Prediction model of potential hepatocarcinogenicity of rat hepatocarcinogens using a large-scale toxicogenomics database

A novel support vector machine-based 1-day, single-dose prediction model of genotoxic hepatocarcinogenicity in rats

The development of a rapid and accurate model for determining the genotoxicity and carcinogenicity of chemicals is crucial for effective cancer risk assessment. This study aims to develop a 1-day, single-dose model for identifying genotoxic hepatocarcinogens (GHCs) in rats. Microarray gene expression data from the livers of rats administered a single dose of 58 compounds, including 5 GHCs, was obtained from the Open TG-GATEs database and used for the identification of marker genes and the construction of a predictive classifier to identify GHCs in rats. We identified 10 gene markers commonly responsive to all 5 GHCs and used them to construct a support vector machine-based predictive classifier. In the silico validation using the expression data of the Open TG-GATEs database indicates that this classifier distinguishes GHCs from other compounds with high accuracy. To further assess the model's effectiveness and reliability, we conducted multi-institutional 1-day single oral administration studies on rats. These studies examined 64 compounds, including 23 GHCs, with gene expression data of the marker genes obtained via quantitative PCR 24 h after a single oral administration. Our results demonstrate that qPCR analysis is an effective alternative to microarray analysis. The GHC predictive model showed high accuracy and reliability, achieving a sensitivity of 91% (21/23) and a specificity of 93% (38/41) across multiple validation studies in three institutions. In conclusion, the present 1-day single oral administration model proves to be a reliable and highly sensitive tool for identifying GHCs and is anticipated to be a valuable tool in identifying and screening potential GHCs.

Advancing Computational Toxicology by Interpretable Machine Learning

Chemical toxicity evaluations for drugs, consumer products, and environmental chemicals have a critical impact on human health. Traditional animal models to evaluate chemical toxicity are expensive, time-consuming, and often fail to detect toxicants in humans. Computational toxicology is a promising alternative approach that utilizes machine learning (ML) and deep learning (DL) techniques to predict the toxicity potentials of chemicals. Although the applications of ML- and DL-based computational models in chemical toxicity predictions are attractive, many toxicity models are "black boxes" in nature and difficult to interpret by toxicologists, which hampers the chemical risk assessments using these models. The recent progress of interpretable ML (IML) in the computer science field meets this urgent need to unveil the underlying toxicity mechanisms and elucidate the domain knowledge of toxicity models. In this review, we focused on the applications of IML in computational toxicology, including toxicity feature data, model interpretation methods, use of knowledge base frameworks in IML development, and recent applications. The challenges and future directions of IML modeling in toxicology are also discussed. We hope this review can encourage efforts in developing interpretable models with new IML algorithms that can assist new chemical assessments by illustrating toxicity mechanisms in humans.

The evolving role of investigative toxicology in the pharmaceutical industry

For decades, preclinical toxicology was essentially a descriptive discipline in which treatment-related effects were carefully reported and used as a basis to calculate safety margins for drug candidates. In recent years, however, technological advances have increasingly enabled researchers to gain insights into toxicity mechanisms, supporting greater understanding of species relevance and translatability to humans, prediction of safety events, mitigation of side effects and development of safety biomarkers. Consequently, investigative (or mechanistic) toxicology has been gaining momentum and is now a key capability in the pharmaceutical industry. Here, we provide an overview of the current status of the field using case studies and discuss the potential impact of ongoing technological developments, based on a survey of investigative toxicologists from 14 European-based medium-sized to large pharmaceutical companies.

A Collaborative Initiative to Establish Genomic Biomarkers for Assessing Tumorigenic Potential to Reduce Reliance on Conventional Rodent Carcinogenicity Studies

There is growing recognition across broad sectors of the scientific community that use of genomic biomarkers has the potential to reduce the need for conventional rodent carcinogenicity studies of industrial chemicals, agrochemicals, and pharmaceuticals through a weight-of-evidence approach. These biomarkers fall into 2 major categories: (1) sets of gene transcripts that can identify distinct tumorigenic mechanisms of action; and (2) cancer driver gene mutations indicative of rapidly expanding growth-advantaged clonal cell populations. This call-to-action article describes a collaborative approach launched to develop and qualify biomarker gene expression panels that measure widely accepted molecular pathways linked to tumorigenesis and their activation levels to predict tumorigenic doses of chemicals from short-term exposures. Growing evidence suggests that application of such biomarker panels in short-term exposure rodent studies can identify both tumorigenic hazard and tumorigenic activation levels for chemical-induced carcinogenicity. In the future, this approach will be expanded to include methodologies examining mutations in key cancer driver gene mutation hotspots as biomarkers of both genotoxic and nongenotoxic chemical tumor risk. Analytical, technical, and biological validation studies of these complementary genomic tools are being undertaken by multisector and multidisciplinary collaborative teams within the Health and Environmental Sciences Institute. Success from these efforts will facilitate the transition from current heavy reliance on conventional 2-year rodent carcinogenicity studies to more rapid animal- and resource-sparing approaches for mechanism-based carcinogenicity evaluation supporting internal and regulatory decision-making.

Comparative Analysis of Transcriptional Responses to Genotoxic and Non-Genotoxic Agents in the Blood Cell Model TK6 and the Liver Model HepaRG

Transcript signatures are a promising approach to identify and classify genotoxic and non-genotoxic compounds and are of interest as biomarkers or for future regulatory application. Not much data, however, is yet available about the concordance of transcriptional responses in different cell types or tissues. Here, we analyzed transcriptomic responses to selected genotoxic food contaminants in the human p53-competent lymphoblastoid cell line TK6 using RNA sequencing. Responses to treatment with five genotoxins, as well as with four non-genotoxic liver toxicants, were compared with previously published gene expression data from the human liver cell model HepaRG. A significant overlap of the transcriptomic changes upon genotoxic stress was detectable in TK6 cells, whereas the comparison with the HepaRG model revealed considerable differences, which was confirmed by bioinformatic data mining for cellular upstream regulators or pathways. Taken together, the study presents a transcriptomic signature for genotoxin exposure in the human TK6 blood cell model. The data demonstrate that responses in different cell models have considerable variations. Detection of a transcriptomic genotoxin signature in blood cells indicates that gene expression analyses of blood samples might be a valuable approach to also estimate responses to toxic exposure in target organs such as the liver.

Thresholds Derived From Common Measures in Rat Studies Are Predictive of Liver Tumorigenic Chemicals

We hypothesized that typical tissue and clinical chemistry (ClinChem) end points measured in rat toxicity studies exhibit chemical-independent biological thresholds beyond which cancer occurs. Using the rat in vivo TG-GATES study, 75 chemicals were examined across chemical-dose-time comparisons that could be linked to liver tumor outcomes. Thresholds for liver weight to body weight (LW/BW) and 21 serum ClinChem end points were defined as the maximum and minimum values for those exposures that did not lead to liver tumors in rats. Upper thresholds were identified for LW/BW (117%), aspartate aminotransferase (195%), alanine aminotransferase (141%), alkaline phosphatase (152%), and total bilirubin (115%), and lower thresholds were identified for phospholipids (82%), relative albumin (93%), total cholesterol (82%), and total protein (94%). Thresholds derived from the TG-GATES data set were consistent across other acute and subchronic rat studies. A training set of ClinChem and LW/BW thresholds derived from a 38 chemical training set from TG-GATES was predictive of liver tumor outcomes for a test set of 37 independent TG-GATES chemicals (91%). The thresholds were most predictive when applied to 7d treatments (98%). These findings provide support that biological thresholds for common end points in rodent studies can be used to predict chemical tumorigenic potential.

