The role of prostaglandins in the utero-ovarian axis of the cycling and early pregnant mare

MISSING IMAGE

Material Information

Title:
The role of prostaglandins in the utero-ovarian axis of the cycling and early pregnant mare
Physical Description:
xii, 146 leaves : ill. ; 28 cm.
Language:
English
Creator:
Vernon, Michael Walter, 1949-
Publication Date:

Subjects

Subjects / Keywords:
Horses -- Reproduction   ( lcsh )
Prostaglandins -- Physiological effect   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Includes bibliographical references (leaves 135-145).
Statement of Responsibility:
by Michael Walter Vernon.
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:
aleph - 000014042
notis - AAB7228
oclc - 06082501
System ID:
AA00003901:00001

Full Text









THE ROLE OF PROSTAGLANDINS IN THE UTERO-OVARIAN
AXIS OF THE CYCLING AND EARLY
PREGNANT MARE







By

MICHAEL WALTER VERNON















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














UNIVERSITY OF FLORIDA


1979






























Dedicated to

Matt and Ann,

my parents.














ACKNOWLEDGEMENTS


The author wishes to express his appreciation to his

major advisor, Dr. Daniel C. Sharp, and the members of his

graduate committee, Dr. Fuller W. Bazer, Dr. Donald Caton,

Dr. Michael J. Fields, and Dr. William W. Thatcher, for their

guidance and direction in the research and preparation of

this manuscript. The author is also indebted to the many

individuals who assisted in surgery and herd maintenance.

He is especially indebted to Dr. Richard L. Asquith, Michael

Zavy, Richard Mayer, Maryann Simonelli, Samuel Strauss, Kathy

and Matthew Seamans, and Thomas Wise. Appreciation is also

expressed to Dr. C.J. Wilcox for his statistical analyses;

Wiley Grubaugh and Debbie Starr for their technical assistance;

and Beverly Martin and Tamara Divito for their continuous

encouragement. Finally, the author wishes to express his

gratitude to Adele Koehler for her expert typing of this

manuscript.


iii
















TABLE OF CONTENTS


Page


ACKNOWLEDGEMENTS. .

LIST OF TABLES. .

LIST OF FIGURES .

ABSTRACT . .


CHAPTER


I INTRODUCTION . .

II LITERATURE REVIEW. . .

Control of Luteolysis. . .
The Uterus as a Source of the Luteolysin
PGF as the Luteolysin. . .
Steroid Modulation of PGF Production .
The Membrane Receptor for Prostaglandins
Mechanism of Action of PGF .

III THE ENDOMETRIAL PGF IN THE
INTACT PREGNANT AND CYCLING MARE .

Materials and Methods. . .
Test Animals. . .
Surgery . .
Incubation Procedure. .
Radioimmunoassay for PGF. .


Statistical Design .. .
Results and Discussion .


IV STEROID MODULATION OF EQUINE ENDOMETRIAL
PGF . . 52


Materials and Methods. .
Results and Discussion .


. 52
. 54


V THE LOCAL EFFECTS OF THE EMBRYO ON THE EQUINE
ENDOMETRIAL PGF . .


Materials and Methods. .
Results and Discussion .


. 66
. 67


iii1


. vi


. viii


S .


. 1


. 21


. 65


: : : :









Page

VI THE LOCALIZATION OF PGF IN THE PERFUSED
EQUINE OVARY. . . ... 73

Materials and Methods . ... 73
Results and Discussion. . .. .79

VII SPECIFIC BINDING OF PGF TO THE EQUINE
CORPUS LUTEUM . ... 85

Materials and Methods . .. 85
PGF Binding to Luteal Tissue .. .85
Receptor Characterization. ... 87
PGF Binding in Pregnant and Non-Pregnant
Mares. . . 90
Results and Discussion. ... ... 91

VIII SUMMARY AND CONCLUSIONS . ... .116

APPENDIX COMMERCIAL SOURCES OF MATERIALS CITED 134

LIST OF REFERENCES . . ... .135

BIOGRAPHICAL SKETCH. . . 146














LIST OF TABLES


Table Page

III-1. Analysis of Variance Table for the Endometrial
PGF Study in Non-Pregnant Mares. .. .32

III-2. Analysis of Variance Table for the Endometrial
PGF Study in Pregnant Mares .. 33

III-3. Endometrial PGF Concentrations in Non-Pregnant
Mares After In Vitro Incubation. .. .34

III-4. Endometrial PGF Concentrations in Pregnant
Mares After In Vitro Incubation. .. ... 35

IV-1. Analysis of Variance Table for the Experiment
on the Steroid Modulation of Endometrial PGF 55

IV-2. Effect of Exogenous and In Vitro Steroid
Treatments upon Endometral PGF Production
Capacities . .. 56

IV-3. Probabilities of a Greater F Statistic for
the Orthogonal Contrasts Between the Exogenous
Treatments . . 57

IV-4. Probabilities of a Greater F Statistic for
the Orthogonal Contrasts Between the In Vitro
Treatments . .. 58

V-l. Analysis of Variance Table for the Experiment
on the Embryonic Influence on Endometrial PGF. 69

V-2. Local Influence of Embryo upon Endometrial
PGF Production Capacities. . .. 70

VI-1. Analysis of Variance Table for the Experiment on
the % Localization of 3H-PGF in the Perfused
Equine Ovary . . 80

VI-2. Amount of 3H-PGF Localized Within Various
Tissues of the Perfused Ovary. ... 81

VII-1. Analysis of Variance Table for the Binding of
PGF to the Corpus Luteum of the Non-Pregnant
Mare . . 92








Table


VII-2. Analysis of Variance Table for the Binding of
PGF to the Corpus Luteum of the Pregnant


VII-3. Receptor Dissociation Constants and Binding
Capacities of Mare Corpora Lutea .. 99

VII-4. Relative Affinities of Various Prostaglandins
for the PGF Receptor . 101

VII-5. Weights of Corpora Lutea (CL), Percent PGF
Bound Specifically to Luteal Membrane Pre-
paration (MP), and Concentration of MP Pro-
tein/CL Throughout the Estrous Cycle and
Early Pregnancy . 102


vii


Page


Mare .


. 93














LIST OF FIGURES


Figure Page

III-1. Recovery of PGF2a added to endometria prior
to extraction. . ... 29

111-2. Production capacities of endometrial PGF
during the estrous cycle and early pregnancy 37

III-3. Regression curves for production capacities
of endometrial PGF during the estrous cycle
and early pregnancy. . ... 39

III-4. Endometrial PGF content during the estrous
cycle and early pregnancy. . ... 41

III-5. Regression curves for endometrial PGF content
during the estrous cycle and early pregnancy 43

III-6. Production capacities of endometrial PGF of
the estrous cycle after in vitro estrogen
and progesterone . ... 45

III-7. Regression curves for endometrial PGF pro-
duction after in vitro estrogen and
progesterone . ... 47

IV-1. Effects of exogenous and in vitro steroid
treatments upon endometriaT PGF production 61

V-l. Local influence of embryo upon endometrial
PGF production . ... 72

VI-1. Diagram of the apparatus utilized in the
perfusion of the equine ovary with tritiated
PGF . . 76

VI-2. Localization of tritiated PGF within various
tissues of the perfused equine ovary ... 84

VII-1. Listing of the chemical structures of the
PGF2 analogues used in the cross-reactivity
study. . . 89


viii









Figure Page

VII-2. (A) Rate of association of PGF to a Day 14 and
Day 8 luteal membrane preparation; (B) Rate of
dissociation of PGF from the luteal membrane
preparation; (C) Dose-response relationship
between the luteal membrane preparation, 3H-
PGF, and various amounts of radioinert PGF .95

VII-3. Scatchard plot of the interactions between
the luteal membrane preparation and various
amounts of 3H-PGF. . .. 97

VII-4. Weight of corpora lutea in non-pregnant and
early pregnant mares . 104

VII-5. Regression curves for the corpora lutea
weights from non-pregnant and pregnant mares 106

VII-6. Concentrations of protein in the luteal
membrane preparation from the non-pregnant
and pregnant mares . 110

VII-7. Regression curves for the concentrations of
protein within the luteal membrane prepara-
tion of the non-pregnant and pregnant mare 112

VII-8. Specific binding of PGF to the luteal cell
membrane preparation of non-pregnant and
early pregnant mare. . ... .114

VIII-1. Concentrations of estrone in the peripheral
plasma and uterine flushings of the non-
pregnant and pregnant mare ... .120

VIII-2. Concentrations of estradiol in the peripheral
plasma and uterine flushings of the non-
pregnant and pregnant mare . 122

VIII-3. Diagram of the proposed interrelationships
between the ovary and uterus of the non-
pregnant mare . . 129

VIII-4. Diagram of the proposed interrelationships
between the ovary and uterus of the pregnant
mare . . 132











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


THE ROLE OF PROSTAGLANDINS IN THE UTERO-OVARIAN
AXIS OF THE CYCLING AND EARLY
PREGNANT MARE

By

Michael Walter Vernon

June, 1979

Chairman: Dr. Daniel C. Sharp
Major Department: Animal Science

The interrelationships between the equine ovary and

uterine prostaglandins were studied. Since little is known

about the endogenous levels of prostaglandin F2a (PGF) in the

equine uterus, endometrial strips (% 300 mg) collected from

pregnant (P) and non-pregnant (NP) pony mares on various days

post-ovulation were incubated (2 hr @ 37 C). With PGF

quantified by radioimmunoassay, endometrial PGF content and

production were found to increase from the early stages of

the estrous cycle until Day 16 and then declined. Thus

maximal PGF concentrations occurred at a time that corresponds

to the expected time of luteolysis, i.e., Day 14. In P

mares, PGF content and production capacity increased during

the first 16 days of pregnancy in a pattern similar to that

of the NP animal; however, PGF continued to increase on Days

18 and 20. Since luteostasis is required for maintenance of

pregnancy, these latter findings appear to be contrary to

our concept of luteal physiology.








When ovariectomized mares were injected, daily, for 3

weeks with (50 pg) 17B-estradiol (E2), (50 mg) progesterone

(P4), or E2 plus P4 (1 week of E2 followed by 2 weeks of P4),

endometrial production of PGF was stimulated by all in vivo

treatments. Additionally, when endometria from these same

animals were incubated in vitro with 3 levels of E2 (1000,

10, 1 ng) and one level of P4 (1.0 pg), P4 was without effect

while E2 dramatically increased PGF production in a dose-

dependent fashion. In this manner, maximal PGF production

occurred in animals receiving exogenous (systemic) P4 and

in vitro E2.

Incubation of NP endometria collected on various days

post-ovulation, with (1.0 pg) E2 or (1.0 pg) P4, indicated

that in vitro E2 had little or no effect on the early stages

of the estrous cycle, but was stimulatory on Days 16 and 20,

while in vitro P4 was without effect on all days studied.

Thus, in vitro stimulation of E2 may be effective only on a

uterus that has had prior prolonged systemic P4 exposure.

The local effect of the embryo on PGF production was

studied by incubating endometria from gravid and non-gravid

horns of Day 12 and 18 NP mares. The PGF production in the

gravid horn of P mares was higher than the non-gravid horn,

while horn differences were not seen in NP mares. A local

embryonic influence on uterine PGF production is apparent.

Specific binding of PGF to CL from P and NP mares was

examined. Studies of the rates of association and dissocia-

tion indicated that 3H-PGF was bound specifically and








reversibly to a luteal cell membrane preparation (MP) that

was isolated by high speed (100,000 g) ultracentrifugation.

Cross reactivity studies implicated the 9a-OH and 5,6 cis

double bond as major contributors to PGF receptor recognition.

The membrane preparation appeared to contain at least two

receptor populations, having dissociation constants (Kd) of:

Kd1 = 4.19 x 10-11 [M]; and Kd2 = 9.11 x 10-10 [M]. Binding

of PGF to CL of the NP mare increased from Day 4 after ovula-

tion, to reach maximal levels by Day 12 and declined there-

after. In pregnancy the binding of PGF continued to increase

until Day 18, before it declined on Day 20. The reduction in

binding by Day 16 in the NP mare may reflect the process of

luteolysis, while high PGF binding capacity of CL between

Days 16 and 18 of pregnancy indicated that luteal maintenance

during pregnancy is not associated with a reduction of PGF

binding capacities.


xii














CHAPTER I

INTRODUCTION



The mare (Equus caballus) is a seasonally polyestrus

mammal that has an estrous cycle of approximately 22 to 24

days. In most cases, the cycle is comprised of 6 days of

estrus followed by 16 days of diestrus with luteal regression

occurring approximately 14 days after ovulation (Hafez, 1974).

