Journal of Applied Physiology

Onset of pulmonary ventilation in fetal sheep produces pial arteriolar constriction dependent on cytochrome p450 ω-hydroxylase activity

Hiroto Ohata, Debebe Gebremedhin, Jayashree Narayanan, David R. Harder, Raymond C. Koehler


With the onset of ventilation at birth, cerebral blood flow decreases as oxygenation increases, but the mechanism of cerebral vasoconstriction is unknown. Cytochrome P-450 ω-hydroxylase activity metabolizes arachidonic acid to 20-HETE, a potent vasoconstrictor, in a physiologically relevant O2-dependent manner. We tested the hypothesis that the ω-hydroxylase inhibitor, 17-octadecynoic acid (17-ODYA), reduces cerebral vasoconstriction during in utero ventilation with O2 in fetal sheep. In anesthetized pregnant sheep near term, the fetal head was exposed with the rest of the body remaining in utero. Pial arteriolar diameter was measured by intravital microscopy through a closed cranial window superfused with vehicle or 17-ODYA. Mechanical ventilation of the fetal lungs with a high O2 mixture to increase arterial Po2 from ∼20 to ∼90 Torr markedly decreased pial arteriolar diameter by 24 ± 3% (±SE) without a change in arterial pressure. In contrast, superfusion of 17-ODYA completely blocked the decrease in diameter (2 ± 3%) with increased oxygenation. Vasoconstriction to hypocapnia was intact after returning to the baseline intrauterine oxygenation state, thereby indicating that the effect of 17-ODYA was selective for increased oxygenation. In cerebral arteries isolated from fetal sheep, increasing oxygenation increased 20-HETE production. We conclude that cytochrome P-450 ω-hydroxylase activity makes an important contribution to cerebral vasoconstriction associated with the onset of ventilation at birth.

  • cerebral circulation
  • 20-HETE
  • hyperoxia
  • neonatal circulation
  • oxygen sensing
  • pial artery

the onset of pulmonary ventilation in mammalian newborns results in a sudden increase in the arterial partial pressure of oxygen (PaO2) and represents the largest increase in oxygenation typically experienced during the lifetime of the organism. The threefold increase in PaO2 can potentially generate an oxidant stress, particularly in organs such as the brain with its high lipid content and high oxidative metabolism.

One possible defense mechanism for limiting potential oxidative stress is vasoconstriction to limit an increase in O2 delivery. Experimental pulmonary ventilation of fetal sheep in utero has been shown to reduce cerebral blood flow by an amount proportional to the increase in arterial O2 content such that cerebral O2 transport and cerebral O2 consumption are unchanged (7). However, the mechanism of the cerebrovascular vasoconstriction is unknown.

One potential mediator of vasoconstriction to increased oxygenation is 20-hydroxyeicosatetraenoic acid (20-HETE). This lipid mediator is produced by ω-hydroxylation of arachidonic acid by specific cytochrome P-450 (CYP) enzymes. Production of 20-HETE in renal cortical microsomes, renal microvessels, and cerebral arteries has been noted to be dependent on Po2 over the physiological range of 20–140 Torr (6, 9). The ω-hydroxylase inhibitor 17-octadecynoic acid (17-ODYA) reduced arteriolar constriction in response to superfusion of high O2-containing solutions over skeletal muscle (3, 9, 11, 15). Cerebral vascular smooth muscle in adult cat and rat express CYP 4A isoforms that possess ω-hydroxylase activity (2, 4, 5). Formation of 20-HETE produces constriction by inhibition of calcium-sensitive potassium channels and by opening l-type calcium channels via a PKC-dependent mechanism (5, 8, 13) and decreases in 20-HETE during hypoxia permit opening of calcium-sensitive potassium channels (6). Furthermore, experiments in adult rats in vivo indicate that 20-HETE contributes to cerebral vasoconstriction during increased oxygenation associated with increased arterial blood pressure and cell-free hemoglobin transfusion (4, 17). Thus O2-dependent production of the vasoconstrictor 20-HETE is a possible candidate mediator for cerebral vasoconstriction during the acute increase in oxygenation at birth.