A set of six Gene expression biomarkers and their thresholds identify rat liver tumorigens in short-term assays

Traditional methods for cancer risk assessment are retrospective, resource-intensive, and not feasible for the vast majority of environmental chemicals. In earlier studies, we used a set of six biomarkers to accurately identify liver tumorigens in transcript profiles derived from chemically-treated rats using either a Toxicological Priority Index (ToxPi) approach or using derived biomarker thresholds for cancer. The biomarkers consisting of 7-113 genes are used to predict the most common liver cancer molecular initiating events: genotoxicity, cytotoxicity and activation of the xenobiotic receptors AhR, CAR, ER, and PPARα. In the present study, we apply and evaluate the performance of these methods for cancer prediction in an independent rat liver study of 44 chemicals (6 h-7d exposures) examined by Affymetrix arrays. In the first approach, ToxPi ranking of biomarker scores consistently gave the highest scores to tumorigenic chemical-dose pairs; balanced accuracies for identification of liver tumorigenic chemicals were up to 89 %. The second approach used tumorigenic thresholds derived in the present study or from our earlier study that were set at the maximum value for chemical-dose exposures without detectable liver tumor outcomes. Using these thresholds, balanced accuracies were up to 90 %. Both approaches identified all tumorigenic chemicals. Almost all of the tumorigenic chemicals activated more than one MIE. We also compared biomarker responses between two types of profiling platforms (Affymetrix full-genome array, TempO-Seq 1500+ array containing ∼2600 genes) and found that the lack of the full set of biomarker genes on the 1500+ array resulted in decreased ability to identify chemicals that activate the MIEs. Overall, these results demonstrate that predictive approaches based on the 6 biomarkers could be used in short-term assays to identify chemicals and their doses that induce liver tumors, the most common endpoint in rodent bioassays.

A Set of Six Gene Expression Biomarkers Identify Rat Liver Tumorigens in Short-term Assays

Chemical-induced liver cancer occurs in rodents through well-characterized adverse outcome pathways. We hypothesized that measurement of the 6 most common molecular initiating events (MIEs) in liver cancer adverse outcome pathways in short-term assays using only gene expression will allow early identification of chemicals and their associated doses that are likely to be tumorigenic in the liver in 2-year bioassays. We tested this hypothesis using transcript data from a rat liver microarray compendium consisting of 2013 comparisons of 146 chemicals administered at doses with previously established effects on rat liver tumor induction. Five MIEs were measured using previously characterized gene expression biomarkers composed of gene sets predictive for genotoxicity and activation of 1 or more xenobiotic receptors (aryl hydrocarbon receptor, constitutive activated receptor, estrogen receptor, and peroxisome proliferator-activated receptor α). Because chronic injury can be important in tumorigenesis, we also developed a biomarker for cytotoxicity that had a 96% balanced accuracy. Characterization of the genes in each biomarker set using the unsupervised TXG-MAP network model demonstrated that the genes were associated with distinct functional coexpression modules. Using the Toxicological Priority Index to rank chemicals based on their ability to activate the MIEs showed that chemicals administered at tumorigenic doses clearly gave the highest ranked scores. Balanced accuracies using thresholds derived from either TG-GATES or DrugMatrix data sets to predict tumorigenicity in independent sets of chemicals were up to 93%. These results show that a MIE-directed approach using only gene expression biomarkers could be used in short-term assays to identify chemicals and their doses that cause tumors.

Extraction of peroxisome proliferator-activated receptor α agonist-induced lipid metabolism-related and unrelated genes in rat liver and analysis of their genomic location

Although peroxisome proliferator-activated receptor α (PPARα) agonists are obviously hepatocarcinogenic in rodents, they have been widely used for dyslipidemia and proven to be safe for clinical use without respect to the species difference. It is established that PPARα acts as a part of the transcription factor complex, but its precise mechanism is still unknown. Using the data of Toxicogenomics Database, reliable genes responsive to PPARα agonists, clofibrate, fenofibrate and WY-14,643, in rat liver, were extracted from both in vivo and in vitro data, and sorted by their fold increase. It was found that there were many genes responding to fibrates exclusively in vivo. Most of the in vivo specific genes appear to be unrelated to lipid metabolism and are not upregulated in the kidney. Fifty-seven genes directly related to cell proliferation were extracted from in vivo data, but they were not induced in vitro at all. Analysis of PPAR-responsive elements could not explain the observed difference in induction. To evaluate possible interaction between neighboring genes in gene expression, the correlation of the fold changes of neighboring genes for 22 drugs with various PPARα agonistic potencies were calculated for the genes showing more than 2.5 fold induction by 3 fibrates in vivo, and their genomic location was compared with that of the human orthologue. In the present study, many candidates of genes other than lipid metabolism were selected, and these could be good starting points to elucidate the mechanism of PPARα agonist-induced rodent-specific toxicity.

Identification of Time-Invariant Biomarkers for Non-Genotoxic Hepatocarcinogen Assessment

Non-genotoxic hepatocarcinogens (NGHCs) can only be confirmed by 2-year rodent studies. Toxicogenomics (TGx) approaches using gene expression profiles from short-term animal studies could enable early assessment of NGHCs. However, high variance in the modulation of the genes had been noted among exposure styles and datasets. Expanding from our previous strategy in identifying consensus biomarkers in multiple experiments, we aimed to identify time-invariant biomarkers for NGHCs in short-term exposure styles and validate their applicability to long-term exposure styles. In this study, nine time-invariant biomarkers, namely A2m, Akr7a3, Aqp7, Ca3, Cdc2a, Cdkn3, Cyp2c11, Ntf3, and Sds, were identified from four large-scale microarray datasets. Machine learning techniques were subsequently employed to assess the prediction performance of the biomarkers. The biomarker set along with the Random Forest models gave the highest median area under the receiver operating characteristic curve (AUC) of 0.824 and a low interquartile range (IQR) variance of 0.036 based on a leave-one-out cross-validation. The application of the models to the external validation datasets achieved high AUC values of greater than or equal to 0.857. Enrichment analysis of the biomarkers inferred the involvement of chronic inflammatory diseases such as liver cirrhosis, fibrosis, and hepatocellular carcinoma in NGHCs. The time-invariant biomarkers provided a robust alternative for NGHC prediction.

Application of transcriptomic and microRNA profiling in the evaluation of potential liver carcinogens

Hepatocarcinogens are agents that increase the incidence of liver cancer in exposed animals or humans. It is now established that carcinogenic exposures have a widespread impact on the transcriptome, inducing both adaptive and adverse changes in the activities of genes and pathways. Chemical hepatocarcinogens have also been shown to affect expression of microRNA (miRNA), the evolutionarily conserved noncoding RNA that regulates gene expression posttranscriptionally. Considerable effort has been invested into examining the involvement of mRNA in chemical hepatocarcinogenesis and their potential usage for the classification and prediction of new chemical entities. For miRNA, there has been an increasing number of studies reported over the past decade, although not to the same degree as for transcriptomic studies. Current data suggest that it is unlikely that any gene or miRNA signature associated with short-term carcinogen exposure can replace the rodent bioassay. In this review, we discuss the application of transcriptomic and miRNA profiles to increase mechanistic understanding of chemical carcinogens and to aid in their classification.

