Evaluation of Existing QSAR Models and Structural Alerts and Development of New Ensemble Models for Genotoxicity Using a Newly Compiled Experimental Dataset

Advances in Activity/Property Prediction from Chemical Structures

Recent technological advancement in AI modeling of molecular property databases has significantly expanded the opportunities for drug design and development. Quantitative structure-activity relationships (QSARs) are shown to provide more accurate predictions with regards to biological activity as well as toxicological assessment. By using a combination of in-silico models or by combining disparate structure-activity databases, researchers have been able to improve accuracy for a variety of drug discovery and analysis methods, generating viable compounds, which in certain cases, can be synthesized and further studied in vitro to find candidates for potential development. Additionally, the development of compounds of determined toxicology can be discontinued earlier, allowing alternative routes to be evaluated, preventing wasted time and resources. Although the progress that has been made is tremendous, expert review is still necessary for most in-silico generated predictions. Regardless, the scientific community continues to move ever closer to completely automated drug discovery and evaluation.

Next generation risk assessment for occupational chemical safety - A real world example with sodium-2-hydroxyethane sulfonate

Next Generation Risk Assessment (NGRA) is an exposure-led approach to safety assessment that uses New Approach Methodologies (NAMs). Application of NGRA has been largely restricted to assessments of consumer use of cosmetics and is not currently implemented in occupational safety assessments, e.g. under EU REACH. By contrast, a large proportion of regulatory worker safety assessments are underpinned by toxicological studies using experimental animals. Consequently, occupational safety assessment represents an area that would benefit from increasing application of NGRA to safety decision making. Here, a workflow for conducting NGRA under an occupational safety context was developed, which is illustrated with a case study chemical; sodium 2-hydroxyethane sulphonate (sodium isethionate or SI). Exposures were estimated using a standard occupational exposure model following a comprehensive life cycle assessment of SI and considering factory-specific data. Outputs of this model were then used to estimate internal exposures using a Physiologically Based Kinetic (PBK) model, which was constructed with SI specific Absorption, Distribution, Metabolism and Excretion (ADME) data. PBK modelling indicated a worst-case plasma maximum concentration (Cmax) of 0.8 μM across the SI life cycle. SI bioactivity was assessed in a battery of NAMs relevant to systemic, reproductive, and developmental toxicity; a cell stress panel, high throughput transcriptomics in three cell lines (HepG2, HepaRG and MCF-7 cells), pharmacological profiling and specific assays relating to developmental toxicity (Reprotracker and devTOX quickPredict). Points of Departure (PoDs) for SI ranged from 104 to 5044 µM. Cmax values obtained from PBK modelling of occupational exposures to SI were compared with PoDs from the bioactivity assays to derive Bioactivity Exposure Ratios (BERs) which demonstrated the safety for workers exposed to SI under current levels of factory specific risk management. In summary, the tiered and iterative workflow developed here represents an opportunity for integrating non animal approaches for a large subset of substances for which systemic worker safety assessment is required. Such an approach could be followed to ensure that animal testing is only conducted as a "last resort" e.g. under EU REACH.

Development and application of consensus in silico models for advancing high-throughput toxicological predictions

