Cells defend their cytosol against pathogen invasion using cell-autonomous immunity. When pathogens enter the cytosol they can damage host endomembranes, causing the mislocalisation of host molecules not normally found in the cytosol that are sensed as Danger Associated Molecular Patterns (DAMPs). Glycans exposed on damaged endomembranes are detected by danger receptors such as Galectin8. Galectin8 is recognised by the autophagy cargo receptor NDP52, specifically targeting the bacteria to autophagy. I hypothesised that other proteins would also be recruited to damaged endomembranes, which may initiate downstream mechanisms involved in cell-autonomous immunity or endomembrane repair. Identifying novel damage recruited proteins (DRPs) is difficult due to the short-lived and dynamic nature of damaged endomembranes. Therefore, I developed an unbiased approach for the identification of novel DRPs by proximity-dependent biotinylation using the ascorbate peroxidise enzyme APEX. This approach preferentially labels proteins located at damaged endomembranes for subsequent identification by TMT mass spectrometry. Four enriched proteins CCDC50, FBXO21, STAMBP and PDCD6 were identified as novel damage recruited proteins, recognising damaged SCVs. An alternative form of cell-autonomous immunity is the induction of cell death, for example by pyroptosis. Cell death destroys the bacteria's replicative niche and exposes them to the extracellular space where they may be phagocytosed. I hypothesised that host cells might tag cytoplasmic bacteria with intracellular opsonins to assist in their phagocytosis following their release from host cells. However, my work revealed that intracellular Salmonella Typhimurium acquire phagocytosis protection, thus becoming internalised by phagocytes less efficiently than control bacteria. Phagocytosis protection was acquired rapidly after S.Typhimurium infection and was not observed with dead bacteria. Phagocytosis protection is only partially reversed by opsonisation in human serum. My results indicate that intracellular S.Typhimurium-induces an evasion mechanism to prevent its subsequent recognition by extracellular phagocytes.

No description available.

Shigella flexneri is a highly adapted pathogen that invades the host cytosol and causes bacillary dysentery. Shigella has evolved powerful countermeasures to disarm host defense mechanisms; amongst them a family of twelve bacterial E3 ubiquitin ligases (IpaH) that are structurally unrelated to eukaryotic enzymes. IpaH ligases are injected into the host cytosol via the bacterial type III secretion system (T3SS) to manipulate the host cell and counteract anti-bacterial defense pathways. My work demonstrated that IFN-induced guanylate-binding proteins (GBPs) are novel targets for IpaH9.8. GBPs inhibit actin-dependent motility and cell-to-cell spread of bacteria unless they are ubiquitylated by IpaH9.8 and consequently degraded by the proteasome. IpaH9.8 targets GBP1, GBP2, and GBP4, thereby causing a transient poly-ubiquitin coat comprising K48 and K27-linked chains around S. flexneri, which leads to the proteasome-dependent destruction of existing GBP coats and the re-establishment of bacterial motility and cell-to-cell spread. So far, ubiquitylation of bacteria has mostly been associated with anti-bacterial autophagy or immune signaling. However, the ubiquitin coat assembled around intracellular Shigella by IpaH effectors, in particular IpaH9.8, serves a pro-bacterial function, the first observed so far. In addition, I characterized IpaH1.4 and IpaH2.5 for their ability to prevent NF-κB activation by targeting LUBAC. I found that IpaH1.4 specifically binds the LUBAC component HOIP and mediates its proteasomal degradation, thus abolishing linear ubiquitylation of bacteria and consecutive NF-κB activation via NEMO and autophagy induction via optineurin. Lastly, I identified novel potential ubiquitylation targets for IpaH effectors in human cells using a mass spectrometry-based approach. The resulting IpaH interactome presents the groundwork for further investigations and will help to identify potentially unknown cellular defense mechanisms that are antagonized by Shigella flexneri.

