A complex system exists to monitor the body’s energy status and regulate food intake and energy expenditure to maintain a constant body weight. However, this homeostatic system is not the sole system regulating appetite. The hedonic system comprised of the mesolimbic reward pathway influences motivation to eat and acts alongside the homeostatic system to control feeding behaviours. It is often assumed that the hedonic system promotes the consumption of palatable, energy-dense foods and this can disrupt homeostatic mechanisms regulating food intake, resulting in energy overconsumption and weight gain in the long term. Yet, it is unclear to what extent the homeostatic system can defend body weight in an environment rich in palatable, energy-dense foods. I hypothesised that the homeostatic system compensates for the energy in palatable foods by reducing subsequent energy consumption, defined as homeostatic caloric compensation. I investigated homeostatic caloric compensation in a rat model of restricted palatable, high-sugar food access. Rats were schedule-fed moderate amounts of sweetened condensed milk (SCM) daily in addition to ad lib bland diet access. Both male and female rats calorically compensated for the energy consumed from moderate amounts of SCM through a robust and accurate reduction in energy consumed from bland diet, resulting in no short-term changes in body weight gain. However, homeostatic responses were limited as male rats were unable to fully calorically compensate for the scheduled-feeding of large amounts of SCM, an apparent loss of homeostatic control. It was not investigated whether female rats are also unable to fully calorically compensate for large amounts of SCM. It is possible that male rats consume these large amounts of SCM due to hedonic drive but continue to eat bland diet to acquire nutrients that are not present in SCM. To determine whether male rats defend bland diet consumption due to nutrient requirements, rats were schedule-fed large amounts of SCM enriched with protein or fibre. However, male rats did not fully calorically compensate for the energy in large amounts of SCM when enriched with protein or fibre. Overall, these findings demonstrate that the homeostatic system is able to respond to the hedonic consumption of palatable food through caloric compensatory mechanisms to defend body weight. However, it appears that the homeostatic system is unable to effectively respond to excessive hedonic palatable food consumption through caloric compensation alone. To shed light on what homeostatic mechanisms may underlie this compensatory behaviour, I used expression of the immediate early gene c-Fos to investigate neuronal activity following the scheduled-feeding of moderate amounts of SCM in male rats. c-Fos expression was increased in the ventral tegmental area of the mesolimbic reward pathway and in the lateral hypothalamus. The lateral hypothalamus has been proposed to act as an interface between homeostatic and hedonic systems. Therefore, in response to the hedonic consumption of palatable food, the homeostatic system and reward pathway may interact. Additionally, c-Fos expression was increased in satiety mediating brain regions of the homeostatic system, including the nucleus of the solitary tract and dorsomedial hypothalamus. This suggests that the homeostatic system may compensate for the energy in the palatable food by reducing subsequent food intake through inducing satiety. Furthermore, following the consumption of SCM, c-Fos expression was increased in magnocellular oxytocin neurons of the hypothalamic supraoptic and paraventricular nucleus. I demonstrated that the oxytocin system was activated by gut-brain signalling potentially involving the nucleus of the solitary tract. Therefore, the oxytocin system may be involved in homeostatic compensatory mechanisms triggered in response to the hedonic consumption of SCM, as part of a pathway mediating satiety. Moreover, I showed that c-Fos expression was also increased in the hypothalamic supramammillary nucleus (SuM) following the consumption of SCM. It has been previously shown that the SuM is involved in reward-related motivated behaviours and was recently implicated in the motivation to acquire and consume palatable food rewards. I also demonstrated that c-Fos expression in the SuM might be specific to the motivated consumption of palatable food, consistent with the SuM being involved in reward-related motivated behaviours. Furthermore, there is additional evidence from these studies that the SuM may functionally communicate with brain regions in the homeostatic and hedonic systems, including the lateral hypothalamus, dorsomedial hypothalamus and ventral tegmental area. Finally, I explored whether the gut-secreted orexigenic hormone ghrelin activates the SuM, as ghrelin may act at the SuM to influence feeding motivation. However, systemic ghrelin administration did not influence SuM c-Fos expression. As the SuM is activated following the consumption of SCM and may act as an interface between the homeostatic and hedonic systems, it is possible that the SuM could be a key component in the regulation of hedonic feeding. Using a rat model, I have shown that homeostatic compensatory mechanisms are triggered in response to the hedonic consumption of palatable, high-sugar food to regulate energy intake. This response is likely to involve homeostatic satiety mechanisms and interactions between multiple brain regions involved in the homeostatic and hedonic control of food intake. Overall, these findings shed light on how the homeostatic system responds to hedonic energy consumption and highlights specific brain regions that may be involved in hedonic feeding or homeostatic compensatory responses.

