Full text of DOI:10.1359/jbmr.2002.17.5.782

JOURNAL OF BONE AND MINERAL RESEARCH
Volume 17, Number 5, 2002
© 2002 American Society for Bone and Mineral Research
Increased Bone Morphogenetic Protein-6 Expression in
Mouse Long Bones After Estrogen Administration
ANDREA PLANT and JONATHAN H. TOBIAS
ABSTRACT
High-dose estrogen administration is known to induce new bone formation in mouse long bones. To study the
role of regulatory proteins in this response, we examined associated changes in femoral messenger RNA
(mRNA) for candidate factors. 17␤-estradiol (E2) 0.5 mg was administered to intact female mice by weekly
injection, and Northern blot analysis was performed 1, 2, 4, 8, 12, and 16 days after the first injection. In
contrast to other factors, an increase was observed in mRNA for bone morphogenetic protein-6 (BMP-6),
which reached significance at day 8 and subsequent time-points. Estrogen-induced changes in BMP-6 protein
expression were assessed by immunocytochemistry in longitudinal femoral sections. In untreated animals,
BMP-6 was expressed by a significant proportion of growth plate chondrocytes and a subpopulation of bone
marrow cells. In contrast, osteoblasts were consistently BMP-6 negative. From as early as 4 days after starting
estrogen, clusters of slightly elongated BMP-6–positive cells were observed within the marrow cavity; the
majority were close to active bone formation surfaces. Double immunolabeling studies revealed that only
approximately 10% of BMP-6–positive bone marrow cells co-expressed the osteoblast transcription factor
Cbfa1 suggesting that they are largely distinct from the osteoblast precursor population generated concur-
rently. BMP-6–positive cells expressed neither leukocyte nor erythroid markers (CD45 and TER-119, respec-
tively), consistent with a stromal origin. We conclude that estrogen-induced osteogenesis in female mice is
associated with increased levels of BMP-6 mRNA in mouse femurs, which seems to reflect the emergence of
clusters of BMP-6 positive stromal cells adjacent to active bone formation surfaces. These findings raise the
possibility that BMP-6 serves as a paracrine mediator of estrogen’s osteogenic action in mice. (J Bone Miner
Res 2002;17:782–790)
Key words: osteoblasts, osteogenesis, RNA, immunocytochemistry, osteoblast transcription factor Cbfa1
ulates osteoblast function in postmenopausal women, par-
ticularly when administered at a relatively high dose.(5,6)
Though estrogen also stimulates osteoblasts in vitro, this
action is restricted to certain experimental conditions,(7,8)
and the physiological relevance of these observations is
unclear. Animal models also exist for studying the stimu-
latory action of estrogen on bone formation, of which the
mouse has been used most widely.(9–11) Limitations in the
latter approach are that previous studies have been restricted
to use of supraphysiological doses of estrogen, and a strik-
INTRODUCTION
T
IS well-recognized that estrogen exerts an important
protective effect on the skeleton; loss of estrogen contrib-
utes to the development of postmenopausal osteoporo-
sis.(1,2) Estrogen’s ability to prevent bone loss is thought to
involve an inhibitory action on bone resorption.(3,4) In ad-
dition, recent evidence suggests that estrogen therapy stim-
I
The authors have no conflict of interest.
Rheumatology Unit, University of Bristol Division of Medicine, Bristol, United Kingdom.
782

ESTROGEN STIMULATES BMP-6 EXPRESSION
783
ing osteogenic response is observed that may be qualita-
tively different from that seen in other mammals treated
with estrogen.(12)
Northern blot analysis
Northern blot analysis was performed on RNA extracted
from whole femurs, including bone marrow, after E2 treat-
ment. For each time-point, femurs were pooled from either
4 animals for preparations of total RNA in the case of
TGF-␤1, or up to 10 animals where polyAϩ RNA was
extracted in the case of other growth factors with less
abundant messenger RNA (mRNA). Femurs were removed
immediately after death, freed from soft tissue, snap-frozen
in liquid nitrogen, and subsequently ground into a fine
powder. Total RNA was obtained after incubation with
Trizol reagent (Life Technologies, Paisley, UK) at room
temperature for 10 minutes. Where required, polyAϩ RNA
was isolated from total RNA using the PolyATract mRNA
isolation kit (Promega, Southampton, UK). Thirty micro-
grams of total RNA or 2–5 ␮g of polyAϩ RNA were loaded
onto denaturing gels, and transcripts were separated by
electrophoresis. RNA was then transferred to a nylon mem-
brane (Perkin Elmer, Beacons Field, UK) and covalently
cross-linked to the membrane by ultraviolet (UV) irradia-
tion.
We recently found that estrogen stimulates bone forma-
tion in female mice at doses within the physiological range
through an estrogen receptor (ER)–dependent mecha-
nism.(13) We subsequently used this model to study in more
detail the mechanisms by which estrogen enhances osteo-
blast function. Inhibition of either prostaglandin or nitric
oxide synthesis significantly reduced the osteogenic re-
sponse to estrogen in female mice, suggesting these path-
ways are involved in mediating this action.(14,15) Consistent
with this conclusion, estrogen-induced osteogenesis has
also been reported to be impaired in mice with a deletion in
the gene for endothelial nitric oxide synthase.(16) Thus,
rather than stimulating osteoblasts directly, estrogen may
enhance osteoblast function by inducing the release of local
regulatory factors within the marrow cavity, which subse-
quently stimulate bone formation through a paracrine inter-
action.