Lignin is the second most abundant biopolymer on earth and is found in the wood (xylem) of vascular land plants. To transport the hydro-mineral sap, xylem forms specialized conduit cells, called tracheary elements (TEs), which are hollow dead cylinders reinforced with lateral secondary cell walls (SCW). These SCWs incorporate lignin to gain mechanical strength, water impermeability and resistance against pathogens. The aim of this thesis is to understand the spatio-temporal deposition of lignin during TE differentiation and the relationship with its neighbouring cells. In vitro TE differentiating cell cultures of Zinnia elegans and Arabidopsis thaliana are ideal tools to study this process: cells differentiate simultaneously into 30-50% TEs while the rest remain parenchymatic (non-TEs). Live-cell imaging of such TEs indicated that lignification occurs after programmed cell death (PCD), in a non-cell autonomous manner, in which the non-TEs provide the lignin monomers. This thesis confirms that lignification occurs and continues long after TE PCD in both in vitro TE cultures and whole plants using biochemical, pharmacological and cytological methods. The cooperative supply of lignin monomers by the non-TEs was demonstrated by using Zinnia and Arabidopsis in vitro TE cultures. Inhibitor experiments revealed further that the non-TEs supply reactive oxygen species (ROS) to TEs and that ROS are required for TE post-mortem lignification. Characterization of the non-TEs showed an enlarged nucleus with increased DNA content, thus indicating that non-TEs are in fact endoreplicated xylem parenchyma cells (XP). The cooperative lignification was confirmed in whole plants by using knock-out mutants in a lignin monomer synthesis gene, which exhibit reduced TE lignification. The XP specific complementation of these mutants led to nearly completely rescuing the TE lignin reduction. Using microscopic techniques, the spatial distribution of lignin was analyzed in TEs from in vitro cultures and whole plants, revealing that lignification is restricted to TE SCWs in both protoxylem and metaxylem. These specific deposition domains were established by phenoloxidases, i.e. laccases localized to SCWs and peroxidases, present in SCWs and the apoplastic space. Laccases were cell-autonomously produced by developing TEs, indicating that the deposition domains are defined before PCD. Altogether, these results highlight that the hydro-mineral sap transport through TEs is enabled by the spatially and temporally controlled lignification of the SCW. Lignification occurs post-mortem by the supply of monomers and ROS from neighbouring XP cells and is restricted to specific deposition domains, defined by the pre-mortem sequestration of phenoloxidases.

Rett Syndrome (RTT) is a neurodevelopmental disorder primarily caused by mutations in the X-linked gene methyl-CpG-binding protein 2 (MECP2). The mosaic brain environment in heterozygous (MECP2+/-) females consists of both MeCP2-wildtype (MeCP2+) and Mecp2-mutant (MeCP2-) neurons. To separate possible cell autonomous and cell non-autonomous effects three-dimensional morphological analysis was performed on individually genotyped layer V pyramidal neurons in the primary motor cortex of heterozygous (Mecp2+/-) and wild-type (Mecp2+/+) mature female mice (>8 months old) from the Mecp2tm1.1Jae line. Mecp2+/+ neurons and Mecp2+ were found to be indistinguishable while Mecp2- neurons have significantly reduced basal dendritic length (p<0.05), predominantly in the region 70-130 μm from the cell body, culminating in a total reduction of 15%. Mecp2- neurons have three (17%) fewer total branch points, lost specifically at the second and third branch orders. Thus the reduced total dendritic length in Mecp2- neurons is a result of fewer higher-order branches. Soma and nuclear areas of 30 Mecp2+/- female mice (5-21 months) with X chromosome inactivation (XCI) ratios ranging from 12% to 56% were analyzed. On average Mecp2- somata and nuclei were 15% and 13% smaller than Mecp2+ neurons respectively. The variation observed in the soma and nuclear sizes of Mecp2- neurons was not due to age, but was found to be correlated with the XCI ratio. Animals with a balanced XCI ratio (approximately 50% Mecp2-) were found to have Mecp2- neurons with a less severe cellular phenotype (11-17% smaller than Mecp2+). Animals with a highly skewed XCI ratio favouring expression of the wild-type allele (less than 30% Mecp2-) were found to have a more severe Mecp2- cellular phenotype (17-22% smaller than Mecp2+). These data support indicate that mutations in Mecp2 exert both cell autonomous and cell non- autonomous effects on neuronal morphology. / Graduate

