0013-7227/08/$15.00/0
Printed in U.S.A.
Endocrinology 149(11):5366 –5373
Copyright © 2008 by The Endocrine Society
doi: 10.1210/en.2008-0857
Low Concentrations of the Soy Phytoestrogen Genistein
Induce Proteinase Inhibitor 9 and Block Killing of
Breast Cancer Cells by Immune Cells
Xinguo Jiang, Nicole M. Patterson, Yan Ling, Jianwei Xie, William G. Helferich, and David J. Shapiro
Departments of Biochemistry (X.J., N.M.P., D.J.S.) and Food Science and Human Nutrition (Y.L., J.X., W.G.H.), University
of Illinois at Urbana-Champaign, Urbana, Illinois 61801
The risks and benefits of diets and supplements containing
the estrogenic soy isoflavone genistein are not well established. We report that 10 nM genistein potently induces the
granzyme B inhibitor, proteinase inhibitor 9 (PI-9) in MCF-7
human breast cancer cells. By inducing PI-9, genistein inhibits the ability of human natural killer (NK) cells to lyse the
target breast cancer cells. In ER␣HA cells, stably transfected
MCF-7 cells, which contain elevated levels of estrogen receptor-␣ (ER␣), 100 pM genistein or 17␤-estradiol potently induce
PI-9 and prevent NK cells from killing the target breast cancer
cells. The concentrations of genistein that fully induce PI-9 in
MCF-7 cells, and in ER␣HA cells, are far lower than those
previously reported to elicit estrogenic responses through
E
STROGENS CONTRIBUTE to the development of breast
cancers in part by stimulating the proliferation and
metastases of tumor cells (1, 2). We and others recently described a different action of estrogens likely to contribute to
tumor development, blocking the ability of immune cells to
induce apoptosis of target cancer cells (3– 6). The immune
cells, cytotoxic T lymphocytes (CTLs) and natural killer (NK)
cells, recognize and destroy early stage developing tumors in
a process sometimes referred to as immunosurveillance (7).
Human CTLs and NK cells lyse target cancer cells and virally
infected cells primarily by elaborating particles called granules containing the protease granzyme B, other granzymes,
and perforin (8, 9). In estrogen receptor- (ER␣)-containing
breast and liver cancer cells, estrogens strongly induce the
granzyme inhibitor, proteinase inhibitor 9 (SerpinB9, PI-9).
By inducing PI-9 in cancer cells, estrogens interfere with the
ability of immune system cells to induce apoptosis of human
cancer cells (3, 4). We proposed a graded response in which
expression of increasing levels of PI-9 in a target cancer cell
progressively blocks NK- and CTL-induced apoptosis
through the granzyme system (3) and also blocks apoptosis through the secondary Fas/Fas ligand apoptosis pathway (5, 10).
First Published Online July 31, 2008
Abbreviations: CTL, Cytotoxic T lymphocyte; DOX, doxycycline; E2,
17␤-estradiol; ER␣, estrogen receptor-␣; ERU, estrogen-responsive unit;
GPR30, G protein-coupled receptor 30; NK, natural killer; OHT, 4-hydroxytamoxifen; PI-9, proteinase inhibitor 9; SERM, selective ER modulator; SPI-6, serine protease inhibitor 6.
Endocrinology is published monthly by The Endocrine Society (http://
www.endo-society.org), the foremost professional society serving the
endocrine community.
ER␣. Because 4-hydroxytamoxifen, raloxifene, and ICI
182,780/Faslodex all block genistein induction of PI-9 and elevated levels of ER␣ enhance induction of PI-9, genistein acts
via ER␣ to induce PI-9. Increasing levels of ER␣ in breast
cancer cells results in a progressive increase in induction of
PI-9 by genistein and in the cell’s ability to evade killing by NK
cells. Moderate levels of dietary genistein and soy flour effectively induce PI-9 in human breast cancers grown in ovariectomized athymic mice. A significant population consumes
levels of genistein in soy products that may be high enough
to induce PI-9, perhaps potentiating the survival of some
preexisting breast cancers by enabling them to evade
immunosurveillance. (Endocrinology 149: 5366 –5373, 2008)
We considered potential sources of estrogens that might
induce PI-9 and enable breast cancer cells to evade immunosurveillance. Genistein is an estrogenic soy isoflavone that
structurally mimics endogenous 17␤-estradiol (E2) (11). In
the United States, consumption of soy-based products has
increased, perhaps because it is thought they may contribute
to the lower rate of incidence of breast cancer seen in Asian
countries where soy consumption is high. Epidemiological
studies suggest that preadolescent consumption of soy protein correlates with a decreased risk of breast cancer (12).
The effect of estrogenic soy isoflavones on the growth and
progression of breast cancer has been controversial (13). Low
concentrations of genistein stimulate the growth of ER␣containing breast cancer cells in vitro and stimulate growth
of preexisting estrogen-dependent breast cancer in mouse
xenografts (14 –17). High concentrations (⬎10 ␮m) genistein
are thought to inhibit growth of cancer cells by inhibiting
tyrosine kinases (18 –21).
In most assays, the reported estrogenicity of genistein in
cells containing ER␣ is quite low (1,000 –100,000-fold lower
than E2) (22, 23). A recent microarray study in MCF-7 cells
found that 3 ␮m genistein and 30 pm E2 produce similar
overall levels of expression of estrogen-regulated mRNAs
(23). Another recent study found that 6 nm genistein did not
induce ER␣-regulated genes in MCF-7 cells (24).
In contrast, we find that 10 nm genistein maximally induces PI-9 mRNA and protein in MCF-7 cells, thereby inhibiting NK cell-induced apoptosis. In ER␣HA cells, which
express elevated levels of ER␣ (3, 25), E2 and genistein show
similar potency, and 100 pm genistein fully induces PI-9 and
elicits a nearly complete block of NK cell-induced apoptosis.
5366
Jiang et al. • Genistein-Induced PI-9 Blocks Immunosurveillance
We also found that dietary genistein and soy flour induce
PI-9 in solid tumors in the MCF-7 mouse xenograft model.
