Can "Hazard-Cost-Effectiveness Analysis" improve the risk management of chemicals under REACH?

Cost-Effectiveness Analysis (CEA) is a decision-making framework to prioritize policy decisions for chemicals. Differences in hazard profiles among chemicals are not integrated in CEA under the EU REACH Regulation, which could limit its relevance. Another concern is that two different economic decision support methods (CEA for chemicals considered as PBTs or vPvBs from a regulatory perspective and Cost Benefit Analysis (CBA) for others) are used under REACH. To address this situation, we define "Hazard" CEA by integrating a hazard score, based on persistence, bioaccumulation and (eco)toxicity, in the effect indicator of CEA. We test different designs and parameterizations of Hazard-CEA on a set of past socio-economic assessments under REACH for PBT and non-PBT chemicals. Weighing and thresholds in hazard scores do not have a significant impact on the outcome of Hazard-CEA but the design of the hazard scoring method does. We suggest using an integrated and unweighted scoring method with a multiplicative formulation based on the notion of risk. Hazard-CEA could be used for both PBT and non-PBT chemicals, to use a single method in REACH and therefore improve consistency in policy decisions. Our work also suggests that using Hazard-CEA could help make decision easier.

Read-across and new approach methodologies applied in a 10-step framework for cosmetics safety assessment - A case study with parabens.

Parabens are esters of para-hydroxybenzoic acid that have been used as preservatives in many types of products for decades including agrochemicals, pharmaceuticals, food and cosmetics. This illustrative case study with propylparaben (PP) demonstrates a 10-step read-across (RAX) framework in practice. It aims at establishing a proof-of-concept for the value added by new approach methodologies (NAMs) in read-across (RAX) for use in a next-generation risk assessment (NGRA) in order to assess consumer safety after exposure to PP-containing cosmetics. In addition to structural and physico-chemical properties, in silico information, toxicogenomics, in vitro toxicodynamic, toxicokinetic data from PBK models, and bioactivity data are used to provide evidence of the chemical and biological similarity of PP and analogues and to establish potency trends for observed effects in vitro. The chemical category under consideration is short (C1-C4) linear chain n-alkyl parabens: methylparaben, ethylparaben, propylparaben and butylparaben. The goal of this case study is to illustrate how a practical framework for RAX can be used to fill a hypothetical data gap for reproductive toxicity of the target chemical PP.

In Silico Prediction of Chemically Induced Mutagenicity: A Weight of Evidence Approach Integrating Information from QSAR Models and Read-Across Predictions.

Information on genotoxicity is an essential piece of information in the framework of several regulations aimed at evaluating chemical toxicity. In this context, QSAR models that can predict Ames genotoxicity can conveniently provide relevant information. Indeed, they can be straightforwardly and rapidly used for predicting the presence or absence of genotoxic hazards associated with the interactions of chemicals with DNA. Nevertheless, and despite their ease of use, the main interpretative challenge is related to a critical assessment of the information that can be gathered, thanks to these tools. This chapter provides guidance on how to use freely available QSAR and read-across tools provided by VEGA HUB and on how to interpret their predictions according to a weight-of-evidence approach.

Integrate mechanistic evidence from new approach methodologies (NAMs) into a read-across assessment to characterise trends in shared mode of action.

Read-across approaches often remain inconclusive as they do not provide sufficient evidence on a common mode of action across the category members. This read-across case study on thirteen, structurally similar, branched aliphatic carboxylic acids investigates the concept of using human-based new approach methods, such as in vitro and in silico models, to demonstrate biological similarity. Five out of the thirteen analogues have preclinical in vivo studies. Three out of them induced lipid accumulation or hypertrophy in preclinical studies with repeated exposure, which leads to the read-across hypothesis that the analogues can potentially induce hepatic steatosis. To confirm the selection of analogues, the expression patterns of the induced differentially expressed genes (DEGs) were analysed in a human liver model. With increasing dose, the expression pattern within the tested analogues got more similar, which serves as a first indication of a common mode of action and suggests differences in the potency of the analogues. Hepatic steatosis is a well-known adverse outcome, for which over 55 adverse outcome pathways have been identified. The resulting adverse outcome pathway (AOP) network, comprised a total 43 MIEs/KEs and enabled the design of an in vitro testing battery. From the AOP network, ten MIEs, early and late KEs were tested to systematically investigate a common mode of action among the grouped compounds. The targeted testing of AOP specific MIE/KEs shows that biological activity in the category decreases with side chain length. A similar trend was evident in measuring liver alterations in zebra fish embryos. However, activation of single MIEs or early KEs at in vivo relevant doses did not necessarily progress to the late KE "lipid accumulation". KEs not related to the read-across hypothesis, testing for example general mitochondrial stress responses in liver cells, showed no trend or biological similarity. Testing scope is a key issue in the design of in vitro test batteries. The Dempster-Shafer decision theory predicted those analogues with in vivo reference data correctly using one human liver model or the CALUX reporter assays. The case study shows that the read-across hypothesis is the key element to designing the testing strategy. In the case of a good mechanistic understanding, an AOP facilitates the selection of reliable human in vitro models to demonstrate a common mode of action. Testing DEGs, MIEs and early KEs served to show biological similarity, whereas the late KEs become important for confirmation, as progression from MIEs to AO is not always guaranteed.

