If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Institute for Dental Research and Oral Musculoskeletal Biology, Center for Biochemistry, and Center for Molecular Medicine Cologne (CMMC), Medical Center, University of Cologne, Cologne, Germany
Institute for Dental Research and Oral Musculoskeletal Biology, Center for Biochemistry, and Center for Molecular Medicine Cologne (CMMC), Medical Center, University of Cologne, Cologne, Germany
Department of Dermatology, University Medical Center Freiburg, Freiburg, GermanyCenter for Biological Signaling Studies (BIOSS), University of Freiburg, Freiburg, Germany
Autoimmunity against laminins has been described in several autoimmune diseases (including mucous membrane pemphigoid, anti–laminin γ1 pemphigoid, and connective tissue diseases), in pregnancy loss, and in infections such as Chagas disease. Except for anti–laminin-332 mucous membrane pemphigoid, adequate evidence has been lacking for the tissue injury potential of laminin-specific antibodies and the pathogenic epitopes. We evaluated the pathogenic potential of antibodies targeting laminin γ1, a major constituent of basement membranes and the main antigen in anti–laminin γ1 pemphigoid. Rabbit antibodies were generated against fragments of the N-terminus and C-terminus of murine laminin γ1, and their ability to disrupt ligand interactions and/or to activate complement and granulocytes was assessed using previously established ex vivo assays. Our findings document a pathogenic potential of antibodies targeting the laminin γ1 N-terminus. These antibodies interfere with the binding of nidogen to laminin and can activate granulocytes and the complement cascade. We detected antibodies with different degrees of reactivity with laminin γ1 N-terminus in patients with anti–laminin γ1 pemphigoid, cutaneous lupus erythematosus, and scleroderma. Our results provide mechanistic insights into the tissue damage associated with laminin autoimmunity and could facilitate development of appropriate diagnostic tools and therapeutic strategies.
Autoimmunity is characterized by a failure or breakdown of self-tolerance mechanisms, leading to a B-cell and/or T-cell–mediated immune response directed against the self that can induce tissue damage and result in autoimmune disease. Among the immunological players involved in autoimmune disease development, the inappropriate production of autoantibodies that mediate tissue injury plays a role in several of the more than 80 described autoimmune diseases.
The interplay between antibodies and self-antigen, either circulating or tissue-bound, may induce damage by several effector mechanisms, including inflammatory and noninflammatory complement-mediated and Fc-mediated reactions, as well as Fc-independent noninflammatory pathways.
Antibodies bound to self-antigens from cell surfaces can trigger tissue damage by Fc-mediated or complement-mediated mechanisms, including induction of cell phagocytosis, cell lysis, and recruitment and activation of inflammatory cells, as is described in autoimmune diseases such as autoimmune hemolytic anemia, rheumatoid arthritis, myasthenia gravis, and bullous pemphigoid. Antibodies may, however, exert direct pathogenic effector functions in an Fc-independent manner, including stimulation of the bound receptor (as in Graves' disease) or disruption of protein–protein interactions through steric hindrance (as in pemphigus).
The in vivo pathogenicity of autoantibodies in several autoimmune disorders has been demonstrated by the passive transfer of antigen-specific IgG from patients or immunized animals.
Ex vivo models have also proved useful in confirming the tissue damage-inducing potential of antibodies in autoimmune diseases, including skin blistering disorders, multiple sclerosis, neuromyelitis optica, and antiphospholipid syndrome.
A retrospective study of antibodies against basement membrane antigens (type IV collagen and laminin) in patients with primary and secondary Raynaud's phenomenon.
Laminin γ1 is an extracellular protein present in all basement membranes as a constitutive chain of almost all laminin isoforms. It has a critical role in embryogenesis; its absence in mice results in lack of basement membrane formation and death at the peri-implantation stage of embryonic development.
Through its short arm located at the N-terminus, laminin γ1 binds nidogen, which cross-links the two meshworks of laminins and collagen IV within the basement membrane.
The requirement of the glutamic acid residue at the third position from the carboxyl termini of the laminin gamma chains in integrin binding by laminins.
It has been recently shown that serum of anti–laminin γ1 pemphigoid patients reacts mainly with the C-terminus of laminin γ1, but epitopes located outside this fragment may also be targeted.
Anti–laminin γ1 pemphigoid is an autoimmune subepidermal blistering disease characterized by circulating and tissue-bound antibodies and a neutrophil-rich inflammatory infiltrate.
Interestingly, despite the widespread distribution of the autoantigen, the pathology remains organ-specific. The pathogenic potential of antibodies targeting laminin-111 is still poorly understood. Immunization with the autoantigen and passive transfer of specific antibodies in animal models had little effect or (depending on the experimental setting) induced limited renal disease, myositis, abortion, or reproductive failure.
Because a skin-specific effect had not yet been demonstrated for autoantibodies against laminin γ1,18 in the present study we addressed the ex vivo pathogenic potential of laminin γ1–specific antibodies. For this purpose, rabbit antibodies were generated against N- and C-terminal fragments of mouse laminin γ1. Subsequently, their direct effect on nidogen and integrin interactions was investigated in a cell adhesion assay and a solid-phase ligand inhibition and disruption assay. Furthermore, the ability of the newly generated antibodies to activate the complement system and/or inflammatory cells was assessed in a complement fixation test and antibody-induced granulocyte-dependent dermal–epidermal separation assay or by measurement of reactive oxygen species (ROS) production. Our findings show that anti–laminin γ1 antibodies, especially those targeting the N-terminus, have pathogenic potential either through direct inhibition and/or disruption of the nidogen 1 binding or by Fc-mediated activation of granulocytes and complement cascade. Interestingly, autoantibodies against laminin γ1 N-terminus were detected not only in cases of anti–laminin γ1 pemphigoid, but also in cutaneous lupus erythematosus (CLE) and scleroderma.
Materials and Methods
Proteins
Recombinant human laminin-511 was purchased from BioLamina (Stockholm, Sweden), natural mouse laminin-111 from Life Technologies–Invitrogen (Darmstadt, Germany; Carlsbad, CA), and recombinant human α6β4 from R&D Systems (Minneapolis, MN). Human laminin γ1 N-terminus was expressed in a mammalian system as described previously.
The murine laminin γ1 fragments and nidogen 1 were recombinantly expressed in a mammalian or a bacterial system, as described below.
Heterologous Expression of Recombinant Forms of Murine Laminin γ1 and Nidogen 1
The DNA sequence data for murine laminin γ1 was retrieved from GenBank (http://www.ncbi.nlm.nih.gov/nuccore, accession number NM_010683) using the accession number NM_010683. The cDNA sequence encoding mouse laminin γ1 N-terminal fragment [mLnγ1-Nterm, amino acids (aa) 30–988] was synthesized from total RNA (SuperScript III kit; Life Technologies–Invitrogen) using primers as presented in Table 1 and cloned via NheI/BamHI restriction sites into a modified pCEP-Pu expression vector containing a thrombin cleavage site next to a 2×StrepII/FLAG tag at the C-terminal end of the protein sequence.
293-EBNA cells were transfected with the expression vector, and the supernatant was collected and supplemented with 1 mmol/L phenylmethylsulfonyl fluoride (Sigma-Aldrich, Munich, Germany; St. Louis, MO). After filtration, the supernatant was passed through a Strep-Tactin Sepharose column (IBA, Göttingen, Germany), and the recombinant protein was eluted with buffer [100 mmol/L Tris, 150 mmol/L NaCl (pH 7.4)] containing 2.5 mmol/L d-desthiobiotin (Sigma-Aldrich). For the removal of the 2×StrepII/FLAG tag, the purified glycosylated protein (Supplemental Figure S1) was incubated with 1 U bovine thrombin (Sigma-Aldrich) per milligram protein overnight, at room temperature [(in Tris-buffered saline with 5 mmol/L CaCl2 (pH 8.2)]. Afterward, the protein was applied on a Strep-Tactin Sepharose column and the digested protein was collected in the flow-through.
