Identification of Epitopes Within Integrin β4 for Binding of Auto-Antibodies in Ocular Cicatricial and Mucous Membrane Pemphigoid: Preliminary Report

Abstract

Purpose.: To identify the epitopes on human β4 integrin to which the sera of patients with ocular cicatricial pemphigoid (OCP) and mucous membrane pemphigoid (MMP) without ocular involvement bind.

Methods.: Fragments of the intracellular domain of the β4 molecule were cloned, expressed, purified and peptides were synthesized. Antibodies to various fragments and peptides were produced in rabbits. Binding specificity was determined via Western blot and blocking experiments. Test sera and controls were injected into neonatal BALB/c mice for in vivo passive transfer.

Results.: Sera from patients with OCP, MMP, and both OCP and MMP were bound to cloned fragments of IC3.0. Its subcloned fragments IC3.4 (1489 aa–1572 aa) and IC3.4.1 (1489 aa–1510 aa) were bound with the sera from patients with OCP only. Subcloned fragments IC3.6 (1573 aa–1822 aa) and IC3.6.1 (1689 aa–1702 aa) were bound with MMP sera only. No cross-reactivity in binding was observed. Immuno-affinity–purified sera from patients with OCP, MMP, and rabbit antibodies to IC3.0, IC3.4, IC3.4.1, IC3.6, and IC3.6.1, when injected in neonatal BALB/c mice, produced subepidermal blisters in their skin.

Conclusions.: These preliminary observations identified IC3.4.1 as the possible epitope for the binding of OCP auto-antibody and IC3.6.1 as the possible epitope for the binding of MMP auto-antibody without ocular disease. Antibodies specific to these peptides produced blisters when injected in mice. Still-unidentified epitopes may exist. These observations may enhance our understanding of the role of β4 integrin in the pathobiology of OCP and MMP. Early diagnosis may be possible if serologic tests with specificity and sensitivity can be developed.

Introduction

Mucous membrane pemphigoid (MMP) is a potentially fatal multisystemic auto-immune blistering disease that affects the mucosa of the eye, nose, oral cavity, pharynx, larynx, esophagus, genitalia, anal canal, and skin. ^1,2 Its most serious aspect is irreversible scarring. When limited to the conjunctiva, or when that is the most prominent involvement, it is referred to as ocular cicatricial pemphigoid (OCP). Scarring of the conjunctiva can result in blindness in 25% of the patients. ¹ Not all patients with MMP will develop OCP. Scarring in the larynx and upper airway can cause asphyxiation and sudden death, scarring in the esophagus can cause stenosis, and vaginal and penile scarring have grave consequences. Scarring of the anal mucosa results in fecal leakage. 

The 2 major issues in MMP are delays in diagnosis and therapy. Patients with MMP produce auto-antibodies to molecules in the basement membrane zone (BMZ). Because the titer of the anti-BMZ antibodies is low, it may not be measurable. Biopsy of these mucous membranes is difficult, partly because of poor access. Fragility can result in inadequate amounts of tissue for a definitive diagnosis. 

Specific serologic assays have not been available, in part owing to the lack of defined antigen(s). Earlier studies have demonstrated that the β4 subunit of integrin heterodimer α₆β4 is the target molecule recognized by OCP and MMP sera. ^3–5 Integrins are a family of cell surface receptors involved in cell adhesion to the extracellular matrix, signal transduction, ^6–8 and hemidesmosomal assembly. ^9,10 The hemidesmosomes provide adhesion of the basal keratinocytes to the basement membranes. ^7,11  

The purpose of this study was to identify epitopes within the β4 integrin to which antibodies from patients with OCP, MMP, and both OCP and MMP bind. 

Materials and Methods

Sera

The sera of 10 patients with active MMP without ocular involvement, 7 patients with active OCP, and 8 patients who had both MMP and OCP were collected at the Center for Blistering Diseases, Boston, Massachusetts, and frozen at −80°C until used. 

