(Received for publication, February 12, 1997, and in revised form, May 16, 1997)
From the Yale University School of Medicine, Section
of Plastic Surgery, New Haven, Connecticut 06520-8041 and the
¶ Division of Molecular Medicine, Oregon Health Sciences
University, Portland, Oregon 97201-3098
Insulin-like growth factor-I (IGF-I)
is a key factor in bone remodeling. In osteoblasts, IGF-I synthesis is
enhanced by parathyroid hormone and prostaglandin E2
(PGE2) through cAMP-activated protein kinase. In rats,
estrogen loss after ovariectomy leads to a rise in serum IGF-I and an
increase in bone remodeling, both of which are reversed by estrogen
treatment. To examine estrogen-dependent regulation of
IGF-I expression at the molecular level, primary fetal rat osteoblasts
were co-transfected with the estrogen receptor (hER, to ensure active
ER expression), and luciferase reporter plasmids controlled by promoter
1 of the rat IGF-I gene (IGF-I P1), used exclusively in these cells. As
reported, 1 µM PGE2 increased IGF-I P1
activity by 5-fold. 17-Estradiol alone had no effect, but
dose-dependently suppressed the stimulatory effect of
PGE2 by up to 90% (ED50 ~0.1
nM). This occurred within 3 h, persisted for at least
16 h, required ER, and appeared specific, since 17
-estradiol was 100-300-fold less effective. By contrast, 17
-estradiol
stimulated estrogen response element (ERE)-dependent
reporter expression by up to 10-fold. 17
-Estradiol also suppressed
an IGF-I P1 construct retaining only minimal promoter sequence required
for cAMP-dependent gene activation, but did not affect the
60-fold increase in cAMP induced by PGE2. There is no
consensus ERE in rat IGF-I P1, suggesting novel downstream interactions
in the cAMP pathway that normally enhances IGF-I expression in skeletal
cells. To explore this, nuclear extract from osteoblasts expressing hER
were examined by electrophoretic mobility shift assay using the
atypical cAMP response element in IGF-I P1. Estrogen alone did not
cause DNA-protein binding, while PGE2 induced a
characteristic gel shift complex. Co-treatment with both hormones
caused a gel shift greatly diminished in intensity, consistent with
their combined effects on IGF-I promoter activity. Nonetheless, hER did
not bind IGF-I cAMP response element or any adjacent sequences. These
results provide new molecular evidence that estrogen may temper the
biological effects of hormones acting through cAMP to regulate skeletal
IGF-I expression and activity.
Although postmenopausal osteoporosis is one of the most prevalent age-related skeletal disorders, our understanding of the role of estrogen, which has a critical role in maintaining bone mass, remains incomplete. The principal laboratory animal model of postmenopausal osteoporosis is the ovariectomized (OVX)1 rat. One consistent observation with this model is an increase in the rate of bone remodeling after OVX-induced estrogen depletion (1-5). Furthermore, consistent with a cause and effect relationship, estrogen administration re-establishes a reduced rate of bone remodeling in the OVX rat.
Bone remodeling consists of two opposing events, i.e. bone
resorption and formation (6). Skeletal integrity in adults relies on
closely coupled remodeling where there is a balance between these
catabolic and anabolic processes. Net bone loss in postmenopausal osteoporosis is thought to result from bone resorption that exceeds formation, even though both processes are accelerated in this condition
(7-9). Recent data indicate the involvement of several interleukins
(IL-1 and IL-6) in estrogen deficiency-induced bone resorption
(10-13). IL-6 has emerged as a pivotal factor in accelerated bone
resorption in the estrogen-depeleted state, due in part to the apparent
absence of bone loss in OVX mice carrying a knockout of the IL-6 gene
(14). While progress has been made in understanding the interactions of
estrogen and interleukins in regulating bone resorption, relatively
little is known about factors involved in bone formation in the
estrogen-depleted state. Recently, several groups reported elevated
serum IGF-I levels that returned to normal following estrogen
supplementation in OVX rats (15, 16). Similarly, postmenopausal women
given estrogen experience a drop in serum IGF-I levels (17). The
endocrine influence of elevated levels of circulating IGF-I is likely
to include a positive effect on bone formation (18-21).
