Bayesian estimation of pharmacokinetic and pharmacodynamic parameters in a mode-of-action-based cancer risk assessment for chloroform

PBPK modeling to evaluate maximum tolerated doses: A case study with 3-chloroallyl alcohol

Introduction: A physiologically based pharmacokinetic (PBPK) model for 3-chloroallyl alcohol (3-CAA) was developed and used to evaluate the design of assays for the in vivo genotoxicity of 3-CAA. Methods: Model development was supported by read across from a published PBPK model for ethanol. Read across was motivated by the expectation that 3-CAA, which like ethanol is a primary alcohol, is metabolized largely by hepatic alcohol dehydrogenases. The PBPK model was used to evaluate how two metrics of tissue dosimetry, maximum blood concentration (Cmax; mg/L) and area under the curve (AUC; mg-hr/L) vary with dose of 3-CAA and with dose route (oral gavage, drinking water). Results: The model predicted that oral gavage results in a 6-fold higher Cmax than the same dose administered in drinking water, but in similar AUCs. Predicted Cmax provided the best correlation with severe toxicity (e.g., lethality) from 3-CAA, consistent with the production of a reactive metabolite. Therefore, drinking water administration can achieve higher sustained concentration without severe toxicity in vivo. Discussion: This evaluation is significant because cytotoxicity is a potential confounder of mutagenicity testing. The PBPK model can be used to ensure that studies meet OECD and USEPA test guidelines and that the highest dose used is not associated with severe toxicity. In addition, PBPK modeling provides assurance of target tissue (e.g., bone marrow) exposure even in the absence of laboratory data, by defining the relationship between applied dose and target tissue dose based on accepted principles of pharmacokinetics, relevant physiology and biochemistry of the dosed animals, and chemical-specific information.

High-throughput PBTK models for in vitro to in vivo extrapolation

INTRODUCTION

Toxicity data are unavailable for many thousands of chemicals in commerce and the environment. Therefore, risk assessors need to rapidly screen these chemicals for potential risk to public health. High-throughput screening (HTS) for in vitro bioactivity, when used with high-throughput toxicokinetic (HTTK) data and models, allows characterization of these thousands of chemicals.

AREAS COVERED

This review covers generic physiologically based toxicokinetic (PBTK) models and high-throughput PBTK modeling for in vitro-in vivo extrapolation (IVIVE) of HTS data. We focus on 'httk', a public, open-source set of computational modeling tools and in vitro toxicokinetic (TK) data.

EXPERT OPINION

HTTK benefits chemical risk assessors with its ability to support rapid chemical screening/prioritization, perform IVIVE, and provide provisional TK modeling for large numbers of chemicals using only limited chemical-specific data. Although generic TK model design can increase prediction uncertainty, these models provide offsetting benefits by increasing model implementation accuracy. Also, public distribution of the models and data enhances reproducibility. For the httk package, the modular and open-source design can enable the tool to be used and continuously improved by a broad user community in support of the critical need for high-throughput chemical prioritization and rapid dose estimation to facilitate rapid hazard assessments.

Using quantitative modeling tools to assess pharmacokinetic bias in epidemiological studies showing associations between biomarkers and health outcomes at low exposures

Biomarkers of exposure can be measured at lower and lower levels due to advances in analytical chemistry. Using these sensitive methods, some epidemiology studies report associations between biomarkers and health outcomes at biomarker levels much below those associated with effects in animal studies. While some of these low exposure associations may arise from increased sensitivity of humans compared with animals or from species-specific responses, toxicology studies with drugs, commodity chemicals and consumer products have not generally indicated significantly greater sensitivity of humans compared with test animals for most health outcomes. In some cases, these associations may be indicative of pharmacokinetic (PK) bias, i.e., a situation where a confounding factor or the health outcome itself alters pharmacokinetic processes affecting biomarker levels. Quantitative assessment of PK bias combines PK modeling and statistical methods describing outcomes across large numbers of individuals in simulated populations. Here, we first provide background on the types of PK models that can be used for assessing biomarker levels in human population and then outline a process for considering PK bias in studies intended to assess associations between biomarkers and health outcomes at low levels of exposure. After providing this background, we work through published examples where these PK methods have been applied with several chemicals/chemical classes - polychlorinated biphenyls (PCBs), perfluoroalkyl substances (PFAS), polybrominated biphenyl ethers (PBDE) and phthalates - to assess the possibility of PK bias. Studies of the health effects of low levels of exposure will be improved by developing some confidence that PK bias did not play significant roles in the observed associations.

