O

AE

Ha

b

a

ARAA

KPPOUR

1

ctaq(caasfFcbidd

tflo

1h

Prostaglandins & other Lipid Mediators 106 (2013) 23– 28

Contents lists available at ScienceDirect

Prostaglandins and Other Lipid Mediators

riginal research article

comprehensive immunohistochemistry of prostaglandins F2� and2 synthetic enzymes in rat ovary and uterus around parturition

ironori Satoha, Kikuko Watanabeb, Mitsumori Kawaminamia, Shiro Kurusua,∗

Laboratory of Veterinary Physiology, Kitasato University School of Veterinary Medicine, Towada, Aomori 034-8628, JapanDepartment of Nutrition, Koshien University, Takarazuka, Hyogo 665-0006, Japan

r t i c l e i n f o

rticle history:eceived 31 May 2013ccepted 26 July 2013vailable online 6 August 2013

a b s t r a c t

A comprehensive immunohistochemistry with the isoform-distinguishable antibodies againstprostaglandin (PG) F2� and PGE2 biosynthetic enzymes was undertaken to identify the cellular types andenzyme isoforms in rat ovary and uterus around parturition. In general ovarian and uterine cells showedpositive immunoreactions for phospholipase A2 groups 4A and 6A, but not group 2A, and cyclooxygen-

eywords:GF2�

GE2

varyterusat

ase (COX)-1 rather than COX-2. Their immunoreactions for PGF2� synthase and PGE2 synthase werecell type-dependently variable. The putative PGF2� and PGE2 producing cell types included, as expected,ovarian luteal cells, uterine endometrial epithelium and myometrium, and cervical connective tissue and,unexpectedly, ovarian stromal cells and basal lamina of cervical endometrium. Obtained data indicatethe generation of PGF2� and PGE2 by multiple sites, which are entirely the same as established sites ofactions, in parturition processes and tissue-dependent differential usage of PG biosynthetic pathway.

. Introduction

Gestation is a mammalian unique reproductive process that isomposed of multiple sequential events and culminates in parturi-ion (reviewed in Refs. [1,2]). This final event is another sequentialnd integrated process composed of progesterone withdrawal fre-uently due to functional regression of ovarian corpus luteumCL), fetal membrane rupture, ripening and dilatation of uterineervix, uterine smooth muscle contraction, placental separation,nd uterine involution (reviewed in Refs. [1–3]). Its successfulccomplishment requires the spatially and temporally regulatedynthesis and actions of a variety of endocrine and auto/paracrineactors. In most mammalian species examined, prostaglandin (PG)2� and PGE2 have essential or critical roles in all of parturition pro-esses described above (reviewed in Refs. [1–3]). PGs act basicallyy auto/paracrine fashion and, exceptionally but very importantly

n ruminants and rodents, by endocrine fashion where uterine-erived PGF2� passes via a counter-current mechanism and actsirectly on CL (reviewed in Refs. [1–4]).

PG is generally generated via three sequential enzymatic reac-ions (reviewed in Refs. [5,6]); (1) arachidonic acid (AA) is cleaved

rom membrane glycerophospholipid by the action of phospho-ipase A2 (PLA2), (2) AA is converted to PGH2 via the actionf prostaglandin endoperoxide synthase (PGHS) also known as

∗ Corresponding author. Tel.: +81 176 23 4371x416; fax: +81 176 23 8703.E-mail address: kurusu@vmas.kitasato-u.ac.jp (S. Kurusu).

098-8823/$ – see front matter © 2013 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.prostaglandins.2013.07.005

© 2013 Elsevier Inc. All rights reserved.

cyclooxygenase (COX), and (3) PGH2 is metabolized to F2�, E2, andother types of PGs by respective PG terminal synthases (Fig. 1).The type(s) of PGs generated is determined by terminal PG syn-thase(s) expressed, whereas the amount of PG is regulated by all ofPLA2, COX, and PG synthases those are under transcriptional con-trol depending on cell types and stimuli [5–7]. There are multipleisoforms in each of PLA2, COX, PGF2� synthase (PGFS), and PGE2synthase (PGES). PLA2 constitutes a large superfamily with over 20members in mammals, among which group 4A PLA2 (PLA2G4A) andseveral others have been demonstrated to be expressed by femalereproductive tissues [5,7]. COX has only two isoforms with COX-1 and COX-2, both of which are present in female reproductiveorgans [5,6]. PGES has three isoforms with cytosolic type (cPGES)and microsomal types 1 and 2 (mPGES-1 and -2) [5,6]. PGFSs arenow found to have several isoforms catalyzing PGH2 to PGF2� orconverting PGE2 to PGF2� (Fig. 1) and to belong to aldo-keto reduc-tase (AKR) superfamily [5].

A great number of studies have accumulated on the expres-sion and localization, at mRNA and/or protein levels, of mainlyCOXs and less or additionally PLA2s in gestational tissues of humanand many other mammalian species. If we limit these referencesto those on rodents and rabbits, CL has been shown to expressPLA2G1B [8], PLA2G4A [8], COX-1 [9,10], and COX-2 [9–11]. Uter-ine myometrial, endometrial and cervical tissues of rodents and

rabbits have been shown to express PLA2G2A [7,12], PLA2G4A[7,13–16], PLA2G5 [7,12], PLA2G6 [7,17], COX-1 [14,18–22,24–31],and COX-2 [14,19–21,23–31]. The reports have shown gestationalage-related changes in expression and/or activity of the enzymes

