Au vu de la relative inefficacité des traitements actuels de la douleur cancéreuse osseuse (DCO) il est devenu nécessaire aujourd'hui d'identifier de nouvelles cibles (cellulaires et/ou moléculaires) pour développer de nouveaux outils thérapeutiques. Dans ce contexte, ces dernières années, de nombreuses études ont suggéré que les cellules gliales, principalement les astrocytes et la microglie, pourraient contribuer au développement et au maintien de la douleur chronique. D'autre part, dans des modèles d'études précliniques de la DCO, plusieurs auteurs ont récemment constaté une réactivité astrocytaire importante dans les cornes dorsales de la moelle épinière et ont montré que, si on empêche cette réactivité, les symptômes douloureux sont diminués. Cependant, la relation exacte existant entre la réactivité des cellules gliales et les symptômes douloureux en condition de DCO est inconnue. Afin de décrypter cette relation, nous avons dans un premier temps étudié le décours temporel des comportements douloureux et caractérisé l’état de sensibilisation centrale dans un modèle de DCO chez le rat induit par l'injection de cellules de carcinome glandulaire mammaire (MRMT-1) dans le tibia. Nous montrons par des approches radiologiques, comportementales (tests de douleur évoquée et de distribution pondérale dynamique) et immunohistochimiques (immunodétection de la protéine Fos après palpation non douloureuse de la patte) que les animaux cancéreux MRMT développent graduellement une tumeur osseuse (premiers signes au 10ème jour post-inoculation), une allodynie et une hyperalgésie mécaniques (à partir du 10ème jour) et thermiques (à partir du 14ème jour), un inconfort de la patte injectée (à partir du 14ème jour ) et des phénomènes de sensibilisation centrale. Dans un deuxième temps, nous avons recherché des indices structuraux et fonctionnels de réactivité gliale spinale dans notre modèle de DCO. L'objectif était donc de dater l'apparition de la réactivité gliale, et de déterminer la nature des cellules gliales impliquées : microglie et/ou astrocytes. Nous montrons par des approches immunohistochimiques qu’aucun signe morphologique de réactivité astrocytaire ni microgliale n’est observable pendant l’établissement et le maintien de la DCO alors que ces signes existent dans un modèle de douleur neuropathique (ligature de nerfs spinaux). De plus, par des approches moléculaire (qRT-PCR) et biochimique (technique du Bio-Plex) nous montrons que, parmi les 20 marqueurs structuraux et fonctionnels de réactivité gliale testés, seule l’expression de l’aquaporine 4 (un canal à eau spécifique des astrocytes) est significativement augmentée en condition de DCO. Nos résultats suggèrent donc que les astrocytes et les cellules microgliales jouent des rôles différents dans la douleur cancéreuse et dans la douleur neuropathique. Enfin, dans un troisième temps, nous avons cherché à mettre en évidence une implication des astrocytes dans la pathologie DCO au travers d’une caractérisation des transmissions glutamatergique et glycinergique, qui sont toutes deux fortement modulées par l’environnement astrocytaire. Par la quantification de l’expression de l’ARNm (qRT-PCR) et par dosage des taux d’acides aminés (électrophorèse capillaire), nous montrons que les principaux acteurs (transporteurs, récepteurs, agonistes et co-agonistes) de la transmission glutamatergique et glycinergique spinale ne subissent pas d’altération significative en condition de DCO. En conclusion, nous montrons que des symptômes douloureux chroniques peuvent se développer et se maintenir (1) sans signe d’astrogliose et de réactivité microgliale spinale ; et (2) sans altération de l’expression des principaux acteurs de la transmission spinale glutamatergique et glycinergique. Nos résultats invitent donc à revoir le lien très fort qui est fait actuellement entre douleur chronique et astrogliose. / The relative lack of efficiency of current treatments used to relieve bone cancer pain prompts to the identification of new molecular and/or cellular targets for the development of new therapeutic strategies. In that context, a large number of recent studies have suggested the involvement of glial cells, among which astrocytes and microglial cells, in the onset and maintenance of chronic pain symptoms. In few animal models of bone cancer pain, several authors have recently evidenced an increased glial reactivity in spinal cord dorsal horn, and demonstrated that preventing astrocytic reactivity was sufficient to reduce pain symptoms in these models. However, the exact relationship of