Segundo a Organização Mundial da Saúde, em 2018, entre 6 e 7 milhões de pessoas estavam infectadas pelo parasita Trypanosoma cruzi, agente etiológico da Doença de Chagas. Dentre as piores consequências para os chagásicos estão os problemas cardíacos causados pela infecção, os quais afetam cerca de 25% dos pacientes na fase crônica da doença. Na busca por biomoléculas envolvidas no processo de invasão celular pelo parasita, uma proteína produzida por ele, denominada P21, foi encontrada. Ensaios biológicos mostraram que a proteína P21 de T. cruzi interage com o receptor de quimiocinas CXCR4 e desencadeia diversos processos bioquímicos, como: indução de fagocitose por macrófagos, indução da polimerização de actina e inibição de angiogênese. Os dados sugerem que a P21 também pode desempenhar um papel na cardiomiopatia induzida pelo parasita. Visando contribuir com a caracterização biofísica e estrutural da proteína P21, foi desenvolvido um protocolo para sua obtenção a partir de corpos de inclusão e reenovelamento, o qual forneceu quantidades necessárias da proteína para os experimentos. Espectros de ressonância Magnética Nuclear (RMN) permitiram avaliar de forma precisa a qualidade estrutural da amostra reenovelada frente a proteína P21 produzida de forma solúvel. Resultados de RMN mostraram também que a proteína possui cinco hélices-α em sua estrutura secundária, uma grande porção desestruturada, além de duas populações conformacionais em equilíbrio. Ainda, corroborando com os dados biológicos encontrados na literatura, um ensaio de interação in vitro indicou que a P21 interage com o receptor CXCR4, sendo a porção N-terminal deste receptor a região envolvida em tal processo. / According to the World Health Organization, in 2018, between 6 and 7 million people were still infected by the parasite Trypanosoma cruzi, the causative agent of Chagas disease. One of the worst aggravations for chagasics is heart disorder caused by the infection, which affects about 25% of patients in the chronic phase of the disease. Searching for biomolecules involved in the parasitic invasion of human cells, a T. cruzi protein called P21 was found. It has a high probability of being secreted or anchored to the membrane. Biological assays have shown that T. cruzi P21 interacts with the CXCR4 chemokine receptor and mediates several biochemical processes, such as: induction of phagocytosis by macrophages, induction of actin polymerization and inhibition of angiogenesis. The data suggests that P21 may also play a role in cardiomyopathy induced by parasite. In order to contribute to the biophysical and structural characterization of P21, a refolding protocol was developed to obtain P21 from inclusion bodies. The protocol provided the necessary amounts of protein required for the experiments. Nuclear Magnetic Resonance (NMR) spectra allowed to accurately evaluate the structural quality of the refolded sample in comparison to the produced soluble P21 protein. NMR spectra also showed five α-helices in the P21 secondary structure, a large unstructured portion and two ensembles of structures in equilibrium. Furthermore, corroborating with the biological data found in the literature, an in vitro interaction assay indicated that P21 interacts with CXCR4 via the N-terminal portion of the receptor.

