Main

EDNO(14) and dilator PG such as prostacyclin (PGI2)(46) are potent modulators of vasomotor tone that contribute to the decrease in PVR at birth(79). Effective attenuation of pulmonary vasoconstriction remains essential during the early postnatal period when constriction of highly muscularized small pulmonary arteries(10, 11) may otherwise lead to a resumption of shunting across fetal cardiovascular channels. In addition, modulation of venous tone likely contributes to the maintenance of normal lung fluid balance(12). Although many studies suggest that both dilator PG(6, 9, 13, 14) and EDNO(2, 15) attenuate pulmonary vasoconstriction in young newborns, others have found that cyclooxygenase inhibition had little effect(16), and EDNO activity was minimal during the early postnatal period(17, 18). This study was designed to identify the relative effects of dilator PG and EDNO on basal and increased pulmonary vasomotor tone within the first few hours of birth. To achieve this goal, total and segmental vascular resistances were measured during normoxia and hypoxia in isolated lungs of 6-h-old lambs studied under control conditions, after inhibiting PG synthesis with the cyclooxygenase inhibitor, indomethacin, after inhibiting EDNO synthesis with the nitric oxide synthase inhibitor Nω-nitro-L-arginine, and after inhibiting synthesis of both modulators.

METHODS

Preparation. The study was approved by the Institutional Animal Care and Use Committee of the University of Maryland School of Medicine. Isolated lungs from approximately 6-h-old (range 4.5-8 h) lambs were prepared as previously described(3, 6, 13). Briefly, after anesthesia with ketamine (50 mg/kg, i.m.), femoral venous and arterial catheters were inserted, and a tracheostomy was performed. Ventilation was initiated with a hyperoxic gas mixture, and the lamb was exsanguinated after i.v. administration of heparin (5000 U) and pancuronium bromide (0.1 mg/kg). During exsanguination the gas mixture was adjusted to 28% O2, 5% CO2, and balance with N2 (termed normoxia in this study). Lungs were ventilated (model 713 animal ventilator, Harvard Apparatus, South Natick, MA) at a tidal volume of ≈15 mL/kg (body weight), rate of 10/min, and end expiratory pressure of 3 mm Hg.

After exsanguination, a midline thoracotomy was performed, the ductus arteriosus was ligated, and the pulmonary artery and left atrium were cannulated and connected to perfusion tubing. The perfusate, a mixture of autologous blood and 3% dextran 70 in Ringer's lactate, exited the left atrial cannula to a reservoir from whence it was pumped (Varistaltic roller pump, Manostat, New York, NY) through a heat exchanger (SciMed, Minneapolis, MN), bubble trap, and flow probe (model EP 300, Carolina Medical Electronics, King, NC), and a plexiglass flow diverter through which perfusate could be directed to the pulmonary artery or back to the reservoir. Ppa, Pla, Paw, and perfusate flow were constantly measured (Statham P23id pressure transducers, Spectramed, Oxnard, CA and model 501 flowmeter, Carolina Medical Electronics, respectively) and recorded (model 7D polygraph, Grass Instrument Co, Quincy, MA). All pressures were zero referenced to the level of the left atrium. Inspired oxygen tension was constantly monitored (model 3A O2 analyzer, Ametek, Pittsburgh, PA), and perfusate blood gases (model 945 blood gas analyzer, AVL Scientific Corp, Roswell, GA) and glucose concentrations were determined at both oxygen tensions. Perfusate temperature was maintained between 38.5 and 39.2°C, glucose was maintained between 5.0 and 7.0 mmol/L by the addition of 50% dextrose, and pH was maintained at 7.35-7.42 by the addition of 1 N NaHCO3.

Protocol. At the start of perfusion, lungs were divided into four study groups: CON (n = 6), INDO (n = 5), LNA (n = 8), and BOTH (n = 7). The concentrations of indomethacin (40 μg/mL perfusate) andNω-nitro-L-arginine (10-3 M) added to the reservoir at the beginning of perfusion have been previously shown to block cyclooxygenase and nitric oxide synthase activity, respectively, in this preparation(3, 6). During the initial 60 min of perfusion, lungs were ventilated with normoxic gas, and perfusate flow was gradually increased to 100 mL/kg·min, while Pla was maintained<0 mm Hg (i.e. zone 2 conditions). At 60 min, Pla was increased to 5 mm Hg so that Ppa > Pla > Paw at end expiration (i.e. zone 3 conditions). After a brief (≈3 min) stabilization period, total and segmental pressure gradients were measured using the inflow-outflow and double occlusion techniques described below. Pla was subsequently reduced below 0 mm Hg, and the gas switched to 4% O2, 5% CO2, and balance N2 (termed hypoxia in this study). This level of hypoxia has previously been shown to elicit maximal HPV in the preparation(6). After 20 min, Pla was increased to 5 mm Hg, and total and segmental pressure gradients were again measured.

