SHORT COMMUNICATION Immune reactivity to type VII collagen: implications for gene therapy of recessive dystrophic epidermolysis bullosa

Transcription

1 (2010) 17, & 2010 Macmillan Publishers Limited All rights reserved /10 $ SHORT COMMUNICATION Immune reactivity to type VII collagen: implications for gene therapy of recessive dystrophic epidermolysis bullosa V Pendaries 1,2, G Gasc 1,2, M Titeux 1,2, C Leroux 3, ZG Vitezica 4, JE Mejía 1,2,ADécha 1,2, P Loiseau 3, C Bodemer 5,6,7, C Prost-Squarcioni 8,9 and A Hovnanian 1,6,7,10 1 Inserm, U563, Toulouse, France; 2 Université Toulouse III Paul Sabatier, Toulouse, France; 3 Département d Immunologie et Histocompatibilité, Hôpital Saint-Louis, Paris, France; 4 Ecole Nationale Supérieure Agronomique de Toulouse, Auzeville-Tolosane; 5 Centre de Référence sur les Maladies Génétiques à Expression Cutanée (MAGEC), Hôpital Necker-Enfants Malades, Paris, France; 6 Département de Dermatologie, Hôpital Necker-Enfants malades, Paris, France; 7 Université Paris V René Descartes, Paris, France; 8 Centre de Référence des Maladies Bulleuses Auto-immunes Ile de France, Hôpital Avicenne, Bobigny, France; 9 UFR Léonard de Vinci, Laboratoire d Histologie et de Thérapie Génique, Université Paris 13, Bobigny, France and 10 Département de Génétique Hôpital Necker-Enfants malades, Paris, France Recessive dystrophic epidermolysis bullosa (RDEB) is a severe genodermatosis caused by loss-of-function mutations in COL7A1 encoding type VII collagen, the component of anchoring fibrils. As exogenous type VII collagen may elicit a deleterious immune response in RDEB patients during upcoming clinical trials of gene therapies or protein replacement therapies, we developed enzyme-linked immunosorbent assay (ELISA) and enzyme-linked immunosorbent spot (ELISPOT) assays to analyze B- and T-cell responses, to the full-length type VII collagen. The ELISA was highly sensitive and specific when tested against sera from 41 patients with epidermolysis bullosa acquisita (EBA), and the IFN-g ELI- SPOT detected a cellular response that correlated with ongoing EBA manifestations. Both tests were next applied to assess the risk of an immune response to type VII collagen in seven RDEB patients with a range of type VII collagen expression profiles. Immune responses against type VII collagen were dependent on the expression of type VII collagen protein, and consequently on the nature and position of the respective COL7A1 mutations. These immunologic tests will be helpful for the selection of RDEB patients for future clinical trials aiming at restoring type VII collagen expression, and in monitoring their immune response to type VII collagen after treatment. (2010) 17, ; doi: /gt ; published online 8 April 2010 Keywords: dystrophic epidermolysis bullosa; type VII collagen; clinical trial; immune response Introduction Dystrophic epidermolysis bullosa (DEB) is a severe inherited skin disorder characterized by fragility of the skin and mucous membranes because of loss-of-function mutations in COL7A1 (OMIM *120120) encoding type VII collagen. DEB has an autosomal dominant or recessive (RDEB) inheritance, the recessive forms being more severe. 1 7 Clinical RDEB ranges in severity from mild localized forms with limited blistering to the particularly severe, generalized form (RDEB-sev gen). Individuals with RDEB-sev gen suffer from lifelong severe skin and mucosal blistering followed by scarring, caused by lossof-adhesion between the epidermis and the dermis. The hands in particular are severely affected and repeated scarring leads to the fusion and retraction of digits. Correspondence: Professor A Hovnanian, Département de Dermatologie, Hôpital Necker-Enfants malades, Paris Cedex 15, France. alain.hovnanian@inserm.fr Received 11 May 2009; revised 7 September 2009; accepted 7 September 2009; published online 8 April 2010 Aggressive squamous cell carcinomas frequently develop in the areas subjected to repetitive blistering and scarring, and represent the most frequent cause of death in these patients. 6 Type VII collagen forms anchoring fibrils, which are attachment structures in the basement membrane zone, that have a major role in epidermal dermal adherence by anchoring the lamina densa to the underlying dermis. No specific treatment is currently available for RDEB, but the recent success of ex vivo gene therapy for a related disorder, junctional EB, has shown that gene transfer to epidermal stem cells is a therapeutic option. 8 We develop a similar ex vivo gene therapy approach for RDEB aimed at grafting autologous skin equivalents genetically corrected with safe COL7A1 retroviral vectors. 9 A potential obstacle to gene therapy is the development of an immune response against the newly introduced gene product that could destroy the transduced cells or neutralize the therapeutic protein. This is particularly of concern when the genetic defect results in the complete absence of the normal protein, but it could also occur in those patients who express a mutated protein. Cytotoxic immune responses against transduced cells after in vivo

