Toxicity of binary metal mixtures to freshwater aquatic organisms: Experimental design, acute and chronic effects

While most metal toxicity assessments in freshwater aquatic organisms focus on single metal exposures, actual environmental exposures occur as metal mixtures. Generally, models used to predict metal mixture toxicity to freshwater organisms assume either dose-additivity or response-additivity. Previous researchers have concluded that responses due to exposure to various metal mixtures can be either more-than-additive or less-than-additive. Most mixture experimental designs suffer from low precision and they lack an objective test that discriminates between additivity and non-additivity. There is a need for more quantitative approaches to characterize mixture effects and further, there is a need for more research into chronic mixture toxicity. The goals of my dissertation were to increase our understanding of the influence of experimental design on the ability to discriminate between additive and non-additive effects of binary metal mixtures and to characterize acute and chronic toxicity of binary metal mixtures to freshwater organisms. Monte Carlo simulations of metal mixture data were used to generate average 48-h LC50 values and average 95% CI from different combination of replicates and organisms/replicate. The simulated results with the smallest 95% CI would provide high increased precision and ability to better differentiate between additivity and non-additivity. A complete factorial experimental design was used to evaluate the acute response of Daphnia magna to zinc-cadmium/cadmium-zinc mixtures and a titration experimental design was used to evaluate the chronic responses of Ceriodaphnia dubia and Pimephales promelas to binary mixtures of copper, cadmium, nickel and zinc. A factorial ANOVA was used to determine mixture effect using a complete factorial experimental design and a slope analysis approach was used to differentiate between additive and non-additive effects using the titration experimental design. Linearized relationships of control-normalized reproduction/growth and dissolved metal concentration were determined for each metal mixture and the slopes were statistically compared. My results showed that metal mixture bioassays with high precision (more replicates) resulted in an average 95% CI that was more than two times smaller than that the standard acute toxicity bioassay. Results using factorial ANOVA indicated a significant (p=0.0024) less-than-additive effect where a low concentration of Zn (111 +/- 24.91 mug/L) had a protective effect on acute Cd toxicity to D. magna. All the chronic exposures to C. dubia showed additive effects at low background exposure concentrations except CdCu, which showed a more-than-additive effect (Cd dose-response 0-93 mug/L and background concentrations of 45, 89, 201 mug/L Cu). Non-additive effects (less-than-additive and more-than-additive) occurred at intermediate and high background exposure concentrations except for NiCd which was less-than additive with the background Cd concentration of 10 mug/L. Additive effects were obtained with P. promelas upon exposure to ZnCd mixtures with background concentrations of 13, 36, and 39 microg/L Cd. As well, additive effects were obtained upon exposure to CdZn mixtures with background concentrations of 110, 412, 680 and 874 microg/L Zn. C. dubia was more sensitive to Cd exposure with high background concentrations of Zn (183, 266, and 392 ug/L) while P. promelas were more sensitive to Zn exposures to a background concentration of 43 ug/L Cd. Overall the results from my dissertation indicate that an increase in the number of replicates in the standard acute 48-h D. magna bioassay would decrease the 95% CI and increase the ability of this test to discriminate between additive and non-additive effects of zinc-cadmium mixtures. Also, quantitative experimental designs are effective in quantitatively characterizing acute and chronic toxicity of binary mixtures of Cu, Cd, Ni and Zn to D. magna, C. dubia and P. promelas. This study will facilitate the development of predictive models for the chronic toxicity of metal mixtures. As well, this study may warrant improvements to the experimental design of mixture studies to allow better discrimination between additive and non-additive effects.