Cyclooxygenase-2-Derived Prostaglandin E2 Protects Mouse Embryonic Stem Cells from Apoptosis

Authors

  • Jun-Yang Liou,

    1. Division of Hematology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    2. Vascular Biology Research Center, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    Search for more papers by this author
  • David P. Ellent,

    1. Division of Hematology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    2. Vascular Biology Research Center, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    Search for more papers by this author
  • Sang Lee,

    1. Division of Hematology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    2. Vascular Biology Research Center, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    Search for more papers by this author
  • Jennifer Goldsby,

    1. Division of Hematology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    2. Vascular Biology Research Center, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    3. Department of Obstetrics and Gynecology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    Search for more papers by this author
  • Bor-Sheng Ko,

    1. Division of Hematology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    2. Vascular Biology Research Center, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    Search for more papers by this author
  • Nena Matijevic,

    1. Division of Hematology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    2. Vascular Biology Research Center, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    Search for more papers by this author
  • Jaou-Chen Huang,

    1. Department of Obstetrics and Gynecology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    Search for more papers by this author
  • Kenneth K. Wu M.D., Ph.D.

    Corresponding author
    1. Division of Hematology, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    2. Vascular Biology Research Center, The University of Texas Health Science Center at Houston, Houston, Texas, USA
    3. National Health Research Institutes, Zhunan, Miaoli, Taiwan
    • Division of Hematology and Vascular Biology Research Center, The University of Texas Health Science Center at Houston, 6431 Fannin, MSB 5.016, Houston, Texas 77030, USA. Telephone: 713-500-6801; Fax: 713-500-6812
    Search for more papers by this author

Abstract

Little is known about prostaglandin synthesis and function in embryonic stem cells. We postulated that mouse embryonic stem (mES) cells possess enzymes to synthesize protective prostaglandins. Compared with differentiated adult cells, mES cells were less susceptible to H2O2-induced apoptosis. However, their apoptosis was enhanced by indomethacin or SC-236, a selective inhibitor of cyclooxygenase (COX)-2. Analysis of COX pathway enzymes by Western blotting revealed expression of COX-2 and cytosolic and microsomal prostaglandin E2 (PGE2) synthases. COX-1 and prostacyclin (PGI2) synthases were undetectable. mES cells produced PGE2 but not PGI2. Importantly, PGE2 rescued mES cells from apoptosis. To elucidate the signaling mechanism by which PGE2 inhibits apoptosis, we analyzed E-type prostaglandin (EP) receptors by Western blots. All EP isoforms were detected except EP4. Butaprost, a specific EP2 agonist, rescued mES cells from apoptosis, whereas sulprostone, an EP1/EP3 agonist, had no effect, suggesting selective interaction of PGE2 with EP2. The antiapoptotic effect of PGE2 was abrogated by Ly-294002 or wortmannin but not H-89 or a specific inhibitor of protein kinase A, suggesting signaling via phosphatidylinositol-3 kinase (PI-3K). Akt was constitutively active in mES cells, which were inhibited by indomethacin and rescued by PGE2. The rescuing effect of PGE2 was abrogated by Ly-294002. These results indicate that mES cells constitutively express COX-2 and PGE synthases and produce PGE2, which confers resistance to apoptosis via EP2-mediated activation of PI-3K to the Akt pathway.

Disclosure of potential conflicts of interest is found at the end of this article.

