This appendix provides case studies that show how 21st century science can be used for chemical assessment, including any component of the risk-assessment process (hazard identification, dose–response assessment, exposure assessment, or risk characterization). The first case study illustrates the use of read-across methods to address gaps in information on a data-poor chemical. The second uses air pollution as a topic to illustrate how 21st century science can be used to address unanswered questions about well-defined hazards or data-rich chemicals and to evaluate emerging concerns about those hazards or chemicals.
APPLICATIONS OF READ-ACROSS FOR A DATA-POOR CHEMICAL
As discussed in Chapters 3 and 5, read-across involves the assessment of a chemical on the basis of its structural similarities to chemicals that have already been tested and takes into account any differences that might influence pharmacokinetics, metabolism, or toxicodynamics. The approach can be coupled with computational and high-throughput data to support or refute the read-across results (see Figure 5-5). Alkylphenols are used as example chemicals for this case study.
Alkylphenols are metabolites or persistent environmental breakdown products of alkylphenol ethoxylates, chemicals that were formerly used in detergents. A few of the more widely used alkylphenols, particularly p-octylphenol and p-nonylphenol, have a rich toxicology dataset. In this case study, p-octylphenol and p-nonylphenol are used as analogues to support the assessment of p-dodecylphenol, a data-poor chemical that has been tested in ToxCast. Both p-octylphenol and p-nonylphenol have weak affinity for estrogen receptors in vitro (Laws et al. 2000). In vivo reproductive-toxicity data on the two chemicals have conflicting results. Multigeneration studies run under good-laboratory-practice (GLP) conditions by National Toxicology Program (NTP) indicate a few effects on reproduction with lowest observed-adverse-effect levels in the oral-intake range of about 30–100 mg/kg-day (p-nonylphenol, Chapin et al. 1999; p-octylphenol, Tyl et al. 1999). Other studies show effects on the reproductive system, although by different routes, such as parenteral injection, or at higher oral doses (see, for example, Hossaini et al. 2003; Mikkilä et al. 2006). Thus, the critical end point for the read-across is reproductive toxicity with estrogenicity as the presumed mechanism.
p-Dodecylphenol is a related chemical on which there are few in vivo toxicity data. The KOW for p-dodecylphenol is higher than those of the other alkylphenols, but all are very hydrophobic (see Table B-1). The chemical structure of p-dodecylphenol is similar to those of p-octylphenol and p-nonylphenol; the difference is that it has four or three more carbons, respectively, on the alkyl chain. Chemical-similarity scores for straight-chain p-octylphenol or p-nonylphenol are in the range of 55–65%. The chemical similarity score is a measure of molecular similarity that is based on atom-by-atom matching and is a good starting point for molecular comparisons. However, there is no bright-line chemical-similarity score for analogue suitability; it should be considered with other factors, such as physical chemistry and specific molecular features that can dramatically change potential reactivity or biological activity. Wu et al. (2010) provide a series of heuristics for determining the suitability of analogues for read-across. The committee notes that the chemical-similarity scores in Table B-1 suggest that the branched p-nonylphenol might be inappropriate for read-across for p-dodecylphenol. However, it is included here because most models of estrogenicity would consider para-substituted phenol moieties to have a potential to interact with the estrogen-receptor binding site—see, for example, the decision-tree scheme of Wu et al. (2013).
ToxCast has data on the chemicals in Table B-1. In each case, the most sensitive assay (the assay that has the lowest AC501 ) was one that measured estrogenic activity, and all chemicals were active in several estrogen–response assays at concentrations below 10 μM. Estrogen response (such as binding to the receptor or activation of an estrogen response element) was by far the most preva-
1 AC50 is the concentration at which a 50% response is elicited in an in vitro assay.
lent response to all four chemicals in ToxCast. Those results are consistent with the predictions from a qualitative structure–activity relationship (SAR) program developed by the US Environmental Protection Agency (EPA) that classifies all the chemicals as having weak estrogenic activity on the basis of the presence of a para-substituted phenol and the known estrogenic activity of p-alkylphenols as a class. A few other assays had a strong positive concentration response and an AC50 at or below 10 μM (see Table B-2). Activity also included interactions with a retinoid X receptor (RXR) isoform, pregnane X receptor (PXR), a vitamin D receptor, and peroxisome proliferator-activated receptor gamma (PPAR-γ), and mitochondrial toxicity (see Table B-2).
In summary, the SAR and ToxCast data support grouping p-dodecylphenol with the other phenols as chemicals that appear to have a common mechanism, weak estrogenicity. The minor bioactivity observed with the other receptors (RXR, PXR, vitamin D receptor, and PPAR-γ) is not unexpected and emphasizes that the toxicant activity is typically multimodal. Even endogenous hormones that are considered to have high specificity for a particular receptor have comparable nonspecificity (Kelce and Gray 1997), and high-throughput assays provide the basis for evaluating other potential or unsuspected toxicities. The interactions at higher concentrations are probably not involved in toxicity. The overall in vitro potency of p-dodecylphenol as an estrogen appears to be higher by a factor of roughly 15 than that of the p-octyl and p-nonyl analogues, and it was active in 3 times as many estrogen-receptor assays. Because p-dodecylphenol is the most hydrophobic of the alkylphenols, its lower AC50 could be inaccurate (see references and discussion in Chapter 2 on challenges in interpreting in vitro test data), but the data indicate that its estrogenicity in vitro is in the range of the other alkylphenols tested.
Estrogenic responses of p-octylphenol and p-nonylphenol have been reported in numerous studies, including in vivo rat multigeneration studies conducted by NTP (Chapin et al. 1999; Tyl et al. 1999). For the present case study, the no-observed-adverse-effect levels (NOAELs)2 identified by the two NTP studies could be used as a starting point to derive a reference dose of p-dodecylphenol, although it should be noted that other published studies reported effects at lower doses. The studies were both feeding studies in which a dietary concentration of 200 mg/kg had no reproductive effects. Because the animals’ growth and food consumption changed over time, a range of doses
|Chemical||CAS Number||Log KOWa||CSSb|
bThe CSSs of analogues to test chemical (p-dodecylphenol) were calculated by using the Tanimoto coefficient from an online source (http://chemmine.ucr.edu). CSSs provide another line of evidence (quantitative) for using (or not using) visual read-across (qualitative) data.
Protein Interactions; AC50 values in μMa
|Chemical||ER||RXR||PXR||Vitamin D Receptor||PPAR-γ||Mitochondrial_Toxicity|
|Branched p-nonylphenol||0.517 (14)||1.4||2.29||1.98||–||6.3|
aNumber in parentheses is the number of estrogen-responsiveness assays with an AC50 less than 10 μM. Abbreviations: ER, estrogen receptor; PPAR-γ, peroxisome proliferator-activated receptor gamma; PXR, pregnane X receptor; RXR, retinoid X receptor.
2 The committee notes that a point of departure identified through benchmark-dose modeling could also be used.
(9-36 mg/kg-day) was associated with that concentration. Using the NOAELs as surrogates for p-dodecylphenol could require an adjustment for potency: the lowest AC50 for p-dodecylphenol was about one-twentieth of the lowest AC50 for p-octylphenol and p-nonylphenol, and this could require a comparable revision of the NOAEL.
Several limitations were identified in this read-across exercise. Improved estimations of the AC50 data by using in vitro mass-balance models could be prudent before adjusting the NOAEL. Adjustments of the NOAEL on the basis of possible differences in the pharmacokinetics of the chemicals should also be considered. Differences in logKOW of 2 orders of magnitude are likely to be important in the rate and extent of absorption and clearance, although in this case the hydrophobicity of all the chemicals is high enough that one would expect high oral absorption of all chemicals. Predicted estimates of absorption and clearance and NOAELs for estrogenic effects could be obtained from targeted testing or similarly focused studies to corroborate the inferences based on read-across. Finally, the uncertainty in read-across should be assessed to ensure consistency and appropriate conservatism (Blackburn and Stuard 2014).
An outcome of this read-across exercise could be classification of p-dodecylphenol as an estrogenic compound potentially more potent than the other alkylphenols. Establishment of a reference dose would be plausible, but additional information on metabolism, absorption, and developmental effects on estrogen-sensitive organs would improve confidence.
AIR-POLLUTION CASE STUDY
There is long-standing concern that exposure to air pollution might lead to chronic health effects, but only in the last several decades have epidemiological studies convincingly linked air-pollution exposure to premature mortality and increased risk of cardiovascular disease and cancer (EPA 2009). Beyond demonstrating hazard, recent studies have refined the characterization of the exposure–response relationship (Beelen et al. 2014). The new evidence reflects the increasing computing power that has enabled refinements in epidemiological methods, especially data-intensive exposure assessment that combines large-scale ambient monitoring of pollutants with advanced geographic information system (GIS) applications, dispersion models, and land-use regression (LUR) models to estimate exposures of large populations. Those methods—and decades of investment in nationwide air-pollution surveillance networks—have allowed researchers to establish long-term exposure models for large prospective cohort studies and to investigate long-term consequences of air pollution, such as cancer and cardiovascular disease, while controlling for major potential confounders. Studies based on those advances—exemplified by recent publications from the European Study of Cohorts for Air Pollution Effects (ESCAPE) consortium (Beelen et al. 2014)—have led a working group of the International Agency for Research on Cancer (IARC) to conclude that there is “sufficient” evidence to conclude that ambient air pollution is carcinogenic to humans and that the evidence is “sufficient” to conclude that airborne PM is carcinogenic to humans (IARC 2015).
The evidence on the causal relationship of air pollution with lung cancer (IARC 2015) is strong, and hazard identification is not at issue with regard to regulatory decision-making, at least in high-income countries with well-established evidence-based air-quality standards. However, there are a number of unanswered scientific questions concerning air pollution and cancer that are still relevant to regulatory decision-making; for these questions, 21st century science has the potential to reduce uncertainty around key issues relevant to tightening and targeting air-quality regulation. This particular case study illustrates how new and emerging science can be used to address lingering questions about well-defined hazards or data-rich chemicals and considers the following key issues:
- Identifying critical air-pollution sources and components. (1) Air pollution is a mixture that reflects its many sources; its composition varies by time and space. (2) The composition of the pollutant mix is not fully characterized, and research suffers from the “lamp-post syndrome” (that is, it has focused on a few target or indicator pollutants, such as EPA’s criteria pollutants, including PM and nitrogen dioxide). (3) There is potential for interaction and synergy among different components of the air pollution mixture with implications for overall mixture toxicity.
- Characterizing the exposure–response relationship. (1) On the basis of available epidemiological evidence, there is no apparent threshold for the long-term effects of air pollution at current levels in the United States and elsewhere, particularly on total mortality and on cancer (Raaschou-Nielsen et al. 2013; Beelen et al. 2014; Hamra et al. 2014). (2) The power to detect effects and characterize risks precisely at low exposures is difficult even in large cohorts, such as the ESCAPE and American Cancer Society cohorts. (3) There are various hypotheses about the possible mechanisms by which air pollution causes long-term adverse effects at current exposures, and the mechanisms are likely to vary by outcome and pollutant mixture. (4) Specific groups might be at greater risk because of particular characteristics, such as genetics, life stage, disease status, or co-exposure to other agents.
