Cyclooxygenase Inhibitors Suppress Aromatase Expression and ...

0021-972X/05/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 90(5):2563–2570 Copyright © 2005 by The Endocrine Society doi: 10.1210/jc.2004-2029

Cyclooxygenase Inhibitors Suppress Aromatase Expression and Activity in Breast Cancer Cells Edgar S. Dı´az-Cruz, Charles L. Shapiro, and Robert W. Brueggemeier Division of Medicinal Chemistry and Pharmacognosy (E.S.D.-C., R.W.B.), College of Pharmacy; and Division of Hematology and Oncology (C.L.S.), Department of Internal Medicine, College of Medicine, The Ohio State University, Columbus, Ohio 43210 Estradiol is biosynthesized from androgens by the aromatase enzyme complex. Previous studies suggest a strong association between aromatase (CYP19) gene expression and the expression of cyclooxygenase (COX) genes. Our hypothesis is that higher levels of COX-2 expression result in higher levels of prostaglandin E2, which, in turn, increases CYP19 expression through increases in intracellular cAMP levels. This biochemical mechanism may explain the beneficial effects of nonsteroidal antiinflammatory drugs on breast cancer. The effects of nonsteroidal antiinflammatory drugs, COX-1 and COX-2 selective inhibitors on aromatase activity and expression were studied in human breast cancer cells. The data from these experiments revealed dose-dependent decreases in aro-

A

DENOCARCINOMA OF THE breast is the most common cancer in women in the United States and ranks second only to lung cancer as a cause of cancer-related mortality. An estimated 215,990 new cases of invasive breast cancer were expected to occur among women in the United States during 2004, and an estimated 40,110 deaths were anticipated from breast cancer in 2004 (1). Approximately 60% of all breast cancer patients have hormone-dependent breast cancer, which contains estrogen receptors and requires estrogen for tumor growth. Estrogens are biosynthesized from androgens by aromatase, the product of the CYP19 gene and a member of the cytochrome P450 enzyme superfamily (2). The concentration of estradiol, the most potent endogenous estrogen, is higher in the tumor tissue than in the normal areas of the breast (3). Aromatase transcript expression (4) and activity (3, 5) in the tumor are greater than that in the normal breast tissue. Regulation of aromatase expression in human tissues is quite complex, involving alternative promoter sites that provide tissue-specific control. In the normal breast cells, aromatase expression is primarily derived by the tissue-specific promoter I.4 for transcription, whereas expression from breast cancer patients switches from promoter I.4 to promoter I.3 and promoter II (6). These results suggest that promoters I.3 and II are the major promoters directing aromatase expression in breast cancer and surrounding stromal cells. This switch in First Published Online February 1, 2005 Abbreviations: COX, Cyclooxygenase(s); DMSO, dimethylsulfoxide; NSAIDs, nonsteroidal antiinflammatory drugs; PGE2, prostaglandin E2; PKA, protein kinase A; PKC, protein kinase C; RT, reverse transcription. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community.

matase activity after treatment with all agents. Real-time PCR analysis of aromatase gene expression showed a significant decrease in mRNA levels when compared with control for all agents. These results were consistent with enzyme activity data, suggesting that the effect of COX inhibitors on aromatase begins at the transcriptional level. Exon-specific realtime PCR studies suggest that promoters I.3, I.4, and II are involved in this process. Thus, COX inhibitors decrease aromatase mRNA expression and enzymatic activity in human breast cancer cells in culture, suggesting that these agents may be useful in suppressing local estrogen biosynthesis in the treatment of hormone-dependent breast cancer. (J Clin Endocrinol Metab 90: 2563–2570, 2005)

the regulatory mechanism of aromatase expression from normal breast tissue to cancerous tissue has been extensively investigated (4, 7, 8). Although promoter I.4 requires the synergistic actions of glucocorticoids and class I cytokines (9), promoters I.3 and II are both transactivated by protein kinase A (PKA) cAMP-dependent signaling pathways (8, 10). Prostaglandin G/H endoperoxide synthase, also known as cyclooxygenase (COX), is a key enzyme that catalyzes the conversion of arachidonic acid to prostaglandins. Two isoforms have been identified, COX-1 and COX-2 (11, 12). Although COX-1 is present at a constant level in most cells and tissues and is believed to play a housekeeping role, some studies have shown that COX-1 activity and expression is elevated in human breast cancer tumors (13, 14). Most studies have shown that COX-2 is present in breast cancer tissue samples but not in the normal breast tissue (14 –16). Prostaglandins produced by COX-2, predominantly prostaglandin E2 (PGE2), induce inflammation and are potent mediators of a number of signal transduction pathways that are implicated in cancer development. Concentrations of PGE2 in the normal tissue are lower than the levels detected in tumor and metastatic tissues (16, 17). Nonsteroidal antiinflammatory drugs (NSAIDs) are used to decrease inflammation by inhibiting COX. A growing body of experimental (18, 19) and epidemiological (19 –22) evidence suggests that the use of NSAIDs may decrease the risk of breast cancer. Aspirin (23) and flurbiprofen (24) were shown to reduce mammary tumorigenesis. SC-560, celecoxib, and indomethacin treatment resulted in statistically significant inhibition of tumor size in comparison with vehicle-treated control animals in a murine model of breast cancer (25). Celecoxib and ibuprofen produced striking reductions in the incidence of mammary cancer, tumor burden,

