Expression of prostaglandin endoperoxide synthase in the fetal brain and interactions of prostanoids and stress hormones...

MISSING IMAGE

Material Information

Title:
Expression of prostaglandin endoperoxide synthase in the fetal brain and interactions of prostanoids and stress hormones during cerebral hypoperfusion
Physical Description:
vii, 144 leaves : ill. ; 29 cm.
Language:
English
Creator:
Tong, Haiyan, 1966-
Publication Date:

Subjects

Subjects / Keywords:
Prostaglandin-Endoperoxide Synthase   ( mesh )
Cerebrovascular Circulation   ( mesh )
Hypotension, Controlled   ( mesh )
Prostaglandins -- biosynthesis   ( mesh )
Prostaglandins -- metabolism   ( mesh )
Corticotropin -- metabolism   ( mesh )
Argipressin -- metabolism   ( mesh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 124-143).
Statement of Responsibility:
by Haiyan Tong.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 49015917
ocm49015917
System ID:
AA00011173:00001

Full Text










EXPRESSION OF PROSTAGLANDIN ENDOPEROXIDE SYNTHASE IN THE
FETAL BRAIN AND INTERACTIONS OF PROSTANOIDS AND STRESS
HORMONES DURING CEREBRAL HYPOPERFUSION














By

HAIYAN TONG


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1998
































This dissertation is dedicated to my parents and brother at home, who encouraged me to
reach high ...














ACKNOWLEDGMENTS


I wish to express my deep gratitude and appreciation to Dr. Charles E. Wood,

chairman of my supervisory committee, for all the guidance, support, and help he has

given me. His personality has made my studies in the USA much easier, and his academic

attitude and working habits have set an invaluable example for me both in school and in

my future career. His many extra hours spent on this work are especially acknowledged.

My sincere thanks are extended to the following committee members: Drs. Pushpa

S. Kalra, Maureen Keller-Wood, Colin Sumners, and James Simpkins for their guidance,

encouragement and understanding throughout this project. I also thank them for providing

unlimited laboratory access during this research.

1 want to express my appreciation to my fellow graduate students. Mr. Pinhas

Orbach, Mr. Scott Purinton, Ms. Harveen Dhillon, Ms. Deanna Deauseault, and Dr.

Darren Roesch for their friendship and help. I also want to express my appreciation to

Drs. Elaine Sumners, Hong Yang, and Hongwei Wang. for their indispensable help and

advice. My special thank goes to Dr. Farahaba Lakhdir for her collaboration in the first

study.

Last, but not least, the unending support from my husband, Dr. Ge Sun, has been

essential to the success of this work. The arrival of our first daughter, Michelle Tong Sun,

has brought much of joy when I am working on this project.















TABLE OF CONTENTS


ACKNOW LEDGMENTS .................... ............. iin

A B ST R A C T ........................................... ........ .vi

CHAPTERS

1 IN TR O D U CTIO N .............................................. 1

H ypothesis ................ .. ........ ..... ..... ... 3
Specific Aims ..................................... ......... 3

2 LITERATURE REVIEW ........................................ 5

Baroreceptors and Chemoreceptors Control of Hormone Secretion ......... 5
Prostanoid Control of Hormone Secretion ............................ 8
The Relationship between Prostanoids and Cerebral Blood Flow .......... 10
Prostanoids Biosynthesis ................... ..... ...... 12
Sum m ary ..... ............... ..... .. ....... ...... ....... .. 26

3 ENDOGENOUS PROSTANOIDS MODULATE THE ACTH AND AVP
RESPONSES TO HYPOTENSION IN LATE-GESTATION FETAL
S H E E P . .. .. . .. .. .. .. .. 2 8

Introduction ..................... ..................... 28
M materials and M ethods .................................. ...... 29
R esu lts . . . . . .. 3 6
Discussion ................. ................................ 39

4 PROSTAGLANDIN ENDOPEROXIDE SYNTHASE-2 IS THE MAJOR
CONTRIBUTOR TO PROSTANOID BIOSYNTHESIS DURING
CEREBRAL HYPOPERFUSION IN LATE-GESTATION FETAL
BRA IN TISSUES ............................................. 48

Introduction ............... ................................ 48
Materials and M ethods .................................. ....... 50









R results ................. ... .... .................... 56
D iscu ssion ............................................... 58

5 EFFECT OF PROSTANOIDS AND INDOMETHACIN ON CEREBRAL
BLOOD FLOW IN RESPONSE TO CEREBRAL ISCHEMIA IN LATE-
GESTATION FETAL SHEEP ......................... .......... 72

Introduction ...................................... .......... 72
M materials and M ethods ................................ .......... 74
R esu lts . . . . . . 7 9
Discussion ................. ........... .......... 83

6 EXPRESSION OF PGHS-2 AND FOS GENE INDUCED BY
CEREBRAL ISCHEMIA IN FETAL SHEEP BRAIN ................. 88

Introduction ..................... ................ ....... 88
Materials and Methods ................. ...................... 90
Results ........ ................ .... ..... ......... 95
Discussion ............... .................... ........ 99

7 SUMMARY AND CONCLUSIONS ........................... 117

REFEREN CES .............................. .................... 124

BIOGRAPHICAL SKETCH .......................................... 144














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

EXPRESSION OF PROSTAGLANDIN ENDOPEROXIDE SYNTHASE IN THE
FETAL BRAIN AND INTERACTIONS OF PROSTANOIDS AND STRESS
HORMONES DURING CEREBRAL HYPOPERFUSION

By

Haiyan Tong

December 1998


Chairperson: Charles E. Wood, Ph.D.
Major Department: Physiology

Fetal sheep respond to hypotension with reflex hormonal and hemodynamic

responses which return blood pressure to normal levels. Some of these responses are

mediated by both baroreceptor- and chemoreceptor- reflexes. However, a significant

portion of the reflex responses are not mediated by either baroreceptors or

chemoreceptors. We hypothesized that cerebral hypoperfusion reduces fetal cerebral

blood flow (CBF), and that reductions in CBF stimulate prostanoids biosynthesis within

the central nervous system which, in turn, stimulate adrenocorticotropin (ACTH), arginine

vasopressin (AVP) and cardiovascular responses.

Fetal sheep were chronically catheterized, either intact or subjected to a prior

carotid sinus denervation between 124 and 136 days gestation. The cyclooxygenase








inhibitor indomethacin (0.2 mg/kg) or its vehicle was injected intravenously 90 min before

the start of a 10-min period of hypotension. Hypotension induced by vena caval occlusion

significantly increased ACTH and AVP secretion, and indomethacin significantly reduced

the magnitude of this increase. Baroreceptor and chemoreceptor denervation attenuated

the ACTH and AVP responses, but these responses were not further inhibited by

indomethacin. This suggested that endogenous prostanoids partially mediate the reflex

hormonal responses to hypotension. Prostaglandin E2 and thromboxane Az were

produced in the brain in response to cerebral hypoperfusion induced by brachiocephalic

occlusion. Cerebral hypoperfusion significantly reduced CBF and indomethacin decreased

the reduction of CBF by inhibiting local prostanoid production. The increase in

prostanoid production was associated with the "suicide" inactivation of prostaglandin

endoperoxide synthase (PGHS), which is the key enzyme for prostanoid biosynthesis.

Western blot and RT-PCR analyses showed that in the brain protein levels and mRNA of

PGHS-2, but not PGHS-1, were induced by cerebral hypoperfusion.

Immunocytochemistry demonstrated that immunoreactive PGHS-2 and Fos induced by

cerebral hypoperfusion, were localized in neurons and the cerebrovasculature.

In summary, these studies indicated that hypotension stimulated ACTH and AVP

secretion, also induced expression of mRNAand protein for prostaglandin endoperoxide

synthase-2 in the fetal brain. We conclude that the induction of PGHS-2 is the major

contributor to elevated prostanoids in the brain which, in turn, modulate the ACTH and

AVP responses to cerebral hypoperfusion in late-gestation fetal sheep.














CHAPTER 1
INTRODUCTION


Hypotension initiates hormonal and hemodynamic responses which act in concert

to return blood pressure to normal levels. Many of these responses are mediated by

cardiovascular mechanoreceptors which change their firing rates in response to changes in

vessel wall or cardiac chamber radius. Cardiovascular mechanoreceptors, situated in the

walls of arteries, in the walls of cardiac atria and ventricles, and elsewhere, respond to

stretch with increased rates of firing. Arterial baroreceptors and chemorecepotors play an

important role in maintaining blood pressure during periods of hypotension in late

gestation fetal sheep (Wood, 1989; Wood et al. 1990). However, baroreceptors and

chemoreceptors do not completely mediate the responses to hypotension in the fetus.

AVP, ACTH, and renin reflex responses to hypotension (produced by vena caval

obstruction) were only 50% inhibited by peripheral baroreceptor and chemoreceptor

denervation (Wood, 1989). Therefore, this observation suggests that some other

mechanisms within the central nervous system are responsible for the remainder of the

reflex endocrine responses to hypotension in the fetus.

Moderate to severe hypotension in the fetus reduces cerebral blood flow and

reductions in fetal cerebral blood flow are effective in increasing cardiac output and

regional vasoconstriction (Backofen et al. 1991). Therefore, it is likely that some of the










reflex responses to moderate to severe hypotension in the fetus are the result of

compromised fetal cerebral perfusion. Koos et al. (1994) demonstrated that fetal hypoxia

stimulates the production of adenosine, and that some of that adenosine could originate

from the cerebral vasculature and, as such, could represent a transduction mechanism for

stimulating AVP responses to hypoxia. Prostanoids are another class of compounds

which could be stimulating hormonal reflex responses to hypotension in the fetus. The

influence ofprostanoids on the central nervous system to alter hormonal and

hemodynamic responses might mediate the homeostatic responses to cerebral ischemia.

At present, the mechanisms of these responses to reductions in cerebral blood flow are not

fully understood. Prostanoids appear to contribute to regulation of the cerebral vascular

tone in human beings and some animals, particularly during the perinatal period

(Pourcyrous et al. 1994). One site of prostanoid production is in the local

cerebrovasculature during cerebral hypoperfusion (Chemtob et al. 1990a). During

hypotension, brain and sagittal sinus concentrations of vasodilator prostaglandins have

been show to increase to maintain cerebral blood flow (Leffler and Busija, 1987a; Leffler

et al. 1986; Chemtob et al. 1990b). But concentrations of the vasoconstrictor

thromboxane are also increased in newborn piglets, suggesting the participation of

thromboxane in the decrease in cerebral blood flow during hypotension (Chemtob et al.

1990b). However, in this research project, we proposed that in addition to their direct

action at the cerebral vasculature, these prostanoids can be generated within the fetal brain

in responses to cerebral hypoperfusion, and that these prostanoids in turn modulate the

reflex cardiovascular and endocrine responses which increase blood pressure and

ultimately help ameliorate the cerebral hypoperfusion.










Hypothesis


Studies from Wood's laboratory and others demonstrate that: 1) prostaglandin E2

and thromboxane A2 act at the central nervous system to increase ACTH and AVP

secretions; 2) reductions in cerebral blood flow cause the local production of these

prostanoids in the brain; 3) baroreceptors and chemoreceptors do not completely mediate

the hormonal reflex responses to hypotension. Other mechanisms responsible for this

reflex response are not fully known yet. In this project, we hypothesized that changes in

fetal cerebral perfusion pressure stimulate increases in ACTH and AVP secretions and

cardiovascular responses, by the induction of prostaglandin endoperoxide synthase,

resulting in increased local generation of prostaglandin E2 and thromboxane A2.


Specific Aims


The overall objective of this research project is to explore the mechanism that

cerebral hypoperfusion stimulates the production ofprostaglandin E2 and thromboxane A2

within the brain and that these prostanoids in turn mediate the reflex endocrine and

cardiovascular responses to this stimulus in the late-gestation fetal sheep. My specific

objectives are as follows:

1) determine if endogenous prostaglandin E2 and thromboxane A2 modulate the ACTH

and AVP responses to hypotension

2) characterize the effects of endogenous prostaglandin E2, thromboxane A2 and

indomethacin on the reduction of cerebral blood flow








4

3) elucidate the mechanism responsible for the prostanoids biosynthesis during cerebral

hypoperfusion

a) test the inducibility of prostaglandin endoperoxide synthase-2 mRNA

and protein by cerebral hypoperfusion.

b) localize and evaluate the immunoreactive prostaglandin endoperoxide

synthase and Fos in the brain in response to cerebral hypoperfusion.














CHAPTER 2
LITERATURE REVIEW


Baroreceptors and Chemoreceptors Control of Hormone Secretion


Baroreceptors and chemoreceptors reflex systems are important mechanisms

involved in the rapid control of blood pressure in both adult and fetal animals. It is known

that arterial baroreceptors in the carotid sinus and aortic arch buffer the minute-to-minute

changes in arterial blood pressure which are produced by common disturbances in blood

pressure, including orthostasis and psychogenic stress (Cowley et al. 1973). A change in

pressure signals are transmitted from each carotid sinus through the Hering's nerve to the

glossophryngeal nerve and hence to the tractus solitarius (NTS) in the medullary area of

the brain stem. Signals from the arch of the aorta are transmitted through the vagus

nerves also into NTS of the medulla (Guyton, 1991). Carotid baroreceptor cell bodies are

located in the inferior ganglion of the glossopharyngeal nerve, and aortic baroreceptor cell

bodies are located in the inferior ganglion of the vagus nerve. Neurons in the NTS

subseving cardiovascular system regulation occupy primarily the densely packed portion

of the nucleus near the obex (Eckberg and Sleight, 1992). Afferent baroreceptor neurons

carrying the incoming signals release chemicals that alter properties of other neurons and

eventually provoke changes that affect parasympathetic and sympathetic outflow. After

the baroreceptor signals have entered the NTS of the medulla, secondary signals modulate










the vasoconstrictor center and the vagal center in the medulla in response to the incoming

signals. Anatomical studies identified monosynatic interconnections within the NTS, as

well as between NTS and the hypothalamus (Spyer, 1972), the rostral ventrolateral

medulla (Sawchenko and Swanson, 1982) and the intermediolateral region of the spinal

cord (Loewy and Burton, 1978). Feedback signals are then sent back through the efferent

autonomic nervous system to the peripheral tissues such as arterioles, veins and heart to

adjust arterial blood pressure toward the normal level on a second to second basis.

Prostaglandin endoperoxide synthase are found in various regions that are described

above, especially in the hypothalamus and in the NTS (Breder et al. 1992, 1995).

It is also well accepted that both baroreceptors and chemoreceptors influence the

secretion of the hormonal systems which are involved in blood pressure regulation. For

example, arterial baroreceptors tonically inhibit the secretion of vasopressin (Share, 1965),

ACTH and cortisol (Wood and Rudolph, 1983; Wood et al. 1984), and renin and

aldosterone (Bartter et al. 1960; Thames et al. 1978). Peripheral chemoreceptors are

thought to control the secretion ofAVP (Hanley et al. 1988) and ACTH (Raffet al.

1984). Increases in plasma Ang II and vasopressin promote vasoconstriction and direct

defense of blood pressure (Iwamoto and Rudolph, 1981; Schwartz and Reid, 1981;

Robillard et al. 1982). Increases in plasma Ang II, aldosterone and vasopressin promote

renal sodium and water conservation (Bie, 1990; Hall and Brands, 1992) and an increase

in plasma cortisol concentration promotes restitution of plasma proteins (Pirkle and Gann,

1975) and intracellular-extracellular compartment fluid shifts (Bartter and Fourman, 1962;

Fan et al. 1978).









7

In late-gestation fetus, decreases in arterial blood pressure increase the secretion of

ACTH, cortisol, AVP, and renin (Robillard et al. 1979; Rose et al. 1981; Wood, 1989).

The magnitudes of the responses are proportional to the magnitude of the decrease in

arterial pressure (Wood et al. 1982), and the responses are attenuated by prior sinoaortic

denervation (Wood, 1989). AVP secretion was not affected by arterial chemoreceptor

activity in the fetus (Raff et al. 1991). Hypercapnia in combination with hypoxia

stimulated increases in fetal plasma renin activity which were blocked by sinoaortic

denervation (Wood et al. 1990). Reductions in blood flow to the arterial chemoreceptors

could stimulate reflex ACTH responses to acute hypotension in fetuses (Wood, 1995a).

Because chemoreceptors can not account for the responses to hypotension, and sinoaortic

denervation attenuates the responses of the three endocrine systems (ACTH, AVP, and

renin), it was concluded that at least some of the endocrine responses to acute

hypotension must be mediated by arterial baroreceptors (Itskovitz and Rudolph, 1982;

Wood, 1995a). Wood (1995a) suggested that the baroreceptors are indeed active in late

gestation, and that unloading of the baroreceptors stimulates reflex responses. In intact

fetal sheep arterial baroreceptors buffer changes in arterial blood pressure caused by

orthostasis, cord compression, and other acute influences, and sinoaortic denervation

eliminates the buffer (Wood, 1995a).

Based on this information, it is concluded that both baroreceptors and

chemoreceptors located in the carotid sinus and/or aortic arch contribute to the overall

reflex responses to acute hypotension. A portion of the AVP response to acute

hypotension is mediated by arterial baroreceptors and not by chemoreceptors, ACTH










responses might also be mediated by arterial baroreceptors. However, AVP and ACTH

reflex responses to hypotension (produced by vena caval obstruction) were only 50%

inhibited by peripheral baroreceptor and chemoreceptor denervation (Wood, 1989).

Therefore, some other mechanism contributed to these reflex responses that need to be

investigated.