Introduction

The inner cell mass (ICM) of blastocysts possesses apoptotic machinery and an antiapoptotic defense program [1, 2]. It has been reported that transforming growth factor α protects ICM from apoptosis [3, 4]. Cultured mouse embryonic stem (mES) cells derived from ICM undergo apoptosis when subjected to prolonged hypoxia or methylglyoxal-induced oxidative stress [5, 6]. However, little is known about the endogenous protective mechanism. The role of prostaglandins in protecting mES cells from apoptosis has not been reported. Prostacyclin (PGI2) and prostaglandin E2 (PGE2) have emerged as important endogenous molecules for protection of somatic cell survival. Endogenously produced PGI2 has been shown to protect endothelial cells from H2O2-induced apoptosis and renal interstitial cells from hypertonicity induced apoptosis via the peroxisome proliferators-activated receptors δ pathway [7, [8]–9]. PGE2 has been reported to protect dendritic and neuronal cells from apoptosis via E-type prostaglandin (EP) receptors EP2 and/or EP4 [10, [11]–12]. PGI2 and PGE2 are synthesized in diverse somatic cells via the cyclooxygenase (COX) pathway. Prostaglandin H2 (PGH2) synthases (also known as COX) are bifunctional enzymes possessing cyclooxygenase activity, which converts arachidonic acid (AA) to prostaglandin G2 (PGG2), and peroxidase activity, which reduces PGG2 to PGH2 [13]. PGH2 is a common precursor for the production of all prostanoids, including PGI2 and PGE2 [14]. Thus, COX enzymes occupy a central position in prostaglandin production. There are two COX isoforms. COX-1 is constitutively expressed in most mammalian somatic cells, whereas COX-2 is inducible by cytokines, growth factors, endotoxins, and hormones [15]. PGH2 produced by COX-1 or COX-2 is converted to PGI2 by a single enzyme, PGI synthase (PGIS) [16]. PGIS is constitutively expressed in several types of somatic cells, such as endothelial cells [17]. Recently, studies have shown that PGIS is expressed in uterus, oviduct, and preimplantation embryo and plays an important role in embryo implantation [18, [19]–20]. PGE2 is generated from PGH2 by PGE synthase (PGES). Three PGES isoforms have been identified. The cytosolic PGES (cPGES) is constitutively expressed in many somatic cells, whereas the microsomal PGES (mPGES)-1 is inducible and expressed in inflammatory cells [21, [22]–23]. mPGES-2 is also inducible [24], but its biology is still being investigated. It is unknown whether embryonic stem cells express PGI2 or PGE2 synthetic enzymes. Nor is it known whether prostaglandins play a role in protecting mES cell survival. Results from this study reveal constitutive expression of COX-2 and PGES, notably cPGES and mPGES-2 in mES cells. COX-2-derived PGE2 protects mES cells form H2O2-induced apoptosis via EP2-mediated activation of phosphatidylinositide-3 kinase (PI-3K) → Akt pathway.

Materials and Methods

Cell Culture and Treatment

CCE mouse embryonic stem cells were obtained from Stem Cell Technologies (Vancouver, BC, Canada, http://www.stemcell.com) [25, 26] with permission from Dr. Robertson and Dr. Keller. Plastic tissue culture dishes were pretreated with 0.2% gelatin. Undifferentiated mES cells were grown at 5% CO2 in Dulbecco's modified Eagle's medium (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) at 37°C containing glutamine (Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com), sodium pyruvate (Sigma-Aldrich), nonessential amino acids (Sigma-Aldrich), β-mercaptoethanol (Sigma-Aldrich), leukemia inhibitory factor (1,000 U/ml; Chemicon, Temecula, CA, http://www.chemicon.com), 15% fetal bovine serum (FBS; Sigma-Aldrich), and penicillin/streptomycin (Gibco-BRL). Cells were passed every 2 days and maintained at a low cell density. For apoptosis analysis, mES cells were treated with H2O2 (2 mM) for 24 hours. For COX inhibitor experiments, mES cells were pretreated with indomethacin (20 μM), SC-560 (10 μM), and SC-236 (5 μM) for 30 minutes followed by H2O2 for 24 hours. For rescuing experiments, cells were pretreated with carbaprostacyclin (cPGI2) (10 μM), PGE2 (10 μM), butaprost (10 μM), or sulprostone (1 μM) for 30 minutes followed by H2O2 for 24 hours. For inhibitor experiments, cells were treated with H89 (10 μM), protein kinase A inhibitor (PKI) 14–22 amide (5 μM), wortmannin (10 μM), or Ly-294002 (20 μM) for 30 minutes followed by indicated treatments.

Western Blot Analysis

Thirty μg of cell lysate proteins were applied to each lane and analyzed by Western blotting as described previously [8]. Rabbit polyclonal antibodies against mouse cleaved poly(ADP-ribose) polymerase (PARP) were purchased from Cell Signaling Technology (Beverly, MA, http://www.cellsignal.com). Rabbit polyclonal antibodies against mouse mPGES-1, mPGES-2, cPGES, EP1, EP2, EP3, and EP4 and monoclonal antibodies against mouse COX-1, COX-2, and PGIS were obtained from Cayman Chemical Company (Ann Arbor, Michigan, http://www.caymanchem.com). Monoclonal antibody against mouse Oct-3/4 was from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, http://www.scbt.com). Donkey anti-rabbit or donkey anti-mouse IgG conjugated with horseradish peroxidase were purchased from Santa Cruz Biotechnology Inc. Rabbit polyclonal antibodies against Akt and phosphorylated Akt (p-Akt) at Ser 473 were obtained from BD Pharmingen (San Diego, http://www.bdbiosciences.com/index_us.shtml). Protein bands were visualized by an enhanced chemiluminescence system (Pierce Biotechnology Inc., Rockford, IL, http://www.piercenet.com).

