Effect of Antiprogesterone RU486 on VEGF Expression and Blood Vessel Remodeling on Ovarian Follicles before Ovulation

Background The success of ovarian follicle growth and ovulation is strictly related to the development of an adequate blood vessel network required to sustain the proliferative and endocrine functions of the follicular cells. Even if the Vascular Endothelial Growth Factor (VEGF) drives angiogenesis before ovulation, the local role exerted by Progesterone (P4) remains to be clarified, in particular when its concentration rapidly increases before ovulation. Aim This in vivo study was designed to clarify the effect promoted by a P4 receptor antagonist, RU486, on VEGF expression and follicular angiogenesis before ovulation, in particular, during the transition from pre to periovulatory follicles induced by human Chorionic Gonadotropins (hCG) administration. Material and Methods Preovulatory follicle growth and ovulation were pharmacologically induced in prepubertal gilts by combining equine Chorionic Gonadotropins (eCG) and hCG used in the presence or absence of RU486. The effects on VEGF expression were analyzed using biochemical and immunohistochemical studies, either on granulosa or on theca layers of follicles isolated few hours before ovulation. This angiogenic factor was also correlated to follicular morphology and to blood vessels architecture. Results and Conclusions VEGF production, blood vessel network and follicle remodeling were impaired by RU486 treatment, even if the cause-effect correlation remains to be clarified. The P4 antagonist strongly down-regulated theca VEGF expression, thus, preventing most of the angiogenic follicle response induced by hCG. RU486-treated follicles displayed a reduced vascular area, a lower rate of endothelial cell proliferation and a reduced recruitment of perivascular mural cells. These data provide important insights on the biological role of RU486 and, indirectly, on steroid hormones during periovulatory follicular phase. In addition, an in vivo model is proposed to evaluate how periovulatory follicular angiogenesis may affect the functionality of the corpus luteum (CL) and the success of pregnancy.

VEGF, during gonadotropin surge, controls the crucial follicles transition from preovulatory to periovulatory stage that precedes ovulation [12,16,27]. In this time period of 24 or 44 h, depending on species, follicle profoundly changes its morphology and function [1,25]. Luteinizing hormone (LH) surge induces a progressive disorganization of follicle basal membrane caused by proteolityc enzymes activation. Moreover, it modifies the steroid enzymatic pathway transforming an estrogen secreting preovula-tory follicle to a P 4 producing periovulatory follicle [28,29]. In addition, follicular blood vessels undergo to dramatic changes. For the first time, large blood vessels appear in follicle walls and a higher blood flow is recorded [6,7,10,27,30]. Experimental evidences show that VEGF controls this vascular remodeling [10,11,21,22,27]. Indeed, its inhibition stops endothelial cell proliferation, impairs follicle angiogenesis preventing preovulatory follicle growth and ovulation [17,22,31,32]. Although VEGF role on follicle development is clearly established, the mechanisms involved in its local expression remain to be clarified. Gonadotropins seem to link follicle growth and blood vessel remodeling through VEGF expression regulation [6,7,27,33,34]. However, it remains to be clarified whether gonadotropins directly affect VEGF expression or if this angiogenic factor is indirectly influenced by follicular steroidogenesis, as suggested by in vitro [33,35] and in vivo experiments [36]. Amongst steroids, P 4 may have a role in inducing VEGF expression [37,38], endothelial cell proliferation [39][40][41], and angiogenesis [19,42]. Indeed, the administration of P 4 antagonist molecule, RU486, significantly decreased VEGF synthesis in vivo, in rat ovary [43] and in monkey endometrium [37]. Even if P 4 influence on ovarian blood flow was demonstrated [44][45][46][47], during the periovulatory stage, when the Figure 1. Experimental Plan. A) A schematic illustration of in vivo hormonal synchronization able to promote follicle growth and ovulation in prepubertal gilts. The periovulatory interval studied is showed in the box. B) Different experimental treatments and animal groups. All groups (n = 5 animals for each group) received eCG injection. Sixty-two hours after the eCG treatment, the first group (Iu) was ovariectomized to obtain preovulatory follicles (no hCG), while the remaining groups received hCG treatment, with or without RU486. In detail, the second group (IIu) received hCG injection dissolved in aqueous solution (*), as reagent control (CTR); the third group (IIIu) received hCG dissolved in corn oil (**), as experimental CTR, followed by administration of a corn oil alone 18 hours later; the fourth group (IVu) received hCG in combination with RU486 dissolved in corn oil and, 18 hours later, a second RU486 injection. Few hours before ovulation (36 hours after hCG treatments) the animals of these experimental groups were ovariectomized to obtain hCG, hCG+vehicle and hCG+RU486 periovulatory follicles. *commercial solvent **solvent for RU486. doi:10.1371/journal.pone.0095910.g001 RU486 In Vivo Effect on Ovarian Follicles PLOS ONE | www.plosone.org LH surge induces a complete inversion of follicular steroid secretion [8], the role exerted by steroids on follicular VEGF synthesis still remains to be investigated.
To this aim, the present in vivo study was designed to study the effects induced by P 4 antagonist administration, RU486 [48][49][50][51][52][53], on VEGF expression and angiogenesis during the transition from preovulatory to periovulatory follicles in gilts. This phase of follicle development was in vivo reproduced using a validated hormonal protocol [7,27] able to promote antral follicular growth until the preovulatory stage (eCG injection) and ovulation (hCG treatment). Granulosa and theca VEGF contents were studied in selected follicles and its expression was correlated with follicle morphology and blood vessel organization assessed through histological, immunofluorescence and biochemical studies.

