Experimental bullous pemphigoid generated in mice with an antigenic epitope of the human hemidesmosomal protein BP230

Journal of Autoimmunity 24 (2005) 1e10 www.elsevier.com/locate/issn/08968411

Experimental bullous pemphigoid generated in mice with an antigenic epitope of the human hemidesmosomal protein BP230 Ma´ria Kissa,*, Sa´ndor Husza, Tama´s Ja´nossyb, Ilona Marczinovitsc, Ja´nos Molna´rc, Irma Koroma, Attila Dobozya,d a

Department of Dermatology and Allergology, University of Szeged, H-6720 Szeged, Kora´nyi fasor 6, Hungary b Institute of Surgical Research, University of Szeged, Hungary c Department of Physiology, University of Szeged, Hungary d Dermatological Research Group of the Hungarian Academy of Sciences, Szeged, Hungary Received 8 October 2003; revised 17 September 2004; accepted 22 September 2004

Abstract Bullous pemphigoid (BP) is an IgG-mediated autoimmune blistering disease that targets the hemidesmosomal proteins BP230 and BP180. To investigate the pathogenic role of anti-BP230 antibodies, rabbit polyclonal antibodies were generated against an antigenic sequence of the human BP230 antigen (BPAG 1, 2479e2499), which shows 67% homology in the human and the mouse BP230. Purified IgG from the rabbit anti-serum was transferred subcutaneously into the dorsal skin of neonatal isogeneic CBA/Ca (CBA) mice in a dose of 5 mg (n Z 7) or 1.2 mg IgG/50 ml (n Z 16). After 24 h, 1 of the mice injected with 5 mg IgG exhibited blisters, but the dorsal skin of all 7 of them was erythematous, and gentle friction produced a fine persistent wrinkling of the epidermis in 4 mice. The mice injected with 1.2 mg IgG developed less severe symptoms. Immunohistological examinations revealed linear rabbit IgG and mouse C3 depositions along the basement membrane of the perilesional skin and subepidermal blister formation. An intradermal inflammatory reaction (granulocyte infiltration) was also detected. None of these symptoms was seen in mice injected with IgG from a control rabbit anti-serum. These findings demonstrate that antibodies against BP230 can elicit the clinical and immunopathological features of BP in neonatal mice, suggesting that anti-BP230 antibodies may possibly play a pathogenic role in this disease. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: Bullous pemphigoid; Subepidermal blister formation; Protein BP230; Experimental mouse model

1. Introduction Bullous pemphigoid (BP) is an IgG-mediated autoimmune blistering skin disease that is seen predominantly in elderly people. The disease is characterized immunologically by tissue-bound and circulating autoantibodies targeting the hemidesmosomal cytoplasmic plaque protein BP230 (BPAG1) and the type II

* Corresponding author. Tel.: C36 62 545278; fax: C36 62 545954. E-mail address: [email protected] (M. Kiss). 0896-8411/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jaut.2004.09.007

transmembrane protein BP180 (also known as collagen XVII or BPAG2) [1e3]. Immunodominant and pathogenic epitopes of BP180 associated with BP have been mapped to the NC16A domain, a membrane proximal non-collagenous stretch of the extracellular portion [4]. Autoantibodies to BP180 associated with BP can release interleukin-6 and interleukin-8 from cultured human keratinocytes [5] and can induce dermaleepidermal separation in cryosections of human skin [6]. Serum levels of autoantibodies to BP180 NC16A have been shown to parallel the disease activity in BP patients [7]. The pathogenic relevance of BP180 autoantibodies has

2

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10

also been established by using experimental mouse model systems [8]. Animal model studies and human investigations have provided strong evidence that IgG antibodies which react with the NC16A domain of BP180 play a direct role in the onset and progression of the disease [8e13]. It has been demonstrated that rabbit polyclonal antibodies directed against a BP180 fusion protein can reproduce the clinical, histologic and immunopathologic features of BP when injected into neonatal mice [8]. The other autoantigen of BP is a hemidesmosomal cytoplasmic plaque protein that belongs to a family of intermediate filament binding plakin proteins. BP230 is a major isoform of the BPAG1 gene expressed in the epidermis. In addition to BP230, the BPAG1 gene also encodes several structurally distinct proteins. These alternatively spliced products exhibit different tissue distributions. BP230 is specifically expressed in basal epithelial cells. It comprises a central coiled-coil rod domain flanked by two globular end domains, with several distinct subregions [14]. Immunodominant B cell epitopes of BP230 have also been characterized, but the pathogenic role of antiBP230 antibodies remains unclear. Major antigenic epitopes of BP230 map within the C-terminal end of the protein. The C-terminal domain of BP230 is associated with keratin intermediate filaments [15]. Nevertheless, the studies did not allow any conclusions as concerns the exact distribution of the major autoantibody-reactive sites on the entire autoantigen [16e19]. Using recombinant fragments encompassing almost the total BP230 molecule, Skaria et al. [20] recently demonstrated that the region bearing the B and C subdomains of its C-terminal end contained the immunodominant sequences recognized by a majority of BP sera (84% of investigated BP patients). However, additional reactive sites are also present over extended regions of the Nterminal end of the BP230 molecule. By means of immunoblot analyses involving the use of bacterial recombinant proteins covering the entire molecule, Hamada et al. [21] confirmed that BP sera react specifically with various domains of BP230, and most frequently with C-terminal domains. In previous work, we established a sensitive and specific ELISA system for the detection of BP autoantibodies, first using synthetic immunogenic epitopes of the relevant autoantigens (BP180 and BP230) and then different biotechnologically prepared recombinant fusion proteins. We demonstrated that use of the homoand hetero-oligomers of the recombinant fusion peptides increased the sensitivity of disease-specific antibody detection in the sera of patients with BP [22]. We have a recombinant protein (GST-BP1112) which contains an antigenic segment (BP1) of the human BP230 in triplicate and an antigenic peptide (BP2) of the human BP180. When tested by an ELISA technique,

