2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Saturday, December 12, 2026
09:30 a.m. - 11:00 a.m.
MCCNO - Hall C
Technology and Procedures, Molecular Testing, Omics Technologies, Bioinformatics, Computational Biology, Gene Editing, Profiling, Biological Processes, Pathogenesis, Genomics, Hematopoiesis, Molecular Biology, Diseases, Myeloid Malignancies, Acute Myeloid Malignancies, Chronic Myeloid Malignancies, Research, Fundamental Science, AML, Translational Research, MDS, MPN, CMML, Study Population, Human, Adult, Elderly
PAPER

97 Spatial atlas of human MDS reveals mutant hematopoietic clones within immunosuppressive niches that evade immunity

Wenjie Cai, Honghao Bi, MD, PhD2, Pan Wang, PhD3, Inci Aydemir, MD4, Julia Galyayeva, BS5, Lisa Tao6, Peng Ji, MD, PhD7
1 Northwestern University, Chicago, Illinois, United States
2 Northwestern University, Chicago, Illinois, United States
3 Northwestern University, Chicago, Illinois, United States
4 Northwestern University, Chicago, Illinois, United States
5 Northwestern University, Chicago, Illinois, United States
6 Northwestern University, Chicago, Illinois, United States
7 Northwestern University, Chicago, Illinois, United States

Myelodysplastic syndromes (MDS) are genetically heterogeneous clonal hematopoietic malignancies characterized by ineffective hematopoiesis, bone marrow (BM) failure, and immune dysregulation. While recurrent genetic mutations drive clonal evolution and disease progression in MDS, how mutant clones remodel the native BM microenvironment and evade the immune surveillance remains poorly understood. Here, we generated the first mutation-resolved multimodal spatial atlas of human MDS by integrating Visium HD and Xenium 5K spatial transcriptomics with customized mutation-specific probes, enabling simultaneous mapping of tissue architecture, single-cell transcriptomes, mutation-defined hematopoietic clones, and their surrounding cellular ecosystems in intact human bone marrow.

The study included 38 human bone marrow specimens, comprising 5 normal controls, 9 MDS-LB, 6 MDS-IB1, 4 MDS-IB2, 5 MDS-del(5q), 5 MDS with biallelic TP53 inactivation, and 4 MDS-LB-SF3B1 samples. Visium HD profiling generated 19,643,421 8-μm bins, with a median of 1,105 sequencing reads and 44.1 UMIs per bin, detecting an average of 17,856 genes per specimen. Xenium profiling analyzed 7,837,732 individual cells, with a median of 99.9 transcripts per cell and 370.8 nuclear transcripts per 100 μm², providing mutation-resolved single-cell spatial maps across genetically diverse MDS subtypes.

This approach first allowed us to map the normal human bone marrow tissue architecture, validated by the literature. Compared to the normal marrow, integrated spatial transcriptomic analysis revealed widespread remodeling of bone marrow architecture in MDS, with altered spatial organization across hematopoietic, vascular, stromal, and immune compartments. Hematopoietic stem and progenitor cells (HSPCs) showed increased association with osteoblasts and decreased proximity to endothelial cells across multiple MDS subgroups, suggesting a shift in HSPC localization from perivascular niches toward endosteal regions.

To investigate how mutant clones remodel their local microenvironment, we next performed permutation-based spatial enrichment analysis centered on mutant cells (50 μm radius). Cellular neighborhoods surrounding mutant cells contained significantly higher proportions of CD8? T cells across multiple recurrent MDS driver mutations in genes including ASXL1, RUNX1, TP53, SF3B1, DNMT3A, and TET2. The strongest enrichment was observed in ASXL1-mutant MDS-IB1, where mutation-associated niches exhibited a 9.59-fold higher proportion of CD8? T cells than randomly generated neighborhoods (Benjamini–Hochberg-adjusted P < 0.001).

To determine the functional state of niche CD8? T cells, we compared niche and non-niche CD8? T cells across all MDS subgroups. Pathway analysis of genes upregulated in niche CD8? T cells demonstrated strong activation of transcriptional programs associated with chronic antigen stimulation, immune suppression, and progressive dysfunction, including PD-1 checkpoint, TGF-β, NF-κB, TNF, T-cell receptor, and apoptotic signaling, indicating a transcriptionally dysfunctional state. De novo niche identification using NicheCompass independently corroborated these findings by demonstrating robust enrichment of T-cell exhaustion-associated gene programs within mutant-cell niches. Consistent with this dysfunctional phenotype, ligand–receptor analysis identified recurrent checkpoint, immune synapse, cytokine, and death receptor interactions between mutant cells and niche CD8? T cells, including CD274–PDCD1, LGALS9–HAVCR2, LGALS9–CD47, PVR–CD96, TNFSF12–TRAF2, CD40–TRAF2, and CD40–FAS, suggesting persistent cell–cell communication that promotes CD8? T-cell dysfunction. In parallel, recurrent adhesion and retention ligand–receptor interactions, including CD55–ADGRE5, LAP–ITGB2, and RELN–ITGA2B, suggest mechanisms that promote the local retention of CD8? T cells within these niches.

Collectively, these findings support a multiscale model of bone marrow remodeling in MDS, in which tissue-wide architectural reorganization is coupled to localized, mutation-defined immune niches. Within these niches, mutant clones engage CD8? T cells through persistent cell–cell interactions that drive progressive dysfunction and local immune tolerance. Together, these data define a mutation-resolved spatial framework linking genotype, microenvironmental remodeling, and immune evasion in human MDS, and they nominate these niches as therapeutic targets for restoring antitumor immunity.

Wenjie Cai: No relevant conflicts of interest to declare.