| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
and ß1
Department of Biochemistry and Molecular Biology, University of Louisville School of Medicine, Louisville, Kentucky 40292
Address all correspondence and requests for reprints to: Carolyn M. Klinge, Department of Biochemistry and Molecular Biology, University of Louisville School of Medicine, Louisville, Kentucky 40292. E-mail carolyn.klinge{at}louisville.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ß (ER and ERß). Resveratrol,
trans-3,5,4'-trihydroxystilbene, is a phytoestrogen in
grapes that is present in red wine. Resveratrol was shown to bind ER in
cytosolic extracts from MCF-7 and rat uteri. However, the contribution
of ER vs. ERß in this binding is unknown. Here we
report that resveratrol binds ERß and ER with comparable affinity,
but with 7,000-fold lower affinity than estradiol (E2).
Thus, resveratrol differs from other phytoestrogens that bind ERß
with higher affinity than ER. Resveratrol acts as an estrogen
agonist and stimulates ERE-driven reporter gene activity in CHO-K1
cells expressing either ER or ERß. The estrogen agonist activity
of resveratrol depends on the ERE sequence and the type of ER.
Resveratrol-liganded ERß has higher transcriptional activity than
E2-liganded ERß at a single palindromic ERE. This
indicates that those tissues that uniquely express ERß or that
express higher levels of ERß than ER may be more sensitive to
resveratrol’s estrogen agonist activity. For the natural, imperfect
EREs from the human c-fos, pS2, and progesterone
receptor (PR) genes, resveratrol shows activity comparable to that
induced by E2. We report that resveratrol exhibits
E2 antagonist activity for ER with select EREs. In
contrast, resveratrol shows no E2 antagonist activity with
ERß. These data indicate that resveratrol differentially affects the
transcriptional activity of ER and ERß in an ERE
sequence-dependent manner. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Resveratrol is a stilbene that exists as cis- and
trans-isomers. The trans-isomer appears to have
greater anticancer and cardio-protective properties than the
cis-isomer (9). Resveratrol has been
characterized as a phytoestrogen based on its ability to bind to and
activate estrogen receptor (ER) (10). ER is a nuclear
steroid receptor that binds estrogens and regulates the transcription
of estrogen-responsive genes by either binding directly to DNA, at
particular sequences called estrogen response elements (EREs), or by
interacting with other transcription factors, e.g. Sp1
(11), bound to their cognate sites on DNA. When activated
by an agonist ligand, ER interacts with coactivators,
e.g. SRC-1 and CBP, that either acetylate lysine residues in
histones to alter chromatin conformation and/or interact with
components of the RNA polymerase II initiation complex to enhance
target gene transcription (12).
There are two known ER subtypes, ER and the more recently identified
ERß (8). Because ER and ERß exhibit different
patterns of tissue distribution and have select differences in
biochemical properties (13), it is important to determine
which form mediates the effects of resveratrol. While both ER isoforms
bind E2 with comparable affinity, some
phytoestrogens, e.g. genistein and coumestrol, show higher
affinity for ERß than ER (5), suggesting that
resveratrol may show selectivity for ERß. Resveratrol was shown to
compete with [125I]E2 for
binding ER in an extract from MCF-7 human breast cancer cells with an
IC50 value of 
10 µM
(10). More recently, resveratrol was reported to compete
with [3H]E2 for binding
to rat uterine ER with an IC50 value of 100
µM (14). Because MCF-7 cells and
uterus reportedly express ERß as well as ER (15, 16),
it is important to determine the affinity of resveratrol interaction
with ERß and with ER.
Ligands that bind ER can act as agonists, antagonists, or mixed
agonist/antagonists. The archetype mixed agonist/antagonist is
tamoxifen (TAM), used clinically to prevent breast cancer promotion and
recurrence. TAM has both estrogen agonist and antagonist activity
depending on the cell type and gene promoter (reviewed in Ref.
17). Resveratrol showed estrogen agonist activity in MCF-7
cells, i.e. activating the expression of progesterone
receptor (PR) and pS2 genes (2). However, resveratrol
elicited only weak agonist activity in both a yeast hER
transcription assay and in transient transfections with ER in COS-1
cells. In another study, resveratrol antagonized
E2-stimulated growth and inhibited transcription
of PR in MCF-7 cells (18). In animal studies, oral
administration of resveratrol to weanling rats had no significant
effect on estrogenic responses such as serum cholesterol or messenger
RNA (mRNA) for insulin-like growth factor I, but gave a slight increase
in uterine wet weight (19). The same study showed that
resveratrol antagonized the effect of E2 on serum
cholesterol (19). Thus, the relative estrogen
agonist/antagonist activity of resveratrol remains to be
determined.
