Search

In the central nervous system, astrocytes are glial cells displaying crucial roles in the brain physiology by regulating synaptic transmission, ion homeostasis, the blood brain barrier integrity or the cerebral blood flow. These functions are sustained by morphological and molecular polarities of astrocytes with specific interfaces contacting synapses, the perisynaptic processes (PAP), and blood vessels, the perivascular processes (PvAP). Recently, my laboratory and others demonstrated the presence of mRNAs, ribosomes and local protein synthesis in PAPs and PvAPs and proposed local translation as a mechanism for setting astrocytic polarity. However, little is known about the mechanisms of translation in astrocytes. My PhD project aimed at deciphering how translation is regulated in astrocytes and their interfaces and the role of these mechanisms in brain physiology. First, to characterize the distribution of mRNAs in astrocytes, I developed AstroDot, a specific in-situ approach to quantify the density and distribution of mRNAs in astrocytes. I determined that in hippocampal astrocytes, the distribution and density of mRNAs encoding α and δ isoforms of the glial fibrillary acidic protein (GFAP) are altered and vary with the presence of amyloid plaques in APP/PS1dE9 mice, a mouse model of Alzheimer’s disease. I also demonstrated for the first time the presence of mRNAs in microglia processes, a resident glial cell involved in particular in brain immunity. Following this study, I aimed at identifying mechanisms setting translation in astrocytes and focused on proteins associated with polysomes in astrocytes. I developed a method combining translating ribosome affinity purification in astrocytes with mass spectrometry. Among immunoprecipitated proteins, I focused on RACK1, a scaffold protein known to associate with the small subunit of ribosomes. Our previous analysis of the translatome in PAPs identified Gnb2l1 encoding RACK1 as one of the most enriched translated mRNAs in PAPs, suggesting its possible role in the regulation of translation at this interface. RACK1 acts as a ribosome formation mediator, a translation regulator and takes part in RNA quality control. Its role in astrocytes was unknown. Using a restricted panel of astrocytic enriched or specific mRNAs, I showed that RACK1 preferentially associates with Kncj10 coding for the potassium channel KIR4.1, one of the main regulator of K+ homeostasis in PAPs and PvAPs, and Slc1a2 coding for a glutamate transporter GLT1 which allows glutamate recycling in PAPs. To further address the role of RACK1 in astrocytes, I developed a mouse model in which RACK1 is specifically inactivated in astrocytes (RACK1 cKO). In this model, I found the level of KIR4.1 upregulated in the brain and in PAPs, whereas GLT1 was unchanged. RACK1 can act on the ribosome and on the RNA levels to regulate translation. In vitro investigations on RACK1 KO HEK cells transfected with luciferase reporters showed that KIR4.1 upregulation does not involve ribosome stalling on the Kncj10 coding sequence but relies on a part of the 5’UTR of Kcnj10 (Collaboration with Clément Chapat, Polytechnique). We finally addressed the physiological consequences of these regulations. Performing electrophysiological recordings (Collaboration with Nathalie Rouach, CIRB), we determined that KIR4.1 increase in absence of RACK1 leads to changes in the electrophysiological properties of the hippocampal neurons with an overall protection against high synaptic activity, certainly due to a higher capacity of PAPs to uptake K+. In summary, I developed during my thesis a new method to characterize mRNA distribution in astrocytes. I identified a pool of proteins associated to polysomes in astrocytes. Finally, I focused on RACK1 and identified it as an important regulator of KIR4.1 translation and astrocytic perisynaptic K+ homeostasis., Dans le système nerveux central, les astrocytes, cellules gliales, jouent des rôles cruciaux dans la physiologie du cerveau en régulant la transmission synaptique, l’homéostasie ionique, l’intégrité de la barrière sang-cerveau et le flux sanguin cérébral. Ces fonctions sont soutenues par des polarités morphologique et moléculaire des astrocytes avec des interfaces contactant les synapses, avec les prolongements périsynaptiques (PAP), et les vaisseaux sanguins, avec les prolongements périvasculaires (PvAP). Nous et d’autres avons démontré la présence d’ARNm, de ribosomes et de synthèse protéique locale dans les PAPs et PvAPs et proposé que la traduction locale soit un mécanisme pour la mise en place de la polarité astrocytaire. Cependant, on connaît peu les mécanismes de traduction dans les astrocytes. Mon projet de thèse a eu pour but de décrypter comment la traduction dans les astrocytes est régulée et le quels sont les rôles de ces mécanismes dans la physiologie du cerveau. Nous avons d’abord caractérisé la distribution des ARNm dans les astrocytes en développant AstroDot, une approche in situ pour quantifier la densité et la distribution des ARNm. Nous avons déterminé dans les astrocytes de l’hippocampe que la distribution et la densité d’ARNm codant pour les isoformes alpha et delta de la GFAP étaient altérées et variaient avec la présence de plaques amyloïdes dans un modèle murin de la maladie d’Alzheimer. Nous avons aussi montré pour la 1ère fois la présence d’ARNm dans les prolongements de microglies, une cellule gliale impliquée dans l’immunité cérébrale. Après, j’ai voulu identifier les mécanismes de mise en place de la traduction dans les astrocytes et je me suis concentré sur les protéines associées aux polysomes de cette cellule. J’ai développé une méthode combinant le TRAP avec la spectrométrie de masse. Parmi les protéines immunoprécipitées, je me suis concentré sur RACK1, une protéine d’échafaudage connue pour s’associer avec la petite sous unité des ribosomes. Notre analyse précédente du translatome des PAPs a identifié Gnb2l1, codant pour RACK1, comme l’un des ARNm traduit les plus enrichis dans les PAPs suggérant son rôle dans la régulation de la traduction à cette interface. RACK1 agit comme un régulateur de la formation des ribosomes, de la traduction et prend part au contrôle qualité des ARNm. Son rôle dans les astrocytes était inconnu. En utilisant un panel restreint d’ARN astrocytaires spécifiques, j’ai montré que RACK1 était préférentiellement associé à Kcnj10 codant pour le canal potassique KIR4.1 et Slc1a2 codant pour le transporteur au glutamate GLT1. Pour adresser le rôle de RACK1 dans les astrocytes, j’ai développé un modèle de souris dans laquelle RACK1 est inactivé dans les astrocytes. Dans ce modèle, le niveau de KIR4.1 était augmenté dans le cerveau et dans les PAPs alors que GLT1 était inchangé. RACK1 peut agir aux niveaux du ribosome et de l’ARN pour réguler la traduction. Des expériences in vitro sur des cellules HEK RACK1 KO transfectées avec des rapporteurs luciférase ont montré que l’augmentation de KIR4.1 n’impliquait pas le blocage des ribosomes sur la séquence codante de son ARNm kcnj10 mais était lié à sa partie 5’ non codante. Nous avons adressé les conséquences physiologiques de ces régulations. Par enregistrements électrophysiologiques, nous avons déterminé que l’augmentation de KIR4.1 en l’absence de RACK1 menait à un changement dans les propriétés électriques des neurones de l’hippocampe avec une protection globale contre des activités synaptiques élevées certainement due à une plus grande capacité des PAPs à capturer le potassium. En résumé, j’ai développé une nouvelle méthode pour caractériser la distribution des ARNm dans les astrocytes. J’ai identifié un pool de protéines associées aux polyribosomes astrocytaires. Enfin, Je me suis concentré sur RACK1 et je l’ai identifié comme un régulateur de la traduction de KIR4.1 et de l’homéostasie du potassium périsynaptique.

