Acute Hepatotoxicity: A Predictive Model Based on Focused Illumina Microarrays

Pharmacodynamic and safety considerations for influenza vaccine and adjuvant design

INTRODUCTION

A novel adjuvant evaluation system for safety and immunogenicity is needed. Vaccination is important for infection prevention, for example, from influenza viruses. Adjuvants are considered critical for improving the effectiveness of influenza vaccines. Adjuvant development is an important issue in influenza vaccine design.

AREAS COVERED

A conventional in vivo evaluation method for vaccine safety has been limited in analyzing phenotypic and pathological changes. Therefore, it is difficult to obtain information on the changes at the molecular level. This review aims to explain the recently developed genomics analysis-based vaccine adjuvant safety evaluation tools verified by AddaVax^TM and polyinosinic-polycytidylic acid (poly I:C) using 18 biomarker genes and whole-virion inactivated influenza vaccine as a toxicity control. Genomics analyzes would help provide safety and efficacy information regarding influenza vaccine design by facilitating appropriate adjuvant selection.

EXPERT OPINION

The efficacy and safety profiles of influenza vaccines and adjuvants using genomics technologies provide useful information regarding immunogenicity, which is related to safety and efficacy. This approach provides important information to select appropriate inoculation routes, combinations of vaccine antigens and adjuvants, and dosing amounts. The efficacy of vaccine adjuvant evaluation by genomics analysis should be verified by various studies using various vaccines in the future.

Repeated Administration of Clinical Doses of Tramadol and Tapentadol Causes Hepato- and Nephrotoxic Effects in Wistar Rats

Tramadol and tapentadol are fully synthetic and extensively used analgesic opioids, presenting enhanced therapeutic and safety profiles as compared with their peers. However, reports of adverse reactions, intoxications and fatalities have been increasing. Information regarding the molecular, biochemical, and histological alterations underlying their toxicological potential is missing, particularly for tapentadol, owing to its more recent market authorization. Considering the paramount importance of liver and kidney for the metabolism and excretion of both opioids, these organs are especially susceptible to toxicological damage. In the present study, we aimed to characterize the putative hepatic and renal deleterious effects of repeated exposure to therapeutic doses of tramadol and tapentadol, using an in vivo animal model. Male Wistar rats were randomly divided into six experimental groups, composed of six animals each, which received daily single intraperitoneal injections of 10, 25 or 50 mg/kg tramadol or tapentadol (a low, standard analgesic dose, an intermediate dose and the maximum recommended daily dose, respectively). An additional control group was injected with normal saline. Following 14 consecutive days of administration, serum, urine and liver and kidney tissue samples were processed for biochemical, metabolic and histological analysis. Repeated administration of therapeutic doses of both opioids led to: (i) increased lipid and protein oxidation in liver and kidney, as well as to decreased total liver antioxidant capacity; (ii) decreased serum albumin, urea, butyrylcholinesterase and complement C3 and C4 levels, denoting liver synthesis impairment; (iii) elevated serum activity of liver enzymes, such as alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase and γ-glutamyl transpeptidase, as well as lipid profile alterations, also reflecting hepatobiliary commitment; (iv) derangement of iron metabolism, as shown through increases in serum iron, ferritin, haptoglobin and heme oxygenase-1 levels. In turn, elevated serum cystatin C, decreased urine creatinine output and increased urine microalbumin levels were detected upon exposure to tapentadol only, while increased serum amylase and urine N-acetyl-β-D-glucosaminidase activities were observed for both opioids. Collectively, these results are compatible with kidney injury. Changes were also found in the expression levels of liver- and kidney-specific toxicity biomarker genes, upon exposure to tramadol and tapentadol, correlating well with alterations in lipid profile, iron metabolism and glomerular and tubular function. Histopathological analysis evidenced sinusoidal dilatation, microsteatosis, mononuclear cell infiltrates, glomerular and tubular disorganization, and increased Bowman's spaces. Although some findings are more pronounced upon tapentadol exposure, our study shows that, when compared with acute exposure, prolonged administration of both opioids smooths the differences between their toxicological effects, and that these occur at lower doses within the therapeutic range.

Using human in vitro transcriptome analysis to build trustworthy machine learning models for prediction of animal drug toxicity

During the development of new drugs or compounds there is a requirement for preclinical trials, commonly involving animal tests, to ascertain the safety of the compound prior to human trials. Machine learning techniques could provide an in-silico alternative to animal models for assessing drug toxicity, thus reducing expensive and invasive animal testing during clinical trials, for drugs that are most likely to fail safety tests. Here we present a machine learning model to predict kidney dysfunction, as a proxy for drug induced renal toxicity, in rats. To achieve this, we use inexpensive transcriptomic profiles derived from human cell lines after chemical compound treatment to train our models combined with compound chemical structure information. Genomics data due to its sparse, high-dimensional and noisy nature presents significant challenges in building trustworthy and transparent machine learning models. Here we address these issues by judiciously building feature sets from heterogenous sources and coupling them with measures of model uncertainty achieved through Gaussian Process based Bayesian models. We combine the use of insight into the feature-wise contributions to our predictions with the use of predictive uncertainties recovered from the Gaussian Process to improve the transparency and trustworthiness of the model.

Inflammation-associated suppression of metabolic gene networks in acute and chronic liver disease

Inflammation has been recognized as essential for restorative regeneration. Here, we analyzed the sequential processes during onset of liver injury and subsequent regeneration based on time-resolved transcriptional regulatory networks (TRNs) to understand the relationship between inflammation, mature organ function, and regeneration. Genome-wide expression and TRN analysis were performed time dependently in mouse liver after acute injury by CCl₄ (2 h, 8 h, 1, 2, 4, 6, 8, 16 days), as well as lipopolysaccharide (LPS, 24 h) and compared to publicly available data after tunicamycin exposure (mouse, 6 h), hepatocellular carcinoma (HCC, mouse), and human chronic liver disease (non-alcoholic fatty liver, HBV infection and HCC). Spatiotemporal investigation differentiated lobular zones for signaling and transcription factor expression. Acute CCl₄ intoxication induced expression of gene clusters enriched for inflammation and stress signaling that peaked between 2 and 24 h, accompanied by a decrease of mature liver functions, particularly metabolic genes. Metabolism decreased not only in pericentral hepatocytes that underwent CCl₄-induced necrosis, but extended to the surviving periportal hepatocytes. Proliferation and tissue restorative TRNs occurred only later reaching a maximum at 48 h. The same upstream regulators (e.g. inhibited RXR function) were implicated in increased inflammation and suppressed metabolism. The concomitant inflammation/metabolism TRN occurred similarly after acute LPS and tunicamycin challenges, in chronic mouse models and also in human liver diseases. Downregulation of metabolic genes occurs concomitantly to induce inflammation-associated genes as an early response and appears to be initiated by similar upstream regulators in acute and chronic liver diseases in humans and mice. In the acute setting, proliferation and restorative regeneration associated TRNs peak only later when metabolism is already suppressed.

