BIOLOGY OF REPRODUCTION (2014) 91(3):73, 1–10 Published online before print 13 August 2014. DOI 10.1095/biolreprod.114.119990

Endocrine Gland-Derived Endothelial Growth Factor (EG-VEGF) Is a Potential Novel Regulator of Human Parturition1 C. Dunand,3,4,5 P. Hoffmann,3,4,5 V. Sapin,8 L. Blanchon,8 A. Salomon,3,6 F. Sergent,3,6 M. Benharouga,5,6,7 S. Sabra,9 J. Guibourdenche,10 S.J. Lye,9 J.J. Feige,3,5,6 and N. Alfaidy2,3,5,6 3

Institut National de la Sante´ et de la Recherche Me´dicale (INSERM U1036), Grenoble, France Department of Obstetrics and Gynaecology, Grenoble Hospital, Grenoble, France 5 Universite´ Grenoble-Alpes, Grenoble, France 6 Commissariat a` l’Energie Atomique (CEA), iRTSV-BCI, Grenoble, France 7 Centre National de la Recherche Scientifique, Unite´ Mixte de Recherche 5249, Laboratoire de Chimie et Biologie des Me´taux, Grenoble, France 8 EA7281 R2D2, Biochemistry Laboratory, Medicine School, Auvergne University, Clermont-Ferrand, France 9 Lunenfeld-Tanenbaum Research Institute Mount Sinai Hospital, University of Toronto, Toronto, Canada 10 Hoˆpital Cochin AP-HP, Inserm U767, Faculte´ de pharmacie, Universite´ Paris Descartes, Paris, France 4

INTRODUCTION

EG-VEGF is an angiogenic factor that we identified as a new placental growth factor during human pregnancy. EG-VEGF is also expressed in the mouse fetal membrane (FM) by the end of gestation, suggesting a local role for this protein in the mechanism of parturition. However, injection of EG-VEGF to gravid mice did not induce labor, suggesting a different role for EG-VEGF in parturition. Here, we searched for its role in the FM in relation to human parturition. Human pregnant sera and total FM, chorion, and amnion were collected during the second and third trimesters from preterm no labor, term no labor, and term labor patients. Primary human chorion trophoblast and FM explants cultures were also used. We demonstrate that circulating EG-VEGF increased toward term and significantly decreased at the time of labor. EG-VEGF production was higher in the FM compared to placentas matched for gestational age. Within the FM, the chorion was the main source of EG-VEGF. EG-VEGF receptors, PROKR1 and PROKR2, were differentially expressed within the FM with increased expression toward term and an abrupt decrease with the onset of labor. In chorion trophoblast and FM explants collected from nonlaboring patients, EG-VEGF decreased metalloproteinase-2 and -9 activities and increased PGDH (prostaglandin-metabolizing enzyme) expression. Altogether these data demonstrate that EG-VEGF is a new cytokine that acts locally to ensure FM protection in late pregnancy. Its fine contribution to the initiation of human labor is exhibited by the abrupt decrease in its levels as well as a reduction in its receptors.

