(Received for publication, May 8, 1996, and in revised form, September 5, 1996)
From the Department of Molecular Genetics, University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030
1,25-Dihydroxyvitamin D3 (1,25-(OH)2D3), a key regulator of mineral metabolism, regulates the expression of several genes that are expressed in osteoblasts. In particular, in rat and human osteoblasts, 1,25-(OH)2D3 increases the expression of Osteocalcin by interacting, through a hormone-receptor complex, with a vitamin D-responsive element present in the promoter of the genes. Here we show that in mouse, 1,25-(OH)2D3 inhibits the expression of both osteocalcin genes, OG1 and OG2. This inhibition was observed in primary osteoblast cultures and in the whole animal. From sequence inspection, DNA transfection experiments, and DNA binding assays, we could not identify a functional vitamin D-responsive element in the promoter of OG2 or in the first 3.3 kilobases of the OG1 promoter. However, we show that 1,25-(OH)2D3 treatment of primary osteoblasts abolishes the binding of OSF2, an osteoblast-specific activator of transcription that binds to OSE2, a critical osteoblast-specific cis-acting element present in OG1 and OG2 promoters. Consistent with these DNA binding data, a mutation in OSE2 in the OG2 promoter abrogated the inhibitory effect of 1,25-(OH)2D3 treatment on this promoter activity. This study illustrates that 1,25-(OH)2D3 can play different roles in the expression of the same gene in various species and indicates that this regulation in mouse occurs through an indirect mechanism, 1,25-(OH)2D3 acting on a gene genetically located upstream of Osteocalcin.
The steroid hormone 1,25-(OH)2D3 1 is a key regulator of mineral homeostasis in mammals. Its main function is to promote intestinal absorption of mineral ions and to mobilize calcium from bone (1, 2). In bone, 1,25-(OH)2D3 favors osteoclast and osteoblast differentiation (1, 2, 3, 4). Although the overall function of 1,25-(OH)2D3 in bone is to promote resorption, it is the osteoblasts and not the osteoclasts that contain vitamin D receptors (VDRs). As a consequence, the expression of several genes in osteoblasts is regulated by 1,25-(OH)2D3. For instance, the type I collagen genes (5) and the bone sialoprotein genes (6) are down-regulated by the hormone, while the osteopontin (7) and osteocalcin (8, 9, 10, 11, 12) genes are up-regulated in human and rat.
The role of 1,25-(OH)2D3 in the regulation of
synthesis of osteocalcin, a noncollagenous protein of the bone
extracellular matrix that regulates bone formation (13), has been
extensively studied in human and rat.
1,25-(OH)2D3 increases the synthesis of this
protein by rat osteosarcoma cells and increases the accumulation of its
mRNA in rat osteoblastic cells (8). It also induces a rapid
increase in Osteocalcin expression in the whole animal early
after 1,25-(OH)2D3 treatment (14). The
promoters of the human and rat osteocalcin genes were the first
promoter elements in which VDREs were identified and characterized
(10, 11, 12). The VDRE is typically made up of two direct repeats separated by a 3-bp spacer. This finding contributed to the definition of the
"3, 4, 5 rule" for DNA binding of various nuclear receptors (15).
The hormone-VDR complex binds to this element as an heterodimer with
the retinoid X receptor (16). Notably, the VDRE is located in most
cases within the first kilobase (kb) of 5-flanking sequence of the
target gene (17). In spite of this regulatory role, the precise
physiological function of 1,25-(OH)2D3 during
osteoblast differentiation in vivo remains unclear. The fact
that normal bone mineralization can be achieved in human in the absence
of a functional VDR, provided that adequate serum calcium and
phosphorus levels are maintained (18), suggests that the main function of the hormone in bone is to mobilize calcium through resorption.
We recently cloned the two mouse osteocalcin genes, OG1 and OG2 (19), and used the OG2 promoter to study the mechanisms controlling the osteoblast-specific expression of this gene (20, 21). In the course of this work, we were surprised to find a total absence of up-regulation of this promoter activity by 1,25-(OH)2D3, even though a VDRE-like sequence was present. Thus, we decided to systematically study this aspect of the regulation of expression of the mouse osteocalcin genes in vivo and in vitro. Here we show that the expression of the mouse osteocalcin genes, unlike their rat and human homologues, is down-regulated by 1,25-(OH)2D3 and that an indirect mechanism is involved in this down-regulation.
