Identification of Potent Quiescent Human Hematopoietic Stem Cells Using Mitochondrial Profile

Quiescence is the main property of the most potent hematopoietic stem cells (HSCs). Recent evidence suggests that mitochondrial activity may be implicated in the maintenance of stem cell quiescence. However, the potential function of mitochondria in the regulation of human HSCs remains largely unknown. To address this, we measured mitochondrial membrane potential (MMP) in subpopulations of human HSCs, using Tetramethylrhodamine, ethyl ester (TMRE) a cationic fluorescent dye sequestered by active mitochondria. We found that MMP profiles progressively shift towards lower levels in subpopulations with phenotypes of higher hematopoietic hierarchy. Subpopulation of CD34+CD38-CD45- cells within the 10% lowest MMP enriched for highly primitive CD90+ HSCs from 17.5% to 49.0% as compared to those within the 10% highest MMP (p=0.0049, n=5). These results are consistent with the murine HSC finding in our laboratory. To analyze the functional activity of HSCs with various MMP levels, subpopulations of CD90+ HSCs (CD34+CD38-CD45-CD90+) with the lowest and the highest 25% MMP were FACS-sorted as MMP-low and MMP-high HSCs respectively and subjected to long-term culture initiating cell (LTC-IC) limiting dilution assay. The results after 5 weeks revealed that the frequency of LTC-IC is much greater in MMP-low HSCs (1 in 7.85 cells) than in MMP-high HSCs (1 in 59.3 cells) (n=2). These results suggest that MMP-low HSCs maintain higher functional potential than MMP-high HSCs. The in vivo analysis of the MMP-low and MMP-high HSCs transplanted into NSG mice is ongoing (will be performed at 5 months). FACS-sorted MMP-low and ^high HSCs were subjected to cell cycle analysis by Pyronin Y/ Hoechst-33342 staining. We found that above 90% of both MMP-low and MMP-high HSCs were in a quiescence state (MMP-low: 4.48 %, MMP-high: 9.38% in G1; 0% in S/G2/M). Additionally, the expression of CDK6 that is associated with HSC activation was not detectable by confocal microscopy in either MMP-low or MMP-high HSCs (low: n=31, high: n=19). To identify potential distinct kinetic of cell cycle entry, FACS-sorted single HSCs were cultured in serum free medium (SFM), STEM SPAN, supplemented with cytokines (SCF 100ng/ml, TPO 50ng/ml, Flt3 50ng/ml). The occurrence of cell division in each well was monitored under microscopy every 12 hours for 6 days. At each time point the percentage of divided HSCs among total initial seeding HSCs were plotted and curve fitted to calculate the kinetics of cumulative first division (low: n=58, R²=0.9981; high: n=58, R²=0.9973). These analyses showed that MMP-low HSCs were delayed by 1.9 hrs as compared to MMP-high HSCs for cumulative 50% cells to complete the first division. The percentage of newly divided cells of first division at each time point was also plotted. Two waves of first division were revealed in both MMP-low and MMP-high HSCs. The peak of the first wave was delayed by 7 hours in MMP-low HSCs, while the second wave was delayed by 14 hours as compared to MMP-high HSCs. These results indicate that even though both MMP-low and MMP-high CD90+ HSCs are mostly in a quiescence state, upon cytokine exposure in vitro, MMP-high HSCs exist G₀ phase more rapidly than MMP-low HSCs. In agreement with this interpretation, MMP-high HSCs express significantly higher level of CDK6 when cultured for 34 hours in the presence of the same cytokines as described above (low: n=14, high n=35; p=0.006). We also investigated whether MMP-low and MMP-high CD90+ HSCs can be maintained in vitro in the absence of cytokines. Both populations were cultured in 96 well plates for 7 days in SFM without cytokines. The percentage of live cells was significantly higher in MMP-low HSCs cultured for 7 days (64.9%) as compared to MMP-high HSCs (44.2%) (p=0.03). Furthermore, morphological analysis by mitochondrial specific probe TOM20 showed that MMP-low contained more fragmented mitochondria as compared to MMP-high HSCs. Our findings suggest that mitochondrial activity may be implicated in the regulation of HSC quiescence. Altogether these results support the notion that human MMP-low CD90+ HSCs are molecularly distinct from MMP-high CD90+ HSCs and maintain quiescence in vitro to a higher degree than MMP-high CD90+ HSCs which are more primed for activation.