Leukemia (1999) 13, 428–437
 1999 Stockton Press All rights reserved 0887-6924/99 $12.00
http://www.stockton-press.co.uk/leu
Bone morphogenetic protein 9 is a potent synergistic factor for murine hemopoietic
progenitor cell generation and colony formation in serum-free cultures
RE Ploemacher1, LJA Engels1, AEM Mayer1, S Thies2 and S Neben2
1
Institute of Hematology, Ersasmus University, Rotterdam, The Netherlands; and 2Genetics Institute, Cambridge, MA, USA
The effect of the recently cloned cytokine bone morphogenetic
protein 9 (BMP-9) on colony formation and generation in vitro
clonable hematopoietic progenitors (CFU-C) in serum-free
liquid cultures (LC) of both normal and post-5-fluorouracil
murine bone marrow cells was studied in the presence of various other cytokines. In LC, BMP-9 concentrations of 100 ng or
more per ml led to complete inhibition of Steel Factor (SF) +
interleukin-11 (IL-11) or IL-12 supported CFU-C generation,
which was partly abrogated when IL-3 was additionally
included. We found this inhibitory effect of BMP-9 to be
mediated by an increased TGF-␤1 elaboration and TGF-␤1
mRNA expression in bone marrow cells with increasing BMP-9
concentrations. In the presence of neutralizing antibodies (Ab)
against TGF-␤1, BMP-9 concentrations of 3 ng or higher synergized with IL-3, SF+IL-3, SF+IL-11/12, or IL-3+SF+IL-11/12 to
increase CFU-C generation. Similarly, high BMP-9 concentrations dramatically inhibited primary colony formation
induced by SF+IL-11/12, whereas in the presence of TGF-␤1
neutralizing Ab only 3 ng or more BMP-9 per ml stimulated both
the time of colony appearance, the colony size and colony numbers in the presence of IL-3, M-CSF, GM-CSF, SF, SF+Flt3-L,
SF+IL-3, SF+IL-11/12 or IL-3+SF+IL-11/12. BMP-9 neither stimulated CFU-C generation nor colony formation as a single factor,
nor did it synergize with thrombopoietin (Tpo), erythropoietin
(Epo), Flt3-L, IL-11, IL-12 or G-CSF. The effect of BMP-9 on its
target cells was direct as demonstrated using single-sorted
stem cells. These observations demonstrate that BMP-9 plays
a dual role in regulating proliferation of primitive hemopoietic
progenitor cells. Thus, in addition to its ability to enhance TGF␤1 elaboration in bone marrow cells, it acts as a potent synergistic activity that is different from SF, Flt3-L, IL-11 or IL-12.
BMP-9 mRNA was exclusively detected in the liver of adult
mice, whilst no expression was found in stromal cell lines
propagated from day-16 fetal liver or neonatal or adult bone
marrow. 125I-BMP-9 bound specifically to a high percentage of
blast cells in lineage-depleted post-fluorouracil bone marrow
cells and to megakaryocytes in normal and post-fluorouracil
bone marrow, indicating that BMP-9R are expressed on these
cells. The dissociation between the site of BMP-9 production
and its target cells in the bone marrow makes BMP-9 a hemopoietic hormone.
Keywords: BMP-9; stem cells; TGF-␤1; mouse
Introduction
Bone morphogenetic protein-9 is a new member of the transforming growth factor (TGF)-␤ superfamily.1,2 Individual members of this family function in a diverse array of various embryonic and adult organ systems.3–7 The bone morphogenetic
proteins (BMPs) are a group of related proteins, many of which
are capable of inducing the formation of new cartilage and
bone. In bone marrow stroma they induce the osteoblast
phenotype and promote osteogenesis, while simultaneously
inhibiting other mesodermal differentiation pathways, such as
Correspondence: RE Ploemacher, Department of Hematology,
Ee1391, Erasmus University, PO Box 1738, 3000 DR Rotterdam, The
Netherlands; Fax: 31 10 436 7601
Received 27 November 1998; accepted 14 December 1998
adipogenesis and myogenesis.8 In contrast to the disperse
expression pattern of individual TGF-␤ family members, BMP9 expression in the fetus has been described to be restricted
to the liver.1 RhBMP-9 binds to high affinity receptors
expressed on primary rat hepatocytes and on the human liver
tumor cell line, HepG2, and additionally induces proliferation
in these cells.2 On the basis of these findings it has been proposed that BMP-9 may play a physiological role in stimulating
hepatocyte growth. The fact that BMP-9 also stimulates alkaline phosphatase activity in a murine osteoprogenitor cell line,
W-20-17, suggests that other tissues may also be physiological
targets for BMP-9.2 A broader role for BMP is also supported
by the observation that 125I-BMP-9 binds to plasma membrane
preparations from adult rat kidney, lung, heart, but not cerebrum, intestine, muscle or blood (S Thies, personal
communication).
Regulation of quiescence and proliferation in the hemopoietic stem cell compartment probably is a complex interplay
between inhibitory and stimulatory activities that may predominantly act locally. Due to the inherent dispersed nature
of hemopoiesis in the body there should exist more systemic
regulatory signals controlling the rate of hemopoiesis in
answer to immediate demands for blood cells. Such controls
seem to be fulfilled by a number of cytokines (IL-1, TNF-alfa,
TGF-␤, IFN-g) that have (either direct or indirect) pleiotropic
activities and are not easily detectable in serum under steady
state conditions. Because many members of the TGF-␤ superfamily show pleiotropism, we set out to test the recently
characterized members of the BMP group. We made use of a
well-calibrated serum-free liquid culture protocol for ex vivo
generation of murine hemopoietic progenitor cells and stem
cells. This assay has allowed us to detect both potent inhibitory and stimulatory synergistic activities of BMP-9 on hemopoietic progenitor cells.
Materials and methods
Mice
Male and female (C57BLxCBA)F1 (BCBA) mice, 12 weeks old,
were bred in the Central Animal Department of the Erasmus
University Rotterdam, The Netherlands, and maintained under
clean conventional conditions. The microbiological status of
the animals was SPF-5. The drinking water was acidified to
pH 2.8. In part of the experiments, bone marrow donor mice
were injected with 150 mg 5-fluorouracil (Sigma, St Louis,
MO, USA) in phosphate-buffered saline (PBS) per kg of body
weight in a lateral tail vein at day 6 before harvesting the
marrow.
Cell preparation
Bone marrow cells (BMC) were prepared by cleaning femora
and tibiae from muscles and tendons and gentle crushing in
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
a mortar using PBS+5% fetal bovine serum (FBS). The cell suspensions were sieved over a nylon filter (mesh size 100 ␮m).