2563

2564

J Clin Endocrinol Metab, May 2005, 90(5):2563–2570

and tumor volume compared with those seen in the control group in animal models (26). Genetic and pharmacological evidence has shown that specific COX-2 inhibition is more effective than traditional NSAIDs in suppressing polyposis in the mouse (27). All of these findings indicate that COX are involved in the promotion of this type of cancer. The relationship between COX and aromatase was examined in a preliminary study of CYP19 gene expression with COX-1 and COX-2 gene expression in breast cancer patient specimens. Regression analysis using a bivariate model showed a strong linear association between the sum of COX-1 and COX-2 expression and CYP19 expression (15). Another study using immunohistochemistry staining for aromatase and COX-2 revealed a marked correlation between COX-2 and aromatase expression in tumor samples (28). The present study examines the activity and expression of aromatase in breast cancer cell lines after the exposure to nonselective and isozyme-selective COX inhibitors, to provide additional information on the association between aromatase, COX, and human breast cancer development. Materials and Methods Chemicals and reagents Radiolabeled [1␤-3H]androst-4-ene-3,17-dione was obtained from NEN Life Science Products (Boston, MA). The following compounds were purchased from Cayman Chemical (Ann Arbor, MI): niflumic acid, nimesulide, NS-398, SC-560, and SC-58125. The following compounds were purchased from Sigma (St. Louis, MO): ibuprofen, indomethacin, and piroxicam. Celecoxib was a gift from Dr. Ching-Shih Chen [The Ohio State University (OSU), College of Pharmacy, Columbus, OH]. Trypsin, TRIzol, and all enzymes were obtained from Invitrogen (Carlsbad, CA). Radioactive samples were counted on a LS6800 liquid scintillation counter (Beckman, Palo Alto, CA). Mixture 3a70B was obtained from Research Prospect International Corp. (Mount Prospect, IL).

Cell culture JAR, MCF-7, MDA-MB-231, and SK-BR-3 cell lines were obtained from American Type Culture Collection (Rockville, MD). Cell cultures were maintained in phenol red-free custom media [MEM, Earle’s salts, 1.5⫻ amino acids, 2⫻ nonessential amino acids, l-glutamine, and 1.5⫻ vitamins (Life Technologies, Inc., Carlsbad, CA)] supplemented with 10% fetal bovine serum, 2 mm l-glutamine, and 20 mg/liter gentamycin. Fetal calf serum was heat inactivated for 30 min in a 56 C water bath before use. Cell cultures were grown at 37 C in a humidified atmosphere of 5% CO2 in a Hereaus CO2 incubator. For all experiments, cells were plated in either T-25 flasks or 100 mm plates and grown to subconfluency. Before treatment, the media was changed to a defined one containing DMEM/F12 media (Sigma) with 1.0 mg/ml human albumin (OSU Hospital Pharmacy, Columbus, OH), 5.0 mg/liter human transferrin, and 5.0 mg/liter bovine insulin.

Tritiated water-release assay Measurement of aromatase enzyme activity was based on the tritium water-release assay (29). Cells in T-25 flasks or 100 mm plates were treated with 0.1% dimethylsulfoxide (DMSO; control), NSAIDs (ibuprofen, piroxicam, and indomethacin), COX-1 selective inhibitor SC-560, and COX-2 selective inhibitors (SC-58125, NS-398, celecoxib, niflumic acid, and nimesulide) at the indicated concentrations. After 24 h, the cells were incubated for 6 h with fresh media along with 50 nm androstenedione including 2 ␮Ci [1␤-3H]androst-4-ene-3,17-dione. Subsequently, the reaction mixture was removed, and proteins were precipitated using 10% trichloroacetic acid at 42 C for 20 min. After a brief centrifugation, the media was extracted three times with an equal amount of chloroform to extract unused substrate and additional dextran-treated charcoal.

Dı´az-Cruz et al. • Suppression of Aromatase by COX Inhibitors

After centrifugation, a 250-␮l aliquot containing the product was counted in 5 ml of liquid scintillation mixture. Results were corrected for blanks and for the cell contents of culture flasks, and results were expressed as picomoles of 3H2O formed per hour incubation time per million live cells (pmol/h䡠106 cells). To determine the amount of live cells in each flask, the cells were trypsinized and analyzed using the diphenylamine DNA assay adapted to a 96-well plate (29, 30).

Enzyme immunoassay of PGE2 To study PGE2 synthesis in cell culture media, experiments were performed in 12-well plates. An aliquot of SK-BR-3 cells (150,000 cells) was added to each well, and plates were incubated overnight to allow the cells to adhere to the plates. After this time, cells were serum starved in defined media for 24 h. This step was followed by replacement of media with fresh media containing either vehicle (DMSO) or the indicated concentration of agents. After 24-h incubation at 37 C, the media were collected, and the amount of PGE2 was determined by ELISA (Cayman Chemical) according to the protocol provided by the manufacturer. PGE2 concentration was normalized to total protein. Total proteins were extracted from adhered cells by 30-min treatment with 0.5 m NaOH at room temperature and shaking. Protein concentrations in these extracts were determined using a protein assay method (Bio-Rad Laboratories, Inc., Hercules, CA).

