Figures
Abstract
Vascular effects of estradiol are being investigated because there are controversies among clinical and experimental studies. DNA microarrays were used to investigate global gene expression patterns in cultured human umbilical vein endothelial cells (HUVEC) exposed to 1 nmol/L estradiol for 24 hours. When compared to control, 187 genes were identified as differentially expressed with 1.9-fold change threshold. Supervised principal component analysis and hierarchical cluster analysis revealed the differences between control and estradiol-treated samples. Physiological concentrations of estradiol are sufficient to elicit significant changes in HUVEC gene expression. Notch signaling, actin cytoskeleton signaling, pentose phosphate pathway, axonal guidance signaling and integrin signaling were the top-five canonical pathways significantly regulated by estrogen. A total of 26 regulatory networks were identified as estrogen responsive. Microarray data were confirmed by quantitative RT-PCR in cardiovascular meaning genes; cyclooxigenase (COX)1, dimethylarginine dimethylaminohydrolase (DDAH)2, phospholipase A2 group IV (PLA2G4) B, and 7-dehydrocholesterol reductase were up-regulated by estradiol in a dose-dependent and estrogen receptor-dependent way, whereas COX2, DDAH1 and PLA2G4A remained unaltered. Moreover, estradiol-induced COX1 gene expression resulted in increased COX1 protein content and enhanced prostacyclin production. DDAH2 protein content was also increased, which in turn decreased asymmetric dimethylarginine concentration and increased NO release. All stimulated effects of estradiol on gene and protein expression were estrogen receptor-dependent, since were abolished in the presence of the estrogen receptor antagonist ICI 182780. This study identifies new vascular mechanisms of action by which estradiol may contribute to a wide range of biological processes.
Citation: Sobrino A, Mata M, Laguna-Fernandez A, Novella S, Oviedo PJ, García-Pérez MA, et al. (2009) Estradiol Stimulates Vasodilatory and Metabolic Pathways in Cultured Human Endothelial Cells. PLoS ONE 4(12): e8242. https://doi.org/10.1371/journal.pone.0008242
Editor: Patrick Callaerts, Katholieke Universiteit Leuven, Belgium
Received: June 18, 2009; Accepted: November 15, 2009; Published: December 14, 2009
Copyright: © 2009 Sobrino et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by grants FIS 06/0589 and Red HERACLES RD06/0009/0005 from the Ministerio de Ciencia e Innovación, Instituto de Salud Carlos III, Madrid, Spain. Pilar J Oviedo holds a post-doc position and Agua Sobrino is a fellowship recipient from the FPI program (BFPI 06/145), both from Conselleria de Educación, Generalitat Valenciana, Spain. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The incidence of coronary heart disease is greater in men than in premenopausal women of the same age, but increases in frequency after menopause, an effect that has been attributed, at least in part, to estrogens [1]. Estrogens have been used as contraceptive agents or as principal constituents of hormone replacement therapy formulations in postmenopausal women, 17β-estradiol being the most widely used molecule. The cardiovascular protective effect detected in a considerable number of observational clinical studies [2] has not been confirmed by more recent randomized placebo-controlled trials designed to study the effects of hormonal therapy in either secondary [3], [4] or primary [5] prevention. It should be stated that the clinical trials of estrogen therapy for the treatment of cardiovascular disease are largely flawed (e.g., hormone replacement therapy started too late in menopause). Moreover, a number of studies have demonstrated a favorable profile for estrogens in both experimental animal as well as in vitro models [6].
Endothelium is crucial to the modulation of vessel tone and to the control of platelet adhesion and aggregation, two key factors in the initiation and development of atherosclerosis [7]. Endothelium, including human umbilical vein endothelial cells (HUVEC), expresses both types of estrogen receptors (ER), α and β, and the actions of estrogens on endothelium have been exhaustively studied [8]. Moreover, clinical and experimental data support the consideration of endothelium as a target for sexual hormones [9]. Estradiol effects on partial gene expression in endothelial cells have frequently been studied, but there is a lack about its effects on the whole gene expression profile.
Microarrays are high-throughput genomic tools that allow the comparison of global expression changes in thousands of genes between different experimental conditions in cell/tissue analysis, and they have been widely adopted for analyzing the global gene expression profiles in vivo and in vitro [10]. Recent studies have demonstrated the ability of this technology for investigating molecular pathophysiological mechanisms involved in a variety of human diseases. For instance, microarray technology has been used as a novel experimental approach to analyze alterations in gene expression in different cardiovascular diseases [11], atherosclerosis [12] and experimental stroke in rats [13].
Microarray technology offers the possibility of exploring a large number of candidate genes which are modified by estrogens. The present study aims to explore gene expression modification, mainly focused on candidate genes that may regulate the vascular effects of estrogens, in cultured human endothelial cells exposed to physiological concentrations of estradiol by using microarrays, thus providing new information to the available body of knowledge about the influence of estradiol on the vascular wall.
Materials and Methods
Ethics Statement
This investigation conforms to the principles outlined in the Declaration of Helsinki, was approved by the Ethical Committee of Clinical Research of the University Clinical Hospital of Valencia, and written informed consent was obtained from all donors.
Cell Culture and Experimental Design
Primary HUVEC were isolated, grown, and identified as described earlier [14] in human endothelial cell-specific Medium EBM-2 (Lonza, Basel, Switzerland), supplemented with EGM-2 (Lonza), in an incubator at 37°C with 5% CO2.
Cells from passages 4 to 6 were seeded onto 25 cm2 flasks for mRNA isolation. When cells were at 75% of confluence, culture medium was exchanged for a phenol red–free Medium 199 (GIBCO, Invitrogen, Barcelona, Spain) supplemented with 20% charcoal/dextran-treated fetal bovine serum (GIBCO), EGM-2, pyruvic acid and antibiotics (“hormone free medium”) to avoid any estrogenic activity and maintained for 24 hours. Then, culture medium was eliminated and replaced by phenol red-free medium 199.
