The application of chemical grouping approaches in toxicology——preparing an analogue justification report

<正>Progress in the use of non-testing methods in chemical risk assessment has accelerated in recent years as global chemical legislations have been updated to highlight the principles of animal welfare and the 3 Rs(Replacement,Reduction,Refinement).Chemical agencies have either adopted specific programmes to facilitate grouping of chemicals(eg OECD Existing Chemicals Database,US EPA HPV Challenge Programme) or are in the process of reviewing category or analogue chemical grouping approaches(eg European Chemicals Agency(ECHA) EU REACH dossier submissions for 2010 and 2013).It is clear that systematic grouping approaches(eg read-across in chemical categories or between chemical analogues) are expected to be applied more regularly in filling data gaps and the submission of a high quality robust read-across report will underpin an alternative non-testing strategy. As there are no specific guidelines available from the CRC-MEP in China for submission of a chemical grouping approach report,we propose a template based on experience in submitting read-across reports from technical guidance provided from ECHA and the European Chemicals Ecotoxicology and Toxicology Centre(ECETOC).The template will focus on preparing a quantitative read-across report for a chemical analogue for human health endpoints.The report should consist of 4 main sections:(i) Hypothesis for the analogue approach The hypothesis is the most critical part of the justification,as it is the basis for why the read-across may be performed.The hypothesis should explain how the target and source substance (s) have molecular structure similarities as well as describing any other evidence that may support the read-across e.g.similar mode of action,toxicokinetics.All functional groups need to be identified and any possible effects of any disparate groups should be discussed.The hypothesis should also clearly state which endpoint the read-across approach applies to as each endpoint has a specific set of complexities that should be addressed,(ii) Substance identity and similarity of target and source substance (s) The target and source substance(s) should be identified clearly with names,chemical structures and chemical identifier numbers provided.The composition and purity/impurity profiles should be detailed, particularly in cases where either the target and source substance(s) are multi-constituents or UVCBs(unknown or variable composition,complex reaction products or biological materials),as these substances pose further challenges for read-across based on their complex characteristics.The form or phase of the target and source substance(s) should be clearly stated,as this may affect which hazards the substances may pose.The structural similarities(e.g.functional groups) between the target and source substance(s) should be discussed in detail;if there are any relevant QSAR profilers that may provide an insight into the mechanistic activity of the substances(e.g.DNA or protein binding in the OECD QSAR Toolbox),they can be referred to here as supporting data.It is also beneficial to consider any common breakdown products here though this point can be further elaborated on in the description of the toxicokinetic profiles,(iii) Analogue approach justification The physicochemical and toxicokinetic profiles of the target and source substance(s) can be used to strengthen the read-across hypothesis.If specific toxicokinetic studies are not available,the absorption,distribution,metabolism and excretion may be assessed based on physicochemical properties and toxicity data.Predicted data derived from QSAR profilers may also be acceptable to use as supporting information(e.g.metabolic profilers in the OECD QSAR Toolbox).If reliable toxicokinetic studies are available,they should be submitted as it may provide substantial evidence to justify/ support the read-across hypothesis.The endpoint-specific data (e.g.specific study results and dose descriptor) should then be discussed in detail to demonstrate how the results from the source substance(s) suggest a similar systemic and local toxicity profile to the target substance.The effect of structural differences between the target and source substance(s) should also be discussed as they may contribute to a difference in the toxicity profiles.It is also recommended to consider how the uncertainty in the read-across prediction may influence the overall conclusion of the endpoint.A robust consideration of the impact of uncertainty and steps taken to account for it(e.g.use of appropriate assessment factors in DNEL derivation) may be included in a qualitative uncertainty assessment and will strengthen the read-across hypothesis.The current classification and labelling of the target and source substance(s),as well as the results of the PBT assessment should also be included in the justification.Overall,the analogue approach justification will clearly indicate that the hypothesis is supported by data available and that the read-across approach will not lead to an underestimation of the hazards of the target substance.(iv) Data matrix The data matrix should report the available study results for the target and source substance(s).The target and source substance(s) are the columns in the matrix and the endpoints are the rows in the matrix.Each cell in the matrix should contain the study result type e.g.key experimental study result,reliable predicted data(e.g.QSAR),read-across from supporting substance.Typically the data matrix should report the physicochemical,mammalian toxicity, ecotoxicological and environmental fate data.The source of all data listed in the matrix should also be included, with a list of comprehensive references provided.Any issues relating to data ownership/sharing here should also be considered,as the data is supporting the read-across justification and therefore data access must be considered.In conclusion,we believe the template described is a useful starting point to generate a robust hypothesis-driven comprehensive report that may be used to fill data gaps using a chemical grouping analogue method.