24 H. Satoh et al. / Prostaglandins & other

Fig. 1. Biosynthetic pathway of PGF2� and PGE2. PLA2 liberates AA from membranephospholipid and this fatty acid is metabolized to PGH2 by COX. This common PGsAi

a1cpeIebaCsfniii

ibp(bA“wa

2

2

wlaf[CtbttparS

ubstrate is further metabolized to PGF2� and PGE2 by PGFS and PGES, respectively.KR1C5 converts PGE2 to PGF2� . Described are isoforms of each enzyme those were

nvestigated in this study.

nd their regulations by progesterone [26,27,29–31], estradiol-7� [12,26,31], glucocorticoid [6], oxytocin [12], pro-inflammatoryytokines [6,7,19,21], or PGF2� [10,11]. There are clearly not a fewotential discrepancies among the results on temporal changes inxpression, cellular distribution and regulation of the enzymes.n contrast to COXs, current knowledge about their down-streamnzymes, PGFSs [10,14,32,33] and PGESs [34,35], in rodent and rab-it gestational tissues is very limited and undefined. The studiesccumulated so far have focused largely on one step (very mostlyOX) or two (additionally PLA2, PGFS, or PGES) of PGF2� or PGE2ynthetic pathway and on their expression at the tissue level. Asar as we know, only two references are available with simulta-eous demonstration of all of PLA2, COX, PGFS, or PGES expression

n human [36] and ovine [37] gestational tissues. Therefore, thenformation on the thorough PGs biosynthetic steps and their local-zations at the cellular level are almost unclear.

The purpose of this study was to identify immunohistochem-cally the cell type(s) and enzyme isoforms of PGF2� and PGE2iosynthesis in the ovary and uterus in late pregnancy and aroundarturition in rats. As for PLA2s we selected three isoformsPLA2G2A, PLA2G4A, and PLA2G6A) from a number of members,ased on the previous findings [5,7]. As for PGFSs, we investigatedKR1C7 (formerly the “lung-type” PGFS or PGFS I), AKR1C11 (theliver-type” PGFS or PGFS II), and AKR1C5 (PGE2 9-keto reductase)hose specific antibodies with sufficient characterization were

vailable.

. Materials and methods

.1. Antibodies and chemicals

The sources and characteristics of the antibodies currently usedere as follows. The anti-PLA2G2A antibody, a kind gift from the

ate Dr. Ichiro Kudo (Showa University, Tokyo, Japan), was gener-ted in rabbits against rat platelet enzyme with 14 kDa and wasractionated to IgG [38]. The anti-PLA2G4A polyclonal antibody13,39], a kind gift from Pfizer Co. (formerly Genetics Institutes,ambridge, MA, USA), was generated in rabbits against dena-ured human PLA2G4A extracted from Escherichia coli inclusionody. The rabbit antisera were generated against bovine lung-ype PGFS (AKR1C7) and bovine liver-type PGFS (AKR1C11) andheir cross-reactivities with each of rat antigens were confirmed

reviously [40,41]. The antiserum against AKR1C5 was gener-ted through immunizing rabbits with synthetic peptide fromat AKR1C5 (amino acids 303–317). This was manufactured byigma–Aldrich Japan (Ishikari, Japan). All other antibodies were

Lipid Mediators 106 (2013) 23– 28

purchased from Cayman Chemical Co. (Ann Arbor, MI, USA). Therabbit anti-PLA2G6A antiserum (Catalog No. 160507) was gener-ated against synthetic peptide from CHO cell-derived PLA2G6A,which had 95% homology with rat enzyme, and did not cross-reactwith either PLA2G2A or PLA2G4A. The affinity-purified IgG of rab-bit polyclonal anti-COX-1 antibody (Catalog No. 160109, Lot No.151667) was generated against synthetic peptide from rat COX-1 (amino acids 274–288) and did not cross-react with COX-2. Therabbit polyclonal anti-COX-2 antibody (Catalog No. 160106, Lot No.155745) was generated against synthetic peptide from rat COX-2(amino acids 570–598) and did not cross-react with COX-1. Therabbit polyclonal anti-cPGES antibody (Catalog No. 160150, Lot No.0410417-1) was generated against synthetic peptide from humancPGES (amino acids 58–67) and cross-reacted with murine cPGES,but not with recombinant human mPGES. The rabbit polyclonalanti-mPGES-1 antibody (Catalog No. 160140, Lot No. 0426637-1)was generated against synthetic peptide from human mPGES-1(amino acids 59–75) and cross-reacted with rat mPGES-1, but notwith human cPGES. The rabbit polyclonal anti-mPGES-2 antibody(Catalog No. 160145, Lot No. 0407020-1) was generated againstsynthetic peptide from human mPGES-2 (amino acids 221–235)and cross-reacted with rat mPGES-2, but not with either humanmPGES-1 or cPGES.

Vectastain Elite ABC staining kit was purchased from VectorLaboratories (Burlingame, CA, USA). All other reagents including3,3′-diaminobenzidine tetrahydrochrolide (DAB) were of analyticalgrade.

