Research ArticleReproductive Biology

Phosphorylation of STAT3 mediates the induction of cyclooxygenase-2 by cortisol in the human amnion at parturition

See allHide authors and affiliations

Science Signaling  27 Oct 2015:
Vol. 8, Issue 400, pp. ra106
DOI: 10.1126/scisignal.aac6151

Labor, STAT!

The glucocorticoid cortisol has anti-inflammatory effects in many tissues of the body. However, in fetal membranes during labor, cortisol is associated with increased activity of the inflammatory enzyme COX-2, which produces prostaglandins that promote reproductive smooth muscle changes in the mother that enable delivery. Using primary human amnion fibroblasts from full-term births from patients who delivered after active labor or through nonlabor cesarean sections, Wang et al. found that the interaction of the glucocorticoid receptor and the transcription factor STAT3 at the gene encoding COX-2 mediates an autocrine feed-forward loop that amplifies cortisol-triggered prostaglandin production. The findings identify a mechanism linking cortisol to inflammatory prostaglandin production in the amnion.

Abstract

The induction of cyclooxygenase-2 (COX-2) and subsequent production of prostaglandin E2 (PGE2) by cortisol in the amnion contrast with the effect of cortisol on most other tissues, but this proinflammatory effect of cortisol may be a key event in human parturition (labor). We evaluated the underlying mechanism activated by cortisol in primary human amnion fibroblasts. Exposure of the amnion fibroblasts to cortisol led to the activation of the cyclic adenosine monophosphate (cAMP)–protein kinase A (PKA) pathway, which induced the phosphorylation of the kinase SRC and STAT3 (signal transducer and activator of transcription 3). STAT3 interacted with the glucocorticoid receptor (GR) and the transcription factor CREB-1 (cAMP response element–binding protein 1) at the promoter of the gene encoding COX-2, which promoted the production of the secreted prostaglandin PGE2. PGE2 activates the prostaglandin receptors EP2 and EP4, which stimulate cAMP-PKA signaling. Thus, cortisol reinforced the activation of cAMP-PKA signaling through an SRC–STAT3–COX-2–PGE2–mediated feedback loop. Inhibiting STAT3, SRC, or the cAMP-PKA pathway attenuated the cortisol-stimulated induction of COX-2 and PGE2 production in amnion fibroblasts. In human amnion tissue, the amount of phosphorylated STAT3 correlated positively with that of cortisol, COX-2, and PGE2, and all were more abundant in tissue obtained after active labor than in tissue obtained from cesarean surgeries in the absence of labor. These results indicated that the coordinated recruitment of STAT3, CREB-1, and GR to the promoter of the gene encoding COX-2 contributes to the feed-forward induction of COX-2 activity and prostaglandin synthesis in the amnion during parturition.

INTRODUCTION

Prostaglandins (PGs), in particularly PGE2 and PGF, play a crucial role in human parturition by stimulating myometrium contraction, inducing cervical ripening, and causing rupture of the fetal membranes (1, 2). The fetal membranes have been recognized as the major source of PGs toward the end of pregnancy: the choriodecidua is a major source of PGF, and the amnion is a major source of PGE2 (3, 4). A large body of evidence strongly suggests that the activation of PG synthesis in the fetal membranes is one of the ultimate pathways leading to parturition at both term and preterm birth in humans (1, 5). The increase in PG synthesis in the fetal membranes is associated with the induction of cyclooxygenase-2 (COX-2), the inducible enzyme that catalyzes the rate-limiting step in the formation of PGs by converting arachidonic acid to PGH2, the precursor for the two-series thromboxanes and prostaglandins (4).

Although both amnion epithelial cells and fibroblasts produce PGE2, amnion fibroblasts are believed to contribute at least five times more PGE2 than the epithelial cells at term (6). We and others have demonstrated that glucocorticoids induce the expression of PTGS2 encoding COX-2, thereby stimulating PG synthesis in human amnion fibroblasts (68). This is in marked contrast to the reduction of COX-2 and PG synthesis by glucocorticoids in most other tissues in the body (9). We have further demonstrated that this induction of COX-2 by glucocorticoids requires the interaction of phosphorylated cyclic adenosine monophosphate (cAMP) response element (CRE)–binding protein 1 (CREB-1) and glucocorticoid receptor (GR) as well as the binding of CREB-1 to the CRE in the promoter of PTGS2 (10, 11).

In addition to prostaglandin production, human fetal membranes have abundant 11β-hydroxysteroid dehydrogenase 1 (11β-HSD1), which generates cortisol (also called Kendall’s compound F) from its biologically inactive counterpart cortisone (also called Kendall’s compound E) (12). We previously showed that cortisol, in turn, induces the expression of HSD11B1, the gene encoding 11β-HSD1, in human amnion fibroblasts and chorion trophoblasts (13, 14), suggesting that a local feed-forward generation of cortisol exists in the fetal membranes toward the end of pregnancy (1315). Given this generation of cortisol, the induction of COX-2 by cortisol in human amnion fibroblasts is possibly one of the key components of the feed-forward mechanisms of parturition at both term and preterm birth in humans.

