Expression profiling in toxicology--potentials and limitations

Interpretable molecular encodings and representations for machine learning tasks

Molecular encodings and their usage in machine learning models have demonstrated significant breakthroughs in biomedical applications, particularly in the classification of peptides and proteins. To this end, we propose a new encoding method: Interpretable Carbon-based Array of Neighborhoods (iCAN). Designed to address machine learning models' need for more structured and less flexible input, it captures the neighborhoods of carbon atoms in a counting array and improves the utility of the resulting encodings for machine learning models. The iCAN method provides interpretable molecular encodings and representations, enabling the comparison of molecular neighborhoods, identification of repeating patterns, and visualization of relevance heat maps for a given data set. When reproducing a large biomedical peptide classification study, it outperforms its predecessor encoding. When extended to proteins, it outperforms a lead structure-based encoding on 71% of the data sets. Our method offers interpretable encodings that can be applied to all organic molecules, including exotic amino acids, cyclic peptides, and larger proteins, making it highly versatile across various domains and data sets. This work establishes a promising new direction for machine learning in peptide and protein classification in biomedicine and healthcare, potentially accelerating advances in drug discovery and disease diagnosis.

Myocardial toxicity induced by silica nanoparticles in a transcriptome profile

The deleterious effects of silica nanoparticles (SiNPs) on human health and the ecological system have gradually gained attention owing to their heavy annual output and extensive global flux. The updated epidemiological or experimental investigations have demonstrated the potential myocardial toxicity triggered by SiNPs, but the underlying mechanisms and long-lasting cardiac effects are still poorly understood. Here, a rat model of sub-chronic respiratory exposure to SiNPs was conducted, and the histopathological analysis and ultrastructural investigation of heart tissues were carried out. More importantly, a comprehensive analysis of whole-genome transcription was utilized in rat heart to uncover key biological and cellular mechanisms triggered by SiNPs. The widening of myocardial space and partial fiber rupture were clearly manifested in rat heart after prolonged SiNPs exposure, particularly accompanied by mitochondrial swelling and cristae rupture. With the aid of Affymetrix GeneChips, 3153 differentially expressed genes (DEGs) were identified after SiNPs exposure, including 1916 down- and 1237 up-regulated genes. GO and KEGG analysis illustrated many important biological processes and pathways perturbed by SiNPs, mainly specializing in cellular stress, energy metabolism, actin filament dynamics and immune response. Signal-net analysis revealed that Prkaca (PKA) plays a core role in the cardiac toxification process of prolonged exposure of SiNPs to rats. Furthermore, qRT-PCR verified that PKA-mediated calcium signaling is probably responsible for SiNPs-induced cardiac injury. Conclusively, our study revealed that SiNPs caused myocardial injury, and particularly, provided transcriptomic insight into the role of PKA-calcium signaling triggered by SiNPs, which would facilitate SiNPs-based nanosafety assessment and biomedicine development.

Impact of pyrethroids and organochlorine pesticides residue on IGF-1 and CYP1A genes expression and muscle protein patterns of cultured Mugil capito

This study aimed to assess the levels of pyrethroids and organochlorine residues in the tissues of cultured Mugil capito and in water samples obtained from three different sites (Al-Hamol, Al-Riad and Sidi Salem; referred to as Area 1, Area 2, and Area 3, respectively) in the Delta region, Egypt. The study also assessed the biochemical markers in exposed mullet and evaluated the impact of these residues on the expression of insulin-like growth factor 1 (IGF-1) in muscle and cytochrome P4501A (CYP1A) in liver tissues using qRT-PCR and SDS-PAGE methods. The results revealed that pesticide residue levels in the water were variable, but were lower than detected levels in fish. Significant (P < 0.05) differences were found across the three study areas in terms of serum ALT, but the serum AST level was not significantly (P > 0.05) elevated in all study regions. Serum creatinine and urea levels were significantly (P < 0.05) elevated in area 3. Furthermore, glutathione and malondialdehyde concentrations significantly increased (P < 0.05) in liver tissues in area 3. Using the qRT-PCR technique, the results revealed that the expression level of IGF-1 was most significant in area 3, while the expression level of CYP1A was most significant in area 1. The protein profile showed some differences in band numbers and molecular weights of protein bands across different regions. Overall, the alteration in biochemical parameters revealed pesticide interference with the metabolic processes of fish. Furthermore, the pesticide pollution had an effect on the expression of IGF-1 and CYP1A genes and led to changes in the protein profile. Therefore, these markers can be used to monitor fish distress following exposure to the pollutant.

