Screening Chemicals for Receptor-Mediated Toxicological and Pharmacological Endpoints: Using Public Data to Build Screening Tools within a KNIME Workflow

Computational Strategies for Assessing Adverse Outcome Pathways: Hepatic Steatosis as a Case Study

The evolving landscape of chemical risk assessment is increasingly focused on developing tiered, mechanistically driven approaches that avoid the use of animal experiments. In this context, adverse outcome pathways have gained importance for evaluating various types of chemical-induced toxicity. Using hepatic steatosis as a case study, this review explores the use of diverse computational techniques, such as structure-activity relationship models, quantitative structure-activity relationship models, read-across methods, omics data analysis, and structure-based approaches to fill data gaps within adverse outcome pathway networks. Emphasizing the regulatory acceptance of each technique, we examine how these methodologies can be integrated to provide a comprehensive understanding of chemical toxicity. This review highlights the transformative impact of in silico techniques in toxicology, proposing guidelines for their application in evidence gathering for developing and filling data gaps in adverse outcome pathway networks. These guidelines can be applied to other cases, advancing the field of toxicological risk assessment.

Making in silico predictive models for toxicology FAIR

In silico predictive models for toxicology include quantitative structure-activity relationship (QSAR) and physiologically based kinetic (PBK) approaches to predict physico-chemical and ADME properties, toxicological effects and internal exposure. Such models are used to fill data gaps as part of chemical risk assessment. There is a growing need to ensure in silico predictive models for toxicology are available for use and reproducible. This paper describes how the FAIR (Findable, Accessible, Interoperable, Reusable) principles, developed for data sharing, have been applied to in silico predictive models. In particular, this investigation has focussed on how the FAIR principles could be applied to improved regulatory acceptance of predictions from such models. Eighteen principles have been developed that cover all aspects of FAIR. It is intended that FAIRification of in silico predictive models for toxicology will increase their use and acceptance.

In Silico Models for Hepatotoxicity

In this chapter, we review the state of the art of predicting human hepatotoxicity using in silico techniques. There has been significant progress in this area over the past 20 years but there are still some challenges ahead. Principally, these challenges are our partial understanding of a very complex biochemical system and our ability to emulate that in a predictive capacity. Here, we provide an overview of the published modeling approaches in this area to date and discuss their design, strengths and weaknesses. It is interesting to note the diversity in modeling approaches, whether they be statistical algorithms or evidenced-based approaches including structural alerts and pharmacophore models. Irrespective of modeling approach, it appears a common theme of access to appropriate, relevant, and high-quality data is a limitation to all and is likely to continue to be the focus of future research.

Computational evaluation of endocrine activity of biocidal active substances

Exposure to endocrine disrupting chemicals is an important public health concern although only a few endocrine disruption chemicals have been identified so far. To speed up their identification, in silico toxicological models appear to be the most appropriate, since the potential endocrine disruption of a large number of compounds can be estimated in a short time. In this study three in silico models (Endocrine disruptome software, VirtualToxLab and COSMOS KNIME) have been used. In silico predictions of the endocrine disruption potential of biocidal active substances have been made and predictions then compared with the available in vitro experimental binding affinities to androgen, estrogen, glucocorticoid and thyroid receptors. The chosen models had similar accuracies (around 60%), while differences were shown between the models in specificity and sensitivity. VirtualToxLab was the most balanced model. Additionally, three combined models were prepared and evaluated. As expected, the majority rule approach model was more accurate and balanced. However, the positive consensus rule model, that improved the specificity of predictions (≥80% for all studied nuclear receptors) was more applicable. This reduction of false positive predictions is especially useful in the search for positive (active) compounds. On the other hand, the novel negative consensus rule model improved the specificity of prediction (≥80% for all studied nuclear receptors), giving good predictions of negative (inactive) compounds that can be excluded from further testing. The results obtained by these combined models have great added value, since they can significantly reduce further experimental testing.

