Le cancer papillaire de la thyroïde (PTC) représente 70-80% des cas de cancers de la thyroïde. Il est principalement caractérisé par des réarrangements chromosomiques affectant le gène RET. Le réarrangement RET/PTC1, dans lequel RET est réarrangé avec un gène proapoptotique H4, représente 30% des cas sporadiques et jusqu’à 60% des cas survenus après irradiation. Afin d’inhiber l’oncogène de fusion RET/PTC1, nous avons utilisé un siRNA ciblant la zone de jonction RET/PTC1 (siRNA RET/PTC1) au niveau de l’ARN messager des cellules tumorales et montré sa spécificité et son efficacité. Néanmoins, le développement des siRNAs comme molécule d’intérêt thérapeutique se heurte in vivo à des difficultés liées à leur administration. Sous forme libre, ces molécules sont, en effet, très vite dégradées par les nucléases extracellulaires et leur pénétration intracellulaire est limitée. C’est la raison pour laquelle il est nécessaire de les vectoriser. Nous avons choisi de le faire par la méthode de « squalénisation » et avons couplé d’une manière covalente le squalène, un lipide naturel précurseur de la biosynthèse du cholestérol, au siRNA RET/PTC1. Le bioconjugué formé s’autoassemble spontanément en milieu aqueux sous forme de nanoparticules stables de 170 nm de diamètre. L’efficacité et la toxicité des nanoparticules siRNA RET/PTC1-squalène ont été étudiées in vitro dans deux lignées de PTC exprimant RET/PTC1 (BHP10-3 et TPC-1) et l’activité antitumorale a été évaluée in vivo sur des souris athymiques xénogreffées par BHP10-3 puis traitées en i.v. par ces nanoparticules. Les nanoparticules siRNA RET/PTC1-squalène ont montré une bonne efficacité antitumorale. En revanche, aucune activité inhibitrice n’a été retrouvée in vitro. En conclusion, nous avons réussi à vectoriser le siRNA RET/PTC1 par la méthode de squalénisation. Cette étude ouvre des perspectives thérapeutiques pour certains patients atteints de PTC et réfractaires au traitement conventionnel. / The papillary thyroid carcinoma (PTC) is the most common type of thyroid malignancy. This tumour is associated with somatic mutations of the RET proto-oncogene, due to gene rearrangements of the proto-RET. RET/PTC1 rearrangement is the most common genetic alteration identified to date, it is formed by an intra chromosomic rearrangement which leads to the juxtaposition of the RET Tyrosine Kinase domain of the proto-RET with the gene H4. The fusion RET/PTC1 oncogene represents an interesting target for small interfering RNA (siRNA) strategies since it is present only in the tumour cells and not in the surrounding normal cells. However, the biological efficacy of the siRNAs is hampered by their short plasma half-life due to poor stability in biological fluids and low intracellular penetration. In order to protect siRNA from degradation, and to improve their intracellular capture, we applied the concept of “squalenoylation”, ie. The bioconjugation of a drug substance to squalene, for the delivery of siRNA targeted toward the RET/PTC1 fusion oncogene. The acyclic isoprenoid chain of squalene was covalently coupled with RET/PTC1 siRNA at the 3’-terminus of the sense strand via a stable thioether linkage. The linkage of RET/PTC1 siRNA to squalene leads to an amphiphilic molecule that self-organise in water as RET/PTC1 siRNA-SQ nanoassemblies of 170 nm and Zeta potential of -26.4 mV. These RET/PTC1 siRNA-SQ NPs did not showed any cytotoxicity in vitro. Interestingly, in vivo, in a mouse xenografted RET/PTC1 experimental model, RET/PTC1 siRNA-SQ nanoparticles inhibited tumour growth, RET/PTC1 oncogene and oncoprotein expression, after intravenous injections of 2.5 mg/kg cumulative dose. In the last of this work, GALA-cholesterol combination with siRNA-SQ NPs further enhanced nucleic acid internalization, promoted their escape into the cytosol and consequently their gene silencing efficiency in vitro. In conclusion, these results showed that the “squalenoylation” offers a new non cationic plate-form for the siRNA delivery.