EG-VEGF (endocrine gland-derived endothelial growth factor), also called prokineticin 1, is the canonical member of a recently described family of cytokines, the prokineticin family [1, 2]. EG-VEGF expression was mainly reported in endocrine tissues, including the placenta [3–8]. Prokineticin family members exert their effects via two G protein-coupled prokineticin receptors, PROKR1 and PROKR2 [1, 9]. Recent data from our group showed that 1) EG-VEGF is abundant in the human placenta during the first trimester of pregnancy, with the strongest expression found in the syncytiotrophoblast [6], 2) its expression is up-regulated by hypoxia [6], 3) it controls trophoblast invasion and placental angiogenesis [10], and 4) its circulating levels are significantly increased in placental pathologies such as preeclampsia and intrauterine growth restriction [10, 11]. Altogether, these findings suggest that EG-VEGF is a new pregnancy factor that might play additional roles to those established during early pregnancy. Former publications that identified EG-VEGF and its analog, prokineticin 2, reported roles in the regulation of contractility for these factors in the gastrointestinal smooth muscle [2], suggesting a potential role for EG-VEGF in the mechanism of parturition, a multistep process that includes myometrial contraction, cervical ripening, and fetal membrane (FM) rupture [12]. In relation to the FM, Lannagan et al. [13] recently showed that EG-VEGF mRNA expression increased significantly in murine FM on Day 18 (as compared with Day 16) but not in the uterus or placenta. Nevertheless, intrauterine injection of EG-VEGF in gravid mice did not result in preterm delivery, suggesting a different role for EG-VEGF in the mechanism of parturition [13]. To date, no studies have examined the expression of EG-VEGF and its receptors, PROKR1 and PROKR2, in the human FM during late human pregnancy and in relation to the labor process. A role in the mechanism of parturition in humans remains to be investigated. Human FM represent a complex multilaminate tissue composed of the amnion and the chorion, two closely adherent layers that consist of several cell types, including epithelial and mesenchymal cells for the amnion and cytotrophoblastic and mesenchymal cells for the chorion. This tissue is characterized by the presence of abundant local metalloproteinase (MMP) activities that contribute to their rupture at the onset and progression of labor [14, 15]. The FMs are also the main intrauterine site that control prostaglandin (PG) synthesis and

EG-VEGF, fetal membranes, human pregnancy, mechanism of parturition, preterm birth, trophoblast 1

Supported by INSERM U1036, University Joseph-Fourier, Commissariat a` l’Energie Atomique (DSV/iRTSV/BCI), the region Rhoˆnes Alpes, and the Canadian Institutes of Health Research (to S.J.L.). Presented in part at the 61st Annual Meeting of the Society for Gynecologic Investigation, Florence, Italy, 25–29 March, 2014. 2 Correspondence: Nadia Alfaidy, HDR, Unite´ INSERM U1036, Laboratoire BCI-iRTSV, CEA, Grenoble 17, rue des Martyrs, 38054 Grenoble cedex 9, France. E-mail: [email protected] Received: 3 April 2014. First decision: 6 May 2014. Accepted: 14 July 2014. Ó 2014 by the Society for the Study of Reproduction, Inc. eISSN: 1529-7268 http://www.biolreprod.org ISSN: 0006-3363

1

Article 73

Downloaded from www.biolreprod.org.

ABSTRACT

DUNAND ET AL. Ethics. Tissues and sera collections were approved by the Ethics Committees of the University Hospital of Grenoble and Mount Sinai Hospital, respectively. Written informed consents were obtained from patients before collection.

metabolism during the third trimester of pregnancy and at the time of labor. In women, PG synthesis and metabolism are discretely compartmentalized in the FM [16]. Human amnion is a major site of PG synthesis because it expresses the highest levels of cyclooxygenase (COX-2), the key PG-synthesizing enzyme. In contrast, the chorion is the major site of PG metabolism as it expresses the highest levels of 15-hydroxyprostaglandin dehydrogenase (PGDH), the key PG-metabolizing enzyme [17, 18]. It is well-established that during the third trimester of human pregnancy, a fine regulation of PG synthesis and metabolism occurs to ensure the quiescence of the uterus until the onset of labor. In the FM, the chorion layer is a key compartment of the FM that ensures this quiescence by preventing active PGs from crossing to the decidua and myometrium before the onset of the labor process [16, 18, 19]. However to date, little is known about local regulators of these processes in the intrauterine tissues and particularly in the FM. Because EG-VEGF exhibits high circulating levels in preterm-associated pregnancies, such as preeclampsia and intrauterine growth restriction [10, 11], and because its local levels increase in the FM during late gestation, we hypothesized that this factor might a play a direct role in human parturition, other than triggering the onset of labor. Here, we investigated the role of EG-VEGF in human parturition during late pregnancy and in relation to the labor process. We compared the levels of circulating EG-VEGF in sera collected from pregnant women at the second and the third trimesters, in the absence or the presence of labor, determined the site of expression of EG-VEGF and its receptors within the human FM, and determined the ontogeny of these proteins in late pregnancy and in relation to the initiation of labor. We have also compared the levels of EG-VEGF in FM versus placental explant cultures and in chorion versus amnion explant cultures. More importantly, we characterized the mechanism by which EG-VEGF contributes to the control of the onset of human labor; this includes the regulation of the key parturition enzymes, MMP-2 and -9 and PDGH.