At the time of luteolysis, peripheral progesterone levels

drop (Sharp & Black, 1973) and histological degeneration of

luteal cells is pronounced (van Niekerk et al., 1975). If

fertilization occurs, the process of luteolysis is averted

and the corpus luteum (CL) with its progesterone secretion

is retained (Hafez, 1974).

Uterine prostaglandin F2a (PGF) has been proposed as the

luteolytic factor of the mare. In support of this hypothesis,

it has been demonstrated that luteal regression coincides

with the time of maximal concentrations of PGF in the uterine

vein (Douglas & Ginther, 1976) and uterine lumen (Zavy et al.,

1978). Furthermore, exogenous PGF will induce premature

luteal regression (Douglas & Ginther, 1972; Ginther &

Meckley, 1972; Oxender et al., 1975) and hysterectomy will

delay luteolysis (Ginther & First, 1971).


-1-









The PGF content of the endometrium has also been shown

to be temporally related to the process of luteolysis in the

non-pregnant cow, ewe, mouse, pig, rat, monkey, and woman

(Chapter II). However, in the early stages of pregnancy

PGF content is high (Lewis et al., 1977; Carminati et al.,

1975). Since luteostasis is required for maintenance of

pregnancy, these latter findings appear to be contrary to

luteal physiology. In contrast, incubated homogenized Day 15

pregnant guinea-pig uteri have been shown to produce only

10% of the PGF that is produced by non-pregnant uteri (Walker

& Poyser, 1974).

The PGF content or production of the equine endometrium

has as of yet not been studied. This lack of information

along with the above perplexing findings of the pregnant

animal prompted some of the work presented in this disserta-

tion. Chapter III deals with experiments performed to deter-

mine the PGF content and production capabilities of equine

endometrium during the estrous cycle and early pregnancy.

The observation that PGF was secreted in a pattern that

was temporally associated with the reproductive status of the

animal implied that uterine PGF production was under the

control of extra-uterine factors. It is now apparent that

the ovarian steroids, estrogen and progesterone, enhance the

ability of the endometrium to produce PGF (Castracane &

Jordan, 1975). Estrogen has been demonstrated to increase

the concentration of PGF within the uterine flushings of the

progesterone primed uterus of the pig (Frank et al., 1978)









and cow (W. Thatcher, G. Lewis, & F. Bartol, personal communi-

cation). To further assess the modulating effect of estrogen

and progesterone, a study was conducted to determine the

effects of exogenous (systemic) and in vitro steroids

on the ability of the equine endometrium to produce PGF

(Chapter IV).

In the pregnant animal, the high levels of PGF contained

within the endometrium may be a reflection of endogenous

steroid enhancement, since this is a time when both systemic

luteal progesterone (Allen & Hadley, 1974) and local fetal

estrogen are high. In our laboratory, the incubated equine

concepts has been shown to produce estrone and estradiol

(Mayer et al., 1977) and to possibly convert 3H-progesterone into

3H-estrone (Seamans et al., 1979). Likewise, Flood and

Marrable (1975) have observed, histochemically, the presence

of a variety of hydroxysteroid dehydrogenase (HSD) enzymes

within the Day 13 equine concepts. But of particular in-

terest was the appearance of high levels of 178-HSD only on

those parts of the uterine endometrium directly opposed to

the trophoblast (Flood et al., 1979). This is suggestive of

a local concepts influence on the endometrium. Since the

PGF production capacity of the equine endometrium may be

locally enhanced by fetal estrogens, the PGF production of

the endometrium associated with the concepts was compared

with the PGF production of the endometrium contralateral to

the concepts (Chapter V).








The PGF induced luteal regression probably involves a

direct biochemical interaction with luteal cells. This has

been supported by the observation that PGF stimulates the

membrane-associated adenyl cyclase system (Marsh, 1971) and

that the addition of cyclic adenosine monophosphate to cul-

tures of porcine (Channing & Seymour, 1970) and monkey

(Channing, 1970) granulosa cells stimulates progesterone

production. Although the mechanism of action of PGF is not

well understood, the involvement of the adenyl cyclase system

suggests that the mode of action of PGF is mediated through

a membrane receptor. Luteal PGF membrane receptors have been

isolated and biochemically characterized in the cow (Kimball

& Lauderdale, 1975; Rao, 1976), ewe (Powell et al., 1974),

mare (Kimball & Wyngarden, 1977), and woman (Rao, 1977). It

is likely that the PGF membrane receptor may play a role in

luteal regression in the cycling mare, or in prevention of

luteal demise in the pregnant mare. Therefore, studies were

conducted to characterize the equine luteal PGF receptor and

to assess its possible role in the cycling and early pregnant

mare (Chapters VI & VII).














CHAPTER II

LITERATURE REVIEW



Control of Luteolysis


Control of luteolysis was first thought to reside solely

within the adenohypophysis. Nalbandov (1961) set forth the

theory that the corpus luteum (CL) appeared to have a limited

life expectancy that was extended (rescued) in pregnancy by

a pituitary luteotrophin. The major experimental evidence

for this theory was the demonstration that an injection of

progesterone into the pregnant guinea-pig was effective in

reducing the size of the CL only if administered within one

day of breeding. Progesterone injection on any other day

post-mating had no effect on CL size. Assuming that pro-

gesterone had a negative feedback on the pituitary, he pro-

posed that "there is discharge of luteotropin hormone over

a relatively short period of time, lasting in pigs or guinea-

pigs not longer than 2 to 3 days after ovulation. This single

release is sufficient to maintain the CL for their normal

lifespan during the cycle. No further release of hypophyseal

luteotropin occurs unless the female becomes pregnant"

(Nalbandov, 1961; p. 138).






-6-


Short (1964) postulated that luteal regression in the

non-pregnant animal was the result of a uterine luteolysin.

He indicated that Nalbandov's theory did not explain the

luteotropic effect of hysterectomy (see next section) nor

did it explain the results of an experiment in ewes by Innskeep

and co-workers (1962) in which CL of different age on the

same ovary regressed in unison at the expected time of luteo-

lysis. Short championed a hypothesis, on scant experimenta-

tion, that has since been shown to be correct. He stated

that "the progesterone secreted by the corpus luteum could

stimulate the uterus, for example the uterine glands, to

secrete a luteolysin. This substance would be liberated into

the circulation and would have a direct effect on the ovary,

suppressing the activity of the corpus luteum" (Short, 1964;

p. 329). Thus control of luteolysis was now felt by some to

be under uterine control.



The Uterus as a Source of the Luteolysin


A relationship between the uterus and CL was first

realized when Leo Loeb (1923) reported that extirpation of the

guinea-pig's uterus increased the lifespan of the CL from 15

days to 60 to 80 days. Since that time, the luteostatic

response of hysterectomy (HYX) has been observed in the rat

(Durrant, 1926), mouse (Anderson et al., 1969), sheep

(Caldwell & Moor, 1971), rabbit (Asdell & Hammond, 1933),

cow (Wiltbank & Casida, 1956), pig (du Mesnil du Buisson,








1966), and horse (Ginther, 1967), but not in the monkey or

human (Wiqvist et al., 1970). Furthermore, uterine trans-

plants to ectopic sites can, at least partially, reverse the

effect of HYX (Caldwell, 1970). In the case of the mare,

HYX on the second day after the end of an estrus period-

maintained CL weight at 3300 mg on Day 30 post-surgery, while

sham HYX mares had a significantly lower luteal weight of

309 mg on Day 30 post-surgery (Ginther & First, 1971).

In nearly all species studied, a unilateral, i.e. local,

relationship has been demonstrated between the uterus and

the CL. After a unilateral hysterectomy (UHYX), luteal re-

gression was observed only when the UHYX was contralateral to

the CL and vascular continuity between the uterus and CL was

unbroken (Ginther, 1967). However, UHYX of the rabbit

(Ginther, 1967) and mare (Ginther & First, 1971) was in-

effective in altering the lifespan of the CL. When mares

were UHYX or HYX on the second day after estrus mean weights

of luteal tissue, 30 days later, in the HYX, UHYX ipsilateral

to CL, UHYX contralateral to CL,and sham HYX were 3337 mg,

2237 mg, 3006 mg, and 503 mg, respectively (Ginther & First,

1971). Since there was no significant difference in the

response of CL to uterine surgery, the uterine influence was

thought to follow a systemic vascular route.

Denamur and co-workers (1966) were able to prolong

secretion of ovine luteal progesterone by HYX at mid-cycle.

However, subsequent hypophysectomy on Day 20 post-ovulation








resulted in a rapid decline in progesterone secretion,

commencing 48 hr after the operation (Denamur et al., 1966).

Thus in the case of the HYX ewe, maintenance of the CL re-

quires daily, basal output of pituitary luteotropin. However,

this luteotropin plays a permissive role since luteal re-

gression can be evoked at any time with exogenous PGF (see

next section).



PGF as the Luteolysin


Luteolytic substances of uterine origin have been shown

to influence CL function in HYX animals. Infusion of re-

constituted freeze-dried uterine venous plasma (Caldwell &

Moor, 1971) and ether-soluble uterine endometrial extracts

(Kiracobe et al., 1966) depressed luteal function in the HYX

ewe, while porcine uterine flushings were shown to have a

cytolytic action upon cultured granulosa cells (Shomberg,

1967).

Crude cell suspensions and aqueous extracts of bovine

endometria collected on Days 10 to 13 of the estrous cycle

were shown to reduce the luteal weight and progesterone

content of hamster CL (Lukaszewska & Hansel, 1970). Most of

the luteolytic activity of the extract was precipitated in

a 55% ammonium sulfate fraction. Further fractionation of

the precipitate on columns of Sephadex G-100 yielded an

active fraction having a partition coefficient of 0.271.

At this time Lukaszewska and Hansel felt, "the results suggest





-9-


that the active factor is a large molecular weight protein,

or some smaller molecule bound to protein" (Lukaszewska &

Hansel, 1970; p. 261). Hansel, Concannon, and Lukaszewska

(1973) later reported that the 55% ammonium sulfate fraction

contained some lipids and that an ethanolic extract of the

precipitate was as effective as the entire precipitate in

reducing luteal weight and progesterone content of pseudopregnant

hamsterCL. However, further partitioning of the ethanolic

extract with thin layer chromatography suggested that only

the fraction containing free fatty acids was luteolytic. The

authors indicated that "the bovine endometrium might exert

its luteolytic effect by providing the CL with one (arachi-

donic acid) or more precursors which are converted into

prostaglandins, or other luteolysin in situ by the luteal

tissue" (Hansel et al., 1973; p. 239). This was further

supported by the finding that arachidonic acid given to

cattle on Day 12 and 13 of the estrous cycle was luteolytic

(Shemesh et al., 1974).

The uterus has been shown to contain large amounts of

prostaglandins (Pickles, 1959) and in 1969 Pharriss and

Wyngarden (1969) were the first to report that exogenous PGF

lowered the progesterone content of pseudopregnant rat

ovaries. Exogenous PGF was luteolytic in the rabbit (Gutknecht

et al., 1966), guinea-pig (Blatchley & Donovan, 1972), mouse

(Saksena & Lau, 1973), gerbil (Chaichareon et al., 1974),

hamster (Johnson & Hunter, 1970), cow (Lauderdale, 1972), pig

(Moeljono et al., 1977), sheep (McCracken et al., 1970), mare





-10-


(Douglas & Ginther, 1972), and monkey (Kirton et al., 1970)

but not in the human (Wiqvist et al., 1970).

The luteolytic effect of exogenous PGF in the mare was

first demonstrated by Douglas and Ginther in 1972. All mares

that were subcutaneously injected with PGF returned to estrus

3 to 4 days after treatment (Douglas & Ginther, 1972).

Since 1972, PGF was found to be luteolytic in the mare when

administered intravenously (Douglas et al., 1976), intra-

muscularly (Spincemaille et al., 1975), intrauterine or

intraluteal (Douglas & Ginther, 1975). Furthermore, PGF was

luteolytic if given to cycling (Douglas & Ginther, 1972),

pregnant or pseudopregnant mares (Kooistra & Ginther, 1976).

