Associer les thérapies ciblées et la chimiothérapie cytotoxique constitue un axe thérapeutique extrêmement prometteur en oncologie clinique. La combinaison des fluoropyridines orales de dernière génération (capécitabine) aux inhibiteurs de tyrosine kinase (erlotinib, lapatinib)a ainsi été proposée dans divers contextes cliniques, avec une efficacité variable, allant de l'échec (pancréas) à la réponse thérapeutique (sein). L'objectif de ce travail de recherche est de dégager, sur des modèles expérimentaux non-cliniques, un rationnel pharmacologique d'utilisation de la capécitabine avec divers TKI. Ce travail se focalise notamment sur l'étude de la modulation des déterminants de réponse aux drogues dans les modèles pancréatiques(association capecitabine + erlotinib) et mammaires (capecitabine + lapatinib). Nous avons ainsi pu montrer les effets délétères de l'association erlotinib-capecitabine dans les modèles pancréatiques chimio-résistants, liés d'une part à un effet represseur de l'expression du récepteur à l'EGF par la capécitabine, combiné à une sécrétion exocrine de VEGF en réponse aux diverses séquences d'association testées. Ces effets se sont soldés par l'obtention de réponses antagonistes aux drogues combinées, comparativement aux monothérapies respectives. Nos données suggèrent que quelle que soit la séquence d'utilisation envisagée,capecitabine et erlotinib ne devraient jamais faire l'objet d'association. A contrario, l'étude de l'association lapatinib-capecitabine sur diverses lignées mammaires a permis de mettre en évidence un effet positif allant de la supra-additivité à la synergie. Nos données suggèrent que c'est d'une part par une dérégulation de la thymidylate syntahse par le lapatinib qu'une plus grande efficacité anti-proliférative est obtenue. Par ailleurs, une meilleure inhibition de la voie p-AKT après exposition à la capécitabine des cellules traitées par le lapatinib pourraitde surcroît expliquer la synergie observée. Nos données suggèrent qu'une utilisation concomitante de ces deux molécules devrait garantir un gain d'efficacité optimal. La synergie de l'association a été conservée sur un modèle de souris porteuses de tumeurs mammaires ectotopiques et traitées de façon concomitante, avec un net ralentissement de la croissance tumorale comparativement aux animaux traités par monothérapie, sans augmentation de la toxicité médicamenteuse.En conclusion, nos données illustrent le fait que les associations thérapiesciblées/antimétabolites doivent faire l'objet systématiquement d'un travail préalable de recherche des déterminants oncopharmacologiques d'efficacité afin d'anticiper le gain d'efficacité qu'on est en droit d'en tirer, et d'identifier au mieux les modalités d'association de ces drogues. / Optimizing cancer treatments remains an on going task. The rise over the last decade of a new class of anticancer agents (the so-called “targeted therapies”) has fuelled agreat hope in the search for achieving better response with lower toxicities. To date,the actual efficacy of targeted therapies used alone looks in some respect below the initial expectations, and associating them with classic cytotoxic drugs represents therefore a new promising combinational strategy in clinical oncology. However, littleis known on the pharmacologic rationale for combing these drugs, and identifying molecular determinants could help to rationalize the way the drugs should beassociated. Anti-EGFR erlotinib (Tarceva®) has been approved for the treatment of patients with metastatic pancreatic cancer, in combination with a standard chemotherapy, gemcitabine, both in the USA and in Europe. Its combination with capecitabine has recently been tested in patients with gemcitabine-refractorypancreatic tumours, with limited success. Lapatinib (Tykerb®) is an innovative small molecule with unique dual-TK inhibitory action against both HER1 and HER2.Lapatinib showed clinical efficacy in various models and has heightened the expectation for new breast cancer therapies. Associating lapatinib with capecitabine(Xeloda®) has been recently approved for the treatment of patients with HER2+metastatic breast cancer, after failure of trastuzumab treatment, both in the USA andin Europe as well. The aim of our work was to study, either in vitro or in tumor bearing animals, the effects of these drugs used either alone or in combination. In this preclinical study, we studied the molecular determinants of this new targeted therapies + antimetabolite association on various, representative human pancreatic and breast cancer cell lines. Efficacy and apoptosis experiments were performed after concomitant and various sequential combinations, along with study ofresponse/resistance determinants such as dihydropyrimidine dehydrogenase (DPD),thymidylate synthase (TS), thymidine phosphorylase (TP), Bax, Bcl-2, P21 andHER1/2 signaling pathway. Determination of Chou & Talalay combination indices (CI)allowed us to determine whether the tested associations led to synergism, additivity or antagonism. The best sequences found were transposed to tumor-bearing animal model for in vivo confirmation of the efficacy observed in vitro. Lapatinib andcapecitabine showed additive to moderate synergistic effects in our breast cancer celllines. Further experiments were performed to identify the mechanisms underlyingthis efficacy. We first found that lapatinib down-regulates thymidylate synthase (-30%) but did not have any effect on dihydropyrimidine dehydrogenase levels, thussuggesting theoretically a better cytotoxic effect of capecitabine within tumor cellspreliminary exposed to Lapatinib. Besides, capecitabine, when added to lapatinib,strongly