A prioritization strategy for functional alternatives to bisphenol A in food contact materials.

The use of bisphenol A (BPA), a substance of very high concern, is proposed to be banned in food contact materials (FCMs) in the European Union. To prevent regrettable substitution of BPA by alternatives with similar or unknown hazardous properties, it is of importance to gain the relevant toxicological information on potential BPA alternative substances and monitor them adequately. We created an inventory of over 300 substances mentioned as potential BPA alternatives in regulatory reports and scientific literature. This study presents a prioritization strategy to identify substances that may be used as an alternative to BPA in FCMs. We prioritized 20 potential BPA alternatives of which 10 are less familiar. We subsequently reviewed the available information on the 10 prioritized less familiar substances regarding hazard profiles and migration potential obtained from scientific literature and in silico screening tools to identify a possible risk of the substances. Major data gaps regarding the hazard profiles of the prioritized substances exist, although the scarce available data give some indications on the possible hazard for some of the substances (like bisphenol TMC, 4,4-dihydroxybenzophenone, and tetrachlorobisphenol A). In addition, very little is known about the actual use and exposure to these substances. More toxicological research and monitoring of these substances in FCMs are, therefore, required to avoid regrettable substitution of BPA in FCM.

New Approach Methodologies (NAMs) for ad hoc human health risk assessment of food and non-food products - Proceedings of a workshop.

RIVM convened a workshop on the use of New Approach Methodologies (NAMs) for the ad hoc human health risk assessment of food and non-food products. Central to the workshop were two case studies of marketed products with a potential health concern: the botanical Tabernanthe iboga which is used to facilitate mental or spiritual insight or to (illegally) treat drug addiction and is associated with cardiotoxicity, and dermal creams containing female sex hormones, intended for use by perimenopausal women to reduce menopause symptoms without medical supervision. The workshop participants recognized that data from NAM approaches added valuable information for the ad hoc risk assessment of these products, although the available approaches were inadequate to derive health-based guidance values. Recommendations were provided on how to further enhance and implement NAM approaches in regulatory risk assessment, specifying both scientific and technical aspects as well as stakeholder engagement aspects.

The Bigger Fish: A Comparison of Meta-Learning QSAR Models on Low-Resourced Aquatic Toxicity Regression Tasks.

Toxicological information as needed for risk assessments of chemical compounds is often sparse. Unfortunately, gathering new toxicological information experimentally often involves animal testing. Simulated alternatives, e.g., quantitative structure-activity relationship (QSAR) models, are preferred to infer the toxicity of new compounds. Aquatic toxicity data collections consist of many related tasks─each predicting the toxicity of new compounds on a given species. Since many of these tasks are inherently low-resource, i.e., involve few associated compounds, this is challenging. Meta-learning is a subfield of artificial intelligence that can lead to more accurate models by enabling the utilization of information across tasks. In our work, we benchmark various state-of-the-art meta-learning techniques for building QSAR models, focusing on knowledge sharing between species. Specifically, we employ and compare transformational machine learning, model-agnostic meta-learning, fine-tuning, and multi-task models. Our experiments show that established knowledge-sharing techniques outperform single-task approaches. We recommend the use of multi-task random forest models for aquatic toxicity modeling, which matched or exceeded the performance of other approaches and robustly produced good results in the low-resource settings we studied. This model functions on a species level, predicting toxicity for multiple species across various phyla, with flexible exposure duration and on a large chemical applicability domain.

ZZS similarity tool: The online tool for similarity screening to identify chemicals of potential concern.

Screening and prioritization of chemicals is essential to ensure that available evaluation capacity is invested in those substances that are of highest concern. We, therefore, recently developed structural similarity models that evaluate the structural similarity of substances with unknown properties to known Substances of Very High Concern (SVHC), which could be an indication of comparable effects. In the current study the performance of these models is improved by (1) separating known SVHCs in more specific subgroups, (2) (re-)optimizing similarity models for the various SVHC-subgroups, and (3) improving interpretability of the predicted outcomes by providing a confidence score. The improvements are directly incorporated in a freely accessible web-based tool, named the ZZS similarity tool: https://rvszoeksysteem.rivm.nl/ZzsSimilarityTool. Accordingly, this tool can be used by risk assessors, academia and industrial partners to screen and prioritize chemicals for further action and evaluation within varying frameworks, and could support the identification of tomorrow's substances of concern.

