A Data-Driven Approach to Estimating Occupational Inhalation Exposure Using Workplace Compliance Data.

A growing list of chemicals are approved for production and use in the United States and elsewhere, and new approaches are needed to rapidly assess the potential exposure and health hazard posed by these substances. Here, we present a high-throughput, data-driven approach that will aid in estimating occupational exposure using a database of over 1.5 million observations of chemical concentrations in U.S. workplace air samples. We fit a Bayesian hierarchical model that uses industry type and the physicochemical properties of a substance to predict the distribution of workplace air concentrations. This model substantially outperforms a null model when predicting whether a substance will be detected in an air sample, and if so at what concentration, with 75.9% classification accuracy and a root-mean-square error (RMSE) of 1.00 log10 mg m-3 when applied to a held-out test set of substances. This modeling framework can be used to predict air concentration distributions for new substances, which we demonstrate by making predictions for 5587 new substance-by-workplace-type pairs reported in the US EPA's Toxic Substances Control Act (TSCA) Chemical Data Reporting (CDR) industrial use database. It also allows for improved consideration of occupational exposure within the context of high-throughput, risk-based chemical prioritization efforts.

Systematic Evidence Mapping of Potential Exposure Pathways for Per- and Polyfluoroalkyl Substances Based on Measured Occurrence in Multiple Media.

Given that human biomonitoring surveys show per- and polyfluoroalkyl substances (PFAS) to be ubiquitous, humans can be exposed to PFAS through various sources, including drinking water, food, and indoor environmental media. Data on the nature and level of PFAS in residential environments are required to identify important pathways for human exposure. This work investigated important pathways of exposure to PFAS by reviewing, curating, and mapping evidence for the measured occurrence of PFAS in exposure media. Real-world occurrence for 20 PFAS was targeted primarily in media commonly related to human exposure (outdoor and indoor air, indoor dust, drinking water, food, food packaging, articles, and products, and soil). A systematic-mapping process was implemented to conduct title-abstract and full-text screening and to extract PECO-relevant primary data into comprehensive evidence databases. Parameters of interest included the following: sampling dates and locations, numbers of collection sites and participants, detection frequencies, and occurrence statistics. Detailed data were extracted on PFAS occurrence in indoor and environmental media from 229 references and on PFAS occurrence in human matrices where available from those references. Studies of PFAS occurrence became numerous after 2005. Studies were most abundant for PFOA (80% of the references) and PFOS (77%). Many studies analyzed additional PFAS, particularly, PFNA and PFHxS (60% of references each). Food (38%) and drinking water (23%) were the commonly studied media. Most studies found detectable levels of PFAS, and detectable levels were reported in a majority of states in the United States. Half or more of the limited studies for indoor air and products detected PFAS in 50% or more of the collected samples. The resulting databases can inform problem formulation for systematic reviews to address specific PFAS exposure queries and questions, support prioritization of PFAS sampling, and inform PFAS exposure measurement studies. The search strategy should be extended and implemented to support living evidence review in this rapidly advancing area.

Integrative exposomic, transcriptomic, epigenomic analyses of human placental samples links understudied chemicals to preeclampsia.

BACKGROUND: Environmental health research has recently undergone a dramatic shift, with ongoing technological advancements allowing for broader coverage of exposure and molecular biology signatures. Approaches to integrate such measures are still needed to increase understanding between systems-level exposure and biology. OBJECTIVES: We address this gap by evaluating placental tissues to identify novel chemical-biological interactions associated with preeclampsia. This study tests the hypothesis that understudied chemicals are present in the human placenta and associated with preeclampsia-relevant disruptions, including overall case status (preeclamptic vs. normotensive patients) and underlying transcriptomic/epigenomic signatures. METHODS: A non-targeted analysis based on high-resolution mass spectrometry was used to analyze placental tissues from a cohort of 35 patients with preeclampsia (n = 18) and normotensive (n = 17) pregnancies. Molecular feature data were prioritized for confirmation based on association with preeclampsia case status and confidence of chemical identification. All molecular features were evaluated for relationships to mRNA, microRNA, and CpG methylation (i.e., multi-omic) signature alterations involved in preeclampsia. RESULTS: A total of 183 molecular features were identified with significantly differentiated abundance in placental extracts of preeclamptic patients; these features clustered into distinct chemical groupings using unsupervised methods. Of these features, 53 were identified (mapping to 40 distinct chemicals) using chemical standards, fragmentation spectra, and chemical metadata. In general, human metabolites had the largest feature intensities and strongest associations with preeclampsia-relevant multi-omic changes. Exogenous drugs were second most abundant and had fewer associations with multi-omic changes. Other exogenous chemicals (non-drugs) were least abundant and had the fewest associations with multi-omic changes. CONCLUSIONS: These global data trends suggest that human metabolites are heavily intertwined with biological processes involved in preeclampsia etiology, while exogenous chemicals may still impact select transcriptomic/epigenomic processes. This study serves as a demonstration of merging systems exposures with systems biology to better understand chemical-disease relationships.

