INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The pulmonary circulation of the fetus is characterized by high pulmonary vascular resistance (PVR), with only 8% to 10% of the combined ventricular output reaching the lung in term fetuses (1, 2). With the onset of ventilation, PVR drops dramatically, and pulmonary blood flow increases accordingly. Failure of these processes in the normal transitional circulation results in persistent pulmonary hypertension of the neonate (PPHN), a disease process with significant morbidity and mortality and for which no optimal therapy exists. Therefore, the transitional circulation is an ideal model for investigating both normal endogenous mediators regulating pulmonary vascular tone and pathologic mechanisms resulting in PPHN. Further insight into control of the transition may permit the development of new therapies for PPHN and other types of pathologic pulmonary hypertension.

The biochemical and molecular mechanisms of regulation of PVR during the transition are complex and not fully understood. A number of physiologic and biochemical mechanisms have been delineated as contributing to the transitional circulation (3). Recent work suggests that the nitric oxide (NO)-guanylate cyclase-3',5'-guanylate cyclic monophosphate (cGMP) system plays an important role in regulating PVR in normal as well as in pathologic states, in both the perinatal and mature pulmonary vasculature (7). Nitric oxide (NO) activates soluble guanylate cyclase in vascular smooth muscle (VSM) cells, resulting in an increase in 3',5'-guanylate cyclic monophosphate (cGMP) levels (12). Increased VSM cGMP results in vasodilation through the activation of cGMP-dependent protein kinases (13, 14). With elucidation of the role of the NO-cGMP vasodilation pathway, a great deal of interest has focused on the use of exogenous NO in treating increased PVR. In multiple studies, NO has been reported to be effective in reducing increased PVR both in animal models and clinical trials (15). Interestingly, in clinical trials there appear to be patients, especially neonates, in whom elevations of PVR are relatively refractory to reductions by inhaled NO. These NO-refractory patients suggest that there is either a defect in NO-stimulated cGMP production, or that the "downstream" mechanisms of cGMP-mediated vasodilation are defective. One possible mechanism for a blunted or absent pulmonary vasodilator response to exogenous NO is an abnormal hydrolysis of VSM cGMP, notwithstanding the existence of normal NO-stimulated cGMP production.

Intracellular cGMP concentrations are determined by a balance between the synthesis of cGMP and its degradation. Cyclic nucleotide phosphodiesterases (PDEs) are the enzymatic activities responsible for the degradation of all cyclic nucleotides (21). The mammalian lung contains a number of PDEs, but a cGMP-binding, cGMP-specific PDE termed PDE5 is especially prevalent in whole-lung homogenates (22, 23). It follows that inactivating PDE5 would cause an increase in VSM cGMP and result in pulmonary vasodilation. Our laboratory and others have reported that inhibition of PDE5 with either zaprinast or dipyridamole results in reductions of increased PVR in perinatal and mature animals (24).

Because PDE5 inhibition induces profound pulmonary vasodilation, we speculated that regulation of the hydrolytic activity of pulmonary PDE5 might be an important modulator of perinatal PVR. PDE5 modulation of perinatal PVR could be conceptualized in the following manner: in utero, pulmonary VSM PDE5 hydrolytic activity would be high, resulting in low levels of vasodilating cGMP and thus contributing to the increased PVR of the fetus. With the initiation of ventilation and the onset of the transitional circulation, the hydrolytic activity of pulmonary-vascular PDE5 would fall, resulting in higher levels of VSM cGMP, and thus in the pulmonary vasodilation and falling PVR that characterize the transition. In this paradigm, the changes in PDE5 hydrolytic activity should approximate the known time course of changes in perinatal PVR, with significant reductions in the hydrolytic activity of PDE5 occurring in the first hours of postnatal life, when the most profound reduction in perinatal PVR is known to occur.

To test the hypothesis that modulation of PDE5 activity may be important in regulation of the transitional circulation, we analyzed PDE5 hydrolytic activity in ovine and murine lung homogenates obtained at time points during the early and late phases of the transitional circulation in utero, and compared the results with those obtained in adult animals. Because perinatal regulation of pulmonary-vascular PDE5 cGMP hydrolytic activity could occur at any of three levels (i.e., transcriptional, translational, or posttranslational modification of PDE5), we measured the amount of PDE5 messenger RNA (mRNA) with a quantitative ribonuclease protection assay (RPA) and the amount of PDE5 protein with a quantitative immunoblot technique at each of the perinatal and adult time points.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Preparation of Lung Tissue

The use of animals in these experiments was approved by the University of Washington Animal Use and Care Committee. Animal care was in accordance with all relevant American Association of Laboratory Animal Care (AALAC) guidelines.