Predicting drug-induced hepatotoxicity based on biological feature maps and diverse classification strategies

Identifying hepatotoxicity as early as possible is significant in drug development. In this study, we developed a drug-induced hepatotoxicity prediction model taking account of both the biological context and the computational efficacy based on toxicogenomics data. Specifically, we proposed a novel gene selection algorithm considering gene's participation, named BioCB, to choose the discriminative genes and make more efficient prediction. Then instead of using the raw gene expression levels to characterize each drug, we developed a two-dimensional biological process feature pattern map to represent each drug. Then we employed two strategies to handle the maps and identify the hepatotoxicity, the direct use of maps, named Two-dim branch, and vectorization of maps, named One-dim branch. The two strategies subsequently used the deep convolutional neural networks and LightGBM as predictors, respectively. Additionally, we here for the first time proposed a stacked vectorized gene matrix, which was more predictive than the raw gene matrix. Results validated on both in vivo and in vitro data from two public data sets, the TG-GATES and DrugMatrix, show that the proposed One-dim branch outperforms the deep framework, the Two-dim branch, and has achieved high accuracy and efficiency. The implementation of the proposed method is available at https://github.com/RanSuLab/Hepatotoxicity.

Identification of p53 Activators in a Human Microarray Compendium

Biomarkers predictive of molecular and toxicological effects are needed to interpret emerging high-throughput transcriptomic data streams. The previously characterized 63 gene TGx-DDI biomarker that includes 20 genes known to be regulated by p53 was previously shown to accurately predict DNA damage in chemically treated cells. We comprehensively evaluated whether the molecular basis of the DDI predictions was based on a p53-dependent response. The biomarker was compared to microarray data in a compendium derived from human cells using the Running Fisher test, a nonparametric correlation test. Using the biomarker, we identified conditions that led to p53 activation, including exposure to the chemical nutlin-3 which disrupts interactions between p53 and the negative regulator MDM2 or by knockdown of MDM2. The expression of most of the genes in the biomarker (75%) were found to depend on p53 activation status based on gene behavior after TP53 overexpression or knockdown. The biomarker identified DDI chemicals that were strong inducers of p53 in wild-type cells; these p53 responses were decreased or abolished in cells after p53 knockdown by siRNAs. Using the biomarker, we screened ∼1950 chemicals in ∼9800 human cell line chemical vs control comparisons and identified ∼100 chemicals that caused p53 activation. Among the positive chemicals were many that are known to activate p53 through direct and indirect DNA damaging mechanisms. These results contribute to the evidence that the TGx-DDI biomarker is useful for identifying chemicals that cause DDI and activate p53.

The Carcinogenome Project: In Vitro Gene Expression Profiling of Chemical Perturbations to Predict Long-Term Carcinogenicity

BACKGROUND

Most chemicals in commerce have not been evaluated for their carcinogenic potential. The de facto gold-standard approach to carcinogen testing adopts the 2-y rodent bioassay, a time-consuming and costly procedure. High-throughput in vitro assays are a promising alternative for addressing the limitations in carcinogen screening.

OBJECTIVES

We developed a screening process for predicting chemical carcinogenicity and genotoxicity and characterizing modes of actions (MoAs) using in vitro gene expression assays.

METHODS

We generated a large toxicogenomics resource comprising [Formula: see text] expression profiles corresponding to 330 chemicals profiled in HepG2 (human hepatocellular carcinoma cell line) at multiple doses and replicates. Predictive models of carcinogenicity and genotoxicity were built using a random forest classifier. Differential pathway enrichment analysis was performed to identify pathways associated with carcinogen exposure. Signatures of carcinogenicity and genotoxicity were compared with external sources, including Drugmatrix and the Connectivity Map.

RESULTS

Among profiles with sufficient bioactivity, our classifiers achieved 72.2% Area Under the ROC Curve (AUC) for predicting carcinogenicity and 82.3% AUC for predicting genotoxicity. Chemical bioactivity, as measured by the strength and reproducibility of the transcriptional response, was not significantly associated with long-term carcinogenicity in doses up to [Formula: see text]. However, sufficient bioactivity was necessary for a chemical to be used for prediction of carcinogenicity. Pathway enrichment analysis revealed pathways consistent with known pathways that drive cancer, including DNA damage and repair. The data is available at https://clue.io/CRCGN_ABC , and a portal for query and visualization of the results is accessible at https://carcinogenome.org .

DISCUSSION

We demonstrated an in vitro screening approach using gene expression profiling to predict carcinogenicity and infer MoAs of chemical perturbations. https://doi.org/10.1289/EHP3986.

Using a gene expression biomarker to identify DNA damage-inducing agents in microarray profiles

High-throughput transcriptomic technologies are increasingly being used to screen environmental chemicals in vitro to provide mechanistic context for regulatory testing. The TGx-DDI biomarker is a 64-gene expression profile generated from testing 28 model chemicals or treatments (13 that cause DNA damage and 15 that do not) in human TK6 cells. While the biomarker is very accurate at predicting DNA damage inducing (DDI) potential using the nearest shrunken centroid method, the broad utility of the biomarker using other computational methods is not fully known. Here, we determined the accuracy of the biomarker used with the Running Fisher test, a nonparametric correlation test. In TK6 cells, the methods could readily differentiate DDI and non-DDI compounds with balanced accuracies of 87-97%, depending on the threshold for determining DDI positives. The methods identified DDI agents in the metabolically competent hepatocyte cell line HepaRG (accuracy = 90%) but not in HepG2 cells or hepatocytes derived from embryonic stem cells (60 and 80%, respectively). DDI was also accurately classified when the gene expression changes were derived using the nCounter technology (accuracy = 89%). In addition, we found: (1) not all genes contributed equally to the correlations; (2) the minimal overlap in genes between the biomarker and the individual comparisons required for significant positive correlation was 10 genes, but usually was much higher; and (3) different sets of genes in the biomarker can by themselves contribute to the significant correlations. Overall, these results demonstrate the utility of the biomarker to accurately classify DDI agents. Environ. Mol. Mutagen. 59:772-784, 2018. Published 2018. This article is a U.S. Government work and is in the public domain in the USA.

Evaluation of 12 mouse marker genes in rat toxicogenomics public data, Open TG-GATEs: Discrimination of genotoxic from non-genotoxic hepatocarcinogens