However, the early CL (< Day 5) of the non-pregnant mare was

refractory to exogenous PGF (Douglas & Ginther, 1972;

Ginther & Meckley, 1972; Oxender et al., 1975). The lowest

effective dose for a luteolytic effect in the non-pregnant

mare was determined to be 1.25 mg or 8.8 pg/kg. This was

not expected, since on a body weight basis, considerably

higher subcutaneous doses are required to induce luteolysis

in other species (144 pg/kg for sheep, Douglas & Ginther,

1973; 30 mg for cattle, Lauderdale, 1972). This high sensi-

tivity is further perplexed by the apparent lack of a local

transfer (counter-current) of PGF between the equine uterus

and ovary (Ginther & First, 1971).

If uterine PGF is the luteolytic agent, its secretary

patterns should be temporally associated with luteal re-

gression. The concentration of PGF in the uterine venous





-11-


drainage of the non-pregnant sheep (Bland et al., 1971), sow

(Gleeson et al., 1974), cow (Shemesh & Hansel, 1975), rat

(Saksena & Harper, 1972a), guinea-pig (Blatchley et al.,

1972), and mare (Douglas & Ginther, 1976) are maximal at the

expected time of luteolysis.

In the mare, luteal regression occurs approximately 14

days after ovulation. At this time, peripheral progesterone

levels drop (Sharp & Black, 1973; Douglas & Ginther, 1976)

and the histological degeneration of the luteal cell is pro-

nounced (van Niekerk et al., 1975). Douglas & Ginther (1976)

demonstrated that uterine venous concentration of PGF in the

anesthetized non-pregnant mare rose from 1.4 ng/ml at the

time of ovulation to a maximal concentration of 14.9 ng/ml

on Day 14 post-ovulation and declined to 6.4 ng/ml on Day 18.

In a similar fashion, Zavy and co-workers (1978) indicated

that the total amount of PGF within the uterine lumen of the

unanesthetized mare rose from 52.19 ng on Day 4 to a maximal

concentration on Day 14 of 1,133.81 ng and declined to

244.06 ng by Day 20. In the early pregnant (anesthetized)

mare, the uterine venous concentration of PGF was lower than

that found in the non-pregnant animal (Day 10 = 4.3, Day

14 = 9.3, and Day 18 = 7.3 ng/ml; Douglas & Ginther, 1976).

Thus at a time of luteal maintenance the uterine output of

PGF appears to be low. This has also been observed in the ewe

(Barcikowski et al., 1974) and pig (Moeljono et al., 1977).

The PGF content of the uterine endometrium has also been

shown to be temporally related to the time of luteolysis





-12-


in the non-pregnant cow (Shemesh & Hansel, 1975), ewe

(Wilson et al., 1972), mouse (Saksena et al., 1974), rat

(Saksena & Harper, 1972a), monkey (Demers et al., 1974), and

woman (Singh et al., 1975); while, the endometrial PGF con-

tent in the ewe (Lewis et al., 1977) and rat (Carminati

et al., 1975) increased during the early stages of pregnancy.

Since luteostasis is required for maintenance of pregnancy,

these latter findings appear to be contrary to in vivo

physiology. In contrast, Walker and Poyser (1974) incubated

homogenized Day 15 pregnant and non-pregnant guinea-pig uteri

and demonstrated that the production of PGF in the pregnant

animal was lower than the non-pregnant (15 ng vs 110 ng/mg

homogenate, pregnant vs non-pregnant). In Chapter III,

experiments are described which were performed to determine

the PGF content and production capabilities of the equine

endometrium during the estrous cycle and early pregnancy.

A possible explanation for the high PGF levels within

the pregnant uterus has been set forth by Bazer & Thatcher

(1977). Based on the secretary patterns of PGF (Moeljono

et al., 1977; Frank et al., 1978) and purple protein (Chen

et al., 1975) by porcine endometrium, they have suggested

that the embryo influences the direction of movement of

uterine secretions. In pregnancy, uterine secretary products,

e.g., PGF, move toward the uterine lumen exocrinee secretion);

while, in the absence of embryonic influence, uterine secre-

tions proceed toward the endometrial stroma and associated

vascular system (endocrine secretion) (Bazer & Thatcher,





-13-


1977). In this fashion, uterine PGF production may be

sequestered within the uterus during pregnancy and may never

reach the CL. The experimental evidence for this hypothesis

was derived primarily from the following three observations:

1) The uterine venous concentrations of PGF in the pregnant

and the estrogen induced pseudopregnant giltwere lower than

those of the non-pregnant animal; while, the uterine luminal

concentration of PGF and porcine purple glycoprotein was

higher in the pregnant and pseudopregnant sow than in the

non-pregnant animal (Moeljono et al., 1976; Moeljono et al.,

1977; Frank et al., 1978); 2) the porcine embryo secretes

estrogen (Perry et al., 1973; Perry et al., 1976); and

3) immunofluorescent studies have indicated that the porcine

purple glycoprotein of the endometrium (Chen et al., 1975)

and the PGF of the primate oviduct (Ogra et al., 1974) were

secreted in a luminal direction when under the in vivo in-

fluence of pregnancy and in a vascular direction in its

absence.



Steroid Modulation of PGF Production


The possibility of steroids controlling endometrial PGF

was first suggested in 1972 when injections (10 pg, S.C.) of

estradiol benzoate into the guinea-pig (Day 4 to 6 of the

estrous cycle) dramatically increased the concentration of

PGF in the utero-ovarian vein (Blatchley et al., 1972).

These findings immediately prompted studies in the ewe





-14-


(Caldwell et al., 1972), hamster (Saksena & Harper, 1972b),

and rat (Castracane & Jordan, 1975) demonstrating that

exogenous estrogen and progesterone were both stimulatory

to PGF production and also that estrogen was a more potent

stimulator than progesterone. Caldwell postulated the possi-

bility that in the case of the ewe, "each estrous cycle is

terminated due to the rise in estradiol (on Day 13) which

causes a release of PGF2a from the uterus which in turn

causes luteal regression and the loss of progesterone"

(Caldwell et al., 1972; p. 226). In the case of the ovariec-

tomized rat, Castracane and Jordan (1975) have shown a posi-

tive synergism between estrogen and progesterone. Optimal

conditions for PGF biosynthesis occurred when 1.0 ig of

estradiol was given to animals that had been pretreated for

two days with progesterone (2 mg/day). Similarly, estrogen

has been shown to increase the concentration of PGF within

the uterine flushings of the luteal phase pig (Frank et al.,

1978) and cow (W. Thatcher, G. Lewis, & F. Bartol, personal

communication). No research data are available on the

interrelationships between ovarian steroids and PGF produc-

tion by equine endometrium. This lack of information was

the stimulus for a portion of this dissertation (Chapter IV).



The Membrane Receptor for Prostaglandins


Kuehl and co-workers (1970) at the Merck Institute fo.r

Therapeutic Research demonstrated a distinct dose-dependent





-15-


relationship between prostaglandins (PG) and cyclic adenosine

monophosphate (AMP) formation in isolated mouse ovary. This

finding propagated a series of experiments that they had

hoped would be the basis for a PG radio-ligand assay. These

binding studies with the mouse ovary were unsuccessful

(Kuehl & Humes, 1972); therefore, the Merck group initiated

experiments on the rat lipocyte. In 1972, competition studies

with PGE -3H demonstrated the presence of a PG receptor on

the rat adipocyte (1000 G) membrane (Kuehl & Humes, 1972).

The affinity of PGE for the adipocyte receptor was greater

than that of PGF and PGA; an observation consistent with their

relative potencies in stimulating cyclic AMP. The dissocia-

tion constants (Kd) for PGE1 and PGE2 were determined by
9
Lineweaver-Burk plots to be 3 x 109 [M].

At this same time, Marsh at the University of Miami

demonstrated that, in the case of the bovine CL, PGs also

stimulated the adenyl cyclase system. Prostaglandins E1 and

E2 elicited a 130% increase in cyclic AMP formation (relative

to controls) and PGF2a a 40% increase (Marsh, 1971). With

the inference that PG interacted with the membrane-associated

adenyl cyclase system, a search was initiated to characterize

a luteal membrane receptor for PG. In 1974, Rao, of the

University of Louisville, isolated and characterized a PGE

receptor in the cow (Rao, 1974) and Powell and co-workers

from Sweden presented evidence for the existence of an ovine

PGF luteal receptor (Powell et al., 1974). Since that time,

a PGF receptor has been found on luteal membranes from the





-16-


cow (Kimball & Lauderdale, 1975; Rao, 1976), woman (Rao,

1977), and mare (Kimball & Wyngarden, 1977).

Scatchard analysis (Scatchard, 1949) of the PGF bovine

CL receptor indicated a heterogenic population of receptors

having Kds of 1.3 x 10-9 [M] and 1.0 x 10-8 [M] (Rao, 1974).

However, a single homogenous receptor population was shown
8
to exist in the cow (Kd = 2.1 x 10-8 [M]; Kimball & Lauder-

dale, 1975), sheep (Kd = 1 x 10-7 [M]; Powell et al., 1974),

and mare (Kd = 3.2 x 10-9 [M]; Kimball & Wyngarden, 1977).

The equine membrane receptor preparation described by

Kimball and Wyngarden (1977) required 1.5 to 2 hrs to reach

equilibrium with PGF at 37 C and was displaceable, in a dose-

dependent fashion with radioinert PGF. Competition studies

with several natural PGs for the PGF receptor indicated

specificity for the 9a-hydroxyl moiety and the 5,6-cis double

bond. However, the specific binding of PGF in luteal tissue

collected on Days (n), 7(2), 8(2), 10(1), 11(1), and 15(1)

of the estrous cycle was determined to be devoid of any

physiological patterns (51258 & 52252, 23629 & 38159, 25548,

15317, & 30585 DPM/mg membrane protein for the above days,

respectively). As a part of this dissertation, a study was

conducted to further characterize the equine luteal PGF

receptor and to assess its possible physiological role in

the cycling and early pregnant mare (Chapter VII).

In addition to binding to plasma membranes, PGF has

recently been demonstrated to bind to several cytoplasmic

inclusions. Utilizing differential and discontinuous sucrose





-17-


gradient centrifugation, PGF was found to bind specifically

to the nuclear, mitochondrial, microsomal, cytosol, and

membrane fractions (4.4, 32.4, 18.9, 5.1, and 192.5 fmol/mg

protein, respectively) (Rao & Mitra, 1977). Since the total

mitochondrial fraction contained both mitochondria and

lysosomes, it was subfractionated and only the lysosomal

enriched fraction bound PGF (Rao & Mitra, 1978). Lysosomal

PGF binding may be interpreted as further evidence for the

role of the lysosome in hormone action (Szego, 1974) or

internalization and catabolism of the membrane receptor within

the lysosome (Chen et al., 1978).



Mechanism of Action of PGF


Although the mechanism of action of the luteolytic

effect of PGF is not well understood, the presence of a

membrane receptor and the involvement of the adenyl cyclase

system suggests that the primary site of action of PGF is at

the luteal cell level. However, some evidence is available

to indicate that PGF may exert its luteolytic effect via

vasoconstriction (Pharriss, 1970) or hypophyseal interaction

(Labsetwar, 1970).

The possibility that PGF may induce luteolysis by re-

stricting the oxygen and nutritive supply to the rat and

rabbit ovary was demonstrated by Pharriss (1970) when a

single intravenous injection of PGF elicited a rapid drop

in ovarian vein drainage. Luteolysis, through anoxia, was





-18-


also indirectly supported by the initial observation that

in vitro PGF stimulated luteal progesterone production

(Behrman et al., 1971; Speroff & Ramwell, 1970). But con-

trary results indicating in vitro PGF inhibition of proges-

terone production (O'Grady et al., 1972; Henderson & McNatty,

1975) and no vascular effect (McCracken et al., 1971) have

severely weakened the argument. Furthermore, Janson et al. (1975)

has indicated that the vascular effect may be an artifact

of low systemic arterial pressure. Thus the lower venous

blood pressure observed by some may just be a result of a

hormonal lowering of the systemic pressure.

Another appealing theory was the possibility of an inter-

relationship between PGF and the anterior pituitary. A poten-

tial role for the equine pituitary was implicated when Pineda

and co-workers (1972) induced CL regression with an antisera

against the equine pituitary. However, the role of the equine

pituitary appears to be permissive since Garcia and Ginther

(personal communication) have indicated that the luteotropic

action of human chorionic gonadotropin (IICG) cannot override

the luteolytic action of PGF. The daily injection of 1500 IU

of HCG from Day 9 to Day 17 lengthened the lifespan of the

CL by four days. This luteotropic response to HCG was in-

effective in altering PGF induced CL regression. In the

rat, PGF has been shown to increase the content of pituitary

LH (Labsetwar, 1970) and LH has been demonstrated to be

luteolytic in rats (Leavitt & Acheson, 1972), rabbits





-19-


(Stormshak & Casida, 1965), and hamsters (Greenwald, 1967).

