Prostaglandin E2 (EP) Receptors Mediate PGE2-Specific Events in Ovulation and Luteinization Within Primate Ovarian Follicles

Abstract

Prostaglandin E2 (PGE2) is a key mediator of ovulation. All 4 PGE2 receptors (EP receptors) are expressed in the primate follicle, but the specific role of each EP receptor in ovulatory events is poorly understood. To examine the ovulatory events mediated via these EP receptors, preovulatory monkey follicles were injected with vehicle, the PG synthesis inhibitor indomethacin, or indomethacin plus PGE2. An ovulatory dose of human chorionic gonadotropin was administered; the injected ovary was collected 48 hours later and serially sectioned. Vehicle-injected follicles showed normal ovulatory events, including follicle rupture, absence of an oocyte, and thickening of the granulosa cell layer. Indomethacin-injected follicles did not rupture and contained oocytes surrounded by unexpanded cumulus; granulosa cell hypertrophy did not occur. Follicles injected with indomethacin plus PGE2 were similar to vehicle-injected ovaries, indicating that PGE2 restored the ovulatory changes inhibited by indomethacin. Additional follicles were injected with indomethacin plus an agonist for each EP receptor. EP1, EP2, and EP4 agonists each promoted aspects of follicle rupture, but no single EP agonist recapitulated normal follicle rupture as seen in follicles injected with either vehicle or indomethacin plus PGE2. Although EP4 agonist-injected follicles contained oocytes in unexpanded cumulus, the absence of oocytes in EP1 agonist- and EP2 agonist-injected follicles suggests that these EP receptors promote cumulus expansion. Surprisingly, the EP3 agonist did not stimulate any of these ovulatory changes, despite the high level of EP3 receptor expression in the monkey follicle. Therefore, agonists and antagonists selective for EP1 and EP2 receptors hold the most promise for control of ovulatory events in women.

Prostaglandin E2 (PGE2) has been recognized as a key prostaglandin in mediating ovulatory processes (1). In primates as in other mammalian species, the midcycle surge of LH stimulates expression of the cyclooxygenase (COX)-2 (gene name PTGS2) enzyme and PGE2 production by granulosa cells, with follicular PGE2 concentrations reaching peak levels just before ovulation (2–5). PGE2 acts within the follicle to regulate ovulatory processes, including follicle basement membrane breakdown, luteinization of the follicle wall, follicle rupture, and oocyte release from the ovary (6–8).

PGE2 alters cell functions by binding to 4 distinct PGE2 receptors (EP receptors): PTGER1, PTGER2, PTGER3, and PTGER4 (commonly referred to as EP1, EP2, EP3, and EP4) (9). Individual EP receptors have been linked to specific ovulatory events. Mice lacking EP2 expression demonstrated severe deficiencies in cumulus expansion, decreased rate of ovulation, and decreased fertility (10, 11). Although EP1- and EP3-deficient female mice were fertile (12), fertility of EP4-deficient mice could not be determined, because these animals died shortly after birth (13, 14). Each EP receptor may contribute to ovulatory events in large animal species. For example, the expression level of EP3 mRNA in bovine cumulus cells has been correlated with the quality of the cumulus-oocyte complex (15). It has also been suggested that the EP3 receptor participates in luteinization of both bovine and monkey follicles (16, 17). All 4 EP receptors are expressed by granulosa cells of monkey follicles (18), and EP1, EP2, and EP3 were shown to regulate components of the plasminogen activator proteolytic pathway in monkey granulosa cells in vitro (19). To date, the role of individual EP receptors in regulating ovulatory events has not been investigated in a large animal species in vivo.

Macaque species are widely used to examine ovarian function with direct applicability to ovulatory events in women. Female adult monkeys have regular menstrual cycles averaging 28 days in length, with cyclical patterns of gonadotropin and ovarian steroid hormones and changing ovarian structures very similar to those of women (20). In the present study, a macaque model of controlled ovulation (21) was used in conjunction with injection of the preovulatory follicle (6) to elucidate the role of each EP receptor in ovulatory events. The dominant follicle was injected with a prostaglandin synthesis inhibitor in combination with an agonist specific for a single EP receptor, and then an ovulatory dose of human chorionic gonadotropin (hCG) was administered. Injected ovaries were removed 48 hours after hCG and assessed for events associated with ovulation and luteinization. The results suggest that EP1 and EP2, with possible support from EP4, mediate PGE2-stimulated ovulatory events in the primate.

