The pulmonary vasoconstrictor response to hypoxia: effects of inhibitors of prostaglandin biosynthesis

Hypoxic Pulmonary Vasoconstriction

It has been known for more than 60 years, and suspected for over 100, that alveolar hypoxia causes pulmonary vasoconstriction by means of mechanisms local to the lung. For the last 20 years, it has been clear that the essential sensor, transduction, and effector mechanisms responsible for hypoxic pulmonary vasoconstriction (HPV) reside in the pulmonary arterial smooth muscle cell. The main focus of this review is the cellular and molecular work performed to clarify these intrinsic mechanisms and to determine how they are facilitated and inhibited by the extrinsic influences of other cells. Because the interaction of intrinsic and extrinsic mechanisms is likely to shape expression of HPV in vivo, we relate results obtained in cells to HPV in more intact preparations, such as intact and isolated lungs and isolated pulmonary vessels. Finally, we evaluate evidence regarding the contribution of HPV to the physiological and pathophysiological processes involved in the transition from fetal to neonatal life, pulmonary gas exchange, high-altitude pulmonary edema, and pulmonary hypertension. Although understanding of HPV has advanced significantly, major areas of ignorance and uncertainty await resolution.

Hypoxia and pulmonary vascular endothelium

The steady development of pulmonary hypertension in the rat between Days 1 and 14 of exposure to hypoxia is based largely on structural remodelling of the pulmonary arterial circulation. The changes induced by hypobaric hypoxia (380 Torr; PaO2 of 40 mmHg) are different in the pre-acinar muscular arteries from the intra-acinar (alveolar wall) arteries, including the effect on the endothelial cell. Quantitative analysis of normal lungs has revealed structural features peculiar to the pulmonary microcirculation. The precapillary arterial unit that lies within the alveolar region includes (i) a non-muscular region and (ii) a partially muscular region as well as (iii) a muscular region. The endothelial cells in these segments or 'compartments' appear different from each other in the normal animal and behave differently in disease. In the various segments the endothelial cell has different neighbours and there are differences in its boundary markers (the basement membrane and elastic lamina). In the normal precapillary arterial unit the neighbour may be a pericyte, intermediate cell or smooth muscle cell. The pericyte and intermediate cell are precursor smooth muscle cells that under the influence of hypoxia develop into mature muscle.

COX and the control of the pulmonary circulation

The lung possesses a low resistance circulation that must, at all times, accommodate the entire cardiac output. Moreover, to ensure proper oxygenation of hemoglobin, blood flow must be directed to well-ventilated lung units, i.e., ventilation and perfusion must be matched. Cyclooxygenase (COX)-mediated arachidonic acid (AA) metabolites can influence both pulmonary vascular resistance and ventilation-perfusion matching. In isolated dog and rabbit lungs, ventilation of the right lung with oxygen (O2) and the left lung with nitrogen (N2) resulted in normal perfusion pressures and blood gas tensions. In both species, administration of COX inhibitors resulted in redistribution of blood flow from hypoxic alveoli (N2-ventilated lung) to the well-oxygenated alveoli (O2-ventilated lung) and an increase in arterial oxygen tension, i.e., a COX-mediated AA metabolite or metabolites opposed hypoxic pulmonary vasoconstriction (HPV), but, by virtue of this activity, prevented optimal matching of ventilation with perfusion. The effects of COX inhibition were reversed by prostacyclin (PGI2). These findings are consistent with the hypothesis that COX-derived products of AA metabolism can influence ventilation-perfusion matching in the lung. Moreover, the results of studies in dogs suggest that one such metabolite of AA is PGI2.

