This study was undertaken to characterize the antigenic determinants recognized by the autoantibodies of patients with ocular cicatricial pemphigoid (OCP). OCP is a subepithelial, blistering, autoimmune disease that mainly affects the conjunctiva and other mucous membranes. We previously demonstrated that a cDNA clone, isolated from a keratinocyte expression library by using immunoaffinity-purified OCP autoantibody, encoded the cytoplasmic domain of β4 integrin subunit. Our subsequent studies showed that sera from all the OCP patients that were tested recognize the human β4 integrin subunit. To identify the prevalent epitopes of the anti-β4 autoantibodies of OCP, we have used cell lines transfected with vectors encoding a wild-type β4 subunit, a tailless β4 subunit, or a β4 subunit lacking the extracellular domain. Nontransfected cell lines were used as controls. Lysates from these cell lines were analyzed with OCP sera, IgG fractions from OCP sera, and immunoaffinity-purified OCP autoantibodies. Abs to extracellular and cytoplasmic domains of human β4 integrin were used as positive controls, whereas normal human sera and normal human IgG fractions were used as negative controls. The reactivity of OCP Abs was determined by using immunoblotting, immunoprecipitation, and FACS analysis. The results of this study indicate that OCP sera, OCP IgG fractions, and immunoaffinity-purified OCP autoantibodies react with the intracellular and not the extracellular domain of human β4 integrin subunit. In vitro cell culture experiments demonstrated that OCP autoantibody binds to the cytoplasm of the cells. The relevance of these findings to the pathogenesis of OCP is discussed.

Ocular cicatricial pemphigoid (OCP)4 is an autoimmune vesiculobullous disease that affects the conjunctiva, and less frequently, other squamous epithelia (1). If untreated or inappropriately treated, OCP can result in blindness. OCP is a form of mucous membrane pemphigoid or cicatricial pemphigoid (CP), a heterogeneous disease with a wide spectrum of clinical phenotypes, depending upon the predominant area affected and the number of loci involved (mouth, eye, skin, genitals, nose, oropharynx, esophagus, and larynx). OCP can also have a diverse array of clinical courses and a spectrum of responses to systemic and topical therapy.

Some investigators have reported that like bullous pemphigoid (BP), mucous membrane pemphigoid or CP patients’ sera recognize 180-kDa proteins termed as BPAg2 when human epidermal lysate and BPAg2 fusion proteins were used as substrates (2, 3). This protein along with BPAg1 (BP230) and α6β4 integrin are found in hemidesmosomes, which are structures that link the underlying basement membranes of stratified as well as other epithelia to the intermediate filament system. They are composed of an outer plaque, containing proteins such as BP180 and α6β4, and an inner plaque containing BP230 and HD1/plectin, which then associate with the cytoskeleton. Defects in these proteins can lead to epithelial blisters in both humans and mouse models (4)

We have observed that after preabsorption of epidermal or conjunctival or tumor cell lysates with BP sera, OCP sera recognize a 205-kDa protein (5, 6), identified as the β4 integrin subunit (7). These findings raised the possibility that a putative target Ag for OCP may be the β4 integrin subunit (7). In fact, we have recently reported that OCP sera, immunoaffinity-purified OCP Abs eluted from the 205-kDa band on nitrocellulose, and anti-β4 Abs cause blistering in an in vitro human conjunctival model (8). This result suggests that Abs from OCP sera against β4 could be pathogenic in vivo.

A subset of CP patients has been reported to be characterized by the presence of autoantibodies against epiligrin, which is now identified as the α3 subunit of laminin-5 (9, 10). Interestingly, many of the OCP sera used in the present study and in our earlier reported studies (5, 6) did not bind to epiligrin (K. Yancey, unpublished observation).

Since CP is a heterogeneous disease, it is possible that clinical subsets may correlate with a variety of anti-BMZ autoantibodies with different specificities that recognize different target molecules within the complex of cell surface proteins and extracellular matrix proteins in the BMZ.

Given the possible pathogenic role of anti-β4 Abs in OCP sera, the purpose of this study was to determine the region of β4 recognized by these Abs. Surprisingly, we have found that the anti-β4 Abs of OCP bind to the intracellular region of human β4 integrin and we were unable to detect, using immunoblotting, immunoprecipitation, and FACS analysis, any binding to the extracellular domain of β4 integrin subunit.

Sera were obtained from 15 patients with active OCP in the acute phase of the disease before the institution of therapy. The diagnosis of OCP was confirmed in each patient by clinical presentation, routine histology, and immunopathological analysis of biopsied conjunctiva. When tested on salt split skin, the anti-BMZ autoantibody bound to the epidermal side of the split.

