Introduction

Dystrophic epidermolysis bullosa (DEB) is a group of heritable mechano-bullous skin diseases characterized by skin fragility, separation of the epidermis from the dermis (blister formation), milia and scarring of varying clinical severity.^1,2 DEB is transmitted in either a dominant (DDEB) or a recessive (RDEB) mode. All forms of DEB are caused by mutations in COL7A1, the gene encoding for type VII collagen (C7).^3,4,5,6 C7 is the major component of anchoring fibrils, attachment structures in the basement membrane zone (BMZ) of skin that adhere the epidermal layer of skin onto the dermis.^7,8,9 The BMZ of patients with DEB is characterized by a paucity or altered morphology of anchoring fibrils.²

DEB affects thousands of families and more than 300 distinct COL7A1 mutations have been identified in DEB patients.^3,4,10,11 The most severe form of DEB is RDEB (the Hallopeau-Siemens type, HS-RDEB) in which both C7 and anchoring fibrils are absent from the skin due to null mutations in the COL7A1 gene. As a result, HS-RDEB is characterized by severe skin blistering, extremely fragile skin, mutilating scarring of the hands and feet, joint contractures, and strictures of the esophagus. In the second or third decade of life, HS-RDEB patients develop aggressive squamous cell carcinomas in chronically wounded areas which often lead to metastasis and death.

The development of therapeutic approaches for DEB has been previously explored using ex vivo and in vivo strategies. Ex vivo gene therapy using either a phi C31 integrase–based nonviral or a lentiviral vector gene transfer approach was explored.^12,13 Restoration of C7 expression and correction of RDEB cellular phenotypes in vitro was achieved in RDEB keratinocytes with either of these vectors. In both cases, formation of anchoring fibrils at the dermal–epidermal junction and stable correction of the RDEB disease hallmarks were observed when human skin regenerated by gene-corrected RDEB keratinocytes were grafted onto immunodeficient mice.^12,13

We and others have also developed more straightforward direct in vivo gene therapy approaches to correct DEB. Specifically, we showed that the intradermal injection of gene-corrected RDEB fibroblasts ("cell therapy"), recombinant human C7 ("protein therapy"), or lentiviral vectors expressing human C7 ("vector therapy") into mouse skin or a human DEB skin equivalent engrafted onto a mouse achieves long-term expression of C7. This protein then incorporates into the BMZ and reverses RDEB disease features, including dermal–epidermal separation and anchoring fibril defects.^14,15,16,17 More recently, we also demonstrated the feasibility of an intravenous injection approach. We showed that intravenously 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 BMZ and formed anchoring fibril structures.¹⁸

A mouse model has been developed for RDEB in immunocompetent mice by targeted inactivation of the COL7A1 gene.¹⁹ These COL7A1 null (Col7a1^–/–) mice have no C7 at the BMZ of their skin, and they entirely lack ultrastructurally recognizable anchoring fibrils. Electron microscopy also reveals sublamina densa bullae precisely like those in human DEB patients. Clinically, the newborn mice exhibit extensive blisters and die within the first week of life, probably from complications due to the extensive blistering. Thus, these Col7a1^–/– mice recapitulate many of the clinical, genetic, and ultrastructural features of severe RDEB patients.

In this study, we sought to determine whether protein therapy with intradermal C7 injections into these mice could reverse their DEB-like disease. We showed that the intradermally injected human C7 translocated and stably incorporated into the mouse's BMZ and formed anchoring fibrils. As a consequence, the DEB murine phenotype was significantly improved with decreased skin fragility and blistering and markedly prolonged survival. Most interestingly, although anti-human C7 antibodies were induced by the injected protein, the antibodies did not exhibit any adverse effects in the animals.

Results

Restoration of C7 in the DEB mouse's BMZ

To evaluate the feasibility of protein therapy for DEB, we used a murine C7 knockout model (Col7a1^–/–) that recapitulates the clinical, genetic, immunohistochemical, and ultrastructural characteristics of severe human RDEB.¹⁹ As shown in Figure 1, at day 1 after birth, the DEB mice had hemorrhagic blisters on their paws and neck and large fluid-filled blisters on their ventral surface (Figure 1a and b). Histology showed dermal–epidermal separation, and immunostaining with anti-C7 antibodies revealed a complete absence of C7 at the BMZ (Figure 1c and d). In contrast, skin obtained from wild-type littermates (Col7a1^+/+) demonstrated strong C7 staining at the BMZ (Figure 1e). Without treatment, the DEB mice die within a week from complications of extensive skin blistering.

Figure 1.

Clinical, histological, and immunological presentation of dystrophic epidermolysis bullosa (DEB) mice. DEB mice show characteristic clinical features of the disease including skin fragility, hemorragic blisters, erosions, and large fluid-filled bullae primarily on the ventral side. (a) Typical large hemorrhagic blisters are seen on paws and (b) large blisters are seen on the ventral side. (c) Hematoxylin and eosin staining of skin from DEB mice revealed separation of the epidermis (e) from the dermis (d). (d and e) Immunofluorescence staining of the tissues obtained from DEB mouse skin (d) and its wild-type littermate's skin (e) with an affinity-purified polyclonal antibody recognizing the NC1 domain of type VII collagen (C7). Note the subepidermal separation and absence of C7 staining in DEB mice compared with the strong C7 staining of the basement membrane zone in normal mice.

Full figure and legend (51K)

Figure 7.