Signal transducer and activator of transcription 3 (STAT3) is a transcription factor that is activated upon phosphorylation at Tyr705 (16). STAT3 classically participates in the signaling events elicited by cytokines or growth factors (17, 18). Once phosphorylated, STAT3 forms a homodimer and translocates to the nucleus, thereby stimulating target gene transcription (16, 19). Sequence analysis of the 5′ flanking region of the PTGS2 gene reveals a potential STAT3 binding site adjacent to the previously identified CRE (10). STAT3 directly binds the PTGS2 promoter to promote its expression in several cell types, including glioblastoma and gastric epithelial cells (20, 21). Moreover, STAT3 also interacts with coactivators, including GR and CREB, to regulate gene expression in other cell types (2226). Together, these findings prompted us to investigate whether STAT3 was involved in the regulation of PTGS2 expression by cortisol in cultured primary human amnion fibroblasts and in the amnion tissue.

RESULTS

Involvement of STAT3 in the induction of COX-2 by cortisol in human amnion fibroblasts

Treating cultured amnion fibroblasts with cortisol for 24 hours increased COX-2 and 11β-HSD1 mRNA and protein abundance in a concentration-dependent manner, with a significant induction observed in response to ≥0.1 μM cortisol (Fig. 1, A and B). The increase in COX-2 mRNA and protein abundance as well as PGE2 abundance by cortisol was significantly attenuated by either inhibiting STAT3 with the antagonist S3I-201 (Fig. 1, C and D) or knocking down STAT3 using small interfering RNA (siRNA) (Fig. 1, E and F). Time course analysis revealed that up to 12 hours of cortisol treatment increased the phosphorylation of STAT3 at Tyr705, with the maximal effect observed at 6 hours (Fig. 2A). The phosphorylation of STAT3 at Tyr705 also displayed a concentration-dependent response to cortisol treatment (Fig. 2A). Immunofluorescence staining (Fig. 2B) and Western blotting analysis (Fig. 2C) revealed that STAT3 translocated from the cytoplasm to the nucleus after cortisol stimulation. Coimmunoprecipitation assays showed that Tyr705-phosphorylated STAT3 was identified in the nuclear protein complex precipitated by either phosphorylated CREB-1 antibody or GR antibody after cortisol treatment (Fig. 2D). Furthermore, by chromatin immunoprecipitation (ChIP) assays, we found that cortisol increased the binding of STAT3, CREB-1, and RNA polymerase II (Pol II) to the PTGS2 promoter region that contains CREB and STAT3 binding sites (Fig. 2, E and F). These data suggest that STAT3 is involved in the induction of COX-2 by cortisol through its interaction with CREB-1 and GR in the nucleus of human amnion fibroblasts.

Fig. 1 STAT3 mediates cortisol-induced COX-2 abundance and PGE2 production in human amnion fibroblasts.

(A and B) COX-2 and 11β-HSD1 mRNA (A) and protein (B) abundance in human amnion fibroblasts in response to increasing doses of cortisol treatment for 24 hours. (C and D) Abundance of COX-2 mRNA (C) and protein and PGE2 (D) in human amnion fibroblasts in response to cortisol (F, 1 μM) treatment for 24 hours in the presence or absence of the STAT3 antagonist S3I-201 (10 μM). (E and F) Amounts of COX-2 mRNA (E) and protein and PGE2 (F) in human amnion fibroblasts in response to cortisol (1 μM) treatment for 24 hours in the presence or absence of siRNA-mediated knockdown of STAT3. Ctr, control; NC, negative control with scrambled siRNA. *P < 0.05, **P < 0.01, ***P < 0.001 against control cells; ##P < 0.01, ###P < 0.001 against cortisol-treated cells (C and D) or cortisol-treated, control-transfected cells (E and F). Blots are representative, and data are means ± SEM from three to six experiments.

Fig. 2 STAT3 is activated by cortisol in human amnion fibroblasts.

(A) Time- and concentration-dependent effect of cortisol on the phosphorylation of STAT3 at Tyr705 in human amnion fibroblasts. (B and C) Immunofluorescence staining (B) and Western blotting (C) to detect intracellular localization of STAT3 in human amnion fibroblasts in response to cortisol (1 μM, 6 hours). Scale bars, 50 μm. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (D) Coimmunoprecipitation assay in amnion fibroblasts from two patients to detect Tyr705-phosphorylated STAT3 with phosphorylated CREB-1 or GR in response to cortisol (1 μM, 12 hours). Immunoglobulin G (IgG) served as negative control. Ab, antibody. (E) Sequence of the PTGS2 promoter spanning −249 to +47 base pairs (bp). Arrows indicate the primer aligning positions in ChIP assay. Boxed bold letters indicate putative transcription factor binding sites, as labeled. TSS, transcription start site. (F) ChIP assay for the enrichment of STAT3, CREB-1, and Pol II at the PTGS2 promoter in human amnion fibroblast in response to cortisol (1 μM, 24 hours). IgG served as negative control. *P < 0.05 against control cells. Blots are representative, and data are means ± SEM from three experiments.

