Biotechnol. Appl. Biochem. (2010) 55, 199208 (Printed in Great Britain) doi:10.
1042/BA20090356
Isolation of a mouse bone marrow population enriched in stem
and progenitor cells by centrifugation on a Percoll gradient
Ana-Maria Rosca and Alexandrina Burlacu1
Laboratory for Adult and Embryonic Stem Cell Biology, Institute of Cellular Biology and Pathology Nicolae Simionescu, Hasdeu
Street, Bucharest 050568, Romania
Given the complex composition of bone marrow, a
cell separation technique that results in populations
enriched in progenitor cells is required for cellular dif-
ferentiation and transplantation studies. In the present
study, we designed a method that allows for the
isolation of a progenitor-enriched population of bone
marrow by exploiting the physical properties of these
cells. Bone marrow aspirate was separated on a discon-
tinuous Percoll gradient (ranging from 1.050 to 1.083
g/cm3 ) that resulted in the recovery of six cell fractions.
The fractions were characterized by FACS and RTPCR
(reverse transcriptionPCR) analyses and evaluated for
their capacity to differentiate into haematopoietic and
mesenchymal cells. Fraction IV, including cells with
a density of 1.0701.076 g/ml, contained 11.68% of
total bone marrow cells and was enriched in c-kit+
and Sca-1+ (stem cell antigen-1) progenitor cells as
compared with total bone marrow. This fraction
demonstrated an increase in clonogenic capacity under
specific conditions as well as a potential to generate a
mesenchymal stem cell culture in a shorter period than
that using bone marrow aspirate. Furthermore, this
fraction lacked differentiated cell types and contained
cells positive for endothelial markers, which further
increases its value in cellular transplant. In conclusion,
a bone marrow subpopulation that is enriched in
progenitor cells and may be valuable in cellular
transplant therapy can be isolated by exploiting the
physical properties of these cells.
Introduction
Adult bone marrow is still recognized as the main source
of stem/progenitor cells throughout the life of an adult.
This tissue houses two stem cell populations with distinct
progenies that are valuable in cellular transplant: HSCs
(haematopoietic stem cells) and MSCs (mesenchymal stem
cells) [1]. HSCs can give rise to all blood cells, whereas
MSCs have been shown to differentiate in vitro, not only
into mesenchymal lineages, such as osteoblasts, adipocytes
and chondrocytes [1,2], but also into neurons and muscle
199
[3]. However, stem/progenitor cells represent rare events
among bone marrow nucleated cells [4,5]. They exist in a
highly organized microenvironment composed of stromal
cells (osteoblasts, adipocytes, endothelial cells and vascular
pericytes), an extracellular matrix rich in fibronectin,
collagens and proteoglycans [6] and a large amount of
haematopoietic cells.
Given the complex composition of bone marrow, a
cell separation technique that results in subpopulations
enriched in stem/progenitor cells is required for cellular
differentiation and transplantation studies. At present,
density gradient centrifugation using Histopaque is employed
by many clinical investigators to separate progenitor cells
from blood; however, it provides a pool of mononuclear cells
containing all cells with a density of less than 1.077 g/ml, but
no substantial improvement in progenitors. Other general
methods to isolate progenitor cells from the mononuclear
cell population include immunological methods [7], which
are time-consuming and expensive because of the use of
monoclonal antibodies and advanced technology.
In the present study, we aimed to find out whether
a bone marrow subpopulation enriched in stem/progenitor
cells can be obtained by exploiting the physical properties
of these cells. For this purpose, we designed a method to
fractionate bone marrow cells, aiming in particular for cells
with a density of less than 1.077 g/ml. The bone marrow
suspension was layered on a discontinuous Percoll gradient,
which allowed the fractionation of cells with a density of
less than 1.077 g/ml into five subpopulations. Cells with
a density higher than 1.077 g/ml remained at the bottom
of the gradient in a distinct fraction. We showed that
a subpopulation of bone marrow, which was enriched in
small, c-kit+ and Sca-1+ (stem cell antigen-1) progenitor
Key words: bone marrow, c-kit, haematopoietic stem cell (HSC),
mesenchymal stem cell (MSC), Percoll, stem cell antigen-1 (Sca-1).
Abbreviations used: CFU, colony forming units; DMEM, Dulbeccos modified
Eagles medium; FBS, fetal bovine serum; HSC, haematopoietic stem cell;
MSC, mesenchymal stem cell; PBS-FBS, PBS containing 2% FBS; PECAM,
platelet endothelial cell adhesion molecule; RTPCR, reverse
transcriptionPCR; Sca-1, stem cell antigen-1; SSC, side scatter; VEGFR2,
vascular endothelial growth factor receptor 2.
1
To whom correspondence should be addressed (email
sanda.burlacu@icbp.ro).