However, this positive feedback system has not been observed

in any of the domestic animals (Cerini et al., 1972).

Therefore, PGF induced luteal regression probably in-

volves a direct biochemical reaction with the luteal cell.

As previously stated, this is supported by the presence of

the PGF receptor and the involvement of the membrane-

associated adenyl cyclase system. Furthermore, the addition

of cyclic AMP to cultures of porcine (Channing & Seymour,

1970) and monkey (Channing, 1970) luteinized granulosa cells

stimulated progesterone production. Henderson and McNatty

(1975) have suggested a biochemical mechanism of action for

the luteolytic process. Using the membrane model of Garren

et al. (1971), they viewed the membrane as a dynamic mosaic

(Singer & Nicholson, 1972), containing regulatory, coupling,

and catalytic units. They postulate that binding of LH to

the regulatory unit enhances production of cyclic AMP, which

in turn stimulates the protein kinase activity of the

catalytic unit. The kinase then phosphorylates and activates

cholesterol esterase to stimulate progesterone production.

During luteolysis, PGF would bind to the membrane (coupling

unit) and prevent the LH regulatory unit from activating the

catalytic unit and thus halt the biochemical cascade leading

to progesterone production (Henderson & McNatty, 1975). The

theory, however, is not compatible with the findings of

Marsh (1971) that PGF stimulates in vitro cyclic AMP produc-

tion by the CL. Henderson and McNatty (1975) also offer an





-20-


explanation for the inability of the early CL to respond

to PGF (Douglas & Ginther, 1972). After the ovulatory LH

surge, the luteal membrane may be saturated with LH and this

high saturation may induce conformational change in the

membrane that interferes with luteal PGF uptake. Due to the

dynamic equilibrium of a receptor, LH could slowly dissociate

from the membrane and facilitate subsequent PGF uptake and

regression (Henderson & McNatty, 1975). The authors support

this second hypothesis with the findings of Hichens and co-

workers (1974) where PGF was demonstrated to decrease the

HCG binding capacity of luteal tissue, "possibly by inducing

conformational changes" (Henderson & McNatty, 1975; p. 791).

However, final proof of this hypothesis awaits experimental

evidence indicating that LH interferes with PGF binding.

Rao (1974) indicated that neither HCG and PGE1, nor PGE and

PGF compete with each other for their respective receptors.














CHAPTER III

THE ENDOMETRIAL PGF IN THE
INTACT PREGNANT AND CYCLING MARE



Prostaglandin F2a (PGF) has been proposed as the uterine

luteolytic substance in the mare (Chapter II). Maximal

levels of PGF occur in the uterine lumen (Zavy et al., 1978)

and uterine venous drainage (Douglas & Ginther, 1976) at the

expected time of luteolysis. Since little is known about the

endogenous levels of PGF in the equine endometrium, this

study was conducted to evaluate PGF content and production

capabilities during the estrous cycle and early pregnancy.



Materials and Methods


Test Animals


Tissues for this experiment were obtained from 41 re-

productively sound pony mares of mixed breeding, having a

mean weight at the time of experimentation, of 175.6 38.7

(X S.D.) kg (range 86-306 kg). During the course of the

experiment, the animals were maintained ad libitum on pasture

and supplemental hay at the Horse Research Center in Lowell,

Florida.


-21-






-22-


The animals used in the non-pregnancy study were derived

from a herd of 68 mares in the summer of 1977 and, in addition

to the following endometrial study, were also involved in a

luteal PGF2a binding study (Chapter VII). The day of ovula-

tion (Day 0) was determined by daily teasing and palpation

per rectum and on the day of ovulation the mares were assigned

randomly for surgery on either Day 4, 8, 12, 16, or 20 post-

ovulation. A total of 20 non-pregnant mares, four per day,

was used.

The animals used in the pregnancy study were derived

from a herd of 42 mares in the summer of 1976 which, in

addition to being in this study and the luteal PGF binding

study (Chapter VII), were also involved in studies concerned

with uterine secretions by Michael T. Zavy and studies of

embryonic steroidogenic capabilities by Richard E. Mayer.

Animals were bred once daily if their ovaries contained a

follicle of ovulatory size (35-60 mm diameter). When

possible, mating was natural; however, artificial insemina-

tion was utilized if there was a shortage of available

stallions. On the day of ovulation (Day 0), mares were

randomly assigned for surgery on either Day 4, 8, 12, 14, 16,

18, or 20 post-ovulation with pregnancy confirmed by the

presence of an embryo at surgery. Conception rate at the

termination of the study was 76%. As determined at surgery,

three mares were assigned to Days 8, 12, and 14, and four mares

to Days 14, 18, and 20 of gestation. No embryos were dis-

covered in the oviducts or uteri of Day 4 bred mares





-23-


and therefore this day was not included in the pregnancy

study.



Surgery


To reduce bowel contents and motility during surgery,

solid food was restricted 24 hr prior to surgery. On the day

of surgery, the animals were weighed and tranquilized with

Acepromazine (see Appendix I for listing of commercial sup-

pliers) (% 4 mg/45 kg body weight). General anesthesia was

induced with a fast-acting barbiturate (1 g Pentothal/150 kg

body weight) and 5% fluothane gas at a flow rate of 4 to 5

1/min. After the proper plane of anesthesia was attained

(2-10 min), general anesthesia was maintained for the re-

mainder of the surgery at 1 to 2% halothane gas, at a flow

rate of 1-3 1/min. The animals were then placed in dorsal

recumbency and the surgical area was shaved and scrubbed with

an iodine solution (Betadine). In the pregnancy study a com-

plete ovario-hysterectomy was performed via a mid-saggital

incision (% 150 mm) in the abdomen; while, in the non-

pregnancy study entry was via a para-lumbar incision and only

the ovulatory ovary and a segment of the adjacent uterine

horn were ablated. The removed tissue was placed on ice and

brought immediately to the laboratory for subsequent pro-

cessing. After the incision area was closed, the wound was

sprayed with a topical sulfa drug and protected temporarily

with surgical tape. Each animal then received 1,000,000 IU of





-24-


penicillin and, after recovery, was placed in a post-operative

stall for observation.



Incubation Procedure


In the pregnancy study, 60 ml of hypertonic saline (0.33

M) were injected into the extirpated uterus at the uterotubal

junction and the embryo and uterine secretions were flushed

through an incision made in the uterine horn. As explained

before, the embryos, uterine secretions, and ovaries were

used in separate studies.

The uterus was cut along its longitudinal axis and the

endometrium was dissected bluntly from the myometrium and

placed into a flask containing chilled (4 C) phosphate buf-

fered saline (PBS) (pH = 7.4). The endometrium was then cut

into strips, weighed (n 300 mg/strip), placed into 25 ml

Erlenmeyer flasks containing PBS and kept on ice until incu-

bation. Endometrial strips from pregnant animals were

derived from the body of the uterus, while the strips from

non-pregnant animals were derived from the medial portion

of the uterine horn, ipsilateral to the ovulatory ovary.

After endometrial strips were collected, two strips each

were transferred to glass test tubes containing 5.0 ml of

Krebs-Ringer Carbonate-Bicarbonate buffer (KRB) and frozen

immediately. The buffer system was constructed according to

the procedure outlined by Umbreit and co-workers (1957).

Unincubated samples were assayed by radioimmunoassay at the





-25-


termination of the experiment for PGF concentration in the

endometrium prior to incubation, i.e., endogenous PGF

content. Two additional strips each were transferred to

25 ml Erlenmeyer flasks containing 5.0 ml KRB. Serum bottle

caps were placed on the flasks containing endometrium and

KRB and the flasks were flushed for 30 sec with 95% oxygen-

5% carbon dioxide. The flushing step was accomplished by

piercing the serum cap with two 18 gauge hypodermic needles,

using one needle as a gas inlet and the second as a vent.

After flushing, the needles were removed and the flasks were

placed in an incubator-shaker (@ 37 C & 1 stroke/sec). The

endometrium was incubated for 2 hr and then placed into glass

test tubes and frozen until it was radioimmunoassayed for

PGF. The concentration of endometrial PGF within these

samples represents the total amount of PGF present after

the two hour incubation (i.e., total production & endogenous

content). In addition to the two treatments just described,

endometrial strips from non-pregnant animals were also in-

cubated, in duplicate, in media containing 1.0 pg estradiol-

178 or 1.0 ug progesterone. To construct the steroid treat-

ments, steroids in a 95% ethanolic solution were added

directly to the 25 ml incubation flasks, dried, and re-

suspended in 5.0 ml KRB.


Radioimmunoassay for PGF

The concentration of PGF in the endometrial incubation

was measured with a double antibody radioimmunoassay. The





-26-


first antibody was donated by Dr. Kenneth T. Kirton of the

Upjohn Company, Kalamazoo, Michigan, and was generated in

rabbits by subcutaneous injections of PGF2a conjugated to

bovine serum albumin at the C1 carbon. The second antibody,

purchased from Cappel Laboratories, Inc., was generated in

goats against light and heavy IgG chains purified from rabbit

serum. Radioactive PGF2a (NET-345, Lot 932-121) was pur-

chased from New England Nuclear. The PGF2a molecule was

labelled at the C9 position with tritium and had a specific

activity of 10.9 Ci/mmol. The radioinert PGF2a was a gift

from Dr. John E. Pike of the Upjohn Company.

Measurement of PGF in the endometrial incubation was

performed after tissues from both pregnant and non-pregnant

animals were collected. Samples were thawed at ambient

temperature (23 C) and homogenized in a Potter-type hand

homogenizer. The homogenizations and all subsequent pro-

cedures were conducted at 4 C with ice baths and mechanical

refrigeration. The homogenate (100 pl) was transferred to

a test tube containing 900 pl of absolute ethanol and vortexed

for 5 sec. The mixture was then centrifuged at 1086 G for

10 min. A portion (100 pl) of the resulting supernatant was

transferred into 10 mm x 75 mm culture tubes, dried, and

resuspended in 100 pl of Tris-HCL buffer (pH = 8.0). Usually

32 endometrial extracts (unknowns) were run in duplicate,

between two standard curves which were composed of ten

reference points ranging from 10 pg to 10 ng of authentic






-27-


PGF2a. The radioimmunoassay procedure, described elsewhere

(Moeljono, 1975), required an incubation period of 48 hr at

4 C. At the end of the incubation, the test tubes were

centrifuged at 1086 G for 20 min and 500 pl of the resulting

supernatant were counted for 5 min in 4.0 ml cocktail (3.9 g

PPO, 0.1 g POPOP/1 toluene-20% triton X) in a Beckman LS-300

liquid scintillation spectrophotometer.

The concentration of PGF in the unknowns was calculated

from the equation of the line derived from a logit transfor-

mation of the standard curve values. The equation for the

line of the standard curve was calculated by a least squares

regression program on a Monroe 1860 calculator. The con-

centration of PGF in the ethanolic extract was corrected for

procedural dilutions and endometrial mass, and expressed as

ng PGF/mg endometrium.

The radioimmunoassay has been previously validated for the

measurement of PGF in porcine plasma (Moeljono, 1975; Moeljono

et al., 1977) and equine uterine secretions (Zavy et al., 1978).

For the purpose of validating the assay for the endometrial

extract, increasing concentrations of authentic PGF2a were added

to an endometrial homogenate (range 1 to 50 ng) and processed

as described previously. The amount of PGF recovered in the

assay (after subtraction of the 0 tube) was graphed against

the amount of PGF added to the homogenate (Fig. III-1) and

the resulting line verified the accuracy of the extraction

procedure and also indicated the possible absence of any

extraneous endometrial competitors. During the course of the





























Fig. III-1. Recovery of PGF2a added to endometrial
homogenates prior to extraction.






