The present study was designed to assess whether or not changes in genomic expression of cyclooxygenases (COX-1, COX-2), endothelial nitric oxide synthase (eNOS), and prostanoid synthases in the endothelium and of prostanoid receptors in vascular smooth muscle contribute to the occurrence of endothelium-dependent contractions during aging and hypertension. Gene expression was quantified by real-time PCR using isolated endothelial cells and smooth muscle cells (SMC) from the aorta of Wistar-Kyoto and spontaneously hypertensive rats. Genes for all known prostanoid synthases and receptors were present in endothelial cells and SMC, respectively. Aging caused overexpression of eNOS, COX-1, COX-2, thromboxane synthase, hematopoietic-type prostaglandin D synthase, membrane prostaglandin E synthase-2, and prostaglandin F synthase in endothelial cells and COX-1 and prostaglandin E2 (EP)4 receptors in SMC. Hypertension augmented the expression of COX-1, prostacyclin synthase, thromboxane synthase, and hematopoietic-type prostaglandin D synthase in endothelial cells and prostaglandin D2 (DP), EP3, and EP4 receptors in SMC. The increase in genomic expression of endothelial COX-1 explains why in aging and hypertension the endothelium has greater propensity to release cyclooxygenase-derived vasoconstrictive prostanoids. The expression of prostacyclin synthase was by far the most abundant, explaining why the majority of the COX-1-derived endoperoxides are transformed into prostacyclin, substantiating the role of prostacyclin as an endothelium-derived contracting factor. The expression of thromboxane synthase was increased in the cells of aging or hypertensive rats, explaining why the prostanoid can contribute to endothelium-dependent contractions. It is uncertain whether the gene modifications caused by aging and hypertension directly contribute to endothelium-dependent contractions or rather to vascular aging and the vascular complications of the hypertensive process.
- endothelium-dependent contractions
- endothelium-derived contracting factors
- prostacyclin synthase
- real-time quantitative polymerase chain reaction
the occurrence of endothelium-dependent contractions is accelerated by aging and hypertension (13, 17, 19, 39, 42). Thus such contractions are observed readily in aortas from adult spontaneously hypertensive rats (SHR) and aged normotensive Wistar-Kyoto rats (WKY) (17, 19, 42). The mediators of endothelium-dependent contractions are produced by endothelial cyclooxygenase (7, 12, 33) and have been termed endothelium-derived contracting factors (EDCFs). EDCF is unlikely to be a single substance but rather is constituted of a mixture of prostanoids, the formation of which depends on the vascular bed, the age, and the condition of the species studied. Thus prostacyclin (9, 10, 27), endoperoxides (7, 14), and thromboxane A2 (10, 11, 29) have been proposed as EDCFs. Cyclooxygenase breaks down arachidonic acid to form endoperoxides, which are subsequently converted into prostacyclin, thromboxane A2, prostaglandin D, prostaglandin E, and/or prostaglandin F by their respective synthases (3, 20). There are two subtypes of prostaglandin D synthases [lipocalin type and hematopoietic type (38)], prostaglandin F synthases [liver and lung prostaglandin F synthase, named after their site of purification (30)], and prostaglandin E synthases [cytosolic and membrane type, based on their location within the cell (23); the membrane prostaglandin E synthase is further subdivided into two isoforms: mPGES-1 and mPGES-2].
Cyclooxygenase-derived EDCFs diffuse to the underlying smooth muscle, where they activate prostanoid receptors (41). Prostanoid receptors are classified into five discrete types based on their sensitivity to the five naturally occurring prostanoids. They are termed P receptors, with a preceding letter indicating the prostanoid [prostacyclin (prostaglandin I2; IP), thromboxane (TP), prostaglandin D2 (DP), prostaglandin E2 (EP), and prostaglandin F2 (FP)] to which they are the most sensitive (6). The EP receptors are divided further into four subtypes denoted as EP1, EP2, EP3, and EP4 (6). These isoforms differ in their structure and G protein coupling and trigger different cellular responses (6). The primary prostanoid receptor responsible for endothelium-dependent contractions is believed to be the TP receptor, because such contractions are prevented by selective TP receptor antagonists (2, 42). However, other prostanoid receptors may also contribute (29).