Materials and Methods
Cell lines and cell culture
MCF-7 human breast cancer cells were cultured in MEM containing
5% calf serum and in phenol red-free MEM containing 5% charcoal
dextran-treated calf serum for at least 4 d before experiments. The
tet-inducible MCF-7 cells, ER␣HA cells, were grown as described (3).
NK92 cells and NKL cells were maintained as we recently described (3).
Endocrinology, November 2008, 149(11):5366 –5373
5367
To prepare estrogen pellets, 2 mg E2 and 18 mg cholesterol were
mixed and packed into a pellet about 4.5 mm in diameter and 2.5 mm
in depth. The E2 pellet was implanted sc in the interscapular region of
each mouse 1 wk before MCF-7 cells were injected into the animal.
American Institute of Nutrition 93 growth diet (AIN-93G) semipurified phytoestrogen-free diet (Dyets, Bethlehem, PA) was selected as a
base diet. Soy oil was removed from all diets and corn oil added to
eliminate any additional components of soy being added to the diets.
Treatment animals were fed AIN-93G diet plus genistein at 500 mg/kg
(500 ppm) or a diet containing 20% soy protein, which contains approximately 400 ppm genistein (29).
Analysis of tumor growth
Quantitative RT-PCR
PI-9 mRNA levels were analyzed by quantitative RT-PCR. Primers
were designed using Primer Express software (Applied Biosystems,
Foster City, CA). PI-9 primers were forward 5⬘-TGGAATGAACCGTTTGACGAA-3⬘ and reverse 5⬘-CATCTGCACTGGCCTTTGCT-3⬘. Actin
primers were forward 5⬘-AAGCCACCCCACTTCTCTCTAA-3⬘ and reverse 5⬘-AATGCTATCACCTCCCCTGTGT-3⬘. pS2 primers were forward 5⬘-ACCGGACACCTCAGACACG-3⬘ and reverse 5⬘-CTGTGTTGTGAGCCGAGGC-3⬘. RNA extraction and cDNA synthesis were as we
recently described (3). Briefly, quantitative RT-PCR were conducted in
96-well plates using a SYBR Green PCR kit (Applied Biosystems), and
samples were amplified with a Bio-Rad iCycler system (Bio-Rad, Hercules, CA). The fold change in expression of each gene was calculated
using the ⌬⌬Ct method with actin mRNA as the internal control.
Western blotting
Whole-cell extracts were prepared in RIPA buffer (1⫻ PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 10⫺6 m sodium orthovanadate, 10 ␮g/ml phenylmethylsulfonyl fluoride, and 30 ␮l/ml
aprotinin). For Western blotting, cell lysates containing 5 ␮g protein
were separated on 10% SDS-PAGE gels and transferred to nitrocellulose
membranes. Because actin is almost the same size as PI-9, calnexin was
used as the loading control. Blots were incubated with the appropriate
secondary antibodies, and signals were detected by PhosphorImager
analysis using ECL Plus (Amersham, Piscataway, NJ).
Cell-mediated cytotoxicity assay
The time-resolved fluorescence assay produces results similar to the
well-established chromium release assay (3, 26). The assay was carried
out following the supplier’s protocol with a few modifications. Target
cells were incubated at 106 cells/ml in phenol red-free MEM plus 5%
charcoal dextran-treated calf serum with BATDA (PerkinElmer, Wellesley, MA) for 10 min at room temperature, followed by four washes.
Cytotoxicity assays were carried out in 96-well V-bottom plates containing 10,000 target cells in a total volume of 200 ␮l/well. The target
cells were mixed with effector cells at different effector cell/target cell
ratios. To facilitate contact between target and effector cells, the plates
were subjected to centrifugation for 60 sec at 1000 rpm. After incubating
the cells for 2 h at 37 C, supernatant from each reaction (20 ␮l) was added
to 200 ␮l of the Eu3⫹ solution. After agitation of the plates for 5 min,
time-resolved fluorescence was measured by using an excitation wavelength of 337 nm and an emission wavelength of 615 nm using the TRF
337 615 optic module on a PHERAstar (BMG Labtech, Durham, NC)
microplate fluorometer. Percent specific lysis was calculated as 100 ⫻
(experimental release ⫺ spontaneous release)/(maximum release ⫺
spontaneous release).
MCF-7 xenograft studies
MCF-7 cells (27), used in xenografts, represent a widely used animal
model for studies of the role of estrogen and genistein in breast cancer
(14, 16, 28). Female athymic BALB/c (nude) mice were purchased from
Charles River Laboratories (Wilmington, MA) and acclimated for a
week. Mice were ovariectomized at 21 d of age by the vendor and
allowed to recover for 7 d. All procedures were conducted with approval
of the Institutional Animal Care and Use Committee (IACUC) of the
University of Illinois at Urbana-Champaign.
One week after the estrogen pellets were inserted into mice, MCF-7
cells were harvested using 500 ␮l trypsin-EDTA (0.5% Trypsin, 5.3
mmol/liter EDTA䡠4Na) (Life Technologies, Inc.-BRL, Grand Island, NY)
per 100-mm culture plate. Cells were adjusted to 1 ⫻ 105 cells per 40 ␮l
Matrigel (Collaborative Biomedical Products, Bedford, MA) and injected
at 40 ␮l per site into each of the two flanks of the athymic mice. Tumors
were grown until their average cross-sectional area reached 38 mm2. One
week before grouping, animal diets were switched to the casein-based
AIN-93G diet. Mice were divided into three treatment groups (13–14
mice per group): negative control (NC), positive control (PC, estradiol
control), and 500 ppm genistein treatment groups. Estradiol pellets were
removed from all mice except the PC group. Mice in the genistein group
started with 500 ppm genistein diet treatment. Tumor growth and body
weight were measured weekly for 23 wk. The cross-sectional area of
tumor was determined by using the formula [length/2 ⫻ width/2␲]
(30). Food intake was measured throughout the study. At the end of the
study, mice were anesthetized and killed by cervical dislocation. Tissues
and blood samples were collected for analysis. Tumors were removed
and stored in liquid nitrogen for RNA analysis.
Statistical analysis
The results are reported as mean ⫾ sem. We compared the ability of
NK92 and NKL cells to induce apoptosis of control and genistein-treated
ER␣HA cells expressing different levels of ER␣. For each panel of Fig.
