La métabolomique par résonance magnétique nucléaire (RMN) permet d’étudier laréponse métabolique globale d’un système biologique à un stimulus ou un événementphysiopathologique (maladie, manipulation génétique, etc.). Cette discipline connaît un essorimportant dans la recherche clinique et biologique, et constitue ainsi un outil à fort potentielpour la découverte de biomarqueurs de maladies, et l’étude de la fonction des gènes.Cette thèse est dédiée à l’application de la métabolomique par RMN à hauts champspour l’étude des pathologies associées à la signalisation thyroïdienne chez la souris. L’objectifglobal est d’identifier des biomarqueurs spécifiques liés aux différentes maladies hormonales :l’hypothyroïdie et la maladie génétique émergente résistance à l’hormone thyroïdienne due àune mutation au niveau du récepteur TRα1 (RTHα). Cette dernière est particulièrementdifficile à diagnostiquer à cause du manque de marqueurs biochimiques et de symptômesspécifiques à cette maladie. De plus, elle présente des similitudes avec l’hypothyroïdie auniveau symptomatique. Des modèles murins de RTHα et de l’hypothyroïdie ont été analysés,et l’investigation a été menée sur l’urine et le plasma sanguin dans le but de différenciermétaboliquement ces maladies et d’identifier des biomarqueurs spécifiques à RTHα. Dessignatures métaboliques liées à chaque maladie ont été identifiées dans l’urine et le plasmasanguin. Cinq métabolites qui varient de façon significative ont été identifiés dans l’urinecomme étant liés à la maladie RTHα : trimethylamine, dimethylamine, isovalerylglycine, Nacetylglucosamineet la choline. Dans le sang, ce sont les lipides insaturés qui varient de façonsignificative chez les souris mimant la maladie RTHα. / Metabolomics by nuclear magnetic resonance (NMR) allows studying the metabolicresponse of a global biological system to a stimuli or a physiopathological even (diseases,genetic modifications, etc.). This discipline is growing especially in the clinical and biologicalfields, and represents a strong potential tool to identify biomarkers related to diseases, andstudy the function of genes.This thesis is dedicated to the application of metabolomics by high field NMR to studythyroid signalisation pathologies in mice. The main goal is to identify biomarkers related tothe emerging genetic disease called resistance to thyroid hormone due to a mutation in thyroidhormone receptor TRα1 (RTHα). This disease is particularly difficult to diagnose because ofthe lack of biochemical markers and specific symptoms. In addition, it presents commonfeatures with hypothyroidism in term of symptoms. Mice models of RTHα andhypothyroidism were analysed, and the investigation were driven on urine and blood plasmain order to differentiate metabolically theses diseases and identify biomarkers related toRTHα. Metabolic fingerprints related to each disease were identified in both urine and bloodplasma. Five metabolites vary significantly in the urine of RTHα mice: trimethylamine,dimethylamine, isovalerylglycine, N-acetylglucosamine and choline. Unsaturated lipids varysignificantly in the blood plasma of RTHα mice.The impact of thyroid hormones (TH) and the thyroid hormone receptor TRβ on theliver metabolism were also studied in the present manuscript through NMR-basedmetabolomics. A mouse model, with a specific knock-out of TRβ gene in hepatocytes (LTRβ-KO), were used to study this question. To understand the function of TH mediated by TRβ,the liver metabolic response to TH, obtained from liver aqueous extracts and intact livertissues, TRβKO and wild-type mice were compared. The results suggest the presence ofdirect and indirect effects of thyroid hormones on the liver metabolism.

In animals, some of the most critical regulators of gene expression are nuclear hormone receptors (NRs) and their coregulators, specifically the Mediator complex. Of particular interest are the NRs implicated in metabolic and developmental regulation and in carcinogenesis: thyroid hormone receptors (TRs) and retinoid X receptors (RXRs). In this work, I venture to elucidate some aspects of gene expression regulation by these NRs: the degree of evolutionary conservation of signalling based on NRs and their coregulators; the mechanisms of negative regulation by NRs; and possible implications of these findings for clinical medicine. State-of-the-art bioinformatical, genome editing and microscopic techniques are applied at three levels of animal evolution to study NRs and Mediator. Reverse genomics in human patients suffering from the syndrome of resistance to thyroid hormones β are used to infer the structure and function of TRβ subdomains. Alignments, binding studies and in vivo experiments in Trichoplax adhaerens allow identification of a close orthologue of human RXR at the basis of metazoan evolution. Employing database queries, genome editing and microscopy, we describe a correct orthologue of the Mediator subunit 28 in Caenorhabditis elegans, indicating a complete homology of the Mediator complex...