Purpose and Experimental Design: New therapies against cancer are based on targeting cyclooxygenase (COX)-2. Activation of the endothelin A receptor (ETAR) by endothelin (ET)-1 is biologically relevant in several malignancies, including ovarian carcinoma. In this tumor, the ET-1/ETAR autocrine pathway promotes mitogenesis, apoptosis protection, invasion, and neoangiogenesis. Because COX-1 and COX-2 are involved in ovarian carcinoma progression, we investigated whether ET-1 induced COX-1 and COX-2 expression through the ETAR at the mRNA and protein level in HEY and OVCA 433 ovarian carcinoma cell lines by Northern blot, reverse transcription-PCR, Western blot, and immunohistochemistry; we also investigated the activity of the COX-2 promoter by luciferase assay and the release of prostaglandin (PG) E2 by ELISA.
Results: ET-1 significantly increases the expression of COX-1 and COX-2, COX-2 promoter activity, and PGE2 production. These effects depend on ETAR activation and involve multiple mitogen-activated protein kinase (MAPK) signaling pathways, including p42/44 MAPK, p38 MAPK, and transactivation of the epidermal growth factor receptor. COX-2 inhibitors and, in part, COX-1 inhibitor blocked ET-1-induced PGE2 and vascular endothelial growth factor release, indicating that both enzymes participate in PGE2 production to a different extent. Moreover, inhibition of human ovarian tumor growth in nude mice after treatment with the potent ETAR-selective antagonist ABT-627 is associated with reduced COX-2 and vascular endothelial growth factor expression.
Conclusions: These results indicate that impairing COX-1 and COX-2 and their downstream effect by targeting ETAR can be therapeutically advantageous in ovarian carcinoma treatment. Pharmacological blockade of the ETAR is an attractive strategy to control COX-2 induction, which has been associated with ovarian carcinoma progression and chemoresistance.
Ovarian cancer represents the leading cause of death among gynecological malignancies. About 26,250 women are diagnosed yearly, with an overall 5-year survival rate of only 47%. Despite recent advances in surgery and chemotherapy, improvement in long-term survival of these patients has been slight (1) . The endothelin (ET) family is composed of three isopeptides, ET-1, ET-2, and ET-3, which are potent mitogens for several human tumors (2, 3, 4, 5, 6, 7, 8, 9) including ovarian carcinoma (3 , 4) . ETs and their receptors have been implicated in tumor progression through autocrine and paracrine pathways. ET-1 is produced by different tumor cells and is biologically active through two distinct subtypes of G protein-coupled receptors (GPCRs), namely, ET A receptor (ETAR) and ET B receptor, endowed with different affinities for ETs (10) . The ET-1/ETAR autocrine pathway plays a key role in the development and progression of prostatic, ovarian, and cervical carcinoma (11) . We have demonstrated previously (4) that ET-1 and ETAR are overexpressed in primary and metastatic ovarian carcinomas compared with normal ovarian tissues. In this tumor, engagement of ETAR by ET-1 promotes tumor cell proliferation (3 , 12) , apoptosis protection (13) , invasiveness (14) , and neovascularization stimulating vascular endothelial growth factor (VEGF) secretion (15 , 16) . ET-1 induces VEGF by increasing hypoxia-inducible factor (HIF)-1α accumulation and activity under both normoxic and hypoxic conditions (17) . Among the many growth-promoting factors known to be present in ovarian cancer ascites, ET-1 is found there at significant levels and may play an important role in the development or progression of ovarian cancer (16) . ET-1, acting through the ETAR, consistently induces the activity of multiple metastasis-related proteinases, such as the matrix-metalloproteinases and the urokinase-type plasminogen activator system, that stimulate cellular invasion (14) . Loss of gap junction intercellular communication is critical for tumor progression because it allows the tumor cells to escape growth control, invade, and metastasize. In this regard, ET-1 inhibits gap junction intercellular communication by inducing phosphorylation/inactivation of connexin 43 (18) . Interestingly, we found that the addition of ETAR antagonist blocked ET-1-induced migration and invasion of ovarian carcinoma cells (14) . These results indicate that the ET-1/ETAR pathway participates in the progression of ovarian carcinoma and could be targeted for antitumor therapy. Thus, ETAR blockade by a selective receptor antagonist, ABT-627, has been shown to inhibit the growth of ovarian carcinoma cell xenografts (19) .