Role of cAMP-PKA-SRC pathway activation in the phosphorylation of STAT3 at Tyr705 in human amnion fibroblasts exposed to cortisol

Treating cultured amnion fibroblasts with cortisol significantly increased the intracellular cAMP level by 3 and 6 hours but not at 1 hour (Fig. 3A). The adenylate cyclase inhibitor SQ22536 and the protein kinase A (PKA) inhibitor H89 prevented cortisol-induced phosphorylation of STAT3 at Tyr705 (Fig. 3B), COX-2 abundance, and PGE2 production (Fig. 3C). In contrast, treating amnion fibroblasts with the adenylyl cyclase stimulator forskolin (FSK) or the cAMP analog dibutyl cAMP (db-cAMP) significantly increased the phosphorylation of STAT3 at Tyr705 (Fig. 3D).

Fig. 3 The cAMP-PKA pathway induces STAT3 phosphorylation, COX-2 abundance, and PGE2 production by cortisol in human amnion fibroblasts.

(A) Amount of intracellular cAMP in human amnion fibroblasts in response to cortisol treatment. (B) Amount of phosphorylated STAT3 in human amnion fibroblasts in response to cortisol (1 μM, 6 hours) in the presence or absence of the PKA inhibitor H89 (10 μM) or the adenylate cyclase inhibitor SQ22536 (SQ, 10 μM). (C) COX-2 abundance by Western blotting and PGE2 production by enzyme immunoassay (EIA) in human amnion fibroblasts in response to cortisol (1 μM, 6 hours) in the presence or absence of H89 (10 μM) or SQ22536 (10 μM). (D) Amount of phosphorylated STAT3 in human amnion fibroblasts in response to the adenylyl cyclase stimulator FSK (100 μM, 2 hours) or the cAMP analog db-cAMP (100 μM, 2 hours). *P < 0.05, **P < 0.01, ***P < 0.001 against control cells; #P < 0.05, ##P < 0.01, ###P < 0.001 against cortisol-treated cells. Blots are representative, and data are means ± SEM from three to four experiments.

Concurrently, treating amnion fibroblasts with cortisol increased the phosphorylation of SRC at Tyr416 and Ser17 (Fig. 4A), an effect blocked by SQ22536 or H89 (Fig. 4B). Likewise, treating amnion fibroblasts with FSK or db-cAMP increased the phosphorylation of SRC at Tyr416 and Ser17 (Fig. 4C). Furthermore, the SRC inhibitor AZD0530 completely blocked the induction of phosphorylation of STAT3 at Tyr705, COX-2 abundance, and PGE2 production by either cortisol (Fig. 4, D and E) or activators of the cAMP-PKA pathway (Fig. 4, F and G). These results suggest that cortisol-induced phosphorylation of STAT3 and subsequent induction of COX-2 abundance and PGE2 secretion are mediated by the activation of the cAMP-PKA-SRC pathway in human amnion fibroblasts.

Fig. 4 SRC induces STAT3 phosphorylation, COX-2 abundance, and PGE2 production by cortisol and cAMP-PKA pathway in human amnion fibroblasts.

(A) Time course of the effect of cortisol on the phosphorylation of SRC in human amnion fibroblasts. (B) Phosphorylation of SRC in human amnion fibroblasts in response to cortisol (1 μM, 6 hours) in the presence or absence of H89 (10 μM) or SQ22536 (10 μM). (C) Phosphorylation of SRC in human amnion fibroblasts in response to FSK (100μM, 2 hours) or db-cAMP (100 μM, 2 hours). (D and E) Abundance of phosphorylated STAT3 (D) and COX-2 and PGE2 (E) in human amnion fibroblasts in response to cortisol (1 μM, 6 hours) in the presence or absence of the SRC inhibitor AZD0530 (20 μM). (F and G) Abundance of phosphorylated STAT3 (F) and COX-2 and PGE2 (G) in human amnion fibroblasts in response to FSK or db-cAMP [as in (C)] in the presence or absence of AZD0530 (20 μM). *P < 0.05, **P < 0.01, ***P < 0.001 against control cells; #P < 0.05, ##P < 0.01, ###P < 0.001 against cortisol-treated cells (D and E) or FSK- or db-cAMP–treated cells (F and G). Blots are representative, and data are means ± SEM from three to five experiments.

Role of PGE2 in the activation of cAMP-PKA pathway and the phosphorylation of STAT3 at Tyr705 in human amnion fibroblasts exposed to cortisol

In addition to the long-term effect on COX-2 abundance as described above, cortisol was also able to induce COX-2 abundance and PGE2 secretion in the relative short term. Treating amnion fibroblasts with cortisol significantly increased COX-2 abundance by 1 hour (Fig. 5A) and PGE2 secretion by 3 hours (Fig. 5B). Exposing cells to a combination of antagonist to EP2 and EP4 receptors (PF-04418948 and L-161982, respectively), the two PGE2 receptor subtypes that are coupled with the activation of the cAMP-PKA pathway (27, 28), significantly reduced basal and cortisol-induced intracellular cAMP abundance (Fig. 5C) and phosphorylation of STAT3 at Tyr705 (Fig. 5D). Consistently, treating amnion fibroblasts with PGE2 significantly increased the phosphorylation of STAT3 at Tyr705 (Fig. 5E). These results suggest that activation of the cAMP-PKA pathway and induction of STAT3 phosphorylation at Tyr705 by cortisol are mediated by increased PGE2 activation of EP2 and EP4 receptors in human amnion fibroblasts.

Fig. 5 PGE2 stimulates the phosphorylation of STAT3 in response to cortisol in human amnion fibroblasts.

