Les CDK/cyclines jouent un rôle majeur dans la progression du cycle cellulaire et dans le maintien de la prolifération des cellules cancéreuses, constituant ainsi des biomarqueurs clés et des cibles pharmacologiques attractives. Plus particulièrement, l’activité de CDK4/cycline D, kinase responsable de la progression de la phase G1 et de la transition G1/S, est dérégulée dans de nombreux cancers dont le mélanome. Cette hyperactivation est associée à des mutations, à l’amplification ou à la surexpression de CDK4, cycline D, p16INK4a ou encore pRb.Comme aucune approche sensible et directe n’existe pour évaluer l’activité de CDK4/cycline D dans des conditions physiologiques et pathologiques, le premier objectif de ma thèse a consisté à développer un biosenseur fluorescent permettant d’étudier cette kinase in vitro et in cellulo. Une fois caractérisé et validé in vitro, le biosenseur a été appliqué à la détection d’altérations de CDK4/cycline D dans des biopsies de peau humaine et de xénogreffes de mélanome dans des essais fluorescents d’activité kinase, ainsi que dans des cellules cancéreuses vivantes par microscopie de fluorescence et vidéo microscopie.Par ailleurs, peu d’inhibiteurs sont actuellement disponibles pour inhiber CDK4/cycline D et la plupart d’entre eux ciblent la poche de fixation de l’ATP. C’est pourquoi le second objectif de ma thèse a consisté à identifier des inhibiteurs non compétitifs de l’ATP, soit par élaboration rationnelle de peptides, soit par criblage de petites molécules. A cette fin, deux biosenseurs fluorescents ont été développés qui permettent d’identifier respectivement des composés ciblant l’interface entre CDK4 et cycline D ou des inhibiteurs allostériques capables de perturber la dynamique conformationnelle de CDK4. Des essais de criblage par fluorescence réalisés avec ces biosenseurs ont conduit à l’identification de touches qui ont été validées et caractérisées in vitro et dans des essais de prolifération cellulaire, et qui constituent des candidats prometteurs pour une chimiothérapie sélective du mélanome. / CDK/cyclins play a central role in coordinating cell cycle progression, and in sustaining proliferation of cancer cells, thereby constituting established cancer biomarkers and attractive pharmacological targets. In particular, CDK4/cyclin D, which is responsible for coordinating cell cycle progression through G1 into S phase, is a relevant target in several cancers including melanoma, associated with mutation of CDK4, cyclin D, p16INK4a and pRb.As there are no sensitive and direct approaches to probe CDK4/cyclin D activity in physiological and pathological conditions, the first goal of my thesis has consisted in engineering a fluorescent biosensor to probe this kinase in vitro and in cellulo. Once characterized and validated in vitro, the biosensor was applied to detect CDK4/cyclin D alterations in biopsies from human skin and melanoma xenografts in fluorescence-based activity assays, and in living cancer cells by fluorescence microscopy and timelapse imaging.Moreover, only few inhibitors are currently available to target CDK4/cyclin D and most of them bind the ATP pocket. As such, the second major goal of my thesis project has consisted in identifying non-ATP competitive inhibitors, either through rational design of peptides or by screening small molecule libraries. To this aim, two fluorescent biosensors were engineered which discriminate compounds that target the interface between CDK4 and cyclin D, or that perturb the conformational dynamics of CDK4, respectively, from ATP-pocket binding compounds. Fluorescence-based screening assays performed with these biosensors lead to identification of hits, which were validated and characterized in vitro and in cell proliferation assays, and which constitute promising candidates for selective chemotherapy in melanoma.

<p>  </p> <p>Thrombosis is the pathological formation of a blood clot in the body that blocks blood circulation, leading to high morbidity and mortality rates. Thrombolytic drugs that offer rapid clot dissolution are promising treatments yet current drugs are often associated with limited efficacy and high bleeding risks. While numerous animal thrombosis models have been developed for drug screening, the translation of therapeutic agents into and through clinical trials remains limited. This is largely due to animal models’ poor reproducibility and distinctive physiology to that of humans. <em>In-vitro</em> flow models that utilize both human blood components and physiologically relevant flow conditions can provide for a more representative testing environment to screen thrombolytic drugs. Developing better <em>in-vitro</em> models may not eliminate the need for preclinical animal testing but can help exclude inefficient agents earlier in the drug development pipeline to expedite the drug evaluation process. Existing <em>in-vitro</em> thrombolysis flow models are not ideal as they either adopt over-simplified clot substrates or utilize small-length-scale geometries that insufficiently mimic native hemodynamics. Thus, we propose to first develop a static fluorescently labeled clot lysis assay for an initial high throughput screening of thrombolytic drugs, and ultimately engineer a highly reproducible, physiological scale, flowing clot lysis model for more human relevant drug efficacy evaluation. Developing the static clot lysis assay not only helps to understand the mechanism of how diversified clotting conditions affect clot properties but also offer a chance to well-characterize fluorescence conjugations to fibrins. The ultimate flow model combines an <em>in-vivo</em>-like fluorescence incorporated synthetic clot (FISC) and a human-relevant flow system. Guided by results from static clotting experiments diversified FISCs are fluorescently optimized and fabricated dynamically using a Chandler loop setup at various conditions. The flow system is a tubing-based structure that comprises of a peristaltic pump, and a well-controlled flow chamber to provide for physiological shear and pulsatile levels. Therefore, the proposed synthetic clot model is a versatile platform that can mimic a variety of thrombosis conditions and offer representative drug testing and dosing results across numerous thrombolytic agents.</p>

No description available.