Human chorionic gonadotropin decreases proliferation and invasion of breast cancer MCF-7 cells by inhibiting NF-kappaB and AP-1 activation.

The epidemiological data suggest that breast cancer risk decreases in women who complete full-term pregnancy at a young age. Studies on a rat breast cancer model indicate that human chorionic gonadotropin (hCG), a hormone that is present in very high levels during pregnancy, could be responsible for this decrease. These findings, as well as those demonstrating the presence of functional luteinizing hormone (LH)/hCG receptors in human breast cells, prompted us to investigate the anti-proliferative and anti-invasive effects of hCG in human breast cancer MCF-7 cells by down-regulating NF-kappaB and AP-1 transcription factors. Treatment of MCF-7 cells with highly purified hCG resulted in a modest dose-dependent and hormone-specific decrease in cell proliferation. hCG treatment also decreased cell invasion, which was more dramatic than the decrease in cell proliferation. These hCG actions were abrogated when receptor synthesis was inhibited by treatment with antisense hCG/LH receptor phosphorothioate oligodeoxynucleotide. hCG treatment prevented the tumor necrosis factor-dependent NF-kappaB and AP-1 activation, which paralleled a decrease in the phosphorylation and degradation of IkappaBalpha. The findings that hCG treatment increased cAMP synthesis and activated cAMP-dependent protein kinase, dibutyryl cAMP mimicked hCG in preventing NF-kappaB activation, and dideoxyadenosine, an adenylate cyclase inhibitor, prevented the hCG effect on NF-kappaB suggested that the hCG actions are mediated via the cAMP-dependent protein kinase A signaling pathway. In summary, our results demonstrate that hCG has anti-proliferative and anti-invasive effects in MCF-7 cells by down-regulating NF-kappaB and AP-1. These findings support the premise that hCG could be responsible for the pregnancy-induced protection against breast cancer in women.

Human chorionic gonadotropin (hCG) 1 belongs to both the glycoprotein hormone and the cystine knot growth factor fam-ilies (1,2). Its levels exponentially increase during the first trimester followed by a rapid decline to low steady state levels during the remainder of pregnancy (3). Contrary to the widely held belief, hCG has pregnancy-maintaining actions other than preventing corpus luteum regression during the first 9 weeks of pregnancy (3). These actions are mediated by a G-proteincoupled receptor, which also binds luteinizing hormone (LH) (4,5). Low levels of these receptors are also present in breast (6 -13) as well as in numerous other nongonadal tissues (14). Activation of nongonadal hCG/LH receptors results in tissuespecific responses (14), which in the case of breast involve a number of them that can be collectively described as anticancer (10,(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29). These data, along with epidemiological findings of a decreased breast cancer risk in women who complete full-term pregnancy at a young age (30 -33) and a protective effect of hCG against carcinogen-induced mammary tumor formation in rodents (34), suggest that hCG may have a direct role in breast cancer protection. Since it is not practical to investigate anti-breast cancer effects of hCG in women and due to limitations in obtaining normal human breast epithelial cells or cells in primary breast cancer lesions, we used MCF-7 cells, which were first isolated from pleural effusion of a 69-year-old woman who had breast adenocarcinoma (35). These cells retained several characteristics of differentiated mammary epithelium and have been extensively used in breast cancer research. Thus, this cell line is a reasonable choice to test our hypothesis that hCG has anti-proliferative and anti-invasive effects via possibly down-regulating the activation of NF-B and AP-1.