Primary osteoblasts were isolated from
calvariae of 2-day-old mice as described previously (20, 22), with the
following modifications. Calvariae were sequentially digested for 20, 40, and 90 min at 37 °C in -modified minimum essential medium
(Life Technologies, Inc.) containing 0.1 mg/ml collagenase P
(Boehringer Mannheim) and 0.25% trypsin (Life Technologies, Inc.).
Cells of the first two digests were discarded, whereas cells released
from the third digestion were plated in
-modified minimum essential medium and 10% fetal bovine serum. After 2 days, this medium was replaced by mineralization medium (
-modified minimum essential medium and 10% fetal bovine serum containing 5 mM
-glycerophosphate and 0.1 mg/ml ascorbic acid), which was changed
every 2 days. After 10 days of culture in mineralization medium, the
cells were treated with 1,25-(OH)2D3 dissolved
in ethanol or with vehicle and harvested at specific intervals before
RNA extraction. ROS 17/2.8 rat osteosarcoma cells (23) were maintained
in Dulbecco's modified Eagle's medium, Ham's F-12 medium, and 10%
fetal bovine serum; MC3T3-E1 mouse cells (24) were maintained in
-modified minimum essential medium and 10% fetal bovine serum. Both
cell lines were treated with 1,25-(OH)2D3 or
with vehicle for the appropriate period of time. RNA was prepared using
RNAzOLTM (Cinna/Biotecx Laboratory) following the
manufacturer's instructions.
All inserts were cloned in the p4luc
promoterless luciferase expression vector (25), upstream of the
luciferase gene (luc). A SalI site was introduced
at position +13 of OG2 by PCR amplification of the sequence
extending from positions 657 to +13, using as primers the
oligonucleotides 5
-CCAAGACCTGGCCCAG-3
for the 5
-side and
5
-TGGTCGACTTGTCTGT-3
for the 3
-side. This initial fragment was used
to generate p657OG2-luc (20). p1316OG2-luc was created by inserting the
EcoRV-NcoI fragment of the OG2
promoter (positions
1316 to
657) at the 5
-end of p657OG2-luc (20).
p147OG2-luc was obtained by an NcoI-PvuII
(positions
343 to
147) deletion of p657OG2-luc (20). p2900OG2-luc
was created by inserting at the
1035 PstI site in
p1316OG2-luc a 1.9-kb PstI-BamHI fragment of the
OG2 promoter. p147OG2mut-luc was generated by cloning in p4luc a PCR fragment resulting from the amplification of p147OG2-luc, using as primers the oligonucleotides
5
-GATCCGCTGCAATCACCAAGAACAGCA-3
for the 5
-side and
5
-TGGTCGACTTGTCTGT-3
for the 3
-side. The existence of the 2-bp
substitution mutation was verified by DNA sequencing. To generate
pOG1-luc constructs, a SalI site was first introduced at
position +13 of OG1 by PCR amplification of the OG1 promoter sequence extending from positions
343 to +13,
using as primers the BglII site-containing oligonucleotide
5
-CCGAATTCAGATCTCT-3
for the 5
-side and the SalI
site-containing oligonucleotide 5
-TGGTCGACTTGTCTGT-3
for the 3
-side.
A PstI-BglII fragment of the OG1
promoter (positions
1035 to
343) was cloned upstream of the initial
BglII-SalI fragment to generate p1049OG1-luc.
p3300OG1-luc was created by cloning a PstI-EcoRI
fragment extending 2.3 kb upstream of the PstI site (position
1035) to this latter construction. PCR product sequences were all verified by DNA sequencing. The rat osteocalcin-luc construct was created by insertion of 1.7 kb of the
HindIII-BamHI rat Osteocalcin promoter
fragment (a gift from Dr. M. Demay), blunt-ended, into the
SmaI site of p4luc. pROP-luc contains 1.7 kb of the rat
Osteopontin promoter fused to luciferase (26);
pMOP-luc contains 1 kb (positions
910 to +90) of mouse
Osteopontin promoter fused to luciferase (a gift
from Dr. A. Ridall). Orientation and conformity of the inserts cloned
in p4luc were confirmed by restriction analysis.
