What is a wetland? What does this category imply? Is it different from a land which is wet? This thesis addresses these questions through a study of environmental conflicts in two protected wetlands in Agro Pontino, Italy. This region, 70 kilometres south of Rome, was affected between the 1920s and the 1930s, by one of the most important operations of land reclamation of that time, conducted by the fascist regime which drained the largest marshland in Italy, i.e. the Pontine Marshes. Based on fieldwork in protected wetlands considered as the remnants of the Pontine Marshes, the thesis includes the analysis of various original case studies, such as the activities of Environmental Interpretation and farming, among others. Moreover, it discloses archive records that cast new light on the social and environmental context of the Pontine Marshes. The thesis explores how farmers and conservationists relate with a particular kind of environment, namely wetlands in conservation areas; their relations with non-humans; their different reading of the history and evolution of these landscapes and the connections between these aspects and the conflicts affecting these areas. These conflicts concern different ideas of what a wetland should be and the appropriate practices aimed at implementing those ideas. The research, which is influenced by the work of Tim Ingold, challenges the established notion of wetland, based on a topographic and spatial understanding of the world, suggesting that wetlands are patterns characterized by movements, tasks and activities: entanglements of innumerable life-lines traced by human and non-human beings in the accomplishment of their tasks. It is in this entanglement which politics emerge, conflicts flourish and life goes on. Accordingly, I argue that following these life-lines and exploring what I call the geography of the meshwork, can disclose new interpretative perspectives to understand environmental conflicts.

Les processus qui mènent à la découverte de nouveaux médicaments sont longs et fastidieux, et les taux de succès sont relativement faibles. L’identification de candidats par le biais de tests expérimentaux s’avère coûteuse, et nécessite de connaître en profondeur les mécanismes d'action de la protéine visée afin de mettre en place des essais efficaces. Le criblage virtuel peut considérablement accélérer ces processus en permettant une évaluation rapide de chimiothèques de plusieurs milliers de molécules afin de déterminer lesquelles sont les plus susceptibles de se lier à une cible. Ces dernières années ont ainsi été témoins de quelques success stories dans ce domaine.Le premier objectif de ce travail était de comparer différents outils et stratégies couramment utilisés dans le criblage virtuel “structure-based”, puis de les appliquer à des cibles protéiques à visée thérapeutique, en particulier dans le cadre du cancer.La protéine kinase GSK3 et un test set de ligands connus ont servi de modèle pour différentes études méthodologiques ayant pour but d’évaluer les programmes de docking et de scoring à notre disposition. En particulier, l’utilisation de plusieurs structures relaxées du récepteur ou l’insertion de torsions sur certains résidus du site actif pendant le docking ont permis d’évaluer l’influence de la flexibilité de la protéine. L’utilité et la pertinence d’outils permettant de générer automatiquement les structures 3D des ligands et de méthodes de consensus scoring ont également été étudiées.Un criblage virtuel de la Pontine, une ATPase impliquée dans la croissance tumorale pour laquelle aucun inhibiteur n’était connu, a permis la sélection de candidats issus de banques de données commerciales. Ces molécules ont été testées dans un essai enzymatique par le biais d’une collaboration, et quatre d’entre elles se sont révélées capable d’inhiber l’activité ATPase de la Pontine. Le criblage de bases de ligands synthétisés et imaginés dans l’équipe a également fourni un inhibiteur original. Au contraire, l’étude de la sPLA2-X humaine, une phospholipase dont l’activité catalytique est dépendante d’un atome de Ca2+ localisé au sein du site actif, a montré les limites de nos outils de docking qui n’ont pas été capables de gérer cet ion métallique et mis en évidence la nécessité de mettre en place d’autres outils. / The process of drug discovery is long and tedious. Besides, it is relatively inefficient in terms of hit rate. The identification of candidates through experimental testing is expensive and requires extensive data on the mechanisms of the target protein in order to develop efficient assays. Virtual screening can considerably accelerate the process by quickly evaluating large databases of compounds and determining the most likely to bind to a target. Some success stories have emerged in the field over the last few years.The objectives of this work were first, to compare common tools and strategies for structure-based virtual screening, and second, to apply those tools to actual target proteins implied notably in carcinogenesis.In order to evaluate the docking and scoring programs available, the protein kinase GSK3 and a test set of known ligands were used as a model to perform methodological studies. In particular the influence of the flexibility of the protein was explored via relaxed structures of the receptor or the insertion of torsions on the side chains of residues located in the binding site. Studies concerning the automatic generation of 3D structures for the ligands and the use of consensus scoring also provided insights on the usability of these tools while performing a virtual screening.Virtual screening of the human protein Pontin, an ATPase implied in tumor cell growth for which no inhibitors were known, allowed the prioritization of compounds from commercial databases. These compounds were tested in an enzymatic assay via a collaboration, and led to the identification of four molecules capable of inhibiting the ATPase activity of Pontin. Additional screens of in-house oriented databases also provided at least one innovative inhibitor for this protein. On the contrary, a study of the human PLA2-X, a phospholipase that requires a Ca2+ atom to bind to its active site in order to catalyze the hydrolysis of its substrate, revealed the limits of our docking tools that could not handle the metal ion and the need for new tools.

