TECHNICAL REPORT
Quantitative imaging of haematopoietic stem and
progenitor cell localization and hypoxic status
in the bone marrow microenvironment
César Nombela-Arrieta1,5 , Gregory Pivarnik1 , Beatrice Winkel1 , Kimberly J. Canty1 , Brendan Harley2 ,
John E. Mahoney3 , Shin-Young Park1 , Jiayun Lu1 , Alexei Protopopov3,4 and Leslie E. Silberstein1,5
The existence of a haematopoietic stem cell niche as a spatially
confined regulatory entity relies on the notion that
haematopoietic stem and progenitor cells (HSPCs) are
strategically positioned in unique bone marrow
microenvironments with defined anatomical and functional
features. Here, we employ a powerful imaging cytometry
platform to perform a comprehensive quantitative analysis of
HSPC distribution in bone marrow cavities of femoral bones.
We find that HSPCs preferentially localize in endosteal zones,
where most closely interact with sinusoidal and non-sinusoidal
bone marrow microvessels, which form a distinctive circulatory
system. In situ tissue analysis reveals that HSPCs exhibit a
hypoxic profile, defined by strong retention of pimonidazole and
expression of HIF-1α, regardless of localization throughout the
bone marrow, adjacency to vascular structures or cell-cycle
status. These studies argue that the characteristic hypoxic
state of HSPCs is not solely the result of a minimally
oxygenated niche but may be partially regulated by cell-specific
mechanisms.
The bone marrow cavities of long bones are the principal sites of
postnatal haematopoiesis, which is sustained by a rare population of
HSPCs (ref. 1). As for other well-defined adult stem cell types, HSPCs
have been proposed to reside in defined anatomical locations, where
they receive and integrate regulatory cues from neighbouring cells,
extracellular matrix components and/or soluble factors2–4 . The precise
definition of the physical localization and physiological features of
HSPC niches has been greatly hampered by the technical difficulties
associated with imaging long bones, the need for complex cell-surface
marker combinations to track rare and dispersed HSPC populations,
and the lack of tools required for the automated quantitative
microscopic analysis of large-scale specimens at a single-cell level.
Previous attempts to visualize HSPCs in their native context
provided relevant information but were limited to the observation
of relatively low numbers of events and often led to controversial
views on the compartmentalization of HSPC niches in the bone
marrow (ref. 5). Analysis of the distribution patterns of ex vivo purified,
transplanted HSPCs or long-term DNA-label-retaining cells suggested
that HSPCs preferentially interact with bone-lining osteoblasts4,6,7 . An
alternative view of HSPC localization was offered by studies visualizing
endogenous HSPC-enriched populations in immunostained bone
marrow tissue sections, which revealed that most HSPCs reside in
bone-distal regions, in direct contact with bone marrow sinusoids and
stromal perisinusoidal populations with mesenchymal stem cell and
osteoprogenitor potential8–12 . Nonetheless, a comprehensive analysis
of the global distribution of phenotypically defined endogenous HSPC
populations in the context of entire bone marrow cavities has not
been attempted so far.
A key niche-related feature of HSPCs is their recently reported
hypoxic profile2,13–15 , which has been described on the basis of
two lines of experimental evidence. First, HSPCs exhibit enhanced
incorporation of pimonidazole (Pimo), the most widely studied
hypoxic marker, which selectively forms adducts with proteins in cells
under low-oxygen conditions16 . Second, HSPCs stably express the α
subunit of hypoxia-inducible transcription factor 1 (HIF-1α; ref. 15),
which normally undergoes degradation by the proteasome when
oxygen levels exceed 5% (refs 17,18). These experimental observations,
together with bone marrow perfusion assays19 , have inspired a model by
1
Division of Transfusion Medicine, Department of Laboratory Medicine, Children’s Hospital Boston, 300 Longwood Avenue, Boston, Massachusetts 02115, USA.
Department for Chemical and Biomolecular Engineering, University of Illinois, Urbana Champaign, Illinois 61801, USA. 3 Division of Hematologic Malignancies,
Dana-Farber Cancer Institute, Boston, Massachusetts 02115, USA. 4 Present address: Institute for Applied Cancer Science, The University of Texas MD Anderson
Cancer Center, Texas 77030, USA.
5
Correspondence should be addressed to C.N-A. or L.E.S. (e-mail: [email protected] or [email protected])
2
Received 4 April 2012; accepted 12 March 2013; published online 28 April 2013; DOI: 10.1038/ncb2730
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
533
TECHNICAL REPORT
which HSPCs localize in areas of the bone marrow with minimal oxygen
content, at a certain distance from vascular structures; a condition
previously attributed to endosteal regions18,20 . Adaptation to hypoxia
is thought to determine the remodelling of the metabolic profile and
induction of quiescence in HPSCs (refs 15,21). Despite the fundamental
physiological implications of this model, evidence demonstrating that
defined poorly oxygenated bone marrow domains are enriched in
hypoxic HSPCs remains indirect and inconclusive so far.
Here we apply two complementary imaging approaches to perform
a comprehensive mapping of the spatial distribution of HSPCs in
the bone marrow and analyse their relationship to bone surfaces, as
well as to a variety of distinct bone marrow vascular structures, of
which we deliver a detailed three-dimensional (3D) characterization.
Finally, we exploit these technologies to demonstrate that the hypoxic
profile of HSPCs, based on Pimo incorporation and HIF-1α expression,
is unrelated to anatomical positioning in defined bone marrow
microenvironments as well as to proximity to vascular structures and
cell-cycle progression.
RESULTS
Global distribution of c-kit+ progenitors in longitudinal bone
marrow tissue sections
We adapted the use of laser scanning cytometry (LSC), a technological platform, which enables quantitative imaging cytometry of
fluorescently labelled discrete cell subsets within tissue sections22 ,
for the analysis of HSPC distribution in the context of whole
longitudinal murine bone marrow femoral sections. Cryopreserved,
non-decalcified, 5-µm-thick sections were systematically scanned
using monochromatic laser light excitation, to generate a sequence of
high-magnification fluorescent digital images that were assembled into
a composite high-resolution image of the entire bone marrow section
(Supplementary Fig. S1a). Individual cells were defined and quantified
through software-based automatic segmentation of DAPI+ nuclei
(Supplementary Fig. S1a, lower right panels). Positional information
and emitted fluorescent signals in different channels were recorded on
a per cell basis, allowing data representation and analysis in the form of
tissue maps and scattergrams or histograms (Supplementary Fig. S1b).
We initially employed LSC to determine the spatial distribution of
a cellular subset enriched in haematopoietic progenitor cell content
and defined by the expression of c-kit, the receptor for stem cell
factor (SCF). As explained in the Methods section, autofluorescent
cells were excluded from this analysis and isotype control stained
sections were systematically scanned to determine base-line levels of
fluorescence and define specific gates (Fig. 1a,b). c-kit+ cells were
detected, labelled in high-resolution images throughout the bone
marrow cavity (Fig. 1c) and plotted in tissue maps (Fig. 1d,e). To avoid
marked differences in cellular content derived from the dissection
plane, we restricted our analysis to bone marrow femoral sections
in which both proximal and distal metaphysis were exposed and the
diameter of the diaphysis was >800 µm. Quantitative analysis revealed
no significant differences in the frequencies of c-kit+ progenitor cells in
areas of trabecular (metaphysis) and cortical (diaphysis) bone marrow
of femurs (Fig. 1d). The presence of c-kit+ cells along the width of
the diaphyseal compartment was investigated by dividing the cavity
into eight longitudinal regions (100–150 µm wide each) spanning the
diameter of the diaphysis. This analysis uncovered a gradient in the
534
relative content of haematopoietic progenitors in different regions of
the diaphysis; highest percentages of c kit+ progenitors were found
in bone-proximal endosteal regions (5.59 ± 0.79%), progressively
decreasing towards most central areas (2.68±0.47%; Fig. 1e).
Given the crucial role of bone marrow microvasculature in the
regulation of early haematopoietic events23,24 , we investigated the
spatial relationship of c-kit+ progenitors and laminin+ microvessels.
Initial observations revealed a strong tendency of c-kit+ progenitors to
closely associate with bone marrow vascular structures (Fig. 1f), which
was verified in 3D reconstructions obtained by confocal imaging of
immunostained thick slices of femoral bone marrow (Supplementary
Fig. S2 and Video S1). We next employed software-based segmentation
of laminin+ vascular structures to create peripheral contours in
which the perimeter of vessel walls was expanded by up to 10 µm
(Fig. 1g). This enabled us to automatically discriminate perivascular
(<10 µm from vessel) from non-perivascular populations (Fig. 1h) and
determine that perivascular spaces are significantly enriched for c-kit+
progenitors (Fig. 1i). Therefore, we uncover a previously unknown
tendency of haematopoietic precursors to spatially arrange according
to a gradient throughout the diaphyseal cavity, and concentrate in
perivascular areas of bone marrow parenchyma.
HSPCs associate with blood vessels and preferentially localize
in endosteal zones
Next we determined the global distribution of three phenotypically
defined populations increasingly enriched in HSPC content. First,
c-kit+ cells lacking lineage marker expression, which encompass a
heterogeneous subset of lineage committed progenitor cells with
limited self-renewal capacity. Second, Sca-1+ c-kit+ (here on referred
to as SK), which are mostly negative for lineage markers, and consist of
multipotent progenitors, short-term and long-term repopulating stem
cells25 . Finally, we tracked very rare Lineage− CD48− CD41lo/− c-kit+
cells, highly enriched for Lin− Sca-1+ c-kit+ CD48− CD150+ , which are
mainly considered as long-term repopulating haematopoietic stem
cells (HSCs) and are highly enriched in early HSPC progenitor activity
(Supplementary Figs. S3a,b). LSC analysis of immunostained sections
permitted rapid screening and confirmation of individual cells falling
under the gates of interest by automatic relocation and visualization of
high-resolution images in every fluorescent channel (Fig. 2a–f), which
was necessary to exclude possible false positive cells. A total of 1,915
Lin− c-kit+ , 1,589 SK cells, and 312 Lin− CD48− CD41lo/− c-kit+ cells,
were visualized and their distance to vascular structures and endosteal
surfaces (detected by bone autofluorescence) was determined.
