Epigenetic regulation of gene expression plays a pivotal role in the orchestration of immune responses. In particular, they have been implicated in the generation of in vitro cytokine-driven T cell polarization and therefore may determine the vigor, quality and/or longevity of such responses in vivo. Chemical allergens form two categories: skin sensitizing chemicals associated with allergic contact dermatitis, such as 2,4-dinitrochlorobenzene (DNCB) that result in type 1/type 17 responses in mice, and chemicals that cause sensitization of the respiratory tract and occupational asthma, for example trimellitic anhydride (TMA) that induces preferential type 2 responses in mice. To explore the regulation and maintenance of these divergent responses generated by polarised T cell populations in vivo, BALB/c strain mice were exposed topically DNCB and TMA. DNA from draining lymph nodes (LN) was processed for methylated DNA (5mC) immunoprecipitation (MeDIP) followed by hybridization to a whole-genome DNA promoter array. A higher number of DNCB-associated differently methylated regions (DMR) were identified and there was significant crossover between allergen treatments. Promoter-associated DMR, unique to either DNCB or TMA, were generally hypomethylated. Pathway analyses highlighted a number of immune related pathways, including chemokine and cytokine signalling. A number of these DMR were hypothesised to be candidate biomarkers of chemical allergy. To confirm this, novel analysis of hydroxymethylated (5hmC) DNA in the in vivo allergen-activated LN was compared to analysis of 5mC to identify LN specific DMR. The Gmpr DMR is suggested as a possible biomarker for contact allergen-induced immune responses and the Nwc DMR was characteristic of TMA treatment, highlighting its possible utility as a biomarker for responses induced by chemical respiratory allergens. These data not only represent novel analysis of 5hmC in response to chemical allergy in vivo, but also provide a possible basis for differentiation between classes of chemical allergens. Finally, a combined population of effector/effector memory T cells (TEff/TEM) was isolated from the CD4+ and CD8+ populations of allergen-activated draining lymph nodes (LN). Levels of 5mC and 5hmC at T cell lineage cytokine prompters was determined and analysed by comparison with concurrently sorted naïve T cells. In CD8+ TEff/TEM from DNCB-stimulated LN, increased expression of Ifng and Gzmb correlated with a reduction 5mC at their respective promoters. There were also reduced levels of 5mC at an Ifng enhancer. In contrast, TMA-simulated CD4+ TEff/TEM were characterised by high levels of Il4 expression which were associated with a decrease in promoter 5mC and an increase in 5hmC, as well as increased 5hmC at an Il4 enhancer region. These data demonstrate that exposure to chemical allergens results in characteristic DNA methylation patterns indicative of epigenetic regulation of divergent T cell populations in vivo. Furthermore, it highlights a particularly important role for DNA hydroxymethylation at the Th2 locus. In conclusion, exposure to chemical allergens results in divergent patterns of 5mC and 5hmC. These provide possible biomarkers for the different classes of chemical allergens and represent an insight into the importance of 5mC and 5hmC in the control of polarised T cell responses in vivo.

Dans cette thèse nous étudions l'influence de l'environnement sur le comportement d'une cellule dans deux situations différentes. Dans chacune de ces deux situations, apparaît un couplage non-linéaire sur le champ d'advection lié à un terme non-local provenant du bord du domaine. Dans une première partie, nous modélisons la polarisation cellulaire durant la conjugaison de la cellule de levure. Nous utilisons un modèle de type convection-diffusion avec un terme de convection non-linéaire et non-local. Ce modèle présente des similarités avec le modèle de Keller-Segel, la source du potentiel attractif étant sur le bord du domaine. Nous étudions le cas de la dimension un en utilisant des inégalités de Sobolev logarithmiques et HWI. En nous appuyant sur un raisonnement heuristique, nous ramenons l'étude de notre modèle en dimension deux au bord du domaine. Nous validons le modèle à l'aide des résultats expérimentaux obtenus par M. Piel en utilisant un bruit dynamique dans nos simulations numériques. Nous étudions ensuite le problème du dialogue cellulaire entre cellules de levure de sexe opposé. Dans une seconde partie, nous étudions la réaction immunitaire durant l'athérosclérose. Nous construisons puis développons un modèle structuré en âge pour décrire l'inflammation. Pour des paramètres particuliers, nous déterminons le comportement en temps long de notre système en utilisant une fonctionnelle de Lyapunov. / We investigate the influence of the environment on the behaviour of a cell in two different situations. In each of these situations, there is a non-linear coupling of the drift due to a non-local term coming from the boundary of the domain.The first part focuses on the modeling of cell polarisation during the mating of yeast. We use a convection-diffusion model with a non-linear and non-local drift. This model is similar to the Keller-Segel model, the source of the attractive potential comes from the boundary of the domain. We study the long time behaviour of the one-dimensional case by using logarithmic Sobolev and HWI inequalities.By relying on a heuristic, we reduce the study of our model in the two-dimensional case to the boundary of the domain. We validate the model with data provided by M. Piel. This validation requires adding a dynamical noise in our numerical simulations. We study then the cell discussion between yeast of opposite gender. In the second part we study the immune response in atherosclerosis. We build and then develop an age structured model in order to describe the inflammation. For specific parameters, we investigate the long time behaviour of our system by using a Lyapunov functional.