Cells were exposed to different concentrations (range: 0,01 – 100 nmol/L) of 17β-estradiol (Sigma, Alcobendas, Spain) by serial dilutions of a stock solution with phenol red-free medium. The pure anti-estrogen ICI182780 (1 µmol/L; Biogen, Madrid, Spain) was used to evaluate whether the observed effects were mediated by ER modulation. Control cells were exposed to the same vehicles of estradiol (0.1% ethanol) or ICI182780 (0.1% DMSO). All treatments were added in hormone free medium and experiments were performed at 75–80 % of confluence.
RNA Isolation and Genechip Expression Analysis
To carry out the microarray experiments, HUVEC from 9 separate cultures were exposed to control (0.1% ethanol) and 1 nmol/L estradiol treatments for 24 h. Total cellular RNA was extracted by using the TRIzol® reagent (Invitrogen, USA) following the manufacturer's instructions. RNA integrity was assayed by means of the 2100 Bioanalizer (Agilent Technologies, Santa Clara, CA, USA). Equal amounts of RNA extracted from 3 control- or 3 estradiol-treated cultured flasks obtained from three different cultures were pooled, achieving three biological replicates of the control and three that were treated with estradiol. Therefore, a total number of 6 microarrays were developed (3 control pools, named C1, C2, C3, and 3 estradiol-treated pools named E1, E2 and E3). Five micrograms of total RNA were amplified and labeled according to GeneChip Expression Analysis Technical Manual (Affymetrix Ltd, UK). The concentration of biotinylated and fragmented cRNA was measured using the 2100 Bioanalizer (Agilent Technologies). Twenty micrograms of fragmented biotinylated cRNA were used to prepare the hybridization cocktail and subsequently hybridized for sixteen hours to the Human Genome U133A plus 2.0 microarrays, which analyzes the expression level of over 47000 transcripts and variants. Arrays were washed and stained according to the EukGene_ws_2v5 in the Fluidics Station (Affymetrix) and scanned using the GeneChip scanner 3000. Affymetrix's GeneChip Operating Software (GCOS, Affymetrix) was used to obtain and analyze images.
Files obtained from GCOS (.cel) were used to analyze significant changes in expression profiles of different experimental groups using the dCHIP Analysis Software and the SpotFire Decision Site software. Data were normalized using the Invariant Set Method described earlier [15] and modeled using the PM/MM model. Then, ANOVA was used to find significant changes among experimental groups. False Discovery Rate (FDR) was used to discriminate false positives in the multivariant system. Only adjusted p-values <0.05 were considered significant. Global differences between different samples were measured by Principal Component Analysis (PCA) and Linear Discrimination Analysis (LDA). Hierarchical Cluster was used to analyze expression profiles of different samples, and was carried out using UPGMA (Unweighted Pair Group Method with Arithmetic Mean) analysis, with an ordering function based on the input rank. Data are represented as a dendrogram, with the closest branches of the tree representing arrays with similar gene expression patterns. Gene Onthology Browser (Nettaffyx Analysis Center, Affymetrix) was used to classify genes according to functionality context. Finally, relationships among data were screened using the Pathway Architect software (Stratagene, La Joya, CA, USA). All data discussed in this publication is MIAME compliant and that the raw data has been deposited in NCBI's Gene Expression Omnibus [16], a MIAME compliant database, and are accessible through GEO Series accession number GSE16683.
Network Identification and Canonical Pathway Analysis
List of genes significantly regulated by estrogen were analyzed using Ingenuity Pathways Analysis (IPA) software (Ingenuity Systems, Redwood City, CA, USA). IPA uses a variety of computational algorithms to identify and establish cellular networks that statistically fit the input gene list and expression values from experiments. Data sets containing the Affymetrix probe set identifiers and fold changes of genes were overlaid onto a global molecular network developed from information contained in the database. Networks were then algorithmically generated based on their connectivity and a score was assigned. The score is used to rank networks according to how relevant they are to the genes in the input dataset. Each network or pathway was arbitrarily set to have a maximum of 35 focus genes. Genes or gene products are represented as nodes, and the biological relationship between two nodes is represented as an edge. The intensity of the node color indicates the degree of up- (red) or down- (green) regulation.
Canonical pathways analysis identified the pathways, which were most significant to the input data set. The significance of the association between the data set and the canonical pathway was determined based on two parameters: (1) a ratio of the number of genes from the data set that map to the pathway divided by the total number of genes that map to the canonical pathway and (2) a P value calculated using Fischer's exact test determining the probability that the association between the genes in the data set and the canonical pathway is due to chance alone.
Quantitative Real Time PCR (QRT-PCR) Assays
Reverse transcription (RT) was carried out using SuperScriptTM II Synthesis System for RT–PCR (Invitrogen) by using a personal Mastercycler Eppendorf Thermocycler (Eppendorf, Hamburg, Germany). Different samples than that used for microarrays experiments were used to perform the QRT-PCR assays. One microgram of total RNA was reverse-transcribed to cDNA following the manufacturer's instructions. For each RT, a blank was prepared using all the reagents except the RNA sample (replaced with an equivalent volume of diethylpyrocarbonate (DEPC)-treated water) and also used as non-template control in real-time PCR experiments.
Quantitative real-time PCR (QRT-PCR) was done with SYBR-Green PCR Master Mix or TaqMan Universal Mastermix (Applied Biosystems, Fosters City, CA, USA). In the case of DHCRA7, PLA2G4A and PLA2G4B, the PCR reaction mix was prepared in 0.2 mL RNase free tubes by adding a volume of TaqMan Universal PCR Master Mix and TaqMan Gene Expression Assay (Table 1). The sample of cDNA obtained from the RT was incorporated into the necessary quantity of DEPC-treated water to get a final concentration of 40 ng approximately (range: 10–100 ng). The GADPH gene was used as endogenous control. The appropriate volume of each reaction mixture was transferred to a reaction plate which was then placed in the 7900HT Fast Real-Time PCR System (Applied Biosystems) with the appropriate thermal cycling conditions (50°C/2 min, 95°C/10 min, 40 Cycles; 95°C/15 s, 6°C/1 min).
