Human Immunology 75 (2014) 354–363

Contents lists available at ScienceDirect

www.ashi-hla.org

journal homepage: www.elsevier.com/locate/humimm

IgG antibodies against immunodominant C-terminal epitopes of BP230 do not induce skin blistering in mice q Vasile Feldrihan a,b, Emilia Licarete b,c, Florina Florea b, Victor Cristea a, Octavian Popescu d,e, Cassian Sitaru b,f,⇑, Mircea Teodor Chiriac c,d,g,⇑ a

Faculty of Medicine, Iuliu-Hatieganu University of Medicine and Pharmacy, Cluj-Napoca, Romania Department of Dermatology, University of Freiburg, Freiburg, Germany Department of Biology, Babes-Bolyai University, Cluj-Napoca, Romania d Molecular Biology Center, Interdisciplinary Research Institute on Bio-Nano-Sciences, Babes-Bolyai University, Cluj-Napoca, Romania e Institute of Biology, Romanian Academy, Bucharest, Romania f BIOSS Centre for Biological Signalling Studies, University of Freiburg, Freiburg, Germany g Department of Medicine 1, University of Erlangen-Nuremberg, Erlangen, Germany b c

a r t i c l e

i n f o

Article history: Received 11 April 2012 Accepted 14 January 2014 Available online 24 January 2014

a b s t r a c t Bullous pemphigoid, the most common autoimmune blistering disease in Western Europe and the USA is characterized by the presence of circulating and tissue-bound autoantibodies against the hemidesmosomal proteins BP230 and BP180/collagen XVII. After binding to their target antigens at the basement membrane of the dermal–epidermal junction these autoantibodies are thought to trigger an inflammatory cascade comprising complement- and granulocyte-dependent reactions that result in tissue damage. Whereas the role of anti-BP180 antibodies has been extensively characterized, few and conflicting data is available on the contribution of anti-BP230 antibodies to bullous pemphigoid pathogenesis. Therefore, we addressed in the present study the role of autoantibodies to BP230 in experimental bullous pemphigoid. Rabbit polyclonal antibodies generated against epitopes of the C-terminal fragment of murine BP230 bound to the basement membrane and activated the complement system ex vivo. Affinity-purified antibodies were subsequently subcutaneously transferred into neonatal and adult BALB/c mice. In vivo, we observed a dose-dependent binding of transferred antibodies in the murine skin; however, there was no complement activation and these mice showed no clinical or histological signs of inflammatory disease, in contrast to mice receiving anti-BP180 antibodies. We further conducted ex vivo experiments and demonstrated that rabbit IgG anti-BP230-specific antibodies, in contrast to antibodies from bullous pemphigoid patients or rabbit IgG anti-BP180 antibodies used as positive controls, did not activate human granulocytes to induce dermal–epidermal separation in skin cryosections. Our present findings demonstrate that antibodies against BP230 are non-pathogenic in experimental models of bullous pemphigoid and suggest that proper activation of the complement and granulocytes represent prerequisites for conferring bullous pemphigoid autoantibodies their tissue destructive potential. Ó 2014 Published by Elsevier Inc. on behalf of American Society for Histocompatibility and Immunogenetics.

q Nonstandard abbreviations used: BP, bullous pemphigoid; BP180/type XVII collagen and BP230, bullous pemphigoid antigens of 180 kDa and 230 kDa, respectively; GST, glutathione-S-transferase; GST-mBP230-C2, fusion protein of GST and the C2-terminal fragment of the murine BP230; IF, immunofluorescence; IgG, immunoglobulin G. ⇑ Corresponding authors. Address: Universitäts-Hautklinik, Hauptstr. 7, 79104 Freiburg, Germany. Fax: +49 (0) 761 270 68290 (C. Sitaru). Address: Interdisciplinary Research Institute on Bio-Nano-Sciences, 42 August Treboniu Laurian Street, 400271 Cluj-Napoca, Romania. Fax: +40 264 414495 (M.T. Chiriac). E-mail addresses: [email protected] (C. Sitaru), [email protected] (M.T. Chiriac).

1. Introduction Bullous pemphigoid (BP) is considered to be the most prevalent autoimmune blistering disease in Western Europe and the USA [1– 4]. The disease is characterized by the presence of autoantibodies circulating in the serum of patients and binding to their target autoantigens localized at the dermal–epidermal junction in the skin and mucous membranes. Previous studies have demonstrated that BP autoantibodies predominantly recognize the bullous

http://dx.doi.org/10.1016/j.humimm.2014.01.005 0198-8859/Ó 2014 Published by Elsevier Inc. on behalf of American Society for Histocompatibility and Immunogenetics.

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

pemphigoid antigen of 180 kDa (BP180) also called type XVII collagen and bullous pemphigoid antigen of 230 kDa (BP230) [5–12]. These two hemidesmosomal proteins that mediate the contact between the intermediate filament network inside the basal cells and the underlying basal membrane are important for the integrity of the skin. Hence, either immuno-pathological conditions directed against these proteins or the spontaneous or triggered mutations in the genes encoding the BP autoantigens may result in different clinical variants of BP [11,13–16]. Compelling evidence on the pathogenetic role of BP180 has been accumulating over the past three decades. It is now well established that anti-BP180 antibodies mediate their pathogenic effects by binding to their target antigen and recruiting and activating innate immune players including the complement cascade and granulocytes that are instrumental for the course of the disease [12,16–19]. Different epitopes within the BP180 protein have been recognized to be triggered by the autoimmune response. Among these, the non-collagenous 16A domain has been shown to be the most relevant, with over 85% of BP sera reacting with this stretch of amino-acids whereas only about 10–15% of the sera recognize exclusively epitopes outside this domain [20–23]. Nevertheless, the majority of BP patients show also reactivity to BP230 [24]. Due to its intracellular location, BP230 is being widely considered of minor pathogenic relevance and few studies addressed the pathogenic relevance of the BP230-specific autoimmunity [25,26]. To investigate the possible pathogenic potential of anti-BP230 antibodies, Kiss et al. immunized rabbits with recombinant peptides from the BP230 protein and passively transferred the purified specific immunoglobulin G (IgG) to neonatal mice [25]. Using the strategy developed by Liu et al., [17] almost two decades ago, Kiss et al., demonstrated that BP230 autoantibodies have the potential to induce disease-characteristic changes in neonatal mice within few hours after their injection [25]. However, since rabbits were immunized with a construct containing both BP180- and BP230derived fragments that were covalently linked, the purified antibodies were specific to both BP180 and BP230. Therefore, an indirect effect of anti-BP230 antibodies that could have been mediated by a prior anti-BP180 effect cannot be completely excluded despite the pre-adsorption of sera against BP180 and the lack of lesions in the control group that received only anti-BP180 antibodies. Using another approach, Hall et al. immunized rabbits with an 18 amino-acid long synthetic peptide from the human BP230 antigen and found that they were resistant to experimental BP. In that study antibodies bound to their target antigens recruited C3 and inflammatory cells leading to subsequent necrosis only after irradiating the animals [26]. While partly reproducing the immunopathological findings seen in BP patients, the major drawback of this active disease model is that BP cannot be triggered without prior radiation-induced tissue injury. To address these shortcomings, in the present study we sought of investigating whether antibodies specific only to a stretch of the murine BP230 protein are capable of inducing ‘‘spontaneous’’ tissue damage by their passive transfer into mice. Immune rabbit IgG antibodies bound to the dermal–epidermal junction and activated C3 when incubated with normal mouse skin as a substrate and mouse serum as a source of complement. When passively transferred into mice, antibodies bound to their target antigens but failed to reproduce the other immuno- and histo-pathological features of BP including complement activation, granulocyte recruitment and induction of dermal–epidermal separation. Furthermore, this lack of pathogenicity was also confirmed in our ex vivo model of BP wherein, in contrast to rabbit IgG antibodies against the BP180 antigen or human BP sera, rabbit IgG antibodies to BP230 were not able to activate granulocytes and induce tissue damage in the cryosections model. Without excluding completely a possible role for anti-BP230 autoantibodies in disease pathogen-