Hypoxic pulmonary vasoconstriction in the adult respiratory distress syndrome

Increased pulmonary vascular resistance (PVR) and microvascular hyperpermeability resulting in lung edema and arterial hypoxemia are mainstays in the development of adult respiratory distress syndrome (ARDS). The proposed pathophysiologic mechanisms include activation of complement and polymorphonuclear leukocytes secreting lysosomal enzymes, toxic oxygen metabolites (TOM) and eicosanoids. Platelets and coagulation factors are also involved, and in the most severe cases even monocytes are activated as reflected in release of thromboplastin. The latter may elicit disseminated intravascular coagulation (DIC). Under physiologic conditions lung blood flow is diverted from poorly to better oxygenated areas by way of hypoxic pulmonary vasoconstriction (HPV), thereby counteracting a decrease in arterial oxygenation. Many vasoactive substances have been proposed and again refuted as possible mediators of HPV. In this study we have focused on the following: histamine, catecholamines, arachidonates, calcium, phosphoinositides and TOM as well as endothelium-derived relaxing and constricting factors. Whether HPV is present in ARDS and whether it is advantageous or not seems to depend on the stage and extent of disease. We discuss possible interactions between HPV and ARDS mediators and between HPV and various vasoactive agents tested for therapeutic effects. Out of the abundance of mediators released, prostacyclin, prostaglandin E1, activated complement and platelet activating factor have been shown explicitly to inhibit HPV whereas others are suspected of doing so. In therapeutical use, prostacyclin has proved to reduce PVR and at the same time enhance cardiac output and oxygen delivery. In mild to moderate ARDS, improvement of arterial oxygenation has also been obtained employing almitrine bismesylate, a potentiator of HPV. Experimentally, adenosine effectively reduces increments in PVR and microvascular permeability with modest effects on systemic circulation. However, further investigations are warranted to decide whether adenosine or more specific blockers as, for instance, monoclonal antibodies against tumor necrosis factor should be integrated in ARDS therapy in the future.

Halothane inhibits hypoxic pulmonary vasoconstriction in the presence of cyclooxygenase blockade

Using an isolated lung the effects of halothane on hypoxic pulmonary vasoconstriction (HPV) were studied in the presence of cyclooxygenase blockade. The pulmonary vasculature can be divided into arterial, middle and venous segment resistances. Analysis of the vascular pressure-flow relationship further separates resistance into a flow dependent resistance (1/slope) and a zero-flow pressure intercept (PCRIT). We ventilated six lobes with control (35 per cent O2) and hypoxic (three per cent O2) gas mixtures with the addition of either 0, 0.5, 1.0, or 2.0 per cent halothane. We found that after addition of indomethacin (5 mg.kg-1), ventilation with three per cent O2 increased total resistance by 87 per cent over baseline with the increase primarily in the middle vascular segment. During normoxic ventilation PCRIT was 7.9 cm H2O and this increased significantly with hypoxia to 11.5 cm H2O). Only 2.0 per cent halothane blocked the increases in middle segment resistance and in PCRIT. We conclude that following cyclooxygenase blockade, halothane inhibits HPV by acting on middle segment vessels.

Relation of TXA2 and PGI2 to the difference in hypoxic pulmonary vasoconstriction between different strains of rats

TXB2 and 6-keto-PGF1 alpha levels in arterial and venous plasma of Wistar and Hilltop rats during hypoxia were measured to investigate the roles of TXA2 and PGI2 in hypoxic pulmonary vasoconstriction (HPV) and responsiveness difference of pulmonary vessels to hypoxia between different strains of rats. The results showed that PGI2 might play an important role in maintaining the low resistance in pulmonary circulation of these two strains of rats. Increased TXA2 during hypoxia may partially mediate HPV in Wistar rats, while augmented PGI2 during hypoxia may modulate HPV in Wistar rats. This might be the important mechanism responsible for more intensive responsiveness of pulmonary vessels to hypoxia in Hilltop rats than in Wistar rats.

Cyclooxygenase inhibition and effects of hypoxia on pulmonary circulation and gas exchange in anesthetized dogs

To investigate whether endogenously produced prostanoids are involved in hypoxic pulmonary vasoconstriction, pulmonary hemodynamic and gas exchange parameters and eicosanoid metabolites were measured in 5 anesthetized, artificially ventilated dogs (mean body weight 27 kg). Hypoxia elicited pulmonary vasoconstriction, but blood plasma levels of thromboxane B2 (TXB2) and 6-keto-prostaglandin F 1 alpha (6kPGF1 alpha) (stable metabolites of TXA2 and prostaglandin I2, respectively) remained unchanged. Administration of the cyclooxygenase inhibitor indomethacin blocked the synthesis of prostanoids, so that 6kPGF1 alpha and TXB2 levels decreased to values below the detection level (10 pg.ml-1) both during normoxia or hypoxia, but did not affect pulmonary vascular resistance or the alveolar-arterial PO2 difference (PAi-Pa)O2. The pulmonary vascular bed remained, however, responsive to TXA2 as evidenced by infusion of the TXA2 mimetic, U 46619, which significantly increased the pulmonary vascular resistance and (PAi-Pa)O2. Our data suggest that prostanoids are not involved in eliciting the effects of hypoxia on pulmonary hemodynamics and gas exchange efficiency.

