Main

The fetal pulmonary circulation is characterized by high PVR and low basal blood flow, with less than 10% of the combined ventricular output entering the pulmonary vascular bed (1). In addition to increased basal tone, the premature fetal lung lacks the capacity for pulmonary vasodilation (2, 3). However, with advancing gestation, the fetal lung progressively develops the ability to respond to endothelium-dependent vasodilators (2, 3). ACh and oxygen, the most extensively studied pulmonary vasodilators in the fetus, markedly and acutely increase pulmonary blood flow in late gestation. However, the response to these agents is transient, and in spite of ongoing exposure to the vasodilator stimulus, flow returns to baseline within 2 h (4–6).

Another group of potential vasoactive mediators whose effects have not been investigated in the fetal pulmonary circulation are hormones, and specifically, E2. The vascular effects of E2 in the perinatal pulmonary circulation are of interest, because E2 levels are high in the fetus and surge just before birth (7). E2 causes vasodilation of numerous adult vascular beds, but the timing of its effects is variable and depends upon the specific experimental model studied (8). At least three different patterns of response to E2 have been reported. E2 causes acute relaxation of explanted coronary artery rings from human males within 5 min of exposure (9). In contrast, the onset of its vasodilator effects in uterine arteries of nonpregnant ewes is delayed for 40 min and peaks at 100 min, before flow gradually returns to baseline over several hours (10, 11). Other studies suggest that its vasodilator effects may be delayed for several days, as in rabbit coronary arteries (12).

Although no studies have investigated the effects of E2 in the intact fetal lung, recent studies report that E2 increases the expression of eNOS, the enzyme responsible for the generation of NO, in isolated PAECs (13). In addition, E2 acutely stimulates eNOS activity in intact isolated PAECs (14). These findings may be of physiologic importance, because the eNOS-NO cascade modulates pulmonary vascular tone in the fetus and newborn (15, 16), and they suggest that E2 has the capacity to rapidly increase blood flow in the fetal and neonatal lung. However, this hypothesis has never been tested in the whole animal.

Therefore, to determine whether E2 causes pulmonary vasodilation in the intact fetal lung, we tested the hemodynamic response to intrapulmonary E2 infusions in chronically catheterized, late-gestation fetal lambs. We found that E2 markedly increases fetal pulmonary blood flow, but that in contrast to the acute response suggested by in vitro studies, the response to E2 occurs only after prolonged exposure. In addition, the vasodilator response is sustained even after withdrawal of the hormone. These results support the possibility of a contribution of endogenous estrogens to the normal pulmonary vasodilation at birth.

METHODS

Pregnant, mixed-breed (Columbia-Rambouillet) ewes were used in this study. All procedures and protocols were reviewed and approved by the Animal Care and Use Committee of the University of Colorado Health Sciences Center.

Surgical Preparation

At 124–127 d gestation (term = 147 d), ewes were sedated with i.v. sodium pentobarbital and intrathecal tetracaine hydrochloride (1%, 3 mL) after fasting for 2 days. Ewes were given intramuscular penicillin G (600,000 units) and gentamicin (80 mg) prophylactically at the time of surgery. Under sterile conditions, a midline abdominal incision was made and the uterus externalized. A hysterotomy was made, and the left fetal forelimb was taken out of the uterus. Twenty-gauge polyvinyl catheters were placed in the left axillary artery and vein and were advanced to the level of the ascending aorta and superior vena cava, respectively. A left thoracotomy and pericardial incision were made to expose the heart and great vessels. The LPA was exposed by blunt dissection. Using a 16-g intravenous placement unit (Angiocath; Travenol, Deerfield, IL), we then placed a 22-g catheter through purse-string sutures in the main pulmonary artery between the ductus arteriosus and the bifurcation of the pulmonary artery and advanced it into the LPA, as previously described (17). Twenty-gauge catheters were similarly placed in the main pulmonary artery between the ductus arteriosus and the pulmonic valve and in the left atrial appendage. An ultrasonic flow transducer (6 mm; Transonics Systems Inc., Ithaca, NY) was then placed around the left pulmonary artery. A catheter was placed in the amniotic cavity to serve as a pressure referent. The uterus was sutured closed and replaced within the maternal abdominal cavity. The catheters and flow transducer cable were externalized to a flank pouch on the ewe after the abdominal wall had been closed. Postoperatively, ewes were allowed to eat and drink ad libitum and were generally standing within 12 h. Fetuses were treated with ampicillin (500 mg) in the venous catheter for the first two postoperative days. Lambs were allowed to recover for at least 48 h before studies were performed.