Relevance of In Vitro Transcriptomics for In Vivo Mode of Action Assessment.

Recently, we reported an in vitro toxicogenomics comparison approach to categorize chemical substances according to similarities in their proposed toxicological modes of action. Use of such an approach for regulatory purposes requires, among others, insight into the extent of biological concordance between in vitro and in vivo findings. To that end, we applied the comparison approach to transcriptomics data from the Open TG-GATEs database for 137 substances with diverging modes of action and evaluated the outcomes obtained for rat primary hepatocytes and for rat liver. The results showed that a relatively small number of matches observed in vitro were also observed in vivo, whereas quite a large number of matches between substances were found to be relevant solely in vivo or in vitro. The latter could not be explained by physicochemical properties, leading to insufficient bioavailability or poor water solubility. Nevertheless, pathway analyses indicated that for relevant matches the mechanisms perturbed in vitro are consistent with those perturbed in vivo. These findings support the utility of the comparison approach as tool in mechanism-based risk assessment.

Evaluating chemical similarity as a measure to identify potential substances of very high concern.

Due to the large amount of chemical substances on the market, fast and reproducible screening is essential to prioritize chemicals for further evaluation according to highest concern. We here evaluate the performance of structural similarity models that are developed to identify potential substances of very high concern (SVHC) based on structural similarity to known SVHCs. These models were developed following a systematic analysis of the performance of 112 different similarity measures for varying SVHC-subgroups. The final models consist of the best combinations of fingerprint, similarity coefficient and similarity threshold, and suggested a high predictive performance (≥80%) on an internal dataset consisting of SVHC and non-SVHC substances. However, the application performance on an external dataset was not evaluated. Here, we evaluated the application performance of the developed similarity models with a 'pseudo-external assessment' on a set of substances (n = 60-100 for the varying SVHC-subgroups) that were putatively assessed as SVHC or non-SVHC based upon consensus scoring using expert elicitations (n = 30 experts). Expert scores were direct evaluations based on structural similarity to the most similar SVHCs according to the similarity models, and did not consider an extensive evaluation of available data. The use of expert opinions is particularly suitable as this is exactly the intended purpose of the chemical similarity models: a quick, reproducible and automated screening tool that mimics the expert judgement that is frequently applied in various screening applications. In addition, model predictions were analyzed via qualitative approaches and discussed via specific examples, to identify the model's strengths and limitations. The results indicate a good statistical performance for carcinogenic, mutagenic or reprotoxic (CMR) and endocrine disrupting (ED) substances, whereas a moderate performance was observed for (very) persistent, (very) bioaccumulative and toxic (PBT/vPvB) substances when compared to expert opinions. For the PBT/vPvB model, particularly false positive substances were identified, indicating the necessity of outcome interpretation. The developed similarity models are made available as a freely-accessible online tool. In general, the structural similarity models showed great potential for screening and prioritization purposes. The models proved to be effective in identifying groups of substances of potential concern, and could be used to identify follow-up directions for substances of potential concern.

Model development for evidence-based prioritisation of policy action on emerging chemical and microbial drinking water risks.

While the burden of disease from well-studied drinking water contaminants is declining, risks from emerging chemical and microbial contaminants arise because of social, technological, demographic and climatological developments. At present, emerging chemical and microbial drinking water contaminants are not assessed in a systematic way, but reactively and incidence based. Furthermore, they are assessed separately despite similar pollution sources. As a result, risks might be addressed ineffectively. Integrated risk assessment approaches are thus needed that elucidate the uncertainties in the risk evaluation of emerging drinking water contaminants, while considering risk assessors' values. This study therefore aimed to (1) construct an assessment hierarchy for the integrated evaluation of the potential risks from emerging chemical and microbial contaminants in drinking water and (2) develop a decision support tool, based on the agreed assessment hierarchy, to quantify (uncertain) risk scores. A multi-actor approach was used to construct the assessment hierarchy, involving chemical and microbial risk assessors, drinking water experts and members of responsible authorities. The concept of value-focused thinking was applied to guide the problem-structuring and model-building process. The development of the decision support tool was done using Decisi-o-rama, an open-source Python library. With the developed decision support tool (uncertain) risk scores can be calculated for emerging chemical and microbial drinking water contaminants, which can be used for the evidence-based prioritisation of actions on emerging chemical and microbial drinking water risks. The decision support tool improves existing prioritisation approaches as it combines uncertain indicator levels with a multi-stakeholder approach and integrated the risk assessment of chemical and microbial contaminants. By applying the concept of value-focused thinking, this study addressed difficulties in evidence-based decision-making related to emerging drinking water contaminants. Suggestions to improve the model were made to guide future research in assisting policy makers to effectively protect public health from emerging drinking water risks.

