Les virus influenza aviaires faiblement pathogènes (IAFP) ciblent principalement les épithéliums des voies respiratoires et intestinales chez les poulets (Gallus gallus) infectés. Cependant, les virus influenza aviaires hautement pathogènes (IAHP) mènent à une maladie systémique fatale avec une localisation particulière aux endothéliums. L’objectif de cette thèse a été d’explorer les relations entre la réplication des virus influenza aviaires (IA) et la réponse antivirale de l’hôte dans deux modèles cellulaires originaux obtenus chez le poulet : des cellules épithéliales pulmonaires (CLEC213) et des cellules endothéliales d’aortes (chAEC). Les résultats clés sont les suivants : (i) la réplication productive des virus IA dans les chAEC dépend du clivage de l’hémagglutinine et de l’échappement viral à la réponse immunitaire innée ; (ii) les CLEC213 sont très permissives aux virus IA et présentent une faible réponse antivirale médiée par la signalisation TLR3 et MDA5 ; (iii) les fonctions régulatrices de SOCS1 et SOCS3, sur le signal des interférons et des cytokines, sont conservées chez le poulet. Nous proposons que certains virus IA peuvent exploiter les fonctions pro-virales de SOCS1 et SOCS3 à leur avantage de manière spécifique au type cellulaire. / Low pathogenic avian influenza (LPAI) viruses essentially target the epithelia of the respiratory and intestinal tract in the infected chicken host (Gallus gallus). However, highly pathogenic avian influenza (HPAI) viruses induce a peracute fatal systemic disease and exhibit a striking endothelial cell tropism. The objective of the present thesis was to explore the interdependencies of AI virus replication and the antiviral host response in two novel avian cell culture models: chicken lung epithelial cells (CLEC213) and chicken aortic endothelial cells (chAEC). The salient findings from this study are that (i) productive AI virus replication in chAEC is dependent on hemagglutinin cleavability and appears to be related to innate immune escape; (ii) CLEC213 are highly permissive to AI virus infection, due to a cell type-specific diminished TLR3- and/or MDA5-mediated antiviral signaling response; (iii) the interferon and cytokine regulatory functions of SOCS1 and SOCS3 are conserved in the chicken. Based on our data, we propose a model that predicts that certain AI viruses may exploit the proviral functions of SOCS1 and SOCS3 in a cell type-specific manner.

京都大学 / 0048 / 新制・課程博士 / 博士(生命科学) / 甲第22596号 / 生博第429号 / 新制||生||57(附属図書館) / 京都大学大学院生命科学研究科統合生命科学専攻 / (主査)教授 藤田 尚志, 教授 永尾 雅哉, 教授 今吉 格 / 学位規則第4条第1項該当 / Doctor of Philosophy in Life Sciences / Kyoto University / DFAM

The innate immune pathways that stimulate the expression of cytokines and proapoptotic factors in response to infection are triggered by the activation of the cytosolic receptors retinoic acid-inducible gene I (RIG-I) and melanoma differentiationassociated gene 5 (MDA5). Activation of both receptors occurs as a result of binding to RNA. MDA5 only recognizes double stranded forms of RNA, whereas RIG-I is capable of recognizing both single and double stranded RNA. In vivo, MDA5 is known to be stimulated by long (>1 kb) strands of RNA, forming filaments along the phosphate backbone. However, the manner in which MDA5 can recognize the terminal end of its RNA ligand is uncertain. I have examined the mechanism of binding of the MDA5 protein by comparing MDA5 binding to short (<18 bp) blunt RNA, 5’ triphosphate RNA, and RNA with a 3’ or 5’ overhang. It is shown that while the MDA5 protein regulatory domain (RD) is essential for RNA recognition, the MDA5 RD only weakly recognizes short double stranded RNA ligands with overhangs or a 5’ triphosphate group. The Cys951 residue was shown to disrupt stability of the MDA5 RD-RNA complex. Binding analyses were performed using a combination of SDS-PAGE, gel filtration analysis, and nondenaturing gel electrophoresis. In addition, structural data was gathered by crystallization of the MDA5 RD-RNA complex using X-ray crystallography. These results help to establish the manner in which MDA5 is regulated predominantly to the binding of long RNA ligands. Also included in this document is structural data on the dimer form of the PepQ protein from E. coli. PepQ is a highly conserved proline peptidase that has a secondary activity of hydrolyzing organophosphorus triesters, toxic compounds found in many pesticides. The PepQ protein was crystallized and analyzed by X-ray diffraction. The dimer interface was clearly defined within the structure and provides insight into how the active dimer forms from the PepQ monomer.