The established clinical diagnoses of MMP and OCP were confirmed via the histology and immunopathology of the perilesional tissue. In patients with OCP, positive confirmatory direct immunofluorescence (DIF) was present on the conjunctival biopsies of all patients. In patients with MMP or both MMP and OCP, DIF of the oral mucosa was present in all patients. Patients lacking DIF studies were not included. On the salt split skin (SSS), all the test sera bound to the epidermal side of the split. In none of the patients in this study did the sera bind to the dermal side of the SSS. Control sera were obtained from 10 healthy individuals, 10 patients with bullous pemphigoid (BP) and mucosal pemphigus vulgaris (PV), 5 patients with other forms of cicatricial conjunctivitis, and 6 patients with noncicatricial conjunctivitis. Serologic confirmation of MMP and OCP was performed via the presence of antibodies to the integrin β4 subunit by using Western blot analysis. ⁴ None of the sera had antibodies to human α6 integrin, as described earlier. ¹⁸ Blood samples were collected after informed consent. The study was approved by the Institutional Review Board and the Institutional Animal Care and Use Committee. Our research adhered to the tenets of the Declaration of Helsinki and was consistent with the guidelines set forth in the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research. 

Antibodies

The services of GenScript (Piscataway, NJ) were used to produce polyclonal antibodies in rabbits. New Zealand rabbits were immunized subcutaneously with 100 μg purified fragments of integrin β4 and synthesized peptides. Pre- and postimmunization sera were collected. 

The mouse anti-human mAb (UMA 9; Ancell Corporation, Bayport, MN), an mAb to β4 integrin, mouse anti-human (GB3) against laminin 5 (Harlan Bioproducts, Indianapolis, IN), and mouse mAb to His-Tag (EMD Millipore, Biosciences Division, Billerica, MA) were purchased. 

Analysis of Antigenic Determinants

The amino acid sequence of the IC3.0 fragment of integrin β4 was analyzed for antigenicity, flexibility, and β turn with PC/GENE software (IntelliGenetics, Mountain View, CA). Peak values were assigned for each criterion. Overlapping regions of the 3 criteria and peak values greater than 1.0 were selected. 

Cloning of Fragments Representing the Extracellular and Intracellular Domains of the Integrin β4 Molecule

The fragments IC2.0 (1075 aa–1488 aa), IC3.0 (1489 aa–1822 aa), IC3.3 (1489 aa–1803 aa), IC3.4 (1489 aa–1572 aa), and IC3.6 (1573 aa–1822 aa) were cloned and expressed as described earlier, ⁵ with some modifications to the purification process. ¹²  

Peptide Synthesis

To identify the epitopes for pathogenic antibodies for MMP, OCP, and both OCP and MMP, short peptides were synthesized, and HPLC was purified to 96.5% purity at GL Biochem, Ltd. (Shanghai, China). Based on PC/GENE analysis (IntelliGenetics), a sequence of a 14-aa residue peptide IC3.4.1 (range, 1498 aa–1510 aa) and a 14-aa residue peptide IC 3.6.1 (range, 1689 aa–1702 aa) were synthesized. A 14-aa residue peptide IC 2.1 (range, 1475 aa–1489 aa) within the fragment IC2.0 was synthesized to serve as a negative control. The peptides were conjugated with bovine serum albumin (BSA) for Western blot analysis. Polyclonal antibodies were raised by using conjugated peptides in rabbits. A schematic representation of the cloned and synthesized peptides is presented in Figure 1. 

Figure 1

 

Schematic representation of the cloned fragments of the human integrin β4 subunit. The fragments IC3.0 (1489 aa–1702 aa) and IC3.3 (1489 aa–1803 aa) represent the intracellular domain. The fragments IC3.1 (1489 aa–1654 aa) and IC3.2 (1655 aa–1822 aa) represent the subfragments of IC3.0. The fragments IC3.4 (1489 aa–1572 aa), IC3.5 (1573 aa–1654 aa), and IC3.6 (1573 aa–1822 aa) represent the subfragments of fragment IC 3.0. The fragments IC3.4.1 (1489 aa–1510 aa) and IC3.6.1 (1689 aa–1702 aa) are synthetic peptides within fragments IC3.4 and IC3.6, respectively.