Inflow-outflow occlusion pressures were determined as previously described(3, 13). Briefly, with the ventilator off at end expiration, inflow to the pulmonary artery was suddenly (<50 ms) occluded, resulting in an abrupt decrease in Ppa to a point termed Ppa', followed by a slower exponential decline in pressure; Ppa - Ppa'= ΔPA. After perfusion had been reestablished, venous outflow was suddenly occluded, resulting in a sharp rise in Pla to a point termed Pla', followed by a slower linear rise in pressure; Pla' - Pla = ΔPV. Because Ppa - Pla =ΔPT, the total pressure gradient across the pulmonary circuit, the pressure gradient across the relatively compliant small arteries and veins of the middle segment (ΔPM) could be calculated asΔPM = ΔPT - (ΔPA + ΔPV).ΔPM was then partitioned into pre- and postcapillary segments using the double occlusion technique. Briefly, after perfusion had been reestablished, the inflow and outflow cannulae were simultaneously occluded, resulting in a single double occlusion (Pdo) previously shown to approximate capillary pressure(19). Pressure gradients across small arteries (Ppa' - Pdo = ΔPSA) and small veins (Pdo - Pla' = ΔPSV) were then calculated. Inflow and outflow occlusions were performed in duplicate under each condition and mean pressure gradients calculated. Because flow was constant in all experiments, differences in total and segmental pressure gradients reflect differences in total and segmental PVR.

After final occlusion measurements, 4.5 mL of perfusate were collected in tubes containing indomethacin (5 μg/mL) and 7.6 mg EDTA and promptly centrifuged at 4500 rpm. The plasma was then frozen at -80°C for later determination of 6-keto-PGF concentrations, the stable metabolite of PGI2 using a commercial enzyme immunoassay kit (Amersham Corp.). Endothelium-dependent relaxation was also measured by adding 10-6 M acetylcholine to the reservoir and noting the change inΔPT.

Drugs and solutions. A stock solution of indomethacin was prepared daily by dissolving 50 mg of indomethacin (Sigma Chemical Co., St. Louis, MO) in 5 mL of normal saline and 5 mL of 1 N NaHCO3. A stock solution of 0.1 M Nω-nitro-L-arginine (Sigma Chemical Co.) was also prepared daily by dissolving 219.2 mg ofNω-nitro-L-arginine in a total of 10 mL of normal saline titrated with 12 N HCl. Perfusate acid-base status was determined after adding the Nω-nitro-L-arginine. The small volume ofNω-nitro-L-arginine added to the reservoir caused a mild decrease in pH to ≈7.25 that was readily corrected with 1 N NaHCO3 before any measurements of pressure gradients were made.

Data analysis. All data shown are mean ± SEM. Differences among groups were determined by one- or two-way ANOVA and LSD test. Results were considered significant at p < 0.05.

RESULTS

There were no differences in body weight, blood or perfusate hematocrits, peak Paw, and blood gas data among the four study groups. Mean results from all lambs revealed a weight of 3.9 ± 0.2 kg, an initial hematocrit of 43.8 ± 1.5%, a perfusate hematocrit of 17.1 ± 0.5%, and a peak Paw of 8.2 ± 0.4 mm Hg. During normoxia, mean Po2 was 185.9 ± 2.9 torr, mean Pco2 was 39.2 ± 0.9 torr, and mean pH was 7.39 ± 0.01. During hypoxia, mean Po2 was 32.9± 0.8 torr, mean Pco2 was 36.1 ± 0.8 torr, and mean pH was 7.41 ± 0.01.

Acetylcholine caused a significant decrease in ΔPT in CON and INDO, but not BOTH or LNA lungs (Fig. 1A, LSD = 7.7 mm Hg). Plasma concentrations of 6-keto-PGF were higher in CON and LNA compared with INDO or BOTH lungs (Fig. 1B, LSD = 0.43 pg/L). In addition, 6-keto-PGF concentrations were significantly higher in the LNA compared with the CON group.

