Animal models of disease are central to discovery of new RA therapies. Their utility is limited, however, by the fact that no model recapitulates all aspects of human disease. Disease activity biomarkers common to animal models and human RA could be used to focus research effort on those molecules most likely to offer clinical benefit, significantly expediting drug development. The goal of this study was to use high throughput proteomics technology to identify peripheral blood biomarkers common to murine models of RA and human disease.
We used multiplex ELISA to assay inflammatory biomolecules associated with RA pathophysiology in mouse model and RA patient sera. RT-PCR was used to measure gene expression in animal joint tissues. In mice (N=31) with anticollagen antibody-induced arthritis (AbIA), disease activity was assessed using a visual scoring system (scale 0-4). Human sera were from 243 patients (82% female, mean age 58, 89% Caucasian) with a mean RA disease duration of 15.3 years. DMARDs were used by 67% of patients (MTX 34%; TNFα inhibitors 22%; both treatments 12%). RA disease activity was quantified using DAS28-CRP.
In mice with AbIA, serum MMP-3 increased 6-fold (p<0.001), and IL-6 8-fold (p<0.001) at peak disease activity. Both markers correlated strongly with arthritis scores across animals: MMP-3 R2=0.84, p<0.001; IL-6 R2=0.5, p<0.01. In mouse joints, MMP-3 RNA level increased up to 7.5-fold (p<0.001) and IL-6 up to 76-fold (p<0.001) during disease progression.
Analysis of patient sera showed both MMP-3 (OR=3.0, p<0.0001) and IL-6 (OR=1.7, p=0.001) to significantly correlate with disease activity as assessed by DAS28-CRP. These findings were confirmed in an independent patient cohort (N=259). Other proteins correlated with RA activity in patients, including TNFR1 (p<0.01), MMP-1 (p<0.01), OPG (p=0.01), IL2R (p=0.01), and TNFR2 (p=0.01); these markers are now leading candidates for cross-validation in animal models.
We have demonstrated that serum levels of MMP-3 and IL-6 are similarly and significantly associated with disease activity in a murine arthritis model and in human RA, and have prioritized five additional human markers for validation in animal systems. These biomarkers hold potential to be a quantitative and objective means of relating disease activity assessment in pre-clinical studies to potential therapeutic impact in human clinical trials.
E.S. Izmailova, Millennium Pharmaceuticals, 3; M.D. Pickard, Millennium Pharmaceuticals, 3; J. Narang, Millennium Pharmaceuticals, 3; N. Paz, Millennium Pharmaceuticals, 3; N.E. Maher, None; N.A. Shadick, Millennium Pharmaceuticals, 5; M.E. Weinblatt, Millennium Pharmaceuticals, 2; Millennium Pharmaceuticals, 5; R. Roubenoff, Millennium Pharmaceuticals, 3; A. Parker, Millennium Pharmaceuticals, 3.