Components of the Vision
The committee foresees pervasive changes in toxicity testing and in interpretive risk-assessment activities. The current approach to toxicity testing focuses on predicting adverse effects in humans on the basis of studies of apical end points in whole-animal tests. In the committee’s vision, in vitro mechanistic tests provide rapid evaluations of large numbers of chemicals, greatly reduced live-animal use, and results potentially more relevant to human biology and human exposures. As discussed in Chapter 2, toxicity testing can be increasingly reconfigured with the accrual of better understanding of biologic pathways perturbed by toxicants and of the signaling networks that control activation of the pathways. The use of systems-biology approaches that integrate responses over multiple levels from molecules to organs will enable a more holistic view of biologic processes, including an understanding of the relationship between perturbations in toxicity pathways and consequences for cell and organism function. The central premise of the committee’s vision is that toxicant-induced responses can be quantified with appropriate cellular assays and that empirical or mechanistic models of pathway perturbations
can be used as the basis of environmental decision-making. Combining a fundamental understanding of cellular responses to toxicants with knowledge of tissue dosimetry in cell systems and in exposed human populations will provide a suite of tools to permit more accurate predictions of conditions under which humans are expected to show pathway perturbations by toxicant exposure. The institutional and infrastructural changes required to achieve the committee’s vision will include changes in the types of tests that support toxicity testing and how toxicity, mechanistic information, and epidemiologic data are used in regulatory decision-making. The regulatory transition from the current emphasis on apical end-point toxicity tests to reliance on perturbations of toxicity pathways will raise many issues. The challenges to implementation and a strategy to implement the vision are discussed in Chapter 5.
This chapter discusses individual components of the vision: chemical characterization (component A), toxicity testing (component B), dose-response and extrapolation modeling (component C), population-based and human exposure data (component D), and risk contexts (component E). Component B is composed of a toxicity-pathway component and a limited targeted-testing component. The toxicity-pathway component will be increasingly dominant as more and more high-throughput toxicity-pathway assays are developed and validated. Surveillance and biomonitoring data will be needed to understand the effects of toxicity-pathway perturbations on humans. Finally, the overall success of the new paradigm will depend on ensuring that toxicity testing meets the information needs of environmental decision-making given the risk contexts.
An overview of component A is provided in Figure 3-1. Chemical characterization is meant to address key questions,
including the compound’s stability in the environment, the potential for human exposure, the likely routes of exposure, the potential for bioaccumulation, the likely routes of metabolism, and the likely toxicity of the compound and possible metabolites based on chemical structure or physical or chemical characteristics. Thus, data would be collected on physical and chemical properties, use characteristics, possible environmental concentrations, possible metabolites and breakdown products, initial molecular interactions of compounds and metabolites with cellular components, and possible toxic properties. A variety of computational methods might be used to predict those properties when data are not available. Decisions could be made after chemical characterization about further testing that might or might not be required. For example, if a chemical were produced in such a manner that it would never reach the environment or were sufficiently persistent and biologically reactive, further toxicity evaluation might not be
necessary for regulatory decision-making. Moreover, computational tools for estimating biologic activities and potency could be useful in assessing characteristics of compounds during their development or in a premanufacturing scenario to rule out development or introduction of compounds that are expected to lead to biologically important perturbations in toxicity pathways. In most cases, chemical characterization alone is not expected to be sufficient to reach decisions about the toxicity of an environmental agent.
The tools for chemical characterization will include a variety of empirical and computational methods. As outlined in the committee’s first report (NRC 2006a), computational approaches that can and most likely will be used are in the following categories: tools to calculate physical and chemical properties, models that predict metabolism and metabolic products of a chemical, structure-activity relationship (SAR) and quantitative SAR (QSAR) models that predict biologic activity from molecular structure, and models that predict specific molecular interactions, such as protein-ligand binding, tissue binding, and tissue solubility. An array of computational tools is available to calculate physical and chemical properties (Volarath et al. 2004; Olsen et al. 2006; Grimme et al. 2007; Balazs 2007). Tools for assessing metabolic fate and biologic activity are continually evolving, and many of the more accurate and refined examples rely on proprietary technology or proprietary databases. Databases that support the most predictive tools may therefore end up being proprietary and substantially different from those available in the public domain. The committee urges the Environmental Protection Agency (EPA) to consider taking a lead role in ensuring public access to the datasets that are developed for predictive modeling and in providing the resources necessary for the continual evolution of methods to develop SAR, QSAR, and other predictive modeling tools.
Many models used to predict hazard are based only on structure and physical and chemical properties and rely on historical datasets. Their reliability is limited by the relevant datasets, which
are continually evolving and increasing in size and accessibility. That is, the predictive value of the structure-activity rules will depend on the chemicals in the dataset from which they are derived—their prevalence, structures, and whether they have the toxic activity of interest (see, for example, Battelle 2002). Computational approaches for predicting toxicity and molecular interactions are available for only a small number of end points, such as estrogen-receptor binding, and their predictive value can be low (Battelle 2002). As approaches improve with time and experience and as the datasets available for model development become larger and more robust, computational tools should become much more useful for chemical characterization, predicting activity in toxicity pathways, and early-stage decision-making.
TOXICITY TESTING OF COMPOUNDS AND METABOLITES
The long-term vision makes the development of predictive toxicity-pathway-based assays the central component of a broad toxicity-testing strategy for assessing biologic activity of new or existing compounds. The assays will be conducted primarily with cells or cell lines, optimally with human cells or cell lines, and as time passes, the need for traditional apical animal tests will be greatly reduced and optimally eliminated. The overview of component B provided in Figure 3-2 indicates that toxicity testing will include both pathway testing and targeted testing, which are discussed further below.
A period of transition is inevitable because of the need to develop the full suite of toxicity-pathway tests that will be required for a comprehensive assessment of toxicity. Challenges related to the transition from the current paradigm oriented to apical end points to that outlined here are addressed separately in Chapter 5.
The committee’s vision focuses on toxicity pathways. Toxicity pathways are simply normal cellular response pathways that are expected to result in adverse health effects when sufficiently perturbed. For example, in early studies of cancer biology, specific genes that were associated with malignant growth and transformation were called oncogenes (those promoting unrestrained cell replication) and tumor-suppressor genes (those restricting replication). Both oncogenes and tumor-suppressor genes were later found to code for proteins that played important roles in normal biology. For example, oncogenes were involved in cell replication, and suppressor-gene products normally halted some key part of
the replication process. However, mutations (such as those which can be induced by some environmental agents) were found to make oncogenes constitutively active or to cause a great reduction in or loss of activity of suppressor genes.
It is the ability of otherwise normal cellular response pathways to be targets for environmental agents that leads to their definition as toxicity pathways. Perturbations of toxicity pathways can be evaluated with a variety of assays, including relatively straightforward biochemical assays, such as receptor-binding or reporter-gene expression, or more integrated cellular response assays, such as assays to evaluate proliferation of an estrogen-responsive cell line after treatment with environmental agents. Cellular responses can be broadly dichotomized as those requiring recognition of the structure of an environmental agent and those occurring because of reactivity of the environmental agent. In the first case, the three-dimensional structure is recognized by macromolecular receptors, as with estrogenic compounds. Accordingly, tests for the structurally mediated responses could be based on binding assays or on integrated cellular-response events, such as proliferation, induction of new proteins, or alteration of phosphorylation status of cells after exposure to environmental agents. In the second case, with reactivity-driven responses, the compound or a metabolite reacts with and damages cellular structures. Reactive compounds have the capacity to be much more promiscuous in their targets in cells, and the initial stress responses to tissue reactivity with these agents may also trigger adaptive changes to maintain homeostasis in the face of increased cellular stress (see Figure 2-2).