Prostanoid Control of Hormone Secretion


Prostanoids are a class of compounds which could be stimulating hormonal reflex

responses to hypotension in the fetus. One site of production of prostanoids is in the local

cerebrovasculature during cerebral hypoperfusion (Chemtob et al. 1990a), but in addition

to being made by and released into the vasculature, prostanoids are synthesized within

neurons in the brain of fetal sheep (Jones et al. 1993; Pace-Asciak and Rangaraj, 1976).

Physiologically, this endogenous production of prostanoids can be demonstrated by

central administration of indomethacin to late-gestation fetal sheep at doses which block

CNS prostaglandin synthesis, but not peripheral prostaglandin synthesis. Prostanoids

might alter cardiovascular function by affecting neuronal processing within the central

nervous system. For example, Breuhaus and coworkers (Breahaus and Chimosky, 1983;

Breahaus and Chimosky, 1985; Breahaus et al. 1989) and Cudd and Wood (1991) have

demonstrated that infusions of PGE2 into the carotid arteries of conscious adult sheep

increase heart rate and blood pressure. This effect is not mediated by an action of PGE,

on arterial baroreceptor or chemoreceptor afferent activity (Breahaus and Chimosky,

1983) and is therefore a direct effect on the brain. PGE2 also has a direct effect on the










fetal brain to alter blood pressure and heart rate, presumably by altering autonomic

efferent tone (Cudd and Wood, 1991; Cudd and Wood, 1992). Wood and coworkers

(1993) have found that in addition to its action in the peripheral circulation, thromboxane

also stimulates increases in arterial blood pressure and heart rate by an action in regions

perfused by carotid artery.

An important component of the effect of prostanoids on the cardiovascular system

is their effect on the hormones which, in turn, influence blood pressure or fluid balance. It

has been demonstrated that PGE2 has direct effect on the fetal sheep pituitary gland by

enhancing AVP-stimulated, but not CRH-stimulated ACTH secretion from dispersed fetal

anterior pituitary cells in culture (Brooks and Gibson, 1992). Young and Thorburn (1994)

have found that PGE2 has potent stimulatory actions on the gestation fetal sheep pituitary

to increase both the absolute concentration and the bioactive fraction of ACTH-containing

peptides in the fetal circulation. An increase in the concentration of PGE2 in the

hypothalamus, therefore, would be expected to stimulate ACTH secretion (Wood, 1995a).

It also directly stimulates glucocorticoid secretion from the fetal adrenal gland. It has also

showed that PGE2 increases AVP secretion (Inoue et al. 1991). Intracerebroventricular

infusion of PGE2 can stimulate ACTH and cortisol secretion in fetal sheep (Brooks, 1992)

and treatment with indomethacin decreases ACTH release (Thompson and Hege, 1978).

Thromboxane A, is also a potential mediator of CRF release. In vitro treatment of adult

rat hypothalami with U46619 (a thromboxane A2 mimic) causes secretion of corticotropin-

releasing factor (CRF) (Bernardini et al. 1989). Wood et al. have found that thromboxane

A, is a powerful and specific stimulator of ACTH secretion in both adult and fetal sheep










(Cudd and Wood, 1993; Cudd and Wood, 1994; Wood et al. 1993). In the adult sheep,

endogenously generated thromboxane A2 stimulates ACTH secretion, and in the fetal

sheep infusions of U46619 into the carotid arterial blood supplying the fetal brain

stimulates increases in fetal ACTH secretion (Wood et al. 1993). The ACTH response to

this stimulus is blocked through the inhibition of cyclooxygenase or through blockade of

TxA2-PGH receptor (Cudd and Wood, 1993; Cudd and Wood, 1994). Whether

endogenous prostanoids could mediate the ACTH and AVP secretions in response to

hypotension in the fetus remains to be found.


The Relationship Between Prostanoids and Cerebral Blood Flow


In fetal and newborn animals, cerebral blood flow (CBF) is maintained during

changes in arterial blood pressure (Chemtob et al. 1990a; Hernandez et al. 1980; Papile et

al. 1985; Tweed et al. 1983). In the fetal sheep, cerebral blood flow is held relatively

constant during changes in arterial blood pressure between 45 and 80 mm Hg(Papile et al.

1985; Purves and James, 1969; Tweed et al. 1983). Therefore, decreases in arterial blood

pressure below the normally regulated level result in a decrease cerebral blood flow.

Prostanoids appear to contribute to regulation of cerebral vascular tone in human

beings and some animals, particularly during the perinatal period. Substantial

physiological data in vivo point to a significant role for dilator prostanoids in the

regulation of cerebral vascular tone and cerebral blood flow in newborns (Leffler et al.

1993). The decreases in CBF occur concomitantly with decreases in CSF dilator

prostaglandin levels (Pourcyrous et al. 1990; Leffler et al. 1989). Chemtob and his










coworkers (1990b) demonstrated that reduction of cerebral blood flow in the newborn

piglet below the autoregulatory range stimulates increases in the local production of PGE2,

PGF2, and thromboxane B2.

Arachidonic acid is liberated from damaged cell membranes during ischemia and is

the source ofvasoactive prostanoids which are believed to adversely influence cerebral

ischemia. When cerebral blood flow is reduced, both dilator and constrictor prostanoids

are produced locally. The physiological effects of arachidonate metabolites often have

opposing actions. For example, PGE, and PGI2 are potent vasodilators, while TxA2 and

PGF2a are vasoconstrictors (Moncada and Vane, 1978; Moskowitz and Coughlin, 1981).

The prostanoids are most likely a component of the control of vascular tone by locally

generated effectors of tone. Chemtob et al. (1990c) have shown that PGE2, PGF2,, and

PGI2 increase CBF in newborn piglets. This effect is in sharp contrast to that observed in

adults of virtually all species studied, in which PGE, and PGF2, are cerebral

vasoconstrictors (White and Hagen, 1982; Murphy and Pearce, 1988). Thus, there exists

a predominant vasodilator action of the major cerebrovascular PGs in newborns. This

may compromise CBF autoregulation as appropriate vasoconstriction is required to

maintain CBF when blood pressure increases.

Inhibition of cyclooxygenase, an enzymatic step in arachidonic acid metabolism,

has an improvement effect on the outcome during cerebral ischemia. Indomethacin, a

cyclooxygenase inhibitor, has been used for pharmacologic closure of the patent ductus

arteriosus (Heymann et al. 1976; Firedman et al. 1976; Gersony et al. 1983), and it has a

potential ability to prevent or attenuate the development of intraventricular hemorrhage










(Ment et al. 1994). Free indomethacin crosses the blood brain barrier (Bannwarth et al.

1990). It can be found in the brain within 30 min of systemic administration (Bannwarth

et al. 1990; Day et al. 1987). It is generally assumed that the duration of action of

indomethacin is long, but the plasma half-life is relatively short and the CSF concentration

is closely related to that of plasma (Bannwarth et al. 1990). Indomethacin has been

reported to have beneficial prophylactic effects in prevention and/or reduction of severity

of intraventricular /periventricular hemorrhage in premature babies (Ment et al. 1985a;

Hanigan et al. 1988; Bada et al. 1989). In piglets, indomethacin decreases CBF at rest

(Leffler et al. 1985). These decreases in CBF occur concomitantly with decreases in CSF

levels of dilator PGs (Leffler et al. 1989).

The effects of prostanoids and indomethacin on cerebral ischemia on fetuses are

not known. We hypothesized that during cerebral hypoperfusion, prostanoids produced in

the microvasculature and neurons might cause vasoconstriction and reduce cerebral blood

flow, and that pretreatment with indomethacin may attenuate this reduction of cerebral

blood flow. The information obtained from this study may result in better understanding

and treatment of fetal cerebral ischemia.


Prostanoids Biosynthesis


When a cell-specific stimulus (hormone or growth factor) interacts with the

appropriate receptors, the subsequent signal transduction results in activation of

phospholipases. Two pathways are postulated to be responsible for the release of

unesterfied arachidonate from cell membrane. One involves the direct hydrolysis of










arachidonic acid from phospholipids by the action of phospholipase A2 (Rittenhouse-

Simmons et al. 1976; McKean et al. 1981; Isakson et al. 1977). The other begins with the

action of phosphatidylinositol-specific phospholipase C, followed by the action of

diglyceride lipase and monoglyceride lipase to liberate arachidonic acid (Chau and Tai,

1981; Borgeat and Samuelsson, 1979).

Once released, arachidonic acid is metabolized through three different pathways.

The first pathway involves a series oflipoxygenases (5, 12 or 15-lipoxygenase) that

peroxidize arachidonic acid at different carbon atoms to form unstable

hydroperoxyeicosatetraenoic acids (HPETEs) which are then converted into lipoxins

(LXs), hydroxyeicodatetraenoic acids (HETEs), leukotrienes (LTs) and related

compounds (Borgeat and Samuelsson, 1979; Fitzpatrick and Murphy, 1988). The second

pathway is known as "epoxygenase branch" of the arachidonate cascade and is mediated

by cytochrome P450 monooxygenases (Shimizu and Wolfe, 1990). The third pathway

involves prostaglandin endoperoxide synthase (PGHS or cyclooxygenase) that leads to the

formation of endoperoxide PGH2. iPG is the common intermediate for the biosynthesis

of PGs, prostacyclin (PGI2) and thromboxane A2 (TxA2) (DeWitt et al. 1981; Hamberg et

al. 1975). After PGH2 is biosynthesized, the various eicosanoids are made in a tissue-

specific fashion before they exit the cell (Fig. 2-1). Each eicosanoid can bind to multiple

receptors on either the cell of origin or neighboring cells. The eicosanoid-binding

receptors have not been fully characterized, but so far the evidence indicates that they are

all G-protein linked (Smith, 1989). Each cell type has various receptors with distinct

functions. Each receptor signals a different G-protein to alter the production of second









14

messengers (Smith and Marnett, 1991; Smith, 1989). The prostaglandins thereby help to

coordinate the cellular response to circulating hormones (Smith and Marnett, 1991; Smith,

1989).

The products of epoxygenase, lipoxygenase and cyclooxygenase pathways have a

wide range of biological activities. Eicosanoid is used to refer to this large family of

compounds. Several epoxygenase metabolites have been identified as endogenous

constituents of mammalian systems. Some of them were found to be potent inhibitors of

Na' K -ATPase in the corneal epithelium or to induce glucagon or vasopressin secretion

(Shimizu and Wolfe, 1990; Whittake et al. 1976). Thus, the epoxygenase metabolites of

arachidonic acid might be involved in the regulation of renal and ocular transport functions

which rely on the sodium pump mechanism. Prostaglandins can influence arterial and

venous blood flow, capillary permeability, tone of smooth muscle, cell chemotaxis and

sensation of pain. Leukotrienes participate in host defense reactions and

pathophysiological conditions such as immediate hypersensitivity and inflammation.

The biological activities of the eicosanoids are mainly limited to their local site of

biosynthesis. Thus, they are so called local hormones or autocoids. These eicosanoids

usually do not exert systemic effects because they are rapidly metabolized to inactive

metabolites. Some eicosanoids such as TxA2 and PGI2 are chemically unstable, with half

life measured in seconds, and undergo rapid non-enzymatic hydrolysis to inactive

compounds TxB2 and 6-keto-PGF, (Hamberg et al. 1975). Others are enzymatically

degraded to inactive compounds. The enzymatic degradation of eicosanoids occurs in

multiple organs of the body including the liver, lung, kidney, blood vessels, platelets and

erythrocytes. Liver and lung are the most important sites of metabolism.










A contemporary view of eicosanoid regulation would consider two additional

processes. The first is transcriptional and translational control of enzyme expression. The

second is the mechanism-based inactivation of eicosanoid biosynthetic enzymes.

Increasing evidence supports the role of gene regulation and enzyme inactivation as

important control mechanisms in eicosanoid metabolism.


Prostaglandin Endoperoxide Synthase


Prostaglandin endoperoxide synthase (PGHS) converts arachidonic acid to PGH2,

the precursor of prostaglandins, thromboxanes, and prostacyclin. This enzyme is present

in all mammalian cells that utilize these classes of eicosanoids (Smith and Marnett, 1991).

Small amounts of eicosanoids are produced locally and have effects only on those cells in

the immediate vicinity which express specific prostaglandin receptors. Eicosanoids have

been shown to have effects on almost every physiological system, including the

reproductive, digestive, cardiovascular, immune, nervous and respiratory systems. In

these systems, cells that make eicosanoids typically synthesize one or two specific

products rather than participate in the entire arachidonic acid cascade. For example,

platelets primarily produce TxA2, and endothelial cells primarily produce prostacyclin

(PGI2). Such bioactive lipids have both autocrine and paracrine functions.

PGHS can be induced in cultured cells by growth factors (EGF), hormones, and

tumor promoters with increases in prostaglandin production (Yokota et al. 1986; Kusaka

et al. 1988; Raz et al. 1989; Pash and Bailey, 1988; Wu et al. 1988; Goerig et al. 1989).

This increased activity is not attributed to increased arachidonate availability and is

diminished by cycloheximide and actinomycin D, indicating that protein synthesis is











necessary for the effect. The mechanism for inducing capability has been attributed to

increases in PGHS message or enzyme mass following stimulation.


Properties of PGHS


PGHS was first purified from microsomes of ovine and bovine vesicular glands

(Miyamoto et al. 1976; Van der Ouderaa et al. 1977; Hemler et al. 1976). Purified PGHS

exhibits a cyclooxygenase activity that oxygenates arachidonic acid to produce the

hydroperoxy endoperoxide PGG2 and a peroxidase activity that reduces the hydroperoxide

to the corresponding alcohol, PGH2 (Miyamoto et al. 1976). Purified protein has a

molecular weight of 70 kDa when analyzed on denaturing electrophoresis gels (Hemler et

al. 1976). However, when purified synthase is applied to a gel filtration column, it elutes

as a homodimer at an apparent molecular weight of 140 kDa. Gel filtration of PGHS can

be carried out in buffer containing 6 N guanidine.HCI to interrupt non-covalent

interactions, but the purified protein still elutes as a homodimer under these conditions

(Ogino et al. 1978).

The DNA sequence for PGHS has been elucidated from sheep, mice, and human

sources (DeWitt and Smith, 1988; Merlie et al. 1988; Yokoyama et al. 1988; Yokoyama

and Tanabe, 1989; DeWitt et al. 1990). Analysis of the protein sequence reveals

approximately 90% sequence similarity between these species. Recent evidence suggests

that there are two different RNA transcripts for PGHS, one of which is mitogen-inducible

(Xie et al. 1991; Rosen et al. 1989). These two transcripts appear to be derived from

different genes (Rosen et al. 1989). The mitogen-inducible form in chicken embryo

fibroblasts shares 59% sequence similarity with the sheep enzyme (Xie et al. 1991).









17

PGHS is a membrane protein that is associated with the endoplasmic reticulum and

nuclear membrane in Swiss mouse 3T3 fibroblasts (Miyamoto et al. 1976; Van der

Ouderaa et al. 1977; Hemler et al. 1976; Rollins and Smith, 1980). The cyclooxygenase

active site is reported to be on the cytosolic side of the endoplasmic reticulum (DeWitt et

al. 1981). Detergents are required to release PGHS from the microsomal membrane,

which has been interpreted to indicate that it is an integral membrane protein. However,

the possibility that it is a peripheral membrane protein can not be ruled out. Five

hydrophobic regions are deduced from a hydrophobicity profile of the primary amino acid

sequence. One of these regions is in the signal peptide (Yokoyama et al. 1988). There is

no obvious transmembrane segment in PGHS because none of the hydrophobic regions

have average hydrophobicities characteristic of a transmembrane sequence nor are they

long enough to span the membrane (DeWitt and Smith, 1988; Merlie et al. 1988;

Yokoyama et al. 1988).

PGHS is a glycoprotein. Approximately 5 mol ofN-acetyl glucosamine and 12

mol of mannose are contained on each 70 kDa subunit (Van der Ouderaa et al. 1977).

The attachment of these carbohydrates accounts for the discrepancy between the 65.5 kDa

molecular weight determined from the primary amino acid sequence and the molecular size

of 70 kDa determined from its electrophoretic mobility (DeWitt and Smith, 1988; Merlie

et al. 1988; Yokoyama et al. 1988). Four potential N-glycosylation sites exits in the

amino acid sequence (Merlie et al. 1988; Yokoyama and Tanabe, 1989). There is firm

evidence that two of these sites are glycosylated in the purified enzyme. Both are in the

N-terminal half of the protein (Smith and Marnett, 1991).










Cvclooxygenase and Peroxidase Activities of PGHS


Cyclooxygenase catalyzes the oxygenation of arachidonate to the prostaglandin

endoperoxide PGG2 (Smith and Marnett, 1991; Marnett and Maddipati, 1991). The

PGHS peroxidase activity catalyzes the reduction of the cyclooxygenase product PGG2 to

PGH2 (Miyamoto et al. 1976). PGH2 is further converted by specific synthases

(isomerases) or reductases to the major biologically active prostanoids PGD2, PGE2,

PGF2., PGI2 and thromboxane A2 (Smith et al. 1991). The peroxidase activity and

cyclooxygenase activity have distinct active sites on PGHS. Aspirin, which inhibits

cyclooxygenase activity, does not inhibit peroxidase activity (Roth et al. 1975; Roth and

Siok, 1978; Roth et al. 1980; Raz and Needleman, 1990).


