Advances in computational methods along the exposure to toxicological response paradigm.

Human health risk assessment is a function of chemical toxicity, bioavailability to reach target biological tissues, and potential environmental exposure. These factors are complicated by many physiological, biochemical, physical and lifestyle factors. Furthermore, chemical health risk assessment is challenging in view of the large, and continually increasing, number of chemicals found in the environment. These challenges highlight the need to prioritize resources for the efficient and timely assessment of those environmental chemicals that pose greatest health risks. Computational methods, either predictive or investigative, are designed to assist in this prioritization in view of the lack of cost prohibitive in vivo experimental data. Computational methods provide specific and focused toxicity information using in vitro high throughput screening (HTS) assays. Information from the HTS assays can be converted to in vivo estimates of chemical levels in blood or target tissue, which in turn are converted to in vivo dose estimates that can be compared to exposure levels of the screened chemicals. This manuscript provides a review for the landscape of computational methods developed and used at the U.S. Environmental Protection Agency (EPA) highlighting their potentials and challenges.

Thresholds Derived From Common Measures in Rat Studies Are Predictive of Liver Tumorigenic Chemicals.

We hypothesized that typical tissue and clinical chemistry (ClinChem) end points measured in rat toxicity studies exhibit chemical-independent biological thresholds beyond which cancer occurs. Using the rat in vivo TG-GATES study, 75 chemicals were examined across chemical-dose-time comparisons that could be linked to liver tumor outcomes. Thresholds for liver weight to body weight (LW/BW) and 21 serum ClinChem end points were defined as the maximum and minimum values for those exposures that did not lead to liver tumors in rats. Upper thresholds were identified for LW/BW (117%), aspartate aminotransferase (195%), alanine aminotransferase (141%), alkaline phosphatase (152%), and total bilirubin (115%), and lower thresholds were identified for phospholipids (82%), relative albumin (93%), total cholesterol (82%), and total protein (94%). Thresholds derived from the TG-GATES data set were consistent across other acute and subchronic rat studies. A training set of ClinChem and LW/BW thresholds derived from a 38 chemical training set from TG-GATES was predictive of liver tumor outcomes for a test set of 37 independent TG-GATES chemicals (91%). The thresholds were most predictive when applied to 7d treatments (98%). These findings provide support that biological thresholds for common end points in rodent studies can be used to predict chemical tumorigenic potential.

Identification of differentially expressed genes and networks related to hepatic lipid dysfunction.

A range of chemical exposures that resulted in the specific pathology of hepatic lipid dysfunction in rats were selected from DrugMatrix, a publicly available toxicogenomic database. Raw microarray data collected from these exposures were further analyzed using bioinformatic tools to generate a differentially expressed genes (DEGs) dataset associated with hepatic lipid dysfunction. Further analysis of the DEGs dataset resulted in 324 upregulated genes, and 275 genes that were down regulated. Meanwhile, 36 genes were either up regulated or down regulated in different chemical treatments. All identified genes were uploaded in the web application for Database for Annotation, Visualization and Integrated Discovery (DAVID) for gene ontology enrichments and to identify Kyoto Encyclopedia of Genes and Genome (KEGG) pathways. Some of the identified pathways included glycolysis/gluconeogenesis, steroid hormone biosynthesis, retinol metabolism, and metabolism of xenobiotics by cytochrome P450. The same DEGs dataset was also analyzed using Ingenuity Pathway Analysis (IPA) software. IPA identified several pathways including PXR/RXR activation, Aryl hydrocarbon receptor signaling, and xenobiotic metabolism signaling. Furthermore, the generated DEGs lists were uploaded into NCATS BioPlanet platform. Some of the identified pathways were related to fatty acid omega oxidation, lipid and lipoprotein metabolism, and adipogenesis. The enrichment and clarification of the pathways and biological networks obtained from the DEGs dataset provide prior knowledge on the underlying biological key events and molecular mechanisms for the computational development of putative adverse outcome pathways (AOPs) for hepatic lipid dysfunction as a precursor to hepatic steatosis.

Application of Biologically Based Lumping To Investigate the Toxicokinetic Interactions of a Complex Gasoline Mixture.

