Read-across in REACH and the read-across assessment framework (RAAF) (#127)
A. Kovari1, N. Andersson1, D. Bell1, G. Cartlidge1, N. Fedtke1, A. Kojo1, P. Papadaki1, E. Stilgenbauer1, F. Temussi1
1 ECHA, Helsinki, Finland
In principle, read-across is a technique for predicting a property or endpoint information for one substance, by using data from the same property/endpoint from (an)other substance(s).
The interpretation and use of grouping and read-across approach may differ depending on the regulatory context in which it is used.
Under REACH, a data gap can be filled by read-across within a group of structurally similar substances under the conditions set out in the requirements of Annex XI, section 1.5 of the REACH Regulation. It is to be noted that structural similarity is a pre-requisite for any grouping and read-across approach under REACH.
Based on the most frequently applied read-across approaches observed in REACH registration dossiers, ECHA developed the RAAF as an internal tool to enhance the assessment of such cases.
The RAAF highlights the critical scientific elements to be addressed when assessing predictions based on read-across in the context of the REACH. The assessment of read-across approaches following the RAAF is conducted through the use of scenarios, assessment elements (AE) and assessment options (AO).
The appropriate scenario is to be selected based on three key principles: (i) type of approach applied (analogue vs category approach), (ii) the type of read-across hypothesis ((bio)transformation to common compound(s) vs different compounds having the same type of effect(s)) and (iii) the type of prediction (similar vs difference in potency).
Each scenario comprises a series of dedicated AE, representing crucial scientific aspects which need to be addressed. To conclude on the adequacy and scientific robustness of the information provided in the dossier for each AE under consideration one of a predefined set of assessment options needs to be selected.
The final outcome of the read-across assessment is established on the basis of the conclusions derived for all of the AEs.
Keywords: REACH, RAAF, scenario, assessment element, (bio)transformation
Scientific considerations for assessing read-across between multi-constituent or UVCB substances (#80)
E. Stilgenbauer1, N. Andersson1, D. Bell1, G. Cartlidge1, N. Fedtke1, A. Kojo1, A. Kovari1, P. Papadaki1, F. Temussi1
1 European Chemicals Agency, Helsinki, Finland
According to Annex XI, Section 1.5 of the REACH Regulation, structural similarity between substances is a pre-requisite for adaptations of information requirements based on grouping of substances and read-across. A substance registered under REACH may comprise a set of chemical structures corresponding to different constituents, impurities, i.e. not only one relevant chemical structure is present. By definition, the composition of multi-constituent substances and “substances of unknown or variable composition, complex reaction products or biological materials” (UVCBs) includes more than one and up to many relevant chemical structures.
Whilst the Read-Across Assessment Framework (RAAF) was developed for mono- constituent substances, its principles may also be applied when assessing read-across approaches involving multi-constituent substances and UVCBs. However, the assessment of such adaptations requires additional considerations on some scientific aspects in order to ensure that all the constituents of the substances are taken into account in the prediction.
As a consequence of the increasing complexity of the composition of the substances, a comprehensive characterisation of the identity and composition of the substances is essential for a sound read-across approach. Detailed information on the composition of source substance and the test material used to perform the source studies is also needed to establish their relation to the target substance in terms of grouping and predictions. The allowed quantitative variations in the concentrations of the different constituents, the impact of combined exposure to several constituents also constitute relevant considerations.
As for read-across approaches on mono-constituents, supporting information is required to increase the robustness of the read-across hypothesis. Bridging studies, i.e. comparable studies on the source and target substance, may constitute precious supporting information in demonstrating that substances have similar properties for a particular endpoint.
