Content uploaded by Toshio Nikaido
Author content
All content in this area was uploaded by Toshio Nikaido on Mar 26, 2014
Content may be subject to copyright.
Estrogen-induced proliferation of normal endometrial glandular cells
is initiated by transcriptional activation of cyclin D1 via binding of c-Jun
to an AP-1 sequence
Tanri Shiozawa*
,1
, Tsutomu Miyamoto
1
, Hiroyasu Kashima
1
, Kohzo Nakayama
2
,
Toshio Nikaido
3
and Ikuo Konishi
1
1
Department of Obstetrics and Gynecology, Shinshu University School of Medicine, 3-1-1 Asahi, Matsumoto 390-8621, Japan;
2
Department of 2nd Anatomy, Shinshu University School of Medicine. 3-1-1 Asahi, Matsumoto 390-8621, Japan;
3
Department
of Organ Regeneration, Shinshu University Graduate School of Medicine, 3-1-1 Asahi, Matsumoto 390-8621, Japan
To explore the mechanism of estrogen-induced growth of
normal endometrium, the transactivation system of the
cyclin D1 gene was analysed using cultured normal
endometrial glandular cells. Estradiol (E2) treatment of
cultured normal endometrial glandular cells induced
upregulation of c-Jun, and then cyclin D1 proteins,
followed by serial expressions of cyclins E, A and B1
proteins. Increase in the mRNA expression of cyclin D1
preceded the protein expression of cyclin D1 under E2
treatment. A luciferase assay using deletion constructs of
the cyclin D1 promoter indicated that E2-induced increase
in transcriptional activity was observed in reporters
containing AP-1-binding site sequence, and that in the
absence of E2, cotransfection of c-Jun also showed
increase of transcriptional activity in the same reporters
with AP-1 sequence. A gel shift assay using nuclear
extract from E2-treated endometrial glandular cells and
AP-1 sequences of the cyclin D1 promoter indicated
specific binding between c-Jun protein and the promoter.
Transfection of c-jun antisense oligonucleotides to the
glandular cells resulted in the suppression of the E2-
induced upregulation of cyclin D1 mRNA and protein.
These findings suggest that E2-induced proliferation of
normal endometrial glandular cells is initiated by
transcriptional activation of cyclin D1 via binding of c-
Jun to the AP-1 sequences.
Oncogene (2004) 23, 8603–8610. doi:10.1038/sj.onc.1207849
Published online 27 September 2004
Keywords: endometrium; endometrial carcinoma; estro-
gen; cyclin; AP-1
Introduction
Proliferation of normal endometrial glandular cells and
endometrial carcinoma cells expressing estrogen recep-
tors (ER) is known to be increased by estrogen (Strauss
and Coutifaris, 1999). To understand the estrogen-
induced growth mechanism of the normal and malig-
nant endometrium is of particular importance for the
prevention of endometrial neoplasia in women under
unopposed estrogen environment, and also for the
development of better hormonal treatment of endome-
trial carcinoma, the most common malignancy in the
female genital tract (Lurain, 2002). It is also true for
appropriate stimulation of endometrial growth in
women with infertility, since the number of infertile
couples especially due to implantation failure is increas-
ing in recent years (Speroff et al., 1999). Although the
mechanisms for estrogen-dependent growth of these
cells have previously been investigated from various
aspects, including of upregulation of steroid receptors
(Gronroos et al., 1987; Lessey et al., 1988), altered
function of steroid metabolizing enzymes (Tseng and
Mazzella, 1980), autocrine or paracrine effect of growth
factors and cytokines (Giudice, 1994), very little is
known about the direct effect of estrogen on the
machineries propelling the cell cycle.
Cell cycle regulators such as cyclins, cyclin-depen-
dent-kinases (cdks) and cdk inhibitors are essential in
cell cycle control, that is, the cyclins form a complex
with the respective cdks, and the kinase activities
generated by these complexes phosphorylate the retino-
blastoma gene products (pRb), resulting in cell cycle
progression (Nurse, 1994; Sherr, 1994). We previously
reported the expression of these cell cycle regulators in
normal cycling endometria, and the overexpression of
various classes of cyclins and cdks in endometrial
carcinomas (Li et al., 1996; Nikaido et al., 1996;
Shiozawa et al., 1996, 1997). In addition, we demon-
strated that progestins upregulate the expression of a
cdk2 inhibitor, p27Kip1 (p27), via post-translational
mechanism, and that p27 plays a crucial role in the
progestin-induced growth suppression of normal and
malignant endometrial glandular cells (Shiozawa et al.,
2001a). On the other hand, the intracellular molecules
mediating between estrogen and cell cycle regulators in
endometrial tissues remain undetermined. The possible
mechanisms for estrogen-induced activation of cyclins
have been proposed from the studies using breast cancer
Received 30 June 2003; revised 29 March 2004; accepted 22 April 2004;
published online 27 September 2004
*Correspondence: T Shiozawa; E-mail: tanri@hsp.md.shinshu-u.ac.jp
Oncogene (2004) 23, 8603–8610
&
2004 Nature Publishing Group
All rights reserved 0950-9232/04 $30.00
www.nature.com/onc
cell lines (Altucci et al., 1996; Lee et al., 1998; Sabbah
et al., 1999; Castro-Rivera et al., 2001; Liu et al., 2002a);
the transcription factors involved have been character-
ized, although the results in breast cancer cells are
controversial probably due to the differences of cell
context. Therefore, the present study was conducted to
identify the molecules directly involved in estrogen-
induced progression of the cell cycle in normal and
neoplastic endometrial glandular cells.