In the case of COX-1, COX-2, DDAH-1 and DDAH-2, a QRT-PCR was performed using an ABI PRISM 7700 Sequence Detection System (Applied Biosystems) with a heated lid (105°C), an initial denaturation step at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. To amplify cDNA, the RT samples were diluted 1/10. In each reaction, a total of 1 µL from each RT tube was mixed with 12.5 µL of SYBR Green PCR master mix (Applied Biosystems) containing nucleotides, Taq DNA polymerase, MgCl2 and reaction buffer with SYBR green; 1.5 µL of 5 µmol/L adequate primers and DEPC-treated water were added to a final volume of 25 µL. In parallel, 5-fold serial dilutions of well-known DNA concentrations were run as calibration curves. Primers (Table 1) were designed using the Primers Express Software (Applied Biosystems) and synthesized by Custom Primers (Life Technologies, Barcelona, Spain).
Data were analysed with the ABI PRISM Sequence Detection v. 1.7 analysis software (Perkin Elmer, Nieuwerkerk, The Netherlands). To validate a QRT-PCR, standard curves with r>0.95 and slope values between −3.1 and −3.4 were required. Gene expression was relative quantified based on the work of Pfaffl [17]. In some samples, PCR bands were purified using a MiniElute PCR Purification Kit (Qiagen, Valencia, CA, USA) and then sequenced to prove that the amplified products corresponded to previously published sequences. Agarose gel electrophoreses were also performed to demonstrate that QRT-PCR yielded a unique band.
Immunoblotting
HUVEC were treated in 25 cm2 flasks for 24 hours with the desired products. A volume of 150 µL of lysis buffer (0.1 % triton X-100, 0.5 % sodium deoxicholate acid, 0.1 % Sodium Dodecyl Sulphate (SDS), 0.1% phenylmethanesulphonylfluoride or phenylmethylsulphonyl fluoride (PMSF), in 100 mL of phosphate saline buffer (PBS) containing protease inhibitors: 1 µg/mL leupeptin, 0.5 µg/mL pepstatin and 1 µg/mL bestatin) was added and maintained at 4°C for 30 minutes. Then, cells were collected using a cell scraper, boiled for 5 minutes and sonicated for 10 seconds. Protein content was measured [18] and samples were frozen at –20°C until assay.
Equal amounts of protein (60–80 µg) were then separated by 10% of SDS-Polyacrylamide gel electrophoresis, and the protein was transferred to PVDF sheets (Biorad, Spain). Immunostaining was achieved using specific antibodies anti-ERα (sc-8002; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-ERβ (sc-8974; Santa Cruz Biotechnology), anti-COX-1 (cat 236003; Calbiochem, Germany), anti-COX-2 (cat 160107; Cayman Chemical), anti-DDAH-I (PC716; Calbiochem) or anti-DDAH-II (PC717; Calbiochem). Development was performed with alkaline-phosphatase-linked appropriate secondary antibodies (from Sigma), followed with nitroblue tetrazolium (NBT)/5-Bromo-4-Chloro-3-Indolyl Phosphate, p-toluidine salt (BCIP) color development reaction. Blots were digitalized using a Gelprinter PLUS (TDI, Madrid, Spain), and the densities of spots were analyzed with the program Image Gauge 4.0 (Science Lab. 2001). Equivalent protein loading and transfer efficiency were verified by staining for β-actin (Sigma).
Prostacyclin Assay
After treatment with the desired products, medium was collected and stored at –20°C until prostacyclin was measured. Culture wells were then washed with PBS and adherent cells were collected in 0.5 N NaOH for protein determination by the modified Lowry's method using bovine serum albumin as standard [18].
The amount of prostacyclin produced, calculated as the concentration of stable hydrolysis product, 6-keto-prostaglandin-F1α, was assessed in duplicate by a commercial EIA kit (Cayman Chemical). Prostacyclin production was expressed as ng prostacyclin/mg protein.
Isolation and Measurement of Asymmetric Dimethylarginine (ADMA)
After 24 hours of incubation with the desired treatments, medium was collected and stored at –20°C until asymmetric dimethylarginine (ADMA), a major endogenous inhibitor of nitric oxide synthase (NOS), quantification. Culture wells were then washed with PBS and adherent cells were collected in 0.5 N NaOH solution for protein determination [18].
Measurement of ADMA was accomplished by high-performance liquid chromatography (HPLC) as described earlier [19]. In brief, ADMA from 1 mL of culture medium was purified with Bond Elut SCX columns (Varian Inc., Palo Alto, CA, USA) and eluted with 4 mL of methanol containing 30% distilled water and 2% triethylamine. The eluent was then evaporated to dryness at 60°C, and the dried extract was redissolved in running buffer. HPLC was carried out on a Shimadzu chromatography system (Shimadzu Corporation, Kyoto, Japan). Separation of ADMA was achieved with a 250×4.6-mm (inner diameter), 5-µm “Kromasil C18” analytical column (Scharlau, Barcelona, Spain) using 25 mM phosphoric acid containing 10 mM hexane sulphonic acid and 1 % [v/v] acetonitrile, pH 5.0. The analysis was carried out at a flow rate of 1.3 mL/min and the absorbance monitored at 200 nm. Concentrations of ADMA in the samples were determined by comparison with standards (Sigma, Alcobendas, Spain). ADMA production was expressed as nmol/mg protein.
Nitric Oxide (NO) Production
After 24 hours of incubation with the desired treatments, cells were washed and incubated with HEPES buffer (5 mM HEPES containing (in mM) 140 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2 and 10 glucose, pH adjusted to 7.4), for 120 min. Then, incubation medium was collected, culture wells were washed with PBS, and adherent cells were collected in 0.5 N NaOH solution for protein determination [18].
