Les récepteurs couplés aux protéines G (RCPG) représentent la plus grande famille de protéines membranaires. Ils sont activés par une grande variété d'hormones, de neurotransmetteurs, et par des stimuli sensoriels. Ces récepteurs jouent un rôle central dans le contrôle de la grande majorité des fonctions physiologiques et constituent une cible thérapeutique majeure. Plusieurs études supportent l'existence de plusieurs états conformationnels de ces récepteurs stabilisés par les ligands. Le caractère dynamique des RCPG est essentiel dans leur fonctionnement. Une question majeure est de déterminer comment un ligand modifie la structure et la fonction de son récepteur. Pour cela, nous avons analysé les changements conformationnels d'un récepteur prototype de la famille des récepteurs à ligands peptidiques, le sous-type V2 de la vasopressine (V2R). Le V2R présente un large éventail de ligands ayant des efficacités différentes (agoniste partiels et complets, agonistes inverses et antagonistes) ainsi que des agonistes biaisés vis-à-vis des voies de signalisation dépendantes de Gs et des arrestines. Afin de déterminer les bases structurales de l'efficacité (amplitude de la réponse biologique) et de la sélectivité fonctionnelle (la capacité d'un RCPG à activer ou à inactiver préférentiellement une voie de signalisation parmi l'ensemble des voies de transduction auxquelles il est couplé), nous avons purifié et stabilisé le V2R par reconstitution en amphipols neutres (Napols). La fonctionnalité du récepteur a été vérifiée par mesure de son interaction directe avec la protéine Gs et les arrestines purifiés. Les profils d'efficacité des ligands vis-à-vis des deux voies de signalisation sont cohérents avec ceux décrits dans des cellules vivantes. Afin d'aborder directement les changements conformationnels dépendants des ligands, nous avons développé deux approches de fluorescence, la fluorescence intrinsèque des tryptophanes et le LRET (Lanthanide Resonance Energy Transfer). La liaison des ligands ayant des efficacités opposées pour la voie Gs ont induit des variations opposées de la fluorescence intrinsèque des tryptophanes, suggèrant l'existence d'états conformationnels distincts. En parallèle, l'analyse des changements des signaux de LRET entre deux domaines fonctionnels du récepteur marqués par deux fluorophores compatibles, le domaine transmembranaire 6 (TM6) côté cytoplasme et l'extrémité C-terminale distale, a permis de calculer une distance moyenne de 33 Å. En accord avec les variations de fluorescence intrinsèque des tryptophanes, les ligands ayant des efficacités opposées pour la voie Gs ont induit un mouvement opposé de ces deux domaines. Les agonistes complets entraînent un éloignement de la boucle i3 et de l'extrémité C-terminale (+2.4 Å) alors qu'un rapprochement des deux domaines est associé à la liaison de l'agoniste inverse (-0.9 Å). Nos résultats démontrent qu'un récepteur à ligands peptidiques répond à la liaison de ses ligands spécifiques par des changements conformationnels dynamiques. Chaque ligand est caractérisé par un ou plusieurs états conformationnels distincts. De plus, les changements conformationnels du V2R jouant un rôle dans le couplage à Gs sont différents de ceux impliqués dans le recrutement des arrestines. Ces données apportent des éléments essentiels de compréhension des mécanismes moléculaires et structuraux de l'activation des RCPG. A plus long terme, une étude plus extensive de la dynamique des RCPG devrait guider le développement de molécules thérapeutiques possédant des propriétés de sélectivité fonctionnelle. / G protein-coupled receptors (GPCR) are seven-transmembrane proteins that mediate most cellular responses to hormones and neurotransmitters, representing the largest group of therapeutic targets. Several studies support the existence of multiple ligand-specific conformational states of GPCR. The dynamic character of GPCR is likely to be essential for their functioning, and a better understanding of this molecular plasticity might facilitate structure-based drug discovery. A major question is to determine how ligands modify receptor structure and function. To this end, we have been studying the structural dynamics of the human vasopressin type 2 receptor (V2R), a prototypical peptide-activated class A GPCR. The V2R is coupled to Gs protein and to β-arrestins, and it has been well characterized pharmacologically using a large panel of ligands with different efficacies. Several display functional selectivity (Gs activation and concomitant β-arrestin inhibition). To demonstrate that