An ensemble machine learning model generates a focused screening library for the identification of CDK8 inhibitors.

The identification of an effective inhibitor is an important starting step in drug development. Unfortunately, many issues such as the characterization of protein binding sites, the screening library, materials for assays, etc., make drug screening a difficult proposition. As the size of screening libraries increases, more resources will be inefficiently consumed. Thus, new strategies are needed to preprocess and focus a screening library towards a targeted protein. Herein, we report an ensemble machine learning (ML) model to generate a CDK8-focused screening library. The ensemble model consists of six different algorithms optimized for CDK8 inhibitor classification. The models were trained using a CDK8-specific fragment library along with molecules containing CDK8 activity. The optimized ensemble model processed a commercial library containing 1.6 million molecules. This resulted in a CDK8-focused screening library containing 1,672 molecules, a reduction of more than 99.90%. The CDK8-focused library was then subjected to molecular docking, and 25 candidate compounds were selected. Enzymatic assays confirmed six CDK8 inhibitors, with one compound producing an IC50 value of ≤100 nM. Analysis of the ensemble ML model reveals the role of the CDK8 fragment library during training. Structural analysis of molecules reveals the hit compounds to be structurally novel CDK8 inhibitors. Together, the results highlight a pipeline for curating a focused library for a specific protein target, such as CDK8.

Interlaboratory Study on Zebrafish in Toxicology: Systematic Evaluation of the Application of Zebrafish in Toxicology's (SEAZIT's) Evaluation of Developmental Toxicity.

Embryonic zebrafish represent a useful test system to screen substances for their ability to perturb development. The exposure scenarios, endpoints captured, and data analysis vary among the laboratories who conduct screening. A lack of harmonization impedes the comparison of the substance potency and toxicity outcomes across laboratories and may hinder the broader adoption of this model for regulatory use. The Systematic Evaluation of the Application of Zebrafish in Toxicology (SEAZIT) initiative was developed to investigate the sources of variability in toxicity testing. This initiative involved an interlaboratory study to determine whether experimental parameters altered the developmental toxicity of a set of 42 substances (3 tested in duplicate) in three diverse laboratories. An initial dose-range-finding study using in-house protocols was followed by a definitive study using four experimental conditions: chorion-on and chorion-off using both static and static renewal exposures. We observed reasonable agreement across the three laboratories as 33 of 42 test substances (78.6%) had the same activity call. However, the differences in potency seen using variable in-house protocols emphasizes the importance of harmonization of the exposure variables under evaluation in the second phase of this study. The outcome of the Def will facilitate future practical discussions on harmonization within the zebrafish research community.

Corrigendum: Tox21Enricher-Shiny: an R Shiny application for toxicity functional annotation analysis.

[This corrects the article DOI: 10.3389/ftox.2023.1147608.].

Tox21Enricher-Shiny: an R Shiny application for toxicity functional annotation analysis.

Inference of toxicological and mechanistic properties of untested chemicals through structural or biological similarity is a commonly employed approach for initial chemical characterization and hypothesis generation. We previously developed a web-based application, Tox21Enricher-Grails, on the Grails framework that identifies enriched biological/toxicological properties of chemical sets for the purpose of inferring properties of untested chemicals within the set. It was able to detect significantly overrepresented biological (e.g., receptor binding), toxicological (e.g., carcinogenicity), and chemical (e.g., toxicologically relevant chemical substructures) annotations within sets of chemicals screened in the Tox21 platform. Here, we present an R Shiny application version of Tox21Enricher-Grails, Tox21Enricher-Shiny, with more robust features and updated annotations. Tox21Enricher-Shiny allows users to interact with the web application component (available at http://hurlab.med.und.edu/Tox21Enricher/) through a user-friendly graphical user interface or to directly access the application's functions through an application programming interface. This version now supports InChI strings as input in addition to CASRN and SMILES identifiers. Input chemicals that contain certain reactive functional groups (nitrile, aldehyde, epoxide, and isocyanate groups) may react with proteins in cell-based Tox21 assays: this could cause Tox21Enricher-Shiny to produce spurious enrichment analysis results. Therefore, this version of the application can now automatically detect and ignore such problematic chemicals in a user's input. The application also offers new data visualizations, and the architecture has been greatly simplified to allow for simple deployment, version control, and porting. The application may be deployed onto a Posit Connect or Shiny server, and it uses Postgres for database management. As other Tox21-related tools are being migrated to the R Shiny platform, the development of Tox21Enricher-Shiny is a logical transition to use R's strong data analysis and visualization capacities and to provide aesthetic and developmental consistency with other Tox21 applications developed by the Division of Translational Toxicology (DTT) at the National Institute of Environmental Health Sciences (NIEHS).