HSPC populations were equally distributed in metaphysis and
diaphysis of femoral cavities (Supplementary Fig. S3c). Within
these regions, more primitive subsets were increasingly enriched in
endosteal zones, defined as 100 µm from the closest bone surface
(42.72 ± 2.6 of Lin− c-kit+ , 61.0 ± 3.3% of SK and 72.64 ± 4.8% of
Lin− CD48− CD41lo/− c-kit+ cells; Fig. 2h). However, only a minority
was found in direct contact or close proximity to endosteal surfaces
(Fig. 2h). As expected, the average distance to endosteum of HSPCs
residing in metaphysis was significantly lower than those in the
diaphysis, where a detectable fraction of these cells resided in central
regions, >200 µm away from the closest bone surface (Supplementary
Fig. S3d,e). Notably, in all subsets (64.46 ± 3.2 of Lin− c-kit+ ,
81.99 ± 2.1% SK and 70.70 ± 5.3% Lin− CD48− CD41lo/− c-kit+ ), most
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
TECHNICAL REPORT
PM
c-kit
PM
c-kit
DAPI
f
Laminin
b
Isotype ctrl
a
200 μm
DAPI
DAPI
c-kit
c-kit
c
20 μm
Gated
c-kit+ gated
Diaphysis
Percentage of
c-kit+ cells
e
DM
500 μm
c-kit+ gated
7
6
5
4
3
2
1
0
50 μm
NS
g
DAPI
c-kit
Laminin
20 μm
MET
DIA
∗∗∗
∗
ER
CMR
h
ER
Non-perivascular
DAPI
Perivascular
i
Percentage of c-kit
Percentage of
c-kit+ cells
d
7
6
5
4
3
2
1
0
c-kit
DIA
6
5
4
3
2
1
0
∗∗
Non- Perivasc.
perivasc.
Figure 1 Quantitative image analysis of the spatial distribution of c-kit+
cells in whole longitudinal femoral sections. (a) Low-magnification image
of a longitudinal femoral bone marrow cryosection immunostained
for DAPI (blue) and c-kit (green) (PM, proximal metaphysis; DIA,
diaphysis; DM, distal metaphysis). (b) Representative LSC scattergrams
of isotype control and c-kit-stained bone marrow sections used to quantify
percentages of c-kit+ cells in different regions of the bone marrow in
d,e. Intensity in the DAPI channel is shown in the x axis. Bone marrow
cells contained in the c-kit+ gate in b can be automatically identified
and visualized in high-resolution field images (c). (c) Representative
image of a bone marrow stained for DAPI (blue) and c-kit (green) (top
panel). Cells gated as positive based on fluorescence intensity in the
c-kit channel are automatically marked by the LSC analysis software
(green rectangles, bottom panel). (d) Representative tissue maps and
frequencies of c-kit+ haematopoietic progenitors in the metaphysis (MET)
and diaphysis (DIA) of femoral bone marrow (data are represented as
mean ± s.e.m. calculated from n = 9 mice, a total of 17 sections were
analysed). (e) Frequencies of c-kit+ progenitors in longitudinal gates,
spanning the width of the diaphysis of femoral sections (representative
gates in left panel). ER, endosteal regions; CMR, central medullary region
(∗ P < 0.05 mean ± s.e.m. calculated from n = 7 mice, a total of 11
sections). (f) Image of the diaphysis of a bone marrow section co-stained
for c-kit and laminin. White arrows label perivascular c-kit+ cells. See
also Supplementary Video S1. (g) LSC analysis of perivascular cells
in bone marrow sections. Laminin fluorescence signal is thresholded
and an integration contour in which the original vascular structure is
expanded by 20 pixels (10 µm in the original ×40 magnification) is
defined to generate spatial gates containing only perivascular spaces
(green outline, right panel). (h) Representative LSC dot plots of total
bone marrow, perivascular and non-perivascular cell populations used
to quantify cells in i. (i) Frequency of c-kit+ cells in perivascular and
non-perivascular fractions (dashed line, frequency of total bone marrow;
∗∗
P < 0.01 mean ± s.e.m. calculated from n = 7 mice, a total of 10
sections analysed).
cells localized in close adjacency to endothelium (<10 µm) throughout
the entire bone marrow cavity, including densely vascularized
periendosteal regions (Fig. 2g). These results offer a comprehensive
view of the distribution of primitive HSPCs in bone marrow cavities,
highlighting perivascular areas of bone-proximal regions as their
preferred microenvironment.
To rule out the possibility that a fraction of HSPCs reside inside blood
vessels (as reported for immature B cells) we employed a previously
described protocol for the in vivo quantification of intravascular
populations26 . Injection of phycoerythrin-coupled anti-CD45 for very
short periods of time before euthanasia results in the specific labelling
of intravascular haematopoietic bone marrow populations, which can
then be quantified by flow cytometry26 . Under steady-state conditions,
the fraction of intravascular Lin− c-kit+ Sca-1+ (LSK) HSPCs and
Lin− c-kit+ Sca-1− (LK+ S− ) is very low or undetectable (Supplementary
Fig. S4–c and Supplementary Video S2).
3D study of bone marrow microvascular compartment
While studying the relationship of HSPCs with blood vessels, we
noticed the previously reported, heterogeneous nature of bone marrow
microvasculature24,27–29 . The presence of diverse types of vascular
structure in thin tissue sections was highly irregular and mostly
dependent on the dissection plane. To gain global insight into
the 3D structural relationship of diverse types of microvessel, we
devised a unique methodology to study whole-mount stained femoral
longitudinal slices (300–600 µm thick) using confocal or multiphoton
microscopy. 3D reconstructions of optical sections of the diaphysis
revealed a highly organized microvascular network in which central
longitudinal Sca-1+ large arteries gave rise to smaller radial arteries,
which in turn branched into arterioles as they migrated towards
endosteal regions (Fig. 3a,b and Supplementary Video S3). Transition
from arterial to venous circulation occurred right along, or in close
proximity to, the endosteum, where the network of arterial vessels
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
535
TECHNICAL REPORT
a
b
Lin
c-kit
*
Lin
Istp ctrl
*
Istp ctrl
*
c-kit
c
*
10 µm
100 µm
*
c-kit
Istp ctrl
d
Isotype ctrl
Istp ctrl
Laminin
DAPI
c-kit
Sca-1
*
*
Sca-1
LinCD41CD48
Istp ctrl
e
Istp ctrl
*
10 µm
100 µm
c-kit
f
Percentage of cells adjacent
to vasculature
**
g
h
100
Laminin
DAPI
c-kit
LinCD48CD41
**
*
80
60
*
40
*
20
0
Lin-c-kit+
SK
10 µm
Lin-CD48–
CD41lo/-c-kit+
100 µm
Laminin
DAPI
40
Freq. of population (%)
Lin–c-kit+
SK
30
Lin–CD48-CD41lo/–c-kit+
20
10
0
0–10
11–20
21–30
31–40
0–100
41–50
51–100
100–200 200–300
300–400
400–500
>500
Distance to bone (µm)
Figure 2 Most HSPCs lie in direct contact with bone marrow microvessels.
Mapping of three subsets of committed progenitors and HSPCs in
bone marrow cavities using LSC. Isotype control (Istp ctrl) stained
sections were used to determine specific gates. Individual HSPCs
falling in the target gates were systematically visualized and confirmed.
(a–f) Representative dot plots of bone marrow femoral sections stained for
different combinations of HSPC markers, and images of examples (marked
by asterisks) of Lin− c-kit+ (a,b), SK (c,d) and Lin− CD48− CD41lo/− c-kit+
(e,f) cells. (g) Frequency of perivascular (<10 µm from nearest blood
vessel) Lin− c-kit+ , SK and Lin− CD48− CD41lo/− c-kit+ cells. (h) Histogram
showing the distribution of the distances of Lin− c-kit+ (mean ± s.e.m.
calculated from n = 3 mice, a total of 3 sections and 1,915 individually
validated cells), SK (n = 8 mice, 8 sections and 1,589 individually
validated cells) and Lin− CD48− CD41lo/− c-kit+ HSPC (n = 4 mice, 8
sections and 312 individually validated cells) populations to nearest bone
surface.
progressively transformed into wider, irregularly shaped Sca-1lo/−
sinusoids (Fig. 3b,c, and Supplementary Video S4). Sinusoidal vessels
extended back into the cavity, forming frequent anastomosis and
eventually coalescing into a big collecting central sinus into which
the venous circulation drains (Fig. 3c and Supplementary Video S5).
The vascular compartment of the metaphysis was less structured. We
536
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
TECHNICAL REPORT
a
c
Diaphysis
Laminin Bone
a
s
cs
s
ev
b
d Laminin
ev
Bone
ev Laminin
25 µm
100 µm
200 µm
Sca-1
Laminin
Metaphysis
Sca-1
Bone
100 µm
Sca-1
200 µm
e
s
a
Laminin
a
s
Sca-1
a
50 µm
50 µm
Figure 3 3D study of bone marrow microvasculature heterogeneity.
(a) Representative 3D reconstruction of the femoral diaphysis stained
with laminin (green) and Sca-1 (red). Sca-1+ arteries run centrally
along the diaphysis, emitting branches of smaller arterial vessels,
which gradually narrow as they run laterally towards endosteal surfaces.
(b,c) Within endosteal regions, lamininhi Sca-1+ endosteal vessels give
rise to Sca-1− sinusoids, which return towards the central area of
the diaphysis and coalesce in the draining central sinus (shown in c).
ev, endosteal vessels; a,arteries; s,sinusoid; cs,central sinus. Bone
surfaces are shown in blue as revealed by second-harmonic generation
signals. (d) 3D overview of arterial and sinusoidal networks in the
femoral metaphysis. (e) Transition from arterial to sinusoidal vessels
(white arrow) in the endosteal surface of trabecular bone areas. See also
Supplementary Videos S3–S7.
observed Sca-1+ arteries entering the bone marrow through the bone
cortex, migrating longitudinally towards the diaphysis (Fig. 3d and
Supplementary Video S6) and giving rise to smaller arterioles that
frequently ran along the surface of the trabecular bone and drained into
sinusoids (white arrow Fig. 3e and Supplementary Video S7). These
experiments revealed that densely vascularized endosteal zones contain
a unique circulatory system.
On the basis of morphological criteria and differential expression
of Sca-1 and laminin, we were able to discriminate between sinusoidal
and non-sinusoidal endothelium in thin tissue sections and quantify
the interactions of HSPCs with different types of bone marrow
microvessel (Fig. 4a,b). Most perivascular Lin− CD48− CD41lo/− ckit+ cells directly associated with sinusoidal endothelium (Fig. 4c).