Le carcinome hépatocellulaire (CHC) est le principal cancer primitif du foie et est associé à un très mauvais pronostic. Notre équipe a mis en évidence que la Reptine et la Pontine, des AAA+ ATPases homologues, sont surexprimées dans le CHC par rapport au foie non tumoral. Au cours de ce travail de thèse, j’ai contribué à démontrer que l’extinction de la Reptine par l’induction de shRNA suffit à arrêter la croissance de tumeurs déjà établies, et même à induire leur régression dans des xénogreffes chez la souris. Ces résultats encourageants suggèrent que la Reptine pourrait être une cible thérapeutique dans le CHC. L’utilisation de siRNA en thérapeutique n’étant pas envisageable actuellement, il parait plus pertinent de tenter de cibler la Reptine via son activité ATPase. Le principal objectif de ma thèse était donc de déterminer l’implication de l’activité ATPase de la Reptine pour ses propriétés oncogéniques dans le CHC. Nos résultats ont montré que des mutants inactifs de la Reptine (D299N et E300G) ont un effet dominant négatif et ne sont pas capables de complémenter l’absence de la Reptine endogène, ce qui conduit à une diminution significative de la croissance des cellules HuH7 et Hep3B, et à une induction de l’apoptose. Ceci indique que l’activité ATPase de la Reptine est nécessaire pour la croissance et la survie des cellules de CHC. Enfin, grâce à une étude transcriptomique, nous avons identifié de nouveaux gènes dont l’expression est régulée par la Reptine et/ou la Pontine. Parmi ces gènes, certains pourraient être impliqués dans les fonctions oncogéniques de la Reptine et/ou de la Pontine dans le CHC. Finalement, ce travail a permis de mettre en évidence l’implication de l’activité ATPase de la Reptine, et d’apporter des éléments permettant de mieux comprendre le mécanisme d’action de la Reptine dans le CHC. / Hepatocellular carcinoma (HCC) is the main primary cancer of the liver and is often associated with poor prognosis. Our team has demonstrated that Reptin and Pontin, two AA+ ATPases, are overexpressed in HCC compared to normal liver. Moreover this overexpression is also associated with poor prognosis. In the course of my PhD, I demonstrated that shRNA-mediated silencing of Reptin is sufficient to inhibit tumor growth and even can promote their regression in xenografted mice. These encouraging results suggest that Reptin might represent a novel therapeutic target in HCC. As the use of siRNA as therapeutic tools is still debated, the targeting of Reptin enzymatic activity might represent a more relevant approach to impair its functions. To this end I first proposed to determine the involvement of Reptin ATPase activity in HCC oncogenesis. My results show that ATPase inactive Reptin mutants (D299N and E300G) play dominant negative roles toward Reptin functions and are unable to complement for the depletion of endogenous Reptin, thereby leading to a significant decrease of cell growth and to a significant increase of apoptosis in HuH7 and Hep3B cells. These results show that Reptin’s ATPase activity is necessary for HCC cell growth and survival. Moreover, using a transcriptomic approach that compared gene expression upon siRNA-mediated Reptin or Pontin silencing, we identified specific genes whose expression is under the control of those proteins and whose functions might provide mechanistic explanation to Reptin’s involvement in HCC. Collectively, the results obtained during my PhD thesis have characterized the contribution of Reptin ATPase activity to HCC growth and development and might represent a founding step in the understanding of Reptin’s biology in cancer development.

