Molecular Basis of Estrogen-Induced Cyclooxygenase Type 1 Upregulation in Endothelial Cells
http://www.100md.com
循环研究杂志 2005年第3期
the Department of Pediatrics (L.L.G., L.H., S.O.-L., Z.G., K.L.C., P.W.S.), University of Texas Southwestern Medical Center at Dallas and the Vascular Biology Research Center (K.K.W.), University of Texas-Houston Health Science Center, Dallas, Tex.
Abstract
Estrogen upregulates cyclooxygenase-1 (COX-1) expression in endothelial cells. To determine the basis of this process, studies were performed in ovine endothelial cells transfected with the human COX-1 promoter fused to luciferase. Estradiol (E2) caused activation of the COX-1 promoter with maximal stimulation at 10eC8 mol/L E2, and the response was mediated by either ER or ER. Mutagenesis revealed a primary role for a putative Sp1 binding motif at eC89 (relative to the ATG codon) and lesser involvement of a consensus Sp1 site at eC111. Electrophoretic mobility shift assays yielded a single complex with the site at eC89, and supershift analyses implicated AP-2 and ER, and not Sp1, in protein-DNA complex formation. In endothelial cells with minimal endogenous ER, the transfection of ER mutants lacking the DNA binding domain or primary nuclear localization signals caused 4-fold greater stimulation of promoter activity with E2 than wild-type ER. In contrast, mutant ER lacking the A-B domains was inactive. Thus, estrogen-mediated upregulation of COX-1 in endothelium is uniquely independent of direct ER-DNA binding and instead entails protein-DNA interaction involving AP-2 and ER at a proximal regulatory element. In addition, the process may be initiated by cytoplasmic ER, and critical receptor elements reside within the amino terminus.
Key Words: cyclooxygenase endothelium estrogen estrogen receptor
Introduction
Evidence from multiple paradigms indicates that estrogen has potent impact on endothelial cell function. The hormone activates signaling by endothelium, and it modifies endothelial cell growth, migration, and apoptosis. These actions are primarily mediated by estrogen receptor (ER) subtypes and , which classically function as transcription factors.1eC3 We have demonstrated that ER and ER are expressed in endothelium, and that subpopulations of ER and ER reside in endothelial caveolae/lipid rafts.4,5 One of the most thoroughly delineated targets of endothelial ER action is endothelial nitric oxide synthase (eNOS). eNOS transcription is upregulated by the activation of classical ER and estrogen-related receptor 1 in processes that involve diverse regulatory elements within the 5' flanking region of the eNOS gene.6,7 In addition, eNOS activity is rapidly stimulated by ER or ER ligand activation in endothelial caveolae.4,5 Through these multiple mechanisms estrogen enhances the production of the important atheroprotective molecule NO.8
In contrast to the detailed understanding of the basis of estrogen action on endothelial NO production, considerably less is known about the equally potent ability of the hormone to stimulate endothelial prostacyclin synthesis. This process is critical to the overall impact of estrogen on vascular health.9 We have previously demonstrated that estrogen acutely activates endothelial prostacyclin synthesis through an ER-dependent, calcium-dependent process.10 We have also shown that estrogen upregulates the expression of the rate-limiting enzyme in prostacyclin synthesis, cyclooxygenase type 1 (COX-1) in endothelial cells, resulting in enhanced basal and agonist-stimulated prostacyclin synthesis. We have further demonstrated that this is mediated by increases in steady-state COX-1 mRNA levels, which are not related to changes in mRNA degradation, thus implicating transcriptional processes.11 However, the molecular mechanisms by which COX-1 is upregulated by estrogen are unknown.
The purpose of the present study was to determine the processes by which COX-1 expression is upregulated in endothelium by estrogen. The 5' flanking sequence of the COX-1 gene lacks estrogen response elements (EREs) or ERE half-palindrome motifs.12 The hypothesis was therefore raised that estrogen activates COX-1 gene transcription via the involvement of putative Sp1 elements, which are prevalent in the COX-1 promoter and which mediate the estrogen responsiveness of certain genes.13 In addition to testing this hypothesis, experiments were performed to determine the role of ER and ER in this process. The specific cis DNA sequences involved in the estrogen response were also identified, along with nuclear proteins that are involved. Furthermore, studies were done to begin to determine the features of ER that play a role in mediating COX-1 expression.
Materials and Methods
Cell Culture
Experiments were performed in primary ovine endothelial cells obtained as previously described and studied at passage 4 to 6.14 These cells have been used previously in studies of COX-1 regulation.11,15 In experiments evaluating the actions of estrogen receptor mutants, primary endothelial cell cultures were specifically selected that had negligible expression of endogenous ER or ER by immunoblot analysis.
Construction of Reporter and Receptor Plasmids and Mutagenesis
The basis of estrogen-induced activation of COX-1 transcription was investigated using a 2075-base pair fragment (eC2095 to eC21 relative to the ATG codon, A as +1) of the human COX-1 5' flanking region inserted into KpnI/HindIII sites of the luciferase reporter gene plasmid, pGL2 (Promega Corp) to yield the full-length promoter-reporter plasmid denoted as eC2095COX-1-Luc. Two 5' deletion mutants were also studied including 1241 (eC1261 to eC21) or 117 (eC137 to eC21) base pairs of 5' flanking sequence of COX-1, designated eC1261COX-1-Luc and eC137COX-1-Luc, respectively. The COX-1 genomic sequences for these plasmids were obtained from COX-1 promoter reporter constructs originally made using pXP1.12 Site-directed mutants were also generated of the putative Sp1 element at eC111, the putative Sp1 element at eC89, or of both elements (see online data supplement available at http://circres.ahajournals.org).12
In studies of the role of estrogen receptors in COX-1 gene transcription, cells were cotransfected with cDNAs for wild-type human ER inserted into pCDNA3.14 or deletion mutants of ER. These included a mutant lacking the two primary NLS, NLS 2 and 3 (ER250eC274), a mutant lacking the DNA binding domain (ER 185eC251), and an N-terminal deletion mutant lacking residues 1 to 175 (ER1eC175) (see online data supplement and online Figure I). In selected experiments, cells were transfected with mouse ER cDNA cloned into pCDNA3.1.5
Cell Transfection and Reporter Activity
Cell transfection was performed using methods that have been previously reported.16 Twenty four hours later, the cells were placed in media containing charcoal-stripped serum and were treated with either vehicle, 17-estradiol (E2, 10eC12 to 10eC6 mol/L), or E2 plus ICI 182,780 (10eC5 mol/L) for 48 hour. The cells were lysed, extracts were centrifuged at 10 000g, and luciferase and -galactosidase activity were measured.17,18 The results are normalized as relative luciferase light units/-galactosidase activity. In selected wells, the cells were transfected with pGL2-Control Vector (Promega Corp) to serve as a positive control for luciferase expression. In preliminary studies, 17-estradiol had no effect on promoter activity. In experiments focused on the role of ER, sham plasmid or cDNAs for wild-type or mutant ER were cotransfected with the promoter-reporter constructs and pSV--Gal, and equal expression of ER forms was confirmed by immunoblot analysis.4
Electrophoretic Mobility Shift Assays
After 0 to 48 hours treatment of endothelial cells with vehicle or 10eC8 mol/L E2, nuclear extracts were prepared.16 Oligonucleotide probes were generated and added to incubations of nuclear extracts using methods that have been previously reported.16 In competition studies, excess wild-type, mutant, or unrelated oligonucleotides were added in 2- to 200-fold molar excess before the addition of the 32P-labeled probe. The wild-type eC111 Sp1 probe was 5'-GAGGGAGGAGCGGGGGTGGAGCCGGGGGAA-3' (upper strand) and the mutant probe was 5'-GAGGGAGGAGCGGTTTTAGAGCCGGGGGAA-3'. The wild-type eC89 Sp1 probe was 5'-GGGGGAAGGGTGGGGAGGGGATGGGCTGGA-3' (upper strand), and the mutant probe was 5'-GGGGGAAGGGTGTTTAAGGGATGGGCTGGA-3'.
