The development of infrared microspectroscopy techniques to examine the effect of monoclonal antibodies to type II collagen in a murine model of rheumatoid arthritis

Type II collagen (CII) is the major protein in cartilage. Collagen induced arthritis (CIA), that occurs in animals immunised with CII, is an experimental model of human rheumatoid arthritis (RA). The autoimmune response that develops in CIA includes the formation of antibodies to CII that provoke an acute destructive arthritis, collagen antibody induced arthritis (CAIA), on transfer to naïve mice. Although CAIA can be transferred by some single monoclonal antibodies (mAbs) to CII and by combinations of mAbs, not all are arthritogenic and arthritogenicity appears to be epitope related. The epitope specificity of arthritogenic mAbs has been mapped to regions of the CII molecule that are involved in structural interactions important to maintain cartilage stability. Several of these arthritogenic mAbs have been shown to affect collagen fibril formation, chondrocyte morphology and the synthesis of new cartilage matrix in vitro. The aim of the present study was to examine the direct effects of combinations of mAbs to CII on pre-existing cartilage in the absence of inflammation in vitro in bovine cartilage explant cultures, and in vivo in mice that had not developed inflammation after injection with arthritogenic combinations of mAbs. Because mAbs penetrate intact cartilage poorly, biochemical techniques used to examine new matrix formation in chondrocyte cultures were not suitable for such analysis. Instead, Fourier Transform Infrared Microspectroscopy (FTIRM) was used to examine chemical changes across the cartilage. FTIRM has been used previously in preliminary studies that indicate that some individual mAbs cause damage to pre-existing cartilage in vitro. To test the effects of combinations of mAbs to CII on pre-existing cartilage in vitro, the morphology and chemical nature of the cartilage was examined using a standard combination of two mAbs M2139 and CIIC1 that are used to induce CAIA in vivo. Histological and FTIRM changes were examined in bovine cartilage explants cultured with or without mAbs for up to two weeks in vitro. The effect of the addition of the non-arthritogenic mAb CIIF4 to this combination of arthritogenic mAbs was tested, noting that in vivo CIIF4 decreases the incidence and severity of CAIA induced by M2139 and CIIC1. To determine whether the effects of the mAbs were mediated by the mAbs alone or by cellular mechanisms, experiments were performed on cartilage containing viable or non-viable chondrocytes. The role of matrix metalloproteinases (MMPs) in mAb mediated cartilage damage and protection was examined by the addition of the MMP inhibitor GM6001 to cultures containing the arthritogenic mAbs. To test the effects of combinations of mAbs to CII on pre-existing cartilage in vivo the morphology and chemical nature of the cartilage was examined in mice injected with arthritogenic mAbs that did not develop inflammation. Two different strains of mice were used, B10.Q mice that usually develop CAIA only after an additional injection of lipopolysaccharide (LPS) and B10.Q C5δ mice that lack complement factor C5 and do not develop CAIA. Mice were injected with the standard mixture of two mAbs, CIIC1 and M2139 or a more arthritogenic combination of four mAbs, M2139, CIIC1, UL-1 and CIIC2. Three days after injection, the mice were killed, and paws were decalcified and embedded in paraffin. The chemical nature of changes to the cartilage matrix and chondrocytes were examined by synchrotron FTIRM, and detailed histological examination of both matrix and chondrocytes was performed using a modified Mankin score, originally developed for scoring cartilage changes in osteoarthritis. FTIRM analysis focused on the identification of changes to CII and proteoglycans. The amide 1 band, which is typically a measure of total protein content, was used for identification of changes to CII. The position of the amide 1 band when native triple helical CII is present is located above 1660 cm-1, with a shift to a lower wavenumber upon denaturation of CII. Proteoglycan content was observed by examination of the height of the peak at 1076 cm-1 and compared with proteoglycan loss measured histologically by toluidine blue staining. The results of the in vitro experiments on cartilage explants showed that the two arthritogenic mAbs M2139 and CIIC1 caused progressive degradation of the cartilage surface, including substantial loss of proteoglycans seen both by FTIRM and by toluidine blue staining, and denaturation and loss of CII demonstrable by FTIRM. The cartilage damage did not require living chondrocytes as the mAbs caused greater cartilage destruction in the absence of viable cells. By contrast, the non-arthritogenic mAb CIIF4 alone had no apparent effect on cartilage, but when given with the arthritogenic