HEMATOPOIESIS
Soluble factor cross-talk between human bone marrow-derived hematopoietic and
mesenchymal cells enhances in vitro CFU-F and CFU-O growth and reveals
heterogeneity in the mesenchymal progenitor cell compartment
Dolores Baksh, John E. Davies, and Peter W. Zandstra
The homeostatic adult bone marrow (BM)
is a complex tissue wherein physical and
biochemical interactions serve to maintain a balance between the hematopoietic
and nonhematopoietic compartments. To
focus on soluble factor interactions occurring between mesenchymal and hematopoietic cells, a serum-free adhesion-independent culture system was developed that
allows manipulation of the growth of both
mesenchymal and hematopoietic human
BM-derived progenitors and the balance
between these compartments. Factorial
experiments demonstrated a role for stem
cell factor (SCF) and interleukin 3 (IL-3) in
the concomitant growth of hematopoietic
(CD45ⴙ) and nonhematopoietic (CD45ⴚ)
cells, as well as their derivatives. Kinetic
tracking of IL-3␣ receptor (CD123) and
SCF receptor (CD117) expression on a
sorted CD45ⴚ cell population revealed the
emergence of CD45ⴚCD123ⴙ cells capable of osteogenesis. Of the total fibroblast colony-forming units (CFU-Fs) and
osteoblast colony-forming units (CFU-O),
approximately 24% of CFU-Fs and about
22% of CFU-Os were recovered from this
population. Cell-sorting experiments dem-
onstrated that the CD45ⴙ cell population
secreted soluble factors that positively
affect the survival and proliferation of
CFU-Fs and CFU-Os generated from the
ⴚ
CD45 cells. Together, our results provide
insight into the intercellular cytokine network between hematopoietic and mesenchymal cells and provide a strategy to
mutually culture both mesenchymal and
hematopoietic cells in a defined scalable
bioprocess. (Blood. 2005;106:3012-3019)
© 2005 by The American Society of Hematology
Introduction
Hematopoiesis in the adult bone marrow (BM) is governed by a
complex interplay between soluble regulatory molecules, cell-tocell contact, and interactions with extracellular matrix components.1-6 Although self-regulation within the hematopoietic niche is
undoubtedly critical, it has also been clearly demonstrated that
hematopoiesis is directly regulated by mesenchymal stem/
progenitor cells (MPCs) and their descendents.7,8 Despite the role
that MPCs play in hematopoiesis, much less is understood about
MPC regulation in vivo by hematopoietic cells (HCs), largely due
to difficulties in prospectively identifying putative MPCs in their
native microenvironment. MPCs can be detected in vitro using
retrospective assays, such as the fibroblast colony-forming unit
(CFU-F) assay.9-11 Interestingly, despite observations suggesting a
role for HCs in MPC development,12,13 HCs are commonly
regarded as “contaminants” in MPC culture and are typically
discarded during MPC propagation.
“Dexter-type” cultures were designed to recapitulate both the
HC and non-HC compartments of the BM microenvironment.14,15
However, limited control of the dynamic balance between HCs and
non-HCs, as well as the inability of Dexter-type cultures to support
long-term hematopoietic and nonhematopoietic progenitor cell
growth,14,16 limits the suitability of this culture system for studying
the regulation of mesengenesis by HCs. Accordingly, an in vitro
model capable of maintaining the developmental potential of both
BM-derived mesenchymal cells (MCs), including MPCs, and HCs,
including hematopoietic progenitor cells (HPCs), would be ideal
for studying the soluble factor interplay between these cell systems.
Understanding this interaction may provide insight into the biologic mechanisms governing MPC regulation.
We previously reported an approach to grow adult BM-derived
HPCs17 and MPCs as single cells in adhesion-independent, stirred
suspension cultures (SSCs).18 This approach supports the growth of
both populations and allows the quantitative examination of the
role of isolated (in the absence of adhesion-mediated signaling)
soluble factor interactions in the interplay between the HC and MC
compartments. Our studies revealed that the addition of specific
growth factors to the serum-containing culture medium influenced
the survival and growth of both MPCs and HPCs in suspension.
Herein, we extend our work to demonstrate that SSC-grown cells
maintain the ability to differentiate along multiple mesenchymal
lineages, including fibroblastic adipogenic, osteogenic, and myogenic cells, and that such multilineage progenitors can proliferate
in SSCs. Furthermore, using high-content screening approaches,
we remove the confounding influences of serum and identify
From the Institute of Biomaterials and Biomedical Engineering, Department of
Chemical Engineering and Applied Chemistry, Faculty of Dentistry, University
of Toronto, Toronto, ON, Canada.
assisted in experimental design and interpretation, as well as manuscript
writing and editing. P.W.Z. assisted in experimental design and interpretation,
as well as manuscript writing and editing.
Submitted February 10, 2005; accepted July 2, 2005. Prepublished online as
Blood First Edition Paper, July 19, 2005; DOI 10.1182/blood-2005-01-0433.
Reprints: Peter W. Zandstra, Institute of Biomaterials and Biomedical
Engineering, Rm 407, Roseburgh Bldg, 4 Taddle Creek Rd, Toronto, ON, M5S
3G9, Canada; e-mail: peter.zandstra@utoronto.ca.
Supported by operating grants from the Natural Sciences and Engineering
Research Council (NSERC) of Canada (P.W.Z.) and the Ontario Research and
Development Challenge Fund (J.E.D.). D.B. was a recipient of a NSERC
postgraduate scholarship and held an Ontario Graduate Scholarship. D.B.
designed and performed research, analyzed data, and wrote the paper. J.E.D.
3012
The publication costs of this article were defrayed in part by page charge
payment. Therefore, and solely to indicate this fact, this article is hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. section 1734.
© 2005 by The American Society of Hematology
BLOOD, 1 NOVEMBER 2005 䡠 VOLUME 106, NUMBER 9
BLOOD, 1 NOVEMBER 2005 䡠 VOLUME 106, NUMBER 9
soluble factor combinations that influence the growth of MPCs in
serum-free (SF) SSCs. This SF-SSC system was then used to study
the mechanism by which factors, directly or indirectly secreted by
HCs, influence MPC growth. Our analyses revealed that supplementation of interleukin 3 (IL-3) is an effective mediator of MPC
survival and, when supplemented with stem cell factor (SCF),
supports the expansion of MPCs in SSCs. SCF did not influence the
growth and expansion of MPCs, but the effects observed when
combined with IL-3 suggest that SCF may act on HCs, stimulating
the secretion of as yet undetermined soluble factors capable of
modulating MPC responses, including survival and growth responses. Kinetic tracking of IL-3␣ receptor (CD123) and SCF
receptor (CD117) expression on the CD45⫺ cell population revealed the culture type-specific emergence of CD45⫺CD123⫹ cells
capable of forming a functional bone phenotype. Cell-sorting
studies demonstrated a supportive role for CD45⫹ cells on MPC
growth during suspension culture. These findings provide insight
into the regulatory role that HCs (via their endogenous release of
factors) have on the survival and growth of MPCs in vitro.