5 mutations responsables de retard mental ont été identifiées dans le gène p21-activated kinase 3 (pak3). Nous avons récemment identifiés dans pak3 deux exons alternatifs très conservés appelés b et c. Ainsi, en plus du variants PAK3a (dépourvu des inserts b ou c), le gène pak3 code pour 3 nouveaux variants d’épissage PAK3b, PAK3c et PAK3cb qui sont constitutivement actifs et insensibles aux GTPases. De plus, contrairement à PAK1 et PAK3a, leur domaine d’auto-inhibition est incapable d’inhiber un domaine kinase. Ainsi, le but de ce projet était de comprendre le mécanisme de régulation de la kinase PAK3. Un modèle de régulation a récemment été proposé dans lequel PAK1 forme des homodimères pouvant être dissociés par les GTPases, permettant ainsi l’activation de la kinase. En se basant sur ces observations j’ai cherché à identifier les dimères PAK3 et j’ai montré que les kinases PAK3a, b, c et cb forment préférentiellement des hétérodimères avec PAK1. J’ai démontré l’existence de ces dimères dans le cerveau et j’ai mis en évidence que ces hétérodimères permettent à chaque monomère de réguler l’activité kinase de son partenaire in vitro. Ce travail permet de proposer un modèle de régulation symétrique pour PAK3a qui forme des hétérodimères avec PAK1 et un nouveau modèle de régulation asymétrique pour les variants d’épissage, également basé sur leur hétérodimérisation avec PAK1. Mes résultats montrant une corégulation des kinases PAK neuronales suggèrent d'une part que leur activation puisse être synchronisée et d'autre part que dans certaines situations physiopathologiques (Cancer et maladies neurologiques) leur dérèglement puissent interférer. / 5 mutations responsible for mental retardation have been identified in p21-activated kinase 3 (pak3) gene. We recently identified in pak3, two highly conserved alternative exons called b and c. In addition to the classical PAK3a variant (without any alternative exon), the pak3 gene encodes 3 new splice variants PAK3b, PAK3c and PAK3cb which are constitutively active and insensitive to GTPase activation. Moreover, unlike PAK1 or PAK3a, their autoinhibitory domain is unable to inhibit a kinase domain. The aim of this project was to understand how PAK3 regulation occurs. A model of regulation was recently proposed in which PAK1 forms homodimers that can be dissociated through GTPase binding, leading to kinase activation. Given these observations, I searched to identify PAK3 dimers and I showed that PAK3a, b, c and cb preferentially form heterodimers with PAK1. I demonstrated the existence of such dimers in the brain and that the different heterodimers allow each monomer to regulate the kinase activity of its partner. Through this study, I propose a symmetric regulation model for PAK3a which heterodimerizes with PAK1 and a new asymmetric regulation model for splice variants, also based on heterodimerization with PAK1. My results showing a co-regulation of neuronal PAK kinases suggest that their activation may be synchronized but also that, in some physiopathological situations (cancers and neurologic diseases), their misregulation may interfere.

Le gène CDKN1A a été initialement décrit comme une cible transcriptionnelle de la protéineoncosuppressive p53. Son produit, p21Cip1 (p21), supprime l’activité des kinases dépendantes descyclines et de la protéine PCNA, ce qui en fait un puissant inhibiteur du cycle et de la proliférationcellulaires. En outre, p21 est fréquemment inactivée dans les cancers épithéliaux. Or, la progressionde ces tumeurs est associée à l’altération de l’organisation tissulaire, au processus invasif et à ladissémination métastatique. Ces phénomènes résultent des modifications cytosquelettiquesconduisant à la transformation des propriétés d’adhésion et de migration cellulaires. Pourtant, le rôlede p21 dans la dynamique cytosquelettique des cellules épithéliales humaines n’a jamais été adressé.Nous montrons ici que p21 contribue à l’adhésion et la migration normale des cellules épithélialesmammaires non transformées. Nos résultats montrent également que l’inactivation de p21 provoquela suppression de l’adhésion focale et des fibres de stress. Ce phénotype est caractérisé parl’inactivation de la GTPase Rho et l’activation de la cofiline, facteur de dépolymérisation de l’actine. Enoutre, la suppression de p21 provoque une désacétylation des microtubules associée à unedéstabilisation microtubulaire globale. La réduction de l’instabilité dynamique, par inhibition de ladésacétylase HDAC6, restaure partiellement l’étalement cellulaire et l’adhésion focale altérés parl’inactivation de p21. L’ensemble de nos données démontre que la régulation de la dynamiquecytosquelettique par p21 est nécessaire au contrôle de l’adhésion des cellules épithéliales humainesnon tumorales. / CDKN1A gene was initially identified as a target of the tumor suppressor p53. The ability to inhibitcyclin-dependant kinase and PCNA activities confers to its product, p21Cip1 (p21), strong antiproliferativeproperties. Moreover, p21 is frequently inactivated in epithelial cancers. Progression ofthese malignancies is also associated with disorganized tissue architecture, invasion and metastaticdissemination. These processes involve major cytoskeletal reorganization associated withtransformation of adhesion and migration properties. Still, the role of p21 in cytoskeletal dynamics innormal epithelial cell has never been addressed. Here we show that p21 contributes to normaladhesion and migration in untransformed human mammary epithelial cells (HMEC). We alsodemonstrate that p21 inactivation in HMEC suppresses focal adhesion and stress fiber assembly. p21depletion is also associated with inactivation of Rho GTPase and activation of the F-actin severingfactor cofilin. In addition, p21 silencing provokes microtubule hypoacetylation associated withincreased dynamic instability. We find that forced microtubule stabilization by HDAC6 inhibitionpartially restore cell spreading and focal adhesion in p21-depleted cells. Collectively, our data showthat regulation of cytoskeletal dynamics by p21 is required for adhesion control in untransformedhuman epithelial cells.