Although ET-1 has also been shown to play a role in modulating cyclooxygenase (COX)-2 and COX-2-derived PGE2 expression in different normal cells (20, 21, 22, 23) , the molecular mechanisms controlling these effects have not been defined. Moreover, the role of ET-1 in the regulation of COX-1 and COX-2 in tumor cells has not been investigated.
The COX-1 and COX-2 enzymes catalyze the rate-limiting step in the conversion of arachidonic acid to PGs (24) and are involved in tumor progression by inducing proliferation, survival, invasion, and metastasis in several solid tumors (25) . It was reported recently (26, 27, 28, 29) that expression of COX-2 is elevated in ovarian cancer. Moreover, in ovarian cancer, elevated COX-2 expression has been identified as an independent prognostic factor (30) that is associated with reduced survival (31) and poor response to standard combination chemotherapy (32) . High COX-1 expression was also reported in ovarian cancer specimens, in which it is associated with elevated levels of proangiogenic factors, such as VEGF and HIF-1α (33) .
In view of these findings, we investigated whether activation of the ETAR by ET-1 leads to up-regulation of COX-2 and COX-1 expression and PGE2 production in two human ovarian carcinoma cell lines in which the ET-1/ETAR autocrine pathway is biologically active. Ligand binding to the ETAR results in activation of a pertussis toxin-insensitive G protein that stimulates phospholipase C activity and increases intracellular Ca2+ levels, activation of protein kinase C, mitogen-activated protein kinase (MAPK), paxillin, and p125 focal adhesion kinase (12) . Among the downstream events after ETAR activation, ET-1 also causes epidermal growth factor receptor (EGFR) transactivation, which is partly responsible for MAPK phosphorylation (34) . Because the activation of ETAR and EGFR autocrine circuits triggers mitogenic signaling in these tumor cells, we examined whether the ET-1-induced effects may derive from cross-communication between ETAR and EGFR signaling pathways, which are amplified in ovarian carcinoma. In this study, we report that (a) ET-1 significantly induces COX-2 and COX-1 expression and PGE2 production through ETAR activation, and (b) signaling progresses through multiple MAPK pathways including p42/44 and p38 MAPK and EGFR transactivation. Furthermore, we demonstrated that both ET-1-induced COX-1 and COX-2 might contribute, to a different extent, to neovascularization by stimulating VEGF levels. In addition, inhibition of human tumor growth in nude mice after treatment with the nonpeptide orally active ETAR antagonist ABT-627 (atrasentan) is associated with a reduction of in vivo COX-2 expression. Thus, the present findings indicate that pharmacologically blocking ETAR signaling may result in targeting COX-2 and related signaling cascades that contribute to ET-1-mediated ovarian cancer progression.
MATERIALS AND METHODS
Human ovarian carcinoma cell lines HEY and OVCA 433, previously characterized for ET-1 receptor expression and ET-1 production (2 , 3) , were cultured in DMEM containing 10% FCS. The cells were serum starved by incubation for 24 h in serum-free DMEM. All culture reagents were from Invitrogen (Paisley, Scotland, United Kingdom). ET-1 (Peninsula Laboratories, Belmont, CA) was added in the cell medium at the indicated concentration and for the indicated time. When the effects of the antagonists (BQ 123 and BQ 788; Peninsula Laboratories) were studied, they were added 15 min before agonists and left for the whole experiment. Pretreatment of cells with 10 μm PD98059, 5 μm SB203580, and 0.1 μm AG1478 (Calbiochem-Novabiochem Corp., San Diego, CA) was performed for 1 h, whereas pretreatment with 1 μm NS-398, 9 nm SC-560, 70 nm SC-58125, 26 μm indomethacin (Cayman, Chemical, Ann Arbor, MI) was performed for 30 min before the addition of ET-1. Epidermal growth factor (Cell Signaling, Beverly, MA) was used at 10 ng/ml.