Anti-type VII collagen (anti-C7) IgG neither bound to the skin directly nor inhibited further basement membrane zone (BMZ) incorporation of newly injected C7. (a) Cryosections were obtained from the back skin of dystrophic epidermolysis bullosa (DEB) mice treated with C7 for various times as indicated and labeled with fluorescein isothiocyanate-conjugated goat anti-mouse IgG (-mIgG), goat anti-mouse C3 antibody (-mC3), or with an anti-NC1 antibody (-NC1), respectively. Please note that injected C7 was found at the BMZ of DEB mice for all the time points (lower panels). In contrast, no deposits of mouse-anti-C7 IgG (top panels) or murine C3 (middle panels) were detected in the BMZ of treated mice. (b) Incorporation of subsequently injected C7 into the mouse's BMZ regardless of the presence of C7 antibodies. DEB mice (n = 4) that had been injected with C7 and were found by enzyme-linked immunosorbent assay to have anti-C7 IgG in their blood were reinjected with 10 g of C7 at remote sites where C7 was not expressed. Immunofluorescence staining with an anti-NC1 antibody was performed on mouse skin 1 week after injection at both reinjected (panel A) and previously uninjected abdominal areas (panel B). Please note that the newly injected C7 still migrated from the dermis and incorporated into the mouse's BMZ.

Full figure and legend (106K)

To test whether protein therapy could correct the DEB mouse phenotype, we intradermally injected purified human recombinant C7 into the DEB mice (n = 45) on day 1 after birth in the skin of the back. Skin biopsies at various distances from the injection site were then obtained 1 week after initial injection and subjected to immunostaining with a polyclonal rabbit antibody that recognized the aminoterminal noncollagenous domain (NC1) of human C7. As shown in Figure 2a, the injected human recombinant C7 translocated to the dermal–epidermal junction and incorporated into the DEB mouse's BMZ. C7 staining was strongest near the injected site and became weaker away from the injected area (compare panel A to C). Nevertheless, in some severely affect animals in which much of the skin was blistered, we detected C7 incorporated into the BMZ of skin distal to the back such as the neck, abdomen, and paws (Figure 2a, panels D–F). This observation suggests that, when human C7 is injected early into the back of severely affected mice with extensive blistering, the widespread poor epidermal–dermal adherence allows the C7 to migrate to skin sites distal from the injection site. Furthermore, as shown in Figure 2b, the injected human C7 was detected at the BMZ as early as 6 hours after the injection. With continued weekly injections, we observed that the mice continued to have human C7 at the BMZ of their back skin for as long as we continued the experiments (7 months) (Figure 2b).

Figure 2.

Presence of type VII collagen (C7) in dystrophic epidermolysis bullosa (DEB) mice following protein injection. (a) DEB mice were intradermally injected on their dorsal surface with 10 g C7 once every day for 4 consecutive days, and tissue sections obtained from skin at various distances from the injection site 1 week after injections were subjected to immunostaining using an anti-NC1 antibody. Panels A, B, and C are biopsies taken from the immediate injected area, near by and far away, respectively. Note that the injected human C7 migrated from the dermis and incorporated into the DEB mouse basement membrane zone at the injected dorsal side. Interestingly, some biopsies obtained from sites remote from the injection sites, such as neck (panel D), abdomen (panel E), and paw (panel F), also revealed C7 staining. (b) Dorsal side of skin from DEB mice injected intradermally with 10 g of purified recombinant human C7 once every day for the first week and then weekly thereafter was stained with an anti-NC1 antibody. Panels A, B, and C are biopsies taken 6 hours, 15 weeks, or 28 weeks, respectively. For panel D, DEB mice (n = 10) were injected with 60 g C7 for the first week only, and tissue sections obtained from the skin at 8 weeks after initial injections were subjected to immunostaining using an anti-NC1 antibody. e, epidermis; d, dermis.

Full figure and legend (82K)

To determine more precisely the in vivo half-life of injected C7, we performed multiple injections into the DEB mice (n = 10) during the first week of life and then stopped the injections. As also shown in Figure 2b (panel D), injected C7 was detected at the injected skin area for at least 2 months after the final injection. However, biopsies taken at the 14 weeks after the final injection showed an absence of C7 in the skin of the animals. These data indicate that C7 has a half-life of at least 2 months or even longer in vivo.

Absence of C7 in the blood and other organs of DEB mice after protein therapy

Because C7 was seen distal to the intradermal injection site in some animals, we wished to determine whether the injected C7 was transported in the animals' circulation. To examine this issue, we performed immunoblot analysis on sera obtained from DEB mice at various times after C7 injections. As shown in Figure 3a, we did not detect any C7 in the blood stream (Figure 3a). Furthermore, to determine whether intradermally injected C7 trafficked to tissues other than skin, we also conducted immunostaining of tissue sections from brain, kidney, liver, lung, spleen, heart, and small intestine with an anti-NC1 antibody. As shown in Figure 3b, C7 was readily observed in the skin, but not in brain, kidney, liver, spleen, heart, liver, or small intestine.

Figure 3.

Absence of type VII collagen (C7) in blood stream and internal organs. (a) Sera were taken from either control wild-type mice (WT) without injection or dystrophic epidermolysis bullosa (DEB) mice that were injected with C7 once every day for the first week and then weekly thereafter at the time indicated and subjected to 6% sodium dodecyl sulfate polyacrylamide sulfate followed by immunoblot analysis using an anti-NC1 antibody. One hundred and fifty nanograms of purified recombinant C7 was run as a control (Con). The positions of full-length 290-kd C7, 70-kd mouse IgG heavy chain (mIgGH), and molecular weight markers are indicated. (b) Tissue distribution of intradermally injected C7. Four weeks after injection of C7, necropsies were performed on the DEB mice (n = 6), and tissue sections obtained from brain, kidney, liver, lung, spleen, heart, small intestine (SI), and skin were subjected to immunostaining using an anti-NC1 antibody. Note that the injected C7 was readily detected in the skin, but not in any other organs.

Full figure and legend (105K)

Formation of anchoring fibrils by injected human C7

C7 is the major component of anchoring fibrils, structures that stabilize the association of the epidermis onto the underlying dermis. The DEB mouse is characterized by a complete absence of anchoring fibrils in the skin. To determine whether the injected recombinant human C7 could form anchoring fibrils at the BMZ in vivo, immunoelectron microscopy was performed on mouse skin that was injected with 60 g of human C7 using a polyclonal antibody that recognizes the NC1 domain of C7. As shown in Figure 4, the injected human C7 incorporated into the mouse's BMZ, oriented correctly and formed anchoring fibril structures, thus demonstrating correction of the major ultrastructural abnormality seen in DEB skin.

