Prediction of volume of distribution in humans: analysis of eight methods and their application in drug discovery.

The performance of eight different methods to predict human volume of distribution (VDss) using a large data set (N > 100) was evaluated.The accuracy was assessed by the end points % within two-fold and absolute average fold error (AAFE). The ability to rank order was accessed by the σ and bias was examined using average fold error. Significance of observed differences was established using statistical permutation testing.The Rodgers-Lukova equation, a tissue composition model, for acids and single species scaling based on rat for other ion classes showed the best results in absence of non-rodent data.The semimechanistic Øie-Tozer model based on all thee preclinical species showed the best performance overall (81% within two-fold, AAFE 1.55, σ 0.62). This was not statistically significantly better at the 95% confidence level than the same model based on two preclinical species or single species scaling from monkey. Thus, the use of primates appears difficult to justify when the sole goal is to extrapolate human volume of distribution.

Comparison of methods for the prediction of human clearance from hepatocyte intrinsic clearance for a set of reference compounds and an external evaluation set.

1. We compared direct scaling, regression model equation and the so-called "Poulin et al." methods to scale clearance (CL) from in vitro intrinsic clearance (CLint) measured in human hepatocytes using two sets of compounds. One reference set comprised of 20 compounds with known elimination pathways and one external evaluation set based on 17 compounds development in Merck (MS). 2. A 90% prospective confidence interval was calculated using the reference set. This interval was found relevant for the regression equation method. The three outliers identified were justified on the basis of their elimination mechanism. 3. The direct scaling method showed a systematic underestimation of clearance in both the reference and evaluation sets. The "Poulin et al." and the regression equation methods showed no obvious bias in either the reference or evaluation sets. 4. The regression model equation was slightly superior to the "Poulin et al." method in the reference set and showed a better absolute average fold error (AAFE) of value 1.3 compared to 1.6. A larger difference was observed in the evaluation set were the regression method and "Poulin et al." resulted in an AAFE of 1.7 and 2.6, respectively (removing the three compounds with known issues mentioned above). A similar pattern was observed for the correlation coefficient. Based on these data we suggest the regression equation method combined with a prospective confidence interval as the first choice for the extrapolation of human in vivo hepatic metabolic clearance from in vitro systems.

A comparative evaluation of models to predict human intestinal metabolism from nonclinical data.

Extensive gut metabolism is often associated with the risk of low and variable bioavailability. The prediction of the fraction of drug escaping gut wall metabolism as well as transporter-mediated secretion (Fg ) has been challenged by the lack of appropriate preclinical models. The purpose of this study is to compare the performance of models that are widely employed in the pharmaceutical industry today to estimate Fg and, based on the outcome, to provide recommendations for the prediction of human Fg during drug discovery and early drug development. The use of in vitro intrinsic clearance from human liver microsomes (HLM) in three mechanistic models - the ADAM, Qgut and Competing Rates - was evaluated for drugs whose metabolism is dominated by CYP450s, assuming that the effect of transporters is negligible. The utility of rat as a model for human Fg was also explored. The ADAM, Qgut and Competing Rates models had comparable prediction success (70%, 74%, 69%, respectively) and bias (AFE = 1.26, 0.74 and 0.81, respectively). However, the ADAM model showed better accuracy compared with the Qgut and Competing Rates models (RMSE =0.20 vs 0.30 and 0.25, respectively). Rat is not a good model (prediction success =32%, RMSE =0.48 and AFE = 0.44) as it seems systematically to under-predict human Fg . Hence, we would recommend the use of rat to identify the need for Fg assessment, followed by the use of HLM in simple models to predict human Fg . © 2017 Merck KGaA. Biopharmaceutics & Drug Disposition Published by John Wiley & Sons, Ltd.

Time- and NADPH-dependent inhibition of cytochrome P450 3A4 by the cyclopentapeptide cilengitide: significance of the guanidine group and accompanying spectral changes.