COX-2 Small Interfering RNA

COX-2 expression was knocked down by transfection of SureSilencing small hairpin RNA plasmids (SuperArray Bioscience Corporation, Frederick, MD, http://www.superarray.com). A plasmid containing control or a COX-2 small interfering (si)RNA sequence (5′ AACCTCGTCCAGATGCTATCT 3′) was transfected to CCE cells by an Effectene Transfection Kit (Qiagen, Hilden, Germany, http://www1.qiagen.com). After 72 hours, COX-2 expression was determined by Western blotting. For cytotoxicity experiments, CCE cells were transfected with COX-2 siRNA plasmids for 48 hours, and the transfected cells were treated with PGE2 and/or H2O2 for an additional 24 hours.

Analysis of Apoptosis by Flow Cytometry

mES cells incubated with indicated reagents were harvested by trypsin, centrifuged at 500 g for 10 minutes, washed with phosphate-buffered saline (PBS), and incubated with fluorescein isothiocyanate-labeled annexin V antibody and propidium iodide (PI) (Beckman Coulter, Fullerton, CA, http://www.beckmancoulter.com) in the dark at room temperature for 30 minutes. The labeled cells were analyzed by flow cytometry. Percentages of cells with positive stain for annexin V and PI were calculated.

Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labeling Assay

mES cells were grown on gelatinized coverslips and treated with indicated reagents for 24 hours. Cells were washed with PBS and fixed by 4% paraformaldehyde at 4°C for 40 minutes followed by 100% methanol for 10 minutes at room temperature. Fixed cells were washed three times with PBS. The terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining was performed by using an assay kit (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com). For cell counting, samples were incubated in 1 μg/ml Hoechst 33258 at room temperature in the dark for 30 minutes. The number of TUNEL positive cells stained by fluorescein isothiocyanate and Hoechst positive cells was determined by immunofluorescent microscopy.

Alkaline Phosphatase

mES cells were evaluated for alkaline phosphatase (AP) activity using the AP substrate kit (Sigma-Aldrich) in accordance with the manufacturer's instruction. Cells were fixed with 4% paraformaldehyde for 40 minutes at 4°C and washed with Tris-PBS. Cells were incubated in freshly prepared nitroblue tetrazolium/5-bromo-4-chloro-3-indoyl phosphate solution in the dark for 15 minutes. The reaction was terminated by adding a stop buffer (20 mM Tris-HCl, pH 8.0, and 5 mM EDTA).

Measurement of PGE2 and 6-Keto-PGF by Enzyme Immunoassay

Medium of mES cells treated with indicated reagents was harvested and stored at −80°C. The enzyme immunoassay (EIA) kits for PGE2 and 6-keto-PGF (a stable metabolite of PGI2) were purchased from R&D Systems Inc. (Minneapolis, http://www.rndsystems.com).

Statistical Analysis

Analysis of variance was used to determine statistical differences of apoptosis between groups. A p < .05 was considered to be statistically significant.

Results

Augmentation of H2O2-Induced Apoptosis by COX-2 Inhibitors

To evaluate the effect of H2O2 on mES cell apoptosis, mES cells were treated with increasing concentrations of H2O2 for 24 hours, and apoptosis was analyzed by flow cytometry for annexin V positive cells and by Western blotting for PARP cleavage. mES cells were resistant to oxidant-induced apoptosis, as H2O2 at 1 mM did not induce significant mES cell apoptosis (data not shown). At 2 mM, it increased annexin V positive cells by two- to threefold over the control and increased PARP cleavage by more than twofold (Fig. 1A, 1B). Indomethacin had no effect on annexin V positive cells or PARP cleavage but augmented H2O2-induced annexin V positive cells and PARP cleavage by more than twofold (Fig. 1A, 1B). These results suggest the involvement of COX in controlling apoptosis. To identify the responsible COX isoform(s), we treated mES cells with SC-560, a selective COX-1 inhibitor, or SC-236, a selective COX-2 inhibitor, and measured H2O2-induced annexin V positive cells. SC-236 but not SC-560 increased annexin V positive cells induced by H2O2 (Fig. 1C). These results indicate that COX-2 is involved in protecting mES cells from oxidative apoptosis.