2 gene transfer have been reported earlier in pre-clinical studies. 10,11 For example, hepatic gene transfer using retroviral vectors in rat models for Criggler Najar disease triggered an immune response resulting in the elimination of tranduced hepatocytes by T cells. 12 The expression or injection of therapeutic proteins can also initiate humoral immune responses, 13,14 and several studies have shown a correlation between antibodies against the therapeutic protein and treatment failure, for example, antibodies against factor VIII in mouse and dog models of hemophilia A. 15 Recently, Remington et al. 16 have shown that injection of recombinant type VII collagen in a type VII collagen knockout mouse model led to an immune response involving circulating type VII collagen antibodies, which was not located at the basement membrane zone. Although in these mice, the antibodies seemed to be no pathogenic, this study confirms the risk of developing an immune response against type VII collagen in patients who do not express this protein, as neutralizing antibodies may also well be raised. 14,17,18 Anticipating an immune response against the therapeutic protein is thus essential in the context of a clinical trial. For this, we have developed an enzyme-linked immunosorbent assay (ELISA) and an enzyme-linked immunosorbent spot (ELISPOT) assay using full-length, recombinant human type VII collagen to analyze the B- and T-cell responses to this protein. The ELISPOT technique was introduced to detect early antigendirected activation of lymphocyte subpopulations at the single cell level, 19,20 as a complement to the identification of circulating antibodies by ELISA. We have assessed the sensitivity and specificity of type VII collagen-elisa and ELISPOT across two cohorts of patients suffering from epidermolysis bullosa acquisita (EBA) and other subepidermal autoimmune bullous dermatoses (AIBD). EBA is an acquired subepidermal AIBD of the skin and mucous membranes caused by autoimmunity against type VII collagen, and is clinically highly reminiscent of hereditary DEB. Next, we have used these ELISA and ELISPOT assays to investigate preimmune reactivity against type VII collagen in seven RDEB patients with different type VII collagen expression patterns, as a prospective study on the risk of a humoral or cytotoxic immune response to this protein after gene therapy of RDEB. Development and validation of ELISA and ELISPOT tests using full-length type VII collagen as antigen Type VII collagen is secreted by keratinocytes and dermal fibroblasts, and has a homotrimeric quaternary structure. Each of the three identical 290 kda a1(vii) chains consists of three major domains: the 145 kda globular, noncollagenous domain 1 (NC1) at the amino terminus, the central helical, collagenous domain of 140 kda, and the 18 kda carboxy-terminal, noncollagenous domain 2 (NC2) (Figure 1a). Despite the fact that the major type VII collagen auto-antigenic determinants are located in the NC1 domain, western blot studies have shown that other epitopes are positioned in the NC2 or in the triple-helical central domain, especially in childhood EBA. 24 Furthermore, our in silico prediction of type VII collagen peptide binding on different Human Leukocyte Antigen (HLA) molecules showed that antigenic epitopes were located on the entire type VII collagen Immune response to collagen VII molecule although consistently at a much higher density in the NC1 domain (data not shown). Owing to potential antibodies to epitopes outside of the NC1 domain, the earlier reported NC1-based ELISA, 25 which showed great sensitivity in EBA and Crohn s disease patients, is not suitable in the context of gene therapy for RDEB. By contrast, the full-length type VII collagen ELISA that we have developed addresses a wide spectrum of potential anti-type VII collagen antibodies. COL7A1 mutations in RDEB can lead to complete nonexpression of type VII collagen, but also to truncated or missense forms of the protein where a subset of potential epitopes are missing and cannot take part to the establishment of selftolerance. This is the reason for our use of full-length recombinant type VII collagen protein in the ELISA and ELISPOT assays, enabling the detection of an immune response against any region of the molecule. First, we verified that recombinant type VII collagen adopts the triple-helical conformation of the natural protein, and assembles into functional anchoring fibrils in a human skin model 9 like native type VII collagen. Recombinant type VII collagen was purified from the conditioned medium of an RDEB keratinocyte cell line transduced with a COL7A1 retroviral vector. Nonreducing SDS PAGE and western blot analysis of FPLCpurified type VII collagen showed a band at 900 kda corresponding to type VII collagen homotrimers, and a band at 290 kda corresponding to a1(vii) monomers. Trimers were converted into monomers under reducing conditions, showing that the recombinant molecules assembled into disulfide-bonded