If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Regulation of fetoplacental blood flow is likely mediated by factors such as prostanoids. Estrogen and its receptors affect prostanoid biosynthesis. Previously, we demonstrated that villous endothelial cells express estrogen receptor-beta (ESR2), and we sought to determine its role in the mediation of fetoplacental vascular function.
Villous endothelial cells from uncomplicated pregnancies were isolated, cultured, and treated with estrogen. RNA interference, real-time polymerase chain reaction, Western blotting, and enzyme immunoassays were performed.
Cyclooxygenase-2 (COX-2) expression levels were not altered consistently by estrogen. RNA interference of ESR2 led to a concomitant decrease in COX-2 messenger RNA (P < .0001) and protein (P < .05) in the presence and absence of estradiol. ESR2 knock-down also led to diminished prostacyclin and thromboxane concentrations in the absence of estradiol (P < .005).
ESR2 mediates COX-2 expression levels and both prostacyclin and thromboxane concentrations in the basal state, which suggests the possibility of ligand-independent regulation of COX-2 activity and prostaglandin H2 substrate availability. Further investigation regarding ESR2 regulation of prostanoid biosynthesis and its effects on the fetoplacental vasculature is warranted.
Adequate fetoplacental blood flow is a critical element in the achievement of successful pregnancy outcome; this circulation remains incompletely understood. Doppler studies have established that, in fetal growth restriction because of uteroplacental insufficiency, aberrant umbilical arterial and venous blood flow may exist, which leads to potential risks of fetal hypoxemia and acidemia.
From a histopathologic standpoint, placentas that are affected by fetal growth restriction with abnormal umbilical artery Doppler indices often demonstrate fetal stem vessel vasoconstriction and luminal obliteration.
Together, these clinical and pathologic findings suggest that regulation of fetoplacental vascular tone is an important component in appropriate uteroplacental function.
For Editors' Commentary, see Table of Contents
The fetoplacental circulation is comprised of umbilical and fetal vessels as well as villous stem vessels and their downstream branches. Unlike other vascular systems within the human body, the placental portion of this circulation lacks innervation, which suggests that autonomic regulation of its tone does not exist.
Prostanoids are 1 group of major regulators of the vasculature. They are derived from metabolism of arachidonic acid by the endothelium and vascular smooth muscle cells and include the prostaglandins (PGD2, PGE2, PGF2α), prostacyclin (PGI2), and thromboxane (TXA2; Figure 1).
These substances are important in the local regulation of vascular tone in both normal and abnormal physiologic states. Their mechanism of action occurs in an autocrine or paracrine fashion by binding to specific receptors in target cells.
With the ability of the vasculature itself to synthesize both vasoconstricting and vasodilating mediators, an imbalance between total TXA2 and PGI2 may lead to inappropriate vasoconstriction and endothelial cell and platelet activation. For instance, PGI2 limits the vasoconstrictive response to TXA2 within the cardiovascular system.
From an obstetric perspective, neonatal PGI2 production is lower in those neonates with intrauterine growth restriction because of chronic placental insufficiency than in gestational age–matched normal neonates.
A balance between total vascular PGI2 and TXA2 may very well regulate the fetoplacental vasculature.
Within the human vasculature, estrogen plays a vital role in blood vessel homeostasis. During pregnancy, the human placenta produces dramatic amounts of the various estrogens (ie, estradiol, estrone, estriol). Estrogens affect numerous basic cellular functions that include gene expression, cellular proliferation, and cellular differentiation. Their function occurs by activation of 1 or both of 2 estrogen receptors (estrogen receptor-alpha [ESR1] and estrogen receptor-beta [ESR2]).
These estrogen receptors are expressed in a wide variety of tissue that includes endothelial cells and vascular smooth muscle cells. With respect to vascular physiologic development, they are known to regulate the expression of multiple vasodilator and vasoconstrictor proteins.
