Main

At birth, the function of gas exchange is transferred from the uterine-placental vessels to the neonatal lung. Associated with this transition is a dramatic increase in pulmonary blood flow and a precipitous decline in pulmonary vascular resistance. Although many factors contribute to the perinatal reduction of pulmonary vascular resistance, there is abundant evidence to suggest that NO, the active moiety of endothelium-derived relaxing factor, plays an important role(1–6). NO acts, in part, by stimulating sGC to synthesize cGMP, an intracellular second messenger. In vascular smooth muscle cells, increased cGMP levels lead to vasodilation.

cGMP is inactivated by cyclic nucleotide PDEs, a complex family of enzymes which catalyze the hydrolysis of 3′:5′-cyclic nucleotides to nucleoside 5′-monophosphates(7). PDEs are multidomain proteins with distinct catalytic and regulatory sites(8). Seven classes of PDEs are categorized on the basis of substrate specificity, kinetics, regulatory and immunologic properties, and selective inhibition(9). Of the five PDE isozymes described in mammalian lung(10), only PDE5, the cGMP-binding, cGMP-specific PDE, is specific for cGMP hydrolysis and is selectively inhibited by zaprinast (M&B 22948)(7, 11). PDE5 is a major cGMP-binding protein present in the lung (particularly the soluble fraction), and appears to be the predominant cGMP-metabolizing PDE of pulmonary tissues(12, 13). Recently, the potential role for PDE5 in modulating pulmonary vascular responsiveness to exogenous NO administration has been demonstrated(14–16).

Our laboratory(17) and others(18) have reported abundant expression of one of the enzymes responsible for NO synthesis, the constitutive endothelial nitric oxide synthase (NOS III), in the lungs of perinatal rats. NOS III mRNA and protein levels were greater in the lungs of neonatal rats than in adult lungs(17). In addition, sGC subunit mRNA levels and enzyme activity were increased in the lungs of neonatal rats compared with adult rat lungs(19). These results are consistent with an important role for increased NOS III and sGC enzyme activity in perinatal pulmonary vasorelaxation. To further investigate the molecular mechanisms modulating perinatal pulmonary vascular NO responsiveness, we examined PDE5 regulation during pulmonary development. PDE5 mRNA levels and zaprinast-sensitive cGMP hydrolytic activity were found to be most abundant in the lungs of neonatal rats with lower levels present in adult lungs.

METHODS

These investigations were approved by the Subcommittee for Research Animal Studies at the Massachusetts General Hospital.

Isolation of a rat PDE5 cDNA. Degenerate oligonucleotides 5′-CGGGATCCTGYACNCCNATHAARAA-3′ and 5′-CGGAATTCCYTGRTCRAARAAYTC-3′, corresponding to amino acid sequences 443-448 (CTPIKN) and 775-780 (EFFDQG), respectively, of bovine lung PDE5 cDNA(7), were used in a PCR with cDNA prepared from RNA extracted from cultured rat aortic smooth muscle cells (40 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 2 min, and extension at 72°C). A 700-bp BamHI/EcoRI restriction fragment of amplified cDNA was ligated into pBS, and its nucleotide sequence was determined.

RNA blot hybridization. RNA was extracted from lung tissues(excluding hilar structures) harvested from fetal, neonatal, and adult Sprague-Dawley rats using the guanidine isothiocyanate/cesium chloride method(20). Time-dated pregnant rats were used to obtain fetal pulmonary tissues. RNA blots containing 10-μg samples of total lung RNA, prepared as described previously(17), were hybridized overnight with the 32P-labeled BamHI/EcoRI restriction fragment of the rat PDE5 cDNA at 42°C. Filters were washed in 0.2 × SSC (1 × SSC = 15 mM sodium citrate and 150 mM sodium chloride) with 0.1% SDS at 65°C and subjected to autoradiography. To quantitate the amount of RNA loaded on the agarose gels, membranes were subsequently hybridized with a 15-fold molar excess of 32P-labeled oligonucleotide complementary to 18 S RNA, as described previously(17). Nonsaturated autoradiograms were quantitated using volume-densitometric scanning (La Cie Silver Scanner II, Seiko Epson Corp., Japan) and National Institutes of Health Image 1.44 software. To estimate PDE5 mRNA concentrations, the PDE5 mRNA:18 S rRNA ratio was determined by dividing the absorbance corresponding to the PDE5 probe hybridization by the absorbance corresponding to the 18 S oligonucleotide probe hybridization.

