Differentiation of DNA reactive and non-reactive genotoxic mechanisms using gene expression profile analysis

TGx-DDI, a Transcriptomic Biomarker for Genotoxicity Hazard Assessment of Pharmaceuticals and Environmental Chemicals

Genotoxicity testing is an essential component of the safety assessment paradigm required by regulatory agencies world-wide for analysis of drug candidates, and environmental and industrial chemicals. Current genotoxicity testing batteries feature a high incidence of irrelevant positive findings—particularly for in vitro chromosomal damage (CD) assays. The risk management of compounds with positive in vitro findings is a major challenge and requires complex, time consuming, and costly follow-up strategies including animal testing. Thus, regulators are urgently in need of new testing approaches to meet legislated mandates. Using machine learning, we identified a set of transcripts that responds predictably to DNA-damage in human cells that we refer to as the TGx-DDI biomarker, which was originally referred to as TGx-28.65. We proposed to use this biomarker in conjunction with current genotoxicity testing batteries to differentiate compounds with irrelevant “false” positive findings in the in vitro CD assays from true DNA damaging agents (i.e., for de-risking agents that are clastogenic in vitro but not in vivo). We validated the performance of the TGx-DDI biomarker to identify true DNA damaging agents, assessed intra- and inter- laboratory reproducibility, and cross-platform performance. Recently, to augment the application of this biomarker, we developed a high-throughput cell-based genotoxicity testing system using the NanoString nCounter® technology. Here, we review the status of TGx-DDI development, its integration in the genotoxicity testing paradigm, and progress to date in its qualification at the US Food and Drug Administration (FDA) as a drug development tool. If successfully validated and implemented, the TGx-DDI biomarker assay is expected to significantly augment the current strategy for the assessment of genotoxic hazards for drugs and chemicals.

Mechanisms of DNA-reactive and epigenetic chemical carcinogens: applications to carcinogenicity testing and risk assessment

Chemicals with carcinogenic activity in either animals or humans produce increases in neoplasia through diverse mechanisms. One mechanism is reaction with nuclear DNA. Other mechanisms consist of epigenetic effects involving either modifications of regulatory macromolecules or perturbation of cellular regulatory processes. The basis for distinguishing between carcinogens that have either DNA reactivity or an epigenetic activity as their primary mechanism of action is detailed in this review. In addition, important applications of information on these mechanisms of action to carcinogenicity testing and human risk assessment are discussed.

A Qualitative Modeling Approach for Whole Genome Prediction Using High-Throughput Toxicogenomics Data and Pathway-Based Validation

Efficient high-throughput transcriptomics (HTT) tools promise inexpensive, rapid assessment of possible biological consequences of human and environmental exposures to tens of thousands of chemicals in commerce. HTT systems have used relatively small sets of gene expression measurements coupled with mathematical prediction methods to estimate genome-wide gene expression and are often trained and validated using pharmaceutical compounds. It is unclear whether these training sets are suitable for general toxicity testing applications and the more diverse chemical space represented by commercial chemicals and environmental contaminants. In this work, we built predictive computational models that inferred whole genome transcriptional profiles from a smaller sample of surrogate genes. The model was trained and validated using a large scale toxicogenomics database with gene expression data from exposure to heterogeneous chemicals from a wide range of classes (the Open TG-GATEs data base). The method of predictor selection was designed to allow high fidelity gene prediction from any pre-existing gene expression data set, regardless of animal species or data measurement platform. Predictive qualitative models were developed with this TG-GATES data that contained gene expression data of human primary hepatocytes with over 941 samples covering 158 compounds. A sequential forward search-based greedy algorithm, combining different fitting approaches and machine learning techniques, was used to find an optimal set of surrogate genes that predicted differential expression changes of the remaining genome. We then used pathway enrichment of up-regulated and down-regulated genes to assess the ability of a limited gene set to determine relevant patterns of tissue response. In addition, we compared prediction performance using the surrogate genes found from our greedy algorithm (referred to as the SV2000) with the landmark genes provided by existing technologies such as L1000 (Genometry) and S1500 (Tox21), finding better predictive performance for the SV2000. The ability of these predictive algorithms to predict pathway level responses is a positive step toward incorporating mode of action (MOA) analysis into the high throughput prioritization and testing of the large number of chemicals in need of safety evaluation.

Genotoxic potential of the binary mixture of cyanotoxins microcystin-LR and cylindrospermopsin

Increased eutrophication of water bodies promotes cyanobacterial blooming that is hazardous due to the production of various bioactive compounds. Microcystin-LR (MCLR) is among the most widespread cyanotoxins classified as possible human carcinogen, while cylindrospermopsin (CYN) has only recently been recognized as health concern. Both cyanotoxins are genotoxic; however, the mechanisms of their action differ. They are ubiquitously present in water environment and are often detected together. Therefore, we studied genotoxic potential of the binary mixture of these cyanotoxins. Human hepatoma cells (HepG2) were exposed to a single dose of MCLR (1 μg/mL), graded doses of CYN (0.01-0.5 μg/mL), and their combinations. Comet and Cytokinesis block micronucleus assays were used to detect induction of DNA strand breaks (sb) and genomic instability, respectively, along with the transcriptional analyses of the expression of selected genes involved in xenobiotic metabolism, immediate/early cell response and DNA-damage response. MCLR induced DNA sb that were only transiently present after 4 h exposure, whereas CYN, after 24 h exposure, induced DNA sb and genomic instability. The MCLR/CYN mixture induced DNA sb after 24 h exposure, but to lesser extent as CYN alone. On the other hand, induction of genomic instability by the MCLR/CYN mixture was comparable to that induced by CYN alone. In addition, patterns of changes in the expression of selected genes induced by the MCLR/CYN mixture were not significantly different from those induced by CYN alone. Our results indicate that CYN exerts higher genotoxic potential than MCLR and that genotoxic potential of the MCLR/CYN mixture is comparable to that of CYN alone.

