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

Paper Details

Abstracts, authors, affiliations, and disclosures.

Sunday, December 13, 2026
09:30 a.m. - 11:00 a.m.
MCCNO - 353-355
Technology and Procedures, Assays, Omics Technologies, Bioinformatics, Computational Biology, Treatment Considerations, Supportive Care, Biological Therapies, Transplantation, Biological Processes, Genomics, Clinical Practice (Health Services and Quality), Diseases, Immunology, GVHD, Myeloid Malignancies, Acute Myeloid Malignancies, Measurable Residual Disease, Research, AML, Translational Research, Clinical Research, Clinical Trials, Health Outcomes Research, Registries, Study Population, Human, Adult, Elderly, Pediatric
PAPER

574 Single-cell proximity network analysis reveals decoupling of physical leukemia-donor CD8+ T cell contact from checkpoint exhaustion at post-transplant relapse in acute myeloid leukemia

Amirthagowri Ambalavanan, PhD, Oyeronke Ayansola, PhD2, Azin Keshavarz, MHSc.3, Danielle Pyne, BSc4, Igor Novitzky-Basso5, Jonas Mattsson, MD PhD6, Ben Wang7, Dennis Dong Hwan Kim, MD, PhD8
1 Hans Messner Allogenic Transplant Program, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada
2 Hans Messner Allogenic Transplant Program, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada
3 Hans Messner Allogenic Transplant Program, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada
4 Hans Messner Allogenic Transplant Program, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada
5 Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada, Hans Messner Allogenic Transplant Program, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada
6 Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada, Hans Messner Allogenic Transplant Program, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada
7 Tumor Immunotherapy Program, Princess Margaret Cancer Centre, Toronto, Ontario, Canada
8 Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada, Hans Messner Allogenic Transplant Program, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada, Institute for Medical Science, University of Toronto, Toronto, Ontario, Canada, Leukemia Tissue Bank, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada, Research Program, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada

Background:

Post-transplant relapse (PTR) remains a major cause of treatment failure in acute myeloid leukemia (AML). Immune escape of leukemic blasts drives graft-versus-leukemia (GvL) failure, linked to blast-intrinsic changes and to donor T-cell dysfunction. However, the direct physical interaction between relapsing blasts and T cells, and its consequences on the T-cell surface, remains poorly defined. While bulk and single-cell transcriptomics fail to capture surface protein organization or intercellular membrane exchange on the cells, the Proximity Network Assay (PNA; Pixelgen®) enables highly multiplexed assay on protein abundance, same-protein clustering, and pairwise protein proximity at single-cell resolution. As a trogocytosis-based method, PNA resolves lymphocytes conjugated to other cells via immune synapse, such as leukemic blasts, and per-cell surface protein abundance and co-localization, enabling detection of leukemia-associated membrane material physically acquired by T cells and of the synapse/checkpoint architecture. This defines how relapsing leukemic blasts engage the donor CD8⁺ T-cell compartment.

Patients and Methods:

A cohort of 22 AML samples, enriched for PTR cases, was profiled using a 155-plex PNA panel. Bulk BM/PB mononuclear cells underwent rolling circle amplification, proximity ligation, and library preparation followed by sequencing. In this preliminary analysis, 11 samples had completed analysis (10 PTR samples and 1 diagnostic sample), while the remainder are still under.

Per-cell surface protein abundance and pairwise molecular co-localization (log₂ observed/expected proximity) were analyzed using the Pixelator and PixelatorR software. Leukemic components were defined by prespecified protein modules including CD34, CD117, C123, CD13, CD64, and CD371, while T-cell states were summarized using modules representing less-differentiated states (SLAMF6, CD28, CD27, CD127); multi-checkpoint dysfunction (PD-1, TIM-3, CD39, TIGIT, CTLA-4, VISTA, 2B4); activation; cytotoxicity; and terminal differentiation. Proximity module scores were computed for differentiation, checkpoint (PD-1/TIM-3/CD39/TIGIT), and synapse-adhesion programs. AML-associated membrane patches on T cells were identified using a patch analysis pipeline.

Results:

Out of 11,803 cells from 11 samples, 10,561 passed QC and were analyzed. Major populations included B-, CD4 T-, CD8 T-, NK-cells, dendritic cells, monocytes, and AML blast-like cells. Compared with the diagnostic sample, PTR samples showed a differentiated, adhesion-enriched CD8 T-cell phenotype. KLRG1, ICAM-1, CD11a, CD18, and NKG2D abundance was higher in all 10 PTR samples, with median relapse-minus-diagnosis differences of +1.51, +0.63, +0.58, +0.55, and +0.41 CLR units, respectively. Conversely, CD38 abundance was lower across all T cells in all 10 (median -1.63 CLR units).

PTR-enriched analysis identified recurrent, non-random organization of T-cell adhesion and differentiation proteins. Across all T cells, CD11a, CD18, CD58, and CX3CR1 self-clustering was positive in all 10 PTR specimens, with median log₂ observed-to-expected ratios of 0.17, 0.16, 0.37, and 0.47 (prespecified self-clustering analysis family; P=0.007).

Patch analysis revealed AML-associated protein patches predominantly within CD8 T-cell PNA graphs, varying markedly among PTR samples; >70% of analyzed CD8 T cells contained a patch. Patch regions were compositionally distinct from surrounding receiver regions and contained heterogeneous AML-associated markers, including CD34, CD13, CD123, CD371 and CD123 without a dominant PD-L1/PD-L2/VISTA phenotype, whereas receiver regions retained CD3ε, TCRαβ, CD2 and CD8. The combined AML-marker set was enriched within patches (median log₂ patch/receiver enrichment=+6.43; P=0.022), while the T-cell receiver-marker set was enriched outside patches (median log₂ patch/receiver enrichment=-4.15; P=0.022).

Canonical checkpoints showed no coordinated multi-checkpoint high pattern. The dominant relapse phenotype featured differentiated, adhesion-enriched T cells and altered immune-synapse architecture, not uniform checkpoint upregulation. Patch burden was unrelated to CD8⁺ functional/exhaustion modules, indicating that acute physical AML contact is decoupled from the standing T-cell exhaustion state.

Conclusion:

PNA shows that donor CD8⁺ T cells acquire AML membrane patches and shift to an adhesion-enriched, differentiated state distinct from checkpoint exhaustion at PTR.
Dennis Dong Hwan Kim, MD, PhD: Novartis - Research Funding; Enliven - Research Funding; Abbvie - Research Funding; Incyte - Research Funding; Sanofi - Research Funding; Ascentage - Research Funding; JAZZ - Consultancy (Includes expert testimony); Abbvie - Consultancy (Includes expert testimony); Ascentage - Consultancy (Includes expert testimony); Daiichi-Sankyo - Consultancy (Includes expert testimony); Novartis - Consultancy (Includes expert testimony); Sanofi - Consultancy (Includes expert testimony); J&J - Consultancy (Includes expert testimony); Astellas - Consultancy (Includes expert testimony)