Immunohistochemistry

Immunofluorescence Tissue sections were rehydrated in serial dilutions of alcohol (100%, 90%, 70%, and 50%) and washed in PBS. Nonspecific binding of antibodies was blocked with 1% bovine serum albumin in PBS for 2 h at RT. The samples were then incubated with rabbit anti-human cytokeratin 7, anti-PROKR1 (33 lg/ml), and anti-PROKR2 (69 lg/ml). Tissue sections were incubated overnight (18–24 h) with the primary antibodies at 48C. After the incubation period, the sections were washed three times in 0.1 M PBS. The secondary antibodies were added and incubated at RT for 2 h. The secondary antibodies used were fluorophore-labeled Alexa Fluor 568 goat anti-rabbit immunoglobulin G (IgG) and Alexa Fluor 488 goat anti-mouse IgG (1:200 dilution in 1% bovine serum albumin; Molecular Probes). Samples were washed again in PBS and then dehydrated in serial dilutions of alcohol (50%, 70%, 90%, and 100%). Anti-fading reagent, 1 mg/ml p-phenylendiamine, in 50% glycerol and PBS was added to the tissue sections, and coverslips were applied before observation.

MATERIALS AND METHODS Tissue and Sera Collection

Western Blot Analysis

Tissue collection. Third-trimester human FMs were collected from women aged 33–39 (n ¼ 23; average age 36 yr) with uncomplicated healthy singleton pregnancies (no infection, no fetal growth restriction, no preeclampsia, and no premature rupture of membranes). Second-trimester human FMs (n ¼ 7; average age 33 yr) were provided by the Mount Sinai Hospital of Toronto, Canada, with the same sampling conditions. These samples were collected for a study published in 2003 by N. Alfaidy et al. [20]; the study examined the in situ ontogeny of the enzyme 11-beta-hydroxysteroid dehydrogenase expression in the FMs throughout human pregnancy. The indications for the preterm nonlaboring patients in this study included antepartum hemorrhage, cervical cancer, and multiple pregnancy. There were no clinical or pathological signs of preeclampsia or other placental disease. Term human placentas and FMs were obtained from uncomplicated pregnancies after elective cesarean section (nonlaboring women, n ¼ 6) and after spontaneous labor (n ¼ 6). To compare EG-VEGF expression in amnion and chorion layers, FMs collected from nonlaboring women were further dissected to separate the amnion (n ¼ 3) and the chorion (n ¼ 3); those samples were used in quantitative RT-PCR experiments. Samples were either stored at 808C shortly after collection or fixed in paraformaldehyde 4% for 24 h at 48C and then stored in 70% alcohol. Sera collection. The sera were obtained from another group of pregnant women during their routine visits at the second and third trimesters of pregnancy at Mount Sinai Hospital of Toronto, Canada. The blood samples were divided into three groups according to the term and type of delivery: term deliveries not in labor (n ¼ 11, average age and term: 35 yr and 39 wk of gestation [wg], respectively), term deliveries in labor (n ¼ 12, average age and term: 32 yr and 39 wg, respectively), and preterm deliveries not in labor (n ¼ 14, average age and term: 30 yr and 26 wg, respectively). There were no clinical or pathological signs of preeclampsia or other placental diseases.