Bayesian population physiologically-based pharmacokinetic model for robustness evaluation of withdrawal time in tilapia aquaculture administrated to florfenicol

The antimicrobial residues of aquacultural production is a growing public concern, leading to reexamine the method for establishing robust withdrawal time and ensuring food safety. Our study aims to develop the optimizing population physiologically-based pharmacokinetic (PBPK) model for assessing florfenicol residues in the tilapia tissues, and for evaluating the robustness of the withdrawal time (WT). Fitting with published pharmacokinetic profiles that experimented under temperatures of 22 and 28 °C, a PBPK model was constructed by applying with the Bayesian Markov chain Monte Carol (MCMC) algorithm to estimate WTs under different physiological, environmental and dosing scenarios. Results show that the MCMC algorithm improves the estimates of uncertainty and variability of PBPK-related parameters, and optimizes the simulation of the PBPK model. It is noteworthy that posterior sets generated from temperature-associated datasets to be respectively used for simulating residues under corresponding temperature conditions. Simulating the residues under regulated regimen and overdosing scenarios for Taiwan, the estimated WTs were 12-16 days at 22 °C and 9-12 days at 28 °C, while for the USA, the estimated WTs were 14-18 and 11-14 days, respectively. Comparison with the regulated WT of 15 days, results indicate that the current WT has well robustness and resilience in the environment of higher temperatures. The optimal Bayesian population PBPK model provides effective analysis for determining WTs under scenario-specific conditions. It is a new insight into the increasing body of literature on developing the Bayesian-PBPK model and has practical implications for improving the regulation of food safety.

Monte-Carlo and multi-exposure assessment for the derivation of criteria for disinfection byproducts and volatile organic compounds in drinking water: Allocation factors and liter-equivalents per day

The probability distributions of total potential doses of disinfection byproducts and volatile organic compounds via ingestion, inhalation, and dermal exposure were estimated with Monte Carlo simulations, after conducting physiologically based pharmacokinetic model simulations to takes into account the differences in availability between the three exposures. If the criterion that the 95th percentile estimate equals the TDI (tolerable daily intake) is regarded as protecting the majority of a population, the drinking water criteria would be 140 (trichloromethane), 66 (bromodichloromethane), 157 (dibromochloromethane), 203 (tribromomethane), 140 (dichloroacetic acid), 78 (trichloroacetic acid), 6.55 (trichloroethylene, TCE), and 22 μg/L (perchloroethylene). The TCE criterion was lower than the Japanese Drinking Water Quality Standard (10 μg/L). The latter would allow the intake of 20% of the population to exceed the TDI. Indirect inhalation via evaporation from water, especially in bathrooms, was the major route of exposure to compounds other than haloacetic acids (HAAs) and accounted for 1.2-9 liter-equivalents/day for the median-exposure subpopulation. The ingestion of food was a major indirect route of exposure to HAAs. Contributions of direct water intake were not very different for trihalomethanes (30-45% of TDIs) and HAAs (45-52% of TDIs).

Evaluation and calibration of high-throughput predictions of chemical distribution to tissues

Toxicokinetics (TK) provides critical information for integrating chemical toxicity and exposure assessments in order to determine potential chemical risk (i.e., the margin between toxic doses and plausible exposures). For thousands of chemicals that are present in our environment, in vivo TK data are lacking. The publicly available R package "httk" (version 1.8, named for "high throughput TK") draws from a database of in vitro data and physico-chemical properties in order to run physiologically-based TK (PBTK) models for 553 compounds. The PBTK model parameters include tissue:plasma partition coefficients (Kp) which the httk software predicts using the model of Schmitt (Toxicol In Vitro 22 (2):457-467, 2008). In this paper we evaluated and modified httk predictions, and quantified confidence using in vivo literature data. We used 964 rat Kp measured by in vivo experiments for 143 compounds. Initially, predicted Kp were significantly larger than measured Kp for many lipophilic compounds (log10 octanol:water partition coefficient > 3). Hence the approach for predicting Kp was revised to account for possible deficiencies in the in vitro protein binding assay, and the method for predicting membrane affinity was revised. These changes yielded improvements ranging from a factor of 10 to nearly a factor of 10,000 for 83 Kp across 23 compounds with only 3 Kp worsening by more than a factor of 10. The vast majority (92%) of Kp were predicted within a factor of 10 of the measured value (overall root mean squared error of 0.59 on log10-transformed scale). After applying the adjustments, regressions were performed to calibrate and evaluate the predictions for 12 tissues. Predictions for some tissues (e.g., spleen, bone, gut, lung) were observed to be better than predictions for other tissues (e.g., skin, brain, fat), indicating that confidence in the application of in silico tools to predict chemical partitioning varies depending upon the tissues involved. Our calibrated model was then evaluated using a second data set of human in vivo measurements of volume of distribution (Vss) for 498 compounds reviewed by Obach et al. (Drug Metab Dispos 36(7):1385-1405, 2008). We found that calibration of the model improved performance: a regression of the measured values as a function of the predictions has a slope of 1.03, intercept of - 0.04, and R2 of 0.43. Through careful evaluation of predictive methods for chemical partitioning into tissues, we have improved and calibrated these methods and quantified confidence for TK predictions in humans and rats.

Optimization of alginate purification using polyvinylidene difluoride membrane filtration: Effects on immunogenicity and biocompatibility of three-dimensional alginate scaffolds

Sodium alginate is an effective biomaterial for tissue engineering applications. Non-purified alginate is contaminated with protein, lipopolysaccharide, DNA, and RNA, which could elicit adverse immunological reactions. We developed a purification protocol to generate biocompatible alginate based on (a) activated charcoal treatment, (b) use of hydrophobic membrane filtration (we used hydrophobic polyvinylidene difluoride membranes to remove organic contaminants), (c) dialysis, and finally (d) ethanol precipitation. Using this approach, we could omit pre-treatment with chloroform and significantly reduce the quantities of reagents used. Purification resulted in reduction of residual protein by 70% down to 0.315 mg/g, DNA by 62% down to 1.28 µg/g, and RNA by 61% down to less than 10 µg/g, respectively. Lipopolysaccharide levels were reduced by >90% to less than 125 EU/g. Purified alginate did not induce splenocyte proliferation in vitro. Three-dimensional scaffolds generated from purified alginate did not elicit a significant foreign body reaction, fibrotic overgrowth, or macrophage infiltration 4 weeks after implantation. This study describes a simplified and economical alginate purification method that results in alginate purity, which meets clinically useful criteria.

Characterization of the human kinetic adjustment factor for the health risk assessment of environmental contaminants

A default uncertainty factor of 3.16 (√10) is applied to account for interindividual variability in toxicokinetics when performing non-cancer risk assessments. Using relevant human data for specific chemicals, as WHO/IPCS suggests, it is possible to evaluate, and replace when appropriate, this default factor by quantifying chemical-specific adjustment factors for interindividual variability in toxicokinetics (also referred to as the human kinetic adjustment factor, HKAF). The HKAF has been determined based on the distributions of pharmacokinetic parameters (e.g., half-life, area under the curve, maximum blood concentration) in relevant populations. This article focuses on the current state of knowledge of the use of physiologically based algorithms and models in characterizing the HKAF for environmental contaminants. The recent modeling efforts on the computation of HKAF as a function of the characteristics of the population, chemical and its mode of action (dose metrics), as well as exposure scenario of relevance to the assessment are reviewed here. The results of these studies, taken together, suggest the HKAF varies as a function of the sensitive subpopulation and dose metrics of interest, exposure conditions considered (route, duration, and intensity), metabolic pathways involved and theoretical model underlying its computation. The HKAF seldom exceeded the default value of 3.16, except in very young children (i.e., <≈ 3 months) and when the parent compound is the toxic moiety. Overall, from a public health perspective, the current state of knowledge generally suggest that the default uncertainty factor is sufficient to account for human variability in non-cancer risk assessments of environmental contaminants.

Application of an updated physiologically based pharmacokinetic model for chloroform to evaluate CYP2E1-mediated renal toxicity in rats and mice

Physiologically based pharmacokinetic (PBPK) models are tools for interpreting toxicological data and extrapolating observations across species and route of exposure. Chloroform (CHCl(3)) is a chemical for which there are PBPK models available in different species and multiple sites of toxicity. Because chloroform induces toxic effects in the liver and kidneys via production of reactive metabolites, proper characterization of metabolism in these tissues is essential for risk assessment. Although hepatic metabolism of chloroform is adequately described by these models, there is higher uncertainty for renal metabolism due to a lack of species-specific data and direct measurements of renal metabolism. Furthermore, models typically fail to account for regional differences in metabolic capacity within the kidney. Mischaracterization of renal metabolism may have a negligible effect on systemic chloroform levels, but it is anticipated to have a significant impact on the estimated site-specific production of reactive metabolites. In this article, rate parameters for chloroform metabolism in the kidney are revised for rats, mice, and humans. New in vitro data were collected in mice and humans for this purpose and are presented here. The revised PBPK model is used to interpret data of chloroform-induced kidney toxicity in rats and mice exposed via inhalation and drinking water. Benchmark dose (BMD) modeling is used to characterize the dose-response relationship of kidney toxicity markers as a function of PBPK-derived internal kidney dose. Applying the PBPK model, it was also possible to characterize the dose response for a recent data set of rats exposed via multiple routes simultaneously. Consistent BMD modeling results were observed regardless of species or route of exposure.

A Bayesian population PBPK model for multiroute chloroform exposure

A Bayesian hierarchical model was developed to estimate the parameters in a physiologically based pharmacokinetic (PBPK) model for chloroform using prior information and biomarker data from different exposure pathways. In particular, the model provides a quantitative description of the changes in physiological parameters associated with hot-water bath and showering scenarios. Through Bayesian inference, uncertainties in the PBPK parameters were reduced from the prior distributions. Prediction of biomarker data with the calibrated PBPK model was improved by the calibration. The posterior results indicate that blood flow rates varied under two different exposure scenarios, with a two-fold increase of the skin's blood flow rate predicted in the hot-bath scenario. This result highlights the importance of considering scenario-specific parameters in PBPK modeling. To demonstrate the application of a probability approach in toxicological assessment, results from the posterior distributions from this calibrated model were used to predict target tissue dose based on the rate of chloroform metabolized in liver. This study demonstrates the use of the Bayesian approach to optimize PBPK model parameters for typical household exposure scenarios.

Development of a quantitative model incorporating key events in a hepatotoxic mode of action to predict tumor incidence

Biologically based dose-response (BBDR) modeling of environmental pollutants can be utilized to inform the mode of action (MOA) by which compounds elicit adverse health effects. Chemicals that produce tumors are typically labeled as either genotoxic or nongenotoxic. Though both the genotoxic and the nongenotoxic MOA may be operative as a function of dose, it is important to note that the label informs but does not define a MOA. One commonly proposed MOA for nongenotoxic carcinogens is characterized by the key events cytotoxicity and regenerative proliferation. The increased division rate associated with such proliferation can cause an increase in the probability of mutations, which may result in tumor formation. We included these steps in a generalized computational pharmacodynamic (PD) model incorporating cytotoxicity as a MOA for three carcinogens (chloroform, CHCl(3); carbon tetrachloride, CCL(4); and N,N-dimethylformamide, DMF). For each compound, the BBDR model is composed of a chemical-specific physiologically based pharmacokinetic model linked to a PD model of cytotoxicity and cellular proliferation. The rate of proliferation is then linked to a clonal growth model to predict tumor incidences. Comparisons of the BBDR simulations and parameterizations across chemicals suggested that significant variation among the models for the three chemicals arises in a few parameters expected to be chemical specific (such as metabolism and cellular injury rate constants). Optimization of model parameters to tumor data for CCL(4) and DMF resulted in similar estimates for all parameters related to cytotoxicity and tumor incidences. However, optimization of the CHCl(3) data resulted in a higher estimate for one parameter (BD) related to death of initiated cells. This implies that additional steps beyond cytotoxicity leading to induced cellular proliferation can be quantitatively different among chemicals that share cytotoxicity as a hypothesized carcinogenic MOA.

Mode of action considerations in the quantitative assessment of tumour responses in the liver

Chemical carcinogenesis is a complex, multi-stage process and the relationship between dose and tumour formation is an important consideration in the risk assessment of chemicals. Extrapolation from empirical dose-response relationships obtained in experimental studies has been criticized, as it fails to take into account information on mode of action. Strategies for incorporating mode of action information into the risk assessment of chemical carcinogens are described, with a focus on hepatic cancer. Either toxicokinetic or toxicodynamic processes can be addressed. Whilst the former have been the focus of more attention to date, for example by using physiologically based modelling, there is increasing interest in the development of mode of action-based toxicodynamic models. These have the advantage that they do not require extreme assumptions, and may be amenable to paramaterization using human data. This is rarely if ever possible when using conventional dose-tumour response relationships. The approaches discussed are illustrated using chloroform as a case study. This compound is converted to a cytotoxic metabolite, phosgene, by CYP2E1 in liver and/or kidney. Cytotoxicity results in proliferative regeneration, with increased probability of tumour formation. Both physiologically based toxicokinetic and toxicodynamic models have been developed, and it is possible to use probabilistic approaches incorporating, for example, data on the distribution of hepatic CYP2E1 levels. Mode of action can provide an invaluable link between observable, experimental data, on both toxicokinetics and toxicodynamics, and chemical-specific risk assessment, based on physiological approaches.

Approaches for assessing risks to sensitive populations: lessons learned from evaluating risks in the pediatric population

Assessing the risk profiles of potentially sensitive populations requires a "tool chest" of methodological approaches to adequately characterize and evaluate these populations. At present, there is an extensive body of literature on methodologies that apply to the evaluation of the pediatric population. The Health and Environmental Sciences Institute Subcommittee on Risk Assessment of Sensitive Populations evaluated key references in the area of pediatric risk to identify a spectrum of methodological approaches. These approaches are considered in this article for their potential to be extrapolated for the identification and assessment of other sensitive populations. Recommendations as to future research needs and/or alternate methodological considerations are also made.

Application of key events analysis to chemical carcinogens and noncarcinogens

The existence of thresholds for toxicants is a matter of debate in chemical risk assessment and regulation. Current risk assessment methods are based on the assumption that, in the absence of sufficient data, carcinogenesis does not have a threshold, while noncarcinogenic endpoints are assumed to be thresholded. Advances in our fundamental understanding of the events that underlie toxicity are providing opportunities to address these assumptions about thresholds. A key events dose-response analytic framework was used to evaluate three aspects of toxicity. The first section illustrates how a fundamental understanding of the mode of action for the hepatic toxicity and the hepatocarcinogenicity of chloroform in rodents can replace the assumption of low-dose linearity. The second section describes how advances in our understanding of the molecular aspects of carcinogenesis allow us to consider the critical steps in genotoxic carcinogenesis in a key events framework. The third section deals with the case of endocrine disrupters, where the most significant question regarding thresholds is the possible additivity to an endogenous background of hormonal activity. Each of the examples suggests that current assumptions about thresholds can be refined. Understanding inter-individual variability in the events involved in toxicological effects may enable a true population threshold(s) to be identified.