glial reactivity with bone cancer pain symptoms remains poorly understood. In order to decipher this link, we have first studied the temporal development of pain symptoms, and characterized the degree of central sensitization in a rat model of bone cancer pain induced by the injection of mammary gland carcinoma cells (MRMT-1) in the tibial bone. Using radiologic assessment of tumor development, behavioral measurements to quantify evoked (von Frey hairs) and spontaneous (dynamic weight bearing) pain and immunodetection of Fos after non nociceptive palpation of cancer bearing limb, we demonstrate that animals injected with MRMT-1 cells gradually develop a bone tumor (first detectable 10 days after inoculation), a mechanical allodynia and hyperalgesia (first noticeable at day 10), and later on a thermal allodynia and hyperalgesia (first detectable at day 14) as well as discomfort of the injected limb (day 14) and finally central sensitization phenomenons. Second, we have investigated the presence of structural and functional markers of spinal glial reactivity in our model of bone cancer pain. Our objectives were to date the onset of spinal glial reactivity, for microglial and astrocytic cells. Using immunohistochemical approaches, we show that none of the classical markers of astrocytic and microglial reactivity can be observed during the onset and the persistent phase of bone cancer pain whereas the markerswere easily identified in a neuropathic pain model (spinal nerve ligation). Furthermore, using molecular (qRT-PCR) as well as biochemical (Bio-Plex) approaches, we show that among the 20 structural and functional markers of glial reactivity tested, only aquaporin-4 displays increased mRNA levels in bone cancer pain model. Hence, our results suggest that astrocytes and microglial cells play different roles in bone cancer and neuropathic pain. Finally, we tried to evidence the involvement of astrocytes in bone cancer pain by characterizing glutamatergic and glycinergic synaptic transmission, both of which are heavily modulated by astrocytic environment. By quantifying mRNA levels (qRT-PCR) and measuring the level of inhibitory and excitatory amino acids (capillary electrophoresis), we show that the main actors (transporters, receptors, agonists and co-agonists) of glutamatergic and glycinergic transmissions in the spinal cord do not undergo any significant alteration in bone cancer pain conditions. We conclude that chronic painful symptoms may develop and persist (1) without any sign of astrogliosis or enhanced microglial reactivity in the spinal cord, and (2) without any alteration in the expression/levels of the main actors involved in glutamatergic and glycinergic transmission. These results therefore question the strong link that is frequently made between astrogliosis and chronic pain.

La Neuromyélite Optique (NMO) est une maladie auto-immune démyélinisante, rare et grave, du système nerveux central (SNC). Elle est caractérisée par une démyélinisation et une perte axonale ciblant principalement le nerf optique et la moelle épinière. La découverte d'un auto-anticorps (IgG-NMO) dirigé contre l'aquaporine-4 (AQP4), un canal hydrique exprimé par l'astrocyte, a été une étape clé dans la compréhension de la physiopathologie de la NMO, actuellement définie comme une astrocytopathie. La pathogénicité de l'IgG-NMO a été démontrée : il induit une internalisation d'AQP4 et des transporteurs au glutamate, provoquant une altération de la fonction astrocytaire. Cependant les mécanismes permettant de lier la dysfonction astrocytaire aux altérations caractéristiques de la NMO, notamment la démyélinisation, restent méconnus. Les astrocytes sont des cellules gliales essentielles à l'établissement et au maintien de l'homéostasie du SNC. Ils permettent la régulation des flux hydriques et ioniques, le contrôle extracellulaire des neuromédiateurs ainsi que l'apport de métabolites énergétiques aux neurones et aux oligodendrocytes. Ils sont aussi caractérisés par une très forte expression de connexines (Cx), des molécules transmembranaires s'assemblant sous une forme hexamérique : le connexon. Les connexines forment soit des hémicanaux, permettant l'échange de petites molécules entre les milieux intra- et extra-cellulaires, soit des jonctions communicantes par la juxtaposition de connexons appartenant à deux cellules, assurant le couplage intercellulaire avec le passage de petites molécules et d'ions (ATP, glutamate, lactate, calcium). Les fonctions hemicanal et jonction communicante sont fortement régulées en condition physiologique