Identification of essential transcription factors for adequate DNA damage response after benzo(a)pyrene and aflatoxin B1 exposure by combining transcriptomics with functional genomics

DNA damage mediates widespread changes in transcription through activation or repression of transcription factors (TFs). However, the consequences of regulating specific TFs for the outcome of the DNA repair process remain incompletely understood. Here, we combined transcriptomics and TF binding prediction with functional genomics to identify TFs essential for adequate DNA repair in HepG2 liver cells after a non-cytotoxic dose of carcinogens benzo(a)pyrene (BaP) (2μM) and aflatoxin B1 (AFB1) (5μM). BaP and AFB1 induced a largely common transcriptional response, mediated by similar TFs. A lentiviral shRNA screen knocking down the top31 identified TFs, was performed to determine their effect on DNA repair by assessing phosphorylation of H2AX (γ-H2AX). In addition to the top candidate p53, we identified several other interesting TFs that modulated γ-H2AX after BaP and AFB1 treatment. Validation studies confirmed the role of p53 in reducing γ-H2AX formation and DNA breaks measured by COMET assay after BaP and AFB1 exposure. Expression of the cell cycle inhibitor p21 was profoundly impaired upon p53 knock-down. In addition, the expression of 2 genes involved in nucleotide exchange repair, DDB2 and XPC was significantly reduced in p53 knock-down cells. Although p63 knock-down affected DNA damage upon BaP treatment this was not associated with altered expression of DDB2 or XPC. Finally, knock-down of ARNT reduced γ-H2AX in response to BaP, which was associated with reduced CYP1A1 expression. Importantly, our results suggest a new role for ARNT and its dimerization partner AHR in the occurrence of H2AX phosphorylation after AFB1 treatment. These data show that modulation of TF activity impacts on the repair of BaP- and AFB1-induced DNA damage. Our study also demonstrates the potential of combining functional genomics with genome-wide expression analysis to identify yet unknown causal relationships, thereby aiding in the interpretation of complex biological systems.

The TGx-28.65 biomarker online application for analysis of transcriptomics data to identify DNA damage-inducing chemicals in human cell cultures

The TGx-28.65 biomarker is a 65-gene expression profile generated from testing 28 model chemicals (13 that cause DNA damage and 15 that do not) in human TK6 cells. It is used to predict whether a chemical induces DNA damage or not. We expanded availability to the biomarker by developing the online TGx-28.65 biomarker application for predicting the DNA damage inducing (DDI) potential of suspect toxicants tested in p53-proficient human cells and assessing putative mode(s) of action (MOA). Applications like this that analyse gene expression data to predict the hazard potential of test chemicals hold great promise for risk assessment paradigms. The TGx-28.65 biomarker interfaces with an analytical tool to predict the probability that a test chemical can directly or indirectly induce DNA damage. User submitted in vitro microarray data are compared to the 28-chemical x 65-gene signature profile and the probability that the data fit the profile for a DDI or a non-DDI (NDDI) chemical is calculated. The results are displayed in the Results Table, which includes the classification probability and hyperlinks to view heatmaps, hierarchical clustering, and principal component analyses of user-input data in the context of the reference profile. The heatmaps and cluster plots, along with the corresponding text data files of fold changes in gene expression and Euclidean distances can be downloaded. Review of the test chemical data in relationship to the biomarker allows rapid identification of key gene alterations associated with DNA damage as well as chemicals in the reference set that produced a similar response. Environ. Mol. Mutagen. 58:529-535, 2017. © 2017 Wiley Periodicals, Inc.

Using Transcriptomics to Evaluate Thresholds in Genotoxicity Dose–Response

Several genotoxic chemicals have been reported to produce threshold-shaped dose–response curves for mutation and genotoxicity assays, both in vivo and in vitro, challenging the current default practice for risk assessment of genotoxic chemicals, which assumes a linear dose–response below the lowest tested dose. Statistical methods cannot determine whether a biological threshold exists with sufficient confidence to overturn this assumption of linearity. Indeed, to truly define the shape of the dose–response curves, we must look to the underlying biology and develop targeted experiments to identify and measure the key processes governing the response of the cell to DNA damage. This chapter describes a series of studies aimed at defining the key transcriptional responses. Two approaches were taken to evaluate transcriptional responses preventing micronucleus induction: (1) comparison of gene signatures for several prototype compounds at a single chemical dose that led to a similar activation of the p53-DNA damage pathway (i.e. 1.5-fold increase in total p53); and (2) evaluation of a subset of chemicals with in-depth dose–response studies. The goal of these efforts was to determine the transcriptional pathways responsible for maintaining homeostasis at low levels of DNA damage, i.e., the biological underpinning of threshold-shaped dose–response curves for mutagenicity.

Randomly amplified polymorphic DNA as a tool for genotoxicity: an assessment

Novel short-term assays are required to substantiate the battery of assessment methods for evaluating the genotoxicity of candidate drugs. In this study, an attempt has been made to evaluate randomly amplified polymorphic DNA (RAPD) analysis for its potential to establish genotoxic effect of a known genotoxicant, ie, ethyl methanesulfonate (EMS) in Swiss mice (Mus musculus). Based on the RAPD profiles, genetic damages were detected in EMS-exposed animals, suggesting its usefulness in scanning whole genome for assessing the genotoxic effects of candidate drugs. The profiles were generated using genomic DNA, isolated from liver prior to treatment and from liver, bone marrow and blood after treatment of the genotoxicant. Measurable differences indicative of genetic damages were observed when the pre- and post-treatment profiles were compared. This suggests that RAPD analysis may be useful for assessing the pre-clinical genotoxic effects of candidate drugs.

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.