Gene Expression Data Based Deep Learning Model for Accurate Prediction of Drug-Induced Liver Injury in Advance

Drug-induced liver injury (DILI), one of the most common adverse effects, leads to drug development failure or withdrawal from the market in most cases, showing an emerging challenge that is to accurately predict DILI in the early stage. Recently, the vast amount of gene expression data provides us valuable information for distinguishing DILI on a genomic scale. Moreover, the deep learning algorithm is a powerful strategy to automatically learn important features from raw and noisy data and shows great success in the field of medical diagnosis. In this study, a gene expression data based deep learning model was developed to predict DILI in advance by using gene expression data associated with DILI collected from ArrayExpress and then optimized by feature gene selection and parameters optimization. In addition, the previous machine learning algorithm support vector machine (SVM) was also used to construct another prediction model based on the same data sets, comparing the model performance with the optimal DL model. Finally, the evaluation test using 198 randomly selected samples showed that the optimal DL model achieved 97.1% accuracy, 97.4% sensitivity, 96.8% specificity, 0.942 matthews correlation coefficient, and 0.989 area under the ROC curve, while the performance of SVM model only reached 88.9% accuracy, 78.8% sensitivity, 99.0% specificity, 0.794 matthews correlation coefficient, and 0.901 area under the ROC curve. Furthermore, external data sets verification and animal experiments were conducted to assess the optimal DL model performance. Finally, the predicted results of the optimal DL model were almost consistent with experiment results. These results indicated that our gene expression data based deep learning model could systematically and accurately predict DILI in advance. It could be a useful tool to provide safety information for drug discovery and clinical rational drug use in early stage and become an important part of drug safety assessment.

A computational toxicogenomics approach identifies a list of highly hepatotoxic compounds from a large microarray database

The liver and the kidney are the most common targets of chemical toxicity, due to their major metabolic and excretory functions. However, since the liver is directly involved in biotransformation, compounds in many currently and normally used drugs could affect it adversely. Most chemical compounds are already labeled according to FDA-approved labels using DILI-concern scale. Drug Induced Liver Injury (DILI) scale refers to an adverse drug reaction. Many compounds do not exhibit hepatotoxicity at early stages of development, so it is important to detect anomalies at gene expression level that could predict adverse reactions in later stages. In this study, a large collection of microarray data is used to investigate gene expression changes associated with hepatotoxicity. Using TG-GATEs a large-scale toxicogenomics database, we present a computational strategy to classify compounds by toxicity levels in human and animal models through patterns of gene expression. We combined machine learning algorithms with time series analysis to identify genes capable of classifying compounds by FDA-approved labeling as DILI-concern toxic. The goal is to define gene expression profiles capable of distinguishing the different subtypes of hepatotoxicity. The study illustrates that expression profiling can be used to classify compounds according to different hepatotoxic levels; to label those that are currently labeled as undertemined; and to determine if at the molecular level, animal models are a good proxy to predict hepatotoxicity in humans.

Transcriptional response of stress-regulated genes to cadmium exposure in the cockle Cerastoderma glaucum from the gulf of Gabès area (Tunisia)

This study investigates cadmium effects on key messenger RNA (mRNA) expression (MT, MnSOD, CuZnSOD, CAT, ABCB1, HSP70, and CO1) by qPCR in the cockle Cerastoderma glaucum after chronic exposure to two high but environmentally relevant concentrations of CdCl2 (50 μg/L and 5 mg/L) for 12 h to 18 days. Cd accumulation measured in cockles' tissues is significantly higher in both treatment conditions compared to controls and in a dose-dependent manner. Stress on stress tests performed at different times of the experiment clearly demonstrated that exposure to both concentrations of Cd significantly affects cockle survival time in air. Important changes in gene transcription were also highlighted. In particular, MT, HSP70, CAT, and CuZnSOD seem to be relevant biomarkers of Cd exposure because (1) their mRNA levels increase upon exposure and (2) they are highly correlated to Cd accumulation in tissues. Results may be useful for control strategies and for the use of cockles as sentinel organisms.