This study examined the relative agonist/antagonist activity of
resveratrol in defined assays using ER and ERß. We examined
whether resveratrol preferentially binds ER or ERß, the effect of
resveratrol on the proliferation of cells expressing ER or ERß,
how resveratrol impacts ER-ERE binding in vitro, and how
resveratrol affects the expression of reporter gene activity from a
consensus and naturally occurring EREs from estrogen- responsive genes
in transiently transfected CHO-K1 cells. Our results show that
resveratrol binds ER and ERß with comparable affinity, but with
much lower affinity than E2. Resveratrol-occupied
ER and ERß bind an ERE in vitro, but resveratrol
inhibits ER-ERE binding in a concentration-dependent manner.
Resveratrol exhibits agonist activity in transiently transfected cells
using a variety of ERE-driven reporter constructs and shows differences
in activity depending on the ERE sequence and on which ER is expressed.
With ER, but not ERß, resveratrol shows E2
antagonist activity from certain EREs, including a consensus ERE. Thus,
the mechanism of action of resveratrol is unique for ER and
ERß.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Plasmid preparation
The sequences of EREs used are: EREc38:
5'-CCAGGTCAGAGTGACCTGAGCTAAAATAACACATT-3';
PR1148:
5'-AGCCCTCCCTCCTGCGAGGTCACCAGCTCTTGGTGCCTGTTT-3';
pS2:
5'-CTTCCCCCTGCAAGGTCAGCGTGGCCACCCCGTGAGCCACT-3';
and Fos-1211:
5'-AGCTTGGGCTGAGCCGGCAGCGTGACCCCGCATG-3'.
The underlined nucleotides correspond to the minimal
core consensus ERE. The nucleotides in bold indicate an
alteration in the consensus ERE. EREc38, PR-1148, pS2, and
Fos-1211 were cloned into the pGL3-promoter luciferase reporter
vector (Promega Corp., Madison, WI) as described
(20). A mammalian expression vector containing the
sequence for recombinant human (rh) rhER was generously supplied by
Dr. Benita Katzenellenbogen (21). An expression vector
containing the sequence for recombinant rat (rr) ERß was generously
provided by Dr. J.-A. Gustafsson (22).
Nuclear extract preparation
Baculovirus-expressed rhER and rrERß were prepared as
nuclear extracts (NE) from Sf-21 cells as previously described
(23, 24).
Competition binding experiments
Aliquots of NE from Sf-21 cells containing 6 nM
(monomer) rhER or rrERß were incubated in TEGK100 buffer (40
mM Tris-HCl (pH 7.4), 1 mM EDTA, 10% (vol/vol)
glycerol, 100 mM KCl, 0.5 mM PMSF, 2 µg/ml
aprotinin, 0.5 µg/ml leupeptin, 0.75 µg/ml pepstatin, 1
mM DTT) with 15.5 nM
[3H]E2 (NEN Life Science Products, Boston, MA) and the indicated concentrations
of trans-resveratrol (generously provided by Pharma Science,
Montréal, Québec, Canada or purchased from
Sigma, St. Louis, MO), or E2
(Sigma). Unless specifically stated, all references to
resveratrol in this manuscript indicate trans-resveratrol
(see Fig. 1
). NE isolated form
Sf-21 cells expressing ER or ERß were used as the source of
receptors and nonspecific ligand binding was determined using a NE
Sf-21 cells expressing alkaline phosphatase. Nonspecific binding varied
between 7.6 and 15.5%. ER-bound and unbound
[3H]E2 were separated by
hydroxyapatite (HAP) (Bio-Rad Laboratories, Inc.,
Hercules, CA) (25). Radioactivity in the HAP pellet was
counted in a liquid scintillation counter (Wallac, Inc.
1409, Turku, Finland). Specific
[3H]E2 binding was
calculated, graphed, and analyzed using a GraphPad Software, Inc. prism (San Diego, CA). Ki was
estimated by the formula described in (26).
|
Electrophoretic mobility shift assay (EMSA)
An EREc38 oligomer, 77 nucleotides in length, which includes
sequences flanking the ERE between the EcoRI and
HindIII sites in pGEM-7Zf(+), which do not bind ER
(27), was prepared and fill-in labeled with
[32P]-dATP (800 Ci/mmol) (NEN Life Science Products) using Klenow large fragment DNA polymerase I
(New England Biolabs, Inc., Beverly, MA)
(20). Unincorporated nucleotides were removed by
centrifugation through a Centri-Spin 20 column (Princeton Separations, Adelphia, NJ). Labeled EREc38 (50,000 cpm) was
incubated for 2 h at 4 C with a nuclear extract of
baculovirus-expressed rhER or rrERß. Binding reactions were
performed in TDPEKG buffer (40 mM Tris-HCl (pH
7.5), 1 mM DTT, 0.5 mM
PMSF, 1 mM EDTA, 111 mM
KCl, 10% vol/vol glycerol) and included 5 µg
poly(deoxyinosine-deoxycytidine) (Midland Certified Reagent Co.,
Midland, TX), 10 µg purified BSA (New England Biolabs, Inc.)/reaction, in a total reaction volume of 30 µl, with a
final salt concentration of 92 mM KCl. An
ER-specific antibody H222, generously provided by Abbott Laboratories (Abbott Park, IL), was diluted 1:10 in TE buffer
(10 mM Tris-HCl (pH 8.0), 1
mM EDTA) and 1 µl of the diluted H222 was added
to selected samples in each experiment to confirm the identity of ER
protein in the shifted ER-ERE complexes. ERß-specific antibodies
PA1–310 (Affinity BioReagents, Inc., Golden, CO) and Y-19
(Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were used
undiluted in select samples to confirm the identity of ERß protein in
shifted ER-ERE complexes. After incubation, the protein-DNA mixture was
loaded onto 4% nondenaturing polyacrylamide gels and electrophoresed
as described (20). Gels were dried under vacuum and
autoradiographed on Kodak X-Omat film with an intensifying
screen (Lightning Plus from DuPont Co., Wilmington, DE).