Molecular sequencing technology has provided the mRNA sequence for most types of prostanoid synthases and prostanoid receptors in the rat. The exceptions are prostaglandin F synthase and cytosolic prostaglandin E synthase, for which only the bovine and murine sequences, respectively, are available (31, 36). Since relative homology exists between both the bovine prostaglandin F synthase and the murine cytosolic prostaglandin E synthase with the corresponding rat enzymes, amplification of rat mRNA is possible with these sequences (5, 15, 31). The relative contribution of prostanoid synthases and receptors in rat nephron (40), primary cultured rat hepatocytes (26), and the human trabecular meshwork (16) is known, but information on their expression is lacking for aortic endothelial and smooth muscle cells of the rat. Indeed, the expression of prostacyclin synthase and of the IP receptor, but not of the other prostanoid synthases and receptors, has been quantified only in whole aortas of the rat (25).
The present study was designed to evaluate whether or not aging or hypertension can cause changes in the diversity of the gene expression of different prostanoid synthases in freshly isolated endothelial cells and of prostanoid receptors in vascular smooth muscle cells and to determine whether or not such alterations in genomic expression can help to explain the occurrence of endothelium-dependent contractions. The effect of hypertension was studied by comparing the differences between 36-wk-old hypertensive SHR and age-matched normotensive WKY. The effect of aging was studied by comparing the differences between adult (36 wk old) and aged (72 wk old) WKY.
MATERIALS AND METHODS
Experiments were performed on thoracic aortas from 36 wk-old SHR, 36 wk-old WKY, and 72 wk-old WKY. The animals were housed in a temperature-controlled room (21 ± 1°C) with a 12:12-h light-dark cycle (0700 lights on, 1900 lights off) and free access to chow (LabDiet 5010, Brentwood, MO) and water. The rats were anesthetized intraperitoneally with pentobarbital sodium (70 mg·ml−1·kg−1; Ganes Chemicals, Pennsville, NJ), and their blood pressure was measured by means of a catheter placed in the carotid artery. The aorta was excised, cleaned of adhering fat and connective tissue, and placed immediately in Krebs-Ringer bicarbonate buffer of the following composition (mmol/l): 118 NaCl, 4.7 KCl, 2.5 CaCl2, 1.2 MgSO4, 1.2 KH2PO4, 25 NaHCO3, and 11.1 glucose (control solution). This study was approved by the Animal Care and Use Committee of the University of Hong Kong.
Aortic rings from SHR or WKY were suspended in organ chambers, which contained control solution (37°C), aerated with 95% O2 and 5% CO2. They were connected to force transducers (ADInstruments, Sydney, Australia) for isometric tension recording (PowerLab, ADInstruments). The rings were allowed to equilibrate for 1 h at their optimal resting tension of 2.5 g (as determined in preliminary study; data not shown). To study endothelium-dependent relaxations, rings were contracted with phenylephrine (10−6 mol/l; Sigma, St. Louis, MO) and then exposed to progressively increasing concentrations of acetylcholine (10−10 to 10−4 mol/l; Sigma). All changes in tension were expressed as a percentage of the phenylephrine-induced contraction. To study endothelium-dependent contractions, quiescent rings were incubated with Nω-nitro-l-arginine methyl ester [l-NAME, a nitric oxide synthase (NOS) inhibitor, 10−4 mol/l; Sigma] for 40 min, to optimize endothelium-dependent contractions (34, 43), and then exposed to progressively increasing concentrations of acetylcholine (10−8 to 10−4 mol/l). All changes in tension were expressed as a percentage of the reference contraction to 60 mmol/l KCl obtained at the start of the experiment. In some preparations the endothelium was removed mechanically by inserting the tip of a syringe needle into the ring and rolling it back and forth in a Sylgard-based container filled with control solution. In other experiments, the rings were treated with indomethacin (cyclooxygenase inhibitor, 10−5 mol/l; Sigma) or terutroban (S18886, a TP receptor antagonist, 10−7 mol/l; a kind gift of the Institut de Recherches Servier) for 40 min before evoking endothelium-dependent contractions.
Total RNA extraction.
