Data-driven identification of structural alerts for mitigating the risk of drug-induced human liver injuries

Hybrid non-animal modeling: A mechanistic approach to predict chemical hepatotoxicity

Developing mechanistic non-animal testing methods based on the adverse outcome pathway (AOP) framework must incorporate molecular and cellular key events associated with target toxicity. Using data from an in vitro assay and chemical structures, we aimed to create a hybrid model to predict hepatotoxicants. We first curated a reference dataset of 869 compounds for hepatotoxicity modeling. Then, we profiled them against PubChem for existing in vitro toxicity data. Of the 2560 resulting assays, we selected the mitochondrial membrane potential (MMP) assay, a high-throughput screening (HTS) tool that can test chemical disruptors for mitochondrial function. Machine learning was applied to develop quantitative structure-activity relationship (QSAR) models with 2536 compounds tested in the MMP assay for screening new compounds. The MMP assay results, including QSAR model outputs, yielded hepatotoxicity predictions for reference set compounds with a Correct Classification Ratio (CCR) of 0.59. The predictivity improved by including 37 structural alerts (CCR = 0.8). We validated our model by testing 37 reference set compounds in human HepG2 hepatoma cells, and reliably predicting them for hepatotoxicity (CCR = 0.79). This study introduces a novel AOP modeling strategy that combines public HTS data, computational modeling, and experimental testing to predict chemical hepatotoxicity.

Design of OASIS 1 and 2: phase 3 clinical trials assessing the efficacy and safety of elinzanetant for the treatment of vasomotor symptoms associated with menopause

OBJECTIVE

Elinzanetant is a selective neurokinin-1,3 receptor antagonist in development for the treatment of vasomotor symptoms (VMS) associated with menopause. The pivotal, double-blind, randomized, placebo-controlled phase 3 studies Overall Assessment of efficacy and Safety of elinzanetant In patients with vasomotor Symptoms (OASIS) 1 and 2 will assess the efficacy and safety of elinzanetant in women with VMS.

METHODS

The OASIS 1 and 2 pivotal studies are designed in accordance with regulatory guidance. Postmenopausal women with moderate/severe VMS are randomized to receive 120 mg elinzanetant or placebo once daily for 12 weeks, followed by a 14-week active treatment extension. Primary endpoints are the mean change in frequency and severity of moderate/severe VMS from baseline to weeks 4 and 12. Key secondary endpoints will assess the onset of action and effects on sleep disturbance and menopause-related quality of life. Primary and key secondary endpoints will be analyzed using a mixed model with repeated measures. Feedback from postmenopausal women with VMS was used during protocol development.

RESULTS

Women confirmed the relevance of endpoints that assess the impact of VMS, sleep disturbance, and mood changes, and the need for new nonhormone treatments. Educational materials around study design, conduct and expected assessments and procedures were developed based on questions and concerns raised by women.

CONCLUSIONS

The OASIS 1 and 2 pivotal phase 3 studies will enable assessment of the efficacy and safety of elinzanetant as a treatment for VMS, together with its effect on sleep disturbances, depressive symptoms, and menopause-related quality of life. Feedback from postmenopausal women with VMS was used to maximize patient centricity in the trials.

GeoDILI: A Robust and Interpretable Model for Drug-Induced Liver Injury Prediction Using Graph Neural Network-Based Molecular Geometric Representation

Drug-induced liver injury (DILI) is a significant cause of drug failure and withdrawal due to liver damage. Accurate prediction of hepatotoxic compounds is crucial for safe drug development. Several DILI prediction models have been published, but they are built on different data sets, making it difficult to compare model performance. Moreover, most existing models are based on molecular fingerprints or descriptors, neglecting molecular geometric properties and lacking interpretability. To address these limitations, we developed GeoDILI, an interpretable graph neural network that uses a molecular geometric representation. First, we utilized a geometry-based pretrained molecular representation and optimized it on the DILI data set to improve predictive performance. Second, we leveraged gradient information to obtain high-precision atomic-level weights and deduce the dominant substructure. We benchmarked GeoDILI against recently published DILI prediction models, as well as popular GNN models and fingerprint-based machine learning models using the same data set, showing superior predictive performance of our proposed model. We applied the interpretable method in the DILI data set and derived seven precise and mechanistically elucidated structural alerts. Overall, GeoDILI provides a promising approach for accurate and interpretable DILI prediction with potential applications in drug discovery and safety assessment. The data and source code are available at GitHub repository (https://github.com/CSU-QJY/GeoDILI).

ADMET property prediction via multi-task graph learning under adaptive auxiliary task selection

It is a critical step in lead optimization to evaluate the absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties of drug-like compounds. Classical single-task learning (STL) has effectively predicted individual ADMET endpoints with abundant labels. Conversely, multi-task learning (MTL) can predict multiple ADMET endpoints with fewer labels, but ensuring task synergy and highlighting key molecular substructures remain challenges. To tackle these issues, this work elaborates a multi-task graph learning framework for predicting multiple ADMET properties of drug-like small molecules (MTGL-ADMET) by holding a new paradigm of MTL, "one primary, multiple auxiliaries." It first adeptly combines status theory with maximum flow for auxiliary task selection. The subsequent phase introduces a primary-task-centric MTL model with integrated modules. MTGL-ADMET not only outstrips existing STL and MTL methods but also offers a transparent lens into crucial molecular substructures. It is anticipated that this work can promote lead compound finding and optimization in drug discovery.

