Search

The EU chemicals strategy for sustainability (CSS) asserts that both human health and the environment are presently threatened and that further regulation is necessary. In a recent Guest Editorial, members of the German competent authority for risk assessment, the BfR, raised concerns about the scientific justification for this strategy. The complexity and interdependence of the networks of regulation of chemical substances have ensured that public health and wellbeing in the EU have continuously improved. A continuous process of improvement in consumer protection is clearly desirable but any initiative directed towards this objective must be based on scientific knowledge. It must not confound risk with other factors in determining policy. This conclusion is fully supported in the present Commentary including the request to improve both, data collection and the time-consuming and bureaucratic procedures that delay the publication of regulations.

The authors regret that the letter inaccurately refers to “the lancet diabetes endocrinology” and “The Lancet” while it should read: “The Lancet Diabetes & Endocrinology”. The correct title of the letter should read: “Critique of the “Comment” entitled: “Pyrethroid exposure: Not so harmless after all” by Demeneix et al. (2020) published in The Lancet Diabetes & Endocrinology”. The authors wish to clarify that the comment was published in The Lancet Diabetes & Endocrinology and not The Lancet. The authors also wish to clarify that The Lancet Diabetes & Endocrinology chose not to publish their criticism of the article because the correspondence content published in The Lancet Diabetes & Endocrinology is only accepted within 8 weeks of publication of the original item.

© 2021 IUPAC & De Gruyter. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. For more information, please visit: http://creativecommons.org/licenses/by-nc-nd/4.0/ 2021.This project originated more than 15 years ago with the intent to produce a glossary of drug metabolism terms having definitions especially applicable for use by practicing medicinal chemists. A first-draft version underwent extensive beta-testing that, fortuitously, engaged international audiences in a wide range of disciplines involved in drug discovery and development. It became clear that the inclusion of information to enhance discussions among this mix of participants would be even more valuable. The present version retains a chemical structure theme while expanding tutorial comments that aim to bridge the various perspectives that may arise during interdisciplinary communications about a given term. This glossary is intended to be educational for early stage researchers, as well as useful for investigators at various levels who participate on today's highly multidisciplinary, collaborative small molecule drug discovery teams.

Paracetamol (acetaminophen, APAP) overdose is a leading cause of acute drug-induced liver failure. APAP hepatotoxicity is mediated by the reactive metabolite N-acetyl-p-benzoquinone imine (NAPQI). NAPQI is inactivated by conjugation with glutathione (GSH) to APAP-GSH, which is further converted into its cysteine derivative APAP-CYS. Before necrosis of hepatocytes occurs, APAP-CYS is measurable in plasma of the affected patient and it has been proposed as an early biomarker of acetaminophen toxicity. APAP-GSH and APAP-CYS can be extruded by hepatocytes, but the transporters involved are unknown. In this study we examined whether ATP-binding cassette (ABC) transporters play a role in the cellular efflux of APAP, APAP-GSH, and APAP-CYS. The ABC transport proteins P-gp/ABCB1, BSEP/ABCB11, BCRP/ABCG2, and MRP/ABCC1-5 were overexpressed in HEK293 cells and membrane vesicles were produced. Whereas P-gp, BSEP, MRP3, MRP5, and BCRP did not transport any of the compounds, uptake of APAP-GSH was found for MRP1, MRP2 and MRP4. APAP-CYS appeared to be a substrate of MRP4 and none of the ABC proteins transported APAP. The results suggest that the NAPQI metabolite APAP-CYS can be excreted into plasma by MRP4, where it could be a useful biomarker for APAP exposure and toxicity. Characterization of the cellular efflux of APAP-CYS is important for its development as a biomarker, because plasma concentrations might be influenced by drug-transporter interactions and upregulation of MRP4. Electronic supplementary material The online version of this article (10.1007/s00204-020-02793-4) contains supplementary material, which is available to authorized users.

Background and purpose The aim of this study was to characterize the human cytochrome P450s (CYPs) involved in oxidative bioactivation of flucloxacillin to 5-hydroxymethyl flucloxacillin, a metabolite with high cytotoxicity towards biliary epithelial cells. Experimental approach The CYPs involved in hydroxylation of flucloxacillin were characterized using recombinant human CYPs, pooled liver microsomes in the presence of CYP-specific inhibitors and by correlation analysis using a panel of liver microsomes from 16 donors. Key results Recombinant CYPs showing the highest specific activity were CYP3A4, CYP3A7 and to lower extent CYP2C9 and CTP2C8. Michaelis-Menten enzyme kinetics were determined for pooled human liver microsomes, recombinant CYP3A4, CYP3A7 and CYP2C9. Surprisingly, sulfaphenazole appeared to be a potent inhibitor of 5'-hydroxylation of flucloxacillin by both recombinant CYP3A4 and CYP3A7. Conclusions and implications The combined results show that the 5'-hydroxylation of flucloxacillin is primarily catalysed by CYP3A4, CYP3A7 and CYP2C9. The large variability of the hepatic expression of these enzymes could affect the formation of 5'-hydroxymethyl flucloxacillin, which may determine the differences in susceptibility to flucloxacillin-induced liver injury. Additionally, the strong inhibition in CYP3A-catalysed flucloxacillin metabolism by sulfaphenazole suggests that unanticipated drug-drug interactions could occur with coadministered drugs.

In yeast, toxicity of acetaminophen (APAP), a frequently used analgesic and antipyretic drug, depends on ubiquitin-controlled processes. Previously, we showed a remarkable overlap in toxicity profiles between APAP and tyrosine, and a similarity with drugs like rapamycin and quinine, which induce degradation of the amino acid permease Tat2. Therefore, we investigated in yeast whether APAP reduced the expression levels of amino acid permeases. The protein levels of Tat2, Tat1, Mup1 and Hip1 were reduced, while the expression of the general permease Gap1 was increased, consistent with a nutrient starvation response. Overexpression of Tat1 and Tat2, but not Mup1, Hip1 and Gap1 conferred resistance to APAP. A tryptophan auxotrophic strain trp1Δ was more sensitive to APAP than wild-type and addition of tryptophan completely restored the growth restriction of trp1∆ upon APAP exposure, while tyrosine had an additive effect on APAP toxicity. Furthermore, intracellular aromatic amino acid concentrations were reduced upon APAP exposure. This effect was less prominent in ubiquitin-deficient yeast strains that were APAP resistant and showed a reduced degradation of high affinity amino acid permeases. APAP-induced changes in intracellular amino acid concentrations were also detected in hepatoma HepG2 cells indicating significance for humans. Electronic supplementary material The online version of this article (10.1007/s00726-018-2613-8) contains supplementary material, which is available to authorized users.