Ethical Committee
The experiments were approved by the Ethics Committee of the Universities of Teramo and Chieti-Pescara (Prot. 81/2011/ CEISA/COM). All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.

Animal Treatments and Ovary Collection
Follicular growth and ovulation were pharmacologically promoted into 20 prepubertal Large White gilts with a mean weight of 90.761.16 Kg (mean 6 Standard Error, SE) using two sequentially intramuscular (i.m.) administrations of 1250 IU of eCG (Folligon; Intervet International B.V.-Boxmeer, Netherlands) and 750 IU of hCG (Corulon; Intervet International-Boxmeer, Netherlands), respectively [6,7,27]. The animals were divided in four experimental groups (5 animals each) as summarized in Fig. 1. Ovariectomies [27] were performed in order to collect preovulatory follicles 62 hours after eCG treatment (no hCG) and periovulatory ones collected 36 hours after hCG injection (Fig. 1B).

Sample Preparation
After ovariectomy, one ovary was fixed in 4% paraformaldehyde/phosphate-buffered saline (PBS; pH 7.4) for 12 h at 4uC, dehydrated and embedded in paraffin wax before performing histological and immunohistochemical analyses. The contra lateral ovary was collected to isolate single healthy preovulatory (8 mm, diameter $7 mm) and periovulatory follicles (11 mm # diameter .8 mm), as previously reported [6,27]. Each single follicle was opened under a stereomicroscope to collect follicular fluid (FF) and isolate follicular wall. Each follicular wall was dissected in order to separate granulosa cells and theca shell [6,7,27]. In detail, the follicle wall obtained from each structure was transferred into dissection medium and, with the aid of a small spatula, the granulosa layer was gently scraped away from the theca shell. The medium containing dispersed granulosa cells was collected and centrifuged, while the theca shell was vigorously vortexed and carefully washed in order to remove any possible granulosa cell contamination (as preliminary demonstrated by histological analysis) [6]. FF, granulosa and theca samples were individually stored in liquid nitrogen.

