Cancer Cell Biology
PGE2-mediated upregulation of iNOS in murine breast cancer cells through the activation of EP4 receptors
Article first published online: 28 OCT 2003
Copyright © 2003 Wiley-Liss, Inc.
International Journal of Cancer
Volume 108, Issue 3, pages 384–389, 20 January 2004
How to Cite
Timoshenko, A. V., Lala, P. K. and Chakraborty, C. (2004), PGE2-mediated upregulation of iNOS in murine breast cancer cells through the activation of EP4 receptors. Int. J. Cancer, 108: 384–389. doi: 10.1002/ijc.11575
- Issue published online: 21 NOV 2003
- Article first published online: 28 OCT 2003
- Manuscript Accepted: 26 AUG 2003
- Manuscript Revised: 28 JUL 2003
- Manuscript Received: 26 MAY 2003
- Canadian Breast Cancer Research Alliance. Grant Number: 012312
- Breast Cancer Society of Canada
- EP receptors;
- nitric oxide;
- nitric oxide synthase;
- prostaglandin E2
We report here that endogenous prostaglandin E2 (PGE2) resulting from cyclooxygenase (COX)-2 expression in a highly metastatic murine breast cancer cell line C3L5 upregulates IFN-γ + LPS-induced nitric oxide (NO) synthase (iNOS) expression and NO production. This action of PGE2 is mediated through the EP4 receptor in a cAMP-dependent manner. Both nonselective and selective COX-2 inhibitors suppressed IFN-γ + LPS-induced NO production, which was largely restored by exogenous PGE2 or EP4 receptor agonist PGE1 alcohol. EP4 antagonist AH-23848B inhibited NO production with a concomitant downregulation of iNOS mRNA in IFN-γ + LPS-stimulated cells. cAMP dependence of NO production by cells under inducible conditions was demonstrated by the use of known modulators of intracellular cAMP. Since both COX-2 and iNOS are implicated in breast cancer progression, our findings of EP4 receptor-mediated upregulation of iNOS in COX-2-expressing breast cancer cells suggest that blocking COX-2 and/or EP4 may provide a simple therapeutic modality in this tumor model. © 2003 Wiley-Liss, Inc.
A molecular cross-talk between cell-derived prostaglandins (PGs) and nitric oxide (NO) is a well-known phenomenon. Depending on the cell type, NO can stimulate or inhibit the activity of cyclooxygenase (COX) enzymes, and conversely, PGs can modulate the activity of NO synthase (NOS) enzymes.1, 2 The inducible forms of COX (COX-2) and NOS (iNOS) can be activated by inflammation-associated cytokines and bacterial products, e.g., IFN-γ and LPS.3 An aberrant upregulation of one or both enzymes in certain cancers has been implicated in tumor progression and metastasis. For example, coexpression of COX-2 and iNOS enzymes was demonstrated in hepatocellular carcinomas,4 ovarian tumors5 and prostate cancer.6 Constitutive overexpression of COX-2 is a salient feature shared by a large variety of invasive human tumors, including breast carcinomas, in which case the degree of expression correlates with poor prognosis.7 Selective inhibition of COX-2 is now in use for chemoprevention and chemointervention of colonic tumors.8 Tumor-derived PGs have been shown to promote tumor progression and metastasis by multiple mechanisms: a stimulation of tumor cell migration, invasiveness and tumor-associated angiogenesis,9 inhibition of tumor cell apoptosis10 and inactivation of host antitumor immune cells.11 The nature of PG-mediated effects on tumor cells or host cells would depend on the expression of the type of PG-binding cell membrane receptors, which are coupled with different G-proteins. For example, a major tumor-derived prostanoid PGE2 acts through 4 different receptors (EP1–EP4 receptors) providing intracellular signaling by different mechanisms, e.g., increase in intracellular level of Ca2+ (EP1), elevation of intracellular cAMP (EP2/EP4) or reduction of cAMP synthesis (EP3).12
