Introduction

Abstract Introduction Materials and Methods Results Discussion References

Prostacyclin (PGI₂) and other vasodilator prostaglandins are key mediators of pulmonary vascular and parenchymal function during late fetal and early postnatal life. PGI₂ administration causes pulmonary vasodilation in the fetus and newborn, and vasoconstriction occurs when its endogenous synthesis is inhibited (1, 2). In addition, PGI₂ plays a major role in successful cardiopulmonary transition at birth, and it attenuates hypoxic pulmonary vasoconstriction in the newborn period (3, 4). There is evidence that PGI₂ also modulates vascular cell growth in the pulmonary circulation (5). Furthermore, endogenous prostaglandins are important stimulants of surfactant synthesis and cell differentiation in the developing lung (6, 7).

Studies with a variety of species indicate that the synthesis of PGI₂ and other vasodilatory prostaglandins in whole lung increases dramatically during late gestation and in the newborn (8, 9). This increase may play an important role in the decline in pulmonary vascular resistance that occurs not only at birth, but also more gradually thereafter, and it may also be involved in maturational changes in pulmonary vascular structure (10). In addition, this increased synthesis of PGI₂ and other vasodilatory proteins may modulate the rise in surfactant synthesis that occurs during late gestation and the early postnatal period (6), thereby being critical to the pulmonary gas exchange that is mandatory for extrauterine existence. We have previously demonstrated that the rate-limiting enzyme in vascular PGI₂ synthesis in fetal and newborn lung is cyclooxygenase (COX) (11), and that the constitutive COX-1 isoform is expressed in the pulmonary vasculature whereas the inducible COX-2 isoform is not. We have also shown that PGI₂ synthesis in ovine intrapulmonary arteries is primarily endothelium-derived, that this synthesis increases markedly during late fetal and early newborn life, and that this is due to an increase in COX activity related to enhanced expression of COX-1 (11). Thus, there is normally a developmental upregulation in COX-1 expression that optimizes the capacity for pulmonary vascular PGI₂ production in the postnatal period.

In an effort to understand better the mechanisms underlying the developmental increase in PGI₂ synthesis in the pulmonary circulation, we designed the present study to compare PGI₂ synthesis in early-passage, cultured pulmonary artery endothelial cells (PAEC) from fetal and newborn lambs. On the basis of the finding that PGI₂ synthesis is greater in newborn than in fetal intrapulmonary arteries (11), we raised the hypothesis that PGI₂ production is enhanced in newborn compared with fetal PAEC. In addition to testing this hypothesis, we designed studies to answer the following questions: (1) Are there also developmental changes in PGI₂ synthesis in early-passage cultured pulmonary vascular smooth-muscle cells (VSM)? (2) Is there comparable stimulation of PGI₂ synthesis in pulmonary endothelial and VSM cells? and (3) What is the basis for developmental changes in PGI₂ production in PAEC and pulmonary VSM cells?

	Materials and Methods

Abstract Introduction Materials and Methods Results Discussion References

Animal Model

Multiple groups of investigators have used the fetal and newborn lamb to assess the role of vasodilator prostaglandins in the regulation of vascular resistance in the developing pulmonary circulation (1). As such, it is an excellent animal model for in vitro study of the mechanisms controlling PGI₂ synthesis in the fetal and newborn pulmonary circulation. In the present investigation, intrapulmonary arteries and PAEC and VSM cells were obtained from mixed-breed fetal lambs at 125 to 135 d gestation (term = 144 d), and from newborn lambs at 4 wk of age. The pregnant ewes used in the investigation were multiparous and had singletons, twins or triplets. The newborn lambs were housed with their mothers. Before killing, the animals were housed in the Animal Resources Center of the University of Texas Southwestern Medical Center and were given standard animal chow and water ad libitum. The procedures followed in the care and euthanasia of the study animals were approved by the Institutional Review Board for Animal Research of the University of Texas Southwestern Medical Center.