Total IgG of OCP autoantibodies was prepared from 10 active OCP patients by EZ-SEP kit obtained from Middlesex Sciences (Foxborough, MA). Immunoaffinity-purified OCP autoantibodies were eluted from the nitrocellulose blots, as described previously (11). In another Western blot, it was confirmed that the immunoaffinity-purified OCP autoantibodies bound to a 205-kDa protein in normal human conjunctiva and epidermis in exactly the same manner as OCP sera. mAb to human β4 (3E1) and polyclonal Ab to extracellular and cytoplasmic domains of human β4 integrin were described elsewhere (12, 13). The 9E10 anti-myc mAb was prepared by the hybridoma core facility at Memorial Sloan-Kettering Cancer Center (New York, NY). Control sera were obtained from 10 healthy individuals and from 1 patient each with active disease with pemphigus vulgaris (PV), BP, oral pemphigoid (OP), and dermatitis herpetiformis (DH). All the control and patient sera were stored at −80°C until used. The study has been reviewed by our Institutional Review Board. Sera were obtained after appropriate informed consent procedures.

The epidermis of the human skin was separated from underlying dermis, and clear lysates of Ags were prepared as reported earlier (14). Briefly, the human epidermis and conjunctiva were homogenized on ice with 1.5% SDS, 62.5 mM Tris-HCl buffer, pH 6.8, supplemented with 5% mercaptoethanol, 2 mM PMSF, and 10 μg/ml of pepstatin A, antipain, leupeptin, and chymostatin (Sigma, St. Louis, MO). The lysates were then boiled for 5 min, vortexed, and centrifuged at 15,000 × g for 30 min. The clear supernatant of human epidermal lysate and human conjunctival lysate was centrifuged again at high speed in a microfuge centrifuge for 30 min, and supernatant was concentrated. Protein content was measured at 280 nm, and aliquots were stored at −80°C until used.

The 804G rat bladder carcinoma cells and 293T human embryonic kidney were grown in DMEM containing 10% bovine calf serum. The 804G cells stably expressing recombinant human full-length β4 (clone A), extracellular domain of β4 (clone F), and intracellular domain of β4 integrin (clone L) were described elsewhere (13, 15). The 293T cells, which do not express endogenous β4, were cotransfected by calcium phosphate precipitation method with cDNAs of human α6 and versions of β4 (13, 15, 16) cloned into the EcoRI site of the PRK5 expression vectors. Typically, subconfluent cells were extracted for 30 min on ice in Triton X-100 lysis buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 1 mM EGTA, 4 mM EDTA, 1% Triton X-100, 0.5 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, and 10 μg/ml each pepstatin A, aprotinin, and leupeptin). Lysates were clarified by centrifugation for 10 min at 14,000 rpm in a refrigerated centrifuge.

SDS-PAGE and Western blot were performed as previously described (17, 18). A sensitive Western blotting assay was used in this study, as described previously (7). In brief, blotted nitrocellulose membrane was blocked with 3% skimmed milk. Blotted proteins (nitrocellulose strips) after extensive washing with TBS-containing 0.05% Tween 20 were incubated with diluted test sera or Abs. After 4× wash, nitrocellulose strips were incubated with HRP-conjugated secondary Abs (anti-human, anti-rabbit, and anti-mouse), and then the final step was detected by using the ECL Western blotting kit (Amersham Life Sciences, Arlington Heights, IL), according to the manufacturer’s protocol.

Immunoprecipitation was performed as described earlier (16). Briefly, 50 μl of packed protein G-agarose washed three times with lysis buffer, resuspended in 500 μl of lysis buffer, then 10–50 μl of OCP sera or Abs to β4 integrin was added to it and incubated for 2 h at 4°C. The protein G was washed three times with lysis buffer. The cell lysates (500 μg/500 μl) suspended in lysis buffer were added to the OCP sera or β4 Ab-coated protein G and incubated overnight at 4°C. The immunoprecipitates were washed five times in cold lysis buffer and then resuspended in 50 μl of sample buffer boiled for 5 min and analyzed by immunoblotting.

In these experiments, we used 804G (rat bladder) and 293T fibroblast cell lysates (untransfected and transfected with full-length, extracellular, and intracellular domains of β4 integrin) and immunoprecipitated with OCP autoantibodies.

The 293T cells expressing full-length β4 were detached briefly with 0.02% trypsin and incubated for 20 min in ice with either 10 μg/ml of 3E1 mAb, or diluted OCP Abs, or normal human IgG, followed by FITC-conjugated goat anti-mouse or anti-human secondary Abs. After fixation in 3.7% formaldehyde, cells were analyzed by flow cytometry using a Becton Dickinson FACS machine and Cell Quest Ware.

This study was done according to the methods described by other investigators (19, 20). Briefly, β4-expressing cell lines (MDA-435) were grown on chamber slides (Nunc, Naperville, IL) and allowed to adhere and proliferate. After 48 h, the cells were nearly confluent and washed with RPMI 1640 medium, then incubated with OCP sera, Abs to cytoplasmic, and extracellular domains of β4 integrin and normal human serum for 24 h. After incubation, the slides were washed three times with PBS, then fixed in methanol acetone at −20OC for 15 min. After fixation, the cells were washed in PBS and blocked with 3% BSA in PBS for 30 min. Then the cells were washed and incubated with fluorescein-labeled goat anti-human or anti-rabbit IgG (Sigma) for 45 min, then washed, mounted, and observed by fluorescent microscope.