Likewise, within endothelial cell cultures from other organs, both estradiol and an ESR2-specific agonist have been found to upregulate cyclooxygenase-2 (COX-2) independently, which leads to increased substrate (ie, PGH2) for the biosynthesis of the various downstream prostanoids.
The role of estrogen and ESR2 in balancing vascular prostanoid biosynthesis remains unclear, both within the fetoplacental compartment and within others. Our objective was to determine the role of ESR2 in the mediation of fetoplacental vascular function, and we hypothesized that ESR2 upregulates COX-2. This, in turn, may lead to an alteration in the proper PGI2:TXA2 ratio in an in vivo setting, which potentially leads to a vicious cycle of vessel dysfunction and injury.
Materials and Methods
Cellular isolation and culture
Human placental villous endothelial cell isolation was performed, as previously described, after approval by the institutional review board at Northwestern University and patient consent.
Cells were isolated from placentas from uncomplicated pregnancies immediately after delivery. None of the subjects were exposed to aspirin or other nonsteroidal medications throughout pregnancy. Immunofluorescence confirmed purity of the cells (data not shown); based on previous data, primary cells were used only through the fifth passage to avoid changes in phenotype.
Cells were cultured and treated with the use of phenol red-free media that was supplemented with 5% fetal bovine serum, bovine brain extract with heparin, epidermal growth factor, hydrocortisone, and gentamicin/amphotericin B (Lonza, Walkersville, MD). Cells were starved in serum-free medium and treated with vehicle (ethyl alcohol 1:1000; Sigma-Aldrich, St. Louis, MO), estradiol (10–11 to 10–6 mol/L; Sigma-Aldrich), the ESR2-specific agonist diarylpropionitrile (10–11to 10–6 mol/L; Tocris Bioscience, Ellisville, MO), or lipopolysaccharide (from Escherichia coli 026:B6 100 ng/mL; Sigma-Aldrich).
All experiments were performed on at least 3 representative subject samples, with each repeated in triplicate, with the use of cells between the first and fifth passage. The results of all the experiments were pooled; Western blots demonstrate representative images from 1 selected subject.
The following antibodies were used for immunohistochemistry: monoclonal antibodies against ESR1 (Dako, Carpinteria, CA) and endothelial cell-specific antigen CD31 (Dako); and polyclonal antibody against ESR2 (BioGenex, San Ramon, CA). For immunoblotting, the following antibodies were used: monoclonal antibodies against beta-actin (Sigma-Aldrich), ESR2 (Millipore, Billerica, MA), and COX-2 (Cell Signaling Technology, Danvers, MA).
Placentas from uncomplicated pregnancies were obtained after delivery and fixed, paraffin-embedded, and processed by the Pathology Core Facility at Northwestern University. Antigen retrieval was performed with citrate buffer, and the primary antibodies that were described earlier were used. Immunoreactivity was determined with the use of horseradish peroxidase-conjugated secondary antibody, followed by addition of diaminobenzidine substrate.
RNA isolation and real-time polymerase chain reaction
Total RNA from primary endothelial cell cultures was extracted with Tri-Reagent (Sigma-Aldrich). One microgram of RNA was reverse transcribed with the Q-script Flex complementary DNA (cDNA) Synthesis Kit (Quanta Biosciences, Gaithersburg, MD). Specific oligodeoxynucleotide primers for COX-2 were synthesized based on its published cDNA sequence (F: 5′-GAATCATTTGAAGAACTTACAGGAG – 3′; R: 5′-GAGGCTTTTCTACCAGAAGG – 3′). Primers against ESR2, aromatase, and the constitutively expressed 36B4 were also used as described in previous reports.
Primer specificity was confirmed by single peaks demonstrated by dissociation curves after amplification of cDNA and a lack of amplification of genomic DNA.
