Uranium is a natural radioactive trace element for which elevated concentrations can be found
in the vicinity of some nuclear fuel cycle facilities or of intensive farming areas. Due its co-occurrence with different trace metals, such as cadmium, in geological ores, U is generally found associated with other contaminants in the environment. The study of their combined effects on ecosystems is of interest to better characterize such multi-metallic polluted sites. The mixture toxicity assessment and the identification of synergistic or antagonistic interaction are generally performed on the basis of additive reference models integrated to descriptive and purely statistical approaches with no real biological basis. Recently, mechanistic models were proposed to better account for the dynamics of biological and toxicological processes. However, such models have only been put into practice in a few number of mixture toxicity case studies. The aim of this PhD project was to assess the chronic U/Cd combined toxicity on the life history traits of the nematode Caenorhabditis elegans using both a descriptive (MixTox) and a mechanistic (DEBtox) approach. To do so, nematodes were exposed during eleven days to different U and Cd concentrations, alone or in mixture. A strong antagonistic interaction between U and Cd was identified for length increase and brood size endpoints on the basis of both approaches. From the study of the U and Cd media-to-food transfer and of the U and Cd bioaccumulation by C. elegans, we showed that the co-presence of U reduced the available Cd fraction for nematodes. To identify a possible interaction at organism level, occurring during the toxicokinetic or toxicodynamic steps, data were re-analyzed on the basis of U/Cd concentrations in food, assumed to be more closely related to available concentrations for C. elegans. Overall additive effects, without interaction, were identified between U and Cd, even if slightly contrasted conclusions were obtained on the basis of the descriptive and mechanistic approaches. The present study underlines the complexity of studying mixture toxicity and identifying chemical interactions. Despite some application problems, the mechanistic approach DEBtox is particularly promising to describe the toxicity of chemical mixtures over time and to test hypothetical interaction mechanisms. In the future, the improvement of tools to analyze the combined toxicity of contaminants would allow to better address the issue of mixtures in ecotoxicological risk assessment processes.