Evaluation of in silico models for the identification of respiratory sensitizers

Identification of true chemical respiratory allergens: Current status, limitations and recommendations

Asthma in the workplace is an important occupational health issue. It comprises various subtypes: occupational asthma (OA; both allergic asthma and irritant-induced asthma) and work-exacerbated asthma (WEA). Current regulatory paradigms for the management of OA are not fit for purpose. There is therefore an important unmet need, for the purposes of both effective human health protection and appropriate and proportionate regulation, that sub-types of work-related asthma can be accurately identified and classified, and that chemical respiratory allergens that drive allergic asthma can be differentiated according to potency. In this article presently available strategies for the diagnosis and characterisation of asthma in the workplace are described and critically evaluated. These include human health studies, clinical investigations and experimental approaches (structure-activity relationships, assessments of chemical reactivity, experimental animal studies and in vitro methods). Each of these approaches has limitations with respect to providing a clear discrimination between OA and WEA, and between allergen-induced and irritant-induced asthma. Against this background the needs for improved characterisation of work-related asthma, in the context of more appropriate regulation is discussed.

Making the Case for Quantum Mechanics in Predictive Toxicology─Nearly 100 Years Too Late?

The use of quantum mechanics (QM) has long been the norm to study covalent-binding phenomena in chemistry and biochemistry. The pharmaceutical industry leverages QM models explicitly in covalent drug discovery and implicitly to characterize short-range interactions in noncovalent binding. Predictive toxicology has resisted widespread adoption of QM, including in the pharmaceutical industry, despite its obvious relevance to the metabolic processes in the upstream of adverse outcome pathways and advances in both QM methods and computational resources, which support fit-for-purpose applications in reasonable timeframes. Here, we make the case for embracing QM as an indispensable part of a toxicologist's toolkit. We argue that QM provides the necessary orthogonality to alert-based expert systems and traditional QSARs, consistent with calls for animal-free integrated testing strategies for safety assessments of commercial chemicals. We outline existing roadblocks to this transition, including the need to train model developers in QM and the shift toward service-based toxicity models that utilize high-performance computing clusters. Lastly, we describe recent examples of successful implementations of QM in hazard assessments and propose how in silico toxicology can be further advanced by integrating QM with artificial intelligence.

Computer-Aided Discovery and Redesign for Respiratory Sensitization: A Tiered Mechanistic Model to Deliver Robust Performance Across a Diverse Chemical Space

Asthma is among the most common occupational diseases with considerable public health and economic costs. Chemicals that induce hypersensitivity in the airways can cause respiratory distress and comorbidities with respiratory infections such as COVID. Robust predictive models for this end point are still elusive due to the lack of an experimental benchmark and the over-reliance of existing in silico tools on structural alerts and structural (vs chemical) similarities. The Computer-Aided Discovery and REdesign (CADRE) platform is a proven strategy for providing robust computational predictions for hazard end points using a tiered hybrid system of expert rules, molecular simulations, and quantum mechanics calculations. The recently developed CADRE model for respiratory sensitization is based on a highly curated data set of structurally diverse chemicals with high-fidelity biological data. The model evaluates absorption kinetics in lung mucosa using Monte Carlo simulations, assigns reactive centers in a molecule and possible biotransformations via expert rules, and determines subsequent reactivity with cell proteins via quantum-mechanics calculations using a multi-tiered regression. The model affords an accuracy above 0.90, with a series of external validations based on literature data in the range of 0.88-0.95. The model is applicable to all low-molecular-weight organics and can inform not only chemical substitution but also chemical redesign to advance development of safer alternatives.