- Addressing emerging concerns. There is an expanding list of possible adverse health effects of long-term exposure to air pollution. For example, some evidence indicates possible adverse effects on neurodevelopment
This case study develops two parallel examples. One is based on lung cancer, and the main concerns are estimating the exposure–response relationship, especially at low exposures, as experienced in the United States and much of Europe and identifying mechanisms involved and key mixture components that might drive cancer risk. The second example, neurodevelopment in children, has been chosen for different reasons. The questions concern mainly hazard identification because causal associations with air pollution for any specific neurodevelopmental outcome are far from well-established. The uncertainties in a number of neurodevelopmental outcomes reflect the challenges in investigating rare but severe outcomes, such as autism, that require large pregnancy cohorts that have detailed air-pollution assessments and the difficulties in comparing results among studies that evaluate a large array of neuropsychological effects and cognitive function at different developmental ages in children exposed to various pollutant mixtures.
Lung Cancer: Characterizing the Exposure–Response Relationship and Identifying Key Mixture Components
Current epidemiological tools are unlikely to offer direct answers to the related problems of characterizing risk precisely at low doses and determining the shape of the exposure–response curve partly because there are limits to the size of cohorts that can be assembled and because exposure-measurement error is unavoidable with the available tools. However, those problems can be addressed with new and emerging approaches and tools described below that help to characterize exposure more precisely and to probe mechanisms more deeply.
One critical issue in characterizing the exposure–response relationship is defining exposures more precisely, particularly at low levels of exposure. New exposure-assessment approaches centered around the concept of the exposome can help to address that issue. As defined in Chapter 1, the term exposome refers to the totality of a person’s exposure. It is discussed here because of the emergence of new tools that provide time-integrated measurements of multiple pollutants at the individual level with greater spatial and temporal resolution than could be achieved previously (see Chapter 2). Such measurements potentially will help to characterize the exposure–response relationship better by reducing exposure-measurement error and by providing the needed inputs for measurement-error correction models.
The new exposure approaches contrast sharply with those used in past studies. Originally, epidemiological studies of air pollution relied on exposure classifications that were based on a few measurements in a few locations. Even the well-known Harvard Six Cities Study (Dockery et al. 1993), initiated in 1974, relied on central site measurements in the six selected cities. The wave of time-series studies that began about 3 decades ago fully incorporated the temporal detail of exposure measures but still used monitoring data that were limited spatially, such as central site monitors. Later cohort studies also incorporated more temporally refined measures, such as hourly or daily ambient monitoring station values, but again were spatially limited, often taken at one or a few stations per city. Citywide average exposures during specified periods were then applied to all residents in a design that would now be recognized as ecological or semiecological (that is, population-level assignment of exposure but with individual-level covariate information) (Künzli and Tager 1997). That approach, reflected in the Six Cities Study, ignores within-city variation and implicitly assumes that there is little spatial heterogeneity of air pollutants or that residents moved around cities enough to be similarly exposed to various pollutant sources. Neither assumption is correct in practice. Thus, measurement error was implicit in those studies, which nonetheless found associations with indicators of PM exposure, most likely because it was possible to exploit the high temporal resolution and fluctuations in air pollutants, especially in assessing short-term effects, such as in the time-series studies of mortality.
New tools are being developed to capture spatial variation in effects better (Coker et al. 2015). Early 21st century advances—such as GIS applications, dispersion models, and LUR models—have added a major refinement of capturing spatial variation in exposure assessment. Before those advances, exposures were generally assigned on the basis of residential location, and that practice accounted for some of the within-city variation. Reliance on residential location, however, did not fully capture or integrate exposures from multiple sources on larger geographic scales. For example, in Europe and the United States, investigators used tailored measurements of PM2.5 in a number of cities with multiple land-use characteristics of each area (traffic, ports, population density, and factories) to predict concentrations at individual addresses with reasonably good performance by using LUR models and sometimes adding a temporal component to the estimates with data from routine ambient monitoring (Raaschou-Nielsen et al. 2013). However, those measurements were affected by measurement error as suggested by comparisons with, for example, personal-exposure monitoring campaigns. The latter are based on the use of backpacks or similar devices containing instruments that measure exposure at the individual level with great temporal and spatial resolution; such campaigns are gen-
erally conducted for shorter periods, such as 2–4 weeks, for feasibility. The external exposome measures showed the complexity of capturing the entirety of personal exposure to PM. For example, cooking was shown to be an important source of exposure to ultrafine particles. Such studies added to earlier understanding that personal exposure to air pollution can vary widely in time and space and be driven by specific time–activity patterns, such as time spent at home, in traffic, at work, and in restaurants. Without an understanding of such variation, exposure estimates can be quite inaccurate and bias risk estimates (Nieuwenhuijsen et al. 2015). New personal devices that will measure a large variety of pollutants are under development, as reported in Chapter 2.
However, none of the new sensor technologies is likely to be feasibly implemented (in terms of data handling and security) at the individual level in large cohorts over the extended periods (decades) necessary to investigate risks of chronic disease outcomes. Studies that are sufficiently large and have the detailed exposure information needed to address the key questions related to lung cancer and air pollution are not likely to be undertaken. Crowd sourcing or anonymous data collection using sensors might be a feasible alternative if implemented within existing or new general cohorts. The resulting data would then be used to refine exposure models and estimates associated with participants in such cohorts. Possible limitations of such data-collection methods include sampling bias and measurement error in the devices that might feasibly be deployed (see NRC 2012 for a more detailed discussion of possible limitations). The committee anticipates further refinements of exposure estimates within cohort studies. The refinements might be achieved by including extensive time–activity data in sophisticated spatiotemporally refined pollution models and by controlling measurement error better, which would reduce one major contributor to uncertainty in the burden of lung cancer attributable to air pollution.
New and emerging approaches also will be helpful for addressing the other challenge noted above that is related to characterizing the specific mixture components and the corresponding sources that drive lung-cancer risk. Most evidence on the health effects of PM air pollution from epidemiological studies—for example, on lung cancer—is based on estimated PM mass as the indicator of exposure. But PM is a complex mixture, and particles of different size and compositions might differ in toxicity and carcinogenic potential. Furthermore, PM exists within the broader air-pollution mixture.
New modeling approaches can provide estimates of concentrations of various PM components and characteristics and facilitate the exploration of the relationships between specific PM components and health risk. Recent studies have comprehensively characterized sources of outdoor air pollution and incorporated LUR models for estimating ambient PM10, PM2.5, and nitrogen dioxide (Raaschou-Nielsen et al. 2016). Models have then been developed for elemental composition (x-ray fluorescence), elemental and organic carbon, polycyclic aromatic hydrocarbons (PAHs), benzene, and ultrafine particles, which have been studied little because of difficulties in exposure assessment (Chang et al. 2015). Exposure estimation for ultrafine PM is now possible with, for example, an innovative mobile monitoring design that has been shown to be reliable and cost-effective (Hudda et al. 2014).
There are opportunties to use new in vitro and in vivo assays to evaluate and compare toxicity of PM samples. One of the properties of particles likely to reflect toxicity is oxidative potential, a property for which novel assays have been developed that measure the reduction of antioxidants in lung-lining fluid (Kelly and Fussell 2015). By analyzing the spatial and temporal variability of the oxidative potential of PM in filters, one can characterize the determinants of that variation and develop new spatially resolved air-pollution models for oxidative potential (Yang et al. 2015).
The air-pollution models alone, however, provide information only on ambient outdoor-pollutant concentrations and do not incorporate data on locations of members of the population needed for an exposure-assessment approach that would integrate data on various spaces. The models do not specifically take into account indoor exposure sources or indoor exposures to outdoor pollutants that have penetrated indoors. Recent advances in GIS (for example, route modeling) and microenvironmental models (for example, indoor-to-outdoor exposures) have led to the development of more detailed personal-exposure models that can be fed by rich sources of detailed data on population time–activity patterns, which should reflect time spent indoors. Regarding outdoor exposures, many cities hold information on origin and destination travel details from prepaid card systems or survey data on travel. Combined with regional or national surveys on time-use, those data constitute a rich additional source for personalized exposure models. Detailed data on personal and population-wide air-pollution exposures and space–time activity patterns from monitoring campaigns are required to evaluate new exposure models and thus support their use in providing improved exposure estimates for epidemiological studies and risk assessment.
The internal exposome can be investigated with two broad approaches: directly with analytical chemistry (as described in Chapter 2) and indirectly with several -omics technologies. Direct measurement focuses on the exogenous chemicals that can be found in internal fluids and measured with great sensitivity given current analytical-chemistry methods. Indirect measurements are based on
changes in DNA, RNA, proteins, or metabolites from which exposure to particular exogenous chemicals can be inferred. Genomics, transcriptomics, epigenetics, and proteomics allow only indirect inferences on exposures, and metabolomics and adductomics might allow direct measurements.
The use of -omics technologies described in this appendix allows the study of changes—for example, in blood or urine—that can help to characterize adverse effects of air pollutants, to refine exposure, to identify mechanisms, and to identify groups at risk. Here, the committee describes the potential contributions of the different -omics technologies in relation to the regulatory issues raised above and provides a few examples intended to show the potential of the rapidly developing science. A systematic review on the topic was not possible, given the scope of the relevant literature and the rapid development of this field. See Chapter 1 for definitions of -omics technologies.
Carcinogenesis is understood to be a multistep process to which genetic and nongenetic changes contribute (see Smith et al. 2016). For lung cancer and air pollution, information on genetic determinants of risk would be useful for public-health protection. Genomics can be based on the systematic investigation of genetic (inherited) variants that lead to or increase susceptibility to air-pollution–related disease or can be based on the study of somatic mutations induced by air pollution in cells. Concerning inherited susceptibility, several genetic variants (such as GSTM1) have been investigated in the candidate gene era; more recently, variants have been identified thanks to genome-wide association studies (see, for example, Kachuri et al. 2016). The associations of genetic variants with lung cancer are mostly weak, but the findings of some variants associated with lung-cancer risk have identified groups in the population that are potentially more susceptible to carcinogens.
A potentially fruitful approach for identifying susceptible groups is to develop profiles of susceptibility that are based on genetic pathways. For example, Bind et al. (2014) used a pathway-analysis approach to investigate whether gene variants that are associated with such pathways as oxidative stress, endothelial function, and metal processing modified the association of PM exposure and fibrinogen, C-reactive protein, intercellular adhesion molecule-1, or vascular-cell adhesion molecule-1.
Concerning somatic (acquired) mutations, the sequencing of several types of cancer tissues has shown that mutational patterns can reflect environmental mutagens (Nik-Zainal et al. 2015). For example, lung cancer has a mutational pattern that strongly resembles that induced by benzo[a]pyrene (B[a]P) in in vitro assays that use immortalized mouse embryo fibroblasts (Nik-Zainal et al. 2015). The results revealed that B[a]P induces a characteristic mutation signature: predominantly G→T mutations for B[a]P as opposed to C→T and CC→TT for ultraviolet radiation and A→T for aristolochic acid, a carcinogenic and mutagenic compound. Thus, the study suggests that the carcinogenicity caused by smoking (and possibly air pollution) could be due to the PAH component in smoke (or ambient air). Mechanistically, that information is of great importance.