Proteomics in pharmaceutical research and development

In the 20 years since its inception, the evolution of proteomics in pharmaceutical industry has mirrored the developments within academia and indeed other industries. From initial enthusiasm and subsequent disappointment in global protein expression profiling, pharma research saw the biggest impact when relating to more focused approaches, such as those exploring the interaction between proteins and drugs. Nowadays, proteomics technologies have been integrated in many areas of pharmaceutical R&D, ranging from the analysis of therapeutic proteins to the monitoring of clinical trials. Here, we review the development of proteomics in the drug discovery process, placing it in a historical context as well as reviewing the current status in light of the contributions to this special issue, which reflect some of the diverse demands of the drug and biomarker pipelines.

Investigation of quinocetone-induced mitochondrial damage and apoptosis in HepG2 cells and compared with its metabolites

Quinocetone (QCT) has been widely used as an animal growth promoter in China. However, amounts of available data indicated that QCT probably had potential toxicity. The present study was aimed to investigate the genotoxicity, mitochondrial damage and apoptosis in HepG2 cells for QCT and its metabolites, DQCT and MQCA. QCT has seriously cytotoxic to HepG2 cells. The cell viability test and cytokinesis-block micronucleus test showed that the micronucleus frequency of cells treated with QCT has increased significantly, compared with DQCT and MQCA. With increasing of QCT concentrations, the genomic template stability and mitochondrial damage of HepG2 cells were aggravated. QCT-induced apoptosis in HepG2 cells were also observed. Data of caspase activities in measurement and real-time RT-PCR possibly suggested both of the mitochondria-dependent and mitochondria-independent pathways participated in the HepG2 cells apoptosis. However, all the results suggested that DQCT and MQCA showed only a little cytotoxic to HepG2 cells. In a word, QCT had toxic effects on HepG2 cells and resulted in the mitochondria-dependent and mitochondria-independent pathways of apoptosis, but the intermediate metabolites of QCT (DQCT and MQCA) were not.

Can systems toxicology identify common biomarkers of non-genotoxic carcinogenesis?

For the rapid development of safe, efficacious chemicals it is important that any potential liabilities are identified as early as possible in the discovery/development pipeline. Once identified it is then possible to make rational decisions on whether to progress a chemical and/or series further; one such liability is chemical carcinogenesis, a highly undesirable characteristic in a novel chemical entity. Chemical carcinogens may be roughly divided into two classes, those that elicit their actions through direct damage to DNA (genotoxic carcinogens) and those that cause carcinogenesis through mechanisms that involve direct damage of the DNA by the agent (non-genotoxic carcinogens). Whereas the former group can be identified by in vitro screens to a good degree of accuracy, the latter group are far more problematic due to their diverse modes of action. This review will focus on the latter class of chemical carcinogens, examining how modern '-omic' technologies have begun to identify signatures that may represent sensitive, early markers for these processes. In addition to their use in signature generation the role of -omic level approaches to delineating molecular mechanisms of action will also be discussed.

Key developments in endocrine disrupter research and human health

Environmental etiologies involving exposures to chemicals that mimic endogenous hormones are proposed for a number of adverse human health effects, including infertility, abnormal prenatal and childhood development, and reproductive cancers (National Research Council, 1999; World Health Organization, 2002). Endocrine disrupters represent a significant area of environmental research with important implications for human health. This article provides an overview of some of the key developments in this field that may enhance our ability to assess the human health risks posed by exposure to endocrine disrupters. Advances in methodologies of hazard identification (toxicogenomics, transcriptomics, proteomics, metabolomics, bioinformatics) are discussed, as well as epigenetics and emerging biological endpoints.

Estrogenic Effects in the Immature Rat Uterus after Dietary Exposure to Ethinylestradiol and Zearalenone Using a Systems Biology Approach

