ORIGINAL ARTICLE 817

Canonical and Non-Canonical

Wnts Differentially Affect the
Development Potential of
Primary Isolate of Human Bone
Marrow Mesenchymal Stem Cells
DOLORES BAKSH AND ROCKY S. TUAN*
Cartilage Biology and Orthopaedics Branch, National Institute of Arthritis, and Musculoskeletal and Skin Diseases,
National Institutes of Health, Department and Health and Human Services, Bethesda, Maryland, 20892

This study examines the role of Wnt signaling events in regulating the differential potential of mesenchymal stem cells (MSCs) from adult
bone marrow (BM). Immunohistochemical analysis of BM revealed co-localization of Wnt5a protein, a non-canonical Wnt, with CD45þ
cells and CD45
STRO-1þ cells, while Wnt3a expression, a canonical Wnt, was associated with the underlying stroma matrix, suggesting
that Wnts may regulate MSCs in their niche in BM. To elucidate the role of Wnts in MSC development, adult human BM-derived
mononuclear cells were maintained as suspension cultures to recapitulate the marrow cellular environment, in serum-free, with the
addition of Wnt3a and Wnt5a protein. Results showed that Wnt3a increased cell numbers and expanded the pool of MSCs capable of
colony forming unit — fibroblast (CFU-F) and CFU — osteoblast (O), while Wnt5a maintained cell numbers and CFU-F and CFU-O
numbers. However, when cells were cultured directly onto tissue culture plastic, Wnt5a increased the number of CFU-O relative to
control conditions. These findings suggest the potential dual role of Wnt5a in the maintenance of MSCs in BM and enhancing osteogenesis
ex vivo. Our work provides evidence that Wnts can function as mesenchymal regulatory factors by providing instructive cues for the
recruitment, maintenance, and differentiation of MSCs.
J. Cell. Physiol. 212: 817–826, 2007. ß 2007 Wiley-Liss, Inc.

The successful ex vivo reconstruction of human bone marrow
(BM) is an important scientific and clinical goal, largely due to
the fact it represents an attractive source of many cell types,
most notably, stem cells, for a number of clinical applications.
The BM represents a microenvironment for the dynamic
interplay between regulatory molecules such as cytokines,
cell–cell and cell–extracellular matrix interactions. It is in this
intricately organized microenvironment where mesenchymal
stem cells (MSC) are thought to reside. MSCs are known to
exist since ex vivo culture of the cellular component of BM
demonstrates that there are mesenchymal stem/progenitor
cells, which are capable of giving rise to a broad spectrum of
mesenchymal tissue derivates, including bone, cartilage, muscle
and fat, upon their isolation and culture expansion on tissue
culture plastic (Caplan, 1991; Pittenger et al., 1999). However,
the exact location of MSCs in BM has yet to be unequivocally
determined and the cell–cell and cell–matrix interactions that
regulate MSCs in their native environment are not fully
understood. In comparison, the interactions, which modulate
hematopoietic stem cells (HSCs) and their progeny in BM have
been clearly documented (Friedenstein et al., 1978; Clark and
Keating, 1995; Koller et al., 1997; Janowska-Wieczorek et al.,
2001). Identifying the factors that regulate MSCs in their niche is
thus a critical step in successfully reconstructing BM ex vivo.
Clinically, successful reconstruction of human BM would permit
the efficient and controlled cultivation of MSCs for a broad
range of cell-based therapies.
In effort to study the factors that regulate MSCs in vivo, we
recognize the key role that Wnts play in mesengenesis. Wnts
have been shown to play an important role in the proliferation
and differentiation of plastic-adherent, cultured-expanded
MSCs derived from adult human BM (Boland et al., 2004; De
Boer et al., 2004; Etheridge et al., 2004; Baksh et al., 2007).
Therefore, in view of the functional involvement of Wnts in
embryonic and postnatal development and in tissue
homeostasis, we hypothesize that Wnt signaling pathways
represent candidate regulatory mechanisms governing the
self-renewal and differentiation of MSCs in their niche. In order
to test this hypothesis, we designed experiments that
maintained the biological and functional properties of MSCs as
comparable to what they are in their in vivo milieu, and thus
minimize the loss of functionality, which has been associated
with the long-term adherent culture-expansion of MSCs
(Bruder et al., 1997). To this end, we used our previously
established suspension culture approach to co-culture both
hematopoietic and mesenchymal cells (Baksh et al., 2003, 2005).
We employed this culture system to study the effects that Wnts
have on MSCs directly isolated from their BM niche in a
non-contact environment. First, using immunofluorescence
microscopy, we revealed the spatial location of prototypical
canonical Wnt, Wnt3a, and prototypical non-canonical Wnt,
Wnt5a, protein expression in sections of adult human
trabecular BM. This analysis demonstrated co-localization of
Wnt5a expression with CD45þ cells (hematopoietic) and
CD45
STRO-1þ (non-hematopoietic) cells, while Wnt3a
Contract grant sponsor: NIH;
Contract grant number: Z01 AR41131.
*Correspondence to: Rocky S. Tuan, Cartilage Biology and
Orthopaedics Branch, National Institute of Arthritis and
Musculoskeletal and Skin Diseases, 50 South Dr., Room 1503, MSC
8022, National Institutes of Health, Bethesda, MD 20892-8022.
E-mail: tuanr@mail.nih.gov
Received 4 January 2007; Accepted 12 February 2007
DOI: 10.1002/jcp.21080