EXPERIMENTAL PROCEDURES
Materials-The following items were obtained: highly purified hCG (CR-127; 14,900 IU/mg), human follicle-stimulating hormone (AFP-87929B, 1,683 IU/mg), human LH (AFP-0264B, 4,015 IU/mg), human thyroid-stimulating hormone (AFP-4314C, 15 IU/mg), ␣ (CR-125) and ␤ (CR-125; 29 IU/mg) subunits of hCG from the NIDDK, National Institutes of Health, National Hormone and Pituitary Program and Dr. A. F. Parlow, Torrance, CA; polyclonal hCG/LH receptor antibody raised in rabbits against a synthetic N terminus amino acid sequence of receptor protein and receptor peptide from Dr. Patrick Roche, who is now at Ventana Medical Systems Inc., Tucson, AZ; and bacteria-derived purified recombinant human TNF from Genentech, Inc., South San Francisco, CA. The following items were purchased from indicated commercial sources: MCF-7 cells from American Type Culture Collection (Manassas, VA); Dulbecco's modified Eagle's medium nutrient mixture F-12 Ham (DMEM/F-12) and antibiotics from Invitrogen; leupeptin, aprotinin, and dibutyryl cAMP from Sigma; enzyme immunoassay kits for cAMP measurements from Cayman Chemical Co. (Ann Arbor, MI); Pre-Tag nonradioactive protein kinase (PKA) activity measurement kits and CellTitre 96® Aqueous One Solution cell proliferation assay kits from Promega Corp. (Madison, WI); avidin-biotin immunoperoxidase kits from Vector Laboratories, Inc. (Burlingame, CA); antibodies against IB␣ and double-stranded oligonucleotide contain-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  ing AP-1 consensus sequence from Santa Cruz Biotechnology (Santa Cruz, CA); phospho-IkB␣ (Ser-32) antibody from New England BioLabs (Beverly, MA); double-stranded oligonucleotide containing NF-B consensus sequence from Invitrogen; Biocoat Matrigel invasion chambers from BD Biosciences. Twenty-one mer phosphorothioate antisense (5Ј-GCC GAG AAC CGC TGC TTC ATG-3Ј) and sense (5Ј-CAT GAA GCA GCG GTT CTC GGC-3Ј) oligodeoxynucleotides (ODNs), designed from human hCG/LH receptor cDNA sequence covering ATG translation initiation codon, were synthesized in our laboratory.
Cell Culture-MCF-7 cells were cultured in DMEM/F-12 containing 10% fetal bovine serum, 100 IU/ml penicillin, and 100 mg/ml streptomycin. The cells were maintained at 37°C in a humidified atmosphere of 5% CO 2 and 95% air. When cells reached ϳ80% confluence, they were then washed twice and cultured in phenol red-free medium and used in various experiments.
Cell Proliferation Assay-Cells were seeded in 96-well culture plates (5 ϫ 10 3 cells/well) and cultured for 5 days in DMEM/F-12 containing various hormones. The culture medium was replaced every other day with medium of the same composition. At the end of culture, the cell proliferation was measured using CellTiter 96® Aqueous One Solution cell proliferation assay kits.
Cell Invasion Assay-The cells (5 ϫ 10 3 ) were incubated for 24 h in the presence or absence of increasing concentrations of hCG. Then they were added to cell culture inserts mounted with Cyclopore 8.0-m polyester membranes and coated with and without (control membranes) Matrigel. The cell inserts were then transferred to a 24-well cell culture cluster in which 1.25 ml of DMEM containing 2.5% NCS was pre-added to each well, and the plates were incubated for either 24 or 48 h at 37°C under a 5% CO 2 atmosphere. The cells and Matrigel were removed by scrubbing the top surface of each insert with a cotton swab. Membranes were washed twice with DMEM/F-12 and then stained with Diffquick solution (Baxter, Chicago, IL). The membrane from each insert was cut, placed on a slide with the bottom side up, and examined under a light microscope. The pictures were taken from representative fields of each membrane, and the cell number was counted in a 49-cm 2 area of the photographs. The percentage of invasion was calculated by a mean number of cells invading through the Matrigel insert membranes, divided by the mean number of cells migrating through the control insert membranes, multiplied by 100.
Cell Culture with Phosphorothioate Antisense and Sense Human hCG/LH Receptor ODNs-MCF-7 cells were plated at a density of 5 ϫ 10 5 cells/well and cultured for 24 h in 6-well plates in serum and phenol red-free medium containing 6 M antisense or sense ODNs. Then immunostaining for hCG/LH receptors was performed on some cells, whereas others were used in different experiments.
Immunocytochemistry-The cell pellets were fixed in Bouin's solution and immunostained by an avidin-biotin immunoperoxidase method using a 1:500 dilution of hCG/LH receptor antibody (7). For the procedural controls, the receptor antibody, preabsorbed with excess receptor peptide, was used.