Residues of illegally used hormones are regularly detected in animal products, feed, or cocktails recovered at farms. In order to better understand the effects of dietary exposure to ethinyl estradiol (EE2, 0.03-1 microg/kg body weight [bw]) and zearalenone (ZEA, 0.03-1 mg/kg bw), an immature rat uterotrophic assay was performed and effects were studied at morphological, histological, and gene expression levels. Ligand-mediated coregulator recruitment by estrogen receptor alpha (ERalpha) was studied in vitro. Uterine weight and epithelial cell height were increased dose dependently after a 3-day oral exposure of rats to the highest tested doses of EE2 or ZEA, respectively. At low doses 0.03 microg/kg EE2 and 0.1 mg/kg ZEA, edema, and vacuolization could already be observed in some animals. Exposure to 1 mg/kg ZEA resulted in severe damage of the uterine epithelial layer. Our study suggests similar coregulator recruitment and gene expression patterns for the two estrogenic compounds. Main regulated pathways were remodeling of extracellular matrix, alternative complement activation, cell proliferation, and estrogen-mediated calcium signaling. The level of regulation differed between EE2 and ZEA, attributing a much lower estrogenic potency to ZEA than to EE2. A major difference was their ability to recruit coregulator inhibitor of kappa B beta and induce expression of the matrix metalloproteinase 7 gene (381.4- and 6.9-fold upregulation by EE2 and ZEA, respectively), which plays an important role in the maintenance of the integrity of the epithelial layer of the uterus during proliferation and growth. This observation may explain the observed differences at the histological level.

Toxicogenomics: a pivotal piece in the puzzle of toxicological research

Toxicogenomics, resulting from the merge of conventional toxicology with functional genomics, being the scientific field studying the complex interactions between the cellular genome, toxic agents in the environment, organ dysfunction and disease state. When an organism is exposed to a toxic agent the cells respond by altering the pattern of gene expression. Genes are transcribed into mRNA, which in turn is translated into proteins that serve in a variety of cellular functions. Toxicogenomics through microarray technology, offers large-scale detection and quantification of mRNA transcripts, related to alterations in mRNA stability or gene regulation. This may prove advantageous in toxicological research. In the present review, the applications of toxicogenomics, especially to mechanistic and predictive toxicology are reported. The limitations arising from the use of this technology are also discussed. Additionally, a brief report of other approaches, using other -omic technologies (proteomics and metabonomics) that overcome limitations and give global information related to toxicity, is included.

Molecular identification and expression of differentially regulated genes of the European flounder, Platichthys flesus, submitted to pesticide exposure

Effects of pesticide exposure on the European flounder, Platichthys flesus, were investigated using a suppression subtractive hybridization method (SSH) to identify up- and down-regulated genes after a 30-day exposure to herbicides (a cocktail of atrazine, diuron, and isoproturon, and a single herbicide, glyphosate). A total of 256 expressed gene sequences were identified as having the potential for being differentially expressed, of which 116 could be identified by homology with databased sequences. The metabolic functions with which they are associated include energy production, general metabolism, signaling, transport, immune system, and structure. Expression of 14 of these genes was analyzed in liver, muscle, and gills by reverse transcriptase-polymerase chain reaction (RT-PCR) under experimental conditions (0, 15, and 30 days of exposure) and under field conditions (sampling in two estuaries displaying different levels of pesticide contamination). This study provides a first basis for studying the response of fish to pesticide exposure and allows the characterization of new potential genetic markers of pesticide contamination in the field.

Analysis of drug-induced effect patterns to link structure and side effects of medicines

The high failure rate of experimental medicines in clinical trials accentuates inefficiencies of current drug discovery processes caused by a lack of tools for translating the information exchange between protein and organ system networks. Recently, we reported that biological activity spectra (biospectra), derived from in vitro protein binding assays, provide a mechanism for assessing a molecule's capacity to modulate the function of protein-network components. Herein we describe the translation of adverse effect data derived from 1,045 prescription drug labels into effect spectra and show their utility for diagnosing drug-induced effects of medicines. In addition, notwithstanding the limitation imposed by the quality of drug label information, we show that biospectrum analysis, in concert with effect spectrum analysis, provides an alignment between preclinical and clinical drug-induced effects. The identification of this alignment provides a mechanism for forecasting clinical effect profiles of medicines.

Stable isotope methods for high-precision proteomics

Stable isotope tagging methods provide a useful means of determining the relative expression level of individual proteins between samples in a mass spectrometer with high precision (coefficients of variation less than 10%). Because two or more samples tagged with different numbers of stable isotopes can be mixed before any processing steps, sample-to-sample recovery differences are eliminated. Mass spectrometry also allows post-translational modifications, splice variations and mutations (often unnoticed in immunoassays) to be detected and identified, increasing the clinical relevance of the assay and avoiding the issues of non-specific binding and cross-reactivity observed in immunoassays. Several stable isotope tagging methods are available for use in proteomics research. We discuss the advantages and disadvantages of each technique with respect to biomarker discovery, target validation, efficacy and toxicology screening and clinical diagnostic applications.

