Résumé
Background Membrane-bound FasL (mFasL) is able to induce fibroblast-like synoviocytes (FLS) apoptosis. In experimental arthritis mouse models, injection of agonistic antibody (Ab) anti-Fas decreased the symptoms. However, soluble FasL (sFasL) is increased in Rheumatoid Arthritis (RA) patients and correlated with disease activity. These results indicated that mFasL could be protective whereas sFasL could be deleterious suggesting that they could have different functions. We analysed the effect of different FasL preparations mimicking sFasL or mFasL on RAFLS proliferation and apoptosis. Methods RAFLS were treated with different FasL preparations (FasL-Flag ± Ab α-Flag, FasL-Fc or sFasL) or with agonistic Ab anti-Fas. Apoptosis was then analysed by cytometry using annexinV-FITC and TOPRO-3. Proliferation was measured using tritiated thymidine. Signaling pathways was analysed by western blot and their influence was assessed using chemical inhibitors. sFasL was quantified in synovial fluids from patients using cytometric bead array. Results FasL-Flag alone (mimicking sFasL) was not able to induced FLS apoptosis while proliferation was significantly activated (3.3 ± 1 fold; n = 5; p < 0.05). Similarly, sFasL was only able to strongly induce RAFLS proliferation (7 ± 5.5 fold, n = 9 p < 0.05). Membrane bound FasL (FasL-Flag + Ab α-Flag) significantly induced RAFLS apoptosis (52% ± 18; n = 5) and a slighter but significant proliferation (2.2 ± 0.3 fold; n = 4). Duality of mFasL was confirmed using agonistic Ab anti-Fas (mimicking mFasL) with pro-apoptotic (38% ± 18; n = 2) and proliferative effect (4.4 ± 2.0 fold). Finally, growing concentration of FasL-Fc leads to aggregation of the protein, mimicking mFas or sFasL at high and low concentration respectively. Dose responses confirmed mFasL and sFAsL effects. FasL activated Akt, JNK and ERK but also activated caspases (n = 5). Inhibition of each pathways block FasL-induced proliferation. However, only JNK inhibition significantly increased FasL-induced apoptosis. We observed that FasL-Fc was able to induce osteoarthritis (OA) FLS apoptosis but neither FasL-Fc nor sFAsL was able to significantly induced proliferation of OAFLS (1.4 ± 1.3 and 2.6 ± 1.1 fold respectively, n = 4). Synovial fluids from patients with RA (n = 16) tends had higher sFasL concentrations compared to those with osteoarthritis (n = 10) (p = 0.06). Conclusions mFasL induces preferentially RAFLS apoptosis, whereas sFasL only induces RAFLS proliferation. Proliferative effect of sFasL was not seen on OAFLS. According to what we have already described for TRAIL, caspases are involved in FasL-induced apoptosis and proliferation. This is the first demonstration that sFasL and mFasL have different effects on RAFLS proliferation. sFasL by enhancing RAFLS proliferation could have a deleterious role in RA. Therefore, its blockage could be a therapeutic tool to prevent RA.