The application of mechanistic modelling as a tool in drug disposition and risk assessment

Abstract

New molecular entities entering the pharmaceutical market are required to adhere to stringent safety, efficacy and quality requirements that often lead to delays in the early-phases of drug development. Pharmacokinetic modelling approaches, such as physiologically-based pharmacokinetic (PBPK) modelling, can cater to most of the critical PK issues and at the same time optimise the utilising of resources. The overall aim of this work was to illustrate, explore and facilitate the application of PBPK modelling in the context of drug disposition and risk assessment. In Chapter 2 of this thesis, we illustrated the concept of developing customisable pharmacokinetic models through the development of a region-specific CNS PBPK model to assess the rodent hippocampus and frontal cortex pharmacokinetics using MATLAB. We then extrapolated the model to predict human regional brain pharmacokinetics, using morphine as a case study for comparison. This successfully proposed a simplified first-principle approach to the development of a regional brain central nervous system (CNS) PBPK model. This approach has significant implications for assessing drug disposition across the human CNS and provides an opportunity for exploring the relationship between regional brain drug concentration, pharmacodynamics effects, and interspecies extrapolation. In Chapter 3 of this thesis, our goal was to develop a population-based PBPK modelling that could explore the potential risk of drug-drug interactions (DDIs) in adults and paediatric populations.We developed a model capable of predicting the impact of efavirenz-mediated DDIs on thepharmacokinetics of the antimalarial drug lumefantrine in Ugandan paediatric population groups,whilst also accounting for the polymorphic nature of CYP2B6. We demonstrated that an extension of the current artemether-lumefantrine treatment regimen from 3-days to 7-days would counteract the reduction in efavirenz metabolism common with the *6/*6 genotype and hence enhance the attainment of the target day-7 lumefantrine concentration in both *1/*1 and *6/*6 genotype groups, thereby reducing the risk of malaria parasite recrudescence. This study demonstrated the capability of PBPK modelling in predicting PK profiles in special population such as paediatrics and dealing with complex DDIs associated with genotype specific effects.  The final part of this work, Chapter 4, focussed on demonstrating the capability of PBPK modelling in addressing inter-ethnicity variability and risk assessments within a mixed population group. We explored the application of PBPK models for specific population data analyses in the context of CYP2C19 polymorphism on clopidogrel in the multi-ethnic populations of Malaysia.We demonstrated a statistically significant difference in the peak concentrations of the active metabolite, clopi-H4, between the extensive metaboliser (EM) and poor metaboliser (PM) phenotypes with either Malay or Malaysian Chinese population groups. The study directly addresses this inter-ethnicity variability and provide a research tool that brings together the complexity of systems-biology with the ease-of-use applicability of pharmacokinetic modelling to provide a robust predictive platform which can easily be adapted and developed as required within a population.