Results
Serial upregulation of cyclins and cdks under treatment
with estrogen in normal endometrial glandular cells
To analyse the involvement of cell cycle regulators in the
growth of normal endometrial glands, we first examined
the immunohistochemical expression of cyclin and cdks
in human endometrial tissues. The expression of cyclin
D1, E, A and B1 was sporadically observed in the nuclei
of glandular cells in the proliferative phase (Figure 1a).
Cyclin D1- and cyclin E-positive cells were often positive
for cdk4 and cdk2, respectively (Figure 1b). Cyclin
A- and cyclin B1-positive cells were also positive for
cdc2. In addition, these cyclin/cdk-positive cells ex-
pressed Ki-67, suggesting that cyclins and cdks are
functionally involved in the growth regulation of
endometrial glandular cells.
We then examined the expression of cyclins and cdks
in primarily cultured endometrial glandular cells from
normal proliferative endometrium. These cells were
positive for ER-aand cytokeratin (data not shown).
Uptake of [
3
H]thymidine was increased by treatment
with 17b-estradiol (E2) in a dose-dependent manner
(Figure 2). The difference in the uptake between the
control and E2 (10
6
M)-treated group reached a
significant difference with a relative ratio 1.4570.20
(P¼0.01) after 72 h of incubation. On the other hand,
E2-induced [
3
H]thymidine uptake was suppressed by
treatment with progesterone (P4) at 10
8
Mafter 48 and
72 h. Cell cycle analysis using flow cytometry showed
that the S-phase fraction of E2-untreared cells was
13.2%, and those of E2-treated cells for 12 and 24 h
were 18.2 and 20.0%, respectively. These findings
indicate that our in vitro model using the cultured
endometrial glandular cells is suitable for investigation
of estrogen-induced cell proliferation.
Expression of cell cycle regulators under E2 treatment
was analysed using these cultured glandular cells.
Western blot analysis demonstrated that expression of
cyclin D1 was first observed 4 h after the E2 (10
6
M)
treatment, followed by serial expressions of cyclin E,
cyclin A and cyclin B1 (Figure 3a). pRb phosphoryla-
tion was observed after 24 h (Figure 3a, upper arrow).
The expression of cdk4, cdk2 and cdc2 was also
increased by E2 treatment. Analyses using semiquanti-
tative reverse transcriptase–polymerase chain reaction
Figure 1 Immunohistochemical expression of cyclin D1 and c-Jun
is observed in the proliferative phase of normal endometrium.
Immunostaining for cyclin D1 (a) and cdk4 (b) in the serial sections
of proliferative phase endometria showed that cyclin D1-positive
cells were also positive for cdk4 ( 150). Immunostaining for c-Jun
was observed in the proliferative (c), but not in the secretory (d)
endometria ( 150)
Figure 2 E2 treatment increases [
3
H]thymidine uptake in cultured
endometrial glandular cells. Cultured normal endometrial gland-
ular cells were incubated with E2 (10
8
M,10
6
M)orE2
(10
6
M)þP4 (10
8
M) for 24, 48 and 72 h. [
3
H]thymidine incorpora-
tion assay showed that cell growth was stimulated by E2 treatment
in a dose-dependent fashion, and inhibited by P4 after 48 and 72 h.
Columns indicate the mean7standard deviation (s.d.)
Estrogen-induced cyclin D1 expression in human endometrium
T Shiozawa et al
8604
Oncogene
(RT–PCR) revealed that the mRNA expression of cyclin
D1 appeared by 2 h, and then increased with a peak at
12 h after E2 treatment (Figure 3b). Regarding the
expression of cyclin E mRNA, it appeared by 4 h and
increased until 24 h after E2 treatment (Figure 3b). The
expression of both cyclin D1 and cyclin E mRNAs
preceded that of cyclin D1 and cyclin E proteins,
respectively. Immunoprecipitation experiments using
cell lysates from E2-stimulated cultured glandular cells
confirmed the complex formation of cyclin D1 with
cdk4, and that of cyclin E with cdk2 (Figure 3c). To
further examine the function of these complexes, we
performed a histone H1 kinase assay. Kinase activity
was observed when cdk4 and cdk2 were immunopreci-
pitated in E2-treated cells, but not in E2-untreated cells
(Figure 3c). These findings indicate that E2 treatment
serially upregulates cyclins and cdks, and that these
events constitute the initial step in E2-induced growth of
normal endometrial glandular cells.
