CD4\(^+\) and CD8\(^+\) T cell populations can be divided into subpopulations based on expression of surface markers CCR7 and CD45RA. The resulting populations are referred to as naive, central memory, effector memory and effector memory RA\(^+\) (EMRA). The aim of this study was to identify potential lineage relationships between these subpopulations for both CD4\(^+\) and CD8\(^+\) T cells through microarray analysis. The genes found to distinguish between these subpopulations include many molecules with known functions in T cell differentiation, including CCR7, CD45RA, granzymes, L-selectin and TNF receptors. Several genes from the tetraspanin family of proteins were found to be differentially expressed at mRNA and protein level; suggesting a possible role for these genes in CD4\(^+\) and CD8\(^+\) T cell activation, migration and lysosomal function. Other genes identified, such as LRRN3 and CXCR5 which were expressed highest on naive and CM T cells respectively, provide interesting gene targets to follow up on their function in these T cell populations. Microarray data was validated through Real Time PCR and suggests that both CD4\(^+\) and CD8\(^+\) T cells differentiate along a linear pathway of naive to central memory to effector memory. The transcriptional programmes responsible for these differentiation steps were distinct between CD4\(^+\) and CD8\(^+\) T cells, although additional elements were common to both subsets.

This project investigated the effect of β2-adrenergic receptor (β2-AR) stimulation on the function of cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells in vitro. Catecholamines have been previously shown to increase efflux of these cells into the blood, but the effects on cell function are unclear. In this thesis three aspects of function have been addressed. The results of the studies presented in this thesis showed that: (1) β2-adrenergic stimulation by salmeterol reduced the percentage of IFN-γ producing CD4+ and CD8+ T cells activated by Staphylococcus Aureus enterotoxin type B (SEB) superantigen, cytomegalovirus lysate (CMV) or CMV pp65 (pp65) recombinant protein. (2) salmeterol, at high concentrations, increased rolling behaviour and decreased stationary behaviour of peripheral blood lymphocytes (PBLs) on human microvascular cell line (HMEC-1) and on vascular cell adhesion molecule-1 (VCAM-1), in both flow and static assays. (3) adrenergic stimulation impaired the activation and cytotoxic function of CD8+ T and NK cells, as indicated by lower expression of CD107a (a marker of CD8+ T and NK cell activation and function) following incubation of peripheral blood mononuclear cells (PBMCs) with human erythromyeloblastoid leukemia (K562) cell line or MHC class I chain-related gene A (MICA*009) transfected Chinese hamster ovary cells (T-CHO) was analysed. The results presented in this thesis showed that adrenergic stimulation influences a number of cellular functions, such as those related to migration, cytokine production and cytotoxic function. Together the above studies may contribute to our understanding about how stress affects the ability of the cytotoxic cells.

TNF receptor/ligand superfamily members signal through pathways giving rise to proteins that regulate lymphocyte proliferation, activation, differentiation and survival. Absence of TNF ligands OX40L and CD30L impairs survival of GC T cells and affinity maturation of antibody responses. Direct effects of these molecules on B cells in antibody responses are not characterised. I dissected roles of OX40 and CD30 for B cells using T-independent type II (TI-II) antigen NP-Ficoll. Humoral immunity is impaired in OX40 deficiency. Defects in class switched and non-class switched antibody production are due to reduced development of antigen-specific switched and non-switched plasma cells. CD30 has an opposing role, deficiency results in similar or higher switched and non-switched antibody titres and higher numbers of antigen-specific plasma cells that develop rapidly. This may explain why in OX40/CD30 double deficiency, there is a less pronounced defect than in OX40 single deficiency. B cell intrinsic roles are revealed for OX40 and CD30 that suggest OX40 on B cells is critical for TI-II plasmablast differentiation or survival and B cell CD30 inhibits onset of plasmablast differentiation.