trimers (Figure 1b; Supplementary Patients and methods). To study the stability and the triple-helical conformation of the central collagenous domain of the molecule, the protein was treated in parallel with collagenase and pepsin, and subjected to SDS PAGE and western blot analysis (Figures 1c and d). Collagenase digestion released the 145 kda NC1 domain, as described by Chen et al., 27 whereas pepsin digestion released the P1 and P2 pepsinresistant collagenous fragments of apparent molecular masses of 105 and 85 kda as well as a 200 kda fragment corresponding to the intact triple-helical domain. These results are consistent with previous studies of native type VII collagen purified from human amnion, which showed that the noncollagenous NC1 and NC2 domains of type VII collagen were degraded by pepsin, whereas the triple-helical collagenous domain was mostly pepsinresistant, except for partial cleavage at the intervening hinge region resulting in the P1 and P2 fragments. 28 The ELISA using immobilized full-length recombinant molecule was evaluated and validated against three cohorts of healthy donors, EBA patients, and other subepidermal AIBD patients, respectively (Supplementary Patients and methods). The performance of this test was analyzed using a receiver operating characteristic plot (Supplementary Figure S1), and the maximization of Youden s index established the cut-off threshold for the ELISA score at 17.9 (in arbitrary units relative to the standard). No age or sex bias was observed in positive ELISA scores. Twenty-nine EBA patients in a cohort of 41 had a score above the selected cut-off value, showing an ELISA sensitivity of 68% (95% confidence interval: 50 80%), whereas the specificity of this assay was 96% (95% confidence interval: 86 99%) (Figure 2a). Using the 17.9-units cut-off, only 3 out of 55 (B5%) other 931

3 Immune response to collagen VII 932 Pepsin Pepsin Pepsin P2 P1 NC1 TH NC2 Collagenase H M 250 kda 160 kda 105 kda M 250 kda M 250 kda 75 kda 50 kda 160 kda 105 kda NC1 160 kda 105 kda 75 kda TH P1 P2 Figure 1 Biochemical analysis of type VII collagen. A keratinocyte cell line was developed for the production of recombinant type VII collagen (BeFa-COL7A1), using SV40-immortalized RDEB keratinocytes devoid of type VII collagen expression. 50 These cells were transduced with a retroviral vector carrying the COL7A1 cdna. The protein was purified from serum-free culture supernatant by anion exchange chromatography according to Chen et al. 27 Parallel digestions of the purified recombinant type VII collagen with pepsin and collagenase (from Sigma, St Louis, MO, USA, P6887 and C0255) were performed as described. 28 (a) Schematic diagram of the type VII collagen a1(vii) chain. The triple helix (TH) domain is cleaved by pepsin digestion into the carboxyl-terminal P1 and amino-terminal P2 fragments. The TH domain is also sensitive to collagenase digestion, but the NC1 and NC2 domains are not. Protease cleavage regions are indicated by arrows. (b, c) Analysis of purified type VII collagen from BeFa-COL7A1 cells by 4 12% SDS PAGE and western blotting with the LH7.2 monoclonal antibody to the NC1 domain. The protein was either in homotrimeric (H) form (b, lane 1) or reduced to the monomeric form (M) with b-mercaptoethanol (b, lane 2). In (c), recombinant type VII collagen was either untreated (lane 1) or treated with collagenase (lane 2). In (d), recombinant type VII collagen was either untreated (lane 1) or treated with pepsin (lane 2) and analyzed by western blotting with a polyclonal antibody to type VII collagen. subepidermal AIBD patients were positive by this test (Figure 2a). These results established that the ELISA detected anti-type VII collagen antibodies with good sensitivity and was highly specific for EBA patients (Supplementary Patients and methods; Figure 2a). By comparison, only 41% of EBA patients were positive by indirect immunofluorescence, which detects circulating auto-antibodies that bind to the dermal floor of a split skin substrate (Supplementary Patients and methods), that is to either type VII collagen or to the basement membrane component, laminin 5. This suggests a higher sensitivity for ELISA relative to indirect immunofluorescence, as described by Chen et al. 25 Next, we developed and tested the first ELISPOT assay to detect Th1 and Th2 lymphocyte activation in response to full-length type VII collagen. ELISPOT immunoreactivity to recombinant type VII collagen was validated in healthy donor cells and six EBA patients for whom T lymphocytes were expected to be activated in the presence of type VII collagen, resulting in the secretion of proinflammatory cytokines such as IL-4 and/or IFN-g. T cells from healthy donors were stimulated with purified recombinant type VII collagen to measure the baseline reactivity to the protein. Although concanavalin A (25 ng ml 1 ), used in parallel as a positive control of T-cell reactivity, induced a strong Th1 (IFN-g) response with 275±159 spots per well, type VII collagen at a concentration of 10 