et altérées en condition pathologique, notamment en contexte neuroinflammatoire. Nous émettons l'hypothèse que les IgG-NMO altèrent l'expression et la fonction des connexines, et conduisent ainsi à la production d'un environnement toxique pour les oligodendrocytes et la myéline, et délétère pour le fonctionnement neuronal. Mon projet de thèse avait trois objectifs : i) la caractérisation du phénotype astrocytaire induit par les IgGNMO ; ii) l'identification d'altérations des connexines et leur implication dans la pathologie ; iii) la mise en évidence d'altérations de la transmission synaptique induites par les IgG-NMO et l'implication de connexines dans cet effet. Des modèles de cultures primaires gliales traitées par des IgG-NMO issue d'une cohorte de patients m'ont permis de caractériser le phénotype acquis par les astrocytes, et de proposer le concept d'un astrocyte réactif spécifique de pathologie. Les astrocytes réactifs spécifiques de la NMO induisent un milieu inflammatoire spécifique et toxique, provoquant une démyélinisation. Grâce au développement d'une coculture gliale et neuronale produisant des neurones myélinisés, et à l'utilisation de peptides inhibiteurs des Cx, j'ai pu montrer que les NMO-IgG ont un effet démyélinisant et que celui-ci implique les Cx. La démyélinisation est en effet associée à des modifications structurales et fonctionnelles des Cx astrocytaires, observées à la fois in vitro et dans notre modèle in vivo, le rat-NMO. Enfin, la mise en place d'une étude électrophysiologique en potentiel de champs local sur des tranches d'hippocampe de rats m'a permis d'étudier l'effet des IgG-NMO sur la transmission glutamatergique basale. J'ai pu mettre en évidence un effet dépresseur des IgG-NMO, partiellement bloqué par un inhibiteur de connexines, la carbenoxolone. Comme il a déjà été démontré par des études cliniques dans des pathologies neurodégénératives, l'utilisation de modulateurs de Cx semble être une voie thérapeutique prometteuse afin de prévenir la démyélinisation et les altérations du fonctionnement neuronal de la NMO / Neuromyelitis Optica (NMO) is a rare and severe auto-immune demyelinating disease of the central nervous system (CNS). It is characterized by demyelination and axonal loss targeted to the optic nerve dans the spinal cord. The identification of a specific autoantibody (NMO-IgG) directed against the astrocytic protein AQP4 was a key step in the understanding of NMO physiopathology: it is now considered as an astrocytopathy. NMO-IgG is also a biomarker of NMO, and its pathogenicity has been demonstrated. NMO-IgG induce an internalization of AQP4 together with other membrane proteins such à glutamate transport GLT1. This could alter astrocyte functions but the mechanisms connecting astrocytopathy and demyelination remain unclear. Astrocytes are abundant glial cells crucial for the establishment and the maintenance of CNS homeostasis. They regulate water flux and ion homeostasis and control extracellular volume and neurotransmitter concentrations. They also provide neurons and oligodendrocytes with energy substrates. Astrocytes are characterized by a high expression of connexins (Cx), transmembrane proteins assembling in hexameric form, called connexon. Cx form either hemichannels, unopposed connexon at the membrane, allowing the exchange of small molecules (<1,2kDa e.g. glutamate, ATP) and ions (Ca2+, K+) between extra- and intra-cellular compartments. Cx also form gap junctions, formed by the juxtaposition of two connexons at the membrane of two different cells, and allow the quick cell to cell exchange of small molecules, metabolites and ions (e.g. glucose, lactate, Ca2+). Hemichannel and gap junction functions are tightly regulated under physiological conditions and can be altered in pathological condition for example during neuroinflammation. We proposed that NMO-IgG by altering connexins expression and/or function could lead to the production of a toxic environment for oligodendrocytes and myelin but also for neuronal functioning. This feature of astrocyte dysfunction could participate to NMO alterations. My thesis project had three main goals: i) the characterisation of astrocyte phenotype induced buy NMO-IgG, ii) the identification of connexins alterations and their implication NMO physiopathology, iii) the highlight of synaptic alterations induced by NMO-IgG and the involvement of connexins in this effect. Primary glial cell cultures treated with NMO-IgG from a cohort of NMO patients, were used to characterize astrocyte phenotype and we proposed the concept of