Understanding the Impact of Pharmacogenetic Differences in Drug-Drug Interactions (DDIs): A Model-Based Approach to Predict Differences in Drug Exposure

Abstract

Genetic diversity between people can lead to differences in their capabilities to metabolize drugs. These pharmacogenomic differences induce variability in drug exposure (amount of drug in our bodies), which can cause both positive and negative effects. Another important contributor to differences in exposure occurs when one drug interacts with and alters how another drug is eliminated (drug-drug interactions, DDIs). Although the effects of pharmacogenetics or DDIs alone on drug exposure are well understood, a major gap in knowledge is how pharmacogenetic diversity might influence DDIs. Cytochrome P450s (CYP) are important superfamily of enzyme which includes many critical drug metabolism enzymes. CYP2C9 is an important drug metabolism enzyme in the CYP 2C subfamilies, which mediate 10-20% of the drugs that undergo CYP mediated metabolism. Several clinically important drugs including warfarin, phenytoin and flurbiprofen are metabolized through CYP2C9. Loss-of-function variants on gene encoding CYP2C9, such as *2 and *3, introduce considerable variabilities in the exposure and response of these drugs. Additionally, since CYP2C9 is inhibitable and inducible, the administration of these drugs together with CYP inhibitor and inducers will cause DDIs, which further complicate their dosing clinically. Two clinical trials sponsored by National Institute of Health (NIH) were conducted to investigate the extent of DDIs in the presence of CYP2C9 genotypes, using warfarin, flurbiprofen, ketoprofen and tolbutamide as probe drugs and fluconazole (a CYP inhibitor) and rifampin (a CYP inducer) as interacting drugs. Drug concentration data of probe drugs and their major metabolites in plasma and urine were collected for subsequent analysis. The aim of this thesis is to quantify the extent of DDIs in the presence of CYP2C9 genetic variants using the collected clinical PK data via mathematical modeling approach. Population PK models were first constructed for each probe drugs including pharmacogenomic differences and drug interactions as covariates. Model-based analysis were then performed based on parameter estimations. Physiology-based PK (PBPK) modeling was also performed to understand the genotype-dependent DDIs from a bottom-up prospective. The developed models will be valuable in quantifying genotype-dependent DDIs clinically, thus achieving another step toward precision medicine.