NF-B Activation Assay-Electrophoretic gel mobility shift assays were carried out as described previously (36). Briefly, nuclear extracts prepared from MCF-7 cells (2 ϫ 10 6 cells/ml) were incubated with 32 P-end-labeled 45-mer double-stranded NF-B oligonucleotide (6 g of protein with 16 fmol of DNA) from the human immunodeficiency virus-long terminal repeat, 5Ј-TTGTTACAAGGGACTTTCCGCTGGG-GACTTTCCAGGGAGGCGTGG-3Ј (bold indicates NF-B binding sites) for 15 min at 37°C, and then the DNA-protein complexes formed were resolved from free oligonucleotide on 6.6% native polyacrylamide gels. A double-stranded mutated oligonucleotide, 5Ј-TTGTTACAACT-CACTTTCCGCTGCTCACTTTCAGGGAGGCGTGG-3Ј, was used to examine the specificity of NF-B binding to DNA. The specificity of binding was also examined by competition assay with the unlabeled oligonucleotide. The gels were dried, and the radioactive bands were visualized and quantitated with a PhosphorImager (Amersham Biosciences) using ImageQuaNT software.
AP-1 Activation Assay-To assay AP-1 activation by electrophoretic gel mobility shift assay (36), 6 g of nuclear extract protein was incubated with 16 fmol of the 32 P-end-labeled AP-1 consensus oligonucleotide 5Ј-CGCTTGATGACTCAGCCGGAA-3Ј (bold indicates the AP-1 binding site) for 15 min at 37°C, and then the DNA-protein complexes formed were resolved from free oligonucleotide on 6% native polyacrylamide gels. The specificity of binding was examined by competition assay with unlabeled oligonucleotide. The radioactive bands were visualized and quantified as indicated above.
Western Blot for IB␣-Cytoplasmic extracts were prepared from MCF-7 cells, and proteins were resolved by 10% SDS-polyacrylamide gel electrophoresis (37,38). The proteins were then electroblotted to nitrocellulose filters, probed with 1:300 dilution of rabbit polyclonal antibodies against nonphosphorylated or phosphorylated IB␣, and detected by chemiluminescence (ECL, Amersham Biosciences).
Measurement of cAMP Levels-The cAMP levels in 50-l medium aliquots were quantified following the instructions provided in the enzyme immunoassay kit. The specificity of cAMP antibody was 100% for acetylated cAMP, 0.3% for cAMP, 0.05% for acetylated cGMP, and 0.01% or less for cGMP, acetylated adenosine, cytidine, guanosine, and uridine. The intra-and interassay coefficients of variation were less than 10%. The detection limit of the assay was 1.1 pmol/ml.
Measurement of PKA Activity-Cells were lysed by sonication at 4°C in 200 l of 25 mM Tris-HCl buffer, pH 7.5, containing 1 mM EDTA, 1 mM dithiothreitol, 20 mM NaCl, 0.5 mM phenylmethylsulfonyl fluoride, 1 M aprotinin, and 50 M leupeptin. PKA activity was determined by incubating 5-7 g of lysate protein for 30 min at 30°C with fluorescentlabeled A1 peptide. Nonphosphorylated and phosphorylated fluorescent peptides were separated on 0.8% agarose gels. Phosphorylated fluorescent peptide bands were excised and eluted, and the optical density at 570 nm was measured using a 96-well plate reader. PKA activity was calculated from the densitometric values using instructions provided by the kit manufacturer. Positive and negative controls supplied in the kits were assayed at the same time as MCF-7 cell samples.
Replication of Experiments and Statistical Analysis-All the experiments were performed in duplicate and repeated at least three times on different occasions. All the data points were pooled for the calculation of means and their standard errors and for one-way analysis of variance and Duncan's multiple comparison test (39).