Aortas were opened and pinned down on sterilized Sylgard-coated plates filled with control solution with the endothelial cell layer facing upwards. The endothelial cells were harvested by scraping with a sharp spatula (Cell Lifter, Corning Costar, New York, NY). Aortic endothelial cells from six individual rats were pooled into 1 ml of Tri Reagent (Molecular Research Center, Cincinnati, OH) to form one endothelial cell RNA sample. To collect vascular smooth muscle cells, each aortic strip was thoroughly scraped to ensure the complete removal of endothelial cells. Subsequently, it was cut into small pieces and homogenized at 4°C in Tri Reagent to form one vascular smooth muscle RNA sample (pooling was not required for collection of smooth muscle cells). To improve the efficiency of RNA isolation, cells were triturated by thrusting the Tri Reagent cell mixture up and down a syringe needle to facilitate cell rupture. Total RNA was isolated with Tri Reagent according to manufacturer's instructions.
Reverse transcription was carried out in a total volume of 20 μl, comprising 1 μg of total endothelial or smooth muscle cell RNA, 4 μl of 5× first-strand buffer (GIBCO-BRL, Rockville, MD), 2 μl of 0.1 M dithiothreitol (DTT; Sigma), 1 μl of 10 mM deoxynucleoside triphosphate (dNTP; GIBCO-BRL), 0.4 μl of oligo(dT) (500 ng/μl; GIBCO-BRL), 0.5 μl of 40 U/μl ribonuclease inhibitor (RNase; Roche, Indianapolis, IN), and 0.1 μl of Moloney murine leukemia virus reverse transcriptase (200 U/μl; GIBCO-BRL). Diethyl pyrocarbonate (DEPC)-treated water was used to make up the total volume. The mixture was left at room temperature for 10 min, followed by incubation at 37°C for 60 min. The reaction was stopped by incubation at 95°C for 2 min, and the cDNA was stored at −80°C until use.
Gene sequence information was obtained from the Entrez Nucleotide database, and all primer sequences were designed to be target specific with the T-coffee alignment program (www.expasy.org). Sense and antisense primers (Table 1) were designed to have a closely matched melting temperature. All primer sets underwent preliminary testing to ensure that they were effective, in that a clean positive band could be obtained. The polymerase chain reaction (PCR) assay was carried out by using 45 μl of PCR SuperMix (Invitrogen, Carlsbad, CA), 1 μl of each primer (1 μM), and 2 μl of reverse-transcribed product with the GeneAmp PCR system 9200 (Applied Biosystems, Warrington, UK). The reaction conditions were as follows: 95°C for 30 s for denaturing, 55°C for 30 s for annealing, and 72°C for 1 min for Taq activity. The PCR was performed for 35 cycles with a final 10-min extension step. The resulting PCR mix was analyzed by electrophoresis on 1.5% agarose gels containing ethidium bromide. The reverse-transcribed product used to initiate the PCR reaction in order to generate a positive control band for each primer set was derived from different tissues (Table 1).
Gene-specific PCR product purification.
After the adequacy of the primer pairs was ensured, gene-specific PCR products were purified with the Wizard PCR preps DNA Purification System (Promega, Madison, WI) according to manufacturer's instructions. Each gene-specific PCR product was quantified by the GeneQuant II spectrophotometer (Pharmacia Biotech, Cambridge, UK) based on isolated product concentration at optical density of 260 nm. The concentration of DNA was converted to molecules per microliter of gene products (28).
Preparation of gene-specific standard curve.
Gene expression was quantified with the standard curve technique using SYBR Green PCR master mix (Applied Biosystems) in real-time PCR, which allowed the quantification of the exact copy number of genes per microliter of cDNA. The purified gene-specific DNAs were aliquoted serially into successive ninefold diminishing dilutions (from 1010 to 101 specific copies/μl cDNA) in DEPC-treated water to construct standard reference curves, which were obtained in the real-time PCR assay concurrently with the unknown samples.
Unknown samples and gene-specific PCR products at each dilution were amplified with the ABI Prism 7000 real-time PCR machine (Applied Biosystems) in 20 μl of total reaction volume containing 10 μl of 2× SYBR Green PCR master mix, 0.4 μl of sense primer (1 μM), 0.4 μl of antisense primer (1 μM), 8.2 μl of DEPC-treated water, plus 1 μl of specific PCR product at one dilution or the unknown cDNA sample. Master mixes were prepared where possible. Blank samples with no template DNA were examined to ensure that there was no contamination. The following cycling conditions were used to amplify PCR products with the real-time PCR system: cycle 1: 95°C for 10 min; cycle 2: 95°C for 30 s, 55°C for 30 s (40×); and cycle 3: 95°C for 15 s, 55°C for 15 s, 95°C for 15 s. Each assay was carried out in duplicate, and the critical threshold (CT) at linearity was determined for each sample or for a dilution from the standard curve for regression analysis. The quality of the PCR products was tested after each run by melt curve analysis and electrophoresis of the end PCR product to ensure that the products had the specific expected size producing a single band with no smears or primer dimers. CT values obtained in the standard curve were plotted against the log of template amount (molecules/μl) and the amount of copy number was calculated (28).