Interpretable Molecular Property Predictions Using Marginalized Graph Kernels

Marginalized graph kernels have shown competitive performance in molecular machine learning tasks but currently lack measures of interpretability, which are important to improve trust in the models, detect biases, and inform molecular optimization campaigns. We here conceive and implement two interpretability measures for Gaussian process regression using a marginalized graph kernel (GPR-MGK) to quantify (1) the contribution of specific training data to the prediction and (2) the contribution of specific nodes of the graph to the prediction. We demonstrate the applicability of these interpretability measures for molecular property prediction. We compare GPR-MGK to graph neural networks on four logic and two real-world toxicology data sets and find that the atomic attribution of GPR-MGK generally outperforms the atomic attribution of graph neural networks. We also perform a detailed molecular attribution analysis using the FreeSolv data set, showing how molecules in the training set influence machine learning predictions and why Morgan fingerprints perform poorly on this data set. This is the first systematic examination of the interpretability of GPR-MGK and thereby is an important step in the further maturation of marginalized graph kernel methods for interpretable molecular predictions.

Supervised chemical graph mining improves drug-induced liver injury prediction

Drug-induced liver injury (DILI) is the main cause of drug failure in clinical trials. The characterization of toxic compounds in terms of chemical structure is important because compounds can be metabolized to toxic substances in the liver. Traditional machine learning approaches have had limited success in predicting DILI, and emerging deep graph neural network (GNN) models are yet powerful enough to predict DILI. In this study, we developed a completely different approach, supervised subgraph mining (SSM), a strategy to mine explicit subgraph features by iteratively updating individual graph transitions to maximize DILI fidelity. Our method outperformed previous methods including state-of-the-art GNN tools in classifying DILI on two different datasets: DILIst and TDC-benchmark. We also combined the subgraph features by using SMARTS-based frequent structural pattern matching and associated them with drugs' ATC code.

Quantitative evaluation of explainable graph neural networks for molecular property prediction

Graph neural networks (GNNs) have received increasing attention because of their expressive power on topological data, but they are still criticized for their lack of interpretability. To interpret GNN models, explainable artificial intelligence (XAI) methods have been developed. However, these methods are limited to qualitative analyses without quantitative assessments from the real-world datasets due to a lack of ground truths. In this study, we have established five XAI-specific molecular property benchmarks, including two synthetic and three experimental datasets. Through the datasets, we quantitatively assessed six XAI methods on four GNN models and made comparisons with seven medicinal chemists of different experience levels. The results demonstrated that XAI methods could deliver reliable and informative answers for medicinal chemists in identifying the key substructures. Moreover, the identified substructures were shown to complement existing classical fingerprints to improve molecular property predictions, and the improvements increased with the growth of training data.

Mechanism-driven modeling of chemical hepatotoxicity using structural alerts and an in vitro screening assay

Traditional experimental approaches to evaluate hepatotoxicity are expensive and time-consuming. As an advanced framework of risk assessment, adverse outcome pathways (AOPs) describe the sequence of molecular and cellular events underlying chemical toxicities. We aimed to develop an AOP that can be used to predict hepatotoxicity by leveraging computational modeling and in vitro assays. We curated 869 compounds with known hepatotoxicity classifications as a modeling set and extracted assay data from PubChem. The antioxidant response element (ARE) assay, which quantifies transcriptional responses to oxidative stress, showed a high correlation to hepatotoxicity (PPV=0.82). Next, we developed quantitative structure-activity relationship (QSAR) models to predict ARE activation for compounds lacking testing results. Potential toxicity alerts were identified and used to construct a mechanistic hepatotoxicity model. For experimental validation, 16 compounds in the modeling set and 12 new compounds were selected and tested using an in-house ARE-luciferase assay in HepG2-C8 cells. The mechanistic model showed good hepatotoxicity predictivity (accuracy = 0.82) for these compounds. Potential false positive hepatotoxicity predictions by only using ARE results can be corrected by incorporating structural alerts and vice versa. This mechanistic model illustrates a potential toxicity pathway for hepatotoxicity, and this strategy can be expanded to develop predictive models for other complex toxicities.

Selective progesterone receptor modulators and reproductive health

PURPOSE OF REVIEW

This review is intended to provide perspective on the history of selective progesterone receptor modulators (SPRMs) and progesterone antagonists, their current availability, therapeutic promise and safety concerns.

RECENT FINDINGS

Despite keen interest in synthesis of these compounds, only a handful have had clinical test results allowing for commercialization. Mifepristone is well tolerated and effective for single dose first trimester at-home pregnancy termination and is available in much of the world. Ulipristal acetate, at single doses, is well tolerated and effective for emergency contraception, with less availability. Chronic use of these agents has been associated with abnormal liver enzymes, and rarely, with hepatic failure; causality is not understood.

SUMMARY

SPRMs and progesterone antagonists have great therapeutic promise for use in other reproductive disorders, including breast cancer, endometriosis, adenomyosis, estrogen-free contraception and cervical ripening but require additional study. Alternative formulations, whether local (topical breast or intrauterine) or extended-release may reduce the incidence of liver function abnormalities and should be explored.

In Silico Prediction of Chemical Aquatic Toxicity by Multiple Machine Learning and Deep Learning Approaches

Fish is one of the model animals used to evaluate the adverse effects of a chemical exposed to the ecosystem. However, its low throughput and relevantly high expense make it impossible to test all new chemicals in manufacture. Hence using in silico models to prioritize compounds to be tested has been widely applied in environmental risk assessment and drug discovery. In this study, we constructed the local predictive models for four fish species, including bluegill sunfish, rainbow trout, fathead minnow, and sheepshead minnow, and the global models with all four fish data. A total of 1874 unique compounds with their labels, i.e. toxic (LC₅₀ < 10 ppm) or nontoxic were collected from ECOTOX and literature. Both conventional machine learning methods and the deep learning architecture, graph convolutional network (GCN), were used to build predictive models. The classification accuracy of the best local model for each fish species was higher than 0.83. For the global models, two strategies including consistency prediction and probability threshold were adopted to improve the predictive capability at the cost of limiting applicability domain. For 63% of compounds in domain, the accuracy was around 0.97. By comparison of the deep learning and machine learning methods, we found that the single-task GCN showed specific advantages in performance and multi-task GCN showed no advantages over the conventional machine learning methods. The data and models are available on GitHub (https://github.com/ChemPredict/ChemicalAquaticToxicity).