In the present study, intravital microscopy of pial arterioles through a closed cranial window on the skull of fetal sheep was used to assess the mechanism of vasoconstriction during increased oxygenation after the initiation of pulmonary ventilation. We tested the primary hypotheses that 1) pial arterioles constrict when fetal pulmonary ventilation increases PaO2 from intrauterine values of ∼20 Torr to postnatal levels of ∼90 Torr but not when PaO2 is maintained at ∼20 Torr during mechanical ventilation and 2) that application of the 20-HETE synthesis inhibitor 17-ODYA in the window reduces pial arteriolar constriction during increased oxygenation. Specificity of the vascular effects was assessed by examining the constrictor response to hypocapnia.


Surgical preparation.

All procedures were approved by the Johns Hopkins University Animal Care and Use Committee and were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Time-dated pregnant ewes were studied at ∼140 days of gestation (term ∼145 days). The ewes were anesthetized with halothane and mechanically ventilated through a tracheostomy with supplemental O2 to maintain arterial O2 saturation >95% and end-tidal CO2 concentration at ∼5% throughout the surgery and experimental protocol. The femoral artery was catheterized for monitoring mean arterial blood pressure (MABP). The femoral vein was catheterized for administering lactated Ringer solution. A heating blanket was used to maintain normothermia. A midline laparotomy was performed and an incision was made in the uterus. The foreleg of the fetus was exposed and catheters were inserted into the axillary artery and vein. After placing the foreleg back within the uterus, the head of the fetus was exteriorized. An endotracheal tube was inserted into the trachea of the fetus through a tracheostomy. The head of the fetus was placed in a molded plaster cast holder that was held just above the uterus by a metal frame. The rest of the body of the fetus was kept within the uterus to minimize disruption of the placental circulation. Consequently, the auditory meatus was ∼10 cm higher than the fetal heart. The fetal head was wrapped in cellophane to reduce evaporative cooling and kept warm with a heating lamp. A closed cranial window was constructed on the fetal skull over the parietal cortex ∼2 cm lateral to the midline. A 1-cm diameter plastic ring was cemented to the skull around a craniotomy. The ring contained an inlet catheter, an outlet catheter, a catheter for monitoring pressure in the window, and a thermistor. The dura was gently incised and retracted to expose the pial surface. The cranial window was filled with artificial cerebrospinal fluid (CSF) and a coverglass was glued to the top of the plastic ring to seal the window. Fluid temperature in the window was maintained at 37–38°C. The closed cranial window technique with intravital microscopy has been used successfully by others (12) on fetal sheep with an intact umbilical circulation to demonstrate vasoactivity to adenosine analogs.

Physiological measurements.

A microscope and video camera was situated above the closed cranial window for measuring the diameter of pial arterioles in vivo with the fetal head exteriorized and secured in the head holder. Diameter measurements of two arterioles were tracked with each intervention in each fetus. Diameter changes of the larger arteriole within each fetus (157 ± 68 μm) were analyzed separately from that of the smaller arteriole (75 ± 41 μm). Fetal arterial blood pressure and cranial window pressure were both measured at the approximate level of the fetal heart. Arterial pH, partial pressure of CO2 (PaCO2), and PaO2 were measured on a blood gas analyzer (Chiron Diagnostics, Halstead, Essex, UK). Arterial hemoglobin concentration, O2 saturation, and O2 content were measured on an OSM3 Hemoximeter (Radiometer, Copenhagen, Denmark).

Experimental protocol.

