Background/Purpose: As biological treatments for rheumatoid arthritis (RA) are only effective in a subpopulation of patients and the treatment can have serious side effects, there is a need to identify patients who will benefit from a particular treatment. The serum Protein Fingerprint, type I collagen degradation mediated by MMP-cleavage (C1M), is a biomarker of connective tissue destruction. We investigated 1) whether baseline (BL) C1M levels were correlated with structural status, progression and pain, 2) how C1M changed with time and treatment with tocilizumab (TCZ) and 3) if C1M could be a potential early surrogate marker of treatment efficacy with TCZ.

Methods: LITHE biomarker study (n=585) was a 1-year phase III, double-blind, placebo (PBO)-controlled, parallel group study of TCZ 4 or 8 mg/kg every 4 weeks, in RA patients on stable doses of methotrexate (MTX). C1M was tested in serum from BL and week 2, 4, 16, 24 and 52. Spearman’s correlations was analysed between BL level of C1M (log transformed) and clinical measures. The associations between BL serum C1M and change in JSN and mTSS were investigated in the PBO group by regressions analysis, including CRP, age, BMI, disease duration and BL JSN/mTSS. Change in C1M levels were studied as a function of time and treatment. Lastly, in patients receiving TCZ the level of change within 1-year in structural progression and pain were investigated in two groups based of the change in C1M level until week 16: 1; <35% change in C1M and 2; >35% change in C1M and analysed by Student’s T-test.

Results: BL C1M was significantly correlated with change in JSN at week 24 (r=0.38 p<0.0001) and at week 52 (r=0.63, p<0.0001) and with changes in mTSS at 24 (r=0.21 p<0.012) and 52 weeks (r=0.58, p<0.0001). BL C1M was weakly negatively correlated with change in VAS pain at 24 weeks (r=-0.1, p<0.05). However, BL C1M was not correlated to BL JSN (rho=0.12) or mTSS (rho=0.14), but highly correlated with BL VAS pain (rho=0.3). Serum C1M was dose-dependently reduced by TCZ8 (p<0.0001) and TCZ4+MTX (p<0.05) as compared to PBO. Change in C1M at 16 weeks after treatment initiation in the TCZ groups was related to the level of radiographic changes over 1 year (p<0.05). Surprisingly, VAS pain had a negative relationship with change in C1M at 16 weeks.

Conclusion: MMP-mediated tissue degradation is imperative in joint destruction, as those patients with the highest levels of C1M were significantly more likely to progress in JSN and mTSS compared to those with a low level. Surprisingly, pain showed the reverse relationship with C1M. C1M may assist in identifying those patients that are rapid progressors. Interestingly, TCZ dose-dependently inhibited C1M level already after 2 weeks, suggesting almost immediate onset of joint protection. C1M may both be a surrogate marker of structural efficacy and enable prognostic identification of those patients that are in most need of treatment.