Towards a Best Practice Approach in PBPK Modeling: Case Example of Developing a Unified Efavirenz Model Accounting for Induction of CYPs 3A4 and 2B6.

In this study, we present efavirenz physiologically based pharmacokinetic (PBPK) model development as an example of our best practice approach that uses a stepwise approach to verify the different components of the model. First, a PBPK model for efavirenz incorporating in vitro and clinical pharmacokinetic (PK) data was developed to predict exposure following multiple dosing (600 mg q.d.). Alfentanil i.v. and p.o. drug-drug interaction (DDI) studies were utilized to evaluate and refine the CYP3A4 induction component in the liver and gut. Next, independent DDI studies with substrates of CYP3A4 (maraviroc, atazanavir, and clarithromycin) and CYP2B6 (bupropion) verified the induction components of the model (area under the curve [AUC] ratios within 1.0-1.7-fold of observed). Finally, the model was refined to incorporate the fractional contribution of enzymes, including CYP2B6, propagating autoinduction into the model (Racc 1.7 vs. 1.7 observed). This validated mechanistic model can now be applied in clinical pharmacology studies to prospectively assess both the victim and perpetrator DDI potential of efavirenz.

Metformin and cimetidine: Physiologically based pharmacokinetic modelling to investigate transporter mediated drug-drug interactions.

Metformin is used as a probe for OCT2 mediated transport when investigating possible DDIs with new chemical entities. The aim of the current study was to investigate the ability of physiologically-based pharmacokinetic (PBPK) models to simulate the effects of OCT and MATE inhibition by cimetidine on metformin kinetics. PBPK models were developed, incorporating mechanistic kidney and liver sub-models for metformin (OCT and MATE substrate) and a mechanistic kidney sub-model for cimetidine. The models were used to simulate inhibition of the MATE1, MATE2-K, OCT1 and OCT2 mediated transport of metformin by cimetidine. Assuming competitive inhibition and using cimetidine Ki values determined in vitro, the predicted metformin AUC ratio was 1.0 compared to an observed value of 1.46. The observed AUC ratio could only be recovered with this model when the cimetidine Ki for OCT2 was decreased 1000-fold or the Ki's for both OCT1 and OCT2 were decreased 500-fold. An alternative description of metformin renal transport by OCT1 and OCT2, incorporating electrochemical modulation of the rate of metformin uptake together with 8-18-fold decreases in cimetidine Ki's for OCTs and MATEs, allowed recovery of the extent of the observed effect of cimetidine on metformin AUC. While the final PBPK model has limitations, it demonstrates the benefit of allowing for the complexities of passive permeability combined with active cellular uptake modulated by an electrochemical gradient and active efflux.

A mechanistic framework for in vitro-in vivo extrapolation of liver membrane transporters: prediction of drug-drug interaction between rosuvastatin and cyclosporine.

BACKGROUND AND OBJECTIVES: The interplay between liver metabolising enzymes and transporters is a complex process involving system-related parameters such as liver blood perfusion as well as drug attributes including protein and lipid binding, ionisation, relative magnitude of passive and active permeation. Metabolism- and/or transporter-mediated drug-drug interactions (mDDIs and tDDIs) add to the complexity of this interplay. Thus, gaining meaningful insight into the impact of each element on the disposition of a drug and accurately predicting drug-drug interactions becomes very challenging. To address this, an in vitro-in vivo extrapolation (IVIVE)-linked mechanistic physiologically based pharmacokinetic (PBPK) framework for modelling liver transporters and their interplay with liver metabolising enzymes has been developed and implemented within the Simcyp Simulator(®). METHODS: In this article an IVIVE technique for liver transporters is described and a full-body PBPK model is developed. Passive and active (saturable) transport at both liver sinusoidal and canalicular membranes are accounted for and the impact of binding and ionisation processes is considered. The model also accommodates tDDIs involving inhibition of multiple transporters. Integrating prior in vitro information on the metabolism and transporter kinetics of rosuvastatin (organic-anion transporting polypeptides OATP1B1, OAT1B3 and OATP2B1, sodium-dependent taurocholate co-transporting polypeptide [NTCP] and breast cancer resistance protein [BCRP]) with one clinical dataset, the PBPK model was used to simulate the drug disposition of rosuvastatin for 11 reported studies that had not been used for development of the rosuvastatin model. RESULTS: The simulated area under the plasma concentration-time curve (AUC), maximum concentration (C max) and the time to reach C max (t max) values of rosuvastatin over the dose range of 10-80 mg, were within 2-fold of the observed data. Subsequently, the validated model was used to investigate the impact of coadministration of cyclosporine (ciclosporin), an inhibitor of OATPs, BCRP and NTCP, on the exposure of rosuvastatin in healthy volunteers. CONCLUSION: The results show the utility of the model to integrate a wide range of in vitro and in vivo data and simulate the outcome of clinical studies, with implications for their design.