Materials and Methods

Animal protocols

All animal protocols and experiments were approved by the Eastern Virginia Medical School (EVMS) Animal Care and Use Committee, and studies were conducted in accordance with the National Institutes of Health's Guide for the Care and Use of Laboratory Animals. Animal husbandry and sample collections were performed as previously described (22). Menstrual cycles of adult, female cynomolgus macaques (Macaca fascicularis) were monitored regularly. Blood samples were obtained by femoral venipuncture after ketamine chemical restraint (5–10 mg/kg body weight). Aseptic surgeries were performed in a dedicated surgical suite by laparotomy under isofluorane anesthesia.

Controlled ovulation model

A controlled ovulation model was used to obtain ovaries, each with a single, naturally selected preovulatory follicle (21). Beginning within 5–10 days of initiation of menstruation, serum estradiol and progesterone levels were assessed daily using the Immulite 1000 immunoassay system (Siemens Medical Diagnostics Solutions). When estradiol levels surpassed 150 pg/mL, the GnRH antagonist acyline (60 μg/kg per day; National Institute for Child Health and Human Development) was administered for 2 days to prevent an endogenous ovulatory LH surge. Concomitantly, recombinant human (rh)FSH (60 IU daily; Merck & Co) and rhLH (60 IU daily; Serono Reproductive Biology Institute) were administered for 2 days to maintain healthy growth of the preovulatory follicle. On the next day, a surgical procedure was performed as previously described (6) to inject the preovulatory follicle (Supplemental Figure 1O, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org). Immediately after follicle injection, animals received 1000-IU rhCG (Serono) to initiate ovulatory events. Injected ovaries were removed 48 hours after hCG administration. Because ovulation typically occurs about 40 hours after the endogenous LH surge in natural menstrual cycle, this time frame correlates with postovulatory and luteinization events (23). The ovarian surface was observed for evidence of an ovulatory stigma at the time of ovary removal.

The strategy for achieving a known concentration of vehicle and test compounds in monkey follicles was previously described (6). Previously, we identified a follicular concentration of the general COX inhibitor indomethacin (Indo, 10⁻⁵M), which prevented ovulation (6). We also determined that ovulatory events were restored with a concentration of PGE2 (10⁻⁶M) similar to that measured in monkey follicular fluid just before ovulation (2, 6). Concentrations of EP agonists were selected based on in vitro studies (data not shown). Briefly, luteinizing granulosa cells were treated in vitro with a single EP agonist at concentrations spanning at least 3 log dilutions. For the EP1 agonist (17-pheny-trinor-PGE2), cells were assessed for changes in intracellular calcium; media cAMP was assessed after treatment with agonists for the EP2 (butaprost), EP3 (sulprostone), and EP4 (PGE1-alcohol) receptors as previously described (24). The lowest concentration of each agonist having maximal effect on second messenger levels was used for intrafollicular injection. For these preliminary studies, luteinizing granulosa cells were obtained from healthy young women undergoing ovarian stimulation and oocyte retrieval as oocyte donors at the Jones Institute for Reproductive Medicine, EVMS. The Institutional Review Board at EVMS determined that this use of discarded human cells does not constitute human subjects research as defined by 45 CFR 46.102(f). All EP receptor-selective agonists were injected in follicles to achieve a final concentration of 10⁻⁵M. Final concentration of vehicle in all follicles was 0.1% dimethyl sulfoxide in PBS. Indomethacin was obtained from Sigma-Aldrich. PGE2 and all EP agonists were obtained from Cayman Chemical.

Histology

Whole ovaries were fixed in 10% formalin for 24 hours, followed by incubation in PBS for 24–48 hours. Tissues were paraffin embedded and oriented to yield sections which included both the follicle apex and follicle wall opposite the apex at the maximal follicle diameter in order to ensure optimal view of the follicle apex. Each ovary was serially sectioned at 5 μm, and every section was retained in order. Every fifth section was deparaffinized in xylene baths, rehydrated through a series of ethanol washes, stained with 25% hematoxylin and 50% eosin (both from Sigma), and permanently coverslipped. Additional slides were stained with Van Gieson's solution, consisting of 0.1% acid fuchsin (Sigma) in picric acid (Sigma) to visualize collagen fibers (25). Stained sections were imaged using an Olympus microscope with DP70 digital camera system and associated software (Olympus America, Inc).

