MRI of the joint and evaluation of the granulocyte–macrophage colony-stimulating factor–CCL17 axis in patients with rheumatoid arthritis receiving otilimab: a phase 2a randomised mechanistic study

BACKGROUND:Otilimab is a human monoclonal antibody that inhibits granulocyte-macrophage colony-stimulating factor (GM-CSF), a driver in many immune-mediated inflammatory conditions. We evaluated the effect of otilimab on the GM-CSF-chemokine (C-C motif) ligand 17 (CCL17) axis and synovitis in patients with rheumatoid arthritis. METHODS:This phase 2a, randomised, double-blind, multicentre, placebo-controlled, parallel-group study was done at nine sites across the USA, Poland, and Germany. Patients aged 18 years or older with rheumatoid arthritis per American College of Rheumatology-European League Against Rheumatism 2010 criteria and receiving stable methotrexate were randomly assigned (3:1) by an interactive response technology system to either subcutaneous otilimab 180 mg or placebo once weekly for 5 weeks, then every other week until week 10 (within a 12-week treatment period), followed by a 10-week safety follow-up. Randomisation was stratified by early rheumatoid arthritis (≤2 years since diagnosis) and established rheumatoid arthritis (>2 years since diagnosis). Patients and study personnel (except for an unblinded coordinator or nurse who prepared and administered the study drug) were blinded to treatment assignment; the syringe was shielded during administration. Patients were enrolled by study investigators and allocated to a treatment by central randomisation on the basis of a schedule generated by the sponsor. The primary endpoint was change over time (assessed at baseline and weeks 1, 2, 4, 6, 8, 12, and 22 of follow-up) in 112 biomarkers, including target engagement biomarkers and those that may be indicative of rheumatoid arthritis disease activity and response to otilimab. Secondary endpoints were change from baseline in synovitis, osteitis and erosion assessed by rheumatoid arthritis MRI scoring system (RAMRIS) and rheumatoid arthritis MRI quantitative score (RAMRIQ), and safety evaluation. The primary, secondary, and safety endpoints were assessed in the intention-to-treat population. Biomarker and MRI endpoints were analysed for differences between treatment groups using a repeated measures model. This study is registered with ClinicalTrials.gov, NCT02799472. FINDINGS:Between Aug 9, 2016, and Oct 30, 2017, 39 patients were randomly assigned and included in the analysis (otilimab n=28; placebo n=11). In the otilimab group, mean serum concentrations of GM-CSF-otilimab complex peaked at week 4 (138·4 ng/L, 95% CI 90·0-212·9) but decreased from week 6-12. CCL17 concentrations decreased from baseline to week 1, remained stable to week 8, and returned to baseline at week 12; least-squares mean ratio to baseline was 0·65 (95% CI 0·49-0·86; coefficient of variation 13·60) at week 2, 0·68 (0·53-0·88; 12·51) at week 4, 0·78 (0·60-1·00; 12·48) at week 6, and 0·68 (0·54-0·85; 11·21) at week 8. No meaningful change in CCL17 concentrations was observed with placebo. In the otilimab group, the least-squares mean ratio to baseline in MMP-degraded type I collagen was 0·86-0·91 over weeks 1-8, returning to baseline at week 12; concentrations remained above baseline at all timepoints in the placebo group. There were no observable differences between otilimab and placebo for all other biomarkers. At week 12, least-squares mean change in RAMRIS synovitis score from baseline was -1·3 (standard error [SE] 0·6) in the otilimab group and 0·8 (1·2) with placebo; RAMRIQ synovitis score showed a least-squares mean change from baseline of -1417·0 μl (671·5) in the otilimab group and -912·3 μl (1405·8) with placebo. Compared with placebo, otilimab did not show significant reductions from baseline to week 12 in RAMRIS synovitis, osteitis and bone erosion, or in RAMRIQ synovitis and erosion damage. Adverse events were reported in 11 (39%) of 28 otilimab-treated and four (36%) of 11 placebo-treated patients, most commonly cough in the otilimab group (2 [7%] of 28; not reported in placebo group), and pain in extremity (four [36%] of 11) and rheumatoid arthritis (two [18%] of 11) in the placebo group (not reported in otilimab group). There were no serious adverse events or deaths. INTERPRETATION:Serum concentrations of GM-CSF-otilimab complex indicated that target engagement was achieved with initial weekly dosing, but not sustained with every other week dosing. CCL17 might be a pharmacodynamic biomarker for otilimab activity in future studies. Otilimab was well tolerated and, despite suboptimal exposure, showed some evidence for improved synovitis over 12 weeks in patients with active rheumatoid arthritis. FUNDING:GlaxoSmithKline.