Sheep. Pregnant, mixed-breed ewes of known gestational dates and Q-fever-negative serology were obtained from the University of Oregon. Ewes that were to undergo cesarean delivery had free access to food and water until 12 h before surgery, at which time feed was removed but free access to water was continued. Ewes that proceeded to spontaneous vaginal delivery were allowed continual access to food and water both before and after delivery. Lambs killed as fetuses or at 1 or 4 h of postnatal age were delivered by cesarean section. These animals were all of 143 to 146 d gestation (term = 147 d). Two anesthesia protocols were used for cesarean sections: 40 mg of subarachnoid tetracaine with 5 mg/kg intravenous ketamine for sedation or, if adequate spinal anesthesia could not be obtained, general anesthesia, in which anesthesia was induced with pentobarbital 5 mg/kg given intravenously, with the trachea intubated and anesthesia maintained with 1 to 2% inspired halothane. In animals killed as fetuses, great care was taken to assure that the fetus never breathed. To assure this, a forelimb was delivered prior to delivery of the fetal head, an intravenous line was begun, and the animal was euthanized by the injection of 3 ml of euthanasia solution (each ml contains 390 mg pentobarbital, 50 mg phenytoin, 10% [vol/vol] ethanol, 18% [wt/vol] propylene glycol, 0.003688 mg rhodamine B, 2% [vol/vol] benzyl alcohol, and water q.s.). After killing, the chest was immediately opened and the lungs removed and prepared as subsequently described. The animals to be killed at 1 and 4 h of postnatal age were delivered from the uterus and the oropharynx was immediately suctioned. The animals were stimulated to breathe. The exact time of onset of the first breath was noted, and this was used as the reference for the time of killing. A forelimb intravenous line was started and an initial bolus of 50 ml of 5% dextrose in lactated Ringer's solution was given. The intravenous solution was continued at a rate of 50 ml/h. Animals continued to spontaneously breathe room air until the time of killing. Animals were kept warm with the aid of heating blankets and warming lights. Animals were killed with 3 ml of intravenous euthanasia solution at 1 or 4 h after their first breath, and the lungs were immediately removed and processed. After delivery of the fetuses or lambs, ewes were killed with 30 ml of euthanasia solution, the chest was immediately opened, and the left lower lobe was removed for tissue processing as subsequently described. Animals killed at 4 or 7 d of postnatal age were born by spontaneous vaginal delivery. Lambs were fed by ad libitum by the ewe to which they had been born. At the appropriate times after delivery, the lambs were killed with 3 ml of euthanasia solution, the chests were opened, and the lungs removed and immediately prepared as subsequently described. The ewes were killed and a left thoracotomy was performed, with removal of the lower left lobe as subsequently described.

Immediately after removal of the lamb heart-lung blocks (or, in the case of ewes, a single left-lung segment), pulmonary parenchymal tissue was rapidly dissected free of airways greater than approximately 1 mm in diameter. Sections of lung tissue approximately 2 to 3 cm in diameter were immediately frozen in dry ice and liquid nitrogen. The lung tissue was stored at 70° C.

Tissue homogenization. Frozen lung samples were weighed and immediately homogenized on ice in precooled (4° C) homogenization buffer in a preparative Polytron blender (Baxter, McGaw Park, IL) for 1 min. Tissue homogenization was done at a weight/volume ratio of 1 g tissue/4 ml of homogenization buffer. The homogenization buffer contained 40 mM Tris-HCl, pH 7.5; 15 mM benzamidine; 15 mM 2-mercaptoethanol; 1 µg/ml pepstatin A; 1 µg/ml leupeptin; 5 mM ethylene diamine tetraacetic acid (EDTA); 1 mM phenylmethylsufonyl flouride (PMSF); and 20 µg/ml antipain. After homogenization, samples were centrifuged at 250 × g for 10 min at 4° C to remove any large unhomogenized material. The pellet from this initial centrifugation did not contain any PDE activity. The supernatant was centrifuged at 100,000 × g for 1 h at 4° C. The supernatants (or the soluble fraction) from the 100,000 × g centrifugation were diluted in a final concentration of 20% glycerol (vol/vol). The pellets were placed in 50% of the initial homogenization-buffer volume, containing a final concentration of 20% glycerol, and were resuspended with three strokes of a 15-ml Wheaton glass-pestle homogenizer. All samples were stored at 70° C until analysis. Studies indicated that using either fresh or frozen lung, as well as freezing the supernatant or pellet fractions for up to 6 mo, did not alter the PDE results described subsequently.

Mice. C57Bl/CJ mice of known gestational age were obtained from BMK Laboratories (Kent, WA). Dams were allowed free access to food and water. Starting at 24 h before gestational term, the dams were observed continuously to determine the exact time of delivery of each pup. Pups were killed by decapitation at 10 min, 1 h, 4 h, 24 h, 48 h, 4 d, 7 d, 14 d, and 21 d after delivery. Animals representing each of these time points were taken from each litter. After killing, the chest was opened, and pulmonary parenchymal tissue was dissected free of the airways and immediately prepared as subsequently described. After the killing of the last pup, the dams were killed by CO₂ asphyxiation and the lungs were removed and prepared. For the fetal samples, the pregnant dams were killed and the uterus was immediately opened and the fetal pups removed and decapitated. The fetal lungs were then immediately removed and prepared. The adult lung homogenates were made from three postpartum dams and two prepartum dams. A subsequent series of experiments with mice of known gestational age was performed in an identical fashion, except that after killing the lung tissue was immediately frozen between blocks of dry ice and then stored at 70° C until samples of the same developmental age (fetal, 1 h, 2 h, and so forth) underwent the procedure for RNA extraction as subsequently described.

Tissue homogenization. Because of the extremely small size of the mouse-lung samples, tissue preparation was slightly different than that described earlier for the sheep lung. Lung samples were weighed and immediately homogenized. Homogenization was done in the same homogenization buffer as described earlier, at a ratio of 10 mg tissue to 300 µl homogenization buffer. After homogenization, samples were centrifuged at 16,000 × g for 20 min at 4° C. The supernatants were diluted in a final concentration of 20% glycerol (vol/vol), aliquoted into 0.15-ml samples, and stored in capped cryogenic tubes at 70° C. The pellets were placed in 0.15-ml of homogenization buffer containing a final concentration of 20% glycerol, and were resuspended with three strokes of a 15-ml Wheaton glass-pestle homogenizer. These pellet or particulate fractions were stored at 70° C.

