Estradiol-17β and its Cytochrome P450- and Catechol-O-Methyltransferase–Derived Metabolites Selectively Stimulate Production of Prostacyclin in Uterine Artery Endothelial CellsNovelty and Significance
Role of Estrogen Receptor-α Versus Estrogen Receptor-β
Metabolism of estradiol-17β to 2-hydroxyestradiol, 4-hydroxyestradiol, 2-methoxyestradiol, and 4-methoxyestradiol contributes importantly to the vascular effects of estradiol-17β in several vascular beds. However, little is known about the role of estradiol-17β metabolites via the different estrogen receptors (ER-α/ER-β) on de novo endothelial prostacyclin and thromboxane production. We hypothesized that estradiol-17β and its metabolites, via ER-α or ER-β, can enhance the prostacyclin/thromboxane ratio through the classic phospholipase A2, cyclooxygenase-1, and prostacyclin synthase pathway in ovine uterine artery endothelial cells (UAECs) derived from pregnant (P-UAECs) versus nonpregnant (NP-UAECs) ewes. Western analyses showed higher expression of phospholipase A2, cyclooxygenase-1, and prostacyclin synthase in UAECs from the pregnant state, whereas thromboxane synthase was lowered in UAECs from the pregnant state. In UAECs from the pregnant state, estradiol-17β, 2-hydroxyestradiol, 4-hydroxyestradiol, 2-methoxyestradiol and 4-methoxyestradiol concentration and time-dependently increased prostacyclin compared with controls. Prostacyclin increases in UAECs from the nonpregnant state were of a lower magnitude. Estradiol-17β and its metabolites stimulated higher prostacyclin/thromboxane ratios in UAECs from the pregnant state compared with UAECs from the nonpregnant state. Estradiol-17β–induced prostacyclin increases were abrogated by the antagonists SC-560 (cyclooxygenase-1), U-51605 (Prostacyclin synthase), ICI 182 780 (ICI; both ER-α/β), and 1,3-bis(4-hydroxyphenyl)-4-methyl-5-[4-(2-piperidinyleth oxy)phenol]-1H-pyrazole dihydrochloride (MPP; ER-α), but not by 4-[2-phenyl-5,7-bis (trifluoromethyl) pyrazolo[1,5-a]pyrim idin-3-yl]phenol (PHTPP; ER-β). Prostacyclin increases induced by its metabolites were abolished by SC-560 and U-51605, but unaltered by ICI, MPP, or PHTPP. Our findings demonstrate that estrogen via primarily ER-α and its metabolites via ER-independent mechanisms influence the de novo endothelial biosynthesis of prostacyclin, which may be important in the regulation of vascular tone. These findings also shed light on the complexities of estrogen signaling via its metabolism and the functional heterogeneity of the ERs.
Prostacyclin (PGI2) and thromboxane (TXA2), 2 major vasoactive prostanoids that exert opposing effects on vascular tone, are end products of the sequential reactions catalyzed by cytosolic phospholipase A2 (cPLA2), cyclooxygenase-1/2 (COX-1/2), and either PGI2 synthase (PGIS) or TXA2 synthase (TXAS) enzymes, respectively.1 Classically, PGI2, a powerful vasodilator, is produced mainly by vascular endothelial cells, whereas TXA2, a potent vasoconstrictor, is principally produced by platelets.1 In this regard, little is known about whether endothelial cells can produce TXA2 and whether this production plays a role in regulating the ratio of PGI2 and TXA2. PGI2/TXA2 ratio is considered of relevance in the regulation of physiological and clinical vascular tone, with several studies showing that an imbalance in the generation of PGI2 relative to TXA2 is associated with hypertension, atherosclerosis, and gestational vascular diseases such as preeclampsia.2–4
The dramatic rises in uterine blood flow during pregnancy are temporally associated with increases in de novo uterine vascular PGI2 secretion and are accompanied by augmented expression of uterine artery endothelial cPLA2, COX-1, and PGIS.5–8 These uterine blood flow rises are also partly mediated by increases in the plasma levels and actions of estrogen via the classical estrogen receptors (ERs).9–11 Infusion of estradiol-17β (E2β) in sheep causes rises in uterine blood flow, increases the uterine artery endothelial expression of cPLA2, COX-1, and PGIS, which leads to increases of the stable PGI2 metabolite, 6-keto-Prostaglandin F1α (6-keto-PGF1α).7,12,13 In human umbilical vein endothelial cells (HUVECs), E2β has also been shown to selectively stimulate PGI2 production in vitro primarily via ER-α.14 However, less is known about the influence of E2β on endothelial TXA2 production and on the endothelial PGI2/TXA2 ratio.
