Changes in Oxygen Tension and Effects on Cyclooxygenase Metabolites: III. Decrease of Retinal Prostacyclin in Kittens Exposed to Hyperoxia

(1)

Changes

in Oxygen

Tension

and Effects

on

Cyclooxygenase

Metabolites:

III. Decrease

of

Retinal

Prostacyclin

in Kittens

Exposed

to

Hyperoxia

Marie J. Stuart, MD, Dale L. Phelps, MD, and B. N. Yamaja Setty, PhD

From the Division of Pediatric Hematology-Oncology, Department of Pediatrics, State University of New York Health Science Center, Syracuse, and the Department of Pediatrics and Ophthalmology, University of Rochester School of Medicine and Dentistry, Rochester, New York

ABSTRACT. The acute phase of oxygen-induced retinop-athy is associated with vasoconstriction and occlusion of the retinal vessels. Because this acute vasoobliterative phase could be due to the inhibition in retinal vessels of the production of the potent vasodilator and antithrom-botic metabolite prostacyclin, animal experiments were performed to assess this possibility. Eight litters of 27 kittens (four to six days of age) were used. Control kittens

were left in room air; hyperoxic kittens were placed in

80% oxygen for 48 hours; recovery kittens were returned to room air for 24 hours following hyperoxic exposure. Following treatments, the animals were killed, retinas isolated, and prostaglandin formation assessed. Retinal tissues produced 6-keto-prostaglandin F,, prostaglandin F2a, prostaglandin E2, and thromboxane B, from exoge-nous arachidonate. A significant (approximately 33%) reduction in retinal 6-keto-prostaglandin Fia (the end product of prostacyclin) was observed both in the

hyper-oxic and recovery litter mates when compared with

con-trols. Both of the experimental groups also demonstrated a reduction in total retinal prostanoids that paralleled the changes observed in prostacyclin, suggesting that the biochemical effect of hyperoxia on retinal vascular ara-chidonic acid metabolism occurred at the level of cycloox-ygenase. A decrease in the local production of prostacy-din during hyperoxia is consistent with the histologic retinal changes observed during the acute phase of oxy-gen-induced retinopathy. Pediatrics 1988;82:367-372; cy-clooxygenase metabolites, retinal prostacyclin, hyperoxia.

ABBREVIATIONS. NDGA, nordihydroguaiaretic acid; 6-keto-PGF1a, 6-keto-prostaglandin F,; PGF2a, prostaglandin F,; PGE2, prostaglandin E2; TXB2, thromboxane B2; PGD2, prosta-glandin D2.

Received for publication June 1, 1987; accepted Oct 14, 1987. Reprint requests to (M.J.S.) Division of Hematology-Oncology, St Christopher’s Hospital for Children, 5th and Lehigh Aye, Philadelphia, PA 19133

Retinopathy of prematurity develops only in

in-fants whose retinal vessels are immature. Although

it has occurred in some infants without the

admin-istration of supplemental oxygen, the majority of

cases develop in premature infants who have re-ceived oxygen for prolonged periods.”2 Additionally,

it has been recognized since the 1950s that the

prolonged use of 50% inspired oxygen without

ad-justment according to the infant’s need causes the incidence of this disorder to increase significantly.3

In the animal model, the acute phase of

oxygen-induced retinopathy is associated with morphologic

alterations of the retinal vascular intima, and

vaso-constriction and occlusion of the vascular lumen in

both arterioles and capillaries.4’5 Because

prosta-cyclin produced by normal blood vessel

endothe-hum is a potent vasodilator and inhibitor of

vas-cular thrombus formation,6 we had previously

hy-pothesized that the pathologic changes observed

during the acute phase of oxygen-induced

retinop-athy might result from an inhibition of the

forma-tion of this potent vasoactive mediator.7 Using

hu-man umbilical vessels in an in vitro study, we

demonstrated that short-term hyperoxic exposure

resulted in the decreased ability of vessels to

pro-duce prostacyclin.7 A logical extension ofthis initial

work would be to translate these in vitro

observa-tions into an in vivo animal study and, moreover, to evaluate the specific microvasculature of inter-est, namely, the retina. In this study, we

demon-strated that prostacyclin production is reduced in

the partially vascularized retina of kittens exposed

to hyperoxia. Retinal prostacyclin was also found

to be decreased in kittens 24 hours into the recovery

phase following an initial period of hyperoxia.