3, we used the single-tailed t test to compare apoptosis in control cells
and cells treated with three ratios of NK cells (P ⬍ 0.05 for control vs.
genistein-treated cells in Fig. 3, A, C, and D–F; panel B, P ⬎ 0.05).
Results
Nanomolar concentrations of genistein induce PI-9 in MCF-7
breast cancer cells
Using quantitative RT-PCR to determine PI-9 mRNA levels, we examined the concentrations of the soy isoflavones
genistein and equol (⫺) (the active metabolite of diadzein)
(31) and the plant estrogen coumestrol required to induce
PI-9 mRNA in MCF-7 human breast cancer cells. Induction
of PI-9 by E2 was used for comparison. The concentrations
of equol (⫺) and coumestrol required to maximally induce
PI-9 mRNA (100 nm) were 1000 times higher than the 100 pm
E2 required for maximal induction. However, only 10 nm
genistein, a level likely to be observed in the blood of women
consuming soy or soy-based products (32), was required to
maximally induce PI-9 mRNA. In contrast, a 1000-fold higher
concentration of genistein, 10 ␮m, was required to induce pS2
mRNA (Fig. 1A, dashed line). The high concentration of
genistein required to induce pS2 mRNA is consistent with
earlier work (33). Previously, 200 pm E2 was shown to potently induce pS2 mRNA (33).
Because of genistein’s high potency as an inducer of PI-9,
we explored whether binding of genistein in the ligand binding pocket of ER is required. The selective ER modulators
(SERMs) 4-hydroxytamoxifen (OHT, the active metabolite of
5368
Endocrinology, November 2008, 149(11):5366 –5373
Jiang et al. • Genistein-Induced PI-9 Blocks Immunosurveillance
to elevate the ER␣ level 7- to 8-fold (Fig. 2, A and B). When
the ER␣ level is about 3-fold higher than in wild-type MCF-7
cells, genistein’s potency equals that of E2, and both maximally induce PI-9 mRNA at 100 pm. The concentrations of
equol (⫺) and coumestrol required to maximally induce PI-9
mRNA are 100- to 1000-fold higher (Fig. 2C). Increasing ER␣
to 7- to 8-fold above the level in MCF-7 cells increased the
induction of PI-9 mRNA to approximately 250-fold without
significantly altering the concentration of genistein required for
induction (Fig. 2D). Western blot analysis demonstrated that
100 pm genistein elicits a robust induction of PI-9 protein
in the ER␣HA cells expressing the highest levels of ER␣
(Fig. 2E).
Increased induction of PI-9 progressively blocks NK
cell-mediated apoptosis
FIG. 1. Low concentrations of genistein induce PI-9 mRNA in MCF-7
cells. A, Dose-response curves for phytoestrogen induction of PI-9 and
pS2 mRNAs. For the PI-9 study, RNA was isolated 4 h after addition
of the indicated concentrations of E2, genistein (Gen), equol (⫺), and
coumestrol to the cell culture medium. To analyze pS2 induction
(dashed line), RNA was isolated 24 h after genistein treatment. The
level of mRNA in cells maintained in ethanol vehicle was set equal to
1. B, Effect of SERMs on induction of PI-9 mRNA. The medium
contained ethanol vehicle (control) or 10 nM genistein (Gen). The
SERMs OHT, raloxifene (RAL), and ICI 182,780 (ICI) were present at
1000 nM in the presence or absence of 10 nM genistein. Cells were
harvested at 4 h, and RNA was isolated and analyzed for PI-9 mRNA
by quantitative RT-PCR. The data represent the mean ⫾ SEM for at
least three separate experiments.
tamoxifen), raloxifene, and ICI 182,780/Faslodex all block
the induction of PI-9 mRNA by genistein (Fig. 1B). Consistent
with our earlier work (3), OHT partially induces PI-9 mRNA.
Because all three SERMS block genistein induction of PI-9,
and the three SERMS differ in their actions at non-ER-binding sites for estrogens (34), and MCF-7 cells contain negligible levels of ER␤ (24), genistein acts via ER␣ to induce PI-9.
Increasing levels of ER␣ increase PI-9 induction
by genistein
A subset of breast cancers containing very high levels of
ER␣ exhibit increased incidence of recurrence and a poor
prognosis (35). In ER␣HA cells, a model for this type of tumor
in which additional ER␣ is expressed from a tetracyclineinducible promoter (25), E2 induced PI-9 mRNA and protein
to very high levels (3, 5). To evaluate the effects of genistein
in ER␣HA cells containing different levels of ER␣, we used
0.5 ␮g/ml doxycycline (DOX) to increase ER␣ mRNA (Fig.
2A) and protein (Fig. 2B) to levels approximately 3-fold
above the level in wild-type MCF-7 cells and 1 ␮g/ml DOX
We next tested whether induction of PI-9 by genistein
reduces the ability of human immune cells to induce apoptosis of breast cancer cells. Human NK cells and CTLs lyse
their target cells by similar mechanisms, so we used established NK cell lines as killer cells. NK cells target MCF-7 cells
through the granzyme pathway, not the Fas/Fas ligand pathway (3). To test the effect of genistein induction of PI-9 on NK
cells inducing cell death via diverse death pathways, we used
NK92 cells, which contain primarily granzyme B, and NKL
cells, which reportedly contain mostly granzyme A (36, 37).
Granzyme A induces rapid cell death by a pathway distinct
from the granzyme B pathway (38), and its sensitivity to
in vitro inhibition by PI-9 has not been established. We compared the ability of NK92 and NKL cells to induce apoptosis
of control and genistein-treated MCF-7 and ER␣HA cells
expressing different levels of ER␣. For each panel of Fig. 3,
we used the one-tailed Student’s t test to compare apoptosis
in control cells and genistein-treated cells with three ratios of
NK cells (P ⬍ 0.05 for control vs. genistein-treated cells in A,
C, and D–F; P ⬎ 0.05 in B).