A ligand-based computational drug repurposing pipeline using KNIME and Programmatic Data Access: case studies for rare diseases and COVID-19

Biomedical information mining is increasingly recognized as a promising technique to accelerate drug discovery and development. Especially, integrative approaches which mine data from several (open) data sources have become more attractive with the increasing possibilities to programmatically access data through Application Programming Interfaces (APIs). The use of open data in conjunction with free, platform-independent analytic tools provides the additional advantage of flexibility, re-usability, and transparency. Here, we present a strategy for performing ligand-based in silico drug repurposing with the analytics platform KNIME. We demonstrate the usefulness of the developed workflow on the basis of two different use cases: a rare disease (here: Glucose Transporter Type 1 (GLUT-1) deficiency), and a new disease (here: COVID 19). The workflow includes a targeted download of data through web services, data curation, detection of enriched structural patterns, as well as substructure searches in DrugBank and a recently deposited data set of antiviral drugs provided by Chemical Abstracts Service. Developed workflows, tutorials with detailed step-by-step instructions, and the information gained by the analysis of data for GLUT-1 deficiency syndrome and COVID-19 are made freely available to the scientific community. The provided framework can be reused by researchers for other in silico drug repurposing projects, and it should serve as a valuable teaching resource for conveying integrative data mining strategies.

In Silico Identification of Chemicals Capable of Binding to the Ecdysone Receptor

The process of molting, known alternatively as ecdysis, is a feature integral in the life cycles of species across the arthropod phylum. Regulation occurs as a function of the interaction of ecdysteroid hormones with the arthropod nuclear ecdysone receptor-a process preceding the triggering of a series of downstream events constituting an endocrine signaling pathway highly conserved throughout environmentally prevalent insect, crustacean, and myriapod organisms. Inappropriate ecdysone receptor binding and activation forms the essential molecular initiating event within possible adverse outcome pathways relating abnormal molting to mortality in arthropods. Definition of the characteristics of chemicals liable to stimulate such activity has the potential to be of great utility in mitigation of hazards posed toward vulnerable species. Thus the aim of the present study was to develop a series of rule-sets, derived from the key structural and physicochemical features associated with identified ecdysone receptor ligands, enabling construction of Konstanz Information Miner (KNIME) workflows permitting the flagging of compounds predisposed to binding at the site. Data describing the activities of 555 distinct chemicals were recovered from a variety of assays across 10 insect species, allowing for formulation of KNIME screens for potential binding activity at the molecular initiating event and adverse outcome level of biological organization. Environ Toxicol Chem 2020;39:1438-1450. © 2020 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

Towards reproducible computational drug discovery

The reproducibility of experiments has been a long standing impediment for further scientific progress. Computational methods have been instrumental in drug discovery efforts owing to its multifaceted utilization for data collection, pre-processing, analysis and inference. This article provides an in-depth coverage on the reproducibility of computational drug discovery. This review explores the following topics: (1) the current state-of-the-art on reproducible research, (2) research documentation (e.g. electronic laboratory notebook, Jupyter notebook, etc.), (3) science of reproducible research (i.e. comparison and contrast with related concepts as replicability, reusability and reliability), (4) model development in computational drug discovery, (5) computational issues on model development and deployment, (6) use case scenarios for streamlining the computational drug discovery protocol. In computational disciplines, it has become common practice to share data and programming codes used for numerical calculations as to not only facilitate reproducibility, but also to foster collaborations (i.e. to drive the project further by introducing new ideas, growing the data, augmenting the code, etc.). It is therefore inevitable that the field of computational drug design would adopt an open approach towards the collection, curation and sharing of data/code.