Pulmonary cGMP PDE activity. Pulmonary cGMP PDE activity was measured using a modification of the methods of Francis and Thomas(11, 21). Lung tissue was harvested, immediately placed on ice, rinsed in cold PBS, frozen in liquid nitrogen, and stored at-80°C. Lung samples were homogenized in 5 volumes of cold PEM buffer (20 mM sodium phosphate, pH 6.8, 2 mM EDTA and 25 mM β-mercaptoethanol) on ice five times for 10 s each. Homogenates were centrifuged at 16 000 ×g for 40 min at 2°C, and supernatants were assayed for PDE activity. Lung extract (50 μL) was added to a reaction mixture containing 50 mM Tris-HCl, pH 7.5, 10 mM MgCl2, 0.33 mg/mL BSA, 20 μM unlabeled cGMP, and 500 nM [3H]cGMP (Amersham Life Sciences, Arlington Heights, IL) in a final volume of 150 μL and incubated at 30°C for 10 min. The reaction mixture was cooled to 0°C, and 20 μL of a “stop solution” containing 10 mM cGMP, 10 mM cAMP, 50 mM EDTA, 30 mM theophylline, and 100 mM Tris-HCl, pH 7.5, was added. The 5′-GMP generated by cGMP hydrolysis was converted to guanosine by the addition of 20μL of a 10 mg/mL solution of Crotalus atrox snake venom 5′-nucleotidase followed by incubation at 30°C for 10 min. The reaction was terminated by the addition of 1 mL of a solution containing 100μM guanosine, 100 μM adenosine, and 0.15 M EDTA. The reaction mixture was then chromatographed at room temperature using QAE-Sephadex A-25(Pharmacia Biotech, Inc., Piscataway, NJ), previously equilibrated in 20 mM ammonium formate, pH 7.5. [3H]Guanosine in the eluate was quantitated by liquid scintillation spectrometry. Where indicated, 2 μM zaprinast(Rhone-Poulenc Rorer, Dagenham, Essex, UK) was included in the reaction mixture with control samples receiving an equal volume of vehicle (0.05 N NaOH). Lung extracts were diluted in PEM to ensure that less than 20% of the[3H]cGMP was hydrolyzed. The protein concentration of the lung extracts was determined by the method of Bradford(22) using BSA as the standard.