RNA extraction Total RNA was isolated using the TRIzol reagent according to the manufacturer’s protocol. Total RNA pellets were dissolved in nucleasefree water and quantitated using a spectrophotometer. The quality of RNA samples was determined by electrophoresis through agarose gels and staining with ethidium bromide; the 18S and 28S rRNA bands were visualized under UV light.

cDNA synthesis Isolated total RNA (2 ␮g) was treated with DNase I, amplification grade, according to the recommended protocol, to eliminate any DNA before reverse transcription (RT). Treated total RNA was denatured at 65 C for 5 min in the presence of 2.5 ng/␮l random hexamers and 0.5 mm deoxynucleotide triphosphate mix. The samples were snap-cooled on ice and centrifuged briefly. cDNA was synthesized using Superscript II reverse transcriptase according to the recommended protocol. Briefly, the reactions were conducted in the presence of 1⫻ first-strand buffer and 20 mm dithiothreitol at 42 C for 50 min and consequently inactivated at 70 C for 15 min. The cDNA generated was used as a template in real-time PCR.

Real-time PCR Real-time PCR was performed using the Opticon 2 system from MJ Research (Waltham, MA). For the CYP19 total gene, the PCR mixture consisted of TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA), 600 nm of each primer (Invitrogen) (Table 1), 250 nm TaqMan probe, 18S rRNA (Applied Biosystems), and 2.5 ␮l of each RT sample in a final volume of 25 ␮l. The TaqMan probe was designed to anneal to a specific sequence of the aromatase gene between the forward and the reverse primers (Table 1). Cycling conditions were 50 C for 2 min and 95 C for 10 min, followed by 50 cycles at 95 C for 15 sec and 60 C for 1 min. For the specific exon I promoter regions and TATA-box-binding protein, the PCR mixture consisted of DyNAmo Hot Start SYBR Green qPCR kit (MJ Research), 600 nm of each primer (Table 1), and 2.5 ␮l of each RT sample in a final volume of 20 ␮l. SYBR Green uses a dye that will bind to double-stranded DNA. In this methodology, the primers are carefully designed to each of the promoter regions of aromatase exon I (Table 1). Cycling conditions were 95 C for 15 min, followed by 50 cycles at 94 C for 10 sec, 60 C for 25 sec, and 72 C for 30 sec.

Statistical analysis Statistical and graphical information were determined using GraphPad Prism software (GraphPad Software Inc., San Diego, CA) and Mi-



2565

TABLE 1. Oligonucleotide primer and probe sequences for real-time PCR Gene

CYP19 I.1a a

I.3

I.4b PIIb TBPa

Oligonucleotide

Primer Primer Probe Primer Primer Primer Primer Primer Primer Primer Primer Primer Primer

Sequences (5⬘–3⬘)

(S) (A)

TGT TCA 6FAM 5⬘-TGC TGT GGT GGG AGA AAC CAT CTC CAT TGC CAC

(S) (A) (S) (A) (S) (A) (S) (A) (S) (A)

CTC CCA AAA GCT TCA CTT GGG GTG CAC TGA CAC ACA ATC

TTT ATA GCA CGG GCA CCT GGC ACC CAG AGC CAG GGA ACA

GTT ACA CCC GAT TTT TGT AAT AAC CAT AAC CAT GCC GCT

CTT GTC TAA CTT CCA TTT TTA TGG CGT AGG CGT AAG CCC

CAT TGG TGT CCA AAA GAC GAG AGC GCC AGC GCC AGT CAC

GCT ATT TGA GAC CCA TTG TCT CTG TG TAT TG GAA CA

ATT TCT C TCC AGA GGC AAT-3⬘TAMRA TC TAA GTT

AGA T

Total CYP19 gene was analyzed using TaqMan methodology and a sequence-specific fluorogenic probe. Exon I promoter-specific regions and TBP gene were analyzed using SYBR Green methodology. S, Sense; A, antisense. a Ref. 36. b Ref. 37. crosoft Excel (Microsoft Corp., Redmond, WA). Determination of IC50 values was performed using nonlinear regression analysis. Statistically significant differences were calculated with the two-tailed unpaired Student’s t test, and P values were reported at 95% confidence intervals.

