Activated but Not Quiescent Hematopoietic Stem Cells Rely Readily on Glycolysis As Their Main Source of Energy

Elucidating mechanisms that regulate hematopoietic stem cells (HSCs) quiescence is critical for improving bone marrow transplantation. It is postulated that quiescent HSCs rely mostly on glycolysis rather than mitochondrial oxidative phosphorylation (OXPHOS) as their energy source. We have identified a population of HSCs with low mitochondrial activity (LSKCD150+CD48- within the 25% lowest mitochondrial membrane potential; MMP-low) that we show are highly enriched in quiescent HSCs (~95% ±2.65 in G₀) as measured by pyronin Y/Hoechst staining in contrast to LSKCD150+CD48- within 25% highest MMP (MMP-high) that are in majority in G₁/S/G₂/M phase of cell cycle (55.4%±16; p<0.05; n=3). MMP-low HSCs exhibit greater in vivo competitive repopulation ability (3.7 fold; p=0.021; n=10 mice) at 16 weeks as compared to MMP-high fractions, show higher self renewal ability and are enriched in label-retaining H2B-GFP⁺ cells (~3 fold; p<0.01; n=3). Conversely, label-retaining GFP⁺HSCs maintain lower MMP than non-label-retaining cells. Altogether these results reinforce the notion that MMP-low HSCs are quiescent whereas MMP-high HSCs are primed/activated. Using single cell RNA-sequencing (scRNA-seq) analysis to interrogate the transcriptome by a Fluidigm C1 platform within MMP-high versus MMP-low HSCs we found major metabolic pathways including OXPHOS, exhibit significantly greater expression within the MMP-high than in the MMP-low HSC fraction. Seahorse analysis confirmed that MMP-high but not -low hematopoietic stem and progenitor cells (HSPCs) use OXPHOS as their source of energy (3.9 fold; n=16 mice). Unexpectedly, glycolytic gene expression was also enriched in the primed MMP-high HSCs and relatively low in quiescent MMP-low HSCs. qRT-PCR analysis further confirmed that the expression of glycolytic enzymes and other genes including glucose transporter 1 (Glut1, slc2a1) that is the main glucose transporter on HSCs is greater in MMP-high relative to -low HSCs. These unforeseen findings raised the potential that despite the current consensus in HSC biology, glycolysis may more readily support activated rather than quiescent HSCs. We thus measured the glucose uptake in MMP-low vs -high HSCs under metabolic (pyruvate, glucose and glutamine)-restricted conditions. Using 2NBDG, a fluorescently tagged glucose analog, we found that MMP-high HSCs uptake 3.3-fold more glucose and contained three times more 2NBDG+ cells as compared to MMP-low HSCs (p<0.001 for each; n=3). Pharmacological inhibition of Glut1 reduced glucose uptake in MMP-high but not -low HSCs suggesting the specific sensitivity of MMP-high HSCs to the glucose inhibition. In addition, activation of tricarboxylic acid cycle TCA cycle with combined methyl-pyruvate and dimethyl-alpha-ketoglutarate led to a greater glucose uptake (~3.5 fold) in MMP-high as compared to MMP low HSCs (~2.2 fold; p<0.0001). To address the degree to which glycolysis is necessary, FACS-purified MMP-low and -high HSCs were incubated with 2-Deoxy D-Glucose (2DG) - a glucose analog - that inhibits glycolysis via its action on hexokinase. While glucose deprivation with 50 mM dose of 2DG within 12 hours did not have much effect on MMP-low HSCs, over 60% of MMP-high HSCs died (p<0.001; n=3), suggesting that MMP-high but not -low HSCs rely readily on glycolysis for their survival. Also, the inhibition of mitochondrial transport of pyruvate that is the end product of glycolysis (α-cyano-4-hydroxycinnamate (CHC)), decreased survival in MMP-high HSCs by 80% with negligible effect on MMP-low HSCs (p<0.01; n=3). Finally 50 FACS-purified HSCs (LSKCD150+CD48-MMP-low or -high) were transplanted in lethally irradiated mice along with 200,000 unfractionated bone marrow cells in a competitive repopulation assay and mice were subsequently treated with 2DG or control vehicle every other day for two weeks. Two months later, 2DG treatment led to significantly enhanced reconstitution levels in MMP-high HSCs with no or little effect on MMP low reconstitution levels (ongoing). These combined results suggest that primed MMP-high rather than quiescent MMP-low HSCs rely on glycolysis as their main source of energy. These findings are consistent with the concept that glycolysis is key in sustaining rapidly dividing cells such as embryonic stem cells and cancer cells.