The stimulatory effects of estrogen on uterine vascular endothelial PGI2 production may be further modulated by its biologically active metabolites. E2β is sequentially metabolized by cytochrome P450s to form the catecholestradiols, 2-hydroxyestradiol (2-OHE2) and 4-hydroxyestradiol (4-OHE2), followed by metabolism of these catecholestradiols by catechol-O-methyltransferase to form 2-methoxyestradiol (2-ME2) and 4-methoxyestradiol (4-ME2).15–18 In cultured HUVECs, 2-ME2 stimulates production of PGI2.19 Despite this knowledge, very little is known about the effects of E2β metabolites on endothelial PGI2 levels. Furthermore, nothing is known about the roles of the classical ERs on E2β metabolite-induced endothelial PGI2 levels. It is also unclear whether E2β metabolites can alter TXA2 production and the PGI2/TXA2 ratio in endothelial cells and whether this is altered in pregnancy.
Thus, we hypothesized that E2β and its metabolites, 2-OHE2, 4-OHE2, 2-ME2 and 4-ME2, can augment the PGI2/TXA2 ratio via ER-α and ER-β, through the classical cPLA2-COX-1-PGIS pathway in uterine artery endothelial cells (UAECs) in the pregnant (P-UAECs) compared with UAECs in the nonpregnant (NP-UAECs) state. Thus, we investigated: (1) the expression of cPLA2, COX-1, PGIS, and TXAS in P-UAECs versus NP-UAECs; (2) whether treatment with E2β, 2-OHE2, 4-OHE2, 2-ME2 or 4-ME2 can increase the PGI2/TXA2 ratio in P-UAECs more than NP-UAECs; (3) the role of COX-1 and PGIS in regulating P-UAEC PGI2 levels; and (4) the roles of ER-α and ER-β on PGI2 levels induced by E2β and its metabolites.
Cell Preparation and Culture
Animal use protocols and procedures were approved by the University of Wisconsin-Madison School of Medicine Research Animal Care Committee. Ovine NP-UAECs and P-UAECs were isolated, validated, and cultured from nonpregnant (luteal n=2 and follicular n=2) and late gestation (120–130 days; term, 147 days; n=4) ewes as previously described.6,17,18 At passage 4 and ≈70% confluence and serum-starved (24 hours), cells were lysed for Western blotting or transferred to 6-well plates for treatments as needed for respective experiments.
Protein Extraction and Western Immunoblotting
Protein extraction and Western Immunoblot analyses were performed as described previously.6,17,18 cPLA2, COX-1, PGIS, and TXAS expressions were detected using mouse anti-cPLA2, rabbit anti-COX-1, rabbit anti-PGIS, or rabbit anti-TXAS (1:1000) and respective secondary antibodies (1:2000). β-actin was used as loading control. Positive control was only used for the expression of TXAS using human platelet lysates. Human platelet lysates were a generous donation from the Platelet and Neutrophil Immunology Laboratory, Blood Center of Wisconsin.
Experimental Treatments, Blockade of Enzymes and Receptors
All experiments were performed in quadruplicate and replicated in at least 4 different P-UAEC (n=4) and NP-UAEC (n=4) preparations. For time and concentration–response studies, P-UAECs in 6-well plates were serum-starved (24 hours) in Endothelial Basal Medium (EBM), washed with serum-free EBM, and the medium was replaced with EBM vehicle (Control) or EBM containing 0.1, 1, 10, or 100 nmol/L of E2β, 2-OHE2, 4-OHE2, 2-ME2, 4-ME2, or Ca2+ Ionophore (A23187; positive control) for 0, 2, 4, 8, 12, or 24 hours. For nonpregnant versus pregnant concentration–response studies, NP-UAECs and P-UAECs in 6-well plates were serum-starved (24 hours) in EBM, washed with serum-free EBM, and medium was replaced with EBM vehicle (Control) or EBM containing 0.1, 1, 10, or 100 nmol/L of E2β, 2-OHE2, 4-OHE2, 2-ME2, 4-ME2, or A23187 for 12 hours based on time course studies. COX-1 or PGIS blockade was performed by pretreating P-UAECs with SC-560 and U-51605 (1 μmol/L, 1 hour), respectively, followed by E2β, 2-OHE2, 4-OHE2, 2-ME2, 4-ME2, or A23187 treatments at optimal dose (determined from concentration–response curves). ERs were blocked by pretreating P-UAECs for 1 hour with 1 µmol/L of the nonselective ER antagonist ICI 182 780, ER-α–selective antagonist 1,3-bis(4-hydroxyphenyl)-4-methyl-5-[4-(2-piperidinyleth oxy)phenol]-1H-pyrazole dihydrochloride (MPP), or the ER-β–selective antagonist 4-[2-phenyl-5,7-bis (trifluoromethyl) pyrazolo[1,5-a]pyrim idin-3-yl]phenol (PHTPP) followed by treatments with EBM vehicle or EBM containing 0.1, 1, 10, or 100 nmol/L of E2β, 2-OHE2, 4-OHE2, 2-ME2, 4-ME2, or A23187 for optimal time determined from time courses.
