Tracing Environmental Exposure from Neurodevelopment to Neurodegeneration

Interplay between genetic and environmental factors during critical time windows can have effects that span from neurodevelopment to neurodegeneration. We present the concept of the ‘neuroexposome’, emphasizing the brain’s distinctive response to environmental exposure, and how current ‘omics’ sciences can inform on both disease pathogenesis and future public health policies. Interplay between genetic and environmental factors during critical time windows can have effects that span from neurodevelopment to neurodegeneration. We present the concept of the ‘neuroexposome’, emphasizing the brain’s distinctive response to environmental exposure, and how current ‘omics’ sciences can inform on both disease pathogenesis and future public health policies. Environmental toxicants can have profound effects on brain function. Even the smallest exposures can cause significant neurological deficits. These can be sudden and severe or delayed and more gradual, manifesting, for instance, as neuropsychiatric disorders decades after exposure [1Selevan S.G. et al.Identifying critical windows of exposure for children’s health.Environ. Health Perspect. 2000; 108: 451-455Crossref PubMed Scopus (410) Google Scholar]. Assessment of delayed neurological effects of exposure is challenging. First, the lag between exposure and symptom onset limits the usefulness of minimally invasive measures (e.g., blood, urine) that provide little information on past or cumulative exposure. Second, the brain is physiologically compartmentalized and direct assessment of toxicant penetration into the brain itself is limited to invasive measurements. In this Forum, we highlight the need to treat the brain as a unique exposure entity in the context of environmental factors and disease risk. We argue that cross-disciplinary efforts, embedding analytical tools into epidemiological studies, are needed to better understand the long-term risks of environmental exposures for brain health and accelerate public health policy changes where appropriate. The exposome was proposed by Christopher Wild in 2005 as a conceptual integration of burgeoning omic sciences [e.g., metabolomics (see Glossary), transcriptomics, proteomics] with exposure biology to complement the growing use of genomics in epidemiology [2Wild C. Complementing the genome with an “exposome”: the outstanding challenge of environmental exposure measurement in molecular epidemiology.Cancer Epidemiol. Biomarkers Prev. 2005; 14: 1847-1850Crossref PubMed Scopus (1241) Google Scholar]. Contemporary exposomics uses a range of measures beginning with exposure assessment to reconstruct exposure history and identify associations between lifestyle factors and measured biological endpoints. The ultimate goal is to characterize exposure cause and effects across lifespan. The brain is an intriguing system for the study of exposure. It has selective vulnerability to blood–brain barrier-permeable chemicals, heightened sensitivity to low-level chronic exposures (particularly during neurodevelopment), an extended period of maturation, and – unlike most other cells in the body – affected neurons are long-lived. Peripheral biomarkers typically have limited relevance to the compartmentalized brain and central nervous system (CNS) or require extensive preclinical validation, although longitudinal monitoring of blood and urine provides a temporal profile of chemical exposure. As little is known about the effects on neurological function for the majority of environmental chemicals, direct associations between exposure and neurological health outcomes are not easily drawn. This is compounded in situations where acute exposure or chronic low-level intake may not be detected, where the neurotoxic species is unknown, and/or where measurable effects on neurological health emerge years after exposure. Diagnosis of neurological disorders per DSM guidelines [3Freedman R. et al.The initial field trials of DSM-5: new blooms and old thorns.Am. J. Psychiatry. 2013; 170: 1-5Crossref PubMed Scopus (182) Google Scholar] introduces additional complexity; a broad range of phenotypes and the inherent subjectivity of neuropsychiatric assessment need to be compared against absolute exposure levels. Neuroexposomics does not refer to a specific new technology; rather, it encompasses how existing methods used to characterize the exposome are tailored to the study of exposure effects on the brain (Figure 1). In the following, we discuss how neuroexposomic workflows can be applied in the context of three scenarios. The first examines how derailing of early brain development via toxicant exposure may predispose individuals to neurodevelopmental disorders. The second places neuroexposomics in the context of ‘multi-hit’ hypotheses for disease susceptibility. The third scenario examines the concept of ‘sleeper effects’, where exposure might impact the emergence of neurodegenerative disorders much later in life. Vulnerability to environmental toxicants begins at conception and extends through gestation, parturition, infanthood, adolescence, and into adulthood. Although traditionally defined by dynamic processes occurring at the cellular and system level rather than specific exposure risk, critical windows of neurodevelopment can also be delineated by heightened susceptibility to even short-term exposure to chemicals that are otherwise considered harmful at any life stage [1Selevan S.G. et al.Identifying critical windows of exposure for children’s health.Environ. Health Perspect. 