In situ hybridization . In situ hybridization was performed as described previously(17). Lung tissues were perfused and fixed in 4% paraformaldehyde and embedded in paraffin. Paraffin sections (8 μm thick) were dewaxed in xylenes, rehydrated through graded alcohol solutions, and equilibrated in PBS. To prepare an antisense cRNA probe, the rat PDE5 cDNA plasmid was digested with BamHI and incubated with T7 RNA polymerase in the presence of digoxigenin-UTP. Before hybridization, sections were digested with 0.2 N HCl and acetylated with 0.025% acetic anhydride in 0.1 M triethanolamine (pH 8.0) at room temperature. RNA was denatured in 50% formamide/2 × SSC for 15 min at 50°C, and hybridized for 12 h at 50°C with 250 ng/mL antisense or sense digoxigenin-UTP-labeled cRNA probe in buffer containing 50% formamide, 4× SSPE (1 × SSPE = 0.18 M NaCl, 10 mM NaH2PO4, and 10 mM EDTA, pH 7.4), 10 mM sodium pyrophosphate, 1 × Denhardt's solution, 200 mM DTT, 300 μg/mL denatured salmon sperm DNA, 100 μg/mL tRNA, 10% PEG 6000, and 1% SDS. Slides were washed for 2 h at 50°C in 2× SSC with 0.5% Triton X-100. Unhybridized cRNA probe was digested with 50 μg/mL RNase A in 0.5 M NaCl, 10 mM Tris-HCl, and 1 mM EDTA, pH 8.0. The final wash was with 0.2 × SSPE/0.5% Triton X-100 at 55°C for 30 min. Hybridized probe was detected using alkaline phosphatase-conjugated sheep F(ab′)2 anti-digoxigenin antibody (Boehringer Mannheim, Inc., Indianapolis, IN), and antibody binding was detected with X-phosphate/5-bromo-4-chloro-3-indolylphosphate/nitro blue tetrazolium chloride. The sections were lightly counterstained with hematoxylin, preserved under coverglass, and viewed by darkfield microscopy.

Statistical analysis. Comparisons between groups were performed by a factorial model of ANOVA. When significant differences were detected, Scheffe's test was used for post hoc comparison. Significance was accepted at the 0.05 level of probability. All results are expressed as mean values ± SEM.

RESULTS

PDE5 gene expression in rat lung. To investigate the regulation of PDE5 gene expression during pulmonary development, PDE5 mRNA levels were measured in the lungs of perinatal and adult rats. A partial PDE5 cDNA was amplified from rat aortic smooth muscle cell RNA using reverse transcriptase-PCR, and degenerate oligonucleotides corresponding to amino acids 443-448 and 775-780 of bovine lung PDE5(7). A 700-bp BamHI/EcoRI restriction fragment of the amplified cDNA, corresponding to amino acid sequences 492-734 of bovine lung PDE5, was used in all subsequent experiments. Comparison of bovine and rat PDE5 cDNAs revealed that the nucleotide sequences were 86% identical and that the encoded amino acid sequences were 96% identical. The nucleotide sequence of the partial PDE5 cDNA from rat was submitted to GenBank (accession no. U76032). The amino acid sequence predicted from the 3′ portion of the rat cDNA exhibited 38-41% identity with rod (α and β subunits) and cone(α subunit) cGMP-specific photoreceptor PDEs, corresponding to the putative conserved catalytic region located in the 3′ portion of mammalian PDEs(7). A search of the SWISS-PROT and GenEmbl data banks revealed that no protein or nucleotide sequences, other than the PDEs, displayed significant homology with the deduced rat lung PDE5 sequence.

Lungs were harvested from 1-d-old, 8-d-old, and adult rats. Using RNA blot hybridization techniques and the rat PDE5 cDNA probe, PDE5 mRNA was readily detected in the lungs from 1- and 8-d-old rats, whereas only low levels of PDE5 mRNA were found in adult rat lung. The concentration of PDE5 mRNA, as estimated using the PDE5 mRNA:18 S rRNA ratio, was 3- and 4-fold greater in the lungs of 1- and 8-d-old rats, respectively, than in the lungs of adult rats (p < 0.05, for both 1-d-old versus adult and 8-d-old versus adult; Fig. 1).

Figure 1
figure 1

PDE5 gene expression in rat lungs. RNA was extracted from the lungs of three 1-d-old, 8-d-old, and adult rats, hybridized with a radiolabeled rat PDE5 cDNA probe and subsequently hybridized with a radiolabeled 18 S RNA oligonucleotide probe. Data shown are mean values± SEM. *p < 0.05 vs adult.

Full size image

To further study the ontogeny of PDE5 gene expression in rat lung, RNA was extracted from the lungs of fetal rats at 18.5, 19.5, 20.5, and 21.5 d of the 22-d term gestation. PDE5 mRNA was readily detected in fetal lung as early as d 18.5, and the concentrations of mRNA appeared to increase by d 21.5 to levels comparable to those detected in lung samples from neonatal animals (Fig. 2).

