Background: Relapse still represents an unsolved problem for children with acute myeloid leukemia (AML), therefore more effective therapeutic strategies are needed for a complete blasts eradication. Leukemic stem cells (LSCs) have been proposed as the therapy-resistant reservoir of cells responsible for failure of antiproliferative chemotherapy and disease recurrence in AML due to their self-renewal capacity and quiescence allowing evasion from chemotherapy. Thus, the identification of LSCs-specific vulnerabilities is needed for more efficacious treatments. Aims: We aim to dissect novel LSCs characteristics to unravel mechanisms of chemoresistance and disease recurrence, and characterize biological properties of LSCs that can be targeted to completely eradicate AML. Methods: To study LSCs derived from primary pediatric AML samples, we based on reactive oxygen species (ROS) content since it was shown that functionally defined LSCs have relatively low levels of ROS. We isolated the lowest 20% of cells fluorescing with CellROX probe (ROS low) and the highest 20% (ROS high) by cell sorting, and analyzed cell cycle, engraftment in mice and RNA sequencing to identify deregulated pathways. Results: After isolating ROS low and ROS high cells, we first assessed classical stemness features. By colony-forming assay, we observed that ROS low and ROS high cells have similar colony-forming capacity, but only ROS low cells preserve this capacity after a second replating whereas the ROS high population does not (p=0.002). By BrdU assay, we found that ROS low cells are relatively dormant, being more in G0/G1 cell cycle phase (41.9%) than the ROS high counterpart (7.9%, p=0.001), also confirmed by CFSE staining, revealing that ROS low cells proliferate more slowly (at day 4, 31% undivided cells in ROS low vs 8% in ROS high). According to our previously published data, we confirmed that ROS low cells express higher levels of CDK6-AS1 (RQ=1.38) and CDK6 genes (RQ=2.69; RQ=1 for ROS high) and that, being more quiescent, are characterized by lower levels of mitochondrial membrane potential by TMRE and lower content of mitochondrial ROS, byproducts of mitochondrial activity, by MitoSOX staining. We also verified that ROS low cells have a lower mitochondrial mass by MitoTracker and TOMM20 staining (mitochondria to nucleus ratio 1.18 vs 1.35 for ROS high, p=0.0005), suggesting that mitochondrial mass and functions might be involved in quiescent cells properties. However, engraftment experiments did not reveal different capacity of ROS low or ROS high cells in recreating leukemia in NSG mice, neither in terms of human CD45+ cells in mice peripheral blood (23.3% vs 41.9% respectively, p=0.12) nor bone marrow infiltration (73% vs 84.8%, p=0.44) at 8 weeks post-engraftment, opening for further dissection of the presence of LSCs within ROS high cells. By RNA sequencing, we found that the top 150 differentially expressed genes between ROS low and ROS high cells are involved in focal adhesion (logP=-13.51), regulation of hemopoiesis (logP=-12.31), cell response to stress (logP=-8.73) and cytokine production (logP=-8.67). Further analyses will confirm the role of these pathways in maintaining ROS low LSCs. Moreover, we identified 5 differentially expressed surface antigens that are under investigation as putative ROS low LSCs markers. Summary/Conclusion: In conclusion, the ROS low strategy permits to identify a quiescent AML subpopulation with peculiar mitochondrial features and transcriptome allowing to pinpoint novel molecular pathways that might be targeted to enhance LSCs clearance.
Read full abstract