Preface to the special issue of Food and Chemical Toxicology on "New approach methodologies and machine learning in food safety and chemical risk assessment: Development of reproducible, open-source, and user-friendly tools for exposure, toxicokinetic, and toxicity assessments in the 21st century".

This Special Issue contains articles on applications of various new approach methodologies (NAMs) in the field of toxicology and risk assessment. These NAMs include in vitro high-throughput screening, quantitative structure-activity relationship (QSAR) modeling, physiologically based pharmacokinetic (PBPK) modeling, network toxicology analysis, molecular docking simulation, omics, machine learning, deep learning, and "template-and-anchor" multiscale computational modeling. These in vitro and in silico approaches complement each other and can be integrated together to support different applications of toxicology, including food safety assessment, dietary exposure assessment, chemical toxicity potency screening and ranking, chemical toxicity prediction, chemical toxicokinetic simulation, and to investigate the potential mechanisms of toxicities, as introduced further in selected articles in this Special Issue.

Exploring the potential and challenges of developing physiologically-based toxicokinetic models to support human health risk assessment of microplastic and nanoplastic particles.

Microplastics (MPs) and nanoplastics (NPs) pollution has emerged as a significant and widespread environmental issue. Humans are inevitably exposed to MPs and NPs via ingestion, inhalation, and dermal contacts from various sources. However, mechanistic knowledge of their distribution, interaction, and potency in the body is still lacking. To address this knowledge gap, we have undertaken the task of elucidating the toxicokinetic (TK) behaviors of MPs and NPs, aiming to provide mechanistic information for constructing a conceptual physiologically based toxicokinetic (PBTK) model to support in silico modeling approaches. Our effort involved a thorough examination of the existing literature and data collation on the presence of MPs in the human body and in vitro/ex vivo/in vivo biodistribution across various cells and tissues. By comprehending the absorption, distribution, metabolism, and excretion mechanisms of MPs and NPs in relation to their physicochemical attributes, we established a foundational understanding of the link between external exposure and internal tissue dosimetry. We observed that particle size and surface chemistry have been thoroughly explored in previous experimental studies. However, certain attributes, such as polymer type, shape, and biofilm/biocorona, warrant attention and further examination. We discussed the fundamental disparities in TK properties of MPs/NPs from those of engineered nanoparticles. We proposed a preliminary PBTK framework with several possible modeling approaches and discussed existing challenges for further investigation. Overall, this article provides a comprehensive compilation of existing TK data of MPs/NPs, a critical overview of TK processes and mechanisms, and proposes potential PBTK modeling approaches, particularly regarding their applicability to the human system, and outlines future perspectives for developing PBTK models and their integration into human health risk assessment of MPs and NPs.

Toxicokinetics, dose-response, and risk assessment of nanomaterials: Methodology, challenges, and future perspectives.

The rapid growth of nanomaterial applications has raised safety concerns for human health. A number of studies have been conducted to assess the toxicokinetics, toxicology, dose-response, and risk assessment of different nanomaterials using in vitro and in vivo animal and human models. However, current studies cannot meet the demand for efficient assessment of toxicokinetics, dose-response relationships, or the toxicological risk arising from the rapidly increasing number of newly synthesized nanomaterials. In this article, we review the methods for conducting toxicokinetics, hazard identification, dose-response, exposure, and risk assessment studies of nanomaterials, identify the knowledge gaps, and discuss the challenges remaining. We provide the rationale behind the appropriate design of nanomaterial plasma toxicokinetic and tissue distribution studies, including caveats on the interpretation and correlation of in vitro and in vivo toxicology studies. The potential of using physiologically based pharmacokinetic (PBPK) models to extrapolate toxicokinetic and toxicity findings from in vitro to in vivo and from animals to humans is discussed, and the knowledge gaps of PBPK modeling for nanomaterials are identified. While challenges still exist, there has been progress in the toxicokinetics, hazard identification, and risk assessment of nanomaterials in the past two decades. Recent advancements in the field are highlighted with relevant examples. We also share latest guidelines as well as our perspectives on future studies needed to characterize the toxicokinetics, toxicity, and dose-response relationship in support of nanomaterial risk assessment. This article is categorized under: Toxicology and Regulatory Issues in Nanomedicine > Toxicology of Nanomaterials Toxicology and Regulatory Issues in Nanomedicine > Regulatory and Policy Issues in Nanomedicine.