The experimental protocol is summarized in Fig. 1. After completion of the surgical preparation, baseline measurements were obtained of arteriolar diameter and other physiological variables. The cranial window was then superfused with either the suicide substrate CYP inhibitor 17-ODYA (10 μM; n = 6) or vehicle (0.1% ethanol; n = 6) for 30 min at a rate of 0.2 ml/min followed by a maintenance rate of 0.1 ml/min for the remainder of the experiment. Measurements were repeated after 30 min of superfusion. Next, fluid was drained from the fetal lungs. The fetal lungs were initially expanded by manual bag compression. Pulmonary ventilation then continued with the use of a positive-pressure ventilator. Because mechanical ventilation of the lungs could potentially change cerebrovascular tone independent of changes in arterial blood gases, the fetal lungs were initially ventilated with a low O2 mixture (2–9%) for 30 min to keep fetal PaO2 unchanged from baseline levels. To increase fetal PaO2 to postnatal levels of 70–100 Torr, the inspired O2 was increased to 30–50% while the umbilical circulation remained intact. To determine if any loss of constrictor responses by 17-ODYA was selective for increased oxygenation, constrictor responses to hypocapnia were tested 30 min after restoring fetal PaO2 to baseline intrauterine levels. The hypocapnic response was tested under the original low oxygenation state so that baseline diameter would be similar to those before increasing the oxygenation state and baseline diameter would be similar in the vehicle and 17-ODYA groups. Hypocapnia was induced by increasing ventilation of the fetal lungs. The degree of hyperventilation was guided by end-tidal CO2 monitoring. To minimize an accompanying increase in fetal PaO2, inspired O2 was decreased further during a 15-min period of hyperventilation.

Fig. 1.

Schematic representation of the timing of measurements of arteriolar diameter in relation to the interventions of cranial window superfusion with vehicle or 17-octadecynoic acid (17-ODYA) in two groups of fetuses, onset of fetal lung ventilation, fetal lung ventilation with a high O2 mixture, and fetal hyperventilation to produce hypocapnia.

Liquid chromatography-mass spectrometry analysis of CYP metabolites.

Fetal sheep were anesthetized by intraperitoneal injection of 65 mg/kg pentobarbital sodium and their brains were removed surgically. Cerebral arterial segments from the major cerebral arteries and pial arteries were dissected out and cleared of blood and adhering tissues. The isolated fetal sheep cerebral arterial segments from six fetal sheep were pooled and homogenized in fresh DPBs buffer on ice. The cerebral vessel homogenate was divided into 1-ml samples each in 35×10 mm Petri dishes. A 50-μl aliquot was taken from each vessel homogenate sample to determine protein concentration using the protein assay kit obtained from Bio-Rad laboratories (Hercules, CA). To determine the level of production of 20-HETE from arachidonic acid under different levels of O2, groups of six samples of fetal sheep cerebral vessel homogenates were incubated in chambers with either 21% O2 or <2% O2 for 45 min at 37°C following addition of 40 μM cold arachidonic acid and 2 mM of the cofactor NADPH. In additional studies, two groups of fetal sheep cerebral vessel homogenate samples were also incubated with 40 μM cold arachidonic acid and 2 mM of the cofactor NADPH in the absence and presence of the suicide substrate inhibitor 17-ODYA (5 μM) under 21% O2 for 45 min at 37°C. Assuming that the intracellular concentration attained in vivo was less than the 10 μM concentration in the superfusate, a concentration less than 10 μM of 17-ODYA was chosen for the in vitro evaluation of 20-HETE synthesis inhibition. The reaction was stopped by placing the sample dishes on ice. The internal standard 1.0 ng [2H2]20-HETE was added to each sample, and the samples were separately extracted twice with 3 ml of diethyl ether by vortexing for 30 s. The extraction mixture was centrifuged at 5,000 g for 5 min, and the diethyl ether layer was separated and transferred to a clean 10-ml centrifuge tube. The pooled diethyl ether extracts for each sample were dried separately under a flow of nitrogen gas. The dry residue of the sample in the centrifuge tube was reconstituted in 200 μl acetonitrile and subjected to high performance LC-MS analysis (6).

Statistical analysis.