Real-time quantitative polymerase chain reaction (PCR) was used to determine the relative amounts of each transcript with the use of the DNA-binding dye SYBR green (Applied Biosystems, Foster City, CA) and the ABI Prism 7900HT Detection System (Applied Biosystems). Cycling conditions started at 50°C for 2 minutes followed by 95°C for 10 minutes, then 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. The cycle threshold (Ct) was placed at a set level at which the exponential increase in PCR amplification was approximately parallel between all samples. Relative fold-change was calculated by a comparison of Ct values between the target gene and 36B4 as the reference guide. The 2–ΔΔCt method was used to analyze these relative changes in gene expression.
Placental endothelial cells were lysed with the use of Mammalian Protein Extraction Reagent (M-PER; Pierce, Rockford, IL). Protein concentrations were determined by colorimetric BCA Protein Assay (Pierce); equal concentrations were loaded in each well. Samples were subjected to polyacrylamide gel electrophoresis (Bio-Rac, Hercules, CA) and transferred onto nitrocellulose membranes (Invitrogen). Membranes were probed with antibodies as described earlier. Antirabbit and antimouse immunoglobulin G that was conjugated to horseradish peroxidase (Cell Signaling Technology) were used as secondary antibodies. Immunoreactive bands were visualized with an enhanced chemiluminescence detection system (GE Healthcare, Piscataway, NJ). Quantification of chemiluminescence signal intensity was performed after completion of all autoradiographic studies with ImageJ software (National Institutes of Health, Bethesda, MD).
RNA oligonucleotides that were directed against ESR2 and a mismatch negative control small interfering RNA (siRNA) were purchased from Invitrogen. Placental endothelial cells were cultured in media as previously described, but lacked gentamicin/amphotericin B. Cells were plated at a density of 4.0 × 106 cells per 10-cm dish 1 day before transfection to achieve approximately 30-50% confluence at the time of transfection. On the day of transfection, the RNAiMAX lipofectamine-based reagent (Invitrogen) was combined in conjunction with 200 nmol/L siRNA duplexes that were diluted in Opti-Mem I (Invitrogen) and applied to the cells. Six hours after the start of transfection, complete growth medium without antibiotics was added, and cells were allowed to recover and proliferate. Of note, additional controls were performed simultaneously on cells from 3 of the subjects. These controls included cells that were exposed only to Opti-Mem I media and cells undergoing mock transfection with exposure to RNAiMAX lipid reagent in the absence of RNAi oligos. Cells were starved overnight, followed by treatment with vehicle or estradiol (10–7 mol/L) for 24 hours. This was timed to allow for RNA and protein isolation at 48 and 72 hours, respectively, from the start of transfection.
Villous endothelial cells were treated with vehicle or estradiol or prepared for RNA interference studies as described earlier. Cells were serum- and supplement-starved in basal media overnight at 36 hours from the start of transfection. The medium was replaced the next morning. Twenty-four hours later, cell culture supernatant was collected, centrifuged to remove any cellular contaminants, and transferred to a fresh tube. Enzyme immunoassays were performed for thromboxane B2 (TXB2; main metabolite of TXA2) and 6-keto-prostaglandin F1 alpha (6-keto-PGF1α; main metabolite of PGI2) with a commercial kit that used competitive substrate-acetylcholinesterase assays (Cayman Chemical, Ann Arbor, MI). Concentrations were normalized to total protein concentrations.
The data from all experiments were pooled; numeric data are reported as means of the 3 replicates performed within 1 subject, with error bars that represent SEM. Statistical analysis for comparison of treatment groups was performed with the Student t test or analysis of variance followed by the Scheffe multiple comparison test, when appropriate.
In vivo distribution of ESR2 in term placentas
Cellular distribution of ESR1 and ESR2 was evaluated by immunohistochemistry in uncomplicated, term placentas. Immunoreactive ESR2 was detectable readily in the nuclei of the villous endothelial cells and syncytiotrophoblast (Figure 1). Surprisingly, ESR1 was not detected. Of note, human breast cancer tissue was used as a positive control for ESR1.
