Species differences in small intestinal exposure-related epithelial vacuolation in rats and dogs treated with a heteroaryldihydropyrimidine molecule.

Small intestinal epithelial vacuolation induced by a heteroaryldihydropyrimidine compound (HAP-1) was observed in rats but not in dogs at termination in screening toxicity studies, despite the plasma exposure being higher in dogs. To understand the species differences, investigational studies with multiple time points following single dose (SD) and 7-day repeated dose (RD) were conducted in both species at doses resulting in comparable plasma exposures. In rats, epithelial vacuolation in the duodenum and jejunum were observed at all time points. In dogs, transient vacuolation was noted at 8 h post-SD (SD_8h) and 4 h post-RD (RD_4 h), but not at termination (RD_24 h). Special stains demonstrated lipid accumulation within enterocytes in both species and intracytoplasmic inclusion bodies in rats. Transmission electron microscopy identified these inclusion bodies as endoplasmic reticulum (ER) membranous structures. Transcriptomic analysis on jejunal mucosa at SD_8 h and RD_24 h revealed perturbations of lipid metabolism-related genes at SD_8 h in both species, but not at RD_24 h in dogs. ER stress-related gene changes at both time points were observed in rats only. Despite comparable HAP-1 plasma exposures, the duodenum and jejunum tissue concentrations of HAP-1 and acyl glucuronide metabolite were >5- and >30-fold higher in rats than in dogs, respectively. In vitro, similar cytotoxicity was observed in rat and dog duodenal organoids treated with HAP-1. In conclusion, HAP-1-induced intestinal epithelial vacuolation was related to lipid metabolism dysregulation in both species and ER-related injuries in rats only. The species differences were likely related to the difference in intestinal exposure to HAP-1 and its reactive metabolite.

European Society of Toxicologic Pathology (Pathology 2.0 Molecular Pathology Special Interest Group): Review of In Situ Hybridization Techniques for Drug Research and Development.

In situ hybridization (ISH) is used for the localization of specific nucleic acid sequences in cells or tissues by complementary binding of a nucleotide probe to a specific target nucleic acid sequence. In the last years, the specificity and sensitivity of ISH assays were improved by innovative techniques like synthetic nucleic acids and tandem oligonucleotide probes combined with signal amplification methods like branched DNA, hybridization chain reaction and tyramide signal amplification. These improvements increased the application spectrum for ISH on formalin-fixed paraffin-embedded tissues. ISH is a powerful tool to investigate DNA, mRNA transcripts, regulatory noncoding RNA, and therapeutic oligonucleotides. ISH can be used to obtain spatial information of a cell type, subcellular localization, or expression levels of targets. Since immunohistochemistry and ISH share similar workflows, their combination can address simultaneous transcriptomics and proteomics questions. The goal of this review paper is to revisit the current state of the scientific approaches in ISH and its application in drug research and development.

Toxicologic Pathology Forum Opinion Piece: Use of Virtual Control Groups in Nonclinical Toxicity Studies: The Anatomic Pathology Perspective.

In the last decade, numerous initiatives have emerged worldwide to reduce the use of animals in drug development, including more recently the introduction of Virtual Control Groups (VCGs) concept for nonclinical toxicity studies. Although replacement of concurrent controls (CCs) by virtual controls (VCs) represents an exciting opportunity, there are associated challenges that will be discussed in this paper with a more specific focus on anatomic pathology. Coordinated efforts will be needed from toxicologists, clinical and anatomic pathologists, and regulators to support approaches that will facilitate a staggered implementation of VCGs in nonclinical toxicity studies. Notably, the authors believe that a validated database for VC animals will need to include histopathology (digital) slides for microscopic assessment. Ultimately, the most important step lies in the validation of the concept by performing VCG and the full control group in parallel for studies of varying duration over a reasonable timespan to confirm there are no differences in outcomes (dual study design). The authors also discuss a hybrid approach, whereby control groups comprised both concurrent and VCs to demonstrate proof-of-concept. Once confidence is established by sponsors and regulators, VCs have the potential to replace some or all CC animals.

Publication Categories in Toxicologic Pathology.