Figure Figure 1..

COX-2 inhibition augmented mouse ESC apoptosis. Cells were treated with H2O2 (2 mM) with or without indomethacin (20 μM) for 24 hours. (A): Percentage of annexin V positive cells was analyzed by flow cytometry. (B): Cleaved PARP was analyzed by Western blots. This figure is representative of three experiments. (C): Cells were incubated with SC-560 (10 μM) or SC-236 (5 μM) for 30 minutes followed by H2O2 for 24 hours. Each bar in (A) and (C) denotes mean ± SD of three independent experiments. Abbreviation: PARP, poly(ADP-ribose) polymerase.

mES Cells Expressed COX-2 and PGES Isoforms

To ascertain the involvement of COX-2 in mES cell survival, we analyzed COX-2 and COX-1 proteins by Western blotting. COX-2 protein was detected, whereas COX-1 was undetectable in resting mES cells (Fig. 2A). Resting mES cells expressed cytosolic PGES and low levels of mPGES-1 and mPGES-2, whereas PGIS was undetectable (Fig. 2B). Medium from 24-hour cultured mES cells was collected, and PGE2 and 6-keto-PGF were analyzed by EIA. PGE2 was detected in the medium, which was not altered by H2O2 but was completely inhibited by indomethacin or SC-236 (Fig. 2C). Six-keto-PGF was undetectable in the medium (Fig. 2C). These results are consistent with a basal production of PGE2 by resting mES cells via COX-2 and PGES.

Figure Figure 2..

Expression of COX-2 and PGE synthase in mouse ESC. (A, B): Analysis of COX and other synthetic enzymes by Western blotting. (C): Analysis of PGE2 and prostacyclin2 (6-keto-PGF) in cultured medium by enzyme immunoassay. Each bar represents mean ± SD of three independent experiments. Abbreviations: COX, cyclooxygenase; cPGES, cytosolic PGE synthase; mES, mouse embryonic stem; mPGES, microsomal PGE synthase; PGE2, prostaglandin E2; PGIS, prostacyclin synthase.

Rescue of mES Cells from Apoptosis by PGE2

To confirm that COX-2-derived PGE2 protects mES cells from apoptosis, we transfected mES cells with COX-2 siRNA plasmids, which reduced COX-2 protein levels by >80% (Fig. 3A), and evaluated the effects of H2O2 and PGE2 on the cytotoxicity of the transfected cells. Transfection of siRNA per se had no effect on cytotoxicity evaluated by trypan blue staining but enhanced susceptibility of cells to H2O2-induced cytotoxicity (Fig. 3B). Pretreatment of PGE2 rescued the cells from cytotoxicity caused by COX-2 knockdown plus H2O2 (Fig. 3B). Similar results were obtained with pharmacological inhibition of COX-2. SC-236 alone did not significantly increase annexin V positive cells but augmented apoptosis induced by H2O2 (Fig. 3C). Addition of PGE2 abrogated the augmenting effect of COX-2 inhibition (Fig. 3C). We further confirmed the results by performing TUNEL assay. H2O2-induced TUNEL positive cells were elevated by SC-236 by several fold, and PGE2 returned the TUNEL positive cells to the basal level induced by H2O2 (Fig. 4). Taken together, these results indicate that COX-2 plays a crucial role in protecting mES cell survival through its production of PGE2. In the absence of COX-2 inhibition, PGE2 but not cPGI2 inhibited H2O2-induced PARP cleavage in mES cells (Fig. 5A). Furthermore, addition of PGE2 to mES cells cultured in serum-free medium for 24 hours, 48 hours, and 72 hours prevented cell death caused by serum depletion at all three time points (Fig. 5B). Consistent with previous reports [27, 28], CCE mES cells expressed constitutively Oct-3/4 proteins (Fig. 5C) and stained positively for alkaline phosphatase (Fig. 5D). Neither PGE2 nor indomethacin had an apparent effect on the level of these stem cell markers (Fig. 5C, 5D).

Figure Figure 3..