Macrophages are present in virtually all tissues and account for approximately 10% of all body mass. Although classically credited as the scavenger cells of innate immune system, ridding a host of pathogenic material and cellular debris though their phagocytic function, macrophages also play a crucial role in embryogenesis, homeostasis, and inflammation. De-regulation of macrophage function is therefore implicated in the progression of many disease states including cancer, arthritis, and atherosclerosis to name just a few. The diverse range of activities of this cell can be attributed to its exceptional phenotypic plasticity i.e. it is capable of adapting its physiology depending on its environment; for instance in response to different types of pathogens, or specific cocktail of cytokines detected. This plasticity is exemplified by the macrophages capacity to adjust rapidly its transcriptional profile in response to a given stimulus. This includes interferons which are a group of cytokines capable of activating the macrophage by interacting with their cognate receptors on the cell. The different classes of interferons activate downstream signalling cascades, eventually leading to the expression (as well as repression) of hundreds of genes. To begin to fully understand the properties of a dynamic cell such as the macrophage arguably requires a holistic appreciation of its constituents and their interactions. Systems biology investigations aim to escape from a gene-centric view of biological systems. As such this necessitates the development of better ways to order, display, mine and analyse biological information, from our knowledge of protein interactions and the systems they form, to the output of high throughput technologies. The primary objectives of this research were to further characterise the signalling mechanisms driving macrophages activation, especially in response to type-I and type- II interferons, as well as lipopolysaccharide (LPS), using a ‘systems-level’ approach to data analysis and modelling. In order to achieve this end I have explored and developed methods for the executing a ‘systems-level’ analysis. Specifically the questions addressed included: (a) How does one begin to formalise and model the existing knowledge of signalling pathways in the macrophage? (b) What are the similarities and differences between the macrophage response to different types of interferon (namely interferon-β (IFN-β) and interferon-γ (IFN-γ))? (c) How is the macrophage transcriptome affected by siRNA targeting of key regulators of the interferon pathway? (d) To what extent does a model of macrophage signalling aid interpretation of the data generated from functional genomics screens? There is general agreement amongst biologists about the need for high-quality pathway diagrams and a method to formalize the way biological pathways are depicted. In an effort to better understand the molecular networks that underpin macrophage activation an in-silico model or ‘map’ of relevant pathways was constructed by extracting information from published literature describing the interactions of individual constituents of this cell and the processes they modulate (Chapter-2). During its construction process many challenges of converting pathway knowledge into computationally-tractable yet ‘understandable’ diagrams, were to be addressed. The final model comprised 2,170 components connected by 2,553 edges, and is to date the most comprehensive formalised model of macrophage signalling. Nevertheless this still represents just a modest body of knowledge on the cell. Related to the pathway modelling efforts was the need for standardising the graphical depiction of biology in order to achieve these ends. The methods for implementing this and agreeing a ‘standard’ has been the subject of some debate. Described herein (in Chapter-3) is the development of one graphical notation system for biology the modified Edinburgh Pathway Notation (mEPN). By constructing the model of macrophage signalling it has been possible to test and extensively refine the original notation into an intuitive, yet flexible scheme capable of describing a range of biological concepts. The hope is that the mEPN development work will contribute to the on-going community effort to develop and agree a standard for depicting pathways and the published version will provide a coherent guide to those planning to construct pathway diagrams of their biological systems of interest. With a desire to better understand the transcriptional response of primary mouse macrophages to interferon stimulation, genome wide expression profiling was performed and an explorative-network based method applied for analysing the data generated (Chapter-4). Although transcriptomics data pertaining to interferon stimulation of macrophages is not entirely novel, the network based analysis of it provided an alternative approach to visualise, mine and interpret the output. The analysis revealed overlap in the transcriptional targets of the two classes of interferon, as well as processes preferentially induced by either cytokine; for example MHC-Class II antigen processing and presentation by IFN-γ, and an anti-proliferative signature by IFN-β. To further investigate the contribution of individual proteins towards generating the type-I (IFN-β) response, short interfering RNA (siRNA) were employed to repress the expression of selected target genes. However in macrophages and other cells equipped with pathogen detection systems the act of siRNA trasfection can itself induce a type-I interferon response. It was therefore necessary to contend with this autocrine production of IFN-β and optimise an in vitro assay for studying the contribution of siRNA induced gene-knock downs to the interferon response (described in Chapter-5). The final assay design incorporated LPS stimulation of the macrophages, as a means of inducing IFN-β autonomously of the transfection induced type-I response. However genome-wide expression analysis indicated the targeted gene knock-downs did not perturb the LPS response in macrophages on this occasion. The optimisation process underscored the complexities of performing siRNA gene knockdown studies in primary macrophages. Furthermore a more thorough understanding of the transcriptional response of macrophages to stimulation by interferon or by LPS was required. Therefore the final investigations of this thesis (Chapter-6) explore the transcriptional changes over a 24 hour time-course of macrophage activation by IFN-β, IFN-γ, or LPS and the contribution of the macrophage pathway model in interpreting the response to the three stimuli. Taken together the work described in this thesis highlight the advances to be made from a systems-based approach to visualisation, modelling and analysis of macrophage signalling.