Isoforms of PGHS


Two different isoforms of PGHS were confirmed at the protein (Habib et al. 1993)

and mRNA levels (O'Neill and Ford-Hutchinson, 1993). There are two distinct genes

which encode for PGHS-1 (or COX-1) and PGHS-2 (or COX-2). Inducible PGHS-2

encoded by a 4 kilobase mRNA, and constitutive PGHS-I encoded by a 2.8 kilobase

mRNA. Promoter analysis suggests that PGHS-1 is a member of the housekeeping gene

family (Wang et al. 1993) and PGHS-2 is a member of immediate early gene (IEG) family

(Xie et al. 1993). The promoter sequence of PGHS-2 gene contained a TATA box and a

variety of enhancer elements, including an activator protein-1 (AP-1) site (Xie et al.

1993). The PGHS-2 gene may be transcriptionally regulated at its AP-1 binding element









19

by transcription factor c-fos (Chiu et al 1988; Xie et al. 1993); however, promoter analysis

of the PGHS-1 gene reveals no putative AP-1 domain (Wang et al. 1993).

PGHS-1 is the isoenzyme which was first purified from bovine and ovine

microsomes and is a homodimer consisting of two subunits each of approximately 70

kilodaltons (Samuelsson et al. 1978). The deduced amino acid sequences of PGHS-1

from sheep, human, and mouse are -90% identical at the amino acid level, with the major

differences being found in the signal peptide and the 12 amino acids immediately preceding

an endoplasmic reticulum retention signal at the COOH-terminal (Smith, 1992). PGHS-1

is constitutively expressed in most tissues. It is linked to basal synthesis ofprostanoids

during cell differentiation and it is not suppressed by corticosteroid hormones (DeWitt,

1991). Recently, however, it has been shown that PGHS-1 can be slightly induced,

contributing to the up-regulation of prostanoid biosynthesis during monocyte

differentiation (Hoff et al. 1993). The other isoenzyme, PGHS-2, is an inducible, transient

form of PGHS. PGHS-2 was originally isolated as a v-src-inducible gene product in

chicken fibroblasts (Simmons et al. 1991; Xie et al. 1991) and as a phorbol ester-inducible

immediate early gene product called TIS10 in murine 3T3 cells (Kujubu et al. 1991). It

can be induced by a variety of stimuli such as mitogens, growth factors (Kujubu and

Herschman, 1992), inflammatory agents (Lee et al. 1992), hormones (Wong and Richards,

1992) and physiological activity (Yamagata et al. 1993), suggesting its involvement in

physiological responses. PGHS-2 contains an 18-amino acid insert very near the COOH-

terminal of the enzyme, which is an insert that is not present in PGHS-1. The remainder

of the sequence of PGHS-2 is quite similar to that of PGHS-1 (Smith, 1992). PGHS-2










mRNA is detectable in brain, prostate, and lung, but it is in very low or undetectable

abundance in kidney and most other major organs (Simmons et al. 1991).

The two PGHS isoforms differ in tissue distribution. Breder et al. (1992; 1995)

have demonstrated that the distribution of PGHS-2 in the central nervous system was

quite different from PGHS-1. One study demonstrated that PGHS-2 is the predominant

form of prostaglandin endoperoxide synthase in the brain of newborn piglet (Peri et al.

1995). PGHS-1 and PGHS-2 mRNAs are expressed equally in most human tissues except

that PGHS-2 mRNA has a higher level in lung tissues (O'Neill and Ford-Hutchinson,

1993).

Prostaglandin Receptors


The diversity of the effects of prostanoids is explained by the existence of a

number of distinct receptors that mediate their actions. Prostanoids exert their effects

through G protein-linked receptors The receptors have been named for the natural

prostaglandin for which they have the greatest apparent affinity and have been divided into

five main types, designated DP (PGD2), FP (PGF2), IP (PGI2), TP (TxA2), and EP (PGE,)

(Campbell and Halushka, 1996). The EP receptors have been further subdivided into EP,

(responsible for smooth muscle contraction), EP2 (responsible for smooth muscle

relaxation), EP3, and EP,, based on physiological and molecular cloning formation

(Coleman et al. 1994; Toh et al. 1995). The prostanoid responses that have been

characterized most extensively at the cellular and molecular levels are the effects of PGE,

on the renal collecting duct tubule. In the cortical collecting tubule cells, PGE, has been










shown to act through pharmacologically distinct stimulatory (EP2) and inhibitory (EP3)

PGE receptors to stimulate and inhibit cAMP formation, respectively (Nakao et al. 1989;

Sonnenburg and Smith, 1988; Sonnenburg et al. 1990); PGE2, presumably acting via an

EP, receptor, has also been demonstrated to stimulate Ca2+ mobilization and protein

kinase C activity in the rabbit collecting tubule (Hebert et al. 1990). Recently, PGE2 was

found to inhibit Na'-K'-ATPase activity in MDCK cells (Cohen-Luria et al. 1992) via G-

protein mediated receptor. The thromboxane A2 effects on Ca2+ mobilization is through

the EP, receptor.

Inhibition of PGHS


Nonsteroidal anti-inflamatory drugs inhibit the cyclooxygenase activity ofPGHS

without inhibiting peroxidase activity (Mizuno et al. 1982). Indomethacin, fluribuprofen,

and meclofenamic acid initially bind to the enzyme reversibly, competing with

arachidonate for the fatty acid binding site (Kulmacz and Lands, 1985; Rome and Lands,

1975). Eventually, a stable one-to-one complex forms between these inhibitors and the 70

kDa subunit (Kulmacz and Lands, 1985; Rome and Lands, 1975). The complex of apo-

PGHS and inhibitor exhibits 40-10% of the initial cyclooxygenase activity upon

reconstitution with heme (Kulmacz and Lands, 1985). Before reconstitution, this complex

is resistant to proteolytic digestion (Kulmacz, 1989). The complex formation which yields

protease resistance forms at a dose-response similar to that required for cyclooxygenase

inhibition, indicating that the complex formation resulted in both cyclooxygenase inhibition

and protease resistance (Kulmacz, 1989). By preincubating PGHS with both heme and










indomethacin, the synthase becomes even more resistant to proteolysis than when it is

preincubated with either compound alone. Since it is unlikely that both of these

compounds directly block the cleavage area, one of them must invoke a conformational

change which makes the cleavage site less accessible (Kulmacz, 1989).

Reversible cyclooxygenase inhibitors of PGHS include ibuprofen and flufenamate.

These compounds also make apo-PGHS more resistant to proteases (Kulmacz, 1989).

Aspirin inhibits PGHS's cyclooxygenase activity by transferring its acetyl group to Ser530

(Roth et al. 1975). Site-directed mutagenesis demonstrates that acetylation of Ser53

inhibits cyclooxygenase activity by placing a bulky group in a position that prevents

substrate binding at the cyclooxygenase active site (DeWitt et al. 1990).


Inactivation of PGHS


In the conventional model, enzyme reacts with substrate resulting in product

formation and the release of enzyme for further turnover. A "suicide" enzyme partitions

substrate into an alternate process leading to inactivation and eventual depletion of

enzyme (Walsh, 1984; Rando, 1984; Fersht, 1985). PGHS undergoes "suicide"

inactivation during catalysis (Smith and Lands, 1972; Smith and Lands, 1971). Both

cyclooxygenase and peroxidase substrates inactivate PGHS (Smith and Lands, 1972).

Incubation of cyclooxygenase with arachidonic acid results in loss of cyclooxygenase

activity but not peroxidase activity, concurrent with PGH2 production. The chemical

basis for inactivation is unknown. Both cyclooxygenase and peroxidase activities of

PGHS are diminished by hydroperoxides (Hemler and Lands, 1980a). Degradation or










modification of the heme prosthetic group could account for inactivation (Hemler and

Lands, 1980b; Chen et al. 1987). It is important to note that inactivation of PGHS has

been demonstrated in vivo as well as in vitro, indicating that it may indeed be a

physiologically relevant process (Lapetina and Cuatrecases, 1979; Kent et al. 1983).

Balancing inactivation is the gene level expression of the protein. Two general

types of regulation have been studied using cultured cells as model systems. The steady-

state level of PGHS in some cells is altered by agents such as PDGF and IL-1.

Prostaglandin synthesis in these systems increased dramatically following stimulation by

various compounds. The mechanism for this increase is presumably increased PGHS

synthesis. In some instances increased prostaglandin synthesis has not been associated

with increased enzyme mass.

The "suicide" inactivation appears to be a function of the cyclooxygenase reaction

which only destroys the enzyme's cyclooxygenase capability. The rate of "suicide"

inactivation proceeds linearly with cyclooxygenase substrate consumption and the rate

differs with the fatty acid used (Smith and Lands, 1972). Therefore, self-destruction of

PGHS is viewed to be intrinsic to the cyclooxygenase activity and dependent on its

reaction intermediates (Hemler and Lands, 1980a).

Limited digestion of PGHS into fragments of 33 kD and 38 kD also has diminished

peroxidase activity whereas the cyclooxygenase activity persists (Chen et al. 1987; Raz

and Needleman, 1990). The clipped enzyme retains its molecular mass of 140 kD in

solution and can still be acetylated by aspirin (Chen et al. 1987; Chen and Marnett, 1989).

Thus, regulation of prostaglandin biosynthesis is complex. The balance of genesis










and "suicide" for metabolic enzymes in the arachidonic acid cascade may play a critical

role in the physiology and pathophysiology of eicosanoids.

Stimulation ( growth factors, cytokines, hormones)

Increased transcription and translation of biosynthetic enzymes

Increased prostaglandin formation

Enzyme "suicide" inactivation

Depletion of enzymatic activity.


Thromboxane Synthase


The biosynthesis of TxA2 involves arachidonic acid liberation, prostaglandin

endoperoxide formation through cyclooxygenase, and isomerization ofPGH2 under the

catalysis of thromboxane synthase. In normal condition free intracellular arachidonic acid

level is very low. Therefore, the rate limiting step of TxA2 biosynthesis is the liberation of

arachidonic acid. However, two additional processes have been identified with regulatory

potential. These include the genesis of enzymes necessary for thromboxane formation and

the "suicide" inactivation ofthromboxane synthase during catalysis, which conforms to the

criteria for mechanism-based inactivation. The enzyme heme prosthetic group appears to

be affected. Enzyme purified from porcine lung and platelets is "suicide" inactivated by its

physiological substrate PGH2 and hydroperoxy fatty acids (Smith and Lands, 1972; Hall et

al. 1986; Jones and Fitzpatrick, 1991).

TxA2 plays an important role in several normal physiological processes. It is

required for normal platelet aggregation which forms clots to stop bleeding. Inhibition of










TxA2 formation by aspirin or blocking of TxA2 action by its antagonist cause prolonged

bleeding time. In addition to platelets, TxA2 is found in many other tissues such as lung,

spleen, brain, kidney, lymphocytes and macrophages (Hamberg, 1976; Davidson et al.

1980; Goldstein et al. 1978; Murota et al. 1978; Weidmann et al. 1978; Wolfe et al. 1976;

Zenser et al. 1977). Several studies indicated that coronary circulation was affected by

TxA2 (Gewirtz et al. 1985; Brezinski et al. 1986; Smith et al. 1980), suggesting that TxA2

might be the mediator responsible for endothelin action on blood vessels. TxA2 may also

participate in neurotransmission. Trachte (1986) reported that U46619, a TxA2 agonist,

potentiated both adrenergic force generation and norepinephrine generation in rabbit vas

deferens. Further, this effect could be antagonized by SQ29548, which is a well known

TxA2 receptor antagonist. Smooth muscles of many tissues have been reported to be

constricted in response to TxA2 action. TxA2 is very unstable with half life of 30 seconds

in physiological solution. It is rapidly hydrolyzed into the chemically stable and

biologically inactive product TxB2. In addition, TxA2 may also be inactivated in the lung

by prostaglandin dehydrogenase to produce 15-keto-TxA2 which is then quickly

hydrolyzed to 15-keto-TxB2. Therefore, measurement of TxB2 in vitro provides a reliable

indication of thromboxane synthase activity or in vivo TxA2 biosynthesis. Due to the

unstable nature of TxA2, much of the physiological properties of this compound have been

inferred from synthetic mimetics, enzyme inhibitors and receptor antagonists (Campbell

and Halushka, 1996).

Thromboxane synthase was first found in platelet microsomes, later in many other

tissues including lung, liver, kidney, brain, heart and placenta. The enzyme has been










purified to homogeneity from lung and platelets of different species, and found to be a

monomeric 53-59 kD hemoprotein. Thromboxane synthase contain a heme prosthetic

group and has been classified as a cytochrome P450 protein (Haurand and Ullrich, 1985).

Western blot studies indicated that the immunoreactive enzyme varied in molecular weight

from different tissues which has been shown to result from a covalent modification by its

substrate.


Summary


The above background provides a basis for studying the neuromodulatory role of

prostanoids on hormonal and cardiovascular responses to cerebral hypoperfusion. We

know that carotid, aortic, and cardiac afferent fibers are not the sole mediators of

endocrine and cardiovascular responses to severe hypotension in the fetus. We also know

that the fetus responds to reduced cerebral perfusion with reflex increases in ACTH and

AVP secretions and redistribution of combined ventricular output. We believe that it is

likely that prostanoids mediate at least a portion of this response. It is possible that

prostaglandin Ez and thromboxane A, are produced locally to adjust cerebral blood flow,

and these prostanoids stimulate hormonal responses and autonomic efferent activity via

paracrine/autocrine fashion. The prostanoids could be synthesized in the vasculature or in

the neurons of the endocrine and cardiovascular control centers within the central nervous

system. We would find that both PGE2 and thromboxane A2 produced in the fetal brain

stimulate ACTH and AVP secretions and stimulate the cardiovascular compensations to

cerebral hypoperfusion.












CELL MEMBRANE

PHOSPHOUPID
ACTIVATION OF
PHPHOHOUPASES /


cSzc"
\ "


ARACHIDOMC ACID (INHIBITATED BY
CCLOOXY INDOMETHACIN)



SHYROPERSYNTHASE
; 6W OEQIDS


PGG2


\/ |COOH
/-\ iS' coo.^. ^
y v v\ SyNTH


SYNTHASE


PGDz


"hAHJ


0.
P


w w,
PGEI HO


.-'.


P6H
PGF2,


I 5 I

Figure 2-1. Pathways for the formation of various prostanoids. Although all products of
cyclooxygenase pathway are shown, usually only one prostanoid is formed as a major
product by a given cell type. PGG2, PGH2, PGE2, PGD2, PGF2,, prostaglandin G2, H2, E2,
D2, F2, respectively; PGI2, prostacyclin; TxA2, thromboxane A2. X, indomethacin inhibits
cyclooxygenase activity. Adapted from Smith (1989).


OW


/


TxAZ














CHAPTER 3
ENDOGENOUS PROSTANOIDS MODULATE THE ACTH AND AVP RESPONSES
TO HYPOTENSION IN LATE-GESTATION FETAL SHEEP


Introduction


Prostanoids, including thromboxane A2, prostaglandin E2 and prostaglandin I,,

are well-known as locally generated mediators of vascular tone within the cerebral

vasculature (Campbell, 1996; Heistad and Kontos, 1983, Pickard et al. 1977). We have

hypothesized that, in addition to their role in the local control of cardiovascular function,

prostanoids partially mediate hormonal and cardiovascular reflex responses to arterial

hypotension in late-gestation fetal sheep. This hypothesis is based on the following

observations: 1) arterial barcreceptors and chemoreceptors partially mediate the reflex

adrenocorticotropin (ACTH) and arginine vasopressin (AVP) responses to hypotension

(Wood, 1989); 2) prostaglandin E, (PGE2) and thromboxane A, (TxA2), when infused into

the blood perfusing the brain, stimulate endocrine and cardiovascular responses

reminiscent of reflex responses to hypotension (Cudd and Wood, 1992; Cudd and Wood,

1993); 3) hypotension stimulates the local synthesis and release of PGE, and TxA2 in the

cerebral vasculature of the newborn piglet, and therefore might have a similar effect in the

late-gestation fetal sheep (Chemtob et al. 1990b).











We designed the present study to test the hypothesis that moderate hypotension,

reduction of fetal arterial blood pressure approximately 50% below normally controlled

levels, stimulates endocrine responses which are mediated, in part, by endogenously

generated prostanoids and, in part, by chemoreceptors and baroreceptors in the carotid

sinus region and in areas innervated by afferent fibers in the vagosympathetic trunks.

Specifically, this study was designed to identify the effects of denervation alone and the

effects of prostaglandin synthesis blockade alone, and to identify any interactions between

these two systems.


Materials and Methods


Fifteen chronically catheterized fetal sheep between 124 and 136 days' gestation

were used in this study. The pregnant ewes were of mixed Western breeds. These

experiments were approved by the University of Florida Institutional Animal Care and Use

Committee and were performed in accordance with the Guiding Principles for the Care

and Use of Animals of the American Physiological Society.