2.2. Animal and tissue preparation

All procedures employed in this study were carried out fol-lowing the Guidelines of the Animal Care and Use Committee ofKitasato University School of Veterinary Medicine. Female rats ofWistar-Imamichi strain were kept in an air-conditioned room withcontrolled lighting. Rat chow and tapped water were available adlibitum. Animals used in this study were 3–5 months old and theirweights ranged from 250 to 400 g. Rats showing a regular 4-dayestrous cycle were mated with a fertile male in the pro-estrousevening. The day of fertilization, when vaginal smear showed theestrous stage and contained sperm, was designated as day 1 ofpregnancy (depicted as PRG1 in this paper). Duration of pregnancyin our colony was 23.0 ± 0.1 days (mean ± SEM, n = 13). The ratswere killed by cervical dislocation under light ether anesthesiaon PRG17, PRG19, PRG21, PRG23 (before parturition on PRG23),day 0 postpartum (PP0, after parturition) and PP3 (n = 3 or 4 ratsat each time point). For pre-parturient animals, uteri and ovarieswere first excised out from the body. Fetal membranes, fetusesand placentas were freed from gravid uteri and excluded. Uterinehorn (mainly myometrium and endometrium near decidua) andcervix were harvested. From post-parturient rats, uterine hornsand cervices harvested as well. Rat preovulatory ovaries followinggonadotropins treatment [39] were sampled and served as positivecontrol of PLA2G2A and COX-2, because immunoreactions for thesetwo antigens were generally modest as described later. All of thetissues were instantly fixed in Bouin’s solution for the subsequenthistological processing.

2.3. Immunohistochemistry

Tissues fixed in Bouin’s solution overnight were dehydratedand embedded in paraffin. Tissues were serially sectioned (5–6 �min thickness), deparaffinized, and examined immunochemically.

Antigen retrieval by the treatment with 20 �g/ml proteinase Kfor 30 min at room temperature was done for immunodetectionof COX-1, AKR1C7 and AKR1C11. Endogenous peroxidase wasquenched by pretreatment with 0.3% H2O2 in methanol for 30 min.

other Lipid Mediators 106 (2013) 23– 28 25

AP(((saNniat

3

fitPte

3p

pttws(Asfaf

s(tPm

FIP(aiC

Fig. 3. Ovarian stromal immunoreactivities for PGs biosynthetic enzymes. Ovarieson PRG19 were shown. Immunostaining of negative control IgG (B), PLA2G2A (C),PLA2G4A (A), PLA2G6A (D), COX-1 (E), COX-2 (F), AKR1C7 (G), AKR1C11 (H), AKR1C5

H. Satoh et al. / Prostaglandins &

ntibodies with final concentrations of their application were anti-LA2G2A (used at 1:200), anti-PLA2G4A (1:200), anti-PLA2G6A1:200), anti-COX-1 (1:200), anti-COX-2 (1:200), anti-AKR1C71:100), anti-AKR1C11 (1:100), anti-AKR1C5 (1:100), anti-cPGES1:200), anti-mPGES-1 (1:200) and anti-mPGES-2 (1:200). Serialections for each sample were incubated with these primaryntibodies at room temperature for 90 min or at 4 ◦C overnight.egative controls were performed with normal rabbit serum oron-specific mouse IgG. The antigen/antibody complex was visual-

zed as a brown color with the Vectastain ABC staining kit and DABs peroxidase substrate. Slides were counter-stained with hema-oxylin (blue).

. Results

All the tissues sampled from PRG17 through PP0 or PP3 wererst immunostained for each antigen and temporal alterations inhe immunoreactive intensity was evaluated semi-quantitatively.hotographic data are shown for each tissue harvested at theime point when the immunoreactivities for 11 different kinds ofnzymes were in general the most evident.

.1. Immunohistochemical localization of PGF2˛ and PGE2roducing enzymes in the ovary (Figs. 2 and 3)

Immunohistochemical localization of a set of PGF2� or PGE2roducing enzymes in the CL of pregnancy was most evident inhat obtained on PP0. No positive signal was present in the sec-ion stained with the non-specific IgG (Fig. 2B). Immunostainingith anti-PLA2G6A antibody demonstrated positive signals in most

teroidogenic cells with differential intensity (Fig. 2A). PLA2G4AFig. 2D) and COX-1 (Fig. 2E) were positive and signals for all ofKR1C7, AKR1C11 and AKR1C5 (Fig. 2G–I) were also evident. Thoseignals were uniformly present throughout the CL tissue. Signalsor PLA2G2A (Fig. 2C), COX-2 (Fig. 2F) and all PGESs (Fig. 2J–L) werelmost absent. There were trends for sustained immunoreactivitiesor PLA2s and PGFSs toward PP3 (data not shown).

Interestingly, there were massive populations of ovariantromal cells showing intense immunoreactivities for PLA2G4A

Fig. 3A), COX-1 (Fig. 3E), and AKR1C7 (Fig. 3G) with preferen-ial presence on the nuclei. Those cells were faintly positive forLA2G6A (Fig. 3D) and AKR1C11 (Fig. 3H). There were also stro-al cells intensely positive for mPGES-1 (Fig. 3K) but with different

ig. 2. CL immunoreactivities for PGs biosynthetic enzymes. CLs on PP0 were shown.mmunostaining (shown as dark brown) of negative control IgG (B), PLA2G2A (C),LA2G4A (D), PLA2G6A (A), COX-1 (E), COX-2 (F), AKR1C7 (G), AKR1C11 (H), AKR1C5I), cPGES (J), mPGES-1 (K) and mPGES-2 (L). (A) Most large luteal cells (whiterrow heads) and small luteal cells (black arrow heads) were positive for PLA2G6Ammunoreactions with different intensities. Scale bars in A and B (applicable to–L) = 25 �m.

