Regulatory Policy Implications of 1,4-Dioxane - Environmental Investigation and Remediation - page 467

metabolites of xenobiotic chemicals and interfacing that information with toxic effect models that
predict chemical binding to estrogen receptors in aquatic species can be used to anticipate the
potential for endocrine disruption in aquatic organisms (USEPA, 2006).
Among the most challenging aspects in conducting risk assessments is relating the effects
detected at the dose level to which the test animal was subjected to the effects that would be caused
by the dose that actually reaches the target organ in humans. Toxicokinetic studies can predict the
internal target organ dose, which must account for the absorption, distribution, metabolism, and
excretion (ADME) of the toxin of interest. PBPK models can be used to predict ADME in vivo by
integrating literature values and computational techniques, and by extrapolating data from in vitro
studies between species. Improved dose-response models of in vivo toxicity can then be developed
from the test animal data (Bhogal et al., 2005).
The work of developing a PBPK model involves tracking down lots of data to estimate the
required input parameters; consequently, PBPK modeling is considered “data hungry” and “resource
intensive,” and thus requires a certain i nancial threshold before a PBPK modeling study can be
launched. Efforts are underway in Britain to make PBPK modeling more efi cient, less expensive,
and more accessible by developing a PBPK model equation generator and a PBPK parameter data-
base, which allow the construction of PBPK models in minutes rather than days (Health and Safety
Laboratories [HSL], 2008). The PBPK parameter database contains physicochemical, biochemical,
anatomical, and physiological data that have been validated for their quality. Parametric data are
retained together with metadata describing the studies that generated the data and the quality of the
parameter data. Developing the key PBPK model equations and populating the model parameters is
then a semiautomatic task, allowing researchers to focus their skill and attention on the toxicologi-
cal validity of the modeled relationships rather than on the laborious and time-consuming task of
collecting and compiling data (HSL, 2008).
Another computational toxicology initiative to accelerate the prediction of chemicals' toxic
effects is USEPA's ToxCast™, which was launched in 2007 to develop rapid and cost-effective
tools for prioritizing the toxicity testing of large numbers of chemicals. Using data from high-
throughput screening in vitro bioassays, the ToxCast™ project is building computational models
to forecast the potential toxicity of chemicals to humans. ToxCast™ is expected to provide
USEPA regulatory programs with science-based information that are helpful in prioritizing
chemicals for more detailed toxicological evaluations lead to more efi cient use of animal testing
(USEPA, 2007).
Predicting carcinogenicity is another toxicological challenge for which computational toxicol-
ogy has made substantial progress using a method called Computer Automated Structure
Evaluation (CASE). The CASE system uses QSAR in the same manner as described in Section
3.1. Molecules of interest are divided into chemical fragments of two to ten heavy atoms, and a
statistical distribution is performed to determine whether the fragments identii ed are capable of
the specii c biological activity necessary to induce toxic effects. The CASE system has developed
predictions of carcinogenicity and mutagenicity endpoints, using a variety of database modules
including Ames mutagenicity and carcinogenicity in male and female rats and mice (Bhogal
et al., 2005).
Overall, the combined application of advanced biotechnology tools to develop cell lines for
toxicological assays, leveraging advances in computer technology and database systems, and
improvements to assembling PBPK models all signify that the quality of toxicological assessments
is likely to continue to grow better. These changes will completely transform the way in which
toxicological data are generated and applied. Taken together, these advances may comprise more
than a paradigm shift; rather, environmental toxicology is probably undergoing a quantum leap.
With sufi cient funding to enable research scientists to focus on advancing the state of toxicology,
the traditional laboratory animal approach to toxicity testing could soon be consigned to history.
Advanced in vitro and in silico methods will become the routine approaches to assess environmen-
tal toxicology (Bhogal et al., 2005).