Surgical Preparation


Aseptic surgery was performed at least 5 days before the start of experiments in

each animal. Nine fetal sheep were subjected to catheterization only. Six fetal sheep were

subjected to catheterization plus denervation of arterial baroreceptors and

chemoreceptors. Fetal hindlimbs were identified and delivered through a small

hysterotomy incision near the tip of one uterine horn. As previously described, we










introduced polyvinyl chloride catheters into the tibial artery (.050" ID; .090" OD) and

saphenous vein (.030"ID, .050" OD) bilaterally and advanced the tips to the abdominal

aorta and inferior vena cava, respectively. After closing the skin incisions in the fetal

hindlimb, amniotic fluid catheters were sutured to the fetal skin and the hindlimb returned

to the amniotic space. After placement of these vascular and amniotic fluid catheters and

closure of the hysterotomy, the fetal head was located, the uterus was incised, and the

head was delivered. After a single midline incision in the skin of the neck was made,

lingual arteries were identified, ligated, and catheterized with polyvinyl chloride catheters

(0.030 in ID, 0.050 in. OD), with the catheter tips advanced retrograde to the lumen of the

common carotid arteries. As previously described, this catheterization technique allows

measurement of common carotid arterial pressure without interruption of carotid arterial

blood flow (Breuhaus and Chimoskey, 1985; Wood et al. 1993). In fetuses subjected to

baro- and chemo-denervation, the common carotid arteries and cervical vagosympathetic

trunks were carefully exposed on both sides. The vagosympathetic trunks were cut

bilaterally. The carotid sinus nerves were identified bilaterally and cut. The walls of the

common carotid arteries in this area, as well as the lingual arteries and common carotid

arteries extending 0.5-1 cm rostral from the lingual-carotid arterial junction were stripped

of all visible nerve fibers. The carotid sinus denervation procedure used in this

investigation was shown in a previous study to block the changes in fetal heart rate during

hypoxia (Wood et al. 1990). Standard tests of the completeness of denervation were not

possible in the present experiments, because the severing of cervical vagosympathetic

trunks interrupts the efferent limb of the reflex decrease in heart rate that would be










observed after phenylephrine injection (Wood, 1989; Wood and Rudolph, 1983). After

performing these denervations in the fetal neck, the fetal skin was closed, the head was

returned to the amniotic cavity, and the uterus was closed.

In some fetuses, we introduced a Swan-Ganz catheter (size 5F) into the saphenous

vein of one hindlimb and advanced the tip to the supradiaphragmatic inferior vena cava.

In other fetuses, we placed an extravascular balloon occluder around the

supradiaphragmatic inferior vena cava using a separate incision through the fourth

intercostal space. Extravascular balloon occluders were either purchased (In Vivo Metric,

Healdsburg, CA) or fabricated in the laboratory. Occluders fabricated in the laboratory

were made from Penrose tubing (0.5 cm diameter), looped around the vena cava (between

the diaphragm and right atrium) and connected to polyvinyl chloride tubing. Results from

fetuses with intravascular versus extravascular balloon occluders were not distinguishable

and therefore were combined. All catheters exited via a small incision in the flank of the

ewes.

Ampicillin trihydrate (Polyflex; Aveco Co, Fort Dodge, IA) 500 mg was

administered to the fetus via the amniotic fluid and to the mother (500-750 mg)

intramuscularly at the time of surgery and again each time the fetus was studied or the

catheters were flushed. Ampicillin (500-750 mg) was administered to the mother

intramuscularly twice daily for five days after surgery. All catheters were flushed and

reheparinized at least once every 3 days.










Experimental Protocol



Indo or PB

VCO


t tttt
-90 -10 0 +10 +20



Figure 3-1. Experimental design. In each experiment, indomethacin (Indo) or phosphate
buffer (PB) was injected intravenously 90 min before the start of vena caval obstruction
(VCO). VCO was achieved for a period of 10 min (between 0 min and 10 min). Fetal
blood samples were drawn at -90 min (prior to injection of Indo or PB), at -10 min (prior
to the start of the VCO), at 10 min (the end of VCO), and at 20 min (10 min after the end
of VCO).

Sheep were transported to the procedure room from their pens within the Health

Center Animal Resources Department at least 1 hour before the start of each experiment.

Each fetus was studied twice. Experiments consisted of a 90 min pre-occlusion control

period (-90 to 0 min), a 10 min occlusion period (0 to 10 min) and a 10 min post-

occlusion recovery period (10 to 20 min) (Figure 3-1). In one experiment on each fetus,

the vehicle for indomethacin (0.1 M phosphate buffer) was injected intravenously, and in

the other experiment 0.2 mg/kg indomethacin (an inhibitor of cyclooxygenase) was

injected intravenously 90 min prior to the 10 min period of hypotension. In each

experiment, the intravascular or extravascular occluder was completely inflated for 10 min

to produce arterial hypotension, which blood pressure was decreased by 50% of normal

level.











Blood Sample Handling


Fetal arterial blood samples were drawn from tibial artery catheters (descending

aorta) at 90 and 10 min before the start of the period of occlusion, at the end of the 10 min

period of occlusion, and 20 min after the start of occlusion. Blood samples (3 ml) were

collected into chilled polystyrene tubes containing 150 pl of 0.5 M EDTA. Separate

blood samples (1 ml) were collected into chilled polypropylene tubes containing 50 Pl 0.5

M EDTA and 40 pg/ml indomethacin for measuring TxB2, 6-keto-PGF1 and PGE2.

Tubes were kept on ice until the end of the experiment and then centrifuged for 20 min at

3000 x g at 4 C. Plasma was separated and stored in separate aliquots at -20 o C.


Radioimmunoassav (RIA) for Hormone Assays


Plasma adrenocorticotropic hormone (ACTH), cortisol, and arginine vasopressin

(AVP) concentrations were measured by specific radioimmunoassay. In ACTH and AVP

assays, all plasma samples were run in one assay, and the standard was extracted with

plasma to correct for recovery. ACTH was measured using 25I-ACTH and rabbit anti-

ACTH antiserum produced in this laboratory. Before assay, ACTH was extracted from

plasma and standard with powdered glass (0.5 ml of 70 mg/ml slurry of glass in 0.5 M

phosphate buffer and 0.1% glass extracted bovine serum albumin; Corning Co., Corning,

NY). The glass was then washed with buffer and the ACTH eluted from the glass with

acid-acetone (1.0 N HCI : 1 acetone). The recovered ACTH (the recovery is about 70-

80%) was dried on an evaporator (Savant Instruments, Farmingdale, NY), then frozen at -










20C until assay. The extracts were reconstituted in 500 pl assay buffer (0.5%

mercaptoethanol in 0.5 M phosphate buffer with extracted BSA, pH 7.4). The standards

in this assay range from 20 to 800 pg/ml. Cortisol was measured using 3H-cortisol

purchased from Amersham Co (Arlington Heights, IL) and rabbit anti-cortisol antiserum

produced in Dr. Wood's laboratory. Before assay, cortisol was extracted from plasma

using 20 volumes of ethanol. AVP was measured using anti-AVP antiserum also

produced in Dr. Wood's laboratory and '25I-labeled AVP purchased from Amersham

(Arlington Heights, IL). Before assay, AVP was extracted from plasma on bentonite

(Sigma Co., St. Louis, MO) and eluted from bentonite with acid-acetone. The recovered

AVP was dried on an evaporator, then frozen at -20C until assay. The ACTH, cortisol,

and AVP assays have been completely described elsewhere (Raff et al. 1991; Wood et al.

1993).


Enzyme-Linked Immunoassay (EIA) for Prostanoids Assays


PGE2, thromboxane B2 (TxB2, a stable metabolite of TxA2), and 6-keto-

prostaglandin Fia( 6-keto-PGF,, a stable metabolite of PGI2) were measured using

enzyme-linked immunoassay kits purchased from Cayman Chemical. Before assay, the

prostanoids were extracted from acidified plasma with 6 volumes of ethyl acetate. The

recovery using this protocol averages approximately 60%, and the extracted prostanoids

diluted parallel to the standard curves. All plasma were run in one assay.











Blood Pressure Recording


Fetal arterial blood pressure and amniotic fluid pressure were measured

continuously during the 110 min experiments using a Grass recorder and Statham P23 ID

pressure transducers. Not all hemodynamic variables were successfully measured in all

experiments. These hemodynamic values were recorded, and analog-to-digital

conversions were performed at 2-s intervals using an IBM PC computer. The data

collection was achieved using ASYSTANT' software (Asyst Technologies, Rochester,

NY). All fetal intravascular pressure were corrected by subtraction of amniotic fluid

pressure. One minute means were analyzed statistically.


Statistical Analyses


Changes in the values of fetal hormonal and prostanoids variables over time and

between groups were analyzed using two-way analysis of variance (ANOVA) corrected

for unequal cell size and for repeated measures in one dimension, time (Winter, 1971).

Multiple comparison of mean values was performed using the Student-Newman-Keuls

test. The hormonal data were not distributed normally. All ANOVAs performed on

hormonal data were calculated after logarithmic transformation to correct for

heteroscedasity of the data. A significance level of P < 0.05 was used to reject the null

hypothesis in all tests. Analyses were performed using SigmaStat software (Jandel

Scientific, San Rafael, CA).











Results

Prostanoids

The mean values of fetal PGE2, TxB2, and 6-keto-PGF, concentrations are

reported in Table 3-1. In intact fetuses, hypotension significantly increased plasma

concentrations of PGE2 (when compared to the preocclusion value), but did not

significantly change plasma concentrations of the other measured prostanoids.

Indomethacin significantly decreased the plasma concentrations ofPGE2, 6-keto-PGFl,,

and TxB2 in response to vena caval occlusion (Table 3-1).

Table 3-1. Plasma concentrations of prostanoids.
Time (min) Intact Denervated

PB Indomethacin PB Indomethacin

-90 min PGE2 105.3 + 29.7 150.9 38.7 108.6 + 53.6 77.0 24.3
TxB2 116.8 31.7 169.8 39.1 62.3 14.3# 70.0 27.2
PGF,, 151.1 68.4 159.0 49.4 106.1 28.8 105.8 33.4

-10 min PGE2 158.3 13.8 40.1 4.5 98.9 29.2 62.8 27.7
TxB, 143.4 26.9 132.2 40.3 53.7 14.0# 48.5 16.7
PGFI, 196.8 37.8 95.0 29.2 100.6 26.2 100.8 31.0

10 min PGE, 235.5 35.2*t 35.1 4.4* 140.4 43.5 60.2 26.5
TxB2 173.3 46.3*# 105.1 33.1* 73.420.6# 34.6 11.9
PGFi, 275.3 45.0*# 112.0 32.4* 166.1 63.7# 134.0 40.2

20 min PGE2 298.5 86.6*t 34.0 5.9* 107.0 26.9 29.2 + 8.4
TxB2 205.5 73.3* 107.1 30.7* 136.8 55.1 34.9 11.1
PGF, 230.3 t77.9* 92.1 28.3* 129.1 41.3 97.3 27.5


Responses are grouped by the time of vena caval occlusion and intact or denervated and
indomethacin or phosphate buffer (PB) treatments. The periods of occlusion were from 0
to 10 min. Prostanoids concentrations were presented as pg/ml SE. indicates
statistical significance between the indomethacin and PB treatments groups at p < 0.05
(Student's Newman-Keuls). # indicates statistical significance between the intact and
denervated groups treated with PB at p < 0.05 (Student's Newman-Keuls). t indicates
statistical significance when compared within-group to the value at -10 min.









37
Plasma concentrations of the prostanoids were reduced by denervation of arterial

baroreceptors and chemoreceptors. In particular, plasma TxB2 concentrations were

significantly lower before and during hypotension in denervated fetuses treated with

phosphate buffer. Plasma PGFi, concentrations during hypotension were lower in

(phosphate buffer-treated) denervated fetuses than in intact fetuses. Indomethacin did not

produce further decreases in plasma prostanoid concentrations in the denervated fetuses.

Cardiovascular Variables


In the intact group (n = 9), vena caval occlusion significantly decreased mean

arterial blood pressure from 42.9 2.2 to 24.8 3.7 mm Hg (mean SEM) in the fetuses

treated with PB; mean arterial blood pressure decreased from 46.5 1.8 to 26.0 + 3.1 mm

Hg in fetuses treated with indomethacin (Fig. 3-2). The effect of the vena caval occlusion

on blood pressure was not significantly different in intact vs. denervated groups. The

values of fetal blood gases and pH are reported in the table 3-2. There are no statistical

differences between the intact and denervated groups and phosphate buffer and

indomethacin treatment groups.

Table 3-2. Fetal blood gases and pH before indomethacin or PB injections


Intact Denervated
PB indomethacin PB indomethacin
PaO2, mmHg 19.70 1.18 18.16 1.18 17.940.92 18.27 2.10

PaCO2, mmHg 51.51 1.81 45.71 1.94 47.43 1.71 49.88 2.34

pH 7.34 0.01 7.36 0.01 7.37 0.01 7.34 0.01

Values are means SE.










Endocrine Variables


The initial values ofACTH, AVP, and cortisol (before the injection of phosphate

buffer or indomethacin) were not significantly different between the intact and denervated

groups. Compared with phosphate buffer-treated fetuses in the intact group,

indomethacin treatment significantly attenuated the fetal plasma ACTH and AVP reflex

responses to vena cava occlusion (Fig. 3-3). Vena caval occlusion increased fetal plasma

ACTH concentration significantly from 83 39 to 3611 774 pg/ml, AVP from 4 0.5

to 1079 549 pg/ml, and cortisol from 5 1 to 10 2 ng/ml in fetuses treated with

phosphate buffer in intact group; in the indomethacin treatment fetuses, ACTH increased

significantly from 156 63 to 1922 674 pg/ml, AVP increased significantly from 5 1

to 226 181 pg/ml, and cortisol from 82 to 10 3 ng/ml. After conversion of data to

common logarithms, analysis by ANOVA indicated that indomethacin attenuated the

ACTH, AVP, and cortisol responses to the vena caval occlusion (significant interaction of

time x group in the two-way ANOVA corrected for unequal cell size and for repeated

measures in one dimension, time).

The ACTH, AVP, and cortisol responses to the vena caval occlusion were also

significantly attenuated in the phosphate buffer-treated denervated fetuses compared to

phosphate buffer-treated intact fetuses (significant main effect of group with significant

interaction of time x group in the two-way ANOVA). Vena caval occlusion increased

fetal plasma ACTH concentration from 201 77 to 456 + 135 pg/ml, and AVP from 4 +

0.5 to 16 6 pg/ml in the phosphate buffer-treated denervated fetuses. In contrast to its









39
effect in intact fetuses, indomethacin had no further effect on the ACTH, AVP, or cortisol

responses to vena caval occlusion in the denervated fetuses (no significant main effect of

group and time x group interaction). After indomethacin treatment, vena caval occlusion

increased plasma ACTH concentration from 282 66 to 1246 560 pg/ml and AVP from

8 3 to 17 6 pg/ml in the denervated fetuses.


Discussion


The results of this study confirm and substantially extend our previous experiments

designed to test the mechanism of the hormonal responses to hypotension in the fetus.

First, these results confirm our earlier observations that vena caval occlusion is a stimulus

to fetal ACTH, AVP, and cortisol secretion (Wood, 1995b), and that the ACTH and AVP

responses are attenuated by arterial baroreceptor and chemoreceptor denervation (Wood,

1989). These results therefore support our earlier conclusions that sinoaortic afferent

fibers partially mediate these responses to hypotension. These data extend our earlier

observations by demonstrating that the baro- and/or chemo-reflex control of ACTH and

AVP secretion is partially dependent upon the endogenous production of prostanoids.

There are several important implications of these results which we will address

individually.

ACTH and AVP Responses to Hypotension in the Sheep Fetus Are Not Completely
Mediated by Carotid Sinus and Vagal Afferent Fibers


In previous studies, we have demonstrated that carotid sinus denervation

attenuated the ACTH and AVP responses to systemic hypotension (produced by vena










caval obstruction) or to carotid arterial hypotension (produced by bilateral carotid

occlusion). Sinoaortic denervation impairs the ability of the fetus to maintain arterial

blood pressure during progressive hemorrhage (Chen and Wood, 1992). On the other

hand, bilateral cervical vagosympathetic nerve section has no apparent effect on the reflex

hormonal responses to hemorrhage or on the ability of the fetus to maintain blood pressure

when blood volume is reduced (Wood et al. 1989). We therefore strongly suspected that

the carotid sinus afferent fibers were more important for reflex hormonal and

cardiovascular responsiveness than afferent fibers in the vagosympathetic trunks carrying

afferent information from cardiopulmonary receptors. It was possible, although perhaps

not likely, that the ACTH and AVP responses to hypotension would be completely

eliminated by total (carotid sinus and vagosympathetic nerve section) denervation. The

present results prove that the ACTH and AVP responses to hypotension are not

completely mediated by neural afferent fibers from these regions.

In the present study, we measured arterial blood gases only at the beginning of

each experiment. In a previous study, we found that vena caval obstruction produced no

statistically significant changes in fetal PaO2 or PaCO2, but did produce moderate decreases

in pH, (approximately 0.04 pH units) (Wood et al. 1982). In another study, we found that

such small changes in pH,, on their own, were not sufficient to stimulate either ACTH or

vasopressin secretion (Wood and Chen, 1989). Finally, we found that sinoaortic

denervation did not alter this pattern of decreasing pH, and unchanging PaO2 or PaCO2

(Wood, 1989). For these reasons, we do not think that the responses that we observe

during vena caval obstruction are caused by changes in arterial blood gases, or that baro-










and chemo-denervation alters arterial blood gases during this manipulation. Increases in

plasma hormone concentrations can also be caused by hypotension-induced decreases in

hormone clearance from plasma. While this can be a complicating factor, even dramatic

changes in clearance cannot account for the large changes in plasma concentrations that

we have measured.



ACTH and AVP Responses to Hypotension Are Mediated, in Part, by Endogenous
Prostanoids



An important result of the present study is the demonstration that the ACTH and

AVP responses to hypotension are partially inhibited by indomethacin in intact fetuses.