mgml 1 induced only few spots (5.5±6 spots) with the control cells (data not shown). This was consistent with tolerance toward type VII collagen expected in healthy donors. Among six EBA patients tested, the Th1 ELISPOT allowed the detection of an ongoing immune response against type VII collagen (10 mgml 1 ) in those with no immunossupressive treatment and different degrees of clinical manifestations (Table 1; Figure 2b). The two EBA patients under immunosuppression (patients 3 and 6) failed to show a Th1 reaction against type VII collagen (scores of 1 and 2.5, respectively). Another EBA patient (patient 1, untreated) showed a low but significant response (score: 23). The last three untreated EBA patients, 2, 4, and 5, showed positive Th1 responses against type VII collagen with high ELISPOT scores of 76, 61, and 72, respectively. All ELISPOT scores from healthy donors were below 20, and a statistically significant difference was found between the ELISPOT scores of untreated EBA and those of healthy controls (P ¼ 0.021, Welch s t-test). The Th2 response (IL-4), by contrast, appeared to be low in this small cohort. Only EBA patient 2, with the highest Th1 score, developed a Th2 immune response against type VII collagen as detected by the ELISPOT assay (Figure 2b). Overall, this was consistent with the expected weak secretion of IL-4 by lymphocytes. 29 RDEB patients show different degrees of reactivity to type VII collagen After validation against EBA patients, we applied the ELISA and ELISPOT methods to seven RDEB patients with a view to predict an immune response elicited by

4 Immune response to collagen VII Th1 P2 80 Th P5 P4 P AU Number of spots P P8 20 P1 P10 20 P Controls n=17 BAID n=55 EBA n=41 P7 P9,11 P10,12,13 RDEB n=7 Figure 2 Anti-type VII collagen ELISA and ELISPOT data for EBA, other subepidermal AIBD, and RDEB patients. (a) Scatter plot representation of anti-type VII collagen antibody ELISA scores. Patient sera (diluted 1:10 and 1:100) were incubated on Nunc MaxiSorp 96-well microtiter plates coated with recombinant type VII collagen (540 ng per well); no-serum and healthy serum controls were also run on each plate. Bound anti-type VII collagen auto-antibodies were detected with peroxidase-labeled antibody to human IgG, IgA, and IgM (ab8504, Abcam, Cambridge, UK). Anti-type VII collagen antibody levels are expressed as a score relative to the standard curve, provided by a serial dilution of a pool of 10 sera positive for the indirect immunofluorescence test. The score of this reference pool is 100 in arbitrary units. Statistical analysis was performed using the statistical software R ( A receiver operating characteristic curve (Supplementary Figure S1) was compiled to evaluate the diagnostic properties of the type VII collagen ELISA and select a cut-off threshold. Note the higher score for RDEB patient 8 (P8), above the 17.9 cut-off point denoted by a horizontal line, relative to her RDEB-affected sister (P7) and the other RDEB patients (P9 P13). (b) Scatter plot representation of ELISPOT results from EBA, RDEB patients and healthy donors. Adherent monocytes and CD4 + and CD8 + T lymphocytes were purified from buffy coats or whole blood using 2 h incubation in Petri dishes and anti-cd4 and anti-cd8 MACS magnetic cell sorting (Miltenyi Biotech, Bergisch Gladbach, Germany), respectively. Patient and control CD4 + /CD8 + T cells were incubated 40 h in RPMI medium with 5% human AB serum with purified recombinant type VII collagen (10 mgml 1 ) or concanavalin A (3 mg ml 1, Sigma C5275) in ELISPOT plates coated with anti-cytokine antibodies (anti-ifn-g, clone NIB42, or anti-il-4, clone 8D4-8; both from BD Pharmingen, San Diego, CA, USA), in the presence of IL-2 (50 UI ml 1 ). Cytokine production was detected in situ with biotinylated anti-ifn-g or anti-il-4 antibody (clones 4S.B3 and MP4-25D2, respectively, from BD Pharmingen), and spots were counted under a Zeiss microscope using an ELISPOT assay software. Negative controls were run in parallel using T cells without antigen, and the corresponding scores were substracted from those of the unknowns Controls n=13 P6 P3 EBA n=6 P12,13 P11 P9 RDEB n= Controls n=8 P1,P4 P3,5,6 EBA n=6 exogenous type VII collagen during gene therapy for RDEB. These patients expressed different levels of immunodetectable type VII collagen (Supplementary Patients and methods; Table 1). Type VII collagen expression not only depends on the COL7A1 mutations involved but also on the individual genetic background of RDEB patients. 