355

esis (e.g. consecutive to and complementing anti-BP180-triggered changes), we provide here evidence that BP230-specific antibodies lack essential pathogenic capacity ex vivo and in mice, primarily due to a lack of complement fixation and granulocyte recruitment and activation. 2. Materials and methods For all experiments conducted with patients’ material we obtained informed consent according to the Helsinki Principles (Institutional Board Projects No. 318/07 and 278/11). Mouse experiments were approved by the Ethics Committee (Babes-Bolyai University, Cluj-Napoca, No. 31458/2011) and performed by qualified personnel. 2.1. Patients’ sera Human sera were collected from patients diagnosed with the disease on the following clinical and immuno-histopathological criteria [27]: (1) blisters on the skin; (2) subepidermal blister with a rich inflammatory infiltrate, consisting predominantly of eosinophils and neutrophils by histopathology; (3) linear deposits of IgG and C3 at the dermal–epidermal junction by direct immunofluorescence (IF) microscopy of perilesional skin; (4) circulating IgG autoantibodies that fixed C3 to the epidermal side of 1 M NaClsplit human skin by indirect IF microscopy and (5) reactivity to BP180 and BP230 as detected by immunoblot analysis and ELISA [28]. 2.2. Heterologous expression of murine BP230 The cDNA sequence coding for the carboxy-terminus of murine BP230 was cloned into a prokaryotic expression vector and expressed in Escherichia coli following published protocols [29]. Briefly, DNA sequence data for the murine BP230 protein was retrieved from GenBank using the accession number AF_396877. Primers used for polymerase chain reaction: forward – mBP230C2: GATCGGATCCCTGCGGCTTGGTCTGAAGACTGTCGAAGAAG; reverse – mBP230-C2: GATCGTCGACTCATCAAACGCCCTACTGGGAA GAATAGTAGAG were from MWG Biotech (Ebersberg, Germany). Restriction sites for BamHI and SalI were introduced and the BP230 fragment was cloned into linearized pGEX-6P-1 (Amersham Biosciences), resulting in the recombinant pGEX-mBP230-C2. Correct ligation and in-frame insertion of the DNA fragment was confirmed by DNA sequence analysis. The recombinant glutathione-S-transferase (GST) fusion protein was expressed in E. coli TOP 10 and purified by glutathione Sepharose affinity chromatography (Amersham Bioscience) following manufacturer’s instruction. 2.3. Generation of polyclonal rabbit IgG antibodies to murine BP230 and affinity purification of rabbit IgG antibodies Rabbit serum to the murine BP230-C2 fragment was generated by immunizing one New-Zealand rabbit with 200 lg of GST fusion protein in complete Freund’s adjuvant. The animal was boosted twice at 15 days interval with the same amount of the protein in incomplete Freund’s adjuvant. Regular bleedings were scheduled every month. IgG from rabbit serum immunized against a GSTmBP230 fusion protein (called hereafter GST-mBP230-C2) and a mix of murine BP180 fragments [30] as well as control preimmune rabbit serum were purified using Protein G Sepharose affinity column chromatography (Amersham Bioscience). Antibodies were eluted with 0.1 mol/L of glycine buffer (pH 2.5), neutralized with Tris–HCl (pH 10), and concentrated under extensive washing with phosphate-buffered saline (pH 7.2) using

356

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

Ultrafree 15 filters (Millipore, Bradford, MA). IgG concentrations were determined by absorbance at 280 nm.

after the last injection, mice were clinically re-examined, sacrificed and tissue samples and blood were collected for immuno-histopathological examination (Table 1).

