INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE DUCTUS ARTERIOSUS (DA) performs two major functions: to remain patent during fetal life and to close rapidly after birth to separate the pulmonary and systemic circulations (20). Prostaglandins, particularly PGE₂, play a major role in maintaining the patency of the fetal DA (12, 14, 15). The relaxant effects of PGE₂ have been attributed to its ability to increase intracellular cAMP concentrations (17, 43). Immediately after birth, the response of DA to PGE₂ is markedly reduced (2, 11) thereby promoting DA closure. The mechanisms for this decreased responsiveness to PGE₂ are not well understood.

PGE₂ exerts its effects through a diverse group of receptors classified as EP₁, EP₂, EP₃, and EP₄ (16). Although pharmacological evidence suggests that EP₄ may be the main functional PGE₂ receptor in fetal rabbit DA (40), genetic disruption of this receptor does not induce DA closure in either fetal or newborn mice (34). Thus at present, the types of PGE₂ receptors that govern DA tone are uncertain. We (5) have recently found that the DA of the fetal pig expresses three EP receptor subtypes that would appear to have different effects on ductus contractile tone. We identified two cAMP-stimulating EP receptors (EP₂ and EP₄) and one cAMP inhibiting receptor (EP₃). In contrast, we detected only EP₂ in the newborn pig (5). However, it remains to be explained how loss of a cAMP-inhibiting (EP₃) and a cAMP-stimulating (EP₄) EP receptor can result in decreased responsiveness of the newborn DA to PGE₂.

Therefore we proceeded to determine the developmental profile of EP receptor expression in another species, the ovine, and to examine the effects of EP receptor stimulation on DA signaling events and contractile responses. Our findings reveal that the fetal lamb expresses, in equal proportions, the same three EP receptor subtypes detected in the fetal pig (5). Similarly, EP₂ is the only EP receptor identified by binding studies in the newborn ovine DA. Although stimulation of cAMP-generating EP₂ and EP₄ receptors resulted in the expected DA relaxation, surprisingly stimulation of cAMP-inhibiting EP₃ receptors also produced relaxation. This effect was mediated via a previously undescribed cAMP-independent pathway for EP₃ involving activation of ATP-sensitive K⁺ (K_ATP) channels. The loss of a relaxant EP₃ receptor in the newborn DA is consistent with decreased responsiveness of the DA to PGE₂ in the immediate neonatal period.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials. AH-6809, AH-23848B, and GR-63799X were generously provided by Dr. Simon Lister (Glaxo-Wellcome), butaprost by Dr. Harold Kluender (Bayer), M&B-28767 by Dr. Jean Hough (Rhone-Poulenc Rorer), and the PGI₂ analog (cicaprost) by Dr. Fiona McDonald (Schering, Berlin, Germany). 16,16-dimethyl-PGE₂ and carbaprostacyclin were purchased from Cayman Chemical (Ann Arbor, MI) and [³H]PGE₂ (165 Ci/mmol) from Amersham Pharmacia, Biotech (Mississauga, ON, Canada); all other chemicals were from Sigma (St. Louis, MO).

Tissue collection. Pregnant ewes were anesthetized with alternating intravenous injections of ketamine HCl (0.3 mg · kg¹ · min¹; Ketalar, Parke-Davies) and diazepam (0.01 mg · kg¹ · min¹; Valium, Hoffman-Laroche) and fetal lambs (mixed Western breed) delivered by cesarean section at 135 ± 3 days gestation (range, 125-140 days; term, 145 days). The fetus was given ketamine (30 mg/kg iv) before exsanguination to obtain DA. The same procedure was used to obtain DA from the newborn (<8 h after birth). Vessels were frozen immediately after removal with liquid N₂ and stored at 80°C. These procedures were approved by the Committee on Animal Research at the University of California, San Francisco.