Methyl methacrylate and respiratory sensitisation: a comprehensive review

Methyl methacrylate (MMA) is classified under GHS as a weak skin sensitiser and a skin and respiratory irritant. It has recently been proposed that MMA be classified as a respiratory sensitiser (a designation that in a regulatory context embraces both true respiratory allergens, as well as chemicals that cause asthma through non-immunological mechanisms). This proposal was based primarily upon the interpretation of human data. This review, and a detailed weight of evidence analysis, has led to another interpretation of these data. The conclusion drawn is that persuasive evidence consistent with the designation of MMA as a respiratory sensitiser is lacking. It is suggested that one reason for different interpretations of these data is that occupational asthma poses several challenges with respect to establishing causation. Among these is that it is difficult to distinguish between allergic asthma, non-allergic asthma, and work-related exacerbation of pre-existing asthma. Moreover, there is a lack of methods for the identification of true chemical respiratory allergens. The characterisation and causation of occupational asthma is consequently largely dependent upon interpretation of human data of various types. Recommendations are made that are designed to improve the utility and interpretation of human data for establishing causation in occupational asthma.

An Explainable Supervised Machine Learning Model for Predicting Respiratory Toxicity of Chemicals Using Optimal Molecular Descriptors

Respiratory toxicity is a serious public health concern caused by the adverse effects of drugs or chemicals, so the pharmaceutical and chemical industries demand reliable and precise computational tools to assess the respiratory toxicity of compounds. The purpose of this study is to develop quantitative structure-activity relationship models for a large dataset of chemical compounds associated with respiratory system toxicity. First, several feature selection techniques are explored to find the optimal subset of molecular descriptors for efficient modeling. Then, eight different machine learning algorithms are utilized to construct respiratory toxicity prediction models. The support vector machine classifier outperforms all other optimized models in 10-fold cross-validation. Additionally, it outperforms the prior study by 2% in prediction accuracy and 4% in MCC. The best SVM model achieves a prediction accuracy of 86.2% and a MCC of 0.722 on the test set. The proposed SVM model predictions are explained using the SHapley Additive exPlanations approach, which prioritizes the relevance of key modeling descriptors influencing the prediction of respiratory toxicity. Thus, our proposed model would be incredibly beneficial in the early stages of drug development for predicting and understanding potential respiratory toxic compounds.

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

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

Asthma-inducing potential of 28 substances in spray cleaning products-Assessed by quantitative structure activity relationship (QSAR) testing and literature review

Exposure to spray cleaning products constitutes a potential risk for asthma induction. We set out to review whether substances in such products are potential inducers of asthma. We identified 101 spray cleaning products for professional use. Twenty‐eight of their chemical substances were selected. We based the selection on (a) positive prediction for respiratory sensitisation in humans based on quantitative structure activity relationship (QSAR) in the Danish (Q)SAR Database, (b) positive QSAR prediction for severe skin irritation in rabbits and (c) knowledge on the substances' physico‐chemical characteristics and toxicity. Combining the findings in the literature and QSAR predictions, we could group substances into four classes: (1) some indication in humans for asthma induction: chloramine, benzalkonium chloride; (2) some indication in animals for asthma induction: ethylenediaminetetraacetic acid (EDTA), citric acid; (3) equivocal data: hypochlorite; (4) few or lacking data: nitriloacetic acid, monoethanolamine, 2‐(2‐aminoethoxy)ethanol, 2‐diethylaminoethanol, alkyldimethylamin oxide, 1‐aminopropan‐2‐ol, methylisothiazolinone, benzisothiazolinone and chlormethylisothiazolinone; three specific sulphonates and sulfamic acid, salicylic acid and its analogue sodium benzoate, propane‐1,2‐diol, glycerol, propylidynetrimethanol, lactic acid, disodium malate, morpholine, bronopol and benzyl alcohol. In conclusion, we identified an asthma induction potential for some of the substances. In addition, we identified major knowledge gaps for most substances. Thus, more data are needed to feed into a strategy of safe‐by‐design, where substances with potential for induction of asthma are avoided in future (spray) cleaning products. Moreover, we suggest that QSAR predictions can serve to prioritise substances that need further testing in various areas of toxicology.

We reviewed whether substances in spray cleaning products constitute a potential risk for asthma induction. For this, we identified 101 spray cleaning products for professional use and prioritised their ingredient substances by use of quantitative structure activity relationship (QSAR). We provide a review of 28 selected substances: we give conclusions on their asthma induction potential, as well as a discussion on the use of QSAR for prioritisation of substances, and the major knowledge gaps we encountered.

