EP2 and EP4 Receptors Regulate Aromatase Expression in Human Adipocytes and Breast Cancer Cells

Cytochrome P450 aromatase (aromatase), a product of the CYP19 gene, catalyzes the synthesis of estrogens from androgens. Because aromatase-dependent estrogen biosynthesis has been linked to hormone-dependent breast carcinogenesis, it is important to elucidate the mechanisms that regulate CYP19 gene expression. The main objective of this study was to identify the receptors (EP) for prostaglandin E2 (PGE2) that mediate the induction of CYP19 transcription in human adipocytes and breast cancer cells. Treatment with PGE2 induced aromatase, an effect that was mimicked by either EP2 or EP4 agonists. Antagonists of EP2 or EP4 or small interference RNA-mediated down-regulation of these receptors suppressed PGE2-mediated induction of aromatase. PGE2 via EP2 and EP4 stimulated the cAMP→protein kinase A pathway resulting in enhanced interaction between P-CREB, p300, and the aromatase promoter I.3/II. Overexpressing a mutant form of p300 that lacks histone acetyltransferase activity suppressed PGE2-mediated induction of aromatase promoter activity. PGE2 via EP2 and EP4 also caused a reduction in both the amounts of BRCA1 and the interaction between BRCA1 and the aromatase promoter I.3/II. Activation of the aromatase promoter by PGE2 was suppressed by overexpressing wild-type BRCA1. Silencing of EP2 or EP4 also blocked PGE2-mediated induction of the progesterone receptor, a prototypic estrogen-response gene. In a mouse model, overexpressing COX-2 in the mammary gland, a known inducer of PGE2 synthesis, led to increased aromatase mRNA and activity and reduced amounts of BRCA1; these effects were reversed by knocking out EP2. Taken together, these results suggest that PGE2 via EP2 and EP4 activates the cAMP→PKA→CREB pathway leading to enhanced CYP19 transcription and increased aromatase activity. Reciprocal changes in the interaction between BRCA1, p300, and the aromatase promoter I.3/II contributed to the inductive effects of PGE2.

Estrogen plays a significant role in the development and progression of breast cancer. Cytochrome P450 aromatase, encoded by the CYP19 gene, catalyzes the synthesis of estrogens from androgens (1). In postmenopausal women, peripheral aromatization in adipose tissue is largely responsible for estrogen production, and in particular mammary adipose tissue is considered an important local estrogen source. In breast cancers, aromatase activity is increased in both tumor and stromal cells (2), suggesting a role in tumor growth (3). In fact, estrogen deprivation is a commonly used approach for both the prevention and treatment of hormone-dependent breast cancer (4,5).
Given the significance of estrogen synthesis in the pathogenesis of hormone-dependent breast cancer, it is important to elucidate the mechanisms that regulate the transcription of CYP19. The expression of CYP19 and thereby aromatase activity is regulated by several distinct promoters (6 -8). In normal breast adipose tissue, the aromatase CYP19 gene is expressed at low levels under the control of promoter I. 4, whereas increased expression occurs in breast cancers and cancer-proximal adipose tissue predominantly from promoters I. 3 and II (7,8). The proximal promoters I. 3 and II are located close to each other and are activated by stimulation of the cAMP3protein kinase A (PKA) 2 3cAMP-responsive element-binding protein (CREB) pathway (9 -12). Several findings suggest a key role for cyclooxygenase (COX)-derived prostaglandin E 2 (PGE 2 ) in stimulating CYP19 transcription via this pathway. PGE 2 is a potent inducer of CYP19 transcription via a cAMP-dependent mechanism in vitro (9,10,13). Positive correlations have been detected between COX and aromatase expression in human breast cancer specimens (14 -16). In mice genetically engineered to overexpress COX-2 in the mammary gland, increased levels of PGE 2 and aromatase were found (17). Finally, use of aspirin, an inhibitor of COX-mediated PGE 2 production, was associated with a reduced risk of hormone receptor-positive but not hormone receptor-negative breast cancer (18). Recently, the tumor suppressor, BRCA1, was found to negatively regulate CYP19 expression via selective inhibition of promoters 1.3 and II (19). This inhibitory effect of BRCA1 on aromatase activity was relieved by PGE 2 .