Materials and methods
Progenitor cell assays
To test the developmental repertoire of BM-derived cells grown under
adhesion-independent SSC conditions, the following functional assays
were performed. In all assay systems, SSCs were initiated according to the
methods described by Baksh et al.18 CFU-F and osteoblast colony-forming
unit (CFU-O) assays were performed according to the methods previously
described.18
Adipogenesis. Suspension cells were removed from SSCs supplemented with SCF and IL-3 at 7, 14, and 21 days and plated at a density of
1 ⫻ 104 cells/cm2 in adipogenic-inducing medium containing Dulbecco
modified Eagle medium (DMEM; high glucose, 4.5 g/L), 1 ␮M dexamethasone, 0.5 mM indomethacin, 10 g/mL insulin, 100 mM 3-isobutyl-1methylxantine (all from Sigma, St Louis, MO) and 10% fetal bovine serum
(FBS). Cells were cultured for 3 weeks and then stained with Oil Red O
(Sigma). The expression of the adipogenic-specific gene, peroxisome
proliferator-activated receptor ␥-2 (PPAR␥-2) was analyzed on day 21 by
reverse transcription-polymerase chain reaction (RT-PCR).
Myoblast-CFU assay. Cell suspensions harvested from adhesionindependent SSCs supplemented with either SCF plus IL-3, SCF plus IL-3
plus 20 ng/mL platelet-derived growth factor (PDGF-BB) or without
cytokines were prepared at a concentration of 1 ⫻ 104 cells/cm2 and plated
in 35-mm tissue culture dishes. The cells were maintained for 2 weeks in
MCDB custom-made molecular and developmental biology 120 medium
(provided by Dr J. Tremblay, Centre de Recherche en Génétique Humaine,
Montreal, QC, Canada) containing 15% FBS, after which the serum level
was decreased to 2% for a further 7 days. At termination, cultures were
incubated with purified mouse antidesmin (1:40 in phosphate-buffered
saline [PBS], BD PharMingen, San Diego, CA) for 1 hour, washed 3 times
with PBS, and incubated with the secondary antibody (Alexa Fluor 488 goat
antimouse antibody [1:200, Molecular Probes, Eugene, OR]) for 1 hour,
washed again with PBS, and resuspended in 0.1 M sodium cacodylate
buffer (pH 7.4). Positive desmin colonies were imaged and enumerated at
⫻ 4 (4 ⫻/0.1 NA objective lens) using an inverted phase microscope.
Serum-free–small-scale suspension culture configuration
To screen multiple growth factors and growth factor combinations, a
small-scale suspension culture configuration was implemented. BM samples
were obtained from healthy donors (n ⫽ 3) and fractionated on a FicollPaque gradient (Sigma) as previously described.18 Approval was obtained
from the University of Toronto Institutional Review Board for these studies.
Informed consent was provided according to the Declaration of Helsinki.
BM-derived mononuclear cells (BM MNCs) were cultured at 37°C with 5%
CROSS-TALK BETWEEN HUMAN BONE MARROW COMPARTMENTS
3013
humidified CO2 in StemSpan medium (StemCell Technologies, Vancouver,
BC, Canada), in polystyrene 6-well plates (Falcon, Becton Dickinson
Labware, Franklin Lake, NJ) with and without the addition of cytokines.
The 6-well plates were placed on a shaker platform (Hoeffer, San Francisco,
CA) rotating at low speed. The cultures were fed according to the methods
previously described.18 Cell suspensions were assayed on days 0, 7, 14, and
21 for CFU-F and CFU-O potential.
Factorial experiments and analysis
A 2-level factorial design (FD) matrix was constructed as described by Box
et al19 to evaluate the effects of basic fibroblast growth factor (FGF),
PDGF-BB, SCF, and IL-3 and their respective 24 possible combinations on
total cell, CFU-F, and CFU-O expansion. In this study, optimal (high: 20
ng/mL) or zero dose levels (0 ng/mL) for PDGF and FGF were chosen
based on previous studies.13,20 A dose level of 100 ng/mL and 20 ng/mL was
chosen for SCF and IL-3, respectively. Each of the 16 cytokine combinations was tested on 3 independent sources of cells in triplicate. All statistical
analyses of the FD experiments was performed using SPSS 11.0 (SPSS,
Chicago, IL). A univariate analysis of variance (ANOVA) was performed
and the significance level of each combination was determined by
performing F tests. The parameter values of the general linear model19
(which contains parameter coefficients denoting the affects (␤) of each
cytokine and cytokine combination on the expansion of MPCs as assayed
by CFU-F and CFU-O development) were estimated by regression analysis;
significance (P ⬍ .05) was determined by performing F tests. To satisfy the
statistical assumption of normality all responses analyzed were logtransformed. Box plots constructed with the log-transformed data confirmed data symmetry.
SF suspension cultures initiated with unsorted, CD45ⴙ, or CD45ⴚ
cells. To study the regulation of MPC growth in suspension and eliminate
confounding influences attributed to the presence of HCs, the CD45⫺ cell
fraction of BM MNCs was isolated. Specifically, BM MNCs (n ⫽ 3) were
incubated with saturating concentrations (1:100) of conjugated mouse
IgG1␬ anti–human CD45-fluorescein isothiocyanate (FITC; BD Biosciences, Mississauga, ON, Canada) and 7-amino-actinomycin D (7-AAD;
viability dye; BD Biosciences). Both viable CD45⫺ and CD45⫹ cells were
sorted on a Beckman-Coulter (Hialeah, FL) EPICS Cell Sorter at a rate of
2000 to 3000 cells/sec at 10.5 psi to 99.5% purity. SSCs (n ⫽ 3, for each
group; Baksh et al18) were initiated with unsorted, CD45⫺, and CD45⫹ cells
and cultured in StemSpan medium supplemented with 20 ng/mL IL-3 and
100 ng/mL SCF. Control suspension cultures (no cytokine addition) were
initiated in parallel. At 7, 14, and 21 days, live cells were determined by
trypan blue exclusion and counted using a hemocytometer. An aliquot of
cells from each condition were initiated in CFU-F and CFU-O assays
(n ⫽ 3, in triplicate).
Flow cytometric sorting of suspension-derived CD45ⴚCD117ⴙ and
CD45ⴚCD123ⴙ cells. On days 7, 14, and 21, cells from SSCs (no cytokine
and SCF plus IL-3 treatment groups) that were initiated with CD45⫺ cells
on day 0 were incubated with saturating concentrations of mouse IgG1␬
anti–human CD45-FITC and CD123-phycoerythrin (PE; IL-3␣ receptor) or
CD45-FITC and CD117-PE (SCF receptor; BD Biosciences) to recover
populations of CD45⫺CD123⫹ and CD45⫺CD117⫹ cells on a BeckmanCoulter EPICS Cell Sorter at a rate of 2000 to 3000 cells/sec at 10.5 psi to
99.5% purity. The CFU-F and CFU-O developmental potential of these
sorted cell populations were assayed as described.