Local osteoblast-derived IGF-I also increases bone cell activity (18, 19, 22-26). The osteotropic hormone, parathyroid hormone (PTH), while stimulating resorption, also stimulates IGF-I synthesis by osteoblasts through a mechanism involving cAMP-dependent protein kinase A (22, 23, 27). PTH administered intermittently in vivo or in vitro increases osteoblast replication and matrix collagen synthesis (23, 28-32). An increase in IGF-I expression is in part responsible for the anabolic effect of PTH. In fetal rat calvarial bone explants, a brief exposure to PTH increases the rate of collagen synthesis, and neutralizing antibodies to IGF-I dramatically reduce this effect (23). Therefore, IGF-I appears to couple this aspect of bone remodeling.
Earlier studies examining the influence of estrogen on IGF-I expression by primary cultures of neonatal rat calvarial osteoblasts indicated a small stimulatory effect (~2-fold), thought to result from changes in the rate of gene transcription (33, 34). However, OVX increases IGF-I mRNA expression in the calvariae of growing rats, and this effect can be reversed by the estrogen agonist diethylstilbesterol (35, 36). These latter observations agree with the elevated serum IGF-I levels seen in OVX rats, indicated earlier (15, 16). Therefore, in vivo and in vitro findings together predict that estrogen may suppress IGF-I expression in skeletal tissue and that this may be focused on the stimulatory effect of agents that induce cAMP. This interpretation is consistent with the lower rate of bone remodeling seen in the estrogen replete versus estrogen-depleted animals.
In this study we wished to evaluate the effect of 17-estradiol on
IGF-I promoter activity. Because we previously demonstrated a
PKA-dependent increase in IGF-I transcription after
exposure to PGE2 in fetal rat osteoblasts (37-40), we used
this model to examine the influence of 17
-estradiol alone and in
combination with PGE2 on IGF-I promoter function.
Furthermore, because other studies suggested that estrogen can suppress
an increase in cAMP in response to PTH (33), we also examined the
effect of 17
-estradiol on PGE2-stimulated cAMP in these
cultures. Although we find a potent regulatory effect of
17
-estradiol on cAMP-dependent IGF-I expression, it
occurs downstream of cAMP accumulation.
Primary osteoblast-enriched cell cultures
were prepared from the parietal bones of 22-day-old Sprague-Dawley rat
fetuses (Charles River Laboratories, Raleigh, NC). Animals were housed
and euthanized by methods approved by the Yale University Animal Care
and Use Committee. Cranial sutures were eliminated during dissection, and the bones were digested with collagenase for five sequential 20-min
intervals. The cell population released during the last three
digestions exhibits biochemical characteristics associated with
differentiated osteoblasts, including PTH receptors, type I collagen
synthesis, and a rise in osteocalcin expression in response to
dihydroxyvitamin D3 (41, 42). Histochemical staining demonstrates that approximately 80% of the cells express alkaline phosphatase,2 although this itself is not
entirely specific for osteoblasts. However, using these criteria, as
well as differential sensitivity to transforming growth factor-,
bone morphogenetic protein-2, various prostaglandins, and the ability
to form mineralized nodules in vitro (43, 44), these cells
are well distinguished from the less differentiated cells released
during earlier collagenase digestions. Cells from the last three
digestions were pooled and then plated at 4,800 per cm2 in
Dulbecco's modified Eagle's medium (DMEM) containing 20 mM HEPES (pH 7.2), 0.1 mg/ml ascorbic acid, penicillin and
streptomycin (all from Life Technologies Inc.) and 10% fetal bovine
serum (Sigma). At the time of transfection, cultures were rinsed twice
with phenol red-free, serum-free DMEM. All subsequent culturing was in
phenol red-free DMEM, with (growth medium) or without (treatment
medium) charcoal-stripped serum, prepared as described by Ernst and
Rodan (34).
