Abstract : During embryonic development, the posterior part of vertebrates brain, or hindbrain, goes through a transient segmentation process along the anterio-posterior axis, leading to the definition of metameric units called rhombomeres. They correspond to neuronal differenciation units and allow the organization of cranial nerves. They also play a major role in cranio-facial structures development through regionalized production of neural crest. The Krox20 gene encodes a transcription factor which is expressed in rhombomeres r3 and r5. It has been shown that its expression in these domains is necessary to define and maintain them, as well as to generate the corresponding cranial nerves properly. Krox20 controls indeed the transcriptional expression of various genes implicated in particular in rhombomeres identity and intactness. Some cofactors of Krox20, as Nab or HCF-1, had been previously described but their role in vivo had not been addressed, in particular during hindbrain development. The first aim of my PhD was to better understand the mechanisms on which Krox20 activity relies in vivo and in particular the regulating roles of some of its cofactors. To address these questions, I performed a structure-function study in vivo using chick embryo as a model. I showed that Krox20 activity during hindbrain development relies on different domains of the protein depending on the transcriptional target considered. I also showed that Nab proteins are implicated in a negative feedback loop regulating Krox20 activation function, whereas HCF-1 is a positive cofactor of Krox20 for the activation of two of its transcriptional targets. Furthermore, this study allowed us to characterize new interaction sites between HCF-1 and Krox20. Krox20 is also a key factor of the myelination process in the peripheral nervous system (PNS), on which the rapid saltatory conduction along nerves relies. Various human pathologies are due to hypomyelination or demyelinating events. A Krox20 knock-out mouse model previously generated in the laboratory had shown a lack of myelination of the PNS in the absence of Krox20. This was due to a defect of differenciation of the myelinating Schwann cells in which Krox20 is normally expressed. Following this study, various mutations in Krox20 had been characterized in hereditary myelinopathies of the PNS, such as Charcot-Marie-Tooth diseases (CMT). In particular, one of these mutations, that leads to a severe early-onset subtype of recessive CMT, appears to abolish the interaction between Krox20 and its cofactors Nab. The second aim of my PhD was to characterize the role of Nab interaction with Krox20 during the myelination process and the hindbrain segmentation in vivo. For this purpose, I generated a mouse model of CMT carrying the mutation abolishing Krox20/Nab interaction. Homozygous mutant mice show a phenotype similar to what is observed in human patients. These mice have locomotion defects evolving progressively to limbs paralysis and are affected by a severe defect in PNS myelination. My studies showed that the myelination process is first delayed, leading subsequently to an hypomyelination of the PNS, followed by demyelination. These defects appear to be linked to a delay in proliferation arrest of Schwann cells and a disregulation of gene expression in these cells. Surprisingly, Nab proteins appear to be positive cofactors of Krox20 in Schwann cells. Furthermore, I showed that these mutant mice present cranial nerves abnormalities which are compatible with functional impairments observed in human patients carrying this mutation. In summary, the studies made during my PhD allowed to show the complexity of Krox20 transcriptional activity, relying on different target-specific domains of the protein, as well as to better characterize the role of two of its cofactors during hindbrain development. This work also lead to the generation of a mouse model for a severe form of CMT, allowing a better understanding of its complex pathomechanisms.