Results
Suspension-cultured cells retain multilineage
developmental capacities
The developmental potential of the suspension-grown cells was
assessed in functional assays. First, the osteogenic potential of
suspension-grown cells was assessed in a CFU-O assay.12,21
Imaging of tetracycline-labeled cultures21 was used to quantify
3014
BAKSH et al
bone nodule colonies (Figure 1A) and to determine the CFU-O
expansion over the time period studied. Bone nodules were also
confirmed by alkaline phosphatase, procollagen Ia, osteopontin,
and osteocalcin gene expression.18 When cells derived from SSCs,
incubated with SCF plus IL-3, were maintained in adipogenic
culture conditions for 3 weeks, cells containing fat globules (Figure
1B) were observed that expressed mRNA transcripts for PPAR␥-2
(Figure 1C). We also observed other phenotypes when our suspension-derived cells were grown under CFU-F assay conditions. For
example, when suspension cells derived from the SCF plus IL-3
plus 20 ng/mL PDGF-BB group were plated in CFU-F assays, in
addition to discrete colonies of fibroblastic-like cells (Figure 1D),
smaller cells with thin cellular protrusions (Figure 1E) and
expressing NeuN22 (not shown) were also observed.
Myogenic (myoblast-CFU [CFU-M]) developmental potential
was also assayed. Plated cells were stained with Giemsa to
visualize the colonies generated (Figure 1F), and staining the cells
with desmin resulted in 73% ⫾ 10% of the colonies labeling
positive (Figure 1G). Kinetic analysis of the number of desminpositive colonies revealed a statistically significant difference
(P ⬍ .05) in CFU-M expansion in the different cytokine treatment
groups on day 21 (Figure 1H). RT-PCR analysis confirmed that the
desmin-positive colonies also expressed mRNA transcripts for
myosin D1 (Figure 1I).
Figure 1. Multidifferentiation potential of suspension-derived cells. (A) In
addition to bone nodule formation detected by UV-fluorescence excitation of
tetracycline-labeled cultures (week 1 SCF plus IL-3 suspension-derived cells cultured
in CFU-O conditions shown here), (B) cells containing fat globules formed under
adipogenic culture conditions (⫻ 64). (C) RT-PCR confirmed that adipogenic cultures
expressed PPAR-␥. (D) Typical cells that grew in a CFU-F assay that were initiated
with cells removed from SSCs of the SCF plus IL-3 treatment group after 1 week
(⫻ 32.1). CFU-F cultures were stained with ␣-naphthyl acetate esterase followed by
a counterstain with hematoxylin solution at 2 weeks. (E) The typical cells from CFU-F
assays that were initiated with suspension-derived cells cultured in the presence of
SCF plus IL-3 plus 20 ng/mL PDGF-BB ( ⫻32.1). (F) At 7, 14, and 21 days of
suspension culture, test populations were removed from each suspension culture
and plated in CFU-M assays; at termination (day 21), the cultures were stained with
Giemsa to visualize the colonies generated. (G) Extent of desmin staining within
individual cells comprising the fibroblastic colonies observed in panel F (⫻ 20). (H)
Graphic display of the fold expansion in CFU-Ms. A statistically significant difference
(P ⬍ .05) in CFU-M expansion achieved in the different cytokine treatment groups on
day 21 was observed. Each bar represents the mean ⫾ SD (n ⫽ 3). (I) RT-PCR
confirmed that CFU-M cultures expressed MyoD1 after 21 days. GAPDH indicates
glyceraldehyde phosphate dehydrogenase.
BLOOD, 1 NOVEMBER 2005 䡠 VOLUME 106, NUMBER 9
A serum-free suspension culture system for the propagation
of mesenchymal and hematopoietic progenitors
The adhesion-independent SSC system used here allows the
independent manipulation of HCs and MPCs using phenotypic
markers. To study the interplay between MPCs and HPCs
effectively, as well as to attribute cell responses to specific
cytokines and not to unknown serum components and cell
adhesion effects, we used SF conditions. A 24 FD was used to
quantitatively explore the effects of exogenously added factors
and to detect synergies among factors. FGF, PDGF, SCF, and
IL-3 and their respective combinations were screened for their
ability to form CFU-Fs and CFU-Os on output cells from SSCs.
At 7, 14, and 21 days, the total number of cells was calculated
and CFU-F and CFU-O assays were initiated with cells derived
from each condition. The greatest expansion effects on total cell,
CFU-F, and CFU-O expansion were achieved in the SCF plus
IL-3 and SCF plus IL-3 plus PDGF treatment conditions,
whereas minimal growth was achieved in the single-factor
groups (Figure 2A, in which day-21 results are shown).
Interestingly, we have reported that the combination of SCF and
IL-3 also supports the growth of HPCs (in serum-containing
medium),17 suggesting that this approach may allow the growth
of both types of marrow-derived cell populations.
Using the standard methods of analyzing FD experiments,
ANOVA and linear regression analysis,19 the significance of the
output responses (ie, total cell, CFU-F, and CFU-O expansion)
from each cytokine combination was determined at each time
point. Statistical analysis obtained from the 2-level (none and
high-dose) FD experiment reported coefficients (␤ values) and
their respective P values for single-factor (main) effects as well as
for second- and higher-order interactions. The ␤ value was used to
determine the type of the effect (ie, negative or positive) that a
treatment condition had on the desired outcomes. This approach
allowed us to score significant differences (at day 21) in the
magnitude of the effect (␤) of each treatment condition on the
measured outcomes (Figure 2B).
Single-factor positive effects on total cell, CFU-F, and
CFU-O development were identified only when SCF or IL-3 was
added (Figure 2B). Despite this, SCF or IL-3 alone was not
sufficient in supporting a significant expansion of total cell,
CFU-O, or CFU-F numbers at any of the time points studied
(Figure 2Aii-iv). However, it would appear that SCF alone was
capable of supporting CFU-F expansion over the study period,
but the extent of this expansion was not statistically significant
(P ⬎ .05) between days 14 and 21. Importantly, CFU-F expansion was not associated with an increase in the total number of
cells in the presence of SCF alone (Figure 2Aii-iii). However,
when combined, SCF and IL-3 supported a significant expansion in total cells, CFU-Fs, and CFU-Os (Figure 2Aii-iv). Our
statistical analysis reported that the combination of SCF plus
IL-3 yielded a negative effect on both CFU-F and CFU-O
expansion (Figure 2B), despite the fact that the actual values
calculated for the fold expansion of CFU-Fs and CFU-Os were
more than 5-fold (Figure 2Ai). The negative parametric values
associated with CFU-F and CFU-O expansion (Figure 2B)
suggest that single treatment with SCF or IL-3, at the doses
examined, elicited overlapping responses and, as such, when
combined did not achieve a result greater than the additive
outputs (relative to no factor supplementation).
