INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

At birth, successful transition from gas exchange by the placenta to gas exchange by the lung depends upon a dramatic decrease in pulmonary vascular resistance (PVR). In some newborn infants the normal decrease in PVR does not occur, resulting in persistent pulmonary hypertension of the newborn (PPHN). Surgical ligation or constriction of the ductus arteriosus in the near-term fetal lamb several days before delivery results in persistent pulmonary hypertension after delivery (1, 2). This lamb model closely resembles the human condition, with pulmonary arterial pressure (Ppa) equal to or greater than aortic pressure, and anatomic changes including muscularization of the normally nonmuscular intra-acinar arteries (3).

Nitric oxide (NO), produced from L-arginine by an endothelial constitutive NO synthase (NOS), is an important factor in the decrease in PVR that occurs at birth (4, 5). NO is believed to relax vascular smooth muscle by stimulation of soluble guanylate cyclase and production of the second messenger guanosine 3',5'-cyclic monophosphate (cGMP) from guanosine 5' triphosphate. Cyclic GMP mediates vascular relaxation and is hydrolyzed by specific phosphodiesterases to form the inactive metabolite guanosine 5'-monophosphate. Nine subtypes or families of phosphodiesterases have been described which differ in primary structure, relative affinity for cyclic adenosine monophosphate (cAMP) versus cGMP, responses to specific effectors and inhibitors, and mechanisms of regulation (6, 7). One family (the cGMP-binding, cGMP-specific phosphodiesterase or PDE5) both binds and hydrolyzes cGMP with high specificity relative to cAMP. PDE5 activity appears to be relatively specific for the lung and platelets. For instance, large amounts of activity for PDE5 have been reported in bovine lung homogenates (8), and in human pulmonary arteries (9). Further, PDE5 activity and expression appear to peak in newborn animals (10, 11), indicating that this enzyme may be of particular importance in the pulmonary vascular transition at birth.

Inhaled NO at doses of 5 to 80 ppm decreases pulmonary artery pressure and improves survival in lambs with PPHN following ductal ligation (12). Multicenter randomized trials have shown that inhaled NO improves oxygenation and decreases the need for extracorporeal membrane oxygenation in infants with PPHN (13, 14). However, these studies also demonstrate that many infants either do not respond or fail to sustain their response to inhaled NO. Further, potential toxicities of NO and its higher oxides indicate the need for therapeutic strategies designed to deliver the lowest effective concentration. Newborn lambs with PPHN after ductal ligation have decreased cGMP production in response to exogenous NO compared with control lambs owing to decreased activity of soluble guanylate cyclase, and decreased activity of endothelial NOS (15).

Several pharmacological agents with inhibitory activity for PDE5 have been studied in the pulmonary circulation. Specific PDE5 inhibitors may increase intracellular cGMP, and enhance pulmonary vasodilatation during inhalation of NO. We previously reported that zaprinast, an experimental inhibitor of PDE5, enhances and prolongs the effect of inhaled NO in newborn lambs with PPHN after ductal ligation (19). Zaprinast also dilates the pulmonary circulation in near-term fetal lambs, as well as adult cats, newborn lambs, and adult sheep with acute pharmacological pulmonary hypertension (20). Dipyridamole, a Food and Drug Administration-approved agent with significant inhibitory activity for PDE5, dilates the pulmonary circulation of fetal lambs (24, 25), and newborn lambs with PPHN after ductal ligation (17). However, the vascular effects of zaprinast and dipyridamole were not selective for the pulmonary circulation because both produced significant systemic hypotension (17, 19, 25). These systemic effects were probably caused by inhibition of other PDE isozymes not selective for cGMP as well as vasoreactive properties not related to phosphodiesterase inhibition.