Toxicologic Pathology is the official journal of the Society of Toxicologic Pathology (STP), the British Society of Toxicological Pathology, and the European STP (ESTP). Toxicologic Pathology publishes articles related to topics in various aspects of toxicologic pathology such as anatomic pathology, clinical pathology, experimental pathology, and biomarker research. Publications include society-endorsed Best Practice/Position and Points to Consider publications and ESTP Expert Workshop articles that are relevant to toxicologic pathology and scientific regulatory processes, Opinion articles under the banner of the STP Toxicologic Pathology Forum, Original Articles, Review Articles (unsolicited/contributed, mini, and invited), Brief Communications, Letters to the Editor, Meeting Reports, and Book Reviews. This article provides details on the various publication categories in Toxicologic Pathology and will serve as a reference for authors and readers.

Artificial Intelligence in Toxicologic Pathology: Quantitative Evaluation of Compound-Induced Hepatocellular Hypertrophy in Rats.

Digital pathology evolved rapidly, enabling more systematic usage of image analysis and development of artificial intelligence (AI) applications. Here, combined AI models were developed to evaluate hepatocellular hypertrophy in rat liver, using commercial AI-based software on hematoxylin and eosin-stained whole slide images. In a first approach, deep learning-based identification of critical tissue zones (centrilobular, midzonal, and periportal) enabled evaluation of region-specific cell size. Mean cytoplasmic area of hepatocytes was calculated via several sequential algorithms including segmentation in microanatomical structures (separation of sinusoids and vessels from hepatocytes), nuclear detection, and area measurements. An increase in mean cytoplasmic area could be shown in groups given phenobarbital, known to induce hepatocellular hypertrophy when compared to control groups, in multiple studies. Quantitative results correlated with the gold standard: observation and grading performed by board-certified veterinary pathologists, liver weights, and gene expression. Furthermore, as a second approach, we introduce for the first time deep learning-based direct detection of hepatocellular hypertrophy with similar results. Cell hypertrophy is challenging to pick up, particularly in milder cases. Additional evaluation of mean cytoplasmic area or direct detection of hypertrophy, combined with histopathological observations and liver weights, is expected to increase accuracy and repeatability of diagnoses and grading by pathologists.

Characterization of BAY 1905254, an Immune Checkpoint Inhibitor Targeting the Immunoglobulin-Like Domain Containing Receptor 2 (ILDR2).

The immunoglobulin-like domain containing receptor 2 (ILDR2), a type I transmembrane protein belonging to the B7 family of immunomodulatory receptors, has been described to induce an immunosuppressive effect on T-cell responses. Besides its expression in several nonlymphoid tissue types, we found that ILDR2 was also expressed in fibroblastic reticular cells (FRC) in the stromal part of the lymph node. These immunoregulatory cells were located in the T-cell zone and were essential for the recruitment of naïve T cells and activated dendritic cells to the lymph nodes. Previously, it has been shown that an ILDR2-Fc fusion protein exhibits immunomodulatory effects in several models of autoimmune diseases, such as multiple sclerosis, rheumatoid arthritis, and type I diabetes. Herein, we report the generation and characterization of a human/mouse/monkey cross-reactive anti-ILDR2 hIgG2 antibody, BAY 1905254, developed to block the immunosuppressive activity of ILDR2 for cancer immunotherapy. BAY 1905254 was shown to promote T-cell activation in vitro and enhance antigen-specific T-cell proliferation and cytotoxicity in vivo in mice. BAY 1905254 also showed potent efficacy in various syngeneic mouse cancer models, and the efficacy was found to correlate with increasing mutational load in the cancer models used. Additive or even synergistic antitumor effects were observed when BAY 1905254 was administered in combination with anti-PD-L1, an immunogenic cell death-inducing chemotherapeutic, or with tumor antigen immunization. Taken together, our data showed that BAY 1905254 is a potential drug candidate for cancer immunotherapy, supporting its further evaluation.

Nonclinical Safety Assessment of a Long-Acting Recombinant PEGylated Factor Eight (BAY 94-9027) With a 60 kDa PEG.