BLOOD, 1 NOVEMBER 2005 䡠 VOLUME 106, NUMBER 9
Figure 2. Fold expansion in total cells, CFU-Fs, and CFU-Os as a function of
cytokine combination and the effect of cytokine combination on total cell,
CFU-F, and CFU-O expansion. (A) The fold expansions in total cell, CFU-Fs, and
CFU-Os on day 21 were calculated relative to the day 0 values of each 24 cytokine
combination. Statistical analysis was performed using the Student t test between the
no cytokine and SCF plus IL-3 treatment conditions and between the SCF plus IL-3
and SCF plus IL-3 plus PDGF treatment conditions. Statistical differences (P ⬍ .05)
were observed in total cell, CFU-F, and CFU-O expansion between the no cytokine
and SCF plus IL-3 groups (denoted by the asterisk) and between the no cytokine and
SCF plus IL-3 plus PDGF treatment groups (denoted by double asterisks). No C
indicates no cytokine; F, FGF; P, PDGF; I, IL-3, and S, SCF. Each bar represents the
mean ⫾ SD (n ⫽ 3, performed in triplicate). Panels ii-iv show the fold expansion in
total cells, CFU-Fs, and CFU-Os as a function of time in the no cytokine, SCF, IL-3,
and SCF plus IL-3 conditions. (B) Graphic display of the effects (␤) of SCF, IL-3,
PDGF, and FGF and their 24 combinations on total cell, CFU-F, and CFU-O
expansion. Bars to the left of the center axis in each graph represent negative ␤
values, whereas positive ␤ values are represented on the right. Statistical significance is denoted by the asterisk (P ⬍ .05).
Subpopulation interactions in heterogeneous SSC conditions
To explore the mechanism by which SCF and IL-3 affect MPC growth,
experiments to examine the cross-talk between the different cell
populations in this heterogeneous culture were undertaken. CD45⫹ and
CD45⫺ cells were cultured in SSCs in the presence of 100 ng/mL SCF
and 20 ng/mL IL-3 in SF conditions (SSCs initiated with unsorted cells
and in the absence of added cytokines served as controls). SSC cells
from each group remained as single suspended cells throughout the
study period as determined by analyzing the cell suspension after
removing an aliquot from each SSC and examining the cells directly
under a microscope. Parallel SSCs were initiated with adherent-derived
MPCs. These cultures were used as positive controls for cell aggregate
formation and generated aggregates measuring greater than 70 to 200
␮m in SSCs, whereas our SSCs initiated with BM MNCs failed to form
such cell aggregates.
CROSS-TALK BETWEEN HUMAN BONE MARROW COMPARTMENTS
3015
The purity of CD45⫺ cells, obtained using cell sorting, was greater
than 99% (Figure 3A). The expression of CD45 within the SSCs
initiated with CD45⫹ and CD45⫺ cells was tracked to ensure that the
starting population remained CD45⫹ and CD45⫺, respectively. Flow
cytometric analysis confirmed that SSCs initiated with CD45⫺ cells
maintained a more than 97% CD45⫺ phenotype and cultures initiated
with CD45⫹ cells remained more than 99% CD45⫹.
At 7, 14, and 21 days, cells were removed and analyzed from each
culture condition (Figure 3B). SSCs initiated with unsorted and CD45⫺sorted cells gave rise to discrete colonies of CFU-Fs and CFU-Os (Figure
3Bi-iv). No detectable CFU-F or CFU-O activity was observed from the
CD45⫹-sorted subpopulation. Representative bone nodules derived from
the unsorted cells (Figure 3Bii) and CD45⫺ cells (Figure 3Biv) were
examined by scanning electron microscopy (SEM) to visualize the
elaborated mineralized collagenous extracellular matrix. Cells derived at
day 7 generated average CFU-F colony sizes of 7.2 ⫾ 1.9 mm and
6.6 ⫾ 1.2 mm from the unsorted and CD45⫺-sorted cell populations,
respectively. No significant difference in the CFU-F colony sizes from
these cell fractions (unsorted versus CD45⫺-sorted cells) was observed at
any time point studied, suggesting that SSC did not compromise the
capacity of suspension-grown MPCs to proliferate upon adhesion.
Quantitatively tracking CFU-F and CFU-O development revealed
that the unsorted cell group (containing ⬎ 60% CD45⫹ cells) resulted in
a higher fold-increase in total, CFU-F, and CFU-O cells when compared
to the numbers obtained from the CD45⫺-sorted cell population (Figure
3C). Significant differences were calculated between the unsorted and
CD45⫺-sorted cell groups with respect to the fold expansion in total
cells, CFU-Fs, and CFU-Os at various time points (Figure 3C), with
minimal CFU-F and CFU-O growth observed in the absence of CD45⫹
cells. On average, a 58% ⫾ 11% increase in CFU-F and CFU-O
development was observed when CD45⫺ cells were cultured in the
presence of CD45⫹ cells. Additionally, CD45⫹ cells cultured in the
absence of CD45⫺ cells demonstrated a significantly lower total cell
expansion (Figure 3Ci) and no capacity to form CFU-Fs (Figure 3Cii) or
CFU-Os (Figure 3Ciii).
As an initial screen to determine if there were factors secreted
by HCs (CD45⫹) that caused the growth of cells capable of CFU-F
and CFU-O development, conditioned medium (CM) from SSCs
supplemented with, or without, SCF and IL-3 were harvested and
used as the basal medium in CFU-F and CFU-O assays. BM MNCs
cultured in CM harvested from day 21 SCF plus IL-3–supplemented SF SSCs that were initiated with sorted BM MNC-derived
CD45⫹ cells yielded significantly greater numbers of CFU-F and
CFU-O cells, relative to CM harvested from parallel day 21 “no
cytokine” suspension cultures (Figure 4). These experiments allow
us to conclude that factors secreted by the CD45⫹ cells, and elicited
by the addition of SCF and IL-3 supplementation, positively affect
CFU-F and CFU-O growth.