a specific reactive dysfunctioning astrocyte induced by NMO-IgG. Those astrocytes, called “NMO-astrocytes” are responsible for the production of a proinflammatory toxic microenvironment for oligodendrocytes and leading to demyelination. With the development of a myelinated culture model, composed of glial cells and neurons with myelinated axons, together with the use of specific inhibitors of Cx functions, we showed that NMO-IgG induced demyelination involved connexin dysfunction. In fact, demyelination was associated with structural and functional alterations of astrocytic connexins observed both in vitro and in vivo in the NMO-rat model. Electrophysiological recording of basal glutamatergic synaptic activity in the rat hippocampus showed a strong depression of synaptic responses induced by NMO-IgG. Connexins could be implicated in this alteration since blocking all connexins with carbenoxolone blocked NMO-IgG effect

Un statut pauvre en Acides Gras Poly-Insaturés ω3 (AGPI ω3), favorisé par une alimentation occidentale comportant un faible ratio en ω3/ω6, semble contribuer au déclin cognitif chez les personnes âgées, mais les mécanismes cellulaires impactés sont encore mal connus. Nous avons donc étudié l’influence du statut en ω3 sur l’évolution de la neurotransmission glutamatergique et des fonctions astrocytaires au cours du vieillissement dans l’hippocampe de rats. Ces processus sont impliqués dans la formation de la mémoire et leurs dérégulations participent aux dommages cérébraux conduisant au déclin cognitif. Nous avons comparé 6 groupes de rats agés de 6 et 22 mois nourris avec un régime déficient en ω3, équilibré en ω3/ω6 ou supplémenté en ω3 (huile de poisson) : Jeunes équilibrés (JEq), déficients (JDef) ou supplémentés (JSup) et Agés équilibrés (AEq), déficients (ADef) ou supplémentés (ASup). Nous avons évalué l’efficacité synaptique et la plasticité (enregistrements électrophysiologiques), les fonctions astrocytaires (capture de glutamate et expression de la GFAP), les marqueurs neuronaux (transporteurs et récepteurs du glutamate), les capacités cognitives (Openfield et Labyrinthe de Barnes) et analysé la composition lipidique cérébrale. Les manipulations nutritionnelles d’apport en ω3 modifient efficacement l’incorporation de l’acide docosahexaénoïque (DHA, principale ω3 des membranes cellulaires) dans le cerveau (-50% deficient vs équilibré, +10% supplementé vs équilibré). Le vieillissement induit une diminution de 35% de l’efficacité synaptique en raison d’une baisse de la libération de glutamate pré-synatique, et une diminution de 30% de la capture de glutamate associé à une astrogliose conséquente (+100% GFAP). La déficience en ω3 acentue les effets du vieillissement (rats ADef vs AEq: -35% efficacité synaptique, -15% capture de glutamate, +30% GFAP). Al’inverse, la supplémentation en ω3 améliore l’efficacité synaptique (rats ASup vs AEq +25%) et semble inhiber l’astrogliose chez le rat âgé (ASup vs JEq : pas de modification de la GFAP). Les tests comportementaux montrent que le vieillissement a des effets plus marqués chez les déficients en ω3 et au contraire atténués chez les supplémentés. Nos résultats révèlent des altérations de la synapse glutamatergique de l’hippocampe au cours du vieillissement aggravées par la déficience en ω3 et atténuées par la supplémentation en ω3. Afin d’évaluer l’influence du statut en ω3 sur l’activation astrocytaire, des modèles in vitro d’astrocytes « âgés » et « activés » par des cytokines inflammatoires dont l’augmentation à bas bruit est caractéristique du vieillissement cérébral, ont été développés. / A poor ω3 polyunsaturated fatty acids (ω3 PUFA) status, favored by the low ω3/ω6 ratio in western diets, seems to contribute to cognitive decline in the elderly, but mechanistic evidence is lacking. We therefore explored the impact of ω3 status on the evolution of glutamatergic transmission and astrocytic functions in the hippocampus during ageing in rats. These processes are involved in memory formation and their dysregulation participates to the age-related brain damage leading to cognitive decline. We have compared 6 groups of rats aged 6 to 22 months fed ω3-deficient, ω3/ω6-balanced, or ω3 (fish oil) supplemented diets: Young ω3 Balanced (YB), Deficient (YD) or Supplemented (YS), and Old ω3 Balanced (OB), Deficient (OD) or Supplemented (OS) rats. We have evaluated synaptic efficacy and plasticity (electrophysiological recording), astroglial regulations (glutamate uptake and GFAP expression), neuronal markers (glutamate transporters and receptors), cognitive