Introduction

The ductus arteriosus (DA), a fetal arterial connection between the pulmonary artery and the descending aorta, is indispensable for fetal life. Considerable evidence supports a major role for the cyclooxygenase (COX) product prostaglandin (PG) E2 in maintaining mammalian fetal patency and for the rapid decline in PGE2 levels as trigger of DA closure after birth (see Clyman 2006; Smith 1998 for reviews).

In blood vessels, PGE2 acts on four different prostanoid receptors: EP1, EP2, EP3 and EP4. Both EP2- and EP4-receptor subtypes mediate PGE2 relaxation through a cyclic AMP-dependent mechanism (Narumiya and FitzGerald 2001). EP1 is a constrictive receptor and EP3 induces a decline in cAMP, inhibiting smooth muscle relaxation (Narumiya and FitzGerald 2001). The main receptor responsible for PGE2-induced DA relaxation varies among mammalian species and was reported as EP4 in rabbits (Smith 1998; Smith et al. 1994; Smith and McGrath 1995), mice (Nguyen et al. 1997), rats (Kajino et al. 2004), lambs (Waleh et al. 2004), baboon (Waleh et al. 2004), and humans (Leonhardt et al. 2003), and as EP2 in pigs (Bhattacharya et al. 1999) and lambs (Bouayad et al. 2001). The cAMP-inhibiting EP3 receptor is also present in the rabbit and piglet DA, modulating the dilating effect of PGE2 (Bhattacharya et al. 1999; Smith 1998; Smith et al. 1994; Smith and McGrath 1995). In contrast, stimulation of EP3 receptor in the lamb DA produces not contraction but relaxation via a cAMP-independent pathway (Bouayad et al. 2001).

Significant progress in our understanding of the ductal effects of PGE2, as well as other aspects of DA physiology and pathophysiology, has been achieved with the use of mammalian models (Sutendra and Michelakis 2007). However, these models are technically complex and experimental manipulations affect both the mother and the fetus. Therefore, there is a need for additional models, addressing these limitations (Sutendra and Michelakis 2007). Recently, we analyzed the ontogeny of the chicken DA response to O2 and other vasoconstrictors (Agren et al. 2007), as well as the relaxation induced by agonists acting through the nitric oxide (NO)-soluble guanylate cyclase (sGC)-cyclic GMP (cGMP) pathway (Agren et al. 2008). We found that the effects of vasoactive mediators on chicken DA tone are developmentally regulated with loss of responsiveness to vasodilators and increase of responsiveness to vasoconstrictors with advancing incubation age.

Most of the physiological functions that prostanoids exhibit in mammals seem to be present in avian and other non-mammalian vertebrates (Rowley 1991). Prostanoids are produced in chicken blood vessels and participate in the regulation of vascular tone (Chand and Eyre 1976; Claeys et al. 1981; Wideman et al. 1999). However, we have observed that in the chicken DA COX inhibitors did not produce significant contractile effects (Agren et al. 2007) and did not affect endothelium-dependent relaxation (Agren et al. 2008), suggesting that locally produced PGs are not important mediators in the control of chicken ductal tone. Accordingly, Dzialowski and Greyner (2008) demonstrated that COX inhibition did not produce any change in the basal tension of DA of another avian species, the emu. However, they also observed that the emu DA was relaxed by exogenous PGE2 (Dzialowski and Greyner 2008). This indicates that PGE2, the most potent vasodilator in the regulation of mammalian DA tone, is also active in the avian DA. In the present study we hypothesized that the chicken DA is responsive to PGE2 and that its responsiveness is developmentally regulated. Therefore, we analyzed, at different ages, the ex vivo response of chicken DA to PGE2 and other vascular smooth muscle relaxing agents acting through the cAMP signaling pathway.

Material and methods

Embryos incubation and vessel isolation

Experiments were performed in accordance with Dutch law for animal experimentation. Fertilized eggs of White Leghorn chickens were incubated at 37.8°C, 45% humidity and rotated once per hour (Incubator model 25HS, Masalles Comercial, Spain). Embryos incubated for 15, 19 and 21 days of the 21-day incubation period were studied. The majority of the experiments were performed in 19-day (non-internally pipped) embryos, while for the study of developmental changes all three ages were compared (see below). On the experimental day the embryos were killed by decapitation and the right DA was carefully dissected free and severed distal to the takeoff of the right pulmonary artery and proximal to the insertion into the dorsal aorta. The average in situ length (in mm) of the right DA was 3.48 (SD 0.27), 4.71 (SD 0.40), and 5.58 (SD 0.40) at embryonic day 15, 19, and 21, respectively, (Agren et al. 2007), whereas the maximal length allowed in the myograph was 2 mm. Therefore, a similar length of ductal tissue was removed from the pulmonary and aortic ends of the vessel to obtain a segment of the central part of the DA with a length of approximately 2 mm. The right DA was selected because its shorter length allowed having a higher proportion of the vessel represented in the myograph.

