There is evidence that hematopoietic stem cells (HSC) are physically localized to specialized areas in the bone marrow termed the vascular and osteoblast niches. It is not clear if there are differences in the capacity of these niches to support HSC function. We and others previously showed that G-CSF treatment suppresses both osteoblast number and function, effectively eliminating the osteoblast niche. In contrast, G-CSF treatment has no apparent effect on the microvasculature in the bone marrow, suggesting that the vascular niche is intact. In this study, we utilized this system to assess the capacity of each niche to support HSC function. We previously reported that the competitive repopulation capacity of bone marrow isolated from G-CSF treated mice is markedly reduced. This is not due to a simple loss of HSC in the bone marrow, as the number of HSC, phenotypically defined as lineage-CD41-CD48-CD150+ (SLAM) or lineage-Kit+Sca+CD34- cells, was comparable to control mice. Moreover, the long-term repopulating activity of sorted SLAM cells from G-CSF treated mice was reduced. This repopulating defect is not secondary to impaired homing to the bone marrow, as direct intrafemoral injection of G-CSF treated bone marrow cells failed to rescue the engraftment defect. Since G-CSF is able to stimulate HSC proliferation, we predicted that the defect in repopulating activity might be secondary to loss of HSC quiescence. Contrary to our prediction, the percentage of quiescent HSC in the bone marrow was actually increased in G-CSF treated mice. Whereas 28.0 ± 3.4% of control SLAM cells were labeled after treatment with BrdU for 48 hours, only 7.5 ± 0.8% of SLAM cells isolated from G-CSF mice were labeled (p < 0.008). Moreover, the percentage of SLAM cells in G0, as determined by Hoechst and pyronin staining, was increased in G-CSF treated mice (80.3 ± 5.0% versus 65.5 ± 6.8% in untreated mice, p=0.104). To elucidate the molecular mechanisms by which disruption of the osteoblast niche leads to a loss of HSC activity, we performed RNA profiling experiments on SLAM cells sorted from G-CSF or saline-treated mice. Consistent with the quiescent phenotype, a significant increase in the expression of the cell cycle inhibitor, Cdkn1a (p21waf1), was observed in G-CSF treated SLAM cells. Collectively, these data show that the osteoblast and vascular niches are not functionally redundant and suggest that it is the osteoblast niche that is key to maintaining long-term repopulating activity of HSC.

Author notes

Disclosure: No relevant conflicts of interest to declare.