8
R. Kusko and H. Hong
In the space of drug development, log regression analysis has predicted druginduced liver injury, which has proven challenging for both the pharmaceutical industry and regulators. Chapter 13 briefed the drug-induced liver injury (DILI) research
efforts at the National Center for Toxicological Research (NCTR), FDA, including
drug-label-based-approach to annotate the DILI risk associated with individual drugs
including a series of models developed to assess the potential of DILI risk.
Alternative methods including computational toxicology have been considered to
inform regulation of drugs, foods, and environmental chemicals. Spanning all three of
these fields, a collaborative project across US governmental agencies known as Tox21
screened 10 k chemicals against a large panel of cell-based assays in a quantitative
high-throughput screen [37]. Chapter 14 described the efforts to build in vivo toxicity
prediction models based on the Tox21 in vitro activity profiles of compounds and
discussed the limitations of the current data and strategies for selection of optimal
assays to improve the performance of the developed models. The Tox21 project
served as powerful fuel for computational predictive modeling across many projects
and institutions including predicting point of departure (POD). Chapter 15 reviewed
common data modeling approaches that use gene expression profiles to estimate the
PODs and compared with the PODs determined using Tox21 data.
From an environmental perspective, endocrine disrupting chemicals (EDCs) are of
grave concern and the MOA has been effectively detailed by target-based molecular
modeling methods. Computational toxicology methods are an essential and powerful tool to elucidate the MOA of endocrine disruptors. Chapter 16 reviewed the
critical processes to perform the molecular modeling of EDCs, including preparation
of three-dimensional (3D) structures of the biomacromolecules and EDCs, generation and optimization of the structures of EDC–biomacromolecule complexes, and
investigation of the underlying interaction mechanism.
The metabolism of xenobiotics by cytochrome P450 enzymes (CYPs) represents an important mechanism for in vivo compound processing via environmental
exposure. Density functional theory (DFT) calculations have been used to highlight
the underpinnings of the mechanisms of various environmental toxicants by CYPs
including brominated flame retardants. Chapter 17 reviewed the recent progress in
molecular simulations of xenobiotic metabolism catalyzed by the typical phase I
enzyme CYPs.
Computational toxicology methods including QSAR and read-across are gaining
acceptance in regulatory science in the USA, Europe, and Japan [38]. To facilitate the
applications of computational toxicology in regulatory science, tools for utilization
of QSAR models and read-across have been developed. Chapter 18 introduced a tool
(VEGA) that was designed to reduce the barriers between the different read-across
and QSAR models for the evaluation of specific chemicals for the assessment of
populations of substances. VEGA provides multiple tools for different purposes.
Rigorous and reproducible in silico workflows are needed for toxicological
databases and analysis to be successful. OpenTox is stepping in to fill this gap. OpenTox advocates the establishment of good practice and guidance for tracking computational toxicology models to enhance reproducibility, a very important parameter
for acceptance of the computational models in regulatory science. Chapter 19 dis-
R. Kusko and H. Hong
In the space of drug development, log regression analysis has predicted druginduced liver injury, which has proven challenging for both the pharmaceutical industry and regulators. Chapter 13 briefed the drug-induced liver injury (DILI) research
efforts at the National Center for Toxicological Research (NCTR), FDA, including
drug-label-based-approach to annotate the DILI risk associated with individual drugs
including a series of models developed to assess the potential of DILI risk.
Alternative methods including computational toxicology have been considered to
inform regulation of drugs, foods, and environmental chemicals. Spanning all three of
these fields, a collaborative project across US governmental agencies known as Tox21
screened 10 k chemicals against a large panel of cell-based assays in a quantitative
high-throughput screen [37]. Chapter 14 described the efforts to build in vivo toxicity
prediction models based on the Tox21 in vitro activity profiles of compounds and
discussed the limitations of the current data and strategies for selection of optimal
assays to improve the performance of the developed models. The Tox21 project
served as powerful fuel for computational predictive modeling across many projects
and institutions including predicting point of departure (POD). Chapter 15 reviewed
common data modeling approaches that use gene expression profiles to estimate the
PODs and compared with the PODs determined using Tox21 data.
From an environmental perspective, endocrine disrupting chemicals (EDCs) are of
grave concern and the MOA has been effectively detailed by target-based molecular
modeling methods. Computational toxicology methods are an essential and powerful tool to elucidate the MOA of endocrine disruptors. Chapter 16 reviewed the
critical processes to perform the molecular modeling of EDCs, including preparation
of three-dimensional (3D) structures of the biomacromolecules and EDCs, generation and optimization of the structures of EDC–biomacromolecule complexes, and
investigation of the underlying interaction mechanism.
The metabolism of xenobiotics by cytochrome P450 enzymes (CYPs) represents an important mechanism for in vivo compound processing via environmental
exposure. Density functional theory (DFT) calculations have been used to highlight
the underpinnings of the mechanisms of various environmental toxicants by CYPs
including brominated flame retardants. Chapter 17 reviewed the recent progress in
molecular simulations of xenobiotic metabolism catalyzed by the typical phase I
enzyme CYPs.
Computational toxicology methods including QSAR and read-across are gaining
acceptance in regulatory science in the USA, Europe, and Japan [38]. To facilitate the
applications of computational toxicology in regulatory science, tools for utilization
of QSAR models and read-across have been developed. Chapter 18 introduced a tool
(VEGA) that was designed to reduce the barriers between the different read-across
and QSAR models for the evaluation of specific chemicals for the assessment of
populations of substances. VEGA provides multiple tools for different purposes.
Rigorous and reproducible in silico workflows are needed for toxicological
databases and analysis to be successful. OpenTox is stepping in to fill this gap. OpenTox advocates the establishment of good practice and guidance for tracking computational toxicology models to enhance reproducibility, a very important parameter
for acceptance of the computational models in regulatory science. Chapter 19 dis-