Degradation of extracellular matrix in articular cartilage is a central event that leads to joint destruction in many arthritic conditions, including rheumatoid arthritis (RA), osteoarthritis (OA), and septic arthritis. Chondrocytes respond to a variety of stimuli, such as proinflammatory cytokines and mechanical loading, by elaborating degradative enzymes and catabolic mediators. Whether the initiation of matrix degradation is cytokine driven or biomechanical, it is likely that the downstream degradative pathways are similar, involving specific matrix metalloproteinases (MMPs) (1).
Degradation of cartilage matrix macromolecules is also associated with increased expression of mediators of inflammation, such as nitric oxide (NO) and prostaglandin E2 (PGE2). NO is involved in the stimulation of MMP messenger RNA (mRNA) expression and activity (2, 3), inhibition of matrix component production (4, 5), and, in the presence of superoxide anions, induction of chondrocyte apoptosis (6). PGE2 produced at high levels at sites of inflammation mediates cartilage resorption by decreasing proliferation, enhancing MMP activity, and inhibiting aggrecan synthesis in chondrocytes, and it can also potentiate the effects of other mediators of inflammation (7, 8). It is well accepted that interleukin-1 (IL-1) and tumor necrosis factor α (TNFα) are key cytokines involved in articular cartilage destruction as well as in the inflammatory response in arthritis. Biologics that inhibit the signaling cascade mediated by both of these cytokines have been effective in the treatment of RA, resulting in alleviation of both inflammation and joint destruction. However, blocking of IL-1 and/or TNFα does not lead to complete protection of the joint structure, indicating that other signaling pathways that mediate joint catabolism remain to be elucidated.
Toll-like receptors (TLRs) are phylogenetically conserved receptors involved in the innate immune response, and they recognize pathogen-associated molecular patterns (9). The mammalian homologs of TLRs belong to a family that currently consists of 10 members in humans (10). The ligands for several of the TLRs (TLRs 2–6 and 9) have been identified and include nonbacterial products, such as Hsp70 and fatty acids, as well as microbial constituents (11). Because ligand recognition by TLRs elicits strong activation of proinflammatory cytokines and up-regulation of costimulatory molecules (12), the role of TLRs in the exacerbation of the inflammatory response and joint destruction in arthritis has been postulated.
Synovial tissues from RA joints express TLR-2 predominantly at sites of cartilage and bone destruction, and expression of TLR-2 in RA synovial fibroblasts was shown to be increased after treatment with proinflammatory cytokines (13). Bacterial peptidoglycan, a TLR-2 ligand, activates synovial fibroblasts to express integrins, MMPs, proinflammatory cytokines, and chemokines, suggesting that signaling through TLR can elicit a proinflammatory response in synoviocytes (14, 15). Mice deficient in myeloid differentiation factor 88 (MyD88), a Toll/IL-1 receptor (IL-1R) domain–containing adaptor molecule known to have a central role in TLR signaling, do not develop a streptococcal cell wall (SCW)–induced arthritis and show significant amelioration of both joint swelling and cartilage degradation (16). This underscores the importance of the TLR-mediated signaling pathway in the regulation of inflammation and joint destruction of arthritis.
Recently, human articular chondrocytes were shown to express TLRs (17, 18). Microcrystals of calcium pyrophosphate dihydrate (CPPD) and monosodium urate monohydrate (MSU) were found to trigger NO generation via TLR-2–mediated signaling in chondrocytes, indicating the potential of TLR-mediated signaling to directly contribute to degradative tissue reactions associated with arthritis. In the current study, we examined the catabolic signaling pathway mediated by TLR ligands, bacterial peptidoglycan, and lipopolysaccharides (LPS), as well as the role played by MAPKs and NF-κB in TLR-mediated catabolic signaling, in human articular chondrocytes.
- Top of page
- MATERIALS AND METHODS
There is a continuing challenge in defining the catabolic triggers of articular chondrocytes in the pathogenesis of cartilage degradation in various arthritides. Findings from this study suggest that TLR ligands play a pivotal role as potent catabolic stimuli in articular chondrocytes. The role of TLR ligands in the pathogenesis of inflammatory arthritis is not well understood. However, the presence of bacterial peptidoglycan and functional TLR-4 ligand in RA joints raises the possibility that TLRs may be involved in the joint degradation of RA as well as in septic arthritis (21, 22). TLRs 2 and 4 are expressed in the synovial tissue of patients with clinically active RA and are associated with the levels of both IL-12 and IL-18 (13, 23). In addition, TLRs 3 and 7 are highly expressed in RA synovium, and combined stimulation of TLR-4 and TLR-7/8 or TLR-3 and TLR-7/8 results in a synergy of the production of mediators of inflammation (24).
In our study, both TLR-2 and TLR-4 mRNA were detected in cultured OA chondrocytes. Our findings differ from those in 2 previous studies that showed TLR-2 mRNA in both normal and OA chondrocytes but did not find TLR-4 mRNA (17, 18). This discrepancy may result from variations in the donors and in the RT-PCR conditions employed. We believe that functional TLR-4 exists in chondrocytes because our results, as well as the results of many others, demonstrate that chondrocytes respond consistently to LPS, a traditional TLR-4 ligand, and this response is effectively blocked by anti–TLR-4 antibody.
