- Reference Pages
- Search Agencies
- Search Topics
Tox21 is a collaboration among groups within four U.S. federal organizations aimed at developing more efficient approaches to predict how chemicals may affect human health. Tox21 studies use assays that are run at higher throughput than traditional tests. The goal of Tox21 is to use data from these assays to prioritize substances for further evaluation, inform understanding of mechanisms of action, and/or develop improved predictive models for toxicity. Test approaches developed and data collected via this initiative may enable agencies to reduce reliance on animal data for assessing chemical safety. Tox21 projects and projects using Tox21 data are described below and throughout this report.
The four groups participating in the Tox21 collaboration are ICCVAM members:
Transcriptomics uses a cell’s overall gene expression to assess many aspects of biology in the cell, including its normal function and response to toxicity. A question of interest is whether the cell types used in high-throughput transcriptomics assays need to reflect human biological diversity to identify different classes of toxicants and clarify the relevant biology for toxicity testing. To evaluate this question, EPA and NIEHS collaboratively used transcriptomics databases and other resources to identify cell lines that maximized biological diversity at the level of gene expression. Using a newer high-throughput transcriptomics technology, gene expression will be assessed in these cells under normal conditions and with chemical treatment. Ultimately, comparisons will identify selected cells for high-throughput transcriptomics chemical screening and also enlighten how future cell lines should be chosen. An abstract providing an update was accepted for presentation at the April 2020 meeting of the Midsouth Computational Biology and Bioinformatics Society.
IVIVE relates chemical concentrations that induce a response in an in vitro assay to chemical exposures that induce relevant effects in vivo. IVIVE typically assumes that chemicals behave in cells in an in vitro system in the same way they behave in blood and tissue in animals. Significant differences between in vitro models and in vivo systems make this assumption inaccurate. While the nature and extent of these differences are not well characterized, it is known that a chemical’s physicochemical properties affect factors such as binding to plastic and partitioning between medium and cells. EPA and NIEHS scientists selected a group of chemicals from the Tox21 library representing a diversity of structural and physicochemical properties and are examining their in vitro disposition to better understand these factors. Work since 2018 has focused on a pilot set of 10-20 chemicals, with anticipated expansion to 100-200 chemicals in 2020. This anticipated work will represent the largest undertaking of empirical measures of in vitro disposition and inform model predictions of in vitro disposition for other chemicals beyond those tested.
High-throughput transcriptomics generates gene expression profiles to rapidly evaluate the effects of large numbers of chemicals on in vitro cell culture systems (Harrill et al. 2019). To provide a basis for characterizing the toxicity potential of chemicals with limited or no available data, scientists at NIEHS and EPA are building a common reference chemical dataset to enhance interpretation of high-throughput transcriptomics screening data. The project systematically identified a robust set of reference chemicals with direct interactions to specific biological targets (e.g., nuclear receptors, enzymes, kinases, ion channels). A subset of approximately 300 of these reference chemicals has been acquired by Tox21 chemistry collaborators for evaluation in two human cell culture models: MCF-7 cells, derived from breast cells, and HepaRG, derived from liver cells (Ramaiahgari et al. 2019). The next stage of the project is to create the reference chemical dataset, analyze both gene-level and pathway-level responses that enable improved interpretation of transcriptomic data with test chemicals, and identify the most efficient conditions to expand coverage to thousands of reference chemicals.
FDA and NICEATM scientists applied IVIVE to evaluate the impact of pharmacokinetics and different modeling approaches on predicting relevant external exposure from in vitro developmental toxicity potential concentrations derived from an in vitro human iPSC-based assay. Previous work showed that the devTOX quick Predict assay ranked the developmental toxicity potential of valproate analogues in a manner that was consistent with observed developmental toxicity potency in vivo. The IVIVE analysis in this project estimated equivalent administered doses that would result in maternal blood concentrations equivalent to the developmental toxicity potential and cytotoxic in vitro concentrations. The estimated equivalent administered doses were compared to published lowest effect levels from in vivo developmental toxicity studies. Preliminary results of this analysis showed close agreement between equivalent administered doses and in vivo rat lowest effect levels for two valproate analogues. This suggested that the devTOX quick Predict assay and IVIVE approaches can quantitatively predict in vivo developmental toxicity potential. An abstract describing this work (Chang et al.) was accepted for a presentation at the Society of Toxicology 2020 annual meeting.
