Le cancer est un problème très présent dans nos sociétés modernes. En effet en 2010 il touchait plus de 10 millions de personnes dans le monde. Aujourd'hui cette maladie est la première cause de mortalité dans les pays industrialisés.Malheureusement, les thérapies envisagées restent fréquemment insuffisantes et possèdent de nombreux effets secondaires qui ternissent les bienfaits du traitement. Pour éviter justement cette toxicité auprès des cellules saines, la recherche développe depuis quelques années des traitements ciblés. La plupart des médicaments antimitotiques actuellement sur le marché présentent de forts effets secondaires notamment cardiologiques, hématologiques et neurotoxiques.Nous nous sommes donc intéressés à une autre cible thérapeutique intervenant toujours au niveau de la mitose mais provoquant moins d'effets néfastes et pouvant être surexprimée dans les cellules cancéreuses: la kinésine humaine Eg5.La kinésine humaine Eg5 est indispensable au bon fonctionnement de la mitose. Elle possède un rôle essentiel dans les premières étapes du cycle cellulaire et est requise pour la séparation des centrosomes à chaque pôle de la cellule.La suppression ou l'inhibition d'Eg5 bloque la cellule en pré-métaphase avec un fuseaumonoastral caractéristique formé de deux centrioles non séparés entourés des chromosomes et des microtubules. Le maintien de ce type de fuseau provoque l'activation des checkpoints du cycle cellulaire et provoque l'apoptose.Notre travail consiste en la synthèse de composés susceptibles d'inhiber la kinésine humaine Eg5 et de bloquer ainsi le développement des cellules cancéreuses.Le recherche de nouveaux ligands potentiels de la kinésine Eg5 est effectué selon un mode de conception rationnel fondé sur l'analyse de la structure tridimensionelle des complexes protéines/ligands ou "structure-based drug design". Ces travaux sont réalisés en utilisant les outils de modélisation moléculaire par la mise en oeuvre de méthode "de novo".L'ensemble des informations recueilli au travers de logiciels très performants permet l'obtention d'un modèle statistiquement significatif destiné à la conception et à la prédiction des activités biologiques.Ces travaux associés à l'expertise chimique du laboratoire, ont permis la conception de trois nouvelles familles, potentielles ligand d'Eg5, de structure: triazoloquinazolinone, triazolométhylquinazolinone, dihydroimidazoquinazolinone. / Cancer is a real problem in our civilization. Indeed, in 2010, it affected more than 10 million people in the world. Today, this disease is the first cause of death in industrialized countries.Unfortunately, the proposed therapies remain frequently insufficient and lead to side effects which remove the benefits of the medical treatment. In order to avoid the toxicity to safe cells, since a few years, researches have been done to develop targeted therapies. Most of the anti-mitotic drugs actually available on the market lead to important side effects such as cardiological, hematological and neurotoxic problems.Thus, we interested to another therapeutic target which still acts at the level of the mitosis but causing fewer side effects and can be overexpressed into the cancer cells: the mitotic kinesin Eg5.The mitotic kinesin Eg5 plays an important role in the early stages of mitosis and is one of the most attractive target enzymes in antimitotic drug development. The modulation of the Eg5 activity has been shown to cause aberrant mitotic spindle formation, cell cycle arrest during mitosis and the inhibition of proliferation of tumor cells in culture. With regard to the potential of Eg5 modulators in the treatment of human cancers, we report the design, synthesis and biological studies of quinazolinone derivatives as mitotic kinesin Eg5 inhibitors. We developed three series of molecules derived from quinazolin-4-one scaffold following a “de novo drug design” strategy.