Genistein treatment induced levels of PI-9 in MCF-7 cells
sufficient to block NK92 cell-induced apoptosis (Fig. 3A) but
not NKL cell-induced cell death (Fig. 3B). We recently reported a similar finding in MCF-7 cells treated with E2 (3).
In ER␣HA cells expressing 2- to 3-fold higher levels of ER␣
than MCF-7 cells (Fig. 2B), genistein induces a higher level
of PI-9 (Fig. 2E). This increased level of PI-9 inhibited both
NK92 (Fig. 3C) and NKL-induced cell death (Fig. 3D). The
very high level of PI-9 induced by genistein in ER␣HA cells
expressing 7- to 8-fold more ER␣ than MCF-7 cells (Fig. 2E)
resulted in a nearly complete inhibition of both NK92 cellinduced cell death (Fig. 3E) and NKL-mediated cytotoxicity
(Fig. 3F). These data demonstrate that increasing levels of
ER␣ in breast cancer cells elicit enhanced genistein induction
of PI-9. The enhanced genistein induction of PI-9 results in
progressive inhibition of NK cell-induced cell death.
Dietary genistein and soy products induce PI-9 mRNA in
MCF-7 tumors in ovariectomized athymic mice
The studies in Figs. 1–3 used human breast cancer cells in
culture to demonstrate that 100 pm and 10 nm genistein,
concentrations of genistein previously thought to be too low
to elicit significant biological effects, potently induce PI-9 and
Jiang et al. • Genistein-Induced PI-9 Blocks Immunosurveillance
Endocrinology, November 2008, 149(11):5366 –5373
5369
FIG. 2. Increasing levels of ER␣ in breast cancer cells enhance genistein’s potency and efficacy as an inducer of PI-9. A and B, In ER␣HA cells,
increased concentrations of DOX induce increased levels of ER␣ mRNA and protein. The cells were maintained in medium lacking DOX (Con)
or containing 0.5 ␮g/ml (Dox 0.5) or 1.0 ␮g/ml (Dox 1) DOX . RNA was isolated at 48 h, and quantitative RT-PCR was carried out (A) or cell
extracts were prepared and analyzed for ER␣ protein level by Western blotting using calnexin as internal standard (B). C and D, Dose-response
curves for induction of PI-9 mRNA in cells expressing different levels of ER␣. RNA was isolated 24 h after addition of 0.5 ␮g/ml DOX alone
(C) or 1.0 ␮g/ml DOX alone (D) plus the indicated concentrations of E2, genistein, Equol (⫺), and coumestrol. The cells in C and D contain 2to 3- and 7- to 8-fold more ER␣ than MCF-7 cells, respectively. E, Breast cancer cells expressing very high levels of ER␣ produce large quantities
of PI-9 protein. Cells were maintained in medium containing 100 pM genistein and the indicated concentrations of DOX for 24 h (MCF-7) or
48 h (ER␣HA), which maximizes PI-9 production, and analyzed for PI-9 protein level by Western blotting. The data in A, C, and D represent
the mean ⫾ SEM for three separate experiments.
block immunosurveillance. We wished to test whether
genistein consumed from more typical dietary sources could
induce PI-9 in tumors. Although studies testing genistein
effects on endogenous breast tumors in mice might seem
attractive, they are not appropriate. The mouse serpin whose
amino acid sequence and tissue distribution most closely
resembles human PI-9 is serine protease inhibitor 6 (SPI-6).
The protein coding regions of PI-9 and SPI-6 are about 60%
identical. However, the promoter regions of human PI-9 and
mouse SPI-6 show virtually no sequence homology, and the
mouse and rat SPI-6 promoter regions do not contain the
downstream estrogen-responsive unit (ERU) critical for estrogen induction of PI-9 in human cells (39 – 41). In contrast,
the ERU is highly conserved in the human, chimpanzee, and
Rhesus monkey PI-9 genes. In studies using control and
E2-treated rat liver (supplied by Prof. J. Katznellenbogen), E2
did not influence SPI-6 levels (Krieg, S., J. A. Katzenellenbogen, and D. J. Shapiro, unpublished). There are also important differences between mouse and human granzyme B
and their granzyme inhibitors (42, 43). We therefore tested
whether mice fed genistein could induce the same levels of
PI-9 in MCF-7 xenografts that we showed inhibit NK cellmediated lysis of MCF-7 cells in our in vitro studies.
To mimic the slow tumor growth observed in postmeno-
pausal women with breast cancer, we used a level of dietary
genistein previously shown to induce a moderate rate of
tumor growth (17). MCF-7 cell tumors were established by
injection of MCF-7 cells and brief exposure of the mice to
estrogen in estrogen pellets. In one group of mice, continued
exposure to high concentrations of E2 in the estrogen pellets
resulted in rapid growth of the MCF-7 cell tumors [Fig. 4A,
positive control (PC estradiol)]. The other groups of mice
received 500 ppm genistein in their diet, a level of genistein
that results in much slower tumor outgrowth over about 5
months (Fig. 4A, Gen 500), or diets supplemented with soy
flour (29). It required about four times longer for the tumors
in mice fed genistein to reach a 100-mm2 surface area than
for the tumors in mice exposed to high E2 (23 vs. 4 –5 wk,
Fig. 4A).
Because our affinity-purified anti-PI-9 detects both human
PI-9 and its closest mouse relative, SPI-6, and RT-PCR indicated the presence of significant SPI-6 mRNA in the tumor
samples (data not shown), we used quantitative RT-PCR to
specifically determine levels of PI-9 mRNA in the MCF-7
tumor samples. In MCF-7 cells, PI-9 mRNA is rapidly induced by estrogens to high levels, peaks at about 3– 4 h, and
then declines to an induction of severalfold that is maintained for at least a few days (3). We used cultured MCF-7
5370
Endocrinology, November 2008, 149(11):5366 –5373
Jiang et al. • Genistein-Induced PI-9 Blocks Immunosurveillance
FIG. 3. Induction of increasing levels of PI-9 progressively block NK cell-mediated cell death. A,
Genistein protects MCF-7 cells against NK92 cellmediated killing. B, Genistein does not protect
MCF-7 cells against NKL cell-mediated killing.
