Page 143 - The Toxicology of Fishes

Toxicokinetics in Fishes 123

The principal advantage of a PBTK model is that it provides descriptions of the chemical concentration
time course in specific tissues of interest. This provides a direct link to studies of toxic effect and in
particular to observations for a specific site of action. As an example, numerous studies have suggested
that TCDD is highly toxic to fish in early life stages. In wild fish, this exposure occurs following maternal
transfer of accumulated residues to the developing ovaries. A PBTK model for the maternal transfer of
TCDD was therefore developed to support studies of TCDD embryotoxicity in brook trout (Nichols et
al., 1998). A second important use of PBTK models is to evaluate competing assumptions about factors
that control chemical uptake and disposition. Nichols et al. (2004a,b), for example, used a PBTK model
to investigate factors that control dietary uptake of hydrophobic organic compounds. Based on these
studies, it was concluded that a log K -dependent kinetic limitation prevents the gut tissues and contents
ow
from attaining an internal equilibrium. The nature of this limitation remains unknown, but it does not
appear to be related to diffusion across the gastrointestinal epithelium. PBTK models can also be used
to estimate important kinetic parameters that may be difficult or impossible to determine otherwise.
Several examples of this approach appear in the preceding text, including the use of models to solve for
biliary elimination rate constants (Bungay et al., 1976; Zaharko et al., 1972), dermal permeability
constants (Nichols et al., 1996), and metabolic rate constants (Law et al., 1991).
The utility of PBTK models for fish was demonstrated particularly well by a linked toxicokinetic and
toxicodynamic model for paraoxon in rainbow trout (Abbas and Hayton, 1997). Paraoxon is produced
in mammals by oxidative metabolism of the insecticide parathion and is a potent inhibitor of acetylcho-
linesterase (AChE) and carboxylesterase (CaE). The activation of parathion to paraoxon is thought to
be insignificant in trout; however, paraoxon can be formed in aquatic environments by nonenzymatic
conversion of parathion and may be available for uptake by fish directly from water. The toxicokinetic
portion of this model accurately simulated the uptake of paraoxon from water and its distribution to
selected tissues. This information was then combined with experimentally determined rates of AChE
and CaE synthesis and degradation, as well as biomolecular inhibition rate constants determined in
previous studies with rodents. The linked model successfully reproduced the observed time course of
AChE inhibition in each of the tissues examined and confirmed the role of CaE in detoxification of
paraoxon by sequestration of active compound (Figure 3.30).
Another use of PBTK models is to evaluate the potential for variability among individuals of a single
species to impact chemical uptake and disposition. A simple approach to this question was provided by
Lien et al. (2001), who simulated the kinetics of three chlorinated ethanes in lake trout using physiological
data from individual animals. Simulations for each animal were then plotted to represent the range of
anticipated outcomes. Alternatively, statistical distributions for individual model inputs can be character-
ized and used in a repeated sampling (Monte Carlo) design to generate a distribution of predicted outcomes.
The principal advantage of the former approach is that it explicitly treats the possible interdependence of
model parameters. The Monte Carlo approach can be adapted to deal with parameter interdependence.
In practice, however, this is difficult because the necessary information is generally lacking.
Mammalian PBTK modeling efforts are largely driven by the needs of human health risk assessment
and in particular by the need to extrapolate data from laboratory test animals to humans. In this context,
it is necessary to consider factors such as aging, health status, and dietary habits that could result in
increased vulnerability among a subset of the human population. For fish and other ecological receptors,
the principal driving force behind PBTK model development is the need to identify species that, for
toxicokinetic or other reasons, may be particularly vulnerable. Simultaneous exposure by multiple routes
is common in the environment. PBTK modeling is useful in this regard as it enables modeling of multiple
chemical fluxes at all relevant exchange surfaces.
The principal disadvantage of using a PBTK modeling approach is that considerable effort is required
to develop models for new species and chemicals. Future applications of the PBTK modeling approach
will depend on the systematic collection of necessary biological and chemical information, as well as
the development of methods for estimating critical parameters when data for a compound and species
are limited or absent altogether. Quantitative structure–activity relationship (QSAR) approaches hold
particular promise as a means of obtaining tissue partitioning estimates, and perhaps also metabolic rate
predictions (McKim and Nichols, 1994; Parham and Portier, 1998; Parham et al., 1997; Poulin and
Krishnan, 1996b; Verhaar et al., 1997).