There are four serotypes of dengue virus designated dengue 1, 2, 3 and 4 (D1, D2, D3 and D4) and epidemiological studies indicate that a severe complication of dengue virus infection - dengue hemorrhagic fever (DHF) is more likely to occur following a secondary infection. DHF is hypothesized to be immunologically mediated and may be triggered by virus-specific T cells. It is also likely that dengue virus-specific cytotoxic T lymphocytes (CTLs) are important for recovery from dengue virus infections. An analysis of the immune response during acute illness and when the patient has recovered from the infection (immune state) is therefore important as it will provide insights into the immunopathological nature of the disease. This thesis initially examines the CD8+CTL responses in volunteers who have received live attenuated dengue vaccines and then investigates acute and immune T cell responses in children following natural infection with dengue. When this project was initiated, there was little available information on the human CD8+ T cell responses to dengue viruses. PBMC from one donor had generated memory CD8+CTL to the nonstructural protein NS3 of dengue virus. Memory CD8+CTL responses were therefore analyzed to determine the diversity of the T cell response to dengue virus and to identify immunodominant proteins using PBMC from eight healthy adult American volunteers who had received monovalent live-attenuated candidate vaccines of the 4 dengue serotypes. All the donors had specific T cell proliferation to dengue viruses and to other flaviviruses that we tested. CTLs were generated from the stimulated PBMC of all donors and in the seven donors tested, dengue virus-specific CD8+CTL activity was demonstrated. The nonstructural proteins NS3 and NS1.2a and the structural protein E were recognized by CD8+CTLs from six, five and three donors respectively. All donors recognized either NS3 or NS 1.2a. In a donor who received a dengue 4 vaccine, CTL killing was seen in bulk culture against the premembrane protein (prM). This is the first demonstration of a CTL response against the prM protein. The CTL responses using PBMC of two donors were serotype-specific whereas all other donors had serotype-cross reactive responses. For one donor, CTLs specific for E, NSl.2a and NS3 proteins were all HLA-B44 restricted. For the three other donors tested the potential restricting alleles for recognition of NS3 were HLA-B38, A24 and/or B62 and B35. These results indicate that the CD8+CTL responses of humans after immunization with a single serotype of dengue virus are diverse and directed against a variety of proteins. The nonstructural proteins NS3 and NSl.2a appear to be immunodominant and should be considered when designing subunit vaccines for dengue. Previously T cell responses had not been examined in people who have had natural infections with dengue. The HLA diversity between North American Caucasians and populations where dengue is a serious health problem, calls for the analysis of immune responses in people who have been infected with natural circulating strains of the virus. We examined the memory cytotoxic T lymphocytic (CTL) responses of peripheral blood mononuclear cells (PBMC) obtained from patients in Thailand 12 months after natural symptomatic secondary dengue infections. In all four patients analyzed, CTLs were detected in bulk culture PBMC against nonstructural dengue proteins. Numerous CD4+ and CD8+ CTL lines were generated from the bulk cultures of two patients, KPP94-037 and KPP94-024, which were specific for the NSl.2a and NS3 proteins respectively. All CTL lines derived from both patients were crossreactive with other serotypes of dengue virus. The CD8+ NS1.2a specific lines from patient KPP94-037 were HLA-B57 restricted and the CD8+ NS3 specific lines from patient KPP94-024 were HLA-B7 restricted. The CD4+ CTL lines from patient KPP94-037 were HLA-DR7 restricted. A majority of the CD8+CTLs isolated from patient KPP94-024 were found to recognize a.a. 221-232 on NS3. These results demonstrate that after symptomatic secondary natural dengue infections in Thai patients, CTLs are mainly directed against nonstructural proteins and are broadly crossreactive. The data correlate with our observations that nonstructural proteins are immunodominant proteins in volunteers who received dengue vaccines. We were interested in examining CTL responses in children during their acute illness and comparing them to memory CTLs obtained from the same children a year or more after the infection. A detailed analysis on samples from nine patients during their acute illness failed to generate any dengue virus-specific CTL responses. We therefore decided to determine if cell mediated responses are altered during acute dengue infection. Decreased proliferative responses to mitogens and recall antigens have been observed in PBMC obtained during several acute human viral infections. All responses of PBMC during acute illness were compared to the same patients PBMC obtained at least 6 months after their infection. Proliferative responses to PHA, anti-CD3, tetanus toxoid and dengue antigens were significantly decreased in PBMC obtained during the acute infection. The proliferative responses to PHA were restored by the addition of gamma-irradiated autologous immune or allogeneic PBMC. Cell contact with the irradiated PBMC was necessary to restore proliferation. Non-T cells from the acute PBMC of dengue patients did not support proliferation of T cells from control donors in response to PHA, but T cells from the PBMC of patients with acute dengue proliferated if accessory cells from a control donor were present. Addition of anti-CD28 antibodies restored anti-CD3-induced proliferation of the PBMC of some patients. The percentage of monocytes was reduced in the acute sample of PBMC of the dengue patients. Addition of IL-2 or IL-7, but not IL-4 or IL-12 also restored proliferation of acute PBMC stimulated with anti-CD3. The results demonstrate that both quantitative and qualitative defects in the accessory cell population during acute dengue illness result in a depression of in vitro T cell proliferation. The data generated from this project shed light on the nature of the immune responses during acute natural dengue infections. It strengthens the existing data on the human memory CD8+CTL responses to dengue viruses and validates the observations by examining memory CTL responses after natural dengue infection in patients from Thailand. In addition, we demonstrate a profound defect in lymphoproliferative responses during dengue illness.