To identify the nuclear proteins that bound to the putative Sp1 domains, supershifting of the DNA-protein complex was performed. For ER, 2 e蘈 of MAB461 (Chemicon International Inc) or unrelated IgG was added. For Sp1, Sp3, AP-2, and AP-2, 2 e蘈 of antiserum to the respective transcription factors was added. The antibodies were 2 e/e蘈 from Santa Cruz Biotechnology, Inc, and with the exception of Sp3, all were monoclonal. To confirm effective supershifting with antibody to Sp1, additional experiments were performed with an Sp1 probe from the eNOS promoter (5'-GGATAGGGGCGGGGCGAGG-3', upper strand).16 All nuclear protein-DNA complexes were resolved on 5% nondenaturing polyacrylamide gels containing 1x Tris borate/EDTA buffer. Dried gels were exposed to Kodak XAR film for autoradiograpy.
Statistical Analysis
Data for promoter activity were analyzed by ANOVA and Neuman-Keuls post hoc testing.19 Results are expressed as mean±SEM. All stated differences achieved statistical significance at the 0.05 level of probability or less.
Results
COX-1 Promoter Activation by Estrogen
To determine whether estrogen activates COX-1 transcription, the effects of the hormone on the activation of the full-length promoter (eC2095COX-1-Luc) were determined. The levels of basal promoter activity obtained with eC2095COX-1-Luc was at least 3-fold greater than with vector alone, and activity with vector alone was not altered by hormone treatment (data not shown). In cells transfected with eC2095COX-1-Luc, treatment with 10eC8 mol/L E2 for 48 hours caused a 2.6-fold increase in promoter activity that was fully reversed by concomitant treatment with ICI 182,780 (Figure 1A). In additional studies, the dose-response to E2 was evaluated (Figure 1B). Activation was observed at 10eC10 mol/L E2, and maximal response occurred with 10eC8 mol/L E2. Minimal stimulation of promoter activity was seen with 10eC6 mol/L E2, indicating a biphasic dose-response.
To further determine the role of ER in COX-1 gene modulation, experiments were done in endothelial cells overexpressing either ER or ER (Figure 1C). Relative to the responses seen with endogenous ER action in sham-transfected cells, there was a 2.3-fold greater E2-mediated promoter activity after ER overexpression. Similarly, there was a 2.2-fold enhancement of the E2 response with overexpression of ER. As such, both ER subtypes are capable of activating COX-1 transcriptional transactivation.
Promoter Elements Required for Regulation by Estrogen
To determine the elements in the COX-1 promoter required for estrogen responsiveness, the activities of 5' deletion mutants were evaluated in endothelial cells. Basal promoter activity fell by 63% with deletion from eC2095 to eC1261, and it fell a comparable 70% with deletion from eC2095 to eC137 (Figure 2A). However, the relative capacity for E2-stimulated promoter activity was conserved with deletion from eC2095 to eC1261, with the hormone causing similar 2.4-fold and 2.6-fold increases in activity over basal, respectively. In addition, there was continued conservation of E2 activation of the promoter with deletion to eC137, with eC137COX-1-Luc displaying a 2.8-fold response to hormone. Importantly, ER antagonism with ICI 182,780 reversed estrogen responses in all cases. To further implicate the proximal promoter in hormone action, additional experiments were done using eC2095COX-1-Luc and eC137COX-1-Luc in cells cotransfected with ER cDNA (Figure 2B). With ER overexpression, E2-stimulated activity of eC2095COX-1-Luc was raised by 2.4-fold. Similarly, with overexpression of the receptor the E2-induced activity of eC137COX-1-Luc increased by 2.2-fold.
Inspection of the proximal COX-1 promoter reveals a lack of estrogen response elements (ERE) or ERE half-palindrome motifs known to mediate the estrogen responsiveness of certain genes.20 However, there is a potential Sp1 binding site at eC111 and another at eC89,12 and Sp1 is known to mediate estrogen responsiveness in the absence of ERE-like motifs in certain paradigms.13 Because E2-induced activation of the promoter in endothelial cells was conserved with 5' deletions, which retained these domains, site-directed mutants of either or both of these putative Sp1 elements within eC2095COX-1-Luc were tested. First, the basal activity of mutated promoter sequences was evaluated (Figure 2C). Compared with full-length wild-type promoter (eC2095COX-1-Luc), mutation of the eC111 Sp1 caused a 58% decline in activity, mutation of the eC89 Sp1 yielded a 72% fall, and mutation of both elements caused a 53% decrease. Estrogen-stimulated promoter activity expressed relative to the level of basal activity for the same construct is shown in Figure 2D. Mutation of the Sp1 element at eC111 caused a decline in E2-induced activity of 43%, and mutation of eC89 Sp1 resulted in a 73% decline in the E2 response. Mutation of both domains yielded an 88% fall in E2-induced promoter activity.
Evaluation of Relevant Nuclear Proteins
To evaluate the nuclear proteins in endothelial cells involved in estrogen actions mediated by the consensus Sp1 elements of the COX-1 promoter at eC111 and at eC89, electrophoretic mobility shift assays were performed. Incubation of nuclear extracts from E2-treated endothelial cells (48 hours) with a double-stranded oligonucleotide probe encompassing the putative Sp1 site at eC111 resulted in the appearance of one major DNA-protein complex that was diminished by 200-fold molar excess of unlabeled probe (Figure 3A, left). In contrast, three minor complexes were formed with mutated eC111 Sp1 site probe. The use of mutated oligonucleotide as a competitor did not prevent the formation of the major complex by the wild-type probe (Figure 3A, right). Incubation of nuclear extracts with a probe containing the putative Sp1 site at eC89 resulted in the appearance of one major DNA-protein complex that were prevented by 200-fold molar excess of unlabeled probe (Figure 3B, left). The complex was minimally formed with mutated eC89 Sp1 DNA probe, and the mutated oligonucleotide only partially prevented the formation of the complex by the wild-type probe (Figure 3B, right).
To identify the endothelial nuclear proteins involved in complex formation with the presumptive eC111 Sp1 and the eC89 Sp1COX-1 promoter elements, supershift analyses were done. Two different antisera to Sp1 did not cause supershifting of either the eC111 Sp1 protein-DNA complex or the eC89 Sp1 protein-DNA complex (Figure 4A and 4B, left). To provide a control for supershifting of Sp1, a study was performed using the same endothelial nuclear proteins and the Sp1 element from the proximal eNOS promoter. With the Sp1 domain from eNOS, supershifting was observed as previously described (data not shown).16 In an effort to identify other potential nuclear proteins relevant to gene regulation by Sp1 elements, the effects of antisera to Sp3 were also assessed, and supershifting was not evident for either the eC111 Sp1 protein-DNA complex or the eC89 Sp1 protein-DNA complex (Figure 4A and 4B, middle). In studies of potential involvement of AP-2 or AP-2, which have consensus binding sequences similar to Sp1,21eC24 antisera to either protein did not cause supershifting of the eC111 Sp1 protein-DNA complex (Figure 4A, right). However, antisera to AP-2 caused supershifting of the eC89 Sp1 protein-DNA complex (Figure 4B, right). In experiments comparing nuclear proteins from endothelial cells treated with vehicle vs 10eC8 mol/L E2 for 48 hours, there was no difference in the overall formation of the eC111 Sp1 protein-DNA complex or the eC89 Sp1-protein DNA complex, and the supershifting of AP-2 in the eC89 Sp1 protein-DNA complex was also similar (data not shown).
The contribution of ER to complex formation was also assessed by supershift analyses. Using nuclear proteins from cells treated with E2 for 48 hours, antiserum to ER shifted both the eC111 Sp1 protein-DNA complex and the eC89 Sp1 protein-DNA complex, whereas an unrelated IgG did not (Figure 5A and 5B, left). In evaluations of the temporal features of ER involvement in complex formation, studies of nuclear proteins from cells treated with E2 for 0 versus 6 hours yielded comparable supershifting of ER from the eC111 Sp1 complex (Figure 5A, right). In contrast, the supershifted bands for ER in the eC89 Sp1 complex differed at 0 and 6 hours of E2 exposure; there was less prevalence of a rapidly migrating supershifted complex and greater abundance of a slowly-migrating supershifted complex after 6 hours of E2 treatment versus control (Figure 5B, right).
