- Email: [email protected]
The impact of new technologies on human population studies
Mutation Research 544 (2003) 349–360
The impact of new technologies on human population studies Michael D. Waters a,∗ , James K. Selkirk a , Kenneth Olden b a
National Center for Toxicogenomics, 111 Alexander Drive, P.O. Box 12233, MD F1-05, Research Triangle Park, NC 27709-2233, USA b National Institute of Environmental Health Sciences, 111 Alexander Drive, P.O. Box 12233, MD F1-05, Research Triangle Park, NC 27709-2233, USA Received 5 May 2003; received in revised form 21 June 2003; accepted 23 June 2003
Abstract Human population studies involve clinical or epidemiological observations that associate environmental exposures with health endpoints and disease. Clearly, these are the most sought after data to support assessments of human health risk from environmental exposures. However, the foundations of many health risk assessments rest on experimental studies in rodents performed at high doses that elicit adverse outcomes, such as organ toxicity or tumors. Using the results of human studies and animal data, risk assessors define the levels of environmental exposures that may lead to disease in a portion of the population. These decisions on potential health risks are frequently based on the use of default assumptions that reflect limitations in our scientific knowledge. An important immediate goal of toxicogenomics, including proteomics and metabonomics, is to offer the possibility of making decisions affecting public health and public based on detailed toxicity, mechanistic, and exposure data in which many of the uncertainties have been eliminated. Ultimately, these global technologies will dramatically impact the practice of public health and risk assessment as applied to environmental health protection. The impact is already being felt in the practice of toxicology where animal experimentation using highly controlled dose–time parameters is possible. It is also being seen in human population studies where understanding human genetic variation and genomic reactions to specific environmental exposures is enhancing our ability to uncover the causes of variations in human response to environmental exposures. These new disciplines hold the promise of reducing the costs and time lines associated with animal and human studies designed to assess both the toxicity of environmental pollutants and efficacy of therapeutic drugs. However, as with any new science, experience must be gained before the promise can be fulfilled. Given the numbers and diversity of drugs, chemicals and environmental agents; the various species in which they are studied and the time and dose factors that are critical to the induction of beneficial and adverse effects, it is only through the development of a profound knowledge base that toxicology and environmental health can rapidly advance. The National Institute of Environmental Health Sciences (NIEHS), National Center for Toxicogenomics and its university-based Toxicogenomics Research Consortium (TRC), and resource contracts, are engaged in the development, application and standardization of the science upon which to the build such a knowledge base on Chemical Effects in Biological Systems (CEBS). In addition, the NIEHS Environmental Genome Project (EGP) is working to systematically identify and characterize common sequence polymorphisms in many genes with suspected roles in determining chemical sensitivity. The rationale of the EGP is that certain genes have a greater than average Abbreviations: AFLP, amplified fragment length polymorphism; CEBS, Chemical Effects in Biological Systems; EPA, US Environmental Protection Agency; EGP, Environmental Genome Project; IPCS, International Programme for Chemical Safety; NCT, National Center for Toxicogenomics; NIEHS, National Institute of Environmental Health Sciences; PCR, polymerase chain reaction; PB/PK, physiologically-based pharmacokinetic; RTI, reverse transcript imaging; SAGE, serial analysis of gene expression; SDS, sodium dodecyl sulfate; SNPs, single nucleotide polymorphisms; TRC, Toxicogenomics Research Consortium ∗ Corresponding author. Tel.: +1-919-316-4589; fax: +1-919-541-1460. E-mail address: [email protected] (M.D. Waters). 1383-5742/$ – see front matter © 2003 Published by Elsevier B.V. doi:10.1016/j.mrrev.2003.06.022
350
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
influence over human susceptibility to environmental agents. If we identify and characterize the polymorphism in those genes, we will increase our understanding of human disease susceptibility. This knowledge can be used to protect susceptible individuals from disease and to reduce adverse exposure and environmentally induced disease. © 2003 Published by Elsevier B.V. Keywords: Toxicogenomics; Microarray; Gene expression; Proteomics; Metabononics; Bioinformatics; Phenotypic anchoring; Phenotype; Molecular expression; Systems biology; Systems toxicology; Database; Knowledge base; Compendia; Ontologies; Global query; Sequence; Transcription factors; Single nucleotide polymorphisms
1. Introduction All individuals are continually exposed to hazardous agents and chemicals in their environment. Thus, there is a certain probability that each individual will suffer an adverse effect, such as disease, as a result of such exposure. This process has been conceptualized by environmental health researchers and expressed as a paradigm that describes the continuum between exposure and disease. Environmental exposures may be derived from several external sources including the environmental media (air, water, and soil), the diet, or from components specific to an individual’s workplace or occupation. The route of exposure varies (inhalation, oral, and dermal), and the concentration of hazardous compounds in the environment also varies over time. Importantly, each individual receives a unique internal dose of the compound, which is determined by the individual’s activities and movements while he or she is being exposed. Most biologically active agents have a much higher probability of causing an adverse effect in one or more “target tissues” in the human body (i.e., liver, skin, intestine, lung, blood, and bone). Thus, the amount of a compound that actually reaches the target tissue is described as the biologically effective dose. A compound can reach this target tissue at a significant concentration and still have no biological consequences, if the compound is efficiently removed or detoxified by cellular defense systems before cellular damage occurs. However, if cellular damage occurs, the damage can have a biological effect that leads to sustained or permanent changes in biological structures or functions. In many cases where no treatment or intervention is introduced, disease may develop and progress in these exposed individuals. A particular focus of this symposium is on the monitoring of human exposure, effects, and susceptibility to toxic agents and particularly genotoxic agents in the environment [1,2]. How can these studies in hu-
man populations assist us: (1) in measuring internal dose, (2) in understanding mode of action, (3) the etiology of environmentally-induced disease, and (4) in improving risk assessment methods and models? The focus of this paper is on new genomic and proteomic technologies and the impact that they can have in studies on human populations. In addition, we briefly consider their applications in risk assessment, with emphasis on detection of human susceptibility genes and genetic polymorphisms. Recent advancements in molecular profiling analysis using microarray and proteomics technologies provide tools for assessing expression of hundreds to thousands of genes in a single experiment. A molecular expression profile provides a complete molecular phenotype of the cell under a specific set of conditions and exposures. With these same tools comparative expression profiles can be generated for representative genomes in both experimental animals and in humans targeting functional gene clusters that represent a variety of physiological and toxicological processes. Among these genes and their corresponding proteins are the molecular biomarkers of the future. A new version of the familiar exposure-to-disease paradigm has been developed by National Institute of Environmental Health Sciences (NIEHS) Environmental Genome Project (EGP) and is reproduced below in Fig. 1. This illustration represents a genomics perspective on the familiar exposure-to-disease paradigm [3]. Such a perspective considers the environmental exposure including its distribution and metabolism, but it also considers its influence on signal transduction pathways, whether it causes genomic damage, and whether, for example, it affects cell differentiation or apoptosis genes, cell cycle control or repair genes. It asks about measurable, reproducible, alterations in gene expression related to the exposure, about the dose–response relationship, the temporal association and the sequence of molecular expression events. It
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
351
Fig. 1. Environmental genomics and the exposure-to-disease paradigm.
also asks about the influence of genetic variability, including single nucleotide polymorphisms (SNPs) on the human gene expression profile. The basic steps of the paradigm are conventional, but the potential information output in terms of toxicological information from global molecular expression studies is greatly increased.
2. Risk assessment The International Programme for Chemical Safety (IPCS) has produced [7] a generic “Framework for Risk Assessment” using many of the concepts presented in April 1996, the Proposed US Environmental Protection Agency (EPA) Cancer Risk Assessment Guidelines [4–6]. The 10 elements of the framework are listed below: • Toxicological or disease endpoint. • Postulated mode of action. • Key events—measurable events related to the mode of action (biomarkers). • Dose–response relationship for the key events. • Temporal association/sequence of events. • Strength, consistency and specificity of association of response with key events. • Biological plausibility and coherence. • Other modes of action.
• Assessment of postulated mode of action. • Uncertainties, inconsistencies and data gaps. Collectively, these elements give an idea of how molecular studies of exposure and effect can be used in risk assessment. The framework is applicable to all endpoints—to cancer as well as non-cancer effects. It places major emphasis on understanding the mode of action of the agent in question. It assigns a central role to “key events”—(molecular biomarkers in the context of the present discussion) that are related to the mode of action. These “key events” should ideally display a dose–response relationship. There should be a definable temporal association of the key events with the disease of concern, and the sequence of events should be consistent with the hypothesized mode of action. Clearly, this framework will require substantial basic and applied research to provide the information needed for risk assessment. It also provides a useful construct for our consideration of applications and impacts of toxicogenomics in human population studies, with subsequent impacts on human health and risk assessment. Using the results of human studies and animal data, risk assessors define the levels of environmental exposures that may lead to disease in a portion of the population. These decisions on potential health risks are frequently based on the use of default assumptions that reflect limitations in our
352
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
scientific knowledge [8]. An important immediate goal of toxicogenomics, including proteomics and metabonomics, is to offer the possibility of making decisions affecting public health and public based on detailed toxicity, mechanistic, and exposure data in which many of the uncertainties have been eliminated.
3. Genomics, proteomics, functional genomics We belatedly define the term “genomics.” An “-ome” is an abstract entity or group. Genomics is the study of genes in the aggregate—DNA, the primary RNA transcript and messenger RNA. Genomics is also used to refer to transcriptomics—the quantitative analysis of transcripts at the mRNA level. Expression data at the mRNA level can be produced using several technologies such as DNA microarrays, amplified fragment length polymorphism (AFLP) [9–18], reverse transcript imaging (RTI) or subtractive hybridization [19–32], serial analysis of gene expression (SAGE) [33–70] and others. For purposes of the present discussion SAGE and microarray analysis are most important. SAGE is a technique designed to take advantage of high-throughput sequencing technology to obtain a quantitative profile of cellular gene expression. Essentially, the SAGE technique measures not the expression level of a gene, but quantifies a “tag” which represents the transcription product of a gene. A tag, for the purposes of SAGE, is a nucleotide sequence of a defined length, directly 3 –adjacent to the 3 –most restriction site for a particular restriction enzyme. DNA microarrays have become a more popular and important approach offering great potential for human environmental studies. On a typical microarray, each gene of interest is represented either by a long DNA fragment (200–2400 bp) typically generated by polymerase chain reaction (PCR) and spotted on a suitable substrate using robotics [71,72], or by several short oligonucleotides (20–70 bp) synthesized directly onto a solid support using photolabile nucleotide chemistry [73–76]. Total RNA or mRNA is isolated from control and treated tissues, and reverse transcribed in the presence of radioactive or fluorescent labeled nucleotides, and the labeled probes are then hybridized to the arrays. The intensity of the array signal is measured for each gene transcript by laser scanning con-
focal microscopy. The ratio between the signals of control and treated samples reflect the relative drug or chemical-induced change in transcript abundance [77]. Current microarray technology allows the simultaneous expression monitoring of 20,000–25,000 genes. In the immediate future, when all human genes have been identified and cloned [78], and technological advances allow further miniaturization of solid supports we will have the capability to examine simultaneously all expressed genes in the human genome. Proteomics [79,80] is defined as the study of protein products in aggregate—it applies to the translation from the mRNA to the primary protein products, and their maturation and modification to yield active proteins as components of a cell, tissue, or organism. The aim of proteomics is to quantify gene expression down-stream, to obtain a “snapshot” of gene regulation in the actual control of cell function. Proteins generally display greater stability than mRNA. The technical problem is that global quantitative expression analysis at the protein level has largely been restricted to the use of 2D gel electrophoresis. In this technique, isoelectric focusing in the first dimension is followed by SDS slab gel electrophoresis-based molecular weight separation [81,82]. Protein spots are identified by mass spectrometry following generation of peptide mass fingerprints [83,84] and sequence tags [85]. Functional genomics or computational biology is the systematic analysis of gene products, including both mRNAs and proteins to determine their function. Gene function can be inferred from sequence homology among predicted proteins, where the function of at least one protein is known [86]. Computational methods have been developed that detect functional linkages among predicted proteins [87]. But sequence analysis alone is insufficient to fully inform us about gene function. The traditional microarray-based approach to functional analysis has been to cluster genes according to their expression behavior under a range of conditions and to assign function according to the set of known genes that fall into these clusters [88,89]. The premise with this “guilt-by-association” approach is that clustered genes will be co-regulated and, therefore, be involved in similar functions. There are several additional methods that do not require clustering to analyze gene expression profiles and predict gene function. The limitations are mainly in the models, and experience necessary to analyze large datasets.