Previously, we proposed 12 marker genes (Aen, Bax, Btg2, Ccnf, Ccng1, Cdkn1a, Gdf15, Lrp1, Mbd1, Phlda3, Plk2 and Tubb4b) to discriminate mouse genotoxic hepatocarcinogens (GTHC) from non-genotoxic hepatocarcinogens (NGTHC). This was determined by qPCR and principal component analysis (PCA), as the aim of an in vivo short-term screening for genotoxic hepatocarcinogens. For this paper, we conducted an application study of the 12 mouse marker genes to rat data, Open TG-GATEs (public data). We analyzed five typical rat GTHC (2-acetamodofluorene, aflatoxin B1, 2-nitrofluorene, N-nitrosodiethylamine and N-nitrosomorpholine), and not only seven typical rat NGTHC (clofibrate, ethanol, fenofibrate, gemfibrozil, hexachlorobenzene, phenobarbital and WY-14643) but also 11 non-genotoxic non-hepatocarcinogens (NGTNHC; allyl alcohol, aspirin, caffeine, chlorpheniramine, chlorpropamide, dexamethasone, diazepam, indomethacin, phenylbutazone, theophylline and tolbutamide) from Open TG-GATEs. The analysis was performed at 3, 6, 9 and 24 h after a single administration and 4, 8, 15 and 29 days after repeated administrations. We transferred Open TG-GATEs DNA microarray data into log₂ data using the "R Project for Statistical Computing". GTHC-specific dose-dependent gene expression changes were observed and significance assessed with the Williams test. Similar significant changes were observed during 3-24 h and 4-29 days, assessed with Welch's t-test, except not for NGTHC or NGTNHC. Significant differential changes in gene expression were observed between GTHC and NGTHC in 11 genes (except not Tubb4b) and between GTHC and NGTNHC in all 12 genes at 24 h and 10 genes (except Ccnf and Mbd1) at 29 days, per Tukey's test. PCA successfully discriminated GTHC from NGTHC and NGTNHC at 24 h and 29 days. The results demonstrate that 12 previously proposed mouse marker genes are useful for discriminating rat GTHC from NGTHC and NGTNHC from Open TG-GATEs.

Chronic dietary toxicity and carcinogenicity studies of dammar resin in F344 rats

Dammar resin is a natural food additive and flavoring substance present in many foods and drinks. The present study evaluates the chronic toxicity and carcinogenicity of dietary dammar resin in F344 rats. Dietary concentrations in the 52-week chronic toxicity study were 0, 0.03, 0.125, 0.5, or 2%. The major treatment-related deleterious effects were body weight suppression, increased relative liver weight, and low hemoglobin levels in males and females. Foci of cellular alteration in the liver were observed in the male 2% group, but not in any other group. The no-observed-adverse-effect level for chronic toxicity was 0.125% for males (200.4 mg/kg b.w./day) and females (241.9 mg/kg b.w./day). Dietary concentrations in the 104-week carcinogenicity study were 0, 0.03, 0.5, or 2%. Dammar resin induced hemorrhagic diathesis in males and females, possibly via the inhibition of extrinsic and intrinsic coagulation pathways. Incidences of hepatocellular adenomas and carcinomas were significantly increased in the male 2% group, but not in any other group. In the 4-week subacute toxicity study, the livers of male rat-fed diet-containing 2% dammar resin had increased levels of protein oxidation and increased the expression of two anti-apoptotic and seven cytochrome P450 (CYP) genes. There was also an increased tendency of oxidative DNA damage. These findings demonstrate that dammar resin is hepatocarcinogenic in male F344 rats and underlines the roles of inhibition of apoptosis, induction of CYP enzymes, and oxidative stress in dammar resin-induced hepatocarcinogenesis.

Isoflurane is a suitable alternative to ether for anesthetizing rats prior to euthanasia for gene expression analysis

Diethyl ether (ether) had been widely used in Japan for anesthesia, despite its explosive properties and toxicity to both humans and animals. We also had used ether as an anesthetic for euthanizing rats for research in the Toxicogenomics Project (TGP). Because the use of ether for these purposes will likely cease, it is required to select an alternative anesthetic which is validated for consistency with existing TGP data acquired under ether anesthesia. We therefore compared two alternative anesthetic candidates, isoflurane and pentobarbital, with ether in terms of hematological findings, serum biochemical parameters, and gene expressions. As a result, few differences among the three agents were observed. In hematological and serum biochemistry analysis, no significant changes were found. In gene expression analysis, four known genes were extracted as differentially expressed genes in the liver of rats anesthetized with ether, isoflurane, or pentobarbital. However, no significant relationships were detected using gene ontology, pathway, or gene enrichment analyses by DAVID and TargetMine. Surprisingly, although it was expected that the lung would be affected by administration via inhalation, only one differentially expressed gene was extracted in the lung. Taken together, our data indicate that there are no significant differences among ether, isoflurane, and pentobarbital with respect to effects on hematological parameters, serum biochemistry parameters, and gene expression. Based on its smallest affect to existing data and its safety profile for humans and animals, we suggest isoflurane as a suitable alternative anesthetic for use in rat euthanasia in toxicogenomics analysis.

Mechanistic roles of microRNAs in hepatocarcinogenesis: A study of thioacetamide with multiple doses and time-points of rats

Environmental chemicals exposure is one of the primary factors for liver toxicity and hepatocarcinoma. Thioacetamide (TAA) is a well-known hepatotoxicant and could be a liver carcinogen in humans. The discovery of early and sensitive microRNA (miRNA) biomarkers in liver injury and tumor progression could improve cancer diagnosis, prognosis, and management. To study this, we performed next generation sequencing of the livers of Sprague-Dawley rats treated with TAA at three doses (4.5, 15 and 45 mg/kg) and four time points (3-, 7-, 14- and 28-days). Overall, 330 unique differentially expressed miRNAs (DEMs) were identified in the entire TAA-treatment course. Of these, 129 DEMs were found significantly enriched for the “liver cancer” annotation. These results were further complemented by pathway analysis (Molecular Mechanisms of Cancer, p53-, TGF-β-, MAPK- and Wnt-signaling). Two miRNAs (rno-miR-34a-5p and rno-miR-455-3p) out of 48 overlapping DEMs were identified to be early and sensitive biomarkers for TAA-induced hepatocarcinogenicity. We have shown significant regulatory associations between DEMs and TAA-induced liver carcinogenesis at an earlier stage than histopathological features. Most importantly, miR-34a-5p is the most suitable early and sensitive biomarker for TAA-induced hepatocarcinogenesis due to its consistent elevation during the entire treatment course.

Identification of consensus biomarkers for predicting non-genotoxic hepatocarcinogens

The assessment of non-genotoxic hepatocarcinogens (NGHCs) is currently relying on two-year rodent bioassays. Toxicogenomics biomarkers provide a potential alternative method for the prioritization of NGHCs that could be useful for risk assessment. However, previous studies using inconsistently classified chemicals as the training set and a single microarray dataset concluded no consensus biomarkers. In this study, 4 consensus biomarkers of A2m, Ca3, Cxcl1, and Cyp8b1 were identified from four large-scale microarray datasets of the one-day single maximum tolerated dose and a large set of chemicals without inconsistent classifications. Machine learning techniques were subsequently applied to develop prediction models for NGHCs. The final bagging decision tree models were constructed with an average AUC performance of 0.803 for an independent test. A set of 16 chemicals with controversial classifications were reclassified according to the consensus biomarkers. The developed prediction models and identified consensus biomarkers are expected to be potential alternative methods for prioritization of NGHCs for further experimental validation.

Detection of non-genotoxic hepatocarcinogens and prediction of their mechanism of action in rats using gene marker sets

Several studies have successfully detected hepatocarcinogenicity in rats based on gene expression data. However, prediction of hepatocarcinogens with certain mechanisms of action (MOAs), such as enzyme inducers and peroxisome proliferator-activated receptor α (PPARα) agonists, can prove difficult using a single model and requires a highly toxic dose. Here, we constructed a model for detecting non-genotoxic (NGTX) hepatocarcinogens and predicted their MOAs in rats. Gene expression data deposited in the Open Toxicogenomics Project-Genomics Assisted Toxicity Evaluation System (TG-GATEs) was used to investigate gene marker sets. Principal component analysis (PCA) was applied to discriminate different MOAs, and a support vector machine algorithm was applied to construct the prediction model. This approach identified 106 probe sets as gene marker sets for PCA and enabled the prediction model to be constructed. In PCA, NGTX hepatocarcinogens were classified as follows based on their MOAs: cytotoxicants, PPARα agonists, or enzyme inducers. The prediction model detected hepatocarcinogenicity with an accuracy of more than 90% in 14- and 28-day repeated-dose studies. In addition, the doses capable of predicting NGTX hepatocarcinogenicity were close to those required in rat carcinogenicity assays. In conclusion, our PCA and prediction model using gene marker sets will help assess the risk of hepatocarcinogenicity in humans based on MOAs and reduce the number of two-year rodent bioassays.