Bone matrix contains abundant quantities of a wide range of
peptide growth factors that are known to induce osteoblast
Specific transcripts were detected by hybridization with
proliferation and differentiation.(17) For example, bone mor- complementary DNA (cDNA) probes labeled with
␣-[32P]deoxycytosine triphosphate (Amersham Pharmacia
phogenetic proteins (BMPs) are known to induce endochon-
Biotech, Little Chalfont, Bucks, UK) by random primer
method. The following probes were supplied by Dr. J. Lean
(St. George’s Hospital Medical School, London, UK):
mouse BMP-2,(28) mouse BMP-6,(28) mouse BMP-7,(29) rat
TGF-␤1,(30) a 0.5 kb StyI/EcoRI fragment of rat ␤-fibroblast
growth factor (bFGF),(31) rat IGF-I,(32) and rat IGF-II.(33)
cDNAs for rat platelet-derived growth factor (PDGF)-A(34)
and PDGF-B (S. P. Templeton and S. C. Sweet, GenBank,
1999) were cloned based on sequence information held in
GenBank. Blots were subsequently exposed to Phosphor
Screens (Version 4; Amersham Pharmacia Biotech), and the
resulting signal quantified using ImageQuant (Molecular
Dynamics, Little Chalfont, Bucks, UK). To correct for gel
loading, blots were stripped and reprobed with mouse
␤-actin cDNA.(35)
dral bone formation in vivo when administered as recombinant
proteins(18,19) and are thought to play important roles in skel-
etal development and fracture repair.(20,21) Previous studies
have reported that estrogen stimulates the release of osteogenic
growth factors from osteoblasts in vitro, including BMPs,(22)
insulin-like growth factor I (IGF-I),(23–25) and transforming
growth factor-␤ (TGF-␤).(25–27) However, whether these serve
as autocrine or paracrine mediators of estrogen’s stimulatory
action on bone formation in vivo has not previously been
addressed. We investigated this possibility in this study by
analyzing the time-course of gene expression for a range of
peptide growth factors after estrogen administration in female
mice, with the initial aim of identifying factor(s) up-regulated
during this response.
Immunolocalization studies
MATERIALS AND METHODS
Based on results of Northern blot analysis, we examined
the effect of estrogen on the expression of BMP-2 and
BMP-6 by immunocytochemistry. Animals were treated
with E2, as above, in a separate experiment, and the femurs
were removed, freed from soft tissue, paraformaldehyde-
fixed, and decalcified in EDTA. The diaphysis and metaph-
ysis were separated, and the latter was paraffin-embedded
after dehydration through graded alcohols. Longitudinal
sections of the distal femoral metaphysis (6 ␮m) were
subsequently cut on a Reichert-Jung 2050 microtome and
air-dried on Polysine slides (BDH, Poole, Dorset, UK).
Sections were dewaxed in Xylene (BDH), rehydrated, and
incubated in 10% rabbit serum (NRS; Vector Laboratories,
Peterborough, UK) to block nonspecific binding of second-
ary antibody.
Animals
Eight-week-old CBA-1 intact female mice from the Uni-
versity of Bristol (Bristol, UK) breeding colony were allo-
cated to weight-matched treatment groups and were admin-
istered with 17␤-estradiol (E2) in corn oil (Sigma, Poole,
Dorset, UK) at 0.5 mg/animal per week by subcutaneous
(sc) injection. This dose was used to induce a maximal
osteogenic response based on results of our previous dose-
response study.(13) Animals were killed by cervical disloca-
tion before or 1, 2, 4, 8, 12, or 16 days after commencement
of treatment. Throughout, mice received a standard diet (rat
and mouse standard diet; B&K Ltd., Humberside, UK) and
water ad libitum, and they were kept on a 12-h light/dark
cycle. All procedures complied with the guiding principles
in Care and Use of Laboratory Animals.
Slides were incubated overnight with goat anti-human
BMP-2 (clone N-14) or BMP-6 (clone N-19; Autogen Bio-

784
PLANT AND TOBIAS
clear, Calne, Wiltshire, UK), washed, and incubated with
fluorescein isothiocyanate (FITC) conjugated rabbit anti-
goat immunoglobulin G (IgG; Sigma) as a secondary anti-
body. Slides were subsequently washed, mounted in
Vectashield mountant (Vector Laboratories), and analyzed
under fluorescent microscopy (Leica) equipped with an I3
filter, using a Neotech image grabber. Control sections were
included where the primary antibody was omitted or after
preincubation with 5-fold excess of human recombinant
BMP-6 (a gift from Genetic Institute, Cambridge, MA,
USA).
To study whether BMP-6 positive bone marrow cells
belong to the osteogenic lineage, we analyzed whether they
co-express Cbfa1. This approach was based on evidence
that Cbfa1 is an essential transcription factor for osteoblast
differentiation,(36,37) and our previous finding that estrogen-
induced osteogenesis in mice is associated with the appear-
ance of large numbers of Cbfa1-expressing bone marrow
cells assumed to represent early osteoblast precursors.(38)
Cbfa1 protein expression was studied using a histochemical
approach to enhance detection of Cbfa1 immunoreactivity.
Sections were processed for BMP-6 immunodetection as
described above except that an alkaline phosphatase (ALP)-
conjugated donkey anti-goat secondary antibody (Jackson
ImmunoResearch, Luton, UK) was used. BMP-6 immuno-
reactivity was detected by incubation in Vector Red sub-
strate with 1.2 mM levamisole (Sigma) to block endogenous
ALP activity. Cbfa1 immunoreactivity was subsequently
detected by incubation with rabbit anti-mouse Cbfa1 pri-
mary antibody (a gift from Dr. G. Karsenty, Baylor College
of Medicine, Houston, TX, USA),(39) followed by biotinyl-
ated donkey anti-rabbit secondary antibody (Jackson Immu-
noResearch). Specific immunoreactivity was amplified us-
ing a Tyramide Signal Amplification kit (Perkin Elmer) and
detected with dimethylaminobenzaldehyde (DAB) substrate
and nickel chloride (Vector Laboratories). Sections were
subsequently mounted in DPX (BDH), and the presence of
BMP-6 and Cbfa1 immunoreactivity identified under light
microscopy by the presence of red and black precipitates
respectively.