Systematic Evaluation of the Application of Zebrafish in Toxicology (SEAZIT): Developing a Data Analysis Pipeline for the Assessment of Developmental Toxicity with an Interlaboratory Study.

The embryonic zebrafish is a useful vertebrate model for assessing the effects of substances on growth and development. However, cross-laboratory developmental toxicity outcomes can vary and reported developmental defects in zebrafish may not be directly comparable between laboratories. To address these limitations for gaining broader adoption of the zebrafish model for toxicological screening, we established the Systematic Evaluation of the Application of Zebrafish in Toxicology (SEAZIT) program to investigate how experimental protocol differences can influence chemical-mediated effects on developmental toxicity (i.e., mortality and the incidence of altered phenotypes). As part of SEAZIT, three laboratories were provided a common and blinded dataset (42 substances) to evaluate substance-mediated effects on developmental toxicity in the embryonic zebrafish model. To facilitate cross-laboratory comparisons, all the raw experimental data were collected, stored in a relational database, and analyzed with a uniform data analysis pipeline. Due to variances in laboratory-specific terminology for altered phenotypes, we utilized ontology terms available from the Ontology Lookup Service (OLS) for Zebrafish Phenotype to enable additional cross-laboratory comparisons. In this manuscript, we utilized data from the first phase of screening (dose range finding, DRF) to highlight the methodology associated with the development of the database and data analysis pipeline, as well as zebrafish phenotype ontology mapping.

Establishment of extensive artificial intelligence models for kinase inhibitor prediction: Identification of novel PDGFRB inhibitors.

Identifying hit compounds is an important step in drug development. Unfortunately, this process continues to be a challenging task. Several machine learning models have been generated to aid in simplifying and improving the prediction of candidate compounds. Models tuned for predicting kinase inhibitors have been established. However, an effective model can be limited by the size of the chosen training dataset. In this study, we tested several machine learning models to predict potential kinase inhibitors. A dataset was curated from a number of publicly available repositories. This resulted in a comprehensive dataset covering more than half of the human kinome. More than 2,000 kinase models were established using different model approaches. The performances of the models were compared, and the Keras-MLP model was determined to be the best performing model. The model was then used to screen a chemical library for potential inhibitors targeting platelet-derived growth factor receptor-ß (PDGFRB). Several PDGFRB candidates were selected, and in vitro assays confirmed four compounds with PDGFRB inhibitory activity and IC50 values in the nanomolar range. These results show the effectiveness of machine learning models trained on the reported dataset. This report would aid in the establishment of machine learning models as well as in the discovery of novel kinase inhibitors.

Differences in neurotoxic outcomes of organophosphorus pesticides revealed via multi-dimensional screening in adult and regenerating planarians.