(2)

80 60 40 z Ui >-0

I 2 3 4 5 6 7

AGE IN DAYS

Fig 1. Study design. Time periods during which control

( ), hyperoxic (- - - -), and recovery animals

(#{149}#{149})

were exposed to 21% and 80% oxygen are

graphically illustrated.

retinal changes observed during the acute phase of

oxygen-induced retinopathy.

MATERIALS AND METHODS

Pregnant queens were obtained from a specific

pathogen-free colony of laboratory cats and housed

individually. The animals were observed daily, and

the day when kittens were first discovered was

defined as day 1. The kittens were then numbered

and randomly assigned to either control, hyperoxia,

or recovery groups. As shown in Fig 1, control

animals stayed continuously in room air. Hyperoxic

animals were placed in 80% inspired oxygen for 48

hours before removal of their eyes. The recovery

kittens were placed in 80% inspired oxygen for 48

hours, after which the animals were removed to

room air for an additional 24-hour period prior to

removing the eyes for evaluation of prostaglandin

metabolite formation. Thus, as shown in Fig 1, the

hyperoxic group was placed in 80% inspired oxygen

one day later than the recovery animals. This study

design ensured that all animals within the

individ-ual litters were at similar postnatal ages when all

litter mates were killed. A total of 27 kittens from

eight litters were used, and the retinal tissues were

obtained at the ages of six to eight days of life.

Thus, the time at which retinal cyclooxygenase

products were assessed varied between the sixth

and eighth postnatal day but remained constant

within each experiment. Oxygen was provided at a

minimum flow of 10 L/min in standard infant

incubators to prevent carbon dioxide accumulation.

The mother cat provided all nutrition for the

kit-tens and was rotated between the two halves of her

litter at 12-hour intervals.

In other studies, arterial blood gas values were

determined on heparinized blood obtained by left

ventricular puncture from three- to six-day-old

kit-- - - I-1YPER0XI

I-,

I#{149}

, .

,

I

#{149}_.RECOVERY - - CONTR0L

tens who were in either 80% oxygen (n = 20) or

room air (n = 8).