We and others have previously demonstrated that exogenous PGE2 and TxA2-mimetic

(U46619) stimulates ACTH secretion (Brooks, 1992; Brooks and Gibson, 1992; Cudd

and Wood, 1991; Cudd and Wood, 1993; Wood et al. 1993). It is also well known that in

newborn piglets the endogenous production ofPGE2, PGI2, and TxA2 is stimulated during

induced reductions in cerebral blood flow (Chemtob et al. 1990b). These prostanoid

responses are thought to be important for the autoregulation of blood flow when perfusion

pressure is changed (Chemtob et al. 1990a). Although the prostanoid generation and

release into venous blood is often thought of as originating in the vasculature, it is known

that the neurons contain the prostaglandin G/H synthase (Breder et al. 1992) and

thromboxane synthase (Husted and Wood, 1998). We designed the present experiments

to test the hypothesis that endogenous production of prostanoids, whether in vasculature









42

or in neurons, might partially mediate the reflex hormonal responses to hypotension. The

results indicate that this hypothesis was correct.

The results reveal an interaction between the afferent fibers and the endogenous

prostanoids. Indomethacin had a statistically significant effect only in the intact fetuses.

The ACTH and AVP responses were attenuated by the denervation procedure, but were

not further attenuated by indomethacin. This suggests that the action of the prostanoids

must be on the afferent pathways or on a central element which is activated by the afferent

pathways. In other words, this selective action appears to rule out an effect solely on the

anterior or posterior pituitary or on the nerve terminals of the median eminence, because

an action on the so-called "final common pathway" would attenuate the response in both

groups. The results therefore demonstrate that indomethacin is not simply a "nonspecific"

inhibitor of ACTH and AVP secretion.

We and others have found that exogenous prostanoids may stimulate hormonal

and hemodynamic responses by affecting the neuronal processing within the central

nervous system. These studies have focused mainly on PGE2 and thromboxane A2 (TxA2).

Infusions of PGE2 into the carotid arteries of conscious adult sheep increase heart rate and

blood pressure (Breuhaus and Chimoskey, 1983; Breuhaus and Chimoskey, 1985;

Breuhaus et al. 1989) and ACTH secretion (Cudd and Wood, 1991). This effect is not

attenuated by carotid sinus baro- or chemo-denervation (Cudd and Wood, 1993) and

therefore is likely to be a direct effect on the brain. Infusions of PGE, also stimulate

increases in fetal blood pressure and heart rate (Cudd and Wood, 1992; Cudd and Wood,

1993). Intracerebroventricular (Brooks and Gibson, 1992) or intravenous (Young and











Thorburn, 1994) infusion of PGE2 can stimulate ACTH and cortisol secretion in fetal

sheep, and treatment with indomethacin decreases ACTH release (Thompson and Hedge,

1978). In addition to its effect on the brain, PGE2 has a direct effect on the fetal sheep

pituitary gland by enhancing AVP-stimulated, but not CRH-stimulated, ACTH secretion

from dispersed fetal anterior pituitary cells in culture (Brooks, 1992). TxA2 stimulates

increases in arterial blood pressure and heart rate and in the rate of ACTH secretion in

both adult and fetal sheep (Cudd, 1997).

The influence ofprostanoids on AVP secretion has been investigated in adult dogs

and rats and in newborn piglets. Intraventricular administration ofPGE2 (Hoffman et al.

1982; Yamamoto et al. 1976) or PGD2 (Brooks et al. 1986) stimulates AVP secretion.

Indeed, the stimulation of AVP secretion in response to intracerebroventricular injections

of angiotensin II or acetylcholine is attenuated by indomethacin (Inoue et al. 1990; Inoue

et al. 1991), suggesting that central pathways affecting AVP secretion by hypothalamic

magnocellular neurons are dependent upon prostanoid generation. AVP responses to

hypertonic saline are significantly inhibited by prior administration ofmeclofenamate,

indomethacin, or offlunixin meglumin (Brooks et al. 1986; Hoffman et al. 1982; Patel et

al. 1997). The effect of endogenous prostanoids on the AVP responses to hemorrhage

were somewhat more variable. Meclofenamate, but not indomethacin, attenuated the

AVP response to hemorrhage in pentobarbital-anesthetized dogs (Brooks et al. 1984).

Indomethacin augmented the AVP response to hemorrhage in morphine sedated and

urethane-chloralose-anesthetized dogs (Brooks et al. 1984). These interesting

experiments suggest that there might be a differential effect of indomethacin and









44
meclofenamate on the relative degrees of inhibition of the prostanoid biosynthetic enzymes

(cyclooxygenase, thromboxane synthase, etc.). Or, perhaps, there might be an effect of

anesthesia on the degree to which prostanoids influence AVP secretion.

We measured plasma concentrations of PGE2, 6-keto-PGF,, and TxB2 before and

after indomethacin injection to confirm our assumption that the production ofprostanoids

would be reduced. Overall, indomethacin did effectively reduce the plasma concentrations

of all three prostanoids, although the effect was more pronounced in the intact fetuses.

The denervated fetuses had lower plasma prostanoid concentrations initially, and

suppression was therefore less pronounced. Although we used plasma concentrations as

an index of overall prostanoid production, it is important to emphasize that these

compounds are most likely not acting as hormones, and that the local concentrations

within the brain were unaccessible to us with the present experimental design. We are

confident that the circulating (plasma) concentrations of PGE2, and probably other

prostanoids as well, are too low to stimulate ACTH secretion. Cudd and Wood have

found that high rate infusions of PGE2 into the carotid artery stimulate ACTH secretion in

both fetal and adult sheep (Cudd and Wood, 1991; Cudd and Wood, 1992). However,

when PGE2 is infused at rates which increase carotid arterial plasma concentrations within

the physiological range, no effect on ACTH was measured (Cudd and Wood, 1991; Cudd

and Wood, 1992). We believe that the response to TxA2 is analogous. We demonstrated

that infusions of U46619 (a stable TxA2 mimetic) into the carotid arterial blood of fetal

sheep stimulates ACTH secretion, whereas similar infusions into the venous blood (and

therefore diluted by the entire cardiac output) does not (Wood et al. 1993). TxA2











therefore acts at the brain to stimulate ACTH secretion. Because we used a stable

mimetic drug (with a relatively long half-disappearance time in blood), we could not assess

the physiological significance of the concentrations of TxA2 naturally occurring in plasma.

However, TxA2 is very unstable in aqueous solution and would be unlikely to reach the

brain in physiologically meaningful amounts.

One interesting result of the present experiments (although probably not related to

the ACTH and AVP secretion) is the increase in plasma prostanoid concentrations during

the period of hypotension. This release of prostanoids into the plasma is reminiscent of

the PGE2 and PGI2 production by the umbilical-placental vasculature in response to

angiotensin II (Yoshimura et al. 1990). Indeed, it is impossible to identify the source of

these circulating prostanoids: they could originate from any of the systemic or pulmonary

vascular beds. Nevertheless, it is tempting to speculate that the prostanoid release into

plasma is secondary to the reflex vasoconstriction during hypotension. If so, the reduction

in prostanoid concentration in the denervated animals would be explained by a lower level

of vasomotor tone in these animals (Jacob et al. 1991).

In conclusion, the results of this study demonstrate that the endogenous

production ofprostanoids modulates the reflex hormonal responses to hypotension in late-

gestation fetal sheep. We speculate that the prostanoids are generated either on the

afferent nerves subserving arterial baroreceptors and chemoreceptors, or that they are

generated within the nucleus of the tractus solitarius or other sites within the central

nervous system receiving afferent input from baroreceptors and chemoreceptors.
















60 -

50 -

40 -

30 -

20 -

10 -

0




60

50

40

30


20 -

10 -

0


Intact


Before


SAD


During


After


EI: PB
EzZ2 INDO


Before During


After


Time relative to hypotension




Figure 3-2. Fetal mean arterial blood pressure (MAP) before, during and after a 10-min
period of vena caval obstructions in the intact (top panel) and denervated (bottom panel)
fetuses. There were no significant differences of MAP between intact and denervated
fetuses or between phosphate buffer or indomethacin treated groups. Mean values are
accompanied by vertical bars representing 1 SEM. *p < 0.05 indicates that VCO
significantly decreased fetal MAP.


I IIL1 I I-II I r I n
















SAD


T



ti t t


p_ a
I I I


25

20

15 -




0 T

-10 -5 0 5 10 15 20


25 -

20 -E

15 -
10 -

5
0 I I
-10 -5 0 5 10 15 20


Time (minutes from onset of occlusion)



Figure 3-3. Fetal plasma ACTH, AVP and cortisol concentrations before, during, and after
a 10-min period of vena caval obstruction in the intact and denervated (SAD) fetuses. The
period of hypotension is from 0 min to 10 min. Mean values of hormones in phosphate
buffer- (1, dashed lines) and indomethacin- (@, solid lines) treated fetuses are
accompanied by vertical bars representing 1 SEM. Two-way ANOVA revealed a
significant interaction of time x treatment (p < 0.05) in the intact but not in the
denervated fetuses.


INTACT






/ T \


3200

2400

1600

800

0 -


3200

2400 -

1600 -

800 -

0 -


--- INDO
-- PB


1000 -
800 -
600 -
400 -
200 -
0 -


1000 -
800 -
600 -
400
200 -
0 -


I \;














CHAPTER 4
PROSTAGLANDIN ENDOPEROXIDE SYNTHASE-2 IS THE MAJOR
CONTRIBUTOR TO PROSTANOID BIOSYNTHESIS DURING CEREBRAL
HYPOPERFUSION IN LATE-GESTATION FETAL BRAIN TISSUES


Introduction


Prostaglandin endoperoxide synthase (PGHS) catalyzes the oxygenation

(cyclooxygenase) of arachidonate to the prostaglandin endoperoxide PGG2 and the

reduction (peroxidase) of PGG2 to PGH2. PGH2 is further converted by specific

synthases (isomerases) or reductases to the major biologically active prostanoids PGD2,

PGE2, PGF2,, PG1, and thromboxane A2 (TxA2) (Smith et al. 1991). A unique feature of

PGHS is the autoinactivation of the enzyme following the conversion of PGG2 to PGH2

(Hemler and Lands, 1980b).

Two different isoforms of PGHS were confirmed at the protein (Habib et al. 1993)

and mRNA levels (O'Neill and Ford-Hutchinson, 1993). There are two distinct genes

which encode for PGHS-1 and PGHS-2. PGHS-1 is the isoenzyme which was first

purified from bovine and ovine microsomes and is a homodimer consisting of two subunits

each of approximately 70 kilodaltons (Samuelesson et al. 1978). PGHS-1 is constitutively

expressed in most tissues. It is linked to basal synthesis of prostanoids during cell

differentiation and it is not suppressed by corticosteroid hormones (DeWitt, 1991).

Recently, however, it has been shown that PGHS-1 can be slightly induced, contributing









49
to the up-regulation of prostanoid biosynthesis during monocyte differentiation (Hoff et al.

1993). The other isoenzyme, PGHS-2, is an inducible, transient form of PGHS. It can be

induced by a variety of stimuli such as mitogens, growth factors (Kujubu et al. 1992),

inflammatory agents (Lee et al. 1992), hormones (Wong and Richards, 1992) and

physiological activity (Ymagata et al. 1993), suggesting its involvement in physiological

responses.

The two PGHS isoforms differ in tissue distribution. Breder et al. (1992, 1995)

have demonstrated that the distribution of PGHS-2 in the central nervous system was

quite different from PGHS-1. One study demonstrated that PGHS-2 is the predominant

form ofprostaglandin endoperoxide synthase in the brain of newborn piglet (Peri et al.

1995). PGHS-1 and PGHS-2 mRNAs are expressed equally in most human tissues except

that PGHS-2 mRNA has a higher level in lung tissues (O'Neill and Ford-Hutchinson,

1993).

Thromboxane synthase catalyzes the formation ofthromboxane A2 from

prostaglandin H2. The enzyme has been purified from lung and platelets. Western blot

studies indicated that the immunoreactive thromboxane synthase molecular weight ranges

from 46 to 130 KD (Wood et al. 1997), depending on the source tissue and analytical

method for determination. This enzyme is widely distributed in the brain (Wood et al.

1997).

Prostaglandins are local hormones which are synthesized by virtually all

mammalian tissues (Smith, 1987) and act in an autocrine and/or paracrine fashion. These

bioactive lipids have important roles in regulating several neurological (Smith, 1987;










Hayaishi, 1991), cerebral hemodynamic functions (Wolfe and Coceani, 1979) and

cardiovascular activities (Siren, 1982; Chemtob et al. 1990b). Our previous studies

demonstrated that endogenous prostanoids had significant effects on modulating

adrenocorticotropic hormone (ACTH) and arginine vasopressin (AVP) secretions from

fetal sheep during moderate arterial hypotension (Tong et al. 1998). The present studies

were conducted to determine the mechanism by which cerebral hypoperfusion enhances

prostanoid secretion by fetal brain tissues. Specifically, we investigated the ability of

cerebral hypoperfusion to modify levels of the constitutive (PGHS-1), inducible (PGHS-2)

isoforms of PGHS, and thromboxane synthase in brain tissues.


Materials and Methods


Materials


Polyclonal antibodies specific to PGHS-1 and PGHS-2 were purchased from

Oxford Biomedical Research, Inc. (Oxford, MI). Polyclonal antibody specific to

thromboxane synthase, and arachidonate, PGH2 and PGE, TxB2 enzyme immunoassay kits

were purchased from Cayman Chemical (Ann Arbor, MI). Diethyldithiocarbamic acid

(DEDTC) were purchased from Sigma Chemical Co. (St. Louis, MO). Protein assay kits,

10% SDS-polyacrylarnide ready gels and electrophoretic reagents, nitrocellulose

membranes, and filter papers were purchased from Bio-Rad (Richmond, CA). Rainbow

molecular weight markers and horseradish peroxidase-conjugated anti-rabbit IgG

antibodies were purchased from Amersham (Arlington Heights, IL). Western blot

chemiluminescence reagent were purchased from DuPont NEN (Boston, MA).











Animals


These experiments were approved by the University of Florida Institutional

Animals Care and Use Committee and were performed in accordance with the Guiding

Principles for the Care and Use of Animals of the American Physiological Society. Time

dated pregnant ewes were of mixed Western breeds obtained from the Institute of Food

and Agricultural Sciences at University of Florida. Fetal sheep between 124 and 136 days'

were used in this study.

One group of fetuses (n = 5) was studied as intact (control) fetuses; the second

group of fetuses (n = 5) was studied after carotid sinus denervation (experimental group).

The experimental group was subjected to a 10-min period of cerebral hypoperfusion

produced by occlusion of the brachiocephalic artery, performed using an extravascular

occluder (In Vivo Metric, Healdsburg, CA).

The fetal sheep were anesthetized using 1.5% halothane. Lingual arteries were

catheterized bilaterally for measurement of blood pressure, blood gases, and plasma

concentrations of prostaglandins in arterial blood. Using our previously described

techniques (Wood, 1989), bilateral carotid sinus denervation was performed on the five

fetuses in the experimental group to remove the confounding effect ofbaro- and

chemoreflexes. Next, an extravascular balloon was placed around the brachiocephalic

artery, exposed using an incision in the left second intercostal space. The sagittal sinus

was catheterized to allow sampling of blood at this site for measurement of plasma

prostaglandin concentrations in venous blood. Microdialysis probes (CMA Corporation,










size CMA-10) were inserted in brain stem and hypothalamus region for measurement of

prostaglandins in interstitial fluid at these sites.

Animals were allowed to stabilize for at least 2 h before starting experiments. The

studies consisted of 10-min preocclusion, 10-min occlusion and 10-min postocclusion

periods. Size 8 to 10 occluder was used dependent on the animal size. The occlusion of

the brachiocephalic artery was complete achieved by saline inflatation. Blood samples (3

mL) and microdialysate (flow rate = 160 pL/min) were collected during the 3 periods.

Blood samples were kept chilled on ice and plasma was stored at -80o C until assay of

prostaglandins.


Tissue Collection


The fetuses were sacrificed 30 min after hypotension using an overdose of

pentobarbital. For Western blot and enzyme activity assay studies, fetal hypothalamus,

brainstem, cerebellum, hippocampus and cerebral cortex tissues were quickly removed,

and snap-frozen in a dry ice and acetone bath and stored at -80o C.


Measurement of PGE, and TxB2


PGE2 and TxB, from blood samples and produced in the incubates were measured

by enzyme-linked immunoassay (EIA). All incubation medium samples were centrifuged

at 2500 x g for 25 min at 40 C and the supernatants were collected for prostanoids assay.

This assay has been described in chapter 3.










PGHS Activity Assay


The microsomal PGHS activity (PGHS-1 and PGHS-2) was determined by

measuring PGE2 production under initial velocity conditions (Wimsatt et al. 1993).

Approximately 1 g frozen tissues were homogenized in 10 ml of homogenization buffer

(50 mM Tris, pH=8.0, 0.25 M sucrose, 2 mM EDTA, 1 mM diethyldithiocarbamate) for

two 30-sec bursts on ice using a homogenizer. The homogenates were centrifuged at

2200 x g at 40 C for 10 min, then supernatants were centrifuged at 100,000 x g for 60 min

at 4 C. The pellet was washed twice with 1 ml reaction buffer (50 mM Tris with 2 mM

EDTA, 1 mM diethyldithiocarbamate) and resuspended in 2 ml of this same buffer

containing 7.5 .l Tween 20/ml. The suspension was further dissociated using a drill-

mounted Teflon-glass homogenizer at moderate speed. The microsome preparation was

kept on ice and assayed immediately.