Integration of metabolic activation with a predictive toxicogenomics signature to classify genotoxic versus nongenotoxic chemicals in human TK6 cells

The use of integrated approaches in genetic toxicology, including the incorporation of gene expression data to determine the molecular pathways involved in the response, is becoming more common. In a companion article, a genomic biomarker was developed in human TK6 cells to classify chemicals as genotoxic or nongenotoxic. Because TK6 cells are not metabolically competent, we set out to broaden the utility of the biomarker for use with chemicals requiring metabolic activation. Specifically, chemical exposures were conducted in the presence of rat liver S9. The ability of the biomarker to classify genotoxic (benzo[a]pyrene, BaP; aflatoxin B1, AFB1) and nongenotoxic (dexamethasone, DEX; phenobarbital, PB) agents correctly was evaluated. Cells were exposed to increasing chemical concentrations for 4 hr and collected 0 hr, 4 hr, and 20 hr postexposure. Relative survival, apoptosis, and micronucleus frequency were measured at 24 hr. Transcriptome profiles were measured with Agilent microarrays. Statistical modeling and bioinformatics tools were applied to classify each chemical using the genomic biomarker. BaP and AFB1 were correctly classified as genotoxic at the mid- and high concentrations at all three time points, whereas DEX was correctly classified as nongenotoxic at all concentrations and time points. The high concentration of PB was misclassified at 24 hr, suggesting that cytotoxicity at later time points may cause misclassification. The data suggest that the use of S9 does not impair the ability of the biomarker to classify genotoxicity in TK6 cells. Finally, we demonstrate that the biomarker is also able to accurately classify genotoxicity using a publicly available dataset derived from human HepaRG cells.

Theoretical considerations for thresholds in chemical carcinogenesis

There is increasing evidence for non-linear relationships for gene mutations, chromosomal aberrations and even tumor incidences in response to low doses of genotoxic carcinogens. To attain the biological relevance of such non-linear responses, there is a need to identify the underlying defense mechanisms that allow tolerance to low doses of genotoxicants. This communication discusses presumptive cancer prevention mechanisms that may contribute to thresholds, i.e. points of departure, for each endpoint, from initial DNA lesion to tumor formation. We discuss a sequential order of genome protection during carcinogenesis where genotoxicant scavenging, cellular efflux, DNA repair, elimination of damaged cells by apoptosis, autophagy, silencing by DNA damage-triggered replicative senescence, and finally, elimination of transformed (premalignant) cells by the immune system are thought to be responsible for a threshold in tumor formation. We highlight DNA repair, for which experimental evidence has been recently provided to dictate a role in PoDs. In conclusion, from a theoretical perspective it is reasonable to posit that tolerance to low dose levels exists for each requisite step of tumor formation and these tolerance mechanisms are critical in determining thresholds in chemical carcinogenesis.

Role of direct reactivity with metals in chemoprotection by N-acetylcysteine against chromium(VI), cadmium(II), and cobalt(II)

The antioxidant N-acetylcysteine (NAC) is widely used for the assessment of the role of reactive oxygen species (ROS) in various biological processes and adverse drug reactions. NAC has been found to effectively inhibit the toxicity of carcinogenic metals, which was attributed to its potent ROS-suppressive properties. However, the absence of redox activity among some metals and findings from genetic models suggested a more diverse, smaller role of oxidative stress in metal toxicity. Here, we examined mechanisms of chemoprotection by NAC against Cd(II), Co(II), and Cr(VI) in human cells. We found that NAC displayed a broad-spectrum chemoprotective activity against all three metals, including suppression of cytotoxicity, apoptosis, p53 activation, and HSP72 and HIF-1α upregulation. Cytoprotection by NAC was independent of cellular glutathione. NAC strongly inhibited the uptake of all three metals in histologically different types of human cells, explaining its high chemoprotective potential. A loss of Cr(VI) accumulation by cells was caused by NAC-mediated extracellular reduction of chromate to membrane-impermeative Cr(III). Suppression of Co(II) uptake resulted from a rapid formation of Co(II)-NAC conjugates that were unable to enter cells. Our results demonstrate that NAC acts through more than one mechanism in preventing metal toxicity and its chemoprotective activity can be completely ROS-independent. Good clinical safety and effectiveness in Co(II) sequestration suggest that NAC could be useful in the prevention of tissue accumulation and toxic effects of Co ions released by cobalt-chromium hip prostheses.

Chromium(VI) causes interstrand DNA cross-linking in vitro but shows no hypersensitivity in cross-link repair-deficient human cells

Hexavalent chromium is a human carcinogen activated primarily by direct reduction with cellular ascorbate and to a lesser extent, by glutathione. Cr(III), the final product of Cr(VI) reduction, forms six bonds allowing intermolecular cross-linking. In this work, we investigated the ability of Cr(VI) to cause interstrand DNA cross-links (ICLs) whose formation mechanisms and presence in human cells are currently uncertain. We found that in vitro reduction of Cr(VI) with glutathione showed a sublinear production of ICLs, the yield of which was less than 1% of total Cr-DNA adducts at the optimal conditions. Formation of ICLs in fast ascorbate-Cr(VI) reactions occurred during a short reduction interval and displayed a linear dose dependence with the average yield of 1.3% of total adducts. In vitro production of ICLs was strongly suppressed by increasing buffer molarity, indicating inhibitory effects of ligand-Cr(III) binding on the formation of cross-linking species. The presence of ICLs in human cells was assessed from the impact of ICL repair deficiencies on Cr(VI) responses. We found that ascorbate-restored FANCD2-null and isogenic FANCD2-complemented cells showed similar cell cycle inhibition and toxicity by Cr(VI). XPA-null cells are defective in the repair of Cr-DNA monoadducts, but stable knockdowns of ERCC1 or XPF in these cells with extended time for the completion of cross-linking reactions did not produce any sensitization to Cr(VI). Our results together with chemical and steric considerations of Cr(III) reactivity suggest that ICL generation by chromate is probably an in vitro phenomenon occurring at conditions permitting the formation of Cr(III) oligomers.

A review on ochratoxin A transcriptomic studies

The mycotoxin Ochratoxin A (OTA) is a potent renal carcinogen in male rats. Transcriptomic studies on OTA (4 in vitro, 6 in vivo, 2 in vitro/in vivo) have been reviewed. The aim of 6 of them was mainly mechanistic whereas the rest had mostly predictive (1) or evaluation (5) purposes. An overall tendency towards gene expression downregulation was observed, probably as a result of protein synthesis inhibition. DNA damage response genes were not deregulated in most of the studies. Genes involved in acute renal injury, cell survival and cell proliferation were upregulated in several in vivo studies. Apoptosis genes were deregulated in vitro but less affected in vivo; activation of several MAPKs has been observed. Many genes related to oxidative stress or involved in cell-to-cell interaction pathways (Wnt) or cytoskeleton structure appeared to be deregulated either in vitro or in vivo. Regucalcin was highly downregulated in vivo and other calcium homeostasis genes were significantly deregulated in vitro. Genes related to OTA transport (OATs) and metabolism (CYPs) appeared downregulated in vivo. Overall, the mechanism of action of OTA remains unclear, however transcriptomic data have contributed to new mechanistic hypothesis generation and to in vitro-in vivo comparison.