COX-2 is not induced by estradiol or diarylpropionitrile in villous endothelial cells
Previous studies have demonstrated estradiol induction of COX-2 within vascular endothelium of certain organs.
Within our model of fetoplacental endothelial cells, there was no consistent COX-2 induction with estradiol treatment in a dose- or time-dependent fashion (Figure 2, A and B). This was reproduced with another less potent ESR2 ligand, estrone, and diarylpropionitrile treatment (data not shown). We did note a 1.62-fold increase in COX-2 messenger RNA (mRNA) after 1 hour of estradiol treatment (P = .05). Similarly, statistical analysis also demonstrated significant changes with 10–7 mol/L treatment at 4 and 8 hours, although the maximal fold-change that was seen was only 1.06. Although statistically significant, this degree of induction was minimal and of uncertain relevance from a mechanistic standpoint. From a protein standpoint, there did not appear to be any COX-2 induction with estradiol or diarylpropionitrile, whereas COX-2 protein levels were inducible when lipopolysaccharide was used as a positive control (Figure 2, C and D).
Human placental villous endothelial cells do not express endogenous aromatase
Endothelial cell aromatase expression has been described previously and appears organ dependent.
With real-time PCR, aromatase cycle threshold expression was undetectable in our villous endothelial cell model in 5 subjects. Of note, cDNA from MCF-7 breast cancer cells that were used as a positive control demonstrated the presence of low, but detectable, expression (average Ct value, 34.98).
COX-2 expression is decreased in the setting of ESR2 knock-down
The lack of COX-2 induction with estradiol treatment or in the setting of an ESR2-specific agonist was surprising. To confirm that estrogen and ESR2 did not mediate COX-2 induction within villous endothelial cells, we used RNA interference to knock down ESR2 within our cultured villous endothelial cells. Transfection of ESR2 siRNA led to a consistent knock-down of ESR2 mRNA by no < 84%, and these results were not altered by estradiol treatment (P < .0001; Figure 3, A). Similarly, COX-2 mRNA levels also were ablated significantly in the setting of ESR2 knock-down; again, this appeared independent of estradiol treatment (P < .001; Figure 3, B). Although ESR2 protein knock-down was less dramatic, pooled image analyses of all Western blots demonstrated an approximate 50% knock-down in ESR2 protein (Figure 3, C and D). This led to concomitant decreases in COX-2 protein levels that occurred in the presence and absence of treatment (P < .05; Figure 3, C and D). Of note, there was no effect noted on COX-2 or ESR2 expression in the absence of transfection conditions or in the presence of lipid transfection reagent alone.
TXB2 and 6-keto-prostaglandin F1α levels are decreased after ESR2 knock-down
TXA2, which is one of the main prostanoids responsible for vasoconstriction and platelet activation, is formed by both endothelial cells and platelets locally and exerts its effect in an autocrine or paracrine manner.
Enzyme immunoassays of cells that were treated with a dose-dependent course of estradiol did not demonstrate any significant changes in TXB2 or 6-keto-PGF1α levels (data not shown). To elucidate downstream effects of ESR2 and COX-2 ablation, we performed enzyme immunoassays for TXB2 and 6-keto-PGF1α in the setting of nontarget siRNA transfection and ESR2 siRNA transfection in the absence of treatment. These results demonstrated corresponding decreases in both TXB2 and 6-keto-PGF1α concentrations in cell culture supernatant (P < .0001 and P < .005, respectively; Figure 4).
In this study, we found that, unlike other endothelial cells such as those within uterine vasculature, COX-2 was not induced reliably by direct estrogenic stimuli within villous endothelial cells. Yet, 1 of the main receptors for the various estrogens, ESR2, appears to be an important mediator in the prostanoid biosynthesis pathway.
We previously demonstrated ESR2 expression as the sole estrogen receptor within total cellular protein of cultured villous endothelial cells.