Chronic use of selective COX-2 inhibitors, prototypic inhibitors of PG synthesis, has been associated with an increased risk of cardiovascular complications (20,21). The mechanism underlying this toxicity is not fully understood, but inhibition of COX-2 results in the loss of all downstream PGs. It has been suggested that selective COX-2 inhibitors block the production of cardioprotective PGI 2 by vascular endothelium, without inhibiting COX-1-dependent platelet thromboxane A 2 synthesis, supporting a pro-thrombotic mechanism (22). To improve the therapeutic index, alternate treatment strategies are being explored. One possibility is to block the actions of PGE 2 rather than to inhibit its production. PGE 2 exerts its effects by binding to G protein-coupled receptors that contain seven transmembrane domains. Four subtypes of PGE 2 receptor (EP 1 , EP 2 , EP 3 , and EP 4 ) have been cloned and defined pharmacologically (23). Different EP receptors have been implicated in regulating cell proliferation, immune function, and angiogenesis (23,24), but little is known about the regulation of aromatase. Hence, the primary objective of this study was to identify the EP receptor(s) that mediate the induction of aromatase by PGE 2 in human adipocytes and breast cancer cells. We show that PGE 2 via EP 2 and EP 4 activates the cAMP3 PKA3 CREB pathway resulting in enhanced CYP19 transcription and increased aromatase activity. This was due, at least in part, to reciprocal changes in FIGURE 1. EP 2 is important for PGE 2 -mediated induction of aromatase in human adipocytes. A, cells were treated with the indicated concentration of PGE 2 ; B-D, cells were treated with the indicated concentration of butaprost, an EP 2 receptor agonist; E and F, cells were treated with vehicle, 500 nM PGE 2 , or 500 nM PGE 2 plus the indicated concentration of AH6809, an EP 2 receptor antagonist. All treatments were for 24 h. In A, B, and E, aromatase activity was determined as under "Experimental Procedures." Enzyme activity is expressed as femtomoles/g of protein/min. Means Ϯ S.D. are shown, n ϭ 6. In A and B, *, p Ͻ 0.001 versus vehicle-treated cells; E, *, p Ͻ 0.01 versus PGE 2 -treated cells. In C, D, and F, total RNA was prepared. In C and F, 10 g/lane RNA was subjected to Northern blotting. The blots were hybridized sequentially with the indicated probes. In D and F, aromatase mRNA was analyzed by real-time PCR as described under "Experimental Procedures." Values were normalized to the expression levels of glyceraldehyde-3-phosphate dehydrogenase. In D and F, means Ϯ S.D. are shown, n ϭ 3. D, *, p Ͻ 0.01 versus vehicle-treated cells and in F, *, p Ͻ 0.01 versus PGE 2 -treated cells.
the interaction between BRCA1, p300, and the aromatase promoter I.3/II.
Tissue Culture-Visceral adipocytes were obtained from ScienCell TM Research Laboratories. These primary cells were alcohol plus ONO AE3-208 or PGE 1 alcohol plus AH6809 as indicated. All treatments were for 24 h. In A, D, and F, aromatase activity was determined as under "Experimental Procedures." Enzyme activity is expressed as femtomoles/g of protein/min. Means Ϯ S.D. are shown, n ϭ 6. *, p Ͻ 0.01. In B, C, and E, total RNA was prepared and in B and E, 10 g/lane was subjected to Northern blotting. The blots were hybridized sequentially with the indicated probes. In C and E, aromatase mRNA was analyzed by real-time PCR. Values were normalized to the expression levels of glyceraldehyde-3-phosphate dehydrogenase. In C and E, means Ϯ S.D. are shown, n ϭ 3. C, *, p Ͻ 0.01 versus vehicle-treated cells and in E, *, p Ͻ 0.01 versus PGE 2 -treated cells.
grown in adipocyte medium containing 10% fetal bovine serum. In experiments requiring transfection, cells were transiently transfected using a system from Amaxa (Gaithersburg, MD). SKBR3 cells were purchased from ATCC and grown in McCoy's 5a medium (modified) with 1.5 mM L-glutamine adjusted to contain 2.2 g/liter sodium bicarbonate, and 10% fetal bovine serum. Experimental treatments of adipocytes and SKBR3 cells were carried out under serum free conditions. Animals-Generation of the mice used in this study has been described in detail previously (27). Briefly, MMTV-COX-2 transgenic mice on an FVB/N background were crossed with Ep2 Ϫ/Ϫ mice in the C57BL/6J background. The resulting MMTV-COX-2 Ep2 ϩ/Ϫ mice in the (FVB/N and C57BL/6) background were backcrossed with FVB/N mice more than nine times. The resulting MMTV-COX-2 Ep2 ϩ/Ϫ female mice were crossed with Ep2 ϩ/Ϫ male mice in same strain yielding the different experimental groups. Mammary tissues were harvested, snap-frozen in liquid nitrogen and stored at Ϫ80°C until analysis.
Northern Blotting-Prior to RNA extraction, frozen tissue was homogenized. Total RNA was prepared from mammary tissues and cell monolayers using an RNA isolation kit from Qiagen. 10 g of total RNA/lane were electrophoresed in a formaldehyde-containing 1% agarose gel and transferred to nylon-supported membranes. EP 1 -EP 4 , BRCA1, aromatase and 18 S rRNA probes were labeled with [ 32 P]CTP by random priming. The blots were probed as described previously (17). All experiments were repeated three times, and representative Northern blot data are shown.