RESULTS
The Effect of hCG on Cell Proliferation-Culturing MCF-7 cells with highly purified hCG resulted in a modest dose-dependent significant decrease in cell proliferation (Fig. 1A). This effect was mimicked by LH but not by others such as folliclestimulating hormone and thyroid-stimulating hormone in the same glycoprotein hormone family (Fig. 1B). Isolated subunits of hCG also had no effect.
The Effect of hCG on Cell Invasion-Treatment of MCF-7 cells with hCG resulted in a dose-dependent significant decrease in invasion across Matrigel membranes at 24 or 48 h. (Fig. 2). This decrease was more dramatic than the decrease in cell proliferation.
Receptor Requirement for the hCG Effects in MCF-7 Cells-The MCF-7 cells contain higher hCG/LH receptor levels than MDA-MB 231 and normal human breast epithelial HMEC cells (7,9). These higher levels were due to increased gene transcription as a result of increased levels of Ap2 and Egr transcription factors (9). As MCF-7 cells are nontraditional targets of hCG/LH action, we felt some verification of receptor requirement in gonadotropins action was needed. This verification was obtained by inhibiting hCG/LH receptor synthesis in MCF-7 cells by treating them with phosphorothioate ODNs synthesized from human receptor sequence before treating them with hCG.
As expected, MCF-7 cells were immunostained for hCG/LH receptor protein (Fig. 3A). This staining was absent when the receptor antibody was preabsorbed with excess receptor peptide (Fig. 3D). Treatment of MCF-7 cells with antisense ( Fig.  3B) but not sense (Fig. 3C) hCG/LH receptor ODN resulted in a decreased receptor immunostaining to the procedural control level (Fig. 3D).

in antisense ODN-treated MCF-7 cells. In sense ODN-treated cells, which had similar receptor levels as non-ODN-treated cells, hCG was quite effective in decreasing both cell proliferation and invasion.
Effect of hCG on NF-B Activation-NF-B is one of the transcription factors that inhibits cell proliferation and invasion by regulating the gene expression (40, 41). Cytokines such as TNF mediate its effects by activating NF-B (40 -42). This led us to investigate the ability of hCG to prevent NF-B activation induced by TNF. Fig. 5A shows that treatment of MCF-7 cells with TNF alone resulted in a time-dependent 8-fold activation of NF-B. This activation was prevented by pretreatment with hCG.
Effect of hCG on Phosphorylation and Degradation of IB␣-Inactive NF-B is bound to inhibitor protein IB␣ in cytoplasm (40,41). This binding masks the nuclear localization signal sequence in NF-B. The phosphorylation and subsequent degradation of IB␣ by the ubiquitin pathway are necessary for the release of NF-B and expose its nuclear localization signal

FIG. 4. Effect of hCG on cell proliferation (A) and invasion (B) in antisense or sense hCG/LH receptor ODN-treated MCF-7 cells.
The cells were treated for 24 h at 37°C with either 6 M antisense or 6 M sense hCG/LH receptor ODNs, and then different concentrations of hCG were added and culturing was continued for another 5 days for cell proliferation assay or for 1 day for cell invasion assay. *, p Ͻ 0.05 as compared with the control. sequence, which allows it to move into the nucleus to bind to its target sequences in the promoter region of responsive genes (40, 41). The ability of hCG to inhibit TNF-dependent NF-B activation suggests that hCG might inhibit TNF-dependent IB phosphorylation and degradation. The studies to investigate this possibility revealed that TNF treatment increased IB phosphorylation between 5 and 15 min of incubation followed by return to basal levels, and hCG treatment inhibited the increase (Fig. 5B). Fig. 5C shows that TNF treatment also decreased IB␣ levels at 10 -15 min as a result of its degradation. This decrease was also prevented by hCG pretreatment.
Effect of hCG on AP-1 Activation-AP-1 is another transcription factor that inhibits cell proliferation and invasion by reg-ulating the gene expression (42,43). Various agents, such as TNF, that activate NF-B can also activate AP-1 (42,43). This led us to test whether hCG could block AP-1 activation induced by TNF. Fig. 6 shows that treatment with TNF alone resulted in a time-dependent 4.8-fold activation of AP-1, and hCG treatment prevented this activation.
