Rhabdomyosarcoma (RMS) is a skeletal muscle-derived tumour and is the most common soft tissue sarcoma of childhood. RMS tumours arise due to defects in differentiation, and as a result proliferate indefinitely. Pannexins consist of three members (pannexin1, 2, and 3) and are known to form single membrane channels. Reports have shown expression of pannexin3 (Panx3) as both the expected molecular weight species at ~43 kDa as well as a ~70 kDa immunoreactive species. While studies have begun to report on the function of the ~43 kDa form, the exact identity and function of the ~70 kDa immunoreactive species of Panx3 remains poorly understood. It has recently been reported that the ~70 kDa immunoreactive species of Panx3 is highly expressed in proliferative, non-differentiated human primary skeletal muscle myoblasts (HSMM), becoming drastically down regulated during differentiation. Indeed, knockdown of the ~70 kDa immunoreactive species of Panx3 inhibited proliferation without inducing differentiation in skeletal muscle myoblasts. We thus hypothesized that the ~70 kDa immunoreactive species of Panx3 would be upregulated in RMS supporting this proliferative phenotype. Here we now show that the ~70 kDa immunoreactive species of Panx3 is increased in RMS cell lines and tumours to a level similar to that seen in undifferentiated HSMM and fetal tissues, respectively. Further characterization of this species revealed that it is indeed a glycoprotein, an intrinsic characteristic of all pannexin members, it is recognized by two Panx3 antibodies targeting distinct epitopes, and it is reduced with Panx3 specific shRNA. Reduction of levels of the ~70 kDa immunoreactive species of Panx3 resulted in a significant decrease in proliferation of RMS cells without inducing differentiation. Taken together, these data suggest that the ~70 kDa immunoreactive species of Panx3 might be involved in keeping undifferentiated RMS cells in a proliferative state and that reduction of its levels or functions may be beneficial for RMS.

Les sarcomes sont des tumeurs malignes des tissus conjonctifs, représentant moins de 1%des tumeurs malignes de l’adulte, mais près de 8% de l’ensemble des cancers pédiatriques. Enraison de leur rareté, de leur grande variété histologique et de leur potentiel évolutifhétérogène, les sarcomes sont des pathologies difficiles à traiter, tant sur le plan diagnostique,pronostique que thérapeutique. Ces dernières années, l’avènement de techniques d’analyse pangénomiques par biologie moléculaire a permis d’améliorer la prise en charge clinique des sarcomes, mais les microARN sont des biomarqueurs émergents encore peu utilisés. au cours de c e travail de thèse, nous avons cjhoisi d'étudier la valeur des profils d'expression des micrfoARN dans les rhabdomyosarcomes et les ostéosarcomes. Les données brutes des profils d'expression ont été obtenues à l'aidre d'une technologie à moyen débit basée sur des réactions de PCR quantitative. Nous avons tout d'abord développé une méthodologie d'ananlyse permettant d'obtenir des données d'expression précises, reproductibles et à forte valeur ajoutée, à partir de matériel biologique hétérogène.. Dans un second temps, nous avons montré que les profils d'expression de microARN permettent d'améliorer la prise en charge clinique des deuc types de sarcomes étudiés : il est possible d'affiner la classification nosologique des rhabdomyosarcomes, et de prédire la réponse des ostéosarcomes à la chimiothérapie néo-adjuvante. La recherche de nouvelles applications cliniques liées aux profils d'expression des micorARN doit donc être poursuivie, et peut désormais l'être grâce à l'outil robuste que nous avons développé au cours de cette thèse. / Sarcomas are malignant soft tissue tumors, accounting for 1% of adult tumors and 8% of all pediatric malignancies. Sarcomas are rare, and display a variety of histological subtypes and clinical characteristics. Therefore, everyday management is difficult in terms of diagnosis,prognosis and treatment. Recently, the development of pangenomic molecular techniquesimproved the clinical management of sarcomas, but the use of microRNAs as biomarkers is still being investigated.In the present work, we studied the value of microRNA expression profiles inrhabdomyosarcomas and osteosarcomas. Raw data of expression profiles were obtained using amedium throughput technology based on quantitative PCR. We first developed an analysismethodology to gain accurate, reproducible and relevant expression data, starting fromheterogeneous samples. Furthermore, we showed that microRNA expression profiles canimprove the clinical management of both sarcoma entities: they are helpful to upgrade the fine nosological classification of rhabdomyosarcomas, and they are able to predict the response of osteosarcomas to neoadjuvant chemotherapy. Searching for new clinical applications tomicroRNA expression profiles must be pursued.