Recording of arterial reactivity

Two stainless steel wires (diameter 40 μm) were inserted into the lumen of the DA, which was mounted as a ring segment between an isometric force transducer and a displacement device in a myograph (Danish Myo Technology A/S model 610 M, Aarhus, Denmark). The myograph organ bath (5 mL vol) was filled with Krebs–Ringer bicarbonate (KRB) buffer maintained at 39°C and continuously aerated with one of the following gas mixtures: 0% O2/95% N2/5% CO2 (pO2 = 2.6–3.3 kPa), 5% O2/90%N2/5% CO2 (pO2 = 6.8–7.2 kPa), or 95% O2/5% CO2 (pO2 = 72–76 kPa). The final pH was 7.38–7.42 and pCO2 was 4.6–5.6 kPa in all solutions. Each DA was stretched to its individual optimal lumen diameter, i.e., the diameter at which it developed the strongest contractile response to 62.5 mM K+, using a diameter-tension protocol as previously described (Agren et al. 2007; Villamor et al. 2002). During the first phase of stabilization and determination of optimal diameter, DA rings were maintained in KRB buffer with a pO2 of ~3 kPa. Afterward, oxygen concentration was switched to 5% (pO2 ~7 kPa), unless the effect of O2 on the relaxant responses was evaluated. When two or more agonists were studied in the same arterial preparation, the vessels were repeatedly washed and allowed to equilibrate for at least 30 min. If the tone did not recover to resting level or 62.5 mM K+ failed to induce an adequate level of contraction, the vessels were discarded for further experiments.

The effects of PGE2 (0.1 nM–10 μM), the direct activator of adenylate cyclase forskolin (0.1–10 μM), the β-adrenoceptor agonist isoproterenol (1 nM–10 μM) and the phosphodiesterase 3 (PDE3) inhibitor milrinone (0.1 μM–0.1 mM) were evaluated in DA rings contracted with K+ (62.5 mM), or the thromboxane (Tx)A2-mimetic (selective TP-receptor agonist) U46619 (0.1 μM). When stable contractions were obtained, agonists were added cumulatively to the bath until a maximal response was achieved. In some experiments, the relaxation evoked by PGE2 was studied in the presence of the TP-receptor antagonist SQ29,548 (10 μM), or the non-selective EP receptor antagonist AH6809 (6-isopropoxy-9-oxoxanthene-2-carboxylic acid; 10 μM) and the relaxations evoked by isoproterenol and forskolin were studied in the presence of the non-selective COX inhibitor indomethacin (10 μM), the COX-1 inhibitor valeryl salicylate (0.5 mM), or the COX-2 inhibitor nimesulide (0.1 μM).

In order to analyze developmental changes on relaxation, the responses to PGE2, forskolin, isoproterenol, and milrinone were studied in the central part of DA from 15-, 19-, and 21-day-old embryos. The 21-day embryos were externally pipped for about 1-3 h. In another set of experiments, aimed to evaluate the modulatory role of O2 in DA relaxation, the responses to PGE2, were studied under three different conditions of pO2: ~3, 7, and 74 kPa. In these experiments, optimal diameter was determined at 3 kPa. Thereafter the O2 concentration was switched to the appropriate level and the vessels were allowed to equilibrate for at least 30 min before the concentration- response curve for the agonists was obtained. Finally and in order to analyze the differences along the DA in the responses to PGE2, forskolin, isoproterenol, milrinone and U46619, we performed a third set of experiments in which the pulmonary and the aortic parts of the vessel were compared in 19-day embryos. The responses to U46619 and milrinone were also analyzed in the pulmonary and the aortic parts of the DA from 15-day embryos. The boundary between pulmonary and aortic segments was determined during the dissection based on the marked differences of diameter observed along the chicken DA (see Agren et al. 2007).