FIG. 1. Effect of treatment with 17␤-estradiol (E2), 0.5 mg/animal
per week for varying durations, on BMP-6 expression of mouse femurs.
Top: picture of phosphor imager screen used for mRNA quantification
in a representative experiment. Bottom: mean Ϯ SEM (closed symbols)
after combining results from three independent experiments in which
RNA was extracted from femurs pooled from 10 (2 experiments) or 4
(1 experiment) mice per time-point. In each experiment, results were
expressed as the BMP-6/actin ratio divided by the time 0 value. Open
symbols represent data from a control group of animals killed after
vehicle administration during the first experiment (four animals per
time-point). *p Ͻ 0.05 versus day 0 by one-way ANOVA.
before starting E2. Whether E2 treatment affected mRNA
levels was assessed by one-way analysis of variance
(ANOVA) and Fisher’s test of least-significant difference
performed where p Ͻ 0.05 (Statview 4.0; SAS Institute,
Cary, NC, USA).
To further characterize BMP-6–expressing cells, double
immunolabeling studies were performed using rat anti-
mouse CD45 (leukocytes, clone 30-F11(40)) and TER-119
(erythroid cells(41); Pharmingen, San Diego, CA, USA).
Tissue sections were processed as above, except both pri-
mary antibodies were added simultaneously overnight fol-
lowed by simultaneous incubation with rabbit anti-goat
IgG-FITC and Cy3 conjugated rabbit anti-rat IgG (Jackson
RESULTS
E2 treatment significantly increased BMP-6 mRNA from
ImmunoResearch). Whether FITC and Cy3 showed evi- day 8 onwards, assessed by combining results from three
dence of colocalization was analyzed by fluorescent micros- independent experiments (Fig. 1). No equivalent increase was
copy using a dual red/green filter. Cy3 alone was detected observed in control animals in which BMP-6 expression was
using an N2 filter for red fluorescence. Control sections analyzed at different time-points after vehicle treatment. Pre-
were included where one or both primary antibodies were liminary findings based on analysis of one experiment sug-
omitted.
gested that other growth factors do not show a similar increase
(Table 1). Indeed, the majority of these other factors, including
TGF-␤1, BMP-2, and IGF-1, showed trends suggestive of
reduced expression after E2 administration.
Statistical analysis
Results for Northern blot analysis were corrected for
To study estrogen-induced changes in expression of
␤-actin expression and expressed as a ratio to the value BMP-2 and BMP-6 at the protein level, immunocytochem-
obtained from animals in the same experiment immediately istry was subsequently performed on longitudinal femoral

ESTROGEN STIMULATES BMP-6 EXPRESSION
785
TABLE 1. EFFECT OF TREATMENT WITH 17〉-ESTRADIOL (E2) 0.5 mg/ANIMAL PER WEEK FOR VARYING DURATIONS ON
GROWTH FACTOR EXPRESSION OF MOUSE FEMURS
17␤-Estradiol treatment
Growth factor
Day 1
Day 2
Day 4
Day 8
Day 12
Day 16
TGF-␤1 (2.5 kb)
BMP-2 (3.8 kb)
PDGF-A (2.0 kb)
PDGF-A (6.3 kb)
PDGF-B (3.5 kb)
IGF-I (7.0 kb)
0.89
1.11
1.13
1.27
0.96
0.50
0.78
1.07
0.50
0.63
0.79
0.67
0.93
0.70
0.85
0.43
0.61
0.61
0.78
1.04
1.25
0.75
0.90
0.42
0.51
0.59
1.49
0.79
1.0
0.89
0.77
0.49
1.05
0.83
0.95
0.67
0.96
1.12
0.65
1.00
1.08
0.64
0.85
0.68
1.02
1.03
0.66
0.85
IGF-I (1.8 kb)
IGF-I (1.0 kb)
Results show the growth factor/actin ratio divided by the time 0 value for RNA samples pooled from femurs of 10 animals per
time-point.
growth plate, presumably related to estrogen’s inhibitory ac-
tion on longitudinal bone growth (Fig. 3D).
Interestingly, E2 led to the appearance of clusters of
BMP-6–positive bone marrow cells within the metaphysis,
which were first identified in sections obtained 4 days after
starting E2 (Fig. 4). By day 8, these clusters included cells
with a slightly elongated morphology and were generally
close to active bone formation surfaces (Figs. 3E–3G). At
day 16, the marrow cavity within the metaphysis had largely
been replaced by new cancellous trabeculae, and relatively
few BMP-6–positive bone marrow cells were present (re-
sults not shown). In contrast, the marrow cavity of the
diaphysis, which had not yet been replaced by new bone,
showed large numbers of BMP-6–positive cells throughout
(Fig. 3H).
We examined the relationship of BMP-6–expressing
FIG. 2. Longitudinal femoral sections incubated with BMP-2 anti-
body and subsequently visualized by indirect immunofluorescence.
Sections from untreated mice show (A) BMP-2 expression by growth
plate chondrocytes (ϫ200) and (B) lack of BMP-2 expression within
bone marrow (ϫ200). (C) Section obtained 8 days after starting E2 0.5
mg/week, in which BMP-2 expression within bone marrow is still
absent (ϫ200). (D) Lack of immunostaining after omission of BMP-2
antibody (ϫ200).
cells to the osteogenic lineage by analyzing whether they
co-express the osteoblast transcription factor, Cbfa1. Sev-
eral growth plate chondrocytes were identified which co-
expressed BMP-6 and Cbfa1 (Fig. 4A). In untreated ani-
mals, a small number of bone marrow cells were identified
that expressed Cbfa1; the occasional cell co-expressed
BMP-6 (Fig. 4B). Within 4 days of E2 treatment, clusters of
BMP-6–positive bone marrow cells could be identified
(Figs. 4C and 4D). Large numbers of Cbfa1-positive bone
marrow cells were also present at this time, distributed
relatively evenly throughout the marrow cavity, as previ-
ously observed.(38) A small number of bone marrow cells
co-expressed BMP-6 and Cbfa1, representing ϳ10% of the
BMP-6–positive bone marrow population. Sections in
which one of the primary antibodies was omitted showed
absent immunostaining as appropriate (Figs. 4E and 4F).