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
Despite the challenges, expression profiles provide a complete molecular phenotype of the cell under a specific set of conditions and have numerous applications in studies on disease etiology [90].
4. Toxicogenomics Toxicogenomics is the study of the response of a genome to environmental stressors and toxicants. It combines genetics, genomic-scale mRNA expression (transcriptomics), cell and tissue-wide protein expression (proteomics), metabolite profiling (metabonomics) and bioinformatics with conventional toxicology in an effort to understand the role of gene–environment interactions in disease. Toxicogenomics seeks to identify and characterize mechanisms of action of known and suspected toxicants [91–102]. The analysis of gene expression patterns for different chemicals under different doses or times of exposure scenarios is used to gain a clearer understanding of genes that are mechanistically linked to a toxicologic response. This process, including the development, perturbation and reassessment of models that relate gene expression profiles to toxic outcomes is referred to as “systems toxicology” [103,104]. An important immediate goal of toxicogenomics, including proteomics and metabonomics, is to offer the possibility of making decisions affecting public health and public based on detailed toxicity, mechanistic, and exposure data in which many of the uncertainties have been eliminated. It will be important to determine how many and which genes should be measured to characterize a toxic response and distinguish it from physiologically adaptive responses that are not linked to toxicity. The use of global gene expression data in hazard identification in the absence of a correct interpretation of the toxicological significance of the data would be unfortunate. Using a combination of laboratory and field studies, comparison of chemicals from different mode of action classes (for example, cytotoxic chemicals, peroxisome proliferators, or estrogenic chemicals) will allow the identification of groups of genes whose expression at multiple doses and times is consistently linked to specific exposures and disease outcomes. The outcome of environmental exposures is influenced by the function of many human genes. Not all
353
of these genes or their functions are known at present. However, many genes have been identified that are likely to be important factors in genetic susceptibility to environmentally induced disease. These genes tend to fall into the following categories: cell cycle control, DNA repair, regulation of cell division, cell signaling, cell structure, apoptosis and metabolism. Several genes that control metabolic pathways are crucial determinants of the outcome of exposure. In many cases, a compound enters a biological system in an inert innocuous form that is metabolically converted into a reactive species that causes cellular damage. The converse is also true; some metabolic pathways destroy toxic compounds by changing their chemical structure, frequently making them more water soluble and easier to excrete. Cell cycle and cell division genes regulate the ability of a cell to proliferate, grow and differentiate. Changes in the progression of a cell through the cell cycle can increase the ability of a cell to survive stress; in most cases, a proliferating cell exposed to stress will enhance its survival by delaying the cell cycle so that cellular damage can be repaired prior to cell division. Cell signaling and gene expression pathways have profound effects on all cellular functions, including cell proliferation and differentiation. Some exogenous agents can activate these pathways in aberrant and deleterious ways (i.e., agents that mimic a biological component), and this can disrupt or alter normal cellular function. DNA repair genes influence the outcome of exposure to environmental agents that cause DNA damage. Individuals with higher or lower capacity for DNA repair have decreased or increased risk of certain types of environmentally induced disease, respectively. Heavily damaged cells often die by a process known as programmed cell death or apoptosis. This process protects the organism by removing aberrant cells and damaged structures. To obtain the most relevant data from gene expression studies with microarrays requires that experiments be performed at multiple doses and varied exposure durations to identify those genes clearly linked to a toxic response. To develop approaches that will maximize the likelihood of detecting true positives for human exposure and minimize false negatives, a substantial matrix of data on chemicals with known exposure-disease outcomes will need to be obtained. This will require the evaluation of gene expression profiles of chemicals not causing health
354
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
effects as well as those known to cause disease and positive chemicals with varying potency for causing disease. When the database on known chemicals becomes comprehensive enough, the fingerprints of new chemicals with unknown toxic properties can be compared to well-characterized gene expression profiles, allowing the new chemical to be provisionally placed into one or more mode of action classes. Then more directed studies may be undertaken to confirm or refute the predicted mode of action and toxic outcome for the new chemical. As with any new science, experience must be gained in toxicogenomics before its promise can be fulfilled. This experience will involve the assembly of massive amounts of data. The difficulty is that each of the new technological approaches being applied in genome and proteome analysis can overwhelm the current information infrastructure. The challenge is not simply managing the data flow that will be generated by these new approaches. New models will be needed to manipulate the data and new analytic strategies will be required to interpret it. The NIEHS National Center for Toxicogenomics (NCT) is working to help the field of environmental health research evolve into a knowledge-based science in which experimental toxicogenomics and clinical data are compiled. The NCT (http://www.niehs.nih. gov/nct/) and its university-based Toxicogenomics Research Consortium (TRC, http://www.niehs.nih.gov/ nct/trc.htm) and resource contracts are engaged in the development, application and standardization of the
science upon which to the build such a knowledge base (Fig. 2) on Chemical Effects in Biological Systems (CEBS, http://www.niehs.nih.gov/nct/cebs.htm). CEBS [104] will be created as a high quality, publicly accessible relational database that is compatible with standard laboratory output platforms. Database development will be integrated with strategic toxicogenomics experimental design and conduct. Standardized procedures, protocols, data formats, and assessment methods will be used to ensure that data meet a uniform high level of quality. Raw data sets from NCT experiments will be available in their entirety. Relational and descriptive compendia will be included on toxicologically important genes, groups of genes, SNPs, and mutants and their functional phenotypes. Information about the biological effects of chemicals and other agents and their mechanism of action will be collected from the literature and stored. CEBS will be fully searchable by compound, structure, toxicity, pathology, gene, gene group, SNP, pathway, and network. Dictionaries and explanatory text will guide researchers in understanding toxicogenomics datasets. CEBS will be linked extensively to other databases and to Web genomics and proteomics resources, providing users the suite of information and tools needed to fully interpret toxicogenomics data. Computational and informatics tools will play a significant role in improving our understanding of toxicant-related disease by creating a system of predictive toxicology.
Fig. 2. Conceptual framework for the CEBS knowledge base.
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
Future promises of CEBS include: 1. Developing large context-annotated datasets that allow precise definition of biological/toxicological pathways and lead to the identification of new biomarkers. 2. Linking genomic sequence to expression data to determine those genes that may be responsible for the coordinated regulation of sets of genes. 3. Increasing the interpretability and dimensionality of expression data by including data from new types of arrays including protein arrays and SNP arrays. 4. Aiding in development of new algorithms and computational tools that allow predictive modeling of gene interactions and networks relevant to human health and disease. 5. Impact on human population studies A challenge for environmental toxicologists in applying these new technologies in studies in humans is to better determine the level of chemical exposure. This includes actual measures of exposure or dose and total human exposure across all relevant pathways. Exposure assessment is the component of risk assessment that is most neglected especially by biologists. The measurement of human environmental exposure is a complicated business. There are many variables that influence what amount of an exposure will become a target tissue dose. All of them come into play before the opportunity to induce an adverse outcome. Exposure-related databases that report actual measures of exposure or dose are very limited and virtually none collect measures of total human exposure across all relevant environmental pathways. Gaining a better estimate of exposure and internal dose is an important direction and challenge in future toxicogenomics research. Accurate exposure information will be essential to understand environmentally induced disease and critical to support the needs of human health risk assessment. An area of concern in approaching this challenge is that assessment of exposure in people who do not have a pre-characterized baseline molecular expression will be difficult. This is because baseline expression among individuals is expected to vary widely with differences in diet, personal habits, and health status. Because of these baseline differences, changes
355
due to environmental chemical exposure may not be greater than the noise of gene expression variability. Thus, in the future, gene expression profiles from exposed individuals will probably be compared to a standard gene expression profile or profiles based on a national or international gene expression database for humans [105]. Assuming that the scientific (as well as ethical) issues of human gene expression databases are resolved, microarray analysis of gene expression patterns could be used to identify the type of chemical exposure experienced by individuals and to confirm its linkage to a mode of action class as described previously. Another major application of gene expression profiling is to understand human susceptibility to disease. Experimental studies in molecular genetics, toxicology and other biomedical fields have shown that genetic differences between two individuals can determine the relative sensitivity of each individual to environmental chemicals or agents. Some human inherited disease susceptibilities are caused by a single inherited trait, and other disease susceptibilities may be determined by multiple genetic traits [106,107]. For many diseases, the genetic complement of the individual does not itself cause disease; instead, disease is the outcome of a complex interplay between multiple genetic and environmental factors [108–110]. The National Institute of Environmental Health Sciences (NIEHS) in Research Triangle Park, NC has initiated the Environmental Genome Project (EGP, http://www.niehs.nih.gov/envgenom/) to systematically identify and characterize common sequence polymorphisms in many genes with suspected roles in determining chemical sensitivity. The rationale of the EGP is that certain genes have a greater influence over human susceptibility to environmental agents than others. If we identify and characterize the polymorphism in those genes, we will increase our understanding of human disease susceptibility. This knowledge can be used to protect susceptible individuals from disease and to reduce adverse exposure and environmentally induced disease.
6. Human polymorphisms There may be as many as 10–30 million SNPs in the human genome, if all variants in all individuals are
356
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
counted, and only about 1% of them alter an amino acid in a protein encoded by a gene. While most SNPs do not directly affect protein function, some affect function indirectly, through alteration of regulatory sequences that control gene expression, or the stability or processing of the mRNA transcript of the gene. It is important to note that the frequency of a polymorphism can vary in different population subgroups. In other words, polymorphic sites can be common or rare in the sequence of different population groups. The common polymorphic sites are found in one of five individuals or higher, but most variants are represented at much lower frequencies. EGP resequencing studies are designed to effectively detect polymorphic sites that are represented at 1% or higher in the sampled population. The EGP has identified a group of human genes that are likely to influence the outcome of environmental exposure. Polymorphic variants in these “environmentally responsive genes” are being identified by systematic resequencing of a predefined diverse, randomly selected set of human DNA samples. The DNA resource resource is maintained by the Coriell Institute for Medical Research and is part of the Coriell Cell Repositories. The repository includes cell lines and DNA samples from a panel of 450 unrelated male and female individuals. The samples are from individuals who belong to diverse ethnic groups including European–Americans, African–Americans, Mexican– Americans, native Americans, and Asian–Americans. Candidate genes are currently being sequenced across the entire panel of 450 individuals to identify common sequence variation for functional analysis and population-based studies. Automated DNA sequencing is being used to identify and genotype SNPs in human candidate genes. All SNPs have been identified using only high quality sequence data (Q > 25) and each SNP reported from the NIEHS-SNPs program has been confirmed in multiple individuals and/or in multiple reactions. The EGP has generated a list of 554 environmentally responsive genes that are potential targets for re-sequencing in subgroups according to a phased timeline. The list is not comprehensive and it is expected that more genes may be added over time. The first phase of resequencing extended from 1998 to 2001 and has been focused on polymorphic variants on a group of 123 genes. The second phase of rese-
quencing, which extends from 2001 to 2004, focuses on approximately 200 DNA repair and cell cycle control genes. Genes regulating metabolism, signal transduction, and apoptosis will also be targeted for resequencing. The project will develop a central database of polymorphisms for these genes, and foster population-based studies of gene–environment interaction in disease etiology. The functional significance of specific gene variants is also being characterized to determine which gene variants are correlated with increased or decreased risk of disease. NIEHS SNPs are available in the GeneSNPs database (http://www.genome.utah.edu/genesnps/) as well as the national database resource, dbSNP (http:// www.niehs.nih.gov/envgenom/snpsdb.htm). GeneSNPs provides a gene-centric map of the genome structure, coding sequences, and identified allelic variation in genes being targeted for a role in disease susceptibility by the NIEHS. This database provides a graphical view of all available SNP data including allele frequencies and genotypes in select populations. This information is key in selecting the polymorphic sites needed to examine disease risk in human population studies. 7. Conclusions Clearly, these sequence-based global technologies and the databases that result from their application will dramatically impact the practice of public health and risk assessment [111] as applied to environmental health protection. These new approaches hold the promise of reducing the costs and time lines associated with animal and human studies designed to assess both the toxicity of environmental pollutants and efficacy of therapeutic drugs. More importantly, they promise to dramatically improve the fidelity of the information that will be available from animal as well as toxicogenomics studies in human populations. These studies will ultimately find their way into the risk assessment process and will play an important role in the protection of human environmental health. Acknowledgements We thank Dr. Douglas Bell for valuable discussions on human epidemiological studies. We thank Nancy
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
Stegman, Jerry Nehls for assistance on information technology. We are indebted to Samuel Wilson, and Lutz Birnbaumer for their continuing support of the NCT program. References [1] R.J. Albertini, Biomarker responses in human populations: towards a worldwide map, Mutat. Res. 428 (1999) 217–226. [2] R.J. Albertini, J.A. Nicklas, J.P. O’Neill, Future research directions for evaluating human genetic and cancer risk from environmental exposures, Environ. Health Perspect. 104 (Suppl. 3) (1996) 503–510. [3] D.A. Bennett, M.D. Waters, Applying biomarker research, Environ. Health Perspect. 108 (2000) 907–910. [4] W.H. Farland, Cancer risk assessment: evolution of the process, Prev. Med. 25 (1996) 24–25. [5] V.L. Dellarco, J.A. Wiltse, US Environmental Protection Agency’s Revised Guidelines for carcinogen risk assessment: incorporating mode of action data, Mutat. Res. 405 (1998) 273–277. [6] M. Andersen, D. Brusick, S. Cohen, Y. Dragan, C. Frederick, J.I. Goodman, G. Hard, B. Meek, E.J. O’Flaherty, US Environmental Protection Agency’s Revised Cancer Guidelines for carcinogen risk assessment, Toxicol. Appl. Pharmacol. 153 (1998) 133–136. [7] C. Sonich-Mullin, R. Fielder, J. Wiltse, K. Baetcke, J. Dempsey, P. Fenner-Crisp, D. Grant, M. Hartley, A. Knaap, D. Kroese, I. Mangelsdorf, E. Meek, J.M. Rice, M. Younes, IPCS conceptual framework for evaluating a mode of action for chemical carcinogenesis, Regul. Toxicol. Pharmacol. 34 (2001) 146–152. [8] M.E. Andersen, M.E. Meek, G.A. Boorman, D.J. Brusick, S.M. Cohen, Y.P. Dragan, C.B. Frederick, J.I. Goodman, G.C. Hard, E.J. O’Flaherty, D.E. Robinson, Lessons learned in applying the US EPA proposed cancer guidelines to specific compounds, Toxicol. Sci. 53 (2000) 159–172. [9] P. Vos, R. Hogers, M. Bleeker, M. Reijans, T. van de Lee, M. Hornes, A. Frijters, J. Pot, J. Peleman, M. Kuiper, et al., AFLP: a new technique for DNA fingerprinting, Nucl. Acids Res. 23 (1995) 4407–4414. [10] T. Money, S. Reader, L.J. Qu, R.P. Dunford, G. Moore, AFLP-based mRNA fingerprinting, Nucl. Acids Res. 24 (1996) 2616–2617. [11] M. Otsen, M. den Bieman, M.T. Kuiper, M. Pravenec, V. Kren, T.W. Kurtz, H.J. Jacob, A. Lankhorst, B.F. van Zutphen, Use of AFLP markers for gene mapping and QTL detection in the rat, Genomics 37 (1996) 289–294. [12] B.L. Roman, V.N. Pham, P.E. Bennett, B.M. Weinstein, Non-radioisotopic AFLP method using PCR primers fluorescently labeled with Cy5, Biotechniques 26 (1999) 236–238. [13] J.J. Lin, J. Ma, J. Kuo, et al., Chemiluminescent detection of AFLP markers, Biotechniques 26 (1999) 344–348. [14] U.G. Mueller, L.L. Wolfenbarger, AFLP genotyping and fingerprinting, Trends Ecol. Evol. 14 (1999) 389–394.
