Neural crest cells are a transient population of cells which differentiate into multiple derivatives. How these derivatives become specified is not well understood but Sox10 is known to be important in many of them. We are interested in defining the precise role of Sox10 in zebrafish melanophores. Current evidence suggests that the only vital function that Sox10 performs in melanophores is to induce expression of the melanocyte master regulator mitfa (Elworthy et al. 2003). We explored a model for Sox10 function in melanophores, based upon a model for Sox10’s role in mouse sympathetic neurons (Kim et al. 2003), and tested the following predictions: as well as inducing expression of mitfa, Sox10 will repress expression of genes downstream of Mitfa thus, Sox10 must be downregulated, via Mitfa, to allow melanophore differentiation. We observed derepression of melanophore marker genes in sox10t3 mutants, supporting the hypothesis that Sox10 represses these genes in wild type melanophores. We documented Sox10/sox10 downregulation in developing melanophores and generated transgenic lines to test whether this is necessary for differentiation. Unfortunately our experimental lines did not express our transgene so we were unable to test this hypothesis. However, transgenic lines, generated as controls, which express CFP in melanophores or xanthophores will be useful tools in their own right. Finally we conducted RNA injection experiments to explore regulation of melanophore genes by Sox10 and Mitfa. We found that injection of mitfa induces expression of all our melanophore markers whereas co-injection of mitfa and sox10 does not. We also found that the 7.2 kb sox10 promoter contains six Mitf binding sites and is Mitfa responsive. Our data broadly support our original model but also suggest that it does not describe the complete network. We propose a modified model for the role of Sox10 in the genetic regulatory network controlling melanophore development.

Over the past two decades, the mammalian Ste20-like kinase (SLK) has been characterized for its role in regulating cellular migration, proliferation and apoptosis in fibroblasts and myoblasts. In mammary epithelial cells, SLK has been shown to be required for efficient epithelial-to-mesenchymal transition and to be activated downstream of the HER2/Neu-oncogene to control chemotactic cellular migration. Here, we assessed the role of SLK in HER2/Neu-induced mammary tumorigenesis in vivo. As SLK is activated downstream of HER2/Neu, we hypothesized that the loss of SLK would significantly delay tumor progression in a mouse model of HER2-positive breast cancer. As we have shown that global attenuation of SLK kinase activity results in embryonic lethality, a conditional SLK knockout mouse model was generated. To study the role of SLK in HER2-positive breast cancer, we crossed these conditional SLK knockout mice with mice expressing HER2/Neu linked to Cre recombinase in the mammary luminal epithelium. Unexpectedly, we have demonstrated that conditional deletion of SLK significantly accelerates Neu-induced mammary tumor onset and decreases overall survival. SLK deletion results in the induction of Sox10 which drives mammary stem/progenitor activity and cooperates with HER2/Neu to drive tumor growth. Using the Cancer Genome Atlas, we have supported previous findings and validated Sox10 as a potential biomarker of the Triple-negative Breast Cancer subtype. Furthermore, we have uncovered that SLK deletion results in enhanced activation of both PDK1 and AKT. We provide evidence that Sox10 induction requires signaling through a novel AKT/Sox9-dependent pathway following SLK deletion. Taken together, our data suggests that SLK may be required to maintain cells in a fully differentiated state and that loss of SLK in HER2/Neu-induced breast cancer drives a more basal/stem-like phenotype through the induction of Sox10.