Molecular approaches to the identification of biomarkers of exposure and effect—report of an expert meeting organized by COST Action B15

In the past, the term biomarker has been used with several meanings when used in human and environmental toxicology as compared to pharmaceutical development. However, with the advent of molecular approaches and their application in the field of drug development and toxicology, the concept of biomarkers has to be newly defined. In the meeting, the experts found consent in defining the term and described the application of biomarkers in toxicology, drug development and clinical diagnostics. Molecular approaches to biomarker identification and selection lead to a large amount of data. Hence, the statistical analysis is challenging and special statistical problems have to be solved in biomarker characterization, of particular interest are attempts aiming at class discovery and prediction. Reliability and biological relevance are to be demonstrated for biomarkers of exposure and effect which is also true for biomarkers of susceptibility. It is envisaged that the application of biomarkers will expand from current use in pre-clinical toxicology to the risk characterization and risk assessment of chemicals and from early clinical phases of drug development to later phases and even into daily clinical use in diagnostics and disease classification.

Toxicogenomics in predictive toxicology in drug development

The goal of toxicology is the assessment of possible risk to man. An emerging technology with the potential to have a major impact on risk assessment is toxicogenomics. In this review, we provide an overview of the many possibilities for toxicogenomics including technology platforms, data interpretation, and regulatory perspective and we give examples of toxicogenomics investigations. Toxicogenomics is a powerful tool for compound classification, for mechanistic studies, and for the detection of toxicity markers. Thus, toxicogenomics helps in the extrapolation of findings across species and increases predictability. Biomarkers are valuable in the evaluation of compounds at earlier development phases, improving clinical candidate selection. Caution regarding the interpretation of the results is still necessary. Nevertheless, toxicogenomics will accelerate preclinical safety assessments and improve the prediction of toxic liabilities, as well as of potential risk accumulation for drug-drug or drug-disease interactions.

Nasal lavage fluid and proteomics as means to identify the effects of the irritating epoxy chemical dimethylbenzylamine

The aims of this study were to describe the changes in the nasal lavage fluid (NLF) protein pattern after exposure to the irritating epoxy chemical dimethylbenzylamine (DMBA) and to identify the affected proteins using a proteomic approach. The protein patterns of NLF from six healthy subjects and eight epoxy workers with airway irritation were analysed using two-dimensional gel electrophoresis (2-DE) before and after exposure to 100 microg m(-3) DMBA for 2 h in an exposure chamber. NLF proteins were identified by (i) comparison with a 2-DE NLF reference database; (ii) N-terminal amino acid sequencing; and (iii) mass spectrometry. In NLF from healthy subjects, the levels of immunoglobulin A increased and the levels of Clara cell protein 16 (CC16) decreased after chamber exposure, while in NLF from epoxy workers, alpha(2)-macroglobulin and caeruloplasmin increased. Two previously unidentified proteins decreased in NLF from epoxy workers after exposure; these were identified as statherin and calgranulin B. In addition, the subjects who developed high counts of eosinophils in their nasal mucosa after chamber exposure had significantly lower levels of immunoglobulin-binding factor (IgBF) before exposure than subjects with low eosinophil infiltration. These results show that short-term exposure to DMBA causes distinct changes in NLF proteins. Moreover, three proteins that have previously not been associated with upper airway irritation were identified: statherin, calgranulin B and IgBF. Further studies are needed to investigate whether these proteins may be used as biomarkers of airway irritation and to give new insight into the ways in which occupational exposure to irritants causes inflammation of the airways.

Human pharmacogenomics: the development of a science

Until about 50 years ago, the altering of a normal drug effect by a genetic deficiency was only rarely observed. Here, my discovery of the genetic variant of butyrylcholinesterase affecting succinylcholine action is described in some detail. Such discoveries led to the combination of the two older sciences, genetics and pharmacology, thereby forming pharmacogenetics. After the discovery of similar examples in the years that followed, pharmacogenetics expanded on the basis of two discoveries. First, the common occurrence of interethnic differences in drug response and, secondly, the fact that most pharmacological differences were multigenic. New methodologies brought a transition to pharmacogenomics; this included detection of clinically important genetic variants and has uncovered potentially new drug targets. The arrival of personalised medicine--where a patient's genes determine the choice of drug to be administered--can be hoped to gradually improve drug safety and efficacy. Efforts to reach this level of perfection are, however, dogged by uncertainties.

