Comprehensive profiling of the bone marrow microenvironment reveals an exhausted immune signature associated with tumor burden in Acute Myeloid Leukemia

Abstract Acute myeloid leukemia (AML) is a heterogeneous and aggressive disease characterized by the uncontrolled clonal expansion of hematopoietic progenitor cells. It has been described that the recruitment of inhibitory immune cells, and the dysfunction and depletion of effector immune cells at bone marrow (BM) niche induces the immune evasion of leukemia cells. The role of the immune system in AML progression and treatment response is under investigation, but a comprehensive phenotypic characterization of the immune microenvironment in the AML BM niche is still lacking. This study aimed to identify an “immune signature” related to tumour burden in the BM of AML patients. BM samples from 38 AML patients at diagnosis and 9 healthy donors (HD) were analyzed by multiparametric flow cytometry. We assessed the composition of selected key immune cell populations (CD4+, CD8+, regulatory T cells (Tregs), B cells, regulatory B cells (Bregs), natural killer cells (NK), monocytes, dendritic cells (DC) and myeloid-derived suppressor cells (MDSC)), along with their respective subtypes. Notably, the expression of a large array of specific immune checkpoints was also evaluated. Additionally, AML cells were quantified and phenotypically characterized based on CD34, CD117, CD33, CD13, HLA-DR. Aberrant lineage markers (CD7, CD56, CD25) expression used to define leukemia-associated immunophenotype (LAIP) in AML were also evaluated. To assess changes in the immune cell population composition of the AML BM niche and the expression of immune checkpoints after therapy, longitudinal analysis was performed in 4 refractory patients after immune- or chemo-therapy (after 1 cycle n=4; after 2 cycles n=2). At diagnosis, AML patients showed a trend toward increased naïve CD4+ and CD8+ T cells, and significantly higher proportions of CD8 effector and intermediate effector T cells, as well as naïve Tregs compared to HD. Plasmacytoid DCs were also expanded. Conversely, double negative B cells and Bregs were markedly reduced. Analysis of immune checkpoint expression revealed upregulation of TIM1, CD244 and TIGIT on naïve B cells, T cells and Tregs, respectively. A significant increase in the expression of LAG3 was also observed in the CD8+ T cells of patients. Furthermore, a trend of increased expression of the inhibitory immune checkpoint LOX1 was observed in early, polymorphonuclear and monocytic MDSC. Notably, AML blast burden (% AML cells) positively correlated with the expression of exhaustion-related checkpoints on immune cell subsets, including TIM3, LAG3, CTLA4, and NKG2D on NK cells; TIM3 and TIGIT on CD4+ T cells; LAG3, BTLA, and CTLA4 on CD8+ T cells; CTLA4 on Tregs; and CD1d on B and Breg cells. Distinct aberrant AML phenotypes also correlated with specific immune exhaustion profiles: CD56+ AML cells correlated with the expression of PD1, CD244, CD160, and TIGIT on CD4; CD244 and LAG3 on CD8+ T cells; LAG3 on Tregs; and TIGIT on B cells; CD7+ AML cells correlated with CD244 and TIM3 on CD4+ T cells and TIM3 on Tregs; CD25+ AML cells correlated with CTLA4 and TIM3 on CD8+ T cells. After therapy, refractory patients showed a trend toward a reduction of naïve CD4+ and CD8+ T cells and a significant decrease in double negative B cells and Bregs compared to baseline, along with an increase in effector Tregs. Our results demonstrate that AML is associated with the sequestration of naïve T cells within the BM and with a tumour burden-related upregulation of specific exhaustion-related immune checkpoints in T, B, regulatory subsets and MDSCs. Therefore, in AML the immune evasion of leukemic cells can be attributed, at least in part, to the expression of a large array of inhibitory checkpoint on key immune cell subtypes. Moreover, refractory disease appears to be characterized by a shift toward a more immunosuppressive immune profile. Overall, this study, pending validation in larger cohorts, provides a deeper understanding of AML immune-pathogenesis and has the potential to identify novel targets for immunotherapeutic strategies. Future investigations will clarify how the immune profile evolves in the context of measurable residual disease, which may further refine prognostic and therapeutic approaches.