Estrogen-induced activation of cyclin D1 via binding
of c-Jun with the AP-1 sequence of the promoter
We then investigated the molecular aspects of estrogen-
induced upregulation of cyclin D1. Screening of the
cyclin D1 promoter revealed that it lacks an estrogen-
response element (ERE), but contains an AP-1-binding
site sequence located at 953 to 947 (Motokura and
Arnold, 1993). Promoters of Jun and Fos are known to
have functional EREs (Weisz and Rosales, 1990; Hyder
et al., 1995), and expressions of c-Jun and c-Fos have
reportedly been induced by estrogen in animal models
(Webb et al., 1993). Our immunohistochemical analysis
in human endometrial tissues revealed that expressions
of c-Jun in the glandular cells was observed only in the
proliferative phase (Figures 1c and d), but c-Fos
expression was observed in both the proliferative and
secretory phases. Western blot analysis using cultured
glandular cells showed an induction of c-Jun expression
2 h after E2 treatment, which preceded the expression of
cyclin D1 (Figure 3a). Slight expression of c-Fos was
also observed after E2 treatment. Hence, we hypothe-
sized that either c-Jun, c-Fos, or both are involved in the
upregulation of cyclin D1 in E2-induced growth of
endometrial glandular cells.
To address this hypothesis, we analysed the cyclin D1
transcription in cultured endometrial glandular cells by
a luciferase assay using plasmids containing various
sizes of deletion constructs ranging from 1793 to 43
of the cyclin D1 promoter (Figure 4a). The results
indicated that, without E2 treatment, the transcriptional
activity was low in most of the promoter constructs. In
contrast, under E2 treatment, there were two reporters
showing elevated activities with significant differences,
being consistent with the E2-induced upregulation of
cyclin D1 mRNA. And, both of these two reporters
contained the AP-1 sequence, suggesting that E2-
induced expression of cyclin D1 is mediated by the
AP-1 sequence of the promoter. To further analyse
whether c-Jun or c-Fos or both were functionally
involved in the transcriptional activation of cyclin D1,
we performed a luciferase assay in cultured glandular
Figure 3 E2 treatment induces serial upregulation of c-Jun, cyclins and cdks in cultured endometrial glandular cells. In Western blot
analyses (a), expression of cyclin D1 was first observed 4 h after E2 treatment, followed by serial expressions of cyclin E, cyclin A and
cyclin B1. The expression of cdk4 and cdk2 also showed an increase after E2 treatment. pRb phosphorylation was observed 24 h after
E2 treatment (upper arrow). Upregulation of c-Jun and slight increase in c-Fos expression were observed 2 h after E2 stimulation. In
semiquantitative RT–PCR for cyclin D1 and cyclin E (b), mRNA expression of cyclin D1 appeared by 2 h, and then increased with a
peak at 12 h after E2 treatment. Cyclin E mRNA appeared by 4 h and increased until 24 h after E2 treatment. In immunoprecipitation
and histone H1 kinase assays (c), complex formation of cyclin D1–cdk4 and cyclin E–cdk2 was observed, and these complexes showed
histone H1 kinase activities in E2-treated endometrial glandular cells
Estrogen-induced cyclin D1 expression in human endometrium
T Shiozawa et al
8605
Oncogene
cells, using the same set of plasmids containing various
sizes of cyclin D1 promoter deletion constructs, along
with cotransfection of either a c-Jun- or a c-Fos-
expression plasmid in the absence of E2. Transcriptional
activity in the two reporters containing the AP-1
sequence was increased when c-Jun was cotransfected
with significant differences, but not when c-Fos was
cotransfected (Figure 4b).
To evaluate the specificity of binding of the AP-1
sequence with the AP-1 protein from the endometrial
glandular cells, we performed a gel shift assay and a
competition assay. In this study, three different DNA
probes were used, that is, PRAD-D1, promoter DNA
containing an AP-1 sequence in the cyclin D1 promoter
(TGAGTCA): PRAD-M, a mutant of the AP-1
sequence and PRAD-C; a consensus AP-1-binding
sequence (TGACTCA). The gel shift assay indicated
that a specific band shift was observed when E2-treated
nuclear extract was applied with PRAD-D1 and PRAD-
C (Figure 5a; lanes 4 and 6). In contrast, the bands were
very weak when a nuclear extract from E2-untreated
cells was applied with PRAD-D1 and PRAD-C
(Figure 5a; lanes 1 and 3). Specific bands were not
detected when PRAD-M was applied (Figure 5a; lanes 2
and 5). The competition assay showed that an excessive
amount of cold PRAD-D1 completely eradicated the
specific band (Figure 5a; lane 10). Addition of cold
PRAD-C also eradicated most of the above specific
bands (Figure 5a; lane 12). In contrast, addition of cold
PRAD-M did not affect the bands (Figure 5a; lane 11).
All of these findings indicate that the protein extracted
from E2-treated endometrial glandular cells binds
specifically to the AP-1 sequence of the cyclin D1
promoter. In the gel shift assay, band supershift was not
observed when anti-c-Jun or anti-c-Fos antibodies used
for immunostaining were added (Figure 5a; lanes 8 and
9). However, when anti-AP-1 antibodies designed for
the binding inhibition assay were applied, band super-
shift was observed with the anti-c-Jun antibody
(Figure 5b; lane 2). These findings indicate that c-Jun
protein specifically binds to the AP-1 sequence of the
promoter. It is likely therefore that E2-induced activa-
tion of cyclin D1 expression is mediated by binding of
c-Jun with the AP-1 sequence of the cyclin D1 promoter
in normal endometrial glandular cells.
