In experimental models of demyelination such as cat optic nerve injected with antibody to galactocerebroside, stepwise and ultimately full repair occurs, starting with recruitment of oligodendrocyte progenitor cells (OP) from surrounding tissue and culminating in remyelination by young competent oligodendrocytes. Endogenous repair of demyelination can also occur in the adult human central nervous system, as evidenced by remyelinated shadow plaques in MS, but ultimately fails in this disease, leading to areas of chronic demyelination where surviving axons are both dysfunctional in terms of conduction and vulnerable to ongoing damage. In order to meaningfully investigate this failure of remyelination in the human situation, an essential prerequisite is to be able to reliably identify the neuroglial cells, and in particular, oligodendrocyte lineage cells, involved in the repair pathway in situ in post mortem tissue. While some marker antigens have been shown to remain demonstrable despite autolytic change and through differing fixation levels, others are far more sensitive and only reliable in freshly obtained tissue with light fixation. For instance, the surface antigens NG2 and PDGFαR, which have been widely used in experimental studies as a marker for OP both in vivo and in vitro, have been shown to be adversely affected by both fixation and autolysis. To this end, the cat optic nerve demyelination model, in which the reparative oligodendrocyte lineage stages have been antigenically defined, was extended to normal optic nerve including lightly fixed tissue. Here, NG2, PDGFαR and the oligodendrocyte lineage transcription factors Olig1 and Olig2 were able to be demonstrated and then correlated with the existing antigenic phenotypes. Subsequently, normal human optic nerve, optimised for both morphological preservation and antigen retention, was used to develop an in vivo staining profile for all neuroglia including OP, that was then applied to conventionally prepared, normal and MS tissue. It was found that, with careful attention to technical parameters such as post mortem interval and details of fixation, OP and other stages of the remyelinating oligodendrocyte lineage could be identified in such material, resulting in meaningful insight into the repair status of the three MS samples studied.

O complexo da esclerose tuberosa (TSC) é um transtorno clínico, com expressividade variável, caracterizado por hamartomas que podem ocorrer em diferentes órgãos. Tem herança autossômica dominante e é devido a mutações em um de dois genes supressores de tumor, TSC1 ou TSC2. Estes codificam para as proteínas hamartina e tuberina, respectivamente, que se associam formando um complexo macromolecular que regula funções como proliferação, diferenciação, crescimento e migração celular. As lesões cerebrais podem ser muito graves em pacientes com TSC e caracterizam-se por nódulos subependimários (SEN), astrocitomas subependimários de células gigantes (SEGA), tuberosidades corticais e heterotopias neuronais, podendo relacionar-se clinicamente à epilepsia refratária à terapia medicamentosa, deficiência intelectual, desordens do comportamento e hidrocefalia. O potencial de crescimento de SEGA até os 21 anos de idade dos pacientes exige acompanhamento periódico por exame de imagem e condutas clínicas ou cirúrgicas, conforme indicação médica. As lesões subependimárias têm sido explicadas por déficits de controle da proliferação, crescimento e diferenciação de precursores neuro-gliais na zona subventricular telencefálica. Embora a capacidade da tuberina em inibir a proliferação celular pela repressão do alvo da rapamicina em mamíferos (mTOR) esteja bem documentada, outros aspectos celulares do desenvolvimento de SEGA ainda não foram examinados. Assim, é importante estabelecer um sistema in vitro para o estudo de células da zona subventricular e testá-lo na análise das proteínas hamartina e tuberina. Neste sentido, o cultivo de neuroesferas em suspensão é muito apropriado. Neste estudo, buscamos relacionar a expressão e distribuição subcelular da hamartina e tuberina à capacidade proliferativa e de diferenciação das células de neuroesferas cultivadas in vitro a partir da dissociação da vesícula telencefálica de embriões de ratos normais. Analisamos a expressão e distribuição subcelular da hamartina e tuberina por imunofluorescência indireta em células entre a primeira e a quarta passagens das neuroesferas, sincronizadas nas fases G1 ou S do ciclo celular e após a reentrada no ciclo celular, através da incorporação de 5-bromo-2\'-desoxiuridina (BrdU) e imunofluorescência com anticorpo anti-BrdU. Em geral, células de neuroesferas apresentaram baixa colocalização entre hamartina e tuberina in vitro. A expressão da tuberina foi elevada em basicamente todas as células das esferas e fases do ciclo celular; ao contrário, a hamartina apresentou-se principalmente nas células da periferia das esferas. A colocalização entre hamartina e tuberina foi observada em células mais periféricas das