Submitted by Michel Batista (mbatista@fiocruz.br) on 2014-11-28T11:47:15Z No. of bitstreams: 1 Tese Guilherme Ferreira Silveira.pdf: 9523097 bytes, checksum: 1944bdd5853c5e3e3e1fe8bda7c59f5b (MD5) / Approved for entry into archive by Michel Batista (mbatista@fiocruz.br) on 2014-11-28T11:50:19Z (GMT) No. of bitstreams: 1 Tese Guilherme Ferreira Silveira.pdf: 9523097 bytes, checksum: 1944bdd5853c5e3e3e1fe8bda7c59f5b (MD5) / Made available in DSpace on 2014-11-28T11:50:19Z (GMT). No. of bitstreams: 1 Tese Guilherme Ferreira Silveira.pdf: 9523097 bytes, checksum: 1944bdd5853c5e3e3e1fe8bda7c59f5b (MD5) / Fundação Oswaldo Cruz. Instituto Carlos Chagas, Fiocruz-PR. Curitiba, PR, Brasil. / A dengue representa na atualidade a mais prevalente arbovirose, caracterizando-se como um grave problema de saúde pública, tanto no Brasil como em todo o mundo. No entanto, o paciente que desenvolve a febre da dengue e/ou dengue com complicações recebe apenas tratamento de suporte, uma vez que não existem medicamentos ou vacinas específicas contra o vírus. Adicionalmente, pouco se sabe sobre os mecanismos imunopatológicos que desencadeiam os sintomas graves. Um dos aspectos que parece ser chave na patogenia é a resposta funcional dos linfócitos ao vírus da dengue (DV). Embora, a susceptibilidade destas células à infecção pelo DV ainda não fora bem caracterizada. Deste modo, os alvos principais deste trabalho foram à determinação da susceptibilidade e a caracterização da resposta funcional de linfócitos humanos de doadores saudáveis, analisadas na infecção por cepas dos quatro sorotipos do DV. Foi demonstrado que três diferentes populações de linfócitos (LT CD4+, LT CD8+ e LB CD19+) são susceptíveis ao DV, especialmente o LT CD8+, sendo o heparam sulfato um receptor de entrada para a infecção destas células pelo DV. Esta infecção foi produtiva, com replicação ativa e liberação de partículas virais viáveis no sobrenadante das culturas, além da produção da proteína não-estrutural 1 (NS1). Contudo, foram observados apenas níveis discretos de ativação celular e produção de citocinas. Adicionalmente, não foram observadas respostas funcionais como: apoptose de linfócitos, degranulação de LT CD8+, infecção diferencial de LT CD8+ “naïve” ou LT CD8+ de memória, resposta de polarização Th1/Th17 de LT CD4+ ou produção de IgM anti-DENV por LB CD19+. Ademais, a infecção dos linfócitos não foi capaz de reduzir a proliferação celular observada em ensaios de Ensaio Misto de Leucócitos e Ensaio de “Prime” de Linfócitos, sendo observada uma discreta elevação na proliferação quando as células foram infectadas com o DV4 360. Do mesmo modo, as infecções de culturas de células mononucleares do sangue periférico (PBMC) pelas cepas DV3 98 e DV4 360 foram capazes de induzir uma polarização de perfil Th2. Portanto, estes achados demonstram que, apesar das infecções produtivas, os linfócitos são pouco responsivos aos DV, sendo que importantes funções destas células não foram reguladas pela infecção. Desta forma, é possível sugerir que os linfócitos seriam uma fonte importante para a manutenção da viremia observada nos casos de dengue. / Dengue fever is the most prevalent arboviral disease, characterized as a serious public health problem, both in Brazil and worldwide. The patient who develops dengue fever, which could culminate in severe complications, receives only live-support therapy, since there are no specific drugs or vaccines against the virus. Furthermore, little is known about the immunopathogenic mechanisms that trigger the severe symptoms in dengue. One aspect that seems to be key in the pathogenesis is the functional response of lymphocytes to the dengue (DV) infection. The susceptibility of these cells to DV infection is not well characterized. Therefore, the main targets of this study were the determination of ex vivo susceptibility by DV infection and the characterization of functional response of human lymphocytes. Cells from healthy donors were analyzed against four DV serotypes. It has been shown that three populations of lymphocytes (LT CD4+, LT CD8+ and LB CD19+) are susceptible to DV, especially LT CD8+, infection by DV which occurs through the participation of heparan sulfate as the receptor. Infection