Further double immunolabeling studies were performed
to examine the relationship of BMP-6–expressing bone
marrow cells to other lineages. BMP-6–positive bone mar-
row cells expressed neither CD45 nor TER-119, suggesting
that they are not of leukocyte or erythroid origin (Figs. 5A
and 5B). Sections incubated with antibodies to CD45 or
TER-119 alone showed the expected tissue distribution for
sections. BMP-2 expression was restricted to growth plate
chondrocytes in untreated animals (Figs. 2A and 2B), which
remained the case after administration of E2 (Fig. 2C).
Controls in which the primary antibody was omitted showed
no immunostaining (Fig. 2D).
A significant proportion of growth plate chondrocytes were
also found to express BMP-6 (Fig. 3A). Preincubation of
BMP-6 antibody with recombinant human BMP-6 led to a
decrease in the intensity of the BMP-6 signal, confirming the
specificity of BMP-6 immunoreactivity (Fig. 3B). No immu-
noreactivity was observed in sections where the primary anti-
body was omitted (Fig. 2D). A small number of bone marrow
cells were also observed to express BMP-6 (Fig. 3C). In
contrast, osteoblasts consistently failed to express BMP-6. E2
treatment seemed to reduce BMP-6 expression within the these lineages (Figs. 5C and 5D).

786
PLANT AND TOBIAS
FIG. 3. Longitudinal femoral sections incu-
bated with BMP-6 antibody and subsequently
visualized by indirect immunofluorescence. Sec-
tions from untreated mice show (A) BMP-6 ex-
pression by growth plate chondrocytes (ϫ100),
(B) reduced immunostaining in the growth plate
of untreated animals after preincubation with
recombinant human BMP-6 (ϫ100), and (C)
BMP-6–positive cells within bone marrow
(ϫ200). Sections obtained 8 days after starting
E2 show (D) reduced BMP-6 immunostaining of
the growth plate (ϫ100), (E) clusters of BMP-
6–positive cells adjacent to active bone for-
mation surfaces (arrow; ϫ200), (F) BMP-6–
expressing cells adjacent to cancellous bone for-
mation surface under higher power (ϫ400), and
(G) BMP-6–positive cells adjacent to endosteal
formation surface (ϫ400). (H) Sections obtained
at day 16 show large numbers of BMP-6–
expressing bone marrow cells throughout the
diaphysis (ϫ200).
The increase in BMP-6 expression after the addition of
estrogen preceded the osteogenic response that was assessed
in our previous time-course studies, performed under iden-
tical experimental conditions.(11,42) For example, type I col-
lagen expression and the extent of mineralizing surfaces,
which were assessed by Northern blot analysis and histo-
morphometry, respectively, were not found to increase until
12 days after starting estrogen. Because BMP-6 expression
seemed to increase during the early phase of the osteogenic
response to estrogen and the majority of clusters of BMP-
6–positive bone marrow cells were in close proximity to
DISCUSSION
We investigated the mechanisms that mediate new bone
formation after estrogen administration in female mice by
analyzing whether this response is associated with increased
expression of candidate growth factors, which were as-
sessed by quantification of femoral mRNA. A significant
rise in BMP-6 mRNA was observed on day 8 and beyond,
whereas in contrast, no increase was observed in other
factors analyzed. Subsequent immunocytochemistry studies
revealed that estrogen treatment led to the appearance of
clusters of BMP-6–positive cells within the marrow cavity sites of active bone formation, our findings raise the possi-
as early as 4 days after starting estrogen. In view of their bility that BMP-6 serves as a paracrine mediator of estrogen-
similar time-courses, we assume that expansion in the BMP- induced bone formation. This suggestion is consistent with
6–positive bone marrow population after estrogen was re- previous evidence that BMP-6 is a potent inducer of osteoblast
sponsible for the rise in femoral BMP-6 mRNA.
differentiation both in vitro and in vivo.(43–47)

ESTROGEN STIMULATES BMP-6 EXPRESSION
787
FIG. 4. Longitudinal femoral
sections incubated with antibodies
to BMP-6 and Cbfa1 and subse-
quently viewed by light micros-
copy. (A) In growth plate sections,
the majority of chondrocytes
showed high levels of Cbfa1 ex-
pression, whereas several chon-
drocytes were identified that co-
expressed BMP-6 (arrowhead;
ϫ200). (B) Section from an un-
treated mouse showing BMP-6–
expressing bone marrow cells
(open arrow), a small number of
bone marrow cells that expressed
Cbfa1 (filled arrow), and the oc-
casional double-positive cell (ar-
rowhead; ϫ200). Sections ob-
tained 4 days after starting E2 0.5
mg/week, under (C) lower power
(ϫ200) and (D) higher power
(ϫ400), showing clusters of
BMP-6–positive cells within
bone marrow and a marked in-
crease in the number of Cbfa1-
positive bone marrow cells;
ϳ10% of BMP-6–positive cells
co-expressed Cbfa1. Control sec-
tions were obtained 4 days after
starting E2 0.5 mg/week: (E) pri-
mary antibody for Cbfa1 omitted,
showing BMP-6–positive cells
within bone marrow (ϫ100) and
(F) primary antibody for BMP-6
omitted, showing expression of
Cbfa1 by bone marrow cells and
osteoblasts (ϫ100).