Data analysis

Results are shown as mean (SD) of measurements in n embryos. For clarity, results are shown in the figures as mean ± SE. Contractions are expressed in terms of active wall tension (N/m), calculated as the force divided by twice the length of the segment, while the relaxant responses are expressed as the percentage of reduction of the contraction induced by K+, or U46619. Sensitivity (expressed as pD2 = −log EC50) and maximal relaxation (E max) to agonists was determined for each artery by fitting individual concentration–response data to a non-linear sigmoidal regression curve and interpolating (Graphpad Prism version 2.01; GraphPad Software Inc, San Diego, CA, USA). When a maximal response was not achieved with a given agonist, the regression curve could not be calculated and the responses for each concentration of the agonist were used for comparison. Differences between mean values were assessed by one-way ANOVA followed by Bonferroni post hoc t test. Non-paired t tests were used if only two groups were compared. Differences were considered significant at a P < 0.05. All analyses were performed using a commercially available statistics package (GraphPad InStat version 3.00; GraphPad Software Inc, San Diego, CA, USA).

Drugs and solutions

PGE2, U46619, SQ29,548, and AH6809 were obtained from Cayman Chemical (Ann Arbor, MI, USA). Valeryl salicylate, nimesulide, and milrinone were obtained from Alexis Biochemicals (Lausen, Switzerland). All the other drugs were obtained from Sigma Chemical Co (St Louis, MO, USA). PGE2, U46619, and AH6809 were dissolved in DMSO and SQ29,548 indomethacin, valeryl salicylate, nimesulide and milrinone in ethanol to prepare adequate stock solutions and further dilutions were made in KRB. The final bath concentration of DMSO and ethanol did not exceed 0.1%, which we have found elsewhere to have little or no effect on mechanical activity (Agren et al. 2007). All the other drugs were dissolved initially in distilled deionised water.

Results

Potassium-induced contraction

Isolated DA obtained from 15-, 19- and 21-day chicken embryos responded to depolarizing high-K+ (62.5 mM) solution with a tonic contraction and this response significantly increased with age (15-day: 0.21 N/m, SD 0.05, n = 30, P < 0.001 from both 19 and 21-day; 19-day: 0.61 N/m, SD 0.19, n = 69; 21-day: 0.99 N/m, SD 0.37, n = 21, P < 0.001 from 19-day old.). The diameter at which a maximal response to 62.5 mM KCl was obtained increased between day 15 (787 μm, SD 89, n = 30) and 19 (830 μm, SD 121, n = 69, P < 0.001 vs.15-day), but decreased at day 21 (483 μm, SD 227, n = 21, P < 0.001 vs.15- and 19-day).

Effects of PGE2 in chicken DA

In the first set of experiments we tested the relaxant effects induced by PGE2 in chicken DA stimulated with KCl 62.5 mM. We found that PGE2 exhibited a relaxant response in 15-day DA (Figs. 1a, 2a; Table 1), but no significant relaxation was observed in 19- and 21-day DA (Figs. 1b, 2a; Table 1). Moreover, high concentrations of PGE2 (≥3 μM in 15-day and ≥1 μM in 19- and 21-day DA) induced contraction of the chicken DA (Figs. 1a, b, 2a). The presence of the TP receptor antagonist SQ29,548, unmasked a relaxant effect of PGE2 in the 19- (Figs. 1c, 2b; Table 1) and 21-day DA (Fig. 2b; Table 1) and increased the relaxation induced by PGE2 in the 15-day DA (Fig. 2b; Table 1). Under these conditions (i.e., presence of SQ29,548), the maximal relaxant response induced by PGE2 was reduced in term (externally pipped 21-day) versus preterm (non-internally pipped 15-day and 19-day) chicken DA, while the sensitivity to PGE2 was not affected by age (Table 1). When the 19-day DA was contracted with the TP receptor agonist U46619 (0.1 μM), PGE2 induced a concentration-dependent relaxation with a threshold concentration of 1 nM (Figs. 1d, 2d) and without achieving the maximal effect with the highest concentration tested (10 μM). The presence of the EP receptor antagonist AH6809 reduced the contraction induced by U46619 from 0.8 N/m (SD 0.11, n = 6) to 0.58 N/m (SD 0.12, n = 6, P < 0.01 vs. control) and abolished PGE2-mediated relaxation of U46619-contracted DA (Fig. 2d). As shown in Fig. 2c, bubbling the organ chamber with 95% O2 (pO2 ~74 kPa) produced a marked reduction of the relaxation evoked by PGE2 in 19-day DA (contracted with 62.5 mM K+ and in the presence of SQ29,548).