Tachykinins and their receptors are present in gustatory centers, but little is known about tachykinin function in gustation. In this study, immunohistochemical localization of substance P and two centrally prevalent neurokinin receptors, NK1 and NK3, was carried out in the rostral nucleus of the solitary tract and the caudal parabrachial nucleus to evaluate regional receptor/ligand correspondences. All three proteins showed regional variations in labeling density that correlated with distinct sites in gustatory centers. In the rostral nucleus of the solitary tract, the relative densities of substance P and NK1 receptors varied in parallel across subnuclei, with both being moderate to dense in the dorsocentral, chemoresponsive zone. NK3 receptors had a distinct distribution in the caudal half of this zone, suggesting a unique role in processing taste input from the posterior tongue. In the caudal parabrachial nucleus, substance P and NK1 receptor immunoreactivities were dense in the pontine taste area, while NK3 receptor labeling was sparse. The external medial subnucleus had substantial NK3 receptor and substance P labeling, but little NK1 receptor immunoreactivity. These findings suggest that distinct tachykinin ligand/neurokinin receptor combinations may be important in local processing of information within brainstem gustatory centers.

Introduction: Astrocytopathy is known to be an early feature of different neuroinflammatory diseases. However, the impact of astrocyte loss and repopulation on the development and progression of demyelinating lesions in complex etiologies, such as multiple sclerosis, is difficult to determine. To more easily analyse astrocyte- oligodendrocyte-interactions during lesion formation and progression in the human brain, diseases like Central pontine myelinolysis (CPM) can be used as a less complex model of demyelinating disorders. CPM is a rare neurological condition characterized by damage to the myelin sheath of pontine nerves after osmotic shifts in serum. Astrocytopathy is regarded to be the first event in the pathogenesis of CPM lesions. Methods: Histological investigation of autopsy tissue from human CPM patients was performed. Lesions were staged considering the myelination and the appearance of different astrocyte subtypes, which was used to judge behaviour of the astrocytic compartment. Further, dynamics of oligodendrocyte loss and repopulation were analysed and compared to the astrocytic repopulation. Results: Early-staged lesions were largely demyelinated and showed an overall reduction of astrocyte densities. The few astrocytes present showed a bipolar morphology and were APQ4-negative, indicating an immature state. Intermediate- stage lesions were still largely demyelinated, but had increased overall densities of astrocytes, which did not yet reflect densities observed in the perilesion. Astrocytes appeared mostly ramified and AQP4-positive, indicating maturity. Nevertheless, bipolar astrocytes were still observable, indicating that repopulation was not yet finalized. Late-stage CPM-lesions were at least partially remyelinated. Astrocytes were detectable in overall densities comparable to the perilesion and showed a ramified (or even reactive morphology), as well as regular expression of AQP4. Investigating the oligodendrocytes, intralesional densities were reduced in early- and intermediate-stage lesions when compared to the perilesion. Re-increase in oligodendrocyte densities was first observable in late-stage lesions, but did not reach perilesional levels. Conclusion: The study at hand indicates that the recovery of demyelinated osmolyte- induced pontine lesions follows a distinct time-course. Repopulation of the lesion with oligodendrocytes is not carried out until lesions are completely repopulated with functional resident astrocytes, as indicated by the ramified morphology and the expression of AQP4. Further studies will be needed to determine, whether the appearance of immature astrocytes, indicating an ongoing repopulation of lesions with astrocytes, correlates with an inefficient repair of demyelinated lesions.:List of Abbreviations.................................................................................................................6 1 Introduction................................................................................................................7 1.1 Osmotic Demyelinating Syndrome......................................................................... 7 1.2 Clinical manifestation............................................................................................. 9 1.3 Diagnosis and Management of CPM.....................................................................11 1.4 Aetiology of Central Pontine Myelinolysis.................................. ......................... 