esferas, sobretudo no citoplasma e, em G1, no núcleo celular. A proteína rheb, que conhecidamente interage diretamente com a tuberina, apresentou distribuição subcelular muito semelhante à desta. Ao carenciamento das células visando à parada do ciclo celular na transição G1/S, tuberina distribuiu-se ao núcleo celular em quase todas as células avaliadas e, de forma menos frequente, a hamartina também. À reentrada no ciclo celular pelo reacréscimo dos fatores de crescimento, avaliaram-se células com incorporação de BrdU ao seu núcleo celular, após 72 e 96 horas. Nestas, tuberina mostrou-se novamente no citoplasma de forma preponderante e hamartina manteve-se citoplasmática, em geral subjacente à membrana plasmática, em níveis mais baixos. Os grupos cujas células reciclaram por 72 ou 96 horas diferiram quanto ao aumento significativo da expressão da hamartina em células proliferativas no último. À diferenciação neuronal, aumentaram-se os níveis de expressão de hamartina observáveis à imunofluorescência indireta, tornando-se equivalentes àqueles da tuberina. Concluímos que as células de neuroesferas cultivadas em suspensão apresentam-se como um sistema apropriado ao estudo da distribuição das proteínas hamartina e tuberina e sua relação com o ciclo celular / The tuberous sclerosis complex (TSC) is a clinical disorder with variable expressivity, characterized by hamartomas that can occur in different organs. It has autosomal dominant inheritance and is due to mutations in one of two tumor suppressor genes, TSC1 or TSC2. These encode for the proteins hamartin and tuberin, respectively, which are associated in a macromolecular complex which functions as a regulator of cell proliferation, differentiation, growth and migration. TSC brain lesions may be severe and are characterized by subependymal nodules (SEN), subependymal giant cell astrocytomas (SEGA), neuronal heterotopias and cortical tubers, and may be clinically related to refractory epilepsy, intellectual disability, behavioral disorders and hydrocephaly. The growth potential of SEGA up to 21 years of age in TSC patients requires regular monitoring by imaging. Clinical and surgical interventions may be medically indicated. Subependymal lesions have been explained by deficient control of proliferation, growth and differentiation of neuro-glial progenitors from the telencephalic subventricular zone. While tuberin ability to inhibit cell proliferation by repressing the mammalian target of rapamycin (mTOR) has been well documented, other cell aspects of SEGA development have not been thoroughly examined. Therefore, it is important to establish conditions for an in vitro system to study the cells from the subventricular zone and to test its suitability for the study of the TSC proteins. In this regard, the neurosphere suspension culture is very appropriate. We evaluated the expression and subcellular distribution of hamartin and tuberin in relation to the proliferation and differentiation capability of neurosphere cells derived in vitro from the dissociation of the telencephalic vesicle of normal E14 rat embryos. These analyses were performed by indirect immunofluorescence in cells from first through fourth passages of neurospheres, synchronized in G1 or S phases of the cell cycle, and after reentry into the cell cycle by the addition of 5-brome-2\'-desoxyuridine (BrdU) and immunolabeling with anti-BrdU antibody. In general, neurosphere cells presented low colocalization between hamartin and tuberin in vitro. Tuberin expression was relatively high in basically all neurosphere cells and cell cycle phases, whereas hamartin distributed mainly to cells from the periphery of the spheres. In these cells, hamartin and tuberin colocalization was evident mostly in the cytoplasm and, in G1, also in the cell nucleus. Rheb, which is known to interact directly with tuberin, had subcellular distribution very similar to tuberin. Cell starvation indicating cell cycle arrest at G1/S redistributed tuberin to the cell nucleus in virtually all cells examined, what was accompanied by nuclear location of hamartin in a small subset of cells. When cells were allowed to reenter cell cycle by adding growth factors, we evaluated BrdU-labeled nuclei 72 and 96 hours later. In the two groups, tuberin was shown to move back to the cytoplasm as well as hamartin, which apparently maintained its lower expression levels distribution underneath the plasma membrane. Group of cells that recycled for 96 hours had significantly more expression of hamartin than those cells that cycled for only 72 hours. After neuronal differentiation, hamartin expression levels observed by immunofluorescence were similar to those of tuberin. We conclude that neurosphere cells cultured in suspension showed to be an appropriate cell system to study hamartin and tuberin distribution in respect to the cell cycle