The effects of aliphatic alcohols and related acid metabolites in zebrafish embryos - correlations with rat developmental toxicity and with effects in advanced life stages in fish.

The zebrafish embryo toxicity test (ZFET) is a simple medium-throughput test to inform about (sub)acute lethal effects in embryos. Enhanced analysis through morphological and teratological scoring, and through gene expression analysis, detects developmental effects and the underlying toxicological pathways. Altogether, the ZFET may inform about hazard of chemical exposure for embryonal development in humans, as well as for lethal effects in juvenile and adult fish. In this study, we compared the effects within a series of 12 aliphatic alcohols and related carboxylic acid derivatives (ethanol, acetic acid, 2-methoxyethanol, 2-methoxyacetic acid, 2-butoxyethanol, 2-butoxyacetic acid, 2-hydroxyacetic acid, 2-ethylhexan-1-ol, 2-ethylhexanoic acid, valproic acid, 2-aminoethanol, 2-(2-hydroxyethylamino)ethanol) in ZFET and early life stage (ELS, 28d) exposures, and compared ZFET results with existing results of rat developmental studies and LC50s in adult fish. High correlation scores were observed between compound potencies in ZFET with either ELS, LC50 in fish and developmental toxicity in rats, indicating similar potency ranking among the models. Compounds could be mapped to specific pathways in an adverse outcome pathway (AOP) network through morphological scoring and gene expression analysis in ZFET. Similarity of morphological effects and gene expression profiles in pairs of alcohols with their acid metabolites suggested metabolic activation of the parent alcohols, although with additional, metabolite-independent activity independent for ethanol and 2-ethylhexanol. Overall, phenotypical and gene expression analysis with these compounds indicates that the ZFET can potentially contribute to the AOP for developmental effects in rodents, and to predict toxicity of acute and chronic exposure in advanced life stages in fish.

A Case Study with Triazole Fungicides to Explore Practical Application of Next-Generation Hazard Assessment Methods for Human Health.

The ongoing developments in chemical risk assessment have led to new concepts building on integration of sophisticated nonanimal models for hazard characterization. Here we explore a pragmatic approach for implementing such concepts, using a case study of three triazole fungicides, namely, flusilazole, propiconazole, and cyproconazole. The strategy applied starts with evaluating the overall level of concern by comparing exposure estimates to toxicological potential, followed by a combination of in silico tools and literature-derived high-throughput screening assays and computational elaborations to obtain insight into potential toxicological mechanisms and targets in the organism. Additionally, some targeted in vitro tests were evaluated for their utility to confirm suspected mechanisms of toxicity and to generate points of departure. Toxicological mechanisms instead of the current "end point-by-end point" approach should guide the selection of methods and assays that constitute a toolbox for next-generation risk assessment. Comparison of the obtained in silico and in vitro results with data from traditional in vivo testing revealed that, overall, nonanimal methods for hazard identification can produce adequate qualitative hazard information for risk assessment. Follow-up studies are needed to further refine the proposed approach, including the composition of the toolbox, toxicokinetics models, and models for exposure assessment.

Testing developmental toxicity in a second species: are the differences due to species or replication error?