Endothelial NO production was determined in culture medium using the ISONOP nitric oxide sensor (World Precision Instruments, Sarasota, FL, USA), an amperometric sensor specific for NO, as described earlier [19]. A chemical titration calibration was performed with use of an acidic iodide solution (0.1 mol/l H2SO4, 0.14 mol/l K2SO4, 0.1 mol/l KI) against varied volumes of KNO2. NO was formed stoichiometrically and measured directly. The quantity of NO was expressed as nmol/mg protein.
Statistical Analysis
Values shown in the text and figures are mean ± SEM. ANOVA test was applied for comparisons of mean, and then Bonferroni's test was performed. P values<0.05 were considered significant. The statistical analysis was carried out using the Prism 4 software (GraphPad Software Inc., San Diego, CA, USA).
Results
Identification of Global Gene Expression Changes in Estradiol–Treated HUVEC
The gene expression profile of human vascular cells treated with or without estradiol was assessed by using the Human Genome U133A plus 2.0 microarray technology from Affymetrix. A total of 1886 genes passed the ANOVA analysis, with fold changes between 2.99 and −5.34. Table 2 (online supporting information) summarizes most differentially expressed genes between control and estradiol treated samples. Only genes with more than a 1.9-fold change were included. As expected, the list of genes became greater as a more permissive fold-change was selected. Only 4 genes (∼14%) were up-regulated and 25 (∼86%) were down-regulated when the fold-change was higher than 2.5. By decreasing the fold-change, the number of genes regulated by estradiol increased, and there was a tendency to equate the percentage of genes up-regulated and down-regulated. For instance, with a fold-change higher than 1.9, 187 genes were significantly regulated: 95 (∼50%) were up-regulated and 92 (∼50%) were down-regulated.
Comparison of Gene Expression Profiles by Hierarchical Clusters and Principal Component Analysis (PCA)
Hierarchical Clusters were used to analyze the expression profile of the different samples (Figure 1). Results identified broad similarities among arrays hybridized with the mRNA of control cells or among arrays hybridized with mRNA of cells treated with estradiol. Even though the overall signal patterns found on the mRNA hybridized arrays were similar, a small subset of regions show differential expression signals between the mRNA of control cells and mRNA of cells treated with estradiol.
263 probe sets of genes significantly regulated by greater than 1.8-fold change were used for 2D hierarchical clustering. Each row represents an individual probe set, and each column represents a pool of cells (C1, C2 and C3: control samples; E1, E2 and E3: estradiol-treated samples). 129 up- (red) or 134 down- (green) were regulated (P value<0.05).
PCA was applied to establish the interrelationships among the samples used in our study. By visualizing projections of these components in low-dimensional spaces, samples were grouped, reflecting underlying patterns in their gene expression profiles. Figure 2 depicts the PCA with the six pools clearly separated into two sets, one set with three control samples, and the other set with three estradiol-treated samples.
Microarray hybridizations were performed using total RNA from HUVEC exposed to 1 nmol/L estradiol for 24 h. The gene expression profiles of 3 pools of control cells (blue) and 3 pools of cells treated with estradiol 1 nmol/L (red) were compared using PCA. The three-dimensional (3D) plot view of gene expression data (including all probe sets on U133 Plus 2.0 GeneChip) is shown, with respect to their correlation to the first three principal components.
Functional Categorization of Genes
HUVEC genes regulated by estradiol were organized by function to better understand their profile. This classification showed that estradiol regulated a great number of genes mainly associated with biological processes that include Cellular Growth and Proliferation; Cell-to-cell Signaling; Cellular Assembly and Organization; Cellular Compromise; Cellular Movement and Cell Death, as shown in Table 3 (online supporting information). The Cardiovascular System Development and Function also appears as one of the main regulated. Genes with a role in Lipid and Carbohydrate Metabolism, Cell Signaling, Endocrine System Disorders or Metabolic Disease appear to be significantly regulated by estrogens as well. Among these biological processes, the greater part of molecules induced by estradiol in HUVEC is related to growth of cells (47 molecules), cell death (47 molecules) and apoptosis (39 molecules), cell movement (32 molecules), growth of eukaryotic cells (25 molecules) adhesion cells (22 molecules), colony formation (16 molecules), development of blood vessels (15 molecules), cell surface receptor linked signal transduction (14 molecules) and angiogenesis (10 molecules).
The functional characterization of data are presented in Figure 3, which lists top ten canonical pathways regulated by estrogen across both tissue types. The top five canonical pathways based on their significance (P value) included Notch Signaling, Actin Cytoskeleton Signaling, Pentose Phosphate Pathway, Axonal Guidance Signaling and Integrin Signaling. Genes included in each group of the top ten signaling pathways presented in Figure 3 are listed in Table 4.
For the functional categorization of genes, Fischer's exact test was used to calculate a p value (shown as bars) determining the probability that each biological function assigned to the network is due to chance alone. The ratio (shown as squares) represents the number of differentially expressed genes in a given pathway divided by total number of genes that make up that canonical pathway.
When the IPA software was used to analyze the probe sets there were 26 significant regulatory networks (score>2), of which 5 were highly significant (score>20). The number one ranked network (score = 62, focus molecules = 33) (Figure 4) is associated with Cardiovascular System Development and Function, Cellular Growth and Proliferation and Cell Morphology. Transforming Growth Factor beta-1 (TGFB1) plays a central role in the formation of this network. Top functions of the other four highly significant networks are associated with Cellular Compromise, Cellular Movement, Hematological System Development and Function, Lipid Metabolism, Molecular Transport and Small Molecule Biochemistry.
Networks of genes were algorithmically generated with the IPA software based on their connectivity and assigned a score. The intensity of the node color indicates the degree of up- (red) or down- (green) regulation. A continuous line means a direct relationship between the two genes, whereas a discontinuous line indicates an indirect association. The most significant network regulated by estradiol includes 33 genes with an assigned score of 62 and is centered on TGFB1.
