Predictive models and computational toxicology

Pluripotent stem cell assays: Modalities and applications for predictive developmental toxicity

This manuscript provides a review focused on embryonic stem cell-based models and their place within the landscape of alternative developmental toxicity assays. Against the background of the principles of developmental toxicology, the wide diversity of alternative methods using pluripotent stem cells developed in this area over the past half century is reviewed. In order to provide an overview of available models, a systematic scoping review was conducted following a published protocol with inclusion criteria, which were applied to select the assays. Critical aspects including biological domain, readout endpoint, availability of standardized protocols, chemical domain, reproducibility and predictive power of each assay are described in detail, in order to review the applicability and limitations of the platform in general and progress moving forward to implementation. The horizon of innovative routes of promoting regulatory implementation of alternative methods is scanned, and recommendations for further work are given.

Sub-lethal Camphor Exposure Triggers Oxidative Stress, Cardiotoxicity, and Cardiac Physiology Alterations in Zebrafish Embryos

Camphor is a terpene ketone with aromatic and volatile properties in nature derived from the bark of Cinnamomum camphora or synthesized from turpentine. Camphor exhibits various biological properties such as anti-microbial, anti-viral, anti-coccidial, and anti-cancer. It is also used as a form of topical medication for skin irritation, joint pain, and as a relief for itching from insect bites. However, even though the high dose of camphor has been documented to be toxic/lethal in humans in different studies, camphor's developmental toxicity has not yet been explored, and its extensive mechanism of action is still unclear. In the present study, we aimed to assess the toxic effects of camphor in zebrafish embryos in the initial developmental stages. The obtained results demonstrated that a sub-lethal dose of camphor caused a decrease in hatching rate, body length, and substantial elevation in malformation rate on zebrafish embryos. On further observation, in the following time frame, curved body and pericardial edema of zebrafish were also observed. Furthermore, exposure to a sub-lethal dose of camphor was also able to trigger cardiotoxicity in zebrafish larvae. Later, on subsequent biochemical analysis, it was found that the antioxidant capacity inhibition and oxidative stress elevation that occurred after camphor exposure might be associated with the inhibition of total superoxide dismutase (SOD) activity and an increase in reactive oxygen species (ROS) and malondialdehyde (MDA) concentration. In addition, compared to the control group, several apoptotic cells in treated zebrafish were also found to be elevated. Finally, after further investigation on marker gene expressions, we conclude that the developmental toxicity of camphor exposure might be associated with apoptosis elevation and oxidative stress. Taken together, the current study provides a better understanding of the developmental toxicity of camphor on zebrafish, a promising alternative animal model to assess the developmental toxicity of chemical compounds.

Genipin induces developmental toxicity through oxidative stress and apoptosis in zebrafish

Genipin, an iridoid substance, is mainly derived from Gardenia jasminoides Ellis of the traditional Chinese medicine and is widely used in raw materials for the food additive gardenia blue and biological materials. The developmental toxicity of genipin has not been investigated, and its underlying mechanism is unclear. Therefore, in this study we attempt to investigate the potential developmental toxicity of genipin in zebrafish embryos/larvae. The results showed zebrafish embryos treated with 50 μg/ml dose of genipin display inhibited hatching rates and body length. The pericardial edema was observed. It was also found that genipin could induce cardio-toxicity, hepatotoxicity and nephrotoxicity in zebrafish larvae. After genipin treatment, the suppression of antioxidant capacity and increase of oxidative stress were showed for the triggered generation of ROS and MDA, and decreased activity of SOD. Compared with the 0.5% DMSO group, a number of apoptotic cells in zebrafish were increased after genipin exposure. By measuring marker gene expression with the using of qRT-PCR, we proposed that developmental toxicity after genipin treatment might be associated with oxidative stress and apoptosis increase. Our research offers a better understanding for developmental toxicity of genipin.

Agrochemicals with estrogenic endocrine disrupting properties: Lessons Learned?

Many agrochemicals have endocrine disrupting properties. A subset of these chemicals is characterized as "estrogenic". In this review, we describe several distinct ways that chemicals used in crop production can affect estrogen signaling. Using three agrochemicals as examples (DDT, endosulfan, and atrazine), we illustrate how screening tests such as the US EPA's EDSP Tier 1 assays can be used as a first-pass approach to evaluate agrochemicals for endocrine activity. We then apply the "Key Characteristics" approach to illustrate how chemicals like DDT can be evaluated, together with the World Health Organization's definition of an endocrine disruptor, to identify data gaps. We conclude by describing important issues that must be addressed in the evaluation and regulation of hormonally active agrochemicals including mixture effects, efforts to reduce vertebrate animal use, chemical prioritization, and improvements in hazard, exposure, and risk assessments.