Genomics could thus prove useful in two ways. First, genetic (inherited) variants that contribute to modulating the cancer risk associated with air-pollution exposure could be identified. Identification of populations at greater (or less) risk would refine understanding of the exposure–response relationship and point to a susceptible population. Second, if a molecular signature in tumor tissue (somatic mutations) were linked specifically to air-pollution exposure, burden could be more effectively quantified and exposure–response models developed for particular phenotypes defined by etiology. The committee notes that substantial research indicates differences in mutational spectra of lung cancers between smokers and never smokers, although markers that are definitive for any specific type of environmental exposure have not yet been identified. Third, even if signatures are not identified, mechanistic insights that support biological plausibility further and perhaps provide insights concerning mixture components could be gained.
Environmental exposures are able to change epigenetic signatures, for example, the methylation pattern of DNA or chromatin. DNA methylation and the associated repressed or activated transcription of genes might affect carcinogenesis (Vineis et al. 2010). Changes in methylation of the aryl-hydrocarbon receptor (AHR) repressor gene show that methylation can be used as a marker of exposure to smoking (Shenker et al. 2013) and to monitor the effect of cessation of exposure (Guida et al. 2015). Some authors have used AHR repressor methylation as a marker for in utero exposure of the fetus to tobacco-smoke components from maternal smoking (Joubert et al. 2012). Epigenetic markers in cord blood and placental tissue could also be used to detect possible effects of air-pollution exposure on the fetus and might be useful in addressing the question of whether maternal exposure to air pollution leads to developmental effects (Novakovic et al. 2014). And epigenetic markers might provide information on exposure to air pollution and even particular components.
How informative epigenetics is in studying risks of disease or health outcomes depends on whether the markers are permanent, whether they develop during a critical age window, and whether the right tissue can be investigated; methylation markers are tissue-specific. A few studies have investigated the effects of air-pollution exposure on DNA-methylation patterns (see, for example, Baccarelli et al. 2009) and focused on methylation of long interspersed element-1 (LINE-1) and Alu elements as measures of whole-genome methylation in blood cells. LINE-1 and Alu elements are retrotransposons, that is, repetitive and mobile sequences in the genome. LINEs comprise a substantial proportion of the genome, and LINE-1 and Alu methylation correlates with overall cellular levels of DNA methylation. Air pollution was found to alter LINE-1 methylation (Baccarelli et al. 2009; Demetriou et al. 2012).
Epigenetic changes might also be integral to carcinogenesis, perhaps to the same extent as genetic mutations. Fasanelli et al. (2015) showed that the same genes (including the AHR respressor gene) for which methylation changes are associated with smoking predict lung-cancer risk. Similar studies are not available for air pollution and lung cancer.
Given the substantial current emphasis on the epigenome and the environment, the committee anticipates that the utility of epigenetics in risk assessment will be determined over the next decade. Studies that span the life course are in progress, and there is opportunity for marker validation over longer times, although this research would require multiple biological samples from well-characterized large cohorts.
Transcriptomics can lead to the identification of perturbations in gene expression relevant to lung carcinogenesis due to environmental exposures, including exposure to air pollution. Thus, transcriptomics is expected to be a key tool in research, for example, for identifying which specific components of an air-pollution mixture are biologically active and might have a role in causing lung cancer. Transcriptomics might also help to reveal interactions of mixture components by showing that the overall effect of a mixture on gene expression is greater than the sum of gene expression of the individual components.
Gene-expression changes have been linked to air-pollution exposures in in vitro and animal experiments. Specifically, exposure to air pollution leads to increased or decreased expression of genes that are relevant to immune or inflammatory actions. Although few observations have been made in humans, Wittkopp et al. (2016) performed an exploratory analysis and tested whether gene expression was associated with air-pollution exposures in a Los Angeles area cohort of elderly subjects who were exposed to PM2.5 at an average of 10-12 μg/m3. The authors found positive associations of traffic-related pollutants (including nitrogen oxides and PAH content in PM0.25–2.5 or PM0.25) with the expression of several candidate genes, particularly Nrf2-mediated genes, which indicated involvement of oxidative stress pathways. A number of genes have been found to be dysregulated by using transcriptomics tools in studying lung cancer (see, for example, Amelung et al. 2010).
As noted in Chapter 1, proteomics refers to the measurement of the whole compartment of proteins in a biological sample with high-throughput techniques. Like transcriptomics, it might be useful in characterizing toxicity of individual air-pollution components, identifying interactions of air-pollution components, and identifying pathways that might be involved in a response to air pollution and possibly related to lung carcinogenesis For example, the association between long-term exposure to air pollution and inflammatory markers was investigated with a proteomic approach (Mostafavi et al. 2015), and immune–inflammatory perturbations were observed at high exposures. Little work has been conducted on the proteome in relationship to air pollution.
DNA and protein adducts have long been measured in relation to air-pollution exposure (Demetriou et al. 2012; Demetriou and Vineis 2015). Specific adducts, such as PAH–DNA adducts, have been measured. Adductomics is a new approach to identifying exposure biomarkers with a systematic, high-throughput search of all potential adducts resulting from external exposures or internally generated compounds. As part of the exposome concept, adductomics typically involves an untargeted investigation that analyzes hydrolysis products of albumin by using mass spectrometry. Electrophilic chemicals or their metabolites that bind to albumin are also likely to bind to DNA. Thus, protein-based adductomics can potentially be used to identify genotoxic, electrophilic components in a mixture. Adductomics might also be helpful in refining exposure–response relationships, including the shape of the exposure–response curve for lung cancer, because the high sensitivity of adductomics reduces misclassification and uncertainty. That research would require repeat samples from prospective cohorts, and one of the pillars of modern epidemiology is the availability of large prospective cohorts with multiple samples that create an opportunity to study the stability of signals. Some of the markers integrate exposures over relatively long periods and would thus be useful for exposure estimation.
Metabolomics can be performed on plasma, serum, or urine samples by several methods, including high-resolution mass spectrometry coupled to ultra-high-performance liquid chromatography for untargeted analyses. Metabolic features that characterize exposed groups are identified by multivariate statistics with appropriate correction for false discovery rate. Metabolites unique to exposed groups are then identified with more targeted investigations. However, metabolomics data are subject to high intraindividual variability, and many metabolites have short lives, which might limit their utility in estimating longer-term exposures. Annotation is another limiting factor; researchers are unable to characterize features detected with, for example, mass spectrometry without additional chemical analyses. In principle and with likely future technical developments, however, metabolomics could become a useful tool for achieving several goals, as suggested in Table B-3: the identification of specific metabolites related to mixture components and their interactions, better characterization of exposure by linking metabolites to external measurements, and reconstruction of molecular and biochemical pathways, which would contribute to mechanistic knowledge and identification of pathways.
Early and still evolving findings from epidemiological research that uses -omics techniques are starting to suggest that air pollutants might act via pathways that involve inflammation and oxidative stress. In addition, there might be mutational signatures that are characteristic of air-pollution exposure vs, for example, smoking, although air pollution and cigarette smoke have several common components, such as PAHs. The small samples of early studies, however, do not allow sound quantitative estimation of pathway perturbations at low doses. Although the evidence is limited, some consistency is emerging among different -omics platforms, such as transcriptomics, epigenomics, and proteomics. The consistency among platforms can be investigated by using statistical techniques known as cross-omics (Vineis et al. 2013). The long-term goal is to couple external exposome approaches to reduce measurement error at the individual level with a suite of -omics investigations that characterize the various steps involved in carcinogenesis by investigating, for example, mutational spectra, epigenetic changes, inflammation, and cell proliferation in human samples. That research is expected to lead to more accurate quantitative risk assessment.
Overall, -omics technologies will facilitate exploration of all the characteristics of carcinogens and the pathways that lead from exposure to diseases. The main challenges are related to the variability of measures due to technical reasons and biological intraindividual variation, the long latency of cancer with decades between exposure and disease onset and the multiple steps involved, and the lack of access to precursor lesions—there is access only to surrogate tissues, such as blood—to study molecular changes that take place in target cells. Regardless of the
|Regulatory Question||-Omics Technologies|
|Identifying Critical Air-Pollutions Sources and Components|
|Characterize toxicity and long-term effects of mixture components|
|Investigate interaction potential of mixture components|
|Characterizing the Exposure–Response Relationship|
|Characterize exposure better|
|Identify groups at greater risk|
aThis table is related to the current knowledge and uses of -omics in the field of lung carcinogenesis. Assignment of checkmarks in the table is likely to change with advances in the science of -omics and in the understanding of lung carcinogenesis.
challenges, the -omics technologies offer opportunities to identify critical components of air-pollution mixtures and to refine the exposure–response relationship as illustrated in Table B-3.
Neurodevelopmental Effects and Particulate Air Pollution: Determining Whether a Causal Relationship Exists
Determining whether there is a causal relationship between neurodevelopmental effects and PM is potentially of great public-health importance. It has long been known that fetuses, infants, and young children are more sensitive than adults to diverse environmental toxicants because of the vulnerability accompanying developmental, growth, and maturation processes (WHO 1986; NRC 1993; Anderson et al. 2000; Perera et al. 2004; Grandjean and Landrigan 2006). One topic of particular concern is neural development. A large body of research has addressed the influences of air pollution on fetal growth, including head circumference (Vrijheid et al. 2011; Stieb et al. 2012; van den Hooven et al. 2012; Backes et al. 2013; Proietti et al. 2013; Smarr et al. 2013). More recently, epidemiologists have become interested in potential effects of PM air pollutants because some combustion components of PM, such as PAHs and their derivatives, have shown neurodevelopmental toxicity in some experimental and small pathology studies (Calderon-Garciduenas et al. 2002; Takeda et al. 2004). In this section, the committee briefly discusses the epidemiological studies that have linked air-pollution exposures to neurodevelopmental effects and offers some suggestions on how ES21 and Tox21 tools and methods could be used to strengthen or improve the epidemiological studies. The committee notes that epidemiological studies that address neuropsychological effects of air pollution have been summarized by Guxens and Sunyer (2012) and Suades-González et al. (2015) and are not discussed here. The section concludes with some general considerations related to developmental neurotoxicity (DNT) and possible approaches for studying DNT.
Epidemiological Evidence of Associations Between Air Pollution and Neurodevelopment in Children
Epidemiological studies have begun to investigate the association between various air pollutants and neurodevelopmental effects in children. The characteristics and designs of the key studies are summarized in Table B-4. Several small cohort studies in the United States, Poland, and China have shown adverse neurodevelopmental effects in children exposed in utero to PAHs (Perera et al. 2006, 2009; Tang et al. 2008, 2014; Edwards et al. 2010; Lovasi et al. 2014). PAH exposure in the studies was measured through short-term (48-hour) personal-exposure measurements during pregnancy or as PAH–DNA adducts in cord blood. The adverse effects reported were decreases in mental function or IQ and motor developmental delays early in childhood, but these effects were not observed consistently at all ages at which the children were examined. An additional cohort study in the United States linked adverse neuro-developmental effects (IQ and attention disorders) in children with increases in children’s lifetime exposure to black carbon, which is related to traffic (Suglia et al. 2008; Chiu et al. 2013); however, only in boys was black-carbon exposure associated with attention disorders, and this suggests possible sex-specific vulnerability. A large European study combined six birth cohorts (Guxens et al. 2014) and reported that nitrogen dioxide, but not other air pollutants, was associated with delayed psychomotor development in children 4 years old and younger; no associations with cognitive or language development were seen. In addition, several Asian studies and a Polish study reported associations of different types of air pollutants and exposure periods with various developmental outcomes (see Table B-4 below). Most of the studies were small, tested children at different developmental ages and for different functions or disorders, and measured exposures prenatally or postnatally, focusing on different pollutants and sources. Thus, additional studies are needed to replicate or confirm some of the reported findings before conclusions about associations of air pollution with adverse neurodevelopment outcomes can be drawn from epidemiological data.