serum from 59% of the BP patients displayed specific reactivity with this recombinant protein [22]. The antigenic peptide BP1 shows high homology in the human and the mouse protein BP230. In contrast, the epitope BP2 is an antigenic part of BP180, but the sequence of BP2 peptide shows complete divergence from that of murine BP180 in this region. On this basis, antibodies generated against this GST-BP1112 recombinant protein seemed to be suitable for an investigation of the possible pathogenic role of BP230 autoantibodies in a passive transfer mouse model. As a negative counterpart in our experiments, we included another recombinant protein (GST-BP22), which contains the antigenic segment of BP2 in duplicate. Our aim was to investigate whether antiBP230 antibodies are capable of inducing the clinical and immunopathologic symptoms of BP in neonatal mice.

2. Materials and methods 2.1. Structures of the recombinant proteins used for immunization The sequences of the BP autoantigens were from the Swiss-Prot and TrEMBL databanks; they were analyzed with the Wisconsin Package, Version 8 (Genetic Computer Group, Madison, Wisconsin, USA) by means of the programs PeptideStructure and PlotStructure. The chosen epitopic fragments were as follows: BP1: WTQEPQPLEEKWQHRVVEQIP (BP230, AC Q03001; 2479e2499), BP2: RSILPYGDSMDRIEKDRLQGMAP (BP180, AC Q02802; 507e528). Schematic diagrams of the recombinant proteins and sequence comparisons of the human and murine proteins BP230 and BP180 are illustrated in Fig. 1. 2.2. Preparation of the recombinant proteins used for immunization Nucleotide sequences coding for peptides BP1 and BP2 were chemically synthesized in vitro. These synthetic DNA sequences were inserted through the BamHI site to the GST gene of the fusion-expression plasmid pGEX-4T-2 (Pharmacia, Sweden) in-frame. As a result of the cloning strategy, a dipeptide of GS (one letter code) was inserted between GST and the particular epitope sequence. In the homologous dimer (GST-BP22) and heterologous tetramer (GST-BP1112) constructs, DNA blocks encoding the peptides without stop codon were linked to one another by BamHI and BgIII restriction sites in-frame through a sequence

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10

2.3. Generation of polyclonal rabbit anti-serum against GST-BP1112 and against GST-BP22

A N NN

Z Y

1

X W

500

V

R1

R2

1000

BP1 peptide

B

1500

2000

C

C

BP230

2500 2649

(human) (mouse)

WTQEPqplEeKWqHRvvEQIP WTQEPhgtEgKWpHRaaEQIP TM

Collagen domains

C BP180

N

1

1000

500

BP2 peptide

1497

rsilpygdsmdriekdrlqgmap (human) vlyhdvqmdksnrdrlqaeapsl (mouse)

GST-BP1112

GST

PPRS

BP1

PPRS

BP1

PPRS

BP1

BP2

GST-BP22 GS

GST

PPRS

BP2

New Zealand white rabbits (4 males, with a mean weight of 3.6 kg at the beginning of the immunization) were immunized with the GST-BP1112 (70 mg peptide/ 250 mg total protein content) or the GST-BP22 (40 mg peptide/250 mg total protein content) recombinant protein, intramuscularly in complete Freund adjuvant (SigmaeAldrich, St. Louis, USA) on the first two occasions (days 0 and 7), and subcutaneously with the same doses in incomplete Freund adjuvant on days 49, 70 and 92. The immunized rabbits were killed on day 109 and sera were collected. 2.4. Testing the immune reactivity of the immunized rabbit sera