For buoyant density centrifugation a modification of a discontinuous Ficoll-400 (Pharmacia Fine Chemicals, Uppsala,
Sweden) gradient designed to isolate thymocyte and bone
marrow subpopulations9 was used. Cells with a density of
1.069–1.078 g/ml were collected, washed twice in PBS and
5% FBS, and properly diluted. This cell preparation is further
designated LD/FU6dBMC. In some experiments, LD normal
BMC and LD/FU6dBMC were further enriched for hemopoietic
progenitor cells by depletion of mature lineage-restricted cells.
Briefly, these cells were incubated with mAbs against CD4,
CD8,B220, Gr-1, Mac-1 and Ter119 antigens (Pharmingen),
followed by goat anti-rat microbeads (Miltenyi) and negative
selection on MACS BS columns (Miltenyi, Bergisch Gladbach,
Germany). This fraction will be referred to as the LIN− preparation.
Fluorescence activated cell sorting (FACS)
FACS was done essentially as described previously.9 Briefly,
LD/FU6dBMC cells were incubated with 0.25 ␮g/ml fluorescinated wheat germ agglutinin (WGA-FITC; Vector, Burlingame, CA, USA) prior to sorting using a FACS Vantage
(Becton Dickinson, Sunnyvale, CA, USA) with a single argon
laser at 488 nm tuned at 350 mW output. FITC intensity were
measured using a 510/515 LP and 540 SP filter set. Cells were
sorted individually into each well of two microtiter plates per
group using a single cell deposition unit. WGA-bright cells
represented the highest 25% of all WGA-binding cells and 8%
of all LD/FU6dBMC cells.
Growth factors
Recombinant human BMP-9 was purified from CHO cells as
previously described6 and batches contained 29–492 ␮g/ml.
Recombinant human erythropoietin (Epo) was obtained from
Merckle (Ulm, Germany). Purified recombinant human thrombopoietin (Tpo) was obtained from Genentech (San Francisco,
CA, USA) and its protein concentration was 480 ␮g/ml.
Recombinant murine IL-3 was used as conditioned medium
from CHO cells containing the appropriate cDNA and its biological activity was determined by stimulation of RB5. Recombinant murine GM-CSF was used as conditioned medium
from transfected COS-1 cells and the activity was determined
by RB5 assay, while recombinant murine G-CSF was also produced in COS-1 and the biological activity was 16 300 U/ml
as determined by RB5 proliferation assay. Recombinant
human M-CSF was purified from CHO cells expressing M-CSF
cDNA and batches contained 860 ␮g/ml with a biological
activity of 1.65 × 106 U/ml as determined by 32Dcfms cell
proliferation using 3H-thymidine incorporation. Purified
recombinant human IL-11 (produced in E. coli) had a specific
activity greater than 1 × 106 U/mg as determined by the T10
cell proliferation assay. Purified recombinant human Flt3-L
was produced in E. coli and contained 700 ␮g/ml. Soluble
recombinant murine c-kit ligand (Steel Factor; SF) was purified
from CHO cells to give batches containing 322 ␮g protein/ml.
Heterodimeric recombinant murine IL-12 was purified from
the conditioned medium of NSO myeloma cells containing
an expression vector producing both murine p35 and p40 subunits. The specific activity of the pure material, measured with
a human PHA-blast cell assay, was approximately 3 × 106
units/mg. BMP-9, IL-3, IL-11, IL-12, GM-CSF and M-CSF were
kindly provided by the Genetics Institute (Cambridge, MA,
USA), SF and Flt3-L by Amgen (Thousand Oaks, CA, USA).
Unless otherwise indicated concentrations of factors were as
follows: SF, 50 ng U/ml; Flt3-L, 50 ng/ml; IL-3, 5 ng/ml; IL11, 50 ng/ml; IL-12, 1 ng/ml; rhG-CSF, 20 ng/ml; M-CSF, 100
ng/ml; GM-CSF, 20 ng/ml; Tpo, 100 ng/ml; Epo, 1 U/ml.
Clonogenic assays
Quantitation of CFU-C was carried out in duplicate to quadruplicate 35-mm bacteriological dishes (627102, Greiner,
Alfen aan de Rijn, The Netherlands) in a fully humidified incubator at 37°C and 5% CO2. One milliliter of culture contained
either 2–10 × 103 LD/FU6dBMC or 5–50 × 103 NBMC in semisolid (0.8% methylcellulose; Methocel AP4 Premium, Dow
Chemical, Rotterdam, The Netherlands) serum-free culture
medium (IMDM with Glutamax-1; Gibco, Breda, The Netherlands, cat No. 31980-022 and penicillin/streptomycin). The
medium was complemented with 1 × 10−4 mol/l 2-mercaptoethanol (Merck, Darmstad, Germany), 1% bovine serum albumin (Sigma), 10 ␮g/ml bovine insulin (Gibco), 15 × 10−3 ␮m
cholesterol (Sigma), 15 × 10−3 ␮m linolic acid (Merck), 154
␮g/ml iron-saturated human transferrin (Behring Institute, Marburg, Germany), a nucleoside mix of 1 ␮g/ml of cytidine,
adenosine, uridine, guanosine, 2⬘-deoxycytidine, 2⬘-deoxyadenosine, thymidine, 2⬘-deoxyguanosine (all from Sigma) and
cytokines as specified per experiment. Colonies (⭓50 cells)
and clusters (8–49 cells) were counted on various time points
between day 5 and 21 of culture with an inverted microscope.
In some experiments colonies were plucked individually from
the cultures on day 14, May–Grünwald-Giemsa stained and
the presence of specific lineages (either monocyte, macrophage, neutrophil, eosinophil, erythroid, megakaryocytic,
blast cell or mast cell) per colony was determined. For quantitation of liquid culture CFU-C output the CFU-C culture
medium included 20% horse serum, while 10% of pokeweed
mitogen-stimulated mouse spleen-conditioned medium
(PWM-MSCM) was used as growth stimulus. Colonies in all
cultures were counted at day 7 and 14 of culture with an
inverted microscope.
Liquid culture
Serum-free liquid suspension cultures were performed in 1-ml
volumes in 35-mm bacteriologic dishes (627102, Greiner)
with one to four dishes per group according to the anticipated
cell number required for replating in the CFU-C assays. The
medium was similar to that used for serum-free primary colony assays. The cells were incubated for 5–7 days at 37°C
and 10% CO2 in a fully humidified incubator. At the end of
the incubation period cells were replated in the CFU-C assay
with 20% horse serum and 10% pokeweed mitogen-stimulated mouse spleen cell conditioned medium (PWM-MSCM)
as colony-stimulating activity.
HepG2 assay
The human liver tumor cell line, HepG2 (ATCC HB8065) was
obtained from the American Type Culture Collection
(Rockville, MD, USA). The cells were plated at 105 cells in
200 ␮l in 96-well plates and cultured for 44 h in IMDM/0.1%
429
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
430
FCS with or without 10 or 300 ng rhBMP-9/ml, 0.3 or 10 ng
TGF-␤1/ml or 1 ␮g of anti-TGF-␤1 antibodies/ml. 3H-thymidine was included in the last 20 h of the incubation period and
cellular DNA was harvested with a 96-well plate cell harvester. Incorporated 3H-thymidine was measured by liquid
scintillation counting.
expression was monitored by using primers for HPRT as a
housekeeping gene.