Similar to the PGs, tumor-derived NO has also been implicated in the progression of many human and experimental tumors. Aberrant expression of endothelial-type (e) NOS as well as iNOS in tumor cells and/or tumor-associated host cells has been shown to serve as the source of NO overproduction in these tumors.13 While the constitutive expression of iNOS in cancer cells may not be uniform, its induction in cancer cells can be triggered, e.g., by inflammatory cytokines released from macrophages within the tumor tissue.14 In the case of human breast cancer, either tumor cells or tumor-associated cells or both express iNOS,15, 16 and the level of NOS activity was shown to have a strong correlation with tumor grade.16, 17 Studies from our laboratory have documented that tumor-derived NO promoted tumor growth and metastasis in an eNOS-expressing murine breast cancer model by stimulating tumor cell migration, invasiveness and angiogenesis.18, 19, 20, 21, 22 The migration-promoting effects of NO were shown to be due to sequential activation of NOS, guanylate cyclase and mitogen-activated protein kinase.23 Induction of iNOS in these eNOS-expressing tumor cells further accentuated their invasiveness by an upregulation of matrix metalloproteinase MMP-2.19 Because of permanently elevated levels of PGE2 in the microenvironment of many tumors with high metastatic potential, PGE2-dependent molecular mechanisms associated with different EP receptors may be involved in the regulation of iNOS activity, which may play an additional role in tumor progression. The present study was therefore designed to test whether PGE2 can modulate IFN-γ + LPS-induced iNOS activity in a constitutively COX-2-expressing, iNOS-negative, highly metastatic murine mammary adenocarcinoma cell line C3L5 and, if so, to understand what EP receptor class and postreceptor signaling molecules are instrumental in the PGE2-mediated effects.
MATERIAL AND METHODS
Tumor cell line
C3L5 is a highly metastatic murine breast cancer cell line clonally derived in this laboratory from a spontaneous mammary adenocarcinoma developing in a C3H/HeJ retired breeder female mouse.24 This cell line constitutively expresses eNOS (but not iNOS)19 and COX-2 enzymes9 and express all the EP receptors except EP2, as detectable by RT-PCR.25 Cells were grown in DMEM media (Invitrogen, Burlington, ON) supplemented with 10% fetal bovine serum, 100 U/ml penicillin G, 100 μg/ml streptomycin and 25 mM HEPES in a humidified incubator with 5% CO2 at 37°C.
PGE2, PGE1 alcohol, 17-phenyl trinor PGE2, butaprost, sulprostone, carboprostacyclin and SC-19220 were purchased from Cayman Chemical (Ann Arbor, MI). AH-6809 was purchased from Biomol (Plymouth Meeting, PA). AH-23848B, a selective EP4 receptor antagonist,26 was a gift from Glaxo/Wellcome (Stevenage, U.K.). Calcium ionophore A23187, concanavalin A (Con A), indomethacin, ionomycin, Nω-nitro-L-arginine methyl ester, hydrochloride (L-NAME), Nω-nitro-D-arginine methyl ester, hydrochloride (D-NAME), forskolin and lipopolysaccharide (LPS) from E. coli 026:B6 were from Sigma (Oakville, ON, Canada). 8-bromoadenosine-3′,5′-cyclic monophosphorothionate, Rp isomer (Rp-8-Br-cAMPS), was from Biolog (Bremen, Germany). Recombinant mouse γ-interferon (IFN-γ) was from Calbiochem (San Diego, CA).
NOx measurements in cell culture media
Measurements of total amount of nitrate + nitrite (NOx) in cell culture media were performed using Nitrate/Nitrite Colorimetric Assay Kit from Cayman Chemical designed to run with 96-well plate.
Enzyme immunoassay of PGE2
The levels of PGE2 in cell culture media were measured using the Prostaglandin E2 EIA Kit-Monoclonal from Cayman Chemical.
