published_or_final_version / Biological Sciences / Doctoral / Doctor of Philosophy

<p>The introduction of high density EEG (hd-EEG) nets for easy application on subjects of all ages has improved the possibilities to investigate the development of the infant neurophysiology. This dissertation consists of three studies (I – III) that investigate the visual motion system and mirror neuron system of the infant, and methodological sections that outline the bioelectrical background and the characteristics of the methods used. </p><p>Study I covers the maturation of cortical areas involved in motion perception in adults and infants using an ERP paradigm. Over three age groups (2, 3 and 5 month olds) the cortical activation increased dramatically. All infant groups showed significant activation when moving displays was contrasted to static displays on a video screen. The study shows that 5-month-old infants and older can be expected to process motion in a similar fashion as adults.</p><p>Study II covers the infant mirror neuron system (MNS). In adults the mu rhythm perturbations is considered a reliable measure of activation of the MNS. This study presented both a mu rhythm analysis and a ERP analysis to detect MNS activity in 6-month-olds and in adults. This study concludes that the infant MNS can be measured using ERPs and that the development of mu rhythm perturbations requires further study.</p><p>Study III focused on exploring the mu rhythm suppressions. 8-month-olds observed a live actor that performed goal directed reaches and non-goal directed hand movements. The results show robust mu rhythm perturbations time-locked to the grasping moment. The study concluded that the MNS activity is possible to evaluate by analysis of mu rhythm perturbations and that the MNS show mature characteristics at the age of 8 months.</p><p>In summary, Study 2 and 3 present new methods to investigate the infant mirror neuron system and shows that the infant MNS is active at 6 months of age. At 8 months of age the infant MNS show mature EEG responses to simple actions such as reaching. How the MNS development relates to the infants’ motor development, and how the MNS interacts with the development of social skills requires further studies that could benefit from the methods presented here.</p>

L'analyse de la fonction cérébrale à de multiples échelles est une étape nécessaire pour comprendre ses complexités. Dans ce travail de thèse, nous avons étudié cet aspect aux niveaux microscopiques et macroscopiques en utilisant des enregistrements invasifs et non-invasifs. Nous avons utilisé une série d'outils d'analyse communicationnels et de corrélation pour étudier l'activité cérébrale pendant l'éveil et le sommeil. Dans une première étude, nous avons analyse les enregistrements simultanés d'electroencephalogramme (EEG) etmagnetoencephalogramme (MEG) dans le cerveau de sujets éveillés. Nous montrons théoriquement, que si le milieu est résistif, le comportement d'échelle en fréquence doit être le même pour les signaux EEG and MEG à basse fréquence (˂10 Hz). Afin de tester cette prédiction, nous avons analyse le spectre d'enregistrements EEG et MEG simultanés de quatre sujets humains. Le comportement d'échelle en fréquence de l'EEG montre des variations cohérentes sur la surface du cerveau, avec des exposants en général compris entre 1/ f et 1/ f 2; ces exposants tendent à être plus faibles dans les régions temporales et pariétales. Dans une région donnée, les exposants de la MEG ont une variabilité plus grande que pour l'EEG, mais les deux signaux ont systématiquement un exposant différent. Dans certains cas, les exposants sont proches, mais ces cas correspondent 'a un mauvais rapport signal/bruit pour la MEG. Plusieurs méthodes de corrections du bruit instrumental et environnemental ont été testées, et dans tous les cas, ces méthodes augmentent la différence de comportement spectral entre l'EEG et la MEG. En conclusion, il y a une différence significative de comportement d'échelle en fréquence entre EEG et MEG, ce qui peut être explique si le milieu extracellulaire (incluant d'autres couches telles que la dure-mère et le crane) est globalement non-résistif. La nature résistive ou non-résistive du milieu extracellulaire est un déterminant important pour la modélisation des potentiels extracellulaires. Au cours d'une seconde étude, nous avons analyse la dynamique spatio-temporelle de l'excitation et de l'inhibition pendant le sommeil 'a partir d'enregistrements intra-crâniens à haute densité. Ces enregistrements à haute densité permettent la séparation efficace entre cellules "regular spiking" (RS) et "fast spiking" (FS). La haute densité des électrodes permet d'obtenir des connections apparemment mono-synaptiques, et de corroborer cette séparation RS-FS avec la nature excitatrice ou inhibitrice de la connexion. Cette procédure