From autoimmunity to inflamed joints : study of risk factors, mediators and cellular targets in rheumatoid arthritis

Abstract

Rheumatoid arthritis (RA) is a chronic autoimmune disease affecting mainly the joints and leading to disability. Circulating antibodies against citrullinated proteins (ACPA) are present in the majority of patients and are considered important to disease pathogenesis, however the mechanisms triggered by these antibodies remain unclear. Although significant progress has been made in refining disease modifying therapies during the last decades, to date there is still no curative treatment. In this context identification of individuals likely to develop the disease from those being at risk can help rethink the therapeutic strategy towards prevention and personalized medicine.

The main objective here was to study the role of ACPA in disease pathogenesis in the very early stages of RA in order to a) set up a predictive model for arthritis development and b) have a better understanding of ACPA effects on cellular targets, especially of the immune system. This work followed a translational approach: the observations from the clinics were assessed in experimental models, in a laboratory setting.

In Paper I we followed-up a cohort of individuals at risk for developing RA (Karolinska RiskRA cohort) i.e. individuals positive for CCP2-antibody test and musculoskeletal symptoms (joint stiffness and pain) in the absence of clinical or ultrasound signs of arthritis or any other rheumatic diagnosis. Baseline data was collected, including clinical and ultrasound assessments of the joints and blood samples were analyzed for 9 ACPA reactivities, a panel of 92 inflammation-related markers and HLA shared epitope. 101/267 individuals progressed to arthritis within a median of 14 months of follow-up. The multivariate analysis showed that the presence of at least one ACPA reactivity, tenosynovitis detected by ultrasound, high levels of IL-6 and low levels of IL-15Rα had independent and significant predictive value in a model for arthritis progression.

Next, in Paper II we tested whether a symptom-based e-questionnaire called “Rheumatic?” could be used as a tool for prediction of arthritis development. This was an international collaboration with two other centers, in Leiden, The Netherlands and in Erlangen, Germany. The tool was primarily aimed to be used by patients and, based on their answers, to suggest visit to the general practitioner or directly to the rheumatologist. In the case of the Karolinska RiskRA cohort the tool had limited usability regarding discrimination for an inflammatory outcome, suggesting that symptomatology alone is not sufficient for prediction. This confirms our previous observations from our cohort where clinical parameters were not significant for risk prediction in univariate analysis.

In Paper III we studied the effects of ACPA on pain-like symptoms, tenosynovitis and bone loss in the preclinical stages of RA. Briefly, ACPA obtained from RA patients were injected in healthy mice. MRI of the joints at 28 days post injection showed signs of tenosynovitis, which was further confirmed by histological analysis. Moreover ACPA-treated mice showed pain-like behavior and had signs of bone resorption in Xray microscopy.

In Paper IV we aimed to further characterize the citrullinated targets of ACPA in neutrophils, cells potentially important in disease pathogenesis. For this purpose we tested ACPA obtained from RA patients and an at-risk individual. Some but not all ACPA could bind to citrullinated targets released by activated neutrophils, suggesting that neutrophils could be important sources for citrullinated antigens. However, ACPA could not induce neutrophil activation nor bind to intact cells, suggesting a limited role in perpetuating ACPA response via boosting autoantigen release from neutrophils. When testing polyclonal ACPA preparations from individual RA patients we observed a highly variable binding capacity to neutrophil-derived antigens and similarly, a variable effect on another target cell for ACPA, specifically the osteoclast. These results suggest high patient-to-patient variability in the ACPA effects in line with the clinical observations on the heterogeneity in RA clinical phenotypes.

In conclusion, on the timeline from systemic autoimmunity to clinical diagnosis of RA, we propose a particularly high at-risk for disease phase, characterized by the presence of ACPA, subclinical tenosynovitis and inflammation-related factors. Digital tools such as Rheumatic? e-questionnaire show promising use and should be further developed to help arthritis prediction. ACPA could directly contribute to symptomatology of the risk phase, where a heterogeneity in targeting different cell types may contribute to the various clinical presentations.