An Algorithm Framework for Drug-Induced Liver Injury Prediction Based on Genetic Algorithm and Ensemble Learning

In the process of drug discovery, drug-induced liver injury (DILI) is still an active research field and is one of the most common and important issues in toxicity evaluation research. It directly leads to the high wear attrition of the drug. At present, there are a variety of computer algorithms based on molecular representations to predict DILI. It is found that a single molecular representation method is insufficient to complete the task of toxicity prediction, and multiple molecular fingerprint fusion methods have been used as model input. In order to solve the problem of high dimensional and unbalanced DILI prediction data, this paper integrates existing datasets and designs a new algorithm framework, Rotation-Ensemble-GA (R-E-GA). The main idea is to find a feature subset with better predictive performance after rotating the fusion vector of high-dimensional molecular representation in the feature space. Then, an Adaboost-type ensemble learning method is integrated into R-E-GA to improve the prediction accuracy. The experimental results show that the performance of R-E-GA is better than other state-of-art algorithms including ensemble learning-based and graph neural network-based methods. Through five-fold cross-validation, the R-E-GA obtains an ACC of 0.77, an F1 score of 0.769, and an AUC of 0.842.

Exploration of structural requirements for azole chemicals towards human aromatase CYP19A1 activity: Classification modeling, structure-activity relationships and read-across study

Human aromatase, also called CYP19A1, plays a major role in the conversion of androgens into estrogens. Inhibition of aromatase is an important target for estrogen receptor (ER)-responsive breast cancer therapy. Use of azole compounds as aromatase inhibitors is widespread despite their low selectivity. A toxicological evaluation of commonly used azole-based drugs and agrochemicals with respect to CYP19A1is currently requested by the European Union- Registration, Evaluation, Authorization and Restriction of Chemicals (EU-REACH) regulations due to their potential as endocrine disruptors. In this connection, identification of structural alerts (SAs) is an effective strategy for the toxicological assessment and safe drug design. The present study describes the identification of SAs of azole-based chemicals as guiding experts to predict the aromatase activity. Total 21 SAs associated with aromatase activity were extracted from dataset of 326 azole-based drugs/chemicals obtained from Tox21 library. A cross-validated classification model having high accuracy (error rate 5%) was proposed which can precisely classify azole chemicals into active/inactive toward aromatase. In addition, mechanistic details and toxicological properties (agonism/antagonism) of azoles with respect to aromatase were explored by comparing active and inactive chemicals using structure-activity relationships (SAR). Lastly, few structural alerts were applied to form chemical categories for read-across applications.

Consensus Prediction of Human Gut Microbiota-Mediated Metabolism Susceptibility for Small Molecules by Machine Learning, Structural Alerts, and Dietary Compounds-Based Average Similarity Methods

The human gut microbiota (HGM) colonizing human gastrointestinal tract (HGT) confers a repertoire of dynamic and unique metabolic capacities that are not possessed by the host and therefore is tentatively perceived as an alternative metabolic ″organ″ besides the liver in the host. Nevertheless, the significant contribution of HGM to the overall human metabolism is often overlooked in the modern drug discovery pipeline. Hence, a systematic evaluation of HGM-mediated drug metabolism is gradually important, and its computational prediction becomes increasingly necessary. In this work, a new data set containing both the HGM-mediated metabolism susceptible (HGMMS) and insusceptible (HGMMI) compounds (329 vs 320) was manually curated. Based on this data set, the first machine learning (ML) model, a new structural alerts (SA) model, and the K-nearest neighboring dietary compounds-based average similarity (AS) model were proposed to directly predict the HGM-mediated metabolism susceptibility for small molecules, and exhibit promising performance on three independent test sets. Finally, consensus prediction (ML/SA/AS) for DrugBank molecules revealed an intriguing phenomenon that a typical Michael acceptor ″α,β-unsaturated carbonyl group″ is a very common warhead for the design of covalent inhibitors and inclined to be metabolized by HGM in anaerobic HGT to generate the reduced metabolite without the reactive warhead, which could be a new concern to medicinal chemists. To the best of our knowledge, we gleaned the first HGMMS/HGMMI data set, developed the first HGMMS/HGMMI classification model, implemented a relatively comprehensive program based on ML/SA/AS approaches, and found a new phenomenon on the HGM-mediated deactivation of an extensively used warhead for covalent inhibitors.

ADENet: a novel network-based inference method for prediction of drug adverse events

Identification of adverse drug events (ADEs) is crucial to reduce human health risks and improve drug safety assessment. With an increasing number of biological and medical data, computational methods such as network-based methods were proposed for ADE prediction with high efficiency and low cost. However, previous network-based methods rely on the topological information of known drug-ADE networks, and hence cannot make predictions for novel compounds without any known ADE. In this study, we introduced chemical substructures to bridge the gap between the drug-ADE network and novel compounds, and developed a novel network-based method named ADENet, which can predict potential ADEs for not only drugs within the drug-ADE network, but also novel compounds outside the network. To show the performance of ADENet, we collected drug-ADE associations from a comprehensive database named MetaADEDB and constructed a series of network-based prediction models. These models obtained high area under the receiver operating characteristic curve values ranging from 0.871 to 0.947 in 10-fold cross-validation. The best model further showed high performance in external validation, which outperformed a previous network-based and a recent deep learning-based method. Using several approved drugs as case studies, we found that 32-54% of the predicted ADEs can be validated by the literature, indicating the practical value of ADENet. Moreover, ADENet is freely available at our web server named NetInfer (http://lmmd.ecust.edu.cn/netinfer). In summary, our method would provide a promising tool for ADE prediction and drug safety assessment in drug discovery and development.