SCF and IL-3 act by different mechanisms to support
suspension culture MPC growth
The results indicated that both IL-3 and SCF supplementation and
the presence of the CD45⫹ cell population were necessary components for maximal propagation of CD45⫺-derived CFU-Fs and
CFU-Os in SSCs. To investigate whether IL-3 or SCF transduces
signals directly on CD45⫺ cells, the expression of CD123 (IL-3␣
receptor) and CD117 (SCF receptor) was tracked using flow
cytometry. Propagation with SCF plus IL-3 resulted in a significant
increase in the percentage of cells expressing CD123 between days
7 and 21 (Figure 5A). The relative expression of CD45⫺CD123⫹
cells increased from 3.5% ⫾ 1.3% on day 7 to 8.9% ⫾ 2.3% by
day 21. Furthermore, the fluorescent intensity of CD45⫺CD123⫹
3016
BAKSH et al
BLOOD, 1 NOVEMBER 2005 䡠 VOLUME 106, NUMBER 9
Figure 3. The influence of CD45ⴙ cells on the yield of CFU-F
and CFU-O development from CD45ⴚ cells. (A) Representative
flow cytometric histogram plots demonstrate the distribution of
both CD45⫺ and CD45⫹ cell populations from input BM-derived
cells. (i) Day 0 BM-derived cells were incubated with saturating
concentrations of CD45-PE and (ii) the CD45⫺ cells were recovered. R1 represents the isotype control gate corresponding to the
CD45⫺ cell fraction, and R2 represents the region used to sort
CD45⫹ cells. The purity of the recovered CD45⫺ fraction was
more than 99%, as determined by flow cytometry. (B) CFU-F and
CFU-O development from unsorted and CD45⫺ suspensionderived cells. Cells harvested from day 7 suspension cultures
initiated with unsorted and CD45⫺ cells were plated in CFU-F and
CFU-O assays. CFU-F assays were stained with Giemsa to
visualize discrete colonies of fibroblastic cells (i,iii), and CFU-O
assays (ii,iv) were incubated with tetracycline to visualize newly
formed bone nodules (seen as white areas) under UV-fluorescence microscopy. Scanning electron micrographs revealed a
layer of globular accretions (v) and mineralized extracellular
matrix (vi). Field width (FW) for panels v and vi ⫽ 70 ␮m and 14
␮m, respectively. Note: CFU-F and CFU-O assays were initiated
with 1 ⫻ 103 cells/cm2. (C) The calculated fold expansion in total
cells, CFU-Fs, and CFU-Os attained in SSCs initiated with
unsorted, CD45⫺, and CD45⫹ cells. Calculated (i) total cell, (ii)
CFU-F, and (iii) CFU-O expansions as a function of suspension
culture time. The cells harvested from the CD45⫹ suspension
cultures did not give rise to CFU-Fs and CFU-Os. Statistically
significant differences (P ⬍ .05) are denoted with an asterisk.
Each bar represents the mean ⫾ SD (n ⫽ 3, performed in
triplicate).
cells also increased with time (Figure 5Aiv), suggesting that a
higher density of CD123 receptor complexes may be present on
single CD45⫺ cells with time in suspension culture.
The total number of CD45⫺CD123⫹ cells that were generated
throughout SSCs was calculated using the percent expression
obtained by flow cytometry (Figure 5A). A decrease in the number
of CD45⫺CD123⫹ cells was initially observed; however, after 7
days, the number of CD45⫺CD123⫹ cells significantly increased
(P ⬍ .05) in number resulting in about a 12-fold increase in total
CD45⫺CD123⫹ cells by day 21 (relative to day 7 values; Figure 5B).
To determine whether IL-3 was acting directly on MPCs, the
CFU-F and CFU-O developmental potential of the CD45⫺CD123⫹
cells was tested. Representative tetracycline-labeled CFU-O cells
were processed for SEM and demonstrated hallmarks of de novo
bone formation, as we have previously described23 (Figure 5Ciiiiv). No CFU-Fs or CFU-Os were detected from input (day 0)
BM-derived CD45⫺CD123⫹ cells at the cell-seeding dilution used
(ie, 1 ⫻ 103 cells/cm2). However, following 7 days in suspension
culture, about 1 in 104 CD45⫺CD123⫹ cells had the capacity to
form CFU-Fs and CFU-Os (Figure 5C). CD45⫺ cells were also
evaluated for their expression of CD117; less than 1% of the total
CD45⫺ population was CD117⫹ (data not shown). Furthermore,
CD45⫺CD117⫹ cells, isolated on day 7 of suspension culture,
failed to give rise to CFU-Fs and CFU-Os. Importantly, a
CD45⫺CD123⫹CD117⫹ or CD45⫺CD123⫺CD117⫹ cell population was not detected in these studies, suggesting that the observed
contribution of SCF to the MPC growth is indirect.
Figure 4. Effects of conditioned medium derived from HCs grown in suspension on CFU-F and CFU-O development. BM-MNCs were plated directly in CFU-F
and CFU-O assay conditions that used conditioned medium harvested from day 21
suspension cultures initiated with CD45⫹ cells that were supplemented with and
without SCF plus IL-3. Statistical significance (P ⬍ .05) is denoted by an asterisk.
Each bar represents the mean ⫾ SD (n ⫽ 2 BM samples, performed in triplicate).
A comparison in the yield of CFU-Fs and CFU-Os from
unsorted, CD45ⴚ, and CD45ⴚCD123ⴙ cell populations
Analysis of the frequency of CFU-Fs and CFU-Os over the 21-day
culture period in the unsorted, CD45⫺-, and CD45⫺CD123⫹-sorted
BLOOD, 1 NOVEMBER 2005 䡠 VOLUME 106, NUMBER 9
CROSS-TALK BETWEEN HUMAN BONE MARROW COMPARTMENTS
3017
cell fractions revealed an increase in the frequency of both CFU-Fs
and CFU-Os by day 21 (relative to day 7 values; Table 1).
Importantly, the frequency of CFU-Fs and CFU-Os measured in
SSCs initiated with CD45⫺ cells (alone) was statistically less than
that obtained in the unsorted cell population, emphasizing, again,
the importance of CD45⫹ cells in the growth and maintenance of
these clonogenic MPCs.
Flow cytometric analysis revealed that the CD45⫺CD123⫹
subpopulation represented less than about 10% of the total CD45⫺
cell population (Figure 5). We also observed a similar percent
expression (⬃10%) of CD45⫺CD123⫹ cells in the total unsorted
BM cell population after 21 days of SSC (not shown). We
determined using clonogenic assays that the CD45⫺CD123⫹
subpopulation, after sorting from the CD45⫺ cell population,
contained about 24% and about 22% of the total CFU-Fs and
CFU-Os, respectively, when compared to the number generated
from the CD45⫺-sorted cell population at day 21 (Table 1).
Additionally, we observed that the CD45⫺CD123⫹ subpopulation
sorted from the total cell population (unsorted cells) generated
about 23% and about 18% of the CFU-Fs and CFU-Os, respectively, relative to the unsorted cell population, by day 21 (Table 1)
and not 100% recovery, which we might have expected. Taken
together, our analysis (Table 1) revealed that CD123⫹CD45⫺ cells
might also require the presence of CD45⫹ cells to attain their
maximal colony number output. Despite this, the observed increase
in the frequency of CFU-Fs and CFU-Os, as a function of time,
suggests that a CD45⫺ CD123⫹ cell subpopulation may contain a
cell type directly responsive to IL-3, which has the capacity to give rise
to both CFU-Fs and CFU-Os and that this population also requires
factors secreted by HCs to support their growth in suspension.