Rat IGF-I promoter 1 constructs have been
described previously (45). A Rous sarcoma virus promoter-driven
expression vector for both the human estrogen receptor (pRSV-hER;
containing a glycine to valine substitution at position 400), and a
recombinant luciferase reporter construct containing two tandem
estrogen response elements (ERE from the vitellogenin gene) along with
the minimal prolactin promoter ((ERE)2PRL36-Luc), were
kindly provided by Dr. Stuart Adler (Washington University School of
Medicine). All plasmids were propagated in Escherichia coli
strain DH5 with ampicillin selection and were prepared with a Wizard
Plus Maxiprep DNA purification kit (Promega Corp., Madison, WI) and the
manufacturer's recommended protocol.
IGF-I promoter 1-luciferase reporter
plasmids (1.25 µg/9.6-cm2 culture well) were
co-transfected with pRSV-hER (1.25 µg/culture well). Cultures at 40%
confluence were rinsed in serum-free medium and exposed to plasmids in
the presence of 0.5% LipofectinTM (Life Technologies Inc.) for 3 h. The solution was then replaced with growth medium containing 5%
charcoal-stripped fetal bovine serum, and cultures were re-incubated
for 48 h. Subconfluent cultures were rinsed with serum-free medium
and treated for the indicated times with vehicle (ethanol-diluted
1/1,000 or greater), PGE2, 17-estradiol, or
17
-estradiol (all from Sigma; each was dissolved in 95% ethanol),
either individually or in combination. After treatment, medium was
aspirated, and cultures were rinsed with phosphate-buffered saline and
then lysed in 100 µl of 25 mM Tris phosphate (pH 7.8), 2 mM dithiothreitol (DTT), 2 mM
1,2-diaminocyclohexane-N,N,N
,N
-tetraacetic acid, 10% glycerol, 1% Triton X-100 (cell lysis buffer, Promega Corp.). Nuclei were pelleted at 12,000 × g for 5 min,
and supernatants were stored at
75 °C until assay. A commercial
kit (Promega Corp.) was used to measure luciferase by single channel
photon counting, and enzyme activity was corrected for protein content
as determined by the Bradford assay (46).
Cultures were transiently transfected
with pRSV-hER. After 48 h, cultures were rinsed twice with
serum-free phenol red-free DMEM and exposed to control medium or 10 nM 17-estradiol for 30 min. Cultures were then treated
for 5 min with 0.5 mM isobutylmethylxanthine, to inhibit
endogenous phosphodiesterase activity, prior to a 10-min treatment (at
room temperature) with control (ethanol vehicle), PGE2 (1 µM), 17
-estradiol (0.1-10 nM), alone or
in combination. Treatments were terminated by aspirating the medium,
and the plates were immediately frozen at
75 °C. Cultures were
extracted with n-propyl alcohol, and the samples were
air-dried and then dissolved in 50 mM sodium acetate (pH
6.2). An aliquot was used to measure cAMP content using a highly
specific commercial radioimmunoassay kit (Biomedical Technologies,
Inc.). The amount of cAMP in each extract was determined from the
linear portion of a standard curve and is expressed as picomoles
of cAMP/2-cm2 culture, as reported previously (27).
Control or pRSV-hER-transfected
cultures were rinsed with serum-free medium and exposed to treatment
solution for 4 h. Medium was aspirated, and cultures were rinsed
twice with phosphate-buffered saline at 4 °C. All subsequent steps
were performed on ice or at 4 °C. Cells were harvested with a cell
scraper and gently pelleted, and the pellets were washed with
phosphate-buffered saline. Nuclear extracts were prepared by the method
of Lee et al. with minor modifications (39, 47). Cells were
lysed in hypotonic buffer (10 mM HEPES (pH 7.4), 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT) with phosphatase inhibitors (1 mM
sodium orthovanadate, 10 mM sodium fluoride), protease
inhibitors (0.5 mM phenylmethylsulfonyl fluoride, 1 µg/ml
pepstatin A, 2 µg/ml leupeptin, 2 µg/ml aprotinin; Sigma), and 1%
Triton X-100. Nuclei were gently pelleted. Nuclei were resuspended in
hypertonic buffer containing 10 mM HEPES, pH 7.4, 0.42 M NaCl, 0.2 mM Na2EDTA, 25%
glycerol, and the phosphatase and protease inhibitors indicated above.