Hemodynamic studies were performed by connecting the externalized catheters to pressure transducers that are connected to a computer-driven system (MP100A, Biopac Systems, Inc., Santa Barbara, CA) that converts the signals to a wave-form display. Simultaneous measurements were made of main pulmonary artery, aorta, left atrial and amniotic pressures, and LPA flow. Arterial blood was drawn from the aorta catheter to measure pH, PaCO2, PaO2 (ABL 500, Radiometer, Copenhagen), and saturation and Hb (OSM3 hemoximeter, Radiometer, Copenhagen).

Drug Preparation

A stock solution of 17β-estradiol (E2, E-8875, Sigma Chemical Co. Pharmaceuticals) was made (1 mg/mL; 100% ethanol) and stored at −4°C. The stock solution was diluted with normal saline to its final concentration on the day of each study. Final concentration of ethanol was 0.1%, 1%, and 10% for each of the three doses (1, 10, and 100 μg, respectively) of E2 tested in protocol #1, and 0.025% for the dose (250 ng/mL) tested in protocol #2. A stock solution of ACh (A-6625, Sigma Chemical Co. Pharmaceuticals) was made in distilled water (15 μg/mL). L-NA (N-1522, Sigma Chemical Co. Pharmaceuticals) and L-Arg (A-5006, Sigma Chemical Co. Pharmaceuticals) were dissolved in distilled water and prepared before each individual study.

Protocol #1: hemodynamic effects of brief (1-min) intrapulmonary E2infusion in late-gestation fetal lambs.

The purpose of protocol #1 was to determine whether brief intrapulmonary administration of E2 causes vasodilation of the fetal lung and whether its effects are immediate or delayed. In addition, the capacity of E2 to enhance endothelium-dependent vasodilation was tested by comparing the response to ACh before and after administration of E2. After an initial period of at least 30 min of stable baseline hemodynamics, fetal lambs (n = 5) were given an infusion of ACh (1.5 μg/min × 10 min) into the LPA catheter. Pressure and flow measurements were recorded every 10 min until values returned to baseline. 17β-estradiol was then infused over 1 min, and hemodynamic measurements were taken at 10-min intervals for the next 2 h. After 2 h, the response to ACh was repeated. In previous studies, we have observed no tachyphylaxis to repeated, brief infusions of ACh separated by 2 h. Twenty-four hours after the initial dose of E2, the response to ACh was again assessed after stable baseline hemodynamics were established. In a second set of animals (n = 3), responses to a lower (1-μg) and higher (100-μg) dose of E2 were assessed. ACh responsiveness was not studied after these doses. Arterial blood gas measurements were made at baseline and at 2 and 24 h after each dose of estradiol.

Protocol #2: hemodynamic effects of prolonged intrapulmonary E2infusion in late-gestation fetal lambs.