Table 1Primer Sequences for PCR Amplification of cDNA Fragments of Murine Laminin γ1 and Nidogen 1
The cDNA sequences coding for the three overlapping fragments of murine laminin γ1 C-terminus (aa 1018–1607) were cloned into prokaryotic vectors and expressed in Escherichia coli as glutathione S-transferase (GST)–fusion proteins according to protocols described previously.
Synthesis of the primers used in PCR (Table 1), as well as cloning of the cDNA encoding the fragments into pUC57 vector, was performed at GenScript (Piscataway, NJ). The cDNA sequences were subcloned into the linearized pGEX-6P-1 vector (Amersham; GE Healthcare, Freiburg, Germany; Little Chalfont, UK) using the BamHI/SalI cutting site, resulting in the recombinant vectors pGEX-LAMC1-Cterm-1, pGEX-LAMC1-Cterm-2, and pGEX-LAMC1-Cterm-3. The correct ligation and in-frame insertion of various DNA fragments were confirmed by DNA sequence analysis performed by GATC Biotech (Konstanz, Germany). Recombinant GST-fusion proteins were expressed in E. coli strain BL21 and purified by glutathione agarose affinity chromatography, as described previously.
The DNA sequence data for murine nidogen 1 was retrieved from GenBank (http://www.ncbi.nlm.nih.gov/nuccore, accession number NP_035047.2). The full-length murine nidogen 1 cDNA sequence obtained by reverse transcription from total RNA (SuperScript III kit, Life Technologies–Invitrogen) was cloned into a modified pCEP-Pu vector containing a thrombin cleavage site next to a 2×StrepII/FLAG tag at the N-terminal end of the protein sequence and expressed in 293-EBNA cells. Primer sequences are presented in Table 1. Protein concentrations were determined by spectrophotometry at 280 nm (NanoDrop 1000 spectrophotometer; Thermo Fisher Scientific, Waltham, MA).
Immunoblot Analysis
Recombinant 2×StrepII/FLAG–mLnγ1-Nterm, GST-tagged C-terminal fragments (GST-mLnγ1-Cterm-1, GST-mLnγ1-Cterm-2, GST-mLnγ1-Cterm-3), and 2×StrepII/FLAG–nidogen 1 were fractionated by SDS-PAGE, transferred onto nitrocellulose, and analyzed by immunoblotting.
For detection of the GST-tagged fragments, after blocking, the membrane was incubated for 2 hours at room temperature with goat polyclonal antibody specific to GST (GE Healthcare) diluted 3000-fold. Subsequently, the bound goat IgG was detected with 5000-fold diluted horseradish peroxidase (HRP)–labeled antibody (Promega, Mannheim, Germany; Madison, WI) and diaminobenzidine. GST alone was used as control. For detection of recombinant 2×StrepII/FLAG-fusion nidogen 1, the membrane was incubated with murine monoclonal antibody specific to StrepII tag (IBA) diluted 2000-fold, for 1 hour at room temperature. Next, the murine antibody was detected with a 3000-fold diluted HRP-coupled antibody (Dako, Hamburg, Germany; Carpinteria, CA), and the signal was visualized with chemiluminescence substrate. The recombinant 2×StrepII/FLAG-fusion mLnγ1-Nterm and the tag-free mLnγ1-Nterm were detected using the 1000-fold diluted serum of rabbits immunized with mLnγ1-Nterm, for 1 hour at room temperature, followed by incubation with the 3000-fold diluted HRP-coupled antibody specific to rabbit IgG (Dako) and then signal detection by chemiluminescence.
Antibodies
Rabbit polyclonal antibodies were produced against the recombinant mLnγ1-Nterm fragment and the GST-tagged mLnγ1-Cterm-1, -2, and -3 fragments according to established protocols.
Rabbit polyclonal antibodies were also generated against the recombinant GST-fusion protein containing sequences of murine collagen VII (GST-mCVII-1, aa 1-300 and GST-mCVII-2, aa 281-594)
The eluted fractions were concentrated by ultrafiltration using Amicon Ultra-15 (15 mL, 30 kDa) centrifugal filter units (EMD Millipore, Billerica, MA) in a centrifuge at 3600 × g and 4°C, 45 minutes per cycle. Protein concentration was measured using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific).
Human Sera
Serum samples were obtained from patients with anti–laminin γ1 pemphigoid (n = 7), CLE (discoid CLE, n = 13; subacute CLE, n = 2; acute CLE, n = 4), dermatomyositis (n = 9), systemic sclerosis, (n = 10); and localized scleroderma, (n = 14). Anti–laminin γ1 pemphigoid patients were characterized by i) blisters on the skin; ii) linear deposits of IgG at the dermal–epidermal junction (DEJ), as observed under direct immunofluorescence (IF) microscopy; iii) circulating IgG binding the dermal side of the 1 mol/L NaCl-split skin, as observed under indirect IF microscopy; and iv) immunoblot reactivity with the 200-kDa protein from dermal extracts. Normal control serum samples were obtained from healthy donors (n = 20). Informed consent was obtained from patients whose material was used in the study, in adherence with Helsinki principles.
Complement Fixation Test
The ability of the antibodies to activate the complement cascade was assessed in vitro by a complement fixation test, according to described protocols.
In brief, cryosections (6 μm thick) of normal mouse skin were incubated with normal rabbit serum, with rabbit serum raised against mLnγ1-Nterm (diluted 1:10) or against GST-mLnγ1-Cterm-1, -2, or -3 (diluted 1:2), or with purified rabbit IgG antibodies targeting the C-terminal fragments of laminin γ1 (diluted 1:10), followed by incubation with serum from a healthy human donor fivefold diluted in gelatin Veronal buffer (Sigma-Aldrich) as a source of complement. Rabbit polyclonal antibody targeting GST-mCVII-2 (diluted 1:10) was used as control. Bound human complement C3 was subsequently detected with 100-fold diluted fluorescein isothiocyanate–labeled antibody (Cappel; MP Biomedicals, Eschwege, Germany; Santa Ana, CA). Sections were counterstained with 1000-fold diluted DAPI (Sigma-Aldrich).
Ex Vivo Cryosection Assay
The blister-inducing potential of antibodies targeting different fragments of murine laminin γ1 was determined by an ex vivo assay of antibody-induced granulocyte-dependent dermal–epidermal separation, as described previously.
In brief, cryosections of murine skin were incubated with rabbit serum against mLnγ1-Nterm (diluted 1:5), GST-mLnγ1-Cterm-1, -2, or -3 (diluted 1:2) or with purified rabbit IgG specific to mLnγ1-Nterm (diluted 1:10), GST-mLnγ1-Cterm-1, -2, or -3 (diluted 1:5). Rabbit polyclonal antibody specific to GST-mCVII-1 (diluted 1:5) was used as control. Leukocytes isolated from healthy human donors were added to cryosections and incubated at 37°C for 3 hours. The extent of dermal–epidermal separation was assessed under a microscope and quantified as the percentage of the split area at the DEJ relative to the total length of the DEJ.
