Predictive toxicology today: the transition from biological knowledge to practicable models

Multi-Endpoint Acute Toxicity Assessment of Organic Compounds Using Large-Scale Machine Learning Modeling

In recent years, alternative animal testing methods such as computational and machine learning approaches have become increasingly crucial for toxicity testing. However, the complexity and scarcity of available biomedical data challenge the development of predictive models. Combining nonlinear machine learning together with multicondition descriptors offers a solution for using data from various assays to create a robust model. This work applies multicondition descriptors (MCDs) to develop a QSTR (Quantitative Structure-Toxicity Relationship) model based on a large toxicity data set comprising more than 80,000 compounds and 59 different end points (122,572 data points). The prediction capabilities of developed single-task multi-end point machine learning models as well as a novel data analysis approach with the use of Convolutional Neural Networks (CNN) are discussed. The results show that using MCDs significantly improves the model and using them with CNN-1D yields the best result (R2train = 0.93, R2ext = 0.70). Several structural features showed a high level of contribution to the toxicity, including van der Waals surface area (VSA), number of nitrogen-containing fragments (nN+), presence of S-P fragments, ionization potential, and presence of C-N fragments. The developed models can be very useful tools to predict the toxicity of various compounds under different conditions, enabling quick toxicity assessment of new compounds.

Technology-based approaches toward a better understanding of neuro-coagulation in brain homeostasis

Blood coagulation factors can enter the brain under pathological conditions that affect the blood-brain interface. Besides their contribution to pathological brain states, such as neural hyperexcitability, neurodegeneration, and scar formation, coagulation factors have been linked to several physiological brain functions. It is for example well established that the coagulation factor thrombin modulates synaptic plasticity; it affects neural excitability and induces epileptic seizures via activation of protease-activated receptors in the brain. However, major limitations of current experimental and clinical approaches have prevented us from obtaining a profound mechanistic understanding of "neuro-coagulation" in health and disease. Here, we present how novel human relevant models, i.e., Organ-on-Chips equipped with advanced sensors, can help overcoming some of the limitations in the field, thus providing a perspective toward a better understanding of neuro-coagulation in brain homeostasis.

Machine learning-based biomarkers identification from toxicogenomics - Bridging to regulatory relevant phenotypic endpoints

One of the major challenges in realization and implementations of the Tox21 vision is the urgent need to establish quantitative link between in-vitro assay molecular endpoint and in-vivo regulatory-relevant phenotypic toxicity endpoint. Current toxicomics approach still mostly rely on large number of redundant markers without pre-selection or ranking, therefore, selection of relevant biomarkers with minimal redundancy would reduce the number of markers to be monitored and reduce the cost, time, and complexity of the toxicity screening and risk monitoring. Here, we demonstrated that, using time series toxicomics in-vitro assay along with machine learning-based feature selection (maximum relevance and minimum redundancy (MRMR)) and classification method (support vector machine (SVM)), an "optimal" number of biomarkers with minimum redundancy can be identified for prediction of phenotypic toxicity endpoints with good accuracy. We included two case studies for in-vivo carcinogenicity and Ames genotoxicity prediction, using 20 selected chemicals including model genotoxic chemicals and negative controls, respectively. The results suggested that, employing the adverse outcome pathway (AOP) concept, molecular endpoints based on a relatively small number of properly selected biomarker-ensemble involved in the conserved DNA-damage and repair pathways among eukaryotes, were able to predict both Ames genotoxicity endpoints and in-vivo carcinogenicity in rats. A prediction accuracy of 76% with AUC = 0.81 was achieved while predicting in-vivo carcinogenicity with the top-ranked five biomarkers. For Ames genotoxicity prediction, the top-ranked five biomarkers were able to achieve prediction accuracy of 70% with AUC = 0.75. However, the specific biomarkers identified as the top-ranked five biomarkers are different for the two different phenotypic genotoxicity assays. The top-ranked biomarkers for the in-vivo carcinogenicity prediction mainly focused on double strand break repair and DNA recombination, whereas the selected top-ranked biomarkers for Ames genotoxicity prediction are associated with base- and nucleotide-excision repair The method developed in this study will help to fill in the knowledge gap in phenotypic anchoring and predictive toxicology, and contribute to the progress in the implementation of tox 21 vision for environmental and health applications.

Large-Scale Modeling of Multispecies Acute Toxicity End Points Using Consensus of Multitask Deep Learning Methods

Computational methods to predict molecular properties regarding safety and toxicology represent alternative approaches to expedite drug development, screen environmental chemicals, and thus significantly reduce associated time and costs. There is a strong need and interest in the development of computational methods that yield reliable predictions of toxicity, and many approaches, including the recently introduced deep neural networks, have been leveraged towards this goal. Herein, we report on the collection, curation, and integration of data from the public data sets that were the source of the ChemIDplus database for systemic acute toxicity. These efforts generated the largest publicly available such data set comprising > 80,000 compounds measured against a total of 59 acute systemic toxicity end points. This data was used for developing multiple single- and multitask models utilizing random forest, deep neural networks, convolutional, and graph convolutional neural network approaches. For the first time, we also reported the consensus models based on different multitask approaches. To the best of our knowledge, prediction models for 36 of the 59 end points have never been published before. Furthermore, our results demonstrated a significantly better performance of the consensus model obtained from three multitask learning approaches that particularly predicted the 29 smaller tasks (less than 300 compounds) better than other models developed in the study. The curated data set and the developed models have been made publicly available at https://github.com/ncats/ld50-multitask, https://predictor.ncats.io/, and https://cactus.nci.nih.gov/download/acute-toxicity-db (data set only) to support regulatory and research applications.