Organophosphorus pesticides (OPs) are a chemically diverse class of commonly used insecticides. Epidemiological studies suggest that low dose chronic prenatal and infant exposures can lead to life-long neurological damage and behavioral disorders. While inhibition of acetylcholinesterase (AChE) is the shared mechanism of acute OP neurotoxicity, OP-induced developmental neurotoxicity (DNT) can occur independently and/or in the absence of significant AChE inhibition, implying that OPs affect alternative targets. Moreover, different OPs can cause different adverse outcomes, suggesting that different OPs act through different mechanisms. These findings emphasize the importance of comparative studies of OP toxicity. Freshwater planarians are an invertebrate system that uniquely allows for automated, rapid and inexpensive testing of adult and developing organisms in parallel to differentiate neurotoxicity from DNT. Effects found only in regenerating planarians would be indicative of DNT, whereas shared effects may represent neurotoxicity. We leverage this unique feature of planarians to investigate potential differential effects of OPs on the adult and developing brain by performing a comparative screen to test 7 OPs (acephate, chlorpyrifos, dichlorvos, diazinon, malathion, parathion and profenofos) across 10 concentrations in quarter-log steps. Neurotoxicity was evaluated using a wide range of quantitative morphological and behavioral readouts. AChE activity was measured using an Ellman assay. The toxicological profiles of the 7 OPs differed across the OPs and between adult and regenerating planarians. Toxicological profiles were not correlated with levels of AChE inhibition. Twenty-two "mechanistic control compounds" known to target pathways suggested in the literature to be affected by OPs (cholinergic neurotransmission, serotonin neurotransmission, endocannabinoid system, cytoskeleton, adenyl cyclase and oxidative stress) and 2 negative controls were also screened. When compared with the mechanistic control compounds, the phenotypic profiles of the different OPs separated into distinct clusters. The phenotypic profiles of adult vs. regenerating planarians exposed to the OPs clustered differently, suggesting some developmental-specific mechanisms. These results further support findings in other systems that OPs cause different adverse outcomes in the (developing) brain and build the foundation for future comparative studies focused on delineating the mechanisms of OP neurotoxicity in planarians.

Exploring the Influence of Experimental Design on Toxicity Outcomes in Zebrafish Embryo Tests.

Compound toxicity data obtained from independent zebrafish laboratories can vary vastly, complicating the use of zebrafish screening for regulatory decisions. Differences in the assay protocol parameters are the primary source of variability. We investigated this issue by utilizing data from the NTP DNT-DIVER database (https://doi.org/10.22427/NTP-DATA-002-00062-0001-0000-1, last accessed June 2, 2022), which consists of data from zebrafish developmental toxicity (devtox) and locomotor response (designated as "neurotox") screens from 3 independent laboratories, using the same set of 87 compounds. The data were analyzed using the benchmark concentration (BMC) modeling approach, which estimates the concentration of interest based on a predetermined response threshold. We compared the BMC results from 3 laboratories (A, B, C) in 3 toxicity outcome categories: mortality, cumulative devtox, and neurotox, in terms of activity calls and potency values. We found that for devtox screening, laboratories with similar/same protocol parameters (B vs C) had an active call concordance as high as 86% with negligible potency difference. For neurotox screening, active call concordances between paired laboratories are lower than devtox screening (highest 68%). When protocols with different protocol parameters were compared, the concordance dropped, and the potency shift was on average about 3.8-fold for the cumulative devtox outcome and 5.8-fold for the neurotox outcome. The potential contributing protocol parameters for potency shift are listed or ranked. This study provides a quantitative assessment of the source of variability in zebrafish screening protocols and sets the groundwork for the ongoing Systematic Evaluation of the Application of Zebrafish in Toxicology effort at the National Toxicology Program.

Accounting for Artifacts in High-Throughput Toxicity Assays.

Compound activity identification is the primary goal in high throughput screening (HTS) assays. However, assay artifacts including both systematic (e.g., compound autofluorescence) and nonsystematic (e.g., noise) complicate activity interpretation. In addition, other than the traditional potency parameter, half-maximal effect concentration [EC50], additional activity parameters (e.g., point-of-departure [POD] and weighted area-under-the-curve [wAUC]) could be derived from HTS data for activity profiling. A data analysis pipeline has been developed to handle the artifacts, and to provide compound activity characterization with either binary or continuous metrics. This chapter outlines the steps in the pipeline using Tox21 estrogen receptor (ER) ß-lactamase assays, including the formats to identify either agonists or antagonists, as well as the counterscreen assays for identifying artifacts as examples. The steps can be applied to other lower throughput assays with concentration-response data.

Identification of a dual FLT3 and MNK2 inhibitor for acute myeloid leukemia treatment using a structure-based virtual screening approach.