Solid-Phase Binding Assay
Recombinant mLnγ1-Nterm [5 μg/mL in 0.1 mol/L carbonate buffer (pH 9.3)] was coated onto 96-well microtiter plates (Greiner Bio-One, Frickenhausen, Germany) overnight, at 4°C. The wells were then blocked with 1% bovine serum albumin (BSA) (fraction V; Sigma-Aldrich) in PBS and subsequently incubated with serial dilutions of recombinant mouse 2×StrepII/FLAG-fusion nidogen 1. The amount of bound nidogen was detected with 5000-fold diluted HRP-tagged mouse antibody against StrepII tag (AbD Serotec, Kidlington, UK; Raleigh, NC).
For enzymatic reaction, 50 μL of orthophenylenediamine (Sigma-Aldrich) solution in water and 0.1% (v/v) H2O2 was added to each well. The reaction was stopped after 10 minutes with 50 μL/well of 0.5 mol/L H2SO4, and absorbance at 490 nm was read using a microplate reader (Labsystems Multiskan Multisoft; Thermo Fisher Scientific).
For the inhibition assay, mLnγ1-Nterm–coated wells were first incubated with serial dilutions of rabbit serum reactive with mLnγ1-Nterm, patient's serum, or normal rabbit or human serum for 1 hour at room temperature. Subsequently the recombinant 2×StrepII/FLAG-tagged nidogen 1, 3.5 nmol/L in blocking buffer was added to the wells and the plate was incubated for 2 hours at room temperature. After washing, the bound nidogen was detected with 5000-fold diluted HRP-tagged antibody against StrepII tag (AbD Serotec).
For the disruption of protein binding assay, wells were first incubated with nidogen 1, followed by incubation with serial dilutions of rabbit serum against mLnγ1-Nterm or normal rabbit serum; subsequently, the nidogen 1 was detected. Inhibition and disruption were calculated as percentages, relative to uninhibited controls.
For integrin binding inhibition assay, published protocols were followed, with minor modification.
Ligand-binding specificities of laminin-binding integrins: a comprehensive survey of laminin-integrin interactions using recombinant alpha3beta1, alpha6beta1, alpha7beta1 and alpha6beta4 integrins.
In brief, 10 nmol/L recombinant human laminin-511 was coated onto 96-well microtiter plates in 0.1 mol/L carbonate buffer (pH 9.3) overnight, at 4°C. Wells were washed with PBS, blocked with 1% BSA in PBS, and then incubated with a 25-fold diluted mix of rabbit serum specific to GST-mLnγ1-Cterm-1, -2, or -3 or with normal rabbit serum in blocking buffer. After washing with PBS containing either 1 mmol/L MnCl2 and 2 mmol/L MgCl2 or 10 mmol/L EDTA, 80 nmol/L recombinant human integrin α6β4 (R&D Systems) in blocking buffer was added in the presence of 2 mmol/L MgCl2 and 1 mmol/L MnCl2 or of 10 mmol/L EDTA and incubated for 2 hours at room temperature. After washing, the bound integrin α6β4 was detected by a 3000-fold diluted HRP-labeled rabbit polyclonal anti-His antibody (Novus Biologicals, Littleton, CO). Readings obtained in the presence of 10 mmol/L EDTA were subtracted. The binding of integrin was calculated as percentage of the value obtained in the absence of rabbit serum.
Cell Culture and Cell Adhesion Assay
HaCaT immortalized human keratinocytes (Dr. Petra Boucamp, DKFZ, Heidelberg, Germany) were cultured in Dulbecco's modified Eagle's medium (Lonza, Basel, Switzerland) supplemented with 10% fetal calf serum, 4 mmol/L l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin (Biochrom; Merck Millipore, Billerica, MA). Cell adhesion assay was performed according to published protocols, with minor modification.
In brief, 1 μg/well of recombinant human laminin-511 or BSA was coated onto 96-well, non–tissue-treated microtiter plates (Nalge Nunc International, Roskilde, Denmark; Penfield, NY). After blocking for 2 hours at 4°C with 1% BSA in PBS, HaCaT cells were added to the wells as a single-cell suspension at a density of 1 × 105 cells per well in serum-free medium supplemented with 2 mmol/L MgCl2 and 1 mmol/L MnCl2, followed by incubation for 60 minutes at 37°C in the presence of 5% CO2. For inhibition assays, a 25-fold diluted mix of rabbit sera raised against the three GST-mLnγ1-Cterm fragments, or normal rabbit serum, or the monoclonal antibody 4C7 (Merck Millipore) (50- or 100-fold diluted) was incubated with laminin-511–coated wells for 60 minutes at room temperature before incubation with cells. The 4C7 antibody can interfere with cell attachment by interacting with the laminin α5 LG domain.
After washing, cells were fixed with 3% formalin solution and stained with 0.1% crystal violet. Adhesion was quantified by lysing the adherent cells with 2% SDS solution and determining the amount of released dye by reading the absorbance at 550 nm with a microtiter plate reader. Readings obtained for nonspecific cell adhesion in the presence of BSA were subtracted. Cell attachment was expressed as percentage of the adhesion in the absence of rabbit serum.
Binding of Anti–Laminin γ1 Antibodies in Murine Skin
Organ culture of skin explants was performed according to published protocols, with minor modification.
In brief, punch biopsies (0.5 cm × 0.5 cm) of mouse tail were incubated for 12 hours at room temperature or at 37°C with 300 μL of rabbit serum specific to mLnγ1-Nterm or normal rabbit serum diluted 1:10 in culture medium. The culture medium consisted of Dulbecco's modified Eagle's medium (Lonza) supplemented with 10% (v/v) fetal bovine serum (FBS), 4 mmol/L l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin (all from Biochrom). After a washing, the samples were frozen and stained for the detection of rabbit IgG deposition by IF microscopy.
For antibody injection into mice, 6- to 8-week-old BALB/c female mice with a body weight of approximately 17 g were obtained from Charles River Laboratories International (Wilmington, MA). All injections were performed on mice narcotized by inhalation of isoflurane or subcutaneous administration of a mixture of 100 μg/g ketamine and 15 μg/g xylazine. The experiments were approved by the local Animal Care and Use Committee (G-02/2012) and were performed by certified personnel. Mice were injected subcutaneously twice with 15 mg of purified rabbit IgG specific to mLnγ1-Nterm (n = 5) or with normal rabbit IgG (n = 3). Skin samples were frozen in Tissue-Tek OCT optimal cutting temperature compound (Sakura Finetek Europe, Zoeterwoude, Netherlands) and analyzed by direct IF microscopy for IgG deposition.
Enzymatic Deglycosylation of Recombinant Mouse Laminin γ1 N-Terminus
Analysis of post-translational glycosylation of the recombinant murine 2×StrepII/FLAG–laminin γ1 N-terminus expressed in a eukaryotic system was performed by enzymatic deglycosylation in 200 mmol/L ammonium hydrogen phosphate buffer. Protein (10 μg) was incubated with 500 units PNGase F (New England Biolabs, Ipswich, MA) overnight, at 37°C. The deglycosylated protein was analyzed on 12% SDS-PAGE gel under reducing conditions and was detected by immunoblotting with polyclonal rabbit antibodies specific to mLnγ1-Nterm.
Indirect IF Microscopy
Detection of rabbit antibodies by indirect IF microscopy was performed according to established protocols, with minor modification.
Alport alloantibodies but not Goodpasture autoantibodies induce murine glomerulonephritis: protection by quinary crosslinks locking cryptic α3(IV) collagen autoepitopes in vivo.
