Prediction of hepatic metabolic clearance: comparison and assessment of prediction models

Usage of In Vitro Metabolism Data for Drug-Drug Interaction in Physiologically Based Pharmacokinetic Analysis Submissions to the US Food and Drug Administration

The key parameters necessary to predict drug-drug interactions (DDIs) are intrinsic clearance (CL_int ) and fractional contribution of the metabolizing enzyme toward total metabolism (f_m ). Herein, we summarize the accumulated knowledge from 53 approved new drug applications submitted to the Office of Clinical Pharmacology, US Food and Drug Administration, from 2016 to 2018 that contained physiologically based pharmacokinetic (PBPK) models to understand how in vitro data are used in PBPK models to assess drug metabolism and predict DDIs. For evaluation of CL_int and f_m , 29 and 20 new drug applications were included for evaluation, respectively. For CL_int , 86.2% of the PBPK models used modified values based on in vivo data with modifications ranging from -82.5% to 2752.5%. For f_m , 45.0% of the models used modified values with modifications ranging from -28% to 178.6%. When values for CL_int were used from in vitro testing without modification, the model resulted in up to a 14.3-fold overprediction of the area under the concentration-time curve of the substrate. When values for f_m from in vitro testing were used directly, the model resulted in up to a 2.9-fold underprediction of its DDI magnitude with an inducer, and up to a 1.7-fold overprediction of its DDI magnitude with an inhibitor. Our analyses suggested that the in vitro system usually provides a reasonable estimation of f_m when the drug metabolism by a given CYP pathway is more than 70% of the total clearance. In vitro experiments provide important information about basic PK properties of new drugs and can serve as a starting point for building a PBPK model. However, key PBPK parameters such as CL_int and f_m still need to be optimized based on in vivo data.

Comparative Assessment of Extrapolation Methods Based on the Conventional Free Drug Hypothesis and Plasma Protein-Mediated Hepatic Uptake Theory for the Hepatic Clearance Predictions of Two Drugs Extensively Bound to Both the Albumin And Alpha-1-Acid Glycoprotein

Bteich and coworkers recently demonstrated in a companion manuscript (J Pharm Sci 109: https://doi.org/10.1016/j.xphs.2020.07.003) that a protein-mediated hepatic uptake have occurred in an isolated perfused rat liver (IPRL) model for two drugs (Perampanel; PER and Fluoxetine; FLU) that bind extensively to the albumin (ALB) and alpha-1-acid glycoprotein (AGP). However, to our knowledge, there is no quantitative model available to predict the impact of a plasma protein-mediated hepatic uptake on the extent of hepatic clearance (CLh) for a drug binding extensively to these two proteins. Therefore, the main objective was to predict the corresponding CLh, which is an extension of the companion manuscript. The method consisted of extrapolating the intrinsic clearance from the unbound fraction measured in the perfusate or the unbound fraction extrapolated to the surface of the hepatocyte membrane by adapting an existing model of protein-mediated hepatic uptake (i.e., the fu_p-adjusted model) to include a binding ratio between the ALB and AGP. This new approach showed a relevant improvement compared to the free drug hypothesis particularly for FLU that showed the highest degree of ALB-mediated uptake. Overall, this study is a first step towards the development of predictive methods of CLh by considering the binding to ALB and AGP.

Animal to human translation: a systematic scoping review of reported concordance rates

Background

Drug development is currently hampered by high attrition rates; many developed treatments fail during clinical testing. Part of the attrition may be due to low animal-to-human translational success rates; so-called “translational failure”. As far as we know, no systematic overview of published translational success rates exists.

Systematic scoping review

The following research question was examined: “What is the observed range of the animal-to-human translational success (and failure) rates within the currently available empirical evidence?”. We searched PubMed and Embase on 16 October 2017. We included reviews and all other types of “umbrella”-studies of meta-data quantitatively comparing the translational results of studies including at least two species with one being human. We supplemented our database searches with additional strategies. All abstracts and full-text papers were screened by two independent reviewers. Our scoping review comprises 121 references, with various units of measurement: compound or intervention (k = 104), study/experiment (k = 10), and symptom or event (k = 7). Diagnostic statistics corresponded with binary and continuous definitions of successful translation. Binary definitions comprise percentages below twofold error, percentages accurately predicted, and predictive values. Quantitative definitions comprise correlation/regression (r²) and meta-analyses (percentage overlap of 95% confidence intervals). Translational success rates ranged from 0 to 100%.

Conclusion

The wide range of translational success rates observed in our study might indicate that translational success is unpredictable; i.e. it might be unclear upfront if the results of primary animal studies will contribute to translational knowledge. However, the risk of bias of the included studies was high, and much of the included evidence is old, while newer models have become available. Therefore, the reliability of the cumulative evidence from current papers on this topic is insufficient. Further in-depth “umbrella”-studies of translational success rates are still warranted. These are needed to evaluate the probabilistic evidence for predictivity of animal studies for the human situation more reliably, and to determine which factors affect this process.

Interpretation of Non-Clinical Data for Prediction of Human Pharmacokinetic Parameters: In Vitro-In Vivo Extrapolation and Allometric Scaling

Extrapolation of pharmacokinetic (PK) parameters from in vitro or in vivo animal to human is one of the main tasks in the drug development process. Translational approaches provide evidence for go or no-go decision-making during drug discovery and the development process, and the prediction of human PKs prior to the first-in-human clinical trials. In vitro-in vivo extrapolation and allometric scaling are the choice of method for projection to human situations. Although these methods are useful tools for the estimation of PK parameters, it is a challenge to apply these methods since underlying biochemical, mathematical, physiological, and background knowledge of PKs are required. In addition, it is difficult to select an appropriate methodology depending on the data available. Therefore, this review covers the principles of PK parameters pertaining to the clearance, volume of distribution, elimination half-life, absorption rate constant, and prediction method from the original idea to recently developed models in order to introduce optimal models for the prediction of PK parameters.

