2025 EHE Research Retreat

Exposure Modeling Selection: Comparison of the extended WMR and NF-FF models using probabilistic modeling simulation studies

Presenter:   Ryan Hines, MS, CIH

Background: A recent set of papers published in JOEH (Hewett and Ganser (2017), Models for Nearly Every Occasion), extended the Well-Mixed Room (WMR) and Near-Field/Far-Field (NF-FF) models to more realistic environmental conditions including general ventilation with return and local exhaust ventilation (LEV). The proper use of these extended models will require the occupational hygienist to estimate several additional parameters, some of which may not be readily available and cumbersome to measure. This work proposes to aid in model selection by suggesting conditions where parameters are such that the various extended models may not produce results that are distinguishable from one another. The model simulation results will be evaluated through the use of experimental trials conducted in a highly controlled exposure chamber equipped to simulate the various conditions.

Objective: This work aims to promote the use of exposure models by simplifying the choice of model, reducing the need to estimate superfluous parameters, and thereby improving the efficiency of the exposure assessment.

Methods: To assess these models under various conditions, a series of probabilistic modeling assessments were run for each of the six conditions for the WMR and NF-FF models: general ventilation (base models), general ventilation with return (GVR), LEV, GVR and LEV, LEV with return (LEVR), and GVR with LEVR. By specifying a range of realistic parameter values and associated uncertainties, the outputs and their confidence intervals were able to determine the environmental conditions under which the models produce significantly different results. Select simulations were then evaluated for their performance through chamber studies in a highly-controlled exposure chamber. In these select situations, a solvent vapor was generated at a controlled rate using a Harvard Apparatus Syringe Pump, and the various ventilation setups were controlled using dedicated fans and ductwork and measured with flow meters and anemometers. The concentration of the solvent generated was monitored using Dräger X-AM 7000 photoionization detectors (PID). The performance of the models were compared to the metrics specified in ASTM D5157 Standard Guide for Statistical Evaluation of Indoor Air Quality Models.

Results/Conclusion: The results of model simulations indicate that across a range of realistic parameter values, including generation rate, general ventilation return percentage, and presence of local exhaust ventilation, the models proposed often have overlapping confidence intervals, indicating that they may not be practically distinguishable from one another. For example, the presence of LEV is almost always worthwhile to include in the model if vented directly outside the room, even in conditions with a flow rate at approximately 10% of the general ventilation and capture efficiencies between 1 to 5%. The inclusion of return ventilation becomes important only at certain combinations of return flow and control efficiency. The overall results will suggest guidance on the selection of models given the presence of ventilation conditions, allowing the occupational hygienist to improve efficiency in the need to estimate the associated parameters in more complex models.

The Impact of the Built Environment and Shift-Work on Depression and Cognitive Impairment among Women in the Sister Study

Oscar Lopez¹, Gurumurthy Ramachandran¹, Dale Sandler² 

¹Department of Environmental Health and Engineering, Johns Hopkins Bloomberg School of Public Health 

²Epidemiology Branch, National Institute of Environmental Health Sciences 

Mental health outcomes such as depression and cognitive impairment are highly prevalent among adults residing in the United States and are among the leading causes of both disability and suicide. The built and workplace environments are likely to contribute to or further exacerbate both depression and cognitive impairment. Despite what is known about the built environment and its impact on mental well-being, there are substantial gaps in the literature. For one, no studies have looked at how aspects of the built environment such as food deserts, alcohol and tobacco environments, access to green space, and sprawl, collectively impact the mental well-being of individuals residing in such neighborhoods. Prior studies looking at the aforementioned indicators have either been cross-sectional in design or have not been performed. Occupational factors such as having a shift-work (SW) schedule have also been shown to impact both mental health and cognitive functioning. However, few studies have assessed how depressive symptoms change over time, and no studies have looked at how cognitive functioning changes over time in response to having a SW schedule. We seek to longitudinally assess how the built environment and SW impact the mental well-being of individuals using data from the Sister Study conducted by the National Institute of 

Environmental Health Sciences. Currently, we have tied the USDA Food Access Research Atlas, the National Neighborhood Data Archive, and the 2010 Sprawl Index to the participants’ census tracts of residence. We intend to tie green space data from the National Land Cover Database. We will then assess how food deserts, alcohol and tobacco swamps, green space, and sprawl impact depression and cognitive impairment utilizing multilevel modeling. Finally, we will assess how SW scheduling impacts depression and cognitive impairment using mixed effect modeling. 

Associations of endocrine disrupting chemicals with vitamin D biomarkers in childhood: The HOME Study 

Katherine M. Marquess¹, Jordan R. Kuiper², Bruce P. Lanphear³, Andrew N. Hoofnagle⁴, Kim M. Cecil^5,6, Aimin Chen⁷, Yingying Xu⁶, Kimberly Yolton⁶, Heidi J. Kalkwarf⁶, Joseph M. Braun⁸, Jessie P. Buckley⁹

¹Department of Environmental Health and Engineering, Johns Hopkins Bloomberg School of Public Health, Baltimore, USA  ²Department of Environmental and Occupational Health, The George Washington University Milken Institute School of Public Health, Washington, D.C., USA ³Faculty of Health Sciences, Simon Fraser University, Vancouver, Canada ⁴Department of Laboratory Medicine, University of Washington, Seattle, USA

⁵Department of Radiology, Cincinnati Children’s Hospital Medical Center, University of Cincinnati College of Medicine, Cincinnati, USA ⁶Department of Pediatrics, Cincinnati Children’s Hospital Medical Center, University of Cincinnati College of Medicine, Cincinnati, USA ⁷Department of Biostatistics, Epidemiology and Informatics, University of Pennsylvania Perelman School of Medicine, Philadelphia, USA ⁸Department of Epidemiology, Brown University, Providence, USA ⁹Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Children are universally exposed to endocrine disrupting chemicals (EDCs) in consumer products. EDCs may disrupt vitamin D metabolism through several mechanisms, including competitive receptor binding, and interfere with various biological processes. As epidemiologic evidence is limited, especially in children, we investigated associations of 25 EDCs measured at ages 8 and 12 years with vitamin D biomarkers at the same ages. We used data from 236 children of mothers enrolled during pregnancy in a prospective cohort from Cincinnati, Ohio (2003-2006). We measured 4 per-/polyfluoroalkyl substances (PFAS), 6 polybrominated diphenyl ethers (PBDEs), and 4 vitamin D biomarkers in serum and metabolites of 4 organophosphate esters (OPEs), 9 phthalates, and 2 environmental phenols (EPs) in urine collected at ages 8 (n=180) and 12 (n=187) years. Based on normality and detection frequency of the vitamin D biomarkers, we used linear/logistic regression g-computation models with generalized estimating equations to estimate covariate-adjusted effect estimates of vitamin D biomarkers per interquartile range (IQR) increase in log2-transformed EDCs for each 