2.4. Characterization of polyclonal GST-mBP230-C2-specific antibodies Circulating autoantibodies from rabbit serum were detected by indirect IF microscopy using normal and salt-split mouse skin as substrates [29]. Briefly, after incubating with diluted rabbit serum, the frozen sections of murine and human skin were treated with 100-fold diluted FITC-labeled antibodies specific to rabbit IgG (DakoCytomation, Denmark). Complement-fixing assays of autoantibodies to the dermal– epidermal junction were performed on cryosections of human and murine skin as previously described [29]. Briefly, section incubated with BP or rabbit IgG anti-GST-mBP230-C2 serum at 37 °C for 30 min were subsequently incubated with fresh human serum from healthy donors or mouse serum, respectively, as a source of mammalian complement diluted 1:5 with Gelatin Veronal Buffer (Lonza). After 60 min of incubation at room temperature, sections were washed twice with phosphate-buffered saline and complement deposition was detected using a monoclonal AlexaFluor-488-conjugated goat antibody to human or to murine C3 (both Invitrogen), respectively. For immunoblot analysis, extracts of cell-derived proteins (PAM212 cells) or recombinant protein (GST-mBP230-C2) were fractionated by 4% and 6% SDS–PAGE, respectively, transferred to nitrocellulose and probed with either a 50-fold dilution of polyclonal rabbit sera or with human BP serum and the specific binding was subsequently visualized using a 2000-fold diluted peroxidase-conjugated anti-rabbit IgG or anti-human antibody (DakoCytomation), respectively, and diaminobenzidine as a chromogenic substrate [29]. 2.5. Induction of disease in vivo and phenotype assessment Anti-GST-mBP230-C2 rabbit sera were pooled and the IgG was purified by protein G Sepharose chromatography. BALB/c neonatal mice with a body weight of 1.5 g were injected subcutaneously four (n = 5) to six times (n = 4) every 12 h with 5 mg of affinitypurified anti-mBP230-C2 antibodies, end-titer 1:3000, or six times with affinity-purified anti-mBP180 antibodies (n = 4) [30,31], endtiter 1:30000. Control animals received similar amounts of pre-immune rabbit IgG antibodies (n = 5). Adult BALB/c mice received 2  7.5 mg/day of rabbit IgG antibodies to either GST-mBP230-C2 (n = 4), BP180 (n = 4) or normal rabbit IgG (n = 2) every second day for a period of 10 days (total dose was 75 mg) (Table 1). Mice were clinically examined every 12 h after the first injection. At 24 h

2.6. The cryosections model The ability of BP patients and rabbit autoantibodies, to activate human leukocytes was assessed using an ex vivo assay of antibodyinduced granulocyte-dependent dermal–epidermal separation in cryosections of human and murine skin, respectively as imagined by Gammon et al. [32,33]. Briefly, 6 lm thin neonatal human foreskin sections and murine skin section were incubated with antibodies from BP patients or rabbit serum and subsequently with leukocytes isolated from the peripheral blood of healthy human volunteers. After 3 h of incubation at 37 °C, sections were stained with hematoxylin and eosin and examined by two investigators to appreciate the separation of the dermal–epidermal junction [27]. In alternative experiments aimed at clarifying the granulocyte-activating potential of antibodies, neonatal human foreskin sections were first incubated with BP sera or rabbit IgG antibodies and then for additional 45–60 min at 37 °C with 5  105/ml granulocytes in the presence of 0.05% nitro-blue tetrazolium as an indicator for the cells’ respiratory burst [34].

3. Results 3.1. Expression of recombinant proteins BP230 is an intracellular protein that connects the keratin filaments within cells to other proteins including BP180 that leave the basal cells and anchor themselves in the basal membrane (Fig. 1a). For the present study, we generated and used a recombinant protein designated mBP230-C2 spanning the C-terminal end of the globular C-terminal domain (amino-acid positions 2076–2611). Using epitopes in this C-terminal region have been previously shown to be the best way to detect anti-BP230 reactivity in the clinic since antibodies in BP patients mostly recognize such epitopes [35–39] (Fig. 1c). The correct sequence of the DNA was verified by sequencing. By SDS–PAGE the translated fusion protein migrated according to its expected weight of 87 kDa (Fig. 1b). After purification using affinity chromatography, the protein was used both as a substrate for immunoblotting and for rabbit immunizations.

Table 1 In vivo passive transfer experiments of the present study. Mouse groupa

a

Antibody specificityb

Dose of IgG (mg)

Observation period (days)

Immuno-reactivity Bound IgG

CBTc

Neo #1 (n = 5) Neo #1a (n = 4) Adult #1 (n = 4)

Anti-GST-mBP203-C2

5 5 7.5

20 30 75

2 3 10

+ + +

– + +

– – –

Neo #2 (n = 2) Neo #2a (n = 3) Adult #2 (n = 2)

Pre-immune IgG

5 5 7.5

20 30 75

2 3 10

– – –

– – –

– – –

Neo #3a (n = 4) Adult #3 (n = 4)

Anti-BP180v

5 7.5

30 75

3 10

+ +

+ +

+ +

per injection

Total amount

Circulating IgG

Neo-natal (Neo) or adult mice were used; n denotes numbers of mice in the respective group. Mice were injected with immune antibodies (anti-GST-mBP230-C2, groups with 1), pre-immune rabbit IgG antibodies (groups with 2) or anti-BP180 antibodies [30] (groups with 3). c CBT, complement binding test. b

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

357

Fig. 1. The generated recombinant fragment of murine BP230 migrated according to its predicted molecular mass by SDS–PAGE. (a) The structure of the hemidesmosome is based on three classes of proteins according to their localization and function, i.e., the cytoplasmic plaque proteins (including BP230, plectin acting as linkers for elements of the cytoskeleton), the transmembrane proteins (a6b4 integrin, BP180/type XVII collagen, serving as cell receptors), and finally, the basement membrane-associated proteins (laminin and type IV collagen); arrows indicate locations of antigens. (b) The recombinant protein migrated correspondingly to its calculated molecular weight of 87 kDa by SDS–PAGE. From lane 1 to 3 we loaded: molecular weight marker, GST, GST-mBP230-C2. The lower arrow indicates GST at 27 kDa while the upper indicates GST-mBP230-C2 at 87 kDa. (c) Schematic diagram of the human BP230 antigen including the amino-acid residues of the fragments that we designed in our laboratory. The C2 fragment used in this study corresponds to the C-terminal residues of the protein. Red bars indicate the sites and amino acid positions of fragments recognized by BP230 antibodies in previous studies; dashed bar [35], thin bar [36], thick bar [37]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.2. Antibodies against the mBP230-C2 fragment recognize the recombinant form of GST-mBP230-C2 by immunoblotting

plement. As shown in Fig. 2e antibodies from immune sera fixed complement ex vivo.

Rabbit IgG antibodies against the mBP230-C2 antigen were purified by affinity chromatography using protein G Sepharose as described in the methods. Both immune rabbit serum (Fig. 2a, second band) and bullous pemphigoid serum (Fig. 2a, third band) blotted the recombinant mBP230-C2 fragment in contrast to the pre-immune rabbit serum (Fig. 2a, forth band).