In Vitro-In Vivo Extrapolation and Hepatic Clearance-Dependent Underprediction

Accurately predicting the hepatic clearance of compounds using in vitro to in vivo extrapolation (IVIVE) is crucial within the pharmaceutical industry. However, several groups have recently highlighted the serious error in the process. Although empirical or regression-based scaling factors may be used to mitigate the common underprediction, they provide unsatisfying solutions because the reasoning behind the underlying error has yet to be determined. One previously noted trend was intrinsic clearance-dependent underprediction, highlighting the limitations of current in vitro systems. When applying these generated in vitro intrinsic clearance values during drug development and making first-in-human dose predictions for new chemical entities though, hepatic clearance is the parameter that must be estimated using a model of hepatic disposition, such as the well-stirred model. Here, we examine error across hepatic clearance ranges and find a similar hepatic clearance-dependent trend, with high clearance compounds not predicted to be so, demonstrating another gap in the field.

Hepatic Clearance Predictions from In Vitro-In Vivo Extrapolation and the Biopharmaceutics Drug Disposition Classification System

Predicting in vivo pharmacokinetic parameters such as clearance from in vitro data is a crucial part of the drug-development process. There is a commonly cited trend that drugs that are highly protein-bound and are substrates for hepatic uptake transporters often yield the worst predictions. Given this information, 11 different data sets using human microsomes and hepatocytes were evaluated to search for trends in accuracy, extent of protein binding, and drug classification based on the Biopharmaceutics Drug Disposition Classification System (BDDCS), which makes predictions about transporter effects. As previously reported, both in vitro systems (microsomes and hepatocytes) gave a large number of inaccurate results, defined as predictions falling more than 2-fold outside of in vivo values. The weighted average of the percentage of inaccuracy was 66.5%. BDDCS class 2 drugs, which are subject to transporter effects in vivo unlike class 1 compounds, had a higher percentage of inaccurate predictions and often had slightly larger bias. However, since the weighted average of the percentage of inaccuracy was still high in both classes (81.9% for class 2 and 62.3% for class 1), it may be currently hard to use BDDCS class to predict potential accuracy. The results of this study emphasize the need for improved in vitro to in vivo extrapolation experimental methods, as using physiologically based scaling is still not accurate, and BDDCS cannot currently help predict accurate results.

Upcyte Human Hepatocytes: a Potent In Vitro Tool for the Prediction of Hepatic Clearance of Metabolically Stable Compounds

In vitro models based on primary human hepatocytes (PHH) have been advanced for clearance (CL) prediction of metabolically stable compounds, representing state-of-the-art assay systems for drug discovery and development. Yet, limited cell availability and large interindividual variability of metabolic profiles remain shortcomings of PHH. Upcyte human hepatocytes (UHH) represent a novel hepatic cell system derived from PHH, exhibiting proliferative capacity for approximately 35 population doublings. UHH from three donors were evaluated during culture for up to 18 days, investigating relative mRNA expression and in situ enzyme activity of cytochrome P450s (P450s), UDP-glucuronosyltransferases, and sulfotransferases. Furthermore, UHH were used for predicting hepatic CL of 21 marketed low to intermediate CL drugs. In a typical experiment, expansion from 3.9 × 10(6) up to 8.5 × 10(7) cells was achieved during subculture. When maintained at confluence, transcripts of major P450s were expressed at donor-specific levels with sustained activities for the majority of isoforms, showing generally low CYP1A2 and high CYP2B6 activity levels. For donor 151-03, CL prediction based on depletion experiments resulted in an average fold error of 2.0, and 80% of compounds being predicted within twofold to in vivo CL for a subset of 10 low CL drugs. UHH showed sustained and consistent activity of drug-metabolizing enzymes (DME), resulting in highly reproducible CL prediction performance. In conclusion, UHH show promising potential as alternative to PHH for standardized in vitro applications in discovery research based on their stable, hepatocyte-like DME phenotype and virtually unlimited cell availability.

Investigation of omeprazole and phenacetin first-pass metabolism in humans using a microscale bioreactor and pharmacokinetic models

A new in vitro microfluidic platform (integrated insert dynamic microfluidic platform, IIDMP) allowing the co-culture of intestinal Caco-2 TC7 cells and of human primary hepatocytes was used to test the absorption and first-pass metabolism of two drugs: phenacetin and omeprazole. The metabolism of these drugs by CYP1A2, CYP2C19 and CYP3A4 was evaluated by the calculation of bioavailabilities and of intrinsic clearances using a pharmacokinetic (PK) model. To demonstrate the usefulness of the device and of the PK model, predictions were compared with in vitro and in vivo results from the literature. Based on the IIDMP experiments, hepatic in vivo clearances of phenacetin and omeprazole in the IIDMP were predicted to be 3.10 ± 0.36 and 1.46 ± 0.25 ml/min/kg body weight, respectively. This appeared lower than the in vivo observed data with values ranging between 11.9-19.6 and 5.8-7.5 ml/min/kg body weight, respectively. Then the calculated hepatic and intestinal clearances led to predicting an oral bioavailability of 0.85 and 0.77 for phenacetin and omeprazole versus 0.92 and 0.78 using separate data from the simple monoculture of Caco-2 TC7 cells and hepatocytes in Petri dishes. When compared with the in vivo data, the results of oral bioavailability were overestimated (0.37 and 0.71, respectively). The feasibility of co-culture in a device allowing the integration of intestinal absorption, intestinal metabolism and hepatic metabolism in a single model was demonstrated. Nevertheless, further experiments with other drugs are needed to extend knowledge of the device to predict oral bioavailability and intestinal first-pass metabolism.

First pass intestinal and liver metabolism of paracetamol in a microfluidic platform coupled with a mathematical modeling as a means of evaluating ADME processes in humans

We developed a microfluidic platform to investigate paracetamol intestinal and liver first pass metabolism. This approach was coupled with a mathematical model to estimate intrinsic in vitro parameters and to predict in vivo processes. The kinetic modeling estimated the paracetamol and paracetamol sulfate permeabilities, the sulfate and glucuronide effluxes in the intestine compartment. Based on a gut model, we estimated intrinsic intestinal clearance of between 26 and 77 L/h for paracetamol in humans, a permeability of 10 L/h, and a gut availability between 0.17 and 0.53 (compared to 0.95-1 in vivo). The role played by the liver in paracetamol metabolism was estimated via in vitro intrinsic clearances of 7.6, 13.6, and 11.5 µL/min/10(6) cells for HepG2/C3a, rat primary hepatocytes, and human primary hepatocytes, respectively. Based on a parallel tube model to describe the liver, the paracetamol hepatic clearance, and the paracetamol hepatic availability in humans were estimated at 6.5 mL/min/kg of bodyweight (BDW) and 0.7, respectively (when compared to 5 mL/min/kg of BDW and 0.77 to 0.88 for in vivo values, respectively). The drug availability was predicted ranging between 0.24 and 0.41 (0.88 in vivo). The overall approach provided a first step in an integrated strategy combining in silico/in vitro methods based on microfluidic for evaluating drug absorption, distribution and metabolism processes.