Immunohistochemistry revealed that OA lesional cartilage expressed abundant TLR-2 and TLR-4 proteins compared with non-OA and nonlesional cartilage. TLR-2 expression increased in response to TLR ligand itself as well as in response to proinflammatory cytokines, suggesting a positive amplification loop contributing to the inflammatory response in articular chondrocytes. This hypothesis is also supported by the finding that the fibronectin fragment, a well-known catabolic mediator in chondrocytes, also up-regulates TLR-2 expression in articular chondrocytes (17). Recent studies show that IL-1β up-regulates TLR-2 mRNA in hepatocytes and epithelial cells via the NF-κB–dependent pathway (25, 26).
Both peptidoglycan and LPS were found to stimulate phosphorylation of all 3 MAPKs as well as NF-κB nuclear translocation in chondrocytes. In addition, both ligands stimulated MMP, NO, and PGE2 release in chondrocytes and increased the release of proteoglycan and collagen degradation product into the explant culture media. Increases in MMPs, NO, and PGE2 were completely inhibited by the NF-κB inhibitor, which is consistent with previous studies showing a link between the cytoplasmic Toll/IL-1R domain and NF-κB activation in terms of iNOS and MMP up-regulation in a variety of cells (27, 28).
Our results show that in addition to NF-κB, the JNK pathway mediates the induction of mediators of inflammation in response to TLR ligands in chondrocytes. The MAPKs p38 and ERK primarily regulated the TLR-mediated PGE2 response. Interaction between TLR signaling and MAPK in the proinflammatory response is not fully understood, and different patterns seem to exist in different cells. TLR-mediated signaling depends upon adaptor molecules such as MyD88 and TRIF and is often classified as an MyD88- or TRIF-dependent pathway (29). TLR interaction with MyD88 leads to recruitment of IL-1R–associated kinase 1 (IRAK-1), activation of IRAK-1, its association with TNF receptor–associated factor 6 (TRAF6), and activation of MAPKs and NF-κB (30). A recent study showed that MSU and CPPD crystals induce NO production in articular chondrocytes, which is dependent on IKK-2–mediated NF-κB activation (18). Upstream of NF-κB, parallel to the canonical MyD88, IRAK, and TRAF6 signaling, the Rac1/phosphatidylinositol 3-kinase pathway has been found to mediate TLR-2–mediated NO production by MSU and CPPD crystals (18).
Despite numerous reports on the stimulatory function of LPS in articular chondrocytes, a paucity of data exists on the pertinent signaling mechanism. Delineation of a catabolic signaling pathway other than MAPK and NF-κB mediated by peptidoglycan and LPS will be the subject of future investigations in our laboratory. Because the IL-1 receptor and TLRs share a common signaling pathway, we examined the induction of IL-1 in peptidoglycan- or LPS-treated chondrocytes. Peptidoglycan and LPS at a concentration up to 1 μg/ml failed to induce significant IL-1 in chondrocytes (data not shown). In addition, the catabolic response mediated by peptidoglycan or LPS was not inhibited by the IL-1 receptor antagonists (data not shown). In a study investigating the in vivo role of TLR-mediated signaling in an SCW-induced arthritis model using TLR and MyD88-knockout mice (16), it was found that signaling via the TLR/MyD88 pathway is responsible for early cartilage damage independent of IL-1 and IL-18 before induction of the inflammatory response.
In our study, chondrocytes were obtained from OA cartilage samples both for monolayer culture and for explant culture. It would have been more relevant to study the effect of TLR ligands in chondrocytes obtained from patients with RA or septic arthritis, because the catabolic signaling of TLR would be more pronounced in these diseases compared with that in OA. For culture, we excluded severely affected areas, and the chondrocytes mostly came from the nonlesional area, which is similar to non-OA cartilage in terms of the low level of TLR expression. However, we cannot exclude the possibility that OA pathology itself might have influenced the chondrocyte responses to TLR ligands. Our results thus may not be extrapolated to the situation in RA or normal cartilage.
In summary, we have shown that TLR is present in articular cartilage and cultured OA chondrocytes and is increased in the presence of OA lesions. TLR-2 expression is increased by catabolic cytokines and by the TLR ligand itself. TLR-2 and TLR-4 ligands strongly induce catabolic responses in chondrocytes. Recently, TNFα blockade was reported to down-regulate the systemic and local expression of TLR-2 and TLR-4 in spondylarthropathy, raising the possibility that modulation of TLR might contribute both to therapeutic aspects and to side effects (31). TLR ligands include components of the extracellular matrix, and TLR-mediated catabolic responses may further trigger chronic inflammation, leading to self-perpetuating loops of cell activation in arthritis, with a minimal requirement for adaptive immune mechanisms (32). Modulation of TLR-mediated signaling as a potential therapeutic strategy for arthritis requires detailed elucidation of the signaling pathways and endogenous ligands involved.