confirme que les cellules classifiées comme FS sont toujours inhibitrices, alors que les RS sont toujours excitatrices, et donc peuvent être classifiées respectivement comme cellules pyramidales ou interneurones inhibiteurs. Finalement, nous investiguons la dynamique des corrélations au sein de chaque classe de neurone. Les corrélations entre excitateurs montrent une décroissance exponentielle avec la distance, tandis que les cellules inhibitrices restent corrélées 'a plus grande distance. L'amplitude des corrélations dépend de l'échelle temporelle du calcul de corrélation, mais pas la constante spatiale. Cette constante est compatible avec la taille typique des colonnes corticales chez l'homme. Ces résultats permettent, pour la première fois, de caractériser l'activité neuronale et l'interaction entre excitation et inhibition dans le néocortex humain. Dans une troisième étude, nous avons investigue les signatures de la dynamique complexe et l'activité auto-organisée, à partir d'enregistrements intra-crâniens chez le chat, le singe et l'homme. Nous utilisons des enregistrements à haute densité dans le cortex moteur du chat (96 électrodes), le cortex moteur et prémoteur du singe et dans le cortex temporal humain (96 électrodes) de patients épileptiques. Lors d'avalanches définies à partir d'unités (jusqu''a 160 neurones), les distributions ne se comportent pas en loi de puissance, mais tendent à être exponentielles ou intermédiaires. Nous analysons également les potentiels de champ (LFPs), et en particulier les pics négatifs (nLFPs) au sein de l'ensemble d'électrodes (de 96 a 128 sites, selon la configuration d'enregistrement). Dans ce cas, les avalanches définies à partir des nLFPs peuvent se comporter en loi d'échelle, comme observé précédemment chez le singe. Cependant, les avalanches définies à partir des pics positifs (pLFPs), qui ne sont pas directement reliées aux décharges des neurones, ont le même comportement. Une analyse plus détaillée en utilisant la représentation cumulée (CDF) ne confirme pas la présence de loi de puissance. Les mêmes résultats s'appliquent au chat, au singe et aux enregistrements humains, pendant différents états cérébraux d'éveil et de sommeil. Nous avons également testé des distributions alternatives, et des processus multi-exponentiels semblent expliquer les distributions obtenues, de fac¸on optimale pour des distributions bi-exponentielles. L'ensemble de ces résultats ne montrent pas d'évidence de loi de puissance ou d'états critiques dans le cerveau éveillé ou en sommeil de différents mammifères, du chat 'a l'homme. Finalement, dans un appendice, nous montrons des résultats préliminaires concernant les relations entre cellules excitatrices et inhibitrices, et les potentiels de champ locaux pendant le sommeil humain. Nous avons pu séparer les cellules entre "regular-spiking" (RS) et "fastspiking" (FS), ce qui a été confirmé par connections monosynaptiques (voir Peyrache et al., PNAS, 2012). Nous analysons ici la décharge spécifique des cellules RS et FS pendant différents états d'éveil et de sommeil, sélectionnés sans activité interictale. Jusqu''a 92 unités enregistrées simultanément, procurent une base solide pour la caractérisation de la dynamique de l'excitation et de l'inhibition pendant ces différents états. Pendant le sommeil lent (Stade III ou IV), domine par les ondes lentes de type delta, tous les neurones déchargent selon des états "Up" ou "Down", en relation avec les ondes lentes du LFP, comme décrit précédemment. Les cellules RS et FS sont toutes silencieuses pendant les états "Down". Pendant le sommeil REM et pendant l'éveil, les neurones déchargent de fac¸on irrégulière alors que le LFP ou l'ECoG sont désynchronisées. Dans tous les états les cellules FS déchargent plus que les cellules RS (4 ou 5 fois plus en moyenne). En conclusion, ces résultats procurent une caractérisation des différents rôles de l'excitation et de l'inhibition pendant l'éveil et le sommeil chez l'homme. En conclusion, nous avons utilisé différentes méthodes de mesure, aux échelles microscopiques (activité unitaire), mésoscopique (LFP) etmacroscopiques (ECoG, EEG,MEG), pour caractériser les états de veille et sommeil chez l'homme (ainsi que chez le chat et le singe dans une étude). Nous concluons que le cerveau suit une dynamique complexe à toutes les échelles. Il n'y a pas d'évidence de dynamique auto-organisée critique, mais l'activité du cerveau manifeste d'autres signes d'auto-organisation, comme l'activité synchrone à grande distance et des processus multi-exponentiels. Nous suggérons que ces résultats peuvent être expliques par l'interaction entre excitation et inhibition. Nous anticipons que des réseaux d'oscillateurs couples, avec interaction entre excitation et inhibition, devraient pouvoir expliquer ces résultats. Cette perspective constitue un défi pour des études futures.