Physiologically based pharmacokinetic modelling to predict single- and multiple-dose human pharmacokinetics of bitopertin

BACKGROUND

Bitopertin (RG1678) is a glycine reuptake inhibitor currently in phase 3 trials for treatment of schizophrenia. This paper describes the use of physiologically based pharmacokinetic (PBPK) modelling and preclinical data to gain insights into and predict bitopertin clinical pharmacokinetics.

METHODS

Simulations of pharmacokinetics were initiated early in the drug discovery stage by integrating physicochemical properties and in vitro measurements into a PBPK rat model. Comparison of pharmacokinetics predicted by PBPK modelling with those measured after intravenous and oral dosing in rats and monkeys showed a good match and thus increased confidence that a similar approach could be applied for human prediction. After comparison of predicted plasma concentrations with those measured after single oral doses in the first clinical study, the human model was refined and then applied to simulate multiple-dose pharmacokinetics.

RESULTS

Clinical plasma concentrations measured were in good agreement with PBPK predictions. Predicted area under the plasma concentration-time curve (AUC) was within twofold of the observed mean values for all dose levels. Maximum plasma concentration (C max) at higher doses was well predicted but approximately twofold below observed values at the lower doses. A slightly less than dose-proportional increase in both AUC and C max was observed, and model simulations indicated that when the dose exceeded 50 mg, solubility limited the fraction of dose absorbed. Refinement of the absorption model with additional solubility and permeability measurements further improved the match of simulations to observed single-dose data. Simulated multiple-dose pharmacokinetics with the refined model were in good agreement with observed data.

CONCLUSIONS

Clinical pharmacokinetics of bitopertin can be well simulated with a mechanistic PBPK model. This model supports further clinical development and provides a valuable repository for pharmacokinetic knowledge gained about the molecule.

Evaluation of a liver microfluidic biochip to predict in vivo clearances of seven drugs in rats

We investigated metabolic clearances of phenacetin, midazolam, propranolol, paracetamol, tolbutamide, caffeine, and dextromethorphan by primary rat hepatocytes cultivated in microfluidic biochips. The levels of mRNA of the HNF4α, PXR, AHR, CYP3A1, and CYP1A2 genes were enhanced in the biochip cultures when compared with postextraction levels. We measured a high and rapid adsorption on the biochip walls and inside the circuit for dextromethorphan and midazolam, a moderate adsorption for phenacetin and propranolol, and a low adsorption for caffeine, tolbutamide, and paracetamol. Drug biotransformations were demonstrated by the formations of specific metabolites such as paraxanthyne (caffeine), paracetamol (phenacetin), 1-OH midazolam (midazolam), paracetamol sulfate (paracetamol and phenacetin), and dextrorphan (dextromethorphan). We used a pharmacokinetic model to estimate the adsorption and in vitro intrinsic drug clearance values. We calculated in vitro intrinsic clearance values of 0.5, 3, 12.5, 83, 100, 160, and 900 μL/min per 10(6) cells for the tolbutamide, caffeine, paracetamol, dextromethorphan, phenacetin, midazolam, and propranolol, respectively. A second model describing the liver as a well-stirred compartment predicted in vivo hepatic clearances of 0.1, 13.8, 30, 44.1, 61, 72, 85, and 61 mL/min per kg of body mass for the tolbutamide, caffeine, paracetamol, midazolam, dextromethorphan, phenacetin, and propranolol, respectively. These values appeared consistent with previously reported data.

Pharmacophore modelling: applications in drug discovery

This review highlights the concept of using pharmacophore models in modern drug research and reviews some important examples as well as success stories. This includes papers from both method-development and application areas. As indicated by the number of publications available, the pharmacophore approach has proven to be extremely useful not only in virtual screening and library design for efficient hit discovery, but also in the optimisation of lead compounds to clinical candidates. Recent studies focus on the use of parallel screening using pharmacophore models for bioactivity profiling and early stage risk assessment of potential side effects and toxicity, due to the interaction of drug candidates with antitargets.

Applications of human pharmacokinetic prediction in first-in-human dose estimation

Quantitative estimations of first-in-human (FIH) doses are critical for phase I clinical trials in drug development. Human pharmacokinetic (PK) prediction methods have been developed to project the human clearance (CL) and bioavailability with reasonable accuracy, which facilitates estimation of a safe yet efficacious FIH dose. However, the FIH dose estimation is still very challenging and complex. The aim of this article is to review the common approaches for FIH dose estimation with an emphasis on PK-guided estimation. We discuss 5 methods for FIH dose estimation, 17 approaches for the prediction of human CL, 6 methods for the prediction of bioavailability, and 3 tools for the prediction of PK profiles. This review may serve as a practical protocol for PK- or pharmacokinetic/pharmacodynamic-guided estimation of the FIH dose.

Using expression data for quantification of active processes in physiologically based pharmacokinetic modeling

Active processes involved in drug metabolization and distribution mediated by enzymes, transporters, or binding partners mostly occur simultaneously in various organs. However, a quantitative description of active processes is difficult because of limited experimental accessibility of tissue-specific protein activity in vivo. In this work, we present a novel approach to estimate in vivo activity of such enzymes or transporters that have an influence on drug pharmacokinetics. Tissue-specific mRNA expression is used as a surrogate for protein abundance and activity and is integrated into physiologically based pharmacokinetic (PBPK) models that already represent detailed anatomical and physiological information. The new approach was evaluated using three publicly available databases: whole-genome expression microarrays from ArrayExpress, reverse transcription-polymerase chain reaction-derived gene expression estimates collected from the literature, and expressed sequence tags from UniGene. Expression data were preprocessed and stored in a customized database that was then used to build PBPK models for pravastatin in humans. These models represented drug uptake by organic anion-transporting polypeptide 1B1 and organic anion transporter 3, active efflux by multidrug resistance protein 2, and metabolization by sulfotransferases in liver, kidney, and/or intestine. Benchmarking of PBPK models based on gene expression data against alternative models with either a less complex model structure or randomly assigned gene expression values clearly demonstrated the superior model performance of the former. Besides accurate prediction of drug pharmacokinetics, integration of relative gene expression data in PBPK models offers the unique possibility to simultaneously investigate drug-drug interactions in all relevant organs because of the physiological representation of protein-mediated processes.