Measurement of PDE5 Activity

The PDE activity in the ovine and murine samples was measured in an identical fashion, using a modification of the method of Mumby and coworkers (27). The standard PDE reaction buffer contained 40 mM 3-(N-morpholino)propanesulfonic acid (MOPS), pH 7.5; 0.8 mM ethylene glycol-bis-(-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA); 15 mM MgOAc; 0.2 mg/ml bovine serum albumin (BSA); cold cGMP 1 µM; and [³H]cGMP 10⁵ cpm. In order to determine PDE5 activity, 1 µM cGMP in the presence and absence of 4 µM zaprinast was used. As an additional control, some assays were done in the presence of EGTA to ensure that the activity observed was not due to PDE1/calcium-calmodulin dependent PDE. In order to assure appropriate linear kinetics of cGMP hydrolysis, a titration curve was generated for each sample to determine the sample volume required for 20% to 30% hydrolysis of cGMP. All samples were assayed in triplicate. Protein concentrations were determined by the method of Bradford (28). The specific cGMP-hydrolytic activity of PDE is expressed as pmol cGMP hydrolyzed/minute/milligram of measured protein.

PDE5 activity was determined by the difference in activity of the enzyme in the presence and absence of 4 µM zaprinast. In initial experiments, the IC₅₀ for zaprinast in both fetal and adult sheep lung tissue was determined by adding zaprinast in final concentrations ranging from 0.001 to 8 mM and measuring the cGMP-hydrolytic activity. The IC₅₀ for inhibition of PDE cGMP hydrolysis was 4 µM for both fetal and adult lung homogenates. Because similar results were obtained with mice, this concentration of zaprinast was used in subsequent PDE assays with tissue from animals of all ages. With this common methodology, we were unable to determine the absolute amounts of PDE5 activity. However, because such low zaprinast concentrations were used, they are unlikely to have interfered with any other known PDE, and were probably very appropriate for comparing relative amounts of PDE5 activity in comparative samples.

Measurement of PDE5 Protein by Immunoblot Analysis

Peptide synthesis. Using the published sequence of bovine PDE5 (29), we selected a peptide of 16 amino acids from the C-terminal region, with the sequence C-R-K-N-R-Q-K-W-Q-A-L-A-E-Q-Q-E-K-OH, for synthesis as a PDE antigen. This peptide region, from amino acids 856 to 872, was chosen on the basis of its antigenicity as predicted by PIR-protein (Genetics Computer Group [GCG], Madison, WI), and because this sequence is substantially conserved in all PDE5 enzymes cloned to date. The peptide was synthesized by the Department of Pharmacology Core Protein synthesis facility at the University of Washington. The peptide was conjugated to keyhole limpet hemocyanin (KLH) (30).

Antibody production and characterization. Rabbits were used for the production of antibodies. The antiserum titer and specificity of the antibodies to PDE5 were determined by immunoblot analysis. Partly purified bovine lung PDE5 and albumin-peptide conjugate were used as controls. Partly purified bovine lung PDE5 was obtained as previously described by Thomas and coworkers (31). Staining with Coomassie brilliant blue after 8% sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) revealed 10% purity. With methods described by Yan and colleagues, immunoblot analysis demonstrated the potency and specificity of the antibody for bovine, sheep, rat, and mouse PDE5 at 93 kD (32). Initial immunoblot studies indicated that concentrations of purified bovine PDE5 as low as 50 ng could be detected at an antibody dilution of 1:10,000. As a control, preimmune serum at 1:10,000 did not reveal either bovine or ovine PDE5 at 93 kD. The ovine protein at 93 kD and the purified bovine PDE5 were competitively blocked by antibody at 1:10,000 in the presence of c-terminal peptide (100 g/ml), demonstrating specificity of the antibody for PDE5. Additional control immunoblots demonstrated that the antibody did not bind to cloned, expressed PDE1a, PDE1b, PDE1c, PDE2, PDE3, or PDE4, and that it was specific for PDE5 (data not shown).

Quantitative immunoblot analysis of pulmonary PDE5. Quantitative PDE5 immunoblot analysis was performed in identical fashion for ovine and murine samples. Sheep supernatant (50 µg) or pellet (except for 25 µg of pellet from fetal sheep of 1 h after delivery) fractions were mixed with 2× sample buffer and boiled for 5 min. SDS-gel electrophoresis and immunoblot procedures were completed as previously described (32). Mouse supernatant (50 µg) and pellet (25 µg) were used in immunoblot analysis. As internal controls for quantitative immunoblotting, both 0.25 µg of bovine PDE5 and 50 µg of the same sample of sheep lung supernatant (a 2-h postnatal-age sample) were run on all gels. Prestained Kaleoscope broad-molecular-weight standards (Bio-Rad, Hercules, CA) were also run on all gels.

In order to delineate the concentration of ovine pulmonary PDE5 protein in supernatant and pellet fractions expressed during development, as well as to account for variation in enhanced chemiluminescence (ECL) results, standard curves of bovine PDE5 at concentrations ranging from 0.0625 to 0.5 µg were developed at 1 through 10 s of ECL measurement. These experiments demonstrated that ECL was linear at 3 and 5 s over the range of added bovine PDE5 protein. Thus, all subsequent Western blots were developed at 3 and 5 s. These immunoblots were scanned on a Hewlett-Packard ScanJet 4C (Boise, ID), using DATA Scan II software, and were analyzed with the NIH Image program. With the NIH-image program, each lane containing bovine PDE or sheep-lung protein was analyzed vertically, with the backgrounds subtracted, and the number of pixels under the curves was determined. A quantitative comparison of ovine PDE5 immunoblots was made, with knowledge that all samples were on the linear portion of the ECL exposure curves and that the same internal standards were present on each immunoblot.

Measurement of PDE5 mRNA by Ribonuclease Protection Assay

Total RNA extraction and mRNA extraction. RNA extraction and quantitative RPA analysis were performed in identical fashion on ovine and murine samples, with the exception that individual RNA preparations were made on ovine lung samples, whereas murine lung samples of the same developmental stage were pooled prior to RNA extraction. Sheep- and mouse-lung RNAs were extracted through a modification of the single-step method described by Chomoczyuski and Sacchi (33).