Biologic systems from single cells to complex plant and animal organisms have evolved many mechanisms to respond to and counter stressors in their environment. Many responses are mediated through coordinated changes in expression of genes in specific patterns, which result in new operational characteristics of affected cells (Ho et al. 2006; Schilter et al. 2006; Singh and DuMond 2007). Many stress-response pathways—such as those
regulated by hsp90-mediated regulation of chaperone proteins, by Nrf2-mediated antioxidant-element control of cellular glutathione, or by steroid-hormone family (for example, PPAR, CAR, and PXR) receptor-mediated induction of xenobiotic metabolizing enzymes—are conserved across many vertebrate species (Aranda and Pascual 2001; Handschin and Meyer 2005; Westerheide and Morimoto 2005; Kobayashi and Yamamoto 2006). Initial responses to stressors represent adaptation to maintain normal function. When stressors are applied at increasingly high concentrations in combination with other stressors, in sensitive hosts, or during sensitive life stages, adaptation fails, and adverse effects occur in the cell and organism (see Figure 2-2).
As stated, the committee’s long-range vision capitalizes on the identification and use of toxicity pathways as the basis of a new approach to toxicity testing and dose-response modeling. An important question for toxicity-testing strategies concerns the number of pathways that might need to be examined as primary targets of chemical toxicants. For example, in the case of reproductive and developmental toxicity, the National Research Council Committee on Developmental Toxicology listed 17 primary intracellular and intercellular signaling pathways that were then known to be involved in normal development (NRC 2000). Those pathways and the various points for toxic interaction with them are potential targets of chemicals whose structures mimic or disrupt portions of them. Some of the pathways are also important at other life stages, and biologically significant perturbations of them might result in long-lasting effects or effects that are manifested later in life. As discussed in Chapter 5, considerable effort will be required to determine which pathways ultimately to include in the suite of toxicity pathways for testing and what patterns and magnitudes of perturbations will lead to adverse effects.
Some examples of toxicity pathways that could be evaluated with high-throughput methods are listed below, where the consequences of pathway activation are also noted. Most tests are expected to use high-throughput methods, but others could include
medium-throughput assays of more integrated cellular responses, such as cytotoxicity, cell proliferation, and apoptosis. Simpler assays, such as receptor binding or reactivity of compounds with targets (for example, tests of inhibition of cholinesterase activity), also could be used as needed.
Nrf2 antioxidant-response pathway (McMahon et al. 2006; Zhang 2006). The activation of antioxidant-response element signaling occurs through oxidation of sentinel sulfhydryls on the protein Keap1. Some agents, such as chlorine, activate Nrf2 signaling in vitro, and the oxidative stress likely is the cause of irritation and toxicity in the respiratory tract.
Heat-shock-response pathway (Maroni et al. 2003; Westerheide and Morimoto 2005). The activation of protein synthesis by HSP1 transcription factor signaling maintains cellular proteins in an active folded configuration in response to stressors that cause unfolding and denaturation.
PXR, CAR, PPAR, and AhR response pathways (Waxman 1999; Handschin and Meyer 2005; Hillegass et al. 2006; Timsit and Negishi 2006; Li et al. 2006). The activation of xenobiotic metabolizing pathways by transcriptional activation reduces concentrations of some biologically active xenobiotics and enhances elimination from the body as metabolites (Nebert 1994); it can also increase the activation of other xenobiotics to more toxic forms. The toxicity and carcinogenicity of some agents, such as polyaromatic hydrocarbons, occur because of production of mutagenic metabolites by inducible oxidative enzymes.
Hypo-osmolarity-response pathway (Subramanya and Mensa-Wilmot 2006). Cellular stressors damage the integrity of the cellular membranes and activate p38 MAP kinase-mediated pathways to counter them (Van Wuytswinkel et al. 2000). The p38 MAP kinase functionality for the stress responses is conserved across eukaryotes.
DNA-response pathways (Nordstrand et al. 2007). Damage to DNA structures induces repair enzymes that act through
GADD45 (Sheikh et al. 2000) and other proteins. Unrepaired damage increases the risk of mutation during cell division and increases the risk of cancer.
Endogenous-hormone-response pathways (NRC 1999; Harrington et al. 2006). Enhancement or suppression of activity of transcriptionally active hormone receptors—including estrogen, androgen, thyroid, and progesterone receptors (Aranda and Pascual 2001)—leads to altered homeostasis and alteration in biologic functions that are controlled by the receptors.
The biologic revolution now making its way into toxicity testing sets the stage for the design of mechanistic cell-based assays that can be evaluated primarily with high-throughput approaches to testing. The promise of the novel cell-system assays is becoming apparent in advances in several areas: genomic studies of cellular signaling networks affected by chemical exposures, identification of common toxicity pathways that regulate outcomes in diverse tissues, and understanding of networks that control cell responses to external stressors. To ensure the value of results for use in environmental decision-making, the toxicity-pathway assays should be amenable to measurements of dose-response relationships over a broad range of concentrations. Chemical concentrations should be measured directly in the media used in the toxicity-pathway assays when administered concentrations might not represent the concentrations in vitro (for example, in the case of volatile compounds).
Finding new assays for assessing the dose-response characteristics of the toxicity pathways will have high priority for research and standardization. Environmental agents on which animal, human, and cellular evidence consistently demonstrates increased risk of adverse health outcomes could serve as positive controls for evaluation of toxicity-pathway assays. Those controls would serve as standards for the evaluation of the ability of other compounds to perturb the assayed toxicity pathways. Negative controls would also be needed to evaluate the specificity of re-
sponses for the key toxicity pathways. For risk implications in specific populations, interpretation of the studies would consider the results of the assays coupled with information on host susceptibility from other human cell or tissue assays and population-based studies. The research needed to implement the toxicity-pathway approach is discussed further in Chapter 5.
As discussed in Chapter 2, an integral part of the committee’s vision is targeted testing, which would be used to complement toxicity-pathway testing and used in the following circumstances:
To clarify substantial uncertainties in the interpretation of toxicity-pathway data.
To understand effects of representative prototype compounds from classes of materials, such as nanoparticles, that may activate toxicity pathways not included in a standard suite of assays.
To refine a risk estimate when the targeted testing can reduce uncertainty, and a more refined estimate is needed for decision-making.
To investigate the production of possibly toxic metabolites of new compounds.
To fill gaps in the toxicity-pathway testing strategy to ensure that critical toxicity pathways and end points are adequately covered.
One of the challenges of developing an in vitro test system to evaluate toxicity is the current inability of cell assays to mirror the metabolism of a whole animal (Coecke et al. 2006). For the foreseeable future, any in vitro strategy will need to include a provision to assess likely metabolites with whole-animal testing. The metabolites would also need to be tested in a suite of in vitro as-
says. For very reactive metabolites, the suite of assays should include cell models that have biotransformation enzymes required for metabolism. Although it may become possible to make comprehensive predictions of metabolism of environmental agents, any plan to implement the vision here will probably have to rely on some metabolite-identification studies in whole animals. Another challenge is adequate development of in vitro assays to identify reliably toxicity pathways that are causally related to neurodevelopment and other physiologic processes that depend on timing and patterns of exposure and the interactions of multiple pathways. In the near term, targeted in vivo testing will most likely be needed to address those types of toxicities.
Targeted testing might be conducted in vivo or in vitro, depending on the conditions and the toxicity tests available. In the case of metabolite studies, one approach might be to dose small groups of animals with radiolabeled compound, to separate and characterize the excreted radioactivity with modern analytic techniques, and to compare the metabolite structure with known chemistries to determine the need for testing specific metabolites. Similar studies might be conducted in tissue bioreactors, especially a liver bioreactor or cocultures of cells from human liver and other tissues that might make the studies more applicable to human metabolism. Concerns raised in evaluations of metabolism could necessitate synthesis of specific metabolites that would then be tested in the main toxicity-pathway assays. In the development of the European Centre for the Validation of Alternative Methods, there has been extensive discussion of the challenges of capturing the possible toxicity of metabolites so as not to miss ultimate toxicities of substances with in vitro testing (Coecke et al. 2005, 2006).