Using Chemical-Induced Gene Expression in Cultured Human Cells to Predict Chemical Toxicity

Chemical toxicity is conventionally evaluated in animal models. However, animal models are resource intensive; moreover, they face ethical and scientific challenges because the outcomes obtained by animal testing may not correlate with human responses. To develop an alternative method for assessing chemical toxicity, we investigated the feasibility of using chemical-induced genome-wide expression changes in cultured human cells to predict the potential of a chemical to cause specific organ injuries in humans. We first created signatures of chemical-induced gene expression in a vertebral-cancer of the prostate cell line for ∼15,000 chemicals tested in the US National Institutes of Health Library of Integrated Network-Based Cellular Signatures program. We then used the signatures to create naı̈ve Bayesian prediction models for chemical-induced human liver cholestasis, interstitial nephritis, and long QT syndrome. Detailed cross-validation analyses indicated that the models were robust with respect to false positives and false negatives in the samples we used to train the models and could predict the likelihood that chemicals would cause specific organ injuries. In addition, we performed a literature search for drugs and dietary supplements, not formally categorized as causing organ injuries in humans but predicted by our models to be most likely to do so. We found a high percentage of these compounds associated with case reports of relevant organ injuries, lending support to the idea that in vitro cell-based experiments can be used to predict the toxic potential of chemicals. We believe that this approach, combined with a robust technique to model human exposure to chemicals, may serve as a promising alternative to animal-based chemical toxicity assessment.

Prediction of Non-Genotoxic Carcinogenicity Based on Genetic Profiles of Short Term Exposure Assays

Non-genotoxic carcinogens are substances that induce tumorigenesis by non-mutagenic mechanisms and long term rodent bioassays are required to identify them. Recent studies have shown that transcription profiling can be applied to develop early identifiers for long term phenotypes. In this study, we used rat liver expression profiles from the NTP (National Toxicology Program, Research Triangle Park, USA) DrugMatrix Database to construct a gene classifier that can distinguish between non-genotoxic carcinogens and other chemicals. The model was based on short term exposure assays (3 days) and the training was limited to oxidative stressors, peroxisome proliferators and hormone modulators. Validation of the predictor was performed on independent toxicogenomic data (TG-GATEs, Toxicogenomics Project-Genomics Assisted Toxicity Evaluation System, Osaka, Japan). To build our model we performed Random Forests together with a recursive elimination algorithm (VarSelRF). Gene set enrichment analysis was employed for functional interpretation. A total of 770 microarrays comprising 96 different compounds were analyzed and a predictor of 54 genes was built. Prediction accuracy was 0.85 in the training set, 0.87 in the test set and increased with increasing concentration in the validation set: 0.6 at low dose, 0.7 at medium doses and 0.81 at high doses. Pathway analysis revealed gene prominence of cellular respiration, energy production and lipoprotein metabolism. The biggest target of toxicogenomics is accurately predict the toxicity of unknown drugs. In this analysis, we presented a classifier that can predict non-genotoxic carcinogenicity by using short term exposure assays. In this approach, dose level is critical when evaluating chemicals at early time points.

Collaborative studies in toxicogenomics in rodent liver in JEMS·MMS; a useful application of principal component analysis on toxicogenomics

Toxicogenomics is a rapidly developing discipline focused on the elucidation of the molecular and cellular effects of chemicals on biological systems. As a collaborative study group of Toxicogenomics/JEMS·MMS, we conducted studies on hepatocarcinogens in rodent liver in which 100 candidate marker genes were selected to discriminate genotoxic hepatocarcinogens from non-genotoxic hepatocarcinogens. Differential gene expression induced by 13 chemicals were examined using DNA microarray and quantitative real-time PCR (qPCR), including eight genotoxic hepatocarcinogens [o-aminoazotoluene, chrysene, dibenzo[a,l]pyrene, diethylnitrosamine (DEN), 7,12-dimethylbenz[a]anthracene, dimethylnitrosamine, dipropylnitrosamine and ethylnitrosourea (ENU)], four non-genotoxic hepatocarcinogens [carbon tetrachloride, di(2-ethylhexyl)phthalate (DEHP), phenobarbital and trichloroethylene] and a non-genotoxic non-hepatocarcinogen [ethanol]. Using qPCR, 30 key genes were extracted from mouse livers at 4 h and 28 days following dose-dependent gene expression alteration induced by DEN and ENU: the most significant changes in gene expression were observed at 4 h. Next, we selected key point times at 4 and 48 h from changes in time-dependent gene expression during the acute phase following administration of chrysene by qPCR. We successfully showed discrimination of eight genotoxic hepatocarcinogens [2-acetylaminofluorene, 2,4-diaminotoluene, diisopropanolnitrosamine, 4-dimethylaminoazobenzene, 4-(methylnitsosamino)-1-(3-pyridyl)-1-butanone, N-nitrosomorpholine, quinoline and urethane] from four non-genotoxic hepatocarcinogens [1,4-dichlorobenzene, dichlorodiphenyltrichloroethane, DEHP and furan] using qPCR and principal component analysis. Additionally, we successfully identified two rat genotoxic hepatocarcinogens [DEN and 2,6-dinitrotoluene] from a nongenotoxic-hepatocarcinogen [DEHP] and a non-genotoxic non-hepatocarcinogen [phenacetin] at 4 and 48 h. The subsequent gene pathway analysis by Ingenuity Pathway Analysis extracted the DNA damage response, resulting from the signal transduction of a p53-class mediator leading to the induction of apoptosis. The present review of these studies suggests that application of principal component analysis on the gene expression profile in rodent liver during the acute phase is useful to predict genotoxic hepatocarcinogens in comparison to non-genotoxic hepatocarcinogens and/or non-carcinogenic hepatotoxins.