Immunoprecipitation and Immunoblotting.
Precleared lysates were immunoprecipitated with anti-EGFR antibody (Cell Signaling) insolubilized on protein A-Sepharose (Amersham Pharmacia Biotech, Little Chalfont, United Kingdom). The blots were then incubated for 1 h with anti-phospho-tyrosine monoclonal antibody (PY20; BD Transduction Laboratories, Heidelberg, Germany) or anti-EGFR (Cell Signaling) and developed with an enhanced chemiluminescence detection system (ECL; Amersham Pharmacia Biotech). For detection of COX-2, COX-1, p42/44 MAPK, phospho-p42/44 MAPK, p38 MAPK, phospho-p38 MAPK, and VEGF, whole cell lysates or homogenized HEY tumor specimens were subjected to 7.5–12.5% SDS-PAGE and examined by Western blotting using anti-COX-2 or anti-COX-1 (1:1000; Cayman Chemical) or anti-p44/42 MAPK, anti-phospho-p44/42 MAPK, anti-p38 MAPK, anti-phospho-p38 MAPK (1:1000; Cell Signaling), or anti-VEGF (Sigma, St. Louis, MO). Blots were developed with enhanced chemiluminescence (Amersham Pharmacia Biotech). After being stripped, the membranes were reprobed with β-actin (Oncogene, CN Biosciences, Inc., Darmastadt, Germany).
Total RNA from HEY cells was extracted using the TRIzol (Invitrogen) method according to the manufacturer’s instructions. RNA samples (30 μg/lane) were separated by electrophoresis on a 2% denaturing formaldehyde agarose gel and transferred to a nylon membrane. The membranes were UV cross-linked and hybridized in QuickHyb hybridization solution (Stratagene, La Jolla, CA). The cDNA probe (Cayman Chemical) used for analysis of COX-2 mRNA was labeled with [α-32P]dCTP using a random primer oligolabeling kit (Amersham Pharmacia Biotech) according to the manufacturer’s instructions. To ensure RNA integrity and confirm equal loading between lanes, the filters were stripped and rehybridized with a probe for 18S rRNA. The relative intensity of signals was quantified using the Scion Image analysis program.
Reverse transcription-PCR was performed using a Superscript One-Step Reverse Transcription-PCR System (Invitrogen) according to the manufacturer’s instructions. Briefly, 1 μg of RNA was reverse transcribed. The primers sets were as follows: COX-1, 5′-TGCCCAGCTCCTGGCCCGCCGCTT-3′ and 5′-GTGCATCAACACAGGCGCCTCTTC-3′; and glyceraldehyde-3-phosphate dehydrogenase, 5′-TGAAGGTCGGTGTCAACGGA-3′ and 5′-GATGGCATGGACTGTGGTCAT-3′. Thirty-five cycles of amplification were performed under the following conditions: melting at 95°C for 30 s; annealing at 54°C for 45 s; and extension at 72°C for 30 s. The PCR products were analyzed by electrophoresis on a 2% agarose gel.
Cytospins of HEY cells and cryostat sections of tumor xenografts were fixed in cold acetone for 10 min, and COX-2 expression was detected by streptoavidin-biotin immunoperoxidase (Vector Laboratories, Burlingame, CA) using anti-COX-2 (Cayman Chemical). 3-Amino-9-ethilcarbazole was used as chromogenic substrate. Nuclear counterstain was obtained with Mayer’s hematoxylin. Negative control staining used reagent absorbed with relevant immunogenic peptide (Cayman Chemical).
The VEGF protein levels in the conditioned media were determined in duplicate by ELISA using the Quantikine Human VEGF immunoassay kit (R&D Systems, Minneapolis, MN). The sensitivity of the assay is <5.0 pg/ml. Intra-assay variation is 5.4%, and interassay variation is 7.3%. Levels of PGE2 released into the medium of treated cells were measured by ELISA using the PGE2 High Sensitivity Immunoassay according to the manufacturer’s instructions (R&D Systems). The sensitivity of the assay is <8.25 pg/ml. Intra-assay variation is 9.5%, and interassay variation is 10.9%.