357
[15] S. Iyoda, A. Wada, J. Weller, S.J. Flood, E. Schreiber, B. Tucker, H. Watanabe, Evaluation of AFLP, a high-resolution DNA fingerprinting method, as a tool for molecular subtyping of enterohemorrhagic Escherichia coli O157:H7 isolates, Microbiol. Immunol. 43 (1999) 803–806. [16] A. Reineke, P. Karlovsky, Simplified AFLP protocol: replacement of primer labeling by the incorporation of alpha-labeled nucleotides during PCR, Biotechniques 28 (2000) 622–623. [17] P.E. Klein, R.R. Klein, S.W. Cartinhour, P.E. Ulanch, J. Dong, J.A. Obert, D.T. Morishige, S.D. Schlueter, K.L. Childs, M. Ale, J.E. Mullet, A high-throughput AFLP-based method for constructing integrated genetic and physical maps: progress toward a sorghum genome map, Genome Res. 10 (2000) 789–807. [18] M. Prochazka, K. Walder, J. Xia, AFLP fingerprinting of the human genome, Hum. Genet. 108 (2001) 59–65. [19] T. Nakagawa, M. Mabry, A. de Bustros, J.N. Ihle, B.D. Nelkin, S.B. Baylin, Introduction of v-Ha-ras oncogene induces differentiation of cultured human medullary thyroid carcinoma cells, Proc. Natl. Acad. Sci. U.S.A. 84 (1987) 5923–5927. [20] G. Tarone, M.A. Russo, E. Hirsch, T. Odorisio, F. Altruda, L. Silengo, G. Siracusa, Expression of beta-1 integrin complexes on the surface of unfertilized mouse oocyte, Development 117 (1993) 1369–1375. [21] C.M. Niessen, O. Cremona, H. Daams, S. Ferraresi, A. Sonnenberg, P.C. Marchisio, Expression of the integrin alpha-6 beta-4 in peripheral nerves: localization in Schwann and perineural cells and different variants of the beta-4 subunit, J. Cell Sci. 107 (Part 2) (1994) 543–552. [22] A.K. Stuart-Tilley, B.E. Shmukler, D. Brown, S.L. Alper, Immunolocalization and tissue-specific splicing of AE2 anion exchanger in mouse kidney, J. Am. Soc. Nephrol. 9 (1998) 946–959. [23] F. Mannello, M. Malatesta, F. Luchetti, S. Papa, S. Battistelli, G. Gazzanelli, Immunoreactivity, ultrastructural localization, and transcript expression of prostate-specific antigen in human neuroblastoma cell lines, Clin. Chem. 45 (1999) 78– 84. [24] S. Alcolea, M. Theveniau-Ruissy, T. Jarry-Guichard, I. Marics, E. Tzouanacou, J.P. Chauvin, J.P. Briand, A.F. Moorman, W.H. Lamers, D.B. Gros, Downregulation of connexin 45 gene products during mouse heart development, Circ. Res. 84 (1999) 1365–1379. [25] Y. Kiyozuka, H. Miyazaki, K. Yoshizawa, H. Senzaki, D. Yamamoto, K. Inoue, K. Bessho, Y. Okubo, K. Kusumoto, A. Tsubura, An autopsy case of malignant mesothelioma with osseous and cartilaginous differentiation: bone morphogenetic protein-2 in mesothelial cells and its tumor, Dig. Dis. Sci. 44 (1999) 1626–1631. [26] D.C. Cross, J.P. Munoz, P. Hernandez, R.B. Maccioni, Nuclear and cytoplasmic tau proteins from human nonneuronal cells share common structural and functional features with brain tau, J. Cell Biochem. 78 (2000) 305–317. [27] C.G. Carlotti Jr., J.M. Drake, J.P. Hladky, I. Teshima, L.E. Becker, J.T. Rutka, Primary Ewing’s sarcoma of the skull
358
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
M.D. Waters et al. / Mutation Research 544 (2003) 349–360 in children: utility of molecular diagnostics, surgery and adjuvant therapies, Pediatr. Neurosurg. 31 (1999) 307–315. R. Leemans, B. Egger, T. Loop, L. Kammermeier, H. He, B. Hartmann, U. Certa, F. Hirth, H. Reichert, Quantitative transcript imaging in normal and heat-shocked Drosophila embryos by using high-density oligonucleotide arrays, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 12138–12143. M.J. O’Sullivan, E.J. Perlman, J. Furman, P.A. Humphrey, L.P. Dehner, J.D. Pfeifer, Visceral primitive peripheral neuroectodermal tumors: a clinicopathologic and molecular study, Hum. Pathol. 32 (2001) 1109–1115. D.K. Graham, L.C. Stork, Q. Wei, J.D. Ingram, F.M. Karrer, G.W. Mierau, M.A. Lovell, Molecular genetic analysis of a small bowel primitive neuroectodermal tumor, Pediatr. Dev. Pathol. 5 (2002) 86–90. M.C. Rissoan, T. Duhen, J.M. Bridon, N. Bendriss-Vermare, C. Peronne, B. de Saint Vis, F. Briere, E.E. Bates, Subtractive hybridization reveals the expression of immunoglobulin-like transcript 7, Eph-B1, granzyme B, and three novel transcripts in human plasmacytoid dendritic cells, Blood 100 (2002) 3295–3303. A. Borchert, N.E. Savaskan, H. Kuhn, Regulation of expression of the phospholipid hydroperoxide/sperm nucleus glutathione peroxidase gene: tissue-specific expression pattern and identification of functional cis- and transregulatory elements, J. Biol. Chem. 278 (2003) 2571–2580. V.E. Velculescu, L. Zhang, B. Vogelstein, K.W. Kinzler, Serial analysis of gene expression, Science 270 (1995) 484– 487. S.L. Madden, E.A. Galella, J. Zhu, A.H. Bertelsen, G.A. Beaudry, SAGE transcript profiles for p53-dependent growth regulation, Oncogene 15 (1997) 1079–1085. J. Powell, Enhanced concatemer cloning—a modification to the SAGE (Serial Analysis of Gene Expression) technique, Nucl. Acids Res. 26 (1998) 3445–3446. S.M. Wang, J.D. Rowley, A strategy for genome-wide gene analysis: integrated procedure for gene identification, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 11909–11914. N.A. Datson, J. van der Perk-de Jong, M.P. van den Berg, E.R. de Kloet, E. Vreugdenhil, MicroSAGE: a modified procedure for serial analysis of gene expression in limited amounts of tissue, Nucl. Acids Res. 27 (1999) 1300–1307. F. Baas, H.F. Tabak, A tale of tags: report on a HUGO/EU SAGE workshop, 29 January–1 February 1999, Hilversum, The Netherlands, Eur. J. Hum. Genet. 7 (1999) 510–512. A. van den Berg, J. van der Leij, S. Poppema, Serial analysis of gene expression: rapid RT-PCR analysis of unknown SAGE tags, Nucl. Acids Res. 27 (1999) e17. D.G. Peters, A.B. Kassam, H. Yonas, E.H. O’Hare, R.E. Ferrell, A.M. Brufsky, Comprehensive transcript analysis in small quantities of mRNA by SAGE-lite, Nucl. Acids Res. 27 (1999) e39. A. Lal, A.E. Lash, S.F. Altschul, V. Velculescu, L. Zhang, R.E. McLendon, M.A. Marra, C. Prange, P.J. Morin, K. Polyak, N. Papadopoulos, B. Vogelstein, K.W. Kinzler, R.L. Strausberg, G.J. Riggins, A public database for gene expression in human cancers, Cancer Res. 59 (1999) 5403– 5407.
[42] M. Larsson, S. Stahl, M. Uhlen, A. Wennborg, Expression profile viewer (ExProView): a software tool for transcriptome analysis, Genomics 63 (2000) 341–353. [43] J.J. Going, B.A. Gusterson, Molecular pathology and future developments, Eur. J. Cancer 35 (1999) 1895–1904. [44] A.E. Lash, C.M. Tolstoshev, L. Wagner, G.D. Schuler, R.L. Strausberg, G.J. Riggins, S.F. Altschul, SAGEmap: a public gene expression resource, Genome Res. 10 (2000) 1051– 1060. [45] J. Powell, SAGE: the serial analysis of gene expression, Methods Mol. Biol. 99 (2000) 297–319. [46] J. Stollberg, J. Urschitz, Z. Urban, C.D. Boyd, A quantitative evaluation of SAGE, Genome Res. 10 (2000) 1241–1248. [47] S.L. Madden, C.J. Wang, G. Landes, Serial analysis of gene expression: from gene discovery to target identification, Drug Discov. Today 5 (2000) 415–425. [48] M. Ishii, S. Hashimoto, S. Tsutsumi, Y. Wada, K. Matsushima, T. Kodama, H. Aburatani, Direct comparison of GeneChip and SAGE on the quantitative accuracy in transcript profiling analysis, Genomics 68 (2000) 136–143. [49] E.H. Margulies, J.W. Innis, eSAGE: managing and analysing data generated with serial analysis of gene expression (SAGE), Bioinformatics 16 (2000) 650–651. [50] V.E. Velculescu, B. Vogelstein, K.W. Kinzler, Analysing uncharted transcriptomes with SAGE, Trends Genet. 16 (2000) 423–425. [51] S.Q. Ye, L.Q. Zhang, F. Zheng, D. Virgil, P.O. Kwiterovich, miniSAGE: gene expression profiling using serial analysis of gene expression from 1 g total RNA, Anal. Biochem. 287 (2000) 144–152. [52] A.H. van Kampen, B.D. van Schaik, E. Pauws, E.M. Michiels, J.M. Ruijter, H.N. Caron, R. Versteeg, S.H. Heisterkamp, J.A. Leunissen, F. Baas, M. van der Mee, USAGE: a web-based approach towards the analysis of SAGE data (Serial Analysis of Gene Expression), Bioinformatics 16 (2000) 899–905. [53] H. Caron, B. van Schaik, M. van der Mee, F. Baas, G. Riggins, P. van Sluis, M.C. Hermus, R. van Asperen, K. Boon, P.A. Voute, S. Heisterkamp, A. van Kampen, R. Versteeg, The human transcriptome map: clustering of highly expressed genes in chromosomal domains, Science 291 (2001) 1289–1292. [54] J. Bartlett, Technology evaluation: SAGE, genzyme molecular oncology, Curr. Opin. Mol. Ther. 3 (2001) 85–96. [55] C. Stoeckert, A. Pizarro, E. Manduchi, M. Gibson, B. Brunk, J. Crabtree, J. Schug, S. Shen-Orr, G.C. Overton, A relational schema for both array-based and SAGE gene expression experiments, Bioinformatics 17 (2001) 300–308. [56] E.H. Margulies, S.L. Kardia, J.W. Innis, Identification and prevention of a GC content bias in SAGE libraries, Nucl. Acids Res. 29 (2001) E60–60. [57] J.C. Moreno, E. Pauws, A.H. van Kampen, M. Jedlickova, J.J. de Vijlder, C. Ris-Stalpers, Cloning of tissue-specific genes using serial analysis of gene expression and a novel computational substraction approach, Genomics 75 (2001) 70–76.