The ratio of ER-bound to free DNA was determined using a Packard
Instruments InstantImager and associated software, Packard Imager for
Windows v 2.04 (Packard Instrument Co., Meriden, CT).
Transient transfection experiments
CHO-K1 cells were plated in 12-well plates at 2 x
105 cells/well with IMDM (-phenol red)
supplemented with 10% charcoal-stripped CS. The cells were transfected
with 0.6 µg reporter construct containing the ERE, 0.1 µg pCMV
ß-gal, 10 ng pCMV-ER or pCMV-ERß, and 0.49 µg pGEM-7Zf(+)
(Promega Corp., Madison, WI) when 80% confluent. The
transient transfection was performed using Transfast (Promega Corp.) according to directions supplied by Promega Corp. Cells were treated, in triplicate, 24 h later with
resveratrol, E2 (Sigma), or
4-hydroxytamoxifen (4-OHT) (Research Biochemicals International, Natick, MA) diluted in phenol-red-free IMDM (-).
The cells were harvested 30 h later, and the luciferase and
ß-galactosidase (ß-gal) activities were assayed (20).
All data for transient transfections were normalized by ß-gal to
account for transfection efficiency. Statistical analyses were
performed using Student’s t test in Microsoft Corp. Excel ‘97.
Cell proliferation
Cell proliferation was determined using the Cell Proliferation
Kit 1 (MTT) according to the directions provided by the supplier
(Roche Molecular Biochemicals, Indianapolis, IN). Briefly,
2 x 104 CHO-K1 cells/well were plated in
96-well plates and treated as described above for the transient
transfection experiments. Mock transfected cells were incubated with
Transfast without added plasmid DNA and then treated with hormones or
vehicle as described above. Cell proliferation experiments were
conducted concomitantly with transient transfection experiments. Cells
were transfected with 60 ng 1EREc38, 10 ng ß-gal, 1 ng rhER or
rrERß, and 49 ng of pGEM-7Zf(+). Cells cotransfected with ER or
ERß and treated with 0.1 nM E2, 50
µM resveratrol, 0.1 nM
E2 and 50 µM resveratrol, 100
µM resveratrol, 1 µM 4-OHT, 1
µM 4-OHT and 1 nM E2, 1
µM 4-OHT and 50 µM resveratrol, and EtOH
for control. At the time that the transfection was harvested, the MTT
assay was performed according to manufacturer’s instructions. The
absorbance of solubilized crystals was measured at 595 nm in a
Molecular Devices SpectraMAX250 plate reader. The means
from three separate experiments were analyzed using Student’s
t test for two samples assuming unequal variances in
GraphPadPrism.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ERß with comparable affinity vs. ERß for this binding is unknown.
Competition binding experiments were used to determine the relative
binding affinity of resveratrol for ER and ERß. Resveratrol binds
to rhER (Fig. 2A) and rrERß (Fig. 2B). The relative binding affinities (RBA) of resveratrol for ER and
ERß are not statistically different (Table 1). The equilibrium dissociation
constants (KI) are different, reflecting higher
affinity of E2 binding to ER than ERß (Table 1). However, since the 95% confidence intervals for
KI overlap, the difference in
KI for resveratrol between ERß and ER is not
statistically significant. This is the first demonstration of
resveratrol interaction with ERß.
|
|
and ERß with lower
affinity than E2 (Table 1
). The estimated
IC50 values for resveratrol of 58.5
µM and 130 µM for ER and ERß,
respectively, are in agreement with the value of 100 µM
reported for ER in MCF-7 cell extract (14). Together,
these data indicate that resveratrol binds ER and ERß with
comparable binding affinity.