Evaluation of an In Silico PBPK Post-Bariatric Surgery Model through Simulating Oral Drug Bioavailability of Atorvastatin and Cyclosporine.

An increasing prevalence of morbid obesity has led to dramatic increases in the number of bariatric surgeries performed. Altered gastrointestinal physiology following surgery can be associated with modified oral drug bioavailability (Foral). In the absence of clinical data, an indication of changes to Foral via systems pharmacology models would be of value in adjusting dose levels after surgery. A previously developed virtual "post-bariatric surgery" population was evaluated through mimicking clinical investigations on cyclosporine and atorvastatin after bariatric surgery. Cyclosporine simulations displayed a reduced fraction absorbed through gut wall (fa) and Foral after surgery, consistent with reported observations. Simulated atorvastatin Foral postsurgery was broadly reflective of observed data with indications of counteracting interplay between reduced fa and an increased fraction escaping gut wall metabolism (FG). Inability to fully recover observed atorvastatin exposure after biliopancreatic diversion with duodenal switch highlights the current gap regarding the knowledge of associated biological changes.CPT: Pharmacometrics & Systems Pharmacology (2013) 2, e47; doi:10.1038/psp.2013.23; advance online publication 12 June 2013.

Prediction of time-dependent CYP3A4 drug-drug interactions by physiologically based pharmacokinetic modelling: impact of inactivation parameters and enzyme turnover.

Predicting the magnitude of time-dependent metabolic drug-drug (mDDIs) interactions involving cytochrome P-450 3A4 (CYP3A4) from in vitro data requires accurate knowledge of the inactivation parameters of the inhibitor (K(I), k(inact)) and of the turnover of the enzyme (k(deg)) in both the gut and the liver. We have predicted the magnitude of mDDIs observed in 29 in vivo studies involving six CYP3A4 probe substrates and five mechanism based inhibitors of CYP3A4 of variable potency (azithromycin, clarithromycin, diltiazem, erythromycin and verapamil). Inactivation parameters determined anew in a single laboratory under standardised conditions together with data from substrate and inhibitor files within the Simcyp Simulator (Version 9.3) were used to determine a value of the hepatic k(deg) (0.0193 or 0.0077h(-1)) most appropriate for the prediction of mDDIs involving time-dependent inhibition of CYP3A4. The higher value resulted in decreased bias (geometric mean fold error - 1.05 versus 1.30) and increased precision (root mean squared error - 1.29 versus 2.30) of predictions of mean ratios of AUC in the absence and presence of inhibitor. Depending on the k(deg) value used (0.0193 versus 0.0077h(-1)), predicted mean ratios of AUC were within 2-fold of the observed values for all (100%) and 27 (93%) of the 29 studies, respectively and within 1.5-fold for 24 (83%) and 17 (59%) of the 29 studies, respectively. Comprehensive PBPK models were applied for accurate assessment of the potential for mDDIs involving time-dependent inhibition of CYP3A4 using a hepatic k(deg) value of 0.0193h(-1) in conjunction with inactivation parameters determined by the conventional experimental approach.

Prediction of in vivo drug clearance from in vitro data. II: potential inter-ethnic differences.

