Citation
Regulation of endometrial prostaglandin synthesis by phospholipases and interferon tau and characterization of an endometrial prostaglandin synthesis inhibitor in the bovine

Material Information

Title:
Regulation of endometrial prostaglandin synthesis by phospholipases and interferon tau and characterization of an endometrial prostaglandin synthesis inhibitor in the bovine
Creator:
Danet-Desnoyers, Guénahel Henri, 1963-
Publication Date:
Language:
English
Physical Description:
vii, 256 leaves : ill., photos ; 29 cm.

Subjects

Subjects / Keywords:
Cattle ( jstor )
Endometrium ( jstor )
Fatty acids ( jstor )
Interferons ( jstor )
Lipids ( jstor )
Nonesterified fatty acids ( jstor )
Pregnancy ( jstor )
Prostaglandins ( jstor )
Secretion ( jstor )
Sheep ( jstor )
Animal Science thesis Ph.D
Cattle -- Physiology ( lcsh )
Dissertations, Academic -- Animal Science -- UF
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1994.
Bibliography:
Includes bibliographical references (leaves 205-255).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Guénahel Henri Danet-Desnoyers.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
030582226 ( ALEPH )
31920216 ( OCLC )

Downloads

This item has the following downloads:


Full Text










REGULATION OF ENDOMETRIAL PROSTAGLANDIN SYNTHESIS
BY PHOSPHOLIPASES AND INTERFERON TAU
AND CHARACTERIZATION OF AN ENDOMETRIAL
PROSTAGLANDIN SYNTHESIS INHIBITOR IN THE BOVINE


21


4


'A


i


By


GUtNAHEL HENRI DANET-DESNOYERS


A

"~ ?
L~~1


* '~:


1~
*
'4 c~


*

1'


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

UNIVERSITY OF FLORIDA

1994














D6die !
Monique et Jean Desnoyers

et

Gregory Seaney



































'1














ACKNOWLEDGMENTS


I would like to express my sincere gratitude to the chairman of my supervisory committee, W.W. Thatcher, for his guidance and support during the course of my program. He made possible this training at the University of Florida that I consider to be personally and educationally rewarding. I would like to thank the remaining members of my supervisory committee, W.C. Buhi, P.J. Hansen, M.J. Lawman and F.A. Simmen, for their guidance and suggestions.
Sincere thanks are extended to Jesse Johnson and Sean O'Keefe for their invaluable assistance with my research and their technical expertise. I would like to extend special thanks to the many students, technical persons and postdoctoral fellows I have worked with throughout my doctoral program: Lokenga Badinga, Thais Diaz, Susan Gottshall, Timothy Gross, Joyce Hayen, Eric Schmitt, Luzbel de la Sota, Marie-Joelle Thatcher, Carla Wetzels and all the others for their friendship, humor and moral support. Thanks are extended to Monte Meyer for providing me with endometrial tissue from interferon tau-treated cows (chapter 4) and Mary-Ellen Hissem for her eternal smile, sense of humor and secretarial expertise.


iii













TABLE OF CONTENTS
page


ACKNOW LEDGM ENTS ...................................................................................... iii

ABSTRACT...................................................................................................... vi

CHAPTERS

1 REVIEW OF LITERATURE........................................................................ 1
Phospholipids ......................................................................................... I
Linoleic and Arachidonic Acids............................................................. 5
Arachidonic Acid Mobilization................................................................ 8
Phospholipases................................................................................... 10
The Cyclooxygenase Pathway.............................................................. 18
Inhibition of Cyclooxygenase ................................................................ 23
Prostaglandin Receptors..................................................................... 26
Estrous Cycle....................................................................................... 27
Follicular Growth, Maturation and Dom inance..................................... 27
Corpus Luteum and Prostaglandins.................................................... 33
Oxytocin and Prostaglandins ............................................................... 37
Regulation of Uterine Prostaglandin Production.................................. 39
M aternal Recognition of Pregnancy.................................................... 49
Trophoblast Interferons....................................................................... 55
Implications.......................................................................................... 68

2 REGULATION OF ENDOMETRIAL PROSTAGLANDIN SYNTHESIS
DURING EARLY PREGNANCY IN CATTLE: EFFECTS OF
PHOSPHOLIPASES AND CALCIUM IN VITRO .............................. 70

Introduction .......................................................................................... 70
M aterial and Methods ......................................................................... 72
Results................................................................................................. 81
Discussion................................................................................................ 90


iv











3 EFFECT OF NATURAL AND RECOMBINANT BOVINE
INTERFERON t AND OXYTOCIN ON IN VITRO SECRETION OF PGF2a AND PGE2 BY ENDOMETRIAL EPITHELIAL AND
STROMAL CELLS ............................................................................ 97

Introduction ......................................................................................... 97
M aterial and Methods .......................................................................... 99
Results................................................................................................... 111
Discussion..............................................................................................122

4 IDENTIFICATION AND QUANTIFICATION OF AN
ENDOMETRIAL PROSTAGLANDIN SYNTHESIS INHIBITOR
(EPSI) IN THE BOVINE .....................................................................129

Introduction ............................................................................................ 129
M aterial and Methods ............................................................................ 131
Results...................................................................................................152
Discussion..............................................................................................179

5 G ENERAL DISCUSSION.......................................................................194


REFERENCES .................................................................................................205


BIOG RAPHICAL SKETCH................................................................................256


V













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy REGULATION OF ENDOMETRIAL PROSTAGLANDIN SYNTHESIS
BY PHOSPHOLIPASES AND INTERFERON TAU
AND CHARACTERIZATION OF AN ENDOMETRIAL
PROSTAGLANDIN SYNTHESIS INHIBITOR IN THE BOVINE By

Gu6nahel Henri Danet-Desnoyers

April, 1994

Chairman: William W. Thatcher
Major Department: Animal Science

The bovine conceptus must prevent uterine luteolytic secretion of prostaglandin F2. (PGF2a) in order to maintain pregnancy. The antiluteolytic signal from the conceptus is trophoblast interferon tau (IFN'r). The present investigations were conducted to extend our understanding of conceptus and IFNt-mediated regulation of endometrial secretion of PGF2a and PGE2

Arachidonic acid, phospholipase A2 (PLA2), phospholipase C (PLC), extracellular calcium (Ca2.) and calcium ionophore were examined for regulatory effects on prostaglandin biosynthesis by endometrial explants from cyclic and pregnant cows at day 17 postestrus. Secretion of PGF2a was lower in endometrium of pregnant versus cyclic cows. Arachidonic acid stimulated prostaglandin endometrial secretion for both reproductive statuses. The stimulatory effect of PLA2 on PGF2. and PGE2 secretions was greater in explants of pregnant cows. However, calcium ionophore inhibited the PLA2 stimulatory


vi









effect on endometrium of pregnant but not cyclic cows. Findings indicate that availability of AA may limit endometrial secretion of prostaglandin in pregnant cows. The effects of natural and recombinant bovine IFNT (nblFNt and rblFNT) on secretion of PGF2a and PGE2 by endometrial epithelial and stromal cells were assessed in a second phase of investigation. Epithelial cells secreted more PGF2a than stromal cells, whereas stromal cells secreted primarily PGE2. Oxytocin stimulated secretion of PGF2a and PGE2 from epithelial cells. However, basal and oxytocin-induced secretions of PGF2a and PGE2 decreased with increasing doses of either nblFNT or rblFNt. Neither nblFNT, rblFNT or oxytocin altered prostaglandin secretion by stromal cells. The antiluteolytic effect of blFNT is exerted on epithelial cells of the endometrium.

Nature of the endometrial prostaglandin synthesis inhibitor (EPSI) present during early pregnancy was characterized. In cytosol, EPSI was associated with serum albumin. Cytosol extraction followed by liquid and gas chromatography indicated that EPSI activity was associated with nonesterified linoleic acid (LA), a competitive inhibitor of cyclooxygenase. Higher EPSI activity in association with higher LA and decreased AA concentrations were found in endometrial microsomes from pregnant versus cyclic cows. In summary, the bovine conceptus, possibly through IFNT secretion, attenuates endometrial secretion of PGF2a by altering endometrial lipid metabolism to enhance the ratio of unesterified LA to AA concentrations in microsomes.











vii













CHAPTER 1
REVIEW OF LITERATURE


Phospholipids


Phospholipids comprise by far the major component of most membranes with the exceptions of nervous tissue and plant chloroplast membranes. Different phospholipid classes are always found in fixed proportions in a given membrane type, although another membrane in the same cell might be quite different in its phospholipid distribution (Wirtz, 1974). Most of the structural phospholipids are derived from the same precursor pools and interconversions of one phospholipid into another are commonplace. Not only is each phospholipid present at a specific percentage of the total, but each maintains its own characteristic distribution of fatty acyl side chains. This is achieved through regulatory factors controlling rates and selectivity of the synthetic and degradative enzymes.

In eukaryotic cells, most of the phospholipids are synthesized in the microsomal membranes (endoplasmic reticulum). However, the rapid exchange of phospholipids between the microsomes and other functionally different membranes renders it impossible to consider the pools as being independent. Phosphatidic acid is an important branching point in the pathway of phospholipid synthesis. The fatty acids (FA) found esterified to the 1-position of phosphatidic acid and other phospholipids are almost all saturated (Lands and Crawford, 1976) and in particular, phosphatidic acid contains large amounts of palmitic acid in position 1. In contrast to the preference for saturated FA at the


1







2


1-position, the 2-acyl ester of phosphatidic acid and other common phospholipids is usually a cis-unsaturated component such as oleic, linoleic or arachidonic acid (AA). It appears that the rate-limiting steps in diglyceride utilization for phospholipid formation are the cytidyl transfer reactions which are, at least in part, controlled by the concentrations of substrates. The apparent Km for chicken brain microsomal 1,2-diacyl-sn-glycerol:CDP-ethanolamine transferase decreases in the presence of diacylglycerols, and unsaturated FA stimulate phosphatidylethanolamine (PE) formation, while saturated FA appear to inhibit it (Mead et al., 1986). At low concentrations of free fatty acids (FFA) normally present in tissues, their nature and the presence of diacylgycerols and phosphorylated bases appear to be controlling phosphoglyceride biosynthesis. However, as the concentration of exogenous FA increases, the formation of triacylglycerols becomes dominant. In a study with cultured dissociated brain cells, Yavin and Menkes (1974) found that added labeled linoleic acid (18:3n-6) is first incorporated into di- and triglycerols. However, as the conversion of 18:3n-6 into longer more highly unsaturated analogues progressed, these tended to be transferred to ethanolamine phosphoglycerides from triacylglycerols and choline phosphoglycerides. In the liver, 80% of phosphatidylinositol (PI) contains stearate and arachidonate at the sn-1 and sn2 positions, respectively. This probably arises from deacylation/reacylation reactions in which the fatty acid constituents are remodeled following synthesis of the lipid molecule.
In 1968, Wirtz and Zilversmit identified a soluble intracellular protein in rat liver that was capable of binding phosphatidylcholine (PC) and transferring it from one population of donor membranes to a second population of acceptor membranes. Since this initial observation, phospholipid transfer proteins have







3

been identified in almost all mammalian tissues (Wirtz and Gadella, 1990). Phospholipid transfer proteins fall into three main categories: (1) those specific for PC, (2) those with high activity for PI and less but significant activity for PC, and (3) those with transfer activity with most phospholipids and cholesterol. The action of these proteins is a one-for-one exchange of lipid molecules between donor and acceptor membranes using ATP dependent and ATP independent mechanisms with time constants that can vary by several orders of magnitude for different lipids (see Voelker, 1991). Some lipids such as free fatty acids (FFA) or phosphatidic acid may have sufficient solubility to allow some passive transport, but most other lipids require these mechanisms. In summary, differences in membrane composition can be accomplished by the selective transport of lipid molecules and/or by specific metabolic events that occur in the acceptor membranes.

It is very clear that a major function of phospholipids is to store precursors of second messengers that activate specific processes in the cell. The eicosanoids were the first phospholipid-derived second messengers identified and will be reviewed later; this section will review the roles of phosphatidylisinositol (PI), diacylglycerol (DAG) and platelet activating factor (PAF) as intracellular second messengers. Hokin and Hokin (1953) observed that incubation of pigeon pancreas with acetylcholine caused the release of the digestive enzyme amylase. If 32P was included in the incubation, there was a rapid labeling of Pl. It is now known that acetylcholine binding to its cell surface receptor results in hydrolysis of inositol lipids to DAG and inositol phosphates. The 32P incorporation was the result of resynthesis and rephosphorylation of the lipids through a sequence of reactions known as the PI cycle (see Hawthorn, 1982). The initial event in inositol lipid metabolism occurs within 20-30 seconds







4


of the binding of the hormone to its receptor and involves primarily the catabolism of phosphatidyl inositol-4,5-biphosphate (PIP2) into DAG and inositol1,4,5-triphosphate (P3). For example, 20% of cellular PIP2 is degraded within 30 seconds of exposure of hepatocytes to vasopressin (Fisher et al., 1984). IP3 causes the release of Ca2+ from the endoplasmic reticulum. A receptor for 'P3 has been identified as a 313 kDa protein that has both Ca2+ channel activity and a ligand binding site (Ferris et al., 1989). The receptor is phosphorylated by cyclic adenosine monophosphate (cAMP)-dependent protein kinase (PKA) which causes a decrease in both 'P3 binding and Ca2. release.
The other product of PIP2 hydrolysis is DAG which activates PKC. This kinase was discovered by Nishizuka and co-workers in 1977 and seven isozymes have been identified, all containing regulatory and kinase domains (Nishizuka, 1988). PKC appears to exist in the cytosol in an inactive form and, when DAG is generated in the plasma membrane, is translocated to this membrane and activated. PKC also requires Ca2+ and phosphatidylserine (PS) for activity.

PAF (1-alkyl-2-acetyl-sn-glycerol-3-phosphocholine) belongs to a large family of ether-linked glycerolipids that serve as both membrane components and as intracellular mediators. Arachidonic acid plays an important role in PAF biosynthesis because it participates at the substrate level (alkyl-arachidonoylglycerophospho-cholines) in the phospholipase A2 (PLA2) catalyzed reaction (Ramesha and Pickett, 1986). Interestingly, it has been shown that PAF production can be blocked by inhibition of acetyltransferase activity in bovine neutrophils by arachidonic and oleic acids (Remy et al., 1989). Some species of ether lipids associated with membranes act as reservoirs for polyunsaturated fatty acids. The protective effect of ether-linked groups is probably due to their







5

ability to slow the rate of hydrolysis of the acyl group at the sn-2 position by PLA2. High affinity binding sites for PAF on the plasma membrane have been shown for platelets, neutrophils, smooth muscle cells, and lung tissue (see Snyder, 1990). It appears that PAF interaction with its receptor is coupled to a G protein, since PAF stimulates GTPase activity (Houslay et al., 1986). PAF is produced by preimplantation mouse (O'Neill, 1985), human (O'Neill et al., 1987) and sheep (Battye et al., 1991) embryos following in vitro fertilization. The ability of human embryos to secrete PAF is correlated with the establishment of pregnancy (O'Neill et al., 1987) and the use of PAF antagonists can inhibit implantation in mice (Spinks and O'Neill, 1987). Basically, PAF is an embryonic autocrine growth factor which can also suppress oxytocin-induced release of prostaglandins and extend estrous cycle length. The reader is referred to O'Neill (1990) for a review of the physiological role of PAF and its interaction with the arachidonic acid cascade in the stimulation of embryonic development.


Linoleic and Arachidonic Acids


All eukaryotic organisms contain polyenoic fatty acids in the complex lipids of their membranes and most mammalian tissues can modify acyl chain composition by introducing one or more double bonds. But animal requirements for polyunsaturated acyl chains can not be met exclusively by de novo metabolic processes. Animals are absolutely dependent on plants for linoleic (18:2n-6) and a-linolenic (18:3n-3) acids which are the two major precursors of the n-6 and n-3 fatty acids. In vitro studies indicate that 18:2n-6 competes with 18:3n-3 for the 2-acyl position of PC and for conversion to long chain metabolites (Brenner and Peluffo, 1966). Severe effects observed in experimental animals and







6


humans (Wene et al., 1975) in the absence of these dietary fatty acids include a dramatic decrease in body weight, dermatitis and increased skin permeability to water, enlarged kidneys, reduced adrenal and thyroid glands, decreased cholesterol accumulation and altered fatty acyl composition in many tissues, impaired fertility (Ziboh and Hsia, 1972), and ultimately death (see Sinclair, 1990). When exposed, in vivo or in vitro, to large amounts of unsaturated FA (LA, AA, 20:5n-3, 22:5n-3), the cells compensate by incorporating the fatty acid in low amounts and by reducing other fatty acids in order to maintain the membrane physico-chemical characteristics (Stubbs and Kisielewski, 1990). It seems reasonable to conclude that the cell will not allow the proportion of saturated and unsaturated FA to change appreciably because the membrane fluidity would change in an undesirable manner. The nature of the compensation mechanism which appears to allow sensing of the membrane's physical properties by the elongase-desaturase system is still unknown (Quinn, 1981). Stubbs and Kisielewski (1990) proposed that such a mechanism may be termed 'homeoviscous adaptation'.

The four n-6 acids in the sequence from LA to AA individually have similar potency to reverse these effects of a deficiency. AA (20:4n-6) can be formed from LA (18:2n-6) by the alternating sequence of A6 desaturation, then chain elongation of the 18:3n-6 intermediate to form 20:3n-6, and A5 desaturation resulting in 20:4n-6. AA can be further metabolized into higher polyunsaturated FA such as 22:4n-6 and 22:5n-6 (Ramwell et al., 1977). The A6-, A5desaturases are membrane bound extrinsic and immunologically distinct proteins bound to the cytoplasmic surface of the endoplasmic reticulum by hydrophobic tails and require an electron transport chain for desaturation (Fujiwara et al., 1984). Frequently, AA is referred to as an essential fatty acid







7

(EFA) since it is required in many tissues; but an adequate dietary supply of LA can be converted to AA in most circumstances. When ingested and absorbed, AA is transported bound to albumin (Spector et al., 1969) or lipoproteins as part of triglycerides, cholesterol esters or phospholipids (Lamberth and Oates, 1976). Normally, human plasma concentration of free AA are very low (~3 pg/ml). Contrarily to non-ruminants, dietary EFA absorption is reduced in cows because of biohydrogenation in the rumen (Viviani, 1970). Most of the AA absorbed is incorporated into phospholipids destined to be integral part of cell membranes (Flower and Backwell, 1976). In the liver and most other tissues of animals in a normal, balanced state, the only members of the n-6 family to accumulate in relatively large quantities are LA and AA; much lower levels of intermediates 18:3n-6 and 20:3n-6 are detected (Cook, 1991). This suggests that A6desaturase is a rate-limiting step in the enzymatic sequence. On the other hand, A5-desaturase activity measured in vitro with 20:3n-6 as substrate is approximately equal to, or lower than the limiting A6-desaturase activity (Cook, 1978; Sprecher, 1981). The activities of A6-, A5-desaturases are under endocrine control. Corticosteroids, either in vivo or when added to hepatocytes in culture, inhibit A5- and A6-desaturase activity trough direct synthesis of a "lipocortin" like (e.g., PLA2 inhibitor) soluble protein (Brenner, 1990). The precise mechanism by which this cytosolic protein inhibits the desaturase reaction is not clear; prevention of substrate access to the enzyme rather than a direct inhibition of the enzyme is a possibility.
Concentrations of free AA in the cytoplasm and plasma are very low compared with esterified levels of AA (Irvine, 1982). In plasma, it is bound electrostatically and hydrophobically to albumin (Spector, 1975). The plasma concentration varies from 3 pg/ml (1 to 2% of total free fatty acid) in humans to







8

22-122 pg/ml in dogs (Ramwell et al., 1977) and 0.15-0.22 pg/ml in guinea pigs (Leaver and Poyser, 1981). In guinea pig uterus, 93% of total AA was esterified to phospholipids. On day 7 of the estrous cycle, 54% was found in PE, 33% in PC and 12% in PS. At day 15, the amount present in PE decreased significantly to 33%, while PS and PC contained about 20% (Leaver and Poyser, 1981). In endometrium of cyclic and pregnant cows at day 18 postestrus, 73% of the total AA was esterified to phospholipids (Lukaszewska and Hansel, 1980). Curl (1988) determined the distribution of AA in the most common phospholipid classes. At day 17 postestrus, the quantity of AA bound to phospholipids in endometrium from pregnant cows was lower than in cyclic cows. Overall, 64% was found in PE, 18% in PC, while PS and PI contained about 10% of total AA esterified to phospholipids (Curl, 1988). The amount of esterified AA in the endometrium declined considerably between days 17 and 19 of the estrous cycle, probably because of prostaglandin production associated with luteolysis. On the other hand, the amount of esterified AA increased between days 17 and 19 of pregnancy (Curl, 1988).

In guinea pig uterus, the 7% of AA not esterified in phospholipids was distributed as follows: triglycerides > cholesterol esters > free = diglycerides > monoglycerides (Leaver and Poyser, 1981). In bovine endometrium, the distribution of non phospholipid associated AA was: free > cholesterol esters > triglycerides (Lukaszewska and Hansel, 1980).


Arachidonic Acid Mobilization


In resting tissues, the levels of free AA are low and controlled by acylCoA hydrolase for release, and by acyl-CoA transferase for esterification into







9

phospholipids. In macrophages from bone marrow, AA liberation could be observed only if acylation was inhibited (Kroner et al., 1981). When human platelets were exposed to a specific inhibitor of lysophosphatide acyltransferase, PGE2, prostacyclin and thromboxane B2 secretions increased significantly (Goppelt-Struebe et al., 1986). These observations suggest that control of AA release can involve acyltransferases and future research may prove acyltransferases to be critical enzymes in the control of free intracellular AA.
The major rate limiting step for production of eicosanoids is AA release. The term eicosanoids is used to designate a group of oxygenated, C20 fatty acids derived from AA. In many tissues, membrane phospholipids are highly enriched with AA at the sn-2 position (Flower and Blackwell, 1976). Importantly, release of arachidonate is a selective process and other fatty acids normally found in phospholipids are not mobilized (Dennis, 1987). Lands and Samuelsson (1968) addressed the legitimate question as to whether AA was converted into prostaglandins (PG) on the phospholipids and stored in an esterified form and released in response to specific stimuli. They concluded that the primary rate limiting step in PG synthesis was the release of AA from phospholipids or triacylglycerols that would result from activation of lipases. Perfusion of the adrenal gland with acetylcholine in the rat was followed by the release of PG into the perfusion fluid. The presence of Ca2. in the perfusion fluid was essential for the release of PG (Shaw and Ramwell, 1967), which suggested that Ca2. may be needed for the activity of the phospholipase that releases the PG precursors from phospholipids. PC and PE, known to carry polyunsaturated FA at sn-2 position of the glycerol moiety, were originally favored as source of PG precursors through the action of PLA2. Release of AA from cellular stores in response to stimuli including hormones, growth factors, physical stress, is a







10

very quick process (see Smith, 1989). However, the rapid turnover of phosphoinositides in various tissues and cells called the attention to phosphoinositides as a possible source of AA for PG synthesis (Dawson and Irvine, 1978). In human platelets, PLC can cleave phosphoinositol from such lipids, yielding a 1,2-diacylglycerol, which after hydrolysis by DAG lipase, provided AA for PG synthesis. Prescott and Majerus (1983) demonstrated that DAG lipase, while equally active for 1 -stearoyl- and 2-arachidonoylgycerol, hydrolyses the 1stearoyl-2-arachidonoyl-glycerol sequentially, releasing first stearic acid and then AA. Domin and Rosengurt (1993) demonstrated that platelet-derived growth factor (PDGF)-stimulated release of AA and its metabolites from Swiss 3T3 cells is a biphasic process. The initial phase involves a rapid activation of PLA2 and is independent of de novo RNA and protein synthesis, whereas the second major phase is dependent upon rapid expression of a protein.


Phospholipases


The phospholipases are a group of enzymes with the common property of hydrolyzing phospholipids. The classification of the phospholipases is based on their site of attack on the phospholipids and several types of phospholipases are involved in liberating fatty acids from phospholipids. Phospholipase A, (PLA1) removes the acyl group from carbon sn-1 and PLA2 from carbon atom 2 of the phospholipid (Figure 1-1). Some phospholipases hydrolyze both acyl groups and are termed phospholipases B (PLB). Cleavage of the glycerophosphate bond is catalyzed by PLC, while the removal of the base group (e.g., choline, ethanolamine, serine) is catalyzed by phospholipase D.


. . :r""








16)
20 MD6 P-Cholbe



OH
20 o6 jjOh 18 , -CoA h

1-AT
C A

18:0
20 ,1o6 P-Chohe

2034o6

1810
O H
18:2o6 -C P-Chobe

2-AT
C QA

181)
18:206 P-Cho]be


Figure 1-1: The fatty acids (palmitic, 16:0; arachidonic, 20:40)6) at the sn-i and
sn-2 positions of phosphatidyicholine (P-choline) can be deacylated by phospholipases (PLA, and PLA2) and other fatty acids (stearic, 18:0; linoleic 18:2&'6) can be reacylated by acyltransferases (acylCoA:lyso Phosphatidyl choline 1-acyl transferase, 1-AT; acylCoA:lyso Phosphatidyl
choline 2-acyl transferase, 2-AT).


Two PLA1 have been purified from Escherichia coli based on their sensitivity or resistance to detergents (see Dennis, 1983). The detergent sensitive enzyme preferentially degrades phosphatidylglycerol and can act as a







12

transacylase. Transacylation occurs when an acyl-enzyme intermediate is formed in a two-step reaction where in the second step, the acyl group is transferred nonspecifivally to a hydroxyl acceptor of a soluble alcohol or of a lipid (Waite, 1991). Transacylation reactions are now known to be important events in many cells and provide a means to redistribute acyl groups between phospholipids without a deacylation-reacylation cycle. Rat liver lysosomes contain a soluble PLA, that is a glycoprotein with optimal activity at pH 4.0 and does not require Ca2+ for activity (Van den Bosch, 1980). The action of PLA, results in the production of a lysophospholipid that can be further cleaved through the action of specific lysophospholipases.
The distinction between PLB and lysophospholipases is not clear since PLB that has been purified and characterized has a high lysophospholipase activity (Kawazaki and Saito, 1973). Such activity would be expected since PLB deacylates at both sn-1 and sn-2 positions, with a lysophospholipid intermediate that is subsequently deacylated. The sum of two activities of the enzyme, PLA2 and lysophospholipase, gives the total activity of PLB; but it is unclear if the enzyme has two distinct binding sites (one for diacylphospholipids and a second one for lysophospholipids).

Phospholipases A2 were the first phospholipases to be described: Bokay (1877-1878) recognized a factor that degrades PC in pancreatic fluid and Kyes (1902) observed that cobra venom had hemolytic activity on erythrocytes. Since then a number of cellular PLA2s from mammals have been purified and compared to their pancreatic and venom counterparts. The nomenclature of the different isoforms of PLA2 has been based on structure and sequence. The 14 kDa enzymes have been categorized into types I (mammalian pancreatic and cobra venom), I (mammalian platelet, synovial fluid and viper venom) and Ill







13

(bee venom). The 85 kDa PLA2 has been given the designation of type IV (mammalian platelet, macrophage).
The most studied mammalian Ca2+ dependent PLA2s are the type 114 kDa PLA2 digestive enzymes which are released as proenzymes from the pancreas (Abita et al., 1972). Pancreatic-like PLA2s exist in a variety of tissues such as lung and kidney (Sakata et al., 1989) and can bind to high affinity cell surface binding sites (Hanasaki and Arita, 1992). Human type I PLA2 has been characterized as active at pH 7-10 and requires Ca2+ as a cofactor. This enzyme does not exhibit a preference for certain FA in the sn-2 position but is fairly selective toward the phospholipid class in that it prefers PE or PS and poorly hydrolyses PC (Kramer et al., 1989; Hara et al., 1989). Crystallization of the enzyme has shown that the catalytic site and mechanism are conserved relative to other type I and 11 14 kDa PLA2, but that subtle differences exist such as variation in the hydrophobic channel that binds substrate (Wery et al., 1991; Scott et al., 1991). The proposed role of Ca2+ catalysis is to orient the phospholipid by binding the sn-3 PO4 and to activate the substrate by binding the carbonyl oxygen of the sn-2 fatty acyl group (Mayer and Marshall, 1993). Optimal activity is observed with aggregated forms of phospholipid such as membranes, vesicles and micelles and considerably weaker activity with monomeric or dispersed substrates. The role of cytokines in the production and release of extracellular type 1114 kDa PLA2 was studied recently using models of septic shock. Injections of endotoxin lipopolysaccharide (LPS) produced an elevation of PLA2 activity in rabbit serum associated with an increase in PLA2 protein level (Pruzanski and Vadas, 1991). This increase occurs much later than the endotoxin-induced secretion of cytokines such as tumor necrosis factor (TNF), interleukin-1 (IL-1) and IL-6, which is consistent with a cytokine-mediated







14

induction of PLA2 synthesis and release. Pruzanski and Vadas (1991) hypothesized that LPS stimulates macrophages to release cytokines and these induce secretion of PLA2 from a variety of sources, propagating the inflammatory process. Type 11 14 kDa PLA2 exists also in a cell-associated form that can serve a number of functions, such as phospholipid turnover, repair of lipid peroxidation, or mobilization of AA for generation of PG or other derivatives (Pernas et al., 1991).
The 85 kDa PLA2 has been identified in cells as an activity with 2 specific characteristics: (1) a strong preference for sn-2-arachidonoyl phospholipids and no preference for the type of sn-1 acyl linkage (acyl versus alkyl) or for the type of phospholipid functional group (ethanolamine versus choline) (Diez et al., 1992); and (2) a translocation activity such that it is always found in the cytosolic fraction when cells are broken in the presence of Ca2+ chelators (EDTA or EGTA), but is largely in the microsomal fraction when broken in the presence of free Ca2+ concentrations (300 to 700 nM) approaching those found in activated cells (Rehfeldt et al., 1991; Krause et al., 1991). Analysis of the sequence revealed a 45 amino acid domain homologous to the Ca2+ binding domain of protein kinase C (PKC) but no other sequence homology including type 1114 kDa PLA2 (Sharp et al., 1991). Ca2. is required for translocation but catalytic activity can be obtained in the absence of Ca2+ using high concentration of NaCl, suggesting that Ca2. is required for association with membrane lipids but not for catalysis (Wijkander and Sundler, 1992). Recently, a side by side in vitro comparison of the 85 kDa PLA2 and type 11 14 kDa PLA2 revealed that both enzymes have similar activity profiles as a function of Ca2+ concentration (Marshall and McCarte-Roshak, 1992). The 85 kDa PLA2 has a clear preference for sn-2-arachidonoyl over sn-2-oleoyl-phospholipids, while sn-2-palmitoyl is a







15

very poor substrate (Clark et al., 1990; Diez et al., 1992). The degree of saturation seems to be the only determinant of good substrate activity. Lin et al. (1992) provided evidence that 85 kDa PLA2 can be involved in hormonestimulated release of AA using transfected CHO cells over-expressing this enzyme. Treatment of these cells with thrombin or ATP resulted in increased AA release compared to control cells without over-expressed enzyme. Interestingly, CHO cells over-expressing 14 kDa PLA2 did not show increased AA release in similar experiments (Clark et al., 1991). In addition, 85 kDa PLA2 has been shown to possess a lysophospholipase Al activity which is not observed with the 14 kDa PLA2 (Leslie, 1991). Coexistence of type 11 14 kDa PLA2 and 85 kDa PLA2 in the same cell or tissue has been reported. For example, the macrophage cell line P388D, was shown to have distinct cytosolic and membrane-associated Ca2+-dependent PLA2 as well as cytosolic lysophospholipase activity (Rose et al., 1985).
Another class of sn-2-acylhydrolase activity has been described in almost all major organs: Ca2+-independent PLA2 which is largely uncharacterized and may represent one or more isozymes (Pierik et al., 1988). A 40 kDa Ca2+independent, sn-2-arachidonoyl selective isozyme has been reported in rabbit and human myocardial tissues (Hanzen et al., 1991; Hanzen and Gross, 1992, respectively). In human ischemic heart tissue, Ca2+-independent PLA2 was found predominantly in the microsomal fraction and accounted for 98-99% of total PLA2 activity (Hanzen et al., 1991). Another Ca2+-independent, sn-2arachidonoyl PE selective activity (58 kDa dimeric isozyme) has been identified in sheep platelet cytosol and was recently cloned (Zupan et al., 1992). For more detailed reviews on phospholipases, the reader is referred to Waite (1991) and Smith (1992).







16

As originally defined, lipocortins are glycoproteins which specifically inhibit PLA2 in vitro and in vivo (Di Rosa et al., 1984). Lipocortins are related to calpactins which fall under the broad umbrella of annexins (Crumpton and Dedman, 1990). Lipocortin I (= p35 = calpactin 11) was the first to be cloned and sequenced (Wallner et al., 1986) followed by lipocortin I (= p36 = calpactin I; Kristensen et al., 1986). The p36 protein is a major substrate for phosphorylation by PKC (Gould et al., 1986). Additional related proteins (lipocortins IlIl-VI) have been identified which all bind Ca2+ and phospholipids (see Pepinsky et., 1988; Hunter, 1988; Klee, 1988; and Goulding and Guyre, 1992 for reviews). A decrease in annexin 1 expression has been found to be associated with the onset of labor (Lynch-Salamon et al., 1992). An antiphospholipase protein named gravidin has been described in amniotic fluid and its production from the amnion decreases with labor (Wilson and Ganendren, 1992). However, the way in which these proteins inhibit phospholipase activity remains controversial. In a commentary, Davidson and Dennis (1989) discussed the foundation of lipocortin's inhibitory activity and suggested that substrate (phospholipids) and/or Ca2+ depletion could explain some of the experimental observations.

Phospholipases C are phosphodiesterases that cleave the glycerophosphate bond on phospholipids resulting in production of DAG. One of the earliest reports of a mammalian PLC came from Sloane-Stanley (1953) who demonstrated the release of inositol from phosphatidylinositol, catalyzed by a brain preparation. PLC have now been purified from the cytosolic fraction of muscle, brain, platelets and ram seminal vesicles. The properties of phospholipases C purified from diverse sources have been compared, and Rhee et al. (1989) developed a nomenclature. So far five basic types are listed (a, 0, y







17

5, e); each has distinct immunological properties and little homology in their predicted amino acid sequences. PLC is part of a second messenger system whereupon activation of cell surface receptors, PLC is activated and hydrolyses PIP2 into two second messengers: IP, is released into the cytoplasm where it interacts with Ca2+ storage sites to mobilize intracellular Ca2+; and DAG acts within the plasma membrane to activate PKC (see review by Rhee et al., 1989). Direct evidence for involvement of trimeric G proteins in P1/Ca2. signaling required observation of GTP-dependent activation of PLC (Cockcroft and Gomperts, 1985; Litosch et al., 1985). Binding of the ligand to the extracellular portion of the receptor results in dissociation of the GDP from the G protein (a subunit) and replacement by GTP; the GTP-bound a subunit then dissociates from the Oy subunits and activates PLC (Bourne et al., 1990; Birnbaumer, 1990). There are several isoforms of the G protein a-subunits (ai, as,aq) and, for example, activation of PLC occurs via aq (Smrcka et al., 1991). Oxytocin stimulates production of prostaglandins via receptor activation of PLC, which could also be in part regulated by activation of PKC (Schrey et al., 1986). There is also evidence that phosphorylation is involved in regulation of PLC activity. PLCy is a substrate for tyrosine kinase and contains a conserved sequence motif known as Src homology 2 (SH2) domain which allows PLCy to bind in a specific manner to certain tyrosine-phosphorylated proteins. Via SH2 domains, PLCy is able to bind to specific tyrosine residues within the cytoplasmic domain of certain receptor tyrosine kinases when these tyrosines become autophosphorylated during the receptor activation (Anderson et al., 1990; Margolis et al., 1990). This could be involved in the translocation process of cytosolic PLC to the membrane for activation. Finally, the major pathway of sphingomyelin degradation involves a special phospholipase C, sphingomyelinase. Although some phospholipases







18
C that act on PC also work on sphingomyelin, a number of distinct sphingomyelinases exist (Waite, 1987). Some interest has focused on sphingomyelinase in the plasma membrane that, when coupled to ceramidase action, yields sphyngosine, a negative regulator of PKC (Merrill, 1989).
Although PLA2and PLC appear to be the primary enzymes responsible for increasing free intracellular AA, release of AA may not be the only rate limiting step in the production of eicosanoids. It seems that phospholipid substrate availability for PLA2and PLC can play an important role in subsequent release of AA. Phospholipids are not distributed randomly among all cellular membranes, but instead different lipid classes (e.g. phospholipids, triglycerides, cholesterol esters, free fatty acids, etc.) are often enriched in different membranes. This enrichment may involve site-specific synthesis or degradation of phospholipids, remodeling through deacylation-reacylation reactions, translocation or some combination of these mechanisms (see Pagano, 1990). Thus the capacity of cells to release AA and synthesize eicosanoids is directly dependent upon an adequate distribution of phospholipids in their membranes. The following sections will present information regarding the metabolism of AA into eicosanoids with a particular emphasis on the cyclooxygenase pathway which is responsible for the production of prostaglandins.


The Cyclooxyqenase Pathway


Once released through the action of phospholipases, free AA can enter the 'arachidonate cascade'leading to the eicosanoids. This cascade includes three major pathways: cyclooxygenase, lipoxygenase and cytochrome P-450 epoxygenase. Prostanoids, which include prostaglandins and thromboxanes, are







19

formed via the cyclooxygenase pathway (Needleman et al., 1986); leukotrienes and lipoxins are formed via the lipoxygenase pathway (Needleman et al., 1986); and epoxides and diols are formed through the P-450 epoxygenase pathway (Fitzpatrick and Murphy, 1989).
Cyclooxygenase is often used as a synonym for prostaglandin endoperoxide synthase [E.C.1.14.99.1] also called PGH synthase (PGHS) but, in fact, this enzyme exhibits two different activities: (1) a cyclooxygenase (bisdioxygenase) activity that generates prostaglandin G (PGG) from polyunsaturated FA; and (2) a hydroperoxidase activity that converts PGG into PGH. These two activities are associated with the same heme-binding glycoprotein that has been purified to homogeneity from microsomes of sheep and bovine seminal vesicles (Hemler et al., 1976; Van der Ouderaa et al., 1977; Miyamoto et al., 1976). Cyclooxygenase and peroxidase activities have distinct binding sites for their lipid substrates which are located on the cytoplasmic side of the endoplasmic reticulum (Marshall and Kulmacz, 1988). Recent evidence points to the existence of multiple isoforms of PGHS (Xie et al., 1991; Kujubu et al., 1991; O'Banion et al., 1992). Constitutively expressed PGHS-1 has a mRNA size of 2.7-3.0 kb and the inducible form (PGHS-2) 4.0-5.5 kb (O'Banion et al., 1992; Diaz et al., 1992). PGHS-1 is an integral membrane homodimer with a subunit molecular weight of 72 kDa. The amino acid sequence deduced from the cDNA for the sheep enzyme indicates a molecular weight of 65.5 kDa and the presence of four consensus sites (Asn-X-Ser/Thr) for N-glycosylation (DeWitt, 1991). The sequences of cDNA clones for PGHS from sheep, mouse and human indicate a high degree of homology between species and that the protein initially contains a signal peptide of 24-26 amino acids which in all species is cleaved to a mature 576 amino acid protein (Smith and Marnett,







20


1991). The presence of the signal peptide suggests that the enzyme crosses the endoplasmic reticulum during its synthesis, but it is unclear how the mature PGHS is associated with the membrane since it contains no obvious transmembrane hydrophobic sequences (Smith and Marnett, 1991). PGHS-1 is probably responsible for the constitutive production of prostaglandins involved in "housekeeping" functions (Smith, 1989). PGHS-2, which has about 62% homology with PGHS-1, is expressed following cell activation (Xie et al., 1991). The synthesis of PGHS-2 is stimulated by serum, growth factors and phorbol esters in fibroblasts (Kujubu et al., 1991) and by chorionic gonadotrophin in granulosa cells (Sirois and Richards, 1992, Sirois et al., 1992).

The substrates for cyclooxygenase activity are two molecules of oxygen and one molecule of FA. A variety of polyunsaturated FA can be substrates for cyclooxygenase. FA containing at least three methylene-interrupted cis-double bonds beginning at n-6 are converted into PGG derivatives (Hamberg and Samuelsson, 1967a). The rates of oxygenation vary considerably with the number of carbons in the FA and the presence of additional double bonds. The best substrates (Km= 2-10 pM) for the cyclooxygenase reaction are AA (20:4n-6) and di-homo-y-linolenic acid (20:5n-6)( Marshall et al., 1987; Odenwaller et al., 1991). In vivo, the most common cyclooxygenase substrate is AA but unsaturated FA containing two methylene-interrupted double bonds (20:2n-6 and 18:2n-6) can be oxygenated to monohydroxy FA (Hemler and al., 1978).

Cyclooxygenase catalysis involves an initial activation of the substrate FA followed by oxygen insertion, carbon skeleton rearrangement and a second oxygen insertion (Hamberg and Samuelson, 1967b). In fact this process is very similar to non-enzymatic autoxidation of some polyunsaturated FA (Nugteren and Van Dorp, 1966). Cyclooxygenase has a strict requirement for a







21

hydroperoxide to activate the 13-pro-S hydrogen on the FA substrate. If hydroperoxide concentration is below 10 nM, cyclooxygenase activity is inhibited. The best peroxide activators are 15-HpETE and PGG which are also the best substrates for the peroxidase activity of PGS (Kulmacz and Lands, 1983).
The peroxide activity of PGHS catalyses the reduction of various hydroperoxides to alcohols and a heme group is essential for this activity. PGG generated by the cyclooxygenase is the major endogenous substrate reduced into PGH2, which is the substrate for production of all PG and thromboxanes (Samuelsson, 1978). The peroxidase activity of PGHS resembles other peroxidases, including horseradish peroxidase and cytochrome c peroxidase, in being non-specific toward the electron donors. The endogenous electron donor is not known but, among naturally-occurring molecules, the best substrates are epinephrine and uric acid (Markey et al., 1987). Uric acid is not an efficient reductant but is present in high concentration (~ 300 pM) in human plasma (Smith and Marnett, 1991), likewise GSH is a relatively poor reducing substrate but is present in very high concentrations in most cells (~ 6 mM) (Marshall et al., 1987).
Prostaglandin synthesis appears to be cell specific considering that often one cell type produces predominantly one type of PG (Smith et al., 1991). The PG are classified as belonging to the '1', '2' or '3' series depending on the number of double bonds they contain. Within each series, PG are further classified as 'A', 'D', 'E', 'F' or 'I' types, depending on the presence of various oxygen-containing substituents. These substituents are critical for the biological activities of the various PG; the most important positions on the carbon skeleton being C-9, C-10 in the ring structure and C-15 in the side chain (Grastr6m,







22

1981). For example, the primary difference between PG of the "F" series and the "E" series is the substitution of a hydroxyl group (F) rather than an oxygen

(E) at position C-9. Little is known about the enzymes converting PGH into the various PG. Prostaglandin F2a (PGF2 ) is formed following (1) reduction of PGH2 by PGF synthase, or (2) reduction of prostaglandin E2 (PGE2) by a 9-ketoreductase. PGF synthase has been purified from bovine lung by Watanabe et al. (1985) and the cDNA sequence of this enzyme contains a 323 amino acid open reading frame coding for a protein with a molecular weight of about 36 kDa (Watanabe et al., 1989). This enzyme also converts PGH2 to prostaglandin D2 and, at a much lower rate, PGE2 to PGF2a. A 9-keto-reductase activity, which catalyses conversion of PGE2 into PGF2a, has been found in renal cortex (Korff and Jaraback, 1982) and placenta (Jaraback et al., 1983) but not in bovine uterine tissue (Wlodawer et al., 1976). However the Km value for PGE2 is quite high (~ 300 pM) suggesting that PGE2 synthesis is not the natural substrate and that in vivo PGF2a synthesis is unlikely to involve this enzyme.

PGE2 is synthesized by non-oxidative rearrangement of the endoperoxide PGH2. This reaction can be catalyzed by a GSH-dependent PGE synthase [E.C.5.3.99.3] (Ogino et al., 1977); but there is also evidence for a nonenzymatic rearrangement (Hamberg and Samuelsson, 1973). Partially purified PGE synthase from ovine vesicular glands was reported to have a molecular weight of 60-70 kDa (Moonen et al., 1982). Immunoprecipitation using monoclonal antibodies indicated that a 17.5 kDa protein is responsible for at least 45% of the total PGE synthesis activity in ovine vesicular microsomes (Tanaka et al., 1987).
Metabolism of PGF2a and PGE2 follows essentially the same pathway, with metabolism occurring primarily in the lungs. First, oxidation of the 15-







23

hydroxyl group to a 15-keto group is catalyzed by the 1 5-hydroxy-prostaglandin dehydrogenase enzyme. Then it undergoes reduction by 13,14-prostaglandin reductase to form 1 5-keto-1 3,14-dihydro-prostaglandin, the main plasma metabolite. A limited amount of PGF2a is metabolized in the endometrium (Curl et al., 1983; Guibault et al., 1984; Knickerbocker et al., 1986a). Urinary metabolites are formed by one or two steps of p or o oxidation (Kindhal, 1980). Both 1 5-keto-prostaglandin and 1 5-keto-1 3,1 4-dihydro-prostaglandin have virtually no biological activity (Behrman and Hall, 1982). However, injections of 10 or 15 mg of 13,14-dihydro-PGF2, at day 10 postestrus can induce premature luteolysis and estrus in the cow (Milvae and Hansel, 1983).

For a more extensive review of enzymes involved in metabolism of PGH, the reader is referred to Needleman et al. (1986) and Smith et al. (1991).


Inhibition of Cyclooxyqenase


There appear to be two different inactivation processes for PGHS: (1) auto inactivation of cyclooxygenase ('suicide' reaction) and (2) peroxidedependent inactivation of cyclooxygenase and peroxidase activities. When PGHS is incubated in presence of AA, the cyclooxygenase activity is inactivated before all the AA is metabolized (Smith and Lands, 1972). The rate of this 'suicide' reaction is linked directly to the rate of cyclooxygenase catalysis but is relatively independent of newly formed metabolites. 'Suicide' inactivation of cyclooxygenase appears to be an intrinsic property of this catalytic activity. Each mole of purified PGHS can form about 1300 moles of PGG2 before the cyclooxygenase activity disappears, although most of the peroxidase activity remains (Marshall et al., 1987).







24

Inhibition of cyclooxygenase activity can potentially involve several factors: (1) reduction in concentrations of either oxygen or FA substrate; (2) reduction in concentration of hydroperoxide necessary for initiation; (3) interference with AA binding to the cyclooxygenase active site.
Molecular oxygen concentration does not appear to be a limiting factor on the rate of cyclooxygenase since the K, for 02 for cyclooxygenase is about 5 pM and 02 concentrations in tissues are greater than 20 pM (Smith and Marnett, 1991). The concentration of unesterified AA in tissues is approximately 20 pM which is well above the Km of AA for cyclooxygenase (Lands et al., 1978; Marshall et al., 1987).
Both cyclooxygenase and peroxidase activities are activated in presence of hydroperoxides (Hemler and Lands, 1980; Markey et al., 1987). In cells that form prostaglandin endoperoxide only upon hormonal stimulation, free AA must be inaccessible to cyclooxygenase. Cyclooxygenase activity is blocked by reduced glutathione (GSH) and in the presence of excess GSH peroxidase (Smith and Lands, 1972). Marshall et al. (1987) described the presence of cytosolic factors in ram seminal vesicles that potentiate the inhibitory effect of exogenous GSH peroxidase. They reported that many cells contain levels of GSH and GSH peroxidase that would inhibit cyclooxygenase activity in vitro. However prostaglandin synthesis may still proceed in such tissues because of cell compartmentalization of cyclooxygenase (membrane-bound) and GSH peroxidase (cytosolic) (Marshall et al., 1987).
Lands et al. (1972) demonstrated that n-3 and n-9 FA having 18 to 22 carbons cause significant inhibition of cyclooxygenase activity. Pace-Asciak and Wolfe (1968) reported that linoleic (18:2n-6) and linolenic (18:3n-6) acids inhibit PGHS from sheep seminal vesicles but concentrations required for inhibition







25

were high (1.8 to 5.0 mM); they also noted that pre-incubation with the FA resulted in a greater degree of inhibition, especially at low substrate concentrations. Other n-3 unsaturated FA (18:3n-3, 20:5n-3 and 22:6n-3), which are not substrates for cyclooxygenase, can still compete for binding at the substrate site. In the case of 20:3n-3, the competitive inhibition constant was 6 pM. This competitive binding was stronger for the more highly unsaturated acids. The activity of cyclooxygenase toward 20:5n-3 is less than 50% of that observed with AA, a fact that may account for the anti-thrombogenic effect of fish oils (Culp et al., 1979). Another constituent of fish oils, docosahexaenoic acid (22:6n-3) is a strong inhibitor of AA oxygenation with a K, of about 0.36 pM (Corey et al., 1983). On the other hand, the n-6 acids such as linoleic cause both an instantaneous competitive inhibition and a progressive and nonreversible loss of cyclooxygenase activity (Lands et al., 1972). Acetylenic analog of AA: eicosa-5,8,11,14-tetraynoic acid (ETYA) is known to irreversibly inhibit cyclooxygenase in the presence of oxygen and hydroperoxide activators (Vanderhoek and Lands, 1973). Two cis-epoxieicosatrienoic acid (EET) isomers, derived from the epoxygenase pathway, have been found to inhibit AA metabolism by the ram seminal vesicle cyclooxygenase (Fitzpatrick et al., 1986). Concentrations required for 50% inhibition were 30 ptM for 14,15-cis-EET and 50 pM for 8,9-cis-EET. The potency of 14,15-cis-EET surpassed that of nonsteroidal anti-inflammatory drugs such as ibuprofen and aspirin. In platelet suspensions, this inhibitory effect of 14,15-cis-EET was confined to the cyclooxygenase enzyme; the lipoxygenase enzyme was not affected (Fitzpatrick et al., 1986).
Inhibition of cyclooxygenase activity can also be achieved using steroidal or non steroidal anti-inflammatory drugs. Anti-inflammatory steroids appear to







26

diminish cyclooxygenase activity by inhibiting translation of the mRNA for PGHS (Koehler et al., 1990; Sebalt et al., 1990). In general, non-steroidal drugs compete with AA for cyclooxygenase active site even though they bear very little or no structural resemblance to AA. Subsequent to competitive inhibition, several of these drugs form covalent (aspirin = salicylic acid ) or non-covalent bonds (indomethacin) with the enzyme resulting in irreversible inactivation of cyclooxygenase but not peroxidase activity (Smith et al., 1991; Smith and Marnett, 1991). Meade et al. (1993) demonstrated that non-steroidal antiinflammatory drugs can exhibit some selectivity toward PGHS-1 or PGHS-2.


Prostaglandin Receptors


Prostaglandin receptors are classified as FP, EP, DP, IP and TP, with subsets of EP-1, EP-2 and EP-3; the first letter corresponding to the type of prostaglandin that is the major ligand (Kennedy et al., 1983). The subset definitions are based on responsiveness to drugs; for example, EP-1 and EP-3 mediate contractile activity of PGE2 whereas EP-2 mediates relaxant activity (Senior et al., 1992). Water reabsorption in the kidney collecting tubule is one of the most extensively characterized systems to study cellular and molecular effects of PGE2. Arginine vasopressin (AVP)-induced water reabsorption is mediated by cAMP and stimulates PGE2 synthesis. The inhibitory and stimulatory effects of PGE2 have been shown to be mediated through distinct stimulatory (EP-2) and inhibitory (EP-3) PGE2 receptors (Smith, 1989). Addition of GDP or GTP increased binding affinity of PGE2 to cell membrane preparations by 2 fold, and this effect of guanine nucleotides on binding was eliminated by treatment with pertussis toxin (Smith, 1989). Specific association of PGE2







27


receptor with G, has been demonstrated in canine and rabbit kidney (Watanabe et al., 1986). Recently, Sugimoto et al. (1992) cloned a functional EP-3 receptor from a mouse lung cDNA library, and the predicted structure suggested that EP3 was a member of the G protein-linked receptor family.


Estrous Cycle


Sexually mature cows display recurring periods of sexual receptivity which defines the boundary of the period referred as the estrous cycle. The sexual receptivity lasts for about 16.9 4.9 h (Schams et al., 1977) and the estrous cycle averages approximately 20 to 21 days (Hansel et al., 1973). The estrous cycle can be can divided into four phases: (1) proestrus, during which final maturation of the preovulatory follicle occurs; (2) estrus, which is the period of sexual receptivity; (3) metestrus, which is the period of ovulation and development of the corpus luteum (CL); and (4) diestrus, the period dominated by a functional CL and as a result high plasma concentrations of progesterone. These four periods involve endocrine and biochemical changes which result in the release of an ovum, and, if fertilized, possible establishment of pregnancy. Failure of the embryo to survive leads to reoccurrence of the cycle. The estrous cycle is the result of complex interactions between hypothalamus, pituitary, ovary and uterus.


Follicular Growth, Maturation and Dominance


In the cow, the number of primordial follicles (~150,000) present in the ovaries is fixed at birth, but there is a natural reduction to approximately 3000 by







28

the age of 15 to 20 years (Erickson, 1966). The regulation of the rate at which the non-growing follicles are stimulated to start development is not well understood but it appears to be partly dependent on the size of the pool of primordial follicles (Krohn, 1966). Once development of a follicle is initiated, it is continuous until the follicle is ovulated or becomes atretic (Peters and Levy, 1966). In sheep, it takes approximately 6 months for a primordial follicle to develop into a large dominant follicle (Cahill and Maul6on, 1980). A primordial follicle consists of the ovum surrounded by a single layer of epithelial cells. Once stimulated, the follicular epithelial cells proliferate, become cuboidal, and form several layers of granulosa cells characteristic of a secondary follicle. The ovarian stroma contributes a vascularized theca interna and theca externa which are separated from the non vascularized granulosa layer by a basal lamina. When the antrum develops, the follicle is termed an antral follicle in which the oocyte is suspended by the cumulus oophorus granulosa cells in the antrum. Mature antral follicles with a diameter greater than 10 mm in diameter are referred to as Graafian follicles. It appears that the development of a primary follicle to the antral stage does not require gonadotropins, since in sheep antral follicles up to 2 mm in diameter were observed in hypophysectomized animals (Driancourt et al., 1979; McNatty et al., 1990). Follicle stimulating hormone (FSH) and luteinizing hormone (LH) are required for further development to ovulatory size (Lohstroh and Johnson, 1966).

Both granulosa and theca cells are necessary for estrogen production by the follicle. The two cell theory for estrogen production was demonstrated by Fortune and Armstrong (1978). Theca cells produce androgens that are converted into estrogens by granulosa cells containing aromatase enzyme. LH stimulates synthesis of androgens from cholesterol in thecal cells, whereas FSH







29

stimulates production of aromatase enzyme (Richards, 1980) and LH receptors (McNatty, 1979). Carson et al. (1981) noted a reduction in the ratio of estradiol to androgen in total follicular fluid was associated with atresia in all sizes of follicles. They suggested that the decrease in aromatase activity was of major importance in the process of atresia. This led to the classification of estrogen active and estrogen inactive follicles (Ireland and Roche, 1982). Estrogen active follicles have much higher capacity to bind gonadotropins than estrogen inactive follicles, and this capacity is correlated to follicular diameter (Ireland and Roche, 1983).
Pierson and Ginther (1984, 1988) reported that there appear to be two waves of large follicle growth and that the ovulatory follicle is selected at about 3 days before ovulation. Other reports (Sirois and Fortune, 1988; Savio et al., 1988; 1990) described three waves of follicular development during bovine estrous cycles. Each wave is characterized by the emergence of several growing follicles (> 5 mm) from a pool of small follicles. One of these emerges as the selected dominant follicle and continues to develop while the others become atretic. The dominant follicle reaches a maximum size of 10 to 15 mm in diameter and remains dominant (i.e., inhibits recruitment of new follicles) for 7 to 10 days, until it becomes atretic, to be replaced by a newly selected dominant follicle that emerged from the recruitment phase of the next follicular wave. If luteal regression occurs during the growth phase or early period of dominance, the dominant follicle can develop into the healthy and estrogenic follicle (up to 20 mm) and ovulate. However a large unhealthy and poorly estrogenic follicle at the time of PGF2a-induced CL regression does not ovulate but regresses and new follicles develop into ovulatory follicles within 48 to 72 h (Ireland and Roche, 1982).







30

The luteal phase lasts until about 17 to 19 days postestrus. High progesterone levels are associated with low secretion of pituitary gonadotropins. The decrease in progesterone associated with luteolysis results in qualitative and quantitative increases in LH and FSH secretion which stimulate final follicular growth and maturation. Decreasing concentrations of progesterone result in increased frequency and amplitude of LH pulses and a gradual decrease of FSH associated with final maturation of the follicle (Foxcroft, 1978; Flowers et al., 1991). There is a rapid rise in estradiol originating from the preovulatory follicle that reaches peak concentration at the onset of estrus (Wetteman et al., 1972). Increasing estradiol causes increased LH receptor number and thus increased responsiveness to LH (Hansel and Convey, 1983). Loss of progesterone negative feedback effects coupled with the stimulus of high estradiol concentrations trigger the release of the preovulatory surge of LH resulting in luteinization and ovulation of the granulosa and theca cells.

Cyclic ovarian follicular development is a complex process that involves proliferation and differentiation of various cellular components of the follicle (see Greenwald and Terranova, 1988 for review). This has been viewed classically as being regulated by gonadotropins and steroid hormones. However additional molecules of ovarian origin are also regulators of these processes. Among these are polypeptides such as inhibin, activin, follistatin and growth factors, including insulin-like growth factors (IGFs), epidermal growth factor and fibroblast growth factor. For comprehensive reviews of the role of these ovarian regulators, the reader is referred to Adashi (1989), Adashi et al. (1989), Geisthovel et al. (1990), Adashi et al. (1991) and Adashi et al. (1992).







31


Following luteinization of the theca and granulosa cells and ovulation, a corpus hemorrhagicum is formed. Luteinization begins approximately 6 h after onset of estrus directed by the effects of LH (Hansel, 1966). Alila and Hansel (1984) determined the origin of large and small cells in the bovine CL using monoclonal antibodies to membrane preparations of theca and granulosa cells collected from ovaries after the LH surge and before ovulation. Large luteal cells were shown to be derived from granulosa cells and small luteal cells from the theca interna cells. Although there is some disagreement as to the absolute numbers of large and small cells at various stages of the estrous cycle, the large cells account for less than 10% of the total luteal cell population in cows (Hansel et al., 1987) and sheep (Rodgers et al., 1984). It appears that large and small luteal cells of the ovine (Harrison et al., 1987) and porcine (Lemon and Mauleon, 1982) CL synergize to promote production of progesterone. Large cells produce greater amounts of progesterone under basal conditions than small cells. However, small ovine and bovine luteal cells are very sensitive to LH, while large luteal cells are relatively insensitive to LH (Alila et al., 1988b; Farin et al., 1989). In vitro, treatment with LH or cAMP analogs can double progesterone production from bovine large cells (Alila et al., 1988b). Purified small luteal cells produce very little progesterone unless stimulated by LH (Alila et al., 1988b, Farin et al., 1989) or other activators of cAMP/PKA transduction system (Hoyer and Niswender, 1985). Small and large cells differ in a number of morphological features such as nuclei and mitochondria positions and presence of oxytocincontaining secretory granules (Fields et al., 1985; Hansel et al., 1987). The characteristics of the small and large bovine luteal cell types have been described extensively by Hansel et al. (1987) and Niswender and Nett (1988).







32

The signal transduction systems involved in regulation of steroidogenesis in luteal cells has generated a great deal of interest. It has become apparent that cAMP/PKA second messenger pathway and the phosphatidylinositol-Ca2PKC effector system are important in regulating luteal cell function. Bovine small luteal cells in vitro contain low resting intracellular [Ca2+] and are able to produce basal levels of progesterone in absence of Ca2. in the culture medium. However, in small luteal cells, Ca2. ions are necessary for LH stimulation of progesterone synthesis and increased cAMP levels are associated with LH action (Alila et al., 1988a). In contrast, bovine large luteal cells contain high levels of resting intracellular [Ca2+] and require Ca2. for basal progesterone production (Alila et al., 1989). In small bovine cells, it appears that the intracellular [Ca2+] increase associated with LH stimulation of progesterone is caused by Ca2. mobilization from intra- and extra-cellular sources; while in large luteal cells, it is primarily due to an extracellular Ca2. influx. In contrast, Wiltbank et al. (1989a) could not find any evidence for LH-induced increase in free [Ca2+] within large or small ovine luteal cells.

PKC also seems to be an essential component of luteal cell function. PKC has been detected in small and large luteal cells from cattle (Davis and Clark, 1983), sheep (Wiltbank et al., 1989b) and pigs (Wheeler and Veldhuis, 1987). Contrasting results have been reported from studies in which cultured luteal cells were treated with phorbol esters to activate PKC. In ovine large luteal cells, phorbol-1 2-myristate-1 3-acetate (PMA) was a potent inhibitor of progesterone secretion (Wiltbank et al., 1989a), but had no effect on large bovine luteal cells (Alila et al., 1988b). In non-stimulated small ovine luteal cells, activation of PKC with PMA had no effect; however when small cells were stimulated with LH, PMA was inhibitory to progesterone production. Wiltbank et







33

al. (1991) and Hansel et al. (1991) have reviewed extensively the second messenger systems involved in the regulation of luteal cell function.
The ability of the CL to produce progesterone is also dependent upon substrate availability (cholesterol) and activity of steroidogenic enzymes. Cholesterol is delivered to ovarian tissues by receptor-mediated uptake of high and low density lipoproteins (see Grummer and Carroll, 1988). Wiltbank et al. (1989b) demonstrated that PMA decreased cholesterol side-chain cleavage activity in large but not in small ovine luteal cells. However, progesterone production was not affected. The 30-hydroxysteroid dehydrogenase / A5-A4 isomerase (3p-HSD) is required for conversion of pregnenolone to progesterone. The activity of 30-HSD in the ovine CL is not a limiting factor (Caffrey et al., 1979) and is not affected by activation of the PKC second messenger system (Hawkins et al., 1993). There is a good correlation between levels of 3P-HSD mRNA and 3p-HSD enzyme activity in the bovine CL (Couet et al., 1990). Hawkins et al. (1993) showed that small ovine luteal cells contained less 30HSD mRNA than large cells and that message for 33-HSD was lowest at day 15 postestrus, a time when luteolysis is initiated.


Corpus Luteum and Prostaglandins


Loeb (1923) made the first observation of the involvement of the uterus with luteal regression in guinea pigs. He noted that "extirpation" of the uterus performed a few days following ovulation, prevented luteolysis and extended CL life span from the normal 14-15 days to 60-80 days. He also noted that maintenance of the CL prevented ovulation and that hysterectomy had no effect on ovulation if performed prior to ovulation. Loeb (1927) suggested three







34

mechanisms to explain this association between the uterus and luteolysis: there could be nervous connections between the CL and uterus; the uterus could compete with the ovary for a limited blood flow and thus removal of the uterus would increase blood flow to the ovary and sustain the CL for a longer period of time; the uterus produces "an internal secretion which exerts a specific, abbreviating effect on the life of the corpus luteum". He rejected the first two possibilities believing the third to be the most likely explanation.
Wiltbank and Casida (1956) demonstrated that hysterectomy of the ewe and cow also resulted in extended CL life span, confirming the concept of uterine control of CL maintenance proposed by Loeb (1923). Studies in which the ovary (Goding et al., 1967) or both the ovary and uterus (Harrison et al., 1968; McCraken et al., 1970) were transplanted to the neck of the ewe demonstrated that uterine control of CL function was not systemic but localized. In cows and sheep, Kiracofe et al. (1966; 1973) demonstrated that uterine luteolytic effect was mediated through local ipsilateral vasculature and transfer of luteolysin involved uterine veins. Passage of the luteolysin from the uterus to the ovary was determined to occur via a counter current transfer from the uterine vein to the ovarian artery and rendered possible by the convolutions of the ovarian artery on the surface of the uterine venous drainage. Surgical anastomoses of uterine veins or ovarian arteries in hemi-hysterectomized cattle and sheep confirmed the existence of the veino-arterial pathway (Ginther, 1974; 1981, Mapletoft and Ginther, 1975).
Pharris and Whyngarden (1969) first investigated the luteolytic role of PGF2. through vasoconstriction of the utero-ovarian vein. Numerous reports have since confirmed PGF2a as a luteolysin in cows, sheep, pigs, horses, rats, rabbits and other species (see Horton and Poyser, 1976); but the







35

vasoconstriction theory has not been supported since direct infusions of PGF2a into the ovary of sheep induced luteolysis without any corresponding decrease in ovarian blood flow (McCracken et al, 1970) and PGF2a inhibited CL progesterone production in vitro (Demers et al., 1973). In other studies, addition of PGF2a to total dispersed bovine luteal cells increased progesterone synthesis (Hixon and Hansel, 1979), and more recently, this stimulation was shown to be limited to small bovine luteal cell. (Alila et al., 1988b). In the cow, intrauterine infusions of PGF2a resulted in higher concentrations of PGF2a in the ovarian artery than the carotid artery (Hixon and Hansel, 1974). Likewise, (3H)-PGF2a infused into the uterine vein of ewes appeared in the ovarian arterial plasma after 30 minutes, whereas the concentration of PGF2a remained very low in the iliac artery (McCracken, 1971). Wolfenson et al. (1985), using frequent blood sampling, observed that the decrease in progesterone at luteolysis (days 18, 19 and 20 postestrus) was associated with increased levels of PGF2a in the ovarian vein but not in the carotid artery.
The lymphatic system also may play an important role in the transfer of PGF2a to the ovary. Abdel Rahim et al. (1984) reported that CL extension occurred in animals with intact utero-ovarian vasculature system but all other utero-ovarian connections cut off. In addition, the concentration of PGF2. in uterine lymph of nonpregnant ewes increased from day 12 postestrus (Abdel Rahim et al., 1983). When (3H)-PGF2a was infused into the uterine lumen of nonpregnant ewes, concentrations of labeled PGF2a in uterine lymph increased and remained high for a longer period of time than in the uterine venous plasma. The transfer rate of (3H)-PGF2a from uterine lymphatics to ovarian artery was actually higher than the transfer rate from uterine vein (0.4% versus 0.3%). A similar transfer cannot be excluded in the cow.







36

The presence of PGF2a receptors on luteal cells is a requirement for PGF2 a luteolytic action. Binding sites for PGF2a have been localized on luteal cell membranes in the cow (Kimball and Lauderdale, 1975), sheep (Powell et al., 1974), pig (Gadsbey et al., 1990), horse (Kimball and Wynegarden, 1977), human (Rao et al., 1977) and rat (Wright et al., 1979). PGF2a receptor has been purified partially (Hammarstr6m et al., 1975; Samuelsson et al., 1978), and the estimated molecular weight is 107 kDa (Samuelsson et al., 1978). High and low affinity binding sites on luteal membranes have been reported, and it appears that ligand binding to the high affinity sites is Ca2+-dependent (Rao, 1975). More recently, Balapure et al. (1989) provided evidence that a high affinity PGF2a receptor is present on the ovine large luteal cells and a low affinity receptor on both large and small luteal cells.
In early studies, addition of PGF2a to total dispersed bovine luteal cells resulted in increased progesterone synthesis (Hixon and Hansel, 1979). This apparent discrepancy between in vivo and in vitro effects of PGF2a has raised the issue of the mechanism of action of PGF2a at the level of luteal cells. Several intracellular biochemical events following binding of PGF2a to its receptor have been described. Alila et al. (1988b) showed that PGF2a increased progesterone synthesis only in small bovine luteal cells. PGF2a had no effect on basal production of large cells, but inhibited LH-stimulated progesterone secretion; a result which suggested that PGF2a, in vivo, acts on these cells. A similar effect of PGF2a was reported for ovine large luteal cells (Schwall et al., 1986). The PGF2a-induced increase in progesterone by bovine small luteal cells appears to be mediated through PLC hydrolysis of PIP2, generation of DAG and activation of PKC rather than by increase in cAMP (Davis et al., 1988). In sheep, PGF2a activates PKC and inhibits progesterone production (Wiltbank et







37

al., 1991). The absence of effect of PGF2. on ovine small luteal cells (no progesterone inhibition, no activation of PKC or increase in intracellular calcium) is consistent with the absence of high affinity PGF2a receptors on this cell type. PGF2a, which unlike LH has no effect on cAMP generation, increases intracellular [Ca2+] in both large and small bovine luteal cells (Alila et al., 1989;1990). PGF2. increases intracellular [Ca2+] to levels sufficiently high to be cytotoxic and to cause transport and exocytosis of oxytocin granules (Chegini and Rao, 1987; Wiltbank et al., 1989b). One possibility to explain the lack of success of in vitro models in reproducing in vivo acute responses of luteal tissue to PGF2a might be the absence of cell-cell contacts. Miyamoto et al. (1993), using a microdialysis system which allows intraluteal application of treatments, observed a plateau in progesterone secretion when PGF2a was applied.
Other biochemical events have been described in response to PGF2a binding to luteal cell membranes: decrease in cell membrane permeability (Riley et al., 1989), increase in synthesis of inositol phosphates (Duncan and Davis, 1991) and increase in production of superoxide radicals (Sawada and Carlson, 1991). Benyo et al. (1991) showed that, in vivo, expression of bovine class 11 major histocompatibility complex (MHC) antigens increased at day 18 postestrus in cyclic cows compared to day 18 pregnant cows. He hypothesized that PGF2a may induce luteolysis by altering MHC antigens expression to stimulate an immune response in luteal tissues.


Oxytocin and Prostaglandins


Evidence for the involvement of oxytocin in the process of luteal regression includes the fact that immunization against oxytocin delays luteolysis







38

(Sheldrick et al., 1980), as does the administration of exogenous oxytocin (Lafrance et Goff, 1985) or oxytocin receptor agonist (Jenkin et al., 1991). In cattle and sheep, oxytocin can be detected in preovulatory follicles (Wathes et al., 1984; 1986; Schams et al., 1985) but reaches a maximum in mid-luteal phase CL and then gradually declines (Wathes et al., 1984; Parkinson et al., 1992b) to reach very low levels by the preovulatory period. In early pregnancy the pattern of oxytocin release by the CL is similar (Parkinson et al., 1992b). In sheep and cattle, mRNA for oxytocin was at a maximum at days 1-3 of the estrous cycle (Jones and Flint, 1988; Ivell et al., 1990) and mRNA levels declined at a steady rate until mid-cycle. Control of oxytocin gene transcription, translation and processing in the ovary has been reviewed in detail by Wathes and Denning-Kendall (1992). Oxytocin and progesterone are secreted in a pulsatile pattern with 97-100% of all oxytocin pulses associated with pulses of progesterone in the cow; however, only 29% and 86% of all progesterone pulses were associated with oxytocin pulses during the early and mid-luteal phases, respectively (Walters et al., 1984). This suggests that ovarian oxytocin and progesterone do not share a common release mechanism.

It is well established that, in vivo, the CL will release oxytocin in response to exogenous PGF2a (Flint and Sheldrick, 1982; Watkins and Moore, 1987) and small elevations of PGF2a in ovarian arterial blood elicit such a response (Lamsa et al., 1989). In vivo, very low infusion rates of PGF2a can selectively release oxytocin, presumably by interacting with the high affinity receptor on the large luteal cell (Lamsa et al., 1989); while much higher infusion rates of PGF2a (1000 pg/min into the ovarian artery) are required to cause the decline of progesterone secretion in vivo (Schramm et al., 1983). In vivo, desensitization and recovery of high affinity PGF2. receptors controlling luteal oxytocin release may contribute to







39

the pusatile pattern of PGF2a release (Lamsa et al., 1992). These effects of PGF2. on the release of oxytocin has been difficult to replicate in vitro using luteal cell cultures. Neither McArdle and Holtorf (1989) using bovine luteal cells, Hirst et al. (1988) using ovine luteal slices, nor Wathes et al. (1988) using perifused bovine luteal explants could demonstrate an acute effect of PGF2a treatment on oxytocin release. Recently, a different culture system has proved more successful; Miyamoto et al. (1993) obtained acute PG stimulation of oxytocin release in bovine CL.


Regulation of Uterine Prostaglandin Production


The action of oxytocin to increase PGF2a secretion is mediated by the uterus. As demonstrated in unilaterally hysterectomized heifers, the uterine horn ipsilateral to the CL is necessary to achieve luteolysis (Ginther et al., 1967). Movement of oxytocin from the ovary to uterus is by countercurrent exchange between the ovarian vein and the uterine branch of the ovarian artery. Transfer efficiency, using (1251)-oxytocin, was estimated to be about 1% (Schramm et al., 1986). Roberts et al. (1976) demonstrated the presence of oxytocin receptors in the uterine endometrium with highest concentrations of receptors present late in the estrous cycle. Heap et al. (1989) reported that in vivo infusion of PGF2a into ovine uterine lymphatics resulted in oxytocin release from both ovaries, although it was considered unlikely that the infused PGF2a could reach the contralateral ovary at a sufficient concentration to produce a response. It also was demonstrated that noradrenaline and acetylcholine can trigger luteal oxytocin release when infused in the ovarian artery (Heap et al., 1989.). However, pretreatment with a- or P-adrenergic blockers could not inhibit PGF2a-induced







40

release of oxytocin, indicating that the PGF2a effect was unlikely due to an adrenergic mechanism (Kotwica et al., 1991).
The release of oxytocin from the CL in response to prostaglandin and the stimulation of uterine prostaglandin secretion by oxytocin suggest that a positive feedback system may function between the ovary and the uterus to cause luteolysis. McCracken (1984) proposed a model for regulation of uterine responsiveness to oxytocin and the release of PGF2a pulses in which steroid receptor dynamics plays a key role. According to the model, a declining endometrial progesterone receptor population permits estradiol from developing follicles to induce oxytocin receptors on the uterine endometrium, making the uterus increasingly responsive to estradiol and oxytocin (Hixon and Flint, 1987). Subsequently, oxytocin released from the pituitary and/or CL binds to newly synthesized endometrial receptors and causes synthesis and pulsatile release of PGF2a. As a consequence, a series of luteolytic pulses of PGF2. initiates a rapid decline in plasma progesterone and causes luteal regression (Hixon and Flint, 1987).

