A novel antibiotic class targeting the lipopolysaccharide transporter.

Carbapenem-resistant Acinetobacter baumannii (CRAB) has emerged as a major global pathogen with limited treatment options1. No new antibiotic chemical class with activity against A. baumannii has reached patients in over 50 years1. Here we report the identification and optimization of tethered macrocyclic peptide (MCP) antibiotics with potent antibacterial activity against CRAB. The mechanism of action of this molecule class involves blocking the transport of bacterial lipopolysaccharide from the inner membrane to its destination on the outer membrane, through inhibition of the LptB2FGC complex. A clinical candidate derived from the MCP class, zosurabalpin (RG6006), effectively treats highly drug-resistant contemporary isolates of CRAB both in vitro and in mouse models of infection, overcoming existing antibiotic resistance mechanisms. This chemical class represents a promising treatment paradigm for patients with invasive infections due to CRAB, for whom current treatment options are inadequate, and additionally identifies LptB2FGC as a tractable target for antimicrobial drug development.

Combining In Vivo and Organotypic In Vitro Approaches to Assess the Human Relevance of Basimglurant (RG7090), a Potential CAR Activator.

Basimglurant (RG7090), a small molecule under development to treat certain forms of depression, demonstrated foci of altered hepatocytes in a long-term rodent-toxicity study. Additional evidence pointed toward the activation of the constitutive androstane receptor (CAR), an established promoter of nongenotoxic and rodent-specific hepatic tumors. This mode of action and the potential human relevance was explored in vivo using rodent and cynomolgus monkey models and in vitro using murine and human liver spheroids. Wild type (WT) and CAR/pregnane X receptor (PXR) knockout mice (CAR/PXR KO) were exposed to RG7090 for 8 consecutive days. Analysis of liver lysates revealed induction of Cyp2b mRNA and enzyme activity, a known activation marker of CAR, in WT but not in CAR/PXR KO animals. A series of proliferative genes were upregulated in WT mice only, and immunohistochemistry data showed increased cell proliferation exclusively in WT mice. In addition, primary mouse liver spheroids were challenged with RG7090 in the presence or absence of modified antisense oligonucleotides inhibiting CAR and/or PXR mRNA, showing a concentration-dependent Cyp2b mRNA induction only if CAR was not repressed. On the contrary, neither human liver spheroids nor cynomolgus monkeys exposed to RG7090 triggered CYP2B mRNA upregulation. Our data suggested RG7090 to be a rodent-specific CAR activator, and that CAR activation and its downstream processes were involved in the foci of altered hepatocytes formation detected in vivo. Furthermore, we demonstrated the potential of a new in vitro approach using liver spheroids and antisense oligonucleotides for CAR knockdown experiments, which could eventually replace in vivo investigations using CAR/PXR KO mice.

In Vitro Assessment of the Hepatotoxicity Potential of Therapeutic Oligonucleotides.

The mechanisms of antisense oligonucleotide-induced liver toxicity are still poorly understood. Assessment of the hepatic safety profile is currently mostly investigated directly in rodent studies. A predictive preclinical in vitro model that is capturing liver liabilities of antisense oligonucleotides can be of great help to be used as a first filter in the screening process of therapeutic oligonucleotides. We describe here an in vitro cytotoxicity assay using freshly isolated mouse hepatocytes or cryopreserved human hepatocytes that recapitulates the hepatotoxic profile of antisense oligonucleotides previously observed in rodents and can be used for the prioritization of molecules prior to in vivo testing.

Simultaneous Assessment of Clearance, Metabolism, Induction, and Drug-Drug Interaction Potential Using a Long-Term In Vitro Liver Model for a Novel Hepatitis B Virus Inhibitor.