E4021 (Sodium 1-[6-chloro-4-(3.4-methylenedioxybenzyl)- aminoquinazolin-2-yl]piperidine-4-carboxylate sesquihydrate), a new experimental agent, is 100 times as potent as zaprinast for PDE5 (26), and has minimal to no inhibitory activity for other PDE isoenzymes. E4021 has been studied in the coronary circulation of adult pigs (26) and in the pulmonary circulation of adult rats with pulmonary hypertension created by chronic hypoxia (27). Both studies found that low doses of E4021 (10 to 300 µg/kg) resulted in pulmonary vasodilation. In the current study, we examined the pulmonary and systemic vascular effects of E4021 in the ductal ligation lamb model. We hypothesized that E4021 would selectively dilate the pulmonary circulation. We first examined the effect of E4021 alone and in combination with the NO donor S-nitrosyl-acetyl-penicillamine (SNAP) in pulmonary arteries isolated from near-term control and pulmonary hypertensive fetal lambs. Then, in fully instrumented control fetal lambs, we studied the hemodynamic effects of E4021. Finally, in newborn lambs with PPHN, we developed a dose-response curve to E4021 and studied the hemodynamic effects of E4021 in combination with inhalation of threshold and subthreshold concentrations of NO.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The Laboratory Animal Care Committee at the State University of New York at Buffalo approved all procedures and protocols in this study.

Fetal Sheep Instrumentation

We prepared late-gestation fetal lambs for chronic vascular access. Time-dated pregnant ewes were operated on at 127 or 128 d gestation. Anesthesia was induced with 20 ml of a 5% solution of sodium thiopental and maintained with 1.5 to 2.0% halothane. A hysterotomy was performed through a midline abdominal incision, which exposed the fetal head, neck, and the left foreleg. A thoracotomy was performed in the left third intercostal space and a 6.0-mm ultrasonic transit time flow transducer was placed around the left pulmonary artery just beyond the bifurcation from the main pulmonary artery. Sufficient room was left between the probe and vessel to allow for growth. Polyvinyl catheters were inserted into the left pulmonary artery and the left atrium. The catheters and leads to the transducer were then secured to the exterior of the chest and the chest was closed. Polyvinyl catheters were then inserted in the right carotid artery and the right jugular vein; the venous catheter was advanced into the superior vena cava. After closing the neck incision, a catheter was sutured to the fetal skin for later monitoring of amniotic liquid pressure. Skin incisions were closed, and all catheters were tunneled through the uterus and the abdominal walls, which were doubly oversewn to prevent fluid leakage. A pouch was sewn to the maternal flank to prevent the ewe from damaging the catheters. Antibiotics were injected into the amniotic sac (700 mg ampicillin and 20 mg gentamicin) and fetal vein (300 mg ampicillin) at the time of surgery and daily thereafter. The ewe also received a daily intramuscular injection of 100 mg of ampicillin and 40 mg of gentamicin. Vascular catheters were flushed with isotonic saline and filled with heparin solution (1,000 U/ml) daily.

Prenatal Ligation of the Ductus Arteriosus

The technique for creating pulmonary hypertension by prenatal ligation of the ductus arteriosus has been described previously (2, 16, 28). Briefly, time-dated pregnant ewes (mixed breed) were operated on at 126 or 127 d of gestation (term = 146 d). Anesthesia was induced with 20 ml of a 5% solution of sodium thiopental and maintained with 1.5 to 2.0% halothane. The fetal head and left foreleg were delivered through a hysterotomy. A left lateral thoracotomy was performed in the fourth intercostal space of the fetus and the ductus arteriosus was ligated. The chest was closed and the fetus returned to the uterus. Postoperatively the ewe was treated with intramuscular ampicillin (300 mg/d) and gentamicin (40 mg/d) for 48 h. The ewe was allowed to recover for 9 d.