Finally, to confirm the functional involvement of c-
jun in E2-induced upregulation of cyclin D1 in normal
endometrial glandular cells, we performed c-jun silen-
cing experiments using antisense oligonucleotides. E2
treatment induced upregulation of both c-jun and cyclin
D1 mRNAs and the respective proteins (Figure 6, lane
2). Transfection of c-jun antisense oligonucleotides
resulted in the suppression of the mRNA expression of
c-jun and cyclin D1 as well as of the protein expression
of c-Jun and cyclin D1 (Figure 6, lane 3). Transfection
of c-jun sense oligonucleotides did not markedly affect
their mRNA and protein expressions (Figure 6, lane 4).
These findings support our hypothesis that c-jun is
actually involved in the E2-induced upexpression of
cyclin D1 in the normal endometrial glandular cells.
Figure 4 AP-1 site and c-Jun are involved in estradiol-induced upregulation of cyclin D1 gene. In luciferase assays, reporter plasmids
containing various sizes of deletion constructs were transfected into normal endometrial glandular cells, with or without E2 treatment
(a), and elevated activities were observed in the two reporters (nos. 1 and 2) containing the AP-1 sequence under E2 treatment. In
luciferase assays using the same reporters with cotransfection of c-Jun or c-Fos in the absence of E2 (b), elevated activities were
observed in the same two reporters containing AP-1 site (nos. 1 and 2), only when cotransfected with c-Jun but not with c-Fos. The
structure of the cyclin D1 promoter deletion constructs fused to the luciferase reporter gene is indicated below the graphs. *P¼0.028
when compared to the activity using the same reporter without E2. **P¼0.031 when compared to the activity using the same reporter
without E2. #Po0.01 when compared to the activity using the same reporter with control vector or with c-Fos. ##Po0.001 when
compared to the activity using the same reporter with control vector or with c-Fos
Estrogen-induced cyclin D1 expression in human endometrium
T Shiozawa et al
8606
Oncogene
Estrogen-induced activation of cyclin D1 in endometrial
carcinoma Ishikawa cells without involvement of the AP-1
sequence
We also examined the effect of E2 treatment on the
expression of cyclins in endometrial carcinoma Ishikawa
cells. Western blot analysis showed that expression of
cyclin D1 was noted even before E2 treatment, and a
slight increase of cyclin D1 expression was observed
after E2 treatment. A luciferase assay using the same set
of plasmids containing various sizes of cyclin D1
promotor deletion constructs revealed that E2-induced
increase in the activity was most prominent in the
reporter containing the E2F-binding site, but was not
observed in the reporters having the AP-1 sequence
(data not shown). These findings suggest that the
mechanism of E2-induced activation of cyclin D1 in
endometrial carcinoma Ishikawa cells is different from
that of normal endometrial glandular cells.
Discussion
In this study, to explore the molecular mechanism of
estrogen-induced cell proliferation, we analysed the
transcriptional activation system of the cyclin D1 gene.
Among various cyclins, cyclin D1 is the only molecule
whose expression is controlled by extracellular signals
(Sherr, 1994). In our primarily cultured glandular cells
from normal human endometrium, expression of cyclin
D1 protein appeared 4 h after E2 treatment, followed by
serial expressions of cyclins E, A and B1. Such
sequential expression of cyclins is the same as that in
other cells stimulated by various mitogens, and expres-
sion of cyclin D1 protein was reportedly increased
approximately 4–6 h after the stimulation. Therefore,
upregulation of cyclin D1 seems to be crucial also in the
estrogen-induced growth of endometrial glandular cells.
The activation of the cyclin D1 gene is predominantly by
transcription (Sherr, 1994), and its promoter contains
Figure 5 Direct binding of c-Jun to AP-1 is shown by gel shift mobility and competition assays. Nuclear extracts from the E2-treated
and untreated normal endometrial glandular cells were applied with three radiolabeled probes: D1 (PRAD-D1), promoter DNA
containing an AP-1 sequence in the cyclin D1 promoter (TGAGTCA); M (PRAD-M), a mutant of the AP-1 sequence; C (PRAD-C), a
consensus AP-1-binding sequence (TGACTCA). Specific band shift (a; arrow) was observed when E2-treated nuclear extract was
applied with PRAD-D1 (lane 4) and PRAD-C (lane 6), but not with PRAD-M (lane 5). The bands were very weak when E2-untreated
extract was applied with PRAD-D1 (lane 1) and PRAD-C (lane 3), and not detected when PRAD-M was applied (lane 2). The
competition assay showed that an excessive amount of cold PRAD-D1 completely eradicated the specific band (lane 10), and cold
PRAD-C also eradicated most of the above specific bands (lane 12). In contrast, cold PRAD-M did not affect the bands (lane 11). The
band supershift was not observed when anti-c-Jun or anti-c-Fos antibodies used for immunostaining were added (a; lanes 8 and 9).
However, when anti-AP-1 antibodies designed for the binding inhibition assay were applied, band supershift was observed with the
anti-c-Jun antibody (b; arrowhead, lane 2)
Figure 6 Treatment with c-jun antisense oligonucleotides sup-
presses mRNA and protein expressions of c-jun and cyclin D1.