Prediction of human pharmacokinetics—evaluation of methods for prediction of hepatic metabolic clearance

Methods for prediction of hepatic clearance (CLH) in man have been evaluated. A physiologically-based in-vitro to in-vivo (PB-IVIV) method with human unbound fraction in blood (fu,bl) and hepatocyte intrinsic clearance (CLint)-data has a good rationale and appears to give the best predictions (maximum ∼2-fold errors; < 25% errors for half of CL-predictions; appropriate ranking). Inclusion of an empirical scaling factor is, however, needed, and reasons include the use of cryopreserved hepatocytes with low activity, and inappropriate CLint- and fu,bl-estimation methods. Thus, an improvement of this methodology is possible and required. Neglect of fu,bl or incorporation of incubation binding does not seem appropriate. When microsome CLint-data are used with this approach, the CLH is underpredicted by 5- to 9-fold on average, and a 106-fold underprediction (attrition potential) has been observed. The poor performance could probably be related to permeation, binding and low metabolic activity. Inclusion of scaling factors and neglect of fu,bl for basic and neutral compounds improve microsome predictions. The performance is, however, still not satisfactory. Allometry incorrectly assumes that the determinants for CLH relate to body weight and overpredicts human liver blood flow rate. Consequently, allometric methods have poor predictability. Simple allometry has an average overprediction potential, > 2-fold errors for ∼1/3 of predictions, and 140-fold underprediction to 5800-fold overprediction (potential safety risk) range. In-silico methodologies are available, but these need further development. Acceptable prediction errors for compounds with low and high CLH should be ∼50 and ∼10%, respectively. In conclusion, it is recommended that PB-IVIV with human hepatocyte CLint and fu,bl is applied and improved, limits for acceptable errors are decreased, and that animal CLH-studies and allometry are avoided.

Prediction of hepatic clearance using cryopreserved human hepatocytes: a comparison of serum and serum-free incubations

Cryopreserved human hepatocytes have been used to predict hepatic in-vivo clearance. Physiologically-based direct scaling methods generally underestimate human in-vivo hepatic clearance. Cryopreserved human hepatocytes were incubated in 100% serum and in serum-free medium to predict the in-vivo hepatic clearance of six compounds (phenazone (antipyrine), bosentan, mibefradil, midazolam, naloxone and oxazepam). Monte Carlo simulations were performed in an attempt to incorporate the variability and uncertainty in the measured parameters to the prediction of hepatic clearance. The intrinsic clearance (CLint) and the associated variability of the six compounds decreased in the presence of serum and the values were reproducible across donors. The predicted CLhep, in-vivo obtained with hepatocytes from donors incubated in serum was more accurate than the prediction obtained in the absence of serum. For example, the CLhep, in-vivo of mibefradil in donor GNG was 4.27 mL min−1 kg−1 in the presence of serum and 0.46 mL min−1 kg−1 in the absence of serum (4.88 mL min−1 kg−1 observed in-vivo). Using the results obtained in this study together with an extended data set (26 compounds), the clearance of 77% of the compounds was predicted within a 2-fold error in the absence of serum. In the presence of serum, 85% of the compounds were successfully predicted within a 2-fold error. In conclusion, cryopreserved human hepatocyte suspensions represented a convenient and predictive model to assess human drug clearance.

Physiologically based pharmacokinetics (PBPK)

Allometric scaling is widely used to predict human pharmacokinetic parameters from preclinical species, and many different approaches have been proposed over the years to improve its predictive performance. Nevertheless, prediction errors are commonly observed in the practical application of simple allometry, for example, in cases where the hepatic metabolic clearance is mainly determined by enzyme activities, which do not scale allometrically across species. Therefore, if good correlation was noted for some drugs, poor correlation was observed for others, highlighting the need for other conceptual approaches. Physiologically based pharmacokinetic (PBPK) models are now a well-established approach to conduct extrapolations across species and to generate simulations of pharmacokinetic profiles under various physiological conditions. While conventional pharmacokinetic models are defined by drug-related data themselves, PBPK models have richer information content and integrate information from various sources, including drug-dependent, physiological, and biological parameters as they vary in between species, subjects, or with age and disease state. Therefore, the biological and mechanistic bases of PBPK models allow the extrapolation of the kinetic behavior of drugs with regard to dose, route, and species. In addition, by providing a link between tissue concentrations and toxicological or pharmacological effects, PBPK modeling represents a framework for mechanistic pharmacokinetic-pharmacodynamic models.

Comparison of clearance predictions using primary cultures and suspensions of human hepatocytes

Various incubation conditions of human hepatocytes were compared for their accuracy in predicting the in vivo hepatic clearance (CL(H)) of model compounds. The test compounds were the highly cleared, low protein bound naloxone (in vivo CL(H) = 25 ml min(-1) kg(-1); free fraction = 0.6), the medium clearance, highly protein bound midazolam (CL(H) = 12 ml min(-1) kg(-1); free fraction = 0.04) and the low clearance, highly protein bound bosentan (CL(H) = 3.9 ml min(-1) kg(-1); free fraction = 0.02). Each compound was tested in three 'hepatocyte systems', using resections from three donors, in the presence and absence of human serum. Those hepatocyte systems were: conventional primary cultures, freshly isolated suspensions and cryopreserved suspended hepatocytes. Except for a twofold overestimated CL(H) for bosentan from conventional primary cultures, and despite variable cryopreservation recoveries, similar predictions of CL(H) were recorded with all hepatocyte systems. Moreover, the CL(H) values obtained with cryopreserved suspended hepatocytes were similar to those obtained with freshly isolated suspensions. For midazolam and bosentan, the predicted in vivo CL(H) was markedly higher in the presence of serum, whereas serum had little influence on the scaled-up CL(H) of naloxone. In vivo, CL(H) was properly approached for naloxone and bosentan (particularly from experiments in the presence of serum), but it was strongly underestimated for midazolam (particularly in the absence of serum). Additional compounds need to be investigated to confirm the above findings as well as to assess why the clearances of some highly protein-bound compounds are still considerably underestimated.