In brief, cryosections of normal mouse or human skin were incubated first with 100-fold diluted rabbit serum or 10-fold diluted Biotin-tagged rat anti–laminin γ1 monoclonal antibody (Thermo Fisher Scientific), followed by incubation with 100-fold diluted Alexa Fluor 488–labeled goat antibody against rabbit IgG or Alexa Fluor 488–labeled streptavidin (both from Life Technologies–Invitrogen). For unmasking cryptic epitopes, cryosections were treated with 6 mol/L urea in 0.1 mol/L glycine (pH 2.2), for 5 minutes on ice, before incubation with rabbit serum, according to published protocols.
Alport alloantibodies but not Goodpasture autoantibodies induce murine glomerulonephritis: protection by quinary crosslinks locking cryptic α3(IV) collagen autoepitopes in vivo.
Images were acquired using a Nikon Eclipse 80i microscope equipped with a Nikon DS-Fi1 camera and Nikon NIS-Elements D software (F 4.00.06).
Measurement of Circulating Serum Autoantibody Reactivity by ELISA
Reactivity of circulating antibodies from rabbit serum with laminin-111 and -511 was tested by enzyme-linked immunosorbent assay (ELISA) assay according to published protocols, with minor modification.
In brief, 96-well plates (Greiner Bio-One) were coated overnight with 500 ng/well of native or heat-denatured (30 minutes, at 80°C) mouse laminin-111 (Life Technologies–Invitrogen) or recombinant human laminin-511 (Biolamina, Sundbyberg, Sweden) in 0.05 mol/L carbonate–bicarbonate buffer (pH 9.3). After blocking, the wells were incubated with 100-fold or serially diluted rabbit serum and bound IgG antibodies were subsequently detected with a 5000-fold diluted HRP-labeled chicken antibody (Novus Biologicals).
For detection of circulating antibodies from patient's serum, ELISA plates were coated with 500 ng/well of recombinant human or mouse laminin γ1 N-terminus and subsequently incubated with 50-fold diluted human serum. Bound antibodies were detected with a 5000-fold diluted HRP-labeled goat antibody (Abcam, Cambridge, UK). For enzymatic reaction, 50 μL of orthophenylenediamine (Sigma-Aldrich) solution in water and 0.1% (v/v) H2O2 was added to each well. The reaction was stopped after 10 minutes with 50 μL/well of 0.5 mol/L H2SO4. Absorbance at 490 nm was read using a microplate reader (Labsystems Multiskan Multisoft; Thermo Fisher Scientific).
ROS Measurement by Luminol Chemiluminescence
Measurement of ROS production was performed in a luminol assay according to published protocols.
Plate-bound rabbit immune complexes were formed using the recombinant mLnγ1-Nterm or natural mouse laminin-111 and serum from rabbits immunized against the laminin γ1 fragments. In brief, 96-well microtiter plates were coated with 10 μg of mouse laminin-111 (Life Technologies–Invitrogen) or 1 μg of recombinant mouse mLnγ1-Nterm in 0.1 mol/L carbonate buffer (pH 9.3). The wells were washed three times with PBS (pH 7.2) containing 0.05% Tween 20, after each step. After blocking for 1 hour with PBS–Tween 20 containing 1% BSA (fraction V; Sigma-Aldrich), wells were incubated for 2 hours with several dilutions of rabbit serum against mLnγ1-Nterm (100-, 200-, 400-fold diluted) or a mix of rabbit sera targeting GST-mLnγ1-Cterm-1 and GST-mLnγ1-Cterm-3 (25-, 50-, and 100-fold diluted). Human leukocytes were isolated from the peripheral blood of healthy donors. After 3% dextran sedimentation, the erythrocytes were lysed using a hypotonic solution of 0.2% NaCl. The leukocytes were further washed in Dulbecco's modified Eagle's medium (Lonza) without supplements and were resuspended at a final density of 3 × 106 cells/mL. Cell viability was tested with Trypan Blue, and only preparations with viability greater than 95% were used. ROS production was measured using luminol chemiluminescence. The leukocytes were preincubated for 5 minutes at 37°C with 150 μmol/L luminol (Sigma-Aldrich), and then 100 μL of the cell suspension was added to each well. All procedures were performed in the dark. Chemiluminescence intensity was recorded continuously for 1 hour. Kinetic measurement of ROS production in leukocytes stimulated with 500 nmol/L phorbol 12-myristate 13-acetate (Sigma-Aldrich) was performed as control.
In Silico Analysis
Amino acid sequences of murine and human laminin γ1 were retrieved from GenBank (http://www.ncbi.nlm.nih.gov/nuccore, accession numbers NP_034813.2 and NM_002284.3), and sequence alignment was performed using an algorithm developed by Smith and Waterman.
In silico prediction of linear B-cell epitopes was performed using publicly available software (http://tools.immuneepitope.org/tools/bcell, last accessed June 26, 2012) that computes an antigenic score based on a prediction algorithm developed by Kolaskar and Tongaonkar,
(http://www.imtech.res.in/raghava/cbtope, last accessed October 22, 2012) was used for prediction of conformational B-cell epitopes of a protein from its amino acid sequence. The sequence corresponding to the short arm is longer than the one for the α-helical C-terminus domain, and therefore the number of epitopes was calculated per 100 amino acid residues.
Statistical Analysis
Statistical analysis was performed using GraphPad Prism software version 6 (GraphPad Software, San Diego, CA). Statistical differences were calculated using two-tailed, paired Student's t-test. Data were considered significantly different if P < 0.05.
Results
Expression of Recombinant Laminin γ1 and Nidogen 1 Forms
The mLnγ1-Nterm fragment (Figure 1 and Supplemental Figure S1) and nidogen 1 (Supplemental Figure S2) were recombinantly expressed in 293-EBNA cells, using a modified eukaryotic expression vector pCEP-Pu (Table 1). The 2×StrepII/FLAG-fusion recombinant proteins were purified by Strep-Tactin Sepharose affinity chromatography. The tag of the recombinant mLnγ1-Nterm fragment was subsequently removed by thrombin digestion. The cDNA sequences coding for the three C-terminal fragments of murine laminin γ1 were cloned into prokaryotic expression vector pGEX-6P-1 and expressed in E. coli (Table 1 and Figure 1A). The GST-fusion mLnγ1-Cterm-1, -2, and -3 recombinant fragments were purified by glutathione-affinity chromatography. With SDS-PAGE separation, all proteins migrated consistently with their calculated molecular masses (Figure 1B and Supplemental Figure S2B). With immunoblotting, the GST-tagged recombinant forms of laminin γ1 C-terminus were stained by an anti-GST antibody, similar to the GST used as control (Figure 1C). The 2×StrepII/FLAG-fusion nidogen 1 (Supplemental Figure S2C) was detected by an anti-StrepII antibody, and the generated polyclonal rabbit antibodies were used for detection of the 2×StrepII/FLAG-fusion mLnγ1-Nterm and tag-free mLnγ1-Nterm (Figure 1C).
Figure 1Characterization of recombinant murine laminin γ1 fragments. A: Laminin γ1 consists of a short arm at its N-terminus, containing globular structures (IV, VI) alternating with rod-like domains composed of epidermal growth factor–like repeats (III, V), and an α-helical domain at its C-terminus (I, II) that forms the coiled-coil domain by binding other laminin chains. The sequence corresponding to mLnγ1-Nterm fragment was cloned into a modified pCEP-Pu vector and expressed in 293-EBNA cells; the fragments of the C-terminus corresponding to the α-helical domain were cloned into pGEX-6P-1 and expressed in E. coli. Amino acid residue numbers are indicated above the fragments. B: When separated by 10% SDS-PAGE, recombinant GST-tagged mLnγ1-Cterm-1, -2, and -3 fragments (lanes 2, 3, and 4, respectively) and recombinant mLnγ1-Nterm (lane 5) migrated consistently to their calculated molecular masses. Migration positions of the molecular weight marker (lane 1) are indicated on the left. C: Recombinant mLnγ1 fragments were electrophoretically transferred to nitrocellulose. GST-tagged recombinant mLnγ1-Cterm-1, -2, and -3 fragments (lanes 1, 2, and 3, respectively) were detected with goat antibody specific to GST. GST (lane 4) was used as control. The 2×StrepII/FLAG-fusion mLnγ1-Nterm fragment (lane 5) and the tag-free mLnγ1-Nterm fragment (lane 6) were detected with rabbit antibodies specific to mLnγ1-Nterm. Migration positions of molecular weight markers are indicated on the left.
