Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Congenital diaphragmatic hernia (CDH) occurs in 1/2,000 live births (1). This malformation, which, for many years was thought to be a simple hole in the diaphragm, potentially curable by surgical closure of the defect after birth, now appears as a complex disease. The main features of CDH include lung immaturity, pulmonary vascular bed structural anomalies causing persistent pulmonary hypertension of the newborn (PPHN), and lung hypoplasia associated with left heart underdevelopment (2). Despite significant advances in pre-and postnatal care the mortality of CDH remains unchanged (40-70%) (1, 2). The two main factors that determine outcome in CDH are lung hypoplasia and PPHN (1). Inhaled nitric oxide (NO), a potent and selective pulmonary vasodilator, improves oxygenation in various respiratory disorders of the newborn associated with PPHN (5, 6) but remains ineffective in CDH (7, 8).

The pathophysiology of PPHN in CDH remains poorly understood. Structural changes of the pulmonary vascular bed are well characterized. These include excessive muscularization of the preacinar arteries, a reduced external diameter, and increased medial wall thickness of prealveolar and intraalveolar arteries, resulting in reduced luminal area of these arteries (9). Altered vasoreactivity in CDH result from a fixed (i.e., structural change) or a dynamic component, or from both (10). The dynamic component may relate to dysfunction of the endothelium, which plays a critical role in the modulation of pulmonary vascular tone through the release of vasoactive mediators. By favoring vasodilatation, NO significantly contributes to the regulation of the transitional circulation (10). NO stimulates soluble guanylyl cyclase (sGC), thereby increasing intracellular cyclic guanosine 3',5' monophosphate (cGMP) levels and causing vasodilatation. cGMP is, in turn, rapidly hydrolyzed and inactivated by cGMP-specific phosphodiesterase (PDE-type V) enzymes, which therefore limit the vasodilatory response to NO (14).

We hypothesized that alterations in NO/cGMP pathway may contribute to PPHN in CDH. We therefore conducted pharmacologic studies in vitro to compare the response of fourth generation pulmonary arteries from fetal lambs with surgically induced CDH and control animals to various drugs modulating the NO/cGMP pathway. To ascertain the presence of endothelial NO synthase (NOS-3), we also performed immunohistochemistry in the lungs of these animals.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

The study was approved by the Animal Care and Use Committee of the "Ecole de Chirurgie, Assistance Publique-Hôpitaux de Paris."

Animal Model: Creation of the Diaphragmatic Hernia

Pregnant ewes between 80-85 d of gestation (term = 147 d) were fasted for 24 h before surgery. Ewes were sedated with intravenous pentobarbital sodium (250 mg) administered through an external jugular vein line and anesthetized with lumbar intrathecal dose of 2 ml of 1% xylocaine. Intramuscular Amoxicillin (1 g) was given. Pentobarbital sodium dosing was adjusted so that ewes were sedated but breathed spontaneously throughout surgery. Under sterile conditions, the uterus was exposed through a midline abdominal approach. The fetal lamb's left forelimb was delivered through a uterine incision. After local infiltration of 1 ml of 1% xylocaine, an incision of the left hemithorax at the level of the ninth intercostal space was made to expose the fetal diaphragm. After a short incision of the left hemidiaphragm, the stomach was pulled manually into the thorax. After fetal chest closure, the fetus was placed back into the uterus. Ampicillin (500 mg) was given into the amniotic cavity and the hysterotomy was closed. The abdominal incision of the ewe was closed in two layers.

This animal model reliably mimics pulmonary hypoplasia and the structural changes of the vascular bed observed in humans with CDH (15).