3.4. Passive transfer of rabbit IgG anti-GST-mBP230-C2 into neonatal and adult mice does not result in clinical phenotype or histological modifications

3.3. Rabbit IgG antibodies to GST-mBP230-C2 bind to the epidermal side of 1M NaCl-split skin and activate the complement ex vivo To assess the capacity of our rabbit anti-mBP230-C2 polyclonal antibodies to recognize the native protein in situ we incubated normal neo-natal mouse skin with immune serum and subsequently with a secondary antibody to rabbit IgG. Immune rabbit IgG antibodies bound to the basal membrane of neonatal mouse skin in a characteristic linear fashion (Fig. 2b). In contrast, antibodies from the pre-immune serum lacked the capacity to recognize the murine protein in situ (Fig. 2c). When tested for their specific localization, the antigens recognized by our immune IgG antibodies proved to be on the epidermal side of the artificial split, in accordance to their expected intracellular, supralaminar, localization (Fig. 2d). The complement fixing potential of rabbit IgG antibodies to BP230 was tested using fresh mouse serum as a source of com-

To test for the potential of rabbit IgG anti-GST-mBP230-C2 antibodies to induce clinical BP phenotype in mice we passively transferred rabbit IgG antibodies into neonatal BALB/c mice every 12 h for a period of 2 (n = 5) or 3 days (n = 4). None of the neonatal mice had any signs of clinical disease over the observation period (Fig. 3b). After the observation period, mice were sacrificed and skin biopsies were examined histologically. We could not find signs of inflammation or tissue damage in any of the mice receiving immune IgG antibodies (Fig. 3e). As expected, control mice receiving pre-immune rabbit IgG antibodies had no clinical (Fig. 3c) or histological modifications (Fig. 3f). In additional experiments, neonatal mice receiving rabbit IgG antibodies generated against murine BP180 (n = 4) did not present clinical disease (Fig. 3a) and we could not detect dermal–epidermal separation by histopathology (Fig. 3d). In contrast however, adult mice (n = 4) receiving rabbit IgG antibodies against murine BP180 that were used as a control group for our in vivo experiments, did show dermal–epidermal separation by histopathology (Fig. 3g). Despite sustained transfer of rabbit IgG antibodies to GST-mBP230-C2 over extended periods of time into adult mice (n = 4), we could not detect any

358

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

Fig. 2. Polyclonal rabbit IgG antibodies recognize the mouse BP230 and activate complement ex vivo. (a) By immunoblotting, both rabbit immune serum (band 2) and human BP serum (band 3) recognized the GST-mBP230-C2 protein (arrow indicates 87 kDa) in contrast to pre-immune rabbit serum which shows no reactivity (band 4); band 1 represents the migration pattern of the molecular weight marker. (b) Immune rabbit IgG antibodies bound to the dermal–epidermal junction of neonatal mouse skin in a characteristic fashion. In contrast, (c) antibodies isolated from rabbit serum before immunization did not recognize the murine BP230 protein in tissue sections. Incubation with 1 M NaCl-split mouse skin led to the binding of the immune antibodies to the (d) roof of the artificial blister, i.e., the place where the BP230 antigen is localized inside the basal keratinocytes. Once the antibodies bound to their target antigen on mouse skin they could fix the complement as revealed by the complement binding assay (e) using fresh mouse serum as a source of complement.

histopathological signs of disease in these animals (Fig. 3h). As expected, adult mice receiving normal rabbit IgG were free of lesions (Fig. 3i). 3.5. Transferred antibodies circulate in the sera of injected mice, bind to their target antigens but do not activate the complement in vivo To further investigate the reason behind the lack of pathogenic potential for our immune antibodies we used immunological tests. The immune rabbit IgG antibodies circulating in the sera of injected mice recognized mouse BP230 when incubated with sections of normal mouse skin by indirect IF (not shown). However, as detected by direct IF, these antibodies were capable to recognize their target antigens in vivo only in mice (n = 4) that were transferred with higher doses over 3 days (Fig. 4c) but not in mice (n = 5) receiving lower doses for only 2 days (not shown). Irrespective of the passive transfer protocol however, rabbit IgG antibodies to GST-mBP230-C2 were unable to activate the complement in vivo (Fig. 4d). Likewise, no deposits of IgG or complement were seen in mice receiving pre-immune rabbit IgG antibodies (Fig. 4e and f, respectively). In contrast, mice receiving antibodies against the BP180 auto-antigen of BP showed both IgG binding (Fig. 4a) and C3 (Fig. 4b) activation at the basal membrane (Table 1). 3.6. Unlike human antibodies, rabbit IgG antibodies to murine BP230C2 do not induce dermal–epidermal separation in the cryosections model We and others have previously shown that antibodies from BP patients have the capacity to activate granulocytes and to induce dermal–epidermal separation in cryosections of human skin. To compare the tissue damage-inducing capacity of anti mBP230-C2 antibodies to that of BP sera in the cryosections model we incubated normal human foreskin sections and murine skin sections

with either antibodies from BP patients or with antibodies from immune or pre-immune rabbit sera. Like BP sera (Fig. 5a), antibodies from the immune rabbit sera bound to the human foreskin (Fig. 5b inset). In contrast, rabbit pre-immune sera did not recognize the human substrate (not shown). When granulocytes isolated from the human peripheral blood of healthy donors were added to the sections incubated with human BP antibodies, they became specifically attached and subsequently activated at the dermal– epidermal junction as demonstrated by the dark blue formazan precipitates, the reduced form of nitro-blue tetrazolium that was used to monitor their activation (Fig. 5d). In contrast, neither immune (Fig. 5e) nor pre-immune (not shown) rabbit IgG antibodies were able to recruit and activate human granulocytes at the dermal–epidermal membrane. With longer incubation times, granulocytes activated by the antibodies from BP patients were able to induce dermal–epidermal separation (Fig. 5g inset), whereas immune (Fig. 5h inset) and pre-immune (not shown) rabbit IgG antibodies were unable to cause damage to the skin. To more specifically address the lack of pathogenicity observed in the passive transfer model using mice, we performed additional cryosections experiments in which we used sections of murine rather than human skin and incubated them with either BP sera or rabbit IgG antibodies against GST-mBP230-C2 and subsequently with human granulocytes and looked for the level of antibody-induced tissues destruction in the form of dermal–epidermal separation. In contrast to human BP antibodies which bound to the murine skin and induced dermal–epidermal separation (Fig. 5g), our immune rabbit IgG antibodies to GST-mBP230-C2 did recognize their cognate antigen (Fig. 5b) but failed to induce dermal–epidermal separation in mouse skin sections (Fig. 5h). As expected, rabbit IgG antibodies generated against murine BP180 that were used as positive controls in the cryosection model, bound to their autoantigen (Fig. 5c), activated granulocytes (Fig. 5f) and induced dermal–epidermal separation (Fig. 5i).