Developing a Multi-Dose Computational Model for Drug-Induced Hepatotoxicity Prediction Based on Toxicogenomics Data

Drug-induced hepatotoxicity may cause acute and chronic liver disease, leading to great concern for patient safety. It is also one of the main reasons for drug withdrawal from the market. Toxicogenomics data has been widely used in hepatotoxicity prediction. In our study, we proposed a multi-dose computational model to predict the drug-induced hepatotoxicity based on gene expression and toxicity data. The dose/concentration information after drug treatment is fully utilized in our study based on the dose-response curve, thus a more informative representative of the dose-response relationship is considered. We also proposed a new feature selection method, named MEMO, which is also one important aspect of our multi-dose model in our study, to deal with the high-dimensional toxicogenomics data. We validated the proposed model using the TG-GATEs, which is a large database recording toxicogenomics data from multiple views. The experimental results show that the drug-induced hepatotoxicity can be predicted with high accuracy and efficiency using the proposed predictive model.

Deep Neural Network Models for Predicting Chemically Induced Liver Toxicity Endpoints From Transcriptomic Responses

Improving the accuracy of toxicity prediction models for liver injuries is a key element in evaluating the safety of drugs and chemicals. Mechanism-based information derived from expression (transcriptomic) data, in combination with machine-learning methods, promises to improve the accuracy and robustness of current toxicity prediction models. Deep neural networks (DNNs) have the advantage of automatically assembling the relevant features from a large number of input features. This makes them especially suitable for modeling transcriptomic data, which typically contain thousands of features. Here, we gaged gene- and pathway-level feature selection schemes using single- and multi-task DNN approaches in predicting chemically induced liver injuries (biliary hyperplasia, fibrosis, and necrosis) from whole-genome DNA microarray data. The single-task DNN models showed high predictive accuracy and endpoint specificity, with Matthews correlation coefficients for the three endpoints on 10-fold cross validation ranging from 0.56 to 0.89, with an average of 0.74 in the best feature sets. The DNN models outperformed Random Forest models in cross validation and showed better performance than Support Vector Machine models when tested in the external validation datasets. In the cross validation studies, the effect of the feature selection scheme was negligible among the studied feature sets. Further evaluation of the models on their ability to predict the injury phenotype per se for non-chemically induced injuries revealed the robust performance of the DNN models across these additional external testing datasets. Thus, the DNN models learned features specific to the injury phenotype contained in the gene expression data.

Establishment of a novel safety assessment method for vaccine adjuvant development

Vaccines effectively prevent infectious diseases. Many types of vaccines against various pathogens that threaten humans are currently in widespread use. Recently, adjuvant adaptation has been attempted to activate innate immunity to enhance the effectiveness of vaccines. The effectiveness of adjuvants for vaccinations has been demonstrated in many animal models and clinical trials. Although a highly potent adjuvant tends to have high effectiveness, it also has the potential to increase the risk of side effects such as pain, edema, and fever. Indeed, highly effective adjuvants, such as poly(I:C), have not been clinically applied due to their high risks of toxicity in humans. Therefore, the task in the field of adjuvant development is to clinically apply highly effective and non- or low-toxic adjuvant-containing vaccines. To resolve this issue, it is essential to ensure a low risk of side effects and the high efficacy of an adjuvant in the early developmental phases. This review summarizes the theory and history of the current safety assessment methods for adjuvants, using the inactivated influenza vaccine as a model. Our novel method was developed as a system to judge the safety of a candidate compound using biomarkers identified by genomic technology and statistical tools. A systematic safety assessment tool for adjuvants would be of great use for predicting toxicity during novel adjuvant development, screening, and quality control.

Repetitive Transcranial Magnetic Stimulation to the Unilateral Hemisphere of Rat Brain

Previous rodent models of repetitive transcranial magnetic stimulation (rTMS) adopted whole-brain stimulation instead of unilateral hemispheric rTMS, which is unlike the protocols used for human subjects. We report a successful application of rTMS to the unilateral hemisphere of rat brain. The rTMS was delivered with a low-frequency (1 Hz), high-frequency (20 Hz), or sham stimulation protocol to one side of the brain by using a small 25-mm figure-8 coil. We placed the center of the coil 1 cm lateral to the vertex on the biauricular line and angulated the coil 45° to the ground to minimize a potential direct effect of rTMS on the contralateral cortex. We also used an in-house water cooling system to enable repetitive magnetic stimulation for more than 20 min, even at a 20-Hz stimulation frequency. Increases in the transcriptions of immediate early genes (Arc, Junb, and Egr2) were greater after rTMS than after sham stimulation. After 5 consecutive days of 20-min 1-Hz rTMS, bdnf mRNA expression was significantly higher in stimulated cortex than in contralateral side. The model presented herein will elucidate the molecular mechanisms of rTMS by allowing analysis of the inter-hemispheric difference in its effect.

Dose-dependent effects of alpha-naphthylisothiocyanate disconnect biliary fibrosis from hepatocellular necrosis

Exposure of rodents to the xenobiotic α-naphthylisothiocyanate (ANIT) is an established model of experimental intrahepatic bile duct injury. Administration of ANIT to mice causes neutrophil-mediated hepatocellular necrosis. Prolonged exposure of mice to ANIT also produces bile duct hyperplasia and liver fibrosis. However, the mechanistic connection between ANIT-induced hepatocellular necrosis and bile duct hyperplasia and fibrosis is not well characterized. We examined impact of two different doses of ANIT, by feeding chow containing ANIT (0.05%, 0.1%), on the severity of various liver pathologies in a model of chronic ANIT exposure. ANIT-elicited increases in liver inflammation and hepatocellular necrosis increased with dose. Remarkably, there was no connection between increased hepatocellular necrosis and bile duct hyperplasia and peribiliary fibrosis, as these pathologies increased similarly in mice exposed to either dose of ANIT. The results indicate that the severity of hepatocellular necrosis does not dictate the extent of bile duct hyperplasia/fibrosis in ANIT-exposed mice.