Characterization of Integrin β4 Fragments and Peptides

The intracellular fragments IC2.0, IC3.0, IC3.3, IC3.4, and IC3.6 of integrin β4 were characterized for their size and homogeneity via SDS-PAGE and Western blot by using polyclonal antibodies raised in rabbits, as described above. ⁵  

Identification of Epitopes in β4 Integrin That Bind to Pathogenic Auto-Antibodies in the Sera of Patients With OCP, MMP, and Both OCP and MMP Studied by Western Blot

Integrin β4 fragments IC2.0, IC3.0, IC3.3, IC3.4, and IC3.6 and BSA-conjugated peptides IC3.4.1, IC3.6.1, and IC2.1 were analyzed via SDS-PAGE on slab gel. Sera from patients with OCP (n = 7), MMP (n = 10), and both MMP and OCP (n = 8) were studied, as described above. ¹²  

Identification of Epitopes in β4 Integrin That Bind to Pathogenic Auto-Antibodies in the Sera of Patients With OCP, MMP, and Both OCP and MMP via Immunoprecipitation

Integrin β4 fragments IC2.0, IC3.0, IC3.3, IC3.4, and IC3.6 and BSA-conjugated peptides IC3.4.1, IC3.6.1, and IC2.1 were diluted to 1 μg/mL, as described above. ¹²  

Binding Specificity and Cross-Reactivity of OCP, MMP, and Both OCP and MMP Sera to Integrin β4 Fragments

The binding specificity of OCP, MMP, and both OCP and MMP sera to integrin β4 fragments was studied by immobilizing the integrin β4 fragments and unconjugated peptides to cyanogen bromide (CNBr)-activated Sepharose 4B, as described elsewhere. ¹²  

The OCP, MMP, and both MMP and OCP sera equilibrated in the binding solution were each passed through columns of integrin β4 fragments IC2.0, IC3.0, IC3.3, IC3.4, and IC3.6, and unconjugated peptides IC2.1.1, IC3.4.1, and IC3.6.1. Column chromatography and antibody elution are described elsewhere. ¹²  

Binding of Antibodies to Cloned Fragments and Synthetic Peptides of Integrin β4 to Normal Human Conjunctiva, Oral Mucosa, and the Skin via Indirect Immunofluorescence

Four-micrometer-thick sections of normal human conjunctiva, oral mucosa, and skin were incubated with OCP, MMP, PV, and NHS and immuno-affinity–purified antibodies against IC3.0, IC3.4, IC3.6, IC3.4.1, IC3.6.1, IC2.0, and IC2.1 by using indirect immunofluorescence. ^1,12,13  

Confocal Microscopy

1M NaCl-produced SSS was incubated with OCP, MMP, and both OCP and MMP sera; UMA9; and rabbit antibodies to fragments IC2.0, IC3.0, IC3.4, and IC3.6 and peptides IC3.4.1, IC3.6.1, and IC2.1 of integrin β4. The controls were BP sera, epidermolysis bullosa acquisita sera, and NHS. Confocal microscopy was performed, as described elsewhere. ¹⁴  

Passive Transfer Studies in Mice

A passive transfer model to create blisters in mice was used, as described elsewhere. ¹⁵  

Neonatal BALB/c and littermate control mice, which were 24 to 36 hours old with a body weight of 1.3 to 1.8 g, were injected intraperitoneally or subcutaneously with rabbit antibodies to IC2.0, IC3.0, IC3.4, IC3.6, IC3.4.1, IC3.6.1, and IC2.1 and with IgG fractions of OCP and MMP sera. Controls included IgG from high-titer pemphigus antibody and NHS. The volume injected for any mouse was 100 μL, with a maximal IgG concentration of 100 mg/mL. The mice were examined daily for the presence of erythema, blisters, or erosions on the skin and mucosae. Each test serum was injected in 3 mice and observed from 1 to 72 hours. 

Results

Expression and Purification of IC2.0, IC3.0, IC3.3, IC3.4, and IC3.6 Fragments of Integrin β4

The purified protein fragments were analyzed via SDS-PAGE for their homogeneity and size. For a Western blot, the membrane was probed with mAb to His-Tag. The protein fragments were of the following sizes: IC2.0 ∼40 kDa, IC3.0 ∼40 kDa, IC3.3 ∼20 kDa, IC3.4 ∼15 kDa, and IC3.6 ∼38 kDa (data not shown). 

Identification of the Epitope for Antibodies in the Sera of Patients With OCP, MMP, and Both OCP and MMP to Cloned and Synthesized Fragments via Western Blot Analysis

These experiments were performed to test the binding of OCP (n = 5), MMP (n = 7), and both MMP and OCP (n = 8) sera with various fragments of integrin β4. The sera of patients with PV (n = 5), BP (n = 5), non-OCP cicatrizing conjunctivitis (n = 5), noncicatrizing conjunctivitis (n = 5), and normal human serum (n = 5) did not bind to fragment IC3.0 (Fig. 2). 