Beyond dermal exposure: The respiratory tract as a target organ in hazard assessments of cosmetic ingredients

Dermal contact is the main route of exposure for most cosmetics; however, inhalation exposure could be significant for some formulations (e.g., aerosols, powders). Current cosmetic regulations do not require specific tests addressing respiratory irritation and sensitisation, and despite the prohibition of animal testing for cosmetics, no alternative methods have been validated to assess these endpoints to date. Inhalation hazard is mainly determined based on existing human and animal evidence, read-across, and extrapolation of data from different target organs or tissues, such as the skin. However, because of mechanistic differences, effects on the skin cannot predict effects on the respiratory tract, which indicates a substantial need for the development of new approach methodologies addressing respiratory endpoints for inhalable chemicals in general. Cosmetics might present a particularly significant need for risk assessments of inhalation exposure to provide a more accurate toxicological evaluation and ensure consumer safety. This review describes the differences in the mechanisms of irritation and sensitisation between the skin and the respiratory tract, the progress that has already been made, and what still needs to be done to fill the gap in the inhalation risk assessment of cosmetic ingredients.

Mapping Chemical Respiratory Sensitization: How Useful Are Our Current Computational Tools?

Chemical respiratory sensitization is an immunological process that manifests clinically mostly as occupational asthma and is responsible for 1 in 6 cases of adult asthma, although this may be an underestimate of the prevalence, as it is under-diagnosed. Occupational asthma results in unemployment for roughly one-third of those affected due to severe health issues. Despite its high prevalence, chemical respiratory sensitization is difficult to predict, as there are currently no validated models and the mechanisms are not entirely understood, creating a significant challenge for regulatory bodies and industry alike. The Adverse Outcome Pathway (AOP) for respiratory sensitization is currently incomplete. However, some key events have been identified, and there is overlap with the comparatively well-characterized AOP for dermal sensitization. Because of this, and the fact that dermal sensitization is often assessed by in vivo, in chemico, or in silico methods, regulatory bodies are defaulting to the dermal sensitization status of chemicals as a proxy for respiratory sensitization status when evaluating chemical safety. We identified a data set of known human respiratory sensitizers, which we used to investigate the accuracy of a structural alert model, Toxtree, designed for skin sensitization and the Centre for Occupational and Environmental Health (COEH)'s model, a model developed specifically for occupational asthma. While both models had a reasonable level of accuracy, the COEH model achieved the highest balanced accuracy at 76%; when the models agreed, the overall accuracy was 87%. There were important differences between the models: Toxtree had superior performance for some structural alerts and some categories of well-characterized skin sensitizers, while the COEH model had high accuracy in identifying sensitizers that lacked identified skin sensitization reactivity domains. Overall, both models achieved respectable accuracy. However, neither model addresses potency, which, along with data quality, remains a hurdle, and the field must prioritize these issues to move forward.

Modeling and insights into molecular basis of low molecular weight respiratory sensitizers

Respiratory sensitization has been considered an important toxicological endpoint, because of the severe risk to human health. A great part of sensitization events were caused by low molecular weight (< 1000) respiratory sensitizers in the past decades. However, there is currently no widely accepted test method that can identify prospective low molecular weight respiratory sensitisers. Herein, we performed the study of modeling and insights into molecular basis of low molecular weight respiratory sensitizers with a high-quality data set containing 136 respiratory sensitizers and 518 nonsensitizers. We built a number of classification models by using OCHEM tools, and a consensus model was developed based on the ten best individual models. The consensus model showed good predictive ability with a balanced accuracy of 0.78 and 0.85 on fivefold cross-validation and external validation, respectively. The readers can predict the respiratory sensitization of organic compounds via https://ochem.eu/article/114857 . The effect of several molecular properties on respiratory sensitization was also evaluated. The results indicated that these properties differ significantly between respiratory sensitizers and nonsensitizers. Furthermore, 14 privileged substructures responsible for respiratory sensitization were identified. We hope the models and the findings could provide useful help for environmental risk assessment.