Cardiovascular effects of leukotrienes in neonatal piglets. Role in hypoxic pulmonary vasoconstriction?

We investigated the effects of exogenous leukotriene D4, synthesis inhibitors, and a leukotriene receptor antagonist upon chloralose anesthetized, mechanically ventilated, neonatal piglets with constant left pulmonary blood flow and upon piglets with uncontrolled pulmonary blood flow. Leukotriene D4 (100-10,000 ng, intravenously) caused dose-dependent increases in peak tracheal pressure, pulmonary vascular resistance, and systemic arterial pressure. In a limited number of experiments, cardiovascular responses to exogenous leukotriene C4 were qualitatively similar but quantitatively less than those to leukotriene D4. Neither treatment with diethylcarbamazine or the lipoxygenase inhibitor nordihydroguaiaretic acid, nor with the leukotriene receptor antagonist, FPL55712, altered any baseline cardiovascular parameter measured, suggesting the absence of any influence of leukotrienes on resting hemodynamics. Hypoxia or hypoxia combined with mild hypercapnia caused pulmonary vasoconstriction. Neither treatment with diethylcarbamazine or the lipoxygenase inhibitor nordihydroguaiaretic acid, nor with the leukotriene receptor antagonist FPL55712, altered the pulmonary vasoconstrictor response to hypoxia or combined hypoxia/hypercapnia. We conclude that endogenous leukotrienes do not appear to have an influence on resting cardiovascular function, neither do they appear to be necessary for hypoxic pulmonary vasoconstriction in the neonatal piglet, although exogenous leukotrienes are capable of producing cardiovascular effects.

Canine nonresponders to alveolar hypoxic vasoconstriction and quantitative restoration of the response by aspirin I-3

Hypoxic pulmonary vasoconstriction is characterized by considerable variability in rate of response (pulmonary vascular resistance [PVR] as a function of time under hypoxia). To further define this response, forty-five closed chest dogs were anaesthetized (pentobarbitone sodium), intubated, and mechanically ventilated. Constant left lower lobar pulmonary artery flow was maintained through a balloon tipped 14F catheter via an extracorporeal pump at a rate to achieve lobar pulmonary artery pressure (Plobar) equal to main pulmonary artery pressure (PPA) and thereafter held constant. Left ventricular end diastolic pressure (PLVED) was measured by left ventricular catheter and lobar pulmonary artery flow rate (Q) by flow meter. Lobar PVR (mmHg/min per 1) was calculated every 15 min. Ventilation with 10% oxygen (O2) separated two groups based on the increase in PVR over time: twenty-two rapid hypoxic responders [HR] (slope= delta PVR/delta min greater than 0.3) and twenty-three slow or nonresponders [NR] (slope less than 0.1). The twenty-three NR were divided into two groups. Ten NR dogs (NR control) had no change in mean PVR (23.9, s.d. = 8.2, to 24.1, s.d. = 9.6) over a mean of 78 min and were used as controls. Thirteen NR dogs (NR ASA) had no change in mean PVR (32.9, s.d. = 9.5, to 32.3, s.d. = 9.8) over 75 min and were given aspirin (ASA), 10-15 mg/kg intra-arterially. The NR ASA group mean PVR then increased from 32.3 (s.d. = 9.8) to 59.1 (s.d. = 23.9, 82.9% increase, P less than 0.01) over a mean of 54 min. The mean PVR for the twenty-two HR rose from 39.8 (s.d. = 34.0) to 64.5 (s.d. = 36.6, 62.1% increase, P less than 0.01) over a mean of 72 min. The slopes of rate of response for HR (0.66) and for NR ASA (0.88) were not significantly different. The absolute values of PVR reached after plateau for HR and for NR ASA (after ASA) were also not different. Aspirin restored the NR capability to develop pulmonary vasoconstriction in response to alveolar hypoxia. The rate of response and the absolute level of response reached were also restored by aspirin.