Figure 4.

Immunoelectron microscopy of mouse skin injected with human type VII collagen (C7). Immunogold labeling of dystrophic epidermolysis bullosa mouse skin after injections with 60 g of C7 at 2 weeks after injection was performed using an anti-NC1 antibody. Note that human C7 incorporated into the mouse basement membrane zone and formed anchoring fibrils, ends of which are decorated by gold particles (arrowheads). D, dermis; E, epidermis; HD, hemidesmosome. Scale bars = 400 nm.

Full figure and legend (174K)

Increased survival of DEB mice by protein therapy

Without treatment, DEB mice die within the first week of life. Therefore, we also used the survival rate as a readout for therapeutic efficacy of the C7 injections. The Kaplan–Meier curve shows that C7 injections significantly increased the survival of DEB mice (Figure 5a). The median overall survival was 5 weeks for 45 C7-treated mice (range: 1–28 weeks) compared to 1 week (range: 0–1 week) for 30 untreated DEB mice (P < 0.001, log-rank test). Of note, 7 of the 45 (13%) C7-injected mice survived beyond 12 weeks. Figure 5b shows one example of a C7-treated mouse which survived 22 weeks after multiple injections with a cumulative total of 250 g of recombinant C7. Immunolabeling with an anti-C7 antibody showed strong C7 staining at the BMZ of its skin. Hematoxylin and eosin staining of biopsies from its skin showed intact dermal–epidermal adherence, essentially similar to normal mouse skin.

Figure 5.

Improved survival of dystrophic epidermolysis bullosa (DEB) mice by type VII collagen (C7) injections. (a) Kaplan–Meier curves showing survival comparison between untreated (n = 30) and C7-treated (n = 45) DEB mice. The curves indicate a significant increase in the survival of C7-treated mice in comparison with untreated DEB mice (P < 0.001). (b) Picture showing 1-day-old DEB mouse with hemorrhagic blisters on paws (panel A)—the same mouse survived past 22 weeks after injections with cumulative 250 g of C7 (panel B). Immunofluorescence staining with anti-NC1 antibody demonstrated strong C7 staining at the basement membrane zone (BMZ) of the 22-week-old treated mouse (panel C) and histological analysis revealed apparently normal association of the epidermis and dermis at the cutaneous BMZ (panel D).

Full figure and legend (88K)

It is possible that injection of any protein could increase the survival of an animal. To examine the specificity of the observed efficacy of injected C7, we injected DEB mice (n = 15) with bovine serum albumin (BSA) using the same dose as C7. We found that 14 of the 15 (93%) BSA-injected mice died within 5 days of birth and 1 of the 15 (6%) died at day 10. Therefore, we conclude that the increased survival seen in C7-treated DEB mice is specifically due to the restoration of C7, anchoring fibrils, and epidermal–dermal adherence. Collectively, we injected a total of 45 DEB mice with C7 at cumulative total doses between 20 and 300 g. Approximately 88% of the injected mice had restored C7 at the BMZ and displayed an increased survival up to 7 months. Therefore, C7 injections into DEB mice not only corrected their skin DEB phenotype, but also markedly improved their survival.

Production of anti-C7 antibodies after injection of human C7

Gene- or protein-based therapy for DEB may provoke unwanted immune responses against the introduced gene product, especially in patients with null mutations and a complete absence of C7. In this case, DEB mice with a complete absence of C7 are ideal for generating preclinical data with regard to immune responses following C7 therapy and for developing strategies to reduce immune responses. We wished to examine whether the injected human C7 would induce an immune response in DEB mice by inducing circulating anti-C7 antibodies. Sera from C7 injected mice (n = 6) were taken at various time points after treatment and then subjected to our NC1-based enzyme-linked immunosorbent assay (ELISA) analysis to determine the presence of anti-C7 antibodies.²⁰ Anti-C7 antibodies were detected in all of the six mice examined at 30 days after the initial injection, and the amount increased steadily thereafter (Figure 6a). C7-injected mice had titers of circulating antibody (1:20–1:100) when assayed by indirect immunofluorescence against either mouse or salt-split human skin. As shown in Figure 6b, serum from the C7-injected mice contained mouse IgG antibodies that bound to the dermal–epidermal junction of normal mouse skin (panel A) and the dermal side of the salt-split human skin (panel B).

Figure 6.

Anti-type VII collagen (anti-C7) IgG production in dystrophic epidermolysis bullosa mice after protein therapy. (a) Anti-C7 IgG production was measured by NC1-based enzyme-linked immunosorbent assay over time after intradermal injetions with recombinant human C7. Anti-C7 IgG was detected at 30 days and increased steadily thereafter in C7-treated mice (mean SD, n = 6). (b) Immunolabeling of mouse and human skin with serum from mice injected with C7. Sections of normal mouse skin (A) and salt-split human skin (B) were stained with mouse serum obtained from mice injected with C7 at a dilution of 1:20. Note that circulating antibodies labeling the basement membrane zone of mouse skin and the dermal floor of salt-split human skin were found in the serum samples from C7-injected mice. OD, optical density.

Full figure and legend (36K)

To determine whether the induced anti-C7 antibodies bound directly to the C7 at the mouse's BMZ, biopsies were taken from the injection sites at various times and examined by direct immunofluorescence for murine IgG or C3 complement deposits at the BMZ. As shown in Figure 7a, no murine IgG or C3 complement deposits were detected at the BMZ of the mice. Nevertheless, we observed abundant human C7 at the BMZ when the same tissue section was probed with an anti-NC1 antibody. Furthermore, we did not detect other murine immunoglobulin isotypes such as IgA or IgM deposits at the BMZ either (data not shown).

