Background: Being a crucial part of the tumor microenvironment (TME) in solid tumors, tumor-associated macrophages are often associated with poor prognosis (Bruni et al. Nat Rev Cancer 2020). Yet, in acute myeloid leukemia (AML) the role of macrophages remains unclear. Aims: Here, we evaluated the impact of M2 macrophages in AML using a patient derived xenograft (PDX) model. Methods: To do so, we first injected 1x105 peripheral blood (PB) derived M2 macrophages into NSGS mice and next transplanted notoriously difficult to engraft primary Acute Promyelocytic Leukemia (APL) cells (1 x 106 cells - n=7 different patient samples) via the retro-orbital vein. Results: As a result, mice with co-injected human M2-macrophages developed full-blown leukemia with increased spleen weight in comparison with the control. Perhaps even more strikingly, ex vivo culture of APL and other favorable AML subtypes such as inv (16) or NPM1 mutant leukemic cells on M2-macrophages for 48h was sufficient to “train” these cells to engraft and induce fatal leukemia. Maintenance of self-renewal was shown in a secondary transplant assay and an enhanced frequency of leukemic stem cells was assessed by in vivo LTC-IC assays (LSC frequency: Control: 1/6x106 cells vs M2 pre-culture: 1/7.2x104 cells). To better understand the biological changes induced on leukemic blast when exposed to M2 macrophages, we performed an RNA sequencing analysis comparing AML/APL samples at diagnosis to cells that were “trained” (48h) on M2-macrophages or on MS5 mesenchymal bone marrow stromal cells, as a control. Gene ontology and gene set enrichment analysis on the genes up regulated upon M2 co-culture were significantly enriched for oxidative phosphorylation (OxPhos) signatures. Evaluation of functional respiration using seahorse measurements, confirmed the increase of oxygen consumption rate (OCR, basal and maximum) in primary AML/APL cells (n=7) after exposure to M2 macrophages compared to MS5. The increase of basal and maximum OCR suggested enhanced mitochondrial metabolism, which prompted us to determine whether macrophages, similar to MSC cells, can transfer mitochondria to primary AML cells (van der Vlist et al. Neuron 2022). Flow cytometry analysis revealed an efficient mitochondrial transfer from M2 macrophages to leukemic blasts, which was more efficient compared to mitochondrial exchange from MS5 cells. Treatment with the CPT1A inhibitor Etomoxir (50 µM), prevented the gain in functional respiration and enhanced proliferation of AML cells when those were co-cultured on M2-macrophages, while no changes were observed for MS5 co-cultures. These results suggest increased fatty acid oxidation to drive the OXPHO-like state in AML. Summary/Conclusion: Overall, we revealed that M2 macrophages can support leukemic growth of favorable AML subtypes that are commonly difficult to engraft in PDX models. Even an in vitro exposure to M2 macrophages suffices to alter the biology of AML cells transforming a non-engraftable cell into a cell with high leukemic potential. In vitro we show that AML blast cells cultured on M2 macrophage adapt a more OxPhos-like state linked to intrinsic changes associated with increased uptake of mitochondria. In summary, our study uncovers how the TME can contribute to leukemic transformation which provides alternative avenues for therapeutic interventions.