Microarray Analysis Verification
To validate the HUVEC gene expression changes induced by estradiol in the microarray analysis, QRT-PCR was performed in a separate series of follow-up studies of HUVEC exposed to different concentrations (0,01–100 nmol/L) of estradiol. Target genes were selected based upon their important cardiovascular functions and were genes encoding for DDAH1, DDAH2, PLA2G4A, PLA2G4B, COX1, COX2 and DHCR7.
Estradiol dose-dependent increased mRNA expression of COX1, DDAH2, PLA2G4B, and DHCR7 (Figure 5). In all cases, the effect afforded by 1 nmol/L estradiol was significantly higher than that of 0,01 nmol/L estradiol. There were no differences between the effects on gene expression induced by the higher tested concentrations (1, 10 and 100 nmol/L), although in the case of DHCR7 the effect of 10 nmol/L was 34% higher than that of 1 nmol/l. The increased gene expressions induced by 1 nmol/L estradiol were similar to change levels obtained in the microarray analysis (probeset 205128_x_at for COX1, - fold change of 1.56 -p = 0.007-, probeset 202262_x_at for DDAH2 -fold change: 1.37, p = 0.045-, probeset 219095_at for PLA2G4B -fold change 1.56, p = 0.019-, and probeset 201790_s_at for DHCR7 -fold change 1.79, p = 0.005-).
HUVEC were exposed to different estradiol concentrations (0,01–100 nmol/L), and to 1 µmol/L ICI182780 alone or plus 1 nmol/L estradiol, and the relative expression of the genes was quantified: (A) COX1, (B) DDAH2, (C) PLA2G4B, and (D) DHCR7. Data are percentage of control values and are mean ± SEM of 5–19 values (4–6 different experiments). * p<0.05, ** p<0.01 or *** p<0.001 vs. control cells, † p<0.05 vs. 0.01 nmol/L estradiol, and ‡ p<0.05 vs. 1 nmol/L estradiol.
The mRNA expression of COX2, DDAH1 and PLA2G4A (Figure 6) remained unaltered under all the estradiol concentrations, as in the microarray analysis (probeset 204748_at for COX2 -fold change: −1.18, p = 0.541-, probeset 209094_at for DDAH1-fold change: −1.03, p = 0.743-, and probeset 210145_at for PLA2G4A -fold change −1.06, p = 0.570-).
HUVEC were exposed to different estradiol concentrations (0,01–100 nmol/L), and to 1 µmol/L ICI182780 alone or plus 1 nmol/L estradiol, for 24 hours. Total cellular RNA was extracted, and the relative expression of the genes was quantified using specific primers for (A) COX2, (B) DDAH1 and (C) PLA2G4A. The GADPH gene was used as control as described in Materials and Methods. Data are expressed as percentage of control values and are mean ± SEM of 5–17 values corresponding to 5 different experiments.
Estradiol genomic effects are mainly mediated through ERα and ERβ. HUVEC express both types of ER (Figure 7), and no changes in protein expression of both types of ER were observed after exposure to estradiol, ICI 182780, or estradiol plus ICI182780 (Figure 7). To study the role of ER on the effects induced by estradiol on gene expression, cells were exposed to the nonselective ER antagonist ICI182780. In cells exposed to different concentrations (0,01 – 10 µmol/L) of ICI 182780 alone, expression of the seven selected genes remained unaltered (Table 5), thus discarding a direct effect of ICI 182780 on gene profile. In cells coexposed to 1 nmol/L estradiol plus 1 µmol/L ICI182780 for 24 h, estradiol-induced effects on gene expression were completely abolished (Figure 5).
Cells were exposed to 1 nmol/L estradiol with or without 1 µmol/L ICI182780 for 24 hours, and protein expression of (A) ERα and (B) ERβ were measured as stated in Materials and methods. A typical immunoblotting image and relative levels assessed by densitometry of bands of 66-kDa (ERα) or 56-kDa (ERβ) are presented. Data are percentage of control values and are mean ± SEM of 6 values (3 different experiments).
To further validate microarray data, COX1 and COX2 protein expression were analyzed by immunoblotting (Figure 8A and 8B). Estradiol increased COX1 protein expression up to 30 % of control values, whereas COX2 protein expression remained unchanged, in sharp agreement to data obtained from microarray analysis and QRT-PCR assays. Moreover, estradiol-induced COX1 up-regulation resulted in an increased production of prostacyclin, the main vasodilatory prostanoid regulated by COX activity (Figure 8C). These stimulatory effects of estradiol on prostacyclin synthesis pathway were mediated through ER activation, since were abolished in the presence of ICI 182780.
HUVEC were exposed to 1 nmol/L estradiol with or without 1 µmol/L ICI182780, and protein expression of (A) COX1 and (B) COX2 and prostacyclin production (C) were measured as stated in Materials and methods. Data are percentage of control values and are mean ± SEM of 6–8 values (3–4 different experiments). * p<0.05 or ** p<0.01 vs. control cells, and † p<0.05 vs. estradiol-alone treated cells.
In a similar way, estradiol-induced changes in the DDAH gene expression were correlated to similar changes in protein expression. DDAH2 protein expression was increased in the presence of estradiol, whereas DDAH1 remained unchanged (Figure 9A and 9B). In vivo, DDAH degrades most of ADMA [20], an endogenous inhibitor of NO synthase. The increased DDAH expression resulted in decreased ADMA production (Figure 9C), which in turn lead to an increased NO production after estradiol exposure (Figure 9D). The effects of estradiol on the DDAH-ADMA-NO pathways were mediated by ER, since were abolished in the presence of ICI 182780.
HUVEC were exposed to 1 nmol/L estradiol with or without 1 µmol/L ICI182780, and protein expression of (A) DDAH1 and (B) DDAH2, along with (C) ADMA levels and NO production, were measured as stated in Materials and methods. Data are percentage of control values and are mean ± SEM of 9–12 values (4 different experiments). * p<0.05 or ** p<0.01 vs. control cells, and † p<0.05 vs. estradiol-alone treated cells.