ß 2 0 0 7 W I L E Y - L I S S , I N C .
818 BAKSH AND TUAN
expression was associated with the underlying extracellular
matrix. Furthermore, we determined the colony forming
unit-fibroblast (CFU-F) and CFU-osteoblast (CFU-O) potential
of BM-derived MSCs directly isolated from marrow in response
to Wnt3a and Wnt5a exposure, and demonstrated a differential
recruitment of MSCs capable of giving rise to fibroblastic and
osteogenic colonies. Using this suspension culture
configuration, BM-derived cells, containing MSCs, were grown
in the presence and absence of Wnt3a and Wnt5a protein
under serum-free conditions, and the CFU-F and CFU-O
development was evaluated from suspension-grown cells. Our
results showed that Wnt3a expanded the total number of cells,
as well as MSCs capable of giving rise to CFU-F and CFU-O,
while exposure to Wnt5a maintained input numbers. These
findings revealed the differential response of BM-derived MSCs
to Wnts, with respect to cell maintenance, self-renewal and
differentiation potential. Our work provides evidence
suggesting that Wnts can function as mesenchymal regulatory
factors in vivo by providing instructive cues for the recruitment,
maintenance, and differentiation of MSCs.
Materials and Methods
Histology of human trabecular bone marrow
Adult human bone trabecular samples harvested from the hips of
consenting patients (following IRB approval, George Washington
University and National Institutes of Health) (N ¼ 4, ages 55–65) were
immersed in 10% buffered formalin (pH 7.2) overnight. Fixed samples
were placed in Immunocal (Decal Corp., Tallman, NY) for
approximately 4–5 weeks until the specimens were decalcified. The
decalcifing bone was kept under orbital shaking and the solution was
changed every second day. Decalcified samples were then
dehydrated through a graded series of alcohols and HistoClear, and
were paraffin-embedded. Sections were cut at 5 mm thickness using a
Leica RM 2135 microtome (Leica Microsystems Inc, Columbia, MD).
Slides were stained in Harris’ Hematoxylin (Sigma) and counter-stained
in Aqueous Eosin (Sigma).
Immunofluorescence for Wnt3a and Wnt5a
Wnt3a and Wnt5a monoclonal antibodies were purchased from R&D
Systems: Wnt3a and Wnt5a were used as primary antibodies at optimal
concentrations of 10 mg/ml. Binding to nonspecific proteins was
inhibited using a blocking buffer consisting of 5% rat and 5% goat serum
for Wnt3a and Wnt5a incubated slides, respectively. All antibody
dilutions were made in phosphate buffered saline (PBS) (Gibco)
supplemented with 2% bovine serum albumin (BSA) (Sigma). Labeling
was done overnight at 48C. Sections were washed three times in PBS.
Rhodamine Red-conjugated mouse anti-rat (Jackson Immuno Research
Laboratories, West Grove, PA) and Texas Red-conjugated mouse anti-
goat (Jackson Immuno Research Laboratories) secondary antibodies
were used at 1:200 dilution from stock at room temperature in a
humidified chamber for 1 h to detect the expression of Wnt3a and
Wnt5a, respectively. Sections were then washed three times in PBS.
Following the initial staining with Wnt3a and Wnt5a, some sections
were also stained with conjugated mouse IgG1,k anti-human
monoclonal antibody for CD45-PE (phycoerythrin) (1:200) (BD
Biosciences), anti-mouse/human monoclonal antibody for collagen
type I (Col I) (1:2000) (Abcam) and STRO-1 (Hybridoma Bank). Slides
incubated with anti-Col 1 were stained with FITC-conjugated rat anti-
mouse monoclonal secondary antibody (Molecular Probes). The
procedure for STRO-1 antibody staining involved incubating the slides
in 200 ml of saturating concentrations of the mouse IgM monoclonal
antibody STRO-1 for 30 min at 48C. This step was preceded with
blocking the slides with 1% human, goat and mouse sera for 10 min. The
slides were then washed twice with PBS containing 2% fetal bovine
serum (FBS) before the addition of PE-conjugated rat anti-mouse IgM
monoclonal secondary antibody (BD Biosciences). All sections were
counterstained with Hoechst dye (Sigma) and mounted in Vectasheild
(Vector) to preserve fluorescence labeling. Negative controls were
obtained by incubating sections only in secondary antibodies.
Immunostained sections were observed with epifluorescence optics
(Leitz, Heerbrugg, Switzerland). The expression of Wnt3a, Wnt5a,
JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP
CD45, Col I and STRO-1 was analyzed in multiple sections from at least
four different human donors.
Bone marrow mononuclear cell harvest
Human BM cells were harvested from the hips of consenting
patients (following IRB approval, George Washington University)
(N ¼ 4, ages 55–65). Ficoll-Paque (Amersham Biosciences) density
gradient separation medium was used to isolate the mononuclear cells
(MNCs). The MNCs were washed twice with PBS and the total viable
cells were counted using trypan blue dye (Sigma) exclusion.
Wnt conditioned media
Wnt3a and Wnt5a proteins were used to test the effects on CFU-F and
CFU-O development of BM MNCs and were applied as conditioned
medium collection from mouse L cells that were stably transfected with
either Wnt3a or Wnt5a plasmid [L cells (Cat#: CRL-2648), L-Wnt3a
cells (Cat#: CRL-2647) and L-Wnt5a cells (Cat#: CRL-2814),
American Type Culture Collection (ATCC) Manassas, VA]. Control
conditioned medium (CM), and Wnt3a and Wnt5a CM were prepared
according to the manufacturer’s protocol (ATCC); however, the
during the CM collection, L-cells were grown in both serum (10% FBS)
and under serum-free conditions [ITS-premix (BD Biosciences) in
DMEM (Invitrogen Corp.)]. Briefly, two rounds of conditioned media
were collected, pooled and was diluted with DMEM (1:2) for
experiments.
CFU-F and CFU-O assay
CFU-F and CFU-O potential of MNCs were determined by plating
aliquots (1000 cells/cm2) of cells in basal media [10% FBS (Gibco) in
DMEM (Invitrogen Corp.)] in 24-well polystyrene tissue culture plates
at 378C with 5% humidified CO2 and grown in the conditions
previously described (Baksh et al., 2003). After 14 days, CFU-F cultures
were terminated and stained with Giemsa modified solution (Sigma) to
visualize the cell nuclei and cytoplasm. Colonies of fibroblasts
containing >50 Giemsa positive colonies were counted. After 3 weeks
of culture, CFU-O cultures were stained with alkaline phosphatase
(Sigma) and after 5 weeks of cultures, CFU-O cultures were stained
with alkaline phosphatase (Sigma) followed by staining with 2.5% silver
nitrate (Sigma) (von Kossa stain). Nodular areas staining positive for
silver nitrate were counted.
Limiting dilution assays
Cells were plated in limiting dilution (4  105, 4  104, 4  103 cells per
well of a six-well plate) under CFU-F and CFU-O assay conditions
described above. After 2 and 5 weeks, the number of CFU-F and
CFU-O colonies generated, respectively, was counted. Giemsa staining
was used to identify CFU-F colonies and von Kossa positive-stained
nodular areas were used to identify CFU-O colonies. In order to
determine the clonality of CFU-F and CFU-O as a function of Wnt
treatment, the total number of colonies counted as function of cell
seeding number was plotted and a trend-line was fit to the data and
the R2 value was obtained. In order to calculate the frequency of CFU-F
and CFU-O as a function of the different Wnt treatments, cells were
plated at 1  104 cells/cm2 in a six-well plate, and after 2 and 5 weeks,
the CFU-F and CFU-O colonies, respectively, were counted per total
cells seeded. The number of cells needed to generate 1 colony was then
determined.
[3H] thymidine incorporation
To estimate cell number, MNCs seeded at 5000 cells/cm2 in
12 well-plates were pulsed-labeled with [3H] thymidine (1 mCi/ml)
(Amersham Biosciences) for 24 h at 2, 4, 6 and 8 days. Incorporated
radioactivity was quantified by a Liquid Scintillation Counter.
Reverse Transcription-polymerase chain reaction (RT-PCR)
The expression of cyclin D1 mRNA transcripts in MSCs was examined
by RT-PCR on cultured day 8 cells. The primers used were: (1) human