was productive, with active replication and release of viable viral particles and NS1 protein in the culture supernatants. Only discrete levels of cellular activation and cytokine production were observed. In addition, lymphocyte apoptosis, LT CD8+ degranulation, differential infection of naïve or memory LT CD8+, LT CD4+ polarization (Th1/Th17) or production of IgM anti-DENV by LB CD19+ were not modulated. Furthermore, the infection of the lymphocytes was not able to reduce cell proliferation induced in Mixed Leukocyte Reaction and Lymphocytes Prime Assay, showing a slight increase in proliferation when the cells were infected with DV4 360. PBMC infection with DV3 98 and DV4 360 strains was able to induce a Th2 polarized response. These findings suggest that, despite the productive infection, lymphocytes are poor responders to DV, and important functions of these cells were not modulated by DV infection. However, it is possible that lymphocytes would be an important source for maintenance of viremia observed in dengue fever.

Título de Ingeniero en Biotecnología Molecular / Las Gap Junctions (GJs) son clústers de canales intercelulares localizados en la membrana plasmática que permiten la comunicación directa entre células adyacentes. Cada canal GJ está compuesto por 2 hemicanales hexaméricos conocidos como conexones, los cuales a su vez están compuestos por seis proteínas de transmembrana llamadas conexinas (Cxs), siendo Cx43 la más representada en el sistema inmune. Se ha vinculado la variación de los niveles normales de expresión de las Cxs con la patogénesis de distintas enfermedades, incluyendo sordera congénita, arritmias cardiacas y cáncer. Dentro de esta última nuestro laboratorio ha desarrollado proyectos de investigación que permitan determinar el papel de las GJs en la respuesta inmune antitumoral. Por ejemplo, publicaciones previas de nuestro laboratorio han demostrado que Cx43 participa en la transferencia directa de antígenos tumorales entre células dendríticas (DCs). Así mismo, Cx43 polariza hacia la sinapsis entre DC y linfocitos T (LT) CD4+, formando canales funcionales que permiten una comunicación bidireccional requerida para la activación de LT mediada por APC. Además, se ha encontrado que Cx43 juega un rol en la activación de células natural killer (NK) por parte de DCs y en la regulación de la citotoxicidad de las células NK contra células tumorales. Así mismo, resultados no publicados indican que Cx43 polariza hacia el sitio de contacto entre LT CD8+ citotóxicos (CTLs) y células de melanoma, sugiriendo que Cx43 participaría en la citotoxicidad de CTLs en contra de éstas. Estos antecedentes sugieren que las GJs podría participar en el transporte de moléculas o señales que se encuentren involucradas en el proceso de activación de LT CD8+ mediada por DCs, tema que aún no ha sido estudiado por las investigaciones actuales. En estudios 2 anteriores nuestro laboratorio ha trabajado con modelos celulares humanos, sin embargo, dado que la obtención de LT vírgenes específicos contra el tumor a partir de donantes sanos es improbable y engorroso, en este trabajo decidimos utilizar el modelo murino específico para melanoma pMEL-1 y validar y optimizar su utilización para estudios de la participación de las GJs en la activación de LT CD8+ por parte de DCs. Nuestros resultados indican que el modelo pMEL-1 no representa alteraciones en la expresión de Cx43 respecto a su contraparte silvestre (wild type (WT)), y que es un modelo muy sensible, encontrando expresión del marcador de activación CD69 incluso desde 1 hora post co-cultivo. Además, se determinó que existe transferencia de calceína entre moDC y LT CD8+ pMEL-1 en un contexto estrictamente antígeno específico. Sin embargo, no pudimos dilucidar si las GJs participan del proceso de activación ya que el método de inhibición utilizado, el cual es un péptido mimético de Cx43, presentó problemas técnicos de funcionamiento, fenómeno que fue comprobado en modelos donde anteriormente se había determinado la comunicación mediada por GJs. Proponemos en el futuro utilizar otros métodos de inhibición, como el uso de siRNA contra Cx43, para determinar si estos canales juegan un rol en este proceso. / Gap Junctions