Our immunohistochemical results in this study confirm endogenous expression of ESR2 within villous endothelial cells and negate the possibility of induction from the isolation process itself. Thus, with the presence of ESR2, it was anticipated that estrogenic compounds would induce COX-2 within villous endothelial cells.
It should be noted that there did appear to be a significant induction of COX-2 at 1 hour. However, this value was driven primarily by results from 2 subjects that demonstrated approximately 5-fold induction of COX-2 at 1 hour. Within the remaining 12 subjects, fold-induction varied from 0.92-1.31 at 1 hour, which was unlikely to be biologically relevant. Similarly, there was statistical significance noted at the 10–7 mol/L dose point, the 4-hour time point, and the 8-hour time point. Of these 3 specific points, maximal fold-induction was 1.06, which again, although statistically significant, likely did not carry biologic significance.
The discrepancy between subjects at 1 hour was surprising, especially because ESR2 was expressed endogenously, and various possible explanations were considered. First, nongenomic induction of COX-2 was considered, and treatment at 5, 15, and 30 minutes was performed, with no induction noted of COX-2 (data not shown). To confirm that COX-2 was indeed inducible in villous endothelial cells of all subjects, lipopolysaccharide was used as a positive control. Even within the 12 subjects without significant COX-2 induction, lipopolysaccharide was able to stimulate a response, thereby demonstrating the potential to capture COX-2 induction within our model.
Another possible explanation was that endogenous estrogen exists within the culture system. Certain endothelial cells are capable of synthesizing estrogen through aromatase; in our system, this may have negated any further exogenous estradiol that was added to the culture medium. To test this hypothesis, we performed real-time PCR using validated aromatase primer-probes on endothelial cell cDNA from 5 separate subjects. Aromatase Ct values were undetectable in all, whereas cDNA from MCF-7 breast cancer cells that were used as a positive control demonstrated the presence of low, but detectable, expression. Without the presence of endogenous estrogen, another possibility was that our in vitro culture system was inadvertently and inconsistently affecting coactivators or corepressors. This could explain the fact that 5-fold induction was seen in 2 subjects. The final possibility was that ESR2 itself was not an actual mediator of COX-2 and prostanoid biosynthesis. To address this, we performed RNA interference studies knocking down ESR2; surprisingly, we found striking decreases in COX-2 mRNA and protein levels with ESR2 knock-down. This occurred both in the basal untreated state and under estradiol-treatment conditions.
The implications of ESR2 mediation of COX-2 within fetoplacental endothelial cells, even in the untreated basal state, are important for a number of reasons. It is possible that ESR2 is a major regulator of COX-2, where a 50% decrement in expression of a major transcription factor may be adequate to induce significant changes to downstream genes.
Furthermore, to our knowledge, ligand-independent regulation of vascular function and vasomotor tone has not been described previously, and this possibility warrants further investigation. To accomplish this, it will be important to rule out less traditional ligands such as estriol.
It will also be essential to elucidate the regions of the COX-2 promoter that are stimulated by ESR2. Several potential candidate DNA-binding sites exist and include multiple estrogen response element (ERE) half-sites (eg, –832/–827, –1493/–1398, –4671/–4666) based on results from the “Transcription Element Search System,” a computer-assisted homology search.
Although there do not appear to be any classic ERE consensus sequences (aGTTCAnnnTGACCt), there is increasing evidence that nonclassic binding of transcription factors to DNA occurs. For instance, 1 group of investigators has shown that many natural EREs deviate substantially from the classic consensus sequence, which suggests that ESR binding to half-sites may occur.
Furthermore, the possibility of transcription factor cross-talk also exists, and multiple authors have found circumstances in which estrogen receptors interact with other transcription factors, which include activating protein-1 and stimulating protein 1.
