Autoimmune disease: Rheumatoid arthritis Animal models of rheumatoid arthritis Darren L. Asquith, Ashley M. Miller, Iain B. McInnes and Foo Y. Liew Division of Immunology, Infection and Inflammation, Glasgow Biomedical Research Centre, University of Glasgow, Glasgow, Scotland, UK DOI 10.1002/eji.200939578 Animal models have been used extensively in studies of rheumatoid arthritis pathogenesis. Despite the inherent limitations of all animal models, several rodent models have significantly progressed our under- standing of the fundamental mechanisms underpinning rheuma- toid arthritis and contributed to several current major advances in treatment. These models include the induced arthritis models such as collagen-induced arthritis, collagen- antibody-induced arthritis, zymosan- induced arthritis, and the methylated BSA model, and the genetically manipulated or spontaneous arthritis models such as the TNF-a-transgenic mouse, K/BxN mouse, and the Skg mouse. Here, we describe these animal models and discuss their advantages and limitations. Rheumatoid arthritis (RA) is a chronic inflammatory disease that affects primar- ily the joints manifesting as pain, stiff- ness, and synovitis (inflammation of the synovial membrane) leading in turn to articular destruction. The etiology of RA is multifactorial, with genetic and environmental components that together lead to early immune perturba- tion in both the innate and adaptive compartments and subsequent chronic inflammation. In particular, studies of sera from pre-disease onset human cohorts reveal the presence of auto- antibodies against citrinullated self- proteins and rheumatoid factors that predate disease onset by up to 10 years. The trigger to articular disease onset is unknown and factors that define chroni- city of the responses are similarly poorly understood. By definition this phase of disease, although potentially crucial for future preventative therapeutics, is impossible to study at present in humans and thus provides powerful incentive for development of relevant animal models of disease, in which mechanisms and pathways can be modeled and explored. Moreover, in established RA synovitis there is ample evidence of innate and adaptive immune pathways that co-exist and interact to perpetuate tissue inflam- mation and destruction. Thus, germinal centers are found in a significant propor- tion of RA synovia, auto-antibody titers are characteristic of poor prognosis disease and recently therapeutics that target both innate cytokines, e.g. TNF-a, IL-6 and adaptive immune pathways, e.g. abatacept (co-stimulatory blockade) and rituximab (B-cell depletion) have proven to be efficacious. Whereas the studies in human tissues are helpful in defining such pathways, they are significantly limited in that true kinetic analyses are impossible, relevant tissues, e.g. lymph node, bone marrow, spleen, are not readily obtainable and even in synovial tissue, the target lesion is accessed only rarely and not normally on an ongoing basis. Therefore, animal models are essential to facilitate understanding of the mechanisms of disease in RA and development of new therapies. Induced arthritis models Collagen-induced arthritis Collagen-induced arthritis (CIA) shares many similarities with human RA. Two characteristics of the CIA model – breach of tolerance and generation of auto- antibodies toward self and collagen – make CIA the gold standard in vivo model for RA studies. CIA was first described in rats [1] and subsequently shown to be inducible in susceptible strains of mice [2], following inoculation with type II heterologous collagen in complete Freund’s adjuvant. Susceptibility has been linked to strains that have MHC Class II I-A q haplotypes; however, it is clear that many mouse strains have variable degrees of susceptibility to CIA. Similarly restricted class II genotypes can be found in RA patients, for whom pathogenesis is associated with HLA-DR1 and HLA-DR4 (reviewed in [3]). CIA can also be initiated in non-human primates, making it a useful model in which to better assess efficacy of novel therapeutic targets and aid their transition through the primary stages of pre-clinical development. DBA/1 mice are most widely used in the CIA model. Clinical signs of disease typically develop 21–25 days after the initial inoculation and presents as a polyarthritis, which is most prominent in the limbs and characterized by synovial inflammatory infiltration, cartilage and bone erosion and synovial hyperplasia similar to human RA. The development of CIA is associated with both B- and T-lymphocyte responses with the production of anti-collagen type II anti- bodies and collagen-specific T cells. The auto-antibody response in CIA is predo- minated by the IgG2 subclass with high levels of both IgG2a and IgG2b present at the peak of arthritis. Disease severity is expected to peak at approximately day 35, after which DBA/1 mice enter remission, marked by increased concen- trations of serum IL-10 and a subsequent decrease in pro-inflammatory Th1 cyto- kines [4]. In a further development of the model, inoculation with homologous type II collagen has been reported to cause chronic relapsing arthritis more akin to human RA, and has been suggested to be Eur. J. Immunol. 2009. 39: 1991–2058 FORUM 2040 & 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu