Elsevier

Science of The Total Environment

Volume 578, 1 February 2017, Pages 1-15
Science of The Total Environment

Review
Toxicokinetic models and related tools in environmental risk assessment of chemicals

https://doi.org/10.1016/j.scitotenv.2016.10.146Get rights and content

Highlights

  • State-of-the-art of toxicokinetic models in environmental risk assessment

  • Empirical compartment models were more developed than mechanistic models.

  • Models were mostly developed for aquatic species, in particular for fish.

  • Models were applied to a diversity of organic and inorganic compounds classes.

  • Further research needs: developing mechanistic models integrating confounding factors.

Abstract

Environmental risk assessment of chemicals for the protection of ecosystems integrity is a key regulatory and scientific research field which is undergoing constant development in modelling approaches and harmonisation with human risk assessment. This review focuses on state-of-the-art toxicokinetic tools and models that have been applied to terrestrial and aquatic species relevant to environmental risk assessment of chemicals. Both empirical and mechanistic toxicokinetic models are discussed using the results of extensive literature searches together with tools and software for their calibration and an overview of applications in environmental risk assessment. These include simple tools such as one-compartment models, multi-compartment models to physiologically-based toxicokinetic (PBTK) models, mostly available for aquatic species such as fish species and a number of chemical classes including plant protection products, metals, persistent organic pollutants, nanoparticles. Data gaps and further research needs are highlighted.

Introduction

Environmental risk assessment (ERA) of chemicals is a well-established regulatory and scientific research field which is undergoing constant development including harmonisation with the human risk assessment (HRA) area (Pery et al., 2013, EFSA, 2015). In this context, mechanistic approaches for risk assessment purposes offer the opportunity to bring together ecological and human dimensions particularly while taking into account taxa-specific traits resulting from ecological, physiological and biochemical features for a quantitative mechanistic understanding of toxicity (EFSA, 2014, EFSA, 2016).

Over the last decade, these approaches have been applied to ERA through a number of in vivo and in vitro models and methodologies, as well as in silico models to investigate the mode of action and adverse outcome pathways of chemicals at different levels of biological organisation (van de Waterbeemd and Gifford, 2003, Raunio, 2011, EFSA, 2014). In the same fashion as for the HRA field, the integration of mechanistic approaches in ERA can describe the processes linking exposure or environmental concentrations (external dose) to internal dose, up to toxicity in a given taxa or the whole ecosystem. Due to the complexity of those processes, the relationship between effective dose and external exposure is not necessarily linear. Two phases can be distinguished in the dose-response relationship of a chemical in an organism: the toxicokinetics (TK), or pharmacokinetics (PK) for pharmaceuticals, that is the fate of the compound in the organism including absorption, distribution, metabolism and excretion (ADME), and the toxicodynamics (TD), or pharmacodynamics (PD) for pharmaceuticals, that is the expression of the toxicity/effect of the substance. Schematically, TK can be viewed as the action of the body on the compound “what the body does to the chemical”, and TD as the action of the compound on the body “what the chemical does to the body” (Spurgeon et al., 2010). Depending on the chemical and its mode of action (TK and TD), the appropriate internal dosimetry can be related to the levels of the active substance/toxic moiety (parent or metabolite) that may reach the target tissues, i.e. the biologically effective dose (Escher and Hermens, 2004), or the quantity of metabolites produced by xenobiotic-metabolising enzymes through bioactivation (e.g. cytochrome P-450), detoxification reactions (e.g. glucuronidation) or renal excretion (Brochot et al., 2007, Dorne, 2010). TK models can be used in ERA and HRA to perform various extrapolations of external dose to internal (effective) dose: extrapolations between exposure levels, exposure routes, individuals or between species (IPCS, 2010). TK models can be used to estimate external exposures or internal dose from indirect biomarker (e.g., metabolites) measurements (EPA, 2006).

The simplest approaches to take into account TK are based on steady-state assumptions which reflect the equilibrium between accumulation of a compound and its elimination from all exposure routes i.e., both the exposure and the environmental/physiological factors affecting the uptake and the elimination of the chemical remain constant. The bioconcentration factor (BCF) is based on this assumption and is defined as the ratio of the chemical concentration in the organism over the chemical concentration in water or any other biological media at steady-state (Landrum et al., 1992). Other steady-state approaches, such as simple empirical statistical relationships (reviewed in Landrum et al. (1992)), or quantitative structure-activity relationships (QSAR) models have also been proposed and applied to the prediction of internal concentrations or TK descriptors (e.g. BCF, reviewed in Pavan et al. (2006)).

In some cases, the steady-state assumption may not hold particularly in environmental conditions under which chemical occurrence and exposure may vary considerably. For example, Ashauer et al. (2006a) review models used to predict effects of pesticides on aquatic organisms from pulses or fluctuating exposure. Historically, one of the first approach to consider time as an explicit variable in risk assessment is through the Haber's rule (for reviews see Bunce and Remillard (2003), Miller et al. (2000) and Rozman and Doull (2000)). Currently, the most common approach to describe the TK of a compound is to represent the organism as a system of compartments. A TK model aims to describe the time course of blood or tissue concentrations resulting from the ADME processes of the chemical. TK models fall generally into two classes: empirical models and models based on the physiology of the organism considered. The first TK/PK modelling was based on physiological descriptions of the human body in Teorell (1937). However, at the time, such mathematical models were too complex to be solved. Later on, research and applications then focused on simpler one-, two- or three-compartment models that were not necessarily based on physiology, but which proved to be adequate for describing and interpolating concentration-time profiles of many compounds in the blood or in other biological matrices (Tarr et al., 1990, Ritterhoff and Zauke, 1997). However, simple models such as one compartmental TK models are insufficient for substances with complex kinetics, or when inter-species extrapolations are required and research is pursued to further develop physiological models (Nichols et al., 1990, Pery et al., 2014, Brinkmann et al., 2016). These models are now recognised and recommended by national and international scientific advisory bodies particularly for evaluation of toxicological properties and chemical risk assessment in general (EPA, 2006, IPCS, 2010). Moreover, approaches based on in vitro testing and PBTK models to evaluate the toxicity of compounds were developed. Such combinatorial approaches are very promising to investigate interspecies differences in exposure patterns and toxicological sensitivity to chemicals in relation to taxa specific traits influencing physiology, level of expression of xenobiotic metabolising enzymes, and life cycle (EFSA, 2016).

This review provides a state-of-the-art review of mechanistic TK models applied to ERA, i.e. risk identification and dose-response analysis or exposure quantification, for relevant terrestrial and aquatic species using the results of extensive literature searches (ELS). Both empirical (data-based) and physiologically based TK (PBTK) models are presented with particular attention to the tools available for their parameterisation, and an overview of applications. Data gaps and further research needs are highlighted.

Section snippets

Empirical compartment TK models

Compartmental models describe toxicant movements between compartments which may or may not have a physiologic or anatomic meaning. The toxicant enters into a compartment according to an uptake rate coefficient and exits according to an elimination rate coefficient (Gibaldi and Perrier, 1982). Two sub-groups of TK compartmental models can be distinguished: (i) models based on a one-compartment assumption, according to which the chemical concentration is the same throughout the organism; (ii)

Description of models

A physiologically based toxicokinetic (PBTK) model is defined as “a model that estimates the dose to target tissue by taking into account the rate of absorption into the body, distribution and storage in tissues, metabolism and excretion on the basis of interplay among critical physiological, physicochemical and biochemical determinants” (U.S. EPA, 2006). A PBTK (or PBPK for physiologically based pharmacokinetic) model subdivides the body in compartments representing real tissues or organs

Tools for the calibration of PBTK models

A PBTK model requires a large amount of information for its parameterisation. Two subsets of parameters can be identified: (i) substance-independent parameters that represent anatomical (e.g., organ volume), physiological (e.g., cardiac output) and biochemical entities (e.g., enzyme concentrations); and (ii) substance-specific parameters which reflect the specific interactions between the body and the substance of interest. Substance-independent parameters can be, in theory, directly measured

Available software to build PBTK models

Many software can actually be used to build a PBTK model and run simulations (Schmitt and Willmann, 2005, Bouzom et al., 2012). Software can be subdivided into two types: (i) general simulation platforms which are not designed for PBTK modelling and require the user to write and code the model equations, e.g. R CRAN, GNU MCSim, Matlab® and Mathematica®; and (ii) “ready to use” tools which have been developed expressly for PBTK modelling, e.g. Simcyp® and MERLIN-Expo for humans. However, in that

Conclusions, data gaps and perspectives

This review has highlighted that a great diversity of TK tools and models have been developed to take into account TK processes for ERA applications e.g., empirical one- and multi-compartment models and mechanistic physiologically based toxicokinetic models. These models have been mostly developed for aquatic species and in particular for fish species, and have been applied to a diversity of compound classes (e.g., metals, persistent organic compounds, biocides, nanoparticles…). Whatever the

Funding information

This work was supported by the European Food Safety Authority (EFSA) and the French Ministry of Ecology and Sustainable Development [Contract number: EFSA/SCER/2014/06 and Program 190, respectively].

References (216)

  • C. Brochot et al.

    Development of a physiologically based toxicokinetic model for butadiene and four major metabolites in humans: global sensitivity analysis for experimental design issues

    Chem. Biol. Interact.

    (2007)
  • J.K. Carlson et al.

    Respiratory and hematological responses of the bonnethead shark, Sphyrna tiburo, to acute changes in dissolved oxygen

    J. Exp. Mar. Biol. Ecol.

    (2003)
  • J. Chang et al.

    Effects of butachlor on reproduction and hormone levels in adult zebrafish (Danio rerio)

    Exp. Toxicol. Pathol.

    (2013)
  • M. Connolly et al.

    Tissue distribution of zinc and subtle oxidative stress effects after dietary administration of ZnO nanoparticles to rainbow trout

    Sci. Total Environ.

    (2016)
  • C.E. Cowan-Ellsberry et al.

    Approach for extrapolating in vitro metabolism data to refine bioconcentration factor estimates

    Chemosphere

    (2008)
  • E. Curtis et al.

    Application of various mathematical models to data from the uptake of methyl mercury in bluegill sunfish (Lepomis macrochirus)

    Ecol. Model.

    (1977)
  • J.W. Deneer

    Bioconcentration of chlorpyrifos by the three-spined stickleback under laboratory and field conditions

    Chemosphere

    (1994)
  • M. Diez-Ortiz et al.

    Influence of soil properties on molybdenum uptake and elimination kinetics in the earthworm Eisenia andrei

    Chemosphere

    (2010)
  • J.L. Dorne

    Metabolism, variability and risk assessment

    Toxicology

    (2010)
  • K.G. Drouillard et al.

    Spatial and temporal variability of PCBs in Detroit River water assessed using a long term biomonitoring program

    Chemosphere

    (2013)
  • R.J. Erickson et al.

    A model for exchange of organic chemicals at fish gills: flow and diffusion limitations

    Aquat. Toxicol.

    (1990)
  • P.N. Fitzsimmons et al.

    Branchial elimination of superhydrophobic organic compounds by rainbow trout (Oncorhynchus mykiss)

    Aquat. Toxicol.

    (2001)
  • P.N. Fitzsimmons et al.

    A compilation of in vitro rate and affinity values for xenobiotic biotransformation in fish, measured under physiological conditions

    Comp. Biochem. Physiol., Part C: Toxicol. Pharmacol.

    (2007)
  • K.L. Forsgren et al.

    The effects of the pyrethroid insecticide, bifenthrin, on steroid hormone levels and gonadal development of steelhead (Oncorhynchus mykiss) under hypersaline conditions

    Gen. Comp. Endocrinol.

    (2013)
  • A. Franco-Uria et al.

    Generic parameterization for a pharmacokinetic model to predict Cd concentrations in several tissues of different fish species

    Chemosphere

    (2010)
  • Y. Gao et al.

    Prediction of acute toxicity of cadmium and lead to zebrafish larvae by using a refined toxicokinetic-toxicodynamic model

    Aquat. Toxicol.

    (2015)
  • J. Garnier-Laplace et al.

    A simple fish physiological model for radioecologists exemplified for Mn-54 direct transfer and rainbow trout (Oncorhynchus mykiss W.)