Isometric Tension Studies

Tissue preparationAnimals were anesthetized with intravenous pentobarbital sodium (20-30 mg/kg) and killed at 140 d of gestation by rapid exsanguination. The heart and lung were removed en bloc from the thorax immediately after death and placed in cold Krebs solution (in mM) 118 NaCl, 5.9 KCl, 1.2 MgSO₄, 2.5 CaCl₂, 1.2 NaH₂PO₄, 25.5 NaHCO₃, and 5.5 glucose. Fourth-generation intralobar arteries with an internal diameter ranging from 1.5 to 3 mm were isolated from the left lung and carefully removed to minimize vascular compression or stretch with handling and cleaned under magnification with gentle removal of remnant surrounding adventitia. Arteries were cut in rings of 2 mm length, placed on horizontally oriented thin steel wires attached to a force displacement transducer (Sigma-Aldrich, St. Quentin-Fallavier, France) and suspended in 20 ml of Krebs-Ringer in a glass-jacket muscle bath at 37°C and continuously aerated with 21% O₂ 5% CO₂, and 74% N₂. A continuous recording of isometric force generation was obtained by connecting the transducer to an analog digital computer system (MacLab; AD Instruments Inc., Medford, MA). After the vessels were mounted, they were allowed to equilibrate for 30 min in the bathing solution. The vessels were then progressively stretched to their optimal tension (600 mg), determined by the maximal contractile response to 60 mM KCl. One hour equilibration was allowed after tissues were brought to their optimal tension.

Pharmacologic preparations and experimental designTo investigate the NO/cGMP pathway at different levels, changes in isometric force of phenylephrine-precontracted rings were measured after the cumulative addition of one of the following pharmacologic agents:

1. 1. The endothelium-dependent vasodilator acetylcholine (ACh, 10⁸-10³ M), and calcium ionophore A23187 (10⁸-10⁵ M) to investigate NOS activity;
2. 2. The endothelium-independent vasodilator sodium nitroprusside (SNP, 10¹²-10⁴ M), an NO donor that directly activates sGC;
3. 3. The specific cGMP-phosphodiesterase type-V inhibitor zaprinast (ZAP, 10⁸-10⁴ M).

To exclude the involvement of endogenous cyclooxygenase products, 10⁵ M indomethacin was added to the bath before phenylephrine precontraction of the vessels.

To determine the role of endogenous NO production, the relaxation was determined in the presence or absence of 10⁴ M of the NOS inhibitor N-nitro-L-arginine (L-NA).

The vessels of 11 CDH lambs and 14 age-matched controls were studied. All vessels were exposed to only one drug. For each drug, n refers to the number of analyzed vessels. For SNP, a second challenge was applied after a 45-min recovery period and several rinses of the bath. All drugs were purchased from Sigma-Aldrich.

Immunohistochemistry

To localize NOS-3 in the ovine lungs, we performed immunohistochemistry with a specific mouse monoclonal antibody (Transduction Laboratories, Paris, France). A fragment from the upper left lobe of the lungs (n = 4 in each group) was flash frozen in liquid monochloro-difluoromethane (40°C), and stored at 20°C. Ten-millimeter sections were thaw-mounted on glass slides (Superfrost; CML, Paris, France) and immediately immersed in isotonic paraformaldehyde (4%) in a phosphate buffer (PBS), 1 h at 4°C as previously described (16). Slides were dehydrated through graded alcohol solutions, dried for 10 min and stored at 80°C. Slides were incubated for 30 min in PBS glycine and partially dried around the tissue. Two hundred ml of a phosphate buffer PBS containing 0.2% of serum albumin (SAB fraction V; Boehringer Ingelheim, Mannheim, Germany) and Tween 20 (0.1%) were disposed on each slide. Slides were then incubated overnight with the NOS-3 mouse monoclonal antibody. NOS-3 antigen-antibody complexes were detected using the avidin-biotin procedure and visualized by peroxydase substrate solution.

Statistical Analysis

Relaxations were expressed as percentage of fall in tension from contraction of tissues to phenylephrine. Values are means ± standard error. Intergroup comparisons were performed with Student's t test or a factorial, repeated-measures analysis of variance, as appropriate. Fisher's probable least significant differences test was used for post hoc comparisons. A P value of < 0.05 was considered to be statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Isometric Tension Studies

The contractile response to 10⁶ M phenylephrine did not differ between groups (3,877 ± 1,035 mg in CDH and 4,593 ± 1,239 mg in control animals).