B

GS

3

BP2

Fig. 1. Sequence comparison of the antigenic peptides (BP1 and BP2) of the human and mouse proteins BP230 and BP180 (A). Schematic diagrams of the recombinant fusion products GST-BP1112 and GSTBP22 used for immunization (B). (A) The schematic diagrams are structural representations of proteins BP230 and BP180. For BP230, the various subdomains are represented in white for the eNH2 terminus, in gray for the coiled-coil domain and in dotted boxes for the eCOOH terminus [14]. For BP180, TM represents the transmembrane domain, the filled ellipse the NC16A domain, harboring the main antigenic epitopes of the protein, and the dotted box the collagenous part. The bottom portions show the amino acid alignment of the antigenic peptides BP1 and BP2 in the human and mouse proteins BP230 and BP180, respectively. Identical amino acids are designated by fat capital letters. Note the high sequence homology of BP1 between the human and the mouse BP230 and the complete sequence divergence of BP2 between the human and the mouse BP180. (B) Schematic diagrams of the recombinant products containing antigenic peptide epitopes of BP230 and BP180. Both recombinant fusion products (GST-BP1112 and GST-BP22) are composed of GST at the eNH2 end, followed by the antigenic epitopes at the eCOOH terminus. A dipeptide GS was inserted between GST and the particular epitope sequence. In the homologous and heterologous constructs, the peptides are linked through a tetrapeptide PPRS. It should be noted that the GST-BP1112 protein contains both the BP230 antigenic epitope BP1 in triplicate and the BP180 specific epitope BP2, whereas the GST-BP22 protein contains only the BP2 epitope in duplicate.

coding for a tetrapeptide PPRS. The identities of the inserted DNA blocks and the flanking plasmid regions were controlled by sequencing. Recombinant plasmids were expressed in E. coli DH5a cells and the induced recombinant proteins were purified by the method of Marczinovits et al. [23].

During the immunization, rabbits were venipunctured after different time intervals and the immune reactivities of the sera were checked. The titers of rabbit anti-GST-BP1112 and anti-GSTBP22 antibodies were tested in serial dilutions (1:30e1:5640) by indirect immunofluorescence (IIF), using normal human skin, human 1 M NaCl-split skin samples and mouse skin cryosections as substrates and FITC-conjugated goat anti-rabbit IgG (DAKO, Glostrup, Denmark). The antibody titers of the sera of the immunized rabbits were also tested by ELISA technique according to Husz et al. [22]. Western blot studies were performed by the method of Hashimoto et al. [24] with a slight modification [25]. The proteins of the epidermal extracts were separated by SDS-PAGE (with 6% separating gel) and then electrotransferred to nitrocellulose filter (BIORAD Laboratories, Hercules, USA). Immune and control rabbit sera were used to probe immunoblots at a dilution of 1:40. The specific binding of the serum was detected by using alkaline phosphatase-labeled antirabbit IgG (SigmaeAldrich, St. Louis, USA) and visualized with BCIP/NBT substrate tablets (Sigmae Aldrich, St. Louis, USA). 2.5. Administration of purified IgG from anti-GSTBP1112 and anti-GST-BP22 rabbit sera to mice Anti-GST-BP1112 and the anti-GST-BP22 rabbit antisera were pooled and the IgG was purified by protein A Sepharose chromatography. Neonatal mice of an inbred CBA/Ca strain (12e24 h old, with body weights between 1.3 and 1.8 g) were obtained from breeding colonies maintained in the laboratory. Four separate induction experiments were performed. Altogether, 7 animals were injected subcutaneously in their dorsal skin with 5.0 mg anti-GSTBP1112 rabbit IgG/50 ml, and 16 animals with 1.2 mg

4

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10

IgG/50 ml. Control mice were injected with anti-GSTBP22 and normal rabbit IgG (Sigma-Aldrich, St. Louis, USA) at the same doses or with PBS alone. 2.6. Animal evaluation The neonatal mice were examined 12 and 24 h after the subcutaneous injections. The extent of the clinical cutaneous disease was scored according to Liu et al. [9]: ÿ no detectable skin disease; C a mild erythematous reaction at the site of the injection with no evidence of the epidermal detachment sign (this sign was elicited by gentle friction of the mouse skin, which when positive, produced a fine, persistent wrinkling of the epidermis); CC intense erythema and the epidermal detachment sign in localized areas; CCC intense erythema with blister formation. Formalin-fixed, paraffin-embedded sections of dorsal skin stained with hematoxylineeosin were studied histologically. Direct immunofluorescence (DIF) investigations on the perilesional skin of the animals were performed with FITC-conjugated goat anti-rabbit IgG (DAKO, Glostrup, Denmark) and FITC-conjugated goat anti-mouse C3 (ICN, Cappel Laboratories, Aurora, Ohio, USA). 3. Results 3.1. Polyclonal rabbit immune sera were suitable for passive transfer of BP in mice The antigenic specificities of the rabbit antibodies prepared against the GST-BP1112 and GST-BP22 were characterized on human and mouse skin by IIF and Western blotting, and also by the ELISA technique with the use of antigenic epitope-coated wells. All 4 rabbit sera (2 immunized with GST-BP1112 and 2 with GST-BP22) showed linear rabbit IgG deposition along the basement membrane on normal human skin cryosections (Fig. 2a) and on the epidermal part of the basement membrane on human salt-split skin sections (Fig. 2b). The antibody titers were reasonably high: 1:2840 for the GST-BP1112-immunized rabbits and 1:5680 for the GST-BP22-immunized animals. The ELISA investigations proved the specific immune reactivity of the rabbit anti-sera. All 4 sera specifically reacted with the recombinant protein (Fig. 3a) via which the rabbits had previously been immunized. All displayed immune reactivity against GST alone (Fig. 3b). On the use of synthetic antigenic epitope-coated wells (BP1 or BP2), it was demonstrated that the sera of rabbits immunized with GST-BP22 exhibited specific immune reactivity only against antigenic peptide BP2, while the rabbit sera immunized with GST-BP1112 reacted predominantly with BP1, but also slightly with BP2 (Fig. 3c and d). The rabbit sera immunized with GST-BP1112 revealed the characteristic