Le récepteur cannabinoique de type 1 (CB1R) est l’un des récepteurs couplés aux protéines G les plus abondants du cerveau mammifère. CB1R a longtemps été décrit comme un récepteur présynaptique régulant de manière rétrograde la transmission synaptique. Cependant, depuis les vingt dernières années, de nouveaux rôles ont été découverts et il est maintenant clairement admis que l’action des endocannabinoides (eCBs) ne se limite pas à la régulationde la neurotransmission au niveau de synapses adultes déjà établies. En effet, les eCBs et le CB1R sont des acteurs majeurs de l’ensemble des phases du développement cérébral. Cependant, les mécanismes moléculaires impliqués n’ont toujours pas été identifiés. Les mécanismes cellulaires auxquels nous nous intéressons ne dépendant pas de l’environnement cellulaire, nous proposons donc de combiner la puissance génétique du modèle drosophile à l’accessibilité et la haute résolution offerte par la culture primaire de neurones. De plus, le récepteur CB1 ne possédant pas d’orthologue parmi les invertébrés, ce système offre la possibilité d’étudier la biologie du récepteur en s’affranchissant de la machinerie endocannabinoide. Cependant, actuellement, aucun protocole de culture primaire de neurones de drosophile ne permet d’obtenir des cellules hautement différenciées et polarisées à basse densité. Ainsi, nous avons tout d’abord développé, optimisé et validé un nouveau protocole permettant de d’obtenir des neurones fonctionnels, hautement différenciés et polarisés en culture de basse densité. Dans un second temps, nous avons démontré que l’activation durécepteur CB1, exprimé ectopiquement dans les neurones de drosophile, entrainait son internalisation, de manière identique à ce qui avait déjà été observé chez les mammifères. Puis, nous avons étudié l’effet de l’expression et de l’activation ectopique de CB1R sur le développement neuronal chez la drosophile. Ainsi, nous avons démontré que l’activation du récepteur module directement la dendritogénèse. Afin de compléter la caractérisation de notremodèle, nous avons démontré que l’activation transitoire du récepteur dans les corps pédonculés (le centre de la mémoire olfactive chez la drosophile) altérait spécifiquement la formation d’une forme consolidée de mémoire après un conditionnement aversif. En conclusion, la validation du modèle drosophile dans l’étude de la biologie cellulaire durécepteur CB1 ouvre de nouvelles perspectives quant à la détermination des mécanismes moléculaires régissant l’action du récepteur sur le fonctionnement neuronal. / The type-1 cannabinoid receptor (CB1R), the neuronal receptor for the major psychoactive substance of marijuana, is one, of the most abundant G-protein coupled receptors in the mammalian central nervous system. CB1R is traditionally described as a presynaptic receptor that retrogradely regulates synaptic transmission. In addition to this now relatively wellcharacterized function, in the last two decades it has become widely recognized that endocannabinoid (eCB) actions in the brain are not limited to the regulation of neurotransmission at established adult synapses. Indeed, eCB and CB1R are now recognized to be involved in brain development at the synaptic, neuronal and network levels. However, precise mechanisms underlying these processes remain poorly described. Since cellular mechanisms that mediate CB1R-activition dependent neuronal remodeling and subneuronal targeting have been demonstrated to be cell-autonomous, we aimed to combine the power of Drosophila genetics with the experimental accessibility and single-cell resolution of lowdensity primary neuronal cultures, a tool currently lacking in Drosophila. Moreover, becauseDrosophila does not have a CB1R ortholog, CB1R cell biology may be observed independently from eCB machinery. Thus, we first developed and validated an in vitro culture protocol that yields mature and fully differentiated Drosophila neurons. Secondly, we showed that activation-dependent endocytosis of