Features of ER Required for Modulation of COX-1 Transcription
There are multiple mechanisms by which ER regulates gene transcription.13,20 To begin to elucidate the features of the receptor necessary for modulation of COX-1, experiments were performed in a primary endothelial cell culture selected to have negligible endogenous ER or ER into which ER mutant constructs were cotransfected along with eC2095COX-1-Luc. In cells transfected with sham plasmid for wild-type ER, E2 caused no change in the activity of eC2095COX-1-Luc (Figure 6A). In contrast, cells transfected with wild-type ER displayed a 2.5-fold increase in promoter activity with E2. In simultaneous comparisons of wild-type ER and ER185eC251, E2 caused a 1.8-fold increase in promoter activity in cells expressing wild-type receptor and there was a 7-fold increase in activity with E2 in cells expressing ER185eC251 (Figure 6B). Additional independent studies revealed a 2.3-fold stimulation of promoter activity with E2 via wild-type ER and a markedly greater 9.2-fold stimulation in cells expressing ER250eC274 (Figure 6C). In contrast, the N-terminal truncation mutant of ER, ER1eC175, was incapable of E2-induced activation of the COX-1 promoter (Figure 6D).
Discussion
Estrogen causes upregulation of COX-1 expression in endothelial cells that is instrumental to the impact of the hormone on vascular function.9 We have previously demonstrated that this is mediated by an increase in steady-state levels of COX-1 mRNA, and that the degradation of the mRNA is unaltered.11 In the present study, we show that E2 activates COX-1 gene transcription in an ER-dependent manner. We further demonstrate that the process entails unique ERE-independent mechanisms, providing new insights into the complex nature of E2 actions in endothelium.
Role of ER and Regulatory Elements on COX-1 Gene
In transient transfection experiments in endothelial cells with a construct containing the human COX-1 promoter fused to luciferase, we have observed E2-induced promoter activation of sequence residing between eC2095 and eC21 of the 5' flanking sequence relative to the ATG codon. Experiments with ICI 182,780 indicated that the process is ER-dependent, and the dose-response was biphasic with a threshold concentration of 10eC10 mol/L and maximal activation at 10eC8 mol/L E2. The biphasic dose-response mimics that found previously for changes in COX-1 mRNA abundance with E2.11 Interestingly, the nongenomic activation of prostacyclin synthesis by E2 does not display a biphasic dose response,10 suggesting quite different kinetics for genomic versus nongenomic E2 modulation of prostacyclin synthesis in endothelial cells. Cotransfection studies further indicated that the COX-1 promoter response can be mediated by either ER or ER. Thus, COX-1 regulation by E2 occurs at a physiological concentration of the hormone8 and via either of the ER subtypes.
To determine the regulatory elements within the COX-1 promoter required for estrogen responsiveness, the activities of progressive 5' deletion mutants were evaluated. The capacity for E2-stimulated promoter activity was fully conserved with 5' deletion from eC2095 to eC1261 and also to eC137. In addition, the enhancement in E2 response with ER overexpression was similar for eC2095COX-1-Luc and eC137COX-1-Luc. These findings indicate that the estrogen responsiveness of the COX-1 gene resides within the proximal promoter.
Because the 5' flanking sequence of the COX-1 gene lacks EREs or ERE half-palindrome motifs, alternative modes of estrogen regulation were contemplated. Within the proximal COX-1 promoter, there are potential Sp1 binding sites at eC111 and at eC89,12 and Sp1 mediates estrogen responsiveness in the absence of ERE-like motifs in certain paradigms. Sp1 physically associates with ER, resulting in increased binding of Sp1 to its DNA site.13 With mutation of eC111Sp1, E2-induced promoter activity was attenuated by 43%, and it was decreased by over 70% with mutation of eC89Sp1 alone and by almost 90% with mutation of both eC111Sp1 and eC89Sp1. These findings indicate that putative Sp elements within the core COX-1 promoter are critical to E2-induced upregulation.
Identification of Nuclear Proteins Mediating COX-1 Response
To evaluate the nuclear proteins involved in binding to the proximal consensus Sp1 domains of the COX-1 promoter, electrophoretic mobility shift assays were performed with nuclei from E2-treated endothelial cells and oligonucleotide probes encompassing either eC111Sp1 or eC89Sp1. With eC111Sp1, one major DNA-protein complex was formed with wild-type probe that was less apparent with mutant probe. Because the identical mutation of eC2095COX-Luc yielded a modest decrease in E2 responsiveness, it suggests that the complex with eC111Sp1 plays a minor role in hormonal regulation of COX-1 expression. With eC89Sp1, a single major DNA-protein complex was formed with wild-type but not mutant probe, and the same mutation in eC2095COX-Luc prevented the majority of E2 responsiveness either alone or in combination with mutation of eC111Sp1. As such, the complex with eC89Sp1 most likely depicts the principal protein-DNA interaction mediating E2 induction of the COX-1 gene in endothelium.
To identify the nuclear proteins involved, supershift analyses were done. Antiserum to Sp1did not shift either the eC111Sp1 or the eC89Sp1 probe-nuclear protein complexes. Because Sp3 can interact with ER to mediate Sp1 responses,13 supershifts for Sp3 were also done that were negative. AP-2 was then considered because Sp1 and AP-2 have similar consensus binding sequences consisting of short GC-rich elements, and Sp1 and AP-2 bind to the same or overlapping regions within a variety of promoters.21eC24 Furthermore, AP-2 and AP-2 are both known to be involved in E2 modulation of gene expression.25 Antisera to AP-2 or AP-2 did not shift the eC111Sp1 probe-nuclear protein complex. However, antiserum to AP-2 uniquely caused supershifting of the eC89Sp1 probe-nuclear protein complex. Thus, AP-2 is likely to play an important role in the modulation of COX-1 gene expression by E2.
The participation of ER was also investigated. Antibody to ER supershifted the eC111Sp1 and the eC89Sp1 probe-nuclear protein complexes, indicating that ER is a component of both complexes. However, differences in overall complex formation were not observed in the absence versus presence of E2 treatment for 6 or 48 hours, there were only modest changes in the relative quantities of supershifted bands for the eC89Sp1 complex with ER antibody at 0 versus 6 hours of E2, and no changes in AP-2 recruitment were apparent. Thus, although the current approaches revealed important roles for the proximal Sp1-like elements and for ER and AP-2, the exact events activated by ER ligand binding are yet to be elucidated. Recent work using chromatin immunoprecipitation (ChIP) suggests that ER and coregulator association in gene promoters is temporally regulated. ChIP and cell imaging studies further show that members of the steroid receptor coactivator (SRC)-1 family and other coactivators may cycle on and off promoters with rapid kinetics.20,26 Further experiments are now indicated using strategies including ChIP to interrogate this unique promodulatory role of AP-2, which is collaborating with ER and other yet-to-be identified coactivators to target proximal regulatory elements on an important endothelial cell gene.
Features of ER-Mediating COX-1 Response
The features of ER required for COX-1 upregulation were also investigated. In studies of primary endothelial cells specifically selected to have minimal endogenous ER, the transient transfection of wild-type receptor caused COX-1 promoter responsiveness to E2 comparable to that observed at endogenous levels of ER expression. In contrast, a mutant form of ER lacking the DNA binding domain (ER185eC251) yielded a 4-fold greater E2 response than wild-type receptor. This finding indicates that E2 upregulation of COX-1 is independent of direct ER-DNA interaction, which is consistent with the involvement of Sp1-like elements and AP-2. In addition, a mutant form of ER lacking the two primary nuclear localization signals (ER250eC274) displayed 5-fold greater E2 responsiveness than wild-type receptor. Because ER250eC274 (HE257G) has attenuated nuclear localization in the absence or presence of E2,27 this finding suggests that the process may be initiated primarily by cytoplasmic ER, and that other mechanisms besides NLS must be responsible for nuclear import of the ER-related transcriptional machinery in this paradigm. Furthermore, we found that mutant ER lacking the A-B domains (ER1eC175) was incapable of inducing COX-1 promoter activity. This observation parallels the findings in prior work indicating that the AF-1 region within the A-B domains of ER is important for Sp1 activation.13,28 Similar domains within AF-1 may bind Sp1 and AP-2, and further detailed mutagenesis will be required to delineate these domains.