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

359

Fig. 3. Mice transferred with pathogenic antibodies show no clinical or histological phenotype. Neonatal mice were injected subcutaneously every 12 h with 5 mg antibodies (10 mg/24 h) and observed for 72 h after the first of the six injections. No clinical alterations were seen in mice receiving (a) immune rabbit IgG antibodies against murine BP180 (b) rabbit IgG antibodies against GST-mBP230-C2 or (c) control pre-immune antibodies. Histologically, none of the mice presented dermal–epidermal separation (d–f). Passive transfer experiments in adult mice revealed the potential of rabbit IgG antibodies to BP180 to induce histopathological changes (g). Nevertheless, adult animals receiving rabbit IgG antibodies to GST-mBP230-C2 or normal rabbit IgG antibodies were not susceptible to tissue damage albeit being injected with high doses over 10 days.

4. Discussion Bullous pemphigoid, the most common autoimmune subepidermal blistering disease, is characterized by the presence of tissue bound and circulating antibodies directed against two proteins of the basal membrane, namely BP180 and BP230 [28,40]. To understand the pathogenetic relevance of these antibodies, different models for experimental bullous pemphigoid have been generated over the past three decades [15–17,27,30–33,41–44]. While the vast majority of the existing models clearly show the pathogenicity of anti-BP180 antibodies, the potential of anti-BP230 antibodies to induce tissue destruction remained poorly understood. In the present study, in order to further our understanding of BP pathogenesis, we addressed the relevance of antibodies against murine BP230 using both in vivo and ex vivo models of experimental BP. In a first set of experiments, we generated a fragment of the murine BP230 corresponding to C-terminal residues of human BP230 that have been previously shown to be preferentially recognized by patients antibodies [36,38]. Moreover, very recent studies confirmed the fact that a C-terminal fragment represents a valuable antigenic substrate in screening patients BP sera for antiBP230 reactivity using enzyme linked immunosorbent assays [35,39]. Following the strategy that has been successfully used in several studies using the passive transfer of antibodies [17,29,45], we immunized rabbits and purified the antibodies from pooled immune sera. According to our previous experience with antibodies generated to other autoantigens [29,31], rabbit IgG anti-

bodies to GST-mBP230-C2 proved to have the preliminary proprieties that would qualify them as having good pathogenic potential in modeling disease: (1) they recognized the recombinant form of the protein they were generated to by immunoblotting; additionally, the fact that human BP serum also recognized our recombinant protein (Fig. 2a, third band) strengthened the suitability of our rabbit IgG antibodies for modeling human BP; (2) they bound to the basal membrane in cryosections of normal neo-natal mouse skin (Fig. 2b); (3) the localization of their recognized antigen proved to be on the roof of the artificial split, in accordance to their expected supralaminar localization (Fig. 4d) and (4) they fixed complement at the basal membrane in cryosections of normal murine skin (Fig. 2e). However, these features did not translate into pathogenicity ex vivo and in vivo as discussed below. To explore the capacity of these antibodies to induce experimental BP in vivo, we passively transferred purified IgG into neonatal and adult mice. Despite sustained injections over longer times and an increased amount of antibodies as compared to the previously reported models in which antibodies to BP proteins were reported to have pathogenic potential [17,25], the animals showed no clinical signs of disease throughout the observation period. The titers of BP230-specific IgG were comparable with those of antibodies generated against the murine BP180 fragments [30,31]. At the end of the observation period, we took skin biopsies and prepared them for histological examination. None of the sections we evaluated had any signs of tissue destruction and no inflammatory modifications were recorded. In line with our

360

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

Fig. 4. Antibodies to GST-mBP230-C2 bind to the dermal–epidermal junction but fail to fix complement. In contrast to rabbit IgG antibodies against murine BP180 that were able to recognize their target antigens (a) and activate complement (b) in vivo, rabbit IgG to GST-mBP230-C2 could solely bind to their specific antigens (c) but lacked the ability to subsequently activate murine complement (d). As expected, mice receiving pre-immune antibodies showed no IgG (e) or murine C3 (f) deposition at the basal membrane by direct IF.

previously published data [30,31] however, adult mice that received rabbit IgG antibodies generated against murine BP180 presented a BP-like phenotype (Fig. 3g). To investigate the causes that could underlie this lack of pathogenic potential, we analyzed the immunoreactants distribution in the blood and tissue of the injected mice. In the serum of all mice injected with immune rabbit IgG antibodies we found circulating antibodies that recognized the murine protein in the normal murine skin. However, only neo-natal mice receiving higher doses over 3 days showed deposition of the injected antibodies at the basal membrane in vivo. Likewise, adult mice injected over 10 days with a total amount of 75 mg immune rabbit IgG antibodies to GSTmBP230-C2 showed positive direct immunofluorescence. Nevertheless, complement fixation was absent in biopsies of all neo-natal and adult mice receiving rabbit IgG antibodies to GST-mBP230-C2. It has been previously shown that whole serum or purified antiBP180 antibodies from patients [27,46] as well as rabbit IgG antibodies generated against extracellular epitopes of BP180 [27] or against type VII collagen [47], another auto-antigen present at the dermal–epidermal junction [29] can induce dermal–epidermal separation when incubated with human granulocytes in the cryosections model [32,33]. In contrast, antibodies directed to an intracellular fragment of human BP180, a C-terminal fragment of hu-

man BP230 or anti-murine type VII collagen chicken IgY, the avian equivalent of mammalian IgG were incapable to activate granulocytes and subsequently induce dermal–epidermal separation in the cryosections model. Moreover, using rabbit or mouse granulocytes instead of human cells proved to be unsuccessful in generating dermal–epidermal splits in this model irrespective of the antibodies we used [48] (and our unpublished observations). Sesarman et al., clearly demonstrated that BP patients’ autoantibodies had a lower capacity to fix murine complement and a reduced ability to activate murine granulocytes when compared to the human complement and cells, respectively thus providing an explanation for their lack of blister-inducing ability [48]. Taken together these results validated the use of human granulocytes for addressing the potential of antibodies to generate tissue damage ex vivo. Therefore, to address the reason for the observed lack of pathogenicity in vivo, we employed the cryosections model using human rather than mouse peripheral blood as a source of granulocytes. Immune rabbit IgG antibodies bound to both murine and human skin in a characteristic linear fashion along the dermal–epidermal junction. However, when incubated with human granulocytes, these antibodies did not have the capacity to activate them and to subsequently induce dermal–epidermal separation. Based on our previous results, as discussed above, we excluded the fact that