ABC gene-ranking for prediction of drug-induced cholestasis in rats

As legacy toxicogenomics databases have become available, improved data mining approaches are now key to extracting and visualizing subtle relationships between toxicants and gene expression. In the present study, a novel “aggregating bundles of clusters” (ABC) procedure was applied to separate cholestatic from non-cholestatic drugs and model toxicants in the Johnson & Johnson (Janssen) rat liver toxicogenomics database [3]. Drug-induced cholestasis is an important issue, particularly when a new compound enters the market with this liability, with standard preclinical models often mispredicting this toxicity. Three well-characterized cholestasis-responsive genes (Cyp7a1, Mrp3 and Bsep) were chosen from a previous in-house Janssen gene expression signature; these three genes show differing, non-redundant responses across the 90+ paradigm compounds in our database. Using the ABC procedure, extraneous contributions were minimized in comparisons of compound gene responses. All genes were assigned weights proportional to their correlations with Cyp7a1, Mrp3 and Bsep, and a resampling technique was used to derive a stable measure of compound similarity. The compounds that were known to be associated with rat cholestasis generally had small values of this measure relative to each other but also had large values of this measure relative to non-cholestatic compounds. Visualization of the data with the ABC-derived signature showed a very tight, essentially identically behaving cluster of robust human cholestatic drugs and experimental cholestatic toxicants (ethinyl estradiol, LPS, ANIT and methylene dianiline, disulfiram, naltrexone, methapyrilene, phenacetin, alpha-methyl dopa, flutamide, the NSAIDs–—indomethacin, flurbiprofen, diclofenac, flufenamic acid, sulindac, and nimesulide, butylated hydroxytoluene, piperonyl butoxide, and bromobenzene), some slightly less active compounds (3′-acetamidofluorene, amsacrine, hydralazine, tannic acid), some drugs that behaved very differently, and were distinct from both non-cholestatic and cholestatic drugs (ketoconazole, dipyridamole, cyproheptadine and aniline), and many postulated human cholestatic drugs that in rat showed no evidence of cholestasis (chlorpromazine, erythromycin, niacin, captopril, dapsone, rifampicin, glibenclamide, simvastatin, furosemide, tamoxifen, and sulfamethoxazole). Most of these latter drugs were noted previously by other groups as showing cholestasis only in humans. The results of this work suggest that the ABC procedure and similar statistical approaches can be instrumental in combining data to compare toxicants across toxicogenomics databases, extract similarities among responses and reduce unexplained data varation.

Applicability of a gene expression based prediction method to SD and Wistar rats: an example of CARCINOscreen®

Recently, the development of several gene expression-based prediction methods has been attempted in the fields of toxicology. CARCINOscreen® is a gene expression-based screening method to predict carcinogenicity of chemicals which target the liver with high accuracy. In this study, we investigated the applicability of the gene expression-based screening method to SD and Wistar rats by using CARCINOscreen®, originally developed with F344 rats, with two carcinogens, 2,4-diaminotoluen and thioacetamide, and two non-carcinogens, 2,6-diaminotoluen and sodium benzoate. After the 28-day repeated dose test was conducted with each chemical in SD and Wistar rats, microarray analysis was performed using total RNA extracted from each liver. Obtained gene expression data were applied to CARCINOscreen®. Predictive scores obtained by the CARCINOscreen® for known carcinogens were > 2 in all strains of rats, while non-carcinogens gave prediction scores below 0.5. These results suggested that the gene expression based screening method, CARCINOscreen®, can be applied to SD and Wistar rats, widely used strains in toxicological studies, by setting of an appropriate boundary line of prediction score to classify the chemicals into carcinogens and non-carcinogens.

Development of an in vitro renal epithelial disease state model for xenobiotic toxicity testing

There is a growing impetus to develop more accurate, predictive and relevant in vitro models of renal xenobiotic exposure. As part of the EU-FP7, Predict-IV project, a major aim was to develop models that recapitulate not only normal tissue physiology but also aspects of disease conditions that exist as predisposing risk factors for xenobiotic toxicity. Hypoxia, as a common micro-environmental alteration associated with pathophysiology in renal disease, was investigated for its effect on the toxicity profile of a panel of 14 nephrotoxins, using the human proximal tubular epithelial RPTECT/TERT1 cell line. Changes in ATP, glutathione and resazurin reduction, after 14 days of daily repeat exposure, revealed a number of compounds, including adefovir dipivoxil with enhanced toxicity in hypoxia. We observed intracellular accumulation of adefovir in hypoxia and suggest decreases in the efflux transport proteins MRP4, MRP5, NHERF1 and NHERF3 as a possible explanation. MRP5 and NHERF3 were also down-regulated upon treatment with the HIF-1 activator, dimethyloxalylglycine. Interestingly, adefovir dependent gene expression shifted from alterations in cell cycle gene expression to an inflammatory response in hypoxia. The ability to investigate aspects of disease states and their influence on renal toxin handling is a key advantage of in vitro systems developed here. They also allow for detailed investigations into mechanisms of compound toxicity of potential importance for compromised tissue exposure.

Gene expression markers in the zebrafish embryo reflect a hepatotoxic response in animal models and humans

The zebrafish embryo (ZFE) is a promising non-rodent model in toxicology, and initial studies suggested its applicability in detecting hepatotoxic responses. Here, we hypothesize that the detailed analysis of underlying mechanisms of hepatotoxicity in ZFE contributes to the improved identification of hepatotoxic properties of new compounds and to the reduction of rodents used for screening. ZFEs were exposed to nine reference hepatotoxicants, targeted at induction of cholestasis, steatosis and necrosis, and two non-hepatotoxic controls. Histopathology revealed various specific morphological changes in the ZFE hepatocytes indicative of cell injury. Gene expression profiles of the individual compounds were generated using microarrays. Regulation of single genes and of pathways could be linked to hepatotoxic responses in general, but phenotype-specific responses could not be distinguished. Hepatotoxicity-associated pathways included xenobiotic metabolism and oxidoreduction related pathways. Overall analysis of gene expression identified a limited set of potential biomarkers specific for a common hepatotoxicity response. This set included several cytochrome P450 genes (cyp2k19, cyp4v7, cyp2aa3), genes related to liver development (pklr) and genes important in oxidoreduction processes (zgc:163022, zgc:158614, zgc:101858 and sqrdl). In conclusion, the ZFE model allows for identification of hepatotoxicants, without discrimination into specific phenotypes.

