PGE2 Promotes Apoptosis Induced by Cytokine Deprivation through EP3 Receptor and Induces Bim in Mouse Mast Cells

Increased mast cell numbers are observed at sites of allergic inflammation and restoration of normal mast cell numbers is critical to the resolution of these responses. Early studies showed that cytokines protect mast cells from apoptosis, suggesting a simple model in which diminished cytokine levels during resolution leads to cell death. The report that prostaglandins can contribute both to recruitment and to the resolution of inflammation together with the demonstration that mast cells express all four PGE2 receptors raises the question of whether a single PGE2 receptor mediates the ability of PGE2 to regulate mast cell survival and apoptosis. We report here that PGE2 through the EP3 receptor promotes cell death of mast cells initiated by cytokine withdrawal. Furthermore, the ability of PGE2 to limit reconstitution of tissues with cultured mast cells is lost in cell lacking the EP3 receptor. Apoptosis is accompanied by higher dissipation of mitochondrial potential (ΔΨm), increased caspase-3 activation, chromatin condensation, and low molecular weight DNA cleavage. PGE2 augmented cell death is dependent on an increase in intracellular calcium release, calmodulin dependent kinase II and MAPK activation. Synergy between the EP3 pathway and the intrinsic mitochondrial apoptotic pathway results in increased Bim expression and higher sensitivity of mast cells to cytokine deprivation. This supports a model in which PGE2 can contribute to the resolution of inflammation in part by augmenting the removal of inflammatory cells in this case, mast cells.


Introduction
Mast cells are long-lived tissue resident cells found throughout the body primarily in association with blood vessels, nerves, and in proximity to surfaces that interface the external environment. Upon activation, mast cells release inflammatory mediators, including histamine, proteases, prostaglandins, leukotrienes and cytokines [1]. Mast cells have an important function in several physiological as well as pathophysiological processes including host defense, especially in response to parasites, allergic reaction and inflammation. It is, therefore not surprising, that mast cell numbers increase at sites of inflammation during the course of the response. For example, elevated numbers of mast cells are observed in the submucosa of the lungs of asthmatics [2,3], allergy [4], rheumatoid arthritis [5,6], and chronic allergic dermatitis [6].
During the resolution phase of inflammation, homeostasis is reestablished in inflamed tissues and mast cell numbers decline. For many immune cells, survival at the site of inflammation is enhanced by cytokines, and therefore a decrease in local levels of these mediators, as the threat to the organism is neutralized, can lead to apoptosis. ''Cytokine withdrawal'' has been reported to activate an intrinsic (mitochondrial) apoptotic pathway in immune cells resulting in compromised mitochondrial integrity [7]. The mitochondrial integrity is guarded by Bcl-2 protein family members including anti-apoptotic proteins Bcl-2, Bcl-X L , Mcl-1, A1 and pro-apoptotic proteins Bax, Bak, Bim, Bid, Puma, Noxa, Bad, Bik, Bmf and Hrk. Mitochondrial outer membrane permeabilisation (MOMP) occurs when the balance of these factors is markedly disturbed. MOMP results in the release of principal killing factors such as cytochrome c and Smac/DIABLO from mitochondria to cytoplasm, where they contribute to the formation of apoptosome and activate aspartate-specific cysteine proteases (caspases) including initiator caspase-9. Caspase-9 in turn cleaves and induces the activation of downstream effector caspases that degrade and disassemble the cell [7].
Mast cell survival is regulated primarily by Stem cell factor (SCF), the ligand of c-kit receptor, through inactivation of the Forkhead transcription factor FOXO3a by MEK/MAPK-and PI3-kinase-mediated phosphorylation. Phosphorylation of FOXO3a also leads to phosphorylation, subsequent ubiquitination and proteasomal degradation of proapoptotic Bim and Puma.
Upon cytokine withdrawal, phosphorylation of FOXO3a decreases, followed by increase in Bim and Puma expression and apoptosis [8].
Prostaglandin E 2 (PGE 2 ), a bioactive mediator elevated at sides of inflammation, exerts its biological function through four distinct membrane-bound G-coupled receptors EP1-EP4. PGE 2 can contribute to resolution of inflammation by stimulating the expression of lipid mediators involved in the regulation of phagocytic clearance of apoptotic cells and by suppressing of the initial inflammatory response [9]. Here we examine the ability of PGE 2 to contribute to the resolution of inflammation, specifically, the removal of mast cells from inflammatory sites.

Mice
The generation of mice deficient in the EP 1 , EP 2 , EP 3 , and EP 4 receptors and mPGES1 has been previously reported [10][11][12][13][14]. All mice used were at least 8 wks old and were bred and maintained in specific pathogen-free animal facilities at the University of North Carolina (Chapel Hill, NC). Mice were killed by exposure to CO 2 follow by physical euthanasia prior to collection of cells. All experiments were carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee guidelines of the University of North Carolina (permit number:13-158).