(h) cyclin D1 (forward, TCC TGT GCT GCG AAG TGG AAAC;
reverse, AAA TCG TGC GGG GTC ATTG) and (2) hGAPDH
(forward, CCA GAA CAT CAT CCC TGC CTC TAC; reverse, GGT
CTC TCT CTT CCT CTT GTG C). Reaction conditions were: 948C
for 3 min, 948C for 308 sec, 558C for 60 sec and 608C for 60 sec;
30 cycles of amplification; and final incubation at 728C for 5 min. The
 Wnts SIGNALING REGULATES MARROW STEM CELLS
amplified products from 1 mg total RNA per sample were resolved on
1.5% Tris-acetic acid-EDTA (TAE; pH 8.0) agarose gel and visualized
with ethidium bromide. All amplified products were detected as single
bands of the expected sizes.
Suspension culture of bone marrow mononuclear cells
BM MNCs were inoculated at a concentration of 1  106 cells/ml into
siliconized (Sigma) 100-ml stirred suspension spinner flasks (Bellco,
Vineland, NJ), as previously described (Baksh et al., 2003), suspended in
50 ml of serum-free Wnt conditioned media (Control, Wnt3a and
Wnt5a CM). At 7, 14, and 21 days, one-half of the medium, including
cells, was removed and replaced with fresh Wnt CM. Cell counts and
CFU-F and CFU-O assays were performed using the aliquot of cells
removed at each time point.
Time course of Wnt effects
Bone marrow MNCs were plated under CFU-F and CFU-O assay
conditions and were assayed for Wnt effects using two experimental
designs. The first experimental design tested the effects of specific
Wnts. Briefly, cells were exposed to Wnt3a and Wnt5a conditioned
media for 2, 4, 6 and 8 days and then replaced with control medium for
the remainder of the assay period. CFU-F and CFU-O assays were
terminated on day 12 and 5 weeks, respectively. The second
experimental design tested the effects of sequential exposure to
different Wnts. Briefly, the cells were first exposed to Wnt3a (or
Wnt5a) for 2, 4, 6 and 8 days, followed by exposure to Wnt5a (or
Wnt3a) for the next 2, 4, and 6 days, respectively. Again, CFU-F assays
were terminated at day 12. For CFU-O assays, the medium was
replaced after day 12 with control medium for the remainder of the
assay period (5 weeks). Cultures in control medium (no Wnt addition)
were performed in parallel.
Autocrine and paracrine effects of Wnts
MSCs were co-transfected (4.8 mg total DNA per 500,000 cells) using a
nucleofection protocol (Amaxa Biosystems; Haleem-Smith et al., 2005)
with a Wnt responsive reporter plasmid (TOP-FLASH) (2 mg) driven
by luciferase (provided by Dr. Bert Vogelstein, Johns Hopkins
University) and with Wnt3a (Upstate Signaling, 2 mg) or Wnt5a
(Upstate Signaling, 2 mg), or transfected with Wnt3a and Wnt5a alone.
Groups transfected with TOP-FLASH were co-transfected with Renilla
luciferase (Upstate Signaling, 0.8 mg) to control for transfection
efficiency. To analyze paracrine effects, equal cell numbers of MSCs
transfected with TOP-FLASH þ Wnt3a or Top TOP-FLASH þ Wnt5a
(Target cells) were co-cultured with Wnt3a- or Wnt5a-transfected
cells (Effector cells). An empty vector (EV) group was used as a control
for Target and Effector cell experiments. To analyze autocrine effects,
TOP-FLASH þ Wnt3a and Top TOP-FLASH þ Wnt5a transfected
MSCs were cultured alone. All groups were cultured in
undifferentiated (basal) and osteogenic conditions. After 4 days,
luciferase activity was assayed using the Dual Luciferase Assay System
according to the manufacturer’s protocol (Promega), and normalized
to Renilla luciferase activity. Activity levels are expressed relative to
that of control (empty vector) conditions. All groups were cultured
long-term under CFU-F and CFU-O assay conditions (as described
above).
Data analysis
Data are presented as mean values  S.D. for at least three
independent experiments performed in triplicate, unless otherwise
stated. Statistical significance is determined by ANOVA and set at
P < 0.05.
Results
Wnt3a and Wnt5a expression in the adult human
bone marrow microenvironment
Immunohistochemical analysis of sections of trabecular BM
revealed the regional arrangement of the cells and extraceullar
matrix (Fig. 1A). Specifically, H&E staining revealed the location
of flattened, elongated cells [expected to be osteoblasts (Obs)
(Calvi et al., 2003)] lining the trabecular bone surface, spindle-
shaped fibroblast-like cells (Fbs) extending throughout the
marrow space and niches of intense H&E positive cells [likely
JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP
819
hematopoietic cells (HCs)] nestled in the network of marrow
sinusoids.
The regions highlighted in Figure 1A were analyzed for the
expression of Wnt3a, Wnt5a, CD45, STRO-1 and Col I using
immunofluorescence microscopy. Imaging of sections revealed
that cells of the marrow sinusoids expressed CD45 (Fig. 1B),
which also co-localized with Wnt5a expression (Fig. 1C and D).
CD45 expressing cells did not express any detectable level of
Wnt3a. The fibroblastic-like cells comprising the bulk of the
marrow cavity did not express CD45 but expressed STRO-1
(Fig. 1E), as well as Wnt5a (Fig. 1F). Trabecular bone lining cells
moderately expressed STRO-1 as well as Wnt5a (Fig. 1G & H).
RT-PCR confirmed the expression of Wnt5a and the lack of
expression of Wnt3a from sorted populations of CD45þ and
CD45
cells directly from bone marrow (data not shown).
Wnt3a expression appeared to be associated with the
underlying stroma matrix, as observed by punctuated staining in
labeled sections (data not shown), which also stained positive
for Col I. Col I was also detected in the connective tissue
comprising the marrow space (Fig. 1I), and also in the bone and
around the marrow vessels (data not shown).
Effects of Wnts on CFU-F and CFU-O recruitment
The spatial distribution of Wnt3a and Wnt5a in the BM
microenvironment suggested a possible regulatory role(s) for
these canonical and non-canonical Wnts in the self-renewal
and/or differentiation of cells residing in the BM. To test this,
BM-derived cells were directly cultured, without further
fractionation onto tissue culture plastic in the presence of
Wnt3a, Wnt5a, or combined Wnt3a/5a conditioned medium
(CM) (from serum-containing CM collection) under CFU-F and
CFU-O assay conditions. After 2 and 5 weeks, CFU-F (Giemsa
positive colonies) (Fig. 2A and B) and CFU-O [(alkaline
phosphatase, day 21) (Fig. 2C and D) and von Kossa positive
colonies) (Fig. 2E)] development, respectively, were assayed.
Enumeration of the CFU-F colony number from each condition
revealed that Wnt3a treatment significantly increased CFU-F
numbers (2-fold), relative to control and Wnt5a treatments
(Fig. 2F). The Wnt3a/5a co-treatment group resulted in
significantly less CFU-F detected relative to Wnt3a treatment.
However, under osteogenic conditions, an 1.5-fold increase
in CFU-O numbers was observed under Wnt5a CM conditions,
relative to both types of controls (Ctrl. and C CM), while the
presence of Wnt3a significantly decreased CFU-O
development both alone and in co-treatment (Fig. 2F). There
was no significant difference in CFU colony number calculated
between the control (Ctrl.) and control CM (C CM) treatment
groups, suggesting that the mouse L-cells did not secrete other
factors that affected CFU-F and CFU-O development.
Clonality of bone marrow-derived MSCs in
response to Wnt exposure
To confirm that each CFU-F and CFU-O arose from a single
cell, and therefore to attribute the effects of Wnt exposure
directly to CFU-F and CFU-O development, BM MNCs were
plated at limiting dilution in Wnt CM under CFU-F and CFU-O
assay conditions (Fig. 3A and B). Regression analyses revealed a
linear relationship between colony number and cell number
(Fig. 3C), thus suggesting that one MSC was limiting for CFU-F
and CFU-O in all Wnt treatment groups (Fig. 3A and B). The
results were used to calculate the CFU-F and CFU-O
frequencies. Quantification of CFU-F and CFU-O frequency as
a function of Wnt exposure revealed that Wnt3a increased
CFU-F frequency (1:4737) relative to control (1:6521), while
Wnt5a increased CFU-O frequency (1:6667) compared to
control (1:10,714) (Fig. 3D).
820 BAKSH AND TUAN