Histologic evaluation of each ovary was performed by at least 2 independent observers. For each ovary, each observer noted the status of the mural granulosa cells, cumulus expansion, oocyte of the ovulatory follicle, and rupture site(s). Granulosa cell hypertrophy and size of rupture sites were measured using ImageJ software (NIH). For each follicle, mural granulosa cell layer width was measured from basement membrane to antrum and perpendicular to the basement membrane in 4 locations opposite the follicle apex; these measurements were averaged for each ovary. The size of each rupture site was quantified by measuring the width on the section with the largest rupture site, counting the number of 5 μm sections where the rupture site was present, and using these measurements to calculate the area of an oval (mm²).

Immunohistochemistry

Ovarian sections were subjected to antigen retrieval by microwaving for a total of 7 minutes (2 min on 1000 W, 5 min on 300 W) in 0.01M sodium citrate buffer (pH 6.0), followed by incubation in 2% hydrogen peroxide (Fisher Scientific) in methanol to block endogenous peroxidase activity. To reduce nonspecific staining, 5% nonimmune goat or horse serum (Vector Laboratories) in PBS containing 0.1% Triton X-100 was applied for 1 hour. Sections were washed and then incubated with a primary antibody generated against 3β-hydroxysteroid dehydrogenase (3βHSD) (polyclonal rabbit antihuman antibody; 1:2000 dilution in 5% nonimmune goat serum; antibody of Dr Ian Mason) or aromatase (monoclonal mouse antihuman antibody; 1:50 dilution in 5% nonimmune horse serum; AbD Serotec) for 1 hour at room temperature. Thereafter, sections were incubated with secondary antirabbit or antimouse antibody labeled with horseradish peroxidase for 1 hour (Vectastain ABC kit; Vector Laboratories). 3,3′-Diaminobenzidine (Vector Laboratories) was used to detect peroxidase. Sections were counterstained with hematoxylin before dehydration and coverslipping. Exclusion of the primary antibody served as a negative control (Supplemental Figure 1, I and N).

Statistical analysis

Data were assessed for heterogeneity of variance by Bartlett's test. Data were log transformed when Bartlett's test yielded P < .05; log-transformed data were subjected to Bartlett's test to confirm that P > .05. All datasets were assessed by ANOVA with no repeated measures, followed by Duncan's multiple range test (StatPak version 4.12 software; Northwest Analytical). Significance was assumed at P < .05. Data are expressed as mean ± SEM; n = 3–4 animals/treatment group, except where noted.

Results

Luteinization

To determine whether luteinization of the primate ovulatory follicle requires PGE2, morphological characteristics of luteinization were assessed. Each vehicle-injected follicle had a thickened granulosa cell layer, with enlarged granulosa cells (Figure 1A). Tracts of stroma containing connective tissue and cells, including theca cells, extended towards the follicle antrum. Remodeling of the connective tissue of the stroma was visualized by histochemical staining for collagen, which both surrounded vehicle-injected follicles and infiltrated into the granulosa cell layer. In contrast, the granulosa cell layer of Indo-injected follicles was much thinner than that observed in vehicle-injected follicles and consisted of small, tightly packed granulosa cells (Figure 1B). Indo-injected follicles were surrounded with collagen-rich stroma but lacked invaginations into the granulosa cell layer. Structural luteinization of Indo+PGE2-injected follicles (Figure 1C) was comparable with that observed in vehicle-injected follicles, with enlarged granulosa cells, a thickened granulosa cell layer, and stromal invaginations staining strongly for collagen. Granulosa cell hypertrophy was assessed quantitatively by measurement of the width of the granulosa cell layer directly opposite the apex of the follicle (Figure 2). Indomethacin prevented thickening of the granulosa cell layer, whereas PGE2 restored granulosa cell layer thickness to the level measured in vehicle-injected follicles. Overall, these results demonstrate that these specific aspects of luteinization are dependent on PGE2.

Figure 1.

Luteinization of injected follicles. Collagenous connective tissue (pink/red) was visualized in follicles injected with vehicle (A), Indo (B), and Indo+PGE2 (C). Immunohistochemical detection (brown) of 3βHSD (D–F) and aromatase (G–I) in the cells of follicles injected with vehicle (D and G), Indo (E and H), and Indo+PGE2 (F and I); hematoxylin counterstain is blue. All images show a portion of the follicle wall in the same orientation, with stroma in the lower left, granulosa cells (gc) central, and the antrum in the upper right portion of each image (see A and B). Arrows indicate cells expressing aromatase (G–I). A–C are at the same magnification (scale bar, 100 μm). D–I are at the same magnification (scale bar, 50 μm).