(A and B) Time course of the effect of cortisol on COX-2 abundance (A) and PGE2 secretion (B) from human amnion fibroblasts. (C and D) Amount of intracellular cAMP (C) and phosphorylated STAT3 (D) in human amnion fibroblasts in response to cortisol (1 μM, 6 hours) in the presence or absence of EP2 and EP4 receptor antagonists PF-04418948 (PF, 10 μM) and L-161982 (L161, 10 μM), respectively. (E) Time course of the effect of PGE2 (1 μM) on the phosphorylation of STAT3 at Tyr705. *P < 0.05, **P < 0.01 against control cells; ###P < 0.001 against cortisol-treated cells. Blots are representative, and data are means ± SEM from three to four experiments.

Correlation among Tyr705-phosphorylated STAT3, COX-2, PGE2, and cortisol and their association to parturition in human amnion tissue

Correlation analysis of Tyr705-phosphorylated STAT3, COX-2, PGE2, and cortisol abundance (Fig. 6) in the amnion tissue collected both at term after spontaneous labor and at elective cesarean section without labor showed a moderate but significantly positive correlation between the amount of Tyr705-phosphorylated STAT3 and that of COX-2 (R = 0.588) or PGE2 (R = 0.6316) in the amnion tissue. Furthermore, the amount of cortisol in the amnion tissue also correlated with the amounts of Tyr705-phosphorylated STAT3 (R = 0.8333), COX-2 (R = 0.6544), and PGE2 (R = 0.693). Moreover, the amounts of Tyr705-phosphorylated STAT3 (but not total STAT3), COX-2 (Fig. 6, A and B), cortisol, and PGE2 (Fig. 6C) were all significantly increased in the amnion tissue collected at term after spontaneous labor as compared with the tissue collected at elective cesarean section without labor. These data suggest that the increases in Tyr705-phosphorylated STAT3, COX-2, cortisol, and PGE2 in the amnion tissue are associated with labor, and support the notion that the phosphorylation of STAT3 at Tyr705 is promoted by cortisol through the induction of COX-2 and PGE2 production.

Fig. 6 Amounts of COX-2, phosphorylated STAT3, cortisol, and PGE2 in human amnion tissue.

(A and B) Western blotting (A) and quantification (B) of the abundance of COX-2, phosphorylated STAT3, and total STAT3 in human amnion tissue collected at term after spontaneous labor [term labor (TL)] and at term by cesarean section without labor [term nonlabor (TNL)]. (C) Amounts of cortisol and PGE2 in human amnion tissues collected from TL and TNL assessed by EIAs. *P < 0.05, **P < 0.01 against TNL. Data are means ± SEM from 10 (TL) or 9 (TNL) patients.

DISCUSSION

It is well known that the amount of cortisol in the maternal circulation rises gradually toward the end of gestation (29). However, the increase in biologically active free cortisol in the maternal circulation is limited because of the increase in corticoid-binding globulins with advancing gestation (30). Therefore, cortisol generated locally in the intrauterine tissues might play a more effective and crucial role in the onset of parturition (31). Earlier studies have found that the ratio of cortisol to cortisone in human amniotic fluid increases with gestation and was considerably higher in the fetal membranes than in other tissues of the body at the end of gestation (32, 33), suggesting that the fetal membranes are an essential extra-adrenal source of cortisol during gestation. Consistently, it has been shown that both 11β-HSD1 abundance and reductive activity increase with gestational age and labor (34, 35). The current study reinforces previous evidence (13, 14) that cortisol can reinforce the expression of HSD11B1, which indicates that the local concentration of cortisol generated by 11β-HSD1 in the fetal membranes can be extremely high toward the end of gestation. Cortisol levels in maternal circulation increase with gestational age and further increase with labor (29, 33, 36). Here, we found that the concentration of cortisol in the amnion tissue at labor increased to about 8.6 times of that in the maternal blood. We and others have demonstrated that one role of this peculiar feature of cortisol generation of the fetal membranes may, at least, be associated with parturition because cortisol within the range of concentrations detected in the amnion can induce COX-2 and subsequent PGE2 production, a widely recognized mediator in parturition, in amnion fibroblasts (68). We demonstrated in this study that cortisol could activate the cAMP-PKA-SRC-STAT3 pathway through increased binding of PGE2 to EP2 and EP4 receptors in human amnion fibroblast. Phosphorylated STAT3 (Tyr705) derived from this activated pathway then entered the nucleus and interacted with CREB-1 and GR at the PTGS2 promoter to further induce its expression. The involvement of STAT3 in the induction of COX-2 by cortisol was confirmed by the attenuation of cortisol-induced COX-2 mRNA and protein abundance by a STAT3 antagonist as well as the siRNA-mediated knockdown of STAT3, and was further illustrated by the positive correlation between the amount of Tyr705-phosphorylated STAT3 and the amounts of COX-2, PGE2, and cortisol in the amnion tissue. Because the concentrations of cortisol used in this study are within the concentration range that is observed toward the end of pregnancy in fetal membranes, our findings may be physiologically relevant to parturition.