Results
BMP-9 inhibits CFU-C generation in LC
BMP-9 binding
For binding experiments, suspension of spleen, thymus, normal BMC (NBMC), FU6dBMC or lineage-depleted (Lin−) fractions of NBMC or FU6dBMC were centrifuged on gelatinized
glass slides. The cells were incubated with 1 ng/ml 125I-BMP9 with or without 100 ng/ml of unlabeled rhBMP-9 in binding
buffer (50 mm Tris, 100 mm NaCl, 5 mm CaCl2, 3% BSA, 0.1
mm bacitracin, 0.15 ␮m aprotinin, 1 mm pefabloc, 8 mm leupeptin, pH 7.4) for 2 h at 25°C following a 30 min preincubation at 37°C in binding buffer alone. Cells were washed
three times in ice-cold binding buffer, fixed in 0.5% glutaraldehyde in binding buffer and dried in cool air. The dried slides
were coated with Ilford K.5 emulsion, exposed for 3 weeks
and stained with toluidine blue. RhBMP-9 was iodinated to a
specific activity of 50–100 ␮Ci/␮g by the methods of Frolik
et al.10 Iodinated rhBMP-9 and unlabeled rhBMP-9 exhibited
similar potencies on HepG2 cell proliferation.2
Among a series of BMPs that we tested in the 10–50 ng ranges
on their ability to affect maintenance and generation of progenitors and stem cells in liquid cultures (LC), rhBMP-9 was
inhibitory. To determine a dose response for the inhibitory
effect of BMP-9 we performed 5-day serum-free LC of 2 × 104
LD/FU6dBMC stimulated by either SF+IL-12, or SF+IL-12+IL-3
(5 U/ml). Previously we have found these cytokine combinations to synergize in progenitor cell generation and colony
formation of murine bone marrow cells,12,13 and to be
extremely useful for detecting the presence of very low concentrations of TGF-␤1 (ie down to 1 pg/ml).14,15 From Figure
1 it appears that BMP-9 concentrations of 100 ng and higher,
completely abrogated CFU-C generation stimulated by SF+IL12, and almost completely when low concentrations of IL-3
(5 U) were included. The same effect could be observed if IL12 was replaced by IL-11 (data not shown). These observations indicate that BMP-9 can act as a potent inhibitory
activity for CFU-C generation. Similarly, 100 ng BMP-9 completely inhibited colony formation by LD/FU6dBMC in the
presence of SF+IL-12 (data not included in Tables or Figures).
RNA assay
For TGF-␤1 mRNA expression, 105 LD/FU6dBMC were cultured for 4 days in 33 mm-diameter bacterial petri dishes containing serum-free medium with 50 ng SF, 1 ng IL-12 with or
without 100 ng BMP-9 and 1 ␮g anti-TGF-␤1 antibody per
ml. FBMD-1 stromal cells were cultured to confluency, the
10% horse serum containing medium changed for the same
medium as used for the LD/FU6dBMC, and the cells were harvested by trypsinization after 4 days. For BMP-9 mRNA
expression, total RNA was prepared from adult mouse liver,
spleen, bone marrow, lymph nodes, thymus, and from bone
marrow derived (FBMD-1, NBM-11F4G, NB-11F10H, NBM6D3D) or fetal liver-derived (FL-7H7B and FL-1E2C) stromal
cells.
Total cellular RNA was isolated from tissues or cell suspensions by extraction with TRIzol Reagent (Gibco, cat No.
15596-026). cDNA synthesis11 was performed starting with 1–
3 ␮g RNA, after heating for 10 min at 65°C. Semi-quantitative
PCR was performed on 3–5 cDNA batches prepared from different experiments using a 100 ␮l mixture containing 1/20 of
the cDNA products (in 50 ␮l). TAq polymerase buffer (100
mm Tris-HCl, 500 mm KCl, 15 mm MgCl2 and 1% gelatin; pH
8.3), 0.2 mm of each dNTP (Pharmacia) and 1 unit of Taq
polymerase (Promega). In addition, 50 pmol of the following
sense and anti-sense primers were added: HPRT, 5⬘-GTA ATG
ATC AGT CAA CGG GGG AC and 3⬘-CCA GCA AGC TTG
CAA CCT TAA CCA, product size 179 bp; TGF-␤1, 5⬘-GAC
GTC AAA AGA CAG CCA CT, 3⬘-GAA GCC ATC CGT GGC
CAG AT, product size 461; BMP-9,5⬘-GAAGGTGGAT TT
CCTACGCA, 3⬘-CACTGTCACC TCGAGCTTAT, product size
504 bp. After 15, 20, 25, or 30 cycles for TGF-␤1, or 45 cycles
for BMP-9, 10 ␮l of the PCR product was removed from the
reaction tube, samples were submitted to electrophoresis in a
2% agarose gel and differences between groups were determined by comparing the number of cycles required to get a
similar electrophoretic density. The level of cellular mRNA
IL-3 partly abrogates the inhibition of CFU-C
generation by BMP-9 in LC
We further investigated whether IL-3 was indeed able to abrogate the inhibition by BMP-9 as it does with TGF-␤1.15 From
Figure 2 it can be seen that increasing IL-3 concentrations in
SF+IL-12 containing cultures did not induce a significant
increased CFU-C generation in 5-day LCs, however, clearly
abrogated part of the 100 ng BMP-9 mediated inhibition of
CFU-C production. In view of our previous observations14,15
that IL-3 partly abrogates the TGF-␤1-induced inhibition of
colony formation and CFU-C generation in the presence of
Figure 1
Dose-response of rhBMP-9 in 5-day serum-free liquid
cultures of 2 × 104 LD/FU6dBMC/ml. Input cells contained 360 day-8
CFU-C, while output was 19 250 when SF (50 ng/ml) + IL-12 (1 ng/ml)
was used as stimulus, and 46 500 when SF+IL-12+IL-3 (5 U/ml) was
included. The data represent the arithmetic mean of duplicate dishes
in a representative experiment.
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
Figure 2
Dose-response of IL-3 in 5-day serum-free liquid cultures
of 2 × 104 LD/FU6dBMC/ml. Input cells contained 370 day-8 CFU-C,
while output was 19 500 CFU-C day-7 when SF+IL-12 was used as
stimulus, and 23 000 when SF+IL-12+IL-3 (100 U/ml) was included.