An ontology for developmental processes and toxicities of neural tube closure

In recent years, the development and implementation of animal-free approaches to chemical and pharmaceutical hazard and risk assessment has taken off. Alternative approaches are being developed starting from the perspective of human biology and physiology. Neural tube closure is a vital step that occurs early in human development. Correct closure of the neural tube depends on a complex interplay between proteins along a number of protein concentration gradients. The sensitivity of neural tube closure to chemical disturbance of signalling pathways such as the retinoid pathway, is well known. To map the pathways underlying neural tube closure, literature data on the molecular regulation of neural tube closure were collected. As the process of neural tube closure is highly conserved in vertebrates, the extensive literature available for the mouse was used whilst considering its relevance for humans. Thus, important cell compartments, regulatory pathways, and protein interactions essential for neural tube closure under physiological circumstances were identified and mapped. An understanding of aberrant processes leading to neural tube defects (NTDs) requires detailed maps of neural tube embryology, including the complex genetic signals and responses underlying critical cellular dynamical and biomechanical processes. The retinoid signaling pathway serves as a case study for this ontology because of well-defined crosstalk with the genetic control of neural tube patterning and morphogenesis. It is a known target for mechanistically-diverse chemical structures that disrupt neural tube closure The data presented in this manuscript will set the stage for constructing mathematical models and computer simulation of neural tube closure for human-relevant AOPs and predictive toxicology.

Endocrine-disrupting chemicals: economic, regulatory, and policy implications

Endocrine-disrupting chemicals (EDCs) substantially cost society as a result of increases in disease and disability but-unlike other toxicant classes such as carcinogens-have yet to be codified into regulations as a hazard category. This Series paper examines economic, regulatory, and policy approaches to limit human EDC exposures and describes potential improvements. In the EU, general principles for EDCs call for minimisation of human exposure, identification as substances of very high concern, and ban on use in pesticides. In the USA, screening and testing programmes are focused on oestrogenic EDCs exclusively, and regulation is strictly risk-based. Minimisation of human exposure is unlikely without a clear overarching definition for EDCs and relevant pre-marketing test requirements. We call for a multifaceted international programme (eg, modelled on the International Agency for Research in Cancer) to address the effects of EDCs on human health-an approach that would proactively identify hazards for subsequent regulation.

Rapid hazard characterization of environmental chemicals using a compendium of human cell lines from different organs

The lack of adequate toxicity data for the vast majority of chemicals in the environment has spurred the development of new approach methodologies (NAMs). This study aimed to develop a practical high-throughput in vitro model for rapidly evaluating potential hazards of chemicals using a small number of human cells. Forty-two compounds were tested using human induced pluripotent stem cell (iPSC)-derived cells (hepatocytes, neurons, cardiomyocytes and endothelial cells), and a primary endothelial cell line. Both functional and cytotoxicity endpoints were evaluated using high-content imaging. Concentration-response was used to derive points-of-departure (POD). PODs were integrated with ToxPi and used as surrogate NAM-based PODs for risk characterization. We found chemical class-specific similarity among the chemicals tested; metal salts exhibited the highest overall bioactivity. We also observed cell type-specific patterns among classes of chemicals, indicating the ability of the proposed in vitro model to recognize effects on different cell types. Compared to available NAM datasets, such as ToxCast/Tox21 and chemical structure-based descriptors, we found that the data from the five-cell-type model was as good or even better in assigning compounds to chemical classes. Additionally, the PODs from this model performed well as a conservative surrogate for regulatory in vivo PODs and were less likely to underestimate in vivo potency and potential risk compared to other NAM-based PODs. In summary, we demonstrate the potential of this in vitro screening model to inform rapid risk-based decision-making through ranking, clustering, and assessment of both hazard and risks of diverse environmental chemicals.

Stacked Generalization with Applicability Domain Outperforms Simple QSAR on in Vitro Toxicological Data

The development of in silico tools able to predict bioactivity and toxicity of chemical substances is a powerful solution envisioned to assess toxicity as early as possible. To enable the development of such tools, the ToxCast program has generated and made publicly available in vitro bioactivity data for thousands of compounds. The goal of the present study is to characterize and explore the data from ToxCast in terms of Machine Learning capability. For this, a large scale analysis on the entire database has been performed to build models to predict bioactivities measured in in vitro assays. Simple classical QSAR algorithms (ANN, SVM, LDA, random forest, and Bayesian) were first applied on the data, and the results of these algorithms suggested that they do not seem to be well-suited for data sets with a high proportion of inactive compounds. The study then showed for the first time that the use of an ensemble method named "Stacked generalization" could improve the model performance on this type of data. Indeed, for 61% of 483 models, the Stacked method led to models with higher performance. Moreover, the combination of this ensemble method with an applicability domain filter allows one to assess the reliability of the predictions for further compound prioritization. In particular we showed that for 50% of the models, the ROC score is better if we do not consider the compounds that are not within the applicability domain.

