Method: Three models for OA were used, collagenase induced OA (CIOA), destabilization of the medial meniscus (DMM) and spontaneous OA in IL-1ra-/- murine knee joints. CIOA was induced by injection of collagenase into murine knee joints. Due to instability, OA-like pathology develops within 42 days. DMM was induced by transsecting the medial meniscotibial ligament, which results in OA-like pathology within 56 days. The role of TLR2 was studied by induction of CIOA and DMM OA in TLR2-/- mice (n=16) and C57BL/6 controls (n=15), and by backcrossing of TLR2-/- mice to IL-1ra-/- mice. At end point, knee joints were isolated and OA-like changes in the cartilage were scored. At several time points after induction, synovial tissue was isolated to study TLR2 expression in the wild type (WT) mice by Q-PCR.
In CIOA and to a lesser extent in the IL-1ra-/- mice, synovial activation, indicated by thickening of the synovial lining layer, was clearly present during the whole course of the disease. However, no synovial activation was observed in the DMM model. At day 7 after induction of CIOA in WT mice, synovial TLR2 levels were upregulated up to 56 fold, probably reflecting cellular influx at this time point. At day 21 TLR2 levels were up 2-fold and 42 days after induction, TLR2 levels in the synovium were down to baseline levels. In the DMM model, TLR2 was not regulated at any of these time points.
During CIOA in TLR2-/- mice, OA cartilage pathology increased from 10,0 in the WT controls to 15,5 (p<0,02). Changes were observed in both medial and lateral joint compartment. Incidence of severe cartilage damage was increased in the medial femur, medial tibia, lateral femur and lateral tibia respectively from 13% to 44%, from 27% to 69%, from 60% to 81% and from 53% to 81%. In knee joints of IL-1ra-/- mice, damage was mild compared to CIOA WT mice with a mean score of 2,2 and showed an increase in the IL-1ra-/-/TLR2-/- up to a score of 4,1. In contrast, cartilage pathology during DMM OA in TLR2-/- was not changed in the medial compartment of the joints, compared to WT mice, respectively 13,7 and 14,7. Severe cartilage damage did not differ between WT and TLR2-/-. In all models, synovitis in TLR-2-/- mice was comparable to the WT, indicating no direct role for TLR2 in synovitis.
Conclusion: Unexpectedly, this study indicates that in TLR2 deficient mice OA pathology is more severe in experimental OA that involves synovial activation. This suggests a protective role for this receptor in OA. This protective role was abolished in the absence of synovial activation, as was found in the DMM model. Stimulation of TLR2 by endogenous ligands such as biglycan, hyaluronan or HMGB-1 has been shown to induce IL-10, TGFb and HGF, which are protective mediators for OA. Although further research is needed, these results indicate that the TLR2 pathway protects cartilage from developing OA and that this effect is mediated by the synovium.
Disclosure: A. Blom, None; P. van Lent, None; S. Abdollahi-Roodsaz, None; P. M. van der Kraan, None; W. B. van den Berg, None.