Although targeted tests could be based on existing toxicity-test systems, they will probably differ from traditional tests in the long term. They could use transgenic species, isogenic strains, new animal models, or other novel test systems (see the committee’s interim report [NRC 2006a] for further discussion) and could include a toxicogenomic evaluation of tissue responses over wide
dose ranges. Whatever system is used, testing protocols would maximize the amount of information gained from whole-animal toxicity testing. For example, routinely used whole-animal toxicity-testing protocols could provide mode-of-action information on toxicity pathways and target tissues in short-term repeat studies. They could emphasize measurement of metabolite formation and applications of transcriptomics and bioinformatics; future designs might include other -omic approaches as the technologies mature and the costs of such studies decrease. Toxicogenomic studies of 14-30 days could provide tissues for microarray analysis and information on pathology. They would harvest a suite of major tissues, mRNA analysis would be performed, and bioinformatics analysis would be conducted to evaluate dose-response relationships in connection with changes in genes and groups of related genes. mRNA from tissues with evidence of pathologic alterations at high doses might also be examined with the major tissues. Thus, the targeted testing in the committee’s vision will not necessarily resemble the standard whole-animal assays now conducted either in the protocol used or in the information gained.
DOSE-RESPONSE AND EXTRAPOLATION MODELING
The committee’s vision includes dose-response and extrapolation modeling modules, which are discussed below; an overview of this component is provided in Figure 3-3.
Empirical Dose-Response Modeling
As they are currently used in toxicity testing with apical end points, empirical dose-response (EDR) models often describe a relationship between the incidence of the end point and either the
dose given to the animal or the concentration of the environmental agent or its metabolite in the target tissue. In the long-range vision, the committee believes that EDR models will be developed for environmental agents primarily on the basis of data from in vitro, mechanistically based assays described in component B. The EDR models would describe the relationship between the concentration in the test medium and the degree of in vitro response; in some cases, they would provide an estimate of some effective concentration at which a specified level of response occurs. The effective concentration could describe, for example, a percentage of maximal response or a statistical increase above background for a more integrated assay, such as an enhanced-cell-proliferation assay. Considerations in the interpretation of in vitro response metrics would include responses in positive and negative controls, their statistical variability, background historical
data, and the experimental dose-response data on the test substance. In general, the toxicity-pathway evaluations require consideration of increases in continuous rather than dichotomous responses.
Dose measures in targeted-testing studies conducted in whole animals could also be expressed in relation to a measure of tissue or plasma concentrations of the parent compound or a metabolite in the organism, such as blood concentration, area under a concentration-time course curve, and rate of metabolism. Preferably, the concentrations would be based on empirical measurements rather than on predictions from pharmacokinetic models. The main reason for insisting that the in vivo studies have a measure of tissue concentration is to permit comparison with the results from the in vitro assays.
In some risk contexts, an EDR model based on in vitro assay results might provide adequate data for a risk-management decision, for example, if host-susceptibility factors of a compound in humans are well understood and human biomonitoring provides good information about its tissue or blood concentrations and about other exposures that affect the toxicity pathway in a human population. Effective concentrations in the suite of in vitro mechanistic assays could be adjusted for host susceptibility and then compared with the human biomonitoring data. In the absence of detailed biomonitoring data and host-susceptibility information, predictions of human response to a toxicant will require building on the data provided by the in vitro EDR models and using physiologically based pharmacokinetic (PBPK) models and perhaps host-susceptibility information on related compounds.
Extrapolation modeling encompasses the analytic tools required to predict exposures that might result in adverse effects
in human populations primarily on the basis of results of hazard testing completed in component B. In the committee’s vision, extrapolation modeling would most likely include PBPK modeling to equate tissue-media concentrations from toxicity testing with tissue doses expected in humans; toxicity-pathway modeling that provides an understanding of the biologic components that control the toxicity-pathway response in vitro; and consideration of human data on host susceptibility and background exposure that provide the context for interpreting the modeling results. As stated in the committee’s interim report (NRC 2006a), the computational approaches must be validated, adequately explained, and made accessible to peer review to be valuable for risk assessment. Models not accessible for review may be useful for many scientific purposes but are not appropriate for regulatory use.
Toxicity-Pathway Dose-Response Models
Models of toxicity-pathway perturbations need to be developed to interpret results from toxicity tests in a mechanistic rather than simply empirical manner; they should be achievable in the near future. Toxicity-pathway models should be more readily configured than models of organism-level toxicity because they describe only the toxicity pathway itself and the initial chemical-related perturbations that are believed to be obligatory but not necessarily sufficient for causing the overt adverse health effect.
Several models of normal signaling pathways have been developed, for example, for heat-shock response (El-Samad et al. 2005; Rieger et al. 2005), platelet-derived growth-factor signaling (Bhalla et al. 2002), and nuclear factor kappa-B-mediated inflammatory signaling in response to cytokines, such as tumor necrosis factor-alpha (Hoffmann et al. 2002; Cho et al. 2003). Also, a screen for anticancer drugs has been developed by using the Nrf2 anti-oxidant-response pathway (Wang et al. 2006), and a preliminary
Nrf2 oxidative-stress model has been developed (Zhang 2006) to examine chlorine as an oxidative stressor and to evaluate both adaptive and overtly toxic responses of cells in culture. Toxicity-pathway dose-response models optimally would describe the interaction of chemicals with cell constituents that activate or repress the pathway (that is, control it) and describe the cellular consequences of activation (that is, the cellular responses, usually altered gene expression, to changes in normal control). Box 3-1 and Figure 3-4 illustrate these concepts in terms of the activation of the Nrf2 antioxidant stress-response pathway.
Although the toxicity-pathway models are discussed here as part of component C of the vision, creation of the models would occur as a natural extension of developing and validating the in vitro toxicity-pathway tests discussed in component B. In other words, the committee envisions that the models would be developed for many assays in component B. The committee recognizes that in the near term there will be continued reliance on default approaches for low-dose extrapolation, such as the linear dose-response model and application of uncertainty factors to benchmark doses or no-observed-adverse-effect levels. The application of uncertainty and adjustment factors to precursor biologic responses from perturbations will not necessarily involve the same factors as currently used in EPA risk assessments for noncancer end points.
The committee emphasizes the important distinction between models for toxicity-pathway perturbations and biologically based dose-response (BBDR) models for apical responses. Approaches to BBDR modeling for complex apical responses—such as cancer (Moolgavkar and Luebeck 1990; Conolly et al. 2003), developmental toxicity (Leroux et al. 1996), and cytotoxicity (Reitz et al. 1990; el-Masri et al. 1996)—have focused on integrated processes, such as proliferation, apoptosis, necrosis, and mutation. Experimental studies and biologic and toxicologic research are still
Example of Components of Signaling Pathway That Could Be Modeled
In nontoxic environments, antioxidant genes are repressed through inactivation of the transcriptional regulator Nrf2. The cytoplasmic protein Keap-1 binds Nrf2 and sequesters Nrf2 in the cytoplasm, where it cannot activate transcription of antioxidant genes (see Figure 3-4). Nrf2 bound to Keap-1 is then quickly degraded through the Cul3-based E3 ligase system (Kobayashi et al. 2004). In toxic environments, some oxidants interact with thiol groups on Keap-1, causing Nrf2 to be released and translocated to the nucleus. Once in the nucleus, Nrf2 heterodimerizes with a small Maf protein and binds to antioxidant response elements; this leads to expression of antioxidant-stress proteins and phase 2-detoxifying enzymes (Motohashi and Yamamoto 2004).
The negative-feedback response loop has two major portions, each of which could be the target of model development. First, the inactivation of Keap-1 by oxidants and the later formation of the Nrf2-Maf heterodimer are response circuits that can be mathematically modeled to predict low-dose toxic responses. Second, the expression of antioxidant-stress proteins and phase 2-detoxifying enzymes can also be modeled to predict low-dose toxic responses.
required to guide the development and validation of such models. Although toxicity-testing strategies would be enhanced by availability of quantitative BBDR models for apical responses, this type of modeling is still in its infancy and probably will not be available for risk-assessment applications in the near future. Progress in developing the models will rely heavily on biologic studies of disease processes in whole animals and mathematical descriptions of the processes. The committee sees BBDR-model development for apical end points as part of a much longer-range research program and does not see routine development of the models from toxicity-pathway testing data in the foreseeable future.
