Toxicogenomics Research Consortium Cooperative Research Program
The Cooperative Research Program of the Toxicogenomics Research Consortium represents the coordinated, dependent research projects performed through the Toxicology Research Cores of each CRM, as well as interactions with the Resource Contractors. There are two types of ongoing cooperative research: standardization experiments and toxicology STAR projects.
The goal of the standardization experiments is to identify and address sources of technical variation in gene expression experiments across multiple technology platforms and research centers. The standardization experiments address different aspects of the microarray experiment, including:
Experiment 1: Variation in RNA labeling and hybridization
Experiment 2: Variation in data analysis (bioinformatics)
Experiment 3: Variation in RNA extraction
Experiment 4: Variation in animal husbandry
The standardization experiments will result in the development of research standards for scientists within the Toxicogenomics Research Consortium and the scientific community as a whole to generate high quality data that are reproducible and comparable. The standardization experiments will also lay the foundation for the toxicology STAR Projects.
The STAR Projects involve investigators from two or more CRMs conducting collaborative toxicology research using gene expression profiling. The STAR projects are focused largely on characterizing cross-specifies genomic responses to environmental stressors. These projects involve a variety of environmental stressors, ranging from metals to alkylating agents, and a number of model species, ranging from yeast to humans.
Project: DNA Alkylation In Neurodegenerative Disease And Cancer
Project Prinicpal Investigator: Peter Spencer, Ph.D., OHSU
Toxicogenomics Research Consortium Members: OHSU-MIT-FHCRC-NTP
This 3-CRM + NTP Star Project employs wild-type and DNA-repair knock-out mice to define differential genome responses of adult liver and brain tissues to a range of alkylating agents linked to cancer and neurodegeneration, respectively. We hypothesize that the genome of adult mice will show differential responses by tissue type and method of alkylation. Since hepatic cells repair alkylation DNA adducts more rapidly than non-dividing neurons, we anticipate a differential gene response in liver versus brain tissue independent of the fact that these tissues are certain to have different basal transcriptional profiles. This project will allow us simultaneously to examine a scientific question relevant to exposures to environmental agents as we address critical technical issues pertinent to the Toxicogenomics Research Consortium /NCT mission.
Chemicals selected for study include the SN1 alkylating agent N-methyl-N-nitrosourea (MNU) (Jin et al., 1996), the SN2 agent methyl methanesulfonate (MMS) or methyl iodide (MI) (Sobel et al., 2002), and methylazoxymethanol (MAM) glycone (MAM-β-D-glucoside or cycasin), which alkylates DNA by an unknown mechanism. Cycasin is of special interest because, in addition to its ability to induce hepatic and other tumors in rodents (Cavanna et al., 1979; Fiala et al., 1987), it is also the leading etiologic candidate for a prototypical neurodegenerative disease known as western Pacific amyotrophic lateral sclerosis (ALS) and parkinsonism-dementia complex (PDC) (Zhang et al., 1996; Esclaire et al., 1999). Because brain (unlike liver) tissue repairs DNA damage inefficiently, the accrual of DNA damage from repeated exposures may establish conditions that eventuate in tardive neuronal degeneration (Kisby et al., 1999). Two radically different phenotypes of long-latency disease – cancer and neurodegeneration – may therefore conceivably be linked through a common pattern of DNA damage that differentially impacts dividing and non-dividing cells, respectively.