PGI2 and TXA2 Assays
After steroid treatments, media from individual 6-well plates were collected to measure production levels of PGI2 or TXA2 by using enzyme immunoassay kits (Cayman Chemical, Ann Arbor, MI). Indices of both PGI2 and TXA2 levels were calculated from measuring their stable nonenzymatic hydrolysis products, 6-keto-PGF1α and thromboxane B2 (TXB2), respectively, in duplicates. Productions were calculated per manufacturer’s instructions after subtracting the value of the blank wells to remove background absorbance values. The levels of PGI2 or TXA2 in NP-UAECs and P-UAECs are expressed as the fold or ratio increases over untreated control in pg/mL corrected against a standard curve, nonspecific binding, maximum binding and normalized to the amount of protein per well in mg.
Data are presented as a fold change of untreated control and expressed as means±SEM. For the PGI2/TXA2 ratio, data are presented as the ratio of PGI2/TXA2 calculated from the absolute pg/mL concentrations from the same treatment wells in duplicates from all cell lines studied. Data were analyzed using a 2-way ANOVA (SigmaPlot 11 Statistical Software). When appropriate, an analysis of the simple effect was performed using 1-way ANOVA followed by post hoc Student-Newman-Keuls test. Pairwise comparisons were performed using Student-Newman-Keuls test. Level of significance was established a priori at P<0.05.
Expression of cPLA2, COX-1, PGIS, and TXAS in NP-UAECs Versus P-UAECs
Western immunoblotting revealed the protein expression of cPLA2, COX-1, PGIS, and TXAS in both NP-UAECs and P-UAECs (Figure 1A). Densitometric analyses showed that the protein expressions of cPLA2, COX-1, and PGIS were significantly higher in P-UAECs compared with NP-UAECs (Figure 1B). In contrast, the expression of TXAS was significantly reduced in P-UAECs compared with NP-UAECs (Figure 1B).
Basal and Calcium Ionophore Stimulated PGI2 and TXA2 Production in NP-UAECs and P-UAECs
Basal production of PGI2 was higher by P-UAECs compared with NP-UAECs (Figure 2A). Unstimulated basal production of PGI2 by P-UAECs at 8, 12, and 24 hours was 1.31±0.23-, 1.43±0.23-, and 1.53±0.23-fold, respectively, corresponding to 47±0.75, 58±0.70, and 68±0.16 pg/mg of protein, respectively, at the same time points. A23187 stimulated significant production of PGI2 in P-UAECs compared with NP-UAECs (Figure 2B). Maximum production by P-UAECs at 12 hours was 10.8±0.39-fold of control compared with maximum production by NP-UAECs of 6.0±0.37-fold of control.
Basal production of TXA2 by P-UAECs was not different compared with production by NP-UAECs (Figure 2C). Unstimulated basal production of TXA2 by P-UAECs at 8, 12, and 24 hours was 1.18±0.19-, 1.20±0.19-, 1.33±0.36-fold, respectively, corresponding to 13±1.0, 20±1.1, and 23±0.22 pg/mg of protein, respectively, at the same time points. However, A23187 stimulated significantly lower concentration-dependent production of TXA2 in P-UAECs compared with NP-UAECs (Figure 2D). The maximum production by P-UAECs at 12 hours was 1.49±0.09-fold of control compared with maximum production by NP-UAECs at 1.93±0.64-fold of control.
