Introduction: We recently proposed that B-cell depletion in immune thrombocytopenia (ITP) promotes the generation of long-lived plasma cells in the spleen, some of them being auto-reactive; but it remained possible that this observation was related to ITP itself rather than to B-cell depletion. Primary warm autoimmune hemolytic anemia (wAIHA) is a rare disease characterized by IgG auto-antibodies directed against antigens at the surface of red blood cells (RBCs) antigens, leading to their accelerated destruction. The use of B-cell depletion in wAIHA leads to 60 % to 70% of overall response at one-year and beyond. Nevertheless, 30-40 % of patients may resist to rituximab and then require a splenectomy. Nothing is known about antibody secreting cells (ASC) in the spleen of wAIHA patients, who have previously been treated or not with rituximab. In this study, we analyzed at the single cell level the splenic ASC from patients with chronic and active wAIHA, previously treated or not with rituximab (RTX), and we compared them with splenic ASC from ITP patients, and with splenic and bone marrow plasma cells from healthy donors (HD).Methods: We took advantage of the different therapeutic outcomes to analyze the splenic B-cell compartment of wAIHA patients, not previously treated with RTX (n=6), or after failure RTX treatment (n=3).Splenic tissues from organ donors and bone marrow from cardiovascular thoracotomy were used as controls. Blood samples from wAIHA (n=19), and (HD) (n=8) enrolled in this study were obtained after giving written informed consent in accordance with the Declaration of Helsinki.Results: We observed by flow cytometry and microscopy that the spleen from wAIHA patients who received less than 3 months of steroid therapy was the site of a B-cell response characterized by the presence of Bcl6+ germinal-center (GC) B-cells. Furthemore, splenic ASC secreted anti-red blood cell IgG in vitro. In line with this observation, we observed in the peripheral blood from patients with a newly diagnosed wAIHA (n=11), that short-lived IgG plasmablasts were increased compared with HD (n=8) (Mean 4.2 ± 0.84 % vs 0.99 ± 0.19% of CD19+ cells, p< 0.01). Moreover, for patients receiving long term steroid therapy (> 6 months) the plasmablast response was suppressed in the peripheral blood (Mean: 0.68 ± 0.2 % of CD19+ cells, n=8) and the splenic GC B-cell reaction was impaired (n=3). We conclude that short-lived IgG ASC result from an over-activity of GC reactions in wAIHA. We then analyzed the spleen of 3 patients who failed to respond to RTX, and observed a residual population of CD19+B-cells (median: 0.9% of lymphoid cells), including non-proliferative memory B-cells and plasma cells (PC). A fraction of these residual PC secreted anti-red blood cells IgG in vitro, thus accounting for the faillure of the B-cell depletion therapy. By using a single cell multiplex quantitative RT-PCR (Fluidigm dynamic arrays), we showed that such RTX-resistant plasma cells display a long-lived transcriptional program, which differs from PC from untreated wAIHA patients or HD, as well as from plasmablasts. Interestingly, the gene expression profile of wAIHA long-lived plasma cells segregated with long-lived PC previously observed in the spleen of ITP patients treated with rituximab. By a principal component analysis, we observed a gradient of maturation from plasmablasts to bone marrow plasma cells in which PC from RTX-treated spleens segregated close to bone marrow PC. We also observed that the cytokine BAFF was increased in the supernatants of spleen cell cultures from wAIHA patients treated with rituximab compared with controls (p< 0.05), suggesting, in keeping with our previous report in ITP, a role for BAFF in the differentiation of short-lived plasma cells into long-lived plasma cells.Conclusion: The presence of splenic long-lived autoreactive PC in wAIHA may explain why some patients cannot achieve a response after RTX. Our results show that, the B-cell depletion induced by rituximab itself, as opposed to the underlying auto-immune condition, promotes a suitable environment for the maturation of auto-immune long-lived plasma cells in the spleen. Targeting specifically some factors such as BAFF right after rituximab injection could be an interesting therapeutic option in the future. DisclosuresNo relevant conflicts of interest to declare.