Page 126 - The Toxicology of Fishes

106 The Toxicology of Fishes

agents (Gerlowski and Jain, 1983); subsequently, this approach was adopted by toxicologists interested
in extrapolating data from laboratory test animals to humans. Much of the early work in this field was
focused on modeling volatile organic compounds (Andersen, 1981; Ramsey and Andersen, 1984).
Additional models have been developed for a variety of nonvolatile toxicants, including several metals
(Medinsky and Klaassen, 1996). Increasingly, PBTK models are being linked with biologically based
dose–response (BBDR) models to provide a complete (toxicokinetic and toxicodynamic) toxicological
description (Connolly et al., 1988).
One early application of the PBTK modeling approach was to compare chemical distribution and
elimination in fish and rodents. Models for methotrexate in stingrays (Zaharko et al., 1972) and phenol
red in the dogfish shark (Bungay et al., 1976) were designed to simulate experiments in which compounds
were injected intravascularly, and elimination was limited to urinary and fecal routes. The first PBTK
model for fish that incorporated an environmentally relevant route of exposure (branchial) was developed
by Nichols et al. (1990). Later models have incorporated both dermal and dietary routes of exposure
(Nichols et al., 1996, 1998, 2004a). In this section, we review the principles of PBTK modeling with
fish. Mathematical details pertaining to mass-balance tissue descriptions are deemphasized, as this
information is available in several reviews (Gerlowski and Jain, 1983; Krishnan and Andersen, 2001;
Rowland, 1985). This section focuses instead on route of exposure considerations as they relate to fish.
An emphasis is placed on factors that distinguish fish and mammalian models, either in terms of model
structure or the ways in which the models are used. In several instances, empirical data are reviewed to
define the chemical kinetic behavior that successful models must simulate.

Model Structure
A PBTK model consists of several tissue compartments, connected in a manner that is consistent with
the cardiovascular system of the exposed organism (Figure 3.22). Additional details follow from the
type of exposure, characteristics of the test compound, and goals of the modeling exercise. Often, it is
necessary to ignore known or suspected complexity to simplify the task of parameter estimation. In
general, a modeler strives to create the simplest possible structure that will satisfactorily represent the
behavior of the system under study. The number and identity of tissue compartments can vary greatly.
An emphasis is usually placed on tissues that control the kinetics of uptake and elimination, as well as
those that are important toxicologically.
Physiological and anatomical attributes of fish that influence chemical uptake and disposition were
reviewed earlier in this chapter. Several attributes are singled out here because they tend to distinguish
fish and mammalian PBTK models. First, chemical uptake by inhalation exposure can be adequately
simulated in mammals by modeling the lungs as a mixed chemical reactor. Chemicals taken up across
the lung travel first to the heart and then to the rest of the body. In contrast, the gills of many fish species
are more appropriately modeled as a counter-current exchange system, and chemicals taken up across
the gills travel throughout the body before reaching the heart. Second, the kidney in many fish is supplied
by venous blood draining both the trunk musculature and skin. Although this portal blood does not
appear to be available for glomerular filtration, it may contribute to urinary elimination of compounds
that are actively secreted in the proximal tubule. Because fish are poikilothermic, physiological and
metabolic parameters such as ventilation volume and cardiac output can vary greatly with ambient
temperature, thereby having a large impact on chemical kinetics. Finally, it is important to recognize
that important physiological differences exist among fish species because of differences in life history
and reproductive strategy, dietary preferences, and the need to adapt to different environmental conditions.
Incorporation of these differences into any given model depends on whether they influence the kinetics
of the chemical.
The compartments in a PBTK model correspond to tissues and organs with similar kinetic character-
istics. For a well-mixed compartment that receives only arterial blood, a mass-balance equation may be
written as:

dX dt = ( C vi) (3.83)
Q C a −
i
i