The use of diclofenac is associated with rare but severe drug-induced liver injury (DILI) in a very small number of patients. The factors which predispose susceptible patients to hepatotoxicity of diclofenac are still incompletely understood. Formation of protein-reactive metabolites by UDP-glucuronosyl transferases and cytochromes P450 is commonly considered to play an important role, as indicated by the detection of covalent protein adducts and antibodies in the serum of patients suffering from diclofenac-induced liver injury. Since no associations have been found with HLA-alleles, polymorphisms of genes encoding for proteins involved in the disposition of diclofenac may be important. Previous association studies showed that possession of the UGT2B7*2 and CYP2C8*4 alleles is more common in cases of diclofenac-induced DILI. In the present study, the metabolism of diclofenac by UGT2B7*2 and CYP2C8*4 was compared with their corresponding wild-type enzymes. Enzyme kinetic analysis revealed that recombinant UGT2B7*2 showed an almost 6-fold lower intrinsic clearance of diclofenac glucuronidation compared to UGT2B7*1. The mutant CYP2C8*4 showed approximately 35% reduced activity in the 4′-hydroxylation of diclofenac acyl glucuronide. Therefore, a decreased hepatic exposure to diclofenac acyl glucuronide is expected in patients with the UGT2B7*2 genotype. The increased risk for hepatotoxicity, therefore, might be the result from a shift to oxidative bioactivation to cytotoxic quinoneimines.

CYP130 belongs to the subset of cytochrome P450s from Mycobacterium tuberculosis (Mtb) that have been structurally characterized. Despite several efforts for its functional characterization, CYP130 is still considered an orphan enzyme for which no endogenous or exogenous substrate has been identified. In addition, functional redox-partners for CYP130 have not been clearly established yet, hampering the elucidation of its physiological role. In the present study, a catalytically active fusion protein involving CYP130 and the NADPH reductase-domain of CYP102A1 from Bacillus megaterium was created. By screening a panel of known substrates of human P450s, dextromethorphan N-demethylation was identified as a reaction catalyzed by CYP130. The fusion enzyme showed higher catalytic activity, when compared to CYP130 reconstituted with a selection of non-native redox-partners. Molecular dynamics simulation studies based on the crystal structure of CYP130 revealed two primary docking poses of dextromethorphan within the active site consistent with the experimentally observed N-demethylation reaction during the entire molecular dynamics simulation. The dextromethorphan N-demethylation reaction was strongly inhibited by azole-drugs and maybe applied to identify mechanism-based inhibitors of CYP130. Furthermore, the present active CYP130-fusion protein may facilitate the identification of endogenous substrates from Mtb.

Detoxicating enzymes NAD(P)H:quinone oxidoreductase 1 (NQO1) and NRH:quinone oxidoreductase 2 (NQO2) catalyze the two-electron reduction of quinone-like compounds. The protective role of the polymorphic NQO1 and NQO2 enzymes is especially of interest in the liver as the major site of drug bioactivation to chemically reactive drug metabolites. In the current study, we quantified the concentrations of NQO1 and NQO2 in 20 human liver donors and NQO1 and NQO2 activities with quinone-like drug metabolites. Hepatic NQO1 concentrations ranged from 8 to 213 nM. Using recombinant NQO1, we showed that low nM concentrations of NQO1 are sufficient to reduce synthetic amodiaquine and carbamazepine quinone-like metabolites in vitro. Hepatic NQO2 concentrations ranged from 2 to 31 μM. NQO2 catalyzed the reduction of quinone-like metabolites derived from acetaminophen, clozapine, 4′-hydroxydiclofenac, mefenamic acid, amodiaquine, and carbamazepine. The reduction of the clozapine nitrenium ion supports association studies showing that NQO2 is a genetic risk factor for clozapine-induced agranulocytosis. The 5-hydroxydiclofenac quinone imine, which was previously shown to be reduced by NQO1, was not reduced by NQO2. Tacrine was identified as a potent NQO2 inhibitor and was applied to further confirm the catalytic activity of NQO2 in these assays. While the in vivo relevance of NQO2-catalyzed reduction of quinone-like metabolites remains to be established by identification of the physiologically relevant co-substrates, our results suggest an additional protective role of the NQO2 protein by non-enzymatic scavenging of quinone-like metabolites. Hepatic NQO1 activity in detoxication of quinone-like metabolites becomes especially important when other detoxication pathways are exhausted and NQO1 levels are induced.

Cytochrome P450 aromatase (CYP19A1) plays a key role in the development of estrogen dependent breast cancer, and aromatase inhibitors have been at the front line of treatment for the past three decades. The development of potent, selective and safer inhibitors is ongoing with in silico screening methods playing a more prominent role in the search for promising lead compounds in bioactivity-relevant chemical space. Here we present a set of comprehensive binding affinity prediction models for CYP19A1 using our automated Linear Interaction Energy (LIE) based workflow on a set of 132 putative and structurally diverse aromatase inhibitors obtained from a typical industrial screening study. We extended the workflow with machine learning methods to automatically cluster training and test compounds in order to maximize the number of explained compounds in one or more predictive LIE models. The method uses protein-ligand interaction profiles obtained from Molecular Dynamics (MD) trajectories to help model search and define the applicability domain of the resolved models. Our method was successful in accounting for 86% of the data set in 3 robust models that show high correlation between calculated and observed values for ligand-binding free energies (RMSE < 2.5 kJ mol(-1)), with good cross-validation statistics.

Acetaminophen (APAP), although considered a safe drug, is one of the major causes of acute liver failure by overdose, and therapeutic chronic use can cause serious health problems. Although the reactive APAP metabolite N-acetyl-p-benzoquinoneimine (NAPQI) is clearly linked to liver toxicity, toxicity of APAP is also found without drug metabolism of APAP to NAPQI. To get more insight into mechanisms of APAP toxicity, a genome-wide screen in Saccharomyces cerevisiae for APAP-resistant deletion strains was performed. In this screen we identified genes related to the DNA damage response. Next, we investigated the link between genotype and APAP-induced toxicity or resistance by performing a more detailed screen with a library containing mutants of 1522 genes related to nuclear processes, like DNA repair and chromatin remodelling. We identified 233 strains that had an altered growth rate relative to wild type, of which 107 showed increased resistance to APAP and 126 showed increased sensitivity. Gene Ontology analysis identified ubiquitin homeostasis, regulation of transcription of RNA polymerase II genes, and the mitochondria-to-nucleus signalling pathway to be associated with APAP resistance, while histone exchange and modification, and vesicular transport were connected to APAP sensitivity. Indeed, we observed a link between ubiquitin levels and APAP resistance, whereby ubiquitin deficiency conferred resistance to APAP toxicity while ubiquitin overexpression resulted in sensitivity. The toxicity profile of various chemicals, APAP, and its positional isomer AMAP on a series of deletion strains with ubiquitin deficiency showed a unique resistance pattern for APAP. Furthermore, exposure to APAP increased the level of free ubiquitin and influenced the ubiquitination of proteins. Together, these results uncover a role for ubiquitin homeostasis in APAP-induced toxicity.