(I), cPGES (J), mPGES-1 (K) and mPGES-2 (L). (A) Stromal cells below the ovariansurface epithelium presented marked nuclear reactions (black arrow heads) forPLA2G4A specifically on PRG19. Scale bars in A and B (applicable to C–L) = 12.5 �m.

cellular localization from those described above. Other enzyme iso-forms (Fig. 3C, F, I, J and L) were almost negative in these cells.The intensely positive reactions for PLA2G4A, COX-1, and AKR1C7were present at multiple stromal sites just beneath the ovarian sur-face epithelium but not in central stroma and the periphery of CLs.This result excluded out the possibility of the marginal effect inthe immunohistochemical reaction. These commonly intense reac-tions appeared specifically on PRG19, as no such staining behaviorwas found in samples of any other periods examined. The intenseimmunoreactions were confirmed to occur in every ovarian speci-men harvested from 4 different rats at this time period.

3.2. Immunohistochemical localization of PGF2˛ and PGE2producing enzymes in delivering uterine endometrium andmyometrium (Figs. 4 and 5)

Next, the uterus (Figs. 4 and 5) and cervix (Fig. 6) obtained onPP0 were also the most evident among samples from differenttime periods. Endometrial epithelial cells near the decidua wereobviously positive for PLA2G4A (Fig. 4D), COX-1 (Fig. 4A), AKR1C7(Fig. 4G), AKR1C11 (Fig. 4H) and AKR1C5 (Fig. 4I) immunoreactions.

In this tissue, PLA2G6A (Fig. 4D), PLA2G2A (Fig. 4C), and three PGESs(cPGES, Fig. 4J; mPGES-1, Fig. 4K; mPGES-2, Fig. 4L) were slightlypositive while COX-2 (Fig. 4G) was almost negative. A population of

Fig. 4. Uterine endometrial immunoreactivities for PGs biosynthetic enzymes. Uterion PP0 were shown. Immunostaining of negative control IgG (B), PLA2G2A (C),PLA2G4A (D), PLA2G6A (E), COX-1 (A), COX-2 (F), AKR1C7 (G), AKR1C11 (H), AKR1C5(I), cPGES (J), mPGES-1 (K) and mPGES-2 (L). (A) Endometrial epithelial cells (EE) nearthe decidual tissue presented showed evident reactions for COX-1. Scale bars in Aand B (applicable to C–L) = 50 �m.

26 H. Satoh et al. / Prostaglandins & other

Fig. 5. Uterine myometrial immunoreactivities for PGs biosynthetic enzymes. Uterion PP0 were shown. Immunostaining of negative control IgG (B), PLA2G2A (C),PLA2G4A (A), PLA2G6A (D), COX-1 (E), COX-2 (F), AKR1C7 (G), AKR1C11 (H), AKR1C5((P

mt

giP(f((

3p

siotfp(

Fc(APblt

I), cPGES (J), mPGES-1 (K) and mPGES-2 (L). (A) Longitudinal smooth muscle cellsLM) but not circular muscle cells (CM) showed moderate immunoreactions forLA2G4A. Scale bars in A and B (applicable to C–L) = 50 �m.

PGES-2-positive cells was found in the interstitial tissue beneathhe endometrium (Fig. 4L).

In uterine smooth muscle tissue, immunoreactive signals wereenerally evident in the longitudinal myometrium and modestn the circular layer. Myometrium was evidently positive forLA2G4A (Fig. 5A), PLA2G6A (Fig. 5D), COX-1 (Fig. 5E), and all PGFSsFig. 5G–I). The signal in myometrial tissue was the most intenseor 9-KR (Fig. 5I). The tissues were slightly positive for PLA2G2AFig. 5C) and mPGES-1 (Fig. 5K) and almost negative for COX-2Fig. 5F), cPGES (Fig. 5J) and mPGES-2 (Fig. 5L).

.3. Immunohistochemical localization of PGF2˛ and PGE2roducing enzymes in delivering uterine cervix (Fig. 6)

In the uterine cervix of mothers just after parturition (PP0), theite of the most intense immunoreactive signals was the basal lam-na of the endometrium, which was composed of two populationsf cell types (Fig. 6). One was stratified squamous cells positive inhe cytoplasm. The other was mono- and bi-layer columnar cells,

acing the connective tissue, with more intense signals in the cyto-lasm. Enzymes immunoreactive in this region included PLA2G4AFig. 6A), PLA2G6A (Fig. 6D), COX-1 (Fig. 6E), AKR1C7 (Fig. 6G),

ig. 6. Uterine cervical immunoreactivities for PGs biosynthetic enzymes. Uterineervices on PP0 were shown. Immunostaining of negative control IgG (B), PLA2G2AC), PLA2G4A (A), PLA2G6A (D), COX-1 (E), COX-2 (F), AKR1C7 (G), AKR1C11 (H),KR1C5 (I), cPGES (J), mPGES-1 (K) and mPGES-2 (L). (A) Immunoreactions forLA2G4A were most evident in two distinct cell types, separated by the dashedlack line, in the basal lamina of the cervical endometrium (CE), the surface epithe-

ial cells of endometrium (black arrow heads) and scattered in stromal connectiveissue (SC). Scale bars in A and B (applicable to C–L) = 25 �m.

Lipid Mediators 106 (2013) 23– 28

AKR1C11 (Fig. 6H), AKR1C5 (Fig. 6I). The signal intensities seemedmaximal around parturition. Signals of mPGES-1 (Fig. 6K), cPGES(Fig. 6J), and mPGES-2 (Fig. 6L) were also positive. PLA2G2A (Fig. 6C)and COX-2 (Fig. 6F) were negative. Secondly, cervical connectivetissues had positive signals for PLA2G4A, PLA2G6A, COX-1, AKR1C7,AKR1C5, cPGES and mPGES-1 sparsely. The surface epithelial cell ofthe endometrium was also positive for PLA2G4A, PLA2G6A, COX-1and AKR1C7.