M.D. Waters et al. / Mutation Research 544 (2003) 349–360 [58] N.A. Datson, J. van der Perk, E.R. de Kloet, E. Vreugdenhil, Expression profile of 30,000 genes in rat hippocampus using SAGE, Hippocampus 11 (2001) 430–444. [59] J. Colinge, G. Feger, Detecting the impact of sequencing errors on SAGE data, Bioinformatics 17 (2001) 840–842. [60] L. Cheval, B. Virlon, E. Billon, J.C. Aude, J.M. Elalouf, A. Doucet, Large-scale analysis of gene expression: methods and application to the kidney, J. Nephrol. 15 (Suppl. 5) (2002) S170–S183. [61] S.Q. Ye, T. Lavoie, D.C. Usher, L.Q. Zhang Microarray, SAGE and their applications to cardiovascular diseases, Cell Res. 12 (2002) 105–115. [62] T. Clark, S. Lee, L. Ridgway Scott, S.M. Wang, Computational analysis of gene identification with SAGE, J. Comput. Biol. 9 (2002) 513–526. [63] S.P. Mathupala, A.E. Sloan, In-gel purified ditags direct synthesis of highly efficient SAGE libraries, BMC Genomics 3 (2002) 20. [64] P. Liang, SAGE Genie: a suite with panoramic view of gene expression, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 11547– 11548. [65] J.M. Ruijter, A.H. Van Kampen, F. Baas, Statistical evaluation of SAGE libraries: consequences for experimental design, Physiol. Genomics 11 (2002) 37–44. [66] K.R. Boheler, M.D. Stern, The new role of SAGE in gene discovery, Trends Biotechnol. 21 (2003) 55–57. [67] M.D. Stern, S.V. Anisimov, K.R. Boheler, Can transcriptome size be estimated from SAGE catalogs, Bioinformatics 19 (2003) 443–448. [68] S. Kaneko, K. Kobayashi, Clinical application of a DNA chip in the field of liver diseases, J. Gastroenterol. 38 (Suppl. 15) (2003) 85–88. [69] C. Vilain, F. Libert, D. Venet, S. Costagliola, G. Vassart, Small amplified RNA–SAGE: an alternative approach to study transcriptome from limiting amount of mRNA, Nucl. Acids Res. 31 (2003) e24. [70] S.M. Gorski, S. Chittaranjan, E.D. Pleasance, J.D. Freeman, C.L. Anderson, R.J. Varhol, S.M. Coughlin, S.D. Zuyderduyn, S.J. Jones, M.A. Marra, A SAGE approach to discovery of genes involved in autophagic cell death, Curr. Biol. 13 (2003) 358–363. [71] M. Schena, D. Shalon, R.W. Davis, P.O. Brown, Quantitative monitoring of gene expression patterns with a complementary DNA microarray, Science 270 (1995) 467– 470. [72] D. Shalon, S.J. Smith, P.O. Brown, A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization, Genome Res. 6 (1996) 639– 645. [73] S.P. Fodor, J.L. Read, M.C. Pirrung, L. Stryer, A.T. Lu, D. Solas, Light-directed, spatially addressable parallel chemical synthesis, Science 251 (1991) 767–773. [74] M. Chee, R. Yang, E. Hubbell, A. Berno, X.C. Huang, D. Stern, J. Winkler, D.J. Lockhart, M.S. Morris, S.P. Fodor, Accessing genetic information with high-density DNA arrays, Science 274 (1996) 610–614. [75] J.G. Hacia, F.S. Collins, Mutational analysis using oligonucleotide microarrays, J. Med. Genet. 36 (1999) 730–736.
359
[76] I. Shmulevich, J. Astola, D. Cogdell, S.R. Hamilton, W. Zhang, Data extraction from composite oligonucleotide microarrays, Nucl. Acids Res. 31 (2003) e36. [77] Y. Chen, E.R. Dougherty, M.L. Bittner, Ratio-based decisions and the quantitative analysis of cDNA microarray images, J. Biomed. Opt. 2 (1997) 364–374. [78] T. Woodage, S. Broder, The human genome and comparative genomics: understanding human evolution, biology, and medicine, J. Gastroenterol. 38 (Suppl. 15) (2003) 68–77. [79] E. Phizicky, P.I. Bastiaens, H. Zhu, M. Snyder, S. Fields, Protein analysis on a proteomic scale, Nature 422 (2003) 208–215. [80] S.D. Patterson, R.H. Aebersold, Proteomics: the first decade and beyond, Nat. Genet. 33 (Suppl.) (2003) 311–323. [81] N.L. Anderson, R. Esquer-Blasco, J.P. Hofmann, N.G. Anderson, A two-dimensional gel database of rat liver proteins useful in gene regulation and drug effects studies, Electrophoresis 12 (1991) 907–930. [82] N.L. Anderson, R. Esquer-Blasco, J.P. Hofmann, L. Meheus, J. Raymackers, S. Steiner, F. Witzmann, N.G. Anderson, An updated two-dimensional gel database of rat liver proteins useful in gene regulation and drug effect studies, Electrophoresis 16 (1995) 1977–1981. [83] M. Mann, P. Hojrup, P. Roepstorff, Use of mass spectrometric molecular weight information to identify proteins in sequence databases, Biol. Mass Spectrom. 22 (1993) 338– 345. [84] R. Aebersold, M. Mann, Mass spectrometry-based proteomics, Nature 422 (2003) 198–207. [85] M.R. Wilkins, E. Gasteiger, J.C. Sanchez, R.D. Appel, D.F. Hochstrasser, Protein identification with sequence tags, Curr. Biol. 6 (1996) 1543–1544. [86] S.A. Chervitz, L. Aravind, G. Sherlock, C.A. Ball, E.V. Koonin, S.S. Dwight, M.A. Harris, K. Dolinski, S. Mohr, T. Smith, S. Weng, J.M. Cherry, D. Botstein, Comparison of the complete protein sets of worm and yeast: orthology and divergence, Science 282 (1998) 2022–2028. [87] R.S. Eisenberg, Genomics in the public domain: strategy and policy, Nat. Rev. Genet. 1 (2000) 70–74. [88] M.B. Eisen, P.T. Spellman, P.O. Brown, D. Botstein, Cluster analysis and display of genome-wide expression patterns, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 14863–14868. [89] P. Tamayo, D. Slonim, J. Mesirov, Q. Zhu, S. Kitareewan, E. Dmitrovsky, E.S. Lander, T.R. Golub, Interpreting patterns of gene expression with self-organizing maps: methods and application to hematopoietic differentiation, Proc. Natl. Acad. Sci. U.S.A. 96 (1999) 2907–2912. [90] S. Hanash, Disease proteomics, Nature 422 (2003) 226–232. [91] M.J. Aardema, J.T. MacGregor, Toxicology and genetic toxicology in the new era of “toxicogenomics”: impact of “-omics” technologies, Mutat. Res. 499 (2002) 13–25. [92] G.A. Boorman, S.P. Anderson, W.M. Casey, R.H. Brown, L.M. Crosby, K. Gottschalk, M. Easton, H. Ni, K.T. Morgan, Toxicogenomics, drug discovery, and the pathologist, Toxicol. Pathol. 30 (2002) 15–27. [93] H.K. Hamadeh, R.P. Amin, R.S. Paules, C.A. Afshari, An overview of toxicogenomics, Curr. Issues Mol. Biol. 4 (2002) 45–56.
360
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
[94] H.K. Hamadeh, P. Bushel, R. Paules, C.A. Afshari, Discovery in toxicology: mediation by gene expression array technology, J. Biochem. Mol. Toxicol. 15 (2001) 231–242. [95] Y. Hirabayashi, T. Inoue, Kokuritsu Iyakuhin Shokuhin Eisei Kenkyusho Hokoku, Toxicogenomics—a new Paradigm of Toxicology and Birth of Reverse Toxicology, Cellular and Molecular Toxicology Division, Center for Biological Safety and Research, National Institute of Health Sciences, Tokyo, Japan (Annual Report) 120 (2002) 39–52 (in Japanese). [96] E.F. Nuwaysir, M. Bittner, J. Trent, J.C. Barrett, C.A. Afshari, Microarrays and toxicology: the advent of toxicogenomics, Mol. Carcinog. 24 (1999) 153–159. [97] W.D. Pennie, I. Kimber, Toxicogenomics; transcript profiling and potential application to chemical allergy, Toxicol. In Vitro 16 (2002) 319–326. [98] J.C. Rockett, D.J. Dix, Application of DNA arrays to toxicology, Environ. Health Perspect. 107 (1999) 681–685. [99] T. Storck, M.C. von Brevern, C.K. Behrens, J. Scheel, A. Bach, Transcriptomics in predictive toxicology, Curr. Opin. Drug Discov. Devel. 5 (2002) 90–97. [100] R.W. Tennant, The National Center for Toxicogenomics: using new technologies to inform mechanistic toxicology, Environ. Health Perspect. 110 (2002) A8–A10. [101] R. Ulrich, S.H. Friend, Toxicogenomics and drug discovery: will new technologies help us produce better drugs, Nat. Rev. Drug Discov. 1 (2002) 84–88. [102] A.L. Castle, M.P. Carver, D.L. Mendrick, Toxicogenomics: a new revolution in drug safety, Drug Discov. Today 7 (2002) 728–736.
[103] T. Ideker, T. Galitski, L. Hood, A new approach to decoding life: systems biology, Annu. Rev. Genomics Hum. Genet. 2 (2001) 343–372. [104] M.D. Waters, G. Boorman, P. Bushel, M. Cunningham, R. Irwin, A. Merrick, K. Olden, R. Paules, J. Selkirk, S. Stasiewicz, B. Weis, B. Van Houten, N. Walker, R. Tennant, Systems toxicology and the chemical effects in biological systems knowledge base, Environ. Health Perspect. 111 (2003) 15–28. [105] O. Ermolaeva, M. Rastogi, K.D. Pruitt, G.D. Schuler, M.L. Bittner, Y. Chen, R. Simon, P. Meltzer, J.M. Trent, M.S. Boguski, Data management and analysis for gene expression arrays, Nat. Genet. 20 (1998) 19–23. [106] R.B. Hayes, Genetic susceptibility and occupational cancer, Med. Lav. 86 (1995) 206–213. [107] P. Grandjean, Individual susceptibility in occupational and environmental toxicology, Toxicol. Lett. 77 (1995) 105–108. [108] A. Hirvonen, Genetic factors in individual responses to environmental exposures, J. Occup. Environ. Med. 37 (1995) 37–43. [109] M.R. Spitz, Risk factors and genetic susceptibility, Cancer Treat. Res. 74 (1995) 73–87. [110] N. Rothman, R.B. Hayes, Using biomarkers of genetic susceptibility to enhance the study of cancer etiology, Environ. Health Perspect. 103 (Suppl. 8) (1995) 291– 295. [111] C.J. Henry, Evolution of toxicology for risk assessment, Int. J. Toxicol. 22 (2003) 3–7.