Total RNA was extracted from cell monolayers using TRIzol reagent and reverse-transcribed utilizing the Superscript II RNase H− Reverse Transcriptase (Invitrogen) following the protocols of the manufacturer. Primers to amplify transcripts of iNOS (TTTGACCAGAGGACCCAGAG, sense; AAGACCAGAGGCAGCACATC, antisense) and GAPDH (AATGCATCCTGCACCACCAA, sense; GTAGCCGTATTCATTGTCATA, antisense) genes were synthesized locally at the UWO Oligo Factory (University of Western Ontario, London, Canada). PCR reaction was run in a final volume of 20 μl containing 0.2 mM dNTP, 10 pmol of each sense and antisense primers, 5% DMSO, 1.5 mM MgCl2 and 0.8 units of Taq DNA Polymerase (Promega, Madison, WI) in the Eppendorf Mastercycler Gradient (Eppendorf Scientific, Westbury, NY). PCR products were separated on 1% agarose gel containing 0.25 μg/ml ethidium bromide, visualized under UV light and recorded by a digital camera; the density of bands was quantified using Kodak Digital Science 1D software.
Data were analyzed by Student's t-test considering p < 0.05 as an indicator of significant difference between means. Each experiment was performed at least 3 times to confirm reproducibility.
NOS inhibition does not influence PGE2 synthesis by murine C3L5 breast cancer cells
Treatments with L-NAME, a nonselective inhibitor of NOS enzymes, and D-NAME, an inactive enantiomer of L-NAME, were used to evaluate the role of endogenous NO in any cellular function. L-NAME (but not D-NAME) inhibited, in a dose-dependent manner, the accumulation of NOx in culture media of C3L5 cells stimulated with 200 U/ml IFN-γ and 10 μg/ml LPS for 48 hr (Fig. 1a). It should be noted that treatment of C3L5 cells with a single agent, either IFN-γ or LPS, did not induce accumulation of NOx in cell culture media (data not shown). L-NAME at the concentration of 0.5 mM was found to be sufficient to block NO production resulting from iNOS activity (Fig. 1a).
C3L5 cells released and accumulated relatively high basal levels of PGE2 in cell culture media (2–3 nM) and this level was elevated further after stimulation with phospholipase A2-inducing agents Con A, A23187, ionomycin, or iNOS/COX-2 inducing agents IFN-γ + LPS (Fig. 1b and c). The level of PGE2 biosynthesis depends on the availability of arachidonic acid (liberated from cell membrane phospholipids by the activity of multiple phospholipase A2 enzymes) and the activity of COX enzymes. The observed differences in effects of Ca2+ ionophores (A23187 and ionomycin) and Con A on the levels of PGE2 release from cancer cells can be due to stimulus-dependent activation of different phospholipase A2 isoforms,27 which may have differential effects on different phospholipid pools.28 To test whether the basal or induced levels of PGE2 could be affected by endogenously produced NO, we examined the effects of L-NAME or D-NAME on PGE2 accumulation. At a concentration of 0.5 mM, both L-NAME and D-NAME resulted in a small decrease of extracellular PGE2, especially under the treatment with Con A, A23187 and ionomycin (Fig. 1b). However, there were no significant differences between L-NAME and D-NAME effects in each case, indicating that the minor effect of L-NAME on PGE2 production was not a consequence of NOS inhibition. Thus, inhibition of the eNOS enzyme by L-NAME failed to affect the activity of COX enzymes in C3L5 cells. Furthermore, neither Con A nor A23187 nor ionomycin was found to have any significant effect on NOx production (results not presented).
While constitutive eNOS expression in C3L5 cells resulted in relatively low levels of NOx, treatment of cells with IFN-γ + LPS leads to both upregulation of iNOS and COX-2, resulting in the release of high levels of NO and PGE2, respectively.9, 19, 25 In spite of the relatively high level of NOx in the culture medium, these cells resisted apoptosis most likely due to high COX-2 expression.13 Although L-NAME inhibited IFN-γ + LPS-induced synthesis of NO by C3L5 cells (Fig. 1a), it failed to affect PGE2 synthesis (Fig. 1c). Nonselective COX inhibitor indomethacin as well as selective COX-2 inhibitor NS-389, however, readily inhibited the basal as well as the IFN-γ/LPS-induced production of PGE2 (Fig. 1c). These results indicate that NO production by murine C3L5 breast cancer cells under both basal and inducible conditions do not influence PGE2 synthesis.