Glucocorticoids, such as cortisol, are known to operate through the intracellular GR to execute their transcriptional activities (9). However, this study demonstrated that cortisol also increased the level of intracellular cAMP and the phosphorylation of SRC and STAT3 in human amnion fibroblasts. Accumulating evidence suggests that glucocorticoids can exert rapid effects through one or more membrane-associated GRs coupled to the downstream G protein–dependent signaling cascades, thereby increasing the amount of cAMP in several cell types (37). This is unlikely in this study because we observed that increases in cAMP and phosphorylated SRC by cortisol were not detected at 1 hour after cortisol treatment in human amnion fibroblasts. Moreover, previous studies have demonstrated that the increases in COX-2 abundance and phosphorylation of CREB-1 by cortisol can be blocked by RU486 (10), an antagonist to the cytosolic GR (38, 39). All these findings indicate that the activation of the cAMP-PKA pathway by cortisol is very likely an indirect effect subsequent to the transcriptional activity of GR.

Because the increases in intracellular cAMP and Tyr705-phosphorylated STAT3 after exposure to cortisol were blocked by treating amnion fibroblasts with a combination of EP2 and EP4 receptor antagonists, the two Gs protein–coupled PGE2 receptor subtypes that use cAMP as the second messenger (27, 28), we believe that activation of the cAMP-PKA-SRC-STAT3 pathway by cortisol might be the consequence of the early onset of induction of COX-2 and subsequent PGE2 production by cortisol. This notion is in accord with previous studies showing that PGE2 induces PTGS2 expression through the activation of the cAMP-PKA pathway (40, 41) and is further supported by our observation that cortisol increased COX-2 abundance and PGE2 secretion in human amnion fibroblasts with a relatively short incubation time. The initial induction of COX-2 by cortisol may be a joint effect of activated cAMP-PKA-SRC-STAT3 pathway maintained by basal PGE2 secretion and activated GR upon cortisol stimulation, which may be enhanced when PGE2 production is increased (Fig.7). The phosphorylation of STAT3 and SRC in the absence of cortisol treatment seen in this study also supports that there is tonic stimulation of this signaling pathway driven by basal PGE2 secretion. Although the expression of PTGS2 is relatively low with basal PGE2, the transcription of PTGS2 can be amplified in the presence of GR activation by cortisol, resulting in increased PGE2 production, which further intensifies the activation of this signaling pathway (Fig.7).

Fig. 7 Signaling pathways underpinning the induction of COX-2 by cortisol through PGE2 synthesis in human amnion fibroblasts.

By binding to EP2 and EP4 receptors, PGE2 activates the cAMP-PKA pathway, which phosphorylates SRC and CREB and, in turn, SRC phosphorylates STAT3. Phosphorylated CREB and STAT3 interact at PTGS2 promoter, thereby maintaining basal PTGS2 transcription and PGE2 production. Feed-forward generation of cortisol by 11β-HSD1 results in GR activation, which amplifies PTGS2 transcription by PGE2 by interacting with phosphorylated CREB and STAT3, in turn further enhancing PGE2 synthesis. Gs, stimulatory G protein; AC, adenylate cyclase; ATP, adenosine 5′-triphosphate; E, cortisone; F, cortisol.

STAT3 is activated by phosphorylation at Tyr705, which is required for STAT3 dimerization, nuclear translocation, and DNA binding (42). This study demonstrated that phosphorylation of STAT3 at Tyr705 was increased by cortisol and activation of the cAMP-PKA pathway. Therefore, there must be other tyrosine kinases activated by PKA to phosphorylate STAT3 because PKA is a well-known serine/threonine kinase. Although phosphorylation of STAT3 at Tyr705 is typically activated by cytokines or growth factor receptors through Janus kinases (JAKs) (16), SRC, a nonreceptor tyrosine kinase, is also implicated in the phosphorylation of STAT3 (4346). PKA increases the phosphorylation of SRC at Ser17 and subsequent autophosphorylation at Tyr416, which activates SRC activity (4749). Our study showed that both cortisol and the activation of the cAMP-PKA pathway induced the phosphorylation of SRC at Ser17 and Tyr416 in human amnion fibroblasts. Furthermore, the phosphorylation of STAT3 at Tyr705 by cortisol-induced activation of the cAMP-PKA pathway was suppressed in the presence of the SRC inhibitor AZD0530. Thus, the phosphorylation of STAT3 at Tyr705 by the activation of cAMP-PKA pathway appears to be mediated by the activation of SRC. Our findings indicate a complex signaling network involved in the paradoxical induction of COX-2 by cortisol in human amnion fibroblasts. Nevertheless, our in vitro findings in cultured human amnion fibroblasts with in vivo observations in human amnion tissue strongly support a role of STAT3 in the regulation of PTGS2 expression toward the end of gestation, thereby participating in parturition. However, the moderate correlation between phosphorylated STAT3 and COX-2 or PGE2 abundance in the amnion tissue also suggests that in addition to the cortisol-activated pathway, there may also be other pathways operated by nonglucocorticoid factors that regulate PTGS2 expression in the amnion at parturition.

In conclusion, we have demonstrated in this study that phosphorylation of STAT3 is required for the induction of COX-2 by cortisol in human amnion fibroblasts and that the phosphorylation of STAT3 at Tyr705 by cortisol is mediated by the activation of the cAMP-PKA-SRC pathway through induction of PGE2 production and PGE2 binding to EP2 and EP4 receptors in human amnion fibroblasts (Fig. 7). We believe that there exists a feed-forward loop of COX-2 induction and PGE2 production under the control of cortisol generation in human amnion fibroblasts toward the end of pregnancy, which may comprise a crucial component of the feed-forward mechanisms of parturition.