Identification of biomarkers that distinguish chemical contaminants based on gene expression profiles

Background

High throughput transcriptomics profiles such as those generated using microarrays have been useful in identifying biomarkers for different classification and toxicity prediction purposes. Here, we investigated the use of microarrays to predict chemical toxicants and their possible mechanisms of action.

Results

In this study, in vitro cultures of primary rat hepatocytes were exposed to 105 chemicals and vehicle controls, representing 14 compound classes. We comprehensively compared various normalization of gene expression profiles, feature selection and classification algorithms for the classification of these 105 chemicals into14 compound classes. We found that normalization had little effect on the averaged classification accuracy. Two support vector machine (SVM) methods, LibSVM and sequential minimal optimization, had better classification performance than other methods. SVM recursive feature selection (SVM-RFE) had the highest overfitting rate when an independent dataset was used for a prediction. Therefore, we developed a new feature selection algorithm called gradient method that had a relatively high training classification as well as prediction accuracy with the lowest overfitting rate of the methods tested. Analysis of biomarkers that distinguished the 14 classes of compounds identified a group of genes principally involved in cell cycle function that were significantly downregulated by metal and inflammatory compounds, but were induced by anti-microbial, cancer related drugs, pesticides, and PXR mediators.

Conclusions

Our results indicate that using microarrays and a supervised machine learning approach to predict chemical toxicants, their potential toxicity and mechanisms of action is practical and efficient. Choosing the right feature and classification algorithms for this multiple category classification and prediction is critical.

A mouse model of mitochondrial complex III dysfunction induced by myxothiazol

Myxothiazol is a respiratory chain complex III (CIII) inhibitor that binds to the ubiquinol oxidation site Qo of CIII. It blocks electron transfer from ubiquinol to cytochrome b and thus inhibits CIII activity. It has been utilized as a tool in studies of respiratory chain function in in vitro and cell culture models. We developed a mouse model of biochemically induced and reversible CIII inhibition using myxothiazol. We administered myxothiazol intraperitoneally at a dose of 0.56 mg/kg to C57Bl/J6 mice every 24 h and assessed CIII activity, histology, lipid content, supercomplex formation, and gene expression in the livers of the mice. A reversible CIII activity decrease to 50% of control value occurred at 2 h post-injection. At 74 h only minor histological changes in the liver were found, supercomplex formation was preserved and no significant changes in the expression of genes indicating hepatotoxicity or inflammation were found. Thus, myxothiazol-induced CIII inhibition can be induced in mice for four days in a row without overt hepatotoxicity or lethality. This model could be utilized in further studies of respiratory chain function and pharmacological approaches to mitochondrial hepatopathies.

Metabolomics in toxicology and preclinical research

Metabolomics, the comprehensive analysis of metabolites in a biological system, provides detailed information about the biochemical/physiological status of a biological system, and about the changes caused by chemicals. Metabolomics analysis is used in many fields, ranging from the analysis of the physiological status of genetically modified organisms in safety science to the evaluation of human health conditions. In toxicology, metabolomics is the -omics discipline that is most closely related to classical knowledge of disturbed biochemical pathways. It allows rapid identification of the potential targets of a hazardous compound. It can give information on target organs and often can help to improve our understanding regarding the mode-of-action of a given compound. Such insights aid the discovery of biomarkers that either indicate pathophysiological conditions or help the monitoring of the efficacy of drug therapies. The first toxicological applications of metabolomics were for mechanistic research, but different ways to use the technology in a regulatory context are being explored. Ideally, further progress in that direction will position the metabolomics approach to address the challenges of toxicology of the 21st century. To address these issues, scientists from academia, industry, and regulatory bodies came together in a workshop to discuss the current status of applied metabolomics and its potential in the safety assessment of compounds. We report here on the conclusions of three working groups addressing questions regarding 1) metabolomics for in vitro studies 2) the appropriate use of metabolomics in systems toxicology, and 3) use of metabolomics in a regulatory context.

An Exploratory Evaluation of the Utility of Transcriptional and Urinary Kidney Injury Biomarkers for the Prediction of Aristolochic Acid–Induced Renal Injury in Male Rats

The predictive value of different urinary and transcriptional biomarkers was evaluated in a proof-of-principle toxicology study in rats using aristolochic acid (AA), a known nephrotoxic agent. Male Wistar rats were orally dosed with 0.1, 1, or 10 mg/kg for 12 days. Urine was collected on days 1, 5, and 12 over 24 hours. Gene expression analysis was also conducted using quantitative real-time polymerase chain reaction and Illumina whole-genome chips. Protein biomarkers (Kim-1, Timp-1, vascular endothelial growth factor, osteopontin, clusterin, cystatin C, calbindin D-28K, β2-microglobulin, α–glutathione S-transferase, GSTY1b, RPA-1, and neutrophil gelatinase-associated lipocalin) were measured in these urine samples. Treatment with AA resulted in a slight dose- and/or time-dependent increase in urinary β2-microglobulin, lipocalin 2, and osteopontin before an increase in serum creatinine or serum urea nitrogen was observed. A strong decrease in urinary calbindin D-28K was also detected. The Compugen Ltd. prediction model scored both the 1- and 10-mg/kg AA dose groups as positive for nephrotoxicity despite the absence of renal histopathological changes. In addition, several previously described transcriptional biomarkers were identified as early predictors of renal toxicity as they were detected before morphological alterations had occurred. Altogether, these findings demonstrated the predictive values of renal biomarkers approved by the Food and Drug Administration, European Medicines Agency, and Pharmaceuticals & Medical Devices Agency in AA-induced renal injury in rats and confirmed the utility of renal transcriptional biomarkers for detecting progression of compound-induced renal injury in rats. In addition, several transcriptional biomarkers identified in this exploratory study could present early predictors of renal tubular epithelium injury in rats.