30 Of the seven RDEB patients studied, four showed no detectable type VII collagen by immunohistochemistry on skin sections (patients 7 10), whereas three patients (11, 12, and 13) showed positive but (markedly) reduced immunostaining at the basement membrane zone. Western blot analysis showed the absence of detectable type VII collagen protein in patients 7 and 8, who are sisters carrying a very early frameshift mutation (Supplementary Figure S2), whereas patient 9, who is a compound heterozygote for nonsense mutations, showed very low levels of full-length collagen VII when 100 mg of protein extracts from fibroblast primocultures (Supplementary Figure S2) were probed using a highly sensitive polyclonal antibody (a kind gift from Dr Mei Chen). RT PCR and sequence analysis of COL7A1 mrna showed no evidence of skipping of the mutated exons during pre-mrna splicing that could account for the observed protein expression. Ribosomal read-through of the nonsense mutations could underlie weak protein expression as shown earlier for other genetic diseases. 31 Patient 10 did express a short, 65 kda protein (Supplementary Figure S2), consistent with the expected truncation within the NC1 domain. The ELISA test detected no antibody reactivity for RDEB patients 7, 9, 10, 11, 12, and 13. By contrast, patient 8 showed detectable type VII collagen-binding antibodies, with an elevated ELISA score of 68 (Figure 2b; Table 1). These antibodies, however, failed to recognize anchoring fibrils in salt-split skin. The ELISPOT assay detected a T-cell-mediated immune response against type VII collagen in three of the seven RDEB patients. Two of them, patients 7 and 8 with elevated Th1 ELISPOT scores of 63 and 31 (Table 1), respectively, carried a genetic defect that resulted in the complete absence of the normal protein. This suggested that the immune system of these patients recognizes type VII collagen as a foreign antigen and points to potential rejection of type VII collagen-based gene or protein therapy. Patient 8, who showed also an elevated ELISA score, had a healthy dizygotic twin who died at birth from umbilical cord strangulation, suggesting that she may have developed type VII collagen-reactive T-cell clones through her contact with this protein in utero owing to microchimerism. Materno-fetal microchimerism may arise from the passage of small numbers of blood cells, such as hematopoietic stem cells, from the

5 Immune response to collagen VII 934 Table 1 Summary of the clinical, molecular, and immunological characterization of the EBA and RDEB patients in this study Clinical observations Defective exons or intron COL7A1 mutations AFs on transmission electron microscopy Type VII collagen immunostaining at the BMZ Type VII collagen WB IIF ELISA score Th1 ELISPOT score Th2 ELISPOT score Immunosuppressive treatment Patient 1 Localized EBA Mucosal lesions NA NA NA NA NA Negative No Patient 2 Inflammatory EBA NA NA NA NA NA No since 2 months Patient 3 EBA ongoing blistering NA NA NA NA NA Yes Patient 4 EBA ongoing blistering NA NA NA NA NA No Patient 5 EBA cutaneous fragility NA NA NA NA NA No Patient 6 EBA in remission for 5 years Patient 7 RDEB-sev gen 2 Homozygous c.189delg p.leu64trpfsx40 Patient 8 RDEB-sev gen 2 Homozygous c.189delg p.leu64trpfsx40 Patient 9 RDEB-sev gen 5, 19 c.[553c4t]+[2527c4t] p.[arg185x]+[arg843x] NA NA NA NA NA Cancer chemotherapy Patient 10 RDEB-sev gen 12 Homozygous c.1593_1603delins ACCGCATCATTGCCACCCAGTGCCG p.gly533hisfsx87 Patient 11 RDEB-sev gen 80, 114 c.[6527_6528insc] +[8440C4T] p.[gly2177trpfsx112] +[Arg2814X] Patient 12 RDEB nongeneralized 73, 94 c.[7249c4t] +[6072C4T] p.[gln2417x] +[Arg2008Gly] Patient 13 RDEB generalized other 103, intron 91 c.[7765g4t]+[7068+5g4a] p.[gly2569cys]+splice defect Negative Negative Negative Negative ND NA Negative Negative Negative Negative ND NA Negative Negative Weak Negative 10 1 ND NA Negative Negative 65 kda peptide Negative 6 22 ND NA Positive Positive ND Negative ND NA Positive Positive ND Negative ND NA Positive Positive ND Negative 5 4 ND NA Abbreviations: Afs, anchoring fibrils; BMZ, basement membrane zone; EBA, epidermolysis bullosa acquisita; ELISA, enzyme-linked immunosorbent assay; ELISPOT, enzyme-linked immunosorbent spot; IIF, indirect immunofluorescence; NA, not applicable; ND, not done; RDEB, recessive dystrophic epidermolysis bullosa; WB, western blotting.

6 mother to the fetus across the placenta, 32 and chimerism between dizygotic twins has been shown. 33 Recent experiments have shown that the mammalian fetus can develop antibody responses to antigenic stimuli. 34,35 Our observations are compatible with these recent studies, with patient 8 possibly recognizing type VII collagen as a nonself-protein by memory and/or reactive T-cell clones as a result of the suspected in utero exposure to this antigen. The likely in utero sensitization of patient 8 shows the importance of testing all RDEB patients independent of their COL7A1 mutations or type VII collagen expression, before introduction of the exogenous protein during gene or protein therapy. The third RDEB-sev gen patient (patient 10), who expressed a truncated protein of 65 kda, showed an ELISPOT score slightly above the cut-off point (22 versus 20). The 612-codon COL7A1 open reading frame in this patient encodes the most immunogenic part of the type VII collagen NC1 domain, namely the cartilage matrix protein subdomain Expression of this protein region in patient 10, and the development of immune tolerance to the major epitopes therein, may thus explain this moderate naive