BMP-9 was added at 100 ng/ml. The data represent the arithmetic
mean of quaduplicate dishes in a representative experiment.
ing cultures during the course of the culture we performed 7day cultures of 2 × 104 LD/FU6dBMC in the presence of SF+IL12+BMP-9 (100 ng) and added anti-TGF-␤1 Ab either on day
0, 2 or 4. Cultures were replated at 4 h, and on days 1, 2, 3,
4 and 7 for estimation of the CFU-C number recovered. On
a per 105 input cell basis about 2500 CFU-C were inoculated
at day 0, and 6 × 105 CFU-C were generated in 7 days when
only SF+IL-12 was present (Figure 3). The more delayed the
anti-TGF-␤1 Ab was added to the culture, the less CFU-C
were finally retrieved on subsequent days, including day 7.
However, in the absence of anti-TGF-␤1 Ab, CFU-C could
always be detected, although at lower numbers than when
started. These data show that the progenitors generating the
CFU-C were not immediately killed by BMP-9, or by the
induced TGF-␤1, and that upon neutralization of TGF-␤1
growth curves ensued in a manner roughly parallel to the
SF+IL-12 (‘none’ in Figure 3) stimulated cultures. Remarkably,
the CFU-C growth curve in the group that contained BMP-9
and anti-TGF-␤1 Ab for the whole culture period started earlier than in the SF+IL-12 group and after day 2–3 parallelled
the other growth curves at a higher level. This indicates that
SF+IL-12, the present data suggested to us that either TGF-␤1
mediates the inhibition of BMP-9, or that TGF-␤1 and BMP9 could have identical effects in that respect.
TGF-␤1 mediates the inhibition of CFU-C generation
by BMP-9 in LC
To study this possibility, we performed similar 5-day LC with
2 × 104 LD/FU6dBMC per 1 ml culture stimulated by SF+IL-12,
with or without IL-3, and added neutralizing concentrations
of antibodies (Ab) against TGF-␤1. Table 1 shows that not only
could 1 ␮g of anti-TGF-␤1 Ab prevent the inhibition by 100
ng of BMP-9, but it led to a further increased CFU-C generation both when SF+IL-12, or SF+IL-12+IL-3 were included.
Antibodies against IFN-gamma did not have any effects (data
not shown). These data strongly suggested that BMP-9 not
only induces or enhances TGF-␤1 elaboration by the cultured
BMC, but in addition may stimulate CFU-C generation driven
by SF+IL-12 with or without IL-3.
To investigate the fate of progenitor cells in BMP-9-containTable 1
Figure 3
Effect of delayed addition of anti-TGF-␤1 antibody (Ab)
(1 ␮g/ml) to liquid cultures of 2 × 104 LD/FU6dBMC/ml containing SF
(50 ng/ml) + IL-12 (1 ng/ml) + BMP-9 (100 ng/ml). BMP-9 was present
during the whole culture period, while Ab was added to the cultures
at the beginning of the culture period (Ab), or at day 2 [(2d)Ab] or
day 4 [(2d)Ab]. The data represent the arithmetic mean of quaduplicate dishes in a representative experiment.
Anti-TGF-␤1 antibodies abrogate the BMP-9 induced inhibition of CFU-C generation stimulated by SF+IL-12 with or without IL-3
Cytokines
Input
SF+IL-12
SF+IL-12
SF+IL-12
SF+IL-12
SF+IL-12+IL-3b
SF+IL-12+IL-3
SF+IL-12+IL-3
SF+IL-12+IL-3
BMP-9
(300 ng)
anti-TGF-␤1
AB
−
+
−
+
−
+
−
+
−
−
+
+
−
−
+
+
No. CFU-C-7
generated/105 nc
5 daysa
No. CFU-C-8
generated/105 nc
6 daysa
No. CFU-C-8
generated/105 nc
7 daysa
900
13 750
12.5
32 500
56 250
62 500
7 000
110 000
235 000
1225
125 000
800
210 000
462 500
−
−
−
−
2175
160 000
688
270 000
900 000
−
−
−
−
2 × 104 LD/FU6dBMC were cultured in 1 ml serum-free LC for 5, 6 or 7 days in three separate experiments. Data represent arithmetic means
of quadruplicate dishes.
a
Culture time.
b
100 U IL-3/ml.
431
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
432
BMP-9 initiates early proliferation in hemopoietic progenitor
cells in the presence of SF+IL-12, rather then stimulating CFUC expansion over the whole culture period. Our data also suggest that BMP-9 activates additional progenitors that had
otherwise remained quiescent.
The observation that BMP-9 could increase CFU-C generation by SF+IL-12 with or without IL-3 (see Table 1) in the
presence of anti-TGF-␤1 Ab also shows that the antibody did
not neutralize the BMP-9. To obtain additional evidence for
this notion we tested whether BMP-9 in the presence of the
Ab could still stimulate proliferation of BMP-9 sensitive
HepG2 cells, which it could. TGF-␤1 significantly inhibited
their growth as measured by 3H-thymidine uptake (Figure 4).
BMP-9 increases steady state TGF-␤1 mRNA levels in
LD/FU6dBMC
To confirm the ability of BMP-9 to induce TGF-␤1 in cultures
of bone marrow cells, we cultured either LD/FU6dBMC at
2 × 104 ml, or confluent FBMD-1 stromal cells, for 4 days in
dishes in the presence of SF+IL-12 and anti-TGF-␤1 Ab, with
or without 100 ng of BMP-9/ml. Total RNA was extracted on
day 4 and RT-PCR performed with intermittent sampling
between cycles. Figure 5 shows that at 25–30 PCR cycles a
visible TGF-␤1 band was present in the LD/FU6d BMC
exposed to BMP-9, whereas without BMP-9 no TGF-␤1 signal
was observed. As stromal cells are well known producers of
TGF-␤1, we tested whether BMP-9-induced TGF-␤1 mRNA
over-expression also occurred in the stromal cell line FBMD1, which it did not. Therefore, these data suggest that the BMC
themselves have BMP-9 receptors and that BMP-9 stimulation
either upregulates TGF-␤1 mRNA, or stabilizes TGF-␤1 message, which may lead to an increased elaboration of TGF-␤1.
The data also point to the possibility that the TGF-␤1mediated inhibitory action of BMP-9 may not involve the
bone marrow stroma.
Figure 4
Anti-TGF-␤1 antibody (Ab) does not abrogate the BMP9-induced proliferation of HepG2 cells. HepG2 cells (105/ml) were
plated in 100 ␮l in 96-well plates and cultured for 44 h with or without rhBMP-9, TGF-␤1 or anti-TGF-␤1 antibodies. 3H-thymidine was
included in the last 24 h of the incubation period. The data represent
the arithmetic mean of triplicate wells in two separate experiments as
a percentage of cultures without Ab or cytokine addition (none).
Figure 5
BMP-9 increases expression of TGF-␤1 mRNA in
LD/FU6dBMC, but not in FBMD-1 stromal cells. LD/FU6dBMC
4
(2 × 10 /ml) or confluent FBMD-1 stromal cells were incubated for 4
days in serum-free medium containing SF (50 ng) + IL-12 (1 ng/ml) +
anti-TGF-␤1 Ab (1 ␮g/ml) with (lanes 5–8) or without (lanes 1–4) 100
ng BMP-9/ml. Total RNA was isolated at day 4 and RT-PCR performed
on cDNA. Lanes 1–4 and lanes 5–8: 15, 20, 25 and 30 cycles, respectively.