Proteome analysis by bio-active ceramic water in rat liver: contribution to antioxidant enzyme expression, SOD I

The purpose of this study was to investigate the protective effect of bio-active ceramic water on rat liver. Male Wistar rats were divided into 4 groups of 15 animals each. Groups 1 and 2 were fed bio-active ceramic water and tap water for 4 months, respectively. Groups 3 and 4 were treated with the same condition for 12 months. The changes of protein expression of these four groups were investigated using two-dimensional gel electrophoresis and matrix-assisted laser desorption/ionization-time of flight mass spectrometry. Eleven proteins were significantly up-regulated in bio-active ceramic water treated rat liver including aldehyde dehydrogenase I and II, albumin, fructose-1,6-bisphosphatase, and superoxide dismutase I (SOD I). The most highly expressed protein, SOD I with up-regulated enzyme activity, was confirmed by immunoblots as a major antioxidant capable of detoxifying normally generated reactive oxygen species. These data suggest that modified protein expression of the liver contributes to enhance liver function.

Differential gene expression in anthracene-exposed mummichogs (Fundulus heteroclitus)

The polycyclic aromatic hydrocarbon (PAH) anthracene is present in many estuarine systems at concentrations believed to cause sublethal adverse effects, although its exact mode of toxicity remains unclear. Knowledge of the induction or suppression of specific genes as a result of exposure may be useful in explaining these effects. We have generated a fingerprint of anthracene exposure using the mummichog (Fundulus heteroclitus), a non-migratory estuarine fish species. The fish were exposed in 7-day static renewal tests to environmentally relevant concentrations of 0, 27, 50, and 80 microg/l of anthracene. Total RNA was extracted from the livers and differential display reverse transcription polymerase chain reaction (DD RT-PCR) was used to recover 26 differentially expressed cDNA fragments. These cDNAs were isolated, sequenced, and compared to sequences of known genes in order to identify possible physiological consequences of exposure to anthracene. We then constructed macroarrays using these fragments and probed them with RNA from both anthracene-exposed fish and fish from a known PAH-impacted site. Three genes appear to be good indicators of exposure to anthracene in the range of concentrations tested, which included CYP2N2 and two expressed sequence tags (ESTs) termed 15C1 and 18C2. The expression of nine genes was altered in fish collected from a site with multiple PAHs. Band 15C1 and CYP2N2 again showed statistically significant upregulation in the field-caught fish, while a trypsin precursor and fatty acid-binding protein (FABP) all showed similar trends in induction as the laboratory-exposed fish. Further insight into the mechanism of toxicity of contaminants will be gained by the ability to identify and use differentially expressed genes as markers of exposure and effects.

Proteomics in biomarker discovery and drug development

Proteomics is a research field aiming to characterize molecular and cellular dynamics in protein expression and function on a global level. The introduction of proteomics has been greatly broadening our view and accelerating our path in various medical researches. The most significant advantage of proteomics is its ability to examine a whole proteome or sub-proteome in a single experiment so that the protein alterations corresponding to a pathological or biochemical condition at a given time can be considered in an integrated way. Proteomic technology has been extensively used to tackle a wide variety of medical subjects including biomarker discovery and drug development. By complement with other new technique advances in genomics and bioinformatics, proteomics has a great potential to make considerable contribution to biomarker identification and to revolutionize drug development process. This article provides a brief overview of the proteomic technologies and their application in biomarker discovery and drug development.

Genomics and the search for novel biomarkers in toxicology

The advent of 'genomics' technology, in particular transcript profiling, has already had a measurable impact on the drug discovery process in the areas of target identification and validation. This review is concerned with the potential application of this technology to toxicology and drug safety assessment, with particular emphasis on biomarker discovery and characterization. An advantage (or possibly a drawback!) of transcript profiling is that candidate biomarkers of toxicity can be speedily identified, with the caveat that a significant amount of subsequent experimental and bioinformatic effort needs to be expended in order to evaluate and validate them. Attention is also drawn to the critical need for robust experimental design with studies of this type and to issues associated with the analysis of large data sets. In summary, while genomics technology undoubtedly offers much that can assist drug safety assessment, its potential has yet to be realized fully in this area. However, a large amount of resource continues to be applied to 'toxicogenomics'. Tangible benefits, in terms of new biomarkers of toxicity and reduced numbers of adverse drug effects, remain realistic objectives.