Contamination by smooth muscle.
To validate the successful isolation of endothelial cells, the ratio between α- and β-actin was determined. α-Actin is expressed in smooth muscle cells only, while β-actin is expressed in many cell types (37). Samples with an α- over β-actin ratio >0.5% were excluded from the study, because this presumably reflected poor scraping with too much contamination by smooth muscle cells (37). The positive detection of the endothelial NOS (eNOS) gene within the samples reinforced the conclusion that they were composed of endothelial cells.
All final values for specific gene expression are given as copy number in molecules per microliter of cDNA. Results are presented as group means ± SE, with n being the number of individual (or pooled) observations. Statistical analysis was performed by unpaired Student's t-test or one-way ANOVA, wherever appropriate, with Prism version 3a (GraphPad Software, San Diego, CA). A difference was accepted as statistically significant when probability (P) values were <0.05.
Body weight and blood pressure.
The body weight for 36 wk-old SHR, 36 wk-old WKY, and 72 wk-old WKY averaged 414 ± 4, 405 ± 4, and 470 ± 6 g, respectively. The body weights of 36 wk-old WKY and SHR were not statistically significantly different.
The mean arterial blood pressure for 36 wk-old SHR, 36 wk-old WKY, and 72 wk-old WKY averaged 201 ± 5, 116 ± 5, and 112 ± 6 mmHg, respectively. Mean arterial blood pressure was significantly higher in 36 wk-old SHR compared with age-matched WKY. Aging did not increase mean arterial blood pressure in the normotensive rats.
In aortic rings of 36-wk-old WKY contracted with phenylephrine, acetylcholine caused concentration-dependent relaxation. However, in aortic rings of 36-wk-old SHR, acetylcholine caused a triphasic response. A concentration-dependent relaxation was achieved from 10−10 to 10−7 M. This was followed by a rebound in tension at concentrations between 10−7 and 3 × 10−6 M. The third phase was a secondary fall in tension from 3 × 10−6 M onwards. The concentration-response curve to acetylcholine in contracted rings of aged WKY (72 wk old) mimicked that obtained in aortas of 36-wk-old SHR (Fig. 1).
Endothelium-dependent contractions to acetylcholine were apparent in quiescent aortas of 36 wk-old SHR, but not in those from age-matched WKY. The aortas of normotensive WKY at the age of 72 wk also developed endothelium-dependent contractions. The magnitude of such contraction was significantly lower than that observed in arteries of 36 wk-old SHR (Fig. 1). The contractions elicited by acetylcholine in the aorta of 36-wk-old SHR and 72-wk-old WKY were prevented by removal of the endothelium or by treatment with indomethacin or terutroban (data not shown).
Standard curve and confirmation of PCR products.
Standard curves were generated for each gene with the purified PCR product (Fig. 2A). Linear regression analysis confirmed that PCR amplification with SYBR Green in real-time analysis provided highly reproducible results, because duplicates gave closely matched values (Fig. 2B). The average r2 value for all genes assayed in this study was 0.995. Detection limits achievable could be as low as 1 molecule per 20 μl of PCR reaction. Reaction with no template gave negative values, representing the absence of specific cDNA. Melt curve analysis was carried out at the end of the PCR run by ramping the temperature of the samples from 55°C to 95°C while continuously collecting fluorescence data. The curves of the melting profiles of prostanoid synthases and prostanoid receptors did not reveal an accumulation of primer dimers or nonspecific products (Fig. 2C). Furthermore, end PCR products were run in a 1.5% agarose electrophoresis gel to confirm the presence of a single band demonstrating the absence of contamination by genomic DNA or nonspecific products (Fig. 2D).
α-Actin over β-actin ratio.