Tissues exposed to these methylating agents develop DNA adducts such as O6-methylguanine (O6MeG), N7-methylguanine (7MeG), and 3-methyladenine (3MeA). 7MeG is the predominant adduct produced by exogenous methylating agents (Asagoshi et al., 1999). DNA adducts are repaired by a variety of enzyme-specific mechanisms, such as O6MeG methyltransferase (Mgmt) in the case of O6MeG. 3MeA is repaired by 3MeA alkyladenine DNA glycosylase (Aag), which initiates DNA base-excision repair by removing 3MeA. Mouse dividing cells lacking Aag are susceptible to 3MeA-induced S phase arrest, chromosome aberrations and apoptosis. Smith and Engelward (2000) have shown that repair of MNU-induced 3MeA is at least 10 times slower in Aag (-/-) cells than in Aag (+/+) cells. Consequently, 24 h after exposure to [3H]MNU, 30% of the original 3MeA burden is intact in Aag (-/-) cells, while 3MeA is undetectable in Aag (+/+) cells. MNU-induced 7MeG is removed with equal efficiency in Aag (+/+) and Aag (-/-) cells, suggesting that another DNA glycosylase acts on MNU-induced 7MeG lesions previously thought to be repaired by Aag (Smith and Engelward, 2000).
While understanding is growing of the differential functional roles of DNA repair proteins in mammalian tissues exposed to DNA-damaging alkylating agents, little is known about the protein-specific and general genomic responses to these substances. The three participating CRMs will conduct common experiments that will involve assessment of genomic profiles, data sharing and data analysis to carry out the following specific aims:
Obtain consistent gene expression profiles for wild-type, Aag-, and Mgmt-deficient mice treated with vehicle (year 1);
Assess differential gene expression in wild-type and knock-out mice treated with selected alkylating agents (MNU, MI/MMS and/or MAM-β-D-glucoside, time-permitting) versus vehicle (years 2-3). Each CRM will study animals treated with two of three agents and vehicle, such that Star-wide comparisons may be made for all three agents.
Project: Genetic Modulation Of Hydrocarbon Solvent Neurotoxicity
Project Prinicpal Investigator: Peter Spencer, Ph.D., OHSU
Toxicogenomics Research Consortium Members: OHSU, UW
This 2-CRM Star Project uses wild-type and transgenic mice to examine a specific gene-environment interaction that is postulated to modulate the neurotoxic potency of the active γ-diketone metabolites of certain hydrocarbon solvents — both aliphatic (n-hexane, methyl n-butyl ketone) and aromatic (1,2-diethylbenzene) – that induce CNS-PNS axonopathy, testicular “atrophy” and hepatic enzyme changes in rodents (Gagnaire et al., 1990; Boekelheide and Schoenfeld, 2001, Spencer et al., 2002). Aliphatic γ-diketones also cause human axonal neuropathies that are presently refractory to therapeutic intervention (Spencer et al., 2002). This study seeks to determine whether genetic resistance to oxidative stress renders animals less susceptible to the toxic effects of γ-diketones.
We and others have demonstrated that n-hexane and methyl n-butyl ketone are metabolized to the aliphatic γ-diketone 2,5-hexanedione (2,5-HD), an agent that induces nerve fiber (axonal) degeneration in humans and animals (DeCaprio, 2000). Neurotoxic 2,5-HD, but not non-neurotoxic 2,4-HD, reacts primarily with e-amino groups of lysine in tissue proteins (such as neurofilaments, NF) to form pyrroles that polymerize spontaneously in situ by an oxidative mechanism (Zhu et al., 1997). A similar mechanism is proposed for the formation of isoindole polymers that result from the reaction between NF and 1,2-diacetylbenzene (but not between NF and 1,3-diacetylbenzene), the apparent proximate neurotoxic γ-diketone metabolite of 1,2-diethylbenzene (Kim et al., 2001, 2002; Zhan et al., 2002). Polymer formation appears to underly the accumulation of NF in focal axonal swellings in the central nervous system (brain, spinal cord) and peripheral nervous system (spinal roots). The polymers formed by the reaction between γ-diketones and proteins are chromogenic, and the formation of colored pigment with test proteins and NFs is blocked in vitro by reduced glutathione (GSH) (Sabri MI et al., unpublished observations). The biosynthesis of GSH, which plays a crucial role in protecting cells against oxidative stress, proceeds via two steps that are catalyzed by the rate-limiting enzyme glutamate-cysteine ligase (GCL) and by GSH synthetase (Diaz et al., 2002, Yang et al., 2002). Mice and rats rapidly upregulate GCL and GSH synthesis in response to prooxidant chemical exposures (Woods et al., 1999, Thompson et al., 1999). Kavanagh and colleagues have recently generated inducible GCL transgenic mice which show resistance to CCl4-induced