The purpose of protocol #2 was to determine whether prolonged administration of E2 by a continuous infusion causes vasodilation of the fetal lung, and to characterize the response. After the animals had recovered from surgery, baseline hemodynamic and arterial blood gas measurements were taken. Following baseline measurements, a continuous infusion of vehicle (0.025% ethanol, 1 mL/h, n = 4) or E2 (250 ng/h at 1 mL/h, n = 9) was given into the LPA. Daily arterial blood gas and hemodynamic measurements were taken on each subsequent day until the end of the study period. The end of the study period was defined as a sustained doubling of LPA flow over baseline, an absolute flow greater than 150 mL/min, or 8 total days of infusion without response. The E2 dose was increased to 1.25 μg/h after 6 d of infusion if there was no response. A “sustained” increase in flow was defined as lasting continuously for >2 h. Once a sustained increase in flow was documented, the E2 infusion was discontinued, and the lamb was continuously monitored for another hour before sacrifice. Ewes were killed at the end of the study period with an overdose of euthanasia solution (Fort Dodge Laboratories, Fort Dodge, IA). At autopsy, correct catheter position was confirmed and fetal sex was noted. Peripheral lung tissue was rapidly frozen in liquid nitrogen for later molecular studies.

Protocol #3: hemodynamic effects of NOS inhibition in a subset ofE2-responsive fetal lambs.

The purpose of this protocol was to determine whether release of endogenous NO maintains high pulmonary blood flow in those animals responsive to E2 infusion. Three of the E2-responsive animals were studied with L-NA, a nonspecific inhibitor of nitric oxide synthase, which was used to block release of endogenous NO. After the sustained rise in pulmonary blood flow caused by E2 treatment, L-NA (1 mg/min × 20 min) was infused into the LPA during continuous hemodynamic monitoring. After pressure and LPA flow measurements had stabilized in response to L-NA, L-Arg (10 mg/min; total dose 100–400 mg) was infused into the LPA to determine the specificity of the effect of L-NA on NOS activity.

Protocol #4: eNOS protein content and endothelin-1 peptide levelsin whole lung tissue from E2-responsive and control fetal lambs.

The purpose of this protocol was to determine whether E2 treatment increased eNOS protein content or endothelin-1 peptide levels in the lungs of E2-responsive lambs. Adjacent pieces of frozen left peripheral lung tissue from E2-responsive (n = 4) and control (n = 5) fetal lambs were used. Because tissue was available from only two vehicle-treated lambs, three age-matched, untreated fetal lambs were included in the control group.

Western blots were performed with 25 μg of lung protein, in accord with a previously published method (32), with a MAb to eNOS (Transduction Laboratories, Inc., Lexington, KY). Densitometry was performed with a scanner and National Institutes of Health “Image” software.

ET-1 peptide levels were determined from 100 mg of tissue, as previously reported (18). Briefly, tissue was homogenized in 1 M acetic acid (0.1% Triton X-100; Sigma Chemical Co., St. Louis, MO) and immediately boiled for 7 min. Forty microliters of homogenate were applied to pre-equilibrated C2 columns (Amersham, Arlington Heights, IL), washed with 0.1% trifluoroacetic acid, and eluted with 80% methanol/0.1% trifluoroacetic acid. The volume was reduced to near dryness on a Sped-Vac Concentrator (Savant, Farmingdale, NY) and then was reconstituted to 250 μL. One hundred microliters were then applied to replicate wells of an ET-1 peptide ELISA kit (Cayman Chemical, Ann Arbor, MI) according to the manufacturer's recommendation. Samples were incubated overnight with an acetylcholinesterase-linked ET antibody, and activity was quantified by determination of OD units at 410 nm with a DynaTech MR 700 plate reader (Bio-Tek, Winooski, VT).

To analyze statistical data, the StatMost Statistical Analysis and Graphics Program (DataMost Corporation, Salt Lake City, UT) was used. Data from protocols #1 and #2 were compared by repeated measures ANOVA with Student-Neuman-Keuls posthoc testing. Data from protocol #3 were compared by one-way ANOVA with Student-Neuman-Keuls posthoc testing. Data from protocol #4 were compared by t test. Results obtained are presented as mean ± SEM. Significance was set at p < 0.05.

RESULTS

Protocol #1: hemodynamic effects of brief intrapulmonary E2infusion in late-gestation fetal lambs.