Real-time Reverse Transcription-PCR-Total RNA was isolated using TRIzol reagent and reverse transcribed with the Superscript III First-Strand Synthesis System (Invitrogen). For total aromatase mRNA, the forward and reverse primers were 5Ј-CACATCC-TCAATACCAGGTCC-3Ј and 5Ј-CAGAGATCCAGACTCG-CATG-3Ј, and the fluorescence-labeled probe was 5Ј-CCCTCA-TCTCCCACGGCAGATTCC-3Ј. For BRCA1, the forward and reverse primers were 5Ј-AAGAGGAACGGGCTTGGAA-3Ј, 5Ј-AAAATAATCAAGAAGAGCAAAGCATGGATTCAAA-CTTA-3Ј, and the fluorescence-labeled probe was 5Ј-CACAC-CCAGATGCTGCTTCA-3Ј. For glyceraldehyde-3-phosphate dehydrogenase mRNA (control), the forward and reverse primers were 5Ј-GAAGGTGAAGGTCGGAGTC-3Ј and 5Ј-GAA-GATGGTGATGGGATTTC-3Ј, and the fluorescence-labeled probe was 5Ј-CAAGCTTCCCGTTCTCAGCC-3Ј. Real-time FIGURE 3. EP 2 and EP 4 receptors mediate the induction of aromatase by PGE 2 . A, adipocytes were transfected with 2 g of siRNAs to GFP or EP 1-4 and allowed to grow for 36 h prior to analysis. In B, cells were transfected with 2 g of siRNAs to GFP or EP 1-4 . 36 h later, cells were treated with vehicle (control) or 500 nM PGE 2 for 24 h. In A and B, total RNA was prepared from cells and subjected to Northern blotting (10 g/lane). The blots were hybridized sequentially with the indicated probes. In B, aromatase activity was determined in cell lysates as under "Experimental Procedures." Enzyme activity is expressed as femtomoles/g of protein/min. Means Ϯ S.D. are shown, n ϭ 6. *, p Ͻ 0.01 versus PGE 2 -treated cells.

FIGURE 4. PGE 2 via EP 2 and EP 4 stimulates the cAMP3PKA3CREB pathway resulting in induction of aromatase in adipocytes.
A and B, adipocytes were transfected with 2 g of siRNAs to GFP, EP 2 , or EP 4 . C, cells received 0.9 g of siRNA to GFP, EP 2 , or EP 4 and 0.9 g of CYP19 promoter-luciferase. All cells also received 0.2 g of pSV␤gal. In A-C, 36 h after transfection, cells received fresh medium containing vehicle (control) or 500 nM PGE 2 . Treatments were for 24 h. Subsequently, cellular levels of cAMP (A) and PKA activity (B) were determined. Means Ϯ S.D. are shown, n ϭ 6. *, p Ͻ 0.01 versus PGE 2 -treated cells. In C, luciferase activity was measured in cell lysates, and the activities represent data that have been normalized to ␤-galactosidase activity. Means Ϯ S.D. are shown, n ϭ 6. *, p Ͻ 0.01 compared with PGE 2 plus siRNA GFP. D, ChIP assays were performed. Top, cells were treated as indicated with vehicle (C), PGE 2 (500 nM), butaprost (1 M), or PGE 1 alcohol (0.2 M). Middle, cells were treated as indicated with vehicle, PGE 2 (500 nM), PGE 2 plus EP 2 antagonist (50 M AH6809), or PGE 2 plus EP 4 antagonist (0.1 M ONO AE3-208). Bottom, cells were transfected with 2 g of siRNAs to GFP, EP 2 , or EP 4 . 36 h after transfection, cells received fresh medium containing vehicle (C) or 500 nM PGE 2 . All treatments were for 2 h. Chromatin fragments were immunoprecipitated with antibodies against phospho-CREB, and the CYP19 I.3/II promoter was amplified by PCR (panel 1) or real-time PCR (panel 2). DNA sequencing was carried out, and the PCR product was confirmed to be the CYP19 I.3/II promoter. The CYP19 promoter was not detected when normal IgG was used or antibody was omitted from immunoprecipitation step (data not shown). For panel 2 (top, middle, and bottom), means Ϯ S.D. are shown, n ϭ 3. *, p Ͻ 0.01. PCR was carried out using an ABI Prism 7900 apparatus (Applied Biosystems).
Transient Transfections-Cells were seeded at a density of 5 ϫ 10 4 per well in six-well dishes and grown to 50 -60% confluence. For each well, 2 g of plasmid DNA was introduced into cells using 8 g of Lipofectamine as per the manufacturer's instructions. After 24 h of incubation, the medium was replaced with basal medium. The activities of luciferase and ␤-galactosidase were measured in cellular extract.