Robotic fluidic coupling and interrogation of multiple vascularized organ chips

Organ chips can recapitulate organ-level (patho)physiology, yet pharmacokinetic and pharmacodynamic analyses require multi-organ systems linked by vascular perfusion. Here, we describe an 'interrogator' that employs liquid-handling robotics, custom software and an integrated mobile microscope for the automated culture, perfusion, medium addition, fluidic linking, sample collection and in situ microscopy imaging of up to ten organ chips inside a standard tissue-culture incubator. The robotic interrogator maintained the viability and organ-specific functions of eight vascularized, two-channel organ chips (intestine, liver, kidney, heart, lung, skin, blood-brain barrier and brain) for 3 weeks in culture when intermittently fluidically coupled via a common blood substitute through their reservoirs of medium and endothelium-lined vascular channels. We used the robotic interrogator and a physiological multicompartmental reduced-order model of the experimental system to quantitatively predict the distribution of an inulin tracer perfused through the multi-organ human-body-on-chips. The automated culture system enables the imaging of cells in the organ chips and the repeated sampling of both the vascular and interstitial compartments without compromising fluidic coupling.

Deep Neural Network Models for Predicting Chemically Induced Liver Toxicity Endpoints From Transcriptomic Responses

Improving the accuracy of toxicity prediction models for liver injuries is a key element in evaluating the safety of drugs and chemicals. Mechanism-based information derived from expression (transcriptomic) data, in combination with machine-learning methods, promises to improve the accuracy and robustness of current toxicity prediction models. Deep neural networks (DNNs) have the advantage of automatically assembling the relevant features from a large number of input features. This makes them especially suitable for modeling transcriptomic data, which typically contain thousands of features. Here, we gaged gene- and pathway-level feature selection schemes using single- and multi-task DNN approaches in predicting chemically induced liver injuries (biliary hyperplasia, fibrosis, and necrosis) from whole-genome DNA microarray data. The single-task DNN models showed high predictive accuracy and endpoint specificity, with Matthews correlation coefficients for the three endpoints on 10-fold cross validation ranging from 0.56 to 0.89, with an average of 0.74 in the best feature sets. The DNN models outperformed Random Forest models in cross validation and showed better performance than Support Vector Machine models when tested in the external validation datasets. In the cross validation studies, the effect of the feature selection scheme was negligible among the studied feature sets. Further evaluation of the models on their ability to predict the injury phenotype per se for non-chemically induced injuries revealed the robust performance of the DNN models across these additional external testing datasets. Thus, the DNN models learned features specific to the injury phenotype contained in the gene expression data.

Applications of Chemoinformatics in Predictive Toxicology for Regulatory Purposes, Especially in the Context of the EU REACH Legislation

Chemoinformatics methodologies such as QSAR/QSPR have been used for decades in drug discovery projects, especially for the finding of new compounds with therapeutic properties and the optimization of ADME properties on chemical series. The application of computational techniques in predictive toxicology is much more recent, and they are experiencing an increasingly interest because of the new legal requirements imposed by national and international regulations. In the pharmaceutical field, the US Food and Drug Administration (FDA) support the use of predictive models for regulatory decision-making when assessing the genotoxic and carcinogenic potential of drug impurities. In Europe, the REACH legislation promotes the use of QSAR in order to reduce the huge amount of animal testing needed to demonstrate the safety of new chemical entities subjected to registration, provided they meet specific conditions to ensure their quality and predictive power. In this review, the authors summarize the state of art of in silico methods for regulatory purposes, with especial emphasis on QSAR models.

Endocrine Disruptors: Data-based survey of in vivo tests, predictive models and the Adverse Outcome Pathway

The protection from endocrine disruptors is a high regulatory priority. Key issues are the characterization of in vivo assays, and the identification of reference chemicals to validate alternative methods. In this exploration, publicly available databases for in vivo assays for endocrine disruption were collected and compared: Rodent Uterotrophic, Rodent Repeated Dose 28-day Oral Toxicity, 21-Day Fish, and Daphnia magna reproduction assays. Only the Uterotrophic and 21-Day Fish assays results correlated with each other. The in vivo assays data were viewed in relation to the Adverse Outcome Pathway, using as a probe 18 ToxCast in vitro assays for the ER pathway. These are the same data at the basis of the EPA agonist ToxERscore model, whose good predictivity was confirmed. The multivariate comparison of the in vitro/in vivo assays suggests that the interaction with receptors is a major determinant of in vivo results, and is the critical basis for building predictive computational models. In agreement with the above, this work also shows that it is possible to build predictive models for the Uterotrophic and 21-Day Fish assays using a limited selection of Toxcast assays.