Physiologically Based Pharmacokinetic Modeling
PBPK models assist in extrapolations of dosimetry among doses, dose routes, animal species, and classes of similar chemicals (Clark et al. 2004). They also support risk assessment, aid in designing and interpreting the results of biomonitoring studies (Clewell et al. 2005), and facilitate predictions of human body burden based on use and exposure patterns in specific populations. The development of PBPK models requires variable investment, depending on the chemical. For well-studied classes of compounds, PBPK-model development might require collection of compound-specific characteristics or statistical analysis to incorporate descriptions of human variability and to describe uncertainty (see, for example, Bois et al. 1996; Fouchecourt et al. 2001; Poulin and Theil 2002; Theil et al. 2003). For less well-studied classes of chemicals, model development might require collection of time-course data on tissue concentrations (see, for example, Sarangapani et al. 2002). Validation of existing models is an important consideration. The possibility of studying the pharmacokinetics of low concentrations in environmentally or occupationally exposed humans provides many opportunities for check-
ing the validity of PBPK models. Advances in analytic chemistry permit kinetic studies at extremely low doses that enable opportunities for such studies.
In the future, QSAR should allow estimation of such parameters as blood-tissue partitioning, metabolic rate constants, and tissue binding and could give rise to predictive PBPK models validated with a minimal research investment in targeted studies in test animals. The goal of developing predictive PBPK models dates back to efforts to develop in vitro tools to measure model parameters or to develop QSAR models to predict model parameters on the basis of physical and chemical characteristics or properties (Gargas et al. 1988, 1989).
POPULATION-BASED AND HUMAN EXPOSURE DATA
Population-based and human exposure data will be crucial components of the new toxicity-testing strategy. They will be critical for selecting doses in in vitro and targeted in vivo testing, for interpreting and extrapolating from high-throughput test results, for identifying and understanding toxicity pathways, and for identifying toxic chemical hazards. Figure 3-5 provides an overview of component D, and the following sections discuss how population-based and exposure data can be integrated with toxicity testing.
Population-Based Data and the Toxicity-Testing Strategy
The new toxicity-testing strategy emphasizes the collection of data on the fundamental biologic events involved in the activation of toxicity pathways after exposure to environmental agents. The
collection of mechanistic data on fundamental biologic perturbations will provide new opportunities for greater integration of toxicity testing and population-based studies. In some cases, coordination of the tests will be required; interpretation of toxicity-test results will require an understanding of how human susceptibility factors and background exposures affect the toxicity pathway and how those factors and exposures vary among people.
Genetic epidemiology provides an excellent example of the integration of information from toxicity testing in the long-range vision and population-based studies. It seeks to determine the relationship between specific genes in the population and disease. The finding of genetic loci associated with susceptibility potentially can inform biologists of important cellular proteins that affect disease and can uncover novel disease pathways. Toxicity-testing assays can then be designed to investigate and evaluate the finding and the effects of exogenous chemicals on the disease pathways. For example, human studies have provided information on DNA damage in arsenic-exposed people and motivated laboratory studies on cultured human cells to determine specific DNA-repair pathways affected by arsenic (Andrew et al. 2006).
Conversely, as understanding of toxicity pathways grows, specific genetic polymorphisms that increase or decrease susceptibility to adverse effects of exposure to environmental agents can be more accurately predicted. For example, genetic polymorphisms in some DNA repair and detoxification genes result in higher levels of chromosomal and genomic damage based on the micronuclear centromere content in tissue samples from welders occupationally exposed to welding fumes (Iarmarcovai et al. 2005). Although a substantial amount of normal genetic variation has been identified, only a small fraction of the variation may play a substantive role in influencing differences in human susceptibility. Understanding the biology of the toxicity pathways provides insight into how genetic susceptibility may play an important role. Specifically, a toxicity-testing strategy with a mechanistic focus should define pathways and indicate points that are rate-limiting or are critical signaling nodes in cellular-response systems. Identifying those nodes will allow the potential effects of genotypic variation to be better determined and integrated into chemical-toxicity assessments.
Another example of the interplay between toxicity testing and epidemiology is the generation of potentially important data
on biomarkers. The committee’s vision emphasizes studies conducted in human cells that indicate how environmental agents can affect human biologic response. The studies will suggest biomarkers of early biologic effects that could be monitored in human populations (NRC 2006b). Studying the markers in a variety of cellular systems will help to determine the biomarkers that are best for systematic testing and for use in population-based studies.
Population-health surveillance may indicate human health risks that were not detected in toxicity tests. For example, although pharmaceutical products are subject to extensive toxicologic and clinical testing before their introduction into the marketplace, pharmacovigilance programs have identified adverse health outcomes that were not detected in preclinical and clinical testing (Lexchin 2005; IOM 2007). Food-flavoring agents provide another illustrative example. In 2000, several cases of bronchiolitis obliterans, a severe and rare pulmonary disorder, were described in former workers at a microwave-popcorn plant (Akpinar-Elci et al. 2002). Exposure to vaporized flavoring agents used in the production process was associated with decreased lung function (Kreiss et al. 2002). Flavoring-associated respiratory disease was also documented among food-product workers and among workers in facilities that manufactured the flavoring agents (Lockey et al. 2002). Although the toxicity of the flavoring agents was confirmed in animal studies (Hubbs et al. 2002), their inhalation hazards during manufacture and food-product production was not recognized at the time of product approval. Situations in which toxicity testing is not adequately conducted or fails to identify an important human health risk emphasize the need to integrate population-based studies into any toxicity-testing paradigm and the need to collect human data in a structured manner so that they can be used effectively by the toxicology community.
Human-Exposure Data and the Toxicity-Testing Strategy
Human-exposure data may prove to be pivotal as toxicity testing shifts from the current apical end-point whole-animal testing to cell-based testing. Several types of information will be useful. The first is information collected by manufacturers, users, agencies, or others on exposures of employees in the workplace or on environmental exposures of the population at large. Such exposure information would be considered in the setting of dose ranges for in vitro toxicity testing and of doses for collecting data in targeted pharmacokinetic studies and in selecting concentrations to use in human PBPK models.
Other valuable information will come from biomonitoring surveys of the population that measure environmental agents or their metabolites in blood, urine, or other tissues. New sensitive analytic tools that allow measurement of low concentrations of chemicals in cells, tissues, and environmental media enable tracking of biomarkers in the human population and the environment (Weis et al. 2005; NRC 2006b). Comparison of concentrations of agents that activate toxicity pathways with concentrations of agents in biologic media in human populations will help to identify populations that may be overexposed, to guide the setting of human exposure guidelines, and to assess the cumulative impact of chemicals that influence the same toxicity pathway. The ability to make such comparisons will be greatly strengthened by a deeper understanding of the pharmacokinetic processes that govern the absorption, distribution, metabolism, and elimination of environmental agents by biologic systems. The enhanced ability to identify media concentrations that can evoke biologic responses will help to reduce the uncertainties associated with a focus on apical effects observed at high doses in animal testing.
The importance of biomonitoring data emphasizes the need to support and expand such programs as the National Biomonitoring Program conducted by the Centers for Disease Control and
Prevention (CDC 2001, 2003, 2005). Those programs have greatly increased the understanding of human population exposure and have provided valuable information to guide toxicity testing. In time, biomonitoring will enable assessment of the status of the toxicity-pathway activation in the population. That information will be critical in understanding the implications of high-throughput results for the population and for identifying susceptible populations.
Toxicity testing is valuable only if it can be used to make more informed and more efficient responses to public-health concerns faced by regulators, industry, and the public. In Chapter 1, the committee identified five broad risk contexts requiring decisions about environmental agents, which are listed in Figure 3-6. Each decision-making context creates a need for toxicity-testing information that, if fulfilled, can help to identify the most effective ways to reduce or eliminate health risks posed by environmental agents.
Some of the risk contexts require rapid screening of environmental agents numbering in the tens of thousands. Others require highly refined dose-response information on effects at environmental concentrations, the ability to test chemical mixtures, or the use of focused assays targeted to specific toxicity pathways or end points. Some risk contexts may require the use of population-based approaches, including population health surveillance and biomonitoring. The committee believes that its vision for a new toxicity-testing paradigm will help to respond to decision-making needs, whether regulatory or nonregulatory, and will allow evaluation of all substances of concern whatever their origin
might be. Specific implications of the vision for risk management can be illustrated by considering the five risk contexts identified in Chapter 1.