The 804G cells are derived from a rat bladder carcinoma and are widely used for studies on hemidesmosomes. These cells were transfected with cDNA constructs encoding wild-type β4, a β4 subunit lacking the ectodomain (headless), or one lacking the cytoplasmic domain (tailless). Stable clones expressing comparable surface levels of each recombinant protein were analyzed by immunoblot and immunoprecipitation experiments. Untransfected 804G cells were used as controls.

Extracts from various cell lines were separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with either OCP sera or control Abs against either the cytoplasmic or extracellular domains of β4 (Fig. 1). OCP sera reacted with wild-type (205 kDa) and headless β4 (140 kDa), but did not react with tailless β4 (100 kDa). Immunoblotting with rabbit polyclonal Abs against the cytoplasmic and extracellular domains of β4 confirmed the expression of the transfected constructs. Wild-type β4 (205-kDa band) is observed in all lanes because it is endogenously expressed by 804G cells.

FIGURE 1.

Immunoblot assays using 804G cell lines. Extracts of 804G cells stably transfected with wild-type β4 (lanes 1, 5, and 9), tailless β4 (lanes 2, 6, and 10), headless β4 (lanes 3, 7, and 11), or nontransfected cells (lanes 4, 8, and 12) were separated by SDS-PAGE and transferred to nitrocellulose. Lanes 1–4 were blotted with OCP serum. Note binding to wild-type β4 (205-kDa) and headless β4 (140-kDa) proteins. Lanes 5–8 were blotted with rabbit polyclonal Ab to the cytoplasmic domain of human β4 integrin. Note binding to wild-type (205 kDa) and headless (140 kDa) proteins. Lanes 9–12 were blotted with rabbit polyclonal Ab to the extracellular domain of human β4 integrin. Note binding to wild-type (205-kDa) and tailless β4 (140-kDa) proteins.<.>

FIGURE 1.

Immunoblot assays using 804G cell lines. Extracts of 804G cells stably transfected with wild-type β4 (lanes 1, 5, and 9), tailless β4 (lanes 2, 6, and 10), headless β4 (lanes 3, 7, and 11), or nontransfected cells (lanes 4, 8, and 12) were separated by SDS-PAGE and transferred to nitrocellulose. Lanes 1–4 were blotted with OCP serum. Note binding to wild-type β4 (205-kDa) and headless β4 (140-kDa) proteins. Lanes 5–8 were blotted with rabbit polyclonal Ab to the cytoplasmic domain of human β4 integrin. Note binding to wild-type (205 kDa) and headless (140 kDa) proteins. Lanes 9–12 were blotted with rabbit polyclonal Ab to the extracellular domain of human β4 integrin. Note binding to wild-type (205-kDa) and tailless β4 (140-kDa) proteins.<.>

Close modal

Lysates from the various 804G cell lines were immunoprecipitated with OCP sera. These immunoprecipitates were then subjected to SDS-PAGE, transferred to nitrocellulose membrane, and immunoblotted with polyclonal Ab to extracellular and cytoplasmic domains of β4 integrin (Fig. 2). OCP sera was able to immunoprecipitate the wild-type and headless versions of β4, but did not immunoprecipitate with the tailless β4. Wild-type β4 is observed in all lanes because 804G cells express endogenous β4. Similar results were seen using IgG fractions of OCP sera and immunoaffinity-purified OCP Abs (data not shown). Sera from patients with PV, BP, OP, DH, or normal human serum did not immunoprecipitate with β4 (data not shown). These experiments demonstrate that autoantibodies from OCP sera react predominantly with the cytoplasmic, and not extracellular domain of β4 integrin.

FIGURE 2.

Immunoprecipitation experiment using 804G cell lines. Extracts from 804G cells stably transfected with headless β4 (lanes 1 and 5), tailless β4 (lanes 2 and 6), wild-type β4 (lanes 3 and 7), or nontransfected 804G cells (lanes 4 and 8) were immunoprecipitated with OCP sera. Immunoprecipitates were separated by SDS-PAGE and transferred to nitrocellulose. Lanes 1–4 were blotted with rabbit polyclonal Ab to the cytoplasmic domain of β4 integrin. Note binding to wild-type and headless β4. Lanes 5–8 were blotted with rabbit polyclonal Ab to the extracellular domain of β4 integrin. Note binding to wild-type β4 only.

FIGURE 2.

Immunoprecipitation experiment using 804G cell lines. Extracts from 804G cells stably transfected with headless β4 (lanes 1 and 5), tailless β4 (lanes 2 and 6), wild-type β4 (lanes 3 and 7), or nontransfected 804G cells (lanes 4 and 8) were immunoprecipitated with OCP sera. Immunoprecipitates were separated by SDS-PAGE and transferred to nitrocellulose. Lanes 1–4 were blotted with rabbit polyclonal Ab to the cytoplasmic domain of β4 integrin. Note binding to wild-type and headless β4. Lanes 5–8 were blotted with rabbit polyclonal Ab to the extracellular domain of β4 integrin. Note binding to wild-type β4 only.