Mast cell culture and cytokine depletion
Mast cells were derived from bone marrow isolated from mouse tibia and cultured in complete medium (Iscove's Modified Dulbecco's Medium supplemented with 12% fetal bovine serum (FBS), 100 U/ml of Penicillin and Streptomycin, 4 mM Lglutamine, 10 mM HEPES, 1 mM sodium pyruvate, 100 mM non-essential amino acid (all from GIBCO Life Technologies, Grand Island, NY, USA)), 10 ng/ml of recombinant mouse IL-3 and 20 ng/ml of recombinant mouse SCF (both from Sigma-Aldrich, St. Louis, MO, USA) for 4 weeks at 37uC in a humidified incubator with 5% CO 2 atmosphere. Medium was changed weekly and adherent cells were removed regularly from cultures by transfer of non-adherent cells to a new culture dish. Maturity of non-adherent BMMC was assessed by measuring surface expression of FceRI and c-Kit. Only cultures with .95% cells positive for both receptors and ,12 weeks in culture were used for experiments.
For depletion, BMMC were washed twice in phosphate buffered saline (PBS) and re-suspended in depletion medium (the same as the culture medium but without IL-3 and SCF) at a concentration 1610 6 cells per 1 ml. Cells were cultured at 37uC in a humidified incubator with 5% CO 2 atmosphere.

Cell death and apoptosis analysis
7AAD. Cells depleted of cytokines for various times were stained for 15 min in 7-AAD (BD Bioscience, Franklin Lakes, NJ USA), 5 ml per test (1610 6 cells) and analyzed using Fluorescenceactivated cell sorting (FACS) on a CyAn ADP analyzer (Beckman Coulter, Inc., Brea, CA).
Mitochondrial membrane potential. Cells were stained using JC-1 from a Flow cytometry mitochondrial membrane potential detection kit (BD Biosciences, Franklin Lakes, NJ USA) according to the manufacturer's protocol. Green and red fluorescence ratios were measured on CyAn ADP analyzer (Beckman Coulter, Inc., Brea, CA).
Caspase-3 activation. Caspase activation was measured with the APO LOGIX carboxyfluorescein (FAM) caspase detection kit (Cell Technology, Inc., Mountain View, CA, USA) specific for caspase-3 (FAM500-2) according to the manufacturer's protocol. Fluorescence was measured on a CyAn ADP analyzer (Beckman Coulter, Inc., Brea, CA) and results were verified by western blot using an antibody specific for cleaved caspase-3 antibody (#9661, Cell Signaling Technology, Danvers, MA, USA).
Low molecular weight cleavage assessment. 3610 6 BMMC depleted of cytokines for 24 h were lysed in lysis buffer (1% SDS, 0.1 M NaCl, 0.1 M EDTA, 0.05 M Tris base, 0.5 mg/ ml proteinase K, 0.1 mg/ml RNase A, pH 8) for 2 h at 55uC. 135 ml of saturated NaCl solution was added to each sample, mixed and centrifuged for 10 min at 16,000 g. Supernatant (400 ml) was removed, added to 800 ml 100% ethanol and precipitated for 20 min at 220uC. Precipitated DNA was pelleted by centrifugation at 16,000 g and washed with 0.5 ml 70% ethanol and 0.5 ml 100% ethanol. The pellet of DNA was air dried for 10 min at room temperature to remove trace of ethanol and then dissolve in 30 ml TE buffer (10 mM Tris base, 1 mM EDTA). Loading buffer was then added. Samples were separated by electrophoreses on 1.2% agarose gel 5 V/cm for 1-2 h. The gel was stained with ethidium bromide and visualized by transillumination with UV light and photographed.

Isolation of peritoneal mast cells
Peritoneal cells from 5 congenic mice were pooled and washed once with PBS. Cells were re-suspended in 8 ml of 70% Percoll solution (7 ml Percoll (GE healthcare Bioscience AB, Upsala, Sweden), 1 ml 106 concentrated PBS, 1.9 ml H 2 O, 0.1 ml FBS) and overlaid with 2 ml of peritoneal mast cell medium (PMC: DMEM, 20 mM HEPES, 5% FBS). Cells were centrifuged 650 g at room temperature for 15 min. Supernatant was discarded and mast cell pellets were re-suspended in 0.5 ml PMC medium, transferred to new tubes and washed once in 10 ml pre-warmed PMC medium.  . Calcium concentrations were calculated as described previously [15]. For inhibition of intracellular calcium with BAPTA was added to cell suspension 5 min before stimulation with PGE 2 .

Gene expression assay
RNA was isolated with RNAbee (Tel-Test Inc., Friendswood, TX) according to the manufacturer's protocol and cleaned with the Total RNA mini purification kit (Denville Scientific Inc., Metuchen, NJ, USA). Reverse transcription was carried out with the High Capacity cDNA transcription kit (Applied Biosystems, Life Technologies, Grand Island, NY, USA). Obtained cDNAs were used for relative quantitative analysis by real-time PCR using comparative CT method. The following primers and detection TaqMan MGB probes were used (Bcl2l11 (Bcl-2-like protein 11/ Bim) # Mm00437796, Bbc3 (Bcl-2-binding component 3/Puma) #Mm00519268) for cDNA amplification by ABI Prism 7 900HT detection system in TaqMan universal PCR master mix (all from Applied Biosystems, Life Technologies, Grand Island, NY, USA) according to the manufacturer's protocol. The gene expression level was expressed as a relative expression of the gene of interest to GAPDH expression level in each sample.
Reconstitution of mast cells in ear of W sh/sh mice W sh/sh mice were lightly anesthetized and the pinna of the ears reconstituted intradermally (i.d.) with various numbers of 5-weekold cultured BMMC in 40 ml PBS. In other experiments mice received intradermally to both ears 5610 5 BMMC, treated or not with 1610 26 M PGE 2 for 20 min and washed twice with PBS to remove PGE 2 before injection.