Fig. 1. Spatial distribution of Wnt3a and Wnt5a in adult human trabecular bone marrow. Sections of human trabecular bone specimens
were prepared for histochemical and immunofluorescence examination for the expression of Wnt3a, Wnt5a, CD45, STRO-1 and Col 1.
A: Hematoxylin and eosin stained section of human trabecular bone marrow revealing the spatial arrangement of morphological different
cell types, including osteoblasts (Ob), fibroblasts (Fb), and hematopoietic cells (HCs). Highlighted sections were further analyzed for the
expression ofWnt3a, Wnt5a, CD45, STRO-1 and Col Iusing immunofluorescence microscopy. BandC:depict cells comprising sinusoids ofmarrow
staining positive for CD45 and Wnt5a, respectively, while similar cells (D), closer to trabecular bone, also expressed Wnt5a. F: Fibroblastic
cells comprising the marrow cavity in (A) stained positive for STRO-1 and Wnt5a, respectively. G and H: Trabcular bone lining cells also
expressed STRO-1, and Wnt5a, respectively. The cells shown in (E–H) did not express CD45. I: Collagen type I(Col I) expression in the bone marrow
cavity. All images (B–I) counterstained with Hoechst dye. Scale bar U 50 mm.

JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP

 Wnts SIGNALING REGULATES MARROW STEM CELLS
Fig. 2. Wnt effects on CFU-F and CFU-O development from bone
marrow mononuclear cells. Bone marrow mononuclear cells (BM
MNCs) were plated directly onto tissue culture plastic under CFU-F
and CFU-O assays conditions and grown in the presence or absence of
Wnt3a and Wnt5a conditioned media. A and B: Low and high power
magnification, respectively, of a representative CFU-F colony, stained
with Giemsa, generated from bone marrow MNCs after 14 days. C and
D: Low and high power magnification, respectively, of an alkaline
phosphatase positive CFU-O colony after 21 days of culture under
osteogenic conditions. E: A representative alkaline phosphatase and
von Kossa stained CFU-O colony at 5 weeks of culture. F: Graphic
representation of the number of CFU-F (A) and CFU-O (E) generated
from exposure to Wnt treatment. Analysis revealed the differential
effects of Wnt exposure of CFU-F and CFU-O development. (n U 3,
triplicate; MP < 0.05).
Effects of Wnts on proliferation of bone
marrow-derived MSCs
We next studied the effects of Wnt exposure on the
proliferation of BM MNCs, which contained MSCs by
quantifying [3H]-thymidine incorporation. [3H]-thymidine
incorporation was first measured after 2 days, post BM harvest,
on the adherent cells of the marrow fraction, directly cultured
in different Wnt CM. By day 8 of culture, cells exposed to
Wnt3a demonstrated the greatest [3H]-thymidine
JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP
821
incorporation in basal media (basal) conditions relative to the
other groups (Fig. 4A), which was concomitant with the
increase in cyclin D1 gene expression (Fig. 4B). However, under
osteogenic conditions (OS), Wnt3a did not have the same
[3H]-thymidine incorporation as was observed in basal
conditions. Overall, the addition of Wnt3a or Wnt5a did
not affect the [3H]-thymidine in OS conditions (Fig. 4A).
Quantification of population kinetic parameters from
these proliferation studies revealed that Wnt3a
exposure substantially decreased the population doubling
time (PDt) of MSCs relative to other treatment groups
(Fig. 4C) and therefore, enhanced the cumulative number of
population doublings (Fig. 4D). Interestingly, the presence of
Wnt5a combined with Wnt3a significantly increased the
population doubling time relative to Wnt3a and control
conditions.
Wnt3a increases the number of bone marrow-derived
MSCs capable of CFU-F and CFU-O development
To further understand how Wnts might directly affect MSCs
in vivo, we employed our suspension culture system, which aims
to recapitulate the bone marrow cellular microenvironment,
supporting both hematopoiesis and mesengenesis (Baksh et al.,
2005). Tracking the total number of viable cells as a function of
time revealed that BM MNCs cultured in the presence of
Wnt3a generated higher cell numbers than that observed in
control conditioned media (C CM) and Wnt5a conditions
(Fig. 5A). CFU-F colony number was also the highest in
Wnt3a-supplemented conditions (Fig. 5B). Interestingly,
suspension cells grown in the presence of Wnt3a generated
significantly more CFU-O than from cells cultured in Wnt5a
(Fig. 5B). In contrast, suspension cells grown in the presence of
Wnt5a maintained the total cell number throughout the culture
period and gave rise to similar numbers of CFU-F and CFU-O as
the control condition (Fig. 5B). In all conditions, suspension
cells, when plated under CFU-F and CFU-O assay conditions,
generated discrete colonies, staining positive for Giemsa in
CFU-F assays, after 14 days (Fig. 5C), and alkaline phoshpatase
and von Kossa positive in CFU-O assays, after 5 weeks (Fig. 5D).
Temporal effects of Wnts on CFU-F and CFU-O
development of bone marrow-derived MSCs
Our results strongly suggest a competitive interaction between
canonical Wnt3a and non-canonical Wnt5a in mesenchymal
colony formation from MNCs directly isolated from BM. In light
of these observations, two experimental schemes were
designed to elucidate the mechanism for the competitive
phenomenon that might exist between Wnt3a and Wnt5a
during colony development. The results of the experimental
schemes implemented revealed that Wnt3a increased CFU-F
numbers in a time-dependent manner with maximal colonies
obtained as early as 2 days of Wnt3a exposure (Fig. 6Ai), while
Wnt3a completely abrogated CFU-O formation as early as day
2 (Fig. 6Bi). Short-term exposure to Wnt5a did not significantly
reduce the total attainable CFU-F numbers when cells were
first exposed to Wnt3a (Fig. 6Aii, 2d–2d dataset). However,
long-term, first-time exposure to Wnt5a resulted in
significantly less CFU-F numbers (Fig. 6Aii, 6d–6d dataset). The
initial presence of Wnt3a, both short- and long-term exposure,
significantly inhibited CFU-O development, with significantly
less CFU-O recovered even after exposure to Wnt5a (Fig. 6Bii,
2d–2d, 4d–4d, and 6d–6d datasets).
Autocrine and paracrine effects of Wnts
We observed from our studies that Wnt3a resulted in
increased numbers of CFU-F from bone marrow-derived cells
relative to control or Wnt5a conditions. As a result of these
findings, we ascertained whether the responses of MSCs to
822 BAKSH AND TUAN