Investigation of the endothelium-dependent vasodilatation: vasorelaxation to ACh

ACh induced dose-dependent relaxation in pulmonary artery rings that was not significantly different between fetal lambs with CDH and control animals in the absence (Figure 1A) or presence (Figure 1B) of the NOS inhibitor L-NA.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Effects of acetylcholine on phenylephrine precontracted fourth-generation pulmonary arteries from 140-d gestational age fetal lambs in control (open circles) and CDH (shaded squares) animals. ACh induced a dose-dependent vasorelaxation in both groups. There was no difference between the two groups in the absence (A) or presence (B) of the NOS inhibitor L-NA.

Vasorelaxation to A23187A23187, in the absence of L-NA, induced a dose-dependent relaxation in pulmonary artery rings from fetal lambs, which was more pronounced in control than in CDH animals (Figure 2A). In the presence of L-NA, the response to A23187 was similar in both groups (Figure 2B).

View larger version (17K): [in this window] [in a new window]   Figure 2.   Effects of A23187 in control and CDH animals on phenylephrine precontracted fourth-generation pulmonary arteries from 140-d gestational age fetal lambs with and without the NOS-inhibitor L-NA. (A) In the absence of L-NA, A23187 induced a dose-dependent vasorelaxation that was more pronounced in control (open circles) than in CDH (shaded squares) animals. *P < 0.05, CDH versus controls. (B) In the presence of L-NA, A23187-induced vasorelaxation was similar in both groups.

Investigation of the endothelium-independent vasodilatation: vasorelaxation to SNP. First stimulation (SNP1)

Relaxation to SNP was impaired in pulmonary artery rings from fetal lambs with CDH as compared with control animals (Figure 3A). L-NA did not affect vasorelaxation in response to SNP in both groups. Second stimulation (SNP2). After washout and 45 min of recovery, pulmonary artery rings were challenged with a second stimulation by SNP. During this second SNP challenge, control rings showed a reduced response as compared with CDH rings (Figure 3B). Moreover, the initial impaired response to SNP in rings from CDH animals disappeared with the second challenge. As a result, the dose-response curve to the second challenge with SNP was superimposable upon that of the first challenge in control rings (Figure 3C).

View larger version (27K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Figure 3.   Comparative dose-response curves to the endothelium- independent vasodilator sodium nitroprusside (SNP) on phenylephrine precontracted fourth-generation pulmonary arteries of control (open circles) and CDH (shaded squares) animals. (A) A first stimulation by SNP induced a dose-dependent relaxation that was more pronounced in control compared with CDH animals. *P < 0.05, CDH versus control. (B) By contrast, a second stimulation by SNP induced a more pronounced relaxation in CDH animals compared with controls. *P < 0.05, CDH versus control. (C) Relaxation of CDH rings to a second SNP stimulation (shaded squares) was not different from the relaxation obtained in control rings with the first SNP stimulation (open circles).

Investigation of PDE-V: vasorelaxation to ZAP.

ZAP induced a dose-dependent relaxation in pulmonary artery rings that was not significantly different between fetal lambs with CDH and control animals (Figure 4). L-NA did not affect vasorelaxation in response to ZAP in all pulmonary artery rings.

View larger version (14K): [in this window] [in a new window]   Figure 4.   Effects of the phosphodiesterase type V inhibitor zaprinast (ZAP) on phenylephrine precontracted fourth-generation pulmonary arteries from 140-d gestational age fetal lambs in control (open circles) and CDH (shaded squares) animals. ZAP induced a dose-dependent vasorelaxation in both groups. There was no difference between the two groups.

Immunohistochemistry

Immunostaining with NOS-3 antibodies revealed the presence of NOS-3 in the endothelium of pulmonary arteries in both control (Figures 5B and 5C) and CDH animals (Figures 5D and 5E). Negative control used to assess selectivity shows absence of immunostaining in slides, which were not treated with the NOS-3 antibody (Figure 5A).