O complexo da esclerose tuberosa (TSC) é um transtorno clínico, com expressividade variável, caracterizado por hamartomas que podem ocorrer em diferentes órgãos. Tem herança autossômica dominante e é devido a mutações em um de dois genes supressores de tumor, TSC1 ou TSC2. Estes codificam para as proteínas hamartina e tuberina, respectivamente, que se associam formando um complexo macromolecular que regula funções como proliferação, diferenciação, crescimento e migração celular. As lesões cerebrais podem ser muito graves em pacientes com TSC e caracterizam-se por nódulos subependimários (SEN), astrocitomas subependimários de células gigantes (SEGA), tuberosidades corticais e heterotopias neuronais, podendo relacionar-se clinicamente à epilepsia refratária à terapia medicamentosa, deficiência intelectual, desordens do comportamento e hidrocefalia. O potencial de crescimento de SEGA até os 21 anos de idade dos pacientes exige acompanhamento periódico por exame de imagem e condutas clínicas ou cirúrgicas, conforme indicação médica. As lesões subependimárias têm sido explicadas por déficits de controle da proliferação, crescimento e diferenciação de precursores neuro-gliais na zona subventricular telencefálica. Embora a capacidade da tuberina em inibir a proliferação celular pela repressão do alvo da rapamicina em mamíferos (mTOR) esteja bem documentada, outros aspectos celulares do desenvolvimento de SEGA ainda não foram examinados. Assim, é importante estabelecer um sistema in vitro para o estudo de células da zona subventricular e testá-lo na análise das proteínas hamartina e tuberina. Neste sentido, o cultivo de neuroesferas em suspensão é muito apropriado. Neste estudo, buscamos relacionar a expressão e distribuição subcelular da hamartina e tuberina à capacidade proliferativa e de diferenciação das células de neuroesferas cultivadas in vitro a partir da dissociação da vesícula telencefálica de embriões de ratos normais. Analisamos a expressão e distribuição subcelular da hamartina e tuberina por imunofluorescência indireta em células entre a primeira e a quarta passagens das neuroesferas, sincronizadas nas fases G1 ou S do ciclo celular e após a reentrada no ciclo celular, através da incorporação de 5-bromo-2\'-desoxiuridina (BrdU) e imunofluorescência com anticorpo anti-BrdU. Em geral, células de neuroesferas apresentaram baixa colocalização entre hamartina e tuberina in vitro. A expressão da tuberina foi elevada em basicamente todas as células das esferas e fases do ciclo celular; ao contrário, a hamartina apresentou-se principalmente nas células da periferia das esferas. A colocalização entre hamartina e tuberina foi observada em células mais periféricas das esferas, sobretudo no citoplasma e, em G1, no núcleo celular. A proteína rheb, que conhecidamente interage diretamente com a tuberina, apresentou distribuição subcelular muito semelhante à desta. Ao carenciamento das células visando à parada do ciclo celular na transição G1/S, tuberina distribuiu-se ao núcleo celular em quase todas as células avaliadas e, de forma menos frequente, a hamartina também. À reentrada no ciclo celular pelo reacréscimo dos fatores de crescimento, avaliaram-se células com incorporação de BrdU ao seu núcleo celular, após 72 e 96 horas. Nestas, tuberina mostrou-se novamente no citoplasma de forma preponderante e hamartina manteve-se citoplasmática, em geral subjacente à membrana plasmática, em níveis mais baixos. Os grupos cujas células reciclaram por 72 ou 96 horas diferiram quanto ao aumento significativo da expressão da hamartina em células proliferativas no último. À diferenciação neuronal, aumentaram-se os níveis de expressão de hamartina observáveis à imunofluorescência indireta, tornando-se equivalentes àqueles da tuberina. Concluímos que as células de neuroesferas cultivadas em suspensão apresentam-se como um sistema apropriado ao estudo da distribuição das proteínas hamartina e tuberina e sua relação com o ciclo celular / The tuberous sclerosis complex (TSC) is a clinical disorder with variable expressivity, characterized by hamartomas that can occur in different organs. It has autosomal dominant inheritance and is due to mutations in one of two tumor suppressor genes, TSC1 or TSC2. These encode for the proteins hamartin and tuberin, respectively, which are associated in a macromolecular complex which functions as a regulator of cell proliferation, differentiation, growth and migration. TSC brain lesions may be severe and are characterized by subependymal nodules (SEN), subependymal giant cell astrocytomas (SEGA), neuronal heterotopias and cortical tubers, and may be clinically related to refractory epilepsy, intellectual disability, behavioral disorders and hydrocephaly. The growth potential of SEGA up to 21 years of age in TSC patients requires regular monitoring by imaging. Clinical and surgical interventions may be medically indicated. Subependymal lesions have been explained by deficient control of proliferation, growth and differentiation of neuro-glial progenitors from the telencephalic subventricular zone. While tuberin ability to inhibit cell proliferation by repressing the mammalian target of rapamycin (mTOR) has been well documented, other cell aspects of SEGA development have not been thoroughly examined. Therefore, it is important to establish conditions for an in vitro system to study the cells from the subventricular zone and to test its suitability for the study of the TSC proteins. In this regard, the neurosphere suspension culture is very appropriate. We evaluated the expression and subcellular distribution of hamartin and tuberin in relation to the proliferation and differentiation capability of neurosphere cells derived in vitro from the dissociation of the telencephalic vesicle of normal E14 rat embryos. These analyses were performed by indirect immunofluorescence in cells from first through fourth passages of neurospheres, synchronized in G1 or S phases of the cell cycle, and after reentry into the cell cycle by the addition of 5-brome-2\'-desoxyuridine (BrdU) and immunolabeling with anti-BrdU antibody. In general, neurosphere cells presented low colocalization between hamartin and tuberin in vitro. Tuberin expression was relatively high in basically all neurosphere cells and cell cycle phases, whereas hamartin distributed mainly to cells from the periphery of the spheres. In these cells, hamartin and tuberin colocalization was evident mostly in the cytoplasm and, in G1, also in the cell nucleus. Rheb, which is known to interact directly with tuberin, had subcellular distribution very similar to tuberin. Cell starvation indicating cell cycle arrest at G1/S redistributed tuberin to the cell nucleus in virtually all cells examined, what was accompanied by nuclear location of hamartin in a small subset of cells. When cells were allowed to reenter cell cycle by adding growth factors, we evaluated BrdU-labeled nuclei 72 and 96 hours later. In the two groups, tuberin was shown to move back to the cytoplasm as well as hamartin, which apparently maintained its lower expression levels distribution underneath the plasma membrane. Group of cells that recycled for 96 hours had significantly more expression of hamartin than those cells that cycled for only 72 hours. After neuronal differentiation, hamartin expression levels observed by immunofluorescence were similar to those of tuberin. We conclude that neurosphere cells cultured in suspension showed to be an appropriate cell system to study hamartin and tuberin distribution in respect to the cell cycle

La dystrophie myotonique de type 1 (DM1), ou maladie de Steinert, est une maladie qui touche plusieurs tissus, dont le système nerveux central (SNC). L’atteinte neurologique est variable et inclut des troubles de la fonction exécutive, des changements de comportement et une hypersomnolence dans la forme adulte, ainsi qu’une déficience intellectuelle marquée dans la forme congénitale. Dans leur ensemble, les symptômes neurologiques ont un fort impact sur le parcours académique, professionnel et les interactions sociales. Aujourd’hui aucune thérapie n’existe pour cette maladie. La DM1 est due à une expansion anormale d’un triplet CTG non-codant dans le gène DMPK. Les ARN messagers DMPK, porteurs de l’expansion, s’accumulent dans le noyau des cellules (sous forme de foci) et perturbent la localisation et la fonction de protéines de liaison à l’ARN, notamment des familles MBNL et CELF, ce qui entraîne des défauts d’épissage alternatif, d’expression, de polyadenylation et de localisation d’autres ARN cibles. Malgré le progrès récent dans la compréhension des mécanismes de la maladie, les aspects cellulaires et moléculaires de l’atteinte neurologique restent méconnus: nous ne connaissons ni la contribution de chaque type cellulaire du cerveau, ni les voies moléculaires spécifiquement dérégulées dans chaque type cellulaire. L’objectif de ma thèse a été de répondre à ces deux questions importantes en utilisant un modèle de souris transgéniques et des cellules primaires dérivées de celui-ci. Pour mon projet, j’ai utilisé les souris DMSXL générées par mon laboratoire. Ces souris reproduisent des caractéristiques importantes de la DM1, notamment l’accumulation des ARN toxiques et la dérégulation de l’épissage alternatif dans plusieurs tissus. L’impacte fonctionnel des transcrits DMPK toxiques dans le SNC des souris DMSXL se traduit par des problèmes comportementaux et cognitifs et par des défauts de la plasticité synaptique. Afin d’identifier les mécanismes moléculaires associés à ces anomalies, une étude protéomique globale a montré une dérégulation de protéines neuronales et astrocytaires dans le cerveau des souris DMSXL. De plus, l’étude de la distribution des foci d’ARN dans les cerveaux des souris et des patients a montré un contenu plus élevé dans les astrocytes par rapport aux neurones. Ensemble, ces résultats suggèrent une contribution à la fois neuronale et gliale dans la neuropathogenèse de la DM1. L’étude protéomique globale des cerveaux des souris DMSXL, a aussi montré des défauts de protéines synaptiques spécifiques des neurones, que nous avons par la suite validés dans le cerveau des patients. SYN1 