Fms-like tyrosine kinase 3 (FLT3) is considered a promising therapeutic target for acute myeloid leukemia (AML) in the clinical. However, monotherapy with FLT3 inhibitor is usually accompanied by drug resistance. Dual inhibitors might be therapeutically beneficial to patients with AML due to their ability to overcome drug resistance. Mitogen-activated protein kinase (MAPK)-interacting kinases (MNKs) phosphorylate eukaryotic translation initiation factor 4E (eIF4E), which brings together the RAS/RAF/ERK and PI3K/AKT/mTOR oncogenic pathways. Therefore, dual inhibition of FLT3 and MNK2 might have an additive effect against AML. Herein, a structure-based virtual screening approach was performed to identify dual inhibitors of FLT3 and MNK2 from the ChemDiv database. Compound K783-0308 was identified as a dual inhibitor of FLT3 and MNK2 with IC50 values of 680 and 406 nM, respectively. In addition, the compound showed selectivity for both FLT3 and MNK2 in a panel of 82 kinases. The structure-activity relationship analysis and common interactions revealed interactions between K783-0308 analogs and FLT3 and MNK2. Furthermore, K783-0308 inhibited MV-4-11 and MOLM-13 AML cell growth and induced G0/G1 cell cycle arrest. Taken together, the dual inhibitor K783-0308 showed promising results and can be potentially optimized as a lead compound for AML treatment.

Structure-based virtual screening and biological evaluation of novel small-molecule BTK inhibitors.

Bruton tyrosine kinase (BTK) is linked to multiple signalling pathways that regulate cellular survival, activation, and proliferation. A covalent BTK inhibitor has shown favourable outcomes for treating B cell malignant leukaemia. However, covalent inhibitors require a high reactive warhead that may contribute to unexpected toxicity, poor selectivity, or reduced effectiveness in solid tumours. Herein, we report the identification of a novel noncovalent BTK inhibitor. The binding interactions (i.e. interactions from known BTK inhibitors) for the BTK binding site were identified and incorporated into a structure-based virtual screening (SBVS). Top-rank compounds were selected and testing revealed a BTK inhibitor with >50% inhibition at 10 µM concentration. Examining analogues revealed further BTK inhibitors. When tested across solid tumour cell lines, one inhibitor showed favourable inhibitory activity, suggesting its potential for targeting BTK malignant tumours. This inhibitor could serve as a basis for developing an effective BTK inhibitor targeting solid cancers.

Identification and analysis of a selective DYRK1A inhibitor.

The dysregulation of DYRK1A is implicated in many diseases such as cancer, diabetes, and neurodegenerative diseases. Alzheimer's disease is one of the most common neurodegenerative disease and has elevated interest in DYRK1A research. Overexpression of DYRK1A has been linked to the formation of tau aggregates. Currently, an effective therapeutic treatment that targets DYRK1A is lacking. A specific small-molecule inhibitor would further our understanding of the physiological role of DYRK1A in neurodegenerative diseases and could be presented as a possible therapeutic option. In this study, we identified pharmacological interactions within the DYRK1A active site and performed a structure-based virtual screening approach to identify a selective small-molecule inhibitor. Several compounds were selected in silico for enzymatic and cellular assays, yielding a novel inhibitor. A structure-activity relationship analysis was performed to identify areas of interactions for the compounds selected in this study. When tested in vitro, reduction of DYRK1A dependent phosphorylation of tau was observed for active compounds. The active compounds also improved tau turbidity, suggesting that these compounds could alleviate aberrant tau aggregation. Testing the active compound against a panel of kinases across the kinome revealed greater selectivity towards DYRK1A. Our study demonstrates a serviceable protocol that identified a novel and selective DYRK1A inhibitor with potential for further study in tau-related pathologies.

Discovery of a novel cyclin-dependent kinase 8 inhibitor with an oxindole core for anti-inflammatory treatment.