The exact nature of BMP-6–positive bone marrow cells
identified in this study is unclear. Although growth plate
chondrocytes have previously been shown to express
BMP-6,(47) we are not aware of any previous reports of
BMP-6 expression in bone marrow cells. Because stromal
cells regulate other bone marrow lineages and secrete a
wide range of growth factors,(48) BMP-6–positive bone
marrow cells may represent a stromal population involved
in stimulation of osteoblast formation through a paracrine
interaction. Preliminary analysis was consistent with this
possibility because BMP-6–positive cells lacked markers
for other marrow populations such as erythroid or leukocyte
lineages. However, definitive identification of stromal cells
is difficult due to the lack of specific markers.
In view of previous findings that cells of the osteogenic
lineage secrete BMP-6,(22) BMP-6–positive bone marrow
cells may alternatively represent early osteoblast precursors
in which BMP-6 acts to stimulate further differentiation
through an autocrine mechanism. Consistent with this pos-
sibility, we previously found that estrogen induces a marked
increase in the population of early osteoblast precursors
FIG. 5. Longitudinal femoral sections from untreated mice incubated
with antibodies to (A) BMP-6 and CD45 or (B) TER-119 and subse-
quently viewed by indirect immunofluorescence (ϫ200). BMP-6–
positive bone marrow cells expressed neither CD45 nor TER-119. (C)
within bone marrow of mouse femurs, as assessed by an ex CD45 antibody alone (ϫ100). (D) TER-119 antibody alone (ϫ100).

788
PLANT AND TOBIAS
vivo CFUf assay and number of Cbfa1-positive bone mar- is suggested to play a role in estrogen-induced osteogenesis
row cells.(49,50) However, relatively few BMP-6–positive in mice by stimulating early osteoblast precursors, in other
bone marrow cells seemed to co-express Cbfa1, suggesting species, estrogen may act predominantly to prolong osteo-
that BMP-6–positive cells are largely distinct from the blast survival.(56,57) Previous clinical studies have reported
Cbfa1-positive osteoblast precursor population. Because that cells of the osteoblast lineage express estrogen recep-
Cbfa1 expression is restricted to the nucleus, the number of tors,(58,59) and that estrogen administration leads to an in-
Cbfa1-positive cells may have been underestimated, be- crease in serum levels of IGF-I.(60) However, to our knowl-
cause the nucleus may not have been included in the plane edge, no previous study has analyzed possible influences of
of section for a significant number of bone marrow cells. estrogen on expression of BMP-6 or other growth factors
However, BMP-6– and Cbfa1-positive cells also showed within the human skeleton. Our findings may also reflect a
differing distributions within bone marrow, suggesting that wider role of BMP-6 in bone formation. For example, the
these represent essentially separate populations. For exam- permissive effect of glucocorticoids on osteoblast differen-
ple, BMP-6–positive cells were initially seen to form clus- tiation has been found to be mediated by BMP-6 specifi-
ters of relatively elongated cells, which was not the case for cally,(45) and BMP-6 expression has been suggested to play
cells which expressed Cbfa1. Furthermore, large numbers of an important osteoinductive role in osteoblastic metastases
BMP-6–positive cells continued to be expressed throughout associated with prostate adenocarcinoma.(61)
the marrow cavity as assessed up until 16 days after starting
Although this study focused on the role of BMP-6 in
estrogen, whereas at these later time-points, Cbfa1-positive mediating estrogen-induced bone formation, other paracrine
bone marrow cells are generally located adjacent to bone mediators are likely to be involved in this response. For
forming surfaces.(38)
example, we recently reported that estrogen-induced bone
Although our results suggest that BMP-6– and Cbfa1- formation is partially suppressed by inhibition of prosta-
positive cells represent distinct populations, it is likely that glandin and nitric oxide synthesis.(14,15) Though growth
important functional relationships exist between these. For factors such as TGF-␤1 and IGF-1 were also predicted to
example, previous findings indicate that considerable plas- play a role in this response based on results of studies with
ticity exists in mesenchymal-derived lineages,(51) and Cbfa1 isolated osteoblasts,(23–27) our preliminary findings sug-
expression is necessary but not sufficient for osteoblast gested that if anything, expression of these factors tended to
differentiation.(52) Hence, paracrine growth factors such as fall after estrogen administration. Interestingly, administra-
BMP-6 may contribute to bone formation by stimulating tion of diethylstilbestrol (DES) has also been reported to
and maintaining osteogenic commitment of Cbfa1-positive suppress IGF-I expression in rat bones.(62) However, rela-
early osteoblast precursors. This possibility is supported by tively few growth factors were examined in the present
previous evidence that BMP-6 is distinct from other BMPs investigation, and whether estrogen selectively up-regulates
because it acts principally at the level of early osteoprogeni- expression of BMP-6 as opposed to other regulatory factors
tor cells.(44)
is currently unclear. It would also be informative to deter-
To our knowledge, this is the first report of increased mine whether BMP-6 is solely responsible for the increased
BMP-6 expression in vivo in response to administration of bone formation after estrogen administration by examining
estrogen or any other stimulus of bone formation. Estrogen whether this response is impaired in BMP-6 knockout
has previously been found to induce BMP-6, but not other mice.(63)
BMPs, in fetal human osteoblast cell lines.(22) Although this
A potential limitation of this study is that growth factor
observation implies that estrogen can directly induce mRNA expression was assessed at the level of whole bone.