Arterial Segment Preparation

The procedures used generally followed those we have previously reported (11). Briefly, the ewe and fetus(es) were euthanized with sodium pentobarbital (120 mg/kg) given intravenously to the ewe, and the fetuses were delivered by cesarean section. The newborn lambs were killed in a similar manner. The lungs were immediately removed and placed in ice-cold phosphate-buffered saline (PBS; 0.01 M PO₄³; 0.15 M NaCl, pH 7.4). Further tissue preparation was performed in a cold room at 4°C. The intrapulmonary arterial tree was rapidly dissected from the lung parenchyma and placed in fresh ice-cold PBS. Remaining fatty and connective tissue were gently removed and the adventitia was grossly dissected from the arteries, taking care not to disrupt the endothelium. Fourth-generation intrapulmonary arteries (0.5 to 1.0 mm external diameter) (12) were isolated and placed in freshly prepared Krebs- Henseleit buffer gassed with 95% O₂/5% CO₂ at 37°C. The Krebs-Henseleit buffer contained 4.8 mM KCl, 2.0 mM CaCl₂, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 11.0 mM dextrose, 118 mM NaCl, and 25 mM NaHCO₃ at pH 7.4. The generation of arteries studied was chosen in an effort to examine PGI₂ production in freshly obtained intact arterial segments as close to the level of the resistance vessels as possible (12). Segments with wet weights of 1 to 4 mg were cut, rinsed in fresh oxygenated Krebs-Henseleit buffer, and equilibrated in the oxygenated buffer at 37°C for 2 h. In selected experiments the endothelium was removed by repeated passage of knotted silk suture through the lumen of the arteries. The presence or absence of intact, functional endothelium was confirmed in randomly chosen segments by (1) light microscopy of 5-µm sections of the arteries; (2) examinations of endothelium-dependent relaxation with acetylcholine; and (3) quantification of cyclic guanosine monophosphate (cGMP) production with acetylcholine stimulation (13).

Endothelial and VSM Cell Culture

PAEC and VSM cells were harvested from third-generation intrapulmonary arteries of fetal and newborn lambs and maintained with methods previously described (14). The cells were cultured from the same animals from which the fourth-generation intrapulmonary artery segments were obtained. PAEC were propagated in RPMI medium containing 10% iron-supplemented calf serum, 10% lamb serum, 1% L-glutamine, 1% antibiotic/antimycotic mixture, 0.15% nystatin, 0.15% gentamycin, and 0.10% tylosin in a humidified incubator with 5% CO₂ in air at 37°C. The identity of the endothelial cells was confirmed by phenotype (cobblestone appearance and contact inhibition), by immunofluorescence studies with antibody to Factor VIII antigen, and by examinations of acetylated low-density lipoprotein (LDL) uptake. VSM cells were grown in M199 medium containing 5% iron-supplemented calf serum, 5% lamb serum, 1% L-glutamine, and the antimicrobial agents listed previously. The identity of the VSM cells was confirmed by phenotype ("hill-and-valley" appearance) and immunofluorescence studies with antibody to -actin. Both PAEC and VSM cells were studied at passages 3 to 6 in 24-well plates or 75 cm² flasks, at near confluence.

Incubations for PGI₂ Synthesis

The incubation procedures used for the intrapulmonary arterial segments were similar to those we have used previously in studies of ovine fetal arteries (15). The segments were placed into sealed polypropylene chambers containing 2.0 ml of oxygenated Krebs-Henseleit buffer (PO₂ = 680 mm Hg) at 37°C for 1 h. Following this preincubation period, the medium was replaced with fresh Krebs- Henseleit buffer and 20 min incubations were performed. At the end of the incubation, the medium was placed into ice-cold tubes containing 100 µg of acetylsalicylic acid (ASA) and stored at 20°C until the time of assay for PGI₂. The segments were placed in ice-cold 7% trichloroacetic acid (TCA) and were stored at 20°C until protein content was determined with a modification of the method of Lowry and associates (16), using bovine serum albumin (BSA) as the standard. We have demonstrated in ovine fetal arteries that the PGI₂ measured with this technique is newly synthesized, and that synthesis is linear with time for at least 120 min (15). Duplicate segments were studied from four lambs in each age group.