Morphological and Morphometric Analyses of the Follicular Response
Paraffin sections of 5 mm of thickness were serially collected on poly-L-lysine-coated slides before processing them for morphological, immunohistochemical (IHC) and morphometric investigations, according to Martelli et al. [11,12,27]. More in detail, at least three different sections/follicle were processed to analyze: Table 3. Evaluation of preovulatory follicles (no hCG) and periovulatory follicles, 36 hours after hCG, with or without RU486, collected from hormonally-induced prepubertal gilts. -VEGF distribution within granulosa and theca compartment using IHC detection [11]; -endothelial cell proliferation using vWF and Ki-67 double IHC (dIHC) detection [11,27]; -blood vessel maturation and mural cell recruitment using vWF and alpha Smooth Muscle Actin (a-SMA) dIHC analysis [27]. The used antibodies are summarized in Table 1.
VA values were determined on the follicular wall (Tot VA) or on the inner and outer blood vessel network (iN and oN, respectively) by quantifying the relative vWF-positive area (mm 2 ) determined within a defined field of 15000 mm 2 [11,12,27]. The results were recorded on at least five different sections/follicle and expressed as mean values 6 SE.
Morphometric analyses were performed at x 400 Magnification according to Martelli et al. [11,27] with an Axioscop-2plusepifluorescence microscope (Zeiss) provided with an interactive and automatic image analyzer (Axiovision, Zeiss). Morphometrical analysis was then performed by using a KS300 computed image analysis system (Zeiss).

VEGF Biochemical Analysis
The levels and distribution of VEGF protein were analyzed in FF, granulosa and theca compartments.

VEGF Content in FF
VEGF content in FF samples was analyzed using a specific ELISA assay (Quantikine; R&D Systems, Minneapolis, MN) as previously described [6,7,27]. VEGF soluble levels recorded on at least five different samples of FF/hormonal treatment were expressed as ng/ml (mean values) 6 SE.

VEGF Protein Content in Granulosa and Theca Compartments
Proteins extracted from granulosa or theca layers of each single isolated follicle (n = 5 follicles/gilt/treatment) were processed according to Martelli et al. [27]. Proteins (75 mg) were separated by 12% SDS-PAGE and electrophoretically transferred into a nitrocellulose membrane (Hybon C Extra; Amersham Pharmacia, Piscataway, NJ, USA) for Western Blot analysis [55]. Protein detection was performed by incubating the membranes with the primary polyclonal anti-human VEGF-antibody (2.5 mg/ml); Calbiochem, La Jolla, CA). The goat anti-rabbit immunoglobulins peroxidase-conjugated (IgG-HRP; 0.1 mg/ml, Santa Cruz Biotechnology Inc, Heidelberg, Germany) were finally used as secondary antibody. The membranes were re-probed with a monoclonal anti a-Tubulin (2 mg/ml, Sigma, St Louis, USA) and a goat anti-mouse IgG-HRP secondary antibody (0.1 mg/ml, Santa Cruz Biotechnology Inc, Heidelberg, Germany). The signals were detected with ECL Western Blot analysis system (Amersham Pharmacia, Amersham Pharmacia, Piscataway, NJ, USA). VEGF protein expression quantitative data were determined as the mean ratio of the optical density of specific bands normalized for the a-Tubulin expression through Advanced Image Data Analyzer (Rai Test; GMBH, Straubenhardt, Germany).

VEGF mRNA Content Recorded In Granulosa and Theca Layers Using Real Time PCR
Total RNA was extracted from granulosa and theca tissues of each single isolated follicle using TRI-Reagent (Sigma, S Louis, USA) as previously described [56]. Purified RNA was resuspended in RNAse-free water and spectrophotometrically quantified (A 260 nm ). One mg of RNA was electrophoretically separated in 1% Agarose gel in order to determine RNA quality. The RNA was treated with DNase I digestion (Sigma) for 15 min at room temperature. One mg of total RNA was used for reverse Real-time quantitative PCR was performed with Stratagene Mx3000P instrument, (Stratagene, La Jolla, CA) using SYBR Green I dye detection accordingly to Barboni et al [56]. Detailed information on VEGF and b-Actin [57] gene primers, product length and cycles are reported in Table 2. The reaction components were prepared at a concentration of: 2.5 ml forward (0.25 mM) primer, 2.5 ml reverse (0.25 mM) primer (Table 2), 3.5 ml water and 12.5 ml Brilliant SYBR Green QPCR Master  sample/follicle was run in triplicate, and normalized to the housekeeping b-Actin gene mRNA level. The amplicons specificity was confirmed by dissociation curve and by 2% agarose gel electrophoresis. Each sample was run in triplicate and for quantitation of VEGF gene target the Comparative Ct Method was applied for all samples normalized to the control housekeeping b-Actin gene using the formula: 2 2DDC(t ) = 2 2(DCt control gene 2 DCt target gene) . Statistical analysis was performed considering the last 3 experiments with a mean of 5 follicles/gilts/experiment.