Isolated Vessel Preparation

After delivery by cesarean section under the same thiopental and halothane anesthesia described previously, the umbilical cord was clamped. In each case, a twin fetus served as the control. Before the first breath, lambs were killed by rapid exsanguination through a direct cardiac puncture. We have described the subsequent techniques previously in detail (16, 17). Briefly, the heart and lungs were removed en bloc from the thorax immediately after death and placed in Krebs-Ringer solution (in mM = NaCl 118, KCl 4.7, CaCl₂ 2.5, MgSO₄ 1.2, KH₂PO₄ 1.2, NaHCO₃ 25.5, glucose 5.6, and calcium disodium ethylenediaminetetraacetate 0.026). Fifth-generation intralobar pulmonary arteries (29) with inside diameters of 0.5 to 1.0 mm were isolated, dissected with care to preserve the integrity of the endothelium, and cut into rings approximately 2 mm wide and 0.7 to 2.6 mg in weight. In some rings the endothelium was removed by rubbing the luminal surface with the tip of a curved forceps. Endothelium removal was confirmed by absence of response to the endothelium-dependent agonist acetylcholine, and by light microscopy examination. Wet tissue weights were obtained at the end of each experiment after blotting the rings dry on gauze pads. The force of contraction was normalized by the weight of each ring and expressed as grams/gram (g/g) of tissue. Vessel rings were mounted on stainless steel hooks and placed in water-jacketed chambers. Tissues were bathed with 6 ml of the Krebs-Ringer solution, which was maintained at 37° C, and aerated with a gas mixture of 94% O₂ and 6% CO₂ to maintain a pH of 7.40, a PCO₂ of 38 mm Hg and a PO₂ of greater than 500 mm Hg. A continuous recording of isometric force generation was obtained by tying each vessel ring to a force-displacement transducer (Statham UC 2; Statham Instruments, Hato Rey, PR) that was connected to an oscillographic recorder. Once mounted, the vessel rings were allowed to equilibrate for 20 min in the bathing solution. A micrometer was then used to stretch the tissues repeatedly in small increments over the next 45 min until resting tone remained stable at a passive tension of 1.0 g for control arteries and 1.2 g for arteries from hypertensive lambs. Previous experiments determined that this is the optimal length for generation of active tone in response to exogenous norepinephrine (NE) (16, 17).

The following pharmacological agents were used: E4021, indomethacin, L-norepinephrine, DL-propranolol hydrochloride, SNAP, and N-nitro-L-arginine (L-NA). E4021 manufactured by Eisai Tsukuba Research Laboratory was a gift from the Animal Health Trust (Newmarket, Suffolk, UK). The other drugs were purchased from Sigma Chemical Company (St. Louis, MO). All drugs were dissolved in distilled H₂O except for indomethacin, and E4021 which were dissolved in ethanol and 0.01 N NaOH, respectively. Preliminary experiments showed that in the concentrations used in the tissue bath studies, neither vehicle had any effect. Drugs were made fresh daily.

Experimental Protocols

Isolated vessels. Pulmonary arteries from control and hypertensive lambs were pretreated for 20 min with 10⁶ M propranolol to block -adrenergic receptors, and with 10⁵ M indomethacin to prevent the formation of vasoactive prostaglandins. They were preconstricted with a median effective concentration (EC₅₀) of norepinephrine (NE, 10⁶ M) as determined from preliminary studies in which cumulative concentration-response curves for NE (10⁸ to 10⁵ M) were developed (16). Once the response to NE had reached a steady level, cumulative concentration-response curves to E4021 (10⁹ to 10⁵ M) were obtained by increasing its bath concentration in successive steps. The next concentration was added only when the response to the prior concentration had reached a plateau. In order to determine the role of endogenous NO production, some vessels were incubated for 30 min with 10³ M L-NA before constriction with NE.

In the second protocol, pulmonary arteries isolated from control and hypertensive lambs were pretreated with 10³ M L-NA to prevent formation of endogenous NO, in addition to preconstriction with 10⁶ NE as described previously. Cumulative concentration-response curves to the NO donor SNAP were then performed in arteries with and without preincubation with E4021 (10⁷ M). Each vessel ring was used for one experimental protocol and then discarded.

Intact Animal Preparations

Materials. For the in vivo experiments, E4021 was dissolved in 0.01 N NaOH. The concentration of E4021 was adjusted to maintain a constant 6 ml/h infusion rate for all protocols. NO was obtained as 1,000 ppm in nitrogen (Matheson Gas Products Inc., Twinsburg, OH), and blended with oxygen to obtain 0.5 or 5 ppm. NO was continuously analyzed from the inspired gas immediately before the endotracheal tube by a chemiluminescence analyzer (Model 42 H/42S; Thermoenvironmental Instruments, Franklin, MA).