To determine whether the presence of these circulating anti-C7 antibodies would prevent newly injected human C7 from incorporating into the BMZ of the animals' skin after subsequent treatments, we selected four DEB mice that had been dorsally injected with C7 for 1 month and were found by ELISA to have anti-C7 antibodies in their blood. These mice had C7 at the dermal–epidermal junction of their back, but no C7 in the skin of the ventral abdominal surface. We then reinjected these antibody-positive mice with 20 g of purified recombinant C7 in the abdominal area, far away from the initial injection sites on the back. Skin biopsies were obtained from the newly injected ventral areas 1 week after injection and then subjected to immunofluorescence staining for C7. Interestingly, the newly injected C7 was still able to incorporate into the mouse's BMZ (Figure 7b, panel A). In contrast, there was no C7 staining in the DEB mice that did not receive new C7 injection in the abdominal area (Figure 7b, panel B). These data indicate that the presence of anti-C7 antibodies in the animal's circulation do not prevent additional BMZ incorporation of subsequently injected C7. Furthermore, despite the anti-C7 antibody formation, none of the injected mice developed skin blisters. Therefore, the anti-C7 antibodies induced by C7 injections may not be pathogenic, unlike those in human patients who have epidermolysis bullosa acquisita (EBA).^21,22

Inhibition of anti-C7 antibody formation by MR1

We next determined whether the transient blockade of the CD40L-CD40 interaction with an anti-CD40L monoclonal antibody, MR1, could prevent anti-C7 antibody production in DEB mice. We chose to utilize MR1 as an immunosuppressant because this agent has been shown to inhibit anti-Dsg3 antibody production when normal skin containing Dsg3 was grafted onto Dsg3 null mice.²³ DEB mice were either treated with MR1 (n = 3) or control IgG (n = 3) at days 0, 2, 4, 7, 14, 21, and 28 following C7 injection. Serum samples were collected at 30, 60, and 90 days and analyzed by a NC1-based ELISA, as described.²⁰ As shown in Figure 8, the production of anti-C7 IgG was detected in all three mice treated with control IgG at day 30 and was maintained thereafter. In contrast, the mice receiving MR1 did not have detectable anti-C7 antibodies at any of the time points tested. These data indicate that MR1 is effective in preventing murine anti-C7 antibody production in DEB mice upon C7 injection.

Figure 8.

Suppression of anti-type VII collagen (anti-C7) antibody production by MR1 treatment. All of the control IgG-treated mice (n = 3) developed anti-C7 IgG about 4 weeks after protein injection, and this IgG was maintained at high levels for >3 months. This immune reaction was effectively suppressed when dystrophic epidermolysis bullosa mice were treated with MR1 (n = 3) at the time of protein injection and continued at specific time points (arrows) for 30 days, as described under Materials and Methods.

Full figure and legend (14K)

Discussion

In this study, we used a C7 knockout mouse model (Col7a1^–/–) to evaluate the feasibility of direct administration of recombinant human C7 for protein-based therapy of DEB. The feasibility of this study rested upon our ability to purify large quantities of recombinant human C7 from gene-corrected RDEB cells.¹⁷ Our study demonstrated that intradermally injected recombinant human C7 stably incorporated into the BMZ of the DEB mouse skin and formed anchoring fibrils. We showed that protein therapy corrected the DEB phenotype as evidenced by decreased skin fragility and blistering and a markedly prolonged survival of the animals. Our studies provide the first evidence for using protein therapy to correct a genetic skin disease in a preclinical animal model. This is also the first time that human C7 is shown to substitute for its murine counterpart. This latter finding is not entirely unexpected because there is an 85% identity and 90.4% homology between mouse and human C7 at the amino acid level.²⁴ This allowed us to validate the protein therapy approach in a preclinical animal model.

Our previous studies showed that the direct intradermal injection of recombinant C7 into intact mouse skin or a human RDEB skin equivalent transplanted onto mice achieved long-term protein expression and correction of the RDEB phenotype.¹⁷ However, in those studies, we found that intradermally injected C7 did not migrate to other parts of skin and remained mainly localized near the injected area. In this study, when DEB mice were injected in the back shortly after birth, in ~25% of the injected mice, we could detect C7 by immunostaining at the BMZ in sites remote from the injection site such as paw, neck, and abdomen at 1 week after injection. This is because our "intradermal" injections were actually a systemic administration of C7 into the animals due to the fact that the adherence between their epidermis and dermis was compromised over the entire animals' body. In keeping with this observation, the mice in which we detected C7 staining in sites remote from the injection were the animals with the most severe clinical phenotype at birth. However, the human C7 that was initially observed at skin sites distant from the back injection site did not persist past about 2 months (data not shown). We believe that this is because as the epidermal–dermal adherence of the mice improved and the animals developed hair that additional human C7 intradermally injected into the back skin no longer had access to these distal skin sites. It is likely that much of the C7 intradermally injected into the back skin of older pups (which had hair) bound to the BMZs surrounding the myriad of hair follicles because C7 avidly binds to type IV collagen and laminin-5 in the BMZ of hair follicles. The persistence of human C7 in murine skin is ~8 weeks. Therefore, without additional contributions from the back injections, the C7 at distal sites such as the abdomen eventually disappeared.

Without treatment, DEB mice die within the first week of life. We believe this is due to the severely compromised skin barrier which subjects the pups to fluid and salt loss and infection. The blister fluid in some young mice is equivalent to 20% of their total body weight.¹⁹ With our C7 protein therapy, however, skin blisters resolve and new blister formation diminishes. The animals have a more intact skin and >88% of these mice exhibit significantly prolonged survival. The injected human C7 improves the dermal–epidermal adherence of the animals and allows the mice to live beyond this critical first week of life. Once past this period, the animals begin to grow hair which may be protective and lessen new blister formation.