Passive cutaneous anaphylaxis
Animals were lightly anesthetized, and the pinna of the right ears injected i.d. with 8 ng of murine monoclonal anti-DNP IgE in 20 ml of PBS. The left ears received 20 ml of PBS, i.d. Twenty-four hours later, animals were injected i.v. with 100 ml of PBS containing 100 mg of DNP-albumin and 1% Evan's Blue dye. Animals were killed 90 min after i.v. injection, and pinna of the ears were removed close to the base and incubated in 1 ml of formamide at 54uC for 48 h. Quantitative analysis of formamide extracts was determined by measuring the absorbance of Evan's Blue at 610 nm with a spectrophotometer.

Statistical analysis
Data are represented as means 6 SEM. Statistical significance was assessed by the Student's two-tailed t test. When three or more groups or variables were compered, statistical significance was determined by ANOVA. A P value of ,0.05 was considered as statistically significant

PGE 2 increases cell death during cytokine deprivation
Interleukin-3 (IL-3) and SCF are critical to mast cell survival, both in vitro and in vivo. We took advantage of this to model cell death of mast cells in vitro. Mast cells derived from bone marrow of C57BL/6J mice (BMMC) in complete medium containing IL-3 and SCF, were deprived of these cytokines to initiate cell death, either in the presence or absence of PGE 2 . Cell death was evaluated as a percentage of 7-AAD-stained cells at various times after cytokine withdrawal. Cytokine deprivation resulted in an increase in the number of 7-AAD-stained cells, reaching 30% 24 h and 44% 36 h after cytokine removal. Addition of 1610 26 M PGE 2 to the cytokine depleted medium led to a further increase in number of 7-AAD stained cells compared to untreated BMMC with 48% and 76% of the cells staining with the dye 24 h and 36 h after cytokine removal, respectively. Nevertheless, addition of 1610 26 M PGE 2 to complete medium did not result in an increase of 7-AAD stained cells (Fig.1A). The difference in numbers of 7-AAD-stained cells during cytokine withdrawal in the presence or absence of PGE 2 was significant (P,0.05) in concentrations from 1610 26 M to 1610 27 M ( Fig 1B). Interestingly, pretreatment of mast cells with PGE 2 for only 10 min in complete medium follow by cytokine deprivation without further addition of PGE 2 to the medium was sufficient to increase the number of 7-AAD positive cells to a similar magnitude as observed when PGE 2 was present throughout the experiment (Fig 1C). This suggests that PGE 2 can affect cell survival even after its concentration return to homeostatic levels.

Increased cell death after PGE 2 treatment is caused by apoptosis
To further characterize the PGE 2 mediated increase in BMMC death, we measured loss of plasma membrane integrity and extracellular translocation of phosphatidylserine (PS) with APClabeled Annexin V. Twenty hours after cytokine withdrawal, about 40% of BMMC stained with Annexin V. Treatment with PGE 2 increased this number to 45% of BMMC ( Fig. 2A). A characteristic feature of cell apoptosis is a loss of mitochondrial membrane potential (DY m ), which can be measured by cationic dye JC-1 as a potential-dependent ratio of red and green fluorescence intensity. As shown in Fig. 2B, the decrease of red fluorescence intensity, a sign of dissipation of DY m during cytokine deprivation, was significantly higher in PGE 2 treated BMMC compared to un-treated BMMC. Caspase-3 is a critical executioner of apoptosis. Its activation requires proteolytic processing of the inactive zymogen into activated p17 and p12 fragments. Western blot analysis with antibody specific to the cleaved p17 fragment showed higher levels of this fragment in lysates prepared from the cytokine deprived PGE 2 treated BMMC compared to similar cultures without PGE 2 ( Fig 2C). Consistent with this, PGE 2 treatment of cytokine-deprived cells increased the number of BMMC that stained with the fluorescently labeled inhibitor of caspase-3 (FLICA) FAM-DEVD-FMK (Fig 2D). Similarly to double staining with Annexin V in Fig. 2A, the majority of 7-AAD-stained cells are also stained with FLICA ( Fig. 2D), suggesting that the increase in cell death shown in Fig 1 is directly related to increase in caspase-3 activation and apoptosis.
Characteristic hallmarks of the late phase of apoptotic execution are low molecular weight cleavage (LMW) of DNA and chromatin condensation. Cytokine deprivation of both PGE 2 treated and untreated BMMC showed a pattern of DNA fragmentation consistent with apoptosis ( Fig 2E). Similarly, the number of cells with condensed chromatin was higher in BMMC treated with PGE 2 compare to untreated BMMCs ( Fig 2F). The observation of multiple events characteristic of apoptosis in BMMC after cytokine withdrawal, and the increase in the magnitude of these changes in PGE 2 treated cells, is consistent with a model in which PGE 2 augments the apoptotic pathway initiated by cytokine withdrawal.