Resveratrol has no effect on CHO-K1 cell proliferation unless cells
are cotransfected with ER or ERß
Resveratrol has been reported to inhibit the proliferation of ER
positive and negative cultured human breast cancer cells (10, 28). To determine the effect of resveratrol on CHO-K1 cell
proliferation in a receptor-isoform-dependent assay, cells were
transfected with expression vectors for either ER or ERß. The
proliferation of untransfected, "mock-transfected," and ER or
ERß transfected CHO-K1 cells treated with EtOH,
E2, or resveratrol was determined using the MTT
assay (Fig. 3 and data not shown). The
mock transfected cells showed no alteration in proliferation regardless
of cell treatment, indicating that TransFast is not toxic to the cells
(data not shown). Cells transfected with ER or ERß showed no
alteration in cell proliferation with or without treatment with
E2 or 4-OHT (Fig. 3 and data not shown). However,
cells transfected with ER or ERß and treated with 100
µM resveratrol showed decreased proliferation.
Untransfected or mock-transfected cells showed no decrease in
proliferation with 100 µM resveratrol treatment,
indicating the effect is dependent upon ER expression (Fig. 3 and data
not shown). Similarly, cells transfected with pCMV-ß-gal and treated
100 µM resveratrol show no decrease in cell
proliferation. Cells treated with E2 showed no
decrease in viability. These data indicate that 100 µM
resveratrol decreases cell proliferation only when the CHO-K1 cells are
transfected with ER or ERß.
|
or
ERß with an ERE, ER or ERß were incubated with
E2, 4-OHT, or resveratrol and ERE binding was
measured by EMSA (Fig. 4). ER bound
specifically to EREc38, as indicated by the supershift of the entire
bound complex with ER-specific antibody H222 (Fig. 4A). The multiple
ER bands are due to truncated ER present in the NE. As
anticipated from previous reports (29, 30, 31), preincubation
of ER with E2 accelerated the migration of the
ER-EREc38 complex, while addition of 4-OHT slowed the migration of
the 4-OHT-ER-EREc38 complex (compare Fig. 4, lanes 2, 3, and 4).
Resveratrol had no apparent effect on the migration of the ER-EREc38
complex. These data indicate differences in ER conformation in the
presence of these ligands. Likewise, resveratrol did not affect the
binding of H222 to ER-EREc38, generating a supershifted complex
similar to that of the E2-ER-EREc38
complexed with H222 (Fig. 4, lanes 6 and 1).
|
, lanes 1, 4, and 5).
Resveratrol had no apparent effect on the migration of the ERß-EREc38
complex (compare Fig. 4B, lanes 1 and 6). These data indicate
differences in ERß conformation in the presence of these ligands.
Resveratrol did not affect the interaction of antibody PA1–310 with
the resveratrol-ERß-EREc38 complex (Fig. 4B, lane 7).
While the concentration of resveratrol added was expected to fully
occupy ER or ERß, based on the data shown in Fig. 2, we cannot
exclude the possibility that the receptor was not fully saturated with
resveratrol in this experiment. Because resveratrol was recently
reported to inhibit the binding of the arylhydrocarbon receptor
(AHR)/AHR-nuclear translocator (ARNT) heterodimer to a xenobiotic
response element in EMSA (32), we examined the effect of
increasing concentrations of resveratrol and E2
on ER-ERE binding in vitro using EMSA (Fig. 5 and data not shown). Quantitation of
the amount of ER-EREc38 or ERß-EREc38 complexes shows that
resveratrol inhibits ER-ERE binding at concentrations > 10
µM, but that identical concentrations of
E2 resulted in less inhibition of ER-EREc38
binding, until the final concentration reached 200
µM (Fig. 5C). This inhibition of ER-ERE binding
by resveratrol can also be observed by noting the increased amount of
free EREc38 at the bottom of the gel (Fig. 5, A and B). Resveratrol had
a more pronounced inhibitory effect on ERß-EREc38 binding compared
with ER-EREc38 binding. Like resveratrol, E2
had more of an inhibitory effect on ERß-EREc38 binding compared with
ER-EREc38 binding. This observation indicates that ERß-EREc38
binding is more labile than ER-EREc38 binding.
|
from a
consensus ERE and
ERß. This cell line was selected because it requires exogenous ER
or ERß to activate ERE-driven reporter gene expression and thus
allows evaluation of the transcriptional response of each ER isoform in
isolation with each ERE. First, we examined the effect of resveratrol
on reporter gene activity driven by one copy of a consensus ERE called
EREc38 (sequence in Materials and Methods) with ER.
Resveratrol stimulated luciferase activity in a concentration-dependent
manner up to 50 µM, but 100
µM resveratrol inhibited luciferase activity
(Fig. 6A). Because results from the MTT
assay indicate that 100 µM resveratrol
decreased cell proliferation in ER-expressing cells (Fig. 3), this
is likely the explanation for the decreased reporter activity at this
resveratrol concentration. E2, at concentrations
of 1 or 10 nM, has higher agonist activity with
ER than 50 µM resveratrol. Resveratrol at
concentrations of 1, 10, or 50 µM had no effect
on basal luciferase activity from the pGL3-pro-luciferase parental
vector (data not shown). Addition of 100 nM 4-OHT
blocked resveratrol activity with ER (Fig. 6A and data not shown).