How to assess respiratory sensitization of low molecular weight chemicals?

There are no validated and regulatory accepted (animal) models to test for respiratory sensitization of low molecular weight (LMW) chemicals. Since several decades such chemicals are classified as respiratory sensitizers almost exclusively based on observations in workers. However, both respiratory allergens (in which process the immune system is involved) as well as asthmagens (no involvement of the immune system) may induce the same type of respiratory symptoms. Correct classification is very important from a health's perspective point of view. On the other hand, over-classification is not preferable in view of high costs to overdue workplace engineering controls or the chemical ultimately being banned due to Authorities' decisions. It would therefore be very beneficial if respiratory sensitizers can be correctly identified and distinguished from skin sensitizers and non-sensitizers/respiratory irritants. The purpose of this paper is to consider whether LMW chemicals can be correctly identified based on the currently available screening methods in workers, and/or via in silico, in vitro and/or in vivo testing. Collectively, based on the available information further effort is still needed to be able to correctly identify respiratory sensitizers and to distinguish these from skin sensitizers and irritants, not at least because of the far-reaching consequences once a chemical is classified as a respiratory sensitizer.

Progress with Structure-Activity Relationship modelling of occupational chemical respiratory sensitizers

PURPOSE OF REVIEW

This appraises currently available computer-based ('in silico') models relating the molecular structure of low molecular weight compounds to their respiratory sensitization hazard. The present review places focus on the two main applications of such structure--activity relationship (SAR) models: hypotheses on disease mechanisms and toxicological prediction.

RECENT FINDINGS

Analyses of the chemical structures of low molecular weight organic compounds known to have caused occupational asthma has led to the development of mechanistic alerts usually based on electrophilic reaction chemistry and protein cross-linking potential. Protein cross-linking potential has also been found to be a consistent feature of chemicals that have caused human cases of hypersensitivity pneumonitis. Stepwise iteration of quantitative SAR (QSAR) modelling has shown appreciable improvements in predictivity for occupational asthma hazard and useful prospects for practical application. A good case has also been made for the potential use of structural alert-based mechanistic SARs in predictive toxicology.

SUMMARY

Further understanding of the molecular interactions between chemical respiratory sensitizers and components of human proteins have been obtained from in-vitro and in-silico techniques. There have been developments in both qualitative (mechanistic) SARs and QSARs, which offer potential for use in a predictive algorithm for the toxicological screening of industrial chemicals for respiratory sensitization potential.

Chemical-induced asthma and the role of clinical, toxicological, exposure and epidemiological research in regulatory and hazard characterization approaches

Uncertainties in understanding all potential modes-of-action for asthma induction and elicitation hinders design of hazard characterization and risk assessment methods that adequately screen and protect against hazardous chemical exposures. To address this challenge and identify current research needs, the University of Cincinnati and the American Cleaning Institute hosted a webinar series to discuss the current state-of-science regarding chemical-induced asthma. The general consensus is that the available database, comprised of data collected from routine clinical and validated toxicological tests, is inadequate for predicting or determining causal relationships between exposures and asthma induction for most allergens. More research is needed to understand the mechanism of asthma induction and elicitation in the context of specific chemical exposures and exposure patterns, and the impact of population variability and patient phenotypes. Validated tools to predict respiratory sensitization and to translate irritancy assays to asthma potency are needed, in addition to diagnostic biomarkers that assess and differentiate allergy versus irritant-based asthmatic responses. Diagnostic methods that encompass the diverse etiologies of asthmatic responses and incorporate robust exposure measurements capable of capturing different temporal patterns of complex chemical mixtures are needed. In the absence of ideal tools, risk assessors apply hazard-based safety assessment methods, in conjunction with active risk management, to limit potential asthma concerns, proactively identify new concerns, and ensure deployment of approaches to mitigate asthma-related risks.