Exogenous oxytocin increases PGF2a concentrations in uterine veins in the ewe (Milvae and Hansel, 1980) and peripheral PGFM concentrations of sheep (Fairclough et al., 1984) and cow (Lafrance and Goff, 1985). Likewise, exogenous PGF2 causes release of oxytocin from the CL in the ewe (Flint and Scheldrick, 1982; Watkins and Moore, 1987) and cow (Schallenberger et al., 1984) in vivo and in vitro (Abdelgadir et al., 1987). In the ewe, 66% of detected PGF2a surges during the estrous cycle were accompanied by pulses of oxytocin from the ovary (Flint and Scheldrick, 1983). Thus, the release of oxytocin from the CL in response to prostaglandin and the stimulation of prostaglandin secretion from the uterus by oxytocin indicate that a positive feedback system







41

may function between the ovary and the uterus to cause regression of the CL (Schallenberger et al., 1984). Oxytocin could amplify the luteolytic signal and ensure rapid completion of luteal regression (Flint and Scheldrick, 1983).
The pulsatile pattern of PGF2a secretion in association with luteolysis is common to all large domestic species (Thatcher et al., 1986a). However, the factors that initiate and terminate PGF2a secretion are not known. Endometrium continuously exposed to oxytocin becomes refractory to further oxytocin challenges. McCracken et al. (1984) proposed that this refractory period was due to down regulation of oxytocin receptors and the pulsatile pattern of PGF2a release was a reflection of alternating periods of down-regulation and resynthesis of oxytocin receptors in the endometrium. Poyser (1991) suggested that uterine refractoriness to oxytocin may be related to the ability of PLA2 to stimulate AA release from phospholipids. Superfusion of the guinea pig uterus from day 7 of the estrous cycle with AA increased PGF2a and PGE2 production when repeated at 0, 1, 3 or 5 h intervals. In contrast, superfusion with PLA2 at the same intervals failed to stimulate further production. Response to PLA2 was restored after 5 h following initial stimulation. Poyser (1991) suggested that at least 5 h are required to replenish releasable pools of AA after an initial stimulation with PLA2 and that this may explain, in part, the pulsatile nature of uterine release of PGF2a. It appears that the activation of the positive feed-back between uterus and ovary is initiated at the level of the uterus since, during spontaneous PGF2. surges, concentrations of PGF2a rise in the utero-ovarian veins before oxytocin (Moore et al., 1986).
In ovariectomized ewes, progesterone can induce a frequency of PGF2a pulses similar to the one observed in intact ewes (Silvia and Raw, 1993) suggesting that ovarian products other than progesterone (such as oxytocin)







42

were not necessary to initiate pulsatile secretion of PGF2, but could be required to achieve full pulse magnitude. Administration of estradiol during the midluteal phase of the estrous cycle stimulates uterine release of PGF2a, increases peripheral concentrations of PGFM and causes CL regression (Thatcher et al., 1986b; Lafrance and Goff, 1985; Hixon and Flint, 1987). Furthermore, Hughes et al. (1987) demonstrated that destruction of ovarian follicles decreased concentrations of E2 in the utero-ovarian vein and delayed luteolysis. Destruction of follicles also prevented full regression of CL in response to an injection of PGF2a. However, the role of estradiol in regulating secretion of PGF2a is not clear. Estradiol appears to have a stimulatory effect on the frequency of neurohypophyseal oxytocin secretion (McCracken et al., 1991). In intact ewes, estradiol treatment increased the frequency of PGFM pulses (Zhang et al., 1991). Yet, destruction of ovarian follicles by X-irradiation had no effect on the frequency or magnitude of PGFM pulses. Silvia and Raw (1993) could not detect any beneficial effect of estradiol on PGF2a pulse magnitude in ovariectomized ewes.
The first 10-12 days of progesterone dominance are characterized by an inhibition of endometrial synthesis of estradiol and oxytocin receptors. During the estrous cycle in the cow, endometrial estrogen and progesterone receptors are at their highest levels during the first 10-12 days postestrus and decline to their lowest levels on about day 13 (Meyer et al., 1986). Progesterone downregulates its own receptor after about day 12 of the estrous cycle, and the decrease in endometrial progesterone receptor is followed by an increase in estradiol receptors (between days 14 and 21) and in oxytocin receptors (between days 17 and 21). McCracken et al. (1984) detected an increase in nuclear estradiol receptors in endometrium by 6 h after stopping a 5 day







43
progesterone infusion in ovariectomized ewes continuously infused with estradiol. Accompanying the rise in estradiol receptors was a parallel increase in membrane bound oxytocin receptors. The concentration of oxytocin receptors rise approximately 500-fold in the uterine endometrium at the time of luteolysis (Roberts et al., 1976; Scheldrick and Flint, 1985). Premature induction of the oxytocin receptor by administration of estradiol results in premature luteolysis (Hixon and Flint, 1987). However, estradiol treatment in ovariectomized cows did not increase uterine responsiveness to oxytocin (Lafrance and Goff, 1988).
Recent results suggest that this increase in oxytocin receptors following the progesterone block may not be an estradiol-mediated event. Vallet et al. (1990) and Lamming et al. (1991) reported that concentrations of uterine oxytocin receptors increased in ovariectomized ewes after 10 days of progesterone treatment. In absence of ovarian hormone replacement, oxytocin receptors are high in ovariectomized ewes probably due to the absence of the inhibitory effect of progesterone. However in such condition, oxytocin is unable to stimulate uterine prostaglandin secretion (Vallet et al., 1990) suggesting that the oxytocin signal transduction system is not functional. When these ewes are treated with progesterone and estradiol, the frequency of PGF2a pulses is enhanced. Application of these findings to the natural cycle suggest that the duration of the estrous cycle is the result of the period of exposure during which progesterone exerts an inhibitory effect on uterine oxytocin receptor expression. Acute treatments with estradiol in intact ewes, which induced premature luteal regression, also induced an increase in endometrial concentrations of oxytocin receptors (Hixon and Flint, 1987). However, estradiol has little effect on oxytocin receptor concentrations in uterine tissues from ovariectomized ewes previously treated with progesterone (Vallet et al., 1990). The role of estradiol is not well







44

established but it may enhance oxytocin receptor synthesis and/or oxytocin transduction pathway to increase the frequency of PGF2a pulses from sheep endometrium during luteolysis (Flint et al., 1989). Sheldrick and Flick-Smith (1993) showed that estradiol enhances oxytocin-stimulated but not basal PGF2a secretion from ovine endometrial explants without affecting oxytocin receptor binding activity. The reader is referred to Silvia et al. (1991) for a review of hormonal regulation of PGF2a secretion during luteolysis in ruminants.

Modulation of uterine prostaglandin production by steroids may also involve regulation of PGH synthase, PLA2 and/or free intracellular AA content. The amount of PGHS activity in ovine uterine tissue changes during the estrous cycle. Maximum activity was achieved late in the estrous cycle when high levels of PGF2 are secreted to induce luteal regression (Huslig et al., 1979). This was due to an increase in PGHS concentration in the tissues and not to an increase in the specific activity of the enzyme. Using an immunohistochemical procedure, Salamonsen and Findlay (1990) have described changes in the concentration of PGHS in uterine tissue throughout the ovine estrous cycle. They found that concentrations were low on days 4 and 17 postestrus and high on days 10, 14, 15 and 16 postestrus; suggesting that levels of PGHS could be controlled by progesterone. Changes in the cellular localization of immunoreactive PGHS were observed: on day 4 of the cycle, PGHS was located primarily in the stromal cells, whereas by day 14 to 16 the enzyme was primarily located in the luminal and glandular epithelial cells (Salamonsen and Findlay, 1990). Administration of progesterone increased amounts of immunoreactive PGHS in the uterine epithelial cell layer of ovariectomized sheep (Raw et al., 1988). Using in situ hybridization, Eggleston et al. (1990) reported that progesterone can induce an increase in the concentration of PGHS mRNA in the endometrium of intact ewes.







45

Data presented by Salamonsen et al. (1991) suggest that expression of mRNA for PGHS in ovine endometrium appears to be constitutive, estradiol having an overall suppressive role which is overcome in the presence of progesterone. PGHS gene expression can be stimulated by factors other than steroids which include: epidermal growth factor (EGF) in amnion cells (Casey et al., 1988) and osteoblasts (Yokota et al., 1986); PDGF in mouse 3T3 cells (Lin et al., 1989); interleukin-1 (IL-1) in rabbit chondrocytes (Lyons-Giordano et al., 1993); and tumor necrosis factor (TNF) in human endothelial cells (Jones et al., 1993). The reader is referred to DeWitt (1991) and Smith et al. (1991) for extensive reviews on factors affecting PGHS gene expression.
Phospholipase enzymes, notably PLA2, can be regulated by steroid hormones. Estradiol has been shown to stimulate PLA2 activity in the endometrium of ovariectomized rats in vivo (Dey et al., 1982). In this study, administration of progesterone alone inhibited PLA2 activity, but treatment with progesterone followed by estradiol caused a stimulation of PLA2 activity above that of estradiol alone. A relationship between PLA2 activity and ovarian steroids has also been demonstrated by Downing and Poyser (1983) who showed that, in the guinea-pig, activity of endometrial PLA2 was higher on day 16 of the estrous cycle than on day 7. The increase followed the rise in estradiol concentrations which occurs on day 10. Bonney (1985) demonstrated the presence of a calcium-dependent PLA2 in human endometrium and showed that there are critical changes in activity which are dependent on stage of the menstrual cycle. The pattern of PLA2 activity described in this study suggested that estradiol stimulates and progesterone inhibits endometrial PLA2 activity. Bonney et al. (1987) demonstrated the presence of two types of PLA2 enzymes in human endometrium, one which is active maximally in presence of Ca2. and another







46


which is inhibited in presence of Ca2 . Interestingly, the two enzymes appeared to be located at different sites within the endometrium. Stromal tissue was shown to contain predominantly the Ca2+-dependent form of PLA2 while the Ca2+-independent form of PLA2 was present mainly in glandular epithelium (Bonney et al., 1987). Pretreatment of human endometrial explants with progesterone followed by incubation with estradiol caused a two-fold stimulation of the Ca2+-dependent PLA2 activity but not of the Ca2+-independent activity. Several studies, using protein synthesis inhibitors in the guinea-pig, indicated that stimulation of endometrial PGF2. synthesis by estradiol acting on a progesterone-primed uterus is dependent upon increased endometrial protein synthesis (Poyser and Riley, 1987; Riley and Poyser, 1989; Leckie and Poyser, 1993). Similarly, ongoing protein synthesis appeared to be essential for both basal and oxytocin-stimulated PGF2a secretion from bovine endometrial explants (Lafrance and Goff, 1990). It was hypothesized that estradiol, acting on a progesterone-primed uterus, may 'switch on' PGF2a production by stimulating the synthesis of a protein that activates PLA2. A PLA2-stimulating protein, which is antigenically and functionally related to the PLA2-stimulating melittin (a bee venom peptide), has been isolated from various cell lines and has been described as an 'intracellular messenger' in the stimulation of prostaglandin synthesis (Clark et al., 1987; Clark et al., 1988). Johnson and Poyser (1991) showed that melittin stimulates PGF2a synthesis by the guinea-pig uterus. Recently, Leckie and Poyser (1993) were unsuccessful in isolating endometrial proteins from guinea-pig at day 15 of the cycle that would stimulate PLA2 activity and prostaglandin synthesis. The Ca2+ ionophore A23187 has been shown to increase the outputs of PGF2a and PGE2 from superfused endometrium (Poyser and Brydon, 1983; Poyser, 1987). The guinea-pig uterus showed complete







47

refractoriness to PLA2, partial refractoriness to A23187, but no refractoriness to exogenous AA with regards to the stimulation of prostaglandin when the same treatment is re-applied after an interval of 1 h (Poyser, 1991). This refractoriness lasted for 3-5 h, and it was suggested that there may be one or several pools of bound AA which were readily releasable but which take 3-5 h to be replenished. Poyser and Ferguson (1993) hypothesized that if PLA2 and A23187 were producing refractoriness by the same mechanism (slow refilling of bound AA pool), then cross refractoriness between the two compounds should occur. No cross refractoriness was observed indicating that the intracellular processes involved in the stimulation of uterine prostaglandin secretion by PLA2 and A23187 were different (Poyser and Ferguson, 1993).
Progesterone may also regulate prostaglandin production by enhancing cellular content of AA. CL development is paralleled by increased endometrial content of lipid droplets which attain maximal concentrations on days 14 to 15 of the cycle (Boshier et al., 1987). In ovariectomized ewes, long term exposure to progesterone increased endometrial content in lipids (Brinsfield and Hawk, 1973). Addition of estradiol to progesterone-treated ewes decreased endometrial lipid content compared to progesterone-treated controls, suggesting that estradiol inhibited progesterone-dependent lipid accumulation or mobilized lipids faster than their accumulation could be stimulated by progesterone (Brinsfield and Hawk, 1973).

McCracken et al. (1981) originally postulated that, in the ewe, cAMP was responsible for the oxytocin stimulation of AA release. However, recent studies suggest that oxytocin-stimulated production of prostaglandin involves the phosphoinositide signal transduction system in ovine (Flint et al., 1986; Hixon and Flint, 1987; Mirando et al., 1990a) and bovine (Mirando et al., 1990b;







48

Mirando et al., 1993a) endometrium. Binding of oxytocin to its endometrial membrane receptor stimulates PLC that can cleave PIP2 into DAG and 'P3 (Flint et al., 1986; Silvia and Homanics, 1988). Agonists of PLC-generated second messengers have been used to replicate the effect of oxytocin on prostaglandin production. PMA (DAG agonist) and A23187 stimulated production of PGF2. from ovine endometrial tissue in vitro (Silvia et Homanics, 1988; Raw and Silvia, 1991). In bovine endometrial explants from heifers at day 19 or 20 of the estrous cycle, PMA induced PGF2a release to the same extent as oxytocin while PDD, a phorbol ester with little PKC-stimulating activity, did not stimulate prostaglandin release (Lafrance and Goff, 1990). The effect of PMA would appear to be specific for PKC activation rather than a non specific effect such as alteration of the plasma membrane. Neither basal nor oxytocin-stimulated PGF2 . release were affected by A23187, dibutyryl cAMP, dibutyryl cGMP or a PKA inhibitor (Lafrance and Goff, 1990). However activation of PKC does not stimulate prostaglandin release from guinea pig endometrium (Poyser, 1987). This is interesting considering that oxytocin does not appear to be involved in PGF2 release during the estrous cycle of the guinea pig (Poyser and Bryon, 1983). Mirando et al. (1990a) showed that the ability of oxytocin to stimulate PLC activity is much greater at day 16 than day 12 postestrus. Such changes in the ability of oxytocin to stimulate PLC, or in the ability of the second messengers to activate PGF2. synthesis, could account for the changes in uterine responsiveness to oxytocin observed during the estrous cycle (Silvia and Raw, 1993). Formation of DAG, following PLC activation, was suggested to serve as a substrate to mono- and diacylglycerol lipases which could cleave AA from DAG. It was suggested that liberation of AA from DAG resulted in increased production of PGF2 (Flint et al., 1986). It appears that this source of







49

AA is not essential to PGF2a production since PMA, which activates PKC but cannot be a source of AA, induced PGF2a secretion to the same extent as oxytocin (Lafrance and Goff, 1990).
The conceptus could prevent luteolysis in early pregnancy by production of a secretory product to maintain progesterone action on the endometrium (McCracken et al., 1984). If the progesterone receptors are maintained, nuclear estradiol receptors could be suppressed and thus terminate the chain of events leading to luteolysis. The conceptus could also produce a product that directly inhibits formation of endometrial estradiol or oxytocin receptors.


Maternal Recognition of Pregnancy


In order for pregnancy to be maintained in ruminants, regression of the CL must be prevented since ovarian progesterone is necessary for maintenance of pregnancy. Therefore, the luteolytic influence of the uterus must be neutralized in some manner. The signal for maintenance of the CL must come from the conceptus and could be luteotrophic (increase the rate of production and/or release of progesterone from the CL), antiluteolytic, or some combination. Antiluteolytic effects include those that are anti-PGF2a in nature (alter the dynamics of secretion of PGF2a) and those that are luteoprotective (directly protect the CL) (Rothchild, 1981; Thatcher et al., 1986a). The processes by which the conceptus signals its presence to the maternal unit, resulting in CL maintenance and other events, have been termed maternal recognition of pregnancy (Short, 1969).
Studies in cattle and sheep have determined the critical time by which the conceptus must signal its presence to the maternal uterus. Embryo removal







50

experiments have shown this. The critical period for conceptus-induced maintenance of the CL appeared to occur on day 12 in sheep (Moor and Rowson, 1966) and between days 15 and 17 postestrus in cattle (Northey and French, 1980; Humblot and Dalla Porta, 1984). This led to a search for the embryonic "signals" responsible for the alteration of the uterine luteolytic mechanism resulting in CL maintenance and continuation of progesterone secretion.
Intrauterine infusion of conceptus extracts or homogenates extended luteal lifespan in cattle (Northey and French, 1980; Humblot and Dalla Porta, 1984) and sheep (Ellinwood et al., 1979; Martal et al., 1979). The active component in ovine conceptus homogenates was reported to be proteasesensitive, heat-labile, absent beyond day 21 of gestation (Martal et al., 1979) and to contain neither LH/hCG-like nor prolactin like activity (Ellinwood et al., 1979). Intrauterine administration and reciprocal interspecies (ovine-bovine) transfer of trophoblastic vesicles extended recipients CL lifespan indicating that the active proteinaceous component was secreted by trophoblastic tissues and that it was similar or closely related in these two species (Heyman et al., 1984; Martal et al., 1984). The total array of secreted proteins by day 16-18 bovine or day 15-16 ovine conceptuses was referred to as bovine or ovine conceptus secretory proteins (bCSP or oCSP). Infusion of CSP into the uterine lumen extended CL lifespan and lowered release of PGFM in response to luteolytic doses of estradiol in the cow (Knickerbocker et al., 1986a, 1986b) and the ewe (Godkin et al., 1984b; Fincher et al., 1986). These results suggested that bovine and ovine conceptuses secreted a proteinaceous factor of trophectoderm origin with antiluteolytic-anti-PGF2, activity.







51

The active component found in ovine conceptus homogenates was first named "trophoblastin" by Martal et al. (1979) and later as "ovine trophoblast protein-I" (oTP-1) since this protein was the major secreted protein from conceptus trophectoderm (Godkin et al., 1984a). oTP-1 has a molecular weight of about 18 kDa and isoelectric points (pl) ranging from 5.3 to 5.7 (Godkin et al., 1982; 1984b; Hansen et al., 1985) and is not glycosylated (Anthony et al., 1988). Shortly after, a low molecular weight (22-26 kDa), acidic (pl = 6.6.-5.6) protein complex (Bartol et al., 1985a) secreted by bovine conceptuses (days 16-24) was shown to cross-react with an antibody raised against oTP-1 and was called bovine trophoblast protein-1 (bTP-1) (Helmer et al., 1987). The bTP-1 complex consists of at least 7-9 isomers grouped in three molecular weight classes (21, 23 and 25 kDa) (Bartol et al., 1985a; Helmer et al., 1987; 1989a; Anthony et al., 1988; Plante et al., 1990). This variation in molecular weight and isoelectric points is due to differential N-linked glycosylation (Helmer al., 1988) and to translation of various bTP-1 gene transcripts (Anthony et al., 1988).
Trophoblast protein-1 could be identified in uterine fluids from day 15 to 25 in pregnant cows, and from day 14 to 22 in pregnant ewes (Kazemi et al., 1988). Secretion of oTP-1 increased as conceptus morphology changed from spherical (312 ng/uterine flushing) to tubular (1380 ng) to filamentous (4450 ng) on days 12-13 (Nephew et al., 1991). Presence of bTP-1 mRNA could be detected as early as day 11 by in situ hybridization (Farin et al., 1990; Guillomot et al., 1990), or day 8 by reverse transcription-polymerase chain reaction (RTPCR) (Hernandez-Ledezma et al., 1992). Immunochemical studies have localized bTP-1 in the cytoplasm of mono- and binuclear cells of the trophectoderm (Lifsey et al., 1989). The increase in trophoblast protein-1 production coincides with a dramatic elongation of the conceptus (Geisert et al.,







52
1988; Nephew et al., 1989). High rates of oTP-1 synthesis occur between day 13 and 21 of pregnancy, with production increasing over one hundred times from day 12 to day 15-16 when total synthesis is maximum (Hansen et al., 1985; Farin et al., 1990; Roberts et al., 1992). Ott et al. (1989) described a second period of oTP-1 secretion by chorion between days 25 and 45 of pregnancy. Bovine TP-1 displays a similar pattern of expression with synthesis peaking between days 17 and 19 (Farin et al., 1990). Expression of oTP-1 was absent in the trophoblastic regions showing cellular contacts with the uterine epithelium (Guillomot et al., 1990). Interestingly, c-fos mRNA expression was shown to follow the same expression pattern in trophoblastic cells as oTP-1 (Xavier et al., 1991).
N-terminal sequencing of purified oTP-1 (Stewart et al., 1987; Charpigny et al., 1988) and molecular cloning of the cDNA for oTP-1 (Imakawa et al., 1987; 1989; Stewart et al., 1989b; Charlier et al., 1989; 1991; Klemann et al., 1989; 1990) and later bTP-1 (Imakawa et al., 1989; Stewart et al., 1990; Hansen et al., 1991) revealed that both trophoblast proteins were related to the type I interferon (IFN) family. Three subtypes of type I interferons have been described: IFNa, IFNP and IFNco which differ considerably in amino acid sequence but share a common membrane receptor (see Roberts et al., 1992). The various isoforms encoded by multiple genes within each one of these subtypes can present significant differences in amino acid sequence and biological activity . In particular, the IFNa subtype represents a group of 165166 amino acid proteins and contain 10-12 subtypes, whereas the IFNo (formerly known as IFNt11) subtype represents a group of 172-174 amino acid proteins and contain 15-20 subtypes (Capon et al., 1985; ). Bovine TP-1 and oTP-1 mRNA code for polypeptides of 195 amino acid, including a 23 animo







53

acid signal peptide. The length of mature oTP-1 and bTP-1 (172 amino acids) and their high sequence homology (70%) with bovine IFNO) (Imakawa et al., 1989) suggested that they belonged to the IFNco subtype. However, bTP-1 and oTP-1 resemble each other in both amino acid and nucleotide sequence more than other IFNo do within their own subtype. Their high sequence homology to each other and a number of additional unique features such as poor virus inducibility led Roberts et al. (1992) to recommended the creation a new structural subtype of type I interferons: interferon (IFN) tau ('). This designation was sanctioned by the International Committee on the Interferon Nomenclature (Roberts, 1993). Genes coding for IFNt have been identified in other related ruminants such as goat, and musk ox, but not in horse, pig, mouse, rabbit and human (Leaman et al., 1992). The high level of homology across ruminant species and presence of IFNT within a limited subset of mammals suggests that IFNt derived from IFNo about 30-50 million years ago (Roberts et al., 1992).

IFNr gene expression is ephemeral, intensive and specific to the trophectoderm during the preimplantation period. These characteristics generated a number of studies designed to better understand the unique features of IFNr transcriptional regulation. Very low levels of IFNr mRNA has been detected in bovine leukocytes exposed to Sendai virus (Cross and Roberts, 1991). IFNT expression was not observed when a bovine IFNt (blFNt) gene or promoter constructs coupled to a reporter gene was transfected into non-trophoblastic cells. However IFNt gene was constitutively expressed when transfected into trophoblast tumor derived cells (Cross and Roberts, 1991). A 450 base pair region of the IFNT gene promoter was necessary to obtain maximal constitutive expression in these cells, whereas an IFNO gene promoter was inactive. Ko et al. (1991) demonstrated a good correlation of uterine luminal







54

insulin-like growth factors (IGFs) with otFNtproduction, and observed a beneficial effect of culturing blastocysts in presence of IGFs with respect to maintenance of olFNT secretion. Recently, Imakawa et al. (1993) provided evidence that olFNT mRNA and protein production was stimulated by granulocyte macrophage-colony stimulating factor (GM-CSF), and that GM-CSF mRNA was found in the endometrial epithelium. They proposed that uterine GM-CSF was in part responsible for the large increase in olFNt secretion observed at the time of maternal recognition of pregnancy. The reader is referred to Roberts et al. (1991), Roberts et al. (1992) and Leaman et al. (1992) for reviews on the molecular biology of IFNT.
High affinity (Kd=0.1x10-10 to 0.4x10-10 M) and low affinity (Ke1O-10 M) receptors for ovine IFNT (olFNT) were reported in pregnant and cyclic sheep endometrium (Godkin et al., 1984a; Hansen et al., 1989; Knickerbocker and Niswender, 1989). On days 8 and 12, the number of unoccupied olFNT receptors was similar between cyclic and pregnant ewes, but diminished following day 12 in pregnant ewes (Knickerbocker and Niswender, 1989). The ability of olFNt to displaced radiolabelled human IFNa from its receptors was indicative of olFNt affiliation with type I interferons (Stewart et al., 1987). Cross-linking studies indicated that olFNt was associated intimately with two proteins when bound to its receptor (Hansen et al., 1989). These two proteins (70 and 100 kDa) are not specific to endometrial tissue since olFNt cross-linked similar proteins in spleen membranes (Hansen et al., 1989); however they could be part of a high affinity, multiple subunit complex responsible for mediation of IFNt biological effects.







55


Trophoblast Interferons


Antiluteolytic Activity

In vivo intrauterine infusion of olFNt into cyclic ewes from day 12 to 20 (Godkin et al., 1984b) or from day 11 to 15 (Vallet et al., 1988) extended CL lifespan. Similarly, infusion of blFNt into the uterine lumen of cyclic cows from days 14 to 17 prolonged CL lifespan and inhibited PGF2a release into the posterior vena cava (Helmer et al., 1989b). Because of the expense of purifying large amounts of "natural" IFNt from conceptuses, the availability of recombinant interferon bovine IFNa, (50% amino acid sequence homology with bIFNt and olFNT), and later recombinant ovine and bovine IFNT became a key factor in the study of the antiluteolytic properties of type I interferons. Intrauterine infusion of milligram quantities of recombinant blFNa, into cyclic ewes (Stewart et al., 1989a; Parkinson et al., 1992a), or intrauterine or intramuscular injections of this protein into cyclic cows (Plante et al., 1988, 1989, 1991; Newton et al., 1990) during the time of maternal recognition of pregnancy extended luteal lifespan of the treated animals. Nevertheless, these studies demonstrated that considerably more recombinant blFNa, had to be introduced into the uterus to achieve an extension of interestrous interval than if natural IFNt had been used instead. For example, daily (days 9 to 19) intrauterine administration of 2,000 p1 g of recombinant blFNa, were necessary to extend interestrous intervals of ewes (Parkinson et al., 1992a), compared to a requirement of only 100 pag olFNt injected daily (days 12 to 15) to extend interestrous intervals of ewes to greater than 19 days (Vallet et al., 1988). Subsequently, recombinant olFNT became available (Martal et al., 1990; Ott et al., 1991) and was tested for its antiluteolytic activity. Intrauterine injection of recombinant oIFNr in cyclic ewes between days







56

11 and 15 (Martal et al., 1990), and in cyclic cows between days 14 and 24 (Meyer et al., 1992), extended CL lifespan and interestrous interval .
Fertility studies have been conducted using recombinant blFNa, to improve pregnancy rate in sheep and cattle. The rational for these studies was to supplement inseminated animals with interferon during the period of maternal recognition of pregnancy in order to "rescue" conceptuses lagging in their development and therefore unlikely to produce sufficient amounts of IFNT to prevent luteolysis. Although recombinant olFNT and blFNt (Klemann et al., 1990) were produced and physiologically active, the amounts necessary to conduct fertility trials were not available. Encouraging results were obtained in sheep where conception rates were improved significantly with two daily intramuscular injections (2 mg) of recombinant blFNa, (Nephew et al., 1990; Schalue-Francis et al., 1991; Martinod et al., 1991). Francis et al. (1992) verified that intramuscularly injected recombinant blFNX, could reach the uterus and induce synthesis of a pregnancy-specific protein (p70) by endometrium in cyclic ewes. In contrast, there have been no reports of improved conception rates in cattle using recombinant blFNal. In fact, administration of blFNa, in cows, regardless of the injection pattern, decreased conception rates by about 10% (Barros et al., 1992b). This unexpected effect was attributed to hyperthermia and acute lowering of serum progesterone concentrations induced by recombinant blFNa, (Barros et al., 1992a).


Regulation of Endometrial Prostaqlandin Synthesis
During early pregnancy, pulsatile secretion of PGF2a is suppressed in sheep (Thornburn et al., 1973; McCracken et al., 1984; Zarco et al., 1988) and cattle (Kindahl et al., 1976; Thatcher et al., 1984a; Basu and Kindahl, 1987a), as







57

is the ability of oxytocin to stimulate uterine secretion of PGF2a (Fairclough et al., 1984; Silvia et al., 1992; Lafrance and Goff, 1985). The induction of PGFM pulses and luteolysis in response to estradiol administration is absent in pregnant cows (Rico et al., 1981; Thatcher et al., 1984a) or ewes (Kittok and Britt, 1977; Lacroix and Kann, 1986; Fincher et al., 1986) as compared to cyclic animals at the same day postestrus. In vitro studies characterized endometrial prostaglandin secretion during early pregnancy. Endometrial explants (Thatcher et al., 1984b) or perifused endometrial tissue (Gross et al., 1988b) from cows at day 17 postestrus synthesized less PGF2. than those of cyclic cows. The absence of oxytocin responsiveness observed in vivo and in vitro in pregnant cattle and sheep is directly related to the fact that endometrial oxytocin receptors are decreased considerably in pregnant compared to cyclic ewes (McCracken et al., 1984; Flint et al., 1986) and cows (Fuchs et al., 1990; Jenner et al., 1991) during the luteolytic period.
The antiluteolytic role of IFNt observed in vivo is supported by a number of in vitro studies examining the effects of trophoblast interferon on endometrial secretion of prostaglandins. PGF2a secretion was decreased when endometrial explants from cows at day 17 postestrus were incubated for 24 h in presence of bCSP (Gross et al., 1988b) or bovine IFNT (Helmer et al., 1989b). Vallet et al. (1989) reported lower PGF2a secretion by perifused endometrium when collected from ewes that received intrauterine infusions of ovine IFN-r for 3 days. Culture of highly purified endometrial epithelial and stromal cells have demonstrated marked differences in prostaglandin release by the two cell types. Epithelial cells were the major source of PGF2a, and PGE2 was primarily secreted by stromal cells from sheep (Cherny and Findlay, 1990; Charpigny et al., 1991) and cattle (Fortier et al., 1988). Salamonsen et al. (1988) reported an inhibitory







58

effect of ovine IFNt on release of both PGF2a and PGE2 from mixed ovine endometrial cells in primary culture. Using highly purified cultures of endometrial epithelial and stromal cells from cyclic sheep at day 15 postestrus, Charpigny et al. (1991) observed that ovine IFN' inhibited basal and oxytocininduced secretion of PGF2a and PGE2 in both cell types.


Regulation of Endometrial Steroid and Oxytocin Receptors
The point(s) in the intracellular cascade regulating uterine secretion of PGF2a that is affected by the conceptus has not been identified. Bazer (1992) hypothesized that IFNr antiluteolytic effect could be mediated through several potential mechanisms: (1) stabilization or up-regulation of endometrial progesterone receptors to maintain the progesterone block and prevent synthesis of oxytocin receptors or up-regulation of estrogen receptors; (2) direct inhibition of estrogen receptors which would attenuate oxytocin receptor synthesis and pulsatile release of PGF2a ; (3) direct inhibition of oxytocin receptor synthesis; (4) alteration of post-oxytocin receptor mechanisms to inhibit oxytocin-induced release of PGF2a.

During maternal recognition of pregnancy in sheep, endometrial progesterone receptor (PR) concentrations are maintained at concentrations characteristic of mid-luteal phase, whereas in non-pregnant animals, PR are high during mid-diestrus and decrease during late diestrus (Findlay et al., 1982; Cherny et al., 1991; Ott et al., 1993). Endometrial estrogen receptor (ER) concentrations decreased from day 10 to 16 postestrus in pregnant ewes, whereas in cyclic ewes, ER remained relatively constant from day 10 to 14 and then increased dramatically at day 16 postestrus (Ott et al., 1993). In vivo, progesterone was necessary for olFNt to exert an inhibitory effect on oxytocin-







59

stimulated production of prostaglandin in the ewe (Ott et al., 1992). The antiluteolytic effects of IFNT could be mediated through direct transcriptional regulation of uterine progesterone and/or estrogen receptor genes. Neither ovine nor bovine progesterone and estrogen receptor genes have been cloned and therefore no interferon stimulated response elements (ISRE) have been described in the regulatory regions of these genes. Structures of the human estrogen receptor (Ponglikitmongkol et al., 1988), and human (Kastner et al., 1990) and rabbit (Milgrom et al., 1988) progesterone receptors have been described. Analysis of the 5' flanking regions of these genes revealed only putative but no functional ISREs. Intrauterine administration of oCSP (containing 100 pg of oIFNr/day) from day 11 to 15 postestrus could not totally replicate the PR and ER protein and mRNA concentrations observed in pregnant ewes at day 16 postestrus. Ovine CSP appeared to have a negative effect on PR protein and mRNA concentrations compared to serum albumin treated ewes at day 16 postestrus (Mirando et al., 1993b).
In vivo, oxytocin was unable to stimulate uterine PGF2a secretion when administered to ewes on days 12 to 16 of pregnancy (Fairclough et al., 1984; McCracken et al., 1984; Silvia et al., 1992). This lack of oxytocin responsiveness was paralleled by an almost total suppression of oxytocin's ability to stimulate PLC activity in pregnant ewes (Mirando et al., 1990a; 1990b; Ott et al., 1993). However, oxytocin was equally effective in stimulating release of PGF2a from endometrial tissue collected from pregnant and non pregnant ewes on days 14 and 16 postestrus (Vallet et al., 1989; Silvia and Raw, 1993). The oxytocin-induced PGF2a secretion by endometrial tissue from pregnant ewes appears to be mediated through a pathway other than PLC-induced release of phophoinositides. Silvia et al. (1992) suggested that the disparity between in







60

vivo and in vitro models with regards to oxytocin responsiveness of endometrium from pregnant ewes could be due to a requirement for continuous presence of the conceptus to maintain oxytocin non-responsiveness. Incubation of uterine tissue in the absence of the conceptus for just 2.5 h appeared to be sufficient for the suppressive influence of the conceptus to wear off; implying that a very labile factor could be involved in mediating the conceptus effect on oxytocin responsiveness.
Acute in vitro exposure of endometrial tissue to olFNT (Vallet and Bazer, 1989) had very little effect on the ability of oxytocin to stimulate activity of PLC. Vallet et al. (1989) showed that olFNT did not compete with oxytocin for its receptor. Interestingly, oxytocin effect on prostaglandin secretion by endometrial cells in primary culture was inhibited only if cells were exposed to ovine IFNT before (24 h pretreatment) and during the oxytocin challenge (Charpigny et al., 1991). When cyclic ewes were exposed to olFNT from days 12 to 14 postestrus, PGF2a secretion was reduced (Vallet et al., 1988; Mirando et al., 1990a) as well as phosphoinositide metabolism (Mirando et al., 1990a; 1990b) indicating a reduced ability of oxytocin to stimulate PLC activity. Intrauterine injections of olFNT from day 11 to 15 decreased endometrial number and affinity of oxytocin receptors (Mirando et al., 1993b) possibly by down-regulating endometrial ER and/or stabilizing endometrial PR.


Endometrial Prostaplandin Synthesis Inhibitor
The regulation of endometrial steroid receptors is not the only possible mechanism to explain the absence of oxytocin-responsiveness and pulsatile release of PGF2a in endometrium from pregnant cows and ewes. No differences in detectable PGHS enzyme (Salamonsen and Findlay, 1990) or mRNA







61

(Salamonsen et al., 1991) were found in endometrial tissue from pregnant and cyclic ewes on day 15 after estrus. Consequently, the conceptus does not appear to suppress PGF2a secretion by reducing the concentration of PGHS in this tissue.
Another possibility is that the conceptus induces an intracellular endometrial factor which inhibits enzymes involved in prostaglandin synthesis. Endogenous inhibitors of prostaglandin synthesis have been reported in bovine follicular fluid (Shemesh, 1979), human amniotic fluid (Saeed et al., 1982) and human serum and plasma (Saeed et al., 1977). A heat-labile, inhibitor of prostaglandin synthesis was described in fetal cotyledon tissue (Shemesh et al., 1984b) and maternal caruncle tissue (Shemesh et al., 1984a) of bovine placentomes. Inhibitory activity was present in caruncle extract from days 120150 of gestation, but was not detectable at term (260-280 days) (Shemesh et al., 1984a). The exposure of bovine luteal cells to the active placental extract was sufficient to suppress their prostaglandin synthesis (Shemesh et al., 1984a).
The conversion of (I4C)-AA by endometrial microsomes and cytosol of cyclic and pregnant cows has been examined to investigate their ability to synthesize prostaglandins (Wlodawer et al., 1976; Basu and Kindahl, 1987b). Wlodawer et al. (1976) showed that bovine microsomes had a relatively low capacity to convert AA into PGF2a. However, when PGH2 was substrate for the microsomal enzymes, the conversion into PGF2a was much higher. This observation suggested that PGHS activity was a limiting factor in the ability of endometrial microsomes to synthesize PGF2a, possibly due to the presence of a PGHS inhibitor. In a comparative study, Basu and Kindahl (1987b) described the presence of a potent inhibitory factor that controls prostaglandin biosynthesis in bovine endometrial tissue by acting on PGHS activity. The inhibitory capacity







62

was highest in microsomes and much higher in pregnant than nonpregnant animals. When comparing the potency of this inhibitory factor, as calculated by IC50 values, the potency was about 8 times higher for day 17 pregnant than day 17 cyclic endometrial cytosol (Basu and Kindahl, 1987b). This inhibitory factor was not destroyed when boiled and was present in both uterine horns suggesting that a humoral or local factor from the conceptus could have influenced the non pregnant horn (Basu and Kindahl, 1987b).
Gross et al. (1988b) examined the ability of endometrial cytosol and microsomes, from day 17 cyclic and pregnant cows, to modulate prostaglandin synthesis by cotyledonary microsomes from parturient cows. The endometrial prostaglandin synthesis inhibitor (EPSI) appeared to be present in much higher amounts in pregnant cows than cyclic cows, to be primarily cytosolic, proteinaceous, and to act as a non-competitive inhibitor with regard to AA metabolism (Gross et al., 1988b). Bovine IFNT induced EPSI activity when administered to endometrial explants from cyclic cows at day 17 postestrus (Helmer et al., 1989a).


Antiviral, Anti proliferative and Immunosuppressive Activities

In common with other type I interferons, IFNT has a high antiviral activity of 2 to 3 x 108 IU/mg of protein (Pontzer et al., 1988) and can protect type I receptor-bearing cells from lysis induced by a variety of viruses (see Roberts et al., 1992). Among the enzymes induced by IFNT binding to its receptor are 2',5'oligoadenylate synthetase (2-5A synthetase), protein kinase (p68) and endoribonuclease (Jacobsen et al., 1988). In the presence of viral double stranded RNA and ATP, the protein kinase phosphorylates and inactivates an initiation factor (elF-2) necessary for the synthesis of viral protein (Esteban and







63

Paez, 1985). Ovine and bovine IFN'r, as well as bovine interferon a,1, induced synthesis of 2-5A synthetase in endometrial tissue (Barros et al., 1991; Mirando et al., 1991; Short et al., 1991). Three isoforms of 2-5A synthetase have been reported (see Hovanessian, 1991) which catalyze the synthesis of an activator of a ribonuclease involved in degradation of viral RNAs. It appears that PKC is not essential for the antiviral activity of a interferons (Cernescu et al., 1989). The antiviral activity of IFNt may play an important role in protection of the periimplantation conceptus when exposed to viral infection.
Interferons a are known to inhibit proliferation of many cell types (see Clemens and McNurlan, 1985; Salzberg et al., 1990). The IFNT was a potent inhibitor of (3H)-thymidine incorporation into lymphocytes after exposure to mitogens (Newton et al., 1989; Fillion et al., 1991; Skopets et al., 1992). Ovine IFNT exhibited a potent anticellular activity across species and appeared to decrease the rate of progression through the cell cycle (transition from GO/G1 into and through S phase) (Pontzer et al., 1991). However , considering the enormous amounts secreted by the conceptus at the time of maternal recognition of pregnancy, it is remarkable that IFNT lacked the cytotoxic effects observed with high doses of other types of interferons (Pontzer et al., 1991).


Regulation of Endometrial Protein Synthesis

Trophoblast IFN stimulated the synthesis and secretion of specific proteins by endometrial explants from day 12 cyclic sheep but did not induce changes in the type of proteins secreted (Godkin et al., 1984a). Quantitative studies revealed that olFNt stimulated synthesis of four or five types of proteins, and decreased the synthesis of another two (Vallet et al., 1987; Salamonsen et al., 1988). The pattern of proteins synthesized by ofFN'-treated endometrium







64

was similar to that of endometrium on day 13 of pregnancy and to endometrium from cyclic ewes treated with day 15 oCSP or human IFNc (Salamonsen et al., 1988). One of the major stimulated proteins has a molecular weight of 70 kDa and pl~4. The same protein was identified in sheep endometrium between days 14 and 20 of pregnancy (Sharif et al., 1989), and could be induced in endometrium from cyclic sheep by in vivo administration of recombinant bovine IFNa1l (Francis et al., 1991). In bovine endometrial explants from bCSP-treated cyclic cows or from pregnant cows, proteins in the 14 kDa molecular weight range were upregulated (Geisert et al., 1988; Gross et al., 1988c; Helmer et al., 1989a) compared to endometrium from untreated cyclic cows. Rueda et al. (1993) described two recombinant blFNT-stimulated proteins (12 kDa, pl > 7.5 and 28 kDa, pl 4.5-5.5) in secretion from bovine endometrial explants; these two proteins have not been identified.
The 70 kDa "pregnancy-specific" protein has not been identified, but another lower molecular weight olFNT-induced protein has been isolated and the N-terminal amino acid sequence determined (Vallet et al., 1991). The N-terminal sequence of this protein had a 40-55% sequence homology with the N-terminus of 02-icroglobulin. Stewart et al. (1992) presented evidence that expression of 02-microglobulin and major histocompatibility complex (MHC) class I were correlated with expression of olFNT, suggesting that the trophoblast interferon induced expression of these proteins. A recent report described that, in neuroblastoma cells, IFNa increased MHC class I and 02-microglobulin gene expression by inducing factor binding to ISREs present in both genes (Drew et al., 1993).







65

Signal Transduction System
Using biochemical and electron microscopic techniques, it was shown that, following binding, the IFNa-receptor complex was rapidly internalized via receptor-mediated endocytosis (see Langer and Pestka, 1988; Mogensen et al., 1989; Grossberg et al., 1989). Binding of IFNa to type I receptors is a necessary but not sufficient condition for cell activation. Through analysis of mutants unresponsive to IFNct, it was determined that lack of receptors was only one reason for lack of biological response (Dron et al., 1986).
A rapid and transient activation of PLA2 and release of AA was noted following treatment of 3T3 fibroblasts with IFNa, and inhibition of PLA2 activity specifically blocked the binding of nuclear factors to the 2-5A synthetase gene ISRE (Hannigan and Williams, 1991). A consequence of PLA2 activation is the release of AA from membrane phospholipids. Free AA can be metabolized through three pathways: cyclooxygenase (PGHS), lipoxygenase and epoxygenase. Perturbation of AA metabolism, using inhibitors of cyclooxygenase and lipoxygenase, enhanced the IFNa stimulation of factor binding to the ISRE (Hannigan and Williams, 1991). The potentiation of IFNaC effect observed, when AA metabolism was redirected to epoxy-derivatives, was not restricted to nuclear factor binding but also was evident at the levels of 2-5A synthetase mRNA and protein (Williams, 1991). While AA metabolism seems to play an important role in the signal transduction of IFNa, AA itself is not the second messenger. Addition of exogenous AA did not affect the transcription of IFN-stimulated genes (Yan et al., 1989). AA metabolism has been implicated in signal transduction of other systems such as induction of tumor necrosis factor gene expression by phorbol esters (Horiguchi et al., 1989).







66

Several reports have indicated the importance of protein phosphorylation and possibly a role for PKC in IFNa action (Yap et al., 1986; Tiwari et al., 1988; Reich and Pfeffer, 1990; Decker et al., 1991). Pfeffer et al. (1990) proposed that the transient increase in DAG, resulting from phosphatidylcholine hydrolysis, may induce selectively translocation and activation of the 0 isoform of PKC. Interferon a was shown to induce phosphatidylcholine hydrolysis, but not inositol phospholipid turnover, therefore providing a source of DAG without increasing the intracellular free Ca 2 concentration (Rosoff et al., 1988). However, agonists of PKC activity did not mimic IFNa action and PKC down-regulation did not alter the transcriptional response to IFNa (Reich and Pfeffer, 1990; Pfeffer et al., 1990). These observations are consistent with the potential role of AA metabolites in the signal transduction of IFNa.

Recently, direct evidence for the involvement of a protein kinase distinct from PKC has been presented (Kessler and Levy, 1991). This protein kinase was discovered by functional complementation of mutant cells unresponsive to IFNa. The only message encoded within the complementing cosmid was found to be tyk2, a non-receptor protein tyrosine kinase (Firmbach-Kraft et al., 1990). Since, JAK-1 and JAK-2 have been added to this new family of protein kinases, which is not related to the src family of tyrosine kinases (Hunter, 1991). Velazquez et al. (1992) speculated that tyk protein kinase could be activated by interaction of its large extracatalytic domain with IFNa receptors and, subsequently phosphorylate interferon-stimulated gene factor 3 (ISGF3).

Among ISRE-binding proteins, ISGF3 has been shown to be the primary transcriptional activator for interferon a-induced genes and to recognize a 15nucleotide cis-acting DNA element (ISRE) located in the promoter region of IFNa
-stimulated genes. ISGF3 is the only ISRE-binding factor rapidly induced by







67

interferon a and its activity profile parallels the process of receptor binding and transcription activation (Dale et al., 1989; Bandyopadhyay et al., 1990). It was shown that an ISGF3-like activity was sensitive to inhibition of PLA2 in mouse fibroblasts (Hannigan and Williams, 1991). ISGF3 is a complex shown to be composed of four (113, 91, 84, and 48 kDa) distinct proteins (Fu et al., 1990). The 48 kDa (ISGF3y) itself can bind to the ISRE with low affinity (Kessler et al., 1990). ISGF3cc, composed of the three other polypeptides, is immediately activated and translocated to the nucleus in response to interferon a and forms an active complex with ISGF3y which has a 20-fold higher binding affinity for ISRE than ISGF3y by itself (Bandyopadhyay et al., 1990; Kessler et al., 1990). Fu et al. (1992) presented evidence that ISGF3a complex was activated by a direct tyrosine phosphorylation at the level of SH2 and SH3 domains, which are commonly found in a number of tyrosine kinase-regulated proteins (see Koch et al., 1991). Deletion analysis of ISGF3y revealed that two domains confer the two known activities of this protein, DNA recognition and multimer assembly (Veals et al., 1993). Using a combination of specific tyrosine kinase and phosphatase inhibitors, David et al. (1993) were able to demonstrate that a tyrosine kinase and a membrane-bound tyrosine phosphatase lead to modification of ISGF3a and subsequent formation of the complete ISGF complex.

The reader is referred to Levy and Darnell (1990) and Stark and Kerr (1992) for recent reviews on the early events in interferon transduction system.







68


Implications


In ruminants, the role of uterine PGF2a in the demise of the corpus luteum is well established. However, the nature of the relationships between uterus and peri-attachment conceptus at the time of maternal recognition of pregnancy is still under intensive investigation. Potential mechanisms by which the embryo and bovine IFNT exert an antiluteolytic effect include regulation of endometrial steroid receptors (see Bazer, 1992) and oxytocin responsiveness (Mirando et al., 1993b). Alternative antiluteolytic mechanisms induced by the embryo and IFNt may involve regulation of some components of the endometrial prostaglandin biosynthetic pathway. The current concepts consider arachidonic acid availability and PGH synthase activity as the two main physiological constraints on the uterine capacity to synthesize prostaglandins (Marshall et al., 1987).
AA availability is dependent upon a variety of regulating factors such as PLA2 for the release of AA from membrane phospholipids, acylCoA-synthase and acyl transferases for the control of AA turnover in phospholipids (Norman and Poyser, 1993). Other metabolic pathways (e. g., lipoxygenase, epoxygenase, 0-oxydases) compete with PGHS for unesterified AA. In addition, AA accessibility to these enzymes can be controlled by intracellular AA-binding factors (Spector, 1975).
The cyclooxygenase activity of PGH synthase requires nanomolar concentrations of hydroperoxides as activator (Kulmacz and Lands, 1983). Glutathione peroxidase and other peroxidases can decrease intracellular concentrations of hydroperoxides and thus reduce cyclooxygenase activity (Marshall et al., 1987). The irreversible self-inactivation of PGHS after a number







69

of catalytic turnovers is another limiting factor in prostaglandin production (Smith and Lands, 1972). Finally, other unsaturated fatty acids can compete with AA for the substrate binding site on PGHS and inhibit prostaglandin synthesis (Lands et al., 1972).
There are numerous examples of endogenous inhibitory activities to prostaglandin synthesis in a variety of other tissues: sheep endometrium (Basu, 1989) and allantoic fluid (Leach-Harper and Thornburn, 1984), renal cortex (Terragano et al., 1978), placenta of the rat (Harrowing and Williams, 1977), human amniotic fluid (Saeed et al., 1982), ovarian follicular fluid of humans (Carson et al., 1986) and cows (Shemesh, 1977), and human decidua (Ishihara et al., 1990). However, none of these inhibitors has been fully characterized and isolated. The presence of an endogenous factor inhibiting prostaglandin synthesis in bovine endometrium has been described in a number of studies (Wlodawer et al., 1976; Shemesh et al., 1984a; Basu and Kindahl, 1987; Gross et al., 1988b). This inhibitor(s) decreased synthesis of PGF2. and all other identifiable cyclooxygenase metabolites including PGE2 (Basu and Kindahl, 1987). The differential effect of pregnancy on PGF (inhibitory) versus PGE2 (no effect) may reflect cellular differences (epithelium versus stroma) in prostaglandin secretion. High levels of inhibitor in epithelial cells (major source of PGF2a) but not in stromal cells (major source of PGE2) could explain the antiluteolytic effect of pregnancy.

In view of these findings, the objective of research in this dissertation is to further characterize the cellular mechanisms by which the conceptus, through IFNt secretion, can regulate endometrial prostaglandin synthesis at the time of maternal recognition of pregnancy in cattle.













CHAPTER 2
REGULATION OF ENDOMETRIAL PROSTAGLANDIN SYNTHESIS
DURING EARLY PREGNANCY IN CATTLE: EFFECTS OF
PHOSPHOLIPASES AND CALCIUM IN VITRO


Introduction


Maintenance of the corpus luteum (CL) during early pregnancy in cattle involves interactions between the conceptus and endometrium (Thatcher et al., 1984b). Increased plasma concentrations of 13,14-dihydro-15-ketoprostaglandin F2, (PGFM) associated with luteolysis are reduced or absent during early pregnancy (Kindahl et al., 1976; Betteridge et al., 1984). Furthermore, in vitro endometrial synthesis of prostaglandin F2, (PGF2a) and PGFM are lower for pregnant than cyclic cows at Day 17 postestrus (Thatcher et al., 1984b; Gross et al., 1988b; 1988c); secretion of prostaglandin E2 (PGE2), in contrast, is either unaltered (Thatcher et al., 1984b; Gross et al., 1988b) or increased slightly during early pregnancy (Gross et al., 1988c; Helmer et al., 1989a). Presence of an endometrial inhibitor of prostaglandin synthesis during the estrous cycle and an increase in inhibitory activity during early pregnancy were demonstrated (Basu and Kindahl, 1987b; Gross et al., 1988a). The conceptus secretory proteins (bCSP) and bovine interferon tau (blFNt) in particular, are believed to be responsible for many of the changes in prostaglandin metabolism that are associated with early pregnancy. For example, blFNt induces a decrease in PGF2a secretion and an increase in PGE2 secretion by endometrial explants from cyclic cows (Helmer et al., 1989a);


70







71

bCSP increase intracellular activity of an inhibitor of prostaglandin synthesis (Gross et al., 1988c). Therefore, a decrease in the ratio of PGF2a/PGE2 is one of the effects of bovine conceptus secretory proteins (Gross et al., 1988b; 1988c) that is observed during early pregnancy. These results indicate a differential regulation of endometrial prostaglandin biosynthesis during early pregnancy in cattle.
Prostaglandin biosynthesis is regulated by a variety of factors including phospholipases and Ca2+ (Marshall et al., 1987). Phospholipase (PL) C is involved in prostaglandin biosynthesis through oxytocin-induced release of two second messengers: inositol 1,4,5-triphosphate (P3) which initiates a rise in intracellular Ca2+, and diacylglycerol (DAG), which subsequently can activate protein kinase C or can be processed to obtain arachidonic acid (AA) (Schrey et al., 1986). Arachidonic acid is the rate-limiting precursor of prostaglandin biosynthesis (Marshall et al., 1987). Phospholipase A-2 (PLA2) induces release of AA from the second carbon of phospholipids, which also results in increased substrate availability for prostaglandin biosynthesis (Vogt, 1978). In addition, Ca2+-dependent PLA2 is a primary focus for the action of Ca2+ (Vogt, 1978; Schrey and Rubin, 1979). Riley and Poyser (1987) reported increased prostaglandin secretion from endometrium of guinea pigs in response to Ca2+. In addition, PGF2. secretion was dependent upon presence of extracellular Ca2+, such that Ca2+ preferentially increased PGF2a as compared to PGE2 secretion. Calcium ionophore A23187 (Cal) increases Ca2+ cycling across cell membranes (Pressman, 1976) and, therefore, is capable of altering prostaglandin biosynthesis (Gemsa et al., 1979; Knapp et al., 1977). Indeed, Cal increases prostaglandin secretion by perifused endometrium from cyclic gilts (Gross et al., 1990). Altered responses to regulators of prostaglandin synthesis due to







72

pregnancy may account for the distinct differences in prostaglandin secretion between endometrial tissues of pregnancy versus the estrous cycle.

Current experiments examined whether PLA2, PLC, extracellular Ca2. and calcium ionophore A23187 have regulatory effects on bovine endometrial PG biosynthesis. In addition, these factors were examined for their ability to exert preferential effects on PGF2a or PGE2 biosynthesis and whether these effects differ between endometrium of cyclic and pregnant cows.


Materials and Methods


Materials

Isotopically labeled [5,6,8,11,12,14,15-3H]-PGF2a (specific activity = 180 Ci/mmole) and [5,6,8,12,14,15-3H]-PGE2 (specific activity = 184 Ci/mmole) were from Amersham Corporation (Arlington Heights, IL). Arachidonic acid was from Sigma Chemical Company (St. Louis, MO). Antiserum to PGF2a was provided courtesy of T.G. Kennedy and antiserum to PGE2 was a gift from the E.L. Lilly Co. (Indianapolis, IN.). Phospholipase A (PLA2, from Naja naja venom, 1300 U/mg protein), phospholipase C (PLC, Type IX, from C. perfringens, 50-150 U/mg protein), and calcium ionophore A23187 (Cal) were from Sigma Chemical Company (St. Louis, MO).
A modified minimum essential medium (MEM; custom formula #87-5007) and other medium ingredients were purchased from Gibco (Grand Island, New York). Medium was prepared as described by Basha et al. (1980) except that phenol red was included and 10 ml of Gibco MEM vitamin solution was added per liter of medium. Additional MEM without Ca2+ (Ca2+ free-MEM) also was purchased from Gibco.







73


Experiment 1: Effects of PLA, PLC and Calcium lonophore on Endometrium from Cyclic Cows
Dairy cows (Holstein and Jersey) were observed for estrous behavior (cyclic, n = 3) and slaughtered at Day 17 postestrus. Reproductive tracts were removed within 30 min after stunning and endometrium from the uterine horn ipsilateral to the CL was isolated from myometrium and minced. Tissue (0.5 g) was placed into duplicate petri dishes containing 15 ml MEM (containing 200 mg/I of CaCl2) for each of the following treatments:
1. Control
2. AA (0.2 mg)
3. PLA2 (1U/ml, 2U/ml, 5U/ml)
4. PLC (1 U/ml)
5. Calcium ionophore A23187 (Cal; 2, 4 and 10 pg/ml).
Medium samples (0.5 ml) were collected from each dish at 4, 6, 8, 10 and 12 h of incubation at 370C on a rocker platform under an atmosphere of 45% 02, 50% N2, 5% CO2. Medium samples were stored immediately at -200C until analyzed.


Experiment 2: Effects of PLA2 PLC and Calcium lonophore in the Presence or Absence of Calcium on Endometrium from Cyclic and Pregnant Cows

Holstein cows were observed for estrous behavior and either bred by artificial insemination (pregnant, n = 4) or not bred (cyclic, n = 3). Cows were slaughtered at Day 17 postestrus. All reproductive tracts were removed rapidly (within 30 min after stunning) and flushed with 40 ml MEM to collect conceptus tissues and confirm pregnant or cyclic statuses. Endometrium from the uterine horn ipsilateral to the CL was isolated from myometrium and minced. Tissue (0.25 g) was placed into duplicate petri dishes containing 7.5 ml MEM with Ca2.







74


(CaC2 = 200 mg/I) and into duplicate petri dishes containing 7.5 ml Ca2+ freeMEM for each of the following treatments:

1. Control 5. PLA2(1 U/ml) + Cal (7.5 pg/ml)
2. AA (0.1 mg) 6. AA (0.1 mg) + Cal (7.5 pg/ml)
3. PLA2 (1 U/ml) 7. PLC (1 U/ml) 4. Cal (7.5 pg/ml) 8. PLC (2 U/ml)
Medium samples (0.3 ml) were collected from each dish at 4, 8 and 12 h of incubation at 370C on a rocker platform under atmosphere of 45% 02, 50% N2, 5% CO2. Medium samples were stored immediately at -200C until analyzed.


Experiment 3: Ability of Endometrium from Cyclic Cows to Interconvert PGF9a and PGEHolstein cows were observed for estrous behavior and slaughtered at Day 17 postestrus (cyclic, n = 4). Reproductive tracts were removed rapidly (within 30 min. after stunning) and flushed with 40 ml MEM to confirm cyclic status. Endometrium from the uterine horn ipsilateral to the CL was isolated from myometrium and minced. Tissue (0.5 g) was placed into duplicate Petri dishes containing 15 ml of MEM for each of the following treatments: 1) (3H)-PGF2a (5 pCi) ; 2) (3H)-PGE2 (5 pCi). Control cultures utilized similar treatments but in the absence of endometrial tissue. Tissues were incubated for 12 h at 390C on a rocker platform under an atmosphere of 45% 02, 50% N2, 5% CO2. Following incubation, medium was stored immediately at -200C until analyzed.
Medium (1 ml) was extracted twice with ethyl acetate (3 ml), frozen at -70 IC, and the organic phase collected. Extraction efficiencies were 82.1 + 4.3% and 62.6 + 6.2% for medium containing (3H)-PGF2. and (3H)-PGE2 respectively. Ethyl acetate was evaporated and samples reconstituted with HPLC mobile







75

phase (1 ml, acetic acid/water/ acetonitrile [0.5/70.5/29]). Samples were stored at -700C until HPLC analysis.
Prostaglandins were separated for each sample with a HPLC system (PerkinElmer Series 4; Perkin-Elmer Corp., Norwalk, CT) using a C18 column (Whatman Partsil, ODS-3, 5 ptm particle size, 10% carbon load), a mobile phase of acetic acid/water/acetonitrile (0.5/70.5/29), and a flow rate of 1 ml/min for 60 min. The mobile phase was then changed to 100% acetonitrile at a flow rate of 1 ml/min for 20 min (to clean the column). Fractions were collected every 30 sec. (0.5 ml) and radiolabel counted using scintillation spectroscopy. Elution profiles were determined for each sample as well as for control samples, and results expressed as percent conversion (% of total elution profile cpm within each peak of interest) for PGF2a, PGE2, PGFM, and unidentified prostaglandins. Elution profiles were validated (peaks identified) by HPLC analysis of non-incubated

(3H)-PGF2a (fractions 57-68), (3H)-PGE2 (fractions 75-85) and (3H)-PGFM (fractions 108-120) standards.


Radioimmunoassay Procedures

Samples of medium were analyzed for PGF2. using a direct radioimmunoassay (RIA) procedure (Gross et al., 1988c) with an antibody characterized by Kennedy et al. (Kennedy, 1985). The antibody had a crossreactivity of 94% with PGF1a; therefore PGF2a experimental responses are designated as PGF. Inter- and intra-assay coefficients of variation were 11.4% and 13.1%, respectively. A similar assay for PGE2 (Gross et al., 1988c) utilized an antibody characterized by Lewis et al. (1978). Inter- and intra-assay coefficients of variation were 9.9% and 12.6%, respectively.







76


Statistical Analyses
Data were analyzed statistically using least squares analysis of variance in the General Linear Models procedure of the Statistical Analysis System (SAS, 1985). Treatments initially were analyzed for Experiment 1 utilizing the model components of cow, treatment, cow x treatment, replicate (cow x treatment), time, cow x time, treatment x time, cow x treatment x time, and residual. Treatments were then analyzed separately to evaluate dose effects. Dose effects of AA, PLA2, PLC and Cal on endometrial PGF and PGE2 secretions were analyzed using the model components of cow, dose, cow x dose, replicate (cow x dose), time, cow x time, dose x time, cow x dose x time and residual. Prostaglandin secretion rates over the 12 h incubation period were linear for all responses, and accumulation rates for each treatment-dose combination were determined using the model components of cow, replicate (cow), treatment x time (linear) and residual.
For Experiment 2, treatments initially were analyzed to determine status (pregnant vs cyclic) effects by using the model components (Table 2-1) of status, cow (status), Ca2+, treatment, status x Ca2+, Ca2+ x cow (status), status x treatment, treatment x cow (status), Ca2+ x treatment, status x Ca2+ x treatment, Ca2+ x treatment x cow (status) , replicate [cow (status) x Ca2+ x treatment], time, status x time, Ca2+ x time, treatment x time, and residual. There were significant effects of status (e.g., status x trt, status x Ca x trt, status x time). Therefore, remaining analyses were separated for each reproductive status. PG secretion rates over the 12 h incubation period were linear for all responses and the accumulation rates for each treatment-dose combination were determined by sorting data for status, Ca2+, and using the model components of cow, replicate (cow), time (linear) x treatment, and residual.




.2/.


~'


77
Results in Experiment 2 were analyzed further as a series of factorial

experiments to examine the effect of PLA2, Ca2+ and Cal (2 x 2 x 2, Tables 2-2,

2-3, 2-4), and the effects of AA, Ca2+ and Cal (2 x 2 x 2 and 2 x 2, Tables 2-5, 26, 2-7). These factorial designs, for both statuses, utilized the model

components of cow, treatment, cow x treatment, time, cow x time, treatment x

time, cow x treatment x time and residual.




Table 2-1. Least squares analysis of variance of PGF and PGE2 secretions by
endometrial explants (Experiment 2) from cyclic and pregnant (status = stat) cows (cow) sampled at different times (time) after treatments (trt; AA,
PLA2 and calcium ionophore) in presence or absence of Ca2+ (Ca).

Source df PGF PGE2 Error term
SS SS
stat 1 47371 15343 cow (stat)
cow(stat) 5 994309 612345 residual
Ca 1 669884* 319476* cow (stat) x Ca
Trt 7 734915** 344750** cow (stat) x trt
stat x Ca 1 76833 6 cow (stat) x Ca
stat x trt 7 352416** 30729 cow (stat) x trt
Ca x cow(stat) 5 345056 138364 residual
trt x cow(stat) 35 547502 115855 residual
Ca x trt 7 111018* 13552 cow (stat) x Ca x trt
stat x Ca x trt 7 88929 27494* cow (stat) x Ca x trt
Ca x trt x cow (stat) 35 230529 51467 residual rep[cow(stat) x Ca x trt] 112 214217 942245 residual time 2 867287** 227119* residual
stat x time 2 22699** 8129** residual
Ca x time 2 33103** 36120** residual
trt x time 14 149088** 29257** residual
residual 426 588741 174960
** P<0.01; * P<0.05







78


Table 2-2. Least squares analysis of variance of PGF and PGE2 basal
secretions by endometrial explants (Experiment 2) from cyclic and pregnant (status = stat) cows (cow) sampled at different times (time) after
treatments (trt): PLA2, Cal and Ca2+ (2 x 2 x 2 factorial).

Source df PGF PGE2 Error term
SS SS
stat 1 6911 5067 cow (stat)
cow (stat) 5 546909 334138 residual
trt 7 449885** 172069** cow (stat) x trt
stat x trt 7 164427* 24872 cow (stat) x trt
trt x cow (stat) 35 338705 156702 cow (stat) x trt time 2 592855** 121356** cow (stat) x time
stat x time 2 12650 1123 cow (stat) x time
cow (stat) x time 10 85727 20316 residual trt x time 14 79926** 30587** cow (stat) x trt x time
stat x trt x time 14 32313 7024 cow (stat) x trt x time
cow (stat) x trt x time 70 133858 21538 residual residual 168 203608 53763
** P<0.01; * P<0.05


Table 2-3. Least squares analysis of variance of PGF and PGE2 secretions by
endometrial explants (Experiment 2) from cyclic cows (cow) sampled at different times (time) after treatments (trt): PLA2, Cal and Ca2+ (2 x 2 x 2
factorial).

Source df PGF PGE2 Error term
SS SS
cow 3 290625 286777 residual
trt 7 382731* 129175* cow (stat) x trt
cow x trt 21 213226 111979 cow (stat) x trt
time 2 422896** 79657** cow (stat) x time
cow x time 6 56001 18164 residual
trt x time 14 77502** 25291 cow (stat) x trt x time
cow x trt x time 42 89028 17771 residual
residual 96 152450 39961
** P<0.01; * P<0.05







79


Table 2-4. Least squares analysis of variance of PGF and PGE2 secretions by
endometrial explants (Experiment 2) from pregnant cows (cow) sampled at different times (time) after treatments (trt): PLA2, Cal and Ca2+ (2 x 2 x 2
factorial).

Source df PGF PGE2 Error term
SS Ss
cow 2 256283 47361 residual
trt 7 231581* 67766* cow (stat) x trt
cow x trt 14 125479 44722 cow (stat) x trt
time 2 182609* 42822** cow (stat) x time
cow x time 4 29725 2151 residual
trt x time 14 34738 12321* cow (stat) x trt x time
cow x trt x time 28 44829 9766 residual
residual 72 51158 13802
** P<0.01; * P<0.05


Table 2-5. Least squares analysis of variance of PGF and PGE2 secretions by
AA-treated endometrial explants (Experiment 2) from cyclic and pregnant (stat) cows (cow) sampled at different times (time) after treatments (trt):
Cal and Ca2. (2 x 2 factorial).

Source df PGF PGE2 Error term
SS SS
stat 1 443872 5428 cow (stat)
cow (stat) 5 605197 209296 residual
trt 3 467925** 128986** cow (stat) x trt
stat x trt 3 55929 994 cow (stat) x trt
trt x cow (stat) 15 450367 53199 cow (stat) x trt time 2 434672** 129875** cow (stat) x time
stat x time 2 12108 1585 cow (stat) x time
cow (stat) x time 10 82688 5771 residual
trt x time 6 11728 25720** cow (stat) x trt x time
stat x trt x time 6 13828 3693 cow (stat) x trt x time
cow (stat) x trt x time 30 43205 18043 residual residual 84 190566 59011
** P<0.01







80


Table 2-6. Least endometrial explants (Experiment 2) from cyclic cows (cow)
sampled at different times (time) after treatments (trt): Cal and Ca2+ (2 x 2 factorial). squares analysis of variance of PGF and PGE2 secretions by
AA-treated

Source df PGF PGE2 Error term
SS Ss
cow 3 93564 204857 residual
trt 3 452503* 80334* cow (stat) x trt
cow x trt 9 383935 37116 cow (stat) x trt
time 2 339197* 89100** cow (stat) x time
cow x time 6 75734 3106 residual
trt x time 6 24973 24035** cow (stat) x trt x time
cow x trt x time 18 37022 11252 residual
residual 48 150625 36594
** P<0.01; ** P<0.05


Table 2-7. Least squares analysis of variance of PGF and PGE2 secretions by
AA-treated endometrial explants (Experiment 2) from pregnant cows (cow) sampled at different times (time) after treatments (trt): Cal and Ca2. (2 x 2
factorial).

Source df PGF PGE2 Error term
Ss SS
cow 2 511632 4439 residual
trt 3 108137 53302* cow (stat) x trt
cow x trt 6 66431 16082 cow (stat) x trt
time 2 129381* 46620** cow (stat) x time
cow x time 4 6953 2665 residual
trt x time 6 3028 7172 cow (stat) x trt x time
cow x trt x time 12 6182 6791 residual
residual 36 39941 22417
* P<0.05; ** P<0.01


Effects of PLC on endometrial prostaglandin secretion also were analyzed
using the model components of status, cow(status), Ca2+, status x Ca2+, cow
(status) x Ca2+, PLC, status x PLC, cow (status) x PLC, Ca2+ x PLC, status x Ca2.







81

x PLC, cow (status) x Ca2+ x PLC, replicate [cow (status) x Ca2+ x PLC], time, status x time, Ca2+ x time, status x Ca2+ x time, PLC x time, status x PLC x time, Ca2+ x PLC x time, status x Ca2+ x PLC x time and residual.
Differences in metabolism of either radiolabeled PGF2a or PGE2 between presence or absence of endometrial tissue in Experiment 3 were evaluated by least squares analysis of variance considering effects of cow and tissue.


Results


Experiment 1
Endometrial PGF and PGE2 secretion increased linearly throughout 12 h of incubation regardless of treatment. Arachidonic acid (AA) increased (P<0.01) PGF and PGE2 secretion rates 103% and 85%, respectively (Table 2-8). In addition, AA increased (P<0.05) the ratio of PGF/PGE2 (Table 2-8). Phospholipase C (PLC) did not alter endometrial prostaglandin secretion (Table 2-8).

Phospholipase A2 (PLA2) increased endometrial PGF secretion (25% increase) in a non-dose-dependent manner (P<0.01); increases were similar regardless of dose (1, 2, and 5 U/ml; Table 1). In contrast, only the lowest dose of PLA2 (1 U/ml) increased (49% increase; P<0.01) endometrial PGE2 secretion, whereas the higher doses (2 and 5 U/ml) had no effect (Table 2-8). Calcium ionophore A23187 (Cal) increased (P<0.01) endometrial secretion of PGF and PGE2 in a dose-dependent manner (Table 2-8). However, PGE2 secretion was increased (14, 77, and 140% for 2, 4, and 10 ptg/ml, respectively) to a slightly greater degree than PGF secretion rates (0, 67, and 107% for 2, 4, and 10 pg/ml







82


respectively). Neither PLC, PLA2 or Cal altered the ratio of PGF/PGE2 (Table 28).


Table 2-8. Accumulation rates (slopes) for PGF and PGE2 by endometrial
explants (0.5 g) from cyclic cows for Experiment 1 during 12 h of
incubation. Results are expressed as ng PGF or PGE2/h.

Accumulation Rate (ng/h) Ratio
Treatment PGF PGE2 PGF / PGE2
Control 56.4 24.7 2.58

Arachidonic Acida 114.6 45.6 3.55

Phospholipase A 2
1 U/mI 71.9 36.9 2.18
2 U/mI 69.1 24.9 2.85
5 U/ml 69.9 26.7 2.60

Ca2. lonophorec
2 pg/ml 52.8 28.0 2.07
4 pg/ml 94.3 43.8 2.13
10 pg/ml 116.7 59.5 2.29

Phospholipase Cd 53.5 28.2 2.43

Intercept (ng) 56.4 31.6

Pooled SE ( ) 8.1 5.2 0.24
a For PGF, PGE2 and ratio, treatment x time interaction (P<0.01). b For PGF and PGE2, treatment x time effects (P<0.01). Effects were dosedependent.
c For PGF and PGE2, treatment and time effects (P<0.01). Effects were not dose-dependent.
d No treatment effects. Effect of time (P<0.01).


Experiment 2

After having established effective doses in Experiment 1 for the stimulation of endometrial prostaglandin secretion by PLA2, Cal and AA in cyclic







83

cows, effects of and interactions between these regulators of prostaglandin were examined. Secretion rates (Table 2-9) of PGF were reduced for endometrial explants from pregnant compared to cyclic cows for control and AA either in the presence or absence of Ca2+ (P<0.05). Arachidonic acid increased (P<0.01) both PGF and PGE2 secretion rates in the presence or absence of exogenous Ca2+ (321 and 99% respectively for endometrium of cyclic cows; 446 and 156% respectively for endometrium of pregnant cows). Furthermore, AA increased PGF secretion rate to a greater degree than PGE2 secretion rate, which resulted in an increase (P<0.05) in the PGF/PGE2 ratio (without AA, 1.8 and 1.5 for cyclic and pregnant cows, respectively; with AA, 3.7 and 3.0 for cyclic and pregnant cows, respectively). Stimulated increases in PGF and PGE2 in response to PLA2 and Cal were not of the magnitude obtained with AA.
Regulation of basal Prostaglandin secretion. Variation in prostaglandin secretion associated with responses to control, PLA2, Cal and PLA2 + Cal in the presence or absence of Ca2+ (2 x 2 x 2 factorial analysis) were considered as measurements of and regulators of basal prostaglandin secretion (Figure 2-1). Mean basal secretion of prostaglandin was higher in cultures with Ca2+ (PGF: 144 > 94 [P<0.01] and PGE2: 91 > 51 [P<0.01] ng per 0.25 g per 8 h). Basal secretion rates of PGF were reduced from endometrial explants of pregnancy compared with the estrous cycle, and a differential response to treatments was detected (Status x Treatment, P<0.05; Figure 2-1). Although Ca2+ stimulated PGF secretion, responsiveness to Cal differed between each reproductive status. The Cal stimulated PGF basal secretion either in the presence or absence of Ca2+ in endometrium from cyclic and pregnant cows. However, endometrial tissue from pregnant cows expressed a hyperstimulation of PGF and PGE2 secretion when treated with Cal in the absence of Ca2+.







84


Table 2-9. Accumulation rates (slopes) for PGF and PGE2 by endometrial
explants(0.25 g) from Experiment 2 during 12 h of incubation. Results are
expressed as ng PGF or PGE2/h.


Reproductive Status
Cyclic Pregnant
Treatmenta Calcium PGF PGE2 PGF PGE2


Control


Phospholipase A2 Arachidonic acid Ca2. lonophore PLA2 + Cal AA + Cal Phospholipase C:
1 U/mI

2 U/ml

Intercept (ng)


+


+

-+

+

-+

+

-+

+

+


13.3 5.8 47.8 32.9 17.3 9.6 20.8 8.2 19.2 7.3 50.6 36.7

9.6 5.3 7.9
2.2 67.5 56.7


9.7
4.3


18.1 9.8 11.1 5.8
12.2 5.2
12.0 5.0 16.2 8.9

6.9
4.3
5.3 3.0
32.4 29.4


6.2
4.5


35.2 23.8 18.8 16.9 11.5 17.7 11.7 6.9 31.8
28.4

8.3 3.1 6.9 2.6 32.2 33.0


4.9 3.2


13.6 7.2 9.9 3.9
11.4 8.7 6.9
4.0 13.6 7.8

7.1
2.4 8.0 1.3 25.6 25.3


Pooled SE ( ) + 1.79 1.00 1.88 0.83
- 1.44 0.61 1.54 0.64
a Treatment x time (linear) interaction (P<0.01) detected within each reproductive status (cycle, pregnant) and Ca2+ (Ca2+, Ca2+-free) group for PGF and PGE2 accumulation rates.







85

Addition of PLA2 to the explants was an attempt to examine potential differential responses in the mobilization of endogenous pools of AA from phospholipids for synthesis of prostaglandins. Addition of PLA2 to endometrial explants from cyclic cows did not stimulate basal PGF secretion, whereas secretion was stimulated in explants from pregnant cows (Status x Treatment, P<0.05). Indeed PLA2 amplified basal secretion of PGF to the basal secretion level of explants from cyclic cows (Figure 2-1, A-B). The stimulatory effect of PLA2 on basal PGF secretion in pregnancy was blocked by the addition of Cal. Since PLA2 was ineffective in cyclic cows, Cal did not exert a negative effect on PGF basal secretion. Basal secretion of PGE2 was considerably less than that of PGF. As observed for PGF, Ca2+ and Cal amplified secretion of PGE2 (Figure 2-1, C-D). Clear effects of pregnancy on PGE2 secretion were not detected in contrast to the attenuating effects of pregnancy on PGF secretion. PLA2 failed to stimulate PGE2 secretion in explants from either cyclic or pregnant cows.

Regulation of stimulated secretion. An index of stimulated secretion of prostaglandin was the response to exogenous arachidonic acid (AA) that was evaluated by analyses of groups containing AA in the presence and absence of Ca2+ and Cal (2 x 2 factorial analysis; Figure 2-2). Availability of AA appears to be limiting in endometrial explants of both cyclic and pregnant cows as noted by the marked increase in both PGF and PGE2 secretion (Figure 2-2) compared to the basal secretion responses observed in Figure 2-1. The stimulatory effects of exogenous AA are far greater than stimulation of basal secretion with PLA2 (release of endogenous AA). Both PGF and PGE2 stimulated secretions were enhanced when Ca2+ was added in the presence of AA (P<0.01) for endometrium of both cyclic and pregnant cows. In the presence of AA, Cal had no effect on stimulated secretion of either PGF or PGE2. The rate











CYCLIC


PGF (ng/0.25g tissue)
-/A


I,


PREGNANT


PGF (ng/0.25g tissue)
B







/ /


+PLA2-Ca +PLA2+Ca -PLA2-ra -PLA2+Ca +PLA2-Ca +PLA2+Ca


* + Ca lonophore [ - Ca lonophore


PGE2 (ng/0.25g tissue)


PGE2 (ng/0.25g tissue)


100
100


50 50



S/ / /-0
-PLA2-Ca -PLA2+Ca +PLA2-Ca +PLA2+Ca -PLA2-Ca -PLA2+Ca +PLA2-Ca +PLA2+Ca


Figure 2-1. (Experiment 2) Interactive effects of calcium (Ca2+), phospholipase
A2 (PLA2), and calcium ionophore A23187 (Cal) on mean endometrial PGF (Panels A & B) and PGE2 (Panels C & D) in medium from cyclic and pregnant cows. Results expressed as ng/0.25 g tissue. There was a Ca2+ x PLA2 X
Cal interaction (P<0.05) for each reproductive status.