The impact of new technologies on human population studies Michael D. Waters a,∗ , James K. Selkirk a , Kenneth Olden b a
National Center for Toxicogenomics, 111 Alexander Drive, P.O. Box 12233, MD F1-05, Research Triangle Park, NC 27709-2233, USA b National Institute of Environmental Health Sciences, 111 Alexander Drive, P.O. Box 12233, MD F1-05, Research Triangle Park, NC 27709-2233, USA Received 5 May 2003; received in revised form 21 June 2003; accepted 23 June 2003
Abstract Human population studies involve clinical or epidemiological observations that associate environmental exposures with health endpoints and disease. Clearly, these are the most sought after data to support assessments of human health risk from environmental exposures. However, the foundations of many health risk assessments rest on experimental studies in rodents performed at high doses that elicit adverse outcomes, such as organ toxicity or tumors. Using the results of human studies and animal data, risk assessors define the levels of environmental exposures that may lead to disease in a portion of the population. These decisions on potential health risks are frequently based on the use of default assumptions that reflect limitations in our scientific knowledge. An important immediate goal of toxicogenomics, including proteomics and metabonomics, is to offer the possibility of making decisions affecting public health and public based on detailed toxicity, mechanistic, and exposure data in which many of the uncertainties have been eliminated. Ultimately, these global technologies will dramatically impact the practice of public health and risk assessment as applied to environmental health protection. The impact is already being felt in the practice of toxicology where animal experimentation using highly controlled dose–time parameters is possible. It is also being seen in human population studies where understanding human genetic variation and genomic reactions to specific environmental exposures is enhancing our ability to uncover the causes of variations in human response to environmental exposures. These new disciplines hold the promise of reducing the costs and time lines associated with animal and human studies designed to assess both the toxicity of environmental pollutants and efficacy of therapeutic drugs. However, as with any new science, experience must be gained before the promise can be fulfilled. Given the numbers and diversity of drugs, chemicals and environmental agents; the various species in which they are studied and the time and dose factors that are critical to the induction of beneficial and adverse effects, it is only through the development of a profound knowledge base that toxicology and environmental health can rapidly advance. The National Institute of Environmental Health Sciences (NIEHS), National Center for Toxicogenomics and its university-based Toxicogenomics Research Consortium (TRC), and resource contracts, are engaged in the development, application and standardization of the science upon which to the build such a knowledge base on Chemical Effects in Biological Systems (CEBS). In addition, the NIEHS Environmental Genome Project (EGP) is working to systematically identify and characterize common sequence polymorphisms in many genes with suspected roles in determining chemical sensitivity. The rationale of the EGP is that certain genes have a greater than average Abbreviations: AFLP, amplified fragment length polymorphism; CEBS, Chemical Effects in Biological Systems; EPA, US Environmental Protection Agency; EGP, Environmental Genome Project; IPCS, International Programme for Chemical Safety; NCT, National Center for Toxicogenomics; NIEHS, National Institute of Environmental Health Sciences; PCR, polymerase chain reaction; PB/PK, physiologically-based pharmacokinetic; RTI, reverse transcript imaging; SAGE, serial analysis of gene expression; SDS, sodium dodecyl sulfate; SNPs, single nucleotide polymorphisms; TRC, Toxicogenomics Research Consortium ∗ Corresponding author. Tel.: +1-919-316-4589; fax: +1-919-541-1460. E-mail address: [email protected] (M.D. Waters). 1383-5742/$ – see front matter © 2003 Published by Elsevier B.V. doi:10.1016/j.mrrev.2003.06.022
350
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
influence over human susceptibility to environmental agents. If we identify and characterize the polymorphism in those genes, we will increase our understanding of human disease susceptibility. This knowledge can be used to protect susceptible individuals from disease and to reduce adverse exposure and environmentally induced disease. © 2003 Published by Elsevier B.V. Keywords: Toxicogenomics; Microarray; Gene expression; Proteomics; Metabononics; Bioinformatics; Phenotypic anchoring; Phenotype; Molecular expression; Systems biology; Systems toxicology; Database; Knowledge base; Compendia; Ontologies; Global query; Sequence; Transcription factors; Single nucleotide polymorphisms
1. Introduction All individuals are continually exposed to hazardous agents and chemicals in their environment. Thus, there is a certain probability that each individual will suffer an adverse effect, such as disease, as a result of such exposure. This process has been conceptualized by environmental health researchers and expressed as a paradigm that describes the continuum between exposure and disease. Environmental exposures may be derived from several external sources including the environmental media (air, water, and soil), the diet, or from components specific to an individual’s workplace or occupation. The route of exposure varies (inhalation, oral, and dermal), and the concentration of hazardous compounds in the environment also varies over time. Importantly, each individual receives a unique internal dose of the compound, which is determined by the individual’s activities and movements while he or she is being exposed. Most biologically active agents have a much higher probability of causing an adverse effect in one or more “target tissues” in the human body (i.e., liver, skin, intestine, lung, blood, and bone). Thus, the amount of a compound that actually reaches the target tissue is described as the biologically effective dose. A compound can reach this target tissue at a significant concentration and still have no biological consequences, if the compound is efficiently removed or detoxified by cellular defense systems before cellular damage occurs. However, if cellular damage occurs, the damage can have a biological effect that leads to sustained or permanent changes in biological structures or functions. In many cases where no treatment or intervention is introduced, disease may develop and progress in these exposed individuals. A particular focus of this symposium is on the monitoring of human exposure, effects, and susceptibility to toxic agents and particularly genotoxic agents in the environment [1,2]. How can these studies in hu-
man populations assist us: (1) in measuring internal dose, (2) in understanding mode of action, (3) the etiology of environmentally-induced disease, and (4) in improving risk assessment methods and models? The focus of this paper is on new genomic and proteomic technologies and the impact that they can have in studies on human populations. In addition, we briefly consider their applications in risk assessment, with emphasis on detection of human susceptibility genes and genetic polymorphisms. Recent advancements in molecular profiling analysis using microarray and proteomics technologies provide tools for assessing expression of hundreds to thousands of genes in a single experiment. A molecular expression profile provides a complete molecular phenotype of the cell under a specific set of conditions and exposures. With these same tools comparative expression profiles can be generated for representative genomes in both experimental animals and in humans targeting functional gene clusters that represent a variety of physiological and toxicological processes. Among these genes and their corresponding proteins are the molecular biomarkers of the future. A new version of the familiar exposure-to-disease paradigm has been developed by National Institute of Environmental Health Sciences (NIEHS) Environmental Genome Project (EGP) and is reproduced below in Fig. 1. This illustration represents a genomics perspective on the familiar exposure-to-disease paradigm [3]. Such a perspective considers the environmental exposure including its distribution and metabolism, but it also considers its influence on signal transduction pathways, whether it causes genomic damage, and whether, for example, it affects cell differentiation or apoptosis genes, cell cycle control or repair genes. It asks about measurable, reproducible, alterations in gene expression related to the exposure, about the dose–response relationship, the temporal association and the sequence of molecular expression events. It
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
351
Fig. 1. Environmental genomics and the exposure-to-disease paradigm.
also asks about the influence of genetic variability, including single nucleotide polymorphisms (SNPs) on the human gene expression profile. The basic steps of the paradigm are conventional, but the potential information output in terms of toxicological information from global molecular expression studies is greatly increased.
2. Risk assessment The International Programme for Chemical Safety (IPCS) has produced [7] a generic “Framework for Risk Assessment” using many of the concepts presented in April 1996, the Proposed US Environmental Protection Agency (EPA) Cancer Risk Assessment Guidelines [4–6]. The 10 elements of the framework are listed below: • Toxicological or disease endpoint. • Postulated mode of action. • Key events—measurable events related to the mode of action (biomarkers). • Dose–response relationship for the key events. • Temporal association/sequence of events. • Strength, consistency and specificity of association of response with key events. • Biological plausibility and coherence. • Other modes of action.
• Assessment of postulated mode of action. • Uncertainties, inconsistencies and data gaps. Collectively, these elements give an idea of how molecular studies of exposure and effect can be used in risk assessment. The framework is applicable to all endpoints—to cancer as well as non-cancer effects. It places major emphasis on understanding the mode of action of the agent in question. It assigns a central role to “key events”—(molecular biomarkers in the context of the present discussion) that are related to the mode of action. These “key events” should ideally display a dose–response relationship. There should be a definable temporal association of the key events with the disease of concern, and the sequence of events should be consistent with the hypothesized mode of action. Clearly, this framework will require substantial basic and applied research to provide the information needed for risk assessment. It also provides a useful construct for our consideration of applications and impacts of toxicogenomics in human population studies, with subsequent impacts on human health and risk assessment. Using the results of human studies and animal data, risk assessors define the levels of environmental exposures that may lead to disease in a portion of the population. These decisions on potential health risks are frequently based on the use of default assumptions that reflect limitations in our
352
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
scientific knowledge [8]. An important immediate goal of toxicogenomics, including proteomics and metabonomics, is to offer the possibility of making decisions affecting public health and public based on detailed toxicity, mechanistic, and exposure data in which many of the uncertainties have been eliminated.
3. Genomics, proteomics, functional genomics We belatedly define the term “genomics.” An “-ome” is an abstract entity or group. Genomics is the study of genes in the aggregate—DNA, the primary RNA transcript and messenger RNA. Genomics is also used to refer to transcriptomics—the quantitative analysis of transcripts at the mRNA level. Expression data at the mRNA level can be produced using several technologies such as DNA microarrays, amplified fragment length polymorphism (AFLP) [9–18], reverse transcript imaging (RTI) or subtractive hybridization [19–32], serial analysis of gene expression (SAGE) [33–70] and others. For purposes of the present discussion SAGE and microarray analysis are most important. SAGE is a technique designed to take advantage of high-throughput sequencing technology to obtain a quantitative profile of cellular gene expression. Essentially, the SAGE technique measures not the expression level of a gene, but quantifies a “tag” which represents the transcription product of a gene. A tag, for the purposes of SAGE, is a nucleotide sequence of a defined length, directly 3 –adjacent to the 3 –most restriction site for a particular restriction enzyme. DNA microarrays have become a more popular and important approach offering great potential for human environmental studies. On a typical microarray, each gene of interest is represented either by a long DNA fragment (200–2400 bp) typically generated by polymerase chain reaction (PCR) and spotted on a suitable substrate using robotics [71,72], or by several short oligonucleotides (20–70 bp) synthesized directly onto a solid support using photolabile nucleotide chemistry [73–76]. Total RNA or mRNA is isolated from control and treated tissues, and reverse transcribed in the presence of radioactive or fluorescent labeled nucleotides, and the labeled probes are then hybridized to the arrays. The intensity of the array signal is measured for each gene transcript by laser scanning con-
focal microscopy. The ratio between the signals of control and treated samples reflect the relative drug or chemical-induced change in transcript abundance [77]. Current microarray technology allows the simultaneous expression monitoring of 20,000–25,000 genes. In the immediate future, when all human genes have been identified and cloned [78], and technological advances allow further miniaturization of solid supports we will have the capability to examine simultaneously all expressed genes in the human genome. Proteomics [79,80] is defined as the study of protein products in aggregate—it applies to the translation from the mRNA to the primary protein products, and their maturation and modification to yield active proteins as components of a cell, tissue, or organism. The aim of proteomics is to quantify gene expression down-stream, to obtain a “snapshot” of gene regulation in the actual control of cell function. Proteins generally display greater stability than mRNA. The technical problem is that global quantitative expression analysis at the protein level has largely been restricted to the use of 2D gel electrophoresis. In this technique, isoelectric focusing in the first dimension is followed by SDS slab gel electrophoresis-based molecular weight separation [81,82]. Protein spots are identified by mass spectrometry following generation of peptide mass fingerprints [83,84] and sequence tags [85]. Functional genomics or computational biology is the systematic analysis of gene products, including both mRNAs and proteins to determine their function. Gene function can be inferred from sequence homology among predicted proteins, where the function of at least one protein is known [86]. Computational methods have been developed that detect functional linkages among predicted proteins [87]. But sequence analysis alone is insufficient to fully inform us about gene function. The traditional microarray-based approach to functional analysis has been to cluster genes according to their expression behavior under a range of conditions and to assign function according to the set of known genes that fall into these clusters [88,89]. The premise with this “guilt-by-association” approach is that clustered genes will be co-regulated and, therefore, be involved in similar functions. There are several additional methods that do not require clustering to analyze gene expression profiles and predict gene function. The limitations are mainly in the models, and experience necessary to analyze large datasets.
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
Despite the challenges, expression profiles provide a complete molecular phenotype of the cell under a specific set of conditions and have numerous applications in studies on disease etiology [90].