All retinal tissues from a single litter were

ob-tamed on the same day and analyzed together. The

animals were anesthetized with intraperitoneal

bar-biturate, the aorta was canalized and ligated

dis-tally, and the cerebral vasculature was perfused

with warmed isotonic saline until the effluent from

the jugular veins was essentially clear. The time

required to reach this condition averaged

approxi-mately seven minutes following euthanasia. The

eyes were then enucleated, the vitreous, retina, and

choroid were carefully separated under a dissecting

microscope, and the retinal tissue was suspended

in 1 mL of 25 mmol/L of

tris(hydroxymethyl)-aminomethane-buffered calcium- and

magnesium-free Hank’s balanced salt solution, pH 7.4,

contain-ing 2 mmol/L of calcium chloride and 1 mmol/L of

magnesium chloride. The retinas from a single

an-imal were pooled, [‘4C]arachidonic acid (20 zmol/

L) was added, and the retinal tissues were incubated

at 37#{176}Cfor 30 minutes (a time by which maximum

prostacyclin production is achieved). The reaction

was terminated by the addition of 2 vol of

chioro-form-methanol (1:2, by volume), and the

metabo-lites were extracted as described by Bligh and Dyer.8

Extracts were dried over anhydrous magnesium

sulfate, evaporated under a stream of nitrogen, and

dissolved in 100 L of chloroform for prostaglandin

analysis by thin-layer chromatography. Thin-layer

chromatography was performed on silica gel G

plates (Anal Tech, Inc, Newark, DE) in an

ethyla-cetate-acetic acid (99:1, by volume) solvent system9

or in the upper phase of

ethylacetate-isooctane-acetic acid-water (90:50:20:100, by volume).’0

Ref-erence prostaglandins (6-keto-prostaglandin-F,a,

[6-keto-PGF,a}, prostaglandin F2a [PGF2a],

pros-taglandin E2, [PGE2J, thromboxane B2,

prostaglan-din D2 [PGD2]) and arachidonic acid were run on

the same plate and were visualized following

expo-sure to iodine vapor. Following radioautography,

the prostaglandin bands were identified by

compar-ing their retention factor values with those of

reference prostaglandin standards. Radiolabeled

bands were scraped into scintillation vials,

ex-tracted with 500 L of methanol, and counted in

an LKB Mini-Beta Counter using Liquiscint. The

counting efficiency in Liquiscint was 94% and 33%

for the ‘4C and 3H labels, respectively. Recovery of

individual prostaglandins determined using

3H-la-beled products was found to be between 95% and

98% for all metabolites. Prostanoid production in

picomoles per milligram of dry tissue weight was

calculated from the specific radioactivity of the

added substrate, and necessary corrections were

(3)

-

PGB2

-

PGD2

LEU KOT RIE N ES

and

HETEs Retinal tissues from an additional two control

kittens were preincubated for ten minutes in the presence or absence of 30 .tmol/L indomethacin (a cyclooxygenase inhibitor) or 30 mol/L of

nordihy-droguaiaretic acid (NDGA, a lipoxygenase

inhibi-tor) prior to the addition of [‘4C]arachidonic acid

to ascertain whether the arachidonate metabolites

produced were enzymatic products (Fig 2).

Materials used and their sources were as follows: prostaglandin standards (Upjohn Co, Kalamazoo,

MI), arachidonic acid (purity >99%) (Nu-Chek,

Inc, Elysian, MN), and [‘4C]arachidonic acid (50 to

60 mCi/mmol) and 3H-labeled prostaglandins

(Amersham, Arlington Heights, IL, or New

Eng-land Nuclear, Boston, MA).

Statistical significance was determined by

ran-domized complete block analysis of variance (F

ratio), and individual treatments were compared

with their litter mate controls by the Dunnett’s

test.1’

RESULTS

As shown in Fig 3, in response to exogenous

arachidonic acid, kitten retinal tissues produced

several prostaglandin metabolites, including

6-keto-PGF,. (the end product of prostacyclin

hy-drolysis), PGF2a, PGE2, and thromboxane B2 (the

end product of thromboxane A2) with retention

factor values of 0.08 ± 0.02 (mean ± SD), 0.12 ±

0.02, 0.23 ± 0.04, and 0.3 ± 0.02, respectively, in

the ethylacetate-acetic acid system. The second

thin-layer chromatography system, ie,

ethylacetate-ARACHIDONIC ACID

*

CYCLIC

EN DOPEROX IDES

Isomerase Thromboxone A2

RVJYflSnt<tose

PGE2ond PGF2 PROSTACYCLIN THROMBOXANE A2

6-KETO-PGF1 THROMBOXANE B2

Fig 2. Arachidonic acid metabolism via cyclooxygenase

and lipoxygenases. Enzymatic steps in conversion of

ar-achidonic acid to its various end products is outlined. Indomethacin is an inhibitor of cyclooxygenase, prevent-ing the production from arachidonic acid of metabolites 6-keto-prostaglandin F, (6KETOPGF,), prostaglan-din F2, (PGF2), prostaglandin E2 (PGE2), and

thrombox-ane B2. In contrast, nordihydroguaiaretic acid is an

inhibitor of lipoxygenases. HETES, hydroxyeicosa-tetraenoic acid. Asterisks indicate site of action of indo-methacin (*) and nordihydroguaiaretic acid (**).