Cilengitide is a stable cyclic pentapeptide containing an Arg-Gly-Asp motif responsible for selective binding to αVß3 and αVß5 integrins. The candidate drug showed unexpected inhibition of cytochrome P450 (P450) 3A4 at high concentrations, that is, a 15-mM concentration caused attenuation of P450 3A4 activity (depending on the probe substrate): 15-19% direct inhibition, 10-23% time-dependent inhibition (30-minute preincubation), and 54-60% metabolism-dependent inhibition (30-minute preincubation). The inactivation efficiency determined with human liver microsomes was 0.003 ± 0.001 min(-1) mM(-1) and was 0.04 ± 0.01 min(-1) mM(-1) with baculovirus-based microsomes containing recombinant P450 3A4. Neither heme loss nor covalent binding to apoprotein could explain the observed reductions in residual activity. Slowly forming type II difference spectra were observed, with maximum spectral changes after 2 hours. Binding to both reduced and oxidized P450 3A4 was observed, with apparent Kd values of 0.66 µM and 6 µM. The significance of the guanidine group in inhibition was demonstrated using ligand binding spectral changes and inactivation assays with guanidine analogs (debrisoquine, N-acetylarginine-O-methyl ester) and the acetylated ornithine derivative of cilengitide. The observed inhibition could be explained by direct inhibition, plus by formation of stable complexes with both ferric and ferrous forms of heme iron and to some extent by the formation of reactive species capable to react to the protein or heme. Formation of the complex required time and NADPH and is attributed to the guanidino group. Thus, the NADPH-dependent inhibition is considered to be mainly due to the formation of a stable complex rather than the formation of reactive species.

Utility of in vitro systems and preclinical data for the prediction of human intestinal first-pass metabolism during drug discovery and preclinical development.

A growing awareness of the risks associated with extensive intestinal metabolism has triggered an interest in developing robust methods for its quantitative assessment. This study explored the utility of intestinal S9 fractions, human liver microsomes, and recombinant cytochromes P450 to quantify CYP3A-mediated intestinal extraction in humans for a selection of marketed drugs that are predominantly metabolized by CYP3A4. A simple competing rates model is used to estimate the fraction of drug escaping gut wall metabolism (fg) from in vitro intrinsic clearance in humans. The fg values extrapolated from the three in vitro systems used in this study, together with literature-derived fg from human intestinal microsomes, were validated against fg extracted from human in vivo pharmacokinetic (PK) profiles using a generic whole-body physiologically-based pharmacokinetic (PBPK) model. The utility of the rat as a model for human CYP3A-mediated intestinal metabolism was also evaluated. Human fg from PBPK compares well with that from the grapefruit juice method, justifying its use for the evaluation of human in vitro systems. Predictive performance of all human in vitro systems was comparable [root mean square error (RMSE) = 0.22-0.27; n = 10]. Rat fg derived from in vivo PK profiles using PBPK has the lowest RMSE (0.19; n = 11) for the prediction of human fg for the selected compounds, most of which have a fraction absorbed close to 1. On the basis of these evaluations, the combined use of fg from human in vitro systems and rats is recommended for the estimation of CYP3A4-mediated intestinal metabolism in lead optimization and preclinical development phases.

Comparison of the in vitro and in vivo metabolism of Cladribine (Leustatin, Movectro) in animals and human.

New insight into the in vitro and in vivo metabolism of Cladribine (2-chloro-2'-deoxyadenosine, [2-CdA]) are presented. Following incubation of [(14)C]-2-CdA in mouse, rat, rabbit, dog, monkey and human hepatocyte cultures, variable turnover was observed with oxidations and direct glucuronidation pathways. The oxidative cleavage to 2-chloroadenine (2-CA, M1) was only observed in rabbit and rat. Following incubation of [(14)C]-2-CdA in whole blood from mouse, monkey and human, a significant turnover was observed. The main metabolites in monkey and human were 2-chlorodeoxyinosine (M11, 16% of total radioactivity) and 2-chlorodeoxyinosine (M12, 43%). In mouse, 2-CA was the major metabolite (2-CA; M1, 73%). After single intravenous and oral administration of [(14)C]-2-CdA to mice, 2-chlorodeoxyinosine (M11) was confirmed in plasma, while 2-chlorohypoxanthine (M12) and 2-CA (M1) were found in urine. Overall, the use of [(14)C]-2-CdA both in vitro (incubations in mouse, monkey and human whole blood) and in vivo (mouse) has confirmed the existence of an additional metabolism pathway leading to the formation of 2-chlorodeoxyinosine (M11) and 2-chlorohypoxanthine (M12). Formation of these two metabolites demonstrates that Cladribine as free form is not fully resistant to adenosine deaminase as suggested earlier, an enzyme involved in its mode of action.

The right compound in the right assay at the right time: an integrated discovery DMPK strategy.