Transfection of Reporter Construct and Luciferase Assay.
Reporter constructs phPES2 (−1432/+59, −372/+59, −220/+59, −124/+59) containing the 5′-flanking region of the human COX-2 gene (35) were kindly provided by Dr. R. N. DuBois (Vanderbilt University Medical Center, Nashville, TN). For transient transfection, 1 × 105 cells were plated in 6-well plates 48 h before transfection. The cells were cotransfected with 0.5 μg of COX-2 firefly luciferase plasmid construct and 0.05 μg of the pCMV-β-galactosidase plasmid (Promega, Madison, WI) using LipofectAMINE reagent (Invitrogen) as described in the manufacturer’s protocol. The cells were lysed, and their luciferase activities were measured (luciferase assay system; Promega). The results were normalized to β-galactosidase activity. For each experiment, the mean of three independent experiments performed in triplicate is reported.
HEY Xenografts in Nude Mice.
Female athymic (nu+/nu+) mice, 4–6 week of age, were purchased from Charles River Laboratories (Milan, Italy). The treatment protocol followed the guidelines for animal experimentation of the Italian Ministry of Health. Mice were given s.c. injection in one flank with 1.5 × 106 viable HEY cells, as determined by trypan blue staining, resuspended in 200 μl of PBS. After 7 days, when established tumors of approximately 0.2–0.3 cm3 in volume were detectable, mice were randomized into groups (n = 10) to receive different treatments. One group was treated i.p. for 21 days with 2 mg/kg/day ABT-627 (atrasentan; provided by Abbott Laboratories, Abbott Park, IL). Control mice were injected with drug vehicle. On day 40 after tumor injection, tumors were removed and snap frozen in liquid nitrogen for immunohistochemical and Western blot analyses.
All statistical analysis was assessed using a two-tailed Student’s t test and by two-way ANOVA as appropriate.
ET-1 Stimulates COX-1 and COX-2 Expression.
We first investigated whether ET-1 can regulate COX-1 and COX-2 expression in two ovarian carcinoma cell lines, HEY and OVCA 433, that express functional ETAR and secrete high levels of ET-1 (3 , 4) . ET-1 significantly induced time-dependent up-regulation of COX-1 and COX-2 mRNA in HEY cells. Northern blot analysis indicated that the steady-state COX-2 mRNA levels increased in ET-1-treated cells compared with control by 1-, 3.8-, 4-, 5.6-, and 2.8-fold at 2, 4, 6, 12, and 24 h, respectively, as quantified by the Scion Image analysis program (Fig. 1A)⇓ . Reverse transcription-PCR analysis showed that COX-1 mRNA levels were enhanced for increasing time periods (1–12 h) in ET-1-treated cells (Fig. 1A)⇓ . Moreover, ET-1 treatment evoked a time-dependent increase in COX-2 and COX-1 protein levels. Western blots exhibited low expression of both proteins in untreated HEY and OVCA 433 cells but showed a 2-fold increase in COX-2 and COX-1 protein by 2 h and a 4.5- and 5-fold increase in COX-2 and COX-1 protein, respectively, after 24 h of ET-1 stimulation (Fig. 1B)⇓ . ET-1 also increased COX-1 and COX-2 protein levels in a dose-dependent fashion. Treatment of HEY and OVCA 433 cells with 0.1 and 1 nm ET-1 for 24 h revealed 3.2- and 4.8-fold increases in COX-2 protein expression, respectively, reaching maximum responses (6.8-fold) at 10 and 100 nm ET-1. COX-1 protein levels increased 1.2-fold at 0.01 nm ET-1, reaching maximum levels (5-fold) at 100 nm ET-1 (Fig. 1C)⇓ . To examine whether the increase of COX-1 and COX-2 expression is associated with elevated PG synthesis, the level of PGE2 released from HEY cells was measured using ELISA. ET-1 increased PGE2 levels in the conditioned medium of HEY cells in a time-dependent manner (Fig. 1D)⇓ . To assess whether ET-1 up-regulates COX-2 cytoplasmic expression, HEY cells were analyzed by immunohistochemistry using anti-COX-2. Although a faint cytoplasmic stain was detected in untreated cells (Fig. 1E)⇓ , cells treated with ET-1 for 24 h displayed a significantly stronger and homogeneous immunoreactivity.
