Bone marrow immune and stromal niche remodeling in myelodysplastic neoplasm patients under erythropoietic therapies by 29-color full-spectrum flow cytometry

Abstract

Abstract INTRODUCTION Despite immune and stromal components of the bone marrow (BM) niche are increasingly recognized as key regulators in myelodysplastic neoplasms (MDS), their dynamic remodeling under therapy remains poorly characterized. Available data is scarce and often lacks single-cell resolution. To address this, we developed a full-spectrum cytometry panel to simultaneously characterize hematopoietic and non-hematopoietic BM populations in low-risk MDS (LR-MDS) patients under treatment. METHODS Cryopreserved BM mononuclear cells from three patients with LR-MDS with ring sideroblasts and multilineage dysplasia (MDS-RS-MLD) were analyzed at diagnosis and after 6 months of therapy. Patient 1 received an erythropoiesis-stimulating agent (ESA), while patients 2 and 3 received TGF-β ligand traps. An age-matched control BM aspirate from a lymphoma patient without BM infiltration was used as a non-MDS reference. A 29-color full-spectrum flow cytometry panel was designed to characterize hematopoietic (monocytes, macrophages, dendritic cells, and myeloid-derived suppressor cells) and non-hematopoietic (endothelial precursors, mesenchymal cells) populations. Panel optimization was based on antigen density, co-expressions, and spectral spread. Dead cells and lymphoid populations (CD3, CD7, CD19, CD56) were excluded using a viability dye and a dump channel. Staining was performed in three steps: surface markers, viability dye, and intracellular CD68 after permeabilization. Data was acquired on a 5-laser Aurora cytometer. RESULTS Baseline immune profiling of the non-hematopoietic compartment, revealed a relative increase (+65%) in endothelial progenitor cells (EPCs: CD45⁻ DUMP⁻ CD34⁺ CD133⁺ CD309⁺) among the three patients compared to the control, while primitive mesenchymal stromal cells (pMSCs: CD45⁻ DUMP⁻ CD34⁻ CD271⁺) and their functional phenotype (CD90⁺CD105⁺CD73⁺) were reduced (−52% and −51%, respectively). In the hematopoietic compartment, the granulocytic/monocytic myeloid derived suppressor cells ratio (MDSC-G/MDSC-M) was markedly decreased (−93%), and non-classical monocytes (CD14⁻ CD16⁺) were augmented (+261%). Total macrophages increased moderately (+16%), with divergent M1/M2 ratios across patients. Plasmacytoid dendritic cells (pDCs: CD45⁺ DUMP⁻ CD34⁻ CD14⁻ HLA-DR⁺ CD123⁺ CD11c⁻ CD303⁺) were consistently elevated (+810%) but showed reduced activation profile (CD86⁺, −61%). Conventional DC (cDC, CD45⁺ DUMP⁻ CD34⁻ HLA-DR⁺ CD123⁻/dim CD11c⁺) subsets exhibited heterogeneous changes, with robust cDC-type 2 expansion (+660% to +2715%) and slightly reduction in the activated fraction cDC-type 2 and cDC-type 3. All patients achieved complete hematologic response after 6 months of treatment, allowing analysis of intra-individual immune remodeling. Patients P2 and P3, treated with TGF-β ligand traps, exhibited greater modulation of the BM microenvironment compared to P1. Both showed a marked reduction in EPCs (−79% and −55%, respectively), along with significant increases in pMSCs (+23% and +1161%), and their functional phenotype (+390% and +232%). Both patients experienced reductions in classical monocytes (CD14⁺ CD16⁻) and M1/M2 ratios, alongside enhanced activation of macrophages, particularly M2/CD169⁺, suggesting a shift toward a supportive niche that may promote erythroid maturation and contribute to the reduction of ring sideroblasts. In P2, a reduction of MDSC-Ms (-92%) and increase in MDSC-Gs (+82%), led to a large rise in the MDSC-G/MDSC-M ratio (2.9 to 68.5, +2270%), indicating reduced immunosuppression and improved blast clearance. pDCs increased in all patients, potentially reflecting a compensatory immune mechanism. However, activation diverged: P2 showed increased pDC activation (+63%), while P3 showed reduced activation (−49%). Conventional dendritic cell subsets (cDC1–3) showed a general reduction in P2 and P3, with selective increases in activation. In contrast, P1, treated with ESA, displayed milder changes in MSCs and monocytes, with an increase in cDC-type 2 and cDC-type 3 populations but limited activation, and less pronounced modulation of the immunosuppressive microenvironment. CONCLUSIONSImmunological profiling of three MDS-RS-MLD patients in complete hematologic response revealed treatment-dependent microenvironmental remodeling. These findings suggest that treatment differentially influences immune architecture, with potential impact on clinical efficacy and therapeutic personalization.