To enhance therapeutic development strategies for Rheumatoid arthritis (RA), we have developed and qualified a QSP model for RA. This model encompasses 10 immune cell types, 11 cytokines, and autoantibodies involved in RA pathophysiology, and incorporates the mechanisms of action for 14 RA therapies. It links drug actions on proximal biomarkers with various longitudinal clinical endpoints in RA, utilizing in vitro, ex vivo experiments and drug pharmacokinetics, receptor/target engagement, and immune/inflammatory biomarker data. The model is further calibrated using diverse aggregated and individual data from trials across major RA therapies resulting in a virtual population consisting of methotrexate-inadequate responders (MTX-IR) and TNF-inhibitor-inadequate responders (TNF-IR). This QSP model allows: (i) simultaneous characterization of reported efficacies across existing therapies, (ii) differentiation of responses among MTX-IR and TNF-IR subpopulations, and (iii) successful a priori prediction of therapies not used in model development. This work demonstrates a robust approach to QSP model development, calibration, and qualification.