Résumé non transmis / Summary not transmitted

Résumé non transmis / SOX10 is a transcription factor with well-known functions in neural crest and oligodendrocyte development. Mutations in SOX10 were first associated with Waardenburg-Hirschsprung disease (WS4; deafness, pigmentation defects and intestinal aganglionosis). However, variable phenotypes that extend beyond the WS4 definition are now reported. The neurological phenotypes associated with some truncating mutations are suggested to be the result of escape from the nonsense-mediated mRNA decay pathway; but, to date, no mechanism has been suggested for missense mutations, of which approximately 20 have now been reported, and about half of which are redistributed in vitro to nuclear bodies of undetermined nature and function. Here, we reported that the paraspeckle protein p54NRB, which plays a crucial role in the regulation of gene expression during many cellular processes including differentiation, and is a member of the Drosophila behavior Human Splicing (DBHS) protein family, interacts and acts synergistically with SOX10 to regulate several target genes. Interestingly, this multifunctional protein, as well as two other members of the DBHS protein family, co-localized with SOX10 mutants in nuclear bodies, suggesting the possible paraspeckle nature of these foci or re-localization of the DBHS members to other subnuclear compartments. Remarkably, the co-transfection of wild-type and mutant SOX10 constructs led to the sequestration of wild-type SOX10 in mutant-induced foci. However, only foci forming mutants exclusively found in the nucleus altered synergistic activity between SOX10 and p54NRB. We proposed that such a dominant negative effect may contribute to or be at the origin of the progressive neurological phenotype observed in affected patients.One of the roles of p54NRB is the regulation of gene expression via nuclear retention, by binding to hyperedited IRAlu sequences this protein blocks their efficient export to the cytoplasm (Zhang and Carmichael., 2001), we then decided to get into the world of editing. Editing, is a molecular mechanism characterized by the deaminase conversion of adenosines into inosines (A-to-I). In mammals, this molecular modification, is performed by a cluster of three enzymes named Adenosine deaminases acting on RNA (ADARs 1-3) (Wagner RW et al., 1989).In order to evaluate the role of ADAR1 in NC development, we decided to conditionally invalidate the expression of this enzyme using the NC specific HtPA-Cre line. Two main crossing strategies were followed, one including the Rosa26R-LacZ marker (RADR crossing) to track the NCCs and one not (CADR crossing). Globally, the Adar1 deficient pups harvested from the CADR crossing presented with 100% mortality within the first three days after birth. The survival rate of the mutants generated using the second strategy (RADR) was higher, however, none of the mutants survived up to P30. In general, the mutants of the latest crossing, presented with pleiotropic NC phenotype: abnormal melanocyte, ENS and sciatic nerve defects were observed.

The neural crest is a transient population of cells that forms a diverse range of derivatives in vertebrate embryos. Neural crest cells also migrate extensively throughout the embryo. The specification of a number of neural crest derivatives, including pigment cells and neurons and glia of the peripheral nervous system, is dependent on the transcription factor Sox10. In sox10 mutant zebrafish embryos, these neural crest derivatives fail to specify and subsequently the cell differentiation and migration fails leading to apoptosis. Sox10 mutant embryos also display an ear defect although the precise role of Sox10 in the ear is less well defined. Thus Sox10 controls an extensive gene regulatory network that drives the development of an important subset of neural crest derivatives and also functions during ear development. This gene regulatory network is currently poorly defined. The aim of this project was to identify genes that are both direct and indirect targets of Sox10 to further elucidate this gene regulatory network. To achieve this, a microarray approach was adopted. Initially, fluorescence activated cell sorting was employed to enrich for sox10 expressing cells from 24 hours post fertilization sox10:GFP transgenic embryos. The transcriptomes of WT and sox10 mutant cells were compared by microarray analysis to identify differentially regulated genes. A large number of target genes were identified by this method and by an unbiased in situ hybridization screen, 28 genes were validated. Of these, 23 genes were expressed in cells of the neural crest and down-regulated in sox10 mutant embryos. The majority of these genes were expressed in cells of the melanocyte and xanthophore lineages. 5 genes were expressed in the ear (otic vesicle) of which three otic vesicle genes were down-regulated while two otic vesicle genes were up-regulated in sox10 mutant embryos. Unfortunately due to time constraints, a study into the function of one of these target genes could not be completed. The series of validated genes identified during this project has opened new opportunities for research and has identified a number of highly expressed marker genes that will be useful in future studies. In addition, the microarray data presented will be a useful resource to aid the identification of further targets of Sox10.