Measurements. Phasic pulmonary arterial, left atrial, aortic, and airway pressures were measured by Gould Statham physiologic pressure transducers (P-23 XL; Gould Electronics, Cleveland, OH) which were calibrated at the start of each experiment with a mercury column manometer. Pulmonary blood flow (p) was measured by an ultrasonic transit time flow transducer (6.0 mm around the left pulmonary artery for fetuses, 10.0 mm around the main pulmonary artery for newborn lambs; Model T 101, Transonic Systems, Ithaca, NY) and processed by a digital flowmeter. These data were recorded continuously on a physiologic amplifier-recorder system (Gould Electronics). Pulmonary vascular resistance (PVR) was calculated as: PVR = (  Pla)/p and reported in units of (mm Hg · kg · min/ml), where = mean pulmonary arterial pressure and Pla = left atrial pressure. Aortic blood was collected for measurement of pH, PO₂, PCO₂, hemoglobin, and base status (Acid-Base Laboratory 3; Radiometer Medical A/S, Copenhagen, Denmark).

Postnatal care. At 135 to 136 d (9 d after ductal ligation) the pregnant ewe was fasted and anesthetized as described previously, and the fetal head exposed through a hysterotomy. The subsequent techniques have previously been described in detail (17, 19). The fetal trachea was intubated with a 4.0-mm cuffed endotracheal tube. The carotid artery and jugular vein were exposed in the neck, polyvinyl catheters inserted, and advanced into the aorta and the right atrium, respectively. The fetal chest was exposed and a left thoracotomy was performed. Polyvinyl catheters were placed in the main pulmonary artery and left atrium, and a 10.0-mm ultrasonic transit time flow transducer was placed around the main pulmonary artery. The umbilical cord was ligated and the lamb delivered. The lamb was wrapped in a homeothermic servo-controlled warming blanket (Harvard Apparatus, Edenbridge, KY) and placed under an infant warmer (Air Shields, Vickers, Hatboro, PA) to maintain temperature at 39° C. Ventilation was initiated with a time-cycled pressure control ventilator (Siemens Servo 900C; Solna, Sweden) at a fraction of inspired oxygen (FI_O2) 0.95, ventilator rate [IMV] = 60 breaths/min, peak inspiratory pressure (PIP) = 30 cm H₂O, peak end-expiratory pressure (PEEP) = 4 cm H₂O, and an inspiratory time (TI) = 33%. The IMV and PIP were adjusted to maintain between 35 and 50 mm Hg. Tris-hydroxymethyl aminomethane (THAM) was used to correct any metabolic acidosis defined as a base deficit greater than 10. Systemic hypotension, defined as a decrease in mean systemic blood pressure of greater than 15 mm Hg, or an initial hemoglobin of less than 11 g/dl was corrected by administration of 10 ml/kg of maternal blood. The newborn lamb was sedated with 10 mg/kg of ketamine hydrochloride as needed. The animal was stabilized for 90 to 120 min before an experimental protocol was begun. During each protocol no blood or THAM infusions were given and ventilator settings remained constant. After each study the lambs were killed by a lethal dose of sodium pentobarbital.

Experimental protocol 1: late-gestation fetal lambs (n = 6). Animals were allowed to recover from surgery for at least 72 h before experiments began. Experiments were performed on unanesthetized fetuses that were judged to be healthy by arterial pH and blood gas tensions. Fetuses were studied while their ewes stood upright in a cage with free access to food and water. E4021 was infused in the left pulmonary artery at 31 µg/min for 20 min, with hemodynamics measured before the infusion, every 10 min during the infusion, and every 15 minutes after completion of the infusion until hemodynamics returned to baseline. Subsequently, 30 mg of L-NA was infused over 30 min followed immediately by 31 µg/min of E4021 for 20 min. Arterial blood gases were measured before and after each experimental protocol.