The percent change in arteriolar diameter was analyzed for each of four interventions in each animal: 1) cranial window superfusion of vehicle or 17-ODYA as a percent of the baseline diameter before superfusion, 2) mechanical ventilation with a low O2 mixture during maintenance superfusion of vehicle or 17-ODYA as a percent of the value during superfusion of vehicle or 17-ODYA before initiation of fetal pulmonary ventilation, 3) mechanical ventilation with a high O2 mixture as a percent of the value obtained during mechanical ventilation with the low O2 mixture, and 4) hypocapnia with a low O2 mixture as a percent of the value during the previous normocapnic period with a low O2 mixture. The percent change in diameter for each intervention was compared between the vehicle- and 17-ODYA-treated groups by t-test. In addition, the change in diameter and other physiological variables during each intervention were compared within each animal by paired t-test. The effect of the two O2 concentrations on 20-HETE concentrations in vitro and the effect of 17-ODYA on 20-HETE concentrations in vitro were each analyzed by t-tests. Values are presented as means ± SE.


Measurements were made after 30 min of superfusion of the cranial window with vehicle or 17-ODYA to determine if 17-ODYA altered baseline pial arteriolar diameter. Diameters were not significantly changed for large arterioles (vehicle 8 ± 6%; 17-ODYA 4 ± 5%) or for small arterioles (vehicle 15 ± 8%; 17-ODYA −1.1 ± 2.0%). Fetal MABP and PaCO2 were unchanged over this time period, although arterial pH, PaO2, and arterial O2 saturation decreased slightly in the vehicle group (Table 1).

View this table:
Table 1.

Fetal physiological variables during in utero ventilation

The effect of mechanical ventilation of the fetal lungs without increased oxygenation was evaluated next. After instituting mechanical ventilation with a low O2 mixture, PaCO2, PaO2, arterial O2 saturation, and MABP were unchanged in both groups (Table 1). A small, significant increase in arterial pH and hemoglobin concentration occurred in the 17-ODYA group, although the values were similar to those in the vehicle group. Mechanical ventilation did not increase CSF pressure in the cranial window. The absolute diameter values of the large and small arterioles are presented in Table 1 and the percent changes are illustrated in Fig. 2. The diameters of large and small arterioles were not significantly changed during fetal pulmonary ventilation in the absence of arterial blood gas changes.

Fig. 2.

Percent change in large and small pial arteriolar diameter (±SE) during the onset of mechanical ventilation of fetal lungs with a low O2 mixture to maintain arterial blood gases at baseline intrauterine levels in fetal sheep with cranial windows superfused with vehicle (n = 6) or 17-ODYA (n = 6). Diameters were not significantly changed and were not different between groups for large vessels (P = 0.60; 1-β = 0.51 for group difference of 20% and α of 0.05) and for small vessels (P = 0.59; 1-β = 0.57 for group difference of 20% and α of 0.05).

The inspired O2 in the fetal lungs was then elevated to increase fetal PaO2 to postnatal levels of ∼90 Torr in both groups. Ventilation with the elevated O2 mixture resulted in an arterial O2 saturation of ∼95% and a doubling of arterial O2 content with no significant differences between groups, whereas PaCO2, MABP, and CSF window pressure were unchanged (Table 1). The increased oxygenation produced 23% constriction of large arterioles and 25% constriction of small arterioles in the vehicle group (Fig. 3). In contrast, pial arteriolar constriction to increased oxygenation was completely blocked in large and small arterioles of the 17-ODYA-superfused group. The diameter responses differed significantly between groups (P < 0.001), whereas arterial blood values were not different between groups.

Fig. 3.

Diameter response of large and small pial arterioles during mechanical ventilation of fetal lungs with a high O2 mixture to increase arterial oxygenation to postnatal levels in fetal sheep with cranial windows superfused with vehicle (n = 6) or 17-ODYA (n = 6). Values are percent change (±SE) from diameter during low O2 ventilation. *P < 0.001 between groups.