Fig. 2. The rabbit immune sera against GST-BP1112 exhibited the characteristic indirect immunofluorescence (IIF) pattern on human skin cryosections and the specific pattern of BP on human epidermal protein extracts by the Western blot technique. The titers and specificities of rabbit antibodies anti-GST-BP1112 and anti-GSTBP22 were tested by IIF, using normal human skin and human 1 M NaCl-split skin samples. Rabbit immune serum against GST-BP1112 showed linear rabbit IgG deposition along the basement membrane on normal human skin cryosections (a; dilution: 1:240) and on the epidermal part of the basement membrane on human salt-split skin sections (b; dilution: 1:480). Magnification: 400!. Strong specific immune reactivity of the rabbit immune serum was demonstrated at 230 kDa on the Western blot, using human epidermal protein extract as substrate (c). Rabbits 1 and 2 were immunized with GST-BP1112.

pattern of BP on human epidermal protein extracts by the Western blotting technique (Fig. 2c). The sera of the immunized rabbits were then tested on series of CBA mice epithelial cryosections, on lip and dorsal skin samples. The sera of the rabbits immunized with GST-BP1112 showed the characteristic linear rabbit IgG deposition along the basement membrane on both epithelial mouse substrates. However, in line with the sequence differences between the BP2 human antigenic epitope and its mouse counterpart of BP180,

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10

5

Fig. 3. The rabbit immune sera showed immune reactivity against the recombinant fusion proteins GST-BP1112 and GST-BP22 used for immunization, against the GST alone and against the synthetic antigenic peptides (BP1 and BP2) by the ELISA technique. Rabbits 1 and 2 were immunized with GST-BP1112. Rabbits 3 and 4 were immunized with GST-BP22. Immobilized fusion constructs (GST-BP1112, GST-BP22 or GST) and synthetic antigenic peptides (BP1 or BP2) were incubated with each rabbit serum in serial dilutions. Binding of antibodies to immobilized antigens was visualized with o-phenylenediamine and H2O2, after incubation with a peroxidase-labeled anti-rabbit IgG [17]. The immune reactivities of the rabbit sera were expressed in optical density (OD) values. (a) The GST-BP1112 fusion product served as antigen (1 mg/100 ml/well) for rabbits 1 and 2, as did GST-BP22 (1 mg/100 ml/well) for rabbits 3 and 4. The sera of all 4 rabbits reacted specifically with the recombinant protein via which were previously immunized. (b) GST alone served as antigen (1 mg/100 ml/well). All of the sera showed immune reactivity against GST. (c) When synthetic BP1-coated wells (0.1 mg/100 ml) were used, antibodies against synthetic BP1 were detected only in rabbits 1 and 2. (d) When synthetic BP2coated wells (0.1 mg/100 ml) were used, strong immune reactivity against BP2 was observed in rabbits 3 and 4, but also slightly in rabbits 1 and 2.

the sera of the GST-BP22-immunized rabbits did not display any specific immune reactivity on mouse epithelial substrates (data not shown). Our data indicated that the specificities and titers of the rabbit GST-BP1112 and GST-BP22 immune sera were suitable for passive transfer in the experimental mouse model. 3.2. Induction of clinical and immunohistological symptoms of BP in neonatal mice Neonatal CBA mice (12e24 h old) were given subcutaneous injections of different doses (5.0 mg or 1.2 mg

IgG/50 ml) of purified IgG fractions prepared from the rabbit anti-GST-BP1112 and anti-GST-BP22 antisera and control rabbit IgG. Twenty-four hours after the injections, the mice were investigated clinically; they were then sacrificed, and histological and DIF assays were performed. Clinical, immunohistological and histological data are presented in Table 1. After 24 h, in all the control animals injected with PBS, anti-GST-BP22 or normal rabbit IgG, the sites of the injections had cleared up (Fig. 4b). However, the dorsal skin areas of most of the anti-GST-BP1112injected animals were constantly erythematous and gentle friction produced a fine persistent wrinkling of

6

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10

Table 1 Clinical, immunohistological and histological data of CBA neonatal mice treated with rabbit IgG anti-GST-BP1112 and anti-GST-BP22 Dose Number Clinical diseasea (mg IgG) of mice

