Despite being the leading cause of neurodegeneration and dementia in the aging brain, the cause of Alzheimer's disease (AD) remains unknown in most patients. The terminal pathological hallmarks of abnormal protein aggregation and neuronal cell death are well-known from the post-mortem brain tissue of Alzheimer's disease patients, but research into the earliest stages of disease development is hindered by limited model systems. In this thesis, an in vitro human neuronal system was derived from induced pluripotent stem (iPS) cell lines reprogrammed from dermal fibroblasts of AD patients and age-matched controls. This allows us to investigate the cellular mechanisms of AD neurodegeneration in the human neurons of sporadic AD (SAD) patients, whose development of the disease cannot be explained by our current understanding of AD. We show that neural progenitors and neurons derived from SAD patients show an unexpected expression profile of enhanced neuronal gene expression resulting in premature differentiation in the SAD neuronal cells. This difference is accompanied by the decreased binding of the repressor element 1-silencing transcription/neuron-restrictive silencer factor (REST/NRSF) transcriptional inhibitor of neuronal differentiation in the SAD neuronal cells. The SAD neuronal cells also have increased production of \(amyloid-\beta\) and higher levels of tau protein, the main components of the plaques and tangles in the AD brain.

京都大学 / 0048 / 新制・課程博士 / 博士(農学) / 甲第19759号 / 農博第2155号 / 新制||農||1039(附属図書館) / 学位論文||H28||N4975(農学部図書室) / 32795 / 京都大学大学院農学研究科応用生命科学専攻 / (主査)教授 植田 充美, 教授 宮川 恒, 教授 栗原 達夫 / 学位規則第4条第1項該当 / Doctor of Agricultural Science / Kyoto University / DFAM

Fundamental to the complexity of the nervous system is the precise regulation in space and time of the production, maturation, and migration of neurons in the developing embryo. This is eloquently seen in the forming cranial sensory ganglia (CSG) of the peripheral nervous system. Placodes, which are transient pseudostratified neuroepithelia in the surface ectoderm of the embryo, are responsible for generating most of the neurons of the CSG. Placodal progenitors commit to the neuronal fate and delaminate from the epithelium as immature, multipolar neuroblasts. These neuroblasts reside in a staging area immediately outside the placode. Differentiation of the neuroblasts is intimately coupled to their adoption of a bipolar morphology and migration away from the staging area to the future site of the CSG. Thus the forming CSG is a highly tractable model to anatomically separate the three phases of a neuroblast’s lifetime: from neuroepithelial progenitor (in the placode), to immature neuroblast (in the staging area), to mature neuron (in the migratory stream). In this thesis, I used the forming CSG as a model to investigate the role of Regulator of G protein Signalling 3 (RGS3) in neuroblast commitment and differentiation. Promoters within introns of the RGS3 locus generate isoforms in which N-terminal sequences are sequentially truncated, but C-terminal sequences are preserved. Intriguingly, I found that expression of these isoforms in the forming CSG is temporally co-linear with their genomic orientation: longer isoforms are exclusively expressed in the progenitor placode; a medium isoform is expressed exclusively in the neuroblast staging area; and the shortest isoforms are expressed in the neuronal migratory stream. Furthermore, through loss- and gain-of-function experiments, I demonstrated that each of these isoforms plays a specific role in the differentiation state in which it is expressed: placode-expressed isoforms negatively regulate neurogenesis; the neuroblast-expressed isoform negatively regulates differentiation; and the neuron-expressed isoforms negatively regulate neuronal migration. The negative regulatory role which all isoforms play in different cell-biological contexts is intriguing in light of the fact that they all share a C-terminal RGS domain, which canonically negatively regulates G protein signalling. Through domain mutation and deletion, I showed that the RGS and N-terminal domains are important for the function of each isoform. Thus temporally co-linear expression within the RGS3 locus generates later-expressed isoforms which lack the regulatory N-terminal domains of the earlier-expressed isoforms, giving them new license to perform different biochemical functions. Lastly, I investigated the conservation and evolution of RGS3 and its isoforms. RGS3 was found to be present in all extant metazoans, and results from this thesis implicate it as the founding member of the R4 subfamily of RGS proteins. Furthermore, in the early vertebrate lineage, a critical domain was lost. This is intriguing in light of the fact that placodes in their stereotypic forms also emerged early in the vertebrate lineage. Ectopic overexpression of the full-length invertebrate RGS3 protein prevented pseudostratification of the vertebrate placode, suggesting that the domain loss in the early vertebrate lineage was important for the evolution of pseudostratified placodes and the expansion of the vertebrate nervous system. In summary, the work in this thesis has uncovered a previously unseen model of transcriptional regulation of a single locus: intragenic temporal co-linearity. Furthermore, the demonstrated functions of this regulation have profound implications on the generation and differentiation of vertebrate neurons, as well as the evolution of the vertebrate nervous system.