down-regulates P-AKT and p-P42/44, thus allowing a deeper inhibition of the downstream signaling pathway after inhibition of HER1 and HER2. Muchinterestingly, we showed as well an increase of the Bax/Bcl2 ratio in cells exposed toboth drugs, leading to the enhancement of apoptosis induction and cell death. Thisfinding is fully consistent with the fact that the association of lapatinib andcapecitabine led to an increase of the P21 protein expression and consequently tothe arrest of cell cycle indeed. Besides, in pancreatic cancer cell lines, wedemonstrated that erlotinib up-regulated thymidine phosphorylase (+50%) anddown-regulated dihydropyrimidine dehydrogenase (-55%), thus suggestingtheoretically a better activation of capecitabine within tumor cells preliminaryexposed to erlotinib. We checked as well that Bax, P27, p-AKT were fully functionalin our models. Nevertheless, when not antagonist, only moderate additive effectswere achieved. To understand these rather surprising results, further experimentswere performed to identify the putative causes of this lack of efficacy. We first foundthat capecitabine down-regulates EGFR (-35%) in capan-1 cells with no effect on itsdownstream signalisation pathway nor antiproliferative related effects, thussuggesting a possible loss of efficacy when cells were subsequently treated witherlotinib due to the diminution of its target. Much interestingly, we showed as well adramatic overexpression of both cytosolic (+100 up to +1900%) and extra-cellular(+100 up to 3500%) VEGF secretion in response to our treatments, leading possibly to a boost in cell proliferation. Taken together, this in vitro study suggests that themore-than-additive efficacy observed when associating lapatinib and capecitabine inbreast cancer is based upon TS modulation and down-regulation of HER1/2 signalingpathways, resulting in apoptosis and cell death eventually. In vivo experiments are ongoing to fully confirm these data. Concerning the erlotinib/capecitabineassociation in pancreatic cancers, we suggest that VEGF secretion could be a new,major mechanism of resistance to the erlotinib + capecitabine association inpancreatic cancers, and, possibly, to other targeted therapy + antimetabolitescombinations. To test this hypothesis, further experiments including the use of anti-VEGF bevacizumab will be undertaken.

Comprendre les mécanismes de sénescence ou développer de nouveaux anti-cancéreux impose d’utiliser des modèles biologiques divers et complémentaires. De par sa facilité de mise en œuvre, la culture cellulaire est une des méthodes les plus employées. Cependant, travailler sur des cellules isolées ne permet pas toujours d’extrapoler les résultats à des conditions plus complexes comme un organisme entier. Les biologistes disposent alors de modèles intégrés plus ou moins évolués : levures, poissons (zebrafish …), grenouille (xénope …), insectes (drosophile …), rongeurs (souris, rats …), chien, primates, etc. Cependant leur mise en œuvre est souvent difficile, tant d’un point de vue administratif que technologique.L’objectif de notre travail a été d’initier le développement de deux nouveaux modèles biologiques intégrés (Caenorhabditis elegans et Hypsibius dujardini), beaucoup plus facilement manipulables, pour des études en oncopharmacologie et sénescence.Le potentiel de Caenorhabditis elegans a été évalué au travers de deux axes : (i) nous avons utilisé le nématode pour déterminer les potentiels toxiques, mutagènes et cancérogènes de gaz candidats au remplacement du gaz SF6 (ii) à l’aide du Seahorse XF24 Analyzer, nous avons mis au point un test fonctionnel original : la mesure de la respiration mitochondriale du nématode entier.Nous avons choisi de développer comme second modèle, Hypsibius dujardini, beaucoup moins travaillé que C. elegans. C’est un tardigrade connu pour sa résistance aux conditions extrêmes, notamment à la dessication. Afin de mieux comprendre cette résistance, nous avons choisi d’analyser le fonctionnement mitochondrial de cet organisme en sortie d’anhydrobiose, en utilisant des marquages mitochondriaux et en mesurant sa respiration à l’aide du Seahorse XF24 Analyzer. / Biological models are necessary to understand how organisms works, how evolution of living run, to test new treatments or to perform toxicological assessments as well. Cellular culture is one of the methods employed because it is easy to use. However, working on isolated cells doesn’t always allow to challenge of the results with more complex conditions as found in full organisms. Biologist needs to develop new biological models for new assessments but the new model choice can be a problem. Murine model, frog, drosophila, yeast, zebrafish,… have each other some benefits and limits. Their choice is directly linked with their use and with the type of research we intend to make.The aim of our work was to develop biological models to oncopharmacology and aging studies.The nematode model with Caenorhabditis elegans was used in several projects. One study was made on gases. They were tested in terms of toxicity, mutagenicity and cancerogenicity. On the other hand, a new tool was developed for prospective studies on either toxicology or mechanistic with the mitochondrial respiration measure with the Seahorse XF24 Analyzer device.The second biological model studied is the tardigrade and more exactly Hypsibius dujardini. Tardigrades are extremely resistant organisms to harshest conditions. They can resist to desiccation. To gain insights on tardigrade resistance, we have choice to analyze the mitochondrial dynamics in the course of anhydrobiosis exit by using mitochondrial dyes and respiration measurements.