The potential for neurotoxicity in children following exposure to environmental chemicals remains a high public priority due to concerns about recent increases in the prevalence of neurological disorders such as attention deficit hyperactivity disorder and autism. Neurotoxicity risk for an individual depends on a number of factors, including interactions between an individual’s variation in genetic makeup and exposures to neurotoxic chemicals in the environment. To investigate the role of genetic diversity in susceptibility to neurotoxicity, scientists at NIEHS, EPA, and FDA are using a genetically diverse set of cells to evaluate a curated set of chemicals with neurotoxic potential. Neural progenitor cells were derived from a set of mice bred to maximize genetic diversity, yielding 200 male and female genetically different cell lines. The panel of cell lines will be exposed to varying concentrations of the chemical test set and assessed using a high-content imaging assay called cell painting. The compiled dataset will be used to identify chemicals with a range of developmental neurotoxicity potencies. These data will inform data-driven uncertainty factors that account for interindividual variability, allowing for adequate protection of genetically sensitive subpopulations.
To use data generated by HTS initiatives such as Tox21 and the EPA ToxCast program in regulatory applications, the assays and models built from the assays must be validated based on their performance against the biological targets they query. This requires developing sets of reference chemicals that consistently yield reproducible results when assayed for these biological targets. A process also needs to be established for assessing and reporting the performance evaluation of an assay in a standardized format that provides consumers of the data the ability to interpret the appropriate context for use of the assay. Finally, both the development of reference chemical sets and the validation process need to be streamlined and rapid enough to manage the tens to hundreds of assays that can help inform regulatory toxicity endpoints. To address these needs, scientists at EPA and NIEHS developed a process to identify reference chemicals that consistently produce positive or negative results when assayed in defined assays (Judson et al 2019). Additional work under this project has focused on identifying chemicals that consistently produce false signals by interfering with specific technology types and backgrounds. Models to predict interferent chemicals for luciferase inhibition and autofluorescence are available on the NTP website. Current efforts are focused on using these data to identify reference chemicals and establish protocols for evaluating the performance of specific Tox21 and ToxCast assays.
The HTS assays that have been run in the Tox21 testing program to date generally lack the metabolic activity found in living systems, which can potentially increase or decrease the toxicity of chemicals. As a result, HTS results may not accurately reflect in vivo activity. Scientists at EPA, NCATS, and NIEHS are using several approaches to address this problem: adding human or rat liver microsomes into the existing assays, transfecting cells with mRNAs encoding human metabolic enzymes, or using metabolically capable human HepaRG cells. The addition of metabolic capacity to HTS assays is expected to improve characterization of the in vivo activity of chemicals in the Tox21 collection. Current efforts focus on retrofitting three types of assays for which a massive amount of data have already been generated: cellular stress-related assays, endocrine disruption assays, and CYP450 enzyme inhibition assays. The retrofitted assays are being used to screen the Tox21 10K chemical library to identify chemicals that are either bioactivated or detoxified by metabolic activity.
To date, Tox21 HTS assays have focused primarily on selected nuclear receptor and stress response pathways. This relatively limited focus suggests that activity in other toxicity pathways has not been adequately assessed; it is likely that some unexplored pathways relate to unanticipated adverse drug effects. Therefore, expanding the coverage of biological responses by adding assays that probe under-represented pathways in the current Tox21 assay portfolio may improve the predictivity of Tox21 data. Scientists at FDA, NIEHS, and NCATS are systematically identifying these under-represented pathways in a data-driven approach and nominating assays for development and Tox21 chemical screening. The data generated (Huang et al. 2018) will be used to build models for human toxicity prediction, focusing on common adverse drug effects such as drug-induced liver injury and cardiotoxicity. An initial panel of targets and pathways has been identified (Huang et al. 2019) using existing drug-target annotations and adverse effect information obtained from the literature and public databases, such as DrugBank. In parallel, human toxicity data are being collected and curated from the literature. These data will be used to target additional cellular pathways for future assay development and validation.
Acetylcholinesterase inhibitors cause a variety of adverse effects in the nervous system. Some acetylcholinesterase inhibitors serve as drugs, while others are used as pesticides or found in natural products. Scientists at FDA and NCATS developed acetylcholinesterase inhibition assays (Li et al. 2017) and screened the Tox21 10K chemical library to identify environmental and drug- or food-related chemicals that inhibit acetylcholinesterase activity. The screening study also provided an opportunity to evaluate performance characteristics of HTS assays intended to measure acetylcholinesterase inhibition. An accomplishment of the study included incorporating metabolism into the assays (Li et al. 2019), an important consideration as some acetylcholinesterase inhibitors become more potent when metabolized. Selected inhibitors were further characterized using stem cells and computational models to gain insights on their inhibitory mechanisms.