BMP-6 expression by osteoblasts, the latter were not found This approach was successful in detecting changes in ex-
to express BMP-6 in this study. Furthermore, our results pression of BMP-6 that were subsequently confirmed by
suggest that estrogen is more likely to have increased ex- immunocytochemistry. However, the lack of any overall
pression of BMP-6 in mouse femurs by expanding the increase in expression of other growth factors during
BMP-6–positive population than by direct activation of estrogen-induced osteogenesis may have reflected the fact
gene transcription. For example, a relatively close correla- that several compartments were being analyzed concur-
tion was found between the increase in BMP-6 mRNA rently and estrogen may have exerted opposing effects. For
content after estrogen administration and the appearance of example, TGF-␤1 is expressed by hemopoietic cells,(64)
clusters of BMP-6–positive cells. In addition, any direct which undergo a rapid decline after high-dose estrogen
tendency of estrogen to induce BMP-6 gene transcription administration in female mice.(49) Similarly, any tendency
may be expected to have caused a more rapid increase in for BMP-2 mRNA to fall after E2 treatment may have
mRNA levels than observed here.
reflected estrogen’s suppression of longitudinal bone
Estrogen is also thought to stimulate osteoblast function growth, because estrogen seemed to reduce chondrocyte
in humans(5,6) and other mammals.(53–55) In view of our expression of BMP-2 to a similar extent observed for
findings, it is tempting to speculate that BMP-6 is involved BMP-6 (results not shown).
in mediating stimulatory effects of estrogen on osteoblast
In conclusion, we found that estrogen-induced osteogen-
function in other species. Although this study used a phar- esis in female mice is associated with increased femoral
macological dose of estrogen to induce a maximal response, content of BMP-6 mRNA. Further analysis revealed that
it is likely that lower doses of estrogen, which we have this change reflects the appearance of clusters of BMP-6
previously found to stimulate bone formation in mice,(13) expressing bone marrow cells adjacent to active bone for-
act through a similar mechanism. However, whereas BMP-6 mation surfaces. These findings raise the possibility that

ESTROGEN STIMULATES BMP-6 EXPRESSION
789
and anabolic response to exogenous estrogen in mice with
targeted disruption of endothelial nitric oxide. Endocrinology
142:760–766.
BMP-6 expression seems to be indirectly increased in re-
sponse to estrogen administration. Further studies are re-
quired to determine whether BMP-6 may perhaps mediate
the action of estrogen on the skeleton in other species
including humans.
17. Baylink DJ, Finkelman RD, Mohan S 1993 Growth factors to
stimulate bone formation. J Bone Miner Res 8:S2;S565–S572.
18. Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ,
Kriz RW, Hewick RM, Wang EA 1988 Novel regulators of
bone formation: Molecular clones and activities. Science 242:
1528–1534.
ACKNOWLEDGMENTS
19. Sampath TK, Rashka KE, Doctor JS, Tucker RF, Hoffmann
FM 1993 Drosophila transforming growth factor beta super-
family proteins induce endochondral bone formation in mam-
mals. Proc Natl Acad Sci USA 90:6004–6008.
We are grateful to Dr. S. Colley for his help in cloning the
PDGF-A and PDGF-B cDNA probes and to Ms. R. Gibson
for her assistance in processing femurs.
20. Hogan BL 1996 Bone morphogenetic proteins in development.
Curr Opin Genet Dev 6:432–438.
21. Nakase T, Nomura S, Yoshikawa H, Hashimoto J, Hirota S,
Kitamura Y, Oikawa S, Ono K, Takaoka K 1994 Transient and
localized expression of bone morphogenetic protein 4 messen-
ger RNA during fracture healing. J Bone Miner Res 9:651–
659.
22. Rickard DJ, Hofbauer LC, Bonde SK, Gori F, Spelsberg TC,
Riggs BL 1998 Bone morphogenetic protein-6 production in
human osteoblast cell lines. J Clin Invest 101:413–422.
23. Ernst M, Heath JK, Rodan GA 1989 Estradiol effects on
proliferation, messenger ribonucleic acid for collagen and
insulin-like growth factor-I, and parathyroid hormone-
stimulated adenylate cyclase activity in osteoblastic cells from
calvariae and long bones. Endocrinology 125:825–833.
24. Bankson DD, Rifai N, Williams ME, Silverman LM, Gray TK
1989 Biochemical effects of 17␤-estradiol on UMR 106 cells.
Bone Miner 6:55–63.
25. Slater M, Patava J, Kingham K, Bason RS 1994 Modulation of
growth factor incorporation into ECM of human osteoblast-
like cells in vitro by 17beta-estradiol. Am J Physiol 267:
E990–E1001.
26. Komm BS, Terpening CM, Benz DJ, Graeme KA, Gallegos A,
Korc M, Greene GL, O’Malley BW, Haussler MR 1988 Es-
trogenic binding, receptor mRNA and biologic response in
osteoblast-like osteosarcoma cells. Science 241:81–83.
27. Oursler MJ, Cortese C, Keeting P, Anderson MA, Bonde SK,
Riggs BL, Spelsberg TC 1991 Modulation of transforming
growth factor-␤ production in normal human osteoblast-like
cells by 17␤-estradiol and parathyroid hormone. Endocrinol-
ogy 129:3313–3320.
28. Pereira RC, Rydziel S, Canalis E 2000 Bone morphogenetic
protein-4 regulates its own expression in cultured osteoblasts.
J Cell Physiol 182:239–246.
29. Ozkaynak E, Rueger DC, Drier EA, Corbett C, Ridge RJ,
Sampath TK, Oppermann H 1990 OP-1 cDNA encodes an
osteogenic protein in the TGF-beta family. EMBO J 9:2085–
2093.
30. Qian SW, Kondiah P, Roberts AB, Sporn MB 1990 cDNA
cloning and PCR of rat transforming growth factor beta-1.
Nucleic Acids Res 18:3059.
31. Kurokawa T, Seno M, Igarashi K 1988 Nucleotide sequence of
rat basic fibroblast growth factor cDNA. Nucleic Acids Res
16:5201.