PGE 2 activation of the EP3 receptor increases apoptosis
BMMC express all four PGE 2 receptors [16]. Therefore BMMC were derived from mice lacking each of these receptors and the appropriate controls. The impact of the loss of each of the four PGE 2 receptors on the ability of 1610 26 M PGE 2 to augment apoptosis initiated by cytokine withdrawal was assessed by determining the number of 7AAD positive cells in the cultures 20 hours after initiation of the experiment. The absence of the EP1, EP2 or EP4 receptors did not alter the augmentation of 7AAD positive cells in the PGE 2 treated cultures (Fig. 3A). In contrast, no increase in the number of 7AAD staining cells was observed in PGE 2 treated EP32/2 BMMC deprived of cytokines, compared to untreated cultures (Fig 3A). During the experiment we noted some variability between different clones of mast cells. Although, mast cells generally do not produce large amounts of  , and its quantitative analysis expressed as a ratio of red PGE 2 , it is possible that during cytokine deprivation and cell death, endogenous PGE 2 is released and thus PGE 2 produced by the mast cell itself augments apoptosis. This can contribute to differences in cell death seen in various mast cell cultures. To determine the contribution of such an autocrine loop to mast cell apoptosis, BMMC were derived from mice lacking mPGES1 (mPGES12/2), the primary PGE 2 synthase. mPGES12/2 BMMC showed a similar rate of cell death during cytokine withdrawal and were equally sensitive to a PGE 2 mediated increase in cell death compare to wild type (WT) cells ( Fig 3B). These results show that in this model exogenous rather than endogenous PGE 2 mediates the increase in mast cell death during cytokine withdrawal.
The absence of EP3 on BMMC eliminated PGE 2 mediated augmentation of the loss in mitochondrial membrane potential during cytokine depletion (Fig 3C). The activity of caspase-3, measured by presence of the cleaved p17 fragment after cytokine depletion, was not increase by PGE 2 treatment of EP32/2 BMMC, while WT, EP22/2, and EP42/2cells showed higher levels of caspase-3 activation (Fig. 3D). Similarly, caspase-3 activity measured by FLICA staining showed no increase in EP32/2 BMMC when treated with PGE 2 (Fig.3E). Sulprostone, a metabolism resistant synthetic analog of PGE 2 , is a EP3 receptor preferring agonist. Pretreatment of mast cells with various concentration of sulprostone in complete medium prior to cytokine depletion resulted in concentration dependent increase in cell death measure by 7-AAD. The similar actions of PGE 2 and sulprostone support the role of EP3 in mast cell apoptosis following cytokine withdrawal (Fig. 3F).
Ex vivo and in vivo evaluation of mast cell survival after exposure to PGE 2 BMMCs are widely used as a model for studying mast cells in vitro; however they are considered less mature compared to residential mast cells. To verify our finding in primary mast cells, peritoneal mast cells (PMC) from WT and EP3 deficient animals were isolated on Percoll gradient with more than 98% efficiency as describe previously [17]. PMC were cultured either in complete medium, in cytokine-depleted medium, or in cytokine-depleted medium with 1610 26 M PGE 2 for 16 h. Viability was assessed by determining the number 7AAD positive cells. As expected, a marked increase in the number of 7-AAD staining cells was observed in cytokine-deprived cultures. Similarly to BMMCs the number of 7-AAD cells in the cytokine deprived cultures was significantly increased by addition of PGE 2 . This action of PGE 2 was dependent on the EP3 receptor as PGE 2 treatment did not increase the number of 7-AAD stained cells in PMC isolated from EP3 deficient animals (Fig 4A). PGE 2 treatment of WT PMC, but not EP32/2 PMC also increased the number of cells staining with caspase-3 specific FLICA compared to untreated cells (Fig. 4B).
Study of the role of PGE 2 in apoptosis in vivo is complicated by the fact that in healthy tissue PGE 2 may act to recruit circulating mast cells progenitors and this action may also be mediated by the EP3 receptor [18,19]. After PGE 2 treatment of mice it is not possible to distinguished cells mobilized to the site towards the PGE 2 gradient and cells already present at the site of PGE 2 application. Use of sulprostone does not alleviate the problem, since some of the pro-migratory actions of PGE 2 are mediated through the EP3 receptor. Therefore, to begin to address the in vivo role of EP3 in apoptosis, we developed the following model. The skin of mast cells deficient mice (W sh/sh ) is reconstituted locally by injection of excessive numbers of BMMC, with the anticipation that only small number of cells will survive and stably repopulate tissue. We use the decrease in the amount of transferred BMMC to the ear of W sh/sh mice to evaluate the pro-apoptotic actions of PGE 2 and assess the ability of the EP3 pathway to contribute to mast cell death in vivo. EP32/2 and wild type BMMC were briefly exposed to PGE 2 washed twice with PBS and then used to reconstitute the pinna of the ear of mast cells deficient W sh/sh mice. After 10 days the survival of the mast cells was compared indirectly by examining the susceptibility of the tissue to suboptimal passive cutaneous anaphylaxis. PCA mediated by exogenous IgE is entirely dependent on mast cells [20]. Accordingly, the extravasation of serum proteins characteristic of this response shows a linear relationship with the number of mast cells introduced into tissue (Fig. 4C). Treatment of wild type but not EP32/2 BMMC with PGE 2 decreased the anaphylactic response in the reconstituted animals (Fig. 4D). This decrease is consistent with a decrease in survival of mast cells after activation of the EP3 receptor in tissue lacking inflammatory cytokines.