Radioligand binding assays. [³H]PGE₂ binding and displacement studies were performed as described previously on DA membranes (5, 25). Frozen ductuses (with endothelium) were homogenized with a tissue grinder in 10 mM PBS buffer (pH 7.4) containing soybean trypsin inhibitor (1 mg/ml) and phenylmethylsulfonyl fluoride (PMSF, 5%). The homogenate was then centrifuged twice at 10,000 g for 15 min at 4°C to remove nuclei, undisrupted cells, and fibrous tissue. The combined supernatants were recentrifuged at 100,000 g for 90 min at 4°C. The membrane pellet was stored at 80°C (necessary to pool tissues) and used for receptor binding assay within 1 wk.

cAMP measurements. DA homogenates (100 µg protein) were incubated at 37°C for 10 min in an assay mixture (100 µL) containing: 10 mM Tris · HCl buffer (pH 8.0), 1 mM ATP, 7.5 mM MgCl₂, 15 mM creatine phosphate, 0.5 mM EGTA, 0.5 mM IBMX, 1 mM dithiothreitol, 1 mM benzamidine, 0.1 mM PMSF, and 185 U/ml creatine phosphokinase, 200 µg/ml aspirin, and 100 µg/ml soybean trypsin inhibitor (5, 25). The reaction was terminated with 200 µl of acidic ethanol. After centrifugation, cAMP was measured by radioimmunoassay as described by the manufacturer (Diagnostic Products). The assay is based on the competition between unlabeled cAMP and a fixed quantity of tritium-labeled compound ([³H]cAMP) for binding to an antiserum which has a high specificity for cAMP. The amount of [³H]cAMP bound to the antiserum is inversely related to the amount of cAMP present in the assay sample. Separation of the unbound nucleotide (including radiolabeled) is achieved by dextran-coated charcoal adsorption. The radioactivity of the supernatant (which contains the antibody-[³H]cAMP complex) is determined by liquid -scintillation counting. The concentration of unlabeled cAMP in the sample is then determined from a standard curve.

Isometric tension in vitro. The freshly collected DA was divided into 1-mm thick rings placed in separate 10-ml organ baths in a darkroom as previously described (12, 13). The rings were suspended between two stainless steel hooks at 38°C in a modified Krebs buffer (pH 7.4) of the following composition (in mM): 118 NaCl, 4.7 KCl, 2.5 CaCl₂, 0.9 MgSO₄, 1 KH₂PO₄, 11.1 glucose, and 23 NaHCO₃ (pH 7.4). The buffer was equilibrated with 5% CO₂-30% O₂-65% N₂. In some rings the luminal endothelium was removed by scraping the surface with a fine wire as previously described (13). The bath solution was changed every 20 min. Isometric tension was measured by Grass FTO3C force transducers (Quincy, MA). The tissues were equilibrated with 30% O_2, 65% N₂, and 5% CO₂ until the tension reached a plateau (~100-120 min). Indomethacin at 5.6 µM [a dose that was previously reported to cause maximal inhibition of PGE₂ or 6-keto-PGF₁ production in the ductus (10,13)] was then added to the bath solution, and the rings were allowed to reach a new steady-state tension over the next 60-90 min. The nitric oxide (NO) synthase inhibitor, N^G-nitro-L-arginine methyl ester (L-NAME) (0.1 mM) was then added to a dose previously found to cause the maximal inhibition of NO synthesis in the ductus (10, 13). The rings were exposed to indomethacin and L-NAME for the remainder of the study protocol. Maximal contraction was determined from the response to a 100 mM KCl that gave the maximal contraction of the ductus (13).

Effects of different agents on DA tone in vivo. Pregnant ewes (120-140 days of gestation) were anesthetized with ketamine (30 mg/kg iv). A cesarean section was performed and the fetus with intact placental circulation was exteriorized and its head submerged in warm saline to prevent breathing. Body temperature was maintained at 38.5°C by an overhead lamp. The femoral artery was cannulated for blood pressure recording using a pressure transducer (Gould, Valley View, OH), and the jugular vein was cannulated for the administration of drugs. Arterial blood gas was measured with an ABL300 blood gas analyzer (Radiometer, Copenhagen, Denmark) before and after each drug infusion.