COX inhibitors suppress IFN-γ + LPS-induced NO production and iNOS expression in C3L5 cells
Next, we examined whether NO production by C3L5 cells was affected by cell-derived PGE2 under basal or inducible conditions. NOx concentration in the medium was measured in the presence of nonselective COX inhibitor indomethacin and selective COX-2 inhibitor NS-398. While the basal production of NOx by the C3L5 cells incubated for 48 hr in serum-free DMEM was less than 1 μM (Fig. 2a), treatment of these cells with 200 U/ml IFN-γ and 2–10 μg/ml LPS led to a large increase of NOx, reaching a level of ∼ 100 μM during 48 hr (Fig. 2a) associated with an upregulation of steady-state iNOS mRNA expression as detected by RT-PCR (Fig. 2b and c). Nontoxic doses of COX inhibitors, 20 μM indomethacin or 50 μM NS-398, significantly (p < 0.01) reduced IFN-γ + LPS-induced accumulation of NOx in cell culture media (Fig. 2a). The reduced levels of NOx in the presence of COX inhibitors were correlated with decreased steady-state levels of iNOS mRNA expression in these cells. Since PGE2 is a major prostanoid released from C3L5 cells,25 these findings suggest that NO production by C3L5 cells under inducible conditions is mediated at least in part through PGE2-dependent molecular mechanisms. This possibility was then tested by examining the effects of PGE2 and some selective EP receptor agonists on NO production by C3L5 cells in the presence of COX-2 inhibitor.
Partial restoration of iNOS-associated NO production by PGE2 and an EP4 agonist in NS-398 treated cells
We have earlier shown that treatment of C3L5 cells with the selective COX-2 inhibitor NS-398 leads to a drastic reduction in PGE2 release from these cells (Fig. 1c). Since PGE2 acts through 4 different types of G-protein-coupled receptors (EP receptors), we examined whether PGE2 and selective EP receptor agonists12 (17-phenyl trinor PGE2, butaprost, sulprostone and PGE1 alcohol) affect iNOS activity in C3L5 cells. Effect of the IP receptor agonist carbaprostacyclin was also studied in this experiment. We observed that only PGE1 alcohol (mouse EP4/EP3 receptor ligand)12 and PGE2 at a concentration of 100 nM caused a significant increase in IFN-γ + LPS-induced NOx accumulation in culture media of C3L5 cells, which had also been treated with 50 μM NS-398; on the other hand, other agonists, e.g., for EP1, EP2, EP3 and IP receptors, failed to affect extracellular level of NOx significantly (Fig. 3). These results suggest that EP4 receptors are primarily responsible for mediating PGE2 action in regulating iNOS activity.
EP4 receptor antagonist downregulates iNOS induction in C3L5 cells
To verify the role(s) of different EP receptors in endogenous PGE2-mediated regulation of iNOS activity, we measured the accumulation of NOx by C3L5 cells under inducible conditions in the presence of selective antagonists of EP1 receptors (SC-19220), EP1/EP2/DP receptors (AH-6809) and EP4 receptors (AH-23848B). The abilities of these compounds at a concentration range of 0.1–10 μM to block the actions of elevated levels of PGE2 (∼ 12.5 nM) resulting from treatment with IFN-γ + LPS were examined. EP4 antagonist AH-23848B exhibited a strong inhibition of NOx production, EP1/EP2/DP antagonist AH-6809 exhibited modest effect, but EP1 antagonist SC-19220 was without any effect (Fig. 4a). To determine whether EP4 antagonist affected iNOS mRNA expression in C3L5 cells, we further performed RT-PCR analysis using the RNA isolated from the cells treated with AH-23848B. A concentration-dependent decrease of iNOS mRNA expression was observed (Fig. 4b and c), and this decrease correlated strongly with the decreased levels of NOx in cell culture media. This strong correlation of the 2 parameters (mRNA and NOx) indicates clearly that an EP4 receptor-mediated signaling pathway is crucial for iNOS upregulation in these cells.