During the first 8 weeks of life, the growth rate of the C7-treated animals and their wild-type littermates was essentially the same. After 8 weeks of life, however, the C7-treated DEB mice began to grow slower than their wild-type littermates. Because our intradermally injected C7 migrated and incorporated into the BMZ of the animal's skin, but not into their oral mucosa, it is possible that the C7-treated animals continued to have oral blisters and erosions and once weaned off of mother's milk they had to eat solid food which was difficult for them. Compared with their wild-type littermates, the C7-treated DEB mice were nutritionally deprived.

One potential side effect of administering purified proteins is the induction of hypersensitivity or autoimmunity. Preventing immune responses against the new exogenous protein is potentially a central issue in the development of either protein-based or gene-based therapy for skin diseases. The problem becomes a major issue when molecular therapy corrects a recessively inherited disease caused by a null mutation, because the patient's immune system has never encountered the missing protein and may recognize all or part of the exogenously delivered C7 as a foreign protein; this would likely result in an unwanted immune response. The importance of preventing anti-C7 antibody production is further underscored by the fact that autoantibodies to C7 are associated with an acquired autoimmune blistering disease, EBA.^21,22 In this group of patients, autoantibodies bind to C7 within anchoring fibrils and induce a diminution of anchoring fibrils in the patient's skin with subsequent epidermal–dermal disadherance. One scenario would then be that we might cure genetic dystrophic DEB by protein therapy, but in the process may create an acquired autoimmune disease, EBA. In this study, we showed that C7-injected DEB mice developed circulating anti-C7 antibodies within 1 month after injection, and that the titers of these antibodies remained the same for >6 months. However, these mouse anti-C7 antibodies did not deposit in the BMZ of the mouse skin despite an abundance of human C7 at the mouse's BMZ. It is known that tissue-bound anti-C7 autoantibodies induce EBA.²⁵ These data suggest, therefore, that anti-C7 antibodies produced after C7 injection may not be pathogenic. Furthermore, we demonstrated that the presence of these circulating anti-C7 antibodies in the DEB mice did not prevent newly injected human C7 from incorporating into the BMZ of the mouse's skin.

In our previous studies, we demonstrated the pathogenicity of anti-C7 antibodies by passively transferring human or rabbit antibodies against C7 into hairless mice and inducing an EBA-like bullous disease.^26,27 In those studies, we found that the injected mice which developed an EBA-like disease had high titers of circulating antibody (1:10,000–1:20,000) when assayed by indirect immunofluorescence against either normal or salt-split human skin.^26,27 In this study, the titers of circulating anti-C7 antibody in the blood of the C7-injected mice were extremely low (1:20–1:100). This may explain why there is no detectable mouse anti-C7 IgG deposited in the skin of these mice when examined by direct immunofluorescence.

The nature of the host immune response to protein therapy depends on the immunogenicity of the protein and the underlying gene mutation. DEB patients with missense mutations may still express some of the protein but with an altered function. Nevertheless, the presence of the mutant protein may serve as a self-protein to establish self-tolerance. In this regard, the majority of DEB patients express a mutated C7 protein and have milder clinical manifestations than those associated with HS-RDEB.²⁸ Furthermore, recent studies reported that >60% of RDEB patients retain the NC1 domain of C7,²⁹ which contains the most antigenic regions within C7 for antibody production.^30,31 As discussed earlier, even with production of anti-C7 antibodies after protein therapy in the DEB mice, these antibodies neither bound directly to the BMZ of the mouse skin nor prevented the further therapeutic efficacy of newly injected human C7. Therefore, the production of specific anti-C7 antibodies in RDEB patients may not impose a major problem after protein injection.

The CD40–CD40L pathway, which upregulates B7 on antigen-presenting cells, is one of the critical costimulatory signals for T-cell activation, B-cell proliferation, and immunoglobulin secretion.^{32,33,34,35,36} Blocking this interaction with anti-CD40L antibody has been used to inhibit immune responses against gene therapy transgene products in lung, liver, or brain.^37,38 In a recent study, when skin expressing normal Dsg3 was grafted onto Dsg3 null mice, antibodies against Dsg3 were generated that caused rejection of the Dsg3 (^+/+) skin grafts. However, anti-Dsg3 IgG production was successfully suppressed in animals treated with the monoclonal MR1 antibodies that bound CD40L.²³ In this study, we found that MR1 can efficiently inhibit murine anti-C7 antibody production in C7 injected DEB mice. Furthermore, we did not observe any adverse effects of MR1 suppression on the treated mice. The advantages of anti-CD40L therapy are that it does not require knowledge of the structure of the autoantigen, and it is relatively specific because the CD40L molecule is expressed primarily on activated T cells. Our studies suggest that blockade of the CD40–CD40L interaction may be an effective way to prevent autoantibody formation in RDEB patients treated with C7 protein.

Proteinaceous agents, in general, have limited half-lives in vivo, and require repeated administration to maintain therapeutic efficacy. The current treatment of hemophilia and arthritis-related diseases requires frequent injections or infusions. Nevertheless, our data showed that injected human recombinant C7 stably incorporated into the DEB mouse BMZ and persisted there for at least 2 months. These results are consistent with our previous studies showing sustained incorporation of human C7 at the BMZ of human RDEB skin equivalents grafted onto mice after a single injection. In this regard, there may be an advantage working with collagen as a therapeutic agent, because collagens in general have slow turnover times and are stable long-lived molecules.^9,39 Therefore, some therapeutic improvement could be expected in the skin of RDEB patients by injecting sufficient recombinant collagen to make a significant allotment of anchoring fibrils. Once formed, these stable structures could persist, perhaps, for months or years.