Discussion
This study summarizes changes in complete gene expression in human endothelial cells exposed to estradiol. We have identified new genes that are up-regulated in endothelium by exposure to a physiological concentration of estradiol (1 nmol/L) for 24 hours, a time and a concentration selected according to previous work of our group [19]. We have identified 1886 genes differentially expressed. Taking advantage of ranked gene expression pathways, results have shown that pathways related to cellular growth and proliferation, cell-to-cell signaling and cellular organization, movement and death were among the most differentially expressed.
Canonical pathway analysis revealed Notch signaling as the most significant signaling pathway modulated by estradiol. Aberrant Notch signaling is implicated in carcinogenesis and tumor angiogenesis [21], and interestingly with human pathologies involving cardiovascular abnormalities [22]. Recently, it was reported that Notch pathway regulates cell-cell or cell-matrix interaction, contributing hence, to cell migration in situations of tissue remodeling [23]. Also, Notch1 has been implicated in the estradiol-induced increase in microvessel density in vivo and therefore in estradiol-increased tumor angiogenesis in MCF7 cells and HUVEC [24]. Our findings provide further support for the important role that Notch signaling pathway plays on endothelial effects of estradiol.
Estradiol has also important effects on other signaling pathways, mainly in Actin Cytoskeleton Signaling, Integrin Signaling, and Vascular Endothelial Growth Factor (VEGF) Signaling. These pathways exert important vascular actions, such as maintaining vascular integrity, regulating cell cycle, and promoting vasculogenesis. Moreover, four metabolic pathways are among the first 10 pathways significantly modulated by estradiol: Pentose Phosphate Pathway, Galactose Metabolism, N-Glycan Biosynthesis and Inositol Phosphate Metabolism. Some of these effects of estradiol have already been described. Estradiol, for instance, has already reported to increase HUVEC attachment to extracellular matrix proteins, mainly up-regulating surface expression of β1, α5 and α6 integrins [25]. Estradiol directly regulates the glucose-6-phosphate dehydrogenase (G6PDH) expression [26], the enzyme that directs glucose carbons into the pentose phosphate pathway. Moreover, estradiol-stimulated breast cancer cells have also increased pentose phosphate pathway activity, suggesting that this pathway is essential for estrogen-dependent cell proliferation [27]. Nevertheless, the majority of genes affected by estradiol treatment have been described for the first time in our results and our data open new approaches to discover unexplored estrogen-regulated pathways and new vascular actions.
The IPA software outlined the most changed pathways in the microarray data. Among them, TGFB1 plays a central role in the formation of the number-one-ranked network, which contains 33 genes (Figure 4) and is associated with other important cardiovascular networks, such as Cardiovascular System Development and Function, Cellular Growth and Proliferation and Cell Morphology. TGFB1 is a multifunctional peptide that controls proliferation, differentiation, and other functions in many cell types. In our study, TGFB1 was significantly up-regulated by estradiol as a main effect, supporting its important role in cardiovascular function. According to our results, estradiol exerts an important role in vessel assembly and stabilization through TGFB signaling pathways [28]. Moreover, TGFB pathway status determines the antiatherogenic effect of estradiol in apoE-/- hypercholesterolemic mice [29]. Furthermore, estradiol administration to postmenopausal women increases circulating levels of the active form of TGFB1 [30]. Altogether, these findings led to the conclusion that TGFB1 is one of the main targets of estradiol stimulation.
With the use of ICI 182780 in some experiments, our study demonstrates that activation of ER by E2 modifies the expression of several genes in HUVEC. In spite of some authors have found that HUVEC do not express ERα [31], other investigators have demonstrated the presence of both ERα and ERβ mRNA in HUVEC [32]. Data presented in Figure 7 demonstrate the expression of both ERα and ERβ protein in HUVEC, thus confirming previous reports [33], [34].
The extensive information gained from this first analysis has resulted in the collection of new data and new genes that provide other opportunities of study not explored so far, for example, the increased expression of the DHCR7 gene when HUVEC were exposed to different estradiol concentrations. This gene is responsible for the last step in cholesterol synthesis, and its inhibition results in hypocholesterolemia and accumulation of 7-dehidrocholesterol [35], while different mutations of this gene cause the Smith-Lemli-Opitz syndrome [36]. In our study, DHCR7 expression induced by 1 nmol/L estradiol was completely abolished in the presence of ICI182780 (Figure 5D). This is similar to the unique description of the relationship between this gene and estradiol, in which the expression of DHCR7 on human osteosarcoma cells was increased in response to estradiol through receptor beta [37].
Other cardiovascular-relevant genes confirmed the consistency of microarray data. Thus, the results are in accordance with similar effects observed in endothelial cells, both measuring the gene or the protein expression. COX are the rate-limiting step in the formation of vasoactive prostanoids, such as prostacyclin and thromboxane, from arachidonic acid [38]. Our results point to an estradiol-induced, dose-dependent gene expression, resulting in increased protein expression of COX-1 without effect on COX-2, which in turn resulted in increased prostacyclin production. These data, mediated through ER activity, have already been reported in some studies performed in ovine pulmonary artery endothelial cells [39], but not in others [40]. Related to COX-mediated prostanoid production, cytosolic phospholipase A2 activity is the initial step which liberates arachidonic acid from the cell membrane. In our study, PLA2G4B expression was reported to be dose-dependent increased by estradiol, while the main subtype PLA2G4A remained unaltered. Previous studies have reported an increase in cytosolic phospholipase A2 protein expression, without subtype differentiation, in ovine [41] and rat [42] uterine arteries exposed to estradiol.