In Silico Analysis of Antigenicity of Murine Laminin γ1
For the murine and human forms of laminin γ1 (with a length of 1607 and 1609 amino acids, respectively), alignment analysis using the Smith–Waterman algorithm indicated an identity of 92.9% and positivity of 96.7%. The epitopes targeted by antibodies from anti–laminin γ1 pemphigoid patients have been suggested to be located mainly within the C-terminus of the protein.
By in silico prediction analysis, B-cell epitopes are distributed throughout the entire mouse laminin γ1, with an average antigenic score slightly higher for the short arm, compared with the C-terminus forming the coiled-coil domain (Supplemental Table S1 and Supplemental Figure S3). Kolaskar and Tongaonkar antigenicity analysis predicted 42 linear epitopes for the N-terminus sequence (aa 1–1028) and 18 for the C-terminus (aa 1029–1607).
The sequence for the nidogen binding site (domain III 3–5, aa 769–932) contains six linear antigenic determinants and five probable conformational epitopes.
Characterization of Reactivity and Specificity of Antibodies Generated against Laminin γ1
The specificity of antibodies from immune rabbit serum was determined by indirect IF microscopy on normal mouse skin and ELISA using native and heat-denatured mouse laminin-111. Serum from rabbits immunized with the recombinant mLnγ1-Nterm exhibited linear binding at the DEJ and also staining of dermal blood vessels (Figure 2A), similar to the monoclonal rat anti–laminin γ1 antibody used as control (Figure 2E). Circulating antibodies from rabbits immunized against the three fragments of C-terminus laminin γ1 chain (GST-mLnγ1-Cterm-1, -2, and -3) weakly stained the DEJ and dermal blood vessels (Figure 2, B–D). By ELISA, rabbit serum raised against C- or N-terminal fragments of laminin γ1 exhibited different degrees of reactivity against mouse laminin-111 (Figure 2F).
Figure 2Characterization of rabbit antibodies specific to laminin γ1 fragments by indirect IF microscopy and ELISA. A–E: Cryosections of murine skin were incubated with serum from rabbits immunized with mLnγ1-Nterm (A), GST-mLnγ1-Cterm-1 (B), GST-mLnγ1-Cterm-2 (C), or GST-mLnγ1-Cterm-3 (D), followed by staining with anti-rabbit IgG Alexa Fluor 488–labeled secondary antibody (E). Skin basement membrane and dermal capillaries are indicated by arrows. F: Serial dilutions of rabbit serum were tested by ELISA using murine laminin-111 as a substrate. Rabbit serum raised against C- or N-terminal fragments of laminin γ1 differed in degree of reactivity against mouse laminin-111. Data are expressed as means ± SD. Experiments were conducted in duplicate. Original magnification, ×400.
Murine and human laminin γ1 are highly homologous and share predicted B-cell epitopes to a large extent (Supplemental Table S2). We used indirect IF to further test the cross-reactivity of the newly generated antibodies against murine laminin γ1 with the human protein on cryosections of normal human skin. With one exception (rabbit serum raised against mLnγ1-Cterm-2), all the sera exhibited linear binding in the human epidermal basement membrane and also in the walls of dermal blood vessels (Figure 3, A–D).
Figure 3Cross-reactivity of rabbit antibodies specific to murine laminin γ1 fragments with the human protein by indirect IF microscopy. A–D: Cryosections of human skin were incubated with serum from rabbits immunized with mLnγ1-Nterm (A), GST-mLnγ1-Cterm-1 (B), GST-mLnγ1-Cterm-2 (C), or GST-mLnγ1-Cterm-3 (D). After a PBS washing, the respective sections were incubated with anti-rabbit IgG Alexa Fluor 488–labeled secondary antibody. Skin basement membrane and dermal capillaries are indicated by arrows. Original magnification, ×400.
Laminin γ1 C-Terminus-Specific Antibodies React with Cryptic Epitopes
The C-terminus of laminin γ1 consists of an α-helix that trimerizes into a coiled coil, and we therefore assessed the availability of the epitopes within the C-terminus for rabbit antibodies by comparing reactivity with the native and denatured forms of laminin. Indirect IF microscopy of murine skin cryosections revealed that the serum raised against laminin γ1 C-terminal fragments stained stronger the skin treated with acid urea for unmasking epitopes than the native skin (Supplemental Figure S4, A and B). Similarly, ELISA revealed a higher reactivity with heat-denatured laminin-111 or -511 than with the native proteins (Supplemental Figure S4, C and D).
Laminin γ1 N-terminus–Specific Antibodies Bind in Mouse Skin
A prerequisite for developing a mouse model for antibody-induced skin disease is binding of antibodies in the skin. We therefore assessed whether the laminin γ1 N-terminus–specific antibodies are able to bind in the mouse skin both in an ex vivo model of skin organ culture and in vivo. Incubation of normal mouse skin explants with anti–laminin γ1 N-terminus–specific rabbit serum led to the deposition of rabbit IgG at the DEJ and in the walls of dermal blood vessels (Figure 4A). The same binding pattern was observed after injection of purified rabbit antibodies into normal mice (Figure 4C), but no binding was observed for normal rabbit IgG (Figure 4, B and D).
Figure 4Laminin γ1-Nterm–specific antibodies bind in murine skin explants and in vivo. A and C: After incubation of murine skin samples with anti–mLnγ1-Nterm rabbit serum, the antibodies stained the DEJ and dermal vessel walls (A). The same antibodies bound in a similar pattern in the skin after injection in mice (C). B and D: Normal rabbit serum (NRS) exhibited no binding in the skin samples (B) or in vivo (D). Skin basement membrane and dermal capillaries are indicated by arrows. Original magnification, ×400.
Laminin γ1 N-Terminus–Specific Antibodies Interfere With the Ligand Function of Nidogen 1 Binding of Laminin γ1
We further asked whether antibodies against laminin γ1 influence its binding to nidogen 1, a well-described ligand for laminins containing the γ1 chain. The interaction site is located within a single epidermal growth factor–like domain from the N-terminal part of laminin γ1. Indeed, the recombinant purified mLnγ1-Nterm fragment bound with high-affinity nidogen 1 in our solid-phase assay (Figure 5A). In this system, antibodies generated against mLnγ1-Nterm inhibited the binding of nidogen 1 to mLnγ1-Nterm in a concentration-dependent manner (Figure 5B). These antibodies also exhibited a concentration-dependent ability to disrupt binding of nidogen to mLnγ1-Nterm, although to a lesser extent than observed in the inhibition assay (Figure 5C).