ADMET Evaluation in Drug Discovery. Part 17: Development of Quantitative and Qualitative Prediction Models for Chemical-Induced Respiratory Toxicity

As a dangerous end point, respiratory toxicity can cause serious adverse health effects and even death. Meanwhile, it is a common and traditional issue in occupational and environmental protection. Pharmaceutical and chemical industries have a strong urge to develop precise and convenient computational tools to evaluate the respiratory toxicity of compounds as early as possible. Most of the reported theoretical models were developed based on the respiratory toxicity data sets with one single symptom, such as respiratory sensitization, and therefore these models may not afford reliable predictions for toxic compounds with other respiratory symptoms, such as pneumonia or rhinitis. Here, based on a diverse data set of mouse intraperitoneal respiratory toxicity characterized by multiple symptoms, a number of quantitative and qualitative predictions models with high reliability were developed by machine learning approaches. First, a four-tier dimension reduction strategy was employed to find an optimal set of 20 molecular descriptors for model building. Then, six machine learning approaches were used to develop the prediction models, including relevance vector machine (RVM), support vector machine (SVM), regularized random forest (RRF), extreme gradient boosting (XGBoost), naïve Bayes (NB), and linear discriminant analysis (LDA). Among all of the models, the SVM regression model shows the most accurate quantitative predictions for the test set (q²_ext = 0.707), and the XGBoost classification model achieves the most accurate qualitative predictions for the test set (MCC of 0.644, AUC of 0.893, and global accuracy of 82.62%). The application domains were analyzed, and all of the tested compounds fall within the application domain coverage. We also examined the structural features of the compounds and important fragments with large prediction errors. In conclusion, the SVM regression model and the XGBoost classification model can be employed as accurate prediction tools for respiratory toxicity.

Can the Direct Peptide Reactivity Assay Be Used for the Identification of Respiratory Sensitization Potential of Chemicals?

Prospective identification of low molecular weight respiratory sensitizers is difficult due to the current lack of adequate test methods. The direct peptide reactivity assay (DPRA) seems to be a promising method to determine the sensitization potential of chemicals because it determines the intrinsic characteristic of sensitizers to bind to proteins. It is already applied in the field of skin sensitization, and adaptation to respiratory sensitization has started recently. This article further evaluates the ability of the DPRA to predict the respiratory sensitization potential of chemicals. In addition, the added value of applying High Performance Liquid Chromatography (HPLC)-MS and measurements after 20 minutes and 24 hours of incubation was evaluated. Eighteen respiratory sensitizers (10 haptens, 3 prehaptens, and 5 prohaptens) and 14 nonsensitizers were tested with 2-model peptides. Based on peptide depletion, a prediction model was proposed for the identification of (respiratory) sensitizers. Application of mass spectrometry and measurements at 2 time-points increased prediction accuracy of the assay by resolving discordant results. The prediction model correctly identified all haptens and prehaptens as sensitizers. The 5 prohaptens were not identified as sensitizers, most likely due to lack of metabolic activity in the DPRA. All but 1 nonsensitizer was correctly predicted. The model, therefore, shows an accuracy of 78% for the tested dataset. Unfortunately, this assay cannot be used to distinguish respiratory from skin sensitizers. To make this distinction, the DPRA needs to be combined with other test methods that are able to identify respiratory sensitizers.

Development of an in vitro test to identify respiratory sensitizers in bronchial epithelial cells using gene expression profiling

Chemicals that induce asthma at the workplace are substances of concern. At present, there are no widely accepted methods to identify respiratory sensitizers, and classification of these substances is based on human occupational data. Several studies have contributed to understanding the mechanisms involved in respiratory sensitization, although uncertainties remain. One point of interest for respiratory sensitization is the reaction of the epithelial lung barrier to respiratory sensitizers. To elucidate potential molecular effects of exposure of the epithelial lung barrier, a gene expression profile was created based on a DNA microarray experiment using the bronchial epithelial cell line 16 HBE14o(-). The cells were exposed to 12 respiratory sensitizers and 10 non-sensitizers. For statistical analysis, we used a class prediction approach that combined three machine learning algorithms, leave-one-compound-out cross validation, and majority voting per tested compound. This approach allowed for a prediction accuracy of 95%. Identified predictive genes were mainly associated with the cytoskeleton and barrier function of the epithelial cell. Several of these genes were reported to be associated with asthma as well. Taken together, this indicates that pulmonary barrier function is an important target for respiratory sensitizers and associated genes can be used to predict the respiratory sensitization potential of chemicals.