Benchmarking Molecular Feature Attribution Methods with Activity Cliffs

Feature attribution techniques are popular choices within the explainable artificial intelligence toolbox, as they can help elucidate which parts of the provided inputs used by an underlying supervised-learning method are considered relevant for a specific prediction. In the context of molecular design, these approaches typically involve the coloring of molecular graphs, whose presentation to medicinal chemists can be useful for making a decision of which compounds to synthesize or prioritize. The consistency of the highlighted moieties alongside expert background knowledge is expected to contribute to the understanding of machine-learning models in drug design. Quantitative evaluation of such coloring approaches, however, has so far been limited to substructure identification tasks. We here present an approach that is based on maximum common substructure algorithms applied to experimentally-determined activity cliffs. Using the proposed benchmark, we found that molecule coloring approaches in conjunction with classical machine-learning models tend to outperform more modern, graph-neural-network alternatives. The provided benchmark data are fully open sourced, which we hope will facilitate the testing of newly developed molecular feature attribution techniques.

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.

Preclinical models of idiosyncratic drug-induced liver injury (iDILI): Moving towards prediction

Idiosyncratic drug-induced liver injury (iDILI) encompasses the unexpected harms that prescription and non-prescription drugs, herbal and dietary supplements can cause to the liver. iDILI remains a major public health problem and a major cause of drug attrition. Given the lack of biomarkers for iDILI prediction, diagnosis and prognosis, searching new models to predict and study mechanisms of iDILI is necessary. One of the major limitations of iDILI preclinical assessment has been the lack of correlation between the markers of hepatotoxicity in animal toxicological studies and clinically significant iDILI. Thus, major advances in the understanding of iDILI susceptibility and pathogenesis have come from the study of well-phenotyped iDILI patients. However, there are many gaps for explaining all the complexity of iDILI susceptibility and mechanisms. Therefore, there is a need to optimize preclinical human in vitro models to reduce the risk of iDILI during drug development. Here, the current experimental models and the future directions in iDILI modelling are thoroughly discussed, focusing on the human cellular models available to study the pathophysiological mechanisms of the disease and the most used in vivo animal iDILI models. We also comment about in silico approaches and the increasing relevance of patient-derived cellular models.

Clinical Pharmacokinetics and Pharmacodynamics of the Selective Progesterone Receptor Modulator Vilaprisan: A Comprehensive Overview

Vilaprisan is a highly potent selective progesterone receptor modulator in development for the treatment of symptomatic uterine fibroids and endometriosis. Its pharmacokinetics are characterized by rapid absorption, almost complete bioavailability, and dose-proportional exposure. The intrinsic factors of age, bodyweight, and race have no clinically relevant effect on the pharmacokinetics and pharmacodynamics of vilaprisan and do not warrant a dose adjustment. Similarly, vilaprisan can be used in patients with mild or moderate renal or hepatic impairment without dose adjustment, but its use is not recommended in patients with severe organ impairment. Vilaprisan has no perpetrator potential on cytochrome P450 (CYP) enzymes or transporters and therefore restrictions in the concomitant use of their substrates are not required. Nonetheless, because it is a sensitive CYP3A4 substrate itself, concomitant use of vilaprisan with strong CYP3A inhibitors or inducers is not recommended. However, there is no risk for QTc prolongation when vilaprisan and a strong CYP3A inhibitor are administered concomitantly, as indicated by a vilaprisan concentration–QTc response analysis across all studies with triplicate electrocardiogram measurements. Furthermore, due to its mode of action, vilaprisan is also not recommended to be used together with progestin-containing oral contraceptives. Vilaprisan shows a steep exposure–response relationship for inducing amenorrhea in patients with uterine fibroids experiencing heavy menstrual bleeding. Based on simulations, a dose of 2 mg/day is expected to induce a maximum bleeding reduction and was thus selected for phase III.

In-silico evaluation of Malawi essential medicines and reactive metabolites for potential drug-induced toxicities

BACKGROUND

Drug-induced toxicity is one of the problems that have negatively impacted on the well-being of populations throughout the world, including Malawi. It results in unnecessary hospitalizations, retarding the development of the country. This study assessed the Malawi Essential Medicines List (MEML) for structural alerts and reactive metabolites with the potential for drug-induced toxicities.

METHODS

This in-silico screening study used StopTox, ToxAlerts and LD-50 values toxicity models to assess the MEML drugs. A total of 296 drugs qualified for the analysis (those that had defined chemical structures) and were screened in each software programme. Each model had its own toxicity endpoints and the models were compared for consensus of their results.

RESULTS

In the StopTox model, 86% of the drugs had potential to cause at least one toxicity including 55% that had the potential of causing eye irritation and corrosion. In ToxAlerts, 90% of the drugs had the potential of causing at least one toxicity and 72% were found to be potentially reactive, unstable and toxic. In LD-50, 70% of the drugs were potentially toxic. Model consensus evaluation results showed that the highest consensus was observed between ToxAlerts and StopTox (80%). The overall consensus amongst the three models was 57% and statistically significant (p < 0.05).

CONCLUSIONS

A large number of drugs had the potential to cause various systemic toxicities. But the results need to be interpreted cautiously since the clinical translation of QSAR-based predictions depends on many factors. In addition, inconsistencies have been reported between screening results amongst different models.