Discussion
Figure 5. Developmental potential of CD45ⴚCD123ⴙ cells. (A) Representative
flow cytometric dot plots demonstrating the relative increase in the percentage of
CD45⫺CD123⫹ cells throughout the 21-day culture period. CD45⫺ cells cultured in
suspension were dual-labeled for CD45 and CD123 to track CD45⫺CD123⫹expressing cells. The CD45⫺CD123⫹ cells, represented by R1, were sorted on day 0
(i), 7 (ii), 14 (iii), and 21 (iv) and the developmental potential of these isolated
populations were determined in CFU-F and CFU-O assays. Representative quad
plots of day 0, 7, 14, and 21 suspension-derived cells (of a suspension culture
initiated with CD45⫺ cells) stained with CD45 and CD123 are shown here. Positive
expression was determined as more than 99% of isotype control and about 15 000
gated events were collected. (B) Total number of CD45⫺CD123⫹ cells generated in
SF suspension culture conditions supplemented with 100 ng/mL SCF and 20 ng/mL
IL-3. Dual fluorescence labeling for CD45 and CD123 was used to track the
percentage of CD45⫺CD123⫹cells in suspension cultures initiated with CD45⫺ cells.
These values were used to calculate the total number of CD45⫺CD123⫹cells
generated throughout suspension culture. Statistical significance (P ⬍ .05) is denoted by an asterisk. Each data point represents the mean ⫾ SD (n ⫽ 3). (C) CFU-F
and CFU-O development from CD45⫺CD123⫹ suspension-derived cells.
CD45⫺CD123⫹-sorted cells (grown in CD45⫺ SSCs) were plated in CFU-F and
CFU-O assays following 7 days of suspension culture. (i) CFU-F assays were stained
with Giemsa to visualize discrete colonies of fibroblastic cells; (ii) CFU-O assays were
incubated with tetracycline (TC), to visualize newly formed bone nodules (seen as
white areas) under UV-fluorescence microscopy. Representative scanning electron
micrographs of revealing (iii) cement line matrix deposition (arrow) and (iv) collagen
mineralization (arrow) derived from CD45⫺CD123⫹ sorted cells grown in CFU-O
assay conditions (FW for iii and iv ⫽ 53 ␮m and 21 ␮m, respectively). Note: CFU-F
and CFU-O assays were initiated with 1 ⫻ 103 cells/cm2.
Despite increasing experimental and clinical interest in MSCs, our
understanding of the underlying mechanisms regulating MSC fate
remains elusive. The established approach to the isolation and
culture of MSCs involves their propagation on tissue culture plastic
with the concomitant removal of HCs during adherent culturemediated expansion. However, one important unanswered question
in this approach is whether HCs play a role in signaling to MPCs,
regulating their growth and development, similar to the converse
that is seen in hematopoiesis by MC types.7 To address this
question, we investigated the regulatory role that HCs may play in
MPC development by designing a suspension culture system that
would enable the coculture of both HCs and MCs.
First, we confirmed that our SSC maintained MPCs capable of
developing into multiple cell types by demonstrating their ability to
differentiate into the fibroblast, osteogenic, adipogenic, and myogenic lineages. These findings provided motivation to examine
specific factors that might influence the survival and growth of
multipotential MPCs in suspension such as cytokine supplementation and accessory cell involvement, specifically HCs. Therefore,
to limit the effects of adhesion and focus on the regulatory role of
specific cytokines in supporting the growth and expansion of MPCs
in suspension, as well as to determine the influence of HCs (via
endogenous secretion of soluble factors) on the growth of MPCs,
experiments were performed in a SF SSC.
FD was used rather than “one-factor-at-a-time” experiments to
evaluate the effects of multiple growth factors on total cell and
mesenchymal colony-forming potential. Our factorial also allowed
interactions between factors to be detected. Analysis revealed that
3018
BLOOD, 1 NOVEMBER 2005 䡠 VOLUME 106, NUMBER 9
BAKSH et al
Table 1. CFU-F and CFU-O colony numbers/104 cells from suspension-derived cells of various cell fractions
CFU-F
CFU-O
CD123ⴙCD45ⴚ
cells*
CD45ⴚ cells
sorted†
CD123ⴙCD45ⴚ
cells‡
Unsorted cells
CD123ⴙCD45ⴚ
cells*
CD45ⴚ cells
sorted†
CD123ⴙCD45ⴚ
cells‡
Time, d
Unsorted cells
0
10 ⫾ 2
0⫾0
8⫾0
0⫾0
5⫾2
4⫾0
0⫾0
7
14 ⫾ 2
2.3 ⫾ 0.6
1⫾0
0.3 ⫾ 0.6
8.6 ⫾ 1.5
0.67 ⫾ 0
0.7 ⫾ 0.6
0.3 ⫾ 0.6
14
11 ⫾ 4
3⫾1
1.3 ⫾ 0.6
1.0 ⫾ 0
15 ⫾ 3
1.33 ⫾ 0.6
2.3 ⫾ 0.6
0.3 ⫾ 0.6
21
19 ⫾ 5
4.3 ⫾ 0.6
7⫾1
1.7 ⫾ 0.6
20 ⫾ 3.5
3 ⫾ 0.6
0.67 ⫾ 0.6
0⫾0
3.6 ⫾ 1
*CD123⫹CD45⫺ cells isolated from unsorted cells that were grown in stirred suspension.
†CD45⫺ cells were sorted from input BM samples and were directly grown in stirred suspension.
‡CD123⫹CD45⫺ cells isolated from CD45⫺-sorted cells that were grown in stirred suspension.
SCF and IL-3 had positive effects on the survival of total cells,
CFU-Fs, and CFU-Os in SF SSCs initiated with (unsorted)
BM-derived cells. IL-3 alone demonstrated the capacity to maintain input numbers throughout the 21-day culture period; however,
this growth condition was insufficient in stimulating cell and
progenitor expansion. Expansion was obtained with the combination of SCF plus IL-3 (⬎ 5-fold). Interestingly, the combined effect
of SCF plus IL-3 was less than additive (Figure 2B), suggesting
that these factors may elicit their effects via overlapping mechanisms. This observation, along with the cell-sorting and cell surface
receptor analysis studies, provide insight into the complex intercellular network regulating MPC and HPC responses in vitro.
In HSC culture systems, it has been well established that the
survival and proliferation of HSCs is controlled by a number of
interacting cytokines including SCF and IL-3,24 which act directly
on progenitor cells expressing receptors for these ligands.25 IL-3 is
considered an early acting growth factor necessary for the proliferation of multipotential colony-forming cells that include granulocytes, erythroid cells, monocytes, and the like,26 whereas SCF has
been shown to promote the survival of hematopoietic stem/
progenitor cells.27 In combination, these 2 cytokines are thought to
act synergistically to inhibit apoptosis and promote proliferation.25,28 Based on the presumed colocalization of putative MSCs
and their descendents and HSCs in the BM compartment, it is
plausible that these cytokines play similar roles in vivo.