Frozen placental samples were homogenized on ice with a Polytron 1200 mixer (Bioblock Scientific) in 1 ml of RIPA lysis buffer (Sigma-Aldrich) and 10 ll of protease inhibitor cocktail (Sigma-Aldrich). The homogenates were centrifuged (15 000 3 g at 48C) for 15 min, and the supernatants were collected. Protein concentrations were determined using the Bradford assay with a Multiscan EX (Thermo LabSystems) and Accent Software 2.6. Fifty micrograms of proteins extracted from total FM, amnion or chorion were electorophoretically separated in 12% acrylamide gels, and electrically transferred onto 0.45 lm nitrocellulose membranes (BioRad). Appropriate transfer was confirmed by protein staining with Ponceau S (Sigma-Aldrich). The blots were washed with PBS and 0.1% Tween (PBS-T) and incubated 1 h in blocking solution (5% skimmed milk in PBS-T). Membranes were immunoblotted overnight with a rabbit antibody against PROKR1 or PROKR2 (1:400 of 33 lg/ml anti-PROKR1 or 1:300 of 69 lg/ml anti-PROKR2) (Covalab). Blots were then rinsed three times with PBS-T and incubated with biotinylated goat anti-rabbit IgG (450 ng/ml, 1:2000 dilution in blocking solution; Dako) for 30 min. After three PBS-T washes, the membranes were incubated with a peroxidase-conjugated extravidin (1:2000 dilution in blocking solution; Sigma-Aldrich) for 30 min. Blots were washed four times with PBST, and the antibody-antigen complex was detected using the ECL detection system (GE Healthcare Life Science). The membranes were then exposed to radiographic films (Kodak Scientific Imaging Products). The intensities of immunoreactive bands were measured by scanning the photographic film and analyzing the image on a desktop computer using Image J software. The PROKR1 and PROKR2 protein were detected between 47–50 kDa. The mean pixel density for each band was analyzed to obtain relative optical density units for PROKR1 and PROKR2 protein expression. To standardize for sample

2

Article 73

Downloaded from www.biolreprod.org.

FMs were fixed for 24 h at 48C in 4% (vol/vol) paraformaldehyde, then in 70% alcohol. They were embedded in paraffin and cut into 5 lm sections as described previously [21]. FM sections were deparafinized and rehydrated. Endogenous peroxidase activity was quenched using 40 ml of pure methanol mixed with 1.3 ml of 30% hydrogen peroxide solution (Sigma-Aldrich) for 30 min at room temperature (RT). After two washes using PBS, nonspecific protein binding was blocked by 5% normal goat serum for 1 h at RT (Dako). Sections were then incubated with rabbit anti-human EG-VEGF, PROKR1, or PROKR2 polyclonal antibodies for 24 h at 48C (45 lg/ml anti-EG-VEGF, 33 lg/ml anti-PROKR1, 69 lg/ml anti-PROKR2) (Covalab). Anti-cytokeratin 7 antibody, a marker of trophoblast cells, was also used to identify the chorion leaf in first-trimester specimens. After two washes with PBS, a biotinylated secondary antibody (1.1 g/L polyclonal goat anti-rabbit biotinylated immunoglobulin) (Dako) was applied for 1 h at 1:400 dilution followed by peroxidaseavidin-biotin complex for 1 h (Vectastain ABC Kit; Vector Laboratories). Peroxidase activity was detected by incubation with 3,3 0 -diaminobenzidine for 30 sec to 1 min (liquid 3,3 0 -diaminobenzidine and substrate chromogen system) (Dako); the reaction was stopped in deionized water. The sections were then washed, counterstained with hematoxylin, dehydrated, and mounted with Merckoglass. Negative controls were performed using PBS 13 buffer instead of the primary antibody. Photographs were taken using a Zeiss AxioCam HRc coupled to a Zeiss Axioplan microscope.

EG-VEGF CONTROLS HUMAN PARTURITION TABLE 1. Primers used for real-time RT-PCR. Forward, 5 0 to 3 0

Reverse, 5 0 to 3 0

TG AGG AAA CGC CAA CAC CAT ACC CAG AAG ACT GTG GAT GG

CC GGG AAC CTG GAG CAC TTC TAG ACG GCA GGT CAG GT

Genes EG-VEGF GAPDH

loading, the blots were stripped in 5 ml of PBS-T mixed with 500 ll of Re-blot Plus Mild solution 103 (Millipore) and reprobed with an anti-b-actin antibody (Sigma-Aldrich) as an internal control for protein loading. The b-actin antibody detects a 37 kDa protein.

