A Deeply Quiescent Subset of CML LSC depend on FAO yet Avoid Deleterious ROS by Suppressing Mitochondrial Complex I

Background and Objective: Disease relapse and therapy resistance remain serious impediments to treating cancer. Leukemia stem cells (LSC) are therapy resistant and the cause of relapse. A state of deep quiescence appears to enable cancer stem cells (CSC) to acquire new somatic mutations essential for disease progression and therapy resistance. Both normal hematopoietic stem cells (HSC) and LSC share many common features, thereby complicating the safe elimination of LSC. A recent study demonstrated that long lived normal oocytes exist without mitochondrial complex I (MC-1), expressing it in a developmentally regulated fashion, thereby mitigating their vulnerability to ROS. Quiescent CSC rely on mitochondrial FAO, without complex I expression, thereby avoiding the generation of damaging ROS, similar to long lived normal human stem cells. A deeper understanding of the biology of therapy resistance is important for the development of optimal strategies to attain complete leukemia cures. Methods: Here, using scRNA-sequencing and ATAC-seq on primary chronic myelogenous leukemia (CML) patient samples, combined with bioinformatics analyses, we further examine the heterogeneity of a previously characterized in vitro imatinib-selected CD34-CD38- CML LSC population. We utilized a series of functional analyses, including single-cell metabolomic and Seahorse analyses, to validate the existence of the deepest quiescent leukemia initiators (LI) subset. Results: Current study revealed heterogeneity of therapy resistant LSC in CML patients and their existence of two functionally distinct states. The most deeply quiescent LI suppress the expression of MC-1, yet are highly dependent on fatty acid oxidation (FAO) for their metabolic requirements and ATAC-seq demonstrated increased chromatin accessibility in this population, all consistent with an extremely primitive, quiescent stemness transcriptional signature. Importantly, the specific CREB binding protein (CBP)/β-catenin antagonist ICG-001 initiates the differentiation of LSC, including LI, decreases chromatin accessibility with differentiation and increasing expression of MC-1, CD34, CD38 and BCR-ABL1, thereby re-sensitizing them to imatinib. Conclusion: We investigated the biological aspects related to LSC heterogeneity in CML patients and demonstrated the ability of specific small molecule CBP/β-catenin antagonists to safely eliminate deeply quiescent therapy resistant CSC. These observations may represent an attractive generalizable therapeutic strategy that could help develop better protocols to eradicate the quiescent LSC population.