There was abundant mRNA encoding for β-actin and relatively little expression of α-actin in the pooled endothelial cell samples (Fig. 3). The average α-actin over β-actin ratio for all endothelial cell samples was <0.08%. There was no significant difference in the genomic level of α-actin among the three animal groups. The gene expression of β-actin was similar between 36-wk-old WKY and SHR. However, the genomic expression of β-actin in 72-wk-old WKY was significantly greater than that in 36-wk-old WKY.
COX-1 and eNOS.
The genomic expression of COX-1 in freshly isolated endothelial cells was significantly higher in 36-wk-old SHR and 72-wk-old WKY compared with 36-wk-old WKY (Fig. 4). There was no difference in gene expression of COX-1 in smooth muscle cells between 36-wk-old WKY and SHR (Fig. 4). However, aging of the WKY resulted in higher genomic expression of COX-1 in smooth muscle (Fig. 4).
There was no statistically significant difference in the gene expression of COX-2 between endothelial cells of 36-wk-old WKY and SHR. However, aging of the WKY resulted in higher genomic expression of COX-2 in these cells (Fig. 4). COX-2 transcripts were also detected in the smooth muscle cells of the different rats studied. Neither aging nor hypertension altered the gene expression of COX-2 in smooth muscle cells (Fig. 4). Overall, COX-2 was expressed in both endothelial and smooth muscle cells in significantly lower levels than COX-1 (Fig. 4).
The expression of eNOS in freshly isolated endothelial cell samples was significantly lower in 36-wk-old SHR compared with age-matched WKY (Fig. 4). Aging of the WKY resulted in higher genomic expression of eNOS (Fig. 4).
Prostanoid synthases in endothelial cells.
The gene expression of all known subtypes of prostanoid synthases was detected in endothelial cells of adult SHR, adult WKY, and aged WKY (Fig. 5). There was a significant increase in the genomic level of prostacyclin synthase, thromboxane synthase, and hematopoietic-type prostaglandin D synthase in 36-wk-old SHR compared with age-matched WKY. By contrast, the genomic expression of lipocalin-type prostaglandin D synthase was significantly decreased in 36-wk-old SHR (Fig. 5). Hypertension did not modify the genomic expression of prostaglandin F synthase, cytosolic prostaglandin E synthase, membrane prostaglandin E synthase-1, and membrane prostaglandin E synthase-2. The genomic expression of thromboxane synthase, hematopoietic-type prostaglandin D synthase, prostaglandin F synthase, and membrane prostaglandin E synthase-2 was increased in 72-wk-old WKY compared with 36-wk-old WKY. Aging did not change the genomic expression of prostacyclin synthase, lipocalin-type prostaglandin D synthase, and cytosolic and membrane prostaglandin E synthase-1 (Fig. 5). Prostacyclin synthase was consistently the most abundant prostanoid synthase expressed in endothelial cells from the different groups of rats being studied (Table 2). The genes of membrane prostaglandin E synthase-1 and prostaglandin F synthase were expressed at a relatively low level in the endothelial cells of the rats studied (Table 2).
Prostanoid receptors in vascular smooth muscle cells.
The gene expression of all known types of prostanoid receptors was detected in smooth muscle cells of adult SHR, adult WKY, and aged WKY (Fig. 6). There was an increase in the genomic expression of DP, EP3, and EP4 receptors in 36-wk-old SHR compared with age-matched WKY. Hypertension did not modify the genomic expression of IP, TP, FP, EP1, and EP2 receptors (Fig. 6). There were no genomic changes of prostanoid receptors with aging, except for the EP4 receptor, which was increased in 72-wk-old WKY compared with 36 wk-old WKY (Fig. 6). In 36-wk-old SHR, EP3 was the most abundantly expressed prostanoid receptor (Table 3). Its genomic expression was 15-fold greater than in age-matched WKY. For 36- and 72 wk-old WKY, the genomic expression of EP4 receptor dominated over the other prostanoid receptors. The EP2 was the least expressed receptor at the genomic level in smooth muscle cells of the rats studied (Table 3).