Integration of Toxicogenomics and Physiologically Based Pharmacokinetic Modeling in Human Health Risk Assessment of Perfluorooctane Sulfonate.

Toxicogenomics and physiologically based pharmacokinetic (PBPK) models are useful approaches in chemical risk assessment, but the methodology to incorporate toxicogenomic data into a PBPK model to inform risk assessment remains to be developed. This study aimed to develop a probabilistic human health risk assessment approach by integrating toxicogenomic dose-response data and PBPK modeling using perfluorooctane sulfonate (PFOS) as a case study. Based on the available human in vitro and mouse in vivo toxicogenomic data, we identified the differentially expressed genes (DEGs) at each exposure paradigm/duration. Kyoto Encyclopedia of Genes and Genomes and disease ontology enrichment analyses were conducted on the DEGs to identify significantly enriched pathways and diseases. The dose-response data of DEGs were analyzed using the Bayesian benchmark dose (BMD) method. Using a previously published PBPK model, the gene BMDs were converted to human equivalent doses (HEDs), which were summarized to pathway and disease HEDs and then extrapolated to reference doses (RfDs) by considering an uncertainty factor of 30 for mouse in vivo data and 10 for human in vitro data. The results suggested that the median RfDs at different exposure paradigms were similar to the 2016 U.S. Environmental Protection Agency's recommended RfD, while the RfDs for the most sensitive pathways and diseases were closer to the recent European Food Safety Authority's guidance values. In conclusion, genomic dose-response data and PBPK modeling can be integrated to become a useful alternative approach in risk assessment of environmental chemicals. This approach considers multiple endpoints, provides toxicity mechanistic insights, and does not rely on apical toxicity endpoints.

In vitro-in silico-based probabilistic risk assessment of combined exposure to bisphenol A and its analogues by integrating ToxCast high-throughput in vitro assays with in vitro to in vivo extrapolation (IVIVE) via physiologically based pharmacokinetic (PBPK) modeling.

Combined risk assessment of endocrine effects of bisphenol A (BPA) and its analogues, such as bisphenols S, F, and AF (BPS, BPF, and BPAF), is challenging due to lack of related common toxicity metrics. This study conducted a population-based in vitro-to-in vivo extrapolation using physiologically based pharmacokinetic (PBPK) models coupled with Monte Carlo simulations to convert ToxCast in vitro estrogen receptor (ER) assays to human equivalent doses (HEDs). The ER pathway-based HEDs were compared with HEDs from animal studies and used to assess the combined risks for different populations across different countries/regions in a probabilistic manner. The estimated ER pathway-based HEDs for the four bisphenols (BPs) matched the animal-derived HEDs. The HEDs for the ER gene transcription (the common biological process target among BPs) were 0.40 (2.5th-97.5th percentiles: 0.06-5.42), 4.43 (0.69-53.84), 3.30 (0.51-626.57), and 1.12 (0.16-9.73) mg/kg/day for BPA, BPS, BPF, and BPAF, respectively. Results suggest a potentially moderate concern for combined risks of activating the ER pathway for toddlers and adults with high dietary exposures. This study presents in vitro-based credible HEDs for the four BPs and represents an advancement in the application of in vitro-in silico-based alternative approaches in human health risk assessment.

Probabilistic human health risk assessment of perfluorooctane sulfonate (PFOS) by integrating in vitro, in vivo toxicity, and human epidemiological studies using a Bayesian-based dose-response assessment coupled with physiologically based pharmacokinetic (PBPK) modeling approach.

BACKGROUND: Environmental exposure to perfluorooctane sulfonate (PFOS) is associated with various adverse outcomes in humans. However, risk assessment for PFOS with the traditional risk estimation method is faced with multiple challenges because there are high variabilities and uncertainties in its toxicokinetics and toxicity between species and among different types of studies. OBJECTIVES: This study aimed to develop a robust probabilistic risk assessment framework accounting for interspecies and inter-experiment variabilities and uncertainties to derive the human equivalent dose (HED) and reference dose for PFOS. METHODS: A Bayesian dose-response model was developed to analyze selected 34 critical studies, including human epidemiological, animal in vivo, and ToxCast in vitro toxicity datasets. The dose-response results were incorporated into a multi-species physiologically based pharmacokinetic (PBPK) model to reduce the toxicokinetic/toxicodynamic variabilities. In addition, a population-based probabilistic risk assessment of PFOS was performed for Asian, Australian, European, and North American populations, respectively, based on reported environmental exposure levels. RESULTS: The 5th percentile of HEDs derived from selected studies was estimated to be 21.5 (95% CI: 10.6-36.3) ng/kg/day. After exposure to environmental levels of PFOS, around 50% of the population in all studied populations would likely have >20% of increase in serum cholesterol, but the effects on other endpoints were estimated to be minimal (<10% changes). There was a small population (~10% of the population) that was highly sensitive to endocrine disruption and cellular response by environmental PFOS exposure. CONCLUSION: Our results provide insights into a complete risk characterization of PFOS and may help regulatory agencies in the reevaluation of PFOS risk. Our new probabilistic approach can conduct dose-response analysis of different types of toxicity studies simultaneously and this method could be used to improve risk assessment for other perfluoroalkyl substances (PFAS).

Incorporating Exogenous and Endogenous Exposures into Dietary Risk Assessment of Nitrates and Nitrites in Vegetables: A Probabilistic Integrated Toxicokinetic Modeling Approach.

This study aimed to estimate the dietary risk of nitrates and nitrites in vegetables based on internal dose in a probabilistic manner by integrating exogenous exposure based on measured concentrations in vegetables with endogenous exposure using a toxicokinetic (TK) model. We optimized and validated a previous TK model and incorporated Monte Carlo simulations to account for variability across different age populations for predicting internal dose. High levels of nitrates were detected in leafy vegetables (from 545 ± 274 to 1641 ± 873 mg/kg). Nitrite contents of vegetables were generally low (from 1.26 ± 1.40 to 8.20 ± 14.1 mg/kg). The dietary risk was found to be different based on internal versus external dose, suggesting that it is critical to include endogenous nitrite formation into risk assessment. Nitrate and nitrite exposure from vegetables is unlikely to result in appreciable risks for most populations but may be a potential risk for preschoolers.

Bayesian evaluation of a physiologically based pharmacokinetic (PBPK) model for perfluorooctane sulfonate (PFOS) to characterize the interspecies uncertainty between mice, rats, monkeys, and humans: Development and performance verification.

A challenge in the risk assessment of perfluorooctane sulfonate (PFOS) is the large interspecies differences in its toxicokinetics that results in substantial uncertainty in the dosimetry and toxicity extrapolation from animals to humans. To address this challenge, the objective of this study was to develop an open-source physiologically based pharmacokinetic (PBPK) model accounting for species-specific toxicokinetic parameters of PFOS. Considering available knowledge about the toxicokinetic properties of PFOS, a PBPK model for PFOS in mice, rats, monkeys, and humans after intravenous and oral administrations was created. Available species-specific toxicokinetic data were used for model calibration and optimization, and independent datasets were used for model evaluation. Bayesian statistical analysis using Markov chain Monte Carlo (MCMC) simulation was performed to optimize the model and to characterize the uncertainty and interspecies variability of chemical-specific parameters. The model predictions well correlated with the majority of datasets for all four species, and the model was validated with independent data in rats, monkeys, and humans. The model was applied to predict human equivalent doses (HEDs) based on reported points of departure in selected critical toxicity studies in rats and monkeys following U.S. EPA's guidelines. The lower bounds of the model-derived HEDs were overall lower than the HEDs estimated by U.S. EPA (e.g., 0.2 vs. 1.3 µg/kg/day based on the rat plasma data). This integrated and comparative analysis provides an important step towards improving interspecies extrapolation and quantitative risk assessment of PFOS, and this open-source model provides a foundation for developing models for other perfluoroalkyl substances.

Assessing Global Human Exposure to T-2 Toxin via Poultry Meat Consumption Using a Lifetime Physiologically Based Pharmacokinetic Model.

Residue depletion of T-2 toxin in chickens after oral gavage at 2.0 mg/kg twice daily for 2 days was determined in this study. A flow-limited physiologically based pharmacokinetic (PBPK) model was developed for lifetime exposure assessment in chickens. The model was calibrated with data from the residue depletion study and then validated with independent data. A local sensitivity analysis was performed, and 16 sensitive parameters were subjected to Monte Carlo analysis. The population PBPK model was applied to estimate daily intake values of T-2 toxin in different countries based on reported consumption factors and the guidance value of 0.25 mg/kg in feed for chickens by the European Food Safety Authority (EFSA). The predicted daily intakes in different countries were all lower than the EFSA's total daily intake, suggesting that the EFSA's guidance value has minimal risk. This model provides a foundation for scaling to other mycotoxins and other food animal species.

Probabilistic risk assessment of gold nanoparticles after intravenous administration by integrating in vitro and in vivo toxicity with physiologically based pharmacokinetic modeling.

This study aimed to conduct an integrated and probabilistic risk assessment of gold nanoparticles (AuNPs) based on recently published in vitro and in vivo toxicity studies coupled to a physiologically based pharmacokinetic (PBPK) model. Dose-response relationships were characterized based on cell viability assays in various human cell types. A previously well-validated human PBPK model for AuNPs was applied to quantify internal concentrations in liver, kidney, skin, and venous plasma. By applying a Bayesian-based probabilistic risk assessment approach incorporating Monte Carlo simulation, probable human cell death fractions were characterized. Additionally, we implemented in vitro to in vivo and animal-to-human extrapolation approaches to independently estimate external exposure levels of AuNPs that cause minimal toxicity. Our results suggest that under the highest dosing level employed in existing animal studies (worst-case scenario), AuNPs coated with branched polyethylenimine (BPEI) would likely induce ∼90-100% cellular death, implying high cytotoxicity compared to <10% cell death induced by low-to-medium animal dosing levels, which are commonly used in animal studies. The estimated human equivalent doses associated with 5% cell death in liver and kidney were around 1 and 3 mg/kg, respectively. Based on points of departure reported in animal studies, the human equivalent dose estimates associated with gene expression changes and tissue cell apoptosis in liver were 0.005 and 0.5 mg/kg, respectively. Our analyzes provide insights into safety evaluation, risk prediction, and point of departure estimation of AuNP exposure for humans and illustrate an approach that could be applied to other NPs when sufficient data are available.

Development and application of a population physiologically based pharmacokinetic model for penicillin G in swine and cattle for food safety assessment.

Penicillin G is a widely used antimicrobial in food-producing animals, and one of the most predominant drug residues in animal-derived food products. Due to reduced sensitivity of bacteria to penicillin, extralabel use of penicillin G is common, which may lead to violative residues in edible tissues and cause adverse reactions in consumers. This study aimed to develop a physiologically based pharmacokinetic (PBPK) model to predict drug residues in edible tissues and estimate extended withdrawal intervals for penicillin G in swine and cattle. A flow-limited PBPK model was developed with data from Food Animal Residue Avoidance Databank using Berkeley Madonna. The model predicted observed drug concentrations in edible tissues, including liver, muscle, and kidney for penicillin G both in swine and cattle well, including data not used in model calibration. For extralabel use (5× and 10× label dose) of penicillin G, Monte Carlo sampling technique was applied to predict times needed for tissue concentrations to fall below established tolerances for the 99th percentile of the population. This model provides a useful tool to predict tissue residues of penicillin G in swine and cattle to aid food safety assessment, and also provide a framework for extrapolation to other food animal species.

Performance Assessment and Translation of Physiologically Based Pharmacokinetic Models From acslX to Berkeley Madonna, MATLAB, and R Language: Oxytetracycline and Gold Nanoparticles As Case Examples.

Many physiologically based pharmacokinetic (PBPK) models for environmental chemicals, drugs, and nanomaterials have been developed to aid risk and safety assessments using acslX. However, acslX has been rendered sunset since November 2015. Alternative modeling tools and tutorials are needed for future PBPK applications. This forum article aimed to: (1) demonstrate the performance of 4 PBPK modeling software packages (acslX, Berkeley Madonna, MATLAB, and R language) tested using 2 existing models (oxytetracycline and gold nanoparticles); (2) provide a tutorial of PBPK model code conversion from acslX to Berkeley Madonna, MATLAB, and R language; (3) discuss the advantages and disadvantages of each software package in the implementation of PBPK models in toxicology, and (4) share our perspective about future direction in this field. Simulation results of plasma/tissue concentrations/amounts of oxytetracycline and gold from different models were compared visually and statistically with linear regression analyses. Simulation results from the original models were correlated well with results from the recoded models, with time-concentration/amount curves nearly superimposable and determination coefficients of 0.86-1.00. Step-by-step explanations of the recoding of the models in different software programs are provided in the Supplementary Data. In summary, this article presents a tutorial of PBPK model code conversion for a small molecule and a nanoparticle among 4 software packages, and a performance comparison of these software packages in PBPK model implementation. This tutorial helps beginners learn PBPK modeling, provides suggestions for selecting a suitable tool for future projects, and may lead to the transition from acslX to alternative modeling tools.

Estimation of tulathromycin depletion in plasma and milk after subcutaneous injection in lactating goats using a nonlinear mixed-effects pharmacokinetic modeling approach.

BACKGROUND: Extra-label use of tulathromycin in lactating goats is common and may cause violative residues in milk. The objective of this study was to develop a nonlinear mixed-effects pharmacokinetic (NLME-PK) model to estimate tulathromycin depletion in plasma and milk of lactating goats. Eight lactating goats received two subcutaneous injections of 2.5 mg/kg tulathromycin 7 days apart; blood and milk samples were analyzed for concentrations of tulathromycin and the common fragment of tulathromycin (i.e., the marker residue CP-60,300), respectively, using liquid chromatography mass spectrometry. Based on these new data and related literature data, a NLME-PK compartmental model with first-order absorption and elimination was used to model plasma concentrations and cumulative excreted amount in milk. Monte Carlo simulations with 100 replicates were performed to predict the time when the upper limit of the 95% confidence interval of milk concentrations was below the tolerance. RESULTS: All animals were healthy throughout the study with normal appetite and milk production levels, and with mild-moderate injection-site reactions that diminished by the end of the study. The measured data showed that milk concentrations of the marker residue of tulathromycin were below the limit of detection (LOD = 1.8 ng/ml) 39 days after the second injection. A 2-compartment model with milk as an excretory compartment best described tulathromycin plasma and CP-60,300 milk pharmacokinetic data. The model-predicted data correlated with the measured data very well. The NLME-PK model estimated that tulathromycin plasma concentrations were below LOD (1.2 ng/ml) 43 days after a single injection, and 62 days after the second injection with a 95% confidence. These estimated times are much longer than the current meat withdrawal time recommendation of 18 days for tulathromycin in non-lactating cattle. CONCLUSIONS: The results suggest that twice subcutaneous injections of 2.5 mg/kg tulathromycin are a clinically safe extra-label alternative approach for treating pulmonary infections in lactating goats, but a prolonged withdrawal time of at least 39 days after the second injection should be considered to prevent violative residues in milk and any dairy goat being used for meat should have an extended meat withdrawal time.

A computational framework for interspecies pharmacokinetics, exposure and toxicity assessment of gold nanoparticles.

AIM: To develop a comprehensive computational framework to simulate tissue distribution of gold nanoparticles (AuNP) across several species. MATERIALS & METHODS: This framework was built on physiologically based pharmacokinetic modeling, calibrated and evaluated with multiple independent datasets. RESULTS: Rats and pigs seem to be more appropriate models than mice in animal-to-human extrapolation of AuNP pharmacokinetics and that the dose and age should be considered. Incorporation of in vitro and/or in vivo cellular uptake and toxicity data into the model improved toxicity assessment of AuNP. CONCLUSION: These results partially explain the current low translation rate of nanotechnology-based drug delivery systems from mice to humans. This simulation approach may be applied to other nanomaterials and provides guidance to design future translational studies.

Estimation of residue depletion of cyadox and its marker residue in edible tissues of pigs using physiologically based pharmacokinetic modelling.

Physiologically based pharmacokinetic (PBPK) models are powerful tools to predict tissue distribution and depletion of veterinary drugs in food animals. However, most models only simulate the pharmacokinetics of the parent drug without considering their metabolites. In this study, a PBPK model was developed to simultaneously describe the depletion in pigs of the food animal antimicrobial agent cyadox (CYA), and its marker residue 1,4-bisdesoxycyadox (BDCYA). The CYA and BDCYA sub-models included blood, liver, kidney, gastrointestinal tract, muscle, fat and other organ compartments. Extent of plasma-protein binding, renal clearance and tissue-plasma partition coefficients of BDCYA were measured experimentally. The model was calibrated with the reported pharmacokinetic and residue depletion data from pigs dosed by oral gavage with CYA for five consecutive days, and then extrapolated to exposure in feed for two months. The model was validated with 14 consecutive day feed administration data. This PBPK model accurately simulated CYA and BDCYA in four edible tissues at 24-120 h after both oral exposure and 2-month feed administration. There was only slight overestimation of CYA in muscle and BDCYA in kidney at earlier time points (6-12 h) when dosed in feed. Monte Carlo analysis revealed excellent agreement between the estimated concentration distributions and observed data. The present model could be used for tissue residue monitoring of CYA and BDCYA in food animals, and provides a foundation for developing PBPK models to predict residue depletion of both parent drugs and their metabolites in food animals.