Pro-inflammatory mediators in arthritic pain and disease

© 2016 Dr Reem Abdu SalehRheumatoid arthritis (RA) is a systemic autoimmune disease that affects predominantly the synovium of the joints, and ultimately leads to articular cartilage damage, bone erosion, joint destruction and biomechanical dysfunction. Arthritic pain has been identified as one of the most severe symptoms associated with RA, and is a major reason for the disability and poor quality of life in such patients. Pro-inflammatory cytokines and chemokines released in the RA synovial microenvironment have been thought to play key functions in the development of arthritic pain, as well as disease. Targeted therapies against certain inflammatory mediators are providing benefit for arthritic pain and disease. Tumor necrosis factor (TNF) and granulocyte macrophage colony-stimulating factor (GM-CSF) are pro-inflammatory cytokines which have been implicated in the pathogenesis and progression of RA, as well as in the generation of arthritic pain. Biologics targeting both of these cytokines and their receptors have been developed for the treatment of RA; however, the mechanisms by which TNF or GM-CSF mediate pain and arthritis have not been fully elucidated, and whether these cytokines are linked in pain development or whether they act via separate pathways has not been determined. C-C chemokine ligand 17 (CCL17), a chemokine, can be induced by GM-CSF via the transcription factor, interferon regulatory factor-4 (IRF-4), in monocytes and macrophages in vitro. Interleukin-23 (IL-23) is another cytokine that has been implicated in RA pathogenesis but its potential role in arthritis pain has not been investigated. Colony-stimulating factor-1 (CSF-1), another cytokine, and its receptor CSF-1R (c-Fms), have also been implicated in RA and in the development of several arthritis models; however their potential involvement in arthritic pain has received limited attention. The objectives of this thesis were as follows: (i) to examine the kinetics by which TNF or GM-CSF can mediate arthritic pain and disease, and to identify downstream mediators and cell types required for pain and arthritis driven either by TNF or GM-CSF; (ii) to investigate the potential interaction between TNF and GM-CSF in arthritic pain development; (iii) to determine the contribution of CCL17 and IRF-4 to pain and arthritis development driven by either TNF or GM-CSF; (iv) to examine the involvement of IL-23, CSF-1 and CSF-1R in arthritic pain development. The acute monoarticular methylated bovine serum albumin (mBSA)/cytokine-induced arthritis model, whereby mBSA is injected intraarticularly (i.a.) into the knee joint on day 0 followed by subcutaneous (s.c.) injections of a cytokine on days 0-2, was utilized to study the kinetics of pain and disease development driven by either TNF or GM-CSF. Arthritic pain was monitored using an Incapacitance Meter, which measures the differential weight distribution between arthritic and non-arthritic hind limbs. Arthritis development and severity were assessed by histology. Early signs of arthritic pain and disease were evident in the mBSA/TNF-driven model (6 hours post injection), in contrast to the mBSA/GM-CSF-driven model where pain and disease development were delayed (day 3 or 4 post injection). Although the mBSA/TNF and mBSA/GM-CSF models displayed different kinetics of arthritic pain and disease progression, they shared the synovial composition of arthritis (identified by flow cytometric (FACS) analysis), the requirement for the same cell types (identified by the use of genetic approaches and neutralizing monoclonal antibodies (mAbs)) and similar downstream inflammatory mediators (identified by gene expression analysis), which suggests that the mBSA/TNF- and mBSA/GM-CSF-driven pain and inflammatory pathways are linked. mBSA/TNF- and mBSA/GM-CSF-driven pain and arthritis were found to be dependent on neutrophils and CSF-1R signalling (using neutralizing mAbs). By utilizing gene knockout mice, results suggested that mast cells may also contribute to pain and arthritis in the mBSA/TNF- and mBSA/GM-CSF models, with no contribution of T and B lymphocytes. Based on this strategy (gene knockout mice), CCL17 and IRF-4 were identified as downstream mediators required for pain and arthritis driven by mBSA/TNF or mBSA/GM-CSF; also, IL-23 is another downstream mediator required for mBSA/GM-CSF-driven pain and arthritis. Using neutralizing mAb-based strategies, mBSA/TNF-driven pain and arthritis were found to be dependent on GM-CSF, and, conversely, mBSA/GM-CSF-driven pain and arthritis were found to be dependent on TNF, suggesting that both cytokines act in an interdependent manner. For the first time, the ability of exogenous CCL17 or IL-23 to induce arthritic pain and exacerbate disease, following an i.a. injection of mBSA, was demonstrated. mBSA/CCL17- and mBSA/IL-23-driven pain and arthritis occurred independently of T and B lymphocytes, GM-CSF, TNF and IRF-4 (using gene knockout mice); however, both were dependent on a cyclooxygenase (COX) product(s) (determined by the systemic administration of indomethacin, a non-selective inhibitor for COX enzymes), which includes a class of well-known pain mediators called prostaglandins. Pain and arthritis in the mBSA/CCL17 model were also dependent on IL-23, while the mBSA/IL-23 model did not require CCL17 (using gene knockout mice). In the monoarticular antigen-induced arthritis (AIA) model, the deletion of CCL17 was effective in blocking pain (based on weight distribution measurements) and ameliorating arthritis pathology (assessed by histology), predominantly cartilage damage and joint destruction, further highlighting the importance of CCL17 for arthritic pain development. Given that these data suggest a novel pathway, (TNF ↔ GM-CSF → IRF-4 → CCL17→ IL-23 → a COX product(s)), by which the interdependence of TNF and GM-CSF initiates a signalling cascade that involves IRF-4, CCL17, IL-23 and a COX product(s) for the development of arthritic pain and disease, there could be implications with CCL17 and IL-23 being potential targets for treating inflammatory pain in arthritis, and perhaps in other inflammatory diseases. Similar to the mBSA/TNF model, using similar methodological approaches, mBSA/CSF-1 (i.a. mBSA on day 0, followed by s.c. injections of CSF-1 on days 0-2) could induce early signs of an algesic response and arthritis, which were independent of T and B lymphocytes. Unlike the mBSA/TNF model, pain and arthritis in the mBSA/CSF-1 model did not require neutrophils. The blockade of either CSF-1 or the CSF-1R, using neutralizing mAbs, had no analgesic (based on weight distribution measurements) or anti-inflammatory effects (histology) in the AIA model, in contrast to the neutralization of GM-CSF. This finding, in conjunction with data from the mBSA/CSF-1-driven model, suggests that, although exogenous CSF-1 can contribute to the development of arthritic pain (perhaps via indirect mechanisms), it may not be a good therapeutic target for treating pain or arthritis