ligand efficacy and functional selectivity are achieved through the stabilization of multiple conformational states, we have purified and reconstituted the V2R in amphipathic polymers (amphipol) and developed fluorescence-based approaches. The functionality of the V2R was monitored by direct activation of the purified Gs protein and interaction with purified β-arrestin 1. In these two assays, the effect of ligands correlated well with their known efficacy in cellular systems. Binding of V2R ligands with opposite efficacies toward Gs pathway led to opposite variations in the tryptophan intrinsic fluorescence of the receptor, suggesting the presence of different conformational states of the receptor. In parallel, we used Lanthanide-based resonance energy transfer (LRET) to directly analyze dynamics of the V2R and more particularly conformational changes between fluorophore-labeled extreme C-terminus and transmembrane domain 6. We calculated a basal mean distance of 33 Å between these domains. Interestingly, ligands with different efficacies towards Gs protein elicited opposite LRET changes as for tryptophan fluorescence spectroscopy. Indeed, the two labeled domains moved away upon full agonist binding (+2.4 Å), and closer in presence of inverse agonist (-0.9 Å). These data provide the first evidence of ligand-specific conformational changes in a peptide-activated receptor, and demonstrate that receptor conformational changes involved in Gs coupling are different from those responsible for arrestin recruitment. The results shed some light into the molecular and structural mechanisms of GPCR activation that may be relevant to the design of signaling pathway-selective drugs.

Thèse diffusée initialement dans le cadre d'un projet pilote des Presses de l'Université de Montréal/Centre d'édition numérique UdeM (1997-2008) avec l'autorisation de l'auteur.

The R7 regulators of G protein signaling (R7 RGS), namely RGS6, RGS7, RGS9 and RGS11, are expressed in the retina along with its binding partner Gβ5. The RGS9-1 isoform is expressed only in retinal photoreceptors and rate-limits the recovery of rod phototransduction by acting as a member of the transducin GAP complex (RGS9-1/Gβ5L/R9AP). The Gβ5L isoform is also only expressed in retinal photoreceptors and acts by stabilizing the GAP complex. The Gβ5S isoform differs from Gβ5L by the absence of exon 1 due to alternative splicing and is expressed in many other retinal cells. Gβ5L is barely detectable in RGS9-/- mice suggesting that Gβ5L has a protein degradation signal conferring instability in the absence of RGS9. To study the role of exon 1 of Gβ5L, we replaced Gβ5L with Gβ5S in rods by expressing transgenic Gβ5S under the control of the rhodopsin promoter within a Gβ5-/- mouse and determined that exon 1 of Gβ5L has two previously unidentified functions: (1) to increase the capacity of Gβ5L to bind to RGS9-1 and (2) to serve as a signal for rapid degradation of Gβ5L in RGS9-/- photoreceptors. Several groups have reported that RGS7 and RGS11 with Gβ5S reside in the dendritic tips of depolarizing bipolar cells (DBCs) and that they are involved in the mGluR6/Gαo/TRPM1 pathway, which mediates DBC light responses. The exact role of RGS7 in DBCs has not been unequivocally determined. We have contributed by making a true RGS7 null mouse line and found the RGS7-/- mice are viable and fertile, but have a small body size. Electroretinogram (ERG) b-wave implicit time in young RGS7-/- mice is prolonged at eye opening, but the phenotype disappears by 2 months of age. Expression levels of RGS6 and RGS11 are unchanged in RGS7-/- retina, but the Gβ5S level is significantly reduced. We further generated a RGS7 and RGS11 double knockout (711dKO) mouse line and found that Gβ5S expression in the retinal outer plexiform layer is eliminated, as well as the ERG b-wave. Ultrastructural defects similar to those of Gβ5-/- mice are present in 711dKO. Furthermore, in retinas of mice lacking RGS6, RGS7, and RGS11, Gβ5S becomes undetectable, while the photoreceptor-specific Gβ5L remains unaffected. Whereas RGS6 alone sustains a significant amount of Gβ5S expression in the retina, the DBC-related defects found in Gβ5-/- mice appear to be caused solely by a combined loss of RGS7 and RGS11. The notion that the role of Gβ5 in the retina, and likely in the entire nervous system, is mediated exclusively by R7 RGS proteins is firmly established in this work. The availability of all four R7 RGS single knockout mouse lines enables future studies to further elucidate the roles of R7 RGS proteins in vision.