Recent advances in 2D and 3D in vitro systems using primary hepatocytes, alternative hepatocyte sources and non-parenchymal liver cells and their use in investigating mechanisms of hepatotoxicity, cell signaling and ADME

This review encompasses the most important advances in liver functions and hepatotoxicity and analyzes which mechanisms can be studied in vitro. In a complex architecture of nested, zonated lobules, the liver consists of approximately 80 % hepatocytes and 20 % non-parenchymal cells, the latter being involved in a secondary phase that may dramatically aggravate the initial damage. Hepatotoxicity, as well as hepatic metabolism, is controlled by a set of nuclear receptors (including PXR, CAR, HNF-4α, FXR, LXR, SHP, VDR and PPAR) and signaling pathways. When isolating liver cells, some pathways are activated, e.g., the RAS/MEK/ERK pathway, whereas others are silenced (e.g. HNF-4α), resulting in up- and downregulation of hundreds of genes. An understanding of these changes is crucial for a correct interpretation of in vitro data. The possibilities and limitations of the most useful liver in vitro systems are summarized, including three-dimensional culture techniques, co-cultures with non-parenchymal cells, hepatospheres, precision cut liver slices and the isolated perfused liver. Also discussed is how closely hepatoma, stem cell and iPS cell-derived hepatocyte-like-cells resemble real hepatocytes. Finally, a summary is given of the state of the art of liver in vitro and mathematical modeling systems that are currently used in the pharmaceutical industry with an emphasis on drug metabolism, prediction of clearance, drug interaction, transporter studies and hepatotoxicity. One key message is that despite our enthusiasm for in vitro systems, we must never lose sight of the in vivo situation. Although hepatocytes have been isolated for decades, the hunt for relevant alternative systems has only just begun.

Delineation of the key aspects in the regulation of epithelial monolayer formation

The formation, maintenance, and repair of epithelial barriers are of critical importance for whole-body homeostasis. However, the molecular events involved in epithelial tissue maturation are not fully established. To this end, we investigated the molecular processes involved in renal epithelial proximal-tubule monolayer maturation utilizing transcriptomic, metabolomic, and functional parameters. We uncovered profound dynamic alterations in transcriptional regulation, energy metabolism, and nutrient utilization over the maturation process. Proliferating cells exhibited high glycolytic rates and high transcript levels for fatty acid synthesis genes (FASN), whereas matured cells had low glycolytic rates, increased oxidative capacity, and preferentially expressed genes for beta oxidation. There were dynamic alterations in the expression and localization of several adherens (CDH1, -4, and -16) and tight junction (TJP3 and CLDN2 and -10) proteins. Genes involved in differentiated proximal-tubule function, cilium biogenesis (BBS1), and transport (ATP1A1 and ATP1B1) exhibited increased expression during epithelial maturation. Using TransAM transcription factor activity assays, we could demonstrate that p53 and FOXO1 were highly active in matured cells, whereas HIF1A and c-MYC were highly active in proliferating cells. The data presented here will be invaluable in the further delineation of the complex dynamic cellular processes involved in epithelial cell regulation.

Current challenges and controversies in drug-induced liver injury

Current key challenges and controversies encountered in the identification of potentially hepatotoxic drugs and the assessment of drug-induced liver injury (DILI) are covered in this article. There is substantial debate over the classification of DILI itself, including the definition and validity of terms such as 'intrinsic' and 'idiosyncratic'. So-called idiosyncratic DILI is typically rare and requires one or more susceptibility factors in individuals. Consequently, it has been difficult to reproduce in animal models, which has limited the understanding of its underlying mechanisms despite numerous hypotheses. Advances in predictive models would also help to enable preclinical elimination of drug candidates and development of novel biomarkers. A small number of liver laboratory tests have been routinely used to help identify DILI, but their interpretation can be limited and confounded by multiple factors. Improved preclinical and clinical biomarkers are therefore needed to accurately detect early signals of liver injury, distinguish drug hepatotoxicity from other forms of liver injury, and differentiate mild from clinically important liver injury. A range of potentially useful biomarkers are emerging, although so far most have only been used preclinically, with only a few validated and used in the clinic for specific circumstances. Advances in the development of genomic biomarkers will improve the prediction and detection of hepatic injury in future. Establishing a definitive clinical diagnosis of DILI can be difficult, since it is based on circumstantial evidence by excluding other aetiologies and, when possible, identifying a drug-specific signature. DILI signals based on standard liver test abnormalities may be affected by underlying diseases such as hepatitis B and C, HIV and cancer, as well as the concomitant use of hepatotoxic drugs to treat some of these conditions. Therefore, a modified approach to DILI assessment is justified in these special populations and a suggested framework is presented that takes into account underlying disease when evaluating DILI signals in individuals. Detection of idiosyncratic DILI should, in some respects, be easier in the postmarketing setting compared with the clinical development programme, since there is a much larger and more varied patient population exposure over longer timeframes. However, postmarketing safety surveillance is currently limited by the quantity and quality of information available to make an accurate diagnosis, the lack of a control group and the rarity of cases. The pooling of multiple healthcare databases, which could potentially contain different types of patient data, is advised to address some of these deficiencies.

Microarray analysis of the gene expression profile of HMC-1 mast cells following Schizonepeta tenuifolia Briquet treatment

It has long been believed that mast cells play a crucial role in the development of many physiological changes during immediate allergic responses. This study was conducted to evaluate the anti-inflammation mechanism of Schizonepeta tenuifolia (ST) extract and ST purified chemicals on the PMA plus A23187-induced stimulation of HMC-1 human mast cells. ST, rosmarinic acid, pulegone, and 2α,3α,24-thrihydrooxylen-12en-28oic acid treatment of HMC-1 cells led to significant suppression of pro-inflammatory cytokines (IL-6, IL-8, and TNF-α) in a dose dependent manner. In addition, the results of the microarray and real-time RT-PCR analyses revealed that ST regulates several pathways, including the cytokine-cytokine receptor interaction (CCRI), MAPK, and the Toll-like receptor (TLR) signaling pathways. ST may be useful for the treatment of inflammation disease via anti-inflammation activity that occurs through inhibition of the CCRI, MAPK, and TLR signaling pathways.