Fig. 1
figure 1

Typical tracings illustrating the vascular effects of prostaglandin E2 (PGE2) in chicken DA. During contraction induced by K+ (62.5 mM), PGE2 induced a concentration-dependent relaxation of the 15-day DA (a) but a contraction of the 19-day DA (b). In the presence of the TP receptor antagonist SQ 29,548 (10 μM, c) or when the vessel was contracted with the TP receptor agonist U46619 (0.1 μM, d), a relaxant effect of PGE2 was observed in the 19-day DA

Full size image
Fig. 2
figure 2

Concentration-dependent relaxant effects of prostaglandin E2 (PGE2,) in chicken DA during contraction induced by K+ (62.5 mM, a), K+ in the presence of the TP receptor antagonist SQ 29,548 (b and c), or the TP receptor agonist U46619 (0.1 μM, d). The organ chambers were areated with 5% O2/90%N2/5% CO2 (a, b, d). The effect of different oxygen concentrations on the relaxations induced by PGE2 in the 19-day DA is depicted in c, in which the DA rings were aerated with 0% O2/95% N2/5% CO2, or 95% O2/5% CO2. For clarity, the curve to PGE2 in 5% O2 from panel B is repeated in c (dashed line). The effects of the EP receptor antagonist AH6809 (10 μM) on PGE2-induced relaxation is depicted d (19-day DA). * P < 0.05 for difference from the other ages (a, b) and for difference from other oxygen concentrations (c)

Full size image
Table 1 Concentration-response parameters for relaxant substances in chicken ductus arteriosus
Full size table

Relaxant effects of isoproterenol, forskolin, and milrinone in chicken DA

In K+-stimulated DA, the addition of the β-adrenoceptor agonist isoproterenol, the direct activator of adenylate cyclase forskolin, or the PDE3 inhibitor milrinone caused a concentration-dependent relaxation (Fig. 3a, c, e; Table 1). As expected the β-adrenergic receptor blocker propanolol drastically reduced the relaxation response induced by isoproterenol (Fig. 3a). Similar to PGE2, the relaxation induced by isoproterenol was markedly diminished in 21-day versus 15- or 19-day chicken DA (Fig. 3a). In contrast, the relaxant effect of forskolin did not change with age (Fig. 3c; Table 1) and milrinone was less potent (but similarly efficacious) in relaxing the 15-day than the 19-, or 21-day chicken DA (Fig. 3e; Table 1). Finally, as shown in Fig. 3b and d, the relaxation evoked by isoproterenol and forskolin were not significantly affected by the presence of the COX inhibitors indomethacin, valeryl salicylate, or nimesulide, compared to control vessels (incubated with vehicle, ethanol 0.1%).

Fig. 3
figure 3

Concentration-dependent relaxant effects of isoproterenol (a, b), forskolin (c, d) and milrinone (e) in 15, 19 and 21-day DA rings aerated with 5% O2/90%N2/5% CO2. The effects of the β-adrenergic receptor blocker propanolol (prop) on isoproterenol-evoked relaxation are shown in a. b, d Depict the effects of the nonselective COX inhibitor indomethacin (10 μM), the COX-1 inhibitor valeryl salicylate (Val. salic., 0.5 mM), or the COX-2 inhibitor nimesulide (0.1 μM) on the relaxations induced by isoproterenol (b) and forskolin (d) in 19-day DA. Potassium (62.5 mM) was used as precontractile agent. * P < 0.05 for difference in E max from the other ages. ‡ P < 0.05 for difference in E max from 19 day. † P < 0.05 for difference in pD2 from the other ages