BAY 94-9027 (Jivi) is a site-specifically PEGylated human B-domain-deleted (BDD) recombinant factor VIII (FVIII), with a 60 kDa branched PEG molecule attached. The nonclinical safety of BAY 94-9027 was evaluated in a toxicology program that included 2 weeks intravenous (IV) toxicity studies in rats and rabbits, a juvenile toxicity study in rats as well as a 26-week chronic study in rats. Doses of 75, 750, or 2250 IU/kg given every other day for 2 weeks did not elicit any findings related to BAY 94-9027. Specifically, no thrombus formation or histological changes such as cellular vacuolation were seen. In the chronic toxicity study, 40, 400, and 1200 IU/kg of BAY 94-9027 given twice weekly did not induce adverse effects related to BAY 94-9027, and no tissue vacuolation was observed. There was no PEG detected in choroid plexus or other areas of the brain, cerebrospinal fluid or in spleen or kidneys. These results were supported by toxicity studies in rats and rabbits treated with PEG 60 kDa attached to the maleimide linker (PEG-60-Mal-Cys). No findings related to PEG-60-Mal-Cys were seen. These results demonstrate the safety of BAY 94-9027 for long-term use.

Histology and Gadolinium Distribution in the Rodent Brain After the Administration of Cumulative High Doses of Linear and Macrocyclic Gadolinium-Based Contrast Agents.

OBJECTIVES: Retrospective studies in patients with primary brain tumors or other central nervous system pathologies as well as postmortem studies have suggested that gadolinium (Gd) deposition occurs in the dentate nucleus (DN) and globus pallidus (GP) after multiple administrations of primarily linear Gd-based contrast agents (GBCAs). However, this deposition has not been associated with any adverse effects or histopathological alterations. The aim of this preclinical study was to systematically examine differences between linear and macrocyclic GBCAs in their potential to induce changes in brain and skin histology including Gd distribution in high spatial resolution. MATERIALS AND METHODS: Fifty male Wistar-Han rats were randomly allocated into control (saline, n = 10 rats) and 4 GBCA groups (linear GBCAs: gadodiamide and gadopentetate dimeglumine, macrocyclic GBCAs: gadobutrol and gadoteridol; n = 10 rats per group). The animals received 20 daily intravenous injections at a dose of 2.5 mmol Gd/kg body weight. Eight weeks after the last GBCA administration, the animals were killed, and the brain and skin samples were histopathologically assessed (hematoxylin and eosin; cresyl violet [Nissl]) and by immunohistochemistry. The Gd concentration in the skin, bone, brain, and skeletal muscle samples were analyzed using inductively coupled plasma mass spectroscopy (ICP-MS, n = 4). The spatial Gd distribution in the brain and skin samples was analyzed in cryosections using laser ablation coupled with ICP-MS (LA-ICP-MS, n = 3). For the ultra-high resolution of Gd distribution, brain sections of rats injected with gadodiamide or saline (n = 1) were assessed by scanning electron microscopy coupled to energy dispersive x-ray spectroscopy and transmission electron microscopy, respectively. RESULTS: No histological changes were observed in the brain. In contrast, 4 of 10 animals in the gadodiamide group but none of the animals in other groups showed macroscopic and histological nephrogenic systemic fibrosis-like skin lesions. The Gd concentrations observed in the skin/brain samples (in nanomole Gd per gram of tissue) for each agent were as follows: gadodiamide: 1472 ± 115/11.1 ± 5.1, gadopentetate dimeglumine: 80.8 ± 6.2/13.1 ± 7.3, gadobutrol: 1.1 ± 0.5/0.7 ± 0.4, and gadoteridol: 1.7 ± 0.8/0.5 ± 0.2. The average detected residual Gd concentration in the brain was approximately 15-fold higher for linear than for macrocyclic GBCAs. The highest amounts of Gd found in brain corresponded to less than 0.0002% of the injected dose per gram of tissue. Using LA-ICP-MS, high Gd concentrations in the deep cerebellar nuclei and in the granular layer of the cerebellar cortex were detected only for linear gadodiamide and gadopentetate dimeglumine but not for gadoteridol or gadobutrol. The energy dispersive x-ray spectroscopy analysis revealed Gd-containing spots in the skin of animals administered gadodiamide and gadopentetate dimeglumine. Transmission electron microscopy revealed several Gd-containing spots in the region of the dentate nuclei in the brain of 1 animal injected with gadodiamide. CONCLUSIONS: After repeated high dosing, nephrogenic systemic fibrosis-like macroscopic and histopathological lesions of the skin were observed only in some of the gadodiamide-treated animals. No histopathological findings were detected in the rodent brain. The administration of linear GBCAs was associated with significantly higher Gd concentrations in the brain and skin compared with macrocyclic GBCA administration. The results of LA-ICP-MS demonstrated local accumulation of Gd within the deep cerebellar nuclei and the granular layer only after the administration of linear agents. In summary, the detected low Gd concentrations in the skin and brain were well correlated with the higher kinetic stability of macrocyclic GBCA.