Nine-week-old mice (28 g) were injected intraperitoneally with 24 ng of 1,25-(OH)2D3 or with vehicle as described previously (27) and killed 24 h later. Calvariae were dissected out and cleaned of soft tissues, and RNA was prepared by extraction with guanidine isothiocyanate, followed by cesium chloride density gradient ultracentrifugation (28).
RNA AnalysisNorthern blot analysis was performed as
described previously (28). Denatured RNA samples were loaded on 1%
formaldehyde-agarose gels and transferred onto Hybond N+
nylon filters (Amersham Corp.). The filters were then hybridized with
either the Osteocalcin (19) or Osteopontin (29)
probes labeled by random priming. After autoradiography and signal
quantification using a PhosphorImager, filters were stripped and
reprobed with a mouse -actin probe that was used as a control of RNA
loading. Reverse transcription was conducted on 5 µg of DNase-treated
total RNA from 10-day-old primary osteoblast cultures treated with
1,25-(OH)2D3 (10
7 M)
or vehicle 60 h before harvest. cDNA synthesis was performed by oligo(dT) priming using avian myeloblastosis virus reverse transcriptase (Boehringer Mannheim) under conditions suggested by the
supplier. PCR amplification was conducted on one-tenth of the reverse
transcription reaction using as 5
-primer 5
-CAAGTCCCACACAGCAGCTT-3
(positions +8 to +27) and as 3
-primer 5
-AAAGCCGAGCTGCCAGAGTT-3
(positions +359 to +378) to generate a 370-bp PCR product for osteocalcin cDNAs (19). The primers 5
-GTTGAGAGATCATCTCCACC-3
and
5
-AGCGATGATGAACCAGGTTA-3
were used to generate a 320-bp PCR product
corresponding to exon 2 of the hypoxanthine-guanine phosphoribosyltransferase cDNA. For the analysis of reverse
transcription-PCR products, aliquots of the PCR mixtures were separated
on a 1% agarose gel, and the gel was dried. Two synthetic
oligonucleotides were used as probes to detect mRNA specifically
transcribed from either OG1 (5
-AGGACCATCTTTCTGCTC-3
,
positions +35 to +52) or OG2 (5
-AGGACCCTCTCTCTGCTC-3
,
positions +35 to +52) (19). Hybridizations were carried out at
42 °C directly on the gel as described previously (19,
30).
DNA transfections of ROS 17/2.8 and
MC3T3-E1 cells were performed by calcium phosphate precipitation (31).
Cells were treated with 1,25-(OH)2D3 or vehicle
for 64 h before harvest as described previously (11). Luciferase
activities were assayed by using a Monolight 2010 luminometer
(Analytical Luminescence Laboratory) and D-luciferin
substrate (Analytical Luminescence Laboratory) in 100 mM
Tris-HCl (pH 7.8), 5 mM ATP, 15 mM
MgSO4, and 1 mM dithiothreitol. -Galactosidase activities in each lysate, measured by a colorimetric enzyme assay using resorufin
-D-galactopyranoside
(Boehringer Mannheim) as a substrate, were used to normalize
transfection efficiency between experiments.
Nuclear extracts from ROS 17/2.8 cells and primary
osteoblasts were prepared as described (20). Cells were treated with 1,25-(OH)2D3, retinoic acid, BMP7
(one
orphogenetic
rotein
),
-estradiol, or vehicle 48 h before harvest. For
the electrophoretic mobility shift assay, double-stranded mouse
VDRE-like, rat VDRE, OSE2 (
steoblast-
pecific
lement
), and mOCE1 oligonucleotides were
labeled and purified as described previously (20). Five fmol of labeled
double-stranded oligonucleotides were incubated in binding buffer (50 mM KCl, 12 mM HEPES (pH 7.9), 1 mM
EDTA, 1 mM dithiothreitol, and 12% glycerol) with 8 µg
of nuclear extracts and 2 µg of poly(dI-dC) for 15 min at room
temperature. For competition experiments, the indicated amount of
unlabeled double-stranded oligonucleotide was added to the binding
reaction with the other components. The samples were loaded on a 5%
polyacrylamide gel in 0.5 × Tris borate/EDTA (44.5 mM
Tris-HCl, 44.5 mM boric acid, and 4 mM EDTA).
DNA binding and electrophoresis conditions for OSE2 and mOCE1 were as
described previously (20).