Results Aromatase enzymatic activity and expression in human breast cancer cell lines

Aromatase activity was determined using the “in-cell” tritiated water-release assay and normalized to the number of live cells in each flask. Different human breast cancer cell lines (MCF-7, MDA-MD-231, and SK-BR-3) were compared for aromatase activity using the tritiated water-release assay (Table 2). Cell line SK-BR-3 showed the highest levels of aromatase activity, followed by MCF-7 and MDA-MB-231. The aromatase activity in the SK-BR-3 cell line is approximately 20 times higher than that in MCF-7 cells and 35 times higher than that in MDA-MB-231 cells. These results agree with previous studies from other researchers (31–33). Cell lines JAR, MCF-7, MDA-MB-231, and SK-BR-3 were used to compare CYP19 gene expression by real-time PCR (Table 2). CYP19 expression was normalized relative to 18s rRNA and compared with expression in JAR cells. The choriocarcinoma placental JAR cell line, a cell line that expresses high levels of CYP19, was selected as the calibrator. Results show that basal levels of CYP19 gene expression in MCF-7 cells are slightly higher than those in SK-BR-3 cells but considerably higher than those in MDA-MB-231 cells. The discrepancy in aromatase activity and CYP19 mRNA expression

in MCF-7 cells has also been reported by other groups (33), and the mechanism is not well understood. For these studies, the SK-BR-3 cell line was selected as our breast cancer cell model because this cell line had the better ratio of aromatase activity and CYP19 expression than the other breast cancer cell lines. Decrease of aromatase enzymatic activity

The effects of NSAIDs, COX-1, and COX-2 selective inhibitors on aromatase activity were determined, and the concentration producing a 50% decrease in activity (IC50) for each agent was calculated. All agents decreased aromatase activity under control (vehicle) levels in a dose-dependent manner. NSAIDs (nonselective COX inhibitors) decreased aromatase activity only at high micromolar concentrations (Fig. 1). Indomethacin was the most potent NSAID in decreasing aromatase activity, followed by piroxicam and ibu-

TABLE 2. Aromatase activitya and expressionb in cell lines Cell line

Aromatase activity (pmol/h䡠106 cells)

CYP19 expression relative to JAR cells (2⫺⌬⌬Ct)

JAR MCF-7 SK-BR-3 MDA-MB-231

6.31 ⫾ 2.17c 0.00015 ⫾ 0.00004 0.0028 ⫾ 0.0002 0.00008 ⫾ 0.00001

1.0 ⫾ 0.1 0.12 ⫾ 0.02 0.07 ⫾ 0.01 0.004 ⫾ 0.001

Values are expressed as pmol/h䡠106 cells and reported as mean ⫾ (n ⫽ 3). b CYP19 expression was normalized relative to 18S rRNA. Values are expressed as CYP19 expression relative to a calibrator (JAR cells) and reported as mean ⫾ SD (n ⫽ 9). c Ref. 38. a

SD

FIG. 1. Suppression of aromatase activity in SK-BR-3 breast cancer cells by NSAIDs and COX-1 selective inhibitor. SK-BR-3 cells were treated with indomethacin (E), piroxicam (F), ibuprofen (f), or SC560 (⽧), and aromatase activity was measured as described in Materials and Methods. Values are expressed as picomoles 3H2O formed per hour incubation time per million live cells. The results were normalized against a control treatment with vehicle. The value of 100% is equal to 0.003 pmol/h䡠106 cells. Each data point represents the mean results of three independent determinations.

2566


profen, with IC50 values 157 ⫾ 5.7, 408 ⫾ 23, and 809 ⫾ 90 ␮m, respectively. Treatment of SK-BR-3 cells for 24 h with SC-560, a COX-1 selective agent, resulted in a decrease of aromatase activity and an IC50 value of 5.8 ⫾ 1.2 ␮m (Fig. 1). PGE2 is a powerful stimulator of aromatase activity and expression in human breast adipose stromal cells (29). Administration of exogenous PGE2 (1 ␮m) resulted in a 1.4-fold increase in aromatase activity (0.0040 ⫾ 0.0005 pmol/h䡠106 cells). To examine the hypothesis that prostaglandin inhibition could result in aromatase activity suppression, SK-BR-3 cells were treated for 24 h with COX-2 selective inhibitors. In fact, all COX-2 selective inhibitors decreased aromatase activity in SK-BR-3 cells in a dose-dependent manner (Fig. 2). NS-398 was the most potent agent in decreasing aromatase activity, with an IC50 value of 1.0 ⫾ 0.4 ␮m. Celecoxib and SC-58125, a celecoxib-like agent, showed IC50 values of 37 ⫾ 4.5 and 24 ⫾ 1.8 ␮m, respectively. Nimesulide and niflumic acid also decreased aromatase activity with IC50 values of 27 ⫾ 4.7 and 97 ⫾ 8.3 ␮m, respectively. Aromatase activity was expressed as picomoles of 3H2O formed per hour incubation time per million live cells. Some agents (such as ibuprofen, indomethacin, and niflumic acid) showed some cell toxicity at very high concentrations, but aromatase activity was normalized by the number of live cells in each assay, to assure that the loss of cells due to any toxicity effects by the agents was taken into account. Levels of PGE2 production

The production of PGE2 was measured in cells treated with NSAIDs, a COX-1 selective inhibitor, and COX-2 selective inhibitors (Fig. 3). The levels of COX activity in SK-BR-3 cells are low but detectable, consistent with other studies (34). Although exogenous arachidonic acid would have increased


FIG. 3. Effect of cyclooxygenase inhibitors on PGE2 production of SK-BR-3 cells. Cells were treated for 24 h with the indicated concentrations of agents. Results are expressed as means of the concentration of PGE2 produced per microgram protein ⫾ SEM. *, P ⬍ 0.05 vs. control by unpaired t test (n ⫽ 6).