Figure 2.

Granulosa cell hypertrophy. A, Example of replicate measurements (straight lines) of granulosa cell layer thickness in a vehicle-injected follicle. Scale bar, 100 μm. B, Granulosa layer thickness of the follicle wall opposite the follicle apex. Data are expressed as mean + SEM; n = 3–4 monkeys/each treatment group. Data were assessed by ANOVA, followed by Duncan's post hoc test; groups with different superscripts are different, P < .05.

To identify the individual EP receptors that mediate these structural changes associated with luteinization, additional follicles were injected with Indo plus an agonist for a single EP receptor. Agonists selective for EP1, EP2, and EP4 stimulated granulosa cell hypertrophy and collagen remodeling comparable with that seen in vehicle- and Indo+PGE2-injected ovaries (Figure 2B and Supplemental Figure 1, A, B, and D). Injection of the EP3 agonist did not stimulate granulosa cell enlargement or overall granulosa cell layer thickening, and collagen remained restricted to the stroma surrounding the follicle without integrating into the granulosa cell layer, similar to that seen after injection of Indo alone (Figure 2B and Supplemental Figure 1C).

A shift in steroid hormone production from primarily estrogen to primarily progesterone is an essential feature of luteinization in primate follicles. To determine whether PGE2 is involved in this transition, expression of the key steroidogenic enzymes 3βHSD and aromatase and serum steroid levels were assessed. Follicles injected with vehicle, Indo, and Indo+PGE2 showed similar, high levels of immunostaining for 3βHSD (Figures 1, D–F) and consistently lower levels of immunostaining for aromatase (Figures 1, G–I). In addition, there were no effects of these treatments on serum estradiol or progesterone on the day of ovary removal (Supplemental Figure 2).

Further studies confirmed that intrafollicular injection of agonists selective for individual EP receptors did not alter enzyme expression or serum steroid levels. 3βHSD and aromatase immunostaining were observed in each EP receptor agonist-injected follicle, and staining was similar to that seen in follicles injected with vehicle, Indo, and Indo+PGE2 (Supplemental Figure 1, E–N). Serum estradiol and progesterone levels were not different from those measured in monkeys before or after follicle injection with vehicle, Indo, or Indo+PGE2 (Supplemental Figure 2).

Cumulus-oocyte complex

The presence (or absence) of an oocyte and the condition of the cumulus granulosa cells (eg, unexpanded or expanded) were observed in each injected follicle to identify which EP receptors participate in cumulus expansion, necessary for oocyte release. In vehicle-injected follicles, oocytes were never observed among luteinizing granulosa cells, within the antrum or near the rupture sites (Table 1). These observations are consistent with cumulus expansion to facilitate oocyte release. In contrast, each Indo-injected follicle contained an oocyte surrounded by unexpanded cumulus cells near the mural granulosa cell layer (Figure 3A and Table 1). Oocytes were not observed in follicles injected with Indo+PGE2 (Table 1). Most follicles injected with either Indo+EP1 agonist or Indo+EP2 agonist lacked an oocyte (Table 1). Oocytes retained within follicles injected with Indo+EP1 agonist (Figure 3B) or Indo+EP2 agonist (Figure 3C) were surrounded by expanded cumulus cells and located in the follicle antrum, not near the follicle wall as observed in Indo-injected follicles. Oocytes surrounded by unexpanded cumulus cells were identified in 2 of 3 follicles injected with Indo+EP3 agonist (Table 1), similar to those observed in Indo-injected follicles. The majority (3 of 4) of Indo+EP4 agonist-injected follicles contained an oocyte (Table 1) surrounded by unexpanded cumulus cells, which was detached from the follicle wall and located within the follicle antrum.

Table 1.

Presence of Attached vs Detached Oocyte

Treatment	Vehicle	Indo	PGE2	EP1 agonist	EP2 agonist	EP3 agonist	EP4 agonist
Oocyte identified	0/3	3/3	0/3	1/3	1/4	2/3	3/4
Attached oocytea	0/3	3/3	0/3	0/3	0/4	2/3	0/4

For each treatment group, the denominator is the total number of ovaries observed.

^aOocyte-cumulus complex was in contact with mural granulosa cells of the follicle wall.

Figure 3.