Characterization of formaldehyde's genotoxic mode of action by gene expression analysis in TK6 cells

Gene expression analysis has been established as a tool for the characterization of genotoxic mechanisms of chemical mutagens. It has been suggested that expression analysis is capable of distinguishing compounds that cause DNA damage from those that interfere with mitotic spindle function. Formaldehyde (FA) is known to be a DNA-reactive substance which mainly induces chromosomal damage in cultured mammalian cells. However, there has been concern that FA might also induce leukemia-specific aneuploidies, although recent cytogenetic studies excluded a relevant aneugenic potential of FA. We now investigated whether gene expression profiling can be used as a molecular tool to further characterize FA's genotoxic mode of action and to differentiate between clastogenic and aneugenic activity. TK6 cells were exposed to FA for 4 and 24 h, and changes in gene expression were analyzed using a whole-genome human microarray. Results were compared to the expression profiles of two DNA-damaging clastogens (methyl methanesulfonate and ethyl methanesulfonate) and two aneugens (colcemid and vincristine). The genotoxic activity of FA, MMS and EMS under these conditions was confirmed by comet assay experiments. The gene expression profiles indicated that clastogens and aneugens induce discriminable gene expression patterns. Exposure of TK6 cells to FA led to a discrete gene expression pattern, and all toxicogenomics analyses revealed a closer relationship of FA with clastogens than with aneugens.

Effect of chemical mutagens and carcinogens on gene expression profiles in human TK6 cells

Characterization of toxicogenomic signatures of carcinogen exposure holds significant promise for mechanistic and predictive toxicology. In vitro transcriptomic studies allow the comparison of the response to chemicals with diverse mode of actions under controlled experimental conditions. We conducted an in vitro study in TK6 cells to characterize gene expression signatures of exposure to 15 genotoxic carcinogens frequently used in European industries. We also examined the dose-responsive changes in gene expression, and perturbation of biochemical pathways in response to these carcinogens. TK6 cells were exposed at 3 dose levels for 24 h with and without S9 human metabolic mix. Since S9 had an impact on gene expression (885 genes), we analyzed the gene expression data from cells cultures incubated with S9 and without S9 independently. The ribosome pathway was affected by all chemical-dose combinations. However in general, no similar gene expression was observed among carcinogens. Further, pathways, i.e. cell cycle, DNA repair mechanisms, RNA degradation, that were common within sets of chemical-dose combination were suggested by clustergram. Linear trends in dose–response of gene expression were observed for Trichloroethylene, Benz[a]anthracene, Epichlorohydrin, Benzene, and Hydroquinone. The significantly altered genes were involved in the regulation of (anti-) apoptosis, maintenance of cell survival, tumor necrosis factor-related pathways and immune response, in agreement with several other studies. Similarly in S9+ cultures, Benz[a]pyrene, Styrene and Trichloroethylene each modified over 1000 genes at high concentrations. Our findings expand our understanding of the transcriptomic response to genotoxic carcinogens, revealing the alteration of diverse sets of genes and pathways involved in cellular homeostasis and cell cycle control.

A transcriptomics-based in vitro assay for predicting chemical genotoxicity in vivo

The lack of accurate in vitro assays for predicting in vivo toxicity of chemicals together with new legislations demanding replacement and reduction of animal testing has triggered the development of alternative methods. This study aimed at developing a transcriptomics-based in vitro prediction assay for in vivo genotoxicity. Transcriptomics changes induced in the human liver cell line HepG2 by 34 compounds after treatment for 12, 24, and 48 h were used for the selection of gene-sets that are capable of discriminating between in vivo genotoxins (GTX) and in vivo nongenotoxins (NGTX). By combining transcriptomics with publicly available results for these chemicals from standard in vitro genotoxicity studies, we developed several prediction models. These models were validated by using an additional set of 28 chemicals. The best prediction was achieved after stratification of chemicals according to results from the Ames bacterial gene mutation assay prior to transcriptomics evaluation after 24h of treatment. A total of 33 genes were selected for discriminating GTX from NGTX for Ames-positive chemicals and 22 for Ames-negative chemicals. Overall, this method resulted in 89% accuracy and 91% specificity, thereby clearly outperforming the standard in vitro test battery. Transcription factor network analysis revealed HNF3a, HNF4a, HNF6, androgen receptor, and SP1 as main factors regulating the expression of classifiers for Ames-positive chemicals. Thus, the classical bacterial gene mutation assay in combination with in vitro transcriptomics in HepG2 is proposed as an upgraded in vitro approach for predicting in vivo genotoxicity of chemicals holding a great promise for reducing animal experimentations on genotoxicity.

Differentially expressed genes associated with cisplatin resistance in human ovarian adenocarcinoma cell line A2780

Ovarian cancer cells are usually initially sensitive to platinum-based chemotherapy, such as cisplatin (CDDP), but typically become resistant over time. Such drug resistance is a serious impediment to successful disease treatment, and the molecular mechanisms responsible for resistance are not fully understood. In search of novel mechanisms that may lead to the development of CDDP chemoresistance, we used subtractive hybridization to identify differentially expressed genes in CDDP resistant CP70 and C200 cells vs. CDDP sensitive A2780 human ovarian adenocarcinoma cells. We analyzed 256 randomly selected clones. Subtraction efficiency was determined by dot blot and DNA sequencing. Confirmation of differentially expressed cDNAs was done by virtual northern blot analysis, and 17 genes that were differentially expressed in CDDP resistant cell lines vs. CDDP sensitive A2780 cells were identified. The expression of 10 of these genes was low or undetectable in sensitive A2780 cells in comparison to resistant cells and an additional seven genes were more highly expressed in resistant CP70 and C200 vs. A2780 cells. Our identified genes are involved in numerous and diverse cellular processes, such as inhibition of apoptosis (ARHGDIB), stress response (HSPCA, TRA1), chromatin condensation (CNAP1, RanBP2), invasiveness of cells (MMP10), alteration of Ca(2+) homeostasis (ASPH, ATP2B1) and others. Further characterization of these genes and gene products should yield important insights into the biology of CDDP resistance in ovarian carcinoma.