Phospholipase C enzymes are important cell signalling enzymes that catalyse the cleavage of phosphatidylinositol 4,5-bisphophate PI(4,5)P₂ into two biologically active second messenger molecules. These are the inositol 1,4,5-trisphosphate which initiates Ca²⁺ release from the endoplasmic reticulum and the diacylglycerol that activates protein kinase C. Although this basic function is shared between the different isoforms, the PLC family encompasses a diverse collection of proteins with various domain structures in addition to the PLC-specific domains. The neuron-specific “6th family” of these enzymes, PLCηs have most recently been identified with two members, PLCη1 and PLCη2. The aim of the thesis is to characterise the PLCη2 variant from several aspects. Firstly, it describes that PLCη2 possesses a high sensitivity towards Ca²⁺. Secondly, it investigates how the Ca²⁺-induced enzymatic activity of PLCη2 is controlled by its different domains. Also it provides evidence that the pleckstrin homology domain targets PLCη2 to membranes by recognising PI(3,4,5)P₃. Moreover, the uniquely structured EF-hand is responsible for the Ca²⁺-sensitivity of the enzyme. Finally, it is demonstrated that the C2 domain is important for activity. The initial biochemical characterisation is followed by the description of a physiological role for PLCη2. It is shown using a neuroblast model that PLCη2 is crucial for neuronal differentiation and neurite growth. Further efforts were made to assess how PLCη2 is responsible for this effect. It was revealed that it might be involved in regulating intracellular Ca²⁺ dynamics, transcriptional activity and actin reorganisation in differentiating neurons. As the functions of PLCη2 are just beginning to come to light, more aspects for future research are also suggested in the thesis. Hopefully, this and the data presented within the thesis will stimulate even greater interest in this fascinating new field of research.

Muitos subtipos de receptores são ativados pelo mesmo ligante, mas estão acoplados a diferentes mensageiros secundários podendo produzir sinalização divergente em uma célula, enquanto receptores ativados por diferentes ligantes, mas que compartilham o mesmo mensageiro secundário, podem produzir sinalização convergente. Para examinar as bases mecanísticas que influenciam a proliferação e a diferenciação celular determinamos as funções de liberação intracelular de Ca2+ e a excitabilidade celular mediada pelos receptores purinérgicos e colinérgicos utilizando imageamento de cálcio por microscopia confocal. Para tanto, caracterizamos a participação dos subtipos P2X1-7 e P2Y1,2,4,6 de receptores purinérgicos aos níveis dos transcritos de mRNA e de expressão protéica, assim como pela atividade de induzir os transientes de [Ca2+]i, aumento na concentração livre de cálcio intracelular, durante a diferenciação neuronal de células P19 de carcinoma embrionário, que foram utilizadas como modelo in vitro para o desenvolvimento neuronal precoce. Em células embriônicas os receptores P2Y1,2, P2X4 ou os heteromultímeros de P2X com farmacologia semelhante ao do receptor P2X4 foram os responsáveis pelos transientes de [Ca2+]i induzidos pelo ATP e seus análogos. Ao término da diferenciação neuronal, os receptores P2Y2,6 e P2X2 foram os principais mediadores das respostas de [Ca2+]i. Obtivemos evidências do envolvimento destes receptores na indução da proliferação tanto de células embriônicas como de progenitores neuronais, por ensaios de incorporação de BrdU, e da indução da diferenciação neuronal das células progenitoras, na presença de vários agonistas e antagonistas de receptores purinérgicos. Como resultado desses estudos, a regulação da proliferação e diferenciação celular foi principalmente devida aos subtipos de receptores P2Y1 e P2Y2, já que estes efeitos foram eliminados após a depleção dos depósitos intracelulares de cálcio e pela demonstração de que estes eram os possíveis receptores funcionais. Entre os receptores colinérgicos, fornecemos evidências para a expressão de receptores nicotínicos (nAChRs) e muscarínicos (mAChRs) funcionais durante a diferenciação de células P19. Detectamos a expressão e a atividade dos subtipos de nAChRs formados pelos subtipos α2-α7, β2, β4 e M1-M3 e M5 de mAChRs durante a diferenciação neuronal. As respostas de [Ca2+]i induzidas pelos agonistas dos nAChRs foram observadas em células P19 embriônicas e neuronais. As respostas de [Ca2+]i mediadas pelos receptores muscarínicos, em níveis próximos aos basais em células embriônicas, aumentaram durante a diferenciação. As elevações na [Ca2+]i induzidas pelos nAChRs em células indiferenciadas foram devidas ao influxo de Ca2+ do meio extracelular. Em células diferenciadas em neurônios, os aumentos de transientes de [Ca2+]i induzidos pelos nAChRs foram parcialmente inibidas após o pré-tratamento das células com a rianodina, enquanto as respostas de [Ca2+]i mediadas pelos mAChR não foram afetadas na presença deste composto, sugerindo uma