Soluble proteins released by a 30-min incubation at 4 °C were
collected by centrifugation at 12,000 × g for 5 min,
and the supernatant was collected and aliquoted for storage at
75 °C.
Using previously
published methods (40, 48, 49), radiolabeled double-stranded probe
(Universal DNA, Inc., Tigard, OR) was prepared by annealing
complementary oligonucleotides, followed by fill-in of single-stranded
overhangs with dCTP, dGTP, dTTP, and [-32P]dATP, using
the Klenow fragment of DNA polymerase I. Five to ten µg of nuclear
extract protein was preincubated for 20 min on ice with 2 µg of
poly(dI-dC) with or without unlabeled specific or nonspecific
competitor DNAs in 60 mM KCl, 25 mM HEPES (pH
7.6), 7.5% glycerol, 0.1 mM EDTA, 5 mM DTT,
and 0.025% bovine serum albumin. After the addition of 5 × 104 cpm of DNA probe (0.1-0.2 ng) for 30 min on ice,
samples were applied to a 5% nondenaturing polyacrylamide gel that was
pre-electrophoresed for 30 min at 12.5 V/cm at 25 °C in 45 mM Tris, 45 mM boric acid, 1 mM
EDTA. Electrophoresis proceeded for 2.5 h under identical conditions. Dried gels were exposed to x-ray film at
75 °C with intensifying screens. Electrophoretic mobility supershift analysis was
performed in select experiments using normal rabbit serum as a negative
control and monoclonal anti-rhER antibody (Santa Cruz Biotechnologies
Inc., Santa Cruz, CA) for its ability to produce a supershift of
hER-bound probe. Sequences of sense strands of oligonucleotides used in
the electrophoretic mobility shift assays are as follows: HS3D,
5
-GAGCAGATAGAGCCTGCGCAATCGAAATA-3
(the minimal binding
sequence of the atypical cyclic response element within the HS3D
footprint site of the rat IGF-I promoter 1 is underlined (40)) and ERE,
5
-GATCCAAAGTCAGGTCACAGTGACCTGATCAAAGA-3
(the estrogen
response element is underlined in this sequence).
17-Estradiol, 17
-estradiol, and
PGE2 were obtained from Sigma and were 99% pure as
determined by TLC.
Data were assessed by one way analysis of variance with SigmaStat® software, with the Student-Newan-Keuls method for post hoc comparison.
The IGF-I gene has two promoters, and promoter 1 (P1) is used
exclusively in cultures of fetal rat osteoblasts (38). Initial studies
were conducted with osteoblast cultures transiently transfected with
the longest IGF-I P1-luciferase reporter construct, IGF1711b/Luc (containing 1711 bp of upstream sequence, and the initial 328 bp of
transcribed but untranslated exon 1), which we previously found to be
most responsive to treatment with the cAMP-inducing agent
PGE2 (39). To ensure expression of active estrogen
receptors, cultures were co-transfected with the hER expression vector
pRSV-hER. Incubation with 17-estradiol at 0.01-10 nM
for 16 h did not alter basal IGF-I P1 activity (Fig.
1, left panel). As a positive control, parallel cultures were co-transfected with pRSV-hER and a recombinant reporter plasmid containing tandem copies of the consensus estrogen response element (ERE) from the vitellogenin gene, and a short segment
of the prolactin promoter (plasmid (ERE)2PRL-Luc). Unlike
the IGF-I P1 recombinant plasmid, treatment with 17
-estradiol
enhanced expression of this ERE-containing reporter gene up to
10-fold (Fig. 1, center panel). Co-transfection with pRSV-hER was required for all 17
-estradiol-dependent
responses, including activation of the (ERE)2PRL-Luc
construct (data not shown).
