Computational Toxicology最新文献

The use of in silico modeling tools for predictive toxicology has potential to improve force health protection in the military by helping to efficiently evaluate the risk of adverse health effects from operational exposures. Thus, a systematic review was performed to understand if existing quantitative structure–activity relationship (QSAR) models for tissue-specific toxicity were potentially adaptable for use in risk assessments of military-relevant exposures. Within this systematic review, we assessed 563 peer-reviewed publications in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 guidelines and Organization for Economic Co-operation and Development (OECD) 2023 quantitative structural-activity relationship Assessment Framework. From these publications, we further evaluated 129 existing models that utilize QSAR and tissue-specific data for predicting toxicity in the following tissues: liver (i.e., hepatotoxicity), heart (i.e., cardiotoxicity), lung (i.e., respiratory toxicity), the central nervous system (neurotoxicity), and kidney (i.e., nephrotoxicity). The methodology, performance, and accessibility of these models and analysis code were thoroughly documented and then assessed to determine advancements and inadequacies for occupational and military application. While ∼ 58 % of the 129 tissue-specific QSAR approaches followed at least 3 OECD guidelines, there were only 8 tissue-specific models that satisfied all screening criteria. The most common criteria not satisfied was mechanistic interpretation of the model (i.e., OECD criteria number five). Furthermore, while the greatest number of publications and models were available for the liver, many of them were for pharmaceutical applications. Moreover, there were limited available models for heart and kidney for any application. In conclusion, our findings underscore the necessity for additional and updated tissue-specific QSAR models to predict various organ-specific targets while addressing military specific needs. Furthermore, increased publication of model workflows or user-friendly applications are crucial to enhancing model accessibility. In this systematic review, we provide an overview of the databases, resources, and future strategies to advance tissue-specific QSAR model development.

In silico systems can reduce the need for (animal) testing, increase human safety and support critical decisions. They are increasingly being cited in regulatory guidance documents and are forming a key element of New Approach Methodologies (NAMs). Performance is being improved through a combination of new methodologies, increased understanding of mechanistic toxicology and access to experimental data from new assays. Trust and acceptance of in silico methodologies requires them to be accurate and transparent while also providing an explanation and confidence-assessment for each prediction. This paper summarises the state-of-art of in silico models and provides an action plan for further advances in this field.

Per- and Polyfluoroalkyl substances (PFAS) are a class of manufactured chemicals that are in widespread use and many present concerns for persistence, bioaccumulation and toxicity. Whilst a handful of PFAS have been characterized for their hazard profiles, the vast majority have not been extensively studied. Herein, a chemical category approach was developed and applied to PFAS that could be readily characterized by a chemical structure. The PFAS definition as described in the Toxic Substances Control Act (TSCA) section 8(a)(7) rule was applied to the Distributed Structure-Searchable Toxicity (DSSTox) database to retrieve an initial list of 13,054 PFAS. Plausible degradation products from the 563 PFAS on the non-confidential TSCA Inventory were simulated using the Catalogic expert system, and the unique predicted PFAS degradants (2484) that conformed to the same PFAS definition were added to the list resulting in a set of 15,538 PFAS. Each PFAS was then assigned into a primary category using Organisation for Economic Co-operation and Development (OECD) structure-based classifications. The primary categories were subdivided into secondary categories based on a chain length threshold (>=7 vs < 7). Secondary categories were subcategorized using chemical fingerprints to achieve a balance between total number of structural categories vs. level of structural similarity within a category based on the Jaccard index. A set of 128 terminal structural categories were derived from which a subset of representative candidates could be proposed for potential data collection, considering the sparsity of relevant toxicity data within each category, presence on environmental monitoring lists, and the ability to identify plausible manufacturers/importers. Refinements to the approach taking into consideration ways in which the categories could be updated by mechanistic data and physicochemical property information are also described. This categorization approach may be used to form the basis of identifying candidates for data collection with related applications in QSAR development, read-across and hazard assessment.