C 2010 Portland Press Limited
200 A.-M. Rosca and A. Burlacu
cells, could be isolated by Percoll gradient centrifugation.
This subpopulation, containing cells with a density varying
between 1.067 and 1.070 g/ml, had an increased capacity
to generate haematopoietic cells under specific conditions
and, moreover, produced MSC cultures that were free of
haematopoietic cells in a shorter period compared with
those from bone marrow aspirate.
Materials and methods
Isolation of bone marrow cells
Adult RAP mice (8 weeks old) were killed by cervical
dislocation in accordance with the rules of the Institutional
Ethical Board for Experimental Procedures. The tibiae and
femurs were harvested, and the medullar channels were
flushed with 5 ml PBS-FBS [PBS containing 2% FBS (fetal
bovine serum)]. The single cell suspension obtained by
mechanical dissociation (passing the cells through needles
of decreasing gauges, i.e. 20, 22 and 25 gauge) was then
centrifuged at 400 g for 5 min at 20 C and resuspended
in 1 ml of DMEM (Dulbeccos modified Eagles medium;
Invitrogen). This volume of suspension was layered on the
Percoll gradient or analysed by flow cytometry.
Separation on the Percoll gradient
An isotonic Percoll solution was made by mixing Percoll (25
mOs/kg, 1.130 g/ml; SigmaAldrich) and 10 concentrated
DMEM at a 9:1 volume ratio. Densities of 1.050, 1.057,
1.067, 1.070, 1.076 and 1.083 g/ml were prepared by diluting
the isotonic Percoll solution in DMEM at concentrations
of 35, 40, 45, 50, 55 and 60% respectively. A six-layered
Percoll gradient was obtained by layering 2 ml of each
Percoll solution. Finally, the cell suspension obtained as
described above was placed on the top of the gradient and
centrifuged at 1500 g for 25 min at 4 C. Five fractions of
2 ml each were collected after centrifugation and designated
IV (see Figure 2A). The lower part of the gradient (3
ml), which contained the cells with a density higher than
1.076g/ml and included the erythrocytes, was collected as
fraction VI. The cells in each fraction were washed three
times in PBS-FBS and then either analysed or cultivated
in DMEM supplemented with 10% MSC-Qualified FBS
(Invitrogen).
In situ detection of progenitor cells
Tibiae were fixed in formalin and then decalcified in
5% trichloroacetic acid, dehydrated in a graded series
of alcohols and embedded in paraffin. Longitudinal
sections (5 m) on poly-L-lysine-coated glass slides were
deparaffinized and microwaved in citrate buffer for antigen
retrieval. Anti-mouse c-kit or Sca-1 antibody (R&D Systems)

C 2010 Portland Press Limited
was used at 10 g/ml. Incubation with the primary antibody
was followed by quenching in 3% H2 O2 and incubation
with horseradish peroxidase-conjugated anti-mouse IgG.
Ultimately, the proteins were visualized by the DAB
(diaminobenzidine) reaction and examined under a Nikon
microscope.
Flow cytometry assay
FACS was performed using a MoFlo Cell Sorter. Cells
from fractions IVI and bone marrow aspirate were
incubated with a phycoerythrin-labelled anti-c-kit or anti-
Sca-1 antibody (R&D Systems) for 1 h at 4 C, washed in
PBS-FBS and analysed. In order to exclude dead cells from
the analysis, 1 g/ml propidium iodide was added just before
the analysis. At least 20000 events were considered for each
sample. Results were analysed using Summit v4.3 software
(Cytomation, Inc).
RTPCR (reverse transcriptionPCR)
Total RNA was extracted using the RNeasy Micro kit (Qia-
gen), and cDNA was synthesized from 0.2 g of total RNA
employing MMLV Reverse transcriptase (Invitrogen). PCR
was performed using optimized amplification conditions, and
the products were visualized on a 1.5% agarose gel with
ethidium bromide staining. For each gene, DNA primers
were derived from different exons whenever possible, and
cDNAs were treated with DNase I to ensure that the PCR
product represented the specific messenger RNA species
and not genomic DNA. The primer sequences are available
upon request.
Colony-forming unit assay
The functionality of the haematopoietic stem/progenitor
cells within each fraction was estimated in HSC-CFU
complete with Epo medium (Miltenyi Biotec). This medium
allows the enumeration of haematopoietic stem and
progenitor cells characterized as CFU (colony forming units)
and has been reported to support the growth of granulocyte
(CFU-G), macrophage (CFU-M), granulocyte/macrophage
(CFU-GM) and erythroid [BFU-E (erythroid burst-forming
units) and CFU-E] colonies, as well as mixed colonies (CFU-
GEMM) [8]. Cells were cultivated at 1500 cells/cm2 in
duplicates and clusters of cells were scored after 7 days.
Adipogenic and osteogenic differentiation
Cells from each fraction were seeded into two wells of a
24-well plate. After 1 week in culture, the cells were
incubated for an additional week in either adipogenic
[DMEM with 10% FBS, 106 M dexamethasone, 100 M
indomethancin and 1% ITS (insulin-transferrin-sodium
selenite supplement)] [9] or osteogenic (DMEM with 10%
 Isolation of a progenitor-enriched population of bone marrow 201