Potential differences in drug clearance between Japanese and Caucasians were investigated by integrating data on demography, liver size, the abundance of the major cytochromes P450 and in vitro metabolic parameters. Eleven drugs (alprazolam, caffeine, chlorzoxazone, cyclosporine, midazolam, omeprazole, sildenafil, tolbutamide, triazolam, S-warfarin and zolpidem) fulfilled the entry criteria of the study (i.e. the necessary in vitro metabolism data were available and clearance values had been reported both in Caucasians and Japanese). Values of relevant biological variables were obtained from the literature, and clearance predictions were made using the Simcyp Population-Based ADME Simulator. The ratios of observed oral clearance (CLp.o.) values in Caucasians compared with Japanese ranged from 0.6 to 2.8 (integrating data from 82 sources). The CLp.o. values for alprazolam, caffeine and zolpidem were not statistically different between Caucasian and Japanese (p>0.05), whereas those for chorzoxazone, cyclosporine, omeprazole, tolbutamide and triazolam were higher in Caucasians (p<0.05), and those for midazolam, sildenafil and S-warfarin were higher in Japanese (p<0.05). CLp.o. values, predicted from in vitro data, were within 3-fold of observed in vivo values for seven of the 11 drugs in Japanese. Values for the predicted ratios ranged from 1.6 to 4.9. The predicted ratios were not significantly different from observed ratios for cyclosporine, omeprazole, tolbutamide and triazolam. Only partial success in predicting ethnic differences in clearance indicates the need for larger and more reliable databases on relevant variables. With such information, in silico predictions might be used with more confidence to decrease the need for repeating pharmacokinetic studies in different ethnic groups.

Prediction of metabolic drug clearance in humans: in vitro-in vivo extrapolation vs allometric scaling.

Previously in vitro-in vivo extrapolation (IVIVE) with the Simcyp Clearance and Interaction Simulator has been used to predict the clearance of 15 clinically used drugs in humans. The criteria for the selection of the drugs were that they are used as probes for the activity of specific cytochromes P450 (CYPs) or have a single CYP isoform as the major or sole contributor to their metabolism and that they do not exhibit non-linear kinetics in vivo. Where data were available for the clearance of the drugs in at least three animal species, the predictions from IVIVE have now been compared with those based on allometric scaling (AS). Adequate data were available for estimating oral clearance (CLp.o.) in 9 cases (alprazolam, sildenafil, caffeine, clozapine, cyclosporine, dextromethorphan, midazolam, omeprazole and tolbutamide) and intravenous clearance in 6 cases (CLi.v.) (cyclosporine, diclofenac, midazolam, omeprazole, theophylline and tolterodine). AS predictions were based on five different methods: (1) simple allometry (clearance versus body weight); (2) correction for maximum life-span potential (CL x MLP); (3) correction for brain weight (CL x BrW); (4) the use of body surface area; and (5) the rule of exponents. A prediction accuracy was indicated by mean-fold error and the Pearson product moment correlation coefficient. Predictions were considered successful if the mean-fold error was

Prediction of in vivo drug clearance from in vitro data. I: impact of inter-individual variability.

The Simcyp Population-Based ADME Simulator was used to predict median drug clearances and their associated variance from in vitro data. Fifteen drugs satisfied the entry criteria for the study and the relevant information (in vitro metabolism data and in vivo human clearance values) were collated from the literature. Predicted values of median clearances fell within 2-fold of observed values for 73% of the drugs (oral route) and 78% of the drugs (intravenous route) when microsomal binding was disregarded, and for 93% (oral) and 100% (intravenous) when it was considered. Irrespective of whether microsomal binding was considered, the predicted fold variability fell within 2-fold of the observed variability for 80% (oral) and 67% (intravenous) of the drugs.

A critical evaluation of the experimental design of studies of mechanism based enzyme inhibition, with implications for in vitro-in vivo extrapolation.

The published literature on mechanism based inhibition (MBI) of CYPs was evaluated with respect to experimental design, methodology and data analysis. Significant variation was apparent in the dilution factor, ratio of preincubation to incubation times and probe substrate concentrations used, and there were some anomalies in the estimation of associated kinetic parameters (k(inact), K(I), r). The impact of the application of inaccurate values of k(inact) and K(I) when extrapolating to the extent of inhibition in vivo is likely to be greatest for those compounds of intermediate inhibitory potency, but this also depends on the fraction of the net clearance of substrate subject to MBI and the pre-systemic and systemic exposure to the inhibitor. For potent inhibitors, the experimental procedure is unlikely to have a material influence on the maximum inhibition. Nevertheless, the bias in the values of the kinetic parameters may influence the time for recovery of enzyme activity following re-synthesis of the enzyme. Careful attention to the design of in vitro experiments to obtain accurate kinetic parameters is necessary for a reliable prediction of different aspects of the in vivo consequences of MBI. The review calls for experimental studies to quantify the impact of study design in studies of MBI, with a view to better harmonisation of protocols.