A purified mRNA preparation was made from total RNA, using reagents and protocols supplied in the FastTrak 2.0 mRNA Isolation Kit. The mRNA was then quantitated via UV spectrophotometry, and was used in the preparation of a complementary DNA (cDNA) library. Sheep-lung cDNA was made using the Marathon kit (Clontech, Palo Alto, CA). This cDNA pool was used to obtain amplified sheep- or mouse-lung PDE5 and glyceraldehyde-3-phosphate dehydrogenase (G3PDH) gene fragments via the polymerase chain reaction (PCR), as subsequently described.

Generation of ovine PDE5 PCR fragments. Ovine PDE5 gene fragments were made with the PCR. PDE5-specific oligonucleotides based on bovine PDE5 sequences (GCG accession no. L16545) were synthesized in the University of Washington Department of Pharmacology Core Molecular Biology Facility. Two sets of PCR primers were developed that would produce nonoverlapping PCR products of PDE5. The first set of primers was designed to produce a PCR product corresponding base pairs 1687 to 2399 of bovine PDE5. The primers used to generate this product were sense: 5'-TGCAGTCCTTAGCGGCTGCT-3'; antisense: 5'-CCGTTGTTGAATAGGCCAGG-3'. The product of this PCR reaction was named PDE5-1. A second set of primers were designed to cover base pairs 1 to 1046 of bovine PDE5. The primers used to generate this ovine PCR product were sense: 5'-GGGAGGGTCTCGAGGCGAGTTC-3'; antisense: 5'-CTCCAGCAGTGAAGTCTCATAG-3'. The product of this PCR reaction was named PDE5-2. Sense and antisense primers (20 pmol each) were mixed with 5 µl of the cDNA pool described earlier. Fragments generated were purified on low molecular protein (LMP) agarose-gel, and were resuspended in 10 µl of 10 mM tricine-KOH (pH 7.5)-0.1 mM EDTA buffer.

Generation of murine PDE5 PCR fragments. Murine PDE5 gene fragments were also made via PCR. One PCR primer was developed. The primer was designed to produce a PCR product corresponding to base pairs 2140 to 2613 of bovine PDE5. The primers used to generate this product were sense: 5'-TGATCCTTAATAGTCCTGGC-3'; antisense: 5'-TCTTTCTGCAGCCGTCCAGC-3'. The PCR reaction was completed as previously described.

Cloning of ovine and murine PCR fragments. Polyadenine (poly-A) tails were added to PCR fragments to increase ligation efficiency, using Tailing Reaction Mix (Promega, Madison, WI). Escherichia coli (INVaf; Invitrogen, Carlsbad, CA) was transformed through the ligation reactions described previously, using the protocol and reagents of the Original TA Cloning Kit (Invitrogen). Plasmids generated in this way were analyzed via restriction mapping and sequencing. Two clones of ovine PDE5 were obtained. The clone that covers a portion of the catalytic domain was named PDE5-1, whereas a nonoverlapping clone located 5' to PDE5-1, and which covers the cGMP-regulatory domains, was denoted as PDE5-2. Clone PDE5-1 was found to be 95% homologous to a 3' section of the catalytic domain of bovine PDE5. Clone PDE5-2 was 96% homologous to the cyclic-G-binding domain of bovine PDE5. The positions of these clones and the position of the peptide sequence to which the antibody was made relative to those of bovine PDE5 are shown in Figure 1.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Position of sheep and mouse PDE5 probes and PDE5 antibody relative to bovine PDE5.

Production of probes for internal control of total RNA loading for analysis of mRNA levels. To produce G3PDH clones, consensus sequence (human, mouse, and rat) primers for PCR amplification of G3PDH were purchased from Clontech. The 5' primer had the sequence 5'-ACCACAGTCCATGCCATCAC-3'. The 3' primer had the sequence 5'-TCCACCACCCTGTTGCTGTA-3'. Using the method described previously for the production of PDE5 clones, ovine and murine G3PDH clones were obtained. These G3PDH clones were confirmed by sequencing. The ovine lung G3PDH cDNA clone was 93% homologous to mouse G3PDH. These clones were used for the production of G3PDH RPA probes as described subsequently.

Production of Bluescript RNA controls. Because it was not known whether G3PDH levels were indeed constant throughout development in either ovine or murine lung, a second control was used to confirm both RNA sample recovery and to provide a standardized gray scale for autoradiography exposure. For this procedure, unlabeled Bluescript RNA (BSSK-RNA) was transcribed from Bluescript plasmid (Stratagene, La Jolla, CA) that had been cut with Tfi1 restriction enzyme. This RNA was then treated with deoxyribonuclease-1 (DNase 1), pherol/chloraform/isoamyl alcohol (PCI) (25:24:1)-extracted, and precipitated with ethanol as previously described. The RNA was then resuspended in 0.5% SDS in diethylpyrocarbonate (DEPC)-treated H₂O and quantitated with UV spectrophotometry. The RNA produced in this fashion was added to sheep lung RNA samples prior to RPA analysis. By adding and protecting a known quantity of exogenous BSSK-RNA in every RPA reaction, the efficiency of experimental manipulations could be tracked, and each RPA reaction had an internal standard that thereby permitted quantitative comparison of the amounts of protected G3PDH, PDE5-1, or PDE5-2. In addition, in all RPA gels, a gray scale was also made with BSSK-RNA, in order to determine the linearity of autoradiography images. This was done by running 6.5 ng of BSSK-RNA in an RPA reaction as subsequently described. The resultant, protected RNA was resuspended in Gel Loading Buffer (RPAII Kit, Ambion, TX), and five dilutions of protected BSSK RNA, spanning a 20-fold range of concentrations, were run in the lanes adjacent to experimental samples. This resulted in all RPA gels having a standardized gray scale that permitted quantitative comparison of RPA autoradiographic results among different gels and exposures.