Full size image

Heterogeneity of relaxant responses along the DA

In the last set of experiments the responses to the drugs studied above were compared in the pulmonary versus the aortic side of the DA. The contractions induced by KCl were not significantly different when the pulmonary and the aortic side of the DA from the 19-day (pulmonary: 0.72 N/m, SD 0.16, n = 14; aortic: 0.67 N/m, SD 0.2, n = 14) and 15-day embryos (pulmonary: 0.21 N/m, SD 0.03, n = 6; aortic: 0.24 N/m, SD 0.05, n = 6) were compared. Similarly, the optimal diameters were not significantly different between the pulmonary and the aortic side of the 19-day DA (pulmonary: 889 μm, SD 88, n = 14; aortic: 936 μm, SD 87, n = 14) and 15-day DA (pulmonary: 742 μm, SD 38, n = 6; aortic: 792 μm, SD 49, n = 6). As shown in Fig. 4a, the relaxation induced by PGE2 (vessels were contracted with 62.5 mM K+ in the presence of SQ29,548) was similar in both sides of the DA. In contrast, the relaxing potency and efficacy of isoproterenol (Fig. 4b) was significantly higher (P < 0.05) in the pulmonary (pD2 7.71, SD 0.26, E max 117.95%, SD 8.3, n = 6) than in the aortic side (pD2 7.37, SD 0.15; E max 72.77%, SD 3.1, n = 6). Forskolin (Fig. 4c) also induced a higher relaxation in the pulmonary than in the aortic side of the 19-day DA, but significant differences were only observed with the highest concentration tested (10 μM). Similarly to isoproterenol, the relaxing potency and efficacy of milrinone (Fig. 4d) was significantly (P < 0.05) higher in the pulmonary (pD2 5.96, SD 0.16, E max 80.2%, SD 9.8, n = 6) than in the aortic side (pD2 5.0, SD 0.16; E max 59.2%, SD 14, n = 4). The effects of milrinone were also studied in the 15-day DA and a similar significantly higher relaxation was observed in the pulmonary (pD2 5.4, SD 0.18, E max 77.1%, SD 9.4, n = 6) than in the aortic side (pD2 4.9, SD 0.17, E max 47.2%, SD 11, n = 6). Finally, as shown in Fig. 4e, the contractile potency and efficacy of U46619 was also significantly higher (P < 0.05) in the pulmonary than in the aortic side of the 19-day (pulmonary: pD2 6.52, SD 0.15, E max 1.58 N/m, SD 0.61, n = 6; aortic: pD2 6.1, SD 0.13, E max 0.34 N/m, SD 0.18, n = 6) and the 15-day (pulmonary: pD2 6.84, SD 0.185, E max 0.33 N/m, SD 0.24, n = 6; aortic: pD2 5.3, SD 0.16, E max 0.08 N/m, SD 0.06, n = 6) chicken DA.

Fig. 4
figure 4

Heterogeneity of relaxant and contractile responses along chicken ductus arteriosus (DA). Response of segments from the pulmonary and aortic side of 19-day chicken DA to PGE2 (a), isoproterenol (b), and forskolin (c), milrinone (d), and U46619 (e). The responses to milrinone and U46619 in the pulmonary and aortic side of 15-day chicken DA are also shown. Relaxations were studied during contraction induced by K+ (62.5 mM). The TP receptor antagonist SQ 29,548 (10 μM) was present in the experiments involving PGE2. The vessels were aerated with 5% O2/90%N2/5% CO2 (pO2 ~7 kPa). * P < 0.05 for difference in E max aortic versus pulmonary side. ‡ P < 0.05 for difference in the response to a given concentration. † P < 0.05 for difference in pD2 aortic versus pulmonary side

Full size image

Discussion

PGE2 plays a major role in prenatal patency and postnatal closure of the mammalian DA, namely during fetal life by exerting a potent relaxant effect (Clyman 2006; Smith 1998), possibly in conjunction with nitric oxide (NO, Coceani et al. 1994; Momma and Toyono 1999; Seidner et al. 2001; Takizawa et al. 2000), and after birth by abruptly withdrawing its action. We have previously shown that the transition to ex ovo life is accompanied by dramatic changes in chicken DA reactivity (Agren et al. 2007, 2008). In the present paper, we have characterized, for the first time, the effects of PGE2 and other drugs acting through the adenylate cyclase/cAMP pathway on chicken ductal tone. Our main findings are: (1) PGE2 induced an EP receptor-mediated relaxation of the chicken DA that was partially or wholly masked by a simultaneous contractile effect mediated through TP receptors; (2) the relaxant responses induced by PGE2 and the β-adrenoceptor agonist isoproterenol, but not those elicited by the adenylate cyclase activator forskolin or the PDE3 inhibitor milrinone, decreased with maturation, (3) relaxations to PGE2 are decreased with increasing oxygen concentrations and (4) the effects of PGE2 were similar in the pulmonary and aortic sides of the DA, whereas isoproterenol and milrinone caused a more efficient relaxation in the pulmonary side of the chicken DA.