4. Toxicogenomics Toxicogenomics is the study of the response of a genome to environmental stressors and toxicants. It combines genetics, genomic-scale mRNA expression (transcriptomics), cell and tissue-wide protein expression (proteomics), metabolite profiling (metabonomics) and bioinformatics with conventional toxicology in an effort to understand the role of gene–environment interactions in disease. Toxicogenomics seeks to identify and characterize mechanisms of action of known and suspected toxicants [91–102]. The analysis of gene expression patterns for different chemicals under different doses or times of exposure scenarios is used to gain a clearer understanding of genes that are mechanistically linked to a toxicologic response. This process, including the development, perturbation and reassessment of models that relate gene expression profiles to toxic outcomes is referred to as “systems toxicology” [103,104]. An important immediate goal of toxicogenomics, including proteomics and metabonomics, is to offer the possibility of making decisions affecting public health and public based on detailed toxicity, mechanistic, and exposure data in which many of the uncertainties have been eliminated. It will be important to determine how many and which genes should be measured to characterize a toxic response and distinguish it from physiologically adaptive responses that are not linked to toxicity. The use of global gene expression data in hazard identification in the absence of a correct interpretation of the toxicological significance of the data would be unfortunate. Using a combination of laboratory and field studies, comparison of chemicals from different mode of action classes (for example, cytotoxic chemicals, peroxisome proliferators, or estrogenic chemicals) will allow the identification of groups of genes whose expression at multiple doses and times is consistently linked to specific exposures and disease outcomes. The outcome of environmental exposures is influenced by the function of many human genes. Not all
353
of these genes or their functions are known at present. However, many genes have been identified that are likely to be important factors in genetic susceptibility to environmentally induced disease. These genes tend to fall into the following categories: cell cycle control, DNA repair, regulation of cell division, cell signaling, cell structure, apoptosis and metabolism. Several genes that control metabolic pathways are crucial determinants of the outcome of exposure. In many cases, a compound enters a biological system in an inert innocuous form that is metabolically converted into a reactive species that causes cellular damage. The converse is also true; some metabolic pathways destroy toxic compounds by changing their chemical structure, frequently making them more water soluble and easier to excrete. Cell cycle and cell division genes regulate the ability of a cell to proliferate, grow and differentiate. Changes in the progression of a cell through the cell cycle can increase the ability of a cell to survive stress; in most cases, a proliferating cell exposed to stress will enhance its survival by delaying the cell cycle so that cellular damage can be repaired prior to cell division. Cell signaling and gene expression pathways have profound effects on all cellular functions, including cell proliferation and differentiation. Some exogenous agents can activate these pathways in aberrant and deleterious ways (i.e., agents that mimic a biological component), and this can disrupt or alter normal cellular function. DNA repair genes influence the outcome of exposure to environmental agents that cause DNA damage. Individuals with higher or lower capacity for DNA repair have decreased or increased risk of certain types of environmentally induced disease, respectively. Heavily damaged cells often die by a process known as programmed cell death or apoptosis. This process protects the organism by removing aberrant cells and damaged structures. To obtain the most relevant data from gene expression studies with microarrays requires that experiments be performed at multiple doses and varied exposure durations to identify those genes clearly linked to a toxic response. To develop approaches that will maximize the likelihood of detecting true positives for human exposure and minimize false negatives, a substantial matrix of data on chemicals with known exposure-disease outcomes will need to be obtained. This will require the evaluation of gene expression profiles of chemicals not causing health
354
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
effects as well as those known to cause disease and positive chemicals with varying potency for causing disease. When the database on known chemicals becomes comprehensive enough, the fingerprints of new chemicals with unknown toxic properties can be compared to well-characterized gene expression profiles, allowing the new chemical to be provisionally placed into one or more mode of action classes. Then more directed studies may be undertaken to confirm or refute the predicted mode of action and toxic outcome for the new chemical. As with any new science, experience must be gained in toxicogenomics before its promise can be fulfilled. This experience will involve the assembly of massive amounts of data. The difficulty is that each of the new technological approaches being applied in genome and proteome analysis can overwhelm the current information infrastructure. The challenge is not simply managing the data flow that will be generated by these new approaches. New models will be needed to manipulate the data and new analytic strategies will be required to interpret it. The NIEHS National Center for Toxicogenomics (NCT) is working to help the field of environmental health research evolve into a knowledge-based science in which experimental toxicogenomics and clinical data are compiled. The NCT (http://www.niehs.nih. gov/nct/) and its university-based Toxicogenomics Research Consortium (TRC, http://www.niehs.nih.gov/ nct/trc.htm) and resource contracts are engaged in the development, application and standardization of the
science upon which to the build such a knowledge base (Fig. 2) on Chemical Effects in Biological Systems (CEBS, http://www.niehs.nih.gov/nct/cebs.htm). CEBS [104] will be created as a high quality, publicly accessible relational database that is compatible with standard laboratory output platforms. Database development will be integrated with strategic toxicogenomics experimental design and conduct. Standardized procedures, protocols, data formats, and assessment methods will be used to ensure that data meet a uniform high level of quality. Raw data sets from NCT experiments will be available in their entirety. Relational and descriptive compendia will be included on toxicologically important genes, groups of genes, SNPs, and mutants and their functional phenotypes. Information about the biological effects of chemicals and other agents and their mechanism of action will be collected from the literature and stored. CEBS will be fully searchable by compound, structure, toxicity, pathology, gene, gene group, SNP, pathway, and network. Dictionaries and explanatory text will guide researchers in understanding toxicogenomics datasets. CEBS will be linked extensively to other databases and to Web genomics and proteomics resources, providing users the suite of information and tools needed to fully interpret toxicogenomics data. Computational and informatics tools will play a significant role in improving our understanding of toxicant-related disease by creating a system of predictive toxicology.
Fig. 2. Conceptual framework for the CEBS knowledge base.
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
Future promises of CEBS include: 1. Developing large context-annotated datasets that allow precise definition of biological/toxicological pathways and lead to the identification of new biomarkers. 2. Linking genomic sequence to expression data to determine those genes that may be responsible for the coordinated regulation of sets of genes. 3. Increasing the interpretability and dimensionality of expression data by including data from new types of arrays including protein arrays and SNP arrays. 4. Aiding in development of new algorithms and computational tools that allow predictive modeling of gene interactions and networks relevant to human health and disease. 5. Impact on human population studies A challenge for environmental toxicologists in applying these new technologies in studies in humans is to better determine the level of chemical exposure. This includes actual measures of exposure or dose and total human exposure across all relevant pathways. Exposure assessment is the component of risk assessment that is most neglected especially by biologists. The measurement of human environmental exposure is a complicated business. There are many variables that influence what amount of an exposure will become a target tissue dose. All of them come into play before the opportunity to induce an adverse outcome. Exposure-related databases that report actual measures of exposure or dose are very limited and virtually none collect measures of total human exposure across all relevant environmental pathways. Gaining a better estimate of exposure and internal dose is an important direction and challenge in future toxicogenomics research. Accurate exposure information will be essential to understand environmentally induced disease and critical to support the needs of human health risk assessment. An area of concern in approaching this challenge is that assessment of exposure in people who do not have a pre-characterized baseline molecular expression will be difficult. This is because baseline expression among individuals is expected to vary widely with differences in diet, personal habits, and health status. Because of these baseline differences, changes
355
due to environmental chemical exposure may not be greater than the noise of gene expression variability. Thus, in the future, gene expression profiles from exposed individuals will probably be compared to a standard gene expression profile or profiles based on a national or international gene expression database for humans [105]. Assuming that the scientific (as well as ethical) issues of human gene expression databases are resolved, microarray analysis of gene expression patterns could be used to identify the type of chemical exposure experienced by individuals and to confirm its linkage to a mode of action class as described previously. Another major application of gene expression profiling is to understand human susceptibility to disease. Experimental studies in molecular genetics, toxicology and other biomedical fields have shown that genetic differences between two individuals can determine the relative sensitivity of each individual to environmental chemicals or agents. Some human inherited disease susceptibilities are caused by a single inherited trait, and other disease susceptibilities may be determined by multiple genetic traits [106,107]. For many diseases, the genetic complement of the individual does not itself cause disease; instead, disease is the outcome of a complex interplay between multiple genetic and environmental factors [108–110]. The National Institute of Environmental Health Sciences (NIEHS) in Research Triangle Park, NC has initiated the Environmental Genome Project (EGP, http://www.niehs.nih.gov/envgenom/) to systematically identify and characterize common sequence polymorphisms in many genes with suspected roles in determining chemical sensitivity. The rationale of the EGP is that certain genes have a greater influence over human susceptibility to environmental agents than others. If we identify and characterize the polymorphism in those genes, we will increase our understanding of human disease susceptibility. This knowledge can be used to protect susceptible individuals from disease and to reduce adverse exposure and environmentally induced disease.
6. Human polymorphisms There may be as many as 10–30 million SNPs in the human genome, if all variants in all individuals are
356
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
counted, and only about 1% of them alter an amino acid in a protein encoded by a gene. While most SNPs do not directly affect protein function, some affect function indirectly, through alteration of regulatory sequences that control gene expression, or the stability or processing of the mRNA transcript of the gene. It is important to note that the frequency of a polymorphism can vary in different population subgroups. In other words, polymorphic sites can be common or rare in the sequence of different population groups. The common polymorphic sites are found in one of five individuals or higher, but most variants are represented at much lower frequencies. EGP resequencing studies are designed to effectively detect polymorphic sites that are represented at 1% or higher in the sampled population. The EGP has identified a group of human genes that are likely to influence the outcome of environmental exposure. Polymorphic variants in these “environmentally responsive genes” are being identified by systematic resequencing of a predefined diverse, randomly selected set of human DNA samples. The DNA resource resource is maintained by the Coriell Institute for Medical Research and is part of the Coriell Cell Repositories. The repository includes cell lines and DNA samples from a panel of 450 unrelated male and female individuals. The samples are from individuals who belong to diverse ethnic groups including European–Americans, African–Americans, Mexican– Americans, native Americans, and Asian–Americans. Candidate genes are currently being sequenced across the entire panel of 450 individuals to identify common sequence variation for functional analysis and population-based studies. Automated DNA sequencing is being used to identify and genotype SNPs in human candidate genes. All SNPs have been identified using only high quality sequence data (Q > 25) and each SNP reported from the NIEHS-SNPs program has been confirmed in multiple individuals and/or in multiple reactions. The EGP has generated a list of 554 environmentally responsive genes that are potential targets for re-sequencing in subgroups according to a phased timeline. The list is not comprehensive and it is expected that more genes may be added over time. The first phase of resequencing extended from 1998 to 2001 and has been focused on polymorphic variants on a group of 123 genes. The second phase of rese-
quencing, which extends from 2001 to 2004, focuses on approximately 200 DNA repair and cell cycle control genes. Genes regulating metabolism, signal transduction, and apoptosis will also be targeted for resequencing. The project will develop a central database of polymorphisms for these genes, and foster population-based studies of gene–environment interaction in disease etiology. The functional significance of specific gene variants is also being characterized to determine which gene variants are correlated with increased or decreased risk of disease. NIEHS SNPs are available in the GeneSNPs database (http://www.genome.utah.edu/genesnps/) as well as the national database resource, dbSNP (http:// www.niehs.nih.gov/envgenom/snpsdb.htm). GeneSNPs provides a gene-centric map of the genome structure, coding sequences, and identified allelic variation in genes being targeted for a role in disease susceptibility by the NIEHS. This database provides a graphical view of all available SNP data including allele frequencies and genotypes in select populations. This information is key in selecting the polymorphic sites needed to examine disease risk in human population studies. 7. Conclusions Clearly, these sequence-based global technologies and the databases that result from their application will dramatically impact the practice of public health and risk assessment [111] as applied to environmental health protection. These new approaches hold the promise of reducing the costs and time lines associated with animal and human studies designed to assess both the toxicity of environmental pollutants and efficacy of therapeutic drugs. More importantly, they promise to dramatically improve the fidelity of the information that will be available from animal as well as toxicogenomics studies in human populations. These studies will ultimately find their way into the risk assessment process and will play an important role in the protection of human environmental health. Acknowledgements We thank Dr. Douglas Bell for valuable discussions on human epidemiological studies. We thank Nancy
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
Stegman, Jerry Nehls for assistance on information technology. We are indebted to Samuel Wilson, and Lutz Birnbaumer for their continuing support of the NCT program. References [1] R.J. Albertini, Biomarker responses in human populations: towards a worldwide map, Mutat. Res. 428 (1999) 217–226. [2] R.J. Albertini, J.A. Nicklas, J.P. O’Neill, Future research directions for evaluating human genetic and cancer risk from environmental exposures, Environ. Health Perspect. 104 (Suppl. 3) (1996) 503–510. [3] D.A. Bennett, M.D. Waters, Applying biomarker research, Environ. Health Perspect. 108 (2000) 907–910. [4] W.H. Farland, Cancer risk assessment: evolution of the process, Prev. Med. 25 (1996) 24–25. [5] V.L. Dellarco, J.A. Wiltse, US Environmental Protection Agency’s Revised Guidelines for carcinogen risk assessment: incorporating mode of action data, Mutat. Res. 405 (1998) 273–277. [6] M. Andersen, D. Brusick, S. Cohen, Y. Dragan, C. Frederick, J.I. Goodman, G. Hard, B. Meek, E.J. O’Flaherty, US Environmental Protection Agency’s Revised Cancer Guidelines for carcinogen risk assessment, Toxicol. Appl. Pharmacol. 153 (1998) 133–136. [7] C. Sonich-Mullin, R. Fielder, J. Wiltse, K. Baetcke, J. Dempsey, P. Fenner-Crisp, D. Grant, M. Hartley, A. Knaap, D. Kroese, I. Mangelsdorf, E. Meek, J.M. Rice, M. Younes, IPCS conceptual framework for evaluating a mode of action for chemical carcinogenesis, Regul. Toxicol. Pharmacol. 34 (2001) 146–152. [8] M.E. Andersen, M.E. Meek, G.A. Boorman, D.J. Brusick, S.M. Cohen, Y.P. Dragan, C.B. Frederick, J.I. Goodman, G.C. Hard, E.J. O’Flaherty, D.E. Robinson, Lessons learned in applying the US EPA proposed cancer guidelines to specific compounds, Toxicol. Sci. 53 (2000) 159–172. [9] P. Vos, R. Hogers, M. Bleeker, M. Reijans, T. van de Lee, M. Hornes, A. Frijters, J. Pot, J. Peleman, M. Kuiper, et al., AFLP: a new technique for DNA fingerprinting, Nucl. Acids Res. 23 (1995) 4407–4414. [10] T. Money, S. Reader, L.J. Qu, R.P. Dunford, G. Moore, AFLP-based mRNA fingerprinting, Nucl. Acids Res. 24 (1996) 2616–2617. [11] M. Otsen, M. den Bieman, M.T. Kuiper, M. Pravenec, V. Kren, T.W. Kurtz, H.J. Jacob, A. Lankhorst, B.F. van Zutphen, Use of AFLP markers for gene mapping and QTL detection in the rat, Genomics 37 (1996) 289–294. [12] B.L. Roman, V.N. Pham, P.E. Bennett, B.M. Weinstein, Non-radioisotopic AFLP method using PCR primers fluorescently labeled with Cy5, Biotechniques 26 (1999) 236–238. [13] J.J. Lin, J. Ma, J. Kuo, et al., Chemiluminescent detection of AFLP markers, Biotechniques 26 (1999) 344–348. [14] U.G. Mueller, L.L. Wolfenbarger, AFLP genotyping and fingerprinting, Trends Ecol. Evol. 14 (1999) 389–394.