86


200 100


A


200


100


0


-PLA2-Ca -PLA2+Ca











CYCLIC


PGF (ng/O.25g tissue)

500/A

400300

200 100

0-


+AA-Ca


+AA+Ca


PREGNANT


PGF (ng/0.25g tissue) lB


+AA-Ca


+AA+Ca


0 + Ca lonophore F - Ca lonophore


PGE2 (ng/0.25g issue)
-C


+AA-Ca


+AA+Ca


PGE2 (ng/0.25g issue)
D






7-


+AA-Ca


+AA+Ca


Figure 2-2. (Experiment 2) Interactive effects of calcium (Ca), and calcium
ionophore A23187 (Cal) on mean PGF (Panels A & B) and PGE2 (Panels C & D) in medium from endometrial explants of cyclic and pregnant cows that received arachidonic acid (AA). Results expressed as ng/0.25 g tissue. There was a Ca2* x Cal interaction (P<0.05) for each
reproductive status.


87


500

400 300

200 100

0


150 100


50


0


150 100 50


0




-







88


of AA stimulated secretion of PGF was less for endometrial explants from pregnant cows versus cyclic cows (P<0.01).


Phospholipase C. The effect of PLC on endometrial PG secretion differed (P<0.05) between pregnant and cyclic statuses (Figure 2-3).


PGF (ng/M.25g tissue)
.A











- PLC-Ca - PLC+Ca +PLC-Ca +PLC+Ca

N Cyclic


PGE2 (ng/0.25g tissue)
B










0
- PLC-Ca - PLC+Ca +PLC-Ca +PLC+Ca

ED Pregnant


Figure 2-3. Interactive effects of calcium (Ca2+) and phospholipase C (PLC)
on mean endometrial PGF (Panel A) and PGE2 (Panel B) in medium from endometrial explants from cyclic and pregnant cows (Experiment 2). Results expressed as ng/0.25 g tissue. There were Ca2+ x, PLC x reproductive status interaction (P<0.01) for mean responses of PGF
and PGE2 from pregnant cows (P<0.05).



Phospholipase C caused a clear decrease (P<0.05) in PGF and PGE2 secretion
rates for endometrium from cyclic cows either in the presence or absence of
Ca2+ (Figure 2-3, Panel A). Similar PLC-induced decreases in PGF and PGE2


100






0


)0







89

were detected for endometrial tissue from pregnant cows in the absence of Ca2. (Figure 2-3) but not in the presence of Ca2+ (Ca2+ x PLC interaction, P<0.05).


Experiment 3
Experiments 1 and 2 demonstrate a differential regulation of PGF and PGE2 and that pregnancy decreases PGF but not PGE2 secretion (Experiment 2). To determine if relative differences in PGF and PGE2 secretion may be due to interconversion of these molecules, endometrial tissue from cyclic cows was evaluated for its ability to convert (3H)-PGF2a to PGE2 and (3H)-PGE2 to PGF2a.



Table 2-10. Percent conversions of (3H)-PGF2a and (3H)-PGE2 by endometrial
explants (0.5 g) from Experiment 3 during 12 h of incubation. Results are expressed as percent of total chromatographic profile CPM (mean + SE)
for PGF2a, PGE2, PGFM and unidentified PG products.

Percent Conversion
Substrate Treatment PGF2a PGE2 PGFM Unidentified

(3H)-PGF2aa Control 88 4.2 7 2.2 3 1.1 2 1.2 Tissue 75 3.1 9 1.1 9 2.3 7 0.9

(3H)-PGE2' Control 4 1.0 86 5.3 1 0.4 9 3.0 Tissue 3 0.5 60 4.2 2 0.6 35 8.5
a Endometrial tissue did not convert (3H)-PGF2a to PGE2; however, there was an increase (P<0.05) in PGFM and unidentified PG products compared to control incubations.
b Endometrial tissue did not convert (3H)-PGE2 to PGF2a or PGFM; however, there was an increase (P<0.01) in unidentified PG products compared to control incubations.


Endometrium was unable to convert (3H)-PGF2a to PGE2. However, (3H)-PGF2a was metabolized to PGFM and unidentified PG metabolites compared to results for control (without tissue) incubations (Table 2-10). Endometrial tissue also







90

was unable to convert (3H)-PGE2 to PGF2a or PGFM as compared to control incubations. However, (3H)-PGE2 was metabolized to unidentified PG metabolites (Table 3) which bound to the HPLC column and were detectable following column cleaning (elution with 100% acetonitrile).


Discussion


Secretion of PGF is reduced, but secretion of PGE2 is not lowered for endometrial explants from pregnant as compared to cyclic cows at Day 17 postestrus (Thatcher et al., 1984b; Gross et al., 1988b; 1988c). Present results confirm these findings and demonstrate a differential regulation of endometrial prostaglandin biosynthesis for PGF and PGE2 during early pregnancy in cattle.
Prostaglandin biosynthesis is regulated by a variety of factors including phospholipases and Ca2+ (Marshall et al., 1987). Endometrial prostaglandin secretion is dependent upon the presence of extracellular Ca2+, and Ca2. preferentially increases PGF as compared to PGE2 secretion by endometrial explants from guinea pigs (Riley and Poyser, 1987). The present results demonstrate a Ca2+ dependency of PG biosynthesis for endometrial explants from cyclic and pregnant cows at Day 17 postestrus. Calcium ionophore A23187 (Cal) increases Ca2+ cycling across plasma membranes and can induce release of Ca2. stored in the vesicles of the sarcoplasmic reticulum (Pressman, 1976). Therefore, this agent is capable of altering prostaglandin secretion (Gemsa et al., 1979; Knapp et al., 1977). Cal (A23187) can be used as an agonist of a PLC-generated second messenger 'P3, to replicate the effect of oxytocin on prostaglandin production (Raw and Silvia, 1991). Cal is capable of increasing PG secretion by perifused endometrium from cyclic pigs (Basha et al., 1980),







91


sheep (Raw and Silvia, 1991) and guinea pigs (Poyser, 1987). However, Lafrance and Goff (1990), using a 2.6 pg/ml concentration of A23187, failed to detect any effect of A23187 on PGF secretion by endometrial explants from heifers at day 19 or 20 postestrus. In the present study, PG secretion increased in response to 4 and 10 pg/ml in Experiment 1 and 7.5 pg/ml in Experiment 2; in agreement with Lafrance and Goff, a 2 p.g/ml dose did not alter PG secretion (Experiment 1). Cal altered basal PG secretion by endometrial explants and the effect of Cal differed depending on the presence or absence of Ca2+ between cyclic and pregnant statuses. In the absence of Ca2+, for instance, PGF secretion was unchanged essentially in endometrium from cyclic cows exposed to Cal and increased markedly in endometrium from pregnant cows. The membrane perturbation induced by A23187 in the absence of Ca2+ appears to initiate a clear response only in endometrium from pregnant cows. This may mirror the differences in membrane content of phospholipids and/or enzymes between the two reproductive statuses.

Calcium ions can activate certain types of PLA2 resulting in increased AA release (Vogt, 1978; Schrey and Rubin, 1979). Addition of PLA2 to the explants was an attempt to mobilize the endogenous pools of AA for prostaglandin synthesis. At day 17 postestrus, the quantity of esterified AA in endometrial phospholipids is lower in pregnant than in cyclic cows (Curl, 1988). Exogenous PLA2 stimulated PGF secretion only in explants from pregnant cows (status x treatment, P<0.05), regardless of the presence or absence of Ca2+. PLA2 increased preferentially PGF secretion to attain levels comparable to PGF production in explants from cyclic cows. The status-specific effect of PLA2 may reflect a greater substrate accessibility for exogenous PLA2 in endometrium from pregnant compared to cyclic cows. The concept of compartmentalization of







92

unsaturated fatty acids (UFA) such as AA was illustrated in a study by Shands and Noble (1984). Mitochondrial and microsomal fractions contained high proportions of UFA compared to plasma membranes and cytosolic factions. It is conceivable that UFA intracellular trafficking may be altered during early pregnancy. AA could be preferentially re-routed away form microsomal membranes rich in phospholipases and prostanoid biosynthetic enzymes. As a consequence, endometrial plasma membranes from pregnant cows would contain a higher proportion of AA which could be more accessible to exogenous PLA2
Although PLA2 from Naja naja venom is a Ca2+-dependent enzyme (Verheij et al., 1981), PLA2 stimulatory effect on prostaglandin secretion was observed even in the absence of extracellular Ca2+, possibly due to activation by intracellular Ca2+ after integration into plasma membranes. Of interest was the marked differential response to Cal between tissues of cyclic and pregnant cows treated with PLA2 (PLA2 x Cal, P<0.01). Recently, Poyser and Fergusson (1993) indicated that stimulation of prostaglandin secretion by PLA2 and Cal may involve different intracellular processes in the guinea pig. However, the negative interaction of Cal with PLA2 in this study of endometrium from pregnant cows suggests common or interdependent intracellular pathways, possibly involving protein kinase C (PKC) and phosphorylation of PLA2 regulating factors. PKC isozymes a, 0 and y require Ca2+ and activating lipid (DAG and/or AA) to be translocated from the cytoplasm to the cell membrane and become activated (Kikkawa et al., 1989). A treatment with PLA2 (source of AA) + Cal (source of Ca2+) may alter prostaglandin secretion potentially by PKC activation. Nishizuka (1992) suggested that, in physiological conditions, the activity of PKC may be sustained, even after the concentration of intracellular Ca2+ is no longer







93

increased, if DAG and cis-UFA both become available. Lafrance and Goff (1990) observed that synthetic PKC activators (PMA and OAG) stimulated PGF2a secretion by endometrial explants from cyclic cows at day 19-20 postestrus. Alternatively, Cal perturbation of the cell membrane could be also detrimental to the stimulatory effect of exogenous PLA2. Nonetheless, this negative interaction with PGF secretion appears to be pregnancy-specific. In the bovine endometrium, the epithelial tissue is the major source of PGF and stromal tissue mainly secretes PGE2 (Fortier et al., 1988). Moreover, Bonney et al. (1987) demonstrated the presence of two PLA2 enzyme types in human endometrium. One is located in stromal tissue and maximally active in the presence of Ca2+, and the other is located mainly in glandular epithelium and inhibited in the presence of Ca2+. Such a tissue-specific regulation of PLA2 activity within the bovine endometrium could also play a role in the regulation of prostaglandin synthesis. In addition, PLA2 activity appears to be regulated by steroids since, in human endometrium, estradiol increased PLA2 activity 6-fold (Bonney and Franks, 1987). Overall, results imply that a decrease in concentration and/or activity of PLA2 may be responsible partially for the preferential decrease in endometrial PGF secretion observed during early pregnancy in cattle.
Availability of AA appears to be limiting in endometrial explants of cyclic and pregnant cows since addition of AA increased prostaglandin secretion dramatically in both tissues. The stimulatory effect of AA on PGF secretion was far greater than the response to exogenous PLA2. The absolute increases in PGE2 secretion due to AA appeared to be equivalent for the two statuses. However, in tissue from pregnant cows, AA did not stimulate PGF secretion to the same extent as in cyclic cows. This implies that, when AA availability is not limiting, endometrial tissue from pregnant cows has a lower PGF production




Full Text

PAGE 1

REGULATION OF ENDOMETRIAL PROSTAGLANDIN SYNTHESIS BY PHOSPHOLIPASES AND INTERFERON TAU AND CHARACTERIZATION OF AN ENDOMETRIAL PROSTAGLANDIN SYNTHESIS INHIBITOR IN THE BOVINE By GU^NAHEL HENRI DANET-DESNOYERS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1994

PAGE 2

Dedie S Monique et Jean Desnoyers et Gregory Seaney

PAGE 3

ACKNOWLEDGMENTS I would like to express my sincere gratitude to the chairman of my supervisory committee, W.W. Thatcher, for his guidance and support during the course of my program. He made possible this training at the University of Florida that I consider to be personally and educationally rewarding. I would like to thank the remaining members of my supervisory committee, W.C. Buhi, P.J. Hansen, M.J. Lawman and F.A. Simmen, for their guidance and suggestions. Sincere thanks are extended to Jesse Johnson and Sean O'Keefe for their invaluable assistance with my research and their technical expertise. I would like to extend special thanks to the many students, technical persons and postdoctoral fellows I have worked with throughout my doctoral program: Lokenga Badinga, Thais Diaz, Susan Gottshall, Timothy Gross, Joyce Hayen, Eric Schmitt, Luzbel de la Sota, Marie-Joelle Thatcher, Caria Wetzels and all the others for their friendship, humor and moral support. Thanks are extended to Monte Meyer for providing me with endometrial tissue from interferon tau-treated cows (chapter 4) and Mary-Ellen Hissem for her eternal smile, sense of humor and secretarial expertise. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS iii ABSTRACT vi CHAPTERS 1 REVIEW OF LITERATURE 1 Phospholipids 1 Linoleic and Arachidonic Acids 5 Arachidonic Acid Mobilization 8 Phospholipases 10 The Cyclooxygenase Pathway 18 Inhibition of Cyclooxygenase 23 Prostaglandin Receptors 26 Estrous Cycle 27 Follicular Growth, Maturation and Dominance 27 Corpus Luteum and Prostaglandins 33 Oxytocin and Prostaglandins 37 Regulation of Uterine Prostaglandin Production 39 Maternal Recognition of Pregnancy 49 Trophoblast Interferons 55 Implications 68 2 REGULATION OF ENDOMETRIAL PROSTAGLANDIN SYNTHESIS DURING EARLY PREGNANCY IN CATTLE: EFFECTS OF PHOSPHOLIPASES AND CALCIUM IN VITRO 70 Introduction 70 Material and Methods 72 Results 81 Discussion 90

PAGE 5

3 EFFECT OF NATURAL AND RECOMBINANT BOVINE INTERFERON x AND OXYTOCIN ON IN VITRO SECRETION OF PGF2„ AND PGE2 BY ENDOMETRIAL EPITHELIAL AND STROMAL CELLS 97 Introduction 97 Material and Methods 99 Results 111 Discussion 122 4 IDENTIFICATION AND QUANTIFICATION OF AN ENDOMETRIAL PROSTAGLANDIN SYNTHESIS INHIBITOR (EPSI) IN THE BOVINE 129 Introduction.... 129 Material and Methods 131 Results 152 Discussion 179 5 GENERAL DISCUSSION 194 REFERENCES '.. 205 BIOGRAPHICAL SKETCH 256

PAGE 6

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy REGULATION OF ENDOMETRIAL PROSTAGLANDIN SYNTHESIS BY PHOSPHOLIPASES AND INTERFERON TAU AND CHARACTERIZATION OF AN ENDOMETRIAL PROSTAGLANDIN SYNTHESIS INHIBITOR IN THE BOVINE By Guenahel Henri Danet-Desnoyers April, 1994 Chairman: William W. Thatcher Major Department: Animal Science The bovine conceptus must prevent uterine luteolytic secretion of prostaglandin Fj^ (PGFjJ in order to maintain pregnancy. The antiluteolytic signal from the conceptus is trophoblast interferon tau (IFNx). The present investigations were conducted to extend our understanding of conceptus and IFNx-mediated regulation of endometrial secretion of PGFj^ and PGEj . Arachidonic acid, phospholipase A2 (PLA2), phospholipase C (PLC), extracellular calcium (Ca^*) and calcium ionophore were examined for regulatory effects on prostaglandin biosynthesis by endometrial explants from cyclic and pregnant cows at day 17 postestrus. Secretion of PGFj^ was lower in endometrium of pregnant versus cyclic cows. Arachidonic acid stimulated prostaglandin endometrial secretion for both reproductive statuses. The stimulatory effect of PLAj on PGFj,, and PGEj secretions was greater in explants of pregnant cows. However, calcium ionophore inhibited the PLAj stimulatory vl

PAGE 7

effect on endometrium of pregnant but not cyclic cows. Findings indicate that availability of AA may limit endometrial secretion of prostaglandin in pregnant cows. The effects of natural and recombinant bovine IFNx (nblFNx and rblFNt) on secretion of RGFj^j and PGEj by endometrial epithelial and stromal cells were assessed in a second phase of investigation. Epithelial cells secreted more PGFja than stromal cells, whereas stromal cells secreted primarily PGEj. Oxytocin stimulated secretion of PGFj^ and PGEj from epithelial cells. However, basal and oxytocin-induced secretions of PGFj^ and PGEj decreased with increasing doses of either nblFNx or rblFNx. Neither nblFNx, rblFNx or oxytocin altered prostaglandin secretion by stromal cells. The antiluteolytic effect of bIFNx is exerted on epithelial cells of the endometrium. Nature of the endometrial prostaglandin synthesis inhibitor (EPSI) present during early pregnancy was characterized. In cytosol, EPSI was associated with serum albumin. Cytosol extraction followed by liquid and gas chromatography indicated that EPSI activity was associated with nonesterified linoleic acid (LA), a competitive inhibitor of cyclooxygenase. Higher EPSI activity in association with higher LA and decreased AA concentrations were found in endometrial microsomes from pregnant versus cyclic cows. In summary, the bovine conceptus, possibly through IFNx secretion, attenuates endometrial secretion of PGFj^jj by altering endometrial lipid metabolism to enhance the ratio of unesterified LA to AA concentrations in microsomes. , ^ vii

PAGE 8

CHAPTER 1 REVIEW OF LITERATURE Phospholipids Phospholipids comprise by far the major component of most membranes with the exceptions of nervous tissue and plant chioroplast membranes. Different phospholipid classes are always found in fixed proportions in a given membrane type, although another membrane in the same cell might be quite different in its phospholipid distribution (Wirtz, 1974). Most of the structural phospholipids are derived from the same precursor pools and interconversions of one phospholipid into another are commonplace. Not only is each phospholipid present at a specific percentage of the total, but each maintains its own characteristic distribution of fatty acyl side chains. This is achieved through regulatory factors controlling rates and selectivity of the synthetic and degradative enzymes. In eukaryotic cells, most of the phospholipids are synthesized in the microsomal membranes (endoplasmic reticulum). However, the rapid exchange of phospholipids between the microsomes and other functionally different membranes renders it impossible to consider the pools as being independent. Phosphatidic acid is an important branching point in the pathway of phospholipid synthesis. The fatty acids (FA) found esterified to the 1 -position of phosphatidic acid and other phospholipids are almost ail saturated (Lands and Crawford, 1976) and in particular, phosphatidic acid contains large amounts of palmitic acid in position 1. In contrast to the preference for saturated FA at the 1

PAGE 9

2 1 -position, the 2-acyl ester of phosphatidic acid and other common phospholipids is usually a c/s-unsaturated component such as oleic, linoleic or arachidonic acid (AA). It appears that the rate-limiting steps in diglyceride utilization for phospholipid formation are the cytidyl transfer reactions which are, at least in part, controlled by the concentrations of substrates. The apparent for chicken brain microsomal 1,2-diacyl-s/7-glyceroi:CDP-ethanolamine transferase decreases in the presence of diacylglycerols, and unsaturated FA stimulate phosphatidylethanolamine (PE) formation, while saturated FA appear to inhibit it (Mead et al., 1986). At low concentrations of free fatty acids (FFA) normally present in tissues, their nature and the presence of diacylgycerols and phosphorylated bases appear to be controlling phosphoglyceride biosynthesis. However, as the concentration of exogenous FA increases, the formation of triacylglycerols becomes dominant. In a study with cultured dissociated brain cells, Yavin and Menkes (1974) found that added labeled linoleic acid (18:3n-6) is first incorporated into diand triglycerols. However, as the conversion of 18:3/7-6 into longer more highly unsaturated analogues progressed, these tended to be transferred to ethanolamine phosphoglycerides from triacylglycerols and choline phosphoglycerides. In the liver, 80% of phosphatidylinositol (PI) contains stearate and arachidonate at the sn-^ and sn2 positions, respectively. This probably arises from deacylation/reacylation reactions in which the fatty acid constituents are remodeled following synthesis of the lipid molecule. In 1968, Wirtz and Zilversmit identified a soluble intracellular protein in rat liver that was capable of binding phosphatidylcholine (PC) and transferring it from one population of donor membranes to a second population of acceptor membranes. Since this initial observation, phospholipid transfer proteins have

PAGE 10

been identified in almost all mammalian tissues (Wirtz and Gadella, 1990). Phospholipid transfer proteins fall into three main categories: (1) those specific for PC, (2) those with high activity for PI and less but significant activity for PC, and (3) those with transfer activity with most phospholipids and cholesterol. The action of these proteins is a one-for-one exchange of lipid molecules between donor and acceptor membranes using ATP dependent and ATP independent mechanisms with time constants that can vary by several orders of magnitude for different lipids (see Voelker, 1991). Some lipids such as free fatty acids (FFA) or phosphatidic acid may have sufficient solubility to allow some passive transport, but most other lipids require these mechanisms. In summary, differences in membrane composition can be accomplished by the selective transport of lipid molecules and/or by specific metabolic events that occur in the acceptor membranes. It is very clear that a major function of phospholipids is to store precursors of second messengers that activate specific processes in the cell. The eicosanoids were the first phospholipid-derived second messengers identified and will be reviewed later; this section will review the roles of phosphatidylisinositol (PI), diacylglycerol (DAG) and platelet activating factor (PAF) as intracellular second messengers. Hokin and Hokin (1953) observed that incubation of pigeon pancreas with acetylcholine caused the release of the digestive enzyme amylase. If 32p was included in the incubation, there was a rapid labeling of PI. It is now known that acetylcholine binding to its cell surface receptor results in hydrolysis of inositol lipids to DAG and inositol phosphates. The 32p incorporation was the result of resynthesis and rephosphorylation of the lipids through a sequence of reactions known as the PI cycle (see Hawthorn, 1982). The initial event in inositol lipid metabolism occurs within 20-30 seconds

PAGE 11

of the binding of the hormone to its receptor and involves primarily the catabolism of phosphatidyl inositol-4,5-biphosphate (PIPj) into DAG and inositol1,4,5-triphosphate (IP3). For example, 20% of cellular PIPj is degraded within 30 seconds of exposure of hepatocytes to vasopressin (Fisher et al., 1984). IP3 causes the release of Ca^* from the endoplasmic reticulum. A receptor for IP3 has been identified as a 313 kDa protein that has both Ca^* channel activity and a ligand binding site (Ferris et al., 1989). The receptor is phosphorylated by cyclic adenosine monophosphate (cAMP)-dependent protein kinase (PKA) which causes a decrease in both IP3 binding and Ca^* release. The other product of PIP2 hydrolysis is DAG which activates PKC. This kinase was discovered by Nishizuka and co-workers in 1977 and seven isozymes have been identified, all containing regulatory and kinase domains (Nishizuka, 1988). PKC appears to exist in the cytosol in an inactive form and, when DAG is generated in the plasma membrane, is translocated to this membrane and activated. PKC also requires Ca^* and phosphatidylserine (PS) for activity. PAF (1-alkyl-2-acetyl-sn-glycerol-3-phosphocholine) belongs to a large family of ether-linked glycerolipids that serve as both membrane components and as intracellular mediators. Arachidonic acid plays an important role in PAF biosynthesis because it participates at the substrate level (alkyl-arachidonoylglycerophospho-cholines) in the phospholipase Aj (PLA2) catalyzed reaction (Ramesha and Pickett, 1986). Interestingly, it has been shown that PAF production can be blocked by inhibition of acetyltransferase activity in bovine neutrophils by arachidonic and oleic acids (Remy et al., 1989). Some species of ether lipids associated with membranes act as reservoirs for polyunsaturated fatty acids. The protective effect of ether-linked groups is probably due to their

PAGE 12

5 ability to slow the rate of hydrolysis of the acyl group at the sn-2 position by PLAj. High affinity binding sites for PAF on the plasma membrane have been shown for platelets, neutrophils, smooth muscle cells, and lung tissue (see Snyder, 1990). It appears that PAF interaction with its receptor is coupled to a G protein, since PAF stimulates GTPase activity (Houslay et al., 1986). PAF is produced by preimplantation mouse (O'Neill, 1985), human (O'Neill et al., 1987) and sheep (Battye et al., 1991 ) embryos following in vitro fertilization. The ability of human embryos to secrete PAF is correlated with the establishment of pregnancy (O'Neill et al., 1987) and the use of PAF antagonists can inhibit implantation in mice (Spinks and O'Neill, 1987). Basically, PAF is an embryonic autocrine grov^h factor which can also suppress oxytocin-induced release of prostaglandins and extend estrous cycle length. The reader is referred to O'Neill (1990) for a review of the physiological role of PAF and its interaction with the arachidonic acid cascade in the stimulation of embryonic development. Linoleic and Arachidonic Acids All eukaryotic organisms contain polyenoic fatty acids in the complex lipids of their membranes and most mammalian tissues can modify acyl chain composition by introducing one or more double bonds. But animal requirements for polyunsaturated acyl chains can not be met exclusively by de novo metabolic processes. Animals are absolutely dependent on plants for linoleic (18:2n-6) and a-linolenic (18:3n-3) acids which are the two major precursors of the n-6 and n-3 fatty acids. In vitro studies indicate that 18:2/7-6 competes with 18:3a7-3 for the 2-acyl position of PC and for conversion to long chain metabolites (Brenner and Peluffo, 1966). Severe effects observed in experimental animals and

PAGE 13

humans (Wene et al., 1975) in the absence of these dietary fatty acids include a dramatic decrease in body weight, dermatitis and increased skin permeability to water, enlarged kidneys, reduced adrenal and thyroid glands, decreased cholesterol accumulation and altered fatty acyl composition in many tissues, impaired fertility (Ziboh and Hsia, 1972), and ultimately death (see Sinclair, 1990). When exposed, in vivo or in vitro, to large amounts of unsaturated FA (LA, AA, 20:5n-3, 22:5/7-3), the cells compensate by incorporating the fatty acid in low amounts and by reducing other fatty acids in order to maintain the membrane physico-chemical characteristics (Stubbs and Kisielewski, 1990). It seems reasonable to conclude that the cell will not allow the proportion of saturated and unsaturated FA to change appreciably because the membrane fluidity would change in an undesirable manner. The nature of the compensation mechanism which appears to allow sensing of the membrane's physical properties by the elongase-desaturase system is still unknown (Quinn, 1981). Stubbs and Kisielewski (1990) proposed that such a mechanism may be termed 'homeoviscous adaptation'. The four n-6 acids in the sequence from LA to AA individually have similar potency to reverse these effects of a deficiency. AA (20:4a7-6) can be formed from LA {^8:2n-6) by the alternating sequence of A6 desaturation, then chain elongation of the 18:3/7-6 intermediate to form 20:3n-6, and A5 desaturation resulting in 20:4/7-6. AA can be further metabolized into higher polyunsaturated FA such as 22:4/7-6 and 22:5n-6 (Ramwell et al., 1977). The A6-, A5desaturases are membrane bound extrinsic and immunologically distinct proteins bound to the cytoplasmic surface of the endoplasmic reticulum by hydrophobic tails and require an electron transport chain for desaturation (Fujiwara et al., 1984). Frequently, AA is referred to as an essential fatty acid

PAGE 14

7 (EFA) since it is required in many tissues; but an adequate dietary supply of LA can be converted to AA in most circumstances. When ingested and absorbed, AA is transported bound to albumin (Spector et al., 1969) or lipoproteins as part of triglycerides, cholesterol esters or phospholipids (Lamberth and Gates, 1976). Normally, human plasma concentration of free AA are very low («3 ^ig/ml). Contrarily to non-ruminants, dietary EFA absorption is reduced in cows because of biohydrogenation in the rumen (Viviani, 1970). Most of the AA absorbed is incorporated into phospholipids destined to be integral part of cell membranes (Flower and Backwell, 1976). In the liver and most other tissues of animals in a normal, balanced state, the only members of the n-6 family to accumulate in relatively large quantities are LA and AA; much lower levels of intermediates 18:3/7-6 and 20:3/7-6 are detected (Cook, 1991). This suggests that A6desaturase is a rate-limiting step in the enzymatic sequence. On the other hand, A5-desaturase activity measured in vitro with 20:3/7-6 as substrate is approximately equal to, or lower than the limiting A6-desaturase activity (Cook, 1978; Sprecher, 1981). The activities of A6-, A5-desaturases are under endocrine control. Corticosteroids, either in vivo or when added to hepatocytes in culture, inhibit A5and A6-desaturase activity trough direct synthesis of a "lipocortin" like (e.g., PLAj inhibitor) soluble protein (Brenner, 1990). The precise mechanism by which this cytosolic protein inhibits the desaturase reaction is not clear; prevention of substrate access to the enzyme rather than a direct inhibition of the enzyme is a possibility. Concentrations of free AA in the cytoplasm and plasma are very low compared with esterified levels of AA (Irvine, 1982). In plasma, it is bound electrostatically and hydrophobically to albumin (Spector, 1975). The plasma concentration varies from 3 ^ig/ml (1 to 2% of total free fatty acid) in humans to

PAGE 15

8 22-122 ng/ml in dogs (Ramwell et al., 1977) and 0.15-0.22 ng/ml in guinea pigs (Leaver and Poyser, 1981). In guinea pig uterus, 93% of total AA was esterified to phospholipids. On day 7 of the estrous cycle, 54% was found in PE, 33% in PC and 12% in PS. At day 15, the amount present in PE decreased significantly to 33%, while PS and PC contained about 20% (Leaver and Poyser, 1981). In endometrium of cyclic and pregnant cows at day 18 postestrus, 73% of the total AA was esterified to phospholipids (Lukaszewska and Hansel, 1980). Curl (1988) determined the distribution of AA in the most common phospholipid classes. At day 17 postestrus, the quantity of AA bound to phospholipids in endometrium from pregnant cows was lower than in cyclic cows. Overall, 64% was found in PE, 18% in PC, while PS and PI contained about 10% of total AA esterified to phospholipids (Curl, 1988). The amount of esterified AA in the endometrium declined considerably between days 17 and 19 of the estrous cycle, probably because of prostaglandin production associated with luteolysis. On the other hand, the amount of esterified AA increased between days 1 7 and 19of pregnancy (Curl, 1988). J In guinea pig uterus, the 7% of AA not esterified in phospholipids was distributed as follows: triglycerides > cholesterol esters > free = diglycerides > monoglycerides (Leaver and Poyser, 1981). In bovine endometrium, the distribution of non phospholipid associated AA was: free > cholesterol esters > triglycerides (Lukaszewska and Hansel, 1980). i Arachidonic Acid Mobilization * In resting tissues, the levels of free AA are low and controlled by acylCoA hydrolase for release, and by acyl-CoA transferase for esterification into

PAGE 16

9 phospholipids. In macrophages from bone marrow, AA liberation could be observed only if acylation was inhibited (Kroner et al., 1981). When human platelets were exposed to a specific inhibitor of lysophosphatide acyltransferase, PGE2, prostacyclin and thromboxane 63 secretions increased significantly (Goppelt-Struebe et al., 1986). These observations suggest that control of AA release can involve acyltransferases and future research may prove acyltransferases to be critical enzymes in the control of free intracellular AA. The major rate limiting step for production of eicosanoids is AA release. The term eicosanoids is used to designate a group of oxygenated, C20 fatty acids derived from AA. In many tissues, membrane phospholipids are highly enriched with AA at the sn-2 position (Flower and Blackwell, 1976). Importantly, release of arachidonate is a selective process and other fatty acids normally found in phospholipids are not mobilized (Dennis, 1987). Lands and Samuelsson (1968) addressed the legitimate question as to whether AA was converted into prostaglandins (PG) on the phospholipids and stored in an esterified form and released in response to specific stimuli. They concluded that the primary rate limiting step in PG synthesis was the release of AA from phospholipids or triacylglycerols that would result from activation of lipases. Perfusion of the adrenal gland with acetylcholine in the rat was followed by the release of PG into the perfusion fluid. The presence of Ca^* in the perfusion fluid was essential for the release of PG (Shaw and Ramwell, 1967), which suggested that Ca^* may be needed for the activity of the phospholipase that releases the PG precursors from phospholipids. PC and PE, known to carry polyunsaturated FA at sn-2 position of the glycerol moiety, were originally favored as source of PG precursors through the action of PLAj. Release of AA from cellular stores in response to stimuli including hormones, growth factors, physical stress, is a

PAGE 17

' 10 very quick process (see Smith, 1989). However, the rapid turnover of phosphoinositides in various tissues and cells called the attention to phosphoinositides as a possible source of AA for PG synthesis (Dawson and Irvine, 1978). In human platelets, PLC can cleave phosphoinositol from such lipids, yielding a 1 ,2-diacylglycerol, which after hydrolysis by DAG lipase, provided AA for PG synthesis. Prescott and Majerus (1983) demonstrated that DAG lipase, while equally active for 1-stearoyland 2-arachidonoylgycerol, hydrolyses the 1stearoyl-2-arachidonoyl-glycerol sequentially, releasing first stearic acid and then AA. Domin and Rosengurt (1993) demonstrated that platelet-derived grovAU factor (PDGF)-stimulated release of AA and its metabolites from Swiss 3T3 cells is a biphasic process. The initial phase involves a rapid activation of PLA2 and is independent of de novo RNA and protein synthesis, whereas the second major phase is dependent upon rapid expression of a protein. Phospholipases The phospholipases are a group of enzymes with the common property of hydrolyzing phospholipids. The classification of the phospholipases is based on their site of attack on the phospholipids and several types of phospholipases are involved in liberating fatty acids from phospholipids. Phospholipase A^ (PLA^) removes the acyl group from carbon sn-^ and PLAj from carbon atom 2 of the phospholipid (Figure 1-1). Some phospholipases hydrolyze both acyl groups and are termed phospholipases B (PLB). Cleavage of the glycerophosphate bond is catalyzed by PLC, while the removal of the base group (e.g., choline, ethanolamine, serine) is catalyzed by phospholipase D.

PAGE 18

11 20}K06 160 P-Cho2be PLA. 2O>4(06 1830 -CcA 1-AT 20>4C06 PLA, OH 182©6 -CoA 2-AT 182(D6 OH P-Cho2fae CoA 18 O P-Cho]be 20>lC06 18 O P-eho3ne CoA -18 O -p-ChoSje Figure 1-1: The fatty acids (palmitic, 16:0; arachidonic, 20:4co6) at the sn-^ and sn-2 positions of phosphatidylcholine (P-choline) can be deacylated by phospholipases (PLA, and PLA,) and other fatty acids (stearic, 18:0; linoleic 18:2co6) can be reacylated by acyltransferases (acylCoA:lyso Phosphatidyl choline 1-acyl transferase, 1-AT; acylCoA:lyso Phosphatidyl choline 2-acyl transferase, 2-AT). Two PLA, have been purified from Escherichia coli based on their sensitivity or resistance to detergents (see Dennis, 1983). The detergent sensitive enzyme preferentially degrades phosphatidylglycerol and can act as a

PAGE 19

12 transacylase. Transacylation occurs when an acyl-enzyme intermediate is formed in a two-step reaction where in the second step, the acyl group is transferred nonspecifivally to a hydroxyl acceptor of a soluble alcohol or of a lipid (Waite, 1991). Transacylation reactions are now known to be important events in many cells and provide a means to redistribute acyl groups between phospholipids without a deacylation-reacylation cycle. Rat liver lysosomes contain a soluble PLA^ that is a glycoprotein with optimal activity at pH 4.0 and does not require Ca^* for activity (Van den Bosch, 1980). The action of PLA, results in the production of a lysophospholipid that can be further cleaved through the action of specific lysophospholipases. The distinction between PLB and lysophospholipases is not clear since PLB that has been purified and characterized has a high lysophospholipase activity (Kawazaki and Saito, 1973). Such activity would be expected since PLB deacylates at both sn-1 and sn-2 positions, with a lysophospholipid intermediate that is subsequently deacylated. The sum of two activities of the enzyme, PLAj and lysophospholipase, gives the total activity of PLB; but it is unclear if the enzyme has two distinct binding sites (one for diacylphospholipids and a second one for lysophospholipids). Phospholipases were the first phospholipases to be described: Bokay (1877-1878) recognized a factor that degrades PC in pancreatic fluid and Kyes (1902) observed that cobra venom had hemolytic activity on erythrocytes. Since then a number of cellular PLAjS from mammals have been purified and compared to their pancreatic and venom counterparts. The nomenclature of the different isoforms of PLAj has been based on structure and sequence. The 14 kDa enzymes have been categorized into types I (mammalian pancreatic and cobra venom), II (mammalian platelet, synovial fluid and viper venom) and III

PAGE 20

13 (bee venom). The 85 kDa PLAj has been given the designation of type IV (mammalian platelet, macrophage). The most studied mammalian Ca^* dependent PLAjS are the type 1 14 kDa PLAj digestive enzymes which are released as proenzymes from the pancreas (Abita et al., 1972). Pancreatic-like PLAjS exist in a variety of tissues such as lung and kidney (Sakata et al., 1989) and can bind to high affinity cell surface binding sites (Hanasaki and Arita, 1992). Human type II PLAj has been characterized as active at pH 7-10 and requires Ca^* as a cofactor. This enzyme does not exhibit a preference for certain FA in the sn-2 position but is fairly selective toward the phospholipid class in that it prefers PE or PS and poorly hydrolyses PC (Kramer et al., 1989; Hara et al., 1989). Crystallization of the enzyme has shown that the catalytic site and mechanism are conserved relative to other type I and II 14 kDa PLA2, but that subtle differences exist such as variation in the hydrophobic channel that binds substrate (Wary et al., 1991; Scott et al., 1991). The proposed role of Ca^* catalysis is to orient the phospholipid by binding the sn-3 PO4 and to activate the substrate by binding the carbonyl oxygen of the sn-2 fatty acyl group (Mayer and Marshall, 1993). Optimal activity is observed with aggregated forms of phospholipid such as membranes, vesicles and micelles and considerably weaker activity with monomeric or dispersed substrates. The role of cytokines in the production and release of extracellular type 1114 kDa PLAj was studied recently using models of septic shock. Injections of endotoxin lipopolysaccharide (LPS) produced an elevation of PLAj activity in rabbit serum associated with an increase in PLAj protein level (Pruzanski and Vadas, 1991). This increase occurs much later than the endotoxin-induced secretion of cytokines such as tumor necrosis factor (TNF), interleukin-1 (IL-1) and IL-6, which is consistent with a cytokine-mediated

PAGE 21

. M " -'^ 14 induction of PLAj synthesis and release. Pruzanski and Vadas (1991) hypothesized that LPS stimulates macrophages to release cytokines and these induce secretion of PLAj from a variety of sources, propagating the inflammatory process. Type II 14 kDa PLAj exists also in a cell-associated form that can serve a number of functions, such as phospholipid turnover, repair of lipid peroxidation, or mobilization of AA for generation of PG or other derivatives (Pernasetal., 1991). The 85 kDa PLAj has been identified in cells as an activity with 2 specific characteristics: (1) a strong preference for sn-2-arachidonoyl phospholipids and no preference for the type of sn-1 acyl linkage (acyl versus alkyi) or for the type of phospholipid functional group (ethanolamine versus choline) (Diez et al., 1992); and (2) a translocation activity such that it is alv\^ays found in the cytosolic fraction when cells are broken in the presence of Ca^* chelators (EDTA or EGTA), but is largely in the microsomal fraction when broken in the presence of free Ca^* concentrations (300 to 700 nM) approaching those found in activated cells (Rehfeldt et al., 1991; Krause et a!., 1991). Analysis of the sequence revealed a 45 amino acid domain homologous to the Ca^* binding domain of protein kinase C (PKC) but no other sequence homology including type I1 14 kDa PLA2 (Sharp et al., 1991). Ca^* is required for translocation but catalytic activity can be obtained in the absence of Ca^* using high concentration of NaCI, suggesting that Ca^* is required for association with membrane lipids but not for catalysis (Wijkander and Sundler, 1992). Recently, a side by side in vitro comparison of the 85 kDa PLA2 and type II 14 kDa PLAj revealed that both enzymes have similar activity profiles as a function of Ca^* concentration (Marshall and McCarte-Roshak, 1992). The 85 kDa PLAj has a clear preference for sn-2-arachidonoyl over SA7-2-oleoyl-phospholipids, while s/7-2-palmitoyl is a

PAGE 22

very poor substrate (Clark et al., 1990; Diez et al., 1992). The degree of saturation seems to be the only determinant of good substrate activity. Lin et al. (1992) provided evidence that 85 kDa PLAj can be involved in hormonestimulated release of AA using transfected CHO cells over-expressing this enzyme. Treatment of these cells with thrombin or ATP resulted in increased AA release compared to control cells v/ithout over-expressed enzyme. Interestingly, CHO cells over-expressing 14 kDa PLAj did not show increased AA release in similar experiments (Clark et al., 1991). In addition, 85 kDa PLAj has been shown to possess a lysophospholipase A^ activity which is not observed with the 14 kDa PLA2 (Leslie, 1991). Coexistence of type II 14 kDa PLA2 and 85 kDa PLA2 in the same cell or tissue has been reported. For example, the macrophage cell line P388Di was shown to have distinct cytosolic and membrane-associated Ca^*-dependent PLAj as well as cytosolic lysophospholipase activity (Rose et al., 1985). Another class of sn-2-acylhydrolase activity has been described in almost all major organs: Ca^*-independent PLAj which is largely uncharacterized and may represent one or more isozymes (Pierik et al., 1988). A 40 kDa Ca^*independent, sn-2-arachidonoyl selective isozyme has been reported in rabbit and human myocardial tissues (Hanzen et al., 1991; Hanzen and Gross, 1992, respectively). In human ischemic heart tissue, Ca^*-independent PLA2 was found predominantly in the microsomal fraction and accounted for 98-99% of total PLA2 activity (Hanzen et al., 1991). Another Ca^^-independent, sn-2arachidonoyl PE selective activity (58 kDa dimeric isozyme) has been identified in sheep platelet cytosol and was recently cloned (Zupan et al. , 1 992). For more detailed reviews on phospholipases, the reader is referred to Waite (1991) and Smith (1992).

PAGE 23

As originally defined, lipocortins are glycoproteins which specifically inhibit PLA2 in vitro and in vivo (Di Rosa et al., 1984). Lipocortins are related to calpactins which fall under the broad umbrella of annexins (Crumpton and Dedman, 1990). Lipocortin I (= p35 = calpactin II) was the first to be cloned and sequenced (Wallner et al., 1986) followed by lipocortin II (= p36 = calpactin I; Kristensen et al., 1986). The p36 protein is a major substrate for phosphorylation by PKC (Gould et al., 1986). Additional related proteins (lipocortins lll-VI) have been identified which all bind Ca2* and phospholipids (see Pepinsky et.. 1988; Hunter, 1988; Klee, 1988; and Goulding and Guyre, 1992 for reviews). A decrease in annexin 1 expression has been found to be associated with the onset of labor (Lynch-Salamon et al., 1992). An antiphospholipase protein named gravidin has been described in amniotic fluid and its production from the amnion decreases with labor (Wilson and Ganendren, 1992). However, the way in which these proteins inhibit phospholipase activity remains controversial. In a commentary, Davidson and Dennis (1989) discussed the foundation of lipocortin's inhibitory activity and suggested that substrate (phospholipids) and/or Ca2* depletion could explain some of the experimental observations. Phospholipases C are phosphodiesterases that cleave the glycerophosphate bond on phospholipids resulting in production of DAG. One of the earliest reports of a mammalian PLC came from Sloane-Stanley (1953) who demonstrated the release of inositol from phosphatidylinositol, catalyzed by a brain preparation. PLC have now been purified from the cytosolic fraction of muscle, brain, platelets and ram seminal vesicles. The properties of phospholipases C purified from diverse sources have been compared, and Rhee et al. (1989) developed a nomenclature. So far five basic types are listed (a, p, y

PAGE 24

17 , 6, e); each has distinct Immunological properties and little homology in their predicted amino acid sequences. PLC is part of a second messenger system whereupon activation of cell surface receptors, PLC is activated and hydrolyses PIP2 into two second messengers: IP3 is released into the cytoplasm where it interacts with Ca^* storage sites to mobilize intracellular Ca^*; and DAG acts within the plasma membrane to activate PKC (see review by Rhee et al., 1989). Direct evidence for involvement of trimeric G proteins in Pl/Ca^* signaling required observation of GTP-dependent activation of PLC (Cockcroft and Gomperts, 1985; Litosch et al., 1985). Binding of the ligand to the extracellular portion of the receptor results in dissociation of the GDP from the G protein (a subunit) and replacement by GTP; the GTP-bound a subunit then dissociates from the py subunits and activates PLC (Bourne et al., 1990; Birnbaumer, 1990). There are several isoforms of the G protein a-subunits (aj, as,aq) and, for example, activation of PLC(5 occurs via (Smrcka et al., 1991). Oxytocin stimulates production of prostaglandins via receptor activation of PLC, which could also be in part regulated by activation of PKC (Schrey et al., 1986). There is also evidence that phosphorylation is involved in regulation of PLC activity. PLCy is a substrate for tyrosine kinase and contains a conserved sequence motif known as Src homology 2 (SH2) domain which allows PLCy to bind in a specific manner to certain tyrosine-phosphorylated proteins. Via SH2 domains, PLCy is able to bind to specific tyrosine residues within the cytoplasmic domain of certain receptor tyrosine kinases when these tyrosines become autophosphorylated during the receptor activation (Anderson et al., 1990; Margolis et al., 1990). This could be involved in the translocation process of cytosolic PLC to the membrane for activation. Finally, the major pathway of sphingomyelin degradation involves a special phospholipase C, sphingomyelinase. Although some phospholipases

PAGE 25

C that act on PC also work on sphingomyelin, a number of distinct sphingomyelinases exist (Waite, 1987). Some interest has focused on sphingomyelinase in the plasma membrane that, when coupled to ceramidase action, yields sphyngosine, a negative regulator of PKC (Merrill, 1989). Although PLAjand PLC appear to be the primary enzymes responsible for increasing free intracellular AA, release of AA may not be the only rate limiting step in the production of eicosanoids. It seems that phospholipid substrate availability for PLAjand PLC can play an important role in subsequent release of AA. Phospholipids are not distributed randomly among all cellular membranes, but instead different lipid classes (e.g. phospholipids, triglycerides, cholesterol esters, free fatty acids, etc.) are often enriched in different membranes. This enrichment may involve site-specific synthesis or degradation of phospholipids, remodeling through deacylation-reacylation reactions, translocation or some combination of these mechanisms (see Pagano, 1990). Thus the capacity of cells to release AA and synthesize eicosanoids is directly dependent upon an adequate distribution of phospholipids in their membranes. The following sections will present information regarding the metabolism of AA into eicosanoids with a particular emphasis on the cyclooxygenase pathway which is responsible for the production of prostaglandins. The Cvclooxygenase Pathway Once released through the action of phospholipases, free AA can enter the 'arachidonate cascade' leading to the eicosanoids. This cascade includes three major pathways: cyclooxygenase, lipoxygenase and cytochrome P-450 epoxygenase. Prostanoids, which include prostaglandins and thromboxanes, are

PAGE 26

19 formed via the cyclooxygenase pathway (Needleman et al., 1986); leukotrienes and lipoxins are formed via the lipoxygenase pathway (Needleman et al., 1986); and epoxides and diols are formed through the P-450 epoxygenase pathway(Fitzpatrick and Murphy, 1989). ' Cyclooxygenase is often used as a synonym for prostaglandin endoperoxide synthase [E.C.1. 14.99.1] also called PGH synthase (PGHS) but, in fact, this enzyme exhibits two different activities: (1) a cyclooxygenase {bisdioxygenase) activity that generates prostaglandin G (PGG) from polyunsaturated FA; and (2) a hydroperoxidase activity that converts PGG into PGH. These two activities are associated with the same heme-binding glycoprotein that has been purified to homogeneity from microsomes of sheep and bovine seminal vesicles (Hemler et al., 1976; Van der Ouderaa et al., 1977; Miyamoto et al., 1976). Cyclooxygenase and peroxidase activities have distinct binding sites for their lipid substrates which are located on the cytoplasmic side of the endoplasmic reticulum (Marshall and Kulmacz, 1988). Recent evidence points to the existence of multiple isoforms of PGHS (Xie et al., 1991; Kujubu et al., 1991; O'Banion et al., 1992). Constitutively expressed PGHS-1 has a mRNA size of 2.7-3.0 Kb and the inducible form (PGHS-2) 4.0-5.5 kb (O'Banion et al., 1992; Diaz et al., 1992). PGHS-1 is an integral membrane homodimer with a subunit molecular weight of 72 kDa. The amino acid sequence deduced from the cDNA for the sheep enzyme indicates a molecular weight of 65.5 kDa and the presence of four consensus sites (Asn-X-Ser/Thr) for A/-glycosylation (DeWitt, 1991). The sequences of cDNA clones for PGHS from sheep, mouse and human indicate a high degree of homology between species and that the protein initially contains a signal peptide of 24-26 amino acids which in all species is cleaved to a mature 576 amino acid protein (Smith and Marnett,

PAGE 27

I 20 1991). The presence of the signal peptide suggests that the enzyme crosses the endoplasmic reticulum during its synthesis, but it is unclear how the mature PGHS is associated with the membrane since it contains no obvious transmembrane hydrophobic sequences (Smith and Marnett, 1991). PGHS-1 is probably responsible for the constitutive production of prostaglandins involved in "housekeeping" functions (Smith, 1989). PGHS-2, which has about 62% homology with PGHS-1, is expressed following cell activation (Xie et al., 1991). The synthesis of PGHS-2 is stimulated by serum, growth factors and phorbol esters in fibroblasts (Kujubu et al., 1991) and by chorionic gonadotrophin in granulosa cells (Sirois and Richards, 1992, Sirois et al., 1992). The substrates for cyclooxygenase activity are two molecules of oxygen and one molecule of FA. A variety of polyunsaturated FA can be substrates for cyclooxygenase. FA containing at least three methylene-interrupted c/s-double bonds beginning at n-6 are converted into PGG derivatives (Hamberg and Samuelsson, 1967a). The rates of oxygenation vary considerably with the number of carbons in the FA and the presence of additional double bonds. The best substrates (K^= 2-10 ^M) for the cyclooxygenase reaction are AA (20:4a7-6) and di-homo-/-linolenic acid (20:5a7-6)( Marshall et al., 1987; Odenwaller et al., 1991). In vivo, the most common cyclooxygenase substrate is AA but unsaturated FA containing two methylene-interrupted double bonds (20:2n-6 and 18:2aj-6) can be oxygenated to monohydroxy FA (Hemler and al., 1978). Cyclooxygenase catalysis involves an initial activation of the substrate FA followed by oxygen insertion, carbon skeleton rearrangement and a second oxygen insertion (Hamberg and Samuelson, 1967b). In fact this process is very similar to non-enzymatic autoxidation of some polyunsaturated FA (Nugteren and Van Dorp, 1966). Cyclooxygenase has a strict requirement for a

PAGE 28

21 hydroperoxide to activate the 13-pro-S hydrogen on the FA substrate. If hydroperoxide concentration is below 10 nM, cyclooxygenase activity is inhibited. The best peroxide activators are 1 5-HpETE and PGG which are also the best substrates for the peroxidase activity of PGS (Kulmacz and Lands, 1983). The peroxide activity of PGHS catalyses the reduction of various hydroperoxides to alcohols and a heme group is essential for this activity. PGG generated by the cyclooxygenase is the major endogenous substrate reduced into PGHj, which is the substrate for production of all PG and thromboxanes (Samuelsson, 1978). The peroxidase activity of PGHS resembles other peroxidases, including horseradish peroxidase and cytochrome c peroxidase, in being non-specific toward the electron donors. The endogenous electron donor is not known but, among naturally-occurring molecules, the best substrates are epinephrine and uric acid (Markey et a!., 1987). Uric acid is not an efficient reductant but is present in high concentration (« 300 |xM) in human plasma (Smith and Marnett, 1991), likewise GSH is a relatively poor reducing substrate but is present in very high concentrations in most cells (« 6 mM) (Marshall et al., 1987). Prostaglandin synthesis appears to be cell specific considering that often one cell type produces predominantly one type of PG (Smith et al., 1991). The PG are classified as belonging to the "I", '2* or '3' series depending on the number of double bonds they contain. Within each series, PG are further classified as 'A', 'D', 'E', 'P or T types, depending on the presence of various oxygen-containing substituents. These substituents are critical for the biological activities of the various PG; the most important positions on the carbon skeleton being C-9, C-10 in the ring structure and C-15 in the side chain (Grastrom,

PAGE 29

22 1981). For example, the primary difference between PG of the "F" series and the "E" series is the substitution of a hydroxy! group (F) rather than an oxygen (E) at position C-9. Little is known about the enzymes converting PGH into the various PG. Prostaglandin Fj^ (PGFjJ is formed following (1 ) reduction of PGHj by PGF synthase, or (2) reduction of prostaglandin Ej (PGEj) by a 9-ketoreductase. PGF synthase has been purified from bovine lung by Watanabe et al. (1985) and the cDNA sequence of this enzyme contains a 323 amino acid open reading frame coding for a protein with a molecular weight of about 38 kDa (Watanabe et al., 1989). This enzyme also converts PGHj to prostaglandin D2 and, at a much lower rate, PGEj to PGFj^. A 9-keto-reductase activity, which catalyses conversion of PGEj into PGFj^, has been found in renal cortex (Korff and Jaraback, 1982) and placenta (Jaraback et al., 1983) but not in bovine uterine tissue (Wlodawer et al., 1976). However the K„ value for PGEj is quite high (« 300 ^iM) suggesting that PGEj synthesis is not the natural substrate and that in vivo PGFj^ synthesis is unlikely to involve this enzyme. PGEj is synthesized by non-oxidative rearrangement of the endoperoxide PGHj. This reaction can be catalyzed by a GSH-dependent PGE synthase [E.C.5.3.99.3] (Ogino et a!., 1977); but there is also evidence for a nonenzymatic rearrangement (Hamberg and Samuelsson, 1973). Partially purified PGE synthase from ovine vesicular glands was reported to have a molecular weight of 60-70 kDa (Moonen et al., 1982). Immunoprecipitation using monoclonal antibodies indicated that a 17.5 kDa protein is responsible for at least 45% of the total PGE synthesis activity in ovine vesicular microsomes (Tanaka et al., 1987). Metabolism of PGF 2a ^^'^ PGEj follows essentially the same pathway, with metabolism occurring primarily in the lungs. First, oxidation of the 15-

PAGE 30

23 hydroxyl group to a 15-keto group is catalyzed by the 15-hydroxy-prostaglandin dehydrogenase enzyme. Then it undergoes reduction by 13,14-prostaglandin reductase to form 15-keto-13,14-dihydro-prostaglandin, the main plasma metabolite. A limited amount of PGFj^ is metabolized in the endometrium (Curl et al., 1983; Guibault et al., 1984; Knickerbocker et al., 1986a). Urinary metabolites are formed by one or two steps of p or co oxidation (Kindhal, 1980). Both 15-keto-prostaglandin and 15-keto-13,14-dihydro-prostaglandin have virtually no biological activity (Behrman and Hall, 1982). However, injections of 10 or 15 mg of 13,14-dihydro-PGF2„ at day 10 postestrus can induce premature luteolysis and estrus in the cow (Mllvae and Hansel, 1983). For a more extensive review of enzymes involved in metabolism of PGM, the reader is referred to Needleman et al. (1986) and Smith et al. (1991). Inhibition of Cvclooxvoenase There appear to be two different inactivation processes for PGHS: (1) auto inactivation of cyclooxygenase ('suicide' reaction) and (2) peroxidedependent inactivation of cyclooxygenase and peroxidase activities. When PGHS is incubated in presence of AA, the cyclooxygenase activity is Inactivated before all the AA is metabolized (Smith and Lands, 1972). The rate of this 'suicide' reaction is linked directly to the rate of cyclooxygenase catalysis but is relatively independent of newly formed metabolites. 'Suicide' inactivation of cyclooxygenase appears to be an intrinsic property of this catalytic activity. Each mole of purified PGHS can form about 1300 moles of PGGj before the cyclooxygenase activity disappears, although most of the peroxidase activity remains (Marshall et al., 1987).

PAGE 31

. :> , 24 Inhibition of cyclooxygenase activity can potentially involve several factors: (1) reduction in concentrations of either oxygen or FA substrate; (2) reduction in concentration of hydroperoxide necessary for initiation; (3) interference with AA binding to the cyclooxygenase active site. Molecular oxygen concentration does not appear to be a limiting factor on the rate of cyclooxygenase since the for O2 for cyclooxygenase is about 5 \iM and O2 concentrations in tissues are greater than 20 iiM (Smith and Marnett, 1991). The concentration of unesterified AA in tissues is approximately 20 which is well above the of AA for cyclooxygenase (Lands et al., 1978; Marshall etal.. 1987). Both cyclooxygenase and peroxidase activities are activated in presence of hydroperoxides (Hemler and Lands, 1980; Markey et a!., 1987). In cells that form prostaglandin endoperoxide only upon hormonal stimulation, free AA must be inaccessible to cyclooxygenase. Cyclooxygenase activity is blocked by reduced glutathione (GSH) and in the presence of excess GSH peroxidase (Smith and Lands, 1972). Marshall et al. (1987) described the presence of cytosolic factors in ram seminal vesicles that potentiate the inhibitory effect of exogenous GSH peroxidase. They reported that many cells contain levels of GSH and GSH peroxidase that would inhibit cyclooxygenase activity in vitro. However prostaglandin synthesis may still proceed in such tissues because of cell compartmentalization of cyclooxygenase (membrane-bound) and GSH peroxidase (cytosolic) (Marshall et al., 1987). . Lands et al. (1972) demonstrated that n-3 and n-9 FA having 18 to 22 carbons cause significant inhibition of cyclooxygenase activity. Pace-Asciak and Wolfe (1968) reported that linoleic (18:2a7-6) and linolenic (18:3n-6) acids inhibit PGHS from sheep seminal vesicles but concentrations required for inhibition

PAGE 32

25 were high (1.8 to 5.0 mM); they also noted that pre-incubation with the FA resulted in a greater degree of inhibition, especially at low substrate concentrations. Other n-3 unsaturated FA (18:3n-3, 20:5a7-3 and 22:6n-3), which are not substrates for cyclooxygenase, can still compete for binding at the substrate site. In the case of 20:3n-3, the competitive inhibition constant was 6 ^iM. This competitive binding was stronger for the more highly unsaturated acids. The activity of cyclooxygenase toward 20:5/7-3 is less than 50% of that observed with AA, a fact that may account for the anti-thrombogenic effect of fish oils (Gulp et al., 1979). Another constituent of fish oils, docosahexaenoic acid (22:6/7-3) is a strong inhibitor of AA oxygenation with a Kj of about 0.38 \iM (Corey et al., 1983). On the other hand, the n-6 acids such as linoleic cause both an instantaneous competitive inhibition and a progressive and nonreversible loss of cyclooxygenase activity (Lands et al., 1972). Acetylenic analog of AA: eicosa-5,8,11,14-tetraynoic acid (ETYA) is known to irreversibly inhibit cyclooxygenase in the presence of oxygen and hydroperoxide activators (Vanderhoek and Lands, 1973). Two c/s-epoxieicosatrienoic acid (EET) isomers, derived from the epoxygenase pathway, have been found to inhibit AA metabolism by the ram seminal vesicle cyclooxygenase (Fitzpatrick et al., 1986). Concentrations required for 50% inhibition were 30 ^iM for 14,15-c/s-EET and 50 nM for 8,9-c/s-EET. The potency of 14,15-c/s-EET surpassed that of nonsteroidal anti-inflammatory drugs such as ibuprofen and aspirin. In platelet suspensions, this inhibitory effect of 14,15-c/s-EET was confined to the cyclooxygenase enzyme; the lipoxygenase enzyme was not affected (Fitzpatrick etal., 1986). Inhibition of cyclooxygenase activity can also be achieved using steroidal or non steroidal anti-inflammatory drugs. Anti-inflammatory steroids appear to

PAGE 33

26 diminish cyclooxygenase activity by inhibiting translation of the mRNA for PGHS (Koehler et al., 1990; Sebalt et al., 1990). In general, non-steroidal drugs compete with AA for cyclooxygenase active site even though they bear very little or no structural resemblance to AA. Subsequent to competitive inhibition, several of these drugs form covalent (aspirin = salicylic acid ) or non-covalent bonds (indomethacin) with the enzyme resulting in irreversible inactivation of cyclooxygenase but not peroxidase activity (Smith et al., 1991; Smith and Marnett, 1991). Meade et al. (1993) demonstrated that non-steroidal antiinflammatory drugs can exhibit some selectivity toward PGHS-1 or PGHS-2. Prostaglandin Receptors Prostaglandin receptors are classified as FP, EP, DP, IP and TP, with subsets of EP-1, EP-2 and EP-3; the first letter corresponding to the type of prostaglandin that is the major ligand (Kennedy et al., 1983). The subset definitions are based on responsiveness to drugs; for example, EP-1 and EP-3 mediate contractile activity of PGEj whereas EP-2 mediates relaxant activity (Senior et al., 1992). Water reabsorption in the kidney collecting tubule is one of the most extensively characterized systems to study cellular and molecular effects of PGE2. Arginine vasopressin (AVP)-induced water reabsorption is mediated by cAMP and stimulates PGEj synthesis. The inhibitory and stimulatory effects of PGEj have been shown to be mediated through distinct stimulatory (EP-2) and inhibitory (EP-3) PGEj receptors (Smith. 1989). Addition of GDP or GTP increased binding affinity of PGEj to cell membrane preparations by 2 fold, and this effect of guanine nucleotides on binding was eliminated by treatment with pertussis toxin (Smith, 1989). Specific association of PGEj

PAGE 34

27 receptor with G, has been demonstrated in canine and rabbit kidney (Watanabe et al., 1986). Recently, Sugimoto et al. (1992) cloned a functional EP-3 receptor from a mouse lung cDNA library, and the predicted structure suggested that EP3 was a member of the G protein-linked receptor family. . ^ , _ Estrous Cycle Sexually mature cows display recurring periods of sexual receptivity which defines the boundary of the period referred as the estrous cycle. The sexual receptivity lasts for about 16.9 + 4.9 h (Schams et al., 1977) and the estrous cycle averages approximately 20 to 21 days (Hansel et al., 1973). The estrous cycle can be can divided into four phases: (1) proestrus, during which final maturation of the preovulatory follicle occurs; (2) estrus, which is the period of sexual receptivity; (3) metestrus, which is the period of ovulation and development of the corpus luteum (CL); and (4) diestrus, the period dominated by a functional CL and as a result high plasma concentrations of progesterone. These four periods involve endocrine and biochemical changes which result in the release of an ovum, and, if fertilized, possible establishment of pregnancy. Failure of the embryo to survive leads to reoccurrence of the cycle. The estrous cycle is the result of complex interactions between hypothalamus, pituitary, ovary and uterus. Follicular Grov^h. Maturation and Dominance In the cow, the number of primordial follicles («1 50,000) present in the ovaries is fixed at birth, but there is a natural reduction to approximately 3000 by

PAGE 35

28 the age of 15 to 20 years (Erickson, 1966). The regulation of the rate at which the non-growing follicles are stimulated to start development is not well understood but it appears to be partly dependent on the size of the pool of primordial follicles (Krohn, 1966). Once development of a follicle is initiated, it is continuous until the follicle is ovulated or becomes atretic (Peters and Levy, 1966). In sheep, it takes approximately 6 months for a primordial follicle to develop into a large dominant follicle (Cahill and Mauleon, 1980). A primordial follicle consists of the ovum surrounded by a single layer of epithelial cells. Once stimulated, the follicular epithelial cells proliferate, become cuboidal, and form several layers of granulosa cells characteristic of a secondary follicle. The ovarian stroma contributes a vascularized theca interna and theca externa which are separated from the non vascularized granulosa layer by a basal lamina. When the antrum develops, the follicle is termed an antral follicle in which the oocyte is suspended by the cumulus oophorus granulosa cells in the antrum. Mature antral follicles with a diameter greater than 10 mm in diameter are referred to as Graafian follicles. It appears that the development of a primary follicle to the antral stage does not require gonadotropins, since in sheep antral follicles up to 2 mm in diameter were observed in hypophysectomized animals (Driancourt et al., 1979; McNatty et al., 1990). Follicle stimulating hormone (FSH) and luteinizing hormone (LH) are required for further development to ovulatory size (Lohstroh and Johnson, 1966). Both granulosa and theca cells are necessary for estrogen production by the follicle. The two cell theory for estrogen production was demonstrated by Fortune and Armstrong (1978). Theca cells produce androgens that are converted into estrogens by granulosa cells containing aromatase enzyme. LH stimulates synthesis of androgens from cholesterol in thecal cells, whereas FSH

PAGE 36

stimulates production of aromatase enzyme (Richards, 1980) and LH receptors (McNatty, 1979). Carson et al. (1981) noted a reduction in the ratio of estradiol to androgen in total follicular fluid was associated with atresia in all sizes of follicles. They suggested that the decrease in aromatase activity was of major importance In the process of atresia. This led to the classification of estrogen active and estrogen inactive follicles (Ireland and Roche, 1982). Estrogen active follicles have much higher capacity to bind gonadotropins than estrogen inactive follicles, and this capacity is correlated to follicular diameter (Ireland and Roche, 1983). ; Pierson and Ginther (1984, 1988) reported that there appear to be two waves of large follicle growth and that the ovulatory follicle is selected at about 3 days before ovulation. Other reports (Sirois and Fortune, 1988; Savio et al., 1988; 1990) described three waves of follicular development during bovine estrous cycles. Each wave is characterized by the emergence of several growing follicles (> 5 mm) from a pool of small follicles. One of these emerges as the selected dominant follicle and continues to develop while the others become atretic. The dominant follicle reaches a maximum size of 10 to 15 mm in diameter and remains dominant (i.e., inhibits recruitment of new follicles) for 7 to 10 days, until it becomes atretic, to be replaced by a newly selected dominant follicle that emerged from the recruitment phase of the next follicular wave. If luteal regression occurs during the grovi^h phase or early period of dominance, the dominant follicle can develop into the healthy and estrogenic follicle (up to 20 mm) and ovulate. However a large unhealthy and poorly estrogenic follicle at the time of PGFj^-induced CL regression does not ovulate but regresses and new follicles develop into ovulatory follicles within 48 to 72 h (Ireland and Roche, 1982). i

PAGE 37

30 The luteal phase lasts until about 17 to 19 days postestrus. High progesterone levels are associated with low secretion of pituitary gonadotropins. The decrease in progesterone associated with luteolysis results in qualitative and quantitative increases In LH and FSH secretion which stimulate final follicular growth and maturation. Decreasing concentrations of progesterone result in increased frequency and amplitude of LH pulses and a gradual decrease of FSH associated with final maturation of the follicle (Foxcroft, 1978; Flowers et al., 1991). There is a rapid rise in estradiol originating from the preovulatory follicle that reaches peak concentration at the onset of estrus (Wetteman et al., 1972). Increasing estradiol causes increased LH receptor number and thus increased responsiveness to LH (Hansel and Convey, 1983). Loss of progesterone negative feedback effects coupled with the stimulus of high estradiol concentrations trigger the release of the preovulatory surge of LH resulting in luteinization and ovulation of the granulosa and theca cells. Cyclic ovarian follicular development is a complex process that involves proliferation and differentiation of various cellular components of the follicle (see Greenwald and Terranova, 1988 for review). This has been viewed classically as being regulated by gonadotropins and steroid hormones. However additional molecules of ovarian origin are also regulators of these processes. Among these are polypeptides such as inhibin, activin, folllstatin and growth factors, including insulin-like growth factors (IGFs), epidermal growth factor and fibroblast growth factor. For comprehensive reviews of the role of these ovarian regulators, the reader is referred to Adashi (1989), Adashi et al. (1989), Geisthovel et al. (1990), Adashi et al. (1991) and Adashi et al. (1992).

PAGE 38

31 Following luteinization of the theca and granulosa cells and ovulation, a corpus hemorrhagicum is formed. Luteinization begins approximately 6 h after onset of estrus directed by the effects of LH (Hansel, 1966). Alila and Hansel (1984) determined the origin of large and small cells in the bovine CL using monoclonal antibodies to membrane preparations of theca and granulosa cells collected from ovaries after the LH surge and before ovulation. Large luteal cells were shown to be derived from granulosa cells and small luteal cells from the theca interna cells. Although there is some disagreement as to the absolute numbers of large and small cells at various stages of the estrous cycle, the large cells account for less than 10% of the total luteal cell population in cows (Hansel et a!., 1987) and sheep (Rodgers et al., 1984). It appears that large and small luteal cells of the ovine (Harrison et al., 1987) and porcine (Lemon and Mauleon, 1982) CL synergize to promote production of progesterone. Large cells produce greater amounts of progesterone under basal conditions than small cells. However, small ovine and bovine luteal cells are very sensitive to LH, while large luteal cells are relatively insensitive to LH (Alila et al., 1988b; Farin et al., 1989). In vitro, treatment with LH or cAMP analogs can double progesterone production from bovine large cells (Alila et al., 1988b). Purified small luteal cells produce very little progesterone unless stimulated by LH (Alila et al., 1988b; Farin et al., 1989) or other activators of cAMP/PKA transduction system (Hoyer and Niswender, 1985). Small and large cells differ in a number of morphological features such as nuclei and mitochondria positions and presence of oxytocincontaining secretory granules (Fields et al., 1985; Hansel et al., 1987). The characteristics of the small and large bovine luteal cell types have been described extensively by Hansel et al. (1987) and Niswender and Nett (1988).

PAGE 39

32 The signal transduction systems involved in regulation of steroidogenesis in luteal cells has generated a great deal of interest. It has become apparent that cAMP/PKA second messenger pathway and the phosphatidylinositol-Ca^*PKC effector system are important in regulating luteal cell function. Bovine small luteal cells in vitro contain low resting intracellular [Ca^*] and are able to produce basal levels of progesterone in absence of Ca^* in the culture medium. However, in small luteal cells, Ca^* ions are necessary for LH stimulation of progesterone synthesis and increased cAMP levels are associated with LH action (Alila et al., 1988a). In contrast, bovine large luteal cells contain high levels of resting intracellular [Ca^*] and require Ca^* for basal progesterone production (Alila et al., 1989). In small bovine cells, it appears that the intracellular [Ca^*] increase associated with LH stimulation of progesterone is caused by Ca^* mobilization from intraand extra-cellular sources; while in large luteal cells, it is primarily due to an extracellular Ca^* influx. In contrast, Wiltbank et al. (1989a) could not find any evidence for LH-induced increase in free [Ca^*] within large or small ovine luteal cells. PKC also seems to be an essential component of luteal cell function. PKC has been detected in small and large luteal cells from cattle (Davis and Clark, 1983), sheep (Wiltbank et al., 1989b) and pigs (Wheeler and Veldhuis, 1987). Contrasting results have been reported from studies in which cultured luteal cells were treated with phorbol esters to activate PKC. In ovine large luteal cells, phorbol-1 2-myristate-1 3-acetate (PMA) was a potent inhibitor of progesterone secretion (Wiltbank et al., 1989a), but had no effect on large bovine luteal cells (Alila et al., 1988b). In non-stimulated small ovine luteal cells, activation of PKC with PMA had no effect; however when small cells were stimulated with LH, PMA was inhibitory to progesterone production. Wiltbank et

PAGE 40

33 al. (1991) and Hansel et al. (1991) have reviewed extensively the second messenger systems involved in the regulation of luteal cell function. The ability of the CL to produce progesterone is also dependent upon substrate availability (cholesterol) and activity of steroidogenic enzymes. Cholesterol is delivered to ovarian tissues by receptor-mediated uptake of high and low density lipoproteins (see Grummer and Carroll, 1988). Wiltbank et al. (1989b) demonstrated that PMA decreased cholesterol side-chain cleavage activity in large but not in small ovine luteal cells. However, progesterone production was not affected. The 33-hydroxysteroid dehydrogenase / A^-A^* isomerase (3P-HSD) is required for conversion of pregnenolone to progesterone. The activity of Sp-HSD in the ovine CL is not a limiting factor (Caffrey et al., 1979) and is not affected by activation of the PKC second messenger system (Hawkins et al., 1993). There is a good correlation between levels of 3P-HSD mRNA and 3p-HSD enzyme activity in the bovine CL (Couet et a!., 1990). Hawkins et al. (1993) showed that small ovine luteal cells contained less 33HSD mRNA than large cells and that message for 3P-HSD was lowest at day 15 postestrus, a time when luteolysis is initiated. v;;v Corpus Luteum and Prostaglandins " , i " Loeb (1923) made the first observation of the involvement of the uterus with luteal regression in guinea pigs. He noted that "extirpation" of the uterus performed a few days following ovulation, prevented luteolysis and extended CL life span from the normal 14-15 days to 60-80 days. He also noted that maintenance of the CL prevented ovulation and that hysterectomy had no effect on ovulation if performed prior to ovulation. Loeb (1927) suggested three

PAGE 41

I , . " . 34 mechanisms to explain this association between the uterus and luteolysis: there could be nervous connections between the CL and uterus; the uterus could compete with the ovary for a limited blood flow and thus removal of the uterus would increase blood flow to the ovary and sustain the CL for a longer period of time; the uterus produces "an internal secretion which exerts a specific, abbreviating effect on the life of the corpus luteum". He rejected the first two possibilities believing the third to be the most likely explanation. Wiltbank and Casida (1956) demonstrated that hysterectomy of the ewe and cow also resulted in extended CL life span, confirming the concept of uterine control of CL maintenance proposed by Loeb (1923). Studies in which the ovary (Goding et al., 1967) or both the ovary and uterus (Harrison et al., 1968; McCraken et al., 1970) were transplanted to the neck of the ewe demonstrated that uterine control of CL function was not systemic but localized. In cows and sheep, Kiracofe et al. (1966; 1973) demonstrated that uterine luteolytic effect was mediated through local ipsilateral vasculature and transfer of luteolysin involved uterine veins. Passage of the luteolysin from the uterus to the ovary was determined to occur via a counter current transfer from the uterine vein to the ovarian artery and rendered possible by the convolutions of the ovarian artery on the surface of the uterine venous drainage. Surgical anastomoses of uterine veins or ovarian arteries in hemi-hysterectomized cattle and sheep confirmed the existence of the veino-arterial pathway (Ginther, 1974; 1981, Mapletoft and Ginther, 1975). Pharris and Whyngarden (1969) first investigated the luteolytic role of PGFja through vasoconstriction of the utero-ovarian vein. Numerous reports have since confirmed PGRj^ as a luteolysin in cows, sheep, pigs, horses, rats, rabbits and other species (see Horton and Peyser, 1976); but the

PAGE 42

vasoconstriction theory has not been supported since direct infusions of PGFj^ into the ovary of sheep induced luteolysis without any corresponding decrease in ovarian blood flow (McCracken et al, 1970) and PGFj^ inhibited CL progesterone production in vitro (Demers et al., 1973). In other studies, addition of PGFja to total dispersed bovine luteal cells increased progesterone synthesis (Hlxon and Hansel, 1979), and more recently, this stimulation was shown to be limited to small bovine luteal cell. (Alila et al., 1988b). In the cow, intrauterine infusions of PGFja resulted in higher concentrations of PGFj^ in the ovarian artery than the carotid artery (Hixon and Hansel, 1974). Likewise, (3H)-PGF2„ infused into the uterine vein of ewes appeared in the ovarian arterial plasma after 30 minutes, whereas the concentration of PGFj^ remained very low in the iliac artery (McCracken, 1971). Woifenson et al. (1985), using frequent blood sampling, observed that the decrease in progesterone at luteolysis (days 18, 19 and 20 postestrus) was associated with increased levels of PGFj^ in the ovarian vein but not in the carotid artery, i The lymphatic system also may play an important role in the transfer of PGFja to the ovary. Abdel Rahim et al. (1984) reported that CL extension occurred in animals with intact utero-ovarian vasculature system but all other utero-ovarian connections cut off. In addition, the concentration of PGF2„ in uterine lymph of nonpregnant ewes increased from day 12 postestrus (Abdel Rahim et al., 1983). When (3H)-PGF2„ was infused into the uterine lumen of nonpregnant ewes, concentrations of labeled PGFj^ in uterine lymph increased and remained high for a longer period of time than in the uterine venous plasma. The transfer rate of (3H)-PGF2„ from uterine lymphatics to ovarian artery was actually higher than the transfer rate from uterine vein (0.4% versus 0.3%). A similar transfer cannot be excluded in the cow.