Fetal rat osteoblasts respond to PGE2 with a rapid increase
in IGF-I expression, mediated by PKA-dependent activation
of IGF-I P1 (39). This occurs through an atypical CRE located at bp
+202 to +209 of exon 1 (40), and PTH similarly requires this element to
enhance IGF-I promoter function (data not shown). As we reported previously, PGE2 (1 µM, 16 h) increased
IGF-I P1 activity 5-fold (39). While 17-estradiol alone was
ineffective, co-treatment with PGE2 led to a
dose-dependent suppression of IGF-I promoter activation
over a concentration range of 0.1-10 nM (ED50
~0.1 nM) (Fig. 1, left panel). This response
required expression of hER (Fig. 1, right panel), and an
active form of estradiol, since 17
-estradiol was 100-300-fold less
potent than 17
-estradiol at blocking PGE2-mediated IGF-I
P1 activity (Fig. 1, left panel). By contrast, expression of
the ERE-driven reporter gene was not inhibited by co-treatment with
17
-estradiol and PGE2 (see Fig. 3).
Time course experiments next were performed to examine the kinetics of
inhibition of IGF-I P1 activity by 17-estradiol. Cells were treated
with PGE2 (1 µM) or 17
-estradiol (10 nM), alone or in combination for 3, 6, or 16 h.
PGE2-activated expression of the IGF-I P1 reporter gene was
rapidly blunted by 17
-estradiol. Both the stimulatory effect of
PGE2 on IGF-I promoter activity and the opposing,
suppressive influence of 17
-estradiol were observed within 3 h
of hormone treatment. Both responses were maintained throughout a
16 h incubation, while the ERE-containing plasmid
((ERE)2PRL-Luc) was activated progressively by
17
-estradiol (Fig. 2).
IGF-I P1 lacks a consensus ERE, and yet 17-estradiol suppressed
cAMP-dependent promoter activation. Therefore, an effort was made to localize the region(s) of the promoter responsible for the
opposing effect of estrogen on cAMP-mediated gene activation. Deletion
of 5
flanking DNA from nucleotides
1711 to
123 (with respect to
the most 5
transcription start site of rat IGF-I exon 1 (45)) did not
alter either PGE2-stimulated promoter activation (39), or
the opposing inhibitory effect of 17
-estradiol (Fig. 3). All three of the plasmids tested retained the same
328 bp of exon 1, which preserves the unconventional CRE (located at bp
+202 to +209) required for the stimulatory effect of PGE2
(40).
Next, several approaches were taken to determine the mechanism by which
estrogen might suppress IGF-I promoter activation by PGE2.
Previous conflicting reports suggested a possible inhibitory effect by
estrogen on cAMP accumulation in response to PTH in osteoblasts (33,
50). We examined cAMP levels in cells that were treated with 1 µM PGE2 alone or in combination with 10 nM 17-estradiol (as in the promoter/reporter assays) and
in cells preincubated with 17
-estradiol for 30 min followed by
PGE2 treatment. All cultures were transfected 48 h
earlier with pRSV-hER and pretreated with isobutylmethylxanthine for 5 min to inhibit phosphodiesterase activity. In all experiments, cAMP
accumulation increased by 60-fold after 10-min treatment with
PGE2, even in the presence of 17
-estradiol co-treatment
or pretreatment (Fig. 4). In addition, regardless of the
presence of 17
-estradiol, PGE2-induced accumulation of cAMP was accompanied by pronounced morphological changes, as seen previously with these cells (data not shown). These results indicate that the adenylate cyclase pathway was intact and functional in cells
exposed to 17
-estradiol.
While cAMP levels were not altered by 17-estradiol, estrogen may
interfere with other steps in the process of PKA-dependent IGF-I gene activation that lie downstream of the generation of cAMP.
One possibility is that the ligand-activated ER may modulate the
binding of cAMP-dependent transcription factor(s) to the
IGF-I CRE. To examine this possibility, nuclear extracts prepared from control and PGE2-treated cells were tested in the
electrophoretic mobility shift assay using a 32P-labeled
double-stranded oligonucleotide containing the IGF-I CRE (designated
HS3D, based on the notation for the footprinted site (45)). In
agreement with our earlier observations, a gel-shifted doublet band was
seen with this oligonucleotide using nuclear extracts from
PGE2 (1 µM, 4 h)-treated cultures (40).
The PGE2-inducible gel shift that occurred for extracts
from cells previously transfected with the hER expression plasmid was
identical to that for untransfected cultures (Fig. 5).