contribuição da liberação de Ca2+ a partir dos depósitos de Ca2+ sensíveis à rianodina para as elevações mediadas pelos nAChRs. Demonstramos também, que a nicotina, agindo através dos nAChRs, inibiu a proliferação em células embriônicas, porém, a induziu em células progenitoras neuronais pela mobilização de Ca2+ dos depósitos intracelulares. A muscarina induziu em células embriônicas o aumento na proliferação via mAChRs acoplados às proteínas Gαq/11, e promoveu a diferenciação neuronal via M2 mAChRs em células precursoras neuronais. Estes dados sugeriram que a acetilcolina agindo via mAChR funciona como um mitógeno que ativa as proteínas quinases de trifosfato de inositol (IP3) e que poderia estar envolvida na síntese de DNA durante os estágios iniciais da neurogênese. Nós ainda provemos evidências que as oscilações de [Ca2+]i são características para cada estágio da diferenciação e são iniciadas pela liberação de Ca2+ mediada pelo IP3. As análises da determinação do fenótipo neuronal na presença de vários inibidores da transdução do sinal induzido pelo cálcio residem na liberação de Ca2+ induzida pelo IP3 é necessária para o progresso da diferenciação neuronal. Assim, os sinais espontâneos de [Ca2+]i são propriedades intrínsecas das células em diferenciação. A modulação de sua freqüência e amplitudes especifica a aquisição de um fenótipo de célula neuronal. / Various receptors subtypes are activated by the same ligand although coupled to different second messengers. These receptors act either by inducing divergent signaling in one cell, whereas in another cell different receptors may stimulate the very same pathways producing convergent signaling. We have characterized intracellular Ca2+- release and -influx mediated by purinergic and cholinergic receptors using calcium imaging by confocal microscopy to evaluate the mechanistic bases which influence cell proliferation and differentiation We have characterized the participation of purinergic subtypes P2X1-7 and P2Y1,2,4,6 receptor subtypes at mRNA transcription and protein expression levels as well as receptor-induced changes in free intracellular calcium concentration ([Ca2+]i) during differentiation of P19 embryonal carcinoma cells as an in vitro model for early neuronal development. The participation of individual P2X and P2Y receptor subtypes in the differentiation process was studied by employing different available purinergic receptor agonists and antagonists. In embryonic cells, P2Y1,2, P2X4 receptors, or P2X-heteromultimers with similar P2X4 pharmacology were responsible for ATP and ATP-analog-induced [Ca2+]i transients. Following completion of neuronal differentiation, P2Y2,6 receptors and P2X2 subtypes were the major mediators of the [Ca2+]i-response. Regulation of cell proliferation and differentiation of P19 embryonic and progenitor cells was mostly due to P2Y1 and P2Y2 receptor activation, as these effects were abolished following depletion of intracellular calcium stores, and they are probably the unique functional P2Y receptors at these stages of differentiation. We also provide evidence for expression of functional nicotinic (nAChRs) and muscarinic acetylcholine receptors (mAChRs) during neuronal differentiation of P19 cells. We have detected expression and activity of nAChRs formed by the subunits α2-α7, β2, β4, and M1-M3 and M5 mAChR subtypes along the differentiation process. Receptor response in terms of nicotinic agonist-evoked Ca2+ flux was observed in embryonic and neuronal-differentiated cells. However, mAChRs-induced calcium responses, merely present in undifferentiated P19 cells, increased during neuronal differentiation. The nAChR-induced [Ca2+]i response in undifferentiated cells was due to Ca2+ influx. However, in differentiated P19 neurons the nAChR-induced [Ca2+]i response was partially inhibited following pretreatment of the cells with ryanodine, while the mAChR-induced response remained unaffected, suggesting the contribution of Ca2+ release from ryanodine-sensitive stores to nAChR- but not mAChR-mediated Ca2+ responses. The presence of functional nAChRs in embryonic cells suggests that these receptors are involved in triggering Ca2+ waves during initial neuronal differentiation. In the present study we have also shown that nicotine, acting via nAChRs, inhibited proliferation in embryonic cells, but induced cell division of progenitor cells by Ca2+ mobilization from internal stores. Stimulation of progenitor cells by muscarine led to an increase in DNA synthesis mainly resulting from activation of Gαq/11-coupled mAChRs. Muscarine as well promoted differentiation of neural precursor cells by activation of M2 mAChRs subtypes. These data suggest that acetylcholine, acting via mAChRs, functions as a mitogen during early neurogenesis. We also provide evidence that oscillations of [Ca2+]i as characteristics for the respective stage of differentiation are initiated by triphosphate inositol (IP3)-mediated Ca2+-release. Neuronal cell fate determination analysis in the presence of various inhibitors of calcium-induced signal transduction underlined that IP3-mediated Ca2+-release is necessary for neuronal differentiation progress. Thus, spontaneous Ca2+-signals are an intrinsic property of differentiating neural precursor cells. Modulation of their frequency and amplitude is believed to direct the acquisition of a defined neuronal phenotype.