New and emerging technologies for genetic toxicity testing

The International Life Sciences Institute (ILSI) Health and Environmental Sciences Institute (HESI) Project Committee on the Relevance and Follow-up of Positive Results in In Vitro Genetic Toxicity (IVGT) Testing established an Emerging Technologies and New Strategies Workgroup to review the current State of the Art in genetic toxicology testing. The aim of the workgroup was to identify promising technologies that will improve genotoxicity testing and assessment of in vivo hazard and risk, and that have the potential to help meet the objectives of the IVGT. As part of this initiative, HESI convened a workshop in Washington, DC in May 2008 to discuss mature, maturing, and emerging technologies in genetic toxicology. This article collates the abstracts of the New and Emerging Technologies Workshop together with some additional technologies subsequently considered by the workgroup. Each abstract (available in the online version of the article) includes a section addressed specifically to the strengths, weaknesses, opportunities, and threats associated with the respective technology. Importantly, an overview of the technologies and an indication of how their use might be aligned with the objectives of IVGT are presented. In particular, consideration was given with regard to follow-up testing of positive results in the standard IVGT tests (i.e., Salmonella Ames test, chromosome aberration assay, and mouse lymphoma assay) to add weight of evidence and/or provide mechanism of action for improved genetic toxicity risk assessments in humans.

Transcriptomic network analysis of micronuclei-related genes: a case study

Mechanistically relevant information on responses of humans to xenobiotic exposure in relation to chemically induced biological effects, such as micronuclei (MN) formation can be obtained through large-scale transcriptomics studies. Network analysis may enhance the analysis and visualisation of such data. Therefore, this study aimed to develop a 'MN formation' network based on a priori knowledge, by using the pathway tool MetaCore. The gene network contained 27 genes and three gene complexes that are related to processes involved in MN formation, e.g. spindle assembly checkpoint, cell cycle checkpoint and aneuploidy. The MN-related gene network was tested against a transcriptomics case study associated with MN measurements. In this case study, transcriptomic data from children and adults differentially exposed to ambient air pollution in the Czech Republic were analysed and visualised on the network. Six genes from the network, i.e. BAX, DMNT1, PCNA, HIC1, p21 and CDC20, were retrieved. Based on these six genes and in combination with p53 and IL-6, a dedicated network was created. This dedicated network is possibly suited for the development of a reporter gene assay that could be used to screen populations complementary to the current MN test assay. In conclusion, we have shown that network analysis of transcriptomics data in relation to the formation of MN is possible and provides a novel mechanistic hypothesis by indicating which genes are regulated and influence others.

DNA damage and toxicogenomic analyses of hydrogen sulfide in human intestinal epithelial FHs 74 Int cells

Hydrogen sulfide (H(2)S), a metabolic end product of sulfate-reducing bacteria, represents a genotoxic insult to the colonic epithelium, which may also be linked with chronic disorders such as ulcerative colitis and colorectal cancer. This study defined the early (30 min) and late (4 hr) response of nontransformed human intestinal epithelial cells (FHs 74 Int) to H(2)S. The genotoxicity of H(2)S was measured using the single-cell gel electrophoresis (comet) assay. Changes in gene expression were analyzed after exposure to a genotoxic, but not cytotoxic, concentration of H(2)S (500 muM H(2)S) using pathway-specific quantitative RT-PCR gene arrays. H(2)S was genotoxic in a concentration range from 250 to 2,000 microM, which is similar to concentrations found in the large intestine. Significant changes in gene expression were predominantly observed at 4 hr, with the greatest responses by PTGS2 (COX-2; 7.92-fold upregulated) and WNT2 (7.08-fold downregulated). COX-2 was the only gene upregulated at both 30 min and 4 hr. Overall, the study demonstrates that H(2)S modulates the expression of genes involved in cell-cycle progression and triggers both inflammatory and DNA repair responses. This study confirms the genotoxic properties of H(2)S in nontransformed human intestinal epithelial cells and identifies functional pathways by which this bacterial metabolite may perturb cellular homeostasis and contribute to the onset of chronic intestinal disorders.

Human cell toxicogenomic analysis of bromoacetic acid: a regulated drinking water disinfection by-product

The disinfection of drinking water is a major achievement in protecting the public health. However, current disinfection methods also generate disinfection by-products (DBPs). Many DBPs are cytotoxic, genotoxic, teratogenic, and carcinogenic and represent an important class of environmentally hazardous chemicals that may carry long-term human health implications. The objective of this research was to integrate in vitro toxicology with focused toxicogenomic analysis of the regulated DBP, bromoacetic acid (BAA) and to evaluate modulation of gene expression involved in DNA damage/repair and toxic responses, with nontransformed human cells. We generated transcriptome profiles for 168 genes with 30 min and 4 hr exposure times that did not induce acute cytotoxicity. Using qRT-PCR gene arrays, the levels of 25 transcripts were modulated to a statistically significant degree in response to a 30 min treatment with BAA (16 transcripts upregulated and nine downregulated). The largest changes were observed for RAD9A and BRCA1. The majority of the altered transcript profiles are genes involved in DNA repair, especially the repair of double strand DNA breaks, and in cell cycle regulation. With 4 hr of treatment the expression of 28 genes was modulated (12 upregulated and 16 downregulated); the largest fold changes were in HMOX1 and FMO1. This work represents the first nontransformed human cell toxicogenomic study with a regulated drinking water disinfection by-product. These data implicate double strand DNA breaks as a feature of BAA exposure. Future toxicogenomic studies of DBPs will further strengthen our limited knowledge in this growing area of drinking water research.