Baseline PAP, aortic pressure, LPA flow, and left lung PVR were similar for each group of animals studied with each dose of E2 (Table 1). None of the hemodynamic parameters were changed during the continuous 2-h study period or at 24 h after E2 treatment with each of the three doses studied (1, 10, and 100 μg). Figure 1 shows that a 10-μg infusion of E2 resulted in a lack of acute effect on LPA flow. E2 (10 μg) did not change the fall in left lung PVR in response to ACh after 2 or 24 h (48 ± 8% and 49 ± 6% at 2 and 24 h versus 50 ± 6% before E2) (Fig. 2 and Table 2).

Table 1 Acute effects of E2 (1, 10, and 100 μg) on hemodynamic and arterial blood gas parameters (BL = baseline)
Full size table
Figure 1
figure 1

Effects of brief intrapulmonary bolus administration of 17β-estradiol (E2) (10 μg over 1 min) on LPA flow in late-gestation fetal lambs. No difference exists at any time point.

Full size image
Figure 2
figure 2

Fractional change in PVR in response to an endothlium-dependent vasodilator, acetylcholine, in late-gestation fetal lambs. There was no difference between the responses tested at baseline (BL) and those tested 2 and 24 h after intrapulmonary 17β-estradiol (10 μg).

Full size image
Table 2 Hemodynamic effects of endothelial-dependent vasodilation with acetylcholine (ACh; 3.0 μg/min × 10 min) in E2-treated lambs
Full size table

Protocol #2: hemodynamic effects of prolonged intrapulmonary E2infusion in late-gestation fetal lambs.

Baseline hemodynamic and blood gas parameters were similar in control and E2-treated animals (Table 3). E2 treatment increased left PBF compared with both baseline (73 ± 6 versus 188 ± 44 mL/min, p < 0.05) and vehicle-treated controls (188 ± 44 versus 55 ± 12 mL/min, p < 0.05) (Fig. 3). Vehicle infusion decreased flow in the control group (79 ± 6 versus 55 ± 12 mL/min, p < 0.05). Flow remained elevated for at least 2 h in each E2-responder before E2 was discontinued. After discontinuation of the E2 infusion in each of the E2-responders, pulmonary blood flow remained elevated until sacrifice (at least 1 h) in every case. Left lung PVR fell, compared with both baseline (0.62 ± 0.05 versus 0.32 ± 0.07 mm Hg/mL/min, p < 0.05) and control (0.32 ± 0.07 versus 0.93 ± 0.20 mm Hg/mL/min, p < 0.05). pH and saturation fell and PaCO2 rose in E2-treated lambs during the study period. Six of the nine E2-treated animals responded to treatment, and the timing of the response varied from 2 to 8 d. Only one animal responded after the dose of E2 was increased on the 6th day.

Table 3 Effects of continuous intrapulmonary infusion with vehicle (0.025% ethanol) or estradiol (250 ng/h) in near-term fetal lambs * Indicates p < 0.05 vs. baseline. † Indicates p < 0.05 vs. control.
Full size table
Figure 3
figure 3

LPA blood flow measured at baseline and after intrapulmonary treatment with 17β-estradiol (250 ng/h, continuous infusion) in fetal lambs. Treatment time point defined as sustained increase in flow or 8 d of infusion without response. *p < 0.05.

Full size image

The change in left lung PVR over the first 48 h of treatment was different when E2-treated animals were grouped by sex (15 ± 10%versus −44 ± 16%, male versus female, p < 0.05) (Fig. 4). Baseline hemodynamic and blood gas parameters (Table 4) and gestational age (131 ± 2 versus 130 ± 1 d, male versus female) were the same in the two groups. Compared with baseline values, PAP fell at 48 h in females but not in males.

Figure 4
figure 4

Fractional change in PVR after 48 h of intrapulmonary treatment (250 ng/h, continuous infusion) with 17β-estradiol in male and female fetal lambs. *p < 0.05.