Síndrome de Angelman (AS - MIM 105830) é causada pela ausência de função do gene UBE3A que codifica uma proteína ubiquitina - ligase (E6-AP). Esse gene é expresso bialelicamente em vários tecidos exceto no cérebro, onde a expressão é preferencialmente materna. O RNA anti-senso de UBE3A é considerado o regulador dessa expressão diferencial entre os alelos, e faz parte de um transcrito grande que só o alelo paterno é expresso devido ao imprinting genômico; no cérebro, esse transcrito se entende até a região anti-senso de UBE3A, mas nos demais tecidos o transcrito é menor e não engloba a região anti-senso. Este trabalho visa obter um modelo para estudo da AS. Células-tronco da polpa do dente (SHEDs) de pacientes com deleção do segmento 15q11-q13 ou mutação no gene UBE3A foram caracterizadas e submetidas à diferenciação neuronal. A diferenciação foi analisada através do estudo de RNA e proteínas para marcadores neuronais e, também, por testes funcionais. As SHEDs são células-tronco mesenquimais e constituem uma população heterogênea. Essas células ou algumas dessas células já expressam algumas proteínas neuronais ou de células excitáveis como nestina, β-tubulina III, MAP2 e proteína de canais dependentes de voltagem de sódio e potássio. Um ponto interessante é que as SHEDs apresentam baixa expressão do UBE3A anti-senso e a expressão do UBE3A nas células de pacientes é menor que 50% da expressão encontrada nas células de controles, que pode indicar a ocorrência de expressão preferencial materna desse gene em outros tipos celulares além de neurônios maduros. Quando induzidas à diferenciação neurogênica, a maioria das linhagens controles apresentou aumento da expressão de MAP2 e, principalmente, β-tubulina III; e a maioria das linhagens de pacientes com AS não apresentou aumento notável na expressão dessas proteínas, exceto uma linhagem de paciente que aumentou a expressão de β-tubulina III. As células induzidas à diferenciação apresentaram aumento estatisticamente significativo da condutância de sódio através de canais de sódio dependentes de voltagem. Com a análise de expressão de UBE3A e do UBE3A anti-senso é possível afirmar que a expressão deles não alterou com a diferenciação neuronal. Assim, é possível concluir que as células-tronco da polpa do dente, com o protocolo de diferenciação neurogênica, progrediram na via de diferenciação, mas a maioria das células não atingiu o estágio de maturação necessário para que ocorresse o imprinting do UBE3A ou a via de diferenciação não ia em direção a neurônios que apresentam imprinting do UBE3A. / Angelman syndrome (AS - MIN 105830) is caused by the loss of function of the maternal UBE3A gene, which encodes an ubiquitin protein ligase (E6-AP). UBE3A displays biallelic expression in most of tissues, but maternal predominant expression is observed in the brain. A RNA antisense that is paternally expressed in some regions in the brain is considered to be responsible for this tissue-specific imprinting; UBE3A antisense is part of a large transcript that starts at SNURF-SNRPN gene and is paternally expressed, and in the brain this transcript includes UBE3A antisense region however in other tissues this region is not included. The aim of the present study is to develop a new model for studying AS. Dental pulp stem cells (SHEDs) were characterized and differentiated by an already described protocol. SHEDs intrinsically express some neuronal proteins as nestin, β-tubulin III, MAP2 and voltage-gated sodium channels and potassium channels. Interestingly, SHEDs also present a low expression of UBE3A antisense, and UBE3A expression in cells from patients with AS is lower than 50% of the cells from normal control, so it is possible that preferential maternal expression of this gene might occur in some cells beyond mature neurons. After the neuronal differentiation, most control lineages and one lineage of AS patients had an increase of MAP2 and β-tubulin III expression. Two control lineages and most lineages from AS patients did not have a notable increase of expression of these proteins. Neuronal differentiated cells displayed an increase in conductance through voltage-gated sodium channels. Analysis of UBE3A and UBE3A antisense expression in SHEDs and cells induced to differentiate into neurons indicated no changes in their expression. Thus, after neuronal differentiation induction, dental pulp stem cells progressed through neuronal differentiation pathway. However, most cells did not reach the stage which UBE3A imprinting occurs or the neuronal differentiation is resulting in a cell that do not present UBE3A imprinting.