Western Blotting-Lysates were prepared by treating cells with lysis buffer (150 mM NaCl, 100 mM Tris (pH 8.0), 1% Tween 20, 50 mM diethyldithiocarbamate, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml trypsin inhibitor, and 10 g/ml leupeptin). Lysates were sonicated for 20 s on ice and centrifuged at 10,000 ϫ g for 10 min to sediment the particulate material. The protein concentration of the supernatant was measured by the method of Lowry et al. (28). SDS-PAGE was done under reducing conditions on 10% polyacrylamide gels as described by Laemmli (29). The resolved proteins were transferred onto nitrocellulose sheets as detailed by Towbin et al. (30). The nitrocellulose membrane was then incubated with primary antisera. Secondary antibody to IgG conjugated to horseradish peroxidase was used. The blot was probed with the ECL Western blot detection system according to the instructions of the manufacturer.
Measurements of cAMP Levels-Cells were plated at 5 ϫ 10 4 / well in six-well dishes and grown to 60% confluence before treatment. Amounts of cAMP were measured by enzyme immunoassay. Production of cAMP was normalized to protein concentration.
PKA Activity-Cells were plated at 5 ϫ 10 4 /well in six-well dishes and grown to 70% confluence before treatment. PKA activity was measured according to the instructions of the manufacturer. PKA activity was normalized to protein concentration.
Aromatase Assays-To determine aromatase activity, microsomes were prepared from cell lysates and mammary tissues by differential centrifugation using established methods (17). Briefly, tissue was sliced into small pieces and then ground in liquid nitrogen before being suspended in buffer (20 mM Tris-HCl (pH 7.5), 1 mM EDTA, 10% glycerol, 5 M pepstatin and 5 M leupeptin). The tissue lysate was then centrifuged for 20 min at 800 ϫ g, and the pellet containing the nuclear fraction was discarded. The supernatant was subjected to ultracentrifugation (1 h at 100,000 ϫ g) to separate microsomes from cytosol. To determine aromatase activity, microsomal protein was added to a 0.5-ml reaction mixture containing 50 mM Tris-HCl (pH 7.5), 5 mM MgCl 2 , 5 mM glucose 6-phosphate, 5 units of glucose-6-phosphate dehydrogenase, 2 M rotenone, and 1␤-[ 3 H]androstenedione (200 pM for microsomes for tissue, 12.5 nM for microsomes from cells). Following preincubation for 3 min, the reaction was initiated by the addition of 0.5 M NADPH and allowed to run for up to several hours at 37°C. Adding 3 ml of ice-cold chloroform, and applying vigorous shaking and brief centrifugation terminated the reaction. The resulting aqueous layer was further extracted with 3 ml of chloroform and treated with 0.5 ml of 5% activated charcoal/0.5% dextran. Following centrifugation of the mixture, the radioac-tivity in the supernatant was counted. Aromatase activity was quantified by measurement of the tritiated water released from 1␤-[ 3 H]androstenedione. The reaction was also performed in the presence of a specific aromatase inhibitor, as a specificity control and without NADPH as a background control. Aromatase activity was normalized to protein concentration.
Chromatin Immunoprecipitation Assay-ChIP assay was performed with a kit (Upstate) according to the manufacturer's instructions. 2 ϫ 10 6 cells were cross-linked in a 1% formaldehyde solution for 10 min at 37°C. Cells were then lysed in 200 l of SDS buffer and sonicated to generate 200-to 1000-bp DNA fragments. After centrifugation, the cleared supernatant was diluted 10-fold with ChIP buffer and incubated with 1.5 g of the indicated antibody at 4°C. Immune complexes were precipitated, washed, and eluted as recommended. DNA-protein cross-links were reversed by heating at 65°C for 4 h, and the DNA fragments were purified and dissolved in 50 l of water. 10 l of each sample was used as a template for PCR amplification. CYP19 oligonucleotide sequences for PCR primers were forward 5Ј-AAC CTG CTG ATG AAG TCA CAA-3Ј; and reverse, 5Ј-TCA GAC ATT TAG GCA AGA CT-3Ј (19). This primer set encompasses the CYP19 promoter I.3/II segment from nucleotide Ϫ302 to Ϫ38. PCR was performed at 94°C for 30 s, 60°C for 30 s, and 72°C for 45 s for 30 cycles. The PCR products generated from the ChIP template were sequenced, and the identity of the CYP19 promoter was confirmed. For real-time PCR analysis, ChIP-qPCR assay kits from Superarray Bioscience Corp. (Frederick, MD) were used. Real-time PCR was performed as described above. Statistics-Comparisons between groups were made by Student's t test. A difference between groups of p Ͻ 0.05 was considered significant.