The pulmonary endothelial or VSM cells grown in 24-well plates were preincubated for 15 min in room air with 1 ml of serum-free RPMI added per well. The preincubation medium was replaced with fresh, serum-free RPMI, and 15-min incubations were performed. At the end of the incubation, the medium was placed into ice-cold tubes containing 100 µg of ASA and stored at 20°C until time of assay for PGI₂. The plates were air dried and stored at 20°C until cell protein content was determined with the methods described previously (16).

In experiments designed to determine the reaction in the PGI₂ synthetic cascade that may be developmentally regulated, selected wells of PAEC or VSM cells were incubated in RPMI medium alone, reflecting basal (nonstimulated) synthesis, and others were treated with agents that activate the synthetic pathway at various steps. Incubations were performed in the presence of bradykinin to assess developmental changes in PGI₂ synthesis stimulated by receptor-mediated mobilization of arachidonic acid from phospholipids (17). Incubations with the calcium ionophore A23187 were performed to determine maturational changes in PGI₂ production stimulated by an increase in cytosolic free calcium, which activates arachidonic-acid mobilization by a non-receptor-mediated process (18). Exogenous arachidonic acid was used to stimulate PGI₂ synthesis to reveal whether developmental changes in PGI₂ synthesis are related to changes in COX activity (18). Preliminary studies were performed of concentration-related responses and time-courses of activation of the synthetic cascade. Maximal stimulation of PGI₂ synthesis in fetal and newborn PAEC and VSM was found with bradykinin, A23187, and arachidonic acid at 10⁵ M. As a result, this concentration was used for all agents. Basal and stimulated PGI₂ production in PAEC and VSM cells was linear with time for at least 15 min; ensuing incubations were 15 min in duration. In all experiments, n = 4 to 6 for each determination, and findings were replicated in two or three studies, using cells from different primary cultures.

PGI₂ Assays

Samples of arterial-segment or cell-incubation media were assayed for the stable metabolite of PGI₂, 6-keto-prostaglandin F₁ (6-keto-PGF₁), by radioimmunoassay as previously reported (15, 19). Briefly, the assay procedure used duplicate aliquots of standard (0 to 200 pg) or samples placed into a mixture of Krebs-Henseleit solution and 0.1 M PBS plus 0.1% polyvinylpyrrolidone (1:1). Antiserum (0.1 ml, 1:4,000 titer) and 0.1 ml of [³H]6-keto-PGF₁ (12,000 dpm) were added, and the tubes were incubated at 4°C for 12 to 18 h. Bound and free ligand were separated with dextran-coated charcoal, and bound ligand was counted by liquid scintillation spectrometry. The unknown quantities of 6-keto-PGF₁ were determined from the standard curves generated. The intra- and interassay coefficients of variation (CVs) were 5.8% and 8.9%, respectively, at 250 pg/ml, and 2.8% and 8.7%, respectively, at 1,000 pg/ml.

Immunoblot Analysis for COX Protein

The abundance of COX-1 protein was compared in fetal and newborn PAEC and VSM cells through immunoblot analysis. Cells grown in 75-cm² flasks were harvested in ice-cold PBS, pelleted, resuspended in 50 mM KH₂PO₄ buffer (pH 7.8) containing 250 mM mannitol, 5 mM disodium ethylenediamine tetraacetic acid (EDTA), 0.1 mM diethyldithiocarbamate, 0.1 mM indomethacin, and 1% Tween-20, and were ultrasonically disrupted (Sonifier Cell Disruptor 200; Branson Ultrasonics, Chicago, IL). The protein content of the preparation was determined by the method of Bradford (20), using BSA as the standard.

Sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed with 7% acrylamide by the method of Laemmli (21), and the proteins were electrophoretically transferred to nitrocellulose filters overnight. The nitrocellulose filters were blocked for 1 h in buffer containing 100 mM NaCl, 20 mM NaH₂PO₄, and 80 mM Na₂HPO₄ (pH 7.5) with 1% Tween-20 and 5% dried milk, and were then incubated with 1:1,000 COX-1 antiserum (Cayman Chemical Co., Ann Arbor, MI) for 2 h at room temperature. After incubation with primary antiserum, the nitrocellulose filters were washed in the 100-mM NaCl buffer with 1% Tween-20, and were incubated for 1 h with a 1:10,000 dilution of antirabbit immunoglobulin antibody-horseradish peroxidase conjugate raised in donkey. The filters were washed in 100 mM NaCl buffer with Tween-20, and the band for COX-1 was visualized by chemiluminescence and quantitated densitometrically. Purified COX-1 protein (Cayman Chemical Co.) was used as a positive control. Similar techniques were used to evaluate COX-2 protein expression.