Characterization and interlaboratory comparison of a gene expression signature for differentiating genotoxic mechanisms

The genotoxicity testing battery is highly sensitive for detection of chemical carcinogens. However, it features a low specificity and provides only limited mechanistic information required for risk assessment of positive findings. This is especially important in case of positive findings in the in vitro chromosome damage assays, because chromosome damage may be also induced secondarily to cell death. An increasing body of evidence indicates that toxicogenomic analysis of cellular stress responses provides an insight into mechanisms of action of genotoxicants. To evaluate the utility of such a toxicogenomic analysis we evaluated gene expression profiles of TK6 cells treated with four model genotoxic agents using a targeted high density real-time PCR approach in a multilaboratory project coordinated by the Health and Environmental Sciences Institute Committee on the Application of Genomics in Mechanism-based Risk Assessment. We show that this gene profiling technology produced reproducible data across laboratories allowing us to conclude that expression analysis of a relevant gene set is capable of distinguishing compounds that cause DNA adducts or double strand breaks from those that interfere with mitotic spindle function or that cause chromosome damage as a consequence of cytotoxicity. Furthermore, our data suggest that the gene expression profiles at early time points are most likely to provide information relevant to mechanisms of genotoxic damage and that larger gene expression arrays will likely provide richer information for differentiating molecular mechanisms of action of genotoxicants. Although more compounds need to be tested to identify a robust molecular signature, this study confirms the potential of toxicogenomic analysis for investigation of genotoxic mechanisms.

Adenoviral-mediated gene transfer of Gadd45a results in suppression by inducing apoptosis and cell cycle arrest in pancreatic cancer cell

BACKGROUND

The extremely poor prognosis of patients with pancreatic ductal adenocarcinoma indicates the need for novel therapeutic approaches. The growth arrest and DNA damage-inducible (Gadd) gene Gadd45a is a member of a group of genes that are induced by DNA damaging agents and growth arrest signals.

METHODS

We evaluated the biological activity of Gadd45a in pancreatic ductal adenocarcinoma cancer-derived cell lines and assessed the efficacy of a combined treatment with adenoviral-mediated expression of Gadd45a (Ad-G45a) and anticancer drug (Etoposide, cisplatin, 5-fluorouracil, respectively) for the PANC1 cell line.

RESULTS

Gadd45a is variously expressed in cell lines derived from pancreatic ductal adenocarcinoma cancer and adenoviral-mediated expression of Gadd45a (Ad-G45a) in these cells results in apoptosis via caspase activation and cell-cycle arrest in the G2/M phase. Gadd45a significantly increased the chemosensitivity of PANC1, which may be due to abundant apoptosis induction and cell cycle arrest. By combinational treatment of Ad-G45a infection and chemotherapeutics, Gadd45a expression was elevated to a higher extent in cancer cells with wild-type p53 than in that with knocked-out p53 status, indicating a higher chemosensitivity to cancer chemotherapy.

CONCLUSIONS

Gadd45a may be a promising candidate for use in cancer gene therapy in combination with chemotherapeutic agents.

Genetic toxicity assessment: employing the best science for human safety evaluation Part VII: Why not start with a single test: a transformational alternative to genotoxicity hazard and risk assessment

A transformational alternative for genotoxicity hazard and risk assessment is proposed to the current standard regulatory test battery. In principle, the proposed approach consists of a single in vitro test system with high genomic sequence homology to humans that addresses the relevant principal genetic lesions assessed in the current test battery. The single test system also possesses higher throughput attributes to permit the screening of large numbers of compounds and allow for an initial differentiation of genotoxic mechanisms (i.e., direct vs. indirect mechanisms) by how the hazard end point is measured. To differentiate compounds showing positive results, toxicogenomic analysis can be conducted to evaluate genotoxic mechanisms and further support risk assessment. Lastly, the results from the single test system can be followed up with a complementary in vivo assessment to establish mechanistic relevance at potential target tissues. Here, we propose the in vitro (yeast) DNA deletion (DEL) recombination assay as a single test alternative to the current genotoxicity test battery with a mechanistic follow up toxicogenomic analysis of genotoxic stress response as one approach that requires broader evaluation and validation. In this assay, intrachromosomal recombination events between a repeated DNA sequence lead to DNA deletions, which have been shown to be inducible by a variety of carcinogens including those both negative and positive in the standard Salmonella Ames assay. It is hoped that the general framework outlined along with this specific example will provoke broader interest to propose other potential test systems.

Characterization of DNA reactive and non-DNA reactive anticancer drugs by gene expression profiling

Gene expression profiling technology is expected to advance our understanding of genotoxic mechanisms involving direct or indirect interaction with DNA. We exposed human lymphoblastoid TK6 cells to 14 anticancer drugs (vincristine, paclitaxel, etoposide, daunorubicin, camptothecin, amsacrine, cytosine arabinoside, hydroxyurea, methotrexate, 5-fluorouracil, cisplatin, 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (CCNU), 1,3-bis (2-chloroethyl)-1-nitrosourea (BCNU), and bleomycin) for 4-h and examined them immediately or after a 20-h recovery period. Cytotoxicity and genotoxicity, respectively, were evaluated by cell counting and by in vitro micronucleus assay at 24h. Effects on the cell cycle were determined by flow cytometry at 4 and 24h. Gene expression was profiled at both sampling times by using human Affymetrix U133A GeneChips (22K). Bioanalysis was done with Resolver/Rosetta software and an in-house annotation program. Cell cycle analysis and gene expression profiling allowed us to classify the drugs according to their mechanisms of action. The molecular signature is composed of 28 marker genes mainly involved in signal transduction and cell cycle pathways. Our results suggest that these marker genes could be used as a predictive model to classify genotoxins according to their direct or indirect interaction with DNA.