Developmental toxicity studies for chemical and pharmaceutical safety are primarily performed in rats. Regulatory frameworks may require testing in a second, non-rodent species, for which the rabbit is usually chosen. This study shows that differences in NOAELs or LOAELs (N(L)OAELs) observed between rat and rabbit developmental toxicity studies performed according to OECD guidelines could just as well be caused by study replication errors, and not necessarily by differences in species sensitivity. This conclusion follows from an analysis of a database with rat and rabbit developmental toxicity studies for over 1000 industrial chemicals, pesticides, veterinary drugs and human pharmaceuticals, which included 143 compounds with multiple oral rat studies and 124 compounds with multiple oral rabbit studies. Our analysis confirms earlier findings that, on average over all compounds, rat and rabbit do not differ in sensitivity to developmental effects. There is substantial scatter in the correlation plots comparing rat and rabbit developmental N(L)OAELs, which is easily interpreted as species differences for individual compounds. However, for compounds tested twice in the same species, these N(L)OAELs may differ up to a factor of 25. Thus, potential interspecies differences in developmental N(L)OAEL will be overwhelmed by the reproducibility error, rendering the added value of a second species study questionable. As N(L)OAELs serve as point of departure (POD) for setting health-based guidance values in risk assessment, the large reproducibility error of N(L)OAELs should be taken into account by the introduction of an additional uncertainty factor. It is recommended to aim for reducing the reproducibility error by applying dose-response (BMD) analysis, optimize study designs and harmonize study protocols.

Is current risk assessment of non-genotoxic carcinogens protective?

Non-genotoxic carcinogens (NGTXCs) do not cause direct DNA damage but induce cancer via other mechanisms. In risk assessment of chemicals and pharmaceuticals, carcinogenic risks are determined using carcinogenicity studies in rodents. With the aim to reduce animal testing, REACH legislation states that carcinogenicity studies are only allowed when specific concerns are present; risk assessment of compounds that are potentially carcinogenic by a non-genotoxic mode of action is usually based on subchronic toxicity studies. Health-based guidance values (HBGVs) of NGTXCs may therefore be based on data from carcinogenicity or subchronic toxicity studies depending on the legal framework that applies. HBGVs are usually derived from No-Observed-Adverse-Effect-Levels (NOAELs). Here, we investigate whether current risk assessment of NGTXCs based on NOAELs is protective against cancer. To answer this question, we estimated Benchmark doses (BMDs) for carcinogenicity data of 44 known NGTXCs. These BMDs were compared to the NOAELs derived from the same carcinogenicity studies, as well as to the NOAELs derived from the associated subchronic studies. The results lead to two main conclusions. First, a NOAEL derived from a subchronic study is similar to a NOAEL based on cancer effects from a carcinogenicity study, supporting the current practice in REACH. Second, both the subchronic and cancer NOAELs are, on average, associated with a cancer risk of around 1% in rodents. This implies that for those chemicals that are potentially carcinogenic in humans, current risk assessment of NGTXCs may not be completely protective against cancer. Our results call for a broader discussion within the scientific community, followed by discussions among risk assessors, policy makers, and other stakeholders as to whether or not the potential cancer risk levels that appear to be associated with currently derived HBGVs of NGXTCs are acceptable.

In vitro to in vivo extrapolation of effective dosimetry in developmental toxicity testing: Application of a generic PBK modelling approach.

Incorporation of kinetics to quantitative in vitro to in vivo extrapolations (QIVIVE) is a key step for the realization of a non-animal testing paradigm, in the sphere of regulatory toxicology. The use of Physiologically-Based Kinetic (PBK) modelling for determining systemic doses of chemicals at the target site is accepted to be an indispensable element for such purposes. Nonetheless, PBK models are usually designed for a single or a group of compounds and are considered demanding, with respect to experimental data needed for model parameterization. Alternatively, we evaluate here the use of a more generic approach, i.e. the so-called IndusChemFate model, which is based on incorporated QSAR model parametrization. The model was used to simulate the in vivo kinetics of three diverse classes of developmental toxicants: triazoles, glycol ethers' alkoxyacetic acid metabolites and phthalate primary metabolites. The model required specific input per each class of compounds. These compounds were previously tested in three alternative assays: the whole-embryo culture (WEC), the zebrafish embryo test (ZET), and the mouse embryonic stem cell test (EST). Thereafter, the PBK-simulated blood levels at toxic in vivo doses were compared to the respective in vitro effective concentrations. Comparisons pertaining to relative potency and potency ranking with integration of kinetics were similar to previously obtained comparisons. Additionally, all three in vitro systems produced quite comparable results, and hence, a combination of alternative tests is still preferable for predicting the endpoint of developmental toxicity in vivo. This approach is put forward as biologically more plausible since plasma concentrations, rather than external administered doses, constitute the most direct in vivo dose metric.