are clusters of intercellular channels located in the plasma membrane that allow direct communication between adjacent cells. Each GJ channel is composed of 2 hexameric hemichannels known as connexions, which are composed by six transmembrane proteins called conexins (Cxs), being Cx43 the most represented one in the inmmune system. Variation in normal expression levels on Cxs have been linked to the development of different diseases, including congenital deafness, cardiac arrhythmias and cancer, being the last the one where our laboratory has developed investigation lines that allow to determine the role of GJs in the antitumoral response. In example, previous publications of our group have demonstrated that Cx43 participates in the direct transfer of tumoral antigens between dendritic cells (DCs) and that polarizes to the synapsis between DCs and CD4+ T lymphocytes, forming functional channels that allow bidirectional communication required for the DCs mediated activation of LT. Besides this, we have found that Cx43 plays a role in the activation of natural killer (NK) cells mediated by DCs, and in the regulation of the cytotoxicity of NK cells against tumor cells. Likewise, unpublished data indicate that Cx43 polarizes to the contact site between cytotoxic CD8+ T cells (CTLs) and melanoma cells, suggesting that Cx43 participates in the cytotoxicity of these cells. These backgrounds suggest that GJs may participate in the transport of molecules or signals that may be involved in the process of DC mediated activation of CD8+ T cells, subject that hasn’t been studied by the actual investigations. In previous studies our group has worked with human cellular models, however, given that obtaining naïve T cells from healthy donors is unlikely and difficult, in this work we decided to use the melanoma specific murine model pMEL-1 and valid and optimize it’s use for studies of the participation of GJs in the DCs 4 activation of T cells. Our results indicate that the pMEL-1 model doesn’t have alterations in the Cx43 expression levels compared with the wild type background and that it is a very sensible model, in which we can evaluate the activation of T cells even an hour after co-culture. Besides, it was determined the existence of calcein transfer between moDCs and CD8+ pMEL-1 T cells in a strictly antigen specific context. However, we could not elucidate if the GJs participate or not in the CD8+ T cell activation process, as our inhibition method, which is a Cx43 mimetic peptide, presented technical problems, phenomenon that was proven in models where GJ mediated communication has previously been determinate. We propose in the future to use a different inhibition method, as the use of anti-Cx43 siRNA, in order to determine if these channels play a role in this process.

abstract: CD8+ T-lymphocytes (CTLs) are central to the immunologic control of infections and are currently at the forefront of strategies that enhance immune based treatment of a variety of tumors. Effective T-cell based vaccines and immunotherapies fundamentally rely on the interaction of CTLs with peptide-human leukocyte antigen class I (HLA-I) complexes on the infected/malignant cell surface. However, how CTLs are able to respond to antigenic peptides with high specificity is largely unknown. Also unknown, are the different mechanisms underlying tumor immune evasion from CTL-mediated cytotoxicity. In this dissertation, I investigate the immunogenicity and dysfunction of CTLs for the development of novel T-cell therapies. Project 1 explores the biochemical hallmarks associated with HLA-I binding peptides that result in a CTL-immune response. The results reveal amino acid hydrophobicity of T-cell receptor (TCR) contact residues within immunogenic CTL-epitopes as a critical parameter for CTL-self/nonself discrimination. Project 2 develops a bioinformatic and experimental methodology for the identification of CTL-epitopes from low frequency T-cells against tumor antigens and chronic viruses. This methodology is employed in Project 3 to identify novel immunogenic CTL-epitopes from human papillomavirus (HPV)-associated head and neck cancer patients. In Project 3, I further study the mechanisms of HPV-specific T-cell