Using 2D Structural Alerts to Define Chemical Categories for Molecular Initiating Events

Molecular initiating events (MIEs) are important concepts for in silico predictions. They can be used to link chemical characteristics to biological activity through an adverse outcome pathway (AOP). In this work, we capture chemical characteristics in 2D structural alerts, which are then used as models to predict MIEs. An automated procedure has been used to identify these alerts, and the chemical categories they define have been used to provide quantitative predictions for the activity of molecules that contain them. This has been done across a diverse group of 39 important pharmacological human targets using open source data. The alerts for each target combine into a model for that target, and these models are joined into a tool for MIE prediction with high average model performance (sensitivity = 82%, specificity = 93%, overall quality = 93%, Matthews correlation coefficient = 0.57). The result is substantially improved from our previous study (Allen, T. E. H., Goodman, J. M., Gutsell, S., and Russell, P. J. 2016. A history of the molecular initiating event. Chem. Res. Toxicol. 29, 2060-2070) for which the mean sensitivity for each target was only 58%. This tool provides the first step in an AOP-based risk assessment, linking chemical structure to toxicity endpoint.

Avoiding hERG-liability in drug design via synergetic combinations of different (Q)SAR methodologies and data sources: a case study in an industrial setting

In this paper, we explore the impact of combining different in silico prediction approaches and data sources on the predictive performance of the resulting system. We use inhibition of the hERG ion channel target as the endpoint for this study as it constitutes a key safety concern in drug development and a potential cause of attrition. We will show that combining data sources can improve the relevance of the training set in regard of the target chemical space, leading to improved performance. Similarly we will demonstrate that combining multiple statistical models together, and with expert systems, can lead to positive synergistic effects when taking into account the confidence in the predictions of the merged systems. The best combinations analyzed display a good hERG predictivity. Finally, this work demonstrates the suitability of the SOHN methodology for building models in the context of receptor based endpoints like hERG inhibition when using the appropriate pharmacophoric descriptors.

A nanoinformatics decision support tool for the virtual screening of gold nanoparticle cellular association using protein corona fingerprints

The increasing use of nanoparticles (NPs) in a wide range of consumer and industrial applications has necessitated significant effort to address the challenge of characterizing and quantifying the underlying nanostructure - biological response relationships to ensure that these novel materials can be exploited responsibly and safely. Such efforts demand reliable experimental data not only in terms of the biological dose-response, but also regarding the physicochemical properties of the NPs and their interaction with the biological environment. The latter has not been extensively studied, as a large surface to bind biological macromolecules is a unique feature of NPs that is not relevant for chemicals or pharmaceuticals, and thus only limited data have been reported in the literature quantifying the protein corona formed when NPs interact with a biological medium and linking this with NP cellular association/uptake. In this work we report the development of a predictive model for the assessment of the biological response (cellular association, which can include both internalized NPs and those attached to the cell surface) of surface-modified gold NPs, based on their physicochemical properties and protein corona fingerprints, utilizing a dataset of 105 unique NPs. Cellular association was chosen as the end-point for the original experimental study due to its relevance to inflammatory responses, biodistribution, and toxicity in vivo. The validated predictive model is freely available online through the Enalos Cloud Platform ( http://enalos.insilicotox.com/NanoProteinCorona/ ) to be used as part of a regulatory or NP safe-by-design decision support system. This online tool will allow the virtual screening of NPs, based on a list of the significant NP descriptors, identifying those NPs that would warrant further toxicity testing on the basis of predicted NP cellular association.

3D-e-Chem: Structural Cheminformatics Workflows for Computer-Aided Drug Discovery

eScience technologies are needed to process the information available in many heterogeneous types of protein-ligand interaction data and to capture these data into models that enable the design of efficacious and safe medicines. Here we present scientific KNIME tools and workflows that enable the integration of chemical, pharmacological, and structural information for: i) structure-based bioactivity data mapping, ii) structure-based identification of scaffold replacement strategies for ligand design, iii) ligand-based target prediction, iv) protein sequence-based binding site identification and ligand repurposing, and v) structure-based pharmacophore comparison for ligand repurposing across protein families. The modular setup of the workflows and the use of well-established standards allows the re-use of these protocols and facilitates the design of customized computer-aided drug discovery workflows.

