A comparative computational analysis of IFN-alpha pharmacokinetics and its induced cellular response in mice and humans

Drug effects are difficult to investigate in detail in vivo. However, a mechanistic understanding of drug action is clearly beneficial for both pharmaceutical development as well as for optimization of treatment designs. We here established a quantitative systems pharmacology (QSP) mouse model which simultaneously describes whole-body pharmacokinetics of murine IFN-α as well as the cellular pharmacodynamic effect through the antiviral response biomarker Mx2. To this end, a dynamic model of intracellular IFN-α signalling in the JAK/STAT pathway was combined with a whole-body physiologically-based pharmacokinetic model of IFN-α in mice. The pharmacodynamic behaviour of the resulting mouse IFN-α QSP model was first compared to a cellular model of the JAK/STAT pathway to compare in vitro and in vivo drug effects and to identify functional differences. It was found that the in vitro drug effect in the cellular model overestimates the in vivo response in mice at least by a factor of two which is partly due to the missing drug clearance in vitro. Also, the drug responses in the in vitro model were time delayed. Interspecies analyses in murine and a previously published human QSP model of IFN-α next show a similar dynamic behavior. However, our models demonstrate eight to 16-fold stronger response levels in mice than in humans due to more efficient interferon binding. Our analysis supports a mechanistic analysis of both upstream pharmacokinetic as well as downstream pharmacodynamic drug effects through the combination of physiological knowledge and quantitative computational models. The study hence shows potential applications for QSP modelling in terms of study planning, for example by choosing physiologically relevant in vitro concentrations. Also, the QSP model allows inter-species comparisons of the effect strength in specific functional readouts, which in humans are otherwise not possible due to the limited sampling possibilities. We expect QSP modelling to play an increasingly important role in drug development and research in the future.