Glioblastoma is one of the most devastating forms of primary brain tumor, with a survival of 14.6 months from the diagnosis. Despite the aggressive treatments used in the clinic, consisting of surgical resection followed by radiotherapy and concurrent chemotherapy using temozolomide, the absence of novel treatments and the development of resistance to standard-of-care therapies continue to position glioblastoma as the most challenging brain tumor in adults. The main factor contributing to the incurability of glioblastoma is the extensive heterogeneity. This heterogeneity, which is evident both at intratumoral and inter-patient levels, represents a substantial obstacle to the achievement of effective treatments. In this context, the use of single-cell resolution appears one of the most powerful means for understanding the intricacies and unravelling the heterogeneity of glioblastoma. Although single-cell RNA sequencing studies have provided and still provide valuable insights, they lack the essential spatial context that is critical for unravelling the heterogeneity of glioblastoma. This limitation impedes the understanding of interactions among distinct subpopulations and their intricate relationships with the neuronal microenvironment. To gain insights into the heterogeneity of glioblastoma, we used the spatially resolved RNA sequencing technology to analyse glioblastoma samples deriving from 3 different patients. For each patient, we focused on four distinct tumor regions, i. e. the proliferating tumor area, the necrotic core, the infiltrating area, and a distal healthy area. Cancer cells identified in the infiltrating regions exhibited a unique pattern of cell subpopulations, with the oligodendrocytes as the most represented in this area. In addition, we managed to generate patient-derived glioblastoma organoids from nearly all areas, with the tumor regions displaying the highest growth rates. These patient-derived organoids, which represent fundamental models useful to faithfully replicate the disease in vitro, may be employed in future analyses. Finally, we also derived glioblastoma stem cell cultures from the different tumor regions, with the proliferating tumor area showing the highest rate of success. In the second part, we aimed to gain a deeper understanding of the molecular mechanisms that underlie the development of glioblastoma, which is crucial for developing effective treatments, and characterise ELAVL2 role, highlighting its potential role as a potential tumor suppressor in glioblastoma. Collectively, our findings suggest that ELAVL2 promotes the exit of glioblastoma stem cells from quiescence, boosting their self-renewal capacity while also facilitating neuronal differentiation. The in vivo validation of our results using an orthotopic human glioblastoma stem cell xenograft model and a D. melanogaster genetic model strongly supports our findings and points to deletion of ELAVL2 as a factor that increases aggressiveness in glioblastoma stem cells and in vivo.

Cette étude a été conçue pour identifier les protéines qui interagissent potentiellement avec Déméthylase 4A spécifique de la lysine (KDM4A) dans le contexte du cancer en utilisant l’essai d'identification de la biotine dépendante de la proximité 2 (BioID2). KDM4A est une lysine déméthylase et un régulateur épigénétique qui joue un rôle dans la carcinogenèse en favorisant la prolifération. Nous avons cherché à identifier l'interactome protéique de KDM4A dans la lignée cellulaire du cancer du col de l'utérus HeLa. Ces interactions protéiques ont été caractérisées en fonction de leur dépendance à l’activité catalytique de KDM4A et / ou au domaine Tandem Tudor. De nouveaux interactants de KDM4A ont été détectés, tout en observant des partenaires protéiques précédemment identifiés, comme FBXO22. KDM4A semble interagir avec certains membres du complexe de remodelage de la chromatine pBAF, en particulier, ARID2, BRD7 et SMARCA2. Le complexe pBAF facilite ou empêche l'accessibilité à l'ADN en restructurant le nucléosome. Une analyse plus approfondie est nécessaire pour valider si l'interaction complexe KDM4A-pBAF est directe ou indirecte. Cette étude suggère également l’importance du domaine Tandem Tudor dans le rôle de KDM4A dans la réparation de bris double brin. Enfin, nous proposons également une implication potentielle de KDM4A dans l'épissage de l'ARNm et le transport d'anions organiques. Cette étude fournit de nouvelles informations sur le rôle de KDM4A dans le développement du cancer. / This study was designed to identify potential interacting proteins of Lysine-specific demethylase 4A (KDM4A) in the context of cancer using proximity-dependent biotin identification 2 (BioID2) assay. KDM4A is a lysine demethylase and an epigenetic regulator that plays a role in carcinogenesis by promoting proliferation. Herein, we sought out to identify the protein interactome of KDM4A in cervical cancer cell line HeLa. These protein interactions were characterized by their dependency on KDM4A’s catalytic activity and/or Tandem Tudor domain. It succeeded at detecting novel interactors of KDM4A as well as previously studied interactions, such as FBXO22. KDM4A seems to be interacting with some members of the pBAF chromatin remodeling complex, specifically, ARID2, BRD7, and SMARCA2. The pBAF complex facilitates or prevents accessibility to DNA by restructuring the nucleosome. Further analysis is required to validate whether the KDM4A-pBAF complex interaction is direct or indirect. This study also implied the importance of the Tandem Tudor domain in KDM4A’s role in the double stranded break repair. Finally, we also propose the potential involvement of KDM4A in mRNA splicing and organic anion transport. This study provides new insights into KDM4A’s role in cancer development.