As mentioned elsewhere, COX inhibitors did not induce significant changes in the basal tension of the chicken (Agren et al. 2007) or the emu (Dzialowski and Greyner 2008) DA and did not alter the vasoactive effects of other agonists in the chicken DA (Agren et al. 2007, 2008), suggesting that locally produced PGs are not important mediators in the control of ductal tone in these avian species. However, Dzialowski and Greyner (2008) observed that exogenous PGE2 relaxed the oxygen-constricted emu DA with a threshold concentration of 0.1–1 μM and a full relaxation at 10 μM, the highest concentration tested. In the present paper, we demonstrate that exogenous PGE2 is indeed a DA vasodilator in the chicken but its relaxant response is masked by its ability to activate TP receptors. We found that TP receptor occupancy either by an agonist (U46619) or an antagonist (SQ29,548) results in the appearance or the increase of the relaxant response to PGE2. This is consistent with a stimulatory action of PGE2 on TP receptors. The PGE2 promiscuous interaction with TP receptors, especially at concentrations higher than 1 μM, has also been observed in several vascular tissues including human umbilical arteries (Boersma et al. 1999) and veins (Daray et al. 2003), human uterine arteries (Baxter et al. 1995) and aorta of the spontaneously hypertensive rat (Tang et al. 2008). An alternative explanation for the effects of SQ29,548 on PGE2-induced relaxation might be the blockade of a baseline TxA2- or isoprostane-induced tone (Gonzalez-Luis et al. 2005), allowing greater relaxatory actions of PGE on its dilatory receptors. However, this possibility is unlikely, since SQ29,548 have no effect on DA mechanical activity, suggesting the lack of basal release of a TP receptor stimulator.

PGE2-induced relaxation of the chicken DA was abolished by AH6809 suggesting the involvement of EP receptors. However, AH6809 is a non selective EP receptor antagonist, which may also block DP and TP receptors (Tang et al. 2008). Accordingly, we observed that AH6809 impaired the contraction induced by U46619. Future functional experiments with more selective EP antagonists, as well as investigations addressing the expression and binding activity of prostanoid receptors in the chicken DA are, therefore, warranted. Nevertheless, we observed that the relaxant potency and efficacy of PGE2 in the chicken DA, even after blocking its effect on the TP receptor, was very low. Dzialowski and Greyner (2008) also described a low relaxant potency of PGE2 in the emu DA. In contrast, PGE1 and PGE2 in the picomolar and low nanomolar range relaxed the rat, rabbit, pig and lamb DA (Bhattacharya et al. 1999; Clyman et al. 1983, 1980; Smith 1998; Waleh et al. 2004), while the only mammalian species studied where PGs have not been shown to exert a significant DA relaxation is the guinea pig (Bodach et al. 1980; Smith 1998). With our present results we can only speculate about the functional consequences of the low relaxant potency of PGE2 in the chicken DA. In the mammalian fetus, circulating concentrations of PGE2 appear to be of placental origin (Clyman 2006). To the best of our knowledge, there is no study in the literature examining prostanoids levels in avian embryos or the role of the choriallantoic membrane, the analogous of mammalian placenta, in the production of prostaglandins. These issues remain to be investigated.

In mammals, the immature fetal DA is more sensitive than the late-gestation fetal DA to the relaxing effects of PGE2 (Bhattacharya et al. 1999; Clyman et al. 1983, 1980; Waleh et al. 2004). Herein, we have found that chicken DA is more sensitive to PGE2 in preterm than in term. The activation of TP receptors by PGE2 is, at least partially, responsible for these changes with age. Thus, in the absence of TP antagonist a relaxant response is observed in the 15-day DA, whereas no relaxation, but contraction, is found in the 19-day DA. However, in the presence of the TP antagonist the relaxant responses to PGE2 are similar in 15 and 19 days. These results indicate that the contribution of TP receptor activation in the response to PGE2 increases during maturation. Accordingly, we have previously demonstrated a developmental increase in the sensitivity of chicken DA to the specific TP agonist U46619 (Agren et al. 2007). In addition, from 19- to 21-day there is an additional reduction in the relaxant responses to PGE2 when SQ29,548 is present (i.e., independent of TP receptor activation). Similar to PGE2, the β-adrenoceptor agonist isoproterenol induced less relaxation in 21-day when compared with 15- or 19-day DA. PGE2 and isoproterenol induced relaxation, at least partially, through the adenylate cyclase/cAMP pathway. Therefore, our results suggest a developmental decrease in the relaxant efficacy of this pathway during the transition to ex ovo life. Nevertheless, direct activation of adenylate cyclase with forskolin induced a similar relaxation at all stages of incubation. Thus, the developmental differences in the responses to PGE2 and isoproterenol from 19- to 21-day seem to be due to changes in the coupling between their receptors and adenylate cyclase. Accordingly, Waleh et al. (2004) found that the EP receptor agonists PGE2 (nonselective), butaprost (EP2 selective) and M&B 28767 (EP3 selective) produced a greater increase in cAMP production in the immature than in the mature lamb DA. However, there were no differences between the immature and mature ductus in the relaxation or cAMP production induced by forskolin (Waleh et al. 2004). In the emu, the other avian DA that has been studied, Dzialowski and Greyner found similar PGE2-relaxations in the day 45 and day 49 DA (total incubation in the emu: 50 days) (Dzialowski and Greyner 2008). Unfortunately, they did not analyze the response to PGE2 of the DA of less mature emu ductus.