EDC class and the overall mixture. An IQR increase in all EDCs was associated with 6.7 ng/mL (95% CI: 2.7, 

10.6) higher total 25-dihydroxyvitamin D (driven by perfluorooctanoic acid, bis(1,3-dichloro-2-propyl) phosphate 

[BDCPP], and PBDE-47) and 19.2% (95% CI: -9.6%, 57.3%) higher 24,25-dihydroxyvitamin D3 (driven by PBDE-47, BDCPP, and PBDE-28). Associations with other biomarkers were null. Class-based results were similar except the phthalate mixture was associated with lower vitamin D biomarker levels. EDCs may interfere with vitamin D metabolism and have downstream effects on bone and other health outcomes. 

Using machine learning to accelerate air quality data assessment for rapid community access 

Salvatore Milletich¹, Kirsten Koehler¹ 

¹ Department of Environmental Health and Engineering, Johns Hopkins Bloomberg School of Public Health 

Air pollution is a prominent environmental justice (EJ) concern in the United States and around the globe. Regulatory monitors are often sparsely distributed throughout the United States and are often not in EJ communities. A solution to this has been the deployment of low-cost air monitoring networks that can be easily deployed to provide broad spatial coverage and assist with informing the community on pollutant concentrations on a more granular level. However, there are drawbacks, as data from low-cost air monitors often requires manual data assessment to ensure quality control (QC). This is a time-intensive task that requires the training of individuals and many hours to QC the data. Due to this, many low-cost air monitoring systems are not available in real-time and therefore are not fully optimized to prevent inhalation exposures and protect health. To address this problem, we have applied an ensemble machine learning model containing a random forest and convolutional-long short-term memory (CNN-LSTM) neural network to provide a probability of whether data would be classified as acceptable or unacceptable. This model was trained and evaluated sequentially using seven nitrogen dioxide sensors, with an eighth sensor being used for validation. This strategy was taken to create a generalized model for nitrogen dioxide sensors within the network. Within the validation dataset, 95% of the unacceptable data was selected with a negative predictive value of 0.54. Preliminary results, while modest, indicate the potential for this method and underscore the need for further investigation for optimization.  

Evaluating Low-Cost Sensors for Measurement of Total Volatile Organic Compounds (TVOC) in a Laboratory Chamber

Shilpi Misra¹, Sandra Albornoz Martin¹, Salvatore Milletich¹, Kirsten Koehler¹

¹Department of Environmental Health and Engineering, Johns Hopkins Bloomberg School of Public Health, Baltimore, USA 

Background: Volatile Organic Compounds (VOCs) are widespread in indoor air, originating from sources such as building materials, cleaning agents, and personal care products. While some VOCs cause mild health effects, including headaches and respiratory irritation, others are classified as human carcinogens. Advances in low-cost sensor technology have increased the availability of tools for air quality monitoring. However, despite growing concerns about VOC exposure, few exposure assessments have tested the accuracy and precision of these low-cost sensors. Designed to detect a broad spectrum of VOCs, these sensors report a single aggregated value known as Total Volatile Organic Compounds (TVOC). 

Objective: Our research study aimed to evaluate the accuracy and precision of different low-cost TVOC sensors using controlled laboratory chamber experiments across varying environmental conditions to develop calibration curves. 

Methods: We employed five units each from three brands of TVOC monitors, (for a total of 15 low-cost sensors) with a metal-oxide based sensor: BLATN BR-Smart, BRWISSEN BR-A18, and Purple Air Zen Air Quality Monitor, and a reference-grade photoionization detector (RAE Systems MiniRAE 3000+). 

Results/Conclusion: Our results showed moderate correlation (r² = 0.6–0.8) between two low-cost TVOC sensors 

(BLATN BR-SMART and BRWISSEN BR-A18) and the reference photoionization detector, while the Purple Air Zen 

Air Quality Monitor exhibited weak correlation (r² = 0.0–0.1). Insights from this study can support regulatory evaluations of volatile organic compounds. 

Can I get sick from my animals? Characterizing the risk perception of zoonotic diseases at the livestock-human interface 

Eri Togami¹, Meghan F. Davis¹

¹ Department of Environmental Health & Engineering, Johns Hopkins Bloomberg School of Public Health 

Zoonotic diseases are infectious diseases that are transmitted between humans and animals. Zoonotic diseases, including COVID-19, avian influenza, and Ebola virus, impact our health, society, and livelihoods. Risk perception toward zoonotic disease is highly context dependent and poorly characterized. We will conduct a mixed-methods study to characterize the community members’ perception of risk toward zoonotic diseases in Rwanda. This study, embedded in the Community One Health Empowerment in Senegal and Rwanda (COHERS), will take place in Nyamagabe district, southern Rwanda. We expect that the risk perception toward zoonotic diseases will be low among community members and that we will be able to triangulate findings from the survey with key informant interviews. Understanding the risk perception of zoonotic diseases will support health authorities in developing useful risk communication messages (e.g., risk-elevating or riskmitigating messages) for the preparedness for and response to zoonotic diseases. This could influence community members’ health behavior that may hinder or accelerate disease transmission. 

Associations of Particulate Matter (PM_2.5) and Nitrogen Dioxide (NO₂) with Radiologic Sinus Inflammation Identified by Sinus CT Scan   

Tomann, M.M., Hirsch, A.G., Pollak, J.S., Dewalle, J.J., Blake, E., Piepmeier, L., Lehmann, A.E., Casey, J.A., Bandeen-Roche, K., Schwartz, B.S. 

Background. Chronic rhinosinusitis is a disease of the nasal and sinus mucosa with a significant risk of transition to lower airway diseases. While exposure to air pollutants has been consistently associated with the onset and exacerbation of lower respiratory diseases in many studies, investigations of air pollutants and the development, progression, and exacerbation of chronic rhinosinusitis have been less common and findings have been variable. Using electronic health records data from Geisinger in Pennsylvania, we evaluated associations of short and long duration exposures to particulate matter (PM_2.5) and nitrogen dioxide (NO₂) with radiologic sinus inflammation, an objective finding of chronic rhinosinusitis. 