Upregulation of iNOS in C3L5 cells is cAMP-dependent
Since EP4 receptor-mediated signaling utilizes cAMP as a second messenger and C3L5 cells response to exogenous PGE2 by elevation of intracellular cAMP,25 we examined whether IFN-γ + LPS-induced NO production in C3L5 cells is affected by intracellular cAMP. These experiments were performed in cells treated with NS-398 in order to minimize interference of cAMP resulting from endogenous PGE2 activity. We found that a potent activator of adenylate cyclase forskolin stimulated NOx production by IFN-γ + LPS-treated cells in a dose-dependent manner (Fig. 5a), whereas cAMP antagonist and metabolically stable inhibitor of PKA Rp-8Br-cAMPS exhibited dose-dependent inhibition of NOx accumulation in cell culture media (Fig. 5b). Similar effects were also produced by a cell-permeable analogue of cAMP dibutyryl-cAMP and an adenylate cyclase inhibitor SQ 22536 (data not presented). These observations demonstrate that cAMP-mediated signaling cascades may be responsible for the regulation of iNOS activity in C3L5 breast cancer cells.
In this article, we describe for the first time the functional role of EP receptors in PGE2-mediated regulation of iNOS activity in breast cancer cells. Tumor-derived NO has been shown to promote the progression of both human15, 16, 17 and murine18, 19, 20, 21, 22, 23 breast cancer. In the former case, the source of NO has been shown to be iNOS-expressing tumor cells16 and/or tumor-associated macrophages.15 In the latter case, represented by a C3H/HeJ murine breast cancer model inclusive of spontaneous mammary tumors and their clonal derivatives, the primary source of NO was eNOS-expressing tumor cells.18, 19, 20, 21, 22, 23 Using a highly metastatic and high eNOS-expressing clonal derivative C3L5, it was shown that eNOS-derived NO promoted tumor cell migration,21 invasiveness19 and angiogenesis.21, 22 However, additional induction of iNOS in C3L5 cells in the presence of IFN-γ + LPS was shown to increase further its invasive capacity by upregulation of MMP-2.19 Thus, constitutive and induced iNOS expression appears to play an important role in breast cancer progression.
Tumor-derived PGE2 has also been shown to promote tumor progression in this murine breast cancer model by stimulation of tumor cell migration, invasiveness and angiogenesis.9 C3L5 cells utilized in the above and the present study were shown to express COX-29 and the prostaglandin receptors EP1, EP3 and EP4 but not EP2.25 Whether modulation of the prostaglandin (PG) system could influence the NO system, or modulation of NO system could influence the PG system in breast cancer cells was never investigated before. Using C3L5 cell line, the present study has revealed that endogenous PGE2 can upregulate iNOS expression and NO production in IFN-γ + LPS-treated cells through EP4 receptor-mediated signaling, which utilizes cAMP. This conclusion is based on the following observations. First, IFN-γ + LPS-induced accumulation of NOx in cell culture media and iNOS mRNA expression in cells were inhibited by selective COX-2 and nonselective COX inhibitors as well as EP4 antagonist. Second, this inhibition was partially abrogated with exogenous PGE2 and an EP4 agonist PGE1 alcohol. Third, modulation of intracellular cAMP with forskolin, dibutyryl-cAMP, SQ 22536 and Rp-8Br-cAMPS resulted in a parallel modulation of NOx production by these cells under inducible conditions. To the best of our knowledge, this is the first evidence for EP4 receptor-mediated upregulation of iNOS gene expression and NO production by endogenous PGE2 in breast cancer cells.
Unstimulated C3L5 cells do not express any detectable level of iNOS mRNA as measured by RT-PCR in the present study or as had earlier been shown by Northern blot analysis.19 Our data do not answer the question why constitutive expression of COX-2 in C3L5 cells associated with elevated endogenous PGE2 production fails to upregulate iNOS gene expression in the absence of iNOS-inducing agents. COX inhibitors as well as EP4 receptor agonists were found to affect iNOS activity only after the enzyme had been induced. Thus, it is likely that transcription factors induced by IFN-γ + LPS may act synergistically with those induced by PGs for maximal expression of iNOS. For example, heat shock protein 60 (HSP60) can be induced by IFN-γ.29 HSP60 then synergizes with IFN-γ to induce iNOS but not COX-2. Furthermore, it has been demonstrated that p38 MAP kinase differentially regulates COX-2 and iNOS expression.30
Data reported in the literature regarding the effects of PGE2 on iNOS expression are conflicting. Depending on the cell type, PGE2 has been shown to potentiate as well as inhibit iNOS expression.1, 2, 3 In the case of stimulation, similar to the present results, the effects were believed to be associated with an increase in intracellular cAMP.31 While the promoter region of iNOS gene contains binding sites for numerous transcription factors, some functional relevance to cAMP actions has been examined for only 3 factors, namely, NF-κB, CREB and C/EBP.31 Whether activation of any of these transcription factors is important in stimulating cAMP-mediated iNOS transcription in breast cancer cells remains to be tested.