In summary, our studies demonstrate that intradermal injection of recombinant human C7 into DEB mice can lead to restoration of C7 and anchoring fibrils at the BMZ of murine DEB skin with significant beneficial effects including decreased skin fragility and blistering and improved survival. Because currently there is no effective treatment for DEB except supportive care, the therapeutic benefits observed in these DEB mice support the notion of using C7 injections for potential therapy of DEB patients. Recombinant human C7 can be easily purified in large quantities from conditioned culture media from either gene-corrected RDEB fibroblasts or stably transfected 293 cells (2–5 mg/l).^17,40 In the long run, intradermal injections of recombinant human proteins may prove to be a technically simpler and safer procedure than other in vivo or ex vivo gene therapy approaches contemplated today. Similar to injecting bovine type I collagen (Zyderm) into humans for rhytide effacement, one could inject human C7 into the affected skin of patients with DEB. This methodology may also prove to be a therapeutic strategy applicable to other skin disorders due to defects in genes encoding for structural proteins in the skin.

Materials and Methods

Animal studies. Col7a1^+/– animals were developed as described previously¹⁹ and maintained at the animal facilities of the University of Southern California, Los Angeles, under guidelines for the care and use of animals in research. All animal studies were conducted using protocols approved by the University of Southern California Institutional Animal Use Committee. The genotype of the Col7a1^+/– animals was verified by PCR of the C7 gene with a template of genomic DNA from tail samples. Col7a1^+/– animals were clinically normal and indistinguishable from the wild-type littermates (Col7a1^+/+). Heterozygous mice were intercrossed to produce Col7a1 null (Col7a1^–/–) offspring. DEB mice were readily identified at birth by the large fluid-filled blisters developed primarily on the ventral side of the animals and the large hemorrhagic blisters on their paws. Their genotype was further confirmed by PCR.¹⁹ For protein therapy, we intradermally injected 10 g of purified recombinant human C7 suspended in 30 l of phosphate-buffered saline into the dorsal back skin of DEB mice (n = 45) once every day for the first week and then weekly thereafter using a 28 1/2 gauge needle. Therefore, depending on how many days each of these DEB mice survived (range from 4 days to 7 months), the cumulative total amounts of C7 injected into each DEB mouse varied from 20 to 300 g. The injected animals were photographed daily and assessed for therapeutic efficacy by monitoring weight gain, survival, and reduced skin blistering. At various times after injections, mouse skin biopsies were obtained from the whole body (including injected and uninjected areas) and subjected to immunostaining using a rabbit polyclonal antibody recognizing NC1 domain of human C7,⁴¹ as described in the following text. Histological sections of mouse's skin were fixed in 10% buffered formalin and stained with hematoxylin and eosin.

Blood samples were taken at the indicated times from the retro-orbital plexus using a sterile 200-l filter tip and then stored at 4 °C overnight.

Immunofluorescence staining and ultrastructural analysis of tissue. Five-micrometer thick sections of the OCT-embedded tissues were cut on a cryostat, fixed for 5 minutes in cold acetone, and air dried. Sections of skin from mice intradermally injected with C7 were incubated with a rabbit polyclonal antibody recognizing both mouse and human C7, followed by a Cy3-conjugated goat anti-rabbit IgG (Sigma, St. Louis, MO). Working dilutions were 1:500 for the primary antibody and 1:200 for the secondary antibody. Immunolabeling of the tissue was performed using standard immunofluorescence methods as described previously.⁴² Slides were mounted with 40% glycerol. Pictures from 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.⁴³ To assess human anchoring fibril formation and ultrastructure, 40-m sections were fixed in 0.1% glutaraldehyde, rinsed in 0.15 mol/l Tris, pH 7.5, and then incubated in our polyclonal anti-NC1 antibody followed by 5 nm gold secondary antibody and enhancement as described.^44,45

ELISA using recombinant NC1. The production of circulating anti-C7 antibodies was evaluated by ELISA, using the recombinant NC1 domain of C7, as previously described.²⁰ Briefly, 96-well microtiter plates (Immulon-4; Dynatch Laboratory, Alexandria, VA) were coated with purified recombinant NC1 at a concentration of 1.5 g/ml (0.15 g/well) in 20 mmol/l carbonate buffer, pH 9.3, overnight at 4 °C. The plates were washed three times with 20 mmol/l phosphate, pH 7.4, 150 mmol/l NaCl containing 0.05% Tween-20 (PBST). Nonspecific binding was reduced by blocking the plates with PBST containing 1% BSA at room temperature for 2 hours. Coated wells were subsequently incubated with sera obtained from DEB mice (dilution at 1:100 in PBST with 1% BSA) at room temperature for 2 hours. Wells were washed three times and incubated with alkaline phosphatase–conjugated goat antimouse IgG (Cappel, Aurora, OH) diluted in PBST with 1% BSA (1: 1,000) for 1 hour. The plates were then washed with PBST three times, and p-nitrophenylphosphate (Bio-Rad, Melville, NY) substrate was added and allowed to react for 4–8 minutes. Optical density (OD) was measured by absorbance at 405 nm (Bio-Tek Instruments, Winooski, VT). We calculated the mean and standard deviation for this ELISA using optical density values of control sera from 12 normal mice. On the basis of the mean, we set the cutoff values for definite positive reactivity as 0.2.

To evaluate whether there are any anti-C7 antibodies deposited directly in the skin, DEB mouse skin tissues from both injected and uninjected areas were subjected to direct immunofluorescence staining using fluorescein isothiocyanate–conjugated goat anti-mouse IgG, IgM, or IgA (Sigma, St. Louis, MO) as previously described.²¹

MR1 and control IgG treatment. Hamster anti-mouse CD40L monoclonal antibody MR1 was purchased from Taconic Farms (Taconic Farms. Germantown, NY). The DEB mice were injected intraperitoneally with either MR1 (n = 3) or control hamster IgG (n = 3) (Cappel Product, Aurora, OH). The dose of MR1 was determined based on previous reports as well as our dose–response study.^23,37,46 We used 20 g/g body weight at day 0 and 10 g/g body weight per mouse at days 2, 4, 7, 14, 21, and 28. Serum samples from these mice were collected once a month and analyzed with ELISA using human NC1, as described earlier.²⁰