However, 12.0 ± 2.6% of this population and 36.72 ± 5.0% of
SK cells were found to lie adjacent to non-sinusoidal endosteal
vessels, or central and radial arteries expressing high levels of
Sca-1 and laminin (Fig. 4b.c). Thus, HSPCs, in particular SK
multipotent progenitors, directly interact with structurally diverse bone
marrow microvessels.
hypoxic probe Pimo (refs 13,14). Interestingly, not only long-term
HSCs (LSKCD34− ), but Lin− Sca-1+ c-kit+ (LSK) and LK+ S− cells,
containing mostly committed progenitors, avidly incorporated Pimo
as measured by flow cytometry (Fig. 5a,b). The hypoxic nature
of HSPCs has been proposed to derive from their residency in
endosteal regions and relatively distant from blood vessels13,18,20 , in
apparent contradiction with our demonstration of their perivascular
localization. We sought to clarify this discrepancy by employing
5-colour imaging cytometry to simultaneously analyse the spatial
distribution of different bone marrow cell subsets and levels of Pimo
incorporation in situ. Pimohi cells were scattered throughout the
bone marrow exhibiting no apparent spatial patterning (Fig. 5c,d).
In accordance with our fluorescence-activated cell sorting (FACS)
results, the Pimo-specific signal was significantly increased in c-kit+
progenitors when compared with total bone marrow or B220+ cells
(Fig. 5e,f). Strikingly, throughout the bone marrow we frequently
observed Pimohi perivascular c-kit+ progenitors, directly adjacent to
Pimolo/neg B220+ cells sharing the same extracellular environment
(Fig. 5d). A detailed analysis revealed that Pimo incorporation of c-kit+ ,
B220+ or total bone marrow cells was not defined by their residency
in the metaphysis or defined regions of the diaphysis, nor by their
distance to blood vessels (Fig. 5f–h and data not shown). Thus, the
hypoxic phenotype of haematopoietic progenitor populations is cell
specific rather than location dependent.
Haematopoietic progenitors exhibit a hypoxic state irrespective
of their localization in the bone marrow
HSPCs are considered relatively hypoxic when compared with other
bone marrow populations, as measured by incorporation of the
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
537
TECHNICAL REPORT
Laminin
a
c
2
2
1
DAPI
1
b
DAPI
c-kit
∗
Artery
1
3
3
Sca-1
∗
∗
∗
Laminin
∗
Merge
∗
∗
∗
∗
Endosteal
vessel
Sinusoid
2
Percentage of perivascular cells
c-kit Sca-1
500 μm
100
20 μm
90
Lin – CD48 – CD41lo/–c-kit +
80
SK
70
60
50
40
30
20
10
0
Sinusoidal
Non-sinusoidal
2
*∗
∗
*∗
20 μm
3
∗
∗
∗
20 μm
Figure 4 HSPCs directly interact with structurally and phenotypically
diverse types of bone marrow microvessel. (a) Low-magnification
immunofluorescence micrograph of a whole femoral section stained
for laminin (red), DAPI (blue), c-kit (green) and Sca-1 (magenta). Lower
panel shows the enlarged image of a section of the diaphysis in which central
arteries (Sca-1+ laminin+ ), sinusoids (laminin+/lo Sca-1−/lo ) and endosteal
vessels (Sca-1+ laminin+ ) can be visualized. (b) High-resolution images
enlarged from the outlined areas in a depicting examples of SK progenitors
(marked by white asterisks), interacting with arteries, sinusoids and
endosteal vessels. (c) Quantification of the percentages of perivascular SK
progenitors and Lin− CD48− CD41lo/− c-kit+ cells in contact with sinusoidal
and non-sinusoidal endothelium (mean ± s.e.m. calculated from n = 8 mice,
a total of 8 sections and 1,589 cells and n = 4 mice, a total of 8 sections
and 312 individually validated cells, respectively).
Pimo incorporation of primitive HSPC populations is unrelated to
localization in bone marrow endosteal zones or to cycling status
To further confirm the independence of the hypoxic status of HSPCs
from their positioning near endosteal surfaces, we performed an
in-depth study of SK cell distribution and levels of Pimo incorporation.
Again, perivascular SK cells exhibited an enhanced Pimo signal when
compared with that of bone marrow cells (Fig. 6a,b). Of note, the mean
fluorescence intensity (MFI) in the Pimo channel was consistently
increased in SK cells found at distant sites from bone, mainly in the
diaphysis, compared with those in the proximity of endosteal surfaces
(Fig. 6c). Therefore, contrary to what has been proposed18,20 , endosteal
areas of bone marrow do not harbour the most hypoxic HSPCs.
Notably, AMD3100-mobilized circulating, as well as splenic HSPCs,
incorporated high levels of Pimo, comparable to those of bone marrow
HPSCs (Fig. 6d), further reinforcing the notion that the hypoxic status
of HSPCs is preserved throughout different environments of the entire
organism. Of note, despite the proposed link between hypoxia and
quiescence, we found that non-dividing HSPCs in G0 incorporate
similar levels of Pimo to those in the G1 phase of the cell cycle (Fig. 6e,f).
cells exhibited homogeneous expression of HIF-1α, as measured by
flow cytometry (Fig. 7a). In accordance, robust HIF-1α expression
of c-kit+ progenitors was detected in immunostained bone marrow
sections using LSC (Fig. 7b). c-kit+ cells localized throughout the
different regions of the bone marrow expressed HIF-1α at comparable
levels (Fig. 7c,d), regardless of their perivascular localization (Fig. 7e).
Similarly, we visualized perivascular SK HSPCs expressing high levels
of HIF-1α, irrespective of their distance to the nearest endosteal
surface (Fig. 7f). Moreover, splenic HSPCs conserved similar HIF-1α
expression levels to bone marrow-residing HSPCs (Fig. 8a). Altogether,
our results demonstrate that HIF-1α stabilization in HSPCs occurs
independently of the possible differences in oxygenation registered at
different anatomical sites.
Finally, we examined HIF-1α expression levels in cycling and
non-cycling cells. Entry into cell cycle did not result in significant
changes in HIF-1α expression of HSPCs (Fig. 8b). Remarkably, highly
proliferating HSPCs found in murine bone marrow during the recovery
phase post-treatment with the cytotoxic agent 5-fluorouracil (5-FU; 7–9
days), exhibited increased levels of Pimo incorporation and HIF-1α
expression (Fig. 8c,e). As observed during homeostasis, Pimohi and
HIF-1αhi cells were found in direct contact with disrupted vascular
structures throughout the entire bone marrow cavity after 5-FU
treatment (Fig. 8d,f). In conjunction with the Pimo incorporation
data, these findings clearly demonstrate that the characteristic hypoxic
profile of HSPCs is not exclusive of quiescent HPSCs but retained
during cycling and on proliferative stress.
HSPCs exhibit stable expression of HIF1-α independently of
proximity to blood vessels, localization in bone marrow and
cell-cycle status
As for Pimo incorporation, stable expression of HIF-1α in HSPCs is
thought to result from residency in specific hypoxic regions of the bone
marrow. Notably, not only LSKCD34− and LSKCD34+ , but LK+ S−
538
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
TECHNICAL REPORT
a
∗∗
b
LSKCD34+
LSKCD34–
Laminin/Pimo
e
Endosteal region Central marrow
Laminin/Pimo
200
200 μm
Pimo injected
e
PBS injected
200 μm
Pimo injected
f
c-kit+
Pimo
Laminin/Pimo/c-kit/B220 Laminin/Pimo/c-kit/DAPI
50 μm
s
s
50 μm
∗
B220+
∗∗
1,000
MFI Pimo (a.u.)
B220+
s
BMC
BMC
Laminin/Pimo/c-kit
BMC B220+ LK+S- CD34+ CD34LSK
s
c-kit+
∗
800
600
400
h
1,600
Non-perivasc
1,200
800
400
0
200
0
g
MFI Pimo (a.u.)
PBS njected
d
e
Cell number
400
0
e
e
600
PM
DIA
ER
CMR
ER
Perivasc
Pimo
MFI Pimo (a.u.)
c
LK+S–
800
Cell number
B220+
Total born marrow
ΔMFI Pimo (AU)
PBS
Pimo
1,200
1,000
800
600
400
200
0
NS
Non- Perivasc.
perivasc.
DM
Figure 5 c-kit+ progenitors exhibit a hypoxic status regardless of their
perivascular localization and distribution in the bone marrow cavity.
(a) Representative flow cytometry histograms of Pimo incorporation
(green) of total bone marrow cells, B220+ cells, LS− K (Lin− c-kit+ Sca1− ) progenitors, short-term HSCs (LSKCD34+ ) and long-term HSCs
(LSKCD34− ). Grey histograms depict background levels of Pimo staining
in PBS-injected control mice. (b) Quantification of the shift in the
mean fluorescence intensity of the different bone marrow populations
in the Pimo channel (1MFI, in fluorescence arbitrary units) from mice
injected with Pimo compared with non-injected Pimo controls (n = 7
mice; ∗∗ P < 0.01). (c) Images of femoral sections of control PBS-injected
mice and Pimo-injected mice stained with anti-Pimo (e, endosteum).
(d) Representative images showing multiple examples of perivascular c-kit+
cells brightly stained for Pimo (white arrows) in both endosteal and central
marrow regions. Pimohi c-kit+ cells are found in many cases adjacent
to B220+ Pimo− cells (orange arrows). s, sinusoids. (e) LSC analysis of
sections stained for Pimo. Representative histograms used to quantify
the MFI in the Pimo-specific channel of total bone marrow cells, B220+
cells and c-kit+ cells in total bone marrow sections. (f) MFI of the Pimo
fluorescence signal of total bone marrow cells, B220+ cells and c-kit+ cells
localized in the diaphysis (DIA) and the proximal and distal metaphysis
(PM and DM, respectively; mean ± s.e.m. calculated from n = 5 mice and
a total of 9 sections; ∗ P < 0.05, ∗∗ P < 0.01). (g) Quantification of Pimo
signal (MFI) of c-kit+ cells present in consecutive longitudinal regions of
the diaphysis as shown in Fig. 1e. ER, endosteal regions; CMR,central
medullary region (mean ± s.e.m. calculated from n = 6 mice and a total of
9 sections). (h) Representative histograms of the fluorescence intensity (left)
and quantification of the MFI (right) in the Pimo channel of non-perivascular
and perivascular c-kit+ cells (mean ± s.e.m. calculated from n = 7 mice and
a total of 14 sections; NS, not significant).