Activity-dependent and synapse-specific translation of mRNAs is required for long-term changes in synaptic strength (or efficacy). However, many of the components mediating repression, transport and activation of mRNAs are unknown. Translational control in neurons is a highly conserved process and mediated by a ribonuclear particle (RNP). This study shows that RNPs in Drosophila neurons are similar not only to mammalian neuronal RNA granules but also to yeast P-bodies, cytoplasmic foci involved in translational repression and RNA decay. The evolutionarily conserved proteins Me31b and Trailer Hitch localize to RNA granules. Me31b and Trailer Hitch are required for normal dendritic growth. Mutations in Me31b and Trailer Hitch suppress phenotypes resulting from overexpression of Fragile X Mental Retardation protein, suggesting that both proteins may act as translational repressors. In addition, this study reports the identification of novel translational repressors in neurons. Using the overexpression phenotype of Fragile X Mental Retardation protein in a candidate-based genetic screen, I identified dominant suppressor mutations in five genes, including Doubletime/Discs Overgrown, Orb2/CPEB, PolyA Binding Protein, Rm62/Dmp68 and SmD3. Like Me31b and Trailer Hitch, all five proteins localize to neuronal RNPs. Overexpression of each proteins affects dendritic branching of sensory neurons in Drosophila. Identification and further characterization of these novel RNP granule components and dFMR1-interacting proteins may provide further insights into the mechanisms controlling translational in dendrites.

The aim of this work is to introduce modular processing mechanisms for cortical functions implemented in networks of spiking neurons. Neural maps are a feature of cortical processing found to be generic throughout sensory cortical areas, and self-organisation to the fundamental properties of input spike trains has been shown to be an important property of cortical organisation. Additionally, oscillatory behaviour, temporal coding of information, and learning through spike timing dependent plasticity are all frequently observed in the cortex. The traditional self-organising map (SOM) algorithm attempts to capture the computational properties of this cortical self-organisation in a neural network. As such, a cognitive module for a spiking SOM using oscillations, phasic coding and STDP has been implemented. This model is capable of mapping to distributions of input data in a manner consistent with the traditional SOM algorithm, and of categorising generic input data sets. Higher-level cortical processing areas appear to feature a hierarchical category structure that is founded on a feature-based object representation. The spiking SOM model is therefore extended to facilitate input patterns in the form of sets of binary feature-object relations, such as those seen in the field of formal concept analysis. It is demonstrated that this extended model is capable of learning to represent the hierarchical conceptual structure of an input data set using the existing learning scheme. Furthermore, manipulations of network parameters allow the level of hierarchy used for either learning or recall to be adjusted, and the network is capable of learning comparable representations when trained with incomplete input patterns. Together these two modules provide related approaches to the generation of both topographic mapping and hierarchical representation of input spaces that can be potentially combined and used as the basis for advanced spiking neuron models of the learning of complex representations.

Oxidative stress is a feature of many chronic neurodegenerative diseases as well as a contributing factor in acute disorders including stroke. Fork head class of transcription factors (Foxos) play a key role in promoting oxidative stress-induced apoptosis in neurons through the upregulation of a number of pro-apoptotic genes. Here I demonstrate that synaptic NMDA receptor activity not only promotes Foxos nuclear exclusion but also suppresses the expression of Foxo1 in a PI3K-dependent fashion. I also found that Foxo1 is in fact, a Foxo target gene and that it is subject to a feed-forward inhibition by synaptic activity, which is thought to result in longerterm suppression of Foxo downstream gene expression than previously thought. The nuclear factor (erythroid 2-related) factor 2 (Nrf2) is another transcription factor involved in oxidative stress and the key regulator of many genes, whose products form important intrinsic antioxidant systems. In the CNS, artificial activation of Nrf2 in astrocytes has been shown to protect nearby neurons from oxidative insults. However, the extent to which Nrf2 in astrocytes could respond to endogenous signals such as mild oxidative stress is less clear. The data presented herein, demonstrate for the first time that endogenous Nrf2 could be activated by mild oxidative stress and that this activation is restricted to astrocytes. Contrary to the established dogma, I found that mild oxidative stress induces the astrocytic Nrf2 pathway in a manner distinct from the classical Keap1 antagonism employed by prototypical Nrf2 inducers. The mechanism was found to involve direct regulation of Nrf2's transactivation properties. Overall these results advance our knowledge of the molecular mechanism(s) associated with the control of endogenous antioxidant defences by physiological signals.