Les gliomes sont les tumeurs cérébrales les plus fréquentes chez l’adulte. Il s’agit d’un groupe de tumeurs hétérogène pouvant être classées dans différents sous-groupes selon des critères moléculaires et histopathologiques, donnant une indication sur leur agressivité. Les p21-activated kinase (PAK) sont des sérine-thréonine kinases effectrices des GTPases Rac et Cdc42 et font ainsi partie de nombreuses voies de signalisation. Elles régulent de nombreuses voies d’intérêt dans la biologie des cancers, comme les voies Mek/Erk, PI3K/Akt et Wnt/b-caténine. PAK1 est fréquemment surexprimé et/ou suractivé dans plusieurs cancers, notamment du sein, des ovaires de la prostate et du cerveau, alors que PAK3 induit la sortie du cycle cellulaire et la différenciation des cellules au cours de la neurogenèse dans plusieurs modèles animaux. Durant ma thèse, j’ai pu observer que l’expression de PAK3 était de bon pronostic pour les patients atteints de gliomes, et était plus élevée dans un sous-type de tumeurs caractérisé par la codélétion 1p/19q. In vitro, l’expression de PAK3 était plus élevée dans une lignée de cellules de gliome de signature proneurale, non tumorigène après xenogreffe chez la souris. Au cours de la différenciation des cellules de gliome, l’augmentation de l’expression de PAK3 est associée avec des marqueurs de lignage neural et neuronal. L’inhibition de l’expression de PAK3 entraîne l’augmentation de la capacité d’auto-renouvellement et de la tumorigénicité d’une lignée de cellules de gliome, et favorise la différenciation gliale des cellules. / Gliomas are the most common and lethal adult primary brain tumors. Their complex heterogeneity is evidenced by numerous genomic studies showing distinct molecular entities in glioma. P21-activated-kinases (PAK) are serine threonine kinases involved in multiple signal transduction pathways as downstream effectors of Rac and Cdc42. They regulate several key cancer-relevant pathways like cell division and movement. PAK1 and PAK3 are highly expressed in the brain; PAK1 is frequently overexpressed and/or over-activated in several human cancers whereas PAK3 is involved in neural differentiation and the developmental proneural pathway. The role of these two kinases in brain tumor pathophysiology is unknown. We have observed that PAK3 expression was associated with a longer survival for patients with glioma and was higher in 1p/19q gliomas. In vitro, PAK3 was highly expressed in a glioma cell line with a proneural signature that did not induce tumor after xenograft. Its increasing expression upon a set of differentiation paradigms was correlated with those of neural and/or neuronal markers in glioma cell lines. Inhibition of PAK3 expression increased cell renewal and tumorigenicity. It impaired cell differentiation, promoting the glial lineage.