Melanocytes, pigment-producing cells, derive from the neural crest (NC), a population of pluripotent cells that arise from the dorsal aspect of the neural tube during embryogenesis. Many genes required for melanocyte development were identified using mouse pigmentation mutants. The deletion of the transcription factor Ets1 in mice results in hypopigmentation; nevertheless, the function of Ets1 in melanocyte development is unknown. The goal of the present study was to establish the temporal requirement and role of Ets1 in murine melanocyte development. In the mouse, Ets1 is widely expressed in developing organs and tissues, including the NC. In the chick cranial NC, Ets1 is required for the expression of Sox10, a transcription factor critical for the development of melanocytes, enteric ganglia, and other NC derivatives. Using a combination of immunofluorescence and cell survival assays Ets1 was found to be required between embryonic days 10 and 11, when it regulates NC cell and melanocyte precursor (melanoblast) survival. Given the requirement of Ets1 for Sox10 expression in the chick cranial NC, a potential interaction between these genes was investigated. Using genetic crosses, a synergistic genetic interaction between Ets1 and Sox10 in melanocyte development was found. Since Sox10 is essential for enteric ganglia formation, the importance of Ets1 on gut innervation was also examined. In mice, Ets1 deletion led to decreased gut innervation, which was exacerbated by Sox10 heterozygosity. At the molecular level, Ets1 was found to activate a Sox10 enhancer critical for Sox10 expression in melanoblasts. Furthermore, mutating Ets1 at a site I characterized in the spontaneous variable spotting mouse pigmentation mutant, led to a 2-fold decrease in enhancer activation. Overexpression and knockdown of Ets1 did not affect Sox10 expression; nonetheless, Ets1 knockdown led to a 6-fold upregulation of the transcription factor Sox9, a gene required for melanocyte and chondrocyte development, but which impairs melanocyte development when its expression is prolonged. Together, these results suggest that Ets1 is required early during melanocyte development for NC cell and melanoblast survival, possibly acting upstream of Sox10. The transcription factor Ets1 may also act indirectly in melanocyte fate specification by repressing Sox9 expression, and consequently cartilage fate.

Le mélanome est un cancer très agressif en raison de sa capacité rapide à former des métastases et de développer une résistance aux traitements existants.
 MITF (Micropthalmia-associated Transcription Factor) est un facteur de transcription clé à toutes les étapes de développement du lignage mélanocytaire et dans la physiopathologie du mélanome. Afin de comprendre les mécanismes impliqués dans la régulation de l’activité et de la stabilité de MITF, nous avons identifié ses partenaires protéiques parmi lesquels figurent de nombreuses sous-unités des complexes de remodelage de la chromatine ATP-dépendant PBAF et NURF. Ce travail caractérise le rôle et l’étendue de la coopération entre BRG1/PBAF et plusieurs facteurs de transcription clés tels que MITF et SOX10 dans le fonctionnement des cellules de mélanome, qui recrutent activement de BRG1 à la chromatine et contribuent ainsi à la mise en place de la signature épigénétique caractéristique des cellules de mélanome prolifératives. Par ailleurs, l’utilisation de différents modèles murins a permis de révéler in vivo la contribution fonctionnelle distincte mais complémentaire de ces deux complexes de remodelage associé à MITF aux cours de trois stades majeurs du lignage mélanocytaire : le développement embryonnaire des mélanocytes, leur différentiation ainsi que lors de l’initiation et la progression du mélanome. Ce travail contribue ainsi à une meilleure compréhension du fonctionnement biologique des mélanocytes, du mélanome et du remodelage de la chromatine chez les eucaryotes. / Malignant melanoma is the most deadly form of skin cancer due to its quick metastatic spread and the development of resistance to available treatments.
MITF (Micropthalmia-associated Transcription Factor) is a transcription factor and master regulator of melanocyte lineage development and melanoma physiopathology. In order to investigate the mechanisms involved in the regulation of MITF activity and stability, we identified its numerous partners by tandem affinity purification coupled to mass spectrometry, which include several subunits of the PBAF and NURF ATP-dependant chromatin remodelling complexes. The present work characterizes the role and extent of cooperation between BRG1/PBAF and several key transcription factors including MITF and SOX10 in melanoma cell function, that actively recruit BRG1 to chromatin to establish the epigenetic landscape of proliferative melanoma cells. Furthermore, using different mouse models we revealed the distinct but complementary functional contribution of these two MITF-associated chromatin remodelers in vivo at three majors stages of melanocyte lineage development: embryonic development of melanocytes, their differentiation and during melanomagenesis. Thus, this work contributes to a better understanding of processes regulating the biological function of melanocytes, melanoma and more widely chromatin remodelling events in eukaryotes.