4. Discussion

We still know limitedly about the amazingly complicated mech-anism of mammalian parturition, because this process involvesnumerous interactions at all of organ, tissue, cell, and sub-stance levels and a great difference among animal species [1,2].This immunohistochemical study targeting a total of 11 differentenzymes was a trial to define, as comprehensively as possible, theovarian and uterine sources of the two primary PGs in rat spon-taneous parturition. We here provide no biochemical evidence onactivities of the enzymes examined and 15-hydroxy-PG dehydro-genase and the contents of two PGs at the cellular level. Therefore,our findings do not indicate relative contribution of isoforms to the“net” activity of PGs synthesis. However, we value that the currentstrategy is useful to define the PG source at cellular level and iso-forms of PG synthesizing enzymes at the protein level, providingthe basic information to further implicate PGs role in regulation ofparturition.

Our current findings are largely summarized as follows. Firstly,many cell types that had previously been considered with PGsynthetic potential, representatively epithelial cells of uterineendometrium and luteal steroidogenic cell, are demonstrated tohave a full set of immunoreactive enzymes for three metabolicsteps. Secondly, multiple enzyme isoforms are found present andlikely employed for each step of PG synthesis. Relating to this point,the frequent existence of AKR1C5 at several tissues suggests theregulated conversion of PG ligands and thus modulation of biologiceffects. Thirdly, ovarian stromal cells and basal lamina of cervi-cal endometrium are suggested to be novel sites of potential PGsproduction.

PGF2� binding to its specific receptor (FP) on luteal and vas-cular cells in CL causes functional regression, an initial criticalevent of parturition cascade in animal species whose gestationis maintained primarily by progesterone of CL origin [3,4]. Theuterus, especially endometrium, is widely accepted as a primarysource of PGF2� in rodents and ruminants [4,5] and in gravidstates placenta and fetal membranes may be additional contrib-utors of the luteolysin. Endometrial PG biosynthetic pathway hasbeen partially characterized in ewes [42,43] and cows [44–48]. CLis an additional or supplementary source of PGF2� in almost allof the non-primate species examined and very likely an alterna-tive in primates including human [4,5]. Biosynthetic pathway ofPG has also been incompletely demonstrated in CL of cow [49],pig [50] and rhesus monkey [51]. We find spatial association ofapparent immunoreactions of PLA2 (G2A and G4A)/COX-1/AKR(1C7, 1C11, and 1C5) in uterine endometrium and PLA2 (G4A andG6A)/COX-1/AKR (1C7, 1C11, and 1C5) in luteal cells on the deliv-ering day when PGF2� level in uterine and CL tissues is markedlyincreased [8,13]. Gene knockout mice studies have demonstratedan indispensable role for PLA2G4A/COX-1/FP pathway in induc-ing progesterone withdrawal, but not downstream events, leadingfinally to timely delivery of intact neonates [1–3]. Our histologi-cal data of good association of PLA2G4A and COX-1 supports this

concept and further reveals the candidate(s) of PGFS linking to theenzymes. The dual source of the luteolysin is also demonstratedhere and placenta and fetal membrane [25,28,34] as well as fetuses[52] are shown to be potential sites of increased PGs during rodent

other L

pp

bissrtr(t

bilpfpblirtieAlriietihePo

onocaiptodtfhoi

aotPPii

lri

[

[

[

[

[

[

[

[

H. Satoh et al. / Prostaglandins &

arturition. Further investigation on these gestational tissues is inrogress in our laboratory.

Uterine myometrial quiescence and activation is regulated byoth relaxatory and contractile PGs signaling [1,2]. The myometrial

mmunoreactivities for a set of PG synthetic enzymes are almostimilar to those in the endometrium. It is additionally notable thatubstantial existence of AKR1C5 with mPGES-1 and cPGES suggestsegulated alteration of the ligand (PGE2 to PGF2�) balance. Consis-ent with this, the rat uterine myometrium also has a shift at theeceptor level involving decreased expression of relaxatory PGE2EP) receptor EP2 and increased expression in contractory FP aterm [30,53].

Among 3 or 4 steps of PG biosynthetic pathway, COX step haseen most intensively studied and is attributed to as having phys-

ologic and pathophysiologic implications of normal and pretermabor [1,2]. COX-1 is assumed a constitutively expressed isoformlaying “housekeeping” roles, while COX-2 is assumed an iso-orm that is induced and functions under dramatic changes inhysiologic and pathologic states [5]. Contrary to this generallyelieved concept, a number of reports documented somewhat ear-

ier increase in COX-1 expression with unaltered COX-2 expressionn rodent uterine tissues at term [14,25,28,30,31]. Some otherseported increased expression of both COXs [19,24,27]. From func-ional studies using gene knockout mice and COX isoform-selectivenhibitors, generation of luteolytic signal PGF2� is attributed almostxclusively to COX-1 of CL and/or uterine endometrium [24,54,55].fter natural progesterone withdrawal and during induced preterm

abor, generation of contracting PGs seems complemented oreplaced dominantly but never completely by COX-2 that isnduced in uterine and cervical myometria [24,26,30,56]. COX-1mmunoreactivity is currently obvious in ovarian cells and uterinendometrium/myometrium and cervix, while COX-2 immunoreac-ivity is generally modest compared to COX-1 signal and much lessntense than that of COX-2 induced in preovulatory follicles afterCG challenge [39]. Taken together with these findings, anothervidence that COX-1 localization is very frequently associated withLA2G4A, the most efficient AA supplier [5,7], favors the critical rolef COX-1 in normal parturition in rodents.