Fibrocytes are monocyte-derived cells that morphologically look like fibroblasts, express both stromal and haematopoietic markers, and have been reported as being involved in wound healing and fibrosis. In-vitro derived fibrocytes can be differentiated in both serum-containing and serum-free environments and we wanted to study the relationship between these two fibrocytes; which potentially could be involved at different time points at a wound healing site. To investigate the relationship between serum-free and serum-containing derived fibrocytes monocytes were differentiated without serum. When these cells were placed in a serum-containing environment they became round, losing their fibroblast-like morphology. However when the reverse experiment was done on fibrocytes derived in a serum-containing environment there was no apparent effect on their morphology. The relationships between these two fibrocytes, as well as macrophages and fibroblasts was also examined using transcriptome analysis of 37000 genes, clustering the samples based on all the genes, and identifying those that were significantly different between the populations. This demonstrated that both fibrocyte populations are distinct from each other, as well as from both fibroblasts and macrophages. These data demonstrate that these two fibrocytes have different characteristics, suggesting that they may have different roles in the modulation of fibrosis in inflammation.

The immune system has long been attributed cognitive capacities such as "recognition" of pathogenic agents; "memory" of previous infections; "regulation" of a cavalry of detector and effector cells; and "adaptation" to a changing environment and evolving threats. Ostensibly, in preventing disease the immune system must be capable of discriminating states of pathology in the organism; identifying causal agents or ``pathogens''; and correctly deploying lethal effector mechanisms. What is more, these behaviours must be learnt insomuch as the paternal genes cannot encode the pathogenic environment of the child. Insights into the mechanisms underlying these phenomena are of interest, not only to immunologists, but to computer scientists pushing the envelope of machine autonomy. This thesis approaches these phenomena from the perspective that immunological processes are inherently inferential processes. By considering the immune system as a statistical decision maker, we attempt to build a bridge between the traditionally distinct fields of biological modelling and statistical modelling. Through a mixture of novel theoretical and empirical analysis we assert the efficacy of competitive exclusion as a general principle that benefits both. For the immunologist, the statistical modelling perspective allows us to better determine that which is phenomenologically sufficient from the mass of observational data, providing quantitative insight that may offer relief from existing dichotomies. For the computer scientist, the biological modelling perspective results in a theoretically transparent and empirically effective numerical method that is able to finesse the trade-off between myopic greediness and intractability in domains such as sparse approximation, continuous learning and boosting weak heuristics. Together, we offer this as a modern reformulation of the interface between computer science and immunology, established in the seminal work of Perelson and collaborators, over 20 years ago.

Injury, infection and tissue malfunction are triggers of inflammation which if not regulated may acquire new characteristics that result in pathological outcomes. Since innate immunity plays a key role in the resolution of acute inflammation knowledge of the regulation of this component of the host response is relevant to understanding processes in disease progression and therefore has potential clinical benefits. In this thesis I have applied zebrafish as a model organism to investigate the response of innate immune cells to qualitatively distinct inflammatory signals in the absence of adaptive immunity. Using a zebrafish embryo wound injury model I have investigated leukocyte migration profiles by in vivo imaging. In response to wound alone leukocytes migrated to the site of injury with predominantly random walk behaviour. However, the addition of lipopolysaccharide (LPS) enhanced recruitment and influenced the directionality of leukocyte migration to the wound. I demonstrate that leukocyte dynamic behaviour is also dependent on the location of the cells. The LPS enhanced directionality and reduced the random walk behaviour of the leukocytes, and these effects were ablated in the presence of the p38 mitogenactivated protein kinase (MAPK) specific inhibitor SB203580. Cytokine gene profiling in adult zebrafish leukocytes reveals that LPS can stimulate a pro-inflammatory response via the activation of p38 MAPK characteristic of mammalian innate immune responses. It is documented in mammalian innate immune cells that LPS can modulate Notch mediated signalling and thereby cell function. Using zebrafish with null mutations in Notch, which provide an unbiased in vivo model, I have investigated the influence of Notch signalling on leukocyte recruitment and demonstrate that migration to a wound injury is reduced. However, this effect is due to decreased cell numbers and not altered function as the Notch signalling inhibitor DAPT had no effect of recruitment to wound injury. The defect in myelomonocyte numbers was also present in adult zebrafish and this was partially compensated for by an increase in lymphocytes. The experimental results that I report here highlight zebrafish as a model 2 organism for studying the function and regulation of innate immunity. The unique optical translucency, which permits in vivo imaging of host responses in real-time, facilitates the analysis of the innate immune response to different inflammatory signals and immune modulators.

