In western countries, human concern for stress, welfare and health of livestock is growing up. In aquaculture, stress reduction is important from an ethical point of view but also results in an increased fish welfare and productivity. Stress response is quite complex phenomenon including release of several hormones. The most studied hormone during stress response in teleosts is cortisol, also called stress hormone. It is the most important corticosteroid in fish, modulating a vast range of physiological processes including general metabolism, growth, reproduction and immunity. Pathogen outbreaks are frequently reported in fish husbandry. Stress resulting from husbandry is suspected to favour these pathogen outbreaks. Furthermore, cortisol is reported to modulate all levels of the immune system; both cellular and humoral immune defences involved in both innate and adaptive immune systems. Most of the functions of corticosteroids are mediated through nuclear receptors called corticosteroid receptors. There are two types of corticosteroid receptors in teleosts as well as in mammals; the gluco- (GR) and mineralocorticoid receptor (MR). In mammals, as well as in teleosts, a bi-directional communication is reported between corticotropic axis and immune system. This bi-directional communication is essential to maintain efficient responses of both immune system and corticotropic axis, preventing adverse effects of too high or too long responses of those systems. This communication is well described in mammals, but several differences are reported between teleosts and mammals regarding corticotropic axis. Indeed, the mammalian mineralocorticoid aldosterone is not present at relevant physiological levels in teleosts. In 2005, 11-deoxycorticosterone (DOC) was identified to bind to and activate the mineralocorticoid receptor in several teleost species. This opens the door for a new putative active corticosteroid in teleosts. Furthermore, corticosteroid receptor diversity is higher in teleosts than in mammals, indicating that some differences in corticotropic response should exist between these two groups. Nowadays in teleosts, the bi-directional communication existing between both immune system and corticotropic axis is more frequently characterised by the corticosteroid-induced regulation of the immune system than by the regulation of the corticotropic axis by the immune response. Moreover, regulations of the immune system by other corticosteroids than cortisol, as well as the implication of the corticosteroid receptors in these regulations, are poorly documented in teleosts. Eurasian perch is an attractive model from different perspectives. Indeed, high basal levels of cortisol have been reported in this species. We also previously reported that one minute of exondation induced a high increase in cortisol plasma level, modulated some immune parameters in spleens and this immune modulation in spleen was accompanied by an increase in abundance of MR, and not of GR at the transcript levels. Additional in vitro experiments suggested that cortisol could not be the most important factor in the stress-induced modulation of the immune response in spleen. Several infectious outbreaks affecting Eurasian perch have been reported in aquaculture as well as in natural lakes, indicating the sensitivity of this species to some infectious agents. Based on this rationale, we chose to use Eurasian perch for better characterising the bi-directional communication between immune system and corticotropic axis in teleosts. The objective of the present thesis was to increase knowledge on the bi-directional communication between the immune system and corticotropic axis in a non-model species, the Eurasian perch. To reach this goal, we first sequenced and measured the constitutive expression of the different corticosteroid receptors in this species. We observed that Eurasian perch possess two GRs and one MR as most of the teleost species. These receptors are closely related to those of phylogenetically close species such as different Perciforms. An analysis of these receptors revealed that the hinge region, or D domain, of Eurasian perch GR-1 is shorter than that of all other teleost species analysed to date. This difference in the length of hinge region of GR-1 may suggest differences of conformational changes for this receptor after ligand binding. Secondly, we characterised the implication of two corticosteroids, cortisol and DOC in the immune response. During an in vivo experiment, we confirmed the role of cortisol as a strong modulator of cellular and humoral immunity, by increasing blood neutrophil populations, decreasing that of lymphocytes and increasing plasma lysozyme activity as well as the expression of C-type lysozyme in spleen. Moreover, the anti-inflammatory nature of cortisol was confirmed by the cortisol-induced decrease in the expression of the proinflammatory cytokine TNF-α. Moreover, we identified the implication of another corticosteroid in the modulation of the immune response in teleosts. Indeed, DOC, an MR ligand, increased the expression of C-type lysozyme and apolipoprotein A1 in both spleen and gills. Moreover, this corticosteroid was identified to favour its putative signalling by increasing MR and 11β-hydroxysteroid dehydrogenase type-2 (11β-HSD-2) expression. 11β-HSD-2 is an enzyme responsible for cortisol inactivation, which is essential to avoid occupation of MR by cortisol. An in vitro experiment was conducted to study the implication of each corticosteroid receptor in these regulations. This in vitro study indicated that cortisol induced its own inactivation by increasing the expression of 11β-HSD-2 in gills. We also identified the implication of GR, and not that of MR, in this process. The results collected during these in vivo and in vitro studies confirmed the implication of cortisol in immune regulation, as well as the implication of GR in some cortisol-mediated processes. But, we also reported a possible implication of another corticosteroid, namely DOC, in the interactions between corticotropic axis and immune system. To better characterise the bi-directional communication of the immune system and corticotropic axis, we challenged Eurasian perch with lipopolysaccharide (LPS), a component of Gram-negative bacteria cell walls. This component is well described to induce an immune response in teleosts. After having determined the dose of LPS necessary to induce an immune response in Eurasian perch, we challenged fish with this bacterial compound in association with either cortisol or DOC, the two corticosteroids previously reported to modulate the immune response. The results collected during this in vivo study indicate that LPS increased blood neutrophil population, decreased that of lymphocytes and increased plasma lysozyme activity. This immune response was associated with a modulation of the expression of all corticosteroid receptors in spleen. Moreover, cortisol was able to balance the LPS-induced modulation of TNF-α expression. All these results confirmed the close interaction existing between the immune system and the corticotropic axis in teleosts. To conclude, we confirmed by using different approaches the extensive bi-directional communication between immune system and corticotropic axis in a non-model species, the Eurasian perch. Moreover, we identified for the first time the putative implication of another corticosteroid, DOC, in this communication. The implication of this corticosteroid in the modulation of immune response suggests that the bi-directional communication between immune system and corticotropic axis in teleosts may be more complex than previously hypothesised.