Ultrasonography of DA. Real-time and Doppler echographic studies were performed with an Acuson 128 XP/10c real-time ultrasonographic imaging system that used 7.5- and 5-MHz transducers as previously described (6). Two-dimensional imaging of the DA was obtained through a left parasternal approach (second intercostal space), and the angulation of the transducer was such that the ultrasonographic beam was always parallel to or within 20 degrees of the orientation of the blood flow. The DA was preconstricted with indomethacin (0.75 mg/kg) to <50% of the original diameter. Vasorelaxant responses to sulprostone (EP₃ agonist) were determined by measuring the smallest diameter of the vessel on a two-dimensional representation of the echocardiogram before and every 5-10 min after the drug injections. Each measurement of the DA diameter was repeated twice and expressed in mm.

Preparation of total RNA, reverse transcription, and polymerase chain reaction. Total RNA was isolated from fetal DA tissue using the QIAGEN RNeasy mini kit (QIAGEN; Valencia, CA) according to manufacturer instructions. Total RNA (3 µg) was reverse transcribed with 100 units of Moloney murine leukemia virus Rnase H reverse transcriptase (Life Technologies) in the presence of random hexanucleotide primers as described elsewhere (34, 45). The cDNAs were used for the amplification of specific fragments of EP₂, EP₃, and EP₄ receptors by PCR following standard procedures (35 amplification cycles) (34, 45). To amplify a fragment composed of the transmembrane domain I and the second intracellular loop of the EP₂ receptor, the forward primer 5'-ATCTTGGGGTGGTGGGCAA-3' and the reverse primer 5'-CGCTTGTCCACGTAGTGGCT-3' were used. To amplify an EP₃ receptor fragment composed of the transmembrane domains IV and V, the forward 5'-GTGCTCGCCTTCGCCCTGTT-3 and reverse 5'-GCCTTGGCCCTGCAGCGGGA-3' primers were used. To amplify COOH-termini of the EP_3A,B,C isoforms the forward primer 5'-ATAATGATGTTGAAAATGAT-3' was used with the reverse primers R3 5'-CTACTGATGCTCAAGTGTATG-3' and R4 5'-GCCCCCTTCCTCTCCTTGCTT-3'. Primer R5 5'-ATTTCATTGGATAGTGAGATAGTC-3' was used to amplify the EP_3D COOH-terminal. The forward and reverse primers used to amplify the EP₄ receptor were 5'-AAGTCGCGCAAGGAGCAGAA-3' and 5'-CTTGTCCACGTAGTGGCTGT-3', respectively. The PCR products were subcloned into the pPCR-Script Amp SK(+) plasmid (Stratagene, La Jolla, CA) and sequenced with the use of an ABI prism 310 Genetic Analyzer sequencer (PE Applied Biosystems, Foster City, CA).

Statistical analysis. Data were analyzed by Student's t-test and by two-way ANOVA factoring for time or concentration and treatment. The Bonferroni correction was used for comparison among means. Statistical significance was set at P < 0.05. Data are expressed as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Competitive displacement of [³H]PGE₂ in the fetal DA. EP₂, EP₃, and EP₄ receptor ligands caused comparable displacement of specifically bound [³H]PGE₂ from fetal DA membrane preparations. Accordingly, the estimated density of EP₂, EP₃, and EP₄ {product of the proportion of each EP receptor deduced from competitive binding studies and B_max of EP receptors (total [³H]PGE₂ binding)} on fetal DA membranes was similar (Fig. 1B). The EP₁ receptor antagonist AH-6809 was virtually ineffective (Fig. 1, A and B, Table 1). IC₅₀ values were consistent with those previously reported in the literature (1, 16, 32).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Representative displacement curves of specifically bound [3H]PGE2 from lamb ductus arteriosus (DA) membrane preparations from fetal (A and B) and newborn (C and D) lambs. Estimated density of each EP receptor subtype (Bmax) (B and D) was calculated as total maximum binding to all EP receptor subtypes times the proportion of displacement caused by ligands selective to each EP receptor subtype. Values are means ± SE of 3-4 experiments. 16,16-dimethyl (DM)-PGE2, nonselective EP receptor agonist; AH-6809, EP1 selective antagonist; butaprost, EP2 selective agonist; M&B-28767, EP3 selective agonist; AH-23848B, EP4 antagonist.