Evaluation of new environmental agents. Two issues arise in the testing of new chemicals or products. First, emerging technologies might require novel testing approaches. For example, nanotechnology, which focuses on materials in the nanometer range, will present challenges in toxicity testing that might not be easily addressed with existing approaches (IOM 2005; Borm et al. 2006; Gwinn and Vallyathan 2006; Nel et al. 2006; Powell and Kanarek 2006). Specifically, the toxic properties of a nanoscale material will probably depend on its physical characteristics, not on the toxic properties of the substance or element itself (such as
titanium or carbon) that makes up the material. The nanoscale material might be evaluated with new in vitro tests specially designed to identify biologic perturbations that might be expected from exposure to it. As discussed earlier in this chapter, nanoscale materials may require some targeted whole-animal testing to ensure that all biologically significant effects are identified. Second, because many new commercial chemicals are developed each year, there is a need for a mechanism to screen them rapidly for potential toxicity. With an emphasis on high- and medium-throughput screens, the committee’s vision for toxicity testing accommodates screening a large number of chemicals.
Evaluation of existing environmental agents. Two issues arise in the testing of existing environmental agents. For widespread and persistent environmental agents that cannot be easily removed from the human environment and can have potentially significant health effects, an in-depth evaluation of toxic properties is important. The committee’s vision, with its emphasis on toxicity-pathway analysis, will provide the deep understanding needed for refined evaluation of the potential human health effects and risks. As in the evaluation of new environmental agents, there is a need for effective screening methods so that the potential toxicity of the tens of thousands of agents already in the environment can be evaluated. The committee’s toxicity-testing strategy, with high-throughput toxicity-pathway assays, should permit greater coverage of the existing environmental agents that have not been adequately tested for toxicity.
Evaluation of a site. Sites invariably contain a mixture of chemical agents. Evaluation of mixtures has proved to be difficult in the existing toxicity-testing strategy (see Chapter 2). High-throughput assays, as emphasized by the committee, may be the best approach for toxicity assessment of mixtures because they are more easily used to assess combinations of chemicals. Biomonitoring data—whose collection is highlighted in the committee’s vi-
sion—can be especially useful in site investigations to identify problematic exposures.
Evaluation of potential environmental contributors to a specific disease. Public-health problems, such as clusters of cancer cases or outbreaks of communicable diseases, can have an environmental component. Asthma has distinct geographic, temporal, and demographic patterns that strongly suggest environmental contributions to its incidence and severity (Woodruff et al. 2004) and provides an excellent illustration of how the committee’s vision could help to elucidate the environmental components of a disease. First, animal models of asthma have been plagued by important species differences, which limit the utility of standard toxicity-testing approaches (Pabst 2002; Epstein 2004). Second, substantial data are available on toxicity pathways involved in asthma (Maddox and Schwartz 2002; Pandya et al. 2002; Lutz and Sulkowski 2004; Lee et al. 2005; Chan et al. 2006; Nakajima and Takatsu 2006; Abdala-Valencia et al. 2007); the pathways should be testable with high-throughput assays, which could permit the evaluation of many environmental agents for a potential etiologic role in the induction or exacerbation of asthma. Third, environmental agents that raise concern in the high-throughput assays could have high priority in population-based studies for evaluation of their potential link to asthma in human populations, such as workers. The high-throughput assays that are based on evaluation of toxicity pathways can survey large numbers of environmental agents and identify those which operate through a mechanism that may be relevant to a disease of interest, as in the case of asthma, and may help to generate useful hypotheses that can then be examined in population-based studies.
Evaluation of the relative risks posed by environmental agents. It is often useful to assess the relative risks associated with different environmental agents, such as pesticides or pharmaceutical products, that could have been developed for the same purpose. The new toxicity-testing paradigm will provide information on rela-
tive potencies established by computational toxicology, toxicity-pathway analysis, dose-response analysis, and targeted testing.
The future toxicity-testing strategy envisioned by the committee will be well suited to providing the relevant data needed to make the critical risk-management decisions required in the long term.
TOXICITY-TESTING STRATEGIES IN PRACTICE
To illustrate how the results of the tests envisioned by the committee may be applied in specific circumstances, two hypothetical examples of environmental agents that may pose risks to human health are considered. The first example is an irritant gas, and the second is an environmental agent that acts by interactions with estrogen receptors. The committee emphasizes that these examples are intended not to recommend definitive procedures for conducting human health risk assessment but simply to show how assessment might be approached. As the research discussed in Chapter 5 is conducted, much will be learned, and new tests and methods to incorporate results into assessments will emerge.
Toxicity Testing and Empirical Dose-Response Analysis
Among a larger group of gases tested in multiple high-throughput assays, the agent caused dose-related responses in test assays for glutathione depletion, Nfr2 oxidative-stress pathway activation, inflammatory pathway responses, and general cytotoxicity. Most other human toxicity-pathway tests had negative results, but the test gas was routinely cytotoxic in systems in which
gases were easily tested. Nrf2 pathway activation proved to be the most sensitive end point, with an EC101 of 10 ppm and a lower bound on the EC10 of 6.5 ppm.
A known hydrolysis product of the test gas—one produced in stoichiometric equivalents on hydrolysis of the gas—produced similar responses in vitro when tested over a thousand-fold concentration range (0.001-1 mM). The test provided a lower bound ED101 of 0.12 mM for Nrf2 pathway activation. The hydrolysis product was tested in a broad suite of toxicity pathways and showed little evidence of pathway specific responses, but consistently showed toxic responses at concentrations much above 1.0 mM.
At nontoxic concentrations, the compound showed no evidence of mutagenicity.
Low dose. With positive-control oxidants, low-dose behavior of the Nrf2 pathway was shown to be nonlinear because of high gain in the feedback loops that control activation of this adaptive stress-response pathway. A concentration of one-tenth the lower bound on the EC10 would not be expected to cause substantial pathway activation. That concentration would serve as a starting point for consideration of susceptibility factors, preexisting disease in the human population, and possible co-exposures to similarly acting compounds.
In vitro to in vivo. Extrapolation from the in vitro system used a human pharmacokinetic model derived from a computational fluid-dynamics approach. Model inputs derived partially from SAR included reaction rates of the gas in tissues and species-specific breathing rates. The pharmacokinetic dosimetry model
was used to calculate the exposure concentrations that would yield 0.012 mM hydrolysis product (that is, 0.12 mM/10) in the nose and lungs during a continuous human inhalation exposure. The pharmacokinetic model, run in Markov-chain Monte Carlo fashion to account for variability and uncertainty, provided lower-bound estimates of 2.5 ± 0.6 ppm for the lungs and 15 ± 3 ppm for the nose. Sensitivity analysis of the combined toxicity-pathway dosimetry model indicated key biologic and pharmacokinetic factors that had important roles in dose delivery and the circuitry governing Keap1 and Nrf2 signaling.
Susceptibility. Susceptibility would depend heavily on polymorphisms in critical portions of the Nrf2 pathway. People with higher than average Keap1 or lower than average Nrf2 could fail to have an adaptive response to oxidative stressors and could progress to toxicity at lower exposure concentrations. The observed polymorphisms in the human population and sensitivity with pre-existing diseases suggest that estimates arising from the dose-response analysis should be reduced by a factor of 10.
The exposure concentration derived from the high-throughput toxicity-pathway screens and the associated interpretive tools could be used in setting reference standards. The assessment would indicate that the concentration should ensure that an exposure would not lead to biologically significant responses to the compound. In addition, the risk narrative would state that this exposure limit should be protective of other downstream responses—such as respiratory tract toxicity—that might be of concern at higher concentrations, because even adaptive, precursor responses are being avoided.
Estimates of cumulative risk should be considered for situations with simultaneous exposures to the irritant gas and other gases that affect Nrf2 signaling.
Surveillance studies of workers or other human populations potentially exposed to the irritant gas could test for evidence of Nfr2 oxidative-stress pathway activation and inflammatory pathway responses, possibly using induced sputum samples. To evaluate the results, any increases in activation in the exposed population could be compared with pathway activation in control human populations.