RT–PCR (upper) and Western blot (lower) analyses showed that
E2-induced increase of mRNA and protein expressions of c-jun
and cyclin D1 (lane 2) was suppressed by transfection of c-jun
antisene oligonucleotides to cultured endometrial glandular cells
(lane 3). On the other hand, transfection of c-jun sense
oligonucleotides did not affect the E2-induced upexpression of
c-iun and cyclin D1 (lane 4)
Estrogen-induced cyclin D1 expression in human endometrium
T Shiozawa et al
8607
Oncogene
multiple regulatory elements such as TRE, E2F, Oct,
Sp1 and a cAMP response element (Motokura and
Arnold, 1993; Herber et al., 1994). Previous promoter
analyses suggested that the assembly of transcription
factors is highly variable and dependent on multiple
factors including the mitogen and cell context. Over-
expression of p60v-src in MCF-7 breast cancer cells
activates cyclin D1 and involves activation of a cAMP-
response element-binding protein (CREB) and an
activating transcription factor-2 (ATF-2), which inter-
acts with a CRE at 66 in the cyclin D1 promoter (Lee
et al., 1998). The enhanced expression of cyclin D1 by
TGF-ais mediated by early growth response protein
(Egr-1) via a cis-regulatory region spanning nucleotides
144 to 104 of the cyclin D1 promoter (Yan et al.,
1997). Expression of p21ras and p300 activated the
cyclin D1 gene promoter in JEG-3 human trophoblasts
through interaction of proteins at a distal AP-1 sequence
at 954 in the promoter (Albanese et al., 1995; Albanese
et al., 1999). However, the promoter of cyclin D1 lacks
an ERE, and the mechanism for estrogen-induced
activation of cyclin D1 remains to be elucidated.
The present study demonstrated that estrogen-in-
duced activation of the cyclin D1 gene is mediated by
binding of c-Jun to the AP-1 sequence of the cyclin D1
promoter. A luciferase assay using plasmids containing
various sizes of deletion constructs indicated that the
AP-1 sequence of the promoter is essential in E2-induced
activation. After E2 treatment, expression of c-Jun
preceded the cyclin D1 expression, and cotransfection of
c-Jun and a gel shift mobility assay indicated the specific
binding of the AP-1 sequence with c-Jun, but not with c-
Fos. Generally, the effect of AP-1 transcription factors,
such as c-Jun and c-Fos, encompasses a variety of
reactions to the target gene activation (Curran and Franza,
1988). AP-1 mediates basal level enhancer function and
transcriptional activation in response to serum or phorbol
ester (Ryseck et al., 1988). c-Jun is important in promoting
progression of the cell cycle, including the DNA synthesis
ortheSphase(Riabowolet al., 1992). On the other hand,
AP-1 is also a negative trans-acting factor for several genes
including c-Fos, MyoD (Li et al., 1992), insulin (Inagaki
et al., 1992) and human chorionic gonadotropin (hCG)
(Pestell et al., 1994). Both c-jun and c-fos have functional
ERE, and E2 treatment caused the expression of c-Jun or
c-Fos at the mRNA and protein levels in the uteri of
various animals (Webb et al., 1993). In our study of the
human endometrium in vivo, expression of c-Jun was
observed in the proliferative, but not in the secretory phase
of the menstrual cycle. This is consistent with a previous
study for localization of c-jun mRNA in the proliferative
endometria (Salmi et al., 1998). Thus, although the
transcriptional action of c-Jun is either positive or
negative, possibly due to the combination of structure of
c-Jun and a target gene context (Baichwal et al., 1992), our
study showed that c-Jun is important as a positive
mediator for transcriptional activation of the cyclin D1
gene in estrogen-induced proliferation of endometrial
glandular cells.
Mechanisms of estrogen-induced growth of the target
cells have been studied exclusively in breast cancer cells.
Although the results from the previous studies are
controversial, several mechanisms have been proposed.
Estrogen-induced expression of c-Fos and its complex
formation with pS1 was reported in MCF-7 cells (Duan
et al., 1998), and specific inhibition of AP-1 activity was
reported to block the growth of these cells (Liu et al.,
2002b). Altucci et al. (1996) reported that estrogen-
induced expression of cyclin D1 was mediated by a 944
to 136 region of the promoter in simvastatin-treated
MCF-7 cells, and that estrogen-induced pS2 preceded
the expression of cyclin D1, suggesting the possible
involvement of the AP-1 site and of pS2 as a mediator
for cyclin D1 expression. On the other hand, estrogen-
induced expression of cyclin D1 in MCF-7 and HeLa
cells has been reported to occur through a cAMP-
responsive element located at 96 to 29 sites (Sabbah
et al., 1999; Liu et al., 2002a), and this response was
mediated by the binding of ATF-2 homodimers or ATF-
2/c-Jun heterodimers (Sabbah et al., 1999). Further-
more, they also suggested that the AP-1 sequence of the
promoter was rather inhibitory for cyclin D1 transcrip-
tion in estrogen stimulation. Another report suggested
the involvement of GC-rich Sp1-binding sites at 143 to
110 in E2-induced upregulation of cyclin D1 (Castro-
Rivera et al., 2001). Our study has proposed a
mechanism that E2-induced upregulation of cyclin D1
is mediated through binding of c-Jun to the AP-1 site of
the promoter. We presume that upregulation of c-jun
occurs transcriptionally and the following phenomena
occur sequentially, that is, E2–ER complex first binds to
the functional ERE located at the promoter of c-jun
gene, and then c-Jun binds to the AP-1 sequence of the
cyclin D1 gene. However, it should be noted that c-Jun
levels rise quickly with a peak at 4 h and then drop off,
and that cyclin D1 mRNA and protein levels do not
peak until 12 and 24 h in the present study. Therefore, it
may also be possible that there is another pathway of
AP-1 activation, that is, E2–ER acts directly at the AP-1
element by serving as a coactivator for c-Jun, as
suggested by Kushner et al. (2000).