Production of probes for ribonuclease protection assay. Both PDE5-1 and PDE5-2 clones were used to make ovine probes for RPA analyses. Plasmids were cut with appropriate enzymes and ³²P-labeled transcripts were made with the MaxiScript system (Ambion). The respective sizes of the RPA protected fragments for PDE5-1, PDE5-2, G3PDH, and BSSK are shown in Figure 1. Purified probes (1 µl) were counted on a liquid scintillation counter (probes with counts < 300,000 CPM/µl were discarded).

RPAs were begun by adding 1.5 g Bluescript RNA to each 15-µg sample of sheep- or mouse-lung RNA. This was mixed with 1 × 10⁶ cpm of one or more of the probes described previously. Nucleic acid was precipitated with 0.5 M NH₄OAc and 70% ethanol for 15 min at 70° C, and was centrifuged (16,000 × g, for 15 min at 4° C). Pellets were resuspended in 20 µl of hybridization buffer (80% deionized formamide; 100 mM sodium citrate, pH 6.4; and 1 mM EDTA) and denaturated (3 min at 90° C and 10 sec of vortexing), and were incubated overnight at 45° C. The next day, selected samples were digested for 1 h at 37° C with a mixture of ribonuclease A (RNase A) (0.9 U/ml) and RNase T1 (36 U/ml) in a total volume of 220 µl. The RNA was then precipitated, air-dried, and resuspended in 4 µl of gel-loading buffer, subjected to SDS-PAGE (5% acrylamide/8 M urea), and visualized by autoradiography.

Materials

[³H]cGMP (1 Ci/mmol) was purchased from Amersham Corp. (Arlington Heights, IL). Zaprinast was obtained from May and Baker. All other chemicals were obtained from Sigma Chemical Co. (St. Louis, MO), except for those noted otherwise.

Data Analysis

Statistical analysis was done by analysis of variance (ANOVA) with Tukey's test for significance. A value of p < 0.05 was regarded as significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Measurement of PDE5 Activity

Total cGMP hydrolytic activity (under the assay conditions of 1 µM cGMP, EGTA, and absence of zaprinast) at the different developmental time points in ovine samples are shown in Figure 2A. The cGMP hydrolytic activity under identical assay conditions but in the presence of 4 µM zaprinast is shown in Figure 2B. Comparable analyses of total cGMP hydrolytic activity and zaprinast-suppressible cGMP hydrolytic activity at similar developmental time points in the mouse are shown in Figure 3A and B, respectively. Several facts emerge from the analysis of these data. First, comparison of Figure 2A with B and 3A with B shows that at all developmental time points, zaprinast inhibited approximately 50% of the cGMP hydrolytic activity. Thus, under the assay conditions used, the majority of cGMP hydrolytic activity at all time points was attributable to PDE5 in both the ovine and murine lungs. Second, at all developmental time points in the ovine model, the majority of the cGMP hydrolytic activity resided in the soluble fraction and the marked developmental changes in cGMP hydrolytic activity occurred only in the soluble fraction. Moreover, most importantly, both the ovine and murine models showed strikingly similar patterns of developmental change in cGMP hydrolytic activity in lung homogenate supernatants. As an additional control, PDE5 activity for nonpregnant ewes was determined (data not shown). No difference in PDE5 activity was observed between pregnant and nonpregnant ewes.

View larger version (40K): [in this window] [in a new window]   View larger version (25K): [in this window] [in a new window]   Figure 2.   (A) Developmental changes in sheep cGMP PDE activity in the presence of 1 µM cGMP and EGTA. cGMP PDE activity (pmol/min/mg tissue) was determined in 140-gestational-day fetuses (n = 6), at 1 h after delivery (n = 4), at 4 h after delivery (n = 4), at 4 d (n = 4), at 7 d (n = 4), and in adults (n = 17). The presence of EGTA suppressed PDE1/calcium-calmodulin-dependent activity. Hatched bars represent soluble cGMP PDE activity; solid bars represent particulate cGMP PDE activity. Alterations in cGMP PDE activity during development occurred only in the soluble fractions. A significant reduction in cGMP PDE activity occurred during the early transition from fetus to 1 h after delivery (Tukey B test with significance level = 0.05). A secondary increase or "rebound" in cGMP activity occurred at 4 to 7 d (p < 0.05), with a fall in adulthood. (B) Developmental changes in sheep cGMP PDE activity (pmol/min/mg tissue) in the presence of 1 µM cGMP, EGTA, and zaprinast at 4 µM. Zaprinast-sensitive cGMP PDE activity is significant only in soluble fractions. Under the conditions of PDE1 suppression, these data demonstrates that most of the cGMP-PDE activity during sheep-lung development is due to PDE5.