The reaction mixture contained 200 gM free arachidonate, 1 mM hematin, 5 mM

and enzyme preparation in a final volume of 250 pl incubated in a water bath at 380 C for

15 sec. The reaction was terminated by the addition of 200 pl 2.65 mM SnCl2 in 50 mM

HCI made up immediately before use. The reaction mixture without enayme preparation

was for control. All reaction tubes were stored at -800 C until assayed for PGE2.


Thromboxane Synthase Activity Assay


The microsomal thromboxane synthase activity was determined by measuring the

ability to convert PGH2 to TxB2 (Shen and Tai, 1986). The tissues were homogenized for










six 20-sec bursts on ice with 3 vol (W/V) of 25 mM Tris-HCl buffer (pH 7.5), then the

homogenates were centrifuged at 8,000xg for 30 min at 40 C. The supernatants were

centrifuged again at 100,000xg for 1 hr at 40 C. Then the microsome pellets were

resuspened in 1/4 vol (v/w) of 25 mM Tris-HCI buffer (pH 7.5).

Thromboxane synthase is a membrane-associated enzyme for which the membrane

needs to be solubilized before assay. Therefore, the microsomal preparations were

solubilized by stirring the microsomal fraction with solubilization buffer (25 mM Tris-HC1,

20% glycerol, 0.2 mM DTT, 2 mM EDTA, 2% lubrol-PX) at 4 C for Ihr, and the

mixtures were centrifuged for 1 hr at 100,000xg. The reaction mixture contained 10 /M

PGH2 and enzyme preparation in a final volume of 200 ul of 50 mM Tris-HCI buffer (pH

7.5). The reaction mixture without enzyme preparation was for control. All reaction

tubes were stored at -800 C until assayed for TxB2.


Western Blot of PGHS-1, PGHS-2 and Thromboxane Synthase


Approximately 200 mg frozen tissues were homogenized in 1 ml of Laemmli buffer

(0.5 M Tris-HCI, 5% glycerol, 10% sodium dodecyl sulfate, and 2.5% 2-3-

mercaptoethanol) (Laemmli, 1970) for two 30-sec bursts on ice using a homogenizer. The

homogenates were centrifuged at 15,000 x g at 40 C for 30 min and the supernatants were

stored at -40 C. Total protein concentration of the supernatant was determined by using

the Bradford method (Bradford, 1976) and bovine serum albumin as the standard.

The supernatants were boiled in sample loading buffer for 5 min before loading on

10% SDS-polyacrylamide gels. An equal amount of protein was loaded in each lane.










Samples from control and experimental group were run on the same gel. The proteins

were electrophoretically transferred to nitrocellulose membranes, and the nonspecific

binding sites on the membranes were blocked with blocking buffer containing non-fat milk

(15% for PGHS-1, 10% for PGHS-2 and 7.5% for thromboxane synthase respectively),

and 10 mM Tris-HCl (pH 7.5), 100 mM NaCI, 0.1% Tween 20 for lh. The membranes

were incubated in blocking buffer containing PGHS-1, PGHS-2 and thromboxane

synthase specific polyclonal rabbit antibodies (dilution of 1:100, 1:1000 and 1:3000

respectively) for 1h, then incubated in horseradish peroxidase-conjugated anti-rabbit IgG

antibodies in blocking buffer for lh. Finally, the immunoreactive bands were visualized by

a Western blot chemiluminescence reagent (DuPont NEN, Boston, MA) and analyzed by

densitometry. The ovine PGHS-1 or PGHS-2 standard was used in each experiment to

confirm the right molecular weight size. Each experiment was performed three to five

times using different homogenates.


Statistical Analysis


Data were analyzed by paired Student's t test, and significance was established at

p < 0.05. Data are expressed as mean SEM. Statistical procedures were accomplished

using SigmaStat (Jandel Scientific, San Rafael, CA).











Results



Blood Pressure and Blood Gases


Brachiocephalic occlusion decreased carotid arterial blood pressure by

approximately 56%. Mean arterial blood pressure decreased from 40.6 mm Hg to 17.7

mm Hg, then returned to baseline levels after the occlusion (p < 0.01). Hypotension did

not affect arterial pH, PO2, or PaCO2 (Table 4-1).


Table 4-1. Fetal mean blood pressure and blood gases during the experiment

Before During After
MAP (mm Hg) 40.6 3.5 17.7 2.2 34.3 2.8

Pao, (mm Hg) 27.80 + 1.91 29.08 1.59 N/D

Paco, (mm Hg) 64.40 2.05 64.93 3.58 N/D

pH 7.23 0.01 7.24 0.01 N/D

Values are means SE. N/D: not determined.


Prostanoid Production in the Fetal Brain


Plasma thromboxane B2 (TxB2) concentration in sagittal sinus blood increased

significantly after the onset of hypotension and remained high in the 10-min recovery

period. The concentration of TxB, was higher in sagittal sinus plasma than in the arterial

plasma. Plasma prostaglandin E2 (PGE2) concentration did not change during hypotension

and there was no difference between the plasma concentrations in arterial and sagittal

sinus blood (Fig. 4-1). In the microdialysate, TxB2 concentrations in hypothalamus and









57
brainstem decreased during hypotension then returned to pre-hypotension concentrations

(p < 0.05). PGE2 concentration in brainstem and hypothalamus microdialysates increased

in response to hypotension although this response was statistically significant only in the

brainstem (Fig. 4-2; p < 0.05).


PGHS-1 and PGHS-2 Protein Expression and Distribution in Brain Tissues and PGHS
Enzymatic Activity


To examine the effect of carotid sinus denervation and cerebral hypoperfusion on

PGHS levels in fetal brain, Western blot analysis was performed on the constitutive

isoform of PGHS-1 and inducible isoform of PGHS-2 in intact and experimental brain

tissues. Figure 4-3 (top panel) shows that the immunoreactive PGHS-1 protein (70 kDa)

was expressed equally in all regions of the brain and was not affected by the hypotension.

The results of these experiments demonstrated no effect of hypotension on

immunoreactive PGHS-1 protein levels (Fig. 4-3 bottom panel).

The expression of the immunoreactive PGHS-2 protein was relatively region-

specific. Immunoreactive PGHS-2 was detected as a band of 72 kDa and, in some tissues,

a band of about 140 kDa. We suspect that this band was a homodimer which may be

found in fetal tissues. We observed minor doublet bands of 46 kDa which are likely to be

proteolytic fragments of the intact enzyme (data not shown). The 72 kDa molecular

weight of PGHS-2 was expressed in higher abundance in brainstem and hippocampus, but

less in hypothalamus, cortex, and cerebellum (Fig. 4-4 top panel). Hypotension and

denervation increased 72 kDa PGHS-2 protein abundance in all regions (Fig. 4-4 bottom










panel). The expressions of the immunoreactive 140 kDa of PGHS-2 were variable in

different brain regions (Fig. 4-5 top panel), but were increased significantly in all regions

by hypotension and denervation (Fig. 4-5 bottom panel; p < 0.05). There was no

significant effect on the lower molecular weight fragments.

The microsomal PGHS activity assay showed that activity decreased significantly

in all regions of the brain by hypotension and denervation compared to the intact control

(p < 0.05) (Fig. 4-6).


Thromboxane Synthase Protein Expression and Distribution in Brain Tissues and
Thromboxane Synthase Enzymatic Activity


Western blot analysis was used to test the effect of carotid sinus denervation and

cerebral hypoperfilsion on the protein expression ofthromboxane synthase. Fig. 4-7 (top

panel) shows that the expression of the major band on 46 kDa immunoreactive

thromboxane synthase protein was variable in the different brain regions. The

immunoreactive enzyme was most abundant in brainstem, and more abundant in

hippocampus, hypothalamus and cerebellum, than in cerebral cortex. Apparent increases

in thromboxane synthase protein levels in response to hypotension were not statistically

significant (Fig. 4-7 bottom panel). Microsomal thromboxane synthase activity was not

affected by hypotension in the different brain tissues (Fig. 4-8).


Discussion


The present study was designed to test the hypothesis that high prostaglandin

levels in fetal brain during cerebral hypoperfusion may be due to increased PGHS-2










activity and expression. It has been found that prostanoid concentrations in the brain

changed during ischemia (Stevens and Yaksh, 1988; Hanamura et al. 1989). Studies in

adult animals suggested that during ischemia, the changes in prostanoid concentrations in

the brain were important for cerebrovascular control (Dempsey et al. 1986; Chen et al.

1986; Hallenbeck and Furlow, 1979; Shohami et al. 1982). Chemtob et al. (1990c)

demonstrated a relationship between changes in sagittal sinus prostanoid concentrations

and changes in cerebral blood flow (CBF), suggesting that autoregulation of cerebral

blood flow is in part dependent upon local production ofprostanoids. Our previous

studies demonstrated that endogenous prostanoids modulated the adrenocorticotropic

hormone (ACTH) and arginine vasopressin (AVP) responses to moderate hypotension

fetal sheep. Both of these hormonal systems are involved in the homeostatic control of

blood pressure in fetal animals (Tong et al. 1998).

Our experiments demonstrated that, in the sheep fetus, reduction of cerebral

perfusion pressure increases the production of prostanoids, both in the blood draining the

brain and in the extracellular fluid in the brain. In this study, blood pressure was decreased

approximately 60% by brachiocephalic trunk occlusion to reduce the blood flow to the

brain. In separate experiments, we used colored microspheres to demonstrate that

brachiocephalic occlusion decreased cerebral blood flow dramatically (unpublished data).

In the present experiments, we found that the concentration ofthromboxane B2 (the stable

metabolite of TxA2), a cerebral vasoconstrictor (Dempsey et al. 1986), increased

significantly in sagittal sinus blood vs arterial blood during and 10 min after hypotension.

But the concentration of prostaglandin E2, a cerebral vasodilator, was not affected by the









60
hypotension and there was no difference between the arterial and sagittal sinus blood. The

elevated plasma levels of TxA2 in sagittal sinus blood during hypotension was consistent

with finding in other laboratories (Chemtob et al. 1990c). Chemtob et al. (1990a)

suggested that the increase in TxB2 concentration in response to reduced cerebral blood

flow was consistent with thromboxane contributing to the lower limit of CBF

autoregulation in newborn piglets. Our present results are also consistent with this

hypothesis.

Our microdialysis data demonstrated that thromboxane B2 concentrations in the

hypothalamus and brainstem microdialysates decreased during hypotension, and returned

to normal level after hypotension. But prostaglandin E2 concentrations in brainstem and

hypothalamus microdialysate increased during and 10 min after the hypotension (Fig. 4-

2). The results of the present study are consistent with the notion that TxA2 may

contribute to cerebrovascular control during hypotension and cerebral hypoperfusion.

Prostaglandin E2 may be responsible for mediating the hormonal secretions by acting on

the secretary neurons in the brain via a autocrine and/or paracrine mechanism. Our

speculation concerning dual cerebrovascularr and neuromodulatory) effects of brain

prostanoids was based on observations made by others using immunocytochemistry,

Western blot and Northern analysis (Breder et al. 1992; Breder et al. 1995; Peri et al.

1995). Peri et al. (1995) demonstrated that the immunoreactive PGHS-1 and PGHS-2

protein and mRNA expressed in newborn and juvenile pig cerebral microvasculature and

brain. Breder et al. (1992, 1995) showed that immunoreactive PGHS-1 and PGHS-2

were in neurons and microvasculature in the brain. We cannot distinguish the source of










immunoreactive PGHS protein in our study, although we speculate that the increased

prostanoid in sagittal sinus blood was released from cerebral microvasculature, and the

increased prostanoid in brain microdialysate was probably released from neurons

containing PGHS.

PGHS undergoes "suicide" inactivation during catalysis (Smith and Lands, 1971;

Smith and Lands, 1972). Incubation of cyclooxygenase with arachidonic acid results in

loss of cyclooxygenase but not peroxidase activity, concurrent with PGH2 production.

The chemical basis for inactivation is unknown. Degradation or modification of the heme

prosthetic group could account for inactivation (Hemler and Lands, 1980; Chen et al.

1987). Another possibility is that the PGHS gene is an "immediate early gene", its

translation product may act to inhibit the transcription of the PGHS gene (Smith et al.

1991). It is important to note that inactivation of PGHS has been demonstrated in vivo as

well as in vitro, indicating that it may indeed be a physiologically relevant process

(Lapetina and Cuatrecasas, 1979; Kent et al. 1983). Balancing inactivation is the gene

level expression of the protein. Thromboxane synthase is also a "suicide" enzyme (Jones

and Fitzpatrick, 1990). Inactivation of thromboxane synthase is a mechanism-based

process and the enzyme heme prosthetic group appears to be affected.

Ovine PGHS-2 is more resistant to trypsinization than PGHS-1 (Wimsatt et al.

1993). Our results showed that PGHS-1 was measured as a 70 kDa molecular weight in

fetal tissues. Immunoreactive PGHS-2 was expressed as a 140 kDa and 72 kDa molecule

in various fetal brain tissues. We speculated that the 140 kDa exists as a homodimer of 72

kDa PGHS-2. Some lower molecular weight proteolytic fragments appeared in the










Western blot. Various factors can affect the susceptibility to enzyme digestion. This

difference could result from increased glycosylation and glycosylation sites on the enzyme

(Hla and Neilson, 1992). Alternatively, PGHS-2 may exist with a large amount of the

enzyme in the heme-bound state, as heme binding can inhibit trypsin degradation of the

enzyme (Kulmacz and Lands, 1982). Existence of the enzyme in its unoxidized state is

also more susceptible to cleavage by trypsin, suggesting that PGHS-2 might be less readily

oxidized than PGHS-1 (Chen et al. 1987). Purified PGHS protein has a molecular weight

of 70 kDa when analyzed on denaturing electrophoresis gels. However, when purified

PGHS is applied to a gel filtration column, it elutes as a homodimer at an apparent

molecular weight of 140 kDa (Hemler et al. 1976). Gel filtration of PGHS can be carried

out in buffer containing 6 N guanidine.HCI to interrupt non-covalent interactions, but the

purified protein still elutes as a homodimer under these conditions (Ogino et al. 1978).

The clipped enzyme retains its molecular mass of 140 kD in solution and can still be

acetylated by aspirin (Chen et al. 1987; Chen and Marnett, 1989). Overall, trypsin-

digested PGHS appears to behave just like native PGHS, except for its diminished

capacity for peroxidase activity. Limited digestion of PGHS into fragments of 33 kD and

38 kD also has diminished peroxidase activity whereas the cyclooxygenase activity persists

(Chen et al. 1987; Dildy et al. 1994).

Western blot and enzyme activity assay of PGHS demonstrated that a significant

quantitative difference of the two isoforms was noted in different brain regions. In

contrast to the newborn piglet (Peri et al. 1995), PGHS-1 was predominant in the intact

animals and was not appreciably increased by cerebral hypoperfusion. PGHS-2










immunoreactivity was increased by cerebral hypoperfusion while PGHS enzyme activity

was decreased, suggesting that PGHS-2 is responsible for the brain prostanoids

production when the fetus suffered from cerebral ischemia. Thromboxane synthase

protein levels and activity were not altered by cerebral hypoperfusion and carotid sinus

denervation in this study.

Prostaglandins have been shown to exhibit neuroprotective properties (Gazevieille

et al. 1994). It is likely that their increased levels during hypotension may provide

protection to the fetal brain from low oxygen tension (Dildy et al. 1994) and the risk of

hypoxic brain injury increases. Since PGHS-2 is induced rapidly, it would be suited for

such a temporary but important role in response to stress.

In summary, this study demonstrates that in late-gestation fetal sheep: 1)

prostanoids are produced in response to cerebral hypoperfusion; 2) the increase in the

production of prostanoids responses to cerebral hypoperfusion is associated with the

decrease in activity of, and therefore the "suicide" inactivation of, prostaglandin

endoperoxide synthase, and 3) PGHS-2 is the predominant form of prostaglandin

endoperoxide synthase whose synthesis is induced by cerebral hypoperfusion in the fetal

brain. We concluded that PGHS-2 is likely the main contributor to the brain prostaglandin

levels during cerebral hypoperfusion in the fetuses.









64



4000
I I arterial **
Ssagittal sinus
3000


0)
,-I

C 2000

1-
1000



0
-10 0 10 20 30


4000


T
3000



S 2000 -
w





0 --
-10 0 10 20 30

Time from onset of brachiocephalic occlusion (min)




Figure 4-1. TxB2 and PGE2 concentrations (mean SE) in fetal arterial (open bar) and
sagittal sinus plasma (shaded bar) before, during and 10 min after a 10-min period of
brachiocephalic artery occlusion. ** p <0.01, indicates difference from blood sampling
sites. N=5.












Hypothalamus


100-

80-


-10 0 10 20 30

Brainstem


-10 0 10 20 30


-10 0 10 20


Brainstem


-10 0 10 20 30


Time from onset of brachiocephalic occlusion (min)



Figure 4-2. TxB2 and PGE2 concentrations (mean SE) in the hypothalamus (upper
panels) and brainstem (lower panels) microdialysates from denervated fetuses before,
during and 10 min after a 10-min period of brachiocephalic artery occlusion. p < 0.05,
indicates difference of PGE2 concentration during hypotension compared to that before
hypotension. N=5.