Computational toxicology models have been successfully implemented to prioritize and screen chemicals. There are numerous in silico (quantitative) structure-activity relationship ([Q]SAR) models for the prediction of a range of human-relevant toxicological endpoints, but for a given endpoint and chemical, not all predictions are identical due to differences in their training sets, algorithms, and methodology. This poses an issue for high-throughput screening of a large chemical inventory as it necessitates several models to cover diverse chemistries but will then generate data conflicts. To address this challenge, we developed a consensus modeling strategy to combine predictions obtained from different existing in silico (Q)SAR models into a single predictive value while also expanding chemical space coverage. This study developed consensus models for nine toxicological endpoints relating to estrogen receptor (ER) and androgen receptor (AR) interactions (i.e., binding, agonism, and antagonism) and genotoxicity (i.e., bacterial mutation, in vitro chromosomal aberration, and in vivo micronucleus). Consensus models were created by combining different (Q)SAR models using various weighting schemes. As a multi-objective optimization problem, there is no single best consensus model, and therefore, Pareto fronts were determined for each endpoint to identify the consensus models that optimize the multiple-criterion decisions simultaneously. Accordingly, this work presents sets of solutions for each endpoint that contain the optimal combination, regardless of the trade-off, with the results demonstrating that the consensus models improved both the predictive power and chemical space coverage. These solutions were further analyzed to find trends between the best consensus models and their components. Here, we demonstrate the development of a flexible and adaptable approach for in silico consensus modeling and its application across nine toxicological endpoints related to ER activity, AR activity, and genotoxicity. These consensus models are developed to be integrated into a larger multi-tier NAM-based framework to prioritize chemicals for further investigation and support the transition to a non-animal approach to risk assessment in Canada.

Evaluation of QSAR models for predicting mutagenicity: outcome of the Second Ames/QSAR international challenge project

Quantitative structure-activity relationship (QSAR) models are powerful in silico tools for predicting the mutagenicity of unstable compounds, impurities and metabolites that are difficult to examine using the Ames test. Ideally, Ames/QSAR models for regulatory use should demonstrate high sensitivity, low false-negative rate and wide coverage of chemical space. To promote superior model development, the Division of Genetics and Mutagenesis, National Institute of Health Sciences, Japan (DGM/NIHS), conducted the Second Ames/QSAR International Challenge Project (2020-2022) as a successor to the First Project (2014-2017), with 21 teams from 11 countries participating. The DGM/NIHS provided a curated training dataset of approximately 12,000 chemicals and a trial dataset of approximately 1,600 chemicals, and each participating team predicted the Ames mutagenicity of each trial chemical using various Ames/QSAR models. The DGM/NIHS then provided the Ames test results for trial chemicals to assist in model improvement. Although overall model performance on the Second Project was not superior to that on the First, models from the eight teams participating in both projects achieved higher sensitivity than models from teams participating in only the Second Project. Thus, these evaluations have facilitated the development of QSAR models.

Quantification of pharmaceuticals in hospital effluent: Weighted ranking of environmental risk using a fuzzy hybrid multicriteria method

An analytical method for quantification of seventeen pharmaceuticals and one metabolite was validated and applied in the analysis of hospital effluent samples. Two different sampling strategies were used: seasonal sampling, with 7 samples collected bimonthly; and hourly sampling, with 12 samples collected during 12 h. Thus, the variability was both seasonal and within the same day. High variability was observed in the measured concentrations of the pharmaceuticals and the metabolite. The quantification method, performed using weighted linear regression model, demonstrated results of average concentrations in seasonal samples ranged between 0.19 μgL-1 (carbamazepine) and higher than 61.56 μgL-1 (acetaminophen), while the hourly samples showed average concentrations between 0.07 μgL-1 (diazepam) and higher than 54.91 μgL-1 (acetaminophen). It is described as higher because the maximum concentration of the calibration curve took into account the dilution factor provided by DLLME. The diurnal results showed a trend towards higher concentrations in the first and last hours of sampling. The risk quotient (RQ) was calculated using organisms from three different trophic levels, for all the analytes quantified in the samples. Additionally, in order to understand the level of importance of each RQ, an expert panel was established, with contributions from 23 specialists in the area. The results were analyzed using a hybrid decision-making approach based on a Fuzzy Analytic Hierarchy Process (FAHP) and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method, in order to rank the compounds by environmental risk priority. The compounds of greatest concern were losartan, acetaminophen, 4-aminoantipyrine, sulfamethoxazole, and metoclopramide. Comparison of the environmental risk priority ranking with the potential human health risk was performed by applying the same multicriteria approach, with the prediction of endpoints using in silico (Q)SAR models. The results obtained suggested that sulfamethoxazole and acetaminophen were the most important analytes to be considered for monitoring.