Statistical Analysis
The data were expressed as mean 6 SE of at least 3 independent experiments. The quantitative data obtained from the different hormonal treatments were, firstly assessed for normalcy of distribution, by D9Agostino and Pearson test. Then, the data sets were compared using ANOVA test followed, when necessary, by post-hoc Tukey test (GraphPad Prism 5, GraphPad Software, USA). The data were considered significant for p#0.05.

Ovarian Response to Hormonal Treatment
Macroscopic ovary evaluation showed that preovulatory and periovulatory follicular rate was not influenced by RU486 treatments as summarized in HE staining (Fig. 2) confirmed that all isolated follicles were healthy and contained an oocyte surrounded by a cumulus cells continuous layer. They displayed a nucleus with a condensed chromatin and an uniform cytoplasm appearance (data non shown). Preovulatory follicles (n = 24; no hCG) showed a microscopic spherical regular architecture with a compact mural granulosa layer apposed on the basal membrane ( Fig. 2A). Differently, periovulatory follicles isolated after hCG (data not shown) or hCG+vehicle (Fig, 2B) treatments displayed a typical dispersed granulosa layer, with infoldings of the theca projected towards the antrum. No vehicle effect was observed on periovulatory follicles. This periovulatory morphology was observed only in , 30% of follicles isolated from RU486 treated gilts (n = 28; Fig, 2C, left image), while the remaining 70% maintained a structural organization that was similar to preovulatory follicles (Fig. 2C, right image).

VEGF Protein Content-in Follicular Fluid and Granulosa/ theca Compartments
FF, granulosa cells and theca layer were separately collected by each follicle with the aid of a stereomicroscope. As summarized in Table 4, VEGF content in FF was 15.9660.53 ng/ml in preovulatory follicles (no hCG), resulted undetectable after both hCG treatments (hormone alone or solubilized in vehicle), and significantly lower in RU486 periovulatory follicles (0.9860.07 ng/ml; Table 4). Granulosa VEGF protein expression, was similar in all analyzed preovulatory and periovulatory follicles (Fig. 3A). Instead, theca VEGF levels increased in both periovulatory hCG and hCG+vehicle follicles (Fig. 3B), while it did not change in hCG+RU486 follicles and it resulted similar to those recorded in preovulatory ones (Fig. 3B).

VEGF Protein Distribution in Follicular Compartments
VEGF protein distribution within granulosa and theca layers was investigated using IHC, as shown in Fig. 4. In all categories of follicles, VEGF was present within several granulosa cells or accumulated amongst them. By contrast, VEGF localization in theca compartment resulted influenced by hormonal treatments. More in detail, in preovulatory follicles (no hCG, Fig. 4A) VEGF was exclusively localized near the blood vessels. Differently, both periovulatory hCG (data not shown) and hCG+vehicle follicles (Fig. 4B) showed a preferential extracellular matrix VEGF distribution close to the basal membrane and to blood vessels. A similar theca VEGF distribution was observed in RU486 periovulatory follicles (Fig. 4C), even if in those follicles that maintained a preovulatory-like organization its distribution close to the blood vessels appeared lower, if not absent (Fig. 4C, right image).

VEGF mRNA Expression in Granulosa and Theca Layers
As shown in Fig. 5, VEGF mRNA expression significantly decreased in granulosa cells after hCG injection, regardless of the treatments used (Fig. 5A). Differently, VEGF mRNA levels increased in theca layer after hCG injection alone or with vehicle (Fig. 5B). By contrast, VEGF up regulation was absent in theca layers of follicles obtained from RU486 treated animals (Fig. 5B, hCG+RU486 vs hCG+vehicle, p#0.05) and its level remained similar to those of preovulatory ones (Fig. 5B, hCG+RU486 vs. no hCG, p$0.05).