Fig. 3. Wnt effects on the clonality of bone marrow-derived MSCs. A and B: Bone marrow mononuclear cells were plated at limiting dilution in
CFU-F and CFU-O assay conditions and exposed to Wnt3a and Wnt5a conditioned media. Colony formation is expressed as a function of cell
density (n U 3, triplicate, mean W SD). C: Regression analyses revealed a linear relationship between cell number and colony number, suggesting
that one cell was limiting for CFU-F and CFU-O development as a function of Wnt exposure. Importantly, Wnt3a and Wnt5a did not affect the cell
autonomous behavior of MSCs to differentiate. D: Wnt3a exposure increased CFU-F frequency (1:4737) relative to control (1:6521), while Wnt5a
increased CFU-O frequency (1:6667) compared to control (1:10,714). Our analyses revealed that one cell type is limiting for CFU-F or CFU-O in the
presence of Wnt3a or Wnt5a, respectively.

Fig. 4. Wnt effects on bone marrow-derived MSC on [3H]-thymidine incorporation. A: Bone marrow-derived cells were plated directly onto
tissue culture plastic and the proliferative potential of the adherent cell population was assessed by [3H]-thymidine incorporation as a function of
exposure to Wnt3a and Wnt5a under basal medium and osteogenic (OS) conditions. Control (Ctrl) were without Wnt supplementation for both
basal and OS conditions. Wnt3a enhanced [3H]-thymidine incorporation, while long-term exposure to Wnt5a suppressed cell proliferation. (n U 3,
triplicate; mean W SD). B: Wnt3a treated cells showed increased cyclin D1 expression, whereas no cyclin D1 was detected in Wnt5a-treated cells
(Day 8 results shown). C and D: Population doubling. Wnt3a treatment resulted in a three-fold reduction in population doubling time (C) relative to
control (26 W 3 h vs.9 W 9 h), with a cumulative population doubling (D) of  6-fold relative to 2.5-fold. Results in B, C, D were all in basal medium
conditions.

JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP

 Wnts SIGNALING REGULATES MARROW STEM CELLS 823

Fig. 5. Wnts effects on the developmental potential of suspension-grown bone marrow-derived cells. Bone marrow MNCs were grown under
suspension culture conditions in the presence and absence of Wnt3a and Wnt5a or in control conditioned media (C CM). A: Total cell number,
(B) CFU-F and CFU-O potential of cells from aliquots of suspension culture at 7, 14 and 21 days, which were plated into CFU-F and CFU-O assay
conditions. C and D: Representative Giemsa positive (CFU-F) and ALP R von Kossa positive (CFU-O) colonies generated from suspension-grown
cells (Wnt3a-exposed suspension cells shown). (n U 3; mean W SD). [Color figure can be viewed in the online issue, which is available at
www.interscience.wiley.com.]

Fig. 6. Temporal effects of Wnt signal transduction on CFU-F and CFU-O development in bone marrow-derived MSCs. Bone marrow-derived
MNCs were plated under CFU-F and CFU-O assay conditions. Ai and Bi: Cultures were treated with Wnt3a and Wnt5a conditioned medium (CM)
for 2, 4, 6, 8 and 12 days and then switched to the control medium for the remainder of the assay period. CFU-F and CFU-O assays were terminated
on day 12 and 5 weeks, respectively. A control group (Ctrl) involved exposing the cells to only control medium for the entire assay culture period
(white bar). Aii and Bii: In parallel studies, bone marrow MNCs were plated in CFU-F and CFU-O assays and cultured in the presence of Wnt3a
or Wnt5a for 2, 4 and 6 days, and then for the subsequent 2, 4 and 6 days, switched to the other Wnt conditioned medium (i.e., Wnt3a to
Wnt5a and Wnt5a to Wnt3a). Bars represent the mean from three donors W SD. These results show that Wnt3a increased CFU-F numbers in a
time-dependent manner with maximal colonies obtained as early as 2 days of Wnt3a exposure, while Wnt3a completely abrogated CFU-O
formation as early as day 2.

JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP

824 BAKSH AND TUAN

Wnts were mediated in a paracrine and/or autocrine fashion,
leading to changes in their developmental potential. For this
purpose, MSCs were transfected with either Wnt3a or Wnt5a
(Effector cells) and co-cultured with MSCs co-transfected with
a Wnt responsive reporter (TOP-FLASH) and with Wnt3a or
Wnt5a (Target cells). Parallel cultures were initiated with MSCs
co-transfected with the TOP-FLASH and with Wnt3a or Wnt5a
in order to test the autocrine effects of Wnts on TOP-FLASH
activity in MSCs. Under undifferentiated conditions, Wnt3a
effector cells enhanced Wnt transcriptional activity in empty
vector (EV)-transfected target cells and significantly enhanced
transcriptional activity in Wnt3a target cells (Fig. 7Ai).
However, Wnt5a effector cells resulted in no significant
enhancement in Wnt transcriptional activity in EV, Wnt3a and
Wnt5a target cells (Fig. 7Ai). There was no significant difference
detected in reporter activity between paracrine and autocrine
effects from Wnt3a effector cells (Fig. 7Aii), suggesting that
MSCs could respond to Wnt3a both in an autocrine and
paracrine fashion with similar efficiency. Interestingly, the
presence of Wnt5a effector cells significantly decreased
reporter activity in Wnt3a effector cells, suggesting that Wnt5a
acted in a paracrine fashion to regulate canonical Wnt signaling
in undifferentiated MSCs. Under osteogenic conditions, basal
levels of Wnt transcriptional activity were observed; however,
surprisingly, Wnt3a effector cells enhanced canonical Wnt
signaling in Wnt5a target cells (Fig. 7Bi).
To study the functional significance of these findings, we
carried out co-cultures of effector and target cells under CFU-F
and CFU-O assay conditions. Co-culture of Wnt3a transfected
cells resulted in significantly more CFU-F than in EV-Wnt3a co-
cultures (Fig. 7C). However, there was no significant difference
between the number of CFU-O detected in EV-Wnt5a and
Wnt5a-Wnt5a co-cultures, suggesting that Wnt5a might
already be operating at saturating levels to induce osteogenic
differentiation in EV-Wnt5a co-cultures. In striking contrast,
EV-Wnt3a co-cultures decreased CFU-O development,
while co-culture of Wnt3a-Wnt3a cells dramatically decreased
CFU-O development, suggesting a dose-dependence for
Wnt3a suppression of osteogenic development.
Discussion
The bone marrow microenvironment is a complex and highly
organized structure, and thus represents a difficult organ to
reconstruct ex vivo (Emerson et al., 1991). Research efforts
have focused on identifying components that are critical in
recapitulating the niche environment for stem cells
maintenance, specifically cues which are required for HSC

Fig. 7. Autocrine and paracrine effects of Wnts on canonical Wnt responsive transcriptional activity. A and B: Effector cells (Wnt-transfected
MSCs) and target cells (MSCs co-transfected with either Wnt3a or Wnt5a and TOP-Flash reporter plasmid) were co-cultured under
undifferentiated and osteogenic conditions. Aii and Bii: In parallel, target cells were cultured alone. Luciferase activity was determined on day 4
and the results were expressed relative to EV conditions. C: CFU-F and CFU-O development. Equal numbers (1:1) of Wnt-transfected MSCs
(i.e., empty vector (EV)-Wnt3a, EV-Wnt5a, Wnt3a-Wnt3a, Wnt5a-Wnt5a, and Wnt3a-Wnt5a) were co-cultured under CFU-F and CFU-O assays
conditions. Bars represent the mean W SD from three donors. These results showed the paracrine and autocrine action of Wnt3a in enhancing
canonical Wnt signaling and CFU-F development, while the paracrine action of Wnt5a acts to antagonize canonical Wnt signaling and enhance
CFU-O development.

JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP

 Wnts SIGNALING REGULATES MARROW STEM CELLS
self-renewal and differentiation (Long et al., 1992; Mayani et al.,
1992; Calvi et al., 2003; Moore, 2004). Many types of factors and
signaling pathways have been shown to affect HSC development
(Bearman, 1997; Larsson and Karlsson, 2005; Ruscetti et al.,
2005) and more recently, Wnts (Brandon et al., 2000), all of
which are critical in supporting HSC propagation ex vivo.
Recent evidence supports the role of Wnts in regulating MSCs
(Boland et al., 2004; Etheridge et al., 2004), however, these
studies were performed on ex vivo, adherent cultures of bone
marrow-derived MSCs. Thus, little is understood about the
involvement of Wnts in supporting the maintenance, growth
and maturation of MSCs in the BM. Our work, therefore,
focused on elucidating the role that Wnts play in MSC
development in BM. In order to claim that Wnts play a
regulatory role in BM, we first demonstrated the presence of
Wnts, specifically Wnt3a and Wnt5a, in the bone BM niche
environment. Our studies revealed the regional association of
Wnts in isolated BM specimens and therefore, supports the
hypothesis that Wnts have the potential to regulate cells in the
marrow environment. Our analysis revealed that both non-
hematopoietic (lacking CD45 expression) and hematopoietic
cells (expressing CD45) expressed Wnt5a protein, while
neither of these populations expressed any detectable levels of
Wnt3a. Sorting CD45þ and CD45
cells, confirmed the
expression of Wnt5a and the lack of expression of Wnt3a, as
determined by RT-PCR (data not shown). These findings
corroborate with those of Van Den Berg et al. (1998) who
showed that Wnt5a is expressed by highly enriched populations
of hematopoietic progenitor cells (CD34þ Lin
). We also
confirmed that those cells expressing Wnt5a, also expressed
STRO-1, a marker typically used to identify MSCs in culture
(Dennis et al., 2002; Stewart et al., 2003). STRO-1 was not
detected in hematopoietic cells. Taken together, these findings
indicate that Wnt5a is ubiquitously expressed in
morphologically and functionally different populations of cells
in bone marrow. On-going work is focused on elucidating
the functional role of Wnt5a in MSCs using small hairpin (sh)
RNA approaches.
While Wnt3a protein expression was not expressed in any
cell type of the marrow, a moderate level of Wnt3a expression
was detected in the underlying stroma matrix, which we have
shown to express collagen type I (Fig. 1I). Willert et al. (2003)
reported that exogenous supplementation of Wnt3a to growth
medium enhances the ex vivo self-renewal potential of HSCs
(Willert et al., 2003). Thus, we can hypothesize that the
expression of Wnt3a protein in the marrow sinusoids (Fig. 1A),
may play a functional role in modulating the self-renewal of
HSCs in vivo. The lack of Wnt3a gene expression in both
hematopoietic and non-hematopoietic cells raises questions
regarding which cells are responsible for the deposition of
Wnt3a in the bone marrow, whether or not Wnt3a is
constitutively present or temporally deposited, or whether
these Wnts regulate their target cells in an autocrine and/or
paracrine fashion. Future studies are underway to address many
of these types of questions.
Given the expression profile of Wnt5a and Wnt3a in BM, we
next aimed to confirm whether these Wnts modulated the
developmental potential of MSCs. To explore this hypothesis,
we employed a suspension culture approach, which aims to
recapitulate the BM cellular environment (Baksh et al., 2003) in
order to study the effects of Wnts on MSCs directly isolated
from BM, prior to adherent culture on tissue culture plastic. In
suspension culture, Wnt3a enhanced the proliferation of BM
MNCs and increased the number of CFU-F and CFU-O
generated from suspension-grown cells, while Wnt5a
maintained input cell numbers (Fig. 5). The proliferative effects
which we observed from exposure to Wnt3a corroborate our
previous findings of the ability of Wnt3a to enhance the
proliferation of adherent-cultured MSCs (Boland et al., 2004).
JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP
825
In contrast, exposure of MNCs to Wnt5a maintained the
number of MSCs capable of both CFU-F and CFU-O
development throughout suspension culture. Our findings
indicate that exposure to Wnt3a affects the self-renewal and
proliferation of MSCs in the BM population capable of
fibroblastic and osteogenic development, while Wnt5a
maintains the steady-state number of MSCs, and thereby, does
not affect the proliferation or differentiation potential MSCs in a
simulated BM cellular microenvironment. Even when BM
MNCs were plated directly onto tissue culture plastic and
cultured in the presence of Wnt3a, a significant increase in the
recruitment of CFU-F colonies and cell proliferation was
observed relative to control conditions (absence of Wnt
supplementation). Many studies have shown the critical role
that growth factors, such as, platelet-derived growth factor,
play in regulating the growth and development of CFU-F from
primitive bone marrow-derived MSCs grown under serum-free
conditions (Gronthos and Simmons, 1995). To this end, Wnt3a
may play a similar role in enhancing CFU-F recruitment and
development and thus, may be used as realistic approach to
generate increased numbers of MSCs from BM for therapeutic
applications.
Under osteogenic conditions, Wnt5a increased CFU-O
colony number but did not affect CFU-F numbers in
undifferentiated conditions and cell proliferation directly on
tissue culture plastic relative to control conditions. These
findings suggest that Wnt5a is unlikely to affect the
differentiation potential of bone marrow-derived MSCs under
contact-dependent growth conditions. Thus, Wnt5a may play
environmentally-dependent roles in modulating MSC
development, that is, we hypothesize that in BM, Wnt5a
maintains the MSC pool in a quiescent state, while in ex vivo
culture, Wnt5a enhances osteogenesis of MSCs. Interestingly,
co-treatment of Wnt3a and Wnt5a results in inhibition of CFU-
O and CFU-F development, relative to Wnt-alone treatment,
respectively, suggesting a potential cross-talk between
canonical and non-canonical Wnt signaling pathways, such as
that shown in our previous work (Baksh et al., 2006).
Importantly, we confirmed in these studies that one cell was
limiting for colony formation under these Wnt conditions
(Fig. 3), without the involvement of cooperativity of cell types
necessary in colony formation, as proposed by Aubin (Aubin,
1999). Taken together, our results from these studies reveal the
possibility that Wnts might regulate the self-renewal and
differentiation potential of MSCs in the bone marrow, a concept
which is proposed for hematopoietic stem cells (Brandon et al.,
2000).
Results from the time-course study revealed that 2-day
exposure to Wnt3a was sufficient to recruit the maximal
number of CFU-F from BM MNCs relative to constant Wnt3a
exposure, while prolonged exposure to Wnt5a significantly
decreased the total number of attainable CFU-F (Fig. 6Ai). The
ability to recruit maximal CFU-F colonies with only 2 days of
Wnt3a exposure suggests that the action of Wnt3a in
producing a biological effect is rapid and efficient. These findings
provide insight into the clinical utility of Wnt3a in enhancing the
number of MSCs from BM. In contrast, under osteogenic
conditions, exposure to Wnt5a for 4-days was necessary to
recruit enhanced numbers of CFU-O, while short-term
exposure (2 days) to Wnt3a resulted in significantly less CFU-O
(Fig. 6Bi). These results suggest a different time scale for Wnt5a
action compared to Wnt3a action in the recruitment of
colongenic MSCs. Furthermore, six days of exposure to Wnt5a
followed by 6 days of Wnt3a exposure did not significantly
impact the CFU-O development from BM MNCs (Fig. 6Bii),
suggesting that long-term exposure to Wnt5a is necessary for
the irreversible commitment of bone marrow-derived MSCs to
osteogenesis. Early, initial exposure to Wnt3a significantly
impaired the long-term ability of the cells to form a functional
826 BAKSH AND TUAN
osteogenic phenotype (Fig. 6Bii). Specifically, subsequent 6 days