Application of toxicogenomics to genetic toxicology risk assessment

Based on the assumption that compounds having similar toxic modes of action induce specific gene expression changes, the toxicity of unknown compounds can be predicted after comparison of their molecular fingerprints with those obtained with compounds of known toxicity. These predictive models will therefore rely on the characterization of marker genes. Toxicogenomics (TGX) also provides mechanistic insight into the mode of toxicity, and can therefore be used as an adjunct to the standard battery of genotoxicity tests. Promising results, highlighting the ability of TGX to differentiate genotoxic from non-genotoxic carcinogens, as well as DNA-reactive from non-DNA reactive genotoxins, have been reported. Additional data suggested the possibility of ranking genotoxins according to the nature of their interactions with DNA. This new approach could contribute to the improvement of risk assessment. TGX could be applied as a follow-up testing strategy in case of positive in vitro genotoxicity findings, and could contribute to improve our ability to identify the molecular mechanism of action and to possibly better assess dose-response curves. TGX has been found to be less sensitive than the standard genotoxicity end-points, probably because it measures the whole cell population response, when compared with standard tests designed to detect rare events in a small number of cells. Further validation will be needed (1) to better link the profiles obtained with TGX to the established genotoxicity end-points, (2) to improve the gene annotation tools, and (3) to standardise study design and data analysis and to better evaluate the impact of variability between platforms and bioinformatics approaches.

Toxicogenomics: overview and potential applications for the study of non-covalent DNA interacting chemicals

Non-covalent DNA interacting agents, DNA-groove binding chemicals and DNA intercalators, are generally considered less cytotoxic than agents producing covalent DNA adducts and other DNA damage. Although the impact of non-covalent compound-DNA interactions on convoluted molecular and biochemical pathways is not well characterized, the most prominent effects include DNA conformational and related structural perturbations, interference with normal DNA protein interactions, such as topoisomerases, as well as effects on mitochondrial DNA and function. The cellular responses to such perturbations would be expected to include changes in transcription of a variety of genes. The emerging field of toxicogenomics seeks to exploit gene responses to define expression profiling signatures for various types of drugs and toxicants, and to provide mechanistic insight into their cellular effects. There are a variety of examples whereby different classes of genotoxicants and non-genotoxic agents can be distinguished by gene expression profiling using functional genomics approaches, which survey global transcriptional responses. In this review, we will discuss the promises and precautions in the use of functional genomics approaches to characterize stress agents including non-covalent DNA interacting agents.

In vitro gene expression data supporting a DNA non-reactive genotoxic mechanism for ochratoxin A

Ochratoxin A (OTA) is a mycotoxin often found in cereals and agricultural products. There is unequivocal evidence of renal carcinogenicity of OTA in male rats, although the mechanism of action is unknown. At present, available data support an epigenetic mechanism (DNA non-reactive) resulting from oxidative stress and cytotoxicity, because a direct OTA interaction with DNA has not been demonstrated. Genotoxic mechanism (DNA-reactive vs. DNA non-reactive) may have implications on human risk assessment. Therefore, the aim of the present work was to identify biological pathways modulated by OTA in vitro in a human renal cell line (HK-2) to contribute to the elucidation of the mechanism of OTA toxicity. For that purpose, cells were exposed to 50 microM OTA during 6 and 24 h, and gene expression profiles were analyzed using Affymetrix Human Genome U133 A 2.0 Gene Chips. Under the same experimental conditions, genotoxicity was evaluated by the modified comet assay using FPG and Endo III to detect oxidative DNA damage, and intracellular ROS level by the H(2)DCF assay. After 6 h, with slight cytotoxicity (83% survival), genes involved in mitochondrial electron transport chain were up-regulated; and after 24 h, with a more pronounced cytotoxicity (51% survival), genes implicated in oxidative stress response were also up-regulated. Increase in intracellular ROS level and oxidative DNA damage was evident at both exposure times being more pronounced with high cytotoxicity. On the contrary, up-regulation of genes implicated in DNA damage response, as cell cycle control or apoptosis, was not detected at any exposure time. In conclusion, these results support a DNA non-reactive mechanism of OTA genotoxicity.

Genome-wide differential gene expression in children exposed to air pollution in the Czech Republic

The Teplice area in the Czech Republic is a mining district where elevated levels of air pollution including airborne carcinogens, have been demonstrated, especially during winter time. This environmental exposure can impact human health; in particular children may be more vulnerable. To study the impact of air pollution in children at the transcriptional level, peripheral blood cells were subjected to whole genome response analysis, in order to identify significantly modulated biological pathways and processes as a result of exposure. Using genome-wide oligonucleotide microarrays, we investigated differential gene expression in children from the Teplice area (n=23) and compared them with children from the rural control area of Prachatice (n=24). In an additional approach, individual gene expressions were correlated with individual peripheral blood lymphocyte micronuclei frequencies, in order to evaluate the linkage of individual gene expressions with an established biomarker of effect that is representative for increased genotoxic risk. Children from the Teplice area showed a significantly higher average micronuclei frequency than Prachatice children (p=0.023). For considerable numbers of genes, the expression differed significantly between the children from the two areas. Amongst these genes, considerable numbers of genes were observed to correlate significantly with the frequencies of micronuclei. The main biological process that appeared significantly affected overall was nucleosome assembly. This suggests an effect of air pollution on the primary structural unit of the condensed DNA. In addition, several other pathways were modulated. Based on the results of this study, we suggest that transcriptomic analysis represents a promising biomarker for environmental carcinogenesis.

Gene expression profile analysis: an emerging approach to investigate mechanisms of genotoxicity

The response to stress triggers transcriptional activation of genes involved in cell survival and/or cell death. Thus, the monitoring of gene expression levels in large gene sets or whole genomes in response to various agents (toxicogenomics) has been proposed as a tool for investigating mechanisms of toxicity. Although standard in vitro genetic toxicity testing provides relatively simple and accurate hazard detection, interpretation of positive findings, i.e., in vitro chromosome aberrations, in terms of relevant risk to humans is difficult, due to the limited insight into the underlying mechanisms. Therefore, the development of experimental approaches capable of differentiating a wide range of genotoxic mechanisms is expected to significantly improve risk assessment. The goal of this review is to summarize current developments in toxicogenomic analysis of genotoxic stress, and to provide a perspective on the application of gene expression profile analysis in genetic toxicology.