14 1.5 The brain, its adaptation to hyponatraemia and response to correction – pathophysiology of CPM............................................................................................16 1.6 Pathology of myelin............................................................................................. 19 1.6.1 Astrocytopathy and oligodendrocytopathy.................................................................................................20 1.7 Aims of the study................................................................................................. 23 2 Material und Methods............................................................................................. 24 2.1 Patient tissue........................................................................................................ 24 2.2 Histology and immunohistochemistry................................................................................................24 2.2.1 Basic concepts........................................................................................... ......24 2.2.2 Hematoxylin and Eosin (HE)............................................................................. 26 2.2.3 Luxol Fast Blue/ Periodic Acid Schiff stain........................................................27 2.2.4 Immunohistochemistry. Application and Protocol.............................................28 2.3 Implementation.................................................................................................... 31 2.4 Estimation of demyelination................................................................................. 32 2.5 Analysis of cell density and proliferation.............................................................. 32 2.6 Data plotting and statistical analysis.................................................................... 32 3 Results..................................................................................................................... 33 3.1 Patient cohort....................................................................................................... 33 3.2 Characteristics of demyelination.......................................................................... 35 3.3 CPM lesion and disease staging.......................................................................... 37 3.4 Astrocytes within human CPM lesions................................................................. 42 3.4.1 Astrocyte densities are decreased in early CPM lesions....................................42 3.4.2 Astrocytes in CPM– morphological distinctions.................................................45 3.5 Oligodendrocyte densities within human CPM lesions.........................................48 3.6 Macrophages and activated microglia.................................................................. 54 3.6.1 KiM1P – a marker for infiltrating macrophages and activated microglia............54 3.6.2 Proliferating Iba1+ cells are observed in all lesion stages..................................58 4 Discussion................................................................................................................ 61 4.1 Lesion Staging...................................................................................................... 61 4.2 Astrocytes in the pathogenesis of CPM............................................................... 65 4.3 Oligodendrocyte pathology in CPM..................................................................... 69 4.4 Mechanisms of regeneration in human CPM lesions............................................72 4.5 Summary, interpretation and limitations of our study............................................78 5 Conclusion and Outlook.......................................................................................... 80 6 Bibliography............................................................................................................. 82 7 List of Tables.............................................................................................................91 8 List of Figures.......................................................................................................... 92 9 Appendix.................................................................................................................. 94 9.1 Declaration of Authenticity.....................................................................................94 9.2 Acknowledgements...............................................................................................95