dysfunction, and I demonstrate that combination inhibition of Indoleamine 2, 3-dioxygenase (IDO-1) and programmed cell death protein (PD-1) can be a potential immunotherapy against HPV+ head and neck cancers. Lastly, in Project 4, I develop a single-cell assay for high-throughput identification of antigens targeted by CTLs from whole pathogenome libraries. Thus, this dissertation contributes to fundamental T-cell immunobiology by identifying rules of T-cell immunogenicity and dysfunction, as well as to translational immunology by identifying novel CTL-epitopes, and therapeutic targets for T-cell immunotherapy. / Dissertation/Thesis / Doctoral Dissertation Biological Design 2017

Ce travail de thèse a été axé sur l’étude de la réponse immunitaire T chez des patients atteints d’artérite à cellules géantes (ACG) et de pseudo-polyarthrite rhizomélique (PPR). Plusieurs études cliniques successives interrégionales ont permis d’inclure de nombreux patients (57 ACG et 27 PPR) des Centres Hospitaliers (CH) Universitaires et des CH de l’interrégion Est. Les échantillons sanguins ont été étudiés dans le laboratoire de l’unité INSERM U1098. Tout d’abord, nous avons confirmé l’implication des lymphocytes Th17 dans la pathogénie de l’ACG et avons montré pour la première fois leur implication au cours de la PPR. De plus, notre étude des lymphocytes T (LT) CD4+CD161+ a permis de mieux comprendre les mécanismes de plasticité entre les réponses Th1 et Th17 au cours de ces deux pathologies. Nous avons complété ces travaux par l’étude de la réponse T régulatrice en montrant qu’il existe un déficit quantitatif en Treg au cours de l’ACG et la PPR. Dans la suite de ce travail, nous avons mis en évidence, chez des patients atteints de polyarthrite rhumatoïde, que le blocage de la voie de signalisation de l’IL-6 par un anticorps monoclonal dirigé contre le récepteur de l’IL-6 permet de corriger le déséquilibre de la balance Th17/Treg, en diminuant la réponse Th17 et en augmentant simultanément la réponse T régulatrice, à l’inverse des corticoïdes qui diminuent le pourcentage de Th17 sans corriger le déficit en Treg. Enfin, dans la dernière partie de ce travail, nous avons montré pour la première fois que les LT CD8+ étaient également impliqués dans la pathogénie de l’ACG et la PPR. Ces résultats ont permis de progresser dans les connaissances physiopathologiques de l’ACG et la PPR en évoluant d’un modèle articulé autour d’un déséquilibre de la balance Th1/Th2 vers celui d’un déséquilibre de la balance Th17/Treg et permettent de proposer des thérapeutiques mieux ciblées pour l’ACG et la PPR. / The aim of this thesis was to investigate the T-cell immune response in the course of giant-cell arteritis (GCA) and polymyalgia rheumatica (PMR). Several studies conducted by our team allowed us to obtain blood samples from many patients affected by GCA (n=57) and PMR (n=28). Immunological studies were performed in INSERM U1098, University Of Burgundy, Dijon, France. We firstly demonstrated the implication of Th17 and CD4+CD161+ T cells in the pathogenesis of these two diseases, thus extending the knowledge in the plasticity mechanisms arising between Th1 and Th17 cell-immune responses in GCA and PMR. Furthermore, we investigated the regulatory T cell immune response in these two affections, demonstrating that although being functional, the percentage of circulating Treg was decreased in GCA and PMR patients. As interleukin-6 (IL-6) had been shown to control the Th17/Treg balance, we studied Th17 and Treg frequencies in rheumatoid arthritis patients treated with an anti-IL-6 receptor antibody (tocilizumab). We showed that the blockade of the IL-6 pathway was able to correct the Th17/Treg imbalance by decreasing the number of Th17 cells and simultaneously increasing that of Treg. Finally, we demonstrated for the first time the implication of CD8+ T cells in the pathogenesis of GCA and PMR. This thesis allowed us to progress in the knowledge of the pathogenesis of GCA making the pathogenesis model progress from a Th1/Th2 to a Th17/Treg imbalance model. Altogether, these data deciphering the immune response in the pathogenesis of GCA and PMR bring new knowledge which will lead to better targeted therapies.