E2β, 2-OHE2, 4-OHE2, 2-ME2, and 4-ME2 Stimulated a Time- and Concentration-Dependent PGI2 Increase in P-UAECs
Time- and concentration-dependent PGI2 increases were observed in P-UAECs after E2β treatment with highest responses observed at a concentration of 100 nmol/L at 12 hours treatment time (Figure 3A). Increases of PGI2 were noted after 4 hours in response to almost all concentration of E2β studied with the exception of 0.1 nmol/L. However, at 4, 8, 12, 24 hours, there were significant differences in PGI2 levels in response to 10 and 100 nmol/L E2β which were higher compared with 0.1 or 1 nmol/L concentrations.
Similarly, exposure of P-UAECs to 2-OHE2 and 4-OHE2 treatments also stimulated a time- and concentration-dependent PGI2 increases by P-UAECs with maximum responses observed at a concentration of 100 nmol/L at 12 hours of treatment (Figure 3B and 3C). Nevertheless, PGI2 increases were noted after 4 hours in response to all concentration of 2-OHE2 and 4-OHE2 studied, and this increased in a time- and concentration-dependent manner compared with untreated control. No further increases in 2-OHE2- and 4-OHE2 -induced PGI2 responses by P-UAECs were seen after 12 hours at all concentrations studied. There were significant differences in PGI2 increases in response to 10 and 100 nmol/L of 2-OHE2 and 4-OHE2, which were higher compared with 0.1 or 1 nmol/L concentrations at 2, 4, 8, 12, and 24 hours. However, 0.1 or 1 nmol/L concentrations stimulated significant increases of PGI2 compared with untreated control at 2, 4, 8, 12, and 24 hours.
The time course and concentration responses of PGI2 levels in P-UAECs induced by 2-ME2 and 4-ME2 are shown in Figure 3D and 3E. Both methoxyestradiols increased PGI2 in a time- and concentration-dependent manner, with maximal effects observed at a concentration of 100 nmol/L at 12 hours of treatment. However, increases of PGI2 by P-UAECs were seen at all concentrations studied at 2 hours of treatment time or greater. There were significant differences in PGI2 increases in response to 1, 10, and 100 nmol/L of 2-ME2 and 4-ME2, which were higher compared with 0.1 nmol/L concentration at 4 hours of treatment or greater.
E2β, 2-OHE2, 4-OHE2, 2-ME2, and 4-ME2 Stimulated a Greater Concentration-Dependent PGI2 Increase by P-UAECs More Than NP-UAECs
A concentration-dependent increase in production of PGI2 was observed by P-UAECs after E2β treatment with highest responses observed at a concentration of 100 nmol/L (Figure 4A). On the contrary, E2β also induced PGI2 increases by NP-UAECs, however the increase by NP-UAECs was significantly less compared with P-UAECs and no further concentration response was observed after the 10 nmol/L concentration (Figure 4A). Differences in PGI2 increases between NP-UAECs and P-UAECs were observed at 1 nmol/L, 10 nmol/L, 100 nmol/L, and 1 µmol/L concentrations.
Similar to E2β, 2-OHE2 and 4-OHE2 treatments also stimulated greater concentration-dependent PGI2 production increases by P-UAECs compared with NP-UAECs (Figure 4B and 4C). Maximum responses were observed at a concentration of 100 nmol/L, with no further increases seen with higher concentrations. 2-OHE2 and 4-OHE2 treatments did stimulate elevations of the production of PGI2 in NP-UAECs; however, these responses were significantly lower compared with the P-UAEC responses, with no further increases seen after the 10 nmol/L concentration.
2-ME2 and 4-ME2 treatments also stimulated greater concentration-dependent PGI2 production increases by P-UAECs compared with NP-UAECs (Figure 4D and 4E). Maximum responses of P-UAECs to 2-ME2 were seen at 100 nmol/L, with no further increases at greater doses. On the contrary, 4-ME2 induced maximum PGI2 production increases by P-UAECs at a concentration of 10 nmol/L, with no significant further increases with higher concentrations. Interestingly, 4-ME2 was the only estrogen metabolite that stimulated a difference in PGI2 production responses between NP-UAECs and P-UAECs at a concentration as low as 0.1 nmol/L. Elevated PGI2 production was also noted by NP-UAECs in response to 2-ME2 and 4-ME2; however, these responses were of lower magnitude compared with P-UAEC responses, with no further increases seen after the 10 nmol/L concentration.
The levels of TXA2 production stimulated by E2β, 2-OHE2, 4-OHE2, 2-ME2, and 4-ME2 increased linearly but did not reach statistical significance compared with untreated control in either NP-UAECs or P-UAECs. Therefore, these TXA2 productions are discussed within the context of the PGI2/TXA2 ratio.