Cumulus expansion and follicle rupture in injected follicles. Cumulus-oocyte complexes (arrows) are shown in follicles injected with Indo (A), Indo+EP1 agonist (B), and Indo+EP2 agonist (C). Scale bar in each image, 25 μm. All images in A–C are at the same magnification; oocytes are not necessarily imaged at maximal diameter. Representative ruptured follicles (D, E, and H–M) and unruptured follicles (F and G) are shown at ovariectomy (D and F) and in histologic sections (E, G, and H–M) from follicles injected with vehicle (E), Indo (G), Indo+PGE2 (D), Indo+EP1 agonist (H and K), Indo+EP2 agonist (I and L), Indo+EP3 agonist (F), or Indo+EP4 agonist (J and M). Arrow in D indicates a prominent ovulatory stigma. Arrowheads in E indicate granulosa cells at the boundaries of the rupture site. Double arrows in F indicate thinning tissue over enlarged, unruptured follicle; arrowhead indicates fimbria. Location of antrum, stroma, and granulosa cells (gcs) are indicated in G–J. Clusters of red blood cells (rbc) are indicated in H and I. In H, thick arrow indicates rupture to the exterior of the ovary, with the presence and absence of granulosa cells near the rupture site indicated with thin arrows. In I, arrowheads indicate the rupture site connecting the antrum (antrum) of the injected follicle with the antrum of a neighboring small antral follicle (*). In J, the antrum of the injected follicle and granulosa cells bordering a rupture (arrowheads) into the surrounding stroma are shown; presence of ovarian surface epithelium (ose) is indicated. K, L, and M are details of rupture sites in H, I, and J, respectively. Scale bars, 100 μm (E and G) and 200 μm (H–M). Ovarian tissue sections were stained with hematoxylin (blue) and eosin (pink).

Follicle rupture

The ovarian surface was assessed visually at the time of ovariectomy for the presence or absence of a prominent ovulatory stigma; serial sections of each ovary were also assessed for evidence of a rupture site connecting the follicle antrum and the exterior of the ovary. For each vehicle-injected follicle, a classic ovulatory stigma was observed protruding from the surface of the ovary above the injected follicle (Table 2). Ovarian sections showed rupture sites connecting the collapsing antrum directly to the ovarian exterior in all vehicle-injected follicles, with luteinizing granulosa cells protruding through the rupture site to the surface of each of these ovaries (Figure 3E). Indo injection resulted in an enlarged follicle lacking evidence of a rupture site or ovulatory stigma (Table 2). All follicles injected with Indo only were very large in diameter, with thinning tissue on the ovarian surface, consisting of both stroma and granulosa cells (Figure 3G). Follicles injected with Indo+PGE2 were very similar to vehicle-injected follicles, with a prominent ovulatory stigma and histologic evidence of rupture (Figure 3D and Table 2). There was no difference in the size of rupture sites between vehicle- and Indo+PGE2-injected follicles (Figure 4).

Table 2.

Presence or Absence of Ovulatory Stigmata and Histological Rupture Sites

Treatment	Vehicle	Indo	PGE2	EP1 agonist	EP2 agonist	EP3 agonist	EP4 agonist
Ovulatory stigmata	3/3	0/3	3/3	0/4	0/4	0/3	0/4
Rupture site	3/3	0/3	3/3	3/4	4/4	0/3	3/4

For each treatment group, the denominator is the total number of ovaries observed.

Figure 4.

Rupture site area. For each ovary, rupture site was modeled as an oval as described in Materials and Methods and expressed as mm². For follicles with multiple rupture sites, the area of each rupture site was determined, and an average was used for analysis. No rupture sites were observed in Indo and Indo+EP3 agonist-injected ovaries (N/A). Data are expressed as mean + SEM; n = 3–4 monkeys/each group except for Indo+EP1 agonist (n = 2). Vehicle, Indo+PGE2, Indo+EP2 agonist, and Indo+EP4 agonist were not different by ANOVA (P > .05); other groups were excluded from analysis due to complete failure of rupture (Indo and Indo+EP3 agonist) or the inability to accurately assess the rupture site on 1 ovary (Indo+EP1 agonist).