Zymography Twenty microliters of harvested explant culture medium were electrophoresed under nonreducing conditions in a 10% acrylamide gel containing 1 mg/ ml gelatin (Sigma) as previously described, [23]. After electrophoresis, the gels were washed at RT for 1 h in 2.5% Triton X-100 and 50 mM Tris-HCl, pH 7.5, and then incubated at 378C overnight in buffer containing 150 mM NaCl, 5 mM CaCl2, and 50 mM Tris-HCl, pH 7.6. Thereafter, gels were stained with 0.1% (wt/vol) Coomassie brilliant blue R-250 in 30% (vol/vol) isopropyl alcohol and 10% glacial acetic acid for 60 min and destained in 10% (vol/vol) methanol and 5% (vol/vol) glacial acetic acid. Semiquantification of the bands corresponding to 72- and 92-kDa gelatinases was performed with a densitometer.

EG-VEGF Quantification by ELISA EG-VEGF in the collected sera and in conditioned media collected from explant cultures was measured by ELISA (PeproTech). Two separate standard curves were constructed to allow accurate readings of samples at upper and lower ranges of the assay. All the samples were in the linear range of the standard curves. The detection limit of the assay was 16 pg/ml. Data were expressed as pg/ml of sera or as pg/mg of protein extracted from total FM or from separated chorion and amnion tissues.

Statistical Analysis

Real-Time RT-PCR Analysis Total RNAs were extracted from the placenta, FM, chorion, and amnion using the Nucleospin RNA kit (Macherey-Nagel). First-strand cDNAs were synthesized from 1 lg total RNA by reverse transcription using the iScript system (Bio-Rad) according to the manufacturer’s instructions. Quantitative real-time RT-PCR, using SYBER-green real-time PCR assay, was performed on a Bio-Rad CFX96 apparatus and quantitative PCR Master Mix (Promega). Relative quantification of gene expression was normalized to GAPDH mRNA expression level. Sequences of the PCR primers used are listed in Table 1.

RESULTS EG-VEGF Circulating Levels Increase During the Third Trimester of Pregnancy

Placental and FM Explant Cultures

To get more insight into the role of EG-VEGF during late pregnancy, we compared its levels in second- and thirdtrimesters sera collected from nonlaboring women only. Figure1 shows that circulating EG-VEGF levels significantly increased toward the end of pregnancy, suggesting a potential role of this factor in the late processes of parturition. The source of EG-VEGF could not be of placental origin because both our group and Lannagan et al. [7, 13] showed that placental EG-VEGF production does not increase toward the

Samples were collected and stored at 48C in culture medium (DMEM:Ham F-12; 1:1) with an addition of gentalline (1:1000) and penicillin and streptomycin (1:100). Forty-eight-well plates were coated with 150 ll/well of an equal mix of Matrigel (BD) and culture medium and preincubated at 378C for 30 min. Pieces of placenta or FM tissue were dissected, weighed, and placed on Matrigel for 3 h at 378C to ensure their adhesiveness. Culture medium (300 ll) with or without EG-VEGF (25 and 50 ng/ml) was then added. Each condition was done in triplicate. We incubated the samples at 378C for 1, 3, 8, 12, or 24 h. The supernatant was collected and stored at 208C. Tissue samples were homogenized in RIPA and stored at 808C.