ADMA is an analogue of arginine, which is synthesized endogenously and can act as inhibitor of nitric oxide synthase [20]. Both DDAH are responsible in vivo for ADMA degradation to citrulline and dimethylamine. According to the results obtained in the microarray analysis and confirmed by QRT-PCR and inmunoblotting, DDAH2 is increased in HUVEC exposed to different concentrations of estradiol, whereas DDAH1 remains unaltered. DDAH2, the main subtype in the cardiovascular system, has already been reported to be increased by estradiol in endothelium [19]. Moreover, the increased DDAH2 expression resulted in decreased ADMA concentration and therefore, increased NO release. Results of the present work further support that increased DDAH2 expression is dependent on ER-dependent genomic activity.
The strength of the current study is the careful design of the experiments and the use of sample pools which contribute to minimizing inter-individual variations. The average fold-change induced by estradiol is relatively low, but it should be taken into account that cells were exposed to estradiol concentrations that were within physiological levels in premenopausal women [43]. Moreover, fold-changes and number of up-regulated genes in our study were within the same range as that obtained in similar studies performed with higher estradiol concentrations (10–50 nmol/L) in different human breast cancer cell types [44].
Care should be taken determining clinically relevant consequences. Much in vitro and in vivo experimental data support a beneficial effect of estrogens on the cardiovascular system [6]. Observational studies have also consistently shown a benefit of hormone replacement therapy on cardiovascular disease, but some randomized studies have shown even some deleterious effects [6], [45]. New experimental approaches, such as the present study, should contribute to conciliate the divergences observed between clinical and experimental data.
In summary, our study generates a comprehensive estradiol-mediated gene expression profile in HUVEC and characterizes in detail the considerably different responses of control and estradiol-treated endothelial cells. The present study provides the first quantitative large-scale gene expression analysis of estradiol–stimulated human vascular endothelial cells. Identification of pathways regulated by estradiol may add to the knowledge base of how estradiol contributes to a wide range of biological processes. These results could lead to a deeper understanding of fundamental insights of pathophysiological mechanisms involved in cardiovascular diseases such as stroke or atherosclerosis at the level of gene expression and provide a foundation for the development of better therapeutic strategies for vascular disease.
Acknowledgments
The authors are indebted to Mrs. Rosa Aliaga, Mrs. Elvira Calap and Mr. Carlos Bueno for their excellent technical assistance.
Author Contributions
Conceived and designed the experiments: AS CH. Performed the experiments: AS MM ALF SN PJO CH. Analyzed the data: AS MM SN PJO MAGP JJT CH. Contributed reagents/materials/analysis tools: ALF MAGP JJT AC CH. Wrote the paper: AS MM SN JJT AC CH.
References
- 1. Mendelsohn ME, Karas RH (2005) Molecular and cellular basis of cardiovascular gender differences. Science 308: 1583–1587.
- 2. Barrett-Connor E, Grady D (1998) Hormone replacement therapy, heart disease, and other considerations. Annu Rev Pub Health 19: 55–72.
- 3. Hulley S, Grady D, Bush T, Furberg C, Herrington D, et al. (1998) Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. Heart and Estrogen/progestin Replacement Study (HERS) Research Group. JAMA 280: 605–613.
- 4. Grady D, Herrington D, Bittner V, Blumenthal R, Davidson M, et al. (2002) Cardiovascular disease outcomes during 6.8 years of hormone therapy: Heart and Estrogen/progestin Replacement Study follow-up (HERS II). JAMA 288: 49–57.
- 5. Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, Kooperberg C, et al. (2002) Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women's Health Initiative randomized controlled trial. JAMA 288: 321–333.
- 6. Turgeon JL, Carr MC, Maki PM, Mendelsohn ME, Wise PM (2006) Complex Actions of Sex Steroids in Adipose Tissue, the Cardiovascular System, and Brain: Insights from Basic Science and Clinical Studies. Endocr Rev 27: 575–605.
- 7. Ross R (1999) Atherosclerosis: an inflammatory disease. N Engl J Med 340: 115–126.
- 8. Mendelsohn ME (2000) Mechanisms of estrogen action in the cardiovascular system. J Steroid Biochem Mol Biol 74: 337–343.
- 9. Mendelsohn ME (2002) Protective effects of estrogen on the cardiovascular system. Am J Cardiol 89: 12E–18E.
- 10. Trevino V, Falciani F, Barrera-Saldana HA (2007) DNA microarrays: a powerful genomic tool for biomedical and clinical research. Mol Med 13: 527–541.
- 11. Archacki SR, Wang QK (2006) Microarray analysis of cardiovascular diseases. Methods Mol Med 129: 1–13.
- 12. Hiltunen MO, Tuomisto TT, Niemi M, Bräsen JH, Rissanen TT, et al. (2002) Changes in gene expression in atherosclerotic plaques analyzed using DNA array. Atherosclerosis 165: 23–32.
- 13. Graham D, McBride MW, Gaasenbeek M, Gilday K, Beattie E, et al. (2007) Candidate genes that determine response to salt in the stroke-prone spontaneously hypertensive rat: congenic analysis. Hypertension 50: 1134–1141.
- 14. Hermenegildo C, Oviedo PJ, Garcia-Perez MA, Tarin JJ, Cano A (2005) Effects of phytoestrogens genistein and daidzein on prostacyclin production by human endothelial cells. J Pharmacol Exp Ther 315: 722–728.
- 15. Li C, Hung Wong W (2001) Model-based analysis of oligonucleotide arrays: model validation, design issues and standard error application. Genome Biology 2: research0032.
- 16. Edgar R, Domrachev M, Lash AE (2002) Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucl Acids Res 30: 207–210.
- 17. Pfaffl MW, Horgan GW, Dempfle L (2002) Relative expression software tool (REST(C)) for group-wise comparison and statistical analysis of relative expression results in real-time PCR. Nucl Acids Res 30: e36.
- 18. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) Protein measurement with the Folin phenol reagent. J Biol Chem 196: 265–275.
- 19. Monsalve E, Oviedo PJ, Garcia-Perez MA, Tarin JJ, Cano A, et al. (2007) Estradiol counteracts oxidized LDL-induced asymmetric dimethylarginine production by cultured human endothelial cells. Cardiovasc Res 73: 66–72.