Figure 2

 

Binding specificity of OCP, MMP, normal human serum (NHS), cicatrizing conjunctivitis (non-OCP), noncicatrizing conjunctivitis (allergic conjunctivitis), and PV sera with fragment IC3.0 of integrin β4. Fragment IC3.0 of integrin β4 was run on 4% to 20% SDS-PAGE, transferred to a nitrocellulose membrane, and reacted with OCP, MMP, NHS, cicatrizing conjunctivitis (non-OCP toxic epidermal necrolysis), noncicatrizing conjunctivitis (allergic conjunctivitis), and PV sera. Binding was observed with only OCP and MMP sera.

The binding pattern of OCP, MMP, and both OCP and MMP sera to fragments IC2.0, IC3.0, IC3.3, IC3.4, and IC3.6 is shown in Table 1. Five of 7 OCP sera, 7 of 10 MMP sera, and 6 of 8 MMP and OCP sera demonstrated binding to fragment IC3.0 and IC3.3. None of the OCP, MMP, and both OCP and MMP sera showed binding with fragments IC2.0 and IC2.1. The OCP sera showed binding with fragment IC3.4 and IC3.4.1, but not with IC3.6 and IC3.6.1. Mucous membrane pemphigoid sera showed binding with IC3.6 and IC3.6.1, but not with IC3.4 and IC3.4.1. The OCP and MMP sera showed binding with IC3.4, IC3.4.1, IC3.6, and IC3.6.1. 

Table 1

 

Immunoblot Identification of the Fragments to Which Antibodies From OCP, MMP, and OCP + MMP Sera of Patients Bind to Cloned and Synthetic Fragment of Human β4 Integrin Subunit

Sera Tested	IC 2.0 40 kDa	IC 2.1 67 kDa	IC 3.0 40 kDa	IC 3.3 20 kDa	IC 3.4 10 kDa	IC 3.4.1 67 kDa	IC 3.6 40 kDa	IC 3.6.1 67 kDa
OCP,* N = 5	Neg†	Neg	+‡	+	+	+	Neg	Neg
MMP,§ N = 7	Neg	Neg	+	+	Neg	Neg	+	+
MMP + OCP,‖ N = 8	Neg	Neg	+	+	+	+	+	+

*   OCP sera of 5 patients show specific binding to IC 3.4 and IC 3.4.1.

†   Neg indicates no binding.

‡   + indicates positive binding. The molecular weight of the observed band is provided below the identification of the peptide.

§   MMP (no ocular involvement) sera of 7 patients show specific binding to IC 3.6 and IC 3.6.1.

‖   MMP and OCP sera from 8 patients show specific binding to IC 3.4 and IC 3.4.1 as well as IC 3.6 and IC 3.6.1.

Identification of the Epitope for Antibodies in the Sera of Patients With OCP, MMP, and Both MMP and OCP via Immunoprecipitation

The OCP sera (n = 5) immunoprecipitated fragments IC3.0, IC3.3, and IC3.4 and peptide IC3.4.1 from the solution (Table 2). The MMP sera (n = 7) immunoprecipitated fragments IC3.0, IC3.3, and IC3.6 and peptide IC3.6.1 from the solution (Table 2). The OCP and MMP (n = 6) sera immunoprecipitated fragments IC3.0, IC3.3, IC3.4, and IC3.6 and peptides IC3.4.1 and IC3.6.1 (Fig. 3C). None of the OCP, MMP, and both OCP and MMP sera immunoprecipitated fragment IC2.0 or peptide IC2.1. Positive OCP, MMP, OCP and MMP sera bound to the IC3.0 and IC3.3 fragments. In the patients in whom the binding of the sera to the peptide could not be observed, one potential reason for this could be the use of immunosuppressive therapies and the ability to reduce auto-antibody titers. 