2000; 108: 451-455Crossref PubMed Scopus (410) Google Scholar]. The duration of critical windows varies depending on both the chemical species and the neurodevelopmental stages affected, and the consequences of exposure during these periods can be neuropathologies with latency periods extending well beyond childhood. Rapid brain growth occurring in the first 1000 days (conception to roughly second birthday) has drawn the most attention with respect to long-term exposure effects, although appreciation of the influence of biochemical and lifestyle factors on neurodevelopment extending into adolescence is increasing [4Patton G.C. et al.Adolescence and the next generation.Nature. 2018; 554: 458-466Crossref PubMed Scopus (184) Google Scholar]. In many cases it is timing rather than level of exposure that is the largest determinant of delayed-onset effects. Retrospective association of neurodevelopmental outcomes with a specific historical exposure is hampered by the scarcity of suitable biomarkers. One innovative method involves microchemical analysis of deciduous teeth, which begin calcifying in the second trimester and incorporate circulating toxicants in incremental growth lines that can be used as a temporal record of exposure. However, methodological limitations of this approach highlight the risks in drawing parallels between low-level chemical exposure and divergent mechanisms of neurotoxicity and neurodevelopmental disorders with broad psychiatric phenotypes. For instance, in a study of 289 twins, pre- and postnatal lead exposure and zinc and manganese deficiency measured in teeth were associated with a discordant autism spectrum disorder (ASD) diagnosis in seven twin pairs [5Arora M. et al.Fetal and postnatal metal dysregulation in autism.Nat. Commun. 2017; 815493Crossref PubMed Scopus (98) Google Scholar]. Inferences regarding stressors that influence placental transfer of metals and divergent downstream effects of essential metal deficiencies on epigenetic modifications without mechanistic evidence or quantitative exposure levels limit meaningful conclusions. Additionally, control for potential neurodevelopmental comorbidities associated with chronic lead exposure (e.g., Bayley Scales of Infant Development scores) was unclear. This example reiterates that, despite previous demonstrated validation as a marker of prenatal lead exposure, retrospective cohort design and reliance on a peripheral biomarker in a small population with high variance cannot adequately assess ASD risk, and the authors of the abovementioned study appropriately called for replication in larger cohorts. Conducting a comprehensive assessment of possible links between metal dysregulation and ASD etiology would require a prospective cohort integrating tooth biomarkers with direct multi-omics analysis of CNS exposure, which is practical only through invasive sampling of cerebrospinal fluid (CSF). Herein lies the ethical dilemma: the necessity of invasive sampling for biomonitoring of the CNS, especially in a vulnerable population, is difficult to justify when: (i) the adverse effects of the proposed toxicant are divergent or even unknown; (ii) CSF remains an indirect measure of toxicant levels with respect to affected neuroanatomy; and/or (iii) the precision of diagnostic criteria is in itself a contentious issue [6Maenner M.J. et al.Potential impact of DSM-5 criteria on autism spectrum disorder prevalence estimates.JAMA Psychiatry. 2014; 71: 292-300Crossref PubMed Scopus (140) Google Scholar]. Multi-hit hypotheses accounting for a combination of genetic risk, environmental exposure, and lifestyle factors in the etiology of neurological disorders are increasingly used to explain individual susceptibility. For example, chronic cannabis use during adolescence is broadly associated with altered brain development and poor educational performance, and when coupled with certain genetic polymorphisms (a ‘two-hit’ hypothesis) appears to enhance the risk of delayed-onset psychosis and schizophrenia [7Volkow N.D. et al.Effects of cannabis use on human behavior, including cognition, motivation, and psychosis: a review.JAMA Psychiatry. 