Combination of molecular mimicry and aberrant autoantigen expression is important for development of anti-Fas ligand autoantibodies in patients with systemic lupus erythematosus

Abstract

We have reported previously that circulating anti-Fas ligand (FasL) autoantibodies able to inhibit Fas/FasL-mediated apoptosis were present in patients with systemic lupus erythematosus (SLE). In the present study, we describe the epitopes recognized by these anti-FasL autoantibodies. Rabbit antihuman antibody, raised against a FasL fragment consisting of amino acids (aa) 103-179 (fragment 2.0), inhibited Fas/FasL-mediated apoptosis, whereas an antibody against a FasL aa 103-146 fragment (fragment 1.0) did not. This suggested that an epitope around aa 146-179 was important for Fas/FasL interaction. Epitope mapping of anti-FasL autoantibodies using deletion mutants indicated that the epitope was located around aa 163-179. Three-dimensional molecular modelling of the Fas/FasL complex revealed that the aa 162-169 region was located on the outermost side of FasL, which suggested that the anti-FasL autoantibody would easily have access to the epitope. FasL point mutants involving aa positions 162-169 resulted in complete loss of apoptosis-inducing capability, which suggested that the aa 162-169 region was important for Fas/FasL interaction. A synthetic FasL peptide consisting of aa 161-170 blocked the binding of anti-FasL autoantibodies to FasL fragment 2.0 (aa 103-179). The FasL aa 161-170 sequence was found to be highly homologous with aa sequences from several infectious agents. Synthetic peptides derived from some of these microorganisms cross-reacted with the epitope recognized by the autoantibodies, suggesting that several foreign infectious agent-derived proteins may share an epitope with human FasL. As lymphocytes from SLE patients aberrartly expressed FasL, it is possible that infection by one of several infectious agents may trigger cross-reactive antibody responses, after which aberrantly expressed endogenous FasL might induce the shift from a cross-reactive response to an authentic autoimmune response. Therefore, a combination of molecular mimicry and aberrant autoantigen expression may be important for the development of anti-FasL autoantibodies in SLE patients.

Fig. 1

Schematic representation of wild-type FasL and FasL deletion-mutant proteins. FasL deletion-mutant proteins were produced in E. coli using the pQE30 bacterial expression vector. The aa numbers in each mutant FasL are indicated. MW represents the molecular weight (kDa) of the mutant FasL proteins tagged with 6 × histidine.

Fig. 2

Epitope mapping of the anti-FasL autoantibodies in SLE patients. Recombinant human FasL fragments 1, 2, 3, 4 and 5 were produced and purified using Ni-NTA resin. The proteins were blotted onto polyvinylidene difluoride membranes, stained with Ponceau-S (a), and reacted with immunization-induced rabbit anti-FasL antibody (d) or anti-FasL autoantibodies from SLE patients (b,c). Data shown are representative of seven independent experiments. Note that the immunization-induced antihuman FasL whole extracellular domain (fragment 5) antibody from three rabbits showed the same reaction pattern to the FasL deletion-mutant proteins (d). Among 21 SLE patients studied in this experiment, anti-FasL autoantibodies were present in the seven patients. The seven SLE patients showed the same reaction pattern (b,c).

Fig. 3

Inhibition by anti-FasL fragment antibodies of membrane FasL-induced apoptotic cell death of Fas-expressing Jurkat cells. Wild-type FasL cDNA transfected COS cells were incubated with PBS, control rabbit IgG (1 μg/ml), or affinity-purified rabbit anti-FasL fragment antibodies (1 μg/ml). Jurkat-NU cells were then co-cultured with the COS cells. Cell death was estimated by two-colour fluorescence analysis employing TUNEL staining with FITC and OKT3-PE antibodies. Data shown are representative of three independent experiments. Addition of rabbit anti-FasL fragment antibodies into the Jurkat-NU cell culture without FasL exerted no effects on cell death of the target cells (data not shown).

Fig. 4

Precise epitope mapping of anti-FasL autoantibody from SLE patients by immunoblotting analysis. Recombinant human FasL fragments 1·5, 1·8 and 2 were produced and purified using Ni-NTA resin. The proteins were blotted onto polyvinylidene difluoride membranes and the parallel gels stained with Quick-CBB. The membranes were reacted with anti-FasL autoantibodies from SLE patients. Seven patients were positive for the anti-FasL autoantibodies out of 21 patients studied. The autoantibodies from seven patients showed the same reaction pattern and inhibited the Fas/FasL-mediated apoptosis.

Fig. 5

Molecular modelling of Fas/FasL trimolecular complex. (a–c) A molecular model of the FasL/Fas complex was generated by a knowledge-based protein modelling method and the known tridimensional structures of lymphotoxin α (TNF-α) and TNF receptor. The computer-mediated structure model predicted the tertiary structure of the complex. Regions coloured red indicate FasL aa 162–169. Regions coloured yellow show aa 206 and aa 218, both of which were reported to be involved in Fas/FasL molecular interaction. FasL aa 162–169 is located on the outermost side of FasL facing towards its receptor, Fas. (d) Schematic representation of Fas/FasL trimolecular complex. FasL aa 162–169 (FasL 1) faces towards the Fas molecule (Fas 3), as opposed to Fas (Fas 2), to which aa 206 and aa 218 of the same FasL molecule (FasL 1) faces.

Fig. 6

Apoptosis-inducing activities of recombinant wild-type and mutant FasL proteins secreted by COS cells. COS cells were transfected with pME18S empty vector, pME18S carrying the full-length human FasL cDNA, or FasL constructs carrying point mutants. The culture supernatants were recovered 52 h after transfection and assayed for sFasL by ELISA. Cytotoxic activity of the mutant sFasL was determined per ng of protein. Fas-expressing Jurkat-NU cells were cultured with the supernatants containing wild-type and mutant sFasL proteins for 24 h. Cell death was then estimated by DNA staining with PI. Data shown are representative of three independent experiments. ▪, Soluble FasL 0 ng/ml; , 0·8 ng/ml; , 1·6 ng/ml; , 3·2 ng/ml.

Fig. 7

FasL-derived and infectious agent-derived synthetic peptides inhibit the binding of SLE patient anti-FasL autoantibodies to FasL fragment 2 by immunoblot analysis. An equal amount of recombinant FasL fragment 2 was run on a SDS-PAGE gel, transferred onto a membrane and reacted with anti-FasL autoantibodies in the presence of synthetic peptide. FasL peptide aa 161–170 and the HIV-derived peptide efficiently blocked autoantibody binding to FasL fragment 2, while FasL peptide aa 165–174 and the Staphylococcus-derived peptide moderately inhibited binding; eq. denotes equivalent amounts of solvent. We found that the peptide 161–170 inhibited the binding to the fragment 2 of the anti-FasL autoantibodies in four of four patients tested.

Fig. 8

Aberrant FasL protein expression on peripheral blood lymphocytes from SLE patients. Freshly isolated PBMC from four SLE patients were used either immediately (0) or stimulated by 3 h culture with PHA (3). As a control, normal PBMC were stimulated with PHA for up to 15 h. An equal amount of proteins from the cell lysates was analysed by immunoblotting using affinity-purified antihuman FasL fragment 5 antibody. Results shown are representative of six independent experiments.