Karen TILMANT, jeudi 29 novembre à 10h30, Au Grenier, Laboratoire LIEBE, Metz Evaluation de nouvelles technologies pour le criblage toxicologique de molécules chimiques Thèse de doctorat : Ecotoxicologie, Biodiversité, Ecosystèmes: Université de Lorraine - Metz, Laboratoire LIEBE-CNRS UMR 7146, ED 410 RP2E Direction de thèse

Carole Cossu-Leguille, directrice de thèse, UL, LIEBE UMR CNRS 7146 Franck Atienzar, co-directeur de thèse, UCB pharma, Belgique Membres du jury Rapporteurs :

Roger Rahmani, directeur de recherche INRA, UMR 1331, Sophia Antipolis Lucien Hoffmann, directeur scientifique du Lab. EVA, Centre de Recherche publique Gabriel Lipmann, Luxembourg Examinateurs

Pascal Poupin, maître de conférences, LIEBE CNRS UMR 7146, Université de Lorraine (Metz) Carole Cossu-Leguille, professeur, LIEBE CNRS UMR 7146, Université de Lorraine (Metz) Franck Atienzar, directeur associé, UCB pharma, Braine l’Alleud, Belgique Résumé In the pharmaceutical industry, there is a clear need to improve human risk assessment by using early and more diverse toxicity assays. Although the tendency is still to rely on whole animal experiments, mechanistic assays using in vitro approaches are increasingly being used early in the drug discovery phase. In vitro investigations can be used to guide medicinal chemistry prior to the conduct of in vivo toxicology studies designed to support initial human trials. Indeed, in vitro profiling of compounds promises to provide information on molecular mechanisms of toxicity and may allow the creation of algorithms for predictive in vivo toxicology. It is thus of prime importance to use innovative approaches in pharmaceutical industry to gain predictive information in order to reduce the attrition rate as well as the development costs of candidate compounds. Under the European REACH legislation, toxicity has to be assessed for all chemicals that are sold in the EU in annuals quantities of more than one ton. Thus, beyond the pharmaceutical industry, the necessity of predictive information is of importance for a number of other sectors, including the chemical industry, environmental pollution and food sector. The main objective of this research performed in a research laboratory in the pharmaceutical industry was to evaluate new technologies for early toxicity screening of chemicals. Indeed a number of new technologies are becoming commercially available but still need thorough validation studies. The thesis work mostly relies on microarray, cellular imaging and Real Time Analysis (RTCA) to assess toxicity effects from the gene expression level to the cellular level. RTCA is a promising impedance-based labelfree technology, that performs non-invasive kinetic measurements of the cell status. This technology brings information on cell proliferation, arrest, and death as well as on cell signaling pathways. We explored the optimal conditions for the use of the RTCA for toxicity assessment in a number of cellular models. Cells from different organs and species were used to cover the diverse organspecific toxicities. On the other hand, two studies to set up a medium throughput screen were performed, one for drug-induced phospholipidosis (PLD) and one for micronucleus assessment. These assays

were developed for screening purposes in the early stage of drug development, when still many compounds have to be screened. For PLD screening, our goal was to select the optimal method between a genomic method and a fluorescent imaging assay. The genomic approach, based on real-time PCR analysis of biomarkers for PLD, had been previously validated in our laboratory. Due to the workload and cost, associated to PCR, another method based on fluorescent automated imaging was validated and compared to the genomic method. For genotoxicity assessment, the standard micronucleus assay consists of visual evaluation by microscopy. More recently, fast processing automated systems such as cellular imaging and flow cytometry were used to measure micronuclei in an automated way. We optimized and implemented an in vitro and in vivo medium throughput method for micronuclei assessment.