DISCUSSION
In this study, we employ a quantitative imaging approach to
systematically map phenotypically defined HSPC subsets in the context
of entire bone marrow femoral sections. Haematopoietic progenitors
exhibit a tendency to accumulate in endosteal zones, which becomes
increasingly pronounced in primitive HSPC subsets and therefore
seems to be hierarchically regulated. Of note, only a minority of
HSPCs is observed in direct contact with endosteal surfaces. Contrary
to findings of previous studies, which investigated the localization
of transplanted HSPCs or long-term DNA-label-retaining cells4,6,7 ,
our results indicate that direct adjacency to bone-lining osteoblasts
is not required. Regardless of their spatial relationship to bone
surfaces, most HSPCs, as well as haematopoietic progenitors with
restricted self-renewal capacity, lie adjacent (<10 µm) to bone marrow
microvessels. Our findings further extend previous observations8–10 , by
demonstrating that vascular adjacency is not a feature only of primitive
HSCs but also of their immediate progeny, including SK cells, as well
as Lin− c-kit+ multipotent progenitors. In conjunction with previous
work30 , these results may indicate a functional role of the interaction
of c-kit with its ligand SCF, expressed mostly by perivascular stromal
cells12 , in the functional positioning of progenitors in the vicinity of
bone marrow microvessels.
The widespread distribution of HSPCs throughout substantially
large areas of the bone marrow raises the possibility that, unlike
what happens in other model systems such as intestinal crypt stem
cells, no rigorous spatial constraints apply for HSPC localization
within bone marrow. In this scenario, HSPC maintenance capacity
would not be restricted to a limited number of spatially confined
physical entities, but rather correspond to an extended property of
bone marrow parenchyma, which is pervasively infiltrated by dense
vascular and stromal cell networks with HSPC supportive capacity9,12 .
Alternatively, HSPCs could be strictly required to physically interact
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
539
TECHNICAL REPORT
a
c-kit
Sca-1
∗
Pimo
∗
Laminin
DAPI
Merge
∗
∗
10 μm
∗
∗
∗
∗
10 μm
2,000
d
∗
LSK
∗
1,500
1,600
MFI Pimo (AU)
1,250
1,000
750
500
1,200
800
Pimo
Born marrow
Blood
400
250
0
BMC
c-kit+
Spleen
100–200 200–300 300–400 400–500
Distance to endosteum (μm)
f
LSKCD34–
800
G1
S/G2/M
Ki67
104
105
103
0
Pimo
103 104 105
DAPI (DNA)
NS
S/G2/M
G0
G1
600
103
G0
0
G1
104
NS
700
ΔMFI (AU)
105
Ki67
0–100
LSKCD34+
e
Percentage of max
0
SK
G0
0
0
Percentage of max
MFI Pimo (AU)
∗
c
∗∗
Percentage of max
b
103 104 105
DAPI (DNA)
500
400
300
200
G0
G1
100
0
LSKCD34+
LSKCD34–
Figure 6 The hypoxic status of HSPCs is independent of their localization
in bone marrow regions and their cycling status. (a) Representative
images of perivascular SK cells that strongly stained for Pimo (marked
by white asterisks). (b) LSC-quantified Pimo-specific signal for total bone
marrow cells, c-kit+ progenitors and SK cells (mean ± s.e.m. calculated
from n = 3 mice and a total of 3 sections; ∗ P < 0.05, ∗∗ P < 0.01).
(c) Quantification of Pimo fluorescence signal of SK cells located at
various distances from nearest bone surface (mean ± s.e.m. calculated
from n = 3 mice, a total of 3 sections and 742 individually validated SK
cells, ∗ P < 0.05). (d) Pimo incorporation of LSK HSPCs present in bone
marrow, circulating blood and spleen after mobilization with AMD3100.
(e,f) Flow cytometric analysis of Pimo incorporation of LSKCD34+ and
LSKCD34− at different stages of the cell cycle. (e) Bottom panels
show representative histograms of Pimo-specific fluorescence signal for
fractions of each population at G0 and G1 . (f) Quantification of Pimo
intensity for LSKCD34+ and LSKCD34− HSPCs in G0 and G1 (n = 6
mice). NS, not significant. For source data for graphs presented in b,c,
see Supplementary Table S3.
with rare niche cell subsets, not investigated in this report, which would
provide specific cues to instruct maintenance of HSPCs. Supporting
this hypothesis is the recent description of bone marrow GFAP+
non-myelinating Schwann cells, which lie adjacent to a fraction of
HSPCs and promote TGF-ß-dependent HSPC quiescence31 . New
imaging methods enabling in situ multidimensional analysis will be
instrumental to elucidate whether HSPC-specific properties vary as a
result of anatomical localization.
Of particular relevance is our finding that hypoxic features seem
to be physiologically inherent to HSPCs and independent of their
confinement in defined tissue regions and organs or their spatial
relationship to blood supply. In line with previous reports14,32 , our
results challenge the pervasive view that the hypoxic status of HSPCs
directly results from their residence at sites of minimal oxygen
availability of the bone marrow, traditionally proposed to correspond
to endosteal surfaces as opposed to well-perfused and oxygenated
perivascular niches18,20 . Pimo is the most widely used and best
characterized hypoxic marker in terms of molecular mechanism16 .
Pimo competes with intracellular oxygen to bind electrons released
from the electron transport chain, converting into a reduced derivative
540
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
TECHNICAL REPORT
104
3
3
10
10
102
0
102
0
0
LSKCD34+
100 K
200 K
105
104
104
103
103
2
2
100 K
200 K
10
0
0
100 K
c
0
200 K
5
100 K
200 K
5
10
10
104
104
103
103
HIF-1α
Isotype ctrl
BMCs
0
105
10
0
LSKCD34–
b
2
10
0
0
100 K
102
0
0
200 K
FSC-H
100 K
200 K
FSC-H
3,000
c-kit+
Isotype ctrl
∗∗∗
∗∗∗
HIF-1α
∗∗∗
d
MFI HIF-1α (a.u.)
105
104
Cell number
LK+S–
105
MFI HIF-1α (a.u.)
a
2,500
2,000
1,500
1,000
500
0
PM
e
isolectinB4
HIF1-α
c-kit
DIA
2,800
2,400
2,000
1,600
1,200
800
400
0
DM
ER
CMR
ER
DAPI
20 μm
c-kit
Sca-1
∗
∗
∗
HIF1-α
∗
∗
isolectinB4
DAPI
Merge
∗
∗
∗
∗
∗
13 μm
f
10 μm
∗
∗
∗
∗
67 μm
∗
Endosteal
200 μm
20 μm
∗ ∗
∗ ∗
∗
∗
∗∗
∗∗
∗
∗
∗
290 μm
10 μm
Central
∗ ∗
401 μm
10 μm
10 μm
Figure 7 Stable HSPC expression of HIF-1α is not dictated by localization in
the bone marrow. (a) Flow cytometric analysis of intracellular HIF-1α
expression in subsets of haematopoietic progenitors and HSPCs.
(b) Representative LSC-obtained histograms depicting HIF-1α expression
of total bone marrow cells and c-kit+ progenitors in immunostained bone
marrow femoral sections. (c) MFI of the Pimo fluorescence signal of total
bone marrow cells, and c-kit+ cells localized in the diaphysis (DIA) and the
proximal and distal metaphysis (PM and DM, respectively; mean ± s.e.m.
calculated from n = 5 mice, a total of 10 sections ∗∗∗ P < 0.005).
(d) Quantification of HIF-1α expression (MFI) of c-kit+ cells present along
consecutive longitudinal regions of the diaphysis (mean ± s.e.m. calculated
from n = 5 mice and a total of 10 sections). (e) Representative images
depicting multiple examples of HIF-1α+ c-kit+ cells (white arrows) in direct
contact with isolectin B4+ vessels in the context of a large area of a bone
marrow femoral section. (f) Images of perivascular HIF-1α-expressing,
Sca-1+ c-kit+ cells in central and endosteal areas of the bone marrow
(marked by asterisks). Distances to the closest endosteal surface are
indicated.
that in turn incorporates into proteins. Thus, Pimo incorporation
levels indirectly reflect an intracellular state in which oxygen influx
is insufficient to absorb the production of electron release16 . This
status is prominently induced in most cells subjected to extremely
low oxygen supply, such as those present in avascularized areas of solid
tumours. However, we find that within the range of homeostatic oxygen
tensions estimated to exist in bone marrow tissues (1–6%; refs 18,20),
different cell types exhibit significant disparities in Pimo binding,
despite being physically adjacent and sharing similar extracellular
oxygen availability. Thus, it is conceivable that the relatively increased
Pimo incorporation capacity of HSPCs compared with other cell
subsets derives not directly from different oxygenation, but from their
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
541
TECHNICAL REPORT
LSKCD34–
LSK
HIF-1α
HIF-1α
c
1,400
CTRL
5-FU
Pimo
400
0
CTRL
5-FU
DAPI
5,000
CTRL
Endoglin
4,000
3,000
2,000
50 μm
1,000
0
5-FU
600
f
6,000
HIF-1α
800
CTRL
5-FU
HIF-1α
5-FU
∗
7,000
ΔMFI HIF-1α (a.u.)
50 μm
Percentage
of max
CTRL
CTRL
Percentage
of max
DAPI
e
1,000
200
Pimo
d
∗∗∗
1,200
ΔMFI Pimo (a.u.)
LSKCD34+
LK+S–
Laminin
G0
G1
Percentage
of max
b
Born marrow
Spleen
Percentage
of max
a
5-FU
50 μm
50 μm
Figure 8 HIF-1α expression on cycling and actively proliferating HSPCs.
(a) Expression of HIF-1α in HSPCs and haematopoietic progenitors from
bone marrow and spleen. (b) Flow cytometric analysis of HIF-1α expression
in HSPCs in different stages of the cell cycle. (c) Representative histograms
and quantification of Pimo incorporation of LSK cells in control and
5-FU-treated mice (7–9 days post-treatment), n = 8 mice ∗∗∗ P < 0.005.
(d) Images of bone marrow sections from control and 5FU-treated mice
stained for Pimo and vascular markers. (e) Representative histograms and
quantification of HIF-1α expression of LSK cells in control and 5-FU-treated
mice (7–9 days post-treatment). n = 5 mice ∗ P < 0.05. (f) Images of bone
marrow sections from control and 5-FU-treated mice stained for HIF-1α
and vascular markers. Pimo+ and HIF-1α+ cells localize in close contact
with disrupted vascular structures during the recovery phase post 5-FU
treatment.
distinctive metabolic dynamics, which result in a relative deficiency of
intracellular electron acceptors.