Allogeneic hematopoietic stem cell transplantation (allo-HCT) has been a successful cellular therapy for patients suffering from hematological malignancies for many decades; however, the beneficial effects of graft-versus-leukemia (GVL) are classically offset by graft-versus-host disease (GVHD). GVHD occurs when major and/or minor human leukocyte antigen (HLA) mismatches between donor and recipient cause rapid expansion and activation of donor effector T cells (Teffs) resulting in end organ damage to the recipient’s epithelial tissues. Given the lymphoproliferative nature of this disease, the standard treatment option is broad immunosuppression, which can result in primary disease relapse, steroid refractory GVHD, and/or opportunistic infection. A more targeted therapy that can selectively suppress GVH responses with maintained GVL responses would achieve the optimal goal of allo-HCT. Regulatory T cells (Tregs) both natural (nTregs) or induced (iTregs) could be potential cellular therapies for the treatment of GVHD, given their innate suppressive function. Initial clinical trials using nTregs have yielded positive results; however, nTreg cellular therapy has been cumbersome due to the necessity for large scale ex vivo expansion given their low yield within an apheresis product and non-specific suppression. Conversely, iTregs can be generated from naïve T cells thus decreasing ex vivo culture times and can be educated with specific antigen thus providing targeted suppression, but a consensus on their efficacy for GVHD therapy has not been reached. Therefore, we investigated the efficacy of antigen specific iTreg therapy for the prevention of GVHD while maintaining GVL responses. In Chapter 2, we evaluated the effectiveness of monoclonal HY-specific iTregs in GVHD attenuation. We chose HY as a target antigen because it is a naturally processed, ubiquitously expressed minor mismatch antigen carried by only male donors/recipients cited to increase GVHD prevalence when donor and recipient are sex-mismatched. Utilizing HY-transgenic mice in which all T cells recognize HY antigen exclusively, we generated HY specific iTregs which effectively attenuating GVHD in male, but not female recipients in three murine bone marrow transplantation (BMT) models (major mismatch, parent to F1, and miHAg mismatch). We found HY specific iTregs lost stability in female recipients but remained stable and suppressive in male recipients suggesting expression of HY antigen was required for their suppressive function and stability. GVL responses were not compromised with the addition of HY specific iTregs in recipient mice using a pre-established tumor model. Thus, HY-specific iTregs can be generated and suppress GVHD in an antigen-dependent manner while sparing the GVL effect. In Chapter 3, we extend our findings in Chapter 2, which provided proof of principle that antigen specific iTregs effectively control GVHD; however, this therapy has a limited translational potential. Therefore, we generated alloreactive CD4 and CD8 iTregs and evaluated GVHD attenuation and GVL preservation in either full or haplo-MHC mismatched BMT models. We found alloreactive CD4 iTregs significantly suppress lethal GVHD, but completely abrogated the GVL effect against aggressive tumors. Conversely, alloreactive CD8 iTregs moderately attenuated GVHD and possessed direct cytotoxicity against tumor cells. Therefore, to rescue the impaired GVL effect mediated by CD4 iTregs, we established a combinational therapy with CD8 iTregs. Indeed we found combination CD4 and CD8 iTreg therapy significantly suppressed GVHD while sparing GVL responses compared to either CD4 or CD8 singular therapy. Mechanistically, this was achieved by potent suppression of both CD4 and CD8 Teffs coupled with preserved cytolytic molecule expression by both CD8 iTregs and Teffs. Taken together, we propose antigen specific iTreg therapy can effectively attenuate GVHD while preserving GVL responses. We further uncovered unique characteristics of CD4 and CD8 iTregs that can be exploited to achieve the optimal cellular therapy following allo-HCT.

京都大学 / 新制・課程博士 / 博士(医科学) / 甲第23109号 / 医科博第120号 / 新制||医科||8(附属図書館) / 京都大学大学院医学研究科医科学専攻 / (主査)教授 河本 宏, 教授 中島 貴子, 教授 小川 誠司 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

No description available.

During an acute virus infection, antigen-specific CD8 T cells undergo clonal expansion and differentiation into effector cells in order to control the infection. Efficient clonal expansion and differentiation of CD8 T cells are required to develop protective memory CD8 T cells. Antigen specific cells require 3 distinct signals for their activation: TCR engagement of peptide-MHC (signal 1), costimulation between B7 and CD28 (signal 2), and inflammatory cytokines including IL-12 or type 1 IFN (signal 3). CD8 T cells that encounter antigen and costimulation undergo programmed cell division, but these two signals alone are not sufficient for full effector cell differentiation and survival into memory. CD8 T cells need a third signal for efficient clonal expansion, differentiation into various effector populations, acquisition of cytolytic effector functions, and memory formation. The requirements for signal 3 cytokines in CD8 T cell activation have only been recently described; however, the timing of exposure to these signals has yet to be investigated. During the course of an immune response not all T cells will see antigen, costimulation, and inflammatory cytokines at the same time or in the same order. I sought to examine how the timing of signal 3 cytokines affected CD8 T cell activation. I questioned how the order of these signals effected CD8 T cell priming and subsequent activation, expansion and differentiation. In order to study the in vivo effects of out-of-sequence signaling on CD8 T cell activation, I utilized poly(I:C), a dsRNA analogue, which is known to induce a strong type 1 IFN response. Through the