In the promoter of one
rat osteocalcin gene, a functional VDRE is located between positions
455 and
441 (10, 11). A similar sequence, but containing a 2-bp
substitution, is located at approximately the same place (positions
465 to
451) in the promoter elements of OG1 and
OG2. To determine whether this was a functional VDRE, we
performed DNA transfection experiments in ROS 17/2.8 rat osteosarcoma
cells using four different constructs. The p1049OG1-luc and p657OG2-luc
promoter-luciferase chimeric genes both contain the VDRE-like sequence;
p147OG2-luc, which lacks any recognizable VDRE, served as a negative
control, and the 1.7-kb rat Osteopontin promoter-luciferase
chimeric gene (pROP-luc), which contains two well characterized
functional VDREs, served as a positive control (26). ROS 17/2.8 cells
were transfected with these vectors and treated with or without
1,25-(OH)2D3 (10
7 or
10
8 M) for 64 h. In agreement with
previous findings (32), the expression of OG1-luc and OG2-luc chimeric
genes was never up-regulated by treatment of transfected cells with
1,25-(OH)2D3 (Fig. 1). Instead,
their expression was always reduced by this treatment, whereas the
expression of pROP-luc, the positive control vector, was always
up-regulated (Fig. 1A).
To demonstrate that the absence of up-regulation of the OG1 and OG2 promoter activities was not due to the choice of a rat cell line for studying the activity of mouse promoters, we performed two additional experiments. First, we tested whether the mouse Osteopontin promoter activity was increased by 1,25-(OH)2D3 treatment in ROS 17/2.8 cells. As shown in Fig. 1A, 1,25-(OH)2D3 treatment of ROS 17/2.8 cells transfected with the pMOP-luc construct, which contains 1 kb of the mouse Osteopontin promoter, increased the activity of this promoter 3.2 times. Second, the regulation of the OG1 and OG2 promoter activities by 1,25-(OH)2D3 was tested in a mouse osteoblastic cell line (MC3T3-E1). The expression of p147OG2-luc was always reduced by treatment of these cells with 1,25-(OH)2D3, while the activity of pMOP-luc exhibited at least 10 times the control activity following the same treatment. These results indicate that the absence of up-regulation of OG1 and OG2 promoter-luciferase constructs by 1,25-(OH)2D3 is not due to the absence of a mouse specific cofactor in ROS 17/2.8 cells.
The results of these experiments also implied that the 2-bp
substitution in the VDRE-like sequence of OG1 and
OG2 promoters abolished or greatly diminished the binding of
the hormone-VDR complex to this site. To test this hypothesis, we
performed an electrophoretic mobility shift assay using as probes an
oligonucleotide encompassing the VDRE of the rat Osteocalcin
promoter (rat VDRE) and an oligonucleotide containing the VDRE-like
element present in OG1 and OG2 promoters (mouse
VDRE-like) (Fig. 2A). As a source of
proteins, we used nuclear extracts of ROS 17/2.8 cells, either untreated or treated for 48 h with retinoic acid
(107 M),
-estradiol (10
8
M), or 1,25-(OH)2D3
(10
8 M). Incubation of untreated or retinoic
acid- or
-estradiol-treated ROS 17/2.8 nuclear extracts with rat
VDRE or mouse VDRE-like probes did not generate a protein-DNA complex
(Fig. 2B, lanes 1-3 and 5). In
contrast, when we used nuclear extracts of
1,25-(OH)2D3-treated ROS 17/2.8 cells, there
was a marked increase in the amount of binding to the labeled rat VDRE
oligonucleotide, but not to the labeled mouse VDRE-like oligonucleotide
(Fig. 2B, compare lanes 1 and 4 and
lanes 5 and 6). This result suggests that the
mouse VDRE-like sequence has a much lower affinity for the
1,25-(OH)2D3-VDR complex than does the rat
VDRE. This was confirmed by DNA competition experiments in the
electrophoretic mobility shift assay. The binding of hormone-treated
ROS 17/2.8 nuclear extracts to the labeled rat VDRE oligonucleotide was
substantially inhibited by excess molar amounts of the unlabeled rat
VDRE oligonucleotide, but not by the same molar amounts of the
unlabeled mouse VDRE-like oligonucleotide (Fig. 2C).