-29-


R =0.9996

Y 0.02+1.021X


OJ 1.0 2.5 5.0 10

AMOUNT ADDED (ng)


S100
w
w
0
C.,
w
cr 10



0
I i





-30-


assay, the interassay coefficient of variability was 11.63%

(n = 7 @ 500 pg) and the intrassay coefficient of variability

was 5.89% (n = 6 @ 500 pg). Since the recovery of radio-

active PGF2a in this extraction system was found to be

102.99 t 1.2% (X t S.D.), no corrections were made for pro-

cedural losses. The specificity of the assay (Cornette

et al., 1972), as shown by cross reactivity studies, in-

dicated that the only major competitor for the first anti-

body was PGFl (% 8%). Additionally, Fenwick and co-workers

(1977) reported the presence of large amounts of 6 keto PGFla

within the rat uterus, and Kimball of the Upjohn Company

suggests that 6 keto PGFla cross reacts (0 1%) with the PGF2a

antibody (Kimball, personal communication). For this reason

the results of the assay are interpreted as a measure of a

PGF-like molecule and not specifically PGF2a.



Statistical Design


Analysis of data for Treatment and Day effects was by

analysis of the variance. Treatments were Tl = content (no

incubation), T2 = 2 hr incubation, T3 = T2 + 1.0 pg pro-

gesterone, and T4 = T2 + 1.0 ig estradiol. Since variability

due to animal was random and since there was an unequal

distribution of animals within treatments, the data were

representative of a mixed model with unequal subclass

numbers (Steel & Torrie, 1960). With this in mind, the data

were analyzed with the LSML76 computer program developed by





-31-


Dr. Walter R. Harvey of Ohio State University. The statis-

tical models submitted to the computer are shown in Tables

III-1 and III-2. Since the tissue collection was completely

confounded by year and locus of sampling, statistical in-

ferences between these two studies were impossible and the

data were thus analyzed separately.



Results and Discussion


The content and production capabilities of endometrial

PGF in the intact pregnant and non-pregnant mare are shown

(Tables III-3 & III-4; Figs. III-2-7). In the non-pregnant

mare a significant Day (P < .0001) and in vitro Treatment

effect (P < .0001) as well as a Treatment x Day interaction

(P < .0077) were detected (Table III-1). For the non-pregnant

study, the Treatments were T1 = content, T2 = incubation,

T3 = T2 with progesterone, T4 = T2 with estradiol. Due to

the Treatment x Day interaction, the in vitro Treatments

within each Day were compared by orthogonal contrasts.

Additionally, the time trends within each Treatment (reduced

statistical model) were analyzed with polynomial regression.

These subsequent analyses were performed with the Statis-

tical Analysis System (SAS) computer program (procedure GLM)

and least squares means and SEM from this reduced statis-

tical model are shown in Tables III-3 and III-4. The pro-

duction capacity of the non-pregnant endometrium increased

(P < .0001) from estrus until Day 16 post-ovulation and






-32-


P,









4H.-







AO4-)
00r10000000
000000000














0. .



0 0 E E E











4- ;:3 0 o 0l
C 0 m00000


0 l f**L* -* 0 '



onE u ,,.o --l





SC. UC). .0 H C*H



V)f E- E E-I II o
O O LfH C

H







*H
U)


04 U c V) O
^ E --
















C1 *HU, Z -' aH ,- 0

C'Nsc 0 0 Z
Hri H Hr
C- S( t->(V








-33-


00 c00co o
Lr C\OC1) C


*iH


c2 C) m c) CIS S
04<<000
222222%


01
4-,



-4






















ct
4-e




























0 0i
C4
c-








C)
r- *







+CC







*r-i
;-!


U
*r-



v< u
b u
C) 0 -~
0 ,.0
'-4 O
-H C U


.t LO0 00'
V) r-4

t-
--l


E- E-

E-




C)
O~ *H


acu















1n Lt cO Cq
CD oD rli

oH o ^


U U n -


r- C ) CQ :



C 0)

^Q f fC
ZE-ao


MLOl r-1

COO
co c


\(NJoLQ


C) tr--


n o LO r-fl
H- t-irN


0
*H
-o



U



0
C

'-i 0


C):

4-3
CU
0
UH

II II




II














0













0








-34-


0
ot
>




























O
rtj











0








0
U

O
!-4













U






to
-4>
C



















HO
0ed



a,0


O*
U









ct















i-i


a)






uo 4
-U)

.0 0 0


P, P; c
ur u


4r-








0) 0
4- 0 0)->
+ *H 0 (
41 4j 4 a(L)


r ct E












0 O
u U; a)

H *l .H









CM
?- ce E















H



0




















0 1


c qco cN o 0
r- 4


oo

I--,




+1

0


o










to
un

+1























O
CN
t-















00







rs

^-1





-35-


Table III-4.


Endometrial PGF Concentration* in Pregnant
Mares after In Vitro Incubation


PGF PGF after
Day Content Incubation
Post-Ovulation (Treatment 1) (Treatment 2)

8 1.12 1.72(3) 10.05 14.66(5)

12 1.31 1.72(3) 16.75 13.38(6)

14 2.09 1.72(3) 26.87 + 14.66(5)

16 2.38 1.49(4) 49.77 12.39(7)

18 3.07 1.49(4) 67.60 13.38(6)

20 8.79 1.49(4) 92.38 12.39(7)


*ng PGF/mg endometrium; least squares
within each treatment.


X SEM (n) calculated































tn




C0 >
4-JH








4 -e

cnnc









)r-I 4 Uf
C) C)
V)) .-4


Cl. 0 H
Ucro


Ue U r.4



o O


U r/-

'*0 q 0
o C/ OU r
;-q .,



r+i U+ -'









'H





-37-


z
0

O
3
? O

... J

(0


0 0 0 0 0 0 0
03 nO O I- Bu
3nfSSI1 BW//jd bu


0 0 0
ro OC -






























C))



tz C) +co

C3 U I
cn >< k II
1) -H 0 Zo1j-
$-4 uZ~c


V 0 a)
V) "o (A .
do~i ~
r~'*4'-: U "0


-, U rcX
4- 0 0)C



U00) LHfu\c


uQUO *.c~
00 -
(3) 'r 4-
r-i u rl

0 ) U I)+ 0

tn-4:: -C)P
(3) L) 0VO ~

u, u (1)w L
r, 11 --t -





"i r- >-,(1) CD t


04 a u I m~O






-39-








C C
0
cC o
c c a o














0d
0

Q. .
C
o




Szz








o 0D





OOD
S




I-









a O o I \\ 0
)3nSSl B d \,
Q 0\

LLJ 1:



H
































C) C


0 u
uc)a~


v 1--, (1)

V) u~r
kOU




C)o







4-) 0 to



0 4-) ;-4
-TV Q-






4-J V-) CM4


4- CZ


4-3e




0 u
r( U




I::






-41-


o
cu






















0\
I-














--J
C z










a z



LL



-.J
I-I




















anssu uO /_-9d Bu

























Lf)


0


c'Jc.uc
E4-4 cli
U) () L r)
P: "J uI>


CaQ Lf~o
;) -4 1.0-~O

U)-C C)J+
M) CD() C)
(3) (L) c)
4 0 > Lf

C3 1- C '
cMk N

V) C) <

U)k C11~-
cl Ili l C) 00
rnk cbC)
C) r, t C-4 r-,

,4k .N

c-lC)U)
*H 4~ )~ +
03 0 Z'~~-

V) *H X-
::Ll ro (1)
U C)CerO



r-Ur--4 Znr-
M 0U U U
o -H r-4 -4
*H U H 4-)
UE 4-J U V)r:


$i0 r-A (D r4ib


a ) u -+ P.
LI*)







-43-


3nssiI


Ow /.9d u


z
o


-_J

0
I-
co


>-


CO


10 CMJ


( 1 I I





























0
o

C) 4-)
ZC)


;-4 0 b
Cd V)
c)0



~4-J +J
V)






Q) 0 w
r~0








CZ 0



~ H t4





on (1 4-


lz --I C'
0 U u (1)

P+I U 4-j
Ia,







-45-


E E


iee
--


0 F -
- Q-
CD









Mo *
* 0 *


0 0 0 0 0


O O O O

3nsSI. i w / J.)d


z
0




0
I

0
a.


0 0
Nc
























(N rl

cO**r--I

0 C) Lo r, +1



4 *04- a)C:)


0 t'-) r= C
tA C)4-" 4-1+ I
r C 0 c'