Les cellules de la crête neurale se caractérisent par leur capacité de migration dansl’embryon et la variété des types cellulaires qu’elles sont capables de générer (mélanocytes,système nerveux entérique (SNE) et périphérique). Chez l’homme, plusieurs maladiescongénitales affectant des organes et tissus divers, ont pour origine une anomalie demigration, prolifération, survie ou différenciation de ces cellules. Au laboratoire, nousétudions deux d’entre elles, le syndrome de Waardenburg-Hirschsprung (WS4- anomalie depigmentation, surdité et maladie de Hirschsprung (HSCR : anomalie entérique)) et lesyndrome de Mowat et Wilson (MWS – retard mental sévère, dysmorphie faciale avec ousans HSCR). A l’heure actuelle, quatre gènes ont été impliqués : l’endothéline 3 (EDN3) etson récepteur à sept domaines transmembranaires EDNRB et les deux facteurs de transcriptionZEB2 et SOX10. Au cours de ma thèse, nous avons montré que des délétions de SOX10 sontégalement responsables de 15% des cas de WS2 (défauts de pigmentation et surdité sansHSCR), élargissant le spectre des phénotypes liés à une mutation au sein de ce gène(Bondurand, Dastot-Le Moal, Stanchina et al. Am. J Hum. Genet, 2007).Parallèlement à ces études génétiques, nous avons souhaité mieux définir la fonction et lesinteractions entre les différents gènes impliqués dans le WS4 (SOX10, EDN3 et EDNRB).Pour cela, nous avons croisé les modèles murins invalidés pour ces gènes, et comparé lephénotype des simples et doubles mutants. A travers cette analyse phénotypique, nous avonsdémontré qu’une interaction entre ces molécules est nécessaire au développement normal duSNE et des mélanocytes dérivés de la crête neurale. En effet, par rapport aux simples mutants,les doubles mutants Sox10;Edn3 et Sox10;Ednrb présentent une augmentation de ladépigmentation, et une forte aggravation du phénotype entérique. Le suivi du devenir descellules formant le SNE au cours du développement nous a permis de montrer quel’aggravation du phénotype entérique est due à une diminution du pool de cellulesprogénitrices par apoptose (Stanchina et al. 2006).Dans la continuité des travaux déjà réalisés, nous avons voulu améliorer notrecompréhension du rôle joué par le gène du MWS : ZEB2, et étudier ses interactions avec lesgènes du WS4. Dans un premier temps, nous avons analysé l’effet de l’expression constitutiveou l’inhibition de ce facteur sur la survie, prolifération et différenciation des cellulesprogénitrices du SNE à l’aide d’un système de culture de progéniteurs entériques disponibleau laboratoire. Nos résultats suggèrent un effet répresseur de ZEB2 sur la différenciationneuronale. Ce facteur pourrait donc être nécessaire au maintien du pool de cellulesprogénitrices dans un état indifférencié. Nous avons ensuite étudié les interactions entre ZEB2et les gènes du WS4. Nous avons croisé les souris portant une invalidation du gène ZEB2 avecles souris invalidées pour SOX10 ou portant une mutation de EDN3 ou EDNRB, et démontréqu’une interaction entre ZEB2 et SOX10 est nécessaire au développement normal du SNE. Eneffet, par rapport aux simples mutants, les doubles mutants présentent une forte aggravationdu phénotype entérique, due à une diminution de la prolifération des cellules progénitrices et àune augmentation de la différenciation neuronale. L’analyse phénotype des mutantsZeb2;Edn3 et Zeb2;Ednrb suggère également l’existence d’une interaction entre ces troismolécules, mais l’origine du défaut entérique reste inexplorée.Ces études nous ont permis de mieux appréhender les réseaux moléculaires mis en place aucours du développement du SNE, de comprendre l’origine des anomalies entériques observéeschez les patients, améliorant leur prise en charge. / To understand in more details the molecular and cellular bases of hereditary diseases resulting from defects of neural crest (NC) development, we study several neurocristopathies, in particular Waardenburg syndrome (WS – pigmentary abnormalities and hearing loss), and Mowat-Wilson syndrome (MWS, severe mental retardation, facial dysmorphy, with or without HSCR (congenital megacolon)). To date, about ten causative genes have been identified, among which are the seven transmembrane domain receptor EDNRB and its ligand endothelin 3 (EDN3), the two transcription factors SOX10 and ZEB2.We contributed to the research efforts engaged to unravel these disorders. In particular, we identified the first mutations of SOX10 in patients presenting with WS4 (association of WS with HSCR disease) and WS2 (Bondurand, Dastot-Le Moal, Stanchina et al. Am. J Hum. Genet, 2007), and participated to functional studies describing its role during enteric nervous system (ENS) development. More recently, we identified the gene ZEB2 as responsible for MWS. The goal of my thesis was to understand the function of these genes and their interaction during the development of NC and ENS in particular. For this purpose, we combined an in vitro approach (isolation of ENS progenitors) to in vivo experiments (phenotype analysis of simple and double mutant mice). We demonstrated that an interaction between SOX10, EDN3 and EDNRB is necessary for the normal development of the ENS and melanocytes (Stanchina et al. 2006), and then focused our efforts in understanding the function of ZEB2 during the development of the ENS as well as its interactions with WS4 genes. Preliminary results suggest that ZEB2 inhibition accelerates neuronal differentiation in vitro. In the same time, generation of Zeb2;Sox10, Zeb2;Edn3 and Zeb2;Ednrb have been realized. Through phenotype analysis of Sox10;Zeb2 double mutants, we showed that a coordinated and balanced interaction between these two genes is required for normal ENS development. Indeed, double mutants present with more severe ENS defects due to decreased proliferation of enteric progenitors and increased neuronal differentiation from E11.5 onwards. These data revealed that crosstalks between these two transcription factors are crucial for proper ENS development. Analysis of Zeb2;Edn3 and Zeb2;Ednrb double mutant suggest also an interaction between these genes. Future experiments will help us to confirm these results and to determine the cellular and molecular origin of these interactions. These studies will enable us to better apprehend the molecular bases of these diseases, and to understand the origin of the enteric anomalies observed in patients. This knowledge may also help to develop new therapeutic strategies

