Les tumeurs cortico-surrénaliennes bénignes et malignes sont associées à une morbidité élevée résultant de l’hypersécrétion des hormones cortico-surrénaliennes, retrouvée chez près de 60% des patients. Au delà des perturbations endocrines, les carcinomes cortico-surrénaliens (CCS) sont des tumeurs de mauvais pronostic avec 16 à 38% de survie à 5 ans. Cette agressivité résulte à la fois de la présence de métastases chez de nombreux patients, au moment du diagnostic (30 à 40% des cas) et de l’absence d’approches thérapeutiques, au delà de la résection chirurgicale de la tumeur primaire. Au début de ma thèse, les mécanismes moléculaires impliqués dans le développement des tumeurs bénignes et malignes de la cortico-surrénale, étaient largement méconnus. L’activation anormale de la voie de signalisation Wnt/ßcaténine dans 48% des tumeurs bénignes et 37% des tumeurs malignes, suggérait que cette voie pouvait, comme dans d’autres tissus, participer au développement tumoral dans la cortico-surrénale. Afin de confirmer cette hypothèse, nous avons développé et caractérisé un modèle de souris transgéniques dans lesquelles la ßcaténine est constitutivement activée, spécifiquement dans le cortex surrénalien (souris ∆Cat). Grâce à ces souris, nous avons démontré pour la première fois que la ßcaténine agit comme un oncogène dans la cortico-surrénale, mais que son activation constitutive ne suffit pas à déclencher systématiquement le développement de tumeurs malignes. Chez plus de 90% des patients, la formation des CCS est associée à la surexpression du facteur de croissance IGF2. Grâce à des modèles de souris transgéniques qui surexpriment Igf2, nous avons pu montrer que cette surexpression n’a que peu d’effet sur l’initiation ou la progression tumorale, suggérant que d’autres altérations sont requises pour favoriser la transition maligne. Des résultats préliminaires encourageants suggèrent que la surexpression de l’histone méthyl-transférase EZH2 et les altérations épigénétiques résultantes, pourraient être la clé du développement des CCS. Parallèlement, nous avons montré que l’activation constitutive de la ßcaténine conduit au développement d’un hyperaldostéronisme primaire chez les souris ∆Cat, suggérant que l’activation de la voie Wnt/ßcaténine pourrait participer à la formation d’adénomes surrénaliens producteurs d’aldostérone (APA) chez les patients. Effectivement, nous avons mis en évidence que l’activation constitutive de la ßcaténine est l’altération moléculaire la plus fréquente dans les APA, avec une prévalence de 68%. Des analyses in vitro m’ont permis de montrer que la ßcaténine stimule la production d’aldostérone en contrôlant directement et indirectement l’expression de deux enzymes clés de la synthèse d’aldostérone – CYP21 et CYP11B2 – et du récepteur à l’angiotensine II (le sécrétagogue naturel de l’aldostérone), AT1R. Nous avons par ailleurs montré que la production excessive d’aldostérone chez les souris ∆Cat, pouvait être maîtrisée par un régime enrichi en quercétine, un inhibiteur naturel de l’activité transcriptionnelle de la ßcaténine. L’ensemble de ces résultats démontre l’importance de la voie Wnt/ßcaténine dans la tumorigenèse surrénalienne et dans l’hypersécrétion d’aldostérone ce qui fait d’elle une nouvelle cible thérapeutique potentielle. / Benign and malignant adrenocortical tumours are associated with a high morbidity caused by the hypersecretion of adrenocortical hormones found in approximately 60% of patients. Moreover, adrenocortical carcinomas (ACC) have poor prognosis with a 5 years survival rate of 16 to 38%. This aggressiveness results from both the presence of metastases at diagnosis in most patients (30 to 40% of cases) and the absence of therapeutic approaches apart from surgical resection of primary tumours. At the start of my thesis, the molecular mechanisms involved in the development of benign and malignant adrenocortical tumours were largely unknown. Abnormal activation of the Wnt/ßcatenin pathway found in 48% of benign tumours and 37% of malignant tumours suggests that as in other tissues, this pathway could participate in tumour development in the adrenal cortex. To confirm this hypothesis, we developed and characterized a transgenic mouse model with constitutive activation of ßcatenin, specifically in the adrenal cortex (∆Cat mice). With this model, we demonstrated for the first time that ßcatenin acted as an adrenocortical oncogene but that this activation was insufficient to systematically induce the development of adrenocortical carcinomas. In almost 90% of patients, CCS formation is associated with the overexpression of the growth factor IGF2. However, the development of a model of Igf2 overexpression in transgenic mice, allowed us to demonstrate that this overexpression could not initiate tumour formation and that it had a mild effect on tumour progression. This suggested that other alterations were necessary for malignant progression. Our encouraging preliminary results suggest that upregulation of the histone methyltransferase EZH2 and the resulting epigenetic defects could be the cause of ACC development. In parallel, we demonstrated that constitutive ßcatenin activation induced primary hyperaldosteronism development in ∆Cat mice suggesting that aberrant activation of the Wnt pathway could be involved in formation of aldosterone-producing adenomas (APA) in patients. Indeed, we showed that constitutive activation of ßcatenin was the most frequent molecular alteration in APA with a prevalence of 68%. In vitro analysis allowed us to demonstrate that ßcatenin stimulates aldosterone production by controlling directly and indirectly the expression of two key enzymes of aldosterone synthesis –CYP21 and CYP11B2- and of the angiotensin II receptor, AT1R. Furthermore, we showed that excessive aldosterone production in ∆Cat mice could be controlled by a diet enriched in quercetin, a natural inhibitor of the transcriptional activity of ßcatenin. Altogether these results demonstrate the essential role of the Wnt/ßcatenin pathway in adrenocortical tumorigenesis and aldosterone secretion. Consequently, this pathway could be a new potential therapeutic target for the treatment of most adrenal tumours.