response to full-length type VII collagen. Conversely, patients 9, 11, 12, and 13 produced no Th1 response, with low scores of 1 4 (Figure 2b). Indeed, surprisingly, RDEB-sev gen patient 9, a compound heterozygote for two nonsense COL7A1 mutations, expressed low levels of apparently normal-size type VII collagen. The absence of an immune response in patient 9 suggests that marginal expression of normal-size type VII collagen is sufficient to prevent an immune response against this protein after genetic correction. In the same way, the missense mutations in patients 12 and 13 are compatible with the synthesis of full-length type VII collagen permitting the development of immune tolerance toward this protein. Finally, RDEB-sev gen patient 11, who did not show a response against type VII collagen, was found to express a near-complete protein. Indeed, the nonsense mutation in one of the COL7A1 alleles predicted a 280 kda protein lacking the carboxyterminal NC2 domain, a defect thought to prevent type VII collagen from dimerizing and forming anchoring fibrils. This is consistent with the absence of these structures observed by transmission electron microscopy. Conclusion The cell- and serum-based tools that we describe are useful prospective and surveillance tools for measuring the immune response against type VII collagen. The ELISA test affords highly sensitive and specific detection of anti-type VII collagen auto-antibodies. In the EBA cohort that we studied, the Th1 response measured by the ELISPOT assay correlated with clinical signs of an autoimmune response against type VII collagen in the absence of T immunosuppressive treatment. The immune response against type VII collagen in RDEB patients was dependent on RDEB patient protein expression and, consequently, on the nature and position of the respective COL7A1 mutations. These immunologic tests have the potential to identify patients who will react against a new epitope on the wild-type type VII collagen molecule, which is still possible despite the expression of a truncated or full-length mutated protein. Immunological Immune response to collagen VII investigations will thus be essential for the selection and monitoring of RDEB patients who will be candidates to gene therapy. In fact, we are developing an ex vivo gene therapy approach for RDEB based on grafting genetically engineered skin equivalents made of patient s keratinocytes and fibroblasts transduced with a selfinactivating COL7A1-expressing retroviral vector. 9 In the perspective of the development of a gene therapy phase I/II clinical trial for RDEB, these genetically engineered skin equivalents have been granted the orphan drug designation by the European Medecine Agency (EU/3/ 09/630). To minimize the risk of immune response, we focused on patients suffering from moderately severe RDEB and showing residual type VII collagen expression. Among an international cohort of about 400 patients, 30 patients have been pre-selected on the basis of their COL7A1 genotype and their type VII collagen expression. The ELISPOT and ELISA assays that we have developed will be essential to further select three or four patients to be included in this phase I/II clinical trial, that is a positive reaction to these tests will be an exclusion criterium. Finally, a role for HLA antigen presentation molecules in determining the response to exogenous type VII collagen cannot be excluded. Therefore, an HLA-based risk assessment of the antigenicity of the therapeutic protein will be performed as well for each ELISA- and ELISPOT-negative patient using peptide-binding algorithms, taking into account the patient s HLA genotype and his specific COL7A1 mutations. In addition, periodic ELISA and ELISPOT tests will be performed during the clinical trial to monitor the potential development of an immune response to type VII collagen in the weeks and months after treatment. Should the ELISA and ELISPOT tests detect an immune response after grafting, this may allow an early immunosuppressive treatment to be implemented. In later clinical trials, immunosuppression or induction of immune tolerance may also allow the inclusion of patients who show a positive naive reaction against type VII collagen in ELISPOT and ELISA assays. More widely, these predictive and monitoring tests may be applied to other therapy approaches under consideration for RDEB, involving injection of allogenic or genetically engineered autologous fibroblasts or administration of recombinant type VII collagen protein. 16,47 49 Conflict of interest The authors declare no conflict of interest. Acknowledgements We thank Dr Mei Chen for the gift of the polyclonal anti- NC1 antibody, and Professor Yann Barrandon for sharing the immortalized RDEB keratinocyte cell line and for the protein extract of patient 10 cells. We are grateful to Drs Joost van Meerwijk and Paola Romagnoli for fruitful discussion, and Drs Gilbert Fournié and Abdelhadi Saoudi for their critical reading of the manuscript. This work was supported by the Epidermolyse Bulleuse Association d Entraide (EBAE), the Association Française contre les Myopathies (AFM), the Geneskin European coordination action project, the Therapeuskin 935