est hyperphosphorylée d’une façon CELF-dépendante et RAB3A est surexprimé en réponse à l’inactivation de MBNL1. Les protéines MBNL et CELF régulent l’épissage alternatif d’un groupe de transcrits au cours du développement, et leur dérégulation dans la DM1 entraîne l’expression anormale d’isoformes d’épissage embryonnaires dans le tissu adulte. Dans ce contexte, j’ai étudié si les défauts des protéines RAB3A et SYN1 sont associés à une dérégulation d’épissage, et si les anomalies des protéines synaptiques identifiées dans la DM1 reproduisent des évènements embryonnaires de la régulation de RAB3A et SYN1. Mes résultats indiquent que les défauts de ces protéines dans les cerveaux adultes ne sont pas dus à une altération de l’épissage alternatif des transcrits et ne recréent pas des évènements embryonnaires. La neuropathogenèse de la DM1 va, donc, au delà de la dérégulation de l’épissage et d’autres voies moléculaires restent à explorer dans les cerveaux DM1. Afin d’identifier des sous-populations cellulaires susceptibles à l’accumulation des ARN toxiques, nous avons étudié la distribution des foci dans plusieurs régions cérébrales. (...) / Myotonic dystrophy type 1 (DM1) is a severe disorder that affects many tissues, including the central nervous system (CNS). The degree of brain impairment ranges from executive dysfunction, attention deficits, low processing speed, behavioural changes and hypersomnia in the adult form, to pronounced intellectual disability in the congenital cases. The neurological manifestations have a tremendous impact on the academic, professional, social and emotional aspects of daily life. Today there is no cure for this devastating condition. DM1 is caused by the abnormal expansion of a CTG trinucleotide repeat in the 3’UTR of the DMPK gene. Expanded DMPK transcripts accumulate in RNA aggregates (or foci) in the nucleus of DM1 cells, disrupting the activity of important RNA-binding proteins, like the MBNL and CELF families, and leading to abnormalities in alternative splicing, gene expression, RNA polyadenylation, localisation and translation. In spite of recent progress, fundamental gaps in our understanding of the molecular and cellular mechanisms behind the neurological manifestations still exist: we do not know the contribution of each cell type of the CNS to brain dysfunction, or the molecular pathways specifically deregulated in response to the CTG expansion. The aim of my PhD project has been to gain insight into these two important questions using a relevant transgenic mouse model of DM1 and cell cultures derived thereof. In my studies I used the DMSXL mice, previously generated in my host laboratory. The DMSXL mice express expanded DMPK mRNA with more than 1,000 CTG repeats. They recreate relevant DM1 features, such as RNA foci and missplicing in multiple tissues. The functional impact of expanded DMPK transcripts in the CNS of DMSXL mice translates into behavioural and cognitive abnormalities and defective synaptic plasticity. To identify the molecular mechanisms behind these abnormalities, a global proteomics analysis revealed changes in both neuron-specific and glial-specific proteins in DMSXL brain. We also investigated RNA foci in DMSXL and human DM1 brains and found non-homogenous distribution between cell types, with a higher foci content in astrocytes relative to neurons. Together these results suggest that both neuronal and glial defects contribute to DM1 neuropathogenesis. The global proteomics analysis of DMSXL brains also identified abnormalities in neuronal synaptic proteins that we have validated in human brain samples. SYN1 is hyperphosphorilated in a CELF-dependent manner while RAB3A is upregulated in association with MBNL1 depletion. CELF and MBNL proteins regulate the alternative splicing of a subset of transcripts throughout development, and their deregulation in DM1 leads to abnormal expression of fetal splicing isoforms in adult DM1 brains. In this context, I have studied if RAB3A and SYN1 deregulations observed in adult brains are associated with splicing abnormalities or if they recreated embryonic expression and phosphorylation events. My results indicate that the synaptic proteins abnormalities observed in adult DMSXL brains are not caused by defective alternative splicing and do not recreate embryonic events. Thus, DM1 neuropathogenesis goes beyond missplicing and other molecular pathways must be explored in DM1 brains. To better understand the cellular sub-populations susceptible of accumulating toxic RNA foci we have studied foci distribution in different brain regions. We identified pronounced accumulation of toxic RNAs in Bergman astrocytes of DMSXL mice cerebellum and DM1 patients, associated with neuronal hyperactivity of Purkinje cells. A quantitative proteomics analysis revealed a significant downregulation of GLT1 – a glial glutamate transporter expressed by the Bergmann cell in the cerebellum. I have confirmed the GLT1 downregulation in other brain regions of mouse and human brain. (...)