hepatocellular necrosis (Shi et al, 2002). We hypothesize that these oxidative-stress resistant mice are less susceptible to γ-diketone-induced neuro-, testicular, and hepatic toxicity. Neurotoxic γ-diketones (1,2-DAB and 2,5-HD) will be used to assess whether the anticipated protection demonstrated in the whole animal (tissue color, behavior) and in tissue (glutathione levels) is reproduced at the genome level (oligonucleotide microarrays). The non-chromogenic, non-protein reactant and non-neurotoxic isomers (1,3-DAB and 2,4-HD), along with the common vehicle, will serve as negative controls (Kim et al., 2001, in press; Zhan et al., 2002). While there is a wealth of published research on the molecular and cellular toxicity of aliphatic and aromatic hydrocarbon solvents, we are the first to have reported data on the actions of these agents (1,2-DAB, 1,3-DAB) on tissue gene expression, as well as on biochemical and morphological indices of neural function (Kim et al., 2002).
The two participating CRMs will conduct common, interdependent experiments to assess genomic profiles, data sharing and data analysis to carry out the following specific aims:
Obtain consistent gene expression profiles for wild-type and GCL mice treated with vehicle;
Assess differential gene expression in wild-type and GCL transgenic mice treated with selected agents (1,2-DAB and 1,3-DAB, 2,5-HD and 2,4-HD, time-permitting) versus vehicle;
Correlate changes in gene expression with tissue GCL and GSH levels, behavior and histopathological endpoints.
The long-term goal is to conduct coordinated gene- and protein-expression studies to provide a comprehensive view of the response of organs that are vulnerable to organic solvents with protein-reactive metabolites. This work will likely contribute to the development of biomarkers of susceptibility/exposure/effect under study in NIEHS-sponsored research at the two CRM sites.
Project: Analytical Cytology Enhancements to Toxicogenomic Analyses
Project Prinicpal Investigator: Terry Kavanagh, Ph.D., UW
Toxicogenomics Research Consortium Members: FHCRC/UW , Duke, NCT, UNC-CH
cDNA and oligonucleotide microarrays have been used for comparative analyses of mRNA expression at the organism and organ level. However, there are multiple cell types within each target organ, which, because of their differentiation status, may vary dramatically in their response to toxicants. For instance in the liver, centrolobular hepatocytes are often more effective in metabolizing toxicants than hepatocytes in periportal regions. Similarly, in the kidney, proximal tubular cells are often more active towards toxicants than distal tubular cells.
In this STAR project, we propose to make comparisons of differential gene expression in liver and kidney cells obtained from mice that have been treated with agents that induce oxidative stress (e.g. carbon tetrachloride (CCl4), methylmercury (MeHg)). To enhance our ability to detect changes in gene expression that occur with these treatments, we will use high-speed fluorescence activated cell sorting to enrich subpopulations of cells with different metabolic profiles (including cytochrome P450 activities, glutathione (GSH) levels, glutamate-cysteine ligase (GCL) levels and other markers of differentiation-dependent biotransformation capability). We will then compare mRNA expression profiles in these cells using cDNA or oligonucleotide microarray analyses. The mRNA levels for genes found to be differentially expressed by microarray analyses will be further examined by real-time RT-PCR analyses to verify that these mRNA species are in fact differentially expressed, and to assess the relative accuracy of cDNA microarrays. In order to test the hypothesis that these gene expression changes are due to oxidative stress, we will manipulate the level of GSH synthesis using a transgenic mouse strain we have developed which has inducible overexpression of GCL. In order to test the hypothesis that these oxidative stress responses are highly conserved across species, gene expression profiles obtained in these mouse studies will be compared with profiles obtained for the nematode C. elegans (Duke University , Dr. Freedman), rats (NIEHS, Dr. Paules), and PPARα and CAR deficient mice (UNC, Dr. Swenberg) exposed to toxicants which induce oxidative stress. These toxicants include CCl4 (rats and nematodes), MeHg (nematodes), and WY-14,643 and phenobarbital (mice). These expression changes will then be compared to established biomarkers of oxidative tissue and molecular injury in these species. Collectively, these comparisons will form a rational basis for the selection and validation of gene expression biomarkers of oxidative stress.