Experimental protocol 2: dose-response curve in PPHN lambs (n = 7 newborn lambs with prenatal ligation of ductus arteriosus, gestational age 135 to 136 d). The amount of E4021 infused was selected randomly in a block fashion. The three doses of E4021 tested were: 1, 10, and 100 µg/kg/min. After initial stabilization, E4021 was infused through the right atrial catheter for 30 min using a continuous infusion pump (Medfusion Systems Inc., Norcross, GA). Baseline hemodynamics and arterial blood gases were measured before starting the infusion of E4021, every 10 min during the infusion, and every 15 min for a total of 1 to 2 h after completion of the infusion. These measurements were repeated with each dose of E4021. Each infusion lasted 30 min and the following dose was started 60 to 120 min after the completion of the previous dose and return to baseline hemodynamic values. If necessary, ventilator changes, volume infusion, or THAM were given between doses of E4021.

Experimental protocol 3: E4021, inhaled NO, or their combination (n = 26 newborn lambs with prenatal ligation of ductus arteriosus, gestational age 135 to 136 d). After stabilization, lambs were randomly assigned to one of five treatment protocols, each delivered for a total of 20 min: (1) E4021 alone (10 µg/kg/min infused through right atrial catheter), (2) inhaled NO alone (0.5 ppm), (3) inhaled NO alone (5 ppm), (4) E4021 (10 µg/kg/min) combined with 0.5 ppm inhaled NO, or (5) E4021 (10 µg/kg/min) combined with 5 ppm inhaled NO. Baseline hemodynamics and arterial blood gases were measured before starting a treatment protocol, every 10 min during the protocol, and every 15 min for a total of 2 h after completion of the protocol.

Data Analysis

All data are expressed as mean ± SE, with n representing the number of animals studied. Statistical analysis was performed with the Statview 4.5 software package (Abacus Concepts, Berkley, CA). Statistical comparisons for normally distributed data within groups were performed using analysis of variance (ANOVA) for repeated measures, followed if necessary by Student-Newman-Keuls post hoc testing for multiple comparisons. A Wilcoxon signed rank test was used to compare groups of data that were not normally distributed. A one-way ANOVA was performed to determine differences in hemodynamic responses between groups of intact lambs. A p value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolated Vessel Studies

Plateau contractile responses to 10⁶ M norepinephrine were similar for pulmonary arteries from control and hypertensive lambs. Figure 1 demonstrates that increasing concentrations of E4021 significantly relaxed pulmonary arteries isolated from control lambs. This relaxation to E4021 was completely blocked by pretreatment with L-NA. Additional experiments in endothelium-rubbed arteries revealed a similar complete block in relaxation to E4021 (n = 4). In marked contrast, E4021 had no effect on pulmonary arteries isolated from lambs with pulmonary hypertension.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Increasing concentrations of E4021 significantly relax pulmonary arteries isolated from control late-gestation fetal lambs. This relaxation was completely blocked by incubation with 103 M L-NA. E4021 had no effect on pulmonary arteries isolated from lambs after prenatal ligation of the ductus arteriosus. All rings were pretreated with indomethacin and propranolol. Relaxations are expressed as percentage of a plateau NE constriction (100%). Data are mean ± SEM, n = 6 lambs for each data point. *p < 0.05 versus baseline NE constriction.

Figure 2 demonstrates that in arteries from control and hypertensive lambs, pretreatment with 10⁷ M E4021 significantly enhanced relaxations to the NO donor SNAP.

View larger version (16K): [in this window] [in a new window]   Figure 2.   E4021 pretreatment (107 M) significantly enhances cumulative concentration-response curves to the NO donor SNAP in pulmonary arteries isolated from control and hypertensive fetal lambs. All rings were pretreated with L-NA as well as indomethacin and propranolol. Relaxations are expressed as percentage of a plateau NE constriction (100%). Data are mean ± SEM, n = 6 lambs for each data point. *p < 0.05 versus without E4021 pretreatment.

Intact Animal Studies

Experimental protocol 1: late-gestation fetal lambs. E4021 increased p and decreased PVR (p < 0.05 for both variables). Aortic pressure, heart rate, and Pla did not change following E4021. Infusion of L-NA increased baseline PVR from 3.4 ± 0.3 to 5.1 ± 0.6. As illustrated in Figure 3, E4021-induced pulmonary vasodilation was significantly attenuated by L-NA pretreatment.