Full size image
Table 4 Hemodynamic and blood gas data at baseline and after 48 h of E2 infusion from male and female fetal lambs * Indicates p < 0.05 between animals at same time point. † Indicates p < 0.05 between animals of same sex.
Full size table

Protocol #3: hemodynamic effects of NOS inhibition in E2-responsive fetal lambs.

L-NA administration increased PAP (53 ± 4 versus 42 ± 2, L-NA versus E2 response, p < 0.05), and PVR (0.51 ± 0.15 versus 0.15 ± 0.05, L-NA versus E2 response, p < 0.05) (Fig. 5). Flow stabilized in each of the three animals within 15 min in response to L-NA. L-Arg reversed the rise in PVR caused by the administration of L-NA (p < 0.05) (Table 5).

Figure 5
figure 5

Left PVR in estradiol-responsive fetal lambs (n = 3) after estradiol infusion (E2) and in response to endothelial nitric oxide synthase antagonist L-NA (20 mg) and L-Arg (100–400 mg). *p < 0.05 vs both baseline (BL) and L-NA.

Full size image
Table 5 Hemodynamic parameters in E2-responsive animals (n = 3) treated with eNOS inhibitor L-NA * Indicates p < 0.05 vs. E2-response. † Indicates p < 0.05 vs. both baseline and L-NA.
Full size table

Protocol #4: eNOS protein content and ET-1 peptide levels in whole lung tissue from E2-responsive and control fetal lambs.

There was no difference between eNOS protein content in E2-responsive and control fetal lambs (Fig. 6). Neither was there a difference between ET-1 peptide levels in E2-responsive and control fetal lambs (Fig. 7).

Figure 6
figure 6

eNOS protein content in whole lung homogenates (25 μg) from control (n = 5) and E2-responsive (n = 4) fetal lambs.

Full size image
Figure 7
figure 7

ET-1 peptide levels in whole lung homogenates from control (n = 4) and E2-responsive (n = 4) fetal lambs.

Full size image

DISCUSSION

We report that prolonged infusions of E2 markedly increase pulmonary blood flow in the late-gestation fetus. The vasodilator response to E2 is sustained for at least several hours, and its effects persist even after discontinuation of the infusion, a pattern that is unique among previously reported fetal pulmonary vasodilators. In contrast, brief infusions of E2 do not acutely increase pulmonary blood flow or augment endothelium-dependent vasodilation in the ovine fetal lung.

Recent studies report that physiologic concentrations of E2 up-regulate both eNOS and cyclooxygenase-1 gene expression in isolated fetal PAECs (13, 19). Because the respective products of those two enzyme systems, NO and prostacyclin (PGI2), both modulate pulmonary blood flow in the normal fetus and contribute to the normal fall in PVR at birth (15, 16, 20, 21), those studies have suggested that E2 might have vasodilator properties in the fetal lung. However, our findings are the first to demonstrate that estrogens have vasoactive properties in the intact fetal pulmonary circulation. In addition, the delayed nature of the response to E2 is not predicted by those in vitro studies, in which E2 causes acute eNOS activation in isolated fetal PAECs (14). In the intact fetal lamb, the response to E2 occurred only after prolonged exposure. Furthermore, the vasodilation was sustained for the duration of the experiment. Although lambs were killed for tissue collection, they were first observed for at least 3 h with no decrement in PBF. Moreover, in the three animals treated with L-NA, PBF rapidly returned to E2-stimulated levels after L-Arg infusion without being re-exposed to E2. This response is in striking contrast to the fetal pulmonary vasodilation caused by previously reported agents. ACh and oxygen, perhaps the best characterized vasodilators of the fetal lung, rapidly increase pulmonary blood flow within minutes of exposure. However, the vasodilation is transient, and flow returns to baseline within 1–2 h in spite of ongoing exposure to the stimulus (4, 5), suggesting that the normal fetal lung actively opposes sustained vasodilation during late gestation. The persistence of the vasodilation caused by E2, even after discontinuation of the infusion, suggests that E2 may induce fundamental changes in the fetal pulmonary circulation that favor lower basal tone.