Blood Vessel Remodeling during RU486 Treatment
The effects of RU486 on blood vessel organization (Fig. 6), VA and endothelial cell proliferation (Fig. 7) were evaluated using IHC. Preovulatory follicles (no hCG) displayed a blood vessel network organized into two concentric structures connected to each other by anastomotic vessels. The iN VA (2661.82689.46 mm 2 /field, Fig. 7A), characterized by capillaries close to the basal membrane (Fig. 6), was lower than the oN VA (3751.07689.44 mm 2 /field, Fig. 7A). The Tot VA significantly increased after both hCG and hCG+vehicle treatments (Fig. 7A,  p#0.01). In particular, the oN VA dramatically increased (Fig. 7A) becoming ,70% of Tot VA for the presence of new large blood vessels (Fig. 6).
Total and oN VAs similarly increased in follicles that acquired after RU486 treatment a periovulatory organization (Fig. 6 insert image and Fig. 7A, ''peri''). On the contrary, VAs did not change in those follicles that maintained a preovulatory-like architecture which showed few large vessels within the theca ( Fig. 6 and Fig. 7A, ''pre-like'').
The vWF (red stain) and a-SMA (green stain) doubleimmunolabeling, performed to describe blood vessel maturity and perivascular recruited cells, showed a clear RU486 treatment effect. (no hCG) Blood vessels preovulatory follicles (iN and oN) did not display a-SMA immunopositivity (Fig. 6). Instead, perivascular a-SMA positive cells appeared in iN and oN of periovulatory follicles obtained after hCG (data not shown) and hCG+vehicle treatments (Fig. 6). In particular, the iN displayed an a-SMA immunopositivity close to the endothelial cells (vWF positive cells) forming a ''chain-like'' organization. By contrast, a-SMA positivity was observed at the periphery of the large blood vessel walls localized in oN (Fig. 6).
A similar distribution of a-SMApositive cells was observed in those follicles isolated by RU486 treated animals that acquired a periovulatory organization (Fig. 6 insert image, ''peri''). By contrast, the follicles maintaining the preovulatory-like architecture (Fig. 6, ''pre-like'') did not display a-SMA positivity into the iN, and showed a reduced a-SMA perivascular immunoreaction in the oN large blood vessels. Rarely, these large blood vessels displayed a continuous a-SMA perivascular layer (Fig. 6, ''prelike'').

vWF and Ki-67 Staining
The vWF and Ki-67 double immunolabeling was used to quantify the proliferating endothelial cells in the different classes of follicles (Fig. 7B). Preovulatory (no hCG) and periovulatory follicles (hCG and hCG+vehicle) displayed endothelial Ki-67 positive cells in the iN and oN. The incidence of proliferating endothelial cells significantly decreased only in follicles isolated from RU486 treated animals that maintaining the preovulatorylike architecture (Fig. 7B, ''pre-like'').

