Single-Cell Dissection of the B-ALL Niche Reveals Distinct Mesenchymal Stromal Cell Subtypes with Differential Leukemic Cell Support and Relapse Implications

Despite the advancements in treatment, relapse remains a major hurdle in B-cell acute lymphoblastic leukemia (B-ALL). The bone marrow microenvironment, particularly the stromal compartment, is implicated in B-ALL progression, yet its cellular heterogeneity and contribution to disease pathogenesis remain elusive. To elucidate these mechanisms, we employed single-cell RNA-sequencing to map the bone marrow landscape in B-ALL patients (n=9). Our analysis revealed 11 distinct cell populations, including leukemic cells, mesenchymal stromal cells (MSCs), and immune cell subsets. In-depth analysis of the MSC population revealed two distinct subtypes: MSC-1, with a stem cell-like signature, and MSC-2, characterized by adipogenic markers. Using a cell atlas of the human bone marrow stroma (Bandyopadhyay et al., Cell 2024), we mapped these to fibroblastic and adipogenic stromal cells, respectively. A FACS-sorting strategy based on our scRNA-seq data allowed us to isolate MSC-1 and MSC-2 populations. Functional characterization in terms of immunophenotype, colony-forming unit fibroblast (CFU-F) and multipotency in vitro confirmed the mesenchymal nature of both subtypes, with MSC-1 demonstrating higher proliferative capacity and MSC-2 displaying a stronger adipogenic potential. Subsequent single-cell derived cell-to-cell interaction analysis was conducted to delineate the communication dynamics in the leukemic bone marrow ecosystem. Our results revealed robust cell-cell interactions between leukemic cells and both MSC populations, which showed the highest expression of key niche factors (IL7, SCF and CXCL12). Co-culture experiments using patient-derived bone marrow mononuclear cells validated these findings, demonstrating that both MSC subtypes promote leukemic cell growth, with MSC-2 exhibiting a more pronounced supportive effect ex vivo. Further characterization revealed distinct niche-mediated support mechanisms. Notably, MSC-1 primarily interacts with leukemic cells through growth factor signaling (e.g., Osteopontin-CD44), while MSC-2 relies on extracellular matrix interactions (e.g., collagen type I - GP6), providing potential axes contributing to leukemia support, which may be leveraged for future therapeutic use. To determine the clinical relevance of our findings, we analyzed bulk RNA-seq data from an independent patient cohort (n=260). Our analysis revealed significant associations between distinct stromal cell patterns and specific molecular subtypes of B-ALL. Notably, we observed an enrichment of the MSC-2 signature in patients with intrachromosomal amplification of chromosome 21 (iAMP21). Furthermore, the MSC-2 transcriptomic fingerprint was more prevalent in patients experiencing relapse, suggesting its potential involvement in disease progression linked to specific ALL genomic aberrations. In conclusion, our study unveils the B-ALL microenvironment as a complex ecosystem governed by distinct stromal populations, offering crucial understanding of how leukemia-stromal cell interactions may contribute to disease progression and relapse. The findings hold potential for therapeutic exploitation to eventually improve patient outcomes in B-ALL.