BMP-9 is a potent stimulator of CFU-C generation
Because the stimulatory ability of BMP-9 could occur in a different dose range than its inhibitory action, a BMP-9 dose
response was investigated under conditions where the BMP9-induced TGF-␤1 production is efficiently neutralized. To
this end we performed 6-day LC of 2 × 104 LD/FU6dBMC in
the presence of SF+IL-12 and either added 4, 1 or 0.1 ␮g of
anti-TGF-␤1 Ab to these LC. In the presence of 1 or 4 ␮g
anti-TGF-␤1 Ab increasing BMP-9 concentrations led to an
increased CFU-C generation with a plateau reached at only
3–30 ng BMP-9/ml (Figure 6). At higher BMP-9 concen-
Figure 6
Dose-response of the stimulatory effect of BMP-9 on
CFU-C generation in 5-day liquid cultures of 2 × 104 LD/FU6dBMC/ml
containing anti-TGF-␤1 antibody (Ab). All cultures were stimulated
with SF (50 ng/ml) + IL-12 (1 ng/ml). Input cells contained 2025 CFUC day-14/105 nc. Data represent arithmetic means (error bars, 1 s.d.)
of quadruplicate dishes in a representative experiment. Pilot experiments where part of these factors were tested yielded similar results.
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
trations, less CFU-C were generated, probably as a result of
increasing TGF-␤1 elaboration, which could not be completely neutralized by this dose of Ab. In the presence of only
0.1 ␮g of anti-TGF-␤1 Ab the BMP-9 dose response was similar, albeit that at 3–300 ng of BMP-9 even less CFU-C were
generated. These data show that BMP-9 proper stimulates
CFU-C generation in the presence of SF+IL-12. In addition,
our findings strongly suggest that the BMP-9 evoked TGF-␤1
elaboration increases when BMP-9 concentrations are
increased.
BMP-9 synergizes with various (combinations of)
cytokines to increase progenitor generation
Next we investigated the mechanism of the BMP-9 stimulation
of CFU-C generation, ie whether BMP-9 had growth stimulating activity as a single factor, and/or displayed synergistic
activity. To this end LC of 2 × 104 LD/FU6dBMC were set up
and either an array of single growth factors or some potent
cytokine combinations were included with or without 10 ng
of BMP-9 and always in the presence of neutralizing concentrations of anti-TGF␤1 Ab (1 ␮g/ml). The results of these
experiments have been summarized in Table 2. In the presence of BMP-9 alone, no CFU-C were recovered from the LC.
Few or no CFU-C were generated in cultures containing IL11, IL-12, G-CSF, GM-CSF, SF, Flt3-L, Tpo or IL-3 as single
factors. BMP-9 synergized only with IL-3. In multicytokinecontaining cultures BMP-9 showed potent synergy with all
combinations that contained either SF+IL-3, SF+IL-11 or
SF+IL-12. BMP-9 could not increase CFU-C generation stimulated with potent cytokine combinations, ie IL-3+SF+IL-12
with or without Flt3-L+Tpo possibly because these factor con-
ditions already maximally stimulated progenitor growth.
Indeed, in all these cultures 100% of the output cells were
demonstrated to be clonogenic. Because all of these cytokines
were used at plateau concentrations, these data suggest that
the synergy of BMP-9 is different from that of SF, IL-11 or
IL-12.
BMP-9 synergizes with various (combinations of)
cytokines to increase primary colony formation
As shown in Table 2 LD/FU6dBMC show a low or absent
responsiveness to single cytokine stimulation. In order to get
a better insight of possible synergies of BMP-9 with single factors we studied primary colony formation using normal BMC
and only included LD/FU6dBMC to study particular multicytokine combinations. As expected, single factors stimulated less
colony formation then did multiple factor combinations (Table
3). In NBMC, BMP-9 did not stimulate any colony formation
as a single factor, but significantly synergized to increase colony formation by SF, GM-CSF, M-CSF, IL-3, SF+Flt3-L, SF+IL11, SF+IL-12, SF+IL-3, SF+IL-11+IL-3 or SF+IL-12+IL-3. As far
as we have determined, this was also observed in
LD/FU6dBMC. In all of these cultures BMP-9 accelerated colony development, which was particularly evident in the first
5 days of culture. No synergy of BMP-9 was seen in combination with IL-11, IL-12, G-CSF, Flt3-L, Tpo, Flt3-L+IL-12, or
SF+Epo+Tpo. Some of these observations could have been due
to the (virtual) lack of colony formation that prohibited analysis of synergistic effect, eg with IL-11, Il-12, Flt3-L and Tpo.
However, in two out of three combinations that included Flt3L (ie with either SF, IL-12 or IL-3) BMP-9 did not synergize,
suggesting that BMP-9 and Flt3-L could act via similar or over-
Table 2
Synergistic effects of BMP-9 on CFU-C generation in 6-day liquid cultures of LD/FU6dBMC in the presence of neutralizing anti-TGF␤1 antibodies
Cytokines
Mean number of CFU-C (1 sd) generated per 105 input nucleated cells
− BMP-9
+ BMP-9
No. exp = 6
input (day 0)b
none
anti-TGF-␤1 Ab
SF
Flt3-L (FL)
G-CSF
GM-CSF
IL-11
IL-12
Tpo
IL-3c
SF+IL-12
SF+IL-12+IL-3
1393 (721)
0
0
0
0
0
0
0
0
0
0
374 875 (32 673)
1 408 333 (98 181)
1393
0
0
0
0
0
0
0
0
0
750
1 001 250
1 458 333
(150)
(90 346)
(40 104)
No. exp = 2
input (day 0)b
SF+IL-3
SF+IL-11
SF+IL-12+IL-3+FL+Tpo
1888
443 750
612 500
2 025 000
1888
675 000
2 668 750
1 687 500
(53)
(141,421)
(26,517)
(53,033)
(53)
(79 549)
(70 711)
(282 843)
(721)
Fold-increase with
BMP-9a
−
−
−
−
−
−
−
−
−
⬎3.00
2.76
1.04
1.00
1.52
4.36
0.83
2 × 104 LD/FU6dBMC were cultured in 1 ml serum-free LC in the presence of cytokines and 10 ng BMP-9/ml. All cytokine-containing groups
contained antiTGF-␤1 antibodies.
a
The mean of the fold increase with BMP-9 as determined in individual experiments.
b
Input and output CFU-C cultures included 10% PWM-MSCM as a growth stimulus.