Ab initio chemical safety assessment: A workflow based on exposure considerations and non-animal methods

We describe and illustrate a workflow for chemical safety assessment that completely avoids animal testing. The workflow, which was developed within the SEURAT-1 initiative, is designed to be applicable to cosmetic ingredients as well as to other types of chemicals, e.g. active ingredients in plant protection products, biocides or pharmaceuticals. The aim of this work was to develop a workflow to assess chemical safety without relying on any animal testing, but instead constructing a hypothesis based on existing data, in silico modelling, biokinetic considerations and then by targeted non-animal testing. For illustrative purposes, we consider a hypothetical new ingredient x as a new component in a body lotion formulation. The workflow is divided into tiers in which points of departure are established through in vitro testing and in silico prediction, as the basis for estimating a safe external dose in a repeated use scenario. The workflow includes a series of possible exit (decision) points, with increasing levels of confidence, based on the sequential application of the Threshold of Toxicological (TTC) approach, read-across, followed by an "ab initio" assessment, in which chemical safety is determined entirely by new in vitro testing and in vitro to in vivo extrapolation by means of mathematical modelling. We believe that this workflow could be applied as a tool to inform targeted and toxicologically relevant in vitro testing, where necessary, and to gain confidence in safety decision making without the need for animal testing.

In Silico Prediction of Organ Level Toxicity: Linking Chemistry to Adverse Effects

In silico methods to predict toxicity include the use of (Quantitative) Structure-Activity Relationships ((Q)SARs) as well as grouping (category formation) allowing for read-across. A challenging area for in silico modelling is the prediction of chronic toxicity and the No Observed (Adverse) Effect Level (NO(A)EL) in particular. A proposed solution to the prediction of chronic toxicity is to consider organ level effects, as opposed to modelling the NO(A)EL itself. This review has focussed on the use of structural alerts to identify potential liver toxicants. In silico profilers, or groups of structural alerts, have been developed based on mechanisms of action and informed by current knowledge of Adverse Outcome Pathways. These profilers are robust and can be coded computationally to allow for prediction. However, they do not cover all mechanisms or modes of liver toxicity and recommendations for the improvement of these approaches are given.

A Semi-automated Approach to Create Purposeful Mechanistic Datasets from Heterogeneous Data: Data Mining Towards the in silico Predictions for Oestrogen Receptor Modulation and Teratogenicity

The need to find an alternative to costly animal studies for developmental and reproductive toxicity testing has shifted the focus considerably to the assessment of in vitro developmental toxicology models and the exploitation of pharmacological data for relevant molecular initiating events. We hereby demonstrate how automation can be applied successfully to handle heterogeneous oestrogen receptor data from ChEMBL. Applying expert-derived thresholds to specific bioactivities allowed an activity call to be attributed to each data entry. Human intervention further improved this mechanistic dataset which was mined to develop structure-activity relationship alerts and an expert model covering 45 chemical classes for the prediction of oestrogen receptor modulation. The evaluation of the model using FDA EDKB and Tox21 data was quite encouraging. This model can also provide a teratogenicity prediction along with the additional information it provides relevant to the query compound, all of which will require careful assessment of potential risk by experts.

Automated workflows for modelling chemical fate, kinetics and toxicity

Automation is universal in today's society, from operating equipment such as machinery, in factory processes, to self-parking automobile systems. While these examples show the efficiency and effectiveness of automated mechanical processes, automated procedures that support the chemical risk assessment process are still in their infancy. Future human safety assessments will rely increasingly on the use of automated models, such as physiologically based kinetic (PBK) and dynamic models and the virtual cell based assay (VCBA). These biologically-based models will be coupled with chemistry-based prediction models that also automate the generation of key input parameters such as physicochemical properties. The development of automated software tools is an important step in harmonising and expediting the chemical safety assessment process. In this study, we illustrate how the KNIME Analytics Platform can be used to provide a user-friendly graphical interface for these biokinetic models, such as PBK models and VCBA, which simulates the fate of chemicals in vivo within the body and in vitro test systems respectively.

•

The VCBA is a mathematical model that simulates in vitro fate of chemicals and the corresponding cellular effect.

•

The VCBA has been implemented in an open access web-based KNIME platform for ease of use.

•

KNIME Analytics Platform can be used to provide a user-friendly graphical interface for biokinetic models.