Células tronco mesenquimais de tecido adiposo, são uma promissora ferramenta para aplicações clínicas em terapias celular e regenerativa, em vista da facilidade de sua extração e da maior quantidade de células por unidade de massa de tecido quando comparado a outras fontes clássicas de células mesenquimais como medula óssea. O protocolo clássico de extração e purificação dessas células, depende de sua adesão em plástico e xeno-materiais demandando muito tempo para ser utilizado por médicos para auxiliar pacientes em procedimentos de emergência. Estas células são capazes se diferenciar em diversos tipos celulares, o que as torna boas candidatas para terapia celular, embora sua capacidade de transdiferenciação para fenótipos neuronais seja ainda discutida. Neste trabalho demonstramos um novo processo para isolar essas células na base de epitopos específicos expressos (assinatura molecular de superfície) utilizando aptâmeros como ligantes de alta afinidade para estes sitios. Aptâmeros, moléculas de DNA simples fita identificadas a partir de uma biblioteca combinatória de sequencias de DNA simples-fita foram identificados por ciclos reiterativos de seleção in vitro (SELEX) utilizando células tronco do lipoaspirado como alvo. Dois aptâmeros isolados, denominados APT9 e APT11, foram capazes de identificar subpopulações (15,8 e 23,7% respectivamente) dentre as células tronco mesenquimais (classicamente CD29+/CD90+/CD45-) e separá-las usando nano-partículas magnéticas acopladas aos aptâmeros. Além disso, seguindo uma indução para diferenciação neuronal, as células tronco mesenquimais passam a apresentar morfologia neuronal e apresentam expressão e atividade de diversos receptores de neurotransmissores, avaliados por PCR real-time e imageamento de variações da concentração de cálcio intracelular ápos stimulação com vários agonistas de receptores metatrópicos e ionotrópicos. Ao longo da diferenciação, os níveis transcricionais de mRNA de receptores de cininas (B1 e B2), nicotínicos (alfa 7), muscarínicos (M1, M3 e M4), glutamatérgicos (AMPA2 e mGluR2), purinérgicos (P2Y1 e P2Y4) e GABAergicos (GABA-A, subunidade 3) e da óxido nítrico sintase neural aumentaram quando comparados aos níveis das células não diferenciadas, enquanto que os níveis de expressão de outros receptores incluindo purinérgicos P2X1, P3X4, P2X7 e P2Y6 e muscarínico M5 diminuíram. Os níveis de atividade das classes dos receptores estudados, por imageamento de variações da concentração de cálcio intrac, aumentaram para a maioria dos agonistas analisados durante a diferenciação neuronal com exceção para respostas induzidas por glutamato e NMDA. Células diferenciadas expressavam altos níveis de antígenos específicos de neurônios como β3-tubulina, NF-H, NeuN e MAP-2 indicando uma diferenciação em fenótipo neuronal bem sucedida. Desta maneira, esta tese, ao identificar aptâmeros, prove uma inovadora solução para médicos usarem as células tronco mesenquimais dentro de uma sala de cirurgia, através de um método que é capaz de purificar essas células em um tempo clínico viável, com pureza e sem contato com contaminantes. Além disso, nós mostramos aqui que com um protocolo como o proposto para diferenciação neuronal, nós poderíamos induzir essas células para se diferenciar em neurônios, através da ativação de fatores de transcrição específicos, levando às células tronco mesenquimais a serem possivelmente utilizadas em terapias celulares de reparo neuronal. / Adipose mesenchymal stem cells are promising tools for clinical applications in cellular and regeneration therapies, in view of easiness of extraction and higher amount of isolated stem cells per mass of tissue when compared to other classical mesenchymal stem cell sources including bone marrow. The classical protocol to extract and purify these cells, depending on plastic adherence and xeno-materials, is too time consuming to be used by physicians to help patients at emergency procedures. These cells are able to differentiate into various cell types, making them good candidates for cell therapy, however their capability for transdifferentiation into neural phenotypes is yet discussed. Here