<p>The ß-carbolines norharman and harman structurally resemble the synthetic compound 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) that is known for its ability to damage neuromelanin-containing dopaminergic neurons of the substantia nigra and thereby induce parkinsonism. MPTP is, however, not normally present in the environment whereas the ß-carbolines are present in cooked food and tobacco smoke. </p><p>In this thesis it was demonstrated that norharman and harman had affinity to melanin and were retained in neuromelanin-containing neurons of frogs up to 30 days post-injection (the longest survival time examined). It was also demonstrated that norharman induced neurodegeneration, activation of glia cells and motor impairment in mice. Furthermore, this compound induced ER stress and cell death in PC12 cells. An in vitro model of dopamine melanin-loaded PC12 cells was developed in order to study the effect of melanin on norharman-induced toxicity. In this model, melanin seemed to attenuate toxicity induced by low concentrations of norharman. After exposure to the highest concentration of norharman, melanin clusters were disaggregated and there was an increased expression of stress proteins and caspases-3, known to be involved in apoptosis.</p><p>The polycyclic aromatic hydrocarbon, 7,12-dimethylbenz[<i>a</i>]anthracene was demonstrated to have a CYP1A1-dependent localization in endothelial cells in the choroid plexus, in the veins in the leptomeninges and in the cerebral veins of mice pre-treated with CYP1-inducers. </p><p>These results demonstrate that the distribution of environmental compounds could be influenced by the presence of neuromelanin and expression of CYP enzymes in the brain and that norharman may induce neurotoxic effects in vivo and in vitro.</p>

The ß-carbolines norharman and harman structurally resemble the synthetic compound 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) that is known for its ability to damage neuromelanin-containing dopaminergic neurons of the substantia nigra and thereby induce parkinsonism. MPTP is, however, not normally present in the environment whereas the ß-carbolines are present in cooked food and tobacco smoke. In this thesis it was demonstrated that norharman and harman had affinity to melanin and were retained in neuromelanin-containing neurons of frogs up to 30 days post-injection (the longest survival time examined). It was also demonstrated that norharman induced neurodegeneration, activation of glia cells and motor impairment in mice. Furthermore, this compound induced ER stress and cell death in PC12 cells. An in vitro model of dopamine melanin-loaded PC12 cells was developed in order to study the effect of melanin on norharman-induced toxicity. In this model, melanin seemed to attenuate toxicity induced by low concentrations of norharman. After exposure to the highest concentration of norharman, melanin clusters were disaggregated and there was an increased expression of stress proteins and caspases-3, known to be involved in apoptosis. The polycyclic aromatic hydrocarbon, 7,12-dimethylbenz[a]anthracene was demonstrated to have a CYP1A1-dependent localization in endothelial cells in the choroid plexus, in the veins in the leptomeninges and in the cerebral veins of mice pre-treated with CYP1-inducers. These results demonstrate that the distribution of environmental compounds could be influenced by the presence of neuromelanin and expression of CYP enzymes in the brain and that norharman may induce neurotoxic effects in vivo and in vitro.