The high rate of attrition during drug development and its associated high research and development (R&D) cost have put pressure on pharmaceutical companies to ensure that candidate drugs going to clinical testing have the appropriate quality such that the biological hypothesis could be evaluated. To help achieve this ambition, drug metabolism and pharmacokinetic (DMPK) science and increasing investment have been deployed earlier in the R&D process. To gain maximum return on investment, it is essential that DMPK concepts are both appropriately integrated into the compound design process and that compound selection is focused on accurate prediction of likely outcomes in patients. This article describes key principles that underpin the contribution of DMPK science for small-molecule research based on 15 years of discovery support in a major pharmaceutical company. It does not aim to describe the breadth and depth of DMPK science, but more the practical application for decision making in real-world situations.

Evaluation of the use of static and dynamic models to predict drug-drug interaction and its associated variability: impact on drug discovery and early development.

Simcyp, a population-based simulator, is widely used for evaluating drug-drug interaction (DDI) risks in healthy and disease populations. We compare the prediction performance of Simcyp with that of mechanistic static models using different types of inhibitor concentrations, with the aim of understanding their strengths/weaknesses and recommending the optimal use of tools in drug discovery/early development. The inclusion of an additional term in static equations to consider the contribution of hepatic first pass to DDIs (AUCR(hfp)) has also been examined. A second objective was to assess Simcyp's estimation of variability associated with DDIs. The data set used for the analysis comprises 19 clinical interactions from 11 proprietary compounds. Except for gut interaction parameters, all other input data were identical for Simcyp and static models. Static equations using an unbound average steady-state systemic inhibitor concentration (I(sys)) and a fixed fraction of gut extraction and neglecting gut extraction in the case of induction interactions performed better than Simcyp (84% compared with 58% of the interactions predicted within 2-fold). Differences in the prediction outcomes between the static and dynamic models are attributable to differences in first-pass contribution to DDI. The inclusion of AUCR(hfp) in static equations leads to systematic overprediction of interaction, suggesting a limited role for hepatic first pass in determining inhibition-based DDIs for our data set. Our analysis supports the use of static models when elimination routes of the victim compound and the role of gut extraction for the victim and/or inhibitor in humans are not well defined. A fixed variability of 40% of predicted mean area under the concentration-time curve ratio is recommended.

In vitro approach to assess the potential for risk of idiosyncratic adverse reactions caused by candidate drugs.

Idiosyncratic adverse drug reactions (IADRs) in humans can result in a broad range of clinically significant toxicities leading to attrition during drug development as well as postlicensing withdrawal or labeling. IADRs arise from both drug and patient related mechanisms and risk factors. Drug related risk factors, resulting from parent compound or metabolites, may involve multiple contributory mechanisms including organelle toxicity, effects related to compound disposition, and/or immune activation. In the current study, we evaluate an in vitro approach, which explored both cellular effects and covalent binding (CVB) to assess IADR risks for drug candidates using 36 drugs which caused different patterns and severities of IADRs in humans. The cellular effects were tested in an in vitro Panel of five assays which quantified (1) toxicity to THLE cells (SV40 T-antigen-immortalized human liver epithelial cells), which do not express P450s, (2) toxicity to a THLE cell line which selectively expresses P450 3A4, (3) cytotoxicity in HepG2 cells in glucose and galactose media, which is indicative of mitochondrial injury, (4) inhibition of the human bile salt export pump, BSEP, and (5) inhibition of the rat multidrug resistance associated protein 2, Mrp2. In addition, the CVB Burden was estimated by determining the CVB of radiolabeled compound to human hepatocytes and factoring in both the maximum prescribed daily dose and the fraction of metabolism leading to CVB. Combining the aggregated results from the in vitro Panel assays with the CVB Burden data discriminated, with high specificity (78%) and sensitivity (100%), between 27 drugs, which had severe or marked IADR concern, and 9 drugs, which had low IADR concern, we propose that this integrated approach has the potential to enable selection of drug candidates with reduced propensity to cause IADRs in humans.

In vitro characterization of sarizotan metabolism: hepatic clearance, identification and characterization of metabolites, drug-metabolizing enzyme identification, and evaluation of cytochrome p450 inhibition.