In cells expressing hER, incubation with 17
-estradiol alone did not
induce nuclear protein binding to the HS3D oligonucleotide probe.
However, the amount of nuclear protein complex that occurred with
extract from hER tranfected cells co-treated with PGE2 and 17
-estradiol was dose-dependently reduced in intensity
compared with PGE2 treatment alone, but there was no change
in the banding pattern. In contrast 17
-estradiol was not similarly
inhibitory.
We then examined the possibility that ER directly associates with the
IGF-I CRE and therefore competes with cAMP-stimulated transcription
factor(s) for binding at this element. First, we tested recombinant hER
(rhER) for its ability to bind a 32P-labeled consensus ERE
oligonucleotide as a positive control and to determine if excess
unlabeled HS3D oligonucleotide could compete for rhER binding. Without
rhER, no gel-shifted band was detected with 32P-labeled ERE
oligonucleotide. However, the association of rhER with
32P-labeled ERE oligonucleotide produced strong gel shift
complexes that were significantly reduced in intensity by 50-fold molar excess of unlabeled homologous ERE, but were unaffected by excess unlabeled HS3D oligonucleotide. To confirm the specificity of the
hER/ERE interaction further, a monoclonal anti-hER antibody was used in
a supershift assay. As shown, normal rabbit serum was ineffective,
while anti-hER antibody altered the mobility of the
hER-dependent gel-shifted band (Fig. 6,
left panel). Next, to examine binding of hER to the IGF-I
promoter directly, two 32P-labeled double-stranded DNA
probes generated from the 122 bp to +328 bp portion of rat IGF-I P1
were combined with recombinant hER. No gel shift complex was seen with
either fragment. These findings indicate that no direct interactions
occur between rhER and the regions of IGF-I P1 that contain or flank
HS3D (Fig. 6, right panel), even though HS3D is required for
stimulation of reporter gene expression by PGE2 and for the
opposing influence of 17
-estradiol.
Estrogen deprivation has long been associated with a decline in skeletal integrity, but its role in the normal maintenance of bone mass is still poorly understood. Several lines of evidence indicate that estrogen deficiency increases the total process of bone remodeling in such a way that the rate of bone resorption significantly outpaces formation and results in osteoporosis (1-5). Studies with OVX rats, where surgical estrogen depletion incurs a decrease in bone mass, strongly connect IL-1 and IL-6 to accelerated bone resorption. In vivo studies using an IL-1 receptor antagonist demonstrate a bone sparing effect similar to that observed with estrogen replacement therapy, while IL-6 knockout mice do not experience the OVX-induced bone loss seen in their heterozygous littermates (14, 51). In vitro studies indicate that both lymphokines are associated with activation of mature osteoclasts and stimulation of osteoclastogenesis. These results have increased our understanding of the high rates of bone resorption that follow the loss of endogenous estrogen. Nevertheless, little is still known about the mechanisms by which bone formation rates also increase, albeit to a lesser extent, under similar conditions.
Estrogen has a positive influence on the growth of some tissues, such as uterine tissue. In rats, OVX causes weight gain and increased longitudinal bone growth, both of which are reversed by estrogen treatment. Consistent with this, OVX increases the circulating levels of IGF-I, and estrogen treatment can normalize this change (15, 16). Similar results were observed in postmenopausal women (17). Liver is a primary source of serum IGF-I. In this regard, estrogen suppresses growth hormone-stimulated hepatic IGF-I synthesis in hypophysectomized and OVX animals, even though it increases serum growth hormone levels in OVX animals (52, 53). These latter findings may help explain some in vivo results with OVX rats. Interestingly, estrogen treatment is also associated with accelerated skeletal maturation and closure of growth plates in both rats and humans (54). Local IGF-I synthesis by bone cells is also induced by growth hormone and contributes significantly to longitudinal bone growth (4), predicting that estrogen can oppose its stimulatory effect on IGF-I synthesis in liver as well as peripheral tissue. OVX also alters several markers of osteoblast activity in intramembranous bone such as calvariae (the source of osteoblastic cells used in our studies). Within 1 week of OVX, transcript levels for osteocalcin, osteonectin, and IGF-I increase at this site. Subsequent treatment with the estrogen analog diethylstilbesterol rapidly decreases the appearance of these mRNAs, reduces periosteal mineral apposition rate, and also suppresses osteoblast number (8, 35, 36, 55).