Les maladies inflammatoires de l’intestin (MII), incluant la maladie de Crohn et la colite ulcéreuse, sont des maladies chroniques qui résultent d’un dérèglement de la réponse immunitaire aux microorganismes de la lumière intestinale. Des études de séquençages réalisées par le laboratoire du Dr. Rioux, avec ses collègues du International IBD Genetics Consortium ont identifié quatre variants rares, indépendants et causals pour les MII, dans le gène IFIH1. La protéine d’IFIH1 (MDA5) interagit avec certains virus à ARN, afin de déclencher une réponse antivirale de l’immunité innée. Nous avions émis l’hypothèse que ces variants dans IFIH1 diminuaient la réponse antivirale de l’épithélium intestinal, suite à une infection. Nous avons d’abord travaillé avec des lignées cellulaires lymphoblastoïdes (LCLs) obtenues à partir d’individus atteints de MII et qui sont homozygotes ou hétérozygotes composés pour ces variants, ainsi qu’à partir d’individus contrôles (IFIH1 wt). Ces LCLs ont été reprogrammées en cellules souches pluripotentes induites humaines, avant d’être différenciées en cultures épithéliales intestinales. Nos résultats ont d’abord confirmé l’impact des variants sur la structure génique et protéique (IFIH1/MDA5), dans ces modèles cellulaires. Puis, la réponse antivirale a été induite, grâce à la stimulation avec des agents (moléculaire et viral) connus pour stimulés la protéine MDA5. Nous avons démontré que ces variants dans IFIH1 induisaient effectivement une moins grande réponse antivirale, caractérisée par une plus faible expression d’IFNs, comparativement aux contrôles. Finalement, la modulation des IFNs constituerait une avenue potentiellement intéressante pour le traitement des patients atteints des MII et porteurs des variants causals. / Inflammatory Bowel Disease (IBD), including Cronh’s disease and ulcerative colitis, are chronic inflammatory diseases of the gastro-intestinal tract. IBD is associated with a disturbance of the immune response to the microorganisms of the intestinal lumen. Sequencing studies conducted by the laboratory of Dr. John Rioux, in collaboration with his colleagues of the International IBD Genetics Consortium, identified four rare and independent variants in IFIH1, associated to IBD. The protein of IFIH1 (MDA5) interacts with certain RNA viruses to trigger the innate mechanism of antiviral defense. Our hypothesis was that these IFIH1 variants decreased the intestinal epithelial antiviral response, following an infection. We first worked with lymphoblastoid cell lines (LCLs) obtained from IBD patients who are homozygotes or compound heterozygotes for the different variants, as well as from control individuals (IFIH1 wt). These LCLs were reprogrammed into human induced Pluripotent Stem Cells (hiPSCs), before being differentiated into intestinal epithelial cultures. Our results first confirmed the impact the variants on the genetic and protein structure for these models. Then, the antiviral response was triggered by the stimulation of LCLs and intestinal epithelial cells, with agents (molecular and viral) known to stimulate MDA5. We have demonstrated that these IFIH1 variants did indeed induce a lower antiviral response, characterized by lower IFNs expression, compared to control cell lines. Finally, modulation of IFNs could be an interesting avenue for the treatment of IBD patients with the causal variants.