In vitro biotransformation studies of sarizotan using human liver microsomes (HLM) showed aromatic and aliphatic monohydroxylation and dealkylation. Recombinant cytochromes P450 (P450) together with P450-selective inhibitors in HLM/hepatocyte cultures were used to evaluate the relative contribution of different P450s and revealed major involvement of CYP3A4, CYP2C9, CYP2C8, and CYP1A2 in sarizotan metabolism. The apparent K(m, u) and V(max) of sarizotan clearance, as investigated in HLM, were 9 microM and 3280 pmol/mg/min, predicting in vivo hepatic clearance of 0.94 l/h, which indicates that sarizotan is a low-clearance compound in humans and suggests nonsaturable metabolism at the targeted plasma concentration (< or =1 microM). This finding is confirmed by the reported human clearance (CL/F of 3.6-4.4 l/h) and by the dose-linear area under the curve increase observed with doses up to 25 mg. The inhibitory effect of sarizotan toward six major P450s was evaluated using P450-specific marker reactions in pooled HLM. K(i, u) values of sarizotan against CYP2C8, CYP2C19, and CYP3A4 were >10 microM, whereas those against CYP2D6 and CYP1A2 were 0.43 and 8.7 microM, respectively. Based on the estimates of sarizotan concentrations at the enzyme active sites, no clinically significant drug-drug interactions (DDIs) due to P450 inhibition are expected. This result has been confirmed in human DDI studies in which no inhibition of five major P450s was observed in terms of marker metabolite formation.

Prediction of active human dose: learnings from 20 years of Merck KGaA experience, illustrated by case studies.

High-quality dose predictions based on a good understanding of target engagement is one of the main translational goals in drug development. Here, we systematically evaluate active human dose predictions for 15 Merck KGaA/EMD Serono assets spanning several modalities and therapeutic areas. Using case studies, we illustrate the value of adhering to the translational best practices of having an exposure-response relationship in an appropriate animal model; having validated, translatable pharmacodynamic (PD) biomarkers measurable in Phase I populations in the right tissue; having a deeper understanding of biology; and capturing uncertainties in predictions. Given the gap in publications on the subject, we believe that the learnings from this unique diverse data set, which are generic to the industry, will trigger actions to improve future predictions.

Translational strategies in drug development for knee osteoarthritis.

Osteoarthritis (OA) is a common disease worldwide with large unmet medical needs. To bring innovative treatments to OA patients, we at Merck have implemented a comprehensive strategy for drug candidate evaluation. We have a clear framework for decision-making in our preclinical pipeline, to design our clinical proof-of-concept trials for OA patients. We have qualified our strategy to define and refine dose and dosing regimen, for treatments administered either systemically or intra-articularly (IA). We do this through preclinical in vitro and in vivo studies, and by back-translating results from clinical studies in OA patients.

Evaluating the Role of Solubility in Oral Absorption of Poorly Water-Soluble Drugs Using Physiologically-Based Pharmacokinetic Modeling.

Poor aqueous solubility and dissolution of drug candidates drive key decisions on lead series optimization during drug discovery, on formulation optimization, and clinical studies planning during drug development. The interpretation of the in vivo relevance of early pharmaceutical profiling is often confounded by the multiple factors affecting oral systemic exposure. There is growing evidence that in vitro drug solubility may underestimate the true in vivo solubility and lead to drug misclassification. Based on 10 poorly water-soluble tyrosine kinase inhibitors, this paper demonstrates the use of physiologically-based pharmacokinetic (PK) analysis in combination with early clinical PK data to identify drugs whose absorption is truly limited by solubility in vivo and, therefore, expected to exhibit food effect. Our study supports a totality of evidence approach using early clinical data to guide decisions on conducting drug interaction studies with food and acid-reducing agents.

In vitro and in silico identification and characterization of thiabendazole as a mechanism-based inhibitor of CYP1A2 and simulation of possible pharmacokinetic drug-drug interactions.

Thiabendazole (TBZ) and its major metabolite 5-hydroxythiabendazole (5OH-TBZ) were screened for potential time-dependent inhibition (TDI) against CYP1A2. Screen assays were carried out in the absence and presence of NADPH. TDI was observed with both compounds, with k(inact) and K(I) values of 0.08 and 0.02 min(-1) and 1.4 and 63.3 microM for TBZ and 5OH-TBZ, respectively. Enzyme inactivation was time-, concentration-, and NADPH-dependent. Inactivation by TBZ was irreversible by dialysis and oxidation by potassium ferricyanide, and there was no protection by glutathione. 5OH-TBZ was a weak TDI of CYP1A2, and enzyme activity was recovered by dialysis. IC(50) determination of TBZ and 5OH-TBZ showed both compounds to be potent inhibitors, with IC(50) values of 0.83 and 13.05 microM, respectively. IC(50) shift studies also demonstrated that TBZ was a TDI of CYP1A2. In silico methods identified the thiazole group as a TDI fragment and predicted it as the site of metabolism. The observation pointed to epoxidation of the thiazole and the benzyl rings of TBZ as possible routes of metabolism and mechanisms of TDI. Drug-drug interaction (DDI) simulation studies using SimCyp showed good predictions for competitive inhibition. However, predictions for mechanism-based inhibition (MBI)-based DDI were not in agreement with clinical observations. There was no TBZ accumulation upon chronic administration of the drug. The in vitro MBI findings might therefore not be capturing the in vivo situation in which the proposed bioactivation route is minor. This might be the case for TBZ in which, in vivo, UDP glucuronosyltransferases and sulfanotransferase metabolize and eliminate the 5OH-TBZ.