Evaluation of the OECD QSAR toolbox automatic workflow for the prediction of the acute toxicity of organic chemicals to fathead minnow.

Regulatory frameworks require information on acute fish toxicity to ensure environmental protection. The experimental assessment of this property relies on a substantial number of fish to be tested and it is in conflict with the current drive to replace in vivo testing. For this reason, alternatives to in vivo testing have been proposed during the past years. Among these alternatives, there are Quantitative Structure-Activity Relationships (QSAR) that require the sole knowledge of chemical structure to yield predictions of toxicities. In this context, the OECD QSAR Toolbox is one of the leading QSAR tools for regulatory purposes that enables the prediction of fish toxicities. The aim of this work is to provide evidence about the predictive reliability of the automated workflow for predicting acute toxicity in fish which is embedded within this toolbox. The results herein presented show that the logic underpinning this automated workflow can predict with a reliability that, in the majority of cases, is comparable to inter-laboratory variability and, in a significant number of cases, is also comparable with intra-laboratory variability. Moreover, considerations on the toxic mode of action provided by the OECD tool proved to be helpful in refining predictions and reducing the number of prediction outliers.

Prediction of the Partition Coefficient between Adipose Tissue and Blood for Environmental Chemicals: From Single QSAR Models to an Integrated Approach.

The adipose tissue:blood partition coefficient is a key-endpoint to predict the pharmacokinetics of chemicals in humans and animals, since other organ:blood affinities can be estimated as a function of this parameter. We performed a search in the literature to select all the available rat in vivo data. This approach resulted into two improvements to existing models: a homogeneous definition of the endpoint and an expanded data collection. The resulting dataset was used to develop QSAR models as a function of linear and non-linear algorithms. Several applicability domain definitions were assessed and the definition corresponding to a good balance between performance and coverage was retained. We assessed the pertinence of combining single models into integrated approaches to increase the accuracy in predictions. The best integrated model outperformed the single models and it was characterized by an external mean absolute error (MAE) equal to 0.26, while preserving an adequate coverage (84 %). This performance is comparable to experimental variability and it highlights the pertinence of the integrated model.

QSAR Modeling of ToxCast Assays Relevant to the Molecular Initiating Events of AOPs Leading to Hepatic Steatosis.

Nonalcoholic hepatic steatosis is a worldwide epidemiological concern since it is among the most prominent hepatic diseases. Indeed, research in toxicology and epidemiology has gathered evidence that exposure to endocrine disruptors can perturb cellular homeostasis and cause this disease. Therefore, assessing the likelihood of a chemical to trigger hepatic steatosis is a matter of the utmost importance. However, systematic in vivo testing of all the chemicals humans are exposed to is not feasible for ethical and economical reasons. In this context, predicting the molecular initiating events (MIE) leading to hepatic steatosis by QSAR modeling is an issue of practical relevance in modern toxicology. In this article, we present QSAR models based on random forest classifiers and DRAGON molecular descriptors for the prediction of in vitro assays that are relevant to MIEs leading to hepatic steatosis. These assays were provided by the ToxCast program and proved to be predictive for the detection of chemical-induced steatosis. During the modeling process, special attention was paid to chemical and toxicological data curation. We adopted two modeling strategies (undersampling and balanced random forests) to develop robust QSAR models from unbalanced data sets. The two modeling approaches gave similar results in terms of predictivity, and most of the models satisfy a minimum percentage of correctly predicted chemicals equal to 75%. Finally, and most importantly, the developed models proved to be useful as an effective in silico screening test for hepatic steatosis.

Adverse outcome pathways: opportunities, limitations and open questions.

