Dans ce manuscrit, nous démontrons par le biais de dimères forcés que toutes les propriétés oncogéniques de la protéine TWIST1, telles qu'évaluées par le biais de nombreux tests in vitro (tests de complémentation, inhibition de la sénescence oncogénique, induction de l'EMT et tests de coopération oncogénique…), sont spécifiquement attribués au complexe TWIST1-E12. L'insertion de mutations ponctuelles, définies d'après l'analyse de modèles de simulation in silico développés au sein du laboratoire (Bouard et al., 2013) et perturbant la dimérisation du complexe ou encore son interaction avec l'ADN conduit à une perte complète de l'activité, validant l'importance des deux partenaires dans l'activité oncogénique de la protéine de fusion. La détection du complexe TWIST1-E12 dans des carcinomes mammaires canalaires in situ humains, récapitulant les phases précoces de l'initiation tumorale, par la technique de Proximity Ligation Assay (PLA) ainsi que la sensibilité accrue de souris transgéniques à développer des carcinomes mammaires lorsque le complexe hétérodimérique est exprimé dans les cellules épithéliales luminales mammaires, renforce la conclusion que le complexe TWIST1-E12 est la (ou l'une des) forme(s) active(s) de la protéine TWIST1 dans le cadre de la carcinogenèse mammaire / Among embryonic transcription factors, TWIST proteins (TWIST1 and TWIST2) display the particularity to behave as master regulators of both RB and p53-oncosuppressive pathways. These embryonic transcription factors annihilate the induction of numerous cyclin-kinase inhibitors, including p16INK4A, p15INK4B and p21CIP1, abrogating thereby cell commitment to a senescence program or their death through apoptosis in response to an oncogenic activation. By doing so, TWIST proteins cooperate with mitogenic oncoproteins such as Ras in promoting cell transformation in vitro and breast and lung carcinogenesis in vivo (Ansieau S. et al. Cancer Cell ; Morel A-P. et al., PLoS ONE ; Tran P.T. et al., PLoS Genetics). Strikingly, TWIST depletion in numerous cancer cell types associates with a reactivation of failsafe programs, suggesting that some tumor cells remain addictive to TWIST for survival and proliferation (Ansieau et al., 2008). In support of this hypothesis, silencing TWIST expression in a TWIST + RAS-driven lung carcinogenesis mouse model displays a cytostatic effect (Tran P.T. et al., PLoS Genetics). Based on these observations, we aim at identifying TWIST specific inhibitors and evaluate the efficiency of such molecules in eradicating tumor cells in vitro as well as in vivo. TWIST proteins either behave as homodimeric (TT) or heterodimeric (TE) complexes (in association with E2A proteins), both complexes displaying distinct and sometimes even antagonistic functions during the embryonic development (Firulli B.A. et al., Nature Genetics ; Connerney F. et al., Dev. Dynamics). Ongoing experiments, comparing activities of tethered TWIST dimers, strongly support the assumption that an oncogenic potential is specifically allotted to the heterodimer. To screen for specific chemical inhibitors, we first established the in silico structure of the bHLH domain of TWIST complexes bound to their cis-responsive elements, by analogy with the NeuroD/E47 crystallographic structure. Relevance of these models has been confirmed through analysis of Twist1 variants associated with a loss of function in Saethre-Chotzen patients (Bouard et al., J. Biomolecular Structure & Dynamics). Strikingly, structural analysis highlights the importance of lateral loops in stabilizing the protein-DNA complex and in specifying the DNA sequence targeted

The central nervous system is formed from epiblast precursor cells through Neurulation. Neural induction can be studied in its main aspects in vitro. However, the process is poorly understood, especially in regard to when and how a cell becomes specified, and then committed, to be a neural cell. It is, on the other hand, well established that neural formation requires absence or, inhibition of the BMP signalling both in vivo and in vitro. ID1 is a direct target of BMP signalling with major influence on in vitro neural differentiation. A cDNA library screen, looking for transcription factors negatively regulated by ID1, reported TWIST1, along with only two other proteins. Twist1 expression is upregulated during in vitro neural differentiation. Furthermore, targeted deletion of Twist1 has dramatic consequences on anterior neural development. Twist1 knock-out mice fail to form the closed neural tube in the prospective brain, followed by exencephaly and, early embryonic death. In this thesis I investigate the influence on in vitro neural differentiation of a TWIST1 constitutively active form, insensitive to ID1 inhibition. I report that this transcriptionally active TWIST1 accelerates neural differentiation, in vitro and, biases it, towards dorsal phenotypes. I provide, for the first time, evidence for Twist1 expression in the neural tissue, observed weakly in a restricted domain, temporally and spatially, in the dorsal part of the neural tube. I propose a new model for TWIST1 influence at this level. I also investigate how TWIST1 actions depend on levels of expression and dimer choice. I found that, TWIST1 can exert its neural modulating actions only at low levels, as high levels divert a cell fate towards non-neural lineages.