Machine Learning-Based Prediction of Drug-Induced Hepatotoxicity: An OvA-QSTR Approach

Drug-induced hepatotoxicity, also known as drug-induced liver injury (DILI), is among the possible adverse effects of pharmacotherapy. This clinical condition is accepted as one of the factors leading to patient mortality and morbidity. The LiverTox database was built by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) to predict potential liver damage from medications and take appropriate precautions. The database has classified medicines into seven risk categories (A, B, C, D, E, E*, and X) to avoid medicine-induced liver toxicity. The hepatic damage risk decreases from group A to group E. This study did not include the E* and X classes because they contained unverified and unknown data groups. Our study aims to predict potential liver damage of new drug molecules without using experimental animals. We predict which of the LiverTox risk category drugs with unknown liver toxicity potential will fall into using our one-vs-all quantitative structure-toxicity relationship (OvA-QSTR) model. Our dataset, consisting of 678 organic drug molecules from different pharmacological classes, was collected from LiverTox. The OvA-QSTR models implemented by Bayesian Network (BayesNet) performed well based on the selected descriptors, with the precision-recall curve (PRC) areas ranging from 0.718 to 0.869. Our OvA-QSTR models provide a reliable premarketing risk evaluation of pharmaceutical-induced liver damage potential and offer predictions for different risk levels in DILI.

Per- and polyfluoroalkyl substances (PFASs) registered under REACH-What can we learn from the submitted data and how important will mobility be in PFASs hazard assessment?

The EU is planning to restrict the manufacture, placing on the market and use of per- and polyfluoroalkyl substances (PFASs) as a class. For such a broad regulatory approach, a lot of different data are required, including data on the hazardous properties of PFASs. Here, we analyze substances that fulfill the OECD definition of PFASs and that are registered under the regulation on Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) in the EU to obtain a better data basis for PFASs and to elucidate the range of PFASs on the market in the EU. As of September 2021, at least 531 PFASs had been registered under REACH. Our hazard assessment of the PFASs registered under REACH shows that the currently available data are not sufficient to identify those PFASs that are persistent, bioaccumulative and toxic (PBT) or very persistent and very bioaccumulative (vPvB). Using some basic assumptions - which are 1) PFASs or their metabolites do not mineralize, 2) neutral hydrophobic substances bioaccumulate unless they are metabolized and 3) all chemicals exhibit baseline toxicity, and effect concentrations cannot be above effect concentrations for baseline toxicity - shows that at least 17 of the 177 PFASs with full registration are PBT substances, 14 more than currently identified. Moreover, if mobility is considered as a hazard criterion, at least 19 additional substances will need to be considered hazardous. The regulation of persistent, mobile and toxic (PMT) and very persistent and very mobile (vPvM) substances would therefore also affect PFASs. However, many of the substances that have not been identified as PBT, vPvB, PMT or vPvM are either persistent and toxic, persistent and bioaccumulative or persistent and mobile. The planned PFASs restriction will therefore be important for a more effective regulation of these substances.

Application of systematic evidence mapping to identify available data on the potential human health hazards of selected market-relevant azo dyes

BACKGROUND

Azo dyes are used in textiles and leather clothing. Human exposure can occur from wearing textiles containing azo dyes. Since the body's enzymes and microbiome can cleave azo dyes, potentially resulting in mutagenic or carcinogenic metabolites, there is also an indirect health concern on the parent compounds. While several hazardous azo dyes are banned, many more are still in use that have not been evaluated systematically for potential health concerns. This systematic evidence map (SEM) aims to compile and categorize the available toxicological evidence on the potential human health risks of a set of 30 market-relevant azo dyes.