Hypothalamus













H N
H N


H N
H N


4 1
H N


/
C4
C)


H
H N


70kD -*


/a


CCB~
C, d


Figure 4-3. Western blot analysis of PGHS-1 protein level with PGHS-1 antibody in the
intact (open bar) and denervated hypotensive (shaded bar) fetal brain tissues.op panel:
representative immunoblot showing that in each brain tissue the anti-PGHS-I aibod-y
recognized a 70 kDa protein band. bottom panel.quantification of PGHS-1 protein band
as optical density (OD) and values are means SE from five fetuses. H: represents tissue
from hypotensive animal; N: represents tissue from intact animal.













HN
H N H


N H
[ N H


I

N


H N


H
H N


72kD


sC
c^ ^'


0
/3 c'


Figure 4-4. Western blot analysis of PGHS-2 protein level with PGHS-2 antibody in the
intact (open bar) and denervated hypotensive (shaded bar) fetal brain tissues. Top panel:
representative immunoblot showing that in each brain tissue the anti-PGHS-2 antibody
recognized a single band of 72 kDa protein. Bottom panel: quantification of PGHS-2
protein bands as optical density (OD) and values are means SE from five fetuses. H:
represents tissue from hypotensive animal; N: represents tissue from intact animal.














H N


140kD *


HN H
H N H


N


-L~ *I- -C I


HN
UN


H N
I N


-I


0
C,~


Figure 4-5. Western blot analysis of PGHS-2 protein level with PGHS-2 antibody in the
intact (open bar) and denervated hypotensive (shaded bar) fetal brain tissues. Top panel:
representative immunoblot showing that in each brain tissue the anti-PGHS-2 antibody
recognized a single band of 140 kDa protein. Bottom panel: quantification of PGHS-2
protein as optical density (OD) and values are means SE from five fetuses. H: represents
tissue from hypotensive animal; N: represents tissue from intact animal. *p < 0.05 indicates
difference between groups.


/


~
1~9"









69




3

EI I Hypotension
w0 Intact




CL2 -
C
e
.I-


0)
0




C,



0









Figure 4-6. PGHS activity in the intact (shaded bar) and denervated hypotensive (open
bar) fetal brain tissues. *p < 0.05, indicates difference between groups. Values are means
SE from five fetuses.

















N
H N


46kD -







20-



16-



12-



8



4



0


HN HN H NH N


C,


Figure 4-7. Western blot analysis ofthromboxane synthase protein level with thromboxane
synthase antibody in the intact (open bar) and denervated hypotensive (shaded bar) fetal
brain tissues. Top panel: representative immunoblot showing that in each brain tissue the
thromboxane synthase antibody recognized a 46 kDa protein band. Bottom panel:
quantification ofthromboxane synthase protein bands as optical density (OD). Values are
means SE from five fetuses. H: represents tissue from hypotensive animal; N: represents
tissue from intact animal.


--. o ft oel lil O l I













SI Hypotension
1 Intact


Figure 4-8. Thromboxane synthase activity in the intact (shaded bar) and denervated
hypotensive (open bar) fetal brain tissues. Values are mean + SE. N=5.


6-


5-



4-



3-



2-


O
o



r.L




C-DO
E P

EE
0 C-


77







//0/
I.






I" I














CHAPTER 5
EFFECT OF PROSTANOIDS AND INDOMETHACIN ON CEREBRAL BLOOD
FLOW IN RESPONSE TO CEREBRAL ISCHEMIA IN LATE-GESTATION FETAL
SHEEP


Introduction


Substantial physiological data in vivo point to a significant role for dilator

prostanoids in the regulation of cerebral vascular tone and cerebral blood flow in

newborns (Leffler et al. 1993). The physiological effects of arachidonate metabolites

often have opposing actions. For example, PGE2 and PGI2 are potent vasodilators, while

TxA2 and PGF,, are vasoconstrictors (Moncada and Vane, 1978; Moskowitz and

Coughlin, 1981). Arachidonic acid is liberated from damaged cell membranes during

ischemia and is the source of vasoactive prostanoids which are believed to adversely

influence cerebral ischemia.

Inhibition of cyclooxygenase, an enzymatic step in arachidonic acid metabolism,

has an improvement effect in the outcome during cerebral ischemia. Free indomethacin

crosses the blood brain barrier (Bannwarth et al. 1990). It is generally assumed that the

duration of action of indomethacin is long, but the plasma half-life is relative short and the

CSF concentration is closely related to that of plasma (Bannwarth et al. 1990).

Indomethacin, a cyclooxygenase inhibitor, has a potential ability to prevent or attenuate

the development of intraventricular hemorrhage (Ment et al. 1994). A pharmacological








73

dose ofindomethacin is associated with a selective hypoperfusion to certain brain regions

during normoxia and this result is associated with changes in the prostanoid system (Coyle

et al. 1995). Indomethacin has been reported to have beneficial prophylatic effects in

prevention and/or reduction of severity of intraventricular/periventricular hemorrhage in

premature babies (Ment et al. 1985b; Hanigan et al. 1988; Bada et al. 1989). High dose

indomethacin (5mg/kg) had been found to decrease regional and total cerebral blood flow

for 120 min after administration in newborn pig (Pourcyrous et al. 1994). In piglets,

indomethacin decreases CBF at rest (Leffler et al. 1985). These decreases in CBF occur

concomitantly with decreases in CSF dilator PG levels (Leffler et al. 1989).

No information exists with regard to the effect ofprostanoids and indomethacin on

cerebral ischemia in fetuses. Our previous studies showed that prostaglandin E2

concentrations increased in the interstitial fluid of the brain stem and hypothalamus, but

thromboxane B2 (TxB2) decreased in the brain stem and hypothalamus during cerebral

hypoperfusion. TxB2 increased in sagittal sinus blood compared to arterial blood (Tong

and Wood, 1998b). In the present study, we hypothesized that during cerebral

hypoperfusion, prostanoids produced in the microvasculature and in neurons might cause

vasoconstriction and reduce cerebral blood flow, and that pretreatment with indomethacin

may attenuate this reduction of cerebral blood flow. To test these hypotheses, we

measured regional brain blood flow in fetal sheep subjected to a 10-min of cerebral

ischemia with and without pretreatment with indomethacin.










Materials and Methods


Eight chronically catheterized fetal sheep between 126 and 136 days' gestation

were used in this study. The pregnant ewes were of mixed Western breeds. These

experiments were approved by the University of Florida Institutional Animal Care and Use

Committee and were performed in accordance with the Guiding Principles for the Care

and Use of Animals of the American Physiology Society.


Surgical Preparation


Aseptic surgery was performed at least 5 days before the start of experiments in

each animal. The fetuses were chronically instrumented with tibial artery and saphenous

vein catheters and amniotic fluid catheters as previously described. Fetal hindlimbs were

identified and delivered through a small hysterotomy incision near the tip of one uterine

horn. We introduced polyvinyl chloride catheters into the tibial artery (.050" ID; .090"

OD) and saphenous vein (0.030"ID, 0.050" OD) bilaterally and advanced the tips to the

abdominal aorta and inferior vena cava, respectively. After closing the skin incisions in the

fetal hindlimb, amniotic fluid catheters were sutured to the fetal skin and the hindlimb

returned to the amniotic space. After placement of these catheters and closure of the

hysterotomy, the fetal head was located, the uterus was incised, and the head was

delivered. After a single midline incision in the skin of the neck was made, lingual arteries

were identified, ligated, and catheterized with polyvinyl chloride catheters (0.050 in. OD,

0.030 in ID), with the catheter tips advanced retrograde to the lumen of the common











carotid arteries. As previously described, this catheterization technique allows

measurement of common carotid arterial pressure without interruption of carotid arterial

blood flow (Wood et al. 1993). The carotid sinus denervation was performed using

methods described by Wood (1989). The carotid sinus nerves were identified bilaterally

and cut. The walls of the common carotid arteries in this area, as well as the lingual

arteries and common carotid arteries extending 0.5-1 cm rostral from the lingual-carotid

arterial junction were stripped of all visible nerve fibers. After performing these

denervations in the fetal neck, the fetal skin was closed. An extravascular balloon

occluder (size 10) (In Vivo Metric, Healdsburg, CA) was placed around the

brachiocephalic artery using a separate incision through the fourth intercostal space on the

left side of chest. Then the head was returned to the amniotic cavity, and the uterus was

closed. All catheters exited via a small incision in the flank of the ewes.

Ampicillin trihydrate (Polyflex; Aveco Co, Fort Dodge, IA) 500 mg was

administered to the fetus via the amniotic fluid and to the mother (500-750 mg)

intramuscularly at the time of surgery and again each time the fetus was studied or the

catheters were flushed. Ampicillin (500-750 mg) was administered to the mother

intramuscularly twice daily for five days after the surgery. All catheters were flushed and

reheparinized at least once every 3 days.


Experimental Protocol


Sheep were transported to the procedure room from their pens within the Health

Center Animal Resources Department at least 1 hour before the start of each experiment.











Each fetus was studied twice. Experiments consisted of a 90 min pre-occlusion control

period (-90 to 0 min), a 10 min occlusion period (0 to 10 min) and a 10 min post-

occlusion recovery period (10 to 20 min). The occlusion of brachiocephalic artery was

complete. In one experiment on each fetus, the vehicle ofindomethacin, 0. 1 M phosphate

buffer (PB) was injected intravenously, and in the other experiment 0.2 mg/kg

indomethacin (an inhibitor of cyclooxygenase) was injected intravenously 90 min prior to

the 10 min period of hypotension. PB and indomethacin were randomly administrated to

the fetal sheep. In each experiment, the extravascular occluder was inflated for 10 min to

produce arterial hypotension. Fetal arterial blood samples were drawn from tibial artery

catheters (descending aorta) at 90 and 10 min before the start of the period of occlusion,

at the end of the 10 min period of occlusion, and 20 min after the onset of occlusion. One

ml arterial blood sample was drawn before the experiment to measure blood gases. Blood

samples (3 ml) were collected into chilled polystyrene tubes containing 150 pl of 0.5 M

EDTA. Separate blood samples (1 ml) were collected into chilled polypropylene tubes

containing 50 pl 0.5 M EDTA and 40 ig/ml indomethacin for measuring TxB, and PGE2.

Tubes were kept on ice until the end of the experiment and then centrifuged for 20 min at

3000 x g at 4 o C. Plasma was separated and stored in separate aliquots at -20 o C.

Plasma adrenocorticotropic hormone (ACTH), cortisol, and arginine vasopressin

(AVP) concentrations were measured by specific radioimmunoassay. These assays have

been described in detail in chapter 3 and (Wood et al. 1993; Raffet al. 1991). PGE2 and

thromboxane B2 (TxB2, a stable metabolite of TxA2) were measured using enzyme-linked

immunoassay kits purchased from Cayman Chemical. Before assay, the prostanoids were








77

extracted from acidified plasma with 6 volumes of ethyl acetate. The recovery using this

protocol averages approximately 60%, and the extracted prostanoids dilute parallel to the

standard curves. This assay has been described in chapter 3.

Fetal arterial blood pressure and amniotic fluid pressure values were measured

continuously during the 110 min experiments using a Grass recorder and Statham P23 ID

pressure transducers. These hemodynamic variables were sampled, and analog-to-digital

conversions were performed at 2-s intervals using an IBM PC computer. The data

collection was achieved using ASYSTANT' software (Asyst Technologies, Rochester,

NY). All fetal intravascular pressure were corrected by subtraction of amniotic fluid

pressure.

Immediately before and 5 minutes after the onset of hypotension, colored

microspheres (15 tm, Dye-Track, Triton Technologies, San Diego, CA) were injected

through the venous catheter and simultaneous reference blood samples were drawn from

the lingual arterial catheters at the rate of 3 ml/min for 1 min 30 sec. Four colors of

microspheres were used in each animal for measuring cerebral blood flow. These

techniques have been described in detail (Lakhdia et al. 1998).

At the end of the second experiment, the pregnant ewes (and the fetus) were

euthanized with an overdose of pentobarbital. The fetal brains were dissected for

microsphere extraction.


Estimation of Cerebral Blood Flow Using Colored Microspheres


The fetal brains were dissected into cerebral cortex, brain stem, hippocampus,

hypothalamus, cerebellum, and pituitary. Same amount of tissues were used in this study.








78

The dissected tissues were dissolved in 4 ml of 4 M potassium hydroxide and 0.2% Tween

80 for 48 hours and filtered through an 8 pm thickness polyester membrane filter. The

membrane (containing the microspheres) was placed in a microcentrifuge tube to air dry

overnight. Dimethylformamide (150 pl) was added to the tube to leach the color

microsphere from the membrane. Then, the tube was centrifuged and 80 pl of the

supernatant was pipetted to a cuvette. The absorbance spectra (ABS) was quantified

using a spectrophotometer and the appropriate software for least-squares solution of

absorbance of each dye. The reference blood samples were processed similarly, except

that 16 M potassium hydroxide was used.

Cerebral blood flow in each tissue was calculated by the following equation:

CBF = (tissues ABS/reference ABS) reference flow rate



Statistical Analyses


Change in mean arterial blood pressure over time was analyzed by one-way

ANOVA. Changes in the values of fetal hormonal and prostanoid variables over time and

between groups were analyzed using two-way analysis of variance (ANOVA). A multiple

comparison of mean values was performed using Student's Newman-Keuls test. The

hormonal data were not distributed normally. All ANOVAs performed on hormonal data

were calculated after logarithmic transformation to correct heteroscedasity of the data.

Fetal CBFs with PB/indomethacin treatment and occlusion were analyzed using three-way

analysis of variance. A multiple range test was performed using Duncan's test. A










significance level ofP < 0.05 was used to reject the null hypothesis in all tests. Analyses

were performed using SigmaStat software (Jandel Scientific, San Rafael, CA).



Results


Blood Gases and Blood Pressure


The blood gases measured before phosphate buffer and indomethacin injections

indicated that all fetuses were healthy during the experiments (table 5-1). Injection of

indomethacin did not alter fetal blood gases. Fetal mean arterial blood pressure before,

during and after brachiocephalic occlusion were presented in Fig. 5-1. Lingual arterial

blood pressure decreased significantly (p < 0.001) by the brachiocephalic occlusion in both

PB and indomethacin treated groups. In the PB group, mean lingual arterial blood

pressure decreased from 41.9 3.1 to 16.5 3.3 mm Hg (mean SE). In the

indomethacin group, blood pressure decreased from 41.6 2.5 to 14.6 3.0 mm Hg.

Blood pressure returned to baseline level after the occluder was released. Femoral arterial

blood pressure increased significantly (p < 0.05) by the brachiocephalic occlusion in PB-

treated group, from 42.9 3.7 to 64.9 7.0 mm Hg. In the indomethacin-treated group,

mean femoral arterial blood pressure during occlusion (51.1 4.2 mm Hg) was higher (P

< 0.05) than that after occlusion (39.3 1.7 mm Hg). These data suggested that the

blood flow shunted to peripheral tissues when the cerebral perfusion pressure was

reduced.










Plasma Prostanoid Concentrations


The mean values of fetal plasma prostaglandin E2 and thromboxane B2

concentrations are summarized in Table 5-2. There was no difference in the initial values

of prostanoid concentrations before PB and indomethacin injections. After indomethacin

injection, plasma PGE2 and TxB2 concentrations decreased significantly (p < 0.05),

indicating that indomethacin effectively inhibited cyclooxygenase activity. Compared to

the preocclusion value, hypotension increased plasma PGE2 and TxB2 concentrations in

PB-treated group, but there was no statistically significant difference.

Table 5-1. Arterial blood gases before PB and indomethacin injections in denervated
fetuses

PB Indomethacin

pH 7.35 + 0.01 7.35 0.01
PaO2 (mm Hg) 19.28 0.91 20.42 0.68

PaCO2 (mm Hg) 51.88 2.43 49.53 1.72

Values are means SEM.

Regional Cerebral Blood Flow


The mean values of fetal regional cerebral blood flow (rCBF) are reported in table

5-3. In PB-treated group, brachiocephalic occlusion significantly decreased rCBF in

cerebral cortex (from 1.42 0.22 to 0.30 + 0.09 ml/g.min) (mean SE), brain stem (from

2.43 0.39 to 0.86 0.25 ml/g.min), hippocampus (from 1.37 0.18 to 0.49 + 0.17

ml/g.min), hypothalamus (from 2.20 + 0.33 to 1.05 0.40 ml/g.min) and cerebellum (from

1.85 0.27 to 0.41 0.15 ml/g.min) (P < 0.05).











Table 5-2. Fetal plasma prostanoid concentrations in denervated fetuses


PB (n = 6)


Indomethacin (n = 7)


PGE2 (-90 min)

PGE2 (-10 min)

PGE2 (10 min)

PGE2 (20 min)

TxB2 (-90 min)

TxB2 (-10 min)

TxB2 (10 min)


192.2 79.8

163.2 74.5

229.8 119.5

265.7 142.4

85.8 43.4

65.6 25.9

118.9 47.7


181.4 31.9

21.1 8.5"*

44.3 25.1*

17.3 5.4*t

92.2 34.6

20.5 5.6*

10.5 3.9*t


TxB2 (20 min) 87.2 31.2 22.9 8.1*t
Values are means SEM. Time is relative to the start of hypotension (t = 0 min). *p <
0.01, indicates difference due to the time difference within indomethacin group (Student-
Newman-Keuls test). tp < 0.05, indicates difference between PB and indomethacin group
(Duncan's test).