Les noyaux supraoptiques (NSO) et paraventriculaires (NPV) de l’hypothalamus montrent un phénomène réversible de plasticité structurale neurono-gliale dans diverses conditions physiologiques telles que la parturition, l’allaitement ou lors d’une surcharge en sel. En effet, les feuillets astrocytaires qui enveloppent normalement les somas et dendrites des neurones à ocytocine (OT) ou à vasopressine (AVP) se rétractent alors, autour des neurones à OT, laissant place à la formation de nouvelles synapses, surtout GABAergiques. Nous avons émis l’hypothèse voulant que ces mouvements cellulaires soient régulés par des molécules connues pour leurs rôles dans l’adhérence et la motilité cellulaires, notamment les récepteurs Eph et les éphrines (Efn). Nous avons étudié le rôle de l’un de ces récepteurs, EphA4, un récepteur à tyrosine kinase reconnaissant l’ensemble des Efn, A ou B, puis tenté d’identifier les Efn partenaires dans le NSO, à la suite d’une surcharge en sel. Pour démontrer la présence d’EphA4 dans le NSO et déterminer l’effet d’une surcharge en sel sur son expression et sa localisation, nous avons utilisé l’hybridation in situ et l’immunohistochimie en microscopie électronique, sur des coupes de cerveaux de souris ou rats traités ou non à l’eau salée pendant 1-7 j, avec des ribosondes ou des anticorps spécifiques pour EphA4. Ces travaux ont démontré une augmentation de l’expression d’EphA4 dans le NSO, notamment dans des dendrites, après le régime salé. La distribution de cette expression correspondait à celle des neurones OT et était absente de la glia limitans. Nous avons ensuite déterminé l’effet d’une absence d’EphA4 sur les mouvements astrocytaires et la synaptogènese autour des dendrites à OT et AVP, en utilisant des souris EphA4 knockouts et des souris de type sauvage des mêmes portées. Nous avons ainsi mesuré la couverture astrocytaire des dendrites OT ou AVP, identifiées par immunocytochimie anti-OT ou anti-AVP, en microscopie électronique. Ces mesures ont confirmé la rétraction des feuillets astrocytaires et la synaptogenèse autour des dendrites OT, mais pas autour des dendrites AVP, chez les souris de type sauvage, et démontré que la rétraction des feuillets astrocytaires et la synaptogenèse sur les dendrites OT ne se produisait pas chez les souris knockouts soumises à la surcharge en sel. L’ensemble de ces résultats démontre un rôle d’EphA4 dans cette plasticité structurale neurono-gliale. Afin d’identifier l’Efn partenaire d’EphA4 dans cette fonction, nous avons utilisé l’hybridation in situ et l’immunohistochimie pour les EfnB3 et -A3. L’hybridation in situ n’a pas démontré d’expression de l’EfnB3 dans le NSO, tandis que les résultats pour l’EfnA3 restent à quantifier. Cependant, l’immunohistochimie anti-EfnA3 montre un marquage d’astrocytes dans le NSO et la glia limitans, marquage qui semble augmenter après surcharge en sel, mais il reste à démontrer que l’anticorps anti-EfnA3 est bien spécifique et à quantifier les éventuels changements sur un plus grand nombre d’animaux. L’ensemble de ces observations démontre un rôle du récepteur EphA4 dans les mécanismes à la base des changements structuraux neurono-gliaux du NSO et pointe vers l’EfnA3 comme partenaire d’EphA4 dans ce modèle. / The supraoptic (SON) and paraventricular (PVN) nuclei of the hypothalamus display reversible neurono-glial structural plasticity in various physiological conditions, such as parturition, lactation, or following salt loading. In such conditions, astrocytic leaflets that normally envelop the somas and dendrites of ocytocin (OT) or arginin-vasopressin (AVP) neurons retract from OT processes where they are replaced by new synapses, mainly GABAergic. Our hypothesis proposes that these cellular movements are regulated by molecules known for their roles in cell adhesion and motility, notably Eph receptors and ephrins (Efn). We have examined the role of one of these receptors, EphA4, a tyrosine-kinase receptor recognizing all ephrins, A or B, and then tried to identify the Efn interacting with EphA4 in these functions, following salt-loading. To demonstrate the presence of EphA4 in the SON and determine the effect of salt