Chronic inflammation is an underlying cause in a number of diseases. Cyclin-dependent kinase 8 (CDK8) has been implicated as an inflammatory mediator, indicating its potential as an anti-inflammatory target. Herein, we performed structure-based virtual screening (SBVS) to identify novel CDK8 inhibitors. The pharmacological interactions for CDK8 were identified and incorporated into a SBVS protocol. Selected compounds were tested in enzymatic assays, and one compound was confirmed to be a CDK8 inhibitor with a 50% inhibitory concentration (IC50) value of 1684.4 nM. Comparing structural analogs identified a compound, F059-1017, with greater potency (IC50 558.1 nM). When tested in cell lines, the compounds displayed low cytotoxicity. Cellular assays revealed that the identified CDK8 inhibitors can reduce phosphorylation and expression of signaling mediators associated with inflammation. In addition, results of kinase profiling showed that compound F059-1017 is selective towards CDK8. These findings suggest that the new inhibitors have great potential as lead compounds for developing novel anti-inflammatory therapeutics.

Predictive modeling of biological responses in the rat liver using in vitro Tox21 bioactivity: Benefits from high-throughput toxicokinetics.

Computational methods are needed to more efficiently leverage data from in vitro cell-based models to predict what occurs within whole body systems after chemical insults. This study set out to test the hypothesis that in vitro high-throughput screening (HTS) data can more effectively predict in vivo biological responses when chemical disposition and toxicokinetic (TK) modeling are employed. In vitro HTS data from the Tox21 consortium were analyzed in concert with chemical disposition modeling to derive nominal, aqueous, and intracellular estimates of concentrations eliciting 50% maximal activity. In vivo biological responses were captured using rat liver transcriptomic data from the DrugMatrix and TG-Gates databases and evaluated for pathway enrichment. In vivo dosing data were translated to equivalent body concentrations using HTTK modeling. Random forest models were then trained and tested to predict in vivo pathway-level activity across 221 chemicals using in vitro bioactivity data and physicochemical properties as predictor variables, incorporating methods to address imbalanced training data resulting from high instances of inactivity. Model performance was quantified using the area under the receiver operator characteristic curve (AUC-ROC) and compared across pathways for different combinations of predictor variables. All models that included toxicokinetics were found to outperform those that excluded toxicokinetics. Biological interpretation of the model features revealed that rather than a direct mapping of in vitro assays to in vivo pathways, unexpected combinations of multiple in vitro assays predicted in vivo pathway-level activities. To demonstrate the utility of these findings, the highest-performing model was leveraged to make new predictions of in vivo biological responses across all biological pathways for remaining chemicals tested in Tox21 with adequate data coverage (n = 6617). These results demonstrate that, when chemical disposition and toxicokinetics are carefully considered, in vitro HT screening data can be used to effectively predict in vivo biological responses to chemicals.

Profiling the Tox21 Chemical Collection for Acetylcholinesterase Inhibition.

BACKGROUND: Inhibition of acetylcholinesterase (AChE), a biomarker of organophosphorous and carbamate exposure in environmental and occupational human health, has been commonly used to identify potential safety liabilities. So far, many environmental chemicals, including drug candidates, food additives, and industrial chemicals, have not been thoroughly evaluated for their inhibitory effects on AChE activity. AChE inhibitors can have therapeutic applications (e.g., tacrine and donepezil) or neurotoxic consequences (e.g., insecticides and nerve agents). OBJECTIVES: The objective of the current study was to identify environmental chemicals that inhibit AChE activity using in vitro and in silico models. METHODS: To identify AChE inhibitors rapidly and efficiently, we have screened the Toxicology in the 21st Century (Tox21) 10K compound library in a quantitative high-throughput screening (qHTS) platform by using the homogenous cell-based AChE inhibition assay and enzyme-based AChE inhibition assays (with or without microsomes). AChE inhibitors identified from the primary screening were further tested in monolayer or spheroid formed by SH-SY5Y and neural stem cell models. The inhibition and binding modes of these identified compounds were studied with time-dependent enzyme-based AChE inhibition assay and molecular docking, respectively. RESULTS: A group of known AChE inhibitors, such as donepezil, ambenonium dichloride, and tacrine hydrochloride, as well as many previously unreported AChE inhibitors, such as chelerythrine chloride and cilostazol, were identified in this study. Many of these compounds, such as pyrazophos, phosalone, and triazophos, needed metabolic activation. This study identified both reversible (e.g., donepezil and tacrine) and irreversible inhibitors (e.g., chlorpyrifos and bromophos-ethyl). Molecular docking analyses were performed to explain the relative inhibitory potency of selected compounds. CONCLUSIONS: Our tiered qHTS approach allowed us to generate a robust and reliable data set to evaluate large sets of environmental compounds for their AChE inhibitory activity. https://doi.org/10.1289/EHP6993.