In this regard, it is important to note that HIF-1α is a master
regulator driving the remodelling of the metabolic machinery of
HSPCs towards anaerobic glycolysis15 . Our study shows that HIF1α expression in HSPCs is again not determined by their spatial
localization in bone marrow. This finding can be explained by the
fact that oxygen-dependent molecular mechanisms targeting HIF-1α
for rapid degradation are gradually activated at oxygen concentrations
above 5% (refs 18,33), and are therefore probably not the limiting factor
regulating HIF-1α expression levels in the context of bone marrow
microenvironmental conditions. Furthermore, oxygen-independent
mechanisms are known to promote stabilization of HIF-1α in HSPCs
even under saturating oxygen conditions34,35 . Our findings are further
supported by the fact that human circulating HSPCs conserve HIF-1α
expression36 and that HSPCs maintain a typical hypoxic transcriptional
program several hours after incubation at ambient oxygen pressures15 .
In light of our results, it seems reasonable to conclude that the
so-called hypoxic status of HSPCs is related to cell-specific mechanisms
derived from their glycolytic metabolic profile, rather than measurable
differences in intracellular oxygen levels or environmental oxygen
availability. Thus, instead of hypoxic, HSPCs would be best described as
low oxidative phosphorylation cells. Notably, although quiescence has
been proposed to relate to the hypoxic/metabolic phenotype of HSPCs
(refs 37,38), we demonstrate that hypoxic features are not exclusive
to quiescent HSPCs. Further studies are required to determine the
physiological implications of the referred metabolic profile, as well as
how it functionally relates to other HSPC-specific properties such as
self-renewal and multipotency.
Although not crucial for the metabolic phenotype of HSPCs, it is still
reasonable to assume that regional differences in bone marrow environmental oxygen tensions may participate in the control of differentiation
pathways, cell cycle and oxidative stress in haematopoietic cells in
general21,39 . Nonetheless, the assumption that endosteal areas are less
oxygenated relative to central portions of the marrow is not sustained
from an anatomical standpoint. As previously reported for bone
marrow cavities of mouse calvaria40 , we provide evidence that endosteal
surfaces of long bones are as highly vascularized as central bone marrow
regions, and contain the transition from arteriolar to sinusoidal/venous
circulation, where most of the oxygen exchange from blood to tissues
is thought to occur41 . A future challenge lies in determining the thus
far poorly understood dynamics of oxygen availability in bone marrow
parenchyma and its impact in haematopoietic physiology.
Anatomical subdivision in diverse functional microenvironments
is thought to underlie the complexity and plasticity of diverse blood
cell production in the bone marrow. Here we employ a technology
that proves invaluable to quantitatively study cell-specific niche
distributions and analyse cellular features of rare populations in highly
complex tissues. We offer unequivocal anatomical evidence to support
an alternative view to the widely accepted notion of a super hypoxic
HSPC niche in the bone marrow. The application of these quantitative
and 3D imaging techniques to bone marrow and other organs sustained
542
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
TECHNICAL REPORT
by rare stem cell populations will provide important insights into the
sophisticated regulation of tissue homeostasis.
METHODS
Methods and any associated references are available in the online
version of the paper.
Note: Supplementary Information is available in the online version of the paper
ACKNOWLEDGEMENTS
We are grateful to D. Rossi, L. Purton and C. P. Lin for critical reading of the
manuscript and thank the Compucyte Corporation team for helpful advice. C.N-A.
was a recipient of Human Frontiers in Science Program long-term fellowship
00194/2008-L. This work was financially supported in part by a seed grant from the
Harvard Stem Cell Institute. L.E.S. is supported by grants P01 HL095489 and R01
HL093139, and contract HHSN268201000009C from the National Heart Lung and
Blood Institute, USA.
AUTHOR CONTRIBUTIONS
C.N-A. designed and performed experiments, analysed data and wrote the
manuscript. G.P., B.W., K.J.C., S-Y.P. and J.L. performed experiments. B.H.
participated in the design of the study. J.E.M. and A.P. provided technical help. L.E.S.
designed the study and wrote the manuscript.
COMPETING FINANCIAL INTERESTS
The authors declare no competing financial interests.
Published online at www.nature.com/doifinder/10.1038/ncb2730
Reprints and permissions information is available online at www.nature.com/reprints
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Orkin, S. H. & Zon, L. I. Hematopoiesis: an evolving paradigm for stem cell biology.
Cell 132, 631–644 (2008).
Wang, L. D. & Wagers, A. J. Dynamic niches in the origination and differentiation of
haematopoietic stem cells. Nat. Rev. Mol. Cell Biol. 12, 643–655 (2011).
Adams, G. B. & Scadden, D. T. The hematopoietic stem cell in its place. Nat.
Immunol. 7, 333–337 (2006).
Wilson, A. et al. c-Myc controls the balance between hematopoietic stem cell
self-renewal and differentiation. Genes Dev. 18, 2747–2763 (2004).
Kiel, M. J. & Morrison, S. J. Uncertainty in the niches that maintain haematopoietic
stem cells. Nat. Rev. Immunol. 8, 290–301 (2008).
Zhang, J. et al. Identification of the haematopoietic stem cell niche and control of
the niche size. Nature 425, 836–841 (2003).
Xie, Y. et al. Detection of functional haematopoietic stem cell niche using real-time
imaging. Nature 457, 97–101 (2009).
Kiel, M. J., Yilmaz, O. H., Iwashita, T., Terhorst, C. & Morrison, S. J. SLAM family
receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial
niches for stem cells. Cell 121, 1109–1121 (2005).
Sugiyama, T., Kohara, H., Noda, M. & Nagasawa, T. Maintenance of the
hematopoietic stem cell pool by CXCL12-CXCR4 chemokine signalling in bone
marrow stromal cell niches. Immunity 25, 977–988 (2006).
Mendez-Ferrer, S. et al. Mesenchymal and haematopoietic stem cells form a unique
bone marrow niche. Nature 466, 829–834 (2010).
Omatsu, Y. et al. The essential functions of adipo-osteogenic progenitors as the
hematopoietic stem and progenitor cell niche. Immunity 33, 387–399 (2010).
Ding, L., Saunders, T. L., Enikolopov, G. & Morrison, S. J. Endothelial
and perivascular cells maintain haematopoietic stem cells. Nature 481,
457–462 (2012).
Parmar, K., Mauch, P., Vergilio, J. A., Sackstein, R. & Down, J. D. Distribution of
hematopoietic stem cells in the bone marrow according to regional hypoxia. Proc.
Natl Acad. Sci. USA 104, 5431–5436 (2007).
Takubo, K. et al. Regulation of the HIF-1α level is essential for hematopoietic stem
cells. Cell Stem Cell 7, 391–402 (2010).
Simsek, T. et al. The distinct metabolic profile of hematopoietic stem cells reflects
their location in a hypoxic niche. Cell Stem Cell 7, 380–390 (2010).
16. Krohn, K. A., Link, J. M. & Mason, R. P. Molecular imaging of hypoxia. J. Nucl. Med.
49 (Suppl 2), 129S–148S (2008).
17. Brahimi-Horn, M. C. & Pouyssegur, J. Oxygen, a source of life and stress. FEBS Lett.
581, 3582–3591 (2007).
18. Eliasson, P. & Jonsson, J. I. The hematopoietic stem cell niche: low in oxygen but a
nice place to be. J. Cell Physiol. 222, 17–22 (2010).
19. Winkler, I. G. et al. Positioning of bone marrow hematopoietic and stromal cells
relative to blood flow in vivo: serially reconstituting hematopoietic stem cells reside
in distinct nonperfused niches. Blood 116, 375–385 (2010).
20. Mohyeldin, A., Garzon-Muvdi, T. & Quinones-Hinojosa, A. Oxygen in stem cell
biology: a critical component of the stem cell niche. Cell Stem Cell 7,
150–161 (2010).
21. Suda, T., Takubo, K. & Semenza, G. L. Metabolic regulation of hematopoietic stem
cells in the hypoxic niche. Cell Stem Cell 9, 298–310 (2011).
22. Harnett, M. M. Laser scanning cytometry: understanding the immune system in situ.
Nat. Rev. Immunol. 7, 897–904 (2007).
23. Butler, J. M. et al. Endothelial cells are essential for the self-renewal and
repopulation of Notch-dependent hematopoietic stem cells. Cell Stem Cell 6,
251–264 (2010).
24. Hooper, A. T. et al. Engraftment and reconstitution of hematopoiesis is dependent
on VEGFR2-mediated regeneration of sinusoidal endothelial cells. Cell Stem Cell 4,
263–274 (2009).
25. Yang, L. et al. Identification of Lin(-)Sca1(+)kit(+)CD34(+)Flt3- short-term
hematopoietic stem cells capable of rapidly reconstituting and rescuing
myeloablated transplant recipients. Blood 105, 2717–2723 (2005).
26. Pereira, J. P., An, J., Xu, Y., Huang, Y. & Cyster, J. G. Cannabinoid receptor 2
mediates the retention of immature B cells in bone marrow sinusoids. Nat. Immunol.
10, 403–411 (2009).
27. Kopp, H. G., Avecilla, S. T., Hooper, A. T. & Rafii, S. The bone marrow vascular
niche: home of HSC differentiation and mobilization. Physiology (Bethesda) 20,
349–356 (2005).
28. Kopp, H. G., Hooper, A. T., Avecilla, S. T. & Rafii, S. Functional
heterogeneity of the bone marrow vascular niche. Ann. NY Acad. Sci. 1176,
47–54 (2009).
29. Li, X. M., Hu, Z., Jorgenson, M. L. & Slayton, W. B. High levels of acetylated
low-density lipoprotein uptake and low tyrosine kinase with immunoglobulin and
epidermal growth factor homology domains-2 (Tie2) promoter activity distinguish
sinusoids from other vessel types in murine bone marrow. Circulation 120,
1910–1918 (2009).
30. Kimura, Y. et al. c-Kit-mediated functional positioning of stem cells to their niches
is essential for maintenance and regeneration of adult hematopoiesis. PLoS One 6,
e26918 (2011).
31. Yamazaki, S. et al. Nonmyelinating Schwann cells maintain hematopoietic stem cell
hibernation in the bone marrow niche. Cell 147, 1146–1158.
32. Kubota, Y., Takubo, K. & Suda, T. Bone marrow long label-retaining cells
reside in the sinusoidal hypoxic niche. Biochem. Biophys. Res. Commun. 366,
335–339 (2008).
33. Jiang, B. H., Semenza, G. L., Bauer, C. & Marti, H. H. Hypoxia-inducible factor 1
levels vary exponentially over a physiologically relevant range of O2 tension. Am. J.
Physiol. 271, C1172–C1180 (1996).
34. Pedersen, M. et al. Stem cell factor induces HIF-1α at normoxia in hematopoietic
cells. Biochem. Biophys. Res. Commun. 377, 98–103 (2008).