Ce travail de thèse fait partie d’un effort pour le développement des biomarqueurs, actuellement largement inexistants, pour une meilleure classification, pour une détermination plus précise du pronostic, et pour la prédiction de la réponse au traitement des tumeurs gliales malignes de l’enfant (épendymomes et gliomes de haut grade). Dans certains cas, ces biomarqueurs peuvent aussi devenir des cibles thérapeutiques.Dans ces études, nous avons pu montrer que la surexpression fréquente des marqueurs neuronaux distingue les épendymomes supratentoriels des formes infratentorielles. Parmi les épendymomes supratentoriels, une forte expression de neurofilament 70 (NF) est corrélée avec une meilleure survie sans récidive. La tenascine C (TNC) est surexprimée dans les épendymomes infratentoriels. Une étude de collaboration européenne multi-institutionnelle a été mise en place, permettant d’analyser une cohorte pédiatrique de 250 patients atteints d’un épendymome, et de démontrer la forte immunoexpression de TNC comme un marqueur robuste, associé a des survies globale et sans récidive plus péjoratives, particulièrement parmi les enfants âgés de moins de trois ans. Ceci a été validé dans une cohorte indépendante. Des immunomarquages pour NF et TNC pourraient être utilisés en clinique pour aider à déterminer le pronostic des épendymomes chez l’enfant.Une analyse des marqueurs pour la prédiction de la réponse à une thérapie ciblée anti-EGFR (erlotinib) a été réalisée par imunnohistochimie et FISH. La perte fréquente de PTEN dans les gliomes infiltrants du tronc cérébrale et la confirmation des caractéristiques des certains sous groupes (avec une forte expression de EGFR ou avec une différentiation oligodendrogliale) nous a permis de dessiner le protocole pour la prochaine étude de phase III pour cette maladie dans le cadre d’un futur essai thérapeutique de phase I/II. Nous rapportons par ailleurs des mutations du gène PI3KCA dans certains gliomes infiltrants du tronc cérébral, qui comme la perte d’expression de PTEN pourrait entrainer une activation de la voie mTOR qui devient donc une cible thérapeutique majeure théorique dans cette maladie. Des études ultérieures seront nécessaires pour définir le rôle de l’interaction entre la perte de PTEN, la surexpression de EGFR, la différentiation oligodendrogliale, les mutations de PI3KCA et d’autres altérations récemment décrites, (gains et amplifications de PDGFRA/MET mutations de TP53) et leur relation avec le devenir des patients sous traitement ciblé et les thérapies ciblées les plus intéressantes dans cette maladie uniformément fatale.Ce travail de thèse nous a permis d’explorer le rôle de la pathologie moléculaire dans la prise en charge des épendymomes et gliomas de haut grade chez l’enfant. / Biomarkers for the classification, clinical management and prognosis of pediatric brain tumors (ependymoma and high grade glioma, (HGG)) are lacking. To address this, biomarkers were developed and explored in view of classification, prognostication, target identification and prediction of the efficacy of treatment for patients with such tumors.We show that overexpression of neuronal markers distinguishes supratentorial from infratentorial ependymoma, and among the former higher immunoexpression of neurofilament 70 (NEFL) is correlated with better progression free survival (PFS). Tenascin-C (TNC) is significantly overexpressed in infratentorial ependymoma. A multi-institutional European ependymoma collaboration group was established and analyses were performed in a pediatric cohort of 250 patients, where immunohistochemistry (IHC) for TNC showed to be a robust marker of poor overall survival (OS) and PFS, particularly among children under 3 years, this being further validated in an independent cohort. Techniques and scoring performed in different laboratories were highly reproducible. IHC for NEFL and TNC could be used for prognostication of pediatric ependymoma.The analysis of putative predictive markers for the response to targeted therapies in pediatric HGG in the setting of a clinical trial with the anti-EGFR agent erlotinib was performed by IHC and fluorescent in situ hybridization. The frequent loss of PTEN in diffuse intrinsic pontine glioma (DIPG) and the confirmation of the biological singularity of the certain subgroups (expressing EGFR, displaying oligodendroglial differentiation) which seem to be associated with better response to erlotinib have helped our group to establish the design of the next Phase III protocol for this disease at our institution. We report mutations in PI3KCA constituting the first identification of oncogene mutations in some DIPG, which further highlight their biological heterogeneity. Further studies are needed to define the interaction between PTEN loss, EGFR overexpression, oligodendroglial differentiation, PI3KCA mutations and other recent findings such as PDGFRA/MET gains/amplification and TP53 mutations in these heterogeneous lesions and their relationship to the outcome of patients under new targeted therapies for this largely fatal disease.This thesis has allowed us to explore the molecular pathology in the context of biology and clinical setting of pediatric brain tumors.