PGE2 rescued H2O2-induced apoptosis due to COX-2 knockdown or pharmacological inhibition. (A): Mouse ESCs were transfected with plasmids containing a COX-2 siRNA sequence or control. COX-2 proteins in transfected cells were analyzed by Western blots. This figure is representative of two experiments with similar results. (B): siRNA expressed mES cells were treated with H2O2 in the presence or absence of PGE2. (C): Mouse ESCs were pretreated with PGE2 for 30 minutes followed by H2O2 and/or SC-236. Each bar in (B) and (C) represents mean ± SD of three independent experiments. Abbreviations: COX, cyclooxygenase; PGE2, prostaglandin E2; siRNA, small interfering RNA.

Figure Figure 4..

PGE2 suppressed TUNEL positive cells. Cells were pretreated with PGE2 and/or SC-236 for 30 minutes followed by H2O2 for 24 hours. (A): Photomicrographs show TUNEL and Hoechst staining. (B): Quantitative analysis of TUNEL positive cells. Each bar represents mean ± SD of three independent experiments. Abbreviations: PGE2, prostaglandin E2; TUNEL, terminal deoxynucleotidyl transferase dUTP nick-end labeling.

Figure Figure 5..

PGE2 inhibited apoptosis without altering stem cell markers. (A): mES (mouse embryonic stem) cells were pretreated with PGE2 (10 μM) or cPGI2 (50 μM) followed by H2O2 (2 mM). PARP cleavage was analyzed by Western blots. Only cleaved PARP and actin control are shown. (B): mES cells were cultured in serum-free medium for 24 hours, 48 hours, or 72 hours and cytotoxicity was determined at each time point. Each bar denotes mean ± SD of three independent experiments. (C): mES cells were treated with indomethacin with or without PGE2 and Oct-3/4 proteins were analyzed by Western blots. This blot is representative of three experiments. (D): mES cells treated with indomethacin or PGE2 were stained for alkaline phosphatase. The photomicrographs are quantified. Each bar shows mean ± SD of three independent experiments. Abbreviations: cPGI2, carbaprostacyclin; h, hours; PARP, poly(ADP-ribose) polymerase; PGE2, prostaglandin E2.

Involvement of EP2 Receptor in the PGE2 Protective Action

PGE2 exerts its effect through G-protein-coupled EP receptors. Four EP isoforms have been identified [29]. Expression of EP isoforms in mES cells had not been reported. Therefore, we analyzed EP isoforms in resting mES cells by Western blotting using isoform-specific antibodies. EP1, -2, and -3 were detected, and EP4 was undetected (Fig. 6A). Butaprost, an agonist of EP2, and sulprostone, an agonist of EP1 and EP3, were employed to determine the EP isoform responsible for the rescuing effect of PGE2. Butaprost significantly reduced annexin V positive cells induced by SC-236 plus H2O2 (Fig. 6B, lane 6 vs. lane 4), whereas sulprostone had no effect on the annexin V positive cells (Fig. 6B, lane 8 vs. lane 4). Thus, PGE2 exerts its antiapoptotic action via EP2.

Figure Figure 6..

Prostaglandin E2 protected apoptosis via EP2. (A): EP receptors in mouse ESCs were analyzed by Western blots. This figure is representative of three experiments. (B): Cells were pretreated with butaprost (10 μM), sulprostone (1 μM), and/or SC-236 (5 μM) for 30 minutes followed by H2O2 for 24 hours. Each bar represents mean ± SD of three independent experiments. Abbreviations: EP, E-type prostaglandin; mES, mouse embryonic stem.

Inhibition of the Antiapoptotic Action of PGE2 by PI-3K Inhibitors

EP2 receptors are coupled to adenylyl cyclase and signal via the cyclic AMP (cAMP)-dependent kinase (PKA) [29]. Recent studies suggest that EP2 may also signal via the PI-3K → Akt pathway [10, 12]. To gain insight into the signaling pathway of the antiapoptotic action of PGE2, we treated mES cells with PKA inhibitors H-89 (10 μM) and PKI (5 μM) or PI-3K inhibitors Ly-294002 (20 μM) and wortmannin (20 μM) and analyzed cytotoxicity and apoptosis. Cytotoxicity induced by H2O2 plus indomethacin was blocked by PGE2 (Fig. 7A). Neither H-89 nor PKI increased cytotoxicity or abrogated the protective effect of PGE2 (Fig. 7A). By contrast, wortmannin and Ly-294002 enhanced H2O2-induced cytotoxicity and abrogated the protective effect of PGE2 to a similar extent (Fig. 7B). Results from cleaved PARP analysis were comparable to those of cytotoxicity. H-89 had no effect on H2O2-induced PARP cleavage and did not influence the protective effect of PGE2, whereas Ly-294002 enhanced H2O2-induced PARP cleavage to an extent comparable to that of cytotoxicity enhancement and abrogated the protective action of PGE2 (Fig. 7C). These results suggest that endogenous PI-3K plays a crucial role in protecting mES cells from H2O2-induced apoptosis, and exogenous PGE2 lost its rescuing effect once PI-3K was blocked by inhibitors such as Ly-294002 and wortmannin. Since Akt is the key effector of PI-3K, we determined whether COX-2-derived PGE2 activates Akt by measuring p-Akt. Indomethacin suppressed p-Akt without an effect on total Akt (Fig. 7D). PGE2 restored p-Akt to the basal level. p-Akt was completely suppressed by Ly-294002, consistent with PI-3K-mediated Akt activation by COX-2-derived PGE2 (Fig. 7D).