of exposure to Wnt5a was insufficient to allow the cells to
recover from exposure to Wnt3a, and thereby commit to
osteogenesis. Collectively, these findings reveal the
importance of the temporal regulation of the Wnt signal
transduction pathway in maintaining the developmental
potential of MSCs.
Another important aspect is the mode by which these effects
of canonical and non-canonical Wnts signals are transmitted,
i.e., autocrine versus paracrine. To address this issue, equal
numbers of effector and target cells were co-cultured under
differentiated and osteogenic conditions. Our findings revealed
that Wnt3a effector cells enhanced the endogenous Wnt
transcriptional activity in control (empty vector) MSC target
cells, while Wnt3a effector cells synergistically increased Wnt
transcriptional activity in Wnt3a target cells in a paracrine
fashion (Fig. 7A). However, there was no significant difference
observed between the Wnt transcriptional activity observed
from Wnt3a effector cells on Wnt3a target cells in either
paracrine and autocrine condition (Fig. 7A), suggesting that
there is a saturation limit in the dose required for Wnt3a to
exert its effects. Nevertheless, our findings indicate that Wnt3a
can act in both an autocrine and paracrine fashion to enhance
Wnt transcriptional activity. In contrast, under undifferentiated
and osteogenic conditions, we observed that Wnt5a negatively
regulated canonical Wnt transcriptional activity of Wnt3a
effector cells (Fig. 7A and B). Interestingly, under osteogenic
conditions, Wnt3a effector cells upregulated Wnt
transcriptional activity in Wnt5a effector cells, suggesting that
Wnt3a has a potent effect on altering canonical Wnt signaling
on cells induced to differentiate and those which express
Wnt5a. Furthermore, the observed autocrine and paracrine
effects on Wnt transcriptional effects also affected the CFU-F
and CFU-O development in a similar manner (Fig. 7C). Taken
together, these findings suggest that Wnt3a can affect the
development of differentiated cells types in BM, such as those
identified to express Wnt5a (Fig. 1), in a paracrine fashion and
thus affects the CFU-F and CFU-O development of MSCs.
In summary, the spatial association of Wnt3a and Wnt5a in
adult human BM in situ suggests that Wnts may represent
microenvironmental cues that govern the maintenance,
proliferation and differentiation of MSCs in BM. In vitro
observations support this possibility. Thus, Wnt3a enhances
the recruitment of CFU-F and MSC proliferation, both in
suspension culture and in adherent growth conditions, while
Wnt5a maintains the number of BM-derived MSCs capable of
CFU-F and CFU-O development under suspension culture
conditions. However, our work also shows that Wnt5a
suppresses CFU-F and MSC proliferation, but increases CFU-O
development in BM MNCs directly cultured in osteogenic
medium under plastic adherent conditions, thus revealing a
potential dual role for Wnt5a on MSC development. Elucidating
how Wnts and other factors act as instructive cues for the
recruitment, maintenance and maturation of bone marrow-
derived MSCs is of primary interest given the potential use of
these cells in regenerative medicine.
Acknowledgments
This work is supported by the Intramural Research Training
Program of the National Institute of Arthritis, and
JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP
Musculoskeletal and Skin Diseases, National Institutes of Health
(Z01 AR41131).
Literature Cited
Aubin JE. 1999. Osteoprogenitor cell frequency in rat bone marrow stromal populations: role
for heterotypic cell-cell interactions in osteoblast differentiation. J Cell Biochem 72:396–
410.
Baksh D, Davies JE, Zandstra PW. 2003. Adult human bone marrow-derived mesenchymal
progenitor cells are capable of adhesion-independent survival and expansion. Exp Hematol
31:723–732.
Baksh D, Davies JE, Zandstra PW. 2005. Soluble factor crosstalk 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.
Blood 106:3012–3019.
Baksh D, Boland G, Tuan RS. 2007. Cross-talk between Wnt signaling pathways in human
mesenchymal stem cells leads to functional antagonism during osteogenic differentiation.
J Cell Biochem In Press.
Bearman SI. 1997. Use of stem cell factor to mobilize hematopoietic progenitors. Curr Opin
Hematol 4:157–162.
Boland GM, Perkins G, Hall DJ, Tuan RS. 2004. Wnt 3a promotes proliferation and suppresses
osteogenic differentiation of adult human mesenchymal stem cells. J Cell Biochem
93:1210–1230.
Brandon C, Eisenberg LM, Eisenberg CA. 2000. WNT signaling modulates the diversification
of hematopoietic cells. Blood 96:4132–4141.
Bruder SP, Jaiswal N, Haynesworth SE. 1997. Growth kinetics, self-renewal, and
the osteogenic potential of purified human mesenchymal stem cells during
extensive subcultivation and following cryopreservation. J Cell Biochem 64:278–
294.
Calvi LM, Adams GB, Weibrecht KW, Weber JM, Olson DP, Knight MC, Martin RP,
Schipani E, Divieti P, Bringhurst FR, Milner LA, Kronenberg HM, Scadden DT.
2003. Osteoblastic cells regulate the haematopoietic stem cell niche. Nature
425:841–846.
Caplan AI. 1991. Mesenchymal stem cells. J Orthop Res 9:641–650.
Clark BR, Keating A. 1995. Biology of bone marrow stroma. Ann N Y Acad Sci
770:70–78.
De Boer J, Wang HJ, Van Blitterswijk C. 2004. Effects of Wnt signaling on proliferation
and differentiation of human mesenchymal stem cells. Tissue Eng 10:393–401.
Dennis JE, Carbillet JP, Caplan AI, Charbord P. 2002. The STRO-1þ marrow cell population is
multipotential. Cells Tissues Organs 170:73–82.
Emerson SG, Palsson BO, Clarke MF. 1991. The construction of high efficiency human bone
marrow tissue ex vivo. J Cell Biochem 45:268–272.
Etheridge SL, Spencer GJ, Heath DJ, Genever PG. 2004. Expression profiling and
functional analysis of wnt signaling mechanisms in mesenchymal stem cells. Stem Cells
22:849–860.
Friedenstein AJ, Ivanov-Smolenski AA, Chajlakjan RK, Gorskaya UF, Kuralesova AI,
Latzinik NW, Gerasimow UW. 1978. Origin of bone marrow stromal
mechanocytes in radiochimeras and heterotopic transplants. Exp Hematol
6:440–444.
Gronthos S, Simmons PJ. 1995. The growth factor requirements of STRO-1-positive human
bone marrow stromal precursors under serum-deprived conditions in vitro. Blood
85:929–940.
Haleem-Smith H, Derfoul A, Okafor C, Tuli R, Olsen D, Hall DJ, Tuan RS. 2005. Optimization
of high-efficiency transfection of adult human mesenchymal stem cells in vitro. Mol
Biotechnol 30:9–20.
Janowska-Wieczorek A, Majka M, Ratajczak J, Ratajczak MZ. 2001. Autocrine/paracrine
mechanisms in human hematopoiesis. Stem Cells 19:99–107.
Koller MR, Manchel I, Palsson BO. 1997. Importance of parenchymal:stromal cell ratio for the
ex vivo reconstitution of human hematopoiesis. Stem Cells 15:305–313.
Larsson J, Karlsson S. 2005. The role of Smad signaling in hematopoiesis. Oncogene 24:5676–
5692.
Long MW, Briddell R, Walter AW, Bruno E, Hoffman R. 1992. Human
hematopoietic stem cell adherence to cytokines and matrix molecules. J Clin Invest
90:251–255.
Mayani H, Guilbert LJ, Janowska-Wieczorek A. 1992. Biology of the hemopoietic
microenvironment. Eur J Haematol 49:225–233.
Moore KA. 2004. Recent advances in defining the hematopoietic stem cell niche. Curr Opin
Hematol 11:107–111.
Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, Mosca JD, Moorman MA,
Simonetti DW, Craig S, Marshak DR. 1999. Multilineage potential of adult human
mesenchymal stem cells. Science 284:143–147.
Ruscetti FW, Akel S, Bartelmez SH. 2005. Autocrine transforming growth factor-beta
regulation of hematopoiesis: many outcomes that depend on the context. Oncogene
24:5751–5763.
Stewart K, Monk P, Walsh S, Jefferiss CM, Letchford J, Beresford JN. 2003. STRO-1, HOP-26
(CD63), CD49a and SB-10 (CD166) as markers of primitive human marrow stromal cells
and their more differentiated progeny: a comparative investigation in vitro. Cell Tissue Res
313:281–290.
Van Den Berg DJ, Sharma AK, Bruno E, Hoffman R. 1998. Role of members of the Wnt gene
family in human hematopoiesis. Blood 92:3189–3202.
Willert K, Brown JD, Danenberg E, Duncan AW, Weissman IL, Reya T, Yates JRIII, Nusse R.
2003. Wnt proteins are lipid-modified and can act as stem cell growth factors. Nature
423:448–452.