Project: Mechanisms of Selenium–Induced Chemoprevention
Project Prinicpal Investigator: Terry Kavanagh, Ph.D., UW
Toxicogenomics Research Consortium Members: FHCRC/UW, Duke
The unifying theme of the FHCRC/UW Toxicogenomics Research Consortium is that variations in the response to stressors and toxicants will be crucial to the interpretation of gene expression profiles induced by these exposures. The comparison of gene expression profiles between sensitive and resistant strains will facilitate the identification from among all genes whose expression is altered by exposure, the set of those genes that are related to toxicity, mutagenesis or carcinogenesis. Many dietary components including, but not limited to vitamins, phytoestrogens, organic forms of trace metals, terpenes, and reduced fat and caloric intake have been shown to reduce or prevent carcinogen-induced tumorigenesis. Thus, chemopreventive agents, like animals with differential genetic sensitivities, can be powerful tools in dissecting the molecular mechanisms of chemical mutagenesis and carcinogenesis. By comparing gene expression profiles in the presence and absence of a chemopreventive agent, it will be possible to gain insight into the mechanism of chemically induced toxicity and carcinogenesis, as well as mechanisms of chemoprevention. Furthermore, by comparing the altered patterns of gene expression observed in different organisms following carcinogen treatment in the absence or presence of chemopreventives it will be possible to identify highly conserved pathways that play a fundamental role in the suppression of mutagenesis and/or carcinogenesis.
Selenium-enriched garlic (organic Se) is a chemopreventive agent that suppresses tumor formation when administered during the promotion/progression phase of mammary carcinogenesis (shortly before or after NMU treatment). Mechanisms through which this dietary chemopreventive agent inhibits mammary carcinogenesis remains unknown, but could include the inhibition of spontaneous or chemically induced mutagenesis, or carcinogen-induced epigenetic changes that contribute to outgrowth of these pre-existing Hras1 mutants, induction of apoptosis in initiated cells, etc. The overall goal of the collaborative studies proposed in this STAR proposal is to use organic selenium in comparative genomic studies to understand mechanisms of chemically induced carcinogenesis and chemoprevention. The specific aims of this proposal are:
Specific Aim #1. To compare gene expression profiles in rat mammary cells following exposure to a carcinogenic dose of NMU, in the presence and absence of chemopreventive regimen of dietary organic selenium.
Specific Aim #2. To use C. elegans as a model organism for chemopreventive studies examining mechanisms of Se-mediated chemoprevention by analyzing changes in the nematode genome associated with exposure to NMU and Se, and by generating loss-of function mutants in genes that are key regulators of the pathways implicated in mutagenesis and carcinogenesis.