Novel Mps1 Kinase Inhibitors with Potent Antitumor Activity.

Monopolar spindle 1 (Mps1) has been shown to function as the key kinase that activates the spindle assembly checkpoint (SAC) to secure proper distribution of chromosomes to daughter cells. Here, we report the structure and functional characterization of two novel selective Mps1 inhibitors, BAY 1161909 and BAY 1217389, derived from structurally distinct chemical classes. BAY 1161909 and BAY 1217389 inhibited Mps1 kinase activity with IC50 values below 10 nmol/L while showing an excellent selectivity profile. In cellular mechanistic assays, both Mps1 inhibitors abrogated nocodazole-induced SAC activity and induced premature exit from mitosis ("mitotic breakthrough"), resulting in multinuclearity and tumor cell death. Both compounds efficiently inhibited tumor cell proliferation in vitro (IC50 nmol/L range). In vivo, BAY 1161909 and BAY 1217389 achieved moderate efficacy in monotherapy in tumor xenograft studies. However, in line with its unique mode of action, when combined with paclitaxel, low doses of Mps1 inhibitor reduced paclitaxel-induced mitotic arrest by the weakening of SAC activity. As a result, combination therapy strongly improved efficacy over paclitaxel or Mps1 inhibitor monotreatment at the respective MTDs in a broad range of xenograft models, including those showing acquired or intrinsic paclitaxel resistance. Both Mps1 inhibitors showed good tolerability without adding toxicity to paclitaxel monotherapy. These preclinical findings validate the innovative concept of SAC abrogation for cancer therapy and justify clinical proof-of-concept studies evaluating the Mps1 inhibitors BAY 1161909 and BAY 1217389 in combination with antimitotic cancer drugs to enhance their efficacy and potentially overcome resistance. Mol Cancer Ther; 15(4); 583-92. ©2016 AACR.

Glomerulonephritis-induced changes in kidney gene expression in rats.

We investigated a glomerulonephritis (GN) model in rats induced by nephrotoxic serum (NTS) which contains antibodies against the glomerular basement membrane (GBM). The anti-GBM GN model in rats is widely used since its biochemical and histopathological characteristics are similar to crescentic nephritis and Goodpasture's disease in humans (Pusey, 2003[2]). Male Wistar Kyoto (WKY) and Sprague-Dawley (SD) rats were dosed once with 1, 2.5 and 5 ml/kg nephrotoxic serum (NTS) or 1.5 and 5 ml/kg NTS, respectively. GN and tubular damage were observed histopathologically in all treated rats after 14 days. To obtain insight into molecular processes during GN pathogenesis, mRNA expression was investigated in WKY and SD kidneys using Affymetrix's GeneChip Rat genome 230_2.0 arrays (GSE64265). The immunopathological processes during GN are still not fully understood and likely involve both innate and adaptive immunity. In the present study, several hundred mRNAs were found deregulated, which functionally were mostly associated with inflammation and regeneration. The ß-chain of the major histocompatibility complex class II RT1.B (Rt1-Bb) and complement component 6 (C6) were identified as two mRNAs differentially expressed between WKY and SD rat strains which could be related to known different susceptibilities to NTS of different rat strains; both were increased in WKY and decreased in SD rats (Pavkovic et al., 2015 [1]). Increased Rt1-Bb expression in WKY rats could indicate a stronger and more persistent cellular reaction of the adaptive immune system in this strain, in line with findings indicating adaptive immune reactions during GN. The complement cascade is also known to be essential for GN development, especially terminal cascade products like C6.

Application of a Patient Derived Xenograft Model for Predicative Study of Uterine Fibroid Disease.