Le cholestérol est le principal stérol présent dans la rétine. Dans sa forme libre, le cholestérol est distribué dans toutes les couches cellulaires de la rétine, alors que le cholestérol estérifié s’accumule essentiellement à la base de l’épithélium pigmentaire rétinien. La capacité intrinsèque de la rétine à synthétiser le cholestérol paraît limitée, ce qui implique nécessairement que des voies extra-rétiniennes participent activement à suppléer la rétine en cholestérol. Les cellules gliales de Müller contribueraient à l’apport de cholestérol aux neurones de la rétine, en particulier pour la formation des synapses. Les conséquences délétères d’une accumulation ou à l’inverse d’un déficit en cholestérol dans les neurones sur leur survie souligne l’importance de maintenir l’équilibre entre l’apport et la néosynthèse du cholestérol d’une part et son élimination d’autre part. Pour cela, la rétine neurale a en particulier la capacité de convertir, pour l’éliminer, le cholestérol en 24S-hydroxycholestérol. En effet, le transport du 24S-hydroxycholestérol au travers des membranes est facilité par la présence d’un groupe hydroxyle supplémentaire, lui conférant une polarité plus importante par rapport au cholestérol. L’enzyme qui catalyse cette réaction est la cholestérol-24S-hydroxylase (CYP46A1). Des liens ont été établis entre CYP46A1, 24S-hydroxycholestérol et processus neurodégénératifs dans le cerveau, suggérant un rôle potentiel dans certaines pathologies comme la maladie d’Alzheimer. CYP46A1 est exprimée dans la rétine neurale, et plus particulièrement dans les cellules ganglionnaires de la rétine. Le rôle de CYP46A1 dans la rétine reste pour l’instant inconnu. Cependant, par analogie avec le cerveau, nous pouvons supposer une fonction dans le contrôle de l’homéostasie du cholestérol dans les neurones et envisager une association avec des pathologies dégénératives de la rétine comme la Dégénérescence Maculaire Liée à l’Âge (DMLA) ou le glaucome. Dans ce contexte, l’objectif de nos travaux a consisté à évaluer le rôle de la cholestérol-24S-hydroxylase dans la rétine en conditions physiologiques et pathologiques. Par une approche clinique, nous avons trouvé qu’un polymorphisme génétique dans CYP46A1 était un facteur de risque de glaucome (Risque relatif=1,26, intervalle de confiance à 95%=1,006-1,574, p<0,05) (Fourgeux et al. 2009, Invest Ophthalmol Vis Sci 50:5712-7). Par contre, ce polymorphisme génétique n’a pas été retrouvé, en tant que tel, comme facteur de risque chez des patients DMLA, mais pourrait l’être chez les patients non porteurs d’allèles à risque dans les gènes CFH et LOC388715 (Fourgeux et al. 2012, Invest Ophthalmol Vis Sci 53:7026-33). Deux approches expérimentales nous ont permis de suggérer qu’il existe un lien entre le stress des cellules de la rétine et le 24S-hydroxycholestérol. En effet, dans une étude in vivo faite chez le rat, après avoir reproduit une caractéristique principale du glaucome par l’augmentation de la pression intraoculaire, nous avons suggéré le rôle crucial de la glie dans le maintien de l’expression de CYP46A1 au cours de la neurodégénérescence de la rétine (Fourgeux et al. 2012, Acta Ophthalmol, Sep 23 ; doi: 10.1111/j.1755-3768.2012.02490.x.). Enfin, l’inhibition pharmacologique de l’activité CYP46A1 dans la rétine par le voriconazole injecté in vivo chez le rat nous a permis de mettre en évidence que la diminution du contenu en 24S-hydroxycholestérol de la rétine était associée à une dysfonction des cellules ganglionnaires, évaluée par électrorétinographie. En parallèle, nous avons observé une activation gliale, dont l’amplitude était amplifiée par l’inhibition de l’activation microgliale induite par la minocycline [...] / Cholesterol is the major sterol found in the retina. In its free form, cholesterol is present in all cell layers of the retina, whereas cholesteryl esters mainly accumulate at the basement of the retinal pigment epithelium. The intrinsic capacity of the retina to synthetize cholesterol appears limited. Some extra-retinal pathways actively participate to cholesterol uptake to the retina. Müller glial cells may contribute to cholesterol supply to retinal neurons, especially for synaptic formation. Cholesterol accumulation or conversely deficiency have deleterious consequences on neuron survival. Maintaining the equilibrium between cholesterol supply and neosynthesis in the one hand and cholesterol elimination in the other hand is crucial. For that purpose, the inner retina converts cholesterol into 24S-hydroxycholesterol. The transport of 24S-hydroxycholesterol across membranes is facilitated by the addition of the hydroxyle group to cholesterol at position 24 of carbon chain since it renders cholesterol more hydrophilic. CYP46A1 (cholesterol 24S-hydroxylase) is the enzyme which catalyzes this reaction. Some links between CYP46A1, 24S-hydroxycholesterol and neurodegenerative processes have been reported in the brain, suggesting a potential role in several pathologies such as Alzheimer’s