Figure Figure 7..

Phosphatidylinositide-3 kinase inhibitors blocked PGE2 effect. (A, B): Cells were treated with H-89 (10 μM), PKI (5 μM), wortmannin (10 μM), or Ly-294002 (20 μM) ± PGE2 and indomethacin for 30 minutes followed by H2O2 for 24 hours. Cytotoxicity was measured by trypan blue staining. (C): Cells treated as indicated and PARP cleavage was analyzed by Western blots. (D): Cells were treated with the indicated agents for 24 hours. Each bar in (A), (B), and (D) shows mean ± SD of three independent experiments. Statistical differences are shown as p values. Abbreviations: Ly, Ly-294002; NS, not significant; p-Akt, phosphorylated Akt; PARP, poly(ADP-ribose) polymerase; PGE, prostaglandin E; PKI, protein kinase A inhibitor; Wort, wortmannin.

Discussion

Our results show that COX-2 plays a crucial role in conferring mES cell resistance to H2O2-induced apoptosis. COX-2 is constitutively expressed in mES cells and its inhibition by indomethacin, a nonselective COX-2 and COX-1 inhibitor, or SC-236, a selective COX-2 inhibitor, results in increased sensitivity to H2O2-induced apoptosis. We have provided strong evidence that the antiapoptotic action is mediated by COX-2-derived PGE2. First, PGE2 is produced and released into medium by mES cells and its production is completely abolished by selective COX-2 inhibitor. Second, exogenously added PGE2 rescues mES cells from apoptosis augmented by COX-2 suppression. Third, mES cells express PGES isoforms consistent with the presence of enzymes for PGE2 synthesis. It was reported that expression of COX-2 and mPGES-1 in inflammatory cells are coinduced by cytokines and endotoxins and are functionally coupled to catalyze robust PGE2 production [30]. As the level of mPGES-1 is barely detectable in mES cells, it is unlikely that COX-2 generates PGE2 via mPGES-1. COX-2 has been reported to couple to mPGES-2 [21]. It is possible that PGE2 may be generated through this isozyme. It was reported that cPGES is constitutively coexpressed with COX-1 and functionally coupled to COX-1 [22]. There are no reported data regarding coupling of COX-2 to cPGES. Since COX-1 is not detected in mES cells and cPGES is expressed in abundance, cPGES may act as a downstream enzyme of COX-2 for PGE2 biosynthesis. Further studies are needed to determine the PGES isoform that is responsible for converting COX-2-derived PGH2 to PGE2.

PGE2 was reported to protect against apoptosis in several cell types, including bone-marrow-derived and monocyte-derived dendritic cells [10, 31], neurons [32], epithelial cells, and colon cancer cells [12, 34]. However, PGE2 was also reported to induce apoptosis in articular chondrocytes and hippocampal neurons [33, 34]. The reason for the opposite actions of PGE2 is unclear. It may be due to different cell types or different apoptotic stimuli and may reflect the actions of different EP receptor isoforms. Our findings suggest that the antiapoptotic action of PGE2 in mES cells is most likely mediated through EP2 receptors. Protein analysis shows that mES cells express all receptor isoforms but EP4. Since an EP2 agonist but not an EP1/3 agonist protects mES cells from H2O2-induced apoptosis, it may be concluded that PGE2 interacts with the EP2 receptor through which it elicits the cellular effects.