Le gliome infiltrant du tronc cérébral est une tumeur gliale pédiatrique de très mauvais pronostic. La survie médiane des patients varie entre 9 et 12 mois. Son aspect typique en imagerie permet le diagnostic sans confirmation histologique mais a limité la compréhension de la physiopathologie de ces tumeurs pendant plusieurs années. L'identification des mutations des histones H3 K27M comme identité moléculaire des gliomes infiltrants du tronc cérébral définit cette tumeur comme une entité unique en 2012 et la sépare des autres tumeurs cérébrales. Il y a deux variantes des histones H3 à l’origine des tumeurs, la variante H.3.1 et la variante H3.3 qui forment deux sous-groupes présentant des phénotypes et des pronostics différents. L'objectif de cette thèse était d'utiliser les techniques d'imagerie multimodale pour mieux comprendre la physiopathologie des gliomes infiltrants du tronc cérébral, en observant l'évolution des effets des différents traitements à différents temps. Nous avons tout d'abord montré une augmentation des index de perfusion associée à une diminution de l'oedème lésionnel en post- radiothérapie. Puis nous avons décrit en imagerie le phénomène de pseudo-progression. Ce phénomène, qu’il ne faut pas prendre pour une vraie progression tumorale, est caractérisé par une augmentation du volume et du rehaussement tumoral associé à une augmentation importante des index de perfusion. Le ratio d’augmentation du flux sanguin lésionnel dans les suites de la radiothérapie permet identifier une pseudo-progression avec une sensibilité et une spécificité élevées. Ces 2 groupes de patients (sans et avec pseudo-progression) ont un comportement en imagerie très différents en fin de vie. Nous avons ensuite comparé les paramètres d'imagerie aux données histologiques pour mieux appréhender les deux types de mutation d'histones. Les tumeurs H3.1 présentent plus d'oedème, plus de nécrose et une perfusion moins élevée. Les tumeurs H3.3 ont une perfusion plus élevée et moins d'oedème. La valeur du volume sanguin lésionnel est positivement corrélée à la charge tumorale. L’imagerie multimodale est essentielle dans le suivi et la compréhension des mécanismes physiopathologiques des DIPG. Elle devra être incluse dans les protocoles de recherches (efficacité des nouvelles chimiothérapies ciblées et nouvelles techniques d’application, modèles murins, radiomics, corrélation avec les nouvelles quantifications histologiques) dans le but d’avancer puis un jour de guérir ces petits patients. / Diffuse Intrinsic Pontine Glioma is a pediatric tumor with very poor prognosis. Median survival is 9-12 months. Its typical MRI appearance allows for diagnosis without histological confirmation, but has limited understanding its pathophysiology for years. DIPGs molecular identity identified as histone H3K27M mutations in 2012 defined them as a separate entity. Two subgroups with different phenotypes and prognoses are associated with mutations in one of the two variants H3.1/H3.3 of histone H3. The aim of this thesis was to use multimodal imaging techniques to better understand the pathophysiology of DIPG, by observing the evolution of the effects of different treatments over time. 1st, we showed an increase in perfusion indices associated with decreased edema after radiotherapy. Secondly, we have described the phenomenon of pseudo-progression in MRI, an increase in tumor volume and enhancement associated with a significant increase in perfusion index. This shouldn’t be confused with true tumor progression. The ratio of increase in lesion’s blood flow after radiotherapy has high sensitivity and specificity to identify pseudo-progression. 3rd, we compared the imaging parameters with the histological data to better understand the differences between the histone mutations. The mutated tumors H3.1 have more edema, more necrosis and a lower perfusion. While mutated H3.3 tumors have a higher tumor burden. Tumor burden is positive correlated to lesion blood volume. Multimodal imaging is essential in monitoring and understanding the physiopathological mechanisms of DIPG. It should be included in the research protocols in order to advance and then one day to cure these small patients.