For cancer cells to initiate cell migration and progress to metastasize, epithelial genes must be silenced and the expression of mesenchymal genes must be upregulated. During prostate carcinogenesis, E-cadherin expression is downregulated through multiple mechanisms, the majority of which combine to silence E-cadherin expression through transcriptional regulation at the level of the E-cadherin promoter. Recently it has been discovered that there is transcriptional upregulation of the mesenchymal cadherin, N-cadherin during prostate cancer metastasis. Although N-cadherin expression can be detected in human prostate cancer and in prostate carcinoma cell lines, the mechanisms controlling the transcriptional regulation of N-cadherin in cancer are uncharacterized. This body of work offers the first evidence for the mechanisms controlling the transcriptional upregulation of N-cadherin expression in prostate carcinoma. We utilized anchorage independent culture to induce downregulation of N-cadherin expression, and then analyzed the necessary events for N-cadherin upregulation when cells attached to Fibronetin (FN). In order to determine the functional regions of the N-cadherin proximal promoter that were involved in the upregulation of N-cadherin expression, we cloned regions of the human N-cadherin 5’ proximal promoter, and regions of the first intron of the N-cadherin gene into a luciferase reporter vector. It was determined that the bHLH transcription factor Twist1 controlled the upregulation of N-cadherin transcription in PC-3 cells, through β1 integrin dependent nuclear localization of Twist1. A cis-element located in the first intron of the N-cadherin gene was shown to be necessary for Twist1 mediated effects on the N-cadherin promoter. We then determined the requirements for cell-type specific expression of the N-cadherin promoter. It was determined that an additional cis-element located in the first intron of the N-cadherin gene was necessary to repress N-cadherin promoter activity in cells lacking N-cadherin. Through deletion analysis of the N-cadherin promoter luciferase construct, a DNA binding site for the transcription factor FoxP1 was discovered. FoxP1 binds to the repressive cis-element in vitro, and mutation of the FoxP1 DNA binding site eliminated cell-type specific activity of the N-cadherin promoter. Therefore, we have documented that the aberrant expression of N-cadherin in prostate carcinoma involves alterations in both positive and negative transcriptional regulators.

Ovarian cancer results in more deaths per year than any other cancer of the female reproductive system. The low survival rate is partly due to the lack of early detection and the susceptibility to relapse. The AKT serine threonine kinase plays a pivotal role in hallmark cellular processes for the progression of ovarian cancer, including tumor cell growth and migration. Therapeutic targeting of pan-AKT has been problematic, in part due to feedback mechanisms and crosstalk with other pathways. The hypothesis for this study is that AKT 1, -2 and -3 isoforms may have different roles and regulate cell processes in uniquely varied ways. A transgenic mouse model that expresses the SV40 T-antigen viral oncogene and is known to have dramatically increased susceptibility to ovarian cancer was utilized, and it had genetic inactivation of either AKT1 or AKT2 through targeted deletion of the individual genes because these isoforms have been implicated in this cancer. Primary ovarian tumor cell cultures were established and found to exhibit different morphology, proliferation and migration that may indicate a different role for the AKT1 and AKT2 isoforms in these contexts. Ovarian tumor cells with absence of AKT1 predominantly exhibited reduced cell migration when compared to cells with retention of AKT1 and absence of AKT2. Since AKT is known to be important for epithelial-mesenchymal transition (EMT), a process potentially associated with tumor cell metastasis, the expression of transcription factors implicated in EMT was assessed by real-time array analysis in ovarian tumor cells knocked-out for either AKT1 or AKT2. Twist1, one of the major players in EMT, was not detectable in the cells missing the AKT1 isoform. Results indicate an association of Twist1 with AKT1 in EMT and migration of ovarian tumors cells. This finding is significant because AKT2 has been implicated as the major player of cell migration in human breast cancer iv cells. Collectively, these findings support a tissue specific role of the AKT isoforms, and may provide insights regarding the most useful cell context in order to target components of the AKT signaling pathway indirectly affecting EMT in order to prevent tumor progression in patients with ovarian and perhaps other types of cancers