International audience; Together with the compartmentalization of mRNAs in distal regions of the cytoplasm, local translation constitutes a prominent, evolutionarily conserved mechanism for cellular polarization and the regulation of protein delivery in space and time. The translational regulation of gene expression enables a rapid response to stimuli or to a change in the environment, since the use of pre-existing mRNAs can bypass time-consuming nuclear control mechanisms. In the brain, the translation of distally-localized mRNAs has been mainly studied in neurons, whose cytoplasmic protrusions may be more than 1000 times longer than the diameter of the cell body. Importantly, alterations in local translation in neurons have been implicated in several neurological diseases. Astrocytes, the most abundant glial cells in the brain, are voluminous, highly ramified, project long processes to neurons and brain vessels, and dynamically regulate distal synaptic and vascular functions. Recent research has demonstrated the presence of local translation at these astrocytic interfaces that might regulate the functional compartmentalization of astrocytes. In this Review, we summarize our current knowledge about the localization and local translation of mRNAs in the distal perisynaptic and perivascular processes of astrocytes and discuss their possible contribution to the molecular and functional polarity of astrocytes.

Cells with a complex shape often use mRNA distribution and local translation to regulate distal functions. These mechanisms have recently been described in astrocytes, the processes of which contact and functionally modulate neighbouring synapses and blood vessels. In order to study the distribution of mRNA in astrocytes, we developed a three-dimensional histological method that combines mRNA detection viain situhybridization with immunostaining of the astrocyte-specific intermediate filament glial fibrillary acidic protein (GFAP). Three-dimensional confocal images were analyzed using AstroDot, a custom Image J plug-in developed in-house for the identification and quantification of mRNAs in GFAP-immunolabelled astrocyte somata, large processes and fine processes. The custom R package AstroStat was used to analyze the AstroDot results. Taking the characterization of mRNAs encoding the astrocyte-specific GFAP α and δ isoforms in the hippocampus as a proof of concept, we showed thatGfapα andGfapδ mRNAs mainly colocalized with GFAP in astrocyte processes.Gfapα mRNA was more abundant thanGfapδ mRNA, and was predominantly found in fine processes. Upon glial activation in the APPswe/PS1dE9 mouse model of Alzheimer’s disease, the same overall patterns were found but we noted strong variations inGfapα andGfapδ mRNA density and distribution as a function of the part of the hippocampus and the astrocyte’s proximity to beta-amyloid (Aβ) plaques. In astrocytes not associated with Aβ, Gfap α mRNA levels were only slightly elevated, and Gfap δ mRNA was distributed within the fine processes; these effects were more prominent in CA3 than in CA1. In contrast, levels of both mRNAs were markedly elevated in the fine processes of Aβ-associated astrocytes in both CA1 and CA3. In order to validate our new method, we confirmed thatRpl4mRNA (a ubiquitously expressed mRNA encoding the large subunit ribosomal protein 4) was present in large and fine processes in both astrocytes and microglia. In summary, we have developed a novel, reliable set of tools for characterizing mRNA densities and distributions in the somata and processes of astrocytes and microglia in physiological or pathological settings. Furthermore, our results suggest that intermediate filaments are crucial for distributing mRNA within astrocytes and for modulating specificGfapmRNA profiles in Alzheimer’s disease.