References

REFERENCES

1. Lin, AN and Carter, DM (eds) (1992). Epidermolysis Bullosa: Basic and Clinical Aspects. Springer-Verlag: New York.
2. Briggaman, RA and Wheeler, CE Jr (1975). The epidermal-dermal junction. J Invest Dermatol 65: 71–84. | Article | PubMed | ISI | ChemPort |
3. 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. | Article | PubMed | ISI | ChemPort |
4. Uitto, J and Christiano, AM (1992). Molecular genetics of the cutaneous basement membrane zone. Perspectives on epidermolysis bullosa and other blistering skin diseases. J Clin Invest 90: 687–692. | Article | PubMed | ISI | ChemPort |
5. Parente, MG, Chung, LC, Ryynanen, J, Woodley, DT, Wynn, KC, Bauer, EA et al. (1991). Human type VII collagen-cDNA cloning and chromosomal mapping of the gene. Proc Natl Acad Sci USA 88: 6931–6935. | Article | PubMed | ChemPort |
6. Christiano, AM, Greenspan, DS, Lee, S and Uitto, J (1994). Cloning of human type VII collagen: complete primary sequence of the (1(VIII) chain and identification of intragenic polymorphisms. J Biol Chem 26: 20256–20262.
7. 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. | Article | PubMed | ISI | ChemPort |
8. Lunstrum, GP, Kuo, H-J, Rosenbaum, LM, Keene, DR, Glanville, RW, Sakai, LY et al. (1987). Anchoring fibrils contain the carboxyl terminal globular domain of type VII procollagen, but lack the amino terminal globular domain. J Biol Chem 262: 13706–13712. | PubMed | ISI | ChemPort |
9. Burgeson, RE (1993). Type VII collagen, anchoring fibrils, and epidermolysis bullosa. J Invest Dermatol 101: 252–255. | Article | PubMed | ISI | ChemPort |
10. Uitto, J, Hovnanian, A and Christiano, AM (1995). Premature termination codon mutations in the type VII collagen gene (COL7A1) underlie severe recessive dystrophic epidermolysis bullosa. Proc Assoc Am Phys 107: 245–252. | PubMed | ChemPort |
11. Varki, R, Sadowski, S, Uitto, J and Pfendner, E (2007). Epidermolysis bullosa. II. Type VII collagen mutations and phenotype-genotype correlations in the dystrophic subtypes. J Med Genet 44: 181–192. | Article | PubMed | ISI | ChemPort |
12. Ortiz-Urda, S, Thyagarajan, B, Keene, DR, Lin, Q, Fang, M, Calos, MP et al. (2002). Stable nonviral genetic correction of inherited human skin disease. Nat Med 8: 1166–1170. | Article | PubMed | ISI | ChemPort |
13. Chen, M, Keene, DR, Chan, LS, Hoeffler, WK, Finlay, D, Kasahara, N et al. (2002). Restoration of type VII collagen expression and function in dystrophic epidermolysis bullosa. Nat Genet 32: 670–675. | Article | PubMed | ISI | ChemPort |
14. Woodley, DT, Krueger, GG, Jorgensen, CM, Fairley, JA, Atha, T, Huang, Y 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. | Article | PubMed | ISI | ChemPort |
15. 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. | Article | PubMed | ISI | ChemPort |
16. Woodley, DT, Keene, DR, Atha, T, Huang, Y, Ram, R, Kasahara, N et al. (2004). Intradermal injection of lentiviral vectors corrects regenerated human dystrophic epidermolysis bullosa skin tissue in vivo. Mol Ther 10: 318–326. | Article | PubMed | ISI | ChemPort |
17. Woodley, DT, Keene, DR, Atha, T, Huang, Y, Lipman, K, Li, W et al. (2004). Injection of recombinant human type VII collagen restores collagen function in dystrophic epidermolysis bullosa. Nat Med 10: 693–695. | Article | PubMed | ISI | ChemPort |
18. Woodley, DT, Remington, J, Huang, Y, Hou, YP, Li, W, Keene, DR et al. (2007). Intravenously injected human fibroblasts home to skin wounds, deliver type VII collagen and promote wound healing. Mol Ther 15: 628–635. | Article | PubMed | ChemPort |
19. Heinonen, S, Mannikko, M, Klement, JF, Whitaker-Menezes, D, Murphy, GF and Uitto, J (1999). Targeted inactivation of the type VII collagen gene (Col7a1) in mice results in severe blistering phenotype: a model for recessive dystrophic epidermolysis bullosa. J Cell Sci 112: 3641–3648. | PubMed | ISI | ChemPort |
20. Chen, M, Chan, LS, Cai, X, O'Toole, EA, Samples, JC and Woodley, DT (1997). Development of an ELISA for rapid detection of anti-type VII collagen autoantibody in epidermolysis bullosa acquisita. J Invest Dermatol 108: 68–72. | Article | PubMed | ISI | ChemPort |
21. Woodley, DT, Briggaman, RA, O'Keefe, EJ, Inman, AO, Queen, LL and Gammon, WR (1984). Identification of the skin basement membrane autoantigen in epidermolysis bullosa acquisita. N Engl J Med 310: 1007–1013. | PubMed | ISI | ChemPort |
22. Woodley, DT, Burgeson, RE, Lunstrum, G, Bruckner-Tuderman, L, Reese, MJ and Briggaman, RA (1988). Epidermolysis bullosa acquisita antigen is the globular amino terminus of type VII procollagen. J Clin Invest 81: 683–687. | Article | PubMed | ISI | ChemPort |
23. Ohyama, M, Ota, T, Aoki, M, Tsunoda, K, Harada, R, Koyasu, S et al. (2003). Suppression of the immune response against exogenous desmoglein 3 in desmoglein 3 knockout mice: an implication for gene therapy. J Invest Dermatol 120: 610–615. | Article | PubMed | ChemPort |
24. Kivirikko, S, Li, K, Christiano, AM and Uitto, J (1996). Cloning of mouse type VII collagen reveals evolutionary conservation of functional protein domains and genomic organization. J Invest Dermatol 106: 1300–1306. | Article | PubMed | ISI | ChemPort |