Aberrant cell cycle regulation in rat liver cells induced by post-initiation treatment with hepatocarcinogens/hepatocarcinogenic tumor promoters

The present study aimed to determine the onset time of hepatocarcinogen/hepatocarcinogenic tumor promoter-specific cell proliferation, apoptosis and aberrant cell cycle regulation after post-initiation treatment. Six-week-old rats were treated with the genotoxic hepatocarcinogen, carbadox (CRB), the marginally hepatocarcinogenic leucomalachite green (LMG), the tumor promoter, β-naphthoflavone (BNF) or the non-carcinogenic hepatotoxicant, acetaminophen, for 2, 4 or 6 weeks during the post-initiation phase using a medium-term liver bioassay. Cell proliferation activity, expression of G2 to M phase- and spindle checkpoint-related molecules, and apoptosis were immunohistochemically analyzed at week 2 and 4, and tumor promotion activity was assessed at week 6. At week 2, hepatocarcinogen/tumor promoter-specific aberrant cell cycle regulation was not observed. At week 4, BNF and LMG increased cell proliferation together with hepatotoxicity, while CRB did not. Additionally, BNF and CRB reduced the number of cells expressing phosphorylated-histone H3 in both ubiquitin D (UBD)(+) cells and Ki-67(+) proliferating cells, suggesting development of spindle checkpoint dysfunction, regardless of cell proliferation activity. At week 6, examined hepatocarcinogens/tumor promoters increased preneoplastic hepatic foci expressing glutathione S-transferase placental form. These results suggest that some hepatocarcinogens/tumor promoters increase their toxicity after post-initiation treatment, causing regenerative cell proliferation. In contrast, some genotoxic hepatocarcinogens may disrupt the spindle checkpoint without facilitating cell proliferation at the early stage of tumor promotion. This suggests that facilitation of cell proliferation and disruption of spindle checkpoint function are induced by different mechanisms during hepatocarcinogenesis. Four weeks of post-initiation treatment may be sufficient to induce hepatocarcinogen/tumor promoter-specific cellular responses.

An integrative test strategy for cancer hazard identification

Assessment of genotoxic and carcinogenic potential is considered one of the basic requirements when evaluating possible human health risks associated with exposure to chemicals. Test strategies currently in place focus primarily on identifying genotoxic potential due to the strong association between the accumulation of genetic damage and cancer. Using genotoxicity assays to predict carcinogenic potential has the significant drawback that risks from non-genotoxic carcinogens remain largely undetected unless carcinogenicity studies are performed. Furthermore, test systems already developed to reduce animal use are not easily accepted and implemented by either industries or regulators. This manuscript reviews the test methods for cancer hazard identification that have been adopted by the regulatory authorities, and discusses the most promising alternative methods that have been developed to date. Based on these findings, a generally applicable tiered test strategy is proposed that can be considered capable of detecting both genotoxic as well as non-genotoxic carcinogens and will improve understanding of the underlying mode of action. Finally, strengths and weaknesses of this new integrative test strategy for cancer hazard identification are presented.

Application of the TGx-28.65 transcriptomic biomarker to classify genotoxic and non-genotoxic chemicals in human TK6 cells in the presence of rat liver S9

In vitro transcriptional signatures that predict toxicities can facilitate chemical screening. We previously developed a transcriptomic biomarker (known as TGx-28.65) for classifying agents as genotoxic (DNA damaging) and non-genotoxic in human lymphoblastoid TK6 cells. Because TK6 cells do not express cytochrome P450s, we confirmed accurate classification by the biomarker in cells co-exposed to 1% 5,6 benzoflavone/phenobarbital-induced rat liver S9 for metabolic activation. However, chemicals may require different types of S9 for activation. Here we investigated the response of TK6 cells to higher percentages of Aroclor-, benzoflavone/phenobarbital-, or ethanol-induced rat liver S9 to expand TGx-28.65 biomarker applicability. Transcriptional profiles were derived 3 to 4 hr following a 4 hr co-exposure of TK6 cells to test chemicals and S9. Preliminary studies established that 10% Aroclor- and 5% ethanol-induced S9 alone did not induce the TGx-28.65 biomarker genes. Seven genotoxic and two non-genotoxic chemicals (and concurrent solvent and positive controls) were then tested with one of the S9s (selected based on cell survival and micronucleus induction). Relative survival and micronucleus frequency was assessed by flow cytometry in cells 20 hr post-exposure. Genotoxic/non-genotoxic chemicals were accurately classified using the different S9s. One technical replicate of cells co-treated with dexamethasone and 10% Aroclor-induced S9 was falsely classified as genotoxic, suggesting caution in using high S9 concentrations. Even low concentrations of genotoxic chemicals (those not causing cytotoxicity) were correctly classified, demonstrating that TGx-28.65 is a sensitive biomarker of genotoxicity. A meta-analysis of datasets from 13 chemicals supports that different S9s can be used in TK6 cells, without impairing classification using the TGx-28.65 biomarker.

Characterization of Conserved Toxicogenomic Responses in Chemically Exposed Hepatocytes across Species and Platforms

BACKGROUND

Genome-wide expression profiling is increasingly being used to identify transcriptional changes induced by drugs and environmental stressors. In this context, the Toxicogenomics Project-Genomics Assisted Toxicity Evaluation system (TG-GATEs) project generated transcriptional profiles from rat liver samples and human/rat cultured primary hepatocytes exposed to more than 100 different chemicals.

OBJECTIVES

To assess the capacity of the cell culture models to recapitulate pathways induced by chemicals in vivo, we leveraged the TG-GATEs data set to compare the early transcriptional responses observed in the liver of rats treated with a large set of chemicals with those of cultured rat and human primary hepatocytes challenged with the same compounds in vitro.

METHODS

We developed a new pathway-based computational pipeline that efficiently combines gene set enrichment analysis (GSEA) using pathways from the Reactome database with biclustering to identify common modules of pathways that are modulated by several chemicals in vivo and in vitro across species.

RESULTS

We found that some chemicals induced conserved patterns of early transcriptional responses in in vitro and in vivo settings, and across human and rat genomes. These responses involved pathways of cell survival, inflammation, xenobiotic metabolism, oxidative stress, and apoptosis. Moreover, our results support the transforming growth factor beta receptor (TGF-βR) signaling pathway as a candidate biomarker associated with exposure to environmental toxicants in primary human hepatocytes.

CONCLUSIONS

Our integrative analysis of toxicogenomics data provides a comprehensive overview of biochemical perturbations affected by a large panel of chemicals. Furthermore, we show that the early toxicological response occurring in animals is recapitulated in human and rat primary hepatocyte cultures at the molecular level, indicating that these models reproduce key pathways in response to chemical stress. These findings expand our understanding and interpretation of toxicogenomics data from human hepatocytes exposed to environmental toxicants.

Using activation status of signaling pathways as mechanism-based biomarkers to predict drug sensitivity

Many complex traits, as drug response, are associated with changes in biological pathways rather than being caused by single gene alterations. Here, a predictive framework is presented in which gene expression data are recoded into activity statuses of signal transduction circuits (sub-pathways within signaling pathways that connect receptor proteins to final effector proteins that trigger cell actions). Such activity values are used as features by a prediction algorithm which can efficiently predict a continuous variable such as the IC50 value. The main advantage of this prediction method is that the features selected by the predictor, the signaling circuits, are themselves rich-informative, mechanism-based biomarkers which provide insight into or drug molecular mechanisms of action (MoA).

Prediction of human population responses to toxic compounds by a collaborative competition

The ability to computationally predict the effects of toxic compounds on humans could help address the deficiencies of current chemical safety testing. Here, we report the results from a community-based DREAM challenge to predict toxicities of environmental compounds with potential adverse health effects for human populations. We measured the cytotoxicity of 156 compounds in 884 lymphoblastoid cell lines for which genotype and transcriptional data are available as part of the Tox21 1000 Genomes Project. The challenge participants developed algorithms to predict interindividual variability of toxic response from genomic profiles and population-level cytotoxicity data from structural attributes of the compounds. 179 submitted predictions were evaluated against an experimental data set to which participants were blinded. Individual cytotoxicity predictions were better than random, with modest correlations (Pearson's r < 0.28), consistent with complex trait genomic prediction. In contrast, predictions of population-level response to different compounds were higher (r < 0.66). The results highlight the possibility of predicting health risks associated with unknown compounds, although risk estimation accuracy remains suboptimal.