Lipid lowering in patients with diabetes mellitus: what coronary heart disease risk threshold should be used?

OBJECTIVE: To examine the impact for the UK population of providing statin treatment for diabetic patients for the primary prevention of coronary heart disease at a coronary event risk lower than currently recommended by the National Service Framework (NSF) for coronary heart disease. DESIGN: Cross sectional survey. SETTING: England 1998. PARTICIPANTS: Nationally representative sample of 6879 subjects aged 35-74 years living in private households. MAIN OUTCOME MEASURES: The proportion of the UK population recommended for statin treatment according to the NSF for coronary heart disease, and the proportion of the population with diabetes at a coronary disease event risk of > or = 15% over 10 years. RESULTS: Of the 6879 subjects with total cholesterol measurements, 218 (3.2%) had diabetes mellitus. In this nationally representative sample, 6.3% of the subjects (95% confidence interval (CI), 5.7% to 6.9%) were candidates for statin treatment for the secondary prevention of coronary heart disease, including 0.7% (95% CI 0.5% to 0.9%) with diabetes. A further 2.4% (95% CI 2.0% to 2.8%), including 0.4% (0.2% to 0.6%) with diabetes, were identified as candidates for primary prevention of coronary heart disease according to the NSF for coronary heart disease. Lowering the primary prevention threshold for statin treatment to a coronary event risk of > or = 15% over 10 years in diabetic patients identified an additional 0.5% of the population. CONCLUSIONS: Extending statin treatment to diabetic patients at a coronary heart disease risk of > or = 15% over 10 years would have a relatively small numerical impact in the UK population. Thus patients with diabetes mellitus should, as a minimum, be targeted for statin treatment at this level of risk.

Coronary and cardiovascular risk estimation for primary prevention: validation of a new Sheffield table in the 1995 Scottish health survey population.

OBJECTIVE: To examine the accuracy of a new version of the Sheffield table designed to aid decisions on lipids screening and detect thresholds for risk of coronary heart disease needed to implement current guidelines for primary prevention of cardiovascular disease. DESIGN: Comparison of decisions made on the basis of the table with absolute risk of coronary heart disease or cardiovascular disease calculated by the Framingham risk function. The decisions related to statin treatment when coronary risk is >/=30% over 10 years; aspirin treatment when the risk is >/=15% over 10 years; and the treatment of mild hypertension when the cardiovascular risk is >/=20% over 10 years. SETTING: The table is designed for use in general practice. SUBJECTS: Random sample of 1000 people aged 35-64 years from the 1995 Scottish health survey. MAIN OUTCOME MEASURES: Sensitivity, specificity, and positive and negative predictive values of the table. RESULTS: 13% of people had a coronary risk of >/=15%, and 2. 2% a risk of >/=30%, over 10 years. 22% had mild hypertension (systolic blood pressure 140-159 mm Hg). The table indicated lipids screening for everyone with a coronary risk of >/=15% over 10 years, for 95% of people with a ratio of total cholesterol to high density lipoprotein cholesterol of >/=8.0, but for <50% with a coronary risk of <5% over 10 years. Sensitivity and specificity were 97% and 95% respectively for a coronary risk of >/=15% over 10 years; 82% and 99% for a coronary risk of >/=30% over 10 years; and 88% and 90% for a cardiovascular risk of >/=20% over 10 years in mild hypertension. CONCLUSION: The table identifies all high risk people for lipids screening, reduces screening of low risk people by more than half, and ensures that treatments are prescribed appropriately to those at high risk, while avoiding inappropriate treatment of people at low risk.

Evidence that serine 304 is not a key ligand-binding residue in the active site of cytochrome P450 2D6.