Signaling in the hCG Actions in MCF-7 Cells-Although hCG is known to use more than one second messenger system, cAMP/PKA is the most common among them. Consistent with the possibility that hCG uses cAMP as a second messenger, hCG treatment of MCF-7 cells resulted in an increase in cAMP secretion (Fig. 7A). This response was abolished by pretreatment of cells with antisense hCG/LH receptor ODN, which decreased receptor protein levels. The treatment with sense ODN, which had no effect on the receptor protein, did not influence the ability of hCG to increase cAMP secretion.
Since the cAMP increase results in PKA activation, the hCG effect on PKA activity was next investigated. As shown in Fig.  7B, treatment of MCF-7 cells with hCG resulted in a significant increase in PKA activity. This increase was not seen in antisense hCG/LH receptor ODN-treated cells. Sense ODN-treated cells, on the other hand, responded just as the hCG-treated cells, as compared with control cells.
Further efforts to determine the cAMP involvement in the hCG action revealed that dibutyryl cAMP, a stable cAMP analog, mimicked hCG in blocking TNF-dependent NF-B activation in a dose-dependent manner (Fig. 8A). In addition, although dideoxyadenosine, an adenylate cyclase inhibitor, had no effect on its own nor could it affect the TNF-dependent activation of NF-B, it was able to block the ability of hCG to prevent NF-B activation in a dose-dependent manner (Fig. 8B). DISCUSSION Breast tissue and cells from a number of species contain hCG/LH receptors (6 -13). Recently, human male breast tissue has also been shown to contain these receptors. 2 The physiological significance of breast receptors is largely unknown. It is possible, however, that breast LH/hCG receptors may play a role in the growth and differentiation in concert with ovarian steroid hormones. Consistent with this possibility, knockout of LH receptors resulted in a gross underdevelopment of mammary glands (44). The development of glands in knockout animals can be stimulated by estradiol and progesterone replacement therapy; however, the branching pattern was not restored (44). This finding suggests that LH and hCG might possibly control branching morphogenesis in mammary glands.
Epidemiological studies demonstrated a decrease in breast cancer risk in women who complete full-term pregnancy at a young age (30 -33). Since pregnancy is associated with the changes in numerous hormones, it was not clear which hormone(s) could be responsible for this protection. The studies on the rodent model of carcinogen-induced mammary tumor formation revealed that it could be hCG because its administration before, during, or after carcinogen treatment prevented or moderated mammary tumor formation (34).
Until mammary glands were shown to contain hCG/LH receptors, the beneficial effects of hCG in rodent mammary tumor model were thought to be due to stimulation of ovaries to secrete higher levels of estradiol, progesterone, and inhibin (15)(16)(17)34). Although it is possible that some of the beneficial effects of hCG treatment could be mediated by the changing levels of these hormones (45,46), the mere presence of receptors that hCG could activate suggests that some of the effects could also come from the direct mammary gland actions of hCG. Supporting the possibility of direct actions, hCG treatment can decrease breast cell proliferation (7,8,16,22), ER␣ levels (7), IGF-1 (21), and HOXAl transcripts (10) and increase cell differentiation (19), gap junctional protein, connexin-26 (24), IGFBP (21) apoptosis index, and apoptosis-related proteins such as TRMP2, interleukin-1 ␤-converting enzyme, bCl- XS, c-Myc, and P 53 (20,23), inhibin (16,17), DNA repair mechanisms (25), and acetylated histones (28), and activate or inhibit gene expression (17,21,22). Moreover, hCG treatment enhances radiosensitivity of MCF-7 cells (29), and treatment with doxorubicin-hCG conjugate results in a selective decrease in the number of viable breast cancer cells (12). All of these findings strongly support the possibility that hCG was at least partly responsible for pregnancy-induced protection against breast cancer.
The present study investigating the anti-breast cancer effects of hCG in MCF-7 cells demonstrated an inhibition of both cell proliferation and cell invasion in a dose-dependent manner. The effect on cell invasion, however, was much more dramatic and occurred at a time when the effect on cell proliferation was modest to nonexistent. Both of these effects were abrogated in cells in which hCG/LH receptor synthesis was inhibited, suggesting that the receptor activation is required for the hCG effects. The recent evidence demonstrating that women with hCG/LH receptor-positive tumors had a longer metastasis-free survival (13) supports the premise from the present findings that the hCG-induced protection against breast cancer may primarily involve preventing tumor cell invasion.
