Journal of chromatography. A 2015-10-30

Tadpole toxicity prediction using chromatographic systems.

Alejandro Fernández-Pumarega, Susana Amézqueta, Elisabet Fuguet, Martí Rosés

Index: J. Chromatogr. A. 1418 , 167-76, (2015)

Full Text: HTML

Abstract

Toxicity has been emulated in tadpole species through chromatographic systems. The parameter studied to evaluate the non-specific toxicity of a compound is the narcosis concentration (Cnar), which is defined as the concentration needed for the immobilization of the organism. Because experimental investigation with animals is lengthy, costly, technically difficult, and ethically questionable, there is a great interest in developing surrogate physicochemical systems able to emulate biological systems to obtain the same information in a faster, more economic, and easier manner. In order to see which chromatographic systems would be able to emulate tadpole narcosis, both, tadpole narcosis data and data in several chromatographic and electrophoretic systems, were fitted to a linear solvation energy relationship (LSER) model. Thus, by comparison of the models it was possible to see which of the chromatographic systems were more similar to the biological one. The physicochemical systems that best emulate tadpole narcosis were an HPLC system based on an immobilized artificial membrane (IAM) column, and two micellar electrokinetic chromatography (MEKC) systems based on sodium taurocholate (STC) and a mixture of sodium dodecylsulphate (SDS) and Brij 35 as surfactants. A system based on a RP18 HPLC column also was selected for comparison because it is a common column in most analytical laboratories. To establish the models, a set of compounds with known Cnar values were analyzed in the chromatographic, and electrophoretic selected systems and, then, the retention factor (k) was correlated to the concentration of narcosis. Statistics showed that the system based on STC micelles was the best to emulate toxicity in tadpoles. The robustness and predictive ability of the developed models were validated. Copyright © 2015 Elsevier B.V. All rights reserved.