Dissecting the immune landscape in AML relapse versus complete remission post allo-HSCT
Relapse in AML patients after allogenic hematopoietic stem cell transplantation (allo-HSCT) remains a major challenge, often driven by poorly understood mechanisms of immune evasion that compromise the graft-versus-leukemia (GvL) effect. Previous bulk RNA-seq analyses of AML patients at diagnosis indicated enhanced expression of immunoregulatory molecules, particularly in RUNX1-mutated subtypes, suggesting pre-existing inhibitory programs that may contribute to post-transplant immune evasion. We hypothesize that interactions between residual AML blasts and donor immune cells post-transplant lead to immune dysfunction and GvL failure. This study aims to define the immune and transcriptional mechanisms associated with immune evasion and donor immune cell impairment in AML relapse post allo-HSCT.
Bone marrow samples collected three months post-allo-HSCT from 3 patients that will relapse within 12 months post-transplant (REL) and 3 long-term complete remission (CR) AML patients, carefully matched for RUNX1 mutation and minimal residual disease (MRD), were analyzed. Mononuclear cells (MNCs) were sorted into CD34⁺CD45⁺ HSPCs, CD33⁺CD117⁺ myeloid, CD56⁺ NK, and non-NK lymphoid fractions, including CD3⁺ T cells, and subjected to 10x single-cell RNA sequencing. In parallel, a portion of unsorted MNCs was analyzed by multicolor flow cytometry for immune phenotyping. Cellular composition, transcriptional profiles, and T cell exhaustion signatures were assessed, with particular focus on CD4⁺ T cell subsets enriched in CR patients.
Immune profiling revealed notable differences between REL and CR patients. CR patients exhibited higher frequencies of CD4⁺ T cells, alongside enrichment of monocyte and NK populations, suggesting a more balanced and potentially protective immune environment. In contrast, REL patients showed lower CD4⁺ T cell abundance, indicating early immune perturbation. Differential expression analysis of CD4⁺ T cells revealed upregulation of inflammatory genes such as IFNG and multiple inhibitory/exhaustion markers PDCD1, CTLA4, LAG3, HAVCR2, NR4A2, BATF in REL patients, reflecting functional impairment. CR CD4⁺ T cells maintained expression of homeostatic and survival-associated genes (BCL2, SELL). Gene set enrichment analysis confirmed enrichment of inflammation and stress pathways (TNFα–NFκB, interferon-γ response, IL2–STAT5, TGF-β signaling, apoptosis, and unfolded protein response) in REL CD4⁺ T cells. A curated T cell exhaustion score was significantly higher in REL CD4⁺ T cells, suggesting functional impairment.
CR patients exhibited higher CD4⁺ T cell abundance, potentially supporting effective immune surveillance. REL patients displayed inflamed and exhausted CD4⁺ T cells, indicative of early immune dysfunction that may compromise GvL activity. Together, these findings suggest that immune exhaustion in CD4⁺ T cells, along with altered myeloid and NK compartments, may contribute to relapse, whereas preserved CD4⁺ T cell function supports durable remission. Validation in larger cohorts is required, but these results provide early insights into immune evasion mechanisms and inform strategies to restore anti-leukemic immunity.