par Tourneur, Adrien 
Président du jury Langer, Ingrid
Promoteur Detours, Vincent
Publication Non publié, 2024-07-02
Président du jury Langer, Ingrid
Promoteur Detours, Vincent
Publication Non publié, 2024-07-02
Thèse de doctorat
| Résumé : | Thyroid cancer is the most prevalent endocrine malignancy. Among thyroid cancers, papillary thyroid cancer (PTC) is the most abundant form. Anaplastic thyroid cancer (ATC) is the most aggressive form. PTC displays a high variability in its phenotype and clinical outcomes. Variable characteristics in papillary tumors include their purity, and composition. Those two characteristics impact the extent of those interactions and the underlying mechanisms of tumor cell interactions with their microenvironment. Yet those interactions are still poorly understood. Similarly, anaplastic thyroid cancer displays a high potential for interactions between tumor cells and their microenvironment as most of those tumors are of low purity, with macrophages representing as much as 50% of the cells in some tumors. Again, the interactions between ATC cancer cells and macrophages are poorly understood.To investigate those open questions in an unbiased way we profiled ATC and PTC samples with whole transcriptome sequencing technologies. Single nuclei RNA sequencing provided us a single cell-level resolution, and spatial transcriptomics provided us with spatial context. The human tumor samples of our cohort were profiled in parallel with these two high throughputs technologies and were enriched with follow-up multiplex immunofluorescence and DNA sequencing.Based on this dataset, we observed unexpectedly high amounts of B lymphocytes in thyroiditis associated samples and expectedly high amounts of macrophages in ATC samples. But, surprisingly, ATC macrophages had highly dissimilar transcriptomes when comparing different ATC samples. This indicates a major influence of the sample’s unique characteristics on the transcriptome of those macrophages.We studied the epithelial cell states and identified three main clusters of epithelial origin expressing varying levels of epithelial markers. We defined the epithelial cell states by order of dedifferentiation using four canonical thyroid differentiation markers. Those markers display a subset inclusion relationship. They gradually disappear as cells dedifferentiate, with fully differentiated cells expressing all of them and fully dedifferentiated cells expressing none of them.We also analyzed a single cell RNA sequencing dataset from human thyroid organoids generated from pluripotent stem cells. This dataset included thyrocytes at different levels of differentiation. The least differentiated were progenitor cells expressing low levels of thyroid differentiation markers. The most differentiated were the fully differentiated and functional thyrocytes expressing all thyroid differentiation markers. We observed that those differentiating organoids gradually expressed higher levels of the same four canonical thyroid 8 differentiation markers mentioned above. The order of their gain in differentiating organoids was mirrored by the order of their loss in dedifferentiating cancer cells. This indicates the fundamental nature of the dedifferentiation process mentioned above.Beyond this fundamental ordering, we also observed that the next source of variation in those epithelial clusters was the sample of origin. A sub-clustering of each of those main cancer clusters clearly showed that the specificities of each tumor made those cancer cells different between samples, i.e. idiosyncratic. Some of these idiosyncrasies were linked to specific genetic events, others remained unexplained.PTC cancer cells harboring the BRAFV600E mutation did not display signs of full EMT as observed in ATC, but still acquired some mesenchymal features. Those include the expression of fibronectin 1 (FN1), a fibrillar protein excreted in the extracellular matrix. Ligand- receptor analysis showed that those PTC cancer cells both produced the highest amounts of FN1 compared to other cell types, but also that they expressed the receptors required to receive signals from it. Considering that FN1 has previously been associated with some aggressiveness features in thyroid cancer, this potential autocrine positive feedback loop was investigated in vitro. A mouse organoid model previously developed in the institute was used to produce fully functional thyroid organoids and transform them to cancerous organoids with or without a medium enriched in FN1. Those experiments showed that FN1 has an impact on cancerous thyrocyte proliferation. |