The rapid fall in PVR and resultant increase in pulmonary blood flow at birth is of crucial importance to a successful transition from intrauterine to extrauterine life. Nonetheless, factors preparing the lung circulation to respond to birth-related stimuli with a vasodilation that is sustained indefinitely remain unclear. Our findings lend further support to the possibility that high endogenous levels of E2 promote the rise in pulmonary blood flow at birth and the maintenance of lower basal PVR postnatally. Estrogen levels rise through gestation in fetal sheep and surge within 48 h of birth (7). Limited data suggest a similar rise in E2 during the latter half of gestation in human fetuses (22). Our study suggests that the perinatal surge in endogenous estrogens, which peak at precisely the time the pulmonary circulation undergoes its rapid vasodilation at birth, may contribute to that vasodilation. In addition, the prolonged exposure to high estrogen levels in late gestation may serve to “prime” the lung circulation and enhance its vasodilator response to several birth-related stimuli (e.g., rhythmic lung distension, exposure to increased oxygen, and the development of a gas-liquid interface) Although estrogen levels fall relatively rapidly after birth, neonatal estrogen levels remain elevated at several-fold higher than childhood levels for the first 24–48 h postnatally (23). We hypothesize that ongoing exposure to these high levels of endogenous estrogen immediately after birth helps facilitate the pulmonary circulation in its transition to a low-resistance vascular bed postnatally. Nonetheless, direct extrapolation of our data on exogenous estrogens to the response to high endogenous estrogen levels is limited by our use of large local infusions of E2, which may have produced a pharmacologic rather than a physiologic effect.

The mechanism by which E2 causes vasodilation in this model remains unclear. Previous studies of primarily adult circulations have extensively investigated the effects of E2 on the NO-cGMP cascade. Elevated estrogen levels could be responsible for the increased endothelium-dependent or NO-mediated vasodilation noted during pregnancy (24). Exogenous estrogen administration increases endothelium-dependent vasodilation in some beds (25, 26), and these effects are often blocked by competitive inhibitors of NOS. Interestingly, E2 may enhance endothelium-dependent vasodilation in some, but not all, vascular beds in the same animal (27), although the basis for this differential effect is not well understood. Several studies have demonstrated enhanced eNOS activity by using the arginine-to-citrulline conversion assay in estrogen-treated tissues (28–30). However, definitive molecular evidence that estrogens actually up-regulate eNOS expression is limited. Weiner et al. reported that E2 treatment of nonpregnant adult guinea pigs increases eNOS mRNA and activity (31). Hishikawa et al. found increased eNOS protein in human aortic endothelial cells after in vitro estrogen treatment (32), and Goetz et al. demonstrated that increased eNOS mRNA was associated with both pregnancy and E2 replacement in adult rats (33).

To determine the effects of chronic E2 treatment on the NOS-cGMP axis in the present study, we treated a subset of E2-responsive lambs with a selective NOS antagonist, L-NA. L-NA increased PVR by 340%, suggesting that NO makes an important contribution in sustaining the vasodilator response to E2. We did not perform L-NA studies in control animals because of the marked difference in pulmonary blood flow after treatment (5.8-fold higher in E2-responsive animals). Although previous studies from our lab demonstrate that L-NA generally causes only a 35–40% increase in basal PVR in chronically catheterized, late-gestation fetal lambs (15), we believe that meaningful comparison of the L-NA response of untreated animals to E2-responsive animals is impossible, given the marked difference in pulmonary blood flow. In contrast to the effects of E2 in isolated pulmonary artery endothelial cells, we further found that eNOS protein expression was not increased in those animals that responded to the E2 infusion. The pronounced response to L-NA with unchanged eNOS expression suggests that E2 increases NO activity by a posttranslational mechanism in responders. Previous studies have shown that E2 treatment might increase physiologic NO bioavailability by reducing superoxide anion generation and resultant NO inactivation (34). In addition, preliminary studies suggest that E2 may change the intracellular distribution of eNOS to favor increased bioactive NO release (35). E2-induced changes in expression of phosphodiesterases or of other NOS isoforms are additional potential mechanisms that have not been studied. Alternatively, E2 may increase NO release by a secondary mechanism, such as increased shear stress produced by changes in other primary mediators. Direct determination of whether enhanced NO release mediates the pulmonary vasodilator response to E2 in the fetal lamb will require concurrent treatment with L-NA and E2.

