Metabolism matters
Dr Hadwen Trust Research Fellowship 2002 — 2004
Comparison between allometric scaling and in vitro-in vivo extrapolation to predict human xenobiotic clearance
N Proctor, A Rostami-Hodjegan and G Tucker, Molecular Pharmacology and Pharmacogenetics, Sheffield University
The world’s most powerful drug agency, the US Food and Drug Administration (FDA), has pointed out that 92% of novel drugs entering clinical trials never reach the market, failing mainly because of safety or efficacy problems [1]. The absorption, distribution, metabolism and excretion (ADME) characteristics of a drug are key influences on safety and efficacy and, as the FDA says, “… in the past, failure to predict unfavorable human metabolism of candidate drugs has led to costly failures in the clinic as well as multiple drug market withdrawals.”
Traditionally, the prediction of the ADME properties of drugs and chemicals has relied on tests involving several hundred animals including rodents, dogs and primates. As models they are neither fully representative of the human population nor of the individuals most likely to suffer adverse side effects; most obviously because of species differences, for example in the metabolic enzymes involved.
The results of animal tests are extrapolated to humans by means of allometric scaling, based on relationships between human and animal body weights, body surface areas, lifespans or similar parameters. Clearance is not predicted very well by basic allometric scaling, the error between predicted and observed clearance being greater than 30% in most cases [2]. Allometric scaling has been criticised for its empirical nature, and is affected by interspecies differences in metabolism. It also only provides point estimates for a homogeneous population, obscuring the important genetic polymorphisms found in individual humans.
In 2002 the Dr Hadwen Trust awarded a grant for a study of the predictive accuracy for drug clearance of conventional allometric scaling compared with in vitro/in vivo extrapolation (IVIVE) using Simcyp. Simcyp was originally developed to allow prediction of drug-drug interactions and aim of this research project was to adapt it to predict human drug clearances and thereby replace some animal tests.
First, 13 drugs were selected from the literature for which enzyme kinetic data and human pharmacokinetic parameters were available. Simcyp was used to produce 200 age- and gender-matched simulations for each compound and this showed that Simcyp predictions of drug clearance in humans were generally accurate and the technique showed great promise.
Next, the team compiled published in vitro enzyme kinetic data from recombinant human cytochrome P450 systems and human liver microsomes for a number of drugs for which human pharmacokinetic parameters were also available. Using statistical tools, intersystem extrapolation factors were derived to greatly improve the accuracy of predictions of human drug clearances [3].
Simcyp software was used to predict the clearances of 15 drugs on the basis of in vitro metabolism data and incorporating inter-individual variability. This study showed that predicted values of median clearance were within a two-fold range for 93% of drugs given orally and 100% when given intravenously [4].
Of these 15 drugs, 11 had data from at least three animal species for oral and intravenous clearance. Simcyp extrapolations from in vitro to in vivo were compared with allometric scaling. The results suggested that for drugs mainly metabolised by cytochrome P450 enzymes, human in vitro-in vivo extrapolation is more reliable than allometric scaling from animal tests, and can be used to assess the likely variability in clearance.
The authors concluded that in drug development, “…once [in vitro] data become available (particularly information on in vitro drug metabolism in human systems), the need for further in vivo or in vitro animal studies may be minimised by the appropriate application of IVIVE based purely on human data”. They also suggested that the place of allometric scaling of animal data in drug development should be more closely scrutinised [5].
Combining IVIVE with clinical trial simulations is another useful approach. It can be used to investigate, in silico, differences in the pharmacokinetics and pharmacodynamics of a drug in people who are extensive or poor metabolisers.
The team has developed user-friendly software and databases specifically designed to predict drug absorption, clearance, distribution and metabolic drug-drug interactions from in vitro information. By simulating pharmacokinetics in virtual patient populations and identifying individuals at high risk, Simcyp can accelerate drug discovery and development. Simcyp is now used by many global pharmaceutical companies, leading academic institutes and regulatory authorities [6].
The Simcyp scientists are members of the BioSim network of excellence, supported by the European Commission to show how modern simulation techniques can lead to more rational drug development and better treatments, and reduce the need for animal experiments [7].
References and information
1. Food and Drug Administration (2004). Innovation and Stagnation: Challenge and Opportunity on the Critical Path to New Medical Products. See website: www.fda.gov/oc/initiatives/criticalpath/
2. Mahmood I & Balian JD (1999). The pharmacokinetic principles behind scaling from preclinical results to phase I protocols. Clin Pharmacokinet 36:1-11.
3. Proctor NJ, Tucker GT & Rostami-Hodjegan A (2004). Predicting drug clearance from recombinantly expressed CYPs: intersystem extrapolation factors. Xenobiotica 34:151-178.
4. Howgate EM, Rowland-Yeo K, Proctor NJ et al (2006). Prediction of in vivo drug clearance from in vitro data. I: impact of inter-individual variability. Xenobiotica 36:473-497.
5. Shiran MR, Proctor NJ, Howgate EM et al (2006). Prediction of metabolic drug clearance in humans: in vitro-in vivo extrapolation vs allometric scaling. Xenobiotica 36:567-580.
Professor Geoff Tucker is Head of the Academic Unit of Clinical Pharmacology at Sheffield University and Chair of Simcyp Ltd. Amin Rostami-Hodjegan is a Reader in clinical pharmacokinetics and drug metabolism at Sheffield University and also the Director of R&D at Simcyp Ltd.