Adverse outcome pathways (AOPs) are a recent toxicological construct that connects, in a formalized, transparent and quality-controlled way, mechanistic information to apical endpoints for regulatory purposes. AOP links a molecular initiating event (MIE) to the adverse outcome (AO) via key events (KE), in a way specified by key event relationships (KER). Although this approach to formalize mechanistic toxicological information only started in 2010, over 200 AOPs have already been established. At this stage, new requirements arise, such as the need for harmonization and re-assessment, for continuous updating, as well as for alerting about pitfalls, misuses and limits of applicability. In this review, the history of the AOP concept and its most prominent strengths are discussed, including the advantages of a formalized approach, the systematic collection of weight of evidence, the linkage of mechanisms to apical end points, the examination of the plausibility of epidemiological data, the identification of critical knowledge gaps and the design of mechanistic test methods. To prepare the ground for a broadened and appropriate use of AOPs, some widespread misconceptions are explained. Moreover, potential weaknesses and shortcomings of the current AOP rule set are addressed (1) to facilitate the discussion on its further evolution and (2) to better define appropriate vs. less suitable application areas. Exemplary toxicological studies are presented to discuss the linearity assumptions of AOP, the management of event modifiers and compensatory mechanisms, and whether a separation of toxicodynamics from toxicokinetics including metabolism is possible in the framework of pathway plasticity. Suggestions on how to compromise between different needs of AOP stakeholders have been added. A clear definition of open questions and limitations is provided to encourage further progress in the field.

In Silico Prediction of Chemically Induced Mutagenicity: How to Use QSAR Models and Interpret Their Results.

Information on genotoxicity is an essential piece of information gathering for a comprehensive toxicological characterization of chemicals. Several QSAR models that can predict Ames genotoxicity are freely available for download from the Internet and they can provide relevant information for the toxicological profiling of chemicals. Indeed, they can be straightforwardly used for predicting the presence or absence of genotoxic hazards associated with the interactions of chemicals with DNA.Nevertheless, and despite the ease of use of these models, the scientific challenge is to assess the reliability of information that can be obtained from these tools. This chapter provides instructions on how to use freely available QSAR models and on how to interpret their predictions.

Evaluation of QSAR models for the prediction of ames genotoxicity: a retrospective exercise on the chemical substances registered under the EU REACH regulation.

We evaluated the performance of seven freely available quantitative structure-activity relationship models predicting Ames genotoxicity thanks to a dataset of chemicals that were registered under the EU Registration, Evaluation, Authorization and Restriction of Chemicals (REACH) regulation. The performance of the models was estimated according to Cooper's statistics and Matthew's Correlation Coefficients (MCC). The Benigni/Bossa rule base originally implemented in Toxtree and re-implemented within the Virtual models for property Evaluation of chemicals within a Global Architecture (VEGA) platform displayed the best performance (accuracy = 92%, sensitivity = 83%, specificity = 93%, MCC = 0.68) indicating that this rule base provides a reliable tool for the identification of genotoxic chemicals. Finally, we elaborated a consensus model that outperformed the accuracy of the individual models.

Photolysis of estrone generates estrogenic photoproducts with higher activity than the parent compound.

In the present study, we aimed to evaluate the effect of UV-visible irradiation on the estrogenicity of an estrone aqueous solution by using chemical analysis associated with an in vitro bioassay and in silico analysis. An estrone aqueous solution was irradiated with an UV-visible high-pressure mercury lamp. By using the MELN in vitro cellular bioassay, based on the induction of a luciferase reporter gene upon the activation of the estrogen receptor by chemicals, we showed that the estrogenic potency of the solution increased after irradiation. High-performance liquid chromatography fractionation of the photolyzed solution followed by in vitro testing of fractions allowed the quantitation of the estrogenic potency of each fraction. Nine photoproducts were detected and characterized by liquid chromatography-mass spectrometry coupling. The observed estrogenic activity is mediated by mono- and multi-hydroxylated photoproducts; it is influenced by the position of hydroxyl groups on the steroidal skeleton. In addition, a structure-activity analysis of the hydroxylated photoproducts confirmed their ability to act as estrogen receptor ligands.

A physiologically based toxicokinetic model for the zebrafish Danio rerio.

Zebrafish (Danio rerio) is a widely used model for toxicological studies, in particular those related to investigations on endocrine disruption. The development and regulatory use of in vivo and in vitro tests based on this species can be enhanced by toxicokinetic modeling. For this reason, we propose a physiologically based toxicokinetic (PBTK) model for zebrafish describing the uptake and disposition of organic chemicals. The model is based on literature data on zebrafish, other cyprinidae and other fish families, new experimental physiological information (volumes, lipids and water contents) obtained from zebrafish, and chemical-specific parameters predicted by generic models. The relevance of available models predicting the latter parameters was evaluated with respect to gill uptake and partition coefficients in zebrafish. This evaluation benefited from the fact that the influence of confounding factors such as body weight and temperature on ventilation rate was included in our model. The predictions for six chemicals (65 data points) yielded by our PBTK model were compared to available toxicokinetics data for zebrafish and 88% of them were within a factor of 5 of the corresponding experimental values. Sensitivity analysis highlighted that the 1-octanol/water partition coefficient, the metabolism rate, and all the parameters that enable the prediction of assimilation efficiency and partitioning of chemicals need to be precisely determined in order to allow an effective toxicokinetic modeling.