Direct immunofluorescence IgG

C3

No treatment

ÿ

7

ÿ (7)

ÿ

ÿ

Normal

PBS

ÿ

5

ÿ (5)

ÿ

ÿ

Normal

Normal rabbit IgG

5

6

ÿ (6)

ÿ

ÿ

Normal

Treatment

a

Histology

a

Anti-GST-BP22 rabbit IgG

5

5

ÿ (5)

ÿ

ÿ

Normal

Anti-GST-BP22 rabbit IgG

1.2

4

ÿ (4)

ÿ

ÿ

Normal

7

C (2) constant erythema CC (4) epidermal detachment CCC (1) blister formation

G C C C

ÿ (2) C (3) ÿ (1) C (1)

Diffuse inflammatory reaction Subepidermal blister formation with granulocyte infiltration Subepidermal blister formation with intensive granulocyte infiltration

ÿ (4) ÿ C (6) constant erythema C C CC (5) epidermal C detachment C CCC (1) blister C formation

ÿ (4) C (2) ÿ (4) C (4) ÿ (1) C (4)

Normal Diffuse granulocyte infiltration

Anti-GST-BP1112 rabbit IgG 5

Anti-GST-BP1112 rabbit IgG 1.2

a

16

Subepidermal blister formation with granulocyte infiltration Subepidermal blister formation with intensive granulocyte infiltration

Number of mice in brackets.

the epidermis, and 2 animals exhibited blisters clinically (Fig. 4a; Table 1). The epidermal detachment from the underlying dermis is illustrated in Fig. 4b. The mice injected with the higher dose of anti-GST-BP1112 developed clinical symptoms that were more severe than those in the animals injected with the lower dose. Among the 7 mice injected with 5 mg IgG/50 ml of antiGST-BP1112, 1 showed blistering clinically (CCC), 4 the epidermal detachment sign (CC) and 2 constant erythema (C). Sixteen animals were injected with the lower dose of anti-GST-BP1112 (1.2 mg IgG/50 ml). Among them, 1 animal exhibited blistering clinically (CCC), 5 the epidermal detachment sign (CC), and 6 mice erythema (C), while 4 mice remained completely symptom-free (ÿ). The animals were sacrificed 24 h after the injections and skin sections were taken for DIF to detect rabbit IgG and mouse C3 deposition in the basement membrane zone. These immunohistological examinations revealed linear rabbit IgG deposition along the basement membrane of the perilesional mouse skin (Fig. 5a) in 19 of the 23 mice injected with anti-GSTBP1112, and in 11 animals subepidermal blister formation was also demonstrated (Fig. 5c). The rabbit IgG deposition labeled the epidermal roof of the vesicles (Fig. 5c). In situ bound mouse C3 was detected in 11 animals. The C3 deposition was more marked in the lesional skin (Fig. 5d), but the non-lesional skin displayed only a faint reactivity (Fig. 5b). No rabbit IgG or mouse C3 deposition was seen in the skin of the mice injected with normal or anti-GST-BP22 rabbit IgG (Table 1).

Skin sections of neonatal mice injected with IgG fractions (anti-GST-BP1112, anti-GST-BP22 and normal rabbit sera) were taken for light microscopy, with hematoxylin and eosin staining. In the mice injected with anti-GST-BP22 or normal rabbit IgG, the histological examination demonstrated a normal epidermis and dermis without any pathological alterations. In the neonatal mice injected with anti-GST-BP1112, the histological examinations revealed zones of subepidermal blister formation (Fig. 6), a majority of which was slightly inflammatory. In certain areas there was granulocyte infiltration close to the dermal surface, associated with some lesions (Fig. 6).

4. Discussion The aim of this study was to investigate the role of anti-BP230 antibodies in subepidermal blister formation in an experimental mouse model of BP. The findings demonstrated that antibodies against an antigenic epitope (BP1) of the protein BP230 (highly homologous in the human and the mouse) can produce the characteristic clinical, immunopathological and histological signs of BP in neonatal CBA mice. Since the BP230 is an intracellular molecule and BP180 is a transmembrane protein, it has been suggested that antibodies directed against BP180 should represent the pathogenic antibodies as a result of their accessibility for autoantibody induction, whereas antibodies directed against BP230 may represent an epiphenomenon. It has been speculated that patients first develop

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10

Fig. 4. Neonatal CBA mice injected subcutaneously with a purified IgG fraction prepared from rabbit anti-GST-BP1112 or anti-GSTBP22 antisera. (a) Subepidermal blister formation after 24 h (white arrow) on a CBA mouse injected with 5 mg IgG/50 ml rabbit IgG purified from rabbit anti-GST-BP1112 serum. (b) The bottom mouse, injected with 5 mg IgG/50 ml anti-GST-BP22 rabbit IgG, showed no sign of inflammation as compared with the upper mouse, which was injected with 5 mg/50 ml rabbit IgG purified from rabbit anti-GSTBP1112 serum. Epidermal detachment from the underlying dermis was demonstrated on the dorsal area of the latter animal (white arrow).