PAGE 43

36 The presence of PGFj^ receptors on luteal cells is a requirement for PGFj „ luteolytic action. Binding sites for POFj^ have been localized on luteal cell membranes in the cow (Kimball and Lauderdale, 1975), sheep (Powell et al., 1974), pig (Gadsbey et al., 1990), horse (Kimball and Wynegarden, 1977), human (Rao et al., 1977) and rat (Wright et al., 1979). PGFj^ receptor has been purified partially (Hammarstrom et al., 1975; Samuelsson et al., 1978), and the estimated molecular weight is 107 kDa (Samuelsson et al., 1978). High and low affinity binding sites on luteal membranes have been reported, and it appears that ligand binding to the high affinity sites is Ca^*-dependent (Rao, 1975). More recently, Balapure et al. (1989) provided evidence that a high affinity PGFj^ receptor is present on the ovine large luteal cells and a low affinity receptor on both large and small luteal cells. In early studies, addition of PGFj^ to total dispersed bovine luteal cells resulted in increased progesterone synthesis (Hixon and Hansel, 1979). This apparent discrepancy between in vivo and in vitro effects of PGFj^ has raised the issue of the mechanism of action of PGFj^ at the level of luteal cells. Several intracellular biochemical events following binding of PGFj^ to its receptor have been described. Alila et al. (1988b) showed that PGFj^ increased progesterone synthesis only in small bovine luteal cells. PGFj^ had no effect on basal production of large cells, but inhibited LH-stimulated progesterone secretion; a result which suggested that PGFj^, in vivo, acts on these cells. A similar effect of PGF2„ was reported for ovine large luteal cells (Schwall et al., 1986). The PGFj^-induced increase in progesterone by bovine small luteal cells appears to be mediated through PLC hydrolysis of PIPj, generation of DAG and activation of PKC rather than by increase in cAMP (Davis et al., 1988). In sheep, PGF2„ activates PKC and inhibits progesterone production (Wiltbank et

PAGE 44

al., 1991). The absence of effect of PGFj^ on ovine small luteal cells (no progesterone inhibition, no activation of PKC or increase in intracellular calcium) is consistent with the absence of high affinity PGFj^ receptors on this cell type. PGF2„, which unlike LH has no effect on cAMP generation, increases intracellular [Ca^*] in both large and small bovine luteal cells (Alila et al., 1989; 1990). PGFj^ increases intracellular [Ca^*] to levels sufficiently high to be cytotoxic and to cause transport and exocytosis of oxytocin granules (Chegini and Rao, 1987; Wiltbank et al., 1989b). One possibility to explain the lack of success of in vitro models in reproducing in vivo acute responses of luteal tissue to PGFja might be the absence of cell-cell contacts. Miyamoto et al. (1993), using a microdialysis system which allows intraluteal application of treatments, observed a plateau in progesterone secretion when PGFj^ was applied. Other biochemical events have been described in response to PGFja binding to luteal cell membranes: decrease in cell membrane permeability (Riley et al., 1989), increase in synthesis of inositol phosphates (Duncan and Davis, 1991) and increase in production of superoxide radicals (Sawada and Carlson, 1991). Benyo et al. (1991) showed that, in vivo, expression of bovine class II major histocompatibility complex (MHC) antigens increased at day 18 postestrus in cyclic cows compared to day 18 pregnant cows. He hypothesized that PGF2„ may induce luteolysis by altering MHC antigens expression to stimulate an immune response in luteal tissues. Oxytocin and Prostaglandins Evidence for the involvement of oxytocin in the process of luteal regression includes the fact that immunization against oxytocin delays luteolysis

PAGE 45

38 (Sheldrick et al., 1980), as does the administration of exogenous oxytocin (Lafrance et Goff, 1985) or oxytocin receptor agonist (Jenkin et a!., 1991). In cattle and sheep, oxytocin can be detected in preovulatory follicles (Wathes et al., 1984; 1986; Schams et al., 1985) but reaches a maximum in mid-luteal phase CL and then gradually declines (Wathes et al., 1984; Parkinson et al., 1992b) to reach very low levels by the preovulatory period. In early pregnancy the pattern of oxytocin release by the CL is similar (Parkinson et al., 1992b). In sheep and cattle, mRNA for oxytocin was at a maximum at days 1-3 of the estrous cycle (Jones and Flint, 1988; Ivell et al., 1990) and mRNA levels declined at a steady rate until mid-cycle. Control of oxytocin gene transcription, translation and processing in the ovary has been reviewed in detail by Wathes and Denning-Kendall (1992). Oxytocin and progesterone are secreted in a pulsatile pattern with 97-100% of all oxytocin pulses associated with pulses of progesterone in the cow; however, only 29% and 86% of all progesterone pulses were associated with oxytocin pulses during the early and mid-luteal phases, respectively (Walters et al., 1984). This suggests that ovarian oxytocin and progesterone do not share a common release mechanism. It is well established that, in vivo, the CL will release oxytocin in response to exogenous PGFj^ (Flint and Sheldrick, 1982; Watkins and Moore, 1987) and small elevations of PGFj^ in ovarian arterial blood elicit such a response (Lamsa et al., 1989). In vivo, very low infusion rates of PGF2„ can selectively release oxytocin, presumably by interacting with the high affinity receptor on the large luteal cell (Lamsa et al., 1989); while much higher infusion rates of PGFj^ (1000 pg/min into the ovarian artery) are required to cause the decline of progesterone secretion in vivo (Schramm et al., 1983). In vivo, desensitization and recovery of high affinity PGFj^ receptors controlling luteal oxytocin release may contribute to

PAGE 46

. 39 the pusatile pattern of PGFja release (Lamsa et al., 1992). These effects of PGFja on the release of oxytocin has been difficult to replicate in vitro using luteal cell cultures. Neither McArdle and Holtorf (1989) using bovine luteal cells, Hirst et al. (1988) using ovine luteal slices, nor Wathes et al. (1988) using perifused bovine luteal explants could demonstrate an acute effect of PGFj^ treatment on oxytocin release. Recently, a different culture system has proved more successful; Miyamoto et al. (1993) obtained acute PG stimulation of oxytocin release in bovine CL. ^ , . Regulation of Uterine Prostaglandin Production The action of oxytocin to increase PGFj^ secretion is mediated by the uterus. As demonstrated in unilaterally hysterectomized heifers, the uterine horn ipsilateral to the CL is necessary to achieve luteolysis (Ginther et al., 1967). Movement of oxytocin from the ovary to uterus is by countercurrent exchange between the ovarian vein and the uterine branch of the ovarian artery. Transfer efficiency, using (i25|)-oxytocin, was estimated to be about 1% (Schramm et al., 1986). Roberts et al. (1976) demonstrated the presence of oxytocin receptors in the uterine endometrium with highest concentrations of receptors present late in the estrous cycle. Heap et al. (1989) reported that in vivo infusion of PGFj^ into ovine uterine lymphatics resulted in oxytocin release from both ovaries, although It was considered unlikely that the infused PGF2„ could reach the contralateral ovary at a sufficient concentration to produce a response. It also was demonstrated that noradrenaline and acetylcholine can trigger luteal oxytocin release when infused in the ovarian artery (Heap et al., 1989.). However, pretreatment with aor p-adrenergic blockers could not inhibit PGFj^-induced

PAGE 47

40 release of oxytocin, indicating that the PGFj^ effect was unlikely due to an adrenergic mechanism (Kotwica et al., 1991). The release of oxytocin from the CL in response to prostaglandin and the stimulation of uterine prostaglandin secretion by oxytocin suggest that a positive feedback system may function between the ovary and the uterus to cause luteolysis. McCracken (1984) proposed a model for regulation of uterine responsiveness to oxytocin and the release of PCFj^ pulses in which steroid receptor dynamics plays a key role. According to the model, a declining endometrial progesterone receptor population permits estradiol from developing follicles to induce oxytocin receptors on the uterine endometrium, making the uterus increasingly responsive to estradiol and oxytocin (Hixon and Flint, 1987). Subsequently, oxytocin released from the pituitary and/or CL binds to newly synthesized endometrial receptors and causes synthesis and pulsatile release of PGF2„. As a consequence, a series of luteolytic pulses of RGFj^ initiates a rapid decline in plasma progesterone and causes luteal regression (Hixon and Flint, 1987). Exogenous oxytocin increases PGFj^ concentrations in uterine veins in the ewe (Milvae and Hansel, 1980) and peripheral PGFM concentrations of sheep (Fairclough et al., 1984) and cow (Lafrance and Goff, 1985). Likewise, exogenous PGFj^ causes release of oxytocin from the CL in the ewe (Flint and Scheldrick, 1982; Watkins and Moore, 1987) and cow (Schallenberger et al., 1984) in vivo and in vitro (Abdelgadir et al., 1987). In the ewe, 66% of detected PGFja surges during the estrous cycle were accompanied by pulses of oxytocin from the ovary (Flint and Scheldrick, 1983). Thus, the release of oxytocin from the CL in response to prostaglandin and the stimulation of prostaglandin secretion from the uterus by oxytocin indicate that a positive feedback system

PAGE 48

41 may function between the ovary and the uterus to cause regression of the CL (Schallenberger et al., 1984). Oxytocin could amplify the luteolytic signal and ensure rapid completion of luteal regression (Flint and Scheldrick, 1983). The pulsatile pattern of PCFj^ secretion in association with luteolysis is common to all large domestic species (Thatcher et al., 1986a). However, the factors that initiate and terminate PGRj^ secretion are not known. Endometrium continuously exposed to oxytocin becomes refractory to further oxytocin challenges. McCracken et al. (1984) proposed that this refractory period was due to down regulation of oxytocin receptors and the pulsatile pattern of PGFja release was a reflection of alternating periods of down-regulation and resynthesis of oxytocin receptors in the endometrium. Poyser (1991) suggested that uterine refractoriness to oxytocin may be related to the ability of PLAj to stimulate AA release from phospholipids. Superfusion of the guinea pig uterus from day 7 of the estrous cycle with AA increased PGFj^ and PGEj production when repeated at 0, 1 , 3 or 5 h intervals. In contrast, superfusion with PLAj at the same intervals failed to stimulate further production. Response to PLA2 was restored after 5 h following initial stimulation. Poyser (1991) suggested that at least 5 h are required to replenish releasable pools of AA after an initial stimulation with PLAj and that this may explain, in part, the pulsatile nature of uterine release of PGFj^. It appears that the activation of the positive feed-back between uterus and ovary is initiated at the level of the uterus since, during spontaneous PGFj^ surges, concentrations of PGFj^ rise in the utero-ovarian veins before oxytocin (Moore et al., 1986). ,c » ' \n ovariectomized ewes, progesterone can induce a frequency of PGFj^ pulses similar to the one observed in intact ewes (Silvia and Raw, 1993) suggesting that ovarian products other than progesterone (such as oxytocin)

PAGE 49

• 42 were not necessary to initiate pulsatile secretion of PGFj^, but could be required to achieve full pulse magnitude. Administration of estradiol during the midluteai phase of the estrous cycle stimulates uterine release of PGFja, increases peripheral concentrations of PGFM and causes CL regression (Thatcher et a!., 1986b; Lafrance and Goff, 1985; Hixon and Flint, 1987). Furthermore, Hughes et al. (1987) demonstrated that destruction of ovarian follicles decreased concentrations of in the utero-ovarian vein and delayed luteolysis. Destruction of follicles also prevented full regression of CL in response to an injection of RGFj^. However, the role of estradiol in regulating secretion of PGFja is not clear. Estradiol appears to have a stimulatory effect on the frequency of neurohypophyseal oxytocin secretion (McCracken et al., 1991). In intact ewes, estradiol treatment increased the frequency of PGFM pulses (Zhang et al., 1991). Yet, destruction of ovarian follicles by X-irradiation had no effect on the frequency or magnitude of PGFM pulses. Silvia and Raw (1993) could not detect any beneficial effect of estradiol on PGFj^ pulse magnitude in ovariectomized ewes. ' The first 10-12 days of progesterone dominance are characterized by an inhibition of endometrial synthesis of estradiol and oxytocin receptors. During the estrous cycle in the cow, endometrial estrogen and progesterone receptors are at their highest levels during the first 10-12 days postestrus and decline to their lowest levels on about day 13 (Meyer et al., 1986). Progesterone downregulates its own receptor after about day 12 of the estrous cycle, and the decrease in endometrial progesterone receptor is followed by an increase in estradiol receptors (between days 14 and 21) and in oxytocin receptors (between days 17 and 21). McCracken et al. (1984) detected an increase in nuclear estradiol receptors in endometrium by 6 h after stopping a 5 day

PAGE 50

43 progesterone infusion in ovariectomized ewes continuously infused with estradiol. Accompanying the rise in estradiol receptors was a parallel increase in membrane bound oxytocin receptors. The concentration of oxytocin receptors rise approximately 500-fold in the uterine endometrium at the time of luteolysis (Roberts et a!., 1976; Scheldrick and Flint, 1985). Premature induction of the oxytocin receptor by administration of estradiol results in premature luteolysis (Hixon and Flint, 1987). However, estradiol treatment in ovariectomized cows did not increase uterine responsiveness to oxytocin (Lafrance and Goff, 1988). Recent results suggest that this increase in oxytocin receptors following the progesterone block may not be an estradiol-mediated event. Vallet et al. (1990) and Lamming et al. (1991) reported that concentrations of uterine oxytocin receptors increased in ovariectomized ewes after 10 days of progesterone treatment. In absence of ovarian hormone replacement, oxytocin receptors are high in ovariectomized ewes probably due to the absence of the inhibitory effect of progesterone. However in such condition, oxytocin is unable to stimulate uterine prostaglandin secretion (Vallet et al., 1990) suggesting that the oxytocin signal transduction system is not functional. When these ewes are treated with progesterone and estradiol, the frequency of PGFj^ pulses is enhanced. Application of these findings to the natural cycle suggest that the duration of the estrous cycle is the result of the period of exposure during which progesterone exerts an inhibitory effect on uterine oxytocin receptor expression. Acute treatments with estradiol in intact ewes, which induced premature luteal regression, also induced an increase in endometrial concentrations of oxytocin receptors (Hixon and Flint, 1987). However, estradiol has little effect on oxytocin receptor concentrations in uterine tissues from ovariectomized ewes previously treated with progesterone (Vallet et al., 1990). The role of estradiol is not well J-

PAGE 51

established but it may enhance oxytocin receptor synthesis and/or oxytocin transduction pathway to increase the frequency of PGFja pulses from sheep endometrium during luteolysis (Flint et al., 1989). Sheldrick and Flick-Smith (1993) showed that estradiol enhances oxytocin-stimulated but not basal PGFj^ secretion from ovine endometrial explants without affecting oxytocin receptor binding activity. The reader is referred to Silvia et al. (1991) for a review of hormonal regulation of PGFj^ secretion during luteolysis in ruminants. Modulation of uterine prostaglandin production by steroids may also involve regulation of PGH synthase, PLAj and/or free intracellular AA content. The amount of PGHS activity in ovine uterine tissue changes during the estrous cycle. Maximum activity was achieved late in the estrous cycle when high levels of PGFja are secreted to induce luteal regression (Huslig et al., 1979). This was due to an increase in PGHS concentration in the tissues and not to an increase in the specific activity of the enzyme. Using an immunohistochemical procedure, Salamonsen and Findlay (1990) have described changes in the concentration of PGHS in uterine tissue throughout the ovine estrous cycle. They found that concentrations were low on days 4 and 17 postestrus and high on days 10, 14, 15 and 16 postestrus; suggesting that levels of PGHS could be controlled by progesterone. Changes in the cellular localization of immunoreactive PGHS were observed: on day 4 of the cycle, PGHS was located primarily in the stromal cells, whereas by day 14 to 16 the enzyme was primarily located in the luminal and glandular epithelial cells (Salamonsen and Findlay, 1990). Administration of progesterone increased amounts of immunoreactive PGHS in the uterine epithelial cell layer of ovariectomized sheep (Raw et al., 1988). Using in situ hybridization, Eggleston et al. (1990) reported that progesterone can induce an increase in the concentration of PGHS mRNA in the endometrium of intact ewes.

PAGE 52

45 Data presented by Salamonsen et al. (1991) suggest that expression of mRNA for PGHS in ovine endometrium appears to be constitutive, estradiol having an overall suppressive role which is overcome in the presence of progesterone. PGHS gene expression can be stimulated by factors other than steroids which include: epidermal grov\/th factor (EOF) in amnion cells (Casey et al., 1988) and osteoblasts (Yokota et a!., 1986); PDGF in mouse 3T3 cells (Lin et al., 1989); interleukin-1 (IL-1) in rabbit chondrocytes (Lyons-Giordano et al., 1993); and tumor necrosis factor (TNF) in human endothelial cells (Jones et al., 1993). The reader is referred to DeWitt (1991) and Smith et al. (1991) for extensive reviews on factors affecting PGHS gene expression. Phospholipase enzymes, notably PLAj^® regulated by steroid hormones. Estradiol has been shown to stimulate PLAj activity in the endometrium of ovariectomized rats in vivo (Dey et al., 1982). In this study, administration of progesterone alone inhibited PLAj activity, but treatment with progesterone followed by estradiol caused a stimulation of PLAj activity above that of estradiol alone. A relationship between PLAj activity and ovarian steroids has also been demonstrated by Downing and Poyser (1983) who showed that, in the guinea-pig, activity of endometrial PLA2 was higher on day 16 of the estrous cycle than on day 7. The increase followed the rise in estradiol concentrations which occurs on day 10. Bonney (1985) demonstrated the presence of a calcium-dependent PLAj in human endometrium and showed that there are critical changes in activity which are dependent on stage of the menstrual cycle. The pattern of PLA2 activity described in this study suggested that estradiol stimulates and progesterone inhibits endometrial PLAj activity. Bonney et al. (1987) demonstrated the presence of two types of PLAj enzymes in human endometrium, one which is active maximally in presence of Ca^* and another

PAGE 53

which is inhibited in presence of Ca^*. Interestingly, the two enzymes appeared to be located at different sites within the endometrium. Stromal tissue was shown to contain predominantly the Ca^*-dependent form of PLA2 while the Ca^*-independent form of PLAj was present mainly in glandular epithelium (Bonney et al., 1987). Pretreatment of human endometrial explants with progesterone followed by incubation with estradiol caused a two-fold stimulation of the Ca^*-dependent PLAj activity but not of the Ca^*-independent activity. Several studies, using protein synthesis inhibitors in the guinea-pig, indicated that stimulation of endometrial POFj^ synthesis by estradiol acting on a progesterone-primed uterus is dependent upon increased endometrial protein synthesis (Poyser and Riley, 1987; Riley and Poyser, 1989; Leckie and Poyser, 1993). Similarly, ongoing protein synthesis appeared to be essential for both basal and oxytocin-stimulated PGFj^ secretion from bovine endometrial explants (Lafrance and Goff, 1990). It was hypothesized that estradiol, acting on a progesterone-primed uterus, may 'switch on' PGFj,, production by stimulating the synthesis of a protein that activates PLAj. A PLAj-stimulating protein, which is antigenically and functionally related to the PLAj-stimulating melittin (a bee venom peptide), has been isolated from various cell lines and has been described as an 'intracellular messenger' in the stimulation of prostaglandin synthesis (Clark et al., 1987; Clark et al., 1988). Johnson and Poyser (1991) showed that melittin stimulates PGFj^ synthesis by the guinea-pig uterus. Recently, Leckie and Poyser (1993) were unsuccessful in isolating endometrial proteins from guinea-pig at day 15 of the cycle that would stimulate PLAj activity and prostaglandin synthesis. The Ca^* ionophore A23187 has been shown to increase the outputs of PGFj^ and PGEj from superfused endometrium (Poyser and Brydon, 1983; Poyser, 1987). The guinea-pig uterus showed complete

PAGE 54

refractoriness to PLAj, partial refractoriness to A23187, but no refractoriness to exogenous AA with regards to the stimulation of prostaglandin when the same treatment is re-applied after an interval of 1 h (Poyser, 1991). This refractoriness lasted for 3-5 h, and it was suggested that there may be one or several pools of bound AA which were readily releasable but which take 3-5 h to be replenished. Poyser and Ferguson (1993) hypothesized that if PLAj ai^d A23187 were producing refractoriness by the same mechanism (slow refilling of bound AA pool), then cross refractoriness between the two compounds should occur. No cross refractoriness was observed indicating that the intracellular processes involved in the stimulation of uterine prostaglandin secretion by PLAj and A231 87 were different (Poyser and Ferguson, 1993). Progesterone may also regulate prostaglandin production by enhancing cellular content of AA. CL development is paralleled by increased endometrial content of lipid droplets which attain maximal concentrations on days 14 to 15 of the cycle (Boshier et al., 1987). In ovariectomized ewes, long term exposure to progesterone increased endometrial content in lipids (Brinsfield and Hawk, 1973). Addition of estradiol to progesterone-treated ewes decreased endometrial lipid content compared to progesterone-treated controls, suggesting that estradiol inhibited progesterone-dependent lipid accumulation or mobilized lipids faster than their accumulation could be stimulated by progesterone (Brinsfield and Hawk, 1973). McCracken et al. (1981) originally postulated that, in the ewe, cAMP was responsible for the oxytocin stimulation of AA release. However, recent studies suggest that oxytocin-stimulated production of prostaglandin involves the phosphoinositide signal transduction system in ovine (Flint et al., 1986; Hixon and Flint, 1987; Mirando et al., 1990a) and bovine (Mirando et al., 1990b;

PAGE 55

... 48 Mirando et al., 1993a) endometrium. Binding of oxytocin to its endometrial membrane receptor stimulates PLC that can cleave PIPj into DAG and IP3 (Flint et a!., 1986; Silvia and Homanics, 1988). Agonists of PLC-generated second messengers have been used to replicate the effect of oxytocin on prostaglandin production. PMA (DAG agonist) and A23187 stimulated production of PGF2„ from ovine endometrial tissue in vitro (Silvia et Homanics, 1988; Raw and Silvia, 1991). In bovine endometrial explants from heifers at day 19 or 20 of the estrous cycle, PMA induced PGFj^ release to the same extent as oxytocin while PDD, a phorbol ester with little PKC-stimulating activity, did not stimulate prostaglandin release (Lafrance and Goff, 1990). The effect of PMA would appear to be specific for PKC activation rather than a non specific effect such as alteration of the plasma membrane. Neither basal nor oxytocin-stimulated PGFj „ release were affected by A23187, dibutyryl cAMP, dibutyryl cGMP or a PKA inhibitor (Lafrance and Goff, 1990). However activation of PKC does not stimulate prostaglandin release from guinea pig endometrium (Poyser, 1987). This is interesting considering that oxytocin does not appear to be involved in PGF20 release during the estrous cycle of the guinea pig (Poyser and Bryon, 1983). Mirando et al. (1990a) showed that the ability of oxytocin to stimulate PLC activity is much greater at day 16 than day 12 postestrus. Such changes in the ability of oxytocin to stimulate PLC, or in the ability of the second messengers to activate PGFj^ synthesis, could account for the changes in uterine responsiveness to oxytocin observed during the estrous cycle (Silvia and Raw, 1993). Formation of DAG, following PLC activation, was suggested to serve as a substrate to monoand diacylglycerol lipases which could cleave AA from DAG. It was suggested that liberation of AA from DAG resulted in increased production of PGFj^, (Flint et al., 1986). It appears that this source of

PAGE 56

AA is not essential to PGFja production since PMA, which activates PKC but cannot be a source of AA, induced PGFj^ secretion to the same extent as oxytocin (Lafrance and Goff, 1990). The conceptus could prevent luteolysis in early pregnancy by production of a secretory product to maintain progesterone action on the endometrium (McCracken et al., 1984). If the progesterone receptors are maintained, nuclear estradiol receptors could be suppressed and thus terminate the chain of events leading to luteolysis. The conceptus could also produce a product that directly inhibits formation of endometrial estradiol or oxytocin receptors. Maternal Recognition of Pregnancy In order for pregnancy to be maintained in ruminants, regression of the CL must be prevented since ovarian progesterone is necessary for maintenance of pregnancy. Therefore, the luteolytic influence of the uterus must be neutralized in some manner. The signal for maintenance of the CL must come from the conceptus and could be luteotrophic (increase the rate of production and/or release of progesterone from the CL), antiluteolytic, or some combination. Antiluteolytic effects include those that are anti-PGF2„ in nature (alter the dynamics of secretion of PGFjJ and those that are luteoprotective (directly protect the CL) (Rothchild, 1981; Thatcher et al., 1986a). The processes by which the conceptus signals its presence to the maternal unit, resulting in CL maintenance and other events, have been termed maternal recognition of pregnancy (Short, 1969). Studies in cattle and sheep have determined the critical time by which the conceptus must signal its presence to the maternal uterus. Embryo removal

PAGE 57

50 experiments have shown this. The critical period for conceptus-induced maintenance of the CL appeared to occur on day 12 in sheep (Moor and Rowson, 1966) and between days 15 and 17 postestrus in cattle (Northey and French, 1980; Humblot and Dalla Porta, 1984). This led to a search for the embryonic "signals" responsible for the alteration of the uterine luteolytic mechanism resulting in CL maintenance and continuation of progesterone secretion. Intrauterine infusion of conceptus extracts or homogenates extended luteal lifespan in cattle (Northey and French, 1980; Humblot and Dalla Porta, 1984) and sheep (Ellinwood et a!., 1979; Martal et al., 1979). The active component in ovine conceptus homogenates was reported to be proteasesensitive, heat-labile, absent beyond day 21 of gestation (Martal et al., 1979) and to contain neither LH/hCG-like nor prolactin like activity (Ellinwood et al., 1979). Intrauterine administration and reciprocal interspecies (ovine-bovine) transfer of trophoblastic vesicles extended recipients CL lifespan indicating that the active proteinaceous component was secreted by trophoblastic tissues and that it was similar or closely related in these two species (Heyman et al., 1984; Martal et al., 1984). The total array of secreted proteins by day 16-18 bovine or day 15-16 ovine conceptuses was referred to as bovine or ovine conceptus secretory proteins (bCSP or oCSP). Infusion of CSP into the uterine lumen extended CL lifespan and lowered release of PGFM in response to luteolytic doses of estradiol in the cow (Knickerbocker et al., 1986a, 1986b) and the ewe (Godkin et al., 1984b; Fincher et al., 1986). These results suggested that bovine and ovine conceptuses secreted a proteinaceous factor of trophectoderm origin with antiluteolytic-anti-PGF2„ activity.

PAGE 58

The active component found in ovine conceptus homogenates was first named "trophoblastin" by Martal et al. (1979) and later as "ovine trophoblast protein-1" (oTP-1) since this protein was the major secreted protein from conceptus trophectoderm (Godkin et a!., 1984a). oTP-1 has a molecular weight of about 18 kDa and isoelectric points (pi) ranging from 5.3 to 5.7 (Godkin et al., 1982; 1984b; Hansen et al., 1985) and is not glycosylated (Anthony et al., 1988). Shortly after, a low molecular weight (22-26 kDa), acidic (pi = 6.6.-5.6) protein complex (Bartol et a!., 1985a) secreted by bovine conceptuses (days 16-24) was shown to cross-react with an antibody raised against oTP-1 and was called bovine trophoblast protein-1 (bTP-1) (Helmer et al., 1987). The bTP-1 complex consists of at least 7-9 isomers grouped in three molecular weight classes (21 , 23 and 25 kDa) (Bartol et al., 1985a; Helmer et al., 1987; 1989a; Anthony et al., 1988; Plante et al., 1990). This variation in molecular weight and isoelectric points is due to differential N-linked glycosylation (Helmer al., 1988) and to translation of various bTP-1 gene transcripts (Anthony et al., 1988). Trophoblast protein-1 could be identified in uterine fluids from day 15 to 25 in pregnant cows, and from day 14 to 22 in pregnant ewes (Kazemi et al., 1988). Secretion of oTP-1 increased as conceptus morphology changed from spherical (312 ng/uterine flushing) to tubular (1380 ng) to filamentous (4450 ng) on days 12-13 (Nephew et al., 1991). Presence of bTP-1 mRNA could be detected as early as day 11 by in situ hybridization (Farin et al., 1990; Guillemot et al., 1990), or day 8 by reverse transcription-polymerase chain reaction (RTPCR) (Hernandez-Ledezma et al., 1992). Immunochemical studies have localized bTP-1 in the cytoplasm of monoand binudear cells of the trophectoderm (Lifsey et al., 1989). The increase in trophoblast protein-1 production coincides with a dramatic elongation of the conceptus (Geisert et al..

PAGE 59

.52 1988; Nephew et al., 1989). High rates of oTP-1 synthesis occur between day 13 and 21 of pregnancy, with production increasing over one hundred times from day 12 to day 15-16 when total synthesis is maximum (Hansen et al., 1985; Farin et al., 1990; Roberts et al., 1992). Ott et al. (1989) described a second period of oTP-1 secretion by chorion between days 25 and 45 of pregnancy. Bovine TP-1 displays a similar pattern of expression with synthesis peaking between days 17 and 19 (Farin et al., 1990). Expression of oTP-1 was absent in the trophoblastic regions showing cellular contacts with the uterine epithelium (Guillemot et al., 1990). Interestingly, c-fos mRNA expression was shown to follow the same expression pattern in trophoblastic cells as oTP-1 (Xavier et al., 1991). N-terminal sequencing of purified oTP-1 (Stewart et al., 1987; Charpigny et al., 1988) and molecular cloning of the cDNA for oTP-1 (Imakawa et al., 1987; 1989; Stewart et al., 1989b; Charlier et al., 1989; 1991; Klemann et al., 1989; 1990) and later bTP-1 (Imakawa et al., 1989; Stewart et al., 1990; Hansen et al., 1991) revealed that both trophoblast proteins were related to the type I interferon (IFN) family. Three subtypes of type I interferons have been described: IFNa, IFNp and IFNco which differ considerably in amino acid sequence but share a common membrane receptor (see Roberts et al., 1992). The various isoforms encoded by multiple genes within each one of these subtypes can present significant differences in amino acid sequence and biological activity . In particular, the IFNa subtype represents a group of 165166 amino acid proteins and contain 10-12 subtypes, whereas the IFN© (formerly known as IFNan) subtype represents a group of 172-174 amino acid proteins and contain 15-20 subtypes (Capon et al., 1985; ). Bovine TP-1 and oTP-1 mRNA code for polypeptides of 195 amino acid, including a 23 animo

PAGE 60

53 acid signal peptide. The length of mature oTP-1 and bTP-1 (172 amino acids) and their high sequence homology (70%) with bovine IFNco (Imakawa et al., 1989) suggested that they belonged to the IFNco subtype. However, bTP-1 and oTP-1 resemble each other in both amino acid and nucleotide sequence more than other IFN© do within their own subtype. Their high sequence homology to each other and a number of additional unique features such as poor virus inducibility led Roberts et al. (1992) to recommended the creation a new structural subtype of type I interferons: interferon (IFN) tau (x). This designation was sanctioned by the International Committee on the Interferon Nomenclature (Roberts, 1993). Genes coding for IFNx have been identified in other related ruminants such as goat, and musk ox, but not in horse, pig, mouse, rabbit and human (Leaman et al., 1992). The high level of homology across ruminant species and presence of IFNx within a limited subset of mammals suggests that IFNx derived from IFN© about 30-50 million years ago (Roberts et al., 1992). IFNx gene expression is ephemeral, intensive and specific to the trophectoderm during the preimplantation period. These characteristics generated a number of studies designed to better understand the unique features of IFNx transcriptional regulation. Very low levels of IFNx mRNA has been detected in bovine leukocytes exposed to Sendai virus (Cross and Roberts, 1991). IFNx expression was not observed when a bovine IFNx (bIFNx) gene or promoter constructs coupled to a reporter gene was transfected into non-trophoblastic cells. However IFNx gene was constitutively expressed when transfected into trophoblast tumor derived cells (Cross and Roberts, 1991). A 450 base pair region of the IFNx gene promoter was necessary to obtain maximal constitutive expression in these cells, whereas an IFN© gene promoter was inactive. Ko et al. (1991) demonstrated a good correlation of uterine luminal '• %i:f ' ' '' * '\ ' ' f' " '

PAGE 61

54 insulin-like growth factors (IGFs) with olFNxproduction, and observed a beneficial effect of culturing blastocysts in presence of IGFs with respect to maintenance of olFNx secretion. Recently, Imakawa et al. (1993) provided evidence that olFNx mRNA and protein production was stimulated by granulocyte macrophage-colony stimulating factor (GM-CSF), and that GM-CSF mRNA was found in the endometrial epithelium. They proposed that uterine GM-CSF was in part responsible for the large increase in olFNx secretion observed at the time of maternal recognition of pregnancy. The reader is referred to Roberts et al. (1991), Roberts et al. (1992) and Leaman et al. (1992) for reviews on the molecular biology of IFNx. High affinity (Kd=0.1x10-io to 0.4x10-1° m) and low affinity (K^^IO-io M) receptors for ovine IFNx (olFNx) were reported in pregnant and cyclic sheep endometrium (Godkin et al., 1984a; Hansen et al., 1989; Knickerbocker and Niswender, 1989). On days 8 and 12, the number of unoccupied olFNx receptors was similar between cyclic and pregnant ewes, but diminished following day 12 in pregnant ewes (Knickerbocker and Niswender, 1989). The ability of olFNx to displaced radiolabeiled human IFNa from its receptors was indicative of olFNx affiliation with type I interferons (Stewart et al., 1987). Cross-linking studies indicated that olFNx was associated intimately with two proteins when bound to its receptor (Hansen et al., 1989). These two proteins (70 and 100 kDa) are not specific to endometrial tissue since olFNx cross-linked similar proteins in spleen membranes (Hansen et al., 1989); however they could be part of a high affinity, multiple subunit complex responsible for mediation of IFNx biological effects.

PAGE 62

55 Trophoblast Interferons T . , 1 . " ' ' Antiluteolvtic Activity i In vivo intrauterine infusion of olFNx into cyclic ewes from day 12 to 20 (Godkin at al., 1984b) or from day 11 to 15 (Vallet et al., 1988) extended CL lifespan. Similarly, infusion of bIFNx into the uterine lumen of cyclic cows from days 14 to 17 prolonged CL lifespan and inhibited PGF2„ release into the posterior vena cava (Helmer et al., 1989b). Because of the expense of purifying large amounts of "natural" IFNx from conceptuses, the availability of recombinant interferon bovine IFNa, (50% amino acid sequence homology with bIFNx and olFNx), and later recombinant ovine and bovine IFNx became a key factor in the study of the antiluteolytic properties of type I interferons. Intrauterine infusion of milligram quantities of recombinant bIFNa, into cyclic ewes (Stewart et al., 1989a; Parkinson et al., 1992a), or intrauterine or intramuscular injections of this protein into cyclic cows (Plante et al., 1988, 1989, 1991; Newton et al., 1990) during the time of maternal recognition of pregnancy extended luteal lifespan of the treated animals. Nevertheless, these studies demonstrated that considerably more recombinant bIFNa, had to be introduced into the uterus to achieve an extension of interestrous interval than if natural IFNx had been used instead. For example, daily (days 9 to 19) intrauterine administration of 2,000 n g of recombinant blFNa, were necessary to extend interestrous intervals of ewes (Parkinson et al., 1992a), compared to a requirement of only 100 [ig olFNx injected daily (days 12 to 15) to extend interestrous intervals of ewes to greater than 19 days (Vallet et al., 1988). Subsequently, recombinant olFNx became available (Martal et al., 1990; Ott et al., 1991) and was tested for its antiluteolytic activity. Intrauterine injection of recombinant olFNx in cyclic ewes between days

PAGE 63

-tm--' ^ f V \i -J* ^fr^ « ' s ' ' ' . ' ' "' ' m 11 and 15 (Martal et al., 1990), and in cyclic cows between days 14 and 24 (Meyer et al., 1992), extended CL lifespan and interestrous interval . Fertility studies have been conducted using recombinant bIFNa, to improve pregnancy rate in sheep and cattle. The rational for these studies was to supplement inseminated animals with interferon during the period of maternal recognition of pregnancy in order to "rescue" conceptuses lagging in their development and therefore unlikely to produce sufficient amounts of IFNx to prevent luteolysis. Although recombinant olFNx and bIFNx (Klemann et al., 1990) were produced and physiologically active, the amounts necessary to conduct fertility trials were not available. Encouraging results were obtained in sheep where conception rates were improved significantly with two daily intramuscular injections (2 mg) of recombinant blFNa, (Nephew et al., 1990; Schalue-Francis et al., 1991; Martinod et al., 1991). Francis et al. (1992) verified that intramuscularly injected recombinant blFNa, could reach the uterus and induce synthesis of a pregnancy-specific protein (p70) by endometrium in cyclic ewes. In contrast, there have been no reports of improved conception rates in cattle using recombinant blFNa,. In fact, administration of blFNa, in cows, regardless of the injection pattern, decreased conception rates by about 10% (Barros et al., 1992b). This unexpected effect was attributed to hyperthermia and acute lowering of serum progesterone concentrations induced by recombinant blFNa, (Barros et al., 1992a). Regulation of Endometrial Prostaglandin Svnthesis During early pregnancy, pulsatile secretion of PGFj,, is suppressed in sheep (Thornburn et al., 1973; McCracken et al., 1984; Zarco et al., 1988) and cattle (Kindahl et al., 1976; Thatcher et al., 1984a; Basu and Kindahl, 1987a), as 1

PAGE 64

57 is the ability of oxytocin to stimulate uterine secretion of PGF2„ (Fairclough et al., 1984; Silvia et al., 1992; Lafrance and Goff, 1985). The induction of PGFM pulses and luteolysis in response to estradiol administration is absent in pregnant cows (Rico et al., 1981; Thatcher et al., 1984a) or ewes (Kittok and Britt, 1977; Lacroix and Kann, 1986; Fincher et al., 1986) as compared to cyclic animals at the same day postestrus. In vitro studies characterized endometrial prostaglandin secretion during early pregnancy. Endometrial explants (Thatcher et al., 1984b) or perifused endometrial tissue (Gross et al., 1988b) from cows at day 17 postestrus synthesized less PGFj^ than those of cyclic cows. The absence of oxytocin responsiveness observed in vivo and in vitro in pregnant cattle and sheep is directly related to the fact that endometrial oxytocin receptors are decreased considerably in pregnant compared to cyclic ewes (McCracken et al., 1984; Flint et al., 1986) and cows (Fuchs et al., 1990; Jenner et al., 1991) during the luteolytic period. The antiluteolytic role of IFNt observed in vivo is supported by a number of in vitro studies examining the effects of trophoblast interferon on endometrial secretion of prostaglandins. PGFj^ secretion was decreased when endometrial explants from cows at day 17 postestrus were incubated for 24 h in presence of bCSP (Gross et al., 1988b) or bovine IFNx (Helmer et al., 1989b). Vallet et al. (1989) reported lower PGFj^ secretion by perifused endometrium when collected from ewes that received intrauterine infusions of ovine IFNx for 3 days. Culture of highly purified endometrial epithelial and stromal cells have demonstrated marked differences in prostaglandin release by the two cell types. Epithelial cells were the major source of PGFj^, and PGEj was primarily secreted by stromal cells from sheep (Cherny and Findlay, 1990; Charpigny et al., 1991) and cattle (Fortler et al., 1988). Salamonsen et al. (1988) reported an inhibitory

PAGE 65

effect of ovine IFNx on release of both PGFj^ and PGEj from mixed ovine endometrial cells in primary culture. Using highly purified cultures of endometrial epithelial and stromal cells from cyclic sheep at day 15 postestrus, Charpigny et al. (1991) observed that ovine IFNx inhibited basal and oxytocininduced secretion of PGFj^ and PGEj in both cell types. Regulation of Endometrial Steroid and Oxytocin Receptors The point(s) in the intracellular cascade regulating uterine secretion of PGFja that is affected by the conceptus has not been identified. Bazer (1992) hypothesized that IFNx antiluteolytic effect could be mediated through several potential mechanisms: (1) stabilization or up-regulation of endometrial progesterone receptors to maintain the progesterone block and prevent synthesis of oxytocin receptors or up-regulation of estrogen receptors; (2) direct inhibition of estrogen receptors which would attenuate oxytocin receptor synthesis and pulsatile release of PGFj^ ; (3) direct inhibition of oxytocin receptor synthesis; (4) alteration of post-oxytocin receptor mechanisms to inhibit oxytocin-induced release of PGFj^jj. i During maternal recognition of pregnancy in sheep, endometrial progesterone receptor (PR) concentrations are maintained at concentrations characteristic of mid-luteal phase, whereas in non-pregnant animals, PR are high during mid-diestrus and decrease during late diestrus (Findlay et al., 1982; Cherny et al., 1991; Ott et al., 1993). Endometrial estrogen receptor (ER) concentrations decreased from day 10 to 16 postestrus in pregnant ewes, whereas in cyclic ewes, ER remained relatively constant from day 10 to 14 and then increased dramatically at day 16 postestrus (Ott et al., 1993). In vivo, progesterone was necessary for olFNx to exert an inhibitory effect on oxytocin-

PAGE 66

59 stimulated production of prostaglandin in the ewe (Ott et al., 1992). The antiluteolytic effects of IFNx could be mediated through direct transcriptional regulation of uterine progesterone and/or estrogen receptor genes. Neither ovine nor bovine progesterone and estrogen receptor genes have been cloned and therefore no interferon stimulated response elements (ISRE) have been described in the regulatory regions of these genes. Structures of the human estrogen receptor (Ponglikitmongkol et al., 1988), and human (Kastner et al., 1990) and rabbit (Milgrom et al., 1988) progesterone receptors have been described. Analysis of the 5' flanking regions of these genes revealed only putative but no functional ISREs. Intrauterine administration of oCSP (containing 100 ^ig of olFNx/day) from day 11 to 15 postestrus could not totally replicate the PR and ER protein and mRNA concentrations observed in pregnant ewes at day 16 postestrus. Ovine CSP appeared to have a negative effect on PR protein and mRNA concentrations compared to serum albumin treated ewes at day 16 postestrus (Mirando et al., 1993b). In vivo, oxytocin was unable to stimulate uterine PGFj^ secretion when administered to ewes on days 12 to 16 of pregnancy (Fairclough et al., 1984; McCracken et al., 1984; Silvia et al., 1992). This lack of oxytocin responsiveness was paralleled by an almost total suppression of oxytocin's ability to stimulate PLC activity in pregnant ewes (Mirando et al., 1990a; 1990b; Ott et al., 1993). However, oxytocin was equally effective in stimulating release of PGFj^ from endometrial tissue collected from pregnant and non pregnant ewes on days 14 and 16 postestrus (Vailet et al., 1989; Silvia and Raw, 1993). The oxytocin-induced PGFj^ secretion by endometrial tissue from pregnant ewes appears to be mediated through a pathway other than PLC-induced release of phophoinositides. Silvia et al. (1992) suggested that the disparity between in i i

PAGE 67

wVo and in vitro models with regards to oxytocin responsiveness of endometrium from pregnant ewes could be due to a requirement for continuous presence of the conceptus to maintain oxytocin non-responsiveness. Incubation of uterine tissue in the absence of the conceptus for just 2.5 h appeared to be sufficient for the suppressive influence of the conceptus to wear off; implying that a very labile factor could be involved in mediating the conceptus effect on oxytocin responsiveness. | Acute in vitro exposure of endometrial tissue to olFNx (Vallet and Bazer, 1989) had very little effect on the ability of oxytocin to stimulate activity of PLC. Vallet et al. (1989) showed that oIFNt did not compete with oxytocin for its receptor. Interestingly, oxytocin effect on prostaglandin secretion by endometrial cells in primary culture was inhibited only if cells were exposed to ovine IFNx before (24 h pretreatment) and during the oxytocin challenge (Charpigny et al., 1991). When cyclic ewes were exposed to olFNx from days 12 to 14 postestrus, PGF2„ secretion was reduced (Vallet et al., 1988; Mirando et al., 1990a) as well as phosphoinositide metabolism (Mirando et al., 1990a; 1990b) indicating a reduced ability of oxytocin to stimulate PLC activity. Intrauterine injections of olFNx from day 11 to 15 decreased endometrial number and affinity of oxytocin receptors (Mirando et al.. 1993b) possibly by down-regulating endometrial ER and/or stabilizing endometrial PR. , Endometria l Prostaglandin Svnthesis Inhibitor The regulation of endometrial steroid receptors is not the only possible mechanism to explain the absence of oxytocin-responsiveness and pulsatile release of PGF^^ in endometrium from pregnant cows and ewes. No differences in detectable PGHS enzyme (Salamonsen and Findlay, 1990) or mRNA

PAGE 68

61 (Salamonsen et al., 1991) were found in endometrial tissue from pregnant and cyclic ewes on day 15 after estrus. Consequently, the conceptus does not appear to suppress PGFj^ secretion by reducing the concentration of PGHS in this tissue. Another possibility is that the conceptus induces an intracellular endometrial factor which inhibits enzymes involved in prostaglandin synthesis. Endogenous inhibitors of prostaglandin synthesis have been reported in bovine follicular fluid (Shemesh, 1979), human amniotic fluid (Saeed et al., 1982) and human serum and plasma (Saeed et al., 1977). A heat-labile, inhibitor of prostaglandin synthesis was described in fetal cotyledon tissue (Shemesh et al., 1984b) and maternal caruncle tissue (Shemesh et al., 1984a) of bovine placentomes. Inhibitory activity was present in caruncle extract from days 120150 of gestation, but was not detectable at term (260-280 days) (Shemesh et a!., 1984a). The exposure of bovine luteal cells to the active placental extract was sufficient to suppress their prostaglandin synthesis (Shemesh et al., 1984a). The conversion of (i^C)-AA by endometrial microsomes and cytosol of cyclic and pregnant cows has been examined to investigate their ability to synthesize prostaglandins (Wlodawer et al., 1976; Basu and Kindahl, 1987b). Wlodawer et al. (1976) showed that bovine microsomes had a relatively low capacity to convert AA into PGF^^. However, when PGHj was substrate for the microsomal enzymes, the conversion into PGF2„ was much higher. This observation suggested that PGHS activity was a limiting factor in the ability of endometrial microsomes to synthesize PGFj,,, possibly due to the presence of a PGHS inhibitor. In a comparative study, Basu and Kindahl (1987b) described the presence of a potent inhibitory factor that controls prostaglandin biosynthesis in bovine endometrial tissue by acting on PGHS activity. The inhibitory capacity

PAGE 69

62 was highest in microsomes and much higher in pregnant than nonpregnant animals. When comparing the potency of this inhibitory factor, as calculated by IC50 values, the potency was about 8 times higher for day 17 pregnant than day 17 cyclic endometrial cytosol (Basu and Kindahl, 1987b). This inhibitory factor was not destroyed when boiled and was present in both uterine horns suggesting that a humoral or local factor from the conceptus could have influenced the non pregnant horn (Basu and Kindahl, 1987b). Gross et al. (1988b) examined the ability of endometrial cytosol and microsomes, from day 17 cyclic and pregnant cows, to modulate prostaglandin synthesis by cotyledonary microsomes from parturient cows. The endometrial prostaglandin synthesis inhibitor (EPSI) appeared to be present in much higher amounts in pregnant cows than cyclic cows, to be primarily cytosolic, proteinaceous, and to act as a non-competitive inhibitor with regard to AA metabolism (Gross et al., 1988b). Bovine IFNx induced EPSI activity when administered to endometrial explants from cyclic cows at day 17 postestrus (Helmeretal., 1989a). | Antiviral. A ntiproliferative and Immunosuppressive Activities In common with other type I interferons, IFNx has a high antiviral activity of 2 to 3 X 10^ lU/mg of protein (Pontzer et al., 1988) and can protect type I receptor-bearing cells from lysis induced by a variety of viruses (see Roberts et al., 1992). Among the enzymes induced by IFNx binding to its receptor are 2',5'oligoadenylate synthetase (2-5A synthetase), protein kinase (p68) and endoribonuclease (Jacobsen et al., 1988). In the presence of viral double stranded RNA and ATP, the protein kinase phosphorylates and inactivates an initiation factor (elF-2) necessary for the synthesis of viral protein (Esteban and

PAGE 70

., , • ... r . , , , ^ ^ 63 Paez, 1985). Ovine and bovine IFNx, as well as bovine interferon a^^, induced synthesis of 2-5A synthetase in endometrial tissue (Barros et al., 1991; Mirando et al., 1991; Short et al., 1991). Three isoforms of 2-5A synthetase have been reported (see Hovanessian, 1991 ) which catalyze the synthesis of an activator of a ribonuclease involved in degradation of viral RNAs. It appears that PKC is not essential for the antiviral activity of a interferons (Cernescu et al., 1989). The antiviral activity of IFNx may play an important role in protection of the periimplantation conceptus when exposed to viral infection. Interferons a are known to inhibit proliferation of many cell types (see Clemens and McNurlan, 1985; Salzberg et al., 1990). The IFNx was a potent inhibitor of (3H)-thymidine incorporation into lymphocytes after exposure to mitogens (Newton et al., 1989; Fillion et al., 1991; Skopets et al., 1992). Ovine IFNx exhibited a potent anticellular activity across species and appeared to decrease the rate of progression through the cell cycle (transition from G0/G1 into and through S phase) (Pontzer et al., 1991). However . considering the enormous amounts secreted by the conceptus at the time of maternal recognition of pregnancy, it is remarkable that IFNx lacked the cytotoxic effects observed with high doses of other types of interferons (Pontzer et al., 1 991 ). 5 Regulation of Endometrial Protein Synthesis Trophoblast IFN stimulated the synthesis and secretion of specific proteins by endometrial explants from day 12 cyclic sheep but did not induce changes in the type of proteins secreted (Godkin et al., 1984a). Quantitative studies revealed that olFNx stimulated synthesis of four or five types of proteins, and decreased the synthesis of another two (Vallet et al., 1987; Salamonsen et al., 1988). The pattern of proteins synthesized by olFNx-treated endometrium

PAGE 71

64 was similar to that of endometrium on day 1 3 of pregnancy and to endometrium from cyclic ewes treated with day 15 oCSP or human IFNa (Salamonsen et al., 1988). One of the major stimulated proteins has a molecular weight of 70 kDa and pla4. The same protein was identified in sheep endometrium between days 14 and 20 of pregnancy (Sharif et al., 1989), and could be induced in endometrium from cyclic sheep by in vivo administration of recombinant bovine IFNttjl (Francis et a!., 1991). In bovine endometrial explants from bCSP-treated cyclic cows or from pregnant cows, proteins in the 14 kDa molecular weight range were upregulated (Geisert et al., 1988; Gross et al.. 1988c; Helmer et al., 1989a) compared to endometrium from untreated cyclic cows. Rueda et al. (1993) described two recombinant blFNx-stimulated proteins (12 kDa, pi > 7.5 and 28 kDa, pi 4.5-5.5) in secretion from bovine endometrial explants; these two proteins have not been identified. ' The 70 kDa "pregnancy-specific" protein has not been identified, but another lower molecular weight olFNx-induced protein has been isolated and the N-terminal amino acid sequence determined (Vallet et al., 1991). The N-terminal sequence of this protein had a 40-55% sequence homology with the N-terminus of p2-microglobulin. Stewart et al. (1992) presented evidence that expression of 32-microglobulin and major histocompatibility complex (MHC) class I were correlated with expression of olFNx, suggesting that the trophoblast interferon induced expression of these proteins. A recent report described that, in neuroblastoma cells, IFNa increased MHC class I and Ps-microglobulin gene expression by inducing factor binding to ISREs present in both genes (Drew et al., 1993). I

PAGE 72

65 Signal Transduction System Using biochemical and electron microscopic techniques, it was shown that, following binding, the IFNa-receptor complex was rapidly internalized via receptor-mediated endocytosis (see Langer and Pestka, 1988; Mogensen et al., 1989; Grossberg et al., 1989). Binding of IFNa to type I receptors is a necessary but not sufficient condition for cell activation. Through analysis of mutants unresponsive to IFNa, it was determined that lack of receptors was only one reason for lack of biological response (Dron et al., 1986). A rapid and transient activation of PLAj and release of AA was noted following treatment of 3T3 fibroblasts with IFNa, and inhibition of PLAj activity specifically blocked the binding of nuclear factors to the 2-5A synthetase gene ISRE (Hannigan and Williams, 1991). A consequence of PLAj activation is the release of AA from membrane phospholipids. Free AA can be metabolized through three pathways: cyclooxygenase (PGHS), lipoxygenase and epoxygenase. Perturbation of AA metabolism, using inhibitors of cyclooxygenase and lipoxygenase, enhanced the IFNa stimulation of factor binding to the ISRE (Hannigan and Williams, 1991). The potentiation of IFNa effect observed, when AA metabolism was redirected to epoxy-derivatives. was not restricted to nuclear factor binding but also was evident at the levels of 2-5A synthetase mRNA and protein (Williams, 1991). While AA metabolism seems to play an important role in the signal transduction of IFNa, AA itself is not the second messenger. Addition of exogenous AA did not affect the transcription of IFN-stimulated genes (Yan et al., 1989). AA metabolism has been implicated in signal transduction of other systems such as induction of tumor necrosis factor gene expression by phorbol esters (Horiguchi et al., 1989).

PAGE 73

66 Several reports have indicated the importance of protein phosphorylation and possibly a role for PKC in IFNa action (Yap et al., 1986; Tiwari at al., 1988; Reich and Pfeffer, 1990; Decker et al., 1991). Pfeffer et al. (1990) proposed that the transient increase in DAG, resulting from phosphatidylcholine hydrolysis, may induce selectively translocation and activation of the p isoform of PKC. Interferon a was shown to induce phosphatidylcholine hydrolysis, but not inositol phospholipid turnover, therefore providing a source of DAG without increasing the intracellular free Ca^* concentration (Rosoff et al., 1988). However, agonists of PKC activity did not mimic IFNa action and PKC down-regulation did not alter the transcriptional response to IFNa (Reich and Pfeffer, 1990; Pfeffer et al., 1990). These observations are consistent with the potential role of AA metabolites in the signal transduction of IFNa. Recently, direct evidence for the involvement of a protein kinase distinct from PKC has been presented (Kessier and Levy, 1991). This protein kinase was discovered by functional complementation of mutant cells unresponsive to IFNa. The only message encoded within the complementing cosmid was found to be tyk2, a non-receptor protein tyrosine kinase (Firmbach-Kraft et al., 1990). Since, JAK-1 and JAK-2 have been added to this new family of protein kinases, which is not related to the src family of tyrosine kinases (Hunter, 1991). Velazquez et al. (1992) speculated that tyk protein kinase could be activated by interaction of its large extracatalytic domain with IFNa receptors and, subsequently phosphorylate interferon-stimulated gene factor 3 (ISGF3). Among ISRE-binding proteins, ISGF3 has been shown to be the primary transcriptional activator for interferon a-induced genes and to recognize a 15nucleotide c/s-acting DNA element (ISRE) located in the promoter region of IFNa -stimulated genes. ISGF3 is the only ISRE-binding factor rapidly induced by

PAGE 74

67 interferon a and its activity profile parallels the process of receptor binding and transcription activation (Dale et al., 1989; Bandyopadhyay et al., 1990). It was shown that an ISGF3-like activity was sensitive to inhibition of PLAj 'n mouse fibroblasts (Hannigan and Williams, 1991). ISGF3 is a complex shown to be composed of four (113, 91, 84, and 48 kDa) distinct proteins (Fu et al., 1990). The 48 kDa (ISGF3y) itself can bind to the ISRE with low affinity (Kessler et al., 1990). ISGF3a, composed of the three other polypeptides, is immediately activated and translocated to the nucleus in response to interferon a and forms an active complex with ISGF3y which has a 20-fold higher binding affinity for ISRE than ISGF3y by itself (Bandyopadhyay et al., 1990; Kessler et al., 1990). Fu et al. (1992) presented evidence that ISGF3a complex was activated by a direct tyrosine phosphorylation at the level of SH2 and SH3 domains, which are commonly found in a number of tyrosine kinase-regulated proteins (see Koch et al., 1991). Deletion analysis of ISGF3y revealed that two domains confer the two known activities of this protein, DNA recognition and multimer assembly (Veals et al., 1993). Using a combination of specific tyrosine kinase and phosphatase inhibitors, David et al. (1993) were able to demonstrate that a tyrosine kinase and a membrane-bound tyrosine phosphatase lead to modification of ISGF3a and subsequent formation of the complete ISGF complex. The reader is referred to Levy and Darnell (1990) and Stark and Kerr (1 992) for recent reviews on the early events in interferon transduction system. *

PAGE 75

Implications In ruminants, the role of uterine PGFj^ in the demise of the corpus luteum is well established. However, the nature of the relationships between uterus and peri-attachment conceptus at the time of maternal recognition of pregnancy is still under intensive investigation. Potential mechanisms by which the embryo and bovine IFNx exert an antiluteolytic effect include regulation of endometrial steroid receptors (see Bazer, 1992) and oxytocin responsiveness (Mirando et al., 1993b). Alternative antiluteolytic mechanisms induced by the embryo and IFNx may involve regulation of some components of the endometrial prostaglandin biosynthetic pathway. The current concepts consider arachidonic acid availability and PGH synthase activity as the two main physiological constraints on the uterine capacity to synthesize prostaglandins (Marshall et al., 1987). i AA availability is dependent upon a variety of regulating factors such as PLA2 for the release of AA from membrane phospholipids, acylCoA-synthase and acyl transferases for the control of AA turnover in phospholipids (Norman and Poyser, 1993). Other metabolic pathways (e. g., lipoxygenase, epoxygenase, 3-oxydases) compete with PGHS for unesterified AA. In addition, AA accessibility to these enzymes can be controlled by intracellular AA-binding factors (Spector, 1975). The cyclooxygenase activity of PGH synthase requires nanomolar concentrations of hydroperoxides as activator (Kulmacz and Lands, 1983). Glutathione peroxidase and other peroxidases can decrease intracellular concentrations of hydroperoxides and thus reduce cyclooxygenase activity (Marshall et al., 1987). The irreversible self-inactivation of PGHS after a number

PAGE 76

69 of catalytic turnovers is another limiting factor in prostaglandin production (Smith and Lands, 1972). Finally, other unsaturated fatty acids can compete with AA for the substrate binding site on PGHS and inhibit prostaglandin synthesis (Lands etal., 1972). There are numerous examples of endogenous inhibitory activities to prostaglandin synthesis in a variety of other tissues: sheep endometrium (Basu, 1989) and allantoic fluid (Leach-Harper and Thornburn, 1984), renal cortex (Terragano et al., 1978), placenta of the rat (Harrowing and Williams, 1977), human amniotic fluid (Saeed et al., 1982), ovarian follicular fluid of humans (Carson et al., 1986) and cows (Shemesh, 1977), and human decidua (Ishihara et a!., 1990). However, none of these inhibitors has been fully characterized and isolated. The presence of an endogenous factor inhibiting prostaglandin synthesis in bovine endometrium has been described in a number of studies (Wlodawer et al., 1976; Shemesh et al., 1984a; Basu and Kindahl, 1987; Gross et al., 1988b). This inhibitor(s) decreased synthesis of POFj,, and all other identifiable cyclooxygenase metabolites including PGEj (Basu and Kindahl, 1987). The differential effect of pregnancy on PGF (inhibitory) versus PGEj (no effect) may reflect cellular differences (epithelium versus stroma) in prostaglandin secretion. High levels of inhibitor in epithelial cells (major source of PGF2„) but not in stromal cells (major source of PGEj) could explain the antiluteolytic effect of pregnancy. In view of these findings, the objective of research in this dissertation is to further characterize the cellular mechanisms by which the conceptus, through IFNt secretion, can regulate endometrial prostaglandin synthesis at the time of maternal recognition of pregnancy in cattle.

PAGE 77

CHAPTER 2 REGULATION OF ENDOMETRIAL PROSTAGLANDIN SYNTHESIS DURING EARLY PREGNANCY IN CATTLE: EFFECTS OF PHOSPHOLIPASES AND CALCIUM IN VITRO Introduction Maintenance of the corpus luteum (CL) during early pregnancy In cattle involves interactions between the conceptus and endometrium (Thatcher et al., 1984b). Increased plasma concentrations of 13,14-dihydro-15-ketoprostaglandin F2a (PGFM) associated with luteolysis are reduced or absent during early pregnancy (Kindahl et a!., 1976; Betteridge et al., 1984). Furthermore, in vitro endometrial synthesis of prostaglandin Fj,, (PGFjJ and PGFM are lower for pregnant than cyclic cows at Day 17 postestrus (Thatcher et al., 1984b; Gross et al., 1988b; 1988c); secretion of prostaglandin Ej (PGEj), in contrast, is either unaltered (Thatcher et al., 1984b; Gross et al., 1988b) or increased slightly during early pregnancy (Gross et al., 1988c; Helmer et al., 1989a). Presence of an endometrial inhibitor of prostaglandin synthesis during the estrous cycle and an increase in inhibitory activity during early pregnancy were demonstrated (Basu and Kindahl, 1987b; Gross et al., 1988a). The conceptus secretory proteins (bCSP) and bovine interferon tau (bIFNx) in particular, are believed to be responsible for many of the changes in prostaglandin metabolism that are associated with early pregnancy. For example, bIFNx induces a decrease in PGF2„ secretion and an increase in PGEj secretion by endometrial explants from cyclic cows (Helmer et al., 1989a); 70

PAGE 78

•' : : " 71 bCSP increase intracellular activity of an inhibitor of prostaglandin synthesis (Gross et al., 1988c). Therefore, a decrease in the ratio of PGF2a/PGE2 is one of the effects of bovine conceptus secretory proteins (Gross et al., 1988b; 1988c) that is observed during early pregnancy. These results indicate a differential regulation of endometrial prostaglandin biosynthesis during early pregnancy in cattle. Prostaglandin biosynthesis is regulated by a variety of factors including phospholipases and Ca2* (Marshall et al.. 1987). Phospholipase (PL) C is involved in prostaglandin biosynthesis through oxytocin-induced release of two second messengers: inositol 1,4,5-triphosphate (IP3) which initiates a rise in intracellular Ca^*, and diacylglycerol (DAG), which subsequently can activate protein kinase C or can be processed to obtain arachidonic acid (AA) (Schrey et al., 1986). Arachidonic acid is the rate-limiting precursor of prostaglandin biosynthesis (Marshall et al., 1987). Phospholipase A-2 (PLAj) induces release of AA from the second carbon of phospholipids, which also results in increased substrate availability for prostaglandin biosynthesis (Vogt, 1978). In addition, Ca2+-dependent PLA2 's a primary focus for the action of Ca2+ (Vogt, 1978; Schrey and Rubin, 1979). Riley and Poyser (1987) reported increased prostaglandin secretion from endometrium of guinea pigs in response to Ca^*. In addition, PGFza secretion was dependent upon presence of extracellular Ca2+, such that Ca2* preferentially increased PGF2a as compared to PGEj secretion. Calcium ionophore A23187 (Cal) increases Ca2* cycling across cell membranes (Pressman, 1976) and, therefore, is capable of altering prostaglandin biosynthesis (Gemsa et al., 1979; Knapp et al., 1977). Indeed, Cal increases prostaglandin secretion by perifused endometrium from cyclic gilts (Gross et al., 1990). Altered responses to regulators of prostaglandin synthesis due to

PAGE 79

72 pregnancy may account for the distinct differences in prostaglandin secretion between endometrial tissues of pregnancy versus the estrous cycle. Current experiments examined whether PLAj, PLC, extracellular Ca^* and calcium ionophore A23187 have regulatory effects on bovine endometrial PG biosynthesis. In addition, these factors were examined for their ability to exert preferential effects on PGF2a or PGEj biosynthesis and whether these effects differ between endometrium of cyclic and pregnant cows. Materials and Methods ' ' . ' " Materials ! Isotopically labeled [5,6,8,1 1,1 2,1 4, IS-^HJ-PGFza (specific activity = 180 Ci/mmole) and [5.6,8, 12,1 4, IS-^HJ-PGEj (specific activity = 184 Ci/mmole) were from Amersham Corporation (Arlington Heights, IL). Arachidonic acid was from Sigma Chemical Company (St. Louis, MO). Antiserum to PGFza was provided courtesy of T.G. Kennedy and antiserum to PGE2 was a gift from the E.L. Lilly Co. (Indianapolis, IN.). Phospholipase A (PLAj, from Naja naja venom, 1300 U/mg protein), phospholipase C (PLC, Type IX, from C. perfringens, 50-150 U/mg protein), and calcium ionophore A23187 (Cal) were from Sigma Chemical Company (St. Louis, MO). , A modified minimum essential medium (MEM; custom formula #87-5007) and other medium ingredients were purchased from Gibco (Grand Island, New York). Medium was prepared as described by Basha et al. (1980) except that phenol red was included and 10 ml of Gibco MEM vitamin solution was added per liter of medium. Additional MEM without Ca2+ (Ca2+ free-MEM) also was purchased from Gibco. ,

PAGE 80

. n Experiment 1: Effects of PLAo, PLC and Calcium lonophore on Endometrium from Cyclic Cows Dairy cows (Holstein and Jersey) were observed for estrous behavior (cyclic, n = 3) and slaughtered at Day 17 postestrus. Reproductive tracts were removed within 30 min after stunning and endometrium from the uterine horn ipsilateral to the CL was isolated from myometrium and minced. Tissue (0.5 g) was placed into duplicate petri dishes containing 15 ml MEM (containing 200 mg/l of CaCy for each of the following treatments: 1. Control 2. AA(0.2mg) 3. PLA2 (1U/ml, 2U/ml, 5U/ml) 4. PLC(1U/ml) 5. Calcium ionophore A23187 (Cal; 2, 4 and 10 i^g/ml). Medium samples (0.5 ml) were collected from each dish at 4, 6, 8, 10 and 12 h of incubation at 37°C on a rocker platform under an atmosphere of 45% O2, 50% Nj, 5% CO2. Medium samples were stored immediately at -20°C until analyzed. Experiment 2: Effects of PLAo PLC and Calcium lonophore in the Presence or Absence of Calcium on Endometrium from Cyclic and Pregnant Cows Holstein cows were observed for estrous behavior and either bred by artificial insemination (pregnant, n = 4) or not bred (cyclic, n = 3). Cows were slaughtered at Day 17 postestrus. All reproductive tracts were removed rapidly (within 30 min after stunning) and flushed with 40 ml MEM to collect conceptus tissues and confirm pregnant or cyclic statuses. Endometrium from the uterine horn ipsilateral to the CL was isolated from myometrium and minced. Tissue (0.25 g) was placed into duplicate petri dishes containing 7.5 ml MEM with Ca2*

PAGE 81

(CaClz = 200 mg/l) and into duplicate petri dishes containing 7.5 ml Ca^* free• \ ' t * ' MEM for each of the following treatments: _ ; , , •.. I 1 . Control 5. PLAj (1 U/ml) + Cal (7.5 ^ig/ml) 2. AA (0.1 mg) 6. AA (0.1 mg) + Cal (7.5 ^g/ml) 3. PLA2 (1 U/ml) 7. PLC (1 U/ml) 4. Cal (7.5 ^ig/ml) 8. PLC (2 U/ml) Medium samples (0.3 ml) were collected from each dish at 4, 8 and 12 h of incubation at 37°C on a rocker platform under atmosphere of 45% O2, 50% N2, 5% CO2. Medium samples were stored immediately at -20°C until analyzed. j Experiment 3: Ability of Endometrium from Cyclic Cows to Interconvert PGFo ^ and PGE, Holstein cows were obseryed for estrous behayior and slaughtered at Day 17 postestrus (cyclic, n = 4). Reproductiye tracts were removed rapidly (within 30 min. after stunning) and flushed with 40 ml MEM to confirm cyclic status. Endometrium from the uterine horn ipsilateral to the CL was isolated from myometrium and minced. Tissue (0.5 g) was placed into duplicate Petri dishes containing 15 ml of MEM for each of the following treatments: 1) (^H)-PGF2a (5 |iCi) ; 2) (^H)-PGE2 (5 piCi). Control cultures utilized similar treatments but in the absence of endometrial tissue. Tissues were incubated for 12 h at 39°C on a rocker platform under an atmosphere of 45% O2, 50% N2, 5% CO2. Following incubation, medium was stored immediately at -20°C until analyzed. Medium (1 ml) was extracted twice with ethyl acetate (3 ml), frozen at -70 °C, and the organic phase collected. Extraction efficiencies were 82.1 + 4.3% and 62.6 + 6.2% for medium containing (^H)-PGF2a and (^H)-PGE2 respectively. Ethyl acetate was evaporated and samples reconstituted with HPLC mobile

PAGE 82

76 phase (1 ml, acetic acid/water/ acetonitrile [0.5/70.5/29]). Samples were stored at -70°C until HPLC analysis. Prostaglandins were separated for each sample with a HPLC system (PerkinElmer Series 4; Perkin-Elmer Corp., Norwalk, CT) using a Cis column (Whatman Partsil, ODS-3, 5 |j,m particle size, 10% carbon load), a mobile phase of acetic acid/water/acetonitrile (0.5/70.5/29), and a flow rate of 1 ml/min for 60 min. The mobile phase was then changed to 100% acetonitrile at a flow rate of 1 ml/min for 20 min (to clean the column). Fractions were collected every 30 sec. (0.5 ml) and radiolabel counted using scintillation spectroscopy. Elution profiles were determined for each sample as well as for control samples, and results expressed as percent conversion (% of total elution profile cpm within each peak of interest) for PGF2a, PGE2, PGFM, and unidentified prostaglandins. Elution profiles were validated (peaks identified) by HPLC analysis of non-incubated (3H)-PGF2„ (fractions 57-68), (^H)-PGE2 (fractions 75-85) and (^H)-PGFM (fractions 108-120) standards. i I Radioimmunoassay Procedures ! Samples of medium were analyzed for PGF2a using a direct radioimmunoassay (RIA) procedure (Gross et al., 1988c) with an antibody characterized by Kennedy et al. (Kennedy, 1985). The antibody had a crossreactivity of 94% with PGFi„; therefore PGF2„ experimental responses are designated as PGF. Interand intra-assay coefficients of variation were 1 1 .4% and 13.1%, respectively. A similar assay for PGE2 (Gross et al., 1988c) utilized an antibody characterized by Lewis et al. (1978). Interand intra-assay coefficients of variation were 9.9% and 12.6%, respectively.

PAGE 83

statistical Analyses Data were analyzed statistically using least squares analysis of variance in the General Linear Models procedure of the Statistical Analysis System (SAS, 1985). Treatments initially were analyzed for Experiment 1 utilizing the model components of cow, treatment, cow x treatment, replicate (cow x treatment), time, cow x time, treatment x time, cow x treatment x time, and residual. Treatments were then analyzed separately to evaluate dose effects. Dose effects of AA, PLA^, PLC and Cal on endometrial PGF and PGEj secretions were analyzed using the model components of cow, dose, cow x dose, replicate (cow x dose), time, cow x time, dose x time, cow x dose x time and residual. Prostaglandin secretion rates over the 12 h incubation period were linear for all responses, and accumulation rates for each treatment-dose combination were determined using the model components of cow, replicate (cow), treatment x time (linear) and residual. , For Experiment 2, treatments initially were analyzed to determine status (pregnant vs cyclic) effects by using the model components (Table 2-1) of status, cow (status), Ca2*, treatment, status x Ca^*, Ca^* x cow (status), status x treatment, treatment x cow (status), Ca^* x treatment, status x Ca^* x treatment, Ca2* X treatment x cow (status) , replicate [cow (status) x Ca^* x treatment], time, status X time, Ca^* x time, treatment x time, and residual. There were significant effects of status (e.g., status x trt, status x Ca x trt, status x time). Therefore, remaining analyses were separated for each reproductive status. PG secretion rates over the 12 h incubation period were linear for all responses and the accumulation rates for each treatment-dose combination were determined by sorting data for status, Ca^*, and using the model components of cow, replicate (cow), time (linear) x treatment, and residual.