A Gene Expression Biomarker Identifies Chemical Modulators of Estrogen Receptor α in an MCF-7 Microarray Compendium.

Identification of chemicals that affect hormone-regulated systems will help to predict endocrine disruption. In our previous study, a 46 gene biomarker was found to be an accurate predictor of estrogen receptor (ER) α modulation in chemically treated MCF-7 cells. Here, potential ERα modulators were identified using the biomarker by screening a microarray compendium consisting of ∼1600 gene expression comparisons representing exposure to ∼1200 chemicals. A total of ∼170 chemicals were identified as potential ERα modulators. In the Connectivity Map 2.0 collection, 75 and 39 chemicals were predicted to activate or suppress ERα, and they included 12 and six known ERα agonists and antagonists/selective ERα modulators, respectively. Nineteen and eight of the total number were also identified as active in an ERα transactivation assay carried out in an MCF-7-derived cell line used to screen the Tox21 10K chemical library in agonist or antagonist modes, respectively. Chemicals predicted to modulate ERα in MCF-7 cells were examined further using global and targeted gene expression in wild-type and ERα-null cells, transactivation assays, and cell-free ERα coregulator interaction assays. Environmental chemicals classified as weak and very weak agonists were confirmed to activate ERα including apigenin, kaempferol, and oxybenzone. Novel activators included digoxin, nabumetone, ivermectin, and six progestins. Novel suppressors included emetine, mifepristone, niclosamide, and proscillaridin. Our strategy will be useful to identify environmentally relevant ERα modulators in future high-throughput transcriptomic screens.

In silico approaches in organ toxicity hazard assessment: current status and future needs in predicting liver toxicity.

Hepatotoxicity is one of the most frequently observed adverse effects resulting from exposure to a xenobiotic. For example, in pharmaceutical research and development it is one of the major reasons for drug withdrawals, clinical failures, and discontinuation of drug candidates. The development of faster and cheaper methods to assess hepatotoxicity that are both more sustainable and more informative is critically needed. The biological mechanisms and processes underpinning hepatotoxicity are summarized and experimental approaches to support the prediction of hepatotoxicity are described, including toxicokinetic considerations. The paper describes the increasingly important role of in silico approaches and highlights challenges to the adoption of these methods including the lack of a commonly agreed upon protocol for performing such an assessment and the need for in silico solutions that take dose into consideration. A proposed framework for the integration of in silico and experimental information is provided along with a case study describing how computational methods have been used to successfully respond to a regulatory question concerning non-genotoxic impurities in chemically synthesized pharmaceuticals.

In silico approaches in organ toxicity hazard assessment: Current status and future needs for predicting heart, kidney and lung toxicities.

The kidneys, heart and lungs are vital organ systems evaluated as part of acute or chronic toxicity assessments. New methodologies are being developed to predict these adverse effects based on in vitro and in silico approaches. This paper reviews the current state of the art in predicting these organ toxicities. It outlines the biological basis, processes and endpoints for kidney toxicity, pulmonary toxicity, respiratory irritation and sensitization as well as functional and structural cardiac toxicities. The review also covers current experimental approaches, including off-target panels from secondary pharmacology batteries. Current in silico approaches for prediction of these effects and mechanisms are described as well as obstacles to the use of in silico methods. Ultimately, a commonly accepted protocol for performing such assessment would be a valuable resource to expand the use of such approaches across different regulatory and industrial applications. However, a number of factors impede their widespread deployment including a lack of a comprehensive mechanistic understanding, limited in vitro testing approaches and limited in vivo databases suitable for modeling, a limited understanding of how to incorporate absorption, distribution, metabolism, and excretion (ADME) considerations into the overall process, a lack of in silico models designed to predict a safe dose and an accepted framework for organizing the key characteristics of these organ toxicants.