REFERENCES
1. Richelson LS, Heinz HW, Melton LJ III, Riggs BL 1984
Relative contributions of aging and estrogen deficiency to
postmenopausal bone loss. N Engl J Med 311:1273–1275.
2. Lindsay R, Hart DM, Forrest C, Baird C 1980 Prevention of
spinal osteoporosis in oophorectomized women. Lancet
2:1151–1153.
3. Christiansen C, Christiansen MS, Larsen NE, Transbol I 1982
Pathophysiological mechanism of estrogen effect on bone me-
tabolism: Dose-response relationship in early postmenopausal
women. J Clin Endocrinol Metab 55:1124–1130.
4. Wronski TJ, Cintron M, Doherty AL, Dann LM 1988 Estrogen
treatment prevents osteopenia and depresses bone turnover in
ovariectomized rats. Endocrinology 123:681–686.
5. Vedi S, Purdie DW, Ballard P, Bord S, Cooper AC, Compston
JE 1999 Bone remodeling and structure in postmenopausal
women treated with long-term, high-dose estrogen therapy.
Osteoporos Int 28:52–58.
6. Khastgir G, Studd J, Holland N, Alaghband-Zadeh J, Fox S,
Chow J 2001 Anabolic effect of estrogen replacement on bone
in postmenopausal women with osteoporosis: Histomorpho-
metric evidence in a longitudinal study. J Clin Endocrinol
Metab 86:289–295.
7. Qu Q, Peral-Heape M, Kapanen A, Dahhllund J, Salo J,
Vaananen HK, Harkonen P 1998 Estrogen enhances differen-
tiation of osteoblasts in mouse bone marrow culture. Bone
22:201–209.
8. Robinson JA, Harris SA, Riggs BL, Spelsberg TC 1997 Es-
trogen regulation of human osteoblast cell proliferation and
differentiation. Endocrinology 138:2919–2927.
9. Urist MR, Budy AM, McLean FC 1950 Endosteal bone for-
mation in estrogen-treated mice. J Bone Joint Surg Am 32:
143–162.
10. Bain SD, Bailey MC, Celino DL, Lantry MM, Edwards MW
1993 High-dose estrogen inhibits bone resorption and stimu-
lates bone formation in the ovariectomised mouse. J Bone
Miner Res 8:435–442.
11. Samuels A, Perry MJ, Tobias JH 1999 High-dose estrogen
induces de novo medullary bone formation in female mice.
J Bone Miner Res 14:178–186.
12. Turner RT 1999 Mice, estrogen and postmenopausal osteopo-
rosis. J Bone Miner Res 14:187–191.
32. Murphy LJ, Bell GI, Duckworth ML, Friesen HG 1987 Iden-
tification, characterisation and regulation of a rat complemen-
tary deoxyribonucleic acid which encodes insulin-like growth
factor-I. Endocrinology 121:684–691.
33. Whitfield HJ, Bruni CB, Frunzio R, Terrell JE, Nissley SP,
Rechler MM 1984 Isolation of a cDNA clone encoding rat
insulin-like growth factor-II precursor. Nature 312:277–280.
34. Mercola M, Wang CY, Kelly J, Brownlee C, Jackson-Grusby
L, Stiles C, Bowen-Pope D 1990 Selective expression of
PDGF A and its receptor during early mouse embryogenesis.
Dev Biol 138:114–122.
13. Samuels A, Perry MJ, Goodship AE, Fraser WD, Tobias JH
2000 Is high-dose estrogen-induced osteogenesis in the mouse
mediated by an estrogen receptor? Bone 27:41–46.
14. Samuels A, Perry MJ, Tobias JH 1999 High-dose oestrogen
induced osteogenesis in female mice is partially suppressed by
indomethacin. Bone 25:675–680.
15. Samuels A, Perry MJ, Gibson RL, Colley S, Tobias JH 2001
Role of endothelial nitric oxide synthase in estrogen-induced
osteogenesis. Bone 29:24–29.
16. Armour KE, Armour KJ, Gallagher ME, Godecke A, Helfrich
MH, Reid DM, Ralston SH 2001 Defective bone formation

790
PLANT AND TOBIAS
35. Leader DP, Gall I, Campbell P, Frischauf AM 1986 Isolation 50. Samuels A, Perry MJ, Gibson RG, Tobias JH 1999 Cbfa1
and characterization of cDNA clones from mouse skeletal
muscle actin mRNA. DNA 5:235–238.
expression is induced in the metaphysis of female mice by
high-dose estrogen. J Bone Miner Res 14:1032. (abstract).
36. Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K, 51. Nuttall ME, Patton AJ, Olivera DL, Nadeau DP, Gowen M
Deguchi K, Shimizu Y, Bronson RT, Gao Y-H, Inada M, Sato
M, Okamoto R, Kitamura Y, Yoshiki S, Kishimoto T 1997
Targeted disruption of Cbfa1 results in a complete lack of bone
1998 Human trabecular bone cells are able to express both
osteoblastic and adipocytic phenotype: Implications for os-
teopenic disorders. J Bone Miner Res 13:371–382.
formation owing to maturational arrest of osteoblasts. Cell 52. Satomura K, Krebsbach P, Bianco P, Robey PG 2000 Osteo-
89:755–764.
genic imprinting upstream of marrow stromal cell differentia-
tion. J Cell Biochem 78:391–403.
37. Otto F, Thornell AP, Crompton T, Denzel A, Gilmour KC,
Rosewell IR, Stamp GWH, Beddington RSP, Mundlos S, 53. Whitson SW 1972 Estrogen-induced osteoid formation in the
Olsen BR, Selby PB, Owen MJ 1997 Cbfa1, a candidate gene osteon of mature female rabbits. Anat Rec 173:417–436.
for cleidocranial dysplasia syndrome, is essential for osteoblast 54. Takano-Yamamoto T, Rodan GA 1990 Direct effects of 17␤-
differentiation and bone development. Cell 89:765–771.