35. Kirito, K., Fox, N., Komatsu, N. & Kaushansky, K. Thrombopoietin enhances
expression of vascular endothelial growth factor (VEGF) in primitive hematopoietic
cells through induction of HIF-1α. Blood 105, 4258–4263 (2005).
36. Piccoli, C. et al. The hypoxia-inducible factor is stabilized in circulating
hematopoietic stem cells under normoxic conditions. FEBS Lett. 581,
3111–3119 (2007).
37. Eliasson, P. et al. Hypoxia mediates low cell-cycle activity and increases the
proportion of long-term-reconstituting hematopoietic stem cells during in vitro
culture. Exp. Hematol. 38, 301–302 (2010).
38. Guitart, A. V., Hammoud, M., Dello Sbarba, P., Ivanovic, Z. & Praloran, V. Slowcycling/quiescence balance of hematopoietic stem cells is related to physiological
gradient of oxygen. Exp. Hematol. 38, 847–851 (2010).
39. Keith, B. & Simon, M. C. Hypoxia-inducible factors, stem cells, and cancer. Cell
129, 465–472 (2007).
40. Lo Celso, C. et al. Live-animal tracking of individual haematopoietic stem/progenitor
cells in their niche. Nature 457, 92–96 (2009).
41. Pittman, R. N. Oxygen transport and exchange in the microcirculation.
Microcirculation 12, 59–70 (2005).
NATURE CELL BIOLOGY VOLUME 15 | NUMBER 5 | MAY 2013
© 2013 Macmillan Publishers Limited. All rights reserved.
543
METHODS
DOI: 10.1038/ncb2730
METHODS
Mice. Adult C57/B6 mice (6–12 weeks old) were purchased from Jackson
Laboratories or Taconic. All animal experiments were approved and monitored by
the Children’s Hospital Animal Care and Use Committee. In all experiments mice
were euthanized by CO2 asphyxiation.
Antibodies. A complete list of antibodies and protocols used for flow cytometry
and immunofluorescent staining of HSPCs, vascular structures and hypoxyprobe
Pimo in bone marrow cryosections is provided in Supplementary Tables S1 and S2.
Immunofluorescence microscopy of bone marrow sections. To achieve
rapid in situ fixation of tissues, mice were perfused post-mortem with 10 ml
paraformaldehyde–lysine–periodate (PLP) fixative through the vena cava. Bones
were isolated, further fixed in PLP for 4–8 h, rehydrated in 30% sucrose solution
for 48 h and snap frozen in OCT (TissueTek). Whole longitudinal single-cell-thick
(5 µm) femoral cryosections were obtained using a Leica Cryostat and the Cryojane
tape transfer system (Leica Microsystems). Detailed experimental procedures
employed for immunofluorescent staining of different HSPC populations, hypoxic
markers and vascular structures are provided in Supplementary Table S1.
LSC analysis. LSC was performed with an iCys Research Imaging Cytometer
4-laser system (Compucyte) equipped with 4 laser lines (405, 488, 561 and 633 nm)
and 4 detectors with the following filter sets 450/40, 521/15, 575/50 and 650/LP.
Cryosections exhibiting damage or alteration of the microstructure of the bone
marrow parenchyma were discarded for scanning or analysis. Sections exhibiting
preserved integrity in all regions of the bone marrow (both metaphysis and
diaphyseal cavity) were selected, scanned and subsequently analysed entirely. Each
section was first scanned with a ×10 objective using the 405 nm laser to generate
low-resolution images of the DAPI-stained nuclei and obtain a general view of the
bone marrow. Subsequently, sections were divided into small regions that were
scanned with a ×40 dry objective lens at a spatial resolution of 0.25 mm stage
step size to create high-resolution field images of the whole bone marrow cavity
(Supplementary Fig. S1). Lasers were selected according to fluorescent dyes used in
each individual staining (Supplementary Table S1) and used to scan bone marrow
sections in independent passes to avoid the need for signal compensation. Detector
levels were set according to background fluorescence of isotype control stained
sections, which were always scanned in parallel to stained samples. Data were
analysed from raw, unprocessed images, post-acquisition using iCys Cytometric
Analysis Software (Compucyte). Bone marrow autofluorescent cells were typically
identified by collection of the red signal (filter set 575/50) after excitation with the
405 nm laser. This strategy allowed exclusion of autofluorescent events (∼3–5%) of
total bone marrow from our analysis. Images were saved in iCys and exported into
Photoshop (Adobe) for processing.
Generation and whole-mount immunostaining of femoral bone marrow
slices for microscopy. To generate thick femoral bone slices, femurs were collected
and processed as described above. OCT-embedded frozen femurs were iteratively
sectioned using a cryostat until the bone marrow cavity was fully exposed along the
longitudinal axis. Bone was then reversed and the procedure was repeated on the
opposite face of the bone until a thick slice of bone marrow was obtained. OCT
freezing medium covering the sample was removed, and slices were washed with
PBS and blocked overnight at 4 ◦ C in blocking solution (0.2% Triton/1%BSA/10%
donkey serum/PBS). Bone marrow slices were stained with rabbit anti-laminin and
rat anti-Sca-1 for 3 days in blocking solution, washed overnight in PBS and stained
with DyLight488 donkey anti-rabbit IgG and DyLight549 or DyLight649 donkey
anti-rat IgG. Whole-mount stained slices were then washed in PBS and incubated
overnight in FocusClear (CelExplorer Lab). For observation under the confocal
microscope, bone marrow slices were embedded in FocusClear and mounted on
glass sldes.
Confocal and multiphoton microscopy. Multiphoton microscopy was performed with an Olympus BX50WI microscope equipped with a Bio-Rad radiance
2,000 MP confocal/multiphoton microscopy system controlled by Laser-Sharp
software (Bio-Rad Laboratories). For multiphoton excitation a MaiTai Ti:sapphire
laser (Spectra-Physics) was tuned to 800 nm and stacks of x–y optical sections
were sequentially acquired along the z axis at variable step sizes. Second-harmonic
generation of bone tissue was detected through a 400/40 nm band-pass filter and
fluorescent emitted light through 525/50 nm and 620/100 nm band-pass filters with
non-descanned detectors to generate multi-colour images. Confocal microscopy was
performed with a Zeiss LSM510 system. Image stacks were processed and rendered
into 3D volumes using Volocity Software (Improvision) and ImageJ.
Flow cytometric assessment of the hypoxic profile of bone marrow, spleen
and peripheral blood populations. The hypoxic status of bone marrow cells was
assessed using Hypoxyprobe-1 Plus Kit (HPI) as previously described13 . In brief,
6–12-week-old mice were intraperitoneally injected with 60–120 mg kg−1 Pimo
90 min before euthanasia. One femur and one tibia were prepared for cryosectioning
and immunofluorescent staining; the contra-lateral limb bones were used for FACS
analysis. Bone marrow cell suspensions were stained for cell-surface markers and
permeabilized using IntraPrep Kit (Beckman Coulter) or Cytofix/cytoperm Kit
(BD Biosciences). Intracellular Pimo adducts were labelled with FITC-conjugated
mouse anti-Pimo. PBS-injected mice were used as controls to detect baseline
levels of anti-Pimo antibody binding. Intracellular expression of HIF-1α was
determined by staining with PE-conjugated mouse anti-HIF-1α (R&D) compared
with PE-conjugated mouse IgG. Stained cell suspensions were analysed by FACS.
For simultaneous cell-cycle analysis, PE or APC anti-Ki67 (BD Biosciences) were
added after permeabilization and cells were incubated in DAPI (2 µM) 20 min
before analysis. In some cases, for assessment of blood and spleen LSK populations,
mice were treated with AMD3100 (10 mg kg−1 ) 60 min before euthanasia. In some
experiments mice were treated with 250 µg g−1 of 5-FU (APP Pharmaceuticals) to
determine the effect of chemotherapy on the hypoxic profile of HSPCs.
Colony-forming unit assay. For colony-forming unit in culture (CFU-C)
assays, freshly collected bone marrow cells were fluorescently stained for HSPC
markers and populations of interest were sorted using a BD FACSAria sorter.
Lin− CD48− c-kit+ (300 cells per 35 mm Petri dish) and Lin-Sca-1+ c-Kit+ (LSK,
300 cells per 35 mm Petri dish), were cultured for CFU-Cs in methylcellulose
IMDM medium (MethoCult GF M3434, StemCell Technologies) containing
glutamine, 2-mercaptoethanol, IL-3, SCF, IL-6 and erythropoietin according to the
manufacturer’s instructions. Fresh whole bone marrow (WBM, 20,000 cells per
35 mm Petri dish) cells were used as a control. Colonies were counted after 7–12
days of culture at 37 ◦ C, 5% CO2 in a humidified incubator.
Labelling of intravascular populations of the bone marrow. Specific labelling
of bone marrow intravascular populations in vivo was performed as previously
described26 . In brief, mice were injected with 1 µg PE- or PECy7-conjugated rat
anti–mouse CD45 (BD Biosciences) in a total volume of 100 µl of PBS. Mice were
euthanized 2 min after injection and bone marrow from femurs and tibiae was
collected. Bones were crushed with a mortar and pestle, single-cell suspensions were
obtained in PBS 2% FCS and red blood cells were removed with ACK lysis buffer.
Bone marrow cell suspensions were then blocked using anti-mouse CD16/32 (BD
Pharmingen) and stained with monoclonal antibodies against HSPC markers. In
some cases, bones were processed for analysis under the confocal microscope.
Statistical analysis. Statistical significance was determined using a two-tailed
unpaired Student t -test (Prism Software, Graphpad). Data are always expressed as
mean±s.e.m.. For graphs representing LSC data, single femurs from different mice
analysed were considered as independent data points (n = number of mice). In
cases in which multiple sections from a single femur were analysed, the results were
treated as technical replicates, which were averaged and considered as one single
independent sample for statistical purposes. The number of independent samples,
and the total number of sections analysed in each experiment are provided in the
figure legends. Statistical significance and P values are indicated by asterisks and
specified in the figure legends.
NATURE CELL BIOLOGY
© 2013 Macmillan Publishers Limited. All rights reserved.
S U P P L E M E N TA R Y I N F O R M AT I O N
DOI: 10.1038/ncb2730
Nombela-Arrieta and Silberstein
Supplementary Figure 1
a
10x
DAPI
field images
200µm
Tissue map
X
contoured cells
50µm
Dot plot
antigen X
Y
b
40x
DAPI
Figure S1 Laser Scanning Cytometry of bone marrow microenvironments.