mAbs prevented damage and caused apparent matrix regeneration, a process which only occurred in the presence of viable chondrocytes. The addition of the MMP inhibitor GM6001 reduced the cartilage damage caused by the arthritogenic mAbs M2139 and CIIC1 but did not mimic the protective effect of CIIF4. These results showed clearly that in addition to their adverse effects on the new synthesis of cartilage, CII reactive antibodies can cause direct damage to pre-existing cartilage. Moreover some antibodies could also be protective, depending on their epitope specificity, but the protection did not merely result from blocking binding of the damaging antibodies, as it required viable cells. To determine whether the arthritogenic mAbs likewise cause non-inflammatory cartilage damage in vivo, strains of mice with impaired capacity to develop inflammation were injected intravenously with either two or four potently arthritogenic mAbs, or with saline only as controls. Although none of the mice injected with mAbs showed visual or histological evidence of inflammation, in the presence of the mAbs that cause CAIA in vivo there were histological changes in the articular cartilage, including loss of proteoglycan and altered chondrocyte morphology. The changes seen were consistent with those previously observed in vitro in cartilage explants. Synchrotron FTIRM at high lateral resolution revealed loss of CII and the appearance of a new peak at 1635 cm-1 at the surface of the cartilage within the cells and in the matrix, that coincided with regions in which chondrocytes were ringed with strong toluidine blue staining, consistent with new proteoglycan synthesis. The composition of the 1635 cm-1 band may represent the appearance of a second band in the amide 1 region as a result of new synthesis of an unknown protein or from β-secondary structures within the telopeptides of newly synthesized collagen produced by activated chondrocytes; such telopeptides are normally removed prior to fibril formation. Detailed histological analysis was performed on all paws from the mice. The cartilage in the joints was scored for cartilage structure, chondrocyte integrity and organization, and proteoglycan loss using toluidine blue. Cartilage damage was shown by marked loss of cartilage structure, chondrocyte hyperplasia and/or loss, loss of proteoglycans in all joints, and protrusion of the chondrocytes from the cartilage surface, particularly in the small phalangeal joints. These data demonstrate that arthritogenic mAbs to CII cause direct cartilage damage in mice in vivo in the absence of inflammation. Such changes in inflammation-intact mice would readily enhance the exposure of damaged collagen fibrils to immune complexes containing collagen antibodies, and to the degradative enzymes released by inflammatory cells. Hence antibody-induced cartilage damage is a potential mechanism for initiation and perpetuation of cartilage damage in both murine arthritis and, possibly, in human RA wherein anti-CII antibodies of the same specificity are known to be present. These results amply demonstrate the utility of FTIRM for the analysis of localised changes in tissue. It has proved to be particularly useful for the analysis of cartilage, which is an unusual tissue made up of a small number of cells in an abundant extracellular matrix. It is the structure of this extracellular matrix that gives the cartilage its capacity to cushion the ends of joints. In contrast to most other tissues, the small number of chondrocytes in the tissue makes it difficult to examine changes in gene activity, and it is difficult to map changes based on histological techniques. Although there are other techniques available for the analysis of cartilage such as immunohistochemistry, this requires prior knowledge of the mechanisms involved in order to identify suitable neoepitopes of interest. In the present study the use of FTIRM has allowed the definitive demonstration that mAbs to CII can cause direct cartilage damage in the absence of inflammation both in vitro and in vivo. The use of FTIRM has allowed examination of the localised chemical changes within the cartilage matrix without any prior knowledge of the likely mechanism of action of the mAbs. These studies suggest that antibodies to CII may participate in the cartilage damage that accompanies articular inflammation and that such cartilage damage is independent of, and possibly precedes damage that results from inflammation. . Since antibodies of similar epitope specificity to the mAbs used in these studies occur in RA the data presented in this thesis provide new evidence demonstrating the involvement of CII specific antibodies in the pathogenesis and possibly initiation of human RA. A striking observation demonstrated in this thesis, is the protective effect of the non-arthritogenic mAb CIIF4, which suppressed the degradative effects of the arthritogenic mAbs. (...)