- 20. Vallance P, Leiper J (2004) Cardiovascular biology of the asymmetric dimethylarginine:dimethylarginine dimethylaminohydrolase pathway. Arterioscler Thromb Vasc Biol 24: 1023–1030.
- 21. Leong KG, Karsan A (2006) Recent insights into the role of Notch signaling in tumorigenesis. Blood 107: 2223–2233.
- 22. Uyttendaele H, Ho J, Rossant J, Kitajewski J (2001) Vascular patterning defects associated with expression of activated Notch4 in embryonic endothelium. PNAS 98: 5643–5648.
- 23. Lindner V, Booth C, Prudovsky I, Small D, Maciag T, et al. (2001) Members of the Jagged/Notch gene families are expressed in injured arteries and regulate cell phenotype via alterations in cell matrix and cell-cell interaction. Am J Pathol 159: 875–883.
- 24. Soares R, Balogh G, Guo S, Gartner F, Russo J, et al. (2004) Evidence for the Notch signaling pathway on the role of estrogen in angiogenesis. Mol Endocrinol 18: 2333–2343.
- 25. Cid M, Esparza J, Schnaper H, Juan M, Yague J, et al. (1999) Estradiol enhances endothelial cell interactions with extracellular matrix proteins via an increase in integrin expression and function. Angiogenesis 3: 271–280.
- 26. Thomas M, Bader C, Monet JD (1990) Sex steroid hormone modulation of NADPH pathways in MCF-7 cells. Cancer Res 50: 1195–1200.
- 27. Forbes NS, Meadows AL, Clark DS, Blanch HW (2006) Estradiol stimulates the biosynthetic pathways of breast cancer cells: Detection by metabolic flux analysis. Metabolic Engineering 8: 639–652.
- 28. Soares R, Guo S, Gärtner F, Schmitt F, Russo J (2003) 17β-Estradiol-mediated vessel assembly and stabilization in tumor angiogenesis requires TGFβ and EGFR crosstalk. Angiogenesis 6: 271–281.
- 29. Gourdy P, Schambourg A, Filipe C, Douin-Echinard V, Garmy-Susini B, et al. (2007) Transforming growth factor activity is a key determinant for the effect of estradiol on fatty streak deposit in hypercholesterolemic mice. Arterioscler Thromb Vasc Biol 27: 2214–2221.
- 30. Djurovic S, Os I, Hofstad AE, Abdelnoor M, Westheim A, et al. (2000) Increased plasma concentrations of TGF-beta1 after hormone replacement therapy. J Int Med 247: 279–285.
- 31. Toth B, Scholz C, Saadat G, Geller A, Schulze S, et al. (2008) Estrogen receptor modulators and estrogen receptor beta immunolabelling in human umbilical vein endothelial cells. Acta Histochemica.
- 32. Wagner AH, Schroeter MR, Hecker M (2001) 17beta-estradiol inhibition of NADPH oxidase expression in human endothelial cells. FASEB J 15: 2121–2130.
- 33. Oviedo PJ, Hermenegildo C, Tarin JJ, Cano A (2007) Raloxifene increases proliferation of human endothelial cells in association with increased gene expression of cyclins A and B1. Fertil Steril 88: 326–332.
- 34. Harris HA (2007) Estrogen Receptor-{beta}: Recent Lessons from in Vivo Studies. Mol Endocrinol 21: 1–13.
- 35. Kolf-Clauw M, Chevy F, Wolf C, Siliart B, Citadelle D, et al. (1996) Inhibition of 7-dehydrocholesterol reductase by the teratogen AY9944: a rat model for Smith-Lemli-Opitz syndrome. Teratology 54: 115–125.
- 36. Witsch-Baumgartner M, Loffler J, Utermann G (2001) Mutations in the human DHCR7 gene. Hum Mutat 17: 172–182.
- 37. Monroe DG, Secreto FJ, Subramaniam M, Getz BJ, Khosla S, et al. (2005) Estrogen Receptor {alpha} and {beta} Heterodimers Exert Unique Effects on Estrogen- and Tamoxifen-Dependent Gene Expression in Human U2OS Osteosarcoma Cells. Mol Endocrinol 19: 1555–1568.
- 38. Smith WL, DeWitt DL, Garavito RM (2000) Cyclooxygenases: structural, cellular, and molecular biology. Annu Rev Biochem 69: 145–182.
- 39. Jun SS, Chen Z, Pace MC, Shaul PW (1998) Estrogen upregulates cyclooxygenase-1 gene expression in ovine fetal pulmonary artery endothelium. J Clin Invest 102: 176–183.
- 40. Hermenegildo C, Oviedo PJ, Cano A (2006) Cyclooxygenases regulation by estradiol on endothelium. Curr Pharm Des 12: 205–215.
- 41. Rupnow HL, Phernetton TM, Modrick ML, Wiltbank MC, Bird IM, et al. (2002) Endothelial vasodilator production by uterine and systemic arteries. VIII. Estrogen and progesterone effects on cPLA2, COX-1, and PGIS protein expression. Biol Reprod 66: 468–474.
- 42. Farina MG, Billi S, Leguizamon G, Weissmann C, Guadagnoli T, et al. (2007) Secretory and cytosolic phospholipase A2 activities and expression are regulated by oxytocin and estradiol during labor. Reproduction 134: 355–364.
- 43. Hermenegildo C, Garcia-Martinez MC, Tarin JJ, Cano A (2002) Estradiol reduces F2alpha -isoprostane production in cultured human endothelial cells. Am J Physiol Heart Circ Physiol 283: H2644–H2649.
- 44. Cicatiello L, Scafoglio C, Altucci L, Cancemi M, Natoli G, et al. (2004) A genomic view of estrogen actions in human breast cancer cells by expression profiling of the hormone-responsive transcriptome. J Mol Endocrinol 32: 719–775.
- 45. Stevenson JC (2007) HRT and the primary prevention of cardiovascular disease. Maturitas 57: 31–34.