METHODS

Peer-reviewed and gray literature was searched and over 20,000 studies were identified. These were filtered using Sciome Workbench for Interactive computer-Facilitated Text-mining (SWIFT) Review software with evidence stream tags (human, animal, in vitro) yielding 12,800 unique records. SWIFT Active (a machine-learning software) further facilitated title/abstract screening. DistillerSR software was used for additional title/abstract, full-text screening, and data extraction.

RESULTS

187 studies were identified that met populations, exposures, comparators, and outcomes (PECO) criteria. From this pool, 54 human, 78 animal, and 61 genotoxicity studies were extracted into a literature inventory. Toxicological evidence was abundant for three azo dyes (also used as food additives) and sparse for five of the remaining 27 compounds. Complementary search in ECHA's REACH database for summaries of unpublished study reports revealed evidence for all 30 dyes. The question arose of how this information can be fed into an SEM process. Proper identification of prioritized dyes from various databases (including U.S. EPA's CompTox Chemicals Dashboard) turned out to be a challenge. Evidence compiled by this SEM project can be evaluated for subsequent use in problem formulation efforts to inform potential regulatory needs and prepare for a more efficient and targeted evaluation in the future for human health assessments.

MS2Tox Machine Learning Tool for Predicting the Ecotoxicity of Unidentified Chemicals in Water by Nontarget LC-HRMS

To achieve water quality objectives of the zero pollution action plan in Europe, rapid methods are needed to identify the presence of toxic substances in complex water samples. However, only a small fraction of chemicals detected with nontarget high-resolution mass spectrometry can be identified, and fewer have ecotoxicological data available. We hypothesized that ecotoxicological data could be predicted for unknown molecular features in data-rich high-resolution mass spectrometry (HRMS) spectra, thereby circumventing time-consuming steps of molecular identification and rapidly flagging molecules of potentially high toxicity in complex samples. Here, we present MS2Tox, a machine learning method, to predict the toxicity of unidentified chemicals based on high-resolution accurate mass tandem mass spectra (MS²). The MS2Tox model for fish toxicity was trained and tested on 647 lethal concentration (LC₅₀) values from the CompTox database and validated for 219 chemicals and 420 MS² spectra from MassBank. The root mean square error (RMSE) of MS2Tox predictions was below 0.89 log-mM, while the experimental repeatability of LC₅₀ values in CompTox was 0.44 log-mM. MS2Tox allowed accurate prediction of fish LC₅₀ values for 22 chemicals detected in water samples, and empirical evidence suggested the right directionality for another 68 chemicals. Moreover, by incorporating structural information, e.g., the presence of carbonyl-benzene, amide moieties, or hydroxyl groups, MS2Tox outperforms baseline models that use only the exact mass or log K_OW.

Migration of styrene oligomers from food contact materials: in silico prediction of possible genotoxicity

Styrene oligomers (SO) are well-known side products formed during styrene polymerization. They consist mainly of dimers (SD) and trimers (ST) that have been shown to be still residual in polystyrene (PS) materials. In this study migration of SO from PS into sunflower oil at temperatures between 5 and 70 °C and contact times between 0.5 h and 10 days was investigated. In addition, the contents of SD and ST in the fatty foodstuffs créme fraiche and coffee cream, which are typically enwrapped in PS, were measured and the amounts detected (of up to 0.123 mg/kg food) were compared to literature data. From this comparison, it became evident, that the levels of SO migrating from PS packaging into real food call for a comprehensive risk assessment. As a first step towards this direction, possible genotoxicity has to be addressed. Due to technical and experimental limitations, however, the few existing in vitro tests available are unsuited to provide a clear picture. In order to reduce uncertainty of these in vitro tests, four different knowledge and statistics-based in silico tools were applied to such SO that are known to migrate into food. Except for SD4 all evaluated SD and ST showed no alert for genotoxicity. For SD4, either the predictions were inconclusive or the substance was assigned as being out of the chemical space (out of domain) of the respective in silico tool. Therefore, the absence of genotoxicity of SD4 requires additional experimental proof. Apart from SD4, in silico studies supported the limited in vitro data that indicated the absence of genotoxicity of SO. In conclusion, the overall migration of all SO together into food of up to 50 µg/kg does not raise any health concerns, given the currently available in silico and in vitro data.