An integrative test strategy for cancer hazard identification.

Assessment of genotoxic and carcinogenic potential is considered one of the basic requirements when evaluating possible human health risks associated with exposure to chemicals. Test strategies currently in place focus primarily on identifying genotoxic potential due to the strong association between the accumulation of genetic damage and cancer. Using genotoxicity assays to predict carcinogenic potential has the significant drawback that risks from non-genotoxic carcinogens remain largely undetected unless carcinogenicity studies are performed. Furthermore, test systems already developed to reduce animal use are not easily accepted and implemented by either industries or regulators. This manuscript reviews the test methods for cancer hazard identification that have been adopted by the regulatory authorities, and discusses the most promising alternative methods that have been developed to date. Based on these findings, a generally applicable tiered test strategy is proposed that can be considered capable of detecting both genotoxic as well as non-genotoxic carcinogens and will improve understanding of the underlying mode of action. Finally, strengths and weaknesses of this new integrative test strategy for cancer hazard identification are presented.

Alternatives for skin sensitisation: Hazard identification and potency categorisation: Report from an EPAA/CEFIC LRI/Cosmetics Europe cross sector workshop, ECHA Helsinki, April 23rd and 24th 2015.

In the two years since the last workshop report, the environment surrounding the prediction of skin sensitisation hazards has experienced major change. Validated non-animal tests are now OECD Test Guidelines. Accordingly, the recent cross sector workshop focused on how to use in vitro data for regulatory decision-making. After a review of general approaches and six case studies, there was broad consensus that a simple, transparent stepwise process involving non-animal methods was an opportunity waiting to be seized. There was also strong feeling the approach should not be so rigidly defined that assay variations/additional tests are locked out. Neither should it preclude more complex integrated approaches being used for other purposes, e.g. potency estimation. All agreed the ultimate goal is a high level of protection of human health. Thus, experience in the population will be the final arbiter of whether toxicological predictions are fit for purpose. Central to this is the reflection that none of the existing animal assays is perfect; the non-animal methods should not be expected to be so either, but by integrated use of methods and all other relevant information, including clinical feedback, we have the opportunity to continue to improve toxicology whilst avoiding animal use.

A critical appraisal of the process of regulatory implementation of novel in vivo and in vitro methods for chemical hazard and risk assessment.

Regulatory toxicology urgently needs applicable alternative test systems that reduce animal use, testing time, and cost. European regulation on cosmetic ingredients has already banned animal experimentation for hazard identification, and public awareness drives toward additional restrictions in other regulatory frameworks as well. In addition, scientific progress stimulates a more mechanistic approach of hazard identification. Nevertheless, the implementation of alternative methods is lagging far behind their development. In search for general bottlenecks for the implementation of alternative methods, this manuscript reviews the state of the art as to the development and implementation of 10 diverse test systems in various areas of toxicological hazard assessment. They vary widely in complexity and regulatory acceptance status. The assays are reviewed as to parameters assessed, biological system involved, standardization, interpretation of results, extrapolation to human hazard, position in testing strategies, and current regulatory acceptance status. Given the diversity of alternative methods in many aspects, no common bottlenecks could be identified that hamper implementation of individual alternative assays in general. However, specific issues for the regulatory acceptance and application were identified for each assay. Acceptance of one-in-one replacement of complex in vivo tests by relatively simple in vitro assays is not feasible. Rather, innovative approaches using test batteries are required together with metabolic information and in vitro to in vivo dose extrapolation to convincingly provide the same level of information of current in vivo tests. A mechanistically based alternative approach using the Adverse Outcome Pathway concept could stimulate further (regulatory) acceptance of non-animal tests.

Evaluating the performance of integrated approaches for hazard identification of skin sensitizing chemicals.

