Central role for protein kinase C in oxytocin and epidermal growth factor stimulated cyclooxygenase 2 expression in human myometrial cells

Prostaglandins are important mediators of uterine contractility and cervical ripening during labour. Cyclooxygenase-2 (COX-2), also known as prostaglandin-endoperoxide synthase 2, is a rate limiting enzyme involved in the conversion of arachidonic acid into prostaglandins at parturition. In this paper, the pathways underlying agonist-induced cyclooxygenase-2 expression in human myometrial cells were studied. Myometrial cells were stimulated with different agonists: oxytocin (OXT), epidermal growth factor (EGF), interleukin-1β (IL1β), and phorbol-12-myristate-13-acetate (PMA) alone and in the presence of specific signalling pathway inhibitors. The nuclear factor kappa-light-chain-enhancer of activated B cells (NFKB) pathway was inhibited by means of the IKK-2 inhibitor TPCA-1. Signalling through extracellular signal-regulated kinases (ERK) was inhibited using the MEK1/2 inhibitor PD-184352. Bisindolylmaleimide-I was used to inhibit protein kinase C (PKC) signalling. COX-2 expression and ERK phosphorylation were measured using immunoblotting. OXT induced COX-2 expression by activating PKC and ERK. EGF increased COX-2 expression via stimulation of PKC, ERK and NFKB. As expected, the pro-inflammatory cytokine IL1β induced COX-2 expression by activating PKC- and NFKB-dependent pathways. Stimulation of PKC directly with PMA provoked strong COX-2 expression. PKC plays a central role in OXT and EGF induced COX-2 expression in human myometrial cells. However, other pathways, notably ERK and NFKB are also involved to an extent which depends on the type of agonist used.


