Osteoarthritis (OA) is a multifaceted degenerative joint disease, where inflammation of the synovial membrane leads to accelerated joint destruction, resulting in elevated release of matrix metalloproteinase (MMP)-mediated type I and III collagen degradation fragments, C1M and C3M. The aim of the study was to investigate if C1M and C3M were associated with joint inflammation in OA, by i) measuring the release of the biomarkers in response to pro-inflammatory factors in an ex vivo human OA synovial explant model and ii) investigate the level of the biomarkers in patients in association with radiographic and inflammatory knee OA.

  Methods

Synovial membrane explants (SME); Synovial membrane from 4 OA patients undergoing total knee replacement at Gentofte Hospital, Denmark, were cultured as SMEs (30±2mg) for 3 weeks with media alone (w/o), 10 ng/mL TNFα, IL-1β or TGFβ-2, or metabolic inactivated. Supernatant was collected 3 times a week and stored at -20°C. In-house neo-epitope biomarkers; C1M, C3M, and active MMP-3 were assessed by ELISA. MMP-2 and -9 were detected by gelatin zymography. Knee OA cohort: A cross-sectional study including 332 subjects with knee pain; 52% women, mean age: 65 (8.4) years; BMI: 28 (4.4). Patients were divided into 4 groups based on Kellgren-Lawrence (KL) score Mild; KL 0, (n=12), Moderate: KL 1+2 (n=202); Severe: KL 3+4 (n=57); and Terminal OA (n=61). The degree of inflammation was evaluated by VAS pain and level of hsCRP. C1M and C3M were measured in serum. The study was approved by The Ethical Committee of Northern Jutland.

  Results

SME: C1M and C3M were increased (10 and 100-fold, respectively) at day 7 in response to TNFα compared to w/o (C1M: p<0.05, C3M: p<0.0001). IL-1β showed similar pattern. TGFβ-2 did not affect C1M or C3M. Activated MMP-9 and activated MMP-3 (p<0.0001) was increased in SMEs treated with IL-1β or TNFα throughout the study for SMEs, while activation of MMP-2 was not affected. Knee OA cohort: Serum C1M were significant elevated in the Terminal group compared to the Moderate (p<0.0001) and Severe OA (p<0.01) groups, but no significant difference in serum C3M. Patients with inflammatory OA (high hsCRP and pain) had a higher level of C1M and C3M as compared to non-inflammatory patients (table). There was no difference in KL between the two groups.

  Conclusion

In OA patients with the same KL score, C1M and C3M were significantly higher in those with an inflammation-driven phenotype. C1M and C3M were released from the SMEs upon treatment with the pro-inflammatory cytokines IL-1β and TNFα, but not TGFβ-2, which indicates that C1M and C3M are direct measures of synovial turnover and inflammation. This was supported by the detection of active MMP-3 and -9, which act down-stream of the cytokines and up-stream of the biomarkers. These biomarkers may be part of the identification of the important and possibly treatable inflammation-driven OA phenotype.