Harnessing In Silico, In Vitro, and In Vivo Data to Understand the Toxicity Landscape of Polycyclic Aromatic Compounds (PACs).

Polycyclic aromatic compounds (PACs) are compounds with a minimum of two six-atom aromatic fused rings. PACs arise from incomplete combustion or thermal decomposition of organic matter and are ubiquitous in the environment. Within PACs, carcinogenicity is generally regarded to be the most important public health concern. However, toxicity in other systems (reproductive and developmental toxicity, immunotoxicity) has also been reported. Despite the large number of PACs identified in the environment, research attention to understand exposure and health effects of PACs has focused on a relatively limited subset, namely polycyclic aromatic hydrocarbons (PAHs), the PACs with only carbon and hydrogen atoms. To triage the rest of the vast number of PACs for more resource-intensive testing, we developed a data-driven approach to contextualize hazard characterization of PACs, by leveraging the available data from various data streams (in silico toxicity, in vitro activity, structural fingerprints, and in vivo data availability). The PACs were clustered on the basis of their in silico toxicity profiles containing predictions from 8 different categories (carcinogenicity, cardiotoxicity, developmental toxicity, genotoxicity, hepatotoxicity, neurotoxicity, reproductive toxicity, and urinary toxicity). We found that PACs with the same parent structure (e.g., fluorene) could have diverse in silico toxicity profiles. In contrast, PACs with similar substituted groups (e.g., alkylated-PAHs) or heterocyclics (e.g., N-PACs) with varying ring sizes could have similar in silico toxicity profiles, suggesting that these groups are better candidates for toxicity read-across analysis. The clusters/regions associated with certain in silico toxicity, in vitro activity, and structural fingerprints were identified. We found that genotoxicity/carcinogenicity (in silico toxicity) and xenobiotic homeostasis and stress response (in vitro activity), respectively, dominate the toxicity/activity variation seen in the PACs. The "hot spots" with enriched toxicity/activity in conjunction with availability of in vivo carcinogenicity data revealed regions of either data-poor (hydroxylated-PAHs) or data-rich (unsubstituted, parent PAHs) PACs. These regions offer potential targets for prioritization of further in vivo assessment and for chemical read-across efforts. The analysis results are searchable through an interactive web application (https://ntp.niehs.nih.gov/go/pacs_tableau), allowing for alternative hypothesis generation.

Pose Filter-Based Ensemble Learning Enables Discovery of Orally Active, Nonsteroidal Farnesoid X Receptor Agonists.

Farnesoid X receptor (FXR) agonists can reverse dysregulated bile acid metabolism, and thus, they are potential therapeutics to prevent and treat nonalcoholic fatty liver disease. The low success rate of FXR agonists' R&D and the side effects of clinical candidates such as obeticholic acid make it urgent to discover new chemotypes. Unfortunately, structure-based virtual screening (SBVS) that can speed up drug discovery has rarely been reported with success for FXR, which was likely hindered by the failure in addressing protein flexibility. To address this issue, we devised human FXR (hFXR)-specific ensemble learning models based on pose filters from 24 agonist-bound hFXR crystal structures and coupled them to traditional SBVS approaches of the FRED docking plus Chemgauss4 scoring function. It turned out that the hFXR-specific pose filter ensemble (PFE) was able to improve ligand enrichment significantly, which rendered 3RUT-based SBVS with its PFE the ideal approach for FXR agonist discovery. By screening of the Specs chemical library and in vitro FXR transactivation bioassay, we identified a new class of FXR agonists with compound XJ034 as the representative, which would have been missed if the PFE was not coupled. Following that, we performed in-depth biological studies which demonstrated that XJ034 resulted in a downtrend of intracellular triglyceride in vitro, significantly decreased the serum/liver TG in high fat diet-induced C57BL/6J obese mice, and more importantly, showed metabolic stabilities in both plasma and liver microsomes. To provide insight into further structure-based lead optimization, we solved the crystal structure of hFXR complexed with compound XJ034, uncovering a unique hydrogen bond between compound XJ034 and residue Y375. The current work highlights the power of our pose filter-based ensemble learning approach in terms of scaffold hopping and provides a promising lead compound for further development.