abilities (Barnes maze and Openfield) and analyzed brain fatty acids composition. Dietary modulation of ω3 intakes efficiently modified the incorporation of docosahexaenoic acid (DHA, the main ω3 in cell membranes) in brain (-50% deficient vs balanced, +10% supplemented vs balanced). Ageing induced a 35% reduction of synaptic efficacy due to decreased pre-synaptic glutamate release, and a 30% decrease in the astroglial glutamate uptake associated to a marked astrogliosis (+100% GFAP). ω3 deficiency further decreased these hallmarks of ageing (OD vs OB rats: -35% synaptic efficacy, -15% glutamate uptake, +30% GFAP). On the opposite, ω3 supplementation increased synaptic efficacy (+25% OS vs OD) and seems to abolish astrogliosis (OS vs YS : no change in GFAP). Behavioural tests showed some increased effects of age in deficient rats and attenuated effects in supplemented ones. Our results characterize some specific age-related alterations of the glutamatergic synapse in the hippocampus that are aggravated by a dietary deficit in ω3 and attenuated by ω3 supplementation. In order to explore ω3 status on astrocytic activation, in vitro models of “old” astrocytes and “activated” by inflammatory cytokines which characterize the low-grade inflammation in brain aging, have been developed.

Le lactate est considéré comme un métabolite déchet depuis de très nombreuses années. Cependant, cette vision semble revisitée depuis quelques temps, avec l'apparition de la notion de glycolyse aérobie et de navettes lactate dans différents types cellulaires (muscle, cerveau et sperme). Concernant le cerveau, des études in vitro, ex vivo et in vivo réalisées ces 20 dernières années ont montré, d'une part, que les astrocytes produisent du lactate et d’autre part que le lactate pouvait être un substrat énergétique pour le système nerveux central (SNC), et plus particulièrement les neurones. Cette notion de navette lactate entre astrocyte et neurone a été proposée pour la première fois en 1994 par Pellerin et Magistretti (ANLS, pour astrocyte-neuron lactate shuttle). Malgré de nombreuses recherches depuis, l'existence d'un transfert net de lactate entre les astrocytes et les neurones n'a toujours pas pu être démontrée in vivo. Dans cet optique, la visualisation de la production de lactate in vivo dans le cerveau activé est essentielle. Le rôle des transporteurs au lactate, MCTs (Monocarboxylate Transporters), dans la détection de ce signal est également un point capital. L’objectif de cette thèse a été de développer la spectroscopie de RMN in vivo localisée dans le cortex somato-sensoriel du rat en condition d’activation cérébrale. Dans un premier temps, un travail de développement a été effectué afin de mettre au point le protocole de stimulation neuronale et d’obtenir un rapport signal sur bruit suffisant pour pouvoir quantifier de façon fiable le lactate. Une fois le protocole établi sur des rats contrôles, l’étude a été réalisée sur des rats modifiés génétiquement et réprimés pour le MCT, soit neuronal, soit astrocytaire. Le but était de déterminer si ce partenaire clef de l’ANLS avait une influence sur les fluctuations de lactate lors de l'activation cérébrale. En plus de la spectroscopie proton in vivo et de l’IRM fonctionnelle, des études de RMN du carbone-13 ont été réalisées ex vivo. Le résultat majeur de cette thèse montre qu’en l’absence du transporteur de lactate neuronal, non seulement on perd l’augmentation de lactate lors de la stimulation cérébrale mais on perd également le signal BOLD sur l’IRMf. Ce résultat suggère, et ce pour la première fois, que l’activité neuronale est fortement dépendante du transporteur au lactate. / Lactate has been considered as a waste metabolite for many years. However, this vision has been reconsidered recently, with the appearance of the notion of aerobic glycolysis and lactate shuttles in different cell types (muscle, brain, and sperm). Concerning the brain, in vitro, ex vivo and in vivo studies carried out over the last 20 years have shown, on the one hand, that astrocytes produce lactate and, on the other hand, that lactate can be an energetic substrate for the central nervous system (CNS), and more particularly neurons. This lactate shuttle between astrocyte and neuron was first proposed in 1994 by Pellerin and Magistretti (called ANLS, for astrocyte-neuron lactate shuttle). Despite many studies since then, the existence of a net transfer of lactate between astrocytes and neurons has still not been demonstrated in vivo. In this regard, visualization of lactate production in vivo in the