MCF-7 target (T) cells were maintained in medium
containing ethanol vehicle or 10 nM genistein for
24 h, followed by incubation with NK92 effector (E)
cells (A) or NKL effector cells (B) at different effector
cell/target cell ratios. Cytotoxicity assays were performed as described in Materials and Methods and
Ref. 3. C and D, Genistein protects ER␣HA cells
expressing high levels of ER␣ against NK92 cell and
NKL cell-mediated killing. ER␣HA cells were
treated with ethanol vehicle or with 0.5 ␮g/ml DOX
plus 100 pM genistein for 48 h, followed by the cytotoxicity assay. E and F, In ER␣HA cells expressing
very high levels of ER␣, genistein nearly abolishes
NK92 and NKL cell-induced cell death. ER␣HA cells
were treated with ethanol vehicle or with 1.0 ␮g/ml
DOX plus 100 pM genistein for 48 h, followed by the
cytotoxicity assay. Data in A–F represent the
mean ⫾ SEM for three separate experiments. For each
panel, we used the single-tailed t test to compare
apoptosis in control cells and cells treated with three
ratios of NK cells (P ⬍ 0.05 for control vs. genisteintreated cells in A, C, and D–F; B, P ⬎ 0.05).
cells as control cells because MCF-7 tumors require estrogen
for growth, and the regressing tumor samples did not
contain significant numbers of viable MCF-7 cells. Because
the tumors were continuously exposed to E2, genistein, or
genistein in soy products, we compared levels of PI-9 mRNA
in the frozen tumor samples to PI-9 mRNA levels in control
MCF-7 cells and in MCF-7 cells treated with E2 or genistein
for 24 h, the time used in NK cell killing experiments using
MCF-7 cells (Fig. 3, A and B). In the fast-growing tumors from
mice exposed to high concentrations of E2, PI-9 was induced
to a slightly higher level in the tumor than in MCF-7 cells
maintained in medium containing saturating E2 (Fig. 4B, E2).
The genistein and soy flour diets (⬃400 ppm genistein) resulted in slow-growing tumors (Fig. 4A). In the mice fed
these diets, PI-9 mRNA was induced to about the same level
as is seen in MCF-7 cells maintained in saturating genistein
(Fig. 4B).
Discussion
In contrast to the pS2 gene and most other genes that
require micromolar genistein for induction (23, 24, 33), induction of PI-9 requires nanomolar to picomolar concentrations of genistein. Several lines of evidence indicate that the
induction of PI-9 in MCF-7 cells, and in ER␣HA cells, is
mediated by genistein acting via ER␣. Although genistein
binds with much higher affinity to ER␤ than to ER␣ and
lower concentrations of genistein activate transcription
through ER␤ (24, 44), MCF-7 cells contain negligible ER␤
(24). Also, in ER␣HA cells, which are MCF-7 cells stably
transfected to express additional ER␣ under the control of
a tetracycline-regulated promoter, the concentrations of
genistein required for robust induction of PI-9 is reduced
from 10 nm in MCF-7 cells to 100 pm. Several studies indicate
that G protein-coupled receptor 30 (GPR30) is a potential
membrane ER (34). However, GPR30 is an integral membrane protein thought to participate mostly in nongenomic
effects of estrogens. Because OHT is an agonist on GPR30 and
ICI 182,780 also displays some agonist effects (34), our observation that a 100-fold molar excess of OHT, raloxifene,
and ICI 182,780 all block genistein induction of PI-9 (Fig.
1B) is inconsistent with induction of PI-9 by genistein
acting through GPR30. Although genistein is a tyrosine
kinase inhibitor, this requires micromolar concentrations
of genistein, not the nanomolar to picomolar concentrations we use. Increasing levels of ER␣ in ER␣HA cells
enhance the magnitude of genistein induction of PI-9 and
Jiang et al. • Genistein-Induced PI-9 Blocks Immunosurveillance
FIG. 4. Dietary genistein induces PI-9 in MCF-7 tumors in mice. A,
Growth rates of MCF-7 tumors in ovariectomized athymic mice. At wk 0,
the estrogen pellets were removed, and the mice were divided into
three treatment groups: PC (positive control, E2) (13 mice; n ⫽ 51
tumors), GEN 500 (genistein 500 ppm, 14 mice; n ⫽ 56 tumors), and
NC (negative control, regressing tumor) that were fed AIN-93G diet
alone (13 mice; n ⫽ 52 tumors). Tumor size was then measured weekly
for 23 wk. Data are expressed as means ⫾ SEM cross-sectional tumor
area for all tumors in each group. B, Dietary genistein and soy induce
PI-9 in MCF-7 solid tumors. Mice were exposed to high (2 mg) E2 in
implanted cholesterol pellets and tumors harvested (n ⫽ 4) at about
11 wk (black bar) and frozen. Mice were fed diets containing genistein
(500 ppm; n ⫽ 5) or soy flour (diet containing 20% soy protein, ⬃400
ppm genistein; n ⫽ 3) (black bars). Tumors were harvested at about
23 wk when they reached the same size as the E2 tumors (average
cross-sectional area ⫽ 38 mm2). Control MCF-7 cells were maintained
in medium lacking hormones or in 10 nM E2 or 10 nM genistein for 24 h
and harvested (n ⱖ 3; white bars). RNA was extracted and analyzed
for PI-9 mRNA by quantitative RT-PCR as we describe. Data represent the mean ⫾ SEM for at least three samples.
reduce the concentrations of genistein required for maximal induction (Fig. 2). Thus, our data support the view
that genistein induces PI-9 by binding in the ligand binding pocket of ER␣.
The PI-9 gene’s sensitivity to activation by low concentrations of genistein may result from the ability of genistein
to induce an ER conformation highly conducive to binding
to the PI-9 ERU. We previously reported that binding of
intracellular ER to the ERU that controls PI-9 transcription
involves a balance between ER conformations induced by the
DNA sequence and the bound ligand (41). Because the PI-9
ERU contains two directly adjacent weak ER␣ binding sites
Endocrinology, November 2008, 149(11):5366 –5373
5371
(39, 41), it is also possible that the two ER␣ dimers mutually
stabilize each other’s binding to the ERU, making the PI-9
gene highly responsive to small increases in the intracellular
concentration of ER. Although the factors that make the PI-9
gene sensitive to activation by extremely low concentrations
of genistein are unclear, the concentrations of genistein that
fully induce PI-9 in MCF-7 cells, and in ER␣HA cells, are far
lower than those previously reported to elicit estrogenic effects through ER␣ (23, 24, 33).
