Integrated analysis of autoantibody profiles and clonal distribution of RAS mutations in juvenile myelomonocytic leukemia

Abstract Background: Juvenile myelomonocytic leukemia (JMML) is a rare and aggressive myelodysplastic/myeloproliferative neoplasm occurring in infancy and early childhood. It is characterized by excessive myelomonocytic cell proliferation and often co-occuring autoimmunity. More than 90% of patients harbor germline and somatic mutations in RAS pathway genes (PTPN11, NF1, KRAS, NRAS, and CBL). RAS-associated autoimmune lymphoproliferative disorder (RALD) shares molecular features with JMML but generally follows a more indolent clinical course and is considered a non-malignant condition. Distinguishing between these entities can be challenging due to overlapping genetic and clinical characteristics. To enhance our understanding of the molecular and immunological landscape of JMML, we conducted comprehensive autoantibody profiling, cytometry by time-of-flight (CyTOF), and single-cell DNA sequencing (scDNAseq) analyses. Methods: Autoantibody profiling targeting 13,352 recombinant human proteins (HuPEX protein array, Proteo-Bridge) was performed using diagnostic plasma samples of 55 patients and five healthy individuals. Peripheral blood mononuclear cells were analyzed by CyTOF using the Maxpar Direct Immune Profiling Assay (Fluidigm), which features a 30-marker antibody panel. To concurrently assess somatic mutations and the expression of 42 hematopoietic antigens, scDNAseq was performed on bone marrow mononuclear cells using the Tapestri platform (Mission Bio) with integration of the TotalSeq-D panel (BioLegend). Results: A total of 55 patients were clinically diagnosed with either JMML (n = 47) or Noonan syndrome–associated myeloproliferative disorder (NS/MPD; n = 8), which is caused by germline RAS pathway mutations. Among the 13,352 measurable autoantigens, the number of detectable autoantibodies per patient ranged from 10–563 (median 62). After final analysis, patients were stratified into three groups based on autoantibody count: low (L), intermediate (I) and high (H), with <62 autoantibodies (n = 28), 62-200 autoantibodies (n = 17), and >200 autoantibodies (n = 10), respectively. Group H showed a significantly higher prevalence of hypergammaglobulinemia (p = 0.027) and a universal presence of KRAS, NRAS, or CBL mutations (p = 0.022) compared to Group I/L. CyTOF analysis (n = 30) revealed that Group H/I (n = 13) had significantly elevated effector memory CD8+ T cells (2.36% vs 0.46%, p < 0.001) and reduced naive CD4+ T cells (1.64% vs 5.18%, p = 0.017) compared to Group L (n = 17). Next, to evaluate the distribution of RAS pathway mutations across hematopoietic lineages, we performed scDNAseq analysis on three samples derived from two patients in Group H/I and one in Group L. Group H/I patients had similar RAS pathway mutation burden in B-cell and myeloid compartments, establishing broad clonal expansion capacity. In contrast, in Group L patients, RAS mutations made only a minimal contribution to the B-cell compartment. Discussion: We identified an mechanistically distinct JMML subgroup, characterized by RAS-mutated B-cell clonal expansion linked to extensive autoantibody production. Notably, in Group H/I, RAS-mutated clones demonstrated clonal expansion into B-lineage cells at levels comparable to myeloid compartments, suggesting that immune dysregulation may contribute to the heterogeneity of JMML. Autoantibody burden serves as a biomarker of B-lineage clonal involvement, providing a potential diagnostic criterion to distinguish JMML from RALD based on clonal architecture.