activated brain is essential. The role of lactate transporters, MCTs (Monocarboxylate Transporters), in detecting this signal is also a key issue. The objective of this thesis was to develop in vivo NMR spectroscopy located in the somato-sensory cortex of rats under brain activation conditions. First, experiments were carried out to develop the neural stimulation protocol and to obtain a sufficient signal-to-noise ratio to be able to quantify lactate. Once the protocol was established on control rats, the study was performed on genetically modified rats and down-regulated for MCT, either neuronal or astrocytic. The aim was to determine whether this key partner of the ANLS has an influence on lactate fluctuations during brain activation. In addition to in vivo proton spectroscopy and functional MRI, carbon-13 NMR studies were performed ex vivo. The major result of this thesis shows that in the absence of the neuronal lactate transporter, not only is the increase in lactate lost during brain stimulation but the BOLD signal on the fMRI is also lost. This result suggests, for the first time, that neural activity is highly dependent on the lactate transporter.

The prostaglandins comprise a group of bioactive lipids generated from arachidonic acid by cyclooxygenases and cell type-specific prostaglandin and thromboxane synthases. Prostaglandins mediate local cell signaling interactions by activation of G-protein coupled prostanoid receptors. Because the prostaglandins and their receptors are active in all tissues, they have an extraordinarily broad spectrum of physiological and pathophysiological functions that have hampered the development of safe prostanoid-based medications. This situation has emphasized the importance of understanding the functional properties of the prostanoid receptors and developing selective ligands capable of being used in patient care.The aims of this project were to identify novel regulatory functions of endogenous EP and FP prostanoid receptors in cultured human cells. Our results show that activation of EP₂ receptors in human microglia and astrocytes led to increased secretion of BDNF, a growth factor that regulates the survival of neurons. In the same cell lines, FP receptors regulate the induction of TNF-&alpha; gene expression through a classic G_q-PKC pathway. In microglia these FP receptors also stimulate a novel signaling crosstalk mechanism involving the up-regulation of TCF transcriptional function by Raf kinases, which culminates in the expression of the angiogenic inducer Cyr61. FP receptors also regulate the induction of angiogenic immediate early genes in cultured ciliary muscle cells, which may constitute the early steps in a mechanism by which commercial FP agonists reduce intraocular pressure in glaucoma therapy.The up-regulation of BDNF through glial EP₂ receptors constitutes a mechanism by which elevated PGE₂ in the inflamed brain might elicit either healing processes in the brain or neuronal apoptosis. On the other hand, induction of TNF-&alpha; and Cyr61 by glial FP receptors may mediate neuroinflammation and may also contribute to glioma tumor growth. Stimulation of FP receptors in the ciliary muscle leads to the induction of immediate early genes capable of coordinating tissue remodeling processes that have been previously documented. The results of these studies suggest novel regulatory functions of the prostanoid receptors in the brain and eye. Furthermore, these findings provide insight on how the selective modulation of the EP₂ and FP receptors might be therapeutically advantageous.

Several research groups have applied elevated hydrostatic pressure to ONH astrocytes cultured on a rigid substrate as an in vitro model for glaucoma. These studies have shown significant biological effects and this hydrostatic pressure model is now becoming generally accepted in the ophthalmic community. However, since the applied pressures were modest the finding of significant biological effects due to pressure alone is surprising. We hypothesized that the application of hydrostatic pressure as described in these studies also altered gas tensions in the culture media. Our goal was to design equipment and carry out experiments to separate the biologic effects of pressure from those of hypoxia on cultured astrocytes. We designed equipment and carried out experiments to subject cultures of DITNC1 astrocytes to the four combinations of two levels of each parameter. We explored the morphology and migration rates of astrocytes, but observed no significant change in any of these properties.