Ovaries containing follicles injected with Indo+EP-selective agonists were also evaluated to examine the role of each EP receptor in follicle rupture. A large opening in the ovarian surface was observed on 3 of 4 Indo+EP1 agonist-injected ovaries (Supplemental Figure 1P). These openings did not protrude above the ovarian surface like a typical stigma but instead were flush with the ovarian surface and were covered with semitransparent clotted follicular fluid and/or blood. Although 1 ovary did not experience follicle rupture, rupture sites in 2 Indo+EP1 agonist-injected ovaries were apparently larger than those measured in vehicle- and Indo+PGE2-injected ovaries (Figure 4). An unknown number of tissue sections was lost from the region containing the rupture site in an additional Indo+EP1 agonist-injected ovary, but qualitative observation suggests that the rupture site for this ovary was also quite large. Luteinizing granulosa cells were notably absent from a large area of the follicle wall around the rupture site of Indo+EP1 agonist-injected follicles (Figure 3, H and K), in contrast with the luteinizing tissue protruding through the rupture site as seen in vehicle- and Indo+PGE2-injected ovaries (Figure 3E).

Each Indo+EP2 agonist-injected ovary had a small reddened area on the ovarian surface and lacked the protruding luteinizing tissue of the classic ovulatory stigmata (Table 2 and Supplemental Figure 1Q). However, histological examination showed that rupture indeed occurred in each of the Indo+EP2 agonist-injected ovaries. These rupture sites were comparable in size with rupture sites in vehicle- and Indo+PGE2-injected follicles (Figure 4). One of 4 follicles injected with Indo+EP2 agonist contained 2 rupture sites: one into an adjacent small antral follicle (Figure 3, I and L) and a second rupture from the antrum of the injected follicle directly to the ovarian exterior (data not shown). Both rupture sites were in the apical region of the follicle and breached the granulosa cell layer, follicle basement membrane, and stroma directly surrounding the injected follicle. No rupture site was observed between the ovarian exterior and the small antral follicle shown in Figure 3, I and L.

Indo+EP3 agonist-injected ovaries showed no evidence of follicle rupture at surgery or upon histological evaluation. Indo+EP3 agonist-injected ovaries were very similar in appearance to Indo-injected ovaries, with no ovulatory stigmata and very thin layers of granulosa cells and stroma between the follicle antrum and the ovarian exterior (Figure 3F and Table 2).

Each Indo+EP4 agonist-injected ovary possessed a disruption of the surface over the injected follicle, but a prominent ovulatory stigma was never observed (Table 2 and Supplemental Figure 1R). Histological examination showed that most Indo+EP4 agonist-injected follicles possessed rupture sites connecting the follicle antrum and the ovarian exterior. Rupture sites were similar in size to those measured in vehicle- and Indo+PGE2-injected follicles (Figure 4). Multiple rupture sites were observed in 1 of 4 Indo+EP4 agonist-injected ovaries, with rupture occurring into an adjacent antral follicle (data not shown) and to the ovarian exterior (data not shown) as well as into the surrounding stroma (Figure 3, J and M). Granulosa cells were present along the follicle wall near each of these ovulatory openings (Figure 3J and data not shown). The rupture site connecting the follicle antrum and stroma was located along a thin portion of stromal tissue near the other rupture sites (Figure 3J). A continuous layer of ovarian surface epithelium was present in the tissue sections containing the rupture between the follicle and stroma (Figure 3J). However, the pocket of follicular fluid seen in Figure 3J was continuous with the ovarian exterior in a neighboring section (data not shown).

Discussion

This study identifies roles for multiple EP receptors in primate ovulatory processes. Although studies in rodents have focused on EP2, we examined the role of each EP receptor in PGE2-mediated ovulatory events. Previous studies by our laboratory showed that all 4 PGE2 receptors are expressed in the granulosa cells of monkey ovulatory follicles (18, 24). These granulosa cells can be categorized into subpopulations based on their location within the follicle and their functions in ovulatory events. Cumulus granulosa cells support the oocyte and undergo expansion to facilitate oocyte release at ovulation, whereas mural granulosa cells line the follicle wall and remain behind at ovulation to form the corpus luteum. Mural granulosa cells at the follicle apex participate in follicle rupture, whereas those away from the apex may be more involved in steroidogenesis. In monkey ovulatory follicles, levels of EP1 protein were higher in mural granulosa cells opposite the follicle apex when compared with apical granulosa cells; EP2 and EP3 levels were higher in cumulus cells when compared with mural granulosa cells just before ovulation (18). Other ovarian cell types also express EP receptors. For example, EP2 and EP4 expression has been reported in oocytes from mice, cows, and monkeys (26, 27). Expression of EP2, EP3, and EP4 receptors in bovine theca cells has also been described (28). Clearly, each cell type and region of the follicle differentially express each individual EP receptor. This may allow EP receptors to play different and complementary roles in the overall ovulatory process.