The PGI2/TXA2 Ratio: E2β, 2-OHE2, 4-OHE2, 2-ME2, and 4-ME2 Stimulated a Greater Concentration-Dependent PGI2/TXA2 Ratios in P-UAECs More Than NP-UAECs
As illustrated in the Table, the basal PGI2/TXA2 ratio was higher in P-UAECs compared with NP-UAECs at 12 hours treatment. There was a concentration-dependent increase in the PGI2/TXA2 ratio when P-UAECs were exposed to E2β and its metabolites.
E2β, 2-OHE2, 4-OHE2, 2-ME2, and 4-ME2 Stimulated De Novo PGI2 Increases in P-UAECs via COX-1 and PGIS
Antagonism with the COX-1 inhibitor SC-560 and the PGIS inhibitor U-51605 was tested at all concentrations (0.1, 1, 10, and 100 nmol/L) of E2β and its metabolites studied; however, because all concentrations yielded similar results; only the responses from the optimal concentration of 100 nmol/L at 12 hours of treatment are shown here. Neither SC-560 nor U-51605 (1 μmol/L, 1 hour) pretreatments significantly altered basal control PGI2 increases in P-UAECs (Figure 5A and 5B). However, both antagonists completely abrogated the PGI2 increases in P-UAECs stimulated by 100 nmol/L of E2β, 2-OHE2, 4-OHE2, 2-ME2, 4-ME2, and the Ca2+ Ionophore (A23187; data not shown), indicating that the PGI2 increases seen is de novo via the activities of COX-1 and PGIS (Figure 5A and 5B).
E2β, but Not its Metabolites, Stimulated PGI2 Production in P-UAECs via the Classic ERs
Antagonism with ICI 182 780 was tested at 0.1, 1, 10, and 100 nmol/L of E2β and its metabolites studied; however, because all concentrations yielded similar results, only the responses from the optimal concentration of 100 nmol/L are shown. Furthermore, because the NP-UAECs exhibited lower nonsignificant responses, antagonism studies were only carried out in P-UAECs. ICI 182 780 alone did not affect basal PGI2 production in P-UAECs; however, it totally abrogated PGI2 increases in response to E2β, indicating possible involvement of either ER-α and ER-β (Figure 6). In contrast, ICI 182 780 did not have an effect on P-UAEC PGI2 increases in response to 2-OHE2, 4-OHE2, 2-ME2, or 4-ME2, indicating ER-independent mechanisms. ICI 182 780 did not have an effect on PGI2 production by P-UAECs in response to nonreceptor stimulation with A23187 (data not shown), validating specific ER-mediated E2β responses (Figure 6).
E2β, but Not its Metabolites, Stimulated PGI2 Production in P-UAECs Selectively via ER-α and Independent of ER-β
Similar to ICI 182 780 studies, antagonism with MPP and PHTPP were tested at all concentrations (0.1, 1, 10, and 100 nmol/L) of E2β and its metabolites studied. All concentrations examined yielded similar results; therefore, only the data from the optimal concentration of 100 nmol/L are shown. In P-UAECs, ER-α blockade with 1 µmol/L of MPP completely abolished the PGI2 increases stimulated by 100 nmol/L of E2β indicating a role for ER-α (Figure 7A). In contrast, MPP did not abrogate the PGI2 increases stimulated by 100 nmol/L of 2-OHE2, 4-OHE2, 2-ME2, or 4-ME2 similar to the above ICI 182 780 (Figure 7A). Moreover, E2β-induced PGI2 production responses were not inhibited by 1 µmol/L of the ER-β–selective antagonist PHTPP, demonstrating a lack of requirement for ER-β in these responses (Figure 7B). Similar to MPP, PHTPP did not inhibit the PGI2 production stimulated by 100 nmol/L of 2-OHE2, 4-OHE2, 2-ME2, or 4-ME2 (Figure 7B). Both MPP and PHTPP did not inhibit PGI2 production by P-UAECs induced by the nonreceptor stimulation using Ca2+ Ionophore (A23187; data not shown), further validating specific ER-α mediated E2β responses.
