Evans syndrome is defined by the occurence of autoimmune cytopenias, either at the same time or sequential, mainly autoimmune hemolytic anemia and immune thrombocytopenia. In children, it may be secondary to infections, systemic autoimmune disease, or primary immune deficiency, though in most patients, its etiology isn't obvious. Patients affected with Evans syndrome can also present other features, such as autoimmunity toward a particular organ, benign lymphoproliferation or immunodeficiency. The main goal of this work was to identify genetic causes in children presenting an Evans syndrome without a known underlying etiology. We focused our study on severe, early onset forms of the disease, with the hypothesis that a monogenic disease would be more frequent in this group of patients. Taking advantage of high throughput "Next Generation" sequencing (NGS) techniques, we sequenced and analyzed exome from patients and their relatives in search for adequate candidate genes. We identified 4 candidate genes: LRBA, CTLA-4, STAT3 (gain-of-function mutations), and NFKBA. Implication of the first 3 genes in new monogenic diseases with autoimmunity as a key feature was also confirmed by others during the course of this work. For each gene, we pursued 2 complementary goals: First, we sought to validate the implication of the gene in the patients' disease. To do so, we used various techniques and approaches: biochemistry and proteomics to identify protein partners, confocal microscopy to localize proteins and interactions, in vitro cellular assays to bring to light functional defect, flow cytometry to identify changes in lymphocytes subpopulations. We also looked for other mutations of each gene in patients with a similar clinical presentation. Hence we created and explored 3 cohorts of patients presenting with mutations of LRBA, CTLA-4 or STAT3. We constituted a cohort of 18 patients with LRBA mutations within 11 families. We then were able to precise and extend the clinical spectrum of this recently described disease. In particular, we observed patients with severe chronic arthritis associated with diabetes mellitus or enteropathies. We identified 15 new mutations of autosomal recessive transmission in the LRBA gene, coding a protein of unknown function, which absence is responsible for a disease mainly characterized by autoimmune features. We identified 29 candidate protein partners of LRBA and precized LRBA localisation in cell compartiments. We also established a cohort of 12 patients within 10 families presenting CTLA-4 haploinsufficiency. Beyond describing 9 new mutations, we report a family with autosomal recessive transmission.In LRBA and CTLA-4 deficiencies, we showed a decrease of regulatory T lymphocyte subset proportion among PBMC and a decrease of CTLA-4 expression in activated T cells. These results support the interaction between these 2 proteins, described concurrently by another team. We showed that the clinical spectra of these 2 diseases, although widely overlapping in first published reports, could be different despite a role of regulatory T cells in both. Hence, organ-specific autoimmunity and lymphoproliferation are more frequent in LRBA deficiency whereas granuloma and hypogammaglobulinemia are more present in CTLA-4 deficiency. Theses results suggests a role of genetic modifyers, which remain to identify. Among our cohort of patients with Evans syndrome, we also identified 5 patients within 5 families presenting gain-of-function mutations of STAT3. 3 of those mutations were reported by others during our work and appeared de novo in our patients. Functional validation of the 4th one is in progress. The last mutation follows a recessive transmission and could exemplify a new transmission modality of this disease. (...)