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

361

Fig. 5. Rabbit IgG anti-BP230 antibodies fail to induce dermal–epidermal separation in cryosections of human skin. (a) Human BP serum, bound to the basal membrane of human foreskin sections. Likewise rabbit IgG antibodies to the murine BP230 bound to the dermal–epidermal junction of murine (b) and human (b inset) skin in the same characteristic manner as did rabbit IgG antibodies against BP180 incubated with mouse skin (c) that were used as controls. When a fresh source of granulocytes was added to the antibodies bound at this level, nitro-blue tetrazolium (NBT) was reduced to formazan precipitates only by human antibodies from BP patients (d) and rabbit IgG against BP180 (f) but not by rabbit IgG against BP230 (e). With longer incubations times, the activated granulocytes led to the characteristic dermal–epidermal split formation in slides incubated with human serum (g) and rabbit IgG anti BP180 (i), in contrast to immune rabbit IgG antibodies to GST-mBP230-C2 (h). Black arrows on panels d and f indicate the level of formazan precipitates’ formation and on panels g (g inset, red arrowheads) and i the dermal–epidermal separation (DES). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

rabbit IgG antibodies are not able to activate human granulocytes. Moreover, our previous findings were reinforced in the present study by demonstrating the ability of rabbit IgG antibodies against the murine BP180 [30,31] to induce dermal–epidermal separation in the cryosections model. One possible explanation might reside in their specificity, which sterically impedes the binding and proper activation of the granulocytes [5]. In our present study, we did not confirm previous findings by Kiss et al. in a neonatal model of autoimmunity against BP230 [25]. In that work, authors were able to induce a BP-like phenotype by injecting 5 mg of rabbit IgG anti-mBP230 into neonatal CBA mice. 24 h after injections one of the mice had blisters and more than 2/3 of the remaining ones showed persistent wrinkling of the epidermis and an erythematous background. Additionally, at the histological examination, these mice showed an intradermal granulocytic infiltration whereas their skin showed IgG and C3 deposition. The reasons we obtained such divergent result may be multiple. Some of the explanations provided below could account alone or most probably in a combined fashion for these differences: (1) the proteins used for rabbit immunizations were

different, i.e., Kiss et al. used a construct with three repetitive peptides of the BP230 human protein which shows 67% homology with the mouse protein linked to a fragment of BP180; containing a fragment of the BP180 could confer these antibodies a pathogenic relevance despite the fact that sera were pre-adsorbed before being passively transferred to neonatal mice. In addition, the authors reported that a second construct containing the BP180-derived fragment in duplicate was not able to induce disease. This may be explained by the fact that the duplicated BP180-derived fragment adopts a conformation that is not pathogenic in vivo; (2) the mouse background we used was different; (3) the clinical evaluation of mice used different methods/scores. Nevertheless, our findings are in line with previous data on the lack of pathogenesis of BP230 antibodies in a rabbit model of BP. Hall et al. were unable to induce ‘‘spontaneous’’ disease in rabbits after passive transfer of anti-human BP230 antibodies. Antibodies bound to their target antigens only in the skin of UVB-irradiated animals, radiation that caused a focus of tissue damage; subsequently these bound antibodies activated C3 and recruited inflammatory cells in situ triggering the development of BP-like lesions on a necrotic background.

362

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

However, frank blistering was not observed and animals completely recovered within 3 weeks [26]. Our present investigation demonstrated that antibodies to mBP230 are not pathogenic when passively transferred into neonatal or adult mice. The results were confirmed by the lack of tissue damaging capacity of anti-GST-mBP230-C2 antibodies in the ex vivo cryosections model. The reasons for the lack of pathogenicity of GST-mBP230-C2-specific antibodies are not known. However, a first possibility is that autoantibodies cannot bind readily in vivo to the intracellularly localized BP230 antigen and instead need a potent trigger to expose the antigen. Further, BP230-specific antibodies may lack the capacity to activate the complement system and granulocytes. Such antibodies may need a much longer administration time in vivo to reach the threshold needed for the activation of the mentioned innate immune pathways. Our recently developed adult mouse models for BP [30,31] might be the right systems to address such questions. This scenario may also be plausible for the situation in patients where the pathogenic potential of anti-BP230 has not been either proved or disproved [12] but could support ongoing disease initiated by anti-BP180 antibodies. Nevertheless, antibodies to BP230 might trigger disease by a more subtle non-inflammatory mechanism i.e., the depletion of the autoantigen from keratinocytes as suggested for anti-BP180 antibodies [44]. In conclusion, our data show that antibodies generated to an antigenic fragment of the murine BP230 cannot induce tissue destruction in an inflammatory manner in our experimental BP models. However, we cannot completely exclude the potential capacity of such antibodies to induce disease in other experimental settings or in patients. Further experiments aimed at clarifying the pathogenetic potential of anti-BP230 antibodies should prove helpful for understanding disease pathogenesis and creating a solid base for developing novel therapeutic strategies for autoantibodymediated inflammatory blistering diseases.

Funding This work was supported by Grants from the Romanian National Council for Scientific Research in Higher Education IDEI_1146/2009 (M.T.C.) and PCCE_ID_312/2008 and PCCE_ID_129/2008 (OP) and from the Deutsche Forschungsgemeinschaft SI-1281/4-1 (C.S.). V.F. received financial support from the University of Medicine and Pharmacy through an Erasmus scholarship. E.L. received financial support from the Sectoral Operational Programme for Human Resources Development 2007– 2013, co-financed by the European Social Fund, under the project number POSDRU 6/1.5/S/3 (Doctoral studies: through science towards society). M.C.T. has been financially supported through the project POSDRU/89/1.5/S/61104, co-financed by the European Social Fund under the Sectoral Operational Program for Human Resources Development 2007–2013. The funders had no role in study design, data collection and analysis, manuscript preparation and/or publication decisions.

Conflict of interest No conflict of interest has been declared.