Requirements to Establishing Confidence in Physiologically Based Pharmacokinetic (PBPK) Models and Overcoming Some of the Challenges to Meeting Them.

When scientifically well-founded, the mechanistic basis of physiologically based pharmacokinetic (PBPK) models can help reduce the uncertainty and increase confidence in extrapolations outside the studied scenarios or studied populations. However, it is not always possible to establish mechanistically credible PBPK models. Requirements to establishing confidence in PBPK models, and challenges to meeting these requirements, are presented in this article. Parameter non-identifiability is the most challenging among the barriers to establishing confidence in PBPK models. Using case examples of small molecule drugs, this article examines the use of hypothesis testing to overcome parameter non-identifiability issues, with the objective of enhancing confidence in the mechanistic basis of PBPK models and thereby improving the quality of predictions that are meant for internal decisions and regulatory submissions. When the mechanistic basis of a PBPK model cannot be established, we propose the use of simpler models or evidence-based approaches.

Species differences in the response of liver drug-metabolizing enzymes to (S)-4-O-tolylsulfanyl-2-(4-trifluormethyl-phenoxy)-butyric acid (EMD 392949) in vivo and in vitro.

Induction of drug-metabolizing enzymes (DMEs) is highly species-specific and can lead to drug-drug interaction and toxicities. In this series of studies we tested the species specificity of the antidiabetic drug development candidate and mixed peroxisome proliferator-activated receptor (PPAR) alpha/gamma agonist (S)-4-O-tolylsulfanyl-2-(4-trifluormethyl-phenoxy)-butyric acid (EMD 392949, EMD) with regard to the induction of gene expression and activities of DMEs, their regulators, and typical PPAR target genes. EMD clearly induced PPARalpha target genes in rats in vivo and in rat hepatocytes but lacked significant induction of DMEs, except for cytochrome P450 (P450) 4A. CYP2C and CYP3A were consistently induced in livers of EMD-treated monkeys. Interestingly, classic rodent peroxisomal proliferation markers were induced in monkeys after 17 weeks but not after a 4-week treatment, a fact also observed in human hepatocytes after 72 h but not 24 h of EMD treatment. In human hepatocyte cultures, EMD showed similar gene expression profiles and induction of P450 activities as in monkeys, indicating that the monkey is predictive for human P450 induction by EMD. In addition, EMD induced a similar gene expression pattern as the PPARalpha agonist fenofibrate in primary rat and human hepatocyte cultures. In conclusion, these data showed an excellent correlation of in vivo data on DME gene expression and activity levels with results generated in hepatocyte monolayer cultures, enabling a solid estimation of human P450 induction. This study also clearly highlighted major differences between primates and rodents in the regulation of major inducible P450s, with evidence of CYP3A and CYP2C inducibility by PPARalpha agonists in monkeys and humans.

A Novel PBPK Modeling Approach to Assess Cytochrome P450 Mediated Drug-Drug Interaction Potential of the Cytotoxic Prodrug Evofosfamide.

Evofosfamide is a cytotoxic small-molecule prodrug preferentially activated under hypoxic conditions. The cytotoxicity of evofosfamide impacted the generation of in vitro drug-drug interaction (DDI) data, especially in vitro induction results. Therefore, a novel physiologically based pharmacokinetic (PBPK) approach was used, which involved available in vitro and clinical data of evofosfamide and combined it with induction data from the prototypical cytochrome P450 (CYP)3A inducer rifampicin. The area under the concentration-time curve (AUC) ratios of midazolam were above 0.80, indicating that induction of CYP3A by evofosfamide administered weekly is unlikely to occur in humans. Moreover, static and PBPK modeling showed no clinically relevant inhibition via CYP2B6, CYP2D6, and CYP3A4. In conclusion, PBPK models were used to supplement in vitro information of a cytotoxic compound. This approach may set a precedent for future studies of cytotoxic drugs, potentially reducing the need for clinical DDI studies and providing more confidence in the clinical use of approved cytotoxic compounds for which DDI information is sparse.