autoantibodies directed against BP180 and that, after these antibodies have bound to and damaged the epidermal cells, BP230 is made accessible and antibodies against it are formed. However, there is clinical and experimental evidence that anti-BP230 antibodies may also play an important role in the initiation of clinical symptoms and blister formation [26e29]. The pathogenic role of anti-BP230 antibodies therefore, demands further investigations. The recombinant protein we used for immunization contained the BP1 antigenic epitope of BP230 in a linear triplicate coupled to GST. In previous investigations, we demonstrated that the tandem repetitions of antigenic epitopes increased the ligand affinity and the immune reactivity with the sera of BP patients [22]. We have observed that GST-fusion forms with the antigenic epitopes exhibited an increasing contribution of alfahelix conformation in the multimer form, thereby increasing the immunogenicity of the construct [30]. Our recombinant protein also contained an antigenic epitope (BP2) of the human BP180, but the sequence of

7

this peptide is completely different from that of its mouse counterpart. In the present study, we proved that polyclonal antibodies generated against the GST-BP22 recombinant protein neither give any specific immune reactions on mouse epithelial substrates nor induce BPlike alterations in neonatal mice. Earlier attempts to induce BP by the passive transfer of patients’ autoantibodies to mice and guinea pigs failed [31,32]. Later, it was shown that the human and the murine BP180 antigen have an unusually high degree of sequence divergence in the region previously defined as the human antigenic site. Moreover, autoantibodies that specifically react with this immunodominant epitope on the human BP180 ectodomain lacked the ability to cross-react with the murine form of this protein. Using the mouse ortholog of human BP180, Liu et al. [8] have established an experimental mouse model of human BP. They have shown that anti-mouse BP180 antibodies can induce subepidermal blistering in experimental animals. When antibodies generated against a murine ortholog of BP180 are injected into neonatal BALB/c mice, the mice develop a skin disease with many of the key immunopathologic features of BP [8]. The development of lesions is dependent upon the activation of complement [9], the recruitment of neutrophils to the dermaleepidermal junction [10], and the release of proteases, including neutrophil elastase, at the lesion site [11,12]. It has further been demonstrated that mast cells and macrophages might play a role in neutrophil infiltration, which is essential for the blister formation in the mouse model [33,34]. Sitaru et al. used another model system to demonstrate the pathogenic role of BP autoantibodies. They revealed that the IgG fraction isolated from BP patients induced subepidermal splits in cryosections of normal human skin, whereas the IgG of patients depleted of reactivity to the NC16A domain of BP180 had no such ability [6]. Overall, the above data strongly characterize the direct pathogenic relevance of the autoantibody to the BP180 NC16A domain. However, the pathogenic role of the BP230 autoantibodies, directed against an intracellular autoantigen, remains unclear. The clinical and experimental findings are controversial. It has been widely demonstrated that 60e80% of BP sera contain autoantibodies against the protein BP230. Korman [28] has evaluated the antigenic specificity of eluted skinbound autoantibodies in patients with BP and has found a predominance of anti-BP230 specific autoantibodies. His study raised the possibility that these antibodies could play an initiating role in the disease development. However, it remained questionable as to how patients could develop antibodies directed against the intracellular BP230 and whether these antibodies could really cause direct basement membrane damage. Hall et al. demonstrated that rabbits immunized with a synthetic peptide of BP230 (86% sequence homology with our

8

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10

Fig. 5. Immunofluorescence analysis of the skin of neonatal CBA mice injected subcutaneously with rabbit anti-GST-BP1112. Cryosections of mouse skin treated with FITC-goat anti-rabbit IgG (a) and FITC-goat anti-mouse C3 (b), respectively (magnification: 400!). Linear staining of the basement membrane was observed in both cases. Subepidermal blister formation. Rabbit IgG (c) and mouse C3 (d) depositions labeled the epidermal roof of the vesicle.

BP1 epitope) developed antibodies to that antigen and, on exposure to UVB light, exhibited an enhanced inflammatory response, with the deposition of IgG and C3 at the basement membrane. Without UVB irradiation,

spontaneous skin disease in the rabbits did not develop [35]. We have had similar experience. Our immunized rabbits were clinically healthy and an immunohistological check on their epithelial tissues (esophagus, skin and

Fig. 6. Histological examination of the lesional skin of neonatal CBA mice injected subcutaneously with rabbit anti-GST-BP1112. Hematoxylineeosin-stained section showing the separation of the epidermis from the dermis producing a subepidermal blister. Granulocyte infiltration can be observed near the subepidermal vesicle, in the dermis.