Structural luteinization is one of the earliest processes to begin after the ovulatory gonadotropin surge (29). Granulosa cells cease proliferating and undergo hypertrophy. The granulosa cell basement membrane is degraded, and stromal cells, including theca and endothelial cells, infiltrate into the granulosa cell layer. Previous studies have shown that indomethacin can reduce granulosa cell hypertrophy in ovulatory follicles (7, 30). We have confirmed and extended these observations to show that follicles injected with indomethacin alone or in combination with EP3 agonist showed little evidence of granulosa cell hypertrophy or stromal reorganization. In contrast, replacement with PGE2 or agonists for the EP1, EP2, or EP4 receptor restored both hypertrophy and reorganization of collagen, similar to that seen in vehicle-injected follicles. These observations are the first to indicate that structural luteinization of the primate follicle requires PGE2 action via EP1, EP2, and/or EP4 receptors. In granulosa cells, cell cycle arrest is likely a prerequisite for cellular hypertrophy and acquisition of differentiated cell function (31). For example, the ovulatory LH surge decreased cyclin B1 while increasing p21^Cip1 in monkey granulosa cells, consistent with the observed slowing of or exit from the cell cycle (32). Our data suggest that PGE2 may act through its EP1, EP2, and/or EP4 receptors to mediate the ability of the LH surge to control cell cycle exit and hypertrophy in luteinizing granulosa cells. Although this concept has not been directly tested in granulosa cells, PGE2 has been reported to promote hypertrophy of rat gastrointestinal epithelium and regulate expression of cell cycle proteins, including cyclins and cyclin-dependent kinase inhibitors, in rodent hepatocytes (33–35).

Perhaps surprisingly, PGE2 ablation or replacement did not alter steroidogenesis in luteinizing monkey follicles. Elevation of progesterone production and reduction of androgen aromatization to estrogens are features of functional luteinization (36). Key enzymes that regulate progesterone and estrogen synthesis are 3βHSD and aromatase, respectively. In the present study, prostaglandin ablation with indomethacin and replacement with PGE2 or any EP receptor-selective agonist had no effect on serum levels of estradiol and progesterone, consistent with previous studies in monkeys and women (6, 37, 38). Furthermore, 3βHSD and aromatase were expressed by all injected follicles, regardless of the effects of these treatments on structural luteinization. Previous studies in monkey granulosa cells showed that 3βHSD expression was low before hCG but rapidly increased after administration of an ovulatory dose of hCG, whereas aromatase expression was highest just before hCG administration and then declined in a time-dependent manner after hCG (39). This pattern of enzyme expression is consistent with increasing progesterone and declining estrogen production around the time of ovulation. PGE2 has been reported to regulate expression of these key enzymes in nonovarian cells (40, 41), and PGE2 regulation of steroid hormone production by primate luteal cells is well established (42). Although PGE2 weakly promoted steroid hormone production by monkey granulosa cells luteinizing in vitro, this effect was minimal when compared with the ability of an ovulatory dose of gonadotropins to enhance steroidogenesis (43, 44). Moreover, inhibition of endogenous PG production did not prevent LH-stimulated progesterone production by granulosa cell in vitro (43). Taken together, these findings support the concept that PGs, including PGE2, are not key regulators of steroidogenesis in the primate ovulatory follicle.

The present study identifies specific EP receptors that mediate PGE2-stimulated cumulus expansion in primates. In this study, indomethacin blocked and PGE2 replacement restored cumulus expansion in monkey ovulatory follicles. Stimulation of either the EP1 or EP2 receptor restored cumulus expansion, because oocytes were either missing from ruptured follicles or identified in the follicle antrum surrounded by expanded cumulus. Follicles injected with the EP4 agonist contained oocytes with detached but unexpanded cumulus. In contrast, the EP3 receptor appears to play little or no role in this process. The role of gonadotropins, PGE2, and epidermal growth factor (EGF) family members to promote cumulus expansion is well established in rodents. Deletion of COX-2 expression and the resulting loss of follicular PGE2 synthesis caused a failure of cumulus expansion, which was restored by PGE2 treatment (8). PGE2 has been shown to regulate cumulus cell production of hyaluronic acid and TNF-α-induced protein 6 (Tnfaip6), both of which are required for successful cumulus expansion (45, 46). EGF family members, such as amphiregulin and epiregulin, also promote cumulus expansion by increasing hyaluronic acid synthase 2 (Has2) and TNF-α-induced protein 6 (Tnfaip6) expression (47). PGE2 mediates the ability of the LH surge to increase expression of EGF family members in both rodent and human granulosa cells (48–51), so PGE2 may provide an essential link between the ovulatory gonadotropin surge and EGF family member expression to promote cumulus expansion in primate follicles as well. In the present study, the EP2 agonist promoted cumulus expansion. EP2 is highly expressed in monkey cumulus cells and responds to agonist stimulation with increased cAMP (18, 24), consistent with studies in mice identifying key roles for EP2 and cAMP to regulate cumulus expansion (52, 53). The present studies show that EP1 can also promote cumulus expansion in primate follicles, which contrasts with findings of normal fertility in mice lacking EP1 expression (12, 54). In monkey follicles, EP1 is most abundant in mural granulosa cells not at the follicle apex, but these receptors do not regulate cAMP accumulation (18, 19). Additional studies will be needed to determine whether EP1 stimulation can increase expression of the EGF family members to promote cumulus expansion. Although the EP4 receptor is expressed in monkey mural granulosa cells and is coupled to Gαs to increase cAMP, stimulation of this receptor in the monkey follicle resulted in detachment of the cumulus-oocyte complex from the ovarian wall in the absence of cumulus expansion, suggesting that these 2 processes may be regulated independently.