Cervical tissue ripening and softening in accordance with labornset occurs through PGE2-mediated matrix degradation [57]. Con-ective stromal tissue is found to possess any one or two isoformsf each of PLA2, COX-1, and PGFS or PGES and thus supposed toontribute PGs directly in situ. We find more massive and associ-ted existence of PLA2 (G4A and G6A)/COX-1/PGFSs and mPGESsn basal lamina of the cervical endometrium. This finding unifiesrevious demonstration of PLA2G4A [13] and both COXs [19] inhe very cell layer and shows the full set and additional elementsf PGs synthetic machinery. Hinton et al. [58] have provided evi-ence that all of EP1-4 localized more broadly but most abundantlyo basal lamina of rat cervical endothelium at term. As the principalunction of this tissue layer might be contribution to endometrialyperplasia associated with cervical ripening, abundant presencef PGs synthesizing enzymes specific to this layer may have a rolen proliferation of the constituent cells.

Finally but interestingly, ovarian stroma is revealed to possess set of PGF2� producing machinery, the immunostaining behaviorf which is highly specific spatially and temporally. Fibroblas-ic, non-steroidogenic cells just beneath ovarian epithelia haveLA2G4A/COX-1/AKR1C7 system exclusively on PRG19. LutealGF2� content begins to increase on PRG21 [8]. This stromal PGF2�

s very likely to have different temporal change. Exploring its phys-ological significance needs intensive mechanistic studies.

In conclusion, the current findings should provide the histo-ogical basis of cell types and biosynthetic pathway of PGs in aat model of spontaneous delivery. A set of PLA2G4A and COX-1s likely the general pathway for generating PGH2 in gestational

[

ipid Mediators 106 (2013) 23– 28 27

tissues and associated with various isoforms of downstream PGFSand/or PGES. Taken with the others’ findings from Western and/orRT-PCR analyses, we suggest candidate isoforms for three steps ofPG synthesis in the tissues previously defined to make PGs such asuterine endometrium/myometrium and CL. We also reveal ovarianstroma on PRG19 and basal lamina of cervical endothelium aroundparturition as the probable potential site of PGF2� and PGE2 pro-duction whose biologic significances are unknown. The findingsextend our knowledge of PGs synthetic mechanism and are thebasic information to next explore the multiple implications of PGsin the parturition cascade.

Acknowledgments

We thank H. Yoshio for experimental help, T. Yonezawa for valu-able discussion, and M. Nakata for help in manuscript preparation.We value many articles of early original studies and the impor-tant ones with non-rodent models that we have not referred toin the text. This work was supported by Grants-in-Aid for Scien-tific Research (No. 23580392) from the Japanese Society for thePromotion of Science and the Kieikai Research Foundation (to S.K.).

References

[1] Olson DM, Ammann C. Role of the prostaglandins in labor and prostaglandinreceptor inhibitors in the prevention of preterm labor. Front Biosci2007;12:1329–43.

[2] Khan AH, Carson RJ, Nelson SM. Prostaglandins in labor – a translationalapproach. Front Biosci 2008;13:5794–809.

[3] Ratajczak CK, Muglia LJ. Insights into parturition biology from geneticallyaltered mice. Pediatr Res 2008;64:581–9.

[4] Niswender GD, Juengel JL, Silva PJ, Rollyson MK, McIntush EW. Mechanismscontrolling the function and life span of the corpus luteum. Physiol Rev2000;80:1–29.

[5] Funk CD, Song W-C, FitzGerald GA. Prostaglandins and other lipid mediatorsin reproductive medicine. In: Strauss JF, Barbieri RL, editors. Yen and Jaffereproductive endocrinology. San Diego: Elsevier; 2009. p. 121–37.

[6] Lindstrom T, Bennett P. Transcriptional regulation of genes for enzymes of theprostaglandin biosynthetic pathway. Prostaglandins Leukot Essent Fatty Acids2004;70:115–35.

[7] Lappas M, Rice GE. Phospholipase A2 isoenzymes in pregnancy and parturition.Prostaglandin Leukotr Essent Fatty Acids 2004;70:87–100.

[8] Kurusu S, Kaizo K, Ibashi M, Kawaminami M, Hashimoto I. Luteal phospholipaseA2 activity increases during functional and structural luteolysis in pregnantrats. FEBS Lett 1999;454:225–8.

[9] Arend A, Masso R, Masso M, Selstam G. Electron microscope immunocyto-chemical localization of cyclooxygenase-1 and -2 in pseudopregnant rat corpusluteum during luteolysis. Prostaglandins Other Lipid Mediat 2004;74:1–10.

10] Zerani M, Dall’Aglio C, Maranesi M, et al. Intraluteal regulation of prostaglandinF2�-induced prostaglandin biosynthesis in pseudopregnant rabbits. Reproduc-tion 2007;133:1005–16.

11] Taniguchi K, Matsuoka A, Kizuka F, et al. Prostaglandin F2� (PGF2�) stimulatesPTGS2 expression and PGF2� synthesis through NFKB activation via reactiveoxygen species in the corpus luteum of pseudopregnant rats. Reproduction2010;140:885–92.

12] Farina MG, Billi S, Leguizamón G, et al. Secretory and cytosolic phospholipase A2

activities and expression are regulated by oxytocin and estradiol during labor.Reproduction 2007;134:355–64.

13] Kurusu S, Ishii S, Kawaminami M, Hashimoto I. Enhanced activity in cytosolicphospholipase A2 in the rat uterus and cervix around parturition. J Reprod Dev2002;48:65–73.

14] Winchester SK, Imamura T, Gross GA, et al. Coordinate regulation ofprostaglandin metabolism for induction of parturition in mice. Endocrinology2002;143:2593–8.