The currently available animal-free methods for the detection of skin sensitizing potential of chemicals seem promising. However, no single method is able to comprehensively represent the complexity of the processes involved in skin sensitization. To ensure a mechanistic basis and cover the complexity, multiple methods should be integrated into a testing strategy, in accordance with the adverse outcome pathway that describes all key events in skin sensitization. Although current majority voting testing strategies have proven effective, the performance of individual methods is not taken into account. To that end, we designed a tiered strategy based on complementary characteristics of the included methods, and compared it to a majority voting approach. This tiered testing strategy was able to correctly identify all 41 chemicals tested. In terms of total number of experiments required, the tiered testing strategy requires less experiments compared to the majority voting approach. On the other hand, this tiered strategy is more complex due the number of different alternative methods required, and predicted costs are similar for both strategies. Both the tiered and majority voting strategies provide a mechanistic basis for skin sensitization testing, but the strategy most suitable for regulatory decision-making remains to be determined.

The OSIRIS Weight of Evidence approach: ITS for skin sensitisation.

Within the EU FP6 project OSIRIS approaches to Integrated Testing Strategies (ITSs) were developed, with the aim to facilitate the use of non-test and non-animal testing information in regulatory risk assessment of chemicals. This paper describes an analytical Weight-of-Evidence (WoE) approach to an ITS for the endpoint of skin sensitisation. It specifically addresses the European chemicals legislation REACH, but the concept is readily applicable to ITS and WoE procedures in other regulatory frameworks, and for other toxicological endpoints. Bayesian statistics are applied to estimate the reliability of a conclusion on the sensitisation potential of a chemical, combining evidence from different information sources such as QSAR model predictions, in vitro and in vivo test results. The methodology allows for adaptation of the weight of individual information sources to account for the different levels of reliability of the individual ITS components. The calculated reliability of the WoE conclusion gives an objective, transparent and reproducible measure to decide if the information requirements for data evaluation are satisfied. Furthermore, in case the WoE is not sufficient, it gives the possibility to evaluate a priori if and how it will be possible to fulfil the information requirements with additional tests and/or model predictions.

OSIRIS, a quest for proof of principle for integrated testing strategies of chemicals for four human health endpoints.

Chemical substances policies in Europe are aiming towards chemical safety and at the same time a reduction in animal testing. These goals are alleged to be reachable by mining as many relevant data as possible, evaluate these data with regard to validity, reliability and relevance, and use of these data in so-called Integrated Testing Strategies (ITS). This paper offers an overview of four human health endpoints that were part of the EU-funded OSIRIS project, aiming to develop ITS fit for the EU chemicals legislation REACH. The endpoints considered cover their categorical as well as continuous characteristics: skin sensitisation, repeated dose toxicity, mutagenicity and carcinogenicity. Detailed papers are published elsewhere in this volume. The stepwise ITS approach developed takes advantage of existing information, groups information about similar substances and integrates exposure considerations. The different and possibly contradictory information is weighted and the respective uncertainties taken into account in a weight of evidence (WoE) approach. In case of data gaps, the ITS proposes the most appropriate method to acquire the missing information. Each building block for the ITS, i.e. each in vivo test, in vitro test, (Q)SAR model or human evidence, is evaluated with regard to quality.

The OSIRIS Weight of Evidence approach: ITS mutagenicity and ITS carcinogenicity.

Risk assessment of chemicals usually implies data evaluation of in vivo tests in rodents to conclude on their hazards. The FP7 European project OSIRIS has developed integrated testing strategies (ITS) for relevant toxicological endpoints to avoid unnecessary animal testing and thus to reduce time and costs. This paper describes the implementation of ITS mutagenicity and carcinogenicity in the public OSIRIS webtool. The data requirements of REACH formed the basis for these ITS. The main goal was to implement procedures to reach a conclusion on the adequacy and validity of available data. For the mutagenicity ITS a quantitative Weight of Evidence approach based on Bayesian statistics was developed and implemented. The approach allows an overall quality assessment of all available data for the five types of mutagenicity data requirements: in vitro bacterial mutagenicity, in vitro and in vivo chromosome aberration, in vitro and in vivo mammalian mutagenicity. For the carcinogenicity ITS a tool was developed to evaluate the quality of studies not conforming (entirely) to guidelines. In a tiered approach three quality aspects are assessed: documentation (reliability), study design (adequacy) and scope of examination (validity). The quality assessment is based on expert and data driven quantitative Weight of Evidence.