Nature | Vol 578 | 13 February 2020 | 281extravascular pO 2 in the bone marrow (Fig. 2e). Thus, although we
detected regions in the bone marrow with pO 2 as low as 10 mm Hg,
HSPCs and MFG-HSCs were not found in these regions, suggesting that
localization to the regions of deepest hypoxia is not a prerequisite for
maintenance of MFG-HSC quiescence.Heterogeneous HSC response to activation
We next examined the dynamic behaviours of HSPCs and LT-HSCs
in their native niches. In vivo time-lapse imaging of the calvarium
revealed that MFG-HSCs displayed low baseline motility whereas
HSPCs showed enhanced motility (Fig. 3a, b, Supplementary Vid-
eos 3–6). To assess whether HSC behaviours would be affected in the
context of activation, we used a cyclophosphamide (Cy)/G-CSF proto-
col that leads to expansion and subsequent mobilization of LT-HSCs^28
(Extended Data Fig. 7a). Fluorescence-activated cell sorting (FACS)
and imaging analysis demonstrated a tenfold increase in the num-
ber of MFG cells after treatment (Fig. 3c–e, Supplementary Video 7,
Extended Data Fig. 7b, c). MFG cells are still enriched in the phenotypi-
cal LT-HSC fraction in this activated state (Extended Data Fig. 7b) and
display a concomitant increase in cell cycle activity (Extended Data
Fig. 7d). Time-lapse imaging of treated animals for about 6 h after the
third dose of G-CSF (Fig. 3c, Supplementary Video 8) showed that, on
average, MFG-HSCs became significantly more motile (P = 0.0001),
with displacement measurements even higher than those of steady-
state HSPCs (Extended Data Fig. 7e, Supplementary Video 8), but the
response was heterogeneous, ranging from cells displaying limited
displacement to a small number of cells that fully exited the bone
marrow into the blood stream (Supplementary Video 8, Extended Data
Table 1). Similarly, following treatment with the myeloablative agent5-fluorouracil (5-FU), the number of MFG+ cells increased (Extended
Data Fig. 8a, Supplementary Videos 9, 11) and a subset exhibited higher
motility, particularly on day 20 after treatment (Extended Data Fig. 8b,
Supplementary Videos 10, 12). These data suggest that enhanced motil-
ity is a common feature of the HSC response to injury, although we
cannot rule out the possibility that the response is a result of indirect
action on the niche by Cy/G-CSF or 5-FU.
Activated MFG-HSCs were found, on average, to be further away
from the endosteum than native MFG-HSCs (Extended Data Fig. 7f ).
Notably, they were even closer to the vasculature, with an average
distance of about 1 μm (Extended Data Fig. 7g), and maintained their
sinusoidal proximity (Fig. 3f). By assessing the distribution of acti-
vated MFG-HSCs in the entire calvarial region, we identified unique
patterns of HSC proliferation. First, native MFG-HSCs were found as
rare single cells within the bone marrow, whereas activated MFG-HSCs
appeared as clusters (Fig. 3c–e), suggesting that that MFG-HSC prolif-
eration occurs within spatially restricted domains. Second, a subset
of MFG+ cells remained as single cells while others formed clusters in
both the Cy/GCSF and 5-FU models (Fig. 3e, Extended Data Fig. 8c),
suggesting that the proliferative response is heterogeneous among
HSCs. To assess whether these clusters were clonal, we generated
Mds1CreER/+Rosa26Confetti/+ mice^29 (Extended Data Fig. 8d–f ). In untreated
mice, labelled cells were usually found as rare single cells of differ-
ent colours dispersed throughout the bone marrow (Extended Data
Fig. 8g). Following treatment with Cy/GCSF, we observed labelled cell
clusters made up predominantly of cells of a single colour (Extended
Data Fig. 8g, h). Quantitative analysis confirmed that labelled nearby
cells were more likely to have the same colour than a mixture of colours
(Extended Data Fig. 8h), providing evidence for clonal HSPC prolifera-
tion within confined physical domains.Y-displacement (μm)–20–15–10–505101520X-displacement (μm)–20 –15 –10 –5 0510 15 20a cControl Coronal suture Cy/GCSFSagittal suturefbMds1GFPFlt3CreCy/GCSF***
P = 0.0002Mds1GFPMds1GFPFlt3Cre020406080100Displacement(μm) over 2.5 hdeArteriole
Transitional
17% Sinusoid83%50 μmIdentity of nearest vesselFig. 3 | Increased motility, expansion, and localization of activated MFG-HSC
s. a, In vivo motility measurements of HSPCs (n = 12 cells) and MFG-HSCs
(n = 16 cells) at steady-state over a 2.5-h imaging period. Red bars, mean. P value
calculated using two-tailed Mann–Whitney test. b, Cell tracks for 16 MFG-HSCs
over a 2.5-h imaging period. Images were acquired every 30 min.
c, Representative intravital image of a Cy/GCSF-treated MFG mouse 4.5 days
after the beginning of treatment. MFG cells (green), vasculature (red,
Angiosense 680EX), auto-f luorescence (blue). Arrows, GFP+ cells. The
experiment was performed four times with similar results. d, e, Graphical map
of the locations of MFG-HSCs in the calvaria of untreated and Cy/GCSF-treated
mice (n = 3 and 4, respectively). Location data from individual mice are
indicated by different colours. f, Identity of nearest vessel for each MFG+ cell
(n = 12 cells) after treatment with Cy/GCSF. Compare to untreated mice in
Fig. 2d.