The assessment of target protein molecular structure provides a distinct advantage in the rational drug design process. The increasing number of available G protein-coupled receptor crystal structures has enabled utilization of a varied number of computational approaches for understanding the ligand-receptor interactions, ligand selectivity and even receptor response upon ligand binding. The following dissertation examines the results from three different projects with varied objectives – i) structural modeling of human C-C chemokine receptor type 5 (CCR5) and assessment of the ligand binding pocket of the receptor, ii) assessment of the selectivity profile of naltrexone derivatives on the three opioid receptors (μ-opioid, κ-opioid, δ-opioid) with an aim towards designing selective μ-opioid receptor antagonists, and iii) structural modeling of the ‘active’ state conformation of the κ-opioid receptor in response to agonist binding and determination of a plausible molecular mechanism involved in activation ‘switch’ of the κ-opioid receptor. In absence of a crystal-based molecular structure of CCR5, a homology model of the receptor was built and the ligand binding pocket was validated. On the basis of evaluation of the ligand-receptor interactions on the validated binding pocket, structural and chemical modifications to anibamine, a natural plant product, were proposed to enhance its receptor binding. The selectivity of naltrexone (a universal antagonist) was assessed with respect to the three opioid receptors by employing ligand docking studies and the ‘message-address’ concept. Multiple address sites were identified on the opioid receptors and structural modifications were proposed for the naltrexone derivatives for their enhanced selectivity. In the third project, structural modeling of the active state conformation of the κ-opioid receptor covalently bound to a salvinorin A derivative (agonist) was attempted via molecular dynamics simulations. Although the obtained molecular model lacked the signature ‘agonist-like’ conformations, the result provides a template for such studies in the future.

The simplest phospholipid, lysophosphatidic acid (LPA), is a heat stable component of serum known for its proliferative and migratory activities in cancer cells. Strong evidence suggests that LPA production and expression of its receptors are dysregulated in multiple human malignancies. The mechanism behind LPA-mediated tumor cell growth and oncogenesis remains poorly understood. In this thesis project I used ovarian and other cancer cells as a model system to examine the hypothesis that LPA present in the tumor microenvironment is a pathophysiological determinant of hyperactive de novo lipogenesis and aerobic glycolysis, two hallmarks of cancer cells. We demonstrated that LPA induced proteolytic activation of sterol regulatory element binding proteins (SREBPs) in a cancer specific manner, leading to activation of the SREBP-FAS (fatty acid synthase) lipogenic pathway. Treatment of cancer cell lines with LPA also led to dephosphorylation and inhibition of AMP-activated kinase (AMPK), thereby activating acetyl CoA carboxylase (ACC). Moreover, these effects of LPA were mediated by LPA2, a receptor subtype overexpressed in multiple cancers, providing an explanation for the cancer specific regulation of FAS and ACC by LPA. Downstream of the LPA2 receptor, we identified the Gα12-Rho-Rock pathway to activate SREBPs and the Gαq-PLC (phospholipase C) pathway to inactivate AMPK. Consistent with LPA mediated activation of the key lipogenic enzymes FAS and ACC, LPA stimulated de novo lipid synthesis via LPA2, leading to accumulation of intracellular triacylglycerol and phospholipids. Pharmacological and molecular inhibition of LPA2, FAS or ACC attenuated LPA-dependent cell proliferation, indicating that upregulation of lipid synthesis is an integral component of the proliferative response to LPA. In further support of this, downregulation of LPA2 expression led to dramatic inhibition of anchorage-dependent and –independent growth of ovarian cancer cells. To support increased biomass generation, rapidly proliferating