c
100 U IL-3/ml.
433
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
434
Table 3
Synergistic effects of BMP-9 on primary colony formation by normal bone marrow cells and LD/FU6dBMC in the presence of neutralizing anti-TGF-␤1 antibodies
Number of day-14 colonies per 105 input nucleated cells
Normal BMC
LD/FU6dBMC
Cytokines
− BMP-9
None
Anti-TGF-␤1 Ab (1 ␮g)
PWM-MSCM + HSb + Ab
PWM-MSCM + HS − Ab
IL-11
IL-12
SF
Flt3-L (FL)
G-CSF
GM-CSF
M-CSF
IL-3 (100 U)
Tpo
SF+Epo (1 U) + Tpoc (red)
SF+Epo (1 U) + Tpoc (white)
SF+FL
FL+IL-12
SF+IL-11
SF+IL-12
SF+IL-3
FL+IL-3
SF+IL-11+IL-3
SF+IL-12+IL-3
SF+IL-12+IL-3+FL+Tpo
0
0
420
410
2
1
46
3
52
226
124
256
2
92
78.5
91
39
60
92.5
360
366
665
390
770
+ BMP-9d
0
0
470a
ND
2
0
82a
0
57
294a
169a
293a
2
84
81.5
122a
15
95a
190a
439a
344
725
835a
780
− BMP-9
0
0
1850
ND
ND
ND
0
ND
ND
ND
ND
50
ND
ND
ND
ND
ND
1125
22.5
1350
ND
ND
850
ND
+ BMP-9d
0
0
ND
ND
ND
ND
0
ND
ND
ND
ND
80a
ND
ND
ND
ND
ND
2335a
100a
1965a
ND
ND
2100a
ND
Two to 20 000 LD/FU6dBMC or 104–105 normal BMC were cultured in 1 ml serum-free quadruplicate methylcellulose cultures. Data represent
the mean of two separate experiments. All cytokine-containing groups contained anti-TGF-␤1 antibodies.
ND, not determined
a
Significantly different from similar cultures without BMP-9 as judged by the deviation of counts in quadruplicate dishes.
b
HS, 20% horse serum.
c
Haemin was included. Red and white colonies were separately scored.
d
BMP-9 was included at 10 ng/ml.
lapping mechanisms. Similarly, BMP-9 and Tpo may show
overlapping activities. As all cytokines were used at plateau
concentrations, we conclude that the synergistic activity of
BMP-9 is different from that of SF, IL-11 or IL-12. The differential analyses of plucked colonies from cultures that included
SF+IL-11, SF+IL-12, SF+IL-3, SF+IL-11+IL-3 or SF+IL-12+IL-3
with or without BMP-9 did not reveal any effect of BMP-9 on
lineage differentiation in such colonies (data not shown).
BMP-9 acts directly on hemopoietic progenitor cells
In order to investigate if BMP acted on the colony-forming
cells proper, or via assessory cells, we sorted LD/FU6dBMC on
the basis of WGA high affinity in a scatter blast cell window
as previously described.9,16 This cell population is almost
homogenous for day-12 spleen colony-forming units in the
spleen (CFU-S-12), that we have previously shown to represent predominantly transient repopulating stem cells in
vivo.9,16 In cultures containing SF+IL-11, SF+IL-3 or SF+IL3+IL-11, single plated cells in individual wells of microtiter
plates formed 51–112% more clones in the presence of 10
ng/ml BMP-9 as without (Table 4), demonstrating that BMP-9
acted directly on these cells. BMP-9 did not stimulate cluster
or colony formation as a single agent in these cultures.
Table 4
BMP-9 has direct synergistic effects on clone formation
of single sorted WGA-binding target cells from LD/FU6dBMC in the
presence of neutralizing anti-TGF-␤1 antibodies
Cyotkines
None
SF+11
SF+3
SF+11+3
Number of wells containing a clone of 50
cells or more
− BMP-9
+ BMP-9a
0/360
49/360
64/360
91/360
0/360
104/360
97/360
164/360
Data represent the compiled data of two separate experiments.
a
BMP-9 was included at 10 ng/ml. All cultures contained anti-TGF␤1 Ab at 1 ␮g/ml.
BMP-9 receptors are expressed on hemopoietic cells
As all the previous data extensively indicated that at least
some BMC have BMP-9 receptors, we performed 125I-BMP
binding in order to quantitate the percentage of cells expressing BMP-9 receptors. Cells were centrifuged on glass slides
and incubated with 1 ng/ml 125I-BMP-9 with or without excess
cold BMP-9 and processed as detailed in Materials and
methods. Following exposure cells were stained with toluidine
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
blue. In normal BMC specific and abundant 125I-BMP-9 labelling was observed on megakaryocytes and small blast cells.
About 10% of neutrophils and immature myeloid cells also
showed 125I-BMP-9 binding which could, however, not be
competed away with cold BMP-9, strongly suggesting that
binding to these cells was nonspecific. Both in spleen and
thymus preparations lymphocytes showed weak but specific
125
I-BMP-9 binding that could be completely blocked by
excess cold BMP-9. Interestingly, in a LIN− preparation of
FU6dBMC the majority of cells, mainly consisting of small
blasts, were specifically and abundantly labelled with 125IBMP-9 (Figure 7). The small blast cells are likely to include
many CFU-C and more primitive stem cells as few lymphocytes and erythroblasts are present in the LIN⫺ preparations.
These observations indicate that BMP-9 receptors are highly
expressed on immature bone marrow cells and on megakaryocytes.
BMP-9 mRNA is expressed in the liver
Using 45 cycles of RT-PCR, BMP-9 mRNA expression was
exclusively observed in the liver, but not in spleen, bone mar-
Figure 7
Blast cells in a lineage-depleted fraction of day-6 post
5FU bone marrow show dense labelling with iodinated 125I-BMP-9
(A, B). Note that many of the smallest cells are not labelled. This binding is specific for these cells as excess cold BMP-9 abrogates labelling
(C, D). Photomicrographs show toluidine-blue-stained smears in
bright field (A, C) and dark field (B, D) illumination.
row, lymph nodes or thymus. BMP-9 mRNA could also not
be detected in a series of fetal liver-derived and bone marrowderived stromal cell lines (data not shown).