The limitations of the epidemiological studies might be addressed by ES21 and Tox21 approaches. The following paragraphs summarize the challenges and possible approaches to addressing them.
- Studies testing children’s neuropsychological function at different ages are time-consuming and expensive, and researchers have to balance various factors, such as the extent and variety of functional assessments, cohort size, and length of follow-up. Feasibility and costs are major concerns. Those problems are exemplified in the most recent review of epidemiological studies (Suades-González et al. 2015), which still did not identify sufficient data to conduct quantitative meta-analyses because of heterogeneity in the methods used to assess exposures and outcomes. With respect to cognitive and psychomotor development, Suades-González et al. (2015) decided that for only one exposure (PAHs) were there enough high-quality studies available to conclude that there was “sufficient evidence” of an association but not a causal relationship. For other air pollutants, modern exposure assessment and modeling—GIS or dispersion modeling supported by satellite data and ground-level monitoring networks—might facilitate adding comparable air-pollution exposure measures to those completed or current expensive human studies of neurodevelopment (for example, studies using neuroimaging or extensive functional testing). Eventually, the research conducted might
|Study Characteristics||Exposure Details||Principal Outcomes Investigated||Selected Findings||Reference|
|N = 46,039 singleton births in Japan on January 10–17 or July 10–17, 2001||Evaluated maternal exposure to air pollution related to municipality-level traffic, including PM, NO2, CO, and SO2 in the 9 months before birth. Air-pollution measurements were taken from general and roadside stations nationally.||Milestone delays were measured through a series of questions administered at ages 2.5 and 5.5 years. Questions were not validated or selected from an established scale, but have been used in previous studies.||Estimated air-pollution exposure during gestation was positively associated with some risk of several developmental milestone delays at both ages—verbal and fine motor development at age 2.5 years and behaviors related to inhibition and impulsivity at 5.5 years.||Yorifuji et al. 2016|
|N = 183 children, 3 years old, born to black and Dominican women in New York, NY, mother–child pairs recruited in 1998-2003||
Evaluated prenatal exposure to airborne PAHs, secondhand tobacco smoke, and pesticides; PAHs were monitored during pregnancy with personal air sampling.
Umbilical cord blood was taken at delivery, and maternal blood within 2 days postpartum was analyzed for cotinine, heavy metals, and pesticides.
|The Bayley Scales of Infant Development- Revised were used to assess cognitive and psychomotor development at ages 12, 24, and 36 months to generate an MDI and corresponding PDI. Behavioral problems were measured on the Child Behavior Checklist.||Prenatal exposure to PAHs of the mothers was not associated with PDI or behavioral problems. However, high prenatal exposure to PAHs (the upper quartile of the distribution) was associated with lower MDI at the age of 3 years, but not 1 or 2 years.||Perera et al. 2006|
N = 249 children, 5 years old, born to black and Dominican women in New York, NY, mother–child pairs recruited 1998–2003.
Note: This cohort is the same as Perera et al. 2006.
|PAHs were measured in women in their third trimester with a personal monitoring device during the daytime hours for 2 consecutive days; monitor was placed near the bed at night. Pumps operated continuously during this period, collecting vapors and particles ≤ 2.5 μm in diameter.||The WPPSI-R was used to determine verbal, performance, and full-scale IQ scores.||Women who had higher exposure to PAHs during pregnancy were significantly more likely to have infants with lower full-scale and verbal IQ scores tested at the age of 5 years. After adjustment for maternal intelligence, quality of the home caretaking environment, environmental tobacco-smoke exposure, and other potential confounding factors, high PAH levels (above the median of 2.26 ng/m3) were significantly and inversely associated with full-scale and verbal IQ scores but not with performance IQ scores.||Perera et al. 2009|
N = 326 children, born to black and Dominican women in New York, NY in 1998–2006.
Note: This cohort is the same as Perera et al. 2006.
|PAH exposures were measured with personal ambient air monitors worn for 2 consecutive days and placed at the bedside at night during the third trimester of pregnancy. Housing disrepair was self-reported by mothers, and neighborhood characteristics were estimated within a 1-km network from the prenatal address overlaid with data from the 2000 US Census. Indicators measured included number of residents below the federal poverty line, high-school diploma or equivalent degree attained, and low neighborhood English-language proficiency.||The WPPSI-R was used to assess intelligence and neurodevelopment at of age 5 years. Spanish scores were excluded because of difference in the Spanish- and English-language versions.||Prenatal PAH exposure above the median was significantly associated with lower total WPPSI-R and verbal scores. The mean differences were 3.5 total points and 3.9 verbal points between high and low PAH exposure groups, respectively.||Lovasi et al. 2014|
|N = 214 children born to women in Krakow, Poland||Exposure to eight PAHs was measured with personal air monitors carried over a 48-hour period during the second or third trimester of pregnancy; monitors were kept at the bedside at night during this period.||At age 5 years, RCPM were used to assess a child’s nonverbal reasoning ability.||A higher prenatal exposure (above the median of 17.96 ng/m3) to airborne PAHs (range, 1.8–272.2 ng/m3) was significantly associated with decreased RCPM scores at the age of 5 years, after adjustment for potential confounding variables. This corresponds to an estimated average decrease of 3.8 IQ points.||Edwards et al. 2010|
N = 1,257 US children, 6–15 years old; data collected from 2001–2004 cycles of NHANES.
|PAH exposure was based on urinary metabolite concentrations measured in the 2001–2002 and 2003–2004 cycles.||Outcomes were measured by parental reporting of (1) ever doctor-diagnosed ADHD (2) ever doctor- or school representative-identified LD and (3) receipt of SE or early intervention services.||Higher concentrations of fluorine PAH metabolites in children were associated with 2-fold increased odds of needing SE, somewhat more in males than in females.||Abid et al. 2014|
|N = 202 children in Boston, MA, participating in a prospective birth cohort study (1986–2001)||Exposure to BC was estimated with a model on the basis of child’s residence during study follow-up. Data collected from more than 80 locations in the greater Boston area were used to complete a spatiotemporal LUR model to predict 24-hour measures of traffic exposure.||Cognitive tests were administered at ages 8–11 years and included the K-BIT (assesses verbal and nonverbal intelligence) and the WRAML (evaluates a child’s ability to actively learn and memorize a variety of information).||With adjustment for sociodemographic factors, birth weight, blood lead concentration, and tobacco smoke, BC exposure was associated with decreases in the vocabulary (-2.2), matrices (-4.0), and composite intelligence quotient (-3.4) scores of the K-BIT and visual subscale (-5.4) and general index (-3.9) of the WRAML.||Suglia et al. 2008|
N = 174 children, 7–14 years old in Boston, MA.
Note: This cohort is the same cohort as Suglia et al. 2008
|Traffic-related black carbon (BC) concentrations were estimated over child’s lifetime using a spatiotemporal model for 24-hour measures of BC based on 6,021 observations from >2,079 unique exposure days at 82 locations in greater Boston area. Models took into consideration warm (May–October) and cold (November–April) seasons.||The Conners’ CPT was used to assess attention disorders and neurological functioning at ages 7–14 years.||In this population of urban school-aged children, there was a positive association between higher BC and increased commission errors and lower HRT, even after adjustment for child IQ, age, sex, and other variables. Sex-stratified analysis showed statistically significant associations between BC and both commission errors and HRT in boys, but BC was not significantly associated with any outcomes in girls.||Chiu et al. 2013|
|N = 9,482 children in six European population-based birth cohorts: the Netherlands, Germany, France, Italy, Greece, and Spain; mother–infant pairs recruited in 1997-2008.||
LUR models were used to estimate NOx in all study regions and PM with diameter <2.5, <10, and 2.5–10 μm, and PM2.5 absorbance in subregions. Monitoring campaigns took place primarily from October 2008 to January 2011.
NOx was measured at least three times per week for 2 weeks within 1 year. PM2.5 absorbance was measured in a subgroup of regions by reflectance of PM2.5 filters. To obtain final analyses, a back-extraction procedure was used to estimate the concentrations during each pregnancy of each woman.
|Cognitive and psychomotor development was assessed at ages 1–6 years. Different neuropsychological tests for cognitive and psychomotor development were administered, including McArthur Communicative Development Inventory, Bayley Scales of Infant Development I–III editions, Denver Developmental Screening Test II, McCarthy Scales of General Cognition, and Ages and Stages Questionnaire.||Air-pollution exposure during pregnancy, particularly NO2 (of which traffic is a major source) and PM2.5, was associated with delayed psychomotor development in children (-0.68 points in the global development score) for each 10 μg/m3 increase in NO2). Cognitive development measured at similar ages was not related to air-pollution exposure during pregnancy.||Guxens et al. 2014|
|N = 520 mother–child pairs in three regional centers in South Korea studied in January 1, 2006–December 31, 2008||Exposure to PM10 and NO2 during pregnancy was estimated with inverse distance-weighting method. A mini-volume air sampler was used to measure outdoor ambient PM10; a passive sampler was used to measure outdoor ambient NO2; sampling was performed over 24 hours.||The Korean Bayley Scale of Infant Development II was used to measure neurodevelopment progress. Results were expressed as MDI and PDI at 6, 12, and 24 months.||There was a negative association between maternal exposure to PM10 and MDI and PDI throughout the first 24 months of life. Maternal NO2 exposure was associated with impairment of PDI but not with cognitive function. A multiple-linear-regression model showed significant effects of prenatal air-pollution exposure (PM10 and NO2) on MDI and PDI at 6 months, but no significant associations were found at 12 and 24 months.||Kim et al. 2014|
|Study Characteristics||Exposure Details||Principal Outcomes Investigated||Selected Findings||Reference|
|N = 533 mother–infant pairs in 29 villages or cities in Taiwan selected in October 2003–January 2004; followed up at 6 and 18 months.||Hourly ambient concentrations of CO, O3, PM10, SO2, NO2, THCs, and NMHCs were measured at the Taiwan Air Quality Monitoring Network. Participant exposure was considered to be the average taken during the period 7 am to 7 pm. Air-pollutant exposure for each child was measured by linking data from the air-quality monitoring stations of the town to the exposure period from the beginning of gestation to 18 months after birth. The gestational period was divided into 3 trimesters, and the postpartum ranges were birth–6 months, 7–12 months, and 13–18 months.||Neurodevelopmental performance was measured by parent responses to a screening instrument, the TBCS. The scale consists of four developmental divisions: gross motor, fine motor, language/communication, and social/self-care abilities. Parents completed two neurobehavioral development scales at each interview; responses consisted of never, sometimes, and all the time. Scales have good predictive validity, and dimensions correlate with the Bayley Scales of Infant Development.||Various indexes of ambient air pollution, even low SO2 exposure, during pregnancy and up to the age of 12 months were associated with poor subclinical neurodevelopment (neurobehavioral effects and poor gross motor development) in early childhood.||Lin et al. 2014|
|N = 133 children born March 4, 2002–June 19, 2002, in three Tongliang, China county hospitals; followed for 2 years||Study carried out in an area in China with a seasonally operated coal-fired power plant. PAH–DNA adducts, Pb, and Hg were measured in umbilical-cord blood samples collected at delivery. HPLC was used to analyze B[a]P–DNA adducts in extracted white blood cell DNA. A PE-800 Zeeman atomic absorption spectrometer with background correction system was used to measure Pb in samples.||Physical, emotional, and behavioral development of 2-year-old children was measured with the GDS. Children received DQs for each of motor behavior, language behavior, personal behavior, and social behavior.||Increased cord adduct concentration was inversely associated with decreases in the motor area DQ, language area DQ, and average DQ after adjustment for cord lead concentration, environmental tobacco smoke, sex, gestational age, and maternal education level. High cord blood lead was also significantly associated with decreased social area DQ and average DQ. The frequency of developmental delay ranged from 9.1% (social) to 13.6% (motor), with an average score of 6.4%.||Tang et al. 2008|
N = 150 children born March 4, 2002–June 19, 2002, compared with a cohort of 158 children born March 2, 2005–May 23, 2005; both cohorts consisted of children born in Tongliang, China.