View larger version (57K): [in this window] [in a new window]   View larger version (27K): [in this window] [in a new window]   Figure 3.   (A) Developmental changes in mouse cGMP PDE activity in the presence of 1 µM cGMP and EGTA. cGMP PDE activity (pmol/min/mg tissue) was determined in 140-d gestational-day fetuses (n = 8), 10 min after delivery (n = 3), at 1 h after delivery (n = 6), at 4 h after delivery, (n = 4), at 24 h (n = 6), at 2 d (n = 4), at 4 d (n = 4), at 7 d (n = 4), at 14 d (n = 3), at 21 d (n = 3), and in adults (n = 10). The presence of EGTA suppressed PDE1/ calcium-calmodulin dependent activity. Hatched bars represent soluble cGMP PDE activity; solid bars represent particulate cGMP PDE activity. As with sheep development, changes in cGMP PDE activity occurred only in the soluble fractions. As with sheep, a significant reduction in cGMP PDE activity occurred during the early transition from fetus to 1 h after delivery (Tukey B test with significance level = 0.05); however, this reduction in PDE activity was observed by 10 min after delivery. As with sheep-lung development, a secondary rise or "rebound" of cGMP activity occurred in late phases of development (p < 0.05); however, the decline of cGMP PDE activity reached adult level by 21 d. (B) Developmental changes in mouse cGMP PDE activity (pmol/min/mg tissue) in the presence of 1 µM cGMP, EGTA, and zaprinast at 4 µM. As with sheep studies, zaprinast-sensitive cGMP PDE activity was significant only in soluble fractions. Under the conditions of PDE1 suppression, these data demonstrate that most of the cGMP-PDE activity during mouse-lung development is due to PDE5.

Measurement of PDE5 Protein by Immunoblot Analysis

In Figure 4A and B, sheep PDE5 protein as measured by immunoblot analysis was determined in supernatant and pellet fractions at various developmental times. As with PDE5 enzymatic activity, PDE5 protein expression in the supernatant was also increased in near-term fetal sheep, and fell significantly by 1 h. Similarly, in late perinatal development, PDE5 protein expression in the supernatant rose and subsequently fell during adulthood. The PDE5 protein expression of the particulate fraction did not change significantly with maturation of the sheep.

View larger version (45K): [in this window] [in a new window]   View larger version (34K): [in this window] [in a new window]   Figure 4.   (A) Representative immunoblots of sheep PDE5 protein during development in soluble and particulate fractions. Immunoblots at 3 min of exposure were analyzed with DATASCAN II and NIH Image Software, and were compared with PDE5 curves as discussed in METHODS. The molecular weights (Mr × 103) are indicated on the left side of the blots. Lane 1 expressed as STD (standard) contains partly purified PDE5 bovine standard. The 2-h standard is the same in all blots. In the particulate immunoblot, the fetal sample was loaded with 25 µg as compared with all other samples, in which the loading was 50 µg. The soluble fractions expressed the only changes in PDE5 protein. In the soluble immunoblot, a significant decrease in PDE5 protein was observed from the fetus to 1 h after delivery, with a secondary increase in PDE5 protein at 4 and 7 d. (B) Sheep PDE5 protein as determined by immunoblot analysis during development. PDE5 protein (µg/mg tissue) was determined in 140-gestational-day fetuses (n = 6), at 1 h after delivery (n = 4), at 4 h after delivery (n = 4), at 4 d (n = 4), at 7 d (n = 4), and in adults (n = 6), as described in METHODS. PDE5 protein alterations were observed only in the soluble fractions, correlating with PDE5 activity changes. PDE5 protein was elevated in late gestational fetuses, and decreased significantly by 1 h after delivery (p < 0.05). Although a "rebound" in PDE5 protein occurred at 4 to 7 d, the increase did not correlate with PDE5 activity, suggesting a posttransitional regulation of PDE5.

In order to determine whether the foregoing developmental changes were species specific, mouse PDE5 protein was also determined in supernatant and pellet fractions at various developmental time points, using immunoblot analysis (Figure 5A and B). As in the case of sheep development, PDE5 protein expression in the supernatant was also increased in near-term fetal mice and fell significantly by 1 h. Additionally, in late perinatal development, PDE5 protein expression in the supernatant rose, and subsequently fell during adulthood. However, in the mouse PDE5 particulate fraction, the total protein did not change significantly, but the SDS mobility changed after 14 d, suggesting a possible isotype change or alteration in posttranslational activity.

View larger version (48K): [in this window] [in a new window]   View larger version (53K): [in this window] [in a new window]   Figure 5.   (A) Mouse PDE5 protein as determined by immunoblot analysis during development. PDE5 protein (µg/mg tissue) was determined in 140-gestational-day fetuses (n = 6), at 10 min after delivery (n = 3), at 1 h after delivery (n = 6), at 4 h after delivery (n = 4), at 24 h (n = 4), at 48 h (n = 4), at 4 d (n = 4), at 7 d (n = 4), at 14 d (n = 3), at 21 d (n = 3), and in adults (n = 6), as described in METHODS. As with sheep PDE5 protein, mouse PDE5 protein alterations were observed only in the soluble fractions. PDE5 protein was increased in late gestational fetuses and decreased significantly by 10 min after delivery (p < 0.05). Although, a "rebound" of PDE5 protein occurred at 4 to 7 d, with a decline to adult values by 21 d, the protein level correlated with changes in activity. (B) Representative immunoblots of mouse PDE5 protein during development in soluble and particulate fractions. The immunoblot at 3 min exposures was analyzed with DATASCAN II and NIH Image Software and compared with PDE5 curves as discussed in METHODS. The molecular weights (Mr × 103) are indicated on the left side of the blots. Lane 1, expressed as STD (standard) contains partially purified PDE5 bovine standard. In the immunoblot containing the soluble fractions, the fetal PDE5 protein is increased, with a decrease observed by 10 min and lasting until the secondary rise, observed at 7 d. In the particulate immunoblot, PDE5 protein does not change until a slight rise occurs at 7 and 14 d. However, there is a shift in the molecular weight of PDE5 at 7 d, which is recognized by the antibody and is specific. This alteration in particulate PDE5 may be due to alterations in isozymes or in phosphorylation.