7 936 European specific targeted research project, the Midi- Pyrénées Region, the French Ministry of Health through the reference centers for genetic skin diseases, and the Agence Nationale pour la Recherche (ANR, DEBCURE, No. ANR-07-BLAN-0105). References Immune response to collagen VII 1 Uitto J, Pulkkinen L, Christiano AM. Molecular basis of the dystrophic and junctional forms of epidermolysis bullosa: mutations in the type VII collagen and kalinin (laminin 5) genes. J Invest Dermatol 1994; 103: 39S 46S. 2 Uitto J, Chung-Honet LC, Christiano AM. Molecular biology and pathology of type VII collagen. Exp Dermatol 1992; 1: Hovnanian A, Hilal L, Blanchet-Bardon C, de Prost Y, Christiano AM, Uitto J et al. Recurrent nonsense mutations within the type VII collagen gene in patients with severe recessive dystrophic epidermolysis bullosa. Am J Hum Genet 1994; 55: Aumailley M, Has C, Tunggal L, Bruckner-Tuderman L. Molecular basis of inherited skin-blistering disorders, and therapeutic implications. Expert Rev Mol Med 2006; 8: Varki R, Sadowski S, Uitto J, Pfendner E. Epidermolysis bullosa. II. Type VII collagen mutations and phenotype-genotype correlations in the dystrophic subtypes. J Med Genet 2007; 44: Fine JD, Eady RA, Bauer EA, Bauer JW, Bruckner-Tuderman L, Heagerty A et al. The classification of inherited epidermolysis bullosa (EB): Report of the Third International Consensus Meeting on Diagnosis and Classification of EB. J Am Acad Dermatol 2008; 58: Dang N, Murrell DF. Mutation analysis and characterization of COL7A1 mutations in dystrophic epidermolysis bullosa. Exp Dermatol 2008; 17: Mavilio F, Pellegrini G, Ferrari S, Di Nunzio F, Di Iorio E, Recchia A et al. Correction of junctional epidermolysis bullosa by transplantation of genetically modified epidermal stem cells. Nat Med 2006; 12: Titeux M, Pendaries V, Zanta-Boussif A, Décha A, Pironon N, Tonasso L et al. SIN retroviral vectors expressing COL7A1 under human promoters for ex vivo gene therapy of recessive dystrophic epidermolysis bullosa. Mol Ther 2010 (in press). 10 Chen H, Berard J, Luo H, Landers M, Vinet B, Bradley WE et al. Compromised allograft rejection response in transgenic mice expressing antisense sequences to retinoic acid receptor beta2. J Immunol 1997; 159: Tripathy SK, Black HB, Goldwasser E, Leiden JM. Immune responses to transgene-encoded proteins limit the stability of gene expression after injection of replication-defective adenovirus vectors. Nat Med 1996; 2: Aubert D, Menoret S, Chiari E, Pichard V, Durand S, Tesson L et al. Cytotoxic immune response blunts long-term transgene expression after efficient retroviral-mediated hepatic gene transfer in rat. Mol Ther 2002; 5: Chirino AJ, Ary ML, Marshall SA. Minimizing the immunogenicity of protein therapeutics. Drug Discov Today 2004; 9: Koren E, Zuckerman LA, Mire-Sluis AR. Immune responses to therapeutic proteins in humans clinical significance, assessment and prediction. Curr Pharm Biotechnol 2002; 3: Chuah MK, Schiedner G, Thorrez L, Brown B, Johnston M, Gillijns V et al. Therapeutic factor VIII levels and negligible toxicity in mouse and dog models of hemophilia A following gene therapy with high-capacity adenoviral vectors. Blood 2003; 101: Remington J, Wang X, Hou Y, Zhou H, Burnett J, Muirhead T et al. Injection of recombinant human type VII collagen corrects the disease phenotype in a murine model of dystrophic epidermolysis bullosa. Mol Ther 2009; 17: Madoiwa S, Yamauchi T, Kobayashi E, Hakamata Y, Dokai M, Makino N et al. Induction of factor VIII-specific unresponsiveness by intrathymic factor VIII injection in murine hemophilia A. J Thromb Haemost 2009; 7: Barbosa MD, Celis E. Immunogenicity of protein therapeutics and the interplay between tolerance and antibody responses. Drug Discov Today 2007; 12: Czerkinsky CC, Nilsson LA, Nygren H, Ouchterlony O, Tarkowski A. A solid-phase enzyme-linked immunospot (ELISPOT) assay for enumeration of specific antibody-secreting cells. J Immunol Methods 1983; 65: van der Meide PH, Groenestein RJ, de Labie MC, Heeney J, Pala P, Slaoui M. Enumeration of lymphokine-secreting cells as a quantitative measure for cellular immune responses in rhesus macaques. J Med Primatol 1995; 24: Roenigk Jr HH, Ryan JG, Bergfeld WF. Epidermolysis bullosa acquisita. Report of three cases and review of all published cases. Arch Dermatol 1971; 103: Woodley DT, Burgeson RE, Lunstrum G, Bruckner-Tuderman L, Reese MJ, Briggaman RA. Epidermolysis bullosa acquisita antigen is the globular carboxyl terminus of type VII procollagen. J Clin Invest 1988; 81: Sitaru C, Mihai S, Zillikens D. The relevance of the IgG subclass of autoantibodies for blister induction in autoimmune bullous skin diseases. Arch Dermatol Res 2007; 299: Woodley DT, Remington J, Chen M. Autoimmunity to type VII collagen: epidermolysis bullosa acquisita. Clin Rev Allergy Immunol 2007; 33: Chen M, Chan LS, Cai