Reverse Transcription-Polymerase Chain Reaction Assay for COX-1

To determine the basis for developmental differences in COX-1 expression in the cell type primarily responsible for vascular PGI₂ synthesis (11), studies of COX-1 mRNA abundance were performed in fetal and newborn PAEC. A semiquantitative reverse-transcription-polymerase chain reaction (RT-PCR) assay was established because the mRNA for COX-1 was not detectable in PAEC by Northern analysis of poly A(+) RNA. Total cellular RNA was obtained from fetal and newborn PAEC grown in 75-cm² flasks, through a single-extraction method with an acid guanidinium thiocyanate-phenol-chloroform mixture (22). RT was performed with methods previously reported, using 5 µg of total RNA (22). Briefly, complementary DNA (cDNA) synthesis was done with 200 U Maloney murine-leukemia-virus (MMLV) reverse transcriptase, 5 µM oligodeoxythymidine (oligo-[dT]), 1 mM deoxyribonucleoside triphosphate (dNTP), and 3 mM Mg²⁺ in a volume of 20 µl. In selected tubes the reverse transcriptase was omitted to control for amplification from contaminating cDNA or genomic DNA. The temperature profile was (1) annealing at room temperature (25°C) for 5 min; (2) extension at 42°C for 45 min; and (3) termination at 99°C for 5 min.

PCR was performed on the resulting RT product with specific oligonucleotide primers for sheep COX-1 (23). The sequence of the sense primer was 5'-ATGAGTACCGCAAGAGGTTTGG-3', and that of the antisense primer was 5'-ACGTGGAAGGAGACATAGG-3'. The PCR reactions contained 1.5 mM Mg²⁺, 1 µM primers, 200 µM dNTPs, reaction buffer, and 5 µl cDNA in a final volume of 50 µl. To minimize nonspecific amplification, a "hot-start" procedure was used in which the PCR tubes were placed in a thermal cycler (Model 480; Perkin-Elmer, Norwalk, CT) prewarmed to 94°C. After 2 min, each tube was opened sequentially, and 2.5 U (in 0.25 µl) of Taq DNA polymerase was added. The PCR temperature profile consisted of 30 cycles of 94°C for 45 s (denaturation), 57°C for 45 s (annealing), and 72°C for 1 min (extension), followed by an additional 5 min final extension at 72°C. The primer location, primer concentration, Mg²⁺ concentration, and annealing temperature were optimized to produce the greatest amount of a single PCR product.

The PCR products were size-fractionated by agarose gel electrophoresis. The identity of the PCR products was confirmed, and they were quantitated by transferring the DNA to nylon filters, probing with a ³²P end-labeled internal oligonucleotide primer specific for sheep COX-1, and performing densitometric analysis on the resulting autoradiographs. PCR product identity was also confirmed by direct double-stranded sequencing. To control for the RT step and RNA stability, RT-PCR was also done for the housekeeping gene malate dehydrogenase (MDH), using published oligonucleotide primer sequences (24). The PCR temperature profile for MDH was identical to that described previously for COX-1. During the establishment of the technique, experiments were performed to determine the relationship between the quantity of total RNA subjected to RT-PCR and the amount of PCR product generated. In addition, COX-1 mRNA stability was compared in fetal and newborn PAEC in experiments with cells treated with 25 µg/ml actinomycin D for varying periods up to 4 h. We have previously used RT-PCR assays done in this semiquantitative manner in studies of pulmonary endothelial nitric oxide synthase expression; Northern blot analyses were performed in parallel with RT-PCR assays, and identical results were obtained with the two techniques (22, 25).