Background
Preterm birth is defined as birth occurring prior to 37 weeks of gestation. It affects 5-9% of all pregnancies and is associated with more than 50% of long-term morbidity and 75% of neonatal mortality [1,2]. Preterm birth is a worldwide problem affecting developed and developing countries. 15 million babies are born before 37 weeks gestation every year, with one million babies dying from complications [3]. Approximately 7 million children under 12 are affected by sequelae of prematurity, ranging from mild to severe depending on their gestational age at delivery. For example, infants born before 34 weeks gestation have a 40 times higher frequency of cerebral palsy than infants born at term, and the incidence is 90 times higher in infants born before 28 weeks [4,5].
Preterm birth can occur spontaneously (idiopathic) or can be medically induced (iatrogenic) for maternal or fetal indications, such as preeclampsia or intrauterine growth restriction [6]. The underlying mechanisms resulting in spontaneous preterm labour have not been fully elucidated; it is thought that it is either caused by a pathological insult or by an early activation of the mechanism occurring during normal labour at term [1,7].
Inflammation may play an important role in both term and preterm parturition. During labour there is increased synthesis of prostaglandins (PGs) within the uterus via the induction of prostaglandin-endoperoxide synthase (PTGS), also known as cyclooxygenase (COX) [8]. COX is a rate limiting enzyme that is responsible for the conversion of arachidonic acid to prostaglandin endoperoxides, which are further converted into different prostaglandins (PGs) by cell specific terminal synthases [9][10][11]. Besides relaxation and contraction of the myometrium, PGs also stimulate cervical ripening. This remodelling process involves softening and dilatation of the cervical channel facilitating fetal delivery [12].
OXT is a nanopeptide produced by the hypothalamus and released from the posterior pituitary nerve terminals into the circulation in a pulsatile manner [17]. As labour progresses the frequency of OXT release increases [18,19]. In the myometrium, OXT binds to a specific G protein-coupled receptor (GPCR) linked through Gα q/11 to phospholipase Cβ (PLCβ). Activation of PLCβ results in the production of inositol 1, 4, 5-trisphosphate (IP 3 ) and diacylglycerol (DAG) [17,20]. IP 3 releases Ca 2+ from intracellular stores. In addition, OXT receptor (OXTR) activation increases intracellular Ca 2+ by stimulating extracellular Ca 2+ entry through cell membrane channels in human myometrial cells [21]. Increases in intracellular Ca 2+ are essential to stimulate myometrial contractility; moreover OXTR mediated Ca 2+ entry stimulates nuclear transcription in human myometrial cells through calcineurin/NFAT (nuclear factor of activated T cells) signalling [22].
DAG is a lipid second messenger that is capable of activating conventional and novel protein kinase C (PKC) isoforms [23]. The phorbol ester, PMA activates PKC by mimicking the function of DAG. Activation of conventional PKC isoforms can however only happen in the presence of both DAG and Ca 2+ [24]. PKC can stimulate different pathways e.g. mitogen-activated protein kinases (MAPK) such as extracellular signal-regulated kinase (ERK), and nuclear factor kappa-light-chain-enhancer of activated B cells (NFKB) both shown to be involved in labour [15,25]. Activation of these pathways can happen under control of PKC or independently of this kinase [26,27]. It has been reported that OXT induces myometrial COX-2 expression and stimulates PG synthesis through activation of ERK and that this process is entirely PKC independent [15,28]. On the other hand, PKC activity has been demonstrated to be important during OXT-mediated myometrial contractions [29]. Recently, we have shown that OXT induces COX-2 mRNA expression in myometrial cells through activation of the calcineurin/ NFAT pathway [22].
The growth factor EGF accumulates in the amniotic fluid throughout pregnancy and is present in maternal and fetal blood [30,31]. EGF exerts its function by binding to the EGF receptor (EGFR, the first described member of the proto-oncogene ErbB tyrosine kinase family). Binding of EGF to the EGFR stimulates its tyrosine kinase activity via an allosteric mechanism involving the formation of asymmetric intracellular kinase domains. The phosphorylated EGFR recruits adaptor proteins and activates several major signalling cascades including MAPK, PLCγ, PKC, PI3 kinase-AKT and NFKB [32]. EGF stimulates COX-2 expression in human amnion cells and prostaglandin production in cultured myometrial cells [33,34]. The mechanism of action of EGF has been investigated in many cell types, including amnion cells, and the consensus is that EGFR operates primarily through ERK activation to stimulate COX-2 gene expression, with little involvement of PKC [14,35]; however, the mediators involved in EGFR signalling in myometrial cells have not been described.
The pro-inflammatory cytokine IL1β is a potent stimulator of COX-2 expression in human myometrial cells [36]. During pregnancy the concentration of this cytokine increases in the amniotic fluid [37]. It has been reported that IL1β induces myometrial COX-2 expression via activating PKC and NFKB [27,36].
The aim of this study was to investigate the pathways mediating OXTR and EGFR-induced COX-2 expression in human myometrial cells, and in particular the role of the ERK and PKC pathways. We have also compared the effect of OXT and EGF with the pro-inflammatory cytokine IL1β under the same conditions. The results demonstrate that OXT, EGF and IL1β induced COX-2 expression in myometrial cells is mainly PKC dependent. Our findings contribute towards a better understanding of the mechanisms leading to spontaneous term and preterm labour.

Results
Oxytocin stimulates myometrial COX-2 expression via signalling through PKC and ERK Myometrial cells were stimulated with 1 μM OXT for 6 hours in the absence and in the presence of different inhibitors. Signalling through NFKB was inhibited by means of an IKK-2 inhibitor, TPCA-1 (2 μM). The ERK pathway was inhibited using the MEK1/2 inhibitor, PD-184352 (2 μM). Bisindolylmaleimide-I (BIS1; 5 μM) was used to inhibit PKC signalling. After stimulating the cells with OXT they were lysed and subjected to immunoblotting with a COX-2 antibody. Control wells were incubated with inhibitors in the absence of OXT. None of the inhibitors alone had any effect on COX-2 expression (data not shown). As shown in Figure 1, stimulation of myometrial cells with OXT resulted in a fourfold increase (P < 0.001) in COX-2 expression. This increase was significantly inhibited in the presence of PD-184352 (P <0.001) and BIS1 (P <0.001). When both inhibitors were used in combination there was a further decrease in OXT induced COX-2 expression to control levels, but this decrease did not reach statistical significance ( Figure 1).
To demonstrate that the ERK pathway was active in response to OXT, we measured ERK phosphorylation directly. Cells were stimulated with 1 μM OXT for different time points (2-, 5-, 10-, and 20-minutes) with and without pre-treatment with 5 μM BIS1. Lysates were used for immunoblotting with a pERK antibody. Blots were stripped and re-probed for ERK, to which the phosphorylation of ERK was normalised. As shown in Figure 2, ERK phosphorylation increased with OXT stimulation with a maximal response at 5 minutes (P < 0.001). The response to OXT was transient and at 20 minutes pERK levels were below control levels, possibly due to OXT receptor desensitization and phosphatase induction. The PKC inhibitor BIS1 provoked a significant decrease in ERK phosphorylation (P < 0.001).