Homology models of cytochrome P450 2D6 (CYP2D6) have identified serine 304 as an active-site residue and implicated a putative role for this residue in substrate enantioselectivity and the differential inhibition of enzyme activity by the diastereoisomers quinine and quinidine. The role of serine 304 in selectivity is thought to be achieved through a preferential hydrogen-bond interaction between the hydroxyl group of the residue and one of the stereoisomers of each ligand. We have tested this hypothesis by substituting serine 304 with alanine, a non-hydrogen-bonding residue, and compared the properties of the wild-type and mutant enzymes in microsomes prepared from yeast cells expressing the appropriate cDNA-derived enzyme. The Ser(304)Ala substitution did not alter the enantioselective oxidation of metoprolol; the O-demethylation reaction remained R-(+)-enantioselective (wild-type, R/S, 1.7; mutant, R/S, 1.6), whereas alpha-hydroxylation remained S-(-)-enantioselective (wild-type and mutant, R/S, 0.7). Similarly, the selective oxidation of the R-(+) and S-(-) enantiomers of propranolol to the major 4-hydroxy metabolite was identical with both wild-type and mutant forms of the enzyme (R/S 0.9), although the formation of minor metabolites (5-hydroxy and deisopropylpropranolol) did show some slight alteration in enantioselectivity. The differential inhibition of enzyme activity by quinine and quinidine was also identical with both forms of CYP2D6, the IC(50) values for each enzyme being approx. 10 microM and 0.1 microM for quinine and quinidine, respectively. The kinetics of formation of alpha-hydroxymetoprolol and 4-hydroxydebrisoquine by wild-type and the Ser(304)Ala mutant was also very similar. However, modest changes in the regioselective oxidation of metoprolol and debrisoquine were observed with the Ser(304)Ala mutant. The regio- and enantioselective oxidation of an analogue of metoprolol, in which the hydroxyl group attached to the chiral carbon was replaced by a methyl moiety, was again identical with both wild-type and Ser(304)Ala mutant. However, the observed selectivity was the reverse of that observed with metoprolol. Collectively, these data indicate that Ser(304) is unlikely to be a key ligand-binding residue, although the residue may indeed be located in the active-site cavity. The reversal of selectivity with the methyl analogue of metoprolol indicates that the hydroxyl group attached to the chiral centre of ligands, such as metoprolol, is important in defining the enzyme's selective properties, and that a hydrogen-bonding residue, other than Ser(304), may be involved in this interaction. Current homology models of the active site of CYP2D6 that predict a hydrogen-bond interaction between Ser(304) and specific ligands will need to be re-evaluated, and other candidate residues capable of such an interaction nominated and tested by site-directed mutagenesis studies.

Protein binding of aspirin and salicylate measured by in vivo ultrafiltration.

OBJECTIVE: Methods for measuring protein binding of drugs generally require direct measurement of the concentration of unbound drug and thus may require a highly sensitive assay. In vivo ultrafiltration has been used to determine protein binding of endogenous substances. We have examined its value for measuring protein binding of drugs because it requires measurement of only the concentration of total drug, not unbound drug, in plasma. METHODS: The protein binding of aspirin and its metabolite salicylate was measured in 29 healthy subjects 20 minutes after a single oral dose of 600 mg soluble aspirin, by the new method, in vivo ultrafiltration, as well as by a standard method, in vitro ultracentrifugation. RESULTS: The data for salicylate were examined systematically to determine the optimal method of determining estimates of protein binding by in vivo ultrafiltration. Estimates of protein binding of salicylate were 81.7% +/- 10.1% (mean +/- SD) by the in vivo method and 81.6% +/- 11.3% by in vitro ultracentrifugation. Bland-Altman analysis of agreement showed that within-individual differences in percentage of protein binding determined by the two methods did not differ significantly from zero (mean difference, 0.07%; 95% confidence interval, -2.33 to +2.46). There was a highly significant correlation between estimates of protein binding by the two methods (r = 0.82; p = 0.001). Protein binding of aspirin was estimated of protein binding by the two methods (r = 0.82; p = 0.001). Protein binding of aspirin was estimated at 58.3% +/- 9.6% by in vivo ultrafiltration and could not be estimated by in vitro ultracentrifugation because the concentration of unbound aspirin in plasma was below the limit of detection for the assay. CONCLUSION: In vivo ultrafiltration can be used to measure protein binding of drugs and has potential advantages over conventional methods. A sensitive assay may not be required because the unbound drug need not be measured, measurement in vivo may maintain more physiologic conditions, and it may be useful in measuring protein binding of drugs that are degraded rapidly in vitro.