Amodiaquine (AQ), an antimalarial drug, widely prescribed in endemic areas of Africa and Asia, is used in combination with artesunate as recommended by the WHO. However, due to its idiosyncratic hepatotoxicity and agranulocytosis, the therapeutic use has been discontinued in most countries. Oxidative bioactivation to protein-reactive quinonimines (QIs) by hepatic cytochrome P450s and myeloperoxidase (MPO) have been suggested to be important mechanisms underlying AQ idiosyncratic toxicity. However, the inactivation of the reactive QIs by detoxifying enzymes such as human glutathione S-transferases (GSTs) and NAD(P)H:quinone oxidoreducatase 1 (NQO1) has not been characterized yet. In the present study, the activities of 15 recombinant human GSTs and NQO1 in the inactivation of reactive QIs of AQ and its pharmacological active metabolite, N-desethylamodiaquine (DEAQ) were investigated. The results showed that GSTP1-1, GSTA4-4, GSTM4-4, GSTM2-2 and GSTA2-2 (activity in decreasing order) were active isoforms in catalyzing GSH conjugation of reactive QIs of AQ and DEAQ. Additionally, NQO1 was shown to inactivate these QIs by reduction. Simulation of the variability of cytosolic GST-activity based on the hepatic GST contents from 22 liver donors, showed a large variation in cytosolic inactivation of QIs by GSH, especially at a reduced GSH-concentration. In conclusion, the present study demonstrates that a low hepatic expression of the active GSTs and NQO1 may increase the susceptibility of patients to AQ idiosyncratic hepatotoxicity.

Computational protein binding affinity prediction can play an important role in drug research but performing efficient and accurate binding free energy calculations is still challenging. In the context of phase 2 of the Drug Design Data Resource (D3R) Grand Challenge 2 we used our automated eTOX ALLIES approach to apply the (iterative) linear interaction energy (LIE) method and we evaluated its performance in predicting binding affinities for farnesoid X receptor (FXR) agonists. Efficiency was obtained by our pre-calibrated LIE models and molecular dynamics (MD) simulations at the nanosecond scale, while predictive accuracy was obtained for a small subset of compounds. Using our recently introduced reliability estimation metrics, we could classify predictions with higher confidence by featuring an applicability domain (AD) analysis in combination with protein–ligand interaction profiling. The outcomes of and agreement between our AD and interaction-profile analyses to distinguish and rationalize the performance of our predictions highlighted the relevance of sufficiently exploring protein–ligand interactions during training and it demonstrated the possibility to quantitatively and efficiently evaluate if this is achieved by using simulation data only. Electronic supplementary material The online version of this article (doi:10.1007/s10822-017-0055-0) contains supplementary material, which is available to authorized users.

Cytochrome P450 BM3 (CYP102A1) mutant M11 is able to metabolize a wide range of drugs and drug-like compounds. Among these, M11 was recently found to be able to catalyze formation of human metabolites of mefenamic acid and other nonsteroidal anti-inflammatory drugs (NSAIDs). Interestingly, single active-site mutations such as V87I were reported to invert regioselectivity in NSAID hydroxylation. In this work, we combine crystallography and molecular simulation to study the effect of single mutations on binding and regioselective metabolism of mefenamic acid by M11 mutants. The heme domain of the protein mutant M11 was expressed, purified, and crystallized, and its X-ray structure was used as template for modeling. A multistep approach was used that combines molecular docking, molecular dynamics (MD) simulation, and binding free-energy calculations to address protein flexibility. In this way, preferred binding modes that are consistent with oxidation at the experimentally observed sites of metabolism (SOMs) were identified. Whereas docking could not be used to retrospectively predict experimental trends in regioselectivity, we were able to rank binding modes in line with the preferred SOMs of mefenamic acid by M11 and its mutants by including protein flexibility and dynamics in free-energy computation. In addition, we could obtain structural insights into the change in regioselectivity of mefenamic acid hydroxylation due to single active-site mutations. Our findings confirm that use of MD and binding free-energy calculation is useful for studying biocatalysis in those cases in which enzyme binding is a critical event in determining the selective metabolism of a substrate.

Nevirapine (NVP) is a non-nucleoside reverse transcriptase-inhibitor, which is associated with severe idiosyncratic skin rash and hepatotoxicity. These adverse drug reactions are believed to be mediated by the formation of epoxides and/or quinone methide formed by oxidative metabolism by P450s and 12-sulfoxyl-NVP formed by sequential 12-hydroxylation and O-sulfonation. Although different GSH-conjugates and corresponding mercapturic acids have been demonstrated previously in vitro and in vivo, the role of the glutathione S-transferases in the inactivation of the different reactive metabolites has not been studied so far. In the present study the activity of 10 recombinant human glutathione S-transferases (GSTs) in the detoxification of the different reactive metabolites of NVP was studied. The results show that GSTP1-1 is a highly active catalyst of GSH-conjugation of the oxidative metabolites of NVP, even at high GSH-concentration. Experiments with trideuterated NVP suggest involvement of a reactive epoxide rather than quinone methide in the formation of the GSH-conjugate formed after oxidative bioactivation. GSH-conjugation of 12-sulfoxyl-NVP forming NVP-12-GSH was only catalyzed by GSTM1-1, GSTA1-1, and GSTA3-3. Although the exact expression levels of these enzymes in the skin is unknown, the relatively low activity of this catalysis makes it unlikely that GSTs can provide significant protection against this metabolite. However, since NVP-12-GSH is specifically formed via the 12-sulfoxyl-NVP, its corresponding urinary mercapturic acid can be considered as a biomarker for recent internal exposure to this protein-reactive sulfate. However, it has to be taken into account that 12-sulfoxyl-NVP is not completely trapped by GSH and that rates of bioinactivation will differ between patients due to variability in expression of GSTM1, GSTA1, and GSTA3.