Figure 1 Quantification and localization of c-kit+ and Sca-1+ cells in bone marrow

(A) Flow-cytometry quantification of c-kit+ and Sca-1+ cells in bone marrow. C-kit was identified on 11.8 +
2.5% of the bone marrow cells with varying granularity,
whereas Sca-1+ cells represented 9.8 + 2.8% of bone marrow cells and were almost exclusively characterized by low granularity. Diagrams are representatives of
four independent experiments, whose mean values are illustrated in the adjoined table. (B) Immunohistochemistry for in situ localization of c-kit+ (b and c) and
Sca-1+ (d, e and f) cells within bone marrow. Staining was completely absent in tissue sections in which the primary antibody was omitted (a). One can notice the
epiphysis as preferential site for Sca-1+ cells (df), and the diaphysis for c-kit+ cells (b and c).

FBS, 107 M dexamethasone, 10 mM -glycerophosphate Results

and 0.3 mM ascorbic acid) [6] differentiation medium.
All chemicals were from Sigma, unless otherwise stated.
Lipid droplets were stained with Oil Red (Sigma), and
lipid accumulation was quantified as previously described
[10] using the TECAN iGenios spectrofluorimeter. Calcium
deposits were stained by the von Kossa method [6].
Statistics
All quantitative results were expressed as the
means + S.E.M. from at least three experiments. Statistical
analysis was performed using one-way ANOVA. A p-value
less than 0.05 was considered significant.
Localization and quantification of c-kit- and
Sca-1-positive cells in mouse bone marrow
Quantification of progenitor cells from the bone marrow
aspirate was determined on the basis of c-kit and Sca-1
surface marker expression. Flow cytometry results revealed
that 11.8% of total bone marrow cells were c-kit+ cells
and 9.8% were Sca-1+ cells (Figure 1A). To localize their
preferential site within bone marrow, longitudinal sections
of decalcified tibiae were analysed by immunohistochemistry
(Figure 1B). c-kit+ cells were predominantly localized in
the diaphyseal region (Figure 1B, b and c), whereas Sca-
1+ cells were detected within the epiphyseal region of


C 2010 Portland Press Limited
202 A.-M. Rosca and A. Burlacu
Figure 2 Separation of bone marrow cells into six fractions by centrifugation on a discontinuous Percoll gradient
(A) Schematic illustration of the discontinuous Percoll gradient used for fractionation and the percentage of cells segregated in each fraction after centrifugation.
(B) Dot-plot histograms resulting from FACS analysis illustrating the heterogeneity of cells before separation and the morphology of cells in each fraction.
the bone (Figure 1B, df). These results indicate that
colonies of progenitor cells can be found in both the
epiphysis and diaphysis, but differ in their expression of cell
surface markers. The segregation of the two populations
of progenitor cells in different sites within bone marrow
ruled out the existence of c-kit+ /Sca-1+ double-positive
cells in this tissue, indicating that the large majority of
progenitor cells were either Sca-1 or c-kit single-positive
cells. Our results corroborate other studies that revealed
very low percentages of Sca-1+ /c-kit+ double-positive cells
within bone marrow [11,12].
Separation of bone marrow cells on a Percoll
gradient
Because c-kit+ and Sca-1+ cells have a low level of granularity
[as revealed by low SSC (side scatter), see Figure 1A], we
considered the possibility of exploiting physical properties
in order to separate these cells. Centrifugation of cells on
a discontinuous Percoll gradient (ranging from 1.050 to
1.083 g/cm3 ) resulted in the recovery of six cell fractions
(IVI), corresponding to the six densities of the gradient.
The total number of cells recovered in the six fractions
was 35 106 + 1.4 10 cells, meaning a recovery rate
6
+
of 70% 2.8% (considering an average of 50 106 cells