15] Rajabi MR, Cybulsky AV. Phospholipase A2 activity is increased in guinea piguterine cervix in late pregnancy and at parturition. Am J Physiol EndocrinolMetabol 1995;269:E940–7.

16] Sato TA, Gupta DK, Keelan JA, Marvin KW, Mitchell MD. Cytosolic phospholipaseA2 and 15-hydroxyprostaglandin dehydrogenase mRNA expression in murineuterine and gestational tissues during late pregnancy. Prostaglandins LeukotEssent Fatty Acids 2001;64:247–51.

17] Brant K, Guan W, Tithof P, Caruso RL. Gestation age-related increase in50-kDa rat uterine calcium-independent phospholipase A2 expression influ-

ences uterine sensitivity to polychlorinated biphenyl stimulation. Biol Reprod2006;74:874–80.

18] Myatt L, Langdon G, Brockman DE. Identification and changes in concentra-tions of prostaglandin H synthase isoforms in rat myometrium at parturition.Prostaglandins 1994;48:285–96.

2 other

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

8 H. Satoh et al. / Prostaglandins &

19] Dong Y-L, Rettori V, Gangula PRR, Fang L, Yallampalli C. Differential expressionof cyclooxygenase-1 and -2 proteins in rat uterus and cervix during the estrouscycle, pregnancy, labor and in myometrial cells. Prostaglandins 1996;52:13–34.

20] Arslan A, Zingg HH. Regulation of COX-2 gene expression in rat uterus in vivoand in vitro. Prostaglandins 1996;52:463–81.

21] Fang L, Chatterjee S, Dong YL, Gangula PR, Yallampalli C. Immunohistochemicallocalization of constitutive and inducible cyclo-oxygenases in rat uterus duringthe oestrous cycle and pregnancy. Histochem J 1998;30:383–91.

22] Gross GA, Imamura T, Luedke C, et al. Opposing actions of prostaglandinsand oxytocin determine the onset of murine labor. Proc Natl Acad Sci USA1998;95:11875–9.

23] Cook JL, Zaragoza DE, Sung DH, Olson DM. Expression of myometrialactivation and stimulation genes in a mouse model of preterm labor: myome-trial activation, stimulation, and preterm labor. Endocrinology 2000;141:1718–28.

24] Tsuboi K, Sugimoto Y, Iwane A, Yamamoto K, Yamamoto S, Ichikawa A. Uterineexpression of prostaglandin H2 synthase in late pregnancy and parturition inprostaglandin F receptor-deficient mice. Endocrinology 2000;141:315–24.

25] Gupta DK, Sato TA, Keelan JA, Marvin KW, Mitchell MD. Expression ofprostaglandin H synthase-1 and -2 in murine intrauterine and gestationaltissues from mid pregnancy until term. Prostaglandins Other Lipid Mediat2001;66:17–25.

26] Tsuboi K, Iwane A, Nakazawa S, Sugimoto Y, Ichikawa A. Role of prostaglandinH2 synthase 2 in murine parturition: study on ovariectomy-induced parturitionin prostaglandin F receptor-deficient mice. Biol Reprod 2003;69:195–201.

27] Farina M, Ribeiro ML, Weissmann C, et al. Biosynthesis and catabolism ofprostaglandin F2� are controlled by progesterone in the rat uterus during preg-nancy. J Steroid Biochem Mol Biol 2004;91:211–8.

28] Welsh T, Mitchell CM, Walters WA, Mesiano S, Zakar T. Prostaglandin Hsynthase-1 and -2 expression in guinea pig gestational tissues during latepregnancy and parturition. J Physiol 2005;569:903–12.

29] Marx SG, Wentz MJ, MacKay LB, et al. Effects of progesterone on iNOS, COX-2,and collagen expression in the cervix. J Histochem Cytochem 2006;54:623–39.

30] Arthur P, Taggart MJ, Zielnik B, Wong S, Mitchell BF. Relationship between geneexpression and function of uterotonic systems in the rat during gestation, uter-ine activation and both term and preterm labour. J Physiol 2008;586:6063–76.

31] St-Louis I, Singh M, Brasseur K, Leblanc V, Parent S, Asselin E. Expression ofCOX-1 and COX-2 in the endometrium of cyclic, pregnant and in a modelof pseudopregnant rats and their regulation by sex steroids. Reprod BiolEndocrinol 2010;8:103.

32] Wintergalen N, Thole HH, Galla H-J. Schlegel W. Prostaglandin E2 9-reductasefrom corpus luteum of pseudopregnant rabbit is a member of the aldo-ketoreductase superfamily featuring 20�-hydroxysteroid dehydrogenase activity.Eur J Biochem 1995;234:264–70.

33] Unezaki S, Sugatani J, Masu Y, Watanabe K, Ito S. Characterization ofprostaglandin F2� production in pregnant and cycling mice. Biol Reprod1996;55:889–94.

34] Kubota K, Kubota T, Kamei D, et al. Changes in prostaglandin E synthases(PGESs) in microsomal PGES-1 knockout mice in a preterm delivery model.J Endocrinol 2005;187:339–45.

35] Sun T, Deng W-B, Diao H-L, et al. Differential expression and regulation ofprostaglandin E synthases in the mouse ovary during sexual maturation andluteal development. J Endocrinol 2006;189:89–101.

36] Phillips RJ, Al-Zamil H, Hunt LP, Fortier MA, López Bernal A. Genes forprostaglandin synthesis, transport and inactivation are differentially expressedin human uterine tissues, and the prostaglandin F synthase AKR1B1 isinduced in myometrial cells by inflammatory cytokines. Mol Human Reprod2011;17:1–13.

