the lung (data not shown). It is possible that fibroblasts trafficking to the lung will be eliminated by the host’s immune system, which is missing in the immunodeficient mice. Our hope is that fibroblasts IV injected into immune competent DEB patients would be rapidly cleared as well. In summary, our studies demonstrate that IV injected genecorrected RDEB fibroblasts were capable of homing to skin wounds and synthesizing and delivering human C7 that incorporated into the newly regenerated BMZ and promoted wound healing. Defining the potential advantages and drawbacks of IV injected cell-based therapy will be of significant interest for future studies. As the exogenous application of biological products to skin wounds has been very disappointing in terms of promoting wound healing owing to the excess proteolytic enzymes in the wound bed that degrade and nullify the applied agent, we believe that IV delivery of pro-wound healing agents by the method detailed in this study could be used to deliver extracellular matrix molecules and growth factors continuously to healing wounds. For example, physicians could IV inject molecularly engineered fibroblasts that overexpress products that promote various steps in wound healing such as re-epithelialization, fibroplasia, or angiogenesis. In addition to being a therapy for healing chronic wounds, this strategy may be particularly useful for patients with DEB who have multiple open wounds and skin fragility due to a paucity of C7 and anchoring fibrils in their skin.
MATERIALS AND METHODS Cell culture. Normal human fibroblasts were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. Primary RDEB fibroblasts (gift of Dr Scott Herron, Stanford University) and RDEB fibroblasts genetically modified to re-express C7 by lentiviral transduction, as described previously, were cultured in Dulbecco’s modified Eagle’s medium /Ham’s F12 (1:1) supplemented with 10% fetal bovine serum.12 Protein analysis. For immunoblot analysis, normal human dermal
fibroblasts, RDEB fibroblasts, and lentivirus-transduced RDEB fibroblasts were grown to confluence, the medium was changed to serum-free medium containing 150 mm ascorbic acid, and the cultures were maintained for an
634
& The American Society of Gene Therapy
additional 24 h. The media were collected, equilibrated to 5 mM EDTA, 50 mM N-ethylmaleimide, and 50 mM phenylmethylsulfonyl fluoride, concentrated 10- to 15-fold (Centricon-100, Amicon, Beverly, MA) and subjected to 6% sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Proteins were electrotransferred onto a nitrocellulose membrane. The presence of C7 was detected by Western blot analysis using a polyclonal antibody to the NC1 domain of C7.34 Immunofluorescence staining of cells. RDEB or normal human cells
were plated in TissueTek chamber slides (Nunc Inc., Naperville, IL) on polylysine at 371C for 18 h. Cells were immersed in periodate-lysineparaformaldehyde fixative for 10 min at room temperature, washed several times with phosphate buffered saline (PBS) to remove fixative, and then permeabilized and blocked by incubating in PBS with 3% bovine serum albumin, 1% saponin, and 10% normal goat serum for 15 min at room temperature. The cells were incubated with an affinitypurified polyclonal antibody to the NC1 domain of C7 (ref. 34) at a dilution of 1:200 in a humidified chamber for 2 h, washed three times with PBS and 1% saponin, counterstained with a fluorescein isothiocyanate-conjugated goat antibody to rabbit immunoglobulin IgG (1:200 dilution) for 1 h (Organon Teknika-Cappel, MP Biomedicals Inc, Aurora, Ohio) and washed again. The cells were examined and photographed on a Zeiss microscope equipped with an epiluminating fluorescent light source. IV injection of fibroblasts into mice. To prepare mice for injection, we first made a 1.0 1.0-cm full-thickness excision wound by lifting the skin with forceps and removing full-thickness skin using a curved scissors on the mid-back of 8- to 10-week-old athymic nude mice (The Jackson Laboratory, Bar Harbor, ME) and then covered the wound with a Band-Aid and a Coban, self-adherent wrap, to prevent desiccation. Within 8 h after wounding, 1.5 106 gene-corrected RDEB fibroblasts, normal human fibroblasts, and parental uncorrected RDEB fibroblasts in PBS (100 ml PBS with 50 U heparin/ml) were injected into the tail vein of wounded mice using a 30 G needle. We then re-injected mice one more time with 1.5 106 cells 8 h later after the initial injection to increase the number of injected cells. We injected 40 mice with genecorrected RDEB fibroblasts, 20 mice with normal human fibroblasts, and 20 mice with uncorrected parental RDEB fibroblasts. Two to 8 weeks after injection, mouse skin biopsies from wounded or unwounded areas were obtained and subjected to immunostaining using an antibody specific for human C7 (clone LH 7.2; Sigma, St Louis, MO) or a rabbit polyclonal antibody recognizing both human and mouse C7 as described below. To measure the wound size, standardized digital photographs were taken of the wounds at 0, 7, 9, 11, 14, and 17 days after wounding and open wound areas determined with an image analyzer (AlphaEase FC version 4.1.0, Alpha Innotech Corporation, San Leandro, CA) on a Macintosh computer. For evaluating wound healing of human skin, a 1.5 1.5-cm of fullthickness human skin was grafted onto nude mice as described previously.19 Eight weeks after grafting, the human skin was wounded using an 8 mm punch biopsy tool, bandaged, and then injected with gene-corrected RDEB fibroblasts as described above. The assessment of wound healing was performed by area planimetry in the same way as in the assessment of the murine wounds described above. All animal studies were conducted using protocols approved by the University of Southern California Institutional Animal Use Committee. Immunofluorescence staining and ultrastructural analysis of tissue.