Toxicity Testing and Empirical Dose-Response Analysis
A large group of commercial chemicals were tested in multiple high-throughput in vitro assays. One of them triggered dose-related activation of estrogenic signaling in receptor-binding as-says and increased DNA replication—indicative of cell proliferation—in human breast-cancer cells in vitro. Binding assays for this compound had the lowest ED10 values; assay indicators of gene transcription and DNA replication occurred at much higher concentrations. QSAR methods also predicted an estrogenic effect on the basis of a library of tested compounds. All other human toxicity-pathway tests were negative or showed responses at much higher concentrations. The test compound had low cytotoxicity in most screens and produced estrogen-receptor activation at concentrations one-tenth of those which produce signs of cell toxicity.
A short-term, mechanistic in vivo study with ovariectomized female rats confirmed mild estrogenic action in vivo and moderate evidence of gene expression for responses in utero or in breast tissues. Predicted conjugated metabolites of the compound were without activity in those assays.
Experience with estrogen and other estrogenic chemicals indicates the existence of susceptible populations—such as pubescent girls, fetuses, and infants—that require additional protection and attention. In addition, chemicals that bind to and activate the estrogen receptor may act additively with one another. The extrapolation needs to consider the compound uses, subpopulations that are likely to be exposed to it, other background exposures to estrogenic agents in these subpopulations, and the estimated tissue dose in pregnant and nonpregnant women, fetuses, and infants.
Research on estrogen and estrogen agonists reveals that if receptor occupancy in the most sensitive tissues in susceptible humans is increased by less than x % by this exposure or any combined exposure to estrogenic compounds, an appreciable activation of downstream responses or a biologically significant increase in their activation would be unlikely. An alternative assessment would be based on a functional response in a toxicity-pathway assay, such as transcriptional activation.
Human PBPK models for the compound would be used to model absorption, distribution to sensitive tissues, and elimination of active parent compound. The models (for example, Markov Chain Monte Carlo PBPK model) would be designed to account for human variability in pharmacokinetics and modeling un-
certainty. The PBPK models could generate a point-of-departure exposure concentration or a daily intake at which there would be less than x % increase in receptor occupancy or less than x % change in transcriptional activation in susceptible populations (for example, fetuses) and in 95% to 99% of the exposed general population. The PBPK models could also provide the blood concentration associated with the change in receptor occupancy or transcriptional activation. That blood concentration could be expressed in units of “estrogen equivalence” to simplify comparisons with estrogen and similarly acting estrogen agonists. Also, on the basis of estrogen equivalence, the models could be used to assess the effects of cumulative exposure to exogenous estrogenic compounds and could be checked against biologic monitoring data in the human population for validity and to ensure that the point of departure is not overestimated.
Reference doses and concentrations used in decision-making could be based on a point of departure derived as described above. The reference dose would consider factors, such as susceptibility, that could be altered by polymorphisms in critical portions of downstream estrogen-response pathways or in conjugation with enzymes that clear the compound before it reaches the systemic circulation.
Human surveillance of workers exposed to the compound could detect subtle indications of early effects in humans if they were to occur.
TOXICITY TESTING AND RISK ASSESSMENT
A major application of the results of toxicity testing is in the risk assessment of environmental agents. As illustrated in Figure 3-7, the committee’s vision for toxicity testing is consistent with the risk-assessment paradigm originally put forward by the National Research Council in 1983. Chemical characterization and toxicity-pathway evaluation would be involved in hazard identification. Pharmacokinetic models would be used to calibrate in vitro and human dosimetry and thereby facilitate the translation of dose in cellular systems to dose in human organs and tissues. Population-based studies would be used to confirm or explore effects observed in cellular systems to suggest biologic perturbations that require clarification in in vitro tests and to interpret findings in in vitro studies in the context of human populations. All would work together to permit establishment of human exposure guidelines based on risk avoidance, which could be used to enforce scientifically based regulatory standards or support non-regulatory risk-management strategies.
Mode-of-action information is important for informing the dose-response component of the risk-assessment paradigm. A deep understanding of mode of action involves studying the mechanistic pathways by which toxic effects are induced, including the key molecular and other biologic targets in the pathways. Thus, the committee’s vision, outlined in Chapters 2 and 3 of this report, is a shift away from traditional toxicity testing that focuses on demonstrating adverse health effects in experimental animals toward a deeper understanding of biologic perturbations in key toxicity pathways that lead to adverse health outcomes. The committee believes that its vision of toxicity testing would better inform the assessment of the potential human health risks posed by exposure to environmental agents and ensure efficient testing methods.
Abdala-Valencia, H,, J. Earwood, S. Bansal, M. Jansen, G. Babcock, B. Garvy, M. Wills-Karp, and J.M. Cook-Mills. 2007. Non-hematopoietic NADPH oxidase regulation of lung eosinophilia and airway hyperresponsiveness in experimentally-induced asthma. Am. J. Physiol. Lung Cell Mol. Physiol. 292(5):L1111-L1125.
Akpinar-Elci, M., R. Kanwal, and K. Kreiss. 2002. Bronchiolitis obliterans syndrome in popcorn plant workers. Am. J. Respir. Crit. Care Med. 165:A526.
Andrew, A.S., J.L. Burgess, M.M. Meza, E. Demidenko, M.G. Waugh, J.W. Hamilton, and M.R. Karagas. 2006. Arsenic exposure is associated with decreased DNA repair in vitro and in individuals exposed to drinking water arsenic. Environ. Health Perspect. 114(8):1193-1198.
Aranda, A., and A. Pascual. 2001. Nuclear hormone receptors and gene expression. Physiol. Rev. 81(3):1269-1304.
Balazs, A.C. 2007. Modeling self-assembly and phase behavior in complex mixtures. Annu. Rev. Phys. Chem. 58:211-233
Battelle. 2002. Evaluation of SAR Predictions of Estrogen Receptor Binding Affinity. EPA Contract Number 68-W-01-023, Work Assignment 2-3. Prepared for U.S. Environmental Protection Agency, by Battelle, Columbus, OH. August 1, 2002.
Bhalla, U.S., P.T. Ram, and R. Iyengar. 2002. MAP kinase phosphatase as a locus of flexibility in a mitogen-activated protein kinase signaling network. Science 297(5583):1018-1023.
Bois, F.Y., A. Gelman, J. Jiang, D.R. Maszle, L. Zeise, and G. Alexeef. 1996. Population toxicokinetics of tetrachloroethylene. Arch Toxicol 70(6):347-355.
Borm, P.J., D. Robbins, S. Haubold, T. Kuhlbusch, H. Fissan, K. Donaldson, R. Schins, W. Kreyling, J. Lademann, J. Krutmann, D. Warheit, and E. Oberdorster. 2006. The potential risks of nanomaterials: A review carried out for ECETOC. Part. Fibre Toxicol. 3:11.
CDC (Centers for Disease Control and Prevention). 2001. National Report on Human Exposure to Environmental Chemicals. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Atlanta, GA [online]. Available: http://www.noharm.org/details.cfm?ID=745&type=document [accessed July 25, 2006].
CDC (Centers for Disease Control and Prevention). 2003. Second National Report on Human Exposure to Environmental Chemicals. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Atlanta, GA [online]. Available: http://www.serafin.ch/toxicreport.pdf [accessed July 25, 2006].
CDC (Centers for Disease Control and Prevention). 2005. Third National Report on Human Exposure to Environmental Chemicals. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Atlanta, GA [online]. Available: http://www.cdc.gov/exposurereport/pdf/third_report_chemicals.pdf [accessed July 25, 2006].
Chan, R.C., M. Wang, N. Li, Y. Yanagawa, K. Onoe, J.J. Lee, and A.E. Nel. 2006. Pro-oxidative diesel exhaust particle chemicals inhibit LPS-induced dendritic cell responses involved in T-helper differentiation. J. Allergy Clin. Immunol. 118(2):455-465.
Cho, K.H., S.Y. Shin, H.W. Lee, and O. Wolkenhauer. 2003. Investigations into the analysis and modeling of the TNFα-mediated NF-κB-signaling pathway. Genome Res. 13(11):2413-2422.