Les épilepsies mésiales du lobe temporal (ELTM) sont parmi les formes les plus fréquentes d’épilepsies partielles pharmaco-résistantes de l’adulte et l’enfant. Dans ces épilepsies les études électrocliniques et expérimentales indiquent que la zone épileptogène, qui désigne un ensemble de neurones nécessaire et suffisant à l’organisation d’une décharge anormale, ne peut être réduite à la seule formation hippocampique (FH) et impliquerait une réorganisation mettant en jeu plusieurs structures au sein du système limbique. Dans ce travail de thèse, nous nous sommes intéressés à la connectivité structurale entre le noyau supramammillaire (SuM) et la FH chez le rat dans le but de déterminer l’identité neurochimique de la voie de projection supramammillaro-hippocampique qui n’avait pas été clairement identifiée et, vérifier l’hypothèse d’une éventuelle réorganisation de cette voie de projection dans le modèle d’ELTM induit par l’injection intrapéritonéale de pilocarpine chez le rat. Chez les rats naïfs, nous mettons en évidence deux voies de projection distinctes. La première a pour origine les neurones localisés dans la partie latérale du SUM (SuML) qui innervent le champ CA2-CA3a et principalement la couche supragranulaire du gyrus dentelé dorsal. Cette voie est essentiellement ipsi-latérale et a la caractéristique de présenter un profil neurochimique unique, à la fois GABAergique et glutamatergique. La seconde voie de projection a pour origine les neurones localisés dans la partie plus postérieure et médiane du SuM (SuMM) qui innervent la région CA2-CA3a et la région ventrale du gyrus dentelé exclusivement ; cette voie est purement glutamatergique. Chez les rats traités à la pilocarpine, nos résultats montrent une réorganisation structurale des afférences des noyaux SuML et SuMM qui innervent le gyrus dentelé. Cette réorganisation est caractérisée par une distribution aberrante et une augmentation du nombre de fibres et terminaisons axonales en provenance des noyaux SuML et SuMM dans la couche moléculaire interne du gyrus dentelé. Cette réorganisation commence à la fin de la période de latence, et évolue pendant l’épilepsie induite par la pilocarpine. Avec ce travail, nous montrons pour la première fois : 1) l’hétérogénéité à la fois anatomique et neurochimique des voies de projection supramammillaro-hippocampiques ; 2) dans le gyrus dentelé des animaux traités à la pilocarpine, une réorganisation structurale d’origine extra-hippocampique, en provenance des noyaux SuML et SuMM. Cette connectivité aberrante pourrait contribuer avec la réorganisation des circuits intrinsèques de l’hippocampe à l’émergence des premières crises spontanées et à l’installation de l’épilepsie. / Mesial temporal lobe epilepsies (MTLE) are among the most common forms of pharmacoresistant partial epilepsies in adults and children. In these epilepsies, spontaneous seizures likely originate from a multi-structural epileptogenic zone including several structures of the limbic system connected to the hippocampal formation (HF). In this thesis, we investigate the structural connectivity between the supramammillary nucleus (SuM) and the HF in rat, in order to determine the not yet known neurochemical identity of the supramammillaro-hippocampal pathway and, to test the hypothesis of a potential reorganization of this pathway in the rat pilocarpine model of MTLE. In naïve rats, our results highlight two distinct pathways. The first pathway originates in the lateral part of the SuM (SuML) and innervates the supragranular layer of the dorsal dentate gyrus mainly, and the CA2-CA3a pyramidal cell layer of the hippocampus. This pathway is mainly ipsilateral and displays a unique dual phenotype for GABAergic and glutamatergic neurotransmission. The second pathway originates in the most posterior and medial part of the SuM (SuMM) and innervates exclusively the inner molecular layer of the ventral dentate gyrus and the CA2-CA3a subfield and is glutamatergic only.In pilocarpine-treated