Parallel Detection of Single-Cells and Extracellular Vesicles from Bone Marrow and Peripheral Blood Reveals Asymmetric Immunometabolic Profile in Acute Myeloid Leukemia

Introduction. Acute myeloid leukemia (AML) is an aggressive disease with a high relapse rate. Alteration in cell metabolism and metabolic adaptation are crucial hallmarks in AML. Although several studies have addressed the metabolic profile of leukemic cells, the metabolic cross-modulation of immune cells according to their resident (bone marrow, BM) and circulating (peripheral blood, PB) status has been poorly investigated. Growing evidence has highlighted a key role for extracellular vesicles (EVs), shed from a broad variety of cells and cargo of crucial factors, in modulating several functions of leukemic cells, including bioenergetic metabolism. Again, few studies have addressed the impact of EVs in regulating the metabolism of immune cells from leukemic microenvironment. Therefore, in our work, we dissected the immunometabolic profile of immune cells and EV signature profile in parallel from BM versus PB from AML patients.

Methods. BM/PB samples and plasma were collected from AML patients at diagnosis (n=40). Leukemic EVs (herein referred to as AML EVs) were purified from platelet-poor plasma BM/PB by size exclusion chromatography and ultrafiltration. MACSPlex Exosome Kit was used for screening surface EV markers. Flow cytometry-based Single Cell ENergetic metabolism by profilIng Translation inhibition (SCENITH) method was applied in fresh whole blood. In addition, BM/PB mononuclear cells from AML patients were characterized by SCENITH after co-culture with circulating AML EVs.

Results. To explore the metabolic reprogramming within the leukemic immune microenvironment, we firstly applied the SCENITH method in multiple immune cell subpopulations from AML patients. In comparison with the paired PB, we observed a significant increase in glucose dependence of AML BM CD8⁺ T cells. By contrast, the capacity to use fatty acids and amino acids as sources for ATP production when glucose oxidation is inhibited was prominently reduced in AML BM CD8⁺ T cells and increased in AML PB CD8⁺ T cells. Also, mitochondrial dependence of CD8⁺ T cells was higher in the BM compartment, suggesting a switch from oxidative phosphorylation to glycolysis in circulation. Interestingly, in comparison to paired PB, only BM CD3⁺ and CD4⁺ T cells from female AML patients revealed a marked increase in their glucose dependence reporting gender-dependent metabolic differences. Towards clinically relevant implication, according to European LeukemiaNet (ELN) 2017 risk classification, SCENITH analysis showed a critical reduction in mitochondrial dependence at the expense of glycolytic capacity in circulating CD4⁺ T cells from favorable-risk AML compared to intermediate/adverse-risk patients.

Given the immunometabolic profile between BM and PB cell compartments, we hypothesized that EVs may differentially contribute to immunometabolism according to their BM versus PB localization. Preliminarily, we explored the immune profiling of plasma-derived EVs from AML BM versus PB. We found that BM EVs exhibited higher expressions of specific markers including CD4, CD40, ROR-1 and HLA-II/-ABC compared to PB equivalents. Of note, we observed a novel link between metabolic regulation and EV profiling. Indeed, the glucose dependence of PB CD8⁺ T cells was negatively correlated to the expression of CD133-1 (stem cell marker) and CD209 (dendritic marker) depicted on circulating (PB) EVs. Finally, we co-cultured AML circulating EVs with BM/PB mononuclear cells from AML patients in pairs to investigate the EV's role in metabolic regulation. Notably, BM CD8⁺ T cells in monocultures showed a reduction in glucose dependence compared to their PB equivalents. In parallel, mitochondrial dependence was more pronounced in PB CD4⁺ T cells compared to other circulating immune subsets. However, PB EVs caused a metabolic shift in the PB compartment towards high glycolytic capacity in CD8⁺ T cells and high mitochondrial dependency in CD34⁺ cells suggesting an EV-driven metabolic rewiring. Apparently, the co-cultures with PB EVs did not influence the energy profile of resident (BM) immune subsets.

Conclusions. Overall, this study shows the immunometabolic asymmetry orchestrated in AML at diagnosis, driven by distinct immune cell subsets according to their location (BM versus PB). We also describe the putative role of AML EVs in subverting the immune cell metabolism revealing critical implications for immunotherapy in AML.