-

TXB2

-

PGE

-

PGF2a

-

6KPGF1a

Origin

Fig 3. Thin-layer chromatogram (with ethylacetate-acetic acid system) of arachidonate metabolites produced by normal kitten retina at 1 week of age in response to ‘4C-arachidonic acid. Positions of standards are shown to right of thin-layer chromatography panel. Abbreviations:

AA, arachidonic acid; PGB2, prostaglandin B2; PGD2,

prostaglandin D2; TXB2, thromboxane B2; PGE2, pros-taglandin E2; PGF2, prostaglandin F2; 6KPGF,, 6-ke-toprostaglandin F,.

isooctane-acetic acid-water (which was used to

con-firm the identity of the retinal prostanoids) also

showed the formation of 6-keto-PGF,, PGF2,

PGE2, and thromboxane B2. This latter thin-layer

chromatography system did not resolve PGE2 from

thromboxane B2, however, such that the estimation

of all quantitative metabolite values were calculated from the ethylacetate-acetic acid system. Inhibition

studies with indomethacin and NDGA showed that

all retinal prostanoids were products of vascular

cyclooxygenase, in that the production of

6-keto-PGF,, PGE2, PGE2a, and thromboxane B2 was

abolished by indomethacin but not by pretreatment

with NDGA.

A significant reduction in prostacyclin was

oh-served (Table and Fig 4) when control kittens in

each litter were compared with their hyperoxic and

(4)

Thromboxane B2 Total CO Metabolites

* Values are given as picomoles per milligram of dry tissue weight.

(I)

Ui

II

HYPEROXIA CONTROL RECOVERY

TABLE. Cyclooxygenase Metabolites From Individual Experiments

6-Keto-Prostaglandin .

F, Prostagland,n E2 Prostaglandin F2..

Experiment 02 Control Recovery 02 Control Recovery 02 Control Recovery 02 Control Recovery 02 Control Recovery No. Litter Litter Litter Litter Litter Litter Litter Litter Litter Litter Litter Litter Litter Litter Litter

1 1.6 2.5 1.8 0.06 0.05 0.03 1.1 1.49 1.38 2.04 2.0 2.0 4.8 6.0 5.2

2 3.5 3.9 2.6 2.08 2.16 2.4 1.87 1.83 2.31 3.4 3.5 3.9 10.9 11.4 11.2

3 1.5 1.9 2.1 1.05 1.25 1.34 0.84 1.50 0.97 1.4 1.8 1.7 4.8 6.5 6.1

4 1.4 2.5 1.5 5.41 6.45 5.47 1.38 2.63 1.38 3.0 3.7 2.5 11.1 15.3 10.9

5 1.2 1.5 1.4 3.32 3.24 3.16 1.07 1.20 1.22 1.6 1.9 1.7 7.1 7.8 7.5

6 0.4 0.6 0.37 0.27 0.0 0.08 0.1 0.24 0.15 0.8 0.5 0.5 1.5 1.3 1.1

7 0.1 0.4 0.4 0.0 1.41 0.81 0.06 0.46 0.41 0.04 0.1 0.28 0.2 2.3 1.9

8 1.0 2.6 0.2 0.32 0.39 0.32 0.58 1.26 0.38 1.1 1.7 1.3 3.0 6.0 2.2

Fig 4. Effects of hyperoxia on production of 6-keto-prostaglandin F, (6 K PGF,) by retinal tissues. Retinal tissues from control, hyperoxic, or recovery kittens were

incubated with 20 mol/L of ‘4C-arachidonic acid for 30

minutes at 37#{176}Cin 1 mL of

tris(hydroxymethyl)-aminomethane-Hanks’ balanced salt solution, pH 7.4.

Metabolites were extracted and analyzed by thin-layer

chromatography. Connected points represent values for

control

(#{149}),

hyperoxic and recovery (A) kittens from same litter. Values in hyperoxic and recovery groups were significantly different from their litter mate controls by Dunnett’s test at P <.05.