Statistical Analysis

Analysis of variance (ANOVA) with Neuman-Keuls post hoc testing was used to compare mean values between groups (26). Significance was accepted at P < 0.05. All results are expressed as means ± SEM.

	Results

Abstract Introduction Materials and Methods Results Discussion References

Arterial-Segment PGI₂ Synthesis

The result of comparison of PGI₂ synthesis in the fetal and newborn intrapulmonary arteries is shown in Figure 1. PGI₂ production in intact arteries was sevenfold greater in the newborn than in fetal group. Removal of the endothelium resulted in a 79% decrease in PGI₂ synthesis in the fetal arteries and a 67% decrease in the newborn arteries. In the segments denuded of endothelium there was a 12-fold increase in PGI₂ synthesis from fetal to newborn, paralleling the developmental increase demonstrated in the intact segments.

View larger version (15K): [in this window] [in a new window]   Figure 1.   Comparison of PGI2 synthesis in fetal and newborn intrapulmonary arteries. Intact (solid bars) and endothelium-denuded (open bars) arterial segments were incubated for 20 min, and PGI2 (6-keto-PGF1) synthesis was measured by radioimmunoassay. Values are means ± SEM for duplicate segments from four lambs in each group. *P < 0.05 versus fetal and P < 0.05 versus intact segments.

PGI₂ Synthesis in Cultured Vascular Cells

The capacity for stimulated PGI₂ synthesis in newborn PAEC and pulmonary VSM cells is shown in Figure 2. In the PAEC (Figure 2A), maximally stimulating concentrations of bradykinin and A23187 caused a 2.5-fold increase in PGI₂. Maximally stimulating concentrations of arachidonic acid caused a fivefold increase in PGI₂. In contrast, in studies of newborn VSM cells (Figure 2B), bradykinin did not stimulate PGI₂ production, and the increases with A23187 and arachidonic acid were comparable, at 67% and 56%, respectively. Similar observations were made with fetal PAEC and VSM cells (data not shown).

View larger version (15K): [in this window] [in a new window]   Figure 2.   Stimulated PGI2 synthesis in newborn PAEC (A) and pulmonary VSM cells (B). Cells were incubated for 15 min in the absence of exogenous stimulation (basal), or in the presence of bradykinin (BK), A23187, or arachidonic acid (AA) at 105 M. PGI2 (6-keto-PGF1) synthesis was measured by radioimmunoassay. Values are means ± SEM, n = 4 to 6. *P < 0.05 versus basal, and P < 0.05 versus bradykinin or A23187.

The comparison of PGI₂ synthesis in fetal versus newborn PAEC is shown in Figure 3A. Basal PGI₂ synthesis was threefold greater in newborn than in fetal cells. In a parallel manner, PGI₂ with A23187 stimulation was fivefold higher in newborn than in fetal PAEC. Similarly, arachidonic acid-stimulated synthesis was sixfold greater in newborn than in fetal PAEC. The comparison of PGI₂ synthesis in fetal versus newborn pulmonary VSM cells is presented in Figure 3B. Basal PGI₂ synthesis was 2.5-fold greater in newborn than in fetal cells. In a parallel manner, PGI₂ with A23187 stimulation was 2.6-fold higher in newborn than in fetal pulmonary VSM. Similarly, arachidonic acid-stimulated synthesis was 2.8-fold greater in newborn than in fetal pulmonary VSM.

View larger version (20K): [in this window] [in a new window]   Figure 3.   Comparison of basal and stimulated PGI2 synthesis in fetal versus newborn PAEC (A) and VSM cells (B). Cells were incubated for 15 min in the absence of exogenous stimulation (basal), or in the presence of A23187 or arachidonic acid at 105 M. PGI2 (6-keto-PGF1) synthesis was measured by radioimmunoassay. Values are means ± SEM, n = 4 to 6. *P < 0.05 versus fetal, P < 0.05 versus basal, and P < 0.05 versus A23187.

