Ventilatory drive is mediated by respiratory central pattern generators that are located in the brainstem, which are continuously modulated by specialised peripheral and central chemoreceptors to adjust ventilatory patterns according to changes in arterial PO2. These specialised oxygen-sensing chemoreceptors are activated in response to acute reductions in arterial PO2 and ultimately trigger a respiratory response that acts to restore oxygen-levels. However, the molecular mechanism by which mammals are able to regulate their breathing pattern in such a manner during hypoxia remains controversial. Therefore, the studies performed in this thesis aimed to investigate the possibility that this process may be mediated by the liver kinase B 1 (LKB1)/ AMP-activated protein kinase (AMPK) signalling pathway, which is central to cellular adaptations to metabolic stress. This first involved the development of transgenic mice in which Lkb1 or AMPK were deleted. Global knockout of Lkb1 (Sakamoto, 2006) or AMPK activity (Viollet et al., 2009) are embryonic lethal. Thus, the Cre/loxP system was used to develop transgenic mice that had either Lkb1 or both isoforms of the AMPK catalytic α- subunit (α1 and α2) conditionally knocked out in catecholaminergic cells (including therein hypoxia-activated cells of the brainstem and carotid body) by driving Cre expression through a tyrosine-hydroxylase-specific promoter region. The consequent effects on the ventilatory response to hypoxia were then examined using unrestrained whole-body plethysmography. This demonstrated that, in contrast to the hyperventilation evoked in controls, increased ventilation was virtually abolished in the Lkb1 and AMPK α1 and α2 double knockouts during hypoxia. Both knockout mice also exhibited periods of hypoventilation with frequent apnoeas during hypoxia. Additionally, studies on single AMPK α1 and AMPK α2 knockouts identified that the ventilatory dysfunction in AMPK α1 and α2 double knockouts was primarily caused by AMPK α1 deletion. In contrast, the severe ventilatory abnormalities exhibited during hypoxia following the deletion of Lkb1 and AMPK in catecholaminergic cells were mostly reversed upon exposure of mice to hypoxia with hypercapnia. Also, the ventilatory response to hypercapnia alone was without any major effect as a result of Lkb1 deletion or the dual-deletion of AMPK α1 and α2 catalytic subunits in catecholaminergic cells. This thesis therefore demonstrates, for the first time, that the LKB1-AMPK signalling pathway is key to respiratory adaptations during hypoxia, by regulating catecholaminergic oxygen-sensing cells, thus protecting against hypoventilation and apnoeas. The LKB1-AMPK signaling pathway can thereby determine oxygen and energy supply at both a cellular and whole-body level.

Les leucémies aigues myéloïdes (LAM) sont des hémopathies malignes de mauvais pronostic dont les thérapies actuelles ne permettent d’obtenir des taux de survie à 5 ans chez les adultes que d’environ 40%. Par conséquent, il est nécessaire d’approfondir nos connaissances concernant les mécanismes d’oncogenèse pour développer de nouvelles approches thérapeutiques. Malgré leur hétérogénéité clinique et biologique, les LAM ont certaines caractéristiques communes comme l’activation de la voie de signalisation mTORCl qui est détectée dans la plupart des échantillons de LAM. MTORCl contrôle la survie, la croissance et la prolifération cellulaire, notamment via le contrôle de la traduction des ARNm et donc de la synthèse protéique. Au cours de ce travail, nous montrons qu’il existe, dans les LAM, une dérégulation de mTORCl qui explique les limites des effets anti-leucémiques observés avec la rapamycine (un inhibiteur allostérique de mTORCl) et qui est médiée en partie par l’activité de la sérine thréonine kinase Pim2, qui contrôle la phosphorylation de la cible de mTORCl, la protéine 4E-BP1. Cependant, cibler directement la traduction produit des effets anti-leucémiques importants, ce que nous avons montré en utilisant une molécule inhibant spécifiquement le complexe d’initiation de la traduction, le 4EGI-l. EIF4E est essentiel à l’initiation de la traduction et nous avons montré sa surexpression au niveau protéique dans la plupart des échantillons de LAM au diagnostic par comparaison à des cellules hématopoïétiques normales CD34+. Bien que son niveau d’expression n’ait pas de valeur pronostique intrinsèque, ce résultat suggère un potentiel important au blocage de la traduction dans la plupart des cas de LAM. Dans la perspective d’inhiber mTORCl, nous avons activé la voie LKBl/AMPK par la metformine, ce qui a induit des effets anti-leucémiques in vitro et in vivo via une modification du métabolisme cellulaire avec en particulier une inhibition de la synthèse de protéines oncogéniques. La metformine n’étant pas un candidat en thérapeutique dans les LAM du fait d’un index thérapeutique trop étroit, de nouvelles molécules modulant la voie LKBl/AMPK sont en cours de développement. Enfin, nous avons étudié le rôle de la sérine thréonine kinase Pim2, qui contrôle la traduction protéique et la survie dans les cellules de LAM Flt3-ITD+. Nous avons de plus montré que la sur-expression de Pim2 constitue un nouveau mécanisme de résistance aux inhibiteurs de Flt3 et représente donc une cible thérapeutique prometteuse dans cette catégorie de LAM. L’étude de la voie mTORCl et de la traduction permet donc d’envisager de multiples perspectives thérapeutiques dans les LAM dont certaines sont déjà en cours de développement clinique. / Acute myeloid leukemia (AML) are hematological malignancies with adverse prognosis in which therapies only gives 40% survival within 5 years in adults. Hence, it is important to increase our knowledge regarding oncogenesis to further develop new therapeutic approaches. Despite their clinical and biological heterogeneity, AML have in common the constitutive activation of mTORC1 signaling which is detected in most AML samples. MTORC1 controls cell survival, growth and proliferation, in particular through control of mRNA translation and protein synthesis. During this work, we show, in AML, that mTORC1 is deregulated which explain the poor effects observed with rapamycin (a mTORC1 allosteric inhibitor) and is partially mediated by the serine/threonine kinase Pim-2 which controls the mTORC1 target 4E-BP1. Nevertheless, directly targeting translation, using a specific translation initiation inhibitor named 4EGI-1, have important anti leukemic effects. EIF4E is described as essential in translation initiation and we show its protein overexpression in most AML samples at diagnosis compared with normal hematopoietic CD34+ cells. Whereas eIF4E level expression has no prognostic impact, this result suggests an important potential for treatment targeting translation initiation in AML. In our purpose of inhibiting mTORC1, we were able to activate LKB1/AMPK signaling pathway with metformin, which induces anti leukemic effects in vitro and in vivo through in particular oncogenic protein translation inhibition. Metformin is not a good AML therapeutic candidate because of a narrow therapeutic index, new compound targeting LKB1/AMPK are in development. Finally, we studied the role of the serine/threonine kinase Pim-2 and show that it controls protein translation and FLT3-ITD+ AML cells survival. Furthermore, we show that Pim-2 overexpression is a new mechanism of Flt3 inhibitors resistance and represent a new promising therapeutic target in this AML subtype. Overall, mTORC1 and protein translation study in AML show multiple therapeutics perspective, some of them are already in clinical development.