Figure 5Binding of nidogen 1 to mouse laminin γ1 N-terminus and interference of specific antibodies with the binding. A: Serial dilutions of recombinant 2×StrepII/FLAG-fusion nidogen 1 were incubated with 500 ng/well of laminin γ1-Nterm or BSA coated onto 96-well microtiter plates. Binding was detected by an ELISA-based solid-phase assay with an antibody against StrepII tag. B: Recombinant mLnγ1-Nterm (500 ng/well) coated onto 96-well microtiter plates was incubated first with serial dilutions of rabbit serum anti–mLnγ1-Nterm (RSa-mLnγ1-Nterm) or the same dilutions of NRS, followed by incubation with 3.5 nmol/L mouse nidogen 1–2×StrepII/FLAG. The bound nidogen was detected with an anti-StrepII tag antibody. C: Recombinant mLnγ1-Nterm (500 ng/well) coated onto 96-well microtiter plates was incubated first with 3.5 nmol/L mouse nidogen 1–2×StrepII/FLAG. Subsequently, serial dilutions of rabbit serum anti–mLnγ1-Nterm or the same dilutions of NRS were added. Mouse nidogen 1 was detected with an anti-StrepII tag antibody. Data are expressed as means ± SD. Experiments were conducted in triplicate. OD, optical density.
Antibodies Specific to Laminin γ1 C-Terminus Do Not Interfere With Integrin-Mediated Adhesion to Laminin
The binding of laminin to integrins through the LG domains, which plays a major role in cell adhesion, is modulated by the C-terminal fragment of laminin γ1 chain.
The requirement of the glutamic acid residue at the third position from the carboxyl termini of the laminin gamma chains in integrin binding by laminins.
We therefore addressed the question of whether the antibodies targeting the C-terminus of laminin γ1 inhibit laminin–integrin interaction. In a solid-phase assay, a mix of the antibodies targeting laminin γ1 C-terminal fragments did not influence the binding of integrin α6β4 to laminin-511 (Figure 6A). Furthermore, we observed no significant inhibition of HaCaT cell adhesion on recombinant human laminin-511 in the presence of the same antibodies (Figure 6B).
Figure 6Interference of laminin γ1–specific antibodies on laminin–integrin interaction and cell adhesion. A: ELISA wells coated with 10 nmol/L laminin-511 were first incubated with a 25-fold diluted mix of rabbit antibodies against GST-mLnγ1-Cterm-1, -2, and -3 (RSa-mLnγ1-Cterm) or with NRS; subsequently, recombinant integrin α6β4 was added in the presence of 2 mmol/L MgCl2 and 1 mmol/L MnCl2, or 10 mmol/L EDTA. The bound integrin was detected by incubation with an HRP-labeled rabbit polyclonal anti-His antibody. OD readings obtained in the presence of EDTA were subtracted as background. Binding was expressed as a percentage relative to results obtained in the absence of rabbit antibodies. B: Recombinant laminin-511 or BSA (1 μg/well) was immobilized on microtiter plates and further incubated with a 25-fold diluted mix of rabbit antibodies against GST-mLnγ1-Cterm-1, -2, and -3 (RSa-mLnγ1-Cterm), with NRS, or with 50-fold diluted monoclonal antibody 4C7. HaCaT cells were then allowed to adhere to the wells for 60 minutes at 37°C in the presence of 2 mmol/L MgCl2 and 1 mmol/L MnCl2, followed by staining with crystal violet. Readings obtained for the nonspecific cell adhesion in the presence of BSA were subtracted. Cell adhesion was expressed as a percentage, relative to results obtained in the absence of rabbit antibodies. Data are expressed as means ± SD. Experiments were conducted in duplicate (A) or triplicate (B).
Laminin γ1–Specific Antibodies Activate Complement ex Vivo
To test the Fc-dependent pathogenic potential of antibodies against laminin γ1, we assessed the ability of skin-bound antibodies to activate complement, using an ex vivo complement fixation assay. Lnγ1-Nterm–specific rabbit antibodies bound to normal mouse skin were able to activate complement in the walls of dermal blood vessels and, to a lesser extent, at the DEJ (Figure 7A). In contrast, no complement activation ability was detected for the antibodies specific to any of the three GST-mLnγ1-Cterm fragments (Figure 7, B–D). Antibodies raised against collagen VII, used as a positive control, exhibited linear fixation of complement at the epidermal basement membrane (Figure 7E). No complement activation was observed in samples incubated with normal rabbit serum (Figure 7F).
Figure 7Complement fixation ability of laminin γ1–specific rabbit serum. Cryosections of normal mouse skin were incubated with rabbit serum; fresh normal human serum was added, as a source of complement. Human C3 was detected with a fluorescein isothiocyanate–labeled antibody. The sections were counterstained with DAPI. A: Antibodies against mLnγ1-Nterm were able to activate the human complement in dermal blood vessels and, to a lesser extent, linearly along the DEJ. B–D: In contrast, no complement activation ability was detected for antibodies specific to GST-mLnγ1-Cterm-1, -2, or -3 fragments. E: Serum from a rabbit immunized against GST-mCVII-2 strongly fixed the complement linearly along the DEJ. F: No complement activation was observed on the cryosections incubated with NRS. Skin basement membrane and dermal capillaries are indicated by arrows. Original magnification, ×400.
Laminin γ1–Antibody Complexes Elicit ROS Production in Granulocytes
In a further set of experiments, we tested whether laminin γ1–specific antibodies from immune complexes are able to elicit ROS production in granulocytes isolated from healthy human donors. Immobilized immune complexes formed by antibodies specific to mLnγ1-Nterm bound to mouse laminin-111 or recombinant mLnγ1-Nterm coated onto ELISA plates were able to activate human granulocytes, thereby inducing the release of ROS, as detected by chemiluminescence assay. In contrast, rabbit sera raised against the GST-mLnγ1-Cterm fragments failed to induce ROS production in the same assay (Figure 8).
Figure 8Antibody-elicited production of ROS measured by chemiluminescence. Immobilized immune complexes were formed by rabbit antibodies with mouse laminin-111 or mLnγ1-Nterm coated onto ELISA plates. Granulocytes isolated from healthy human donors were preincubated with luminol and subsequently added to the wells. Chemiluminescence intensity was recorded continuously for 1 hour. Immune complexes formed by rabbit antibodies specific to mLnγ1-Nterm and recombinant mLnγ1-Nterm (mLnγ1-Nterm+RSa-mLnγ1-Nterm), or with laminin-111 (ln111+RSa-mLnγ1-Nterm), elicited ROS production, whereas no release of ROS was detected in the laminin-111–coated wells incubated with a mix of rabbit antibodies targeting GST-mLnγ1-Cterm-1 and -3 (ln111+RSa-mLnγ1-Cterm). No activation of granulocytes was observed in wells incubated with NRS (mLnγ1-Nterm +NRS, ln111+NRS). Data are expressed as means ± SD. Experiments were conducted in duplicate. RLU, relative luminescence units.
Clinical and laboratory findings in patients suggest that, to fully unfold their pathogenic potential, autoantibodies against laminin γ1 may need to recruit and activate inflammatory cells, in addition to their complement-fixing capacity. We therefore investigated the ability of laminin γ1–specific antibodies to induce granulocyte-dependent dermal–epidermal separation in an ex vivo assay, using sections of normal skin incubated with antibodies and normal leukocytes. When coincubated with leukocytes purified from healthy donors, antibodies specific to the N-terminus of laminin γ1 bound at the murine DEJ induced subepidermal splits (Figure 9A), but to a lesser extent than the antibodies against collagen VII used as control (Figure 9E). In contrast, no blister formation was observed in sections incubated with antibodies against the C-terminal fragments (Figure 9, B–D).