use of various congenic transgenic and polyclonal CD8 T cell populations, in conjunction with adoptive transfer models, specific T cells which had been exposed to poly(I:C) induced environments could be identified and tracked over time. I wanted to characterize how out-of-sequence signaling affected T cell activation immediately after cognate antigen stimulation (4-5hours), and after prolonged exposure to cognate antigen (days-weeks). Considering type 1 IFN can have both inhibitory and stimulatory effects on CD8 T cell proliferation, and when type 1 IFN provides signal 3 cytokine activity, it has positive effects on CD8 T cell expansion, I wanted to investigate the role of type 1 IFN as an out-of-sequence signal during CD8 T cell activation. We identified a transient defect in the phosphorylation of downstream STAT molecules after IFNβ signaling within poly(I:C) pretreated CD8 T cells. The inability of poly(I:C) pretreated CD8 T cells to respond to IFNβ signaling makes these cells behave in a manner more similar to T cells that only received 2 signals, rather than ones that received all 3 signals in the appropriate order. Consequently, poly(I:C) pretreated, or out-of-sequence, CD8 T cells were found to have defects in clonal expansion, effector differentiation and function as well as memory generation resulting in reduced efficacy of viral clearance. Out-of-sequence CD8 T cells showed suppression of CD8 T cell responses after prolonged exposure to cognate antigen, but naïve CD8 T cells pre-exposed to poly(I:C) exhibited immediate effector function within hours of cognate antigen stimulation, prior to cell division. Poly(I:C) pretreated naïve CD8 T cells acquired an early activated phenotype associated with alterations of transcription factors and surface markers. Changes in naïve CD8 T cell phenotype are thought to be mediated by poly(I:C)-induced upregulation of self-MHC and costimulatory molecules on APCs through direct type 1 IFN signaling. Inoculating with poly(I:C) enabled naive CD8 T cells to produce effector functions immediately upon stimulation with high density cognate antigen, reduced affinity altered peptide ligands (APLs), and in response to reduced concentrations of cognate antigen. Unlike conventional naïve CD8 T cells, poly(I:C) pretreated naïve CD8 T cells acquired the ability to specifically lyse target cells. These studies identified how the timing of activation signals can dramatically affect the acquisition of CD8 T cell effector function. This thesis describes how CD8 T cell exposure to activation signals in an unconventional order may result in altered response to antigen stimulation. Exposure of naïve CD8 T cells to type 1 IFN and costimulatory molecules in the presence of self-peptides enabled them to respond immediately upon antigen stimulation. Primed naïve CD8 T cells produced multiple cytokines in response to low-affinity, and low-density antigens, and gained ability to specifically lyse target cells. However, immediate effector function may come at the expense of clonal expansion and effector cell differentiation in response to prolonged antigen exposure as out-of-sequence CD8 T cells showed reduced proliferation, effector function and memory formation. The findings presented here may seem contradictory because out-of-sequence signaling can prime T cells to produce immediate effector functions and yet cause defects in T cell expansion and effector differentiation. However, these two models ascertained T cell function at different points after antigen exposure; one where functions were evaluated within hours after seeing cognate antigen, and the other showing T cell responses after days of antigen stimulation. Studies described in this thesis highlight the growing complexity of CD8 T cell activation. Not only do the presence or absence of signals 1-3 contribute to T cell activation, but the timing of these signals also proves to be of great importance. These studies may describe how both latecomer and third party antigen specific T cells behave when and if they encounter cognate antigen in the midst of an ongoing infection. Out-of-sequence exposure to IFN initially stimulates effector function but at the expense of efficient clonal expansion and subsequent memory formation. The immediate effector function that naïve T cells gain during out-of-sequence priming may explain how some individuals are more resistant to superinfections, whereas the impairment in proliferation describes a universal mechanism of virus-induced immune suppression, explaining how other individuals can be more susceptible to secondary infections. Ultimately, results identified here can be applied to developing better and more effective vaccines.

Heterologous immunity is a mechanism by which immunological memory within an individual, developed in response to a previous infection, plays a role in the immune response to a subsequent unrelated infection. In murine studies, heterologous immunity facilitated by cross-reactive CD8 T-cell responses can mediate either beneficial (protective immunity) or detrimental effects (e.g. enhanced lung and adipose immunopathology and enhanced viral titers) (Selin et al., 1998; Chen et al., 2001; Welsh and Selin, 2002; Nie et al., 2010; Welsh et al., 2010). Protective heterologous immunity results in enhanced clearance of virus during a subsequent infection with an unrelated pathogen. Such is the case when mice are immunized with lymphocytic choriomeningitis virus (LCMV) and subsequently challenged with Pichinde virus (PV) or vaccinia virus (VACV) (Selin et al., 1998). However, heterologous immunity may also mediate enhanced immunopathology as mice immunized with influenza A virus (IAV) and challenged with LCMV show increased viral titers and enhanced lung immunopathology (Chen et al., 2003). The role heterologous immunity plays during infection is not limited to the murine system. In fact, there have now been several reports of enhanced