The present functional studies confirm that the development of endothelium-dependent contractions is prominent in the aorta of SHR (19, 34, 42) and that they appear in arteries of aging animals (12, 17). Although the magnitude of the endothelium-dependent contractions observed in the aorta of 72-wk-old WKY was relatively small in quiescent preparations, the blunting of the relaxation to higher concentrations of acetylcholine was comparable to that observed in the aorta of the hypertensive strain. These findings reinforce the earlier conclusions that the greater occurrence of endothelium-dependent contractions observed in the adult SHR is a result of the premature aging of the blood vessel caused by the chronic hypertensive process (39).
Loss of muscarinic receptors (37), inositol 1,4,5-trisphosphate receptors, and sarco(endo)plasmic reticulum Ca2+- ATPase occurs in cultured endothelial cells (22). These are essential for the accumulation of high intracellular calcium, required to initiate the production of EDCF (35). This makes cultured endothelial cells inappropriate for the study of cellular mechanisms that may underlie EDCF-mediated responses. Thus, in the present study, endothelial cells were separated freshly from the underlying vascular smooth muscle and each cell type was studied independently. The number of endothelial cells freshly harvested per aorta is limited. To generate enough RNA for eNOS detection, determination of α- over β-actin ratio, and quantification of the expression of the various genes studied, endothelial cells from six individual aortas were pooled to form one sample. Studies could only be carried out at the mRNA level, because under such isolation conditions the quantity of endothelial cells obtained was insufficient for protein analysis.
Actins are commonly used as an internal standard to normalize patterns of gene expression. An ideal internal standard should be expressed at a constant level among all stages of development and in different strains of rats being studied. The present findings demonstrated that the level of β-actin expression is altered by aging. Hence, the use of this gene to normalize expression in the aorta of the rat is probably inappropriate. Therefore, all genomic data in this study were expressed as absolute copy number per microliter of cDNA.
There are two isoforms of cyclooxygenase, COX-1 and COX-2 (12, 33). Both isoforms were detected, but the amount of COX-2 transcripts in endothelial or smooth muscle cells was less than that of COX-1. These findings are in favor of the notion that COX-1 is a constitutive and COX-2 an inducible enzyme in the rat aorta. The greater abundance of COX-1 further reinforces the notion that this enzyme is the prominent isoform for the production of EDCF, as has been suggested in the rat by pharmacological studies (7) and in the mouse by molecular interventions (33). With aging, the genomic expression of both endothelial and smooth muscle COX-1 increases, as demonstrated by its upregulation in aged WKY compared with adult WKY. Age-dependent modifications of COX-1 expression have been suggested previously (8). In that previous study, there was no difference in the gene expression of COX-1 between whole aortas of 20-wk-old WKY and SHR (8). However, when 35-wk-old rats were used, a significantly higher gene expression of COX-1 was found in the whole aortas of the SHR (7). This earlier work showed no differences in COX-1 expression between preparations with and without endothelium in both strains (7), implying that the genomic changes occur only within the smooth muscle cells. By contrast, the present study, which dissected the two layers of cells, showed that smooth muscle cells of 36-wk-old WKY and SHR contained similar levels of COX-1 mRNA but that the endothelial cells of 36-wk-old SHR expressed approximately twofold higher COX-1 mRNA than those of age-matched WKY. In disagreement with the previous study (7), this implies that the elevation of COX-1 expression in hypertensive rats is specific to the endothelial cells and that an increase of expression does not occur within the smooth muscle cells. The ratio of endothelial to smooth muscle cells per aorta is low. Thus a difference may not be measurable when endothelial COX-1 expression is compared in whole aortas, with or without endothelium (7). In the present study, the separation of the two cell types, the pooling of endothelial cells from six individual rats, and the high sensitivity of real-time PCR may have made the small difference in gene expression of endothelial COX-1 between the two strains obvious and thus detectable. The present data, however, provide no explanation for the divergence between the present and previous findings (7) in terms of the gene expression of COX-1 in the aortic smooth muscle of adult SHR.
Nitric oxide suppresses the development of endothelium-dependent contractions through immediate functional antagonism and by causing long-term inhibition of the response (34, 43). The augmented expression of eNOS in aged WKY may serve as a protective strategy against endothelium-dependent contractions. This may explain why the magnitude of endothelium-dependent contractions is low in aortas of aged normotensive rats. By contrast, the genomic expression of eNOS is reduced in the aortic endothelial cells of adult SHR compared with age-matched WKY. This limited amount of eNOS may result in a reduced ability to offset endothelium-dependent contractions. The overexpression of COX-1, together with the decrease of eNOS, in endothelial cells of SHR helps us to understand why vasoconstrictor cyclooxygenase-derived prostaglandin-mediated contractions rather than nitric oxide-mediated relaxations are observed on stimulation with acetylcholine in the aortic preparation of the hypertensive rat (19, 42).