357
[15] S. Iyoda, A. Wada, J. Weller, S.J. Flood, E. Schreiber, B. Tucker, H. Watanabe, Evaluation of AFLP, a high-resolution DNA fingerprinting method, as a tool for molecular subtyping of enterohemorrhagic Escherichia coli O157:H7 isolates, Microbiol. Immunol. 43 (1999) 803–806. [16] A. Reineke, P. Karlovsky, Simplified AFLP protocol: replacement of primer labeling by the incorporation of alpha-labeled nucleotides during PCR, Biotechniques 28 (2000) 622–623. [17] P.E. Klein, R.R. Klein, S.W. Cartinhour, P.E. Ulanch, J. Dong, J.A. Obert, D.T. Morishige, S.D. Schlueter, K.L. Childs, M. Ale, J.E. Mullet, A high-throughput AFLP-based method for constructing integrated genetic and physical maps: progress toward a sorghum genome map, Genome Res. 10 (2000) 789–807. [18] M. Prochazka, K. Walder, J. Xia, AFLP fingerprinting of the human genome, Hum. Genet. 108 (2001) 59–65. [19] T. Nakagawa, M. Mabry, A. de Bustros, J.N. Ihle, B.D. Nelkin, S.B. Baylin, Introduction of v-Ha-ras oncogene induces differentiation of cultured human medullary thyroid carcinoma cells, Proc. Natl. Acad. Sci. U.S.A. 84 (1987) 5923–5927. [20] G. Tarone, M.A. Russo, E. Hirsch, T. Odorisio, F. Altruda, L. Silengo, G. Siracusa, Expression of beta-1 integrin complexes on the surface of unfertilized mouse oocyte, Development 117 (1993) 1369–1375. [21] C.M. Niessen, O. Cremona, H. Daams, S. Ferraresi, A. Sonnenberg, P.C. Marchisio, Expression of the integrin alpha-6 beta-4 in peripheral nerves: localization in Schwann and perineural cells and different variants of the beta-4 subunit, J. Cell Sci. 107 (Part 2) (1994) 543–552. [22] A.K. Stuart-Tilley, B.E. Shmukler, D. Brown, S.L. Alper, Immunolocalization and tissue-specific splicing of AE2 anion exchanger in mouse kidney, J. Am. Soc. Nephrol. 9 (1998) 946–959. [23] F. Mannello, M. Malatesta, F. Luchetti, S. Papa, S. Battistelli, G. Gazzanelli, Immunoreactivity, ultrastructural localization, and transcript expression of prostate-specific antigen in human neuroblastoma cell lines, Clin. Chem. 45 (1999) 78– 84. [24] S. Alcolea, M. Theveniau-Ruissy, T. Jarry-Guichard, I. Marics, E. Tzouanacou, J.P. Chauvin, J.P. Briand, A.F. Moorman, W.H. Lamers, D.B. Gros, Downregulation of connexin 45 gene products during mouse heart development, Circ. Res. 84 (1999) 1365–1379. [25] Y. Kiyozuka, H. Miyazaki, K. Yoshizawa, H. Senzaki, D. Yamamoto, K. Inoue, K. Bessho, Y. Okubo, K. Kusumoto, A. Tsubura, An autopsy case of malignant mesothelioma with osseous and cartilaginous differentiation: bone morphogenetic protein-2 in mesothelial cells and its tumor, Dig. Dis. Sci. 44 (1999) 1626–1631. [26] D.C. Cross, J.P. Munoz, P. Hernandez, R.B. Maccioni, Nuclear and cytoplasmic tau proteins from human nonneuronal cells share common structural and functional features with brain tau, J. Cell Biochem. 78 (2000) 305–317. [27] C.G. Carlotti Jr., J.M. Drake, J.P. Hladky, I. Teshima, L.E. Becker, J.T. Rutka, Primary Ewing’s sarcoma of the skull
358
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
M.D. Waters et al. / Mutation Research 544 (2003) 349–360 in children: utility of molecular diagnostics, surgery and adjuvant therapies, Pediatr. Neurosurg. 31 (1999) 307–315. R. Leemans, B. Egger, T. Loop, L. Kammermeier, H. He, B. Hartmann, U. Certa, F. Hirth, H. Reichert, Quantitative transcript imaging in normal and heat-shocked Drosophila embryos by using high-density oligonucleotide arrays, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 12138–12143. M.J. O’Sullivan, E.J. Perlman, J. Furman, P.A. Humphrey, L.P. Dehner, J.D. Pfeifer, Visceral primitive peripheral neuroectodermal tumors: a clinicopathologic and molecular study, Hum. Pathol. 32 (2001) 1109–1115. D.K. Graham, L.C. Stork, Q. Wei, J.D. Ingram, F.M. Karrer, G.W. Mierau, M.A. Lovell, Molecular genetic analysis of a small bowel primitive neuroectodermal tumor, Pediatr. Dev. Pathol. 5 (2002) 86–90. M.C. Rissoan, T. Duhen, J.M. Bridon, N. Bendriss-Vermare, C. Peronne, B. de Saint Vis, F. Briere, E.E. Bates, Subtractive hybridization reveals the expression of immunoglobulin-like transcript 7, Eph-B1, granzyme B, and three novel transcripts in human plasmacytoid dendritic cells, Blood 100 (2002) 3295–3303. A. Borchert, N.E. Savaskan, H. Kuhn, Regulation of expression of the phospholipid hydroperoxide/sperm nucleus glutathione peroxidase gene: tissue-specific expression pattern and identification of functional cis- and transregulatory elements, J. Biol. Chem. 278 (2003) 2571–2580. V.E. Velculescu, L. Zhang, B. Vogelstein, K.W. Kinzler, Serial analysis of gene expression, Science 270 (1995) 484– 487. S.L. Madden, E.A. Galella, J. Zhu, A.H. Bertelsen, G.A. Beaudry, SAGE transcript profiles for p53-dependent growth regulation, Oncogene 15 (1997) 1079–1085. J. Powell, Enhanced concatemer cloning—a modification to the SAGE (Serial Analysis of Gene Expression) technique, Nucl. Acids Res. 26 (1998) 3445–3446. S.M. Wang, J.D. Rowley, A strategy for genome-wide gene analysis: integrated procedure for gene identification, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 11909–11914. N.A. Datson, J. van der Perk-de Jong, M.P. van den Berg, E.R. de Kloet, E. Vreugdenhil, MicroSAGE: a modified procedure for serial analysis of gene expression in limited amounts of tissue, Nucl. Acids Res. 27 (1999) 1300–1307. F. Baas, H.F. Tabak, A tale of tags: report on a HUGO/EU SAGE workshop, 29 January–1 February 1999, Hilversum, The Netherlands, Eur. J. Hum. Genet. 7 (1999) 510–512. A. van den Berg, J. van der Leij, S. Poppema, Serial analysis of gene expression: rapid RT-PCR analysis of unknown SAGE tags, Nucl. Acids Res. 27 (1999) e17. D.G. Peters, A.B. Kassam, H. Yonas, E.H. O’Hare, R.E. Ferrell, A.M. Brufsky, Comprehensive transcript analysis in small quantities of mRNA by SAGE-lite, Nucl. Acids Res. 27 (1999) e39. A. Lal, A.E. Lash, S.F. Altschul, V. Velculescu, L. Zhang, R.E. McLendon, M.A. Marra, C. Prange, P.J. Morin, K. Polyak, N. Papadopoulos, B. Vogelstein, K.W. Kinzler, R.L. Strausberg, G.J. Riggins, A public database for gene expression in human cancers, Cancer Res. 59 (1999) 5403– 5407.
[42] M. Larsson, S. Stahl, M. Uhlen, A. Wennborg, Expression profile viewer (ExProView): a software tool for transcriptome analysis, Genomics 63 (2000) 341–353. [43] J.J. Going, B.A. Gusterson, Molecular pathology and future developments, Eur. J. Cancer 35 (1999) 1895–1904. [44] A.E. Lash, C.M. Tolstoshev, L. Wagner, G.D. Schuler, R.L. Strausberg, G.J. Riggins, S.F. Altschul, SAGEmap: a public gene expression resource, Genome Res. 10 (2000) 1051– 1060. [45] J. Powell, SAGE: the serial analysis of gene expression, Methods Mol. Biol. 99 (2000) 297–319. [46] J. Stollberg, J. Urschitz, Z. Urban, C.D. Boyd, A quantitative evaluation of SAGE, Genome Res. 10 (2000) 1241–1248. [47] S.L. Madden, C.J. Wang, G. Landes, Serial analysis of gene expression: from gene discovery to target identification, Drug Discov. Today 5 (2000) 415–425. [48] M. Ishii, S. Hashimoto, S. Tsutsumi, Y. Wada, K. Matsushima, T. Kodama, H. Aburatani, Direct comparison of GeneChip and SAGE on the quantitative accuracy in transcript profiling analysis, Genomics 68 (2000) 136–143. [49] E.H. Margulies, J.W. Innis, eSAGE: managing and analysing data generated with serial analysis of gene expression (SAGE), Bioinformatics 16 (2000) 650–651. [50] V.E. Velculescu, B. Vogelstein, K.W. Kinzler, Analysing uncharted transcriptomes with SAGE, Trends Genet. 16 (2000) 423–425. [51] S.Q. Ye, L.Q. Zhang, F. Zheng, D. Virgil, P.O. Kwiterovich, miniSAGE: gene expression profiling using serial analysis of gene expression from 1 g total RNA, Anal. Biochem. 287 (2000) 144–152. [52] A.H. van Kampen, B.D. van Schaik, E. Pauws, E.M. Michiels, J.M. Ruijter, H.N. Caron, R. Versteeg, S.H. Heisterkamp, J.A. Leunissen, F. Baas, M. van der Mee, USAGE: a web-based approach towards the analysis of SAGE data (Serial Analysis of Gene Expression), Bioinformatics 16 (2000) 899–905. [53] H. Caron, B. van Schaik, M. van der Mee, F. Baas, G. Riggins, P. van Sluis, M.C. Hermus, R. van Asperen, K. Boon, P.A. Voute, S. Heisterkamp, A. van Kampen, R. Versteeg, The human transcriptome map: clustering of highly expressed genes in chromosomal domains, Science 291 (2001) 1289–1292. [54] J. Bartlett, Technology evaluation: SAGE, genzyme molecular oncology, Curr. Opin. Mol. Ther. 3 (2001) 85–96. [55] C. Stoeckert, A. Pizarro, E. Manduchi, M. Gibson, B. Brunk, J. Crabtree, J. Schug, S. Shen-Orr, G.C. Overton, A relational schema for both array-based and SAGE gene expression experiments, Bioinformatics 17 (2001) 300–308. [56] E.H. Margulies, S.L. Kardia, J.W. Innis, Identification and prevention of a GC content bias in SAGE libraries, Nucl. Acids Res. 29 (2001) E60–60. [57] J.C. Moreno, E. Pauws, A.H. van Kampen, M. Jedlickova, J.J. de Vijlder, C. Ris-Stalpers, Cloning of tissue-specific genes using serial analysis of gene expression and a novel computational substraction approach, Genomics 75 (2001) 70–76.