    J. Environ. Radioact.

    (2000)
  • F. Gobas et al.

    Dynamics of dietary bioaccumulation and faecal elimination of hydrophobic organic chemicals in fish

    Chemosphere

    (1988)
  • F. Gobas et al.

    Bioaccumulation of chlorinated hydrocarbons by the mayfly (Hexagenia limbata) in Lake St. Clair

    J. Great Lakes Res.

    (1989)
  • G. Gorge et al.

    Kinetics and metabolism of 14C-lindane and 14C-atrazine in early life stages of zebrafish (Brachydanio rerio)

    Chemosphere

    (1990)
  • G.M. Hughes

    General anatomy of the gills

    Fish Physiology

    (1984)
  • O.A. Alvarez et al.

    Physiological modes of action of toxic chemicals in the nematode Acrobeloides nanus

    Environ. Toxicol. Chem.

    (2006)
  • R. Aravindhan et al.

    Biosorption of cadmium metal ion from simulated wastewaters using Hypnea valentiae biomass: a kinetic and thermodynamic study

    Bioresour. Technol.

    (2010)
  • J.A. Arnot

    Mass balance models for chemical fate, bioaccumulation, exposure and risk assessment

  • R. Ashauer et al.

    Advantages of toxicokinetic and toxicodynamic modelling in aquatic ecotoxicology and risk assessment

    J. Environ. Monit.

    (2010)
  • R. Ashauer et al.

    Predicting effects on aquatic organisms from fluctuating or pulsed exposure to pesticides

    Environ. Toxicol. Chem.

    (2006)
  • R. Ashauer et al.

    Uptake and elimination of chlorpyrifos and pentachlorophenol into the freshwater amphipod Gammarus pulex

    Arch. Environ. Contam. Toxicol.

    (2006)
  • R. Ashauer et al.

    Toxicokinetic-toxicodynamic modeling of quantal and graded sublethal endpoints: a brief discussion of concepts

    Environ. Toxicol. Chem.

    (2011)
  • R. Ashauer et al.

    Significance of xenobiotic metabolism for bioaccumulation kinetics of organic chemicals in Gammarus pulex

    Environ. Sci. Technol.

    (2012)
  • R. Ashauer et al.

    A method to predict and understand fish survival under dynamic chemical stress using standard ecotoxicity data

    Environ. Toxicol. Chem.

    (2013)
  • Baas, J., M. Vijver, J. Rambohul, M. Dunbar, M. van 't Zelfde, C. Svendsen, and D. Spurgeon, 2016. Comparison and...
  • M.C. Barber

    A review and comparison of models for predicting dynamic chemical bioconcentration in fish

    Environ. Toxicol. Chem.

    (2003)
  • M.G. Barron et al.

    Temperature-dependence of cardiac output and regional blood flow in rainbow trout, Salmo gairdneri Richardson

    J. Fish Biol.

    (1987)
  • M.G. Barron et al.

    Pharmacokinetics of sulphadimethoxine in the lobster, Homarus americanus, following intrapericardial administration

    Xenobiotica

    (1988)
  • M.G. Barron et al.

    Pharmacokinetic modeling in aquatic animals. 1. Models and concepts

    Aquat. Toxicol.

    (1990)
  • M.G. Barron et al.

    Tissue carboxylesterase activity of rainbow trout

    Environ. Toxicol. Chem.

    (1999)
  • A. Barsi et al.

    Considerations for test design to accommodate energy-budget models in ecotoxicology: a case study for acetone in the pond snail Lymnaea stagnalis

    Environ. Toxicol. Chem.

    (2014)
  • D. Baudiffier et al.

    Effect of in vivo chronic exposure to clotrimazole on zebrafish testis function

    Environ. Sci. Pollut. Res. Int.

    (2013)
  • T. Baussant et al.

    Bioaccumulation of polycyclic aromatic compounds: 2. Modeling, bioaccumulation in marine organisms chronically exposed to dispersed oil

    Environ. Toxicol. Chem.

    (2001)
  • R. Beaudouin et al.

    An individual-based model of zebrafish population dynamics accounting for energy dynamics

    PLoS One

    (2015)
  • Cited by (91)

    View all citing articles on Scopus
    View full text