PAGE 84

77 Results in Experiment 2 were analyzed further as a series of factorial experiments to examine the effect of PLAj, Ca2+ and Cal (2x2x2, Tables 2-2, 2-3, 2-4), and the effects of AA, Ca2* and Cal (2 x 2 x 2 and 2x2, Tables 2-5, 26, 2-7). These factorial designs, for both statuses, utilized the model components of cow, treatment, cow x treatment, time, cow x time, treatment x time, cow x treatment x time and residual. Table 2-1 . Least squares analysis of variance of PGF and PGEj secretions by endometrial explants (Experiment 2) from cyclic and pregnant (status = „ Stat) cows (cow) sampled at different times (time) after treatments (trt; AA, PLAj and calcium ionophore) in presence or absence of Ca^* (Ca). Source df PGF PGE2 Error term SS SS Stat 1 47371 15343 cow (stat) cow(stat) 5 994309 612345 residual Ca 1 669884* 319476* cow (stat) X Ca Trt ; V 7 734915** 344750** cow (stat) X trt Stat X Ca 1 76833 6 cow (stat) X Ca stat X trt 7 352416** 30729 cow (stat) X trt Ca X cow(stat) 5 345056 138364 residual trt X cow(stat) 35 547502 115855 residual Ca X trt 7 111018* 13552 cow (stat) X Ca X trt stat X Ca X trt 7 88929 27494* cow (stat) X Ca X trt Ca X trt X cow (stat) 35 230529 51467 residual rep[cow(stat) x Ca x trt] 112 214217 942245 residual time 2 867287** 227119* residual stat x time 2 22699** 8129** residual Ca X time 2 33103** 36120** residual trt X time 14 149088** 29257** residual residual 426 588741 174960 P<0.01; * P<0.05

PAGE 85

Table 2-2. Least squares analysis of variance of PGF and PGEj basal secretions by endometrial explants (Experiment 2) from cyclic and pregnant (status = stat) cows (cow) sampled at different times (time) after treatments (trt): PLAj. Cal and Ca^* (2x2x2 factorial). Source df PGF PGE2 Error term SS SS stat 1 6911 5067 cow (stat) cow (stat) 5 546909 334138 residual trt 7 449885** 172069** cow (stat) X trt stat X trt 7 164427* 24872 cow (stat) xtrt trt X cow (stat) 35 338705 156702 cow (stat) X trt time 2 592855** 121356** cow (stat) X time stat X time 2 12650 1123 cow (stat) X time cow (stat) X time 10 85727 20316 residual trt xtime 14 79926** 30587** cow (stat) X trt X time stat X trt X time 14 32313 7024 cow (stat) X trt X time cow (stat) X trt X time 70 133858 21538 residual residual 168 203608 53763 P<0.01; * P<0.05 ! Table 2-3. Least squares analysis of variance of PGF and PGE2 secretions by endometrial explants (Experiment 2) from cyclic cows (cow) sampled at different times (time) after treatments (trt): PLAj, Cal and Ca^* (2x2x2 factorial). Source df PGF PGE2 Error term SS SS cow 3 290625 286777 residual trt 7 382731** 129175* cow (stat) X trt cow xtrt 21 213226 111979 cow (stat) X trt time 2 422896** 79657** cow (stat) X time cow X time 6 56001 18164 residual trt xtime 14 77502** 25291** cow (stat) X trt X time cow X trt X time ^ 89028 17771 residual residual 96 152450 39961 P<0.01; * P<0.05

PAGE 86

Table 2-4. Least squares analysis of variance of PGF and PGEj secretions by endometrial explants (Experiment 2) from pregnant cows (cow) sampled at different times (time) after treatments (trt): PLA2, Cal and Ca^* (2x2x2 factorial). Source df PGF PGE2 Error term SS SS cow 2 256283 47361 residual trt 7 231581* 67766* cow (stat) X trt cow X trt 14 125479 44722 cow (stat) X trt time 2 182609* 42822** cow (stat) X time cow X time 4 29725 2151 residual trt xtime 14 34738 12321* cow (stat) X trt X time cow X trt X time 28 44829 9766 residual residual 72 51158 ' 13802 P<0.01; * P<0.05 Table 2-5. Least squares analysis of variance of PGF and PGEj secretions by AA-treated endometrial explants (Experiment 2) from cyclic and pregnant (stat) cows (cow) sampled at different times (time) after treatments (trt): Cal and Ca^* (2x2 factorial). Source df PGF PGE2 Error term SS SS stat 1 443872 5428 cow (stat) cow (stat) 5 605197 209296 residual trt 3 467925** 128986** cow (stat) X trt stat X trt 3 55929 994 cow (stat) X trt trt X cow (stat) 15 450367 53199 cow (stat) X trt time 2 434672** 129875** cow (stat) X time stat X time 2 12108 1585 cow (stat) X time cow (stat) X time 10 82688 5771 residual trt xtime 6 11728 25720** cow (stat) X trt X time stat X trt X time 6 13828 3693 cow (stat) X trt X time cow (stat) X trt X time 30 43205 18043 residual residual 84 190566 59011 P<0.01

PAGE 87

80 Table 2-6. Least endometrial explants (Experiment 2) from cyclic cows (cow) sampled at different times (time) after treatments (trt): Cal and Ca^* (2x2 factorial), squares analysis of variance of PGF and PGEj secretions by AA-treated Source df PGF PGEj Error term SS ' SS cow 3 93564 204857 residual trt 3 452503* 80334* cow (stat) X trt cow X trt 9 383935 37116 cow (stat) X trt time 2 339197* 89100** cow (stat) X time cow X time 6 75734 3106 residual trt xtime B 24973 24035** cow (stat) X trt X time cow X trt X time 18 37022 11252 residual residual 48 150625 36594 P<0.01; ** P<0.05 Table 2-7. Least squares analysis of variance of PGF and PGEj secretions by AA-treated endometrial explants (Experiment 2) from pregnant cows (cow) sampled at different times (time) after treatments (trt): Cal and Ca^* (2x2 factorial). Source df PGF PGE2 Error term SS SS cow 2 511632 4439 residual trt 3 108137 53302* cow (stat) X trt cow X trt 6 66431 16082 cow (stat) X trt time 2 129381** 46620** cow (stat) X time cow X time 4 6953 2665 residual trt xtime 6 3028 7172 cow (stat) X trt X time cow X trt x time 12 6182 6791 residual residual 36 39941 22417 * P<0.05; ** P<0.01 Effects of PLC on endometrial prostaglandin secretion also were analyzed using the model components of status, cow(status), Ca2*, status x Ca2*, cow (status) X Ca2*, PLC, status x PLC, cow (status) x PLC, Ca2* x PLC, status x Ca2* 1

PAGE 88

81 X PLC, cow (status) x Ca2+ x PLC, replicate [cow (status) x Ca2+ x PLC], time, status x time, Ca2* x time, status x Ca^* x time, PLC x time, status x PLC x time, Ca2* x PLC X time, status x Ca^* x PLC x time and residual. Differences in metabolism of either radiolabeled PGFza or PGEj between presence or absence of endometrial tissue in Experiment 3 were evaluated by least squares analysis of variance considering effects of cow and tissue. Results Experiment 1 Endometrial PGF and PGEj secretion increased linearly throughout 12 h of incubation regardless of treatment. Arachidonic acid (AA) increased (P<0.01) PGF and PGEj secretion rates 103% and 85%, respectively (Table 2-8). In addition, AA increased (P<0.05) the ratio of PGF/PGEj (Table 2-8). Phospholipase C (PLC) did not alter endometrial prostaglandin secretion (Table 2-8). Phospholipase A2 (PLAj) increased endometrial PGF secretion (25% increase) in a non-dose-dependent manner (P<0.01); increases were similar regardless of dose (1, 2, and 5 U/ml; Table 1). In contrast, only the lowest dose of PLA2 (1 U/ml) increased (49% increase; P<0.01) endometrial PGEj secretion, whereas the higher doses (2 and 5 U/ml) had no effect (Table 2-8). Calcium ionophore A23187 (Cal) increased (P<0.01) endometrial secretion of PGF and PGE2 in a dose-dependent manner (Table 2-8). However, PGEj secretion was increased (14, 77, and 140% for 2, 4, and 10 |ag/ml, respectively) to a slightly greater degree than PGF secretion rates (0, 67, and 107% for 2, 4, and 10 ug/ml

PAGE 89

82 respectively). Neither PLC, PLAj or Cal altered the ratio of PGF/PGEj (Table 2Table 2-8. Accumulation rates (slopes) for PGF and PGEj by endometrial explants (0.5 g) from cyclic cows for Experiment 1 during 12 h of incubation. Results are expressed as ng PGF or PGEj/h. Accumulation Rate (ng/h) Ratio Treatment PGF PGE2 PGF / PGE2 uontrol 56.4 24.7 2.58 Arachidonic Acid' 1 14.6 AC ^ 45.6 3.55 rnospnoiipase A2 1 U/ml 2 U/ml 5 U/ml 71.9 69.1 69.9 36.9 24.9 26.7 2.18 2.85 2.60 Ca^* lonophore*' 2 ng/ml 4 pig/ml 10ng/ml 52.8 94.3 116.7 28.0 43.8 59.5 2.07 2.13 2.29 Phospholipase C** 53.5 28.2 2.43 Intercept (ng) 56.4 31.6 Pooled SE (±) 8.1 5.2 0.24 ' For PGF, PGE2 and ratio, treatment x time interaction (P<0.01). " For PGF and PGE2, treatment x time effects (P<0.01). Effects were dosedependent. For PGF and PGEj, treatment and time effects (P<0.01). Effects were not dose-dependent. " No treatment effects. Effect of time (P<0.01). Experiment 2 I After having established effective doses in Experiment 1 for the stimulation of endometrial prostaglandin secretion by PLAj, Cal and AA in cyclic

PAGE 90

cx)ws, effects of and interactions between these regulators of prostaglandin were examined. Secretion rates (Table 2-9) of PGF were reduced for endometrial explants from pregnant compared to cyclic cows for control and AA either in the presence or absence of Ca2+ (P<0.05). Arachidonic acid increased (P<0.01) both PGF and PGEj secretion rates in the presence or absence of exogenous Ca2* (321 and 99% respectively for endometrium of cyclic cows; 446 and 1 56% respectively for endometrium of pregnant cows). Furthermore, AA increased PGF secretion rate to a greater degree than PGEj secretion rate, which resulted in an increase (P<0.05) in the PGF/PGEj ratio (without AA, 1.8 and 1.5 for cyclic and pregnant cows, respectively; with AA, 3.7 and 3.0 for cyclic and pregnant cows, respectively). Stimulated increases in PGF and PGEj in response to PLAj and Cal were not of the magnitude obtained with AA. Regulation of basal prostaglandin secretion . Variation in prostaglandin secretion associated with responses to control, PLAj, Cal and PLAj + Cal in the presence or absence of Ca^* (2x2x2 factorial analysis) were considered as measurements of and regulators of basal prostaglandin secretion (Figure 2-1 ). Mean basal secretion of prostaglandin was higher in cultures with Ca^* (PGF: 144 > 94 [P<0.01] and PGEj: 91 > 51 [P<0.01] ng per 0.25 g per 8 h). Basal secretion rates of PGF were reduced from endometrial explants of pregnancy compared with the estrous cycle, and a differential response to treatments was detected (Status x Treatment, P<0.05; Figure 2-1). Although Ca2* stimulated PGF secretion, responsiveness to Cal differed between each reproductive status. The Cal stimulated PGF basal secretion either in the presence or absence of Ca2* in endometrium from cyclic and pregnant cows. However, endometrial tissue from pregnant cows expressed a hyperstimulation of PGF and PGE2 secretion when treated with Cal in the absence of Ca^*.

PAGE 91

.J1 84 Table 2-9. Accumulation rates (slopes) for PGF and PGEj by endometrial explants(0.25 g) from Experiment 2 during 12 h of incubation. Results are expressed as ng PGF or PGE2/h. Reproductive Status Cyclic Pregnant Treatment' Calcium PGF PGE2 PGF PGE2 Control + 13.3 9.7 6.2 4.9 5 8 4 3 4 5 32 Phospholipase Aj + A7.8 18.1 35.2 13.6 32 9 9 8 23 8 7 2 Avracniaonic acia T 1 /.O 11.1 •1 Q D 9.9 9 6 5.8 16.9 3.9 lonopnore 1 OA 0 20.0 12.2 1 1 .0 1 1 .4 8.2 5.2 17.7 8.7 PI A -tPal 19.2 12.0 11.7 7.3 5.0 6.9 4.6 AA + Cal 50.6 16.2 31.8 13.6 36.7 8.9 28.4 7.8 Phospholipase C: 1 U/ml + 9.6 6.9 8.3 7.1 5.3 4.3 3.1 2.4 2 U/ml + 7.9 5.3 6.9 8.0 2.2 3.0 2.6 1.3 Intercept (ng) + 67.5 32.4 32.2 25.6 56.7 29.4 33.0 25.3 Pooled SE (±) + 1.79 1.00 1.88 0.83 ^A4 0.61 1.54 0.64 'Treatment x time (linear) interaction (P<0.01) detected within each reproductive status (cycle, pregnant) and Ca2* (Ca2*, Ca2*-free) group for PGF and PGEj accumulation rates.

PAGE 92

65 Addition of PLA2 to the explants was an attempt to examine potential differential responses in the mobilization of endogenous pools of AA from phospholipids for synthesis of prostaglandins. Addition of PLAj to endometrial explants from cyclic cows did not stimulate basal PGF secretion, whereas secretion was stimulated in explants from pregnant cows (Status x Treatment, P<0.05). Indeed PLAs amplified basal secretion of PGF to the basal secretion level of explants from cyclic cows (Figure 2-1, A-B). The stimulatory effect of PLA2 on basal PGF secretion in pregnancy was blocked by the addition of Cal. Since PLA2 was ineffective in cyclic cows, Cal did not exert a negative effect on PGF basal secretion. Basal secretion of PGEj was considerably less than that of PGF. As observed for PGF, Ca^* and Cal amplified secretion of PGE2 (Figure 2-1 , C-D). Clear effects of pregnancy on PGEj secretion were not detected in contrast to the attenuating effects of pregnancy on PGF secretion. PLAg failed to stimulate PGEj secretion in explants from either cyclic or pregnant cows. Reoulation of stimulated secretion . An index of stimulated secretion of prostaglandin was the response to exogenous arachidonic acid (AA) that was evaluated by analyses of groups containing AA in the presence and absence of Ca2* and Cal (2x2 factorial analysis; Figure 2-2). Availability of AA appears to be limiting in endometrial explants of both cyclic and pregnant cows as noted by the marked increase in both PGF and PGEj secretion (Figure 2-2) compared to the basal secretion responses observed in Figure 2-1 . The stimulatory effects of exogenous AA are far greater than stimulation of basal secretion with PLAj (release of endogenous AA). Both PGF and PGEj stimulated secretions were enhanced when Ca2+ was added in the presence of AA (P<0.01) for endometrium of both cyclic and pregnant cows. In the presence of AA, Cal had no effect on stimulated secretion of either PGF or PGEj. The rate

PAGE 93

86 CYCLIC PREGNANT 200 PGF (ng/0.25g tissue) PGF (ng/0.25g tissue) 100 100 -PLA2-Ca -PLA2+Ca +PLA2-Ca +PLA2+Ca -PLA2-Ca -PLA2+Ca +PLA2-Ca +PLA2+Ca + Ca lonophore Ca lonophore 150-/ PGE2 (ngA).25g tissue) PGE2 (ng/0.25g tissue) Ad 100 100 -50 -PLA2-Ca -PLA2+Ca +PLA2-Ca +PLA2+Ca -PLA2-Ca -PLA2+Ca +PLA2-Ca +PLA2+Ca Figure 2-1 . (Experiment 2) Interactive effects of calcium (Ca^*), phospholipase A2 (PLA2), and calcium lonophore A23187 (Cal) on mean endometrial PGF (Panels A & B) and PGEj (Panels C & D) in medium from cyclic and pregnant cows. Results expressed as ng/0.25 g tissue. There was a Ca^* x PLA2 x Cal interaction (P<0.05) for each reproductive status.

PAGE 94

87 CYCLIC PGF (ng/0.25g tissue) PREGNANT PGF (ng/0.25g tissue) B " 500 400 300 200 100 0 +AA-Ca +AA+Ca +AA-Ca +AA4Ca H + Ca lonophore Ca lonophore PGE2 (ng/0.25g tissue) 150 -r ^ 100 50 PGE2 (ng/0.25g tissue) 150 100 +AA-Ca +AA+Ca +AA-Ca +AA+Ca Figure 2-2. (Experiment 2) Interactive effects of calcium (Ca), and calcium lonophore A23187 (Cal) on mean PGF (Panels A & B) and PGEj (Panels C & D) in medium from endometrial explants of cyclic and pregnant cows that received arachidonic acid (AA). Results expressed as ng/0.25 g tissue. There was a Ca2* x Cal interaction (P<0.05) for each reproductive status.

PAGE 95

of AA stimulated secretion of PGF was less for endometrial explants from pregnant cows versus cyclic cows (P<0.01). Phospholipase C . The effect of PLC on endometrial PG secretion differed (P<0.05) between pregnant and cyclic statuses (Figure 2-3). PGR (ngyo.25g tissue) ^GE2 (ng/0.25g tissue) 100 100 -PLC-Ca -PLC+Ca +PLC-Ca +PLC+Ca -PLC-Ca -PLC+Ca +PLC-Ca +PLC+Ca Cyclic Pregnant Figure 2-3. Interactive effects of calcium (Ca2*) and phospholipase C (PLC) on mean endometrial PGF (Panel A) and PGEj (Panel B) in medium from endometrial explants from cyclic and pregnant cows (Experiment 2). Results expressed as ng/0.25 g tissue. There were Ca^* x, PLC x reproductive status interaction (P<0.01) for mean responses of PGF and PGEj from pregnant cows (P<0.05). Phospholipase C caused a clear decrease (P<0.05) in PGF and PGEj secretion rates for endometrium from cyclic cows either in the presence or absence of Ca2* (Figure 2-3, Panel A). Similar PLC-induced decreases in PGF and PGEj

PAGE 96

89 were detected for endometrial tissue from pregnant cows in the absence of Ca^* (Figure 2-3) but not in the presence of Ca^* (Ca2+ x PLC interaction, P<0.05). Experiment 3 Experiments 1 and 2 demonstrate a differential regulation of PGF and PGEj and that pregnancy decreases PGF but not PGEj secretion (Experiment 2). To determine if relative differences in PGF and PGEj secretion may be due to interconversion of these molecules, endometrial tissue from cyclic cows was evaluated for its ability to convert (3H)-PGF2„ to PGEj and (3H)-PGE2 to PGFj^. Table 2-10. Percent conversions of (^H)-PGF2a and (^H)-PGE2 by endometrial explants (0.5 g) from Experiment 3 during 12 h of incubation. Results are expressed as percent of total chromatographic profile CPM (mean + SE) for PGFsa, PGE2, PGFM and unidentified PG products. Percent Conversion Substrate Treatment PGF2„ PGE2 PGFM Unidentified ('H)-PGF2„^ Control Tissue 88 ± 4.2 75 + 3.1 7 ±2.2 9 + 1.1 3± 1.1 9 + 2.3 2±1.2 7 ±0.9 (^H)-PGE2'' Control Tissue 4±1.0 3 + 0.5 86 ± 5.3 60 ± 4.2 1 ±0.4 2 ±0.6 9 ±3.0 35 ± 8.5 " Endometrial tissue did not convert (^H)-PGF2„ to PGE2; however, there was an increase (P<0.05) in PGFM and unidentified PG products compared to control incubations. Endometrial tissue did not convert (^H)-PGE2 to PGF2a or PGFM; however, there was an increase (P<0.01) in unidentified PG products compared to control incubations. Endometrium was unable to convert (3H)-PGF2„ to PGEj. However, (3H)-PGF2„ was metabolized to PGFM and unidentified PG metabolites compared to results for control (without tissue) incubations (Table 2-10). Endometrial tissue also

PAGE 97

90 was unable to convert (3H)-PGE2 to PGFja or PGFM as compared to control incubations. However, (3H)-PGE2 was metabolized to unidentified PG metabolites (Table 3) which bound to the HPLC column and were detectable following column cleaning (elution with 100% acetonitrile). Discussion Secretion of PGF is reduced, but secretion of PGEj is not lowered for endometrial explants from pregnant as compared to cyclic cows at Day 17 postestrus (Thatcher et al., 1984b; Gross et al., 1988b; 1988c). Present results confirm these findings and demonstrate a differential regulation of endometrial prostaglandin biosynthesis for PGF and PGE2 during early pregnancy in cattle. Prostaglandin biosynthesis is regulated by a variety of factors including phospholipases and Ca2+ (Marshall et al., 1987). Endometrial prostaglandin secretion is dependent upon the presence of extracellular Ca^*, and Ca^* preferentially increases PGF as compared to PGEj secretion by endometrial explants from guinea pigs (Riley and Poyser, 1987). The present results demonstrate a Ca^* dependency of PG biosynthesis for endometrial explants from cyclic and pregnant cows at Day 17 postestrus. Calcium ionophore A23187 (Cal) increases Ca2* cycling across plasma membranes and can induce release of Ca^* stored in the vesicles of the sarcoplasmic reticulum (Pressman, 1976). Therefore, this agent is capable of altering prostaglandin secretion (Gemsa et al., 1979; Knapp et al., 1977). Cal (A23187) can be used as an agonist of a PLC-generated second messenger IP3, to replicate the effect of oxytocin on prostaglandin production (Raw and Silvia, 1991). Cal is capable of increasing PG secretion by perifused endometrium from cyclic pigs (Basha et al., 1980),

PAGE 98

sheep (Raw and Silvia, 1991) and guinea pigs (Poyser, 1987). However, Lafrance and Goff (1990), using a 2.6 jag/ml concentration of A23187, failed to detect any effect of A23187 on PGF secretion by endometrial explants from heifers at day 19 or 20 postestrus. In the present study, PG secretion increased in response to 4 and 10 ng/ml in Experiment 1 and 7.5 \ig/m\ in Experiment 2; in agreement with Lafrance and Goff, a 2 ^ig/ml dose did not alter PG secretion (Experiment 1 ). Cal altered basal PG secretion by endometrial explants and the effect of Cal differed depending on the presence or absence of Ca^* between cyclic and pregnant statuses. In the absence of Ca2+, for instance, PGF secretion was unchanged essentially in endometrium from cyclic cows exposed to Cal and increased markedly in endometrium from pregnant cows. The membrane perturbation induced by A23187 in the absence of Ca2+ appears to initiate a clear response only in endometrium from pregnant cows. This may mirror the differences in membrane content of phospholipids and/or enzymes between the two reproductive statuses. Calcium ions can activate certain types of PLAj resulting in increased AA release (Vogt, 1978; Schrey and Rubin, 1979). Addition of PLA2 to the explants was an attempt to mobilize the endogenous pools of AA for prostaglandin synthesis. At day 17 postestrus, the quantity of esterified AA in endometrial phospholipids is lower in pregnant than in cyclic cows (Curl, 1988). Exogenous PLA2 stimulated PGF secretion only in explants from pregnant cows (status x treatment, P<0.05), regardless of the presence or absence of Ca2+. PLAj increased preferentially PGF secretion to attain levels comparable to PGF production in explants from cyclic cows. The status-specific effect of PLAj reflect a greater substrate accessibility for exogenous PLAj in endometrium from pregnant compared to cyclic cows. The concept of compartmentalization of

PAGE 99

"/., . 92 unsaturated fatty acids (UFA) such as AA was illustrated in a study by Shands and Noble (1984). Mitochondrial and microsomal fractions contained high proportions of UFA compared to plasma membranes and cytosolic factions. It is conceivable that UFA intracellular trafficking may be altered during early pregnancy. AA could be preferentially re-routed away form microsomal membranes rich in phospholipases and prostanoid biosynthetic enzymes. As a consequence, endometrial plasma membranes from pregnant cows would contain a higher proportion of AA which could be more accessible to exogenous PLA2Although PLA2 from Naja naja venom is a Ca2+-dependent enzyme (Verheij et a!., 1981), PLAj stimulatory effect on prostaglandin secretion was observed even in the absence of extracellular Ca^*, possibly due to activation by intracellular Ca^* after integration into plasma membranes. Of interest was the marked differential response to Cal between tissues of cyclic and pregnant cows treated with PLAj (PLAj ^ Cal, P<0.01). Recently, Poyser and Fergusson (1993) indicated that stimulation of prostaglandin secretion by PLAj si^ci Cal may involve different intracellular processes in the guinea pig. However, the negative interaction of Cal with PLAj this study of endometrium from pregnant cows suggests common or interdependent intracellular pathways, possibly involving protein kinase C (PKC) and phosphorylation of PLAj regulating factors. PKC isozymes a, p and y require Ca^* and activating lipid (DAG and/or AA) to be translocated from the cytoplasm to the ceil membrane and become activated (Kikkawa et al., 1989). A treatment with PLAj (source of AA) + Cal (source of Ca^*) may alter prostaglandin secretion potentially by PKC activation. Nishizuka (1992) suggested that, in physiological conditions, the activity of PKC may be sustained, even after the concentration of intracellular Ca^* is no longer

PAGE 100

increased, if DAG and c/s-UFA botli become available. Lafrance and Goff (1990) observed that synthetic PKC activators (PMA and OAG) stimulated PGFj^ secretion by endometrial explants from cyclic cows at day 19-20 postestrus. Alternatively, Cal perturbation of the cell membrane could be also detrimental to the stimulatory effect of exogenous PLAjNonetheless, this negative interaction with PGF secretion appears to be pregnancy-specific. In the bovine endometrium, the epithelial tissue is the major source of PGF and stromal tissue mainly secretes PGEj (Fortier et al., 1988). Moreover, Bonney et al. (1987) demonstrated the presence of two PLAj enzyme types in human endometrium. One is located in stromal tissue and maximally active in the presence of Ca^*, and the other is located mainly in glandular epithelium and inhibited in the presence of Ca^*. Such a tissue-specific regulation of PLA2 activity within the bovine endometrium could also play a role in the regulation of prostaglandin synthesis. In addition, PLA2 activity appears to be regulated by steroids since, in human endometrium, estradiol increased PLAj activity 6-fold (Bonney and Franks, 1987). Overall, results imply that a decrease in concentration and/or activity of PLAj may be responsible partially for the preferential decrease in endometrial PGF secretion observed during early pregnancy in cattle. Availability of AA appears to be limiting in endometrial explants of cyclic and pregnant cows since addition of AA increased prostaglandin secretion dramatically in both tissues. The stimulatory effect of AA on PGF secretion was far greater than the response to exogenous PLAjThe absolute increases in PGE2 secretion due to AA appeared to be equivalent for the two statuses. However, in tissue from pregnant cows, AA did not stimulate PGF secretion to the same extent as in cyclic cows. This implies that, when AA availability is not limiting, endometrial tissue from pregnant cows has a lower PGF production

PAGE 101

capacity compared with endometrium from cyclic cows. In other studies, endometrial explants (Thatcher et al., 1984a) or perifused endometrial tissue (Gross et al., 1988b) from pregnant cows at day 17 postestrus synthesized less PGF than those of cyclic cows. Ca2+ or Cal did not enhance the stimulatory effect of AA suggesting that, in presence of non-limiting amounts of substrate, the activity of enzymes involved in prostaglandin synthesis become a limiting factor. No differences in detectable prostaglandin synthase enzyme were found in endometrial tissue from pregnant and cyclic ewes on day 15 postestrus (Salamonsen and Findlay, 1990). If this is the case in cows, the current observations are consistent with the presence of increased amounts of an intracellular endometrial prostaglandin synthesis inhibitor (EPSI) during early pregnancy (Basu et al., 1987b; Gross et al., 1988b). The differential effect of pregnancy on PGF (inhibitory) versus PGEj (no effect) may reflect the cellular (epithelium versus stroma) differences in prostaglandin secretion. High levels of EPSI in epithelial but not in stromal cells could, at least in part, explain the lower PGF production in the presence of non-limiting amounts of AA in endometrial explants from pregnant cows. The role of PLC in mediating oxytocin-induced release of prostaglandins by the endometrium has been well documented (Silvia and Homanics, 1988; Mirando et al., 1990a). The exogenous PLC used in Experiments 1 and 2 (type IX from C. perfringens) is most active on phosphatidylserine and does not cleave phosphatidylinositol (Takahashi et al., 1981). Therefore this enzyme will not induce release of IP3 and trigger an increase in intracellular Ca^* by acting on phosphatidylinositol in a manner that may mimic the signal transduction system activated by oxytocin. In Experiment 1, PLC did not alter PG secretion from cyclic endometrium. However, in Experiment 2, PLC caused a clear decrease in

PAGE 102

9§ PGF and PGEj secretion in endometrium from cyclic cows. This inconsistency between experiments raises questions regarding the relevance of exogenous type IX PLC on prostaglandins secretion form bovine endometrium. At the time these experiments were conducted, phosphatidylinositol-specific PLC was not available commercially. In addition, G proteins are involved in the activation of PLC (see Harden, 1990). Integration of exogenous PLC into endometrial cell membranes does not guaranty an appropriate association with endogenous G proteins. The specificity of the exogenous PLC or a membrane perturbation may be responsible for the effect observed in Experiment 2. Ca2* is an activator of PLC from C. perfringens (Takahashi et al., 1981). Interestingly, tissue from pregnant cows exposed to Ca^* failed to have a decrease in PGF and PGEj secretion in response to PLC (status x Ca2+ x PLC interaction, P<0.01). This reinforces the concept of differential regulation of prostaglandin production in early pregnancy. Tissue from cyclic cows was not able to convert (^H)-PGF2a to PGEj implying that the g-keto-PGEj reductase activity required for such a conversion is absent in endometrial cells. However, the ability of the endometrium to metabolize (^H)-PGF2„ to PGFM was confirmed (Guibault et al., 1984). The early pregnancy-associated decrease in PGF secretion from the bovine endometrium is unlikely to be due to a conversion of PGF2„ into PGEj. In conclusion, the present experiment indicates that AA, Ca2* and PLA2 exert regulatory effects on bovine endometrial secretion of PGF20 and PGEj and that these factors, in concert with an inhibitor of prostaglandin synthesis, may be involved in the differential responsiveness and regulation of PGF secretion during early pregnancy in cattle. ^

PAGE 103

To further define the regulatory effects of pregnancy on endometrial secretion of prostaglandin, the objectives of the following experiments were to separate endometrial epithelial cells from stromal cells, and examine if the antiluteolytic factor bovine interferon x can alter prostaglandin secretion by endometrial cells in the absence or presence of oxytocin.

PAGE 104

CHAPTER 3 EFFECT OF NATURAL AND RECOMBINANT BOVINE INTERFERON t AND OXYTOCIN ON IN VITRO SECRETION OF PGFj^ AND PGEj BY ENDOMETRIAL EPITHELIAL AND STROMAL CELLS • I ^ • V Introduction In cyclic cows, regression of the CL is caused by pulsatile uterine secretion of PGFj^ (Nancarrow et al., 1973; Peterson et a!., 1975; Inskeep and Murdoch, 1980). Oxytocin is secreted by the CL (Wathes et a!., 1983; Flint and Sheldrick, 1986), and injection of oxytocin stimulates prostaglandin secretion from the uterus of cows during the luteal phase of the estrous cycle (Lafrance and Goff, 1985). Presence of oxytocin receptors in the endometrium determines the PGFja secretory responsiveness of the uterus (Roberts et al., 1976). Oxytocin receptor number progressively increases during the estrous cycle and is highest from Day 17 to 21 post-estrus in cows (Meyer et al., 1988; Fuchs et al., 1990). I Several in vivo experiments demonstrated that increased plasma concentrations of PGFj^, which normally occur just prior to onset of luteal regression, are reduced during early pregnancy (Kindahl et al., 1976; Betteridge et al., 1984). In vitro studies also showed that PGFj^ secretion (release and synthesis) from perifused endometrium and endometrial explants were lower for pregnant than cyclic cows at Day 17 post-estrus. However, secretion of prostaglandin Ej (PGEj) was unaltered for the perifused endometrium (Gross et al., 1988b) or increased in the case of endometrial explants (Gross et al..

PAGE 105

98 1988c). Clearly, PGEj secretion by endometrium from pregnant cows was not decreased. In early pregnancy, the day 15-17 conceptus secretes a protein complex, bovine trophoblast protein-1 (bTP-1), that prevents luteolysis. The bTP-1 complex has been identified as the antiluteolytic agent in cattle (Helmer et al., 1989a; Thatcher et al., 1989). bTP-1 complex consists of a group of acidic, low molecular weight polypeptides (22 to 26 kDa) (Helmer et al., 1987; Helmer et al., 1988b; Anthony et al., 1988; Plante et al., 1990). bTP-1 has an inferred amino acid sequence homology of 80 % with the ovine trophoblast protein-1 (oTP-1; Imakawa et al., 1989) and is immunologically related to oTP-1 (Helmer et al., 1987). Both bTP-1 and oTP-1 are classified as interferon tau (IFNx) proteins (Roberts, 1992), and were shown to have potent antiviral (Pontzer et al., 1988; Klemann et al., 1990; Plante et al., 1990) and immunosuppressive activities (Newton et al., 1989; Skopets et al., 1992). The major synthesis of blFNx by the conceptus is limited to the short period of Day 16 to 26 and increases with elongation of the conceptus (Bartol et al., 1985a; Geisert et al., 1988). In vivo, intrauterine infusions of bIFNx extended functional lifespan of the CL by inhibiting endometrial release of PGF2„ (Helmer et al., 1989a). In vitro, a 24 h-incubation with blFNx reduced PGFj^ secretion from endometrial explants of cyclic cows at Day 17 post-estrus (Helmer et al., 1989b). In sheep and cow endometrium, epithelial cells appear to be the major source of PGFj^, and PGEj is primarily secreted by stromal cells from sheep (Cherny and Findlay, 1990; Charpigny et al., 1991; Fortier et al., 1988). Salamonsen et al. (1988) reported an inhibitory effect of ovine IFNx on release of both PGF2„ and PGE2 from mixed ovine endometrial cells in primary culture. Using highly purified cultures of endometrial epithelial and stromal cells from cyclic sheep at day 15 post-estrus, Charpigny et al. (1991) observed that ovine

PAGE 106

99 IFNx inhibited basal and oxytocin-induced secretion of PGFja and PGE2 in both cell types. . It is of critical importance to better characterize the effects of regulatory agents such as blFNx and oxytocin on prostaglandin secretion from the two major endometrial cell types (epithelial versus stromal) in order to provide a better understanding of the molcular mechanisms involved in maternal recognition of pregnancy in cattle. Objectives of the following experiments were to (1) determine if bIFNx, either natural or recombinant, can modulate basal and oxytocin-induced secretion of PGFj^ and PGEj from endometrial epithelial and stromal cells; (2) compare potencies of natural bIFNx (nblFNx) versus recombinant bIFNx (rblFNx) to modulate secretion of PGF2„ and PGEj. Material and Methods Animals I Eight non-lactating cyclic Holstein dairy cows were synchronized for estrus, using Syncro-Mate-B ear implants (Sanofi Animal Health Inc., Overland Park, KS) given for 9 days and 25 mg Lutalyse (The Upjohn Company, Kalamazoo, Ml) injected i.m. 48 h prior to removal of the implants. Cows were observed for estrus and slaughtered at Day 1 5 post-estrus. Reproductive tracts were collected 15 min after slaughter within the abattoir. Endometrial tissue was obtained from the uterine horn ipsilateral to the CL for the purpose of in vitro cell culture (Figure 3-1). , I Preparation and Culture of Endometrial Cells Dissection of the tracts . Harvesting of endometrial tissue took place under sterile conditions in a laminar flow hood (Forma Scientific Inc., Marietta,

PAGE 107

100 OH). The ipsilateral uterine horn was cut open lengthwise across the antimesometrial border. Endometrium was separated from the myometrium with sharp curved Endometrium dissection (ipsilateral horn) Wash in Hank's medium (Ca^* and Mg 2* free) Mincing of endometrium with razor blades i Enzymatic digestion in Earle's medium for 2 h at 38 C collagenase type IV; 150 lU/ml hyaluronidase type III; 150 lU/ml 2% BSA centrifugation, wash and homogenization Gauze filtration Nylon mesh filtration (30 ^im pores) Epithelial clumps | Isolated cells , (stromal, epithelial, I blood cells,...) i Culture separately in MEM HAM F-1 2 medium (vA/) with 10% calf serum, 10% horse serum, insulin (0.2 lU/ml), antibiotics and antimycotic at 38 C under a 95% air, 5% CO atmosphere Figure 3-1 . Outline of the procedure used to separate and culture epithelial and stromal cells from endometrium of cyclic cows at day 15 post-estrus.

PAGE 108

101 scissors and placed into a medium of Hanks' Balanced Salts (Modified, Ca^* and Mg2* free; Sigma Chemical Company, St. Louis, MO). Endometrial tissue was minced using two scalpel blades until a smooth homogenous paste was formed. Enzymatic dioestion . Minced endometrial tissue from one horn was placed into 150 ml of Earle's Balance Salts medium (Sigma), which also contained: 150 lU/ml of collagenase type IV (Sigma); 150 lU/ml of hyaluronidase type III (Sigma); 2% Bovine Serum Albumin (BSA, Sigma). Tissue was dispersed in this solution with incubation at 38°C for 2 h with periodical swirling. At termination of incubation, enzymatic digestion was verified by microscopic observation. For homogenization, tissue was poured into 50 ml conical centrifuge tubes and triturated with a 20 ml syringe and canulae (14 ga., 10 cm). The digestion was then stopped by centrifugation of the tissue digest (500 x g for 5 min). After discarding the supernatant, the pellet was resuspended in 200 ml Earle's medium, triturated and centrifuged once more. After the second centrifugation, pellets were resuspended in 400 ml of Earle's medium and triturated to form air bubbles. The debris floating to the top was aspirated off, and the cell suspension was then ready for filtration. Filtration . In order to eliminate non-digested pieces of tissue, the cell suspension was filtered first through four layers of gauze. A second filtration through a nylon mesh (30 nm pores; Spectrum, Los Angeles, CA) separated the isolated cells, which went through the mesh, from the retained aggregated cells. The isolated cells (red blood cells, leukocytes, lymphocytes, macrophages, non-adherent epithelial cells, stromal cells...) contained mostly stromal cells, whereas the aggregated cells contained mostly epithelial cells. Aggregated cells were resuspended in 50 ml of Earle's medium, and both solutions were centrifuged (500 x g for 5 min). The pellets were dispersed into culture medium. The culture medium consisted of 39.5% Minimum Essential Medium Eagle 'pi"

PAGE 109

102 (Earle's Salts; Sigma), 39.5% Nutrient Mixture F-12 HAM (Sigma), 10% Fetal Bovine Serum (heat inactivated; Sigma), 10% Horse Serum (heat inactivated; Sigma), insulin (0.2 lU/ml; Sigma), 1% antibiotic and antimycotic solution (Sigma). The antibiotic and antimycotic solution (ABAM) contained 10,000 ID penicillin, 10 mg streptomycin and 25 |jg amphotericin-B per ml. The serum supplied a source of growth and adhesion factors. . , Plating of the cells . After dispersing the respective cells types into the culture medium, cells were plated into 6-well culture plates (diameter = 30 mm, 2 ml/well; Falcon, Becton Dickinson & Company, Lincoln Park, NJ). Isolated cells were plated at an average concentration of 10^ cells per ml. This was done to achieve cell confluency at about 6 to 7 days after plating. Incubation and cell purification . All plated cells were incubated at 38.5°C in a microprocessor controlled COj incubator (5% COj; Lab Line Instruments Inc., Melrose Park, IL). Epithelial clumps (nylon mesh retentate) required 48 h to attach to the culture plates. At 48 h after plating, the endometrial epithelial cells were attached and the medium was replaced. Stromal cells (nylon mesh filtrate) attached to the culture plates within 24 h after plating, whereas contaminating isolated epithelial cells required 48 h to attach. Based on this difference in time for adhesion between the two cell types, stromal cells could be separated from contaminating cells (e.g., epithelial cells, red blood cells, lymphocytes, macrophages) by changing medium at 24 h after plating. Original medium was aspirated and new medium (2ml/well) was added. The new medium consisted of MEM/F-12 HAM (v/v), 2% Ultroser SF (IBF Biotechnics, Columbia, MD), bovine insulin (0.2 lU/ml) and 1% ABAM. This latter medium was changed every 2 to 3 days until cells reached confluency and were ready to undergo the experimental treatment (Table 3-1 ).

PAGE 110

f 103 Table 3-1. Times of medium changes (MEM/F-12 HAM (v/v), 2% Ultroser SF, insulin (0.2 lU/ml) and 1% ABAM) for epithelial and stromal cells cultured separately in 6-well plastic culture plates (30 mm diameter, 2 ml/well). Day of culture 0 Separate plating of epithelial clumps and isolated cell mixture 1 Medium change for isolated cells only stromal cells remain 2 Medium change for epithelial cells 4 Medium change for both cell types 6 Start experimental treatment at early confluency Purity of the stromal and epithelial cell cultures were verified microscopically based on differences of morphology of the two cell types (Cherny and Findlay, 1990). In a preconfluent stage, stromal cells were fibroblast-like in appearance with numerous extensions; at confluency they exhibited 'streaming' and formed loops and twirls. The epithelial cells formed initial islands of cuboidal or columnar shaped cells deriving from the originally plated clumps of epithelial cells and developed a cobble stone appearance at confluency (Photos 1-4). Endometrial epithelial and stromal cells from four cows were also plated on plastic two-chamber Lab-Tek slides (21.3 x 20 mm/chamber; 2 ml/chamber; Nunc Inc., Naperville, IL) for immunocytochemistry of the two cell types. Epithelial and stromal cells were treated with primary antibodies (monoclonal anti-cytokeratin 37, clone KB-37 (Sigma; dilution 1:50) and monoclonal antivementin. Clone V9 (Sigma; dilution 1:500) on separate slides, followed by an immunoperoxidase staining procedure. All antibody treated slides and control slides (negative controls without primary antibody) were stained, using

PAGE 111

104 immunoperoxidase staining kits (Biomeda Corp., Foster City, CA). Epithelial cells stained positively for cytokeratin but not for vimentin, whereas stromal cells stained positively for vementin, but not for cytokeratine. (Figure 3-3). A B C Figure 3-2. Confluent live endometrial epithelial cells (A) at day 7 of culture; cobblestone shaped daughter cells surrounding the initially plated clump of epthelial cells (x 100). Confluent live endometrial stromal cells (B) at day 7 of culture (x 100). Example of a live epithelial cell culture "contaminated" with stromal cells (C; x 100).

PAGE 112

105 Experimental design Two independent experiments were conducted with in vitro cultured endometrial cells. Endometrial epithelial and stromal cells were harvested from two different sets of four cyclic cows each (day 15 post-estrus). Epithelial and stromal cells were cultured in 30 mm wells until they reached early confluency at approximately 6-7 days after plating. Then cells were left untreated or were treated with natural or recombinant bIFNx (nblFNt or rblFNx) for 24 h at the following doses: untreated at 0 ng/ml; nblFNx at 2, 10 and 50 ng/ml; rblFNx at 0.4, 2, 10, 50 and 250 ng/ml. Each dose was administered to 10 wells (for two cows) or 6 wells (for two cows) of cultured epithelial cells; and to 6 wells (for all four cows) of cultured stromal cells. After the 24 h incubation period, a sample (200 |al) was taken from the medium from each well (Time = 0 min). Oxytocin (2.0 x 10-^ M or 201 ng/ml; Sigma) was then added to half of the wells of each dose. All wells were sampled (200 pi) again at 30 and 90 min after administration of oxytocin or at Time = 0 min. The control group consisted of those wells which were not treated with either bIFNx or rblFNx. This 210 ng/ml dose of oxytocin (200 x 10-9 M) was given to insure saturation of the oxytocin receptors (Kq = 0.94 x lO^^ M; Fuchs et al., 1991). Natural bIFNx was purified by Plante et al. (1990) and recombinant bIFNx was produced and purified by Klemann et al. (1990). , ' ^ . , ' , Radioimmunoassay of PGFon and PGE o Diluted medium samples were analyzed for PGFj^ and PGEj by direct radioimmunoassay (RIA) according to a procedure described by Knickerbocker

PAGE 113

106 A B C D Figure 3-3. Immunocytochemistry of endometrial epithelial (A-B) and stromal (C-D) cells from cyclic cows (day 15 post-estrus). Cells were cultured for 7 days then treated with primary antibodies against cytokeratin (clone KB-37; dilution 1:50) and vimentin (clone \/9; dilution 1:500); followed by an immunoperoxidase staining procedure. Epithelial cells stained for cytokeratin (A) but not stromal cells (C), whereas stromal cells stained positively for vimentin (D), but not epithelial cells (B). et al. (1986b). The assay for PGFja utilized antibody which was characterized by Dubois and Bazer (1991) and tritiated PGF^^ ([5,6,88,9,1 1,12,14,15-3H]RGFj^; specific activity of 160-180 Ci/mmole; Amersham Corp., Arlington Heights, IL). Standard curves were prepared in Tris buffer (100 pi; 0.05 M Trizma Base (Sigma); pH = 7.5) with known amounts of radio-inert PGF2„ (5, 10,

PAGE 114

25, 50, 100, 250 and 500 pg/tube). An antiserum dilution of 1:5000 (0.1 ml/tube) was used and the minimum concentration which was distinguishable from zero was 5 pg per tube. Medium samples were diluted to be in the range of the standard curves. Intraand inter-assay coefficients of variation were 7.0 % and 10.1 % respectively. A similar assay was used for PGEj, utilizing antiserum which was characterized by Lewis et al. (1978). Tritiated PGE2 ([5,6,8, 11, 12, 14,1 5-3H]PGEj; specific activity of 140-170 Ci/mmole) was purchased from Amersham Corp. Antiserum was used at a dilution of 1 :6000 and the assay sensitivity was 5 pg/tube. For PGEj similar standard curves were used as for PGFj^. Intraand inter-assay coefficients of variation were 5.2 % and 12.5 % respectively. Pooled samples with 10,000 pg/ml (high pool), 2,500 pg/ml (medium pool) and 500 pg/ml (low pool) added PGFj,, (or PGEj) were assayed serially in 2.5, 5, 10, 25, 50 and 100 \i\ volumes. The three inhibition lines for PGEj,, and PGEj did not differ for the standard curve when tested for heterogeneity of regression. The assays were further characterized by measuring known amounts (2.5, 5, 10, 25, 50 and 100 pg) of PGFj^ (or PGEj) in a pooled sample of culture medium (Y = 1 .009X + 1 1 .1 1 ; Y = pg of PGFj^ measured and X = pg of PGFj^ added; R2 = 0.98; P<0.01; Z = 1.009W + 11.11; Z = pg of PGE2 measured and W = pg of PGE2 added; R2 = 0.99; P<0.01 ). DNA Assav At the end of the experiments, culture medium was drained from the wells. Demineralized water (I ml/well) was added to each well, cells were scraped with a plastic spatula from the well surface and pipetted into tubes. Wells were washed with additional 1 ml demineralized water, which also was added to the tubes. Cell suspensions were sonicated for 30 sec. Samples (50 ^jI) were taken

PAGE 115

108 from the cell suspensions for quantitative determination of DNA by fluorometry according to a procedure described by Labarca and Paigen (1980). The method was based on enhancement of fluorescence when bisbenzimidazole (Hoechst-Roussel, Sommerville, NJ) binds to DNA. Standard curves of calf thymus DNA (Type V; Sigma) were prepared in phosphate buffered saline (PBS; 0.2 g KCI, 0.2 g KH3PO4, 8 g NaCI, 2.16 g Na2PO4.7H20; pH = 7.4). Standards were at 0, 50, 100, 250, 500, 1000, 2500 and 5000 ng DNA/50 ^1 PBS. Dye solution was prepared by diluting 10 pi of Hoechst H33258 stock (1 mg/ml in demineralized water) to 100 ml of TNE buffer (0.01 M Trizma Base, 0.001 M EDTA-Naj, 0.1 M NaCI; pH = 7.4). Dye solution was added to all standards (in triplicate) and samples (in duplicate) and incubated at room temperature for 45 min. Fluorescence of the samples and standards increased with time until it reached a plateau (within 45 min after administration of the dye). When fluorescence had stabilized, samples were analyzed using a DNA-Fluorometer (Hoefer Scientific Instruments, San Francisco, CA). Statistical Analysis Data were analyzed statistically by least squares analysis of variance using Analysis of Variance and General Linear Models procedures of the Statistical Analysis System (SAS, 1989). The mathematical models for epithelial and stromal cells included cow (n = 4), treatment (nblFNx versus rblFNx), dose (0, 0.4, 2, 10, 50, and 250 ng bIFNx/ml), oxytocin (-/+ oxytocin; 2.0 x 10-7 m), time of sampling (0, 30 and 90 min) and all higher order interactions (Tables 3-2 and 3-3). Effects of treatment, dose, oxytocin and time of sampling were considered fixed effects, while the cow effect was regarded as a random effect. Effects were tested using appropriate error terms according to the expectation of mean squares. Contrasts were used to determine differences between

PAGE 116

109 Table 3-2. Analysis of variance of PGFj,, and PGEj secretion by endometrial epithelial cells from cyclic cows (C) after treatment with natural and recombinant IFNx (T) at various doses (D) with or without oxytocin (0) and sampled at various times after treatment (S). OUUl rlf 1 v3i 2(3[ IVIO PGF„ MS Frror term C 3 47.23** 15.68** residual T 1 1.53 21.50 CxT CxT 3 11.89 3.47 residual 0 1 655.01 208.52 CxO CxO 3 196.87** 47.41** residual TxO 1 3.34 0.31 C xTxO CxTxO 3 4.67** 0.44 residual D 3 390.14** 138.51** CxD CxD 9 36.45** 4.14** residual TxD 3 22.22* 11.92** C xTx D OxD 3 114.13 14.34** C X 0 x D CxTxD 9 3.25** 0.77** residual CxOxD 9 45.00** 1.59** residual CxTxOxD 12 1.01 0.56** residual S 2 375.18** 123.48** CxS CxS 6 26.20** 2.96 residual TxS 2 0.16 0.48 C xTx S OxS 2 244.64* 57.71** C X 0 X S D X S 6 42.12 5.73 C X D X S CxTxS 6 1.14 0.10 residual CxOxS 6 29.65** 5.15** residual Cx D X S 18 8.20** 0.46 residual TxOxS 2 0.24 0.11 CxTxOxS TxDxS 6 0.92 0.36** C xTx D X S Ox DxS 6 36.30* 3.29* C X 0 X D X S TxOxDxS 9 1.10 0.56 CxTxOxDxS CxTxOxS 6 0.87 0.07 residual CxTxDxS 18 0.38 0.08 residual CxOxDxS 18 9.36** 0.46** residual CxTxOxDxS 15 0.00 0.00 residual residual 576 1.02 0.20 • P<0.05; ** P<0.01

PAGE 117

110 Table 3-3. Analysis of variance of PGFja and PGEj secretion by endometrial stromal cells (experiment 2) form cyclic cows (C) after treatment with natural and recombinant IFNx (T) at various doses (D) with or without oxytocin (0) and sampled at various times (S) after treatment. ) aftPjf trfiatment Source df PGF2„ MS PGEj MS Error term c 3 80.63** 146.22** residual T 1 1 4.76 9.14 C x T C xT 3 4.75 8.08 residual 0 1 0.12 0.00 CxO C X 0 3 0.17 0.82 residual TxO 1 1.26 11.43 CxTxO C xTx 0 3 0.88** 3.89** residual D 3 10.52 23.15 CxD CxD 9 10.10** 29.36** residual TxD 3 4.25 11.34 C xTx D OxD 3 1.20 4.05 CxOxD C xTx D 9 6.22** 21.75** residual CxOxD 9 6.74** 1.54** residual CxTxOxD 12 4.17** 2.25** residual S 2 1.34 39.44 CxS CxS 6 0.56** 11.99** residual TxS 2 0.03 0.72 CxTxS OxS 2 0.17 0.08 CxOxS DxS 6 0.32 1.35 CxDxS C xTx S 6 0.17* 0.20 residual CxOxS 6 0.22** 0.19 residual CxDxS 18 7.44** 1.27** residual TxOxS 2 0.03 2.29 CxTxOxS TxDxS 6 0.16 0.52 CxTxDxS OxDxS 6 0.16 0.11 0 X 0 x D X S TxOxDxS 9 0.09 0.95 CxTxOxDxS CxTxOxS 6 0.19 0.59 residual CxTxDxS 18 0.11 0.60 residual CxOxDxS 18 0.16 0.69 residual CxTxOxDxS 15 0.07 1.88** residual residual 576 0.07 0.60 * P<0.05; ** P<0.01

PAGE 118

111 treatments and doses. To examine the effects of IFNx on basal and oxytocinstimulated secretion of PGF and PGEj , the data set was sorted by oxytocin for the samples taken at 30 and 90 min (after ± oxytocin). The reduced mathematical models included cow, treatment (0, 0.4, 2, 10, 50, 250 ng rblFNx /ml; 2, 10, 50 ng nblFNx/ml) and time (30, 90 min). . . ^ Results Wells containing only epithelial ceils, that presented no signs of stromal cell contamination as determined by microscopic observation, were used in the first set of experiments. Epithelial cells evolved from original aggregates of epithelium; when contamination occurred, stromal cells were surrounding the epithelial cell areas. In the second set of experiments, contamination of stromal cells with epithelial cells was not observed. Epithelial cells j ' ' The DNA content per well of epithelial cells averaged 32.4 ^g (= 5.4x10* cells/well; Table 3-4) and was fairly consistent among all wells (SE = 0.09; c.v. = 7.8%). There was no effect of treatments (nblFNx versus rblFNx) on DNA, indicating that cell proliferation/survival was not altered by bIFNx over a 25.5 h period. Epithelial cells secreted more PGFj^ than PGEj (Table 3-4) and each PG accumulated over time (P<0.01), but the predominance of PGFj^ over PGEj was maintained. Both natural and recombinant bIFNx suppressed PGFje, and PGEj (p<0.01 ) after 24 hr of treatment (Time = 0 min; Figure 3-4). At this time, mean secretion levels of PGFj^ for the nblFNx and rblFNx treated wells (averaged over

PAGE 119

112 all doses) versus the control wells were 2.41 and 2.00 versus 3.46 ng/^g DNA (SE = 0.05) respectively. The respective means for PGEj were 1.84 and 1.40 versus 2.41 ng/^ig DNA (nblFNx and rblFNx versus control; SE = 0.03). Basal secretion of PGFaoj and PGEj (without oxytocin) decreased with increasing doses of either nblFNx or rblFNx (P<0.01 ); differences in suppression between doses were sustained throughout the 90 min period (Figure 3-4). Table 3-4. Least squares for PGF2„ and PGEj in conditioned culture media and DNA content for bovine endometrial epithelial and stromal cells (day 1 5 post-estrus). Epithelial cells Stromal cells overall^ control" overall control PGF2„ (ng/ng DNA) 3.22 2.9 • 0.89 0.59 well variance 0.86 0.2 0.08 0.01 C.V.*'(%) ' 28.7 13.9 32.5 17.1 PGEj (ng/^ig DNA) 2.22 2.3 53.3 42.2 well variance 0.18 0.2 618.6 113.2 C.V. (%) 19.2 14.1 46.6 25.2 DNA (iig/well) 32.4 1 17.4 well variance 6.4 6.7 C.V. (%) 7.8 14.9 ' overall = overall mean secretion of PGFj^ and PGEj and mean cell DNA content per well, calculated as an average over all samples of all experimental units. " basal secretion of PGFj^ and PGEj after 24 h of incubation of the control wells. > *^ C.V. = coefficient of variation. ^ To study the effect of oxytocin, all data from samples taken before the administration of oxytocin (i.e. samples taken at Time 0 min) were excluded from the analysis. The interaction between oxytocin (+/-) and time of sampling was

PAGE 120

significant for epithelial cell secretion of both PGFja (P<0.05) and PGEj (P<0.01). Secretion of POFj^and PGEj were higher when epithelial cells were exposed to oxytocin and the increase was greater during the first 30 min than between 30 and 90 min following oxytocin treatment (Figures 3-4 and 3-5). Furthermore, there were oxytocin x treatment x dose x time interactions (P<0.01) for PGFja and PGE2. Not only was basal secretion of PGFj^ and PGEj suppressed by both types of bIFNx but the high secretory responses of PGF2
PAGE 121

114 Natural IFNt wthout Oxytocin Recomttinant IFNt wthout Oxytocin 60 time (min) 90 30 60 time (min) -+— control -O— 0.4 ng/ml 2 ng/ml 10 ng/ml -a— 50 ng/ml -©— 250 ng/ml Natural IFNt viith Oxytocin 12 10 Recomtiinant IFNt wtti Oxytocin 30 60 time (min) 90 30 60 time (min) Figure 3-4. Least squares means for PGF secretion by epithelial cells from samples taken at 0, 30 and 90 min after 24 h incubation period with control or bIFNx (natural vs recombinant) at various doses, in absence of oxytocin (basal secretion) or presence of oxytocin (2 x 10-^ M; stimulated secretion).

PAGE 122

115 030 60 time (min) 90 control 0.4 ng/ml NatufcJ IFNt wth Oxytocin 30 60 time (min) 2 ng/ml -B— 10 ng/ml -B— 50 ng/ml -e— 250 ng/ml 30 60 time (min) Recombinant IFNt wtti Oxytocin 30 60 time (min) 90 Figure 3-5. Least squares means for PGEj secretion by epithelial cells from samples taken at 0, 30 and 90 min after 24 h incubation period with control or blFNx (natural vs recombinant) at various doses, in absence of oxytocin (basal secretion) or presence of oxytocin (2 x 10-^ M; stimulated secretion).

PAGE 123

116 8n control nIFNx rIFNx oxytocin + oxytocin 0 2 10 50 0.4 2 10 50 250 control nlFNi rlFNr Figure 3-6. Least squares means for PGR and PGEj secretion by epitlielial cells (average of 30 and 90 min sampling times) after 24 h incubation period with control medium or bIFNx (natural and recombinant) at various doses, in absence of oxytocin (basal secretion) or presence of oxytocin (2 x 10-^ M; stimulated secretion).

PAGE 124

117 0 2 10 50 0 2 10 50 dose (ng/ml) dose (ng/ml) K — natural IFNx 0 recombinant IFN X 0 2 10 50 0 2 10 50 dose (ng/ml) dose (ng/ml) Figure 3-7. Dose responses of PGF and PGEj by endometrial epithelial cells to natural and recombinant blFNx at various doses, in absence of oxytocin (basal secretion) or presence of oxytocin (2 x 10-^ M; stimulated secretion). Plotted least squares means are comprised of samples taken at 30 and 90 min after treatment.

PAGE 125

118 Stromal Cells ' The average amount of DNA per well for the stromal cells was 17.4 ^ig (« 2.9x10^ cells/well). Stromal cells secreted 60 times more PGEj (55.4 ng/^g DNA; SE = 16.9) than PGFj^ (0.94 ng/|^g DNA; SE = 0.31; Table 3-4). Secretion levels of PGF2„ and PGEj did not show significant increases over the 90 min sampling period (Figures 3-8 and 3-9). Natural or recombinant bIFNx failed to exhibit any regulatory effect on the secretion of PGFj^ and PGE2 by stromal cells (Figure 3-10); prostaglandin secretion was higher slightly for bIFNx treated cells (all doses) than for control cells, but differences were not significant. The average PGF2„ concentration for bIFNx-treated stromal cells versus control stromal cells was 0.99 versus 0.57 ng/^g DNA (SE = 0.31) respectively; the respective average PGEj secretion were 58.1 versus 34.3 ng/|ig DNA (SE = 16.9). The oxytocin effects were examined in samples collected at 30 and 90 min (i.e., after oxytocin had been administered). Stromal cells were not responsive to oxytocin treatment (Figure 3-10). The average level of PGF2„ secretion was 0.92 versus 0.84 ng/|ag DNA (SE = 0.31) for oxytocin treated versus control wells. The average PGE2 secretion was 54.2 ng/iig DNA (SE = 0.9) for oxytocin-treated wells versus 56.6 ng/fig DNA (SE = 16.9) for the control wells. Appreciable variability was found among cows for all PGF2„ and PGEj responses (for example: there was a seven-fold difference in overall secretion of both PGFsa and PGEj between the cows with the highest and lowest prostaglandin secretion). There was no evidence that stromal cells were responsive to bIFNx or oxytocin.

PAGE 126

1 Natural IFNt wthout Oxytocin Recombinant IFNt wthout Oxytocin 60 time (min) 90 30 60 time (min) -t — control — X-Q— 0.4 ng/ml — b2 ng/ml —a— 50 ng/ml 10 ng/ml o 250 ng/ml A/afura/ /FA/r M/f/7 Oxytocin 1.5 RecomtMnant IFNt vith Oxytocin 30 60 time (min) 90 30 60 time (min) 90 Figure 3-8. Least squares means for PGF secretion by stromal cells from samples taken at 0, 30 and 90 min after 24 h incubation period with control or blFNt (natural vs recombinant) at various doses, in absence of oxytocin (basal secretion) or presence of oxytocin (2 x 1 0-^ M; stimulated secretion).

PAGE 127

Natural IFNr wthout Oxytocin RecomtHnant IFNt vithout Oxytocin 30 60 time (min) -» — control -O— 0.4 ng/ml Natuial IFNt wth Oxytocin 30 60 time (min) 30 60 time (min) 2 ng/ml 10 ng/ml -a— 50 ng/ml -O— 250 ng/ml Recombinant IFNt wth Oxytocin 30 60 time (min) Figure 3-9. Least squares means for PGEj secretion by stromal cells from samples taken at 0, 30 and 90 min after 24 h incubation period with control or blFNx (natural vs recombinant) at various doses, in absence of oxytocin (basal secretion) or presence of oxytocin (2 x 10-^ M; stimulated secretion).

PAGE 128

121 S.E. =0.31 0 control 2 10 50 nlFNi 0.4 2 10 50 250 rlFNx oxytocin + oxytocin 0 control S.E. = 16.92 2 10 50 nlFNx 0.4 2 10 50 250 rIFNx Figure 3-10. Least squares means for PGF and PGEj secretion by stromal cells (average of 30 and 90 min sampling times) after 24 h incubation period with control medium or bIFNx (natural and recombinant) at various doses, in absence of oxytocin (basal secretion) or presence of oxytocin (2 x 10-^ M; stimulated secretion).

PAGE 129

' 122 Discussion In vitro studies of prostaglandin secretion from perifused endometrium and endometrial explants confirmed that PGFj^ secretion was lower in pregnant than in cyclic cows, whereas secretion of PGEj was either unaltered (Thatcher et al., 1984b; Gross et al., 1988b) or increased (Gross et al., 1988c). Incubation of endometrial explants of day 17 cyclic cows with IFNx or bovine IFN alpha inhibited PGF2„ secretion (Thatcher et al., 1989; Barros et al., 1991). Studies with these tissue models evaluated effects of pregnancy on endometrial tissue but failed to discern between epithelial and stromal cells relative to PG response. Current results indicate that treatment with bIFNx (natural or recombinant) for 24 h reduced basal secretion of PGFj,, and PGEj from primary culture of endometrial epithelial cells prepared from cyclic cows at day 15 postestrus, whereas stromalcells were not responsive to bIFNx. The variation in responsiveness to IFNx observed between cows may be a reflection of the disparity between animals to respond to conceptus signals in vivo. The present results agree with those of Salamonsen et al. (1988; 1989) and Charpigny et al. (1991), in which olFNx reduced both PGFj^ and PGEj secretion by ovine epithelial cells. The present study confirmed that epithelial cells are the major source of PGFj^ secretion, whereas stroma is the major source of PGEj (Fortier et al., 1988). The experiments reported here provide a model for the differential effect of pregnancy on prostaglandin secretion observed for perifused and explant endometrial tissue. Pregnancy or treatment with bIFNx apparently affects epithelial cells but not stromal cells. Therefore, bIFNx treatment and pregnancy attenuates overall IFNx secretion, whereas the prominent secretion of PGEj from stromal cells was not altered markedly. 'A ....,...,< • . -i''' t -

PAGE 130

123 In vivo, injection of oxytocin simulates uterine secretion of PGFj^ during the luteal phase of the estrous cycle (Lafrance and Goff 1985; Silvia and Taylor 1989). Our in vitro experiment indicates that epithelial cells from day 15 cyclic cows are responsive to oxytocin; secretion of both PGFj^ and PGEj increase significantly after administration of oxytocin. However, stromal cells were not responsive to oxytocin. Responsiveness of epithelial cells to oxytocin suggests that oxytocin receptors were present when the cell culture was initiated and were expressed in daughter cells, after 7 days of culture. The rapid increase in prostaglandin secretion within the first 30 min following addition oxytocin was not sustained between 30 and 90 min. This may reflect a rapid activation of the oxytocin transduction system followed by internalization of the oxytocin receptor. Treatment with bIFNx for 24 h attenuated both basal and oxytocin-induced secretion of both PGFj^ and PGEj from epithelial cells; reduction of prostaglandin secretion increased with increasing doses of bIFNx. The mechanisms by which bIFNx inhibits epithelial secretion of both PGF2„ and PGEj are yet to be elucidated. This reduction in prostaglandin biosynthesis could be the result of a decrease in arachidonic acid availability and/or the inhibition of PGH synthase. Membrane phospholipids are considered to be the source of arachidonic acid for prostaglandin synthesis (Dennis 1987; Dennis et al., 1993). However, low-density lipoproteins can provide cholesterolesters and phospholipids that also can be used as prostanoid precursors (Habenicht et al., 1990). Arachidonic acid can even be supplied by neighboring endothelial cells in the case of adipocytes (Richelsen et al., 1989). It is not known which lipases are involved, but it is believed widely that, in most cell types, PLA2 hydrolyzes phospholipids such as phosphatidylcholine (PC) or phosphatidylethanolamine (PE) to release arachidonic acid (Dennis et al., 1993). The addition of PLA2 to endometrial explants from pregnant cows (Day 17

PAGE 131

124 postestrus) stimulates prostaglandin secretion (Chapter 2). There are multiple isoforms of PLAj (Smith 1992) and there are no highly specific inhibitors for any of these enzymes. Recently, Norman and Poyser (1993) demonstrated that the activities of other enzymes regulating unesterified AA availability (e.g., acylCoA synthase, acyltransferase) are modulated by steroids. The diversity in the enzymes, potential sources of arachidonate precursors, and the existence of various isoforms of PLA2 illustrate the complexity of the possible regulatory mechanisms that could be involved if bIFNx regulates arachidonate availability. The possibility that the bovine embryo can regulate overall prostaglandin secretion at the level of PGH synthase has been explored. Gross et al. (1988b) described a proteinaceous factor that inhibits prostaglandin synthesis by bovine cotyledonary microsomes. This endometrial prostaglandin synthesis inhibitor (EPSI) had a higher activity at Day 17 post estrus in pregnancy compared with the estrus cycle (Basu and Kindahl, 1987b; Gross et al., 1988b). Other studies have described the presence of prostaglandin synthesis inhibitors in various systems such as human plasma (Saeed et al., 1977), human amniotic fluid (Saeed et al., 1982), bovine placenta (Shemesh et al., 1984) and ovine endometrium (Basu 1989). It seems likely that, in bovine endometrium, bIFNx (natural and recombinant) may induce EPSI in epithelial cells, leading to reduced secretion of PGFj^ and PGEj. It would be necessary to test for EPSI activity in epithelial cells treated with bIFNx to confirm this hypothesis. Bovine IFNx attenuated oxytocin-induced prostaglandin secretion. In sheep, it has been demonstrated that treatment with olFNx for several hours did not reduce oxytocin receptors, based on the continued induction by oxytocin of phosphatidylinositol turnover, as monitored by accumulation of inositol phosphates (IP3, IPj, and IP; Vallet and Bazer, 1989). However, when sheep received twice-daily intrauterine injections of ovine conceptus proteins, the

PAGE 132

125 number of endometrial oxytocin receptors decreased (Mirando et al., 1993). Abayasekara et al. (1992) suggested that the IFNx-induced decrease in oxytocin receptor expression was mediated by a decrease in PKC activity. In cattle, Mirando et al. (1990) suggested that reduced endometrial secretion of PGFj^ during maternal recognition of pregnancy may involve reduced IP turnover. In the present experimental model, endometrial epithelial cells in primary culture from day 15 of the estrous cycle contain oxytocin receptors based on their inaeased prostaglandin secretion in response to oxytocin. It is unlikely that bIFNx blocks oxytocin receptors since olFNx does not bind to oxytocin receptors in the endometrium of sheep. Thus, some alteration in signal transduction following oxytocin binding appears to be operational. The fact that bovine epithelial cells are responsive to oxytocin treatment while stromal cells failed to respond, implies that oxytocin receptors were present on epithelial, but not on stromal cells at day 15 post-estrus. This hypothesis is supported by Ayad et al. (1991), who indicated that oxytocin receptors were consistently located in endometrial epithelial cells from sheep. Furthermore, presence of oxytocin receptors on epithelial cells seems essential, since epithelial cells would be the main source of oxytocin-induced secretion of POFj^ for luteolysis. Hansel et al. (1973) reported that endometrial epithelial cell aplasia in vivo led to maintenance of CL function. Such a prolonged lifespan of the CL was likely due to absence of oxytocin receptors, which prevented oxytocin-induced PGFj^ secretion and subsequent luteolysis. The observed differential cellular responses to bIFNx treatment imply that receptors for bIFNx reside on epithelial cells, but not on stromal cells. At this time, presence of receptors for bIFNx in bovine endometrium has not been determined. Receptors for bIFNx have been detected and located on the surface of endometrial epithelium in sheep (Godkin et al., 1984a). According to

PAGE 133

V; ; 126 Knickerbocker and Niswender (1989), expression of receptors for oIFNt occurs uniformly throughout the endometrium (i.e., caruncular and intercaruncular regions of both uterine horns). However, epithelial and stromal tissues were not separated. Recombinant bovine interferon alpha (rBolFN-a|1) has been shown to bind to olFNx receptors in ovine endometrium, but binding parameters of olFN T and rBolFN-a|1 to endometrial receptors were different (Hansen et al., 1989). The natural bIFNx complex consists of several protein variants which are all glycosylated (Helmer et al., 1988; Anthony et al., 1988; Plante et al. 1990), whereas recombinant bIFNx consists of only one form which is not glycosylated (Klemann et al., 1990). Physiologically, meaningful doses of biologically active compounds are in the range of the Kq estimates for the receptors. The Kq estimates for olFNx and interferon alpha-like molecules range from 0.01 to 0.5 nM (Godkin et al., 1984; Stewart et al., 1987; Knickerbocker and Niswender, 1989; Hansen et al., 1989), which would lead to an estimated physiological dose for bIFNx of 0.2 to 1 1 .6 ng/ml (molecular weight of the two major forms of bIFNx is 23.2 kDa; Plante et al. 1990). Indeed, the minimum effective dose of rIFNx was 0.4 ng/ml and 50% inhibition was obtained with 1 ng/ml. These effective doses for inhibition of either basal or oxytocin-induced secretion of PCFj^ and PGEj from epithelial cells are within the concentration range associated with the Kq estimates for olFNx receptors in sheep endometrial tissue. Current results show a greater inhibitory response of recombinant than natural bIFNx at the physiological dose of 2 ng/ml. This difference in inhibitory activity may be due to the fact that nblFNx is a mixture of several molecular weight proteinaceous variants with possibly different inhibitory activities. Also, it is possible that nIFNx lost some activity in the purification process. This study suggests that bIFNx glycosylation is not required to bind to type I receptors and to induce inhibition of prostaglandin synthesis. There are numerous examples of conserved biological

PAGE 134

127 activities (antiviral, antiproliferative) when carbohydrates are removed from interferons (see Bocci, 1983). Interestingly, the absence of extracellular matrix in this system of primary culture did not prevent endometrial epithelial cells from responding to IFNx and oxytocin. Classical culture techniques (simple plastic or glass surface) as the one used in this experiment, failed to provide uterine responsive to steroid honnones. The development of a polarized epithelial cell phenotype is initiated by cell-cell and cell-substratum contacts (Thiery et al., 1985; Klein et al., 1988). It appears that responsiveness to oxytocin and IFNx is not (or less) affected by the lack of a basal membrane-like substratum, whereas steroid responsiveness is affected. Recently, an array of new techniques (extracellular matrix coating, permeable surfaces, dual sided chambers...) have allowed culture of functionally polarized uterine epithelial cells that are responsive to steroid hormones (Julian et al. 1992a; 1992b). Receptivity of uterine epithelial cells to embryo implantation is a prime example of the importance of cell polarization (see Glasser and Mulholland, 1993). Indeed, during the non-receptive phase, the apical domain is not adhesive and cannot be invaded by trophoblast. During the receptive phase, the apical domain is remodeled and becomes adhesive (see Denker, 1990). Cortizo et al. (1992) demonstrated on MDBK polarized cells that the direction (apical vs basal) of AA and prostaglandin release was determined by the type of agonist. Thrombin induced an apical secretion of prostaglandin, whereas bradykinin induced predominantly a basolateral release of AA and its metabolites, it would be interesting to examine if a polarized culture system alters epithelial responsiveness to IFNx and oxytocin and if stromal unresponsiveness is maintained when cocultured with epithelial cells. In conclusion, this study provides a better understanding of the roles of the two major endometrial cell types (epithelial and stromal cells) and their

PAGE 135

128 interaction with regulatory agents such as blFNx and oxytocin in cattle. Further research is necessary to elucidate the intracellular regulatory mechanisms by which bIFNx controls prostaglandin secretion by the bovine endometrium. The interactions between epithelial and stromal cells, blFNx regulation of arachidonate availability and cell compartmentalization of arachidonate, the exact nature and localization of EPS!, interaction of EPS! with PGH synthase and expression of PGH synthase(s) are fields that warrant further investigation. In the following chapter the molecular nature and the mechanism of action of the endometrial prostaglandin synthesis inhibitor will be examined. Furthermore, the ability of intrauterine infusions of bovine IFNx to increase EPSI activity will be investigated.

PAGE 136

CHAPTER 4 IDENTIFICATION AND QUANTIFICATION OF AN ENDOMETRIAL PROSTAGLANDIN SYNTHESIS INHIBITOR (EPSI) IN THE BOVINE Introduction In order to maintain pregnancy in cattle, the iuteolytic influence of the uterus must be attenuated or blocked during the usual period of luteolysis that would occur in the absence of a conceptus (Hansel et al., 1973; Thatcher et al., 1984b). Plasma concentrations of PGFM are lower during early pregnancy (Betteridge et al., 1984; Thatcher et al., 1984a; Wolfenson et al., 1985), and in vitro endometrial PGFj^ secretion is lower in pregnant than cyclic cows at day 17 postestrus (Thatcher et al., 1984b; Gross et al., 1988b). In chapter 2, the limited capacity of endometrium from pregnant cows to produce prostaglandin in the presence of exogenous AA suggested the action of an endometrial prostaglandin synthesis inhibitor or a decreased PGH synthase concentration. No differences in detectable PGHS enzyme (Salamonsen and Findlay, 1990) or mRNA (Salamonsen et al., 1991 ) were found in endometrial tissue from pregnant and cyclic ewes on day 1 5 after estrus. Consequently, the conceptus does not appear to suppress PGFj^ secretion by reducing the concentration of PGHS in endometrial tissue of sheep. The possibility that the conceptus induces an intracellular endometrial factor to inhibit prostaglandin synthesis has been explored. Wlodawer et al. (1976) showed that bovine microsomes had a relatively low capacity to convert AA into PGF2a. However, when PGH2 was substrate for the microsomal " ' * 129

PAGE 137

130 enzymes, conversion into RGFj^ was much higher. This observation suggested that PGHS activity was a limiting factor in the ability of endometrial microsomes to synthesize PGF2„, possibly due to the presence of a PGHS inhibitor. There are numerous examples of endogenous inhibitors of prostaglandin synthesis in various tissues and fluids and these may be important regulators of PG-induced events. Takeguchi et al. (1971) reported that cytosol of bovine seminal vesicles inhibited prostanoid biosynthesis by vesicular microsomes. Since then, other authors have reported PG synthesis inhibitors in sheep endometrium (Basu, 1989) and allantoic fluid (Leach-Harper and Thornburn, 1984), renal cortex (Terragano et al., 1978), placenta of the rat (Harrowing and Williams, 1977), human amniotic fluid (Saeed et al., 1982), ovarian follicular fluid of humans (Carson et al., 1986) and cows (Shemesh, 1977), and human decidua (Ishihara etal., 1990). Basu and Kindahl (1987b) detected the presence of a prostaglandin inhibitor in microsomal and cytosolic fractions of bovine endometrium extracts from day 16-31 of pregnancy, with highest inhibitory activity occurring on day 18 and lowest on day 25 postestrus. However in endometrium from cyclic cows, inhibitory activity was lowest on day 17. The capacity of microsomes from bovine endometrium to convert exogenous AA to prostaglandins was 50% lower than in microsomes from cyclic animals (day 17 postestrus). When comparing the inhibitory potency of this endometrial factor, as calculated by IC50 values (inhibitory concentration resulting in 50% inhibition), the potency was about eight times higher for day-17 pregnant than day 17 non-pregnant endometrial cytosolic fractions (Basu and Kindahl, 1987b). Gross et al. (1988a) examined the ability of endometrial cytosol and microsomes, from day 17 cyclic and pregnant cows, to modulate prostaglandin synthesis by cotyledonary

PAGE 138

131 microsomes from parturient cows. The endometrial prostaglandin synthesis inhibitor (EPS!) appeared to be present in much higher amounts in pregnant cows than cyclic cows, to be primarily cytosolic, proteinaceous, and to act as a non-competitive inhibitor with regard to AA metabolism (Gross et al., 1988a). Bovine IFNx induced EPSI activity when administered to endometrial explant cultures from cyclic cows at day 17 postestrus (Helmer et al., 1989b). The objective of this study was to identify the factor(s) responsible for the prostaglandin synthesis inhibitory activity, and quantify this factor(s) in endometrium from pregnant and cyclic cows at day 1 7 postestrus. Materials and Methods Animals Non-lactating cyclic Holstein dairy cows (n = 12) were synchronized for estrus, using Syncro-Mate-B ear implants (Sanofi Animal Health Inc., Overland Park, KS) for a period of 9 days and 25 mg Lutalyse (The Upjohn Company, Kalamazoo, Ml) injected i.m. 48 h prior to removal of the implants. Cows were observed for estrous behavior and either bred twice (12 h apart; n = 9) by artificial insemination or not bred (cyclic, n = 3) . Cows were slaughtered at Day 17 postestrus. Reproductive tracts were collected 15 min after slaughter within the abattoir. Tracts from bred cows were flushed with MEM (40 ml; Sigma Chemical Company, St. Louis, MO) to collect conceptus tissue and confirm pregnancy. Reproductive tracts from bred but non-pregnant cows were not included in this study. Endometrial tissues from a total of 4 pregnant and 3 cyclic cows were processed. The uterine horn ipsilateral to the CL was opened longitudinally along the antimesometrial border, and endometrial tissue was

PAGE 139

132 dissected free from myometrium over the entire length of the uterine horn. Tissues were frozen at -70°C until further analysis. Preparation of Bovine Endometrial Tissue Endometrial tissues from cyclic and pregnant cows were washed three times in 100 ml of 0.05 M Tris-buffered 0.25 M sucrose, drained, finely chopped with scissors and added to 0.05 M Tris-buffered 0.25 M sucrose (40 g/ 80 ml), pH 7.4 and homogenized over ice using a Polytron (Brinkmann, Luzern, Switzerland) homogenizer for three 1 0 second bursts. The homogenates were centrifuged (9000 x g) for 20 min at 4°C to pellet tissue fragments, mitochondria and nuclei. Supernatants were centrifuged (100,000 x g) for 60 min at 4°C to yield microsomal (pellet) and cytosolic (supernatant) fractions. Microsomes were resuspended in 0.1 M potassium phosphate, pH 7.4 (10 ml). Endometrial microsomes and cytosol were stored at -70°C in aliquots of 1 ml and 40 ml, respectively. Preparation of the Cotvledonarv-Prostaqlandin-Generatinq System Placentomes were collected from parturient dairy cows within 15 min following a non-induced calving. Placentomes were placed on ice and fetal cotyledonary villous tissues were manually peeled away from caruncles and stored at -70°C in 40 g aliquots. Cotyledonary tissues were homogenized, and then subjected to differential centrifugation as described for the endometrial tissue to yield microsomal and cytosolic fractions. The cytosol was discarded and microsomes, containing the prostaglandin generating enzymes, were resuspended in 0.1 M sodium phosphate, pH 7.4 (10 ml) and stored in 1 ml aliquots at -70°C. ..^ i ' ' i

PAGE 140

, ; ; 133 >.•.. ' ' Assay for the Inhibition of Prostaglandin Synthesis Cotyledonary microsomes (10 |xl in 0.1 M potassium phosphate; pH 7.4; 40 mg tissue equivalent) were incubated with 160 nl of sample (endometrial cytosol or microsomes or HPLC fraction) and 40,000 dpm of [5,6,8,9,12,14,153H] arachidonic acid (specific activity of 204 Ci/mmol; Amersham Corp., Arlington Heights, IL) diluted in 50 |j,l of resuspension buffer (0.2 M sodium phosphate containing 0.1% [w/v] bovine BSA, fatty acid free; Sigma Chemical Company, St. Louis, MO). Incubations were conducted at 38°C for 1 h. Flunixin meglamine (Banamine; 160 ^1 of a 1 mg/ml solution; Schering Corp., Kenilworth, NJ), a potent cyclooxygenase inhibitor, was utilized as a positive control for inhibition of prostaglandin synthesis, and 0.2 M potassium phosphate (160 nl) was utilized as a negative control. Incubations were terminated with addition of 125 |al of acetonitrile, and each sample was vortexed and centrifuged at 12,000 x g for 5 min. Supernatants were transferred into glass vials for analysis. Tritiated arachidonic acid (substrate) and its tritiated metabolites were separated using a HPLC system consisting of a Series 4 Liquid Chromatography Microprocessor Controlled Solvent Delivery System (Perkin-Elmer; Norwalk, CT), a series 4 Control Module to control the delivery system (Perkin-Elmer), an LKB 2140 Rapid Spectral Detector (LKB; Bromma, Sweden), a Flo-One Series A-200 in-line p detector (Radiomatic, Tampa, FL), and a LKB 2211 SupeRac fraction collector custom modified to include refrigerated collection tray. A Whatman 0,8 (Parsitil 5 ODS-3; 4.6 x 250 mm; 5 ^im particle size; 10.5% carbon load) column was used for separation of AA and its metabolites. Elution was accomplished with a mobile phase of 45% acetonitrile, 2% glycerol, 0.05% trifluoroacetic acid (TFA) and 52.95% water at a flow rate of 1 ml/min. The array of (3H)-AA metabolites was examined first using an extended elution scheme:

PAGE 141

134 samples were eluted isocratically for 5 min then the original mobile phase was changed to 100% acetonitrile over a 15 min period followed by a 10 min equilibration time (Figure 4-1). For each sample of supernatant from the EPS! activity assay, an aliquot of 100 nl was injected into the HPLC system. The column was calibrated using (3H)-AA, (3H)-PGF2„, (3H)-PGE2 and (3H)-PGFM standards. After identification of the major (3H)-AA metabolite as being (^H)PGF2„, a compressed elution scheme of 15 min was developed in order to analyze samples more rapidly (Figure 4-1, panel D). Samples were eluted isocratically for 4 min then the mobile phase was stepped to 100% acetonitrile for 5 min followed by 6 min of equilibration time. Radioactivity was monitored using a Radiometric p-detector in line with the HPLC system. An arbitrary unit of inhibition was created in order to compare inhibitory activities between samples. This unit is directly proportional to the percentage of counts found at the (3H)-AA retention time (Figure 4-1 ). The minimum AA peak area (Z% of total cpm injected) was obtained with the buffer control (no inhibition. Figure 4-1 panel A). The maximum AA peak area (Y% of total cpm injected) was obtained with the Banamine control (total inhibition; Figure 4-1 panel B). A sample with intermediate inhibitory activity generated an elution profile with X% of total cpm (Z < X < Y) in the AA peak area (Figure 4-1 panel C). For each sample, the |3detector software (Radiomatic) provided the number of cpm in the AA peak and the total number of cpm eluting off the CI 8 column. The AA peak was characterized as a percentage of the total number of cpm detected in order to eliminate the slight variation in total cpm between samples of the same assay. Units of inhibition (Ul) were expressed as: Ul = { ( [X Z] / [Y Z] ) X 100 } / mg protein in the sample ,

PAGE 142

135 Figure 4-1. Representative chromatograms of radiolabeled arachidonic acid (PH]-AA) and its metabolites eluted on a CI 8 HPLC column following conversion of (3H)-AA by cotyledonary microsomes in presence of: sodium phosphate buffer (no inhibition; panel A), Banamine (control for maximal inhibition; panel B) and sample (partial inhibition; panel C). Panel D represents a compressed (15 min) elution profile of the same sample represented on panel C (25 min elution scheme). For each sample, the AA peak was characterized as a percentage (AA = X or Y or Z%) of the total number of cpm detected during the entire elution time.