EP 2 and EP 4 Are Important for PGE 2 -mediated Induction of
Aromatase Expression-Initially, we showed that PGE 2 caused dose-dependent induction of aromatase activity in human adipocytes (Fig. 1A). A pharmacological approach was then used to identify the EP receptors that mediated this inductive effect of PGE 2 . Similar to PGE 2 , butaprost, an EP 2 agonist, caused a dosedependent increase in aromatase activity and expression (Fig. 1,  B-D). Next we evaluated AH6809, a nonselective EP 2 receptor antagonist. This compound blocked PGE 2 -mediated induction of aromatase activity and expression (Fig. 1, E and F). Because both EP 2 and EP 4 signal via the cAMP3 PKA3 CREB pathway, we next evaluated the role of EP 4 in mediating the induction of aromatase. Treatment with PGE 1 alcohol, an EP 4 agonist, caused dose-dependent induction of aromatase activity and expression (Fig. 2, A-C). ONO AE3-208, an EP 4 antagonist, suppressed PGE 2 -mediated induction of aromatase activity and expression (Fig. 2, D and E). The inductive effects of the EP 2 receptor agonist (butaprost) were blocked by the EP 2 receptor antagonist (AH6809) but not by the EP 4 receptor antagonist (ONO AE3-208) (Fig. 2F). Similarly, the EP 4 antagonist blocked the inductive effects of PGE 1 alcohol but not butaprost.
To complement this pharmacological approach, a genetic strategy was employed. As shown in Fig. 3A, siRNAs to EP 2 and EP 4 suppressed levels of EP 2 and EP 4 , respectively. Consistent with the pharmacological findings, silencing of EP 2 or EP 4 blocked PGE 2 -mediated induction of aromatase expression (Fig. 3B) and activity (Fig. 3B). Although siRNAs to EP 1 and EP 3 suppressed levels of these EP receptors (Fig.  3A), we did not observe an effect on PGE 2 -mediated induction of aromatase expression and activity (Fig.  3B). Taken together, these data indicate that EP 2 and EP 4 mediate the inductive effects of PGE 2 on aromatase expression in human adipocytes.
Signal Transduction Pathway by which EP 2 and EP 4 Regulate CYP19 Transcription-As mentioned above, both EP 2 and EP 4 can signal via the cAMP3 PKA3 CREB pathway. Accordingly, we investigated whether this pathway was involved in PGE 2 -mediated induction of aromatase. The induction of cAMP levels and PKA activity by PGE 2 was suppressed by siRNAs to either EP 2 or EP 4 (Fig. 4, A and B). Moreover, siRNA to EP 2 or EP 4 blocked PGE 2 -mediated activation of the CYP19 promoter (Fig. 4C). ChIP assays were performed to evaluate the role of EP 2 and EP 4 in modulating the binding of p-CREB to the CYP19 promoter. Treatment with PGE 2 , an EP 2 agonist (butaprost) or an EP 4 agonist (PGE 1 alcohol) stimulated the association between p-CREB and the CYP19 promoter (Fig. 4D, top  panel). This interaction was blocked by pharmacological antagonists to EP 2 (AH6809) or EP 4 (ONO AE3-208) (Fig. 4D, middle panel) or siRNAs to these receptors (Fig. 4D, bottom panel).
Recently, PGE 2 was found to suppress levels of BRCA1 resulting in enhanced CYP19 expression (19). Here we attempted to identify the EP receptors that are responsible for these effects. First, we confirmed that PGE 2 -mediated suppression of BRCA1 mRNA and protein levels occurred over a concentration range that also induced aromatase (Fig. 5, A and B, and supplemental Fig. S1A). Similar to PGE 2 , both the EP 2 and EP 4 receptor agonists suppressed BRCA1 levels and induced

via EP 2 and EP 4 induces aromatase and suppresses levels of BRCA1 in human breast cancer cells.
A, SKBR3 cells were treated with the indicated concentration of PGE 2 , butaprost or PGE 1 alcohol. B, cells were treated with vehicle, 500 nM PGE 2 , or 500 nM PGE 2 plus 50 M AH6809, an EP 2 receptor antagonist, or 500 nM PGE 2 plus 0.1 M ONO AE3-208, an EP 4 receptor antagonist. In A and B, aromatase activity was determined as under "Experimental Procedures." Enzyme activity is expressed as femtomoles/g of protein/ min. Means Ϯ S.D. are shown, n ϭ 6. *, p Ͻ 0.001. C and D, cells were treated with the indicated concentration of butaprost, an EP 2 agonist, or PGE 1 alcohol, an EP 4 agonist. E, cells were treated with vehicle, 500 nM PGE 2 or PGE 2 plus the indicated concentration of AH6809. F, cells were treated with vehicle, 500 nM PGE 2 , or 500 nM PGE 2 plus the indicated concentration of ONO AE3-208. All treatments were for 24 h. In C-F, total RNA was prepared and 10 g/lane was subjected to Northern blotting. The blots were hybridized sequentially with the indicated probes. aromatase (Fig. 5, C and D; and supplemental Fig. S1, B and C). Consistent with these findings, the effects of PGE 2 on BRCA1 and aromatase were blocked by antagonists to EP 2 and EP 4 ( Fig.  5E and supplemental Fig. S1D) or silencing of these receptors ( Fig. 5F and supplemental Fig. S1E).