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10

lip) revealed only a slight linear IgG deposition in certain areas of the basement membrane (data not shown). It has been observed that under certain circumstances, autoantibodies may penetrate into the intracellular milieu of living epithelial cells [36]. Immunoelectronmicroscopic studies have clearly demonstrated that human autoantibodies directed against BP230 bind only to the intracellular domain of the hemidesmosomes, while human antibodies directed against BP180 bind to both intra- and extracellular portions of the hemidesmosomes [37]. The general accepted dogma that the internal milieu of cells is immunologically privileged and cannot be reached by autoantibodies seems to be broken. It is now established that autoantibodies can penetrate living cells and may subsequently alter their function [38,39]. In summary, we have furnished the first evidence that antibodies against an antigenic epitope of the intracellular protein BP230 can produce the characteristic clinical, immunopathological and histological signs of BP in neonatal CBA mice. Our findings suggest that it is conceivable that, in consequence of a tissue injury, or even by penetrating intact cells, autoantibodies to BP230 may enter the cell, bind to the target antigen, and possibly contribute to subepidermal blister formation and disease perpetuation.

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13] [14]

Acknowledgments This work was supported by the Hungarian National Committee for Technical Development (OMFB grants 95-97-65-0985 and 02211), the Hungarian Scientific Research Fund (OTKA (grants T 032 495, T 034 964 and T 026 014), the Hungarian Ministry of Health (ETT 408 05/2000 and 413/2003) and The Netherlands Organization for Scientific Research (NWO grant 047003040).

References [1] Stanley JR. Bullous pemphigoid. In: Freedberg IM, et al, editor. Fitzpatrick’s dermatology in general medicine. New York, New York, USA: McGraw-Hill; 1999. p. 666e71. [2] Labib RS, Anhalt GJ, Patel HP, Mutasin DF, Diaz LA. Molecular heterogenicity of the bullous pemphigoid antigens as detected by immunoblotting. J Immunol 1986;136:1231e5. [3] Yancey KB, Egan CA. Pemphigoid: clinical, histologic, immunopathologic, and therapeutic considerations. JAMA 2000;284: 350e6. [4] Zillikens D, Rose PA, Balding SD, Liu Z, Olague-Marchan M, Diaz LA, et al. Tight clustering of extracellular BP 180 epitopes recognized by bullous pemphigoid autoantibodies. J Invest Dermatol 1997;109:573e9. [5] Schmidt E, Reimer S, Kruse N, Jainta S, Bro¨cker E-B, Marinkovich MP, et al. Autoantibodies to BP180 associated with

[15]

[16]

[17]

[18]

[19]

[20]

[21]

9

bullous pemphigoid release interleukin-6 and interleukin-8 from cultured human keratinocytes. J Invest Dermatol 2000;115:842e8 doi:10.1046/J.1523-1747.2000.00141.x. Sitaru C, Schmidt E, Petermann S, Munteanu LS, Bro¨cker E-B, Zillikens D. Autoantibodies to bullous pemphigoid antigen 180 induce dermaleepidermal separation in cryosections of human skin. J Invest Dermatol 2002;118:664e71 doi:10.1046/ J.1523-1747.2002.01720.x. Schmidt E, Obe K, Bro¨cker E-B, Zillikens D. Serum levels of autoantibodies to BP180 correlate with disease activity in patients with bullous pemphigoid. Arch Dermatol 2000;136:174e8. 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: 2480e8. 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;95:1539e44. 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;100:1256e63. Liu Z, Shapiro SD, Zhou X, Twining SS, Senior RM, Giudice GJ, et al. A critical role for neutrophil elastase in experimental bullous pemphigod. J Clin Invest 2000;105:113e23. Verraes S, Hornebeck W, Polette M, Borradori L, Bernard P. Respective contribution of neutrophil elastase and matrix metalloproteinase 9 in the degradation of BP180 (type XVII collagen) in human bullous pemphigoid. J Invest Dermatol 2001;117: 1091e6 doi:10.1046/J.0022-202x2001.01521.x. Liu Z. Bullous pemphigoid: using animal models to study the immunopathology. J Invest Dermatol Symp Proc 2004;9:41e6. Green KJ, Virata MLA, Elgart DW, Stanley JR, Parry DAD. Comparative structural analysis of desmoplakin, bullous pemphigoid antigen and plectin: members of a new gene family involved in organization of intermediate filaments. Int J Biol Macromol 1992;14:145e53. Yang Y, Dowling J, Yu Q-C, Kouklis P, Cleveland DW, Fuchs E. An essential cytoskeletal linker protein connecting actin microfilaments to intermediate filaments. Cell 1996;86:655e65. Gaucherand M, Nicolas JF, Paranhos Baccala G, Rouault JP, Re´ano A, Magaud JP, et al. Major antigenic epitopes of bullous pemphigoid 230 kDa antigen map within the C-terminal end of the protein. Evidence using a 55 kDa recombinant protein. Br J Dermatol 1995;132:190e6. Ide A, Hashimoto T, Amagai M, Tanaka M, Nishikawa T. Detection of autoantibodies against bullous pemphioid and pemphigus antigens by an enzyme-linked immunosorbent assay using the bacterial recombinant proteins. Exp Dermatol 1995;5: 112e6. Rico MJ, Korman NJ, Stanley JR, Tanaka T, Hall RP. IgG antibodies from patients with bullous pemphigoid bind to localized epitopes on synthetic peptides encoded by bullous pemphigoid antigen cDNA. J Immunol 1990;145:3728e33. Rico MJ, Hashimoto T, Watanabe K, Hall RP, Clark RB, Nishikawa T. Comparative epitope mapping of sera from United States (US) and Japanese patients with bullous pemphigoid (BP) to fusion proteins encoded by BPAG1. J Dermatol Sci 1996;12: 238e45. Skaria M, Jaunin F, Hunziker T, Riou S, Schumann H, BrucknerTuderman L, et al. IgG antibodies from bullous pemphigoid patients recognize multiple antigenic reactive sites located predominantly within the B and C subdomains of the COOHterminus of BP230. J Invest Dermatol 2000;114:998e1004 doi: 10.1046/J.1523-1747.2000.00893.x. Hamada T, Nagata Y, Tomita M, Salmhofer W, Hashimoto T. Bullous pemphigoid sera react specifically with various domains