Figure 3

 

Direct immunofluorescence study using normal human conjunctiva, normal human oral mucosa, and normal human skin as substrates. The test antibodies included sera from patients with only OCP, only MMP, PV, and normal human serum. Rabbit antibodies to the following fragments were tested on all 3 human substrates: These included antibodies to IC2.0, IC3.0, IC3.3, IC3.4, IC3.4.1, IC3.6, and IC3.6.1. (A) OCP sera are used on conjunctiva. (B) MMP sera are used on oral mucosa. (C) Normal human skin stained with normal human sera. (D) Normal human skin stained with PV sera. In (A, B), binding to the BMZ is seen. In (D), intracellular staining of the epidermis is seen, which is typical of PV. (E–G) Staining with antibodies to IC2.0. (H–J) Staining with antibodies to IC3.0. (K–M) Staining with antibodies to IC3.3. (N–P) Staining with antibodies to IC3.4. (Q–S) Staining with antibodies to IC3.4.1. (T–V) Staining with antibodies to IC3.6. (W–Y) Staining with antibodies to IC3.6.1. Note that in (E–Y), smooth linear binding to the BMZ is observed.

Table 2

 

Immunoprecipitation Procedure to Identify Peptides to Which Sera From Patients With OCP, MMP, and MMP + OCP Bind in Cloned and Synthetic Peptides of Human β4 Integrin Subunit

Sera Tested	IC 2.0	IC 2.1	IC 3.0	IC 3.3	IC 3.4	IC 3.4.1	IC 3.6	IC 3.6.1
	*40 kDa	67 kDa	40 kDa	20 kDa	10 kDa	67 kDa	40 kDa	67 kDa
OCP,† N = 5	Neg‡	Neg	+	+	+	+	Neg	Neg
MMP,§ N = 7	Neg	Neg	+	+	Neg	Neg	+	+
MMP + OCP,‖ N = 8	Neg	Neg	+	+	+	+	+	+

*   Molecular weight of band of the sera binding to the peptide and indicated binding observed.

†   Sera of 5 patients with OCP bind specifically to IC 3.4 and to IC 3.4.1 peptide.

‡   Indicates no binding observed.

§   Sera of 7 patients with MMP specifically bind to IC 3.6 and IC 3.6.1.

‖   Sera of 8 patients with MMP who also showed OCP show specific binding to IC 3.0, IC 3.4.1, IC 3.6, and IC 3.6.1.

Binding Specificity of OCP, MMP, and Both OCP and MMP Sera

The binding specificity was tested via column-binding experiments in which pathogenic auto-antibodies, using their complimentary epitopes, were immobilized on CNBr-activated Sepharose 4B columns and absorbed from patients' sera. Then, these antibodies were eluted from the columns and used to probe their respective antigens. 

This analysis showed that columns bound with fragments IC3.4 and IC3.4.1 and depleted the OCP sera of pathogenic antibodies. Thus, these eluates did not show binding with IC3.4 and IC3.4.1 (Fig. 4A). However, when the antibodies were eluted, they bound to IC3.4 and IC3.4.1. 

Figure 4

 

In vivo passive transfer experiment. In this experiment, the ability of antibodies (IgG) from patients with OCP, MMP, and both OCP and MMP and antibodies to cloned fragments IC3.0, IC3.3, IC3.4, IC3.6, IC3.4.1, and IC3.6.1 to produce subepidermal vesicles in mice was tested. Negative controls included IgG from normal human serum and antibodies to cloned fragment IC2.0. The positive control was IgG from a patient with histologically and serologically proven PV. Figure 7 is representative of only 1 experiment in which a neonatal BALB/c mouse was injected intraperitoneally with purified IgG from patients with MMP only. The control for the experiment was IgG from normal human serum. (A) The back of the mouse injected with IgG from the sera of an MMP-only patient, showing multiple vesicles that are intact on the back, neck, and extremities of the mouse's skin. Some of the vesicles seen during a visual examination have been marked with arrows. (B) The back of a mouse injected with IgG from normal human serum.

Columns immobilized with fragments IC3.6 and IC3.6.1 depleted pathogenic antibodies from MMP sera. The eluates from these columns did not bind with IC3.6 and IC3.6.1 on Western blot. When the antibodies were eluted from this column, binding occurred with fragments IC3.6 and IC3.6.1 on the Western blot (Fig. 4B). 

Figure 5

 

Binding specificity of OCP and MMP auto-antibodies to the epitopes in human β4 integrin subunit. (A) CNBr-activated Sepharose 4B column bound with all study peptides. Sera from OCP and MMP patients passed through the columns. No binding is observed since the antibodies bind to their specific epitopes. (B) OCP and MMP antibodies are eluted from these columns. The eluted sera of OCP and MMP are bound to their respective epitope. (C) CNBr-activated Sepharose 4B columns were bound with peptides IC 3.6 and IC 3.6.1. Sera from an OCP patient passed through the column and bound to IC 3.4 or IC 3.4.1 (left). Similarly, column with peptides IC 3.4 and IC 3.4.1, sera from an MMP patient bound to peptides IC 3.6 and IC 3.6.1 (right).