ectopically expressed CB1R is conserved in Drosophila neurons. Next, we investigated whether ectopic expression and activation of CB1R in Drosophila modulate neuronal development. As observed in mammals, we observed that activation of CB1R impairs dendritogenesis in a cell-autonomous manner. For further characterization of our model, we showed that, as with mammals, transient ectopic CB1R expression and activation in mushroom body neurons (the center of olfactory memory in Drosophila) modulate the formation of a consolidated form of aversive memory. In conclusion, the validation of this new animal model opens new perspectives to better characterize mechanisms underlying modulation of neuronal functions induced by CB1Ractivity

Chez les plantes, le RNA silencing (RNAi) est le principal mécanisme de défense antivirale. Il est opéré par de petites molécules d’ARN (siRNA), de 21-22nt de long, générées à partir de l’ARN viral par DCL4 et DCL2, respectivement. Ces siRNA confèrent la séquence-spécificité des réactions de défense intracellulaire et peuvent se déplacer à longue distance pour immuniser les cellules saines. En conséquence, les virus ont développé des protéines (VSRs) capables de supprimer ces deux aspects du RNAi. Au cours de cette thèse, j’ai pu démontrer in vivo que la protéine P15 du Peanut clump virus (PCV) est capable de séquestrer les siRNA de 21 et 22nt et qu’elle bloque le mouvement de ces derniers plus efficacement que ceux de 21nt. Pour compenser cette faiblesse, au cours de l’infection par le PCV, P15 est transportée à l’intérieur des peroxisomes en association avec les siRNA qu’elle séquestre. Le confinement des siRNA mobiles de 21nt à l’intérieur de ces organelles conduit à une inhibition du RNAi systémique et stimule fortement la propagation du PCV à travers la plante. Ces travaux définissent une nouvelle stratégie de pathogénèse virale au cours de laquelle une organelle est utilisé pour neutraliser des molécules de défense produites par l’hôte. / In plants, RNA interference (RNAi) is the main antiviral defense mechanism. It is initiated through the processing of viral RNA into 21-22nt long siRNA by DCL4 and DCL2, respectively. These siRNA can mediate sequence-specific local defense reactions (cell-autonomous RNAi) or move to distant tissues to prime defenses in naive cells (systemic RNAi). Consequently, viruses have evolved proteins (VSRs) to suppress both aspects of RNAi. In this in vivo study, I show that P15, the VSR of Peanut clump virus (PCV), binds and sequesters both 21nt and 22nt siRNA. Importantly, it stops the movement of 22nt siRNA more efficiently than 21nt siRNA. During infection, P15 is shuttled into peroxisomes, and is able to « piggyback » siRNA into these organelles. By confining mobile DCL4-dependent antiviral 21nt siRNA within peroxisomes, P15 is able to shut down systemic RNAi and strongly promote PCV movement. This work describes a novel pathogenic strategy in which an organelle is used to neutralize host defensive molecules.

Les acides gras à très longues chaine (VLCFA) sont essentiels dans le développement, particulièrement dans les mécanismes de trafic vésiculaires, de différenciation et division cellulaire. Cependant, le rôle de ces VLCFA dans ces différents processus chez les plantes n’est pas encore bien compris. Afin d’identifier de nouveaux acteurs associés à la biosynthèse ou la fonction des VLCFA, un crible suppresseur multicopies a été réalisé dans un mutant d’élongation des VLCFA de levure. La perte de l’activité déshydratase PHS1 chez la levure et de PASTICCINO2 chez les plantes perturbe la croissance et induit des défauts de cytokinèse. La PROTEIN TYROSIN PHOSPHATASE-LIKE (PTPLA) historiquement caractérisée comme une déshydratase inactive est capable de restaurer les défauts de croissance et d’élongation de phs1 mais non de pas2. PTPLA interagit avec plusieurs membres du complexe élongase dans le RE et son absence conduit à l’accumulation 3-hydroxyacyl-CoA, signature des déshydratases impliquées dans l’élongation des acides gras. Cependant, la perte de PTPLA conduit à une augmentation des VLCFA, probablement dépendante de PAS2 montrant que PTPLA serait un répresseur potentiel de l’élongation. Les deux déshydratases ont des profils d’expression divergents dans la racine. PAS2 est majoritairement exprimé dans l’endoderme tandis que PTPLA s’exprime