Theoretically, both synthetic endocrine disrupting chemicals (S-EDCs) and natural (exogenous and endogenous) endocrine disrupting chemicals (N-EDCs) can interact with endocrine receptors and disturb hormonal balance. However, compared to endogenous hormones, S-EDCs are only weak partial agonists with receptor affinities several orders of magnitude lower. Thus, to elicit observable effects, S-EDCs require considerably higher concentrations to attain sufficient receptor occupancy or to displace natural hormones and other endogenous ligands. Significant exposures to exogenous N-EDCs may result from ingestion of foods such as soy-based diets, green tea and sweet mustard. While their potencies are lower as compared to natural endogenous hormones, they usually are considerably more potent than S-EDCs. Effects of exogenous N-EDCs on the endocrine system were observed at high dietary intakes. A causal relation between their mechanism of action and these effects is established and biologically plausible. In contrast, the assumption that the much lower human exposures to S-EDCs may induce observable endocrine effects is not plausible. Hence, it is not surprising that epidemiological studies searching for an association between S-EDC exposure and health effects have failed. Regarding testing for potential endocrine effects, a scientifically justified screen should use in vitro tests to compare potencies of S-EDCs with those of reference N-EDCs. When the potency of the S-EDC is similar or smaller than that of the N-EDC, further testing in laboratory animals and regulatory consequences are not warranted.

Human glutathione transferase T2-2 (GSTT2-2) is one of the enzymes considered to play a role in inactivation of toxicants and carcinogens. The expression level of this enzyme is determined by genetic and environmental factors, which may lead to differences in susceptibility. As a specific assay for GSTT2-2 so far a spectroscopical assay based on GSH-conjugation of menaphthyl sulfate (MSu) was used. This spectrophotometric assay, however, appeared too insensitive to accurately quantify the GSTT2-2 activities in a panel of 20 human liver samples. More recently, expression levels of GSTT2-2 in biological samples are quantified by measuring mRNA levels. Since mRNA-levels do not always correlate well with enzyme activity, a specific and sensitive assay is required. In the present study a highly sensitive high-performance liquid chromatography (HPLC)-based method was developed. By applying the new method, firstly, the specificity of GSTT2-2 among 15 recombinant human GST isoforms in catalyzing GSH-conjugation of MSu was confirmed. In addition, a 65-fold inter-individual variation of GSTT2-2 activity was found from the individual liver fractions. By applying the method to individual liver fractions, a 65-fold inter-individual variation of GSTT2-2 activity was found. As a second application, the role of GSTT2-2 in GSH-conjugation of the environmental carcinogen 1-methylpyrene sulfate (MPS) was studied by correlation analysis with GSTT2-2-catalyzed MSu conjugation. The relatively poor correlation suggested that other GSTs also contribute to MPS-conjugation, as confirmed by incubations with recombinant GSTs.

The 5′-hydroxymethyl metabolite of the penicillin based antibiotic flucloxacillin (FLX) is considered to be involved in bile duct damage occurring in a small number of patients. Because 5′-hydroxymethyl FLX is difficult to obtain by organic synthesis, biosynthesis using highly active and regioselective biocatalysts would be an alternative approach. By screening an in-house library of Cytochrome P450 (CYP) BM3 mutants, mutant M11 L437E was identified as a regioselective enzyme with relatively high activity in production of 5′-hydroxymethyl FLX as was confirmed by mass spectrometry and NMR. In contrast, incubation of M11 L437E and other mutants with oxacillin (OX, which differs from FLX by a lack of aromatic halogens) resulted in formation of two metabolites. In addition to 5′-hydroxymethyl OX we identified a product resulting from aromatic hydroxylation. In silico studies of both FLX and OX with three CYP BM3 mutants revealed substrate binding poses allowing for 5′-methyl hydroxylation, as well as binding poses with the aromatic moiety in the vicinity of the heme iron for which the corresponding product of aromatic hydroxylation was not observed for FLX. Supported by the (differences in) experimentally determined ratios of product formation for OX hydroxylation by M11 and its L437A variant and M11 L437E, Molecular Dynamics simulations suggest that the preference of mutant M11 L437E to bind FLX in its catalytically active pose over the other binding orientation contributes to its biocatalytic activity, highlighting the benefit of studying effects of active-site mutations on possible alternative enzyme-substrate binding poses in protein engineering.

Formation of the reactive amodiaquine quinoneimine (AQ-QI) and N-desethylamodiaquine quinoneimine (DEAQ-QI) plays an important role in the toxicity of the anti-malaria drug amodiaquine (AQ). Glutathione conjugation protects against AQ-induced toxicity and GSTP1 is able to conjugate its quinoneimine metabolites AQ-QI and DEA-QI with glutathione. In this study, HepG2 cells transiently transfected with the human GSTP1 construct were utilized to investigate the protective effect of GSTP1 in a cellular context. HepG2 cells were exposed to synthesized QIs, which bypasses the need for intracellular bioactivation of AQ or DEAQ. Exposure was accompanied by decreased cell viability, increased caspase 3 activity, and decreased intracellular GSH levels. Using high-content imaging-based BAC-GFP reporters, it was shown that AQ-QI and DEAQ-QI specifically activated the endoplasmic reticulum (ER) stress response. In contrast, oxidative stress, DNA damage, or inflammatory stress responses were not activated. Overexpression of GSTP1 resulted in a two-fold increase in GSH-conjugation of the QIs, attenuated QI-induced cytotoxicity especially under GSH-depletion condition, abolished QIs-induced apoptosis but did not significantly inhibit the activation of the ER stress response. In conclusion, these results indicate a protective role of GSTP1 by increasing enzymatic detoxification of AQ-QI and DEAQ-QI and suggest a second protective mechanism by interfering with ER stress induced apoptosis.