C 2010 Portland Press Limited
harvested from one mouse). The segregation of cells within
each fraction is represented in Figure 2(A) as a percentage
of the total amount of cells recovered after separation.
The counting revealed that 0.58%, 0.62%, 1.73%, 11.68%,
66.32% and 19.08% of cells were segregated in fractions
IVI respectively, meaning 2 105 cells in fraction I, 2.2 105
in fraction II, 4.2 105 in fraction III, 4.3 106 in fraction
IV, 23.7 106 in fraction V, and 6.8 106 in fraction VI.
Erythrocytes residing mainly in fraction VI were excluded
from the counting. Thus the majority of cells (66.32%)
segregated in fraction V, corresponding to the density
range 1.0701.076 g/ml, whereas fraction IV contained
only a minor population of cells (11.68%). As expected,
flow cytometry analysis of each subpopulation revealed an
increase in cell granularity [FSC (forward scatter)/SSC] with
increasing Percoll density (Figure 2B).
Quantification of the progenitor cells within
fractions
Cells within each fraction were analysed for the presence of
c-kit+ and Sca-1+ progenitor cell markers by flow cytometry.
As shown in Figure 3(A), the distributions of both c-kit+ and
Sca-1+ progenitor cells were confined to fractions IIIV and
were not present in fractions I, II and VI. The quantification
 Isolation of a progenitor-enriched population of bone marrow 203

Figure 3 Depiction of c-kit+ and Sca-1+ cells within fractions

(A) Flow cytometry diagrams revealing the percentage of c-kit+ and Sca-1+ cells within each fraction obtained after Percoll gradient separation of bone marrow
aspirate. The progenitor cells are mainly found in fractions III, IV and V. The diagram is representative of three experiments, and the average percentages are given
in the bar charts below. (B) The overall representation of flow cytometry data showing the enrichment factor for c-kit+ and Sca-1+ cells in the six fractions relative
to bone marrow aspirate (unfractionated cells). The values denote the average of three independent experiments after extraction of the unspecific binding values.
Note that fractions III and IV show the greatest enrichment for progenitor cells, illustrating 2.7- and 2.4-fold for c-kit+ cells and 1.9- and 2.2-fold for Sca-1+ cells, in
fraction III and IV respectively.

of these cells is given in Figure 3(A) as purity of each fraction aspirate. Analysis of fraction V, which contained the majority
and in Figure 3(B) as enrichment factor for c-kit+ cells (left) of the cells (66.32% of the total bone marrow cells; see
and Sca-1+ cells (right) as compared with bone marrow Figure 2A), revealed no significant increase in the percentage


C 2010 Portland Press Limited
204 A.-M. Rosca and A. Burlacu

Figure 4 Evaluation of the clonogenic potential of the haematopoietic cells within the fractions

(A) Evaluation of differentiation potential within each fraction using the CFU assay. The diagram illustrates the average value of three independent experiments.
Note that the highest capacity of differentiation resides in fractions III and IV (*P< 0.05). A representative field of fraction IV was photographed and displayed
above the diagram. (B) RTPCR illustrating the presence of representative cell types within each fraction immediately after separation (left panel) and after two
weeks in culture (right panel). As a control, c-kit+ and Sca-1+ cell populations isolated by the FACS method are given in the middle panel. (C) Lower panel:
spectrophotometric quantification of lipid accumulation in the six fractions after culturing the cells in adipogenic medium. The upper panel depicts a representative
picture of Oil Red staining in fractions III and IV. Right, light microscopy of von Kossa staining in the six fractions after cultivation in osteogenic medium. Note than
fractions III and IV, but not fraction V (that contained the majority of the cells), were able to generate osteoblasts in appropriate conditions.