Other candidate binding sites have been described in nonvascular tissue. For example, a variant nuclear factor-κB site within the proximal COX-2 promoter has been described to be critical for COX-2 induction by malignant endometrial epithelial cells within endometrial stromal cells.
Up-regulation of cyclooxygenase-2 expression and prostaglandin synthesis in endometrial stromal cells by malignant endometrial epithelial cells A paracrine effect mediated by prostaglandin E2 and nuclear factor-kappa B.
Other response elements, which include a cyclic adenosine monophosphate response element and a CCAAT/enhancer-binding protein regulatory element have also been deemed important in COX-2 transactivation.
Interleukin-1beta elevates cyclooxygenase-2 protein level and enzyme activity via increasing its mRNA stability in human endometrial stromal cells: an effect mediated by extracellularly regulated kinases 1 and 2.
With regard to the effects of ESR2 on prostanoid biosynthesis itself, TXB2 levels were decreased in the setting of ESR2 knock-down, which suggests that ESR2 mediates endothelial cell TXA2 production. This regulation may occur through a direct effect on TXA2 synthase gene expression. However, because 6-keto-PGF1α levels were comparably decreased with ESR2 knock-down, a more likely explanation of ESR2 mediation on prostanoid biosynthesis is its effect on COX-2 and the concomitant decrease in PGH2 formation, which is the main substrate for downstream prostanoid biosynthesis. In the future, it will be important to investigate the direct effects of ESR2 on TXA2 synthase and PGI2 synthase gene expression. Furthermore, even if ESR2 does not mediate these genes directly and affects TXA2 and PGI2 production solely through control of PGH2 substrate, investigation surrounding the effects of altered PGH2 concentration within the vasculature is necessary and will help delineate the paracrine interactions between endothelial cells and its adjacent neighbors, which include platelets and vascular smooth muscle cells.
In summary, ESR2 is expressed endogenously within villous placental endothelial cells and appears to mediate COX-2 expression and TXA2 and PGI2 production. These are novel findings for 2 reasons. First, ESR2 appears to mediate COX-2 expression even in the absence of treatment, which suggests that a ligand-independent mechanism may exist to regulate COX-2. Second, ESR2 regulation of TXA2 production has not been described previously. Although its mediation of TXA2 may occur solely through its effect on COX-2 and substrate availability, it will be critical to further elucidate potential paracrine interactions. For instance, the rate of platelet TXA2 release is highly modifiable and appears to be driven largely by enhanced substrate availability.
Furthermore, within a population of subjects with unstable angina whose platelet COX-1 has been inhibited irreversibly by daily aspirin therapy, 1 group of investigators has demonstrated the ability of platelets to transcellularly convert endothelial cell-derived PGH2 to TXA2.
Thus, it is possible that ESR2 mediation of COX-2 affects PGH2 substrate availability; in an in vivo setting in which multiple cellular interactions occur (including those between endothelial cells and platelets), substrate availability ultimately may have a more significant effect on total TXA2 values in comparison with total PGI2 levels. In total, further investigation of the mechanisms that surround prostanoid biosynthesis and other humoral, vasomotor mediators within the fetoplacental vasculature is warranted and hopefully will help enhance understanding of the pathophysiologic mechanisms behind chronic placental insufficiency states.
Pathophysiology of fetal growth restriction: implications for diagnosis and surveillance.
Interleukin-1beta elevates cyclooxygenase-2 protein level and enzyme activity via increasing its mRNA stability in human endometrial stromal cells: an effect mediated by extracellularly regulated kinases 1 and 2.
Support for this research was provided by the AAOGF/SMFM Scholarship Award 2007-10 and Northwestern Memorial Foundation Private Donor Grant 2007-08.
Cite this article as: Su EJ, Lin Z-H, Zeine R, et al. Estrogen receptor-beta mediates cyclooxygenase-2 expression and vascular prostanoid levels in human placental villous endothelial cells. Am J Obstet Gyncol 2009;200:427.e1-427.e8.