Figure 1
figure 1

Acetylcholine reduced the hypoxic pressure gradient in CON and INDO, but not LNA or BOTH lungs (in A, the asterisk(*) indicates the difference from hypoxic baseline pressure atp < 0.05). Plasma 6-keto-PGF concentrations were markedly reduced by indomethacin (B). The dagger (†) indicates that the plasma 6-keto-PGF concentration was higher in CON compared with BOTH or INDO lungs, and ** indicates that concentrations in LNA lungs exceeded all others (p < 0.05).

ANOVA revealed a significant interaction between oxygen tension and study group (Fig. 2, LSD = 12.4 mm Hg). Hypoxia caused a significant increase in ΔPT of INDO, LNA, and BOTH lungs, but not in the CON lungs. Because four of the six CON lungs showed little or no response to hypoxia, the increase in ΔPT of the CON lungs failed to reach significance even when analyzed by paired t test(p = 0.07). Under normoxic conditions. ΔPT was greater in BOTH compared with CON lungs, and the difference between LNA and CON approached significance (t test, p = 0.06). Under hypoxic conditions, ΔPT was greater in LNA than CON lungs and in INDO and BOTH compared with the LNA group (Fig. 2).

Figure 2
figure 2

Data are mean ± SEM total pressure gradient (mm Hg) measured at constant flow. Hypoxia caused a significant increase inΔPT of INDO, LNA, and BOTH lungs (*). NormoxicΔPT was higher in BOTH than other lungs (**), and hypoxicΔPT was higher in LNA than CON lungs (†) and in INDO and BOTH than LNA or CON lungs (‡). All differences significant atp < 0.05.

Analysis of the segmental pressure gradients (Figs. 3 and4) showed that hypoxia caused a significant increase inΔPA of INDO and BOTH lungs (LSD = 5.0 mm Hg) and a significant increase in ΔPM in all groups (LSD = 9.9 mm Hg). This reflected the marked increase in ΔPSA (Table 1; LSD = 8.7 mm Hg) and the smaller, but significant, increase in ΔPSV(Table 1; LSD = 1.8 mm Hg). Normoxic ΔPA did not differ between groups (Fig. 3), whereas hypoxicΔPA was significantly greater in the INDO and BOTH groups compared with CON lungs (Fig. 4). As withΔPT, normoxic ΔPM was greater in BOTH compared with CON lungs (Fig. 3). Furthermore, hypoxic ΔPM was greater in LNA than CON lungs and in the INDO and BOTH groups compared with the LNA group (Fig. 4).

Figure 3
figure 3

Normoxic mean ± SEM pressure gradients across arterial, middle, and venous segments are shown. ΔPM was significantly greater in BOTH than CON lungs (*) and ΔPV was greater in CON (†) than INDO lungs and in LNA (‡) than either INDO or BOTH lungs. All differences significant at p < 0.05).

Figure 4
figure 4

Hypoxic mean ± SEM pressure gradients across arterial, middle, and venous segments are shown. ΔPA was significantly greater in INDO and BOTH than the other groups (**).ΔPM was greater in LNA than CON (*) and INDO or BOTH than the other groups (**). ΔPV was greater in LNA than INDO or BOTH lungs (‡). All differences significant at p < 0.05).

Table 1 Pressure gradients across small arteries and veins during normoxia and hypoxia

In contrast to the other segmental resistances, hypoxia had no effect onΔPV in any group. Nonetheless, one-way ANOVA showed that normoxicΔPV was significantly greater in the LNA compared with INDO or BOTH groups and in CON compared with INDO (Fig. 3, LSD = 2.3 mm Hg) and hypoxic ΔPV was greater in LNA compared with INDO or BOTH (Fig. 4, LSD = 3.1 mm Hg).

DISCUSSION

Nω-Nitro-L-arginine prevented ACh-induced pulmonary vasodilation (Fig. 1A) and indomethacin caused a marked decrease in plasma 6-keto-PGF concentration(Fig. 1B), indicating effective blockade of nitric oxide synthase and cyclooxygenase activity, respectively. The results of this study suggest that both EDNO and dilator PG are important modulators of pulmonary vasomotor tone during early newborn life. During normoxia, inhibition of both modulators was required to elicit a significant increase in total PVR(Fig. 2). However, the increase in normoxicΔPT in the LNA group approached significance, suggesting that EDNO played the major role in maintaining normoxic pulmonary vasomotor tone(Fig. 2). This finding is consistent with previous studies showing that cyclooxygenase inhibition had little or no effect under baseline normoxic conditions in 12-48-h-old lamb lungs(6, 13), but nitric oxide synthase inhibition increased normoxic PVR in isolated lungs from 1-d-old piglets(15).