View larger version (142K): [in this window] [in a new window]   View larger version (130K): [in this window] [in a new window]   View larger version (127K): [in this window] [in a new window]   Figure 5.   Immunolocalization of NOS-3 in fetal lambs of control and CDH lungs. The brown color indicates NOS-3 localized to the endothelium. These experiments were performed in four animals and the pictures are representative. (A) Negative control. NOS-3 is apparent in the endothelial layer of conductance (B) and resistance pulmonary arteries (C) in 140-d gestational age control fetal lambs. NOS-3 was also localized in the endothelium of conductance (D) and resistance (E) pulmonary arteries of fetal lambs with CDH. Note intense smooth muscle cell proliferation in resistance arteries of fetal lambs with CDH. Endo, endothelium; M, muscle; AR, conductance artery; ar: resistance arteries.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

While investigating in vitro the NO/cGMP pathway, we found no difference in the pulmonary vasodilator response to ACh between CDH and control animals, suggesting that endothelium-dependent relaxation mediated by ACh is intact in fetal herniated lambs. The similar vasodilator response to ZAP also suggests no increase in PDE-V activity. Conversely, vasodilator responses to calcium ionophore A23187 and SNP were impaired in CDH as compared with control animals.

The investigation of the NO/cGMP pathway in CDH revealed conflicting results. In experimental CDH in the rat, NOS expression (17) and NOS activity (18) were found to be decreased. However, in experimental CDH in fetal lambs, NOS was evidenced by immunohistochemistry in the main pulmonary artery (19), and fourth-generation pulmonary arteries of CDH lambs have basal and stimulated release of NO identical to that of control animals (20). Studies in the fetal lambs are in agreement with our data, because ACh-induced vasorelaxation did not differ between groups in this study (Figure 1), suggesting intact endothelium-dependent vasodilatation. However, A23187-induced vasorelaxation was reduced in pulmonary arteries from fetal lambs with CDH as compared with control arteries (Figure 2A). The apparent discrepancy in endothelium-dependent vasorelaxation can be explained by the fact that A23187 causes release of endothelial relaxing factors by one mechanism, i.e., by increasing endothelial cell calcium influx through receptor-independent mechanism. Conversely, ACh induces vasorelaxation through different signaling pathways related to expression of various muscarinic receptors at the cell membrane surface, leading to synthesis of either NO or of an as yet unidentified endothelium-derived hyperpolarizing factor (EDHF) (21). Reduced endothelium-dependent response to the calcium-ionophore A23187 would suggest impaired endothelium-derived NO-mediated vasodilatation in fetal lambs with CDH. Because ACh causes the release of both NO and EDHF from endothelial cells, the similar vasodilator responses to ACh in control and CDH animals can only be explained by EDHF compensating the decreased synthesis of NO in CDH animals. In this respect, immunostaining for NOS-3 (Figure 5) that was evidenced in the endothelial layer of pulmonary arteries of both CDH and control animals would suggest that NOS is expressed, but its function is impaired. NOS immunoreactivity has already been shown to be present in the same animal model in the main pulmonary artery (19).

