Systems model identifies baseline cytokine concentrations as potential predictors of rheumatoid arthritis inflammatory response to biologics.

Abstract

Circulating cytokines are central pathological mediators of inflammatory autoimmune diseases like rheumatoid arthritis. Immunological diversity in patients might contribute to inadequate responses to biological drugs. To address this therapeutic challenge, we developed a mathematical model that simultaneously describes temporal patterns of drug disposition for several biologics and their corresponding targeted cytokines, which were linked to triggering inflammatory responses. A modelling framework was applied to rheumatoid arthritis-relevant cytokines regulating C-reactive protein (CRP) as an inflammatory marker. Clinical data were extracted from the literature for anakinra, canakinumab, infliximab, secukinumab and tocilizumab, along with their corresponding cytokines, interleukin-1 receptor antagonist, IL-1β, tumour necrosis factor α (TNFα), IL-17A and IL-6 receptor (IL-6R). Based on prior knowledge of regulatory mechanisms, cytokines were integrated with CRP profiles. The model well captured all serum concentration-time profiles of cytokines and CRP ratios to respective baselines following drug treatment with good precision. On external validation, reasonable model performance on CRP dynamics, including rebound effects, was confirmed with clinical data not used in model development. Model-based simulations demonstrated that serum infliximab concentrations were accurately recapitulated in both a dose- and baseline TNFα-dependent manner. Furthermore, high baseline profiles of both IL-1β and/or targeted cytokines could be predictors of poor responses to biologics targeting TNFα and IL-6R, although the impact of IL-1β must be carefully interpreted. Our model provides a quantitative platform to guide targeting and dosing strategies, including combination and/or sequential therapy, according to distinct baseline cytokine patterns in rheumatoid arthritis patients.