Discussion
The present study was performed to clarify the role of RU486 on VEGF-dependent ovarian angiogenesis that occurs in vivo in dominant follicles after hCG administration (to mimic LH surge). Our results demonstrate, for the first time, that VEGF production, blood vessel network, and follicle tissue remodeling were all affected by the in vivo administration of RU486 during the phase of transition from pre to periovulatory follicle, characterized by increasing levels of P 4 [28,29], even if the cause-effect correlation between these events remains to be clarified. In addition to pregnancy [52,53,58,59], ovulation [60][61][62][63][64][65], and sperm-oocyte recognition [66], a new RU486 negative effect was demonstrated on reproductive events. Indeed, RU486 administration was able to impair periovulatory follicle development and angiogenesis induced by hCG. Its effect appeared to be compartment and vascular (oN and/or iN) network-dependent. In particular, RU486 treatment down-regulated VEGF expression within the theca compartment, while, it did not affect granulosa and FF VEGF content. Independently of RU486 injection, FF VEGF levels dramatically decreased after hCG treatment. This result could be apparently contradictory, since low FF VEGF levels were recorded both in RU486 and hGC periovulatory follicles characterized by low and high theca VEGF expression respectively. This observation, may be explained hypothesizing that hCG treatment, induces the synthesis of VEGF low bioavailability isoforms in follicles. This suggestion is in agreement with the expression of high molecular weight (i.e. VEGF 189 ) VEGF isoforms with a lower solubility evidence in endometrium stimulated by a modified steroid milieu [67]. Alternatively, VEGF bioavailability could be affected by local changes of heparin or cell surface heparan sulphate concentration, as previously supposed by Robinson & Stringer [68]. Regardless the mechanism, RU486 was able to selectively inhibit VEGF expression, triggered by hCG, within theca compartment of periovulatory follicles. This inhibitory effect may be a consequence of the antiprogesterone action of RU486, as reported in bovine [69], human [70] and mink [71], and in another reproductive tissue such as endometrium [37,38,42,72] exposed to low concentrations of P 4 . As a consequence, the RU486 follicles displayed an incomplete development of the blood vessel network. It is interesting to note as RU486 treatment did not influence the total VA but, specifically, affected oN VA that displayed a reduced extension. This important inhibitory RU486 influence was exclusively observed in 70% of the follicles that did not acquire the typical periovulatory organization. These results confirm that morphological and vascular remodeling recognize controlling cross-talk mechanisms [73][74][75][76][77][78]. On the contrary, RU486 treatment did not affect iN VA that probably remained regularly controlled by the unchanged FF and granulosa compartment VEGF levels. The presence of iN VA may have an essential functional role since it could maintain an adequate trophic supply of the germinal structure. Indeed, its persistence is vital to ensure the correct oxygen, precursors and metabolites supply either to the avascular granulosa compartment or to the oocyte [1,6,7,27].
Though, the RU486 inhibitory effects described on endothelial cell proliferation and mural cell recruitment can not be exclusively explained with the theca VEGF down-regulation. Indeed, RU486 treatment induced these responses either the oN or iN. The inhibition of endothelial cell proliferation could be a consequence either of VEGF lower levels, accordingly to other Authors [75,76], as of RU486 antagonist role on P 4 receptor. This hypothesis seems to be supported by the observations of P 4 receptors expression in Huvec cells [79,80] and by the P 4 -dependent inhibition of endothelial cell proliferation, already demonstrated in endometrial blood vessels in mouse [42] and in uterine vasculature in pig [38]. RU486 inhibitory role on mural cell recruitment affected also the consequent blood vessel maturity. Human CG injection in periovulatory follicles promotes the maturation of blood vessels by recruiting mural cells (a-SMA positive cells). These cells in the oN surround large blood vessels, while in iN capillary network acquire the typical ''chain-like organization'' [27]. The RU486 negative influence on blood vessel maturity could be related to a direct effect of P 4 , as already suggested in mouse [42], or as a consequence of VEGF lower expression. It is known that this angiogenic factor indirectly modulates blood vessel maturation in other systems by influencing the secretion of key factors such as Angiopoietins/Tie-2 receptor [74,75,81,82], platelet-derived growth factor (PDGF)-B/PDGF receptor beta [76], and the transforming growth factor beta (TGF beta) [76]. Regardless the mechanisms involved, blood vessels immaturity after RU486 treatment may impair the physiological transition from periovulatory follicle to CL structure. The presence of smooth muscle in follicle blood vessels is essential for the control exerted by nervous system or local hormones on the ovarian function [27]. In particular, the large vessels with a smooth muscle mural wall absence in the oN could probably impair blood flow required for ovulation and the subsequent CL development [83,84]. Moreover, a-SMA positive cells absence within the capillaries of the iN could negatively affect blood vessel stabilization and haemodynamic processes necessary to support the early stage of CL growth [85]. In this context, further experiments will be required to demonstrate if and how the low maturity of follicular blood vessels could influence ovulation and CL function.
In conclusion, the obtained results in the present work demonstrate RU486 negative influences on VEGF expression, vascular and tissue remodeling, providing important insights on the biological in vivo role of this P 4 antagonist during the transition from the preovulatory to the periovulatory stage. Even if the causeeffect correlation amongst angiogenesis and impaired follicle development has not been demonstrated, these results propose an in vivo model that allows to study the effect of an incomplete follicle maturation on female reproduction success.