immunopathology due to heterologous immunity during human infections, involving viruses such as IAV, Epstein-Barr Virus (EBV), hepatitis C virus (HCV), and dengue virus (DENV) (Mathew et al., 1998; Wedemeyer et al., 2001; Acierno et al., 2003; Nilges et al., 2003; Clute et al., 2005; Urbani et al., 2005). Interestingly, in all reported cases in humans, heterologous immunity mediated enhanced immunopathology. Upon infection with EBV the clinical presentation can range from asymptomatic to severe, occasionally fatal, acute infectious mononucleosis (AIM) (Crawford et al., 2006b; Luzuriaga and Sullivan, 2010) which is marked by a massive CD8 lymphocytosis. This lympho-proliferative effect in AIM was shown to be partially mediated by reactivation of cross-reactive IAV-M1 58-66 (IAV-GIL) specific CD8 memory T-cells in HLA-A2 patients reacting to the EBV-BMLF1 280 (EBV-GLC) epitope (Clute et al., 2005). Interestingly, EBV infects ~90% of individuals globally by the third decade of life, establishing a life-long infection (Henle et al., 1969). However, it is unknown why 5-10% of adults remain EBV-sero-negative (EBV-SN), despite the fact that the virus infects the vast majority of the population and is actively shed at high titers even during chronic infection (Hadinoto et al., 2009). Here, we show that EBV-SN HLA-A2+ adults possess cross-reactive IAV-GIL/EBV-GLC memory CD8 T-cells that show highly unique properties. These IAV-GIL cross-reactive memory CD8 T-cells preferentially expand and produce cytokines to EBV antigens at high functional avidity. Additionally, they are capable of lysing EBV-infected targets and show the potential to enter the mucosal epithelial tissue, where infection is thought to initiate, by CD103 expression. This protective capacity of these cross-reactive memory CD8 T-cells may be explained by a unique T-cell receptor (TCR) repertoire that differs by both organization and CDR3 usage from that in EBV-seropositive (EBV-SP) donors. The composition of the CD8 T-cell repertoire is a dynamic process that begins during the stochastic positive selection of the T-cell pool during development in the thymus. Thus, upon egress to the periphery a naïve T-cell pool, or repertoire, is formed that is variable even between genetically identical individuals. This T-cell repertoire is not static, as each new infection leaves its mark on the repertoire once again by stochastically selecting and expanding best-fit effectors and memory populations to battle each new infection while at the same time deleting older memory CD8 T-cells to make room for the new memory cells (Selin et al., 1999). These events induce an altered repertoire that is unique to each individual at each infection. It is this dynamic and variable organization of the T-cell repertoire that leads to private specificity even between genetically identical individuals upon infection with the same pathogens and thus a different fate (Kim et al., 2005; Cornberg et al., 2006a; Nie et al., 2010). It is this private specificity of the TCR repertoire that helps explain why individuals with the same epitope specific cross-reactive response, but composed of different cross-reactive T-cell clones, can either develop AIM or never become infected with EBV. Our results suggest that heterologous immunity may protect EBV-SN adults against the establishment of productive EBV infection, and potentially be the first demonstration of protective T-cell heterologous immunity between unrelated pathogens in humans. Our results also suggest that CD8 T-cell immunity can be sterilizing and that an individual’s TCR repertoire ultimately determines their fate during infection. To conclusively show that heterologous immunity is actively protecting EBV-SN adults from the establishment of a productive EBV infection, one would have to deliberately expose an individual to the virus. Clearly, this is not an acceptable risk, and it could endanger the health of an individual. A humanized mouse model could allow one to address this question. However, before we can even attempt to address the question of heterologous immunity mediating protection from EBV infection in humanized mice, we must first determine whether these mice can be infected with, and build an immune response to the two viruses we are studying, EBV and IAV. We show here that these mice can indeed be infected with and also mount an immune response to EBV. Additionally, these mice can also be infected with IAV. However, at this time the immune responses that are made to these viruses in our established humanized mouse model are not substantial enough to fully mimic a human immune response capable of testing our hypothesis of heterologous immunity mediating protection from EBV infection. Although the immune response in these mice to EBV and IAV infection is not suitable for the testing of our model the data are promising, as the humanized mouse model is constantly improving. Hopefully, with constant improvements being made there will be a model that will duplicate a human immune system in its entirety. This thesis will be divided into 5 major chapters. The first chapter will provide an introduction to both general T-cell biology and also to the role of heterologous immunity in viral infection. The second chapter will provide the details of the experimental procedures that were performed to test our hypothesis. The third chapter will describe the main scientific investigation of the role of heterologous immunity in providing natural resistance to infection in human subjects. This chapter will also consist of the data that will be compiled into a manuscript for publication in a peer-reviewed journal. The fourth chapter will consist of work performed pertaining to the establishment of a humanized mouse model of EBV and IAV infection. The establishment of this model is important for us to be able to show causation for protection from EBV infection mediated by heterologous immunity.