Author’s contributions V.F., C.S. and M.T.C. designed the study. V.F., E.L., F.F. and M.T.C. performed experiments. V.F., E.L., F.F., V.C., O.P., C.S. and M.T.C. analyzed and interpreted data. C.S. and M.T.C. wrote the paper. All authors read and approved the final manuscript.

Acknowledgment The authors would like to thank Dr. Adrian Baican, Department of Dermatology, University for Medicine and Pharmacy, ClujNapoca, Romania, for providing BP sera.

References [1] Zillikens D, Wever S, Roth A, Weidenthaler-Barth B, Hashimoto T, Brocker EB. Incidence of autoimmune subepidermal blistering dermatoses in a region of central Germany. Arch Dermatol 1995;131:957. [2] Bernard P, Vaillant L, Labeille B, Bedane C, Arbeille B, Denoeux JP, et al. Incidence and distribution of subepidermal autoimmune bullous skin diseases in three French regions. Bullous Diseases French Study Group. Arch Dermatol 1995;131:48. [3] Langan SM, Smeeth L, Hubbard R, Fleming KM, Smith CJ, West J. Bullous pemphigoid and pemphigus vulgaris – incidence and mortality in the UK: population based cohort study. BMJ 2008;337:a180. [4] Marazza G, Pham HC, Scharer L, Pedrazzetti PP, Hunziker T, Trueb RM, et al. Incidence of bullous pemphigoid and pemphigus in Switzerland: a 2-year prospective study. Br J Dermatol 2009;161:861. [5] Sitaru C, Zillikens D. Mechanisms of blister induction by autoantibodies. Exp Dermatol 2005;14:861. [6] Stanley JR, Hawley-Nelson P, Yuspa SH, Shevach EM, Katz SI. Characterization of bullous pemphigoid antigen: a unique basement membrane protein of stratified squamous epithelia. Cell 1981;24:897. [7] Labib RS, Anhalt GJ, Patel HP, Mutasim DF, Diaz LA. Molecular heterogeneity of the bullous pemphigoid antigens as detected by immunoblotting. J Immunol 1986;136:1231. [8] Stanley JR, Tanaka T, Mueller S, Klaus-Kovtun V, Roop D. Isolation of complementary DNA for bullous pemphigoid antigen by use of patients’ autoantibodies. J Clin Invest 1988;82:1864. [9] Diaz LA, Ratrie 3rd H, Saunders WS, Futamura S, Squiquera HL, Anhalt GJ, et al. Isolation of a human epidermal cDNA corresponding to the 180-kD autoantigen recognized by bullous pemphigoid and herpes gestationis sera. Immunolocalization of this protein to the hemidesmosome. J Clin Invest 1990;86:1088. [10] Giudice GJ, Emery DJ, Diaz LA. Cloning and primary structural analysis of the bullous pemphigoid autoantigen BP180. J Invest Dermatol 1992;99:243. [11] Ishiko A, Shimizu H, Kikuchi A, Ebihara T, Hashimoto T, Nishikawa T. Human autoantibodies against the 230-kD bullous pemphigoid antigen (BPAG1) bind only to the intracellular domain of the hemidesmosome, whereas those against the 180-kD bullous pemphigoid antigen (BPAG2) bind along the plasma membrane of the hemidesmosome in normal human and swine skin. J Clin Invest 1993;91:1608. [12] Sitaru C. Bullous pemphigoid: a prototypical antibody-mediated organ-specific autoimmune disease. J Invest Dermatol 2009;129:822. [13] Karpati S, Stolz W, Meurer M, Braun-Falco O, Krieg T. Herpes gestationis: ultrastructural identification of the extracellular antigenic sites in diseased skin using immunogold techniques. Br J Dermatol 1991;125:317. [14] McGrath JA, Gatalica B, Christiano AM, Li K, Owaribe K, McMillan JR, et al. Mutations in the 180-kD bullous pemphigoid antigen (BPAG2), a hemidesmosomal transmembrane collagen (COL17A1), in generalized atrophic benign epidermolysis bullosa. Nat Genet 1995;11:83. [15] Nishie W, Sawamura D, Goto M, Ito K, Shibaki A, McMillan JR, et al. Humanization of autoantigen. Nat Med 2007;13:378. [16] Liu Z, Sui W, Zhao M, Li Z, Li N, Thresher R, et al. Subepidermal blistering induced by human autoantibodies to BP180 requires innate immune players in a humanized bullous pemphigoid mouse model. J Autoimmun 2008;31:331. [17] Liu Z, Diaz LA, Troy JL, Taylor AF, Emery DJ, Fairley JA, et al. A passive transfer model of the organ-specific autoimmune disease, bullous pemphigoid, using antibodies generated against the hemidesmosomal antigen, BP180. J Clin Invest 1993;92:2480. [18] Liu Z, Giudice GJ, Swartz SJ, Fairley JA, Till GO, Troy JL, et al. The role of complement in experimental bullous pemphigoid. J Clin Invest 1995;95:1539. [19] Liu Z, Giudice GJ, Zhou X, Swartz SJ, Troy JL, Fairley JA, et al. A major role for neutrophils in experimental bullous pemphigoid. J Clin Invest 1997;100:1256. [20] Giudice GJ, Emery DJ, Zelickson BD, Anhalt GJ, Liu Z, Diaz LA. Bullous pemphigoid and herpes gestationis autoantibodies recognize a common noncollagenous site on the BP180 ectodomain. J Immunol 1993;151:5742. [21] Matsumura K, Amagai M, Nishikawa T, Hashimoto T. The majority of bullous pemphigoid and herpes gestationis serum samples react with the NC16a domain of the 180-kDa bullous pemphigoid antigen. Arch Dermatol Res 1996;288:507. [22] Zillikens D, Rose PA, Balding SD, Liu Z, Olague-Marchan M, Diaz LA, et al. Tight clustering of extracellular BP180 epitopes recognized by bullous pemphigoid autoantibodies. J Invest Dermatol 1997;109:573. [23] Egan CA, Taylor TB, Meyer LJ, Petersen MJ, Zone JJ. Bullous pemphigoid sera that contain antibodies to BPAg2 also contain antibodies to LABD97 that recognize epitopes distal to the NC16A domain. J Invest Dermatol 1999;112: 148. [24] Yoshida M, Hamada T, Amagai M, Hashimoto K, Uehara R, Yamaguchi K, et al. Enzyme-linked immunosorbent assay using bacterial recombinant proteins of