Uncertainty estimation strategies for quantitative non-targeted analysis.

Non-targeted analysis (NTA) methods are widely used for chemical discovery but seldom employed for quantitation due to a lack of robust methods to estimate chemical concentrations with confidence limits. Herein, we present and evaluate new statistical methods for quantitative NTA (qNTA) using high-resolution mass spectrometry (HRMS) data from EPA's Non-Targeted Analysis Collaborative Trial (ENTACT). Experimental intensities of ENTACT analytes were observed at multiple concentrations using a semi-automated NTA workflow. Chemical concentrations and corresponding confidence limits were first estimated using traditional calibration curves. Two qNTA estimation methods were then implemented using experimental response factor (RF) data (where RF = intensity/concentration). The bounded response factor method used a non-parametric bootstrap procedure to estimate select quantiles of training set RF distributions. Quantile estimates then were applied to test set HRMS intensities to inversely estimate concentrations with confidence limits. The ionization efficiency estimation method restricted the distribution of likely RFs for each analyte using ionization efficiency predictions. Given the intended future use for chemical risk characterization, predicted upper confidence limits (protective values) were compared to known chemical concentrations. Using traditional calibration curves, 95% of upper confidence limits were within ~tenfold of the true concentrations. The error increased to ~60-fold (ESI+) and ~120-fold (ESI-) for the ionization efficiency estimation method and to ~150-fold (ESI+) and ~130-fold (ESI-) for the bounded response factor method. This work demonstrates successful implementation of confidence limit estimation strategies to support qNTA studies and marks a crucial step towards translating NTA data in a risk-based context.

Human exposure pathways to poly- and perfluoroalkyl substances (PFAS) from indoor media: A systematic review.

BACKGROUND: Human exposure to per- and polyfluoroalkyl substances (PFAS) has been primarily attributed to contaminated food and drinking water. There is information indicating other sources and pathways of exposure in residential environments, but few studies report relationships between these indoor media and human biomonitoring measurements. METHODS: This study adapts existing systematic review tools and methodologies to synthesize evidence for PFAS exposure pathways from indoor environment media including consumer products, household articles, cleaning products, personal care products, and indoor air and dust. Studies were identified using innovative machine learning approaches and pathway-specific search strings to reduce time needed for literature search and screening. The included studies and systematic review were evaluated using tools modified specifically for exposure studies. The systematic review was conducted following a previously published protocol (DeLuca et al., 2021) that describes the systematic review methodology used in detail. RESULTS: Only 7 studies were identified that measured the targeted subset of 8 PFAS chemicals in concordant household media (primarily house dust) and participant serum. Data extracted from the included studies were used to calculate exposure intake rates and estimate a percentage of occupant serum concentrations that could be attributed to the indoor exposure pathways. These calculations showed that exposure to PFOA, PFOS, PFNA, and PFHxS from contaminated house dust could account for 13%, 3%, 7%, and 25% of serum concentrations, respectively. Inhalation of PFAS in indoor air could account for less than 4% of serum PFOA concentrations and less than 2% of serum PFOS and PFNA concentrations. A risk of bias was identified due to participant profiles in most of the studies being skewed towards white, female, and higher socioeconomic status. CONCLUSIONS: Along with synthesizing evidence for estimated contributions to serum PFAS levels from indoor exposure media, this systematic review also identifies a consistent risk of bias across exposure study populations that should be considered in future studies. It highlights a major research gap and need for studies that measure concordant data from both indoor exposure media and participant serum and the need for continued research on exposure modeling parameters for many PFAS chemicals.