Five-micrometer thick sections of the optimal cutting temperature–embedded tissues were cut on a cryostat, fixed for 5 min in cold acetone, and air-dried. Immunolabeling of the tissue was performed using standard immunofluorescence methods as www.moleculartherapy.org vol. 15 no. 3, march. 2007
Intravenously Injected Human Fibroblasts
& The American Society of Gene Therapy
described previously.35 Briefly, for single- and double-immunofluorescence staining, sections were blocked with MOM. Mouse IgG Blocking Reagent (Vector Laboratories, Burlingame, CA) for 1 h at room temperature. Primary antibodies were diluted in PBS with 1% bovine serum albumin. For type C7 staining, we used monoclonal antibodies against human C7, clone LH 7.2 (Sigma) or a rabbit polyclonal antibody recognizing both mouse and human C7.34 For double-immunofluorescence, we incubated the mouse monoclonal anti-human C7 together with the following rabbit or mouse anti-human antibodies: a rabbit polyclonal antibody to both human and mouse C7; a mouse monoclonal antibody to human skin fibroblasts (Clone D7-FIB, Acris Antibodies GmbH, Hiddenhausen, Germany). All primary antibody dilutions were 1:100. After incubation for 1 h at room temperature, sections were washed in PBS three times and stained for 1 h with fluorescein isothiocyanate -conjugated goat antimouse IgG1 with or without Cy3-conjugated goat anti-rabbit IgG (Sigma) or Alex fluor 568 goat anti-mouse-Ig G2a (Molecular Probe, Eugene, OR) diluted 1:500 in PBS with 1% bovine serum albumin. Slides were mounted with 40% glycerol. Photographs of stained sections were taken using a Zeiss Axioplan fluorescence microscope equipped with a Zeiss Axiocam MRM digital camera system.. Immunogold electron microscopy was performed on the mouse skin using a standardized method as described previously.36 To assess human anchoring fibril formation and ultrastructure, 40 mm sections were fixed in 0.1% glutaraldehyde, rinsed in 0.15 M Tris (pH 7.5), then incubated in an antibody specific for human C7 (NP185, a gift of Dr Lynn Sakai, Shriners Hospital for Children, Portland, OR) followed by 1 nm gold secondary antibody and enhancement as described previously.37
11. 12. 13.
14.
15.
16.
17. 18. 19.
20.
21.
22.
23.
24. 25. 26. 27.
28. 29.
ACKNOWLEDGMENTS This work was supported by grants RO1 AR47981 (Dr Chen) and RO1 AR33625 (Dr Woodley) from the National Institutes of Health. We thank Sara Tufa for technical support of immuno-EM.
30.
31.
REFERENCES 1. 2. 3. 4. 5.
6. 7. 8.
9. 10.
Stein, J (2000). Chronic wounds on the rise. Internal Medicine World Report. /http:// 207.21.218.87/news1a.asp?cid=27S. Hess, CT (1999). Care for a diabetic ulcer. Nursing 29: 70–71. Brown, GL et al. (1989). Enhancement of wound healing by topical treatment with epidermal growth factor. N Engl J Med 321: 76–79. Goldman, R (2004). Adv Skin Wound Care 17: 24. Review. Fu, X, Li, X, Cheng, B, Chen, W and Sheng, Z (2005). Engineered growth factors and cutaneous wound healing: success and possible questions in the past 10 years. Wound Repair Regen 13: 122–130. Lin, AN and Carter, DM (eds) (1992). Epidermolysis Bullosa: Basic and Clinical Aspects. Springer-Verlag Inc.: NY. Briggaman, RA and Wheeler Jr, CE (1975). The epidermal–dermal junction. J Invest Dermatol 65: 71–84. Uitto, J and Christiano, AM (1994). Molecular basis for the dystrophic forms of epidermolysis bullosa: mutations in the type VII collagen gene. Arch Dermatol Res 287: 16–22. Burgeson, RE (1993). Type VII collagen, anchoring fibrils, and epidermolysis bullosa. J Invest Dermatol 101: 252–255. Sakai, LY, Keene, DR, Morris, NP and Burgeson, RE (1986). Type VII collagen is a major structural component of anchoring fibrils. J Cell Biol 103: 1577–1586.
Molecular Therapy vol. 15 no. 3, march. 2007
32.
33. 34.
35.
36.
37.