The Impact of Novel Assessment Methodologies in Toxicology on Green Chemistry and Chemical Alternatives

The field of experimental toxicology is rapidly advancing by incorporating novel techniques and methods that provide a much more granular view into the mechanisms of potential adverse effects of chemical exposures on human health. The data from various in vitro assays and computational models are useful not only for increasing confidence in hazard and risk decisions, but also are enabling better, faster and cheaper assessment of a greater number of compounds, mixtures, and complex products. This is of special value to the field of green chemistry where design of new materials or alternative uses of existing ones is driven, at least in part, by considerations of safety. This article reviews the state of the science and decision-making in scenarios when little to no data may be available to draw conclusions about which choice in green chemistry is "safer." It is clear that there is no "one size fits all" solution and multiple data streams need to be weighed in making a decision. Moreover, the overall level of familiarity of the decision-makers and scientists alike with new assessment methodologies, their validity, value and limitations is evolving. Thus, while the "impact" of the new developments in toxicology on the field of green chemistry is great already, it is premature to conclude that the data from new assessment methodologies have been widely accepted yet.

A systems biology approach to predictive developmental neurotoxicity of a larvicide used in the prevention of Zika virus transmission

The need to prevent developmental brain disorders has led to an increased interest in efficient neurotoxicity testing. When an epidemic of microcephaly occurred in Brazil, Zika virus infection was soon identified as the likely culprit. However, the pathogenesis appeared to be complex, and a larvicide used to control mosquitoes responsible for transmission of the virus was soon suggested as an important causative factor. Yet, it is challenging to identify relevant and efficient tests that are also in line with ethical research defined by the 3Rs rule (Replacement, Reduction and Refinement). Especially in an acute situation like the microcephaly epidemic, where little toxicity documentation is available, new and innovative alternative methods, whether in vitro or in silico, must be considered. We have developed a network-based model using an integrative systems biology approach to explore the potential developmental neurotoxicity, and we applied this method to examine the larvicide pyriproxyfen widely used in the prevention of Zika virus transmission. Our computational model covered a wide range of possible pathways providing mechanistic hypotheses between pyriproxyfen and neurological disorders via protein complexes, thus adding to the plausibility of pyriproxyfen neurotoxicity. Although providing only tentative evidence and comparisons with retinoic acid, our computational systems biology approach is rapid and inexpensive. The case study of pyriproxyfen illustrates its usefulness as an initial or screening step in the assessment of toxicity potentials of chemicals with incompletely known toxic properties.

Consensus statement on the need for innovation, transition and implementation of developmental neurotoxicity (DNT) testing for regulatory purposes

This consensus statement voices the agreement of scientific stakeholders from regulatory agencies, academia and industry that a new framework needs adopting for assessment of chemicals with the potential to disrupt brain development. An increased prevalence of neurodevelopmental disorders in children has been observed that cannot solely be explained by genetics and recently pre- and postnatal exposure to environmental chemicals has been suspected as a causal factor. There is only very limited information on neurodevelopmental toxicity, leaving thousands of chemicals, that are present in the environment, with high uncertainty concerning their developmental neurotoxicity (DNT) potential. Closing this data gap with the current test guideline approach is not feasible, because the in vivo bioassays are far too resource-intensive concerning time, money and number of animals. A variety of in vitro methods are now available, that have the potential to close this data gap by permitting mode-of-action-based DNT testing employing human stem cells-derived neuronal/glial models. In vitro DNT data together with in silico approaches will in the future allow development of predictive models for DNT effects. The ultimate application goals of these new approach methods for DNT testing are their usage for different regulatory purposes.

•

An increased prevalence of neurodevelopmental disorders in children is observed.

•

There is very limited information on neurodevelopmental toxicity (DNT) induced by environmental chemicals.

•

A new framework is required for assessment of chemicals with the potential to disrupt brain development.

•

In vitro DNT data together with in silico approaches should be used for regulatory purposes.