HIV-infected individuals who abuse opiates have been found to have a higher incidence and a faster progression of HIV encephalitis. Astrocytes, the major support cells in the CNS, are known to play a critical role in the HIV neuropathy. Although astrocytes tend not to be productively infected by the HIV-1 virus, dysregulation of their pro- and anti- inflammatory cytokines/chemokines secretion is usually neurotoxic. Glutamate transport in astrocytes is reported to be impaired as well, which result in extrasynaptic excitatory neurotransmitter accumulation and over stimulation of postsynaptic neurons. It is long known that astrocytes from different brain regions have diverse responses to extracellular stimulants. However, few publications discuss the regional heterogeneity of astrocytes in HIV infected central nervous system. Based on the above information, we hypothesize that astroglia from different brain regions vary in their responses to HIV proteins and the responses could be influenced by co-treatment of morphine. To test this we cultured astrocytes from cerebral cortex, cerebellum and spinal cord, treated them with HIV-1 proteins Tat and gp120 with or without morphine and opioid receptor (mu, delta and kappa receptor) antagonist, naloxone. After 12-18 hours, conditioned medium from each group was analyzed using a Bio-Plex array. Cells from striatal cultures were harvested and lysed; proteins were extracted and evaluated with Western blotting to see whether EAAT2 expression on astrocytes is changed. Results showed that there were significant regional differences among three brain regions in cytokine/chemokine release, both in their basal secretion and in response to viral proteins. Astrocytes from spinal cord and cerebellum had a significantly higher basal secretion than those from cortical glia. All regions had increased cytokine/chemokine secretion when treated with Tat. Astroglia from the cortex showed the highest overall accumulation of cytokines/chemokines. Astroglia from the spinal cord had a slightly lower response overall, although KC expression was highest than other two groups in response to HIV-1 proteins. Astroglia from the cerebellum had a noticeably low response to Tat compared with those from spinal cord and cortex. However, concurrent morphine administration did not have a synergistic effect. No significant change in cytokines/chemokines release was seen when treated by gp120 with or without morphine. No significant change was found in EAAT2 expression on astrocytes either. In conclusion, astrocytes from different brain regions had different baseline secretion pattern and responses to viral protein. Tat had a noticeable effect in inducing cytokines/chemokines production in astrocytes from all brain regions, while limited change could be found with gp120 and morphine treatments. No significant change was found in EAAT2 expression on astrocytes.

This dissertation sheds light on several novel mechanistic findings in astrocyte specific gene regulation and function by the NFI-X transcription factor which can be further extrapolated to astrocyte differentiation and glial tumor invasion. First, we cloned and analyzed human NFI-X3, a novel splice variant of the nfix gene, which contains a unique transcriptional activation (TA) domain completely conserved in primates. In contrast to previously cloned NFI-X1, overexpression of NFI-X3 potently activates NFI reporters, including GFAP reporter, in astrocytes and glioma cells. The expression of NFI-X3 is dramatically upregulated during the differentiation of neural progenitors to astrocytes and precedes the expression of astrocyte markers such as GFAP and SPARCL1. Overexpression of NFI-X3 dramatically upregulates GFAP and SPARCL1 expression in glioma cells, while the knockdown of NFI-X3 diminishes the expression of both GFAP and SPARCL1 in astrocytes. Although activation of astrocyte-specific genes involves DNA demethylation and subsequent increase of histone acetylation, the TA domain of NFI-X3 activates GFAP expression by inducing alteration in the +1 nucleosome architecture that lead to the increased recruitment of RNA polymerase II. Thus, we propose that NFI-X3 is the major isoform of NFI-X regulating astrocyte specific gene expression during their differentiation, likely via nucleosomal remodeling of the astrocyte specific promoters. NFI-X knock-out animals display severe neuroanatomical defects including partial agenesis of the corpus callosum and hydrocephalus, however the target genes of NFI-X in the CNS remained elusive. Here, we show for the first time that YKL-40 is a novel target gene of NFI-X in astrocytes and controls their migration. In addition, we report that YKL-40 expression is activated during mouse brain development and also during the differentiation