Prediction of dose-hepatotoxic response in humans based on toxicokinetic/toxicodynamic modeling with or without in vivo data: a case study with acetaminophen.

In the present legislations, the use of methods alternative to animal testing is explicitly encouraged, to use animal testing only 'as a last resort' or to ban it. The use of alternative methods to replace kinetics or repeated dose in vivo tests is a challenging issue. We propose here a strategy based on in vitro tests and QSAR (Quantitative Structure Activity Relationship) models to calibrate a dose-response model predicting hepatotoxicity. The dose response consists in calibrating and coupling a PBPK (physiologically-based pharmacokinetic) model with a toxicodynamic model for cell viability. We applied our strategy to acetaminophen and compared three different ways to calibrate the PBPK model: only with in vitro and in silico methods, using rat data or using all available data including data on humans. Some estimates of kinetic parameters differed substantially among the three calibration processes, but, at the end, the three models were quite comparable in terms of liver toxicity predictions and close to the usual range of human overdose. For the model based on alternative methods, the good adequation with the two other models resulted from an overestimated renal elimination rate which compensated for the underestimation of the metabolism rate. Our study points out that toxicokinetics/toxicodynamics approaches, based on alternative methods and modelling only, can predict in vivo liver toxicity with accuracy comparable to in vivo methods.

Modelling Structure Activity Landscapes with Cliffs: a Kernel Regression-Based Approach.

Quantitative Structure-Activity Relationship (QSAR) models are increasingly used in hazard and risk assessment. Even when models with linear relationships between activity and a small number of descriptors are built and validated regarding predictivity and statistical assumptions, similar structures can exhibit large differences in activity known as similarity paradoxes or activity cliffs. In order to reduce the impact that similarity paradoxes can have on predictions we have devised a statistical method based on Nadaraya-Watson kernel regression. According to our method, activity cliffs filter out contributions of neighbouring chemicals especially along the cliff axis. Our method decreases density-based certainty in particular for chemicals with strong prediction errors and the implementation of Structure-Activity Landscape Index (SALI) curves shows that our method improves the prediction of activity cliff ranks. We also provide useful indications on the density-based applicability domain and the reliability of individual predictions.

Applying quantitative structure-activity relationship approaches to nanotoxicology: current status and future potential.

The potential (eco)toxicological hazard posed by engineered nanoparticles is a major scientific and societal concern since several industrial sectors (e.g. electronics, biomedicine, and cosmetics) are exploiting the innovative properties of nanostructures resulting in their large-scale production. Many consumer products contain nanomaterials and, given their complex life-cycle, it is essential to anticipate their (eco)toxicological properties in a fast and inexpensive way in order to mitigate adverse effects on human health and the environment. In this context, the application of the structure-toxicity paradigm to nanomaterials represents a promising approach. Indeed, according to this paradigm, it is possible to predict toxicological effects induced by chemicals on the basis of their structural similarity with chemicals for which toxicological endpoints have been previously measured. These structure-toxicity relationships can be quantitative or qualitative in nature and they can predict toxicological effects directly from the physicochemical properties of the entities (e.g. nanoparticles) of interest. Therefore, this approach can aid in prioritizing resources in toxicological investigations while reducing the ethical and monetary costs that are related to animal testing. The purpose of this review is to provide a summary of recent key advances in the field of QSAR modelling of nanomaterial toxicity, to identify the major gaps in research required to accelerate the use of quantitative structure-activity relationship (QSAR) methods, and to provide a roadmap for future research needed to achieve QSAR models useful for regulatory purposes.

A Kernel-Based Method for Assessing Uncertainty on Individual QSAR Predictions.

The assessment of uncertainty attached to individual predictions is now a priority for sound decision-making in risk assessment. QSAR predictive uncertainty is affected by a variety of factors related to the quality of the training set data, the adopted statistical models, and the distance between the query chemical and the training set. We developed a method to quantify uncertainty associated with individual linear QSAR predictions that integrates both model and experimental error uncertainty and that defines an applicability domain based on the density of training set data. Our method is based on chemical spaces defined by latent variables identified by Partial Least Squares (PLS) regressions. The method provides a kernel regression estimate of the activity of interest as well as a measure of predictive uncertainty based on a mathematical estimation of the domain of applicability and on local propagation of uncertainty associated with training set data.