Mechanism-Driven Read-Across of Chemical Hepatotoxicants Based on Chemical Structures and Biological Data

Hepatotoxicity is a leading cause of attrition in the drug development process. Traditional preclinical and clinical studies to evaluate hepatotoxicity liabilities are expensive and time consuming. With the advent of critical advancements in high-throughput screening, there has been a rapid accumulation of in vitro toxicity data available to inform the risk assessment of new pharmaceuticals and chemicals. To this end, we curated and merged all available in vivo hepatotoxicity data obtained from the literature and public resources, which yielded a comprehensive database of 4089 compounds that includes hepatotoxicity classifications. After dividing the original database of chemicals into modeling and test sets, PubChem assay data were automatically extracted using an in-house data mining tool and clustered based on relationships between structural fragments and cellular responses in in vitro assays. The resultant PubChem assay clusters were further investigated. During the cross-validation procedure, the biological data obtained from several assay clusters exhibited high predictivity of hepatotoxicity and these assays were selected to evaluate the test set compounds. The read-across results indicated that if a new compound contained specific identified chemical fragments (ie, Molecular Initiating Event) and showed active responses in the relevant selected PubChem assays, there was potential for the chemical to be hepatotoxic in vivo. Furthermore, several mechanisms that might contribute to toxicity were derived from the modeling results including alterations in nuclear receptor signaling and inhibition of DNA repair. This modeling strategy can be further applied to the investigation of other complex chemical toxicity phenomena (eg, developmental and reproductive toxicities) as well as drug efficacy.

Prediction and mechanistic analysis of drug-induced liver injury (DILI) based on chemical structure

BACKGROUND

Drug-induced liver injury (DILI) is a major safety concern characterized by a complex and diverse pathogenesis. In order to identify DILI early in drug development, a better understanding of the injury and models with better predictivity are urgently needed. One approach in this regard are in silico models which aim at predicting the risk of DILI based on the compound structure. However, these models do not yet show sufficient predictive performance or interpretability to be useful for decision making by themselves, the former partially stemming from the underlying problem of labeling the in vivo DILI risk of compounds in a meaningful way for generating machine learning models.

RESULTS

As part of the Critical Assessment of Massive Data Analysis (CAMDA) "CMap Drug Safety Challenge" 2019 ( http://camda2019.bioinf.jku.at ), chemical structure-based models were generated using the binarized DILIrank annotations. Support Vector Machine (SVM) and Random Forest (RF) classifiers showed comparable performance to previously published models with a mean balanced accuracy over models generated using 5-fold LOCO-CV inside a 10-fold training scheme of 0.759 ± 0.027 when predicting an external test set. In the models which used predicted protein targets as compound descriptors, we identified the most information-rich proteins which agreed with the mechanisms of action and toxicity of nonsteroidal anti-inflammatory drugs (NSAIDs), one of the most important drug classes causing DILI, stress response via TP53 and biotransformation. In addition, we identified multiple proteins involved in xenobiotic metabolism which could be novel DILI-related off-targets, such as CLK1 and DYRK2. Moreover, we derived potential structural alerts for DILI with high precision, including furan and hydrazine derivatives; however, all derived alerts were present in approved drugs and were over specific indicating the need to consider quantitative variables such as dose.

CONCLUSION

Using chemical structure-based descriptors such as structural fingerprints and predicted protein targets, DILI prediction models were built with a predictive performance comparable to previous literature. In addition, we derived insights on proteins and pathways statistically (and potentially causally) linked to DILI from these models and inferred new structural alerts related to this adverse endpoint.

A Robust, Mechanistically Based In Silico Structural Profiler for Hepatic Cholestasis

Owing to the primary role which it holds within metabolism of xenobiotics, the liver stands at heightened risk of exposure to, and injury from, potentially hazardous substances. A principal manifestation of liver dysfunction is cholestasis-the impairment of physiological bile circulation from its point of origin within the organ to the site of action in the small intestine. The capacity for early identification of compounds liable to exert cholestatic effects is of particular utility within the field of pharmaceutical development, where contribution toward candidate attrition is great. Shortcomings associated with the present in vitro methodologies forecasting cholestasis render their predictivity questionable, permitting scope for the adoption of computational toxicology techniques. As such, the intention of this study has been to construct an in silico profiler, founded upon clinical data, highlighting structural motifs most reliably associated with the end point. Drawing upon a list of >1500 small molecular drugs, compiled and annotated by Kotsampasakou, E. and Ecker, G. F. (J. Chem. Inf. Model. 2017, 57, 608-615), we have formulated a series of 15 structural alerts. These describe fragments intrinsic within distinct pharmaceutical classes including psychoactive tricyclics, β-lactam antimicrobials, and estrogenic/androgenic steroids. Description of the coverage and selectivity of each are provided, alongside consideration of the underlying reactive mechanisms and relevant structure-activity concerns. Provision of mechanistic anchoring ensures that potential exists for framing within the adverse outcome pathway paradigm-the chemistry conveyed through the alert, in particular enabling rationalization at the level of the molecular initiating event.

QSAR without borders

Prediction of chemical bioactivity and physical properties has been one of the most important applications of statistical and more recently, machine learning and artificial intelligence methods in chemical sciences. This field of research, broadly known as quantitative structure-activity relationships (QSAR) modeling, has developed many important algorithms and has found a broad range of applications in physical organic and medicinal chemistry in the past 55+ years. This Perspective summarizes recent technological advances in QSAR modeling but it also highlights the applicability of algorithms, modeling methods, and validation practices developed in QSAR to a wide range of research areas outside of traditional QSAR boundaries including synthesis planning, nanotechnology, materials science, biomaterials, and clinical informatics. As modern research methods generate rapidly increasing amounts of data, the knowledge of robust data-driven modelling methods professed within the QSAR field can become essential for scientists working both within and outside of chemical research. We hope that this contribution highlighting the generalizable components of QSAR modeling will serve to address this challenge.

