Thèse de doctorat
Résumé : The thyroid gland is a critical organ whose development is a complex morpho-genetic process during which naïve endodermal cells acquire thyroid cell identity, bud out of the endoderm, and move toward their final location organising into hormone-producing 3D structures named follicles. Defects during any of these developmental stages lead to congenital hypothyroidism. Congenital hypothyroidism represents the most common congenital endocrine disorder in humans, affecting around 1 in every 2000-5000 live births. To date, the cause of most CH cases remains unexplained due to poor understanding of the molecular regulation of thyroid development. Although recent studies started to describe the external signals driving thyroid specification, little is known about the intrinsic factors driving the thyroid specification, the different steps of its morphogenesis, its functional maturation or maintaining the thyroid function once it is fully developed.To deal with this situation, we decided to take two different approaches. In the first, we analysed, at a single-cell resolution, the transcriptome of the adult thyroid gland and its surrounding tissue. Analysis of this dataset revealed the existence of two thyroid cell populations co-existing in the thyroid gland. Interestingly, these two populations showed differential expression of thyroid functional markers (such as slc5a5, duox, duoxa or tpo) and of pax2a. To confirm this observation and better characterise these two populations, we developed a new generation of a pax2a reporter line using Crispr/Cas9-based knock-in strategy. Using this line, we were able to observe the distribution of these two populations and show that they are stochastically distributed across the thyroid follicles forming the adult gland. Overall, these analyses lead us to believe that the thyroid gland consists of two functionally distinct populations of thyroid follicular cells, of which the less active population might consist of resting thyroid cells.For the second approach, we combined thyroid cell sorting with RNA sequencing and generated a dynamic gene expression profile of the developing thyroid at 44, 52 and 72hpf. Analysis of this profile highlighted several hundreds of novel genes with strongly enriched expression and hitherto unknown function during thyroid development. To identify their role in this process, we developed a F0 Crispr/Cas9-based mutagenesis assay. This strategy was successfully validated by targeting genes known be involved in thyroid development and/or function (pax2a, nkx2.4b, tshr, duox and duoxa) and for which high mutagenesis efficiency was confirmed by next generation sequencing.With this assay, we identified nrf2a, a transcription factor implicated in the cell’s oxidative stress response, as a potent actor of the thyroid maturation. Analysis of our nrf2a stable mutant zebrafish line revealed that, despite a normal early thyroid development, the thyroid of mutant fish is unable to produce T4 and to express tg. To translate our findings into a mammalian model and to assess its conservation across species, we generated in vitro differentiated mouse embryonic stem-cells (mESC) derived thyroid organoids lacking functional Nrf2. Differentiation of Nrf2 KO mESC line confirmed the critical role of Nrf2 during thyroid development also in vertebrates. Moreover, in both Nrf2/nrf2a KO models, mutated thyroids are unable to express tg and produce hormones at physiological levels. These results highlight the mechanistic conservation of Nrf2/nrf2a role during thyroid development in non-mammalian and mammalian vertebrates.Our results show that Nrf2 plays a crucial role in thyroid functional maturation across species and suggests that a deregulation of the Nrf2 activity during embryonic development could lead to congenital hypothyroidism.