Specific affinity between fibronectin and the epidermolysis bullosa acquisita antigen.

Autoantibodies in the skin and sera of patients with epidermolysis bullosa acquisita bind to a large matrix molecule within the lamina densa region of skin basement membrane. At the site of these immune complexes, the epidermis separates from the dermis, which creates a subepidermal blister just below the lamina densa. The target molecule for the autoantibodies is in close apposition to fibronectin, a major extracellular matrix molecule that is abundant in the upper dermis of skin. In this report, we show specific affinity between fibronectin and the 290,000-D chain of the epidermolysis bullosa acquisita antigen, and that this affinity is mediated by the gelatin/collagen-binding domain of fibronectin (Mr = 60,000). Since blistering in epidermolysis bullosa acquisita often occurs in the absence of clinical and histological inflammation, a direct interruption in the fibronectin-epidermolysis bullosa acquisita antigen bond may be involved in the pathogenesis of epidermal-dermal disadherence that occurs in this bullous disease.

Epidermolysis bullosa acquisita antigen is the globular carboxyl terminus of type VII procollagen.

Epidermolysis bullosa acquisita (EBA) is a severe, chronic blistering disease of the skin. EBA patients have circulating and tissue-bound autoantibodies to a large (Mr = 290,000) macromolecule that is localized within the basement membrane zone between the epidermis and dermis of skin, the site of blister formation. The "EBA antigen" is known to be distinct from laminin, heparan sulfate proteoglycan, fibronectin, the bullous pemphigoid antigen, elastin, and collagen types I, II, III, IV, and V. Sera from patients with EBA, two monoclonal antibodies to the EBA antigen, and a monoclonal antibody to the carboxyl terminus of type VII procollagen identically label human amnion and skin by immunofluorescent and immunoelectron microscopy. Western immunoblots of the EBA antigen extracted from skin and of type VII procollagen labeled with the above sera and antibodies are identical. None of the sera or antibodies labels Western blots of pepsinized type VII collagen which is missing the globular amino and carboxyl terminal domains. These data show that the EBA antigen is the carboxyl terminus of type VII procollagen.

Human dermal fibroblasts synthesize laminin.

Laminin, a glycoprotein of approximately 900,000 daltons, is a major component of the basement membrane that separates the epidermis from dermis in human skin. Previous studies have shown that keratinocytes and other epithelial cells synthesize laminin and utilize it for attachment to other extracellular matrices such as heparan sulfate proteoglycan and basement membrane collagen. The relationships between phenotypically normal mesenchymal cells and laminin have been much less emphasized in the literature. In this study, we have used antibodies that specifically label the A and B chains of laminin (but not fibronectin or other unrelated proteins) by Western blot analysis to immunoprecipitate biosynthetically derived laminin from [35S] methionine labeled cultures of neonatal and adult human skin fibroblasts. To be sure that the precipitated bands were laminin and not fibronectin, which has a molecular size very close to that of the laminin B chains, experiments were performed in which fibronectin was removed from the radiolabeled proteins by first immunoprecipitating with antifibronectin antibody and then sequentially immunoprecipitating laminin from the fibronectin-depleted supernates with antilaminin antibody. These experiments definitively demonstrate that human dermal fibroblasts synthesize and secrete laminin.

Neonatal foreskin substrate has limitations for the immunofluorescent screening of monoclonal antibodies.

Two monoclonal antibodies to type IV collagen showed a marked decrease in the labeling of the dermal-epidermal junction of neonatal foreskin while the basement membrane around dermal blood vessels was brightly stained. In contrast, these antibodies labeled the junction and dermal blood vessels with approximately equal intensity when adult skin of nonforeskin site was used as substrate. Other antibodies to matrix molecules (bullous pemphigoid antigen, epidermolysis bullosa acquisita antigen, and laminin) showed excellent staining of both the dermal-epidermal junction and dermal blood vessels in both neonatal foreskin and adult skin. Further, the ultrastructural appearance of the substrates appeared identical. The implication is that neonatal foreskin is not a good substrate to use for the routine screening of monoclonal antibodies to matrix components by indirect immunofluorescence since a "false negative" evaluation may occur.

Epidermolysis bullosa acquisita antigen, a major cutaneous basement membrane component, is synthesized by human dermal fibroblasts and other cutaneous tissues.

The epidermolysis bullosa acquisita (EBA) antigen is identified as 2 chains: a 290,000-dalton protein and a less prominent 145,000-dalton protein. The 290,000-dalton chain is synthesized by human keratinocytes in culture. In this study, we show that the 290,000-dalton chain is synthesized by human skin fibroblasts and cutaneous human tumors. In contrast, HT1080 cells, a human sarcoma cell line known to produce matrix molecules (such as laminin and type IV collagen), does not synthesize the EBA antigen. Further, the EBA antigen is absent from serum and blood components, placenta, amnion, lung, and the EHS tumor, a murine sarcoma that produces large amounts of laminin, type IV collagen, nidogen, entactin, and basement membrane proteoglycan but is present in cutaneous tumors of adnexal and epithelial origin. These data suggest that while the EBA antigen is synthesized by both human skin keratinocytes and fibroblasts and is therefore not specific for a primordial germ layer, it does appear to be specific for tissue containing a stratified squamous epithelium.

Epidermolysis bullosa acquisita antigen, a new major component of cutaneous basement membrane, is a glycoprotein with collagenous domains.

The epidermolysis bullosa acquisita antigen is a major constituent of the basement membrane zone beneath stratified squamous epithelium. The antigen which is recognized in extracts of skin basement membrane by Western blot analysis with polyclonal or monoclonal antiepidermolysis bullosa acquisita antigen antibodies as 2 chains (a major chain of 290,000 and a minor chain of 145,000) has a native molecular weight over 800,000. Both epidermolysis bullosa acquisita antigen chains contain carbohydrate and the 290K chain is sensitive to collagenase.