Close modal

To avoid the complexity from the reaction of OCP autoantibodies with the endogenous β4 in 804G cells, we decided to confirm our results in 293T cells, which do not express endogenous β4.

Immunoblot experiments.

Human embryonic kidney 293T cells were transiently transfected with cDNA constructs encoding the wild-type β4, the β4 subunit lacking the ectodomain (headless), or the β4 subunit lacking the cytoplasmic domain (tailless). Extracts from these cells were separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with IgG fractions from OCP sera or immunoaffinity-purified OCP autoantibodies, as well as normal human IgG as a negative control or rabbit polyclonal Abs against the β4 cytoplasmic and extracellular domains as positive controls (Fig. 3). Immunoaffinity-purified OCP autoantibodies and IgG fractions of OCP sera recognized wild-type β4 and headless β4, but not tailless β4. Similar results were observed using whole sera from OCP patients (data not shown). Normal human IgG did not react with β4. Expression of wild-type β4 and β4 mutants was confirmed by immunoblotting with rabbit polyclonal Abs to the cytoplasmic or extracellular domains of β4. Lysates from untransfected 293T cells did not show any reactivity with OCP sera or control anti-β4 Abs (data not shown).

FIGURE 3.

Immunoblot assays using transfected 293T cell. Extracts from cells transfected with wild-type β4 (lane 1), headless β4 (lane 2), or tailless β4 (lane 3) were separated by SDS-PAGE and transferred to nitrocellulose. Membranes were blotted with either immunoaffinity-purified OCP autoantibodies (OCP purif), IgG fraction from OCP sera (OCP IgG), IgG fraction from normal human sera (NH IgG), rabbit polyclonal Abs against the cytoplasmic domain of β4 (anti-cyto), or rabbit polyclonal Abs against the extracellular domain of β4 (anti-extra). Note binding of OCP Abs to full-length β4 and headless β4 only. The low molecular mass band in lane 1 is probably a degradation product of β4.<.>

FIGURE 3.

Immunoblot assays using transfected 293T cell. Extracts from cells transfected with wild-type β4 (lane 1), headless β4 (lane 2), or tailless β4 (lane 3) were separated by SDS-PAGE and transferred to nitrocellulose. Membranes were blotted with either immunoaffinity-purified OCP autoantibodies (OCP purif), IgG fraction from OCP sera (OCP IgG), IgG fraction from normal human sera (NH IgG), rabbit polyclonal Abs against the cytoplasmic domain of β4 (anti-cyto), or rabbit polyclonal Abs against the extracellular domain of β4 (anti-extra). Note binding of OCP Abs to full-length β4 and headless β4 only. The low molecular mass band in lane 1 is probably a degradation product of β4.<.>

Close modal

Immunoprecipitation experiments.

Lysates from 293T cell lines untransfected or transfected with wild-type or mutant versions of β4 were immunoprecipitated with OCP sera. These immunoprecipitates were then subjected to SDS-PAGE, transferred to nitrocellulose, and immunoblotted with polyclonal Abs to the extracellular or cytoplasmic domains of human β4 molecule (Fig. 4). The results demonstrate that OCP sera can immunoprecipitate the wild-type and headless β4, but not tailless β4. When the same lysates were immunoprecipitated with sera from PV, BP, OP, DH, or normal human serum, no reactivity with β4 was observed (data not shown).

FIGURE 4.

Immunoprecipitation using 293T cell lines. Extracts of 293T cells transfected with wild-type β4 (lanes 1 and 5), tailless β4 (lanes 2 and 6), headless β4 (lanes 3 and 7), or nontransfected cells (lanes 4 and 8) were immunoprecipitated with OCP sera. Immunoprecipitates were separated by SDS-PAGE and transferred to nitrocellulose. Lanes 1–4 were blotted with rabbit polyclonal Ab to the cytoplasmic domain of β4. Note binding to wild-type β4 and headless β4. Lanes 5–8 were blotted with rabbit polyclonal Ab to the extracellular domain of β4 integrin. Note binding to wild-type β4 only.

FIGURE 4.

Immunoprecipitation using 293T cell lines. Extracts of 293T cells transfected with wild-type β4 (lanes 1 and 5), tailless β4 (lanes 2 and 6), headless β4 (lanes 3 and 7), or nontransfected cells (lanes 4 and 8) were immunoprecipitated with OCP sera. Immunoprecipitates were separated by SDS-PAGE and transferred to nitrocellulose. Lanes 1–4 were blotted with rabbit polyclonal Ab to the cytoplasmic domain of β4. Note binding to wild-type β4 and headless β4. Lanes 5–8 were blotted with rabbit polyclonal Ab to the extracellular domain of β4 integrin. Note binding to wild-type β4 only.

Close modal

Both the immunoblotting and immunoprecipitation experiments, in which OCP autoantibodies fail to react with the tailless β4, suggest that the predominant epitopes recognized by the anti-β4 autoantibody in OCP sera are in the cytoplasmic rather than the extracellular domain of β4 integrin.