To test if pial arterioles were still capable of constricting in the continued presence of 17-ODYA, the fetuses were returned to their low oxygenation state and hypocapnia was induced by hyperventilation of the fetal lungs. After returning to the low oxygenation state, baseline diameters were similar for large arterioles in the vehicle (167 ± 48 μm) and 17-ODYA (169 ± 26 μm) groups and for small arterioles in the vehicle (84 ± 31 μm) and 17-ODYA (84 ± 16 μm) groups. With a decrease in PaCO2 to 19 Torr, substantial constriction occurred in large and small pial arterioles of both groups (Fig. 4), and the percent change was statistically different from zero. However, the percent decrease was not different between groups in large (P = 0.37) or small vessels (P = 0.071). Although inspired O2 was decreased further during the period of hyperventilation to minimize a concurrent increase in PaO2, arterial O2 saturation and arterial O2 content increased during hyperventilation as a result of the alkalosis (Table 2).

Fig. 4.

A: PaCO2 (±SE) during ventilation of the fetal lungs with a low O2 mixture and during hyperventilation in fetal sheep with cranial windows superfused with vehicle (n = 4) or 17-ODYA (n = 5). *P < 0.001 from low O2 ventilation baseline. B: percent change in large and small pial arteriolar diameter during hyperventilation. Constrictor responses were not significantly different between vehicle and 17-ODYA groups for large vessels (P = 0.37; 1-β = 0.55 for group difference of 20% and α of 0.05) and for small vessels (P = 0.07; 1-β = 0.62 for group difference of 20% and α of 0.05).

View this table:
Table 2.

Fetal physiological variables during hyperventilation

Incubation of fetal sheep cerebral arterial homogenate with the substrate arachidonic acid generated 20-HETE. The amount of 20-HETE generated during incubation under 21% O2 was greater than under <2% O2 (Fig. 5). In a second experiment, addition of 5 μM 17-ODYA to fetal sheep cerebral arterial homogenate under 21% O2 decreased 20-HETE formation by 68 ± 2% (n = 5).

Fig. 5.

Amount of 20-HETE generated in homogenates of fetal sheep cerebral arteries during 45-min incubation with arachidonic acid under ambient conditions of <2% and 21% O2 conditions (±SE; n = 6). *P < 0.05 from low O2.


Previous work in skeletal muscle implicated 20-HETE in vascular O2 sensing under conditions in which superfusate Po2 was increased (9, 11, 15). The present study extends this concept to the physiologically relevant condition of increased oxygenation at birth. The major new finding is that increased oxygenation produced by in utero ventilation produces cerebrovascular constriction that is blocked by an inhibitor of 20-HETE synthesis. The large magnitude of the constriction (∼25%) accompanying the threefold increase in PaO2 and onefold increase in arterial O2 content is consistent with the reduction in cerebral blood flow that maintains a constant cerebral O2 delivery previously reported in fetal sheep during increased oxygenation with in utero ventilation (7). Size dependency of the response among pial arterioles was not apparent.

In mature cerebral vascular smooth muscle, 20-HETE is known to act through a PKC signaling mechanism to produce constriction (13). The constrictor response depends on opening of l-type calcium channels and closing of calcium-activated potassium channels, which would ordinarily counteract the membrane depolarization accompanying calcium influx (4, 5, 8). Similar mechanisms of action are assumed to occur in fetal cerebrovascular smooth muscle. The role of 20-HETE in vascular regulation has not been well investigated in immature animals. A CYP metabolite has been implicated in closure of lamb ductus arteriosus during increased oxygenation (1), 20-HETE synthesis inhibition attenuates hyperoxic-induced decreases in retinal blood flow by 23% in newborn pigs (18), and 20-HETE modulates fetal pulmonary vascular tone after nitric oxide synthase inhibition (16). However, the role of 20-HETE in the adaptation to air breathing after birth in other vascular beds remains largely unexplored.

Mechanical ventilation of the fetal lungs with positive airway pressure could potentially influence pial arteriolar diameter by either mechanical effects related to the increase in intrathoracic pressure or by lung inflation reflexes. However, pulmonary ventilation without increased oxygenation did not produce a change in pial arteriolar diameter or in CSF pressure within the window. Thus the pronounced pial arteriolar constriction seen with increased oxygenation was related to the oxygenation state and not to mechanical ventilation per se.

