Characterization of bone marrow B-cells in patients with warm autoimmune hemolytic anemia by single-cell analysis

Abstract Warm autoimmune hemolytic anemia (wAIHA) is a rare disorder caused by IgG autoantibodies targeting autologous erythrocytes. These antibodies bind specific epitopes on red blood cells, leading to extra-vascular hemolysis, primarily in the spleen. Standard treatment includes corticosteroids as first-line therapy and rituximab as second-line, with an efficacy exceeding 80%. However, responses are often transient, and relapse occurs in approximately 50% of patients within two years. The underlying mechanisms of relapse and progression to chronic disease remain poorly understood. To explore these aspects, we performed single-cell RNA sequencing (scRNA-seq) on bone marrow (BM) samples from n=9 wAIHA patients and n=4 healthy donors (HD) as part of the CYTOPAN protocol, approved by the local ethics committee and conducted in accordance with the Helsinki Declaration. Patient samples included five at diagnosis (under prednisone), three at remission (six months off therapy), and four at relapse following rituximab treatment. B-cell subpopulations were enriched in wAIHA patients compared to HD (15%, IQR 10–19 at diagnosis; 14%, IQR 10–15 at remission; versus 5%, IQR 3–5 in HD). Notably, mature B-cells were higher at diagnosis (72%, IQR 65–95 vs 42%, IQR 36–48 at remission and 37%, IQR 30–48 at relapse), whereas immature B-cells (pre-B and pro-B) were increased at remission and relapse (20%, IQR 17–22 and 28%, IQR 23–31 vs 19%, IQR 4–20 at diagnosis). Plasma cells (PC) were significantly more abundant at relapse. Longitudinal sampling of the three patients at remission further confirmed expansion of immature B-cell subsets. Functional analysis using GO and GSEA revealed stage-specific gene expression: immunoglobulin production, protein synthesis, and adaptive immune response genes were upregulated at diagnosis; cell-cycle and division-related genes were enriched at remission and relapse. These included pathways related to TNF-alpha signalling, estrogen/spermatogenesis, and E2F targets. Comparing remission to relapse highlighted an increased expression of TNF-alpha, IFN-alpha and gamma, apoptotic and hypoxic markers, KRAS, and inflammatory signals at relapse, while remission was marked by upregulation of microbial immunity and allograft rejection pathways, suggesting a more immature autoinflammatory profile at relapse. Marker gene analysis mirrored these findings, with upregulation of IGKV3D-11, IGHV3-74, IGHV3-7, CD27, IFNG-AS1, and FAM177B at diagnosis; and CD19, BTK, SYK, CD38, CD69, IGLL1, RAG1/2, and VPREB1/3 at remission/relapse. Inclusion of PCs reinforced the association with hypoxia, TNF-alpha, and glycolysis pathways at relapse, and confirmed absence of CD20 expression on these cells. Gene usage analysis, excluding IgM and IgD (consistently expressed across all disease stages), identified predominant usage of IGHG1 and IGHA1 constant regions and IGHV4-31 and IGHV4-34 variable genes at relapse. These preliminary findings suggest a shift toward an immature B-cell profile and accumulation of CD20-negative plasma cells in wAIHA patients at relapse post-rituximab, along with activation of genes involved in cell proliferation, TNF-alpha signaling, and CD38 expression, potentially revealing novel therapeutic targets.