While the fold induction of E2-stimulated
luciferase detected is low, we note that others have reported that
single EREs show significantly lower levels of transcriptional
activation in response to E2 compared with
commonly used multiple tandem EREs (24, 33, 34). These
results indicate that the estrogen agonist activity of resveratrol is
mediated by ER-ERE interaction in transiently transfected CHO-K1
cells.
|
from natural,
imperfect EREs binds these EREs with reduced
affinity compared with binding to EREc38 (Refs. 36, 37)
and data not shown). Therefore, to determine whether resveratrol
activates ER-mediated reporter gene activity from imperfect EREs, we
performed transient transfection assays using luciferase reporter
plasmids containing the imperfect EREs and their native flanking
regions from the human c-fos, pS2, and PR genes (Fig. 6B, sequences in Materials and Methods).
E2 did not induce as much luciferase activity
from the natural, imperfect EREs from the human pS2 and PR genes as
from the consensus EREc38 (compare Fig. 6, A and B). However, the
luciferase activity induced by E2 from the
imperfect ERE from the human c-fos gene was similar than
that induced from EREc38. While there was a significant increase in
luciferase activity from pS2 when cells were treated with 10
nM E2, no significant
dose-response relationship was detected for induction from Fos-1211 or
PR1148. As seen for EREc38, resveratrol induced significantly lower
luciferase expression from Fos-1211 and pS2 compared with
E2. The luciferase activity induced by 50
µM resveratrol from PR1148 was comparable to
that induced by 1 nM
E2.
The induction of luciferase activity from the natural EREs by
resveratrol was blocked by cotreatment with 4-OHT, indicating that
ER is responsible for resveratrol-induced reporter activity. As seen
for EREc38, treatment of the CHO-K1 cells with 100 µM
resveratrol inhibited luciferase activity from the pS2, FOS-1211, and
PR1148 EREs (data not shown). In conclusion, the data from these
transient transfection assays indicate that resveratrol acts as an
estrogen agonist with ER. These results are similar to those
detected in transiently transfected, ER-expressing COS-1 cells with
either a vitellogenin ERE or LH-ß promoter-luciferase reporter
plasmid (14).
Resveratrol induces reporter gene activity with ERß from a
consensus ERE
ERß has been shown, in transient transfection assays using a
single or multiple tandem copies of a consensus ERE, to have lower
activity in response to E2 than ER
(31). However, in COS-1 cells, ERß induced higher
reporter activity from the vitellogenin ERE than ER in response to
concentrations of E2 ranging from 0.01–1
µM (14). Here, we observed that
cotransfection of CHO-K1 cells with ERß and EREc38 generated lower
luciferase expression in response to E2 compared
with ER (Fig. 6A). Comparable protein expression levels of ER and
ERß were achieved in these cells (Western blot data not shown). ERß
induced 61% of the activity of ER with 1 nM
E2. These data are consistent with the 59, 52,
and 62% of ERß activity relative to ER detected using reporters
bearing one or 3 tandem consensus EREs in CEF, HeLa, and HepG2 cells,
respectively (31, 38).
With ERß, resveratrol stimulated luciferase expression from EREc38 in
a concentration-dependent manner up to 50 µM, whereas 100
µM resveratrol inhibited luciferase activity. These
results are similar to those reported for ERß with a vitellogenin ERE
reporter in transiently transfected COS-1 cells (14). In
CHO-K1 cells, resveratrol-stimulated activity was inhibited by
cotreatment with 4-OHT (Fig. 6A), which also blocked
E2-induced activity from ERß (data not shown),
indicating that direct interaction of resveratrol with ERß is
responsible for the induction of luciferase activity from EREc38.
Interestingly, in contrast to the differences in luciferase activity
induced by E2 with ER and ERß, resveratrol
induced nearly identical levels of luciferase activity from EREc38 with
either ER or ERß (Fig. 6A). This indicates that
resveratrol-liganded ERß has similar transcriptional activity to
E2-ERß at a single perfect, palindromic ERE.
These data differ from those for ERß expression in COS-1 cells in
which 500 µM resveratrol showed higher induction of
reporter activity from two tandem copies of the vitellogenin ERE than
any of the concentrations of E2 (0.01–1
mM) examined (14). These findings indicate
that cell-specific factors influence the agonist activity of
resveratrol with ERß.
Resveratrol induces reporter gene activity with ERß from natural,
imperfect EREs
Next we examined the induction of luciferase activity from the
EREs from the human c-fos, pS2, and PR genes with ERß
(Fig. 6B). Please note that the scale for fold-induction of luciferase
activity for revised Fig. 6C is one half the scale used for Fig. 6, A
and B. As seen for ER, E2 induced lower
activity from each of the imperfect EREs than from EREc38 (compare Fig. 6, A and C). Unlike ER for which E2 stimulated
more activity from Fos-1211 than the other natural EREs, there was no
difference in the luciferase activity induced by ERß with 10
nM E2 from the imperfect
EREs. These data indicate that ER and ERß transactivate reporter
gene expression differentially in response to E2
from natural imperfect EREs in CHO-K1 cells.