Although the OCP autoantibodies do not bind to the β4 extracellular domain by immunoblot or immunoprecipitation analysis, it is possible that a small, but potentially pathogenic pool of OCP autoantibodies can also recognize the β4 extracellular domain in its native conformation at the cell surface. Therefore, we tested the ability of the OCP autoantibodies to bind the extracellular domain of β4 in intact cells by flow cytometry with nonpermeabilized cells. Under these conditions, we could not detect any binding of OCP autoantibodies to the cell surface, although the control Ab 3E1 against β4 extracellular domain bound strongly (Fig. 5).

FIGURE 5.

FACS analysis. To assess the ability of OCP Abs to bind the extracellular portion of β4 in intact cells, 293T cells expressing full-length β4 were incubated with either anti-β4 mAb 3E1(A), immunoaffinity-purified OCP Ab (B), IgG fractions from OCP sera (C), or IgG fractions from normal human sera (D), followed by FITC-conjugated secondary Abs, and analyzed by FACS. The x-axis represents the fluorescence intensity, and the y-axis the number of cells. The dashed line represents staining with secondary Ab alone. Note that only 3E1 (A) stains significantly above background, indicating that OCP Abs do not react with the β4 extracellular domain in intact cells.

FIGURE 5.

FACS analysis. To assess the ability of OCP Abs to bind the extracellular portion of β4 in intact cells, 293T cells expressing full-length β4 were incubated with either anti-β4 mAb 3E1(A), immunoaffinity-purified OCP Ab (B), IgG fractions from OCP sera (C), or IgG fractions from normal human sera (D), followed by FITC-conjugated secondary Abs, and analyzed by FACS. The x-axis represents the fluorescence intensity, and the y-axis the number of cells. The dashed line represents staining with secondary Ab alone. Note that only 3E1 (A) stains significantly above background, indicating that OCP Abs do not react with the β4 extracellular domain in intact cells.

Close modal

When MDA-435 cells, which endogenously produce full-length β4, were incubated with OCP sera, Abs to cytoplasmic and extracellular domains of β4 integrin, and normal human serum for 24 h, the following observations were made. OCP sera bind only to the cytoplasm of the cultured cells (Fig. 6,C). Ab to cytoplasmic domain of β4 integrin binds in a similar pattern to only the cytoplasm of the cultured cells (Fig. 6,B). No extracellular binding was seen in cells cultured with OCP sera or Ab to cytoplasmic domain of β4 integrin. Ab to extracellular domain of β4 integrin demonstrated binding to the cell surface only (Fig. 6,A). No cytoplasmic binding was seen. No binding to either the cell surface or cytoplasm was seen when the cells were cultured with normal human serum (Fig. 6 D).

FIGURE 6.

Direct immunofluorescence of cultured cells (MDA-435) incubated with OCP sera (C), Ab to cytoplasmic domain of β4 (B), Ab to extracellular domain of β4 integrin (A), and normal human serum (D). Binding to cell membrane observed in A (arrows). Binding to cytoplasm noted in B and C (arrows). No binding observed in D. (Magnification, ×60.)

FIGURE 6.

Direct immunofluorescence of cultured cells (MDA-435) incubated with OCP sera (C), Ab to cytoplasmic domain of β4 (B), Ab to extracellular domain of β4 integrin (A), and normal human serum (D). Binding to cell membrane observed in A (arrows). Binding to cytoplasm noted in B and C (arrows). No binding observed in D. (Magnification, ×60.)

Close modal

This study provides evidence that the large majority of epitopes recognized by OCP autoantibodies reside in the cytoplasmic domain of human β4 integrin, thus confirming and extending our previous observations. We have used immunoblotting and immunoprecipitation to analyze 804G cells stably transfected with cDNAs encoding a wild-type, a tailless, or a headless form of β4. Selective binding of OCP sera to the cytoplasmic domain of β4 was observed. However, OCP sera and rabbit polyclonal Abs to the extracellular and cytoplasmic domains of β4 also showed background binding to the endogenous β4 expressed by these cells. We have therefore also used 293T cells, which do not produce endogenous β4. Extracts from 293T cells transfected with cDNAs encoding wild-type, tailless, and headless β4 were analyzed by immunoblotting and immunoprecipitation with OCP sera. The results of these experiments confirmed that OCP sera bind selectively to the intracellular portion of β4. No binding to the extracellular portion was detected by immunoblot, immunoprecipitation, and using even the most sensitive FACS analysis.

The ability of the OCP sera and Ab to cytoplasmic domain of β4 to bind the cytoplasmic tail of β4 was studied in vitro using cells in culture. We showed that OCP sera and Ab to cytoplasmic domain of β4 demonstrate intracellular binding only.

No extracellular binding was observed. The control Ab to the extracellular domain of β4 bound only to the cell membrane and showed no binding to the cytoplasm.