La dérégulation du système de surveillance qui bloque la prolifération lorsque l'intégrité du génome est compromise fait partie intégrante de la cancérogenèse. Nous cherchons à décortiquer les mécanismes qui, en phase G2, orchestrent l'arrêt du cycle cellulaire, irréversible, en présence des lésions de l'ADN (sénescence) ou réversible (quiescence), en absence de signaux mitogéniques ou confluence. L'objectif du premier volet fut d'élucider les rôles respectifs de l'inhibiteur de CDK (CKI) p21Waf1 et des kinases Chk1 et Chk2 dans l'arrêt en G2 dû au stress génotoxique menant à la sénescence. Nous avons montré que dans les cellules humaines normales cet arrêt nécessite l'action de p21 et Chk1 tandis que Chk2 n'est pas requise. Au contraire, dans plusieurs lignées cancéreuses, malgré la présence de p53, ce rôle de p21 est compromis à cause d'une activation inefficace de la kinase ATM. Par conséquent, en dépit d'une forte activation de Chk1 bloquant la mitose, ces cellules ne parviennent pas à initier la sénescence (Lossaint et al., soumis). L'objectif du deuxième volet fut de mettre en évidence le programme déclenchant la quiescence lors de confluence ou en absence de sérum. Les travaux antérieurs de l'équipe ont montré que cette décision pouvait être prise avant la mitose même si l'arrêt du cycle n'a lieu qu'en phase G1 suivante. En étudiant les fibroblastes synchronisés nous avons trouvé que la quiescence est précédée par l'inhibition pré-mitotique de la phosphorylation de pRb due à une diminution de cycline D1 et une stabilisation du CKI p27Kip1 (Chassot et al., 2008). De plus, nos résultats récents montrent que la présence de sérum entre le point R et la mitose est requise pour initier la réplication de l'ADN au cycle suivant. Les travaux futurs devraient élucider comment différentes voies de signalisation, via la voie cycline D-pRb, affectent divers composants de l'appareil de réplication de l'ADN pour inhiber la progression du cycle de façon réversible ou irréversible. / Cancer is a multi-step process resulting from abrogation of several barriers to uncontrolled proliferation. They include inhibitory pathways with appropriate checkpoints that lead to reversible (quiescence) or irreversible (senescence, apoptosis) block of cell proliferation. We are especially interested in pathways orchestrating cell cycle exit that operate in the G2 phase. The first objective of this thesis was to decipher mechanisms that prevent mitosis in response to DNA damage. We found that Cdk inhibitor p21Waf1 plays a crucial role in blocking mitotic onset in normal cells; acting in tandem with checkpoint kinase Chk1, p21 inactivates mitotic Cdks and inhibits pRb phosphorylation, thereby irreversibly blocking mitotic entry. In contrast, in p53-proficient transformed cells, the induction of p21 in G2 is impaired, most likely because of deficient ATM activation. While, in some cases, Chk1 hyper-activation prevents mitosis, the absence of p21 compromises the senescence program from G2. Finally, we showed that Chk2 is dispensable for G2 arrest in both non-transformed and transformed cells (Lossaint et al., submitted). Our second objective was to elucidate the pathways that induce quiescence (G0). This reversible cell cycle exit occurs in G1, requires pRb family members and p27Kip1-dependent Cdk inactivation. Based on observations obtained in our team and the data in the literature, we hypothesized that reversible cell cycle exit program might be launched before mitosis. By using an in vitro wounding model, we showed that confluence-driven quiescence is preceded by pre-mitotic CDK inhibition by p27, cyclin D1 downregulation and reduced pre-mitotic pRb phosphorylation (Chassot et al., 2008). Moreover, our results obtained in synchronized fibroblasts that were serum-starved after release from G1/S block suggest that cyclin D1 might stimulate proliferation by keeping pocket proteins phosphorylated during G2/M progression (Lossaint et al., in preparation).

Historiquement, SOCS1 a été décrite comme une protéine capable de réguler négativement la signalisation induite par les cytokines. Dès lors, un vif engouement s’est porté sur cette protéine quant à un probable rôle de suppresseur de tumeur, notamment grâce à ses fonctions anti-inflammatoires. De surcroit, son expression est fréquemment perdue dans de nombreux cancers des suites de phénomènes épigénétiques. Récemment, il a même été démontré que SOCS1 pouvait rentrer en complexe protéique avec le suppresseur de tumeur P53, et était requis pour son activité transcriptionnelle. Finalement, notre laboratoire a participé à démontrer que SOCS1 inhibait la signalisation et les fonctions du récepteur MET dans les hépatocytes. Étant donné l’influence bien connue de l’inflammation, de l’épigénétique, de P53 et de MET dans le cancer colorectal (CCR), nous avions émis l’hypothèse que SOCS1 agirait en tant que suppresseur de tumeur dans le CCR. Dans la présente étude, nous avons modulé l’expression de SOCS1 dans un modèle cellulaire de progression tumorale colique, soit les cellules CT36 et CT26. Puis, les caractéristiques biologiques dites cancéreuses de ces lignées ont été évaluées in-cellulo et in-vivo. Contre toute attente, nos résultats révèlent une action pro-tumorale menée par SOSC1 dans le CCR. La surexpression de SOCS1 dans les cellules CT26 provoque une augmentation de la croissance cellulaire, de la survie sans ancrage à la matrice extracellulaire et de la capacité à former des tumeurs in vivo. Et à l’inverse, sa diminution, par ARN interférent, provoque chez les cellules CT36 une baisse de ces mêmes fonctions cancéreuses. Au niveau moléculaire, nous avons validé que SOCS1 conservait bien son aptitude à inhiber les voies de signalisation des cytokines et du récepteur MET ainsi qu’à promouvoir l’activité de p53. Nous avons aussi identifié que SOCS1 diminue l’expression des protéines STAT1, P21cip/kip, et de la protéine adaptatrice P66SHC. En conclusion, ce travail de recherche révèle, pour la première fois, que l’expression de SOCS1 favoriserait le développement de certains CCR. Les protéines STAT1, P21 cip/kip et P66SHC étant connues pour être impliquées dans des processus antiprolifératifs et pro-apoptotiques, leurs régulations négatives par SOCS1 pourraient en être les raisons.