Review of Figures 2-5, demonstrates a developmental change in PDE5 activity and protein in the supernatant fraction, and a developmental alteration in PDE5 protein only in the particulate fraction from mice. Also, there was a correlation between increases in developmental cGMP hydrolytic activity and increases in PDE5 protein expression in the supernatant fraction.

Measurement of PDE5 mRNA by Ribonuclease Protection Assay

RPAs for PDE5 mRNA, were done on fetal sheep (n = 6) and sheep at 1 h (n = 4), 4 h (n = 4), 4 d (n = 4), and 7 d (n = 3) after delivery, and adult sheep (n = 8). As a control in the sheep assays, an internal standard, G3PDH was used. Figure 6 demonstrate that sheep PDE5 message was increased in the fetal state, with a rapid fall by 1 h after delivery. In late transition, a similar rise in message was observed, with a fall to adult low levels by 21 d. These results are similar to the results obtained for PDE5 protein and activity in the supernatant.

View larger version (31K): [in this window] [in a new window]   Figure 6.   Developmental changes in sheep PDE5 mRNA levels as assessed by ribonuclease protection assay (RPA). PDE5 mRNA was determined in 140-gestational-d fetuses (n = 4), at 1 h after delivery (n = 3), at 4 h after delivery (n = 3), at 4 d (n = 3), at 7 d (n = 3), and in adults (n = 4), as described in METHODS. PDE5 mRNA is expressed as the ratio of PDE5 mRNA pixels to G3PDH pixels. Sheep PDE5 mRNA was elevated in late gestational fetuses and decreased significantly by 1 h (p < 0.05). An increase in PDE5 mRNA occurred at 4 to 7 d, with a decline in adults. The developmental changes in sheep PDE5 mRNA parallel the changes in PDE5 protein and activity.

Mouse RPA assays were done on fetal animals (n = 3), and mice at 1 h (n = 3), 4 h (n = 3), 2 d (n = 3), 4 d (n = 3), and 7 d (n = 3) after delivery, and adults (n = 3). As a control in the mouse assays, an internal G3PDH standard was used. As with PDE5 mRNA levels in sheep, mouse PDE5 message was elevated in the near-term fetal state, with a rapid fall by 1 h. Figure 7 also demonstrates the late transitional rise and fall of mouse PDE5 mRNA, which correlated with mouse PDE5 protein and activity in the supernatant.

View larger version (25K): [in this window] [in a new window]   View larger version (51K): [in this window] [in a new window]   Figure 7.   (A) Developmental changes in mouse PDE5 mRNA levels as assessed by ribonuclease protection assay (RPA). PDE5 mRNA was determined in 140-gestational-day fetuses (n = 3 individual RPAs), at 1 h after delivery (n = 3), 4 h after delivery (n = 3), at 4 d (n = 3), at 7 d (n = 3), and in adults (n = 3), as described in METHODS. Only the fetal RPAs were the pooled result of three animals; all other assays were done on individual animals. PDE5 mRNA is expressed as the ratio of PDE5 mRNA pixels to G3PDH pixels. As with sheep PDE5 mRNA, developmental changes in mouse PDE5 mRNA parallel the changes in PDE5 protein and activity. (B) Representative RPA of mouse PDE5 and G3PDH message during development. The RPA was analyzed with DATASCAN II and NIH Image Software.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The studies described here demonstrate a biphasic regulation of pulmonary PDE5 during development. Concurrently with the well-described decreases in PVR in the first minutes to hours of transition, PDE5 activity, protein, and message decrease significantly in early transition (1, 3). In addition, the data presented here also show a secondary increase in PDE5 hydrolytic activity, protein expression, and mRNA occurring late in postnatal development, with a decrease to adult levels by 21 d. The data also show that the regulation of pulmonary PDE5 activity is far more complicated than would be indicated by the simple, monotonic decrease from a high level of activity in fetal lung to a low level in the adult lung that we had anticipated. These data suggest a complex regulatory control of pulmonary PDE5 activity and expression during maturation.

PDE5 activity in the supernatant decreases 2-fold within 1 h after delivery in normal fetuses, as seen in Figures 2 and 3. Using a specific antibody to PDE5, we also demonstrated that the protein level in the supernatant appeared to correlate with activity changes (Figures 4 and 5). PDE5 activity, protein, and message are significantly increased in the fetal state in which pulmonary tone is elevated. The parallel decrease in ovine PDE5 activity, protein, and message after 1 h correlates with known decreases in PVR that occur on delivery. More interesting is the mouse data, which suggest that the decrease in PDE5 activity begins to occur by 10 min after the onset of breathing. This decrease in PDE5 activity would further augment the cGMP-mediated vasodilation by NO-stimulated guanylate cyclase that occurs during this critical period.

In preliminary studies, we showed that PDE5 cGMP hydrolytic activity was increased in the fetus and dramatically decreased in the adult. We speculated that PDE5 would remain low after transition, in accord with the continuing low PVR after completion of the perinatal transitional circulation. However, in late perinatal development, from 4 to 7 d after the transition, a rise occurs in PDE5 activity, protein, and message in both sheep and mice. Since it is known that PVR remains low during late perinatal development, our data do not fit with the original hypothesis that PDE5 remained low postnatally. Two possibilities may exist to explain these late gestational increases in PDE5 activity and protein. Since it is known that cGMP levels are maintained by a homeostasis between cGMP degradation and synthesis, it is possible that synthesis of cGMP via guanylate cyclase increases in late perinatal development. A recent study by Bloch and associates demonstrated a 7-fold increase in soluble guanylate cyclase activity and message in 8-d-old rat lung as compared with adult rat lung (34). An additional study demonstrated a significant increase in atrial natriuretic factor (ANF)-induced guanylate cyclase activity by 4 d, which remained elevated for 14 d, with a return to normal in adult rat lung (35). In order for the system to maintain pulmonary vascular homeostasis, an increase in both PDE activity and guanylate cyclase activity may occur, implying a complex regulation of PVR and VSM-cell proliferation. A second possibility for the secondary increase in PDE5 activity, protein, and message is that it is occurring in a different cell type in the lung. Previous work in our laboratory has shown that PDE5 protein, as determined by immunocytochemistry, and PDE5 mRNA, quantitated by in situ hybridization, are localized to VSM and cilia in adult mouse lung (36). Thus, the late perinatal increase in PDE5 may represent increases in airway epithelial, or more specifically cilial, PDE5 activity, protein, and message.