People are often exposed to complex mixtures of environmental chemicals such as gasoline, tobacco smoke, water contaminants, or food additives. We developed an approach that applies chemical lumping methods to complex mixtures, in this case gasoline, based on biologically relevant parameters used in physiologically based pharmacokinetic (PBPK) modeling. Inhalation exposures were performed with rats to evaluate the performance of our PBPK model and chemical lumping method. There were 109 chemicals identified and quantified in the vapor in the chamber. The time-course toxicokinetic profiles of 10 target chemicals were also determined from blood samples collected during and following the in vivo experiments. A general PBPK model was used to compare the experimental data to the simulated values of blood concentration for 10 target chemicals with various numbers of lumps, iteratively increasing from 0 to 99. Large reductions in simulation error were gained by incorporating enzymatic chemical interactions, in comparison to simulating the individual chemicals separately. The error was further reduced by lumping the 99 nontarget chemicals. The same biologically based lumping approach can be used to simplify any complex mixture with tens, hundreds, or thousands of constituents.

Improving the risk assessment of lipophilic persistent environmental chemicals in breast milk.

Lipophilic persistent environmental chemicals (LPECs) have the potential to accumulate within a woman's body lipids over the course of many years prior to pregnancy, to partition into human milk, and to transfer to infants upon breastfeeding. As a result of this accumulation and partitioning, a breastfeeding infant's intake of these LPECs may be much greater than his/her mother's average daily exposure. Because the developmental period sets the stage for lifelong health, it is important to be able to accurately assess chemical exposures in early life. In many cases, current human health risk assessment methods do not account for differences between maternal and infant exposures to LPECs or for lifestage-specific effects of exposure to these chemicals. Because of their persistence and accumulation in body lipids and partitioning into breast milk, LPECs present unique challenges for each component of the human health risk assessment process, including hazard identification, dose-response assessment, and exposure assessment. Specific biological modeling approaches are available to support both dose-response and exposure assessment for lactational exposures to LPECs. Yet, lack of data limits the application of these approaches. The goal of this review is to outline the available approaches and to identify key issues that, if addressed, could improve efforts to apply these approaches to risk assessment of lactational exposure to these chemicals.

Use of novel inhalation kinetic studies to refine physiologically-based pharmacokinetic models for ethanol in non-pregnant and pregnant rats.

Ethanol (EtOH) exposure induces a variety of concentration-dependent neurological and developmental effects in the rat. Physiologically-based pharmacokinetic (PBPK) models have been used to predict the inhalation exposure concentrations necessary to produce blood EtOH concentrations (BEC) in the range associated with these effects. Previous laboratory reports often lacked sufficient detail to adequately simulate reported exposure scenarios associated with BECs in this range, or lacked data on the time-course of EtOH in target tissues (e.g. brain, liver, eye, fetus). To address these data gaps, inhalation studies were performed at 5000, 10 000, and 21 000 ppm (6 h/d) in non-pregnant female Long-Evans (LE) rats and at 21 000 ppm (6.33 h/d) for 12 d of gestation in pregnant LE rats to evaluate our previously published PBPK models at toxicologically-relevant blood and tissue concentrations. Additionally, nose-only and whole-body plethysmography studies were conducted to refine model descriptions of respiration and uptake within the respiratory tract. The resulting time-course and plethysmography data from these in vivo studies were compared to simulations from our previously published models, after which the models were recalibrated to improve descriptions of tissue dosimetry by accounting for dose-dependencies in pharmacokinetic behavior. Simulations using the recalibrated models reproduced these data from non-pregnant, pregnant, and fetal rats to within a factor of 2 or better across datasets, resulting in a suite of model structures suitable for simulation of a broad range of EtOH exposure scenarios.

Cross species extrapolation of the disruption of thyroid hormone synthesis by oxyfluorfen using in vitro data, physiologically based pharmacokinetic (PBPK), and thyroid hormone kinetics models.

The thyroid hormones play key roles in physiological processes such as regulation of the metabolic and cardiac systems as well as the development of the brain and surrounding sympathetic nervous system. Recent efforts to screen environmental chemicals for their ability to alter thyroid hormone synthesis, transport, metabolism and/or function have identified novel chemicals that target key processes in the thyroid pathway. One newly identified chemical, oxyfluorfen, is a diphenyl-ether herbicide used for control of annual broadleaf and grassy weeds in a variety of tree fruit, nut, vine, and field crops. Using in vitro high-throughput screening (HTS) assays, oxyfluorofen was identified to be a potent inhibitor of the thyroidal sodium-iodide symporter (NIS). To quantitatively assess this inhibition mechanism in vivo, we extrapolated in vitro NIS inhibition data to in vivo disruption of thyroid hormones synthesis in rats using physiologically based pharmacokinetic (PBPK) and thyroid hormone kinetics models. The overall computational model (chemical PBPK and THs kinetic sub-models) was calibrated against in vivo data for the levels of oxyfluorfen in thyroid tissue and serum and against serum levels of thyroid hormones triiodothyronine (T3) and thyroxine (T4) in rats. The rat thyroid model was then extrapolated to humans using human in vitro HTS data for NIS inhibition and the chemical specific hepatic clearance rate in humans. The overall species extrapolated PBPK-thyroid kinetics model can be used to predict dose-response (% drop in thyroid serum levels compared to homeostasis) relationships in humans. These relationships can be used to estimate points of departure for health risks related to a drop in serum levels of TH hormones based on HTS assays in vitro to in vivo extrapolation (IVIVE), toxicokinetics, and physiological principles.

Fate of arsenicals in mice carrying the human AS3MT gene exposed to environmentally relevant levels of arsenite in drinking water.

Although mice are widely used to study adverse effects of inorganic arsenic (iAs), higher rates of iAs methylation in mice than in humans may limit their utility as a model organism. A recently created 129S6 mouse strain in which the Borcs7/As3mt locus replaces the human BORCS7/AS3MT locus exhibits a human-like pattern of iAs metabolism. Here, we evaluate dosage dependency of iAs metabolism in humanized (Hs) mice. We determined tissue and urinary concentrations and proportions of iAs, methylarsenic (MAs), and dimethylarsenic (DMAs) in male and female Hs and wild-type (WT) mice that received 25- or 400-ppb iAs in drinking water. At both exposure levels, Hs mice excrete less total arsenic (tAs) in urine and retain more tAs in tissues than WT mice. Tissue tAs levels are higher in Hs females than in Hs males, particularly after exposure to 400-ppb iAs. Tissue and urinary fractions of tAs present as iAs and MAs are significantly greater in Hs mice than in WT mice. Notably, tissue tAs dosimetry in Hs mice resembles human tissue dosimetry predicted by a physiologically based pharmacokinetic model. These data provide additional support for use of Hs mice in laboratory studies examining effects of iAs exposure in target tissues or cells.

Computational estimation of errors generated by lumping of physiologically-based pharmacokinetic (PBPK) interaction models of inhaled complex chemical mixtures.

Many cases of environmental contamination result in concurrent or sequential exposure to more than one chemical. However, limitations of available resources make it unlikely that experimental toxicology will provide health risk information about all the possible mixtures to which humans or other species may be exposed. As such, utilizing computational models in order to make toxicological predictions is a useful tool in complementing experimental efforts which examine mixtures in health risk assessment. This paper outlines a novel mathematical method which reduces the complexity of a mixtures model and increases computational efficiency via a biologically-based lumping methodology (BBLM). In contrast to previous chemical lumping methodologies, BBLM allows the computation of error as a measure of the difference between the lumped simulation based on BBLM and the full mathematical model. As a consequence, the modeler has the opportunity to find the optimal configuration in the tradeoff between simplification and accuracy in order to determine an acceptable number and composition of lumped chemicals. To demonstrate this method, lumped equations based on a typical inhalation physiologically-based pharmacokinetic (PBPK) model assuming a competitive inhibition interaction mechanism are developed for a mixture of arbitrary size. The novel methodology is further tested using literature data for a mixture of 10 volatile organic chemicals (VOCs). Through simulation of these chemicals, BBLM is shown to produce good approximations when compared to the unlumped simulation and experimental data.

Development of multi-route physiologically-based pharmacokinetic models for ethanol in the adult, pregnant, and neonatal rat.

Biofuel blends of 10% ethanol (EtOH) and gasoline are common in the USA, and higher EtOH concentrations are being considered (15-85%). Currently, no physiologically-based pharmacokinetic (PBPK) models are available to describe the kinetics of EtOH-based biofuels. PBPK models were developed to describe life-stage differences in the kinetics of EtOH alone in adult, pregnant, and neonatal rats for inhalation, oral, and intravenous routes of exposure, using data available in the open literature. Whereas ample data exist from gavage and intravenous routes of exposure, kinetic data from inhalation exposures are limited, particularly at concentrations producing blood and target tissue concentrations associated with developmental neurotoxicity. Compared to available data, the three models reported in this paper accurately predicted the kinetics of EtOH, including the absorption, peak concentration, and clearance across multiple datasets. In general, model predictions for adult and pregnant animals matched inhalation and intravenous datasets better than gavage data. The adult model was initially better able to predict the time-course of blood concentrations than was the neonatal model. However, after accounting for age-related changes in gastric uptake using the calibrated neonate model, simulations consistently reproduced the early kinetic behavior in blood. This work provides comprehensive multi-route life-stage models of EtOH pharmacokinetics and represents a first step in development of models for use with gasoline-EtOH blends, with additional potential applicability in investigation of the pharmacokinetics of EtOH abuse, addiction, and toxicity.

A biologically based model to quantitatively assess the role of the nuclear receptors liver X (LXR), and pregnane X (PXR) on chemically induced hepatic steatosis.

Hepatic steatosis is characterized by the intracellular increase of free fatty acids (FFAs) in the form of triglycerides in hepatocytes. This hepatic adverse outcome can be caused by many factors, including exposure to drugs or environmental toxicants. Mechanistically, accumulation of lipids in the liver can take place via several mechanisms such as de novo synthesis and/or uptake of FFAs from serum via high fat content diets. De novo synthesis of FFAs within the liver is mediated by the liver X receptor (LXR), and their uptake into the liver is mediated through the pregnane X receptor (PXR). We investigated the impact of chemical exposure on FFAs hepatic content via activation of LXR and PXR by integrating chemical-specific physiologically based pharmacokinetic (PBPK) models with a quantitative toxicology systems (QTS) model of hepatic lipid homeostasis. Three known agonists of LXR and/or PXR were modeled: T0901317 (antagonist for both receptors), GW3965 (LXR only), and Rifampicin (PXR only). Model predictions showed that T0901317 caused the most FFAs build-up in the liver, followed by Rifampicin and then GW3965. These modeling results highlight the importance of PXR activation for serum FFAs uptake into the liver while suggesting that increased hepatic FAAs de novo synthesis alone may not be enough to cause appreciable accumulation of lipids in the liver under normal environmental exposure levels. Moreover, the overall PBPK-hepatic lipids quantitative model can be used to screen chemicals for their potential to cause in vivo hepatic lipid content buildup in view of their in vitro potential to activate the nuclear receptors and their exposure levels.

Mechanistic Computational Model for Extrapolating In Vitro Thyroid Peroxidase (TPO) Inhibition Data to Predict Serum Thyroid Hormone Levels in Rats.

High-throughput in vitro assays are developed to screen chemicals for their potential to inhibit thyroid hormones (THs) synthesis. Some of these experiments, such as the thyroid peroxidase (TPO) inhibition assay, are based on thyroid microsomal extracts. However, the regulation of thyroid disruption chemicals is based on THs in vivo serum levels. This necessitates the estimation of thyroid disruption chemicals in vivo tissue levels in the thyroid where THs synthesis inhibition by TPO takes place. The in vivo tissue levels of chemicals are controlled by pharmacokinetic determinants such as absorption, distribution, metabolism, and excretion, and can be described quantitatively in physiologically based pharmacokinetic (PBPK) models. An integrative computational model including chemical-specific PBPK and TH kinetics models provides a mechanistic quantitative approach to translate thyroidal high-throughput in vitro assays to in vivo measures of circulating THs serum levels. This computational framework is developed to quantitatively establish the linkage between applied dose, chemical thyroid tissue levels, thyroid TPO inhibition potential, and in vivo TH serum levels. Once this link is established quantitatively, the overall model is used to calibrate the TH kinetics parameters using experimental data for THs levels in thyroid tissue and serum for the 2 drugs, propylthiouracil and methimazole. The calibrated quantitative framework is then evaluated against literature data for the environmental chemical ethylenethiourea. The linkage of PBPK and TH kinetics models illustrates a computational framework that can be extrapolated to humans to screen chemicals based on their exposure levels and potential to disrupt serum THs levels in vivo.

Development of a quantitative model of pregnane X receptor (PXR) mediated xenobiotic metabolizing enzyme induction.

The pregnane X receptor plays an integral role in the regulation of hepatic metabolism. It has been shown to regulate CYP3A4, which is the most abundant cytochrome P450 in the human liver. With its large and flexible ligand-binding domain, PXR can be activated by an enormous range of relatively small, hydrophobic, exogenous compounds. Upon activation, PXR partners with the retinoid X receptor (RXR) to form a heterodimer. The newly formed heterodimer binds to an appropriate DNA response element, causing increased transcription. This leads to an induction in the level of CYP3A4. These mechanistic steps are included into a biologically-based mathematical model. The quantitative model predicts fold level inductions of CYP3A4 mRNA and protein in response to PXR activation. Model parameter values have been taken from literature when appropriate. Unknown parameter values are estimated by optimizing the model results to published in vivo and in vitro data sets. A sensitivity analysis is performed to evaluate the model structure and identify future data needs which would be critical to revising the model.

Extrapolating In Vitro Screening Assay Data for Thyroperoxidase Inhibition to Predict Serum Thyroid Hormones in the Rat.

Thyroperoxidase (TPO) is an enzyme essential for thyroid hormone (TH) synthesis and a target site for a number of xenobiotics that disrupt TH homeostasis. An in vitro high-throughput screening assay for TPO inhibition, the Amplex UltraRed-TPO (AUR-TPO), has been used to screen the ToxCast chemical libraries for this action. Output from this assay would be most useful if it could be readily translated into an in vivo response, namely a reduction of TH in serum. To this end, the relationship between TPO inhibition in vitro and serum TH decreases was examined in rats exposed to 2 classic TPO inhibitors, propylthiouracil (PTU) and methimazole (MMI). Serum and gland PTU, MMI, and TH levels were quantified using tandem liquid chromatography mass spectrometry. Thyroperoxidase activity was determined in thyroid gland microsomes treated with PTU or MMI in vitro and ex vivo from thyroid gland microsomes prepared from exposed animals. A quantitative model was constructed by contrasting in vitro and ex vivo AUR-TPO results and the in vivo time-course and dose-response analysis. In vitro:ex vivo correlations of AUR-TPO outputs indicated that less than 30% inhibition of TPO in vitro was sufficient to reduce serum T4 by 20%, a degree of regulatory significance. Although further testing of model estimates using other TPO inhibitors is essential for verification of these initial findings, the results of this study provide a means to translate in vitro screening assay results into predictions of in vivo serum T4 changes to inform risk assessment.

Gene Expression Thresholds Derived From Short-term Exposures Identify Rat Liver Tumorigens.

Traditional methods for cancer risk assessment are resource-intensive, retrospective, and not feasible for the vast majority of environmental chemicals. In this study, we investigated whether quantitative genomic data from short-term studies may be used to set protective thresholds for potential tumorigenic effects. We hypothesized that gene expression biomarkers measuring activation of the key early events in established pathways for rodent liver cancer exhibit cross-chemical thresholds for tumorigenesis predictive for liver cancer risk. We defined biomarker thresholds for 6 major liver cancer pathways using training sets of chemicals with short-term genomic data (3-29 days of exposure) from the TG-GATES (n = 77 chemicals) and DrugMatrix (n = 86 chemicals) databases and then tested these thresholds within and between datasets. The 6 pathway biomarkers represented genotoxicity, cytotoxicity, and activation of xenobiotic, steroid, and lipid receptors (aryl hydrocarbon receptor, constitutive activated receptor, estrogen receptor, and peroxisome proliferator-activated receptor α). Thresholds were calculated as the maximum values derived from exposures without detectable liver tumor outcomes. We identified clear response values that were consistent across training and test sets. Thresholds derived from the TG-GATES training set were highly predictive (97%) in a test set of independent chemicals, whereas thresholds derived from the DrugMatrix study were 96%-97% predictive for the TG-GATES study. Threshold values derived from an abridged gene list (2/biomarker) also exhibited high predictive accuracy (91%-94%). These findings support the idea that early genomic changes can be used to establish threshold estimates or "molecular tipping points" that are predictive of later-life health outcomes.

Arsenic Metabolism in Mice Carrying a BORCS7/AS3MT Locus Humanized by Syntenic Replacement.

BACKGROUND: Chronic exposure to inorganic arsenic (iAs) is a significant public health problem. Methylation of iAs by arsenic methyltransferase (AS3MT) controls iAs detoxification and modifies risks of iAs-induced diseases. Mechanisms underlying these diseases have been extensively studied using animal models. However, substantive differences between humans and laboratory animals in efficiency of iAs methylation have hindered the translational potential of the laboratory studies. OBJECTIVES: The goal of this study was to determine whether humanization of the As3mt gene confers a human-like pattern of iAs metabolism in mice. METHODS: We generated a mouse strain in which the As3mt gene along with the adjacent Borcs7 gene was humanized by syntenic replacement. We compared expression of the mouse As3mt and the human AS3MT and the rate and pattern of iAs metabolism in the wild-type and humanized mice. RESULTS: AS3MT expression in mouse tissues closely modeled that of human and differed substantially from expression of As3mt. Detoxification of iAs was much less efficient in the humanized mice than in wild-type mice. Profiles for iAs and its methylated metabolites in tissues and excreta of the humanized mice were consistent with those reported in humans. Notably, the humanized mice expressed both the full-length AS3MT that catalyzes iAs methylation and the human-specific AS3MTd2d3 splicing variant that has been linked to schizophrenia. CONCLUSIONS: These results suggest that AS3MT is the primary genetic locus responsible for the unique pattern of iAs metabolism in humans. Thus, the humanized mouse strain can be used to study the role of iAs methylation in the pathogenesis of iAs-induced diseases, as well as to evaluate the role of AS3MTd2d3 in schizophrenia. https://doi.org/10.1289/EHP6943.

Development of an inhalation physiologically based pharmacokinetic (PBPK) model for 2,2, 4-trimethylpentane (TMP) in male Long-Evans rats using gas uptake experiments.

2,2,4-Trimethylpentane (TMP) is a volatile colorless liquid used primarily to increase the octane rating of combustible fuels. TMP is released in the environment through the manufacture, use, and disposal of products associated with the gasoline and petroleum industry. Short-term inhalation exposure to TMP (< 4 h; > 1000 ppm) caused sensory and motor irritations in rats and mice. Like many volatile hydrocarbons, acute exposure to TMP may also be expected to alter neurological functions. To estimate in vivo metabolic kinetics of TMP and to predict its target tissue dosimetry during inhalation exposures, a physiologically based pharmacokinetic (PBPK) model was developed for the chemical in Long-Evans male rats using closed-chamber gas-uptake experiments. Gas-uptake experiments were conducted in which rats (80-90 days old) were exposed to targeted initial TMP concentrations of 50, 100, 500, and 1000 ppm. The model consisted of compartments for the closed uptake chamber, lung, fat, kidney, liver, brain, and rapidly and slowly perfused tissues. Physiological parameters were obtained from literature. Partition coefficients for the model were experimentally determined for air/blood, fat, liver, kidney, muscle, and brain using vial equilibration methods. Common to other hydrocarbons, metabolism of TMP via oxidative reactions is assumed to mainly occur in the liver. The PBPK model simulations of the closed chamber data were used to estimate in vivo metabolic parameters for TMP in male Long-Evans rats.

Advancing internal exposure and physiologically-based toxicokinetic modeling for 21st-century risk assessments.