The homeobox gene Pitx3 has been implicated as a key regulator for lens development because homozygous mutant aphakia mice, which are hypomorph for Pitx3, fail to develop lenses. One aim of my thesis is to investigate the underlying cellular and molecular mechanism of Pitx3 mediated lens defect by studying knockout mice lacking Pitx3. Chimeric embryos, generated by aggregating the wild type embryos with Pitx3 heterozygous or Pitx3 homozygous mutant ES cells, have been used to analyse lens development. Pitx3 null cells failed to colonise the lens epithelium in Pitx3 null wild type chimeric lens, suggesting that Pitx3 is cell-autonomously required for lens epithelial cells. Further study of Pitx3 null mice revealed an earlier downregulation of the lens epithelial markers PDGFR-alpha and E-cadherin in E11.5 lens epithelium, suggesting the loss of lens epithelial identity in Pitx3 deficient mice. Furthermore, cell cycle inhibitors p27KIP1 and p57KIP2 were ectopically expressed throughout the morphologically normal Pitx3 mutant lens vesicle, suggesting that inactivation of Pitx3 leads to cell cycle exit of epithelial lens cells. In addition, precocious activation of the fibre cell-specific proteins beta- and gamma-crystallins was observed in Pitx3 null lens. Beta-crystallin expression could be observed as early as E10.5 throughout the entire Pitx3 null lens vesicle and gamma-crystallin was detected in the malformed Pitx3 deficient lens at E11.5. RNA in situ hybridisation study revealed that the expression of the transcription factor Foxe3 was lost in Pitx3 null lens at E10.5, suggesting that Pitx3 maintains the lens epithelial cells partly via the regulation of transcription factor Foxe3 during lens development. Accordingly, this study provides the cellular and molecular basis for the lens defect observed in Pitx3 null and Pitx3 hypomorph aphakia mice. Pitx3 is a key transcription factor for the maintenance of lens epithelium and its absence leads to premature activation of fibre cell differentiation programme of lens epithelial cells. In the other part of my PhD, I have further developed the Pitx3-GFP knockin ES cell system with a goal to use this tool for the identification of determinants of midbrain dopaminergic (mDA) neurons, the type of cells lost in Parkinson’s disease (PD) patients. Experimental cell therapy and clinical trials have shown that foetal midbrain tissues, but not tissues from other DA neuron containing regions, can functionally restore the lost mDA neurons when transplanted in Parkinson’s disease patients. Therefore, it is essential to coax mDA properties on stem cell-derived neurons when considering therapeutic development. Within the central nervous system, Pitx3 is expressed exclusively in mDA neurons. Using a Pitx3-GFP knockin mouse line previously generated in the laboratory I have derived heterozygous and homozygous Pitx3-GFP ES cells from mouse blastocysts. In keeping with previous findings in our laboratory, the heterozygous Pitx3-GFP (Pitx3GFP/+) ES cell-derived GFP positive cells of neuronal morphology can be detected after in vitro differentiation using the PA6 coculture system. Furthermore, I have shown that these cells express tyrosine hydroxylase and midbrain markers Engrailed-1 and Nurr-1, demonstrating their midbrain characteristics. I have also generated supertransfectable Pitx3GFP/+ ES cells to offer a rapid and efficient way to express a transgene episomally. The Cre-mediated inducible system of Pitx3-GFP reporter ES cells has also been developed in our laboratory and I have shown that they have high induction efficiency thus allows transgene activation in a temporally controlled manner. The Pitx3 null ES cells showed impaired potential to differentiate into mDA neurons thus they may be used to evaluate candidate Pitx3 downstream target by gain-of-function test. In summary, I have developed a Pitx3-GFP reporter ES cell system to identify mDA regulators functionally by in vitro differentiation.