View larger version (9K): [in this window] [in a new window]   Figure 3.   Pulmonary hemodynamic changes following infusions of E4021 in instrumented late-gestation fetal lambs. Pretreatment with L-NA blocked the increase in p and decrease in PVR in response to E4021. *p < 0.05 versus without L-NA pretreatment.

Experimental protocol 2: dose-response curve in PPHN lambs. Table 1 shows the measured hemodynamic variables and the arterial blood gas results for each dose of E4021. The "during" value represents measurements obtained after 10 min of drug infusion. No difference was found between measurements obtained at 10 and 30 min after initiation of the E4021 infusion for any of the variables. E4021 at 1 and 10 µg/kg/min decreased aortic pressure from baseline, but at 100 µg/kg/min aortic pressure did not significantly change (Figure 4). At 10 and 100 µg/kg/min of E4021, Ppa decreased significantly from baseline and at the highest dose the decrease in Ppa was significantly greater than the decrease in aortic pressure (Figure 4). p tended to increase in a dose-dependent manner but the change from baseline was not significant (Figure 5). PVR decreased significantly from baseline after 10 and 100 µg/kg/min of E4021, and Pa_O2 increased significantly from baseline after the 100 µg/kg/min dose of E4021 (Figure 5).

View larger version (14K): [in this window] [in a new window]   Figure 4.   Pulmonary and systemic arterial pressure responses following infusions of E4021 in newborn lambs with PPHN after ligation of the ductus arteriosus. Values shown were obtained after 10 min of infusion of E4021, and are presented as percent change from baseline after each dose of E4021. Data are mean ± SEM, n = 5 lambs for each data point. *p < 0.05 versus baseline. p < 0.05 versus aortic pressure.

View larger version (10K): [in this window] [in a new window]   Figure 5.   p, PVR, and Pa O2 responses after 30-min infusions of E4021 in newborn lambs with PPHN after ligation of the ductus arteriosus. Data are presented as percent change from baseline after each dose of E4021. Data are mean ± SEM, n = 5 lambs for each data point. *p < 0.05 versus baseline.

Experimental protocol 2: E4021, inhaled NO, or their combination. Table 2 shows the measured hemodynamic variables and the arterial blood gas results for each of the five groups. Aortic pressure did not change in any group of animals. In the group receiving E4021 alone the absolute decrease in aortic pressure approached significance (p = 0.07) and if expressed as a percent change from baseline, as in the dose-response protocol, the decrease was significant (9 ± 3, p = 0.03). Pulmonary hemodynamics improved significantly in all groups except the 0.5 ppm inhaled NO group (Figure 6). Pa_O2 increased significantly in all groups except the 0.5 ppm inhaled NO group. E4021 did not enhance the pulmonary hemodynamic effects of NO inhaled at 0.5 or 5 ppm.

View larger version (13K): [in this window] [in a new window]   Figure 6.   E4021 (10 µg/kg/min) does not enhance the pulmonary vascular effects of NO inhaled at 0.5 or 5 ppm in newborn instrumented lambs with PPHN after ligation of the ductus arteriosus. Data for Ppa, p, and PVR are presented as percent change from baseline for each of the five protocol groups. Data are mean ± SEM, n = 5 lambs for each data point. *p < 0.05 versus baseline.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cyclic nucleotide phosphodiesterases, by changing intracellular levels of cyclic nucleotides and thereby changing protein phosphorylation, play an important role in cell signal transduction. The cyclic GMP binding, cyclic GMP-specific phosphodiesterase, referred to as PDE5, has high tissue specificity for the lung and platelets. Further, PDE5 activity, protein, and message appear to be regulated during perinatal development, and have peak activity at the time of birth (10, 11). We and others have proposed that inhibitors of PDE5 enzymatic activity might be useful in the treatment of PPHN (10, 11, 19). We have previously studied zaprinast and dipyridamole, both of which inhibit PDE5 in vitro, in newborn lambs with persistent pulmonary hypertension (17, 19). However, E4021 has been reported to have greater potency and specificity for PDE5 (27). The inhibitory concentration of 50% (IC₅₀) for PDE5 is approximately 100-fold lower than that for zaprinast, and it has no detectable inhibition of PDE1, a calmodulin-dependent PDE that metabolizes cGMP and cAMP.