Previous studies suggest that E2 can decrease expression of ET-1 (36), which we have previously shown to contribute to the maintenance of high vascular tone in the fetal lung (37). We determined ET-1 levels in E2-responsive animals and found no differences between groups, suggesting that this is not the mechanism by which E2 increases flow in this model. A third enzyme system important in modulating fetal pulmonary vascular tone, the COX-PGI2-cAMP cascade, can also be altered by E2 exposure (38–40). However, we did not investigate that possibility in the current report. Extensive further studies are necessary to delineate the precise mechanism by which E2 causes pulmonary vasodilation in the fetal lamb.

The earlier response of female lambs to E2, compared with their male counterparts, is consistent with the findings of other investigators working with estrogens. Weiner et al. found that 5 d of treatment produced pronounced effects in female guinea pigs, with no change seen in males (31). More prolonged treatment, however, produced a similar response in males at 10 d. The implications of an early selective effect of E2 in female fetal animals is unclear. Although females are known by some physiologic parameters to develop earlier than males (41, 42), no data are available with regard to gender differences in pulmonary vascular development. In addition, there are no reports of gender differences in the incidence of pulmonary hypertension in neonates. The different timing in response may result from gender differences in estrogen receptor density, but little is currently known about estrogen receptor expression in the fetal lung.

The choice of E2 dosing was particularly difficult in this study. The range of doses used in our first protocol was based on the absolute dose used in previous studies of intra-arterial infusions (10), which were given into the uteri of nonpregnant ewes with baseline blood flow similar to that of the fetal ovine lung. The dose of 250 ng/h of E2 used in the second protocol was intended to instantaneously raise left lung E2 levels to approximately the published levels noted just before delivery (7). Because we infused E2 directly into the left pulmonary circulation, we were unable to make E2 measurements that would accurately reflect the concentrations achieved locally in left lung. Further, the unknown rates of conjugation and clearance of exogenous estrogen by the fetus and placenta during this local infusion made anticipation of the changes in E2 levels over time very difficult. Whether systemic i.v. infusions would yield similar changes in pulmonary blood flow is not currently known. Understanding the physiologic role of high endogenous E2 levels is further complicated by the concurrently high levels of other estrogens, including estriol and estrone, which rise during gestation (7) and may have additive biologic effects. Further studies with pharmacologic estrogen receptor blockers should clarify the contribution of endogenous estrogens to the pulmonary circulatory response to birth-related stimuli.

We conclude that E2 has no acute hemodynamic effect on the pulmonary circulation of the ovine fetus, but that prolonged exposure can dramatically increase pulmonary blood flow. The vasodilator response to E2 is unique among the previously studied fetal pulmonary vasodilators. Although up-regulation of NOS may be fundamental to the effects of E2 in other settings, eNOS protein was not increased in those animals responding to E2 in our study. Similarly, ET-1 expression was not changed by E2 treatment. The molecular mechanisms through which E2 acts in this model require further investigation. We speculate that the prolonged exposure to increased endogenous estrogens in most animal species near the end of gestation contributes to the critical and sustained fall in PVR which characterizes the normal transitional circulation.