Food mutagens

Several lines of evidence indicate that diet and dietary behaviors can contribute to human cancer risk. One way that this occurs is through the ingestion of food mutagens. Sporadic cancers result from a gene-environment interactions where the environment includes endogenous and exogenous exposures. In this article, we define environment as dietary exposures in the context of gene-environment interactions. Food mutagens cause different types of DNA damage: nucleotide alterations and gross chromosomal aberrations. Most mutagens begin their action at the DNA level by forming carcinogen-DNA adducts, which result from the covalent binding of a carcinogen or part of a carcinogen to a nucleotide. However the effect of food mutagens in carcinogenesis can be modified by heritable traits, namely, low-penetrant genes that affect mutagen exposure of DNA through metabolic activation and detoxification or cellular responses to DNA damage through DNA repair mechanisms or cell death. There are some clearly identified (e.g., aflatoxin) and suspected (e.g., N-nitrosamines, polycyclic aromatic hydrocarbons or heterocyclic amines) food mutagens. The target organs for these agents are numerous, but there is target-organ specificity for each. Mutagenesis however is not the only pathway that links dietary exposures and cancers. There is growing evidence that epigenetic factors, including changes in the DNA methylation pattern, are causing cancer and can be modified by dietary components. Also DNA damage may be indirect by triggering oxidative DNA damage. When considering the human diet, it should be recognized that foods contain both mutagens and components that decrease cancer risk such as antioxidants. Thus nutritionally related cancers ultimately develop from an imbalance of carcinogenesis and anticarcinogenesis. The best way to assess nutritional risks is through biomarkers, but there is no single biomarker that has been sufficiently validated. Although panels of biomarkers would be the most appropriate, their use as a reflection of target-organ risk remains to be determined. Also even when new biomarkers are developed, their application in target organs is problematic because tissues are not readily available. For now most biomarkers are used in surrogate tissues (e.g., blood, urine, oral cavity cells) that presumably reflect biological effects in target organs. This article reviews the role of food mutagens in mutagenesis and carcinogenesis and how their effects are modified by heritable traits and discusses how to identify and evaluate the effects of food mutagens.

Discriminating two classes of toxicants through expression analysis of HepG2 cells with DNA arrays

Microarray technology provides a rapid and cost-effective method to associate specific cellular responses with unique gene expression patterns. If characteristic expression patterns of a small number of genes could be associated with drug toxicity, this association may be used for toxicity prediction, and thereby to reduce the need for traditional toxicity testing. To test this hypothesis, we have designed an array composed of 92 known human genes of toxicological interest (including seven housekeeping genes) and eight bacterial controls. HepG2 cells were treated with either ethanol or one of two quinone containing anticancer drugs, mitomycin C or doxorubicin. RNA was isolated from treated and untreated cells, differentially labeled with fluorescent dyes, and then hybridized to the array. Our results show that the expression patterns induced by ethanol and the anticancer drugs are different. Both of the anticancer drugs, but not ethanol had a differential effect on the regulation of several genes, including CYP4F2/3, CYP3A3, TNFRSF6 and CHES1, demonstrating that the two drugs might function through a similar mechanism, which differs from that of ethanol. These results suggest that microarray-based expression analysis may offer a rapid and efficient means for assessing drug toxicity.

Proteomic analysis of the rat liver

Rat is a useful, widely used animal model for biological and toxicity studies. We analyzed total and cytosolic rat liver proteins by applying proteomics technologies. The proteins were separated by two-dimensional electrophoresis employing broad and narrow range immobilized pH gradient strips, followed by MALDI-MS analysis of the tryptic digests. Two hundred and seventy-three different gene products were identified, of which approximately 60% were enzymes with a broad spectrum of catalytic activities. Most of the identified proteins were detected in other rat protein samples as well, which were analyzed in our laboratory. Fifteen gene products were detected for the first time. These were represented by one spot each, whereas most of the frequently detected proteins were represented by multiple spots. In average, approximately five to 10 spots corresponded to one gene product. The database includes a large number of proteins known to be involved in toxicology-relevant pathways and may be useful in toxicity prediction studies.

DNA microarrays and toxicogenomics: applications for ecotoxicology?