In the current study, we found that E4021 relaxed pulmonary arteries isolated from control late-gestation fetal lambs. This response could be completely blocked after pretreatment with L-NA or removal of endothelium, indicating that the response was dependent on endothelial NOS activity. In contrast, E4021 alone did not relax pulmonary arteries isolated from hypertensive lambs. As our previous studies show that messenger RNA (mRNA), protein, and activity for endothelial NOS are significantly decreased in parenchyma and pulmonary arteries isolated from fetal lambs with PPHN (15, 16), this finding was not surprising, and indicates that basal production of NO is depressed in these isolated vessels.

In intact, instrumented newborn lambs with PPHN, infusions of E4021 decreased Ppa in a dose-dependent fashion. At the higher infusion doses (10 to 100 µg/kg/min), the dilation was selective for the pulmonary circulation. Our findings in newborn lambs with PPHN are therefore similar to recent studies, which found that E4021 infusions decrease Ppa in healthy adult pigs (26) and adult rats with chronic hypoxia (27). E4021 infusions resulted in a more selective effect on the pulmonary circulation than zaprinast or dipyridamole did in the ductal ligation model of PPHN. This is likely a result of the specificity of E4021 for PDE5, and the high concentration of PDE5 isozyme in the lung. Braner and coworkers (20) were able to show a selective pulmonary dilation to increasing doses of zaprinast in control lambs with pulmonary hypertension induced by thromboxane agonists. However, we have previously shown that in the ductal ligation model, zaprinast alone (50 µg/kg/min × 20 min) had no effect on Ppa and decreased systemic blood pressure by 18%.

We could not determine the mechanism for pulmonary vascular dilation to E4021 alone in newborn lambs with PPHN, because we saw no response in fifth-generation pulmonary arteries isolated from hypertensive lambs. While isolated vessel studies provide a useful pharmacological screening method, our study shows that their results must be generalized to the intact pulmonary vasculature with caution. The E4021 concentrations used in the tissue baths were roughly equivalent to estimated serum concentrations in the intact lamb studies. It is possible that E4021 had its in vivo effects at a different point along the pulmonary tree than we studied in the tissue bath studies. Although we and others have previously shown decreased activity for both endothelial NOS (eNOS) and soluble guanylate cyclase in lung parenchyma from newborn lambs with PPHN, perhaps sufficient enzyme activity and cGMP production remain in resistance vessels to allow for a pulmonary vascular response to E4021. For example, our previous studies indicate that eNOS and soluble guanylate cyclase activity are normal in pulmonary veins isolated from lambs with PPHN (16).

It is also possible that E4021 dilates the pulmonary vasculature in vivo by mechanisms other than inhibition of phosphodiesterase activity. However, we found that E4021 decreases PVR in normal instrumented late-gestation fetal lambs, an effect which was blocked completely by pretreatment with L-NA. We speculate, but cannot prove, that the effect in hypertensive lambs is also dependent on NOS activity. Unfortunately, we have previously shown that administering L-NA to newborn lambs after ductal ligation produces profound hypoxemia and cardiovascular instability (30), and is therefore not a feasible tool for determining the mechanism for dilation to E4021 in this model.

We found more dramatic pulmonary vascular responses to E4021 than have been described by other investigators. Cohen and coworkers (27) found a more modest decrease in Ppa and PVR in adult rats after chronic hypoxia than we found in newborn lambs after ductal ligation. Further, in their study the pulmonary hemodynamic effects did not increase with increasing E4021 doses. However, there are fundamental differences between the two models. cGMP production was increased in lung perfusates prepared from the rat model of chronic hypoxia. In contrast, we reported that cGMP production was decreased at baseline and after stimulation with NO in the ductal ligation model (16, 17). In addition, recent reports indicate that PDE5 activity and expression are significantly higher in the lungs of newborn rats, mice, and lambs when compared with adult lungs (10, 11). This pattern of developmental expression is similar to what has been reported for NOS and soluble guanylate cyclase (31, 32). These developmental patterns are likely due to the critical regulatory role of cGMP in the vascular transition at birth, and further indicate the importance of the phosphodiesterase enzyme system in regulating cGMP concentrations during the perinatal period.