Note: This cohort is the same as Tang et al. 2008
|Two mini-volume samplers were used at three sites in Tongliang in March 2002–February 2003 and in March 2005–February 2006 to collect 72-hour PAH samples. Overall PAH concentrations were measured by analyzing B[a]P–DNA adducts in extracted white blood cells collected from the umbilical cord at delivery and from the mother within 1 day postpartum.||Birth weight, length, and head circumference were measured at birth or more than once after birth if the child was delivered by cesarean section. Neurodevelopment was measured with the GDS at the age of 2 years. As above, DQs were developed for motor, adaptive, language, and social behavior.||The power plant was closed between the recruitment of the two cohorts. Patterns of developmental delay in all DQ areas except language were improved in the 2005 post-shutdown cohort compared with the 2002 cohort.||Tang et al. 2014|
Abbreviations: ADHD, attention deficit hyperactivity disorder; BC, black carbon; CO, carbon monoxide; CPT, Continuous Performance Test; DQ, developmental quotient; GDS, Gesell Developmental Schedules; Hg, mercury; HPLC, high-performance liquid chromatography; HRT, hit reaction time; K-BIT, Kaufman Brief Intelligence Test; LD, learning disability; LUR, land-use regression; MDI, mental-development index; NHANES, National Health and Nutrition Examination Survey; NMHC, nonmethane hydrocarbon; NOx, nitrogen oxides; NO2, nitrogen dioxide; O3, ozone; PAH, polyaromatic hydrocarbon; Pb, lead; PDI, psychomotor-development index; PM, particulate matter; RCPM, Raven Coloured Progressive Matrices; SE, special education; SO2, sulfur dioxide; TBCS, Birth Cohort Study Scale; THC, total hydrocarbon; WPPSI-R, Wechsler Preschool and Primary Scale of Intelligence-Revised; WRAML, Wide Range Assessment of Memory and Learning.
provide sufficient sample size, appropriate exposure gradients, and possibly information about source-specific or chemical-specific pollution components to generate results that allow quantitative or causal evaluation of air pollutants and neurodevelopment.
- Key limitations in many DNT studies of air pollution are that they cannot address multiple air-pollutant exposures (mixtures) and most likely can ascertain potential confounders only incompletely, given the limited knowledge of social and cultural determinants of neurodevelopment and the strong association of neurodevelopment with socioeconomic status (SES). GIS could help to disentangle the role of SES by allowing, for example, area-level adjustment for correlates of SES. Computer-resource–intensive multilevel spatial modeling in a Bayesian framework might also allow addressing spatially correlated confounders and pollutant mixtures (Coker et al. 2015, 2016).
- In future studies with smaller samples, it might be possible to use personal air monitoring or biomarker approaches that include new sensor technologies if instruments are small and lightweight and if measurements are less expensive and thus feasible. The new approaches would allow monitoring over extended periods in pregnancy or early life. With the exception of PAH adducts, there are no good biomarkers for toxic PM components. Monitoring only particles does not allow assessment of the toxicity of their components, and particle composition probably depends on the sources that generate the particles. However, combining continuous particle monitoring with repeated collection of relevant biosamples (such as maternal and infant blood, urine, and placenta) would also allow the use of -omics tools to find new exposure biomarkers in human samples and possibly some biomarkers predictive of outcomes (see, for example, Janssen et al. 2015; Saenen et al. 2015). Nontargeted approaches might be useful for identifying new biomarkers.
General Considerations Related to Developmental Neurotoxicity and Possible Assessment Approaches
Historically, establishing causal linkages between neurodevelopmental disorders and environmental exposures, such as exposure to air pollution, has been difficult for a variety of reasons, including the need for large populations in epidemiological studies, the complexity of capturing the full array of relevant exposures before and during pregnancy, the long latency between exposure and effect (particularly for neurodegenerative disorders), the lack of defining pathology of some disorders (such as schizophrenia or autism spectrum disorder), and inherent limitations of animal models and in vitro assays. Perspectives and strategies for assessing DNT more comprehensively have been published by various stakeholders and will not be recapitulated here (Aschner et al. 2010; Bal-Price et al. 2015; Felter et al. 2015). This discussion highlights the unique challenges associated with trying to assess DNT and provides some possible approaches to doing so.
The most notable challenge unique to brain and behavioral targets is the dynamic complexity of the developing brain and a fundamental lack of understanding of the etiology of complex behavioral disorders, such as intellectual disability and emotional impairment. A disease-centric approach to DNT risk assessment is particularly challenging and unlikely to be feasible because many neural disorders, especially neuropsychiatric disorders, are syndromes with a spectrum of hallmark features and lack defining neuropathology or clear etiology. Thus, it is not plausible or rational to use a framework that attempts to make clear linkages between exposure, DNT mechanisms, and neural disease. Only a few such models have been proposed for DNT, and they are all too general (for example, oxidative stress) and do not explain the pathology well. Furthermore, the evidence does not support their acceptance with confidence, particularly in the neuroscience community. Instead, risk assessment of and chemical screening for DNT will have to be conducted in recognition that in the absence of an extraordinary situation (major accident or industrial exposure) clear linkages between exposure and a clinically diagnosed neural disease will be challenging.
Although perspectives on how to improve DNT risk assessment in a regulatory context differ, there is general agreement that testing for DNT should focus on evolutionarily conserved, fundamental events in neurodevelopment. Those events include neural induction, cell migration, axonal guidance, synapse formation and pruning, and apoptosis. Perturbation of the critical events underlies the primary deficits in neural disorders. Given that perspective, developmental neurotoxicants would be identified by their capacity to alter the fundamental events, regardless of their specific cellular or molecular mechanisms. Examples in which that perspective has yielded critical insight in connection with air pollution include evidence that PM2.5 induces oxidative stress in homogenates of rat brain regions and disrupts blood–brain barrier integrity, thereby enhancing neurotoxicity by activated macrophages and microglia (Fagundes et al. 2015; Liu et al. 2015). In mice, developmental exposure to ultrafine particles induced sex-specific neurotoxicity (including excitotoxicity and glial activation) and behavioral changes indicative of heightened impulsivity and hyperactivity—behavioral changes also associated with exposure of children to air pollution (Allen et al. 2014). Furthermore, in utero exposure to B[a]P during peak periods of neuro-genesis in mice leads to behavioral learning deficits (Mc-Callister et al. 2016).
Rapidly evolving experimental, epidemiological, computational, and toxicity-screening strategies are poised
to assess neurotoxicity and neuroendocrine disruption better and to fill critical testing gaps. Thus, DNT is a topic in which the application of Tox21 approaches would be particularly opportune and advantageous. For example, neuroinflammatory responses to air pollution have now been observed in human, animal, and in vitro studies (Costa et al. 2014); the results suggest the potential for contributions of Tox21 approaches that include the use of animal models and human tissues to assess DNT risks posed by air pollution and other exposures.
Tox21 approaches, including DNT assays, could also be used to address the challenges of identifying the air-pollution components that are contributing to neural disease. They could allow rapid testing of specific particle neurotoxicity and could help to identify markers of particle sources responsible for greater toxicity. For example, little is known about what PAHs are present in exposure mixtures; environmental samples can contain hundreds of individual parent or substituted PAHs, and bioactivity and toxicity of PAHs depend heavily on chemical structure (Wang et al. 2011). New methods could increase our understanding of the structure and toxicity relationships of neurobehavioral deficits if the full suite of chemicals present in samples could be identified and their individual or composite activities understood. Specifically, a suite of in vitro and high-throughput integrated systems could be used to classify PAHs by identifying their biological targets or pathways. Those systems could initially use untargeted global assessments—such as proteomics, metabolomics, transcriptomics, and epigenomics—to identify activity signatures for chemical classification and modeling. Recent studies in zebrafish, for example, evaluated and compared the developmental toxicity of 38 oxy-PAHs and revealed patterns of responses associated with PAH structural features (Knecht et al. 2013). In addition, full-genome RNA-sequence studies in zebrafish revealed that even for PAHs that produce toxicity through binding and activation of the AHR, subtle differences in PAH structure yield different overall developmental gene-expression changes and indicate that measuring P450 induction as a measure of AHR activation might be problematic (Goodale et al. 2015). Once targets of individual PAHs are identified, Tox21 approaches might be exploited further to predict how mixtures of PAH interact to produce neurotoxicity. In vitro functional assays of nervous-system development and function could be implemented to identify chemicals and mixtures that alter end points relevant to the nervous system. High-throughput integrated systems, such as zebrafish, might play a pivotal role in connecting identified molecular-response data with neurobehavioral measures (Truong et al. 2014; Reif et al. 2016). Optimization and scale up of assays that probe more complex behaviors in adult zebrafish (discussed in Chapter 3) should provide new avenues to link chemical exposures to functionally relevant neurobehavioral end points.
Despite enthusiasm for improving testing approaches and the emergence of new assays for DNT, implementation has been slow. For example, lack of assay coverage in EPA’s ToxCast for neurotoxicity end points or neuronal targets is a well-recognized limitation. An initial attempt to use the ToxCast data to rank tested chemicals in terms of neurotoxicity failed because of poor assay coverage of suitable end points and low reliability of existing assays (Filer et al. 2014). Stakeholder meetings and workshops have helped to identify better ways to integrate emerging tools and approaches for DNT but require the inclusion of more neuroscientists and developmental endocrinologists to ensure that fundamental pathways in neurophysiology are evaluated and that sexual dimorphisms, region-specific sensitivity, and dynamic critical windows of exposure are considered in assay development (Crofton et al. 2014; McPartland et al. 2015). A battery of assays that incorporates the most up-to-date neuroscience tools and principles and that provides data relevant for regulatory science and risk-based decision platforms will be needed. Identifying and leveraging the most promising approaches and technologies will require active engagement of experts in disciplines outside traditional toxicology, especially the neurosciences. Accomplishing a multidisciplinary approach and encouraging a multidisciplinary research program for assay development and evaluation can be achieved by coordinating with relevant scientific societies and groups that have the needed expertise and with relevant funding agencies, such as the National Institute of Environmental Health Sciences.