Resveratrol stimulated ERß-driven reporter activity from each
natural-occurring ERE in a concentration-dependent manner, although the
response with PR1148 was not statistically different between
resveratrol concentrations. For both ER and ERß, PR1148 was least
responsive to resveratrol. As anticipated, the luciferase activity from
the three imperfect EREs was lower than that induced from EREc38
(compare Fig. 6, A and C). However, for Fos-1211 and pS2, the
luciferase activity induced by 50 µM resveratrol was
greater than that stimulated by 1 or 10 nM
E2. As seen with ER, treatment of the
ERß-transfected CHO-K1 cells with 100 µM resveratrol
inhibited luciferase activity from all EREs due to decreased CHO-K1
cell proliferation (Fig. 3 and data not shown). The induction of
luciferase activity from each of the natural EREs by resveratrol with
ERß was blocked by cotreatment with 4-OHT, indicating that ERß is
responsible for resveratrol-induced reporter activity.
At 50 µM, resveratrol induced identical levels of
reporter activity from EREc38, Fos-1211, and pS2 with ER and ERß.
However, E2 induced higher reporter activity from
all EREs with ER than ERß. For PR1148, ERß was less active than
ER with resveratrol and E2. Taken together,
these results imply that resveratrol-liganded ER and ERß are
equivalently transcriptionally active with the Fos-1211 and pS2 EREs.
In contrast, E2-liganded ERß interacts with
EREc38 and the imperfect EREs from the human c-fos, pS2, and
PR genes less productively than ER. Finally, these data indicate
that the ER agonist activity of resveratrol is not identical for ER
and ERß and varies with ERE sequence.
Resveratrol antagonizes E2 activity at select EREs
with ER
Resveratrol was reported to display "superagonist" activity in
MCF-7 cells transfected with an ERE-driven luciferase reporter plasmid,
i.e. greater activity than with 0.1
nM E2 treatment (9, 10).
This phenomenon is not understood because these two reports show
different definitions of superagonist activity. In the first report
(10), "superagonist" activity was not the equivalent
of synergy since the responses of E2 and
resveratrol were additive (10). In the more recent report,
the reporter activity detected in MCF-7 cells treated with 0.1
nM E2 plus 25 µM
resveratrol was 2-fold greater than the anticipated additive activity
of the two ligands (9). We evaluated how resveratrol
impacted E2-stimulated luciferase activity from
EREc38 or the natural EREs from the human c-fos, pS2, and PR
genes in CHO-K1 cells expressing either ER (Fig. 7A) or ERß (Fig. 7B). No additive
activity was detected with ER and 0.1 nM
E2. Interestingly, 50 µM
resveratrol had E2 antagonist activity for EREc38
and PR1148, but not Fos-1211 or pS2. Thus, variations in the ERE
sequence appear to influence the agonist/antagonist activity of
resveratrol with ER in E2-treated CHO-K1
cells. We note that the relatively low fold induction that we observed
with single EREs is similar to that reported by other investigators
(33, 34).
|
, with ERß the luciferase activity stimulated by
1 nM E2 from EREc38 and PR1148 was
not antagonized by cotreatment with 50 µM resveratrol. As
seen for ER, resveratrol had no effect on
E2-induced reporter activity from pS2. The
results for Fos-1211 were similar for ER and ERß: increased
activity with E2 plus resveratrol. In no case was
the activity of E2 and resveratrol additive. The
activity induced by E2 plus resveratrol for each
ERE with ERß was blocked by 4-OHT, indicating that it is dependent on
ERß AF-2 (data not shown).
To examine the relative agonist/antagonist activity with a stronger ERE
promoter, we evaluated the induction of luciferase activity from two
tandem, head-to-tail, copies of EREc38, called 2(EREc38), with 10
nM E2, resveratrol, or
E2 + resveratrol in CHO-K1 cells transfected
with ER or ERß (Fig. 8). As seen
with the single EREs (Figs. 6 and 7), ER had higher transcriptional
activity than ERß in response to E2.
Resveratrol showed weak agonist activity for both ER and ERß.