We preferred to use MDA-435 cell lines instead of the 293T-transfected cells because it produced full-length β4, and any cross-reactivity between these Abs to the extracellular and cytoplasmic domains or OCP sera would have been detected. Although we have used all means available to detect the presence of Abs against extracellular domain of β4 integrin, it is important to note that they may exist as a small and undetectable, yet potentially pathogenic pool of such autoantibody. Advances in technology or evaluation of patients in preclinical states may detect them.

The intracellular portion of β4 integrin is large and consists of 1000 aa, and contains four type III fibronectin-like modules (4, 21). The cytoplasmic portion of β4 has been shown to interact with two other hemidesmosomal proteins, BP180 and HD1/plectin (4), and is required for hemidesmosome formation and stable adhesion of stratified epithelia, such as the skin and pyloris (22).

The basic pathology in OCP is a subepithelial vesicle or blister formation. In recent studies, we have demonstrated that when normal human conjunctiva is placed in organ culture with OCP sera, immunoaffinity-purified OCP autoantibodeis, or Abs to human β4 integrin, but not normal human sera, a separation between the conjunctival epithelium and underlying submucosa is observed. This separation is remarkably similar to that observed in human conjunctival pathology in vivo in OCP patients (8). These observations suggest that the autoantibodies to β4 in OCP can contribute to the initiation or progression of the disease.

The experiments in this study strongly suggest that autoantibodies to β4 contribute to the pathogenesis of OCP. The observation in this study presents the dilemma of the role of the extracellular domain of β4 in the pathogenesis of OCP and access of OCP Ab to its intracellular target Ag. Three hypothetical but plausible explanations deserve consideration. First, the disease could be initiated by Abs directed to the ectodomain of β4. However, we did not detect such Abs in the sera of the patients we studied with active disease. Hence, it is likely that such Abs could be transient and are present only in the preclinical stage of the disease, and therefore not detected.

Second, the inciting Ag initially involved may be irrelevant, but epitope spreading may involve epitopes that have homology to β4. Alternatively, autoantibodies could enter the cell to initiate injury that is subsequently propagated by Abs to exposed Ag on the cell surface. Nonetheless, Abs to the cytoplasmic epitopes of β4, BP180, and BP230 are generated and cause disruption of hemidesmosomal assembly-resultant blister formation due to BMZ separation and clinical disease.

Our direct immunofluorescence studies using cultured cells demonstrate that OCP autoantibodies and Ab to cytoplasmic domain of β4 integrin bind intracellularly to cells. A number of previous observations demonstrate that autoantibodies can penetrate into living cells, bind to their intracellular targets, and influence cellular function (23). For example, autoantibodies to ubiquitous intracellular Ags are commonly found in the sera of patients with systemic lupus erythematosus, poly and dermatomyositis, progressive systemic sclerosis, and Sjögren’s syndrome (24). Investigators have demonstrated that specific ant-DNA Abs can penetrate cells in culture (19, 20).

Furthermore, when these anti-DNA antibodies are injected into normal mice, they produce glomerular damage and proteinuria, indicating the functional capability to the autoantibodies (19). Anti-nuclear and anti-cytoplasmic Abs have been detected in vivo in human skin biopsies from the patients with connective tissue diseases (25, 26).

There are several mechanisms by which autoantibodies could penetrate cells and gain access to the intracytoplasmic target Ags. The mechanisms are probably different in different diseases and may be different in different subsets of the same disease. Based on the work of several investigators, one hypothesis has emerged and has gained some acceptance. Intracellular molecules that contain sites targeted by autoantibodies may be presented on cell membranes by surrogate molecules (27). The surrogate membrane molecule mimics the structure of the intracellular Ag (28). The complex containing the autoantibody and surrogate Ag is internalized within cell. The autoantibody is then released and binds to the pathogenic intracellular Ag and initiates the events that may eventually lead to autoimmunity (29, 30).

The observations made in this study may help in providing an explanation for the two divergent observations made by different investigators studying immunoelectron microscopy (IEM) of CP. In the first group of studies, authors report deposition of immunoreactants on the lower lamina lucida and lamina densa (31, 32, 33). The sera used by two investigators bound to the dermal side of salt split skin on indirect immunofluorescence assay and lamina densa on IEM contained Abs to laminin-5 (34). In the second group, investigators demonstrate that immunodeposits occur on hemidesmosomes and basal keratinocyte cytoplasm or the junction between hemidesmosomes and the inner plasma membrane of keratinocytes (33, 34, 35). These investigators observed that sera that produced such immune deposition did not contain Abs to laminin-5 or BP (33). In this study, using transfected cell lines, we demonstrated that patients with CP whose autoantibody binds to the epidermal side of salt split skin have Abs that bind to the cytoplasmic domain of β4 integrin. Hence, we propose that the sera studied by the second group contained Abs to cytoplasmic domain of β4 integrin, which accounts for their IEM pattern.

A definitive role for Abs to human β4 integrin in the pathogenesis of OCP can come from in vivo animal model studies. Further studies on the role of β4 in the generation of autoimmunity are important since β4 is the first integrin to be implicated in the pathogenesis of an autoimmune epithelial blistering disease. In addition, OCP may provide a model to study the possible role of integrins in initiation of progression or disease process.

