Plasmin Plays a Role in the In Vitro Generation of the Linear IgA Dermatosis Antigen LADB97(1)
In bullous pemphigoid (BP), autoantibodies target immunodominant epitopes mainly located within the juxtamembranous NC16a domain of collagen XVII. In linear IgA dermatosis (LAD), a rare subepidermal bullous autoimmune dermatosis with clinical and immunopathological features overlapping with BP, the ectodomain of collagen XVII is specifically targeted by IgA-autoantibodies. The IgA-autoantibodies also recognize a 97kDa antigen, designated LABD97, that represents a truncated form of the collagen XVII ectodomain.
Serine proteases and metalloproteinases, such as neutrophil elastase, plasmin, plasminogen activators, MMP-2 and MMP-9, have been detected in BP blister fluids and in perilesional skin. Plasminogen is mainly produced in the liver. After wounding or blistering of the skin, plasminogen will diffuse from the bloodstream to the skin and will be activated by plasminogen activators or other serine proteases.
Immunoblotting of these recombinant NC16a fragments with the hybridomas revealed two particularly interesting clones, termed NC16a-1 and NC16a-3. Although NC16a-1 recognized the N-terminal NC16a fragment, NC16a-3 bound specifically to C-NC16a. Both mAbs were of the IgG1 isotype and reactive with 180kDa full-length collagen XVII in keratinocyte extracts and with the 120kDa shed ectodomain in conditioned medium. However, mAb NC16a-3 reacted significantly stronger with the shed 120kDa ectodomain than with full-length collagen XVII. For more precise epitope mapping, a pepspot assay with 35 overlapping 13-meric synthetic peptides, covering the entire NC16a domain, was used. NC16a-1 specifically bound to pepspots 13 (AA 513–525 of collagen XVII). Subsequently, after pH shift stripping, the assay was reprobed with NC16a-3, which reacted specifically with pepspot 29 (corresponding to AA 545–557).
In about 60% of the BP blister fluids, but only in 25% of the suction blister fluids, an additional 97kDa protein was precipitated in large quantities by mAb NC16a-1. Immunoprecipitation with mAb NC16a-1 yielded increasing amounts of the 97kDa fragment that corresponds to the C-terminally truncated form of the collagen XVII ectodomain, which was initially designated LABD97.
The purified ectodomain of collagen XVII was treated with different human serine proteases, including plasmin, neutrophil elastase, trypsin, and chymotrypsin. Of these, only plasmin produced a 97kDa fragment. Incubation of HaCaT keratinocytes with human plasmin and trypsin for 2hours resulted in the production of a 97kDa fragment, although trypsin generated additional smaller fragments. Further comparison of plasmin and trypsin treatments of HaCaT cells for over 12 hours revealed that only plasmin generated a stable 97kDa collagen XVII fragment. The 97kDa fragment was also generated in the presence of ADAM inhibitors, which indicates an ADAM-independent cleavage of collagen XVII within the NC16a domain.
The NC16a domain contains 15 potential plasmin cleavage sites. In immunoblots analysis, two different LAD sera intense bound to the 97kDa ectodomain indicating that the plasmin-cleaved 97kDa fragment most likely corresponds to the autoantigen LABD97. Using the polyclonal rabbit-anti-NC16a antibody, the 97kDa fragment from BP blister fluid and plasmin-derived LABD97 from HaCaT medium co-migrate.Both mAbs NC16a-1 and NC16a-3 recognized the fragment, demonstrating that the plasmin cleavage occurred N-terminally of the previously published LABD97 terminus at AA531.
As collagen XVII has only one actively used N-glycosylation site at position 1,421 close to the distal C-terminus, with peptide N-glycosidase F (PNGase F) digestion, a mobility shift was seen when the 120kDa ectodomain was used as a substrate, but not with the 97kDa fragment, further confirming that the C-terminus of collagen XVII is not contained in LABD97. Taken together, our immunomapping and digestion experiments place the N-terminus of plasmin-derived LABD97 at approximately AA 515 and the C-terminus N-terminally of AA 1,421. Heparin-affinity chromatography revealed that the 120kDa ectodomain eluted at 0.56M NaCl, whereas plasmin-derived LABD97 eluted already at 0.39M NaCl, indicating a lower heparin bonding capacity.
Several N-termini are produced by plasmin cleavage, as the NC16a domain contains 15 potential plasmin cleavage sites (6 Lys-X and 9 Arg-X peptide bonds) that are preferentially clustered in four distinct regions. In addition, concomitant ADAM cleavage may contribute to the diversity of generated N-termini. Immunoblot analysis with mAbs predicted the N-terminus at approximately AA 515, and therefore clearly upstream of the previously described N-terminus at Ala531, which does not represent a potential plasmin cleavage site. Binding to heparin is mediated through clusters of basic AAs, that is arginine, lysine and histidine. In particular, clustered arginines seem to be essential for heparin binding. Within the NC16a domain, there are especially within the distal C-terminus in the Col-1 domain as a triple-helical cluster (Gly1468 to Lys1479) and as a cluster of five arginines (Arg1487 to Arg1491).
Cell protein extract preparation: For protein extractions, cell layer and media were processed separately. In addition to 1mM Pefabloc, 4mM EDTA, 10μlml-1 protease inhibitor cocktail set III, 14μgml-1 chymostatin and 10mM 1,10-ortho-phenanthroline were added to the extraction buffer. The cells were incubated in lysis buffer on ice for 10minutes, scraped off and centrifuged at 14,000 × g for 30minutes at 4°C.
Heparin-affinity chromatography: The culture media of normal and plasmin-treated keratinocytes were precipitated with 45% ammonium sulfate and dialysed against 20mM Tris-HCl, pH 7.4 at 4°C overnight. After clearing by centrifugation with 13,000 × g for 10minutes, the supernatant was analyzed by HPLC with a HiTrap Heparin HP column. The bound ectodomain was eluted with a gradient from 0 to 1M NaCl in 20mM Tris-HCl, pH 7.2 at a flow rate of 250μlminute-1.