RNA interference knockdown of PI-9 completely restored
the ability of cytolytic lymphocytes to induces cell death in
estrogen-treated MCF-7 cells and hepatoblastoma cells (3, 4).
Our observation that lower concentrations of genistein (100
pm) in MCF-7-derived ER␣HA cells, which express additional ER␣, more effectively induce PI-9 and inhibit NK cell
apoptosis than the higher genistein concentration (10 nm) in
MCF-7 cells makes it extremely unlikely that genistein’s ability to inhibit NK cell-induced apoptosis through the granzyme system is mediated by genistein’s ability to inhibit
tyrosine kinases or by an as yet undiscovered activity of
genistein. Thus, genistein’s ability to inhibit NK cell-mediated apoptosis through the granzyme pathway likely
results from its acting through ER␣ to induce the granzyme inhibitor PI-9.
Numerous studies show an association between the risk of
breast cancer and elevated blood levels of estrogen. Studies
showing that estrogen-replacement therapy increases the relative risk of breast cancer have led to increased interest and
use of soy products as an alternative therapy to treat peri- or
postmenopausal symptoms. Prolonged dietary supplementation with soy makes about 10 –20 nm genistein available to
cells (32). Because 10 nm genistein fully induces PI-9 and
inhibits NK92-induced apoptosis of MCF-7 breast cancer
cells, circulating concentrations of genistein that can be
reached during isoflavone supplementation, or by consuming a diet rich in soy isoflavones, are probably sufficient to
robustly induce PI-9 and thereby reduce the ability of cytolytic lymphocytes to kill target breast cancer cells. Because
100 pm E2 maximally induces PI-9 (Fig. 2), and 50 pm E2,
which is the average circulating level of estrogen in postmenopausal women (45), partially induces PI-9, genistein
consumed from nearly ubiquitous dietary soy may enhance
the induction of PI-9 in breast cancer cells in a large cohort
of postmenopausal women. Two populations, infants and
young children and breast cancer patients undergoing treatment with aromatase inhibitors, have very low circulating
estrogen levels that should not induce PI-9 (46 – 48). In these
populations, dietary soy may contribute sufficient genistein
to induce PI-9 in some cells and enable the cells to evade
immunosurveillance. The consequences of genistein altering
the susceptibility of target cells to lysis by cytolytic lymphocytes are probably influenced by both the duration and time
of genistein exposure. In some cases, especially in liver cells,
some of the pathology that leads to tumor formation is based
on inflammatory processes that attract cytolytic lymphocytes, which then damage the target cells (49). By inducing
PI-9, genistein might protect cells against cytolytic lymphocyte-mediated damage initiated by inflammatory processes.
However, once a cell is transformed, exposure to genistein
5372
Endocrinology, November 2008, 149(11):5366 –5373
will induce PI-9 and reduce cytolytic lymphocyte-mediated
immunosurveillance of the newly transformed cells.
A small subset of breast cancers express much higher
levels of ER␣ than are seen in MCF-7 breast cancer cells. This
generally results in a poor prognosis (25, 50). ER␣HA cells
serve as a potential model for these types of cancers (25). In
these cells, 100 pm genistein potently induced PI-9, resulting
in a nearly complete inhibition of NK cell-mediated lysis of
the ER␣HA cells. This is the lowest concentration of genistein
shown to elicit a biological effect and suggests that for this
small subgroup of breast cancer patients, it may be appropriate to consider issues related to levels of soy consumption.
Altered expression of PI-9 is associated with a poor prognosis
in several tumors (51–53). Because genistein from dietary soy
reportedly enhances the growth of preexisting breast cancers
(15, 16), the possibility that genistein induction of PI-9 plays
a role in tumor progression in breast cancer patients whose
tumors contain very high levels of ER␣ warrants further
attention.
Acknowledgments
We are grateful to Dr. E. Alarid, Dr. J. Miller, and Dr. M. Robertson
who generously provided the ER␣HA, NK92, and NKL cell lines, respectively. We thank Dr. D. Kranz and Dr. S. Nordeen for critical suggestions and comments on the manuscript.
Received June 9, 2008. Accepted July 21, 2008.
Address all correspondence and requests for reprints to: David J.
Shapiro, Department of Biochemistry, University of Illinois, 413 RAL,
600 South Mathews Avenue, Urbana, Illinois 61801. E-mail: djshapir@
uiuc.edu.
This work was supported by National Institutes of Health Grants DK
071909 (D.G.S.) and Grants CA77355 and AG024387 (W.G.H.).
Disclosure Statement: The authors have nothing to disclose.
References
1. Deroo BJ, Korach KS 2006 Estrogen receptors and human disease. J Clin Invest
116:561–570
2. Gradishar WJ, Cella D 2006 Selective estrogen receptor modulators and prevention of invasive breast cancer. JAMA 295:2784 –2786
3. Jiang X, Ellison SJ, Alarid ET, Shapiro DJ 2007 Interplay between the levels
of estrogen and estrogen receptor controls the level of the granzyme inhibitor,
proteinase inhibitor 9 and susceptibility to immune surveillance by natural
killer cells. Oncogene 26:4106 – 4114
4. Jiang X, Orr BA, Kranz DM, Shapiro DJ 2006 Estrogen induction of the
granzyme B inhibitor, proteinase inhibitor 9, protects cells against apoptosis
mediated by cytotoxic T lymphocytes and natural killer cells. Endocrinology
147:1419 –1426
5. Cunningham TD, Jiang X, Shapiro DJ 2007 Expression of high levels of human
proteinase inhibitor 9 blocks both perforin/granzyme and Fas/Fas ligandmediated cytotoxicity. Cell Immunol 245:32– 41
6. Medema JP, de Jong J, Peltenburg LT, Verdegaal EM, Gorter A, Bres SA,
Franken KL, Hahne M, Albar JP, Melief CJ, Offringa R 2001 Blockade of the
granzyme B/perforin pathway through overexpression of the serine protease
inhibitor PI-9/SPI-6 constitutes a mechanism for immune escape by tumors.
