Evaluation of in silico Models to Predict the Toxicity of Binary Heavy Metal Mixtures on Freshwater Phytoplankton

Abstract

Abstract In aquatic ecotoxicology, predicting the effects of different chemical mixtures on ecosystems is a priority. This aspect acquires special significance considering the diversity of pollutants in general, and heavy metals (HMs) in particular, coexisting in the aquatic environment and interacting with each other, generating different types of toxicological response depending on whether the interaction between them induces the development of antagonistic, additive or synergistic effects. Because the evaluation of HMs mixtures is complex due to the expensive and complex nature of the analyses, this work aimed to evaluate the predictive potential exhibited by the in silico "Toxic Units" (TUpred) and Combination Index (CI) models in cell growth inhibition assays of freshwater green algae Scenedesmus armatus exposed to binary HMs combinations, by comparison with the experimental results obtained (TUexp). For this purpose, cells were in vivo exposed to binary mixtures of cadmium (Cd+ 2) and the selected heavy metals silver (Ag+ 1), copper (Cu+ 2), mercury (Hg+ 2), zinc (Zn+ 2), and chromium (Cr+ 6) for 72 hours. Our results showed an inverse Fa-dependent relationship between TUexp and TUpred, and overall, a high variability in the results obtained for all the binary combinations analyzed. The CI predictive model showed a high correlation with in vivo assays (TUexp) when the affected fraction was high (Fa = 0.5), decreasing as Fa was lower until it did not correlate at the lowest Fa assayed (Fa = 0.1). These results demonstrate the suitability of using the CI model over the predictive TU model, and only at high HM concentrations.