La réactivation aberrante du gène TWIST1 embryonnaire a été identifiée dans le laboratoire comme un mécanisme d'inactivation récurrente des voies dépendantes de p53 et Rb dans de nombreuses tumeurs. En diminuant la sénescence et l’induction de l'apoptose, TWIST1 coopère avec des protéines oncogéniques dans la transformation cellulaire in vitro et intervient dans l'initiation et la progression tumorale in vivo. Comme TWIST1 est très faiblement exprimée dans la plupart des cellules adultes différenciées, elle constitue une cible attrayante pour des thérapeutiques futures. Récemment, l’héterodimère TWIST1/E2A (E12 ou E47, deux produits d'épissage alternatif du gène TCF3) joue un rôle pro-métastatique dans le cancer de la prostate (Gajula et al., 2015), alors que ce complexe est la forme oncogénique de TWIST1 dans des cellules épithéliales mammaires humaines (Jacqueroud et al., 2016). L'hétérodimérisation par l'intermédiaire des domaines HLH est une condition sine qua non pour la formation de domaine de liaison d'ADN (Murre et Massari 2000). Une approche in silico basé que la modélisation par homologie des complexes de TWIST1 met en évidence le rôle déterminant des boucles inter-hélicales dans le maintien de complexes de TWIST1 à l’ADN en étudiant plusieurs insertions de TWIST1 observées chez les patients atteints de syndrome Saethre-Chotzen (Bouard et al., 2014). Ensuite des approches in silico et in vitro nous ont permis de comprendre les mécanismes moléculaires sous-jacents à la reconnaissance des séquences E-box des promoteurs de gènes cibles par l'héterodimère TE et leur stabilisation sur l'ADN (bouard et al., 2016). Nous avons décrit trois états différents du complexe TWIST1/E12 lié à l'ADN à des séquences E-box fonctionnels et modifiés en fonction de l'affinité de reconnaissance des E-box par TWIST1 / E12 (Bouard et al, 2016). Et enfin, cette approche in silico nous a permis de montrer l’impact de la phosphorylation de TWIST1 sur la dimérisation et sur la liaison à l’ADN des complexes de TWIST1. Cette dernière étude met en exergue l’importance des régulations post-traductionnelles dans l’étude de l’activité de la protéine TWIST1 et dans la recherche d’inhibiteurs / Aberrant reactivation embryonic TWIST1 gene has been identified in the laboratory as a recurrent inactivation mechanism dependent pathways of p53 and Rb in many tumors. Decreasing senescence and the induction of apoptosis, TWIST1 cooperates with oncogenic proteins in cell transformation in vitro and is involved in tumor initiation and progression in vivo. As TWIST1 is very weakly expressed in most adult differentiated cells, it is an attractive target for future therapies. Recently, the heterodimer TWIST1 / E2A (E12 or E47, two alternative splice product of the TCF3 gene) plays a pro-metastatic role in prostate cancer (Gajula et al., 2015), while the complex is the oncogenic form of TWIST1 in human mammary epithelial cells (Jacqueroud et al., 2016). The heterodimerization through HLH areas is a prerequisite for DNA binding domain of training (Massari and Murre 2000). An in silico approach that based homology modeling of complex TWIST1 highlights the key role of inter-hélicales loops in the DNA TWIST1 complex retention student TWIST1 several insertions observed in patients Saethre syndrome -Chotzen (Bouard et al., 2014). Then approaches in silico and in vitro have enabled us to understand the molecular mechanisms underlying the recognition of E-box sequences of the promoters of target genes by TE heterodimer and stabilize DNA (Bouard et al. 2016). We have described three different states of the complex TWIST1 / E12 bound to DNA to functional E-box sequences and modified according to the affinity recognition by E-box TWIST1 / E12 (Bouard et al, 2016). Finally, this approach in silico allowed us to show the impact of TWIST1 phosphorylation on dimerization and binding to DNA TWIST1 complexes. This latest study highlights the importance of post-translational regulation in the study of the activity of the protein and TWIST1 in the search for inhibitors

The precise role that genes play in early craniofacial development and postnatal craniofacial growth are essential to understand dento-facial development overall. However, genotype-phenotype correlations between genetic variation of early craniofacial genes and adult craniofacial phenotypes is poorly understood. Thus, this thesis focused on identifying the genetic etiology underlying phenotypic variations present in malocclusion conditions. First, we performed genotype-phenotype association analyses between common variants in 82 craniofacial genes and phenotypic variations extracted from 2D and 3D pre-treatment dental records of individuals with malocclusion. This effort identified that variant rs2189000 upstream of TWIST1 is highly associated with mandibular body length and inclination and cranial base angulations which can lead to malocclusion. Next, via cell based functional assays, we discovered that rs2189000 disrupts a PITX2 binding site and also showed the direct regulation of TWIST1 expression by the PITX2 gene. Finally, we identified abnormal craniofacial phenotypes and malocclusion in Twist1 deleted mice including asymmetric snouts, domed cranial vaults, and changes in size and inclination of the cranial base, palate and mandible resulting in malocclusion and resembling the human phenotypes observed. Also, premature calcification of calvarial sutures and cranial base synchondroses were also observed in the mutant mice indicating a possible biological mechanism for the abnormal phenotypes detected. These results confirm that TWIST1 is an important regulator of postnatal growth and that genetic variation in TWIST1 can result in malocclusion. The continued identification of genetic etiological factors and their role in craniofacial growth will impact treatment and prevention of malocclusion and other craniofacial conditions

No description available.