There is international interest in using alternatives to animal testing, including (Q)SARs, in chemical hazard assessments. The regulatory acceptance of alternative methods requires principles for considering the scientific rigour of methods and their results. The OECD (Q)SAR assessment Framework (QAF) was developed as guidance for regulators when considering (Q)SAR models and predictions in chemical evaluation. The QAF builds on existing principles for evaluating models and, learning from the longstanding regulatory experience in assessing (Q)SAR predictions, establishes new principles for evaluating predictions and results from multiple predictions. Assessment elements, identified for all principles lay out criteria for assessing the confidence and uncertainties in (Q)SAR models and predictions, while maintaining the flexibility necessary to adapt to different regulatory contexts and purposes. Using the QAF, assessors can consistently and transparently evaluate and decide on the validity of (Q)SARs, and model developers and users have clear requirements to meet. The publication of the QAF is expected to increase the regulatory use and acceptance of (Q)SARs and may become an example to build similar prescriptive frameworks for other new approach methodologies (NAMs). This article provides an overview of the main scientific aspects of the QAF guidance and provides context for how this guidance can promote the use of alternative methods in chemical assessments.

Developmental and reproductive toxicity (DART) are key regulatory endpoints for the protection of human health. DART assessments require large numbers of animals, are expensive and often run at late stages of drug development. Therefore, new approach methodologies (NAMs) are being developed to transition away from animal testing. These NAMs (including in silico models) can be used to screen for DART hazards at the early stages of compound development and may in the future be used for regulatory DART assessments. Due to the implications of a mischaracterised developmental toxicant, both high confidence and understanding of the assessments made using NAMs will be required; it is likely that multiple NAMs will be needed in order to replace the current animal-based assessments. Adverse outcome pathways (AOPs) serve as a pragmatic tool for documenting mechanisms of toxicity. NAMs can be associated to key events (KEs) along an AOP, providing context to their outputs, and therefore increasing confidence in their use. It is likely that networks of pathways will be required for a specific toxicity endpoint in order to confidently apply an AOP-based approach to safety assessments. An insufficient number of DART AOPs are currently described within the public domain; therefore, using a literature-based approach, a network consisting of 340 KEs (including 68 MIEs) was developed. This foundation of pathways was made chemically aware through the association of relevant assays, data and expert rule-based structural alerts to appropriate KEs. The use of the network as a hazard screening tool was assessed, and the application of this to aid an ICH S5 workflow investigated. The knowledge captured within this AOP network can also guide the further development and use of DART-relevant NAMs and integrated approaches to testing and assessments (IATAs).

We describe the development of the GSK vEXP (virtual enhanced cross screen panel) for off-target pharmacology alerts. The derivation of a panel of machine learning classification models or QSAR models (Quantitative Structure-Activity Relationship) for off-target safety assessment allows early alerting to risk factors in candidate drugs. The models are matched to an internal in-vitro biochemical screening panel described previously with some updates reported here. The extreme imbalance of some internal GSK datasets and most of the related external ChEMBL datasets is shown when considering potency thresholds relevant to in-vitro screening. The small size and bias to the active class make many ChEMBL datasets un-modellable using such thresholds. Although larger, many GSK datasets remain too imbalanced to give a performant model. The value of merging internal and external data to help rebalance datasets and improve the domain of applicability is demonstrated with improvements in model performance frequently seen on merged data. Efforts to collate public datasets with a far better balance of the missing in-actives would likely do more to improve opensource models than simply increasing dataset size. We investigate the use of moving the probability threshold and applying imbalanced learners to help overcome the imbalance problem. Both methods can produce models with improved performance when applied to imbalanced datasets. Datasets with class imbalance 95:5 % or with <100 compounds were un-modellable. Where datasets had a class imbalance of 90:10 % the imbalanced learn methods were often more performant than standard tree-based classifiers. No one classification algorithm consistently out-performed all others and our approach emphasises a standardised, automated build and evaluate approach across all classifiers to identify the best model. The application of vEXP includes ranking of hit compounds for fast prioritisation, flagging of hit series that contain systematic scaffold or functional group related risks and the confirmation that late-stage optimisation is not introducing new off-target activities in established chemical series.

Cosmetic ingredients of choice are those appropriate for a specific commercial use and deemed safer than existing alternatives. In the LIFE VERMEER project (https://www.life-vermeer.eu/), the ToxEraser Cosmetics software was developed as a platform under which an ingredient is presented with a list of potential substitutes, from an archive of 2233 items. Key information about the safety of each item concerns: (a) the risk assessment addressed by seven regulatory and other specialized European-US authorities; (b) the safety class emerging from the systematic evaluation and integration of each authority’s assessment. Read-across analysis makes the substitution possible even when the ingredient is not included in the archive. The list of alternatives can be extended or reduced flexibly, since the commercial use of cosmetics is dictated by attributes indicating progressively detailed and hierarchically related categories. Finally, the identification of significant validated structural alerts for endpoints of interest serves in detecting which part of the structure is associated with certain hazardous properties. This tool will be joined with VERMEER Cosmolife, the other tool for cosmetics developed as part of the VERMEER project. ToxEraser offers a systematic, flexible approach to explore safer cosmetic substitutes, acknowledging the sources of evidence produced by VERMEER Cosmolife, offering a forward-looking tool for the cosmetic sector. More in general, the novelty is the shift to in silico models, not only to assess possible concern associated with a substance, but also to move towards safer alternatives.