Formation of the reactive amodiaquine quinoneimine (AQ-QI) and N-desethylamodiaquine quinoneimine (DEAQ-QI) plays an important role in the toxicity of the anti-malaria drug amodiaquine (AQ). Glutathione conjugation protects against AQ-induced toxicity and GSTP1 is able to conjugate its quinoneimine metabolites AQ-QI and DEA-QI with glutathione. In this study, HepG2 cells transiently transfected with the human GSTP1 construct were utilized to investigate the protective effect of GSTP1 in a cellular context. HepG2 cells were exposed to synthesized QIs, which bypasses the need for intracellular bioactivation of AQ or DEAQ. Exposure was accompanied by decreased cell viability, increased caspase 3 activity, and decreased intracellular GSH levels. Using high-content imaging-based BAC-GFP reporters, it was shown that AQ-QI and DEAQ-QI specifically activated the endoplasmic reticulum (ER) stress response. In contrast, oxidative stress, DNA damage, or inflammatory stress responses were not activated. Overexpression of GSTP1 resulted in a two-fold increase in GSH-conjugation of the QIs, attenuated QI-induced cytotoxicity especially under GSH-depletion condition, abolished QIs-induced apoptosis but did not significantly inhibit the activation of the ER stress response. In conclusion, these results indicate a protective role of GSTP1 by increasing enzymatic detoxification of AQ-QI and DEAQ-QI and suggest a second protective mechanism by interfering with ER stress induced apoptosis.

Background: Inter-individual variability in hepatic drug metabolizing enzyme (DME) activity is a major contributor to heterogeneity in drug clearance and safety. Accurate data on expression levels and activities of DMEs is an important prerequisite for in vitro-in vivo extrapolation and in silico based predictions. Characterization and assessment of inter-correlations of the major DMEs cytochrome P450s (CYPs) and UDP-glucuronosyltransferases (UGTs) have been extensively documented, but simultaneous quantification including other major DMEs has been lacking. Objective: Assessment of inter-donor variability and inter-correlations of CYPs, UGTs, sulfotransferases (SULTs), glutathione S-transferases (GSTs), NAD(P)H:quinone oxidoreductase 1 (NQO1) and NRH: quinone oxidoreductase 2 (NQO2) in a set of 20 individual liver homogenates. Method: The main drug metabolizing isoforms of CYP and UGT have been reaction phenotype in individual liver microsomes and NQO1, NQO2, GSTT1 and GSTT2 in corresponding cytosol. In addition, we assessed overall SULT activity in liver cytosol using acetaminophen and 7-hydroxycoumarin as non-selective substrates and cytosolic GST activity using the non-selective substrate 1-chloro-2,4-dinitrobenzene (CDNB). Expression of GST isoforms was also assessed. Results and Conclusion: While hepatic NQO1 activity was highly variable, NQO2 activity was more conserved. In addition, we found that of the hepatic GST isoforms, the variation in GSTM3 levels, which is poorly studied, was highest. The majority of significant correlations were found amongst CYP and UGT enzyme activities. The dataset presented provides the absolute quantification of the largest number of hepatic DME activities so far and constitute an essential resource for in silico toxicokinetic and metabolic modelling studies.

In the present study, the whole-cell biotransformation using strains expressing CYP BM3 mutants has been evaluated for the stereoselective hydroxylation of the steroid norethisterone (NET), a widely used contraceptive. First, an in vitro CYP BM3 mutant library screen was performed to identify mutants with high activity, as well asstereoselectivity. Subsequently, two different whole-cell setups (resting and growing cells) were tested in two different host organisms ( Escherichia coli and Saccharomyces cerevisiae ) expressing CYP BM3. It was found that resting E. coli whole cells produced the highest amounts of products and therefore this biocatalytic setup was further optimized for application in a laboratory-scale fermentor. In the optimized fermentor setup, high product yields (0.3 g/L 15β-OH-NET and 0.16 g/L 16β-OH-NET) were achieved while it was also shown that the regio- and stereoselectivities of the steroid hydroxylations, as determined during the in vitro library screen, were preserved in the whole-cell system. The combination of a mutant CYP BM3 library and the optimized whole-cell oxidation system represents a promising and cost-effective alternative to a wide range of in vitro biosynthetic routes.

In the present study, the use of Rhodococcus erythropolis mutant strain RG9 expressing the cytochrome P450 BM3 mutant M02 enzyme has been evaluated for whole-cell biotransformation of a 17-ketosteroid, norandrostenedione, as a model substrate. Purified P450 BM3 mutant M02 enzyme hydroxylated the steroid with >95 % regioselectivity to form 16-β-OH norandrostenedione, as confirmed by NMR analysis. Whole cells of R. erythropolis RG9 expressing P450 BM3 M02 enzyme also converted norandrostenedione into the 16-β-hydroxylated product, resulting in the formation of about 0.35 g/L. Whole cells of strain RG9 itself did not convert norandrostenedione, indicating that metabolite formation is P450 BM3 M02 enzyme mediated. This study shows that R. erythropolis is a novel and interesting host for the heterologous expression of highly selective and active P450 BM3 M02 enzyme variants able to perform steroid bioconversions.

Early prediction of safety issues in drug development is at the same time highly desirable and highly challenging. Recent advances emphasize the importance of understanding the whole chain of causal events leading to observable toxic outcomes. Here we describe an integrative modeling strategy based on these ideas that guided the design of eTOXsys, the prediction system used by the eTOX project. Essentially, eTOXsys consists of a central server that marshals requests to a collection of independent prediction models and offers a single user interface to the whole system. Every of such model lives in a self-contained virtual machine easy to maintain and install. All models produce toxicity-relevant predictions on their own but the results of some can be further integrated and upgrade its scale, yielding in vivo toxicity predictions. Technical aspects related with model implementation, maintenance and documentation are also discussed here. Finally, the kind of models currently implemented in eTOXsys is illustrated presenting three example models making use of diverse methodology (3D-QSAR and decision trees, Molecular Dynamics simulations and Linear Interaction Energy theory, and fingerprint-based QSAR).

Tri-ortho-cresyl phosphate (ToCP) is a multipurpose organophosphorus compound that is neurotoxic and suspected to be involved in aerotoxic syndrome in humans. It has been reported that not ToCP itself but a metabolite of ToCP, namely, 2-(ortho-cresyl)-4H-1,2,3-benzodioxaphosphoran-2-one (CBDP), may be responsible for this effect as it can irreversibly bind to human butyrylcholinesterase (BuChE) and human acetylcholinesterase (AChE). The bioactivation of ToCP into CBDP involves Cytochrome P450s (P450s). However, the individual human P450s responsible for this bioactivation have not been identified yet. In the present study, we aimed to investigate the metabolism of ToCP by different P450s and to determine the inhibitory effect of the in vitro generated ToCP-metabolites on human BuChE and AChE. Human liver microsomes, rat liver microsomes, and recombinant human P450s were used for that purpose. The recombinant P450s 2B6, 2C18, 2D6, 3A4 and 3A5 showed highest activity of ToCP-bioactivation to BuChE-inhibitory metabolites. Inhibition experiments using pooled human liver microsomes indicated that P450 3A4 and 3A5 were mainly involved in human hepatic bioactivation of ToCP. In addition, these experiments indicated a minor role for P450 1A2. Formation of CBDP by in-house expressed recombinant human P450s 1A2 and 3A4 was proven by both LC-MS and GC-MS analysis. When ToCP was incubated with P450 1A2 and 3A4 in the presence of human BuChE, CBDP-BuChE-adducts were detected by LC-MS/MS which were not present in the corresponding control incubations. These results confirmed the role of human P450s 1A2 and 3A4 in ToCP metabolism and demonstrated that CBDP is the metabolite responsible for the BuChE inactivation. Interindividual differences at the level of P450 1A2 and 3A4 might play an important role in the susceptibility of humans in developing neurotoxic effects, such as aerotoxic syndrome, after exposure to ToCP. © 2015 American Chemical Society

Cyclophosphamide (CPA) and ifosfamide (IFA) are widely used anticancer agents that require metabolic activation by cytochrome P450 (CYP) enzymes. While 4-hydroxylation yields DNA-alkylating and cytotoxic metabolites, N-dechloroethylation results in the generation of neuro- and nephrotoxic byproducts. Gene-directed enzyme prodrug therapies (GDEPT) have been suggested to facilitate local CPA and IFA bioactivation by expressing CYP enzymes within the tumor cells, thereby increasing efficacy. We screened bacterial CYP BM3 mutants, previously engineered to metabolize drug-like compounds, for their ability to catalyze 4-hydroxylation of CPA and IFA. Two CYP BM3 mutants showed very rapid initial bioactivation of CPA and IFA, followed by a slower phase of product formation. N-dechloroethylation by these mutants was very low (IFA) to undetectable (CPA). Using purified CYP BM3 as an extracellular bioactivation tool, cytotoxicity of CPA and IFA metabolism was confirmed in U2OS cells. This novel application of CYP BM3 possibly provides a clean and catalytically efficient alternative to liver microsomes or S9 for the study of CYP-mediated drug toxicity. To our knowledge, the observed rate of CPA and IFA 4-hydroxylation by these CYP BM3 mutants is the fastest reported to date, and might be of potential interest for CPA and IFA GDEPT.

UDP-glucuronosyltransferases (UGTs) and cytochrome P450s (CYPs) are the major enzymes involved in hepatic metabolism of drugs. Hepatic drug metabolism is commonly investigated using human liver microsomes (HLM) or primary human hepatocytes (PHH). We describe the development of a sensitive assay to phenotype activities of six major hepatic UGT isoforms (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9 and UGT2B7) in intact PHH by analysis of glucuronidation of selective probe substrates. The non-selective, general substrate 7-hydroxycoumarin was included for comparison. For each liver donor preparation (five donors) UGT activities in cryopreserved suspended and plated PHH were compared to HLM prepared from the same donors. Standard CYP reaction phenotyping of seven major isoforms was performed in parallel. For all donors, CYP- and UGT-isoforms activity profiles were comparable in PHH and HLM, indicating that reaction phenotyping with selective probe substrates in intact cells primarily reflects respective CYP or UGT activity. System-dependent effects on UGT and CYP isoform activity were still found. While UGT activity of UGT1A1 was equivalent in plated and suspended PHH, UGT1A3, UGT1A6 and UGT2B7 activity was higher in suspended PHH and UGT1A9 and UGT1A4 activity was higher in plated PHH. The well-known decrease in activity of most CYP isoforms in plated compared to suspended PHH was confirmed. Importantly, we found a significant loss in CYP2C19 and CYP2B6 in HLM, activity being lower than in intact cells. Taken together, these findings implicate that, dependent on the UGT or CYP isoforms involved in the metabolism of a given compound, the outcome of metabolic assays is strongly dependent on the choice of the in vitro system. The currently described UGT- and CYP- activity profiling method can be used as a standard assay in intact cells and can especially aid in reaction phenotyping of in vitro systems for which a limited number of cells are available.

Post-translational protein modification by addition or removal of the small polypeptide ubiquitin is involved in a range of critical cellular processes, like proteasomal protein degradation, DNA repair, gene expression, internalization of membrane proteins, and drug sensitivity. We recently identified genes important for acetaminophen (APAP) toxicity in a comprehensive screen and our findings suggested that a small set of yeast strains carrying deletions of ubiquitin-related genes can be informative for drug toxicity profiling. In yeast, approximately 20 different deubiquitinating enzymes (DUBs) have been identified, of which only one is essential for viability. We investigated whether the toxicity profile of DUB deletion yeast strains would be informative about the toxicological mode of action of APAP. A set of DUB deletion strains was tested for sensitivity and resistance to a diverse series of compounds, including APAP, quinine, ibuprofen, rapamycin, cycloheximide, cadmium, peroxide and amino acids and a cluster analysis was performed. Most DUB deletion strains showed an altered growth pattern when exposed to these compounds by being either more sensitive or more resistant than WT. Toxicity profiling of the DUB strains revealed a remarkable overlap between the amino acid tyrosine and acetaminophen (APAP), but not its stereoisomer AMAP. Furthermore, co-exposure of cells to both APAP and tyrosine showed an enhancement of the cellular growth inhibition, suggesting that APAP and tyrosine have a similar mode of action.

Mefenamic acid (MFA) has been associated with rare but severe cases of hepatotoxicity, nephrotoxicity, gastrointestinal toxicity, and hypersensitivity reactions that are believed to result from the formation of reactive metabolites. Although formation of protein-reactive acylating metabolites by phase II metabolism has been well-studied and proposed to be the cause of these toxic side effects, the oxidative bioactivation of MFA has not yet been competely characterized. In the present study, the oxidative bioactivation of MFA was studied using human liver microsomes (HLM) and recombinant human P450 enzymes. In addition to the major metabolite 3′-OH-methyl-MFA, resulting from the benzylic hydroxylation by CYP2C9, 4′-hydroxy-MFA and 5-hydroxy-MFA were identified as metabolites resulting from oxidative metabolism of both aromatic rings of MFA. In the presence of GSH, three GSH conjugates were formed that appeared to result from GSH conjugation of the two quinoneimines formed by further oxidation of 4′-hydroxy-MFA and 5-hydroxy-MFA. The major GSH conjugate was identified as 4′-OH-5′-glutathionyl-MFA and was formed at the highest activity by CYP1A2 and to a lesser extent by CYP2C9 and CYP3A4. Two minor GSH conjugates resulted from secondary oxidation of 5-hydroxy-MFA and were formed at the highest activity by CYP1A2 and to a lesser extent by CYP3A4. Additionally, the ability of seven human glutathione S-transferases (hGSTs) to catalyze the GSH conjugation of the quinoneimines formed by P450s was also investigated. The highest increase of total GSH conjugation was observed with hGSTP1-1, followed by hepatic hGSTs hGSTA2-2 and hGSTM1-1. The results of this study show that, next to phase II metabolites, reactive quinoneimines formed by oxidative bioactivation might also contribute to the idiosyncratic toxicity of MFA. (Chemical Presented).

NAD(P)H: quinone oxidoreductase 1 (NQO1) is an enzyme capable of reducing a broad range of chemically reactive quinones and quinoneimines (QIs) and can be strongly upregulated by Nrf2/Keap1-mediated stress responses. Several commonly used drugs implicated in adverse drug reactions (ADRs) are known to form reactive QI metabolites upon bioactivation by P450, such as acetaminophen (APAP), diclofenac (DF), and mefenamic acid (MFA). In the present study, the reductive activity of human NQO1 toward the QI metabolites derived from APAP and hydroxy-metabolites of DF and MFA was studied, using purified bacterial P450 BM3 (CYP102A1) mutant M11 as a bioactivation system. The NQO1-catalyzed reduction of the QI metabolites was quantified relative to spontaneous glutathione (GSH) conjugation. Addition of NQO1 to the incubations strongly reduced the formation of all corresponding GSH conjugates, and this activity could be prevented by dicoumarol, a selective NQO1 inhibitor. The GSH conjugation was strongly increased by adding human GSTP1-1 in a wide range of GSH concentrations. Still, NQO1 could effectively compete with the GST catalyzed GSH conjugation by reducing the QIs. In conclusion, we identified the QI metabolites of the 4'- and 5-hydroxy-metabolites of DF and MFA as novel substrates for human NQO1. NQO1-mediated reduction proves to be an effective pathway to detoxify these QI metabolites in addition to GSH conjugation. Genetically determined deficiency of NQO1 therefore might be a risk factor for ADRs induced by reactive QI drug metabolites.

Idiosyncratic adverse drug reactions due to the anti-inflammatory drug diclofenac have been proposed to be caused by the generation of reactive acyl glucuronides and oxidative metabolites. For the oxidative metabolism of diclofenac by cytochromes P450 at least five different reactive intermediates have been proposed previously based on structural identification of their corresponding GSH-conjugates. In the present study, the ability of four human glutathione S-transferases (hGSTs) to catalyze the GSH-conjugation of the different reactive intermediates formed by P450s was investigated. Addition of pooled human liver cytosol and recombinant hGSTA1-1, hGSTM1-1, and hGSTP1-1 to incubations of diclofenac with human liver microsomes or purified CYP102A1M11 L437N as a model system significantly increased total GSH-conjugation. The strongest increase of total GSH-conjugation was observed by adding hGSTP1-1, whereas hGSTM1-1 and hGSTA1-1 showed lower activity. Addition of hGSTT1-1 only showed a minor effect. When considering the effects of hGSTs on GSH-conjugation of the different quinoneimines of diclofenac, it was found that hGSTP1-1 showed the highest activity in GSH-conjugation of the quinoneimine derived from 5-hydroxydiclofenac (5-OH-DF). hGSTM1-1 showed the highest activity in inactivation of the quinoneimine derived from 4′-hydroxydiclofenac (4′-OH-DF). Separate incubations with 5-OH-DF and 4′-OH-DF as substrates confirmed these results. hGSTs also catalyzed GSH-conjugation of the o-iminemethide formed by oxidative decarboxylation of diclofenac as well as the substitution of one of the chlorine atoms of DF by GSH. hGSTP1-1 showed the highest activity for the formation of these minor GSH-conjugates. These results suggest that hGSTs may play an important role in the inactivation of DF quinoneimines and its minor reactive intermediates especially in stress conditions when tissue levels of GSH are decreased. © 2013 American Chemical Society.

Mutants of bacterial cytochrome P450 BM3 from Bacillus megaterium have gained increasing interest to support drug metabolism studies by producing human relevant biotransformation products and/or to support drug discovery by producing libraries of new chemical entities based on initial lead compounds. In this study, a library of 33 P450 BM3 mutants was used to diversify TAK-715, a representative (highly lipophilic) inhibitor of p38α mitogen-activated protein kinase (p38α). To simultaneously determine the individual bioaffinity and identity of the different products, an analytical high-resolution screening approach was used, based on post-column on-line bioaffinity profiling with parallel mass spectrometric identification. The screening of the product mixtures produced demonstrated that introducing mutations in the active site of the P450 BM3 resulted in different product profiles. Several P450 BM3 mutants were mimicking the metabolic profile obtained by human liver microsomes. The major biotransformation products of TAK-715 could be produced with the most active P450 BM3 mutant in sufficient amounts to enable further structure elucidation by 1H-NMR and quantification of their p38α affinity. © The Royal Society of Chemistry.

Clozapine, an often-prescribed antipsychotic drug, is implicated in severe adverse drug reactions (ADRs). Formation of reactive intermediates by cytochrome P450s (CYPs) has been proposed as a possible explanation for these ADRs. Moreover, a protective role for human glutathione S-transferases (hGSTs) was recently shown using purified enzymes. We investigated the interplay between CYP bioactivation and GST detoxification in a reconstituted cellular context using recombinant yeast expressing a bacterial CYP BM3 mutant (M11), mimicking the drug-metabolizing potential of human CYPs, combined with hGSTA1-1, M1-1 or P1-1. Clozapine and the N-desmethylclozapine metabolite caused comparable growth inhibition and reactive oxygen species (ROS) formation, whereas the clozapine-N-oxide metabolite was clearly less toxic. Clozapine metabolism by BM3 M11 and the hGSTs in yeast was confirmed by identification of stable clozapine metabolites and hGST isoform-specific glutathione-conjugates. Oxidative metabolism of clozapine by BM3 M11 increased ROS formation and growth inhibition. Co-expression of hGSTP1-1 protected yeast from BM3 M11 induced growth inhibition in presence of clozapine, whereas similar expression levels of hGSTA1-1 and hGSTM1-1 did not. ROS formation was not lowered by hGSTP1-1 co-expression and was unrelated to mitochondrial electron transport chain (mETC) activity. We present a novel cellular model to study the effect of CYP and GST interplay in drug toxicity. © 2013 Elsevier Ireland Ltd.

Modification of cellular macromolecules by reactive drug metabolites is considered to play an important role in the initiation of tissue injury by many drugs. Detection and identification of reactive intermediates is often performed by analyzing the conjugates formed after trapping by glutathione (GSH). Although sensitivity of modern mass spectrometrical methods is extremely high, absolute quantification of GSH-conjugates is critically dependent on the availability of authentic references. Although 1H NMR is currently the method of choice for quantification of metabolites formed biosynthetically, its intrinsically low sensitivity can be a limiting factor in quantification of GSH-conjugates which generally are formed at low levels. In the present study, a simple but sensitive and generic method for absolute quantification of GSH-conjugates is presented. The method is based on quantitative alkaline hydrolysis of GSH-conjugates and subsequent quantification of glutamic acid and glycine by HPLC after precolumn derivatization with o-phthaldialdehyde/N-acetylcysteine (OPA/NAC). Because of the lower stability of the glycine OPA/NAC-derivate, quantification of the glutamic acid OPA/NAC-derivate appeared most suitable for quantification of GSH-conjugates. The novel method was used to quantify the concentrations of GSH-conjugates of diclofenac, clozapine and acetaminophen and quantification was consistent with 1H NMR, but with a more than 100-fold lower detection limit for absolute quantification.

Recently an iterative method was proposed to enhance the accuracy and efficiency of ligand-protein binding affinity prediction through linear interaction energy (LIE) theory. For ligand binding to flexible Cytochrome P450s (CYPs), this method was shown to decrease the root-mean-square error and standard deviation of error prediction by combining interaction energies of simulations starting from different conformations. Thereby, different parts of protein-ligand conformational space are sampled in parallel simulations. The iterative LIE framework relies on the assumption that separate simulations explore different local parts of phase space, and do not show transitions to other parts of configurational space that are already covered in parallel simulations. In this work, a method is proposed to (automatically) detect such transitions during the simulations that are performed to construct LIE models and to predict binding affinities. Using noise-canceling techniques and splines to fit time series of the raw data for the interaction energies, transitions during simulation between different parts of phase space are identified. Boolean selection criteria are then applied to determine which parts of the interaction energy trajectories are to be used as input for the LIE calculations. Here we show that this filtering approach benefits the predictive quality of our previous CYP 2D6-aryloxypropanolamine LIE model. In addition, an analysis is performed of the gain in computational efficiency that can be obtained from monitoring simulations using the proposed filtering method and by prematurely terminating simulations accordingly. Electronic supplementary material The online version of this article (doi:10.1007/s00894-015-2883-y) contains supplementary material, which is available to authorized users.

Idiosyncratic drug-induced lever injury (IDILI) is a rare but severe side effect of diclofenac (DF). Several mechanisms have been proposed as cause of DF-induced toxicity including the formation of protein-reactive diclofenac-1',4'-quinone imine (DF-1',4'-QI) and diclofenac-2,5-quinone imine (DF-2,5-QI). Formation of these p-benzoquinone imines result from two-step oxidative metabolism involving aromatic hydroxylation to 4'-hydroxydiclofenac and 5-hydroxydiclofenac followed by dehydrogenation to DF-1',4'-QI and DF-2,5-QI, respectively. Although the contribution of individual cytochrome P450s (CYPs) in aromatic hydroxylation of DF is well studied, the enzymes involved in the dehydrogenation reactions have been poorly characterized. The results of the present study show that both formation of 4'-hydroxydiclofenac and it subsequent bioactivation to DF-1',4'-QI is selectively catalyzed by CYP2C9. However, the two-step bioactivation to DF-2,5-QI appears to be catalyzed with highest activity by two different CYPs: 5-hydroxylation of DF is predominantly catalyzed by CYP3A4, whereas its subsequent bioactivation to DF-2,5-QI is catalyzed with 14-fold higher intrinsic clearance by CYP2C9. The fact that both CYPs involved in two-step bioactivation of DF show large interindividual variability may play a role in different susceptibility of patients to DF-induced IDILI. Furthermore, expression levels of these enzymes and protective enzymes might be important factors determining sensitivity of in vitro models for hepatotoxicity.

Cytochrome P450s (CYPs), UDP-glucuronosyltransferases (UGTs) and sulfotransferases (SULTs) are the most important enzymes for metabolic clearance. Characterization of phase I and phase II metabolism of a given drug in cellular models is therefore important for an adequate interpretation of the role of drug metabolism in toxicity. We investigated phase I (CYP) and phase II (UGT and SULT) metabolism of three drugs related to drug-induced liver injury (DILI), namely acetaminophen (APAP), diclofenac (DF) and tolcapone (TC), in cryopreserved primary human hepatocytes from 5 donors in suspension and monolayer. The general phase II substrate 7-hydroxycoumarin (7-HC) was included for comparison. Our results show that the decrease in CYP, UGT and SULT activity after plating is substrate dependent. As a consequence the phase I/phase II metabolism ratio is significantly affected, with a shift in monolayer towards phase I metabolism for TC and towards phase II metabolism for APAP and DF. Inter-donor variability in drug metabolism is significant, especially in sulfation of 7-HC or APAP. As CYP, UGT and SULT metabolism may lead to bioactivation and/or detoxification of drugs, a changed ratio in phase I/phase II metabolism may have important consequences for metabolism-related toxicity.

The current test systems employed by pharmaceutical industry are poorly predictive for drug-induced liver injury (DILI). The ‘MIP-DILI’ project addresses this situation by the development of innovative preclinical test systems which are both mechanism-based and of physiological, pharmacological and pathological relevance to DILI in humans. An iterative, tiered approach with respect to test compounds, test systems, bioanalysis and systems analysis is adopted to evaluate existing models and develop new models that can provide validated test systems with respect to the prediction of specific forms of DILI and further elucidation of mechanisms. An essential component of this effort is the choice of compound training set that will be used to inform refinement and/or development of new model systems that allow prediction based on knowledge of mechanisms, in a tiered fashion. In this review, we focus on the selection of MIP-DILI training compounds for mechanism-based evaluation of non-clinical prediction of DILI. The selected compounds address both hepatocellular and cholestatic DILI patterns in man, covering a broad range of pharmacologies and chemistries, and taking into account available data on potential DILI mechanisms (e.g. mitochondrial injury, reactive metabolites, biliary transport inhibition, and immune responses). Known mechanisms by which these compounds are believed to cause liver injury have been described, where many if not all drugs in this review appear to exhibit multiple toxicological mechanisms. Thus, the training compounds selection offered a valuable tool to profile DILI mechanisms and to interrogate existing and novel in vitro systems for the prediction of human DILI.