We observed a small (8%) decrease in systemic blood pressure after infusion of E4021. However, in contrast to pulmonary pressure, the change in systemic pressure was similar for all three doses. Therefore, increasing doses of E4021 produced an increasingly selective pulmonary vasodilation. Other investigators have similarly found that similar doses of E4021 produced either a modest or no drop in systemic pressure in healthy conscious pigs (26) or rats with pulmonary hypertension (27). The modest effect on systemic pressure is in striking contrast from what we and others have reported for zaprinast (19, 27) and dipyridamole (17). We speculate that the tissue-specific distribution of PDE5 in the lung (8, 9), and the selectivity of E4021 for this isoenzyme minimize its systemic effects.

In humans with PPHN, baseline cGMP levels are decreased when compared with healthy newborns and levels increase following inhaled NO (33). Specific PDE5 inhibitors may provide more intracellular cGMP to enhance pulmonary vasodilation if there is not an adequate or a sustained response following inhaled NO. PDE5 inhibitors such as zaprinast and dipyridamole have been shown to enhance the pulmonary vascular effects of endogenous and exogenous NO in animal models of pulmonary hypertension (19). However both zaprinast and dipyridamole may significantly decrease aortic pressure at concentrations required to enhance pulmonary vasodilation (17, 19, 25).

In the second part of our study, we tested the hypothesis that E4021 would enhance the pulmonary vascular effects after inhalation of NO, and decrease the required dose of inhaled NO in newborn lambs with PPHN. We previously demonstrated that zaprinast more than quadrupled the hemodynamic and oxygenation changes following inhaled NO in the ductal ligation lamb model (19). As expected, E4021 significantly enhanced responses to the NO donor SNAP in fifth-generation pulmonary arteries isolated from lambs with PPHN. However, in contrast to our expectations, the pulmonary vascular effects of E4021 were not enhanced after inhalation of a subthreshold (0.5 ppm) or threshold (5 ppm) concentration of NO in instrumented lambs with PPHN (34). While the final concentrations of E4021 were roughly equivalent in our in vitro and in vivo studies, it is more difficult to directly compare the amount of NO delivered. The in vitro studies used an NO donor instead of authentic NO, as well as tissue baths that were continuously aerated and open to atmosphere. However, we believe that the most reasonable explanation for our findings is that E4021 alone produced a maximal effect on pulmonary hemodynamics. We speculate that the addition of NO at the concentrations used in this study was not additive because the hemodynamic effects of E4021 alone were equivalent to 50 to 100 ppm of inhaled NO (34). However, an additional possibility is that E4021 enhanced endogenous NOS activity, either directly or by indirect mechanisms such as increasing shear stress, a potent stimulus for NO production (35).

Although NO inhalation did not enhance the pulmonary vascular effects of E4021, oxygenation improved (Table 2). Preliminary experiments in fetal lambs in our laboratory have shown that infusions of E4021 decrease net lung liquid production, and this effect is independent of changes in PVR (36). An increase in net lung liquid clearance in the postnatal lung could explain the improvement in oxygenation we observed in the current study.

PPHN is a condition resulting from a diverse set of circumstances, and caution must be exercised when applying the results from experimental models to humans. However, the ductal ligation model of pulmonary hypertension is a chronic preparation resulting in antenatal vascular remodeling (3), and has striking physiologic and anatomic similarities to human infants with PPHN (37). Conducting studies in newborn animals is important as recent studies propose distinct developmental time lines for the enzymatic steps in the NO/cGMP vasodilatory pathway (10, 31, 32). Our previous studies further indicate that prenatal ligation of the ductus arteriosus alters the expression of critical enzymes in the NO pathway (15, 16), a finding that would not be expected in acute models of pulmonary hypertension following infusion of vasoconstrictors. The current study provides strong rationale to further explore selective PDE5 inhibition as a potential therapy for the treatment of persistent pulmonary hypertension. Phosphodiesterase inhibitors such as E4021 may provide a safe and effective alternative for the treatment for PPHN.