To investigate the mechanism by which IL-3 transduced
stimulatory signals to MPCs, SSCs were initiated under cytokinesupplemented SF conditions with subsets of the whole BM
population. Tests were performed to determine if IL-3 affects
MPCs in an indirect fashion through HCs (CD45⫹) and their
subsequent release of soluble factors or by direct action on CD45⫺
cells. Cells responsive to IL-3 express CD123 (IL-3␣ receptor) and
are typically of the hematopoietic lineage (ie, CD45⫹).26,29,30 To
explore the potential direct action of IL-3 on MPCs, CD45⫺ cells
were cultured in SF SSC supplemented with SCF and IL-3. Flow
cytometric analysis confirmed that more than 97% of the cells
maintained a CD45⫺ phenotype throughout suspension culture and
propagation with SCF and IL-3 resulted in a significant increase in
the percentage and total number of CD45⫺ cells expressing IL-3␣
receptor (CD123) over 21 days (Figure 4A-B). Importantly, we
demonstrated that the CD45⫺ cell fraction included a small
percentage of CD123⫹ cells.
Interestingly, there were no CFU-Fs or CFU-Os detected in the
CD45⫺ CD123⫹ cell fraction at input, although about 1 in 10 000
CD4⫺CD123⫹ cells was capable of giving rise to CFU-Fs and
CFU-Os as early as day 7 of suspension culture (Table 1). These
data suggest that the CD123⫹ cells giving rise to CFU-Fs and
CFU-Os may be at a very low frequency in BM, undetectable in our
assay conditions or, perhaps MPCs “acquire” this phenotype after
an extended time in SSC. In response to IL-3 (with SCF) this rare
cell type may not only selectively grow and thus be readily detected
by immunophenotyping techniques such as flow cytometry, but
also be within the detection limit of conventional functional assays
over time in culture. Surprisingly, despite their low frequency,
about 24% and about 22% of the total attainable CFU-Fs and
CFU-Os, respectively, were recoverable in the CD45⫺CD123⫹ cell
population (at day 21). This represents a significant fraction of the
total number of available CFU-Fs and CFU-Os. These findings
suggest that MPCs may be phenotypically heterogeneous, an
observation corroborated in other studies demonstrating that adherent-derived MSC populations display heterogeneity with respect to
their self-renewal31,32 and multilineage differentiation potential.11,33
However, whether the apparent heterogeneity ascribed to suspension-derived MPCs is attributed to lineage hierarchy that exists in
the putative in vivo MSC compartment (ie, CD45⫺CD123⫺ to
CD45⫺CD123⫹ or vice versa), or whether a functional
CD45⫺CD123⫹ population is present, undetected, in the BM
compartment, remains to be elucidated. Nevertheless, the results
presented provide evidence that IL-3 has the potential to act
directly on a unique subset of MPCs with CFU-F and CFU-O
developmental potential and provides a mechanism by which
subpopulations of MPCs have the capacity to grow in noncontact
suspension culture conditions. Determination of the multilineage
potential of the CD45⫺CD123⫹ cell population (studies underway)
would help to clearly delineate the stem versus progenitor cell
potential of this population.
These studies also revealed a role for HCs in the stimulation of
CFU-F and CFU-O growth in SSCs (Figure 3C). Analysis of the
combined fold expansion of total cells attained in SSCs initiated
with CD45⫹ or CD45⫺ cells alone was calculated to be less than
Figure 6. The differential responsiveness of MPCs to endogenously produced
and exogenously supplemented factors. IL-3 exerts its action (⫹) on both HCs
and non-HCs throughout SF suspension culture, resulting in the expansion of cells in
both these compartments. IL-3 has also been shown to promote the growth of
CD45⫺CD123⫹ cells, which are capable of giving rise to CFU-Fs and CFU-Os.
However, on average, only about 24% and about 22% of the total attainable CFU-Fs
and CFU-Os, respectively, are recovered from a CD45⫺CD123⫹ cell population by
day 21 of the study period, suggesting that the mesenchymal stem cell compartment
is phenotypically heterogeneous. Furthermore, the growth of HCs is also directly
influenced by IL-3 (and SCF). It has also been demonstrated from these studies that
HCs potentiate the growth of MPCs via the release of endogenous factors (factor(s)
X) that target MPCs, resulting in their growth.
BLOOD, 1 NOVEMBER 2005 䡠 VOLUME 106, NUMBER 9
CROSS-TALK BETWEEN HUMAN BONE MARROW COMPARTMENTS
that obtained from suspension cultures initiated with unsorted cells.
Specifically, a 58% ⫾ 11% increase in the number of CFU-F–and
CFU-O–detectable colonies was calculated over the 21-day study
period when CD45⫺ cells were grown in the presence of CD45⫹
cells. These results support the existence of positive interactions
between CD45⫹ and CD45⫺ cells, likely mediated by their
respective endogenous secretion of soluble factors, similar to that
which is observed between early HCs and their progenitors.34
The soluble factor composition of the SSC medium was not
analyzed in this study. However, preliminary experiments revealed
a significant increase in the yield of CFU-Fs and CFU-Os when
BM-derived cells were cultured with conditioned media harvested
from SF suspension cultures (Figure 4). Taken together, we
hypothesize that suspension-grown HCs (⬎ 60% of the unsorted
suspension cell population) secrete numerous regulatory molecules
that form the basis of intercellular cross-talk networks, which then
regulate the survival and growth of MPCs in an autocrine or
paracrine manner.34,35 Assessing the protein composition of the
media from which HCs are grown may provide further insight into
the important role that CD45⫹ cells play on MPC growth and allow
further optimization of stimulatory signals necessary for expanding
an MPC population in this controlled SSC system.
In summary, the development of an SF suspension culture
configuration provided the opportunity to examine soluble factor
3019
interactions between the mesenchymal and hematopoietic compartments. Collectively, these results demonstrate the ability of IL-3 to
enhance MPC expansion in SSCs and suggest that it may act
directly with CD123⫹ MPCs (Figure 6). The observed competitive
emergence of CFU-F– and CFU-O–capable CD45⫺ CD123⫹ cells
in direct response to IL-3 supplementation revealed heterogeneity
in the MSC compartment. The functional relevance of this phenotypic heterogeneity remains to be determined.
Acknowledgments
We would like to thank StemCell Technologies (Vancouver, BC,
Canada) for generously providing the reagents and medium used in
this work and Dr A. Keating (Toronto Hospital Network, Toronto,
ON, Canada) for providing the BM samples. We would also like to
acknowledge Dr Jacque Tremblay (Centre de Recherche en Génétique Humaine, QC, Canada) and Dr Jane Aubin (Department of
Medical Biophysics, University of Toronto) for providing culture
media for and advice concerning the myogenic cultures. Minke
Hoeksma provided technical assistance in the factorial design
portion of this work.
P.W.Z. is the Canada Research Chair in Stem Cell
Bioengineering.
References
1. Clark BR, Keating A. Biology of bone marrow
stroma. Ann N Y Acad Sci. 1995;770:70-78.