During mitosis, replicated chromosomes are equally distributed among two daughter cells by means of a multi-component machine called the mitotic spindle. Spindle formation and function has been shown to involve numerous microtubule associated proteins and molecular motor proteins, including kinesins and dynein. One such kinesin, the plus-end directed, homotetrameric, Eg5, is involved in centrosome separation during spindle formation. In vitro, Eg5 crosslinks parallel and antiparallel microtubules, and localizes to spindle poles and microtubules in vivo. To further understand the function of Eg5 in mammalian cells, we determined its distribution and dynamics throughout mitosis using novel cloning techniques, fluorescence recovery after photobleaching, and total internal reflection fluorescence microscopy. Eg5-GFP was expressed from a mouse bacterial artificial chromosome to ensure the transgene’s expression was at or near endogenous levels. Our results confirm that Eg5 colocalizes with spindle, but not astral, microtubules and is enhanced at the spindle poles during prometaphase and metaphase. In early anaphase, Eg5 is localized near the poles transitioning to interzone microtubules with the exception of a 1 µm gap during late anaphase. Fluorescence recovery after photobleaching shows that Eg5 is rapidly turning over throughout mitosis with a recovery half time less than 10 s and extent of recovery greater than 85%. TIRF microscopy revealed a population of Eg5 that transiently binds to microtubules with a residency time of less than 6 seconds for all stages of mitosis. Eg5 remained stationary while bound to microtubules with no apparent directional motion. Treatment of cells expressing mEg5-GFP-LAP with the Eg5 inhibitor, STLC, caused Eg5 to no longer bind to microtubules and remain diffuse within the cell. TIRF microscopy also revealed Eg5-decorated tracks during interphase, which were abolished by treatment with STLC or Nocodazole suggesting that Eg5 is present on microtubules in interphase. Taken together, fluorescence recovery after photobleaching and TIRF microscopy reveal that Eg5 is highly dynamic in the mammalian spindle throughout mitosis.

The development and maintenance of multicellular organisms depends fundamentally on cell division, a series of events largely mediated by the mitotic spindle. Errors in spindle formation and/or function are often associated with severe consequences, most notably cancer. In order to elucidate the cause of such errors and the potential for therapeutic intervention, it is imperative to attain a clear understanding of how cell division normally operates. In this regard, this dissertation focuses on the activity of two microtubule-based motor proteins, Eg5 and dynein, prior to and immediately following nuclear envelope breakdown during mitosis. I show that prophase microtubules are remarkably more dynamic than their metaphase counterparts, moving both toward and away from centrosomes across a wide distribution of rates. Inhibition of Eg5, dynein and Kif2a revealed that a subset of this motion is consistent with microtubule flux, a well-established phenomenon temporally limited to metaphase and anaphase spindles by the preceding literature. My data indicates that flux is operational throughout all of mitosis, possibly functioning at early stages to collect centrosomal components. Immediately following prophase, cells begin assembling bipolar spindles. While the establishment of spindle bipolarity fails in the physical or functional absence of Eg5, I show that co-inhibition of dynein restores a cell’s ability to organize microtubules into a bipolar structure. Despite inhibition of both Eg5 and dynein, these spindles are morphologically and functionally equivalent to controls. Together, these data suggest that Eg5 and dynein share an antagonistic relationship and that a balance of forces, rather than a definitive set of players, is important for spindle assembly and function. To determine how Eg5- and dynein-mediated forces functionally coordinate to bring about antagonism during spindle assembly, I utilize a nocodazole washout assay. I show, via in vivo imaging and in silico modeling, that spindle collapse in the absence of functional Eg5 requires dynein activity and an initial intercentrosomal distance of less than 5.5μm. These data are consistent with a model in which dynein antagonizes Eg5 by crosslinking and sliding antiparallel microtubules, a novel role for dynein within the framework of spindle assembly.