Combined toxicity of triclosan, 2,4-dichlorophenol and 2,4,6-trichlorophenol to zebrafish (Danio rerio)

Triclosan (TCS), 2,4,6-trichlorophenol (2,4,6-TCP) and 2,4-dichlorophenol (2,4-DCP) are the most prevalent chlorinated phenolic pollutants in aquatic environments. Our results showed LC₅₀ and EC₅₀ values of 0.51, 1.11, 2.45mg/L, and 0.36, 0.74, 1.53mg/L for TCS, 2,4,6-TCP and 2,4-DCP, respectively, to 120hpf zebrafish. The highest TCSD (the mixture of TCS, 2,4,6-TCP and 2,4-DCP) toxicity was observed at a TCS:2,4,6-TCP:2,4-DCP concentration ratio of 1:2:4. LC₅₀ and EC₅₀ values of TCSD mixtures for 120-hpf zebrafish were 2.28 and 1.16mg/L, respectively. Two toxicity assessment methods (Toxic Unit and Mixture Toxicity Index) indicated that TCSD interactions produced partly additive toxicity. TCSD exposure decreased zebrafish hatching rate and led to a series of malformations. Following alkaline phosphatase staining, a large area of vascular ablation was observed with almost complete disappearance of vascular branches and a smaller coverage range. Prominent reddening of the yolk sac and visceral mass after oil red O staining implied that TCSD exposure severely affected fat metabolism. Following acridine orange staining, cell death occurred in eyes while high TCSD concentrations (0.84mg/L) induced cardiovascular circulation dysfunction. Alcian blue staining increased the α angle between Meckel's cartilages and β angle between two ceratobranchial. Basihyal and palatoquadrate became shorter and developmental abnormality or defects occurred in the fifth ceratobranchial. Overall, these results provide a theoretical basis for systematically evaluating the combined toxicity of the prevalent chlorinated phenolic pollutants in real-world aquatic environments.

Machine Learning Methods in Computational Toxicology

Various methods of machine learning, supervised and unsupervised, linear and nonlinear, classification and regression, in combination with various types of molecular descriptors, both "handcrafted" and "data-driven," are considered in the context of their use in computational toxicology. The use of multiple linear regression, variants of naïve Bayes classifier, k-nearest neighbors, support vector machine, decision trees, ensemble learning, random forest, several types of neural networks, and deep learning is the focus of attention of this review. The role of fragment descriptors, graph mining, and graph kernels is highlighted. The application of unsupervised methods, such as Kohonen's self-organizing maps and related approaches, which allow for combining predictions with data analysis and visualization, is also considered. The necessity of applying a wide range of machine learning methods in computational toxicology is underlined.

Predicting Organ Toxicity Using in Vitro Bioactivity Data and Chemical Structure

Animal testing alone cannot practically evaluate the health hazard posed by tens of thousands of environmental chemicals. Computational approaches making use of high-throughput experimental data may provide more efficient means to predict chemical toxicity. Here, we use a supervised machine learning strategy to systematically investigate the relative importance of study type, machine learning algorithm, and type of descriptor on predicting in vivo repeat-dose toxicity at the organ-level. A total of 985 compounds were represented using chemical structural descriptors, ToxPrint chemotype descriptors, and bioactivity descriptors from ToxCast in vitro high-throughput screening assays. Using ToxRefDB, a total of 35 target organ outcomes were identified that contained at least 100 chemicals (50 positive and 50 negative). Supervised machine learning was performed using Naïve Bayes, k-nearest neighbor, random forest, classification and regression trees, and support vector classification approaches. Model performance was assessed based on F1 scores using 5-fold cross-validation with balanced bootstrap replicates. Fixed effects modeling showed the variance in F1 scores was explained mostly by target organ outcome, followed by descriptor type, machine learning algorithm, and interactions between these three factors. A combination of bioactivity and chemical structure or chemotype descriptors were the most predictive. Model performance improved with more chemicals (up to a maximum of 24%), and these gains were correlated (ρ = 0.92) with the number of chemicals. Overall, the results demonstrate that a combination of bioactivity and chemical descriptors can accurately predict a range of target organ toxicity outcomes in repeat-dose studies, but specific experimental and methodologic improvements may increase predictivity.

The Next Generation of Risk Assessment Multi-Year Study-Highlights of Findings, Applications to Risk Assessment, and Future Directions

BACKGROUND

The Next Generation (NexGen) of Risk Assessment effort is a multi-year collaboration among several organizations evaluating new, potentially more efficient molecular, computational, and systems biology approaches to risk assessment. This article summarizes our findings, suggests applications to risk assessment, and identifies strategic research directions.

OBJECTIVE

Our specific objectives were to test whether advanced biological data and methods could better inform our understanding of public health risks posed by environmental exposures.

METHODS

New data and methods were applied and evaluated for use in hazard identification and dose-response assessment. Biomarkers of exposure and effect, and risk characterization were also examined. Consideration was given to various decision contexts with increasing regulatory and public health impacts. Data types included transcriptomics, genomics, and proteomics. Methods included molecular epidemiology and clinical studies, bioinformatic knowledge mining, pathway and network analyses, short-duration in vivo and in vitro bioassays, and quantitative structure activity relationship modeling.