A refined QSAR model for prediction of chemical asthma hazard

BACKGROUND

A previously developed quantitative structure-activity relationship (QSAR) model has been extern ally validated as a good predictor of chemical asthma hazard (sensitivity: 79-86%, specificity: 93-99%).

AIMS

To develop and validate a second version of this model.

METHODS

Learning dataset asthmagenic chemicals with molecular weight (MW) <1 kDa were identified from reports published in the peer-reviewed literature before the end of 2012. Control chemicals for which no reported case(s) of occupational asthma had been identified were selected at random from UK and US occupational exposure limit tables. MW banding was used in an attempt to categorically match the control group for MW distribution of the asthmagens. About 10% of chemicals in each MW category were excluded for use as an external validation set. An independent researcher utilized a logistic regression approach to compare the molecular descriptors present in asthmagens and controls. The resulting equation generated a hazard index (HI), with a value between zero and one, as an estimate of the probability that the chemical had asthmagenic potential. The HI was determined for each compound in the external validation set.

RESULTS

The model development sets comprised 99 chemical asthmagens and 204 controls. The external validation showed that using a cut-point HI of 0.39, 9/10 asthmagenic (sensitivity: 90%) and 23/24 non-asthmagenic (specificity: 96%) compounds were correctly predicted. The new QSAR model showed a better receiver operating characteristic plot than the original.

CONCLUSIONS

QSAR refinement by iteration has resulted in an improved model for the prediction of chemical asthma hazard.

Thresholds in chemical respiratory sensitisation

There is a continuing interest in determining whether it is possible to identify thresholds for chemical allergy. Here allergic sensitisation of the respiratory tract by chemicals is considered in this context. This is an important occupational health problem, being associated with rhinitis and asthma, and in addition provides toxicologists and risk assessors with a number of challenges. In common with all forms of allergic disease chemical respiratory allergy develops in two phases. In the first (induction) phase exposure to a chemical allergen (by an appropriate route of exposure) causes immunological priming and sensitisation of the respiratory tract. The second (elicitation) phase is triggered if a sensitised subject is exposed subsequently to the same chemical allergen via inhalation. A secondary immune response will be provoked in the respiratory tract resulting in inflammation and the signs and symptoms of a respiratory hypersensitivity reaction. In this article attention has focused on the identification of threshold values during the acquisition of sensitisation. Current mechanistic understanding of allergy is such that it can be assumed that the development of sensitisation (and also the elicitation of an allergic reaction) is a threshold phenomenon; there will be levels of exposure below which sensitisation will not be acquired. That is, all immune responses, including allergic sensitisation, have threshold requirement for the availability of antigen/allergen, below which a response will fail to develop. The issue addressed here is whether there are methods available or clinical/epidemiological data that permit the identification of such thresholds. This document reviews briefly relevant human studies of occupational asthma, and experimental models that have been developed (or are being developed) for the identification and characterisation of chemical respiratory allergens. The main conclusion drawn is that although there is evidence that the acquisition of sensitisation to chemical respiratory allergens is a dose-related phenomenon, and that thresholds exist, it is frequently difficult to define accurate numerical values for threshold exposure levels. Nevertheless, based on occupational exposure data it may sometimes be possible to derive levels of exposure in the workplace, which are safe. An additional observation is the lack currently of suitable experimental methods for both routine hazard characterisation and the measurement of thresholds, and that such methods are still some way off. Given the current trajectory of toxicology, and the move towards the use of non-animal in vitro and/or in silico) methods, there is a need to consider the development of alternative approaches for the identification and characterisation of respiratory sensitisation hazards, and for risk assessment.