1

This work was supported by Grant RO1 EY08379 from the National Eye Institute, National Institutes of Health, and in part by grants from Pemphigus Foundation and IOIMS, Kuwait.

4

Abbreviations used in this paper: OCP, ocular cicatricial pemphigoid; BMZ, basement membrane zone; BP, bullous pemphigoid; CP, cicatricial pemphigoid; DH, dermatitis herpetiformis; IEM, immunoelectron microscopy; OP, oral pemphigoid; PV, pemphigus vulgaris.

1
Ahmed, A. R., S. M. Hombal.
1986
. Cicatricial pemphigoid.
Int. J. Dermatol.
25
:
90
2
Bernard, P., C. Prost, N. Durepaire, N. B. Seguin, L. Diderjean, J. H. Saurat.
1992
. The major cicatricial pemphigoid antigen is a 180 kDa protein that shows immunologic cross reactivities with bullous pemphigoid antigens.
J. Invest. Dermatol.
99
:
174
3
Balding, S. D., C. Prost, L. A. Diaz, P. Bernard, C. Bedane, D. Aberdam, G. Giudice.
1996
. Cicatricial pemphigoid autoantibodies react with multiple sites on the BP 180 extracellular domain.
J. Invest. Dermatol.
106
:
141
4
Borradori, L., A. Sonnenberg.
1999
. Structure and function of hemidesmosomes: more than simple adhesion complexes.
J. Invest. Dermatol.
112
:
411
5
Mohimen, A., R. Neumann, C. S. Foster, A. R. Ahmed.
1993
. Detection and partial characterization of ocular cicatricial pemphigoid antigens on COLO and SCaBER tumor cell lines.
Curr. Eye Res.
12
:
741
6
Bhol, K., A. Mohimen, R. Neumann, J. Yunis, C. S. Foster, E. J. Yunis, A. R. Ahmed.
1996
. Difference in the anti-basement membrane zone antibodies in ocular cicatricial pemphigoid.
Curr. Eye Res.
15
:
521
7
Tyagi, S. R., K. Bhol, K. Natarajan, C. L. Rallatos, C. S. Foster, A. R. Ahmed.
1996
. Ocular cicatricial pemphigoid antigen: partial sequence and biochemical characterization.
Proc. Natl. Acad. Sci. USA
93
:
14714
8
Chan, R., K. Bhol, T. Nattaporn, E. Letko, R. K. Simmons, C. S. Foster, A. R. Ahmed.
1999
. The role of antibody to human β4 integrin in conjunctival basement membrane separation: possible in vitro model for ocular cicatricial pemphigoid.
Invest. Ophthalmol. Visual Sci.
40
:
2283
9
Domloge-Hultsch, N., G. J. Anhalt, W. R. Gammon, Z. Lazarova, R. A. Briggaman, M. Welch, D. A. Jabs, C. Huff, K. B. Yancey.
1992
. Anti-epiligrin cicatricial pemphigoid a subepithelial bullous disorder.
Arch. Dermatol.
130
:
1521
10
Gudula Kirtschig, M., P. Marinkovich, R. E. Burgeson, K. B. Yancey.
1995
. Anti-basement membrane autoantibody in-patients with anti-epiligrin cicatricial pemphigoid bind the subunit of laminin 5.
J. Invest. Dermatol.
105
:
543
11
Olmsted, J. B..
1981
. Affinity purification of antibodies from diazotized paper blots of heterogeneous protein samples.
J. Biol. Chem.
256
:
11955
12
Giancotti, F. G., M. A. Stepp, S. Suzuki, E. Engvall, E. Ruoslahti.
1992
. Proteolytic processing of endogenous and recombinant β4 integrin subunit.
J. Cell Biol.
118
:
951
13
Mainiero, F., C. Murgia, K. K. Wary, A. M. Curatola, A. Pepe, M. Blumemberg, J. K. Westwick, C. J. Der, F. G. Giancotti.
1997
. The coupling of α6 β4 integrin to Ras-MAP kinase pathways mediated by Shc controls keratinocyte proliferation.
EMBO J.
16
:
2365
14
Mohimen, A., A. R. Ahmed.
1995
. Comparison of indirect immunofluorescence assay and a modified sensitive immunoblot assay for the study of the autoantibody in pemphigus.
Arch. Dermatol. Res.
287
:
202
15
Spinardi, L., Y.-U. Ren, S. Raymond, F. G. Giancotti.
1993
. The β4 subunit cytoplasmic domain mediates the interaction of α6β4 integrin with the cytoskeleton of hemidesmosomes.
Mol. Biol. Cell
4
:
871
16
Spinardi, L., S. Einheber, T. Cullen, T. A. Milner, F. G. Giancotti.
1995
. A recombinant tailless integrin β4 subunit disrupts hemidesmosome, but suppresses α6β4-mediated cell adhesion to laminins.
J. Cell Biol.
129
:
473
17
Laemmli, U. K..
1970
. Cleavage of structural proteins during assembly of the head of bacteriophage T4.
Nature
277