The key findings observed from this study are as follows: (1) Pregnancy induces higher protein expression of cPLA2, COX-1, and PGIS in P-UAECs compared with NP-UAECs as well as a lower protein expression of TXAS in P-UAECs compared with NP-UAECs; (2) E2β and its metabolites, 2-OHE2, 4-OHE2, 2-ME2, and 4-ME2, stimulate time- and concentration-dependent increases in PGI2 production by P-UAECs; (3) E2β and its metabolites stimulate a concentration-dependent increase in the PGI2/TXA2 ratio more in P-UAECs compared with NP-UAECs; (4) E2β and its metabolites stimulate de novo PGI2 production by P-UAECs via activities of COX-1 and PGIS; and (5) E2β-induced PGI2 production by P-UAECs is mediated primarily via ER-α and independent of ER-β, whereas 2-OHE2, 4-OHE2, 2-ME2, and 4-ME2 stimulate PGI2 production in P-UAECs independent of either ER-α or ER-β.
We demonstrate herein that P-UAECs highly express the prostanoid system enzymes including cPLA2, COX-1, and PGIS. These data are consistent with reports of increased expression of these enzymes in ovine uterine arteries in vivo and ex vivo6,13,20,21 and shows that the elevated expression during gestation is maintained to a great extent even through passaging in culture. However, in the present study we demonstrate, for the first time, that TXAS is expressed in UAECs and that expression is lower in P-UAECs compared with NP-UAECs. Taken together, these data from our studies suggest that during pregnancy, the prostanoid enzyme system shifts its expression pattern in favor of more PGI2 production and less TXA2 production in support of rises and maintenance of uterine blood flow. These results demonstrate that PGI2 production by P-UAECs is induced by gestational programming at the level of increased endothelial cell signaling, supporting earlier reports that pregnancy-specific programming in P-UAECs leads to increased basal and agonist-mediated responsiveness.6 The retention of programming in P-UAECs, perhaps via epigenetic mechanisms, may also be responsible for elevated responses to steroids in subsequent pregnancies in vivo.22 We also show herein that the calcium ionophore (A23187) induces significant increases of PGI2 production by P-UAECs and TXA2 by NP-UAECs demonstrating that the prostanoid enzymes expressed in these endothelial cells are functional and capable of eliciting calcium-dependent and receptor-independent PGI2 and TXA2 production. In this study, we demonstrate that E2β induces concentration-dependent increases of PGI2 and PGI2/TXA2 ratios more in P-UAECs than in NP-UAECs. Although maximum responses are noted at a high concentration of 100 nmol/L, low physiological concentrations of 1 and 10 nmol/L stimulated significant concentration-dependent increases of PGI2 and PGI2/TXA2 ratios more in P-UAECs than in NP-UAECs. In support of these findings are reports that E2β induces production of PGI2 in other cultured endothelial cells such as HUVECs, bovine pulmonary artery, and aortic endothelial cells.14,23–25 We have previously reported that the infusion of E2β in sheep significantly increases the ex vivo uterine arterial production of the stable PGI2 metabolite 6-keto-PGF1α.21 However, to the best of our knowledge, this is the first study to evaluate the in vitro comparison with E2β-induced uterine endothelial PGI2 production and elevated PGI2/TXA2 ratios in the pregnant versus the nonpregnant states. In this regard, these data are also consistent with our previous findings that P-UAECs exhibit pregnancy-specific PGI2 production in response to ATP, basic fibroblast growth factor, and epidermal growth factor.6
Our findings that 2-OHE2 and 4-OHE2 stimulate increases of PGI2 and PGI2/TXA2 ratios more in P-UAECs than in NP-UAECs support our hypothesis that cytochrome P450-derived metabolites of E2β may play a role in the regulation of vascular responsiveness during pregnancy. It has been previously demonstrated that the uterine arterial infusion of 2-OHE2 and 4-OHE2 in sheep and gilts causes vasodilation and increases in uterine blood flow.26,27 2-OHE2 and 4-OHE2 also significantly augment endothelial-dependent vasodilation of preconstricted vascular beds in ZSF1 rats, an animal model for hypertension, type 2 diabetes mellitus, hyperlipidemia, nephropathy, and metabolic syndrome.16,28 Because 2-OHE2 and 4-OHE2 are rapidly converted to their methoxy derivatives in the presence of catechol-O-methyltransferase, it is likely that the actions of the catecholestradiols on endothelial PGI2/TXA2 ratios maybe partly modulated by catechol-O-methyltransferase expression and activity in these endothelial cells. Nevertheless, these observations suggest that the metabolism of E2β to the catecholestradiols, 2-OHE2 and 4-OHE2, may also play an essential role in the regulation of physiological vascular responsiveness via production of endothelial-derived vasodilatory factors.