The role of β-adrenoceptors in the control of mammalian DA tone also varies among species (Smith 1998). Infusion of the β-adrenoceptor antagonist propanolol had no effect on ductal tone in the in vivo lamb (Friedman et al. 1983). Furthermore, in the presence of the α-adrenoceptor antagonist phenoxybenzamine, norepinephrine was able to relax lamb DA preparations, but the effects were modest even at the highest doses (Bodach et al. 1980). In the newborn rat, propanolol paradoxically delayed DA closure (Arishima et al. 1995), but it has been suggested that this is secondary to the effects of β-adrenergic blockade on the perinatal cardiovascular homeostasis and not due to a direct effect on ductal tone (Smith 1998). On the other hand, in the guinea pig DA, phenoxybenzamine converts the contractile effect of norepinephrine to a relaxation and isoproterenol causes a marked relaxation (Bodach et al. 1980). These results together with the lack of a prostaglandin-mediated relaxation could point to a special role for β-adrenoceptors in prenatal patency in the guinea pig. Herein, we have observed a similar pattern in the chicken DA. Previously, we demonstrated that the nonselective adrenergic receptor agonist norepinephrine and the α1-adrenegic receptor agonist phenylephrine induced a developmentally increased contraction of the chicken DA. Secretion of catecholamines plays an important role in several of the adaptations that characterize the transition of the chicken to ex ovo life (Mulder et al. 2000; Wittmann and Prechtl 1991). The augmentation of the α-adrenergic-mediated contraction and the concomitant reduction of the β-adrenergic-mediated relaxation suggest an active participation of catecholamines in the closure of the chicken DA.

The cyclic nucleotides cAMP and cGMP are considered the main intracellular secondary messengers involved in smooth muscle relaxation (Omori and Kotera 2007; Rybalkin et al. 2003). These molecules are degraded by a family of enzymes known as phosphodiesterases (PDEs). There are now known to be 11 different PDE families expressed in mammalian tissues. Depending on the species, the major PDEs present in vascular smooth muscle are PDE1, PDE3 and PDE5 (Moreno et al. 2004; Omori and Kotera 2007; Rybalkin et al. 2003). PDE3 s are termed cGMP-inhibited cAMP PDEs, whereas PDE5 specifically hydrolyzes cGMP (Moreno et al. 2004; Omori and Kotera 2007; Rybalkin et al. 2003). We have demonstrated that the chicken DA is relaxed by the PDE5 inhibitor sildenafil (Agren et al. 2008) and by the PDE3 inhibitor milrinone (present work). When clinically used in premature infants, it appeared that PDE3 inhibitors inhibited closure of the DA (Toyoshima et al. 2006). Toyoshima et al. (2006) demonstrated that the PDE3 inhibitors milrinone and amrinone exhibited dose-dependent DA-dilating effects in rat fetuses and pups and were more effective in the preterm than in the near-term period. Liu et al. (2008) observed, in the ovine DA, a developmental increase in the activity and expression of PDE1, 3, 4, and 5 and reported that the mature ductus required higher concentrations of PDE1, PDE3, and PDE4 inhibitors to inhibit its tension to the same extent as in the immature ductus. In contrast, we have observed a reduced relaxant effect of milrinone in the 15-day chicken DA when compared with the 19- or 21-day DA. This reduced response to milrinone in the 15-day DA was homogeneously present when segments of the central part, the pulmonary end or the aortic end of the DA were analyzed (see below). Our results might suggest that an increase in PDE3-mediated cAMP hydrolysis is involved in the age-related decrease of PGE2- and isoproterenol-induced relaxation observed in the chicken DA. However, if that was the case, the relaxation induced by forskolin should also have exhibited a reduction with maturation.