Sera with both OCP and MMP auto-antibody eluates from the IC3.4 and IC3.4.1 columns did not show binding with the IC3.4 and IC3.4.1 fragments, but showed binding with fragments IC3.6 and IC3.6.1 on the Western blot. Similarly, the OCP and MMP sera passed through the IC3.6 and IC3.6.1 columns and did not bind with the IC3.6 and IC3.6.1 fragments, but showed binding with the IC3.4 and IC3.4.1 fragments on the Western blots (data not shown). 

Cross-absorption experiments demonstrated no cross-reactivity. The eluates showed binding to IC3.4 and IC3.4.1 on a Western blot (Fig. 5). When MMP sera passed through columns containing immobilized fragments IC3.4 and IC3.4.1, the eluates bound to membranes containing fragments IC3.6 and IC3.6.1 (Fig. 5). Ocular cicatricial pemphigoid and MMP sera passed through the column immobilized with fragments IC3.6 and showed binding with fragments IC3.0, IC3.4, and IC3.4.1 (Fig. 5). Ocular cicatricial pemphigoid and MMP sera passed through the column immobilized with fragment IC3.4 and showed binding with fragments IC3.0, IC3.6, and IC3.6.1 (Fig. 5). Thus, there was no cross-reactivity in the antigen-binding pattern between OCP and MMP sera. 

Direct Immunofluorescence Using Normal Human Conjunctiva, Oral Mucosa, and Skin

Sera from a patient with only OCP and only MMP rabbit antibodies to IC2.0, IC3.0, IC3.3, IC3.4, IC3.4.1, IC3.6, and IC3.6.1 bound to the BMZ in a linear, smooth, homogeneous pattern (Fig. 3). 

Passive Transfer Animal Studies

Neonatal BALB/c mice injected subcutaneously or intraperitoneally with test reagent developed blisters on their abdomens, backs, necks, and faces. The number and size of the blisters varied and were often accompanied by erythema. The blisters occurred at sites that were remote from the injection site. Denudation of the skin was not observed. Blisters were observed in the mice injected with immuno-affinity–purified OCP, MMP, and both OCP and MMP sera and rabbit antibodies to IC3.0, IC3.4, IC3.4.1, IC3.6, and IC3.6.1. Blisters were observed on the skin of the mice with the various test reagents (Fig. 4). Mice injected with the IgG fraction from the sera of a mucosal PV patient demonstrated blisters and erosions on the skin of the back, neck, face, and abdomen (data not shown). 

The mice injected with rabbit antibodies to cloned fragment IC2.0 and IC2.1 and NHS did not develop blisters. No blisters were observed in the conjunctiva or oral cavities of any mouse (data not shown). 

The amino acid sequences of IC3.4.1 and IC3.6.1 peptides in humans (National Center for Biotechnology Information [NCBI] reference sequence: NP_000204.3), mice (NCBI reference sequence: EDL34540.1) and rabbits (NCBI reference sequence: XP_002722995.1) are presented in Table 3. These sequences demonstrate that the rabbits would be capable of producing antibodies to injected human peptides. The rabbit's anti-human antibodies would bind to mouse skin and cause blisters because of the significant homology between mouse and human, therefore, integrin. 

Table 3

 

Amino Acid Sequence Analysis of Epitopes for Binding of OCP and MMP Sera in β4 Integrins in Human, Mice, and Rabbit

IC 3.4.1
Human	NP000204	1489 TRDYNSLTRS EESHSTILPRDY 1510
Mice	XP002722995	1432 TRDYHSLTRT EESHSAILPRDY 1454
Rabbit	NP001005608	1425 TRDYHSLTRT EESHSGILPRDY 1448
IC 3.6.1
Human	NP000204.3	1689 AFRVDGDSPESRLT 1702
Mice	XP002722995.1	1624 FQVDGDNPESRLT 1636
Rabbit	NP001005608.2	1670 FRVDGDNPESRLT 1682

Routine Histology (H&E) of Mice Blisters

The blisters that developed in the skin of the mice were subjected to hematoxylin and eosin (H&E) staining. The antibodies to fragment IC3.4 (Fig. 6A), IC3.4.1 (Fig. 6B), IC3.6 (Fig. 6C), and IC3.6.1 (Fig. 6D) showed subepidermal separation in the mouse skin. The mouse skin injected with NHS remained intact (Figs. 6E, 6F), while PV sera demonstrated intra-epidermal vesicle formation (Figs. 6G, 6H). 