L’étude des formes familiales de syndrome néphrotique (SN) ou de protéinurie glomérulaire avec lésions histologiques de hyalinose segmentaire et focale (HSF) a permis d’incriminer plus d’une vingtaine de gènes, majoritairement exprimés par le podocyte, cellule principale de la barrière de filtration glomérulaire (BFG). Parmi ces gènes, près d’une dizaine code des régulateurs du cytosquelette d’actine démontrant ainsi le rôle central de la plasticité et de l’architecture du podocyte dans le fonctionnement du filtre glomérulaire. L’ensemble de ces travaux a permis de définir une nouvelle catégorie de maladies nommées podocytopathies héréditaires. Mon projet de thèse a porté sur la caractérisation de plusieurs gènes (TTC21B, WDR73, TRIM3), dont nous avons identifié des mutations dans des cas de podocytopathies héréditaires isolées ou syndromiques. Les résultats du premier volet de ma thèse ont montré que la mutation faux sens p.P209L dans le gène TTC21B induit à l’état homozygote une nouvelle entité clinique associant à la fois une atteinte glomérulaire et une atteinte tubulaire. TTC21B code l’IFT139 (intraflagellar transport protein 139), une protéine impliquée dans le transport protéique antérograde dans le cil primaire, un organite présent à la surface de la plupart des cellules épithéliales. Ces résultats étaient inattendus car l’identification de mutations dans un gène codant une protéine ciliaire n’avait jamais été démontrée auparavant dans des cas de podocytopathies héréditaires, et surtout, il ne semblait pas exister de cil primaire à la surface des podocytes matures. Effectivement, nous avons montré que le cil primaire est présent dans les podocytes humains indifférenciés, mais disparait au cours de la différenciation. Nos résultats ont permis de comprendre l’apparente contradiction entre la survenue d’une pathologie glomérulaire relativement tardive (protéinurie et SN à l’adolescence) et l’absence de cil dans le podocyte mature. En effet, nous avons montré que la mutation p.P209L est une mutation hypomorphe qui induit des défauts mineurs dans la fonction ciliaire, alors qu’elle provoque, dans le podocyte différencié, une déstructuration importante du réseau d’actine et de microtubules du podocyte. Cette étude montre que la protéine ciliaire IFT139, par sa fonction extra-ciliaire, permet de réguler la dynamique des microtubules. Dans le deuxième volet de mon projet, en collaboration avec l’équipe de D Bonneau (Angers), nous avons identifié des mutations tronquantes dans le gène WDR73, dans deux familles non apparentées présentant un syndrome de Galloway-Mowat (SGM), pathologie de transmission autosomique récessive, très hétérogène cliniquement, associant SN et microcéphalie. Ces travaux ont permis d’identifier le premier gène impliqué dans le SGM, dans un sous-groupe de patients présentant un phénotype neurologique très homogène (microcéphalie post-natale, atrophie corticale avec atrophie cérébelleuse majeure, déficience intellectuelle très sévère), alors que l’atteinte glomérulaire est très variable. Ce gène code WDR73, une protéine à motifs WD40. Nos travaux ont montré que la protéine est exprimée dans les neurones du système nerveux central, en particulier dans les cellules de Purkinje du cervelet et dans les podocytes. Des études fonctionnelles nous ont permis de montrer que WDR73 est impliquée dans la survie cellulaire, puisqu’en son absence, une apoptose accrue est observée dans les fibroblastes de patients. De plus, elle est également nécessaire au maintien de la dynamique des microtubules dans les fibroblastes et dans les podocytes différenciés, alors qu’elle ne semble pas avoir de rôle dans la régulation de l’actine. (...) / The genetic study of familial forms of nephrotic syndrome or proteinuria with focal segmental glomerulosclerosis has permitted the identification of 30 causal genes, mainly expressed in the podocyte, which is the principal actor of the glomerular filtration barrier (GFB). Among those genes, approximately ten encode actin cytoskeleton regulators and components, thus highlighting the dramatic role of the podocyte architecture and plasticity in the function of the GFB. During the last decade, all the accumulating results, has made a new category of disease called hereditary podocytopathies. The aim of my thesis project was to characterize the effect of mutations in three candidate genes (TTC21B, WDR73, WDR73), identified by whole exome sequencing in isolated or syndromic podocytopathies. In the first part of my project, we found a homozygous missense mutation (p.P209L) in TTC21B, which encodes a ciliary gene named Intraflagellar transport protein IFT139. This protein ensures the trafficking of components from the tip to the base of the primary cilium, which is an organelle present on most mammalian epithelial cells. These results were unexpected because until now, the existence of the primary cilium was unknown. Our work demonstrates the presence of the primary cilium in the human immature podocyte that disappears once podocytes have differentiated. We also showed that IFT139 localized at the basal body and then relocalized along the complex microtubule network of differenciated cells. We showed that the hypomorphic mutation p.P209L causes minor ciliary defects in undifferentiated cells that are not responsible for the glomerular phenotype. Indeed, the glomerular lesions are rather due to drastic damage in actin and, microtubular dysregulation, found in differentiated podocytes. The second part of my thesis aimed to characterize the effects of truncating mutations identified in the WDR73 gene, found in two families. WDR73 is the first gene identified in Galloway Mowat syndrome by whole exome sequencing combined with homozygous mapping. This rare disease is defined by the association of microcephaly with nephrotic syndrome. In this study, the phenotypes of patients with WDR73 mutations are homogenous concerning neurological features, and are heterogeneous with regards to the renal defects. Thus, WDR73 mutations are responsible for a subset of particular patients affected with Galloway-Mowat syndrome. The WDR73 gene encodes WDR73, a WD-40 containing protein of unknown function. Our studies demonstrated that this protein is expressed in both neurons and podocytes in human tissues. We demonstrated that in undifferentiated cells, WDR73 is weakly expressed in the cytosol, while strong expression and relocalization to the spindle pole, microtubule asters and in the cleavage furrow occur during mitosis. Patient fibroblasts and WDR73-depleted podocytes displayed defects in nuclear morphology, which was associated with a decrease in cell survival in patient fibroblasts. Furthermore, we showed that patient fibroblasts and differentiated WDR73-depleted podocytes harbored an atypical morphology associated with a disorganized microtubule network, suggesting microtubule polymerization defects. Our functional studies demonstrated that WDR73 is crucial in both cell survival and microtubule polymerization in neurons and podocytes. The final part of my PhD work focused on the characterization of a missense mutation in the TRIM3 gene R28W identified by whole exome sequencing in a non consanguineous family with autosomal dominant focal segmental glomerulosclerosis. TRIM3 encodes TRIM3, an E3 ubiquitin-ligase that plays a role in transferrin endosomal recycling, and in microtubule trafficking via KIF21B, one of its known partners. Interestingly, the polymorphism V801M in ACTN4 co-segrates with the disease. Furthermore, mutations in this gene were already incriminated in autosomal dominant cases of HSF. (...)