Cd O'q ) 1 (1)t~r
C)0 4-J O W -
H X I4 1

E XcLn t









~~~ 0 )-
,3 E -1 (1)










0U 0 4J + I-
r- O <>->E r-
M) 4-) C) C+ )













V) 4J tn 010 0--
;-03 c rl .;4r-
-H C) )4J +(- 00
-3 CD O D

r t Cli

V) (1) to F::;" OC 0 C

0 q cd -
bklO 4- 0) -i J:C:
Q) r- CA X 0
94 D 0 -)







-47-


(D,







z
0
-J
>


0

I.-
C,
0~


4
0






F-F~
4c4E



0 0)



4 0E
0o
CL w

to
S0)
o~~~~ g
0~ 0





co ID LI
3nSSloII I/jod 6I





-48-


decreased on Day 20 (Fig. III-2). This time trend was best

described by the cubic equation (P < .0046; R = 0.4708),

Y = 38.96 15.00X + 1.815X 0.561X3 ; however, visually the

data were described best by the quartic equation (P < .2692;

R2 = 0.4909), Y = -34.98 + 19.76X 3.464X2 + 0.2630X3

0.006647X4 (Fig. III-3). Orthogonal contrasts between Treat-

ments in non-pregnant mares indicated that, on all days, con-

tent was lower than production (Table III-3) but mimicked PGF

production trends (Fig. III-4) and was best described (P <
2
.0007; R = 0.6732) by the quartic equation, Y = -21.25 +

11.43X 1.921X2 + 0.1297X3 0.02965X4 (Fig. III-5). There-

fore, in the non-pregnant animal, maximal endometrial produc-

tion and content occurred at a time corresponding to the ex-

pected time of luteolysis (Chapter II) and support the

hypothesis that PGF is the luteolytic substance of the mare.

Orthogonal contrasts between in vitro Treatments within each

Day (Fig. III-6) indicated that progesterone treatment was

without effect, but that estrogen was stimulatory on Days 8

(P < .0032), 16 (P < .0037), and 20 (P < .0023). The time

trends of the estrogen treatment were best described (P <

.0018; R = 0.5943) by the quartic equation, Y = 288.33 +

147.24X 23.588X2 + 1.522X3 0.033525X4 and the progesterone

treatment (P < .0035; R2 = 0.6098) by, Y = -86.49 + 48.31X -

7.916X + 0.5299X3 0.012101X4. These data suggest that estro-

gen has an in vitro stimulatory effect on the luteal phase

endometria. The question of steroid modulation is addressed





-49-


further in Chapters IV and V. In the pregnant mare, a highly

significant Treatment (Tl = content & T2 = incubation)

(P < .0001) and Day x Treatment (P < .0269) and a less

significant Day effect (P < .0934) were detected (Table

III-2). Due to the interaction, the Day effect of each

Treatment (reduced statistical model) was, as before, analyzed

with polynomial regression. The production capabilities of

the endometrium varied with Day (P < .0001) and increased

continually from Day 0 to Day 20 post-ovulation (Fig. III-3).

The time trend for production due to incubation was best

described (P < .1018; R = 0.4906) by the quadratic equation,

Y = 48.04- 9.623X+0.5945X2 (Fig. III-3). Prostaglandin F

content mimicked the production time trends and was also best
2
described (P < .0417; R = 0.4611) by a quadratic equation,

Y = 12.78 2.165X+0.0956X2 (Fig. III-5). Therefore, in the

pregnant mare maximal potential PGF production occurred at a

time of expected luteal maintenance. Since the CL is capable

of binding PGF at this time (see Chapter VII and Kimball &

Wyngarden, 1977) and undergoing regression, it is possible

that endometrial PGF fails to reach the ovary. Indeed,

uterine venous PGF levels are lower in the pregnant mare

than in the non-pregnant mare (Douglas & Ginther, 1976).

Based on the secretary patterns of the porcine endometrium,

Bazer and Thatcher (1977) suggested that the embryo influences

the direction of movement of uterine secretions. In preg-

nancy, uterine secretary products, such as PGF, move toward





-50-


the uterine lumen exocrinee secretion); while, in the absence

of embryonic influence, uterine secretions proceed toward

the endometrial stroma and associated vascular system

(endocrine secretion) (Bazer & Thatcher, 1977). Alternately,

the absence of luteolysis in pregnancy may also be a result

of sequestering and/or metabolizing endometrial PGF by the

embryo itself. Supportive of this hypothesis are the obser-

vations that the yolk sac fluid of the equine embryo con-

tains PGF (Zavy et al., 1979), the porcine concepts

metabolizes PGF (Walker et al., 1977), and the murine con-

ceptus requires PGF for implantation (Saksena et al., 1976).

Even though the reproductive status (pregnant vs non-

pregnant) was confounded by year and direct quantitative

comparisons were inappropriate, it is interesting to note

that the time trends between status were different. The time

trends for pregnant animals were best described by an upward

directed quadratic equation while the time trends for non-

pregnant animals were best described by a fourth order

equation with a declining slope.

It must be mentioned that in vitro incubations of this

nature may only indicate the endometrium's potential capacity

for producing PGF. It presents no proof of in situ synthesis

and production. Also, the membranes of the Day 18 and 20

embryos ruptured as they were removed from the uterus and

some of the yolk sac fluid may have come in contact with the

endometrium. Therefore, the possibility exists that the






-51-


PGF production capability of the Day 18 and 20 endometrium

may be influenced by fetal fluids. However, the exposure

was short-term and the endometrium was washed in PBS prior

to incubation.















CHAPTER IV

STEROID MODULATION OF EQUINE ENDOMETRIAL PGF




Exogenous estrogen increased endometrial or uterine

luminal concentrations of prostaglandin F2a (PGF) in the

progesterone-primed uterus of the ewe (Lewis et al., 1977),

cow (W.W. Thatcher, G. Lewis, & F. Bartol, personal communi-

cation), hamster (Saksena & Harper, 1972b), guinea-pig

(Blatchley et al., 1972), mouse (Saksena & Lau, 1973), and

pig (Frank et al., 1978). In Chapter III in vitro incuba-

tion of late luteal phase equine endometria in the presence

of estrogen led to enhanced production of PGF. To assess the

steroid modulation of endometrial PGF further, the effects of

in vivo and in vitro steroid treatments on ovariectomized

mares were investigated.



Materials and Methods


Except for the following modifications, the Materials

and Methods described in Chapter III were utilized in this

experiment. Endometrial tissue for this study was collected

in the summer of 1978 from 16 pony mares, having a mean

weight at the time of experimentation of 199.31 + 53.24

(X + S.D.) kg (range 170.1-289.3 kg). These animals were


-52-





-53-


derived from a pool of mares that had been previously uni-

laterally ovariectomized.

The remaining ovary was extirpated and after a two week

recovery period, the bilaterally ovariectomized animals were

randomly assigned to one of four hormone treatments. One

milliliter, each, of the following steroids was administered

intramuscularly in sesame oil vechile, for the indicated

time periods:

Treatment (CN) = Sesame oil for 3 weeks (control)

Treatment (XE) = 50 pg/ml Estradiol Valerate for 3 weeks

Treatment (XP) = 50 mg/ml Progesterone for 3 weeks

Treatment (XS) = 50 pg/ml Estradiol Valerate for 7 days
followed by 14 days of 50 mg/ml
Progesterone

In this manner, the effects of long term exogenous

estrogen and progesterone (EXT = XE & XP) upon the equine

endometrial PGF synthetase system could be evaluated. The

effects of exogenous estrogen followed by progesterone

(EXT = XS) was studied in an attempt to mimic the temporal

sequence of ovarian steroid secretion during early pregnancy.

At the end of the exogenous hormone Treatment (EXT)

period, uteri were surgically removed and endometrial strips

were incubated, in duplicate, to determine the effect of the

following in vitro treatments (IVT) on PGF production.

Treatment (CON) = no incubation, i.e., endogenous PGF
content

Treatment (PRD) = total production, i.e., endogenous PGF
content + 2 hr in vitro production

Treatment (P4) = PRD in the presence of 1.0 ig Pro-
gesterone





-54-


Treatment (E2A) = PRD in the presence of 1.0 pg 178-
Estradiol

Treatment (E2B) = PRD in the presence of 10.0 ng 178-
Estradiol

Treatment (E2C) = PRD in the presence of 1.0 ng 178-
Estradiol

As previously described (Chapter III), incubation was

for 2 hr in KRB at 37 C. Endometrial tissue was frozen at

the end of the incubation period and saved for subsequent

PGF radioimmunoassay. The concentration of PGF was corrected

for procedural dilutions and endometrial mass, and expressed

as ng PGF/mg endometrium.

The statistical test for significance between treatments

was through analysis of variance. Since exogenous treatments

were cross-classified with in vitro treatments, the data were

analyzed in a balanced 4 x 6 split plot design. The analysis

of the data was accomplished with the SASTEST (Statistical

Analysis System) computer program (Procedure GLM; Barr &

Goodnight, SAS Institute, Inc.) and the statistical model is

listed in Table IV-1.



Results and Discussion


Significant in vitro treatment (IVT) (P < .0001) and

exogenous treatment (EXT) effects (P < .0397), as well as an

IVT x EXT interaction (P < .0438) were detected (Table IV-1).

Due to the interaction, orthogonal contrasts among EXT were

done with each IVT group (Table IV-3) and, likewise, orthogonal

contrasts among the IVT were done with each EXT group

(Table IV-4).







-55-


0
*4-




0
:2:



o





0
O
0






0



a






,-


















LO




0c
o
















I.
C

E
*
























?-Q

[--l



4-1 -
0I-iU









rt


00 --( rn \0 0
alf C LfO Zl
-n "o L -) 1- 4 :
i--AcI r -- i -i -i


-x

E-< H




4J) >






x 0 I
0 *C)
io> 0
H



0 US fI 1



[-1 S: ^- '-i r


r- -4 000
1o oo











E- E


03 (U C)





,Cv EC () C) P
C E V E
cja

4-1
CA






4-H


0.
4C)







0 *4
*Ol 0






0 rH r

-0 0 0



SC- i -O -H -H -H
O 4:~
o o

II II II II II II




uc **

U2 Q qU


a
0u

) 0






O
rC) :K:



0

O 0
QH I:::
0 0

-4 0




0



0I 0


0 !) V)
On


II II II II


UZ > X







-56-


U


0 -- 0 0 C l
ozu co o co M
SON


4-J


O0---



o .o d4




SL; 0 0
U) vi LC) Cn +j
o SCPC C 0)

o L O

SII C C -'
0 0
4 0 0 .0 *-H0
F =: O M ^ O "o E





C tin C:) co L : :1 0 WW




to 0 O t bo o

S> 0 0 C O C
C 0 0
o 0 o+, M OCT. e- -o 0 0,

H 0q 0 r- r0 O-
c 0 o c co Z q
-1 I o n co oI H, oo a o 3

S0 0 H' 0 040





0 v) 4 0 P: U C l
C I C 0I HI II II +I 4 I 4 I II 4 1 + +



C) O M 00 CD l) V r 0) O C?. z t 0 4
+4 -. M ZO i- 6 +-C C) i U o 4 3




4J tU> OP.OI U a- aC' J
-w 9si t^ oo L o ( 0 l DOi 4o c? d Izt
i c- ii 11 M (C Ur-i 4 U a.iP i






rrl n -
CI H 4 C 1 *n ^<
1 W U-s X; ; 0 &; 0(- (--








0 L U4

Or 4-.1 4 -
SC) 0 P -
CJ r4C)] z (0 cC







41 =U N- tQ c N C) H-
0 H




CU)-z t.L U)
o et 0 : 0
4,U 4 C
Su E <
I U)-.' 4 E
4O 0 C-C O OZ N f o O u
4 0* C i - u CU tr- LO \o i^ (I E-0 i
U ; O --'
0 "3 U bc t
cO 0 E 3 O


tn C 0) *c
0 7o a to >




C E O( V- b ; tn 0
< 4 Z O 0 1 OO C fl C

: O3 C) Cv OU O O C, O- O
.0 l o o -oo o







-57-


u

S0 Cr
U 0U 0 N -

14-J






o ,-. co r- O
X






0 00 C)q
u oom \ Nr o u



00 C)
o I ) cU,

0 to csi rq


0 x C 0 o4

) 4-) r- 0 0 M 4- ) NN -
r C O O te\ o ;:5 V :I. W W
E +-' H '' '- r --0 'r- M O=;0

o r Ci; M 1 o '-oo o
) ,- 0 O* "- o
4 -. q -H Q
01C)0 C) C)C C
0 0 i 4 r-.

r a> o c oiu ~ -^c:jHH-
.qH 0 rO O *H ,-C 'r I--
4 0 t Hf r4 4. 4- 4 4

aj > +->zr LO r--4 tn g 0 4- c UQQQQ
4-J U)- t) r- 0 O ) 0 Cr C
F-4 ) 0 0u W o-' u p L a. P--

C O II I II II II II II II II II

c 0i Z -am Z0Q0 t
4o yI U r-I W W \




:3 C) C

C0 O **




4-O '-I C'-l E- H
Hi 4- rtQ co d 4-)
r40 to I5 o \ N NH m Q
.+0 0 Q Z o N M )
C0 U *










0
4C 4 vi







S H-UO ) N < C)
(/ W du e N 4-1

H0 0
4X' 4-J > >






O Z M r a


H i U U 0 m> -






-58-


C0





4-) o C:1 C
LO 0



0 1
4- co r:i C) O C) CO0




o C ,. C
o o \oo N ,-









G i0 0 a +O r-
Sn- z 0 o o t *HO
o ) a) 0 CD C) 0C






4 6 P 0.-1 .
0 '- 0 0 O4 0
4-- 0 +) 4- +N 4-
S0 ) S4P:(






, 0 O U 0

HO C0 S +

0O 0 n

- W O.-i ,- M ,- T3F






4- CD C 0 _) 0 C *
Sx -'i 0 .




0(D m -
. IC) ,-4.44 ,J
0 M O ) l
6o C) o o o 4- )



S+ C. CD C C 0
C- rt o r/-Z 3 tn Q
















o -4 z 4-
rO CnZ o a )
r-j c-' (^l i-( -4 3 3 C











0 M0 010 0 ;0O
3 40 (=i r- 4-l r l





















U)
LU 0 4-l r
W( bC 0 HT3 r) ^ C C3 I

































>, M > >
0 '--' +--
C'- 4-4 C)I l r M Ln C4-4 0 0 C UO aQ






















SO O Cn C ) C i C
uX o co P M ua, ui*c
- O o o > o o
fl I d)


















U 01
cd o0 0- i

I ) -0 r C. C" C t
- Cu o0\0-i r^ +
/01 0 o o ] Zo o +




*H U 0
















E- U rj &.P-lF l- C, *





-59-


The PGF content (IVT = CON) did not differ with EXT;

however, PGF basal production (IVT = PRD) was stimulated

(P < .0085) by the exogenous steroid treatments (CN < XS, XP,

& XE) (Table IV-2 & Fig. IV-1). In addition, content was

lower (P < .0001) than all production treatments (CON < PRD,

P4, E2A, E2B, & E2C), thus verifying that PGF production

occurred during incubation.

Orthogonal contrasts among the four EXT (Table IV-3)

indicated that the capacity of the endometrium to produce

PGF in animals treated with estrogen plus progesterone was

elevated above the estrogen treated animals (P < .2220 for

PRD; P < .0773 for E2A; P < .0128 for E2B; P < .0027 for E2C;

P < .1115 for P4) (XP & XS > XE) and that the progesterone