cancer cells enhance carbon influx by activating glycolysis. In the next part of the study, we investigated if LPA signaling was also involved in activating aerobic glycolysis in cancer cells. LPA indeed activated glycolysis in ovarian and other cancer cells but failed to elicit this response in non-transformed cells, suggesting a cancer specific role of LPA in regulation of glucose metabolism. While LPA had no effect on glucose uptake, we found that LPA altered expression of multiple genes involved in glucose metabolism. The most significant observation was that LPA treatment dramatically upregulated expression of HK-2, one of the rate-limiting glycolytic enzymes. We explored the underlying mechanism and found that LPA activates HK-2 transcription through LPA2-mediated activation of SREBP-1. Two sterol regulator elements (SREs) on the human HK-2 promoter were identified to be responsible for LPA activation of the promoter. DNA pulldown and chromatin immunoprecipitation assays confirmed that SREBP-1 bound to these SREs in LPA-treated cells. Although in ovarian cancer cells, LPA treatment also stabilized Hif-1α protein, an established activator of HK-2 and glycolysis, LPA-regulated HK-2 expression and glycolysis was largely independent of Hif-1α. These results established that LPA stimulates glycolysis via the LPA2-SREBP-HK-2 cascade in neoplastic cells. Taken together, this dissertation provides the first evidence for regulation of cancer cell metabolism by LPA. The results indicate that LPA signaling is causally linked to lipogenic and glycolytic phenotypes of cancer cells. Therefore, targeting the key LPA2 receptor could offer a novel and innovative approach to blocking tumor-specific metabolism.

Lysophosphatidic acid (LPA) is a potent bioactive phospholipid mediator that functions through multiple G protein couple receptors (GPCRs). LPA is elevated in ascites of ovarian cancer patients and is involved in growth, survival and metastasis of ovarian cancer cells. Gene promoter analyses revealed that some LPA-target genes share similar sets of binding sites for prominent transcription factors posing the possibility of a general mechanism for activation of their expression by LPA. Detailed investigation of the mechanisms of regulation of cyclooxygenase 2 (Cox-2), a paradigm of LPA-regulated genes, showed that LPA robustly upregulated the expression of Cox-2 in ovarian cancer cells through multiple receptors. LPA induced rapid increase in Cox-2 mRNA and significantly enhanced the stability of Cox-2 transcript with the support of mRNA binding protein HuR. The effects of LPA on Cox-2 transcriptional activation include essential involvement of transcription factor, C/EBP-b. Further studies on mechanisms of activation of C/EBP-b demonstrated that LPA increased phosphorylation, binding and transcriptional activities of C/EBP-b. In addition, activation of C/EBP-b and LPA-target genes required contribution from EGFR. This novel crosstalk between LPA GPCRs and EGFR in mediating transcription factors activation was further explored by investigating the mechanisms of activation of AP-1 and NF-kB by LPA. Activation of AP-1 family of proteins by LPA relied heavily on basal inputs from EGFR as inhibition of EGFR kinase activity with AG1478 caused significant loss of LPA-induced AP-1 expression, binding and transcription activities. Although HGF and other agonists of RTK only weakly stimulate LPA-target genes and transcription factors in ovarian cancer cells, costimulation with HGF in the presence of AG1478 restored LPA signals to both C/EBP-b and AP-1. This suggests an obligatory role for a RTK in LPA-induced transcriptional activation, not necessarily inputs from EGFR. Interestingly, inhibition of EGFR with AG1478 did not interfere with LPA-induced NF-kB activation. Pharmacological inhibition and molecular targeting revealed that only a subset of G proteins participate in the crosstalk between LPA receptors and EGFR. Collectively, these results demonstrate the presence of at least two signals downstream of LPA receptors: one dependent on basal RTK activity and another mediated directly by LPA GPCRs.