Although C3L5 cells express EP1 and EP3 receptor mRNAs, EP1/EP3 agonists 17-phenyl trinor PGE2 and sulprostone in NS-398-treated cells did not increase NO production, but PGE1 alcohol, which was shown to have better selectivity for mouse EP4 than EP3 receptors,12 was as effective as PGE2 in partial restoration of NO production in these cells. Furthermore, while EP4 antagonist AH-23848B showed strong dose-dependent inhibition, EP1/EP2/DP antagonist AH-6809 showed modest inhibition and EP1 receptor antagonist SC-19220 showed no inhibition of NO levels. These results strongly indicate the predominant participation of EP4 receptors in mediating the endogenous PGE2-mediated upregulation of iNOS expression in C3L5 cells. Modest effect of AH-6809 could be due to possible inhibition of DP receptor-mediated action, because EP1 receptor antagonist was without effect, and these cells had no detectable EP2.25
Inhibition of NO production by COX inhibitors indomethacin and NS-398 in IFN-γ + LPS-treated C3L5 cells might be not only due to inhibition of cAMP production resulting from the inhibition of PGE2 production, but also by direct transcriptional inactivation resulting from their ability to inhibit binding some transcription factor(s) with specific nucleotide sequence in iNOS promoter region. Our results of incomplete abrogation of NS-398-induced inhibition of NO production by treatment with PGE2 or PGE1 alcohol may indicate the existence of both the mechanisms. In fact, another COX inhibitor sodium salicylate has been shown to inhibit iNOS transcription directly by binding with C/EBPβ in a mouse macrophage cell line32 and in quiescent human fibroblasts.33 Like iNOS, COX-2 gene in its promoter region also has C/EBP binding site and binding of this element with C/EBPβ has been shown to be blocked by COX inhibitors aspirin and salicylate.32, 33, 34 Whether indomethacin and NS-398 can cause transcriptional inactivation of iNOS and COX-2 genes in a similar fashion needs further investigation.
Administration of COX-2 inhibitors for chemointervention in colon cancer is a significant achievement during the recent years.8 Combined inhibition of COX-2 and NOS has recently been shown to provide additional benefit against the development of aberrant crypt foci in azoxymethane-treated mice, an experimental model for human familial adenomatous polyposis.35 We had earlier shown that EP4 is the predominant receptor responsible for autocrine PGE2 stimulation of migration of metastatic human as well as murine breast cancer cells.25 It may be suggested that administration of EP4 antagonist instead of COX or NOS inhibitors may be more advantageous for chemoprevention in breast cancer because there might be compensatory effects of COX inhibitors on upregulation of EP4 receptors,36 and blocking EP4 receptors may serve dual purpose: blockade of PGE2 action and inhibition of iNOS expression.
Supported by grant 012312 from the Canadian Breast Cancer Research Alliance (to P.K.L.) and a fellowship of the Breast Cancer Society of Canada (to A.V.T.).
- 2Cyclooxygenase: an important transduction system for the multifaceted roles of nitric oxide. In: RubanyiGM, editor. Pathophysiology and clinical applications of nitric oxide. part A. Amsterdam: Harwood Academic, 1999. 155–70..
- 11Prostaglandins and the host immune system: application of prostaglandin inhibitors for cancer immunotherapy. In: HarrisJE, BraunDP, AndersonKM, editors. Prostaglandin inhibitors in tumor immunology and immunotherapy. Boca Raton: CRC Press, 1994. 187–227., .
- 23Nitric oxide-mediated promotion of mammary tumour cell migration requires sequential activation of nitric oxide synthase, guanylate cyclase and mitogen-activated protein kinase. Int J Cancer 2003; 106: 496–504., , .