Mechanistic Based Pharmacokinetic-Pharmacodynamics models for Drug Interactions and Disease Population Predictions

Abstract: Therapeutic dose of a medication refers to the quantity of a drug required to produce a pharmacological effect without causing unacceptable adverse events. Dose selection in the clinical setting is not straight forward due to various factors, including specific patient factors such as age, sex, weight, genetic variants and renal/hepatic function, as well as external factors such as food and co-medication, all of which can influence the efficacy and safety of a drug. In silico mechanistic mathematical models, such as physiologically-based pharmacokinetic (PBPK) models, can aid dose setting by considering all these factors into a mechanistic model. The aim of this thesis was to enhance the application of PBPK modelling in drug development, with specific focus on drug-drug interactions and evaluations of genotype and disease dependencies. Additionally, the thesis assessed the implications of pre-defined PBPK platform differences.PBPK modelling, in conjunction with in vitro generated data, quantitative proteomics, pharmacodynamic models, and/or chronic kidney disease (CKD) population models, were applied to predict drug interactions in various clinically relevant cases. The model drugs used were statins, itraconazole and roxadustat. Additionally, two different PBPK platforms, Simcyp and PK-Sim, were employed throughout the different projects of this thesis providing insights on platform differences and its potential implications.The generated in vitro data and developed PBPK models provided novel insights into: 1) the relative contribution of itraconazole metabolites to the clinical drug-drug interactions, 2) the impact of statins’  distribution to muscle by transporters, 3) the critical pathways involved in the effects of disease on statins and roxadustat pharmacokinetics and drug interactions, 4) the influence of disease not only on statins’ pharmacokinetics but also on their efficacy and safety, 5) differences in software and potential implications when selecting between the Simcyp and PK-Sim platforms. These qualified PBPK models have been successfully applied to predict drug-drug, drug-gene and drug-disease interactions. They can serve as valuables tools in drug discovery and development, guiding dose-adjustments in untested scenarios, particularly in complex drug interactions. For instance, the strategy was used in lieu of clinical trials to identify a suitable dosing strategy for patients with CKD to reduce the risk of side effects or therapeutic failure of statins when combined with roxadustat.