Essential Medicinal Chemistry of Essential Medicines

Since 1977, the World Health Organization publishes a list of essential medicines, i.e., those that satisfy the priority health care needs of the population and are selected with regard to disease prevalence and public health relevance, evidence of clinical efficacy, and safety, as well as comparative costs and cost-effectiveness. The Essential Medicines List (EML) is an invaluable tool for all countries to select those medicines that have an excellent risk/benefit ratio and that are reputed to be of pivotal importance to health. In the present perspective, we describe the chemical composition and the main features of the small molecules that are included in the EML, spanning from their origin, to their stereochemistry and measure of drug-likeness. Most and foremost, we wish to disseminate the importance of the EML, which can be both a helpful teaching tool in an ever-expanding world of medicines and an inspiration for those involved in pharmaceutical R&D.

Metabolism and Bioactivation: It's Time to Expect the Unexpected

Improvements in in vitro ADME tools and pharmacokinetic prediction models have helped to shift attrition rates in early clinical trials from poor exposure to drug safety concerns, such as drug-induced liver injury (DILI). Assessing a new chemical entity's potential for liver toxicity is an important consideration for the likely success of new drug candidates. Reactive intermediates produced during drug metabolism have been implicated as a cause of DILI, and their formation has been correlated to the addition of a black box warning on a drug label. In this work, we will present contemporary examples of the bioactivation of atypical structures usually regarded as benign and often used by medicinal chemists when attempting to avoid bioactivation. Medicinal chemistry strategies used to derisk bioactivation will be discussed, and an emphasis will be placed on the necessity of a multidisciplinary approach.

KnowTox: pipeline and case study for confident prediction of potential toxic effects of compounds in early phases of development

Risk assessment of newly synthesised chemicals is a prerequisite for regulatory approval. In this context, in silico methods have great potential to reduce time, cost, and ultimately animal testing as they make use of the ever-growing amount of available toxicity data. Here, KnowTox is presented, a novel pipeline that combines three different in silico toxicology approaches to allow for confident prediction of potentially toxic effects of query compounds, i.e. machine learning models for 88 endpoints, alerts for 919 toxic substructures, and computational support for read-across. It is mainly based on the ToxCast dataset, containing after preprocessing a sparse matrix of 7912 compounds tested against 985 endpoints. When applying machine learning models, applicability and reliability of predictions for new chemicals are of utmost importance. Therefore, first, the conformal prediction technique was deployed, comprising an additional calibration step and per definition creating internally valid predictors at a given significance level. Second, to further improve validity and information efficiency, two adaptations are suggested, exemplified at the androgen receptor antagonism endpoint. An absolute increase in validity of 23% on the in-house dataset of 534 compounds could be achieved by introducing KNNRegressor normalisation. This increase in validity comes at the cost of efficiency, which could again be improved by 20% for the initial ToxCast model by balancing the dataset during model training. Finally, the value of the developed pipeline for risk assessment is discussed using two in-house triazole molecules. Compared to a single toxicity prediction method, complementing the outputs of different approaches can have a higher impact on guiding toxicity testing and de-selecting most likely harmful development-candidate compounds early in the development process.

Current status and future perspective of computational toxicology in drug safety assessment under ontological intellection

For the safety assessment of pharmaceuticals, initial data management requires accurate toxicological data acquisition, which is based on regulatory safety studies according to guidelines, and computational systems have been developed under the application of Good Laboratory Practice (GLP). In addition to these regulatory toxicology studies, investigative toxicological study data for the selection of lead compound and candidate compound for clinical trials are directed to estimation by computational systems such as Quantitative Structure-Activity Relationship (QSAR) and related expert systems. Furthermore, in the "Go" or "No-Go" decision of drug development, supportive utilization of a scientifically interpretable computational toxicology system is required for human safety evaluation. A pharmaceutical safety evaluator as a related toxicologist who is facing practical decision needs not only a data-driven Artificial Intelligence (AI) system that calls for the final consequence but also an explainable AI that can provide comprehensive information necessary for evaluation and can help with decision making. Through the explication and suggestion of information on the mechanism of toxic effects to safety assessment scientists, a subsidiary partnership system for risk assessment is ultimately to be a powerful tool that can indicate project-vector with data weight for the corresponding counterparts. To bridge the gaps between big data and knowledge, multi-dimensional thinking based on philosophical ontology theory is necessary for handling heterogeneous data for integration. In this review, we will explain the current state and future perspective of computational toxicology related to drug safety assessment from the viewpoint of ontology thinking.

Discovery of Vilaprisan (BAY 1002670): A Highly Potent and Selective Progesterone Receptor Modulator Optimized for Gynecologic Therapies

Progesterone plays an important role in the female reproductive system. However, there is also evidence that gynecologic disorders/diseases such as uterine fibroids and endometriosis are progesterone-dependent. Steroidal and non-steroidal selective progesterone receptor modulators (SPRMs) have shown potential for the treatment of such diseases. Steroidal SPRMs, including mifepristone and ulipristal acetate, have proven effective in clinical trials. However, several steroidal SPRMs containing a dimethylamino substituent have been associated with elevated liver enzymes in patients. An earlier drug discovery program identified lonaprisan as a highly selective SPRM that did not show drug-related change in liver enzyme activity. Building on data obtained from that work, here we describe the research program that culminated in the discovery of a novel steroidal SPRM, vilaprisan, which combines an extremely high potency with very favorable drug metabolism and pharmacokinetic properties. Vilaprisan has entered clinical development and is currently undergoing phase 3 clinical trials.