Membrane domains are an important structure in plasamatic membrane. They concentrate various signaling molecules. Their main structural component is cholesterol and by its removal the membrane domains are disrupted. The aim of our work was to examine the effect of cholesterol depeletion on signaling initiated thyreothropin releasing hormone (TRH). Although its signaling cascade is located within membrane domains the receptor itself is not. We showed that cholesterol depletion by -cyclodextrin caused release of Gq/11 proteins and caveolin 2 from membrane domains. We also discovered that cholesterol depletion decreases potency of TRH to activate G proteins as well as induction of release of intracellular Ca2+ In the last part we investigated the effect of disruption of the cell membrane integrity by cholesterol depletion on thyrotropin-releasing hormone receptor (TRH-R) surface mobility and internalization in HEK293 cells stably expressing TRH-R-eGFP fusion protein. CLSM studies indicated that the internalization of receptor molecules initiated by TRH stimulation was significantly attenuated. The detailed analysis of recovery of TRH-R-eGFP fluorescence in bleached spots of different sizes indicated that cholesterol depletion results in an increase of overall receptor mobility. We suggest that migration of...

CCR5 et CXCR4 sont les corécepteurs d'entrée du VIH utilisés par le virus in vivo en plus du récepteur principal CD4 pour infecter les cellules. Au début et tout le long de l'infection, on retrouve chez les patients infectés par le VIH, des virions R5 utilisant le corécepteur CCR5 pour infecter les cellules. Dans les stades tardifs de l'infection et chez environ la moitié des personnes infectées par le VIH, on observe en plus de ces souches R5, l'apparition de souches X4, utilisant le corécepteur CXCR4 pour infecter les cellules. Cette apparition de souches X4 est un facteur d'aggravation de la maladie. Les causes de cette commutation de R5 vers X4 sont mal définies. Le but de mon travail de thèse a été de trouver de nouvelles stratégies thérapeutiques visant l'un ou l'autre de ces corécepteurs. La première partie de mon travail compare les deux isoformes de CXCR4 en tant que corécepteurs du VIH. Ces deux isoformes, CXCR4-A et CXCR4-B, diffèrent de 9 acides aminés en NH2 terminal suite à un épissage alternatif. Nous avons montré que l'isoforme CXCR4-B est la plus performante en tant que corécepteur du VIH mais que ces deux variants sont équivalents pour la migration vers leur ligand commun SDF-1. Ainsi, nous proposons qu'en ciblant exclusivement l'isoforme B qui est la plus favorable à l'infection, via par exemple des siRNA, il serait possible de limiter les infections par des souches X4 tout en gardant une partie des fonctions essentielles de ce récepteur dans l'organisme, assurées par l'isoforme A. Nos résultats suggèrent également que l'infection par des souches R5 augmente le ratio en ARNm CXCR4-B / CXCR4-A dans des PBMC, et que ce ratio est en partie responsable de la commutation de R5 vers X4 associée à une aggravation de la maladie. Cibler cette isoforme CXCR4-B pourrait donc se révéler bénéfique. La deuxième partie de cette thèse étudie la modulation de la fonction de corécepteur du VIH de CCR5 par S1P1, un autre membre de la famille des récepteurs couplés aux protéines G qui permet la remise en circulation des lymphocytes après leur séjour dans les organes lymphoïdes secondaires par chimiotactisme vers son ligand S1P, abondant dans le sang. Nous montrons que S1P1 interagit physiquement avec CCR5 et gêne l'entrée des virus R5 dans la cellule-hôte. A l'inverse, S1P1 active les étapes post-entrée du cycle viral, notamment l'expression génique virale. La résultante de ces effets opposés est une augmentation de la production virale par des cellules infectées in vitro. Ce projet a également montré que l'utilisation de FTY720, un antagoniste fonctionnel de S1P1, diminue l'infection de cellules dendritiques par des virus HIV-R5 in vitro, ainsi que la virémie dans un modèle de souris