In conclusion, our study showed that estrogen-
induced upregulation of cyclin D1 plays an important
role in the growth of normal endometrial glandular cells,
and its transcriptional activation is mediated by binding
of c-Jun to the AP-1 sequence of the promoter. In
endometrial carcinoma Ishikawa cells, our study sug-
gested that the mechanism for estrogen-induced upre-
gulation of cyclin D1 is different from that in normal
endometrial cells. However, further researches are
needed to clarify the mechanisms how ER and c-Jun
interact at the AP-1 sites in normal endometrial
glandular cells, and to elucidate the molecular pathways
for cyclin D1 activation in endometrial carcinomas.
Materials and methods
Immunohistochemistry
A total of 20 normal endometrial tissue specimens obtained
from 20 women (aged 35–45 years) were subjected to
Estrogen-induced cyclin D1 expression in human endometrium
T Shiozawa et al
8608
Oncogene
immunohistochemical staining with the approval of the Ethical
Committee of Shinshu University, Japan, after obtaining
written consent from the patients. Immunohistochemical
staining was performed using an SAB-PO detector kit
(Nichirei, Tokyo, Japan) as previously described (Shiozawa
et al., 1996). Anti-c-Jun and anti-c-Fos antibodies were
purchased from Santa Cruz Biotechnology (Santa Cruz, CA,
USA) for immunohistochemistry and from Oncogene (San
Diego, CA, USA) for the gel shift assay.
Cell culture
Endometrial carcinoma cell An Ishikawa ER-a-positive en-
dometrial carcinoma cell line was kindly provided by Dr
Nishida at Tsukuba University. Cells were cultured in
Dulbecco’s modified Eagle’s medium (DMEM) supplemented
with 15% FCS (GIBCO, Grand Island, NY, USA).
Normal glandular cells The isolation and culture of glandular
cells was performed as previously described (Shiozawa et al.,
2001a). Only primarily cultured cells were used in all
experiments. Before the E2 stimulation experiments, normal
glandular cells and Ishikawa cells were pretreated with
rapamycin (GIBCO) (500 ng/ml) in a serum-starved medium
for 24 h to synchronize them to the G1 phase of the cell cycle.
Then, the cells were cultured for 24 h without rapamycin with
the same medium to wash out the effect of the agent.
[
3
H]thymidine incorporation assay
To evaluate the growth of normal endometrial glandular cells
with estrogen, a [
3
H]thymidine incorporation assay was
performed as previously described (Shiozawa et al., 2001a).
Cell cycle analysis
The cell cycle was analysed in E2-treated normal endometrial
glandular cells and E2-treated Ishikawa cells, using a
cycleTEST (Becton Dickinson, Franklin Lakes, NJ, USA)
and FACScan flow cytometry (Becton Dickinson).
Western blotting
Expressions of the cell-cycle-related molecules in cultured normal
glandular cells after estrogen treatment were examined by
Western blotting as described previously (Shiozawa et al., 2001b).
RT–PCR
Estrogen-induced expression of cyclin D1 mRNA and cyclin E
mRNA was evaluated by semiquantitative RT–PCR as
described previously (Takano et al., 2000; Kaneuchi et al.,
2003). Glyceraldehyde-3-phosphatase dehydrogenase (GAP)
was used as a control.
c-jun silencing experiment
c-jun antisense oligonucleotides (50-CGTTTCCATCTTTG-
CAGT-30) and c-jun sense oligonucleotide (50-ACTGCAAA-
GATGGAAACG-30) (Miao and Ding, 2003) were transfected
to cultured endometrial glandular cells using the EffecteneTM
Transfection Reagent (Qiagen, Hilden, Germany), according
to the manufacturer’s instructions for 24 h after the rapamycin
treatment. After the medium was changed, E2 of 10
6
Mwas
added. The expression of c-Jun mRNA and protein and was
evaluated by semiquantitative RT–PCR (Miao and Ding,
2003) and Western blotting 3 h after addition of E2, and that
of cyclin D1 mRNA/protein was also evaluated by semiquan-
titative RT–PCR (Kaneuchi et al., 2003) and Western blotting
6 h after addition of E2.
Immunoprecipitation
Complex formation between cyclin D1 and cdk4, as well as
cyclin E and cdk2, in E2-treated (10
6
M, 24 h) and E2-
untreated normal endometrial glandular cells was examined by
immunoprecipitation as previously described (Shiozawa et al.,
2001b). The detection of associated proteins was carried out
using antibodies against cyclin D1 (Progen, Heidelberg,
Germany) and cyclin E (Santa Cruz Biotechnology).