Thus, taken together, the DNA transfection experiments, the DNA binding studies, and the DNA competition experiments demonstrate that the 2-bp substitution in the VDRE-like sequence of OG1 and OG2 promoters nearly abolished the binding of the hormone-VDR complex to this sequence and explain the absence of up-regulation of OG1 and OG2 promoter-luciferase activities by 1,25-(OH)2D3 treatment.
1,25-(OH)2D3 Inhibits Expression of the Mouse Osteocalcin GenesWe then analyzed the regulation of
expression of the endogenous osteocalcin genes by
1,25-(OH)2D3. We first used mouse primary osteoblast cultures as an experimental system. These cells were cultured in the absence or presence of
1,25-(OH)2D3 (107,
10
8, or 10
9 M) for 36, 48, or
60 h and assayed for Osteocalcin expression by Northern
blot analysis. Surprisingly, osteocalcin mRNA was readily
detectable in untreated cells, but was nearly undetectable following
1,25-(OH)2D3 treatment of these cultures. This
inhibition of Osteocalcin expression was
concentration-dependent and occurred at every time point
analyzed (Fig. 3A). In contrast, as
previously shown (33), Osteocalcin expression in ROS 17/2.8
cells was barely detectable before 1,25-(OH)2D3
treatment, but was up-regulated up to 7-fold by the hormonal treatment
(Fig. 3B). To demonstrate that this down-regulation of the
mouse Osteocalcin expression was not due to our culture
conditions, we monitored the expression of Osteopontin, a
gene whose expression is known to be increased by
1,25-(OH)2D3 treatment in mouse (34).
Consistent with our DNA transfection experiments,
Osteopontin expression was always up-regulated by
1,25-(OH)2D3 treatment of these cultures (Fig. 3A).
-Actin expression was used as an internal control in
these experiments since it is not affected by
1,25-(OH)2D3 treatment (27). To show that the
down-regulation of Osteocalcin in
1,25-(OH)2D3-treated primary osteoblasts was an
accurate reflection of the regulation occurring in vivo, we
treated 9-week-old mice with a single dose of
1,25-(OH)2D3 (24 ng), isolated RNA from
calvariae 24 h later, and measured Osteocalcin
expression. In the entire animal, 1,25-(OH)2D3 also led to a marked decrease (6-fold) in osteocalcin mRNA
accumulation (Fig. 3C). Taken together, our results show
that in mouse, 1,25-(OH)2D3 inhibits
Osteocalcin expression.
To determine whether both osteocalcin genes were down-regulated by
1,25-(OH)2D3 treatment, we used an
allele-specific oligonucleotide hybridization of the reverse
transcription-PCR product assay (19). RNA from untreated or
1,25-(OH)2D3-treated mouse primary osteoblasts was used as template. Aliquots of each PCR mixture were electrophoresed on an agarose gel along with aliquots of plasmids containing either OG1 or OG2. The gel was hybridized with a labeled
"OG1" oligonucleotide that recognizes only transcripts originating
from OG1 (19). OG1 expression was virtually
abolished by 1,25-(OH)2D3 (107
M) treatment of these cells (Fig. 4). This
gel was stripped and rehybridized with the "OG2" oligonucleotide
designed to visualize cDNAs originating only from OG2
(19). OG2 expression was also virtually abolished by
1,25-(OH)2D3 treatment of these cultures (Fig.
4). Thus, the expression of the two mouse osteocalcin genes is
down-regulated by 1,25-(OH)2D3
in vivo.
1,25-(OH)2D3 Inhibits the Activity of Long OG1 and OG2 Promoter Fragments
Given this inhibition of
OG1 and OG2 expression by
1,25-(OH)2D3, we then searched for VDREs
located farther upstream in the promoters of these genes. First, we
sequenced 2.9 kb of the OG2 promoter and 3.3 kb of the
OG1 promoter, but no VDRE consensus sequence (15) could be
identified (data not shown). Next, we generated promoter-luciferase
constructs containing the largest OG2 promoter fragment
isolated (2.9 kb) or 3.3 kb of the OG1 promoter. Each of
these constructs was transfected into ROS 17/2.8 or MC3T3-E1 cells,
which were subsequently treated with
1,25-(OH)2D3 or vehicle. We used three
different positive controls in this experiment: pROP-luc, pROC-luc, and
pMOP-luc. The activity of the three positive control constructs in
treated cells was always several times that in untreated cells.