SCID infectées après reconstitution immunologique. La mise en évidence des interactions entre CCR5 et S1P1 ouvre donc des perspectives thérapeutiques. / CCR5 and CXCR4 are the two HIV entry coreceptors used by the virus in addition to the main receptor CD4 in vivo to infect cells. R5 virions, that use CCR5 as a coreceptor to infect cells, are detected in most HIV patients. At late stages of infection and in about half of HIV infected persons, there is an emergence of X4 virions that use CXCR4 as a coreceptor, in addition to R5 virions. This emergence is associated with an increase in disease progression. The reasons for this R5 to X4 switch are poorly understood. The goal of my PhD work was to find new therapeutic strategies that target these coreceptors.The first part of this work compares the two CXCR4 isoforms as HIV coreceptors. Those two isoforms, CXCR4-A and CXCR4-B, differ by 9 amino acids at their NH2 terminal extremity as a consequence of an alternative splicing. We have shown that CXCR4-B isoform is more efficient as an HIV coreceptor but that those two variants are equivalent in terms of chemotaxis toward their common ligand SDF-1. Thus, we propose that by targeting specifically the B isoform that supports infection, via siRNA by example, it is possible to limit X4 development while keeping essential functions of this receptor. Our results also suggest that R5 infection increases CXCR4-B / CXCR4-A mRNA ratio in PBMC and that this ratio is in part responsible for R5 to X4 switch. Thus, targeting CXCR4-B isoform could be beneficial.The second part of this PhD thesis studies the effect on CCR5 coreceptor function of S1P1, another G protein-coupled receptor that enables lymphocytes egress from lymph nodes by chemotaxis toward its ligand S1P that is abundant in blood. We have shown that S1P1 physically interacts with CCR5 and blocks R5 virus entry. On the other hand, S1P1 activates post-entry steps of the viral cycle, in particular gene expression. The resulting effect is an increase in viral production by infected cells in vitro. We also showed that the use of FTY720, a S1P1 functional antagonist, decreases dendritic cell infection by R5 viruses in vitro, and in vivo infection in a SCID mouse model. The emphasis of CCR5 and S1P1 interactions opens new therapeutic strategies.

La protéine GPR50, qui fait partie de la famille des récepteurs de la mélatonine, est classée, avec une centaine d’autres protéines à sept domaines transmembranaires (7TM), dans la catégorie des récepteurs couplés aux protéines G hétérotrimériques (RCPG) orphelins, c’est-à-dire pour lesquels aucun ligand n’a pu être identifié. De plus en plus d’études montrent que les 7TM peuvent avoir des fonctions indépendantes d’un ligand. C’est le cas de GPR50 qui inhibe les fonctions du récepteur de la mélatonine MT1 en interagissant directement avec lui. Nous avons cherché à identifier d’autres partenaires associés à GPR50 en appliquant la technique de purification par affinité en tandem et avons mis en évidence son interaction avec un récepteur du facteur de croissance Transforming Growth Factor ß (TGFβ), le récepteur de type I (TβRI).Nous décrivons ici la formation d’un complexe entre GPR50 et le récepteur TβRI au niveau de la membrane plasmique, avec pour conséquence l’induction d’une activité constitutive du récepteur et des voies de signalisation en aval en l’absence de TGFβ, mais également en l’absence du récepteur TßRII qui est habituellement indispensable pour l’activation de TβRI par phosphorylation. Cette activité constitutive se traduit par la phosphorylation des protéines Smad2 et Smad3, leur intégration dans un complexe avec Smad4, la translocation du complexe dans le noyau et finalement l’activation de la transcription de leurs gènes-cibles. Nous avons décrypté les mécanismes moléculaires de cette activation constitutive en montrant que GPR50 entre en compétition, pour l’interaction avec TβRI, avec le régulateur négatif FKBP12, une protéine inhibitrice de l’activité basale du récepteur en l’absence de