EGF stimulates COX-2 expression via signalling through ERK and PKC
To investigate EGF induced COX-2 expression, cells were stimulated with 25 ng/ml EGF for 6 hours in the absence and in the presence of different inhibitors (2 μM TPCA-1, 2 μM PD-184352 and 5 μM BIS1). As shown in Figure 3, EGF provoked a significant increase in COX-2 expression (3.9 fold; P < 0.001). This increase was significantly inhibited in the presence of TPCA-1 (P < 0.01), PD-184352 (P < 0.001) and BIS1 (P < 0.001). The inhibitory Figure 1 Mechanism of OXT induced COX-2 expression in human myometrial cells. Cells were stimulated with 1 μM OXT for 6 hours in the absence and in the presence of different inhibitors. The NFKB pathway was inhibited using TPCA-1 (2 μM). PD-184352 (2 μM) was used to test the ERK pathway. Signalling through PKC was inhibited by means of Bisindolylmaleimide I (BIS1; 5 μM). (A) Cell lysates were subjected to immunoblotting for COX-2. Blots were stripped and re-probed for GAPDH. (B) COX-2 expression was normalised to GAPDH expression and given as a percentage of the OXT treated value. Data points represent mean ± SEM, n = 5. Statistical significance was determined using one way ANOVA and Dunnett's post hoc test. ***P < 0.001. To study the effect of EGF on ERK activation, cells were stimulated with 25 ng/ml EGF in the absence or presence of 5 μM BIS1 for different time points (2-,5-,10-, and 20minutes). As depicted in Figure 4, stimulation of myometrial cells with EGF resulted in a marked increase in ERK phosphorylation with a maximal response after a stimulation of 20 min. The PKC inhibitor BIS1 was unable to block this effect ( Figure 4) demonstrating its specificity towards PKC in our system.

IL1β increases COX-2 expression via activating NFKB and PKC
The pro-inflammatory cytokine IL1β is a potent activator of COX-2 expression. Cells were stimulated for 6 hours with 5 ng/ml IL1β alone or in combination with different inhibitors (TPCA-1 (2 μM), PD-184352 (2 μM) and BIS1 (5 μM)). Cells were lysed and subjected to immunoblotting for COX-2. As expected, IL1β stimulated COX-2 expression in myometrial cells very strongly (8.5 fold; P < 0.001). This effect was significantly inhibited by TPCA-1 (P < 0.01) and BIS1 (P < 0.001) indicating that both NFKB and PKC are involved in IL1β induced COX-2 expression ( Figure 5). The ERK pathway on the other hand did not seem to contribute to this effect.  In summary, these results demonstrate that PKC signalling plays an important role in agonist-induced COX-2 expression. PMA activates PKC directly and is a potent activator of COX-2 expression in myometrial cells [27]. We used PMA as a tool to further investigate the involvement of PKC in agonist-induced COX-2 expression in our system.
To test for ERK phosphorylation, myometrial cells were stimulated with 2 μM PMA for different time points (2-, 5-, 10-, and 20-minutes). Stimulation of myometrial cells with PMA resulted in an increase in ERK phosphorylation with a maximal response after 10 min of stimulation (P < 0.001; Figure 7). Two clear phosphoERK1 and phosphoERK2 bands were detected by the pERK antibody, with preferential phosphorylation of ERK2. The PKC inhibitor BIS1 inhibited this effect (P < 0.001) and also decreased basal levels of ERK phosphorylation.