we show a novel process to isolate these cells using their surface molecular signature and aptamers, ssDNA molecules identified through the SELEX technique, denominated APT9 and APT11 that are able to identify subpopulations (15,8 and 23,7% respectively) within the mesenchymal stem cells (classically CD29+/CD90+/CD45-) and separate them using magnetic nano-particles attached to the aptamers. Moreover, following induction to neural differentiation, mesenchymal cells presents neuronal morphology and present expression and activity of several neurotransmitter receptors, as evaluated by real-time PCR and calcium imaging. During this process, mRNA transcription levels of bradykinin (B1 and B2), cholinergic (alpha 7), muscarinic (M1, M3 and M4), glutamatergic (AMPA2 and mGlu2), purinergic (P2Y1 and P2Y4) and GABAergic (GABA-A, subunit 3) receptors and neuronal nitric oxide synthase were augmented when compared to levels of undifferentiated cells, while the expression levels of other receptors including purinergic P2X1, P2X4, P2X7 and P2Y6 and muscarinic M5 receptors were down-regulated. Activity levels of the studied receptor classes, as studied by calcium imaging, increased for most of the agonists analyzed during the neuronal differentiation with the exception for glutamate- and NMDA-induced receptor responses. Differentiated cells expressed high levels of neuron-specific antigens such as β3-tubulin, NF-H, NeuN and MAP-2, indicating a successful differentiation into neuronal phenotypes. This thesis, by identifying aptamers, provides a novel solution for physicians to use mesenchymal stem cells inside a surgery room, by using a method that are able to purify the cells in a clinical viable time, with purity and no contact with contaminats. Furthermore, we show here that with a protocol as provided for neuronal differentiation, we could induce these cells to differentiate into neurons, by activating specific transcription factors,making mesenchymal stem cells to possibly be used in neuronal repair cell therapies.

Les p21-activated kinases (PAK) du sous-groupe I sont impliquées dans de nombreux processus cellulaires tels la prolifération, les mouvements cellulaires, l’adhérence et l’apoptose. Ces kinases sont des effecteurs des Rho-GTPases Rac1 et Cdc42 et participent à la régulation du cytosquelette d’actine. Les deux kinases neuronales PAK1 et PAK3, qui présentent de fortes identités de séquence, régulent le cytosquelette d’actine, contrôlant ainsi la dynamique des épines dendritiques, et la plasticité synaptique.Les mutations du gène pak3, localisé sur le chromosome X, sont responsables de déficience intellectuelle chez l’homme, et les mécanismes moléculaires et cellulaires associés aux défauts cognitifs sont mal connus. Il a été montré que PAK3 participe à la voie proneurale au cours de l’embryogénèse précoce du xénope en favorisant la sortie du cycle cellulaire et la différenciation neuronale. Cependant, le rôle de PAK3 dans la neurogenèse adulte n’a pas été étudié. Or depuis maintenant une quinzaine d’années, il est admis que la neurogenèse perdure à l’âge adulte et participe aux processus de mémorisation et d’apprentissage. Nous nous sommes donc intéressés à l’implication de PAK3 dans la régulation de la neurogenèse adulte, posant l’hypothèse qu’un défaut de neurogenèse serait responsable, au moins en partie, des défauts cognitifs chez les patients. Nous avons montré que PAK3 n’est pas exprimée dans les cellules souches neurales/progéniteurs prolifératifs mais son expression augmente fortement dès le retrait des facteurs de croissance, ex vivo, suggérant un rôle dans la neurogenèse adulte. Nous avons montré que l’invalidation de pak3 provoque une augmentation de la fréquence de neurosphères primaires formées ainsi qu’un accroissement de leur taille, ceci sans affecter la taille du réservoir de cellules souches ni les propriétés cardinales de celles-ci (multipotence, auto-renouvellement et prolifération). Toutefois, les cellules progénitrices pak3- poursuivent leur prolifération dans des conditions de culture induisant normalement la différenciation, suggérant un défaut de sortie du cycle