Technical guide for applications of gene expression profiling in human health risk assessment of environmental chemicals

Toxicogenomics promises to be an important part of future human health risk assessment of environmental chemicals. The application of gene expression profiles (e.g., for hazard identification, chemical prioritization, chemical grouping, mode of action discovery, and quantitative analysis of response) is growing in the literature, but their use in formal risk assessment by regulatory agencies is relatively infrequent. Although additional validations for specific applications are required, gene expression data can be of immediate use for increasing confidence in chemical evaluations. We believe that a primary reason for the current lack of integration is the limited practical guidance available for risk assessment specialists with limited experience in genomics. The present manuscript provides basic information on gene expression profiling, along with guidance on evaluating the quality of genomic experiments and data, and interpretation of results presented in the form of heat maps, pathway analyses and other common approaches. Moreover, potential ways to integrate information from gene expression experiments into current risk assessment are presented using published studies as examples. The primary objective of this work is to facilitate integration of gene expression data into human health risk assessments of environmental chemicals.

Asymmetric author-topic model for knowledge discovering of big data in toxicogenomics

The advancement of high-throughput screening technologies facilitates the generation of massive amount of biological data, a big data phenomena in biomedical science. Yet, researchers still heavily rely on keyword search and/or literature review to navigate the databases and analyses are often done in rather small-scale. As a result, the rich information of a database has not been fully utilized, particularly for the information embedded in the interactive nature between data points that are largely ignored and buried. For the past 10 years, probabilistic topic modeling has been recognized as an effective machine learning algorithm to annotate the hidden thematic structure of massive collection of documents. The analogy between text corpus and large-scale genomic data enables the application of text mining tools, like probabilistic topic models, to explore hidden patterns of genomic data and to the extension of altered biological functions. In this paper, we developed a generalized probabilistic topic model to analyze a toxicogenomics dataset that consists of a large number of gene expression data from the rat livers treated with drugs in multiple dose and time-points. We discovered the hidden patterns in gene expression associated with the effect of doses and time-points of treatment. Finally, we illustrated the ability of our model to identify the evidence of potential reduction of animal use.

Comprehensive analysis of DNA methylation and gene expression of rat liver in a 2-stage hepatocarcinogenesis model

Recent studies have shown that epigenetic alterations correlate with carcinogenesis in various tissues. Identification of these alterations might help characterize the early stages of carcinogenesis. We comprehensively analyzed DNA methylation and gene expression in livers obtained from rats exposed to nitrosodiethylamine (DEN) followed by a promoter of hepatic carcinogenesis, phenobarbital (PB). The combination of DEN and PB induced marked increases in number and area of glutathione S-transferase-placental form (GST-P)-positive foci in the liver. In the liver of rats that received 30 mg/kg of DEN, pathway analysis revealed alterations of common genes in terms of gene expression and DNA methylation, and that these alterations were related to immune responses. Hierarchical clustering analysis of the expression of common genes from public data obtained through the Toxicogenomics Project-Genomics Assisted Toxicity Evaluation system (TG-GATEs) showed that carcinogenic compounds clustered together. MBD-seq and GeneChip analysis indicated that major histocompatibility complex class Ib gene RT1-CE5, which has an important role in antigen presentation, was hypomethylated around the promoter region and specifically induced in the livers of DEN-treated rats. Further, immunohistochemical analysis indicated that the co-localization of GST-P and protein homologous to RT1-CE5 was present at the foci of some regions. These results suggest that common genes were altered in terms of both DNA methylation and expression in livers, with preneoplastic foci indicating carcinogenic potential, and that immune responses are involved in early carcinogenesis. In conclusion, the present study identified a specific profile of DNA methylation and gene expression in livers with preneoplastic foci. Early epigenetic perturbations of immune responses might correlate with the early stages of hepatocarcinogenesis.

Detection of initiating potential of non-genotoxic carcinogens in a two-stage hepatocarcinogenesis study in rats

We previously reported a toxicogenomics-based prediction model for hepatocarcinogens in which the expression patterns of signature genes following repeated doses of either genotoxic or non genotoxic compounds were similar. Based on the results of our prediction model, we hypothesized that repeated doses of non-genotoxic carcinogens might have initiating potential. Here, we conducted a two stage hepatocarcinogenesis study in rats exposed to the initiating agent nitrosodiethylamine (DEN), and hepatotoxic compounds thioacetamide (TAA), methapyrilene (MP) and acetaminophen (APAP) for 1-2 weeks followed by the liver tumor promoter phenobarbital (PB). The duration of initial treatment was determined based on positive results from our prediction model. Combined treatment of 3 or 30 mg/kg of genotoxic DEN and PB induced marked increases in altered hepatocellular foci and a DEN dose-dependent increase in the number and area of glutathione S-transferase-placental form (GST-P)-positive foci. A low number of altered hepatocellular foci were also observed in rats treated with TAA at a dose of 45 mg/kg.MP at a dose of 100 mg/kg induced a very low number of foci, but APAP did not. Hierarchical clustering analysis using gene expression data revealed that 2-week treatment with TAA at a dose of 30 mg/kg and MP at 45 mg/kg induced specific expression of DNA damage-related genes, similar to 1-week treatment with DEN at a dose of 30 mg/kg. These results suggest that TAA and MP induce DNA damage, which partially supports our hypothesis. Although this study does not indicate whether tumor growth in response to these compounds can be assessed in this model, our results suggest that cumulative treatment with non genotoxic TAA might have initiating potential in the liver.

Toxicogenomics directory of chemically exposed human hepatocytes

A long-term goal of numerous research projects is to identify biomarkers for in vitro systems predicting toxicity in vivo. Often, transcriptomics data are used to identify candidates for further evaluation. However, a systematic directory summarizing key features of chemically influenced genes in human hepatocytes is not yet available. To bridge this gap, we used the Open TG-GATES database with Affymetrix files of cultivated human hepatocytes incubated with chemicals, further sets of gene array data with hepatocytes from human donors generated in this study, and publicly available genome-wide datasets of human liver tissue from patients with non-alcoholic steatohepatitis (NASH), cirrhosis, and hepatocellular cancer (HCC). After a curation procedure, expression data of 143 chemicals were included into a comprehensive biostatistical analysis. The results are summarized in the publicly available toxicotranscriptomics directory ( http://wiki.toxbank.net/toxicogenomics-map/ ) which provides information for all genes whether they are up- or downregulated by chemicals and, if yes, by which compounds. The directory also informs about the following key features of chemically influenced genes: (1) Stereotypical stress response. When chemicals induce strong expression alterations, this usually includes a complex but highly reproducible pattern named 'stereotypical response.' On the other hand, more specific expression responses exist that are induced only by individual compounds or small numbers of compounds. The directory differentiates if the gene is part of the stereotypical stress response or if it represents a more specific reaction. (2) Liver disease-associated genes. Approximately 20 % of the genes influenced by chemicals are up- or downregulated, also in liver disease. Liver disease genes deregulated in cirrhosis, HCC, and NASH that overlap with genes of the aforementioned stereotypical chemical stress response include CYP3A7, normally expressed in fetal liver; the phase II metabolizing enzyme SULT1C2; ALDH8A1, known to generate the ligand of RXR, one of the master regulators of gene expression in the liver; and several genes involved in normal liver functions: CPS1, PCK1, SLC2A2, CYP8B1, CYP4A11, ABCA8, and ADH4. (3) Unstable baseline genes. The process of isolating and the cultivation of hepatocytes was sufficient to induce some stress leading to alterations in the expression of genes, the so-called unstable baseline genes. (4) Biological function. Although more than 2,000 genes are transcriptionally influenced by chemicals, they can be assigned to a relatively small group of biological functions, including energy and lipid metabolism, inflammation and immune response, protein modification, endogenous and xenobiotic metabolism, cytoskeletal organization, stress response, and DNA repair. In conclusion, the introduced toxicotranscriptomics directory offers a basis for a rationale choice of candidate genes for biomarker evaluation studies and represents an easy to use source of background information on chemically influenced genes.