Clark, L.H., R.W. Setzer, and H.A. Barton. 2004. Framework for evaluation of physiologically-based pharmacokinetic models for use in safety or risk assessment. Risk Anal. 24(6): 1697-1717.
Clewell, H.J., P.R. Gentry, J.E. Kester, and M.E. Andersen. 2005. Evaluation of physiologically based pharmacokinetic models in risk assessment: An example with perchloroethylene. Crit. Rev. Toxicol. 35(5):413-433.
Coecke, S., B.J. Blaauboer, G. Elaut, S. Freeman, A. Freidig, N. Gensmantel, P. Hoet, V.M. Kapoulas, B. Ladstetter, G. Langley, D. Leahy, G. Mannens, A. Meneguz, M. Monshouwer, B. Nemery, O. Pelkonen, W. Pfaller, P. Prieto, N. Proctor, V. Rogiers, A. Rostami-Hodjegan, E. Sabbioni, W. Steiling, and
J.J. van de Sandt. 2005. Toxicokinetics and metabolism. Altern. Lab. Anim. 33(Suppl. 1):147-175.
Coecke, S., H. Ahr, B.J. Blaauboer, S. Bremer, S. Casati, J. Castell, R. Combes, R. Corvi, C.L. Crespi, M.L. Cunningham, G. Elaut, B. Eletti, A. Freidig, A. Gennari, J.F. Ghersi-Egea, A. Guillouzo, T. Hartung, P. Hoet, M. Ingelman-Sundberg, S. Munn, W. Janssens, B. Ladstetter, D. Leahy, A. Long, A. Meneguz, M. Monshouwer, S. Morath, F. Nagelkerke, O. Pelkonen, J. Ponti, P. Prieto, L. Richert, E. Sabbioni, B. Schaack, W. Steiling, E. Testai, J.A. Vericat, and A. Worth. 2006. Metabolism: A bottleneck in in vitro toxicological test development. The report and recommendations of ECVAM workshop 54. Altern. Lab Anim. 34(1):49-84.
Conolly, R.B., J.S. Kimbell, D. Janszen, P.M. Schlosser, D. Kalisak, J. Preston, and F.J. Miller. 2003. Biologically motivated computational modeling of formaldehyde carcinogenicity in the F344 rat. Toxicol. Sci. 75(2):432-447.
el-Masri, H.A., R.S. Thomas, G.R. Sabados, J.K. Phillips, A.A. Constan, S.A. Benjamin, M.E. Andersen, H.M. Mehendale, and R.S. Yang. 1996. Physiologically based pharmacokinetic/pharmacodynamic modeling of the toxicology interaction between carbon tetrachloride and kepone. Arch. Toxicol. 70(11):704-713.
El-Samad, H., H. Kurata, J.C. Doyle, C.A. Gross, and M. Khammash. 2005. Surviving heat shock: Control strategies for robustness and performance. Proc. Natl. Acad. Sci. U.S.A. 102(8):2736-2841.
Epstein, M.M. 2004. Do mouse models of allergic asthma mimic clinical disease? Int. Arch. Allergy Immunol. 133(1):84-100.
Fouchecourt, M.O., M. Beliveau, and K. Krishnan. 2001. Quantitative structure-pharmacokinetic relationship modeling. Sci. Total Environ. 274(1-3):125-135.
Gargas, M.L., P.G. Seybold, and M.E. Andersen. 1988. Modeling the tissue solubilities and metabolic rate constant (Vmax) of halogenated methanes, ethanes, and ethylenes. Toxicol. Lett. 43(1-3):235-256.
Gargas, M.L., R.J. Burgess, D.E.Voisard, G.H. Cason, and M.E. Andersen. 1989. Partition coefficients of low-molecular-weight volatile chemicals in various liquids and tissues. Toxicol. Appl. Pharmacol 98(1):87-99.
Grimme, S., M. Steinmetz, and M. Korth. 2007. How to compute isomerization energies of organic molecules with quantum chemical methods. J. Org. Chem. 72(6):2118-2126.
Gwinn, M.R., and V. Vallyathan. 2006. Nanoparticles: Health effects—pros and cons. Environ. Health Perspect. 114(12):1818-1825.
Handschin, C., and U.A. Meyer. 2005. Regulatory network of lipid-sensing nuclear receptors: Roles for CAR, PXR, LXR, and FXR. Arch. Biochem. Biophys. 433(2):387-396.
Harrington, W.R., S.H. Kim, C.C. Funk, Z. Madak-Erdogan, R. Schiff, J.A. Katzenellenbogen, and B.S. Katzenellenbogen. 2006. Estrogen dendrimer conjugates that preferentially activate extranuclear, nongenomic versus genomic pathways of estrogen action. Mol. Endocrinol. 20(3):491-502.
Hillegass, J.M., K.A. Murphy, C.M. Villano, and L.A. White. 2006. The impact of aryl hydrocarbon receptor signaling on matrix metabolism: Implications for development and disease. Biol Chem. 387(9):1159-1173.
Ho, S.M., W.Y. Tang, J. Belmonte de Frausto, and G.S. Prins. 2006. Developmental exposure to estradiol and bisphenol A increases susceptibility to prostate carcinogenesis and epigenetically regulates phophodiesterase type 4 variant 4. Cancer Res. 66(11):5624-5632.
Hoffmann, A., A. Levchenko, M.L. Scott, and D. Baltimore. 2002. The IκB-NF-κB signaling module: Temporal control and selective gene activation. Science 298(5596):1241-1245.
Hubbs, A.F., L.A. Battelli, W.T. Goldsmith, D.W. Porter, D. Frazer, S. Friend, D. Schwegler-Berry, R.R. Mercer, J.S. Reynolds, A. Grote, V. Castranova, G. Kullman, J.S. Fedan, J. Dowdy, and W.G. Jones. 2002. Necrosis of nasal and airway epithelium in rats inhaling vapors of artificial butter flavoring. Toxicol. Appl. Pharmacol. 185(2):128-135.
Iarmarcovai, G., I. Sari-Minodier, F. Chaspoul, C. Botta, M. De Meo, T. Orsiere, J.L. Berge-Lefranc, P. Gallice, and A. Botta. 2005. Risk assessment of welders using analysis of eight metals by ICP-MS in blood and urine and DNA damage evaluation by the comet and micronucleus assays; influence of XRCC1 and XRCC3 polymorphisms. Mutagenesis 20(6):425-432.
IOM (Institute of Medicine). 2005. Implications of Nanotechnology for Environmental Health Research. Washington, DC: The National Academies Press.
IOM (Institute of Medicine). 2007. The Future of Drug Safety: Promoting and Protecting the Health of the Public. Washington, DC: The National Academies Press.
Kobayashi, A., M.I. Kang, H. Okawa, M. Ohtsuji, Y. Zenke, T. Chiba, K. Igarashi, and M. Yamamoto. 2004. Oxidative stress sensor Keap1 functions as an adaptor for CuI3-based E3 ligase to regulate proteasomal degradation of Nrf2. Mol. Cell. Biol. 24(16):7130-7139.
Kobayashi, M., and M. Yamamoto. 2006. Nrf2-Keap1 regulation of cellular defense mechanisms against electrophiles and reactive oxygen species. Adv. Enzyme Regul. 46:113-140.
Kreiss, K., A. Gomaa, G. Kullman, K. Fedan, E.J. Simoes, and P.L. Enright. 2002. Clinical bronchiolitis obliterans in workers at a microwave-popcorn plant. N. Engl. J. Med. 347(5):330-338.
Lee, C.T., J. Ylostalo, M. Friedman, and G.W. Hoyle. 2005. Gene expression profiling in mouse lung following polymeric hexamethylene diisocyanate exposure. Toxicol. Appl. Pharmacol. 205(1):53-64.
Leroux, B.G., W.M. Leisenring, S.H. Moolgavkar, and E.M. Faustman. 1996. A biologically-based dose-response model for developmental toxicology. Risk Anal. 16(4):449-458.
Lexchin, J. 2005. Drug withdrawals from the Canadian market for safety reasons, 1963-2004. CMAJ 172(6):765-767.