Our data also imply a complex level of regulation of perinatal pulmonary PDE5 activity beyond alterations in protein and message. During early development, PDE5 message, as determined by RPA, correlates with changes in PDE5 activity and protein. This finding would suggest that transcriptional control of PDE5 occurs during early transition. During late transition, PDE5 activity is higher than the protein level, suggesting that posttranslational changes occur in the regulation of PDE5 during late transition. Previous work by Burns and colleagues (38) has shown that phosphorylation of PDE5 by the catalytic subunit of protein kinase A (PKA) may be a posttranslational control mechanism for PDE5 activity. Phosphorylation increases the activity of PDE5. Recently, Pyne and associates demonstrated that an additional factor, referred to as gamma factor (similar to the gamma factor in PDE6 regulation), may be involved in posttranslational control of PDE5 (39). Our data would suggest that either phosphorylation of PDE5 and/or gamma protein-PDE5 interaction may be mechanisms of posttranslational control, since PDE5 activity and the PDE5 protein level do not appear to correlate as well during late development.

Some important caveats about this research need to be considered. First, we measured PDE5 activity, protein, and message in whole-lung homogenates. Although we speculate that the early transitional changes in PDE5 may contribute to changes in PVR, we do not know for certain that these changes actually occurred in VSM of fetal tissue. However, we do know that PDE5 is present in the VSM of adult lung tissue, and if these changes do occur in fetal VSM, the present data will underestimate the changes. Studies are currently underway to localize PDE5 protein and message during development, using immunocytochemistry and in situ hybridization, respectively.

Second although previous studies have shown that developmental alterations or remodeling of pulmonary airway and pulmonary VSM mass occurs during the transitional circulation, this does not occur within the first few hours (39). Thus, although the decrease in PDE5 activity and protein in the supernatant cannot be attributed to pulmonary VSM loss by remodeling in the early transitional stage. This could explain the late transitional changes in PDE5 activity and protein.

Third, as is the common practice, we defined and calculated PDE5 activity on the basis of zaprinast suppression. In our studies, PDE5 activity was determined at 1 µM cGMP, in the presence of 10 µM EGTA and in the presence and absence of 4 µM zaprinast. There are four PDEs that preferentially hydrolyze cGMP: types 1, 2, 5, and 6 (21). Torphy and associates have reported six distinct PDE isozymes in human lung, but found that PDE1a, PDE1b, and PDE5 accounted for the vast majority of the cGMP-hydrolytic activity (23). PDE1c has also recently been isolated from aortic VSM (40, 41). Past data suggest that type 1 PDEs, in the presence of EGTA and absence of calcium-calmodulin, are inhibited (42). Some earlier studies with increased calcium-calmodulin and EGTA demonstrated no changes in PDE5 activity. Although we cannot state with certainty that all of the PDE1 activity is suppressed, the available data would suggest that in the presence of EGTA, most of the cGMP hydrolytic activity of PDE1 is suppressed. Type 2 PDEs have a distinctly higher K_m, of 10 to 16 µM (43), and exist in the lung tissue in small concentrations (23). Therefore, PDE2 would contribute very little to hydrolysis of cGMP. Type 6 PDE is only known to be a photoreceptor-associated PDE. The K_m of PDE5 has been documented to range from 1-5 µM (29). For zaprinast, a PDE5 inhibitor, the concentration of 4 µm was determined to be the IC₅₀ for both adult and fetal sheep. Previous studies have shown the IC₅₀ for zaprinast for guinea pig and bovine PDE5 to be 0.4 to 1 µM, respectively. The IC₅₀ for zaprinast for purified ovine PDE5 is not known. Because our IC₅₀ was determined in whole-lung homogenates, additional factors such as PDE5 phosphorylation in ovine lung, may contribute to the slighter higher zaprinast IC₅₀. According to the data of Burns and associates, PDE5 phosphorylation increases the IC₅₀ for zaprinast (37). If most of the PDE5 is phosphorylated, this may explain the higher IC₅₀ for zaprinast found in the current study. Using the IC₅₀ standard curves generated by Burns and associates for phosphorylated guinea pig lung PDE5, the inhibition at 4 µM zaprinast would be 40%. Using the same standard curves for dephosphorylated guinea pig PDE5, the inhibition would be 70%. Thus, taking all these variables into consideration, we can safely conclude that in our ovine-lung tissue homogenates, the vast majority of cGMP hydrolytic activity in the presence of EGTA, 1 µM cGMP, and 4 µM zaprinast is due to PDE5.

In summary, we report a biphasic regulation of pulmonary PDE5 during development. During the perinatal transitional circulation, PDE5 activity and protein decrease dramatically, and these changes correlate with changes in PDE5 specific message. During late perinatal development, there is a secondary increase in PDE5 activity, protein expression, and mRNA. The anatomic and functional significance of the late perinatal increase in PDE5 remains to be elucidated. Our data imply that PDE5 may be an important mediator in the regulation of PVR in normal and possibly in pathologic states, and that inhibitors of PDE5 enzymatic activity may be useful in the treatment of perinatal pulmonary hypertension.