Implementing in vitro bioactivity data to modernize priority setting of chemical inventories

Internationally, there are thousands of existing and newly introduced chemicals in commerce, highlighting the ongoing importance of innovative approaches to identify emerging chemicals of concern. For many chemicals, there is a paucity of hazard and exposure data. Thus, there is a crucial need for efficient and robust approaches to address data gaps and support risk-based prioritization. Several studies have demonstrated the utility of in vitro bioactivity data from the ToxCast program in deriving points of departure (PODs). ToxCast contains data for nearly 1,400 endpoints per chemical, and the bioactivity concentrations, indicative of potential adverse outcomes, can be converted to human-equivalent PODs using high-throughput toxicokinetics (HTTK) modeling. However, data gaps need to be addressed for broader application: the limited chemical space of HTTK and quantitative high-throughput screening data. Here we explore the applicability of in silico models to address these data needs. Specifically, we used ADMET predictor for HTTK predictions and a generalized read-across approach to predict ToxCast bioactivity potency. We applied these models to profile 5,801 chemicals on Canada’s Domestic Substances List (DSL). To evaluate the approach’s performance, bioactivity PODs were compared with in vivo results from the EPA Toxicity Values database for 1,042 DSL chemicals. Comparisons demonstrated that the bioac­tivity PODs, based on ToxCast data or read-across, were conservative for 95% of the chemicals. Comparing bioactivity PODs to human exposure estimates supports the identification of chemicals of potential interest for further work. The bioac­tivity workflow shows promise as a powerful screening tool to support effective triaging of chemical inventories.

Integrating publicly available information to screen potential candidates for chemical prioritization under the Toxic Substances Control Act: A proof of concept case study using genotoxicity and carcinogenicity

The Toxic Substances Control Act (TSCA) became law in the U.S. in 1976 and was amended in 2016. The amended law requires the U.S. EPA to perform risk-based evaluations of existing chemicals. Here, we developed a tiered approach to screen potential candidates based on their genotoxicity and carcinogenicity information to inform the selection of candidate chemicals for prioritization under TSCA. The approach was underpinned by a large database of carcinogenicity and genotoxicity information that had been compiled from various public sources. Carcinogenicity data included weight-of-evidence human carcinogenicity evaluations and animal cancer data. Genotoxicity data included bacterial gene mutation data from the Salmonella (Ames) and Escherichia coli WP2 assays and chromosomal mutation (clastogenicity) data. Additionally, Ames and clastogenicity outcomes were predicted using the alert schemes within the OECD QSAR Toolbox and the Toxicity Estimation Software Tool (TEST). The evaluation workflows for carcinogenicity and genotoxicity were developed along with associated scoring schemes to make an overall outcome determination. For this case study, two sets of chemicals, the TSCA Active Inventory non-confidential portion list available on the EPA CompTox Chemicals Dashboard (33,364 chemicals, 'TSCA Active List') and a representative proof-of-concept (POC) set of 238 chemicals were profiled through the two workflows to make determinations of carcinogenicity and genotoxicity potential. Of the 33,364 substances on the 'TSCA Active List', overall calls could be made for 20,371 substances. Here 46.67%% (9507) of substances were non-genotoxic, 0.5% (103) were scored as inconclusive, 43.93% (8949) were predicted genotoxic and 8.9% (1812) were genotoxic. Overall calls for genotoxicity could be made for 225 of the 238 POC chemicals. Of these, 40.44% (91) were non-genotoxic, 2.67% (6) were inconclusive, 6.22% (14) were predicted genotoxic, and 50.67% (114) genotoxic. The approach shows promise as a means to identify potential candidates for prioritization from a genotoxicity and carcinogenicity perspective.