L’étude des formes familiales de syndrome néphrotique (SN) ou de protéinurie glomérulaire avec lésions histologiques de hyalinose segmentaire et focale (HSF) a permis d’incriminer plus d’une vingtaine de gènes, majoritairement exprimés par le podocyte, cellule principale de la barrière de filtration glomérulaire (BFG). Parmi ces gènes, près d’une dizaine code des régulateurs du cytosquelette d’actine démontrant ainsi le rôle central de la plasticité et de l’architecture du podocyte dans le fonctionnement du filtre glomérulaire. L’ensemble de ces travaux a permis de définir une nouvelle catégorie de maladies nommées podocytopathies héréditaires. Mon projet de thèse a porté sur la caractérisation de plusieurs gènes (TTC21B, WDR73, TRIM3), dont nous avons identifié des mutations dans des cas de podocytopathies héréditaires isolées ou syndromiques. Les résultats du premier volet de ma thèse ont montré que la mutation faux sens p.P209L dans le gène TTC21B induit à l’état homozygote une nouvelle entité clinique associant à la fois une atteinte glomérulaire et une atteinte tubulaire. TTC21B code l’IFT139 (intraflagellar transport protein 139), une protéine impliquée dans le transport protéique antérograde dans le cil primaire, un organite présent à la surface de la plupart des cellules épithéliales. Ces résultats étaient inattendus car l’identification de mutations dans un gène codant une protéine ciliaire n’avait jamais été démontrée auparavant dans des cas de podocytopathies héréditaires, et surtout, il ne semblait pas exister de cil primaire à la surface des podocytes matures. Effectivement, nous avons montré que le cil primaire est présent dans les podocytes humains indifférenciés, mais disparait au cours de la différenciation. Nos résultats ont permis de comprendre l’apparente contradiction entre la survenue d’une pathologie glomérulaire relativement tardive (protéinurie et SN à l’adolescence) et l’absence de cil dans le podocyte mature. En effet, nous avons montré que la mutation p.P209L est une mutation hypomorphe qui induit des défauts mineurs dans la fonction ciliaire, alors qu’elle provoque, dans le podocyte différencié, une déstructuration importante du réseau d’actine et de microtubules du podocyte. Cette étude montre que la protéine ciliaire IFT139, par sa fonction extra-ciliaire, permet de réguler la dynamique des microtubules. Dans le deuxième volet de mon projet, en collaboration avec l’équipe de D Bonneau (Angers), nous avons identifié des mutations tronquantes dans le gène WDR73, dans deux familles non apparentées présentant un syndrome de Galloway-Mowat (SGM), pathologie de transmission autosomique récessive, très hétérogène cliniquement, associant SN et microcéphalie. Ces travaux ont permis d’identifier le premier gène impliqué dans le SGM, dans un sous-groupe de patients présentant un phénotype neurologique très homogène (microcéphalie post-natale, atrophie corticale avec atrophie cérébelleuse majeure, déficience intellectuelle très sévère), alors que l’atteinte glomérulaire est très variable. Ce gène code WDR73, une protéine à motifs WD40. Nos travaux ont montré que la protéine est exprimée dans les neurones du système nerveux central, en particulier dans les cellules de Purkinje du cervelet et dans les podocytes. Des études fonctionnelles nous ont permis de montrer que WDR73 est impliquée dans la survie cellulaire, puisqu’en son absence, une apoptose accrue est observée dans les fibroblastes de patients. De plus, elle est également nécessaire au maintien de la dynamique des microtubules dans les fibroblastes et dans les podocytes différenciés, alors qu’elle ne semble pas avoir de rôle dans la régulation de l’actine. (...) / The genetic study of familial forms of nephrotic syndrome or proteinuria with focal segmental glomerulosclerosis has permitted