Unilateral cortisol producing adenoma (CPA) is the most common cause of ACTH-independent Cushing’s syndrome and is surgically curable. On the other hand, adrenocortical carcinomas (ACCs) are rare and aggressive tumours. Although the overall survival of the patients with ACC is very poor, the outcome can be heterogeneous and vary significantly between the patients. This thesis comprises studies showing genetic and genomic events occurring in CPAs and ACCs, their functional impact and clinical correlations. The Wnt/β-catenin and cAMP/PKA signalling pathways are crucial in adrenal homeostasis and frequent mutations in members of these pathways (CTNNB1, GNAS, and PRKACA) are found in CPAs. Mutational analysis revealed that ~60% of the CPAs harboured mutations in either of these genes. Transcriptome signature exhibited increased expression of genes involved in steroidogenesis in PRKACA/GNAS mutated (Cluster1) tumours in comparison to CTNNB1 mutated /wildtype (Cluster2) tumours. In addition we have also observed that gain of chromosome arm 9q was the most frequent arm level copy number variation (CNV) occurring in CPAs and were exclusively present in Cluster2 tumours. We also discovered novel PRKACA mutations occurring in ACCs, causing activation of cAMP/signalling pathway.    Comprehensive analysis of Wnt/β-catenin signalling pathway in ACCs revealed novel interstitial deletions occurring in CTNNB1 leading to deletion of the N-terminus of β-catenin. This is a novel and yet another frequent event leading to activated Wnt/β-catenin signalling and downstream targets in ACCs. Both, mutations occurring in CTNNB1 and nuclear expression of its protein were associated with poor overall survival. Through multiregional sampling approach we discovered intra-tumour heterogeneity in ACC tumours. Although all the multiregions within a tumour showed presence of shared basal CNVs, they encompassed private CNVs, different ploidy levels and private mutations in known driver genes. We found intra-tumour heterogeneity in CTNNB1, PRKACA, TERT promoter and TP53 mutations as well as ZNRF3 and CDKN2A/2B homozygous deletions.

Regulated cell death (RCD) is indispensable for homeostasis and plays a crucial role in the pathophysiology of numerous diseases. Adrenocortical carcinomas (ACCs) represent a rare and highly malignant type of cancer. Currently, the most common therapeutic options include the complete surgical removal of the adrenal gland and/or the administration of mitotane, a derivative of the pesticide DDT. Yet patient survival remains poor and the mechanism of action of mitotane remains elusive. In this thesis it is demonstrated that the human ACC cell line NCI-H295R is sensitive to mitotane-induced cell death. In the first part, the involvement of three different RCD pathways, namely apoptosis, necroptosis and ferroptosis, in mitotane induced necrosis was investigated. To this end, different inhibitors were used, which were not able to block mitotane-induced cell death. When the medium was supplemented with insulin, transferrin, sodium selenite and linoleic acid (ITS+1) no cell death of the ACC cells was observed. This phenomenon was attributed to the presence of linoleic acid, since ITS supplementation lacking this component was not able to reverse mitotane-induced necrosis. Identification of new drug targets for alternative options of ACC treatment led to the investigation of key molecules involved in the pathways of necroptosis and ferroptosis. The receptor-interacting protein kinase 1 and 3 (RIPK1 and 3) and the mixed lineage kinase domain-like protein (MLKL) were considered as interesting targets given their crucial role in the execution of necroptosis. A western blot analysis of those molecules revealed the presence only of RIPK1, suggesting that the necroptosis machinery is not present in the NCI-H295R cells. Of interest, evaluation of the expression levels of glutathione peroxidase four (GPX4), one of the main inhibitory molecules of ferroptosis, showed a much higher expression in the ACC cells compared to the standard cell line used for studying ferroptosis, the human fibrosarcoma HT1080 cells. A hypothesis that the NCI-H295R cells are susceptible to ferroptosis induction was formed based on this finding. Compounds representative of all the four classes of ferroptosis inducers (FINs) were tested. Direct inhibition of GPX4 using the small compound RSL3, a type II FIN, led to high necrotic populations. Co-treatment with the ferroptosis inhibitor ferrostatin-1 (Fer-1) completely reversed RSL3-induced ferroptosis. Type IV FIN FINO2, that causes indirect loss of the enzymatic activity of GPX4, lead also to high necrotic populations, while Fer-1 prevented FINO2-induced ferroptosis. Data from public databases concerning gene methylation or mutation status of ACC tissues and normal human adrenal tissues was used to investigate potential key players of ferroptosis that might be either mutated or silenced in ACCs. Of note, glutathione