COX-1 Immunoblots

To determine whether the developmental changes in basal and stimulated PGI₂ synthesis in PAEC and pulmonary VSM cells are related to differences in COX expression, immunoblot analysis was performed (Figure 4). In the representative immunoblots shown, COX-1 was detected in both the PAEC and pulmonary VSM cells, and COX-1 protein abundance was greater in the newborn than in the fetal cells. Quantitative densitometry in three independent experiments confirmed these observations, revealing that COX-1 protein was increased threefold in newborn versus fetal PAEC, and was increased 2.8-fold in newborn versus fetal pulmonary VSM cells. COX-2 protein was not detected in either vascular cell type at any age (data not shown).

View larger version (19K): [in this window] [in a new window]   Figure 4.   Immunoblot analysis for COX-1 in fetal (F) and newborn (NB) PAEC (A) and pulmonary VSM cells (B). Purified COX-1 protein was used as a positive control (C). COX-1 protein was detected at 70 kD. The immunoblots shown are representative of three independent experiments. Summary data for quantitative densitometry for the three experiments are given (C). Mean ± SEM values are depicted for protein abundance expressed as the percentage of abundance in fetal cells. *P < 0.05 versus fetal.

COX-1 mRNA Abundance

To determine the basis for developmental differences in COX-1 expression in the cell type primarily responsible for vascular PGI₂ synthesis, studies of COX-1 mRNA abundance were performed in fetal and newborn PAEC, using semiquantitative RT-PCR. In initial studies, linear-regression analysis showed high correlations between densitometry values for RT-PCR products and the quantity of total RNA used for COX-1 RT-PCR (Figure 5, r = 0.95 to 0.99, n = 3 experiments) and for MDH RT-PCR (r = 0.96 to 0.99, n = 3). In both fetal and newborn PAEC, single PCR products were obtained for COX-1 at the expected size of 355 bp (Figure 6A). The representative Southern blot reveals greater COX-1 mRNA abundance as determined by RT-PCR in newborn than in fetal PAEC. PCR was also performed for MDH to control for the RT step, yielding a single PCR product at the expected size of 369 bp. There was no difference in MDH mRNA abundance in fetal versus newborn cells. Quantitative densitometry for three independent experiments confirmed these results (Figure 6B). There was 2.8-fold more COX-1 mRNA in newborn than in fetal PAEC.

View larger version (19K): [in this window] [in a new window]   Figure 5.   Relationship between quantity of RNA subjected to RT-PCR and resulting PCR products for COX-1. Varying amounts of total RNA (2 to 16 µg) from fetal sheep PAEC were subjected to RT-PCR, using primers specific for COX-1. The cDNA generated was separated by electrophoresis and probed with an internal oligonucleotide primer specific for COX-1. (A) A representative Southern blot analysis. Quantitative densitometry was performed, and the abundance of RT-PCR product (in arbitrary densitometry units) is plotted versus quantity of RNA used (B).

View larger version (13K): [in this window] [in a new window]   Figure 6.   (A) Southern blot analysis of RT-PCR products for COX-1 (top) and malate dehydrogenase (MDH, bottom) in fetal (F) and newborn (N) PAEC. Band sizes were: COX-1, 355 bp; MDH, 369 bp. (B) Summary data for three independent experiments. COX-1 densitometry values corrected for MDH are expressed as percentages in fetal cells (mean ± SEM). *P < 0.05 versus fetal.

COX-1 mRNA stability in fetal and newborn PAEC is shown in Figure 7. RT-PCR was performed on cells treated with 25 µg/ml actinomycin D for varying durations up to 4 h. The representative study shown reveals no difference in COX-1 mRNA degradation in the two study groups. In four independent experiments, COX-1 mRNA half-life was similar in fetal and newborn cells (1.5 ± 0.2 and 1.6 ± 0.2 h [mean ± SEM], respectively).

View larger version (14K): [in this window] [in a new window]   Figure 7.   COX-1 mRNA stability in fetal and newborn PAEC. Cells were treated with 25 µg/ml actinomycin D for varying periods up to 4 h, and COX-1 mRNA abundance was assessed by RT-PCR and Southern blot analysis. Linear regression analysis yielded r = 0.99 for fetal cells and r = 0.97 for newborn cells. Similar findings were obtained in four independent experiments.