The thymic stroma primarily comprises epithelial, mesenchymal and endothelial cells, interspersed with those of haematopoietic origin. Thymic epithelial cells (TECs) are highly heterogeneous, but can be divided into two broad lineages, cortical and medullary, based on phenotype, functionality and location. A population of Plet1+ TEC progenitors have been identified which, when isolated from mouse E12.5 or E15.5 fetal thymus, reaggregated, and grafted, can produce a functional thymus. However, the potential of individual progenitors to form cortex and/or medulla is undefined. The main aim of this thesis was to use retrospective clonal analysis to ascertain the point during thymus ontogeny at which the cortical and medullary lineages diverge. To this end, I used transgenic mice carrying a ubiquitous ROSA26laacZ reporter gene (where a duplication within lacZ encodes non-functional b-galactosidase). Here, rare, random laacZ-lacZ genetic recombinations result in heritable expression of functional b-gal, producing labelled clones. As this occurs at a known frequency, determination of TEC numbers would enable calculation of the expected number of TEC clones present throughout ontogeny. Due to the lack of quantitative data on all thymic cell populations, I determined the size not only of TEC (lin-EpCAM+), but also haematopoietic (CD45+), mesenchymal (lin-PDGFRa+ and/or lin-PDGFRb+) and endothelial (lin-CD31+) populations from E12.5 until E17.5. I then showed that the absolute number of Plet1+ TECs remains constant during this time, although the proportion of Plet1+ cells in cycle decreases. From these collective data, I propose a model for the role of the Plet1+ population in thymus development, in which Plet1+ cells continually give rise to Plet1- TECs in a self-renewing manner. Finally, I present a ‘dual origin coefficient’ strategy for analysis of a library of prospective TEC clones. I calculated the number of TEC lacZ+ clones expected to be present throughout thymus ontogeny, selecting an appropriate developmental stage for analysis. Although I observed several clones of apparent mesenchymal origin, supporting a single origin for intrathymic and capsular mesenchyme at E15.5, I observed no TEC clones in this extensive analysis. The CpG content of the ROSA26 promoter suggests a possibility of methylation-induced silencing brought about by de novo methylation of the lacZ reporter gene.

Les cyclophilines appartiennent à une famille de protéines initialement caractérisées par leur activité peptidyl-prolyl cis/trans isomérase et leur capacité à fixer la cyclosporine A. Les formes sécrétées des cyclophilines A (CyPA) et B (CyPB) induisent la migration de sous-populations leucocytaires, via un mécanisme dépendant de l’interaction avec un récepteur cellulaire commun, le CD147. Des données récentes ont également montré que la CyPA induit la sécrétion de cytokines pro-inflammatoires, telles que le TNF-α, ce qui suggère un rôle des cyclophilines sécrétées dans la régulation de la réponse inflammatoire. Dans ce contexte, l’objectif de mes travaux de thèse a consisté à étudier le rôle de la CyPB dans la modulation des réponses inflammatoires du macrophage. Contrairement à la CyPA, la CyPB n’induit pas la sécrétion de cytokines pro-inflammatoires. A l’inverse, la pré-incubation des cellules en présence de CyPB réduit leur production en réponse au LPS. L’analyse des réponses induites par la CyPB nous a permis de mettre en évidence l’expression de Bcl-3, un régulateur négatif de l’activité du facteur de transcription NF-κB. Nous avons alors montré que la CyPB inhibe la production de TNF-α induite par le LPS en bloquant l’activation transcriptionnelle du gène codant pour cette cytokine. En complément, nous avons mis en évidence un mécanisme de régulation post-transcriptionnelle qui pourrait faire intervenir MKP-1, une phosphatase impliquée dans l’inactivation des MAPKs et dans la déstabilisation de l’ARNm du TNF-α. Dans leur ensemble, ces travaux démontrent que la CyPB pourrait agir comme une protéine régulatrice de la réponse inflammatoire. / Cyclophilins are members of a family of proteins initially characterised for their peptidyl-prolyl cis/trans isomerase activity and their ability to interact with cyclosporine A. Secreted forms of cyclophilin A (CyPA) and B (CyPB) induce the migration of leukocyte subsets by a mechanism dependent on the interaction with a common cellular receptor, CD147. Recently published data have also shown that CyPA induced the expression of pro-inflammatory cytokines, such as TNF-α, suggesting a role for secreted cyclophilins in the regulation of inflammatory responses. In this context, the aim of my thesis was to study the implication of CyPB in the modulation of inflammatory responses of macrophages. Unlike CyPA, CyPB did not induce the secretion of pro-inflammatory cytokines. Conversely, pre-treatment with CyPB reduced the production of cytokines from LPS-stimulated macrophages. Analysis of the responses induced by CyPB has highlighted the expression of Bcl-3, a negative regulator of the transcriptional factor NF-κB. Then, we showed that CyPB inhibited LPS-induced production of TNF-α by blocking the transcriptional activation of the gene coding for this cytokine. In addition, we characterized a mechanism of post-transcriptional regulation that could bring into play MKP-1, a phosphatase involved in down-regulation of MAPKs and destabilization of TNF-α mRNA. Altogether, these findings indicate that CyPB could act as a regulatory protein in the inflammatory responses.

