Le cancer du Pancréas est un cancer extrêmement agressif avec un taux de survie à 5 ans ne dépassant pas les 5%. La détection tardive, le manque de biomarqueur et de thérapie efficace en sont les principales causes. Ce cancer est caractérisé par un microenvironnement très dense composé majoritairement de fibroblastes associés au cancer (CAF). Ce stroma est en constante communication avec la tumeur via la production de facteurs de croissance et métabolites, ce qui favorise le développement tumoral. Notre projet consiste à perturber ce dialogue en ciblant par des anticorps monoclonaux, la Neuréguline 1 (NRG1), ligand du récepteur HER3. Sa surexpression a été montrée dans des tumeurs pancréatiques, par les cellules tumorales et par les CAF. Nous avons confirmé l’implication de la NRG1 dans la croissance des tumeurs pancréatiques puis nous avons généré un anticorps monoclonal (7E3) murin spécifique de ce ligand. Cet anticorps a été caractérisé au niveau de sa spécificité, son affinité ainsi que son épitope sur la NRG1. Nous avons ensuite étudié son effet thérapeutique dans différents modèles de cellules pancréatiques seules ou en co-culture avec des fibroblastes activés provenant de tumeurs humaines. Le ciblage de la NRG1 par le 7E3 inhibe la signalisation de HER3 et la croissance cellulaire in vitro en 2D et en 3D. De plus, son efficacité sur le ralentissement de la croissance tumorale a été montrée dans des modèles originaux de greffes orthotopiques de cellules tumorales seules ou mélangées à des CAF chez la souris nude.Dans le but d’étudier plus précisément l’effet de notre anticorps sur le microenvironnement tumoral ainsi que la toxicité de notre stratégie, un second anticorps ciblant la NRG1 murine et humaine a été produit par phage display et est en cours de caractérisation in vitro et in vivo. Des combinaisons du 7E3 avec la chimiothérapie (Gemciatbine ou Folfirinox) ou d’autres thérapies ciblées sont envisagées.Cette thèse propose une nouvelle solution thérapeutique pour les patients atteints de cancer du Pancréas. Nous avons démontré l’importance de perturber le dialogue entre les cellules tumorales et les CAF en validant l’efficacité d’un tel ciblage par des anticorps monoclonaux. / Pancreatic cancer is an extremely aggressive cancer with a 5-year survival rate of no more than 5%. Late detection, lack of biomarker and effective therapy are the main causes. This cancer is characterized by a very dense microenvironment composed of cancer-associated fibroblasts (CAFs). This stroma is in constant communication with the tumor via the production of growth factors and metabolites, which promotes tumor development. Our project consists in disrupting this crosstalk with monoclonal antibodies targeting Neuregulin 1 (NRG1), ligand of the HER3 receptor. Its overexpression has been shown in pancreatic tumor cells and CAFs. We confirmed the involvement of NRG1 in the pancreatic tumors growth and then we generated a monoclonal antibody (7E3 mAb) specific for this ligand. This antibody has been characterized in terms of its specificity, its affinity and its epitope on NRG1. We then studied its therapeutic effect in different models of pancreatic cells alone or in co-culture with CAFs coming from human samples. The anti NRG1 antibody inhibits HER3 signaling as well as cell growth in 2D and 3D culture. In addition, its efficacy on tumor growth was shown in orthoptic xenograft models of tumor cells mixed with CAFs.In order to study more specifically the antibody effect on tumor microenvironment as well as its toxicity, a second mAb targeting murine and human NRG1 was produced by phage display. Its in vitro and in vivo characterization is ongoing. Combinations of 7E3 with chemotherapy (gemciatbine or Folfirinox) or other targeted therapies are being considered.This thesis proposes a new therapeutic solution for patients with pancreatic cancer. We demonstrated the importance of disrupting the crosstalk between tumor cells and CAF by validating the efficacy of targeting NRG1 by monoclonal antibody.