Discussion
The present data demonstrate that rhBMP-9 is both a stimulator of TGF-␤1 transcription and elaboration in bone marrow
cells and a potent synergistic activity for primary colony formation and progenitor cell generation in vitro. Because BMP9 excerts its effect in combination with saturating concentrations of SF, M-CSF, GM-CSF, IL-3, IL-11 and IL-12 its synergism is probably different from that of these cytokines. This
suggests that BMP-9 acts via a novel mechanism rather than
by induction of endogenous production of the above-mentioned cytokines or by up-regulation of their receptors. As discussed in Results, however, there is a theoretical possibility
that BMP-9 synergizes through mechanisms that overlap with
those of Tpo and/or Flt3-L. BMP-9 is a genuine synergistic factor for hemopoietic cells because it did not show activity as
a single factor. It displays characteristics of competence factors that may both release progenitors from cell cycle quiescence that would either not have been activated or had
required more time to do so.17,18 The proliferation of eurkaryotic cells is tightly regulated through a delicate balance of
positive and negative regulatory proteins that exert their
effects during the first gap phase (G1) of the cell cycle.19,20
Inhibition of cell growth by TGF-␤ occurs in mid G1, largely
through its ability to down-regulate the activity of the cyclindependent kinases cdk2 and cdk4.21 Our present demonstration, that BMP-9 is able to induce TGF-␤1 mRNA and
increase TGF-␤1 elaboration by bone marrow cells, is a novel
finding. This observation is reminiscent of the increased TGF␤1 expression in a human osteoblast cell line by BMP-222 and
in human monocytes by BMP-4.23 Because we did not find
increased TGF-␤1 mRNA levels in a series of stromal cells
stimulated with BMP-9, our data suggest that BMP-9 does not
act on bone marrow stroma. However, a more definitive conclusion requires the determination of TGF-␤1 mRNA levels in
primary stromal cells exposed to BMP-9.
We have shown that the synergistic activities of rhBMP-9
are optimal at only 3 ng, which indicates that BMP-9 is an
extremely potent cytokine, as is for example IL-12. Our data
indicate that the physiological role of BMP-9 is complex and
possibly being realised by subtle changes in its concentration.
Thus, the stimulatory activities of BMP-9 are displayed at low
concentrations whereas increasingly more TGF-␤1 is induced
by increasing levels of BMP-9. In vivo the elaborated TGF-␤1
may not be as inhibitory as in our LC in bacterial dishes that
do not inactivate TGF-␤1. It is conceivable that TGF-␤1, elaborated in the stromal microenvironment, is neutralized by a
variety of mechanisms including TGF binding proteins.24 Cells
may secrete small latent TGF-␤1 as a disulfide-bonded complex with a member of the latent TGF-␤1 binding protein
(LTBP) family. These LTBPs may act locally as autocrine or
paracrine regulators of TGF-␤1 activity in specific tissues and
play an important role for the interaction of the latent TGF␤1 complex with extracellular matrix components.25 It has to
be investigated whether BMP-9 may be less inhibitory to longterm stroma-supported cultures than in LC.
Our data show that the presence of BMP-9 during a 7-day
liquid culture period induced only moderate loss of input
CFU-C. The neutralization of TGF-␤1 on day 2 or 4 led to a
delayed onset of CFU-C production with a growth rate largely
435
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
436
similar to that of cultures wherein TGF-␤1 had been neutralized as of day 0 (see Figure 3). This suggests that the
responsive cells had been quiescent in the presence of active
TGF-␤1. In the presence of anti-TGF-␤1 Ab, BMP-9 affected
the CFU-C growth curves only in the first 2–3 days of culture,
indicating that it activated quiescent CFU-C precursor cells
to start proliferation rather than stimulating CFU-C generation
during the full culture period.
The possibility that the neutralizing Ab against TGF-␤1
would bind to BMP-9 and so abrogate the inhibitory effects
of BMP-9 can be refuted. In the presence of the Ab, BMP-9
exerted stimulatory effects on CFU-C generation and colony
formation of hemopoietic progenitors, and in addition stimulated HepG2 cells. Furthermore, if the amount of neutralizing
Ab is calculated on the basis that TGF-␤1 has a much higher
potency (100–300 pg completely inhibits CFU-C generation
and colony formation)14,15 compared to BMP-9 (full inhibition
at 100–300 ng), it would theoretically require at least 1 mg/ml
of Ab to neutralize the BMP-9 if it cross-reacts at all.
We found BMP-9 to be exclusively expressed in the adult
mouse liver, but not in fetal liver-derived stromal cell lines,
suggesting that BMP-9 may be produced by hepatocytes. In
contrast to growth factors and cytokines as TGF-␤1, PDGF,
IGF-I, IGF-II, and FGF that have effects on many other tissue
and cell types, the bone morphogenetic proteins (BMPs) have
been suggested to represent a unique set of differentiation factors that induce new bone formation at the site of implantation
instead of changing the growth rate of pre-existing bone.26
BMP-9 has been suggested to play a physiological role in
hepatocyte proliferation and possibly in osteogenic differentiation.2 The binding of 125I-BMP-9 to membrane preparations
of rat kidney, lung and heart (S Thies, personal
communication) and our present finding of BMP-9 binding to
bone marrow cells, suggests that BMP-9 may have physiological roles unrelated to bone formation. Cross-linking analysis
has indicated that human HepG2 liver tumor cells bind
recombinant human BMP-9 (rhBMP-9) with high affinity.2
HepG2 cells express two BMP-9 receptors of approximately
54 and 80 kDa, similar in size to the type I and type II receptors reported by others for TGF-␤1 and BMP-4. However,
cross-competition experiments have demonstrated that the
BMP-9 receptors on HepG2 cells do not bind other BMPs or
TGF-␤s, indicating that these are novel receptors with binding
specificity for BMP-9.2
Up to now, the BMP-family members have not been suggested to play a role in regulation of hemopoiesis, but a single
member may rather effect hemopoietic ontogeny. Thus, it has
been demonstrated recently that BMP-4 is a potent centralizing factor and inducer of hematopoietic tissue while
repressing dorsal mesoderm formation in Xenopus embryos.
BMP-4 can also stimulate GATA-2 expression in ventral mesoderm leading to increased globin production.27–29 More
characteristic of most BMP members, however, is that they
regulate morphogenetic processes, including but not limited
to, skeletal development. As an example, BMP-7 may also be
required for nephrogenesis and eye lens development30 while
BMP-6 has been suggested to be neuron-specific and to play
an important role in neuronal maturation and synapse formation.31
Finally, because of its hepatic action and ability to induce
TGF-␤1, BMP-9 is not a likely candidate cytokine to be systemically applied following cytoreductive therapy. However,
because the TGF-␤1 production by BMC can be effectively
blocked in vitro, BMP-9 seems an interesting novel activity for
ex vivo propagation and expansion of hemopoietic progenitors and possibly stem cells.
Acknowledgements
The authors are indebted to Mrs Gerry van Someren and Mrs
Irene Blokland for help with the 3H-thymidine technique and
to Mr James Calvetti for assistance with cell preparation. This
research was supported by the Genetics Institute.
References
1 Celeste AJ, Song JJ, Cox K, Rosen V, Wozney JM. Bone morphogenetic protein-9, a new member of the TGF-B superfamily. J Bone
Min Res 1994; 9 (Suppl. 1): S136.
2 Song JJ, Celeste AJ, Kong FM, Jirtle RL, Rosen V, Thies RS. Bone
morphogenetic protein-9 binds to liver cells and stimulates proliferation. Endocrinology 1995; 136: 4293–4297.
3 Lyons KM, Pelton RW, Hogan BL. Patterns of expression of murine
Vgr-1 and BMP-2a RNA suggest that transforming growth factorbeta-like genes coordinately regulate aspects of embryonic development. Genes Dev 1989; 3: 1657–1668.