Figure 2
figure 2

Pulmonary PDE5 gene expression is regulated during development. RNA extracted from the lungs of individual fetal (d 18.5, 19.5, 20.5, and 21.5 of the 22-d term gestation), neonatal (1 and 8 d after birth), and adult rats was hybridized with a radiolabeled rat PDE5 cDNA probe, and subsequently hybridized with a radiolabeled 18 S oligonucleotide probe. Data shown are representative of five separate experiments.

Full size image

Pulmonary cGMP PDE activity. cGMP hydrolytic activity was measured in lungs obtained from neonatal and adult rats. Lungs from 1- and 8-d-old rats yielded significantly greater cGMP hydrolytic activity (1280± 90 and 300 ± 20 pmol/min/mg of protein, respectively) than lungs from adult rats (40 ± 4 pmol/min/mg of protein, p < 0.01, for both 1-d-old versus adult and 8-d-old versus adult; Fig. 3). In addition, total pulmonary cGMP PDE activity was greater in lungs from 1-d-old rats than lungs from 8-d-old rats(p < 0.01 1-d-old versus 8-d-old; Fig. 3). To estimate the contribution of PDE5 to total cGMP hydrolytic activity, the ability of zaprinast to inhibit pulmonary cGMP hydrolytic activity was measured. Zaprinast, 2 μM, inhibited 52 ± 5% and 56 ± 5% of the cGMP hydrolytic activity in the lungs of 1- and 8-d-old rats, respectively. In contrast, zaprinast inhibited only 18 ± 3% of the cGMP hydrolytic activity of adult lungs (p < 0.01, for both 1-d-old versus adult and 8-d-old versus adult;Fig. 4).

Figure 3
figure 3

Pulmonary cGMP PDE activity is regulated during development. Lung extracts were obtained from 1-d-old (n = 3), 8-d-old (n = 5), and adult (n = 5) rats, and were assayed for cGMP PDE activity. Data shown are mean values ± SEM.+p < 0.01 vs d 8; *p < 0.01vs adult.

Full size image
Figure 4
figure 4

Inhibition of cGMP PDE activity by zaprinast. To estimate the contribution of PDE5 to total pulmonary cGMP hydrolytic activity, the ability of zaprinast to inhibit cGMP hydrolytic activity was measured. Lung extracts were obtained from three 1-d-old, 8-d-old, and adult rats and were assayed for PDE activity in the absence (controls), or presence of 2μM zaprinast. Results are expressed as per cent inhibition of control values. Data shown are mean values ± SEM. *p < 0.01vs adult.

Full size image

Localization of PDE5 mRNA using in situ hybridization. To identify the cells expressing the PDE5 gene in neonatal and adult lung tissues, in situ hybridization was performed with a digoxigenin-labeled antisense rat PDE5 cRNA probe. In neonatal lungs, PDE5 mRNA transcripts were present in vascular smooth muscle cells(Fig. 5, upper and lower left panels). In addition, hybridization signals were abundant in the alveolar walls of the neonatal lung(Fig. 5, lower right panel). In contrast, PDE5 mRNA transcripts were primarily localized to vascular smooth muscle cells in adult lungs (Fig. 5, upper right panel). Hybridization signals were also present in airway epithelium of both newborn and adult lungs(Fig. 5, left and right upper panels). Adjacent sections hybridized with sense cRNA probe or pretreated with RNAase A followed by hybridization with antisense cRNA probe did not exhibit positive in situ hybridization signals (data not shown).