ligand. Nous avons identifié dans la queue intracytoplasmique de GPR50 un motif répétitif similaire à la séquence de FKBP12 impliquée dans son interaction avec TβRI , motif qui constitue la base moléculaire de cette compétition.Nous avons étudié les conséquences fonctionnelles de cette activation en surexprimant GPR50 de manière stable dans la lignée cellulaire MDA-MB-231, dérivée d’un cancer de sein. Nous avons observé dans ces cellules des effets pro-migratoires et anti-prolifératifs similaires à ceux causés par l’administration de TGFβ.En conclusion, ce travail décrit un nouveau mode d’activation du récepteur TβRI en l’absence de ligand, mais identifie également une nouvelle fonction indépendante d’un ligand pour le RCPG orphelin GPR50. En perspective de ce travail, nous allons essayer d’identifier des conditions biologiques où cette interaction pourrait prendre place afin de confirmer ces résultats dans un contexte plus physiologique. / During the last years, it became more and more accepted, that orphan G Protein coupled receptors (GPCRs) with a transmembrane spanning heptahelical core (7TM) can have ligand-independent functions. One of those 100 orphan GPCRs is GPR50, a 7TM protein with a long cytosolic domain. Recently, studies revealed ligand-independent functions for GPR50, where it has the capacity to modulate the activity of other proteins upon complex formation. By applying a tandem affinity purification approach we sought to identify further putative interacting partners of GPR50. One of the identified binding partners is the transforming growth factor β (TGFβ) receptor type I (TβRI).The TGFβ-dependent signal transduction pathway of serine/threonine kinases is a pathway with direct signal flow from ligand over the receptor to its substrates, the Smads which translocate into nucleus where they bind DNA and regulate gene expression. An important question concerns the generation of specificity and fine-tuning of TGFβ-dependent signaling. Throughout the years, an important number of proteins which regulate the activity of the TGFβ signal transduction pathway in a positive or negative manner have been identified. Most of them act in a cell-context-dependent manner, allowing the regulation of TGFβ signaling adapted to the particular circumstances.We report here the complex formation of GPR50 and TβRI on the plasma membrane. The consequence of this interaction is the GPR50-mediated induction of a constitutive activation of the TβRI and its downstream signaling in a TGFβ ligand-independent manner. This has been monitored by Smad2/3 phosphorylation, Smad2/3-Smad4 complex formation and their subsequent translocation into the nucleus, where they activate Smad-dependent gene expression. In order to decipher the molecular mechanism that allows this activation, we showed that GPR50 competes with the negative regulator, that prevents leaky TGFβ signaling, the gatekeeping molecule FKBP12, for binding to the TβRI. We identified a motif in FKBP12 involved in the interaction with TβRI with similarities to a motif in GPR50, providing a molecular basis for the replacement of FKBP12 by GPR50 in the TβRI complex. We showed that GPR50 is capable of activating the TβRI even in the absence of the TβRII, which normally is required for activating the TβRI by phosphorylation. This reveals a previously unknown mode of activation of the TβRI in absence of the TGFβ ligand and TβRII. In order to identify the functional consequences of this crosstalk, we studied migration and growth of MDA-MB-231 breast cancer cells stably overexpressing GPR50. In these cells, TGFβ-like pro-migratory and anti-proliferative effects have been observed.Future research will help to identify tissues and biological circumstances, where this crosstalk could take place for putting this novel mode of regulation of TGFβ signaling pathway into a context-dependent-manner. Additionally our work established another ligand-independent task for the orphan 7TM protein GPR50, consolidating its function as binding partner and activity modulator.