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 <or=2. IVIVE predictions were accurate in 14 of 15 cases (mean-fold error range: 1.02-4.00). All five AS methods were accurate in 13, 11, 10, 10 and 14 cases, respectively. However, in some cases the error of AS exceeded fivefold. On the basis of the current results, IVIVE is more reliable than AS in predicting human clearance values for drugs mainly metabolized by CYP450 enzymes. This suggests that the place of AS methods in pre-clinical drug development warrants further scrutiny.

The metabolism of the 5HT3 antagonists, ondansetron, alosetron and GR87442 II: investigation into the in vitro methods used to predict the in vivo hepatic clearance of ondansetron, alosetron and GR87442 in the rat, dog and human

The in vitro clearances of the 5HT3 antagonists, ondansetron, alosetron and GR87442 were investigated. Intrinsic clearances using either metabolite formation or substrate depletion methods were equivalent (R2 = 0.95). Hepatocytes from preclinical species were superior to microsomes for the prediction of hepatic clearance (CL(H)), whereas the predictions from human microsomes and hepatocytes were similar. Using a non-restrictive model, seven of the nine CL(H) predictions using hepatocytes were within 2-fold of the in vivo CL(H) values. If the unbound fraction was included, the clearance of the compounds was generally under-predicted by both in vitro models. However, for the most metabolically stable compound, GR87442, the non-restrictive model over-predicted CLp. This and the possibility of extrahepatic metabolism indicate that the restrictive model is more appropriate for prediction of CL(H). The rank order of metabolic stability correlated with that in vivo. All three compounds were more metabolically stable in human than in the preclinical animal species examined.

A Database of Historically-Observed Chemical Replacements

A systematic analysis of one-to-one chemical replacements occurring in a set of 50,000 druglike molecules was performed. The frequency of occurrence, as well as the average change in measured and calculated properties, was computed for each observed substitution. The experimental properties considered were solubility, protein binding, and logD. The calculated properties were logP, molecular weight, number of hydrogen bond donors and acceptors, and polar surface area. During this analysis, in which 9000 different functional groups were considered, 0.7 million substitutions were identified and stored in a database. As an application, we present a web interface from which users can identify historically observed replacements of any functional group on their query molecule. The server returns a list of side-chains, as well as the historically observed shift in experimental properties.

Simulation and prediction of in vivo drug metabolism in human populations from in vitro data

The perceived failure of new drug development has been blamed on deficiencies in in vivo studies of drug efficacy and safety. Prior simulation of the potential exposure of different individuals to a given dose might help to improve the design of such studies. This should also help researchers to focus on the characteristics of individuals who present with extreme reactions to therapy. An effort to build virtual populations using extensive demographic, physiological, genomic and in vitro biochemical data to simulate and predict drug disposition from routinely collected in vitro data is outlined.

The Prediction of Drug Metabolism, Tissue Distribution, and Bioavailability of 50 Structurally Diverse Compounds in Rat Using Mechanism-Based Absorption, Distribution, and Metabolism Prediction Tools

The aim of this study was to assess a physiologically based modeling approach for predicting drug metabolism, tissue distribution, and bioavailability in rat for a structurally diverse set of neutral and moderate-to-strong basic compounds (n = 50). Hepatic blood clearance (CL(h)) was projected using microsomal data and shown to be well predicted, irrespective of the type of hepatic extraction model (80% within 2-fold). Best predictions of CL(h) were obtained disregarding both plasma and microsomal protein binding, whereas strong bias was seen using either blood binding only or both plasma and microsomal protein binding. Two mechanistic tissue composition-based equations were evaluated for predicting volume of distribution (V(dss)) and tissue-to-plasma partitioning (P(tp)). A first approach, which accounted for ionic interactions with acidic phospholipids, resulted in accurate predictions of V(dss) (80% within 2-fold). In contrast, a second approach, which disregarded ionic interactions, was a poor predictor of V(dss) (60% within 2-fold). The first approach also yielded accurate predictions of P(tp) in muscle, heart, and kidney (80% within 3-fold), whereas in lung, liver, and brain, predictions ranged from 47% to 62% within 3-fold. Using the second approach, P(tp) prediction accuracy in muscle, heart, and kidney was on average 70% within 3-fold, and ranged from 24% to 54% in all other tissues. Combining all methods for predicting V(dss) and CL(h) resulted in accurate predictions of the in vivo half-life (70% within 2-fold). Oral bioavailability was well predicted using CL(h) data and Gastroplus Software (80% within 2-fold). These results illustrate that physiologically based prediction tools can provide accurate predictions of rat pharmacokinetics.

Cryopreserved human hepatocytes in suspension are a convenient high throughput tool for the prediction of metabolic clearance

Hepatocyte assays, routinely used to assess the metabolic stability of new chemical entities, were recently improved by using hepatocytes in suspension instead of primary cultures [N. Blanchard, L. Richert, B. Notter, F. Delobel, P. David, P. Coassolo, T. Lavé, Impact of serum on clearance predictions obtained from suspensions and primary cultures of rat hepatocytes, Eur. J. Pharm. Sci. 23 (2004) 189-199]. The aim of the present study was to investigate miniaturising the suspension assay by using cryopreserved human hepatocytes, i.e., 150,000 cells/well in 96-well plates, to predict hepatic clearance (CLH) in order to increase compound throughput and decrease cost and tissue requirements. For this, an evaluation was first carried out with rat hepatocytes. Then, human hepatocytes from various donors were used under these predetermined conditions, either immediately after isolation, either after a 20-h-cold storage period in UW or after cryopreservation. The values of CLint and CLH determined using human hepatocytes in suspension in 96-well plates, immediately after isolation, after cold storage or after cryopreservation, were comparable to those obtained with hepatocytes in primary culture. In particular, the use of cryopreserved human hepatocytes in suspension in a 96-well format appeared to be largely satisfactory as a tool for screening and ranking of compounds in the early phase of the drug discovery process.