The CBFbeta-SMMHC fusion protein is expressed in acute myeloid leukemia (AML) samples with the chromosome inversion inv(16)(p13;q22). This fusion protein binds the transcription factor RUNX with higher affinity than its physiological partner CBFbeta and disrupts the core binding factor (CBF) activity in hematopoietic stem and progenitor cells. Studies in the Castilla laboratory have shown that CBFbeta-SMMHC expression blocks differentiation of hematopoietic progenitors, creating a pre-leukemic progenitor that progresses to AML in cooperation with other mutations. However, the combined function of cumulative cooperating mutations in the pre-leukemic progenitor cells that enhance their expansion to induce leukemia is not known. The standard treatment for inv(16) AML is based on the use of non-selective cytotoxic chemotherapy, resulting in a good initial response, but with limited long-term survival. Therefore, there is a need for developing targeted therapies with improved efficacy in leukemic cells and minimal toxicity for normal cells. Here, we used conditional Nras+/LSL-G12D; Cbfb+/56M; Mx1Cre knock-in mice to show that allelic expression of oncogenic N-RasG12D expanded the multi-potential progenitor (MPP) compartment by 8 fold. Allelic expression of Cbfbeta-SMMHC increased the MPPs and short-term hematopoietic stem cells (ST-HSCs) by 2 to 4 fold both alone and in combination with N-RasG12D expression. In addition, allelic expression of oncogenic N-RasG12D and Cbfbeta-SMMHC increases survival of pre-leukemic stem and progenitor cells. Differential analysis of bone marrow cells determined that Cbfb+/MYH11 and Nras+/G12D; vii Cbfb+/MYH11 cells included increased number of blasts, myeloblasts and promyelocytes and a reduction in immature granulocytes, suggesting that expression of N-RasG12D cannot bypass Cbfbeta-SMMHC driven differentiation block. N-RasG12D and Cbfbeta-SMMHC synergized in leukemia, in which Nras+/G12D; Cbfb+/MYH11 mice have a shorter median latency than Cbfb+/MYH11 mice. In addition, the synergy in leukemogenesis was cell autonomous. Notably, leukemic cells expressing N-RasG12D and Cbfbeta-SMMHC showed higher (over 100 fold) leukemia-initiating cell activity in vivo than leukemic cells expressing Cbfbeta-SMMHC (L-IC activity of 1/4,000 and 1/528,334, respectively). Short term culture and biochemical assays revealed that pre-leukemic and leukemic cells expressing N-RasG12D and Cbfbeta-SMMHC have reduced levels of pro-apoptotic protein Bim compared to control. The Nras+/G12D; CbfbMYH11 pre-leukemic and leukemic cells were sensitive to pharmacologic inhibition of MEK/ERK signaling pathway with increasing apoptosis and Bim protein levels but not sensitive to PI3K inhibitors. In addition, knock-down of Bcl2l11 (Bim) expression in Cbfbeta-SMMHC pre-leukemic progenitors decreased their apoptosis levels. In collaboration with Dr. John Bushweller’s and other research laboratories, we recently developed a CBFbeta-SMMHC inhibitor named AI-10-49, which specifically binds to CBFbeta-SMMHC, prevents its binding to RUNX proteins and restores CBF function. Biochemical analysis in human leukemic cells showed that AI-10-49 has significant specificity in reducing the viability of leukemic cells expressing CBFbeta-SMMHC (IC50= 0.83μM), and negligible toxicity in normal cells. Likewise, mouse Nras+/G12D; viii Cbfb+/MYH11 leukemic cells were sensitive to AI-10-49 (IC50= 0.93μM). By using the NrasLSL-G12D; Cbfb56M mouse model, we also show that AI-10-49 significantly prolongs the survival of mice bearing the leukemic cells. Preliminary mechanistic analysis of AI-10-49 activity has shown that AI-10-49 increased BCL2L11 transcript levels in a dose and time dependent manner in murine and human leukemic cells, suggesting that the viability through BIM-mediated apoptosis may be targeted by both oncogenic signals. My thesis study demonstrates that Cbfbeta-SMMHC and N-RasG12D promote the survival of pre-leukemic myeloid progenitors primed for leukemia by activation of the MEK/ERK/Bim axis, and define NrasLSL-G12D; Cbfb56M mice as a valuable genetic model for the study of inv(16) AML targeted therapies. For instance, the novel CBFbeta-SMMHC inhibitor AI-10-49 shows a significant efficacy in this mouse model. This small molecule will serve as a promising first generation drug for targeted therapy of inv(16) leukemia and also a very useful tool to understand mechanisms of leukemogenesis driving by CBFbeta-SMMHC.