The major advantage of absolute quantification in real-time PCR by standard curves over the use of the conventional PCR approach is that it allows the precise copy number of any specific gene present in the samples to be determined. This makes comparison in terms of quantity between different genes possible, despite differences in amplified sequence and primer properties (4). By profiling the complete gene expression of prostanoid synthases in endothelial cells, we hoped to better understand the importance of each prostanoid formed from the precursor endoperoxides in EDCF-mediated responses. Although the various prostanoid synthase genes were expressed in endothelial cells, a major finding was that prostacyclin synthase was by far the most abundant in both strains irrespective of age, explaining why endothelial cells are a major source of prostacyclin (9–11). Earlier measurement of the release of prostanoids from the rat aorta indeed showed that the majority of the endoperoxides produced by cyclooxygenase are transformed into prostacyclin (9–11). The fact that the expression of prostacyclin synthase in endothelial cells of 36-wk-old SHR is approximately three times higher than in age-matched WKY does not imply a vasodilator role for the high levels of this prostanoid. Indeed, in 36-wk-old SHR, prostacyclin no longer causes relaxation of aortic vascular smooth muscle but induces contraction (9, 27). The present findings demonstrate that the genomic expression of prostacyclin synthase is not altered in the normotensive rat by the course of aging.
Dazoxiben, a thromboxane synthase inhibitor, cannot prevent the occurrence of endothelium-dependent contractions to acetylcholine in the aorta of the SHR (2, 7, 17), excluding thromboxane A2 as an EDCF during exposure to the muscarinic agonist. However, thromboxane synthase was expressed at a higher level in 36-wk-old SHR than in WKY aorta. This augmented expression of thromboxane synthase may explain why thromboxane A2 contributes to EDCF-mediated responses triggered by agonists such as adenosine diphosphate, the calcium ionophore A-23187, and endothelin-1 (10, 11, 32). Aging resulted in an increase in the genomic level of thromboxane synthase. Thus thromboxane A2 may be an important mediator in age-dependent endothelium-dependent contractions in normotensive rats.
The genomic expression of hematopoietic-type prostaglandin D synthase was increased but that of lipocalin-type prostaglandin D synthase was decreased in 36-wk-old SHR compared with age-matched WKY. By contrast, aging increased the expression of hematopoietic-type prostaglandin D synthase but did not modify that of lipocalin-type prostaglandin D synthase. These differential changes in genomic expression make it unlikely that prostaglandin D synthase participates in endothelium-dependent contractions. Of the two isoforms, the hematopoietic type appears to be the dominant isoform in the endothelium, because its expression was persistently higher. Since prostaglandin D2 causes predominately relaxation via activation of DP receptors (21), a contribution of this subtype to endothelium-dependent contractions is also unlikely. However, prostaglandin D2 can play important roles in the prevention of platelet aggregation and tumors and contributes to the changes in stability of atherosclerotic plaques (1).
The cytosolic type of prostaglandin E synthase appeared to be the dominant form in the aortic endothelium of the rat, because its expression was higher than that of the membrane type. Aging and hypertension did not alter the genomic expression of cytosolic prostaglandin E synthase or membrane prostaglandin E synthase-1, but the expression of membrane prostaglandin E synthase-2 in WKY was altered by aging. Thus cytosolic prostaglandin E synthase and membrane prostaglandin E synthase-1 are unlikely to contribute to the greater endothelium-dependent contractions in the aorta of the hypertensive rats. However, membrane prostaglandin E synthase-2 may contribute to the augmentation of endothelium-dependent contractions in the aorta of the aging rat. Recent gene knockout studies demonstrated that cytosolic prostaglandin E synthase is not required for the generation of prostaglandin E2 in the mouse embryonic heart and liver, but it is required for the generation of prostaglandin E2 in the embryonic mouse lung (18, 24).