To determine whether 17-ODYA blocked the constrictor response to other physiological stimuli, hypocapnia was induced after restoring the oxygenation state and pial arteriolar diameter to the baseline intrauterine levels. In contrast to the complete block of constriction seen with increased arterial oxygenation, constriction to hypocapnia was still present in the 17-ODYA group. Thus 17-ODYA displayed specificity for inhibiting the constrictor response to increased oxygenation. However, it should be noted that arterial O2 content increased ∼50% because of the Bohr shift in the O2-dissociation curve despite reducing the inspired O2 during hyperventilation to minimize an increase in PaO2. Thus one possibility is that the constrictor response to hyperventilation had two components: one related to hypocapnia and a second related to increased oxygenation. In this case, 17-ODYA may be expected to inhibit the increased oxygenation component and attenuate the constrictor response to hyperventilation. The lack of a statistically significant effect of 17-ODYA may be related to the small sample size and the magnitude and variability of the increase in arterial O2 content during hyperventilation.

Studies of fetal sheep customarily use chronically instrumented animals to avoid the confounding influences of anesthesia and surgical trauma. Although attempts were made to minimize disturbances to the placental circulation by not exteriorizing the entire body of the fetus, the baseline fetal PaO2 of 18–20 Torr is slightly less than the 20–26 Torr range typically reported in unanesthetized, near-term fetal sheep. A slightly lower baseline PaO2 may have augmented the constrictor response observed in the present study, but is unlikely to account for the complete inhibition of the response by 17-ODYA. We elected to directly measure arteriolar diameter through a closed cranial window, which required the use of anesthesia. Thus a limitation of this study is the use of halothane anesthesia.

The present in vitro results demonstrated the ability of cerebral arteries from near-term fetal sheep to generate 20-HETE and that the production was oxygen dependent. These arteries included large cerebral arteries and pial arteries including some as small ∼80 μm. An underlying assumption is that 20-HETE production was O2 dependent in all size segments, including the smaller pial vessels where diameter responses were measured and where the in vivo effect of O2 was not found to be dependent on size. Microvessel smooth muscles from adult brain are capable of generating 20-HETE (5, 8), and we are not aware of evidence that control of 20-HETE synthesis would not be O2 dependent in smaller vessels. Production of 20-HETE in fetal sheep cerebral arteries in vitro was inhibited 68% by 5 μM 17-ODYA. This level of inhibition is somewhat less than expected from data in other systems. At a concentration of 10 μM, 17-ODYA was found to fully block formation of 20-HETE in renal microsomes without reducing prostaglandin formation (9). In isolated cat cerebral microvessels, 17-ODYA inhibited 20-HETE formation with an IC50 ∼0.5 μM, without inhibiting prostaglandin formation (5, 8). Thus ω-hydroxylase activity in fetal sheep cerebral arteries appears to be less sensitive to 17-ODYA, and the in vivo concentration of 10 μM used in the present study may not have completely blocked 20-HETE formation. In addition, it should be noted that high concentrations of 17-ODYA can also inhibit CYP epoxygenase activity (19) and impair dilation-dependent endothelial hyperpolarizing factors (10). Although formation of epoxyeicosatrienoic acids may have been reduced by use of 17-ODYA in the present experiments, these epoxides produce vasodilation (14) and thus are unlikely to contribute to the vasoconstriction seen with increased oxygenation.

Assuming that 20-HETE synthesis in cerebral arteries acts to protect the brain from oxidant stress during the sudden increase in oxygenation at birth, adequate ω-hydroxylase activity may be considered an important homeostatic defense mechanism during the transition to air breathing. Furthermore, we speculate that this vasoconstrictive response may act to reduce the risk for intraventricular hemorrhage by limiting vascular oxidative damage and/or microvascular hydrostatic pressure during postnatal increases in arterial blood pressure.


The project described was supported by Grant Number NS-20020 from National Institute of Neurological Disorders and Stroke and by Grant Number HL-59996 from the National Heart, Lung, and Blood Institute.

The contents of this paper are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.


No conflicts of interest, financial or otherwise, are declared by the author(s).


The authors thank George Kuck for technical assistance and Tzipora Sofare for editorial assistance.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
View Abstract