Predictive modeling of chemical hazard by integrating numerical descriptors of chemical structures and short-term toxicity assay data

Quantitative structure-activity relationship (QSAR) models are widely used for in silico prediction of in vivo toxicity of drug candidates or environmental chemicals, adding value to candidate selection in drug development or in a search for less hazardous and more sustainable alternatives for chemicals in commerce. The development of traditional QSAR models is enabled by numerical descriptors representing the inherent chemical properties that can be easily defined for any number of molecules; however, traditional QSAR models often have limited predictive power due to the lack of data and complexity of in vivo endpoints. Although it has been indeed difficult to obtain experimentally derived toxicity data on a large number of chemicals in the past, the results of quantitative in vitro screening of thousands of environmental chemicals in hundreds of experimental systems are now available and continue to accumulate. In addition, publicly accessible toxicogenomics data collected on hundreds of chemicals provide another dimension of molecular information that is potentially useful for predictive toxicity modeling. These new characteristics of molecular bioactivity arising from short-term biological assays, i.e., in vitro screening and/or in vivo toxicogenomics data can now be exploited in combination with chemical structural information to generate hybrid QSAR-like quantitative models to predict human toxicity and carcinogenicity. Using several case studies, we illustrate the benefits of a hybrid modeling approach, namely improvements in the accuracy of models, enhanced interpretation of the most predictive features, and expanded applicability domain for wider chemical space coverage.

Biomarker classifiers for identifying susceptible subpopulations for treatment decisions

AIM

A main goal of pharmacogenomics is to develop genomic signatures to predict patients' responses to a drug or therapy for treatment decisions. Identification of patients who would have no beneficial effect or have the risk of developing adverse effects from an unnecessary treatment could save enormous cost in the healthcare system and clinical trials. This article presents an approach for developing a biomarker classifier for identifying a fraction of susceptible patients, who should be spared unnecessary treatment prior to treatment.

MATERIALS & METHODS

The identification of susceptible patients involves two steps. The first step is to identify biomarkers of susceptibility from a mixture of biomarkers of susceptibility and biomarkers of response; the second step is to develop a classifier using an ensemble classification algorithm, as the number of susceptible patients is generally much smaller than the number of nonsusceptible patients.

RESULTS

Selection of the biomarkers of susceptibility is essential to achieve good prediction accuracy. The ensemble algorithm significantly improves the prediction accuracy compared with the standard classifiers.

CONCLUSION

The study shows that classifiers developed based on the biomarkers obtained by comparing the genomic profiles of responders to those of nonresponders may lead to a high misclassification error rate. Classifiers to identify a small fraction of the subpopulation should take imbalanced class sizes into consideration. A large sample size may be needed in order to ensure detection of a sufficient number of biomarkers and a sufficient number of susceptible subjects for classifier development and validation.

Identification of urinary microRNA profiles in rats that may diagnose hepatotoxicity

Circulating microRNAs (miRNAs) have emerged as novel noninvasive biomarkers for several diseases and other types of tissue injury. This study tested the hypothesis that changes in the levels of urinary miRNAs correlate with liver injury induced by hepatotoxicants. Sprague-Dawley rats were administered acetaminophen (APAP) or carbon tetrachloride (CCl(4)) and one nonhepatotoxicant (penicillin/PCN). Urine samples were collected over a 24 h period after a single oral dose of APAP (1250 mg/kg), CCl(4) (2000 mg/kg), or PCN (2400 mg/kg). APAP and CCl(4) induced liver injury based upon increased serum alanine and aspartate aminotransferase levels and histopathological findings, including liver necrosis. APAP and CCl(4) both significantly increased the urinary levels of 44 and 28 miRNAs, respectively. In addition, 10 of the increased miRNAs were in common between APAP and CCl(4). In contrast, PCN caused a slight decrease of a different nonoverlapping set of urinary miRNAs. Cluster analysis revealed a distinct urinary miRNA pattern from the hepatotoxicant-treated groups when compared with vehicle controls and PCN. Analysis of hepatic miRNA levels suggested that the liver was the source of the increased urinary miRNAs after APAP exposure; however, the results from CCl(4) were equivocal. Computational analysis was used to predict target genes of the 10 shared hepatotoxicant-induced miRNAs. Liver gene expression profiling using whole genome microarrays identified eight putative miRNA target genes that were significantly altered in the liver of APAP- and CCl(4)-treated animals. In conclusion, the patterns of urinary miRNA may hold promise as biomarkers of hepatotoxicant-induced liver injury.

Genetic Variations in TXNRD1 as Potential Predictors of Drug-Induced Liver Injury

Purpose

Drug-induced liver injury (DILI) is the most common adverse drug reaction; however, it is not easily predicted. We hypothesize that DILI has a common genetic basis. Based on the findings of previous animal studies on toxic hepatitis, we selected the thioredoxin reductase 1 gene (TXNRD1) as a candidate marker of DILI for this genetic association study.

Methods

Records from 118 patients with DILI were extracted from the database of the Adverse Drug Reaction Research Group in South Korea. Causative drugs included antituberculosis drugs (n=68, 57.6%), antibiotics (n=22, 18.6%), antiepileptic drugs (n=7, 5.9%), non-steroidal anti-inflammatory drugs (n=5, 4.2%), and others (n=16, 13.7%). Seven single nucleotide polymorphisms (SNPs) in TXNRD1 (rs10735393, rs4964287, rs4595619, rs10861201, rs11111997, rs4246270, and rs4246271) were scored in 118 DILI patients and in 120 drug-matched controls without liver injury.

Results

No differences were found between the frequencies of any of the 7 SNPs in the cases and controls; however, a significant association was found between a TTA haplotype composed of rs10735393, rs4964287, and rs4595619 and DILI using an allele model (odds ratio, 1.79; 95% confidence interval, 1.18-2.73; P=0.008; Bonferroni corrected P=0.024).

Conclusions

These results suggest that genetic variations in TXNRD1 favor the development of DILI, although a larger confirmative study is needed.