PAGE 143

136 Titration Serial dilutions (160, 100, 50, 25, 10, 5, 2.5 and 1 i^l) of endometrial cytosol from pregnant and cyclic cows were evaluated for their inhibitory activities. Each dilution was run in duplicate and protein concentrations were determined by the bicinchoninic acid (BCA; Pierce, Rockford, IL) protein assay (Smith et al., 1985), utilizing BSA as standard. Units of inhibition were computed for the volumes of endometrial cytosol that induced approximately 50% inhibition (e. g., AA peak surface approximately midway between Banamine and buffer control, as seen on Figure 4-1 panel C and D). Purification of EPSI Fractionation of endometrial cytosol from pregnant cows using Centricon membranes . The pooled cytosols were processed first with the Centricon-1 00 (C-100; Amicon, Danvers, MA) ultrafiltration device to remove high molecular weight compounds (> 80-100 kDa). Two milliliters of cytosol was placed in each Centricon unit and centrifuged at 4,000 x g for 90 min (4°C). After each centrifugation, the filtrate was collected and concentrated using a Centricon-1 0 (C-10) ultrafiltration device (4,000 x g for 90 min at 4°C) to remove low molecular weight compounds (< 10-8 kDa). Retentates («1 ml) from each C-100 and C-10 units were collected after five consecutive loadings (10 ml of cytosol processed) by inversion of the ultrafiltration units and centrifugation at 500 x g for 5 min at 4° C. Integrity of the ultrafiltration membranes was monitored throughout the centrifugation process by periodically (every 30 min) comparing retentate volumes in all centrifuged devices. C-100 retentate, C-10 retentate and filtrate protein concentrations were determined by BCA protein assay and each was tested for their PG synthesis inhibitory activity.

PAGE 144

137 Endometrial Cytosd Day 17 Post-Estnjs i Concentration (Centricon 10and 100) I ClOORetentate I HPLC Anion Exchange ^ see Figure 4^. Mono Q (HR 5/5, Pharmacia) 20 mM Tris, 2.5 % NaCI / min, pH 7.5, 1 ml/min. 1 Concentration of active fractions \ HPLC Gel Filtration »see Figure 4-7. GF250 (ZbrtDax), 0.2 M Sodium Phosphate, 6% Glycerol. pH 7.5, 0.5 ml/min. Concentration of active fractions I HPLC Reverse Phase *see Figure 4-8. C4 column (300 A, Vydac) 35% AC N, 20% water, 5% glycerol. 0.1% TFA, 1.5 ml/min. I 2 active fractions in BSA region Westem Blot for BSA ^ see Figure 4-10. Figure 4-2. Purification scheme used to isolate the proteinaceous factor associated with EPS! activity.

PAGE 145

I: ^ V . . 138 Anion exchange HPLC . In order to be compatible with the anion exchange mobile phase, the active material (C-100 retentate) was added to three volumes of Tris-HCI buffer (20 mM; pH 7) and filtered through a C-100 device (4,000 x g for 90 min). After centrifugation of washed retentate at 10,000 X g for 5 min to pellet any particulate material, the material was loaded (maximum of 500 |il/injection) on a Mono Q HR 5/5 anion exchange column (5 x 50 mm, Pharmacia Inc., Pistakaway, NJ). Separation was accomplished with a linear gradient of 0-200 mM NaCI over a 25 min period, in 20 mM Tris-HCI buffer, pH 7, with a flow rate of 1 ml/min. The column was calibrated using myoglobin (pi 7.33), carbonic anhydrase (pi 6.18), BSA (pi 4.9) and pepsin (pi 2.9). Ultraviolet light absorption at 280 nm was monitored and 30 sec fractions (0.5 ml) were collected in an ice bath. For each fraction, protein concentration was assayed and EPSI activity was tested. Gel filtration HPLC . The EPSI active fractions from the Mono Q column were pooled and concentrated using C-100 ultrafiltration units. Retentate and filtrate were tested for inhibitory activity. C-100 retentate buffer was changed to sodium phosphate (0.2 M) and centrifuged for 5 min at 10,000 x g. The material was loaded on a Zorbax Bioseries GF-250 (9.4 x 250 mm, Dupont, Wilmington, DL) gel filtration column. The column was calibrated using BSA (69 kDa), ovalbumin (45 KDa), carbonic anhydrase (29 kDa), myoglobin (18.5 kDa) and sodium azide (salt volume). A maximum of 195|al was applied per injection, the flow rate of eluent (0.2 M sodium phosphate buffer, pH 7.5, 3% glycerol) was 0.5 ml/min and absorbance at 280 nm was monitored. Fractions (30 sec) were collected on an ice bath, analyzed for protein concentration and ability to inhibit prostaglandin synthesis (EPSI activity).

PAGE 146

: 139 Reverse phase HPLC . The active fractions from the gel filtration column were pooled and concentrated using C-100. Retentate and filtrate were tested for inhibitory activity. C-100 retentate was washed with three volumes of deionized water to remove sodium phosphate salts, and centrifuged for 5 min at 10,000 X g. The material was further purified using a reverse phase C4 column (2.1 X 150 mm, 5 |xm, Vydac), a mobile phase of 35% acetonitrile, 5% glycerol, 0.1% TFA, 59.9% water at a flow rate of 1 ml/min. A maximal volume of 100 nl was applied per injection. Separation was accomplished with a linear gradient to 90% acetonitrile, 0.025% TFA and 1 % glycerol over 30 min. Absorbance at 280 nm was monitored, 30 sec fractions were collected in an ice bath and partially evaporated with a Speed-Vac system in order to eliminate acetonitrile. The remaining TFA in each fraction was neutralized with 10 )il of 2 M NaOH solution in order to be compatible with the prostaglandin generator (cotyledonary microsomes). Fractions were tested for protein concentration and EPS! activity. Identification of the Active Proteinaceous Factor Immunoblottinq for BSA . Based on retention times (Figures 4-3; 4-4; 4-5) and absorption profiles (Figure 4-6) of the active fractions from anion exchange, gel filtration and reverse phase columns, the active factor appeared to be associated with BSA. Aliquots of C-100 (20, 15, 10, 5 jig) and C-10 retentates (10, 5, 2, 1 ng) and active fractions from gel filtration and C4 reverse phase were analyzed by ID SDS-PAGE to examine BSA content of the active fractions. After electrophoresis, proteins were Coomassie stained or transferred electrophoretically to 0.2 ^m nitrocellulose paper using 100 V for 45 min at 4°C in an electroblotting chamber containing 25 mM Tris, 0.2 M glycine and 20% (v/v) methanol. Each blot was analyzed for BSA by Western blotting. After 20 i ' f •'. J : ; ^ ; _ ) , .IS'1._ -

PAGE 147

140 min at room temperature in blocking solution (2% dried milk solution in 20 mM Tris-HCI, pH 7.4), blots were washed in 20 mM Tris-HCI, 0.1% Triton-X saline solution three times and then incubated for 1 h at room temperature with either normal rabbit serum or anti-BSA (1:416 dilution, polyclonal, Sigma) diluted in dried-milk solution. Blots were washed for 30 min at room temperature and incubated for 1 h in a goat anti-rabbit alkaline phosphatase conjugate solution (1:1000 dilution). Blots were washed five times with 20 mM Tris-HCI, 0.1% Triton-X saline solution over a 30 min period and incubated in color development mixture (40 ml 200 mM Tris, pH 8.8, 160 nl 1 M MgClj, 2 mg 5-bromo-4-chloro3-indolyl-phosphate and 4 mg nitro blue tetrazolium) for 5 to 15 min. Finally, blots were rinsed in deionized water at the time of optimal color development. Immunoprecipitation . To determine if BSA found in cytosol was synthesized within the endometrium, endometrial explants (2 x 500 mg per cow) from four Holstein cows at day 1 7 of pregnancy were incubated for 24 h in 1 5 ml leucine-deficient MEM supplemented with 50 i^Ci (3H)-leucine. Endometrial explants were rocked under an atmosphere of 50:45:5 (v/v/v) Nj: ©2:002. Medium was collected after 24 h and replaced. After the second 24 h incubation period, culture medium and endometrial tissue were collected and frozen at -70° C until further analysis. Each endometrial explant was processed using a Polytron homogenizer for a single 10 sec burst and homogenates were centrifuged at 12,000 x g for 10 min in a microcentrifuge. For each cow, two samples of (3H)-labeled culture medium and (3H)-labeled endometrial homogenate containing 200,000 dpm were dialyzed against water, lyophilized and redissolved in immunoprecipitation buffer (50 mM Tris-acetate, pH 7.5, 1 mM PMSF, 1 mM EDTA, 0.3 M NaCI, 2% NP-40 [v/v], 0.02% NaNa). Fifty microliters of undiluted rabbit anti-BSA (polyclonal.

PAGE 148

141 Sigma) or NRS were added and samples were incubated overnight on a tube turner at 4°C. Subsequently, 100 ^il Protein-A Sepharose (10% [v/v] in immunoprecipitation buffer) were added and incubated for 6 h at room temperature on a tube turner. The Sepharose suspension was centrifuged (12,000 X g for 1 min) and washed four times in 1 ml washing buffer (50 mM Trisacetate, pH 7.5, 1 mM PMSF, 0.1% SDS, 0.3 M NaCI, 0.5% NP-40 [v/v], 0.02% NaNg). Antibody-antigen complexes and a BSA control solution were then dissociated and denatured by boiling for 3 min with 100 i^l of loading buffer (62.5 mM Tris-HCI, pH 6.8 containing 5% SDS, 10% sucrose and 5% pmercaptoethanol). Tubes with Protein-A Sepharose were centrifuged (12,000 x g for 1 min) and supernatant analyzed by 1-D SDS-PAGE (12% polyacrylamide) using the buffer system of Laemmli (1970). Radiolabelled proteins were detected by fluorography (up to 4 weeks of exposure) after soaking in 1 M sodium salicylate and drying them. Titration of BSA inhibitory activity . The ability of BSA to inhibit prostaglandin synthesis was examined by adding 500, 250, 125, 62.5, 31.25, 15.6, 7.8 and 3.9 iig (in 160 i^l of 0.2 M sodium phosphate solution) of purified BSA (fatty acid free, Sigma) to the prostaglandin generator (cotyledonary microsomes). BSA activity was compared to pooled endometrial cytosol from day 17 pregnant cows (5500, 2750, 1375, 550, 275, 137.5, 55, 27.5, 13.7 and 5.5 ng of protein in 160 and to p-lactoglobulin (500, 250, 125, 62.5, 31.25, 15.6, 7.8 and 3.9 pig of protein in 160 as a protein control) for EPSI activities . Purification of a Lipophilic Factor Lipid extraction . Since EPSI activity resided with BSA, it was hypothesized that a lipophilic factor, bound to BSA in cytosol, interacted with the

PAGE 149

142 cotyledonary microsomes (prostaglandin-generating system) to inhibit PG synthesis. To examine this possibility, 1 ml of endometrial cytosol from a day 17 pregnant cow was extracted according to a modified "Folch" procedure (Christie, 1985) in which cytosol was homogenized with methanol (50 ml) for 1 min. Endometrial Cytosol Day 17 Post-estms Folch's Extraction ^ see Figure 4-12. Chloroform , Methanol (2:1) Thin Layer Chromatography Hexane, Ethyl-Ether, Acetic Acid (80:20:1) Test polar lipid, Cholesterol, Free Fatty Acid and Triglyceride regions for Epsi activity Free Fatty acids HPLC using FFA HP column (4 |im, 4% phenyl load, Waters) 45% AC N, 20% THF, 35% water, 1.5 ml / min and laser light-scattering detector see Table 4-2. see Figure 4-13 A. 2 consecutive fractions were active m see Figure 4-13 B. Methylatlon Gas Chromatography Carbowax (30 m x 0.25mm i.d., Alltech), see Figure 4-14. 200 to 280 C in 16 min, then 280 C for 10 min. Figure 4-3. Purification scheme used to isolate the lipophilic factor responsible for EPSI activity in cytosol lipid extract.

PAGE 150

Chloroform (100 ml) was added and homogenization continued for 2 min. A solution of 0.88% (w/v) potassium chloride in water (10 ml) was added and the mixture shaken thoroughly for 1 min before being allowed to settle. The lower phase (chloroform) was collected and the upper phase (methanol and water) was mixed again with fresh chloroform (100 ml). This was repeated 3 times and the combined lower phases, containing the lipids, were evaporated using a Kuderna Danish concentrator (Fisher). This type of glass evaporative concentrator with solvent reflux is used for concentration of trace amounts of sample dissolved in organic solvents. The cytosolic extract was resuspended in 1 ml of chloroform. Inhibitory activities of unextracted cytosol and cytosol extract were compared by adding 160, 100, 50, 25 12.5, 6.25 |al of cytosol or cytosol extract (chloroform was evaporated and extract was resuspended in 160 [i\ of resuspension buffer) to the PG generator to determine EPS! activity. Verification of EPSI activitv in lipid extract bv radioimmunoassay . The detection assay for inhibitory activity is based on monitoring the metabolism of (3H)-AA into (3H)-metabolites by cotyledonary microsomes. Of concern was the possibility that endogenous AA in the cytosolic lipid extract was responsible for the apparent inhibitory activity. Indeed, large amounts of AA in the extract would compete with the radiolabelled AA for microsomal cyclooxygenase and generate an isotopic dilution. This decrease in conversion of (3H)-AA could have been misinterpreted as increased EPSI activity. In order to rule out this possibility, cotyledonary microsomes (10 |il) were incubated for 1 h at 38°C with non radiolabelled AA (32 pg and 500 ng in 50 \x\ of resuspension buffer) and cytosolic lipid extract (0 and 160 \x\ in resuspension buffer). Banamine (160 [i\ of a 1 mg/ml solution) was used as a positive control for inhibition. After incubation, experimental samples were centrifuged at 12,000 x g for 10 min.

PAGE 151

144 supernatants and lipid extract (160 [i\ cytosol equivalent) were assayed for PGFja by radioimmunoassay as described in chapter 3. ' Thin layer chromatooraphv . It was clear that the lipid extract contained 75% of the EPS! activity in endometrial cytosol. In order to determine the nature of the lipid responsible for inhibition of prostaglandin synthesis, the lipid classes contained in the endometrial extract were separated by thin layer chromatography. Plates were activated by heating at 70°C overnight and then allowed to cool prior to use. Lipid extracts (4 lanes, 20 ^1 chloroform extract per lane equivalent to 5 ml of cytosol) were applied to 20 x 20 cm plates coated with 250 |im Silica Gel G (Fisher Scientific Co.). Lipids were separated using hexane, ethyl ether and acetic acid (80:20:1 by volume, 40 min development). The position of triglycerides (TG), cholesterol, and free fatty acids (FFA) was ascertained by spraying lanes containing standards (tripalmitin [TG], cholesterol, margaric acid [FFA]; Sigma), located on both sides of the plates, with dichlorofluorescein (0.1% in methanol), and standards were visualized under UV light. According to the Rf values of the standards, sample lanes (not exposed to dichlorofluorescein) were divided into four regions: origin (containing polar lipids), cholesterol, FFA and TG. Each region and a blank lane were scraped, and the silica was extracted using chloroform-methanol (95:5 by volume, 3x2 ml). The extracts were dried under nitrogen and resuspended in 160 jil of resuspension buffer to be tested for EPSI activity. Separation of free fatty acids bv HPLC . Only the FFA region totally inhibited prostaglandin synthesis. The array of FFA contained in the active TLC region were separated using HPLC with a Free Fatty Acid column (3.9 x 150 mm, 4 nm, 4 % phenyl load. Waters, Millipore, Milford, MA). An isocratic elution scheme was used with a mobile phase composed of 45% acetonitrile, 20%

PAGE 152

145 tetrahydrofurane and 35% water at a flow rate of 1.5 ml/min. Column was calibrated using individual and mixed FFA standards (14:0, 16:0, 18:0, 20:0, 22:0, 24:0, 18:1(o9,18:2©6, 18:3co3, 20:1 ©9, 20:2co6, 20:3co6, 20:4©6, 22:1co9 Nu Check Prep Inc., Elysian, MN). Because of their low absorption in the ultraviolet band, a laser light-scattering detector (Varex ELSD IIA) was used to monitor the fatty acids eluted. The detector was operated at a nitrogen flow of 55 mm and a gas pressure of 24 psi. The drift tube was heated to 135°C and the exhaust gas temperature was maintained at 73°C. Utilizing a splitter 30% of the eluate was directed into the light scattering detector and, 70% of the eluate was collected in 10 sec fractions. Each fraction was evaporated with a Speed-Vac system and resuspended in 160 |il of resuspension buffer to be tested for its ability to inhibit prostaglandin synthesis. Identification of EPS! by gas chromatography . The HPLC fraction with the most inhibitory activity was methylated using boron trifluoride catalyzed esterification method (Morison and Smith, 1964) for analysis by gas chromatography (GC). HPLC active fractions were placed into leak proof Teflonlined screw cap tubes; 2 ml of a 6% BF3 solution (in methanol) and 2 ml of benzene were added. Tubes were flushed with nitrogen, capped and placed on a boiling water bath for 15 min. After cooling, methyl esters were extracted by adding 4 ml water and 2 ml hexane. The hexane extraction was repeated, pooled organic fractions were evaporated using a stream of nitrogen and redissolved in 10 nl of chloroform prior to GC. Methyl esters of fatty acids were identified and quantified using a Shimadzu Model 14A gas chromatograph equipped with a split-splitless injector and flame ionization detector. A DB-Wax WCOT column (0.25 mm x 30 m, J&W, Folsom, CA) was used and operated at a helium pressure of 1.0 kg/cm^. Column temperature was held at 170°C with

PAGE 153

146 injector and detector at 300°C. The split injector was operated at a split ratio of 40:1. Methyl esters of fatty acids were identified by comparison of retention times with standards (Nu Check Prep Inc., Elysan, MN). Relative concentrations were determined by coinjection of an internal standard (17:0; 1 mg/ml). Kinetics of Inhibition Analysis by GC indicated that active HPLC fractions were highly enriched in linoleic acid (18:2(d6, LA). The mechanism of inhibition of AA metabolism by linoleic acid was examined using the prostaglandin generating system (cotyledonary microsomes). First, inhibitory activity of increasing amounts of linoleic acid was determined. (3H)-AA (80,000 dpm, specific activity of 204 Ci/mmol, Amersham Corp; 32 pg) was diluted with 500 ng of non-radiolabelled AA (in 50 [i\ of resuspension buffer) and incubated with various amounts of LA (0, 125, 250, 500, 1000, 2000 ng in 160 ^il) and cotyledonary microsomes (20 ul) for 1 h at 38°C. Linoleic acid was mixed first with microsomes, vortexed for 5 sec, then AA was added and the mixture was vortexed again for 5 sec. Each sample was run in duplicate. To determine the type of inhibition exerted by LA on the cyclooxygenase in the microsomes, metabolism of AA across time was characterized in presence or absence of LA. Various amounts of AA (500, 1000, 1500, 2000 ng; combined with (3H)-AA (80,000 dpm [64 pg] in 50 [i\ of resuspension buffer) were incubated with cotyledonary microsomes (20^1) and linoleic acid (0 and 500 ng). Incubation at 38°C was terminated after 10, 20, 30, 40, 50 or 60 min by addition of acetonitrile (125 |il). Each time point was run in duplicate. The AA metabolism rates (ng/min) were determined for the various AA and LA concentrations (4x2 slopes). The slope values (velocity in |iM/min) in absence

PAGE 154

147 or presence of LA were placed on a double-reciprocal plot (X axis = 1/[AA] in nM and Y axis = 1/V in nM/min) and the equations of the two regression lines were determined. Position of the point of intersection of the two regression lines (+ or LA) on the X (non-competitive inhibition) or Y axis (competitive inhibition) was indicative of the type of inhibition exerted by linoleic acid. EPSI Activity and Quantification of Free Linoleic and Arachidonic Acids in Endometrium from Pregnant and Cyclic Cows Holstein cows were observed for estrus and either bred by artificial insemination (pregnant, n = 4) or not bred (cyclic, n = 3). Cows were slaughtered at Day 17 postestrus. All reproductive tracts were removed rapidly (within 30 min after stunning) and flushed with 40 ml MEM to collect conceptus and confirm pregnant or cyclic status. Endometrium from the uterine horn ipsilateral to the CL was isolated from myometrium and frozen at -7G''C until further analysis. EPSI activity . For each cow, endometrial cytosol (80 ml) and microsomes (8 ml) were prepared as described above (page 132). Aliquots of cytosol (10 ml) and microsomes (1 ml) were taken to analyze for EPSI activity and protein concentration. Microsomes were placed in boiling water for 10 min before analysis for EPSI activity in order to denature microsomal enzymes that could interfere with the assay system (cotyledonary microsomes). Different volumes (160, 100, 50, 25, 10, 5, 2.5 and 1 ^1) of cytosol and boiled microsomes were tested for EPSI activity. LA and AA quantification . Lipids were extracted according to the method of Bligh and Dyer (1959). Cytosol (70 ml), microsomes (remaining 7 ml diluted to a final volume of 70 ml with 0.05 M Tris-buffered 0.25 M sucrose, pH 7.4) and plasma (12 ml diluted to a final volume of 70 ml) were blended for 4 min in a

PAGE 155

" 148 Waring blender with 100 ml chloroform and 200 ml methanol. The mixtures were filtered through a Buchner funnel using filter papers (Whatman). The material retained on top of the filters and filter paper were added back to the blender and rehomogenized with 100 ml chloroform. The mixture was filtered again and the two filtrates were combined and transferred to a 1 liter separatory funnel. One hundred ml of 0.88% KCI aqueous solution (w/v) was added, the solution was shaken for 30 sec and allowed to settle. The lower layer (chloroform) was drained and allowed to percolate through a filter paper cone filled with several grams of sodium sulfate in order to remove any trace of water. The filtrate was collected in a round bottom flask, and the solvent was removed using a rotary evaporator (water temperature 60°C). The lipid extracts were clear after being redissolved in 10 ml of chloroform. Aliquots of microsomal (4 of 10 ml), cytosolic (3 of 10 ml) and plasma (3 of 10 ml) extracts were evaporated under nitrogen and redissolved in 250 |il of chloroform for separation of the different lipid classes on Silica Gel G thin layer chromatography plates as described previously (page 144). After extraction of the free fatty acid region from the silica gel using chloroform methanol (95:5, by volume, 3x2 ml), extracts were dried under nitrogen. Free fatty acids were methylated using the boron trifluoride catalyzed esterification method (Morison and Smith, 1964) described earlier (page 145). After extraction and evaporation of the solvent under nitrogen, methylated FFA from cytosol, microsomes and plasma were dissolved in different volumes of chloroform (100, 20 and 10 jxl, respectively) in order to obtain comparable FFA concentrations and therefore comparable peak heights on the gas chromatograms. An internal standard (margaric acid methyl ester, 17:0) was added to each sample at a final concentration of 1 mg/ml to determine FFA concentrations. Fatty acid methyl

PAGE 156

esters were separated and quantified using a Perkin Elmer Sigma 3B gas chromatograph equipped with a split-splitless injector and flame ionization detector. A Carbowax (30 m x 0.25 mm, Alltech, Deerfield, IL) column was operated at a helium pressure of 16 psi. Column was held at a temperature of 200°C and temperature was raised to 280°C at the rate of 5°C per min. The split injector was operated at a ratio of 40:1. Fatty acids were identified by comparison of retention times with standards (14:0, 16:0, 18:0, 20:0, 22:0, 24:0, 14:1 ©5, 16:1 ©7, 18:1©9,18:2©6, 18:3o)3, 20:1 ©9, 20:2©6, 20:3©6, 20:4©6, 22:1© 9; 24:1 ©9, 22:6©3; Nu Check Prep Inc., Elysian, MN). Linoleic and arachidonic methyl esters were quantified by area of each respective peak given by the integrator (HP 3390A; Hewlett Packard). The mass of fatty acid represented in each peak was proportional directly to the known mass represented by the internal standard (17:0, 1 mg/ml) peak. The total amounts of unesterified LA and AA were calculated back to the entire cytosoiic and microsomal fractions for each experimental endometrial tissue and expressed as ng/mg protein. EPSI Activity and Quantification of Free Linoleic and Arachidonic Acids in Endometrium from Recombinant IFNx-Treated and Control Cyclic Cows In vivo treatments . Non-lactating cows were synchronized for estrus with Norgestomet implants for 7 days and PGFj^ injections one day before implant removal. On days 9-10 of the estrous cycle, all cows received sterile plastic catheters (Tygon AAQ04103; inside diameter 0.51 mm; Norton Performance Plastics, Akron, OH) into the tip of each uterine horn by mid ventral laparatomy as described by Knickerbocker et al. (1986b). Catheters were inserted approximately 45 mm into the uterine lumen. Catheters were exteriorized via a small incision in the flank, wrapped in gauze soaked in iodine solution, and placed in a plastic bag that was sutured to the skin. Intrauterine infusions (0.8

PAGE 157

150 ml) contained either 0.2 mg of recombinant IFNi in 1.3 mg BSA (treatment, n = 5 cows) or 1.5 mg BSA (control, n = 5 cows) in 20 mM Tris-HCI-150 mM NaCI buffer (pH 8.0) containing penicillin (100 U/ml) and steptomycin (100 ng/ml). Infusions were given twice daily (0700 and 1900 h) from day 14 to day 17 postestrus. At each administration, the infusate (0.8 ml) was split evenly between each uterine horn. At approximately 0800 h on day 18 postestrus, cows were sacrificed to collect uteri. One uterine horn from each cow was prepared to quantify free linoleic and arachidonic acids in both endometrial cytosol and microsomal fractions. Uterine horns used for analysis were balanced relative to location either ipsilateral or contralateral to the ovary bearing the CL. The procedure for quantification of EPS! activity and of free LA and AA in endometrium was identical to that used for cyclic and pregnant cows. Statistical Analysis Data were analyzed statistically using least squares analysis of variance in the General Linear Models procedure of the Statistical Analysis system (SAS, 1985). The effects of reproductive status (cyclic and pregnant; status) on fatty acids (linoleic and arachidonic acids; FA) concentrations in microsomal and cytosolic fractions (location) of endometrial tissues were examined using the model components listed in Table 4-1. For EPS! activity, LA, AA and the ratio of (LA:AA), the model components were status, location, status x location, cow (status), cow (status) x location. The effects of reproductive status (or IFNx treatment) on fatty acid (FA; LA and AA) concentrations in plasma were examined using the model components: status, cow (status), FA, status x FA.

PAGE 158

151 The effects of in vivo treatments (rblFNx vs control) were analyzed utilizing the model components of treatment, cow (treatment), location, (cytosol vs microsomes), treatment x location, location x cow (treatment), fatty acid (LA vs AA), treatment x fatty acid, and fatty acid x cow (treatment), location x fatty acid and treatment x location x fatty acid (Table 4-2). For each fatty acid and the ratio of (LA:AA), the model components were reduced to treatment, cow (treatment), location, treatment x location, location x cow (treatment). The effects in vivo treatment with IFNx or BSA on fatty acid (FA; LA and AA) concentrations in plasma were examined using the model components: treatment , cow (treatment), FA, treatment x FA. Effects were tested using appropriate error terms according to the expectations of mean squares. Table 4-1 . Least squares analysis of variance of free fatty acid (linoleic and arachidonic; FA) concentrations in endometrial microsomes and cytosol (location) from cyclic and pregnant (status) cows at day 17 postestrus. Source df concentration SS Error term status 1 56.3 cow (status) cow (status) 5 261626.9 residual location 1 924960.0** cow (status) X location status X location 1 6762.6 cow (status) X location cow (status) X location 5 178316.6 residual FA 1 16578.5 cow (status) X FA status X FA 1 265736.2** cow (status) X FA cow (status) X FA 5 88317.7 residual location x FA 1 52966.0 residual status X location x FA 1 152312.1* residual residual 5 72102.0 P<0.01; ** P<0.05

PAGE 159

152 Table 4-2. Least squares analysis of variance of free fatty acid (linoleic and arachidonic; FA) concentrations in endometrial microsomes and cytosol (location) from cyclic cows treated in vivo with two daily intrauterine infusions of BSA or recombinant bovine IFNx (treatment). Source df concentration SS Error term treatment 1 32201.2 cow (treatment) cow (treatment) 8 109601.3* residual location 1 333282.7** cow (treatment) x location treatment x location 1 42130.1 cow (treatment) x location cow (treatment) x location 7 113750.3* residual FA 1 7095.1 cow (treatment) x FA treatment x FA 1 1958.2 cow (treatment) x FA cow (treatment) x FA 8 17822.9 residual location x FA 1 4613.7 residual treatment x location x FA 1 800.9 residual residual 18 19916.4 P<0.01; ** P<0.05 Results The ability of cotyledonary microsomes to convert tritiated AA into metabolites provided a rapid, reliable and sensitive detection system for monitoring and measuring EPSI activity. The use of a C^q HPLC column in combination with extended elution scheme (30 min) indicated that the major metabolite of (3H)-AA synthesized by cotyledonary microsomes under the conditions of this assay was (3H)-PGF2„ (Figure 4-1 ). A decrease in surface area of the (3H)-AA peak was paralleled with an increase in surface area of the (3H)-PGF2„ peakFor routine screening of samples, a shorter elution scheme (15 min) was used to monitor variations in AA peak surface area (Figure 4-1, panel D). The variation in the proportion of total cpm found in the AA peak between replicate tubes containing buffer or banamine was less than 4%.

PAGE 160

153 EPSI Activity in Endometrium from Cyclic and Pregnant Cows The inhibitory activity of non-concentrated endometrial cytosol («250 mg tissue equivalent per ml) from cyclic and pregnant cows at day 17 postestrus was determined by serial dilutions. The volumes of concentrated cytosol exerting approximately 50% inhibition were used to determine inhibitory activity for each cow. EPSI activity in concentrated endometrial cytosol from pregnant cows was greater than in non-concentrated cytosol from cyclic cows (28.7 ± 2.7 versus 17.6 ± 0.12 Ul/mg protein; 5 vs 8.1 mg tissue equivalent; P<0.01; Figure 4-4). Pooled cytosols were fractionated and concentrated using Centricon 100 and 10 ultrafiltration devices. EPSI activity was found in the C-100 retentate (« 2.5 g tissue equivalent per ml) suggesting that the active factor(s) had a molecular weight greater than 80 to 100 kDa. Activity of C-100 retentate as determined by serial dilution was 50.6 Ul/mg of protein. Figure 4-4. EPSI activity in endometrial cytosol from cyclic (n = 3) and pregnant (n = 4) cows determined by serial dilution. Cytosol from pregnant cows had greater inhibitory activity than cytosol from cyclic cows at day 17 postestrus.

PAGE 161

154 Purification of a Proteinaceous Factor Associated with EPSI Activity The active C-100 retentate was resolved by anion exchange using a Mono Q column. The chromatogram (Figure 4-5) had two distinct areas of absorbance at 280 nm. Fractions from the first area, containing basic proteins, did not exhibit EPSI activity when tested with the prostaglandin generating system. In the second area of absorption, fractions with retention times from 12.5 to 15.5 min presented the highest levels of inhibitory acivity. The fraction at 12.5 min retention time was the most active with 94.3 Ul/mg of protein as determined by serial dilution. The regions of EPSI activity were located on either side of the BSA retention time, but the most active fractions appeared to be slightly more basic (Figure 4-5). Absorption spectra (190 to 370 nm) of proteins in the most active fraction presented a second maximum at 275 nm similar to the one observed with BSA (Figure 4-6). Active fractions (retention times 12.5 to 15.0 min) were pooled, concentrated with C-100 and further resolved using a Zorbax Bioseries GF-250 gel filtration HPLC column. The chromatogram had a single region of absorbance almost centered at the retention time for BSA (Figure 4-7). Two fractions (retention times of 11.5 and 12.0 min) displayed EPSI activity with a maximum activity of 117.2 Ul/mg of protein for the fraction with a retention time of 12.0 min. i Active fractions from the gel filtration column were pooled and concentrated using a Centricon 100. The increase in specific activity of the active fractions throughout the purification scheme suggested the existence of a proteinaceous factor. However, anion exchange or gel filtration indicated that the isoelectric point and molecular weight of the active factor(s) were similar to bovine serum albumin. An additional purification step was used as an attempt to

PAGE 162

155 separate BSA from the active factor(s) based on potential differences in hydrophobic interactions with a C4 reverse phase HPLC column. The chromatogram (Figure 4-8) had three major regions of absorption suggesting that the gel filtration active fractions contained at least three components differing in hydrophobicity. Only two fractions (retention times of 3.0 and 3.5 min) inhibited prostaglandin synthesis by cotyledonary microsomes. The maximum activity of 194.2 Ul/mg of protein was observed for the fraction with a retention time of 3.5 min. Once more, the retention time of active fractions was very close to the retention time for BSA, confirming the possibility that EPSI was associated with serum albumin. 120 100 c o S 80 .9 !c: 60 in •S 40 BSA 20 1 MYO . 1 J-Ji^^ — — ^ liiii .1 -..iilil.l 1 1.4 1.2 1 0.8 0.6 0.4 0.2 0 10 15 20 retention time (min) 25 30 o 00 £i o o c (0 o M .Q (D Figure 4-5. Representative HPLC chromatogram of EPSI purification using a Mono Q anion exchange column. Endometrial cytosol from pregnant cows was fractionated and concentrated using a Centricon 100. Retentate was washed with 3 volumes of 20 mM Tns-HCI, pH 7.0, centrifuged for 3 min at 12,000 x g and loaded on the Mono Q column! Separation was accomplished by linear salt gradient (0 to 200 mM NaCI in 20 mM Tris-HCI, pH 7.0, over 25 min) at a flow rate of 1 ml/min («500 ^il fractions). The column was calibrated using horse myoglobin (MYO, pi 7.3), carbonic anhydrase (CA , pi 6.2), bovine serum albumin (BSA ,' pi 4.9) and pepsin (PEP, pi 2.9). Absorption at 280 nm was monitored (line) and fractions were assayed for protein concentration and inhibitory activity (bars).

PAGE 163

156 Figure 4-6. Absorption spectra of a BSA standard and of the Mono Q fraction with the highest EPSI activity. An aliquot of BSA solution (standard from Sigma) was eluted on a MonoQ anion exchange HPLC column for calibration. The absorption spectrum (190 to 370 nm) of a fraction containing BSA (retention time 14.0 min) is represented on panel A. Concentrated endometrial cytosol was eluted on a Mono Q column and fractions were analyzed for EPSI activity. The absorption spectrum (190 to 370 nm) of the Mono Q fraction (retention time 12.5 min) with the most EPSI activity is represented on panel B for comparison. Both spectra presented a second maximum of absorption at 274-275 nm which is characteristic of BSA.

PAGE 164

retention time (min) Figure 4-7. Representative HPLC chromatogram of EPSI purification using a Zorbax GF-250 gel filtration column. Active fractions from the anion exchange column were pooled and concentrated using a Centricon 100 ultrafiltration device. Retentate was washed with 3 volumes of 0.2 M sodium phosphate, pH 7.1, centrifuged for 3 min at 12,000 x g and loaded on a GF-250 column. Separation was accomplished with a flow rate of 0.5 ml/min of 0.2 M sodium phosphate, pH 7.1. The column was calibrated using, bovine serum albumin (BSA, 69 kDa), ovalbumin (OV, 45 kDa), carbonic anhydrase (CA. 29 kDa) and horse myoglobin (MYo! 18.5 kDa). Absorption at 280 nm was monitored (line) and fractions («250 ^il fractions) were assayed for protein concentration and inhibitory activity (bars).

PAGE 165

158 300 1 r 1.4 BSA 1.2 1 1 o 00 0.8 01 8 _ c • 0.2 0 0 5 10 15 retention time (min) Figure 4-8. Representative HPLC chromatogram of EPSI purification using a C4 reverse phase column. Active fractions from the gel filtration column were pooled and concentrated using a Centricon 100. Retentate was washed with 3 volumes of 5% glycerol solution in water, centrifuged for 3 min at 12,000 X g and loaded a C4 column. Separation was accomplished with a linear gradient (from 35% acetonitrile, 5% glycerol, 0.1% TFA, 59.9% water to 90% acetonitrile, 0.025% TFA and 1% glycerol over 30 min) at a flow rate of 1 ml/min («500 i^l fractions). Absorption was monitored at 280 nm (line). Fractions were partially evaporated with a Speed-Vac system to eliminate acetonitrile and the remaining TFA was neutralized with 10 pil of 2 M NaOH solution. Neutralized fractions were assayed for protein concentration and inhibitory activity (bars). Western Blots for Bovine Serum Albumin Immunoblotting using polyclonal anti-BSA indicated that most of the BSA found in endometrial cytosol did not transverse C-100 ultrafiltration membranes. Indeed, C-100 retentate appeared to contain higher concentrations of BSA than C-100 filtrate (Figure 4-9). This was unexpected considering that BSA molecular weight (69 kDa) is below the molecular cut-off («80 kDa; as specified by the manufacturer) of C-100 membranes. Washings (up to 7 x 2 ml) of C-100 retentate with Tris-HCI buffer did not appear to reduce the amount of BSA retained by C-100 membranes. Several possible mechanisms could explain the

PAGE 166

. 159 inability of BSA to cross C-100 ultrafiltration membranes. Two or more BSA molecules could weakly bind and form multimeric complexes to large to pass through the pores of C-100 membranes. Such complexes would be broken when eluted in the conditions used for HPLC gel filtration. Alternatively, BSA could bind the ultrafiltration membrane or other peptides or lipids that could alter the dynamic volume of albumin. Interactions between the dynamic volume of the proteins (e.g., globular, elongated), and the size and shape of the pores in ultrafiltration membranes is critical to the process of fractionation. The capacity of purified BSA (fatty acid free; Sigma) in solution to transverse C-100 membranes was verified using HPLC gel filtration. A solution of BSA (10 mg/ml in sodium phosphate buffer) was filtered through a C-100 device and centrifuged at 4,000 x g for 90 min. The C-100 retentate was washed 3 times with 2 ml of sodium phosphate buffer. The initial filtrate, the three additional filtrates from the washes and the final retentate were eluted on a GF 250 HPLC gel filtration column to compare BSA peak areas. Chromatograms indicated that BSA concentrations in C-100 filtrates were highest after the first centrifugation and decreased in subsequent filtrates. BSA was undetectable in the final retentate. This experiment was consistent with the hypothesis that lipids bound to a fraction of cytosolic BSA were responsible for its inability to transverse C-1 00 ultrafiltration membranes. Chromatographic techniques (anion exchange, gel filtration and reverse phase HPLC) suggested that BSA was associated with EPSI activity. This was confirmed by immunoblotting of the most active fraction eluted from the reverse phase (C4) column (Figure 4-10). This fraction was highly enriched in BSA as identified by molecular weight markers and immunoreactivity.

PAGE 167

1? . A5t 160 1 2 3 4 5 6 7 8 B 106 80 CO 1 50 o X 32.5 27.5 18.5 1 106 80 50 CO 1 o T— 32.5 27.5 18.5 12 3 4 5 6 7 8 Figure 4-9. Coomassie staining and immunoblotting of cytosolic proteins found in Centricon 10 (lanes 1-4; 10, 5, 2, 1 ^ig) and 100 retentates (lanes 5-8; 20, 15, 10, 5 ng). Proteins were separated by one-dimensional polyacrylamide (12.5% [w/v]) electrophoresis and either stained with Coomassie (A) or electrophoretically transferred to a nitrocellulose membrane and incubated with rabbit anti-BSA (1:416 dilution; B).

PAGE 168

161 A 1 2 3 B 12 3 Figure 4-10. Coomassie staining and immunoblotting of proteins (2 , 5, 10 ^ig; lane 1-3) found in the C4 fraction with the highest EPSI activity (retention time 2.4 min). Proteins were separated by one-dimensional polyacrylamide (12.5% [w/v]) electrophoresis and either stained with Coomassie (A) or electrophoretically transferred to a nitrocellulose membrane and incubated with rabbit anti-BSA (1:416 dilution; B) immunoprecipitation of BSA Intracellular and secreted radiolabeled proteins from endometrial explants were immunoprecipitated with anti-BSA to determine if serum albumin could be synthesized within the bovine endometrium and play the role of an endogenous endometrial prostaglandin synthesis inhibitor. Following dissociation of immunoprecipitated proteins, one-dimensional electrophoresis and development of fluorograms, no radiolabeled serum albumin could be detected after 3 or 6 weeks of exposure suggesting that serum albumin in endometrial cytosol was not synthesized by the bovine endometrium. Endometrial cells could take up BSA as a normal cellular function or, in spite of several washes before tissue homogenization, BSA could be a contaminant from the serum found in the endometrial tissue. An examination of radiolabeled proteins within the tissue of

PAGE 169

endometrial explants (Bartol et al., 1985b) revealed a major amount of unlabeled serum albumin supporting that it is a normal component of the intracellular milieu of proteins but not synthesized de novo. Titration of BSA Inhibitory Activity Inhibitory activities of C-100 concentrated cytosol, BSA and plactoglobulin were titrated in order to determine if BSA alone could account for all of the EPSI activity found in endometrial cytosol. Purified and fatty acid freeBSA appeared to inhibit prostaglandin synthesis by cotyledonary microsomes, but for equal masses of protein, BSA possessed only 50% of the inhibitory activity present In endometrial cytosol (Figure 4-11). p-lactoglobulin only marginally inhibited (< 4%) prostaglandin synthesis by cotyledonary microsomes and increasing amounts of p-lactoglobulin did not enhance inhibitory activity. Cytosol inhibition curve appeared to plateau with amounts of proteins greater than 1500 ^g. Clearly, BSA alone accounted only for a fraction of EPSI activity detected in endometrial cytosol. Interestingly, the inhibitory activity associated with BSA (fatty acid free, Sigma) was 32.6 Ul/mg protein, whereas purified BSA from endometrial cytosol (most active fraction from reverse phase column) had a specific activity of 194.2 Ul/mg protein. This line of evidence supported the hypothesis that perhaps EPSI is a lipophilic factor bound to BSA and interacted with the microsomes of the prostaglandin generator to inhibit cyclooxygenase. To verify this hypothesis, endometrial cytosol was extracted with organic solvents and activity of the total lipid extract was determined.

PAGE 170

Mg of protein ' Figure 4-1 1 . Prostaglandin synthesis inhibitory activities of pooled endometrial cytosol from pregnant cows at day 17 postestrus, bovine serum albumin (fatty acid free) and p-lactog!obulin as a function of the protein mass added to the prostaglandin-generating system. Panel A shows a semilogarithmic representation and panel B shows a linear plot. Purification of a Lipid Factor Lipid extraction . When EPS! activity was plotted against the natural log of the assayed volumes, the curves for cytosol (% inhibition = 25.56 x Ln(volume) 5.58; R2 = 0.984) and cytosol extract (% inhibition = 19.50 x Ln(volume) 5.47; R2 = 0.976) were linear (Figure 4-12). The ratio of the slopes (19.5 / 25.56) indicated that 76% of EPS! activity in endometrial cytosol was extractable using chloroform and methanol (Figure 412). This suggested that EPSI is a lipid. When cytosol was extracted with ethyl ether acidified with citric acid, commonly used to extract prostaglandins from biological samples, no EPSI activity was found in the organic phase (data not presented).

PAGE 171

164 volume (|jJ) Figure 4-12. EPSI activity (% inhibition) in increasing volumes of endometrial cytosol (5.3 mg of protein/100 from a day 17 pregnant cow compared to EPSI activity of lipid extracts (chloroform:methanol; 2:1 [v/v]) from equivalent volumes of cytosol. Increasing amounts of cytosol and cytosol lipid extract resulted in increased EPSI activity. The remaining 25% of the total activity that was not extractable with chloroform-methanol could reflect the ability of BSA and other hydrophilic factors to inhibit prostaglandin synthesis by cotyledonary microsomes. Endometrial arachidonic acid was not responsible for inhibitory activity of the lipid extract. In this experiment , microsomes were incubated with nonradiolabeled AA (32 pg and 500 ng) and PGFj^ was measured by radioimmunoassay. The 32 pg dose approximates the mass of AA associated with radiolabeled AA used in routine assays for EPSI activity. Results are presented in Table 4-3. In presence of 32 pg of AA, 300.1 ng of PGF, were

PAGE 172

165 produced during a 1 h incubation. When exogenous AA (500 ng) was added, 484.8 ng of PGFj^ were produced during a 1 h period. The addition of endometrial lipid extract to microsomes with 500 ng of AA reduced the amount of PGFja produced to a lower value than that produced in the presence of 32 pg or 500 ng of AA (183.6 vs 300.1 and 484.8 ng, respectively). Indeed the amount of PGFja detected in the supernatant following incubation of microsomes with the lipid extract of endometrium (183.6 ng) is comparable to the inherent amount of PGFja in the lipid extract of cytosol (193.0 ng). Clearly, the inhibitory activity of lipid extract was not due to endometrial AA in the extract competing with the detection assay substrate (PHJ-AA). The production of PGFj,, in presence of minimal exogenous AA (32 pg) indicates that cotyledonary microsomes metabolized endogenous AA. In presence of 500 ng of AA, PGFj^ production was stimulated. Addition of endometrial lipid extract inhibited production of PGF2„ from endogenous and exogenous AA. The lipophilic factor(s) responsible for inhibition of prostaglandin synthesis was further characterized. Table 4-3. Measurement by radioimmunoassay of PGFj^ production by cotyledonary microsomes in presence of exogenous arachidonic acid (500 ng) and endometrial lipid extract (160 [i\ cytosol equivalent) or Banamine (cyclooxygenase inhibitor). Experimental sample PGF2„ (ng) Microsomes incubated (1 h) in presence of 32 pg 300.1 ±8.0 arachidonic acid Microsomes incubated (1 h) with 500 ng 484.8 ±16.4 arachidonic acid Microsomes incubated (1 h) with 500 ng 183.6 ± 5.4 arachidonic acid and endometrial lipid extract Endometrial lipid extract without microsomes 193.0 ±4.1 Microsomes incubated (1 h) with Banamine non-detectable

PAGE 173

166 Thin layer chromatoaraphv . In order to determine the molecular nature of the lipid responsible for the inhibition of prostaglandin synthesis, the various lipid classes contained in endometrial extract were separated on silica gel thin layer chromatography plates. After extraction from the silica gel, each lipid class was tested for EPSI activity. Results are presented in Table 4-4. The free fatty acid region totally inhibited prostaglandin synthesis, whereas the "origin" (area where samples were spotted on the silica gel) presented low activity. No activity was detected in the regions corresponding to cholesterol, cholesterol esters or triglycerides. Consequently, these results indicate that, in cattle, EPSI appeared to be a free fatty acid(s). In this qualitative study, the activity of FFA region was not titrated in order to save active material for further characterization. Table 4-4. EPSI activity associated with different lipid classes separated by thin layer chromatography. Lipid classes were located on the silica gel plate according to the Rf values of standards: tripalmitin (triglyceride), cholesterol, and margaric acid (free fatty acid). For each region, the silica gel was scrapped, extracted with chloroform and lipid extracts were tested for EPSI activity in duplicate. Lipid Class % inhibition Polar lipids (origin) Cholesterol and cholesterol esters Free fatty acids Triglycerides 96.7 ±6.8 17.6 ±4.2 8.4 ±4.4 6.0 ±2.6 Blank lane 5.8 ±2.0 Identification of the activ e free fattv acid . The array of FFAs found in the TLC plate active region was separated using HPLC. The phenyl column achieved separation of FFAs based on lengths of their carbon chain and degree

PAGE 174

167 of unsaturation. Mobile phase composition and flow rate were recommended by the column manufacturer (Waters, Millipore) and achieved good separation as verified by injection of individual and mixed standard fatty acids (Figure 4-13). 2.2x10 .•| 1.7x10 H S 1.2x10 ^-| Q 7.0x10 CO B 100 n 805 is 60 20co On 3 (jd CO CO o od o I I I — r 1 I — I — I I I — 1—1 — r— I — I — (—1—1—1 — I — I2 3 4 5 retention time (min) -I— T— I— I 6 ' ' I 1 4 h > If 2 3 4 retention time (min) 1—1 — 1—1 — [— 5 ' • I 6 Figure 4-13. Representative HPLC chromatogram of a mixture of fatty acid standards using a Fatty Acid (4% phenyl load) column (panel A). Separation was accomplished isocratically with a mobile phase composed of 45% acetonitrile, 20% tetrahydrofurane and 35% water at a flow rate of 1.5 ml/min. Fatty acids were detected using an evaporative light scattering detector (ELSD). Free fatty acids from cytosolic lipid extract were eluted using the same scheme and fractions were tested for EPS! activity (panel B).

PAGE 175

r:.'^.:-168 Since the light scattering detector is destructive to the sample, only 30% of the mobile phase was directed into the evaporative chamber. Sensitivity of the light scattering detector was too low to provide a chromatogram for the separation of cytosolic free fatty acids from the active TLC region. Only two fractions (retention time 2.4 and 2.6 min) inhibited prostaglandin synthesis by cotyledonary microsomes. According to retention times of FFA standards, the most active fraction (2.4 min) was enriched highly in linoleic (16:2co6). Formal identification of FFA present in the most active fraction (2.4 min; 92 ±3% inhibition) was accomplished by gas chromatography (Figure 4-14). 16:0 retenti on ti me (mi n) 7.59 LNOLEIC ACD ( 18:2co6 ) 18:0 05 3 00 12.41 13.02 14.39 AA (20:4co-6) 27.22 Figure 4-14. Representative gas chromatogram identifying EPS!. Endometrial free fatty acids were isolated from cytosolic total lipid extract and separated with a HPLC Fatty Acid column. The FFA found in the most active fraction were methylated and injected (1:40 split ratio) into a gas chromatograph. Separation was achieved with a DB-Wax WCOT column (0.25 mm X 30 m) operated at a helium pressure of 1.0 kg/cm2. Column temperature was held at 170°C and flame ionization detector was at 300° C. 'i.

PAGE 176

169 There were a number of minor unidentified peaks on the chromatogram. These may represent oxidation products. The HPLC fraction was enriched highly in linoleic acid (47.5% of total mass detected) and contained also palmitic (23.6%), oleic (18:1(o9; 16.8%), and arachidonic (7.5%). The HPLC elution scheme did not achieve complete separation of linoleic and palmitic acids. The presence of palmitic acid in the most active fraction (2.4 min) was a consequence of the proximity of the retention times of linoleic acid (2.4 min) and palmitic acid (2.55 min). There were several indications that linoleic acid was responsible for EPSI activity. First, EPSI activity could be detected only in the linoleic acid region (2.4-2.6 min), suggesting that no other fatty acids other than LA was present in sufficient amount to inhibit prostaglandin synthesis in the conditions of the assay system. Secondly, previous studies indicated that palmitic and oleic acids do not inhibit prostaglandin synthesis by sheep vesicular microsomes (Pace-Asciak and Wolfe, 1968; Lands et al., 1972). These observations suggested that palmitic acid was not responsible for EPSI activity in the active fractions. On the other hand, linoleic acid was reported as a competitive inhibitor of cyclooxygenase (Pace-Asciak and Wolfe, 1968; Lands et al., 1972; Elattar and Lin, 1982). Consequently, the fatty acid responsible for EPSI activity appeared to be linoleic acid. Kinetics of Inhibition Linoleic acid was able to inhibit prostaglandin synthesis by cotyledonary microsomes. Inhibitory activity increased with increasing amounts of linoleic acid (0 to 1000 ng) added to the cotyledonary microsomes. The titration curve of inhibition activity in response to increasing amounts of LA was described by a

PAGE 177

170 third order polynomial regression (% inhibition = 4.04x10"® x[LA] 1.60x10"'x[LA] + 0.19x[LA] + 1.07) with an = 0.999 (Figure 4-15). At higher cx)ncentrations, the curve did not reach 100% inhibition but rather, reached a plateau at 80% inhibition. Fifty percent of the maximum inhibition of tritiated arachidonic metabolism due to linoleic acid was obtained with approximately 250 ng of linoleic acid. In this experiment, the dose of 500 ng of non radiolabeled AA was chosen for its ability to stimulate PGFj^ synthesis as measured by RIA (page 143). For equal amounts of linoleic and arachidonic acids (500 ng), prostaglandin synthesis was inhibited by 63.7% (±5.2%). This experiment suggested that linoleic acid was indeed the inhibitory factor detected in TLC and HPLC active fractions. 1000 1500 linoleic acid (ng) 200 Figure 4-15. Inhibition of prostaglandin synthesis by cotyledonary microsomes in presence of increasing amounts of linoleic acid. Microsomes were incubated with (3H)-arachidonic acid (80,000 dpm with 500 ng non radiolabeled arachidonic acid in 50 ^il), and various amounts of linoleic acid (0, 125, 250, 500, 1000, 2000 ng in 160 ^1) for 1 h at 38°C. The experiment to evaluate the nature of enzymatic inhibition by linoleic acid is depicted in Figure 4-16. The double reciprocal plot indicated that addition of

PAGE 178

171 linoleic acid to cotyledonary microsomes inhibited AA metabolism in a competitive manner. Indeed, the regression line obtained with various amounts of AA in absence of linoleic acid (1A/ = 701096 (± 26590) * [1/AA] + 22171 (± 2507); R2 = 0.98) intersected on the Y axis (1 / V) with the regression line 1/V(nM/min) Km(+linol&c) = 57.2/uM Km (linoleic) = 32.9 ^ Figure 4-16. Double reciprocal plot of linoleic acid inhibition of arachidonic acid metabolism by cotyledonary microsomes. Various amounts of AA (500, 1000, 1500, 2000 ng; combined with (3H)-AA (80,000 dpm in 50 ^1 of resuspension buffer) were incubated with cotyledonary microsomes and linoleic acid (0 and 500 ng). Incubation at 38°C was terminated after 10, 20, 30, 40, 50 or 60 min by addition of acetonitrile (125 |al). Rates of metabolism (iiM/min) were determined for each AA concentration in absence or presence of linoleic acid (4x2 slopes) and plotted on a double reciprocal graph. The presence of linoleic acid did not appear to alter the enzyme maximum velocity (V^) but increased the apparent K^. This type of plot is characteristic of a competitive inhibition.

PAGE 179

172 obtained in presence of linoleic acid (1A/ = 11 98609 (± 10690) x [1/AA] + 21254 (± 1008); R2 = 0.99). Consequently, the presence of linoleic acid did not affect the maximum velocity (V^^ = 0.46 nM/min) of the microsomal cyclooxygenase. However, the (intercept with the X axis) increased from 32.9 \xM in absence of linoleic acid to 57.2 nM in presence of linoleic acid. This suggests that linoleic acid can compete for AA binding site on cyclooxygenase and therefore inhibit prostaglandin synthesis. The and V^ax described for microsomal cyclooxygenase in this assay system are not absolute values, and they cannot be used to characterize purified cyclooxygenase for reasons that will be discussed later. EPSI Activity and Quantification of Free Linoleic and Arachidonic Acids in Endometrium from Pregnant and Cvclic Cows Linoleic acid was identified as an active factor in endometrial cytosol that inhibits cyclooxygenase of cotyledonary microsomes. The objective of this experiment was to examine if reproductive status (cyclic versus pregnant) altered concentrations of free LA and AA in endometrial cytosol and microsomes from pregnant and versus cyclic cows at day 17 postestrus. EPSI activity. Tissues from pregnant cows contained more EPSI activity than tissues from cyclic cows (42.5 ±3.5 vs 23.8 ±4.1 Ul/mg protein; P<0.02; Figure 4-17). This confirmed the higher EPSI activity detected in the first set of pregnant cows utilized for EPSI characterization. The capability to assay EPSI activity in endometrial microsomes was potentially compromised by the presence of a variety of enzymes such as cycio-

PAGE 180

Cytosol Microsomes Figure 4-17. EPSI activity of boiled microsomes and cytosol from endometrial tissues of cyclic (n = 3) and pregnant cows (n = 4) at day 17 postestrus. oxygenases, lipoxygenases, acyl-CoA synthetase, dehydrogenases and elongases that could interfere with our assay system (cotyledonary microsomes). In order to eliminate these enzymatic activities, microsomes were boiled for 10 min and tested for EPSI activity. Such treatment of microsomes was unlikely to destroy EPSI activity since, in the process of identification of EPSI as linoleic acid, cytosolic lipid extracts remained active after being heated to 70°C for over 30 min in a Kuderna-Danish concentrator. Such treatment did not destroy EPSI activity in extracted cytosol and heat-treated microsomes. The increase in EPSI activity detected in cytosol (P<0.05) at day 17 of pregnancy was even more evident in microsomes (P<0.01). LA and AA Quantification . This experiment was not conducted to provide absolute quantities of unesterified LA and AA in endometrial tissues but rather to estimate relative concentrations of free LA and AA in cytosolic and microsomal fractions of endometrial tissue. No direct comparison between microsomes and cytosol were made because, after endometrial tissue homogenization and

PAGE 181

174 ultracentrifugation, larger amounts of proteins were recovered in the cytosol compared to microsomes. Gas chromatograms indicated that endometrial cytosol and microsomes contained a variety of free fatty acids. A number of unidentified peaks were present which may be oxidation products. Appreciable differences in free LA and AA concentrations were observed among cows. Compared to cyclic cows, endometrial microsomes from pregnant cows were characterized by higher LA concentrations (736.9 ±100.1 vs 356.7 ±115.6 ng/mg protein; P<0.01) and lower AA concentrations (258.7 ±49.9 vs 570.3 ±57.7; P<0.05). This effect of pregnancy on microsomal fatty acids was not detected in cytosol (status x location x fatty acid interaction; P<0.03; status x location interaction; P<0.1 for LA and P<0.01 for AA). An inverse relationship was found in the microsomes from endometrial tissue of the cycle. Indeed, cyclic tissues were characterized by higher concentrations of AA than LA. Consequently, the ratio (LA:AA) was higher in pregnant cows than in cyclic cows for both cellular fractions (P<0.01; Figure 4-18, panel C). It appears that increased EPSI activity detected in endometrial microsomes from pregnant cows is associated with high concentrations of free LA. In addition, the amounts of LA measured in endometrial microsomes appear to be sufficient to account for all of the EPSI activity. Indeed, microsomes from pregnant cows contained approximately 700 ng of LA per mg of protein (Figure 4-18, panel A) with a corresponding microsomal EPSI activity of 50 Ul/mg protein (Figure 4-17). The titration curve of LA inhibitory activity (using purified LA; Figure 4-15) indicates that 400 ng of LA are necessary to generate 50 units of inhibition. Thus the amount of free linoleic acid extracted from endometrial microsomes exceeds the amount necessary to achieve an EPSI activity of 50

PAGE 182

A 800 q 700600^ c 500400I O) E 300^ 2001000-^ Free Lindeic add T C 2.5n 21.5 H 0.5 Cytosol Microsomes B 800 q 700^ 600 T 500 H 400^ 300 2001000 Free Arachidonic add Cytosol Microsomes Cyclic Pregnant Ratio Linoleic to Aradiidonic add Cytosol Microsomes Figure 4-18. Concentrations of free linoleic (panel A) and arachidonic (panel B) acids in endometrial cytosol and microsomes of cyclic (n = 3) and pregnant (n = 4) cows at day 17 postestrus. Concentrations of free fatty acids expressed as ng per mg of cytosolic and microsomal protein extracted for quantification.

PAGE 183

176 units. Perhaps not all of the extracted linoleic acid is available for regulation of cyclooxygenase. It is possible that the some of the linoleic is associated with microsomal proteins (e.g., binding proteins) and less available to inhibit prostaglandin synthesis. In pregnant cows, the total amount of microsomal proteins recovered after homogenization and ultracentrifugation of endometrial tissue was approximately 37 ±6.6 mg that would contain 25.2 ^g of LA. Interestingly, it took only 1% (« 250 ng) of the total LA recovered in microsomes prepared from the endometrium ipsilateral to the CL and/or embryo to induce a 40% inhibition of prostaglandin synthesis by cotyledonary microsomes. The effect of reproductive status in endometrial tissues was particularly obvious with the ratio of free LA to AA concentrations. This experiment supports the role of linoleic acid as a prostaglandin synthesis inhibitor in bovine endometrium. These observations support the role of unesterified linoleic acid as a prostaglandin synthesis inhibitor in the bovine endometrium and that availability of AA may be less at the time of recognition of pregnancy. These effects appeared to be specific to endometrial tissues since, at day 17 postestrus, reproductive status had no effect on plasma concentrations of free LA, free AA or ratio LA:AA (Figure 4-19). Consequently, EPSI activities in endometrial microsomes and cytosol likely were not due to a "lipid contamination" from the plasma. EPSI Activitv and Quantification of Free Linoleic and Arachidonic Acids in Endometrium from Recombinant Bovine IFNx-Treated and Control Cyclic Cows EPSI activitv . In vivo treatment with rblFNx had no effect on EPSI activity in endometrial cytosol or microsomes (Figure 4-20; P<0.5).

PAGE 184

177 80 70^ c 60 « 5040^ 30^ 20100D) E Cyclic ^ Pregnant linoleic arachidonic Figure 4-19. Plasma concentrations of free linoleic and arachidonic acids in cyclic (n = 3) and pregnant (n = 4) cows at day 17 postestrus. 60-1 |50 i40 I 30 la ^20 CO •110 Z3 Control M rblFN X Cytosol Microsomes Figure 4-20. EPS! activity in boiled microsomes and cytosol from endometrial tissues of cows treated in vivo with rblFNx (n = 4; two daily intrauterine infusions containing 0.2 mg IFNx in 1.3 mg BSA) or BSA (n = 4; two daily intrauterine infusions containing 1.5 mg BSA) from day 14 to 17 postestrus. LA and AA quantification . Recombinant bovine IFNx had no significant effect on free AA content in either cellular fractions (Figure 4-21, panel B). There was no significant difference in the ratio (LA:AA) between the two treatments (P<0.38; Figure 4-21, panel C).

PAGE 185

178 1 450400350300250200? 150 100 50 0FreeUnoleic add B 450 q 400 350^ www 300 i 250 -i 200 i 150H 100H 50 i 0^ Free Arachidonic add Cytosol Microsomes Cytosol Microsomes c 3n 2.5-; 2 1.5 0.5 0 Control M rbFNx Ratio Linoleic to Aradiidonic add Cytosol Microsomes Figure 4-21. Concentrations of free linoleic (panel A) and arachidonic (panel B) acids in endometrial cytosol and microsomes of cows treated with two daily intrauterine infusions of BSA (n = 5) or recombinant bovine IFNx (n = 5) from day 14 to 17 postestrus. Concentrations (ng) of free fatty acids expressed per mg of cytosolic and microsomal protein equivalent that was extracted for quantification. There was no significant difference between reproductive statuses.

PAGE 186

179 In both experimental groups, cytosolic concentrations of LA and AA were comparable to the ones observed in cyclic and pregnant cows. Control cows had microsomal concentrations of LA similar to that of cyclic cows. However, in rblFNx-treated cows, there was no increase in LA concentrations as observed in pregnant cows (Figure 4-21 panel A). In addition, microsomal concentrations of AA in control cows were not as high as in cyclic cows (332.1 ±39.2 vs 570 ±57.7 ng/mg protein; Figures 4-18 and 4-21 , panels B). In vivo treatment with rblFNx failed to increase EPSI activity and concentrations of free LA in endometrial tissue. In this experiment, low microsomal concentrations of free LA in endometrium from IFNx-treated cows were associated with low EPSI activity. In pregnant cows, high LA concentrations were associated with high EPSI activity. Although IFNx was unable to replicate the effects of pregnancy on endometrial FFAs concentrations, these results confirmed the direct relationship between concentrations of free LA and EPSI activity observed in cyclic and pregnant cows. The reasons for IFNx inability to induce higher EPSI activity in this experimental model will be discussed later. Discussion Present results indicate that prostaglandin synthesis is reduced, at least in part, by the presence of unesterified linoleic acid in endometrial tissue. Free linoleic acid is found primarily in the microsomal fraction of endometrial cells and present in higher amounts in endometrium from pregnant cows than from cyclic cows at day 1 7 postestrus.

PAGE 187

180 Several studies (Basu and Kindahl, 1987b; Gross et al., 1988a) have reported higher EPSI activity in endometrium from pregnant cows compared to cyclic cows at day 17 postestrus. Data presented in this chapter are in agreement with these studies. So far, only partial biochemical characteristics of EPSI have been reported. Basu and Kindahl (1987b) assayed inhibitory activity by conversion of {'^^C)-AA to (i'*C)-prostaglandins and separation of AA metabolites by thin layer chromatography. Using this assay, they observed that boiling lowered, but did not eradicate the inhibitory capacity of endometrial tissue. Shemesh et al. (1984) reported a proteinaceous (a60 kDa) inhibitor of prostaglandin synthesis from bovine placental tissue during mid-to late pregnancy which was heat labile and ammonium sulfate precipitable. Ishihari et al. (1990) partially purified an inhibitor of prostaglandin synthesis from human decidua. The active factor was found in the cytosolic fraction and was thought to be a heat-sensitive acidic protein with a molecular weight of about 55 to 60 kDa. These protein characteristics could represent serum albumin. A study by Gross et al. (1988a) described EPSI as a proteinaceous factor, sensitive to heat and proteases, and acting as a non-competitive inhibitor of cyclooxygenase. A critical difference between the present study and the work reported by Gross et al. (1988a) resides in the detection assay for EPSI activity. The assay used by Gross et al. (1988a) was based on conversion of non-radiolabelled AA (100 ng) by cotyledonary microsomes (500 nl; 500 mg tissue equivalent) and measurement of PGRj^ by radioimmunoassay. The detection assay developed for the present study monitored conversion of (3H)-AA (32 pg) by cotyledonary microsomes (lOjxl; 40 mg tissue equivalent) and separated (3H)-AA metabolites from (3H)-AA by HPLC. This assay system is significantly more sensitive since only 2.5 mg tissue equivalent were necessary to inhibit prostaglandin synthesis

PAGE 188

181 by 50%, compared to 270 mg tissue equivalent when monitored by RIA (Gross et al., 1988a). Cotyledonary microsomes from parturient cows were chosen for their capacity to synthesize prostaglandins and to avoid testing bovine EPS! activity with a prostaglandin generator from a different species. The difference in EPS! assay systems and nature of the inhibition tested (cytosol versus linoleic acid) between Gross et al. (1988a) and the present study may explain the apparent differences in kinetics of inhibition (non-competitive versus competitive). For example, trapping of linoleic and arachidonic acids by BSA in cytosol could be responsible for the apparent non-competitive inhibition reported by Gross etal. (1988a). The attempt to characterize a proteinaceous factor was based on the preliminary data obtained in our laboratory by Gross et al. (1988a), who described the presence of inhibitory activity in the 25-30 and 70-75 kDa regions when endometrial cytosol was separated by HPLC gel filtration. In spite of a more sensitive assay, no proteinaceous inhibitory factors were detected in the 25-30 kDa region in the present study. However, the second region of inhibition (70-75 kDa) could correspond to the active BSA region identified in this study. The possibility that EPSI is a lipophilic factor was explored initially by cytosol extraction with acidified ethyl ether (using citric acid). No activity was found in the organic phase using this extraction system. Retrospectively, this result was not surprising considering that EPSI activity is associated with a free fatty acid (linoleic acid). Indeed, a free fatty acid needs to bear a neutral carboxyl group (R-COOH) in order to be soluble in a non polar solvent such as ethyl ether. Citric acid was probably to weak of an acid to induce protonation of the carboxyl groups.

PAGE 189

Fractionation of endometrial cytosol using ultrafiltration membranes (C100) suggested that a high molecular weight (> 80 kDa; as specified by the manufacturer) proteinaceous factor was associated with EPSI activity. However, separation of the C-100 retentate by HPLC anion exchange and gel filtration indicated that EPSI activity was found in the BSA region. The apparent contradiction between the molecular weight estimate given by liquid chromatography and ultrafiltration membranes could be for several reasons. One of them could be the presence of at least two populations of serum albumin with different physicochemical properties in relation to ultrafiltration membranes. The capacity of fatty acid free BSA (Sigma) in solution to transverse C-100 membranes was verified using HPLC gel filtration to monitor BSA concentrations in C-100 filtrates and retentates. It is possible that when a certain amount of fatty acids and/or other lipids bind to BSA, the conformation of the albumin-lipid complex becomes to voluminous to transverse freely the pores of C-100 membranes. Alternatively, under the experimental conditions used in this study, BSA could form multimeric aggregates with a molecular weight greater than 100 kDa. The poor capacity of serum albumin from endometrial tissue to cross C100 membranes was also confirmed by Western blot. Increase in specific activity throughout the protein purification scheme (from 28.7 to 194.2 Ul/mg protein) was indicative of EPSI association with a proteinaceous factor. Each chromatographic step indicated that EPSI activity was in the BSA region but never at the exact retention time of the BSA standard (fatty acid free) used to calibrate the HPLC columns. However, absorption spectra (190 to 370 nm) of the active material presented a second maximum at 275 nm characteristic of BSA. Fatty acids and other lipids complexed with BSA may be responsible for the small shift in retention times observed with anion

PAGE 190

183 exchange and reverse phase HPLC. After gel electrophoresis and immunoblotting of the most active reverse phase fraction, it was clear that the protein associated with EPSI activity was BSA. Extraction with chloroform and methanol indicated that 75% of the activity was soluble in organic solvents. The remaining 25% was associated probably with the ability of BSA itself to inhibit prostaglandin synthesis by cotyledonary microsomes. When fatty acid free BSA (Sigma) and cytosolic activities were compared on an equal mass basis, albumin had approximately 50% activity cytosol. The mechanism by which BSA inhibits prostaglandin synthesis in this assay system is unknown, but BSA is likely to bind radiolabelled AA, making it unavailable for metabolism by cotyledonary microsomes. It has long been known that serum albumin binds long-chain fatty acids. Goodman (1958) described three classes of FA binding sites: the first class consists of two high affinity sites, the second of five intermediate affinity sites, and the third of a larger number of low affinity sites. The effect of BSA on the synthesis of prostaglandins and the trapping of arachidonic acid released from cellular lipid stores are not well understood. BSA can inhibit bradykinin and ionophore A23187-stimulated synthesis of prostaglandins in human embryo lung fibroblasts (Heinsohn et al., 1987). In isolated perfused rat heart, Paul and Kinsella (1983) observed that bradykinin stimulated AA and prostaglandin release, whereas BSA stimulated AA release but had no effect on prostaglandin secretion. It was suggested that bradykinin acted on a specific pool of AA available to cyclooxygenase, whereas BSA stimulated a separate pool of AA which apparently was not available to cyclooxygenase.