Long-term in vitro liver models are now widely explored for human hepatic metabolic clearance prediction, enzyme phenotyping, cross-species metabolism, comparison of low clearance drugs, and induction studies. Here, we present studies using a long-term liver model, which show how metabolism and active transport, drug-drug interactions, and enzyme induction in healthy and diseased states, such as hepatitis B virus (HBV) infection, may be assessed in a single test system to enable effective data integration for physiologically based pharmacokinetic (PBPK) modeling. The approach is exemplified in the case of (3S)-4-[[(4R)-4-(2-Chloro-4-fluorophenyl)-5-methoxycarbonyl-2-thiazol-2-yl-1,4-dihydropyrimidin-6-yl]methyl]morpholine-3-carboxylic acid RO6889678, a novel inhibitor of HBV with a complex absorption, distribution, metabolism, and excretion (ADME) profile. RO6889678 showed an intracellular enrichment of 78-fold in hepatocytes, with an apparent intrinsic clearance of 5.2 µl/min per mg protein and uptake and biliary clearances of 2.6 and 1.6 µl/min per mg protein, respectively. When apparent intrinsic clearance was incorporated into a PBPK model, the simulated oral human profiles were in good agreement with observed data at low doses but were underestimated at high doses due to unexpected overproportional increases in exposure with dose. In addition, the induction potential of RO6889678 on cytochrome P450 (P450) enzymes and transporters at steady state was assessed and cotreatment with ritonavir revealed a complex drug-drug interaction with concurrent P450 inhibition and moderate UDP-glucuronosyltransferase induction. Furthermore, we report on the first evaluation of in vitro pharmacokinetics studies using HBV-infected HepatoPac cocultures. Thus, long-term liver models have great potential as translational research tools exploring pharmacokinetics of novel drugs in vitro in health and disease.

Use of early phenotypic in vivo markers to assess human relevance of an unusual rodent non-genotoxic carcinogen in vitro.

Foci of altered hepatocytes (FAH) are considered putative, pre-neoplastic lesions that can occur spontaneously in aging rodents, but can also be induced by chemicals or drugs. Progression of FAH to hepatocellular neoplasms has been reported repeatedly but increases in foci in rodents do not necessarily lead to tumors in carcinogenicity studies and the relevance for humans often remains unclear. Here we present the case of RG3487, a molecule which induced FAH and, later on, tumors in rats. Because the molecule was negative in genotoxicity assays it was classified as a non-genotoxic carcinogen. In order to assess the potential for liver tumor formation in humans, we analyzed treatment-induced changes in vivo to establish a possible mode of action (MoA). In vivo and in vitro gene expression analysis revealed that nuclear receptor signaling was unlikely to be the relevant MoA and no other known mechanism could be established. We therefore took an approach comparing phenotypic markers, including mRNA changes, proliferation and glycogen accumulation, in vitro using cells of different species to assess the human relevance of this finding. Since the alterations observed in rats were not seen in the liver of mice or dogs in vivo, we could validate the relevance of the cell models chosen by use of hepatocytes from these species in vitro. This ultimately allowed for a cross-species comparison, which suggested that the formation of FAH and liver tumors was rat specific and unlikely to translate to human. Our work showed that phenotypic species comparison in vitro is a useful approach for assessment of the human relevance of pre-clinical findings where no known mechanism can be established.

Comparison of Liver Cell Models Using the Basel Phenotyping Cocktail.

Currently used hepatocyte cell systems for in vitro assessment of drug metabolism include hepatoma cell lines and primary human hepatocyte (PHH) cultures. We investigated the suitability of the validated in vivo Basel phenotyping cocktail (caffeine [CYP1A2], efavirenz [CYP2B6], losartan [CYP2C9], omeprazole [CYP2C19], metoprolol [CYP2D6], midazolam [CYP3A4]) in vitro and characterized four hepatocyte cell systems (HepG2 cells, HepaRG cells, and primary cryopreserved human hepatocytes in 2-dimensional [2D] culture or in 3D-spheroid co-culture) regarding basal metabolism and CYP inducibility. Under non-induced conditions, all CYP activities could be determined in 3D-PHH, CYP2B6, CYP2C19, CYP2D6, and CYP3A4 in 2D-PHH and HepaRG, and CYP2C19 and CYP3A4 in HepG2 cells. The highest non-induced CYP activities were observed in 3D-PHH and HepaRG cells. mRNA expression was at least four-fold higher for all CYPs in 3D-PHH compared to the other cell systems. After treatment with 20 µM rifampicin, mRNA increased 3- to 50-fold for all CYPs except CYP1A2 and 2D6 for HepaRG and 3D-PHH, 4-fold (CYP2B6) and 17-fold (CYP3A4) for 2D-PHH and four-fold (CYP3A4) for HepG2. In 3D-PHH at least a two-fold increase in CYP activity was observed for all inducible CYP isoforms while CYP1A2 and CYP2C9 activity did not increase in 2D-PHH and HepaRG. CYP inducibility assessed in vivo using the same phenotyping probes was also best reflected by the 3D-PHH model. Our studies show that 3D-PHH and (with some limitations) HepaRG are suitable cell systems for assessing drug metabolism and CYP induction in vitro. HepG2 cells are less suited to assess CYP induction of the 2C and 3A family. The Basel phenotyping cocktail is suitable for the assessment of CYP activity and induction also in vitro.

Establishment of a Predictive In Vitro Assay for Assessment of the Hepatotoxic Potential of Oligonucleotide Drugs.

Single stranded oligonucleotides (SSO) represent a novel therapeutic modality that opens new space to address previously undruggable targets. In spite of their proven efficacy, the development of promising SSO drug candidates has been limited by reported cases of SSO-associated hepatotoxicity. The mechanisms of SSO induced liver toxicity are poorly understood, and up to now no preclinical in vitro model has been established that allows prediction of the hepatotoxicity risk of a given SSO. Therefore, preclinical assessment of hepatic liability currently relies on rodent studies that require large cohorts of animals and lengthy protocols. Here, we describe the establishment and validation of an in vitro assay using primary hepatocytes that recapitulates the hepatotoxic profile of SSOs previously observed in rodents. In vitro cytotoxicity upon unassisted delivery was measured as an increase in extracellular lactate dehydrogenase (LDH) levels and concomitant reduction in intracellular glutathione and ATP levels after 3 days of treatment. Furthermore, toxic, but not safe, SSOs led to an increase in miR-122 in cell culture supernatants after 2 days of exposure, revealing the potential use of miR122 as a selective translational biomarker for detection of SSO-induced hepatotoxicity. Overall, we have developed and validated for the first time a robust in vitro screening assay for SSO liver safety profiling which allows rapid prioritization of candidate molecules early on in development.

Biomarkers of Flutamide-Bioactivation and Oxidative Stress In Vitro and In Vivo.

The nonsteroidal androgen-receptor antagonist flutamide is associated with hepatic injury. Oxidative stress and reactive metabolite formation are considered contributing factors to liver toxicity. Here we have used flutamide as a model drug to study the generation of reactive drug metabolites that undergo redox cycling to induce oxidative stress (OS) in vitro and in vivo. Lipid peroxidation (LPO) markers, as well as genes regulated by the redox-sensitive Nrf2 pathway, have been identified as surrogates for the characterization of OS. These markers and metabolism biomarkers for drug bioactivation have been investigated to characterize drug-induced hepatic damage. Rat hepatocytes and in vivo studies showed that several LPO markers, namely the isoprostanes 15R-PD2, dihydro keto PE2, and iPF(2α)-VI, as well as hydroxynonenal mercapturic acid metabolites, had increased significantly by 24 hours after flutamide treatment from 4.9 to 15.3-fold in hepatocytes and from 2.6 to 31.0-fold in rat plasma. Induction of mRNA expression levels for Nrf2-regulated genes was evident as well, with heme oxygenase 1, glutathione-S-transferase π1 and NAD(P)H dehydrogenase showing a 3.6-, 4.1-, and 1.9-fold increase in hepatocytes and 5.6-, 7.5-, and 94.1-fold in rat liver. All effects were observed at drug concentrations that did not show overt liver toxicity. Addition of an in situ hydrogen peroxide-generating system to in vitro experiments demonstrated the formation of a reactive di-imine intermediate as the responsible metabolic pathway for the generation of OS. The dataset suggests that hepatic oxidative stress conditions can be mediated via metabolic activation and can be monitored with suitable biomarkers preceding the terminal damage.

Pathway reporter genes define molecular phenotypes of human cells.

BACKGROUND: The phenotype of a living cell is determined by its pattern of active signaling networks, giving rise to a "molecular phenotype" associated with differential gene expression. Digital amplicon based RNA quantification by sequencing is a useful technology for molecular phenotyping as a novel tool to characterize the state of biological systems. RESULTS: We show here that the activity of signaling networks can be assessed based on a set of established key regulators and expression targets rather than the entire transcriptome. We compiled a panel of 917 human pathway reporter genes, representing 154 human signaling and metabolic networks for integrated knowledge- and data-driven understanding of biological processes. The reporter genes are significantly enriched for regulators and effectors covering a wide range of biological processes, and faithfully capture gene-level and pathway-level changes. We apply the approach to iPSC derived cardiomyocytes and primary human hepatocytes to describe changes in molecular phenotype during development or drug response. The reporter genes deliver an accurate pathway-centric view of the biological system under study, and identify known and novel modulation of signaling networks consistent with literature or experimental data. CONCLUSIONS: A panel of 917 pathway reporter genes is sufficient to describe changes in the molecular phenotype defined by 154 signaling cascades in various human cell types. AmpliSeq-RNA based digital transcript imaging enables simultaneous monitoring of the entire pathway reporter gene panel in up to 150 samples. We propose molecular phenotyping as a useful approach to understand diseases and drug action at the network level.

Effect of GLP1R agonists taspoglutide and liraglutide on primary thyroid C-cells from rodent and man.

Glucagon-like peptide 1 (GLP1) analogs have been associated with an increased incidence of thyroid C-cell hyperplasia and tumors in rodents. This effect may be due to a GLP1 receptor (GLP1R)-dependent mechanism. As the expression of GLP1R is much lower in primates than in rodents, the described C-cell proliferative lesions may not be relevant to man. Here, we aimed to establish primary thyroid cell cultures of rat and human to evaluate the expression and function of GLP1R in C-cells. In our experiments, GLP1R expression was observed in primary rat C-cells (in situ hybridization) but was not detected in primary human C-cells (mRNA and protein levels). The functional response of the cultures to the stimulation with GLP1R agonists is an indirect measure of the presence of functional receptor. Liraglutide and taspoglutide elicited a modest increase in calcitonin release and in calcitonin expression in rat primary thyroid cultures. Contrarily, no functional response to GLP1R agonists was observed in human thyroid cultures, despite the presence of few calcitonin-positive C-cells. Thus, the lack of functional response of the human cultures adds to the weight of evidence indicating that healthy human C-cells have very low levels or completely lack GLP1R. In summary, our results support the hypothesis that the GLP1R agonist-induced C-cell responses in rodents may not be relevant to primates. In addition, the established cell culture method represents a useful tool to study the physiological and/or pathological roles of GLP1 and GLP1R agonists on normal, non-transformed primary C-cells from rats and man.

A long-term three dimensional liver co-culture system for improved prediction of clinically relevant drug-induced hepatotoxicity.

Drug-induced liver injury (DILI) is the major cause for liver failure and post-marketing drug withdrawals. Due to species-specific differences in hepatocellular function, animal experiments to assess potential liabilities of drug candidates can predict hepatotoxicity in humans only to a certain extent. In addition to animal experimentation, primary hepatocytes from rat or human are widely used for pre-clinical safety assessment. However, as many toxic responses in vivo are mediated by a complex interplay among different cell types and often require chronic drug exposures, the predictive performance of hepatocytes is very limited. Here, we established and characterized human and rat in vitro three-dimensional (3D) liver co-culture systems containing primary parenchymal and non-parenchymal hepatic cells. Our data demonstrate that cells cultured on a 3D scaffold have a preserved composition of hepatocytes, stellate, Kupffer and endothelial cells and maintain liver function for up to 3months, as measured by the production of albumin, fibrinogen, transferrin and urea. Additionally, 3D liver co-cultures maintain cytochrome P450 inducibility, form bile canaliculi-like structures and respond to inflammatory stimuli. Upon incubation with selected hepatotoxicants including drugs which have been shown to induce idiosyncratic toxicity, we demonstrated that this model better detected in vivo drug-induced toxicity, including species-specific drug effects, when compared to monolayer hepatocyte cultures. In conclusion, our results underline the importance of more complex and long lasting in vitro cell culture models that contain all liver cell types and allow repeated drug-treatments for detection of in vivo-relevant adverse drug effects.

Gene expression-based in vivo and in vitro prediction of liver toxicity allows compound selection at an early stage of drug development.

We have analyzed gene expression and histopathology of rat liver treated with a histamine-3 receptor inverse agonist under development for the treatment of obesity 24 h after a single acute administration. While histopathology did not identify a clear liver toxicity, analysis of gene changes strongly suggested the development of toxicity. This prediction was confirmed in a 2-week repeat-dose rat study where prominent liver pathology occurred, while gene changes that lead to the prediction persisted. A subset of these genes was analyzed in vitro in both rat and human hepatocytes to reveal the potential relevancy of the findings for the situation in humans. This comprehensive analysis of the development compound at the gene expression level allowed interpretation of findings of the follow-up compound in a frontloaded 24-h single-dose acute study that was initiated before regular 2-week repeat-dose studies started. The high similarity of the follow-up compound to the lead compound based on gene expression lead to the immediate termination of the development program for this compound series. Our data demonstrate the value of genomics-based early toxicity prediction in short-term in vivo studies for the characterization of compounds to allow prioritization and selection of suited candidates before compound-, animal-, and cost-intensive longer term studies are undertaken.

Gene expression analysis of the hepatotoxicant methapyrilene in primary rat hepatocytes: an interlaboratory study.

Genomics technologies are used in several disciplines, including toxicology. However, these technologies are relatively new, and their applications require further investigations. When investigators apply these technologies to in vitro experiments, two major issues need to be clarified: a) can in vitro toxicity studies, in combination with genomics analyses, be used to predict the toxicity of a compound; and b) are the generated toxicogenomics data reproducible between laboratories? These questions were addressed by an interlaboratory study with laboratories of four pharmaceutical companies. We evaluated gene expression patterns from cultured rat primary hepatocytes after a 24-hr incubation with methapyrilene (MP). Extensive data analysis showed that comparison of genomics data from different sources is complex because both experimental and statistical variability are important confounding factors. However, appropriate statistical tools allowed us to use gene expression profiles to distinguish high-dose-treated cells from vehicle-treated cells. Moreover, we correctly identified MP in an independently generated in vitro database, underlining that in vitro toxicogenomics could be a predictive tool for toxicity. From a mechanistic point of view, despite the observed site-to-site variability, there was good concordance regarding the affected biologic processes. Several subsets of regulated genes were obtained by analyzing the data sets with one method or using different statistical analysis methods. The identified genes are involved in cellular processes that are associated to the exposure of primary hepatocytes to MP. Whether they are specific for MP and are cause or consequence of the toxicity requires further investigations.

Assessment of hepatotoxic liabilities by transcript profiling.

Male Wistar rats were treated with various model compounds or the appropriate vehicle controls in order to create a reference database for toxicogenomics assessment of novel compounds. Hepatotoxic compounds in the database were either known hepatotoxicants or showed hepatotoxicity during preclinical testing. Histopathology and clinical chemistry data were used to anchor the transcript profiles to an established endpoint (steatosis, cholestasis, direct acting, peroxisomal proliferation or nontoxic/control). These reference data were analyzed using a supervised learning method (support vector machines, SVM) to generate classification rules. This predictive model was subsequently used to assess compounds with regard to a potential hepatotoxic liability. A steatotic and a non-hepatotoxic 5HT(6) receptor antagonist compound from the same series were successfully discriminated by this toxicogenomics model. Additionally, an example is shown where a hepatotoxic liability was correctly recognized in the absence of pathological findings. In vitro experiments and a dog study confirmed the correctness of the toxicogenomics alert. Another interesting observation was that transcript profiles indicate toxicologically relevant changes at an earlier timepoint than routinely used methods. Together, these results support the useful application of toxicogenomics in raising alerts for adverse effects and generating mechanistic hypotheses that can be followed up by confirmatory experiments.

Discriminating different classes of toxicants by transcript profiling.

Male rats were treated with various model compounds or the appropriate vehicle controls. Most substances were either well-known hepatotoxicants or showed hepatotoxicity during preclinical testing. The aim of the present study was to determine if biological samples from rats treated with various compounds can be classified based on gene expression profiles. In addition to gene expression analysis using microarrays, a complete serum chemistry profile and liver and kidney histopathology were performed. We analyzed hepatic gene expression profiles using a supervised learning method (support vector machines; SVMs) to generate classification rules and combined this with recursive feature elimination to improve classification performance and to identify a compact subset of probe sets with potential use as biomarkers. Two different SVM algorithms were tested, and the models obtained were validated with a compound-based external cross-validation approach. Our predictive models were able to discriminate between hepatotoxic and nonhepatotoxic compounds. Furthermore, they predicted the correct class of hepatotoxicant in most cases. We provide an example showing that a predictive model built on transcript profiles from one rat strain can successfully classify profiles from another rat strain. In addition, we demonstrate that the predictive models identify nonresponders and are able to discriminate between gene changes related to pharmacology and toxicity. This work confirms the hypothesis that compound classification based on gene expression data is feasible.

In vitro investigation on the impact of Solutol HS 15 on the uptake of colchicine into rat hepatocytes.

In the current investigation, the impact of the surface-active formulation ingredient Solutol HS 15 on the uptake of colchicine into freshly isolated rat hepatocytes was investigated using a centrifugal filtration technique through a silicone oil layer. Colchicine is taken up into the cells by an active transport mechanism. When conducting the experiment at 37 degrees C, it was found that at concentrations below its critical micellar concentration (CMC) of 0.021% (0.0003 and 0.003%, w/v), Solutol HS 15 did not impact the uptake of colchicine. By contrast, at a Solutol HS 15 concentration above its CMC (0.03%, w/v), the amount of colchicine taken up into the cells as well as its uptake velocity were significantly decreased. However, in control experiments performed at 4 degrees C, a temperature at which active transport processes should be significantly slowed down, Solutol HS 15 at 0.03% did not affect colchicine uptake and/or its association with the cells. The described findings might be rationalized by inhibition of colchicine transport either due to direct interaction at the transport site or due to alterations of membrane properties in the presence of Solutol HS 15 at concentrations above its CMC. Moreover, a strong molecular interaction between Solutol HS 15 and colchicine as well as an incorporation of colchicine into micelles formed by Solutol HS 15, this way resulting in a limited contact of colchicine with the cells, cannot be excluded as contributors to the observed effect.

In vitro investigation on the impact of the surface-active excipients Cremophor EL, Tween 80 and Solutol HS 15 on the metabolism of midazolam.

The impact of the surface-active formulation ingredients Cremophor EL, Tween 80 and Solutol HS 15 on the intrinsic clearance (Clint) of midazolam (MDZ) was investigated in rat hepatocytes and microsomes. In rat hepatocytes with 0.003%, 0.03% and 0.3% (w/v) Solutol HS 15 already present in the incubation medium, the Clint was significantly reduced in a dose-dependent manner by about 25%, 30% and 50%, respectively. In the presence of Cremophor EL and Tween 80 a significant reduction in Clint by about 30% and 25%, respectively, was observed at 0.03% surfactant concentration. At 0.3% of Cremophor EL and Tween 80, Clint was reduced by about 50% and 20%, respectively. A reduction in Clint was also observed in experiments with rat liver microsomes. At surfactant concentrations up to 0.03%, cytotoxicity assays (lactate dehydrogenase release, adenosine triphosphate content) as well as light microscope investigations did not reveal any cytotoxic impact of the surfactants on the hepatocyte monolayer. A potential interaction of the surfactants with biological membranes was determined using phosphatidylcholine-cholesterol liposomes loaded with self-quenching concentrations of carboxyfluorescein. No marked release of carboxyfluorescein from the liposomes (that would be an indication for a surfactant-dependent disruption of membrane integrity) was observed up to concentrations of 0.03% of the different surfactants. It is concluded that cytochrome P450 3A mediated metabolism of MDZ seems to be prevented by all surfactants at concentrations above 0.03%. In our experiments the surfactants did not show toxic effects at concentrations that resulted in a decreased Clint of MDZ. Thus, a direct inhibition of the metabolizing enzymes, a molecular interaction with the microsomes as well as an alteration of membrane properties that did not yet result in a release of LDH have to be taken into consideration as reasons for the observed changes in the metabolism of MDZ.

Gene expression in two hepatic cell lines, cultured primary hepatocytes, and liver slices compared to the in vivo liver gene expression in rats: possible implications for toxicogenomics use of in vitro systems.

Microarray technology allows the simultaneous analysis of mRNA expression levels of thousands of genes. In the field of toxicogenomics, this technology could help to identify potentially unsafe compounds based on the changes in mRNA expression patterns they induce. Rodent in vivo and in vitro systems are currently the experimental models of choice for predictive toxicology, especially in early phases of development. This study characterizes several hepatic in vitro systems based on mRNA expression profiles, comparing them to gene expression in liver tissue. The in vitro systems investigated comprise two rat liver cell lines (BRL3A and NRL clone 9), primary hepatocytes in conventional monolayer or in sandwich culture, and liver slices. The results demonstrate that liver slices exhibit the strongest similarity to liver tissue regarding mRNA expression, whereas the two cell lines are quite different from the whole liver. We were able to identify genes with strong changes in expression levels in all or at least one of the in vitro systems relative to whole liver. In particular, for some cytochrome P450s the differences observed on the mRNA expression level were paralleled by protein expression and enzymatic activity. In addition, the effect of time in culture was assessed. We were able to show a profound effect of the duration of culture. Expression patterns change most rapidly soon after cell isolation and culture initiation and stabilize with time in culture. The findings are discussed with respect to the usefulness of the various hepatic in vitro systems for microarray-based toxicological testing of compounds.