The EP2 receptor belongs to the G-coupled prostanoid receptor superfamily [29]. It is classically coupled to G proteins, which signal via PKA. In an attempt to confirm signaling via PKA, we evaluated the effect of PKA inhibitors on the protective action of PGE2. To our surprise, neither PKI nor H89 blocked the antiapoptotic action of PGE2. By contrast, the antiapoptotic action of PGE2 was reversed by both PI-3K inhibitors to a similar extent, suggesting signaling via PI-3K. Our results reveal basal Akt activation, which is blocked completely by Ly-294002, consistent with constitutive activation of PI-3K to Akt in mES cells. Constitutive activation of the PI-3K/Akt system is likely induced by PGE2, as exogenous PGE2 is capable of restoring Akt activation to the basal level in indomethacin-treated cells. It is interesting to note that, in the absence of indomethacin treatment, exogenous PGE2 induces only a modest but significant increase in p-Akt (approximately 30%–40% over the basal level), suggesting that the constitutively produced PGE2 already exerts such a large effect on PI-3K/Akt activation that further supplement of exogenous PGE2 has only a modest additive effect. Taken together, these results suggest that PGE2 protects mES cells from cell death via an EP2-mediated PI-3K to Akt signaling pathway. Several recent reports have proposed coupling of EP2 to the PI-3K/Akt pathway. For example, the protective actions of PGE2 against radiation- and ceramide-induced apoptosis are considered to be mediated by EP2/EP4-mediated Akt activation [10, 12]. It is not entirely clear how EP2 is signaled via PI-3K/Akt. One possibility is that EP2 may signal via a cAMP-stimulated GDP exchange factor, Epac, which has been reported to serve as a novel intracellular cAMP receptor [35]. cAMP-bound Epac activates Rap1, which in turn activates PI-3K. It should be noted that neither our results nor the reported data provide evidence for a direct signaling relationship between PGE2-EP2 binding and the EP2-coupled PI-3K signaling. Further studies are needed to elucidate the mechanism by which PGE2 protects mES cells from apoptosis in a PI-3K/Akt dependent fashion. Regardless of the signaling mechanism, PGE2-induced Akt activation contributes to its antiapoptotic action. Activated Akt possesses several antiapoptotic properties [36]. One of its important properties is to phosphorylate Bad and thereby enhance Bad binding to 14–3-3 proteins and facilitate Bad sequestration in the cytosol. This results in reduced Bad translocation to mitochondria and prevents Bad-induced apoptosis via the mitochondria pathway [37].

In most differentiated somatic cells, COX-1 is constitutively expressed, whereas COX-2 is expressed after stimulation with diverse agents including cytokines, growth factors, and endotoxins. Our results show that COX-2 is constitutively expressed in mES cells, whereas COX-1 is undetectable. It has been reported that COX-2 is expressed in large abundance in hematopoietic progenitor cells, mesenchymal stem cells, and cancer cells [38, [39]–40]. Furthermore, constitutively expressed COX-2 declines when leukemia cells are induced to differentiate by phorbol 12-myristate 13-acetate [40]. Constitutive expression of COX-2 in mES cells may reflect its important physiological role in stem cells. One key role is to enable stem cells to resist damage by oxidative stress, serum depletion, and other cytotoxic conditions and maintain their unique properties of self-renewal and differentiation. Our results suggest that mES cells possess a transcriptional program that directs a cohesive expression of COX-2, cPGES, and EP2 to ensure abundant production of PGE2. The transcriptional mechanisms are unclear and require further studies. It is also unclear whether a similar program exists in human ES and adult stem cells. Work is in progress to evaluate this. When a similar program is confirmed in human stem cells, PGE2 should be valuable for expanding stem cells and for protecting stem cells from death during their differentiation into various types of cells for therapeutic use.

Summary

In summary, findings from this study demonstrate for the first time the expression of COX-2, PGES, and PGE receptors (EP1, EP2, and EP3) in mES cells and provide strong evidence for protection of mES cells from apoptosis by PGE2 produced by constitutively expressed COX-2. Our results suggest that the antiapoptotic action of PGE2 is mediated by EP2, which signals via the PI-3K to Akt pathway instead of the classic PKA pathway.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Acknowledgements

We thank Susan Mitterling and Nathalie Huang for editorial assistance. This work was supported by grants from National Institutes of Health, Grants R01-HL-50675 and P50-NS-23327, and Taiwan National Health Research Institutes. Bor-Sheng Ko is on leave from the National Taiwan University Hospital, Taipei, Taiwan.

Ancillary