Les interférons de type I (IFN I) sont des cytokines antivirales aux propriétés puissantes. L’induction, la transmission et la résolution de la réponse immunitaire engendrée par les IFN I est minutieusement régulée. Le concept d’interféronopathie de type I, récemment individualisé par notre équipe, repose sur l’hypothèse que certaines pathologies seraient secondaires au déséquilibre de ces voies de signalisation complexes et à la sécrétion excessive et inappropriée d’IFN I. L’inhibition de celle-ci par des thérapeutiques ciblées permettrait de valider cette hypothèse, si les symptômes allégués s’amélioraient, voire disparaissaient. Ce travail de thèse s’est initialement concentré sur la caractérisation clinique et biologique des interféronopathies monogéniques et polygéniques, et secondairement sur l’identification moléculaire de nouvelles mutations du gène TMEM173 à l’origine de l’interféronopathie liée à STING, également appelée SAVI (STING associated vasculopathy with onset in infancy), syndrome auto-inflammatoire associant une atteinte sévère cutanée et pulmonaire. De nouvelles techniques ont permis la sélection de patients présentant une augmentation de l’IFN I en comparaison à des contrôles sains : la signature IFN I, qPCR de 6 gènes stimulés par l’IFN (IFN stimulated genes – ISGs) et le dosage d’IFN alpha sérique ou plasmatique par méthode du SIMOA (single molecule array) permettant la détection de molécules d’IFN de l’ordre du femtogramme (10-18g). Ces méthodes nous ont ainsi permis d’élargir le spectre clinique phénotypique des interféronopathies de type I, initialement considéré comme essentiellement neurologique. Les patients atteints du syndrome d’Aicardi-Goutières, première interféronopathie monogénique décrite, présentaient les signes suivants : dystonie, spasticité, décalage des acquisitions, calcifications intra-cérébrales et anomalies de la substance blanche. Cependant, l’utilisation systématique de nos méthodes de criblage associée à l’avènement des technologies de séquençage à haut débit (next generation sequencing – NGS) a permis de révéler un phénotype plus large, caractéristique des interféronopathies de type I : sur le plan cutané (engelures, vascularite nécrosante des extrémités, sclérodermie), pulmonaire (pneumopathie interstitielle isolée ou non), musculo-squelettique (arthralgies, arthrites, arthropathie de Jaccoud, myalgies et myosites), ophtalmologique (glaucome), néphrologique (néphropathies lupiques), gastro-entérologique (maladies inflammatoires chroniques intestinales précoces), associées à de l’auto-immunité ou un déficit immunitaire inconstants. Notre méthode de sélection nous a notamment permis d’identifier des patients présentant de manière variable des signes cardinaux de SAVI et une de trois nouvelles mutations activatrices dans une région spécifique du gène TMEM173 (codant pour STING). Ces mutations circonscrivent une région de la protéine à ce jour encore jamais impliquée dans le contrôle de la voie de l’IFN I. STING est une protéine du réticulum endoplasmique qui agit comme adaptateur cytosolique de senseurs intracellulaires d’ADN viral dans une voie de signalisation de l’IFN I. STING active TBK1 (TANK-binding kinase) et permet la transcription des IFN I par la phosphorylation d’IRF3. La Janus Kinase 1 (JAK1) et la tyrosine kinase 2 (TYK2) sont activées suite à la stimulation des récepteurs de l’IFN I et phosphorylent les facteurs de transcription STAT1 et STAT2, conduisant à l’expression de nombreux ISGs. Les analyses génétiques, de conformation tridimensionnelle, sur un modèle cellulaire in vitro (HEK293T) et ex vivo sur cellules mononuclées périphériques des patients nous ont ainsi permis de mettre en évidence pour ces mutations un caractère constitutionnellement activé, indépendant de la liaison au ligand cGAMP, mais transmettant ce signal à travers la voie d’aval par TBK1. (...) / Type I interferons (IFN I) are antiviral cytokines with potent properties. Hence, the induction, transmission and resolution of the immune response generated by IFN I is tightly regulated. The concept of the type I interferonopathies, recently formulated by our team, rests on the assumption that some diseases arise from a disturbance of this complex signalling pathway, leading to excessive and inappropriate IFN I secretion. On this basis, targeted therapeutics should improve or cure features of such type I interferonopathies, thereby providing a validation of the underlying hypothesis. This PhD project initially focused on the clinical and biological characterisation of monogenic and polygenic interferonopathies, and secondarily on the molecular identification of novel mutations in the gene TMEM173 causing the interferonopathy called STING associated vasculopathy with onset in infancy (SAVI), an auto-inflammatory syndrome with severe cutaneous and pulmonary features. Our selection of patients in comparison to healthy controls was made possible through the use of novel screening tools: IFN signature (qPCR of 6 IFN stimulated genes – ISGs), and measurement of IFN alpha protein levels in serum or plasma (SIMOA-single molecule array - enabling the detection of molecules of IFN in the femtogram [10-18g]) range. In this way, we have been able to expand the phenotypic spectrum of the interferonopathies, which