The present study attempted to estimate the contribution of the prostanoid receptors to EDCF-mediated responses by evaluating the presence and abundance of their genes in aortic smooth muscle cells. TP receptors are thought to be the primary prostanoid receptors in vascular smooth muscle cells responsible for endothelium-dependent contractions, since such contractions are abolished by TP receptor antagonists (2, 42). The observation that aging or hypertension does not alter the genomic expression of TP receptors implies that alterations in TP receptor number are unlikely to explain the enhanced occurrence of endothelium-dependent contractions in SHR. The amount of TP mRNA detected actually was low. In fact, the mRNA level of all prostanoid receptors was low compared with the copy number of other genes, implying either that the mRNA for prostanoid receptors is unstable or that prostanoid receptors have a low turnover rate.
Aging and hypertension did not modify the genomic expression of IP receptors. Despite the detection of the IP receptor in the rat aorta, prostacyclin does not induce relaxation of this artery (9, 27). Prostacyclin when formed in excess can cross-activate other prostanoid receptors, in particular TP receptors (9, 27). As more and more prostacyclin occupies TP receptors, contractions are initiated that are unmatched by the insensitive IP receptors.
EP1 and EP3 receptor activation causes inhibition of adenylyl cyclase and/or an increase of calcium mobilization, favoring contractions (1, 6). By contrast, EP2 and EP4 receptors are coupled to the activation of adenylyl cyclase and increase the concentration of cyclic adenosine monophosphate concentration, favoring relaxation (1, 6). The most abundant prostanoid receptor in the aorta of 36-wk-old SHR is the EP3 receptor, while EP4 is the dominant prostanoid receptor in normotensive rats. The present study thus is compatible with a contribution of prostaglandin E2 to endothelium-dependent contractions, which is released during EDCF-mediated response (9–11). Since aging did not increase the expression of EP3 receptors in the aorta of the normotensive rat, if prostaglandin E2 were an EDCF, this would help to explain the lower amplitude of the contractions observed in aortas of aged normotensive rats. The genomic expression of EP1 receptors was comparable in the different groups of rats studied; thus it is unlikely that changes in the density of this receptor subtype facilitate the occurrence of endothelium-dependent contractions in aging or hypertensive rats.
The present study demonstrates that the mRNA encoding for EP4 and DP receptors was upregulated in the 36-wk-old SHR compared with age-matched WKY and there was an age-dependent increase in the genomic expression of EP4 receptors. These receptors are coupled to the adenylate cyclase system, which leads to an increase in intracellular cyclic adenosine monophosphate concentration causing relaxation of the vascular smooth muscle (1, 6). Thus it is unlikely that these modifications facilitate the occurrence of endothelium-dependent contractions. The overexpression of these “relaxant” receptors in hypertensive or aging rats may serve as a counterregulatory system to maintain local vascular tone.
In summary, the genes for all known prostanoid synthases and prostanoid receptors were present in rat aortic endothelial and smooth muscle cells, respectively. Aging caused overexpression of eNOS, COX-1, COX-2, thromboxane synthase, hematopoietic-type prostaglandin D synthase, and prostaglandin F synthase in endothelial cells and COX-1 and EP4 receptors in smooth muscle cells. Hypertension accelerated the expression of COX-1, prostacyclin synthase, thromboxane synthase, and hematopoietic-type prostaglandin D synthase in endothelial cells and DP, EP3, and EP4 receptors in smooth muscle cells. The increase in the genomic expression of endothelial COX-1 explains why in aging and hypertension the endothelium has a greater propensity to release cyclooxygenase-derived vasoconstrictive prostanoids. The expression of prostacyclin synthase was exceedingly abundant in the endothelium of the rat, suggesting that the majority of the endoperoxides are transformed into prostacyclin and substantiating the role of this prostanoid as an EDCF. The present study focused its interpretation on the gene changes that could explain the alternations of EDCF-mediated responses due to aging and/or hypertension. However, it must be emphasized that it is unlikely that the reported changes of all these genes participate only in the changes in EDCF-mediated contractions, and they could also contribute to other important physiopathological alterations.
This work was supported by the Hong Kong Research Grant Council (University of Hong Kong-777507M).
Address for reprint requests and other correspondence: P. M. Vanhoutte, Dept. of Pharmacology, 2/F, Laboratory Block, Faculty of Medicine Bldg., 21 Sassoon Rd., Pokfulam, Univ. of Hong Kong, Hong Kong (e-mail:).
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
- Copyright © 2008 the American Physiological Society