uniquement dans les tissus vasculaires et le péricycle. La comparaison de l’expression ectopique de PAS2 et PTPLA dans leur tissus respectif confirme l’existence de deux complexe élongase indépendant associé à PAS2 ou PTPLA et interagissant de manière non cellule autonome. Les cytokinines pourraient constituer le signal entre les deux complexes élongase du fait que la biosynthèse de ces hormones est réprimée par les VLCFA. Les VLCFA répriment ainsi l'expression d'IPT3 dans les racines comme observées pour la partie apicale. Les cytokinines semblent aussi réguler la teneur en VLCFA dans la racine suggérant la présence de boucles de rétrocontrôles entre ces hormones et les VLCFA / Very long chain fatty acids (VLCFA) are involved in plant development and particularly in several cellular processes such as membrane trafficking, cell division and cell differentiation. However, the precise role of VLCFA in these different cellular processes is still poorly understood in plants. In order to identify new factors associated with the biosynthesis or function of VLCFA, a yeast multicopy suppressor screen was carried out in a yeast mutant strain defective for fatty acid elongation. Loss of function of the elongase dehydratase PHS1 in yeast and PASTICCINO2 in plants prevents growth and induces cytokinesis defects. PROTEIN TYROSIN PHOSPHATASE-LIKE (PTPLA) previously characterized as an inactive dehydratase was able to restore yeast phs1 growth and VLCFA elongation but not the plant pas2 defects. PTPLA interacted with elongase members in the ER and its absence induced the accumulation of 3-hydroxyacyl-CoA as expected from a dehydratase involved in fatty acid (FA) elongation. However, loss of PTPLA function led to increased VLCFA levels, effect that was dependent of the presence of PAS2 indicating that PTPLA activity repressed FA elongation. The two dehydratases have specific expression profiles in the root with PAS2, mostly restricted in the endodermis, while PTPLA was confined in the vascular tissue and pericycle cells. Comparative ectopic expression of PTPLA and PAS2 in their respective domains confirmed the existence of two independent elongase complexes comprising PAS2 or PTPLA that were functionally interacting in a non-cell autonomous manner. A putative regulating signal could involve cytokinins that were described to be regulated by VLCFA. VLCFA were indeed found to repress IPT3 expression in roots like in leaves. Cytokinins were also found to regulate VLCFA levels suggesting the existence of regulatory feedback loops between cytokinins and VLCFA

Les inhibiteurs de poly(ADP-ribose) polymérase (PARPi) ciblent sélectivement les cellules porteuses de défauts des voies de réparation de l’ADN tels que les mutations de BRCA1/2 et les défauts d’ERCC1. Sur le plan clinique, plusieurs PARPi ont été approuvés pour le traitement des cancers BRCA-mutés ou platine-sensibles du sein et de l’ovaire, et des essais cliniques sont en cours pour évaluer l’efficacité des PARPi dans le cancer bronchique non-à-petites cellules (CBNPC) platine-sensible. Alors que les PARPi ont un fort potentiel thérapeutique dans les cancers comportant des défauts de réparation de l’ADN, de plus en plus d’essais cliniques évaluent également l’efficacité de ces médicaments en combinaison avec les « inhibiteurs d’immune checkpoints » (ICI) dans diverses populations de patients. Dans ce contexte, il est essentiel de mieux comprendre comment les PARPi modulent la réponse immunitaire anti-tumorale, et d’étudier le potentiel immunologique inhérent de ces médicaments.Dans cette étude, nous avons établi que les cellules de CBNPC déficientes en ERCC1 expriment fortement la signature interféron (IFN) de type I, et que les tumeurs de CBNPC ayant une faible expression d’ERCC1 ont un infiltrat lymphocytaire renforcé. En utilisant des lignées cellulaires isogéniques et des xénogreffes dérivées de patients, nous avons montré que plusieurs PARPi, notamment l’olaparib et le rucaparib, ont des propriétés immunomodulatrices dans les modèles de CBNPC ERCC1-déficients et de cancers du sein triple-négatifs (CSTN) BRCA1-mutés. D’un point de vue mécanistique, les PARPi génèrent des fragments d’ADN cytoplasmiques ayant les caractéristiques de micronoyaux ; ceux-ci activent la voie cGAS/STING et déclenchent une réponse IFN de type I, associée à la sécrétion de