Localization of the alpha 3 (V) chain of type V collagen in human skin.

Serum from goats immunized with human type V collagen chains that were cut out of polyacrylamide gels contained an antibody that recognized only type V collagen in an enzyme-linked immunoabsorbent assay and did not label laminin, fibronectin, or types I and IV collagen. Western blot analysis of the antibody showed that its determinant was the alpha 3 (V) chain of type V collagen. Indirect immunofluorescent staining of intact human skin with the antibody produced staining of the dermal blood vessels but not of the dermal-epidermal junction (DEJ). In contrast, both the dermal blood vessels and the DEJ were labeled by the antibody if the skin substrate was first split through the lamina lucida region of the DEJ by incubation in 1 M NaCl solution. Indirect immunoelectron microscopy confirmed the staining pattern found by immunofluorescence and defined the ultrastructural localization of type V collagen in skin. Type V collagen is localized within the DEJ to the lamina lucida region and polar aspects of the basal cell keratinocyte plasma membrane.

Differential neutralizing activity of human antibodies in interferon antiviral and natural killer cell bioassays.

The ability of interferon-alpha (IFN-alpha) to augment the cytotoxicity of human natural killer (NK) cells was used to probe the neutralization capacity of human antibodies to IFN-alpha. Sera from patients treated with IFN-alpha were tested for antibodies that could bind to IFN-alpha, neutralize IFN-alpha antiviral activity, or neutralize IFN-alpha-mediated augmentation of NK cytolytic activity. The NK-augmenting activity of IFN-alpha was measured in a chromium-release cytotoxicity assay using K562 targets and human peripheral blood mononuclear cells in the presence of a constant amount of antibody from patients treated with IFN-alpha. Neutralization of IFN-alpha-mediated antiviral activity did not correlate with neutralization of NK augmentation. However, all sera that neutralized a biologic activity of IFN-alpha also bound IFN-alpha. Conversely, sera that did not bind to IFN-alpha also did not neutralize either biologic activity. The data suggest that some immunogenic epitopes of IFN-alpha reside in distinct domains that mediate different biologic activities of IFN-alpha. The identification of neutralizing antibodies for the NK immunomodulating function of IFN-alpha but not antiviral function suggests that assessment of antiviral neutralization alone may be an incomplete evaluation of the potential significance of binding antibodies that occur subsequent to the administration of therapeutic IFNs.

A kinetic enzyme immunoassay for the quantitation of antibodies to a humanized monoclonal antibody in human serum.

A kinetic enzyme immunoassay was developed and validated to quantitate human antibodies to the humanized monoclonal antibody CAMPATH1-1H (C1H) in human serum. The assay was configured using C1H-coated 96-well plates which were blocked with bovine serum albumin, and incubated with dilutions of human serum containing anti-C1H antibody. Antibody was detected using biotinylated C1H followed by streptavidin-conjugated alkaline phosphatase and p-nitrophenyl phosphate. Absorbance data were collected for 10 min, and mOD min-1 data were exported to MultiCalc data analysis software. A 4-parameter logistic-log algorithm was shown to model the data through the range of the standard curve within 15% of nominal values. The overall assay performance coefficient of variation by ANOVA was 9.2%. The lower limit of detection was defined at 160 Units ml-1. The anti-idiotype antibody standard stock solution is stable at 4 degrees C and at -80 degrees C for at least 11 months in buffer. The anti-idiotype antibody controls are stable for at least seven freeze-thaw cycles and at least 6 months in human serum stored at -20 degrees C. A strategy was devised by which to establish the specific antibody potency for any given batch of anti-C1H antibody standard relative to the Reference Standard. This EIA has been used to quantify and characterize anti-C1H antibody in human serum in support of clinical safety and efficacy studies.

Immunochemical characterization of human antibodies to lymphoblastoid interferon.

Antibodies produced in recurrent respiratory papillomatosis (RRP) patients treated with lymphoblastoid interferon (lyIFN) may neutralize antiviral activity or may only bind to lyIFN. These antibodies were characterized for immunoglobulin class, IgG subclass, and light chain type by an indirect immunoassay. Serum dilutions were incubated on lyIFN-coated plates and the presence of antibody detected using peroxidase-conjugated goat antibodies to each human immunoglobulin class and light chain isotype, or using MoAbs to each human IgG subclass. Neutralizing activity was measured as the inhibition of lyIFN antiviral activity for Vervet monkey cells challenged with Semliki Forest Virus. Among antibody-positive patients, 12% produced IgM coincident with IgG, and 25% produced IgA coincident with IgG. Thus, antibody responses in patients treated with lyIFN are not exclusively of IgG class. The predominant lyIFN-specific subclasses were IgG1 and IgG3, which occurred in 70% and 83% of patients, respectively. An IgG4 response was detected in two patients who also had antibody of other isotypes; no IgG2 antibody was detected in any patient. Antibodies were not IgG subclass-restricted, a trend which was more pronounced in patients having neutralizing antibody than non-neutralizing antibody. Light chain molecules of lyIFN-specific antibody were of both kappa and lambda isotypes, with kappa chains occurring most frequently. Among patients having non-neutralizing antibodies, monotypic light chains occurred in 65% of the patients, whereas no patient with neutralizing antibody had monotypic light chain antibody. Sera from 599 normal human volunteers were assayed for antibody, and seven were found to be immunoreactive to lyIFN. Only one serum of the seven was positive for neutralizing activity.