Effects of different agents on cAMP production in the fetal DA. Both the nonselective EP receptor agonist 16,16-dimethyl-PGE₂ and the selective EP receptor agonist butaprost increased cAMP production in the fetal lamb DA (Table 2). In contrast, the selective EP₃ receptor agonist GR-63799X had no effect on cAMP production by itself but inhibited forskolin-stimulated cAMP production (Table 2). In the absence of currently available EP₄ agonists, the role of EP₄ was tested with the use of the EP₄ antagonist AH-23848B in the presence of 16,16-DM PGE₂. AH-23848B alone did not alter cAMP production but decreased 16,16-DM PGE₂- induced stimulation of cAMP formation (Table 2), which suggests that stimulation of EP₄ receptors accounts for some of the increased cAMP production induced by the nonselective EP receptor agonist.

Effects of different agents on fetal DA tension in vitro. The fetal DA contracted spontaneously (77 ± 8% maximal active tension) after addition of indomethacin and L-NAME (in the presence of 30% oxygen). As anticipated, agents that led to elevations in cAMP levels, namely forskolin and 8-Br-cAMP, relaxed the fetal DA (Fig. 2, A and B) with relatively low potency as previously described (37, 41). Analogs of PGI₂, the effects of which are also coupled to an increase in cAMP (16), relaxed the DA albeit with a potency <1,000 time less than PGE₂ (Fig. 2C, Table 3). Similarly, stimulation of EP receptors that increase cAMP relaxed the DA; the nonselective EP agonist PGE₂ and the selective EP₂ agonist butaprost relaxed the DA (Fig. 2, C and D), and inhibition of the EP₄ receptor with AH-23848B inhibited the relaxant effects of PGE₂ (Fig. 2E, Table 3). In contrast, although EP₃ stimulation decreased cAMP production (Table 3), activation of EP₃ receptors with M&B-28767 or GR-63799X caused relaxation of the (preconstricted) fetal DA by >70% (Fig. 2F). To assess whether effects of EP₃ were G protein-dependent, tissues were treated with pertussis toxin, which diminished the inhibitory effect of M&B-28767 on cAMP formation but did not alter its relaxant actions.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Vasomotor response of fetal ductus to prostaglandins, prostaglandin analogs, and cAMP-elevating agents. DA rings were precontracted with 30% oxygen, indomethacin (5.6 µM), and NG-nitro-L-arginine methyl ester L-NAME (0.1 mM). Difference in tension between the one induced by indomethacin + L-NAME and that generated after sodium nitroprusside (0.1 mM) was considered the net tension (15 ± 3 g, n = 120). Dose responses to forskolin, 8-bromoadenosine 3',5'-cyclic monophosphate (8-Br-cAMP) PGE2, PGI2 analogs cicaprost and carbaprotacyclin, butaprost, M&B-28767, and GR-63799X were studied; effects of PGE2 were also tested in the presence of EP4 antagonist AH-23848B. Tension is expressed as a percentage of the net tension. Values are means ± SE of 3-11 experiments.

Effects of PGE₂ analogs on the diameter of the fetal DA in vivo. We also tested the effects of EP₃ stimulation on DA tone in vivo. Infusion of the EP₃ receptor agonist sulprostone relaxed the indomethacin-induced contraction of the fetal DA in vivo (Fig. 3) consistent with our in vitro results (Fig. 2F); a similar effect was seen with 16,16-dimethyl-PGE₂ (data not shown).

