Extreme climatic events and water eutrophication are intensifying the occurrence of cyanobacterial blooms. Microcystis is a ubiquitous and highly proliferating genus in freshwater ecosystems that impacts human, animal, and environmental health. Microcystis senescence releases a wide range of potentially toxic metabolites. Microcystins, a structural family (329 variants) mainly produced by Microcystis, have been extensively studied for their toxicity. Other cyanopeptide families are attracting scientific interest, with some variants exhibiting toxic activities in mammals and vertebrate aquatic organisms.
The ToxCell project aims to characterize the potential cellular toxicity of these emerging contaminants by considering the families independently and in the form of complex mixtures, which are more representative of natural environmental conditions. Toxicity assessments of complex cyanopeptide mixtures require characterizing the underlying synergistic, antagonistic, and additive effects that have already been documented. 
This characterization of cellular toxicity employs an original combined Effect-Directed Analysis (EDA) and Weight-of-Evidence (WoE) approach. EDA consists of fractionating complex cyanopeptide mixtures using in vitro cellular bioassays (3R approach) to identify toxic cyanopeptides. Three freshwater fish cell lines were selected: RTL-W1 (Rainbow Trout Liver-Waterloo 1), rainbow trout liver epithelial cells; CLC (Carp Leucocyte Culture), common carp leucocytes; CCB (Common Carp Brain), common carp brain fibroblasts. The cellular toxicity of cyanopeptides is evaluated through an initial cytotoxicity screening step, followed by dose-response analysis for high cytotoxicity, or analysis of sublethal effects (cell cycle, energy metabolism, ROS, phagocytosis, cytochrome activity) when cytotoxicity is lower. This multi-line and multi-response strategy is an appropriate approach for identifying potential biological targets of emerging cyanopeptides whose mechanism of toxicity remains undetermined. An integrative Weight-of-Evidence (WoE) approach facilitates the interpretation of this complex dataset. This approach integrates weighted lines of evidence to obtain overall toxicity indices, divided into different toxicity classes, and bounded between a negative control (0% effect) and a positive control (100% effect), both of which have been validated in cellular bioassays. This procedure enables a global interpretation of toxicity associated with a toxicity level, as well as the selection of toxic compositions in the EDA fractionation procedure.