Methods. In a nested case-control study we included individuals aged 18–80 years from 2008–2018, with two encounters in the four years prior to their index date, and residence in a 38-county study region. Cases (n=2382) with radiologic sinus inflammation were identified using a validated text algorithm applied to sinus computed tomography (CT) scan reports. Controls (n=11,910), without prior CRS diagnosis, were frequency-matched on age, sex, and year of encounter. Air pollutant metrics were assigned in a 1-kilometer residential buffer and within pre-specified short and long latency (0 months, 3 months) and duration (1, 2, 3 months; 1, 2, 3 years) windows. Logistic regression with robust standard errors clustered on community was used to estimate odds ratios (OR) and 95% confidence intervals (95% CI) while adjusting for potential confounders. 

Results. Cases and controls had a mean (SD) age of 49.5 (15.3) years, were predominantly non-Hispanic White (96%), and resided across a range of community types defined within census tracts: higher-density urban (n, %: 

1009, 7.06%), lower-density urban (1970, 13.78%), suburban/small-town (4395, 30.75%), and rural (5889, 

48.40%). Particulate matter (PM_2.5) and nitrogen dioxide (NO₂) had expected variation across years (2008 – 2018), seasons (fall, winter, spring, summer), and geographies. We found strong associations of the fourth quartile of NO₂ with increased odds of radiologic sinus inflammation across both short (0 latency, 2 months exposure, 1.61 [1.29 - 2.00]) and long (3 months latency, 2 years exposure, 1.48 [1.21 - 1.80]) latencies and durations. We found null associations of PM_2.5 with radiologic sinus inflammation across all latency and duration windows. 

Conclusions. In this population-based, large scale study, we found that exposure to the highest quartiles of NO₂ was associated with higher odds of radiologic sinus inflammation across both short and long latency and duration windows. Our NO₂ findings are consistent with studies that show air pollutants contribute to both the onset and exacerbation of lower respiratory disease. We also found null associations of PM_2.5 with radiologic sinus inflammation, which conflicts with studies showing that PM_2.5 is associated with eosinophilic aggregation, chronic rhinosinusitis prevalence, and risk of sinus surgery. Air pollutants have been shown to influence the allergenicity and health effects of atmospheric allergens. Our previous work found independent associations of surrogates for aeroallergens (higher greenness, lower precipitation, higher cumulative growing degree days, and greater urbanization) with increased odds of radiologic sinus inflammation. Ongoing analyses will evaluate air pollutants as potential confounders and effect modifiers of the association of these environmental and community variables with radiologic sinus inflammation. Additional analyses will also evaluate ozone (O₃) and coal particulate matter. 

Acute Kidney Injury and Exposure to Pesticides among Chilean Agricultural Workers in the Maule Cohort (MAUCO) Study 

Grant Tore, MPH;¹ Melissa DeSantiago, MPH;¹Wassim Obeid, PhD;² Chirag R. Parikh, MD, PhD;² Margarita Valenzuela, MSN;³ Catterina Ferreccio, MD, MPH;³ Sandra Cortés A., PhD, MS;³ Lesliam Quirós-Alcalá, PhD, MSc.¹   

1. Department of Environmental Health & Engineering, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA. 
2. Division of Nephrology, Department of Medicine, Johns Hopkins Medicine, Baltimore, MD, USA.
3. Departamento de Salud Pública, Escuela de Medicina, Pontificia Universidad Católica de Chile, Santiago, Chile. 

Globally, growing evidence suggests that occupational exposures to pesticides are linked to chronic kidney disease of nontraditional etiology (CKDnt) among agricultural workers, but few studies have investigated subclinical indicators of kidney damage. Here, we examine associations between pesticide exposures and acute kidney injury in a nested cross-sectional study of 43 male agricultural workers from the Maule Cohort (MAUCO) study in Chile.     

Participant demographic and occupational characteristics were captured through questionnaire. Five biomarkers of kidney injury, inflammation, and repair, neutrophil gelatinaseassociated lipocalin (NGAL), kidney injury marker-1 (KIM-1), interleukin-18 (IL-18), monocyte chemoattractant protein-1 (MCP-1), and chitinase-3-like-protein-1 (YKL-40), were quantified. Serum creatinine was used to calculate estimated glomerular filtration rate (eGFR), the standard measure of kidney function, using the 2021 Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation. We measured urinary concentrations of ten pesticide biomarkers for organophosphorus and pyrethroid insecticides and select herbicides, and twenty pesticides in silicone wristbands, which were worn by participants for 7-10 days. We modeled associations of pesticide exposures with kidney injury biomarkers using covariate-adjusted linear regression. 

On average, participants were 53(±6) years of age and reported 24 years of agricultural work experience. Biomarkers of acute kidney injury (AKI) and eGFR were all positively correlated (Spearman correlations: 0.00-0.60). Six pesticide biomarkers and eight pesticides were detected in 40% or greater in participant urine samples and silicone wristbands, respectively. Doubling of urinary concentrations of 4-nitrophenol (PNP, biomarker of parathion) and 3-phenoxybenzoic acid (3-PBA, biomarker of pyrethroid insecticides) were associated with increased levels of IL-18 (2^β: 1.31; 95%CI: 1.09, 1.57) and KIM-1 (2^β: 1.39; 95%CI: 1.05, 1.87), respectively. Similarly, a twofold increase in the 7-day average concentration of trans-Permethrin was associated with an increase in KIM-1 (2^β: 1.18; 95%CI: 1.01, 1.37), while cypermethrin was associated with decreased MCP-1 (2^β: 0.91; 95%CI: 0.85, 0.97) in the urine. Exposure to cyprodinil, a fungicide, was also associated with increased urinary levels of NGAL (2^β: 1.12; 95%CI: 1.01, 1.22). 

Findings provide suggestive evidence of how pesticide exposures contribute to the development of kidney disease through acute injury and will help to inform future epidemiologic investigations of CKDnt among agricultural workers. 

Fungal biodegradation of pharmaceutically active compounds

Background. Biosolids are a waste byproduct of wastewater treatment facilities and have been utilized for decades as land filler and agriculture fertilizer. However, during the treatment process, many organic contaminants sorb onto the surfaces of biosolids and become recalcitrant to degradation. Little is known regarding the fate of these biosolid-associated organic compound (BOCs) when they are spread on the environment. 

Objective. The main objective of this project is to determine if white-rot fungi can facilitate the biodegradation of BOCs en situ. 

Methods. Trametes versicolor fungi was exposed to a multiple pharmaceutically active compounds and growth media was samples six times over 20 days. biosolid slurry and sacrificially tested in triplicate over a five-day experiment. Samples were analyzed with LC-HRMS and the data were collected for nontarget analysis to look for degradation and transformation products.  Results. The fungi T. versicolor was able to degrade seven of the nine compounds tested. Antidepressants desvenlafaxine, trazodone, and vilazodone were fully degraded (>97%). Data analysis from LC-HRMS is still ongoing to identify potential transformation products from the fungal metabolism of the compounds tested. 

Conclusion. T. versicolor fungi can degrade some pharmaceutically active compounds during a twenty-day exposure period which is promising for environmental remediation strategies. 

U.S. emissions of methyl bromide have declined since the Montreal Protocol phaseout, but not to zero 

Dylan C. Gaeta  

The use of methyl bromide (CH₃Br) as a fumigant has been largely phased out under the Montreal Protocol on Substances That Deplete the Ozone Layer. In the U.S., CH₃Br usage was phased out by 2005 for most uses, and further reduced by 2015 to meet the global phase-out deadline. However, there are continuing exemptions for Quarantine and PreShipment (QPS) applications of CH₃Br, as well as annually renewed Critical Use Exemptions (CUEs) for pre-plant soil fumigation and post-harvest crop fumigation. The continued use of CH₃Br is an environmental and human health concern because CH₃Br is a stratospheric ozone-depleting substance (ODS) and is toxic to human respiratory health. To better understand trends and patterns in U.S. emissions of CH₃Br, we use long-term atmospheric measurements of CH₃Br collected by the NOAA Global Monitoring Laboratory’s Global Greenhouse Gas Reference Network (GGGRN) and a geostatistical inverse model (GIM) to estimate monthly emissions of CH₃Br across the contiguous U.S. We link atmospheric measurements to upwind surface emissions in the GIM using an atmospheric transport model (WRF-STILT). Overall, we see a decline in U.S. CH₃Br emissions from 2007 to 2018 that is consistent with the Montreal Protocol phase-out deadlines. We find that the largest emissions of CH₃Br are from California, both before and after the 2015 phase-out deadline. While CH₃Br emissions from California have largely declined since 2005, they have not yet declined to zero, likely due to continued QPS fumigation permitted under the Montreal Protocol. This result may be expected given California’s unique agricultural sector and numerous international shipping ports but highlights current agricultural and environmental sustainability challenges facing the state. 

Non-Target Identification of Organic Contaminants in Legacy and Modern Land-Applied Biosolids

Tracking and Locating Recent CH₄ Emissions from China’s Energy System Using Geostatistical Inverse Model 

Author: Ziting Huang 

Background and methodology: The general trend of anthropogenic methane emissions from China has been debatable since 2010s. Such divergent conclusions can be explained by various inventory data coverage and by the adoption of different estimation methodologies (top-down, bottom-up, or an integration of both). By using geostatistical inverse modeling approach together with TROPOMI atmospheric observations, the study aims to summarize the general trend of anthropogenic CH₄ emissions in China for a five-year period (2019~2023) at a spatial resolution of 0.5x0.625 degree. 

Results: As an example, this poster introduces the results of CH₄ emissions estimates in 2019 and shows the estimate efficiency of the geo-statistical model by comparing with the estimates of existing studies. It lays a methodological foundation of moving forward with future years. The study also explores the spatial and seasonal characteristics of the emissions estimates and further correlates the emission patterns with the energy production events (including coal, oil, and gas) as well as agricultural plantation activities, which are the two major methane emission sources in China. 

Uptake and metabolism of wastewater-derived psychoactive pharmaceuticals in edible crop plants 

The use of reclaimed wastewater for agricultural irrigation exposes fields to anthropogenic organic pollutants that crops subsequently uptake and metabolize, leading to human exposure. Most plant- derived transformation products of xenobiotics remain unknown, resulting in incomplete risk assessments for these uptaken contaminants. Differences in metabolism between species must also be considered, and this was explored by exposing lettuce, tomatoes and carrots to four psychoactive pharmaceuticals frequently found in wastewater: carbamazepine, lamotrigine, amitriptyline and fluoxetine. Crops were grown in semi-hydroponic conditions and exposed to the selected chemicals individually for 2-6 weeks with an internal concentration of 500 µg/L. The leaves and edible portions of each plant were analyzed separately via LC-HRMS. Non-targeted analysis revealed extensive fluoxetinemetabolism; in addition to the well-documented metabolites norfluoxetine and 4- (trifluoromethyl)phenol, multiple conjugated species novel to plants were identified. Oxidized, demethylated and glycosylated metabolites of carbamazepine, amitriptyline and lamotrigine were also detected.

Estimating and Attributing Short-term Basal Processes at Petermann Glacier, Greenland from Satellites

Christian J. Taubenberger (1), Harihar Rajaram (1), Ciaran Harman (1), Denis Felikson (2), Lauren Andrews (2)

1. Johns Hopkins University, Environmental Health and Engineering Department, Baltimore MD;
2. NASA Goddard Space Flight Center, Cryospheric Science Laboratory, Greenbelt MD. 

Understanding the processes that drive subseasonal variations in Greenland Ice Sheet dynamics are critical for interpreting past behavior and narrowing uncertainty in projections of future change. Specifically, one challenge is understanding how changes in the subglacial system evolve and govern basal mass balance. To address this at the ice sheet scale, we established proof-of-concept at Petermann Glacier and have developed a novel approach that compares dynamic ice thickness change derived from ICESat-2 altimetry minus modeled surface mass balance and firn components with strain rates derived from ITS_LIVE ice surface velocities to estimate basal mass balance. We use ICESat-2 ground-track crossovers, which reduce the time interval between successive measurements to the order of weeks, well below the 91-day repeat interval of ICESat-2 tracks; and employ a robust methodology for estimating vertical strain rates at these locations from ITS_LIVE. Statistically significant differences between the two estimates provide an estimate of the magnitude and duration of basal processes such as basal uplift, basal melt, and till dilation. We applied our method to several regions of interest across Petermann Glacier, where our technique indicates ~5 cm per day of basal melt in winter 10 km from the grounding line and ~15 cm per day of basal uplift in summer 40 km from the grounding line. Our results offer the possibility of measuring short term basal processes from satellite observations at any location of interest and in future work will improve our understanding of subseasonal dynamics across the Greenland Ice Sheet. This study acknowledges differing uncertainties during both winter and summer seasons and explores the potential for future work in both limiting and refining uncertainty calculations at Petermann and in other regions of the GrIS. The first approach for this future work is by comparing model data with in-situ observations from automatic weather stations across the GrIS, with initial analysis shown.

Spatial Measurements of Volatile Organic Compounds in Southeastern Louisiana’s Industrial Corridor to Assess Communities’ Exposures and Risks

Amira Yassine, Andrea A. Chiger, Ellis S. Robinson, Benjamin A. Nault, Mina W. Tehrani, Shivang Agarwal, Benjamin Werden, Carolyn Gigot, Sara N. Lupolt, Conner Daube, Anita M. Avery, Manjula Canagaratna, Harald Stark, Scott C. Herndon, Megan S. Claflin, Kirsten Koehler, Ana M. Rule, Thomas Burke, Tara I. Yacovitch, Keeve Nachman, and Peter F. DeCarlo. 

Inhalation of air pollutants released from industrial activities can pose serious health risks. Louisiana’s industrial corridor is an 85-mile stretch along the Mississippi River home to a large number of chemical facilities with high levels of reported hazardous air pollutant emissions. This study reports on measurements of Volatile Organic Compounds (VOCs) that are also classified as hazardous air pollutants. At present, limited ambient measurements exist for VOCs in the study region. This dearth of measurements limits assessments of communities’ VOC exposures to estimated values from the EPA Air Toxics Screening Assessment (AirToxScreen) model. Using these modeled data, which relies on industry self-reporting, typically misrepresents the true burden of air pollutant exposure. Improved spatiotemporal measurements of VOCs are therefore needed. 

In February 2023, the mixing ratios of 56 VOCs were measured in 15 census tracts of the industrial corridor, using real-time, in situ gas chromatography-mass spectrometry and optical spectroscopy instruments, deployed aboard the Aerodyne mobile laboratory. Spatial measurements were used to estimate the various VOC mixing ratios at the census tract level for comparison with the 2020 AirToxScreen modeled data. Additionally, the mean of the measured VOCs mixing ratios were used to calculate non-cancer health risk estimates. 

The estimated mixing ratios for 80% of the measured VOCs compounds were higher (range:1.2227000×) than those from the 2020 AirToxScreen values in all of the sampling census tracts. The cumulative risk assessment for non-cancer health endpoints indicated risk exceeding acceptable levels for neurological, respiratory, systemic, and renal health effects. Near threshold risk was observed for immunological, developmental, ocular, dermal, reproductive, hematological, and hepatic health endpoints. Formaldehyde and acrolein were identified as the main drivers of the risk across the sampling area. 

These findings highlight the need for more comprehensive measurements of VOCs to better understand the health consequences of industrial emissions in fenceline communities.

Advancing Validation Methodologies and Regulatory Acceptance of Microphysiological Systems

¹Yiguang Zhu (Presenting author) and ¹Paul A. Locke, 

¹Johns Hopkins University, Bloomberg School of Public Health and Whiting School of Engineering, Department of Environmental Health and Engineering, Baltimore, MD, USA

Microphysiological systems (MPS), as a category of alternative methods to animal testing, are gaining recognition for their application in drug development, toxicity assessment, and disease modeling. MPS offer an innovative approach by providing human physiologically relevant models, improving efficiency, and enabling higher throughput compared to conventional animal models. However, their integration into the policymaking process remains a work in progress. A comprehensive validation framework is critical to facilitate MPS adoption while ensuring scientific rigor and regulatory relevance. 

We conducted a literature review and analysis of the current landscape of MPS in regulatory contexts, focusing on existing MPS validation methodologies. This review is grounded in an assessment of the framework presented in a recent publication by the Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM), titled "Validation, Qualification, and Regulatory Acceptance of New Approach Methodologies". Notably, the framework does not specify which federal agencies or parties should take responsibility for implementing these validation processes. Building upon this foundation and drawing from additional relevant literature, we identified significant bottlenecks in the framework, particularly the overemphasis on the fit-for-purpose validation strategy. Defining the context-of-use of MPS models in the early validation stage may constrain their broader adoption and application in diverse research and regulatory settings, ultimately limiting their overall impact. 

Our research introduces a two-phase validation framework designed to address these limitations. The framework consists of a Validation Phase, which ensures scientific and engineering robustness, and a Qualification Phase, which aligns validated models with specific regulatory and end-user needs. This dual approach tackles the narrow fit-for-purpose constraints by balancing flexibility with rigor. By providing MPS developers, end-users, and policymakers with a clear and adaptable pathway, our framework facilitates the integration of MPS into biomedical research and regulatory decision-making while maximizing their potential to advance human health. 

Diplomacy Disrupted: A Mixed-Methods Analysis of Russian Disinformation at the Ninth Review Conference of the Biological and Toxin Weapons Convention.

Annie Sundelson^1,2, Gigi Gronvall^1,2, Gary Ackerman³, Rupali Limaye⁴, Crystal Watson^1,2, Tara Kirk Sell1,2

¹Johns Hopkins Bloomberg School of Public Health, Department of Environmental Health and Engineering ²Johns Hopkins Center for Health Security ³College of Emergency Preparedness, Homeland Security, and Cybersecurity at SUNY Albany ⁴Johns Hopkins Bloomberg School of Public Health, Department of International Health   

In 2022, Russia invoked Articles V and VI of the Biological Weapons Convention (BWC), requesting a formal meeting to discuss, and subsequent investigation of, alleged U.S.-funded biological weapons laboratories in Ukraine. Such allegations have been dismissed as false by scholars and diplomats alike, many of whom have argued that Russia’s actions represented an abuse of BWC provisions and risked undermining the Convention. However, few scholars have assessed the implications of Russia’s ongoing efforts to level false allegations in BWC meetings following the Article V and VI procedures. Using mixed-methods analysis of BWC meeting recordings, transcripts, and documents, we assessed the volume, consequences, and framing of Russian false allegations at the BWC Ninth Review Conference. Analysis revealed that discussion of Russian allegations took over three hours and contributed to a stunted Final Document. Additional potential consequences are discussed, including increased division among states parties and the erosion of non-proliferation norms.

Improving National Stockpiles of Personal Protective Equipment for Pandemic Preparedness

Caitlin Walker^1,2, Tara Kirk Sell^1,2 

¹Department of Environmental Health and Engineering, Johns Hopkins Bloomberg School of Public Health ²Center for Health Security, Johns Hopkins University 

Personal protective equipment (PPE) is a critical countermeasure for immediate outbreak response. Despite its importance, PPE supply has been limited in previous pandemics, leading to higher mortality and worse societal and economic outcomes. Stockpiles of PPE are required in order to protect healthcare workers and the broader critical workforce until additional supplies can be obtained. Several nations have PPE stockpiles, however capabilities and policies vary. The diversity in approaches has not been well characterized and no guidelines or best practices exist. This research aimed to review of current national PPE stockpiling practices described in published literature. Seven electronic databases were searched using terms related to PPE and stockpiles, alongside a review of 111 WHO Joint External Evaluation reports and 16 National Action Plans for Health Security. Data regarding PPE stockpiles was extracted and inductively coded. Of 392 records identified by the scoping review, and 11 were eligible for inclusion. The scoping review identified literature primarily focused on high income nations in Asia, Europe and North America, with few in-depth descriptions. JEE reports described several features of PPE stockpiles which differ among nations, including the use of regional and facility-level stockpiles, risk-mapping exercises, inventory monitoring, and practice exercise. Both JEE and NAPHS reports identified a range of challenges and recommended improvements to stockpiles. Further research will include conducting national PPE stockpile case studies through semi-structured interviews with subject matter experts, and a modified Delphi study to generate best practices for PPE stockpiles. Recommendations generated can therefore assist with complex decisions involved in stockpile development. 

Toward Risk Analysis of the Impact of Artificial Intelligence on the Deliberate Biological Threat Landscape

The perception that the convergence of biological engineering and artificial intelligence (AI) could enable increased biorisk has recently drawn attention to the governance of biotechnology and artificial intelligence. The 2023 Executive Order, Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, requires an assessment of how artificial intelligence can increase biorisk. Within this perspective, quantitative and qualitative frameworks for evaluating biorisk are presented. Both frameworks are exercised using notional scenarios and their benefits and limitations are then discussed.  

Chronic Low-Dose Domoic Acid Exposure Influences Spontaneous and Evoked Neuronal Network Dynamics in hiPSC Derived Neural Organoids 

Authors: D. M. Alam El Din¹, A. Lysinger¹, T. Hartung¹, L. Smirnova¹ 

¹Center for Alternatives to Animal Testing, Johns Hopkins University, Baltimore MD 

Background and Purpose: Domoic Acid (DA) is a naturally occurring neurotoxin produced by pseudo-nitzschia that can lead to cellular and structural damage in the human brain. Oral exposure through consumption of contaminated shellfish is becoming increasingly more common, as climate change drives a rise in the number of harmful algae blooms. While acute high-dose DA exposure in humans can cause temporary cognitive impairments, the impact of chronic low-dose DA exposure below the regulatory limit on developing and mature nervous systems is poorly understood. Due to the ethical and technical limitations of using primary human cells, most studies to date have been conducted using animal models, therefore, it has been challenging to understand the relationship between chronic low-dose DA exposure and cognitive impairments in the human brain. Since neuronal networks are central to cognitive processes, our primary goal is to explore how chronic, low-dose DA exposure affects neuronal network development and function in human iPSC-derived neural organoids. 

Methods: Neural organoids were differentiated from healthy donor iPSCs following an in-house protocol, producing electrically active organoids containing excitatory and inhibitory neurons, astrocytes, and myelinating oligodendrocytes. Organoids were then chronically exposed to DA at two concentrations (0.10 µM, 0.32 µM, and 3.16 µM) from weeks 4 to 8 of development. Based on an estimated acute daily tolerable intake of DA of 0.075 mg/kg BW (0.24 µM), concentrations above and below this value were selected. Acetonitrile/water (1:19 v/v) was used as the vehicle control. Oligodendrocyte populations were quantified using an endogenous proteolipid protein (PLP) reporter line. Organoids were fixed at 8 weeks post-differentiation and imaged using a confocal microscope. The morphology of PLP+ cells was analyzed using the 3D Suite plugin in ImageJ, which enabled precise quantification of their three-dimensional structural properties. In addition, we assessed spontaneous and evoked neuronal network function under chronic low-dose DA exposure using calcium imaging and a high-density microelectrode array (HD-MEA). Spontaneous network spiking and bursting were quantified by time-series recordings of network activity. Cortical excitability was measured using an input-output experiment, applying input voltages ranging from 100 mV to 800 mV and recording the corresponding electrical output. To assess synaptic plasticity, brief, repetitive bursts of thetafrequency electrical stimuli (theta-burst stimulation) were applied. Additionally, markers of excitatory and inhibitory synapses were quantified using qRT-PCR and immunohistochemistry. 

Results: Four-week exposure to 0.1 µM and 3 .16 µM led to a phenotype of oligodendrocytes that were significantly more compact and spherical. While exposure to 0.32 µM led to a phenotype of oligodendrocytes that were significantly less compact and spherical and significantly larger in volume. Suggesting that exposure to 0.1 µM and 3 .16 µM caused a 

decrease in oligodendrocyte maturity while exposure to 0.32 µM increased oligodendrocyte maturity. In addition, four-week exposure to 0.1 µM and 3.16 µM DA led to a significant increase in spontaneous bursting in neural organoids (P ≤ 0.01 and P ≤ 0.0001, respectively). 

Additionally, the marked increase in spiking activity in the 3.16 µM DA organoids was associated with a significant increase in the number of spike-sorted neurons in this group (P ≤ 0.01). Despite these increases in spontaneous activity, the organoids’ functional output in response to an electrical stimuli was significantly reduced: evoked mean spikes, representing cortical excitability, were significantly decreased in both the 0.1 µM and 3.16 µM DA groups (P 

≤ 0.05 and P ≤ 0.005, respectively). Short-term plasticity, measured by the evoked activity, total spikes, and active area driven by a theta-burst stimulation, was also significantly reduced in the 3.16 µM DA group (P ≤ 0.05). 

Conclusions: These findings suggest that DA exposure, even below the current regulatory limits, poses potential risks to human health by impairing neuronal network development and function. Our future work will focus on elucidating DA’s specific effects on cognitive processes such as learning and memory, using an open-loop "learning assay" with HD-MEA and neural organoids. This project has contributed to a deeper understanding of how domoic acid exposure can disrupt neural network dynamics, with implications for cognitive function. The insights gained hold significant implications for public health, potentially informing updates to regulatory guidelines on chronic, low-dose DA exposure. 

Uncovering Unseen Interactions in Autism Genetic Risk 

Presenter: Alan Kim 

Background: The increasing prevalence of neurodevelopmental disorders, such as ASD, ADHD, and schizophrenia, is of major public health relevance. A prevailing theory is that this increase may be in part mediated by the interaction between environmental exposures and transcriptional-level changes. However, previous analyses of GWAS and EWAS studies have relied on surface-level analyses of single-nucleotide polymorphisms (SNPs) and methylation patterns to identify genes and pathways of interest, which ignore the contributions of network degree and connectedness in exploring potential nodes of interest in disease development.   