Alterations in serum PTH levels have also been reported after OVX (5,
56). We first identified that IGF-I is one coupling factor in the
bone-remodeling process. At least part of the anabolic action of PTH on
skeletal tissue is mediated by its stimulatory effect on IGF-I
expression by way of an increase in cAMP (22, 23). Since then we have
examined the cellular and molecular mechanisms that regulate IGF-I
synthesis in this system. We recently identified an atypical cAMP
response element that controls PKA-dependent (PGE2- or PTH-activated) IGF-I gene expression in fetal rat
bone cells (40). Even more recently we found that this
can occur through activation of one of several previously characterized trans-activating factors that associate with this atypical
CRE.3 Like PTH, PGE2 can prevent or
significantly restore bone loss due to OVX or immobilization (57-59).
PGE2 is synthesized locally in the skeleton following
exposure to cytokines (transforming growth factor- and IL-1),
hormones such as PTH, and mechanical strain (59). In osteoblasts,
PGE2 also activates IGF-I expression through a
PKA-dependent mechanism, where it stimulates IGF-I gene transcription, mRNA accumulation, and protein production
(37-40).
Our current studies revealed that 17-estradiol alone does not
directly alter IGF-I promoter activity. However, it suppresses PKA-dependent IGF-I gene activation in a rapid and
sustained, estrogen receptor-dependent, isoform-specific,
and dose-dependent way. The ED50 of
17
-estradiol was ~0.1 nM, within its physiological concentration range (60), while 17
-estradiol was 100-300-fold less
effective. This suppressive effect was not due to hormone toxicity,
since the same doses of 17
-estradiol stimulated expression of a
reporter gene containing a consensus ERE. Furthermore, neither estrogen
pretreatment or co-treatment altered cAMP accumulation, indicating that
its effect on IGF-I expression occurs downstream of cAMP itself.
Reporter gene expression, under control of the smallest fragment of
IGF-I promoter DNA, which retains PKA sensitivity, was completely
suppressed by 17
-estradiol. This observation co-localizes the
responsive sequences for each event to one element or a small region of
DNA. The rat IGF-I promoter 1 contains no consensus ERE or CRE DNA
sequences. However, electrophoretic mobility shift assays using the
HS3D region of the IGF-I promoter, where the atypical IGF-I promoter
CRE resides, revealed that 17
-estradiol reduced the intensity of the
cAMP-inducible gel shift. No additional bands or changes in relative
mobility occurred with estrogen treatment. Therefore, interactions
between hER and the cAMP-activated nuclear factor(s) or co-activator(s)
appear to limit nuclear factor binding to the atypical IGF-I CRE within
the HS3D region itself. Complex kinetics for the suppressive effect of
17
-estradiol on PKA-activated IGF-I promoter function may indicate
multiple actions of 17
-estradiol on activation, binding, and
synthesis of the responsible cAMP-activated transcription factor. Once
we have identified with certainty the cAMP-activated transcription
factor responsible for this effect, we will be better positioned to
resolve the mechanism by which its binding is reduced at this
element.
In contrast to our findings, earlier studies found that estrogen alone
may enhance IGF-I mRNA levels ~2-fold in neonatal rat osteoblasts
in vitro (33, 34). We do not yet understand this obvious
difference with our results in fetal bone cells or with other in
vivo studies where IGF-I mRNA in intramembranous bone increases after OVX (35, 36). Our findings clearly indicate that
estradiol reduces nuclear factor binding to the atypical CRE found in
the IGF-I promoter, which is much less evident in the basal state.