The ability of estrogen to upregulate COX-1 expression in endothelium through transcriptional transactivation represents a novel mode of regulation of a gene for which there are few means of dynamic regulation.29,30 Our collective observations indicate that this process is independent of direct ER-DNA binding, and that it instead entails protein-DNA interaction involving AP-2 and ER at proximal regulatory elements within the COX-1 promoter. In addition, cytoplasmic ER may play a key role in this process, and critical receptor elements reside within the amino terminus. As such, ER modulation of COX-1 entails mechanisms that are both shared and unique to the known processes whereby estrogen regulates a variety of gene targets.20,26 Further studies in this realm will increase our understanding of both COX-1 regulation and the potent, diverse capacities of estrogen to alter endothelial cell function.
Acknowledgments
This work was supported by National Institutes of Health Grants HL53546 (to P.W.S.), and HL50675 and NS23327 (to K.K.W.). Partial support was also provided by the Crystal Charity Ball Center for Pediatric Critical Care Research and the Lowe Foundation. We are indebted to Marilyn Dixon for preparing the manuscript.
Both authors contributed equally to this work.
References
Mendelsohn ME, Karas RH. The protective effects of estrogen on the cardiovascular system. N Engl J Med. 1999; 340: 1801eC1811.
Razandi M, Pedram A, Levin ER. Estrogen signals to the preservation of endothelial cell form and function. J Biol Chem. 2000; 275: 38540eC38546.
Carson-Jurica MA, Schrader WT, O‘Malley BW. Steroid receptor family: structure and functions. Endocr Rev. 1990; 11: 201eC220.
Chambliss KL, Yuhanna IS, Mineo C, Liu P, German Z, Sherman TS, Mendelsohn ME, Anderson RG, Shaul PW. Estrogen receptor alpha and endothelial nitric oxide synthase are organized into a functional signaling module in caveolae. Circ Res. 2000; 87: e44eCe52.
Chambliss KL, Yuhanna IS, Anderson RG, Mendelsohn ME, Shaul PW. ERbeta has nongenomic action in caveolae. Mol Endocrinol. 2002; 16: 938eC946.
Kleinert H, Wallerath T, Euchenhofer C, Ihrig-Biedert I, Li H, Forstermann U. Estrogens increase transcription of the human endothelial NO synthase gene: analysis of the transcription factors involved. Hypertension. 1998; 31: 582eC588.
Sumi D, Ignarro LJ. Estrogen-related receptor alpha 1 up-regulates endothelial nitric oxide synthase expression. Proc Natl Acad Sci U S A. 2003; 100: 14451eC14456.
Chambliss KL, Shaul PW. Estrogen modulation of endothelial nitric oxide synthase. Endo Rev. 2002; 23: 665eC686.
Orshal JM, Khalil RA. Gender, sex hormones, and vascular tone. Am J Physiol Regul Integr Comp Physiol. 2004; 286: R233eCR249.
Sherman TS, Chambliss KL, Gibson LL, Pace MC, Mendelsohn ME, Pfister SL, Shaul PW. Estrogen acutely activates prostacyclin synthesis in ovine fetal pulmonary artery endothelium. Am J Respir Cell Mol Biol. 2002; 26: 610eC616.
Jun SS, Chen Z, Pace MC, Shaul PW. Estrogen upregulates cyclooxygenase-1 gene expression in ovine fetal pulmonary artery endothelium. J Clin Invest. 1998; 102: 176eC183.
Xu XM, Tang JL, Chen X, Wang LH, Wu KK. Involvement of two Sp1 elements in basal endothelial prostaglandin H synthase-1 promoter activity. J Biol Chem. 1997; 272: 6943eC6950.
Sanchez R, Nguyen D, Rocha W, White JH, Mader S. Diversity in the mechanisms of gene regulation by estrogen receptors. Bioessays. 2002; 24: 244eC254.
Shaul PW, Wells LB. Oxygen modulates nitric oxide production selectively in fetal pulmonary endothelial cells. Am J Respir Cell Molec Biol. 1994; 11: 432eC438.
Shaul PW, Pace MC, Chen Z, Brannon TS. Developmental changes in prostacyclin synthesis are conserved in cultured pulmonary endothelium and vascular smooth muscle. Am J Respir Cell Molec Biol. 1999; 20: 113eC121.
German Z, Chambliss KL, Pace MC, Arnet UA, Lowenstein CJ, Shaul PW. Molecular basis of cell-specific endothelial nitric-oxide synthase expression in airway epithelium. J Biol Chem. 2000; 275: 8183eC8189.
Ow DW, Wood KV, DeLuca M, DeWet JR, Helinski DR, Howell SH. Transient and stable expresion of the firefly luciferase gene in plant cells and transgenic plants. Science. 1986; 234: 856eC859.
Rosenthal N. Identification of regulatory elements of cloned genes with functional assays. Methods Enzymol. 1987; 152: 704eC720.
Snedecor GW, Cochran WG. Statistical Methods. Ames. Iowa: Iowa State University Press; 1980.
Moggs JG, Orphanides G. Estrogen receptors: orchestrators of pleiotropic cellular responses. EMBO Rep. 2001; 2: 775eC781.
Descheemaeker KA, Wyns S, Nelles L, Auwerx J, Ny T, Collen D. Interaction of AP-1-, AP-2-, and Sp1-like proteins with two distinct sites in the upstream regulatory region of the plasminogen activator inhibitor-1 gene mediates the phorbol 12-myristate 13-acetate response. J Biol Chem. 1992; 267: 15086eC15091.
Piao YS, Peltoketo H, Vihko P, Vihko R. The proximal promoter region of the gene encoding human 17beta-hydroxysteroid dehydrogenase type 1 contains GATA, AP-2, and Sp1 response elements: analysis of promoter function in choriocarcinoma cells. Endocrinology. 1997; 138: 3417eC3425.
Chen TT, Wu RL, Castro-Munozledo F, Sun TT. Regulation of K3 keratin gene transcription by Sp1 and AP-2 in differentiating rabbit corneal epithelial cells. Mol Cell Biol. 1997; 17: 3056eC3064.
Ries S, Buchler C, Langmann T, Fehringer P, Aslanidis C, Schmitz G. Transcriptional regulation of lysosomal acid lipase in differentiating monocytes is mediated by transcription factors Sp1 and AP-2. J Lipid Res. 1998; 39: 2125eC2134.
Newman SP, Bates NP, Vernimmen D, Parker MG, Hurst HC. Cofactor competition between the ligand-bound oestrogen receptor and an intron 1 enhancer leads to oestrogen repression of ERBB2 expression in breast cancer. Oncogene. 2000; 19: 490eC497.
Smith CL, O’Malley BW. Coregulator function: a key to understanding tissue specificity of selective receptor modulators. Endocr Rev. 2004; 25: 45eC71.
Ylikomi T, Bocquel MT, Berry M, Gronemeyer H, Chambon P. Cooperation of proto-signals for nuclear accumulation of estrogen and progesterone receptors. EMBO J. 1992; 11: 3681eC3694.
Safe S. Transcriptional activation of genes by 17 beta-estradiol through estrogen receptor-Sp1 interactions. Vitam Horm. 2001; 62: 231eC252.
Smith WL. Molecular biology of prostanoid biosynthetic enzymes and receptors. Adv Exp Med Biol. 1997; 400: 989eC1011.