Discussion
This study provides the first direct comparison of the effect of OXT and EGF in human myometrial cells and demonstrated that both agonists require PKC activity to stimulate COX-2 expression. Given the importance of COX-2 up-regulation for the onset of labour and the involvement of GPCR and tyrosine kinase receptors in the control of uterine activation [7], we decided to investigate the mechanism by which OXT and EGF increase COX-2 expression in myometrial cells, in particular the role of PKC which has been questioned in previous studies [15].
OXT plays an important role during parturition by stimulating myometrial contractility and promoting the expression of labour-associated genes [22]. Moreover OXT has paracrine interactions with IL1β in human amnion to increase COX-2 expression and stimulate PG production [38]. Increased PG release leads to further paracrine and autocrine effects facilitating uterine activation, for instance PGF2α stimulates OXTR and COX-2 expression in human myometrial cells [39]. Molnar et al. [15] reported that OXT induces COX-2 expression in myometrial cells by activating the ERK pathway and suggested that this is entirely PKC independent. We have confirmed that OXT increases myometrial COX-2 expression via ERK stimulation. However, in contrast to Molnar et al. our study reveals that the stimulatory effect of OXT on COX-2 expression is dependent on PKC activation as shown by the blocking effect of BIS1, a PKC inhibitor. BIS1 had an inhibitory effect on basal levels of ERK phosphorylation at the concentration used, however this is likely to be due to inhibition of constitutive PKC activity in myometrial cells. BIS1 had no effect on EGF-stimulated ERK phosphorylation; hence its inhibitory effect on OXTinduced COX-2 expression is likely to be due to its PKC blocking effect and not due to non-specific inhibition of ERK. It is possible that during pregnancy the myometrium becomes more responsive to PKC signalling [40] and this is reflected in our results. MEK1/2 and PKC inhibitors were capable of blocking OXT-induced COX-2 expression almost completely when given separately. When they were used in combination, a small additional inhibitory effect was visible. This indicates that both ERK and PKC signalling mediate OXT stimulated COX-2 expression. Moreover, OXT increased ERK phosphorylation and this was significantly inhibited in the presence of BIS1.
The concentration of EGF increases in the amniotic fluid during pregnancy [31]. EGF exerts its function by binding to EGF-receptor dimers and promoting tyrosine phosphorylation of other proteins. It has been shown that EGF activates PKC in human myometrial cells and that it can induce PG production in these cells [33,41]. In human amnion EGF induces COX-2 expression [34,35]. The involvement of EGF in COX-2 expression in human myometrial cells has not been studied before. Our results show that EGF stimulates COX-2 expression in myometrial cells and that this is dependent on the activation of PKC, ERK and NFKB. The inhibitory effect of PD-184352 was the strongest, followed by that of BIS1, with TPCA-1 having only a moderate effect. The addition of BIS1 to PD-184352 produced an inhibitory effect similar to that with PD-184352 alone. This suggests that the ERK pathway is the most important, but there is also a strong PKC component. There is interaction between the two pathways in EGFR and OXT receptor mediated COX-stimulation in  bovine endometrial cells [42]. Nevertheless in our human myometrial cells EGF stimulated ERK phosphorylation largely in a PKC independent manner. The role of EGF in stimulating COX-2 expression is worth investigating further, in particular given the capacity of OXT receptors to trans-activate EGFR tyrosine kinase leading to ERK1/2 activation in myometrial cells [43].
IL1β is a potent stimulator of COX-2 expression and exposure of myometrial cells to this agonist resulted in a 9-fold increase in COX-2 expression. It has been shown that IL1β induces COX-2 expression in human myometrial cells by activating the PKC and NFKB pathways [27,36] and we have confirmed this here. Moreover, we have shown that compared to OXT, IL1β is relatively more dependent on the NFKB pathway to stimulate COX-2 expression; however, both agonist rely mainly on PKC signalling.
A potential limitation of this study is that we did not have a positive control for the inhibitor in each individual experiment; however controls are presented throughout the work: for the IKK-2 inhibitor TPCA-1 positive control in Figure 5 and negative control in Figure 6; for the MEK1/2 inhibitor PD-184352 positive control in Figure 3 and negative control in Figure 5; for the PKC inhibitor BIS1 positive control in Figure 6 and negative control in Figure 4. This confirms the consistency of the data.
TPCA-1 is a selective IKK-2 inhibitor that has been extensively characterised in human intrauterine tissues; it inhibits p65 translocation and reduces the expression of genes associated with inflammatory signalling via the NFKB pathway [44]. However IKK-2 also phosphorylates other targets in addition to those involved in NFKB activation, such as beta-catenin and BCL10 which play important roles in cell division and apoptosis. Silva et al. (2010) [44] reported no side effects of TPCA-1 in choriodecidual cells and we did not observe signs of cytotoxicity in our myometrial cell cultures at the doses of TPCA used. PD 184352 is a highly potent non-competitive inhibitor of MEK1 and MEK2; it suppresses the activation of MEK1, rather than blocking its activity. The selectivity of PD 184352 for MEK against a wide range of kinases is excellent [45]. BIS1 is a PKC inhibitor that shows high selectivity for PKC α-, β1-, β2-, γ-, δ-, and ε-isozymes in uterine tissues [46]. However, as with other bisindolylmaleimides, the possibility that it also inhibits protein kinases of the RSK and GSK3 families cannot be ruled out [45].
Overall, our results show that PKC plays a central role in agonist-induced COX-2 expression in myometrial cells. When PKC was stimulated directly using PMA there was a large increase in COX-2 expression in the cells. In the presence of PD-184352 a significant decrease in COX-2 expression was observed. BIS1 on the other hand completely inhibited the effect of PMA on COX-2 expression, which shows that PKC can stimulate COX-2 expression independently of ERK. Little or no additive inhibitory effect was observed when the MEK1/2 and PKC inhibitors were used in combination, indicating that the PKC-effect is not entirely ERK mediated.

Conclusion
In conclusion, there is strong evidence for the involvement of PKC in agonist-induced COX-2 expression in human myometrial cells (Figure 8). This is the most important pathway for both physiological and inflammatory agonists. The effect of PKC appears to lie upstream of ERK but we cannot rule out that the two kinase pathways act independently of each other to convert receptor signalling into COX-2 stimulation. The relative prominence of each pathway in myometrial cells depends on the type of agonist used. These findings contribute to our understanding of myometrial preparation for parturition and respond to the need to investigate myometrial specific mechanisms in order to develop tocolytic drugs with better uterine selectivity. The data will inform further research not only into the prevention of preterm labour, but also into the induction or augmentation of labour where increased expression of COX-2 is important.

Tissue collection
This study was approved by the South West Research Ethics Committee and specimens were collected after obtaining informed written consent. Human myometrial smooth muscle samples were obtained from pregnant women at term (not in labour) undergoing caesarean section at St Michael's Hospital (Bristol, UK). Caesarean section was performed following maternal request, fetal malposition or previous section. Women with signs of infection were excluded. Myometrial samples were taken from the upper margin of the incision and washed in ice-cold isotonic saline before being transported to the laboratory.
For some experiments, cells were pre-incubated with BIS1 for 30 minutes followed by stimulation with agonists for different time points (2-, 5-, 10-, and 20-minutes). Before using for immunoblotting, cells were rinsed in PBS and lysed in boiling SDS-PAGE sample buffer. Cells that were not pre-incubated with inhibitors received 0.1% (v/v) dimethyl sulphoxide (DMSO; vehicle for BIS1) in the presence or absence of agonist.

Statistical analysis
Statistical analysis was performed by means of GraphPad Prism (GraphPad Software, La Jolla, CA, USA) using one-way ANOVA followed by the Dunnett post hoc test and two-way ANOVA with a Bonferroni post hoc test. Values are shown as the mean ± SEM of 3 to 5 replicate experiments using tissue from different donors. Differences were considered statistically significant at P < 0.05. *P < 0.05, **P < 0.01 and ***P < 0.001.