Adaptation to osmotic imbalance is crucial for cell survival and many organisms have developed complex mechanisms to counteract the changes induced by aniosmolarity. In mammals, the central nervous system is one of the most vulnerable organs after sudden changes in osmolarity. This is best exemplified by two common clinical disorders, brain edema resulting from acute hyponatremia and brain dehydration after hypernatremia. In clinical practice, hyponatremia is the most common electrolyte disorder and carries a significant mortality and morbidity. However, correction of hyponatremia should be undertaken with great caution as failure to adapt to rapid changes in chronic hyponatremia will cause rapid and fatal demyelination of the central nervous system. This syndrome is called osmotic demyelination syndrome (ODS) or central pontine myelinolysis. In this work, we investigated the pathophysiology and new diagnostic and treatments tools for osmotic demyelination syndrome. Using a rat model, we demonstrated the efficacy of the neuroprotective agent minocycline in osmotic demyelination syndrome. We also compared treatment with dexamethasone and re-lowering of serum sodium after rapid correction of hyponatremia and we showed that re-lowering of serum sodium is better than administration of dexamethasone. We explored the mechanisms underlying brain demyelination in ODS and demonstrated that the rupture of the blood brain barrier is not necessary for demyelination and that activated microglia does not play a key role in brain demyelination but can potentiate the lesions induced by rapid correction of hyponatremia. We also investigated the role of astrocytes during the development of osmotic demyelination and demonstrated that astrocytes are a major component of the physiopathology of osmotic demyelination. We showed that early and massive astrocyte apoptosis delineates the regions of future myelin damage and found that astrocyte death induces severe upregulation of myelinolytic cytokines and destruction of astrocyte oligodendrocyte junctions with subsequent disruption of panglial syncitium. <p>Finally, as there are currently no markers of demyelination, we investigated the astroglial protein S100B in ODS and found a significant release of S100B during development of ODS, which correlated with astrocyte damage. We also showed that the increase in S100B is prevented by protective treatment of hyponatremia with urea and demonstrated that serum levels of S100B could be used as a prognosis factor in ODS .<p>All together, our work has revealed a central role of astrocytes in the pathophysiology of ODS and clarified the importance of blood barrier dysfunction and microglial activation. This work also proposes new diagnostic and treatment tools for ODS. <p> / Doctorat en Sciences médicales / info:eu-repo/semantics/nonPublished

Purpose: The aim of this study is to investigate the spectrum of neurological triad improvement in patients with diffuse intrinsic pontine glioma (DIPG) treated by reirradiation (re-RT) at first progression. Methods: We carried out a re-analysis of the SIOP-E retrospective DIPG cohort by investigating the clinical benefits after re-RT with a focus on the neurological triad (cranial nerve deficits, ataxia, and long tract signs). Patients were categorized as “responding” or “non-responding” to re-RT. To assess the interdependence between patients’ characteristics and clinical benefits, we used a chi-square or Fisher’s exact test. Survival according to clinical response to re-RT was calculated by the Kaplan–Meier method. Results: As earlier reported, 77% (n = 24/31) of patients had any clinical benefit after re- RT. Among 25/31 well-documented patients, 44% (n = 11/25) had improvement in cranial nerve palsies, 40% (n = 10/25) had improvement in long-tract signs, and 44% (11/25) had improvement in cerebellar signs. Clinical benefits were observed in at least 1, 2, or 3 out of 3 symptoms of the DIPG triad, in 64%, 40%, and 24%, respectively. Patients irradiated with a dose ≥20 Gy versus <20 Gy may improve slightly better with regard to ataxia (67% versus 23%; p-value = 0.028). The survival from the start of re-RT to death was not different between responding and non-responding DIPG patients (p-value = 0.871). Conclusion: A median re-irradiation dose of 20 Gy provides a neurological benefit in twothirds of patients with an improvement of at least one symptom of the triad. DIPG patients receiving ≥20 Gy appear to improve slightly better with regard to ataxia; however, we need more data to determine whether dose escalation up to 30 Gy provides additional benefits.