the identification of 30 causal genes, mainly expressed in the podocyte, which is the principal actor of the glomerular filtration barrier (GFB). Among those genes, approximately ten encode actin cytoskeleton regulators and components, thus highlighting the dramatic role of the podocyte architecture and plasticity in the function of the GFB. During the last decade, all the accumulating results, has made a new category of disease called hereditary podocytopathies. The aim of my thesis project was to characterize the effect of mutations in three candidate genes (TTC21B, WDR73, WDR73), identified by whole exome sequencing in isolated or syndromic podocytopathies. In the first part of my project, we found a homozygous missense mutation (p.P209L) in TTC21B, which encodes a ciliary gene named Intraflagellar transport protein IFT139. This protein ensures the trafficking of components from the tip to the base of the primary cilium, which is an organelle present on most mammalian epithelial cells. These results were unexpected because until now, the existence of the primary cilium was unknown. Our work demonstrates the presence of the primary cilium in the human immature podocyte that disappears once podocytes have differentiated. We also showed that IFT139 localized at the basal body and then relocalized along the complex microtubule network of differenciated cells. We showed that the hypomorphic mutation p.P209L causes minor ciliary defects in undifferentiated cells that are not responsible for the glomerular phenotype. Indeed, the glomerular lesions are rather due to drastic damage in actin and, microtubular dysregulation, found in differentiated podocytes. The second part of my thesis aimed to characterize the effects of truncating mutations identified in the WDR73 gene, found in two families. WDR73 is the first gene identified in Galloway Mowat syndrome by whole exome sequencing combined with homozygous mapping. This rare disease is defined by the association of microcephaly with nephrotic syndrome. In this study, the phenotypes of patients with WDR73 mutations are homogenous concerning neurological features, and are heterogeneous with regards to the renal defects. Thus, WDR73 mutations are responsible for a subset of particular patients affected with Galloway-Mowat syndrome. The WDR73 gene encodes WDR73, a WD-40 containing protein of unknown function. Our studies demonstrated that this protein is expressed in both neurons and podocytes in human tissues. We demonstrated that in undifferentiated cells, WDR73 is weakly expressed in the cytosol, while strong expression and relocalization to the spindle pole, microtubule asters and in the cleavage furrow occur during mitosis. Patient fibroblasts and WDR73-depleted podocytes displayed defects in nuclear morphology, which was associated with a decrease in cell survival in patient fibroblasts. Furthermore, we showed that patient fibroblasts and differentiated WDR73-depleted podocytes harbored an atypical morphology associated with a disorganized microtubule network, suggesting microtubule polymerization defects. Our functional studies demonstrated that WDR73 is crucial in both cell survival and microtubule polymerization in neurons and podocytes. The final part of my PhD work focused on the characterization of a missense mutation in the TRIM3 gene R28W identified by whole exome sequencing in a non consanguineous family with autosomal dominant focal segmental glomerulosclerosis. TRIM3 encodes TRIM3, an E3 ubiquitin-ligase that plays a role in transferrin endosomal recycling, and in microtubule trafficking via KIF21B, one of its known partners. Interestingly, the polymorphism V801M in ACTN4 co-segrates with the disease. Furthermore, mutations in this gene were already incriminated in autosomal dominant cases of HSF. (...)