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Effects of EP3 agonist sulprostone on percent change in DA diameter of fetal lambs in vivo. Animal preparation is described in MATERIALS AND METHODS. Briefly, the fetal lamb with intact placental circulation was exteriorized and its head submerged in warm saline to prevent breathing. The femoral artery and jugular vein were cannulated. DA diameter was measured by echocardiogram. DA was preconstricted with indomethacin (0.75 mg/kg iv) and once its diameter was <50% of the original diameter, sulprostone (0.083 µg · kg1 · min1 iv) was infused (indicated by shaded area along the abscissa). Values are means ± SE of 3 experiments. *P < 0.05 compared with presulprostone infusion.

Role of K+ channels on EP receptor-induced fetal DA relaxation. Because EP₃ receptor stimulation caused DA relaxation, despite reducing cAMP generation, we examined whether EP receptor stimulation might relax the DA through other signaling pathways. K⁺ channels have previously been shown to play a role in DA tone (30, 42). We performed PGE₂ dose-response curves in the presence and absence of specific K⁺ channel inhibitors. The relaxant effects of PGE₂ were not affected by the voltage-gated potassium- and calcium-activated potassium-channel blockers 4-aminopyridine and iberiotoxin, respectively (Table 3). However, the K_ATP channel blocker glibenclamide significantly inhibited the relaxation caused by the EP₃ agonist M&B-28767. Glibenclamide did not affect relaxation caused by agents that increase cAMP (butaprost, forskolin, and 8-Br-cAMP) (Table 3). Removal of luminal endothelial cells neither altered the PGE₂- or M&B-28767-induced relaxation of the DA nor did it modify the inhibitory effects of glibenclamide on PGE₂ and M&B-28767-induced relaxation (Table 4).

Detection of EP receptor subtypes and isoforms in the fetal DA by PCR. We amplified an EP₃ receptor fragment from the ovine DA that encompassed transmembrane domains IV and V of the receptor. Because its sequence was 100% homologous with the corresponding bovine sequence (31), we designed oligonucleotide primers on the basis of the bovine sequence to identify which of the different EP₃ receptor carboxyterminal isoforms might be present in the fetal DA (31). Only the EP_3D isoform fragment was successfully amplified from the ovine fetal DA. The PCR product sequence identified was 100% homologous with that of the bovine EP_3D (for relevant bases 895-1145 from the cDNA sequence) (31).

Competitive displacement of [³H]PGE₂: comparison between the fetal and newborn DA. In contrast with the fetus, the EP₂ receptor agonist butaprost was capable of displacing virtually all of the [³H]PGE₂ from newborn (<8 h old) DA membranes (Fig. 1, C and D, Table 1). EP₁, EP₃, and EP₄ receptors were essentially undetectable. The calculated B_max for the EP₂ receptor in the newborn was the same as that calculated for the EP₂ receptor in the fetus (Fig. 1, B and D).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Results presented help to identify the types of PGE₂ receptors that govern DA tone and how developmental changes in their expression may explain the decreased responsiveness of the DA to PGE₂ in the immediate newborn period. We have previously reported the presence of EP₂, EP₃, and EP₄ receptors in the porcine fetal DA, and the loss of EP₃ and EP₄ in the newborn (5). Stimulation of EP₃ and EP₄ led to opposite effects on cAMP formation (5). Decreases and increases in cAMP are usually associated with constriction and relaxation, respectively. Therefore, it was difficult to reconcile how the loss of receptors with apparently opposing actions could account for the changes in vasomotor tone that occurred in the DA immediately after birth (2, 11). The present study in the ovine DA demonstrates that these perinatal changes in EP receptor profile are not just limited to the pig. More importantly, the findings unveil a previously unreported vasomotor response to EP₃ stimulation, namely, a robust relaxation mediated by activation of K_ATP channels and not by increases in cAMP. The loss of two EP receptors coupled to relaxation (EP₃ and EP₄) in the early postnatal DA provides an explanation for its decreased responsiveness to PGE₂.