PAGE 191

' 184 The hypothesis that albumin could be synthesized in bovine endometrium to regulate free arachidonic acid availability and prostaglandin synthesis was tested. The major site of serum albumin synthesis is the liver (Rothschild et al., 1980). There was no evidence of endogenous synthesis of albumin within the endometrium. However, the possibility that BSA is involved in regulation of prostaglandin synthesis cannot be ruled out. Specter (1968) described albumin as the major carrier of FFA in adult plasma. FFA are transferred rapidly and reversibly between albumin in the extracellular space and cells via an unbound intermediate. In addition to a role as a regulator of exogenous FFA availability, BSA could modulate also FFA availability within endometrial cell compartments. Endocytosis of serum albumin has been demonstrated in various tissues. Schwegler et al. (1991) demonstrated that albumin undergoes receptor-mediated endocytosis in cultured opossum kidney cells using fluorescein and gold-labeled albumin. Ultrastructural studies have shown binding and endocytosis of serum albumin in bovine cardiac endothelial cells (Manduteanu et al., 1988) and aortic smooth muscle cells (Sprague et al., 1985). Gel electrophoresis of proteins represented in uterine flushing or in endometrial tissue homogenates during early pregnancy indicated the presence of significant amounts of serum albumin (Bartol et al., 1985b). The bovine embryo may promote endometrial uptake of luminal proteins by regulation of the net charge of microdomains within the glycocalyx of endometrial epithelial cells. Increase in the number of cationic microdomains in the glycocalyx of muscle cells were shown to promote endocytosis of anionic plasma proteins and in particular albumin (Sprague et al., 1985). ^ Thin layer chromatography indicated clearly that EPSI activity was found in the FFA region. It was likely that a portion of the FFA extracted from

PAGE 192

185 endometrial cytosol was associated with serum proteins infiltrated in the extracellular space. This raised the possibility that the activity found in cytosolic extracts was not specific to endometrial tissue but rather due to lipids associated with serum proteins. Quantification of free LA and AA in plasma, endometrial cytosol and microsomes provided two lines of evidence indicating that this was not the case. First, for each cow, concentrations of unesterified LA and AA were lower in plasma than in endometrial cytosol and microsomes. Secondly, reproductive status had no effect on plasma concentrations of free LA and AA. Requirement for polyunsaturated fatty acids cannot be met solely by de novo synthesis within animal tissues. Animals are absolutely dependent on plants for the major precursors of the 006 fatty acids: e.g., linoleic acid (Cook, 1991). Fatty acids are transported between organs either as unesterified fatty acids complexed to serum albumin or in the from of triacylglycerols associated with lipoproteins. Triacylglycerol-rich lipoprotein particles such as chylomicron and very-low-density lipoproteins (VLDL) represent the major source of fatty acids for tissues (Fielding and Fielding, 1 991 ). Triacylglycerols are hydrolyzed at the luminal surface of endothelial cells of capillaries by lipoprotein lipase (LPL) to yield free fatty acids. LPL is synthesized by mesenchymal cells and then transferred to the luminal surface of endothelial cells (Saxena et al., 1991) where it can hydrolyze triacylglycerol-rich lipoproteins. The mechanism by which FFA enter cells remain poorly understood despite a number of studies performed with isolated cells from heart, liver, and adipose tissue (Schultz, 1991a). Evidence has been obtained for both a saturable and non-saturable uptake of FFA. The saturable uptake which predominates at low concentrations of FFA is thought to be carrier-mediated and possibly involves a 40 kDa plasma membrane protein (Schultz, 1991b). The

PAGE 193

186 nonsaturable uptake, which is significant only at higher concentrations of FFA, has been attributed to nonspecific diffusion across the membrane (Schultz, 1991b). Once long-chain fatty acids have crossed the plasma membrane, they either diffuse or are transported to endoplasmic reticulum or mitochondria where they are activated by conversion to acylCoA. The identification of low molecular weight (14-15 kDa) fatty acid binding proteins (FABPs) in the cytosol of various animal tissues has led to the suggestion that these proteins may be carriers of FFA in the cytosolic compartment. FABPs may also be involved in the cellular uptake of fatty acids, their intracellular storage and/or the delivery of fatty acids to their sites of utilization (Veerkamp et al., 1991). In endometrial cells, uptake, storage and delivery of fatty acids could be influenced by the presence of conceptus secretory products. Separation of cytosolic FFA by HPLC indicated that EPSI activity was associated with linoleic acid. The assay system failed to detected EPSI activity in other lipid classes e.g., phospholipids, triglycerides or cholesterol esters. However, in vivo, additional FFA such as linolenic acid (18:3(o6) or ©3 unsaturated FA (18:3(o3, 20:5co3, 22:6(d3) may contribute to cyclooxygenase inhibition. Epoxy derivatives of AA (c/s-epoxy-eicosatrienoic acids; EpEtrEs) and in particular 14,15-c/s-EpETrE are potent natural inhibitors of prostaglandin synthesis but their isolation is difficult because of enzymatic and spontaneous hydration at low pH (Smith et al., 1991). The protocol for endometrial tissue preparation did not include inhibitors of enzymatic hydration. Epoxy derivatives could be responsible for the low EPSI activity detected in the TLC region corresponding to polar lipids. EpEtrEs have been described in liver, kidney, lung and in urine (Catella et al., 1990) but not in endometrial tissues.

PAGE 194

The EPSI activity assay used bovine cotyledonary microsomes as a source of cyclooxygenase enzyme. Tlie apparent values for and V^gx obtained using this assay cannot be considered absolute values characteristic of a pure enzyme. Only purified bovine PGH synthase from endometrial tissue could provide specific estimates of cyclooxygenase and Using purified PGH synthase from ovine and bovine seminal vesicles, cyclooxygenase activity can be monitored continuously by measuring uptake with an oxygen electrode (Smith and Lands, 1972). The values for AA and Oj are both about 5 i^M (Lands et al., 1978; Van der Ouderaa et al., 1979; Kulmacz and Lands, 1983). The apparent K„ for AA in the conditions of EPSI assay was 32.9 jiM. Numerous factors could explain this difference in estimates. Presence of hydroperoxides is required for the oxygenation reaction (Hemler et al., 1979). When AA, Oj and PGH synthase are mixed, the enzyme catalyses a slow formation of PGGj, which is permitted by the presence of trace amounts of hydroperoxide. Hydroperoxides act as an activating cofactor and not as a cosubstrate (Hemler and Lands, 1980). Hydroperoxide concentrations can be regulated by peroxidases and in particular the presence of glutathione (GSH), and GSH peroxidase can reduce cyclooxygenase activity (Smith and Lands, 1972; Hemler and Lands, 1980). Concentrations of hydroperoxides, GSH and GSH peroxidase in cotyledonary microsomes are unknown. Although PGFj^ appeared to be the major AA metabolite, cotyledonary microsomes are likely to contain enzymes using AA as a substrate for lipoxygenases, P-450 epoxygenase, acylCoA transferases, elongases and desaturases. Selfcatalyzed destruction of cyclooxgenase has been observed with both microsomal and purified PGH synthase (Hemler and Lands, 1980). The rates of self-inactivation vary with different fatty acid substrates (Smith and Lands, 1972).

PAGE 195

188 Not only could high endometrial concentrations of linoleic acid compete with AA for cyclooxygenase but also could accelerate self-inactivation of the enzyme. This could represent an additional level of regulation of endometrial cyclooxygenase activity by free LA. The ability of linoleic to competitively inhibit prostaglandin synthesis was reflected in the increase of the apparent value from 32.9 to 57.2 ^M. Linoleic acid has been described to cause both an instantaneous competitive inhibition and a progressive and non-reversible loss of cyclooxygenase activity (Lands et al., 1972). The products formed by oxygenation of linoleic acid by sheep PGH synthase are 9-hydroxy-10,12-octadecadienoic acid (9-HODE; 82%) and 13hydroxy-9,11-octadecadienoic acid (13-HODE; 18%) (Hamberg and Samuelson, 1967a). Several investigations have demonstrated that 18:2co6 (linoleic acid) is metabolized by vascular cells or tissue to monohydroxylated derivatives 13and 9-HODE (Funck and Powell, 1983; 1985; Kaduce et al., 1989). Since cyclooxygenase undergoes a self-inactivation after a certain number of oxygenation cycles (Marshall et al., 1987), formation of HODE (from LA) instead of PGHj (from AA) may contribute also to a deaease in cyclooxygenase activity. Intracellular levels of 13-HODE influence interactions of several circulating cells and tumor cells with the vascular endothelium (Hass et al., 1988). The 9and 13-HODE and their hydroperoxy precursors are weak vasoconstrictors of isolated rabbit vessels and inhibitors of AA-induced platelet aggregation (Bult et al., 1987). To date, the presence of monohydroxy derivatives of linoleic acid and their effects on bovine endometrial prostaglandin synthesis have not been reported. Several studies indicate that lipid composition of endometrial cell membranes differ in pregnant and cyclic cows at day 17 postestrus. Thatcher et

PAGE 196

189 al. (1984a) reported a decreased efficiency of POFj^ transfer from the uterine lumen to the ovarian vein during early pregnancy. Uptake of RGFj^ by the endometrium was not saturable over the range of doses used in this experiment, suggesting a nonspecific diffusion process. Diffusion of a lipophilic substance across cell membranes is dependent upon many factors including concentration gradient and lipid composition of the membrane (Guyton, 1971). Curl (1988) measured relatively low amounts of esterified arachidonic acid in endometrium of day 17 pregnant cows, possibly due to AA utilization by the conceptus to synthesize prostaglandins and/or its own pool of phospholipids. Lukaszewska and Hansel (1980) found no difference in the amount of arachidonic acid esterified to phospholipids in endometrium from pregnant versus cyclic cows at day 18 postestrus. A recent study suggest that in pregnant sheep, fatty acid distribution in uterine arteries is specifically modified compared to nonpregnant animals (Hoffman et al., 1994). Arachidonic acid concentrations were significantly lower in uterine versus systemic arteries, whereas linoleic acid concentrations were elevated. This change in ratio of linoleic to arachidonic acid supports the pregnancy-induced alteration in microsomal fatty acids composition observed in the present study. Determination of free arachidonic and linoleic acid concentrations (as opposed to total AA or LA) may provide a better picture of the actual free fatty acid dynamics in endometrial tissues. These concentrations could represent the end result of the numerous mechanisms involved in regulating AA availability for prostaglandin synthase during early pregnancy. The pregnancy-induced increase in endometrial free linoleic acid concentrations supports the role of this fatty acid as a prostaglandin synthesis inhibitor. The regulation of unesterified linoleic acid concentrations involves a number of intracellular enzymatic

PAGE 197

190 pathways. Release of arachidonic acid from phospholipid storage pools by PLAj appears to be the rate limiting factor in prostaglandin synthesis (Marshall et al., 1987). Characteristically, the second position on phospholipid molecules is occupied by polyunsaturated fatty acids such as 18:2co6, 20:4a)6 (arachidonic acid) and 22:6co6 (Vance, 1991). After a lysophospholipid is formed by release of AA under the action of PLAj, lysophospholipid can either be degraded or reacylated to reform a phospholipid. If the latter reaction occurs and a new fatty acid is introduced, the fatty acid composition of phospholipids is said to have been remodeled. It is possible that during early pregnancy, the population of phospholipids accessible to cyclooxygenase is remodeled in favor of linoleic acid in the sn-2 position instead of AA. In chapter 2, exogenous PLAj stimulated PGFj,, secretion of endometrial explants from pregnant cows but not from cyclic cows, suggesting that AA availability and perhaps PLAj activity is limiting prostaglandin synthesis in pregnancy. Various forms of PLAj could be differentially regulated during early pregnancy. It is conceivable that its 85 kDa PLAj isoform, which exhibits a strong preference for s/7-2-arachidonoyl phospholipids (Diez et al., 1992), could be down regulated which, consequently, would decrease AA availability for prostaglandin synthesis. On the other hand, other Isoforms of PLAj with no preference for the type of unsaturated FA in the sn-2 position (14 kDa PLAj; Kramer et al., 1989; Hara et al., 1989) would be unaffected or even up-regulated during early pregnancy. The 14 kDa isoform of PLAj has been shown to hydrolyze preferentially certain classes of phospholipids (PE > PS » PC). Curl (1988) estimated that 65% of total endometrial arachidonic acid was bound to PE. A preferential remodeling of PE in which linoleic acid would replace AA in the sn-2 position in addition to a differential regulation of PLAj isoforms are two

PAGE 198

191 possible mechanisms contributing to the increase in endometrial free LA and decrease in free AA concentrations observed during early pregnancy. Another mechanism that may contribute to the pregnancy-induced increase in free linoleic acid (and decrease in free AA) involves elongase and desaturase enzymes. Arachidonic acid (20:4©6) can be formed from linoleic acid (18:2q6) by alternating sequences of A6 desaturation, chain elongation of the 18:3co6 intermediate, and A5 desaturation of 20:3co3 (Cook, 1978). In the liver and most tissues, the only members of the ©6 family to accumulate are arachidonic and linoleic acids; much lower levels of the intermediates are detected (Cook, 1978; Sprecher, 1981). Such observations support the role of A 6 desaturation as the rate-limiting step. Downregulation of endometrial A6 desaturase by the conceptus would decrease arachidonic acid synthesis and increase concentrations of linoleic acid. The pattern of high LA and low AA (ratio LA:AA « 2) has been associated with impaired A6 and A5 desaturase activities in rats, human and primates (Brenner, 1990). The activity of A6 desaturase is under endocrine regulation. Glucocorticoids and 1 7p-estradiol have been shown to decrease A6 desaturase activity (Brenner, 1990). The bovine conceptus actively synthesizes steroids by day 16 (Eley et al., 1983), it is conceivable that endometrial A6 desaturase activity could be downregulated by conceptus-secreted steroids. In vivo treatment with recombinant IFNt was an attempt to replicate the pregnancy-induced increase in endometrial free linoleic acid and associated EPSI activity. Two daily intrauterine infusions of rblFNx failed to increase EPSI activity and concentrations of free LA in the endometrium. Interestingly, AA concentrations in BSA-treated cows were lower than in cyclic cows. Daily infusions of 3 mg of bovine serum albumin in the uterus could be partially

PAGE 199

responsible for this decrease In endometrial FFA concentrations. The additional BSA introduced In the uterine lumen might alter the dynamics of endometrial fatty acid exchange In the endometrium. It may be more appropriate to use a protein that does not bind FFA as a control. The presence of 1.3 mg of BSA in each infusion of 0.2 mg of rblFNx might be responsible for the ineffectiveness of rblFNx to Increase EPSI activity and free LA concentrations. Alternatively, IFNt may not be the embryonic factor responsible for the regulation of endometrial free LA and AA concentrations during early pregnancy. However, In chapter 3, natural and recombinant IFNx decreased prostaglandin secretion from endometrial epithelial cells. It would be interesting to examine if, in this primary cell culture system, lower prostaglandin secretion is accompanied with an increase In free LA to AA ratio. Helmer et al. (1989b) reported that natural blFNx induced EPSI activity and decreased PGFj^ secretion when administered to endometrial explants from cyclic cows at day 17 postestrus. In conclusion, nonesterified linoleic acid appears to be responsible for the prostaglandin synthesis Inhibitory activity detected in endometrium from cyclic and pregnant cows at day 17 postestrus. Linoleic acid is a competitive Inhibitor of the cyclooxygenase activity of PGH synthase. The increase in EPSI activity observed In pregnant cows Is concomitant with an increase in free LA and a decrease In free AA. This pregnancy-induced regulation of nonesterified fatty acids Is particularly obvious in the microsomal fraction of the endometrium. The mechanisms by which lipid metabolism, and in particular free LA and AA concentrations, are regulated are very complex. The nature and function of lipid mediators such as fatty acids and their derivatives as a component the Interferons a signal transduction system is still largely unknown, but a number of

PAGE 200

193 studies indicate that lipid mediators may play a significant role in mediating some of the effect induced by interferons a (see general discussion). 1

PAGE 201

CHAPTER 5 GENERAL DISCUSSION Prostaglandin Fj^ produced by the uterine endometrium in response to ovarian secretion of estradiol and oxytocin appears to be the luteolytic agent in the bovine (Lafrance and Goff, 1985; Thatcher et al., 1986a). To maintain pregnancy, the bovine conceptus must signal its presence to the dam and prevent the occurrence of luteolysis by day 15 to 17 (Northey and French, 1980). During early pregnancy, pulsatile secretion of PGFj^ is suppressed in sheep (Thornburn et al., 1973; McCracken et al., 1984; Zarco et al., 1988) and cattle (Kindahl et al., 1976; Thatcher et al., 1984a; Basu and Kindahl, 1987a), as is the ability of oxytocin to stimulate uterine secretion of PGF2„ in sheep (Fairclough et al., 1984; Silvia et al., 1992) and cattle (Lafrance et Goff, 1985; Wolfenson et al., 1993). The conceptus appears to delay the normal development of uterine responsiveness to oxytocin that occurs during the estrous cycle. Associated v^ith this pregnancy effect are a multiplicity of conceptus-secreted proteins (Bartol et al., 1985a), steroids (Eley et al., 1983) and prostaglandins (Lewis et al. 1982). The factor blocking luteolysis by the bovine conceptus was identified as a complex of proteins designated as bovine trophoblast protein-1 (bTP-1; Helmer et al., 1987;1989a; 1989b). Molecular cloning of cDNA (Imakawa et al., 1987; 1989) and protein sequencing (Stewart et al., 1987; Charpigny et al., 1988) identified bTP-1 as type I interferon and later designated as trophoblast interferon x (Roberts et al., 1992). As for other type I interferons, IFNx possesses antiviral, antiproliferative and immunomodulatory properties (see 194

PAGE 202

195 Roberts et al., 1992). The most remarkable biological effect of IFNx is to prevent luteolysis and thereby prolong luteal lifespan (Godkin et al., 1984a; Knickerbocker et al., 1986b; Vallet et al., 1988; Helmer et al., 1989a; Ott et al., 1991). In vitro studies of prostaglandin secretion from perifused endometrium and endometrial explants confirmed that PGFj^ secretion was lower in pregnant than in cyclic cows, whereas secretion of PGE2 was either unaltered (Thatcher et al., 1984b; Gross et a!., 1988b) or increased (Gross et a!., 1988c). Incubation of endometrial explants of day 17 cyclic cows with IFNx or bovine IFNa inhibited PGFja secretion (Thatcher et al., 1989; Barros et al., 1991). Studies with these tissue models evaluated effects of pregnancy or IFNx on endometrial tissue but failed to discern between epithelial and stromal cells relative to PG response. Results presented in Chapter 3 confirm the report by Fortier et al. (1988) describing epithelial cells as the major source of endometrial PGFj^ and stromal cells as the major source of endometrial PGEj. Present results (Chapter 3) indicate that IFNx can decrease prostaglandin secretion from endometrial epithelial cells but not from stromal cells. In addition, oxytocin-induced stimulation of prostaglandin secretion was observed only in epithelial cells, the major source of PGFj^; furthermore, oxytocin-induced secretion of PGFja was attenuated with IFNx treatment. The differential effects of IFNx and oxytocin on endometrial epithelial and stromal cells in primary culture provide a model for the differential effect of pregnancy on endometrial secretion of PGFj,, and PGEj. Responsiveness to IFNx and oxytocin may be used as physiological markers for the development of an endometrial epithelial cell line. Such model would be extremely valuable for the study of IFNx-induced intracellular events.

PAGE 203

196 The points in the intracellular cascade regulating uterine secretion of PGFja that are affected by the conceptus have not been identified clearly. Bazer (1992) hypothesized that the antiluteolytic effect of IFNx could be mediated through several potential mechanisms: (1) stabilization or up-regulation of endometrial progesterone receptors to maintain the progesterone block that prevents the synthesis of oxytocin receptors or the up-regulation of estrogen receptors; (2) direct inhibition of estrogen receptors which would attenuate oxytocin receptor synthesis and pulsatile release of PGFj^ ; (3) direct inhibition of oxytocin receptor synthesis; (4) alteration of post-oxytocin receptor mechanisms to inhibit oxytocin-induced release of PGF^aHowever, the nature of the relationship between IFNx secretion from the conceptus and the dynamics steroid receptors during early pregnancy is not totally clear. Between days 6 and 18, plasma concentrations of estradiol are higher in pregnant cows as compared to cyclic cows (Lukaszweska and Hansel, 1980; Hansel, 1981). In contrast, Schallenberger et al. (1989) reported lower estradiol concentrations in plasma of the aorta in pregnant cows compared to cyclic cows, from days 4-7 to at least day 35. During maternal recognition of pregnancy in sheep, endometrial progesterone receptor (PR) concentrations are maintained at concentrations characteristic of the mid-luteal phase, whereas in non-pregnant animals, PR are high during mid-diestrus and decrease during late diestrus (Findlay et al., 1982; Cherny et al., 1991; Ott et al., 1993). Endometrial estrogen receptor (ER) concentrations decreased from day 10 to 16 postestrus in pregnant ewes, whereas in cyclic ewes, ER remained relatively constant from day 10 to 14 and then increased dramatically at day 16 postestrus (Ott et al., 1993). In vivo, progesterone was necessary for olFNx to exert an inhibitory effect on oxytocinstimulated production of prostaglandin in the ewe (Ott et al., 1992). Intrauterine

PAGE 204

197 administration of oCSP (containing 100 i^g of olFNt/day) from day 11 to 15 postestrus could not totally replicate the PR and ER protein and mRNA concentrations observed in pregnant ewes at day 16 postestrus. Ovine CSP appeared to have a negative effect on PR protein and mRNA concentrations compared to serum albumin treated ewes at day 16 postestrus (Mirando et al., 1993b). Collectively, these observations do not define a clear sequence of events indicative of a direct regulation of endometrial steroid receptors by IFNx during early pregnancy. The absence of oxytocin responsiveness observed in vivo and in vitro in pregnant cattle and sheep is related directly to the fact that endometrial oxytocin receptors are decreased considerably in pregnant compared to cyclic ewes (McCracken et al., 1984; Flint et al., 1986) and cows (Fuchs et al., 1990; Jenner et al., 1991) during the luteolytic period. In Chapter 3, incubation of endometrial epithelial cells with IFNx for 24 h abolished oxytocin-induced stimulation of prostaglandin secretion. Furthermore, results in chapter 3 indicate that endometrial epithelial cells from day 15 cyclic cows are responsive to oxytocin. Responsiveness of epithelial cells to oxytocin suggests that receptors were present on their plasma membranes. Yet, 24 h of exposure to IFNx blocked responsiveness to oxytocin. Thus some alteration in the signal transduction system of oxytocin must be present. Vallet et al. (1989) showed that olFNx did not compete with oxytocin for its receptor. The mechanisms by which the conceptus suppresses oxytocin effects are still unclear. Endometrium from day 13-16 pregnant sheep contains fewer oxytocin receptors compared to cyclic ewes (McCracken et al.; 1984; Flint et al., 1986). Abayasekara et al. (1992) suggested that the ovine IFNx-induced decrease In endometrial oxytocin receptor expression was mediated by a

PAGE 205

198 decrease in PKC activity. PLC is part of the second messenger system activated upon oxytocin binding to its receptor. In pregnant ewes, the lack of oxytocin responsiveness was associated with a total suppression of oxytocin's ability to stimulate PLC activity (Mirando et al., 1990a; 1990b; Ott et al., 1993). The ovine conceptus, through secretion of ovine IFNx, may inhibit the synthesis of endometrial estrogen receptors that would inhibit induction of PGFj^ responsiveness to oxytocin during early pregnancy (Ott et al., 1993). Intrauterine injections of olFNx from day 11 to 15 decreased endometrial number and affinity of oxytocin receptors (Mirando et al., 1993b) possibly by downregulating endometrial ER and/or stabilizing endometrial PR. This would reduce the number of oxytocin receptors since they are induced by estradiol (Soloff, 1975). Understanding the regulation of oxytocin receptors by bIFNx is confounded by potential differences in steroids availability between pregnant and cyclic animals prior to the increase in bIFNx Silvia et al. (1993) suggested that a continuous presence of the conceptus was required to maintain oxytocin non-responsiveness in the endometrium, and suggested that a very labile factor could be involved in mediating the conceptus effect on oxytocin-responsiveness. PLCp activation being GTP-dependent, specific isoforms of G proteins may play a role in the lack of oxytocin responsiveness in the endometrium from pregnant cows. There is also evidence that tyrosine kinase-mediated phosphorylation is involved in regulation of PLCy activity (Anderson et al., 1990; Margolis et al., 1990). Limiting arachidonic acid availability (Chapter 2), possibly due to low PLA2 activity, could represent an additional mechanism regulating prostaglandin secretion in endometrium from pregnant cows. Availability of AA appears to be limiting to prostaglandin secretion in endometrial explants from pregnant cows

PAGE 206

199 (Chapter 2) since both treatment of explants with PLAj or addition of AA increased PGFj. secretion. The potential regulators of AA availability for metabolism by cyclooxygenase are numerous, and many interdependent strategies could result in decreased concentrations of nonesterified AA in the endometrium (Figure 5-1 ). Chapter 4 provided a more direct piece of evidence that pregnancy can induce a decrease in free AA concentrations in endometrial microsomes. Reduced PLAj activity is a possible mechanism that has not been explored extensively. The various isoforms of PLAj associated with different substrate specificity represent many candidates for regulation by the conceptus during early pregnancy. Lin et al. (1992) provided evidence that the 85 kDa PLAj can be involved in hormone-stimulated release of AA using transfected CHO cells over-expressing this enzyme. The 85 kDa PLAj has a strong preference for sn-arachidonoyi phospholipids, and decreased activity could be involved in the pregnancy-induced reduction in AA availability. The possibility that the conceptus induces synthesis of lipocortins in the endometrium cannot be ruled out. An antiphospholipase protein named gravidin has been described in amniotic fluid, and its production from the amnion decreases with labor (Wilson et al., 1992). Curl (1988) reported that the quantity of esterified AA in endometrial phospholipids is lower in pregnant than in cyclic cows. The mechanisms of synthesis and degradation of phospholipids involve numerous enzymatic activities. In addition, phospholipids are not distributed randomly among all cellular membranes, but instead different lipid classes are often enriched in different membranes. This enrichment may involve site-specific synthesis or degradation of phospholipids, remodeling through deacylation-reacylation reactions, translocation or some combination of these mechanisms (see Pagano,

PAGE 207

200 1990). One possibility is that AA is made selectively inaccessible to microsomal membranes containing cyclooxygenase enzymes. This could be achieved by selective targeting of AA-containing phospholipids to other cell membranes. The levels of free fatty acids are controlled by acyl-CoA transferase for esterification, and by acyl-CoA hydrolase for esterification into phospholipids. Uterine concentrations in fatty acyl-CoA were shown to be regulated by estradiol and progesterone in the rat (Young and Barker, 1991). Figure 5-1: The regulation of prostaglandin secretion in bovine endometrial epithelial cells by interferon x (IFNx) could involve interactions with • multiple signal transduction and metabolic pathways. The proposed effects of IFNx upon binding to type I interferon receptors illustrate possible mechanisms by which arachidonic acid (AA) availability could be reduced, linoleic acid (LA) levels could be increased and PGH synthase (PGHS) could be inhibited. Such effects could also be part of the IFNx signal transduction system leading ultimately to regulation of gene transcription. DAG, 1,2-diacylglycerol; IP3, inositol-triphosphate; OT, oxytocin; PI, phosphatidylinositol; PLA2, phospholipase A2; PLC, phospholipase C; PL, phospholipids; PKC, protein kinase C.

PAGE 208

201 The regulation of phospholipids and free fatty acid metabolism appear particularly relevant in view of the results reported in Chapter 4. EPSI activity was associated with free linoleic acid which is a competitive inhibitor of AA metabolism by cyclooxygenase (Figure 5-1). Increased free linoleic acid in microsomes appeared to be responsible for the higher EPSI activity detected in endometrium from pregnant versus cyclic cows. In addition, free AA concentrations were lower in microsomes from pregnant cows. The pregnancyinduced decrease in the ratio of LA:AA could be a consequence of a lower A6 desaturase activity in the endometrium. High affinity receptors for ovine IFNt have been described in the endometrium (Godkin et al., 1984). Interferon x appears to bind to the same receptor as other type I interferons (e.g., IFNa; Stewart et al., 1987). It is not clear whether some of the properties specific to IFNt involve accessory proteins at the level of the IFN type I receptor. The transduction signal(s) of IFNa and x are unknown but rapid changes in metabolism of phosphoinositides or cyclic nucleotides concentrations do not appear to be involved (Vallet et al., 1987; Pestkaetal., 1987). * The factor responsible for IFNa-induced transcriptional activation of certain genes is a complex of four proteins termed ISGF3 (interferon-stimulated gene factor 3) (Levy and Darnell, 1990; Figure 5-2). Three of these proteins, collectively termed ISGFSa, are found in the cytoplasm of non-stimulated cells (Levy et al., 1989). Upon stimulation by IFNa, the ISGFSa complex is translocated to the nucleus, where it binds a DNA-binding protein called ISGF3y to form the ISGF3 complex (Kessler et al., 1990). ISGF3 binds to a specific DNA

PAGE 209

sequence, the ISRE, and directs IFNa-dependent gene transcription in the nucleus (Dale et al., 19889; Bandypadhyay et a!., 1990). . , •„ " 1 There is growing evidence that unsaturated fatty acids such as linoleic and arachidonic acids, are involved in signal transduction of growth factors and interferons. Recently, Bandyopadhaya et al. (1993) reported that linoleic acid enhanced the EGF-induced proliferation of mammary epithelial cells. Linoleic acid can activate PKC and induce phosphorylation of specific proteins (Lester, 1990; Chen et al., 1992). Interestingly, Pfeffer et al. (1990) observed that Interferon a causes a rapid and selective activation of the isozyme p of PKC. Perturbation of AA metabolism, using inhibitors of cyclooxygenase and lipoxygenase, enhanced the IFNa stimulation of factor binding to the ISRE and increased the levels of 2-5 A synthetase mRNA and protein (Hannigan and Williams, 1991; Figure 5-1). This report suggested that redirection of AA metabolism in favor of the formation of epoxy derivatives had a synergistic effect on IFNa-induced gene expression (Figure 5-2). Arachidonic acid itself is not the second messenger since addition of exogenous AA did not effect the transaiption of IFN-stimulated genes (Yan et al., 1989). However, Tiwari et al. (1991) provided evidence that linoleic acid can regulate gene expression in vitro and can induce second messenger signals similar to the ones generated by interferons. In vivo, administration of cyclooxygenase inhibitors (aspirin and indomethacin) in combination with IFNa enhanced the antitumor effect of interferon a (Creagan et al., 1988; Kim and Warnaka, 1989). Recombinant bovine IFNx failed to increase EPSI activity and endometrial concentrations of free linoleic acid (Chapter 4). Possibly, IFNx was unable to mimic the effect of pregnancy on free LA and AA dynamics in the endometrium because of co-infusions with significant amounts of serum albumin in the uterine

PAGE 210

203 lumen. The use of a protein control with no fatty acid-binding properties may allow IFNx to increase EPSI activity and linoleic acid concentrations in the endometrium. Although, two daily infusions of rblFNx failed to increase EPSI activity and linoleic acid concentrations in the endometrium, an extension of the CL lifespan was detected in IFNx-treated cows (Meyer and W.W. Thatcher, unpublished observations). These observations suggests that the capability of rblFNx to extend CL lifespan is not necessarily totally dependent upon its ability to increase levels of linoleic acid in the endometrium. r r LA ARACHIDONIC AQD C^oxyg«T^^ IT EPOXY DERIVATIVES f LEUKOTRIENES (EETS and di-HETE'S) PROSTAGLANDWS j THROMBOXANES ^ "~^® /^""^^acUve ISGF3 ISGf3a-(p) + ISGF3y ^^Vcomplex \ 0 IFN a-stimulated IH » gene transcriptio n (2-5 A synthase) ISGF3 ISRE a Figure 5-2: Redirection of AA metabolism from prostaglandins and leukotrienes to epoxy derivatives has a synergistic effect on IFNa-stimulated expression of 2'-5'-oligoadenylate synthase (2-5 A synthase). ISGF. interferon-stimulated gene factor; ISRE, interferon-stimulated response element; LA, linoleic acid; PGH synthase, PGHS; PL, phospholipidsTK tyrosine kinase.

PAGE 211

In cx)nclusion, the conceptus appears to use multiple strategies to alter uterine secretion of PGRj^ in order to prevent luteolysis. Epithelial cells, the major source of endometrial PCFj^, are probably the prime target for the conceptus-secreted antiluteolytic factor: interferon x. Several lines of evidence indicate that pregnancy can induce significant changes in endometrial lipid metabolism. Indeed, endometrium from cows at day 17 postestrus is characterized by a limited availability of arachidonic acid. This is probably due to a dovi^nregulation of PLAj activity, decreased concentrations of unesterified AA and increased concentrations of unesterified linoleic acid. Since linoleic acid is a competitive inhibitor of cyclooxygenase, it appears also to be a regulatory component that directly reduces synthesis of PGF2. from AA. Altered routes of AA metabolism (e.g., epoxygenase) may lead to an enhancement in the regulation of interferon a-induced gene expression that contributes to a reduction in endometrial responsiveness to luteolytic agents (estradiol and oxytocin). This knowledge may ultimately benefit reproductive management systems and enhance embryo survival. For example, in approximately 20% of the animals, IFNx and IFNa fail to prolong CL lifespan. Perhaps their uterine capability to respond to antiluteolytic proteins (interferon x) is compromised by a non-optimal lipid environment in the endometrium that reduces the efficacy of the interferon signal transduction system. Since EPSI appears to be linoleic acid, it is probably feasible to alter lipid composition (e.g., linoleic acid) of the endometrium by feeding ruminantly inert unsaturated fatty acids that are enriched in linoleic acid. Perhaps such a feeding system may enhance the antiluteolytic mechanism contributing to improved embryo survival in dairy cattle.

PAGE 212

REFERENCES Abayasekara D.R.E., Sheldrick E.L.R., Flick-Smith H.C., Commander H. and Flint A.P.F. 1992. Protein kinase C in the sheep endometrium: mechanism by which interferon inhibits oxytocin receptor expression. J. Reprod. Fertil. 9: 40 (abst). Abdelgadir S.E., Swanson L.V., Oldfield J.E. and Stormshak F. 1987. Prostaglandin Fj^ -induced release of oxytocin from bovine corpora lutea in vitro. Biol. Reprod. 37:550. Abdel-Rahim S.E., Bland K.P., and Peyser N.L. 1983. Prostaglandin F2 alpha and PGE2 in uterine lymph during the oestrous cycle in sheep. Prostaglandins Leukot. Med. 10:157. Abdel-Rahim S.E., Bland K.P., and Poyser N.L. 1984. Surgical separation of the uterus and ovaries with simultaneous cannulation of the uterine vein extends luteal function in the sheep. 1984. J. Reprod. Fertil. 72:231. Abita J. and Lazdunski M. 1972. Zymogen-enzyme transformations. Eur. J. Biochem. 30:37. Adashi E.Y. 1989. Putative intraovarian regulators. Semin. Reprod. Endocrinol 7:1. Adashi E.Y., Resnick C.E., Hernandez E.R., Hunvitz A. and Rosenfeld R.G. 1989. Rodent studies on the potential relevance of IGF-I to ovarian physiology. In: Hirshfield A.N. (Ed) Growth factors and the ovary, p 95. Plenum, New York. Adashi E.Y., Resnick C.E., Hurwitz A, Ricciarelli E., Hernandez E.R., Roberts C.T., LeRoith D. and Rosenfeld R.G. 1991. Insulin-like growth factors: the ovarian connection. Hum. Reprod. 6:1213. Adashi E.Y., Resnick C.E., Hurwitz A., Ricciarelli E., Hernandez E.R., Roberts C.T., LeRoith D. and Rosenfeld R.G. 1992. The inter-ovarian IGF system. Growth Regul. 2:10. 205

PAGE 213

206 Alila H.W., Corradino R.A. and Hansel W. 1989. Differential effects of luteinizing hormone on intracellular free Ca^* in small and large bovine luteal cells. Endocrinology 124:2314. Alila H.W., Corradino R.A. and Hansel W. 1990. Arachidonic acid and its metabolites increase cytosolic free calcium in bovine luteal cells. Prostaglandins 39:481 . Alila H.W., Dowd J. P. and Corradino R.A. 1988a. The role of calcium in progesterone synthesis by small and large bovine luteal cells. Biol. Reprod. (Suppl. 1):55 (abst). Alila H.W., Dowd J. P., Corradino R.A., Harris W.V. and Hansel W. 1988b. Control of progesterone production in small and large bovine luteal cells separated by flow cytometry. J. Reprod. Fertil. 83:645. Alila H.W. and Hansel W. 1984. Origin of different cell types in the bovine corpus luteum as characterized by specific monoclonal antibodies. Biol. Reprod. 31:1015. Anderson D., Koch C.A., Grey L., Ellis C, Moran M.F. and Pawson T. 1990. Binding of SH2 domains of phospholipase Cyl, GAP, and Src to activated growth factor receptors. Science 250:979. Anthony R.V., Helmer S.D., Sharif S.F., Roberts R.M., Hansen P.J., Tahtcher W.W. and Bazer F.W. 1988. Synthesis and processing of ovine trophoblast protein-1 and bovine trophoblast protein-1, conceptus secretory proteins involved in the maternal recognition of pregnancy. Endocrinology 123:1274. Ayad V.J., Guldenaar S.E., and Wathes D.C. 1991. Characterization and localization of oxytocin receptors in the uterus and oviduct of the nonpregnant ewe using iodinated receptor antagonist. J. Endocrinol. 128:187. Balapure A.K., Caicedo I.C., Kawada K., Watt D.S., Rexroad C.E. and Fitz T.A. 1989. Multiple classes of PGFj,, biding sites in subpopulations of ovine luteal cells. Biol. Reprod. 41:385. Bandyopadhyay G.K., Hwang S.I., Imagawa W. and Nandi W. 1993. Role of polyunsaturated fatty acids as signal transducers: amplification of signals from growth factor receptors by fatty acids in mammary epithelial cells. Prostaglandins Leukot. Essent. Fatty Ac. 48:71.

PAGE 214

207 Bandyopadhyay S.K., Kalvakolanu D.V.R. and Sen G.C. 1990. Gene induction by interferons: functional complementation between trans-acting factors induced by alpha interferon and gamma interferon. Mol. Cell. Biol. 10:5055. Barros CM., Betts J.G., Thatcher W.W. and Hansen P.J. 1992a. Possible mechanisms for reduction of circulating concentrations of progesterone by interferon-a in cows: effects on hyperthermia, luteal cells, metabolism of progesterone and secretion of LH. J. Endocrinol. 133:175. Barros CM., Newton G.R., Thatcher W.W., Drost M., Plante C and Hansen P.J. 1992b. The effect of bovine interferon-a, 1 on pregnancy rate in heifers. J. Anim. Sci. 70:1471. Barros CM., Plante C, Thatcher W.W. and Hansen P.J. 1991. Regulation of bovine endometrial secretion of prostaglandins and synthesis of 2', 5'oligoadenylate synthetase by interferon-a molecules. Am. J. Reprod. Immunol. 25:146. Bartol F.F., Roberts R.M., Bazer F.W., Lewis G.S., Godkin J.D. and Thatcher W.W. 1985a. Characterization of proteins produced in vitro by periattachment bovine conceptuses. Biol. Reprod. 32:681. Bartol F.F., Roberts R.M., Bazer F.W. and Thatcher W.W. 1985b. Characterization of proteins produced in vitro by bovine endometrial explants. Biol. Reprod. 33:745. Basha S.M.M., Bazer F.W., and Roberts R.M. 1980. Effect of the conceptus on quantitative and qualitative aspects of uterine secretion in pigs. J. Reprod. Fertil. 60:41. Basu S. and Kindahl H. 1987a. Development of a continuous blood collection technique and a detailed study of PGFj^ during luteolysis and early pregnancy in heifers. J. Vet. Med. A 34:487. Basu S. and Kindahl H. 1987b. Prostaglandin biosynthesis and its regulation in the bovine endometrium: a comparaison between nonpregnant and pregnant status. Theriogenology 28:175. Battye K.M., Ammit A.J., O'Neill C and Evans G. 1991. Production plateletactivating factor by the pre-implantation sheep embryo. J Reprod Fertil. 93:507. Bazer F.W. 1992. Minireview: Mediators of maternal recognition of pregnancy in mammals. Proc. Soc. Exp. Biol. Med. 199:373.

PAGE 215

208 Behrman H.R. and Hall A.K. 1982. Antagonistic interactions of adenosine and prostaglandin F2 alpha modulate acute responses of luteal cells to 5 , h ; luteinization hormone. Endocrinology 110:38. Benyo D.F., Haibel G.K., Laufman H.B. and Pate J.L. 1991. Expression of major histocompatibility complex antigens on the bovine corpus luteum during the estrous cycle, luteolysis, and early pregnancy. Biol. Reprod. 45:229. Betteridge K.J., Randall G.C.B., Eaglesome M.D. and Granstrom E. 1984. The influence of pregnancy on PGFza secretion in cattle. I. Concentrations of 15-keto-13,14-dihydro-prostaglandin and progesterone in peripheral blood of recipients of transferred embryos. Animal Reprod. Sci. 7:195. 1984. Birnbaumer L. 1990. G proteins in signal transduction. Annu. Rev. Pharmacol. Toxicol. 30:675. Bligh E.G. and Dyer W.J. 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37:91 1 . Bocci v. 1983. What is the role of carbohydrates in interferons? Trends Biochem. Sci. 12:432. Bonney R.C. 1985. Measurement of phospholipase A2 activity in human endometrium during the menstrual cycle. J. Endocrinol. 107:183. Bonney R.C. and Franks S. 1987. Modulation of phospholipase A2 activity in human endometrium and amniotic membrane by steroid hormones. J. Steroid Biochem. 26:467. Bonney R.C, Qizilbash S.T. and Franks S. 1987. Endometrial PLAj enzymes and their regulation by steroid hormones. J. Steroid Biochem. 27:1057. Boshier D.P., Fairclough R.J. and Holloway H. 1987. Assessment of sheep blastocyst effects on neutral lipids in the carancular epithelium J Reprod. Fertil. 79:569. Bourne H.R., Sanders D A. and McCormick F. 1990. The GTPase superfamily: A conserved switch for diverse cell functions. Nature 348:125. Brenner R.R. 1990. Endocrine control of fatty acid desaturation. Biochem. Soc. Trans. 18:773.

PAGE 216

209 Brenner R.R. and Peluffo R. 1966. Effect of saturated and unsaturated fatty acids on the desaturation in vitro of palmitic, stearic, oleic, linoleic and linolenic acids. J. Biol. Chem. 241:5213. Brinsfield T.H. and Hawk H.W. 1973. Control by progesterone of the concentration of lipid droplets in epithelial cells of the sheep endometrium. J. Anim. Sci. 36:919. Bult H., Coene M.C., Claeys M., Leakeman and Herman A.G. 1987. Hydroxy and hydroperoxy derivatives of linoleic and arachidonic acid suppress the activation of rabbit blood platelets. Adv. Prostaglandin Thromboxane Leukot. Res. 17:224. Caffrey J.L., Nett T.M., Abel J.H. Jr and Niswender G.D. 1979. Activity of 3phydroxysteroid dehydrogenase / A5-A4 isomerase in the ovine corpus luteum. Biol. Reprod. 20:279. Cahil L.P. and Mauleon P. 1980. Influences of season, cycle and breed on follicular growth rates in sheep. J. Reprod. Fertil. 93:81 . Capon D.J., Shepard H.M. and Goeddel D.V. 1985. Two distinct families of human and bovine interferon-a genes are coordinately expressed and encode functional polypeptides. Mol. Cell. Biol. 5:768. Carson R., Trounson A. and mitchell M. 1986. Regulation of prostaglandin biosynthesis by human ovarian follicular fluid: a mechanism for ovulation? Prostaglandins 32:49. Carson R.S., Findlay J.K., Clarke I.J. and Burger H.G. 1981. Estradiol, testosterone and androstenedione in ovine follicular fluid during growth and atresia of ovarian follicles. Biol. Reprod. 24:105. Casey M., Korte K. and McDonald P.C. 1988. Epidermal growth factor stimulation of prostaglandin E2 biosynthesis in amnion cells. J Biol Chem. 263:7846. Catella F., Lawson J., Fitzgerald D. and Fitzgerald G. 1990. Endogenous biosynthesis of arachidonic acid epoxides in human: increased formation in pregnancy-induced hypertension. Proc. Natl. Acad Sci USA 87:5893. Cernescu C, Constantinescu S.T., Balta F. and Popescu L.M. 1989. Protein kinase C and the antiviral effect of human interferon. Rev. Roum Med Virol. 40:163.

PAGE 217

210 Charlier M., Hue D., Martal J. and Gaye P. 1989. Cloning and expression of cDNA encoding for ovine trophoblastin: Its identity with a class II alpha interferon interferon. Gene 77:341 . Charlier M., Hue D., Martal J. and Gaye P. 1991. Cloning and structural analysis of two distinct families of ovine interferon-alpha genes encoding functional class II and trophoblast (oTP) alpha-interferons. Mol. Cell. Endocrinol. 76:161. Charpigny G., Danet-Desnoyers G., Reinaud P., Guillemot M. and Martal J. 1991. Ovine trophoblast protein (trophoblastin) inhibits basal and oxytocin-induced secretion of PGFj^ by ovine endometrial cells. J. Reprod. Fertil. Suppl. 43:51. Charpigny G., Reinaud P., Huet J.-C, Guillemot M., Charlier M., Pernollet J.-C. and Martal J. 1988. High Homology between a trophoblast protein (trophoblastin) isolated from ovine embryo and a-interferons. FEBS Lett. 228:12. Chegini N. and Rao C.V. 1987. Dynamics of nuclear associated granules in bovine luteal cells after treatment in vitro with PGEj^. Endocrinology 121:1870. Cherny R.A. and Findlay J.K. 1990. Separation and culture of ovine endometrial and stromal cells: evidence of morphological and functional polarity. Biol. Reprod. 43:241. Cherny R.A., Salamonsen L.A. and Findlay J.K. 1991. Immunochemical loccalization of oestrogen receptors in the endometrium of the ewe. Reprod. Fertil. Dev. 3:321. Christie W.H. 1985. Rapid separation and quantification of lipid classes by high performance liquid chromatography and mass (light scattering) detection. Anal. Chem. 57:1342. Clark J.D., Lin L.L., Kriz R.W., Ramesha C.S., Sultzman L.A., Lin A.Y., Milano N., and Knopf J.L. 1991. A novel arachidonic acid-selective cytosolic PLA2 contains a Ca(2+)-dependent translocation domain with homology to PKC and GAP. Cell 65:1043. Clark J.D., Milona N. and Knopf J.L. 1990. Purification of a 110 kDa cytosolic phospholipase Aj from the human monocytic cell line U937 Proc Natl Acad. Sci. USA 87:7708.

PAGE 218

211 Clark M.A., Chen M.-J., Crooke ST. and Bomalaski J.S. 1988.Tumor necrosis factor (cachetin) induces a PLAj-activating protein in endothelial cells. Biochem J. 250:125. Clark M.A., Conway T.M., Shorr R.G.L. and Crooke S.T. 1987. Identification and isolation of a mammalian protein which is antigenically and functionally related to the PLAj stimulating peptide melittin. J. Biol. Chem. 262:4402. Clemens M.J. and McNurlan M.A. 1985. Regulation of cell proliferation and differentiation by interferons. Biochem. J. 226:345. Cockcroft S. and Gomperts B.D. 1985. Role of guanine nucleotide binding protein inthe activation of poly-phosphoinositide breakdown. Nature 314:534. Cook H.W., 1978. In vitro formation of polyunsaturated fatty acids by desaturation in rat brain: some properties of the enzymes in developing brain and liver comparaison. J. Neurochem. 30:1327. Cook H.W. 1991. Phospholipases. In: Vance D.E. and Vance J. (Eds) Biochemistry of lipids, lipoproteins and membranes, p 141. Elsevier, Amsterdam. Corey E.J., Shih C. and Cashman J.R. 1983. Docosahexaenoic acid is a strong inhibitor of prostaglandin but not leukotriene biosynthesis. Proc. Natl Acad. Sci. USA 80:3581. Cortizo A.M., Besterman J.M., Leitner P.P. and Chandrabose K.A. 1992. Stimulated release of arachidonate and prostaglandins is vectorial in MDBK epithelial cells. Prostaglandins 44:357. Couet J., Martel C, Dupont E., The V.L., Sirard M.A., Pelletier G. and Labrie F. 1990. Changes in 3p-hydroxysteroid dehydrogenase / A^-A^ isomerase messenger RNA, activity and protein during the estrous cycle in the bovine ovary. Endocrinology 127:2141. Creagan E.T., Buckner J.C., Hahn R.G., Richardson R.R., Schaid D.J. and Kovach J.B. 1988. An evaluation of recombinant leukocyte A interferon with aspirin in patients with metastatic renal cell cancer Cancer 61:1787. Cross J.C. and Roberts R.M. 1991. Constitutive and trophoblast-specific expression of a class of bovine interferon genes. Proc Natl Acad Sci USA 88:3817.

PAGE 219

212 Crumpton M.J. and Dedman J.R. 1990. Protein terminology tangle (letter). Nature 345:212. Gulp B.R., Titus G.G. and Lands W.E. 1979. Inhibition of prostaglandin biosynthesis by eicosapentaenoic acid. Prostaglandins Med. 3:269. Curl J.S. 1988. Uterine lipid composition and endometrial prostaglandin synthesis and transport in relationship to maternal recognition of pregnancy and luteolysis in the bovine. Dissertation (Ph.D.) University of Florida. Curl J.S., Thatcher W.W. and Bartol F.F. 1983. In vitro production of PGEj, PGFja and PGFM by endometrial tissue from cattle at day 17 of pregnancy or the estrous cycle. J. Anim. Sci. (Suppl. 1 ):327. Dale T.C., Imam A.M.A., Kerr I.M. and Stark G.R. 1989. Rapid activation by interferon a of a latent DNA-binding protein present in the cytoplasm of untreated cells. Proc. Natl. Acad. Sci. USA 86:1203. David M., Romero G., Zhang Z.-Y., Dixon J.E. and Larner A.C. 1993. In vitro activation of the transcription factor ISGF3 by interferon a involves a membrane-associated tyrosine phosphatase and tyrosine kinase. J. Biol. Chem. 268:6593. Davidson F.F. and Dennis E.A. 1989. Biological relevance of lipocortins and related proteins as inhibitors of phospholipase A2. Biochem. Pharmacol 38:3645. Davis J.S., Alila H.W., West L.A., Corradino R.A., Weakland L.L. and Hansel W. 1988. Second messenger systems and progesterone secretion in the small cells of the bovine corpus luteum: effects of gonadotropins and prostaglandin F2„. J. Steroid Biochem. 32:643. Davis J.S. and Clark M.R. 1983. Activation of PKC in the bovine corpus luteum by phospholipid and Ca^* Biochem. J. 214:569. Dawson R.M.C. and Irvine R.F. 1978. Possible role of lysosomal phospholipases in inducing tissue prostaglandin synthesis. In: Galli C. Galli G. and Porcellati G. (Eds) Phospholipases and prostaglandins, p 47. Raven Press, New York. Decker T., Lew D.J. and Darnell J.E. 1991. Two distinct alpha interferondependent signal transduction pathways may contribute to activation of transcription of the guanylate protein binding gene. Mol Cell Biol 11:5147.

PAGE 220

213 Demers L.M., Behrman H.R., and Greep R.O. 1973. Effects of prostaglandins and gonadotrophins on luteal prostaglandin and stroid biosynthesis. Adv. Biosci. 9:701. Danker H.W. 1990. Trophoblast-endometrial interactions at embryo implantation: A cell biological paradox. Troph. Res. 4:3. Dennis E.A. 1983. Phospholipases . In: Boyer P. (Ed) The enzymes . Vol. XVI, p 307. Academic Press, New York. Dennis E.A. 1987. Regulation of eicosanoid production: role of phospholipases and inhibitors. Biotechnology 5:1294. DeWitt D.L. 1991. Prostaglandin endoperoxide synthase: regulation of enzyme expression. Biochim. Biophys. Acta 1083:121. Dey S.K., Hoversland R.C. and Johnson D.C. 1982. Phospholipase A2 activity in the rat uterus: modulation by steroid hormones. Prostaglandins 23:619. Diaz A., Reginato A.M. and Jimenez S.A. 1992. Alternative splicing of human prostaglandin G/H synthase mRNA and evidence of differential regulation of the resulting transcripts by transforming growth factor beta 1, interleukin 1 beta, and tumor necrosis factor alpha. J. Biol. Chem. 267:10816. Diez E., Louis-Flamberg P., Hall R.H., and Mayer R.J. 1992. Substrate specificities and properties of human phospholipases Aj in mixed vesicule model. J. Biol. Chem. 267:18342. Di Rosa M. Flower R.J., Hirata F., Parente L. and Russo-Marie F. 1984. Nomenclature announcement. Antiphospholipase proteins. Prostaglandins 28:441 . Domin J. and Rosengurt E. 1993. Platelet-derived growth factor stimulates a biphasic mobilization of arachidonic acid in Swiss 3T3 cells J Biol Chem. 268: 8927. Downing I. and Poyser N.L. 1983. Estimation of phospholipase A2 activity in guinea pig endometrium on days 7 and 16 of the estrous cycle Prost Leukot. Med. 12:107. Drew P.D.. Lonergan M., Goldstein M.E., Lampson LA., Ozato K. and McFarlin D.E. 1993. Regulation of MHC class I and Pj-f^'croglobulin gene expression in human neuronal cells. Factor binding to conserved cis-

PAGE 221

214 acting regulatory sequences correlates with expression of the genes. J. Immunol. 150:3300. Driancourt M.A., Fry R.C., Clarke I.J. and Cahill L.P. 1979. Follicular growth and regression during the 8 days after hypophysectomy in sheep. J. Reprod. Fertil. 79:635. Dron M., Tovey M.G. and Uze G. 1986. Isolation of Daudi cells with reduced sensitivity to interferon. IV. Characterization of clones with altered binding of human interferon alpha subspecies. J. Gen. Virol. 67:663. Duncan R.A. and Davis J.S. 1991. Prostaglandin Fj^ stimulates inositol 1,4,5trisphosphate and inositol 1,3,4,5-tetrakisphosphate formation in bovine luteal cells. Endocrinology 128:1519. Eggleston D.L., Wilken C, Van Kirk E.A., Slaughter R.G., Ji T.H. and Murdoch W.J. 1990. Progesterone induces expression of endometrial messenger RNA encoding for cyclooxygenase (sheep). Prostaglandins 39:675. Eley R.M., Thatcher W.W., Bazer F.W. and Fields M.J. 1983. Steroid metabolism by the bovine uterine environment and conceptus. Biol. Reprod. 28:804. Ellinwood W.E., Nett T.M. and Niswender G.D. 1979. Maintenance of the corpus luteum of early pregnancy in the ewe. I. Luteotrophic properties of embryonic homogenates. Biol. Reprod. 21:281. Erickson B.H. 1966. Development and senescence of the postnatal bovine ovary. J. Anim. Sci. 25:800. Esteban M. and Paez E. 1985. Antiviral and antiproliferative properties of interferons: mechanism of action. Prog. Med. Virol. 32:159. Fairclough R.J., Moore L.G., Peterson A.J. and Watkins W.B. 1984. Effects of oxytocin on plasma concentrations of 13,14-dihydro-15-keto prostaglandin F and the oxytocin-associated neurophysin during the estrous cycle and early pregnancy in the ewe. Biol. Reprod. 31:36. Farin C.E., Imakawa K., Hansen T.R., McDonnell J.J., Murphy C.N., Farin P.W. and Roberts R.M. 1990. Expression of trophoblastic interferon genes in sheep and cattle. Biol. Reprod. 43:210. Farin C.E., Sawyer H.R. and Niswender G.D. 1989. Analysis of cell types in the corpus luteum of the sheep. J. Reprod. Fertil. Suppl. 37:181.

PAGE 222

215 Ferris CD., Huganir R.L., Supattapone S. and Snyder S.H. 1989. Purified inositol 1, 4, 5-tris-phosphate receptor mediates Ca^* flux in reconstituted lipid vesicles. Nature 342:87. Fielding P.E. and Fielding C.J. 1991. Dynamics of liportein transport in the circulatory system. In: Vance D.E. and Vance J. (Eds) Biochemistry of lipids, lipoproteins and membranes, p 427. Elsevier, Amsterdam. Fields M.J., Dubois W. and Fields P.H. 1985. Dynamic features of secretory granules: ultrastructural changes during the course of pregnancy in the cow. Endocrinology 117:1675. Fillion C, Chaouat G., Reinaud P., Charpigny G. and Martal J. 1991. Immunoregulatory effects of ovine trophoblastin protein (oTP): all five isoforms suppress PHA-induced lymphocyte proliferation. J. Reprod. Immunol. 19:237. Fincher K.B., Bazer F.W., Hansen P. J., Thatcher W.W. and Roberts R.M. 1986. Proteins secreted by the sheep conceptus suppress induction of uterine PGFja release by estradiol and oxytocin. J. Reprod. Fertil. 76:425. Findlay J.K., Clarke I. J., Swaney J., Colvin N. and Dougton B. 1982. Oestrogen receptors and protein synthesis in caruncular and intercaruncular endometrium of sheep before implantation. J. Reprod. Fertil. 64:329. Firmbach-Kraft I., Byers M., Shows T., Dalla-Favera R. and Krolewski J.J. 1990. Tyk2, prototype of a novel class of non-receptor tyrpsine kinase genes. Oncogenes 5:1329. Fisher S.K., Van Rooijen L.A.A. and Agranoff B.W. 1984. Renewed interest in the polyphosphoinositides. Trends Biochem. Sci. 9:53. Fitzpatrick F.A., Ennis M., Baze M., Wynalda M., McGee M. and Ligget W. 1986. Inhibition of cyclooxygenase activity and platelet aggregation by epoxyeicosatrienoic acids; influence of stereochemistry. J. Biol. Chem. 261:15334. Fitzpatrick F.A. and Murphy R.C. 1989. Cytochrome P-450 metabolism of arachidonic acid: formation and biological actions of "epoxygenase"derived eicosanoids. Pharmacol. Rev. 40:229. Flint A.P.F., Leat W.M.F., Sheldrick E.L. and Stewart H.J. 1986. Stimulation of phosphoinositides hydrolysis by oxytocin and the mechanism by which oxytocin controls prostaglandin synthesis in the ovine endometrium. Biochem. J. 237:797.

PAGE 223

216 Flint A.P.F. and Sheldrick E.L. 1982. Ovarian oxytocin is stimulated by prostaglandins. Nature 297:587. Flint A.P.F. and Sheldrick E.L. 1983. Evidence for a sytemic role for ovarian oxytocin in luteal regression in sheep. J. Reprod. Fertil. 67:215. Flint A.P.F. and Sheldrick E.L. 1986. Ovarian oxytocin and maternal recognition of pregnancy. J. Reprod. Fertil. 76:831. Flint A.P.F., Sheldrick E.L, Jones D.S.C. and Auletta F.J. 1989. Adaptation to pregnancy in the interactions between luteal oxytocin and the uterus in ruminants. J. Reprod. Fertil. Suppl. 37:195. Flower R.J. and Backwell G.J. 1976. The importance of phospholipase Aj in prostaglandin biosynthesis. Biochem. Parmacol. 25:285. Flowers B., Canteley T.C., Martin M.J. and Day B.N. 1991. Episodic secretion of gonadotropins and ovarian steroids in jugular and utero-ovarian vein plasma during the follicular phase of the estrous cycle in gilts. J. Reprod. Fertil. 91:101. Fortier M.A., Guibault L.A. and Grasso F. 1988. Specific properties of eptithelial and stromal cells from the endometrium of cows. J. Reprod. Fertil. 83:239. Fortune J.E. and Armstrong D.T. 1978. Hormonal control of 17-pestradiol biosynthesis in proestrous rat follicles: estradiol production by isolated theca versus granulosa. Endocrinology 102:227. Foxcroft G.R. 1978. The development of pituitary gland function. In: Crighton D.B., Foxcroft G.R., Haynes N.B. and Lamming G.E. (Eds) Control of ovulation, p 129. Butterworth Science, London, England. Francis H., Keisler D.H. and Roberts R.M. 1991. Induction of yhe synthesis of the pregnancy-specific protein 70 in the endometrium by intramuscular injection of recombinant bovine interferon-all in nonpregnant ewes. J. Reprod. Fertil. 93:367. Francis H., Keisler D.H. and Roberts R.M. 1992. Intramuscular injection of recombinant bovine interferon-a|1 into nonpregnant ewes induces the synthesis of the pregnancy-specific protein, p70, by endometrium. J. Reprod. Fertil. 93:367.

PAGE 224

217 Fu X.-Y. 1992. A transcription factor with SH2 and SH3 domains is directly activated by an interferon a-induced cytoplasmic protein tyrosin kinase(s). Cell 70:323. Fu X.-Y., Kessler D.S., Veals S.A., Levy D.E., and Darnell Jr. J.E. 1990. ISGF3, the transcriptional activator induced by interferon-a consists of multiple interacting polypeptide chains. Proc. Natl. Acad. Sci. USA 87:8555. Fuchs A.R., Behrens 0., Helmer H., Liu C.-H., Barros CM. and Fields M.J. 1990. Oxytocin and vasopressin in receptors in bovine endometrium and myometrium during the estrous cycle and early pregnancy. Endocrinology 127:629. Fujiwara Y., Ishibashi T. and Imai Y. 1984. Cytoplasmic localtion of linoleoylCoA desaturase in microsomal membranes of rat liver. Arch. Biochem. Biophys. 233:402. Funk CD. and Powell W.S. 1983. Metabolism of linoleic acid by prostaglandin endoperoxide synthase from adult and fetal blood vessels. Biochem. Biophys. Acta 754:57. Funk CD. and Powell W.S. 1985. Release of prostaglandins and monohydroxy and trihydroxy metabolites of linoleic and arachidonic acids by fetal and adult aortae and ductus arteriosus. J. Biol. Chem. 260:7481. Gadsbey J.E., Balpure A., Britt J.H. and Fitz T.A. 1990. Prostaglandin Fj^ receptors on enzyme-dissociated pig luteal cells throughout the estrous cycle. Endocrinology 126:787. Geisert R.D., Zavy M.T., Biggers B.G., Garrett J.E. and Wettemann R.P. 1988. Characterization of the uterine environment during early conceptus expansion in the bovine. Anim. Reprod. Sci. 16:11. Geisthovel F., Moretti-Rojas I., Rojas F.J. and Asch R.H. 1990. IGFs and thecagranulosa-cell function. Hum. Reprod. 5:785 Gemsa, D., Seitz M., Kramer W., Grimm W.. Till G., and Resch K. 1979. lonophore A23187 raises cyclic AMP levels in macrophages by stimulating prostaglandin E formation. Exp. Cell Res. 188:55. Ginther O.J. 1974. Internal regulation of physiological processes through local venoarterial pathways: a review. J. Anim. Sci. 39:550. Ginther O.J. 1981. Local versus systemic uteroovarian relationships in farm animals. Acta Vet. Scand. Suppl.77:103.

PAGE 225

218 Ginther O.J., Woody CO., Mahajan S., Jankiraman K. and Casida L.E. 1967. Effect of oxytocin administration on the oestrous cycle of unilaterally hysterectomized heifers. J. Reprod. Fertil. 14:225. Glasser S.R. and Mulholland J. 1993. Receptivitity is a polarity dependent special function of hormonally regulated uterine epithelial cells. Microsc. Res. Tech. 25:106. Coding J.R., McCracken J.A. and Baird D.T. 1967. The study of ovarian function in the ewe by means of a vasculature autotransplantation technique. J. Endocrinol. 39:37. Godkin J.D., Bazer F.W., Moffatt J., Sessions F. and Roberts R.M. 1982. Purification and properties of a major, low molecular weight protein released by the trophoblast of sheep blastocysts at day 13-21. J. Reprod. Fertil. 65:141. Godkin J.D., Bazer F.W. and Roberts R.M. 1984a. Ovine trophoblast protein-1, an early secreted blastocyst protein, binds specifically to uterine endometrium and affects protein synthesis. Endocrinology 114:120. Godkin J.D., Bazer F.W., Thatcher W.W. and Roberts R.M. 1984b. Proteins released by cultured day 15-16 conceptuses prolong luteal maintenance when introduced into the uterine lumen of cyclic ewes. J. Reprod. Fertil. 71:57. Goodman D.S. 1958. The interaction of human serum albumin with long-chain fatty acid anions. J. Am. Chem. Soc. 80:3892. Goppelt-Strebe M., Kroener C.F., Hausmann G., Gemsa D., and Resch K. 1986. Control of prostanoid synthesis: role of reincorporation of released precursor fatty acids. Prostaglandins 1986:32:373. Gould K.L., Woodgett J.R., Isacke CM. and Hunter T. 1986. The proteintyrosine kinase substrate p36 is also a substrate for PKC in vitro and in vivo. Mol. Cell. Biol. 6:2738. Goulding N.J. and Guyre P.M. 1992. Regulation of inflammation by lipocortin 1. Immunol. Today 13:295. , ., ^ Grastrom E. 1981. Prostaglandin chemistry. Acta Vet. Scand. Suppl. 77:1.

PAGE 226

219 Greenwald G.S. and Terranova P.F. 1988. Follicular selection and its control. In: Knobil E. and Neill J. (Eds) The physiology of reproduction p 387. Raven Press, New York. Gross T.S., Mirando M.A., Young K.H., Beers S., Bazer F.W., and Thatcher W.W. 1990. Reorientation of prostaglandin F (PGF) secretion by calcium ionophore, oestradiol, and prolactin in perifused porcine endometrium. Endocrinology 127:637. Gross T.S., Thatcher W.W., Hansen P.J., Johnson J.W. and Helmer S.D. 1988a. Presence of an intracellular endometrial inhibitor of prostaglandin synthesis during early pregnancy in the cow. Prostaglandins 35:359. Gross T.S.. Thatcher W.W., Hansen P.J. and Lacroix M.-C. 1988b. Prostaglandin secretion by perifused bovine endometrium secretion towards the myometrial and luminal sides at day 17 post-estrus as altered by pregnancy. Prostaglandins 35:343. Gross T.S., TIante C, Thatcher W.W., Hansen P.J., Helmer S.D. and Putney D.J. 1988c. Secretory proteins of the bovine conceptus alter endometrial prostaglandin and protein secretion in vitro. Biol. Reprod. 39:977. Grossberg S.E., Taylor J.L. and Kushnaryov V.M. 1989. Interferon receptors and their role in interferon action. Experientia 45:508. Grummer R.R. and Carroll D.J. 1988. A review of lipoprotein cholesterol metabolism: importance to the ovarian function. J. Anim. Sci. 66:3160. Guibault L.A., Thatcher W.W., Foster D.B. and Caton D. 1984. Relationship of 15-keto-13,14-dihydro-prostaglandin Fj^ concentrations in peripheral plasma with local uterine production of F series prostaglandins and changes in uterine blood flow during postpartum period of cattle. Biol. Reprod. 31:870. Guillemot M., Michel C, Gaye P., Charier M., Trojan J. and Martal J. 1990. Cellular localization of an embryonic interferon, ovine trophoblastin and its mRNA in sheep embryos during early pregnancy. Biol. Cell. 68:205. Guyton A.C. 1971. Transport through the cell membranes. In: Guyton (Ed.), Textbook of medical physiology, p 39. Saunders, Philadelphia. Habenicht A.J., Salbach P., Goerig M., Zeh W., Janssen-Timmen U., Blattner C, King W.C., and Glomset J.A. 1990. The LDL receptor pathway delivers

PAGE 227

220 arachidonic acid for eicosanoid formation in cells stimulated by PDGF. Nature 345:634. Hamberg M. and Samuelsson B. 1967a. Oxygenation of unsaturated fatty acids catalysed by vesicular gland of sheep. J. Biol. Chem. 242:5344. Hamberg M. and Samuelsson B. 1967b. On the mechanism of biosynthesis of prostaglandin E, and Fi„ J. Biol. Chem. 242:5336. Hamberg M. and Samuelsson B. 1973. Detection and isolation of an endoperoxide intermediate in prostaglandin biosynthesis. Proc. Natl. Acad. Sci. USA 70:699. Hammarstrom S., Kylden U., Powell W.S. and Samuelsson B. 1975. Solubilization of prostaglandin receptor in bovine corpora lutea. FEBS Lett. 50:306. Hanasaki K. and Arita H. 1992. Characterization of high affinity binding site for pancreatic-type phospholipase Aj in the rat. Its cellular and tissue distribution. J. Biol. Chem. 265:6414. Hannigan G.E. and Williams B.R.G. 1991. Signal transduction by interferon-a through arachidonic acid metabolism. Science 251 :204. Hansel W. 1966. Luteotrophic and luteolytic mechanisms in bovine corpora lutea. J. Reprod. Fertil. (Suppl. 1):33. . * Hansel W., Alila H.W., Dowd J. P. and Milvae R.A. 1991. Differential mechanisms in small and large bovine luteal cells. J. Reprod. Fertil. Suppl. 43:77. Hansel W., Alila H.W., Dowd J. P. and Yang X. 1987. Control of stroidogenesis in small and large bovine luteal cells. Aust. J. Biol. Sci. 40:331. Hansel W., Concannon P.W. and Lukaszewska J.H. 1973. Corpora lutea of the large domestic animals. Biol. Reprod. 8: 222. Hansel W. and Convey E.M. 1983. Physiology of the estrous cycle. J. Anim. Sci. 57:404. Hansen P.J., Anthony R.V., Bazer F.W., Bauchman G.A. and Roberts R.M. 1985. In vitro synthesis and secretion of ovine trophoblast protein-1 during the period of maternal recognition of pregnancy in the ewe. Endocrinology 117:1424.

PAGE 228

221 Hansen T.R., Kazemi M., Keisler D.H., Malathy P.V., Imakawa K. and Roberts R.M. 1989. Complex binding of the embryonic interferon, ovine trophoblast protein-1, to endometrial receptors. J. Interferon Res. 9:215. Hansen T.R., Leaman D.W., Cross J.C., Mathialagan N., Bixby J.A. and Roberts R.M. 1991. The genes for the trophoblast interferons and the related interferon-a,, possess distinct 5'-promoter and 3'-flanking sequences. J. Biol. Chem. 266:3060. Hanzen S.L., Ford D.A. and Gross R.W. 1991. Activation of a membraneassociated phospholipase Aj during rabbit mycardial ischemia which is highly selective for plasmalogen substrate. J. Biol. Chem. 266:5629. Hanzen S.L. and Gross R.W. 1992. Identification and characterization of human myocardial phospholipase Aj from transplant recipients suffering from end-stage ischemic heart disease. Circ. Res. 70:486. Hara S., Kudo I., Chang H.W., Matsuta K., Miyamoto T. and Inoue K. 1989. Purification and characterization of extracellular phospholipase A2 from human synovial fluid in rheumatoid arthritis. J. Biochem. (Tokyo) 105:395. Harden T.K. 1990. G protein-dependent regulation of phospholipase C by cell surface receptors. Am. Rev. Respir. Dis. 141:8119. Harrison F.A., Heap R.B. and Linzell J.L. 1968. Ovarian function of the sheep after autotransplantation of the ovary and uterus to the neck. J. Endocrinol. 40:xiii (Abst). Harrowing P.D. and Williams K.I. 1977. Homogenates of rat placenta contain a factor(s) which inhibits arachidonic acid metabolism. Br. J. Pharm. 677:482P. Hass T.A., Bastida E., Nakamura K., Hullin F., Admirall L. and Buchanan M.R. 1988. Binding of 13-HODE and 512-, 15 HETE to endothelial cells and subsequent platelet, neutrophil and tumor adhesion. Biochim. Biophys. Acta 961:153. Hawkins D.E., Belfiore C.J., Kile J. P. and Niswender G.D. 1993. Regulation of ribonucleic acid encoding for 3p-hydroxysteroid dehydrogenase / A^-A"^ isomerase in the ovine corpus luteum. Biol. Reprod. 48:1 185. Hawthorn J.N. 1982. Inositol phospholipids. In: Hawthorn J.N. and Ansell G.B. (Eds) Phospholipids, p 263. Elsevier, Amsterdam.

PAGE 229

222 Heap R.B., Fleet I.R., Davis A.J., Goode J.A., Hamon M.H., Walters M.H. and flint A.F.P. 1989. Neurotransmitters and lymphatic vascular transfer of prostaglandin stimulate oxytocin output in sheep. J. Endocrinol. 122:147. Heinsohn C, Polgar P., Fishman J. and Taylor L. 1987. The effect of bovine serum albumin on the synthesis of prostaglandin and incorporation of pH]-acetate into platelet-activating factor. Arch. Biochem. Biophys. 1987 257:251. Helmer S.D., Gross T.S., Hansen P.J. and Thatcher W.W. 1989a. Bovine trophoblast protein-1 complex alters endometrial protein and prostaglandin synthesis and induces an intracellular inhibitor of prostaglandin synthesis in vitro. J. Reprod. Fertil. 87:421. Helmer S.D., Hansen P.J., Anthony R.V., Thatcher W.W., Bazer F.W. and Roberts R.M. 1987. Identification of bovine trophoblast protein-1, a secretory protein immunologically related to ovine trophoblast protein-1 . J. Reprod. Fertil. 79:83. Helmer S.D., Hansen P.J. and Thatcher W.W. 1988. Differential glycolysation of the components of the bovine trophoblast protein-1 complex. Mol. Cell. Endocrinol. 58:103. Helmer S.D. Hansen P.J., Thatcher W.W., Johnson J.W. and Bazer F.W. 1989b. Intrauterine infusion of highly enriched bovine trophoblast protein-1 complex exerts an antiluteolytic effect to extend corpus luteum lifespan in cyclic cattle. J. Reprod. Fertil. 87:89. Hemler M.E and Lands W.E. 1980. Evidence for a peroxide-initiated free radical mechanism of prostaglandin biosynthesis. J. Biol. Chem. 255:6253. Hemler M.E., Cook H.W. and Lands W.E. 1979. Prostaglandin biosynthesis can be triggered by lipid peroxides. Arch. Biochem. Biophys. 193:340. Hemler M.E., Crawford C.G. and Lands W.E. 1978. Lipoxygenation activity of purified prostaglandin-forming cyclooxygenase. Biochemistry 17:1772. Hemler M.E., Lands W.E. and Smith W.L., 1976. Purification of the cyclooxygenase that forms prostaglandins. Demonstration of two forms of iron in the holoenzyme. J. Biol. Chem. 251:5575. Hernandez-Ledezma J.J., Sikes J.D., Murphy C.N., Watson A.J., Schultz G.A. and Roberts R.M. 1992. Expression of bovine trophoblast interferon in conceptuses derived by in vitro techniques. Biol. Reprod. 47:374.

PAGE 230

223 Heyman Y., Camous S., Fevre J., Meziou W. and Martal J. 1984. Maintenance of the corpus luteum after uterine transfer of trophoblastic vesicles to cyclic cows and ewes. J. Reprod. Fertil. 70:533. Hirst J.J., Rice G.E., Jenkins G. and Thorburn G.D. 1988. Secretion of oxytocin and progestrone by ovine corpora lutea in vitro. Biol. Reprod. 35:1 106. Hixon J.E. and Flint A.P.F. 1987. Effects of a luteolytic dose of estradiol benzoate on uterine oxytocin receptor concentrations, phosphoinositides turnover and prostaglandin Fj^ secretion in sheep. J. Reprod. Fertil. 79:457. Hixon J.E. and Hansel W. 1974. Evidence for the preferential transfer of RGFj^ to the ovarian artery following intra-uterine administration to cattle. Biol. Reprod. 11:543. Hixon J.E. and Hansel W. 1979. Effect of PGFj^, estradiol and luteinizing hormone in dispersed cell preparations of bovine corpora lutea. In: Channing CP. and Marsh J.M. (Eds) Ovarian follicular and corpus luteum function, p 61 3. Plenum, New York. Hoffman D.R., Favour S., Uauy R., Rosenfeld C.R. and Magness R.R. 1994. Distribution of unsaturated fatty acids in phospholipids of arteries from nonpregnant, pregnant and fetal sheep. Prostaglandins Leukot. Essent. Fatty Ac. (in press). Hokin M.R. and Hokin L.E. 1953. Enzyme secretion and the incorporation of ^^P into phospholipids of pancreas slices. J. Biol. Chem. 203:967. Horiguchi J., Spriggs D., Imamura K., Stone R., Leubbers R. and Kufe D. 1989. Role of arachidonic acid metabolism in transcriptional induction of tumor necrosis factor gene expression by phorbol ester. Mol. Cell Biol. 9:252. Houslay M.D., Bojanic D. and Wilson A. 1986. PAF and U44069 stimulate a GTPase activity in human platelets which is distinct from the guanine nucleatide regulatory proteins, and Nj. Biochem. J. 234:737. Hovanessian A.G. 1991. Interferon-induced and double-stranded RNA-activated enzymes: a specific protein kinase and 2',5'-oligoadenylate synthetases. J. Interferon Res. 11:199. Hoyer P.B. and Niswender G.D. 1985. The regulation of steroidogenesis is different in the two types of ovine luteal cells. Can. J. Physiol. Pharmacol. 63:240.