Similar to the data for a single copy of EREc38, resveratrol suppressed
E2-ER-stimulated luciferase activity from
2(EREc38) in a concentration-dependent manner. In contrast, resveratrol
did not inhibit E2-ERß activity. This result
indicates that ER and ERß respond differently to resveratrol at
the same ERE. Resveratrol, when combined with E2
exhibits antagonist activity with ER, but no antagonist activity
with ERß.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Resveratrol has chemopreventive and chemotherapeutic activities
(1, 2) and has been classified as a phytoestrogen because
it binds to ER with low affinity (10). Here we
evaluated the estrogen agonist/antagonist activity of resveratrol with
ER and ERß in vitro and in transiently transfected
cells. Our results show that resveratrol binds ER and ERß with
comparable affinity, but with much lower affinity than
E2. This finding contrasts with data showing that
several phytoestrogens bind ERß with higher affinity than ER
(42). One possible explanation for the lower affinity of
resveratrol binding to ERß compared with other phytoestrogens is its
structural similarity with diethylstilbestrol (DES) (10),
which binds ERß with lower affinity than ER (43).
Whether resveratrol-occupied ER has agonist or antagonist activity
has been controversial (10, 14, 18). Resveratrol was first
reported to be a relatively weak ER ligand but showed
"superagonist" activity, i.e. higher reporter activity
than E2, in ER-expressing MCF-7 cells
(9, 10). However, another report found no evidence of
superagonism in COS cells with either ER or ERß (14).
Data from our transient transfection assays in CHO-K1 cells using a
consensus ERE or the natural imperfect EREs from the human
c-fos, pS2, or PR genes indicate that resveratrol acts as an
estrogen agonist with ER and ERß. However, resveratrol does not
display superagonist activity either alone or in combination with
E2. These results are similar to those detected
in COS-1 cells transfected with ER and either a vitellogenin ERE or
LH-ß promoter-luciferase reporter plasmid (14).
Our results demonstrate that the agonist activity of resveratrol with
ERß is fundamentally different from E2 agonist
activity because, in contrast to E2 which induces
higher activity of ER than ERß, resveratrol activated equal
reporter activity from EREc38 with both ER and ERß. In contrast,
500 µM resveratrol was reported to induce higher
ERE-driven reporter activity by ERß than any concentration of
E2 (0.01–1 mM) examined in COS-1
cells (14). Because both we and Ashby et al.
(14) used human ER and rat ERß expression vectors and
pGL3-luciferase vectors in our experiments, we conclude that the
difference between our findings and those reported by Ashy et
al. (14) is likely due to differences in the
expression of coregulators between CHO-K1 and COS-1 cells. Importantly,
our data indicate that resveratrol-liganded ERß has higher
transcriptional activity than E2- liganded ERß
at a single palindromic ERE. This indicates that those tissues that
uniquely express ERß or that express more ERß than ER may be
more sensitive to resveratrol’s estrogen agonist activity.
Biochemical (29, 30, 44, 45) and crystal structure
(46, 47) studies indicate ligand-specific differences in
ER conformation that impact interaction with coactivators. We
observed clear differences in the migration of ERE-bound ER and
ERß either unoccupied or occupied by E2, 4-OHT,
or resveratrol. These data indicate differences in ER conformation in
the presence of resveratrol compared with E2 or
4-OHT. The concentrations at which resveratrol inhibited ERE binding by
ER and ERß in vitro are concentrations at which
resveratrol exhibited agonist activity in transiently transfected
cells. Further experiments, e.g. in vivo DNase I
footprinting, are needed to determine in vivo effects of
resveratrol on ER-ERE binding. The crystal structure of the ERß LBD
occupied by the phytoestrogen genistein showed that helix 12 did not
adopt the distinctive "agonist" position seen with
E2 binding to the ER LBD, but lay in an
orientation similar to that observed when the ERß LBD was occupied by
the select estrogen receptor modulator (SERM) raloxifene
(48). The authors concluded that this positioning of the
transactivational helix 12 was consistent with the partial agonist
activity of genistein (48). Given the pharmacological
activities of resveratrol observed here, we predict that
resveratrol-bound ERß LBD may show a structure similar to that of the
genistein-occupied ERß. Because alterations in LBD conformation
impact the interaction of ER and ERß with coactivators, further
experiments are needed to assess coactivator effects on
resveratrol-liganded ERß activity.
While resveratrol reportedly gave a dose-dependent increase in reporter
activity with concentrations as high as 500 µM with both
ER and ERß in transiently transfected COS cells (14),
we observed that 100 µM resveratrol inhibited ERE-driven
reporter activity in CHO-K1 cells. Because the limit of resveratrol
solubility is 250 µM in 5% EtOH, we are uncertain as to
the soluble concentration of resveratrol in the COS cell assays
(14). Moreover, we observed decreases cell proliferation
in CHO-K1 cells expressing ER or ERß and treated with 100
µM resveratrol. There was no significant decrease in
CHO-K1 cell proliferation in cells treated with 100 µM
resveratrol and expressing ß-galactosidase or in mock-transfected
cells under the same conditions. While others have reported that
E2 treatment of cells stably overexpressing ER
inhibits cell proliferation (49, 50, 51, 52, 53),
E2 at concentrations that stimulate reporter
activity in transiently transfected CHO-K1 cells had no effect on cell
proliferation in the presence or absence of transfected ER or ERß.