	Discussion

Abstract Introduction Materials and Methods Results Discussion References

In the present study, PGI₂ synthesis was compared in fetal and newborn intrapulmonary arteries, and in fetal and newborn early-passage, cultured PAEC and pulmonary VSM cells. It was found that PGI₂ production was increased in the newborn compared with the fetal intrapulmonary arteries, and this developmental increase was also evident in both types of cultured vascular cells. These observations indicate that the maturational change in pulmonary vascular cell PGI₂ synthesis is conserved in culture.

The present finding that PGI₂ synthesis is increased in newborn versus fetal intrapulmonary arteries is consistent with our previous observations in the same arterial segments from this animal model (11). In both the present and previous study, PGI₂ synthesis was greater in intact and endothelium-denuded arterial segments from newborn compared with fetal lambs (11). This indicates that both endothelial and medial, or primarily VSM, PGI₂ production increases with development in the intact intrapulmonary artery. The present experiments with cultured cells reveal that this maturational increase is conserved in both cultured endothelium and pulmonary VSM cells. These observations are also consistent with previous reports indicating that the synthesis of PGI₂ and other vasodilatory prostaglandins in whole lung from a variety of animal species increases dramatically during late gestation and in the newborn (8, 9). Importantly, the current findings further reveal that the developmental differences in whole-lung or in vascular PGI₂ production are not simply related to potential changes in the density of PGI₂-producing cells, but rather to an enhanced capacity for PGI₂ synthesis in the two major vascular cell types.

To determine the mechanism(s) underlying the developmental increase in basal PGI₂ synthesis in the cultured PAEC and pulmonary VSM cells, stimulated synthesis was evaluated. Stimulated PGI₂ production was first compared in endothelial and VSM cells. In endothelial cells, both bradykinin, which acts via receptor-mediated processes (17), and A23187, which stimulates synthesis by non-receptor-mediated activation of arachidonic-acid mobilization (18), caused marked increases in PGI₂ production. In addition, exogenous arachidonate caused further stimulation of PGI₂ synthesis. In contrast, in VSM cells, bradykinin did not stimulate PGI₂ synthesis, and activation was modest and equivalent in response to A23187 versus exogenous arachidonic acid. These findings indicate that bradykinin-receptor-mediated stimulation is absent in the pulmonary VSM, and that the capacity for arachidonic-acid mobilization may also differ in the two vascular-cell types. The noted difference in bradykinin responsiveness in the PAEC and pulmonary VSM cells is similar to that observed in human aortic endothelial and VSM cells (27). Interestingly, there is evidence that the capacity for VSM PGI₂ production in response to bradykinin varies among different vascular beds (28), and that under certain conditions bradykinin-receptor expression can be induced in VSM that does not express the receptor constitutively (29). In addition to potential differences in bradykinin-receptor expression in PAEC and pulmonary VSM cells, differences in bradykininase activity may underlie the disparate responses to bradykinin (30).

Because bradykinin-mediated PGI₂ stimulation was absent in the VSM in our study, we evaluated A23187-stimulated responses in our studies of the developmental changes in the function of the PGI₂ synthetic cascade. The maturation-related enhancement in basal PGI₂ synthesis was mimicked in PAEC and VSM cells stimulated with A23187, indicating that the effect of development on PGI₂ synthesis in both cell types does not involve alterations in arachidonic-acid mobilization. This conclusion is supported by the observation that with development, cells incubated with excess arachidonic acid also exhibited differences in PGI₂ production that were comparable to the changes in basal PGI₂ synthesis. Given this, the effect of maturation on PGI₂ production is beyond arachidonic acid in the synthetic cascade.

Because our previous studies of arterial segments indicated that the rate-limiting enzyme in vascular PGI₂ synthesis in the developing lung is COX (11), we next evaluated COX-1 and COX-2 protein expression in cultured cells from fetal and newborn lambs. Constitutive COX-1 protein expression was evident in fetal and newborn PAEC and pulmonary VSM cells. In contrast, COX-2 protein was not detected in either cell type at any age. In parallel with the developmental increases observed in basal, A23187-stimulated, and arachidonic acid-stimulated PGI₂ synthesis, COX-1 protein expression was enhanced in newborn versus fetal PAEC and pulmonary VSM cells. A similar maturational increase in COX-1 expression was demonstrated previously in our studies of intact intrapulmonary arteries (11). However, the present results further reveal the cell types involved, indicating that the COX-1 upregulation occurs in both the endothelium and VSM.