Keywords: REACH, read-across, UVCB, multi-constituent substance
Scientific Issues That Need to be Addressed for a Successful Read-Across Prediction (#297)
1 Liverpool John Moores University, Liverpool, United Kingdom
The acceptance of a prediction or assessment of toxicological hazard using read-across is dependent on a number of definable issues. The development of case studies for read-across, as well as the retrospective analysis of regulatory dossiers, has informed the definition of issues to be addressed. It is clear that current read-across comprises many different approaches ranging from a simplistic consideration of structural based analogues, to larger mechanistically derived groups and automated approaches to similarity assessment. Whilst appearing to be trivial, the initial challenge for a successful read-across is to utilise the correct approach, or combination of approaches, on the basis of the endpoint required, regulatory context and data available to read-across from. At the outset, to achieve successful read-across the chemical structure must be defined and correctly described. Once chemistry is defined the correct similarity metric must be applied and appropriate analogues selected, this relies on expert judgement as well as the practicalities of the available data. Whilst a large group of analogues may appear desirable, this will add to the complexity of the justification of read-across. To assist with read-across justification, the definition of the uncertainties is required followed by their assessment and relationship to the confidence that can be placed on the read-across. Once defined, further information – ideally from New Approach Methodologies (NAMs) – can be sought in an attempt to reduce the uncertainties. The addition of further information may form a Weight of Evidence (WoE) that gives confidence to the prediction. Justification and interpretation of mechanistic information, e.g. using NAMs from Tox21 / ToxCast data, has the potential to assist in the plausibility that may be associated with read-across hypothesis. Understanding the issues required for successful read-across prediction, and how they can be supported scientifically, will improve the overall quality of the assessment and the likelihood of its acceptance for the intended purpose.
Keywords: read-across, justificationnew appr, new approach methodology, weight of evidence
Quantitative assessment of mitochondrial toxicity and downstream cellular perturbations in Adverse Outcome Pathways. (#799)
B. van de Water1
1 Leiden University, Leiden, Netherlands
Mitochondrial function is essential for cell survival and growth. Many agrochemicals target different complexes of the mitochondrial respiratory chain. For one of these inhibitors, the complex I inhibitor rotenone, epidemiological studies indicate a significant association between rotenone exposure and Parkinson's disease. In the EU-ToxRisk project we made advantage of the recent AOP “Inhibition of the mitochondrial complex I of nigra striatal neurons leads to parkinsonian motor deficit” (AOP#3). Here we report on the systematic evaluation of the effect of 22 different complex I, II and III inhibitors on the activation of the molecular initiation event and various key events of this AOP. We used different test systems (liver, kidney, neuronal cells) and assessed both the mitochondrial function using cellular oxygen consumption as well as high content imaging. Moreover, we assessed the downstream activation of both oxidative stress and unfolded protein response using fluorescent protein reporter systems. Specific perturbations of neuronal dysfunction were assessed using phenotypic assays. We observed that in particular complex I and III inhibitors are highly effective in perturbing mitochondrial activity in association with perturbation of the cellular proteostasis as evaluated e.g. by induction of the unfolded protein response; agrochemicals that inhibit complex II were not effective in causing mitochondrial dysfunction nor cellular stress responses. While all complex I inhibitors effectively inhibited mitochondrial function in neuronal cells, only rotenone and its close analogue deguelin potently interferred with neurite outgrowth in different neuronal cell test systems. The data indicate that all human cell-based test systems demonstrate similar potencies towards the different mitochondrial inhibitors. Moreover, the data provide an understanding of the applicability domain as well as the challenges of AOPs in assessing chemical safety.
Keywords: systems toxicology, high content imaging, quantitative AOPs, liver toxicity, computational biology
Key issues for successful read-across - a user's guide (#236)
N. S. Ball1
1 The Dow Chemical Company, Toxicology, Environment, Research and Consulting, Horgen, Zürich, Switzerland
In the early days of REACH, the use of read-across as a data gap filling technique when preparing a registration dossier for a chemical was often seen as a quick, cost effective way of addressing data gaps while avoiding the need for additional studies; some of which may involve the use of animals. Pre-existing category justifications developed for programs such as the OECD or EPA High Production Volume chemical assessment programs were typically used without modification or supplementation with additional information. In some cases, justifications for read-across were limited to a few sentences in the summary for each endpoint. Over the past 8-9 years it has become clear that when using read-across in a registration dossier there are high expectations from evaluators regarding the quality and validity of the scientific justification, the quality of studies being read-across and the value/relevance of additional supporting evidence. Consequently many pre-existing category justifications and analogue approaches have failed to meet the expectations leading to refinements or requests for additional toxicity studies. Over the same time period both the regulatory and industry scientists have also been on a learning curve, leading in many cases to a shared understanding of what is an adequate justification, what uncertainty is acceptable, and what supporting information is of most utility. This presentation will illustrate this learning curve with some examples, identifying some of the key issues to address in order to have a robust read-across justification and identify when you may need more supporting data.
Keywords: read-across, REACH