10

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

M. Kiss et al. / Journal of Autoimmunity 24 (2005) 1e10 of BP230, most frequently with C-terminal domain, by immunoblot analyses using bacterial recombinant proteins covering the entire molecule. Exp Dermatol 2001;10:256e63 doi:10.1034/J. 1600-0625.2001.100405.x. Husz S, Kiss M, Molna´r K, Marczinovits I, Molna´r J, To´th GK, et al. Development of a system for detection of circulating antibodies against hemidesmosomal proteins in patients with bullous pemphigoid. Arch Dermatol Res 2000;292:217e24. Marczinovits I, Somogyi Cs, Patthy A, Ne´meth P, Molna´r J. An alternative purification protocol for producing hepatitis B virus ! antigen on a preparative scale in Escherichia coli. J Biotechnol 1997;56:81e8 doi:10.1016/S0168-1656(97)00080-1. Hashimoto T, Ogawa MM, Konohana A, Nishikawa T. Detection of pemphigus vulgaris and pemphigus foliaceus antigens by immunoblot analysis using different antigen sources. J Invest Dermatol 1990;94:327e31. Kiss M, Husz S, Molna´r K, Dobozy A. Identification of different circulating autoantibodies in patients with bullous pemphigoid and pemphigus vulgaris by means of immunoblotting. Acta Microbiol Immunol Hung 1996;43:115e23. 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;24:897e903. Mueller SM, Klaus Kovtun V, Stanley JR. A 230 kD basic protein is the major bullous pemphigoid antigen. J Invest Dermatol 1989;92:33e8. Korman NJ. In situ-bound antibodies eluted from the skin of patients with bullous pemphigoid are preferentially directed against the 230-kD bullous pemphigoid antigen. J Invest Dermatol 1995;105:824e30. Borradori L. Autoantibodies against BP230 and BP180 in bullous pemphigoid (BP): which is important? JEADV 2002;16(Suppl. 1): 6(A). Laczko´ I, Vass E, To´th GK, Marczinovits I, Kiss M, Husz S, et al. Conformational consequences of coupling bullous pemphigoid antigenic peptides to glutathione-S-transferase and their diagnostic significance. J Pept Sci 2000;6:378e86.

[31] Sams Jr WM, Gleich GJ. Failure to transfer bullous pemphigoid with serum from patients. Proc Soc Exp Biol Med 1971;136: 1027e31. [32] Gammon WR, Briggaman RA. Absence of specific histologic changes in guinea pig skin treated with bullous pemphigoid. J Invest Dermatol 1988;90:495e500. [33] Chen R, Ning G, Zhao M-L, Fleming MG, Diaz LA, Werb Z, et al. Mast cells play a key role in neutrophil recruitment in experimental bullous pemphigoid. J Clin Invest 2001;108: 1151e8. [34] Chen R, Fairley JA, Zhao M-L, Giudice GJ, Zillikens D, Diaz LA, et al. Macrophages, but not T and B lymphocytes, are critical for subepidermal blister formation in experimental bullous pemphigoid: macrophage-mediated neutrophil infiltration depends on mast cell activation. J Immunol 2002;169:3987e92. [35] Hall III RP, Murray JC, McCord MM, Joyce Rico M, 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; 101:9e14. [36] Golan TD, Ghavari AE, Elkon KB. Penetration of autoantibodies into living epithelial cells. J Invest Dermatol 1993;100: 316e22. [37] 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;91:1608e15. [38] Alarcon-Segovia D, Ruiz-Argu¨elles A, Llorente L. Broken dogma: penetration of autoantibodies into living cells. Immunol Today 1996;17:163e4. [39] Ruiz-Arguelles A, Perez-Romano B, Llorente L, AlarconSegovia D, Castellanos JM. Penetration of anti-DNA antibodies into immature live cells. J Autoimmun 1998;11:547e56 doi: 10.1006/jaut.1998.0216.