Figure 5
figure 5

Localization of PDE5 mRNA using in situ hybridization. Lung sections were hybridized with the antisense PDE5 cRNA probe. Darkfield photomicrographs are shown of representative rat lung sections. White grains represent positive signals. Vascular smooth muscle(sm), airway epithelium (e) and the lumens of arteries(A), veins (V) and bronchioles (B) are labeled. Upper left panel: lung section from a neonatal rat. In situ hybridization signals are present in vascular smooth muscle cells, and to a lesser degree in airway epithelium. Abundant hybridization signals are present in the alveolar walls compared with adult lung (magnification,×4). Lower left panel: higher magnification view of a lung section from a neonatal rat. In situ hybridization signals are present in vascular smooth muscle cells, and to a lesser degree in airway epithelium (magnification, ×10). Upper right panel: lung section from an adult rat. In situ hybridization signals are localized primarily to vascular smooth muscle cells. Hybridization signals are also present in airway epithelium. Scarce hybridization signals are present in alveolar walls compared with neonatal rat lung sections (magnification, 4×). Lower right panel: higher magnification view of a lung section from a neonatal rat. Abundant hybridization signals are present in the alveolar walls (magnification, ×20).

Full size image

DISCUSSION

In the present study, the ontogeny of PDE5 gene expression and enzyme activity was investigated in the lungs of normal rats. PDE5 mRNA levels were significantly greater in the lungs of newborn rats than in those of adult rats (Figs. 1 and 2). Similarly, total pulmonary cGMP hydrolytic activity was greater in lungs obtained from neonatal rats than in lungs obtained from adult rats (Fig. 3). Of the five PDE isozymes described in mammalian lung, PDEs 1, 2, 3, and 5 may contribute to total pulmonary cGMP hydrolytic activity(23); however, only PDE5 is specific for cGMP hydrolysis and is selectively inhibited by zaprinast (M&B 22948)(10). To estimate the contribution of PDE5 to pulmonary cGMP hydrolytic activity, the fraction of total cGMP hydrolytic activity inhibited by zaprinast was measured. PDE5 purified from bovine lung has a reported IC50 for zaprinast of 300 nM(11), whereas PDE1 has a reported IC50 for zaprinast of 4-52 μM(12, 13, 24, 25). PDEs 2, 3, and 4 all have reported IC50 values for zaprinast of>30 μM(26). The fraction of cGMP PDE activity that was sensitive to 2 μM zaprinast, an estimate of PDE5 activity, was 3-fold greater in lungs obtained from neonatal rats than from adult rats (Fig. 4).

Total pulmonary cGMP PDE activity (Fig. 3), zaprinast-inhibitable cGMP PDE activity (Fig. 4), as well as PDE5 mRNA levels (Figs. 1 and 2), were greater in neonatal animals than in adult animals. Of interest, PDE5 gene expression and enzyme activity were incompletely correlated in neonatal animals. Pulmonary PDE5 gene expression did not differ 1 and 8 d postpartum (Figs. 1 and 2), whereas PDE5 enzyme activity was greater in 1-d-old rats than in 8-d-old rats (Figs. 3 and 4). Possible explanations for this apparent discrepancy include the following: during the immediate postpartum period, the specific activity of the PDE5 enzyme may be increased; the mRNA may be translated more efficiently into enzyme, or the stability of the PDE5 protein may be enhanced. Alternatively, a novel zaprinast-sensitive cGMP PDE isozyme may be present.

Approximately 50% of total pulmonary cGMP PDE activity in neonatal animals and 80% of the activity in adult animals was zaprinast-insensitive (Fig. 4). This finding is most likely due to the activity of other cGMP-metabolizing PDEs present in lung (PDEs 1, 2, and 3). Further studies will be necessary to identify which additional PDEs participate in cGMP hydrolysis in rat lung.

Pulmonary PDE5 gene expression was detected as early as d 18.5 of the 22-d term gestation of the rat (Fig. 2). The pattern of increasing PDE5 gene expression during lung development is similar to the pattern of pulmonary gene expression observed for NOS III and the subunits of sGC(17, 19). The abundance of PDE5, as well as NOS III and sGC, in the lung during the perinatal period suggests an important role for the NO/cGMP signal transduction system at birth.

