Background The B cell chemoattractant CXCL13 has recently emerged as a new candidate biomarker of disease activity capable of identifying patients with persistent synovitis and worst radiographic outcomes in early rheumatoid arthritis (RA). However, whether CXCL13 reflects underlying disease processes or simply represents another non-specific marker of inflammation is currently unknown.

Objectives To analyse the clinico-pathologic significance of serum CXCL13 in comparison to routine laboratory markers of disease activity and severity in patients with RA.

Methods Baseline serum levels of CXCL13 were measured by colorimetric ELISA in 205 consecutive early untreated RA patients with disease duration <12 months (median 3 months, IQR 2-5.5). Disease activity was assessed by a comprehensive set of subjective, semi-objective and objective clinical features. Changes in CXCL13 levels were evaluated in 87 patients after 2 months of treatment with methotrexate and low-dose prednisone. An additional study population of 60 RA patients (n=22 with disease duration <12 months) in whom paired serum and synovial samples were collected on the same day was used to assess the pathologic correlates of circulating CXCL13.

Results In cross-sectional analyses at baseline, CXCL13 was moderately correlated with the erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) (rho 0.35 and 0.36 respectively, p<0.001). Similarly to acute phase reactants, CXCL13 correlated with overall disease activity as measured by the DAS28, in particular with physician-derived measures, as well as with ultrasonographic scores for Gray Scale and Power Doppler signals. In contrast to ESR and CRP, no correlation was found with patient-derived measures and functional status. Although increased CXCL13 levels were found in patients with anti-citrullinated protein antibodies (ACPA), high CXCL13 and ACPA were not synonymous. CXCL13 in the 3rd tertile (>100 pg/ml) was found in 27.9% of ACPA(−) patients, and, in turn, 53.4% of ACPA(+) patients had CXCL13 <100 pg/ml. Similar results were observed for rheumatoid factor. After 2 months of treatment, CXCL13 levels were not significantly changed from baseline, as opposite to the significant reduction of acute phase reactants (standardised response mean 0.04, 0.52 and 0.66 for CXCL13, ESR and CRP respectively). In paired serum and tissue samples, circulating CXCL13 was significantly correlated with synovial CXCL13 protein (n=60 patients, rho 0.30, p=0.04) and mRNA (n=19 patients, rho 0.56, p=0.02) expression. Similarly to ESR and CRP, serum CXCL13 was related to synovial inflammatory features such as the degree of sublining macrophage infiltration (rho 0.34, p=0.01). Serum CXCL13, but not acute phase reactants, showed further correlation with specific histopathologic and molecular synovial parameters such as the presence and density of large B cell aggregates (rho 0.28, p=0.03), expression levels of the B cell enzyme activation induced cytidine deaminase (AID) (rho 0.4, p=0.046), and the receptor activator of nuclear factor κB ligand (RANKL)/osteoprotegerin (OPG) ratio (rho 0.72, p<0.01).

Conclusions Serum CXCL13 in RA reflects an immunologically active and potentially persistent pattern of synovial inflammation beyond the levels of conventional inflammatory markers and the ACPA status.

Disclosure of Interest : None declared

DOI 10.1136/annrheumdis-2014-eular.5133