Biomarkers for drug-induced renal damage and nephrotoxicity-an overview for applied toxicology

The detection of acute kidney injury (AKI) and the monitoring of chronic kidney disease (CKD) is becoming more important in industrialized countries. Because of the direct relation of kidney damage to the increasing age of the population, as well as the connection to other diseases like diabetes mellitus and congestive heart failure, renal diseases/failure has increased in the last decades. In addition, drug-induced kidney injury, especially of patients in intensive care units, is very often a cause of AKI. The need for diagnostic tools to identify drug-induced nephrotoxicity has been emphasized by the ICH-regulated agencies. This has lead to multiple national and international projects focusing on the identification of novel biomarkers to enhance drug development. Several parameters related to AKI or CKD are known and have been used for several decades. Most of these markers deliver information only when renal damage is well established, as is the case for serum creatinine. The field of molecular toxicology has spawned new options of the detection of nephrotoxicity. These new developments lead to the identification of urinary protein biomarkers, including Kim-1, clusterin, osteopontin or RPA-1, and other transcriptional biomarkers which enable the earlier detection of AKI and deliver further information about the area of nephron damage or the underlying mechanism. These biomarkers were mainly identified and qualified in rat but also for humans, several biomarkers have been described and now have to be validated. This review will give an overview of traditional and novel tools for the detection of renal damage.

Predictive Power Estimation Algorithm (PPEA)--a new algorithm to reduce overfitting for genomic biomarker discovery

Toxicogenomics promises to aid in predicting adverse effects, understanding the mechanisms of drug action or toxicity, and uncovering unexpected or secondary pharmacology. However, modeling adverse effects using high dimensional and high noise genomic data is prone to over-fitting. Models constructed from such data sets often consist of a large number of genes with no obvious functional relevance to the biological effect the model intends to predict that can make it challenging to interpret the modeling results. To address these issues, we developed a novel algorithm, Predictive Power Estimation Algorithm (PPEA), which estimates the predictive power of each individual transcript through an iterative two-way bootstrapping procedure. By repeatedly enforcing that the sample number is larger than the transcript number, in each iteration of modeling and testing, PPEA reduces the potential risk of overfitting. We show with three different cases studies that: (1) PPEA can quickly derive a reliable rank order of predictive power of individual transcripts in a relatively small number of iterations, (2) the top ranked transcripts tend to be functionally related to the phenotype they are intended to predict, (3) using only the most predictive top ranked transcripts greatly facilitates development of multiplex assay such as qRT-PCR as a biomarker, and (4) more importantly, we were able to demonstrate that a small number of genes identified from the top-ranked transcripts are highly predictive of phenotype as their expression changes distinguished adverse from nonadverse effects of compounds in completely independent tests. Thus, we believe that the PPEA model effectively addresses the over-fitting problem and can be used to facilitate genomic biomarker discovery for predictive toxicology and drug responses.

In vitro transcriptomic prediction of hepatotoxicity for early drug discovery

Liver toxicity (hepatotoxicity) is a critical issue in drug discovery and development. Standard preclinical evaluation of drug hepatotoxicity is generally performed using in vivo animal systems. However, only a small number of preselected compounds can be examined in vivo due to high experimental costs. A more efficient yet accurate screening technique that can identify potentially hepatotoxic compounds in the early stages of drug development would thus be valuable. Here, we develop and apply a novel genomic prediction technique for screening hepatotoxic compounds based on in vitro human liver cell tests. Using a training set of in vivo rodent experiments for drug hepatotoxicity evaluation, we discovered common biomarkers of drug-induced liver toxicity among six heterogeneous compounds. This gene set was further triaged to a subset of 32 genes that can be used as a multi-gene expression signature to predict hepatotoxicity. This multi-gene predictor was independently validated and showed consistently high prediction performance on five test sets of in vitro human liver cell and in vivo animal toxicity experiments. The predictor also demonstrated utility in evaluating different degrees of toxicity in response to drug concentrations, which may be useful not only for discerning a compound's general hepatotoxicity but also for determining its toxic concentration.

Predicting drug-induced hepatotoxicity using QSAR and toxicogenomics approaches

Quantitative structure-activity relationship (QSAR) modeling and toxicogenomics are typically used independently as predictive tools in toxicology. In this study, we evaluated the power of several statistical models for predicting drug hepatotoxicity in rats using different descriptors of drug molecules, namely, their chemical descriptors and toxicogenomics profiles. The records were taken from the Toxicogenomics Project rat liver microarray database containing information on 127 drugs ( http://toxico.nibio.go.jp/datalist.html ). The model end point was hepatotoxicity in the rat following 28 days of continuous exposure, established by liver histopathology and serum chemistry. First, we developed multiple conventional QSAR classification models using a comprehensive set of chemical descriptors and several classification methods (k nearest neighbor, support vector machines, random forests, and distance weighted discrimination). With chemical descriptors alone, external predictivity (correct classification rate, CCR) from 5-fold external cross-validation was 61%. Next, the same classification methods were employed to build models using only toxicogenomics data (24 h after a single exposure) treated as biological descriptors. The optimized models used only 85 selected toxicogenomics descriptors and had CCR as high as 76%. Finally, hybrid models combining both chemical descriptors and transcripts were developed; their CCRs were between 68 and 77%. Although the accuracy of hybrid models did not exceed that of the models based on toxicogenomics data alone, the use of both chemical and biological descriptors enriched the interpretation of the models. In addition to finding 85 transcripts that were predictive and highly relevant to the mechanisms of drug-induced liver injury, chemical structural alerts for hepatotoxicity were identified. These results suggest that concurrent exploration of the chemical features and acute treatment-induced changes in transcript levels will both enrich the mechanistic understanding of subchronic liver injury and afford models capable of accurate prediction of hepatotoxicity from chemical structure and short-term assay results.