Le développement des anticorps monoclonaux thérapeutiques est en plein essor notamment à cause de leur bénéfice important pour le traitement des cancers. Cependant, à l’heure actuelle, aucun anticorps monoclonal sur le marché ou en phase III ne cible de RCPGs, en dépit de l’implication grandissante de ces récepteurs dans la carcinogenèse. Parmi les RCPGs les plus pertinents pour l’oncologie, souvent cités dans la littérature et dont certains inhibiteurs chimiques sont en phase clinique avancée, on trouve les deux sous-types de récepteurs des endothélines ETAR et ETBR. Dans ce contexte, mon projet de thèse a consisté à produire des anticorps monoclonaux capables de lier spécifiquement les récepteurs des endothélines, puis à les caractériser dans le but d’évaluer leur potentiel antitumoral. Grâce à une stratégie d’immunisation génique, un ensemble de 27 anticorps monoclonaux, tous spécifiques de la forme native d’ETBR, a été obtenu. Un de ces anticorps, nommé rendomab-B1, a fait l’objet d’une caractérisation précise et s’est révélé être un puissant inhibiteur allostérique d’ETBR. De plus, cette propriété antagoniste a permis de bloquer l’action autocrine antiapoptotique de l’ET-1 sur des cellules endothéliales vasculaires, suggérant ainsi que le rendomab-B1 pourrait être utilisé comme agent thérapeutique afin d’inhiber les effets tumorigènes liés à la suractivation de l’axe ET1/ETBR au niveau de l’endothélium vasculaire tumoral. Par ailleurs, le rendomab-B1 a également été testé sur des lignées de mélanomes humains ; l’absence de fixation de l’anticorps malgré la présence de récepteurs ETB fonctionnels à la surface de ces cellules suggère l’existence d’une forme moléculaire atypique du récepteur, potentiellement spécifique aux mélanomes. A la lumière de ces résultats, le rendomab-B1 apparaît comme un outil prometteur, à la fois pour l’étude structurale et fonctionnelle d’ETBR, mais aussi pour une éventuelle thérapie anticancéreuse. Enfin, les 26 autres anticorps monoclonaux anti-ETBR, actuellement en cours de caractérisation, constituent également des molécules potentiellement intéressantes pour un usage fondamental ou thérapeutique impliquant ETBR. Pour conclure, ces travaux ont démontré l’intérêt de la méthode d’immunisation génique pour la production d’anticorps monoclonaux anti-RCPGs à visée thérapeutique. / For a decade, monoclonal antibodies have become increasingly important for the biotherapeutic management of cancer. However, none of the monoclonal antibodies currently on the market or in late stage clinical trial do target a G-protein coupled receptor in spite of the emerging role of these receptors in tumor progression. Among the therapeutically relevant GPCRs for oncology, the endothelin receptors (ETAR and ETBR) are particularly attractive considering their overexpression in a wide range of tumors and their involvement in various stages of tumorigenesis. In this context, my PhD project consisted in producing and characterizing monoclonal antibodies directed against endothelin receptors with a view to use them as anti-tumor agents. Using an original DNA immunization strategy, we produced a panel of 27 monoclonal antibodies which selectively recognized ETBR expressed at the surface of transfected cells. One of these antibodies, named rendomab-B1, was extensively characterized and proved to be a potent allosteric antagonist of ETBR. Moreover, rendomab-B1 was able to disrupt the autocrine ET1-mediated survival loop on vascular endothelial cells, suggesting that this antibody could be used to prevent the pro-tumorigenic effect due to ET-1 and ETBR upregulation in the tumor-surrounding endothelium. Furthermore, rendomab-B1 binding onto ETBR was also assessed on melanoma cell lines and revealed that a tumor-specific form of ETBR may exist, as illustrated by the poor fixation of rendomab-B1 on these cells in spite of the presence of functional ETB receptors. Together, these results present rendomab-B1 as promising agent, not only for the structural and functional study of ETBR, but also for its therapeutic modulation in the case of cancer for instance. Finally, the other 26 monoclonal antibodies, whose characterization is still ongoing, also constitute potential tools for fundamental or therapeutic applications involving ETBR. To conclude, this work has highlighted the relevance of the DNA immunization approach to generate monoclonal antibodies against the native form of GPCRs.