PAGE 231

224 Hughes, T.L., Villa-Godoy A., Kesner J.S. and Fogwell R.L. 1987. Destruction of bovine ovarian follicles: effects on the pulsatile release of luteinizing hormone and prostaglandin induced luteal regression. Biol. Reprod. 36: 523. Humblot P. and Dalla Porta M.A. 1984. Effects of conceptus removal and intrauterine administration of conceptus tissue on luteal function in the cow. Reprod. Nutr. Dev. 24:529. Hunter T. 1988. The Ca^Vphospholipid-binding proteins of the submembranous skeleton. Adv. Exp. Med. Biol. 234:169. Hunter T. 1991. Protein kinase classification. Meth. Enzymol. 200:3. Huslig R.L., Fogwell R.L. and Smith W.L. 1979. The prostaglandin forming cyclooxygenase of ovine uterus: relationship to luteal function. Biol. Reprod. 21 :589. Imakawa K., Anthony R.V., Kazemi M., Marotti K.R., Polites H.G. and Roberts R.M. 1987. Interferon-like sequence of ovine trophoblast protein secreted by embryonic trophectoderm. Nature 330:377. Imakawa K., HansenT.R., Malathy P.-V., Anthony R.V., Polites H.G., Marotti K.R. and Roberts R.M. 1989. Molecular cloning and characterization of complementary DNA corresponding to bovine trophoblast protein-1: A comparaison with ovine trophoblast protein-1 and bovine interferon-all. Mol. Endocrinol. 3:127. Imakawa K., Helmer S.D., Nephew K.P., Meka C.R.S. and Christenson R.K. 1993. A novel role for GM-CSF: enhancement of pregnancy specific interferon production, ovine trophoblast protein-1. Endocrinology 132:1869. Inskeep E.K. and Murdoch W.J. 1980. Relation of ovarian functions to uterine and ovarian secretion of prostaglandins during the estrous cycle and early pregnancy in the ewe and the cow. In: Greep R.O. (Ed). Reproductive physiology III: International review of physiology. Vol. 22, p 325. University Park Press, Baltimore. Ireland J.J. and Roche J.F. 1982. Development of antral follicle in cattle after prostaglandin-induced luteolysis: changes in serum hormones, steroids in follicular fluid and gondotrophin receptors. Endocrinology 1 1 1 :2077.

PAGE 232

225 Ireland J.J. and Roche J.F. 1983. Development of nonovulatory antral follicles in heifers: changes in steroids in follicular fluid and receptors for gondotrophins. Endocrinology 112:150. Irvine R.F. 1982. How is the level of free arachidonic acid controlled in mammalian cells? Biochem. J. 204:3. Hishihara 0., Kinoshita K., Satoh K., Mizuno M. and Shimizu T. 1990. An inhibitor of prostaglandin biosynthesis from human decidua: partial purification and properties. Prostaglandins Leukot. Essent. Fatty Ac. 40:223. Ivell R., Hunt N., Abend N., Brackman B., Noumeyer D., Lasma J.C. and McCracken J.A. 1990. Structure and ovarian expression of the oxytocin gene in sheep. Reprod. Fertll. Dev. 2:703. Jacobsen K.L. and Rockwood G.A. 1988. Interferons. I. Their origin and actions. J. Vet. Internal Med. 2:47. Jarabak J., Luncsford A. and Berkowitz D. 1983. Substrate specificity of three prostaglandin dehydrogenases. Prostaglandins 12:915. Jenkin G., Hooper S. and Brouwer D. 1991. Interaction between oxytocin and prostaglandin Fj^ at luteolysis in sheep. Proc. Austr. Soc. Reprod. Biol. 23:24. Jenner L.J., Parkinson T.J. and Lamming G.E. 1991. Uterine oxytocin receptors in cyclic and pregnant cows. J. Reprod. Fertil. 91:49. Johnson F.A. and Poyser N.L. 1991. Effect of melittin on prostaglandin production by guinea-pig uterus. J. Reprod. Fertil. 93:577. Jones D.A., Carlton D.P., Mclntyre T.M., Zimmermann G.A. and Prescott S.M. 1993. Molecular cloning of human prostaglandin endoperoxide synthase type II and demonstration of expression in response to cytokines. J. Biol. Chem. 268:9049. Jones D.S.C. and Flint A.P.F. 1988. Concentrations of oxytocin-neurophysin prohormone mRNA in corpora lutea of sheep during the oestrous cycle and early pregnancy in heifers. J. Endocrinol. 117:409. Julian J.A. Carson D.D. and Glasser S.R. 1992a. Polarized rat uterine epithelium in vitro: responses in defined medium. Endocrinology 1 30:68.

PAGE 233

226 Julian J.A. Carson D.D. and Glasser S.R. 1992b. Polarized rat uterine epithelium in vitro: constitutive expression of estrogen "induced" proteins. Endocrinology 130:79. Kaduce T.L., Figard P.H., Leifur R. and Specter A. A. 1989. Formation of 9hydroxyoctodecadienoic acid from linoleic acid in endothelial cells. J. Biol. Chem. 264:6823. Kastner P., Krust A., Turcotte B., Stropp U., Torra L., Gronemeyer H. and Chambon P. 1990. The distinct estrogen-regulated promoters generate transcripts encoding the two functionally different human progesterone receptor forms A and B. EMBO J. 9:1603. Kawazaki N. and Saito K. 1973. Purification and some properties of lysophospholipase from Penicillium notatum. Biochim. Biophys. Acta 296:426. Kazemi M., Malathy P.V., Keisler D.H. and Roberts R.M. 1988. Ovine trophoblast protein-1 and bovine trophoblast protein-1 are present as specific components of uterine flushings of pregnant ewes and cows. Biol. Reprod. 39:457. Kennedy I., Coleman R.A., Humphrey P.P. and Lumley P. 1983. Studies on the characterization of prostanoid receptors. Adv. Prostaglandin Thromboxane Leukot. Res. 11:327. Kennedy T.G. 1985. Evidence for the involvement of prostaglandins throughout the decidual cell reaction in the rat. Biol. Reprod. 33:140. Kessler D.S. and Levy D.E. 1991. Protein kinase required for for an early step in interferon-a signaling. J. Biol. Chem. 266:23471. Kessler D.S., Veals S.A., Fu X.-Y., and Levy D.E. 1990. IFN-alpha regulates nuclear translocation and DNA-binding affinity of ISGF3, a multimeric transcriptional activator. Genes Dev. 4:1753. Kim B. and Warnaka P. 1989. Indomethacin-enhanced immunotherapy of pulmonary metastases using IL-2 and IFN-alpha. Surgery 106:248. Kimball F.A. and Lauderdale J.W. 1975. Prostaglandin E, and Fj^ specific binding in bovine corpora lutea:comparaison with luteolytic effects. Prostaglandins 10:313. Kimball F.A. and Wyngarden L.J. 1977. Prostaglandin Fj^ binding in equine corpora lutea. Prostaglandins 13:553.

PAGE 234

227 Kindahl H. 1980. Prostaglandin biosynthesis and metabolism. J. Am. Vet. Med. Assoc. 176:1173. Kindahl H., Edqvist L.E., Bane A. and Granstrom E. 1976. Blood levels of progesterone and 15-keto-13,14-dihydro-prostag!andin Fj^ during the normal estrous cycle and early pregnancy in heifers. Acta Endocrinol. (Copenh.) 82:134. Kiracofe G.H., Menzies C.S., Gier T.H. and Spies H.G. 1966. Effect of uterine extracts and uterine or ovarian blood vessels ligations on ovarian function in ewes. J. Anim. Sci. 25:1 159. Kiracofe G.H., Spies H.G. and Gier T.H. 1973. Effect of uterine extracts and ligations in hysterectomized ewes. J. Anim. Sci. 35:862. Kittok R.J. and Britt J.H. 1977. Corpus luteum function in ewes given estradiol during the estrous cycle or early pregnancy. J. Anim. Sci. 45:336. Klee C.B. 1988. Ca2*-dependent phospholipids(and membrane-) binding proteins. Biochemistry 27:6645. Klein G., Langegger M., TimpI R. and Ekblom P. 1988. Role of laminin A chain in the development of epithelial cell polarity. Cell 55:331 . Klemann S.W., Imakawa K. and Roberts R.M. 1989. Sequence variability among ovine trophoblast interferon mRNA. Nucleic Acid Res. 18:6724. Klemann S.W., Li J., Imakawa K., Cross J.C., Francis H.G. and Roberts R.M. 1990. The production, purification, and bioactivity of recombinant trophoblast protein-1 (bovine trophoblast interferon). Mol. Endocrinol. 4:1508. Knapp H.R., Oelz 0., RobertsLJ., Sweetman B.J., Gates J.A., and Reed P.W. 1977. lonophores stimulate prostaglandin and thromboxane biosynthesis. Proc. Natl. Acad. Sci. 74:4251. Knickerbocker J.J. and Niswender G.D. 1989. Characterization of endometrial receptors for ovine trophoblast protein-1 during the estrous cycle and early prgnancy in sheep. Biol. Reprod. 40:361. Knickerbocker J.J., Thatcher W.W., Bazer F.W., Barron D.H. and Roberts R.M. 1986a. Inhibition of uterine PGFj^ production by bovine conceptus secretory protein. Prostaglandins 31:777.

PAGE 235

228 Knickerbocker J.J., Thatcher W.W., Bazer F.W., Drost M., Barron D.H., Fincher K.B. and Roberts R.M. 1986b. Proteins secreted by day 16 to 18 bovine conceptuses extends corpus luteum function in cows. J. Reprod. Fertil. 77:381. ' ' . Ko Y., Lee C.Y., Ott T.L., Davis M.A., Simmen R.C., Bazer F.W., and Simmen F.A. 1991. Insulin-like grovvth factors in sheep uterine fluids: concentrations and relationship to ovine trophoblast protein-1 production during early pregnancy. Biol. Reprod. 45:135. Koch C.A., Anderson D., Moran M.F., Ellis C, and Pawson T. 1991. SH2 and SH3 domains: elements that cobtrol interactions of cytoplasmic signaling proteins. Science 252:668. Koehler L., Hass R., DeWitt D.L., Resch K. and Goppelt-Struebe M. 1990. Glucocorticoid-induced reduction of prostanoid synthesis in TPAdifferentiated U937 cells is mainly due to a reduced cyclooxygenase activity. Biochem. Pharmacol. 40:1307. Korff J.M. and Jarabak J. 1982. The purification and characterization of a prostacyclin dehydrogenase from rabbit kidney. J. Biol. Chem. 257:2177. Kotwica J., Skarzynski D., Jaroszewski J. and Kotwica G. 1991. Effects of PGFja analogue on ovarian oxytocin and progestrone release in heifers: cathecolamines are not involved in this process. Anim. Reprod. Sci. 25:1. Kramer R.M., Hession C, Johansen B., Hayes G., McGray P., Chow E.P., Tizard R. and Pepinsky R.B. 1989. Structure and properties of human nonpancreatic phospholipase A2. J. Biol. Chem. 264:5768. Krause H., Dieter P., Schulze-Specking A., Ballhorn A. and Decker K. 1991. Calcium-induced reversible translocation of phospholipase A2 between the cytosol and the membrane fraction of rat liver macrophages. Eur. J. Biochem. 199:355. . , Kristensen T., Saris C.J.M., Hunter T., Hicks L.J., Nooman D.J., Glenney J.R. Jr. and Tack B.F. 1986. Primary structure of bovine calpactine I heavy chain (p36), a major cellular substrate for retroviral protein-tyrosine kinase: homology with the human PLAj inhibitor lipocortin. Biochemistry 25:4497. Krohn P.L. 1966. Factors influencing the number of oocytes in the ovary. Arch. Anat. Microsc. 56:151.

PAGE 236

229 Kroner E.E., Peskar B.A., Fisher H. and Ferber E. 1981. Control of arachidonic acid accumulation in bone marrow-derived macrophages by acyltransferases. J. Biol. Chem. 256:3690. Kujubu D.A., Fletcher B.S., Varnum B.C., Lim R.W. and Herschman H.R. 1991. TIS10, a phorbol ester tumor promoter-inducible mRNA from Swiss 3T3 cells, encodes a novel prostaglandin synthase/cyclooxygenase homologue. J. Biol. Chem. 266: 12866. Kulmacz R.J. and Lands W.E. 1983. Requirements for hydroperoxide by the cyclooxygenase and peroxidase activities of prostaglandin H synthase. Prostaglandins 25:531 . Kyes P. 1902. Ueber die Wirkungsweise des Cobragiftes, I, II. Berlin Klin. Wochenschr. 39:889-890, 918. Labarca C. and Paigen K. 1980. A simple and sensitive DNA assay procedure. Anal. Biochem. 104:344. Lacroix M.-C. and Kann G. 1986. Aspects of the antiluteolytic activity of the conceptus during early pregnancy in ewes. J. Anim. Sci. 63:1449. Lafrance M. and Goff A.K. 1985. Effect of pregnancy on oxytocin-induced release of prostaglandin Fj^ in heifers. Biol. Reprod. 33:1 1 13. Lafrance M. and Goff A.K. 1988. Effects of progesterone and estradiol -17p on oxytocin induced release of PGFj^ in heifers. J. Reprod. Fertil. 82:429. Lafrance M. and Goff A.K. 1990. Control of bovine uterine PGF2„ release in vitro. Biol. Reprod. 42:288. Lamberth F.L. and Oates J.A. 1976. Plasma lipid pattern in chronic renal failure. Trans. Am. Soc. Artif. Int. Organs 21:473. Lamming G.E., Vallet J.L. and Flint A.P.F. 1991. Progestational control of endometrial oxytocin receptor determines cycle length in sheep. J. Reprod. Fertil. Suppl. 43:53. * Lamsa J.C., Cushman R.A.. Nay M.G. and McCracken J.A. 1992. In vivo desensitization of a high affinity PGFj^ receptor in the ovine corpus luteum. Prostaglandins 43:165. Lamsa J.C., Kot S.J., Elderling J.A., Nay M.G. and McCracken J.A. 1989. ' Prostaglandin F2„-stimulated release of ovarian oxytocin from ovine

PAGE 237

230 uterus during luteolysis and its abrogation during early pregnancy. Biol. Reprod. 40:1215. Lands W.E. and Crawford C.G. 1976. Enzymes of membrane phospholipid metabolism in animals. In: Martonosi A. (Ed) The Enzymes of Biological membranes. Vol. 2, p 3. Plenum, New-York. Lands W.E., LeTellier P.R., Rome L. H. and Vanderhoek J.Y. 1972. Inhibition of prostaglandin synthesis. Adv. Biosci. 9:15. Lands W.E. and Samuelsson B. 1968. Phospholipid precursors of prostaglandins. Biochim. Biophys. Acta. 164:426. Lands W.E., Sauter J. and Stone G.W. 1978. Oxygen requirement for prostaglandin biosynthesis. Prostaglandins Med. 1:117. Langer J.A. and Pestka S. 1988. Interferon receptors. Immunol. Today 9:393. Leach-Harper CM. and Thornburn G.D. 1984. Inhibition of prostaglandin synthesis by ovine allantoic fluid: acute reduction in inhibitory activity during late gestation. Can. J. Physiol. Pharm. 62:1152. Leaman D.W., Cross J.C. and Roberts R.M. 1992. Genes for the trophoblast interferons in sheep, goat and musk ox and their distribution among mammals. J. Interferon Res. 12:1. Leaver H.A. and Poyser N.L. 1981. Distribution of arachidonic acid and other fatty acids in the lipids of guinea-pig uterus and plasma in relation to uterine prostaglandin synthesis. J. Reprod. Fertil. 61:325. Leckie CM. and Poyser N.L. 1993. Effects of guinea-pig endometrial proteins on PLAj activity and prostaglandin synthesis. Prostaglandins Leukot. Essent. Fatty Ac. 48:219. Lemon M. and Mauleon. P. 1982. Interaction between two luteal cell types from the corpus luteum of the cow in progesterone synthesis in vitro. J. Reprod. Fertil. 64:315. Leslie CC 1991. Kinetic properties of a high molecular mass arachidonoylhydrolysing phospholipase Aj that exhibits lysophosphoiipase activity. J. Biol. Chem. 266:11366. Lester D. 1990. In vitro linoleic acid activation of protein kinase C Biochim. Biophys. Acta 1954:297.

PAGE 238

231 Levy D.E. and Darnell Jr. J.E. 1990. Interferon dependent transcriptional activation: signal transduction without a second messenger involvement? New Biol. 2:923. Lewis G.S., Jenkins P.E., Fogwell R.L. and Inskeep E.K. 1978. Concentrations of prostaglandins Ej and and their relationship to luteal function in early pregnant ewes. J. Animal Sci. 47:1314. Lifsey B.J., Baumbach G.A. and Godkin J.D. 1989. Isolation, characterization and immunocytochemical localization of bovine trophoblast protein-1. Biol. Reprod. 40:343. Lin A.H., Bienkowski M.J. and Gorman R.R. 1989. Regulation of prostaglandin H synthetase mRNA levels and prostaglandin biosynthesis by plateletderived growth factor. J. Biol. Chem. 264:17379. Litosch S., Wallis C. and Fain J.N. 1985. 5-Hydroxytryptamine stimulates inositol phosphate production in a cell free system from blowfly salivary glands: evidence for a role of GTP in coupling receptor activation to phosphoinositide breakdown. J. Biol. Chem. 260:5464. Loeb L. 1923. The effect of extirpation of the uterus on the life and function of the corpus luteum of the sexual cycle of the ginea pig. Proc. Soc. Exp. Biol. Med. 20:441. Loeb L. 1927. The effects of hysterectonmy on the system of sex organs and on the periodicity of the sexual cycle in the guinea pig. Am. J. Physiol. 83:202. Lohstroh A.J. and Johnson R.E. 1966. Amounts of interstitial cell-stimulating hormone and follicle-stimulating hormone required for folliclular development, uterine growth and ovulation in the hypophysectomized rat. Endocrinology 79:991. Lukaszewska J. and Hansel W. 1 980. Corpus luteum maintenance during early pregnancy in the cow. J. Reprod. Fertil. 59:485. Lynch-Salamon D.I., Everson W.V. and Myatt L. 1992. Decrease in annexin I messenger RNA expression in human amnion with labor. Am. J. Obstet. Gynecol. 167:1657. . Lyons-Giordano B., Pratta M.A., Galbraith W., Davis G.L. and Arner E.C. 1993. lnterleukin-1 differentially modulates chondrocyte expression of cyclooxygenase-2 and phospholipase A2. Exp. Cell Res. 206:58.

PAGE 239

232 Manduteanu I., Popov D., Radu A. and Simionescu M. 1988. Calf cardiac valvular endothelial cells in culture: production of glycosaminoglycans, prostacyclin and fibronectin. J. Mol. Cell Cardiol. 20:103. Mapletoft R.J. and Ginther O.J. 1975. Adequacy of main uterine vein and the ovarian artery in the local veinoarterial pathway for uterine-induced lutolysis in ewes. Am. J. Vet. Res. 36:957. Margolis B., Li N., Koch A., Mohammadi M., Hurwitz D.R., Zilberstein A., Ulrich A., Pawson T. and Schlessinger J. 1990. The tyrosine phosphorylated carboxyterminus of the EGF receptor is a binding site For GAP and PLC y. EMBO J. 9:4375. Markey CM., Alward A., Weller P.E. and Marnett L.J. 1987. Quantitative studies of hydroperoxide reduction by PGH synthase. J. Biol. Chem. 262:6266. Marshall L.A. and McCarte-Roshak A. 1992. Demonstration of similar calcium dependency by mammalian high an low molecular mass PLA2 Biochem. Pharmacol. 44:1849. Marshall P.J. and Kulmacz R.J. 1988. Prostaglandin H synthase: distinct binding sites for cyclooxygenase and peroxidase substrates. Arch. Biochem. Biophys. 266:162. Marshall P.J., Kulmacz R.J. and Lands W.E. 1987. Constraints on prostaglandin synthesis in tissues. J. Biol. Chem. 262:3510. Martal J., Camous S., Fevre J., Charlier M. and Heyman Y. 1984. Specificity of embryonic signals maintaining corpus luteum in early pregnancy in ruminants. Proc. ^0^^ Int. Cong. Anim. Reprod. Al, Urbana 3:510. Martal J., Degryse E., Charpigny G., Assal N., Reinaud P., Charlier M., Gaye P. and Lecocq J. P. 1990. Evidence for extended maintenance of the corpus luteum by intrauterine infusion of recombinant trophoblast ainterferon (trophoblastin) in sheep. J. Endocrinol. 127:R5. Martal J., Lacroix M.-C, Loudes C, Saunier M. and Wintenberger-Torres S. 1979. Trophoblastin, an anti luteolytic protein present in early pregnancy in sheep. J. Reprod. Fertil. 56:63. Martinod S., Maurer R.R., Siegenthaler B., Gerber C, Hansen P.J. 1991. The effect of recombinant bovine interferon-a on fertility in ewes. Theriogenology 36:231 .

PAGE 240

233 Mayer R.J. and Marshall L.A. 1993. New insights on mammalian phospholipase AjCs); comparaison af arachidonoyl-selective and -nonselective enzymes. FASEB 7(2):339. McArdle M.R. and Holtorf A.-P. 1989. Oxytocin and progestrone release from bovine corpus luteal cells in culture: effects of IGF-I, insulin and prostaglandins. Endocrinology 78:21 1. McCracken J.A. 1971. Prostaglandin Fj^ and corpus luteum regression. Ann. N.Y.Acad. Sci. 180:456. McCracken J.A., Glew M.E. and Scramuzzi R.J. 1970. Corpus luteum regression induced by POFj^. J. Endocrinol. Met. 30:544. McCracken J.A., Schramm W., Barcikowski B. and Wilson L. 1981. The identification of PGFj^ as a uterine luteolytic hormone and the hormonal control of its synthesis. Acta Vet. Scand. Suppl. 77:71. McCracken J.A., Schramm W. and Okulicz W.C. 1984. Hormone receptor control of pulsatile secretion of PGFj^ from the ovine uterus during luteolysis and its abrogation in early pregnancy. Anim. Reprod. Sci. 7:31. McCracken J.A., Smith T.T., Lamsa J.C. and Robinson A.G. 1991. The effect of etradiol-17p and progesterone on the oxytocin pulse generator in the ovariectomized sheep. Biol. Reprod. 44 (Suppl. 1):87 (abst). McNatty K.P. 1979. Follicular determinants of corpus luteum function in the human ovary . Adv. Exp. Med. Biol. 112:465. McNatty K.P., Heath D.A., Hudson N. and Clarke I.J. 1990. Effect of long term hypophysectomy on ovarian follicle populations and gonadotrophininduced cAMP output by follicles from Booroola ewes with or without the F gene. J. Reprod. Fertil. 90:515. Mead E.A., Smith W.L. and DeWitt D.L. 1993. Differential inhibition of prostaglandin peroxide synthase (cyciooxygenase) isozymes by aspirin and other non-steroidal anti-inflammatory drugs. J. Biol. Chem. 268: 6610. Mead J.F., Alfin-Slater R.B.. Howton D.R. and Popjak G. (Eds) 1986. Phosphoglyceride metabolism. In: Lipids chemistry, biochemistry and nutrition, p 405. Plenum, New York. Merrill A.H. 1989. Lipid modulators of cell function. Nutr. Rev. 47:161.

PAGE 241

234 Meyer H.H.D., Mittermeier T. and Schams D. 1986. Dynamics of oxytocin, estrogen and progestin receptors in the bovine endometrium during the estrous cycle. Acta Endocrinol. 118:96. Meyer H.H.D., Mittermeier T. and Schams D. 1988. Dynamics of oxytocin, estrogen and progestin receptors in the bovine endometrium during the estrous cycle. Acta Endocrinol. 118:96. Meyer M.D., Drost M., Ott T.L., Bazer F.W., Badinga L., Hansen P.J. and Thatcher W.W. 1992. Recombinant ovine trophoblast protein-1 (roTP-1) extended corpus luteum lifespan and reduced uterine secretion of PGF2„ on cattle. J. Anim. Sci. 70 (Suppl. 1 ): 270 (abst). Milgrom E., Dessen P., Zerah V., Meriel C, Atger M., Loosfelt H. and Mirashi M. 1988. Organization of the entire rabbit progesterone receptor mRNA and the promoter and 5' flanking region of the gene. Nucleic Acid Res. 16:5459. Milvae R.A. and Hansel W. 1980. Concurrent uterine venous and ovarian arterial prostaglandin F concentrations in heifers treated with oxytocin. J. Reprod. Fertil. 60:7. Milvae R.A. and Hansel W. 1983. Luteolytic effect of 13,14-dihydro-PGF2„ in heifers. J. Reprod. Fertil. 67:203. Mirando M.A., Bazer F.W., Thatcher W.W. and Hansen P.J. 1990b. Endometrial inositol phosphate (IP) turnover is stimulated by oxytocin (OT) and reduced during pregnancy. J. Anim. Sci. 68 (Suppl. 1):419 (abst). Mirando M.A., Becker W.C. and Whiteaker S.S. 1993a. Relashionships among endometrial oxytocin receptors, oxytocin-stimulated phosphoinositide hydrolysis and PGFj^ secretion in vitro, and plasma concentrations of ovarian steroids during corpus luteum regression in cyclic heifers. Biol. Reprod. 48:874. Mirando M.A., Harney J.P., Zhou Y., Ogle T.F., Ott T.L., Moffatt R.J. and Bazer F.W. 1993b. Changes in progesterone and oestrogen receptor mRNA and protein and oxytocin receptors in endometrium of ewes after intrauterine injection of ovine trophoblast interferon. J. Mol. Endocrinol. 10:185. Mirando M.A., Ott T.L., Vallet J.L., Davis M. and Bazer F.W. 1990a. Oxytocinstimulated inositol phosphate turnover in endometrium is influenced by stage of estrous cycle, pregnancy and intrauterine infusion of conceptus secretory proteins. Biol. Reprod. 42:98.

PAGE 242

235 Mirando MA, Short E.G., Geisert R.D., Vallet J.L and Bazer F.W. 1991. Stimulation of 2,5-oligoadenylate synthetase activity in sheep endometrium during pregnancy, by intrauterine infusion of oTP-1 , and by intramuscular administration of recombinant bovine interferon-al1. J. Reprod. Fertil. 93:599. Miyamoto A., Lutzov;/ H.V. and Schams D. 1993. Acute actions of prostaglandin Fjoj, E2 and I2 in microdialyzed bovine corpus luteum in vitro. Biol. Repod. 49:423. Miyamoto T., Ogino N., Yamamoto S. and Hayaishi 0. 1976. Purification of prostaglandin endoperoxide synthetase from bovine vesicular gland microsomes. J. Biol. Chem. 251:2629. Mogensen K.E., Uze G. and Eid P. 1989. The cellular receptor of the alphainterferons. Experientia 45:500. Moonen P., Buytenhek M. and Nugteren D.H. 1982. Purification of PHG-PGE isomerase from sheep vesicular glands. Meth. Enzymol. 86:84. Moor R.M. and Rowson L.E.A. 1966. The corpus luteum of the sheep: effect of the removal of embryos on luteal function. J. Endocrinol. 34:233. Moore L.G., Choy V.J., Elliot R.L. and Watkins W.B. 1986. Evidence for the pulsatile release of PGFj,, inducing the release of ovarian oxytocin during luteolysis in the ewe. J. Reprod. Fertil. 76:159. Nancarrow CD., Buckmaster J., Chamley W., Cox R.I., Gumming I.A., Gummins L, Drinan J. P., Findlay J.K.,. Goding J.R, Restall B.J., Schneider W. and Thorburn G.D. 1973. Hormonal changes around oestrus in the cow. J. Reprod. Fertil. 32: 320. Needleman P., Turk J., Jackschik B.A., Morrison A.R. and Lefkowith J. 1986. Arachidonic acid metabolism. Annu. Rev. Biochem. 55:69. Nephew K.P., McGlure K.E., Day M.L, Xie S., Roberts R.M. and Pope W.F. 1990. Effects of intramuscular administration of recombinant bovine interferon-a|1 during the period of maternal recognition of pregnancy. J. Anim. Sci. 68:2766.

PAGE 243

236 Nephew K.P., McClure K.E., Ott T.L., Bazer F.W. and Pope W.F. 1989. Cumulative recognition of pregnancy by embryonic production of ovine trophoblastic protein-1 . J. Anim. Sci. (Suppl. 1): 404 (abst). Nephew K.P., McClure K.E., Ott T.L., Dubois D.H., Bazer F.W. and Pope W.F. 1991. Relationship between variation in conceptus development and differences in estrous cycle duration in ewes. Biol. Reprod. 44:536. Newton G.R., Martinod S., Hansen P.J., Thatcher W.W., Siegenheiler B., Gerber C. and Voirol M.-J. 1990. Effect of bovine interferon on accute changes in body temperature and serum progesterone concentration in heifers. J. Dairy Sci. 73:3439. Newton G.R., Vallet J.L, Hansen P.J. and Bazer F.W. 1989. Inhibition of lymphocyte proliferation by ovine trophoblast protein-1 and a high molecular weight glycoprotein produced by peri-implantation sheep conceptus. Am. J. Immunol. Microbiol. 19:99. Nishizuka Y. 1988. The molecular heterogeneity of protein kinase C and its implications for cellular regulation. Nature 334:661 . Nishizuka Y. 1992. Intracellular signalling by hydrolysis of phospholipids and activation of protein kinase C. Science 258:607. Niswender G. D. and Nett T.M. 1988. The corpus luteum and its control. In: Knobil E. and Neill J. (Eds) The physiology of reproduction, p 489. Raven Press, New York. Norman S.J. and Poyser N.L. 1993. Regulation of enzymes controlling arachidonic acid trunover in phospholipids of guinea-pig uterus. J. Reprod. Fertil. 11:185 (abst). Northey D.L. and French L.R. 1980. The effect of embryo removal and intrauterine infusion of embryonic homogenates on the lifespan of the bovine corpus luteum. J. Anim. Sci. 50:298. Nugteren D.H. and Van Dorp D.A. 1966. The participation of molecular oxygen in the biosynthesis oof prostaglandins. Biochim. Biophys. Acta 98:654. O'Banion M.K., Winn V.D. and Young D A. 1992. cDNA cloning and functional activity of a glucocorticoid-regulated inflammatory cyclooxygenase. Proc. Natl. Acad. Sci. USA 98:4888. Odenwaller R., O'Gara C.Y. and Marnett L.J. 1991. Properties of trypsin-cleaved PGH synthase. Adv. Prostaglandin Thromboxane Leukot. Res. 21A:81.

PAGE 244

237 Ogino N., Miyamoto T., Yamamoto S. and Hayaishi 0. 1977. Prostaglandin endoperoxide E isomerase from bovine vesicular microsomes, a glutathione-requiring enzyme. J. Biol. Chem. 252:890. O'Neill C. 1985. Examination of the cause of early pregnancy associated thrombopenia. J. Reprod. Fertil. 73:567. O'Neill C. 1990. A physiological role for PAF in the stimulation of mammalian embryonic development. Trends Pharmacol. Sci. 12:82. O'Neill C, Gidley-Baird A. A., Pike I.L. and Saunders D.M. 1987. A bioassay for embryo-derived platelet activating factor as a mean of assessing quality and pregnancy potential of human embryos. Fertil. Steril. 47:969. Ott T.L., Mirando M.A., Davis M.A. and Bazer F.W. 1992. Effects of ovine conceptus secretory proteins and progesterone on oxytocin-stimulated endometrial production of prostaglandin and turnover of inositol phosphate in ovariectomized ewes. J. Reprod. Fertil. 95:19. Ott T.L, Mirando M.A., Davis M.A., Fliss M.F.V. and Bazer F.W. 1989. Characterization of a second period of immunoreactive ovine trophoblast protein-1 secretion in sheep. J. Anim. Sci. 67 (Suppl. 1):370 (abst). Ott T.L., Van Heeke G., Johnson H.M. and Bazer F.W. 1991. Cloning and expression in S. Cerevisiae of a synthteic gene for the type I trophoblast interferon ovine trophoblast protein-1 ; purification and antiviral activity. J. Interferon Res. 11:357. Ott T.L, Zhou Y., Mirando M.A., Stevens C, Harney J. P., Ogle T.F. and Bazer F.W. 1993. Changes in progesterone and oestrogen receptor mRNA and protein during maternal recognition of pregnancy and luteolysis in ewes. J. Mol. Endocrinol. 10:171. Pace-Asciak C. and Wolfe L.S. 1968. Inhibition of prostaglandin synthesis by oleic, linoleicand linolenic acids. Biochim. Biophys. Acta 152:784. Pagano R.E. 1990. Lipid traffic in eukaryotic cells: mechanisms for intracellular transport and organelle-specific enrichment of lipids. Curr. Opin. Cell Biol. 2:652. Parkinson T.J., Lamming G.E., Flint A.F.P. and Jenner L.J. 1992a. Administration of recombinant bovine interferon-alpha I at the time of

PAGE 245

238 maternal recognition of pregnancy inhibits RGFj^ secretion and causes luteal maintenance in cyclic ewes. J. Reprod. Fertil. 94:489. Parkinson T.J., Wathes D.C., Jenner L.J. and Lamming G.E. 1992b. Plasma and luteal concentrations of oxytocin in cyclic and early pregnant cattle. J. Rerod. Fertil. 94:161. Paul J. and Kinsella J.E. 1983. Effects of bradykinin and bovine serum albumin on arachidonic acid and prostaglandin release from perfused rat heart. Prostaglandins Leukot. Med. 11:419. Pepinsky R.B., Tizzard R., Mattaliano R.J., Sinclair L.K., Miller G.T., Browning J.L, Chow E.P., Burne C, Huang K.S., Pratt D., Wachter L., Frey A.Z. and Wallner B.P. 1988. Five distinct calcium and phospholipid binding proteins share homology with lipocortin I. J. Biol. Chem. 263:10799. Pestka S., Langer J.A., Zoon K.C. and Samuel C.E. 1987. Interferons and their actions. Annu. Rev. Biochem. 56:727. Peters H. and Levy E. 1966. Cell dynamics of the ovarian cycle. J. Reprod. Fertil. 11:227. Peterson A.J., Fairclough R.J., Payne E. and Smith J.F. 1975. Hormonal changes around bovine luteolysis. Prostaglandins 10: 675. Pernas P., Masliah J., Olivier J.-L., Salvat C, Rybkine T. and Bereziat G. 1991. Type II phosphollpase A2 recombinant overexpression enhances stimulated arachidonic acid release. Biochem. Biophys. Res. Commun. 178:1298. Pfeffer L.M., StruloviciB. and Saltiel A.R. 1990. Interferon-a selectively activates the 3 isoform of the protein kinase C through phosphatidylcholine hydrolysis. Proc. Natl. Acad. Sci. USA87:6537. Pharris B.B. and Whyngarden L.J. 1969. The effect of prostaglandin F2a on the progesterone content of ovaries from pseudopregnant rats. Proc. Soc Biol. 130:92. Pierik A.J., Nijssen J.G.. Aarsman A.J. and Van den Bosch H. 1988. Calciumindependent phospholipase Aj in rat tissue cytosols. Biochem. Biophys Acta 962:345. Pierson R.A. and Ginther O.J. 1984. Ultrasonography of the bovine ovary. Theriogenology 26:649.

PAGE 246

239 Pierson R.A. and Ginther O.J. 1988. Follicular population during estrus cycle in heifers .III. Time of selection of the ovulatory follicle. Anim. Reprod. Sci. 16:81. Plante C, Hansen P.J., Martinod S., Siegenthaler B., Thatcher W.W. and Leslie M.V. 1989. Effect of intrauterine and intramuscular administration of recombinant bbovine interferon a1 on luteal lifespan in cattle. J. Dairy Sci. 72:1859. Plante C, Hansen P.J. and Thatcher W.W. 1988. Prolongation of luteal lifespan in cows by intrauterine infusion of recombinant bovine a-interferon. J. Endocrinol. 122:2342. Plante C, Hansen P.J., Thatcher W.W., Johnson J.W., Pollard J.W., Mirando M.A. and Bazer F.W. 1990. Purification of bovine protein-1 complex and quantification of its microheterogenous variants as affected by culture conditions. J. Reprod. Immunol. 18:271. Plante C, Thatcher W W. and Hansen P.J. 1991. Alteration of oestrous length, avorian function and oxytocin-induced release of PGFj^ by intrauterine and intramuscular administration of recombinant bovine interferon-a to cows. J. Reprod. Fertil. 93:375. Ponglikitmongkol M., Green S. and Chambon P. 1988. Genomic organization of the human estrogen receptor gene. EMBO J. 7:3385. Pontzer C.H., Bazer F.W. and Johnson H.M. 1991. Antiproliferative activity of a pregnancy recognition hormone, ovine trophoblast protein-1. Cancer Res. 51:5304. Pontzer C.H., Torres B.A., Vallet J.L, Bazer F.W. and Johnson H.M. 1988. Antiviral activity of the pregnancy recognition hormone ovine trophoblast protein-1. Biochem. Biophys. Res. Com. 152:801. Powell W.S., Hammarstrom S. and Samuelsson B. 1974. Prostaglandin Fj,, receptor in ovine corpora lutea. Eur. J. Biochem. 41:103. Poyser N.L. 1987. Effects of various factors on prostaglandin synthesis by the the guinea-pig endometrium. J. Reprod. Fertil. 81:269. Poyser N.L. 1991. A possible explanation for the refractoriness of uterine prostaglandin production. J. Reprod. Fertil. 91:371.

PAGE 247

240 Poyser N.L. and Brydon L.J. 1983. Prostaglandin release from guinea-pig uterus superfused in vitro. Effects of stage of oestrous cycle, progesterone, oestradiol, oxytocin and A23187. Prostaglandins 25:443. Poyser N.L. and Ferguson A. 1993. Is there refractoriness between PLAj and calcium ionophore A23187 in the stimulation of uterine prostaglandin production? Prostaglandins Leukot. Essent. Fatty Ac. 48:327. Poyser and Riley S.C. 1987. Effect of actinomycin D on prostaglandin synthesisby and output from the guinea-pig uterus. Prostaglandin Leukot. Med. 28:153. Prescott S.M. and Majerus P.N. 1983. Characterization of 1 ,2-diacylglycerol hydrolysis in human platelets. Demonstration of an arachidonoylmonoacylglycerol intermediate. J. Biol. Chem. 258:764. Pressman B.C. 1976. Biological applications of ionophores. Annu. Rev. Biochem. 451:501. Pruzanski W. and Vadas P. 1991. Phospholipase Aj-a mediator between proximal and distal effectors of inflammation. Immunol. Today 12:143. Quinn P.J. 1981. Fluidity of cell membranes and its regulation. Prog. Biophys. Mol. Biol. 38:1. Ramesha C.S. and Pickett W.C. 1986. PAF and leukotriene biosynthesis is inhibited in polymorphonuclear leukocytes depleted of arachidonic acid. J. Biol. Chem. 261:7592. Ramwell P.W., Leovey E.M.K. and Sintetos A.L. 1977. Regulation of the arachidonic cascade. Biol. Reprod. 16:70. Rao C.V. 1975. Cationic dependency of high affinity prostaglandin Fj^ receptors in bovine corpora lutea membranes. Biochem. Biophys. Res. Comm. 67:1242. Rao C.V., Griffin L.P. and Carmen F.R. 1977. Prosatglandin F^a binding sites in human corpora lutea. J. Clin. Endocrinol. Metab. 44:1032. Raw R.E., Curry T.E. and Silvia W.J. 1988. Effects of progesterone and estradiol on the concentration and activity of cyclooxygenase in the ovine uterus. Biol. Reprod. 38:104 (abst).

PAGE 248

241 Raw R.E. and Silvia W.J. 1991. Activity of phospholipase C and release of prostaglandin Fj^ by endometrial tissue from ovariectomized ewes receiving progesterone and estradiol. Biol. Reprod. 44:404. Remy E.L., Gerard A., Houben A., Vandesteen C. and Remade J. 1989. Inhibition of PAF biosynthesis via the acetyltransferase by arachidonic and oleic acids in ionophore A23187-stimulated bovine neutrophils. Biochim. Biophys. Acta 1005:87. Rehfeldt W., Hass R. and Goppelt-Struebe M. 1991. Characterization of phospholipase Aj in monocytic cell lines. Functional and biochemical aspaects of membrane association. Biochem. J. 276:631 . Reich N.C. and Pfeffer L.M. 1990. Evidence for the involvement of protein kinase C in the cellular response to interferon a. Proc. Natl. Acad. Sci. USA 87:8761. Rhee S.G., Suh P.-G., Ryu S.-H. and Lee S.Y. 1989. Studies of inositol phospholipid-specific phospholipase C. Science 244:546. Richards J. 1980. Maturation of ovarian follicles: actions and interactions of pituitary and ovahan follicular cell differentiation. Physiol. Rev. 60:51. Richelsen B., Borglum J.D., and Sorensen S.S. 1992. Biosynthetic capacity and regulatory aspects of prostaglandin E2 formation in adipocytes. Mol. Cell. Endocrinol. 85:73. Rico L.W., Thatcher W.W., Drost M., Wolfenson D. and Terqui M. 1981. Plasma PGFM responses to estradiol injection in pregnant and cycling cows. J. Anim. Sci. 53 (Suppl.1):363 (abst). Riley J.C.M., Czirake S.E. and Carlson J.C. 1989. The effects of prolactin and prostaglandin F2„ on plasma membrane changes during luteolysis in the rat. Endocrinology 124:1564. * v Riley S.C. and Poyser N.L. 1987. Prostaglandin production by guinea-pig endometrium: is calcium necessary? J. Endocrinol. 113:463. Riley S.C. and Poyser N.L. 1989. Is protein synthesis necessary for prostaglandin production by guinea-pig endometrium? J. Reprod. Fertil. 86:73. Roberts R.M. 1993. Interferon-tau (letter). Nature 362:583.

PAGE 249

242 Roberts R.M., Cross J.C. and Leaman D.W. 1992. Interferons as hormones of pregnancy. Endocrinol. Rev. 13:432. Roberts R.M., McCracken J. A., Gavagan J.E. and Soloff M.S. 1976. Oxytocinstimulated release of PCFj^ from ovine endometrium in vitro: correlation with estrous cycle and oxytocin-receptor binding. Endocrinology 99:1107. Rodgers R.J., O'Shea J.D. and Bruce N.W. 1984. Morphometric analysis of the cellular composition of the ovine corpus luteum. J. Anat. 138:757. Rose M.I., Deems R.A., Jesaitis A.J., Dennis E.A. and Ulevitch R.J. 1985. Phospholipase activities of the PSSSD^ macrophage-like cell line. Arch. Biochem. Biophys. 238:247. Rosoff P.M., Savage N. and Dinarello C.A. 1988. lnterleukin-1 stimulates diacylglycerol production in T lymphocytes by a novel mechanism. Cell 54:73. Rothchild I. 1981. The regulation of the mammalian corpus luteum. Rec. Prog. Horm. Res. 37:183. Rothschild M.A., Dratz M. and Schneider S.S. 1980. Albumin synthesis. Int. Rev. Physiol. 21:249. Rueda B.R., Naivar K.A., George E.M., Austin K.J., Francis H. and Hansen T.R. 1993. Recombinant interferon-x regulates secretion of two bovine endometrial proteins. J. Interferon Res. 13:303. Saeed S.A, McDonald-Gibson W.J., Cuthbert J., Copas J.L., Schneider C, Gardiner P.J., Butt N.M. and Collier H.O.J. 1977. Endogenous inhibitor of prostaglandin synthetase. Nature 270:32. Saeed S.A., Strickland D.M., Young D.C., Dang A. and Mitchell M.D. 1982. Inhibition of prostaglandin synthesis by amniotic fluid: acute reduction in inhibitory activity of amniotic fluid obtained during labor. J. Clin. Endocrinol. Metab. 55:801. Sakata T., Nakamura E., Tsuruta Y., Tamaki M., Teraoka H., Tojo H., Ono T. and Okamoto M. 1989. Presence of pancreatic-type phospholipase Aj mRNA in rat gastric mucosa and lung. Biochem. Biophys. Acta 1007:124.

PAGE 250

243 Salamonsen B. and Findlay J.K. 1990. Immunohistochemical localization of prostaglandin synthase in the ovine uterus during the oestrous cycle and in early pregnancy. Reprod. Fertil. Dev. 2:31 1 . Salamonsen B., Hampton A.L., Clements J.A. and Findlay J.K. 1991. Regulation of gene expression and cellular localization of prostaglandin synthase by estrogen and progesterone in ovine uterus. J. Reprod. Fertil. 92:393. Salamonsen B., Stuchbery S.J., O'Grady CM., Godkin J.D. and Findlay J.K. 1988. interferon-a mimics the effects of ovine trophoblast protein-1 on prostaglandin and protein secretion by ovine endometrial cells in vitro. J. Endocrinol. 117:R1. Salzberg S., Hacohen D., David S., Dovrat S., Ahwan S., Gamliel H. and Birnbaum M. 1990. Involvement of interferon-system in the regulation of cell growth and differentiation. Scanning Microsc. 84:380. Samuelsson B. 1978. Prostaglandins and thromboxanes. Rec. Prog. Horm. Res. 34:239. SAS 1982. SAS user's guide. SAS Institute Inc., Gary, North Carolina. Savio J.D., Boland M.P. and Roche J.F. 1990. Development of dominant follicles and length of ovarian cycles in post-partum dairy cows. J. Reprod. Fertil. 88:581. Savio J.D., Keeman L., Boland M.P. and Roche J.F. 1988. Pattern of growth of dominant follicles during the oestrous cycle of heifers. J. Reprod. Fertil. 83:663. Sawada M. and Carlson J.C. 1991. Rapid plasma membrane changes in superoxide radical formation, fluidity, and phospholipase A2 activity in the corpus luteum of the rat during induction of luteolysis. Endocrinology 128:2992. , . Saxena U., Klein M.G. and Goldberg I.J. 1991. Transport of lipoprotein lipase across endothelial cells. Proc. Natl. Acad. Sci. USA 88:2254. Schallenberger E., Schams D. and Meyer, H.H.D. 1989. Sequences of pituitary, ovarian and uterine hormone secretion during the first 5 weeks of pregnancy in dairy cattle. J. Reprod. Fertil. Suppl. 37:277. Schallenberger E., Schams D., Buiiermann B. and Walters D.L. 1984. Pulsatile secretion of gonadotrophins, ovarian steroids and ovarian oxytocin

PAGE 251

244 during prostaglandin-induced regression of the corpus luteum in the cow. J. Reprod. Fertil. 71:493. Schallue-Francis T.K., Farin P.W., Cross J.C., Keisler D. and Roberts R.M. 1991. Effect of injected bovine interferon-ajl on oestrous cycle length and pregnancy success in sheep. J. Reprod. Fertil. 91:347. Schams D., Kruip T.A.M. and Koll R. 1985. Oxytocin determination in steroid producing tissues and in vitro production in ovarian follicles. Acta Endocrinol. 109:530. Schams D., Schallenberger E., Hoffman B. and Karg H. 1977. The oestous cycle of the cow: hormonal parameters and time relationships concerning estrous, ovulation, and electrical resistance of the vaginal mucus. Acta Endocrinol. 86:180. Scheldrick E.L. and Flint A.P.F. 1985. Endocrine control of uterine oxytocin receptors in the ewe. J. Endocrinol. 106:249. Schramm W.L., Bovaird L., Glew M.E., Schramm G. and McCracken J.A. 1983. Corpus luteum regression induced by ultra-low pulses of prostaglandin F2„. Prostaglandins 26:347. i Schramm W.L., Einer-Jensen N., Schramm G. and McCracken J.A. 1986. Local exchange of oxytocin from the ovarian vein to the ovarian arteries in sheep. Biol. Reprod. 34:671. Schrey M.P. and Rubin R.P. 1979. Characterization of a calcium-mediated activation of arachidonic acid turnover in adrenal phospholipids by corticotropin. J. Biol. Chem. 254:11234. Schrey M.P., Read A.M. and Steer P.J. 1986. Oxytocin and vasopressin stimulate inositol phosphate production in human gestational myometrium and decidual cells. Biosci. Reports 6:613. Schultz H. 1991a. Beta oxydation of fatty acids. Biochim. Biophys. Acta 1081:109. Schultz H. 1991b. Oxydation of fatty acids. In: Vance D.E. and Vance J. (Eds) Biochemistry of lipids, lipoproteins and membranes, p 87. Elsevier, Amsterdam. Schwall R.H., Sawyer H.R. and Niswender G.D. 1986. Differential regulation of LH by prostaglandins of steroidogenesis in small and large luteal cells of the ewe. J. Reprod. Fertil. 76:821.

PAGE 252

245 Schwegler J.S., Heppelmann B., Mildenberger S. and SilbemagI S. 1991. Receptor-mediated endocytosis of albumin in cultured opossum kidney cells: a model for proximal tubular protein reabsorption. Pflugers Arch. 418:383. Scott D.L, White S.P., Browning J.L, Rosa J.J., Gelb M.H. and Sigler P.B. 1991. Structures of free and inhibited human secretory phospholipase Ajfrom inflamatory exude. Science 254:1007. Sebalt R.J., Sheller J.R., Gates J.A., Roberts LJ. II and Fitzgerald G.A. 1990. Inhibition of eicosanoid biosynthesis by glucocorticoids in humans. Proc. Natl. Acad. Sci. USA 87:6974. Senior J., Sangha R., Baxter G.S., Marshall K. and Clayton J.K. 1992. In vitro characterization of prostanoid FP-, DP-, IPand TP-receptors on the non-pregnant human myometrium. Br. J. Pharmacol. 107:215. Shands J.H. and Noble R.C. 1984. The lipid composition of ovine placental tissue and its subcellular fractions. Comp. Biochem. Physiol. 2:291. Sharif S.F., Francis H., Keisler D.H. and Roberts R.M. 1989. Correlation between the release of ovine trophoblast protein-1 by the conceptus and the production of polypeptides by the maternal endometrium of ewes. J. Reprod. Fertil. 85:471. Sharp J.D., White D.L., Xiou X.G., Goodson T., Gamboa G.C., McClure D., Burgett S., Hoskins J., Scatrud P.L., Sportsman J.R., Becker G.W., Kang L.H., Roberts, E.F. and Kramer R.M. 1991. Molecular cloning and expression of human Ca^*-sensitive cytosolic phospholipase A2. J. Biol. Chem. 266:14850. Shaw J.E. and Ramwell P.W. 1967. Prostaglandin release from the adrenal gland. In: Bergstrom S. and Samuelsson B. (Eds), Nobel symposium 2, Prostaglandins, p 291 . Interscience, New York. Sheldrick E.L. and Flick-Smith H.C. 1993. Oestradiol enhances oxytocinstimulated PGFj^ production by cultured ovine endometrium without affecting oxytocin receptor binding protein activity. J. Reprod Fertil 11:168 (abst). Sheldrick E.L., Mitchell M.D. and Flint A.P.F. 1980. Delayed luteal regression in ewes immunized against oxytocin. J. Reprod. Fertil. 59:37.

PAGE 253

246 Shemesh M. 1979. Inhibitory action of follicular fluid on progesterone and prostaglandin synthesis in bovine follicles. J. Endocrinol. 82:27. Shemesh M., Hansel W. and Strauss J.F. 1984a. Modulation of bovine placental prostaglandin synthesis by an endogenous inhibitor. Endocrinology 115:1401. Shemesh M., Hansel W. Strauss J.F., Rafaeli A., Lavi S. and Mileguir F. 1984b. Control of prostanoid synthesis in bovine trophoblast and placentomes. Anim. Reprod. Sci. 7:177. Short E.G., Geisert R.D.. Helmer S.D., Zavy M.T. and Fulton R.W. 1991. Expression of antiviral activity and induction of 2,5-oligoadenylate synthetase by conceptus secretory proteins enriched in bovine trophoblast protein-1. Biol. Reprod. 44:261. Short R.V. 1969. Implantation and the maternal recognition of pregnancy. In: Wolstenholme G.E.W. and O'Connor M. (Eds), Foetal autonomy, p 2. J. & A. Churchill Ltd., London. Silvia W.J. and Homanics G.E. 1988. Role of phospholipase C in mediating oxytocin-induced release of PGFj^ from ovine endometrial tissue. Prostaglandins 35:535. Silvia W.J. and Raw R.E. 1993. Regulation of pulsatile secretion of RGFj^ from the ovine uterus by ovarian steroids. J. Reprod. Fertil. 98:341. Silvia W.J. and Taylor M.L. 1989. Relationship between uterine secretion of PGF2 alpha induced by oxytocin and endogenous concentrations of estradiol and progesterone at three stages of the bovine estrous cycle. J. Anim. Sci. 67:2347. Silvia W.J., Lewis G.S., McCracken J.A., Thatcher W.W. and Wilson Jr. L. 1991. Hormonal regulation of uterine secretion of PGFj^ during luteolysis in ruminants. Biol. Reprod. 45:655. Silvia W.J., Raw R.E., Aldrich S.L. and Hayes S.H. 1992. Uterine secretion of prostaglandin Fj^ in response to oxytocin in ewes: changes during the estrous cycle and early pregnancy. Biol. Reprod. 46:1007. Sinclair H.M., 1990. Essential fatty acids an historical perspective. Biochem. Soc. Trans. 18: 756.

PAGE 254

Sirois J. and Fortune J.E. 1988. Ovarian follicular dynamics during the estrous cycle in heifers monitored by real-time ultrasonography. Biol. Reprod. 39:308. Sirois J. and Richards J.S. 1992. Purification and characterization of a novel, distinct isoform of prostaglandin endoperoxide synthase induced by human chorionic gonadotropin in granulosa ceils of rat preovulatory follicles. J. Biol. Chem. 267:6382. Sirois J., Simmons D.L. and Richards J.S. 1992. Hormonal regulation of messenger RNA encoding a novel isoform of prostaglandin endoperoxide synthase in rat preovulatory follicles. Induction in vivo and in vitro. J. Biol. Chem. 267:11586. Skopets B, Li J., Thatcher W.W., Roberts R.M. and Hansen P.J. 1992. Inhibition of lymphocyte proliferation by bovine trophoblast protein-1 (Type I trophoblast interferon) and bovine interferon-a|1. Vet. Immunol. Immunopathol. 34:81. Smith P.K., Krohn R.I., Hermanson G.T., Mallia A.K., Gartner F.H., Provenzano M.D., Fujimoto E.K., Goeke N.M., Oson B.J. and Klenk D.C. 1985. Measurement of protein using bicinchoninic acid. Anal. Biochem. 150:76. Smith W.L. 1989. Review article: The eicosanoids and their biochemical mechanisms of action. Biochem. J. 259:315. Smith W.L. 1992. Prostanoid biosynthesis and mechanisms of action. Am. J. Physiol. 263 (Renal Fluid Electrolyte Physiol. 32):F181. Smith W.L. and Lands W.E. 1972. Oxygenation of polyunsaturated fatty acids during prostaglandin biosynthesis by sheep vesicular glands. Biochemistry 11:3276. Smith W.L. and Marnett L.J. 1991. Prostaglandin endoperoxide synthase: structure and catalysis. Biochim. Biophys. Acta 1083:1. Smith W.L., Marnett L.J. and DeWitt D.L. 1991. Prostaglandin and thromboxanes biosynthesis. Pharmac. Ther. 49:153. Smrcka A.V., Hepler J.R., Brown K.O. and Sterweis P.C. 1991. Regulation of polyphosphoinositide-specific phospholipase C activity by purified Gq. Science 251 :804. 247

PAGE 255

248 Snyder F. 1990. PAF and related acetylated lipids as potent biologically active cellular mediators. Am. J. Physiol. 259: C697. Soloff M.S. 1975. Uterine receptor for oxytocin: effects of estrogen. Biochem. Biophys. Res. Comm. 65:205. Specter A.A. 1968. The transport and utilization of free fatty acid. 1968. Ann. N.Y.Acad. Sci. 149:768. Spector A.A. 1975. Fatty acid binding to plasma albumin. J. Lipid Res. 16:165. Specter A.A., John K. and Fletcher J.E. 1969. Binding of long-chain fatty acids to bovine serum albumin. J. Lipid Res. 10:56. Spinks N.R. and O'Neill C. 1987. Embryo-derived PAF is essential for the establishment of pregnancy in the mouse. Lancet 2:1035. Sprague E.A., Kelley J.L, Suenram C.A., Valente A.J., Abreu-Macomber M. and Schwatrz C.J. 1985. Stimulation of albumin endocytosis by cationized ferritin in cultured aortic smooth muscle cells. Am. J. Pathol. 121:433. Sprecher H. 1981. Biochemistry of essential fatty acids. Prog. Lipid Res. 20:13. Stark G.R. and Kerr LM. 1992. Interferon-dependent signaling pathways: DNA elements, transcription factors, mutations, and effect of viral proteins. J. Interferon Res. 12:147. Stewart H.J., Guesdon F.M.J. , Payne J.H., Charleston B., Vallet J.L. and Flint A.P.F. 1992. Trophoblast interferons in early pregnancy of domestic ruminants. J. Reprod. Fertil. Suppl. 45:59. Stewart H.J., McCann S.H.E. and Flint A.F.P. 1990. Structure of an interferon-a2 gene expressed in the bovine conceptus early in gestation. J. Mol. Endocrinol. 4:275. , < Stewart H.J., McCann S.H.E. , Barker P.J., Lee K.E., Lamming G.E. and Flint A.F.P. 1987. Interferon sequence homology and receptor binding activity of ovine trophoblast antiluteolytic protein. J. Endocrinol. 115:R13. Stewart H.J., McCann S.H.E., Lamming G.E. and Flint A.F.P. 1989a. Evidence for a role for interferon in the maternal recognition of pregnancy. J. Reprod. Fertil. (Suppl.) 37:127.

PAGE 256

249 Stewart H.J., McCann S.H.E., Northtrop A.J., Lamming G.E. and Flint A.F.P. 1989b. Sheep antiluteolytic interferon: cDNA sequence and analysis of mRNA levels. J. Mol. Endocrinol. 2:65. Stubbs CD. and Kisielewski A. 1990. Effect of increasing level of co3 fatty acids on rat skeletal muscle sarcoplasmic reticulum. Lipids 25:553. Sugimoto Y., Namba T., Honda A., Hayashi Y., Negishi M., Ichikawa A. and Narumiya S. 1992. Cloning and expression of a cDNA for mouse prostaglandin E receptor EPS subtype. J. Biol. Chem. 267:6463. Takahashi T., Suguhara T. and Ohsaka A. 1981. Phospholipase C from Clostridium perfringens. Meth. Enzymol. 71:710. Takegushi C, Kohne E. and Sih C.J. 1971. Mechanism of prostaglandin biosynthesis. I. Characterization and assay of bovine prostaglandin synthase. Biochemistry 10:2372. Tanaka Y., Ward S. and Smith W.L. 1987. Immunochemical and kinetic evidence for two different prostaglandin H-prostaglandin E isomerases in sheep vesicular gland microsomes. J. Biol. Chem. 262:1374. Terragano N.A., Terragano A., Early J.A., Roberts M.A. and McGiff J.C. 1978. Endogenous prostaglandin synthesis inhibitor in the renal cortex. Effects on prostacyclin by renal blood vessels. Clin. Sci. Mol. Med. 55:199. Thatcher W.W., Bartol F.F., Knickerbocker J. J., Curl J.S., Wolfenson D., Bazer F.W. and Roberts R.M. 1984b. Maternal recognition of pregnancy in cattle. J. Dairy Sci. 67:2797. Thatcher W.W., Bazer F.W., Sharp D.C. and Roberts R.M. 1986a. Interrelationship between uterus and conceptus to maintain corpus luteum function in early pregnacy : sheep, cattle, pigs and horses. J. Anim. Sci. 62:25. Thatcher W.W., Terqui M., Thimonier J. and Mauleon P. 1986b. Effect of estradiol-17p on peripheral olasma concentration of 15-keto-13,14dihydro PGFj^ and luteolysis in cyclic cattle. Prostaglandins 31:745. Thatcher W.W., Wolfenson D., Curl J.S., Rico L.E., Knickerbocker J. J., Bazer F.W. and Drost M. 1984a. Prostaglandin dynamics associated with early development of the bovine conceptus. Anim. Reprod. Sci. 7:149.

PAGE 257

250 Thiery J.P., Duband J.L. and Tucker G.C. 1985. Cell migration in the vertebrate embryo. Annu. Rev. Cell Biol. 1:91. Thornburn G.D., Cox R.I., Currie W.B., Restall B.J. and Schneider W. 1973. Prostaglandin F and progesterone concentrations in the utero-ovarian venous plasma of the ewe during the estrous cycle and early pregnancy. J. Reprod. Fertil. Suppl. 18:151. Tiwari R.K., Kusari J., Kumar R. and Sen G.C. 1988. Gene induction by interferons and double stranded RNA: selective inhibition by 2aminopurine. Mol. Cell. Biol. 8:4289. Tiwari R.K., Mukhopadhyay B., Telang N.T. and Osborn M.P. 1991. Modulation of gene expression by selected fatty acids in human breast cancer cells. Anticancer Res. 11:1 383. Vallet J.L. and Bazer F.W. 1989. Effect of ovine trophoblast protein-1, oestrogen and progesterone on oxytocin-induced phosphatidylinositol turnover in endometrium of sheep. J. Reprod. Fertil. 87:755. Vallet J.L., Barker P. J., Lamming G.E., Skinner N. and Huskisson N.S. 1991. A low molecular weight endometrial secretory protein which is increased by ovine trophoblast protein-1 is a beta 2-microglobulin-like protein. J. Endocrinol. 130:R1. Vallet J.L., Bazer F.W., Fliss M.F.V. and Thatcher W.W. 1990. Control of oxytocin receptor and uterine response to oxytocin by progesterone and estradiol in the ewe. J. Reprod. Fertil. 90:625. Vallet J.L., Bazer F.W., Fliss M.F.V. and Thatcher W.W. 1988. The effect of ovine conceptus secretory proteins and purified ovine trophoblast protein-1 on interestrous interval and plasma concentrations of PGFj^, PGE2 and PGFM in cyclic ewes. J. Reprod. Fertil. 84:493. Vallet J.L., Bazer F.W. and Roberts R.M. 1987. The effect of ovine trophoblast protein-1 on endometrial protein secretion and cyclic nucleotides. Biol. Reprod. 37:1307. Vallet J.L., Gross T.S., Fliss M.F.V. and Bazer F.W. 1989. Effects of pregnancy, oxytocin, ovine trophoblast protein-1 and their interactions on endometrial production of prostaglandin Fj^ in vitro in perifusion chambers. Prostaglandins 38: 1 1 3.

PAGE 258

251 Vance D.E. 1991. Phospholipids metabolism and cell signaling in eucaryotes. In: Vance D.E. and Vance J. (Eds), Biochemistry of lipids, lipoproteins and membranes, p 205. Elsevier, Amsterdam. Van den Bosch H. 1980. Intracellular phospholipases A. Biochim. Biophys. Acta 604:191. Vanderhoek J.Y. and Lands W.E. 1973. Acetylenic inhibitors of sheep vesicular gland oxygenase. Biochim. Biophys. Acta 296:374. Van der Ouderaa F.J., Buytenhek M., Nugteren D.H. and van Dorp D.A. 1977. Purification and characterization of prostaglandin endoperoxide synthetase from sheep vesicular glands. Biochim. Biophys. Acta 487:315. Van der Ouderaa F.J., Buytenhek M., Slikkerveer F.J. and Van Dorp D.A. 1979. On the haemoprotein character of prostaglandin endoperoxide synthetase. Biochem. Biophys. Acta 572:29. Veals S.A., Santa maria T., Levy D.E. 1993. Two domains of ISGF3y that mediate protein-DNA and protein-protein interactions during transcription factor assembly contribute to DNA specific binding. Mol. Cell. Biol. 13:196. Veerkamp J.H., Peeters R.A. and Maatman R.G.H.J. 1991. Structural and , functional features of different types of cytoplasmic fatty acid-binding proteins. Biochim. Biophys. Acta 1081:1. Velazquez L., Fellous M., Stark G.R. and Pelligrini S. 1992. A protein tyrosine kinase in the interferon a/p signaling pathway. Cell 70:313. Verheij H.M., Slotboom A.J. and de Haas G.H. 1981. Structure and function of phospholipase A2. Rev. Physiol. Biochem. Pharmacol. 91:91. Viviani R. 1970. Metabolism of long chain fatty acids in the rumen. Adv. Lipid Res. 8:267. Voelker D.R. 1991. Lipid assembly into cell membranes. In Vance D.E. and Vance J. (Eds), Biochemistry of lipids, lipoproteins and membranes, p 489. Elsevier, Amsterdam. Vogt, W. 1978. Role of phospholipase A-2 in prostaglandin formation. In: Galli C, Galli G. and Porcellati G. (Eds), Advances in prostaglandin and thromboxane research, Vol. 3, p 89. Raven Press, New York.

PAGE 259

252 Waite M. 1987. The phospholipases. In: Hanahan D.J. (Ed), Handbook of lipid research. Vol. 5. Plenum, New York. Waite M. 1991. Phospholipases. In Vance D.E. and Vance J. (Eds), Biochemistry of lipids, lipoproteins and membranes, p 269. Elsevier, Amsterdam. Walters D.L., Schams D. and Schallenberger E. 1984. Pulsatile secretion of gonadotrophins, ovarian steroids and ovarian oxytocin during the luteal phase of the estrous cycle. J. Reprod. Fertil. 71:479. Wallner B.P., Mattaliano R.J. Hession C, Gate R.L, Tizard R., Sinclair L.J., Foeleer C, Chow E.P., Browning J.L, Ramachandran K.L. and Pepinsky R.B. 1986. Cloning and expression of of human lipocortin, a phospholipase inhibitor with potential anti-inflammatory activity. Nature 320:77. Watanabe K., Fuji! Y., Nakayama K., Ohkubo H., Kuramitzu S., Hayashi H., Kagamiyama H., Nakanishi S. and Hayaishi O. 1989. Cloning, nucleotide sequence and gene expression of bovine lung prostaglandin F synthase. Adv. Prostaglandin Thromboxane Leukot. Res. 19:462. Watanabe T., Umegaki H. and Smith W.L. 1986. Association of a prostaglandin Ej binding protein with a pertussis toxin-reactive guanine nucleotide regulatory protein. J. Biol. Chem. 262:13430. Watanabe K., Yoshida R., Shimizu T. and Hayaishi 0. 1985. Enzymatic formation of prostaglandin Fj^ from prostaglandin H2 and D2: purification and properties of prostaglandin F synthase from bovine lung. J. Biol. Chem. 260:7035. Wathes D.C., Barrett J., Gilbert C.L. and Yaron Z. 1988. Regulation of ovarian oxytocin secretion in domestic ruminants. In: Yoshida S. and Share L. (Eds), Recent progress in posterior pituitary hormones, p 227. Elsevier, Amsterdam. Wathes D.C. and Denning-Kendall P.A. 1992. Control of synthesis and secretion of ovarian oxytocin in ruminants. J. Reprod. Fertil. Suppl. 45:39. Wathes D.C, Guldenaar S.E.F., Swann R.W., Webb R., Peter D.G. and Pickering B.T. 1986. A combined radioimmunoassay and immunocytochemical study of ovarian oxytocin production during the periovulatory period in the ewe. J. Reprod. Fertil. 78:167.

PAGE 260

253 Wathes, D.C., Swann R.W., Birkett S.D., Porter D.G. and Pickering B.T. 1983. Characterization of oxytocin, vasopressin and neurophysin from the bovine corpus luteum. Endocrinology 113:693. Wathes D.C., Swann R.W. and Pickering B.T. 1984. Variation in oxytocin, vasopressin and neurophysin concentrations in the bovine ovary during estrus cycle and pregnancy. J. Reprod. Fertil. 71:551. Watkins W.B. and Moore L.G. 1987. Effect of systemic infusion of PGFj,, and 13,14-dihydro-15-keto PGFja on the release of oxytocin-associated neurophysin from the ovary in the ewe. Biol. Reprod. 40:1215. Wene J.D., Connor W.E. and DenBesten L. 1975. The development of essential fatty acid deficiency in healthy men fed fat free diets intravenously and orally. J. Clin. Invest. 56:127. Wery J. P., Schevitz R.W., Clawson D.K., Bobbitt J.L, Dow E.R., Gamboa G., Goodson T.Jr., Hermann R.B., Kramer R.M., McClure D.B., Milhelich E.D., Putnam J.E., Sharp J.D., Stark D.H., Teater C, Warrick M.W. and Jones N.D. 1991. Structure of recombinant human rheumatoid arthritic synovial fluid phospholipidase at 2.2 A resolution. Nature 352:79. Wettemann R.P., Hafs H.D., Edgerton L.A. and Swanson L.V. 1972. Estradiol and progesterone in blood during the bovine estrous cycle. J Anim Sci. 34:1020. Wheeler M.B. and Veldhuis J.D. 1987. Catalytic and receptor-binding properties of the calcium-sensitive phospholipid-dependent protein kinase (protein kinase C) in swine luteal cytosol. Mol. Cell. Endocrinol. 50:123. Wijkander J. and Sundler R. 1992. Macrophage arachidonate-mobilizing phospholipase /K^. role of Ca^*for membrane binding but not for catalytic activity. Biochem. Biophys. Res. Commun. 184:118. Williams B.R.G. 1991. Signal transduction and transcriptional regulation of interferon-a-stimulated genes. J. Interferon Res. 11:207. Wilson T. and Ganendren R. 1992. Serum concentrations of secretory IgA in pregnancies delivering at term or preterm. Prostaglandins 44:373. Wiltbank J.N. and Casida L.E. 1956. Alteration of ovarian activity by hysterectomy. J. Anim. Sci. 15:134.

PAGE 261

254 Wiltbank J.N., Diskin M.G. and Niswender G.D. 1991. Differential actions of second messenger systems in the corpus luteum. J. Reprod. Fertil. Suppl. 43:65. Wiltbank M.C., Kater S.B., Guthrie P.B., Mattson M.P. and Niswender G.D. 1989a. Hormonal regulation of free calcium concentrations in small and large ovine luteal cells. Biol. Reprod. 41:771. Wiltbank M.C., Knickerbocker J.J. and Niswender G.D. 1989b. Regulation of the corpus luteum by protein kinase C. I. Phosphorylation activity and steroidogenic action in large and small ovine luteal cells. Biol. Reprod. 40:1194. Wirtz K.W.A. 1974. Transfer of phospholipids between membranes. Biochem. Biophys. Acta 344:95. . . . . , . Wirtz K.W.A. and Gadella T.W.J. 1990. Properties and modes of action of specific and nonspecific phospholipid transfer proteins. Experientia 46:592. Wirtz K.W.A. and Zilversmit D.B. 1968. Exchange of phospholipids between liver mitochondria and microsomes in vitro. J. Biol. Chem. 243:3596. Wlodawer P.H., Kindahl H. and Hamberg M. 1976. Biosynthesis of prostaglandin from arachidonic acid and prostaglandin endoperoxides in the uterus. Biochim. Biophys. Acta 431:603. Wolfenson D., Bartol F.F., Badinga L., Barros CM., Marple D.N., Cummins K., Wolfe D., Lucy M.C., Spencer T.E. and W.W. Thatcher. 1993. Secretion of PGFja and oxytocin during hyperthermia in cyclic and pregnant heifers. Theriogenology 39:1129. Wolfenson D., Thatcher W.W., Drost M., Caton D., Foster D.B. and LeBlanc M.M. 1985. Characteristics of prostaglandin F measurements in the ovarian circulation during the estrous cycle and early pregnancy in the cow. J. Reprod. Fertil. 75:491 . Wright K., Luborski-Moore J.L. and Behrman H.R. 1979. Specific binding of prostaglandin Fj^ to membranes of rat corpora lutea. Mol. Cell Endocrinol. 13:25. Xavier F., Guillemot M., Charlier M., Martal J. and Gaye P. 1991. Co-expression of the proto-oncogene fos (c-fos) and an embryonic interferon (ovine trophoblastin) by sheep conceptuses during implantation. Biol. Cell. 73:27.

PAGE 262

255 Xie W.L, Chipman J.G., Robertson D.L., Erikson R.L. and Simmons D.L. 1991. Expression of a mitogen-responsive gene coding prostaglandin synthase is regulated by mRNA splicing. Proc. Natl. Acad. Sci. USA 88:2692. Yan C, Sehgal P.B. and Tamm I. 1989. Signal transduction pathways in the induction of 2',5'-oligoadenylate synthetase gene expression by interferon a/p. Proc. Natl. Acad. Sci. USA 86:2243. Yap W.H., Teo T.S. and Tan YH. 1986. An early event in the interferon-induced transmembrane signaling process. Science 234:355. Yavin E. and Menkes J.H. 1974. Polyenoic acid metabolism in cultured dissociated brain cells. J. Lipid Res. 15:152. Yokota K., Kusake M., Ohshima T., Yamanoto S., Kurihara N., Toshina T. and Kumegawa M. 1986. Stimulation of prostaglandin E2 synthesis in cloned osteoblasts cells of mouse (MC3T3-E1) by epidermal growth factor. J. Biol. Chem. 261:15410. Young A.J. and Barker K.L. 1991. Effect of estradiol and progesterone on long chain fatty acyl-coenzyme A levels in the rat uterus. Biochim. Biophys. Acta 1092:211. Zarco L., Stabenfeldt G.H., Basu S., Bradford G.E. and Kindahl H. 1988. Modification of prostaglandin Fj^ and release in the ewe during the initial establishment of pregnancy. J. Reprod. Fertil. 83:527. Zhang J., Weston P.O. and Hixon J.E. 1991. Influence of estradiol on the secretion of oxytocin and PGFj^ during luteolysis in the ewe. Biol. Reprod. 45:395. Ziboh V.A. and Hsia S.L. 1972. Effects of prostaglandin Ej on rat skin: inhibition of sterol ester biosynthesis and clearing of scaly lesions in essential fatty acid deficiency. J. Lipid Res. 13:458. Zupan L.A., Steffens D.L., Berry C.A., Landt M. and Gross R.W. 1992. Cloning and expression of a human 14-3-3 protein mediating phospholipolysis. Identification of an arachidonoyl-enzyme intermediate during catalysis. J. Biol. Chem. 267:8707.

PAGE 263

BIOGRAPHICAL SKETCH Guenahel H. Danet-Desnoyers was bom in Clamecy, Nievre, France on August 17, 1963. In the fall of 1982, he enrolled at the Universite Pierre et Marie Curie in Paris, France, and received a MaTtrise in Biochemistry and Molecular Biology in 1987. In summer of 1988, he completed a 6 month study period in the laboratory of W. W. Thatcher at the University of Florida. He returned to Paris to begin his graduate studies at the Universite Pierre et Marie Curie. His program of research was done under the direction of J. Martal and G. Charpigny in the Unite d'Endocrinologie de I'Embryon, at the Institut National de Recherche Agronomique (INRA). His thesis was entitled "In vitro study of the mechanism of action of an embryonic interferon: ovine trophoblast protein-1 inhibits PGFj^ synthesis by ovine endometrial cells." He received a Diplome d' Etudes Approfondies in Reproductive Biology in fall1989. In January 1990, he began a doctoral program under the direction of W. W. Thatcher in the Department of Dairy Science, University of Florida, Gainesville. Upon completion of his Ph. D. program in April 1994, he will join M. J. Lawman at the Walt Disney Memorial Cancer Institute, Florida Hospital, Orlando, as a postdoctoral fellow. 256

PAGE 264

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William W. Thatcher, Chairman Graduate Research Professor of Animal Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William C. Buhi Associate Professor of Biochemistry and Molecular Biology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Jnsen Professor of Animal Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Kjjtj^ Michael J. Lawman Senior Research Scientist of Walt Disney Memorial Cancer Institute

PAGE 265

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Frank A. Simmen Associate Professor of Animal Science This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to theGraduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. April, 1994 Ilege of Agriculture Dean, Graduate School

PAGE 266

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William W. Thatcher, Chairman Graduate Research Professor of Animal Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William C. Buhi Associate Professor of Biochemistry and Molecular Biology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Jnsen Professor of Animal Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Kjjtj^ Michael J. Lawman Senior Research Scientist of Walt Disney Memorial Cancer Institute

PAGE 267

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Frank A. Simmen Associate Professor of Animal Science This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to theGraduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. April, 1994 Ilege of Agriculture Dean, Graduate School


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EPJTIVBVM_43UHR3 INGEST_TIME 2015-04-03T19:08:00Z PACKAGE AA00030051_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES