ANIMAL MODELS OF RHEUMATOID ARTHRITIS
No animal model is regarded as a high-fidelity replicate of human RA. However, several models that variably use T cells or antibodies that target the joint, or that induce monocyte activation and inflammation, have been used to study distinct aspects of the disease phenotype.
Animal models of arthritis can be used to understand elements of the arthritic process in patients. Recent therapeutic approaches in patients with RA, includ-ing the use of biological agents, are based on initial findings in murine models of experimental arthritis, although final proof of concept must come from clini-cal studies. Animal models are powerful tools for studying pathological changes in articular cartilage and bone in great detail and can be used to evaluate mecha-nisms of erosive processes. Although, in general, more inflammation drives more destruction, the uncoupling of inflamma-tion and erosion can be seen as well, and different mediators are involved in these processes.
Collagen-induced arthritis (CIA) is an example of an inducible animal model for human RA. In this model, the disease is induced in susceptible strains of animals by immunization with cartilage type II collagen. The resulting disease is a T-cell-dependent, antibody-mediated autoim-mune disease directed against the type II collagen immunogen. Similar to RA, CIA is characterized by massive infiltration of synovial joints by inflammatory cells and hyperplasia of the synovial mem-brane (pannus). In addition, the disease is characterized by both cartilage and bone destruction as seen in RA.
Both collagen-specific T and B lym-phocytes are involved in the induction of CIA. The effector mechanisms that lead to joint tissue destruction are less well under-stood, and a number of inflammatory cell types have been implicated. These include fibroblast-like synoviocytes, bone mar-row–derived macrophages, granulocytes, and dendritic cells as well as lymphocytes. Pro-inflammatory cytokines (IL-1β, TNF) and various chemokines are involved in the pathogenesis of the immune-mediated joint damage observed in both CIA and RA. Another similarity between RA and CIA is that the development of synovial inflam-mation is associated with pronounced angiogenesis with growth of high endothe-lial venules.
In contrast to RA, where large num-bers of T lymphocytes accumulate in the inflamed synovium, few T cells are found in the synovium of CIA. Another differ-ence between collagen arthritis and RA is that the former is more highly destructive than RA and in contrast to RA has a marked sensitivity to suppression of inflammation and joint destruction by NSAIDs such as indomethacin. In contrast to the female preponderance in RA, males are more sus-ceptible than females in murine CIA.
Adjuvant arthritis is the oldest stud-ied model of polyarthritis. It is induced by intradermal injection of complete Freund’s adjuvant, containing heat-killed myco-bacteria, in susceptible rat strains, mostly Lewis rats. Arthritis develops within two weeks. The active component in bacteria is the cell wall peptidoglycan.
Histopathological features are primar-ily a periarthritis, with marked periostitis rather than synovitis, and massive inflam-mation in the bone marrow. Immune-complex deposition in the cartilage is not a characteristic feature, and cartilage destruction is limited in early disease.
Adjuvant arthritis has a pure T-cell-driven pathogenesis, and passive trans-fer of T cells from diseased animals can induce arthritis. The joint inflammation is thought to reflect the generation of a T-cell reaction to bacterial epitopes cross-react-ing with endogenous bacterial fragments continuously present in synovial tissues or with cartilaginous antigens. Similar to RA, TNF and IL-1 have been shown to play a role in pathogenesis. Treatment with anti-TNF or anti-IL-1 therapy blocks inflamma-tion and tissue destruction in the model, with the most optimal blockade occurring with combination therapy. As seen in CIA, NSAIDs are effective inhibitors of cartilage and bone destruction in this model and this represents a significant difference from human RA.
Adjuvant-induced arthritis is severe but self-limited, and the rat recovers within a few months. This is a general shortcoming of most animal models when compared with the chronic process of human RA. An additional feature of adjuvant-induced arthritis is that once the rats undergo spontaneous remission the animal can no longer be reinduced to develop disease. This resistance has made the model suit-able for studies in regulation of T-cell tolerance.
Antigen-induced arthritis is a model of RA that is induced by immunizing animals with a foreign antigen, usually bovine serum albumin, and subsequently injecting the same antigen into a joint. As a result, a pronounced T-cell-dependent immune complex-mediated arthritis devel-ops that is severe but self-limited. The advantage of this model is that a defined part of the pathogenesis leading to arthritis is addressed and the arthritis remains con-fined to the injected joint, enabling com-parison with a contralateral control joint of the same animal.
The histopathology of antigen-induced arthritis has many characteristics of human RA. These include granulocyte-rich exudates in the joint space, thicken-ing of the synovial lining layer, and at later stages a predominantly mononuclear infiltrate in the synovium, which later includes numerous T cells and clusters of plasma cells. Intense immune complex formation is seen in superficial layers of the articular cartilage, which may con-tribute to localized cartilage destruction. Early loss of proteoglycan, followed by pannus formation and cartilage and bone erosion, is a common finding, and these characteristics are close to those found in human RA.
The arthritis produced is chronic and is thought to be due to prolonged antigen retention in the joint tissue in combination with antigen-specific T-cell-mediated delayed hypersensitivity. Prolonged antigen retention in the joint occurs by antibody-mediated trapping and charge-mediated binding. This principle is also of importance in the recently devel-oped KRN model of arthritis in which anti-glucose-6-phosphate isomerase (GPI) antibodies stick to GPI antigen trapped at cartilage surfaces and contribute to chro-nicity and destruction.
Unlike the other mouse models described so far, elimination of TNF and IL-1 is poorly effective in suppressing joint inflammation, pointing to a substantial role for other mediators of inflammation in this form of arthritis. Elimination of IL-1, however, did yield impressive protection against cartilage destruction.
The model of antigen-induced arthritis is most suited to studies into the mecha-nism of cartilage destruction, as induced by a mix of immune complexes and T-cell reactivity. The model is useful for the study of the regulation of local T-cell hyperreac-tivity against a retained foreign antigen in comparison with similar events against self antigens as seen in CIA.
This spontaneous mouse model of arthritis was generated by crossing the T-cell receptor (TCR) transgenic mouse line (known as KRNxC56Bl/6) with the nonobese diabetic mouse strain. Autoan-tibodies that arise in the K/BxN mouse recognize the ubiquitously expressed intracellular enzyme GPI. The novel fea-ture of this model is the development of a small-joint arthritis where the driving antigen has been elucidated and where the autoantibodies generated are patho-genic and transfer disease to other strains of mice.
The antibodies recognize endogenous GPI, which seems to associate preferen-tially with the cartilage surface. This may explain the dominance of joint pathology in these mice, although GPI is also abun-dant at other sites of the body. Activation of complement through the alternative path-way is crucial in these mice, and variable susceptibility in different mouse strains may largely be attributed to varying activ-ity of complement and level of expression of Fcγ receptors on phagocytes among the different strains.
Although anti-GPI antibodies are found in some RA patients in moderate lev-els, the role of these antibodies in disease pathogenesis remains to be defined. The role of IL-1 and TNF in the pathogenesis of KRN arthritis is similar to other murine models. IL-1-deficient mice do not develop arthritis and TNF-deficient mice have a more variable course.