ET-1-Induced COX-1 and COX-2 Up-Regulation and PGE2 Production Are Mediated through ETAR.
To investigate which receptor subtype mediates ET-1-induced up-regulation of COX-1 and COX-2 expression, a selective ETA receptor antagonist, BQ 123, and a selective ET B receptor antagonist, BQ 788, were used in the presence or absence of 100 nm ET-1. BQ 123 was able to completely block ET-1-induced COX-1 and COX-2 expression, whereas BQ 788 did not (Fig. 2A)⇓ . To determine whether ETAR blockade was responsible for functional COX activity, PGE2 production was measured in HEY cell conditioned medium. As shown in Fig. 2B⇓ , BQ 123 completely prevented the PGE2 production induced by treatment with 100 nm ET-1, whereas BQ 788 did not exert any effect. Taken together, these findings indicate that ET-1 acts through the ETAR to stimulate COX-1 and COX-2 expression and PGE2 production in ovarian carcinoma cells.
ET-1 Stimulates COX-2 Promoter Activity.
COX-2 expression is controlled at transcriptional and posttranscriptional levels (35, 36, 37) . To determine whether ET-1 regulates COX-2 promoter activity, HEY cells were transfected with different-sized fragments of the human COX-2 promoter coupled with a luciferase reporter gene, and luciferase activity was measured and normalized to β-galactosidase activity. ET-1 (100 nm) induced a 3.1 ± 0.5-fold increase in luciferase activity in cells transiently expressing phPES2(−1432/+59) compared with untreated control cultures. A 2.0-fold induction was observed when using a shorter fragment of the COX-2 promoter phPES2(−327/+59). ET-1 did not stimulate an increase in luciferase activity in cells containing the shortest COX-2 promoter constructs phPES2(−220/+59) and phPES2(−124/+59) (Fig. 3A)⇓ . To study the time course of ET-1-induced COX-2 promoter, phPES2(−1432/+59)-transfected HEY cells were exposed to ET-1 for different times, and luciferase activity was measured (Fig. 3B)⇓ . After 4 h, the ET-1 treatment doubled (a 1.9 ± 0.2-fold increase) the luciferase activity of full-length COX-2 promoter-transfected cells with respect to the control and induced the maximum effect (3.0 ± 0.35-fold increase) at 12 h. Thereafter, this activity decreased, reaching a 1.5 ± 0.2-fold increase over the control. phPES2(−1432/+59) was also used to analyze the ET-1 receptor involved in ET-1-induced COX-2 promoter activity. When BQ 123 was added in combination with ET-1, induction of luciferase activity was fully prevented (from a 3.5 ± 0.2-fold increase to a 1.1 ± 0.2-fold increase; Fig. 3C⇓ ). A 3.5 ± 0.2-fold induction was observed when ET-1 was added to cell medium in combination with BQ 788, demonstrating that ETAR-mediated signaling pathways are linked to regulation of the COX-2 promoter in ovarian cancer cells (Fig. 3C)⇓ .
Signaling Pathways Involved in ET-1-Stimulated COX-2 Expression.