animals, our findings demonstrate a structural reorganization of dentate gyrus afferents originating from the SuM nuclei. Such reorganization is characterized by an aberrant distribution and an increased number of fibers and axon terminals from neurons of the both lateral and medial regions of the SuM, invading the entire inner molecular layer of the dentate gyrus. It starts at the end of the latent period and evolves during the epilepsy induced by pilocarpine. Our findings demonstrate for the first time: 1) the anatomical and neurochemical heterogeneity of the supramammillaro-hippocampal pathways; 2) in pilocarpine-treated animals, a marked reorganization of dentate gyrus afferents originating from the SuM nuclei. This aberrant connectivity could contribute along with the reorganization of hippocampal intrinsic circuitry to the emergence of the first spontaneous seizures and epilepsy installation.

Ce travail s'inscrit dans l'étude du réseau neuronal responsable de l'activation corticale au cours du sommeil paradoxal (SP) chez le rongeur. Dans la première partie de ma thèse, j'ai participé à la démonstration que cette activation est limitée à quelques structures limbiques déterminantes pour l'apprentissage, dont le gyrus dentelé de l'hippocampe (GD). Nous avons ensuite montré que l'activation du GD en SP est due à une projection issue du noyau supramammillaire (Sum). J'ai ensuite montré en combinant l'hybridation in situ d'un marqueur des neurones glutamatergiques et GABAergiques et l'immunohistochimie du FOS que les neurones du Sum latéral actifs en SP sont à la fois glutamatergiques et GABAergiques (GLU/GABA). Enfin, j'ai montré que l'augmentation du nombre de neurones FOS+ dans le GD dorsal en SP est abolie après la lésion neurochimique du Sum. De plus, la lésion du Sum induit une nette réduction de la densité de fibres glutamatergiques dans le GD dorsal. Ces résultats indiquent que les neurones du GD dorsal sont activés en SP par les neurones GLU/GABA du Sum latéral. Le deuxième objectif de ma thèse a été de déterminer la fonction de cette voie en SP. Ainsi j'ai utilisé la technique d'optogénétique afin d'inactiver ou activer les fibres GLU/GABA provenant du Sum localisées dans le GD dorsal au cours du SP. Nos résultats montrent que l'activation de ces fibres en SP induit une augmentation de la fréquence et de la puissance du thêta enregistré dans le GD. Ces résultats indiquent que la voie Sum-GD dorsal contrôle le thêta hippocampique et soutiennent l'hypothèse d'un rôle de cette voie dans les processus de consolidation mnésique prenant place au cours du SP / During my PhD I studied the neuronal network responsible for cortical activation during paradoxical sleep (PS) in rodents. In the first part of my thesis, I participated to the demonstration that this activation is limited to a few limbic structures involved in learning, including the dentate gyrus of the hippocampus (DG). Then, we showed that the activation of DG during PS is due to a projection from the supramammillary nucleus (Sum). Besides, by combining the in situ hybridization of markers of GABAergic and glutamatergic neurons and FOS immunohistochemistry, I demonstrated that lateral Sum neurons active in SP are both glutamatergic and GABAergic (GLU/GABA). Finally, I showed that the increasing number of FOS+ neurons in the dorsal DG during PS is abolished by the neurochemical lesion of the Sum. In addition, the Sum lesion induces a clear reduction of the density of glutamatergic fibers in the dorsal DG. These results indicate that during PS, dorsal DG neurons are activated by GLU/GABA neurons located in the lateral Sum. The second aim of my thesis was to determine the function of this pathway during PS. To realize that, I inactivated or activated by optogenetics the Sum GLU/GABA fibers located in the dorsal GD during SP. Our results show that the activation of these fibers during SP induces an increase in the theta frequency and power recorded in the dorsal DG. These results indicate that the Sum-dorsal DG-pathway modulates the hippocampal theta and supports the hypothesis of a role of this pathway in the memory consolidation process during SP