Inferring pesticide toxicity to honey bees from a field-based feeding study using a colony model and Bayesian inference.

Honey bees are crucial pollinators for agricultural crops but are threatened by a multitude of stressors including exposure to pesticides. Linking our understanding of how pesticides affect individual bees to colony-level responses is challenging because colonies show emergent properties based on complex internal processes and interactions among individual bees. Agent-based models that simulate honey bee colony dynamics may be a tool for scaling between individual and colony effects of a pesticide. The U.S. Environmental Protection Agency (USEPA) and U.S. Department of Agriculture (USDA) are developing the VarroaPop + Pesticide model, which simulates the dynamics of honey bee colonies and how they respond to multiple stressors, including weather, Varroa mites, and pesticides. To evaluate this model, we used Approximate Bayesian Computation to fit field data from an empirical study where honey bee colonies were fed the insecticide clothianidin. This allowed us to reproduce colony feeding study data by simulating colony demography and mortality from ingestion of contaminated food. We found that VarroaPop + Pesticide was able to fit general trends in colony population size and structure and reproduce colony declines from increasing clothianidin exposure. The model underestimated adverse effects at low exposure (36 µg/kg), however, and overestimated recovery at the highest exposure level (140 µg/kg), for the adult and pupa endpoints, suggesting that mechanisms besides oral toxicity-induced mortality may have played a role in colony declines. The VarroaPop + Pesticide model estimates an adult oral LD50 of 18.9 ng/bee (95% CI 10.1-32.6) based on the simulated feeding study data, which falls just above the 95% confidence intervals of values observed in laboratory toxicology studies on individual bees. Overall, our results demonstrate a novel method for analyzing colony-level data on pesticide effects on bees and making inferences on pesticide toxicity to individual bees.

Honey Bees and Neonicotinoid-Treated Corn Seed: Contamination, Exposure, and Effects.

Most corn (Zea mays) seeds planted in the United States in recent years are coated with a seed treatment containing neonicotinoid insecticides. Abrasion of the seed coating generates insecticide-laden planter dust that disperses through the landscape during corn planting and has resulted in many "bee-kill" incidents in North America and Europe. We investigated the linkage between corn planting and honey bee colony success in a region dominated by corn agriculture. Over 3 yr we consistently observed an increased presence of corn seed treatment insecticides in bee-collected pollen and elevated worker bee mortality during corn planting. Residues of seed treatment neonicotinoids, clothianidin and thiamethoxam, detected in pollen positively correlated with cornfield area surrounding the apiaries. Elevated worker mortality was also observed in experimental colonies fed field-collected pollen containing known concentrations of corn seed treatment insecticides. We monitored colony growth throughout the subsequent year in 2015 and found that colonies exposed to higher insecticide concentrations exhibited slower population growth during the month of corn planting but demonstrated more rapid growth in the month following, though this difference may be related to forage availability. Exposure to seed treatment neonicotinoids during corn planting has clear short-term detrimental effects on honey bee colonies and may affect the viability of beekeeping operations that are dependent on maximizing colony size in the springtime. Environ Toxicol Chem 2021;40:1212-1221. © 2020 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

A sensitivity analysis of pesticide concentrations in California Central Valley vernal pools.

Vernal pools are ephemeral wetlands that provide critical habitat to many listed species. Pesticide fate in vernal pools is poorly understood because of uncertainties in the amount of pesticide entering these ecosystems and their bioavailability throughout cycles of wet and dry periods. The Pesticide Water Calculator (PWC), a model used for the regulation of pesticides in the US, was used to predict surface water and sediment pore water pesticide concentrations in vernal pool habitats. The PWC model (version 1.59) was implemented with deterministic and probabilistic approaches and parameterized for three agricultural vernal pool watersheds located in the San Joaquin River basin in the Central Valley of California. Exposure concentrations for chlorpyrifos, diazinon and malathion were simulated. The deterministic approach used default values and professional judgment to calculate point values of estimated concentrations. In the probabilistic approach, Monte Carlo (MC) simulations were conducted across the full input parameter space with a sensitivity analysis that quantified the parameter contribution to model prediction uncertainty. Partial correlation coefficients were used as the primary sensitivity metric for analyzing model outputs. Conditioned daily sensitivity analysis indicates curve number (CN) and the universal soil loss equation (USLE) parameters as the most important environmental parameters. Therefore, exposure estimation can be improved efficiently by focusing parameterization efforts on these driving processes, and agricultural pesticide inputs in these critical habitats can be reduced by best management practices focused on runoff and sediment reductions.