MAPK signal transduction pathways constitute one of the major mechanisms by which extracellular stimuli are converted into specific nuclear responses. ET-1 is known to activate the p42/44 MAPK pathway through the ETAR in ovarian carcinoma cell lines (12) . Moreover, ET-1 regulates COX-2 expression through p42/44 and p38 MAPK in vascular smooth muscle cells (23) and through p42/44 MAPK in osteoblast-like cells (21) . Therefore, we analyzed whether these pathways were involved in ET-1-induced COX-2 expression in HEY ovarian cancer cells. We found that ET-1 treatment stimulated the activation of MAPK (Fig. 4)⇓ . Western blots of HEY cell extracts showed that ET-1 (100 nm) induced a time-dependent increase in phosphorylated p42/44 MAPK that reached a peak after 10 min, before returning to baseline levels by 30 min (Fig. 4A)⇓ . ET-1 also increased the phosphorylation of p38 MAPK. Time course experiments revealed that phosphorylated p38 MAPK expression peaked at 10 min and returned to the baseline after 15 min of exposure to 100 nm ET-1 (Fig. 4B)⇓ . The selective inhibitors of MAPK/extracellular signal-regulated kinase kinase [MEK (PD98059)] and p38 MAPK (SB203580) significantly inhibited ET-1-stimulated COX-2 protein expression (Fig. 4C)⇓ and promoter activity (Fig. 4D)⇓ . Furthermore, in HEY cells, ET-1 (100 nm for 5 min) significantly increased tyrosine phosphorylation of EGFR to an extent similar to that induced by epidermal growth factor (Fig. 4E)⇓ . Treatment of HEY cells with the EGFR tyrosine kinase inhibitor, AG1478 (0.1 μm), significantly inhibited epidermal growth factor- and ET-1-induced EGFR activation (Fig. 4E)⇓ . BQ 123 was able to reduce the EGFR transactivation induced by ET-1 (Fig. 4E)⇓ . The p42/44 MAPK phosphorylation induced by ET-1 was significantly but incompletely reduced by AG1478, indicating that dual signaling pathways converge on p42/44 MAPK activation, one of which is dependent on EGFR kinase activity, and the other is independent of EGFR transactivation (Fig. 4F)⇓ . As shown in Fig. 4, G and H⇓ , AG1478 significantly inhibited the ET-1-induced COX-2 protein levels and COX-2 promoter activity. Taken together, these data indicate that transactivation of the EGFR, p38 MAPK, and p42/44 MAPK-dependent pathways are involved in ETAR-mediated regulation of COX-2 expression.
COX-1 and -2 Inhibitors Decrease ET-1-Induced PGE2 and VEGF Production.
Selective or nonselective COX inhibitors were used to discriminate between COX-2- and COX-1- derived PGE2 production after ET-1 stimulation. COX-2 inhibitors NS-398 (1 μm) and SC-58125 (70 nm), as well as the non-COX isotype-selective inhibitor indomethacin (26 μm), significantly blocked PGE2 production after 24 h of ET-1 (100 nm) treatment, whereas COX-1 inhibitor SC-560 (9 nm) partially decreased ET-1-induced PGE2 production (Fig. 5A)⇓ . These data indicate that both enzymes participate, to a different extent, in PGE2 production in these COX-1/COX-2-positive ovarian carcinoma cell lines. Several lines of evidence indicate that COX-derived PGE2 also plays an important role in tumor angiogenesis (38) . Therefore, we evaluated COX-2- and COX-1-mediated VEGF production in ET-1-stimulated ovarian cancer cells by ELISA. Pretreatment of HEY cells with both COX-2 inhibitors, NS-398 (1 μm) and SC-58125 (70 nm), indomethacin (26 μm), and COX-1 inhibitor (SC-560; 9 nm) resulted in a significant decrease of ET-1-induced VEGF production compared with untreated cells (Fig. 5B)⇓ , suggesting that both COX enzymes participate in the angiogenic effect of ET-1.
ETAR Antagonist-Induced Inhibition of Tumor Growth in HEY Xenografts Is Associated with Decreased COX-2 Expression.
To evaluate whether in vivo COX-2 is a downstream signal during ETAR/ET-1-induced tumor growth and progression, we examined the effect of a potent ETAR antagonist, ABT-627, on COX-2 expression in murine ovarian carcinoma xenografts. The treatment with 2 mg/kg/day dose of ABT-627 produced a 65% inhibition of HEY tumor growth on day 40 after tumor injection (P < 0.001 compared with control) and was generally well tolerated, with no detectable signs of acute or delayed toxicity (19) . Immunohistochemical evaluation of the expression of COX-2 on HEY tumors at day 40 after tumor cell injection revealed a significant and homogeneous decrease of cytoplasmic COX-2 staining in HEY tumor treated with ABT-627 (Fig. 6A)⇓ compared with tumors of untreated animals. Western blot analysis of COX-2 protein expression performed on HEY tumor xenografts revealed a marked reduction of COX-2 in ABT-627-treated mouse compared with control, which paralleled the ability of ABT-627 to inhibit VEGF expression (Fig. 6B)⇓ . These data demonstrate that ETAR antagonist prevents in vivo COX-2 expression and subsequent VEGF expression, indicating that blocking of the ETAR may contribute to tumor growth inhibition by also reducing COX-2-mediated angiogenesis.