la cytokine CCL5. De manière importante, ces effets sont largement diminués dans les cellules de CSTN BRCA1-révertantes et les cellules de CBNPC ré-exprimant ERCC1, ce qui suggère que les défauts de réparation de l’ADN amplifient les phénotypes immunitaires associés au traitement par PARPi. En outre, ces effets sont totalement abrogés dans les cellules de CSTN PARP1-neutralisées, ce qui confirme que les phénotypes observés dépendent d’un effet spécifique des PARPi sur leur cible.Au-delà de leur potentiel d’activation d’une immunité spécifique des cellules cancéreuses via cGAS/STING et la signalisation IFN de type I, nous avons également constaté que les PARPi potentialisent les effets inducteurs de l‘IFN de type II sur l’expression de PD-L1 dans des lignées cellulaires et cellules tumorales fraîches de patients CBNPC, surtout en présence de défauts d’ERCC1. De plus, nous avons montré que certains PARPi, utilisés à des concentrations létales, activent de manière indépendante les éléments moléculaires clés de la mort cellulaire immunogénique, dont l’exposition de la calréticuline à la surface des cellules cancéreuses, la sécrétion d’ATP et le relargage d’HMGB1 en grandes quantités dans le milieu extracellulaire.Dans l’ensemble, ces données précliniques suggèrent que les PARPi ont des propriétés immunomodulatrices intrinsèques qui participent à l’activation de réponses immunitaires anti-tumorales ; ce potentiel pourrait être exploité cliniquement en combinaison avec les ICI dans des populations adéquatement sélectionnées au plan moléculaire. / Poly(ADP-ribose) polymerase inhibitors (PARPi) selectively target cancer cells with DNA repair deficiencies such as BRCA1/2 mutations or ERCC1 defects. Clinically, several PARPi are currently approved for the treatment of BRCA-mutant or platinum-sensitive advanced ovarian and breast cancers, and ongoing clinical trials are investigating the efficacy of PARPi in platinum-sensitive Non-Small Cell Lung Cancer (NSCLC). While PARPi constitute potent targeted therapies for the treatment of DNA repair-deficient malignancies, an increasing number of clinical trials are also evaluating their efficacy in combination with immune checkpoint inhibitor (ICI) in various populations. In this context, it is of critical importance to better understand how PARPi might modulate immune responses against cancer, and to investigate the inherent immunological potential of these agents.In this study, we show that ERCC1-defective NSCLC cells exhibit an enhanced type I interferon (IFN) transcriptomic signature and that low ERCC1 expression correlates with increased lymphocytic infiltration in human NSCLC tumours. Using isogenic cell lines and patient-derived xenografts, we further demonstrate that several clinical PARPi, including olaparib and rucaparib, display cell-autonomous immunomodulatory properties in ERCC1-defective NSCLC and BRCA1-mutant triple-negative breast cancer (TNBC) models. Mechanistically, PARPi generate cytoplasmic chromatin fragments with micronuclei characteristics; this activates the cGAS/STING pathway and elicits downstream type I IFN signalling and CCL5 secretion. Importantly, these effects are suppressed in BRCA1-reverted TNBC cells and ERCC1-rescued NSCLC cells, suggesting that DNA repair defects exacerbate the innate immunity-related phenotypes triggered by PARPi. Similarly, these effects are totally abrogated in PARP1-null TNBC cells, supporting the on-target effect of PARPi in mediating such phenotypes. Besides this potential to activate tumour cell-autonomous immunity through cGAS/STING and type I IFN signalling, we also observed that PARPi synergize with type II IFN to induce PD-L1 expression in NSCLC cell lines and fresh patient tumour cells, especially in the ERCC1-deficient setting. Moreover, we show that lethal concentrations of some PARPi independently activate the key damage-associated molecular patterns dictating the immunogenicity of cancer cell death, including calreticulin exposure at the tumour cell surface, ATP secretion and HMGB1 release in the extracellular compartment.Together, these preclinical data suggest that PARPi have intrinsic immunomodulatory properties that activate anti-cancer immune responses; this could be exploited clinically in combination with ICI in appropriately molecularly-selected populations.

No description available.