Human uterine fibroids, benign tumors derived from the smooth muscle layers of the uterus, impose a major health burden to up to 50% of premenopausal women in their daily life. To improve our understanding of this disease, we developed and characterized a patient-derived xenograft model by subcutaneous transplantation of pieces of human uterine fibroid tissue into three different strains of severe combined immunodeficient mice. Engrafted uterine fibroid tissue preserved the classical morphology with interwoven bundles of smooth muscle cells and an abundant deposition of collagenous matrix, similar to uterine fibroids in situ. The grafts expressed both estrogen receptor 1 and progesterone receptor. Additionally, both receptors were up-regulated by estrogen treatment. Growth of the fibroid grafts was dependent on 17ß-estradiol and progesterone supplementation at levels similar to women with the disease and was studied for up to 60 days at maximum. Co-treatment with the antiprogestin mifepristone reduced graft growth (four independent donors, p<0.0001 two-sided t-test), as did treatment with the mTOR inhibitor rapamycin (three independent donors, p<0.0001 two-sided t-test). This in vivo animal model preserves the main histological and functional characteristics of human uterine fibroids, is amenable to intervention by pharmacological treatment, and can thus serve as an adequate model for the development of novel therapies.

Glomerulonephritis-Induced Changes in Urinary and Kidney MicroRNA Profiles in Rats.

MicroRNAs (miRNAs) regulate gene expression post-transcriptionally and thus are involved in various physiological and pathological states. Due to their stability in biofluids miRNAs have also been proposed as biomarkers (BMs) for tissue injury. We investigated the usefulness of urinary miRNAs for detection of site-specific renal damage in an antiglomerular basement membrane glomerulonephritis (GN) model in rats by comparing GN-induced urinary miRNAs profiles to traditional and newer protein BMs, and to proximal tubular injury-induced urinary miRNA profiles observed previously after cisplatin (Cp) treatment. Male Wistar Kyoto and Sprague Dawley rats were dosed once with 1, 2.5, and 5 ml/kg nephrotoxic serum (NTS) or 1.5 and 5 ml/kg NTS, respectively. GN and tubular damage were observed histopathologically in all treated rats after 14 days. Although serum creatinine and BUN were not changed, several protein BMs and 74 urinary miRNAs were found to be increased 8 and 14 days after NTS administration. Of these 74 miRNAs, 5 were identified as increased after NTS but not after Cp treatment. Using in situ hybridization two of them, miR-10 b and -100, were found to be localized in distal segments of the nephron, potentially reflecting the tubular injury in those regions. Furthermore, evaluation of both miRNA and mRNA expression in the kidney revealed possible miRNA-mRNA interactions mostly associated with fibrotic and transforming growth factor ß signaling. In conclusion, our investigations support the potential of urinary miRNAs as specific BMs for kidney injury, and suggest a role of miRNAs in pathological processes during GN in the kidney.

Comparison of progestin transcriptional profiles in rat mammary gland using Laser Capture Microdissection and whole tissue-sampling.

INTRODUCTION: Investigation of molecular mechanisms by gene expression profiling gains increasingly importance in preclinical safety evaluation. However, assigning expressed genes to specific cell populations is nearly impossible if the investigated RNA originates from whole tissue extracts. In this regard, Laser Capture Microdissection (LCM) can be used to detect changes specific to individual cell types. The objective of this study was to investigate the use of LCM for characterisation of progestin-related gene expression changes in the mammary gland. Thus, transcriptional profiles of the mammary gland of rats treated with a non-steroidal progesterone-receptor ligand, promegestone, medroxyprogesterone acetate, progesterone or vehicle were compared using whole tissue homogenates or LCM-captured epithelial cells. METHODS: Total RNA from 30 mammary glands was isolated from snap-frozen specimen of the whole tissue and from approximately 25.000-30.000 cells of cresyl violet stained frozen sections employing LCM. After amplification of averaged 0.2µg total RNA of LCM-captured samples, RNA was labelled, hybridised to Affymetrix GeneChips and analysed. RESULTS: LCM-captured samples showed up to 3-fold more differentially expressed probe sets (progesterone) and up to 10-fold more downregulated (promegestone) probe sets than whole tissue samples implying high cell specificity. Moreover, mammary gland specific differentiation markers like whey acidic protein, alpha lactalbumin, casein alpha s1 and casein kappa showed up to 3.4-fold (alpha lactalbumin, vehicle) higher expression values. Multivariate data analyses revealed a clear separation of gene expression profiles according to the method used, suggesting an amplification dependent bias. DISCUSSION: LCM transcriptional profiling provides highly cell-specific information. An amplification dependent bias was observed. The technical variability was shown to be smaller than the biological variability. For progestin-related transcriptional profiling of the mammary gland, whole tissue-sampling proved to yield more informative results. Therefore LCM should only be considered when cell-type specific gene expression profiles are necessary for an in depth evaluation.