Both modulators attenuated HPV in 6-h-old lamb lungs (Fig. 2). However, dilator PG appeared to play the dominant modulating role during hypoxic, for hypoxic ΔPT was higher in INDO than LNA lungs and did not increase further in the BOTH group (Fig. 2). These findings are consistent with studies showing that dilator PG such as PGI2 are potent modulators of HPV(4, 5, 20), particularly in lungs from young newborns(6, 9, 13, 14, 21). Indeed, in both this and previous studies of 12-24-h-old lamb lungs, we found that cyclooxygenase inhibition was required before HPV achieved statistical significance(13).

Although several studies have shown that EDNO attenuated pulmonary vasoconstriction due to hypoxia or the thromboxane mimetic, U46619, in young newborns(2, 15), others suggest that EDNO activity is low during the early neonatal period(3, 17, 18). This may explain the relatively small increase in HPV seen in LNA compared with INDO lungs (Fig. 2). Alternatively, the higher plasma 6-keto-PGF concentration in the LNA group(Fig. 1B) suggests that PGI2 synthesis was increased after EDNO synthesis had been blocked and this may have led to the relatively attenuated hypoxic response in this group. A recent study of isolated rat lungs demonstrated that expression of the cyclooxygenase isozyme, cyclooxygenase-2, during hypoxia was enhanced after nitric oxide synthesis inhibition(22). Whether or not this inducible enzyme is upregulated in newborn lungs after nitric oxide synthase inhibition requires further study.

As in previous studies in which the inflow-outflow occlusion technique was used to partition pulmonary vascular resistances(3, 10, 13), hypoxia appeared to act predominantly on the relatively compliant vessels of the middle segment. Indeed, even control lungs demonstrated a significant increase in hypoxicΔPM (Figs. 3 and4). The inflow-outflow and double occlusion techniques do not permit identification of the anatomic limits of the small arteries and veins that comprise ΔPM. However, correlation of data from these techniques with results obtained using other methods to partition pulmonary vascular resistances suggests thatΔPSA includes small arteries <100-200 m. in diameter(10, 23, 24), and ΔPSV is composed of veins of about the same size. Although hypoxia caused a greater percentage increase in resistances across small veins than across small arteries, the absolute increase in ΔPSA was much greater(Table 1), suggesting that HPV occurred mostly in small arteries, as has been shown with other techniques(25, 26).

During normoxia, neither EDNO nor dilator PG appeared to influence resistance across the large noncompliant arteries defined by ΔPA, and both modulators had only modest effects on ΔPM(Fig. 3). During hypoxia, dilator PG appeared to be the dominant modulators of resistance across both large and small arteries(Fig. 4,Table 1), for indomethacin, with or withoutNω-nitro-L-arginine, caused a significant increase inΔPA and ΔPM. EDNO also attenuated the hypoxic response of ΔPM (Fig. 4), but its effect on the small arteries most responsive to hypoxia was significantly less than that of dilator PG (Table 1).

Unlike the other segmental resistances, cyclooxygenase inhibition resulted in a decrease in normoxic ΔPV (Fig. 3). These data suggest that constrictor prostanoids such as thromboxane A2 or PGF had a greater effect on resting venous vasomotor tone than did dilator PG. However, EDNO may have exerted a modest attenuating effect on venous vasomotor tone, for ΔPV of the LNA lungs was significantly greater than that of the INDO lungs during both normoxia and hypoxia(Fig. 3 and4). These data are consistent with previous studies in which cyclooxygenase inhibition reduced venous pressure defined by the micropuncture technique(27), and EDNO activity was greater in isolated veins than in conductance arteries(28).

In conclusion, our data suggest that both dilator PG and EDNO contribute to the maintenance of normal pulmonary vasomotor tone in 6-h-old lamb lungs. Dilator PG attenuated HPV of large and small ateries more than EDNO, but EDNO contributed most to the reduction of normoxic PVR and had a modest attenuating effect on venous tone. In addition, dilator PG synthesis may be enhanced after nitric oxide synthase inhibition. Whether these complementary actions occur in other species is not certain, for cyclooxygenase, but not nitric oxide synthase inhibition increased normoxic PVR in rats, whereas the opposite was true in dogs(4). Furthermore, cyclooxygenase inhibition had no effect on the attenuated hypoxic responses of young newborn rabbit lungs(16). Future studies must further evaluate the roles and interactions between EDNO, dilator PG, and other modulators of pulmonary vasomotor tone during normal and abnormal development of the pulmonary circulation.