Another observation of the present work was that hyperoxia markedly impaired PGE2-induced relaxation. Previously, we reported that high oxygen concentrations decreased the endothelium- dependent and -independent relaxation induced by, respectively, acetylcholine and the nitric oxide donor, sodium nitroprusside and increased the contraction mediated by adrenergic agonists (Agren et al. 2007, 2008). Accordingly, it has been demonstrated that increasing oxygen tension from fetal to neonatal levels desensitizes the rabbit DA to a range of vasodilators, including PGE2 (Smith and McGrath 1991, 1993). Thus the postnatal increase in O2 is not only one of the main triggers for constriction of the DA but also has a profound modulatory effect on other vasoactive systems favoring the action of vasoconstrictors and decreasing the action of vasodilators (Smith 1998).

The chicken DA presents morphological and functional heterogeneity along its path between the pulmonary artery and the aorta (Agren et al. 2007; Belanger et al. 2008; Bergwerff et al. 1999, 1996). Thus, the pulmonary side shows the structure of a muscular artery and responds to O2 with contraction, whereas the aortic part shows the morphology of an elastic artery and relaxes in response to O2 (Agren et al. 2007; Belanger et al. 2008; Bergwerff et al. 1999, 1996). In addition, acetylcholine, SNP and the NO-independent stimulator of sGC BAY 41-2272 induced larger relaxations in the aortic side of the vessel (Agren et al. 2008). In the present work, we observed that isoproterenol, forskolin, and milrinone evoked significantly larger relaxations in the pulmonary than in the aortic side of the chicken DA. This might indicate that the pulmonary side of the vessel is more sensitive to the vasodilators acting through cAMP, whereas the aortic side would be more sensitive to cGMP-mediated relaxation. However, and surprisingly, the relaxation evoked by PGE2 was similar in both regions of the DA. We can only speculate about the nature of this finding. As the contraction induced by the TP receptor agonist U46619 was significantly higher in the pulmonary side, the masking effect of PGE2-induced TP stimulation might be smaller in the aortic than in the pulmonary side. However, the experiments were performed in the presence of TP antagonism. Another speculation might be that cAMP-mediated mechanisms are less involved in the relaxation evoked by PGE2 in the aortic side. In fact, it has been demonstrated that PGE2 dilates the late gestation fetal lamb DA through pathways that involve either cAMP (via EP2 and EP4 receptors) or ATP-dependent potassium channels (via EP3, Bouayad et al. 2001). However, in the experiments that compared the pulmonary and the aortic side of the chicken DA, the vessels were contracted with a high potassium solution which precluded the relaxation induced by potassium channels activation.

The heterogeneity of the chicken DA makes the vessel particularly attractive for the study of its developmental vascular biology but also poses a methodological problem. As we mentioned elsewhere, the maximal length of the rings allowed in the myographs that we used is 2 mm, whereas the length of the chicken DA is ~4–6 mm. Although the pulmonary and the aortic side of the ductus can be clearly differentiated during the dissection (Agren et al. 2007), when mounting the central portion of the DA it is difficult to control for the exact amount of tissue of each part that is finally present in the myograph. In addition, putative developmental changes in the proportion between the two parts of the DA, might influence developmental changes in the responsiveness to the different agonists. This limitation of our experimental setting was particularly relevant when the response to milrinone was analyzed. We observed that milrinone-induced relaxation increased with development but also was higher in the pulmonary than in the aortic side of the DA. Therefore the reduced relaxation in the 15-day DA could be related to the presence of higher amounts of tissue from the aortic side of the vessel. We controlled for this possibility by analyzing the response to milrinone in the pulmonary and aortic sides of the 15-day DA and we observed that the developmental increase was present in both sides of the vessel.

In summary, with the present and our previous studies (Agren et al. 2007, 2008), we have completed an extensive characterization of the pharmacology of the chicken DA and demonstrated its sensitivity to a wide range of vasoactive agonists including oxygen, prostanoids, potassium channels blockers, NO, catecholamines, endothelin-1, adenylate cyclase activators, guanylate cyclase activators and PDE inhibitors. As occurs in the mammalian, the multiplicity of these vasoactive systems seems at odds with the relatively simple physiological role of the DA (Smith 1998). The main vasoconstrictor of the mammalian DA, the postnatal increase in oxygen tension, appears to play also a relevant role in the closure of the chicken DA (Agren et al. 2007; Belanger et al. 2008), whereas the main vasodilator of the mammalian DA, PGE2, is a weak relaxant agent of the chicken DA which even stimulates vasoconstrictive receptors. As the mammalian DA, the chicken DA undergoes a process of maturation to prepare the task of postnatal closure. Due to the above similarities and taking into account the differences, the chicken embryo appears as an extremely interesting model for translational developmental vascular biology (Sutendra and Michelakis 2007).