La maladie d'Alzheimer (MA) est une pathologie neurodégénérative s'exprimant par des troubles de la mémoire et un déclin cognitif évoluant progressivement vers un stade de démence incurable. Elle représente la cause principale de syndrome démentiel puisque l'on estime qu'elle est à l'origine de plus de 70% des cas de démences. Du fait de sa prévalence élevée après 60 ans, la MA représente un problème majeur de santé publique. La MA se caractérise par la présence de deux types de lésions cérébrales : les dégénérescences neurofibrillaires (DNF) et les plaques amyloïdes. Cependant, aucun consensus clair ne se dégage concernant les relations qui lient les deux types de lésions. Leur présence ne peut être mise en évidence que par un examen post-mortem. La MA est par définition une pathologie évolutive, cet examen ne permet donc pas de caractériser de manière adéquate les processus dynamiques qui sous-tendent cette pathologie. La mise au point de techniques d'imagerie non invasives permettant de réaliser un suivi longitudinal in vivo de ces lésions s'avère déterminante dans la compréhension de la physiopathologie de la MA. Les travaux effectués au cours de cette thèse ont pour objectif la mise au point de nouveaux radiotraceurs des lésions amyloïdes et neurofibrillaires pour l'imagerie nucléaire. Cette approche se distingue en trois parties. Dans un premier temps, nous avons validé un modèle animal de la MA: les souris transgéniques 3xTgAD. Dans un second temps, nous avons réalisé l'évaluation biologique de différents radiotraceurs connus sur ce modèle animal : le 99mTc-HMPAO, le 18F-FDG et le 125I-IMPY. Enfin, nous avons initié le développement de plusieurs nouveaux traceurs pour permettre le suivi de la MA in vivo : les para-sulfonato-calixarènes qui présentent une affinité intéressante pour les plaques amyloïdes, les composés COB qui inhibent la formation des agrégats de peptides amyloïdes in vitro ainsi qu'un peptide, A93, associé à un vecteur qui pourrait interagir avec les dégénérescences neurofibrillaires. / Alzheimer's disease (AD) is a neurodegenerative pathology showing cognitive and memory disorders which progress toward an incurable demential state. AD represents the principle cause of the dementia syndrome and it is estimated that AD is involved in 70% of dementia cases. AD prevalence is high in the over 60 years old population. This elevated prevalence is associated with an increasing number of elderly people. AD is therefore a major public health concern. AD is characterized by two types of specific cerebral lesions: neurofibrillary tangles (NFT) and amyloid plaques. However, there is no consensus on the links between these two types of lesions. To date, their presence can only be evaluated by a post-mortem examination. AD being a progressive pathology, this examination cannot be used to fully characterize the dynamic processes involved in AD. In this context, the development of non invasive imaging techniques to monitor the lesions progression in vivo could be determinant in AD pathophysiology understanding. Our objective is to develop new tracers of amyloid and neurofibrillary lesions for nuclear imaging. The first part of this study was dedicated to the validatation of an AD animal model: Transgenic 3xTgAD mice. The second part of this thesis focuses on the appreciatiation of the biological comportement of several known radiotracers of AD on this animal model. In the third part of this work, we initiate the development of several new tracers of AD-specific lesions. The para-sulfonato- calixarenes and the COB compounds for amyloid plaques detection and a peptide named A93, associated to a vector for the study of neurofibrillary tangles.

The prevalence of obesity is growing worldwide partly due to an increase in consumption of diets high in fat. Obesity is known as a risk factor for developing Alzheimer’s disease (AD) later in life. Both obesity and AD are associated with cognitive deficits and experimental high-fat diets can impair memory in cognitively normal rodents but also worsen memory deficits in AD mouse models. What is still unclear is the molecular mechanisms behind the detrimental effects of a high-fat diet on memory and if sex can influence its effect. Data in this thesis demonstrated that compared to females, male control non-transgenic (Non-Tg) mice had earlier deficits in memory after a high-fat diet that were associated with hyperinsulinemia. However, female Non-Tg mice were more vulnerable to ultrastructural changes in mitochondria morphology and loss of synapses after 6 months of a high-fat diet, changes that were similar to those observed in control-fed female triple-transgenic mice (3xTgAD). Finally, the memory deficits observed after a high-fat diet in cognitively normal mice were not associated with obesity and adiposity, as treatment with resveratrol (RSV) an anti-obesogenic compound, attenuated body weight gain and adipose tissue but failed to reverse memory impairment. In control fed 3xTgAD mice, RSV rescued memory deficits. In all experiments a high-fat diet had no detectable effect on cognitive impairment in 3xTgAD mice. In conclusion, the present thesis demonstrates that the sex-dependent differences in the effect a high-fat diet on memory are likely due to hyperinsulinemia and mitochondrial impairment and do not depend on obesity phenotype. These results demonstrate the importance of gender when studying both obesity and AD and are relevant for future clinical trials.