37] Palliser HK, Hirst JJ, Rice GE, et al. Labor-associated regulation of prostaglandinE and F synthesis and action in the ovine amnion and cervix. J Soc GynecolInvest 2006;13:19–24.

38] Murakami M, Kudo I, Natori Y, Inoue K. Immunochemical detection of ‘platelettype’ phospholipase A2 in the rat. Biochim Biophys Acta 1990;1043:34–42.

[

Lipid Mediators 106 (2013) 23– 28

39] Kurusu S, Sapirstein A, Bonventre JV. Group IVA phospholipase A2 optimizesovulation and fertilization in rodents through induction of and metabolic cou-pling with prostaglandin endoperoxide synthase-2. FASEB J 2012;26:3800–10.

40] Suzuki-Yamamoto T, Toida K, Tsuruo Y, Watanabe K, Ishimura K. Immunocy-tochemical localization of lung-type prostaglandin F synthase in the rat spinalcord. Brain Res 2000;877:391–5.

41] Suzuki-Yamamoto T, Toida K, Watanabe K, Ishimura K. Immunocytochemi-cal localization of prostaglandin F synthase II in the rat spinal cord. Brain Res2003;969:27–35.

42] Zhang Q, Wu WX, Brenna JT, Nathanielsz PW. The expression of cytosolicphospholipase A2 and prostaglandin endoperoxide synthase in ovine mater-nal uterine and fetal tissues during late gestation and labor. Endocrinology1996;137:4010–7.

43] Wu WX, Ma XH, Zhang Q, Buchwalder L, Nathanielsz PW. Regulation ofprostaglandin endoperoxide H synthase 1 and 2 by estradiol and progesteronein non-pregnant ovine myometirum and endometrium in vivo. Endocrinology1997;138:4005–12.

44] Tithof PK, Roberts MP, Guan W, Elgayyar M, Godkin JD. Distinct phospholipaseA2 enzymes regulate prostaglandin E2 and F2� production by bovine endome-trial epithelial cells. Reprod Biol Endocrinol 2007;5:16.

45] Xiao CW, Liu JM, Sirois J, Goff AK. Regulation of cyclooxygenase-2 andprostaglandin F synthase gene expression by steroid hormones and interferon-tau in bovine endometrial cells. Endocrinology 1998;139:2293–9.

46] Asselin E, Fortier MA. Detection and regulation of the messenger for a putativebovine endometrial 9-keto-prostaglandin E2 reductase: effect of oxytocin andinterferon-tau. Biol Reprod 2000;62:125–31.

47] Madore E, Harvey N, Parent J, Chapdelaine P, Arosh JA, Fortier MA. An aldosereductase with 20�-hydroxysteroid dehydrogenase activity is most likely theenzyme responsible for the production of prostaglandin F2� in the bovineendometrium. J Biol Chem 2003;278:11205–12.

48] Palliser HK, Ooi GT, Hirst JJ, et al. Changes in the expression of prostaglandinE and F synthases at induced and spontaneous labour onset in the sheep. JEndocrinol 2004;180:469–77.

49] Arosh JA, Banu SK, Chapdelaine P, Madore E, Sirois J, Fortier MA. Prostaglandinbiosynthesis, transport, and signaling in corpus luteum: a basis for autoregu-lation of luteal function. Endocrinology 2004;145:2551–60.

50] Waclawik A, Kaczmarek MM, Kowalczyk AE, Bogacki M, Ziecik AJ. Expression ofprostaglandin synthesis pathway enzymes in the porcine corpus luteum duringthe oestrous cycle and early pregnancy. Theriogenology 2008;70:145–52.

51] Bogan RL, Murphy MJ, Stouffer RL, Hennebold JD. Prostaglandin synthesis,metabolism, and signaling potential in the rhesus macaque corpus luteumthroughout the luteal phase of the menstrual cycle. Endocrinology 2008;149:5861–71.

52] Reese J, Paria BC, Brown N, Zhao X, Morrow JD, Dey SK. Coodinated regula-tion of fetal and maternal prostaglandins directs successful birth and postnataladaptaion in the mouse. Proc Natl Acad Sci USA 2000;97:9759–64.

53] Brodt-Eppley J, Myatt L. Changes in expression of contractile FP and relaxatoryEP2 receptors in pregnant rat myometrium during late gestation, at labor, andpostpartum. Biol Reprod 1998;59:878–83.

54] Loftin CD, Trivedi DB, Langenbach R. Cyclooxygenase-1-selective inhibitionprolongs gestation in mice without adverse effects on the ductus arteriosus.J Clin Invest 2002;110:549–57.

55] Yu Y, Cheng Y, Fan J, et al. Differential impact of prostaglandin H Synthase 1knockdown on platelet and parturition. J Clin Invest 2005;115:986–95.

56] Gross G, Imamura T, Vogt SK, et al. Inhibition of cyclooxygenase-2 preventsinflammation-mediated preterm labor in the mouse. Am J Physiol Regul IntegrCom Physiol 2000;278:R1415–23.

57] Chien EK, Ji H, Feltovich H, Clark K. Expression of matrix metalloproteinase-3

in the rat cervix during pregnancy and in response to prostaglandin E2. Am JObstet Gynecol 2005;192:309–17.

58] Hinton AC, Grigsby PL, Pitzer BA, et al. Hormonal regulation of prostaglandinE2 receptors: localization and expression in rat cervical tissue. Reprod Sci2010;17:136–46.