of progenitor cells. Fraction V contained 14 + 1.6% c-kit

+
increase in c-kit+ cells and a 2.17 + 0.3-fold increase
+ +
and 8 4.6% Sca-1 cells as compared with 11.8 2.5% + in Sca-1+ cells in fraction IV as compared with the
c-kit+ and 9.8 + +
2.8% Sca-1 cells in the total bone marrow unfractionated bone marrow cell population. Thus fraction
cell population. In contrast, fractions III and IV contained IV contained 4.3 106 cells, of which approx. 2 106 cells
important enrichments for progenitor cells (either c-kit+ or were progenitors.
Sca-1+ ). Despite the enrichment factor of progenitor cells in To evaluate the clonogenic potential of the haemato-
fraction III (2.74 +
0.1 for c-kit cells and 1.9 +
+
0.5 for Sca-1
+
poietic cells within fraction IV, the colony-forming assay
cells), the total amount of progenitor cells in fraction III was was performed on the cell fractions. To this purpose, 1.2
inconsistent due to the low number of cells segregated in 104 cells from each fraction were plated in methylcellulose
this fraction (1.73%). Thus fraction III contained 1.8 105 medium, and the cell clusters generated were scored after
progenitor cells out of 4.2 105 cells recovered in this 7 days in culture. As illustrated in Figure 4(A, upper panel),
fraction. the colonies varied in size, probably reflecting the differences
In contrast, 23 + 4.8% of cells in fraction IV were in the proliferation rate and/or differentiation ability of the
c-kit+ and an additional 22 + +
2.3% were Sca-1 as compared various progenitor cells within the fractions. Nevertheless,
with the unfractionated population (11.8% c-kit+ cells and fractions III and IV generated the greatest number of CFU
9.8% Sca-1+ cells). These results suggested a 2.36 + 0.7-fold (Figure 4A). These fractions generated approx. 8.4 + 2.3 and

C 2010 Portland Press Limited

7.8 + 3
0.2 CFU per 10 cells, which was 2-fold higher than
the major fraction, V (3.9 + 3
1.9 CFU per 10 cells). This
result indicates enrichment in haematopoietic progenitors
in fractions III and IV. Conversely, fractions I, II and VI
generated fewer CFU, as could be predicted from the
decreased number of c-kit+ cells and Sca-1+ cells scored
in those fractions.
Expression of differentiated cell markers in the
fractions
To better characterize the content of fraction IV, the
distribution of the main differentiated cell types that are
normally present within bone marrow was evaluated in each
fraction by RTPCR and the results were compared with
c-kit+ cells and Sca-1+ cells purified using FACS. As
illustrated in Figure 4(B, left panel), besides being enriched
in the expression of c-kit and Sca-1, fractions III and IV
showed the highest expression level of the endothelial
markers VEGFR2 (vascular endothelial growth factor
receptor 2) and PECAM (platelet endothelial cell adhesion
molecule) of all six fractions. Furthermore, fraction IV
demonstrated a low level expression of differentiation
markers of other cell types, such as adipocytes (fabp4),
osteoblasts (osteocalcin) and chondrocytes (collagen),
which segregated mainly within fractions IIII. As expected,
the FACS method generated more purified cell populations
that were devoid of other cell types.
After 2 weeks in culture, adherent cells were obtai-
ned from all fractions (results not shown). Specifically,
cells from fraction IV generated a culture of cells with a
fibroblast-like morphology that were distinct from those
derived from fractions III and V, mainly by lacking CD45
mRNA expression (Figure 4B, right panel). Furthermore,
adherent cells in fraction IV expressed Sca-1 and were
not contaminated with haematopoietic cells (as revealed by
the absence of CD45 and c-kit mRNA) or differentiated
cells, such as endothelial cells (PECAM/CD31 and
VEGFR2/Flk-1), osteoblasts (osteocalcin) or chondrocytes
(collagen II).
To evaluate which cell fractions could generate
MSC cultures, the capacity of the cells to differentiate
into adipocytes and osteoblasts was evaluated for each
cell fraction after 1 week of cultivation. Although all
fractions were able to generate adipocytes under the
appropriate conditions, spectrophotometric quantification
of lipid accumulation (Figure 4C, lower panel) suggested
a greater capacity for fractions IIIV as compared with
the other fractions. Nevertheless, the capacity to generate
osteoblasts was limited to cells within fractions III
and IV (Figure 4C, right panel), and fraction V lacked
differentiated cells. These results suggest that fractions III
and IV, but not fraction V, generated multipotent cells
Isolation of a progenitor-enriched population of bone marrow 205
during cultivation. Altogether, our results demonstrate that
fraction IV generated a cell culture that was depleted in
haematopoietic cells, negative for osteoblast, chondrocyte
and endothelial markers, yet possessed a multipotent
capacity. These CD45 /c-kit /Sca-1+ /CD31 /Flk-1 cells are
similar to bone marrow MSCs, which can be obtained
from bone marrow aspirate after several weeks in culture
[13,14].
In conclusion, a Percoll gradient can be used to
obtain a cell fraction that is enriched in progenitors and
can generate an MSC culture after a short period of
cultivation. This fraction is enriched in both haematopoietic
and mesenchymal progenitor cells (c-kit+ and Sca-1+ cells),
whereas it is deprived of other differentiated cell types, and
also contains cells that are positive for endothelial markers,
which further increases its value in cellular transplant studies
where angiogenesis is required.
Discussion
The existence of stem/progenitor cells within the hetero-
geneous population of bone marrow is well documented
[15]. Haematopoietic stem cells are the most primitive
cells that retain multilineage haematopoietic differentiation
potential. In adult bone marrow, most haematopoietic
progenitors express c-kit (CD117), a member of a family
of cell-surface receptors with tyrosine kinase activity
that plays an important role in the regulation of the
early stages of haematopoietic development [16]. c-kit is
expressed in both long-term repopulating HSCs and more
differentiated progenitor cells and is used as a valuable
marker for the identification of these cells [17]. Although
c-kit is an established marker for HSCs, the expression
of Sca-1, a member of the Ly6 family, in these cells is
controversial. To date, Sca-1 expression has been identified
in putative stem/progenitor cell populations within various
tissues and organs including bone marrow, kidney, liver,
lung, pancreas and muscle, as well as the aorta-gonad-
mesonephros, yolk sack and other regions of the embryo
[18]. However, whether these Sca-1+ populations are tissue-
specific precursor/stem cells or represent haematopoietic,
mesenchymal or endothelial precursor/stem cells associated
with these tissues is not known in all cases [18].
We show in the present paper that c-kit+ and Sca-1+
cells segregated within bone marrow with respect to
location, thus indicating that the large majority of progenitor
cells were either Sca-1 or c-kit single-positive cells. This
was also strengthened by the RTPCR analysis on FACS-
separated c-kit+ and Sca-1+ populations that revealed the
absence of c-kit mRNA in a Sca-1+ population, even if
some c-kit+ cells do express Sca-1 at the mRNA level
(Figure 4B). In accordance with these results, very low