The observation that 2-ME2 and 4-ME2 enhance PGI2 levels and PGI2/TXA2 ratios in P-UAECs more than in NP-UAECs supports evidence that methoxyestradiols may positively influence vascular responsiveness during pregnancy. Consistent with these findings is the report that 2-ME2 induces PGI2 production in HUVECs.19 Low 2-ME2 level has been implicated in preeclampsia, a disease characterized by low plasma and urinary PGI2 and impaired uterine blood flow.29,30 These observations suggest that 2-ME2 may be a promising physiological as well as pharmacological agent capable of clinically improving vascular responsiveness. We demonstrate for the first time that 4-ME2 also stimulates in vitro PGI2 increases greater in the pregnant compared with the nonpregnant state and indeed may play a role in positive pregnancy-induced uterine vascular responsiveness. Because 2-ME2 and 4-ME2 induced higher PGI2 increases at very low physiologic concentrations compared to E2β and the catecholestradiols, this suggests that the methoxyestradiols maybe more potent under these conditions and points to the notion that the vascular effects of locally produced or even circulating estrogen metabolites may be more critical than previously thought compared to the effects of the parent substrate.
Demonstrating a role for ER-α and ER-β, ICI 182,780, which nonspecifically blocks both ER-α and ER-β, completely abrogated the E2β-induced PGI2 production in P-UAECs. Previous studies have demonstrated that ICI 182,780 also inhibits E2β-induced production of PGI2 in other cultured endothelial cells, including HUVECs and ovine fetal pulmonary artery endothelial cells.14 However, because of the potential relevance of ER subtype selectivity in vascular function, there is considerable interest in investigating whether the classical ERs exhibit functional heterogeneity in the regulation of E2β-induced endothelial functions. Our data show that E2β-induced PGI2 increases in P-UAECs are completely inhibited by the ER-α–specific MPP and unaffected by ER-β–specific PHTPP, demonstrating that E2β-induced PGI2 production in P-UAECs is primarily mediated by ER-α. These data are in agreement with previous observations that PGI2 production by HUVECs was seen when these cells were treated with an ER-α–selective agonist 4,4′,4″-(4-Propyl-[1H]-pyrazole-1,3,5-triyl)trisphenol.14 These data also support the notion that whereas E2β binds and activates both ER-α and ER-β, it is the molecular and structural-based differences in these receptors that allow for a wide range of functional heterogeneity, which may partly explain the selective actions of E2β.
In the present study, unlike E2β, the effects of its metabolites on PGI2 production by P-UAECs are not inhibited by the ER antagonists used in this study and thus are not mediated via ER-α and ER-β. This is in agreement with several reports that the vascular physiology and pharmacology of E2β metabolites on many cell types, including endothelial cells, occur via ER-independent mechanisms.16–18 Thus, other receptor-mediated mechanisms not involving the classical ERs may mediate E2β metabolites-induced PGI2 production in P-UAECs. Indeed we and others have reported that the genomic effects of E2β metabolites, including proliferation of P-UAECs and suppression of pancreatic islet insulin release, are mediated via the adrenergic receptors.18,31 Nevertheless, the exact mechanism of action of E2β metabolites on nongenomic PGI2 increases in P-UAECs remains to be determined and may likely involve adrenergic receptors and other estrogen-associated receptors like G protein–coupled receptor-30.
In conclusion, the findings from this current study demonstrate that the uterine endothelium exhibits pregnancy-specific increases in cPLA2, COX-1, and PGIS and decreases in TXAS shifting in the endothelial PGI2/TXA2 ratios toward more PGI2 production in association with rises and maintenance of uterine blood flow during pregnancy. Furthermore, we provide evidence that E2β primarily via ER-α and its metabolites via ER-independent mechanisms stimulate a higher de novo increases of the endothelial-derived vasodilator PGI2 and PGI2/TXA2 ratios in the pregnant compared with the nonpregnant state. Although maximum production of PGI2 and the PGI2/TXA2 ratio was noted with the high concentration of 100 nmol/L, low physiologically relevant concentrations of 1 and 10 nmol/L of E2β and its metabolites also stimulated significant synthesis of these prostanoids more in P-UAECs compared with NP-UAECs. Collectively, the selective responses of P-UAECs further illustrate pregnancy-specific programming at the level of the uterine artery endothelium signaling, resulting in enhanced E2β and its metabolites-mediated induction of PGI2 synthesis in P-UAECs without significantly affecting TXA2 production.