Objective: Expanding the identification of SNPs and methylation patterns of interest could illuminate relevant pathways and differentially expressed genes involved in the underlying etiology of these conditions. Therefore, we seek to identify which genes and gene-sets overlap in GWAS and EWAS datasets to identify new pathways of interest disturbed in autism diagnoses. 

Methods: First, we will identify GWAS and EWAS datasets with annotated SNPs and methylation sites of interest in coding and non-coding regions expressed in autism spectrum disorders. Then, taking a network-based approach into our analysis will allow for the identification of critical genes and genesets associated with autism diagnoses and biological processes. Exploring gene functions, such as human transporters and micronutrients, may also provide some clues as to the nature of autism development. Scoring genes based on their predicted loss of function (pLOF) may also provide some insights as to which nodes may be critical targets for further study. 

Results and Conclusion: Preliminary screens of GWAS and EWAS datasets have provided initial directions for trial runs. Further optimization of inclusion/exclusion criteria, a more expansive dataset search, exploration of human transporters and micronutrient interactions, and preliminary analysis of genelists will provide a more robust set of genes and genesets involved in the etiology of autism spectrum disorders. 

Modeling chronic lung disease with microphysiological systems

Presenter: Anuj Lagowala

Rationale

The interactions between the lung microenvironment and epithelial cells influence the progression of many respiratory diseases, including Chronic Obstructive Pulmonary Disease (COPD). Alterations of the extracellular matrix (ECM) is a hallmark of COPD, with characteristic destruction of alveolar walls in emphysema and aberrant collagen deposition in fibrosis. In addition to maintaining the lung’s physical architecture, ECM provides physical cues to cells through the stiffness of the local microenvironment and biochemical cues that can govern proliferation, differentiation, and maintenance of epithelial cell barrier integrity. Characterizing the changes caused by COPD and CS exposure on lung cells and ECM can provide understanding of how ECM alteration contributes to disease pathology, and further identify ECM-related targets for intervention. Traditional in vitro models of lack the complex three-dimensional architecture and native ECM composition of in vivo tissue that can alter the behavior of cells. 3D microphysiological models can recapitulate these conditions and are used to determine the effects of cigarette smoke and respiratory disease on the native lung cells in coculture, including the epithelium and fibroblasts. Using precision-cut lung slices (PCLS), we can maintain the native architecture, cell types, and ECM composition of the human lung while exposing them to cigarette smoke insult to determine how exposures can induce change in both cells and the microenvironment, and how they in turn affect each other in a cycle that causes progressive lung disease. Characterizing ECM-related changes in COPD and exposure, such as stiffness and ECM composition will further inform the creation of ECM-based scaffolds that mimic healthy or COPD microenvironments in order to determine the extent to which they drive the progression of chronic lung disease. 

Methods

Human lungs from normal and COPD affected patients were obtained, filled with agarose, and sliced into 150-500 um slices and maintained in ex-vivo culture to allow for the use of live human cells in their native microenvironment and configuration. During culture, the slices were exposed to direct cigarette smoke or humidified air as a control, and HAPLN1, a hyaluronic acid linker protein. These live slices were placed under mechanical compressive testing, which applied a 20% strain and measured force to identify the Young’s modulus of the tissue slices.

Results

The mechanical testing showed that the CS exposure caused a larger increase in the Young’s modulus of the lung slices. This effect was mitigated by the HAPLN1 treatment, with no significant difference from the control condition. Furthermore, untreated lung slices from COPD patients had a significantly higher Young’s modulus as compared to untreated normal patients. In slices recellularized with fibroblasts after decellularization and exposures, staining for β-Galactosidase showed that the number of senescent fibroblasts was higher in CS-exposed matrix. Additional staining for α-SMA showed greater expression in fibroblasts within the CS-exposed matrix.

Conclusions

Our data indicates that CS exposure alters the mechanical properties of the lung ECM. This effect can be ameliorated with the treatment of an ECM protein, HAPLN1, which indicates that the ECM is a potential target for interventions to alter COPD-related changes in the lung microenvironment. Furthermore, the exposure creates a persistent effect in the lung ECM that can induce senescence and myofibroblast differentiation, phenotypes associated with COPD pathology, within healthy fibroblasts, independent of the original exposure. Ongoing experiments continue to characterize ECM composition of treated PCLS and determining the effects of reseeding epithelial cells on dECM scaffolds.

Multiomics and Exposomics Integration Reveals Molecular Insights into Childhood Asthma 

Background: Asthma is a complex inflammatory disease marked by airway obstruction and hyperresponsiveness and is known to be influenced by genetic and environmental factors ^1,2. How these factors modulate asthma severity is poorly characterized ². 

Objective: To characterize how molecular features of asthma severity relate to environmental exposures using multi-omics data from CD4⁺ T-cell and CD16⁺ monocyte transcriptomics and miRNA, and serum proteomics.

Methods: Missing data were imputed using missForest (omics) and median (exposures). Principal component analysis (PCA) was used to assess variance, and differential abundance analysis was conducted with limma-voom to identify features associated with FEV₁/FVC, adjusting for age, sex, and race. Features with a false discovery rate (FDR)-corrected p-value less than 0.05 and an absolute fold change greater than 1.5 were correlated with exposures using Spearman correlation. Pathway analysis was performed using clusterProfiler.  

Results: Most variance in the feature space was captured by the first two principal components, with three sample outliers removed. CD4⁺ T-cell RNA (2,192 features) and CD16⁺ monocyte RNA (1,982 features) had the highest number of differentially abundant features, followed by serum proteins (288 features), CD16⁺ miRNA (5 features), and CD4⁺ miRNA (2 features). Molecular features were linked to T-cell activation and stress response pathways. Allergens were associated with pathways related to antigen presentation, indoor air pollutants with oxidative stress, and chemical exposures with both.  

Conclusion: Asthma severity is associated with distinct molecular pathways influenced by exposure, including allergens, indoor air pollutants and chemical exposures. These exposures appear to differentially affect antigen presentation, oxidative stress, and immune regulation suggesting complex interactions between environmental factors and molecular responses. Future work will focus on exploring these mechanisms to better understand how exposures shape the progression and severity of asthma.  