2016; 73: 292Crossref PubMed Scopus (503) Google Scholar]. A recent commentary, by contrast, questioned two-hit models of schizophrenia, arguing that they are overly simplistic and downplay cumulative effects of biological, social, and environmental influences [8Davis J. et al.A review of vulnerability and risks for schizophrenia: beyond the two hit hypothesis.Neurosci. Biobehav. Rev. 2016; 65: 185-194Crossref PubMed Scopus (183) Google Scholar]. Given that such influences, while complex, are quantifiable and well defined, we propose that targeted longitudinal neuroexposomics frameworks can be designed to assess schizophrenia risk factors. One could envision a combination of the following assessments: (i) peripheral biofluid measurements employing metabolomics (postnatal vitamin D status, maternal nutrition), transcriptomics (endogenous retroviral activation, genetic risk), and proteomics (downstream dysfunctional cytokines) to monitor system-wide effects; (ii) standardized questionnaires and surveys on diet and lifestyle; (iii) neuropsychiatric assessments; and (iv) consensus diagnosis per DSM criteria. The rapid expansion of the medicinal and recreational cannabis industry underscores the urgency of such prospective cohorts and necessitates the identification of individual risk as a harm-minimization strategy. Genetic risk factors are directly implicated in a small percentage of neurodegenerative diseases. Most diseases are considered sporadic, in that molecular mechanisms of disease remain to be identified, and it is plausible that multi-hit scenarios involving genetic risk also apply to neurodegeneration. In contrast to schizophrenia, where the environment and its relationship to age of onset are better understood, the latency between exposure, molecular pathology, symptoms onset, and clinical diagnosis may be of the order of decades. Prospective cohorts are impractical [9Hare D.J. et al.Excessive early-life dietary exposure: a potential source of elevated brain iron and a risk factor for Parkinson’s disease.NPJ Parkinsons Dis. 2017; 3: 1-5Crossref PubMed Scopus (22) Google Scholar] and the few proposed environmental risk factors are the subject of much debate. With that, much can be learned from postmortem studies with appropriate experimental design or through creative repurposing of existing cohorts. As an example, Morris et al. used a targeted neuroexposomic approach as part of a broader longitudinal study of Alzheimer’s disease [10Morris M. et al.Association of seafood consumption, brain mercury level, and APOE ε4 status with brain neuropathology in older adults.J. Am. Med. Soc. 2016; 315: 489-497Crossref PubMed Scopus (87) Google Scholar]. Evidence supports an association between seafood consumption and neuroprotection, although concomitant bioaccumulation of mercury and effects on cognition may offset any benefit. Morris et al. aligned postmortem brain mercury levels with estimated long-chain fatty acid intake from seafood as well as neuropathology and APOE genotyping collected as part of the two-decade parent study. The authors showed that seafood intake increased brain mercury without affecting disease severity. While the APOE ε4 allele carries a higher risk of Alzheimer’s disease, Morris and colleagues found that ε4 carriers consuming a moderate-seafood diet benefited from increased APOE-mediated transport of protective fatty acids across the blood–brain barrier. In a wider context, this illustrates how deleterious effects of a known neurotoxin can be offset by other, concomitant factors, in this case sharing a single origin – seafood consumption. Neurodegenerative disorders differ from conditions like ASD in that neurological phenotypes are relatively well characterized and more conclusively linked to specific neuropathologies. For instance, loss of nigrostriatal dopaminergic neurons and resulting parkinsonian motor impairment is a feature common to all forms of Parkinson’s disease, although there are multiple molecular pathways that contribute to cell death and neurological dysfunction, including specific environmental exposures. Racette et al. also used a targeted approach in a longitudinal cohort study of parkinsonism in welders that commenced in 2006. Applying a validated model of cumulative manganese exposure extracted from work history surveys and in vivo imaging, Racette and colleagues were able to estimate motor dysfunction progression (per the Unified Parkinson’s Disease Rating Motor Scale) as a function of quantitative annual occupational exposure [11Racette B.A. et al.Dose-dependent progression of parkinsonism in manganese-exposed welders.Neurology. 2017; 88: 344-351Crossref PubMed Scopus (76) Google Scholar]. This demonstrates how a well-designed and validated model can be integrated with an epidemiological study without the need for direct sampling at the time of exposure. Further, it is an example of how the critical-window-of-susceptibility concept can be viewed as dependent on environmental and lifestyle factors as opposed to stages of brain development alone. For sporadic forms of Parkinson’s disease where direct causes of neuron loss are likely to involve a number of endogenous and external factors, elucidating potential environmental contributors is more challenging, particularly for those occurring during early life. Although these are still very much hypothetical and based primarily on animal studies, we direct the reader to our previous discussion of the challenges of and possible