was initially considered as primarily neurological. Patients with Aicardi-Goutières syndrome (AGS), the first described of the monogenic interferonopathies, exhibit dystonia, spasticity, developmental delay, intra-cranial calcifications and white matter abnormalities. However, the systematic use of our interferon screening assays, plus the advent of next-generation sequencing technology, has revealed a much broader set of features relevant to this novel disease grouping – involving the skin (chilblains, necrotising vasculitis, scleroderma), lungs (isolated lung interstitial disease or associated with other signs), musculoskeletal system (joint pain, arthritis, Jaccoud’s arthropathy, muscle pain and myositis), eyes (glaucoma), kidneys (lupus nephritis) and gastro-intestinal tract (early inflammatory bowel disease), as well features of autoimmunity and immunodeficiency. Using our screening assays enabled us to identify three patients variably exhibiting the core features of SAVI, all of whom were found to harbour distinct novel activating mutations in STING. These mutations highlight a protein domain not previously implicated in the control of IFN I signalling. STING is an endoplasmic reticulum protein, acting as a cytosolic adaptor of intracellular sensors of viral DNA in the type I IFN signalling pathway. STING activates TANK-binding kinase (TBK1), allowing transcription of IFN I through phosphorylation of IRF3. Janus kinase 1 (JAK1) and tyrosine kinase 2 (TYK2) are activated following stimulation of the IFN I receptor, leading to phosphorylation of the transcription factors STAT1 and STAT2 and the subsequent induction of a large number of ISGs. Genetic analysis, conformational studies, an in vitro cellular model (HEK293T) and ex vivo experimental data (using patient peripheral blood mononuclear cells - PBMCs) enabled us to confirm the constitutive activating nature of these variants, and show that this activation did not require binding with cGAMP, but was dependent on signalling through TBK1. Ruxolitinib, a JAK1/2 inhibitor, could antagonise this constitutive activation ex vivo. These results indicate a promising therapeutic approach in such patients, and more widely in the monogenic, and perhaps even, polygenic, interferonopathy context.

Human immunodeficiency virus type 1 (HIV-1) is a lentivirus that causes acquired immunodeficiency syndrome (AIDS). Since the first cases of AIDS were described in 1981, HIV-1 has become one of the most serious public health threats in the world. There are approximately 38 million people worldwide currently living with HIV-1. 28 million of these people have access to antiretroviral therapy (ART) that is highly effective in reducing viral load to undetectable levels, thereby curbing the risk of viral transmission and preventing progression to AIDS. Despite their effectiveness in suppressing HIV-1 viremia, systemic inflammation remains as a hallmark of HIV-1 infection in vivo. This persistent immune activation is often associated with non-AIDS related complications, including elevated risk of neurocognitive and cardiovascular disorders. Several different mechanisms may contribute to this chronic immune activation and inflammation in people living with HIV-1 on ART. One of the contributing factors might be HIV-1 RNA expressed from the provirus. Even though ART potently suppresses HIV-1 replication, it fails to eradicate proviruses established prior to initiation of ART. Ongoing activation of CD4+ T cells and macrophages by HIV-1 proviral transcripts might contribute to the persistent inflammation that remains even after HIV-1 suppression by ART. Previously, our laboratory has shown that induction of innate immune signaling after HIV-1 challenge of primary human dendritic cells (DCs), macrophages, or CD4+ T cells requires integration, transcription from the nascent provirus, and nuclear export of intron-containing HIV-1 RNA through the Rev-CRM1 pathway. However, these studies failed to identify the innate immune sensor of intron-containing HIV-1 RNA. Here we conducted a targeted loss-of-function screen, using shRNA-expressing lentivectors in human DCs to identify this innate immune receptor. Of the twenty-one candidate genes targeted for knockdown by shRNA, the innate immune response to HIV-1 was inhibited only by knockdown of IFIH1, MAVS, and XPO1. The effect of IFIH1 and MAVS knockdowns on HIV-1-induced immune activation was confirmed in macrophages, and rescue of the knockdown with non-targetable coding sequence showed that IFIH1 protein was required. IFIH1 mutants that are defective for interaction with MAVS blocked activation, demonstrating that MAVS acts downstream of IFIH1 in this system. Since both IFIH1 and DDX58 signal via MAVS, the specificity of HIV-1 RNA detection by IFIH1 was demonstrated by the fact that DDX58 knockdown had no effect on activation; the IFIH1-specific inhibitor Nipah virus V protein blocked the activation by HIV-1. RNA-Seq showed that IFIH1-knockdown in DCs globally disrupted the induction of IFN-1-stimulated genes. Altogether, results presented in this thesis reveal that IFIH1 is required for innate immune activation by intron-containing RNA from the HIV-1 provirus, and potentially contributes to chronic inflammation in people living with HIV-1.