An evaluation of the utility of physiologically based models of pharmacokinetics in early drug discovery

Generic physiologically-based models of pharmacokinetics were evaluated for early drug discovery. Plasma profiles after intravenous and oral dosing were simulated in rat for 68 compounds from six chemical classes. Input data consisted of structure based predictions of lipophilicity, ionization, and protein binding plus intrinsic clearance measured in rat hepatocytes, single measured values of aqueous solubility, and artificial membrane permeability. LogP of compounds was high with a mean of 3.9 while free fraction in plasma (mean 9%) and solubility (mean 37 microg/mL) were low. Predicted and observed clearance and volume showed mean fold-error and R2 of 1.8, 0.56, and 1.9, 0.25 respectively. Predicted bioavailability showed strong bias to under prediction correlated to very low aqueous solubility and a theoretical correction for bile salt solubilization in vivo brought some improvement in average prediction error (to 31%). Overall, this evaluation shows that generic simulation may be applicable for typical drug-like compounds to predict differences in pharmacokinetic parameters of more than twofold based upon minimal measured input data. However verification of the simulations with in vivo data for a few compounds of each compound class is recommended since recent discovery compounds may have properties beyond the scope of the current generic models.

A Novel Strategy for Physiologically Based Predictions of Human Pharmacokinetics

BACKGROUND

The major aim of this study was to develop a strategy for predicting human pharmacokinetics using physiologically based pharmacokinetic (PBPK) modelling. This was compared with allometry (of plasma concentration-time profiles using the Dedrick approach), in order to determine the best approaches and strategies for the prediction of human pharmacokinetics.

METHODS

PBPK and Dedrick predictions were made for 19 F. Hoffmann-La Roche compounds. A strategy for the prediction of human pharmacokinetics using PBPK modelling was proposed in this study. Predicted values (pharmacokinetic parameters, plasma concentrations) were compared with observed values obtained after intravenous and oral administration in order to assess the accuracy of the prediction methods.

RESULTS

By following the proposed strategy for PBPK, a prediction would have been made prospectively for approximately 70% of the compounds. The prediction accuracy for these compounds in terms of the percentage of compounds with an average-fold error of <2-fold was 83%, 50%, 75%, 67%, 92% and 100% for apparent oral clearance (CL/F), apparent volume of distribution during terminal phase after oral administration (V(z)/F), terminal elimination half-life (t(1/2)), peak plasma concentration (C(max)), area under the plasma concentration-time curve (AUC) and time to reach C(max) (t(max)), respectively. For the other 30% compounds, unacceptable prediction accuracy was obtained in animals; therefore, a prospective prediction of human pharmacokinetics would not have been made using PBPK. For these compounds, prediction accuracy was also poor using the Dedrick approach. In the majority of cases, PBPK gave more accurate predictions of pharmacokinetic parameters and plasma concentration-time profiles than the Dedrick approach.

CONCLUSIONS

Based on the dataset evaluated in this study, PBPK gave reasonable predictions of human pharmacokinetics using preclinical data and is the recommended approach in the majority of cases. In addition, PBPK modelling is a useful tool to gain insights into the properties of a compound. Thus, PBPK can guide experimental efforts to obtain the relevant information necessary to understand the compound's properties before entry into human, ultimately resulting in a higher level of prediction accuracy.

A strategy for preclinical formulation development using GastroPlus™ as pharmacokinetic simulation tool and a statistical screening design applied to a dog study

The aim of this paper is to propose a pharmaceutical risk assessment strategy that goes beyond the usual characterisation of a clinical candidate molecule according to the biopharmaceutical classification system (BCS). This strategy was evaluated for a new CNS drug with poor solubility and good permeability. In a first step, GastroPlus was used to simulate the absorption process based on preformulation data. These input data involved a physicochemical drug characterisation including drug solubility measurements in simulated physiological media, as well as permeability determination. Further computer simulations were conducted to determine the sensitivity to changes of selected input values. Thus, oral bioavailability prediction was studied as a function of the particle size and drug solubility. The second part of the presented strategy for preclinical formulation development was to test specially designed formulations in a 2(3) screening factorial plan using the dog as the animal model. The factors were the dosage form, food effect and dose strength. One of the two experimental formulations was a capsule filled with the micronised drug, whereas the other formulation was a surfactant solution of the drug. Accordingly, a "worst case" formulation was compared with a "best case" drug solution over the clinically relevant dose range in fasted and fed dogs. The results of the computer simulation indicated that a fraction of the dose is dissolved in the stomach and precipitates partially in the small intestine. The simulation predicted almost full drug absorption during the GI transit time. Interestingly, the simulation implies that stomach drug solubility had little impact on overall fraction absorbed. The results also showed that changes of particle size and reference solubility within two orders of magnitude hardly affected the oral bioavailability. This in silico deduction was subsequently compared with the results of the dog studies. Indeed a surfactant drug solution showed no clear biopharmaceutical superiority over a solid capsule formulation on the average of both dose strengths in fasted and fed dogs. Despite the substantial variability of the in vivo data, the factorial screening design indicated marginal significant interaction between the dose level and feeding status. This can be viewed as a flag for the planning of further studies, since a potential effect of one factor may depend on the level of the other. In summary, the GastroPlus simulation together with the statistically designed dog study provided a thorough biopharmaceutical assessment of the new CNS drug. Based on these findings, it was decided to develop a standard granulate in capsules for phase I studies. More sophisticated formulation options were abandoned and so the clinical formulation development was conducted in a cost-efficient way.

Recent advances in pharmacokinetic extrapolation from preclinical data to humans

The early characterisation of drug metabolism and pharmacokinetic (DMPK) properties of new chemical entities plays a key role in the pharmaceutical industry's effort to reduce attrition. Specifically, a major goal of early DMPK studies is to accurately predict the behaviour of new chemical entities in humans, thus allowing likely failures to be terminated rapidly and resource to be placed on molecules most likely to succeed. The present review summarises progress over the past several years in the key technologies used in the pharmaceutical industry to achieve these goals: namely, in vivo, in vitro and in silico/computational tools. The limitations of the various assays are discussed, with attention also given to likely future directions in this field.