:
680
18
Towbin, H., T. Staehlin, J. Gordon.
1979
. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications.
Proc. Natl. Acad. Sci. USA
76
:
4350
19
Viahkos, D., M. H. Foster, A. A. Ucci, K. Barrett, S. K. Datta, M. P. Madaio.
1992
. Murine monoclonal anti-DNA antibodies penetrate cells, bind to nuclei, and induce glomerular proliferation and proteinuria in vivo.
J. Am. Soc. Nephrol.
2
:
1345
20
Yanase, K., R. M. Smith, B. Cizman, M. H. Foster, L. D. Peachey, L. Jarett, M. P. Madaio.
1994
. A subgroup of murine monoclonal anti-deoxyribonucleic acid antibodies traverse the cytoplasm and enter the nucleus in a time and temperature dependent manner.
Lab. Invest.
71
:
52
21
Sonnenberg, A., J. Calafat, H. Janssen, H. Daams, L. M. H. Vander Raaij-Hemler, R. Falcioni, S. J. Kennel, J. D. Aplin, J. Baker, M. Loizidou, D. Garrod.
1991
. Integrin α6β4 complex is located in hemidesmosomes, suggesting a major role in epidermal cell basement membrane adhesion.
J. Cell Biol.
113
:
907
22
Murgia, C., P. Blaikie, N. Kim, M. Dans, H. T. Petrie, F. G. Giancotti.
1998
. Cell cycle and adhesion defects in mice carrying a targeted deletion of the integrin β4 cytoplasmic domain.
EMBO J.
17
:
3940
23
Alarcon-Segovia, D., A. Ruiz-Arguelles, L. Llorente.
1996
. Broken dogma: penetration of autoantibodies into living cells.
Immunol. Today
17
:
173
24
Reichlin, M..
1995
. Cell injury mediated by autoantibodies to intracellular antigens.
Clin. Immunol. Immunopathol.
76
:
215
25
Velthius, P. J., I. Kater, I. Van der Tweel, F. Gmelig-Meyling, R. H. W. M. Derken, R. J. Hene, J. A. Van Geeutselaar, H. Baart de la Faille.
1990
. In vivo antinuclear antibody of the skin: diagnostic significance and association with selective anti-nuclear antibodies.
Ann. Rheum. Dis.
49
:
163
26
Burrows, N. P., B. S. Bhopal, R. R. Jones, M. M. Black.
1993
. Clinicopathological significance of cutaneous epidermal nuclear staining by direct immunofluorescence.
J. Cutan. Pathol.
20
:
159
27
Ioannides, D., J. Bystryn.
1993
. Association of tissue-fixed cytoplasmic deposits of immunoglobulins in epidermal keratinocytes with lupus erythematosus.
Arch. Dermatol.
129
:
1130
28
Koren, E., W. Schnitz, M. Reichlin.
1993
. Concomitant development of chronic active hepatitis and antibodies to ribosomal protein in patient with systemic lupus erythematosus.
Arthritis Rheum.
36
:
1325
29
Jacob, L., F. Tron, J. F. Bach, F. Lauvard.
1984
. A monoclonal anti-DNA antibody also binds to cell surface proteins.
Proc. Natl. Acad. Sci. USA
81
:
3843
30
Raz, E., H. Ben-Bassat, T. Davidi, Z. Shlomai, D. Eliat.
1993
. Cross-reactions of anti-DNA antibodies with cell surface proteins.
Eur. J. Immunol.
23
:
383
31
Fine, J. D., G. R. Neises, S. I. Katz.
1984
. Immunofluorescence and immunoelectron microscopic studies in cicatricial pemphigoid.
J. Invest. Dermatol.
82
:
39
32
Prost, C., B. Labeille, V. Chaussade, J. C. Guillaume, N. Martin, L. Dubertret.
1987
. Immunoelectron microscopy in subepidermal autoimmune bullous disease: a prospective study of IgG and C3 bound in vivo in 32 patients.
J. Invest. Dermatol.
89
:
563
33
Shimizu, H., T. Masunaga, A. Ishiko, K. Matasumura, T. Hashimoto, T. Nishikawa, N. Domloge-Hultsch, Z. Lazarova, K. B. Yancey.
1995
. Autoantibodies from patients with cicatricial pemphigoid target different sites in epidermal basement membrane.
J. Invest. Dermatol.
104
:
370
34
Nieboer, C., D. M. Boorsma, M. J. Woerdeman.
1982
. Immunoelectron microscopic finding in cicatricial pemphigoid: their significance in relation to epidermolysis bullosa aquisita.
Br. J. Dermatol.
106
:
419
35
Bedance, C., C. Prost, P. Bernard, G. Catanzano, J. M. Bonnetblanc, L. Dubertret.
1991
. Cicatricial pemphigoid antigen differs from bullous pemphigoid antigen by its exclusive extracellular localization: a study by indirect immunoelectron microscopy.
J. Invest. Dermatol.
97
:
3