Toxicogenomics attempts to define how the regulation and expression of genes mediate the toxicological effects associated with exposure to a chemical. DNA microarrays are rapidly becoming one of the tools of choice for large-scale toxicogenomic studies. An approach in modern toxicogenomics has been to classify toxicity based on gene transcriptional patterns; comparing the transcriptional responses of a chemical with unknown toxicity to those for which the transcriptional profiles and toxicological endpoints have been well characterized. Recent evidence suggests that gene expression microarrays may be instrumental in defining mechanisms of action of toxicants. However, several assumptions are inherent to a toxicogenomic-based approach in toxicology, many of which remain to be validated. Gene expression profiling using DNA microarrays represents a snapshot of the gene transcriptional responses occurring at a particular time and within a particular tissue. Toxicity, on the other hand, represents a continuum of possible effects governed by both temporal and spatial factors that are inextricably contingent upon the exposure conditions. The perceived toxicological properties of any chemical are dependent on the route, dose, and duration of the exposure, and as such, gene expression patterns are also subject to these variables. Correct interpretation of DNA microarray data for the assessment of the toxicological properties of chemicals will require that temporal and spatial gene expression profiles be accounted for. These considerations are further compounded in ecotoxicological studies, during which altered gene expression patterns induced from exposure to an anthropogenic substance must be discernible over and above the complex effects that phenotypic, genotypic, and environmental variables have on gene expression. To this end, the greatest utility of DNA microarrays in the field of ecotoxicology may be in predicting the toxicological modes of action of anthropogenic substances on host physiology, particularly in non-model organisms. Predictable and accurate assessment of the impacts of a chemical substance in ecotoxicology will require that classical toxicological endpoints be used to validate any effects predicted based on gene expression profiling. Validated expression profiling may subsequently find utility in ecotoxicological-based computer simulation models, such as the Biotic Ligand Model (BLM), in which gene expression information may be integrated with geochemical, pharmacokinetic, and physiological data to accurately assess and predict toxicity of metals to aquatic organisms.

Developing sustainable studies on environmental health

Toxicogenetics, toxicogenomics and proteomics are providing new biomarkers for use in human studies. These, coupled with the more traditional biological responses currently in use, provide a vast armamentarium for assessing exposures, effects and susceptibility factors relating to environmental pollutants. Biomarker availability, however, does not automatically translate to usefulness for studies directed at human health improvement. Transitional studies bridging the gap between laboratory and field, with the biomarker as dependent variable, are required for validation for intended applications. Prospective or cross-sectional studies are usually optimal for validating biomarkers of exposure, where the biomarker response follows the event (exposure) and the entire study population is affected. Biomarkers of effect, where the event (outcome) follows the biomarker response and may be infrequent in the study population, present more complex problems. Identification of genotypes as susceptibility factors requires that additional issues be addressed. Despite these difficulties, new data are emerging from transitional investigations that are providing validated biomarkers for sustainable human studies. It is now possible to envision schemes for integrating the results of molecular epidemiological investigations into the general toxicological evaluations of environmental agents. These will allow intermediate endpoints to be used for making realistic human health assessments and for elucidating pathogenic mechanisms that identify targets for intervention, all with the goal of preventing environmentally mediated human disease. Finally, select biomarker responses that predict the likelihood of disease occurrence will find application in the interpretation of individual medical diagnostic tests, with the goal of improving cancer detection and management.

Two-dimensional database of mouse liver proteins. An update

We updated the two-dimensional protein database for mouse liver. Microsomal and cytosolic fractions of the liver proteins from male mice were separated by two-dimensional electrophoresis. The proteins were identified by Matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) on the basis of peptide mass fingerprinting, following in-gel digestion with trypsin and matching with the theoretical peptide masses of all known proteins from all species. Approximately 5800 spots, excised from 14 two-dimensional gels, were analyzed which resulted in the identification of about 2500 proteins that were the products of 328 different genes. The database includes 112 newly identified gene products. The fractionation prior to two-dimensional electrophoresis was essential for the detection of the new proteins, 55% of which were found in the microsomal and 35% in the cytosolic fraction. The more frequently identified proteins in the various gels were heat shock proteins, house-keeping enzymes, such as ATP synthase chains, disulfide isomerase, and structural proteins, such as tropomyosin. About 45% of the identified proteins were detected 1-3 times, 45% 4-9 times, and the rest 10 or more times. Most proteins were represented by many spots. In average, about 18-20 spots were detected per gene product.