In breast cancer, increased aromatase activity occurs in tumor cells in addition to stromal cells (2). Hence, we also evaluated whether EP 2 and EP 4 were important for regulating aromatase expression and activity in a human breast cancer cell line (SKBR3 cells). Similar to the findings in adipocytes, treat- In A, bars labeled BRCA1 represent cells that also received 0.45 g of BRCA1 expression vector; bars labeled as BRCA1 mutant represent cells that also received 0.45 g of mutant BRCA1 expression vector. In C, bars labeled p300 represent cells that received 0.45 g of wild-type p300 expression vector, whereas bars labeled mutant p300 represent cells that received 0.45 g of mutant p300 (lacks HAT activity) expression vector. In A and C, the total amount of DNA received by cells is limited to 2 g by using vector DNA. 36 h after transfection, cells were treated with vehicle (control), 500 nM PGE 2 , EP 2 agonist (1 M butaprost) or EP 4 agonist (0.2 M PGE 1 alcohol) for another 24 h. Luciferase activity was measured in cell lysates, and the activities represent data that have been normalized to ␤-galactosidase activity. Means Ϯ S.D. are shown, n ϭ 6. *, p Ͻ 0.01 compared with same treatment that received vector alone. B, ChIP assays were performed. In panels 1-4, cells were treated with vehicle (C), PGE 2 (500 nM), butaprost (1 M), or PGE 1 alcohol (0.2 M) for 2 h. Chromatin fragments were immunoprecipitated with antibodies against BRCA1 or p300, and the CYP19 I.3/II promoter was amplified by PCR (panels 1 and 3) or real-time PCR (panels 2 and 4). DNA sequencing was carried out, and the PCR product was confirmed to be the CYP19 1.3/II promoter. The CYP19 I.3/II promoter was not detected when normal IgG was used or antibody was omitted from immunoprecipitation step (data not shown). For panels 2 and 4, means Ϯ S.D. are shown, n ϭ 3. *, p Ͻ 0.01 versus control. D, cells were treated with vehicle or 500 nM PGE 2 for 2h. Cell lysates (500 g) were subjected to immunoprecipitation with p300 antiserum and Western blotting was performed for BRCA1, p300 and phospho-CREB as indicated. These proteins were not immunoprecipitated with control IgG. Input is ␤-actin. ment with PGE 2 , an EP 2 receptor agonist (butaprost) or an EP 4 receptor agonist (PGE 1 alcohol) caused dose-dependent induction of aromatase activity and expression (Fig. 6, A, C and D, and supplemental Fig. S2, A and B). Both the EP 2 (AH6809) and EP 4 (ONO AE3-208) receptor antagonists suppressed PGE 2mediated induction of aromatase activity and expression (Fig.  6, B, E and F, and supplemental Fig. S2, C and D). Importantly, reciprocal changes in amounts of BRCA1 occurred (Fig. 6, C-F,  and supplemental Fig. S2).
To further investigate the role of BRCA1 in repressing CYP19 transcription, transient transfections experiments were carried out. The induction of CYP19 promoter activity by PGE 2 or agonists of EP 2 and EP 4 was suppressed by overexpressing wildtype BRCA1 but not a mutant form of BRCA1 that does not bind to DNA (Fig. 7A). ChIP assays were performed to further explore the interaction between BRCA1 and the CYP19 promoter. Treatment with PGE 2 or agonists of EP 2 and EP 4 suppressed the interaction between BRCA1 and the CYP19 promoter (Fig. 7B). Because p300 is important for CREB-dependent activation of gene expression, the interaction between p300 and the CYP19 promoter was also investigated. Treatment with PGE 2 or agonists of EP 2 and EP 4 increased the interaction between p300 and the CYP19 promoter (Fig. 7B). Hence, PGE 2 via EP 2 or EP 4 reduced the interaction between BRCA1 and the CYP19 promoter but caused a reciprocal increase in the interaction between p300 and the CYP19 promoter. To further evaluate the role of p300, transient transfections were carried out. The induction of CYP19 promoter activity by PGE 2 or agonists of EP 2 and EP 4 was suppressed by overexpressing a histone acetyltransferase mutant form of p300 (Fig. 7C). To further understand the transcriptional regulation of aromatase, we also investigated the interactions between BRCA1, p300, and p-CREB under basal conditions and following treatment with PGE 2 . In untreated cells, immunoprecipitation experiments suggested that BRCA1 and p300 were in a complex (Fig. 7D). Following treatment with PGE 2 , p300, and p-CREB were in the complex, but BRCA1 was not found.