V. Feldrihan et al. / Human Immunology 75 (2014) 354–363

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

human BP230 as a diagnostic tool for bullous pemphigoid. J Dermatol Sci 2006;41:21. Kiss M, Husz S, Janossy T, Marczinovits I, Molnar J, Korom I, et al. Experimental bullous pemphigoid generated in mice with an antigenic epitope of the human hemidesmosomal protein BP230. J Autoimmun 2005;24:1. Hall III RP, Murray JC, McCord MM, Rico MJ, Streilein RD. Rabbits immunized with a peptide encoded for by the 230-kD bullous pemphigoid antigen cDNA develop an enhanced inflammatory response to UVB irradiation: a potential animal model for bullous pemphigoid. J Invest Dermatol 1993;101:9. Sitaru C, Schmidt E, Petermann S, Munteanu LS, Brocker EB, Zillikens D. Autoantibodies to bullous pemphigoid antigen 180 induce dermal–epidermal separation in cryosections of human skin. J Invest Dermatol 2002;118:664. Baican A, Baican C, Chiriac G, Chiriac MT, Macovei V, Zillikens D, et al. Pemphigus vulgaris is the most common autoimmune bullous disease in Northwestern Romania. Int J Dermatol 2010;49:768. Sitaru C, Mihai S, Otto C, Chiriac MT, Hausser I, Dotterweich B, et al. Induction of dermal–epidermal separation in mice by passive transfer of antibodies specific to type VII collagen. J Clin Invest 2005;115:870. Oswald E, Sesarman A, Franzke CW, Wolfle U, Bruckner-Tuderman L, Jakob T, et al. The flavonoid luteolin inhibits Fcgamma-dependent respiratory burst in granulocytes, but not skin blistering in a new model of pemphigoid in adult mice. PLoS One 2012;7:e31066. Chiriac MT, Licarete E, Sas AG, Rados AM, Lupan I, Chiriac AM, et al. Passive transfer of collagen XVII-specific antibodies induces sustained blistering disease in adult mice. Orphanet J Rare Dis 2013;8:17. Gammon WR, Lewis DM, Carlo JR, Sams Jr WM, Wheeler Jr CE. Pemphigoid antibody mediated attachment of peripheral blood leukocytes at the dermal– epidermal junction of human skin. J Invest Dermatol 1980;75:334. Gammon WR, Merritt CC, Lewis DM, Sams Jr WM, Wheeler Jr CE, Carlo J. Leukocyte chemotaxis to the dermal–epidermal junction of human skin mediated by pemphigoid antibody and complement: mechanism of cell attachment in the in vitro leukocyte attachment method. J Invest Dermatol 1981;76:514. Chiriac MT, Roesler J, Sindrilaru A, Scharffetter-Kochanek K, Zillikens D, Sitaru C. NADPH oxidase is required for neutrophil-dependent autoantibody-induced tissue damage. J Pathol 2007;212:56. Blocker IM, Dahnrich C, Probst C, Komorowski L, Saschenbrecker S, Schlumberger W, et al. Epitope mapping of BP230 leading to a novel enzyme-linked immunosorbent assay for autoantibodies in bullous pemphigoid. Br J Dermatol 2012;166:964. Hamada T, Nagata Y, Tomita M, Salmhofer W, Hashimoto T. Bullous pemphigoid sera react specifically with various domains of BP230, most

[37]

[38] [39]

[40] [41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

363

frequently with C-terminal domain, by immunoblot analyses using bacterial recombinant proteins covering the entire molecule. Exp Dermatol 2001;10: 256. Skaria M, Jaunin F, Hunziker T, Riou S, Schumann H, Bruckner-Tuderman L, et al. IgG autoantibodies from bullous pemphigoid patients recognize multiple antigenic reactive sites located predominantly within the B and C subdomains of the COOH-terminus of BP230. J Invest Dermatol 2000;114:998. Kromminga A, Sitaru C, Hagel C, Herzog S, Zillikens D. Development of an ELISA for the detection of autoantibodies to BP230. Clin Immunol 2004;111:146. van Beek N, Knuth-Rehr D, Altmeyer P, Assaf C, Babilas P, Bayerl C, et al. Diagnostics of autoimmune bullous diseases in German dermatology departments. J Dtsch Dermatol Ges 2012;10:492. Mihai S, Sitaru C. Immunopathology and molecular diagnosis of autoimmune bullous diseases. J Cell Mol Med 2007;11:462. Anhalt GJ, Bahn CF, Labib RS, Voorhees JJ, Sugar A, Diaz LA. Pathogenic effects of bullous pemphigoid autoantibodies on rabbit corneal epithelium. J Clin Invest 1981;68:1097. Nishie W, Sawamura D, Natsuga K, Shinkuma S, Goto M, Shibaki A, et al. A novel humanized neonatal autoimmune blistering skin disease model induced by maternally transferred antibodies. J Immunol 2009;183:4088. Ujiie H, Shibaki A, Nishie W, Sawamura D, Wang G, Tateishi Y, et al. A novel active mouse model for bullous pemphigoid targeting humanized pathogenic antigen. J Immunol 2010;184:2166. Iwata H, Kamio N, Aoyama Y, Yamamoto Y, Hirako Y, Owaribe K, et al. IgG from patients with bullous pemphigoid depletes cultured keratinocytes of the 180kDa bullous pemphigoid antigen (type XVII collagen) and weakens cell attachment. J Invest Dermatol 2009;129:919. Lazarova Z, Yee C, Darling T, Briggaman RA, Yancey KB. Passive transfer of antilaminin 5 antibodies induces subepidermal blisters in neonatal mice. J Clin Invest 1996;98:1509. Mihai S, Chiriac MT, Herrero-Gonzalez JE, Goodall M, Jefferis R, Savage CO, et al. IgG4 autoantibodies induce dermal–epidermal separation. J Cell Mol Med 2007;11:1117. Sesarman A, Mihai S, Chiriac MT, Olaru F, Sitaru AG, Thurman JM, et al. Binding of avian IgY to type VII collagen does not activate complement and leucocytes and fails to induce subepidermal blistering in mice. Br J Dermatol 2008;158: 463. Sesarman A, Oswald E, Chiriac MT, Csorba K, Vuta V, Feldrihan V, et al. Why human pemphigoid autoantibodies do not trigger disease by the passive transfer into mice? Immunol Lett 2012;143:92.