Wu KK. Regulation of prostaglandin H synthase-1 gene expression. Adv Exp Med Biol. 1997; 400: 121eC126.(Linda L. Gibson, Lisa Hah)
Abstract
Estrogen upregulates cyclooxygenase-1 (COX-1) expression in endothelial cells. To determine the basis of this process, studies were performed in ovine endothelial cells transfected with the human COX-1 promoter fused to luciferase. Estradiol (E2) caused activation of the COX-1 promoter with maximal stimulation at 10eC8 mol/L E2, and the response was mediated by either ER or ER. Mutagenesis revealed a primary role for a putative Sp1 binding motif at eC89 (relative to the ATG codon) and lesser involvement of a consensus Sp1 site at eC111. Electrophoretic mobility shift assays yielded a single complex with the site at eC89, and supershift analyses implicated AP-2 and ER, and not Sp1, in protein-DNA complex formation. In endothelial cells with minimal endogenous ER, the transfection of ER mutants lacking the DNA binding domain or primary nuclear localization signals caused 4-fold greater stimulation of promoter activity with E2 than wild-type ER. In contrast, mutant ER lacking the A-B domains was inactive. Thus, estrogen-mediated upregulation of COX-1 in endothelium is uniquely independent of direct ER-DNA binding and instead entails protein-DNA interaction involving AP-2 and ER at a proximal regulatory element. In addition, the process may be initiated by cytoplasmic ER, and critical receptor elements reside within the amino terminus.
Key Words: cyclooxygenase endothelium estrogen estrogen receptor
Introduction
Evidence from multiple paradigms indicates that estrogen has potent impact on endothelial cell function. The hormone activates signaling by endothelium, and it modifies endothelial cell growth, migration, and apoptosis. These actions are primarily mediated by estrogen receptor (ER) subtypes and , which classically function as transcription factors.1eC3 We have demonstrated that ER and ER are expressed in endothelium, and that subpopulations of ER and ER reside in endothelial caveolae/lipid rafts.4,5 One of the most thoroughly delineated targets of endothelial ER action is endothelial nitric oxide synthase (eNOS). eNOS transcription is upregulated by the activation of classical ER and estrogen-related receptor 1 in processes that involve diverse regulatory elements within the 5' flanking region of the eNOS gene.6,7 In addition, eNOS activity is rapidly stimulated by ER or ER ligand activation in endothelial caveolae.4,5 Through these multiple mechanisms estrogen enhances the production of the important atheroprotective molecule NO.8
In contrast to the detailed understanding of the basis of estrogen action on endothelial NO production, considerably less is known about the equally potent ability of the hormone to stimulate endothelial prostacyclin synthesis. This process is critical to the overall impact of estrogen on vascular health.9 We have previously demonstrated that estrogen acutely activates endothelial prostacyclin synthesis through an ER-dependent, calcium-dependent process.10 We have also shown that estrogen upregulates the expression of the rate-limiting enzyme in prostacyclin synthesis, cyclooxygenase type 1 (COX-1) in endothelial cells, resulting in enhanced basal and agonist-stimulated prostacyclin synthesis. We have further demonstrated that this is mediated by increases in steady-state COX-1 mRNA levels, which are not related to changes in mRNA degradation, thus implicating transcriptional processes.11 However, the molecular mechanisms by which COX-1 is upregulated by estrogen are unknown.
The purpose of the present study was to determine the processes by which COX-1 expression is upregulated in endothelium by estrogen. The 5' flanking sequence of the COX-1 gene lacks estrogen response elements (EREs) or ERE half-palindrome motifs.12 The hypothesis was therefore raised that estrogen activates COX-1 gene transcription via the involvement of putative Sp1 elements, which are prevalent in the COX-1 promoter and which mediate the estrogen responsiveness of certain genes.13 In addition to testing this hypothesis, experiments were performed to determine the role of ER and ER in this process. The specific cis DNA sequences involved in the estrogen response were also identified, along with nuclear proteins that are involved. Furthermore, studies were done to begin to determine the features of ER that play a role in mediating COX-1 expression.
Materials and Methods
Cell Culture
Experiments were performed in primary ovine endothelial cells obtained as previously described and studied at passage 4 to 6.14 These cells have been used previously in studies of COX-1 regulation.11,15 In experiments evaluating the actions of estrogen receptor mutants, primary endothelial cell cultures were specifically selected that had negligible expression of endogenous ER or ER by immunoblot analysis.
Construction of Reporter and Receptor Plasmids and Mutagenesis
The basis of estrogen-induced activation of COX-1 transcription was investigated using a 2075-base pair fragment (eC2095 to eC21 relative to the ATG codon, A as +1) of the human COX-1 5' flanking region inserted into KpnI/HindIII sites of the luciferase reporter gene plasmid, pGL2 (Promega Corp) to yield the full-length promoter-reporter plasmid denoted as eC2095COX-1-Luc. Two 5' deletion mutants were also studied including 1241 (eC1261 to eC21) or 117 (eC137 to eC21) base pairs of 5' flanking sequence of COX-1, designated eC1261COX-1-Luc and eC137COX-1-Luc, respectively. The COX-1 genomic sequences for these plasmids were obtained from COX-1 promoter reporter constructs originally made using pXP1.12 Site-directed mutants were also generated of the putative Sp1 element at eC111, the putative Sp1 element at eC89, or of both elements (see online data supplement available at http://circres.ahajournals.org).12
In studies of the role of estrogen receptors in COX-1 gene transcription, cells were cotransfected with cDNAs for wild-type human ER inserted into pCDNA3.14 or deletion mutants of ER. These included a mutant lacking the two primary NLS, NLS 2 and 3 (ER250eC274), a mutant lacking the DNA binding domain (ER 185eC251), and an N-terminal deletion mutant lacking residues 1 to 175 (ER1eC175) (see online data supplement and online Figure I). In selected experiments, cells were transfected with mouse ER cDNA cloned into pCDNA3.1.5
Cell Transfection and Reporter Activity
Cell transfection was performed using methods that have been previously reported.16 Twenty four hours later, the cells were placed in media containing charcoal-stripped serum and were treated with either vehicle, 17-estradiol (E2, 10eC12 to 10eC6 mol/L), or E2 plus ICI 182,780 (10eC5 mol/L) for 48 hour. The cells were lysed, extracts were centrifuged at 10 000g, and luciferase and -galactosidase activity were measured.17,18 The results are normalized as relative luciferase light units/-galactosidase activity. In selected wells, the cells were transfected with pGL2-Control Vector (Promega Corp) to serve as a positive control for luciferase expression. In preliminary studies, 17-estradiol had no effect on promoter activity. In experiments focused on the role of ER, sham plasmid or cDNAs for wild-type or mutant ER were cotransfected with the promoter-reporter constructs and pSV--Gal, and equal expression of ER forms was confirmed by immunoblot analysis.4
Electrophoretic Mobility Shift Assays
After 0 to 48 hours treatment of endothelial cells with vehicle or 10eC8 mol/L E2, nuclear extracts were prepared.16 Oligonucleotide probes were generated and added to incubations of nuclear extracts using methods that have been previously reported.16 In competition studies, excess wild-type, mutant, or unrelated oligonucleotides were added in 2- to 200-fold molar excess before the addition of the 32P-labeled probe. The wild-type eC111 Sp1 probe was 5'-GAGGGAGGAGCGGGGGTGGAGCCGGGGGAA-3' (upper strand) and the mutant probe was 5'-GAGGGAGGAGCGGTTTTAGAGCCGGGGGAA-3'. The wild-type eC89 Sp1 probe was 5'-GGGGGAAGGGTGGGGAGGGGATGGGCTGGA-3' (upper strand), and the mutant probe was 5'-GGGGGAAGGGTGTTTAAGGGATGGGCTGGA-3'.