DISCUSSION

NexGen has advanced our ability to apply new science by more rapidly identifying chemicals and exposures of potential concern, helping characterize mechanisms of action that influence conclusions about causality, exposure-response relationships, susceptibility and cumulative risk, and by elucidating new biomarkers of exposure and effects. Additionally, NexGen has fostered extensive discussion among risk scientists and managers and improved confidence in interpreting and applying new data streams.

CONCLUSIONS

While considerable uncertainties remain, thoughtful application of new knowledge to risk assessment appears reasonable for augmenting major scope assessments, forming the basis for or augmenting limited scope assessments, and for prioritization and screening of very data limited chemicals. Citation: Cote I, Andersen ME, Ankley GT, Barone S, Birnbaum LS, Boekelheide K, Bois FY, Burgoon LD, Chiu WA, Crawford-Brown D, Crofton KM, DeVito M, Devlin RB, Edwards SW, Guyton KZ, Hattis D, Judson RS, Knight D, Krewski D, Lambert J, Maull EA, Mendrick D, Paoli GM, Patel CJ, Perkins EJ, Poje G, Portier CJ, Rusyn I, Schulte PA, Simeonov A, Smith MT, Thayer KA, Thomas RS, Thomas R, Tice RR, Vandenberg JJ, Villeneuve DL, Wesselkamper S, Whelan M, Whittaker C, White R, Xia M, Yauk C, Zeise L, Zhao J, DeWoskin RS. 2016. The Next Generation of Risk Assessment multiyear study-highlights of findings, applications to risk assessment, and future directions. Environ Health Perspect 124:1671-1682; http://dx.doi.org/10.1289/EHP233.

Inflammatory Cytokines and White Blood Cell Counts Response to Environmental Levels of Diesel Exhaust and Ozone Inhalation Exposures

Epidemiological observations of urban inhalation exposures to diesel exhaust (DE) and ozone (O3) have shown pre-clinical cardiopulmonary responses in humans. Identifying the key biological mechanisms that initiate these health bioindicators is difficult due to variability in environmental exposure in time and from person to person. Previously, environmentally controlled human exposure chambers have been used to study DE and O3 dose-response patterns separately, but investigation of co-exposures has not been performed under controlled conditions. Because a mixture is a more realistic exposure scenario for the general public, in this study we investigate the relationships of urban levels of urban-level DE exposure (300 μg/m3), O3 (0.3 ppm), DE + O3 co-exposure, and innate immune system responses. Fifteen healthy human volunteers were studied for changes in ten inflammatory cytokines (interleukins 1β, 2, 4, 5, 8, 10, 12p70 and 13, IFN-γ, and TNF-α) and counts of three white blood cell types (lymphocytes, monocytes, and neutrophils) following controlled exposures to DE, O3, and DE+O3. The results show subtle cytokines responses to the diesel-only and ozone-only exposures, and that a more complex (possibly synergistic) relationship exists in the combination of these two exposures with suppression of IL-5, IL-12p70, IFN-γ, and TNF-α that persists up to 22-hours for IFN-γ and TNF-α. The white blood cell differential counts showed significant monocyte and lymphocyte decreases and neutrophil increases following the DE + O3 exposure; lymphocytes and neutrophils changes also persist for at least 22-hours. Because human studies must be conducted under strict safety protocols at environmental levels, these effects are subtle and are generally only seen with detailed statistical analysis. This study indicates that the observed associations between environmental exposures and cardiopulmonary effects are possibly mediated by inflammatory response mechanisms.

Models of germ cell development and their application for toxicity studies

Germ cells are unique in their ability to transfer genetic information and traits from generation to generation. As such, the proper development of germ cells and the integrity of their genome are paramount to the health of organisms and the survival of species. Germ cells are also exquisitely sensitive to environmental influences although the testing of germ cell toxicity, especially in females, has proven particularly challenging. In this review, we first describe the remarkable odyssey of germ cells in mammals, with an emphasis on the female germline, from their initial specification early during embryogenesis to the generation of mature gametes in adults. We also describe the current methods used in germ cell toxicity testing and their limitations in examining the complex features of mammalian germ cell development. To bypass these challenges, we propose the use of alternative model systems such as Saccharomyces cerevisiae, Drosophila melanogaster, Caenorhabditis elegans, and in vitro germ cell methods that have distinct advantages over traditional toxicity models. We discuss the benefits and limitations of each approach, their application to germ cell toxicity studies, and the need for computational approaches to maximize the usefulness of these models. Together, the inclusion of these alternative germ cell toxicity models will be invaluable for the examination of stages not easily accessible in mammals as well as the large scale, high-throughput investigation of germ cell toxicity.