PGE 2 mediated increase in apoptosis is dependent on intracellular calcium release and activation of MAPK
To identify the pathway(s) responsible for the PGE 2 dependent increase in apoptosis during cytokine withdrawal, we compared activation of EP32/2 and WT mast cells, identified pathways activated through this receptor and then evaluated the potential of each of these pathways to contribute to mast cell apoptosis. As we have shown, PGE 2 treatment prior to cytokine removal can augment apoptosis (Fig.1C). We used this protocol to avoid complexities in studying PGE 2 pathways after cytokine withdrawal.
Increase in intracellular calcium mobilization after mast cell activation by PGE 2 is exclusively dependent on the expression of the EP3 receptor [16]. To assess the role of intracellular calcium in apoptosis, we treated BMMC with 100 mM, 50 mM, or 25 mM of BAPTA/AM, a cell-permeating calcium chelator, for 5 min prior to EP3 receptor activation with 1610 26   Activity of caspase-3 in WT, EP2, EP3, and EP4 deficient BMMC identified by specific antibody to cleaved p17 caspase-3 fragment. Data are representative of 3 experiments for EP22/2 and EP42/2 and 4 experiments for WT and EP32/2 using 1 culture of each genotype. E. Staining by cells with either 50 mM or 100 mM BAPTA completely inhibited intracellular calcium release and partial inhibition of the response was observed in 25 mM BAPTA treated cells (Fig. 5A). BMMC were treated with this suboptimal dose of BAPTA for 5 min prior to addition of 1610 26 M PGE 2 . After 10 minutes, cytokines, PGE 2 and BAPTA were removed and cells were cultured in cytokine-depleted medium for 20 h. BAPTA treatment alone did not alter the apoptosis induced by cytokine deprivation, but completely blocked the PGE 2 mediated increase in apoptosis (Fig. 5B).
MAPK Erk 1/2 as well as p38 are activated by PGE 2 [21]. PGE 2 mediated activation of both MAPKs, assessed by phosphorylation at T202/Y204 and T180/Y182 respectively, is EP3 receptor dependent (Fig 5C and data not shown). Treatment of cells with PD98059, an inhibitor specific for kinase of Erk1/2, MEK1 kinase, blocked PGE 2 induced T202/Y204 Erk 1/2 phosphorylation in a dose dependent manner (Fig 5D lower  panel). Erk1/2 inhibition was sufficient to completely inhibit the increase in the number of 7-AAD stained cells in BMMC pretreated with PGE 2 prior to cytokine deprivation (Fig. 5D upper  panel). caspase 3 specific FLICA; example of FACS histogram (upper panel) and its quantitative analysis (lower panel). Data are from 4 independent experiments of 1 culture of WT BMMC and EP32/2. BMMC were cultured in complete medium (CTRL or +), medium without cytokines (no cyt. or -) and medium without cytokines in the presence 1610 26 M PGE 2 (no cyt. + PGE 2 , or P) for 16 h. Student's two-tailed t test was used to evaluate statistical differences between cytokine deprive mast cells and cytokine deprived mast cells treated with PGE 2 in A, C, and E. Statistical significance: * = P,0.05, ** = P,0.01. F. Increase in mast cell death after 20 min treatment with various concentration of sulprostone. Cells cultured for 16 h in complete medium (CTRL), medium without cytokines (no cyt.) and stain with 7AAD. Data from 3 independent experiments of 1 culture of WT BMMC. doi:10.1371/journal.pone.0102948.g003  Activation of mast cells with 1610 26 M PGE 2 resulted in rapid Akt activation, measured by an increase in phosphoryplation at S473 (Fig. 5E upper panels). In EP32/2 BMMC, PGE 2 induced phosphorylation of S473 was completely abolished (Fig. 5E lower  panels). Treatment of WT mast cells with LY294002, an inhibitor of PI3 kinase responsible for activation of Akt, at concentration higher than 10 mM resulted in complete inhibition of PGE 2 induced S473 phosphorylation. However, inhibition of Akt activation by LY294002 did not alter the ability of PGE 2 to augment BMMC apoptosis (Fig. 5F). PGE 2 -stimulated MAPK activation and increase in apoptosis upon cytokine withdrawal is Calmodulin kinase II dependent MAPK activation can be both calcium dependent and calcium independent. To determine the relationship between intracellular calcium mobilization and MAPK phosphorylation triggered by EP3 receptor, we treated BMMC with the intracellular calcium inhibitor BAPTA before addition of 1610 26 M PGE 2 . Inhibition of intracellular calcium resulted in a complete inhibition of PGE 2 induced Erk1/2 and p38 phosphorylation at BAPTA concentrations greater than 25 mM (Fig. 6A, data not shown). Erk 1/2 could be activated by intracellular calcium through several pathways including calcineurin and calmodulin dependent kinase II (CamKII) [22][23][24]. However, we found that activity of calcineurin in WT and EP3 2/2 cells significantly decreased after PGE 2 stimulation. Moreover a specific inhibitor of calcineurin, Cyclosporine A, was not able to inhibit Erk1/2 activation (data not shown).
To investigate a possible role of CamKII in the activation of MAPK in BMMC by PGE 2 , an inhibitor with specificity for CamKII, KN-93, was used. As shown in Fig. 6B, KN-93 effectively inhibited the PGE 2 mediated increase in apoptosis during cytokine withdrawal in a concentration dependent manner ranging from 125 to 500 mM. At higher concentrations this agent was toxic to BMMC. The inhibition of CamKII with KN-93 correlated with inhibition of Erk 1/2 (Fig. 6C).