solutions for the study of postnatal iron overexposure as a risk factor for Parkinson’s disease [9Hare D.J. et al.Excessive early-life dietary exposure: a potential source of elevated brain iron and a risk factor for Parkinson’s disease.NPJ Parkinsons Dis. 2017; 3: 1-5Crossref PubMed Scopus (22) Google Scholar]. By 2030, it is estimated there will be over 100 000 registered compounds in the environment in the USA alone. While the majority are likely to be inert, only a fraction have been properly assessed for acute toxicity, let alone delayed neurological effects. A human-based exposomic approach to address a list of this scale is unrealistic. It is here that preclinical studies using cells and model organisms are crucial for: (i) evaluating potential public health risks; (ii) prioritizing deeper studies into risk factors; and (iii) justifying human monitoring. Between 2006 and 2013, only 12 chemicals were categorically identified as harmful to the human brain via epidemiological studies and even then only within terms of reference as neurodevelopmental impediments [12Grandjean P. Landrigan P.J. Neurobehavioural effects of developmental toxicity.Lancet Neurol. 2014; 13: 330-338Abstract Full Text Full Text PDF PubMed Scopus (1109) Google Scholar]. Several hundred compounds that exhibit developmental neurotoxicity in animals have not been examined in humans. There is, however, hope that compound screening in controlled laboratory settings will accelerate translation as turnkey analytical technologies with ever-improving sensitivity and specificity are taken up by neuroscientists, epidemiologists, and clinicians alike. Although the brain may not be as isolated as dogma previously suggested, we argue that it represents a unique toxin-exposure entity. The mechanisms of toxicity unique to long-lived neurons, in addition to the delayed functional consequences of even low-level exposure, can be considered largely independent from their external effects. The principles of a neuroexposomic workflow and examples we discussed here serve as a guide for future studies of the impact of chemical exposure on brain health, although how they are implemented should fully capitalize on the capabilities of emerging technologies specific to the study of the brain. Successful application in epidemiology is dependent on the same challenges identified by Wild when he introduced the exposome concept [2Wild C. Complementing the genome with an “exposome”: the outstanding challenge of environmental exposure measurement in molecular epidemiology.Cancer Epidemiol. Biomarkers Prev. 2005; 14: 1847-1850Crossref PubMed Scopus (1241) Google Scholar]: a need for greater collaboration across disciplines and a shift in the priorities of funding agencies that recognizes patience as a virtue when assessing the long-term neurological effects of chemical exposures. This work was supported by a joint National Health and Medical Research Council/Australian Research Council Dementia Research Development Fellowship (A.L.H.; 1106911) and a National Health and Medical Research Council Industry Career Development Fellowship (D.J.H.; 1122981). D.J.H. receives research and material support from Agilent Technologies. The authors wish to acknowledge the support provided to The Florey Institute of Neuroscience and Mental Health from the Victorian Government Operational Infrastructure Support Program. assessment of biochemical, physical, and behavioral traits that result from genetic and/or environmental factors. profiling complete sets of epigenetic modifications of the genome (e.g., DNA methylation, phosphorylation) that do not alter the underlying sequence yet affect gene expression. application of various analytical biochemistry techniques to construct the history of exposures to chemicals, diet, and social stressors in totality and spanning conception to death. monitoring of exposure sources and pathways, including nutritive (diet) and non-nutritive (pica) ingestion, inhalation (e.g., air pollution), dermal absorption (e.g., personal-care products), bioavailability, and brain uptake kinetics (typically using model systems) and human biomonitoring of short-to-medium-term exposure using blood, urine, and hair. sequencing of genes, typically to identify specific polymorphisms associated with increased disease risk or altered protein function, including high-throughput sequencing for genome-wide association studies to identify novel risk loci. study of glycans, encompassing all forms of sugars in the cell, including interactions with other biomolecules. measurement of fatty acids, triglycerides, phospholipids, sterols, and other, related organic compounds. characterization of the metabolic pathways of endogenous and exogenous chemicals. application of exposomics with a specific focus on the brain and CNS. study of proteins, including structure, response, regulation, and post-translational modifications. measurement of gene expression at a specific time point used to identify endogenous response to external factors.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.