38. Plant A, Samuels A, Perry MJ, Colley S, Gibson R, Tobias JH
estradiol on trabecular bone in ovariectomised rats. Proc Natl
Acad Sci USA 87:2172–2176.
2001 Estrogen-induced osteogenesis in mice is associated with 55. Chow JWM, Lean JM, Chambers TJ 1992 17␤-estradiol stim-
the appearance of Cbfa1-expressing bone marrow cells. J Cell
ulates cancellous bone formation in female rats. Endocrinol-
Biochem 84:285–294.
ogy 130:3025–3032.
39. Xiao G, Wang D, Benson MD, Karsenty G, Franceschi RT 56. Tomkinson A, Reeve J, Shaw RW, Noble BS 1997 The death
1998 Role of the alpha2-integrin in osteoblast-specific gene
expression and activation of the Osf2 transcription factor.
J Biol Chem 273:32988–32994.
of osteocytes via apoptosis accompanies estrogen withdrawal
in human bone. J Clin Endocrinol Metab 82:3128–3135.
57. Tomkinson A, Gevers EF, Wit JM, Reeve J, Noble BS 1998
The role of estrogen in the control of rat osteocyte apoptosis.
J Bone Miner Res 13:1243–1250.
58. Braidman IP, Davenport LK, Carter DH, Selby PL, Mawer
EB, Freemont AJ 1995 Preliminary in situ identification of
estrogen target cells in bone. J Bone Miner Res 10:74–80.
40. Ledbetter J, Herzenberg A 1979 Xenogeneic monoclonal an-
tibodies to mouse lymphoid differentiation antigens. Immunol
Rev 47:63–90.
41. Ikuta K, Kina T, MacNeil I, Uchida N, Peault B, Chien Y-H,
Weissman IL 1990 A developmental switch in thymic lym-
phocyte maturation potential occurs at the level of haemato- 59. Braidman IP, Hainey L, Batra G, Selby PL, Saunders PTK,
poietic stem cells. Cell 62:863–874.
42. Plant A, Tobias JH 2001 Characterisation of the temporal
sequence of osteoblast gene expression during estrogen-
Hoyland JA 2001 Localisation of estrogen receptor beta pro-
tein expression in adult human bone. J Bone Miner Res 16:
214–220.
induced osteogenesis in female mice. J Cell Biochem 82:683– 60. Garnero P, Tsouderos Y, Marton I, Pelissier C, Varin C,
691.
Delmas PD 1999 Effects of intranasal 17␤-estradiol on bone
turnover and serum insulin-like growth factor I in postmeno-
pausal women. J Clin Endocrinol Metab 84:2390–2397.
43. Gitelman SE, Kirk M, Ye J-Q, Filvaroff EH, Kahn AJ,
Derynck R 1995 Vgr-1/BMP-6 induces osteoblastic differen-
tiation of pluripotent mesenchymal cells. Cell Growth Differ- 61. Bentley H, Hamdy FC, Hart KA, Seid JM, Williams JL,
ent 67:827–836.
44. Hughes FJ, Collyer J, Stanfield M, Goodman SA 1995 The
effects of bone morphogenetic protein-2, -4, and -6 on differ-
Johnstone D, Russell RGG 1992 Expression of bone morpho-
genetic proteins in human prostatic adenocarcinoma and be-
nign prostatic hyperplasia. Brit J Cancer 66:1159–1163.
entiation of rat osteoblast cells in vitro. Endocrinology 136: 62. Turner RT, Kidder LS, Zhang M, Harris SA, Westerlind KC,
2671–2677.
Maran A, Wronski TJ 1999 Estrogen has rapid tissue-specific
effects on rat bone. J Appl Physiol 86:1950–1958.
45. Boden SD, Hair G, Titus L, Racine M, McCuaig K, Wozney
JM, Nanes MS 1997 Glucocorticoid-induced differentiation of 63. Solloway MJ, Dudley AT, Bikoff EK, Lyons KM, Hogan
fetal rat calvarial osteoblasts is mediated by bone morphoge-
BLM, Robertson EJ 1998 Mice lacking BMP-6 function. De-
netic protein-6. Endocrinology 138:2820–2828.
velop Genet 22:321–339.
46. Yamaguchi A, Ishizuya T, Kintou N, Wada Y, Katagiri T, 64. Watson PH, Watson AJ, Hodsman AB 1996 Expression of
Wozney JM, Rosen V, Yoshiki S 1996 Effects of BMP-2,
BMP-4 and BMP-6 on osteoblastic differentiation of bone
marrow-derived stromal cell lines, ST2 and MC3T3–G2/PA6.
Biochem Biophys Res Commun 220:366–371.
growth factor ligand and receptor genes in rat cancellous bone
trabeculae and marrow. J Mol Endocrinol 17:45–54.
47. Gitelman SE, Kobrin MS, Ye J-Q, Lopez AR, Lee A, Derynck
R 1994 Recombinant Vgr-1/BMP-6-expressing tumours in-
duce fibrosis and endochondral bone formation in vivo. J Cell
Biol 126:1595–1609.
48. Sharrock WJ 1998 Bone and the hematopoietic and immune
systems: A report of the proceedings of a scientific workshop.
J Bone Miner Res 13:537–543.
Address reprint requests to:
J. H. Tobias, M.D., Ph.D., F.R.C.P.
Rheumatology Unit
Bristol Royal Infirmary
Bristol BS2 8HW, UK
49. Perry MJ, Samuels A, Bird D, Tobias JH 2000 The effects of
high-dose estrogen on murine hematopoietic marrow precede Received in original form June 6, 2001; in revised form September
those on osteogenesis. Am J Physiol 279:E1159–E1165. 26, 2001; accepted October 30, 2001.