cellulaire.Nous nous sommes ensuite demandé si les mutations de déficience intellectuelle du gène pak3 altèrent la neurogenèse adulte. Nous avons créé pour cela un modèle murin portant la mutation R67C, responsable chez l’homme de la forme la plus sévère de déficience intellectuelle associée aux mutations de ce gène. Nous mettons en évidence, dans cette souris knock-in, une forte diminution du nombre de cellules nouveau-nées dans les deux zones neurogéniques du cerveau (la zone sous-ventriculaire et le gyrus denté de l’hippocampe) et une augmentation de la proportion de neurones nouveau-nés immatures. Ces données suggèrent que la mutation R67C n’induit pas une perte de fonction de la kinase mais un changement de fonction dépendante d’une activation préférentielle par la GTPase Rac1.En conclusion, ce travail de thèse montre que PAK3 participe à la régulation de la neurogenèse adulte chez les mammifères, contrôle la sortie du cycle cellulaire des progéniteurs neuraux et que la mutation R67C impacte la maturation des neurones nouveau-nés. L’ensemble de ces données suggère que les défauts de neurogenèse adulte dus aux mutations de déficience intellectuelle du gène pak3 sont à l’origine de certains dysfonctionnements cognitifs. / The group I p21-activated kinases (PAK) are involved in many cellular processes such as proliferation, cell movement, adhesion and apoptosis. These kinases are effectors of Rho GTPases Rac1 and Cdc42, and participate in the regulation of the actin cytoskeleton. Both neuronal kinase PAK1 and PAK3, which exhibit high sequence identities, regulate the actin cytoskeleton, thereby controlling the dynamics of dendritic spines and synaptic plasticity. Mutations of the X-linked pak3 are responsible for intellectual disability (ID) in humans, and the molecular and cellular mechanisms associated with cognitive defects are poorly described. It was shown that PAK3 participates in the proneural pathway during early Xenopus embryogenic development, by promoting cell cycle exit and neuronal differentiation of neural precursors. However, the role of PAK3 in the adult neurogenesis has not been studied in mammals. It is now generally accepted that neurogenesis persists during human adulthood and is involved in learning and memory. We are therefore interested in the involvement of PAK3 in the regulation of adult neurogenesis, on the assumption that defects in neurogenesis may be responsible, at least in part, for cognitive defects in ID patients.We showed that PAK3 is not expressed in proliferative neural stem/progenitor cells but its expression increased significantly upon growth factor removal, suggesting a role in adult neurogenesis. We showed that the invalidation of pak3 gene causes an increase in the frequency and in size of primary neurospheres. However Pak3 invalidation does not affect the size of the stem cell reservoir nor the NCS cardinal properties (pluripotency, self-renewal and proliferation). However, the pak3- progenitor cells continue their proliferation in culture conditions normally inducing differentiation, suggesting a defect in cell cycle exit. We then asked whether pak3 ID mutations affect adult neurogenesis. We created a knock-in model expressing the pak3-R67C mutation responsible in humans for a severe form of intellectual impairment. We observed in the knock-in mice, a significant decrease in the number of newborn cells in both neurogenic areas of the brain (the subventricular zone inforebrain, and the dentate gyrus of the hippocampus) and an increase in the proportion of immature newborn neurons. These data suggest that the R67C mutation does not induce a loss of function of the kinase but a change of a function dependent on preferential activation by the Rac1 GTPase.In conclusion, we show that PAK3 play an important role in the regulation of adult neurogenesis in mammals by controlling the cell cycle exit of neural progenitors. The R67C ID mutation impacts both newborn cell proliferation and their maturation. Taken together, these data suggest that defects in adult neurogenesis caused by ID mutations in the pak3 gene may be involved in some cognitive dysfunctions.