In Silico Prediction of Drug-Induced Liver Injury Based on Adverse Drug Reaction Reports

Drug-induced liver injury (DILI) is a major cause of drug attrition. Currently existing Quantitative Structure-Activity Relationship models have limited predictive capabilities for DILI. Furthermore, their practical applications were limited by lack of new hepatotoxicity data. In this study, we first collected and curated a novel set of 122 DILI-positive and 932 DILI-negative drugs from online adverse drug reports using proportional reporting ratios as the signal detection method. Second, three strategies (under-sampling the majority class, synthetic minority over-sampling technique, and adjusting decision threshold approach) were employed to develop predictive classification models to cope with the unbalanced dataset. Random forest (RF) models using CDK, MACCS, and Mold2 descriptors based on the under-sampling and over-sampling strategies afforded correct classification ratio (CCR) of ∼0.77 and 0.78, respectively. Recursive RF models based on the last strategy tremendously reduced modeling descriptors (at most 95.4% for Mold2) while apparently improved the predictability with a consensus CCR of 0.84 (sensitivity of 0.88 and specificity of 0.79). Structural analysis showed that pyrimidine derivatives, purine derivatives, and halogenated hydrocarbon were critical for drugs' hepatotoxicity. The reporting frequency of many drugs was gender-dependent (eg, antiviral and anti-cancer drugs for males and antibacterial drugs for females) as well as age-dependent (eg, antiviral and anti-cancer drugs for the middle age group of 20-29, 30-39, and 40-49). Approximately 84% of total cases were reported during the first 6 months of administration. The curated hepatotoxicity dataset along with the predictive classification models presented here should provide insight into future studies of DILI.

Deep Learning Based Regression and Multiclass Models for Acute Oral Toxicity Prediction with Automatic Chemical Feature Extraction

Median lethal death, LD₅₀, is a general indicator of compound acute oral toxicity (AOT). Various in silico methods were developed for AOT prediction to reduce costs and time. In this study, we developed an improved molecular graph encoding convolutional neural networks (MGE-CNN) architecture to construct three types of high-quality AOT models: regression model (deepAOT-R), multiclassification model (deepAOT-C), and multitask model (deepAOT-CR). These predictive models highly outperformed previously reported models. For the two external data sets containing 1673 (test set I) and 375 (test set II) compounds, the R² and mean absolute errors (MAEs) of deepAOT-R on the test set I were 0.864 and 0.195, and the prediction accuracies of deepAOT-C were 95.5% and 96.3% on test sets I and II, respectively. The two external prediction accuracies of deepAOT-CR are 95.0% and 94.1%, while the R² and MAE are 0.861 and 0.204 for test set I, respectively. We then performed forward and backward exploration of deepAOT models for deep fingerprints, which could support shallow machine learning methods more efficiently than traditional fingerprints or descriptors. We further performed automatic feature learning, a key essence of deep learning, to map the corresponding activation values into fragment space and derive AOT-related chemical substructures by reverse mining of the features. Our deep learning architecture for AOT is generally applicable in predicting and exploring other toxicity or property end points of chemical compounds. The two deepAOT models are freely available at http://repharma.pku.edu.cn/DLAOT/DLAOThome.php or http://www.pkumdl.cn/DLAOT/DLAOThome.php .

Predicting drug-induced liver injury: The importance of data curation

Drug-induced liver injury (DILI) is a major issue for both patients and pharmaceutical industry due to insufficient means of prevention/prediction. In the current work we present a 2-class classification model for DILI, generated with Random Forest and 2D molecular descriptors on a dataset of 966 compounds. In addition, predicted transporter inhibition profiles were also included into the models. The initially compiled dataset of 1773 compounds was reduced via a 2-step approach to 966 compounds, resulting in a significant increase (p-value < 0.05) in model performance. The models have been validated via 10-fold cross-validation and against three external test sets of 921, 341 and 96 compounds, respectively. The final model showed an accuracy of 64% (AUC 68%) for 10-fold cross-validation (average of 50 iterations) and comparable values for two test sets (AUC 59%, 71% and 66%, respectively). In the study we also examined whether the predictions of our in-house transporter inhibition models for BSEP, BCRP, P-glycoprotein, and OATP1B1 and 1B3 contributed in improvement of the DILI mode. Finally, the model was implemented with open-source 2D RDKit descriptors in order to be provided to the community as a Python script.

Prediction of hERG Liability - Using SVM Classification, Bootstrapping and Jackknifing

Drug-induced QT prolongation leads to life-threatening cardiotoxicity, mostly through blockage of the human ether-à-go-go-related gene (hERG) encoded potassium ion (K+ ) channels. The hERG channel is one of the most important antitargets to be addressed in the early stage of drug discovery process, in order to avoid more costly failures in the development phase. Using a thallium flux assay, 4,323 molecules were screened for hERG channel inhibition in a quantitative high throughput screening (qHTS) format. Here, we present support vector classification (SVC) models of hERG channel inhibition with the averaged area under the receiver operator characteristics curve (AUC-ROC) of 0.93 for the tested compounds. Both Jackknifing and bootstrapping have been employed to rebalance the heavily biased training datasets, and the impact of these two under-sampling rebalance methods on the performance of the predictive models is discussed. Our results indicated that the rebalancing techniques did not enhance the predictive power of the resulting models; instead, adoption of optimal cutoffs could restore the desirable balance of sensitivity and specificity of the binary classifiers. In an external validation set of 66 drug molecules, the SVC model exhibited an AUC-ROC of 0.86, further demonstrating the utility of this modeling approach to predict hERG liabilities.

Adverse drug reactions triggered by the common HLA-B*57:01 variant: a molecular docking study

BACKGROUND

Human leukocyte antigen (HLA) surface proteins are directly involved in idiosyncratic adverse drug reactions. Herein, we present a structure-based analysis of the common HLA-B*57:01 variant known to be responsible for several HLA-linked adverse effects such as the abacavir hypersensitivity syndrome.

METHODS

First, we analyzed three X-ray crystal structures involving the HLA-B*57:01 protein variant, the anti-HIV drug abacavir, and different co-binding peptides present in the antigen-binding cleft. We superimposed the three complexes and showed that abacavir had no significant conformational variation whatever the co-binding peptide. Second, we self-docked abacavir in the HLA-B*57:01 antigen binding cleft with and without peptide using Glide. Third, we docked a small test set of 13 drugs with known ADRs and suspected HLA associations.