Electronic nicotine delivery systems (ENDS) heat a liquid solution typically containing propylene glycol, vegetable glycerin, water, nicotine, and flavor chemicals to deliver an aerosol to the user. ENDS aerosols are complex, multi-constituent mixtures of droplets and vapors. Lung dosimetry predictions require mechanistic models that account for the physico-chemical properties of the constituents and thermodynamic processes of the aerosol as it travels through the respiratory tract and deposits in lung airways. In this study, a model formulated to predict ENDS aerosol deposition in the oral cavity and lung airways was linked with a physiologically-based pharmacokinetic (PBPK) model to predict nicotine pharmacokinetics (PK) as a function of product characteristics and puff topography. Predicted plasma nicotine PK compared favorably with available experimental data and captured the rapid increase in plasma levels followed by a clearance phase after ENDS use. E-liquid nicotine concentration and puff duration substantially increased nicotine lung deposition and plasma nicotine levels. Increasing the puff duration from 1 to 5 s while assuming a constant aerosol flow rate resulted in an ∼5-fold increase in nicotine lung deposition (45.0 µg to 243.7 µg) and increased maximum plasma nicotine concentrations from 4.7 ng/mL to 25.0 ng/mL; increasing the e-liquid nicotine concentration from 1 % to 5 % yielded increases in nicotine lung deposition (41.0 µg to 204.5 µg) and maximum plasma nicotine concentration (4.2 ng/mL to 21.1 ng/mL). Model predictions demonstrate the sensitivity of ENDS aerosol lung deposition and plasma nicotine profiles to user behavior and allow for quantification of constituent deposition and nicotine absorption after ENDS use.

Monitoring oral exposure to pesticides in wildlife is crucial for assessing environmental risks and preventing adverse effects on non-target species. Traditionally, this requires invasive tissue sampling, raising ethical, regulatory, and economic concerns. To address this gap, our study aims to develop a method for assessing external oral dose levels in rats using physiologically-based kinetic (PBK) modeling based on blood concentration levels of two pesticides, imidacloprid and carbendazim, and one of their primary metabolites. We utilized in vitro metabolic kinetic data from hepatic microsomal and S9 incubations to inform our models. These models were then evaluated by comparing their predictions with existing in vivo experimental data from the literature. Our results demonstrate that the models provide accurate predictions, presenting a novel in vitro and in silico approach for environmental exposure and risk assessment of pesticides. This methodology has the potential for application in wildlife species, advancing the frontier of knowledge in non-invasive pesticide exposure assessment.

There is widespread acceptance that non-animal studies can be used to assess chemical safety in humans. These New Approach Methodologies (NAMs) typically integrate data from multiple sources including in silico and in vitro models. Regulatory guidelines are being updated to recognise that these scientific advances are allowing animal studies to be replaced without compromising human safety. One such regulation, ICH S1B(R1), was updated in 2022 to include the provision for a weight-of-evidence assessment for carcinogenicity, using six factors to determine if there was sufficient evidence to waive the need to run a rat carcinogenicity assay. The volume of data and evidence, however, can be hard to organise and interpret into a cohesive evaluation. To aid such assessments, software has been developed that combines adverse outcome pathways (AOPs) and reasoning, to organise and contextualise knowledge, and provide an outcome based on the data available. Using this framework, a workflow has been developed to assess the initial outcome and structure expert review to investigate the factors, and potential biological mechanisms which could contribute to a compound’s carcinogenic potential (or lack thereof). The framework was used to structure expert review of three examples of differing activity and levels of supporting evidence. This highlighted where AOPs supported expert review by showing 1) the value in using AOPs to analyse data, 2) the importance of expert review to strengthen confidence in outcomes, and 3) how this approach can accurately predict experimental results. Therefore, using this approach to assess evidence for ICH S1B(R1) will give transparent, scientifically robust, and reproducible calls, and thus reduce the need for rat carcinogenicity studies.