The central role of vascular endothelial growth factor (VEGF) in physiological and pathological angiogenesis makes it attractive both as a therapeutic target for anti-angiogenic drugs in cancer treatment and as a pro-angiogenic cytokine to treat ischemic disease. Currently, it is well established that the VEGF but also other growth factors exert their angiogenic effects partly through the activation of endothelial nitric oxide synthase (eNOS). A better understanding of the VEGF/NO signaling pathway could therefore lead to the identification of new therapeutic targets to impact on angiogenesis.This thesis is based on four different articles, the principal findings of which are summarized here below. First, we demonstrated that, chronologically, endothelial cell exposure to VEGF first led to eNOS dissociation from caveolin (a hallmark of the Ca2+/CaM-mediated activation of eNOS), and then to the interaction of eNOS with the heat shock protein hsp90. We also reported that eNOS-bound hsp90 could recruit VEGF-activated (phosphorylated) Akt to the complex, which in turn could phosphorylate eNOS on the serine 1177. Finally, we found that although the VEGF-induced phosphorylation of eNOS led to a sustained production of NO independently of a maintained increase in intracellular [Ca2+], this late stage of eNOS activation was strictly conditional on the initial VEGF-induced Ca2+-dependent stimulation of the enzyme. These data established the critical temporal sequence of events leading to the sustained activation of eNOS by VEGF and suggested new ways of regulating the production of NO in response to this cytokine through the structural protein caveolin and the ubiquitous chaperone protein, hsp90. A second study identified caveolin and Hsp90 as key players in the proangiogenic action of statins and therefore as potential pharmacological targets to modulate NO-dependent angiogenesis. We found that atorvastatin stabilized endothelial tube formation from both outgrowing and isolated macrovascular ECs cultured in Matrigel through a decrease in caveolin abundance and in its inhibitory interaction with eNOS. In a similar angiogenic assay, microvascular endothelial cells appeared also responsive to statins, not through a decrease in the caveolin pool (which is (too) large in these cells) but via the increased recruitment of hsp90 in the eNOS complex and the associated eNOS phosphorylation on the serine 1177. These data provided new mechanistic insights into the NO-mediated effects of statins and underscored the potential of these drugs and other modulators of hsp90 and caveolin abundance to promote neovascularization in disease states associated or not with atherosclerosis. In a third study, we have reported a net decrease in the ability of cultured ECs expressing recombinant caveolin to migrate and to form capillary like networks (e.g. the crucial steps occurring during the angiogenic process) in presence of VEGF. We then exploited the propensity of cationic lipids to target EC lining tumor blood vessels to transfect tumor-bearing mice in vivo. A dramatic tumor growth delay associated to a decrease in tumor microvessel density in the central core of the tumor was observed in mice transfected with caveolin versus sham-transfected animals. Interestingly, we also found that in the early time after lipofection (e.g. when macroscopic effects on tumor growth were not yet detectable), caveolin expression also impaired NO-dependent tumor blood flow. These findings indicated that besides (before) acting as an anti-angiogenic agent, recombinant caveolin can modulate the endothelium phenotype and impact on the tumor blood flow, both effects leading to a decrease in tumor growth. Finally, we showed that the activation of the VEGF/NO signaling pathway led to the down-regulation of adhesion molecules and to the anergy of endothelial cells. We found, indeed, that the adhesion of human CD8+ lymphocytes on microvascular endothelial cells exposed to TNF-a was dramatically reduced in the presence of VEGF. Interestingly, we also documented that the co-administration of the NOS inhibitor L-NAME or the Hsp90 inhibitor geldanamycin could restore this adhesion to the level originally obtained with TNF-a alone. Finally, we confirmed the key role of NO in the VEGF-mediated effects on the CD8+ adhesion by tipping the balance towards more or less angiogenesis through the transfection of caveolin siRNA or caveolin plasmid, respectively. In these experiments, lymphocyte adhesion appeared directly correlated to the extent of caveolin expression, confirming that the so-called anti-angiogenic strategy can directly impact on the phenotype of (tumor) endothelial cells and instead of (before) killing them, be exploited to potentiate cancer immunotherapy. In conclusion, by dissecting the post-translational regulation of eNOS, we have identified major therapeutic targets, namely caveolin and hsp90, that may be exploited either to block or promote angiogenesis. More particulary, cDNA encoding for these proteins or their mutant form, when combined with adequate mode of delivery, appeared to exert profound effects on the vascular compartment of tumors but also of ischemic tissues.