M.D. Waters et al. / Mutation Research 544 (2003) 349–360 [58] N.A. Datson, J. van der Perk, E.R. de Kloet, E. Vreugdenhil, Expression profile of 30,000 genes in rat hippocampus using SAGE, Hippocampus 11 (2001) 430–444. [59] J. Colinge, G. Feger, Detecting the impact of sequencing errors on SAGE data, Bioinformatics 17 (2001) 840–842. [60] L. Cheval, B. Virlon, E. Billon, J.C. Aude, J.M. Elalouf, A. Doucet, Large-scale analysis of gene expression: methods and application to the kidney, J. Nephrol. 15 (Suppl. 5) (2002) S170–S183. [61] S.Q. Ye, T. Lavoie, D.C. Usher, L.Q. Zhang Microarray, SAGE and their applications to cardiovascular diseases, Cell Res. 12 (2002) 105–115. [62] T. Clark, S. Lee, L. Ridgway Scott, S.M. Wang, Computational analysis of gene identification with SAGE, J. Comput. Biol. 9 (2002) 513–526. [63] S.P. Mathupala, A.E. Sloan, In-gel purified ditags direct synthesis of highly efficient SAGE libraries, BMC Genomics 3 (2002) 20. [64] P. Liang, SAGE Genie: a suite with panoramic view of gene expression, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 11547– 11548. [65] J.M. Ruijter, A.H. Van Kampen, F. Baas, Statistical evaluation of SAGE libraries: consequences for experimental design, Physiol. Genomics 11 (2002) 37–44. [66] K.R. Boheler, M.D. Stern, The new role of SAGE in gene discovery, Trends Biotechnol. 21 (2003) 55–57. [67] M.D. Stern, S.V. Anisimov, K.R. Boheler, Can transcriptome size be estimated from SAGE catalogs, Bioinformatics 19 (2003) 443–448. [68] S. Kaneko, K. Kobayashi, Clinical application of a DNA chip in the field of liver diseases, J. Gastroenterol. 38 (Suppl. 15) (2003) 85–88. [69] C. Vilain, F. Libert, D. Venet, S. Costagliola, G. Vassart, Small amplified RNA–SAGE: an alternative approach to study transcriptome from limiting amount of mRNA, Nucl. Acids Res. 31 (2003) e24. [70] S.M. Gorski, S. Chittaranjan, E.D. Pleasance, J.D. Freeman, C.L. Anderson, R.J. Varhol, S.M. Coughlin, S.D. Zuyderduyn, S.J. Jones, M.A. Marra, A SAGE approach to discovery of genes involved in autophagic cell death, Curr. Biol. 13 (2003) 358–363. [71] M. Schena, D. Shalon, R.W. Davis, P.O. Brown, Quantitative monitoring of gene expression patterns with a complementary DNA microarray, Science 270 (1995) 467– 470. [72] D. Shalon, S.J. Smith, P.O. Brown, A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization, Genome Res. 6 (1996) 639– 645. [73] S.P. Fodor, J.L. Read, M.C. Pirrung, L. Stryer, A.T. Lu, D. Solas, Light-directed, spatially addressable parallel chemical synthesis, Science 251 (1991) 767–773. [74] M. Chee, R. Yang, E. Hubbell, A. Berno, X.C. Huang, D. Stern, J. Winkler, D.J. Lockhart, M.S. Morris, S.P. Fodor, Accessing genetic information with high-density DNA arrays, Science 274 (1996) 610–614. [75] J.G. Hacia, F.S. Collins, Mutational analysis using oligonucleotide microarrays, J. Med. Genet. 36 (1999) 730–736.
359
[76] I. Shmulevich, J. Astola, D. Cogdell, S.R. Hamilton, W. Zhang, Data extraction from composite oligonucleotide microarrays, Nucl. Acids Res. 31 (2003) e36. [77] Y. Chen, E.R. Dougherty, M.L. Bittner, Ratio-based decisions and the quantitative analysis of cDNA microarray images, J. Biomed. Opt. 2 (1997) 364–374. [78] T. Woodage, S. Broder, The human genome and comparative genomics: understanding human evolution, biology, and medicine, J. Gastroenterol. 38 (Suppl. 15) (2003) 68–77. [79] E. Phizicky, P.I. Bastiaens, H. Zhu, M. Snyder, S. Fields, Protein analysis on a proteomic scale, Nature 422 (2003) 208–215. [80] S.D. Patterson, R.H. Aebersold, Proteomics: the first decade and beyond, Nat. Genet. 33 (Suppl.) (2003) 311–323. [81] N.L. Anderson, R. Esquer-Blasco, J.P. Hofmann, N.G. Anderson, A two-dimensional gel database of rat liver proteins useful in gene regulation and drug effects studies, Electrophoresis 12 (1991) 907–930. [82] N.L. Anderson, R. Esquer-Blasco, J.P. Hofmann, L. Meheus, J. Raymackers, S. Steiner, F. Witzmann, N.G. Anderson, An updated two-dimensional gel database of rat liver proteins useful in gene regulation and drug effect studies, Electrophoresis 16 (1995) 1977–1981. [83] M. Mann, P. Hojrup, P. Roepstorff, Use of mass spectrometric molecular weight information to identify proteins in sequence databases, Biol. Mass Spectrom. 22 (1993) 338– 345. [84] R. Aebersold, M. Mann, Mass spectrometry-based proteomics, Nature 422 (2003) 198–207. [85] M.R. Wilkins, E. Gasteiger, J.C. Sanchez, R.D. Appel, D.F. Hochstrasser, Protein identification with sequence tags, Curr. Biol. 6 (1996) 1543–1544. [86] S.A. Chervitz, L. Aravind, G. Sherlock, C.A. Ball, E.V. Koonin, S.S. Dwight, M.A. Harris, K. Dolinski, S. Mohr, T. Smith, S. Weng, J.M. Cherry, D. Botstein, Comparison of the complete protein sets of worm and yeast: orthology and divergence, Science 282 (1998) 2022–2028. [87] R.S. Eisenberg, Genomics in the public domain: strategy and policy, Nat. Rev. Genet. 1 (2000) 70–74. [88] M.B. Eisen, P.T. Spellman, P.O. Brown, D. Botstein, Cluster analysis and display of genome-wide expression patterns, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 14863–14868. [89] P. Tamayo, D. Slonim, J. Mesirov, Q. Zhu, S. Kitareewan, E. Dmitrovsky, E.S. Lander, T.R. Golub, Interpreting patterns of gene expression with self-organizing maps: methods and application to hematopoietic differentiation, Proc. Natl. Acad. Sci. U.S.A. 96 (1999) 2907–2912. [90] S. Hanash, Disease proteomics, Nature 422 (2003) 226–232. [91] M.J. Aardema, J.T. MacGregor, Toxicology and genetic toxicology in the new era of “toxicogenomics”: impact of “-omics” technologies, Mutat. Res. 499 (2002) 13–25. [92] G.A. Boorman, S.P. Anderson, W.M. Casey, R.H. Brown, L.M. Crosby, K. Gottschalk, M. Easton, H. Ni, K.T. Morgan, Toxicogenomics, drug discovery, and the pathologist, Toxicol. Pathol. 30 (2002) 15–27. [93] H.K. Hamadeh, R.P. Amin, R.S. Paules, C.A. Afshari, An overview of toxicogenomics, Curr. Issues Mol. Biol. 4 (2002) 45–56.
360
M.D. Waters et al. / Mutation Research 544 (2003) 349–360
[94] H.K. Hamadeh, P. Bushel, R. Paules, C.A. Afshari, Discovery in toxicology: mediation by gene expression array technology, J. Biochem. Mol. Toxicol. 15 (2001) 231–242. [95] Y. Hirabayashi, T. Inoue, Kokuritsu Iyakuhin Shokuhin Eisei Kenkyusho Hokoku, Toxicogenomics—a new Paradigm of Toxicology and Birth of Reverse Toxicology, Cellular and Molecular Toxicology Division, Center for Biological Safety and Research, National Institute of Health Sciences, Tokyo, Japan (Annual Report) 120 (2002) 39–52 (in Japanese). [96] E.F. Nuwaysir, M. Bittner, J. Trent, J.C. Barrett, C.A. Afshari, Microarrays and toxicology: the advent of toxicogenomics, Mol. Carcinog. 24 (1999) 153–159. [97] W.D. Pennie, I. Kimber, Toxicogenomics; transcript profiling and potential application to chemical allergy, Toxicol. In Vitro 16 (2002) 319–326. [98] J.C. Rockett, D.J. Dix, Application of DNA arrays to toxicology, Environ. Health Perspect. 107 (1999) 681–685. [99] T. Storck, M.C. von Brevern, C.K. Behrens, J. Scheel, A. Bach, Transcriptomics in predictive toxicology, Curr. Opin. Drug Discov. Devel. 5 (2002) 90–97. [100] R.W. Tennant, The National Center for Toxicogenomics: using new technologies to inform mechanistic toxicology, Environ. Health Perspect. 110 (2002) A8–A10. [101] R. Ulrich, S.H. Friend, Toxicogenomics and drug discovery: will new technologies help us produce better drugs, Nat. Rev. Drug Discov. 1 (2002) 84–88. [102] A.L. Castle, M.P. Carver, D.L. Mendrick, Toxicogenomics: a new revolution in drug safety, Drug Discov. Today 7 (2002) 728–736.
[103] T. Ideker, T. Galitski, L. Hood, A new approach to decoding life: systems biology, Annu. Rev. Genomics Hum. Genet. 2 (2001) 343–372. [104] M.D. Waters, G. Boorman, P. Bushel, M. Cunningham, R. Irwin, A. Merrick, K. Olden, R. Paules, J. Selkirk, S. Stasiewicz, B. Weis, B. Van Houten, N. Walker, R. Tennant, Systems toxicology and the chemical effects in biological systems knowledge base, Environ. Health Perspect. 111 (2003) 15–28. [105] O. Ermolaeva, M. Rastogi, K.D. Pruitt, G.D. Schuler, M.L. Bittner, Y. Chen, R. Simon, P. Meltzer, J.M. Trent, M.S. Boguski, Data management and analysis for gene expression arrays, Nat. Genet. 20 (1998) 19–23. [106] R.B. Hayes, Genetic susceptibility and occupational cancer, Med. Lav. 86 (1995) 206–213. [107] P. Grandjean, Individual susceptibility in occupational and environmental toxicology, Toxicol. Lett. 77 (1995) 105–108. [108] A. Hirvonen, Genetic factors in individual responses to environmental exposures, J. Occup. Environ. Med. 37 (1995) 37–43. [109] M.R. Spitz, Risk factors and genetic susceptibility, Cancer Treat. Res. 74 (1995) 73–87. [110] N. Rothman, R.B. Hayes, Using biomarkers of genetic susceptibility to enhance the study of cancer etiology, Environ. Health Perspect. 103 (Suppl. 8) (1995) 291– 295. [111] C.J. Henry, Evolution of toxicology for risk assessment, Int. J. Toxicol. 22 (2003) 3–7.