Emerging trends in cell-therapy based tissue repair have focused on the renewable source of adult stem cells including human bone marrow-derived mesenchymal stromal cells (hMSCs). Due to immunomodulatory properties as well as a potential to differentiate into cells characteristic of all three germ layers, hMSCs provide a source of immature cells for utilization in cell-therapy based treatments. Marrow isolated adult multilineage inducible (MIAMI) cells are a homogeneous sub-population of hMSCs which maintain self-renewal potential during ex vivo expansion, in addition to efficiently undergoing trans-differentiation into neuron-like cells in vitro. Even though hMSCs have the potential to be used for neural tissue repair, the molecular mechanisms by which they are stimulated to become neuron-like cells have not been fully characterized. Therefore the work described herein focuses on the molecular mechanisms by which MIAMI cells undergo NT-3 dependent neuronal commitment. MIAMI cells express both the full length (FL-) and tyrosine kinase deficient (TKd-) isoforms of the NTRK3 receptor, the primary NT-3 receptor, at the protein level. NT-3 stimulation of MIAMI cells during neuronal commitment induced the phosphorylation of FL-NTRK3, degradation of TKd-NTRK3, downstream activation of the Mek1/2-Erk1/2 signaling cascade, and subsequent up-regulation of a limited number of pro-neuronal genes. These findings were verified using chemical inhibitors to block NTRK autophosphorylation (K252a) and Erk1/2 activation (U0126). TKd-NTRK3 is hypothesized to activate Rac1 upon NT-3 stimulation. Rac1 was found to suppress NT-3 stimulated Erk1/2 phosphorylation, as well as downstream gene expression, as determined using a Rac1 chemical inhibitor. Further characterization confirmed that Rac1b is the predominant Rac1 isoform in MIAMI cells. Rac1b siRNA mediated knock-down resulted in increased expression of the pro-neuronal genes NGN2, MAP2, NFH and NFL during NT-3 stimulation via regulation of Mek1/2-Erk1/2. Rac1b is also involved in NT-3 stimulated cell proliferation, as well as repression of CCND1 and CCNB1 mRNA expression. In an attempt to enhance neuronal differentiation of MIAMI cells, EGF and bFGF were used to pretreat MIAMI cells prior to NT-3 stimulated neuronal commitment. EGF/bFGF pretreatment increased NTRK3 and NTRK1 protein levels along with NT-3 stimulated Erk1/2 phosphorylation. In addition, bFGF versus EGF/bFGF pretreatment restricted the expression of the pro-neuronal transcription factors Ngn2 and Prox1 versus the neural stem cells self-renewal transcription factor Musashi-1, respectively. The culmination of this work provides a model for the NT-3 induced neuronal commitment of MIAMI cells in vitro, as well as insight into the neurogenic potential of MSCs for future applications in cell-therapy based tissue repair.

In eukaryotes, the first step of interpreting the genetic information is the transcription of DNA into RNA. For protein-coding genes, such transcription is carried out by RNA polymerase II. A special domain of RNA polymerase II, called the C-terminal domain (CTD), functions as a master controller for the transcription process by providing a platform to recruit regulatory proteins to nascent mRNA (Chapter 1-2). The modifications and conformational states of the CTD, termed the 'CTD code', represent a critical regulatory checkpoint for transcription. The CTD, found only in eukaryotes, consists of 26--52 tandem heptapeptide repeats with the consensus sequence, Tyr₁Ser₂Pro₃Thr₄Ser₅Pro₆Ser₇. Phosphorylation of the serines and prolyl isomerization of the prolines represent two major regulatory mechanisms of the CTD. Interestingly, the phosphorylation sites are typically close to prolines, thus the conformation of the adjacent proline could impact the specificity of the corresponding kinases and phosphatases. Understanding how those modifying enzymes recognize and regulate the CTD is important for expanding our knowledge on the transcription regulation and deciphering the 'CTD code'. During my PhD study, I studied the function of CTD phosphatases and prolyl isomerase in the CTD regulation using Scp1, Ssu72 and Pin1 as model regulators. Scp1 and Ssu72 are both Ser5 phosphatases. However, Ssu72 is an essential protein and regulates the global transcription while Scp1 epigenetically silences the expression of specific neuronal genes. Pin1 is a highly conserved phosphorylation-specific prolyl isomerase that recognizes the phospho-Ser/Thr-Pro motif within the CTD as one of its primary substrates in vivo. Among these enzymes, Scp1 is the focal point of this dissertation, as it was studied from different angles, such as enzymatic mechanism (Chapter 3 describes the capture of phospho-aspartyl intermediate of Scp1 as a direct evidence for the proposed two-step mechanism), specific inhibition (Chapter 4 describes the identification and characterization of the first specific inhibitor of Scp1), and its non-active-site contact with the CTD (Chapter 5 describes the structural basis of this contact). These studies are of great importance towards understanding the molecular mechanism of the dephosphorylation process of the CTD by Scp1. / text