Scientifically sound, risk-informed evaluation of chemicals is essential to protecting public health. Systematically leveraging information from exposure, toxicology, and epidemiology studies can provide a holistic understanding of how real-world exposure to chemicals may impact the health of populations, including sensitive and vulnerable individuals and life-stages. Increasingly, public health policy makers are employing toxicokinetic (TK) modeling tools to integrate these data streams and predict potential human health impact. Development of a suite of tools for predicting internal exposure, including physiologically-based toxicokinetic (PBTK) models, is being driven by needs to address large numbers of data-poor chemicals efficiently, translate bioactivity, and mechanistic information from new in vitro test systems, and integrate multiple lines of evidence to enable scientifically sound, risk-informed decisions. New modeling approaches are being designed "fit for purpose" to inform specific decision contexts, with applications ranging from rapid screening of hundreds of chemicals, to improved prediction of risks during sensitive stages of development. New data are being generated experimentally and computationally to support these models. Progress to meet the demand for internal exposure and PBTK modeling tools will require transparent publication of models and data to build credibility in results, as well as opportunities to partner with decision makers to evaluate and build confidence in use of these for improved decisions that promote safe use of chemicals.

How can biologically-based modeling of arsenic kinetics and dynamics inform the risk assessment process? - A workshop review.

Quantitative biologically-based models describing key events in the continuum from arsenic exposure to the development of adverse health effects provide a framework to integrate information obtained across diverse research areas. For example, genetic polymorphisms in arsenic metabolizing enzymes can lead to differences in target tissue dosimetry for key metabolites causative in toxic and carcinogenic response. This type of variation can be quantitatively incorporated into pharmacokinetic (PK) models and used together with population-based modeling approaches to evaluate the impact of genetic variation in methylation capacity on dose of key metabolites to target tissue. The PK model is an essential bridge to the pharmacodynamic (PD) models. A particular benefit of PD modeling for arsenic is that alternative models can be constructed for multiple proposed modes of action for arsenicals. Genomics data will prove useful for identifying the key pathways involved in particular responses and aid in determining other types of data needed for quantitative modeling. These models, when linked with PK models, can be used to better understand and explain dose- and time-response behaviors. This in turn assists in prioritizing modes of action with respect to their risk assessment relevance and future research. This type of integrated modeling approach can form the basis for a highly informative mode-of-action directed risk assessment for inorganic arsenic (iAs). This paper will address both practical and theoretical aspects of integrating PK and PD data in a modeling framework, including practical barriers to its application.

Development of a human physiologically based pharmacokinetic (PBPK) model for inorganic arsenic and its mono- and di-methylated metabolites.

A physiologically-based pharmacokinetic (PBPK) model was developed to estimate levels of arsenic and its metabolites in human tissues and urine after oral exposure to arsenate (As(V)), arsenite (As(III)) or organoarsenical pesticides. The model consists of interconnected individual PBPK models for inorganic arsenic (As(V) and As(III)), monomethylarsenic acid (MMA(V)), and, dimethylarsenic acid (DMA(V)). Reduction of MMA(V) and DMA(V) to their respective trivalent forms also occurs in the lung, liver, and kidney including excretion in urine. Each submodel was constructed using flow limited compartments describing the mass balance of the chemicals in GI tract (lumen and tissue), lung, liver, kidney, muscle, skin, heart, and brain. The choice of tissues was based on physiochemical properties of the arsenicals (solubility), exposure routes, target tissues, and sites for metabolism. Metabolism of inorganic arsenic in liver was described as a series of reduction and oxidative methylation steps incorporating the inhibitory influence of metabolites on methylation. The inhibitory effects of As(III) on the methylation of MMA(III) to DMA, and MMA(III) on the methylation of As(III) to MMA were modeled as noncompetitive. To avoid the uncertainty inherent in estimation of many parameters from limited human data, a priori independent parameter estimates were derived using data from diverse experimental systems with priority given to data derived using human cells and tissues. This allowed the limited data for human excretion of arsenicals in urine to be used to estimate only parameters that were most sensitive to this type of data. Recently published urinary excretion data, not previously used in model development, are also used to evaluate model predictions.