C 2010 Portland Press Limited
206 A.-M. Rosca and A. Burlacu
percentages of Sca-1+ /c-kit+ double-positive cells within
bone marrow have previously been quantified by FACS
analysis [11,12]. For example, less than 2% out of 20%
c-kit+ cell populations and 10% Sca-1+ cell populations
respectively has been found to double-stain for c-kit and
Sca-1 in bone marrow mononuclear cells [12]. Another
study reported that only 0.2% of Lin cells in bone marrow
were Sca-1+ /c-kit+ double-positive [11].
Because c-kit+ and Sca-1+ cells have low levels
of granularity, we considered the possibility of using
this physical property in order to separate these cells.
Our study showed that a bone marrow subpopulation
that is enriched in Sca-1+ and c-kit+ cells can be
obtained through density gradient centrifugation and has
the ability to differentiate under specific conditions. This
fraction can be a source for MSCs (CD45 /c-kit /Sca-1+ /
PECAM /VEGFR2 /osteocalcin /collagenII cells) that can
be used in differentiation and cellular transplantation studies.
The advantage of using a Percoll gradient to enrich
for cell populations
Currently, various techniques based on the physical and
immunochemical characteristics of stem/progenitor cells are
used for their enrichment. Among these, FACS or magnetic
bead separation, which are based on cell surface antigen
expression, are most commonly used for either positive or
negative selection [19,20]. These immunochemical methods
present the disadvantage of being time-consuming and
expensive because of the use of monoclonal antibodies
and advanced technology. Although more purity is given
by immunological assays in isolating cells that express a
certain surface marker, the resulting cell populations are
still heterogeneous. That is because no single specific
stem cell marker has been identified yet. Likewise, stem
cells are known to express a wide variety of antigens
expressed by many other cell types, and these antigens
are variably expressed depending on cellular status. On
the other hand, mesenchymal stem cell cultures isolated by
immunodepletion of haematopoietic cells have been found
to exhibit poor growth mainly due to the absence of other
cells that normally supported their growth by secreting
paracrine factors. Furthermore, our trials to cultivate a
pure population of c-kit+ and Sca-1+ cells isolated by the
FACS method also failed due to the absence of survival
factors provided by supporting cells (results not shown).
Thus physical separation of bone marrow cells remains a
good alternative to immunochemical methods.
Since its introduction in 1977, Percoll has been used
to isolate several cell types. Its nearly ideal physical
characteristics make it especially useful as a first step
in the enrichment of cell populations before attempting
higher resolution. With the use of this method, several