the production of PGE2, the experimental conditions required the SK-BR-3 cells to use the endogenous substrate. This assures that the conditions for the rest of the assays are more consistent and comparable to one another. SK-BR-3 cells were treated for 24 h with the indicated concentration of the agents. All agents resulted in a decrease in PGE2 production in SK-BR-3 cells, but only ibuprofen, piroxicam, indomethacin, celecoxib, nimesulide, NS398, and SC-58125 resulted in significant reductions (P ⬍ 0.05). CYP19 mRNA expression by real-time PCR

FIG. 2. Suppression of aromatase activity in SK-BR-3 breast cancer cells by COX-2 selective inhibitors. SK-BR-3 cells were treated with NS-398 (⽧), nimesulide (E), SC-58125 (f), celecoxib (F), or niflumic acid (䡺), and aromatase activity was measured as described in Materials and Methods. Values are expressed as picomoles 3H2O formed per hour incubation time per million live cells. The results were normalized against a control treatment with vehicle (DMSO). The value of 100% is equal to 0.003 pmol/h䡠106 cells. Each data point represents the mean results of three independent determinations.

Analysis of total CYP19 mRNA transcripts was performed using real-time PCR to determine whether the decrease in aromatase activity by COX inhibitors in SK-BR-3 cells was due to a down-regulation of aromatase expression at the transcriptional level. SK-BR-3 cells were treated with COX inhibitors for 24 h at concentrations at or near the IC50 value for each agent, to assure that the concentration was sufficient to suppress aromatase activity and not result in cell death. In fact, cell cytotoxicity assays showed that the concentrations used for this study were safe and not toxic for the cells. Total RNA was extracted at 24 h, and CYP19 transcript levels were compared with control (vehicle) treatment. NSAIDs and COX-2 selective inhibitors significantly decreased CYP19 gene expression in SK-BR-3 cells relative to the control (vehicle) treatment (Fig. 4A). The COX-1 selective inhibitor, SC-560, also resulted in a significant decrease in expression, suggesting that COX-1 might also be involved in the mechanism. No effect on the expression level of the housekeeping gene 18S rRNA was observed with any of the agents. These results support the aromatase activity results, suggesting that COX inhibitors have an effect on aromatase. To determine whether similar results were observed in other cell



2567

cells, EP1 and EP2. EP2 is coupled to stimulation of cAMP formation, whereas EP1 is coupled to protein kinase C (PKC) activation (35). This suggests that PGE2 is capable of activating both PKA- and PKC-mediated signaling pathways. Exon-specific real-time PCR was performed to determine whether these agents specifically affected expression through promoters I.3 and II, which are the two aromatase promoters directly involved in PKA- and PKC-activation through cAMP. In a separate experiment, administration of exogenous PGE2 (1 ␮m) resulted in elevated transcript levels of CYP19 mRNA through promoter II by approximately 1.5-fold when compared with control in SK-BR-3 cells. This confirms the results observed in the tritiated water-release assay when SK-BR-3 cells were treated with exogenous PGE2 (1 ␮m). The NSAIDs, COX-1 and COX-2 selective inhibitors all demonstrated decreases in aromatase expression specific for promoter II (Fig. 5A). Aromatase expression is also driven through promoter I.3, the other promoter that is directly linked to the CYP19 switch in aromatase expression in breast tumors. The agents tested (COX-1 and COX-2 selective inhibitors) showed significant decreases in aromatase expression through this promoter (Fig. 5B). To study the possibility of other promoter regions being involved in the proposed mechanism, CYP19 promoter I.1- and I.4-specific mRNA expression was analyzed. As expected, these agents had no effect on aromatase expression specific for I.1 (Fig. 5C). Surprisingly, on the other hand, these agents showed a decrease in aromatase expression specific for promoter I.4 (Fig. 5D). Discussion

FIG. 4. Real-time RT-PCR analysis of CYP19 mRNA expression in SK-BR-3 (A) and MCF-7 and MDA-MB-231 cells (B). Cells were treated for 24 h with the indicated concentrations of agents, and total RNA was isolated. Results are expressed as means of CYP19 (normalized to 18S rRNA) ⫾ SEM. *, P ⬍ 0.05 vs. control by unpaired t test (n ⫽ 9).

lines, MCF-7 and MDA-MB-231 cells were treated with selected COX inhibitors. Figure 4B shows similar decreases in CYP19 gene expression in both MCF-7 and MDA-MB-231 cells. CYP19 exon-specific mRNA expression by real-time PCR

PGE2 is a powerful stimulator of aromatase activity and expression in human breast adipose stromal cells (29, 31, 35). PGE2 interacts with two receptor subtypes in adipose stromal