För att diagnostisera malignt melanom och dess undergrupper används immunhistokemi för att färga in celler som uttrycker specifika protein. Särskilt desmoplastiska melanom kan initialt felbedömas baserat på utseendet som ärr, och det är därför viktigt att ha sensitiva antikroppar för att diagnostisera dem. Elva arkiverade patientprover (varav tre epiteloida melanom, fyra spolcelliga, tre desmoplastiska och ett akralt lentiginöst) färgades in med antikroppar mot CK18, HMB-45, Melan-A, S100, SOX10 och synaptophysin. Alla prover var negativa för CK18, nio var positiva för HMB-45 och Melan-A (de negativa var båda desmoplastiska melanom) och alla var positiva för S100 och SOX10. Synaptophysin var positiv i alla epiteloida melanom och det akralt lentiginösa melanomet, två av de fyra spolcelliga melanomen och negativ i alla desmoplastiska melanom. Fördelarna med SOX10, som tidigare studier i ämnet har påvisat, observerades inte i denna studie, troligen på grund av begränsningen i patientmaterial. Trots det verkar SOX10 vara ett användbart tillägg i melanompanelen, eller kanske till och med kan ersätta S100, baserat på tidigare studier.

Genetic interactions that underlie developmental processes such as cell differentiation and pattern formation are complex and difficult to elucidate. Neural Crest (NC) cells and their derivatives offer an optimal system in which to probe for these complex interactions as they acquire different cell fates and constitute a variety of structures. The transcription factors Sox10 and Pax3 as well as the transmembrane receptor Endothelin receptor b (Ednrb) are temporally and spatially co-expressed early in NC cells and mutations in these genes lead to similar hypopigmentation phenotypes due to a reduced number of NC-derived melanocyte precursors, the melanoblasts. The goal of this study was to establish whether Sox10 and Ednrb or Pax3 and Ednrb interact to promote normal murine melanocyte development. Crosses of Sox10 or Pax3 with Ednrb heterozygous mutants showed that the double heterozygous hypopigmentation phenotype was significantly more pronounced than phenotypes of single heterozygotes, implying that a synergistic interaction exists between Sox10 and Ednrb and Pax3 and Ednrb. This interaction was further explored by the attempt to rescue the Sox10 and Pax3 hypopigmentation phenotypes by the transgenic addition of Ednrb to melanoblasts. Pigmentation was completely restored in the Sox10 and partially restored in the Pax3 mutant mice. The comparison of the number of melanoblasts in transgenic and non-transgenic Sox10 mutant embryos showed that the transgenic rescue occurred as early as E11.5, a critical time for melanoblast population expansion. Cell survival assays indicated that the rescue was not due to an effect of the transgene on melanoblast survival. A novel phenotype arose when studying the interaction between Ednrb and Pax3. Newborns appeared normal but by 3.5 weeks of age, the affected pups were smaller than normal littermates and developed a dome-shaped head; some also developed thoracic kyphosis. Affected pups were dead by 4 weeks of age: 80% were Pax3Sp/+ and 75% were female. When compared to normal littermates, affected mice had brains with enlarged 4th ventricles and more glia while skeletal staining showed kyphosis, wider rib cages and pelvic differences. An epistatic interaction resulting from the mixing of genetic backgrounds that is exacerbated in the presence of Pax3 heterozygosity is suspected.