Ortiz-Urda, S et al. (2002). Stable nonviral genetic correction of inherited human skin disease. Nat Med 8: 1166–1170. Chen, M et al. (2002). Restoration of type VII collagen expression and function in dystrophic epidermolysis bullosa. Nat Genet 32: 670–675. Woodley, DT et al. (2003). Normal and gene-corrected dystrophic epidermolysis bullosa fibroblasts alone can produce type VII collagen at the basement membrane zone. J Invest Dermatol 121: 1021–1028. Ortiz-Urda, S, Lin, Q, Green, CL, Keene, DR, Marinkovich, MP and Khavari, PA (2003). Injection of genetically engineered fibroblasts corrects regenerated human epidermolysis bullosa skin tissue. J Clin Invest 111: 251–255. Woodley, DT et al. (2004). Intradermal injection of lentiviral vectors corrects regenerated human dystrophic epidermolysis bullosa skin tissue in vivo. Mol Ther 10: 318–326. Woodley, DT et al. (2004). Injection of recombinant human type VII collagen restores collagen function in dystrophic epidermolysis bullosa. Nat Med 10: 693–695. Singer, AJ and Clark, RAF (1999). Cutaneous wound healing. Rev N Engl J Med 341: 738–746. Li, W, Fan, J, Chen, M and Woodley, DT (2004). Mechanisms of human skin cell motility. Histol Histopathol 19: 1311–1324. Kim, YH, Woodley, DT, Wynn, KC, Giomi, W and Bauer, EA (1992). Recessive dystrohpic bullosa phenotype is preserved in xenografts using SCID mice: development of an experimental in vivo model. J Invest Dermatol 92: 191–197. Gallico III, GG, O’Connor, NE, Compton, CC, Kehinde, O and Green, H (1984). Permanent coverage of large burn wounds with autologous cultured human epithelium. N Engl J Med 311: 448–451. Herzog, SR, Meyer, A, Woodley, DT and Peterson, HD (1988). Wound coverage with cultured autologous keratinocytes: use after burn wound excision, including biopsy followup. J Trauma 28: 195–198. Woodley, DT, Peterson, HD and Herzog, SR (1988). Burn wounds resurfaced by cultured epidermal autografts show abnormal reconstitution of anchoring fibrils. JAMA 259: 2566–2571. Badiavas, EV, Abedi, M, Butmaarc, J, Falanga, JV and Quesenberry, P (2003). Participation of bone marrow derived cells in cutaneous wounding healing. J Cell Physiol 196: 245–250. Badiavas, EV and Falanga, V (2003). Treatment of chronic wounds with bone marrow-derived cells. Arch Dermatol 139: 510–516. Fathke, C et al. (2004). Contribution of bone marrow-derived cells to skin: collagen deposition and wound repair. Stem Cells 22: 812–822. Ito, M et al. (2005). Stem cells in the hair follicle bulge contribute to wound repair but not to homeostasis of the epidermis. Nat Med 11: 1351–1354. Chen, M, O’Toole, EA, Li, YY and Woodley, DT (1999). a2b1 integrin mediates dermal fibroblast attachment to type VII collagen via a 158 amino acid segment of the NC1 domain. Exp Cell Res 249: 231–239. Haapasalmi, K et al. (1995). Expression of epithelial adhesion proteins and integrins in chronic inflammation. Am J Pathol 147: 193–206. Woodley, DT, Peterson, HD and Herzog, SR (1988). Burn wounds resurfaced by cultured epidermal autografts show abnormal reconstitution of anchoring fibrils. JAMA 259: 2566–2571. Chen, M et al. (1997). Interactions of the amino-terminal noncollagenous (NC1) domain of type VII collagen with extracellular matrix components. J Biol Chem 272: 14516–14522. Chen, M, Marinkovich, MP, Jones, JJ, O’Toole, EA and Woodley, DT (1999). NC1 domain of type VII collagen binds to the beta 3 chain of laminin 5 via a unique subdomain within the fibronectin-like repeats. J Invest Dermatol 112: 177–183. Rousselle, P, Keene, DR, Ruggiero, F, Champliaud, MF, Rest, M and Burgeson, RE (1997). Laminin 5 binds the NC-1 domain of type VII collagen. J Cell Biol 138: 719–728. Hopkinson, I, Anglin, IE, Evans, DL and Harding, KG (1997). Collagen VII expression in human chronic wounds and scars. J Pathol 182: 192–196. Chen, M, Petersen, MJ, Li, H-L, Cai, X-Y, O’Toole, EA and Woodley, DT (1997). Ultraviolet A irradiation upregulates type VII collagen expression in human dermal fibroblasts. J Invest Dermatol 108: 125–128. Gammon, WR, Briggaman, RA, Inman III, AQ, Queen, LL and Wheeler, CE (1984). Differentiating anti-lamina lucida and anti-sublamina densa anti-BMZ antibodies by indirect immunofluorescence on 1.0 M sodium chloride-separated skin. J Invest Dermatol 82: 139–144. Keene, DR, Sakai, LY, Lunstrum, GP, Morris, NP and Burgeson, RE (1987). Type VII collagen forms an extended network of anchoring fibrils. J Cell Bio 104: 611–621. Champliaud, MF, Lunstrum, GP, Rousselle, P, Nishiyama, T, Keene, DR and Burgeson, RE (1996). Human amnion contains a novel laminin variant, laminin 7, which like laminin 6, covalently associates with laminin 5 to promote stable epithelial-stromal attachment. J Cell Biol 132: 1189–1198.
635