Proc Natl Acad Sci USA 98:11515–11520
7. Dunn GP, Old LJ, Schreiber RD 2004 The immunobiology of cancer immunosurveillance and immunoediting. Immunity 21:137–148
8. Froelich CJ, Metkar SS, Raja SM 2004 Granzyme B-mediated apoptosis: the
elephant and the blind men? Cell Death Differ 11:369 –371
9. Russell JH, Ley TJ 2002 Lymphocyte-mediated cytotoxicity. Annu Rev Immunol 20:323–370
10. Kummer JA, Micheau O, Schneider P, Bovenschen N, Broekhuizen R, Quadir
R, Strik MC, Hack CE, Tschopp J 2007 Ectopic expression of the serine
protease inhibitor PI9 modulates death receptor-mediated apoptosis. Cell
Death Differ 14:1486 –1496
11. Setchell KD 1998 Phytoestrogens: the biochemistry, physiology, and implications for human health of soy isoflavones. Am J Clin Nutr 68:1333S–1346S
12. Yamamoto S, Sobue T, Kobayashi M, Sasaki S, Tsugane S 2003 Soy, isoflavones, and breast cancer risk in Japan. J Natl Cancer Inst 95:906 –913
Jiang et al. • Genistein-Induced PI-9 Blocks Immunosurveillance
13. McCarty MF 2006 Isoflavones made simple: genistein’s agonist activity for the
␤-type estrogen receptor mediates their health benefits. Med Hypotheses 66:
1093–1114
14. Ju YH, Doerge DR, Allred KF, Allred CD, Helferich WG 2002 Dietary
genistein negates the inhibitory effect of tamoxifen on growth of estrogendependent human breast cancer (MCF-7) cells implanted in athymic mice.
Cancer Res 62:2474 –2477
15. Ju YH, Allred KF, Allred CD, Helferich WG 2006 Genistein stimulates growth
of human breast cancer cells in a novel, postmenopausal animal model, with
low plasma estradiol concentrations. Carcinogenesis 27:1292–1299
16. Allred CD, Allred KF, Ju YH, Virant SM, Helferich WG 2001 Soy diets
containing varying amounts of genistein stimulate growth of estrogen-dependent (MCF-7) tumors in a dose-dependent manner. Cancer Res 61:5045–5050
17. Allred CD, Allred KF, Ju YH, Clausen LM, Doerge DR, Schantz SL, Korol
DL, Wallig MA, Helferich WG 2004 Dietary genistein results in larger MNUinduced, estrogen-dependent mammary tumors following ovariectomy of
Sprague-Dawley rats. Carcinogenesis 25:211–218
18. Pagliacci MC, Smacchia M, Migliorati G, Grignani F, Riccardi C, Nicoletti I 1994
Growth-inhibitory effects of the natural phyto-oestrogen genistein in MCF-7
human breast cancer cells. Eur J Cancer 30A:1675–1682
19. Wang TT, Sathyamoorthy N, Phang JM 1996 Molecular effects of genistein on
estrogen receptor mediated pathways. Carcinogenesis 17:271–275
20. Akiyama T, Ishida J, Nakagawa S, Ogawara H, Watanabe S, Itoh N, Shibuya
M, Fukami Y 1987 Genistein, a specific inhibitor of tyrosine-specific protein
kinases. J Biol Chem 262:5592–5595
21. Cappelletti V, Fioravanti L, Miodini P, Di Fronzo G 2000 Genistein blocks
breast cancer cells in the G2M phase of the cell cycle. J Cell Biochem 79:594 – 600
22. Whitten PL, Patisaul HB 2001 Cross-species and interassay comparisons of
phytoestrogen action. Environ Health Perspect 109(Suppl 1):5–20
23. Shioda T, Chesnes J, Coser KR, Zou L, Hur J, Dean KL, Sonnenschein C, Soto
AM, Isselbacher KJ 2006 Importance of dosage standardization for interpreting transcriptomal signature profiles: evidence from studies of xenoestrogens.
Proc Natl Acad Sci USA 103:12033–12038
24. Chang EC, Charn TH, Park SH, Helferich WG, Komm B, Katzenellenbogen
JA, Katzenellenbogen BS 2008 Estrogen receptors ␣ and ␤ as determinants of
gene expression: influence of ligand, dose, and chromatin binding. Mol Endocrinol 22:1032–1043
25. Fowler AM, Solodin NM, Valley CC, Alarid ET 2006 Altered target gene
regulation controlled by estrogen receptor-␣ concentration. Mol Endocrinol
20:291–301
26. Blomberg K, Hautala R, Lovgren J, Mukkala VM, Lindqvist C, Akerman K
1996 Time-resolved fluorometric assay for natural killer activity using target
cells labelled with a fluorescence enhancing ligand. J Immunol Methods 193:
199 –206
27. Soule HD, Vazguez J, Long A, Albert S, Brennan M 1973 A human cell line
from a pleural effusion derived from a breast carcinoma. J Natl Cancer Inst
51:1409 –1416
28. Soule HD, McGrath CM 1980 Estrogen responsive proliferation of clonal
human breast carcinoma cells in athymic mice. Cancer Lett 10:177–189
29. Allred CD, Allred KF, Ju YH, Goeppinger TS, Doerge DR, Helferich WG
2004 Soy processing influences growth of estrogen-dependent breast cancer
tumors. Carcinogenesis 25:1649 –1657
30. Gottardis MM, Jiang SY, Jeng MH, Jordan VC 1989 Inhibition of tamoxifenstimulated growth of an MCF-7 tumor variant in athymic mice by novel
steroidal antiestrogens. Cancer Res 49:4090 – 4093
31. Muthyala RS, Ju YH, Sheng S, Williams LD, Doerge DR, Katzenellenbogen
BS, Helferich WG, Katzenellenbogen JA 2004 Equol, a natural estrogenic
metabolite from soy isoflavones: convenient preparation and resolution of
R- and S-equols and their differing binding and biological activity through
estrogen receptors ␣ and ␤. Bioorg Med Chem 12:1559 –1567