Failure of acute hypoxia to alter pulmonary prostaglandin metabolism in dogs

We studied the effects of acute hypoxia (Fi02 = 0.09-0.11, 20 min.) on transpulmonary plasma prostaglandin (PG) concentrations in ten anesthetized, paralyzed, artificially ventilated dogs. Concentrations of 6-keto-PGF1 alpha, TxB2, PGE2, PGF2 alpha, and 13,14-dihydro-15-keto-PGF2 alpha were measured from the pulmonary artery and abdominal aorta using radioimmunoassay. In an additional six dogs, the effects of arachidonic acid (AA) infusions (100 mcg/kg/min) during normoxia and acute hypoxia were determined. Compared to normoxic conditions, acute hypoxia increased pulmonary artery pressure (p less than 0.05), decreased both the arterial oxygen tension (PaO2) and the alveolar-to-arterial oxygen tension gradient (A-aDO2) (p less than 0.05), but did not affect transpulmonary plasma PG concentrations. AA infusions significantly (p less than 0.05) increased 6-keto-PGF1 alpha independent of FiO2. Acute hypoxia failed to elicit a pulmonary pressor response in the AA-treated animals although PaO2 and A-aDO2 decreased (p less than 0.05). These data in healthy dogs suggest that (1) acute hypoxia does not alter net pulmonary PG metabolism, (2) prostacyclin synthesis is stimulated by increased plasma AA concentrations and (3) this effect may block normal pressor responses to hypoxic stimuli.

Hypoxia stimulates prostacyclin generation by dog lung in vitro

We have investigated the possibility that pulmonary biosynthesis of prostacyclin and thromboxane A2 (TXA2) may be affected by variations in PO2. Fresh lung homogenated from 8 dogs were incubated for 1 hour at 37 degrees C in Krebs-Ringer solution, at high (492 mmHg) or low (53 mmHg) PO2. After incubation, the stable metabolites of prostacyclin (6-keto-PGF1) and of TXA2 (TXB2) were measured by radioimmunoassay. The basal ("pre-incubation") levels of these metabolites, measured in tubes in which biosynthesis was arrested by the addition of indomethacin (10 g/ml), were 0.76 +/- 0.15 and 0.76 +/- 0.11 ng/mg wet wt. for 6-keto-PGF1 alpha, in high and low PO2, respectively, and 0.97 +/- 0.09 and 0.82 +/- 0.15 x 10(-1) ng/mg wet wt. for TXB2, in high and low PO2 respectively. Synthesis during incubation in other tubes was estimated by subtracting basal values from those measured at the end of incubation. More 6-keto-PGF1 alpha was formed in homogenates exposed to low PO2 (3.01 +/- 0.45) than in those exposed kept at high PO2 (1.89 +/- 0.37, p less than 0.001), but equal amounts of TXB2 were bound under both conditions. These results suggest that hypoxia may stimulate pulmonary prostacyclin synthesis; a pulmonary vasodilator, prostacyclin may help modulate hypoxic vasoconstriction in the lung.

Skeletal muscle lactate dehydrogenase isozymes and fibre composition in viral, mycoplasma and bacterial infections in young and old men

The activities of LD isozymes were studied by thermoinactivation in skeletal muscle biopsies from young and old patients suffering from viral, mycoplasma or bacterial infections and from corresponding controls. The activities of LD2-5 decreased 35-45% and that of LD1 17-29% in the muscle of the patients compared to that of the controls. There were no differences between age groups. In the controls good correlations were demonstrated between the relative activities of the LD isozymes and the muscle fibre composition, but such correlations could not be found in the patients. This indicates that the decrease of the activity of muscle LD in these infections is not associated with alterations in any specific muscle fibre type but rather with an effect of a general reduction of the LD activity.

Release of vasodilator prostaglandin, PGI2, from isolated rat lung during vasoconstriction

Prostaglandins are generated by the lungs and released into the circulation. Since inhibitors of prostaglandin synthesis enhance hypoxic pulmonary vasoconstriction, we wondered whether vasodilator prostacyclin was synthesized by the lungs in response to vasoconstriction. To test this hypothesis, we measured vasoconstriction induced by angiotensin II in isolated rat lungs before and after inhibition of prostaglandin synthetase. We found that sodium meclofenamate enhanced and prostacyclin and its precursor arachidonate abolished pulmonary vasoconstriction. In lungs labeled with 14C-arachidonate, effluent radioactivity increased after angiotensin II-induced vasoconstriction. Hypoxic vasoconstriction, but not hypoxia per se, caused an increase in lung effluent radioactivity. Chromatographic analysis of lung effluent showed that 6-keto-prostaglandin F1 alpha was the major arachidonic acid metabolite released during pulmonary vasoconstriction. We concluded that prostacyclin is produced by the lungs in response to vasoconstriction. Hypoxia per se seems not to be the adequate stimulus for enhanced lung prostacyclin formation. Lung prostacyclin may protect the pulmonary circulation against excessive vasoconstriction.