Figure 6

 

Routine histology (H&E) of the skin of mice injected with various test sera in the in vivo passive transfer experiment. (A–D) Subepidermal separation of the mouse skin. (A) A mouse was injected with rabbit antibody to IC3.4; (B), to IC3.4.1; (C), to IC3.6, and (D), to IC3.6.1. (E, F) Results are representative of mice injected with IgG from normal human serum; no epidermal separation is seen. (G, H) Mice were injected with sera from patients with PV. Note the intra-epidermal vesicle formation with an intact basal cell layer.

Confocal Microscopy

The confocal microscopy of the SSS showed that the antibodies to fragments of integrin β4 IC3.0 (Fig. 7A) and IC3.4 (Fig. 7B) bound to the BMZ. The antibody to the fragments IC3.6 and IC3.4 bound to the roof of the blister (Figs. 7C, 7D, 7F). The UMA9 mAb to laminin 5 showed binding to the base (Fig. 7D). 

Discussion

Mucous membrane pemphigoid is a clinically heterogeneous disease with diverse clinical manifestations. ¹⁶ It is divided into subsets by clinical presentation and antigen specificity. ¹⁷ The version of the disease that is localized to the oral cavity is known as oral pemphigoid. ¹⁸ The target antigen is integrin α6. ¹⁹ When limited to the eye or when conjunctiva is the main site of involvement, the subset is known as ocular cicatricial pemphigoid. ²⁰ The target antigen is β4 integrin. ⁴ Mucous membrane pemphigoid involving all other mucosal surfaces is known as mucous membrane pemphigoid ²¹ ; the target antigen is a β4 integrin subunit. ²² A subset of MMP that is clinically indistinguishable from others has auto-antibodies against laminin 5, currently known as laminin 332. ²³ The unique feature of this subset is its high association with cancer. ^24,25 Recently, another subset, formerly known as anti-p200 pemphigoid and now known as antilaminin pemphigoid, has been described. ²⁶ These subsets are not included in this study. 

Figure 7

 

Confocal microscopy using SSS. (A) Antibody to IC3.0 is used and binds to the BMZ. (B) Antibody to IC3.4 is used and binds to the BMZ. (C) Antibody to IC3.6 is used and binds to the roof of the blister. (D) Antibodies to laminin 5 are used and bind to the base of the blister. (E) Antibodies to IC3.4.1 are used and bind to the roof of the blister. (F) Antibodies to IC3.6.1 are used and bind to the roof of the blister.

A 3-year or longer follow-up has demonstrated that the antigen-binding specificity of the sera remains unchanged. ¹⁷ Mucous membrane pemphigoid, OCP, and both OCP and MMP sera continue to bind to β4 integrin. Cross-reactivity between antibodies to integrin α6 and β4 has not been observed. ^17,19 Furthermore, auto-antibody titers to integrin β4 are high during active disease and decline with improvement. ²⁷ Similarly, a correlation with disease activity has been reported with auto-antibodies to α6 integrin in oral pemphigoid (OP). ²⁸ Interestingly, the major histocompatibility complex class II allele (HLA DQβ1*0301) is associated with all 3 subsets of MMP. ²⁹ Recent computer modeling studies demonstrate that this allele has binding sites for both β4 and α₆ integrin. ³⁰  

Initially, a clone from a cDNA keratinocyte library bound to OCP and MMP sera. ³ The amino acid sequence of the peptide in the identified clone shows 100% homology with the intracytoplasmic domain of human β4 integrin. ³ Several experiments using various tumor cell lines transfected with entire cytoplasmic and extracellular domains of β4 integrin fragments, which indicates that the binding of OCP and MMP sera was limited to the intracytoplasmic domain. ^31,32  

Cloned fragments of the intracellular portion showed that OCP and MMP sera bind to IC3.0, the innermost fragment. ^3,5  