25. Woodley, DT and Chen, M (2007). Epidermolysis bullosa acquisita. In: Freedberg, IM, Eisen, AZ, Wolff, K, Austen, KF, Goldsmith, LA and Katz, SI (eds). Fitzpatrick's Dermatology in General Medicine, 7th edn. McGraw Hill Medical: New York. pp. 494–500.
26. Woodley, DT, Chang, C, Saddat, P, Ram, R, Liu, Z and Chen, M (2005). Evidence that anti-type VII collagen antibodies are pathogenic and responsible for the clinical, histological and immunological features of epidermolysis bullosa acquisita. J Invest Dermatol 124: 958–564. | Article | PubMed | ISI | ChemPort |
27. Woodley, DT, Ram, R, Doostan, A, Bandyopadhyay, P, Huang, Y, Keene, DR et al. (2006). Induction of epidermolysis bullosa acquisita in mice by passive transfer of autoantibodies from patients. J Invest Dermatol 126: 1323–1330. | Article | PubMed | ChemPort |
28. Pulkkinen, L and Uitto, J (1999). Mutation a nalysis and molecular genetics of epidermolysis bullosa. Matrix Biol 18: 29–42. | Article | PubMed | ISI | ChemPort |
29. Ortiz-Urda, S, Garcia, J, Gren, CL, Chen, L, Lin, Q, Veitch, DP et al. (2005). Type VII collagen is required for Ras-driven human epidermal tumorigenesis. Science 307: 1773–1776. | Article | PubMed | ISI | ChemPort |
30. Gammon, WR, Murrell, DF, Jenison, MW, Padilla, KM, Prisayanh, PS, Jones, DA et al. (1993). Autoantibodies to type VII collagen recognize epitopes in a fibronectin-like region of the noncollagenous (NC1) domain. J Invest Dermatol 100: 618–622. | Article | PubMed | ISI | ChemPort |
31. Lapiere, JC, Woodley, DT, Parente, MG, Iwasaki, T, Wynn, KC, Christiano, AM et al. (1993). Epitope mapping of type VII collagen: identification of discrete peptide sequences recognized by sera from patients with acquired epidermolysis bullosa. J Clin Invest 92: 1831–1839. | Article | PubMed | ISI | ChemPort |
32. Stein, CS, Martins, I and Davidson, BL (2000). Long-term reversal of hypercholesterolemia in low density lipoprotein receptor (LDLR) deficient mice by adenovirus-mediated LDLR gene tranfer combined with CD15 blockade. J Gene Med 2: 41–51. | Article | PubMed | ISI | ChemPort |
33. Ranheim, EA and Kipps, TJ (1993). Activated T-cells induce expression of B7/BB1 on normal orleukemic B Cells through a CD40-dependent signal. J Exp Med 177: 925–935. | Article | PubMed | ISI | ChemPort |
34. Roy, M, Aruffo, A, Ledbetter, J, Linsley, P, Kehry, M and Noelle, R (1995). Studies on the interdependence of gp39 and B7 expression and function during antigen-specific immune responss. Eur Immunol 25: 596–603. | Article | ChemPort |
35. Datta, SK and Kalled, SL (1997). CD40-CD40 ligand interaction in autoimmune disease. Arthritis Rheum 40: 1735–1745. | Article | PubMed | ISI | ChemPort |
36. Yamada, A and Sayegh, MH (2002). The CD154-Cd40 costimulatory pathway in transplantation. Transplantation 73: S36–S39. | Article | PubMed | ISI | ChemPort |
37. Yang, Y, Su, Q, Grewal, IS, Schilz, R, Flavell, RA and Wilson, JM (1996). Transient subversion of CD40 ligand function diminishes immune response to adenovirus vectors in mouse liver and lung tissues. J Virol 70: 6370–6377. | PubMed | ISI | ChemPort |
38. Stein, CS, Ghodsi, A, Derksen, T and Davidson, BK (1999). Systemic and centrol nervous system correction of lysosomal storage in mucopolysaccharidosis type VII mice. J Virol 73: 3424–3429. | PubMed | ISI | ChemPort |
39. Pelkonen, R and Kivirikko, KI (1970). Hydroxyprolinemia: an apparently harmless familial metabolic disorder. N Engl J Med 283: 451–456. | PubMed | ChemPort |
40. Chen, M, Costa, FK, Lindvay, CR, Han, YP and Woodley, DT (2002). The recombinant expression of full-length type VII collagen and characterization of molecular mechanisms underlying dystrophic epidermolysis bullosa. J Biol Chem 277: 2118–21124. | Article | PubMed | ISI | ChemPort |
41. Chen, M, Petersen, PJ, Hai-Li, 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. | Article | PubMed | ISI | ChemPort |
42. 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. | Article | PubMed | ISI | ChemPort |
43. Keene, D, Sakai, LY, Lunstrum, GP, Morris, NP and Burgeson, RE (1987). Type VII collagen forms an extended network of anchoring fibrils. J Cell Biol 104: 611–621. | Article | PubMed | ISI | ChemPort |
44. 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. | Article | PubMed | ISI | ChemPort |
45. Sakai, LY and Keene, DR (1994). Fibrillin: monomers and microfibrils. In: Ruoslahti, E and Engvall, E (eds). Methods in Enzymology, vol. 245. Academic. pp. 47–50.
46. Durie, FH, Fava, RA, Foy, TM, Aruffo, A, Ledbetter, JA and Noelle, RJ (1993). Prevention of collagen-induced artiritis with an antibody to gp39, the ligand for CD40. Science 261: 1328–1330. | Article | PubMed | ISI | ChemPort |

Acknowledgments

This work was supported by grants RO1 AR47981 (to M.C.), RO1 AR33625 (to D.T.W.), and P01 AR-38923 (to J.U.) from the National Institutes of Health. We thank Sara Tufa for technical support of immuno-EM and Susan Groshen for analyzing Kaplan–Meier Survival curve.