Les cancers pédiatriques représentent la première cause de décès par maladie chez les enfants puisque 20% des patients présentent des résistances et rechutent. Les traitements actuels sont à leurs limites d'efficacité mais aussi de toxicité. L'amélioration des connaissances de la biologie des tumeurs résistantes et l'identification de nouvelles cibles moléculaires sont nécessaires pour ces pathologies rares, extrêmement diverses et mal comprises. Les oncogènes de fusion sont responsables de l’oncogenèse de ~ 20% des cancers. En raison de la biologie particulière des cancers pédiatriques, nous avons émis l'hypothèse qu’ils seraient dus à des oncogènes de fusion. Leur exploration, particulièrement à la rechute où les données actuelles sont limitées, permettrait de mieux comprendre ces pathologies et d’introduire de nouvelles cibles thérapeutiques. Notre premier objectif était de détecter de nouveaux transcrits de fusion parmi 48 patients pédiatriques en rechute inclus dans l'essai de médecine de précision MOSCATO-01 mené à Gustave Roussy. A partir de données de RNA-seq, nous avons détecté l’ensemble des fusions retrouvées dans ces tumeurs et classé les plus pertinentes selon leur fonction et leur potentiel thérapeutique. Un nouvel oncogène de fusion nommé LMO3-BORCS5 retrouvé chez un patient atteint de sarcome d'Ewing dans les biopsies tumorales à la rechute et au diagnostic, et dans diverses lignées cellulaires a particulièrement retenu notre attention. Nos études fonctionnelles montrent que LMO3-BORCS5 a un rôle important dans la tumorigenèse et la sensibilité au traitement et que BORCS5 fonctionnerait, in vivo, comme un gène suppresseur de tumeur. Ces résultats montrent le rôle critique des transcrits de fusion dans l'initiation et la progression tumorale ainsi que dans l’acquisition de résistance au traitement et que les fusions isolées ne doivent pas être ignorées. A l'avenir, la prise en charge des cancers réside en une médecine moléculaire de précision, les oncogènes de fusions représentant l'un des outils les plus remarquables en clinique. / Pediatric cancers represent the first cause of death by disease in children as 20% of patients harbor resistances and die from recurrence of the malignancy. Nowadays treatment are at their limits of efficacy but also toxicity. Improving the knowledge on the biology of the relapsed/resistant tumors and identifying new molecular targets are needed for these rare, extremely diverse and misunderstood pathologies. Fusion oncogenes are known to be major drivers of oncogenesis, responsible for ~20% of cancers. Due to their particular biology, we hypothesized that pediatric tumors would be more likely to harbor fusions and that their exploration, especially at relapse where very limited data are available, could help to understand the particularity of these malignancies and consequently to provide new therapeutic targets. Our first aim was to detect new fusion transcripts in a cohort of 48 pediatric patients at relapse included in the precision medicine trial MOSCATO-01, previously conducted in Gustave Roussy. From the RNA-seq data, we described the landscape of fusions in these tumors and classified the most confident candidates according to their biological function and druggability predictions. A new fusion named LMO3-BORCS5 found in biopsies of a patient with Ewing's sarcoma, at diagnosis and at relapse as well as in various cell lines particularly caught our attention. Functional studies showed that LMO3-BORCS5 has a high impact on tumorigenesis and treatment sensitivity and that BORCS5 would act, in vivo, as a tumor suppressor gene. Our results strengthened the critical role of fusion transcripts in tumor initiation, progression and resistance and reinforces the idea that a private fusion oncogene should not be ignored. Indeed, the future of cancer care lays in precision medicine, with fusion transcripts being one of the most remarkable tool for treatment decision.