Death domain-associated protein 6 (DAXX) is a multifunctional protein involved in diverse cellular processes. It acts as a histone chaperone or regulator of transcription and apoptosis, in which is its role quite controversial. DAXX also participates in regulation of cell DNA damage response (DDR). DAXX co-creates and stabilizes complex with Mdm2, which negatively regulates the stability of p53, an important tumor suppressor, which is a part of signalling node in the DDR. If DNA damage is not lethal for the cell and unables it to proliferate, the irreversible state of cell cycle called cellular senescence takes place. Under physiological conditions, induction of senescence can prevent the development of tumorigenesis. Therefore, the description of mechanisms involved in the induction of senescence has potential clinical significance. In my thesis, I aimed to determine changes in the level of DAXX protein in senescent cells and to characterize the manner of its regulation. In tumor cells MCF-7 and primary BJ fibroblasts, I observed decrease in DAXX protein level and its regulation. I tested the hypothesis according to which an increase in DAXX level before DNA damage canprevent induction of cellular senescence, or increase in its expression during senescence can cause recovery of cell proliferation....

Cell competition is a form of cell-cell signalling that results in the elimination of less fit cells from a tissue by their fitter counterparts. I take advantage of an established in vitro model of cell competition using Madin-Darby canine kidney (MDCK) cells to shed insight into the molecular basis of cell competition in epithelial cells. In this system, silencing of the tumour suppressor scribble (scribKD) results in a 'loser' phenotype whereby scribKD cells are specifically eliminated from the monolayer by surrounding wild-type cells. More specifically, scribKD cells are compacted into tight clones through activation of a directed, collective migration in the wild-type population: scribKD are 'mechanical losers' and delaminate and die due to an intrinsic hypersensitivity to high cell density. Remarkably, p53 activation is both necessary and sufficient for this mechanical loser cell status. I first investigate the role of E-, N-, and P-cadherin in the directed migration between scribKD and wild-type cells and in scribKD cell loser status. I show that differential expression of E-cadherin between scribKD losers and wild-type winners is required but not sufficient for directed migration and has no impact on loser cell status. I also show that elevation of neither E-cadherin nor N-cadherin is sufficient to induce directed migration or loser status, but that P-cadherin may play a role in both. I next focus on translating findings about the molecular details of competition from the scribKD set-up into a system where p53 differences alone drive the formation and elimination of mechanical losers. I show that the ROCK - P-p38 - p53 pathway activated in response to mechanical compaction in scribKD cells is conserved in p53-driven losers. In the latter part of my thesis, I characterise the directed migration observed during MDCK competition by drawing parallels to canonical leader-follower migration. Canonical leader cells emerge when epithelial sheets are wounded and, by becoming migratory, drive collective cell migration of follower cells, which results in wound closure. It was not known what confers the leader cell fate. I show that p53 and its effector p21 (and potentially other cyclin-dependent kinase inhibitors) are the key drivers of leader cell migration. I demonstrate that p53-induced leaders use the same molecular pathways that have been shown to drive leader cell migration during wound healing and, in fact, p53 and p21 are also elevated in leaders generated by wounding. Importantly, I establish that p53 activity drives efficient wound closure. Lastly, I show that leader cells are often eliminated by cell competition in the final stages of wound closure, as their elevated p53 mediates their hypersensitivity to density. The model incorporating these data proposes that cellular damage during wounding generates cells with elevated p53, which become leaders and drive wound healing, but these are then cleared once the wound is closed because their high p53 levels cause them to become mechanical losers.