An essential feature of ovulation is the removal of cells and connective tissue at the follicle apex, permitting follicle rupture (55, 56). Previous reports demonstrated that indomethacin blocks and PGE2 restores the ability of the ovulatory gonadotropin surge to promote follicle rupture (6, 7, 30). In the present study, we confirm and extend these observations to implicate EP1, EP2, and EP4 in this process. EP1 promoted the formation of extraordinarily large rupture sites with extensive loss of luteinizing granulosa cells near the rupture sites. Both EP2 and EP4 agonist-injected follicles lacked classic ovulatory stigmata but did possess rupture sites connecting the follicle antrum to the ovarian exterior. Interestingly, multiple rupture sites were observed in 2 ovaries, one injected with the EP2 agonist and one injected with the EP4 agonist. Overall, these observations suggest that activation of a single EP receptor can dysregulate this essential process and cause excessive proteolysis, sometimes at multiple or unanticipated sites. Blockade of follicle rupture with COX inhibitors has led to the assumption that PGs regulate specific proteolytic pathways associated with follicle rupture. However, very few reports demonstrate PG regulation of individual proteases, protease inhibitors, or protease activities in ovulatory follicles (43, 57). Our laboratory previously showed that EP2 is a key regulator of tissue-type plasminogen activator (tPA) in monkey granulosa cells (19). Both EP2 and EP4 promote increased cAMP, which can increases tPA expression (58) and may contribute to the multiple rupture sites seen in some EP2- and EP4-injected ovaries. In contrast, EP1 increased expression of tPA's endogenous inhibitor plasminogen activator inhibitor-1 in monkey granulosa cells; high expression of both EP1 and plasminogen activator inhibitor-1 away from the apex supports the concept that EP1 protects most the follicle wall from excessive proteolysis (18, 19). The mechanism by which EP1 stimulates (or fails to prevent) excessive proteolysis specifically at the follicle apex remains to be identified. However, the present study does show that EP1, EP2, and EP4 each promote unique, individual aspects of follicle rupture and suggests that normal proteolytic activity throughout the ovulatory follicle requires the coordinated action of multiple EP receptors.

The studies presented here are the first to demonstrate that EP1, EP2, and EP4 each mediate aspects of PGE2-stimulated ovulatory processes in primates in vivo. Both EP1 and EP2 agonists promote key aspects of ovulation, including structural luteinization, cumulus expansion, follicle rupture, and oocyte release. EP4 does not promote the production of a fully expanded cumulus but does stimulate detachment of the cumulus-oocyte complex from the follicle wall. Although previous studies indicated that the EP3 protein is expressed at very high levels just before ovulation (18, 24), the EP3 receptor agonist alone does not mediate any of these ovulatory processes. Because no single EP receptor recapitulated normal ovulation, it remains possible that EP3 cooperates with other EP receptors to regulate ovulatory events. Although studies in rodents used primarily knockout approaches to identify EP2 as the receptor exclusively responsible for PGE2-dependent ovulatory processes (10, 11), the present study used an ablate-and-replace strategy to identify contributions from multiple EP receptors in PGE2-induced ovulatory events in primates. Further studies to identify essential ovulatory processes stimulated by each EP receptor may indicate that receptor-selective agonists and/or antagonists could be useful to promote or prevent fertility in women.