The effect of iodinated contrast agent properties on renal kinetics and oxygenation.

OBJECTIVE: We analyzed renal kinetics and renal oxygenation in rats after administration of several classes and formulations of contrast agents (CAs) with a focus on the influence of osmolality and substance-specific properties. MATERIALS AND METHODS: We investigated the renal kinetics of a nonionic, dimeric CA (iodixanol) formulated in 3 different osmolalities (hypo-osmolar, iso-osmolar, low-osmolar) and compared it to nonionic, low-osmolar (iopromide), and ionic, low-osmolar CAs (ioxaglate) using computed tomography for a period of 24 hours. The CAs were administered intravenously at a dosage of 4 g iodine/kg body weight. The average exposure was calculated, and urine viscosities were compared before the injection and during the time intervals of 0 to 60 minutes and 60 to 120 minutes after the injection. Renal oxygenation levels of the renal cortex and medulla were estimated using blood-oxygen-level-dependent magnetic resonance imaging. We used histologic methods to systematically analyze the gravity of vacuole formation based on the physicochemical and substance-specific properties of each CA. RESULTS: Iso-osmolar and hypo-osmolar iodixanol and, to a lesser extent, iodixanol/mannitol accumulated rapidly in the kidneys during the first 5 minutes of the injection and remained higher 2, 4, 6, and 24 hours after the injection compared with iopromide and ioxaglate, which showed fast iodine excretion. Similarly, lower renal blood oxygen levels were estimated for all iodixanol formulations as compared with ioxaglate and iopromide. The incidence of vacuole formation was high for all iodixanol formulations and for ioxaglate (6 of 6 rats) and low for iopromide (1 of 6 rats). Moderate severity of vacuoles was determined for the iodixanol solutions; minimal severity, for ioxaglate and iopromide. CONCLUSIONS: We identified a superior profile for the low-osmolar CAs compared with the iso-osmolar CAs regarding rapid excretion, short-term renal exposure, and renal oxygenation.

D-18F-fluoromethyl tyrosine imaging of bone metastases in a mouse model.

UNLABELLED: The presence and localization of metastatic bone lesions is important for the staging of the disease and subsequent treatment decisions. Detecting tumor cells would have additional value over the current indirect bone scintigraphy method for detecting areas of elevated skeletal metabolic activity. d-(18)F-fluoromethyl tyrosine (d-(18)F-FMT) has recently shown good uptake and fast elimination, resulting in good tumor-to-background ratios. The potential of d-(18)F-FMT for imaging bone metastases has been investigated. METHODS: 786-O/luciferase human renal adenocarcinoma cells were injected intracardially, resulting in the formation of bone metastases in mice. Small-animal PET was performed 51 and 65 d after tumor cell inoculation. RESULTS: d-(18)F-FMT showed specific uptake in the bone metastases, giving excellent images with a little background in the pancreas. All imaged metastases were histologically confirmed. A bone scan with (18)F-fluoride showed elevated skeletal metabolic activity in the areas of osteolytic lesions. CONCLUSION: d-(18)F-FMT is a useful PET tracer for the detection of bone metastases and should be evaluated in the clinical setting.

Molecular profiling of hepatotoxicity induced by a aminoguanidine carboxylate in the rat: gene expression profiling.

The hepatotoxicity of the aminoguanidine carboxylate 2-[1-[hydrazino(imino)methyl]hydrazino]acetic acid was characterized using oligonucleotide micro arrays, with the goal to select compounds from the same class with lower toxicity potential. The approach included a 14-day repeated- and a single-dose study in the rat as well as in vitro studies. Common gene expression changes could be followed from in vivo to in vitro studies. Anyhow, comparing the in vivo and in vitro response of the compound on gene expression, significant discrepancies were detected. Many of the genes whose mRNA levels were increased/decreased in the livers of the animals treated with toxic doses of the compound, were expressed at higher/lower levels in control hepatocytes than in control liver. The expression of the majority of these genes was not affected by in vitro treatment. These data question the use of gene expression analysis as a marker for drug response in vitro and illustrate the need of a careful characterization of in vitro systems. The results presented show that array-based gene expression analysis can lead to a better understanding of the molecular basis of drug-induced liver injury and, potentially, be used in the selection process for compounds and in the design of safer drugs.