To determine the basis for the developmental upregulation of COX-1 expression in the cell type primarily responsible for vascular PGI₂ synthesis (11), we performed studies of COX-1 mRNA abundance in fetal and newborn PAEC, using semiquantitative RT-PCR. In contrast to our previous studies of COX-1 mRNA, which were limited to whole lung (11), use of the early-passage PAEC allowed us to assess changes in COX-1 mRNA in a single cell type. In accord with the findings for COX-1 protein, steady-state COX-1 mRNA levels were greater in newborn than in fetal PAEC. The use of the cultured cells also enabled us to evaluate COX-1 mRNA stability, which was similar in fetal and newborn PAEC. This suggests that the developmental upregulation in pulmonary endothelial COX-1 expression is most likely mediated at the level of gene transcription.

Although COX-1 is generally considered to be constitutively expressed in many cell types (31), previous studies with cultured endothelial cells and fibroblasts have shown that COX-1 expression is actually mediated by a variety of factors. There is evidence of COX-1 upregulation by phorbol-12-myristate-13-acetate, transforming growth factor- (TGF-), and interleukin 1 (32, 34), and downregulation has been demonstrated in response to acidic fibroblast growth factor-1 (aFGF-1) (35). Because TGF- and aFGF-1 are produced by the developing lung mesenchyme and influence lung morphogenesis (36), they may be involved in the ontogenic regulation of pulmonary COX-1 expression. Furthermore, COX-1 in human fetal lung fibroblasts is upregulated by prostaglandin E₂ (PGE₂) (34), which is produced in the developing lung along with PGI₂ (11). This suggests that the expression of the enzyme in fetal and newborn lung may also be regulated by the enzyme product.

In addition to its potential regulation by growth factors and prostaglandins, there is evidence that COX-1 gene expression may also be under hormonal control. It has been demonstrated in studies of aortic vascular cells that physiologic concentrations of 17-estradiol cause marked increases in PGI₂ and PGE₂ synthesis in both the endothelium and VSM (37, 38). The effect of estradiol is evident after 2 to 3 d of exposure, suggesting that it may involve an increase in COX expression (37, 38). It has also been shown that chronic estrogen treatment in vivo augments endothelium-dependent responses to arachidonic acid, and that this involves enhanced synthesis of COX products (39). In addition, fetal plasma estrogen levels increase dramatically with advancing gestational age, owing to increased placental production of the hormone (40), and estrogen has been implicated in maturational changes in pulmonary endothelial cell morphology in the late fetus (41). In view of this, we postulate that estrogen plays a role in pulmonary COX-1 upregulation during late fetal life. In contrast to estrogen-mediated upregulation of COX-1, studies with VSM cells indicate that glucocorticoids attenuate COX-1 expression (42). Because there is a dramatic decline in plasma cortisol levels postnatally (43, 44), we further postulate that the ebbing of this inhibitory effect plays a role in the continued upregulation in pulmonary COX-1 expression in the newborn.

On the basis of the findings in the present study, the use of fetal and newborn PAEC and pulmonary VSM cells will facilitate further study of the regulation of COX-1 expression. We have previously demonstrated that the acute effects of varying oxygenation on PGI₂ synthesis are conserved in early-passage, newborn ovine PAEC (45), and the present findings indicate that maturational changes in the mechanisms regulating PGI₂ synthesis are also reliably consistent in cultured pulmonary vascular cells as compared with intact arterial segments. It will now be possible to evaluate the direct effects of growth factors, prostaglandins, estradiol, and glucocorticoids on cultured fetal PAEC and pulmonary VSM cells, avoiding the indirect effects of these agents related to changes in blood flow, pressure, and shear stress in intact animal models. Such experiments will provide fundamental new information about the maturational regulation of COX-1 expression, thereby enhancing basic understanding of the mechanisms underlying both normal and pathologic changes in vascular structure and function in the developing lung.