To identify the nuclear proteins that bound to the putative Sp1 domains, supershifting of the DNA-protein complex was performed. For ER, 2 e蘈 of MAB461 (Chemicon International Inc) or unrelated IgG was added. For Sp1, Sp3, AP-2, and AP-2, 2 e蘈 of antiserum to the respective transcription factors was added. The antibodies were 2 e/e蘈 from Santa Cruz Biotechnology, Inc, and with the exception of Sp3, all were monoclonal. To confirm effective supershifting with antibody to Sp1, additional experiments were performed with an Sp1 probe from the eNOS promoter (5'-GGATAGGGGCGGGGCGAGG-3', upper strand).16 All nuclear protein-DNA complexes were resolved on 5% nondenaturing polyacrylamide gels containing 1x Tris borate/EDTA buffer. Dried gels were exposed to Kodak XAR film for autoradiograpy.
Statistical Analysis
Data for promoter activity were analyzed by ANOVA and Neuman-Keuls post hoc testing.19 Results are expressed as mean±SEM. All stated differences achieved statistical significance at the 0.05 level of probability or less.
Results
COX-1 Promoter Activation by Estrogen
To determine whether estrogen activates COX-1 transcription, the effects of the hormone on the activation of the full-length promoter (eC2095COX-1-Luc) were determined. The levels of basal promoter activity obtained with eC2095COX-1-Luc was at least 3-fold greater than with vector alone, and activity with vector alone was not altered by hormone treatment (data not shown). In cells transfected with eC2095COX-1-Luc, treatment with 10eC8 mol/L E2 for 48 hours caused a 2.6-fold increase in promoter activity that was fully reversed by concomitant treatment with ICI 182,780 (Figure 1A). In additional studies, the dose-response to E2 was evaluated (Figure 1B). Activation was observed at 10eC10 mol/L E2, and maximal response occurred with 10eC8 mol/L E2. Minimal stimulation of promoter activity was seen with 10eC6 mol/L E2, indicating a biphasic dose-response.
To further determine the role of ER in COX-1 gene modulation, experiments were done in endothelial cells overexpressing either ER or ER (Figure 1C). Relative to the responses seen with endogenous ER action in sham-transfected cells, there was a 2.3-fold greater E2-mediated promoter activity after ER overexpression. Similarly, there was a 2.2-fold enhancement of the E2 response with overexpression of ER. As such, both ER subtypes are capable of activating COX-1 transcriptional transactivation.
Promoter Elements Required for Regulation by Estrogen
To determine the elements in the COX-1 promoter required for estrogen responsiveness, the activities of 5' deletion mutants were evaluated in endothelial cells. Basal promoter activity fell by 63% with deletion from eC2095 to eC1261, and it fell a comparable 70% with deletion from eC2095 to eC137 (Figure 2A). However, the relative capacity for E2-stimulated promoter activity was conserved with deletion from eC2095 to eC1261, with the hormone causing similar 2.4-fold and 2.6-fold increases in activity over basal, respectively. In addition, there was continued conservation of E2 activation of the promoter with deletion to eC137, with eC137COX-1-Luc displaying a 2.8-fold response to hormone. Importantly, ER antagonism with ICI 182,780 reversed estrogen responses in all cases. To further implicate the proximal promoter in hormone action, additional experiments were done using eC2095COX-1-Luc and eC137COX-1-Luc in cells cotransfected with ER cDNA (Figure 2B). With ER overexpression, E2-stimulated activity of eC2095COX-1-Luc was raised by 2.4-fold. Similarly, with overexpression of the receptor the E2-induced activity of eC137COX-1-Luc increased by 2.2-fold.
Inspection of the proximal COX-1 promoter reveals a lack of estrogen response elements (ERE) or ERE half-palindrome motifs known to mediate the estrogen responsiveness of certain genes.20 However, there is a potential Sp1 binding site at eC111 and another at eC89,12 and Sp1 is known to mediate estrogen responsiveness in the absence of ERE-like motifs in certain paradigms.13 Because E2-induced activation of the promoter in endothelial cells was conserved with 5' deletions, which retained these domains, site-directed mutants of either or both of these putative Sp1 elements within eC2095COX-1-Luc were tested. First, the basal activity of mutated promoter sequences was evaluated (Figure 2C). Compared with full-length wild-type promoter (eC2095COX-1-Luc), mutation of the eC111 Sp1 caused a 58% decline in activity, mutation of the eC89 Sp1 yielded a 72% fall, and mutation of both elements caused a 53% decrease. Estrogen-stimulated promoter activity expressed relative to the level of basal activity for the same construct is shown in Figure 2D. Mutation of the Sp1 element at eC111 caused a decline in E2-induced activity of 43%, and mutation of eC89 Sp1 resulted in a 73% decline in the E2 response. Mutation of both domains yielded an 88% fall in E2-induced promoter activity.
Evaluation of Relevant Nuclear Proteins
To evaluate the nuclear proteins in endothelial cells involved in estrogen actions mediated by the consensus Sp1 elements of the COX-1 promoter at eC111 and at eC89, electrophoretic mobility shift assays were performed. Incubation of nuclear extracts from E2-treated endothelial cells (48 hours) with a double-stranded oligonucleotide probe encompassing the putative Sp1 site at eC111 resulted in the appearance of one major DNA-protein complex that was diminished by 200-fold molar excess of unlabeled probe (Figure 3A, left). In contrast, three minor complexes were formed with mutated eC111 Sp1 site probe. The use of mutated oligonucleotide as a competitor did not prevent the formation of the major complex by the wild-type probe (Figure 3A, right). Incubation of nuclear extracts with a probe containing the putative Sp1 site at eC89 resulted in the appearance of one major DNA-protein complex that were prevented by 200-fold molar excess of unlabeled probe (Figure 3B, left). The complex was minimally formed with mutated eC89 Sp1 DNA probe, and the mutated oligonucleotide only partially prevented the formation of the complex by the wild-type probe (Figure 3B, right).
To identify the endothelial nuclear proteins involved in complex formation with the presumptive eC111 Sp1 and the eC89 Sp1COX-1 promoter elements, supershift analyses were done. Two different antisera to Sp1 did not cause supershifting of either the eC111 Sp1 protein-DNA complex or the eC89 Sp1 protein-DNA complex (Figure 4A and 4B, left). To provide a control for supershifting of Sp1, a study was performed using the same endothelial nuclear proteins and the Sp1 element from the proximal eNOS promoter. With the Sp1 domain from eNOS, supershifting was observed as previously described (data not shown).16 In an effort to identify other potential nuclear proteins relevant to gene regulation by Sp1 elements, the effects of antisera to Sp3 were also assessed, and supershifting was not evident for either the eC111 Sp1 protein-DNA complex or the eC89 Sp1 protein-DNA complex (Figure 4A and 4B, middle). In studies of potential involvement of AP-2 or AP-2, which have consensus binding sequences similar to Sp1,21eC24 antisera to either protein did not cause supershifting of the eC111 Sp1 protein-DNA complex (Figure 4A, right). However, antisera to AP-2 caused supershifting of the eC89 Sp1 protein-DNA complex (Figure 4B, right). In experiments comparing nuclear proteins from endothelial cells treated with vehicle vs 10eC8 mol/L E2 for 48 hours, there was no difference in the overall formation of the eC111 Sp1 protein-DNA complex or the eC89 Sp1-protein DNA complex, and the supershifting of AP-2 in the eC89 Sp1 protein-DNA complex was also similar (data not shown).
The contribution of ER to complex formation was also assessed by supershift analyses. Using nuclear proteins from cells treated with E2 for 48 hours, antiserum to ER shifted both the eC111 Sp1 protein-DNA complex and the eC89 Sp1 protein-DNA complex, whereas an unrelated IgG did not (Figure 5A and 5B, left). In evaluations of the temporal features of ER involvement in complex formation, studies of nuclear proteins from cells treated with E2 for 0 versus 6 hours yielded comparable supershifting of ER from the eC111 Sp1 complex (Figure 5A, right). In contrast, the supershifted bands for ER in the eC89 Sp1 complex differed at 0 and 6 hours of E2 exposure; there was less prevalence of a rapidly migrating supershifted complex and greater abundance of a slowly-migrating supershifted complex after 6 hours of E2 treatment versus control (Figure 5B, right).