of neural progenitors into astrocytes in vitro. In primary astrocytes, YKL-40 expression is controlled by nuclear factor I-X (NFI-X) and signal transducer and activator of transcription 3 (STAT3), which are known to regulate gliogenesis. Indeed, knock-down of NFI-X and STAT3 significantly reduced YKL-40 expression in astrocytes, while overexpression of NFI-X3 (a splice isoform of NFI-X) dramatically induced YKL-40 expression in glioma cells. In addition, activation of STAT3 by oncostatin M induced YKL-40 expression in astrocytes. Interestingly, STAT3 activated YKL-40 expression via its binding site located in the YKL-40 proximal promoter, whereas direct NFI-X binding had not been detected. Surprisingly, NFI-X and STAT3 physically interact and this complex likely regulates YKL-40 expression in astrocytes. We further show that NFI-X controls migration and invasion of astrocytes and glioma cells, respectively, by regulating YKL-40 expression. These novel data suggest that YKL-40 is expressed by astrocytes during brain development and controls astrocyte migration. Since YKL-40 is used as a shared biomarker for ongoing inflammation and oncogenic transformation and its (high) levels correlate to the severity of disease, we have tested its expression in astrocytes and microglia (CNS macrophage) after treatment of various neuro-inflammatory cytokines. Here we report, that IL-1 and IL-6/OSM synergistically activate YKL-40 expression in astrocytes but not in microglia when added together. Furthermore, induced YKL-40 expression can be detected in the media from astrocytes but not from microglia. Since YKL-40 is a secreted molecule and is highly upregulated in CSF of multiple sclerosis patients, we have tested its role in oligodendrocyte differentiation. Preliminary observations clearly demonstrate that YKL-40 inhibits myelin basic protein (MBP) expression during the in-vitro differentiation of oligodendrocyte progenitor cells into myelin producing oligodendrocytes. Thus, we propose that YKL-40 is produced and secreted by reactive astrocytes during various CNS pathologies, and may inhibit MBP expression in MS. In summary, these studies have identified novel mechanisms in astrocyte gene regulation and functions, and provided new insights into astrocyte biology, with the implications for further understanding of the development and progression of CNS pathology.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Glioblastoma Multiforme (GBM) is an extremely aggressive grade IV brain tumour that is highly infiltrative and can spread to other parts of the brain quickly. It is the most common primary brain tumour where patients have a median survival of 14.6 months. Symptoms vary depending upon the location of the tumour and include seizures, progressive headaches and focal neurological deficit. The poor prognosis is characterised by deregulation of many key signalling pathways involving survival, growth, apoptosis and evasion of immune surveillance. In this study, we investigated whether complement factor H related protein 5 (CFHR5) from primary GBM cells direct from patients exhibited functional activity similar to factor H. The presence of CFHR5 was validated by western blot and ELISA technique from B30, B31 and B33 primary GBM cells. The functional capacity of CFHR5 was examined through the alternative pathway, co-factor, and decay acceleration assay. We demonstrated that CFHR5 was able to inhibit the alternative pathway through the same mechanism as factor H. Emerging evidence had shown that the innate immune protein surfactant protein D (SP-D) and recombinant human SP-D (rhSP-D) were able to induce apoptosis in eosinophilic leukaemic cells. We studied the ability of rhSP-D to induce apoptosis in U87 GBM cells through apoptotic and viability assays. rhSP-D was unable to mediate cell death and instead increased cell viability. This led us to investigate the expression of SP-D in U87 and B30 GBM cells through western blot, ELISA and immuno-fluorescence detection. We demonstrated novel information about the production of SP-D by GBM cells. To extend our study, we investigated the interaction of THP-1 macrophage with rhSP-D bound U87 cells. We carried out live cell imaging, RT-qPCR, and cell viability assays, to study the changes in cytokine expression and viability of cells. THP-1 did not engulf U87 cells; however, it did reduce the number of cells and decrease the expression of pro-tumourigenic cytokines. This study highlights the ability of primary GBM cells to evade innate immune detection by the secretion of functionally active CFHR5. It also demonstrated the ability of U87 to evade destruction by rhSP-D and THP-1 highlighting the extremely aggressive behaviour of the tumour and lack of new treatment to improve prognosis in over a decade.