Aromatase is a key enzyme in the synthesis of estrogens (2) and plays an important role in the process of breast carcinogenesis of hormone-dependent breast cancers (3). COX has also been found to play a key role in this process (13, 14). Furthermore, PGE2 increases aromatase activity through increases in cAMP (29, 31, 35), and COX enzymes are involved in the synthesis of prostaglandins. To study a possible mechanism explaining why NSAIDs exert chemopreventive and antitumor properties in human breast cancers, SK-BR-3 cells were treated with these agents and evaluated for aromatase activity and expression. Aromatase activity was decreased in SK-BR-3 cells by NSAIDs and COX selective inhibitors treatment in a dose-dependent manner. COX selective agents are more effective in suppressing aromatase activity, with significantly lower IC50 concentrations than those required for NSAIDs (or nonselective COX inhibitors). Treatment of SK-BR-3 cells with the various agents at concentrations at or near IC50 values resulted in a decreased production of PGE2, which supports our hypothesis that PGE2 may be involved in CYP19 regulation. Real-time PCR analysis of aromatase gene expression demonstrated that changes in mRNA expression were consistent with enzyme activity data. The data from these experiments showed a significant decrease in mRNA levels when compare to control (vehicle) for all agents. The COX-1 selective inhibitor SC-560 resulted in a significant decrease in aromatase activity and expression in SK-BR-3 cells, suggesting that COX-1 may also produce prostaglandins that are involved in this

2568



FIG. 5. Real-time RT-PCR analysis of CYP19 exon-specific mRNA expression in SK-BR-3 cells. Promoter II (A), exon I.3 (B), exon I.1 (C), and exon I.4 (D). Cells were treated for 24 h with the indicated concentrations of agents, and total RNA was isolated. Results are expressed as means of CYP19 (normalized to TATA-box-binding protein) ⫾ SEM. *, P ⬍ 0.05 vs. control by unpaired t test (n ⫽ 9).

mechanism. Although COX-1 has also been implicated in human breast cancers (13, 14), its levels are relatively constant. Aromatase expression in MCF-7 and MDA-MB-231 cells is low compared with SK-BR-3 cells, and treatment of these cells with COX inhibitors also resulted in a decrease in aromatase expression. These results show that COX inhibitors decrease aromatase mRNA expression and that these drugs may be potent therapeutic agents in the treatment of breast cancer. The effect of NSAIDs and COX selective inhibitors on the exon I-specific promoters for aromatase also was investigated using real-time PCR. SK-BR-3 cells contain high abundance of promoter I.1, followed by promoters II and I.4, and relatively low abundance of promoter I.3. When SK-BR-3 cells were treated with NSAIDs, COX-1 and COX-2 selective inhibitors resulted in decreases in aromatase expression

through promoter II. NS-398 showed the most significant effect followed by SC-58125, consistent with the enzyme activity data. Promoter I.3 is the other promoter involved in the promoter switch from normal breast to cancerous cells (6). Decreases in aromatase expression through promoter I.3 were observed with the COX-2-specific inhibitors and the COX-1-specific inhibitor SC-560 as well, although the levels of promoter I.3 in SK-BR-3 cells are low. As expected, aromatase expression through promoter I.1 was not affected by any of the agents studied. Promoter I.4 was then studied, and, once again, all agents decreased aromatase expression through this promoter as well. The effect was not as significant as promoter II; nevertheless, CYP19 expression through promoter I.4 was affected when treating cells with NSAIDs and COX-specific inhibitors. Expression via promoter I.4 requires the action of


glucocorticoids and class I cytokines or TNF␣ (9). Other studies have shown that PGE2 regulates TNF␣ at the mRNA and protein level. The pleiotropic effect of prostaglandins, in general, may be playing a key role in this process. It is possible that the decrease in prostaglandin production by COX-specific inhibitors results in alterations of other biochemical pathways within these cells affecting glucocorticoid and/or class I cytokine action. This would result in a decrease in CYP19 expression through promoter I.4. Additional studies are underway to better understand this process. These results suggest that PGE2 produced via COX may act locally in an autocrine fashion to increase the local biosynthesis of estrogen by the aromatase enzyme in hormonedependent breast cancer development and lead to growth stimulation. In previous preliminary studies conducted in human placental microsomes, NSAIDs and the COX-2-specific inhibitors NS-398 and celecoxib failed to directly inhibit aromatase activity. These agents showed signs of aromatase activity inhibition only at very high micromolar concentrations. Thus, the effect of NSAIDs and the COX selective inhibitors does not occur through the direct inhibition of the enzyme, but rather as a result of suppressing gene expression. Based on our results, we propose the following model to explain the interrelationship between aromatase and COX enzymes. Higher levels of expression of COX enzymes and increased COX activity would result in higher levels of PGE2. Elevated PGE2 levels increase intracellular cAMP and result in increased aromatase expression via promoters I.3 and II. Higher levels of aromatase would lead to higher levels of estrogens, resulting in increased growth and development of the tumor by both paracrine and autocrine actions. Several studies have demonstrated that certain NSAIDs can cause antiproliferative effects independent of COX activity and prostaglandin synthesis inhibition. The fact that the COX-1-specific inhibitor SC-560 did not significantly inhibit the production of PGE2 suggests that other mechanisms may also be involved in this process. Various NSAIDs can influence the expression of certain transcription factors and other pathway mediators that could result in the suppression of aromatase. Elucidation of the interrelationship between aromatase and COX in breast carcinogenesis could facilitate targeting the enzymes as an effective strategy to prevent and treat breast cancer. We demonstrated that COX inhibitors decrease mRNA expression and aromatase enzyme activity through promoters I.3, I.4, and II. These results support the importance of prostaglandins such as PGE2. Thus, the breast cancer tissue microenvironment can influence the extent of estrogen biosynthesis and metabolism, resulting in altered levels of hormonally active estrogens and therefore influencing breast tumor development and growth. Furthermore, COX selective inhibitors suppress CYP19 expression through promoters I.3 and II in a tissue-selective manner, without affecting CYP19 expression in other tissues that use different promoters, such as brain and bone. Therapy with currently available aromatase inhibitors and COX inhibitors would affect several of these target pathways. The combination of an aromatase inhibitor with a COX inhibitor, such as a