Figure 9Granulocyte activation ability of laminin γ1–specific rabbit serum. Frozen sections of mouse skin were incubated with rabbit serum and, after washing, leukocytes isolated from healthy human donors were added. A–D: Rabbit antibodies specific to mLnγ1-Nterm bound to normal mouse skin recruited and activated granulocytes, inducing subepidermal splits (A). In contrast, no blister formation was observed on sections incubated with rabbit serum against GST-mLnγ1-Cterm-1 (B), GST-mLnγ1-Cterm-2 (C), or GST-mLnγ1-Cterm-3 (D). E: Percentage of dermal–epidermal separation (DES). Data are expressed as means ± SEM. Data represent multiple experiments [cryosections incubated with rabbit serum raised against GST-mCVII-1 (RSa-mCVII-1, n = 18); rabbit antibody against mLnγ1-Nterm (RIgGa-mLnγ1-Nterm, n = 18), GST-mLnγ1-Cterm-1 (RIgGa-mLnγ1-Cterm-1, n = 16), GST-mLnγ1-Cterm-2 (RIgGa-mLnγ1-Cterm-2, n = 16), or GST-mLnγ1-Cterm-3 (RIgGa-mLnγ1-Cterm-3, n = 16); and NRS (n = 15).] ∗∗P < 0.01.
Laminin γ1 N-Terminus–Reactive Antibodies in Patients with Anti–Laminin γ1 Pemphigoid and Collagenoses
Although most of the antibodies in anti–laminin γ1 pemphigoid patients react with the C-terminus of laminin γ1, reactivity with epitopes within the N-terminus is also reported.
We identified serum from one patient with significant reactivity in ELISA to both human (Supplemental Figure S5) and mouse recombinant laminin γ1 N-terminus. In the laminin–nidogen binding assay, the serum did not interfere with the binding of nidogen to laminin (data not shown). Circulating antibodies targeting the N-terminus of laminin γ1 were detected by ELISA in 3/19 (16%) patients with cutaneous lupus erythematosus (2 discoid CLE; 1 subacute CLE) and in 2/14 (8%) patients with scleroderma (1 systemic sclerosis; 1 localized scleroderma) (Supplemental Figure S5).
Discussion
Laminin γ1 has been recently described as the autoantigen in anti–laminin γ1 pemphigoid, a rare autoimmune bullous disease of the skin. In addition, autoantibodies against laminin-111 with unclear pathogenic relevance have been reported in patients with lupus erythematosus,
A retrospective study of antibodies against basement membrane antigens (type IV collagen and laminin) in patients with primary and secondary Raynaud's phenomenon.
Anti–laminin γ1 pemphigoid is characterized by subepidermal blistering, a neutrophil-dominated inflammatory infiltrate in the upper dermis, and tissue-bound and circulating serum antibodies binding the skin basement membrane. Interestingly, indirect IF microscopy revealed that the antibodies label only the DEJ; no staining of the vessel basement membrane was observed. Furthermore, despite the widespread distribution of laminin γ1, the lesions are restricted to skin and mucous membranes.
In skin, laminin γ1 is incorporated in laminin-321/311 and laminin-511; the latter is well expressed in hair follicles, but also in dermal vessels, in addition to the DEJ. It has been proposed that laminin γ1–specific antibodies mediate blistering by interfering with laminin–integrin interaction.
Because their pathogenic potential had not yet been demonstrated, in the present study we investigated the pathogenic potential of laminin γ1–specific antibodies.
Wet-laboratory epitope mapping studies indicate a distribution of antigenic determinants throughout the entire laminin γ1. Although epitopes recognized by antibodies from anti–laminin γ1 pemphigoid patients cluster within the C-terminus of laminin γ1 chain, few epitopes were shown to reside within the N-terminal domain.
Our in silico analysis indicated that linear and conformational antigenic determinants are distributed throughout the entire murine laminin γ1, with the short arm harboring more epitopes, compared with the C-terminus forming the coiled-coil domain. We observed reactivity with the N-terminus of laminin γ1 in 1/7 (14%) patients with anti–laminin γ1 pemphigoid and in 5/52 (10%) patients with connective tissue diseases (CLE, systemic sclerosis, and localized scleroderma).
Because the relevant contribution to tissue damage of the antigenic determinants within the N-terminal laminin γ1 had not been investigated previously, we addressed the pathogenic potential of antibodies directed against epitopes located within both the N- and the C-terminus of laminin γ1. The newly generated laminin γ1–specific antibodies exhibited different degrees of reactivity with the skin basement membrane and dermal blood vessels (as indicated by indirect IF microscopy) and with laminin-111 (as indicated by ELISA). A weaker binding in the skin was observed for the antibodies raised against the C-terminal domain of laminin γ1, perhaps because of lower recognition of the native conformation of laminins (detected also by ELISA). As suggested by our in silico antigenicity analysis of human and murine laminin γ1, the antibodies cross-reacted with the human antigen, as indicated by both indirect IF microscopy and ELISA. Subsequently, the laminin γ1 fragment-specific antibodies were characterized with regard to their ability to mediate tissue damage through Fc-dependent and Fc-independent pathways.
By interacting with other components of the basement membrane, laminins play both an architectural and a functional role. We therefore addressed the ability of the antibodies to directly interfere with the ligand function of laminin. Nidogen 1 and 2 have been shown to connect the laminin network with the collagen IV meshwork in basement membranes, along with perlecan. The high-affinity binding site is located within the domain III 3–5 of the short arm of laminin γ1,
reported inhibition of basement membrane formation by a nidogen-binding recombinant γ1 chain fragment in skin organotypic cocultures. After treatment of the coculture with this fragment, they observed an increased tendency for splitting at the tissue interface. Furthermore, early treatment with the recombinant fragment diminished or abolished deposition of nidogen and laminin γ1, whereas a delayed treatment (after the basement membrane was formed) led to an almost complete, but reversible, loss of nidogen and laminin γ1. Consistent with such report, in the present study the antibodies raised against the N-terminus of laminin γ1 were able to inhibit and disrupt the interaction of nidogen with laminin. Taken together, the findings suggest that this Fc-independent potential of anti–laminin γ1 short-arm antibodies to induce tissue damage might unfold during basement membrane remodeling in both physiological and pathological processes. In our small cohort of patients with anti–laminin γ1 pemphigoid, we found only one serum with significant reactivity with human and murine laminin γ1 N-terminus. This serum did not interfere with the binding of nidogen to laminin. A possible explanation is the low reactivity with epitopes located within the nidogen binding site.
Laminins have a crucial functional role through cell receptor-mediated mechanisms. By interacting with skin integrins such as α6β4 and α3β1, laminin-511 is a potent adhesive ligand for keratinocytes.
Ligand-binding specificities of laminin-binding integrins: a comprehensive survey of laminin-integrin interactions using recombinant alpha3beta1, alpha6beta1, alpha7beta1 and alpha6beta4 integrins.
The interaction site maps to the C-terminal globular domain of laminin α chain, but binding of the integrin requires heterotrimerization with β and γ chains. Furthermore, the C-terminal region of laminin γ1 is important in modulating integrin–laminin-511 interaction through the third glutamic acid residue, although not directly binding to integrin.
The requirement of the glutamic acid residue at the third position from the carboxyl termini of the laminin gamma chains in integrin binding by laminins.
In one study most antibodies from anti–laminin γ1 pemphigoid patients reacted with the 246-amino-acid C-terminus of laminin γ1 chain, and it has been speculated that the mechanism of blister formation might rely on their interference with laminin–integrin binding.
In the present study, however, the laminin γ1 C-terminus–specific antibodies inhibited neither keratinocyte attachment to laminin-511–coated wells, nor the laminin-511–integrin α6β4 interaction.