previously described. Whereas a 6-keto-PGF1 level

of 1.99 ± 0.41 (mean ± SEM) pmol/mg of dry

retinal tissue weight was measured in control

kit-tens, the values of the litter mate hyperoxic and

recovery animals were 1.34 ± 0.36, and 1.30 ± 0.31

pmol, respectively (P <.05). The decrease in retinal

prostacyclin production was accompanied by a

con-comitant decrease in total retinal cyclooxygenase

metabolite production, with the reduction in total

prostanoids mirroring closely the observed

inhibi-tion in prostacyclin (Table and Fig 5). Whereas a

value for total cyclooxygenase metabolites of 7.08

± 1.61 (mean ± SEM) pmol/mg was produced in

HYPEROXIA CONTROL RECOVERY

Fig 5. Effect of hyperoxia on production of total pros-tanoids by retinal tissues. Retinal tissues from control,

hyperoxic, or recovery kittens were incubated with 20

jmol/L of ‘4C-arachidonic acid for 30 minutes at 37#{176}Cin

1 mL of tris(hydroxymethyl)aminomethane-Hanks’

bal-anced salt solution, pH 7.4. Metabolites were extracted

and analyzed by thin-layer chromatography. Total

pros-tanoid production is sum of 6-keto-prostaglandin F1,

prostaglandin F2a, prostaglandin E2, and thromboxane B2

production. Connected points represent values for control

(#{149}),

hyperoxic (i), and recovery (A) kittens from the

same litter. Hyperoxic and recovery values were signifi-cantly different from their litter mate controls by Dun-nett’s test at P <.05 or P <.01.

control kittens, the values for litter mate hyperoxic

and recovery animals were 5.43 ± 1.43 and 5.76 ±

1.39 pmol, respectively (P <.01 and P <.05).

Al-though values fot retinal thromboxane B2 were

minimally lower in the hyperoxic (1.67 ± 0.39) and

recovery groups (1.74 ± 0.4), they were not

signifi-cantly different from those of the controls (1.9 ±

0.44 pmol). As depicted in Figs 4 and 5, there was

marked interlitter variability in the values for both

retinal 6-keto-PGF,a and total cyclooxygenase

(5)

No significant differences in pH and PCo2 have

been observed when arterial blood gases were

com-pared in control and hyperoxic kittens in our other

studies. Kittens in 80% inspired oxygen

demon-strated a pH of 7.45 ± 0.7 (mean ± SD), Pao2 of

305 ± 91 mm Hg, and PCo2 of 26 ± 7 mm Hg. A

pH of 7.45 ± 0.7, Pa02 of 90 ± 26 mm Hg, and

PCo2 of 30 ± 5 mm Hg were measured in control

animals.

DISCUSSION

As shown in Fig 2, arachidonic acid is

metabo-lized via the enzyme cyclooxygenase by vascular

endothelial cells to prostacyclin, a potent

vasodi-lator and platelet antiaggregatory compound that

is in part responsible for the antithrombogenic

po-tential of the vessel wall. Prostacyclin is rapidly

hydrolyzed to 6-keto-PGF,a, which is its stable

metabolite. Other vascular cyclooxygenase products

include PGE2 and PGF2, which are formed from

endoperoxides via the reductase or isomerase

en-zymes. In some species, endothelial cells also

con-tam the enzyme thromboxane synthetase that

con-verts endoperoxide intermediates to the

vasocon-strictor and platelet proaggregatory compound

thromboxane A2. Thromboxane A2 is unstable,

undergoing rapid hydrolysis to thromboxane B2 its

stable metabolite.

We have shown that the microvasculature of the

normal kitten retina at 1 week of age produces the

cyclooxygenase metabolites 6-keto-PGF,a,

throm-boxane B2, PGE2, and PGF2a. In vivo exposure of

newborn animals to hyperoxic conditions (80%

oxy-gen for 48 hours) causes a reduction of

approxi-mately 33% in retinal prostacyclin production in

response to exogenous arachidonate. Litter mates

who were exposed to oxygen for 48 hours and were

then allowed to recover for a day in room air

main-tamed a similar reduction in their retinal

prosta-cyclin values. A reduction in total retinal

prosta-noids that closely paralleled the changes observed

in prostacyclin was seen in both the hyperoxic and

in the recovery animals. Although the results are

statistically significant there was marked interlitter

variability in the values of the cyclooxygenase

prod-ucts. Because maximum intra- and interassay

var-iability were noted to be 8% and 17%, respectively,

the cause for the variability in metabolite

produc-tion appears to reside in other factors. Interlitter

variability could be due to the differences in

post-natal age at which each litter was killed, especially

because retinal vascularity increases rapidly during

the first 4 weeks of postnatal life. Moreover, litters

vary in maturity at birth, and when retinal

perfu-sions are performed in control animals, there is

distinct variability in retinal vascularity between

kittens from the same and from different litters:

The results of the present in vivo study are

consistent with our previous in vitro work. We had

shown that short-term (20 minute) in vitro

expo-sure of umbilical vessels to a 95% oxygen-5%

car-hon dioxide gas mixture caused a 30% inhibition in

the ability of the vessels to produce prostacyclin.7

In the present study, we observed that a nearly

identical effect occurs following the in vivo

expo-sure of retinal microvasculature to hyperoxia of 48

hours’ duration in the newborn kitten model. The

finding that prostacyclin and total prostanoid

formation were inhibited to a similar degree

sug-gests that the effect of oxygen on the microvessels

in vivo occurs at the level of vascular

cyclooxyge-nase. This finding is in keeping with the work of

others’2’4 and with our previous in vitro kinetic

analyses of cyclooxygenase from human umbilical

arteries that were exposed to elevated oxygen

ten-sions.’5 Our previous studies had shown that

de-creased prostacyclin production by hyperoxic tissue

was due to cyclooxygenase inactivation. Although

the maximum rate of conversion of arachidonic acid

to prostanoids was decreased by 30%, the affinity

for arachidonate of the cyclooxygenase enzyme

re-mained unchanged.’5 Our observation that retinal

prostacyclin and total prostanoid formation

re-mains significantly decreased in the recovery

ani-mals who had been placed in room air for 24 hours

following their exposure to hyperoxia suggests the

continued presence of inhibitors of vascular

cy-clooxygenase during the recovery period or a slow

turnover rate of this enzyme in retinal tissues.

Whether a 30% decrease in prostacyclin

forma-tion in the microvasculature can cause

vasoocclu-sion will need further clarification, because no

per-tinent studies have addressed this issue to date.

When compared to the macrovasculature, however,

baseline values for prostacyclin appear to be

de-creased in the microvessels.’6 Thus, although a

30%-decrease in the production of the

antithrom-botic metabolite prostacyclin may not be significant

in large vessels, a similar decrease may be of much

greater importance in vasculature (retinal) in which

baseline production is low.

A role for the metabolites of arachidonic acid in

the pathogenesis of oxygen-induced retinopathy

has been previously suggested. The administration

of the irreversible cyclooxygenase inhibitor aspirin

to oxygen-exposed puppies markedly enhanced the

degree of oxygen-induced retinopathy in a previous

report.’7 In humans, conflicting effects on the

reti-nopathy of prematurity have been attributed to

indomethacin.”.”9 One other animal study has

(6)

in response to hyperoxia.2#{176} Beagle puppies were

used, and retinal tissues were assessed for

prosta-noid formation following exposure of the animal

tissues in vitro to various gaseous combinations.

Changes in retinal prostacyclin levels were found

to be randomly distributed and no characteristic

patterns were noted following hyperoxic exposure.

However, puppies of various ages (two to 29 days)

were used with small numbers (n = 2) in each group

studied. In view of the small sample sizes, the

failure to detect significant differences in retinal

prostacyclin is not surprising. In addition, rather

than an in vivo exposure to oxygen as described in

our study, the puppies were initially killed,

follow-ing which their minced retinal tissues were exposed

to in vitro hyperoxia.

During the acute vasoobliterative phase of

oxy-gen-induced retinopathy, the visualized histologic

retinal pathologic changes are compatible with

those expected from local inhibition of endothelial

cell prostacyclin production, ie, vasoconstriction,

thrombus formation, and occlusion of the retinal

vascular lumen. We have demonstrated that

expo-sure of kittens to hyperoxia does lead to alterations in cyclooxygenase activity that result in a reduction

in retinal prostacyclin and total prostanoid

forma-tion. This change is observed acutely in hyperoxia

and also in animals 24 hours into a room air

recov-ery period following their hyperoxic insult. The

data suggests that inhibition of the potent

vasodi-lator and antithrombotic metabolite prostacyclin

plays a role in the acute vasoobliterative phase of

oxygen-induced retinopathy in the animal model.