Afin d'avancer sur la fonction du sommeil paradoxal, il est nécessaire d'étudier son impact sur le fonctionnement cortical. Nous avons ainsi comparé l'expression génique corticale à l'aide de puces à ADN chez trois groupes de rats présentant différentes quantités de sommeil paradoxal (SP) : témoins, privé de SP ou en hypersomnie de SP. 71 et 83 transcrits montrent un niveau d'expression modifié par notre protocole dans le néocortex et l'hippocampe, respectivement. Ces résultats moléculaires ont été confirmés par PCR quantitative. Dans l'hippocampe l'expression des gènes de plasticité (Fos, Arc, Cox2, Homer1...) augmente en hypersomnie de SP. Au contraire, dans le néocortex le niveau d'expression de ces gènes augmente après privation de SP. Au niveau systémique, les aires limbiques (le gyrus dentelé, le cortex cingulé antérieur et rétrosplénial et le claustrum) contiennent un nombre de neurones immunoréactifs au FOS, un marqueur d'activation indirect, élevé après hypersomnie de sommeil paradoxal. En revanche, le nombre de neurones immunoréactifs au FOS dans les cortex sensoriels est diminué après hypersomnie par rapport à la privation de sommeil paradoxal L'éjection de traceurs rétrogrades dans le gyrus dentelé, le cortex rétrosplénial et le cortex cingulaire antérieur des rats en hypersomnie de SP a permis d'observer des neurones afférents et actifs dans les noyaux supramamillaires et le claustrum. Nous avons ensuite observé que le nombre de neurones immunoréactifs pour FOS, ARC dans le gyrus dentelé, le claustrum et certaines structures limbiques est fortement diminué pendant l'hypersomnie de SP chez des rats porteurs d'une lésion des noyaux supramamillaires. De plus, la lésion du Sum est accompagnée d'une diminution de la puissance du thêta enregistrée par l'électroencéphalogramme pendant le sommeil paradoxal en hypersomnie. Il semble donc que les projections des noyaux supramamillaires soient responsables de l'activation des régions limbiques corticales pendant le SP / To move forward on the PS function, it is necessary to study its impact on the cortical functioning. We so compared the cortical genic expression by using DNA microarrays in three groups of rats with different PS amounts: control, deprived of PS and in PS hypersomnia. 71 and 83 transcripts have an expression level modified by our protocol in the neocortex and the hippocampal formation, respectively. These molecular results were confirmed by quantitative PCR. In the hippocampal formation the genes involved in synaptic plasticity (Fos, Arc, Cox2, Homer1) have an expression level increased after PS hypersomnia. In the contrary, in the neocortex the expression level of these genes increases after PS deprivation. At the systemic level, limbic areas (the dentate gyrus, anterior cingulate and retrosplenial cortex and claustrum) contain a number of FOS immunoreactive neurons, an indirect marker of neuronal activation, increased after PS hypersomnia. On the other hand, the number of FOS immunoreactive neurons in the sensory-motor cortices is decreased after PS hypersomnia compare to PS deprivation. The ejection of retrograde tracers in the dentate gyrus, retrosplenial and anterior cingulate cortex in PS hypersomniac rats showed that active neurons project to the supramammillary nucleus and claustrum. We then observed that the number of FOS and ARC immunoreactive neurons in the dentate gyrus, claustrum and limbic structures is strongly decreased during PS hypersomnia in rats bearing a supramammillary nucleus lesion. Furthermore, the supramammillary nucleus lesion leads to a decrease of the theta power recorded by electroencephalogram during PS in hypersomnia. It thus seems that the supramammillary nucleus projections are responsible for the limbic cortical regions activation during PS