In situ hybridization revealed that PDE5 mRNA transcripts were more abundant and diffusely present in the neonatal lung than in the adult lung, where cRNA probe hybridization signals were primarily localized to vascular smooth muscle cells. It appears that PDE5 has a more vascular-specific role in the adult lung compared with the neonatal lung. These findings are distinct from previous observations of the cellular localization and distribution of other components of the NO/cGMP signal transduction system in the lung(17, 19). NOS III and sGC subunit mRNA transcripts were barely detectable in adult lung compared with neonatal lung. The apparent discrepancy between sGC subunit and PDE5 mRNA levels in vascular tissues of the adult lung suggests that the role of this vascular PDE5 is to metabolize cGMP produced by other forms of guanylate cyclase, such as particulate guanylate cyclases, which are activated by natriuretic peptides(27).

Mammalian postnatal lung growth is characterized by a profound increase in alveolar and arterial number(28, 29). In the rat, the burst of arterial and alveolar multiplication is maximal between birth and postnatal d 11(29). Normal postnatal pulmonary vascular remodeling is further characterized by connective tissue deposition in the subendothelium and media(30). NO and NO-generating compounds have been shown to modulate the synthesis of extracellular matrix components in cultured rat mesangial cells(31) and vascular smooth muscle cells(32). In addition, NO modulates vascular smooth muscle cell proliferation(33), migration(34), and apoptosis(35). Abundant PDE5 gene expression in the alveolar walls of the neonatal rat lung, along with previous data demonstrating NOS III and sGC subunit mRNA in neonatal alveolar and serosal epithelial cells(17, 19), suggests that the NO/cGMP signal transduction system may play an extravascular role in pulmonary development.

The results of multicenter clinical trials to evaluate the effectiveness of inhaled NO in the treatment of pulmonary hypertension in newborns, suggest that although NO is highly effective for some patients, it is not universally successful(36, 37). A number of mechanisms could account for hyporesponsiveness to inhaled NO in patients with pulmonary hypertension, including decreased sGC enzyme activity, increased cGMP PDE activity, or impaired cGMP-dependent signal transduction. Combined therapy of inhaled NO with a selective PDE5 inhibitor may be more effective than NO alone, in obviating the need for extracorporeal membrane oxygenation, a highly invasive rescue therapy for severely ill newborns with persistent pulmonary hypertension(38). A potential clinical role for PDE5 inhibitors in the augmentation of NO responsiveness is provided by the findings that zaprinast is a pulmonary vasodilator in a neonatal ovine model of chemically induced pulmonary hypertension(39) and increases the duration and sensitivity of vasodilatation in response to inhaled nitric oxide(16) and infused nitrovasodilators(14). In addition, Thusu et al.(40) have recently shown that the combination of zaprinast and inhaled NO was more effective than either agent alone in reducing pulmonary vascular resistance and increasing postductal oxygenation in neonatal lambs with persistent pulmonary hypertension, after prenatal ligation of the ductus arteriosus(40). If similar findings are observed in humans, systemic or inhaled zaprinast(15, 16), or other selective inhibitors of PDE5, may provide a useful adjunctive therapy to inhaled NO. Although these data suggest that PDE5 inhibitors may augment NO responsiveness, selective inhibitors of PDE5 should be used with caution, as our observations suggest that they may exhibit extravascular effects in lung, particularly in the perinatal period.

In summary, before and during the normal pulmonary vascular transition to extrauterine life, the gene encoding PDE5, an enzyme responsible for metabolizing cGMP, as well as enzymes responsible for NO production and its receptor, sGC, are abundantly expressed in the lung. PDE5, NOS III and sGC are less abundant in mature lung. These observations suggest a potential role for the NO/cGMP signal transduction system in pulmonary vascular and extravascular development.