Project: Comparative Genomic Responses To Environmental Toxicants
Project Prinicpal Investigator: Jon Freedman, Ph.D., Duke/NIEHS
Toxicogenomics Research Consortium Members: Duke, FHCRC/UW, MIT, NIEHS
When organisms are exposed to toxicants, they defend themselves against intracellular damage by activating the transcription of genes that encode proteins that defend the host, repair the damage, or remove/metabolize the toxicant. Changes in expression of these proteins following toxicant exposure is referred to as the “stress-response.” Many of the proteins induced as part of the stress-response are evolutionarily conserved including metallothioneins, superoxide dismutases, heat shock proteins, glutathione-S-transferases, DNA repair enzymes, and pattern recognition receptors. Likewise, many of the signal transduction cascades that regulate the stress-response are conserved. This includes cAMP-, protein kinase C-, and calmodulin-dependent signaling pathways; as well as mitogen-activated protein kinase and NF-kB pathways. Because of the conserved nature of the stress-response and signal-transduction pathways, novel information on the mechanisms of many human diseases have been elucidated via multi-species comparisons. Much of our current understanding of the organization of the ras signal transduction pathway, the activity of tumor suppressor genes and the processes controlling apoptosis has been elucidated from multi-species studies. In fact, key receptors in innate immunity were first discovered in Drosophila before these receptors were found to profoundly affect human immunity (left figure: a, Drosophila; b, mammals). We propose that because of the conserved nature of the regulatory process controlling the stress-response; species from yeast to mammals will activate the transcription of a similar set of genes following toxicant exposure. Moreover, we believe that discovery of this conserved set of responsive genes will focus our attention on genes and pathways that have greater biological relevance to disease etiology. We will investigate the hypothesis that exposure to toxic concentrations of environmental toxicants modulates the transcription of an evolutionarily conserved set of genes. A codicil to this hypothesis states that conserved regulatory pathways will control the transcription of these genes. To investigate this hypothesis we will compare the expression profiles from four species exposed to the four different classes of toxicants. Specifically, we will generate gene expression profiles of yeast, C. elegans, zebrafish embryos and mice that were exposed to archetypical toxicants, which include cadmium (metal) diquat (pro-oxidant), N-methyl-N’-nitro-nitroso¬guanidine (MNNG; alkylating agent) and endotoxin/mycotoxin (biological). The expression profiles will be analyzed to (a) identify expression signature profiles for specific toxicants, (b) identify those genes whose transcriptional response are conserved and (c) define the conserved regulatory pathways that control the response.
Project: Genomic Profiling In Nuclear Receptor-Mediated Toxicity
Project Prinicpal Investigator: Ivan Rusyn, Ph.D., UNC-CH
Toxicogenomics Research Consortium Members: UNC-CH, FHCRC/UW, NCT, MIT
Understanding the mechanisms of non-genotoxic carcinogenesis is a top priority of contemporary toxicology. Comprehending the cellular and molecular changes will permit a strong scientific basis for or against the extrapolation of findings from animals to humans. Based on our extensive research program on the mechanisms of action of environmental chemicals, we devised a broad plan to expand our studies using genomics approaches. Our original research plan focused on the role of nuclear receptors, a superfamily of proteins that act as ligand-activated transcription factors, in the mechanisms of action of environmental agents. This proved to be an important topic for Toxicogenomics since among many research themes that were put forward by other CRMs and the NIEHS NCT, there are several common areas of interest that are ripe for multi-vector collaborations in the framework established by the Toxicogenomics Research Consortium: (i) the mechanisms of non-genotoxic chemical carcinogenesis in liver (C. Omiecinski, T. Kavanagh, R. Paules); (ii) inter-species, and in vivo/in vitro comparisons (L. Griffith, C. Omiecinski), and (iii) oxidative stress as a mode of action for many diverse classes of environmental agents (T. Kavanagh, R. Paules). Thus, the UNC TRCP, in collaboration with UW/FHCRC, MIT and NIEHS NCT, proposes to establish a research program that is focused on investigation of the mechanisms of action of non-genotoxic carcinogens in rodent and human liver. We strongly believe that this project’s productivity will greatly benefit from tapping into the broad expertise available through the Consortium. By coordinating experiments and sharing samples with other CRMs, this project will enhance the research on environmental stressor-related changes in gene expression, and afford a unique opportunity for cross-platform, cross-laboratory comparison of gene expression and phenotypic data.
Standardizing global gene expression analysis between laboratories and across platforms. (2005) Members of the Toxicogenomics Research Consortium. Nature Methods 2, 351-356.