In vitro and in vivo drug disposition of cilengitide in animals and human.

Cilengitide is very low permeable (1.0 nm/sec) stable cyclic pentapeptide containing an Arg-Gly-Asp motif responsible for selective binding to αvß3 and αvß5 integrins administered intravenously (i.v.). In vivo studies in the mouse and Cynomolgus monkeys showed the major component in plasma was unchanged drug (>85%). These results, together with the absence of metabolism in vitro and in animals, indicate minimal metabolism in both species. The excretion of [(14)C]-cilengitide showed profound species differences, with a high renal excretion of the parent drug observed in Cynomolgus monkey (50% dose), but not in mouse (7 and 28%: m/f). Consistently fecal (biliary) secretion was high in mouse (87 and 66% dose: m/f) but low in Cynomolgus monkey (36.5%). Human volunteers administrated with [(14)C]-cilengitide showed that most of the dose was recovered in urine as unchanged drug (77.5%, referred to Becker et al. 2015), indicating that the Cynomolgus monkey was the closer species to human. In order to better understand the species difference between human and mouse, the hepatobiliary disposition of [(14)C]-cilengitide was determined in sandwich-cultured hepatocytes. Cilengitide exhibited modest biliary efflux (30-40%) in mouse, while in human hepatocytes this was negligible. Furthermore, it was confirmed that the uptake of cilengitide into human hepatocytes was minor and appeared to be passive. In summary, the extent of renal and biliary secretion of cilengitide appears to be highly species specific and is qualitatively well explained using sandwich hepatocyte culture models.

Translational Medicine Guide transforms drug development processes: the recent Merck experience.

Merck is implementing a question-based Translational Medicine Guide (TxM Guide) beginning as early as lead optimization into its stage-gate drug development process. Initial experiences with the TxM Guide, which is embedded into an integrated development plan tailored to each development program, demonstrated opportunities to improve target understanding, dose setting (i.e., therapeutic index), and patient subpopulation selection with more robust and relevant early human-based evidence, and increased use of biomarkers and simulations. The TxM Guide is also helping improve organizational learning, costs, and governance. It has also shown the need for stronger external resources for validating biomarkers, demonstrating clinical utility, tracking natural disease history, and biobanking.

Risk assessment and mitigation strategies for reactive metabolites in drug discovery and development.

Drug toxicity is a leading cause of attrition of candidate drugs during drug development as well as of withdrawal of drugs post-licensing due to adverse drug reactions in man. These adverse drug reactions cause a broad range of clinically severe conditions including both highly reproducible and dose dependent toxicities as well as relatively infrequent and idiosyncratic adverse events. The underlying risk factors can be split into two groups: (1) drug-related and (2) patient-related. The drug-related risk factors include metabolic factors that determine the propensity of a molecule to form toxic reactive metabolites (RMs), and the RM and non-RM mediated mechanisms which cause cell and tissue injury. Patient related risk factors may vary markedly between individuals, and encompass genetic and non-genetic processes, e.g. environmental, that influence the disposition of drugs and their metabolites, the nature of the adverse responses elicited and the resulting biological consequences. We describe a new strategy, which builds upon the strategies used currently within numerous pharmaceutical companies to avoid and minimize RM formation during drug discovery, and that is intended to reduce the likelihood that candidate drugs will cause toxicity in the human population. The new strategy addresses drug-related safety hazards, but not patient-related risk factors. A common target organ of toxicity is the liver and to decrease the likelihood that candidate drugs will cause liver toxicity (both non-idiosyncratic and idiosyncratic), we propose use of an in vitro Hepatic Liability Panel alongside in vitro methods for the detection of RMs. This will enable design and selection of compounds in discovery that have reduced propensity to cause liver toxicity. In vitro Hepatic Liability is assessed using toxicity assays that quantify: CYP 450 dependent and CYP 450 independent cell toxicity; mitochondrial impairment; and inhibition of the Bile Salt Export Pump. Prior to progression into development, a Hepatotoxicity Hazard Matrix combines data from the Hepatic Liability Panel with the Estimated RM Body Burden. The latter is defined as the level of covalent binding of radiolabelled drug to human hepatocyte proteins in vitro adjusted for the predicted human dose. We exemplify the potential value of this approach by consideration of the thiazolidinedione class of drugs.