Analysis of drug-endogenous human metabolite similarities in terms of their maximum common substructures

In previous work, we have assessed the structural similarities between marketed drugs (‘drugs’) and endogenous natural human metabolites (‘metabolites’ or ‘endogenites’), using ‘fingerprint’ methods in common use, and the Tanimoto and Tversky similarity metrics, finding that the fingerprint encoding used had a dramatic effect on the apparent similarities observed. By contrast, the maximal common substructure (MCS), when the means of determining it is fixed, is a means of determining similarities that is largely independent of the fingerprints, and also has a clear chemical meaning. We here explored the utility of the MCS and metrics derived therefrom. In many cases, a shared scaffold helps cluster drugs and endogenites, and gives insight into enzymes (in particular transporters) that they both share. Tanimoto and Tversky similarities based on the MCS tend to be smaller than those based on the MACCS fingerprint-type encoding, though the converse is also true for a significant fraction of the comparisons. While no single molecular descriptor can account for these differences, a machine learning-based analysis of the nature of the differences (MACCS_Tanimoto vs MCS_Tversky) shows that they are indeed deterministic, although the features that are used in the model to account for this vary greatly with each individual drug. The extent of its utility and interpretability vary with the drug of interest, implying that while MCS is neither ‘better’ nor ‘worse’ for every drug–endogenite comparison, it is sufficiently different to be of value. The overall conclusion is thus that the use of the MCS provides an additional and valuable strategy for understanding the structural basis for similarities between synthetic, marketed drugs and natural intermediary metabolites.

In Silico Identification of Proteins Associated with Drug-induced Liver Injury Based on the Prediction of Drug-target Interactions

Drug-induced liver injury (DILI) is the leading cause of acute liver failure as well as one of the major reasons for drug withdrawal from clinical trials and the market. Elucidation of molecular interactions associated with DILI may help to detect potentially hazardous pharmacological agents at the early stages of drug development. The purpose of our study is to investigate which interactions with specific human protein targets may cause DILI. Prediction of interactions with 1534 human proteins was performed for the dataset with information about 699 drugs, which were divided into three categories of DILI: severe (178 drugs), moderate (310 drugs) and without DILI (211 drugs). Based on the comparison of drug-target interactions predicted for different drugs' categories and interpretation of those results using clustering, Gene Ontology, pathway and gene expression analysis, we identified 61 protein targets associated with DILI. Most of the revealed proteins were linked with hepatocytes' death caused by disruption of vital cellular processes, as well as the emergence of inflammation in the liver. It was found that interaction of a drug with the identified targets is the essential molecular mechanism of the severe DILI for the most of the considered pharmaceuticals. Thus, pharmaceutical agents interacting with many of the identified targets may be considered as candidates for filtering out at the early stages of drug research.

A History of the Molecular Initiating Event

The adverse outcome pathway (AOP) framework provides an alternative to traditional in vivo experiments for the risk assessment of chemicals. AOPs consist of a number of key events (KEs) linked by key event relationships across a range of biological organization backed by scientific evidence. The first KE in the pathway is the molecular initiating event (MIE)-the initial chemical trigger that starts an AOP. Over the past 3 years the AOP conceptual framework has gained a large amount of momentum in toxicology as an alternative to animal methods, and so the MIE has come into the spotlight. What is an MIE? How can MIEs be measured or predicted? What research is currently contributing to our understanding of MIEs? In this Perspective we outline answers to these key questions.

Evaluation of OASIS QSAR Models Using ToxCast™ in Vitro Estrogen and Androgen Receptor Binding Data and Application in an Integrated Endocrine Screening Approach

BACKGROUND

Integrative testing strategies (ITSs) for potential endocrine activity can use tiered in silico and in vitro models. Each component of an ITS should be thoroughly assessed.

OBJECTIVES

We used the data from three in vitro ToxCast™ binding assays to assess OASIS, a quantitative structure-activity relationship (QSAR) platform covering both estrogen receptor (ER) and androgen receptor (AR) binding. For stronger binders (described here as AC50 < 1 μM), we also examined the relationship of QSAR predictions of ER or AR binding to the results from 18 ER and 10 AR transactivation assays, 72 ER-binding reference compounds, and the in vivo uterotrophic assay.