Quantitative structure-pharmacokinetic relationships for drug clearance by using statistical learning methods

Quantitative structure-pharmacokinetic relationships (QSPkR) have increasingly been used for the prediction of the pharmacokinetic properties of drug leads. Several QSPkR models have been developed to predict the total clearance (CL(tot)) of a compound. These models give good prediction accuracy but they are primarily based on a limited number of related compounds which are significantly lesser in number and diversity than the 503 compounds with known CL(tot) described in the literature. It is desirable to examine whether these and other statistical learning methods can be used for predicting the CL(tot) of a more diverse set of compounds. In this work, three statistical learning methods, general regression neural network (GRNN), support vector regression (SVR) and k-nearest neighbour (KNN) were explored for modeling the CL(tot) of all of the 503 known compounds. Six different sets of molecular descriptors, DS-MIXED, DS-3DMoRSE, DS-ATS, DS-GETAWAY, DS-RDF and DS-WHIM, were evaluated for their usefulness in the prediction of CL(tot). GRNN-, SVR- and KNN-developed models have average-fold errors in the range of 1.63 to 1.96, 1.66-1.95 and 1.90-2.23, respectively. For the best GRNN-, SVR- and KNN-developed models, the percentage of compounds with predicted CL(tot) within two-fold error of actual values are in the range of 61.9-74.3% and are comparable or slightly better than those of earlier studies. QSPkR models developed by using DS-MIXED, which is a collection of constitutional, geometrical, topological and electrotopological descriptors, generally give better prediction accuracies than those developed by using other descriptor sets. These results suggest that GRNN, SVR, and their consensus model are potentially useful for predicting QSPkR properties of drug leads.

In silico approaches for predicting ADME properties of drugs

Combinatorial chemistry and high-throughput screening have increased the possibility of finding new lead compounds at much shorter time periods than conventional medicinal chemistry. However, too much promising drug candidates often fail because of unsatisfactory ADME properties. In silico ADME studies are expected to reduce the risk of late-stage attrition of drug development and to optimize screening and testing by looking at only the promising compounds. To this end, many in silico approaches for predicting ADME properties of compounds from their chemical structure have been developed, ranging from data-based approaches such as quantitative structure-activity relationship (QSAR), similarity searches, and 3-dimensional QSAR, to structure-based methods such as ligand-protein docking and pharmacophore modelling. In addition, several methods of integrating ADME properties to predict pharmacokinetics at the organ or body level have been studied. In this article, we briefly summarize in silico ADME approaches.

Prediction of Human Drug Clearance from in Vitro and Preclinical Data Using Physiologically Based and Empirical Approaches

PURPOSE

The aim of this study is to compare the accuracy of five methods for predicting in vivo intrinsic clearance (CL(int)) and seven for predicting hepatic clearance (CL(h)) in humans using in vitro microsomal data and/or preclinical animal data.

METHODS

The human CL(int) was predicted for 33 drugs by five methods that used either in vitro data with a physiologic scaling factor (SF), with an empirical SF, with the physiologic and drug-specific (the ratio of in vivo and in vitro CL(int) in rats) SFs, or rat CL(int) directly and with allometric scaling. Using the estimated CL(int), the CL(h) in humans was calculated according to the well-stirred liver model. The CL(h) was also predicted using additional two methods: using direct allometric scaling or drug-specific SF and allometry.

RESULTS

Using in vitro human microsomal data with a physiologic SF resulted in consistent underestimation of both CL(int) and CL(h). This bias was reduced by using either an empirical SF, a drug-specific SF, or allometry. However, for allometry, there was a substantial decrease in precision. For drug-specific SF, bias was less reduced, precision was similar to an empirical SF. Both CL(int) and CL(h) were best predicted using in vitro human microsomal data with empirical SF. Use of larger data set of 52 drugs with the well-stirred liver model resulted in a best-fit empirical SF that is 9-fold increase on the physiologic SF.

CONCLUSIONS

Overall, the empirical SF method and the drug-specific SF method appear to be the best methods; they show lower bias than the physiologic SF and better precision than allometric approaches. The use of in vitro human microsomal data with an empirical SF may be preferable, as it does not require extra information from a preclinical study.

Prediction of human clearance (CL) and volume of distribution (VD)

The crucial pharmacokinetic parameters 'volume of distribution' and 'human clearance' determine the extent and duration a compound remains in an organism. Potential drug candidates will fail to become successful drugs on the market without favorable values for these parameters, even if they are most efficacious at the target in vitro.The prediction of volume of distribution and human clearance in drug research and development is a key technology to assess possible drug candidates.:

A comprehensive analysis of the role of correction factors in the allometric predictivity of clearance from rat, dog, and monkey to humans

This study was conducted to comprehensively evaluate the performance of various allometric scaling methods for the prediction of human clearance. Allometric scaling was used to predict clearance for 103 compounds, for which clearance data in the rat, dog, monkey, and humans were available. Allometry was performed using all three preclinical species and with combinations of any two species. The methods employed included standard allometry and various correction factors, including brain weight, maximum lifespan potential, and glomerular filtration. Scaling was performed on all compounds universally and on segregated subsets based on allometric exponent, clearance, physicochemical property, or route of elimination. 776 allometric combinations with 27,313 individual outcomes were performed. A predicted-to-observed clearance ratio of 0.5 to twofold was preselected as the criterion for predictive success. The success rate of allometric scaling ranged from 18 to 53%; none of the correction factors resulted in substantially improved predictivity. Furthermore, none of the methods attempted in this study achieved a success rate greater than that observed by simply estimating human clearance based on monkey hepatic extraction. Prospective allometric scaling, with or without correction factors, represents a suboptimal technique for estimating human clearance based on in vivo preclinical data.