PGE 2 increases expression of Bim
Our results suggested that PGE 2 augmented apoptosis in mast cells through a signaling pathway that involves intracellular calcium release and Erk 1/2 phosphorylation. As shown in Fig.2 B, an increase in apoptosis is apparent at the mitochondrial level, suggesting that activated Erk 1/2, or events downstream of Erk 1/ 2 are able to upregulate intrinsic apoptotic pathway prior to MOMP. Bcl-2 family members Bim and Puma have been implicated in previous studies examining mast cells apoptosis induced by cytokine withdrawal [8]. Additionally, Erk1/2 was shown to regulate expression of c-Jun, an established positive regulator of Bim expression [25][26][27]. We therefore asked whether the activation of the Erk1/2 pathway by PGE 2 was paralleled by alteration in the expression of genes known in these pathways. Expression of Bim and Puma in BMMC treated with either vehicle or PGE 2 was assessed by qPCR 10 h after cytokine withdrawal. As expected, removal of cytokines from the culture medium led to an increase in expression of both Bim and Puma (genes Bcl2l11and Bcc3) compared to cells growing in complete medium (Fig. 7A). Pretreatment of BMMC with PGE 2 for 20 min prior to cytokine withdrawal significantly increase Bim expression. In contrast, no difference between PGE 2 and vehicle treated cultures was observed in the expression of Puma. Withdrawal of cytokines increased Bim expression also in EP32/2 cells; however, these cells were resistant to a further increase in Bim expression after PGE 2 treatment (Fig. 7B). Inhibition of the Erk 1/2 pathway with PD98059 prior to PGE 2 treatment, also prevent the increase in Bim expression (Fig. 7B). Previous studies have shown that PGE 2 can induce apoptosis of lymphocytes, and that this response is dependent on c-Myc although it is Bcl-2 independent [28]. RNA from mast cells cultured in complete medium, mast cells depleted from cytokines for 10 h, and mast cells pretreated by 10 26 M PGE 2 before 10 h depletion was analyzed for both c-Myc and Bcl-2 expression. Expression of both genes was significantly decreased in RNA samples prepared from cells after removal of cytokines. The presence of PGE 2 did not prevent the decrease in either c-Myc or BCL-2 (data not showed). Therefore, PGE 2 increase apoptosis induced by cytokine withdrawal is likely c-Myc and Bcl-2 independent.
To further delineate mechanism of Bim regulation, we evaluated changes in the protein levels of the various Bim isoforms and assessed cells for the appearance of c-Jun after PGE 2 stimulation. c-Jun levels increased rapidly in mast cells in response to PGE 2 , with maximal level seen by 2 h (Fig. 7C). An increase was also seen after 2 h of cytokine withdrawal and the increase was greater in mast cells exposed to PGE 2 (Fig. 7D, left). This augmented production of c-Jun was EP3 dependent as no observed increase in c-Jun was observed in lysates from EP32/2 cells after EP3 stimulation (Fig 7D, right panel). In the presence of cytokines only 2 isoform of Bim were identified in BMMC lysates (Bim EL and Bim L ) (Fig. 7C). 5 min stimulation with 1610 26 M PGE 2 lead to dramatic change in mobility of Bim EL and Bim L isoforms on western blot as both Bim isoforms showed multiple bands probably due to their phosphorylation (Fig. 7C). However, total Bim protein level did not significantly changed within 6 h of PGE 2 activation (Fig. 7C). In the absence of cytokines, Bim protein level increased within 2 h of depletion and also low expression of Bim S was apparent on long exposure of films (Fig. 7D,E). The presence of PGE 2 resulted in a dramatic increase in the amount of Bim S isoform 2-10 h after cytokine depletion in WT cells. 10 h after cytokine depletion, PGE 2 stimulation also slightly increased levels of other Bim isoforms (Fig. 7E). Cytokine depletion of EP32/2BMMC also resulted in increased levels of Bim, however, no further increase in expression of Bim was observed in cells exposed to PGE 2 (Fig 7D,E). Together this suggest that cytokine withdrawal and EP3 activation work together to modulate the levels of Bim and to alter the ratios of the various Bim isoforms present in the mast cells.

Discussion
In this manuscript we analyzed the role of PGE 2 in mast cell apoptosis during cytokine withdrawal. In our initial observation we found that exposure to PGE 2 increases mast cell death during cytokine deprivation. However, PGE 2 was not able to induce cell death in the presence of anti-apoptotic stimuli (grow factors and cytokines). This suggests that PGE 2 is not an inducer of cell death, but rather it activates signaling pathway(s) which act in synergy independent experiment of 1 culture. F. Pharmacological inhibition of Akt phosphorylation triggered by PGE 2 (lower panel) and increase in apoptosis (upper panel), BMMC were activated by 1610 26 M PGE 2 . For inhibition, cells were pretreated with various concentration of LY294002 for 10 min before PGE 2 activation, n = 6 independent experiments of 2 culture of BMMC. Differences between groups in B were evaluated by Student's two-way t-test, for D and F ANOVA test was used. Statistical significance: * = P,0.05, ** = P,0.01, *** = P,0.001. doi:10.1371/journal.pone.0102948.g005 with pro-apoptotic pathways triggered by withdrawal of cytokine. Analysis of the mitochondrial membrane potential, caspase-3 and late phases of apoptosis indicated that the higher number of dead cells in cultures treated with PGE 2 during cytokine deprivation is caused by an increase in apoptosis. This further supports the synergy of PGE 2 -triggered and apoptotic pathways. Moreover, the up-regulation of all apoptotic events we analyzed suggested that the PGE 2 -triggered signaling pathway modulates apoptotic events before or during MOMP. Interestingly, pretreatment of BMMC with PGE 2 shortly before cytokine deprivation, and subsequent absence of PGE 2 during deprivation, was efficient in potentiation of apoptosis. This raises the possibility that pathways triggered by PGE 2 , or products of this pathway remained activated for a relatively long period of time. A possible explanation for these phenomenons is the induction of transcription of pro-apoptotic factors from Bcl-2 family by PGE 2 which regulate MOMP. This is consistent with our overall finding that the mechanism by which the PGE 2 /EP3 pathway acts is increasing Bim transcription after cytokine withdrawal. This is also supported by the previous findings of others that cross-linking of FceRI and activation of mast cells, before their depletion from cytokines induce Bim and modulate mast cell survival [29]. In neutrophils, GM-CSF treatment temporarily blocks apoptosis by inducing anti-apoptotic factors with rapid turnover and pro-apoptotic factors including Bim, which limit GM-CSF-mediated prolonged survival of neutrophils [30]. Binding sites for several transcription factors including Foxo, c-myb, and c-jun are present in the Bim promoter [31,32]. While c-myb and c-jun can be activated by MAPK, Foxo expression is negatively regulated by Akt phosphorylation. It is interesting to speculate that upregulation of Foxo during cytokine withdrawal cooperates with induction of c-jun and/or c-myb to facilitate increased Bim transcription. Increased expression of cjun and c-myb may be in part secondary to Erk activation by PGE 2 (Fig.8). Bim undergoes alternative splicing to produce three splicing variants (Bim EL , Bim L and Bim S ). The smallest variant Bim S is the most potent inducer of apoptosis [33]. Interestingly, Bim S isoform was upregualated by EP3 dependent PGE 2 activation as well as by cytokine withdrawal. PGE 2 through EP3  receptor cause changes in mobility of Bim L and Bim EL proteins due to phosphorylation, which alter Bim binding capacity with anti-apoptotic members of Bcl-2 family as well as its degradation. Balance between degradation of phosphorylated Bim EL and Bim L isoforms and up regulated expression of all 3 isoforms by transcription factors is dramatically shifted during cytokine withdrawal of PGE 2 treated BMMC toward increased expression and apoptosis.
Sustained intracellular calcium release has been reported as a pro-apoptotic stimulus in many cells [34]. An increase in intracellular calcium is also a distinguishing event during mast cell activation by PGE 2 [16,21]. Indeed, we found that calcium is required for both PGE 2 -triggered up-regulation of apoptosis and Erk1/2 activation in mast cells ( Fig. 4 and 5). Interestingly, partial inhibition of intracellular calcium release (25 mM BAPTA) was sufficient to inhibit both upregulated apoptosis and Erk 1/2 phosphorylation ( Fig. 4 and 5), suggesting a need for strong intracellular calcium activation. Although this data implies that intracellular calcium acts on apoptosis through activation of Erk 1/2, we cannot exclude the possibility that part of the increased apoptosis after PGE 2 treatment is caused directly by intracellular calcium release. However, MEK inhibition of the PGE 2 -depen-dent increase in Bim expression makes this unlikely. Our data support a model in which EP3/calcium/CamKII/Erk/Bim signaling is the PGE 2 -triggered pathway responsible for increased apoptosis (Fig 8).
The role of Erk1/2 in cell survival is well characterized; however, recent evidence suggests that the activation of Erk1/2 also contributes to cell death. Erk 1/2 activation was shown to contribute to apoptosis by cisplatin [35], heavy ion irradiation therapy using a carbon beam [36], exposure to reactive oxygen species [37] or deprivation of pro-survival factors [38]. Erk1/2 can potentiate apoptosis by upregulation of caspase-3 activation, induction of p53 and other pro-apoptotic factor, expression of death ligands, and suppression of pro-survival activity of Akt (for review see [39]). Although these studies support the pro-apoptotic nature of Erk1/2 phosphorylation, the basis for Erk1/2 signaling in those events is not clear. It has been reported that prolonged Erk1/2 activation leads to a pro-apoptotic effect of Erk1/2 while transient activation protects against apoptosis. This is consistent with recent results obtained from mast cells deficient in SHP1 phosphatase, which is responsible for dephosphorylation of Erk1/2 [40]. Stimulation of cells with a low concentration of IL-3 protects against apoptosis in SHP1 negative cells, while a higher independent experiments of 2 culture of WT and EP32/2 BMMC. Student's two-tailed t test was used to evaluate statistical differences between cytokine deprive mast cells and cytokine deprived mast cells treated with PGE 2 in A nad B. Statistical significance: * = P,0.05. doi:10.1371/journal.pone.0102948.g007 Figure 8. Schematic depicting of EP3 induced signaling pathway contribution to apoptosis during cytokine withdrawal. Withdrawal of cytokines results in activation of intrinsic apoptotic pathway due to imbalance between Bcl-2 anti-apoptotic protein family members and BH3 only pro-apoptotic proteins. Intrinsic apoptotic pathway in mast cells is upregulated by PGE 2 via EP3 receptor and Ca 2+ -CaMII-ERK1/2 pathway. While Erk 1/2 mediated phosphorylation of Bim EL and Bim L immediately after PGE 2 stimulation and this leads to ubiquination and degradation of these isoforms, Erk also induces expression of c-Jun, a potent upregulator of Bim expression. Final outcome of Erk regulation of Bim is a balance between Bim degradation after its phosphorylation and increase in its expression due to c-Jun. Data presented here support view that increase expression of Bim and proapoptotic action of EP3-Ca 2+ -CamKII -Erk1/2 pathway is prevalent in mast cells in presence of PGE 2 and during cytokine withdrawal. doi:10.1371/journal.pone.0102948.g008 concentration of IL-3 promotes apoptosis and shows strong prolonged activation of Erk1/2. This suggests that the kinetics and duration of Erk1/2 activation might determine the effect of Erk1/2 on cell fate. Strong phosphorylation of Erk1/2, found after activation of mast cells with PGE 2 supports this view. Several previous reports have described the ability of PGE 2 to either potentiate cell death [28,41,42] or protect against it [43,44]. The contradictory outcomes reported likely reflect that the impact of PGE 2 on cell death depends on cell types and death stimuli. Stimulation of EP2 and EP4 is mostly reported as anti-apoptotic, while stimulation of EP3 and EP1 are reported as pro-apoptotic or without an effect. It is likely that the relative expression of each EP receptor determines the final outcome of PGE 2 . For example, neutrophils express relatively more EP2 and EP4 receptors compared to EP3, while in BMMC, EP3 is the predominant EP receptor. Treatment of neutrophils with PGE 2 protects them from apoptosis [45,46], however treatment of BMMC and peritoneal mast cells by PGE 2 as we report here up-regulates apoptosis. Similarly to mast cells, studies from other cells have found that the EP3 receptor augments apoptosis. In human neutrophils, a selective agonist of EP3, ONO-AE-248, induces death by disruption of mitochondrial potential without chromatin condensation, DNA fragmentation, or expression of PS on the plasma membrane [47]. EP3 was shown to mediate neurotoxicity on exposure to PGE 2 during ischemia, or ischemic stroke through enhancement of inflammatory and apoptotic reactions in the ischemic cortex [41]. In colon carcinogenesis, EP3 plays an important role in suppression of cell growth and its down regulation enhances the late stages of this disease [48].
In summary, even brief and transient exposure to physiologically relevant levels of PGE 2 can increase mast cell apoptosis during cytokine deprivation. We provide evidence that PGE 2 mediates this action solely through the EP3 receptor. This increase in mast cell apoptosis is dependent on synergy between the EP3/ calcium/CamKII/Erk pathway and intrinsic apoptotic pathways. Our analysis suggests that this pathway increases Bim expression and also may alter the relative levels of the various Bim isoforms, in particular the pro-apoptotic Bim S isoform. The unique ability of PGE 2 to increase apoptosis in mast cells could play an important role during resolution of allergic inflammation and suggest that while Nonsteroidal anti-inflammatory drugs (NSAID) can limit many symptoms of inflammation they also may prolong recovery in some instances.