treatment was more stimulatory than the estrogen and estrogen

plus progesterone treatments (XP > XE & XS) (P < .3272 for

PRD; P < .0033 for E2A; P < .0001 for E2B; P < .0146 for E2C;

P < .0034 for P4).

Contrasts of the IVT (Table IV-4) indicated that PGF

production in the incubations containing the three levels of

estrogen was elevated above the progesterone treatment

(IVT = P4) and no treatment (IVT = PRD) (P < .0001 for CN;

P < .0892 for XE; P < .0110 for XP; P < .0062 for XS); while

progesterone treatment was ineffective in stimulating PGF

production above the no treatment groups. An estrogen

response existed among the three levels of estrogen, in that,

the 1.0 ng level of estrogen was lower than the 1.0 ig level

in all cases except the estrogen plus progesterone treatment




























V)
C


()0
o rJ





ou


U)





00


4- 0
ii








tf)









0U() 0"!

bOIZI 4-
X A- 4) MI

0 CLH

:I V
0) 0 LO
4J ) --A
















4-4 -
-4
0 C),
r_
E- -; U






















































C
o
0
0
o a





ui o
- 0-
Z sn "" i-

I-- <0 0
crr-




0 Q


0
z 1[ 11 is
ta X 4 0

0
X
LLJ


0 0

WnlOeuWopu3 Wu /49d


-61-


I-


/


\II


u 0
to clj





-62-


(P < .0181 for CN; P < .0223 for XE; P < .0434 for XP;

P < .4154 for XS) (therefore E2A > E2B > E2C > PRD = P4).

Thus maximal stimulation of endometrial PGF production

occurred in a (systemic) progesterone-primed uterus treated

in vitro with estrogen; however, this effect was absent in

endometria collected from animals treated systemically with

estrogen and in vitro with progesterone. Therefore effects

of steroids upon the endometrial PGF synthetase system may

vary in a manner dependent upon route of administration,

length of exposure, or dosage. Systemically administered

steroids, as the EXT or in vivo ovarian steroids, circulate

in the entire cardiovascular system and may interact with

non-reproductive tissue or organs, as adrenals, kidney, and

liver, and initiate secondary effects that may, in turn, have

a unique effect upon the endometrial PGF synthetase system.

This systemic effect would be absent in IVT where the endo-

metria may interact directly with the steroids during incuba-

tion. Additionally, the incubated endometrium was physically

isolated from any potential inhibitory or stimulatory (bodily)

influences that may or may not be related to steroid action.

Finally, steroids of EXT were diluted by the fluids of the

entire circulatory system and were administered for a longer

time period (3 weeks versus 2 hr) than IVT.

However, the observation that maximal stimulation of

endometrial PGF production occurred in a systemically pro-

gesterone-primed uterus treated in vitro with estrogen is






-63-


consistent with the findings of the endometrial production

studies conducted on the pregnant and non-pregnant mare

(Chapter III). In the pregnant mare, the capacity of the

endometrium to produce PGF increased (P < .0001) from the

beginning of pregnancy to Day 20 (Fig. III-2). This rise in

PGF was observed to occur in a uterus that had been under

the continual influence of systemic progesterone (Allen &

Hadley, 1974) and the influence of embryonic and/or endo-

metrial estrogens locally. The equine concepts was first

shown to be a potential source of estrogens with the histo-

chemical localization of several hydroxysteroid dehydrogenase

(HSD) enzymes within the membranes of the trophoblast (Flood

& Marrable, 1975). However, intense 17B-HSD activity was

found only in the endometrium, and specifically in the endo-

metrium in intimate contact with the concepts (Flood et al.,

1979). Additionally, the equine embryo has been shown to

contain (Zavy et al., 1979) and produce (Mayer et al., 1977)

estradiol and preliminary studies suggest that the blastocyst

may convert 3H-progesterone into 3H-estrone (Seamens et al.,

1979). Likewise, maximal PGF production in the non-pregnant

mare (Fig. III-2) appeared at a time (Day 16) that corres-

ponds to increased systemic progesterone production (Sharp &

Black, 1973; Douglas & Ginther, 1976) and also at a time

when intrauterine (local) estrogens are high (Zavy et al.,

1979; Fig. VIII-2 & Fig. VIII-3).

When the endometrium of the estrous cycle (Chapter III)

was incubated in the presence of steroids, progesterone was






-64-


ineffective in altering PGF production, whereas estrogen had

little or no effect on Days 4 to 12 but was stimulatory on

Days 16 and 20 (Fig. III-6). The in vitro stimulatory effect

of estrogen may be effective only in a uterus that has had

prior (prolonged systemic) progesterone exposure. In this

manner, the CL, through the secretion of its primary product,

progesterone, may ultimately orchestrate its own destruction,

via PGF stimulation.














CHAPTER V

THE LOCAL EFFECT OF THE EMBRYO ON ENDOMETRIAL PGF




The early pregnant equine endometrium has a high capacity

for producing prostaglandin F2a (PGF) (Chapter III) and

endometrial PGF production has been shown to be modulated by

steroids (Chapter IV). Maximal PGF production occurred in

a progesterone-primed uterus stimulated in vitro with estrogen.

The high endometrial levels of PGF in pregnancy may reflect

endogenous steroid enhancement, since this is a time when

both systemic luteal progesterone (Allen & Hadley, 1974) and

local fetal and endometrial estrogens are high. The equine

concepts contained (Zavy et al., 1979) and produced estra-

diol (Mayer et al., 1977) and may convert 3H-progesterone

into H-estrone (Seamens et al., 1979). Flood and co-workers

(1979) have observed histochemically,the presence of high

levels of 178-hydroxysteroid dehydrogenase on those parts of

the uterine endometrium directly apposed to the trophoblast.

This is suggestive of a local concepts influence on the

endometrium. Since the local production of PGF by the

equine endometrium may be enhanced by fetal estrogens, the

PGF production by the endometrium associated with the embryo

was compared with PGF production of the endometrium contra-

lateral to the embryo.


-65-





-66-


Materials and Methods


Endometrial strips were collected from ten pony mares

during the summer of 1978 and incubated according to the pro-

cedures outlined in the Materials and Methods of Chapter III.

Mean weight of the animals at the time of experimentation

was 188.8 28.9 (X t S.D.) kg (range 154.2-231.3 kg).

As before, production and content of PGF were measured

in endometrial strips. Duplicate strips were obtained from

the medial portion of the right and left uterine horns of

three Day 12 non-pregnant mares, three Day 18 non-pregnant

mares, and four Day 18 pregnant mares. In this manner a

comparison could be made of the PGF synthetase system between

the uterine horns of non-pregnant mares, i.e., horn versus

horn variability, as well as a comparison between the uterine

horns of pregnant mares, i.e., local influence of embryo.

The statistical test for significance between Treatments

and Hornsides was by analysis of variance. The Treatments

were T1 = content (no incubation) and T2 = 2 hr in vitro

incubation; and the hornsides were Il = right horn in non-

pregnancy and gravid horn in pregnancy, and H2 = left horn

in non-pregnancy and non-gravid horn in pregnancy. Since

variation due to animal was considered to be random and since

there was an unequal distribution of animals within treat-

ments, the data were representative of a mixed model with

unequal subclass numbers (Steel & Torrie, 1960). As

such the data were analyzed with a SAS (Statistical Analysis





-67-


System) computer program (Procedure GLM; Barr & Goodnight,

SAS Institute, Inc.). The statistical model is listed in

Table V-l.



Results and Discussion


The PGF production and content in opposite uterine horns

are shown in Fig. V-l and Table V-2. As in previous experi-

ments, endometrial PGF content was significantly lower

(P < .0003) than production (Table V-l) (Treatments: Tl =

content, T2 = incubation). The lack of a Day x Hornside

interaction (P < .8313) demonstrated a lack of horn versus

horn variability in the non-pregnant mare while a significant

Status x Hornside interaction (P < .0032) indicated that, in

pregnancy, the gravid horn had a higher capacity for pro-

ducing PGF than the non-gravid horn (Status: P = pregnant,

NP = non-pregnant; Hornside: HI = right horn in NP & gravid

horn in P, H2 = left horn in P & non-gravid horn in NP).

Thus it appears that the embryo has a local influence on the

uterus to stimulate PGF production. Endometrial PGF of

pregnancy may not be involved in ovarian function, since

uterine vein concentrations of PGF were lower in pregnant

than in non-pregnant mares (Douglas & Ginther, 1976); instead,

the PGF may be involved in fetal-placental physiology.

Enhanced production of PGF in only those areas of the equine

endometrium that are in intimate contact with the embryo

suggests that PGF may be utilized by the embryo. Indeed,





-68-


PGF has been demonstrated to be present within equine yolk

sac fluids (Zavy et al., 1979) and required for implantation

in mice (Saksena et al., 1976).







-69-


-- :" m t-IOCDM r-ICDr-i m


000000000000000












E HHiE- H-X E-XX

X X XX XX XE-E-E-H E-

r: 9:1 X X X X r

rcl Cl) ) DO cn cn U) w-
Q MMMMMu n 4 x x x
"Z *'QQQ~ mr Mc*


Z c/ : > Z Z *> *. uuVV-/,/^-<
'-^-^C MU MCO /3LLOQ QQ


,.

u

C)




r-4




C
ct












t-




4-1

















u


.tLi
C)
>r,















0
,U
E
































4-'
aC-
0
















+-4
*-










?-Q
0ue
'+-


K
H=
-K ,-\ U)

0C)
.O C +) X X ,-
C O


C > C) 0 X- X
(^ Qu a- E Q -- Q
3 *1-i E- -12i S &-1~C
+-^~ c -A
    a^^oxx~ki-iXX~
    + 2> Cl rt H o
    LCQSt-^W S2:=OQS%'L


    oo CN] C


    iMHam
    tn tnM


    r-40
    r0Lf0


    PC




    Cc)


    rsi cn t, m i-io "t m mtn m- m

    o \-o 1 I tn m) cc \D


    . -O

    0
    *-H
    -C





    *H












    *H
    m

















    r*H


    CA

    0
    C






    Or(



















    "0
    *l-
    r/
    c;


    0

    4-H



    u
    .0

    U










    II l
    P 3








    U~H
    II II




    II

    0

    C)

    ce

    C)


    4.-

    CO

    d
    +JM

    c o



    4 0


    II II


    z

    II

    U)

    4-J






    -70-


    Table V-2. Local Influence of Embryo
    Production Capacities


    upon Endometrial PGF


    Status* Day Hornside* (n) X ng PGF/mg Tissue
    (I SEM)*

    NP 12 1 3 19.85 2.008

    NP 12 2 3 20.30 2.008

    NP 18 1 3 9.82 2.008

    NP 18 2 3 9.67 2.008

    P 18 1 4 41.24 2.373

    P 18 2 4 19.08 2.373


    *Note: Status =


    Hornside


    P = Pregnant
    NP = Non-Pregnant

    = H1 = Right side in NP, gravid side in P
    H2 = Left side in NP, non-gravid side in P


    The estimate of SEM included
    2
    M (S,D) x T x H.




























    Fig. V-1. Local influence of embryo upon endometrial PGF
    production capacities (least squares mean +
    SEM).






    -72-


    NP = NON-PREGNANT

    P = PREGNANT


    PRODUCTION




    CONTENT


    NP NP P
    DAY 12 DAY 18 DAY 18

    UTERINE HORN SIDE


    S30

    E
    o

    w
    e.,


    IL
    c
    o20

    r-














    CHAPTER VI

    THE LOCALIZATION OF PGF IN THE
    PERFUSED EQUINE OVARY



    Observations that exogenous prostaglandin F2a (PGF)

    induced luteal regression in mares (Douglas & Ginther, 1972;

    Allen & Rowson, 1973), and that an increase in endogenous

    uterine luminal PGF (Zavy et al., 1978), endometrial PGF

    (Chapter III), and uterine venous PGF (Douglas & Ginther,

    1976) occurred at the time of natural luteolysis, strongly

    suggests that uterine PGF is the luteolytic factor in the

    mare. Since the corpus luteum (CL) appears to be a target

    tissue for PGF, it would follow that exogenous PGF may

    localize in the CL. To this end, luteal phase ovaries were

    removed from mares and perfused with radioactive PGF. The

    localization of high levels of PGF within the CL would be

    supportive of the hypothesis that PGF is involved in luteal

    endocrinology.



    Materials and Methods


    Ovaries were collected from four, Day 14 non-pregnant

    pony mares during the summer of 1978 according to the pro-

    cedure outlined in the Materials and Methods of Chapter III.

    The only deviation in the surgical technique was the placement


    -73-





    -74-


    of a loose tie of 2/0 silk suture around the ovarian artery

    as an aid in locating the artery for cannulation after

    surgery.

    After ovariectomy, the ovary was immersed immediately in

    a heparinized 0.9% saline bath maintained at 37 C. A cannula

    (PE 190) was inserted and sutured into the ovarian artery

    and attached to a 3-way valve with a 22 gauge needle (Fig.

    VI-1). The 3-way valve was attached to a 1 cc syringe con-

    taining PGF2- 3H (NET 345, lot 932-121) in 0.9% saline and

    a 20 cc syringe containing heparinized autologous blood that

    was collected at surgery. The 20 cc syringe was removed

    temporarily and the ovarian vasculature was flushed with 2.0

    ml of heparinized 0.9% saline to retard coagulation of endo-