Le gliome infiltrant du tronc cérébral est une tumeur rare, non opérable et inéluctablement fatale. En raison du manque de ressource biologique disponible, aucun progrès dans la compréhension de la biologie de ces tumeurs n’a été fait jusqu’à ces dernières années, laissant la radiothérapie pour seul traitement efficace, et seulement transitoirement. Enfin, grâce à la mise en place de collecte d’échantillons de gliomes infiltrant du tronc cérébral au diagnostic ou à l’autopsie, un nombre sans précédent d’analyses biologiques et génomiques a pu être mené et améliorer la connaissance de ces tumeurs. Si ces études ont montré que ces gliomes pédiatriques étaient bien différents de ceux de l’adulte, elles ont aussi fait apparaître la présence d’anomalies génétiques récurrentes spécifiques de ces tumeurs sous-tentorielles. Ainsi le Platelet-Derived Growth Factor Receptor Alpha (PDGFRα) est apparu comme cible prédominante dans ces tumeurs compte tenu des nombreuses anomalies génétiques constatées. La recherche d’un médicament efficace pouvant inhiber cette voie nous a conduit à évaluer l’effet du dasatinib qui est un inhibiteur multi-ciblé. Nous en rapportons ici l’efficacité in vitro sur de nouvelles lignées cellulaires de gliomes infiltrants du tronc cérébral établies à partir de biopsies stéréotaxiques réalisées au diagnostic. Sachant néanmoins que les thérapies ciblées restent peu efficaces en clinique quand elles sont utilisées seules, nous mettons en évidence l’intérêt de combiner le dasatinib avec un inhibiteur de MET, 2ème oncogène fréquemment amplifié dans ces tumeurs. D’autre part, une stratégie originale de criblage médicamenteux a été mise en œuvre. Celle-ci a permis de définir de manière fonctionnelle de nouveaux médicaments potentiellement efficaces dans les gliomes infiltrants du tronc cérébral, incluant les inhibiteurs d’Histone deacetylases (HDAC), les inhibiteurs des Cyclin-Dependent Kinases (CDK) ou encore les inhibiteurs du protéasome. Enfin par la technique de séquençage génome-entier, de nouvelles anomalies génétiques jamais rencontrées dans aucun autre cancer ont été détectées. Parmi celles-ci se trouvent des mutations d’histone H3K27M dont la fréquence élevée (80%) suggère leur rôle fondamental dans la genèse de ces tumeurs. Des mutations activatrices d’ACVR1/ALK2 ont été également mises en évidence. Celles-ci représentent désormais de nouvelles cibles à explorer.Ce travail de thèse rapporte la recherche de nouvelles cibles thérapeutiques d’une part, via une approche exploratoire par criblage médicamenteux et recherche d’anomalies génétiques par séquençage « génome-entier », et d’autre part, via une approche de validation préclinique sur le plan des thérapies ciblées de type inhibiteurs de tyrosine-kinases. / Diffuse Intrinsic Pontine Glioma (DIPG) is a rare, unresectable and universally fatal tumor. Due to the lack of available material, no improvements have been made in the knowledge of the biology of this tumor until recent years, leaving radiotherapy as the only efficient treatment, and only transiently. Recently, the effort engaged for collecting samples in this disease at the diagnosis or at the autopsy resulted in an unprecedented number of analyses consequently improving our knowledge in DIPG. Those studies bring evidences for their differences with adult gliomas, but also with other pediatric supratentorial glioma showing specific genomic alterations. Thus, Platelet-Derived Growth Factor Receptor Alpha (PDGFRα) appeared to be one of the major target given its frequent aberrations found in those tumors. Investigating an effective drug to inhibit this pathway led us to evaluate the effect of dasatinib, which is known as a multi-targeted inhibitor. We report here the in vitro efficacy of dasatinib on new cell lines of DIPG developed from stereotaxic biopsy at diagnosis. Because therapies are largely inefficient in the clinic when they are used as a monotherapy, we bring out our interest on combining dasatinib with an inhibitor of MET, which is the 2nd most common amplified oncogene in these tumors.Additionally, an innovative strategy of pharmacological screening has been successfully tested. New drugs, potentially efficient in DIPG, have been fonctionnaly-defined, including Histone deacetylase inhibitors (HDACi), Cyclin-Dependent Kinases inhibitors (CDKi) and proteasome inhibitors as well.Finally, by using whole genome sequencing (WGS), we have been able to discover new genetic abnormalities, never encountered before in other cancers. Among those, mutations of histone H3K27M with a high frequency of 80% were found, suggesting that they have a fundamental role in tumors genesis. Moreover, ACVR1/ALK2 activating mutations have been identified as well. And this gene now represents a new target to explore. This work reports the research of new therapeutic targets through an exploratory approach using drug screening and WGS on the one hand, and on the other hand through a preclinical validation approach in terms of targeted therapies with tyrosine-kinases inhibitors.