Le Cancer, l’une des pathologies les plus fréquentes au sein de la population, possède encore actuellement un taux de morbi-mortalité important tous sexes confondus, et ce malgré les importants progrès diagnostiques et thérapeutiques réalisés. D’un point de vue diagnostique, dans une approche dite de « Biologie de systèmes », en complément de l’étude anatomo- pathologique qui reste la référence, de nouvelles techniques ont été développées pour la caractérisation de profils métaboliques (Métabolomique) d’échantillons tissulaires pathologiques ou non, parmi lesquelles la Spectroscopie RMN HRMAS. Après un bref rappel théorique et avoir dressé le bilan des applications de cette technique en Cancérologie, nous avons exposé les différentes étapes du protocole à mettre en place afin d’envisager son implémentation dans un cadre hospitalier. L’ensemble des résultats présentés permettent d’envisager l’utilisation de cette technique en pratique clinique courante. Il faut néanmoins valider la robustesse des modèles statistiques élaborés et confirmer ces résultats sur de plus grandes cohortes d’échantillons. Des développements technologiques, analytiques et statistiques sont également nécessaires. / Cancer, one of the most frequent pathologies among the population, has still an important morbidity-mortality rate all sex confounded, despite the important diagnostical and therapeutical progresses achieved. From a diagnostical point of view, in a so called “Systems Biology approach”, as a complement of the gold standard histopathological study, some new techniques have been developed for the characterization of metabolic profiles (Metabolomics) of tissular samples pathological or not, among which HRMAS NMR Spectroscopy. After some brief theoretical considerations and after reporting the applications of this technique in Cancerology, we exposed the different steps of the protocol to design in order to consider its implementation in a hospital set up. All the results presented allow considering the use of this technique in a clinical routine. Nevertheless, it is necessary to validate the robustness of the statistical models built and to confirm these results on much larger cohorts of samples. Some technical, analytical and statistical developments are also needed.

L’information est un des besoins les plus importants des patients atteints de cancer. Leurs attentes fortes, variables et hétérogènes, en matière d’informations sur les différents aspects de la prise en charge de la maladie, soulèvent la question de la qualité de ces informations mises à disposition des patients. L’élaboration d’outils d’information et d’aide à la décision de qualité, qu’ils soient destinés aux cliniciens, tels que les recommandations pour la pratique clinique (RPC), ou aux patients, tels que des documents écrits d’information, implique la prise en compte des trois composantes essentielles de la décision médicale : données actuelles de la science (evidence-based medicine), expertise professionnelle, et préférences et valeurs des patients. Pour pouvoir baser la décision médicale sur ces trois dimensions, l’enjeu est de partager avec les patients les données actuelles de la science, et d’intégrer les préférences des patients dans les RPC. Notre hypothèse est qu’il est possible d’intégrer les préférences des patients à deux niveaux de production des connaissances : l’élaboration de documents écrits d’information des patients dans le cadre du programme SOR SAVOIR Patient d’une part, et l’élaboration de RPC pour les cliniciens d’autre part. Les résultats de ces travaux fournissent une contribution méthodologique pour améliorer la qualité des documents écrits et impliquer les patients atteints de cancer dans le développement de ces informations / Information is one of the most important needs of cancer patients. Their strong, variable and heterogeneous expectations, in terms of information on different aspects of the management of the disease, raise the question of the quality of the information made available to patients. The development of good-quality information materials and decision aids, dedicated to clinicians, such as clinical practice guidelines (CPG), or dedicated to patients, such as patient information leaflets, requires taking into account the three essential components of medical decision: current scientific data (evidence-based medicine), professional expertise and patient values and preferences. To be able to base medical decisions on these three dimensions, the challenge is to share with the patients current scientific data, and incorporate patient preferences in the development process of CPG. Our hypothesis is that it is possible to integrate the preferences of patients at two levels of knowledge production: the development of written patient information in the French SOR SAVOIR Patient program on the one hand, and development of CPG for clinicians on the other hand. The results of this work provide a methodological contribution to improve the quality of written documents and involve cancer patients in the development of this information.