C 2010 Portland Press Limited
studies showed a successful enrichment of haematopoietic
progenitor cells that were able to repopulate lethally
irradiated mice. Nijhof and Wierenga isolated cells with
a diameter between 7.2 and 8.4 m, which sediment
at a density of 1.065 g/ml, from spleen; these cells had
a high capacity to repopulate lethally irradiated mice
with all haemopoietic precursor cells [21]. In accordance
with that study, our flow cytometry results showed a
preferential distribution of c-kit and Sca-1 progenitor
cell markers in a population characterized by a density
between 1.057 and 1.076 g/ml. By using a discontinuous
gradient centrifugation, mouse bone marrow cells were
separated into six fractions (corresponding to 1.050, 1.057,
1.067, 1.070, 1.076 and 1.083 g/ml), and progenitor cells
were significantly enriched within fractions IIIV. In any
case, fraction IV was particularly important because it
harboured approx. 2 106 stem/progenitor cells out of a
total of 4.3 106 cells. This population is highly enriched
in stem/progenitor cells compared with the initial bone
marrow population and is valuable in various experiments,
including cellular transplantation studies, which usually need
between several thousand and two million cells.
Fraction IV as a source of HSCs
There is always a compromise in choosing the conditions for
isolating HSCs between those that provide the highest yield
and those that provide the highest content of progenitor
cells. As a matter of choice, two techniques, namely
elutriation and labelling with a selective progenitor marker,
are being differentially used for this purpose.
The method described in the present paper provides an
alternative for the isolation of a haematopoietic-progenitor-
cell-enriched population from bone marrow, in which the
physical properties of the progenitor cells are exploited. The
cells in fraction IV, even though they are still heterogeneous,
contain a superior amount of haematopoietic progenitors,
namely 45% of the population, and furthermore, a high
amount of endothelial progenitor cells. Besides these
haematopoietic progenitor cell populations, this fraction
still harbours several other cell types, among which are
chondroblasts and adipocytes. As a proof of its improvement
for HSCs, this subpopulation is able to generate an increased
number of haematopoietic colonies when cultivated under
the appropriate conditions. No prediction can be made
about the short- or long-term reconstituting properties of
these cells, as they are morphologically indistinguishable
and such discrimination can be only based on in vivo
functional assays. It is likely that a number of cells have
reconstituting abilities (as we know that such cells reside
within bone marrow); however, we did not perform in vivo
experiments and consequently we have no evidence on this
issue.