References

1. Guidelines for the Diagnosis and Management of Asthma 2007 (EPR-3). NHLBI, NIH https://www.nhlbi.nih.gov/health-topics/guidelines-for-diagnosis-management-of-asthma. 
2. Lötvall, J. et al. Asthma endotypes: a new approach to classification of disease entities within the asthma syndrome. J. Allergy Clin. Immunol. 127, 355–360 (2011). 

Microglial activation in organoids with 16p11.2 deletion and autism 

Alex Rittenhouse, Lena Smirnova 

Autism spectrum disorder (ASD) is a major public health concern with complex genetic and environmental underpinnings. Deletion of the 16p11.2 chromosome region, the most common ASD-associated mutation, accounts for ~1% of cases and exhibits considerable clinical variability, ranging from ASD diagnosis and/or seizures to no neurodevelopmental symptoms. It currently is unknown why individuals with similar mutations develop such a wide array of symptoms. Environmental factors associated with ASD, such as maternal inflammation, may interact with genetic risk factors, but the associations and relevant cellular mechanisms remain poorly understood. 

In this study, we hypothesized that the 16p11.2 deletion increases susceptibility to environmental influences, leading to exacerbated ASD-related pathology. To investigate this, we generated immune-competent brain assembloids containing microglia using cell lines derived from individuals with 16p11.2 deletions (with and without ASD/seizures) and control lines. 

At baseline, 16p11.2 deletion lines exhibited reduced secretion of pro-inflammatory cytokines compared to controls. However, following exposure to proinflammatory cytokines, fever and Tylenol as a model of maternal inflammation in early pregnancy, all 16p11.2deletion lines showed exaggerated increased pro-inflammatory cytokine secretion. Additionally, microglia within 16p11.2deletion assembloids demonstrated a significant increase in CD68 and TREM2 intensity, a response not observed in controls. CD68 and TREM2 are both markers of microglia activation, and TREM2 is essential for microglia-mediated synapse pruning. This heightened microglial activation in response to inflammation in organoids with 16p.11.2 but not control organoids suggests that the 16p11.2 deletion may amplify vulnerability to environmental stressors during neurodevelopment with potential impacts on synaptic densities, which will be addressed in subsequent following experiments. 

These findings highlight a potential mechanism linking genetic risk and environmental factors in ASD, with microglial activation as a key player. Further investigation is needed to explore longterm impacts of this early immune response on neurodevelopmental outcomes in brain organoids such as excitatory and inhibitory synapse balance, and assembloids neuronal connectivity and activity. 

Gestational arsenic exposure alters one-carbon metabolism enzyme expression in the placenta and maternal liver

Maternal mortality continues to rise in the US and cardiovascular complications currently account for 7-15% of all pregnancy-related deaths. Exposure to environmental toxicants, including inorganic arsenic (iAs), has been implicated in cardiovascular disease development. In prior studies from our group, we found that gestational exposure to iAs altered maternal cardiac remodeling during and after pregnancy, but the mechanism underlying these changes has not been determined; therefore, the objective of this study is to examine one-carbon metabolism (OCM) as a potential target of iAs toxicity during gestation. Six- to eight-week old C57BL6/J mice were exposed to 0, 100, or 1000 ppm iAs in drinking water starting at E2.5.  Placentas and maternal livers were harvested at E17.5 and snap frozen in liquid nitrogen. RNA was extracted and used for qRT-PCR analysis, and protein homogenates were generated for western blotting. Exposure to 1000 ppm iAs significantly decreased protein expression of serine hydroxymethyltransferase 2 (SHMT2), significantly increased protein expression of methylenetetrahydrofolate dehydrogenase 2 (MTHFD2) and thymidylate synthase (TS) in the maternal liver, and trended towards a significant reduction in methionine adenosyltransferase 2A (MAT2A) expression compared to non-exposed control. These same trends were not observed in the placenta, where exposure to 1000 ppm iAs significantly increased MAT2A expression compared to 100 ppm exposure; however, transcriptional changes were not observed. Taken together, our results suggest that gestational iAs exposure is sufficient to alter the expression of OCM enzymes in the placenta and maternal liver during late pregnancy. These findings are of particular interest, as OCM is necessary for generating critical methyl units for arsenic metabolism and detoxification, and the liver is the primary site for iAs metabolism in the body. Future work will investigate whether our observed differences in OCM protein expression are sufficient to alter OCM enzyme activity and iAs metabolism in the maternal liver and placenta. This study also emphasizes the importance of mechanistic studies for environmental contaminants and their effects on maternal health, as these studies are currently lacking. 

Heavy metal mixture elicits less than additive neuronal impairment in human cortical microphysiological system

Breanne Kincaid

Heavy metal exposure can impair neurodevelopment through disrupted cellular proliferation, differentiation, and network formation. While contaminants are regulated individually, exposures don’t occur in isolation. Our objective is to characterize metal mixture toxicity during an early and late window of neurodevelopment in a human cortical microphysiological system (MPS) in order to determine whether metal neurotoxicity is additive at low doses reflective of current exposure levels.

MPS were treated with low doses of Pb, Cr, As, and Cd alone and in combination during early (0-4 week) and late (8-12 week) stages of differentiation. Viability was confirmed with resazurin reduction. Fluorescence microscopy was used to quantify the density of synapses with excitatory neurons (SYN1+/PSD-95+), as well as neurite outgrowth (BTUBIII).

Synaptic density was analyzed with SynapseJ while neurite length and density were quantified with Sholl Analysis. Following late developmental exposure to Cr, pre- and postsynaptic puncta analysis with SynapseJ revealed a slight reduction in the density of synapses with excitatory neurons, attributed to a reduction in presynaptic puncta fluorescence intensity density. No other condition, including mixture, elicited significant effects. Subsequent neurite outgrowth assay revealed no significant effect for Cr during this same timepoint, indicating that impaired synaptogenesis is likely not due to impaired microtubule growth in extending axons or dendrites. However, neurite outgrowth was impaired in Pb exposed 8-to-12-week BrainSpheres, with the same magnitude of impairment observed in mixture-exposed BrainSpheres. Electrophysiology analysis of 8-to-12-week organoids revealed shows a substantial increase in the number and rate of spikes for Cd-exposed samples, with a trend toward dampening of these metrics for Cr-exposed samples.

In assessing deviation from theoretical additivity levels for mixture exposure, the majority of endpoints assessed exhibited a less than additive effect, with the exception of peak amplitude, which exhibited a slightly above-additive response. Taken together, this data provides evidence of an antagonistic relationship between contaminants with confirmed synaptic toxicity when they are administered as a complex mixture.