In an effort to determine the in vivo relevance of these findings, an MMTV-COX-2 transgenic mouse model was employed. Previously, we utilized this model and demonstrated that overexpression of COX-2 led to both increased levels of intramammary PGE 2 and aromatase activity (17). This was confirmed in the current study; overexpression of COX-2 in the mammary gland also led to increased aromatase activity and expression (Fig. 8). To evaluate whether the EP 2 receptor contributed to PGE 2 -mediated induction of aromatase activity, MMTV-COX-2 EP 2 Ϫ/Ϫ mice were generated. Remarkably, when the EP 2 receptor was absent, the COX-2-mediated increase in aromatase activity and expression was abrogated (Fig. 8). Consistent with our in vitro findings, levels of BRCA1 were also reduced in the mammary glands of MMTV-COX-2 transgenic mice, an effect that was reversed by knocking out EP 2 (Fig. 8).
Because aromatase activity can be rate-limiting for the synthesis of estradiol, we also evaluated the role of EP 2 and EP 4 as determinants of estrogen-dependent gene expression. The progesterone receptor, an estrogen target gene, is positively regulated by an estrogen response element (31,32). As shown in Fig.   9, treatment with PGE 2 stimulated ERE-luciferase activity and induced the PR, effects that were suppressed by silencing of either EP 2 or EP 4 .

DISCUSSION
Previous studies indicate that COX-derived PGE 2 can activate CYP19 transcription leading to enhanced aromatase activity and possibly an increased risk of breast cancer (9,10,17,18,33). Recent evidence suggests that the induction of aromatase by PGE 2 is mediated, in part, by suppression of BRCA1, a repressor of CYP19 transcription (19). Given the established link between estrogen biosynthesis and the development and progression of hormone receptor-positive breast cancer, we have attempted to further elucidate the signaling pathway that mediates the activation of CYP19 transcription by PGE 2 . Several lines of evidence indicate that the EP 2 and EP 4 receptors are involved in mediating this inductive effect. First, agonists of EP 2 and EP 4 mimicked PGE 2 in inducing aromatase expression. Moreover, EP 2 and EP 4 receptor antagonists blocked PGE 2mediated induction of aromatase. Consistent with these pharmacological findings, silencing of EP 2 or EP 4 suppressed PGE 2 -mediated induction of aromatase expression and activity. Although both EP 2 and EP 4 play a role in mediating the induction of aromatase, we note that inhibiting either receptor abro- gated the stimulatory effects of PGE 2 . This finding suggests that EP 2 and EP 4 are functionally interdependent. There is growing evidence that G protein-coupled receptors can heterodimerize and that this physical interaction can affect the function of either receptor (34,35). Recently, functionally important heterodimerization of the thromboxane and prostacyclin receptors (34) and the EP 1 and ␤2-adrenergic receptors (35) was observed. Studies are underway to determine whether the current findings can be explained by heterodimerization of the EP 2 and EP 4 receptors.
Because activation of EP 2 or EP 4 induced aromatase activity, we also investigated the role of these receptors in regulating estrogen-dependent gene expression. Silencing of EP 2 or EP 4 suppressed PGE 2 -mediated induction of the progesterone receptor, a prototypic estrogen-response gene. The significance of these in vitro findings is supported by evidence that the increase in aromatase expression and activity in the mammary glands of COX-2 transgenic mice was abrogated by knocking out EP 2 . Because the survival of EP 4 -deficient mice is poor, the role of EP 4 as a determinant of aromatase expression was not assessed in vivo. A previous study suggested that EP 1 was also a determinant of aromatase expression in adipose stromal cells (36). Although it's possible that EP 1 signaling contributes to aromatase expression in this cell type, the earlier study did not utilize a genetic approach to confirm the results suggested by EP receptor agonists and antagonists. Our study was carried out in visceral adipocytes and a breast cancer cell line, because both cell types are believed to be relevant sources of estrogen in breast carcinogenesis. Whether our findings for EP 2 and EP 4 extend to other conditions, e.g. uterine leiomyomata, in which PGE 2 -mediated induction of aromatase appears to be important should be determined (37,38).