How the adult human brain accomplishes complex cognitive and social processing remains mysterious and is the focus of intense research that is using a broad array of tools. Even less is known about when key aspects of the complex systems are organized in development or about how sexual dimorphisms emerge (Reinius and Jazin 2009; Yang and Shah 2014; Hawrylycz et al. 2015; Loke et al. 2015). The role of glia is also gaining substantial attention because these cells, particularly astrocytes and microglia, appear to play a more fundamental role in neural development than previously thought (Schwarz and Bilbo 2012; Schitine et al. 2015). Thus, assessments of neurodevelopmental consequences of chemical exposures must be undertaken with the understanding and acceptance of the fact that fundamental understanding about how the brain develops remains to be achieved, let alone how it enables us to engage in uniquely human behaviors and what contributes to the cognitive and social capacities that define our species. More research is needed on DNT, particularly given its critical consequences and society’s high level of concern about its adverse effects. Addressing the challenges associated with DNT will require collaborative engagement of a broad array of disciplines, from neuroscientists who can address fundamental questions about the vulnerability of the brain to exogenous chemi-
cal exposures to population scientists who can assess the effects of chemical exposures in human populations.
Abid, Z., A. Roy, J.B. Herbstman, and A.S. Ettinger. 2014. Urinary polycyclic aromatic hydrocarbon metabolites and attention/deficit hyperactivity disorder, learning disability, and special education in US children aged 6 to 15. J. Environ. Public Health 2014:628508.
Allen, J.L., X. Liu, D. Weston, L. Prince, G. Oberdörster, N.J. Finkelstein, C.J. Johnston, and A. Cory-Slechta. 2014. Developmental exposure to concentrated ambient ultrafine particulate matter air pollution in mice results in persistent and sex-dependent behavioral neurotoxicity and glial activation. Toxicol. Sci. 140(1):160-178.
Amelung, J.T., R. Bührens, M. Beshay, and M.A. Reymond. 2010. Key genes in lung cancer translational research: A meta-analysis. Pathobiology 77(2):53-63.
Anderson, L.M., B.A. Diwan, N.T. Fear, and E. Roman. 2000. Critical windows of exposure for children’s health: Cancer in human epidemiological studies and neoplasms in experimental animal models. Environ. Health Perspect. 108(Suppl. 3):573-594.
Aschner, M., K.M. Crofton, and E.D. Levin. 2010. Emerging high throughput and complementary model screens for neurotoxicology. Neurotoxicol. Teratol. 32(1):1-3.
Baccarelli, A., R.O. Wright, V. Bollati, L. Tarantini, A.A. Litonjua, H.H. Suh, A. Zanobetti, D. Sparrow, P.S. Vokonas, and J. Schwartz. 2009. Rapid DNA methylation changes after exposure to traffic particles. Am. J. Respir. Crit. Care Med. 179(7):572-578.
Backes, C.H., T. Nelin, M.W. Gorr, and L.E. Wold. 2013. Early life exposure to air pollution: How bad is it? Toxicol. Lett. 216(1):47-53.
Bal-Price, A., K.M. Crofton, M. Sachana, T.J. Shafer, M. Behl, A. Forsby, A. Hargreaves, B. Landesmann, P.J. Lein, J. Louisse, F. Monnet-Tschudi, A. Paini, A. Rolaki, A. Schrattenholz, C. Suñol, C. van Thriel, M. Whelan, and E. Fritsche. 2015. Putative adverse outcome pathways relevant to neurotoxicity. Crit. Rev. Toxicol. 45(1):83-91.
Beelen, R., O. Raaschou-Nielsen, M. Stafoggia, Z.J. Andersen, G. Weinmayr, B. Hoffmann, K. Wolf, E. Samoli, P. Fischer, M. Nieuwenhuijsen, P. Vineis, W. Xun, K. Katsouyanni, K. Dimakopoulou, A. Oudin, B. Forsberg, L. Modig, A.S. Havulinna, T. Lanki, A. Turunen, B. Oftedal, W. Nystad, P. Nafstad, U De Faire, N. Pedersen, C.G. Östenson, L. Fratiglioni, J. Pennell, M. Korek, G. Pershagen, K.T. Eriksen, K. Overvad, T. Ellermann, M. Eeftens, P. H. Peeters, L. Meliefste, M. Wang, B. Bueno-de-Mesquita, D. Sugiri, U. Krämer, J. Heinrich, L. de Hoogh, T. Key, A. Peters, R. Hampel, H. Concin, G.Nagel, A. Ineichen, E. Schaffner, N. Probst-Hensch, N. Künzli, C. Schindler, T. Schikowski, M. Adam, H. Phuleria, A. Vilier, F. Clavel-Chapelon, C. Declercq, S. Grioni, V. Krogh, M. Tsai, F. Ricceri, C. Sacerdote, C. Galassi, E. Migliore, A. Ranzi, G. Cesaroni. C. Badaloni, F. Forastiere, I. Tamayo, P. Amiano, M. Dorronsoro, M. Katsoulis, A. Trichopoulou, B. Brunekreef, and G. Hoek. 2014. Effects of long-term exposure to air pollution on natural-cause mortality: An analysis of 22 European cohorts within the multicentre ESCAPE project. Lancet 383(9919):785-795.
Bind, M.A., B. Coull, H. Suh, R. Wright, A. Baccarelli, P. Vokonas, and J. Schwartz. 2014. A novel genetic score approach using instruments to investigate interactions between pathways and environment: Application to air pollution. PLoS One 9(4):e96000.
Blackburn, K., and S.B. Stuard. 2014. A framework to facilitate consistent characterization of read across uncertainty. Regul. Toxicol. Pharmacol. 68(3):353-362.
Calderón-Garcidueñas, L., B. Azzarelli, H. Acuna, R. Garcia, T.M. Gambling, N. Osnaya, S. Monroy, M. del Rosario Tizapantzi, J.L. Carson, A. Villarreal-Calderon, and B. Rewcastle. 2002. Air pollution and brain damage. Toxicol. Pathol. 30(3):373-389.
Calderón-Garciduenas, L., R. Torres-Jardón, R.J. Kulesza, S. Park, and A. D’Angiulli. 2014. Air pollution and detrimental effects on children’s brain. The need for a multidisciplinary approach to the issue complexity and challenges. Front Hum. Neurosci. 8:613.
Chang, S.Y., W. Vizuete, A. Valencia, B. Naess, V. Isakov, T. Palma, M. Breen, and S. Arunachalam. 2015. A modeling framework for characterizing near-road air pollutant concentration at community scales. Sci. Total Environ. 538:905-921.
Chapin, R.E., J. Delaney, Y. Wang, L. Lanning, B. Davis, B. Collins, N. Mintz, and G. Wolfe. 1999. The effects of 4-nonylphenol in rats: A multigeneration reproduction study. Toxicol. Sci. 52(1):80-91.
Chen, J.C.., X. Wang, G.A. Wellenius, M.L. Serre, I. Driscoll, R. Casanova, J.J. McArdle, J.E. Manson, H.C. Chui, and M.A. Espeland. 2015. Ambient air pollution and neurotoxicity on brain structure: Evidence from Women’s Health Initiative Memory Study. Ann. Neurol. 78(3):466-476.
Chiu, Y.H., D.C. Bellinger, B.A. Coull, S. Anderson, R. Barber, R.O. Wright, and R.J. Wright. 2013. Associations between traffic-related black carbon exposure and attention in a prospective birth cohort of urban children. Environ Health Perspect. 121(7):859-864.
Coker, E., J. Ghosh, M. Jerrett, V. Gomez-Rubio, B. Becker-man, M. Cockburn, S. Liverani, J. Su, A. Li, M.L. Kile, B. Ritz, and J. Molitor. 2015. Modeling spatial effects of PM(2.5) on term low birth weight in Los Angeles County. Environ Res. 142:354-364.
Coker, E., S. Liverani, J.K. Ghosh, M. Jerrett, B. Becker-man, A. Li, B. Ritz, and J. Molitor. 2016. Multi-pollutant exposure profiles associated with term low birth weight in Los Angeles County. Environ. Int. 91:1-13.
Costa, L.G., T.B. Cole, J. Coburn, Y.C. Chang, K. Dao, and P. Roque. 2014. Neurotoxicants are in the air: Convergence of human, animal, and in vitro studies on the effects of air pollution on the brain. Biomed. Res. Int. 2014:736385.
Crofton, K., E. Fritsche, T. Ylikomi, and A. Bal-Price. 2014. International Stakeholder NETwork (ISTNET) for creating a developmental neurotoxicity testing (DNT) roadmap for regulatory process. ALTEX 31(2):223-224.
Demetriou, C.A., and P. Vineis. 2015. Carcinogenicity of ambient air pollution: Use of biomarkers, lessons learnt and future directions. J. Thorac. Dis. 7(1):67-95.
Demetriou, C.A., O. Raaschou-Nielsen, S. Loft, P. Møller, R. Vermeulen, D. Palli, M. Chadeau-Hyam, W.W. Xun, and P. Vineis. 2012. Biomarkers of ambient air pollution and lung cancer: A systematic review. Occup. Environ. Med. 69(9):619-627.
Dockery, D.W., C.A. Pope, III, X. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, B.G. Ferris, Jr., and F.E. Speizer. 1993. An association between air pollution and mortality in six US cities. N. Engl. J. Med. 329(24):1753-1759.
Edwards, S.C., W. Jedrychowski, M. Butscher, D. Camann, A. Kieltyka, E. Mroz, E. Flak, Z. Li, S. Wang, V. Rauh, and F. Perera. 2010. Prenatal exposure to airborne polycyclic aromatic hydrocarbons and children’s intelligence at 5 years of age in a prospective cohort study in Poland. Environ. Health Perspect. 118(9):1326-1331.
EPA (US Environmental Protection Agency). 2009. Integrated Science Assessment for Particulate Matter (Final Report). EPA/600/R-08/139F. US Environmental Protection Agency, Washington, DC [online]. Available: https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=216546 [accessed July 25, 2016].
EPA (US Environmental Protection Agency). 2011. Estimation Programs Interface (EPI) Suite for Microsoft® Windows, Version 4.1. U. S. Environmental Protection Agency, Washington, DC.
Fagundes, L.S., A.daS. Fleck, A.C. Zanchi, P.H. Saldiva, and C.R. Rhoden. 2015. Direct contact with particulate matter increases oxidative stress in different brain structures. Inhal. Toxicol. 27(10):462-467.
Fasanelli, F., L. Baglietto, E. Ponzi, F. Guida, G. Campanella, M. Johansson, K. Grankvist, M. Johansson, M.B. Assumma, A. Naccarati, M. Chadeau-Hyam, U. Ala, C. Faltus, R. Kaaks, A. Risch, B. De Stavola, A. Hodge, G.G. Giles, M.C. Southey, C.L. Relton, P.C. Haycock, E. Lund, S. Polidoro, T.M. Sandanger, G. Severi, and P. Vineis. 2015. Hypomethylation of smoking-related genes is associated with future lung cancer in four prospective cohorts. Nat. Commun. 6:10192.
Felter, S.P., G.P. Daston, S.Y. Euling, A.H. Piersma, and M.S. Tassinari. 2015. Assessment of health risks resulting from early-life exposures: Are current chemical toxicity testing protocols and risk assessment methods adequate? Crit. Rev. Toxicol. 45(3):219-244.
Filer, D., H.B. Patisaul, T. Schug, D. Reif, and K. Thayer. 2014. Test driving ToxCast: Endocrine profiling for 1858 chemicals included in phase II. Curr. Opin. Pharmacol. 19:145-152.