The original Bio-Rad gene gun was unable to transfect acute or organotypic brain slices, as the amount of helium gas used, the distance for the gold-coated microcarriers to travel to target area were not optimised for fragile tissues, such as the brain. Typically, tissues were severely damaged by a helium shock wave and only a few cells were transfected. It was essential to improve gene gun accuracy by restricting the gold particles from being propelled superficially over a wide area. It was also necessary to increase the amount of DNA or dye delivery into intact tissues. Furthermore, for the gene gun to perform successfully on brain slices the helium gas pressure had to be lowered thereby reducing the degree of cell damage incurred during a biolistic delivery. Without knowing it at the time, the modified gene gun had worked particularly well on a variety of other fragile tissues, and not just the brain. However, the modified gun was not optimised for cultured cells as other transfection methods were available. A particularly notable point of this work was the successful labelling of individual Purkinje dendritic spines from live nerve cells in the cerebellum region of the brain. Biolistic images of Purkinje cells show that the distribution of dendritic spines are not random (O’Brien and Unwin, 2006). Spines were shown to grow in elaborate regular linear arrays, that trace short-pitch helical paths around the dendrites. It was apparent that the spines are arranged to maximize the probability that the dendritic arbour would interact with any afferent axon. This was an important discovery as there has been much debate as to how spines develop on a dendritic shaft. There are three general views to this question, each proposing a theory describing a model for spinogenesis. Classification of the three models in relation to our findings is described in chapter six of this thesis. The Investigation of spine morphology by biolistics was further optimized; gold particles were reduced from a micrometre to forty nanometres (O’Brien and Lummis, 2011), demonstrating that it is possible to use gold-coated DNA nanoparticles of this size to transfect tissue revealing exquisite structural detail. It was possible to observe boutons making synaptic contacts with the pyramidal nerve spines in the hippocampal region of the brain. The findings so far have shown spines from the pyramidal shaft are similar to the spines in the cerebellum, forming regular linear arrays. Recent studies had linked defects in the function of presynaptic boutons to the etiology of several neurodevelopment and neurodegenerative diseases, including autism and Alzheimer’s disease. Our discovery could help to understand why there are abnormalities in dendritic spines which are associated with pathological conditions characterized by cognitive decline, such as mental retardation, Alzheimer’s, stroke and schizophrenia (Yuste and Bonhoeffer, 2001). This thesis provides a synthesis of knowledge about biolistic technology. It is presented as a narrative from improving the gene gun transfection efficiency in brain slices to the development of nano-biolistics. The delivery of DNA and fluorescent dyes into living cells by biolistic delivery should enable a detailed map of the anatomical connections between individual cells and groups of cells to be constructed, providing a “wiring diagram” of connections. The implications of this are discussed in Chapter twelve. The original Bio-Rad gene gun was unable to transfect acute or organotypic brain slices, as the amount of helium gas used, the distance for the gold-coated microcarriers to travel to target area were not optimised for fragile tissues, such as the brain. Typically, tissues were severely damaged by a helium shock wave and only a few cells were transfected. It was essential to improve gene gun accuracy by restricting the gold particles from being propelled superficially over a wide area. It was also necessary to increase the amount of DNA or dye delivery into intact tissues. Furthermore, for the gene gun to perform successfully on brain slices the helium gas pressure had to be lowered thereby reducing the degree of cell damage incurred during a biolistic delivery. Without knowing it at the time, the modified gene gun had worked particularly well on a variety of other fragile tissues, and not just the brain. However, the modified gun was not optimised for cultured cells as other transfection methods were available. A particularly notable point of this work was the successful labelling of individual Purkinje dendritic spines from live nerve cells in the cerebellum region of the brain. Biolistic images of Purkinje cells show that the distribution of dendritic spines are not random (O’Brien and Unwin, 2006). Spines were shown to grow in elaborate regular linear arrays, that trace short-pitch helical paths around the dendrites. It was apparent that the spines are arranged to maximize the probability that the dendritic arbour would interact with any afferent axon. This was an important discovery as there has been much debate as to how spines develop on a dendritic shaft. There are three general views to this question, each proposing a theory describing a model for spinogenesis. Classification of the three models in relation to our findings is described in chapter six of this thesis. The Investigation of spine morphology by biolistics was further optimized; gold particles were reduced from a micrometre to forty nanometres (O’Brien and Lummis, 2011), demonstrating that it is possible to use gold-coated DNA nanoparticles of this size to transfect tissue revealing exquisite structural detail. It was possible to observe boutons making synaptic contacts with the pyramidal nerve spines in the hippocampal region of the brain. The findings so far have shown spines from the pyramidal shaft are similar to the spines in the cerebellum, forming regular linear arrays. Recent studies had linked defects in the function of presynaptic boutons to the etiology of several neurodevelopment and neurodegenerative diseases, including autism and Alzheimer’s disease. Our discovery could help to understand why there are abnormalities in dendritic spines which are associated with pathological conditions characterized by cognitive decline, such as mental retardation, Alzheimer’s, stroke and schizophrenia (Yuste and Bonhoeffer, 2001). This thesis provides a synthesis of knowledge about biolistic technology. It is presented as a narrative from improving the gene gun transfection efficiency in brain slices to the development of nano-biolistics. The delivery of DNA and fluorescent dyes into living cells by biolistic delivery should enable a detailed map of the anatomical connections between individual cells and groups of cells to be constructed, providing a “wiring diagram” of connections. The implications of this are discussed in Chapter twelve.