Influence of prostaglandin synthesis inhibitors on pulmonary vasodilatory effects of hydralazine in dogs with hypoxic pulmonary vasoconstriction

TO DETERMINE WHETHER HYDRALAZINE, A SYSTEMIC VASODILATOR, EXERTED A SIMILAR EFFECT ON THE PULMONARY CIRCULATION, WE STUDIED THE CIRCULATORY CHANGES IN DOGS DURING THREE INTERVENTIONS: (a) the control state during room air ventilation; (b) during continuous hypoxic ventilation with 10% oxygen, and maintaining continuous hypoxic ventilation; and (c) after 1 mg/kg hydralazine intravenously. Ventilation with 10% oxygen caused the mean pulmonary artery pressure to increase from 10+/-1.2 to 23+/-2.4 mm Hg (P < 0.01) and the pulmonary arteriolar resistance to increase from 1.51+/-0.19 to 5.87+/-1.10 U (P < 0.01). Hydralazine significantly lowered the pulmonary artery pressure (23.0+/-2.4 to 14.3+/-1.5 mm Hg, P < 0.01) and the pulmonary arteriolar resistance (5.87+/-1.10 to 2.87+/-0.52 U, P < 0.01). Femoral artery pressure, pulmonary artery wedge pressure, heart rate, and cardiac output remained unchanged throughout. To ascertain the contribution of the prostaglandin system to the pulmonary vasodilator effects of hydralazine, we pretreated a group of dogs with the prostaglandin synthetase inhibitor, indomethacin, 5 mg/kg s.c., twice daily for 2 d. These animals then underwent identical studies.The pretreated dogs had comparable base-line and hypoxia hemodynamic data. However, hydralazine had no effect on pulmonary artery pressure (23.3+/-1.6 vs. 21.7+/-2.3 mm Hg, NS) or pulmonary arteriolar resistance (8.03+/-1.09 vs. 7.14+/-1.42, NS) during continuous hypoxic ventilation in the indomethacin-pretreated group. Pretreatment with indomethacin did not, however, block the pulmonary vasodilator effects of intravenous prostacyclin (PGI(2)). Pretreatment with meclofenamate, a cyclo-oxygenase inhibitor structurally unrelated to indomethacin, also blocked the effects of hydralazine during hypoxic ventilation. These data suggest that hydralazine exerts a pulmonary vasodilatory effect during hypoxia-induced pulmonary vasoconstriction, and that this vasodilator effect may be mediated by prostaglandins.

Contribution of prostaglandins to the regulation of pulmonary vascular resistance in adult cats and dogs

In order to determine the role of prostaglandins in the regulation of pulmonary vascular resistance, we investigated effects of prostaglandin cyclo-oxygenase inhibition by indomethacin (2 mg/kg), meclofenamate (2 mg/kg), and aspirin (40 mg/kg) upon canine and feline pulmonary vasoconstriction caused by two stimuli: acute hypoxia (6% O2) and exogenous PGF2 alpha. Measurements were made utilizing an open chest, pump-perfused, in situ lung preparation. Cat pulmonary vasculatures were more responsive to hypoxia and PGF2 alpha than were dog's. In dogs, indomethacin, meclofenamate, and aspirin all increased pulmonary vascular resistance and augmented pulmonary vasoconstrictor responses to hypoxia and PGF2 alpha. Neither indomethacin, meclofenamate, nor aspirin affected baseline pulmonary vascular resistance of cats. Aspirin and indomethacin augmented pulmonary vasoconstriction in the cat. Meclofenamate was without effect. Further, although aspirin and indomethacin strongly inhibited pulmonary vascular pressor responses to exogenous arachidonic acid, meclofenamate decreased pulmonary responses to arachidonic acid very little. All three inhibitors nearly abolished systemic responses to arachidonic acid in the cat. Exogenous PGI2 infusion decreased pulmonary vascular resistance in the cat. Prostaglandin cyclo-oxygenase activity produces a vasodilator influence, probably PGI2, in the canine pulmonary vasculature that may be significant in protecting the animal from pulmonary hypertension and its consequences. The role of the prostaglandin system in the feline pulmonary vasculature appears to be qualitatively similar, but quantitatively less, than in the dog.