MATERIALS AND METHODS

Collection of human amnion and preparation of amnion fibroblast cells

Human fetal membranes were collected from full-term deliveries after spontaneous labor (TL) and at elective cesarean section without labor (TNL) under a protocol approved by the ethics committee of Renji Hospital, School of Medicine, Shanghai Jiao Tong University. Tissues from women that had complications such as preeclampsia, fetal growth restriction, and gestational diabetes were excluded from the study. To examine the involvement of STAT3 in the regulation of COX-2 and PGE2 abundance by cortisol, human amnion fibroblast cells were isolated only from TNL amnion. Briefly, after peeling from the chorion, the amnion tissue was digested with 0.125% trypsin (Sigma) and washed thoroughly with normal saline to remove epithelial cells. The remaining amnion tissue was digested with 0.1% collagenase (Roche), and the dispersed cells were purified with Percoll (GE Healthcare) gradients (5, 20, 40, and 60%). The cells were cultured at 37°C in 5% CO2–95% air in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% newborn calf serum (NCS) and antibiotics (all from Life Technologies Inc.). The identity of cells has been previously verified, and more than 95% of the cells are fibroblasts (6). To study the correlation between the amounts of STAT3, COX-2, PGE2, or cortisol and the correlation between labor and STAT3, COX-2, PGE2, or cortisol, amnion was separated from the fetal membranes collected from both TL and TNL pregnant women and frozen at −80°C for later protein and hormone extraction as described below.

Treatment of the amnion fibroblasts

Three days after plating, the cells were cultured in phenol red–and serum-free medium. The effect of cortisol on COX-2 and 11β-HSD1 abundance was studied by incubating the cells with cortisol (Sigma) at concentrations ranging from 0.01 to 1 μM for 24 hours. To study the role of STAT3 in the regulation of COX-2 and PGE2 by cortisol, the cells were treated with cortisol (1 μM) in the presence or absence of STAT3 antagonist S3I-201 (10 μM; Santa Cruz Biotechnology) or with or without siRNA-mediated knockdown of STAT3 for 24 hours. The method of siRNA transfection is described below. To examine the time course of phosphorylation of STAT3 and SRC, intracellular cAMP accumulation, COX-2 abundance, and PGE2 secretion upon cortisol treatment, the cells were treated with cortisol (1 μM) for 1, 3, 6, and 12 hours. The concentration-dependent effect of cortisol on STAT3 phosphorylation was also examined by incubating the cells with cortisol from 0.1 nM to 1 μM for 6 hours. To test whether the cAMP-PKA-SRC pathway is involved in the induction of STAT3 phosphorylation, COX-2 abundance, and PGE2 production by cortisol, the cells were treated with cortisol (1 μM; Sigma) for 6 hours in the presence or absence of the PKA inhibitor H89 (10 μM; Sigma), adenylate cyclase inhibitor SQ22536 (10 μM; Tocris Bioscience), and SRC inhibitor AZD0530 (20μM; Selleck Chemicals). To further validate the involvement of cAMP-PKA pathway, the cells were treated with FSK (100 μM; Sigma), an adenylyl cyclase stimulator, or db-cAMP (100 μM; Sigma), a cAMP analog, for 2 hours in the presence or absence of the SRC inhibitor AZD0530 (20μM). To determine whether PGE2 is involved in the induction of phosphorylation of STAT3 by cortisol, the cells were treated with cortisol (1 μM) in the presence or absence of the combination of EP2 antagonist PF-04418948 (10 μM; Tocris Bioscience) and EP4 antagonist L-161982 (10 μM; Cayman Chemical) for 6 hours. The effect of PGE2 (1 μM; Cayman Chemical) on STAT3 phosphorylation was also examined after treatment of the cells for 0.5, 1, and 2 hours. All reagents were dissolved in dimethyl sulfoxide (DMSO), and the same amount of DMSO was used as vehicle control. For the antagonist study, antagonists were added 30 min before cortisol, FSK, or db-cAMP treatment.

Transfection of siRNA in amnion fibroblasts with electroporation

Amnion fibroblasts were mixed with 50 nM STAT3 siRNA (5′-CCACUUUGGUGUUUCAUAAtt-3′) (GenePharma Co.) or randomly scrambled siRNA as negative control in Opti-MEM (Life Technologies) and added to 2-mm gap cuvettes. Cells were electroporated at 175 V for 5 ms using a NEPA21 electroporator (Nepa Gene). After dilution with DMEM containing 10% NCS and antibiotics, the cells were transferred to a six-well tissue culture plate and were ready for treatment after incubation for 48 hours.