NF-B and AP-1 are transcription factors, and hCG is known to inhibit their activation in a number of nongonadal cell types (42). This precedent, as well as the anti-breast cancer actions of hCG (10,(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29), led us to test whether hCG can inhibit their activation by TNF, which stimulates cell proliferation and invasion. The results showed that hCG treatment inhibited their activation by TNF. As expected, NF-B consisted of p50 (NF-B1) and p65 (RelA) proteins (data not shown), and its activation was associated with an inhibition of phosphorylation and increase in the degradation of IB␣. Since NF-B and AP-1 activate the transcription of genes involved in promoting cell proliferation and invasion (40 -43), it is possible that hCG treatment inhibited their expression in addition to activating the ones involved in cell differentiation, apopotosis, DNA repair, etc.
The studies on signaling revealed that hCG treatment in- In A, the cells (2 ϫ 10 6 /ml) were incubated for 1 h at 37°C with different concentrations of dibutyryl cAMP followed by an addition of 0.1 nM TNF and continuation of incubation for another 30 min. In B, the cells were incubated for 2 h at 37 C in the presence or absence of increasing concentrations of dideoxyadenosine followed by an addition of 50 ng/ml hCG and continuation of incubation for another 24 h and then an addition of 0.1 nM TNF and incubation for another 30 min. Then NF-B activation was measured in nuclear extracts. The numbers at the bottom represent -fold changes as compared with control (no db cAMP or TNF in panel A and no hCG, TNF, or dideoxyadenosine in panel B), which was set at 1.0. creased cAMP levels and PKA activity in MCF-7 cells. Both of these responses were abolished in antisense hCG/LH receptor ODN-treated cells, suggesting that receptors are required for hCG to activate the cAMP/PKA signaling pathway. If cAMP mediates the hCG action, then it should be able to mimic hCG. Consistent with this expectation, treatment with dibutyryl cAMP prevented TNF-dependent NF-B activation, just as hCG did. Finally, if adenylate cyclase activation is required for the hCG effect, then its inhibition by dideoxyadenosine should prevent hCG from inhibiting TNF-induced NF-B activation. Again, this expectation was met by the demonstration that dideoxyadenosine, which had no effect on its own or on the ability of TNF to activate NF-B, prevented the hCG effect. These findings, along with the data that demonstrate that cAMP arrests the growth of 7,12-dimethyl benz(a)anthraceneinduced rat mammary carcinoma (47,48), cAMP mimics ovariectomy-induced mammary tumor regression (47)(48)(49), and cAMP inhibits breast cancer cell proliferation (50,51), provide strong support for the cAMP/PKA pathway mediating the anticancer actions of hCG in MCF-7 cells.
The actions of hCG can sometimes be pro-breast cancer. For example, hCG treatment seems to hasten breast cancer development in Her-2/neu oncogene overexpressed animals, 3 and hCG can stimulate breast cell growth in the presence of dehydroepiandrosterone (52). These data suggest a possibility that under some circumstances, hCG may enhance instead of retard the tumor progression. This scenario may potentially explain the exacerbation of breast cancer in pregnant women.
There is evidence that breast cancer tissues and cells synthesize hCG and LH, suggesting paracrine and autocrine actions of these hormones (53)(54)(55). These findings suggest that strategies which can increase the local production of these hormones may work in a fight against this malignancy.
In summary, human breast cancer MCF-7 cells contain hCG/LH receptors, and their activation results in modest inhibition of cell proliferation and a dramatic inhibition of cell invasion and prevents TNF-dependent NF-B and AP-1 activation. These effects of hCG are mediated by the cAMP/PKA signaling pathway. The anti-proliferative and anti-invasive effects may be useful in developing therapeutic strategies with hCG to prevent and/or treat breast cancer in women. Using hCG is physiological, nontoxic, cost-effective, and probably comes with fewer side effects, if any.