Consequently, the ~2-fold stimulation by estrogen on basal IGF-I
expression that was reported previously differs in context from our
studies that focus on the ability of this hormone to suppress the more
potent 5-10-fold increase in IGF-I expression that occurs in response
to PKA activation. Transfection-mediated hER expression was required to
observe an effect of estrogen in the studies we present here, including
reporter gene expression by the positive estrogen responsive
(ERE)2-PRL-Luc, indicating low to negligible levels of ER
in our fetal osteoblast cultures. Unlike the earlier results of Ernst
and Rodan (34), we have been unable to detect an effect of
17-estradiol on IGF-I transcript or promoter function in newly
confluent or confluent and aged untransfected primary fetal rat
osteoblasts. The reasons for these discrepancies are unclear, but may
reflect differences in neonatal bone used by Ernst and Rodan (34)
versus fetal bone used in our studies to prepare the primary
cell cultures or subsequent culture conditions.
Other earlier studies identified the importance of an AP-1 binding site
for the hER-dependent stimulatory effect of 17-estradiol on chicken IGF-I promoter activity in human hepatoma cells (61). Again,
we could not detect any effect by 17
-estradiol alone on rat IGF-I
promoter activity with or without co-expression of hER in fetal rat
bone cells. The minimal region of the IGF-I promoter, which is
sensitive to both PGE2 and 17
-estradiol, does not
contain an identifiable AP-1 site (45). We further examined agents such as phorbol myristate acetate, shown to active AP-1, and IGF-I promoter
activity was not enhanced (39). These results may reflect differences
between rat and chicken IGF-I promoters, between human and rat cells
used to assess promoter activity, between liver and bone cells or
certain variations that may sometimes occur when promoters are examined
in heterologous assay systems.
In contrast to the antagonistic effects that we observe within the
IGF-I promoter, activation of PKA can synergistically enhance the
stimulatory effect of estrogen on expression of reporter genes containing a conventional ERE (62), as we observed in osteoblasts for
the (ERE)2PRL-Luc construct. Potentiation of estrogen
action may occur with no detectable changes in ER expression, in
estrogen binding to ER, or in ER binding to the ERE, but may result
from changes in interactions among other components of the
transcriptional apparatus. Additional signals modulated by growth
factor receptors, induced by transforming growth factor , epidermal
growth factor, or IGF-I, also may activate ERE-containing promoters
through estrogen receptor-dependent transcriptional events,
possibly by activation of the Ras-mitogen-activated protein kinase
signal transduction cascade (63). This latter finding further suggests
participation by endogenously synthesized IGF-I in some estrogenic
effects in skeletal tissue.
Osteoclastic bone resorption can also be regulated by local, osteoblast-derived factors whose expression is controlled by yet other local and systemic agents (64, 65). In this context, IL-6 expression is suppressed by estrogen (11-13), while it is induced by IGF-I (66). Consistent with other opposing effects of estrogen and agents that induce cAMP, estrogen may suppress PTH-stimulated osteoclast formation by interfering with a downstream event in the PKA-dependent signal pathway in osteoblasts (67). Furthermore, osteoporotic postmenopausal women with primary hyperparathyroidism are as effectively treated by estrogen replacement as they are by parathyroidectomy (68).
A recent case report may indicate the importance of estrogen in men as well as women. A 28-year-old male with an estrogen receptor mutation exhibited incomplete epiphyseal growth plate closure, continued longitudinal bone growth, and osteoporosis, even in the presence of high levels of serum estrogen and normal levels of testosterone (69). Furthermore, mice with insertional disruption of the estrogen receptor have decreased skeletal mineralization in both females and males (70). Therefore, a functional estrogen signal pathway may be needed for proper skeletal maturation and balanced bone remodeling sequence in either sex.
In conclusion, while 17-estradiol alone had no effect on IGF-I
promoter function, it rapidly and significantly reduced
PGE2-mediated promoter activation. This effect was
estradiol isomer-selective and ER-dependent. While
recombinant hER does not bind directly to the region encompassing the
atypical IGF-I CRE, 17
-estradiol significantly reduced the
activation or binding of the transcription factor(s) responsible for
cAMP inducible binding to this element. These findings provide new
molecular evidence that estrogen may temper PGE2- and
PTH-regulated IGF-I actions in bone and may help to explain the
additional aspects of altered bone remodeling that occurs with estrogen
depletion. These results may have implications for therapeutic regimens
for postmenopausal osteoporosis that include combinations of agents to
activate bone formation while suppressing the even greater rates of
resorption that accompany this event in vivo.