A Comparison of ToxCast Test Results with In Vivo and Other In Vitro Endpoints for Neuro, Endocrine, and Developmental Toxicities: A Case Study Using Endosulfan and Methidathion

INTRODUCTION

The U.S. Environmental Protection Agency's (EPA's) Toxicity Forecaster (ToxCast) is a potential tool for chemical prioritization, hazard identification, and risk assessment. We conducted a case study to compare ToxCast data with endpoints from other in vitro and in vivo studies for two data-rich pesticides: endosulfan and methidathion.

METHODS

ToxCast assays for endocrine disruption, development (zebrafish), and neurotoxicity were qualitatively compared to traditional neurotoxicity, developmental and reproductive toxicity findings. We also used in vitro-in vivo extrapolation to convert half-maximal activity concentrations in active ToxCast assays to rat oral equivalent doses, and quantitatively compared these to the lowest observable effect level (LOEL) from in vivo studies.

RESULTS

Endosulfan was inactive for GABAA R, unlike in vivo; but active with dopamine transporter assays and was neurotoxic in zebrafish as expected. Methidathion was not active for these endpoints in vivo or in vitro. Acetylcholinesterase inhibition was ToxCast-inactive, although both pesticides are inhibitors in vivo. ToxCast results were generally inactive for endosulfan estrogen receptor agonism and androgen receptor antagonism unlike in vivo. Calculated oral equivalent doses for estrogen receptor and androgen receptor pathways and for zebrafish assays for both compounds were generally consistent with in vivo LOELs. Endosulfan showed neurotoxicity and both pesticides showed developmental effects in the zebrafish assays, although methidathion is not developmentally toxic in vivo.

CONCLUSIONS

ToxCast's predictions showed concordance on some endpoints and nonconcordance, consisting mainly of false inactives, in several critical endpoints, likely due to a lack of metabolic activation and limitations in assay design. Zebrafish assays were good predictors of developmental toxicity and neurotoxicity for endosulfan.

Predicting hepatotoxicity using ToxCast in vitro bioactivity and chemical structure

The U.S. Tox21 and EPA ToxCast program screen thousands of environmental chemicals for bioactivity using hundreds of high-throughput in vitro assays to build predictive models of toxicity. We represented chemicals based on bioactivity and chemical structure descriptors, then used supervised machine learning to predict in vivo hepatotoxic effects. A set of 677 chemicals was represented by 711 in vitro bioactivity descriptors (from ToxCast assays), 4,376 chemical structure descriptors (from QikProp, OpenBabel, PaDEL, and PubChem), and three hepatotoxicity categories (from animal studies). Hepatotoxicants were defined by rat liver histopathology observed after chronic chemical testing and grouped into hypertrophy (161), injury (101) and proliferative lesions (99). Classifiers were built using six machine learning algorithms: linear discriminant analysis (LDA), Naïve Bayes (NB), support vector machines (SVM), classification and regression trees (CART), k-nearest neighbors (KNN), and an ensemble of these classifiers (ENSMB). Classifiers of hepatotoxicity were built using chemical structure descriptors, ToxCast bioactivity descriptors, and hybrid descriptors. Predictive performance was evaluated using 10-fold cross-validation testing and in-loop, filter-based, feature subset selection. Hybrid classifiers had the best balanced accuracy for predicting hypertrophy (0.84 ± 0.08), injury (0.80 ± 0.09), and proliferative lesions (0.80 ± 0.10). Though chemical and bioactivity classifiers had a similar balanced accuracy, the former were more sensitive, and the latter were more specific. CART, ENSMB, and SVM classifiers performed the best, and nuclear receptor activation and mitochondrial functions were frequently found in highly predictive classifiers of hepatotoxicity. ToxCast and ToxRefDB provide the largest and richest publicly available data sets for mining linkages between the in vitro bioactivity of environmental chemicals and their adverse histopathological outcomes. Our findings demonstrate the utility of high-throughput assays for characterizing rodent hepatotoxicants, the benefit of using hybrid representations that integrate bioactivity and chemical structure, and the need for objective evaluation of classification performance.

Beyond a means of exposure: a new view of the mother in toxicology research

Toxicological studies generally view pregnant animals as a conduit through which gestational exposure of offspring to chemicals can be achieved, allowing for the study of developmental toxicity.

Databases and collaboration require standards for human stem cell research

Stem cell research is at an important juncture: despite significant potential for human health and several countries with key initiatives to expedite commercialization, there are gaps in capturing and exploiting the results of past and current research. Here, we propose a concerted plan that could be taken to foster a more collaborative approach and ensure that all research efforts can be leveraged across the community. The creation of a definitive centralized database repository, or at least harmonized data repositories, for stem cell groups in academia and industry, enabling secure selective sharing of data when needed, could provide the core structure that is sought globally and protect intellectual property. The development of minimum information about stem cell experiments (MIASCE) could be key to this development.