Rôle des kinésines mitotiques Eg5 et MKLP-2 dans l’angiogenèse physiologique etpathologique.L’angiogenèse est un phénomène biologique complexe qui correspond à la formation de nouveauxvaisseaux à partir de vaisseaux préexistants. Ce processus essentiel est régulé par des nombreuxfacteurs, dont le plus puissant est le facteur de croissance de l’endothélium vasculaire (VEGF).Des inhibiteurs du VEGF sont actuellement utilisés dans le traitement de nombreux cancerssolides. Leur efficacité est constatée dans plusieurs études mais des résistances contre cesmolécules sont fréquemment observées. Afin d’identifier de nouvelles cibles thérapeutiques dansla voie de signalisation de VEGF, nous avons utilisé le modèle de la membrane chorioallantoïdienne(CAM) de l’embryon de poulet. Les CAM traitées au VEGF pendant 24hdéveloppent de nombreux vaisseaux. Ces tissus ont été isolés pour effectuer une analysetranscriptomique. En dehors des gènes endothéliaux déjà connus pour être régulés par le VEGF,de nouveaux gènes ont été identifiés. Nous avons focalisé notre recherche sur des gènes codantpour les kinésines mitotiques KIF11/Eg5 et KIF20A/MKLP-2 qui ont été fortement induites.Nous avons démontré qu’Eg5 et MKLP-2 sont fortement exprimées au niveau de l’endothéliumdans des tissus sains et dans des cancers solides. Des inhibiteurs chimiques spécifiques d’Eg5(dimethylenastron et ispinesib mesylate) et MKLP-2 (paprotrain) bloquent les étapes clés de laformation des vaisseaux sanguins (prolifération, adhérence et migration des cellules endothéliales),la prolifération des cellules tumorales ainsi que la formation de néo-vaisseaux dans des culturesd’anneaux aortiques. De plus, sur la CAM et chez la souris, l’inhibition de cette même kinésinediminue significativement la croissance et la vascularisation des modèles tumoraux utilisés lors dece projet (le glioblastome et le carcinome rénal). En conclusion, Eg5 et MKLP-2 pourraient descibles potentielles dans les thérapies anti-angiogéniques.Mots clés : Eg5, MKLP-2, angiogenèse, kinésine, ispinesib, dimethylenastron, glioblastome,cancer rénal / Role of the mitotic kinesins Eg5 and MKLP-2 in physiologic and pathologic angiogenesis.Angiogenesis is a complex biological phenomenon which corresponds to the formation of newblood vessels from pre-existing vessels. This process is regulated by a plethora of differentmolecules with vascular endothelial growth factor (VEGF) being one of the most important ones.VEGF inhibitors are currently used in the treatment of numerous solid cancers. Even though theefficacy of such treatment is prouven by numerous studies, resistance to anti-angiogenic therapy isa common feature. To identify new therapeutic targets downstream of VEGF, we modelized itsaction on the chick chorioallantoic membrane (CAM). VEGF-treated CAMs develop a densevascular network 24h after application. We used chick microarrays to monitor global geneexpression changes in VEGF-induced CAMs. Beside a consistent number of genes alreadydescribed to be regulated by VEGF, numerous unknown genes have been identified. We havefocused our work on the characterization of Eg5/KIF11 and MKLP-2/KIF20A, members of thekinesin family, both strongly upregulated by VEGF.We demonstrated that Eg5 and MKLP-2 are strongly expressed by blood vessels in normal andcancer tissue sections. KIF20A is involved in the proliferation and migration of endothelial cellsin vitro. We showed that chemical inhibitors specific for KIF11/Eg5 (dimethylenastron andispinesib mesylate) affect key steps in the formation of blood vessels (proliferation, adhesion andmigration of endothelial cells) and proliferation of tumor cells (glioma and renal cancer).Furthermore, in experimental glioblastoma and renal cell carcinoma models (CAM and orthotopicimplantation in mice), anti-Eg5 treatment strongly reduces tumor angiogenesis and growth. Inconclusion, Eg5 and MKLP-2 could be potential targets in anti-angiogenic therapies.Keywords: Eg5, MKLP-2, angiogenesis, kinesin, ispinesib, dimethylenastron, glioblastoma, renalcell cancer

Chromosome Instability (CIN) is defined as an increase in the rate at which whole chromosomes or large parts are gained or lost. CIN is not only associated with virtually all tumor types, but it is associated with aggressive tumors, tumor recurrence, acquisition of multidrug resistance and poor patient prognosis. However, the genes and molecular defects that contribute to CIN are poorly understood. I hypothesize that KIF11 is an essential gene for chromosomes integrity during mitosis and therefore any defect in KIF11 expression or function will induce CIN and contribute to tumorigenesis. Accordingly, KIF11 was either silenced using siRNA or inhibited using monastrol within two distinct human cell lines and was investigated for CIN associated phenotypes. Here, I have identified and validated KIF11 as a novel CIN gene. This study represents the first steps necessary to identify and develop novel treatments design to target origins of CIN in CIN associated cancers. / February 2017