loading on its expression, we used in situ hybridization and immunohistochemistry in light and electron microscopy, on brain sections from rats or mice treated with salted water during 1-7 d, using riboprobes and antibodies specific for EphA4. These experiments demonstrated that EphA4 is expressed in the SON, with a distribution of its mRNA similar to that of OT neurons, and that it was absent from the glia limitans. Its expression increased following salt loading, particularly in dendrites. We then tested the effect of an absence of EphA4 on astrocytic process retraction and on synaptogenesis, using EphA4 kockout mice and wild-type littermates. We measured the ratio of astrocytic contact, and counted the number of synapses on the circumference of OT and AVP dendrites, identified in electron microscopy by immunocytochemistry, after 7 d of salt loading. The results confirmed the retraction of astrocytic processes from OT dendrites in wild-type animals after salt loading, and no change around AVP dendrites. However, there was no retraction from OT dendrites in EphA4 knockout mice, following salt loading. Altogether, these results constitute strong evidence for a role of EphA4 in the astrocyte leaflet retraction and accompanying synaptogenesis, specifically around OT dendrites. In order to identify the Efn interacting with EphA4 in this function, we used in situ hybridization and immunohistochemistry for EfnB3 and –A3. The in situ hybridization did not show the presence of EfnB3 in the SON, while the results for EfnA3 are currently being quantified. Nevertheless, anti-EfnA3 immunohistochemistry showed labelling in astrocytes and in the glia limitans of the SON, a labelling that seemed to increase following salt loading, although the specificity of the anti-EfnA3 antibody remains to be demonstrated on EfnA3 knockout mice, and its expression requires to be measured on a larger number of mice. The latter observations indicate EfnA3 as the potential partner (receptor/ligand) for EphA4 in the neurono-glial structural plasticity occurring in the SON following salt loading.

Les noyaux supraoptiques (NSO) et paraventriculaires (NPV) de l’hypothalamus montrent un phénomène réversible de plasticité structurale neurono-gliale dans diverses conditions physiologiques telles que la parturition, l’allaitement ou lors d’une surcharge en sel. En effet, les feuillets astrocytaires qui enveloppent normalement les somas et dendrites des neurones à ocytocine (OT) ou à vasopressine (AVP) se rétractent alors, autour des neurones à OT, laissant place à la formation de nouvelles synapses, surtout GABAergiques. Nous avons émis l’hypothèse voulant que ces mouvements cellulaires soient régulés par des molécules connues pour leurs rôles dans l’adhérence et la motilité cellulaires, notamment les récepteurs Eph et les éphrines (Efn). Nous avons étudié le rôle de l’un de ces récepteurs, EphA4, un récepteur à tyrosine kinase reconnaissant l’ensemble des Efn, A ou B, puis tenté d’identifier les Efn partenaires dans le NSO, à la suite d’une surcharge en sel. Pour démontrer la présence d’EphA4 dans le NSO et déterminer l’effet d’une surcharge en sel sur son expression et sa localisation, nous avons utilisé l’hybridation in situ et l’immunohistochimie en microscopie électronique, sur des coupes de cerveaux de souris ou rats traités ou non à l’eau salée pendant 1-7 j, avec des ribosondes ou des anticorps spécifiques pour EphA4. Ces travaux ont démontré une augmentation de l’expression d’EphA4 dans le NSO, notamment dans des dendrites, après le régime salé. La distribution de cette expression correspondait à celle des neurones OT et était absente de la glia limitans. Nous avons ensuite déterminé l’effet d’une absence d’EphA4 sur les mouvements astrocytaires et la synaptogènese autour des dendrites à OT et AVP, en utilisant des souris EphA4 knockouts et des souris de type sauvage des mêmes portées. Nous avons ainsi mesuré la couverture astrocytaire des dendrites OT ou AVP, identifiées par immunocytochimie anti-OT ou anti-AVP, en microscopie électronique. Ces mesures ont confirmé la rétraction des feuillets astrocytaires et la synaptogenèse autour des dendrites