2. Koller MR, Manchel I, Palsson BO. Importance of
parenchymal:stromal cell ratio for the ex vivo reconstitution of human hematopoiesis. Stem Cells.
1997;15:305-313.
3. Majumdar MK, Thiede MA, Haynesworth SE,
Bruder SP, Gerson SL. Human marrow-derived
mesenchymal stem cells (MSCs) express hematopoietic cytokines and support long-term hematopoiesis when differentiated toward stromal and
osteogenic lineages. J Hematother Stem Cell
Res. 2000;9:841-848.
4. Haynesworth SE, Baber MA, Caplan AI. Cytokine
expression by human marrow-derived mesenchymal
progenitor cells in vitro: effects of dexamethasone
and IL-1 alpha. J Cell Physiol. 1996;166:585-592.
5. Reese JS, Koc ON, Gerson SL. Human mesenchymal stem cells provide stromal support for efficient CD34⫹ transduction. J Hematother Stem
Cell Res. 1999;8:515-523.
6. Moore KA. Recent advances in defining the hematopoietic stem cell niche. Curr Opin Hematol.
2004;11:107-111.
7. Calvi LM, Adams GB, Weibrecht KW, et al. Osteoblastic cells regulate the haematopoietic stem
cell niche. Nature. 2003;425:841-846.
8. Zhang J, Niu C, Ye L, et al. Identification of the
haematopoietic stem cell niche and control of the
niche size. Nature. 2003;425:836-841.
9. Castro-Malaspina H, Gay RE, Resnick G, et al.
Characterization of human bone marrow fibroblast colony-forming cells (CFU-F) and their progeny. Blood. 1980;56:289-301.
10. Caplan AI. Mesenchymal stem cells. J Orthop
Res. 1991;9:641-650.
11. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage potential of adult human mesenchymal
stem cells. Science. 1999;284:143-147.
12. Aubin JE. Osteoprogenitor cell frequency in rat
bone marrow stromal populations: role for heterotypic cell-cell interactions in osteoblast differentiation. J Cell Biochem. 1999;72:396-410.
13. Gronthos S, Simmons PJ. The growth factor re-
quirements of STRO-1-positive human bone marrow stromal precursors under serum-deprived
conditions in vitro. Blood. 1995;85:929-940.
14. Dexter TM, Wright EG, Krizsa F, Lajtha LG. Regulation of haemopoietic stem cell proliferation in
long term bone marrow cultures. Biomedicine.
1977;27:344-349.
15. Dexter TM, Allen TD, Lajtha LG. Conditions controlling the proliferation of haemopoietic stem
cells in vitro. J Cell Physiol. 1977;91:335-344.
16. Dexter TM, Lajtha LG. Proliferation of haemopoietic
stem cells in vitro. Br J Haematol. 1974;28:525-530.
17. Zandstra PW, Eaves CJ, Piret JM. Expansion of hematopoietic progenitor cell populations in stirred suspension bioreactors of normal human bone marrow
cells. Biotechnol Bioeng. 1994;12:909-914.
18. Baksh D, Davies JE, Zandstra PW. Adult human
bone marrow-derived mesenchymal progenitor cells
are capable of adhesion-independent survival and
expansion. Exp Hematol. 2003;31:723-732.
19. Box GEP, Hunter WG, Hunter JS. Statistics for
Experimenters. New York, NY: John Wiley &
Sons; 1978.
20. Wang QR, Yan ZJ, Wolf NS. Dissecting the hematopoietic microenvironment, VI: the effects of
several growth factors on the in vitro growth of
murine bone marrow CFU-F. Exp Hematol. 1990;
18:341-347.
21. Parker E, Shiga A, Davies JE. Growing human bone
in vitro. In Davies JE, ed. Bone Engineering. Toronto,
ON, Canada: Em Squared Incorporated; 2000:63-80.
22. Mezey E, Chandross KJ, Harta G, Maki RA,
McKercher SR. Turning blood into brain: cells
bearing neuronal antigens generated in vivo from
bone marrow. Science. 2000;290:1779-1782.
26. de Groot RP, Coffer PJ, Koenderman L. Regulation of proliferation, differentiation and survival by
the IL- 3/IL-5/GM-CSF receptor family. Cell Signal. 1998;10:619-628.
27. Smith MA, Court EL, Smith JG. Stem cell factor:
laboratory and clinical aspects. Blood Rev. 2001;
15:191-197.
28. Itakura A, Miura Y, Hikasa Y, Kiso Y, Matsuda H.
Interleukin-3 and stem cell factor modulate cell
cycle regulatory factors in mast cells: negative
regulation of p27Kip1 in proliferation of mast cells
induced by interleukin-3 but not stem cell factor.
Exp Hematol. 2001;29:803-811.
29. Munoz L, Nomdedeu JF, Lopez O, et al. Interleukin-3 receptor alpha chain (CD123) is widely expressed in hematologic malignancies. Haematologica. 2001;86:1261-1269.
30. Huang S, Chen Z, Yu JF, et al. Correlation between IL-3 receptor expression and growth potential of human CD34⫹ hematopoietic cells from
different tissues. Stem Cells. 1999;17:265-272.
31. Banfi A, Muraglia A, Dozin B, et al. Proliferation
kinetics and differentiation potential of ex vivo expanded human bone marrow stromal cells: implications for their use in cell therapy. Exp Hematol.
2000;28:707-715.
32. Gronthos S, Zannettino AC, Hay SJ, et al. Molecular and cellular characterisation of highly purified stromal stem cells derived from human bone
marrow. J Cell Sci. 2003;116:1827-1835.
33. Muraglia A, Cancedda R, Quarto R. Clonal mesenchymal progenitors from human bone marrow
differentiate in vitro according to a hierarchical
model. J Cell Sci. 2000;113(pt 7):1161-1166.
24. Ogawa M. Differentiation and proliferation of hematopoietic stem cells. Blood. 1993;81:2844-2853.
34. Majka M, Janowska-Wieczorek A, Ratajczak J, et al.
Numerous growth factors, cytokines, and chemokines are secreted by human CD34(⫹) cells, myeloblasts, erythroblasts, and megakaryoblasts and
regulate normal hematopoiesis in an autocrine/paracrine manner. Blood. 2001;97:3075-3085.
25. Lemoli RM, Gulati SC. Effect of stem cell factor
(c-kit ligand), granulocyte-macrophage colony
stimulating factor and interleukin 3 on hematopoietic progenitors in human long-term bone marrow
cultures. Stem Cells. 1993;11:435-444.
35. Majumdar MK, Thiede MA, Mosca JD, Moorman
M, Gerson SL. Phenotypic and functional comparison of cultures of marrow-derived mesenchymal stem cells (MSCs) and stromal cells. J Cell
Physiol. 1998;176:57-66.
23. Davies JE. In vitro modeling of the bone/implant
interface. Anat Rec. 1996;245:426-445.