La leucémie myélomonocytaire chronique (LMMC) est une hémopathie myéloïde rare attribuée à l’accumulation d’évènements génétiques et épigénétiques dans une cellule souche ou progénitrice hématopoïétique. Les altérations génétiques somatiques récurrentes qui caractérisent cette maladie ont été identifiées: elles associent des altérations cytogénétiques non spécifiques chez 30% des patients et des mutations des gènes de la régulation épigénétique, de l’épissage, de la signalisation et de la transcription. Si certaines mutations influencent le phénotype (les mutations de RUNX1 génèrent une thrombopénie, celles de la signalisation une maladie proliférative, celles de KIT une mastocytose), elles ne sauraient résumer à elles seules l’expression phénotypique de la maladie. D’ailleurs, les médicaments hypométhylants restaurent une hématopoïèse équilibrée en modifiant le contexte épigénétique des cellules malades sans les éliminer. Il n’existe pas de lignée cellulaire de LMMC et les modèles murins n’en récapitulent que très partiellement les caractéristiques. L’objectif de mon travail de thèse a été de générer des cellules modélisant la maladie. J’ai transformé les cellules CD34+ de 2 patients en clones de cellules souches induites reprogrammées. Les clones obtenus à partir de l’un des patients ont été écartés du fait d’altérations génétiques supplémentaires acquises lors de la reprogrammation. Nous avons focalisé nos travaux sur 5 clones établis à partir des cellules CD34+ de l’autre patient et de 5 clones établis à partir de cellules CD34+ de deux sujets sains. Nous avons capturé 2 étapes de l’évolution moléculaire du clone, sans puis avec mutation KRASV12G. La différenciation hématopoïétique de ces clones en milieu semi-solide ou liquide récapitule les principales caractéristiques phénotypiques de la maladie. Par édition de gènes, nous avons ajouté dans certains clones la mutation SRSF2P95H observée dans les cellules de 50% des patients atteints de LMMC mais absente des cellules de la patiente étudiée. Nous montrons que l’hétérogénéité fonctionnelle et épigénétique des clones obtenus dépasse la seule hétérogénéité génétique et que la decitabine, un agent hypométhylant, a un effet cytotoxique faible mais améliorer l’équilibre de la production des cellules hématopoïétiques matures par des cellules génétiquement altérées. / Chronic myelomonocytic leukemia (CMML) is a rare hematological malignancy that has been related to the accumulation of genetic and epigenetic alterations in a hematopoietic stem or progenitor cell. Somatic recurrent mutations of coding DNA sequences have been in CMML cells, combining non-specific cytogenetic aberrations in 30% of the patients and mutations in epigenetic regulator, signal transduction, spliceosome and transcription factor genes. While some of these mutations directly affect disease phenotype (mutations in RUNX1 and thrombocytopeny, mutations in signaling pathways and proliferative disease, mutations in KIT and mastocytosis), they do not sum up the complex disease phenotype of this pathology on their own. Accordingly, hypomethylating agents restore a balanced hematopoiesis without eliminating clonal cells. There is no CMML cell line and murine models only partially recapitulate the disease. The objective of my thesis work was to reprogram hematopoietic stem/progenitor cells in order to model the disease heterogeneous expression. The clones established from one patient cells were discarded as their genetic background had been altered by reprogramming and cell culture. We analyzed in more details the behavior of 5 induced pluripotent stem cell lines established from a second patient and 5 other clones established from 2 healthy donor cells. We had captured 2 distinct genetic backgrounds of the patient clone, without or with KRASG12D mutation. Hematopoietic differentiation of these clones in semi-solid and liquid medium recapitulated the main characteristics of disease phenotype. With a gene editing tool, we introduced in some clones the SRSF2P95H mutation, observed in 50% of patient with CMML but missing in the studied patient. We noticed that functional and epigenetic heterogeneity of the clones exceeded their genetic heterogeneity and that the demethylating agent decitabine had limited cytotoxic effect but restored a more balanced production of hematopoietic cells by genetically abnormal cells.

No description available.

RNA modifications and processing events are important modulators of global gene expression. Genomic mutations in the RNA methylase NSun2 and the alternative splicing factor Srsf2 are linked to neurological disorders and cancer in humans, respectively. NSun2 methylates cytosine-5 in most tRNAs and, to a lesser extent, other ncRNAs and mRNAs. Srsf2 is a critical component of the spliceosome and interacts with abundant ncRNAs that are methylated by NSun2. However, how precisely these processes effect homeostasis is largely unexplored. Therefore, the main aims of my PhD were (1) to dissect the molecular mechanisms of NSun2-mediated RNA methylation pathways that regulate cell survival under normal conditions and in response to oxidative stress, and (2) to investigate the importance of Srsf2 in stem cells using skin as a model system. In the context of RNA modifications, firstly I described how NSun2-expressing cells enrich for transcripts related to enhanced cell survival. Subsequently, by metabolically profiling wildtype and patient-derived dermal fibroblasts carrying loss-of-function mutations in the NSUN2 gene, I showed that the absence of NSun2 is synonymous to an energy-saving, low-translating and stressed cellular state. I further confirmed that lack of NSun2 was sufficient to instigate a cellular stress response, by monitoring BIRC5, a member of the inhibitor of apoptosis family. To further answer whether lack of NSun2 enhanced the susceptibility of patient cells to external stress stimuli, I next exposed them to oxidative stress and measured transcriptional and translational changes. I discovered that NSun2 is required to adapt global protein synthesis to the stress response, while NSun2-depleted cells failed to do so. This was concurrent with NSun2-depleted cells enriching for transcripts related to mRNA degradation and negative regulators of protein translation in response to stress. Generally, since loss of NSun2-driven methylation in tRNAs triggers their cleavage into small ncRNA fragments by angiogenin, I asked how angiogenin or tRNA-derived ncRNAs affect translation levels. In the presence of NSun2, angiogenin alone did not reduce global protein synthesis, yet tRNA fragmentation was required to modulate translation levels. Finally, to uncover how the lack of NSun2 influenced tRNA cleavage and methylation patterns in response to stress, I exposed wildtype and patient cells to sodium arsenite and measured the abundance of tRNA-derived fragments and occurrence of methylation events. With this I discovered unique tRNA fragmentation patterns and global RNA methylation profiles for wildtype and NSun2-depleted cells, that can account for the underlying molecular and phenotypical differences in response to stress. In the context of alternative splicing, and since the cellular functions of Srsf2 are largely unknown, I explored its role in cellular survival and differentiation. By conditionally deleting SRSF2 in two different stem cell populations of the mouse epidermis, I observed significant thickening of the epidermis, altered expression of cell proliferation and stem cell differentiation markers, and distorted hair follicle structures. Moreover, I demonstrated that lack of Srsf2 promotes skin regeneration following injury, thus strongly indicating that Srsf2 is required for normal skin development and regeneration after injury. In summary, my research suggests that NSun2-mediated RNA methylation pathways orchestrate transcriptional and translational programmes in response to external stress stimuli, and my studies are the first to show that the alternative splicing factor Srsf2 is required for stem cell differentiation in skin.