Development and validation of a quantitative structure-activity relationship for chronic narcosis to fish.

Vertebrate testing under the European Union's regulation on Registration, Evaluation, Authorisation and Restriction of Chemical substances (REACH) is discouraged, and the use of alternative nontesting approaches such as quantitative structure-activity relationships (QSARs) is encouraged. However, robust QSARs predicting chronic ecotoxicity of organic compounds to fish are not available. The Ecological Structure Activity Relationships (ECOSAR) Class Program is a computerized predictive system that estimates the acute and chronic toxicity of organic compounds for several chemical classes based on their log octanol-water partition coefficient (K(OW)). For those chemical classes for which chronic training data sets are lacking, acute to chronic ratios are used to predict chronic toxicity to aquatic organisms. Although ECOSAR reaches a high score against the Organisation for Economic Co-operation and Development (OECD) principles for QSAR validation, the chronic QSARs in ECOSAR are not fully compliant with OECD criteria in the framework of REACH or CLP (classification, labeling, and packaging) regulation. The objective of the present study was to develop a chronic ecotoxicity QSAR for fish for compounds acting via nonpolar and polar narcosis. These QSARs were built using a database of quality screened toxicity values, considering only chronic exposure durations and relevant end points. After statistical multivariate diagnostic analysis, literature-based, mechanistically relevant descriptors were selected to develop a multivariate regression model. Finally, these QSARs were tested for their acceptance for regulatory purposes and were found to be compliant with the OECD principles for the validation of a QSAR.

Development of a biotic ligand model and a regression model predicting acute copper toxicity to the earthworm Aporrectodea caliginosa.

The purpose of this study was to develop a terrestrial biotic ligand model (BLM) for predicting acute copper toxicity to the earthworm Aporrectodea caliginosa. To overcome the basic problems hampering development of BLMs for terrestrial organisms, an artificial flow-through exposure system was developed consisting of an inert quartz sand matrix and a nutrient solution, of which the composition was univariately modified. A. caliginosa was exposed for 7 days under varying concentrations of copper and the major cations modifying toxicity: H+, Ca2+, Mg2+, and Na+. In addition copper speciation was modulated by means of EDTA or dissolved organic carbon (DOC). An increase in pH or pNa resulted in a linear decrease of 7-days median lethal concentrations. Increasing Ca2+ and Mg2+ activities had inconsistent effects. EDTA addition decreased toxicity when the total copper concentration in the pore water was kept the same. This is attributed to the strong complexation capacity of EDTA and shows that total copper is not the toxic species. DOC was more protective than could be explained by its metal complexing properties. The BLM developed incorporates the effects of H+ and Na+. This BLM was validated with the results of a set of bioassays with artificial pore water in quartz sand and by a set of bioassays in spiked field soils. Prediction error was within a factor of 2, but some predictions were not within the 95% confidence interval. Therefore a more widely applicable regression type model was developed that was able to explain >95% of the (lack of) toxicity observed. To our knowledge this is the first report of the successful development of a terrestrial BLM.