Extraction of RNA and analysis with quantitative real-time polymerase chain reaction

Total RNA was extracted from the cells treated with cortisol (1 μM) for 24 hours in the presence and absence of the STAT3 antagonist S3I-201 (10 μM; Santa Cruz Biotechnology) or through siRNA-mediated knockdown of STAT3 using a total RNA kit (Omega Bio-Tek). RNA concentration and quality were determined by measuring OD260 (optical density at 260 nm) and the ratio of OD260/OD280 with NanoDrop ND-2000. mRNA from the total RNA was reverse-transcribed to cDNA using PrimeScript RT Master Mix Perfect Real Time kit (TaKaRa). The expression of PTGS2 and HSD11B1 mRNA was determined with quantitative real-time polymerase chain reaction (qRT-PCR) using the above transcribed cDNA and Power SYBR Premix Ex Taq (TaKaRa). The annealing temperature was set at 61°C. The absolute mRNA in each sample was calculated according to a standard curve set up using serial dilutions of known amounts of specific templates against corresponding cycle threshold (Ct) values. The housekeeping gene Actb was amplified in parallel as an internal loading control. The primer sequences used for amplifying PTGS2, HSD11B1, and Actb were as follows: PTGS2, 5′-TGTGCAACACTTGAGTGGCT-3′ (forward) and 5′-ACTTTCTGTACTGCGGGTG-3′ (reverse); HSD11B1, 5′-GGAGCAGCCTCAGCACACTA-3′ (forward) and 5′-GGCAAGGCAGCTACAGTCAG-3′ (reverse); Actb, 5′-GGGAAATCGTGCGTGACATTAAG-3′ (forward) and 5′-TGTGTTGGCGTACAGGTCTTTG-3′ (reverse). The ratio of the target gene over Actb in each sample was obtained as an indication of the target gene expression.

Extraction of protein and analysis with Western blotting

Total cellular protein was extracted from the above treated cells using ice-cold radioimmunoprecipitation assay (RIPA) lysis buffer (Active Motif) containing a protease inhibitor cocktail (Sigma) and a phosphatase inhibitor (Active Motif). Protein was extracted from the amnion tissue in a similar manner after homogenization in RIPA. The abundance of total STAT3, phosphorylated STAT3 at Tyr705, total SRC, phosphorylated SRC at Tyr416, phosphorylated SRC at Ser17, COX-2, and 11β-HSD1 was determined using a standard Western blotting protocol. Briefly, after determination of protein concentration with Bradford assay, 40 μg of protein from each sample was electrophoresed in 8.5% SDS–polyacrylamide gel and transferred to the nitrocellulose membrane. After blocking with 5% nonfat milk, the membrane was incubated with STAT3 antibody (1:1000) (Cell Signaling), phosphorylated STAT3 Tyr705 antibody (1:1000) (Cell Signaling), SRC antibody (1:1000) (Cell Signaling), phosphorylated SRC Tyr416 (1:500) (Cell Signaling), phosphorylated SRC Ser17 (1:1000) (Cell Signaling), COX-2 antibody (1:200) (Santa Cruz Biotechnology), and 11β-HSD1 antibody (1:1000) (Abcam), respectively, overnight at 4°C. After washing with 1 × Tween 20/tris-buffered salt solution, the membrane was incubated with the appropriate secondary antibody conjugated with horseradish peroxidase (Sigma) for 1 hour. The enhanced chemiluminescent detection system (Millipore) was used to detect the bands with peroxidase activity. To control sampling error, the same blot was also probed for β-actin (1:1000) (Santa Cruz Biotechnology) as an internal loading control. The bands were visualized using a G-Box iChemi Chemiluminescence image capture system (Syngene). The abundance of phosphorylated STAT3 and SRC was expressed as the ratio of their band densities over total STAT3 or SRC, respectively. The ratio of band intensities of COX-2 and 11β-HSD1 over β-actin was obtained as a measure of COX-2 and 11β-HSD1 protein abundance, respectively. To study the nuclear translocation of STAT3 upon cortisol treatment, nuclear and cytoplasmic protein fractions were isolated from the cells using a Nuclear Extract Kit (Active Motif) according to the manufacturer’s protocol. The amounts of STAT3 in the nuclear and cytoplasmic fractions were quantitated by Western blotting as described above. GAPDH (1:10,000) (Proteintech) and histone 3 (1:1000) (Cell Signaling) were used as internal loading controls for cytoplasmic and nuclear protein, respectively.

Immunofluorescence staining of STAT3

To further examine the nuclear translocation of STAT3 upon cortisol stimulation, the localization of STAT3 in the amnion fibroblasts was visualized with immunofluorescence staining. On the third day of cell culture, the cells plated in Chamber Polystyrene Vessel (BD Falcon) were treated with cortisol (1 μM) for 6 hours, and the culture medium was then removed. After washing with phosphate-buffered saline (PBS) three times, the cells were fixed with 4% paraformaldehyde, permeabilized with 0.4% Triton X-100, and blocked with normal goat serum (Jackson ImmunoResearch Laboratories). Subsequently, cells were incubated with antibody against STAT3 (1:100) (Cell Signaling) overnight at 4°C. After washing with PBS, the cells were incubated with Alexa Fluor 598–labeled secondary antibody (red; 1:100) (Proteintech) in darkness at room temperature for 2 hours. The staining was examined using a fluorescence microscope (Zeiss).