2569

NSAID or COX selective inhibitor, may increase efficacy beyond the present treatments for postmenopausal hormone-dependent breast cancer. Acknowledgments Received October 13, 2004. Accepted January 20, 2005. Address all correspondence and requests for reprints to: Robert W. Brueggemeier, College of Pharmacy, The Ohio State University, 500 West 12th Avenue, Columbus, Ohio 43210. E-mail: Brueggemeier.1@ osu.edu. This work was supported by the National Institutes of Health (NIH) Grant R01 CA73698 (to R.W.B.), the NIH Chemistry and Biology Interface Training Program Grant T32 GM08512 (to E.S.D.-C.), and The Ohio State University Comprehensive Cancer Center Breast Cancer Research Fund.

References 1. American Cancer Society 2004 Cancer facts and figures. Atlanta: American Cancer Society 2. Simpson ER, Mahendroo MS, Means GD, Kilgore MW, Hinselwood MM, Graham-Lorence S, Amarneh B, Ito Y, Fisher CR, Michael MD, Mendelson CR, Bulun SE 1994 Aromatase cytochrome P450, the enzyme responsible for estrogen biosynthesis. Endocr Rev 15:342–355 3. Chetrite GS, Cortes-Prieto J, Philippe JC, Wright F, Pasqualini JR 2000 Comparison of estrogen concentrations, estrone sulfatase and aromatase activities in normal, and in cancerous, human breast tissues. J Steroid Biochem Mol Biol 72:23–27 4. Agarwal VR, Bulun SE, Leitch M, Rohrich R, Simpson ER 1996 Use of alternative promoters to express the aromatase cytochrome P450 (CYP19) gene in breast adipose tissues of cancer-free and breast cancer patients. J Clin Endocrinol Metab 81:3843–3849 5. Miller WR, Mullen P, Sourdaine P, Watson C, Dixon JM, Telford J 1997 Regulation of aromatase activity within the breast. J Steroid Biochem Mol Biol 61:193–202 6. Harada N, Utsumi T, Takagi Y 1993 Tissue-specific expression of the human aromatase cytochrome P-450 gene by alternative use of multiple exons 1 and promoters, and switching of tissue-specific exons 1 in carcinogenesis. Proc Natl Acad Sci USA 90:11312–11316 7. Harada N 1997 Aberrant expression of aromatase in breast cancer tissues. J Steroid Biochem Mol Biol 61:175–184 8. Bulun SE, Noble LS, Takayama K, Michael MD, Agarwal V, Fisher C, Zhao Y, Hinshelwood MM, Ito Y, Simpson ER 1997 Endocrine disorders associated with inappropriately high aromatase expression. J Steroid Biochem Mol Biol 61:133–139 9. Zhao Y, Nichols JE, Valdez R, Mendelson CR, Simpson ER 1996 Tumor necrosis factor-␣ stimulates aromatase gene expression in human adipose stromal cells through use of an activating protein-1 binding site upstream of promoter 1.4. Mol Endocrinol 10:1350 –1357 10. Zhou D, Clarke P, Wang J, Chen S 1996 Identification of a promoter that controls aromatase expression in human breast cancer and adipose stromal cells. J Biol Chem 271:15194 –15202 11. Smith WL, Garavito RM, DeWitt DL 1996 Prostaglandin endoperoxide H synthases (cyclooxygenases)-1 and -2. J Biol Chem 271:33157–33160 12. Herschman HR 1994 Regulation of prostaglandin synthase-1 and prostaglandin synthase-2. Cancer Metastasis Rev 13:241–256 13. Hwang D, Scollard D, Byrne J, Levine E 1998 Expression of cyclooxygenase-1 and cyclooxygenase-2 in human breast cancer. J Natl Cancer Inst 90:455– 460 14. Parrett ML, Harris RE, Joarder FS, Ross M, Clausen KP, Robertson FM 1997 Cyclooxygenase-2 gene expression in human breast cancer. Int J Oncol 10: 503–507 15. Brueggemeier RW, Quinn AL, Parrett ML, Joarder FS, Harris RE, Robertson FM 1999 Correlation of aromatase and cyclooxygenase gene expression in human breast cancer specimens. Cancer Lett 140:27–35 16. Badawi AF, Badr MZ 2003 Expression of cyclooxygenase-2 and peroxisome proliferator-activated receptor-␥ and levels of prostaglandin E2 and 15-deoxy␦12,14-prostaglandin J2 in human breast cancer and metastasis. Int J Cancer 103:84 –90 17. Bennett A, Charlier EM, McDonald AM, Simpson JS, Stamford IF, Zebro T 1977 Prostaglandins and breast cancer. Lancet 2:624 – 626 18. Robertson FM, Parrett ML, Joarder FS, Ross M, Abou-Issa HM, Alshafie G, Harris RE 1998 Ibuprofen-induced inhibition of cyclooxygenase isoform gene expression and regression of rat mammary carcinomas. Cancer Lett 122:165– 175 19. Harris RE, Namboodiri KK, Farrar WB 1996 Nonsteroidal antiinflammatory drugs and breast cancer. Epidemiology 7:203–205 20. Khuder SA, Mutgi AB 2001 Breast cancer and NSAID use: a meta-analysis. Br J Cancer 84:1188 –1192