Our results obtained with the NO-donor SNP are consistent with reduced responsiveness of sGC to NO in CDH animals. The first response to SNP showed reduced relaxation in CDH as compared with control animals (Figure 3A), suggesting impaired function of sGC, as found in other experimental models of PPHN (22). When pulmonary arteries were challenged with a second SNP stimulation, pulmonary artery relaxation from control animals was significantly reduced as compared with the first response. This is consistent with studies showing that sustained NO exposure ( 45 min) decreases sGC activity in pulmonary artery smooth muscle (23). By contrast, a second challenge of CDH pulmonary arteries with SNP significantly increased relaxation as compared with the first stimulation (Figure 3B). This second response was superimposable to the first SNP response obtained in control animals (Figure 3C). These findings suggest that sGC hyporesponsiveness to NO stimulation may be reversible in this experimental model. The interpretation of this result is difficult. In the clinical setting, Karamanoukian and coworkers observed that a 20-min trial of inhaled NO in newborns with hypoplastic lungs due to CDH or oligohydramnios failed to improve oxygenation before extracorporeal membrane oxygenation (24). Conversely, after decannulation from extracorporeal membrane oxygenation, inhaled NO increased postductal Pa_O2 and improved oxygenation index (24). One explanation was that signaling of the NO-cGMP pathway has become more efficient over time in pulmonary artery smooth muscle cells. This phenomenon may take place over several days or hours. Alternatively, there is evidence that reactive oxygen species diminish NO bioactivity (25). In our present study, the second SNP stimulation may have unmasked increased levels of oxygen radicals in CDH pulmonary arteries, as has been recently suggested in other animal models (26, 27). Superoxide anions might scavenge NO following the first SNP challenge. As NO derived from the first SNP has cleared the excess of superoxide from the tissue, NO is then free to exert its dilatory effect upon the second SNP challenge (26, 27). Excess of superoxide anions in some instances might also explain the effect of recombinant human superoxide dismutase in improving the efficacy of inhaled NO in PPHN (28).

The apparent discrepancy between our recent in vivo studies (29), showing similar vasodilatory responses to SNP between CDH and control animals, and the present data could be due to the difference in sizes (small versus large) of pulmonary arteries which have been tested in both studies. A more likely explanation still is the fact that we did not perform dose-response curves in the previous in vivo study (29), having tested only one dose of infused SNP in the animals for which no difference in the vasodilatory responses was seen. It is likely that testing the animals in vivo with various concentrations of SNP would have revealed some differences in the two groups, as found in the present in vitro study.

Synthesis of the second messenger, cGMP, also occurs following activation of pulmonary artery smooth muscle cells by natriuretic peptides (30). Unlike NO, which stimulates sGC, the natriuretic petides act by stimulating a membrane-bound particulate guanylyl cyclase associated with natriuretic peptide receptors (30). Recent evidence suggests that this pathway is not perturbed in animals with high pulmonary blood flow and pulmonary hypertension as shown by the similar vasodilatory responses of pulmonary arteries and veins to atrial natriuretic peptides in control and diseased animals (27).

Increased PDE-V activity may also account for increased pulmonary hypertension, as suggested by studies with experimental PPHN induced by ductus arteriosus compression (31, 32). However, in experimental CDH, increased PDE-V activity has not been evidenced so far. The study by Irish and colleagues (20) and our study (Figure 4) show no difference in the vasodilator response to the PDE-V inhibitor ZAP.

Our study was restricted to the investigation of fourth-generation pulmonary arteries. However, pulmonary vasomotor tone is mainly regulated by resistive arterioles. Therefore, we can only speculate whether these arteries may have shown altered responsiveness in the NO/cGMP pathway. The isolated perfused lung technique, which allows reliable monitoring of pulmonary vascular resistance, may overcome the difficulties of studying small, resistive arterioles in vitro. Furthermore, we did not investigate the response of pulmonary veins. Yet pulmonary veins from fetal lambs play an important role in the modulation of endothelium-dependent vasomotor tone (33, 34) that may be more important than in arteries (35). Interestingly, Irish and coworkers (20) showed that pulmonary veins from CDH lambs display enhanced vasoconstrictive response to phenylephrine after L-NA pretreatment and a blunted response to A23187 and ZAP. This suggests dysfunctional pulmonary vein and not artery vasomotor tone regulation as the main factor, which accounts for PPHN in CDH.

In conclusion, the similar response to the cGMP phosphodiesterase inhibitor ZAP in control and in CDH animals suggests intact PDE-V activity in vitro in animals with surgically-induced CDH. By contrast, the diverging responses to the endothelium-dependent vasodilators ACh and A23187 indicate a dysfunction of NOS-3 and suggest compensatory increased release of EDHF. Decreased sGC responsiveness was evidenced by the response to the NO donor SNP. This hyporesponsiveness could be overcome by repeating the challenge, a phenomenon that was not seen in control animals. We speculate that PPHN and decreased inhaled NO responsiveness in CDH may in part be due to decreased sGC activity.