Chemical carcinogenesis experiments and epidemiological and clinical studies have collectively identified PGs and their rate-limiting enzymes, COX-1 and COX-2, as molecules involved in the onset and progression of a variety of malignancies (39) . The development of selective inhibitors of COX-2 clearly adds a novel potential pharmacological target in cancer prevention and treatment. In view of this, studies aimed at identifying metabolic pathways involved in COX induction are of major relevance from a biological as well as a clinical point of view. Different growth factors that bind tyrosine kinase and GPCR have been shown to induce COX-2 expression (36 , 37) . The present study implicates for the first time COX-1 and COX-2 as downstream signals during the ETAR/ET-1 autocrine pathway, providing evidence that ET-1 drastically increases COX-2 promoter activity, COX-1 and COX-2 expression, and, in parallel, PGE2 levels. By using selective and nonselective COX inhibitors, we were able to discern the contribution of each COX enzyme to ET-1-induced PGE2 production. Furthermore, we dissected the signaling pathways mediating COX-2 induction by ET-1 in ovarian tumor cells using different pharmacological inhibitors.
Several signaling pathways, including p42/44 and p38 MAPK, have been implicated in the regulation of COX-2 expression (20 , 23 , 36) . Here we demonstrated that the MEK pathway inhibitor as well as the p38 MAPK inhibitor blocked ET-1-induced COX-2 promoter activity and COX-2 protein expression, indicating that ET-1-mediated effects are likely to be dependent on the MAPK pathway. Moreover, ET-1-induced COX-2 expression and activity require ligand-independent activation of EGFR, as demonstrated by the inhibitory effect exerted by the EGFR tyrosine kinase inhibitor AG1478, indicating that ET-1-induced effects related to COX-2 expression are also mediated by EGFR transactivation.
It should be noted that EGFR transactivation is not exclusively responsible for the mitogenic activity of ET-1 in ovarian carcinoma cells and that cross-talk with other signaling pathways could be relevant for ET-1-induced proliferation. Although tyrphostin AG1478 completely inhibits the activation of EGFR and COX-2 expression and activity, it only partially inhibits p42/44 MAPK phosphorylation in response to ET-1, indicating that an alternative pathway is also involved in ET-1-induced mitogenic signaling and p42/44 MAPK activation. Consistent with this finding, Guo et al. (36) demonstrated that like ET-1, gastrin, another GPCR agonist, stimulates COX-2 expression through multiple signaling pathways including EGFR transactivation in intestinal epithelial cells, thus identifying a mechanism involved in the initiation and progression of colorectal cancer.
Over the past few years, numerous studies have investigated the role of EGFR in GPCR signal transduction. Daub et al. (40) first identified EGFR as a critical element in ET-1-, lysophosphatidic acid-, and trombin-induced MAPK activation. EGFR transactivation is now recognized as a key mechanism that couples distinct signal transduction pathways to diverse cellular responses (41) . Several reports have demonstrated that GPCR-induced EGFR transactivation may be mediated via an intracellular signaling pathway. The cross-signaling mechanism leading to activation of the EGFR kinase may involve protein kinase C or the intracellular tyrosine kinase Src (42) . In our previous report (18) , we demonstrated that ET-1 is able to activate Src by inducing phosphorylation on Tyr418. In view of this, it is possible to speculate that this mechanism could be involved in ET-1-induced EGFR transactivation.