METHODS

NovaScreen binding assay data for ER (human, bovine, and mouse) and AR (human, chimpanzee, and rat) were used to assess the sensitivity, specificity, concordance, and applicability domain of two OASIS QSAR models. The binding strength relative to the QSAR-predicted binding strength was examined for the ER data. The relationship of QSAR predictions of binding to transactivation- and pathway-based assays, as well as to in vivo uterotrophic responses, was examined.

RESULTS

The QSAR models had both high sensitivity (> 75%) and specificity (> 86%) for ER as well as both high sensitivity (92-100%) and specificity (70-81%) for AR. For compounds within the domains of the ER and AR QSAR models that bound with AC50 < 1 μM, the QSAR models accurately predicted the binding for the parent compounds. The parent compounds were active in all transactivation assays where metabolism was incorporated and, except for those compounds known to require metabolism to manifest activity, all assay platforms where metabolism was not incorporated. Compounds in-domain and predicted to bind by the ER QSAR model that were positive in ToxCast™ ER binding at AC50 < 1 μM were active in the uterotrophic assay.

CONCLUSIONS

We used the extensive ToxCast™ HTS binding data set to show that OASIS ER and AR QSAR models had high sensitivity and specificity when compounds were in-domain of the models. Based on this research, we recommend a tiered screening approach wherein a) QSAR is used to identify compounds in-domain of the ER or AR binding models and predicted to bind; b) those compounds are screened in vitro to assess binding potency; and c) the stronger binders (AC50 < 1 μM) are screened in vivo. This scheme prioritizes compounds for integrative testing and risk assessment. Importantly, compounds that are not in-domain, that are predicted either not to bind or to bind weakly, that are not active in in vitro, that require metabolism to manifest activity, or for which in vivo AR testing is in order, need to be assessed differently.

CITATION

Bhhatarai B, Wilson DM, Price PS, Marty S, Parks AK, Carney E. 2016. Evaluation of OASIS QSAR models using ToxCast™ in vitro estrogen and androgen receptor binding data and application in an integrated endocrine screening approach. Environ Health Perspect 124:1453-1461; http://dx.doi.org/10.1289/EHP184.

A semi-automated, KNIME-based workflow for biofilm assays

Background

A current focus of biofilm research is the chemical interaction between microorganisms within the biofilms. Prerequisites for this research are bioassay systems which integrate reliable tools for the planning of experiments with robot-assisted measurements and with rapid data processing. Here, data structures that are both human- and machine readable may be particularly useful.

Results

In this report, we present several simplification and robotisation options for an assay of bacteria-induced biofilm formation by the freshwater diatom Achnanthidium minutissimum. We also tested several proof-of-concept robotisation methods for pipetting, as well as for measuring the biofilm absorbance directly in the multi-well plates. Furthermore, we exemplify the implementation of an improved data processing workflow for this assay using the Konstanz Information Miner (KNIME), a free and open source data analysis environment. The workflow integrates experiment planning files and absorbance read-out data, towards their automated processing for analysis.

Conclusions

Our workflow lead to a substantial reduction of the measurement and data processing workload, while still reproducing previously obtained results in the A. minutissimum biofilm assay. The methods, scripts and files we designed are described here, offering adaptable options for other medium-throughput biofilm screenings.

Electronic supplementary material

The online version of this article (doi:10.1186/s12866-016-0676-9) contains supplementary material, which is available to authorized users.

In Silico Models for Hepatotoxicity

In this chapter we review the challenges of predicting human hepatotoxicity. Principally, this is our partial understanding of a very complex biochemical system and our ability to emulate that in a predictive capacity. We give an overview of the published modeling approaches in this area to date and discuss their design, strengths, and weaknesses. It is interesting to note the shift during the period of this review in the direction of evidenced-based approaches including structural alerts and pharmacophore models. Proposals on how best to utilize the data emerging from modeling studies are also discussed.