peroxidases 3 and 5 (GPX3 and 5) were highly methylated, while the enzyme cystathionine gamma-lyase (CSE) involved in the transsulfuration pathway via the break down of cystathionine into cysteine and α-ketobutyrate and ammonia was found to be highly mutated. Collectively, these data point towards a high sensitivity of ACCs to ferroptosis induction. This could provide a new chapter for the therapeutic approaches of ACCs. Additionally, these findings provide a better understanding of the biology of this type of cancer that highly mutates or silences ferroptosis-related genes. The second part of this thesis focuses on the involvement of RCD in spontaneous cell death in isolated murine tubules. Existing literature points towards an involvement of necroptosis and ferroptosis pathways in the kidney in models of acute kidney injury (AKI). Acute tubular necrosis (ATN) represents a hallmark of AKI. While the work in the Linkermann lab has shown that isolated tubules perfused with type I FIN erastin undergo cell death in a “wave-of-death” manner, no deeper insights into the propagation of tubular necrotic injury exist. A protocol for isolation of murine kidney tubules was established, providing an ex-vivo model for investigation of tubular death. The absence of potentially confounding blood cells as well as immune cells was ensured by extensive washing steps as well as the use of collagenase. Visual observation and staining of isolated tubules with the nucleic acid stain SYTOX green revealed a spontaneous cell death in a “wave-of-death” manner. This wave was running in parallel with a calcium concentration change, indicating its involvement in the spontaneous necrosis. To investigate the potential involvement of mitochondria in this process, electron microscopy images were obtained from parts of the tubules with different levels of damage which revealed highly damaged and ballooned mitochondria. These data provided with a phenotypic characterisation of the spontaneous tubular necrosis. Aiming to approach this type of death genetically, necroptosis and pyroptosis deficient mice (MLKL/GSDMDDKO) were used. Comparison of the LDH release, used as a measure of necrosis, from isolated kidney tubules of the MLKL/GSDMDDKO mice and wild type (WT) mice showed no difference. This indicated that neither necroptosis nor pyroptosis are involved in the tubular necrosis. Therefore, the next step was to investigate the effects of Fer-1 at the levels of LDH of isolated tubules from WT mice. A significantly lower LDH release was observed in tubules treated with Fer-1 compared to the ones treated with vehicle. However, this reduction in the LDH release was not complete, suggesting that ferroptosis is only partially responsible for the spontaneous death of isolated tubules. The difference of male and female mice towards AKI sensitivity has been noted in the literature in that female mice are less susceptible compared to the male mice. Therefore, the next step was to investigate whether this protection of females can be observed at the level of isolated tubules. Indeed, the LDH release from tubules isolated from female mice was significantly less compared to the LDH release of tubules isolated from male mice. Based on the data obtained from isolated tubules from WT male mice treated with Fer-1, a similar experiment was performed with tubules isolated from WT female mice. No difference in the LDH release was observed between the Fer-1-treated tubules and the vehicle-treated ones, indicating that another cell death pathway might be involved. The most obvious difference between male and female organisms is the sex hormones. Whether testosterone or β-estradiol are responsible for the higher susceptibility or protection against cell death has been a debate over the last years. To test this hypothesis, three different cell lines were utilised. A pre-treatment of 16 h with either testosterone or β-estradiol was performed. Treatment with either type I FIN erastin or type II FIN RSL3 followed, and cells were analysed via flow cytometry. Data revealed protective effects of β-estradiol against ferroptosis induction. Next, the effects of β-estradiol in a simultaneous treatment with RSL3 were investigated. Interestingly the protective effects of the hormone were still observed. Among the metabolites of β-estradiol, 2-hydroxyestradiol (2-OHE2) has been reported to exert antioxidant effects. Therefore, 2-OHE2 was used in a simultaneous treatment with RSL3, and the obtained data showed that it was a much more potent inhibitor of necrotic cell death than β-estradiol even at lower concentrations. Collectively these data indicate that the lower susceptibility of female organisms towards cell death might be explained by the presence of β-estradiol and its more potent antioxidant metabolites. Such findings could change the way the two sexes are approached scientifically, while providing new insights on different therapeutic strategies between male and female organisms.