Abstracts / Toxicology Letters 258S (2016) S54–S61
OSC02-005 Nitro and oxy-PAHs derived from Amazon biomass burning and their mutagenicity using different models S. Batistuzzo 1,∗ , M.D. Galvão 1 , N.D. Alves 2 , P.A. Ferreira 3 , S. Caumo 4 , P.D. Vasconcellos 4 , P. Artaxo 5 , S. Hacon 6 , D.A. Roubicek 7 1
Programa de Pós-Graduac¸ão em Bioquímica, Universidade Federal do Rio Grande do Norte, Natal, RN, Brazil 2 Faculdade de Medicina da Universidade de São Paulo, FMUSP, Brazil 3 Universidade Federal do Pará, Campus Universitário de Altamira, Brazil 4 Instituto de Química, Universidade de São Paulo, São Paulo, SP, Brazil 5 Instituto de Física, Universidade de São Paulo, São Paulo, SP, Brazil 6 Escola Nacional de Saúde Pública da Fundac¸ão Oswaldo Cruz, Rio de Janeiro, RJ, Brazil 7 Departamento de Análises Ambientais, CETESB, São Paulo, Brazil
The biomass burning that occurs in the Amazon region causes adverse effect on environmental and human health. The main pollutant emissions source is the forests burning. So, the present study conducted an organic characterization of polycyclic aromatic hydrocarbons (PAHs) including nitro and oxy-PAHs present in the released particulate matter (PM), as well as a comparative assessment between the moderate and intense biomass burning season focusing on the mutagenicity measured by Salmonella/microsome assay and cytokinesis-block micronucleus (CBMN) in A549 cells. Among the PAHs quantified the Retene and dibenzanthracene showed the highest concentrations during intense biomass burning period. Among the oxy-PAHs, 2-metylanthraquinone and 7,12-benzo[a]anthracenquinone were the most abundant while for the nitro-PAHs, 6-nitrochrysene > 9nitroanthracene > 6-nitrobenzo[a]pyrene. The extractable organic matter (EOM) from intense period was more mutagenic than the EOM from moderate period for both TA98 and YG1041 strains. The data for YG1041 were from 5 to 50% higher than TA98 and the most intense responses were obtained in the absence of metabolic activation. Using non-cytotoxic doses, the EOM was able to generate a dose-response increase of micronuclei frequency (p < 0.01) with concentrations ranging from 50 g/mL to 400 g/mL. Among others CBMN biomarkers, as nucleoplasmic bridges and nuclear buds the extract did not induce significant increase. The presence of pollutants with a mutagenic and carcinogenic effect and the increased in the mutagenic activity indicate that the population is potentially exposed to an increased risk of DNA damage and mutation. Financial support: CAPES and CNPq (Brazil). http://dx.doi.org/10.1016/j.toxlet.2016.06.1304 OSC02-006 Evaluation of read-across argumentation according to the ECHA Read-Across Assessment Framework (RAAF) A. Richarz ∗ , E. Berggren, A. Worth European Commission Joint Research Centre, IHCP, Systems Toxicology Unit & EURL ECVAM, Ispra, Italy Read-across approaches are used increasingly in the evaluation of chemicals, for example to fill data gaps, also in the regulatory context. They are based on grouping similar chemicals together and inferring properties including toxicity from substances with
S59
existing data to target substances without available data belonging to the same category. Building the category, and defining and justifying the similarity of the chemicals considered, is the first crucial step. Even though similarity is mostly anchored on structural similarity in the first place, in particular in the context of REACH, chemical, biological as well as toxicodynamic and toxicokinetic similarity have to be evaluated together, possibly leading to subcategorisation. It is important to assess all uncertainties – both in the similarity and in the read-across arguments – to come to an overall conclusion of the applicability and validity of the read-across prediction. The EU research initiative SEURAT-1 (http://www.seurat-1. eu) performed case studies for different scenarios of chemical similarity, specifically aiming at using new approach methodologies (NAM) to reduce uncertainty (Berggren et al., 2015). In an extended case study, the read-across argumentation and the contribution of the NAM where systematically analysed according to the ECHA Read-Across Assessment Framework (RAAF), the Assessment Elements with and without the NAM contribution were evaluated and general conclusions drawn. This case study illustrates the needs for developing read-across arguments in view of a regulatory use for REACH and opportunities for the use of new methods.
Reference Berggren, E., et al., 2015. Environ. Health Perspect. 123, 1232–1240.
http://dx.doi.org/10.1016/j.toxlet.2016.06.1305 OSC02-007 A reliable workflow for in silico assessment of genetic toxicity and application to pharmaceutical genotoxic impurities C.H. Schwab 1,∗ , J.F. Rathman 2 , J. Marusczyk 1 , A. Mostrag 2 , B. Bienfait 1 , V. Gombar 2 , C. Yang 1 1 2
Molecular Networks GmbH, Erlangen, Germany Altamira LLC, Columbus, OH, USA
Chemicals that have a potential to induce genetic mutations and/or clastogenic damage pose a serious human health risk. The ability to identify such chemicals, especially by in silico approaches, not only helps mitigate the risk, but also makes possible the design of safer chemicals in early discovery. This study presents application of the safety assessment workflow reflected in the US FDA CERES (Chemical Evaluation and Risk Estimation System) to wider use cases, in particular, potential pharmaceutical genotoxic impurities. The comprehensive workflow implemented ChemTunes system allows access to expansive, reliable in vivo and in vitro experimental data, and prediction results based on both expert knowledge and mode-of-action driven QSAR models. The prediction results from local and global models as well as chemotype rules (alerts) are combined to predict the outcome while minimizing the overall uncertainty by applying Dempster–Shafer decision theory. This method has been validated at regulatory agencies. Recent results from an open challenge program sponsored by NIHS Japan on Ames mutagenicity are presented. In this challenge with 3950 newly-tested compounds, the validated models in ChemTunes ranked 3 among 17 participants, with a negative predictive value (NPV) of 92.6% (range for all participants being 89.5–93.4%) and the number of false negative predictions at 184 (range being 145–334). We also demonstrate application of ChemTunes to ReadAcross for the safety assessment of potential genotoxic impurities. http://dx.doi.org/10.1016/j.toxlet.2016.06.1306