Histone H1 kinase assay
The kinase activity of the cyclin D1/cdk4 and cyclin E/cdk2
complexes in E2-treated (10
6
M, 24 h) and E2-untreated
normal endometrial glandular cells was evaluated by the
detection of radio-labeled histone H1 protein as previously
described (Shiozawa et al., 2001b).
Luciferase assay
Plasmid cyclin D1 promoter-luciferase reporter constructs
were a kind gift from Professor AK Rutsgi, USA (Yan et al.,
1997). The constructs contained eight different sizes of deleted
cyclin D1 promoter (ranging from 1749 to þ35, Figure 4),
inserted to a promoterless vector, pA3LUC. c-Jun and c-Fos
expression vectors were kind gifts from Professor T Curren,
USA (Sonnenberg et al., 1989).
Luciferase assay Isolated human endometrial glandular cells
or Ishikawa cells (2 10
5
) were dispersed in type IV collagen-
coated 24-well plates and incubated for 24 h with Ham’s F12
þ15%FCS until DNA transfection. Then, 0.5 mg of plasmid
DNAs were transfected into the above cultured normal
glandular cells using the EffecteneTM Transfection Reagent
with or without 10
6
Mof E2. At 72 h after transfection, the
luciferase assays were performed as indicated by the manu-
facturer’s protocol (Dual reporter system, Promega, Madison,
WI, USA). Activities were normalized using the pRL-SV40
vector (Renilla luciferase, Promega), which was cotransfected
as an internal control. A 10 ml of cell lysate was assayed for
luciferase activity using a luminometer (Lumat LB9501, EG &
G Berthold, Bad Wildbad, Germany). The values were means
of three independent experiments. Statistical analyses were
made by Kruskal–Wallis test and Scheffe’s test. A tied P-value
with less than 0.05 was considered as significant.
Gel mobility shift assay and competition study
Nuclear extracts from E2-treated and untreated cultured
endometrial glandular cells were prepared by the modified
Dignam method. The gel shift assay was performed as
described previously (Kawasaki et al., 1999). In brief, three
20 base single-stranded DNA oligomers termed PRAD-D1,
corresponding to regions 958 to 940 (50-AAAAATGAGT-
CAGAATGGAG-30), which included the AP-1 sequence at
953 to 947, PRAD-M (50-AAAAACAAGTTGGAATG-
GAG-30), which contained four introduced mutation sites on
the AP-1 sequence (TGAGTCA-CAAGTTG) and PRAD-C
(50-AAAAATGACTCAGAATGGAG-30), which included a
consensus AP-1-binding domain (TGACTCA), were synthe-
sized, annealed and end-labeled with [
32
P]dCTP using the
Klenow enzyme. For the competition studies, a 10–50-fold
molar excess of the three unlabeled DNA oligomers above
was used (PRAD-D1; 10-fold, PRAD-M; 50-fold, PRAD-C;
10-fold). For the supershift assay, 0.5 mg of antibody against
Estrogen-induced cyclin D1 expression in human endometrium
T Shiozawa et al
8609
Oncogene
for c-Jun or c-Fos (Santa Cruz Biotechnology in Figure 5a,
Oncogene in Figure 5b) was incubated for 15 min at room
temperature after the binding reaction. The bands were
analysed with the MacBAS system (Fuji Film, Tokyo, Japan)
Acknowledgements
This work was supported in part by a Grant-in-aid for
Scientific Research from the Ministry of Education, Science
and Culture (No. 06454468 and No. 07807154), Japan
References
Albanese C, D’Amico M, Reutens AT, Fu M, Watanabe G,
Lee RJ, Kitsis RN, Henglein B, Avantaggiati M, Somasun-
daram K, Thimmapaya B and Pestell RG. (1999). J. Biol.
Chem.,274, 34186–34195.
Albanese C, Johnson J, Watanabe G, Eklund N, Vu D, Arnold
A and Pestell RG. (1995). J. Biol. Chem.,270, 23589–23597.
Altucci L, Addeo R, Cicatiello L, Dauvois S, Parker MG,
Truss M, Beato M, Sica V, Bresciani F and Weisz A. (1996).
Oncogene,12, 2315–2324.
Baichwal VR, Park A and Tjian R. (1992). Genes Dev.,6,
1493–1502.
Castro-Rivera E, Samudio I and Safe S. (2001). J. Biol. Chem.,
276, 30853–30861.
Curran T and Franza Jr BR. (1988). Cell,55, 395–397.
Duan R, Porter W and Safe S. (1998). Endocrinology,139,
1981–1990.
Giudice LC. (1994). Fertil. Steril.,61, 1–17.
Gronroos M, Maenpaa J, Kangas L, Erkkola R, Paul R and
Grenman S. (1987). Ann. Chir. Gynecol.,202, 76–79.
Herber B, Truss M, Beato M and Muller R. (1994). Oncogene,
9, 1295–1304.
Hyder SM, Nawaz Z, Chiappetta C, Yokoyama K and Stancel
GM. (1995). J. Biol. Chem.,270, 8506–8513.
Inagaki N, Maekawa T, Sudo T, Ishii S, Seino and Imura H.