The activity of the OG1 and OG2
promoter-luciferase constructs was constantly reduced by
approximately half by the same treatment in both cell lines (Fig.
5).
1,25-(OH)2D3 Treatment of Primary Osteoblasts Abolishes Binding of OSF2 to OSE2
The moderate effect
observed in DNA transfection experiments suggested that
1,25-(OH)2D3 might act indirectly on
Osteocalcin transcription, possibly by inhibiting the
expression of a gene that encodes a factor controlling
Osteocalcin expression. Although no such gene has been
identified yet, osteoblast-specific nuclear proteins binding to
osteoblast-specific cis-acting elements have been
characterized (20, 21, 35). To determine whether
1,25-(OH)2D3 modulated binding of these
factors, we monitored the amount of binding to two well characterized
cis-acting elements in nuclear extracts of untreated and
1,25-(OH)2D3-treated mouse primary osteoblasts. OSE2 is an osteoblast-specific cis-acting element that binds
OSF2 (osteoblast-pecific
actor
), an osteoblast-specific nuclear protein. OSE2 controls
the osteoblast-specific expression of OG2 (20, 21). mOCE1
binds a ubiquitously expressed protein and does not affect
OG2 promoter activity significantly (20). When we performed
an electrophoretic mobility shift assay using as a probe a labeled OSE2
oligonucleotide, OSF2 binding was nearly abolished in
1,25-(OH)2D3-treated nuclear extracts (Fig.
6A, lane 3). Two lines of evidence
indicate that this effect was specific. First, it could not be obtained
by using nuclear extracts of cells treated with another differentiation
agent of the osteoblasts such as BMP7 (Fig. 6A, lane
2); second, the binding of nuclear proteins to mOCE1 was
unaffected by treatment of the cells with 1,25-(OH)2D3 (Fig. 6B, lane
3). This latter result demonstrated the integrity of the nuclear
extracts. Likewise, 1,25-(OH)2D3 treatment of
primary osteoblasts did not affect the binding of nuclear proteins to
OSE1 (20), another osteoblast-specific cis-acting element
(data not shown). The specific inhibition of binding of OSF2 to OSE2 in
1,25-(OH)2D3-treated nuclear extracts suggests that the regulation of Osteocalcin expression by
1,25-(OH)2D3 occurs through an indirect
mechanism: the hormone either down-regulates OSF2 expression
or induces post-translational modification, preventing binding of OSF2
to DNA.
1,25-(OH)2D3 Treatment of Osteoblastic Cell Lines Does Not Decrease the Activity of the OG2 Promoter Containing a Mutated OSE2 Sequence
As shown in Fig. 1, treatment of ROS 17/2.8
or MC3T3-E1 cells with 1,25-(OH)2D3
down-regulated the activity of the 147-bp OG2 promoter
fragment that contains the originally described OSE2 sequence (20). To
demonstrate the importance of this sequence in the inhibition of
OG2 promoter activity, we introduced in this OG2
promoter fragment a 2-bp substitution mutation in OSE2 that abolishes
OSF2 binding (20). We then performed DNA transfection experiments using constructs containing either the wild-type or mutated
promoters fused to luciferase in ROS 17/2.8 cells or
MC3T3-E1 cells. In agreement with our DNA binding data, the
activity of the mutated promoter was not down-regulated by
1,25-(OH)2D3 treatment of these cells, but
rather slightly up-regulated (Fig. 7).
We have analyzed the regulation of the mouse osteocalcin genes by 1,25-(OH)2D3 and showed that their expression is down-regulated by this hormone. Our results also indicate that this effect is mediated, at least in part, through an indirect mechanism inasmuch as 1,25-(OH)2D3 treatment of osteoblasts abolished OSF2 binding to OSE2, a potent osteoblast-specific cis-activator of transcription (20, 21).