A well-described mechanism of tissue damage by antibodies is their ability to trigger activation of the innate immune system, including the complement cascade and inflammatory cells. In the present study, direct IF microscopy on perilesional skin of anti–laminin γ1 pemphigoid patients revealed deposition of C3 at the DEJ. Furthermore, human and animal anti–laminin-111 antibodies are able to fix complement in vitro.
Our present findings show that antibodies specific to laminin γ1 are able to activate complement ex vivo, to different extents depending on their molecular fine specificity. In the complement fixation test, the complement was activated mainly in dermal blood vessels, and to a lesser extent at the DEJ of mouse skin, suggesting a possible involvement in vascular pathology. Of note, anti–laminin-111 antibodies have been reported in vasculitis patients.
Both the lack of antibody deposition in blood vessels from most anti–laminin γ1 pemphigoid patients and the skin-specific pathology remain puzzling. Although Dainichi et al
reported one patient with evident reactivity with laminin from the skin blood vessels, most of the sera reacted with dermal rather than blood-vessel laminin. As Dainichi et al
speculated, tissue-specific differences in post-translational modifications of the γ1 chain, such as N-glycosylation or laminin trimer conformation, might account for differing antibody reactivity in these patients. Although binding of laminin γ1 N-terminus–specific antibodies in blood vessels might not be relevant for pathogenesis of anti–laminin γ1 pemphigoid, it might have relevance for other autoimmune disorders, including vasculitis and connective tissue diseases, and its contribution to tissue damage should be addressed in further studies.
Inflammatory infiltrates dominated by neutrophils are a common finding in anti–laminin γ1 pemphigoid patients, and we therefore assessed the ability of the laminin γ1–specific antibodies to activate leukocytes. Immune complexes formed by N-terminus–specific antibodies with the targeted antigen elicited ROS production of granulocytes, as shown by luminol assay. In addition, skin-bound antibodies were to some extent able to induce dermal–epidermal separation by activating human leukocytes. The tissue injury might therefore be triggered by antibody-mediated activation of leukocytes and subsequent release of proteases and/or NADPH oxidase, as shown in several inflammatory skin blistering diseases.
antibodies targeting laminin γ1 C-terminal fragments were not able to induce ROS production, nor did they induce split formation in skin cryosections.
Both the antibody-dependent dermal–epidermal separation assay and the complement fixation test require an initial binding step of antibodies in the skin cryosections. The differences in the Fc-dependent pathogenic potential of N- and C-terminus–specific antibodies in both the cryosection assay and the complement binding test may rely on their different affinity for the native skin laminins. However, two recent studies showed that purified laminin γ1 C-terminus–specific patient or rabbit IgG could not induce dermal–epidermal separation ex vivo, nor complement activation and blistering in vivo, although it bound in the skin basement membrane.
Taken together, these results suggest that other factors, in addition to poor binding, might account for the lack of pathogenicity of these antibodies.
Human anti–laminin γ1 antibodies have low availability and are therefore not well suited for use in animal models, which normally require injections of high amounts of specific antibodies over a long period of time. Antibodies of rabbit origin are an appropriate alternative for animal model studies. Laminin γ1 N-terminus–specific antibodies have pathogenic potential ex vivo and, because they bind at the DEJ and in dermal blood vessels in murine skin, they can be used to establish an animal model of laminin autoimmunity.
In conclusion, the present results demonstrate a pathogenic potential of anti–laminin γ1 antibodies ex vivo. The epitopes relevant to tissue damage are located within the N-terminal part of the γ1 chain, and both Fc-mediated and Fc-independent pathways appear to be involved. Thus, antibodies targeting laminin γ1 N-terminus strongly bind in murine skin and are able to inhibit and disrupt the binding of nidogen 1, and also to activate leukocytes and the complement. These results extend our understanding of the pathogenesis of diseases associated with laminin autoimmunity and should greatly facilitate the development of appropriate diagnostic tools and therapeutic strategies.
Enzymatic deglycosylation of recombinant 2×StrepII/FLAG-tagged mouse laminin γ1 N-terminus. A and B: The glycosylated (−) and PNG-ase treated (+) mLnγ1-Nterm were analyzed by SDS-PAGE (A) and detected by immunoblotting with rabbit antibodies specific to mLnγ1-Nterm (B). Migration positions of the molecular weight marker are indicated on the left.
Characterization of recombinant murine nidogen 1. A: Nidogen 1 consists of a polypeptide chain that forms three globular domains (G1, G2, G3) separated by a linker region between G1 and G2, and a rod-like domain between G2 and G3. The sequence corresponding to the full-length murine nidogen 1 was cloned into a modified pCEP-Pu vector and expressed as 2×StrepII/FLAG-fusion protein in 293-EBNA cells. B: By SDS-PAGE analysis the recombinant nidogen 1 migrated consistently to the calculated molecular mass. C: By immunoblot analysis, the recombinant 2×StrepII/FLAG-tagged murine nidogen 1 was stained by an anti-StrepII antibody. Migration positions of the molecular weight marker are indicated on the left.
In silico analysis of B-cell epitopes of murine laminin γ1. Linear and conformational antigenic determinants of laminin γ1 were analyzed based on the algorithm developed by Kolaskar and Tongaonkar
A and B: Cryosections of murine normal (A) or urea-denatured skin (B) were incubated with serum from rabbits immunized with mLnγ1-Cterm-3. After a PBS washing, the sections were incubated with anti-rabbit IgG Alexa Fluor 488–labeled secondary antibody. Skin basement membrane and dermal capillaries are indicated by arrows. C and D: Levels of circulating rabbit IgG were measured by ELISA using native and heat-denatured murine laminin-111 (C) or human laminin-511 (D). OD readings of serum samples from immunized rabbits and from NRS. Data are expressed as means ± SD. Experiments were conducted in duplicate. Original magnification, ×400.
Reactivity of patient serum by ELISA with laminin γ1 N-terminus expressed as optical density (OD) readings of reactivity with recombinant human laminin γ1 N-terminus of serum from CLE, dermatomyositis (DM), scleroderma (SD), anti–laminin γ1 pemphigoid (p200) patients and from healthy volunteers (NHS). The horizontal dashed line represents the cutoff of the assay.
A retrospective study of antibodies against basement membrane antigens (type IV collagen and laminin) in patients with primary and secondary Raynaud's phenomenon.
The requirement of the glutamic acid residue at the third position from the carboxyl termini of the laminin gamma chains in integrin binding by laminins.
Ligand-binding specificities of laminin-binding integrins: a comprehensive survey of laminin-integrin interactions using recombinant alpha3beta1, alpha6beta1, alpha7beta1 and alpha6beta4 integrins.
Alport alloantibodies but not Goodpasture autoantibodies induce murine glomerulonephritis: protection by quinary crosslinks locking cryptic α3(IV) collagen autoepitopes in vivo.
Supported by the German Research Foundation (Deutsche Forschungsgemeinschaft), grants SI-1281/2-1 (C.S.), SFB 829 A2 (M.K.), SI-1281/4-1 (C.S.), and BIOSS-B13 (C.S.), and by the Medical Faculty of the University of Freiburg (C.S.).
00731-1&id=gr1.jpg){kind=link}
00731-1&id=gr2.jpg){kind=link}
00731-1&id=gr3.jpg){kind=link}
00731-1&id=gr4.jpg){kind=link}
00731-1&id=gr5.jpg){kind=link}
00731-1&id=gr6.jpg){kind=link}
00731-1&id=gr7.jpg){kind=link}
00731-1&id=gr8.jpg){kind=link}
00731-1&id=gr9.jpg){kind=link}