Table 5-3. Fetal cerebral blood flow ofPB and indomethacin treatments


PB Indomethacin


Cortex

Brain stem

Hippocampus

Hypothalamus

Cerebellum


Normotension
(n = 5)

1.42 0.22

2.43 + 0.39

1.37 0.18

2.20 0.33

1.85 0.27


Hypotension
(n = 6)

0.30 0.09*

0.86 0.25*

0.49 0.17*

1.05 0.40*

0.41 0.15*


Normotension
(n = 7)

1.22 0.12

2.21 0.29

1.34 0.13

1.78 0.21

1.64 0.20


Hypotension
(n = 7)

0.59 0.11

1.22 0.28t

0.76 0.15

1.06 0.12

0.70 + 0.19t


Values are means + SEM. *p < 0.05, indicates the difference between normotension and
hypotension in the PB-treated group. tp < 0.05, indicates the difference between
normotension and hypotension in the indomethacin group. All multiple range test
procedures were done in Duncan's test.









82

Indomethacin pretreatment (n =7) slightly decreased but did not significantly change basal

rCBFs. Hypotension decreased rCBF in the indomethacin-treated group, especially in

cerebral cortex (from 1.22 0.12 to 0.59 0.11 ml/g.min), brain stem (from 2.21 0.29

to 1.22 0.28 ml/g.min) and cerebellum (from 1.64 0.20 to 0.70 0.19 ml/g.min) (p <

0.05). Regional CBFs in the indomethacin-treated group during brachiocephalic occlusion

were higher than that in the PB group during occlusion. Indomethacin increased rCBF in

cerebral cortex, brain stem, hippocampus, hypothalamus and cerebellum by 49%, 30%,

36%, 18% and 42% respectively during hypotension. Three-way analysis of variance

(ANOVA) performed on fetal CBFs over PB/indomethacin treatment and blood pressure

change, revealed that there was a significant interaction between PB/indomethacin

treatment and blood pressure change (F = 5.224, p < 0.05). Duncan's test was used in the

multiple range test and the results were reported in table 5-3. These data suggested that

indomethacin decreased the reduction in fetal cerebral blood flow during cerebral

hypoperfusion.


Plasma Hormone Concentration


Indomethacin treatment did not alter ACTH and AVP responses to hypotension.

Plasma ACTH concentration was increased significantly from 74 14 to 1811 1009

pg/ml (mean SE) in the fetuses treated with PB, and from 200 + 52 to 1490 482 pg/ml

treated with indomethacin by the brachiocephalic occlusion. Plasma AVP concentration

was increased significantly from 11 2 to 47 18 pg/ml in the fetuses treated with PB,

and from 20 6 to 74 32 pg/ml for those treated with indomethacin by the








83

brachiocephalic occlusion. Plasma cortisol concentration was increased from 11 2 to 21

5 ng/ml in the fetuses treated with PB, and from 16 + 4 to 28 8 ng/ml treated with

indomethacin by the hypotension. There was no significantly difference between PB and

indomethacin treatment (analyzed by two-way ANOVA) (Fig. 5-2). This is consistent

with our previous studies that indomethacin did not further attenuate hormonal responses

to hypotension in denervated fetuses (Tong et al. 1998a).


Discussion


The present study was designed to test the roles of prostanoids and indomethacin

in the control of fetal cerebral blood flow (CBF) during cerebral ischemia. It has been

observed that prostanoids are important in modulating perinatal cerebral hemodynamics in

human beings and some animals (Leffler and Busija, 1987b). Our previous study

demonstrated that TxB2 secretions increased significantly in sagittal sinus blood during

cerebral hypoperfusion, and PGE2 concentrations increased significantly in fetal brain stem

and hypothalamus interstitial fluids. We tested the hypothesis that indomethacin may

compromise fetal CBF by inhibiting these prostaglandins productions in the brain during

cerebral ischemia.

Papile et al. (1985) demonstrated that CBF autoregulation was developed in the

preterm fetal lamb and the range of autoregulation was narrowed in the preterm lamb. It

was speculated that the fetal brain under normal physiological conditions in utero may not

protect against acute fluctuations in blood pressure, particularly hypotension (Papile et al.

1985).










In this study, brachiocephalic occlusion decreased the same degree of lingual

arterial blood pressure which perfuses brain in both PB and indomethacin-treated fetuses.

Also, hypotension decreased cerebral blood flow in both groups. Interestingly, this study

showed that indomethacin increased fetal regional CBF by 30-49% compared to PB-

treated group when blood pressure was reduced by about 60% (below 20 mm Hg). We

believe that the increases in CBF in indomethacin pretreatment were due to inhibiting

constrictor prostaglandin by indomethacin. Indeed, TxB2 concentration decreased

significantly after indomethacin injection in this study. Indomethacin is highly lipid soluble

but also is highly protein-bound; it crosses the intact meninges by simple diffusion

(Bannwarth et al. 1990). Indomethacin can be found in the brain within 30 min of

systemic administration (Bannwarth et al. 1990). In contrast to our findings, Leffler et al.

(1989) found that indomethacin decreased CBF in newborn animals by inhibiting dilator

prostaglandins. Indomethacin possesses several actions other than cyclooxygenase

inhibition (Burch et al. 1983; Jugdutt et al. 1979) that may account for the CBF reduction

during a decrease in blood pressure within the range of CBF autoregulation in newborns

(Leffler et al. 1986). Chemtob et al. (1990a) demonstrated that in newborn piglets,

inhibition of TxB2 synthesis was associated with an extension of the lower blood pressure

limit of CBF autoregulation and an attenuation in the decrease in CBF below this blood

pressure. These studies suggested that there were differences in the cerebrovascular

action of indomethacin in different degree of hypotension.

In this experiment design, all fetuses were carotid sinus denervated in order to be

consistent with the previous study (Tong et al. 1998a) and this may not affect cerebral










circulation in response to ischemia. The fetal regional cerebral blood flow was in the

normal range in PB-treated animals. Itskovitz and Rudolph (1982) demonstrated that

sinoaortic denervation of the preterm fetal lamb dose not alter resting heart rate or blood

pressure and has no effect on CBF within the autoregulation range (75-80 mm Hg). It

appeared that the autonomic nervous system does not appear to have a major role in the

control of cerebral circulation (Traystman, 1981).

The hormonal responses to cerebral hypoperfusion of this study were consistent

with our previous experiments designed to test the endogenous prostanoids modulating

the hormonal responses to arterial hypotension and the effects of denervation on these

responses (Tong et al. 1998a). Carotid sinus denervation attenuated ACTH and AVP

responses to hypotension and endogenous prostanoids did not further affect these

hormonal responses in late-gestation fetal sheep.

In conclusion, this study demonstrated that indomethacin had beneficial effect on

fetal regional cerebral blood flow in response to cerebral ischemia in late-gestation fetal

sheep. This effect was due to inhibiting constrictor prostaglandin synthesis in the brain by

indomethacin. This may be an alternative way of indomethacin to treat severe cerebral

ischemia in fetuses.










86



75

E [- ] PB
E 60 eA Indomethacin


30 -
a)







S15-

0







30 -

15 -
0
Before During After

S75



In
60



-a 45
0

30
t

E 15


M 0
Before During After

Time relative to brachiocephalic occlusion




Figure 5-1. Fetal mean lingual (top panel) and femoral (bottom panel) arterial blood
pressure (MAP) before, during and after a 10-min period of brachiocephalic occlusion
(BCO) in the denervated fetuses. There were no significant differences in MAP between
PB and indomethacin treated groups. Mean values are accompanied by vertical bars
representing 1 SE. **p < 0.01 and *p < 0.05, compared to baseline MAP by BCO.









87




2000

j 1600- 0 NO
NO I N- D

0 1200

I 800
I-
< 400 -

I II I I
-10 -5 0 5 10 15 20
100

80










-10 -5 0 5 10 15 20
40-

=-r
E 30-



20

0
o -------- --------
0 10

0

-10 -5 0 5 10 15 20
Time from onset of brachiocephalic occlusion (min)



Figure 5-2. Fetal plasma ACTH, AVP and cortisol concentrations before, during and after
a 10-min period of brachiocephalic occlusion (BCO) in the denervated fetuses. Mean
values of hormones in phosphate buffer (0, solid lines) and indomethacin (U, dash lines)
treated fetuses are accompanied by vertical bars representing 1 SE. Two-way ANOVA
did not show significant effect between groups.














CHAPTER 6
EXPRESSION OF PGHS-2 AND FOS GENE INDUCED BY CEREBRAL ISCHEMIA
IN FETAL SHEEP BRAIN


Introduction


Prostaglandin endoperoxide synthase (PGHS) is a rate-limiting enzyme in

prostaglandin and thromboxane syntheses. There are two isoforms ofprostaglandin H

synthase (PGHS), inducible PGHS-2 encoded by a 4 kilobase mRNA, and constitutive

PGHS-1 encoded by a 2.8 kilobase mRNA. Promoter analysis suggests that PGHS-1 is a

member of the housekeeping gene family (Wang et al. 1993) and PGHS-2 is a member of

immediate early gene (IEG) family (Xie et al. 1993). The promoter sequence of PGHS-2

gene contained a TATA box and a variety of enhancer elements, including an activator

protein-1 (AP-1) site (Xie et al. 1993). The PGHS-2 gene may be transcriptionally

regulated at its AP-1 binding element by transcription factor c-fos (Chiu et al. 1988; Xie et

al. 1993); however, promoter analysis of the PGHS-1 gene reveals no putative AP-1

domain (Wang et al. 1993).

There are several families of the IEGs. They include the c-fos and c-jun that

encode transcription factors. Products of the fos family bind to members of the jun family

and form AP-1 transcription factor in the promoter of target genes, which in turn regulate

the expression of late response genes that produce long-term changes in cells (Chiu et al.








89

1988). Cerebral ischemia induced proto-oncogene expression in animal models (Kindy et

al. 1991; Onodera et al. 1989). It is suggested that the expression of c-fos after ischemia

may be immediately activated through NMDA receptors and may spread to surrounding

regions in a manner sensitive to reductions in rCBF (Shimazu et al. 1994).

Immunocytochemistry studies showed that PGHS are located intracellularly and

that they function to produce prostaglandins for use as intracellular second messengers or

as paracrine mediators (Breder et al. 1992). PGHS is normally expressed throughout the

brain in discrete populations of neurons (Breder et al. 1995) and detected in several non-

neuronal cells such as macrophages (Lee et al. 1992), astrocytes (O'Banion et al. 1996),

and endothelial cells (Habib et al. 1993). PGHS-2 was seen in many neurons that were

also immunoreactive for Fos in the brain (Holtz et al. 1996). Brain ischemia was reported

to induce PGHS-2 mRNA and protein (Collaco-Moraes et al. 1996). However, Fos was

not involved in the expression of PGHS-1 mRNA during cerebral ischemia (Holtz et al.

1996). PGHS-2 induction in neurons at risk to die could contribute to neuronal death in

brain ischemia, this is supported by our previous study that the cyclooxygenase inhibitor,

indomethacin, reduced brain damage in global ischemia (Tong and Wood, unpublished

data).

Our previous studies demonstrated that cerebral ischemia induced PGHS-2 but not

PGHS-1 protein expression in brain tissues in late-gestation fetal sheep. We also found

that prostanoids are released into cerebrovasculature and brain interstitial fluids in

response to the cerebral ischemia in fetal sheep. One explanation for the rise in

prostaglandin levels in the brain during the cerebral ischemia may be the increased










transcription of the PGHS-2 gene and the translational product. We also hypothesized

that the transcription of the PGHS-2 gene in the ischemic brain would be affected by

protein translated from c-fos mRNA, due to its AP-1 binding domain. We tested these

hypotheses by measuring levels of PGHS-2 mRNA and protein and Fos, and localizing

their expressions in the fetal brain during cerebral ischemia.


Materials and Methods


Materials


Polyclonal rabbit antibodies specific to PGHS-l and PGHS-2 were purchased from

Oxford Biomedical Research (Oxford, MI). Polyclonal rabbit antibody specific to

thromboxane synthase was purchased from Cayman Chemical (Ann Arbor, MI).

Polyclonal rabbit antibody specific to Fos was purchased from Oncogene Science

(Cambridge, MA). Stable DAB was purchased from Research Genetics (Huntsville, AL).

ImmunoPure ABC staining kit was purchased from Pierce (Rockford, IL). Oligo(dT)15

primer, Taq DNA polymerase, DNA ladder, RNAase inhibitor, dNTP were from Promega

(Madison, WI). SuperScript M-MLV reverse transcriptase was purchased from GIBCO-

BRL (Gaithersburg, MD). Guanidine thiocyanate, MOPS, Tris, SDS, EDTA, chloroform,

isopropanol were from Sigma (St. Louis, MO). All other chemicals were from Fisher.

Animals


The pregnant ewes were of mixed Western and Florida native breeds were

obtained from IFAS at University of Florida. The fetal sheep between 126 and 136 days'










of gestation age were used in this study. These experiments were approved by the

University of Florida Institutional Animals Care and Use Committee and were performed

in accordance with the Guiding Principles for the Care and Use of Animals of the

American Physiology Society.


Tissue Preparations


Animals for immunocvtochemistry study

There were three groups in this study: one group of fetuses (n = 3) were intact for

control; carotid sinus denervation was performed in another group of fetuses, they were

subjected to a 10-min period of cerebral hypoperfusion (reduced blood pressure

approximately 60%) and were sacrificed at 30 min (n=4) and 2 hr (n=3) after the onset of

hypotension. An extravascular balloon was put around the brachiocephalic artery after the

left chest was opened to produce cerebral hypoperfusion. All fetal brains were perfused

with heparinized phosphate-buffer saline followed by 4% paraformaldehyde (PFA). The

whole brain (including pituitary) was removed and immersed fixed in 4% PFA for a further

24 h at 4C. Then it was immersed in 70% ethanol with changes every 1-2 days until

processed. The brain was cut into cerebral cortex, cerebellum, hypothalamus,

hippocampus and brain stem in coronal sections and dehydrated with increasing

concentration of ethanol and xylene, and embedded in paraffin. 10 /um thick tissue

sections were cut and affixed on poly-L-lysine coated slides, incubated at 42C for 2 h,

then stored at -20C until processed.










Animals for the expression ofmRNA study


Intact and carotid sinus denervated fetuses (4 each group) were subjected to a 10-

min period of cerebral hypoperfusion (reduced blood pressure approximately 60%) and

were sacrificed 30 min after the onset ofhypotension. The fetal brain was quickly

removed and dissected rapidly into cerebral cortex, cerebellum, hypothalamus,

hippocampus, brain stem and pituitary, snap frozen immediately in a dry ice and acetone

bath, then stored at -800C.


Reverse Transcription and Polymerase Chain Reaction (RT-PCR) Quantification of
mRNAs


The levels ofPGHS mRNA in fetal brain tissues were determined by RT-PCR.

Total cellular RNA was extracted by a guanidine thiocyanate-phenol-chloroform method

(Chomczynski and Sacchi, 1987). Fetal brain tissues (100 mg) were homogenized in 2 ml

of 4 M guanidine thiocyanate denaturing solution, then extracted with phenol-chloroform.

The RNA was precipitated from aqueous phase by the addition ofisopropanol (3 ml) and

ethanol and pelleted by centrifugation at 10,000 g for 20 min. The quantity of total RNA

was measured by spectrophotometer and its integrity was assessed by 1% non-denaturing

agarose gel electrophoresis and visualization of the 18s and 28s ribosomal RNA bands

with ethidium bromide staining.

Total RNA (2 /g) was reversed transcribed to cDNA in 1 X buffer containing 200

U of superScript II reverse transcriptase, 0.5 mM dNTP, 0.01 M dithiothreitol, 40 U

RNase inhibitor and 0.5 ug of oligo(dT),5 primers in a total volume of 20 p1 according to










the manufacturer's protocol (GIBCO). Briefly, the mixture was incubated for 50 min at

42C, and the reaction was terminated by heating at 75C for 15 min. The cDNA sample

(3 /il) was amplified with specific primers for PGHS-1, PGHS-2 and p-actin. PCR

reactions were carried out in 1 X PCR buffer containing 1.5 mM MgCl2, 0.2 mM dNTP, 1

MM primers and 2 U of Taq DNA polymerase. The cycling program for PGHS-1 is 94C

for 1 min for denaturation, 56C for 1 min for annealing and 72C for 1 min for elongation

for 33 cycles; for PGHS-2 is 94C for 1 min, 61C for 1 min and 72C for 1 min for 30

cycles; for p-actin is 94C for 1 min, 56C for 1 min and 720C for 1 min for 30 cycles. The

PCR reaction products were separated by electrophoresis on 2% agarose gels. The gels

were stained with ethidium bromide and visualized under UV transillumination. The

density of PCR bands were analyzed using a densitometer and image analysis software

(Bio-Rad, Hercules, CA). The yield of the amplified product was tested to be linear for

amount of input RNA and PCR cycle number. For RT-PCR, the following optimized

conditions were chosen: 3 gl RT reaction/33 Cycles for PGHS-1 mRNA amplification, 3

,ul RT reaction/30 Cycles for PGHS-2 and 3 /l RT reaction/30 Cycles for p-actin.

PCR primers were selected from the published sheep cDNA sequences (Zhang et

al. 1996; DeWitt and Smith, 1988) with the aid of the PRIMER DESIGN software and

synthesized in ICBR at University of Florida. The expression of the PGHS-1 gene was

determined by amplification of a 618-bp region of the PGHS-1 cDNA sequence (245-

863). Using the sense upstream primer sequence 5'- CTTAGCCGCTACCAATGT-3' and

the antisense downstream primer sequence 5'-GAGCATCTGGTACTTCAGCTTC-3'.

For PGHS-2, expression of the gene was determined by amplification of a 881-bp region