Identification and categorization of liver toxicity markers induced by a related pair of drugs

Drug-induced liver injury (DILI) is the primary adverse event that results in the withdrawal of drugs from the market and a frequent reason for the failure of drug candidates in the pre-clinical or clinical phases of drug development. This paper presents an approach for identifying potential liver toxicity genomic biomarkers from a liver toxicity biomarker study involving the paired compounds entacapone (“non-liver toxic drug”) and tolcapone (“hepatotoxic drug”). Molecular analysis of the rat liver and plasma samples, combined with statistical analysis, revealed many similarities and differences between the in vivo biochemical effects of the two drugs. Six hundred and ninety-five genes and 61 pathways were selected based on the classification scheme. Of the 61 pathways, 5 were specific to treatment with tolcapone. Two of the 12 animals in the tolcapone group were found to have high ALT, AST, or TBIL levels. The gene Vars2 (valyl-tRNA synthetase 2) was identified in both animals and the pathway to which it belongs, the aminoacyl-tRNA biosynthesis pathway, was one of the three most significant tolcapone-specific pathways identified.

Phenotypic and genotypic assessment of concomitant drug-induced toxic effects in liver, kidney and blood

Several studies have characterized drug-induced toxicity in liver and kidney. However, the majority of these studies have been performed with 'individual' organs in isolation. Separately, little is known about the role of whole blood as a surrogate tissue in drug-induced toxicity. Accordingly, we investigated the 'concurrent' response of liver, kidney and whole blood during a toxic assault. Rats were acutely treated with therapeutics (acetaminophen, rosiglitazone, fluconazole, isoniazid, cyclophosphamide, amphotericin B, gentamicin and cisplatin) reported for their liver and/or kidney toxicity. Changes in clinical chemistry parameters (e.g. AST, urea) and/or observed microscopic tissue damage confirmed induced hepatotoxicity and/or nephrotoxicity by all drugs. Drug-induced toxicity was not confined to an 'individual' organ. Not all drugs elicited significant alterations in phenotypic parameters of toxicity (e.g. ALT, creatinine). Accordingly, the transcriptional profile of the organs was studied using a toxicity panel of 30 genes derived from literature. Each of the test drugs generated specific gene expression patterns which were unique for all three organs. Hierarchical cluster analyses of purported hepatotoxicants and nephrotoxicants each led to characteristic 'fingerprints' (e.g. decrease in Cyp3a1 indicative of hepatotoxicity; increase in Spp1 and decrease in Gstp1 indicative of nephrotoxicity). In whole blood cells, a set of genes was derived which closely correlated with individual drug-induced concomitant changes in liver or kidney. Collectively, these data demonstrate drug-induced multi-organ toxicity. Furthermore, our findings underscore the importance of transcriptional profiling during inadequate phenotypic anchorage and suggest that whole blood may be judiciously used as a surrogate for drug-induced extra-hematological organ toxicity.

Induction of vesicular steatosis by amiodarone and tetracycline is associated with up-regulation of lipogenic genes in HepaRG cells

Drug-induced liver injury occurs in general after several weeks and is often unpredictable. It is characterized by a large spectrum of lesions that includes steatosis and phospholipidosis. Many drugs such as amiodarone and tetracycline have been reported to cause phospholipidosis and/or steatosis. In this study, acute and chronic hepatic effects of these two drugs were investigated using well-differentiated human hepatoma HepaRG cells. Accumulation of typical lipid droplets, labeled with Oil Red O, was observed in hepatocyte-like HepaRG cells after repeat exposure to either drug. Amiodarone caused the formation of additional intracytoplasmic vesicles that did not stain in all HepaRG cells. At the electron microscopic level, these vesicles appeared as typical lamellar bodies and were associated with an increase of phosphatidylethanolamine and phosphatidylcholine. A dose-dependent induction of triglycerides (TG) was observed after repeat exposure to either amiodarone or tetracycline. Several genes known to be related to lipogenesis were induced after treatment by these two drugs. By contrast, opposite deregulation of some of these genes (FASN, SCD1, and THSRP) was observed in fat HepaRG cells induced by oleic acid overload, supporting the conclusion that different mechanisms were involved in the induction of steatosis by drugs and oleic acid. Moreover, several genes related to lipid droplet formation (ADFP, PLIN4) were up-regulated after exposure to both drugs and oleic acid.

CONCLUSION

Our results show that amiodarone causes phospholipidosis after short-term treatment and, like tetracycline, induces vesicular steatosis after repeat exposure in HepaRG cells. These data represent the first demonstration that drugs can induce vesicular steatosis in vitro and show a direct relationship between TG accumulation and enhanced expression of lipogenic genes.

FDA-approved drug labeling for the study of drug-induced liver injury

Drug-induced liver injury (DILI) is a leading cause of drugs failing during clinical trials and being withdrawn from the market. Comparative analysis of drugs based on their DILI potential is an effective approach to discover key DILI mechanisms and risk factors. However, assessing the DILI potential of a drug is a challenge with no existing consensus methods. We proposed a systematic classification scheme using FDA-approved drug labeling to assess the DILI potential of drugs, which yielded a benchmark dataset with 287 drugs representing a wide range of therapeutic categories and daily dosage amounts. The method is transparent and reproducible with a potential to serve as a common practice to study the DILI of marketed drugs for supporting drug discovery and biomarker development.

An approach to identifying preclinical biomarkers of susceptibility to drug-induced toxicity

Aim: Drug-induced toxicity that leads to termination of candidate drugs or postmarketing removal of approved drugs can potentially be explained by the existence of susceptible subpopulations. If the susceptible subpopulations are identified in advance, the drug’s benefits could be maximized by optimal treatment decisions. This article presents a statistical model and an approach for identifying pharmacogenomic biomarkers of susceptibility to drug-induced toxicity for detecting the susceptible subpopulations. Materials & methods: Biomarkers are categorized into three disjoint sets: biomarkers of both susceptibility and exposure (A); biomarkers of susceptibility only (B); and biomarkers of exposure only (C). Set B contains the most useful biomarkers to identify susceptible subpopulations prior to drug exposure; these markers demonstrate no change in response before and after drug exposure. We present a sample size analysis to illustrate the issues and challenges facing identifying biomarker set B. Results: The required sample size increases as the proportion of the susceptible subpopulation decreases. The examples demonstrated that at least 75 subjects per group are needed for a population with a 5% susceptible subpopulation and more than 1000 are often necessary. Conclusion: This study demonstrates that the biomarkers identified by common methods are a mixture of biomarkers of exposure and susceptibility (A and C), it further demonstrates that the proposed approach could be used to identify biomarkers of susceptibility (B), where a large sample size may be required for adequate power and low false-positive rate.

Original submitted 14 October 2010; Revision submitted 8 December 2010.