Drought sensitivity of an N2 -fixing tree may slow temperate deciduous forest recovery from disturbance.

Increased drought intensity and frequency due to climate change may reduce the abundance and activity of nitrogen (N2 )-fixing plants, which supply new N to terrestrial ecosystems. As a result, drought may indirectly reduce ecosystem productivity through its effect on the N cycle. Here, we manipulated growing season net rainfall across a series of plots in an early successional mesic deciduous forest to understand how drought affects the aboveground productivity of the N2 -fixing tree Robinia pseudoacacia and three co-occurring nonfixing tree species. We found that lower soil moisture was associated with reduced productivity of R. pseudoacacia but not of nonfixing trees. As a result, the relative biomass and density of R. pseudoacacia declined in drier soils over time. Greater aboveground biomass of R. pseudoacacia was also associated with greater total soil N, extractable inorganic N, N mineralization rates, and productivity of nonfixing trees. These soil N effects may reflect current N2 fixation by R. pseudoacacia saplings, or the legacy effect of former trees in the same location. Our results suggest that R. pseudoacacia promotes the growth of nonfixing trees in early succession through its effect on the N cycle. However, the sensitivity of R. pseudoacacia to dry soils may reduce N2 fixation under scenarios of increasing drought intensity and frequency, demonstrating a mechanism by which drought may indirectly diminish potential forest productivity and recovery rate from disturbance.

A unified approach for protecting listed species and ecosystem services in isolated wetlands using community-level protection goals.

The protection of listed species through the Ecological Risk Assessment (ERA) process is encumbered by the number and diversity of species that need protection and the limited data available to inform assessments. Ecological communities within isolated ecosystems often contain a number of biologically diverse endemic, endangered, and threatened species, as well as providing numerous ecosystem services (ES). We propose an approach that develops community-level protection goals using isolated wetlands that includes both listed species and Service Providing Units (SPUs) that drive ES for ecological risk assessments (ERAs). Community-level protection goals are achieved by developing a protection community and weighing lines of evidence to determine a set of focal species within that community upon which to base the assessment. Lines of evidence include chemical mechanism of action, likely routes of exposure, and taxa susceptibility, as well as relationships among species, and other ecological factors. We demonstrate the process using case studies of chlorpyrifos in California vernal pools and coal ash effluent in Carolina bays. In the California vernal pool case study, listed species were the primary SPUs for the ES provided by the critical habitat. The weight of evidence demonstrated the honey bee as the focal species for the terrestrial environment and the vernal pool fairy shrimp as the focal species for the aquatic environment. The protection community within the Carolina bay case study was more taxonomically diverse than vernal pools for both listed species and SPUs, with amphibians identified as the focal species for which to target mitigation goals and hazard levels. The approach presented here will reduce the time and resource investment required for assessment of risk to listed species and adds an ES perspective to demonstrate value of assessments beyond listed species concerns.

Tolerance or avoidance: drought frequency determines the response of an N2 -fixing tree.

Climate change is increasing drought frequency, which may affect symbiotic N2 fixation (SNF), a process that facilitates ecosystem recovery from disturbance. Here, we assessed the effect of drought frequency on the ecophysiology and SNF rate of a common N2 -fixing tree in eastern US forests. We grew Robinia pseudoacacia seedlings under the same mean soil moisture, but with different drought frequency caused by wet-dry cycles of varying periodicity. We found no effect of drought frequency on final biomass or mean SNF rate. However, seedlings responded differently to wet and dry phases depending on drought frequency. Under low-frequency droughts, plants fixed carbon (C) and nitrogen (N) at similar rates during wet and dry phases. Conversely, under high-frequency droughts, plants fixed C and N at low rates during dry phases and at high rates during wet phases. Our findings suggest that R. pseudoacacia growth is resistant to increased drought frequency because it employs two strategies - drought tolerance or drought avoidance, followed by compensation. SNF may play a role in both by supplying N to leaf tissues for acclimation and by facilitating compensatory growth following drought. Our findings point to SNF as a mechanism for plants and ecosystems to cope with drought.