Fraction IV as a source of MSCs
MSCs are usually isolated on the basis of their adherence
to plastic surfaces. Based on this property, MSCs from
different species have been obtained and proven to be
able to expand in culture while retaining their multipotent
differentiation capacity. Mouse MSC cultures are frequently
contaminated by haematopoietic cells; hence, these cells are
more difficult to isolate and expand as compared with MSCs
from other species [13]. Several approaches for isolating
pure populations of mouse MSCs have been reported. For
instance, immunodepletion of haematopoietic precursors
has been used as a first purification step of bone marrow
aspirate in order to obtain a more homogenous culture
of MSCs [19,20]. However, the cells have been found to
exhibit poor growth due to the dramatic down-regulation
of many genes involved in cell proliferation and cell cycle
progression as a result of immunodepletion [19]. As an
alternative to immunodepletion, mouse MSC cultures can
be obtained from diverse strains by incubating total bone
marrow cells in a high-density culture with medium that
inhibits the growth of haematopoietic cells. In this approach,
the cells are grown under the aforementioned conditions
for the first two passages (56 weeks), after which the most
easily detached cells are expanded in a low-density culture
[21]. In another reported strategy, cells are plated at 106
cells/cm2 and the medium in the primary culture is changed
frequently concomitant with a diminished trypsinization time
[22]. In this procedure, a purified population of MSCs can
be obtained 3 weeks after the initiation of culture (two
passages). In the method described in the present study, a
similar pure MSC population (containing CD45 /c-kit /Sca-
1+ /PECAM /VEGFR2 /osteocalcin /collagenII cells) can be
obtained by cultivating cells from fraction IV for no more
than 1 week in DMEM supplemented with 10% MSC-
qualified FBS.
In conclusion, we designed a method for the
fractionation of mouse bone marrow through the use
of a Percoll gradient that resulted in a progenitor-
cell-enriched population. This population of small cells
segregated at a density of 1.070 g/ml and contained both
mesenchymal and haematopoietic progenitors. These cells
were able to generate more haematopoietic colonies under
the appropriate conditions as compared with total bone
marrow. They were also able to produce an MSC line that
was free of haematopoetic cells in a short time in culture as
compared with classical methods.
Acknowledgements
We thank Eugen Andrei, Ioana Manolescu and Dr Emanuel
Dragan for their technical assistance.
Isolation of a progenitor-enriched population of bone marrow 207
Funding
This work was supported by grants from the Romanian
Ministry of Education and Research, IDEA Program [grant
number 250/2007] and COST (European Co-operation in
Science and Technology) Action BM0602.
References
1 Pittenger, M. F., Mackay, A. M. and Beck, S. C. (1999) Science
284, 143147
2 Nervi, B., Link, D. C. and DiPersio, J. F. (2006) J. Cell. Biochem.
99, 690705
3 Morrison, S. J. and Spradling, A. C. (2008) Cell 132,
598611
4 Friedenstein, A. J., Kailakhyan, R. K., Latsinik, N. V., Panasyuk,
A. F. and Keiliss-Borok, I. V. (1974) Transplantation 17,
331340
5 Jiang, Y., Jahagirdar, B. N., Reinhardt, R. L., Schwartz, R. E.,
Keene, C. D., Ortiz-Gonzalez, X. R., Reyes, M., Lenvik, T., Lund,
T., Blackstad, M. et al. (2002) Nature 418, 4149
6 Tropel, P., Noel, D., Platet, N., Legrand, P., Benabid, A. L. and
Berger, F. (2004) Exp. Cell Res. 295, 395406
7 Baddoo, M., Hill, K., Wilkinson, R., Gaupp, D., Hughes, C.,
Kopen, G. C. and Phinney, D. G. (2003) J. Cell. Biochem. 89,
12351249
8 Hoffman, R., Benz, E., Shattil, S., Furie, B. and Cohen, H. (2008)
Hematology: basic principles and practice, 5th ed., Churchill
Livingstone, Philadelphia
9 Chamberlain, G., Wright, K., Rot, A., Ashton, B. and Middleton,
J. (2008) PLoS One 3, e2934e2939
10 Glassford, A. J., Yue, P., Sheikh, A. Y., Chun, H. J., Zarafshar, S.,
Chan, D. A., Reaven, G. M., Quertermous, T. and Tsao, P. S.
(2007) Am. J. Physiol. Endocrinol. Metab. 293, E1590E1596
11 Calvi, L. M., Adams, G. B., Weibrecht, K. W., Weber, J. M.,
Olson, D. P., Knight, M. C., Martin, R. P., Schipani, E., Divieti, P.,
Bringhurst, F. R. et al. (2003) Nature 425, 841846
12 Yamada, Y., Nishimotoa, E., Mitsuyaa, H. and Yonemurab, Y.
(2006) Exp. Hematol. 34, 97106
13 Peister, A., Mellad, J. A., Larson, B. L., Hall, B. M., Gibson, L. F.
and Prockop, D. J. (2004) Blood 103, 16621668
14 Anjos-Afonso, F. and Bonnet, D. (2008) Curr. Protoc. Stem
Cell. Biol. Chapter 2: Unit 2B.3.
15 Zipori, D. (2005) Stem Cells 23, 719726
16 Perez, C., Moreno, S., Summerfield, A., Domenech, N., Alvarez,
B., Correa, C., Alonso, F., Ezquerra, A., Domnguez, J. and
Revilla, C. (2007) J. Immunol. Methods 321, 7079
17 Briddell, R. A., Broudy, V. C., Bruno, E., Brandt, J. E., Srour, E. F.
and Hoffman, R. (1992) Blood 79, 31593167
18 Holmes, C. and Stanford, W. L. (2007) Stem Cells 25,
13391347

C 2010 Portland Press Limited
208 A.-M. Rosca and A. Burlacu

19 Baddoo, M., Hill, K., Wilkinson, R., Gaupp, D., Hughes, C., 22 Soleimani, M. and Nadri, S. (2009) Nature Protocols 4,
Kopen, G. C. and Phinney, D. G. (2003) J. Cell. Biochem. 89, 102106
12351249
20 Gronthos, S., Graves, S. E., Ohta, S. and Simmons, P. J. (1994)
Blood 84, 41644173 Received 2 December 2009/12 March 2010; accepted 23 March 2010
21 Nijhof, W. and Wierenga, P. K. (1985) Eur. J. Cell. Biol. 39, Published as Immediate Publication 23 March 2010, doi:10.1042/BA20090356

136141


C 2010 Portland Press Limited