A Toxicogenomics Approach to Identify New Plausible Epigenetic Mechanisms of Ochratoxin A Carcinogenicity in Rat

Ochratoxin A (OTA) is a mycotoxin occurring naturally in a wide range of food commodities. In animals, it has been shown to cause a variety of adverse effects, nephrocarcinogenicity being the most prominent. Because of its high toxic potency and the continuous exposure of the human population, OTA has raised public health concerns. There is significant debate on how to use the rat carcinogenicity data to assess the potential risk to humans. In this context, the question of the mechanism of action of OTA appears of key importance and was studied through the application of a toxicogenomics approach. Male Fischer rats were fed OTA for up to 2 years. Renal tumors were discovered during the last 6 months of the study. The total tumor incidence reached 25% at the end of the study. Gene expression profile was analyzed in groups of animals taken in intervals from 7 days to 12 months. Tissue-specific responses were observed in kidney versus liver. For selected genes, microarray data were confirmed at both mRNA and protein levels. In kidney, several genes known as markers of kidney injury and cell regeneration were significantly modulated by OTA. The expression of genes known to be involved in DNA synthesis and repair, or genes induced as a result of DNA damage, was only marginally modulated. Very little or no effect was found amongst genes associated with apoptosis. Alterations of gene expression indicating effects on calcium homeostasis and a disruption of pathways regulated by the transcription factors hepatocyte nuclear factor 4 alpha (HNF4alpha) and nuclear factor-erythroid 2-related factor 2 (Nrf2) were observed in the kidney but not in the liver. Previous data have suggested that a reduction in HNF4alpha may be associated with nephrocarcinogenicity. Many Nrf2-regulated genes are involved in chemical detoxication and antioxidant defense. The depletion of these genes is likely to impair the defense potential of the cells, resulting in chronic elevation of oxidative stress in the kidney. The inhibition of defense mechanism appears as a highly plausible new mechanism, which could contribute to OTA carcinogenicity.

Differentiating mechanisms of toxicity using global gene expression analysis in Saccharomyces cerevisiae

Genotoxic stress triggers a variety of biological responses including the transcriptional activation of genes regulating DNA repair, cell survival and cell death. Genomic approaches, which monitor gene expressions across large numbers of genes, can serve as a powerful tool for exploring mechanisms of toxicity. Here, using five different agents, we investigated whether the analysis of genome-wide expression profiles in Saccharomyces cerevisiae could provide insights into mechanisms of genotoxicity versus cytotoxicity. To differentiate the genotoxic stress-associated expression signatures from that of a general cytotoxic stress, we compared gene expression profiles following the treatment with DNA-reactive (cisplatin, MMS, bleomycin) and DNA non-reactive (ethanol and sodium chloride) compounds. Although each of the tested chemicals produced a distinct gene expression profile, we were able to identify a gene expression signature consisting of a relatively small number of biologically relevant genes capable of differentiating genotoxic and cytotoxic stress. The gene set includes such upregulated genes as HUG1, ECM4 and previously uncharacterized gene, YLR297W in the genotoxic and GAP1, CGR1 in the cytotoxic group. Our results indicate the potential of gene expression profile analysis for elucidating mechanism of action of genotoxic agents.

Comparison of supervised clustering methods to discriminate genotoxic from non-genotoxic carcinogens by gene expression profiling

Prediction of the toxic properties of chemicals based on modulation of gene expression profiles in exposed cells or animals is one of the major applications of toxicogenomics. Previously, we demonstrated that by Pearson correlation analysis of gene expression profiles from treated HepG2 cells it is possible to correctly discriminate and predict genotoxic from non-genotoxic carcinogens. Since to date many different supervised clustering methods for discrimination and prediction tests are available, we investigated whether application of the methods provided by the Whitehead Institute and Stanford University improved our initial prediction. Four different supervised clustering methods were applied for this comparison, namely Pearson correlation analysis (Pearson), nearest shrunken centroids analysis (NSC), K-nearest neighbour analysis (KNN) and Weighted voting (WV). For each supervised clustering method, three different approaches were followed: (1) using all the data points for all treatments, (2) exclusion of the samples with marginally affected gene expression profiles and (3) filtering out the gene expression signals that were hardly altered. On the complete data set, NSC, KNN and WV outperformed the Pearson test, but on the reduced data sets no clear difference was observed. Exclusion of samples with marginally affected profiles improved the prediction by all methods. For the various prediction models, gene sets of different compositions were selected; in these 27 genes appeared three times or more. These 27 genes are involved in many different biological processes and molecular functions, such as apoptosis, cell cycle control, regulation of transcription, and transporter activity, many of them related to the carcinogenic process. One gene, BAX, was selected in all 10 models, while ZFP36 was selected in 9, and AHR, MT1E and TTR in 8. Summarising, this study demonstrates that several supervised clustering methods can be used to discriminate certain genotoxic from non-genotoxic carcinogens by gene expression profiling in vitro in HepG2 cells. None of the methods clearly outperforms the others.

Effects of chlorpromazine with and without UV irradiation on gene expression of HepG2 cells

Damage to DNA can trigger a variety of stress-related signals that alter the expression of genes associated with numerous biological pathways. In this study, we have used a cDNA microarray representing 1089 genes related to DNA damage and repair, cell cycle, transcription, metabolism and other toxicologically important cell functions to identify genes regulated in response to DNA damage in HepG2 cells induced by UV-activated chlorpromazine (CPZ). CPZ itself is not genotoxic but, upon UV irradiation with a non-genotoxic dose in the UVA range, it produces reactive free radical intermediates with DNA damaging properties. Genotoxicity in HepG2 cells was assessed concomitantly to gene expression profiling using the Comet assay. Kinetic studies were performed at a non-cytotoxic but clearly photogenotoxic concentration of CPZ (1.25 microg/ml) to characterize gene expression profiles at four different time points (3, 7, 15, 23 h) post short-term treatment. The results obtained from repeated experiments display a time-dependent expression pattern of up-regulated and repressed genes with distinct peaks in the number of differentially expressed genes at the 7 and 23 h time points. Most of the genes with altered expression belonged to the functional categories of cell cycle regulation and cell proliferation. A comparison with published expression profiles established in response to other genotoxic compounds showed low levels of concordance, which is most likely caused by the fact that extremely different testing conditions were used.