Goodale, B.C., J. La Du, S.C. Tilton, C.M. Sullivan, W.H. Bisson, K.M. Waters, and R.L. Tanguay. 2015. Ligand-specific transcriptional mechanisms underlie aryl hydrocarbon receptor-mediated developmental toxicity of oxygenated PAHs. Toxicol. Sci. 147(2):397-411
Grandjean, P. and P.J. Landrigan. 2006. Developmental neurotoxicity of industrial chemicals. Lancet 368(9553):2167-2178.
Guida, F., T.M. Sandanger, R. Castagné, G. Campanella, S. Polidoro, D. Palli, V. Krogh, R. Tumino, C. Sacerdote, S. Panico, G. Severi, S.A. Kyrtopoulos, P. Georgiadis, R.C. Vermeulen, E. Lund, P. Vineis, and M. Chadeau-Hyam. 2015. Dynamics of smoking-induced genome-wide methylation changes with time since smoking. Hum. Mol. Genet. 24(8):2349-2359.
Guxens, M., and J. Sunyer. 2012. A review of epidemiological studies on neuropsychological effects of air pollution. Swiss Med. Wkly. 141:w13322.
Guxens, M., R. Garcia-Esteban, L. Giorgis-Allemand, J. Forns, C. Badaloni, F. Ballester, R. Beelen, G. Cesaroni, L. Chatzi, M. de Agostini, A. de Nazelle, M. Eeftens, M.F. Fernandez, A. Fernández-Somoano, F. Forastiere, U. Gehring, A. Ghassabian, B. Heude, V.W. Jaddoe, C. Klümper, M. Kogevinas, U. Krämer, B. Larroque, A. Lertxundi, N. Lertxuni, M. Murcia, V. Navel, M. Nieuwenhuijsen, D. Porta, R. Ramos, T. Roumeliotaki, R. Slama, M. Sørensen, E.G. Stephanou, D. Sugiri, A. Tardón, H. Tiemeier, C.M. Tiesler, F.C. Verhulst, T. Vrijkotte, M. Wilhelm, B. Brunekreef, G. Pershagen, and J. Sunyer. 2014. Air pollution during pregnancy and childhood cognitive and psychomotor development: Six European birth cohorts. Epidemiology 25(5):636-647.
Hamra, G.B., N. Guha, A. Cohen, F. Laden, O. Raaschou-Nielsen, J. Samet, P. Vineis, F. Forastiere, P. Saldiva, T. Yorifuki, and D. Loomis. 2014. Outdoor particulate matter exposure and lung cancer: A systematic review and meta-analysis. Environ. Health Perspect. 122(9):906-911.
Hawrylycz, M., J. Miller, V. Menon, D. Feng, T. Dolbeare, A.L. Guillozet-Bongaarts, A. G. Jegga, B. J. Aronow, C. Lee, A. Bernard, M.F. Glasser, D.L. Dierker, J. Menche, A. Szafer, F. Collman, P. Grange, K.A. Berman, S. Mihalas, Z. Yao, L. Stewart, A. Barabási, J. Schulkin, J. Phillips, L. Ng, C. Dang, D.R. Haynor, A. Jones, D.C. Van Essen, C. Koch, and E. Lein. 2015. Canonical genetic signatures of the adult human brain. Nat. Neurosci. 18(12):1832-1844.
Hossaini, A., M. Dalgaard, A.M. Vinggaard, P. Pakarinen, and J.J. Larsen. 2003. Male reproductive effects of octylphenol and estradiol in Fischer and Wistar rats. Reprod. Toxicol. 17(5):607-615.
Hudda, N.T. Gould, K. Hartin, T.V. Larson, and S.A. Fruin. 2014. Emissions from an international airport increase
particle number concentrations 4-fold at 10 km downwind. Environ. Sci. Technol. 48(12):6628-6635.
IARC (International Agency for Research on Cancer). 2015. Outdoor Air Pollution. Monographs on the Evaluation of Carcinogenic Risks to Humans Vol. 109. Lyon: IARC [online]. Available: http://monographs.iarc.fr/ENG/Monographs/vol109/index.php [accessed July 25, 2016].
Janssen, B.G., H.M. Byun, W. Gyselaers, W. Lefebvre, A.A. Baccarelli, and T.S. Nawrot. 2015. Placental mitochondrial methylation and exposure to airborne particulate matter in the early life environment: An ENVIRONAGE birth cohort study. Epigenetics 10(6):536-544.
Joubert, B.R., S.E. Håberg, R.M. Nilsen, X. Wang, S.E Vollset, S.K. Murphy, Z. Huang, C. Hoyo, Ø. Midttun, L.A. Cupul-Uicab, P.M. Ueland, M.C. Wu, W. Nystad, D.A. Bell, S.D. Peddada, and S.J. London. 2012. 450K epigenome-wide scan identifies differential DNA methylation in newborns related to maternal smoking during pregnancy. Environ. Health Perspect.120(10):1425-1431.
Kachuri, L., C.I. Amos, J.D. McKay, M. Johansson, P. Vineis, H.B. Bueno-de-Mesquita, M.C. Boutron-Ruault, M. Johansson, J.R. Quirós, S. Sieri, R.C. Travis, E. Weiderpass, L. Le Marchand, B.E. Henderson, L. Wilkens, G. Goodman, C. Chen, J.A. Doherty, D.C. Christiani, Y. Wei, L. Su, S. Tworoger, X. Zhang, P. Kraft, D. Zaridze, J.K. Field, M.W. Marcus, M.P. Davies, R. Hyde, N.E. Caporaso, M.T. Landi, G. Severi, G.G. Giles, G. Liu, J.R. McLaughlin, Y. Li, X. Xiao, G. Fehringer, X. Zong, R.E. Denroche, P.C. Zuzarte, J.D. McPherson, P. Brennan, and R.J. Hung. 2016. Fine-mapping of chromosome 5p15.33 based on a targeted deep sequencing and high density genotyping identifies novel lung cancer susceptibility loci. Carcinogenesis 37(1):96-105.
Kelce, W.R., and L.E. Gray, Jr.1997. Endocrine disruptors: Effects on sex steroid hormone receptors and sex development. Pp. 435-474 in Drug Toxicity in Embryonic Development, Vol. 2, R.J. Kavlock, and G.P. Daston, eds. Berlin: Springer.
Kelly, F.J., and J.C. Fussell. 2015. Linking ambient particulate matter pollution effects with oxidative biology and immune responses. Ann. N.Y. Acad. Sci. 1340:84-94.
Kim, E., H. Park, Y.C. Hong, M. Ha, Y. Kim, B.N. Kim, Y. Kim, Y.M. Roh, B.E. Lee, J.M. Ryu, B.M. Kim, and E.H. Ha. 2014. Prenatal exposure to PM2.5 and NO2 and children’s neurodevelopment from birth to 24 months of age: Mothers and Children’s Environmental Health (MOCEH) study. Sci. Total Environ. 481:439-445.
Knecht, A.L., B.C. Goodale, L. Truong, M.T. Simonich, A.J. Swanson, M.M. Matzke, K.A. Anderson, K.M. Waters, and R.L. Tanguay. 2013. Comparative developmental toxicity of environmentally relevant oxygenated PAHs. Toxicol. Appl. Pharmacol. 271(2):266-275.
Künzli, N., and I.B. Tager. 1997. The semi-individual study in air pollution epidemiology: A valid design as compared to ecologic studies. Environ. Health Perspect. 105(10):1078-1083.
Laws, S.C., S.A. Carey, J.M. Ferrell, G.J. Bodman, and R.L. Cooper. 2000. Estrogenic activity of octylphenol, nonylphenol, bisphenol a and methoxychlor in rats. Toxicol. Sci. 54(1):154-167.
Lin, C.C., S.K. Yang, K.C. Lin, W.C. Ho, W.S. Hsieh, B.C. Shu, and P.C. Chen. 2014. Multilevel analysis of air pollution and early childhood neurobehavioral development. Int. J. Environ. Res. Public Health 11(7):6827-6841.
Liu, F., Y. Huang, F. Zhang, Q. Chen, B. Wu, W. Rui, J.C. Zheng, and W. Ding..2015. Macrophages treated with particulate matter PM2.5 induce selective neurotoxicity through glutaminase-mediated glutamate generation. J. Neurochem. 134(2):315-326.
Loke, H., V. Harley, and J. Lee. 2015. Biological factors underlying sex differences in neurological disorders. Int. J. Biochem. Cell. Biol. 65:139-150.
Lovasi, G.S., N. Eldred-Skemp, J.W. Quinn, H.W. Chang, V.A. Rauh, A. Rundle, M.A. Orjuela, and F.P. Perera. 2014. Neighborhood social context and individual polycyclic aromatic hydrocarbon exposures associated with child cognitive test scores. J. Child Fam. Stud. 23(5):785-799.
McCallister, M.M., Z. Li, T. Zhang, A. Ramesh, R.S. Clark, M. Maguire, B. Hutsell, M.C. Newland, and D.B. Hood. 2016. Revealing behavioral learning deficit phenotypes subsequent to in utero exposure to benzo(a)pyrene. Toxicol Sci. 149(1):42-54.
McPartland, J., H.C. Dantzker, and C.J. Portier. 2015. Building a robust 21st century chemical testing program at the US Environmental Protection Agency: Recommendations for strengthening scientific engagement. Environ. Health Perspect. 123(1):1-5.
Mikkilä, T.F., J. Toppari, and J. Paranko. 2006. Effects of neonatal exposure to 4-tert-octylphenol, diethylstilbestrol, and flutamide on steroidogenesis in infantile rat testis. Toxicol. Sci. 91(2):456-466.
Mostafavi, N., J. Vlaanderen, M. Chadeau-Hyam, R. Beelen, L. Modig, D. Palli, I.A. Bergdahl, P. Vineis, G. Hoek, S.A. Kyrtopoulos, and R. Vermeulen. 2015. Inflammatory markers in relation to long-term air pollution. Environ. Int. 81:1-7.
Nieuwenhuijsen, M.J., D. Donaire-Gonzalez, I. Rivas, M. de Castro, M. Cirach, G. Hoek, E. Seto, M. Jerrett, and J. Sunyer. 2015. Variability in and agreement between modeled and personal continuously measured black carbon levels using novel smartphone and sensor technologies. Environ. Sci. Technol. 49(5):2977-2982.
Nik-Zainal, S., J.E. Kucab, S. Morganella, D. Glodzik, L.B. Alexandrov, V.M. Arlt, A. Weninger, M. Hollstein, M.R. Stratton, and D.H. Phillips. 2015. The genome as a record of environmental exposure. Mutagenesis 30(6):763-770.
Novakovic, B., J. Ryan, N. Pereira, B. Boughton, J.M. Craig, and R. Saffery. 2014. Postnatal stability, tissue, and time
specific effects of AHRR methylation change in response to maternal smoking in pregnancy. Epigenetics 9(3):377-386.
NRC (National Research Council). 1993. Pesticides in Diet of Infants and Children. Washington, DC: National Academy Press.
NRC (National Research Council). 2012. Exposure Science in the 21st Century: A Vision and a Strategy. Washington, DC: The National Academies Press.
Perera, F.P., D. Tang, Y.H. Tu, L.A. Cruz, M. Borjas, T. Bernert, and R.M. Whyatt. 2004. Biomarkers in maternal and newborn blood indicate heightened fetal susceptibility to procarcinogenic DNA damage. Environ. Health Perspect. 112(10):1133-1136.
