The slow cycling, location and hypoxia-resistance of hematopoietic stem cells are suggestive of a restricted metabolism. We propose that HSC metabolism is adapted to unique metabolic conditions supplied by the stem cell niche, and that a combination of the metabolic and signalling environments acts to support stem cell amplification and to limit it to a narrowly-defined and physiologically rare set of sites. To investigate this possibility, we have established moderate throughput, small scale cultures to examine the metabolic characteristics of primary CD133+ cells isolated from umbilical cord blood. A screen of carbon and energy sources revealed that pyruvate (but neither fatty acids nor amino acids) can replace glutamine as a major substrate. The fact that pyruvate contributes significantly to the cellular metabolism even in the presence of glucose suggests that CD133+ cells employ an unusually low level of glycolysis. Flow cytometric analysis of surface markers before and after culture confirmed that the addition of glucose (0mM, 5mM and 25mM) or insulin (0μg/mL, 4 μg/mL) increased the overall cell yield, but had no effect on the proliferation of early cells (CD133+, CD34+ and c-kit+). In parallel, metabolic profiling of undifferentiated and differentiated FDCPmix cells using gas chromatography and mass spectrometry techniques revealed an accumulation of glucose in the self-renewing (undifferentiated) population. Taken together, these observations suggest that glycolysis makes little contribution to stem cell metabolism, and that hematopoietic stem cells (as has been suggested for germ cells) may instead use glycolytic products supplied by stromal cells. Furthermore, our studies have revealed an unexpected effect of osmolarity on both glucose metabolism and self-renewal, in that an increase in osmolarity from 0,32 Osm/Kg to 0,36 Osm/kg reduced the rate of glucose-dependent proliferation of CD133+ cells without reducing the yield of early (CD133+ CD34+, ckit+) cells. Similarly, both the proportion of self-renewing cells in FDCPmix cultures and the recruitment of these cells to active self-renewal in semi-solid media in a colony forming assay were found to be increased at higher osmolarity. This suggests that high osmolarities suppreses both glycolytic metabolism and proliferation rate, but favour the maintenance of “early” progenitors. Finally, we have applied our assay to test a range of different growth factor combinations (of SCF, Flt3-ligand, TPO, IL3, IL6, IL7, IL11 and TGF-b) in 18% and 1% O2, and found that hypoxia extends markedly the range and magnitude of the proliferation response. Taken together, our results suggest that hematopoietic stem cells are indeed adapted to a rare metabolic microenvironment which includes (but is probably not limited to) low oxygen and glucose concentrations, and that the metabolic environment is likely to influence strongly the response to growth factors. A thorough understanding of stem cell metabolism may therefore provide a basis for more controlled manipulation of HSC in vitro.

Author notes

Corresponding author