32. Nagel SC, vom Saal FS, Welshons WV 1998 The effective free fraction of
estradiol and xenoestrogens in human serum measured by whole cell uptake
assays: physiology of delivery modifies estrogenic activity. Proc Soc Exp Biol
Med 217:300 –309
33. Hsieh CY, Santell RC, Haslam SZ, Helferich WG 1998 Estrogenic effects of
genistein on the growth of estrogen receptor-positive human breast cancer
(MCF-7) cells in vitro and in vivo. Cancer Res 58:3833–3838
34. Prossnitz ER, Oprea TI, Sklar LA, Arterburn JB 2008 The ins and outs of
GPR30: a transmembrane estrogen receptor. J Steroid Biochem Mol Biol 109:
350 –353
35. Gaskell DJ, Hawkins RA, Sangsterl K, Chetty U, Forrest AP 1989 Relation
between immunocytochemical estimation of oestrogen receptor in elderly
patients with primary breast cancer and response to tamoxifen. Lancet 1:1044 –
1046
36. Ko YH, Park S, Jin H, Woo H, Lee H, Park C, Kim K 2007 Granzyme B
leakage-induced apoptosis is a crucial mechanism of cell death in nasal-type
NK/T-cell lymphoma. Lab Invest 87:241–250
37. Mahrus S, Craik CS 2005 Selective chemical functional probes of granzymes
A and B reveal granzyme B is a major effector of natural killer cell-mediated
lysis of target cells. Chem Biol 12:567–577
38. Lieberman J, Fan Z 2003 Nuclear war: the granzyme A-bomb. Curr Opin
Immunol 15:553–559
Jiang et al. • Genistein-Induced PI-9 Blocks Immunosurveillance
39. Krieg SA, Krieg AJ, Shapiro DJ 2001 A unique downstream estrogen responsive unit mediates estrogen induction of proteinase inhibitor-9, a cellular
inhibitor of IL-1␤-converting enzyme (caspase 1). Mol Endocrinol 15:1971–
1982
40. Kannan-Thulasiraman P, Shapiro DJ 2002 Modulators of inflammation use
nuclear factor-␬B and activator protein-1 sites to induce the caspase-1 and
granzyme B inhibitor, proteinase inhibitor 9. J Biol Chem 277:41230 – 41239
41. Krieg AJ, Krieg SA, Ahn BS, Shapiro DJ 2004 Interplay between estrogen
response element sequence and ligands controls in vivo binding of estrogen
receptor to regulated genes. J Biol Chem 279:5025–5034
42. Kaiserman D, Bird CH, Sun J, Matthews A, Ung K, Whisstock JC, Thompson
PE, Trapani JA, Bird PI 2006 The major human and mouse granzymes are
structurally and functionally divergent. J Cell Biol 175:619 – 630
43. Cullen SP, Adrain C, Luthi AU, Duriez PJ, Martin SJ 2007 Human and murine
granzyme B exhibit divergent substrate preferences. J Cell Biol 176:435– 444
44. Kuiper GG, Carlsson B, Grandien K, Enmark E, Haggblad J, Nilsson S,
Gustafsson JA 1997 Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors ␣ and ␤. Endocrinology 138:
863– 870
45. Meldrum DR, Davidson BJ, Tataryn IV, Judd HL 1981 Changes in circulating
steroids with aging in postmenopausal women. Obstet Gynecol 57:624 – 628
46. Yellayi S, Naaz A, Szewczykowski MA, Sato T, Woods JA, Chang J, Segre
M, Allred CD, Helferich WG, Cooke PS 2002 The phytoestrogen genistein
induces thymic and immune changes: a human health concern? Proc Natl Acad
Sci USA 99:7616 –7621
47. Winer EP, Hudis C, Burstein HJ, Wolff AC, Pritchard KI, Ingle JN,
Endocrinology, November 2008, 149(11):5366 –5373
48.
49.
50.
51.
52.
53.
5373
Chlebowski RT, Gelber R, Edge SB, Gralow J, Cobleigh MA, Mamounas EP,
Goldstein LJ, Whelan TJ, Powles TJ, Bryant J, Perkins C, Perotti J, Braun S,
Langer AS, Browman GP, Somerfield MR 2005 American Society of Clinical
Oncology technology assessment on the use of aromatase inhibitors as adjuvant therapy for postmenopausal women with hormone receptor-positive
breast cancer: status report 2004. J Clin Oncol 23:619 – 629
Setchell KD, Cassidy A 1999 Dietary isoflavones: biological effects and relevance to human health. J Nutr 129:758S–767S
Nakamoto Y, Guidotti LG, Kuhlen CV, Fowler P, Chisari FV 1998 Immune
pathogenesis of hepatocellular carcinoma. J Exp Med 188:341–350
Fowler AM, Solodin N, Preisler-Mashek MT, Zhang P, Lee AV, Alarid ET
2004 Increases in estrogen receptor-␣ concentration in breast cancer cells promote serine 118/104/106-independent AF-1 transactivation and growth in the
absence of estrogen. FASEB J 18:81–93
ten Berge RL, Meijer CJ, Dukers DF, Kummer JA, Bladergroen BA, Vos W,
Hack CE, Ossenkoppele GJ, Oudejans JJ 2002 Expression levels of apoptosisrelated proteins predict clinical outcome in anaplastic large cell lymphoma.
Blood 99:4540 – 4546
ten Berge RL, Oudejans JJ, Ossenkoppele GJ, Meijer CJ 2003 ALK-negative
systemic anaplastic large cell lymphoma: differential diagnostic and prognostic aspects: a review. J Pathol 200:4 –15
van Houdt IS, Oudejans JJ, van den Eertwegh AJ, Baars A, Vos W,
Bladergroen BA, Rimoldi D, Muris JJ, Hooijberg E, Gundy CM, Meijer CJ,
Kummer JA 2005 Expression of the apoptosis inhibitor protease inhibitor 9
predicts clinical outcome in vaccinated patients with stage III and IV melanoma. Clin Cancer Res 11:6400 – 6407
Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the
endocrine community.