Perera, F.P., V. Rauh, R.M. Whyatt, W.Y. Tsai, D. Tang, D. Diaz, L. Hoepner, D. Barr, Y.H. Tu, D. Camann, and P. Kinney. 2006. Effects of prenatal exposure to airborne polycyclic aromatic hydrocarbons on neurodevelopment in the first 3 years of life among inner-city children. Environ. Health Perspect. 114(8):1287-1292.
Perera, F.P., Z. Li, R. Whyatt, L. Hoepner, S. Wang, D. Camann, and V. Rauh. 2009. Prenatal airborne polycyclic aromatic hydrocarbon exposure and child IQ at age 5 years. Pediatrics 124(2):e195-e202.
Proietti, E., M. Röösli, U. Frey, and P. Latzin. 2013. Air pollution during pregnancy and neonatal outcome: A review. J. Aerosol. Med. Pulm. Drug Deliv. 26(1):9-23.
Raaschou-Nielsen, O., Z.J. Andersen, R. Beelen, E. Samoli, M. Stafoggia, G. Weinmayr, B. Hoffman, P. Fischer, M.J. Nieuwenhuijsen, B. Brunekreef, W.W. Xun, K. Katsouyanni, L. Dimakopoulou, J. Sommar, B. Forsberg, L. Modig, A. Oudin, B. Oftedal, P.E. Schwarze, P. Nafstad, U. De Faire, N.L. Pedersen, C.G. Ostenson, L. Fratiglioni, J. Penell, M. Korek, G. Pershagen, K.T. Eriksen, M. Sørensen, A. Tjønneland, T. Ellerman, M. Eeftens, P.H. Peeters, K. Meliefste, M. Wang, B. Bueno-de-Mesquita, T.J. Key, K. de Hoogh, H. Concin, G. Nagel, A. Vilier, S. Grioni, V. Krogh, M.Y. Tsai, F. Ricceri, C. Sacerdote, C. Galassi, E. Migliore, A. Ranzi, G. Cesaroni, C. Badaloni, F. Forastiere, I. Tamayo, P. Amiano, M. Dorronsoro, A. Trichopoulou, C. Bamia, P. Vineis, and G. Hoek. 2013. Air pollution and lung cancer incidence in 17 European cohorts: Prospective analyses from the European Study of Cohorts for Air Pollution Effects (ESCAPE). Lancet On-col. 14(9):813-822.
Raaschou-Nielsen, O., R. Beelen, M. Wang, G. Hoek, Z.J. Andersen, B. Hoffmann, M. Stafoggia, E. Samoli, G. Weinmayr, K. Dimakopoulou, M. Nieuwenhuijsen, W.W. Xun, P. Fischer, K.T. Eriksen, M. Sørensen, A. Tjønneland, F. Ricceri, K. de Hoogh, T. Key, M. Eeftens, P.H. Peeters, H.B. Bueno-de-Mesquita, K. Meliefste, B. Oftedal, P.E. Schwarze, P. Nafstad, C. Galassi, E. Migliore, A. Ranzi, G. Cesaroni, C. Badaloni, F. Forastiere, J. Penell, U. De Faire, M. Korek, N. Pedersen, C.G. Östenson, G. Pershagen, L. Fratiglioni, H. Concin, G. Nagel, A. Jaensch, A. Ineichen, A. Naccarati, M. Katsoulis, A. Trichpoulou, M. Keuken, A. Jedynska, I.M. Kooter, J. Kukkonen, B. Brunekreef, R.S. Sokhi, K. Katsouyanni, and P. Vineis. 2016. Particulate matter air pollution components and risk for lung cancer. Environ. Int. 87:66-73.
Reif, D.M., L. Truong, D. Mandrell, S. Marvel, G. Zhang, and R.L. Tanguay. 2016. High-throughput characterization of chemical-associated embryonic behavioral changes predicts teratogenic outcomes. Arch. Toxicol. 90(6):1459-1470.
Reinius, B., and E. Jazin. 2009. Prenatal sex differences in the human brain. Mol. Psychiatry 14(11):987-989.
Saenen, N.D., M. Plusquin, E. Bijnens, B.G. Janssen, W. Gyselaers, B. Cox, F. Fierens, G. Molenberghs, J. Penders, K. Vrijens, P. De Boever, and T.S. Nawrot. 2015. In utero fine particle air pollution and placental expression of genes in the brain-derived neurotrophic factor signaling pathway: An ENVIRONAGE Birth Cohort Study. Environ Health Perspect. 123(8):834-840.
Schitine, C., L. Nogaroli, M.R. Costa, and C. Hedin-Pereira. 2015. Astrocyte heterogeneity in the brain: From development to disease. Front. Cell. Neurosci. 9:76.
Schwarz, J.M., and S.D. Bilbo. 2012. Sex, glia, and development: Interactions in health and disease. Horm. Behav. 62(3):243-253.
Shenker, N.S., S. Polidoro, K. van Veldhoven, C. Sacerdote, F. Ricceri, M.A. Birrell, M.G. Belvisi, R. Brown, P. Vineis, and J.M. Flanagan. 2013. Epigenome-wide association study in the European Prospective Investigation into Cancer and Nutrition (EPIC-Turin) identifies novel genetic loci associated with smoking. Hum. Mol. Genet. 22 (5):843-851.
Smarr, M.M., F. Vadillo-Ortega, M. Castillo-Castrejon, and M.S. O’Neill. 2013. The use of ultrasound measurements in environmental epidemiological studies of air pollution and fetal growth. Curr. Opin. Pediatr. 25(2):240-246.
Smith, M.T., K.Z. Guyton, C.F. Gibbons, J.M. Fritz, C.J. Portier, I. Rusyn, D.M. DeMarini, J.C. Caldwell, R.J. Kavlock, P. Lambert, S.S. Hecht, J.R. Bucher, B.W. Stewart, R. Baan, V.J. Cogliano, and K. Straif. 2016. Key characteristics of carcinogens as a basis for organizing data on mechanisms of carcinogenesis. Environ. Health Perspect. 124(6):713-721.
Stieb, D.M., L. Chen, M. Eshoul, and S. Judek. 2012. Ambient air pollution, birth weight and preterm birth: A systematic review and meta-analysis. Environ. Res. 117:100-111.
Suades-González, E., M. Gascon, M. Guxens, and J. Sunyer. 2015. Air pollution and neuropsychological development: A review of the latest evidence. Endocrinology 156(10):3473-3482.
Suglia, S.F., A. Gryparis, R.O. Wright, J. Schwartz, and R.J. Wright. 2008. Association of black carbon with cognition among children in a prospective birth cohort study. Am. J. Epidemiol. 167(3):280-286.
Takeda, K., N. Tsukue, and S. Yoshida. 2004. Endocrine-disrupting activity of chemicals in diesel exhaust and diesel exhaust particles. Environ. Sci. 11(1):33-45.
Tang, D., T. Li, J.J. Liu. Z. Zhou, T. Yuan, Y. Chen, V.A. Rauh, J. Xie, and F. Perera. 2008. Effects of prenatal exposure to coal-burning pollutants on children’s development in China. Environ. Health Perspect. 116(5):674-679.
Tang, D., T.Y. Li, J.C. Chow, S.U. Kulkarni, J.G. Watson, S.S. Ho, Z.Y. Quan, L.R. Qu, and F. Perera. 2014. Air pollution effects on fetal and child development: A cohort comparison in China. Environ. Pollut. 185:90-96.
Truong, L., D. Reif, L. St. Mary, M. Geier, H.D. Truong, and R.L. Tanguay. 2014. Multidimensional in vivo hazard assessment using zebrafish. Toxicol. Sci. 137(1):212-233.
Tyl, R.W., C.B. Myers, M.C. Marr, D.R. Brine, P.A. Fail, J.C. Seely, and J.P. Van Miller. 1999. Two-generation reproduction study with para-tert-octylphenol in rats. Regul. Toxicol. Pharmacol. 30(2 Pt 1):81-95.
van den Hooven, E.H., F.H. Pierik, Y. de Kluizenaar, S.P. Willemsen, A. Hofman, S.W. van Ratingen, P.Y. Zandveld, J.P. Mackenbach, E.A. Steegers, H.M. Miedema, and V.W. Jaddoe. 2012. Air pollution exposure during pregnancy, ultrasound measures of fetal growth, and adverse birth outcomes: A prospective cohort study. Environ Health Perspect. 120(1):150-156.
Vineis, P., A. Schatzkin, and J.D. Potter. 2010. Models of carcinogenesis: An overview. Carcinogenesis 31(10):1703-1709.
Vineis, P., K. van Veldhoven, M. Chadeau-Hyam, and T.J. Athersuch. 2013. Advancing the application of omicsbased biomarkers in environmental epidemiology. Environ. Mol. Mutagen. 54(7):461-467.
Vrijheid, M., D. Martinez, S. Manzanares, P. Dadvand, A. Schembari, J. Rankin, and M. Nieuwenhuijsen. 2011. Ambient air pollution and risk of congenital anomalies: A systematic review and meta-analysis. Environ. Health Perspect. 119(5):598-606.
Wang, W., N. Jariyasopit, J. Schrlau, Y. Jia, S. Tao, T.W. Yu, R.H. Dashwood, W. Zhang, X. Wang, and S.L. Simonich. 2011. Concentration and photochemistry of PAHs, NPAHs, and OPAHs and toxicity of PM2.5 during the Beijing Olympic Games. Environ. Sci. Technol. 45(16):6887-6895.
WHO (World Health Organization). 1986. Principles for Evaluating Health Risks from Chemicals during Infancy and Early Childhood: The Need for a Special Approach. Environmental Health Criteria 59. Geneva: World Health Organization.
Wittkopp, S., N. Staimer, T. Tjoa, T. Stinchcombe, N. Daher, J.J. Schauer, M.M. Shafer, C. Sioutas, D.L. Gillen, and R.J. Delfino. 2016. Nrf2-related gene expression and exposure to traffic-related air pollution in elderly subjects with cardiovascular disease: An exploratory panel study. J. Expo. Sci. Environ. Epidemiol. 16(2):141-149.
Wu, S., K. Blackburn, J. Amburgey, J. Jaworska, and T. Federle. 2010. A framework for using structural, reactivity, metabolic and physicochemical similarity to evaluate the suitability of analogs for SAR-based toxicological assessments. Regul. Toxicol. Pharmacol. 56(1):67-81.
Wu, S., J. Fisher, J. Naciff, M. Laufersweiler, C. Lester, G. Daston, and K. Blackburn. 2013. Framework for identifying chemicals with structural features associated with the potential to act as developmental or reproductive toxicants. Chem. Res. Toxicol. 26(12):1840-1861.
Yang, C.F., and N.M. Shah. 2014. Representing sex in the brain, one module at a time. Neuron. 82(2):261-278.
Yang, A., M. Wang, M. Eeftens, R. Beelen, E. Dons, D.L. Leseman, B. Brunekreef, F.R. Cassee, N.A. Janssen, and G. Hoek. 2015. Spatial variation and land use regression modeling of the oxidative potential of fine particles. Environ. Health Perspect. 123(11):1187-1192.
Yorifuji, T., S. Kashima, M. Higa Diez, Y. Kado, S. Sanada, and H. Doi. 2016. Prenatal exposure to traffic-related air pollution and child behavioral development milestone delays in Japan. Epidemiology 27(1):57-65.