A linear model to predict chronic effects of chemicals on Daphnia magna.

Chronic toxicity data for Daphnia magna are information requirements in the context of regulations on chemical safety. This paper proposes a linear model for the prediction of chemically-induced effects on the reproductive output of D. magna. This model is based on data retrieved from the Japanese Ministry of Environment database and it predicts chronic effects as a function of acute toxicity data. The proposed model proved to be able to predict chronic toxicities for chemicals not used in the training set. Our results suggest that experiments involving chronic exposure to chemicals could be reduced thanks to the proposed model.

Exploring an ecotoxicity database with the OECD (Q)SAR Toolbox and DRAGON descriptors in order to prioritise testing on algae, daphnids, and fish.

The European regulation on chemicals (REACh) places emphasis on reduction of systematic toxicity testing, thus fostering development of alternative methods. Consequently, we analysed acute toxicity data gathered by the Japanese Ministry of Environment for three species belonging to three different trophic levels (i.e., Pseudokirchneriella subcapitata 72-hour EC50, Daphnia magna 48-hour EC50 and Oryzias latipes 96-hour LC50). This paper investigates the relationships between the chemical structure and both the toxicity of the chemicals and the cross-species differences in sensitivity. The physicochemical properties of the chemicals were represented by the categories they belonged to in several widely-used categorisation schemes implemented by the freely available OECD (Q)SAR Toolbox and by quantitative molecular descriptors using DRAGON software. The outputs of these software products were analysed and compared in terms of quality of prediction and biological interpretation. Amongst the categorisations implemented by the OECD Toolbox, those focussing on bioaccumulation or biotransformation appeared to be the most interesting in terms of environmental prediction on a whole set of chemicals, in particular as the predicted biotransformation half-life is strongly dependent on hydrophobicity. In predicting toxicity towards each species, simple linear regression on logP performed better than PLS regression of toxicity on a very large set of molecular descriptors. However, the predictions based on the interspecies correlations performed better than the QSAR predictions. The results in terms of cross-species comparisons encourage the use of test strategies focussing on reducing the number of tests on fish.

What can be learnt from an ecotoxicity database in the framework of the REACh regulation?

Since REACh applies in all of EU, special emphasis has been put on the reduction of systematic ecotoxicity testing. In this context, it is important to extract a maximum of information from existing ecotoxicity databases in order to propose alternative methods aimed at replacing and reducing experimental testing. Consequently, we analyzed a database of new chemicals registered in France and Europe during the last twenty years reporting aquatic ecotoxicity data with respect to three trophic levels (i.e., Algae EC50 72 h, Daphnia EC50 48 h and Fish LC50 96 h). In order to ensure the relevance of the comparison between these three experimental tests, we performed a stringent data selection based on the pertinence and quality of available ecotoxicological information. At the end of this selection, less than 5% of the initial number of chemicals was retained for subsequent analysis. Such an analysis showed that fish was the least sensitive trophic level, whereas Daphnia had the highest sensitivity. Moreover, thanks to an analysis of the relative sensitivity of trophic levels, it was possible to establish that respective correction factors of 50 and 10 would be necessary if only one or two test values were available. From a physicochemical point of view, it was possible to characterize two significant correlations relating the sensitivity of the aforementioned trophic levels with the chemical structure of the retained substances. This analysis showed that algae displayed a higher sensitivity towards chemicals containing acid fragments whereas fish presented a higher sensitivity towards chemicals containing aromatic ether fragments. Overall, our work suggests that statistical analysis of historical data combined with data yielded by the REACh regulation should permit the derivation of robust safety factors, testing strategies and mathematical models. These alternative methods, in turn, could allow a replacement and reduction of ecotoxicological testing.

An evaluation of the predictive ability of the QSAR software packages, DEREK, HAZARDEXPERT and TOPKAT, to describe chemically-induced skin irritation.

According to the REACH chemicals legislation, formally adopted by the EU in 2006, Quantitative Structure-Activity Relationships (QSARs) can be used as alternatives to animal testing, which itself poses specific ethical and economical concerns. A critical assessment of the performance of the QSAR models is therefore the first step toward the reliable use of such computational techniques. This article reports the performance of the skin irritation module of three commercially-available software packages: DEREK, HAZARDEXPERT and TOPKAT. Their performances were tested on the basis of data published in the literature, for 116 chemicals. The results of this study show that only TOPKAT was able to predict the irritative potential for the majority of chemicals, whereas DEREK and HAZARDEXPERT could correctly identify only a few irritant substances.