Fragment IC3.0 was subfragmented by cloning and labeled as IC3.3, IC3.4, IC3.5, and IC3.6. Pemphigoid sera bound to IC3.3 (1489–1803) and to an 83–amino acid peptide labeled as IC3.4 within it. ⁵ Interestingly, the initial peptide identified from the keratinocyte library was IC3.6, to which sera from only OCP patients did not bind. ⁵ This observation suggested that the auto-antibody binding site for MMP may be in the IC3.6 peptide. In vitro organ culture experiments using normal human conjunctiva and oral mucosa cultured with OCP, MMP sera, and rabbit antibodies to cloned fragments have produced BMZ separation. ^13,33  

In this study, we have attempted to further identify the epitopes within the integrin β4 subunit to which auto-antibodies with OCP and MMP bind. IC3.0 was subfragmented into IC3.1, IC3.2, IC3.3, IC3.4, and IC3.6, and the peptides IC3.6 and IC3.4.1 were synthesized. Indirect immunofluorescence studies using normal human conjunctiva, oral mucosa, and skin showed the binding of antibodies to these peptides to the BMZ in all 3 tissues (Fig. 6). 

Immunoblot and immunoprecipitation experiments demonstrated that antibodies from patients with OCP bind to IC3.0, IC3.3, IC3.4, and IC3.4.1 (Figs. 3A–C; Figs. 5A–C). Binding specificity studies and cross-absorption studies showed that OCP sera bound only to IC3.4.1 and not to IC3.6.1. Sera from patients with MMP bound only to IC3.0, IC3.3, IC3.6, and IC3.6.1 and did not bind to IC3.4.1 (Figs. 4A–C). Sera from patients who had ocular and multiple mucosal involvement bound to IC3.0, IC3.3, IC3.4, IC3.4.1, IC3.6, and IC3.6.1 (Figs. 3C, 5C). The ability of sera to bind to multiple epitopes within the same molecule occurs owing to the phenomenon of epitope spreading. ³⁴  

In the current study, organ culture studies were not done, because in several earlier studies, OCP and MMP sera and rabbit antibodies to cloned fragments of integrin β4 subunit have produced BMZ separation in oral and conjunctival mucosa and human skin. ^4,12,13 Instead, experiments using passive transfer models involving mice were used. 

Immuno-affinity–purified IgG from sera of patients with OCP, MMP, both MMP and OCP, and rabbit antibodies to IC3.0, IC3.4, IC3.4.1, IC3.6, and IC3.6.1 produced blisters on the skin when injected in mice. Upon routine histology, these blisters showed subepidermal separation. This strongly suggests that the antibody to fragments IC3.4 and IC3.6 and peptides IC3.4.1 and IC3.6.1 have the capacity to produce subepidermal blisters in mouse skin that are similar to those produced by injecting immuno-affinity–purified sera from patients with OCP and MMP. The fact that these test reagents did not produce epithelial separation in the conjunctivas or the oral cavities of the mice may in part be due to the concentrations of the injected antibodies and the short duration of postinjection observation. This was intentionally done so that intact vesicles could be observed during physical examination and routine histology adequately performed. This may also explain why sera from mucosal PV produced only skin lesions and not mucosal lesions in the BALB/c mice. 

The binding of antibodies raised in rabbit to normal human tissues on direct immunofluorescence and in the causation of blister in neonatal mice would indicate their specificity and cross-reactivity. 

The binding site for the sera of patients with oral pemphigoid to the extracellular domain of the α6 integrin subunit has been reported. ¹² These studies confirmed that the epitope for the binding of the auto-antibody in the patients with OP is a 14–amino acid peptide. 

The authors recognize that a major limitation of this study is that a small number of patients were studied and that the in vivo animal experiments will require a more extensive and detailed investigation. The authors' present intent was only to demonstrate pathogenicity (BMZ separation) and not the multiple steps involved in blister formation. It is very important to emphasize that the sera used in this study identified IC3.4.1 as the putative OCP antigen and IC3.6.1 as the putative MMP antigen. However, it needs to be stressed that there may be other relevant and important epitopes in the integrin β4 subunit. By reproducing these experiments in a larger cohort of patients, a major hurdle in early diagnosis and the initiation of therapy may be resolved. Such studies have the potential to improve the prognosis of MMP and OCP. 

Acknowledgments

The authors alone are responsible for the content and writing of the paper. 

Disclosure: K.A. Rashid, None; C.S. Foster, None; A.R. Ahmed, None