A critical appraisal of the process of regulatory implementation of novel in vivo and in vitro methods for chemical hazard and risk assessment

Regulatory toxicology urgently needs applicable alternative test systems that reduce animal use, testing time, and cost. European regulation on cosmetic ingredients has already banned animal experimentation for hazard identification, and public awareness drives toward additional restrictions in other regulatory frameworks as well. In addition, scientific progress stimulates a more mechanistic approach of hazard identification. Nevertheless, the implementation of alternative methods is lagging far behind their development. In search for general bottlenecks for the implementation of alternative methods, this manuscript reviews the state of the art as to the development and implementation of 10 diverse test systems in various areas of toxicological hazard assessment. They vary widely in complexity and regulatory acceptance status. The assays are reviewed as to parameters assessed, biological system involved, standardization, interpretation of results, extrapolation to human hazard, position in testing strategies, and current regulatory acceptance status. Given the diversity of alternative methods in many aspects, no common bottlenecks could be identified that hamper implementation of individual alternative assays in general. However, specific issues for the regulatory acceptance and application were identified for each assay. Acceptance of one-in-one replacement of complex in vivo tests by relatively simple in vitro assays is not feasible. Rather, innovative approaches using test batteries are required together with metabolic information and in vitro to in vivo dose extrapolation to convincingly provide the same level of information of current in vivo tests. A mechanistically based alternative approach using the Adverse Outcome Pathway concept could stimulate further (regulatory) acceptance of non-animal tests.

Composition and applications of focus libraries to phenotypic assays

The wealth of bioactivity information now available on low-molecular weight compounds has enabled a paradigm shift in chemical biology and early phase drug discovery efforts. Traditionally chemical libraries have been most commonly employed in screening approaches where a bioassay is used to characterize a chemical library in a random search for active samples. However, robust curating of bioassay data, establishment of ontologies enabling mining of large chemical biology datasets, and a wealth of public chemical biology information has made possible the establishment of highly annotated compound collections. Such annotated chemical libraries can now be used to build a pathway/target hypothesis and have led to a new view where chemical libraries are used to characterize a bioassay. In this article we discuss the types of compounds in these annotated libraries composed of tools, probes, and drugs. As well, we provide rationale and a few examples for how such libraries can enable phenotypic/forward chemical genomic approaches. As with any approach, there are several pitfalls that need to be considered and we also outline some strategies to avoid these.

A short-term in vivo screen using fetal testosterone production, a key event in the phthalate adverse outcome pathway, to predict disruption of sexual differentiation

This study was designed to develop and validate a short-term in vivo protocol termed the Fetal Phthalate Screen (FPS) to detect phthalate esters (PEs) and other chemicals that disrupt fetal testosterone synthesis and testis gene expression in rats. We propose that the FPS can be used to screen chemicals that produce adverse developmental outcomes via disruption of the androgen synthesis pathway more rapidly and efficiently, and with fewer animals than a postnatal one-generation study. Pregnant rats were dosed from gestational day (GD) 14 to 18 at one dose level with one of 27 chemicals including PEs, PE alternatives, pesticides known to inhibit steroidogenesis, an estrogen and a potent PPARα agonist and ex vivo testis testosterone production (T Prod) was measured on GD 18. We also included some chemicals with "unknown" activity including DMEP, DHeP, DHEH, DPHCH, DAP, TOTM, tetrabromo-diethyl hexyl phthalate (BrDEHP), and a relatively potent environmental estrogen BPAF. Dose-response studies also were conducted with this protocol with 11 of the above chemicals to determine their relative potencies. CD-1 mice also were exposed to varying dose levels of DPeP from GD 13 to 17 to determine if DPeP reduced T Prod in this species since there is a discrepancy among the results of in utero studies of PEs in mice. Compared to the known male reproductive effects of the PEs in rats the FPS correctly identified all known "positives" and "negatives" tested. Seven of eight "unknowns" tested were "negatives", they did not reduce T Prod, whereas DAP produced an "equivocal" response. Finally, a dose-response study with DPeP in CD-1 mice revealed that fetal T Prod can be inhibited by exposure to a PE in utero in this species, but at a higher dose level than required in rats.Key words. Phthalate Syndrome, Fetal endocrine biomarkers, Phthalate adverse outcome pathway, testosterone production, fetal rat testis.