Stimulation of EP₂ and EP₄ receptors in the fetal lamb increased cAMP and relaxed the DA as expected (16). Despite reducing cAMP production, stimulation of EP₃ receptors also relaxed the DA (to an extent that was even greater than stimulation of EP₂). This EP₃-induced relaxation was shown in vitro as well as in vivo with the use of distinct selective EP₃ stimulants (M&B-28767, GR-63799X, and sulprostone) (1, 16). The relaxant effect of EP₃ agonists on the fetal DA was due, at least in part, to stimulation of a glibenclamide-sensitive K_ATP channel, whereas that of cAMP-elevating agents was not (Table 3). This failure of glibenclamide to inhibit relaxation by cAMP-elevating agents is not unique to the DA; relaxation of many (but not all) blood vessels to cAMP-elevating agents has been reported to be unaffected by glibenclamide (22, 23, 36). Altogether, relaxation of the DA to PGE₂ seems to be at least partly dependent on cAMP (EP₂ and EP₄) and K_ATP (EP₃) pathways.

The mechanisms responsible for the EP₃-induced, K_ATP-dependent relaxation of the DA are unclear. PGE₂ has been reported to be coupled to K_ATP or calcium-activated potassium channels in different tissues (3, 7); NO-mediated, cGMP-dependent kinase- (8, 19) and protein kinase A-mediated processes have been proposed (43). In our studies a role for protein kinase A appears to be unlikely because cAMP elevating agents (butaprost, forskolin, 8-Br-cAMP) did not relax the DA in a glibenclamide-sensitive manner. Similarly, a role for NO or endothelial derived prostanoids also appears to be unlikely because the relaxant responses of the EP₃ agonists were observed in DA tissues pretreated with NO synthase and cyclooxygenase inhibitors. Furthermore, removal of luminal endothelium did not modify either the EP₃-induced relaxation or its dependence on the K_ATP channel. The EP₃ receptor has numerous isoforms that couple to different G proteins (24, 31). These have the potential to elicit different vasoactive responses. We identified an EP₃ isoform in the ovine DA identical to the bovine EP_3D receptor (31) that can couple to G_i, G_s, and G_q (31). It is unlikely that the EP_3D receptor stimulates the K_ATP channel through a G_s-stimulated process: EP₃ stimulation inhibited rather than stimulated cAMP in the DA and cAMP did not relax the DA in a glibenclamide-sensitive manner (Tables 2, 3). In most tissues, the EP₃ receptor couples with G_i (16, 24, 31, 32), which is responsible for the inhibition of cAMP production; however, pertussis toxin, an inhibitor of G_i, blocked the effects of EP₃ stimulation on cAMP inhibition but did not affect the EP₃-induced relaxation. Hence, G_i may also not be involved in the EP₃-induced DA relaxation. Coupling of the receptors to G_q could lead to relaxation if they were present on the endothelium (9, 27, 39); however, this is not consistent with our findings because EP₃-induced relaxation was not inhibited by removal of the luminal endothelium. Possible explanations for this novel function of EP₃ include activation of K_ATP by NO-independent protein kinase G in response to EP₃ stimulation (44) or heterodimerization of EP₃ with the K_ATP channel as described with other G protein-coupled receptors (26). Extensive studies are in progress to address these issues.

Perspectives. In summary, total EP receptors in the lamb DA decrease after birth. The loss of both a vasodilating EP₄ receptor and a previously undescribed vasodilating EP₃ receptor could explain the loss of vasodilatory response of the DA to PGE₂ in the early newborn period. Present data in the ovine DA support those in the porcine DA (5). They also suggest that a selective EP₂ agonist may be more appropriate than the nonselective agonist PGE₁ for maintaining ductal patency in newborn infants afflicted with certain congenital cardiac malformations.