Proteins of the nuclear lamina provide structural support to the nuclear envelope and participate in a variety of cellular functions, such as chromatin organization and transcriptional regulation. One of these proteins, Lamin A (72kDa), is synthesized as a 74 kDa precursor protein, Prelamin A, which undergoes an unusual maturation pathway that requires two farnesylation-dependent endoproteolytic cleavages. The second cleavage is unique to lamin A in higher vertebrates and is specifically carried out by the endoprotease zmpste24. Although most previous studies have focused mainly on the function of mature Lamin A, recent evidence from our laboratory shows important biological functions for Prelamin A as well. Prelamin A concentration in proliferating cells is very low or undetectable. Conversely, during quiescence induced by mitogen withdrawal or contact inhibition, Prelamin A levels increase dramatically. These variations are directly regulated by changes in expression and enzymatic activity of zmpste24. The central hypothesis of this dissertation is that full-length farnesylated and carboxymethylated prelamin A (FC-PreA) antagonizes both proliferation and apoptosis, therefore playing a role in cellular quiescence/senescence. To accomplish this goal, we studied the transcriptional regulation of zmpste24 and the interaction of FC-preA with proteins that participate in cell cycle control. 1) We identified and characterized a functional site for the E2F1 transcription factor (involved in the control of cell cycle) in the proximal 5’ UTR region of zmpste24. 2) By using proximity-labeling and co-immunoprecipitation-mass spectrometry techniques, we identified a set of proteins that interact preferentially with L467R-Prelamin A (uncleavable mutant) but not with mature Lamin A. Many of these proteins function to regulate progression through cell cycle.

Chronic oxidative injury by allylamine induces proliferative vascular smooth muscle cell (vSMC) phenotypes in the rat aorta similar to those seen in rodent and human atherosclerotic lesions. In this study, we evaluate the potential role of cyclin dependent kinase inhibitors, p21 and p27, and extracellular regulated kinases (ERK1/2) to mediate the proliferative advantage of oxidatively stressed (i.e. allylamine injured) vSMC. Isolated rat aortic SMC from allylamine treated and control rats were cultured on different extracellular matrix (ECM) proteins. Following mitogen restriction, cultures were stimulated with serum with or without inhibitors of NF-kB or MEK. Western blot analysis was performed to identify protein differences between treatment groups. Basal levels of p21 were 1.6 fold higher in randomly cycling allylamine cells than control counterparts seeded on a plastic substrate, a difference lost when cells were seeded on collagen. p27 levels were comparable in both cell types irrespective of substrate. Basal levels of p21 and p27 were 1.4 fold higher in G0 synchronized allylamine cells compared with G0 synchronized control cells seeded on a plastic substrate. Following cell cycle progression, differences in protein levels were not detected. Treatment with 100 nM pyrollidine dithiocarbamate (PDTC) resulted in significant decreases in p21 and p27 in allylamine cells versus control cells following serum stimulation for 9 hours. This decrease was even greater for p21 in allylamine cells when grown on collagen relative to control cells. Alterations in peak and temporal activation of ERK1/2 were observed in allylamine cells seeded on a plastic substrate as compared to control cells, following serum stimulation. Seeding on collagen decreased the enhanced peak phosphorylation of ERK1/2 and increased the sustained activity in allylamine cells compared with control counterparts. Inhibition of ERK1/2 activity resulted in reduced p21 expression in both cells types, but the response was markedly enhanced in allylamine cells, and preferentially observed on a restrictive collagen substrate. We conclude that induction of proliferative (i.e. atherogenic) phenotypes following repeated cycles of oxidative injury involves ERK1/2 activity and modulation of the cyclin dependent kinase inhibitors, p21 and p27, in a matrix-dependent manner.