Les maladies à prions et d’Alzheimer appartiennent à un groupe de maladies neurodégénératives caractérisées par l’accumulation dans le système nerveux central (SNC) de protéines amyloïdes, respectivement la PrPSc et les peptides Aβ. Même si ces maladies ont des étiologies et des manifestations physiopathologiques distinctes, il est suspecté que des mécanismes communs de neurodégénérescence puissent être mobilisés dans ces différentes affections du SNC. Les maladies à prions s’imposent comme un paradigme qui permet l’étude des maladies neurodégénératives amyloïdes. Disposer d’un agent infectieux, la protéine prion scrapie, PrPSc, présente l’avantage de pouvoir initier un processus neurodégénératif et de cerner la nature et la séquence des événements menant à la perte d’homéostasie neuronale. Les mécanismes mis en évidence grâce à l’infection à prions peuvent être objectivés dans d’autres maladies neurodégénératives. Dans les maladies à prions, il est clairement établi que la PrPSc exerce sa toxicité dans les neurones en déviant la/les fonction(s) de la forme non pathologique des prions, la protéine prion cellulaire, PrPC. Les travaux du laboratoire ont permis d’assigner une fonction de signalisation à la PrPC et d’identifier plusieurs intermédiaires de signalisation contrôlés par la PrPC, ce qui a conduit à proposer plusieurs rôles pour la PrPC dans les neurones : régulation de l’équilibre d’oxydoréduction, adhérence, neuritogenèse, survie, contrôle des fonctions associées au neuromédiateur. Une partie de mes travaux de thèse a permis d’illustrer une nouvelle facette de la PrPC dans le contrôle des fonctions neuronales. Au travers d’un couplage à la kinase Lyn et d’une interaction avec la protéine LRP1 et le cuivre, la PrPC du corps cellulaire gouverne l’état d’activation de la kinase GSK3β, qui à son tour, contrôle le trafic et l’activité d’un autorécepteur sérotoninergique, le récepteur 5HT1B. En modulant l’activité de ce récepteur, la PrPC favorise la neurotransmission. A partir de l’infection à prions, mes travaux dévoilent des mécanismes de neurodégénérescence communs aux maladies à prions et à la maladie d’Alzheimer (AD). Dans les neurones infectés par les prions, comme les neurones dérivés de souris modèles pour AD, la suractivation de la kinase PDK1 provoque la phosphorylation et l’internalisation de l’αsécrétase TACE, ce qui annule l’activité neuroprotectrice de TACE à la membrane plasmique. TACE internalisée est découplée de trois de ses substrats, (i) la PrPC, ce qui favorise sa conversion en PrPSc, (ii) la protéine précurseur des peptides amyloïdes APP, ce qui augmente la production des peptides neurotoxiques Aβ et (iii) les récepteurs au TNFα, ce qui rend les neurones malades vulnérables au stress inflammatoire. In vitro comme in vivo, l’inhibition de PDK1 permet de rétablir l’activité neuroprotectrice de TACE et de contrecarrer les effets neurotoxiques de la PrPSc ou de Aβ. Mes travaux établissent également que les Rho kinases (ROCK) sont des régulateurs positifs de l’activité de PDK1. Dans un contexte physiologique, les ROCK interagissent avec PDK1 et phosphorylent PDK1, contribuant à son activité basale. Dans un contexte infectieux, le gain d’activité des ROCK augmente le « pool » de molécules de PDK1 qui interagissent avec et sont phosphorylées par les ROCK, à l’origine de la suractivation de PDK1. Inhiber les ROCK exerce un double effet protecteur dans les neurones infectés par les prions en abaissant le niveau de PrPSc via le module de signalisation PDK1TACE et en préservant la polarité et la connectivité des neurones par action sur le cytosquelette d’actine. Le module ROCKPDK1 émerge comme une cible thérapeutique potentielle pour les maladies à prions et autres maladies neurodégénératives amyloïdes. / Pas de résumé en anglais