Identification of specific mRNA signatures as fingerprints for carcinogenesis in mice induced by genotoxic and nongenotoxic hepatocarcinogens

Long-term rodent carcinogenicity studies for evaluation of chemicals and pharmaceuticals concerning their carcinogenic potential to humans are currently receiving critical revision. Additional data from mechanistic studies can support cancer risk assessment by clarifying the underlying mode of action. In the course of the IMI MARCAR project, a European consortium of EFPIA partners and academics, which aims to identify biomarkers for nongenotoxic carcinogenesis, a toxicogenomic mouse liver database was generated. CD-1 mice were orally treated for 3 and 14 days with 3 known genotoxic hepatocarcinogens: C.I. Direct Black 38, Dimethylnitrosamine and 4,4'-Methylenedianiline; 3 nongenotoxic hepatocarcinogens: 1,4-Dichlorobenzene, Phenobarbital sodium and Piperonyl butoxide; 4 nonhepatocarcinogens: Cefuroxime sodium, Nifedipine, Prazosin hydrochloride and Propranolol hydrochloride; and 3 compounds that show ambiguous results in genotoxicity testing: Cyproterone acetate, Thioacetamide and Wy-14643. By liver mRNA expression analysis using individual animal data, we identified 64 specific biomarker candidates for genotoxic carcinogens and 69 for nongenotoxic carcinogens for male mice at day 15. The majority of genotoxic carcinogen biomarker candidates possess functions in DNA damage response (eg, apoptosis, cell cycle progression, DNA repair). Most of the identified nongenotoxic carcinogen biomarker candidates are involved in regulation of cell cycle progression and apoptosis. The derived biomarker lists were characterized with respect to their dependency on study duration and gender and were successfully used to characterize carcinogens with ambiguous genotoxicity test results, such as Wy-14643. The identified biomarker candidates improve the mechanistic understanding of drug-induced effects on the mouse liver that result in hepatocellular adenomas and/or carcinomas in 2-year mouse carcinogenicity studies.

ToxDBScan: Large-scale similarity screening of toxicological databases for drug candidates

We present a new tool for hepatocarcinogenicity evaluation of drug candidates in rodents. ToxDBScan is a web tool offering quick and easy similarity screening of new drug candidates against two large-scale public databases, which contain expression profiles for substances with known carcinogenic profiles: TG-GATEs and DrugMatrix. ToxDBScan uses a set similarity score that computes the putative similarity based on similar expression of genes to identify chemicals with similar genotoxic and hepatocarcinogenic potential. We propose using a discretized representation of expression profiles, which use only information on up- or down-regulation of genes as relevant features. Therefore, only the deregulated genes are required as input. ToxDBScan provides an extensive report on similar compounds, which includes additional information on compounds, differential genes and pathway enrichments. We evaluated ToxDBScan with expression data from 15 chemicals with known hepatocarcinogenic potential and observed a sensitivity of 88 Based on the identified chemicals, we achieved perfect classification of the independent test set. ToxDBScan is publicly available from the ZBIT Bioinformatics Toolbox.

Open TG-GATEs: a large-scale toxicogenomics database

Toxicogenomics focuses on assessing the safety of compounds using gene expression profiles. Gene expression signatures from large toxicogenomics databases are expected to perform better than small databases in identifying biomarkers for the prediction and evaluation of drug safety based on a compound's toxicological mechanisms in animal target organs. Over the past 10 years, the Japanese Toxicogenomics Project consortium (TGP) has been developing a large-scale toxicogenomics database consisting of data from 170 compounds (mostly drugs) with the aim of improving and enhancing drug safety assessment. Most of the data generated by the project (e.g. gene expression, pathology, lot number) are freely available to the public via Open TG-GATEs (Toxicogenomics Project-Genomics Assisted Toxicity Evaluation System). Here, we provide a comprehensive overview of the database, including both gene expression data and metadata, with a description of experimental conditions and procedures used to generate the database. Open TG-GATEs is available from https://toxico.nibiohn.go.jp/english/index.html.

Data mining reveals a network of early-response genes as a consensus signature of drug-induced in vitro and in vivo toxicity

Gene signatures of drug-induced toxicity are of broad interest, but they are often identified from small-scale, single-time point experiments, and are therefore of limited applicability. To address this issue, we performed multivariate analysis of gene expression, cell-based assays, and histopathological data in the TG-GATEs (Toxicogenomics Project-Genomics Assisted Toxicity Evaluation system) database. Data mining highlights four genes-EGR1, ATF3, GDF15 and FGF21-that are induced 2 h after drug administration in human and rat primary hepatocytes poised to eventually undergo cytotoxicity-induced cell death. Modelling and simulation reveals that these early stress-response genes form a functional network with evolutionarily conserved structure and intrinsic dynamics. This is underlined by the fact that early induction of this network in vivo predicts drug-induced liver and kidney pathology with high accuracy. Our findings demonstrate the value of early gene-expression signatures in predicting and understanding compound-induced toxicity. The identified network can empower first-line tests that reduce animal use and costs of safety evaluation.

Cross-platform toxicogenomics for the prediction of non-genotoxic hepatocarcinogenesis in rat

In the area of omics profiling in toxicology, i.e. toxicogenomics, characteristic molecular profiles have previously been incorporated into prediction models for early assessment of a carcinogenic potential and mechanism-based classification of compounds. Traditionally, the biomarker signatures used for model construction were derived from individual high-throughput techniques, such as microarrays designed for monitoring global mRNA expression. In this study, we built predictive models by integrating omics data across complementary microarray platforms and introduced new concepts for modeling of pathway alterations and molecular interactions between multiple biological layers. We trained and evaluated diverse machine learning-based models, differing in the incorporated features and learning algorithms on a cross-omics dataset encompassing mRNA, miRNA, and protein expression profiles obtained from rat liver samples treated with a heterogeneous set of substances. Most of these compounds could be unambiguously classified as genotoxic carcinogens, non-genotoxic carcinogens, or non-hepatocarcinogens based on evidence from published studies. Since mixed characteristics were reported for the compounds Cyproterone acetate, Thioacetamide, and Wy-14643, we reclassified these compounds as either genotoxic or non-genotoxic carcinogens based on their molecular profiles. Evaluating our toxicogenomics models in a repeated external cross-validation procedure, we demonstrated that the prediction accuracy of our models could be increased by joining the biomarker signatures across multiple biological layers and by adding complex features derived from cross-platform integration of the omics data. Furthermore, we found that adding these features resulted in a better separation of the compound classes and a more confident reclassification of the three undefined compounds as non-genotoxic carcinogens.