Features of ER Required for Modulation of COX-1 Transcription
There are multiple mechanisms by which ER regulates gene transcription.13,20 To begin to elucidate the features of the receptor necessary for modulation of COX-1, experiments were performed in a primary endothelial cell culture selected to have negligible endogenous ER or ER into which ER mutant constructs were cotransfected along with eC2095COX-1-Luc. In cells transfected with sham plasmid for wild-type ER, E2 caused no change in the activity of eC2095COX-1-Luc (Figure 6A). In contrast, cells transfected with wild-type ER displayed a 2.5-fold increase in promoter activity with E2. In simultaneous comparisons of wild-type ER and ER185eC251, E2 caused a 1.8-fold increase in promoter activity in cells expressing wild-type receptor and there was a 7-fold increase in activity with E2 in cells expressing ER185eC251 (Figure 6B). Additional independent studies revealed a 2.3-fold stimulation of promoter activity with E2 via wild-type ER and a markedly greater 9.2-fold stimulation in cells expressing ER250eC274 (Figure 6C). In contrast, the N-terminal truncation mutant of ER, ER1eC175, was incapable of E2-induced activation of the COX-1 promoter (Figure 6D).
Discussion
Estrogen causes upregulation of COX-1 expression in endothelial cells that is instrumental to the impact of the hormone on vascular function.9 We have previously demonstrated that this is mediated by an increase in steady-state levels of COX-1 mRNA, and that the degradation of the mRNA is unaltered.11 In the present study, we show that E2 activates COX-1 gene transcription in an ER-dependent manner. We further demonstrate that the process entails unique ERE-independent mechanisms, providing new insights into the complex nature of E2 actions in endothelium.
Role of ER and Regulatory Elements on COX-1 Gene
In transient transfection experiments in endothelial cells with a construct containing the human COX-1 promoter fused to luciferase, we have observed E2-induced promoter activation of sequence residing between eC2095 and eC21 of the 5' flanking sequence relative to the ATG codon. Experiments with ICI 182,780 indicated that the process is ER-dependent, and the dose-response was biphasic with a threshold concentration of 10eC10 mol/L and maximal activation at 10eC8 mol/L E2. The biphasic dose-response mimics that found previously for changes in COX-1 mRNA abundance with E2.11 Interestingly, the nongenomic activation of prostacyclin synthesis by E2 does not display a biphasic dose response,10 suggesting quite different kinetics for genomic versus nongenomic E2 modulation of prostacyclin synthesis in endothelial cells. Cotransfection studies further indicated that the COX-1 promoter response can be mediated by either ER or ER. Thus, COX-1 regulation by E2 occurs at a physiological concentration of the hormone8 and via either of the ER subtypes.
To determine the regulatory elements within the COX-1 promoter required for estrogen responsiveness, the activities of progressive 5' deletion mutants were evaluated. The capacity for E2-stimulated promoter activity was fully conserved with 5' deletion from eC2095 to eC1261 and also to eC137. In addition, the enhancement in E2 response with ER overexpression was similar for eC2095COX-1-Luc and eC137COX-1-Luc. These findings indicate that the estrogen responsiveness of the COX-1 gene resides within the proximal promoter.
Because the 5' flanking sequence of the COX-1 gene lacks EREs or ERE half-palindrome motifs, alternative modes of estrogen regulation were contemplated. Within the proximal COX-1 promoter, there are potential Sp1 binding sites at eC111 and at eC89,12 and Sp1 mediates estrogen responsiveness in the absence of ERE-like motifs in certain paradigms. Sp1 physically associates with ER, resulting in increased binding of Sp1 to its DNA site.13 With mutation of eC111Sp1, E2-induced promoter activity was attenuated by 43%, and it was decreased by over 70% with mutation of eC89Sp1 alone and by almost 90% with mutation of both eC111Sp1 and eC89Sp1. These findings indicate that putative Sp elements within the core COX-1 promoter are critical to E2-induced upregulation.
Identification of Nuclear Proteins Mediating COX-1 Response
To evaluate the nuclear proteins involved in binding to the proximal consensus Sp1 domains of the COX-1 promoter, electrophoretic mobility shift assays were performed with nuclei from E2-treated endothelial cells and oligonucleotide probes encompassing either eC111Sp1 or eC89Sp1. With eC111Sp1, one major DNA-protein complex was formed with wild-type probe that was less apparent with mutant probe. Because the identical mutation of eC2095COX-Luc yielded a modest decrease in E2 responsiveness, it suggests that the complex with eC111Sp1 plays a minor role in hormonal regulation of COX-1 expression. With eC89Sp1, a single major DNA-protein complex was formed with wild-type but not mutant probe, and the same mutation in eC2095COX-Luc prevented the majority of E2 responsiveness either alone or in combination with mutation of eC111Sp1. As such, the complex with eC89Sp1 most likely depicts the principal protein-DNA interaction mediating E2 induction of the COX-1 gene in endothelium.
To identify the nuclear proteins involved, supershift analyses were done. Antiserum to Sp1did not shift either the eC111Sp1 or the eC89Sp1 probe-nuclear protein complexes. Because Sp3 can interact with ER to mediate Sp1 responses,13 supershifts for Sp3 were also done that were negative. AP-2 was then considered because Sp1 and AP-2 have similar consensus binding sequences consisting of short GC-rich elements, and Sp1 and AP-2 bind to the same or overlapping regions within a variety of promoters.21eC24 Furthermore, AP-2 and AP-2 are both known to be involved in E2 modulation of gene expression.25 Antisera to AP-2 or AP-2 did not shift the eC111Sp1 probe-nuclear protein complex. However, antiserum to AP-2 uniquely caused supershifting of the eC89Sp1 probe-nuclear protein complex. Thus, AP-2 is likely to play an important role in the modulation of COX-1 gene expression by E2.
The participation of ER was also investigated. Antibody to ER supershifted the eC111Sp1 and the eC89Sp1 probe-nuclear protein complexes, indicating that ER is a component of both complexes. However, differences in overall complex formation were not observed in the absence versus presence of E2 treatment for 6 or 48 hours, there were only modest changes in the relative quantities of supershifted bands for the eC89Sp1 complex with ER antibody at 0 versus 6 hours of E2, and no changes in AP-2 recruitment were apparent. Thus, although the current approaches revealed important roles for the proximal Sp1-like elements and for ER and AP-2, the exact events activated by ER ligand binding are yet to be elucidated. Recent work using chromatin immunoprecipitation (ChIP) suggests that ER and coregulator association in gene promoters is temporally regulated. ChIP and cell imaging studies further show that members of the steroid receptor coactivator (SRC)-1 family and other coactivators may cycle on and off promoters with rapid kinetics.20,26 Further experiments are now indicated using strategies including ChIP to interrogate this unique promodulatory role of AP-2, which is collaborating with ER and other yet-to-be identified coactivators to target proximal regulatory elements on an important endothelial cell gene.
Features of ER-Mediating COX-1 Response
The features of ER required for COX-1 upregulation were also investigated. In studies of primary endothelial cells specifically selected to have minimal endogenous ER, the transient transfection of wild-type receptor caused COX-1 promoter responsiveness to E2 comparable to that observed at endogenous levels of ER expression. In contrast, a mutant form of ER lacking the DNA binding domain (ER185eC251) yielded a 4-fold greater E2 response than wild-type receptor. This finding indicates that E2 upregulation of COX-1 is independent of direct ER-DNA interaction, which is consistent with the involvement of Sp1-like elements and AP-2. In addition, a mutant form of ER lacking the two primary nuclear localization signals (ER250eC274) displayed 5-fold greater E2 responsiveness than wild-type receptor. Because ER250eC274 (HE257G) has attenuated nuclear localization in the absence or presence of E2,27 this finding suggests that the process may be initiated primarily by cytoplasmic ER, and that other mechanisms besides NLS must be responsible for nuclear import of the ER-related transcriptional machinery in this paradigm. Furthermore, we found that mutant ER lacking the A-B domains (ER1eC175) was incapable of inducing COX-1 promoter activity. This observation parallels the findings in prior work indicating that the AF-1 region within the A-B domains of ER is important for Sp1 activation.13,28 Similar domains within AF-1 may bind Sp1 and AP-2, and further detailed mutagenesis will be required to delineate these domains.