Therefore, stable expression of ER in ER negative cells appears
to be required for E2-induced inhibition of cell
proliferation in response to E2, but not
resveratrol. We conclude that expression of ER or ERß is involved
in the decrease in CHO-K1 cell proliferation with 100 µM
resveratrol. Others reported that 100 µM resveratrol
inhibited the growth of ER-expressing MCF-7 cells (18).
However, the mechanism for inhibition of MCF-7 cell proliferation may
not be ER-mediated since resveratrol also inhibited the growth of
ER-negative breast cancer cells (28).
In addition to its agonist activity, resveratrol exhibited
antiestrogenic activity in MCF-7 cells (18). Resveratrol
decreased the levels of transcription of PR, insulin-like growth
factor-receptor, and transforming growth factor- genes and
stimulated the expression of transforming growth factor-ß2, results
similar to those elicited by tamoxifen in these cells
(18). While the authors concluded that the most likely
mechanism for the antiestrogenic effect of resveratrol is its direct
competition with E2 for ER binding, they also
suggested that resveratrol might prevent ER binding to EREs
(18). Our data support both suggestions because we
observed that resveratrol competes with E2 for
ER and ERß binding and inhibited ER and ERß binding to EREc38
in vitro. The concentrations at which resveratrol inhibited
ER-ERE binding in vitro are those at which resveratrol
exhibits agonist/antagonist activity in CHO-K1 cells.
We observed both ER isoform-specific and ERE-specific differences in
the agonist activity induced by resveratrol in CHO-K1 cells. For
example with ER, the activity induced by E2
was greater than that stimulated by any concentration of resveratrol
for EREc38, Fos-1211, and pS2, but not PR1148. In contrast, resveratrol
and E2 were equally transcriptionally active with
ERß at all EREs tested. With ER and PR1148, although the induction
levels are low, they are significantly above the ethanol control
values, and resveratrol-induced luciferase activity was comparable to
that induced by E2. This result implies that,
when bound to the PR1148 ERE, resveratrol-liganded and
E2-liganded ER and ERß are equally effective
at recruiting components of the coactivator/RNA polymerase II
preinitiation complex.
In addition to estrogen agonist activity, we also report that
resveratrol has estrogen antagonist activity in CHO-K1 cells. However,
resveratrol’s antagonist activity was only observed with ER and not
ERß. These data are reminiscent of the lack of 4-OHT agonist activity
with ERß (54). We speculate that the antagonist activity
of resveratrol may be mediated by AF-1, which appears to be absent in
ERß. Interestingly, the antagonist activity of resveratrol was only
observed with EREc38, whether as a single or two tandem copies, and
PR1148. These data indicate that the ERE alters the pharmacological
properties of resveratrol mediated by ER. This result agrees with
our postulate that the ERE sequence acts as an allosteric modulator of
ER activity (29). Further experiments are needed to define
exactly what regions of ER and ERß are necessary for resveratrol
agonist and antagonist activity and to define the exact ERE sequence
requirements for resveratrol antagonist activity with ER.
In addition to its direct ER binding, resveratrol has many
non-ER-mediated cellular activities that may influence ER
transcriptional activation through "cross-talk" mechanisms. For
example, resveratrol induces phosphorylation of the mitogen-activated
protein (MAP) kinase family members, extracellular regulated kinase 1
(ERK1), and ERK2, in neuroblastoma SH-SY5Y cells (55)
Activation of the MAP kinase pathway has been shown to activate
unliganded ER through phosphorylation of serine 118 in AF-1
(56). However, this ligand-independent pathway does not
appear to be important in our experiments because 4-OHT blocked both
E2 and resveratrol activity, indicating that
direct activation of ER through the LBD was responsible for the
activity reported here. Moreover, because the transcriptional effects
of resveratrol vary with alterations of the ERE sequence,
resveratrol’s activity appears to be mediated by direct interaction of
the resveratrol-occupied ER with EREs.
While the pharmacokinetics of resveratrol metabolism have not yet been
examined in humans, results from rodent studies indicate that two
servings of red wine may provide two-digit micromolar serum
concentrations of resveratrol (1), i.e.
concentrations identical to those at which the pharmacological
activities of resveratrol were observed here as well as reported by
others (10, 14, 18). The cell-type and ERE-sequence
dependence of the transcriptional activity of resveratrol with ER
and ERß may be related to cell-specific differences in the activity
of enzymes that modulate ER function, e.g. protein kinases,
and in the expression of coactivator or corepressor proteins. Continued
analysis of ER and ERß interaction with estrogenic ligands,
estrogen-regulated genes, and coregulator proteins is necessary to gain
a better understanding of how these receptors regulate estrogenic
activity at the cellular and molecular level and whether the anticancer
and cardioprotective activities of resveratrol are mediated by ER.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received May 9, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ß is cell type- and
promoter-dependent. J Biol Chem 274:32008–32014
