Towards a test tube liver for drug metabolism studies

  • Brahim Achour

Student thesis: Phd


The process of in vitro-in vivo extrapolation (IVIVE) can be used to predict pharmacokinetics of drugs in patients using data from in vitro systems. This process relies on the use of experimentally obtained scaling factors, such as abundances of different drug-metabolising enzymes and microsomal protein content (MPPGL). The use of simulators is dependent on abundances and activities of pharmacokinetically relevant enzymes. The incorporation of inter-individual variability in abundances of enzymes, correlations between enzyme expression patterns, and relationships between genetic, physiological, and environmental factors and enzyme expression and activity can make predictions using IVIVE and simulations of pharmacokinetic experiments in virtual populations more accurate and realistic. Incorporation of variability and correlations can also assist in predicting extreme cases where drug therapy may be ineffective or may cause adverse effects. A meta-analysis of 52 studies was carried out to assess the reported abundances of cytochrome P450 and uridine glucuronosyltransferase (UGT) enzymes in adult Caucasian subjects. Some heterogeneity was found between studies and the weighted means and overall coefficients of variation were calculated. Some strong enzyme expression correlations were identified; CYP3A4/CYP3A5*1/*3 (rs = 0.66, p < 0.0001, n = 37), CYP3A4/CYP2C8 (rs = 0.79, p < 0.0001, n = 107), and CYP2C8/CYP2C9 (rs = 0.71, p < 0.0001, n = 72). A quantitative protocol based on targeted proteomics was used to quantify cytochrome P450 and UGT enzymes in adult liver samples (n = 24). The QconCAT standard used for quantification was successfully expressed in-house after optimisation of the expression protocol, and the utility of two strategies in expressing recalcitrant QconCAT proteins was highlighted; the use of a fusion partner and reshuffling the order of peptides in the sequence. The enzymes quantified in this study were CYP1A2, 2A6, 2B6, 2C8, 2C9, 2C18, 2D6, 2J2, 3A4, 3A5, 3A7, 3A43, and 4F2, and UGT1A1, 1A3, 1A4, 1A6, 1A9, 2B4, 2B7, and 2B15. Correlations of expression identified in the meta-analysis were confirmed and new correlations were demonstrated between UGT enzymes and between enzymes from the two families. Correlations between UGT enzymes were particularly strong and statistically significant. Relationships between enzyme expression levels and genotype, age, sex, smoking, and alcohol consumption were investigated. A significant effect of genotype on expression was seen for CYP3A5 (p < 0.0001). An overall moderate decline of expression with age was observed for all the enzymes under study; however, this relationship was not statistically significant in most cases. Gender did not have a considerable effect on expression, although some differences in expression were observed between male and female donors. Smoking seemed to induce the expression of all enzymes; however, statistically significant induction was demonstrated only in the cases of CYP2A6, CYP3A4, CYP3A7, and UGT1A1 (p < 0.05). Alcohol consumption was not shown to have a considerable effect on enzyme expression.Two pig livers were used to optimise some aspects of the experimental protocol including solubilisation and digestion of proteins. Pig MPPGL was measured and relative hepatic contents of drug-metabolising cytochrome P450 enzymes in pig liver were established using label-free quantification.
Date of Award31 Dec 2013
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorJill Barber (Supervisor) & Amin Rostami-Hochaghan (Supervisor)


  • Cytochrome P450, UGT. abundance, correlation, drug metabolism, variability

Cite this