One of the roles of 1,25-(OH)2D3 in bone is to promote osteoblast differentiation, although the overall physiological consequences of this differentiating action are not clearly understood (1). 1,25-(OH)2D3 regulates the expression of several genes expressed in osteoblasts, among them Osteocalcin (8, 9, 10, 11, 12). This regulation of expression occurs at the transcriptional level, wherein the hormone-receptor complex binds to a VDRE located in the promoter of the rat and human osteocalcin genes (9, 10, 11). However, a growing body of evidence indicates that once the osteoblasts begin to express Osteocalcin, the effect of 1,25-(OH)2D3 on gene expression can vary with the time at which the hormonal treatment is initiated. For instance, when 1,25-(OH)2D3 treatment of rat calvaria cultures was initiated during the proliferative stage of this culture, there was no increase in Osteocalcin expression (36). These data suggest that the hormone may have multiple and complex effects in vivo on osteoblast gene expression varying with the stage of differentiation of the cells.
Several experimental arguments suggest that the inhibitory effect of 1,25-(OH)2D3 on mouse Osteocalcin expression was not due to the stage of differentiation of the osteoblasts under our culture conditions. First, we did not begin 1,25-(OH)2D3 treatment before the osteoblasts reached the mineralizing stage. Second, the expression of Osteopontin, a gene expressed earlier than Osteocalcin during osteoblast differentiation and which was used in these experiments as a positive control (7), was always up-regulated in our cultured osteoblasts. Third and most important, the same inhibition of Osteocalcin expression occurred when we performed this treatment in the entire animal.
What is the mechanism by which 1,25-(OH)2D3
inhibits mouse Osteocalcin expression? Based on our results,
we propose that the hormone achieves this function through an indirect
mechanism. 1,25-(OH)2D3 treatment of mouse
primary osteoblasts selectively abolished the binding of OSF2, an
osteoblast-specific transcription factor that binds to and controls the
activity of OG1 and OG2 promoters (20,
21).2 Our working hypothesis is that
1,25-(OH)2D3 treatment of primary osteoblasts
inhibits transcription of the gene coding for OSF2. Such an indirect
mechanism may also explain why we observed only a 2-fold effect when we
tested the effect of 1,25-(OH)2D3 treatment on
OG1 and OG2 promoter activity in DNA transfection
experiments. Indeed, it is known that other cis-acting
elements are involved in the control of Osteocalcin
expression (20, 35), and the binding of nuclear factors to these
elements was not affected by the hormonal treatment. The fact that the
inhibitory effect of 1,25-(OH)2D3 treatment was
greater on longer promoter fragments suggests that other OSE2-related
sequences could be present farther upstream in OG1 and
OG2 promoters. We (21) and others (37) recently presented
evidence that OSF2 may be a member of the PEBP2 family of
transcription factors. The eventual identification of a cDNA for
OSF2 will allow a direct demonstration of the proposed mechanism.
Broess et al. (38) reported recently that
1,25-(OH)2D3 inhibits Osteocalcin
expression in differentiated osteoblasts from chicken; it will be
important to determine whether 1,25-(OH)2D3 also inhibits OSF2 binding to the promoter of the chicken osteocalcin gene.
This hypothesis does not exclude the possibility that
1,25-(OH)2D3 affects Osteocalcin
expression directly. The hormone-VDR complex may bind to an as yet
unidentified VDRE and, through its interaction with specific cofactors,
inhibit its transcription. This mechanism has been shown recently to
explain the ligand-independent repression of transcription by the
thyroid hormone receptor (39). Although most of the VDREs identified to
date are located within 1 kb of 5-flanking sequence of the target gene
(17), we cannot exclude the possibility that the VDRE is located
farther upstream or downstream in the Osteocalcin locus, as
has been demonstrated for the regulation of the HoxA cluster
by retinoic acid (40). Experiments in progress are designed to test
these two hypotheses, which are not mutually exclusive.
A more general question raised by our studies is, what is the importance of the regulation of Osteocalcin expression by 1,25-(OH)2D3 during osteoblast differentiation in vivo? Since this hormone differently regulates the expression of the same gene in closely related species such as rat and mouse, its role is not likely to be essential for the establishment of the osteoblastic phenotype. Indeed, we have observed in osteocalcin-deficient mice that the osteoblasts do differentiate normally and can lay down an extracellular matrix and mineralize it (13).
We are grateful to Drs. M. Demay, C. Farach-Carson, H. Kronenberg, A. Ridall, S. Rodan, K. Sampath, and M. Uskokovic for providing reagents and plasmids. We thank Dr. L. Etkin and members of the laboratory for critical reading of the manuscript.