Background/Purpose: There are 75,000 children affected by JIA in the United States. Despite recent therapeutic advances, treatment often requires chronic therapy and is associated with considerable cost and morbidity, yet is still not curative as episodic flares are common. However, the pathophysiology of these flares is not well understood. There is a common observation that arthritis flares exhibit a strong tendency to recur in the same joints. This fixed pattern of joint involvement is highly individualized and persists for decades. To study this phenomenon, we established a murine model of recurrent, joint-specific inflammation.

Methods: Methylated bovine serum albumin (meBSA) or ovalbumin (OVA) is injected into the wrist, knee and ankle joints of B6 mice. The contralateral side was injected with vehicle as a negative control. For the OVA conditions, 5×10^6 CD3+ cells collected from peripheral lymph nodes of OT-I mice were transferred IV into the mice 1 day prior to intra-articular OVA injection. Interleukin-1 was concurrently injected into the footpad of the mice to stimulate the immune response. Arthritis flare was triggered by i.p. meBSA or OVA re-stimulation. At days 7 (acute inflammation), 28 (remission) and 31 (flare), the joints were analyzed for inflammation by measuring joint thickness, histological evaluation, and flow cytometry analysis of disaggregated synovium.

Results: 7 days after injection, there is measureable swelling and histological evidence of synovitis, specifically in the joints exposed to the antigen. In the synovium, there is a corresponding increase in CD4 and CD8 T cell populations. At 28 days after injection, inflammation subsides with normalization of both joint size and resolution of synovitis. During this post-inflammatory remission, we found a persistent population of antigen-specific cells with resident memory T cell (TRM) phenotype (CD45.2+CD3+CD8+CD44+CD62L-CD69+CD103+). These TRM cells are preferentially increased in the synovium of the joint exposed to antigen. Upon intraperitoneal challenge with antigen, there is a recurrent inflammation specifically in the joints previously exposed to antigen, and a corresponding expansion of antigen-specific CD8+ TRM cells in the synovium is also seen.

Conclusion: Here, we created an inducible model of recurrent, joint-specific inflammation. Our data suggests synovial resident memory T cells may represent the basis for joint-specific memory in inflammatory arthritis.