RESULTS

In the presence of an endogenous co-binding peptide, we found a significant stabilization (~2 kcal/mol) of the docking scores and identified several modified abacavir-peptide interactions indicating that the peptide does play a role in stabilizing the HLA-abacavir complex. Next, our model was used to dock a test set of 13 drugs at HLA-B*57:01 and measured their predicted binding affinities. Drug-specific interactions were observed at the antigen-binding cleft and we were able to discriminate the compounds with known HLA-B*57:01 liability from inactives.

CONCLUSIONS

Overall, our study highlights the relevance of molecular docking for evaluating and analyzing complex HLA-drug interactions. This is particularly important for virtual drug screening over thousands of HLA variants as other experimental techniques (e.g., in vitro HTS) and computational approaches (e.g., molecular dynamics) are more time consuming and expensive to conduct. As the attention for drugs' HLA liability is on the rise, we believe this work participates in encouraging the use of molecular modeling for reliably studying and predicting HLA-drug interactions. Graphical abstract.

Drug-induced liver injury: Towards early prediction and risk stratification

Drug-induced liver injury (DILI) is a hot topic for clinicians, academia, drug companies and regulators, as shown by the steadily increasing number of publications and agents listed as causing liver damage (http://livertox.nih.gov/). As it was the case in the past decade with drug-induced QT prolongation/arrhythmia, there is an urgent unmet clinical need to develop tools for risk assessment and stratification in clinical practice and, in parallel, to improve prediction of pre-clinical models to support regulatory steps and facilitate early detection of liver-specific adverse drug events. Although drug discontinuation and therapy reconciliation still remain the mainstay in patient management to minimize occurrence of DILI, especially acute liver failure events, different multidisciplinary attempts have been proposed in 2016 to predict and assess drug-related risk in individual patients; these promising, albeit preliminary, results strongly support the need to pursue this innovative pathway.

Protective Roles of N-acetyl Cysteine and/or Taurine against Sumatriptan-Induced Hepatotoxicity

Purpose: Triptans are the drug category mostly prescribed for abortive treatment of migraine. Most recent cases of liver toxicity induced by triptans have been described, but the mechanisms of liver toxicity of these medications have not been clear. Methods: In the present study, we obtained LC₅₀ using dose-response curve and investigated cell viability, free radical generation, lipid peroxide production, mitochondrial injury, lysosomal membrane damage and the cellular glutathione level as toxicity markers as well as the beneficial effects of taurine and/or N-acetyl cysteine in the sumatriptan-treated rat parenchymal hepatocytes using accelerated method of cytotoxicity mechanism screening. Results: It was revealed that liver toxicity induced by sumatriptan in in freshly isolated parenchymal hepatocytes is dose-dependent. Sumatriptan caused significant free radical generation followed by lipid peroxide formation, mitochondrial injury as well as lysosomal damage. Moreover, sumatriptan reduced cellular glutathione content. Taurine and N-acetyl cysteine were able to protect hepatocytes against sumatriptan-induced harmful effects. Conclusion: It is concluded that sumatriptan causes oxidative stress in hepatocytes and the decreased hepatocytes glutathione has a key role in the sumatriptan-induced harmful effects. Also, N-acetyl cysteine and/or taurine could be used as treatments in sumatriptan-induced side effects.

Upgrading HepG2 cells with adenoviral vectors that encode drug-metabolizing enzymes: application for drug hepatotoxicity testing

INTRODUCTION

Drug attrition rates due to hepatotoxicity are an important safety issue considered in drug development. The HepG2 hepatoma cell line is currently being used for drug-induced hepatotoxicity evaluations, but its expression of drug-metabolizing enzymes is poor compared with hepatocytes. Different approaches have been proposed to upgrade HepG2 cells for more reliable drug-induced liver injury predictions. Areas covered: We describe the advantages and limitations of HepG2 cells transduced with adenoviral vectors that encode drug-metabolizing enzymes for safety risk assessments of bioactivable compounds. Adenoviral transduction facilitates efficient and controlled delivery of multiple drug-metabolizing activities to HepG2 cells at comparable levels to primary human hepatocytes by generating an 'artificial hepatocyte'. Furthermore, adenoviral transduction enables the design of tailored cells expressing particular metabolic capacities. Expert opinion: Upgraded HepG2 cells that recreate known inter-individual variations in hepatic CYP and conjugating activities due to both genetic (e.g., polymorphisms) or environmental (e.g., induction, inhibition) factors seems a suitable model to identify bioactivable drug and conduct hepatotoxicity risk assessments. This strategy should enable the generation of customized cells by reproducing human pheno- and genotypic CYP variability to represent a valuable human hepatic cell model to develop new safer drugs and to improve existing predictive toxicity assays.

Alarms about structural alerts

Structural alerts are widely accepted in chemical toxicology and regulatory decision support as a simple and transparent means to flag potential chemical hazards or group compounds into categories for read-across. However, there has been a growing concern that alerts disproportionally flag too many chemicals as toxic, which questions their reliability as toxicity markers. Conversely, the rigorously developed and properly validated statistical QSAR models can accurately and reliably predict the toxicity of a chemical; however, their use in regulatory toxicology has been hampered by the lack of transparency and interpretability. We demonstrate that contrary to the common perception of QSAR models as "black boxes" they can be used to identify statistically significant chemical substructures (QSAR-based alerts) that influence toxicity. We show through several case studies, however, that the mere presence of structural alerts in a chemical, irrespective of the derivation method (expert-based or QSAR-based), should be perceived only as hypotheses of possible toxicological effect. We propose a new approach that synergistically integrates structural alerts and rigorously validated QSAR models for a more transparent and accurate safety assessment of new chemicals.