(1992). Proc. Natl. Acad. Sci. USA,89, 1045–1049.
Kaneuchi M, Sasaki M, Tanaka Y, Sakuragi N, Fujimoto S
and Dahiya R. (2003). Int. J. Oncol.,22, 159–164.
Kawasaki S, Ebara S, Nakayama K and Takaoka K. (1999).
Biochem. Biophys. Res. Commun.,263, 560–565.
Kushner PJ, Agard DA, Greene GL, Scanlan TS, Shiau AK,
Uht RM and Webb P. (2000). J. Steroid. Biochem. Mol.
Biol.,74, 311–317.
Lee RJ, Albanese C, Stenger RJ, Watanabe J, Inghirami G,
Haines III GK, Webster M, Muller WJ, Brugge JS, Davis RJ
and Pestell RG. (1998). J. Biol. Chem.,274, 7341–7350.
Lessey BA, Killam AP, Metzger DA, Haney AF, Greene GL and
McCarty Jr KS. (1988). J. Clin. Endocrinol. Metab.,67, 334–340.
Li L, Chambard JC, Karin M and Olson EN. (1992). Genes
Dev.,6, 676–689.
Li SF, Shiozawa T, Nakayama K, Nikaido T and Fujii S.
(1996). Cancer,77, 321–328.
Liu MM, Albanese C, Anderson CM, Hilty K, Webb P, Uht
RM, Price Jr RH, Pestell RG and Kushner PJ. (2002a). J.
Biol. Chem.,277, 24353–24360.
Liu Y, Ludes-Meyers J, Zhang Y, Munoz-Medellin D, Kim
HT, Lu C, Ge G, Schiff R, Hilsenbeck SG, Osborne CK and
Brown PH. (2002b). Oncogene,21, 7680–7689.
Lurain JR. (2002). Uterine Cancer: Novak’s Gynecology, 13th
edn Berek JS (ed). Lippincott Williams & Wilkins: Philadel-
phia, pp. 1143–1147.
Miao ZH and Ding J. (2003). Cancer Res.,63, 4527–4532.
Motokura T and Arnold A. (1993). Genes Chromosomes
Cancer,7, 89–95.
Nikaido T, Li SF, Shiozawa T and Fujii S. (1996). Cancer,78,
1248–1253.
Nurse P. (1994). Cell,79, 547–550.
Pestell RG, Hollenberg AN, Albanese C and Jameson JL.
(1994). J. Biol. Chem.,269, 31090–31096.
Riabowol K, Schiff J and Gilman MZ. (1992). Proc. Natl.
Acad. Sci. USA,89, 157–161.
Ryseck RP, Hirai SI, Yaniv M and Bravo R. (1988). Nature,
334, 535–537.
Sabbah M, Courilleau D, Mester J and Redeuilh G. (1999).
Proc. Natl. Acad. Sci. USA,96, 11217–11222.
Salmi A, Heikkila P, Lintula S and Rutanen EM. (1998). J.
Clin. Endocrinol. Metab.,83, 1788–1796.
Sherr CJ. (1994). Cell,79, 551–555.
Shiozawa T, Horiuchi A, Kato K, Obinata M, Konishi I,
Fujii S and Nikaido T. (2001a). Endocrinology,142,
4182–4188.
Shiozawa T, Li SF, Nakayama K, Nikaido T and Fujii S.
(1996). Mol. Hum. Reprod.,2, 745–752.
Shiozawa T, Nikaido T, Shimizu M, Zhai YL and Fujii S.
(1997). Cancer,80, 2250–2256.
Shiozawa T, Shiohara S, Kanai M, Konishi I, Fujii S and
Nikaido T. (2001b). Cancer,92, 3005–3011.
Sonnenberg JL, Rauscher III FJ, Morgan JI and Curran T.
(1989). Science,249, 1622–1625.
Speroff L, Glass RH and Kase NG. (1999). Female Infertility:
Clinical Gynecologic Endocrinology and Infertility, 6th edn.
Mitchell C (ed). Lippincott Williams & Wilkins: Philadel-
phia, pp. 1014–1018.
Strauss III J and Coutifaris C. (1999). The Endometrium
and Myometrium. Regulation and Dysfunction: Repro-
ductive Endocrinology, 4th edn Yen SSC, Jaffe RB and
Barbieri R (eds). WB Saunders Co.: Philadelphia, pp.
218–256.
Takano Y, Kato Y, van Diest PJ, Masuda M, Mitomi H and
Okayasu I. (2000). Am. J. Pathol.,156, 585–594.
Tseng L and Mazzella J. (1980). Cyclic Changes of
Estradiol Metabolic Enzymes in Human Endometrium
During the Menstrual Cycle: The Endometrium Kimball
FA (ed). SP Medial and Scientific Books: New York,
pp. 211–226.
Webb DK, Moulton BC and Khan SA. (1993). Endocrinology,
133, 20–28.
Weisz A and Rosales R. (1990). Nucleic. Acids. Res.,18,
5097–5106.
Yan YX, Nakagawa H, Lee MH and Rustgi AK. (1997). J.
Biol. Chem.,272, 33181–33190.
Estrogen-induced cyclin D1 expression in human endometrium
T Shiozawa et al
8610
Oncogene