ACKNOWLEDGMENTS

This work was supported by Public Health Service

grant EY-05614 from the National Institutes of Health.

The authors thank Carolyn Ganley for technical

as-sistance and Nancy Bovalino for secretarial support.

REFERENCES

1. Kinsey VE, Arnold HJ, Kalina RE, et a!: Pao2 levels and retrolental fibroplasia: A report of the cooperative study. Pediatrics 1977;60:655

2. Lucey JF, Dangman B: A reexamination ofthe role of oxygen in retrolental fibroplasia. Pediatrics 1984;73:82-96

3. Kinsey VE: Cooperative study of retrolental fibroplasia and the use of oxygen. Ophthalmology 1956;56:481

4. Ashton A, Ward B, Serpell G: Role of oxygen in the genesis of retrolental fibroplasia: A preliminary report. Br J Ophthalmol 1953;37:513

5. Ashton A, Blach R: Studies on developing retinal vessels. Effecta of oxygen in the retinal vessels of the ratling. Br J Ophthalrnol 1961;45:321

6. Moncada 5, Vane JR: Arachidonic acid metabolites and the interactions between platelets and blood vessel walls. N Engl J Med 1979;300:1142

7. Stuart MJ, Setty BNY, Walenga RW, et al: Effects of hyperoxia and hypoxia on vascular prostacyclin formation in vitro. Pediatrics 1984;74:548-553

8. Bligh EG, Dyer WJ: A rapid method oftotal lipid extraction. Can J Biochem Physiol 1959;37:911

9. Vanderhoek JY, Feinstein MB: Local anesthetics, chlorpro-mazine and propranolol inhibit stimulus-activation of phos-pholipase A2 in human platelets. MolPharmacol 1979;16:171 10. Lapetina EG, Cuatrecasas P: Stimulation of phosphatidic

acid production in platelets proceeds the formation of ar-achidonate and parallels the release of 5HT. Biochim Bio-phys Acta 1979;573:394

11. Steel RGD, and Torrie JH: Principles and Procedures of Statistics, ed 2. New York, McGraw Hill, 1970, p 188 12. Smith WL, Lands WEM: Oxygenation of polyunsaturated

fatty acids during prostaglandin biosynthesis by sheep ye-sicular glands. Biochemistry 1972;11:3276

13. Lands WEM, Sauter J, Stone GW: Oxygen requirement for prostaglandin biosynthesis. Prostaglandins Med 1978;1:117 14. Halliwell B, Gutteridge JMC: Oxygen toxicity, oxygen

rad-icals, transition metals and disease. Biochem J 1984;219:1 15. Setty BNY, Walenga RW, Stuart MJ: Kinetic analyses of

the effects of hyperoxia and hypoxia on vascular cyclooxy-genase activity in vitro. Biochem Biophys Res Commun 1984;125:170

16. Charo IF, Shak 5, Karasek MA, et a!: Prostaglandin 12 is not a major metabolite of arachidonic acid in cultured endo-thelial cells from human foreskin microvessels. J Clin Invest 1984;74:914

17. Flower RW, Blake DA: Retrolental fibroplasia: Evidence for a role ofthe prostaglandin cascade. Pediatr Res 1981;15:1293 18. Sun 5, Aranda Z, Vangvanichyakorn K: Indocin and

reti-nopathy of prematurity (ROP), abstracted. Pediatr Res 1982;16:107

19. Procianoy RS, Garcia-Prats, JA, Hittner HM, et a!: Use of indomethacin and its relationship to retinopathy of prema-turity in very low birth weight infants. Arch Dis Child 1980;55:362

(7)

1988;82;367

Pediatrics

Marie J. Stuart, Dale L. Phelps and B. N. Yamaja Setty

of Retinal Prostacyclin in Kittens Exposed to Hyperoxia