(a) Workflow of global LSC-based analysis of BM populations in femoral
cavities. Whole longitudinal, 5 µm-thick cryosections of femoral bones are
scanned using the iCys research cytometer. Low-resolution digital images of
entire femoral cavities stained with the nuclear marker DAPI, are generated
through scanning with a 405 nm laser and a low magnification objective lens
(10x, upper panel). Dynamic scanning areas are defined to cover the entire
BM cavity, and are subsequently scanned with lasers of choice (405nm,
488nm, 561nm, 633nm) using a 40x objective lens and a step size of
0.25 mm, to acquire high-resolution field images in different fluorescent
Histogram
antigen X
channels. Images are assembled to generate multicolor mosaic images of
the entire BM cavity. Lower panel depicts two grayscale field images (right)
and a composite image of one BM region generated in the DAPI channel.
Automatic image segmentation of DAPI+ nuclei is performed with the iCys
analysis software to generate contours around single nuclei, which define
cellular events. Positional information in a XY Cartesian coordinate system
is recorded for each individual cell enabling the generation of tissue maps
(b) The specific signal in every fluorescent channel is quantified on a per
cell basis and can be displayed and analyzed in dot-plot scattergrams and
histogram charts.
WWW.NATURE.COM/NATURECELLBIOLOGY
1
© 2013 Macmillan Publishers Limited. All rights reserved.
S U P P LNombela-Arrieta
E M E N TA R Y I Nand
F O Silberstein
R M AT I O N
Supplementary Figure 2
c-kit
Laminin
3D confocal
Figure S2 Perivascular accumulation of c-kit+ progenitors in the BM. Representative 3D projection of optical image stacks of thick femoral slices wholemount stained for c-kit (green) and Laminin (red). Multiple c-kit+ cells are shown in direct contact with Laminin+ vessels (see also Supplementary Video 1).
2
WWW.NATURE.COM/NATURECELLBIOLOGY
© 2013 Macmillan Publishers Limited. All rights reserved.
Nombela-Arrieta and Silberstein
S U P P L E M E N TA R Y I N F O R M AT I O N
Supplementary Figure 3
Lin-CD48-c-kit+
10 5
10
4
10
10
3
10 5
10
Sca-1
10
10.1%
4
3
10
84.1%
4
10
5
10
4
CD48
5
c-kit
CD48
a
10
3
10
67.3%
3
10 2
0
0
0
10 3
10 4
10 5
CD3, Ter-119, B220, Gr-1
10 2
10 3
SSC-A
10 4
10 5
150
c
125
100
50
Freq. of population (%)
Freq. of population (%)
10 4
SSC-A
10 5
0
10 3
10 4
CD150
10 5
Lin-c-kit+
0.6
SK
Lin-CD48-CD41lo/-c-kit+
WBMC
0
LSK Lin CD48
+
c-kit
MET
Metaphysis
40
DIA
Lin-c-kit+
SK
Lin-CD48-CD41lo/-c-kit+
30
20
10
0
e
10 3
0.2
25
d
10 2
0.4
75
0
0
Freq. of BMC (%)
CFU-C/500 cells
b
0
0
0
0-10
11-20
21-30
31-40
0-100
41-50
51-100
100-200
200-300 300-400 400-500
>500
Distance to bone (µm)
Diaphysis
40
Lin-c-kit+
SK
Lin-CD48-CD41lo/-c-kit+
30
20
10
0
0-10
11-20
21-30
31-40
0-100
41-50
51-100
100-200
200-300 300-400 400-500
>500
Distance to bone (µm)
Figure S3 Quantification of distribution of HSPCs in the bone marrow.
(a) Flow cytometry analysis of the expression of Sca-1 and CD150
of Lineage -CD48-c-kit+ cells. Although expression of CD41 can be
detected in HSPCs by FACS, our immunofluorescence analysis of BM
sections CD41 labeled exclusively megakaryocytes and platelets, which
express much higher levels than HSPCs. Thus, by LSC we gated on
the CD41lo/- population to exclude both megakaryoctes and platelets
from our HSPC analysis. (b) CFU-C content of Lin-CD48-c-kit+ cells
compared to that of Lin -Sca-1+c-kit+ (LSK) cells and whole BM cells
(WBMC). (c) Absolute frequencies of HSPC subsets in the metaphysis
and diaphysis of femoral BM as measured by LSC. (d and e) Histograms
of the distances Lin -c-kit+ of SK and Lin -CD48-CD41lo/-c-kit+ HSPC
populations to nearest bone surfaces in the metaphysis (d) and
diaphysis (e) of femoral cryosections.
WWW.NATURE.COM/NATURECELLBIOLOGY
3
© 2013 Macmillan Publishers Limited. All rights reserved.
Nombela-Arrieta and Silberstein
S U P P L E M E N TA R Y I N F O R M AT I O N
Supplementary Figure 4
a
10
3
10
4
3
10 2
10 2
0
0
0
50K
100K
150K
FSC-A
10
5
10
4
10
3
10 3
Lineage
10 4
10 5
0
10 2
4
10
4
10
4
10
3
10
3
10
5
10
4
10
4
10
3
10
3
10 2
10 2
0
0
10 4
10 5
10 4
10 5
10 2
0
0
10 5
10 3
Sca-1
LSK
5
5
10 3
2
10
10
0
10
5
0
10
3
10
10 2
10
10
LKS-
0
3
4
0
0
10 2
0
anti CD45-PE
250K
LS-K-
i.v injection
PBS
200K
10
c-kit
10
4
SSC-A
SSC-A
10
10 5
10 5
10
3
10
4
10
5
10
5
10
4
10
3
0
10 3
10 4
10 5
0
10 3
10 4
10 5
SSC-A
10 5
10 2
0
0
10 3
10 4
10 5
CD45-PE
c
CD45-PECy
Laminin
6
****
****
5
% intravascular
b
4
3
2
1
0
Figure S4 HSPCs do not reside in the intravascular compartment of the BM.
Labeling of hematopoietic BM intravascular populations by injection of PEconjugated anti-CD45. Representative FACS dot plots (a) and quantification
(c) of PE+ LSK, LS-K (Lin-c-kit+Sca-1-) and LS-K- (Lin-c-kit+Sca-1-) cells in
the BM of mice after injection of PBS (control) or CD45-PE (upper and lower
4
LSK
LKS- LS-K- B220+
panels) n=10 ****p< 0,001. (b) Representative 3D projection of optical
image stacks of the BM cavities of CD45-PE injected mice. CD45-PE+ cells
(red) are exclusively observed inside vascular structures (green) therefore
validating the specificity of the method as previously reported. See also
Supplementary Video 2.
WWW.NATURE.COM/NATURECELLBIOLOGY
© 2013 Macmillan Publishers Limited. All rights reserved.
S U P P L E M E N TA R Y I N F O R M AT I O N
Supplementary Video legends
Supplementary Video 1. Interaction of c-kit cells with BM microvasculature. 3D-reconstruction of confocal optical sections of thick bone marrow slices
stained for c-kit (green) and the vascular marker Laminin (red). Multiple examples of vessel-adjacent c-kit + cells can be visualized.
Supplementary Video 2. In vivo labeling of intravascular populations. The movie depicts a series of sequential optical sections on the z-axis followed by the
3D-reconstruction of a thick BM femoral slice, 2 mins post-injection of CD45-PE. Intravascular populations (blood vessels in green labeled with Laminin) are
specifically stained with CD45 (red).
Supplementary Video 3. Vascular compartment of the femoral diaphysis. 3D-reconstruction of confocal optical sections of a thick slice of a murine femur
stained with the panvascular marker Laminin (green) and the arterial marker Sca-1 (red).
Supplementary Video 4. BM sinusoidal microvasculature. 3D-reconstruction of multiphoton optical sections of a femoral diaphysis stained for Laminin
(green). The central sinus, as well as an underlying central artery can be visualized running longitudinally across the diaphysis. Bone is visualized in blue as
revealed by second harmonic generation.
Supplementary Video 5. Arteriolar to sinusoidal transition in endosteal regions of the diaphysis. High-magnification 3D-reconstruction of the endosteal
vessels shown in Supplementary Video 4. Arrow depicts the arteriolar to sinusoidal transition in close proximity to endosteal surfaces (bone shown in blue).
Supplementary Video 6. Vascular compartment of the femoral metaphysis. 3D-reconstruction of the vascular network of a femoral metaphysis. All BM
vascular structures are marked by Laminin (green), while the arterial network is positive for Sca-1 (red).
Supplementary Video 7. Endosteal microvascular network of femoral metaphysis. 3D reconstruction of the vascular network of endosteal regions of a femoral
metaphysis. Arrow depicts the arteriolar to sinusoidal transition in close proximity to endosteal surfaces (bone shown in blue).
Supplementary Table Legends
Supplementary Table 1. Complete list and technical specifications of antibodies employed in this study for flow cytometry and immunostaining of BM
cryosections.
Supplementary Table 2. Detailed strategies for immunofluorescent stainings, including antibody and fluorochrome combinations and laser and filter setups,
used for LSC detection and analysis of the different hematopoietic populations in BM sections.
Supplementary Tables 3 and 4. Statistics source data for Figures 6b and 6c.
WWW.NATURE.COM/NATURECELLBIOLOGY
5
© 2013 Macmillan Publishers Limited. All rights reserved.
Protocol for immunofluorescent staining of murine femoral BM cryosections
1)
Fix tissues and rehydrate as described in Methods section. Freeze embedded in OCT.
2)
Section using a Leica Cryostat and the Cryojane tape transfer system (Leica microsystems)
3)
Rehydrate in PBS (2 mins) and 0.1% Tween20/ PBS (2 mins)
4)
Block in PBS/10% donkey serum
5)
Endogenous biotin block using Vector Biotin Blocking Kit according to manufacturerʼs
instructions
6)
Incubate in primary antibodies in PBS /10% donkey serum for 1 hour at room temp or
overnight at 4°C (antibodies in Supplementary Table 1).
7)
Wash in PBS/0.1% Tween20 3x5 minutes
8)
Secondary Antibodies: Room temperature for 1 hour in PBS /10% donkey serum
9)
Wash in PBS/0.1% Tween20 3x5 minutes
10)
Tertiary antibody and additional steps: Room temperature for 1 hour in PBS /10% donkey
serum
11)
Incubation with DAPI 0.5µM – 1.0µM (Invitrogen)
12)
Wash in PBS/0.1% Tween20 3x5 minutes
13)
Dehydrate in EtOH gradient: 70%, 85%, 95%, 100%, 3 minutes each
14)
Rinse in xylene 2x5 minutes
15)
Coverslip with mounting media (Vectashield, Vector labs), store at 4°C until scanned
© 2013 Macmillan Publishers Limited. All rights reserved.