No description available.

Une craniosynostose est une pathologie caractérisée par la fusion prématurée d'une ou plusieurs sutures crâniennes. C'est un défaut de naissance assez fréquent (1/2500 naissances) qui résulte en une forme anormale du crâne et qui peut être accompagné d'une déficience mentale dans certains cas. Des mutations du gène TWIST1, qui encode un facteur de transcription basique Helix-Loop-Helix (bHLH) de classe II, causent le syndrome de Saethre-Chotzen qui est associé à une synostose de la suture coronale (El Ghouzzi et al. 1997; Howard et al. 1997). Un nouveau gène a récemment été découvert comme étant une nouvelle cause du syndrome Saethre-Chotzen ainsi que de synostose coronale asyndromique (Sharma, Fenwick, Brockop, et al., 2013): il s'agit du gène TCF12, qui encode un facteur de transcription bHLH de classe I.Nous démontrons qu'une reduction de l'expression génique de Twist1 et Tcf12 chez la souris cause une synostose coronale, et nous suggérons que les protéines bHLH Twist1 et Tcf12 forment des hétérodimères dont le dosage est critique pour le développement de la suture coronale.Nous nous concentrons aussi sur Twist1 et prouvons que son expression est requise dans les tissus dérivant du mésoderme ainsi que ceux dérivant des crêtes neurales pour le développement normal de la suture coronale.De plus, nous notons que dans la suture coronale, Twist1 exclut Notch2 afin de garder la suture ouverte, et beta-catenin joue un rôle dans la maintenance de l'ouverture de la suture en ciblant Jagged1 lors du développement de la suture coronale chez la souris.Enfin, nous mentionnons de nouveaux gènes qui pourraient avoir un impact sur le développement normal de la suture coronale: Aggrecan, Goosecoid, Gucy1a3 et Gucy1b3. / Craniosynostosis, the premature fusion of one or more cranial sutures, is a common birth defect (1/2500 live births) that results in abnormalities in skull shape and sometimes in neurological deficiencies (Wilkie, 1997; Wilkie and Morriss-Kay, 2001). Mutations in TWIST1, which encodes a class II basic helix-loop-helix (bHLH) transcription factor, cause Saethre-Chotzen syndrome, associated with coronal synostosis (El Ghouzzi et al. 1997; Howard et al. 1997). We recently discovered a new craniosynostosis gene, TCF12, which encodes a class I bHLH transcription factor. Tcf12 causes.Saethre-Chotzen syndrome and asyndromic coronal synostosis. (Sharma, Fenwick, Brockop, et al., 2013). We show that a reduction in the dosage of Twist1 and Tcf12 in mouse causes coronal synostosis, and we suggest that the Twist1 and Tcf12 form heterodimers whose dosage is critical for coronal suture development. We also demonstrate that Twist1 is required in both neural-crest and mesoderm-derived tissues for the normal coronal suture development. Moreover, we show that in the coronal suture, Twist1 excludes Notch2 thus maintaining suture patency. and we show that beta-catenin also plays a role in the maintenance of suture patency by regulating Jagged1. Finally, we identified Aggrecan, Goosecoid, Gucy1a3 and Gucy1b3 as Twist1-regulated genes that could have an impact on the normal development of the coronal suture.

Systemic sclerosis (SSc) is immune-mediated fibrotic disease of unknown aetiology. Among the dominant pathogenic manifestations of SSc belong vascular changes, production of autoantibodies, activation of innate and adaptive immune responses and fibrotic processes. Transforming growth factor beta (TGF-β) has been identified as a central profibrotic factor stimulating fibroblasts to produce collagen. There are, however, a number of other mediators involved in the pathogenesis of SSc. Mutual activation and amplification of these molecules and their cascades may be a central mechanism of the SSc pathogenesis. Hedgehog (Hh) canonical signalling pathway plays an important role in the development and progression of fibrotic diseases. Expression of Hh target genes can be regulated through a canonical or non-canonical signalling cascade. The non-canonical activation of GLI transcription factors by TGF-β has not yet been investigated in SSc. The substantial part of this thesis is focused on the study of the mutual interaction of TGF-β and Hh signalling pathway. In vitro analysis confirmed TGF- β/SMAD3 dependent activation of GLI2 in dermal fibroblasts. Fibroblasts specific knockout of GLI2 prevented the development of experimental fibrosis in vivo. Combined targeting of canonical and non-canonical Hh...