The mechanisms by which estrogens regulate vascular tone and vascular responsiveness during pregnancy are not well understood. However, studies have shown that it likely involves ER-mediated stimulation of endothelial-derived vasodilatory factors including NO and PGI2.1,11 Herein, we demonstrate novel and compelling evidence that the vasoactive/vasoprotective effects of E2β during pregnancy may also involve its sequential conversion to catecholestradiols and methoxyestradiols, which are capable of stimulating ER-independent endothelial PGI2 synthesis. Additional studies are required to understand whether the ER-independent induction of endothelial PGI2 synthesis by estrogen metabolites within uterine vasculature represents unappreciated signaling complexity of estrogens or just simply an evolutionary functional redundancy to maintain uterine blood flow and thus oxygen and nutrient delivery during gestation. Nevertheless, our findings necessitate the evaluation of catecholestradiols and methoxyestradiols in the regulation of vascular tone in physiology via endothelial-derived relaxing factors, as well as dysregulation in the pathophysiology of vascular diseases, such as hypertension, atherosclerosis, and gestational vascular diseases, such as preeclampsia.
We thank members of the Perinatal Research Labs, University of Wisconsin-Madison, and the Platelet and Neutrophil Immunology Laboratory, Blood Center of Wisconsin.
Sources of Funding
This work was supported by National Institutes of Health grants HL49210, HD38843, HL87144 (to Dr Magness), AA19446 (to Dr Ramadoss), R25-GM083252 (to M.L. Carnes), and T32-HD041921-07 (to I.M. Bird).
- Received June 19, 2012.
- Revision received July 8, 2012.
- Accepted November 17, 2012.
- © 2013 American Heart Association, Inc.
- Kobayashi T,
- Tahara Y,
- Matsumoto M,
- Iguchi M,
- Sano H,
- Murayama T,
- Arai H,
- Oida H,
- Yurugi-Kobayashi T,
- Yamashita JK,
- Katagiri H,
- Majima M,
- Yokode M,
- Kita T,
- Narumiya S
- Egan KM,
- Lawson JA,
- Fries S,
- Koller B,
- Rader DJ,
- Smyth EM,
- Fitzgerald GA
- Rupnow HL,
- Phernetton TM,
- Modrick ML,
- Wiltbank MC,
- Bird IM,
- Magness RR
- Bazer F
- Magness R
- Rupnow HL,
- Phernetton TM,
- Modrick ML,
- Wiltbank MC,
- Bird IM,
- Magness RR
- Sobrino A,
- Oviedo PJ,
- Novella S,
- Laguna-Fernandez A,
- Bueno C,
- García-Pérez MA,
- Tarín JJ,
- Cano A,
- Hermenegildo C
- Zhu BT,
- Conney AH
- Dubey RK,
- Tofovic SP,
- Jackson EK
- Jobe SO,
- Ramadoss J,
- Koch JM,
- Jiang Y,
- Zheng J,
- Magness RR
- Jobe SO,
- Fling SN,
- Ramadoss J,
- Magness RR
- Vagnoni KE,
- Magness RR
- Dasgupta C,
- Chen M,
- Zhang H,
- Yang S,
- Zhang L
- Nawroth PP,
- Stern DM,
- Kaplan KL,
- Nossel HL
- Stice SL,
- Ford SP,
- Rosazza JP,
- Van Orden DE
- Tofovic SP,
- Dubey RK,
- Jackson EK
Novelty and Significance
What Is New?
The uterine artery endothelium exhibits pregnancy-specific increases in phospholipase A2, cyclooxygenase-1, and prostacyclin synthase and decreases in endothelial thromboxane synthase expression.
Estradiol-17β primarily acting via estrogen receptor-α and its metabolites 2-hydroxyestradiol, 4-hydroxyestradiol, 2-methoxyestradiol, and 4-methoxyestradiol via estrogen receptor-independent mechanisms to stimulate higher de novo increases of the endothelial-derived prostacyclin/thromboxane ratios in the pregnant compared with the nonpregnant state.
What Is Relevant?
The selective pregnancy-specific endothelial prostanoid enzyme expressions and increases in the prostacyclin/thromboxane ratio demonstrate roles for catecholestradiols and methoxyestradiols in the regulation of vascular responsiveness in physiology via endothelial-derived relaxing factors as well as in the pathophysiology of vascular diseases characterized by blunted endothelium-dependent vasodilator responses such as hypertension, atherosclerosis, and preeclampsia.
These findings demonstrate novel and compelling evidence that the vasoactive/vasoprotective effects of estradiol-17β may involve its sequential conversion to catecholestradiols and methoxyestradiols which are capable of stimulating estrogen receptor-independent endothelial prostacyclin synthesis.