4 Lyons KM, Pelton RW, Hogan BL. Organogenesis and pattern formation in the mouse: RNA distribution patterns suggest a role for
bone morphogenetic protein-2A (BMP-2A). Development 1990;
109: 833–844.
5 Vukicevic S, Helder, MN, Luyten FP. Developing human lung and
kidney are major sites for synthesis of bone morphogenetic protein-3 (osteogenin). J Histochem Cytochem 1994; 42: 869–875.
6 Heine UI, Munoz EF, Flanders KC, Ellingsworth LR, Lam HYP,
Thompson NL, Roberts AB, Sporn MB. Role of transforming
growth factor-B in the development of the mouse embryo. J Cell
Biol 1987; 105: 2861–2876.
7 Wilcox JN, Derynck R. Developmental expression of transforming
growth factors alpha and beta in mouse fetus. Mol Cell Biol 1988;
8: 3415–3422.
8 Wu X, Robinson CE, Fong HW, Gimble JM. Analysis of the native
murine bone morphogenetic protein serine threonine kinase type
I receptor (ALK-3). J Cell Physiol 1996; 168: 453–461.
9 Ploemacher RE, Van der Loo JCM, Van Beurden CAJ, Baert MRM.
Wheat germ agglutinin affinity of murine hemopoietic stem cell
subpopulations is an inverse function of their long-term repopulating ability in vitro and in vivo. Leukemia 1993; 7: 120–130.
10 Frolik CA, Wakefield LM, Smith DM, Sporn MB. Characterization
of a membrane receptor for transforming growth factor beta in
normal rat kidney fibroblasts. J Biol Chem 1984; 259: 10995–
11000.
11 Krug M, Berger SL. First strand cDNA synthesis primed with
oligo(dT). Meth Enzymol 1987; 155: 316–325.
12 Ploemacher RE, Van Soest PL, Voorwinden H, Boudewijn A.
Interleukin-12 synergizes with interleukin-3 and steel factor to
enhance recovery of murine hemopoietic stem cells in liquid culture. Leukemia 1993; 7: 1381–1388.
13 Ploemacher RE, Van Soest PL, Boudewijn A, Neben S. Interleukin12 enhances interleukin-3 dependent multilineage hematopoietic
colony formation stimulated by interleukin-11 or steel factor. Leukemia 1993; 7: 1374–1380.
14 Ploemacher RE, Van Soest PL, Boudewijn A. Autocrine transforming growth factor-␤1 blocks colony formation and progenitor
cell generation by hemopoietic stem cells stimulated with steel
factor. Stem Cells 1993; 11: 336–347.
15 Ploemacher RE, Van Soest PL, Boudewijn A. Transforming growth
factor beta (TGF-␤1) fully inhibits colony formation of murine
hemopoietic stem cells in serum-free medium stimulated by steel
factor (SF) plus IL-11/IL-12, but not plus IL-3/GM-CSF. Br J Hematol 1994; 87 (Suppl. 1): 104.
16 Van der Loo JCM, van den Bos C, Baert MRM, Wagemaker G,
Ploemacher RE. Stable multilineage hemopoietic chimerism in
alpha-thalassemic mice induced by a bone marrow subpopulation
that excludes the majority of day-12 spleen colony-forming units.
Blood 1994; 83: 1769–1776.
Direct synergistic effects of BMP-9 on progenitor cells
RE Ploemacher et al
17 Suda T, Suda J, Ogawa M. Proliferative kinetics and differentiation
of murine blast cell colonies in culture: evidence for variable G0
periods and constant doubling rates of early pluripotent hemopoietic progenitors. J Cell Physiol 1983; 117: 308–318.
18 Tsuji K, Lyman SD, Sudo T, Clark SC, Ogawa M. Enhancement
of murine hematopoiesis by synergistic interactions between steel
factor (ligand for c-kit), interleukin-11, and other early acting factors in culture. Blood 1992; 79: 2855–2860.
19 Sherr CJ. Mammalian G1 cyclins. Cell 1993; 73: 1059–1065.
20 Hunter T. Braking the cycle. Cell 1993; 75: 839–841.
21 Ewen ME, Sluss HK, Whitehouse LL, Livingston DM. TGF-beta
inhibition of Cdk4 synthesis is linked to cell cycle arrest. Cell
1993; 74: 1009–1020.
22 Zheng MH, Wood DJ, Wysocki S, Papadimitriou JM, Wang EA.
Recombinant human bone morphogenetic protein-2 enhances
expression of interleukin-6 and transforming factor-B1 genes in
normal human osteoblast-like cells. J Cell Physiol 1994; 159:
76–82.
23 Cunningham NS, Paralkar V, Reddi AH. Osteogenin and recombinant bone morphogenetic protein 2B are chemotactic for human
monocytes and stimulate transforming growth factor B1 mRNA
expression. PNAS 1992; 89: 11740–11744.
24 Gleizes PE, Beavis RC, Mazzieri R, Shen B, Rifkin DB. Identification and characterization of an eight-cysteine repeat of the latent transforming growth factor-beta binding protein-1 that
mediates bonding to the latent transforming growth factor-beta1.
J Biol Chem 1996; 271: 29891–29896.
25 Taipale J, Saharinen J, Hedman K, Keski-Oja J. Latent transforming
growth factor-beta 1 and its binding protein are components of
extracellular matrix microfibrils. J Histochem Cytochem 1996; 44:
875–889.
26 Wozney JM. The potential role of bone morphogenetic proteins
in periodontal reconstruction. J Periodont 1995; 66: 506–510.
27 Re’em-Kalma Y, Lamb T, Frank D. Competition between noggin
and bone morphogenetic protein 4 activities may regulate dorsalization during Xenopus development. PNAS 1995; 92:
12141–12145.
28 Jones CM, Dale L, Hogan BL, Wright CV, Smith JC. Bone morphogenetic protein-4 (BMP-4) acts during gastrula stages to cause ventralization of Xenopus embryos. Development 1996; 122:
1545–1554.
29 Maeno M, Mead PE, Kelley C, Xu RH, Kung HF, Suzuki A, Ueno
N, Zon LI. The role of BMP-4 and GATA-2 in the induction and
differentiation of hematopoietic mesoderm in Xenopus laevis.
Blood 1996; 88: 1965–1972.
30 Karsenty G, Luo G, Hofmann C, Bradley A. BMP 7 is required for
nephrogenesis, eye development, and skeletal patterning. Ann NY
Acad Sci 1996; 785: 98–107.
31 Tomizawa K, Matsui H, Kondo E, Miyamoto K, Tokuda M, Itano
T, Nagahata S, Akagi T, Hatase O. Developmental alteration and
neuron-specific expression of bone morphogenetic protein-6
(BMP-6) mRNA in rodent brain. Molec Brain Res 1995; 28:
122–128.
437