Coimmunoprecipitation assay

Coimmunoprecipitation assay was carried out to examine whether STAT3 forms a complex with GR and CREB-1. Amnion fibroblasts (1 × 107) were plated in a 100-mm dish and cultured in DMEM containing 10% NCS and 1% antibiotic for 3 days. The cells were then treated with cortisol (1 μM) for 12 hours in serum-free medium. Nuclear protein was extracted from the cells using a Nuclear Exract Kit (Active Motif). Nuclear protein (10 μg) was incubated with 1:50 dilution of rabbit antibody against human GR (Santa Cruz Biotechnology), phosphorylated CREB-1 (Cell Signaling), or preimmune rabbit IgG, respectively, overnight. Protein A agarose beads were added and incubated with the above reaction mixture for 1 hour on ice to pull down the antibody/antigen complex. The antibody/antigen/agarose complex was washed adequately and denatured in Western blot loading buffer at 95°C for the subsequent detection of phosphorylated STAT3 with Western blotting using 1:500 dilution of phosphorylated STAT3 antibody. The bands were visualized as mentioned above.

ChIP assay

The binding of STAT3, CREB-1, and Pol II to PTGS2 promoter after treatment with or without cortisol (1 μM) for 24 hours was measured with ChIP assay. Upon termination of treatment, the amnion fibroblasts were fixed with 1% formaldehyde to cross-link the proteins on chromatin DNA, which was terminated with 1 M glycine. After washing with ice-cold PBS, the cells were scraped off and lysed with 1% SDS lysis buffer supplemented with protease inhibitor cocktail on ice. The lysed cells were sonicated to shear the chromatin DNA to an optimal size around 500 bp. After precleaning with Protein A Agarose/Salmon Sperm DNA (Millipore), sheared chromatin DNA was immunoprecipitated with antibodies against STAT3 (Santa Cruz Biotechnology), CREB-1 (Cell Signaling), or Pol II (Cell Signaling). Equal amounts of preimmune IgG served as negative control. The immunoprecipitate was then incubated with Magna ChIP Protein A Agarose Magnetic Beads (Millipore) and pulled down on magnetic stand. After washing, reverse cross-linking was performed in 5 M NaCl at 65°C overnight. Contaminating RNA was cleaned with ribonuclease A, and protein was digested with proteinase K. Finally, the sheared DNA recovered from reverse cross-linking was extracted using DNA extraction kit for further quantitative analysis with qRT-PCR. The sequences of the primers used for qRT-PCR are as follows: 5′-GGGGGTACGAAAAGGCGGAAAGA-3′ (forward) and 5′-CCTGGACGTGCTCCTGACGC-3′ (reverse), which amplify the region between −88 bp and +47 bp spanning the putative CREB and STAT3 binding sites (Fig. 2D). The same amount of sheared DNA without antibody precipitation after reverse cross-linking served as input control. For qRT-PCR, the absolute DNA levels in each sample were calculated according to a standard curve set up using serial dilutions of known amounts of specific templates against corresponding Ct values. The ratio of DNA precipitated by STAT3, CREB-1, and Pol II antibody over input control was obtained to indicate the amounts of bound transcription factors.

Measurement of intracellular cAMP concentration in human amnion fibroblasts

The intracellular cAMP level in the amnion fibroblasts was measured using an EIA kit (Cayman) after treatment with cortisol (1 μM) in serum-free DMEM for 1, 3, and 6 hours in the presence of the phosphodiesterase inhibitor 3-isobutyl-1-methyl xanthine (IBMX, 500 μM; Sigma). After treatment, the culture medium was removed, and the cells were washed three times with ice-cold PBS (1 ml) and lysed in 0.1 M HCl (1 ml) in the presence of IBMX for 20 min. Lysates were assayed for the cAMP concentration following the manufacturer’s protocol.

Measurements of PGE2 and cortisol with EIA

Cortisol and PGE2 in the amnion tissue collected from pregnant women at TNL and TL were extracted with ethyl acetate after homogenization in PBS. After evaporation of ethyl acetate, the extract was reconstituted in assay buffer. Cortisol and PGE2 were measured with EIA kits (cortisol kit from R&D and PGE2 kit from Cayman) according to the manufacturer’s protocol. PGE2 in the culture medium collected from cultured amnion fibroblasts treated with or without cortisol (1 μM, 6 hours) in the presence or absence of the STAT3 antagonist S3I-201 or siRNA-mediated knockdown of STAT3, the PKA inhibitor H89 (10 μM), the adenylate cyclase inhibitor SQ22536 (10 μM), or SRC inhibitor AZD0530 (20 μM), as well as from the cells treated with or without FSK (100 μM) and db-cAMP (100 μM) for 2 hours in the presence or absence of the SRC inhibitor AZD0530 (20 μM) was measured directly without ethyl acetate extraction.

Statistical analysis

All data are reported as means ± SEM. The number for each study indicates repeated experiments using fetal membranes from different pregnancies. After examination of normal distribution, paired Student’s t test or one-way analysis of variance (ANOVA) test followed by the Newman-Keuls multiple comparison test was used where appropriate to assess significant differences. Spearman correlation analysis was performed to test the correlation between groups. Significance was set at P < 0.05.

REFERENCES AND NOTES

Funding: National Natural Science Foundation of China (grants 81330018 and 81270704) and National Key Basic Research Program of China (grant 2011CB944403). Author contributions: W.W., C.G., and K.S. designed the study. W.W., C.G., C.L., J.L., L.M., and K.S. produced and analyzed data. P.Z., W.L., H.X., and Z.-j.C. performed clinical sample collection and analyses. W.W., L.M., and K.S. contributed to drafting the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Signaling

Navigate This Article