Because EP 2 and EP 4 can signal via the cAMP3 PKA3 CREB pathway (23), we investigated whether this signaling pathway was responsible for PGE 2 -mediated activation of CYP19 transcription. The induction of cAMP levels and PKA activity by PGE 2 was suppressed by silencing of EP 2 or EP 4 . Furthermore, the PGE 2 -mediated increase in binding of p-CREB to the CYP19 I.3/II promoter was suppressed by EP 2 and EP 4 antagonists or silencing of these receptors. Although CYP19 transcription is regulated by CREB (11), little is known about the potential involvement of the coactivator CBP/p300 in regulating aromatase expression (39). We present evidence that PGE 2 via EP 2 and EP 4 stimulated the interaction between p300 and the CYP19 I.3/II promoter. The interaction between p300 and pCREB was also enhanced by treatment with PGE 2 . The interaction between p300 and the CYP19 I.3/II promoter was functionally important, because overexpression of a p300 mutant that lacked histone acetyltransferase activity suppressed activation of the CYP19 I.3/II promoter by PGE 2 . Our observation that p300 is important in the regulation of CYP19 transcription provides mechanistic insights that may help to explain previous findings. More specifically, ligands of nuclear receptors, e.g. retinoids and PPAR-␥ ligands, have been reported to inhibit aromatase expression (40,41). Ligands can stimulate an interaction between nuclear receptors and CBP/ p300 (42,43) that limits, in turn, the availability of relatively low concentrations of CBP/p300 to interact with transcription factors and enhance gene expression. Hence, the current results provide the basis for future experiments to determine whether A, adipocytes were first transfected with 2 g of siRNAs to GFP, EP 2 , or EP 4 . 36 h later, fresh medium containing vehicle (C) or 500 nM PGE 2 was added for an additional 24 h. In the panel on the left, cell lysates were analyzed by Western blotting for progesterone receptor (PR) protein levels using purified PR as a standard (Std). In the panel on the right, total RNA was isolated and PR mRNA was analyzed by real-time PCR. Values were normalized to the expression level of glyceraldehyde-3-phosphate dehydrogenase. Means Ϯ S.D. are shown, n ϭ 3, *, p Ͻ 0.01 versus PGE 2 -treated cells. B, cells were transfected with 0.9 g of a ERE luciferase construct, 0.2 g of pSV␤gal, and 0.9 g of siRNAs to GFP, EP 2 , or EP 4 . Thirty-six hours after transfection, cells were treated with vehicle (control) or 500 nM PGE 2 for another 24 h. ERE luciferase activity represents data that have been normalized to ␤-galactosidase activity. Means Ϯ S.D. are shown, n ϭ 6, *, p Ͻ 0.01 compared with PGE 2 treatment.
squelching of CBP/p300 explains the ability of ligands of nuclear receptors to suppress the expression of aromatase.
The tumor suppressor BRCA1 plays a significant role in repressing aromatase expression (19,44,45). BRCA1 binds directly to the CYP19 I.3/II promoter region and suppresses transcription (19). Previously, PGE 2 and other agents that stimulate cAMP signaling were found to suppress amounts of BRCA1 resulting in enhanced CYP19 transcription (19,44). We extend upon these findings and show that PGE 2 -mediated suppression of BRCA1 is mediated by EP 2 and EP 4 . In fact, EP 2 and EP 4 agonists mimicked PGE 2 in suppressing levels of BRCA1 while inducing aromatase. Conversely, antagonists of EP 2 and EP 4 or silencing of these receptors blocked PGE 2 -mediated suppression of BRCA1. Additionally, PGE 2 via EP 2 and EP 4 caused a decrease in the interaction between BRCA1 and the CYP19 I.3/II promoter. The reduced interaction between repressor and promoter was functionally important because overexpression of wild-type but not mutant BRCA1 abrogated PGE 2 -mediated activation of the aromatase promoter. Consistent with these in vitro findings, overexpression of COX-2 in the murine mammary gland, a known inducer of intramammary PGE 2 (17), led to reduced levels of BRCA1 in association with increased aromatase expression; this effect was reversed by knocking out EP 2 . The mechanism by which PGE 2 inhibits the expression of BRCA1 is likely to be complex. One recent study that utilized a human ovarian granulosa cell line demonstrated proteasome degradation of BRCA1 after a cAMP surge (46). Possibly, PGE 2 via EP 2 and EP 4 acts, in part, by a similar mechanism to suppress levels of BRCA1 protein. Because PGE 2 caused a reduction in BRCA1 mRNA in addition to BRCA1 protein, additional levels of regulation must also exist.
Regardless of the mechanism by which stimulation of EP 2 or EP 4 suppresses BRCA1 levels, it's important to consider the potential implications of this finding. In addition to inhibiting aromatase expression, BRCA1 is involved in several important cellular processes, including DNA damage control, DNA repair, chromatin remodeling, and mitotic spindle formation (47). BRCA1 is also involved in transcriptional regulation through interactions with a number of transcription factors (48). Hence, the reduction in amounts of BRCA1 mediated by activation of EP 2 or EP 4 is likely to affect multiple mechanisms that reduce the threshold for carcinogenesis. Taken together, PGE 2 via EP 2 and EP 4 activates the cAMP3 PKA3 CREB pathway resulting in reciprocal changes in the interaction between BRCA1, p300, and the aromatase promoter I.3/II, which contributes, in turn, to enhanced CYP19 transcription and increased aromatase activity. Inhibitors of this pathway, including antagonists of EP 2 and EP 4 , may reduce the risk of breast cancer.