There is increased evidence that PGE2 contributes to tumor progression also by promoting tumor angiogenesis and that this effect is mediated, at least in part, by modulation of VEGF (38) . We reported previously (16) that ET-1 was expressed in 84% of the ovarian carcinomas, with a strong correlation between ET-1 and VEGF expression and vascularization. Moreover, we demonstrated that, in ovarian carcinoma cell lines, ET-1 stimulates VEGF expression through the ETAR and that this effect is mediated by HIF-1α accumulation and activation (16 , 17) . Here we demonstrated that in HEY and OVCA 433 cells, the ET-1-induced PGE2 release paralleled the increase in VEGF production, indicating that ET-1-induced VEGF expression may be regulated through a dual mechanism including HIF-1 activation and COX expression. The possibility that ET-1-induced up-regulation of VEGF may be regulated through coordinate interdependent mechanisms involving HIF-1 directly and indirectly through COX-2 expression and PGE2 production remains to be explored in the future.
The ability of COX-2-selective inhibitors to reduce ET-1-mediated PGE2 formation indicates that COX-2 plays a major role in PG production in these ovarian cancer cells. These results are further supported by the recent data of Denkert et al. (43) , who demonstrated that COX-2 is the major source of PGE2 in a different COX-1- and COX-2-positive ovarian carcinoma cell line. However, the ability of the COX-1-selective inhibitor SC-560 to induce a partial reduction in ET-1-induced PGE2 and VEGF production suggested that, in HEY cells, the ET-1-induced COX-1 up-regulation represents an active and concomitant mechanism that can, in part, increase PGE2. These data open the intriguing possibility that ET-1 may use both COX enzymes and their products toward the development of ovarian carcinoma.
Interestingly, we found that the addition of a specific ETAR antagonist, BQ 123, blocked ET-1-induced COX-1 and COX-2 expression and COX-2 promoter activity. Moreover, BQ 123 was able to block PGE2 secretion to the same extent as the COX nonselective inhibitor indomethacin, demonstrating the key role of the ETAR-mediated pathway in the induction of both COX-1- and COX-2-derived PGE2 production. We have demonstrated previously that in vivo blockade of the ETAR autocrine pathway using the highly specific antagonist ABT-627 is associated with a significant reduction of microvessel density, VEGF expression, and matrix metalloproteinase 2 and connexin 43 phosphorylation and with increased tumor apoptosis (18 , 19) . In this study, we found that ABT-627 treatment decreases COX-2 expression concomitantly with growth of tumor xenografts, indicating that the antitumoral activity of this small molecule may also be due to the inhibition of COX-2 activity.
In conclusion, the present results implicate ETAR in the induction of COX-1 and COX-2 expression and their downstream effect in ovarian cancer. Given that the two enzymes are overexpressed in the majority of these tumors (26, 27, 28, 29, 30, 31, 32, 33) , the possibility of blocking their activity may have relevant implication in the prevention and treatment of this malignancy. Targeting COX and related signaling cascade via ETAR blockade may be therapeutically advantageous in the treatment of ovarian carcinoma. COX-2 expression has been shown to be a factor indicating poor prognosis (30 , 31 , 44 , 45) and a predictor of chemoresistance in ovarian cancer patients treated with standard chemotherapy (32) . Consistent with these data, a functional link between increased levels of COX-2 and cellular resistance against apoptosis in response to a variety of extracellular and intracellular stimuli has recently been established (46 , 47) . In this regard, combination treatment with COX-2 inhibitors and ETAR antagonist may be warranted to design newer therapeutic approaches to ovarian cancer, which could result in increased susceptibility of the cells to apoptosis.
We gratefully acknowledge Dr. Perry Nisen (Global Oncology Development, Abbott Laboratories) for kindly providing ABT-627, Dr. Raymond DuBois for kindly providing plasmid constructs, Giacomo Elia for excellent technical assistance, and Maria Vincenza Sarcone for secretarial assistance.
Grant support: Grants from the Associazione Italiana Ricerca sul Cancro, Ministero della Salute, and MIUR-Consiglio Nazionale delle Ricerche. F. Spinella is a recipient of a fellowship from Fondazione Italiana Ricerca sul Cancro.
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 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Anna Bagnato, Laboratory of Molecular Pathology and Ultrastructure, Regina Elena Cancer Institute, Via delle Messi d’Oro 156, 00158 Rome, Italy. Phone: 39-06-52662565; Fax: 39-06-52662505; E-mail:
- Received February 18, 2004.
- Revision received April 6, 2004.
- Accepted April 16, 2004.