    genous blood. The ovary, with cannula, was suspended over

    a petri dish and perfused with the tritiated PGF2a (Specific

    Activity = 10.9 Ci/mmol) and heparinized plasma. The scheme

    of perfusion was comprised of injections of 100 pl of PGF2a

    H (11.46 ng) followed by 2.0 ml of heparinized plasma.

    This scheme was repeated five times at a flow rate of

    approximately 2.0 ml/min. Thus, a total of 57.30 ng (0.5

    ml) of labeled PGF2a was injected. The ovary was then

    flushed with 10.0 ml of heparinized 0.9% saline. All fluids

    injected into the ovary were maintained at a temperature of

    37 C.

    At completion of the perfusion, the ovary

    was wiped dry and stored frozen for subsequent analysis. In




































    ct
    C),




    4-4
    Srl -



    C) '
    * 4






    0

    UC)


    ct

    ,I-
    C)4-







    0


    0
    C) ,
    cl V

    W4-
    cd 4 P






    -76-


    L1


    Sa
    a





    0
    ou


    0
    E






    o
    a_-



    ouj 0





    -77-


    addition, 5.0 ml of effluent and 500 Ul of the original
    3
    PGF H stock solution were saved to determine total uptake
    2cv
    and total amount injected.

    The frozen ovaries were thawed at room temperature and

    weighed. The corpus luteum was ablated carefully, so as not

    to disrupt any follicles, then weighed. Each follicle was

    exposed by dissection and the follicular fluid was aspirated

    into syringes and the total volume of fluid noted. One

    hundred microliters of the follicular fluid from each ovary

    were added to liquid scintillation vials containing cocktail

    and counted in a Beckman LS-300 liquid scintillation

    spectrophotometer. The percentage of PGF2- H injected that

    localized within the follicular fluid was calculated.

    The ovary, now minus the CL and follicular fluid, was

    weighed. This weight served as an estimate of stromal

    weight. In addition to the connective tissue and inter-

    stitial cells of the ovarian stroma, this estimate of stromal

    weight also includes the weight of the ovarian vasculature

    and follicular structural elements. Physical separation of

    these ovarian components was found to be impossible. Approxi-

    mately 100 mg, each, of the CL and ovarian stroma were

    homogenized in 2.0 ml of absolute ethanol in a Potter-type

    hand homogenizer. Ovarian stroma samples were collected from

    areas of the ovary that were devoid of large blood vessels

    and follicles. Thus, although the "stromal weight" includes

    follicle structures, the stroma used for estimation of





    -78-


    radioactive label was devoid of any follicle structures. The

    homogenate was centrifuged at 1086 G for 10 min and 100 pi

    of the supernatant were added directly to liquid scintilla-

    tion vials containing cocktail and counted as before. The

    percentage of the total perfusedd) PGF23- H injected that

    localized within the entire CL and ovarian stroma was cal-

    culated. Finally, 100 pl of the effluent and original PGF2a 3H

    stock solution were added to liquid scintillation vials and

    counted.

    The amount of radioactivity (CPM) within all components

    was corrected for quench with a water quench curve and ex-

    pressed as DPM. Quench corrections were determined with the

    aid of an external standard (137Ce) supplied with the

    Beckman LS-300 liquid scintillation spectrophotometer. The

    water quench curve was constructed by adding a known amount

    of radioactivity to liquid scintillation vials containing

    increasing concentrations of water (250 pl to 1500 pl).

    Efficiency (%) of counting was graphed against the external

    standard ratio computed by the liquid scintillation spectro-

    photometer. Corrections for quench within samples were then

    calculated by extrapolation of the external standard ratio

    from each sample on the water quench curve. Statistical

    tests were performed on the amount of radioactivity, ex-

    pressed as DPM.

    The test of significance between Treatments was by one-way

    analysis of variance (Treatments: T1 = % 3H in effluent,
    T2 = 3H in CL, T3 = % 3 in follicular fluid, T4 = 3 in
    T2 = % H in CL, T3 = % H in follicular fluid, T4 = % H in





    -79-


    ovarian stroma). The analysis was computed with the SASTEST

    (Statistical Analysis System) computer Procedure GLM (Barr &

    Goodnight, SAS Institute, Inc.) and the statistical model

    with appropriate orthogonal contrasts is shown in Table VI-1.



    Results and Discussion


    In an attempt to mimic in vivo conditions, PGF 3H was

    perfused through the equine ovarian artery at a rate (2.0

    ml/min) comparable to that detected by electromagnetic flow

    probes (M. Simonelli, D. Caton, and D. Sharp, unpublished

    observations). Furthermore, since endogenous PGF may be

    secreted by the pig in a pulsatile fashion (Moeljono et al., 1977)

    PGF was injected into the ovary in five pulses. Although

    not quantified, the amount of infused heparinized saline

    required to remove endogenous ovarian blood was approximately

    2.0 ml.

    The variances of the four treatments (TI = % 3H in

    effluent, T2 = %3H in CL, T3 = % H in follicular fluid,

    T4 = %3H in ovarian stroma) were heterogeneous as determined

    by the Bartlett's test for homogeneity of variances (Steel &

    Torrie, 1960) and therefore a log transformation of the data

    was performed. Orthogonal contrasts (Table VI-1) of the

    data (Table VI-2) indicated that the percentage of radiation

    (DPM) localized in the ovarian tissues differed (P < .0006)

    from the amount remaining in the effluent (42.2%) (Fig. VI-1)

    (Tl > T2, T3, T4). Furthermore, of the three ovarian tissues


















    o
    t4-4
    0


    r
    *H


    N
    *H
    r-1
    ed
    U
    04-



    uO
    0


    c`.I





    O















    0
    C




    0
















    NC:












    I .
    0



    H
    cri




    U 3

    rt









    *i-i

    rt


    M 1


    +i 4- 75

    0 0



    L) U 0 c3
    2E o


    -80-


    N4 1.0 r- Ln Ln


    00000
    00000







    00000







    Cf) C) C=) C) C
    ct tO mrt 'rl
    EEEEE
















    L/ L N\ 0 Lr) CO0








    M a U)






    -I


    *H



    10 k


    4-) 1





    4tr 0 0 >
    (U O4 0


    *H H *'H *H




    ii II II








    Cd


    Q4)
    0t
    (U


    H



    HH
    iE-



    E-H



    E- E- E-"







    II II II








    O0 0
    " r r,







    -81-


    ) MK t) M rn M
    0 'r 0d 00 01 10 ,j
    ) rC: U
    3 o3 N I).OO. t

    i0 +i +1 +1 +i

    U 00 (14 Ln CD H
    4U oo N .0 r- ,w i
    0 0- ..O N
    0 ** 4*



    o

    ) CO
    0
    S+1
    Un

    H r 0 n





    I .
    m a ): vr-ono
    t0 --

    *H (CU- r4 NJ
    E-i0 co i-n ,-- L i -







    S. (. c
    '- o I 4

    03 I













    0 0 0 L
    U ~ *
    *H- 1 o o- o 0 -













    o odn ant 4- -J
    ?i E ~ t-o 0 0




    m o .-.
    -ed






    m U4-1 MC
    00



    O e e- 0-
    0 0 0
    U- I C o
    0 O U) 0
    i] 0E- I Z
    i ol /




    0 4 1 0* O



    .0 t cr-ct Ot

    et 0 3 2






    E-~a (L U-O -;





    -82-


    studied, the CL (22.88%) sequestered the largest (P < .0007)

    amount of radiation while the follicular fluid (4.62%) con-

    tained more (P < .0195) than the stroma (1.75%) (T2 > T3 > T4)

    (Fig. VI-1). The effluent, CL, ovarian stroma and follicular

    fluid represented 71.54% of the injected PGF2 -3H. Localiza-

    tion of the remaining 28.55% was unknown, but presumed to

    be in ovarian tissues not studied, i.e., granulosa cells

    and epithelial cells.

    The high level of radioactivity within the CL is sup-

    portive of the hypothesis that PGF is involved in luteal

    physiology. However, interpretation of these results is

    mitigated by the possibilities of PGF-3H being metabolized

    prior to binding or the high levels of radiation in the CL

    being a result of vascular pooling.





























    Fig. VI-2. Localization of tritiated PGF within various
    tissues of the perfused equine ovary
    (X + SEM).





    -84-


    OVARIAN
    LOCALIZATION
    OF
    H -PGF2.


    Em FL]


    STROMA FOLLICULAR C L
    FLUID


    45



    40



    35



    30

    L3
    CL
    0


    -r
    2O
    S20
    a-


    EFFLUENT














    CHAPTER VII

    SPECIFIC BINDING OF PGF TO THE
    EQUINE CORPUS LUTEUM



    Exogenous PGF will localize in the equine CL (Chapter VI)

    and induce luteal regression (Douglas & Ginther, 1972; Allen

    & Rowson, 1973). Although the mechanism of action of this

    effect of PGF is not well understood, involvement of the

    membrane-associated adenyl cyclase system in luteolysis

    (Marsh, 1971) suggests that the mode of action of PGF is

    through a membrane receptor. Luteal PGF membrane receptors

    have been isolated and characterized biochemically in the cow

    (Kimball & Lauderdale, 1975; Rao, 1976), sheep (Powell et al.,

    1974), mare (Kimball & Wyngarden, 1977), and woman (Rao,

    1977). Therefore, the membrane receptor may play a role in

    luteal regression in the cycling mare, or in prevention of

    luteal demise in the pregnant mare. This study was designed

    to characterize the equine luteal PGF receptor and to assess

    its possible role in the cycling and early pregnant mare.



    Materials and Methods

    PGF Binding to Luteal Tissue

    Ovaries containing luteal tissue were removed surgically

    from mares according to the procedure outlined in the Materials


    -85-





    -86-


    and Methods of Chapter III. After surgery, the ovary was

    immersed immediately in chilled (4 C) phosphate buffered

    saline (PBS) containing indomethacin (10 pg/ml) and the CL

    was dissected free, weighed, quick frozen in liquid nitrogen,

    and stored at -60 C. Blood clots found within a corpus

    hemorrhagicum were removed prior to weighing.

    Later, the frozen CL was sliced thinly with a scalpel

    and homogenized in a Potter-type hand homogenizer. To remove

    cellular debris, the homogenate was filtered through three

    layers of cheesecloth and centrifuged at 1086 G for 10 min

    at 4 C. The supernatant was recentrifuged at 100,000 G for

    1 hr at 4 C, and the resultant pellet was resuspended in PBS

    (2 ml PBS/g of original tissue). This (particulate) membrane

    fraction was then stored at -60 C until analysis.

    In the equilibrium binding study, 100 il of membrane

    preparation (MP) containing approximately 0.8 mg of protein

    were incubated with 0.38 ng PGF2a- H (specific activity =

    9.2 Ci/mMol; NET 345, Lot 787-250) in either the presence

    (non-specific binding) or absence (total binding) of 1.0 pg

    PGF. Unless otherwise stated, incubations were conducted in

    a reaction volume of 300 pl, in triplicate, for 3 hr at 22 C,

    pH 7.4. Following incubation, the free [F] and bound [B]

    ligands were separated by ultrafiltration as described by

    Kimball and Lauderdale (1975). Separation was accomplished

    by filtering 100 jl of the incubation mixture through HAWP

    0.45 Millipore filters. Filters were washed with 4 x 3 ml





    -87-


    PBS and air dried. The filters were then added to liquid

    scintillation vials, solubilized with 15.0 ml methyl cellu-

    solve, and the tritium counted on a Beckman LS-300 liquid

    scintillation spectrophotometer as described previously.

    Specific binding was calculated by subtracting non-specific

    binding from total binding and was expressed as the percentage

    of the total PGF bound to the membrane preparation.



    Receptor Characterization


    The time required for PGF to equilibrate with its

    receptor was determined by measuring the amount of PGF2a -3H

    specifically bound to the MP after incubation times of 15,

    30, 120, 180, and 240 min at both 4 C and 22 C. The rever-

    sibility of ligand-receptor binding was studied by measuring

    bound PGF at 0, 15, 30, 60, 120, 180, and 240 min after

    addition of 1.0 pg of unlabelled PGF to a 180 min incubation.

    A dose-response study was conducted by performing competi-

    tive binding studies in the presence of 0, 2.5, 5, 10, 25,

    50, and 100 ng unlabelled PGF.

    To further assess receptor specificity and cross-

    reactivity, incubations were conducted in the presence of

    0.1, 1.0, and 10 pg of the various PGF analogues shown in

    Fig. VII-1. The resulting dose-response curves were analyzed

    by least squares regression analysis on a Monroe 1860 cal-

    culator and the theoretical concentration required to displace

    50% of the bound PGF2 -3H was calculated. The ratio of PGF2a



























    Fig. VII-1. Listing of the chemical structures of the PGF2t
    analogues used in the cross-reactivity study.
    Arrows indicate those areas of the analogue
    that differ with PGF2, and the parentheses in-
    dicate the % cross-reactivity between the
    analogues and PGF2 .