OT, mais pas autour des dendrites AVP, chez les souris de type sauvage, et démontré que la rétraction des feuillets astrocytaires et la synaptogenèse sur les dendrites OT ne se produisait pas chez les souris knockouts soumises à la surcharge en sel. L’ensemble de ces résultats démontre un rôle d’EphA4 dans cette plasticité structurale neurono-gliale. Afin d’identifier l’Efn partenaire d’EphA4 dans cette fonction, nous avons utilisé l’hybridation in situ et l’immunohistochimie pour les EfnB3 et -A3. L’hybridation in situ n’a pas démontré d’expression de l’EfnB3 dans le NSO, tandis que les résultats pour l’EfnA3 restent à quantifier. Cependant, l’immunohistochimie anti-EfnA3 montre un marquage d’astrocytes dans le NSO et la glia limitans, marquage qui semble augmenter après surcharge en sel, mais il reste à démontrer que l’anticorps anti-EfnA3 est bien spécifique et à quantifier les éventuels changements sur un plus grand nombre d’animaux. L’ensemble de ces observations démontre un rôle du récepteur EphA4 dans les mécanismes à la base des changements structuraux neurono-gliaux du NSO et pointe vers l’EfnA3 comme partenaire d’EphA4 dans ce modèle. / The supraoptic (SON) and paraventricular (PVN) nuclei of the hypothalamus display reversible neurono-glial structural plasticity in various physiological conditions, such as parturition, lactation, or following salt loading. In such conditions, astrocytic leaflets that normally envelop the somas and dendrites of ocytocin (OT) or arginin-vasopressin (AVP) neurons retract from OT processes where they are replaced by new synapses, mainly GABAergic. Our hypothesis proposes that these cellular movements are regulated by molecules known for their roles in cell adhesion and motility, notably Eph receptors and ephrins (Efn). We have examined the role of one of these receptors, EphA4, a tyrosine-kinase receptor recognizing all ephrins, A or B, and then tried to identify the Efn interacting with EphA4 in these functions, following salt-loading. To demonstrate the presence of EphA4 in the SON and determine the effect of salt loading on its expression, we used in situ hybridization and immunohistochemistry in light and electron microscopy, on brain sections from rats or mice treated with salted water during 1-7 d, using riboprobes and antibodies specific for EphA4. These experiments demonstrated that EphA4 is expressed in the SON, with a distribution of its mRNA similar to that of OT neurons, and that it was absent from the glia limitans. Its expression increased following salt loading, particularly in dendrites. We then tested the effect of an absence of EphA4 on astrocytic process retraction and on synaptogenesis, using EphA4 kockout mice and wild-type littermates. We measured the ratio of astrocytic contact, and counted the number of synapses on the circumference of OT and AVP dendrites, identified in electron microscopy by immunocytochemistry, after 7 d of salt loading. The results confirmed the retraction of astrocytic processes from OT dendrites in wild-type animals after salt loading, and no change around AVP dendrites. However, there was no retraction from OT dendrites in EphA4 knockout mice, following salt loading. Altogether, these results constitute strong evidence for a role of EphA4 in the astrocyte leaflet retraction and accompanying synaptogenesis, specifically around OT dendrites. In order to identify the Efn interacting with EphA4 in this function, we used in situ hybridization and immunohistochemistry for EfnB3 and –A3. The in situ hybridization did not show the presence of EfnB3 in the SON, while the results for EfnA3 are currently being quantified. Nevertheless, anti-EfnA3 immunohistochemistry showed labelling in astrocytes and in the glia limitans of the SON, a labelling that seemed to increase following salt loading, although the specificity of the anti-EfnA3 antibody remains to be demonstrated on EfnA3 knockout mice, and its expression requires to be measured on a larger number of mice. The latter observations indicate EfnA3 as the potential partner (receptor/ligand) for EphA4 in the neurono-glial structural plasticity occurring in the SON following salt loading.