disease. CYP46A1 is expressed in the neural retina and specifically in retinal ganglion cells. The contribution of CYP46A1 in the retina remains unknown. Moreover by analogy with the brain, we can suggest a function for CYP46A1 in the regulation of cholesterol homeostasis in retinal neurons. Possible associations between CYP46A1 and Age-related Macular Degeneration (AMD) and glaucoma were suspected. In this context, we aimed to evaluate the role of CYP46A1 in the retina in physiological and pathological conditions. Through a clinical approach, we found that a genetic polymorphism in CYP46A1 was a risk factor for glaucoma (Odd Ratio = 1.26 ; 95% CI=1.006-1.574, p<0.05) (Fourgeux et al. 2009, Invest Ophthalmol Vis Sci 50:5712-7). By contrast, this genetic polymorphism was not found as a risk factor in AMD patients, but may become an additional risk factor in patients who do not carry risk allele in CFH and LOC387715 genes (Fourgeux et al. 2012, Invest Ophthalmol Vis Sci 53:7026-33). Two experimental approaches suggested that a link between retinal stress and 24S-hydroxycholesterol does exist. Indeed, in a rat model of glaucoma of elevated intraocular pressure, we suggested the crucial role of CYP46A1 in maintaining CYP46A1 expression in the course of retinal neurodegeneration (Fourgeux et al. 2012, Acta Ophthalmol, Sep 23; doi: 10.1111/j.1755-3768.2012.02490.x.). Pharmacological inhibition of CYP46A1 activity in the retina by voriconazole administered in vivo in the rat highlighted that the decrease in retinal 24S-hydroxycholesterol levels was associated with RGC dysfunction evaluated by electroretinography. In parallel, we observed glial activation in which magnitude was exacerbated when microglia activation was inhibited by minocycline at the same time.In conclusion, by a dual clinical and experimental approach, our works suggest a crucial role for CYP46A1 in maintaining cholesterol homeostasis in the retina in physiological and pathological conditions. Müller glial cell intervention in this process may be suspected especially in pathological conditions of glaucoma

<p>Microglia are a population of specialized, tissue-resident immune cells that make up around 10% of total cells in our brain. They actively prune neuronal synapses, engulf cellular debris, and misfolded protein aggregates such as the Alzheimer’s Disease (AD)-associated amyloid-beta (Aβ) by the process of phagocytosis. During AD, microglia are unable to phagocytose Aβ, perhaps due to the several disease-associated changes affecting their normal function. Functional molecules such as lipids and metabolites also influence microglial behavior but have primarily remained uncharacterized to date. The overarching question of this work is, <i>How do microglia become dysfunctional in chronic inflammation</i>? To this end, we developed new chemical tools to better understand and investigate the microglial response to Aβ <i>in vitro</i> and <i>in vivo</i>. Specifically, we introduce three new tools. (1) Recombinant human Aβ was developed via a rapid, refined, and robust method for expressing, purifying, and characterizing the protein. (2) A pH-sensitive fluorophore conjugate of Aβ (called Aβ^pH) was developed to identify and separate Aβ-specific phagocytic and non-phagocytic glial cells <i>ex vivo</i> and <i>in vivo</i>. (3) New lysosomal, mitochondrial, and nuclei-targeting pH-activable fluorescent probes (called LysoShine, MitoShine, and NucShine, respectively) to visualize subcellular organelles in live microglia. Next, we asked, <i>What changes occur to the global lipid and metabolite profiles of microglia in the presence of Aβ in vitro and in vivo</i>? We screened 1500 lipids comprising 10 lipid classes and 700 metabolites in microglia exposed to Aβ. We found significant changes in specific lipid classes with acute and prolonged Aβ exposure. We also identified a lipid-related protein that was differentially regulated due to Aβ <i>in vivo</i>. This new lipid reprogramming mechanism “turned on” in the presence of cellular stress was also present in microglia in the brains of the 5xFAD mouse model, suggesting a generic response to inflammation and toxicity. It is well known that activated microglia induce reactive astrocytes during inflammation. Therefore, we asked, <i>What changes in proteins, lipids, and metabolites occur in astrocytes due to their reactive state? </i>We provide a comprehensive characterization of reactive astrocytes comprising 3660 proteins, 1500 lipids, and 700 metabolites. These microglia and astrocytes datasets will be available to the scientific community as a web application. We propose a final model wherein the molecules secreted by reactive astrocytes may also induce lipid-related changes to the microglial cell state in inflammation. In conclusion, this thesis highlights chemical neuroimmunology as the new frontier of neuroscience propelled by the development of new chemical tools and techniques to characterize glial cell states and function in neurodegeneration.</p>