Premature onset of labor, neonatal patent ductus arteriosus, and prostaglandin synthetase antagonists--a rat model of a human problem

Premature labor and patent ductus arteriosus are two potentially fatal hazards of the human newborn infant. Prostaglandin synthetase antagonists have thus been used to suppress early labor and to close the ductus of the neonate. Indomethacin has been most effective but not free of significant complications. Neuronal necrosis may result from numerous systemic aberrations. A controlled rat model study was therefore devised to investigate fetal neuronal necrosis in relation to maternal indomethacin dose. Dams were given various treatments of 2 mg/kg of 4 mg/kg indomethacin within the last 3 days of gestation. Liquid chromatography was used to assess serum maternal and fetal drug levels. From light microscopy of more than 200 brains it was apparent that fetal neuronal necrosis correlates with maternal dose.

Pulmonary vasodilator activity of prostacyclin (PGI2) in the cat

We studied the pulmonary vascular effects of prostacylin, PGI2, in the cat with intact chest under conditions of controlled blood flow. Intralobar injections of PGI2, 0.03--1 microgram, decreased arterial pressure in the perfused lobe in a dose-dependent manner. Inasmuch as lobar blood flow was held constant and left artrial pressure was unchanged, the fall in lobar arterial pressure reflects a decrease in lobar vascular resistance. Prostaglandin E1 (PGE1) and nitroglycerin also decreased lobar arterial pressure; however, PGI2 had greater vasodilator activity than did these substances. Vasodilator responses to PGI2, PGE1, and nitroglycerin in absolute terms were dependent on the baseline level of tone in the pulmonary vascular bed. Prostacyclin reversed the hypertensive and platelet aggregating effects of ADP in the lobar vascular bed. These data indicate that PGI2 has significant vasodilator activity in the feline pulmonary lobar vascular bed.

Prostaglandin content in blood and lung tissue during alveolar hypoxia

The aim of the present work was to investigate whether prostaglandins (PGs) are synthetized and released from isolated blood-perfused rat and cat lungs secondary to vasoconstriction induced by alveolar hypoxia. The lungs were perfused with autologous blood with constant volume inflow via the pulmonary artery in a recirculating system. They were ventilated with constant volume positive pressure, and acute alveolar hypoxia was induced by ventilation with a gas containing 2% O2. A superfusion bioassay technique was used to measure PG-like activity in the perfusate from the lungs, the blood being re-oxygenated before reaching the assay tissues. The oxygenator prevented the perfusate hypoxia induced by ventilation hypoxia to affect the bioassay tissues. The assay tissues were rat stomach strip, rat colon and chick rectum. They were sensitive to calibrating doses of 0.5--1 ng/ml PGE2 and 1--2 ng/ml PGF2alpha. In another series of experiments PGs of the F-series were measured in lung tissue from normoxic and hypoxic lungs with radioimmunoassay technique. No increase in PG-like activity could be detected in the venous effluent by means of bioassay during hypoxia, nor was the lung tissue content of immunoactive PGF increased by hypoxia. The present findings indicate that alveolar hypoxia does not stimulate PG-synthesis in lungs, refuting that PGs are important mediators of the pulmonary vasoconstrictor response to alveolar hypoxia. It is concluded that PGs play no significant role in producing the pressor response to alveolar hypoxia.

Role of prostaglandins in alveolar hypoxic vasoconstriction

The role of prostaglandins as mediators of alveolar hypoxic vasoconstriction was investigated in dogs with the use of the prostaglandin synthesis inhibitors, aspirin and indomethacin. Alveolar hypoxia was induced by inserting double-lumened endotracheal tube into the carina and ventilating one lung with nigrogen while maintaining normal systemic oxygenation with 100% O(2) ventilation to the other lung. Relative perfusion to each lung was determined with 133Xenon and external counters. Infusions up to 25 mg/kg of indomethacin and up to 250 mg/kg of aspirin did not block the shift in perfusion from the alveolar hypoxic lung. In fact, the shift in perfusion from the alveolar hypoxic lung was slightly augmented by aspirin (P = 0.03). Thus, no positive role was demonstrated in the dog for prostaglandins in producing the vasoconstriction of alveolar hypoxia.