2570


21. Moorman PG, Grubber JM, Millikan RC, Newman B 2003 Association between non-steroidal anti-inflammatory drugs (NSAIDs) and invasive breast cancer and carcinoma in situ of the breast. Cancer Causes Control 14:915–922 22. Sharpe CR, Collet JP, McNutt M, Belzile E, Boivin JF, Hanley JA 2000 Nested case-control study of the effects of non-steroidal anti-inflammatory drugs on breast cancer risk and stage. Br J Cancer 83:112–120 23. Mori H, Sugie S, Rahman W, Suzui N 1999 Chemoprevention of 2-amino1-methyl-6-phenylimidazo [4,5-b]pyridine-induced mammary carcinogenesis in rats. Cancer Lett 143:195–198 24. McCormick DL, Moon RC 1983 Inhibition of mammary carcinogenesis by flurbiprofen, a non-steroidal antiinflammatory agent. Br J Cancer 48:859 – 861 25. Kundu N, Fulton AM 2002 Selective cyclooxygenase (COX)-1 or COX-2 inhibitors control metastatic disease in a murine model of breast cancer. Cancer Res 62:2343–2346 26. Harris RE, Alshafie GA, Abou-Issa H, Seibert K 2000 Chemoprevention of breast cancer in rats by celecoxib, a cyclooxygenase 2 inhibitor. Cancer Res 60:2101–2103 27. Oshima M, Dinchuk JE, Kargman SL, Oshima H, Hancock B, Kwong E, Trzaskos JM, Evans JF, Taketo MM 1996 Suppression of intestinal polyposis in Apc ␦716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 87:803– 809 28. Brodie AM, Lu Q, Long BJ, Fulton A, Chen T, Macpherson N, Dejong PC, Blankenstein MA, Nortier JW, Slee PH, van de Ven J, van Gorp JM, Elbers JR, Schipper ME, Blijham GH, Thijssen JH 2001 Aromatase and COX-2 expression in human breast cancers. J Steroid Biochem Mol Biol 79:41– 47 29. Richards JA, Brueggemeier RW 2003 Prostaglandin E2 regulates aromatase activity and expression in human adipose stromal cells via two distinct receptor subtypes. J Clin Endocrinol Metab 88:2810 –2816


30. Natarajan N, Shambaugh 3rd GE, Elseth KM, Haines GK, Radosevich JA 1994 Adaptation of the diphenylamine (DPA) assay to a 96-well plate tissue culture format and comparison with the MTT assay. Biotechniques 17:166 –171 31. Brueggemeier RW, Richards JA, Petrel TA 2003 Aromatase and cyclooxygenases: enzymes in breast cancer. J Steroid Biochem Mol Biol 86:501–507 32. Kinoshita Y, Chen S 2003 Induction of aromatase (CYP19) expression in breast cancer cells through a nongenomic action of estrogen receptor ␣. Cancer Res 63:3546 –3555 33. Zhou D, Wang J, Chen E, Murai J, Siiteri PK, Chen S 1993 Aromatase gene is amplified in MCF-7 human breast cancer cells. J Steroid Biochem Mol Biol 46:147–153 34. Gilhooly EM, Rose DP 1999 The association between a mutated ras gene and cyclooxygenase-2 expression in human breast cancer cell lines. Int J Oncol 15:267–270 35. Zhao Y, Agarwal VR, Mendelson CR, Simpson ER 1996 Estrogen biosynthesis proximal to a breast tumor is stimulated by PGE2 via cyclic AMP, leading to activation of promoter II of the CYP19 (aromatase) gene. Endocrinology 137: 5739 –5742 36. Girault I, Lerebours F, Tozlu S, Spyratos F, Tubiana-Hulin M, Lidereau R, Bieche I 2002 Real-time reverse transcription PCR assay of CYP19 expression: application to a well-defined series of post-menopausal breast carcinomas. J Steroid Biochem Mol Biol 82:323–332 37. Ellem SJ, Schmitt JF, Pedersen JS, Frydenberg M, Risbridger GP 2004 Local aromatase expression in human prostate is altered in malignancy. J Clin Endocrinol Metab 89:2434 –2441 38. Brueggemeier RW, Gilbert NE, Gu X, O’Reilly JM, Lovely CJ 1997 Aromatase inhibition in JAr choriocarcinoma cells by 7␣-arylaliphatic androgens. J Steroid Biochem Mol Biol 61:73–77

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.