The ability of estrogen to upregulate COX-1 expression in endothelium through transcriptional transactivation represents a novel mode of regulation of a gene for which there are few means of dynamic regulation.29,30 Our collective observations indicate that this process is independent of direct ER-DNA binding, and that it instead entails protein-DNA interaction involving AP-2 and ER at proximal regulatory elements within the COX-1 promoter. In addition, cytoplasmic ER may play a key role in this process, and critical receptor elements reside within the amino terminus. As such, ER modulation of COX-1 entails mechanisms that are both shared and unique to the known processes whereby estrogen regulates a variety of gene targets.20,26 Further studies in this realm will increase our understanding of both COX-1 regulation and the potent, diverse capacities of estrogen to alter endothelial cell function.
Acknowledgments
This work was supported by National Institutes of Health Grants HL53546 (to P.W.S.), and HL50675 and NS23327 (to K.K.W.). Partial support was also provided by the Crystal Charity Ball Center for Pediatric Critical Care Research and the Lowe Foundation. We are indebted to Marilyn Dixon for preparing the manuscript.
Both authors contributed equally to this work.
References
Mendelsohn ME, Karas RH. The protective effects of estrogen on the cardiovascular system. N Engl J Med. 1999; 340: 1801eC1811.
Razandi M, Pedram A, Levin ER. Estrogen signals to the preservation of endothelial cell form and function. J Biol Chem. 2000; 275: 38540eC38546.
Carson-Jurica MA, Schrader WT, O‘Malley BW. Steroid receptor family: structure and functions. Endocr Rev. 1990; 11: 201eC220.
Chambliss KL, Yuhanna IS, Mineo C, Liu P, German Z, Sherman TS, Mendelsohn ME, Anderson RG, Shaul PW. Estrogen receptor alpha and endothelial nitric oxide synthase are organized into a functional signaling module in caveolae. Circ Res. 2000; 87: e44eCe52.
Chambliss KL, Yuhanna IS, Anderson RG, Mendelsohn ME, Shaul PW. ERbeta has nongenomic action in caveolae. Mol Endocrinol. 2002; 16: 938eC946.
Kleinert H, Wallerath T, Euchenhofer C, Ihrig-Biedert I, Li H, Forstermann U. Estrogens increase transcription of the human endothelial NO synthase gene: analysis of the transcription factors involved. Hypertension. 1998; 31: 582eC588.
Sumi D, Ignarro LJ. Estrogen-related receptor alpha 1 up-regulates endothelial nitric oxide synthase expression. Proc Natl Acad Sci U S A. 2003; 100: 14451eC14456.
Chambliss KL, Shaul PW. Estrogen modulation of endothelial nitric oxide synthase. Endo Rev. 2002; 23: 665eC686.
Orshal JM, Khalil RA. Gender, sex hormones, and vascular tone. Am J Physiol Regul Integr Comp Physiol. 2004; 286: R233eCR249.
Sherman TS, Chambliss KL, Gibson LL, Pace MC, Mendelsohn ME, Pfister SL, Shaul PW. Estrogen acutely activates prostacyclin synthesis in ovine fetal pulmonary artery endothelium. Am J Respir Cell Mol Biol. 2002; 26: 610eC616.
Jun SS, Chen Z, Pace MC, Shaul PW. Estrogen upregulates cyclooxygenase-1 gene expression in ovine fetal pulmonary artery endothelium. J Clin Invest. 1998; 102: 176eC183.
Xu XM, Tang JL, Chen X, Wang LH, Wu KK. Involvement of two Sp1 elements in basal endothelial prostaglandin H synthase-1 promoter activity. J Biol Chem. 1997; 272: 6943eC6950.
Sanchez R, Nguyen D, Rocha W, White JH, Mader S. Diversity in the mechanisms of gene regulation by estrogen receptors. Bioessays. 2002; 24: 244eC254.
Shaul PW, Wells LB. Oxygen modulates nitric oxide production selectively in fetal pulmonary endothelial cells. Am J Respir Cell Molec Biol. 1994; 11: 432eC438.
Shaul PW, Pace MC, Chen Z, Brannon TS. Developmental changes in prostacyclin synthesis are conserved in cultured pulmonary endothelium and vascular smooth muscle. Am J Respir Cell Molec Biol. 1999; 20: 113eC121.
German Z, Chambliss KL, Pace MC, Arnet UA, Lowenstein CJ, Shaul PW. Molecular basis of cell-specific endothelial nitric-oxide synthase expression in airway epithelium. J Biol Chem. 2000; 275: 8183eC8189.
Ow DW, Wood KV, DeLuca M, DeWet JR, Helinski DR, Howell SH. Transient and stable expresion of the firefly luciferase gene in plant cells and transgenic plants. Science. 1986; 234: 856eC859.
Rosenthal N. Identification of regulatory elements of cloned genes with functional assays. Methods Enzymol. 1987; 152: 704eC720.
Snedecor GW, Cochran WG. Statistical Methods. Ames. Iowa: Iowa State University Press; 1980.
Moggs JG, Orphanides G. Estrogen receptors: orchestrators of pleiotropic cellular responses. EMBO Rep. 2001; 2: 775eC781.
Descheemaeker KA, Wyns S, Nelles L, Auwerx J, Ny T, Collen D. Interaction of AP-1-, AP-2-, and Sp1-like proteins with two distinct sites in the upstream regulatory region of the plasminogen activator inhibitor-1 gene mediates the phorbol 12-myristate 13-acetate response. J Biol Chem. 1992; 267: 15086eC15091.
Piao YS, Peltoketo H, Vihko P, Vihko R. The proximal promoter region of the gene encoding human 17beta-hydroxysteroid dehydrogenase type 1 contains GATA, AP-2, and Sp1 response elements: analysis of promoter function in choriocarcinoma cells. Endocrinology. 1997; 138: 3417eC3425.
Chen TT, Wu RL, Castro-Munozledo F, Sun TT. Regulation of K3 keratin gene transcription by Sp1 and AP-2 in differentiating rabbit corneal epithelial cells. Mol Cell Biol. 1997; 17: 3056eC3064.
Ries S, Buchler C, Langmann T, Fehringer P, Aslanidis C, Schmitz G. Transcriptional regulation of lysosomal acid lipase in differentiating monocytes is mediated by transcription factors Sp1 and AP-2. J Lipid Res. 1998; 39: 2125eC2134.
Newman SP, Bates NP, Vernimmen D, Parker MG, Hurst HC. Cofactor competition between the ligand-bound oestrogen receptor and an intron 1 enhancer leads to oestrogen repression of ERBB2 expression in breast cancer. Oncogene. 2000; 19: 490eC497.
Smith CL, O’Malley BW. Coregulator function: a key to understanding tissue specificity of selective receptor modulators. Endocr Rev. 2004; 25: 45eC71.
Ylikomi T, Bocquel MT, Berry M, Gronemeyer H, Chambon P. Cooperation of proto-signals for nuclear accumulation of estrogen and progesterone receptors. EMBO J. 1992; 11: 3681eC3694.
Safe S. Transcriptional activation of genes by 17 beta-estradiol through estrogen receptor-Sp1 interactions. Vitam Horm. 2001; 62: 231eC252.
Smith WL. Molecular biology of prostanoid biosynthetic enzymes and receptors. Adv Exp Med Biol. 1997; 400: 989eC1011.
Wu KK. Regulation of prostaglandin H synthase-1 gene expression. Adv Exp Med Biol. 1997; 400: 121eC126.(Linda L. Gibson, Lisa Hah)