Li, T., W. Chen, and J.Y. Chiang. 2006. PXR induces CYP27A1 and regulates cholesterol metabolism in the intestine. J. Lipid Res. 48(2):373-384.
Lockey, J., R. Mckay, E. Barth, J. Dahlsten, and R. Baughman. 2002. Bronchiolits obliterans in the food flavoring manufacturing industry. Am. J. Respir. Crit. Care Med. 165:A461.
Lutz, W., and W.J. Sulkowski. 2004. Vagus nerve participates in regulation of the airways: Inflammatory response and hyperreactivity induced by occupational asthmogens. Int. J. Occup. Med. Environ. Health 17(4):417-431.
Maddox, L., and D.A. Schwartz. 2002. The pathophysiology of asthma. Annu. Rev. Med. 53:477-498.
Maroni, P., P. Bendinelli, L. Tiberio, F. Rovetta, R. Piccoletti, and L. Schiaffonati. 2003. In vivo heat-shock response in the brain: Signaling pathway and transcription factor activation. Brain Res. Mol. Brain Res. 119(1):90-99.
McMahon, M., N. Thomas, K. Itoh, M. Yamamoto, and J.D. Hayes. 2006. Dimerization of substrate adaptors can facilitate cullin-mediated ubiquitylation of proteins by a "tethering" mechanism: A two-site interaction model for the Nrf2-Keap1 complex. J. Biol. Chem. 281(34):24756-24768.
Moolgavkar, S.H. and G. Luebeck. 1990. Two-event model for carcinogenesis: Biological, mathematical, and statistical considerations. Risk Anal. 10(2):323-341.
Motohashi, H., and M. Yamamoto. 2004. Nrf2-Keap1 defines a physiologically important stress response mechanism. Trends Mol. Med. 10(11):549-557.
Nakajima, H., and K. Takatsu. 2006. Role of cytokines in allergic airway inflammation. Int. Arch. Allergy Immunol. 142(4):265-273.
Nebert, D.W. 1994. Drug-metabolizing enzymes in ligand-modulated transcription. Biochem. Pharmacol. 47(1):25-37.
Nel, A., T. Xia, L. Madler, and N. Li. 2006. Toxic potential of materials at the nanolevel. Science 311(5761):622-627.
Nordstrand L.M., J. Ringvoll, E. Larsen, and A. Klungland. 2007. Genome instability and DNA damage accumulation in gene-targeted mice. Neuroscience 145(4):1309-1317.
NRC (National Research Council). 1999. Hormonally Active Agents in the Environment. Washington, DC: National Academy Press.
NRC (National Research Council). 2000. Scientific Frontiers in Developmental Toxicology and Risk Assessment. Washington, DC: National Academy Press.
NRC (National Research Council). 2006a. Toxicity Testing for Assessment of Environmental Agents: Interim Report. Washington, DC: The National Academies Press.
NRC (National Research Council). 2006b. Human Biomonitoring for Environmental Chemicals. Washington, DC: The National Academies Press.
Olsen, L., P. Rydberg, T.H. Rod, and U. Ryde. 2006. Prediction of activation energies for hydrogen abstraction by cytochrome p450. J. Med. Chem. 49(22):6489-6499.
Pabst, R. 2002. Animal models for asthma: Controversial aspects and unsolved problems. Pathobiology 70(5):252-254.
Pandya, R.J., G. Solomon, A. Kinner, and J.R. Balmes. 2002. Diesel exhaust and asthma: Hypotheses and molecular mechanisms of action. Environ. Health Perspect. 110(Suppl. 1):103-112.
Poulin, P., and F.P. Theil. 2002. Prediction of pharmacokinetics prior to in vivo studies. II. Generic physiologically based pharmacokinetic models of drug disposition. J. Pharm. Sci. 91(5):1358-1370.
Powell, M.C., and M.S. Kanarek. 2006. Nanomaterial health effects-Part 1: Background and current knowledge. WMJ 105(2):16-20.
Reitz, R.H., A.L. Mendrala, R.A. Corley, J.F. Quast, M.L. Gargas, M.E. Andersen, D.A. Staats, and R.B. Conolly. 1990. Estimating the risk of liver cancer associated with human exposures to chloroform using physiologically based pharmacokinetic modeling. Toxicol. Appl. Pharmacol. 105(3):443-459.
Rieger, T.R., R.I. Morimoto, and V. Hatzimanikatis. 2005. Mathematical modeling of the eukaryotic heat-shock response: Dynamics of the hsp70 promoter. Biophys. J. 88(3):1646-1658.
Sarangapani, R., J. Teeguarden, K.P. Plotzke, J.M. McKim, Jr., and M.E. Andersen. 2002. Dose-response modeling of cytochrome p450 induction by rats by octamethylcyclotetrasiloxane. Toxicol. Sci. 67(2):159-172.
Schilter, B., M. Marin-Kuan, T. Delatour, S. Nestler, P. Mantle, and C. Cavin. 2006. Ochratoxin A: Potential epigenetic mechanisms of toxicity and carcinogenicity. Food Addit. Contam. 22(Suppl. 1):88-93.
Sheikh, M.S., M.C. Hollander, and A.J. Fornance, Jr. 2000. Role of Gadd45 in apoptosis. Biochem. Pharmacol. 59(1):43-45.
Singh, K.P., and J.W. DuMond, Jr. 2007. Genetic and epigenetic changes induced by chronic low dose exposure to arsenic of mouse testicular Leydig cells. Int. J. Oncol. 30(1):253-260.
Subramanya, S., and K. Mensa-Wilmot. 2006. Regulated cleavage of intracellular glycosylphosphatidylinositol in a trypanosome. Peroxisome-to-endoplasmic reticulum translocation of a phospholipase C. FEBS J. 273(10):2110-2126.
Theil, F.P., T.W. Guentert, S. Haddad, and P. Poulin. 2003. Utility of physiologically based pharmacokinetic models to drug development and rational drug discovery candidate selection. Toxicol. Lett. 138(1-2):29-49.
Timsit, Y.E. and M. Negishi. 2006. CAR and PXR: The xenobiotic-sensing receptors. Steroids 72(3):231-246.
Van Wuytswinkel, O., V. Reiser, M. Siderius, M.C. Kelders, G. Ammerer, H. Ruis, and W.H. Mager. 2000. Response of Saccharomyces cerevisiae to sever osmotic stress: Evidence for a novel activation mechanism of the HOG MAP kinase pathway. Mol. Microbiol. 37(2):382-397.
Volarath, P., H. Wang, H. Fu, and R. Harrison. 2004. Knowledge-based algorithms for chemical structure and property analysis. Conf. Proc. IEEE Eng. Med. Biol. Soc. 4:3011-3014.
Wang, X.J., J.D. Hayes, and C.R. Wolf. 2006. Generation of a stable antioxidant response element-driven reporter gene cell line and its use to show redox-dependent activation of nrf2 by cancer chemotherapeutic agents. Cancer Res. 66(22):10983-10994.
Waxman, D.J. 1999. P450 gene induction by structurally diverse xenochemicals: Central role of nuclear receptors CAR, PXR, and PPAR. Arch. Biochem. Biophys. 369(1):11-23.
Weis, B.K., D. Balshaw, J.R. Barr, D. Brown, M. Ellisman, P. Lioy, G. Omenn, J.D. Potter, M.T. Smith, L. Sohn, W.A. Suk, S. Sumner, J. Swenberg, D.R. Walt, S. Watkins, C. Thompson, and S.H. Wilson. 2005. Personalized exposure assessment: Promising approaches for human environmental health research. Environ. Health Perspect. 113(7):840-848.
Westerheide, S.D., and R.I. Morimoto. 2005. Heat shock response modulators as therapeutic tools for diseases of protein conformation. J. Biol. Chem. 280(39):33097-33100.
Woodruff, T.J., D.A. Axelrad, A.D. Kyle, O. Nweke, G.G. Miller, and B.J. Hurley. 2004. Trends in environmentally related childhood illnesses. Pediatrics 113(Suppl. 4):1133-1140.
Zhang, D.D. 2006. Mechanistic studies of the Nrf2-Keap1 signaling pathway. Drug Metab Rev. 38(4):769-789.