In vitro and modelling approaches to risk assessment from the U.S. Environmental Protection Agency ToxCast programme

A significant challenge in toxicology is the 'too many chemicals' problem. Human beings and environmental species are exposed to tens of thousands of chemicals, only a small percentage of which have been tested thoroughly using standard in vivo test methods. This study reviews several approaches that are being developed to deal with this problem by the U.S. Environmental Protection Agency, under the umbrella of the ToxCast programme (http://epa.gov/ncct/toxcast/). The overall approach is broken into seven tasks: (i) identifying biological pathways that, when perturbed, can lead to toxicity; (ii) developing high-throughput in vitro assays to test chemical perturbations of these pathways; (iii) identifying the universe of chemicals with likely human or ecological exposure; (iv) testing as many of these chemicals as possible in the relevant in vitro assays; (v) developing hazard models that take the results of these tests and identify chemicals as being potential toxicants; (vi) generating toxicokinetics data on these chemicals to predict the doses at which these hazard pathways would be activated; and (vii) developing exposure models to identify chemicals for which these hazardous dose levels could be achieved. This overall strategy is described and briefly illustrated with recent examples from the ToxCast programme.

Teratological effects of a panel of sixty water-soluble toxicants on zebrafish development

The zebrafish larva is a promising whole-animal model for safety pharmacology, environmental risk assessment, and developmental toxicity. This model has been used for the high-throughput toxicity screening of various compounds. Our aim here is to identify possible phenotypic markers of teratogenicity in zebrafish embryos that could be used for the assaying compounds for reproductive toxicity. We have screened a panel of 60 water-soluble toxicants to examine their effects on zebrafish development. A total of 22,080 wild-type zebrafish larvae were raised in 250 μL defined buffer in 96-well plates at a plating density of one embryo per well. They were exposed for a 96-h period starting at 24 h post-fertilization. A logarithmic concentration series was used for range-finding, followed by a narrower geometric series for developmental toxicity assessment. A total of 9017 survivors were analyzed at 5 days post-fertilization for nine phenotypes, namely, (1) normal, (2) pericardial oedema, (3) yolk sac oedema, (4) melanophores dispersed, (5) bent tail tip, (6) bent body axis, (7) abnormal Meckel's cartilage, (8) abnormal branchial arches, and (9) uninflated swim bladder. For each toxicant, the EC50 (concentration required to produce one or more of these abnormalities in 50% of embryos) was also calculated. For the majority of toxicants (55/60) there was, at the population level, a statistically significant, concentration-dependent increase in the incidence of abnormal phenotypes among survivors. The commonest abnormalities were pericardial oedema, yolk sac oedema, dispersed melanophores, and uninflated swim bladder. It is possible therefore that these could prove to be general indicators of reproductive toxicity in the zebrafish embryo assay.

Mechanistic validation

Validation of new approaches in regulatory toxicology is commonly defined as the independent assessment of the reproducibility and relevance (the scientific basis and predictive capacity) of a test for a particular purpose. In large ring trials, the emphasis to date has been mainly on reproducibility and predictive capacity (comparison to the traditional test) with less attention given to the scientific or mechanistic basis. Assessing predictive capacity is difficult for novel approaches (which are based on mechanism), such as pathways of toxicity or the complex networks within the organism (systems toxicology). This is highly relevant for implementing Toxicology for the 21st Century, either by high-throughput testing in the ToxCast/Tox21 project or omics-based testing in the Human Toxome Project. This article explores the mostly neglected assessment of a test's scientific basis, which moves mechanism and causality to the foreground when validating/qualifying tests. Such mechanistic validation faces the problem of establishing causality in complex systems. However, pragmatic adaptations of the Bradford Hill criteria, as well as bioinformatic tools, are emerging. As critical infrastructures of the organism are perturbed by a toxic mechanism we argue that by focusing on the target of toxicity and its vulnerability, in addition to the way it is perturbed, we can anchor the identification of the mechanism and its verification.

Non-monotonic dose responses in studies of endocrine disrupting chemicals: bisphenol a as a case study

Non-monotonic dose response curves (NMDRCs) have been demonstrated for natural hormones and endocrine disrupting chemicals (EDCs) in a variety of biological systems including cultured cells, whole organ cultures, laboratory animals and human populations. The mechanisms responsible for these NMDRCs are well known, typically related to the interactions between the ligand (hormone or EDC) and a hormone receptor. Although there are hundreds of examples of NMDRCs in the EDC literature, there are claims that they are not 'common enough' to influence the use of high-to-low dose extrapolations in risk assessments. Here, we chose bisphenol A (BPA), a well-studied EDC, to assess the frequency of non-monotonic responses. Our results indicate that NMDRCs are common in the BPA literature, occurring in greater than 20% of all experiments and in at least one endpoint in more than 30% of all studies we examined. We also analyzed the types of endpoints that produce NMDRCs in vitro and factors related to study design that influence the ability to detect these kinds of responses. Taken together, these results provide strong evidence for NMDRCs in the EDC literature, specifically for BPA, and question the current risk assessment practice where 'safe' low doses are predicted from high dose exposures.