Impact of serum on clearance predictions obtained from suspensions and primary cultures of rat hepatocytes

The objective of the present study was to compare two configurations of the hepatocyte model namely suspensions (SH) and conventional primary cultures (CPC) for their ability to predict the hepatic clearance in vivo in the rat and, to investigate the impact of serum on the prediction accuracy. The metabolic competences of several cytochrome P450 isoenzymes were investigated both in CPC and SH in the presence or absence of serum. Under the same conditions, the in vitro intrinsic clearance of six test compounds metabolised by a variety of phase I and phase II enzymes (antipyrine, RO-X, mibefradil, midazolam, naloxone and oxazepam) were derived from Vmax/Km scaled up to the corresponding in vivo hepatic metabolic clearance. CYP activities were shown to be stable in both CPC and SH for up to 6 h of incubation, except for the CYP 3A1 activity that decreased in CPC even in the presence of serum. Moreover, the clearances predicted from SH in the presence of serum were closer to the in vivo values than those obtained from CPC. SH represent a convenient model to assess the hepatic metabolism of xenobiotics, the presence of serum in the incubation medium significantly improved in several instances the quality of the predictions.

Predicting drug clearance from recombinantly expressed CYPs: intersystem extrapolation factors

1. Recombinantly expressed human cytochromes P450 (rhCYPs) have been underused for the prediction of human drug clearance (CL). 2. Differences in intrinsic activity (per unit CYP) between rhCYP and human liver enzymes complicate the issue and these discrepancies have not been investigated systematically. We define intersystem extrapolation factors (ISEFs) that allow the use of rhCYP data for the in vitro-in vivo extrapolation of human drug CL and the variance that is associated with interindividual variation of CYP abundance due to genetic and environmental effects. 3. A large database (n = 451) of metabolic stability data has been compiled and used to derive ISEFs for the most commonly used expression systems and CYP enzymes. 4. Statistical models were constructed for the ISEFs to determine major covariates in order to optimize experimental design to increase prediction accuracy. 5. Suggestions have been made for the conduct of future studies using rhCYP to predict human drug clearance.

Progress towards prediction of human pharmacokinetic parameters from in vitro technologies

This review provides an academic view of the current status on using in vitro systems for the prediction of human in vivo drug clearance and inhibition interaction potential. It stresses that although in vitro technology continues to develop in an impressive way and expectations are high within the pharmaceutical industry, the potential of prediction process is yet to be fully realized. The principles of scaling and modeling in vitro parameters have a sound base and have been validated by using animal tissue. However, it is clear that the comparatively simple standard approach developed and validated in animal systems, results in a high incidence of underprediction for parameters describing clearance and inhibition interaction potential when applied to humans. There are several challenges to our ability to interpret the human in vitro data that can now be so readily generated, in particular, accommodating the unusual kinetic properties characteristic of CYP3A4 substrates, namely, positive and negative cooperativity, in the assessment of inhibition potential.

Evaluation of fresh and cryopreserved hepatocytes as in vitro drug metabolism tools for the prediction of metabolic clearance

The intrinsic clearances (CLint) of 50 neutral and basic marketed drugs were determined in fresh human hepatocytes and the data used to predict human in vivo hepatic metabolic clearance (CLmet). A statistically significant correlation between scaled CLmet and actual CLmet was observed (r2 = 0.48, p < 0.05), and for 73% of the drugs studied, scaled clearances were within 2-fold of the actual clearance. These data have shown that CLint data generated in human hepatocytes can be used to provide estimates of human hepatic CLmet for both phase I and phase II processes. In addition, the utility of commercial and in-house cryopreserved hepatocytes was assessed by comparing with data derived from fresh cells. A set of 14 drugs metabolized by the major human cytochromes P450 (P450s) (CYP1A2, 2C9, 2C19, 2D6, and 3A4) and uridine diphosphate glucuronosyltransferases (UGT1A1, 1A4, 1A9, and 2B7) have been used to characterize the activity of freshly isolated and cryopreserved human and dog hepatocytes. The cryopreserved human and dog cells retained on average 94% and 81%, respectively, of the CLint determined in fresh cells. Cryopreserved hepatocytes retain their full activity for more than 1 year in liquid N2 and are thus a flexible resource of hepatocytes for in vitro assays. In summary, this laboratory has successfully cryopreserved human and dog hepatocytes as assessed by the turnover of prototypic P450 and UGT substrates, and both fresh and cryopreserved human hepatocytes may be used for the prediction of human hepatic CLmet.

Inter-individual variability in levels of human microsomal protein and hepatocellularity per gram of liver

AIMS

To determine levels of microsomal protein (MPPGL) and hepatocellularity (HPGL) per gram of human liver and their interindividual variability.

METHODS

Triplicate liver samples were used to determine values of MPPGL (n = 20) and HPGL (n = 7) after accounting for the fractional loss of microsomal protein or hepatocytes during processing. Repeated measurements from each liver sample allowed the estimation of true interindividual variability in MPPGL and HPGL using ANOVA.

RESULTS

The value of MPPGL ranged from 26 to 54 mg g(-1) (mean(geo)= 33 mg g(-1)). The value of HPGL ranged from 65 to 185 x 10(6) cells g(-1) (mean(geo)= 10(7) x 10(6) cells g(-1)).

CONCLUSIONS

There is significant interindividual variability in MPPGL, which has implications for the accurate extrapolation of in vitro data on drug metabolism to predict in vivo metabolic clearance.

Utility of physiologically based pharmacokinetic models to drug development and rational drug discovery candidate selection

The present paper proposes a modeling and simulation strategy for the prediction of pharmacokinetics (PK) of drug candidates by using currently available in silico and in vitro based prediction tools for absorption, distribution, metabolism and excretion (ADME). These methods can be used to estimate specific ADME parameters (such as rate and extent of absorption into portal vein, volume of distribution, metabolic clearance in the liver). They can also be part of a physiologically based pharmacokinetic (PBPK) model to simulate concentration-time profiles in tissues and plasma resulting from the overall PK after intravenous or oral administration. Since the ADME prediction tools are built only on commonly generated in silico and in vitro data, they can be applied already in early drug discovery, prior to any in vivo study. With the suggested methodology, the following advantages of the mechanistic PBPK modeling framework can now be utilized to explore potential clinical candidates already in drug discovery: (i) prediction of plasma (blood) and tissue PK of drug candidates prior to in vivo experiments, (ii) supporting a better mechanistic understanding of PK properties, as well as helping the development of more rationale PK-PD relationships from tissue kinetic data predicted, and hence facilitating a more rational decision during clinical candidate selection, and (iii) the extrapolation across species, routes of administration and dose levels.