Supplementary Table 1. Technical specifications of the Antibodies (dilutions, clones and sources) employed in our study
Antigen/Specificity
Conjugation
Concentration
Host
Clone
Source
Catalogue #
CD117/c-kit
PECy7
1µg/ml
Rat
2B8
BioLegend
105814
Application
FC
Sca-1/Ly6A/E
APCCy7
1µg/ml
Rat
D7
BioLegend
108122
FC
Gr-1
APC
1µg/ml
Rat
RB6-8C5
BioLegend
108412
FC
CD3
APC
1µg/ml
Rat
145-2C11
eBioscience
17-0031-83
FC
B220
APC
1µg/ml
Rat
RA3-6B2
BioLegend
103212
FC
Ter-119
APC
1µg/ml
Rat
Ter-119
eBioscience
17-5921-83
FC
CD34
FITC
1µg/ml
Rat
RAM34
eBioscience
11-0341-85
FC
CD34
PerCpCy5.5
1µg/ml
Rat
HM34
BioLegend
128608
FC
CD48
FITC
2.5µg/ml
Rat
HM48-1
eBioscience
11-0481-85
FC
CD150
PE
1µg/ml
Rat
TC15-12F12.2
BioLegend
115904
FC
CD45
PE
injected i.v 1µg
Rat
30-F11
BD biosciences
553-081
FC
CD45
PECy7
injected i.v 1µg
Rat
30-F11
BioLegend
103113
FC
B220
PerCpCy5.5
1µg/ml
Rat
RA3-6B2
BioLegend
103235
FC
Streptavidin
PerCpCy5.5
1µg/ml
N/A
N/A
BD biosciences
551419
FC
mouse IgG1 isotype ctrl
APC
20µl/test
Mouse
MOPC-21
BD biosciences
557783
FC
Ki67
APC
5µl/test
Mouse
B56
BD biosciences
561126
FC
mouse IgG isotype ctrl
PE
0.5µg/test
Mouse
11711
R&D
IC002P
FC
HIF-1a
PE
0.5µg/test
Mouse
241812
R&D
IC1935P
FC
Pimonidazole
FITC
2.5 µg/ml
Mouse
4.3.11.3
Hypoxyprobe
HP2-100
FC/IC
Gr-1
biotin
1µg/ml
Rat
RB6-8C5
eBioscience
13-5391-75
FC/IC
CD3
biotin
1µg/ml
Hamster
145-2C11
eBioscience
13-0031-75
FC/IC
B220
biotin
1µg/ml
Rat
RA-6B2
eBioscience
13-0452-75
FC/IC
Ter-119
biotin
1µg/ml
Rat
Ter-119
eBioscience
13-5921-75
FC/IC
CD117/c-kit
unconjugated
5µg/ml
Goat
polyclonal
R&D
IEO02
IF
Sca-1
unconjugated
5µg/ml
Rat
E13-161.7
BioLegend
122502
IF
Laminin
unconjugated
10µg/ml
Rabbit
polyclonal
Sigma
L9393
IF
B220
unconjugated
5µg/ml
Rat
RA-6B2
eBioscience
14-0452-82
IF
Ter-119
unconjugated
5µg/ml
Rat
Ter-119
eBioscience
14-5921
IF
Gr-1
unconjugated
6.25µg/ml
Rat
RB6-8C5
BD biosciences
550291
IF
CD48
Biotin
5µg/ml
Hamster
HM48-1
BioLegend
103402
IF
CD41
Biotin
5µg/ml
Rat
eBioMWReg30
eBioscience
13-0411-82
IF
CD4
unconjugated
5µg/ml
Rat
H129.19
BioLegend
130302
IF
CD8
unconjugated
5µg/ml
Rat
53-6.7
BioLegend
100701
IF
Pimonidazole
DyLight549
1 µg/ml
Mouse
4.3.11.3
Hypoxyprobe
HP2-100
IF
HIF-1a
unconjugated
1:500
Rabbit
polyclonal
Thermoscientific
PA1-16601
IF
Endoglin
unconjugated
0.6µg/ml
Goat
polyclonal
R&D
AF1320
IF
rat IgG
DyLight488
7.5µg/ml
Donkey
polyclonal
Jackson immunoResearch
712-485-153*
IF
rat IgG
DyLight549
7.5µg/ml
Donkey
polyclonal
Jackson immunoResearch
712-505-153*
IF
rat IgG
DyLight649
7.5µg/ml
Donkey
polyclonal
Jackson immunoResearch
712-495-153*
IF
goat IgG
Biotin
7.5µg/ml
Donkey
polyclonal
Jackson immunoResearch
705-065-147
IF
goat IgG
DyLight488
7.5µg/ml
Donkey
polyclonal
Jackson immunoResearch
705-485-147*
IF
rabbit IgG
Biotin
7.5µg/ml
Donkey
polyclonal
Jackson immunoResearch
711-065-152
IF
rabbit IgG
DyLight649
7.5µg/ml
Donkey
polyclonal
Jackson immunoResearch
711-495-152*
IF
hamster IgG
Biotin
7.5µg/ml
Goat
polyclonal
Jackson immunoResearch
127-065-160
IF
Streptavidin
Brilliant Violet 650
5µg/ml
N/A
N/A
BioLegend
405213
IF
Streptavidin
Qdot705
0.01µM
N/A
N/A
Invitrogen
Q10161MP
IF
Streptavidin
AF555
4µg/mL
N/A
N/A
Invitrogen
S32355
IF
isotype ctrl goat IgG
unconjugated
5µg/ml
Goat
polyclonal
Southern Biotech
0109-01
IF
isotype ctrl Rabbit IgG
unconjugated
10µg/ml
rabbit
polyclonal
Southern Biotech
0111-01
IF
isotype ctrl Rat IgG
unconjugated
20µg/ml
rat
eBR2a
eBioscience
14-4321-80
IF
IF= Imunofluorescence
FC= flow cytometry
*discontinued by company
© 2013 Macmillan Publishers Limited. All rights reserved.
Supplementary Table 2: Antibodies and strategies used for immunofluorescent staining of BM cryosections
Target
populations
+
c-kit progenitors
and BM
vasculature
+
Lin c-kit HSPCs
and BM
vasculature
+
+
Sca-1 c-kit
hematopoietic
progenitors and
BM vasculature
+
+
Sca-1 c-kit
hematopoietic
progenitors,
pimonidazole and
BM vasculature
+
+
B220 cells, c-kit
hematopoietic
progenitors,
pimonidazole and
BM vasculature
Lin CD48 CD41 c+
kit HSPCs and
BM vasculature
+
c-kit progenitors,
B220 cells and
HIF-1α
+
+
Sca-1 c-kit
HSPCs, HIF-1α
and BM
vasculature
Primary antibodies
Secondary/Tertiary antibodies
- Goat anti-c-kit
- Rabbit anti-Laminin
- Donkey anti-goat DyLight488
- Donkey anti-rabbit DyLight549
- Goat anti-c-kit
- Rabbit anti-Laminin
- Rat anti-CD4 and CD8
- Rat anti-Gr-1
- Rat anti-Ter119
- Rat anti-Sca-1
- Goat anti-c-kit
- Rabbit anti-Laminin
- Donkey anti-goat biotin
- Rat anti-Sca- Goat anti-c-kit
- Rabbit anti-Laminin
- FITC or DyLight549 mouse anti-PIM
- Donkey anti-goat biotin + Streptavidin Qdot 705
- Donkey anti-rat DyLight549 or 488
- Donkey anti-rabbit DyLight649
- Rat anti-B220
- Goat anti-c-kit
- Rabbit anti-Laminin
- FITC or DyLight549 mouse anti-Pimo
- Donkey anti-goat biotin + Streptavidin Qdot 705
- Donkey anti-rat DyLight549 or 488
- Donkey anti-rabbit DyLight649
- Goat
anti-c-kit
- Biotinylated rat anti-B220
- Biotinylated rat anti-Ter110
- Biotinylated rat anti-Gr-1
- Biotinylated hamster-anti CD3
- Biotinylated hamster anti-CD48
- Biotinylated rat-anti CD41
- Rabbit anti-Laminin
- Donkey anti-goat
DyLight 488
- Streptavidin DyLight 488
- Donkey anti-rat DyLight549
- Donkey anti-rabbitDyLight649
- Donkey anti-goat DyLight 488
- Donkey-anti-rat biotin + SA Alexa 555
- Donkey anti-rabbit DyLight 649
-Block with
excess
goat IgG
(50100µg/ml)
- Goat antihamster
biotin
- Donkey anti-rabbit DyLight649
- SA-AF555
Laser-Filter set detectorDye
- 405-450/40-DAPI
- 488-521/15-DyLight488
- 561-575/50-DyLight549
- 405-450/40-DAPI
- 488-521/15-DyLight488
- 561-575/50-DyLight549
- 633-650/LP-DyLight649
- 405-450/40-DAPI
- 488-521/15-DyLight488
- 561-575/50-DyLight549
- 633-650/LP-DyLight649
- 405-450/40-DAPI
- 405-650LP-Qdot705
- 488-521/15-DyLight488
- 561-575/50-DyLight549
- 633-650/LP-DyLight649
- 405-450/40 –DAPI
- 405-650LP- Qdot705
- 488-521/15-FITC/DyLight488
- 561-575/50-DyLight549
- 633-650/LP- DyLight649
- 405-450/40-DAPI
- 488-521/15-FITC/DyLight488
- 561-575/50-SA-AF555
- 633-650/LP-DyLight649
- Goat anti-c-kit
- Rat anti-B220
- Rabbit anti-HIF-1α
- Donkey anti-rabbit biotin
- Donkey anti-goat DyLight488
- SA-AF555 (Invitrogen)
- Donkey anti-rat DyLight649
- 405-450/40-DAPI
- 488-521/15- DyLight488
- 561-575/50-SA-AF555
- 633-650/LP-DyLight649
- Goat anti-c-kit
- Rat anti-Sca-1
- Rabbit anti-HIF-1α
- Donkey anti-goat biotin
- Streptavidin Brilliant Violet (BV) 650
- Isolectin B4 AF589 (injected)
- Donkey anti-rat DyLight488
- Donkey anti-rabbit DyLight649
- 405-450/40-DAPI
- 488-521/15-DyLight488
- 561-575/50-DyLight549
- 633-650/LP-DyLight649
© 2013 Macmillan Publishers Limited. All rights reserved.