X, O Toole EA, Sample JC, Woodley DT. Development of an ELISA for rapid detection of anti-type VII collagen autoantibodies in epidermolysis bullosa acquisita. J Invest Dermatol 1997; 108: Bruckner P, Prockop DJ. Proteolytic enzymes as probes for the triple-helical conformation of procollagen. Anal Biochem 1981; 110: Chen M, Costa FK, Lindvay CR, Han YP, Woodley DT. The recombinant expression of full-length type VII collagen and characterization of molecular mechanisms underlying dystrophic epidermolysis bullosa. J Biol Chem 2002; 277: Morris NP, Keene DR, Glanville RW, Bentz H, Burgeson RE. The tissue form of type VII collagen is an antiparallel dimer. J Biol Chem 1986; 261: Guo L, Hu-Li J, Paul WE. Probabilistic regulation in TH2 cells accounts for monoallelic expression of IL-4 and IL-13. Immunity 2005; 23: Titeux M, Pendaries V, Tonasso L, Decha A, Bodemer C, Hovnanian A. A frequent functional SNP in the MMP1 promoter is associated with higher disease severity in recessive dystrophic epidermolysis bullosa. Hum Mutat 2008; 29: Hein LK, Bawden M, Muller VJ, Sillence D, Hopwood JJ, Brooks DA. alpha-l-iduronidase premature stop codons and potential read-through in mucopolysaccharidosis type I patients. J Mol Biol 2004; 338: Yunis EJ, Zuniga J, Romero V. Chimerism and tetragametic chimerism in humans: implications in autoimmunity, allorecognition and tolerance. Immunol Res 2007; 38: Williams CA, Wallace MR, Drury KC, Kipersztok S, Edwards RK, Williams RS et al. Blood lymphocyte chimerism associated with IVF and monochorionic dizygous twinning: case report. Hum Reprod 2004; 19: Rizzi M, Gerloni M, Srivastava AS, Wheeler MC, Schuler K, Carrier E et al. In utero DNA immunisation. Immunity over tolerance in fetal life. Vaccine 2005; 23:

8 35 Watts AM, Stanley JR, Shearer MH, Hefty PS, Kennedy RC. Fetal immunization of baboons induces a fetal-specific antibody response. Nat Med 1999; 5: Lapiere JC, Woodley DT, Parente MG, Iwasaki T, Wynn KC, Christiano AM et al. Epitope mapping of type VII collagen. Identification of discrete peptide sequences recognized by sera from patients with acquired epidermolysis bullosa. J Clin Invest 1993; 92: Jones DA, Hunt III SW, Prisayanh PS, Briggaman RA, Gammon WR. Immunodominant autoepitopes of type VII collagen are short, paired peptide sequences within the fibronectin type III homology region of the noncollagenous (NC1) domain. J Invest Dermatol 1995; 104: Woodley DT, O Keefe EJ, McDonald JA, Reese MJ, Briggaman RA, Gammon WR. Specific affinity between fibronectin and the epidermolysis bullosa acquisita antigen. J Clin Invest 1987; 79: Chen M, Doostan A, Bandyopadhyay P, Remington J, Wang X, Hou Y et al. The cartilage matrix protein subdomain of type VII collagen is pathogenic for epidermolysis bullosa acquisita. Am J Pathol 2007; 170: Rocino A, Santagostino E, Mancuso ME, Mannucci PM. Immune tolerance induction with recombinant factor VIII in hemophilia A patients with high responding inhibitors. Haematologica 2006; 91: Mingozzi F, Hasbrouck NC, Basner-Tschakarjan E, Edmonson SA, Hui DJ, Sabatino DE et al. Modulation of tolerance to the transgene product in a nonhuman primate model of AAV-mediated gene transfer to liver. Blood 2007; 110: Joffre O, Santolaria T, Calise D, Al Saati T, Hudrisier D, Romagnoli P et al. Prevention of acute and chronic allograft rejection with CD4+CD25+Foxp3+ regulatory T lymphocytes. Nat Med 2008; 14: Immune response to collagen VII 43 Faria AM, Weiner HL. Oral tolerance: therapeutic implications for autoimmune diseases. Clin Dev Immunol 2006; 13: Dobrzynski E, Fitzgerald JC, Cao O, Mingozzi F, Wang L, Herzog RW. Prevention of cytotoxic T lymphocyte responses to factor IX-expressing hepatocytes by gene transfer-induced regulatory T cells. Proc Natl Acad Sci USA 2006; 103: Cao O, Furlan-Freguia C, Arruda VR, Herzog RW. Emerging role of regulatory T cells in gene transfer. Curr Gene Ther 2007; 7: Cao O, Dobrzynski E, Wang L, Nayak S, Mingle B, Terhorst C et al. Induction and role of regulatory CD4+CD25+ T cells in tolerance to the transgene product following hepatic in vivo gene transfer. Blood 2007; 110: Ortiz-Urda S, Lin Q, Green CL, Keene DR, Marinkovich MP, Khavari PA. Injection of genetically engineered fibroblasts corrects regenerated human epidermolysis bullosa skin tissue. J Clin Invest 2003; 111: Wong T, Gammon L, Liu L, Mellerio JE, Dopping-Hepenstal PJ, Pacy J et al. Potential of fibroblast cell therapy for recessive dystrophic epidermolysis bullosa. J Invest Dermatol 2008; 128: Woodley DT, Krueger GG, Jorgensen CM, Fairley JA, Atha T, Huang Y et al. Normal and gene-corrected dystrophic epidermolysis bullosa fibroblasts alone can produce type VII collagen at the basement membrane zone. J Invest Dermatol 2003; 121: Mecklenbeck S, Compton SH, Mejía JE, Cervini R, Hovnanian A, Bruckner-Tuderman L et al. A microinjected COL7A1-PAC vector restores synthesis of intact procollagen VII in a dystrophic epidermolysis bullosa keratinocyte cell line. Human 2002; 13: Supplementary Information accompanies the paper on website (