Primary Chorion Trophoblast Culture Trophoblast cells from choriodecidual tissue were isolated and cultured using a modification of the technique described by Kliman et al. [22]. Briefly, human placentas were obtained from uncomplicated term pregnancies (38–40 wk, n ¼ 3) after elective cesarean section delivery. The chorion with adherent decidua was peeled off the amnion and digested three times for 60 min each with 0.125% trypsin (Gibco), 0.02% deoxyribonuclease I (Sigma), and 0.2% collagenase (Sigma) in DMEM (Gibco). The dispersed choriodecidual cells were filtered through 100 lM nylon gauze and loaded onto a continuous Percoll (Sigma) gradient (5%–70% in 5% steps of 3 ml each) and then centrifuged at 1200 3 g for 20 min at RT to separate different cell types. Cytotrophoblast cells between the density markers of 1.049 and 1.062 g/ml were collected and plated in 24-well plates (Corning) at a density of 106 cells/ml. Cells were also plated in 8-well chamber slides (Lab-Tek) (0.3 3 106 cells/well) for immunostaining. The cells were cultured for 3 days at 378C in 5% CO2/95% air before experimental treatments. Under these conditions, the chorionic trophoblast cells formed small clumps or remained as single cells. The purity of the cell preparation was assessed at the end of the experiment by histochemical staining for cytokeratin, an epithelial cell lineage marker (DAKO Corp.) or vimentin, a mesenchymal cell lineage marker (DAKO Corp.). Cells were counterstained with Carrazzi hematoxylin. After 72 h, representative cultures were stained as previously described [21]. Briefly, the cells were fixed with 4% paraformaldehyde and incubated with 10% normal goat serum that served as a blocking agent for nonspecific binding. Commercial antibodies were used to detect immunoreactive PGDH (Thermo Scientific), cytokeratin (Dako), and vimentin (Dako). The cells were incubated at 48C for 18 h. For detection, we used the avidin-biotinperoxidase technique and Vectastain ABC kit (Vector Laboratories). Immunoreactive proteins were visualized after the addition of 3,30 -diaminobenzidine (Sigma) for 2 min. Cells were counterstained with hematoxylin.

FIG. 1. Comparison of EG-VEGF serum levels in nonlaboring pregnant women during the second and third trimester of pregnancy. A total of 25 serum samples were analyzed (second-trimester samples, n ¼ 14; thirdtrimester samples, n ¼ 11). EG-VEGF levels were measured by ELISA. The box plot demonstrates 10th, 25th, 50th, 75th, and 90th percentiles. Oneway ANOVA was used followed by the Dunn method (*P , 0.05).

3

Article 73

Downloaded from www.biolreprod.org.

Statistical comparisons were made using Student t-test and one-way ANOVA (both parametric and nonparametric) followed by Dunn or Bonferroni tests. All the data were checked for normality and equal variance. (SigmaPlot and SigmaStat; Jandel Scientific Software). All the data are expressed as mean 6 SEM (*P , 0.05 and **P , 0.001).

DUNAND ET AL.

end of pregnancy. These data suggested that another source of EG-VEGF might exist during late pregnancy.

EG-VEGF and PROKRs Expression in Human FM Here, we determined the localization of EG-VEGF and its receptors, PROKR1 and PROKR2, within the FMs in nonlaboring women at term pregnancy. Figure 3Aa shows that EG-VEGF is mainly expressed in the chorio-decidua layer with an intense signal in the chorion layer. A strong expression was also observed in the amnion epithelial layer. Figure 3, Ab and Ac, show PROKR1 and PROKR2 expression, respectively. PROKR1 was localized to all layers of the FM. The strongest expression was observed in the chorion mesenchymal layer. PROKR2 was expressed in the chorion layer, with faint expression in the amnion epithelial cells. Figure 3Ad is a negative control that illustrates FM layers. To further characterize PROKR1 and PROKR2 expression, we used the immunofluorescence technique to search for colocalization of PROKR1 and PROKR2 with cytokeratin 7, a marker of trophoblast and epithelial cells. Figure 3Ba shows that PROKR1 is mainly expressed in the chorion mesenchymal layer, while PROKR2 is mainly expressed in the trophoblast

EG-VEGF Is Produced by Human FMs Using our established explant culture model, we compared EG-VEGF production in human FMs and in their age-matched control placentas. Figure 2A reports a time course secretion of EG-VEGF from total FM and their respective placentas. EGVEGF production was significantly higher in the FM than in the placenta. More importantly, the chorion layer within the FM was the main source of EG-VEGF (Fig. 2B). Further comparisons of EG-VEGF mRNA content in the placenta, total FM, chorion, and amnion is reported in Figure 2C. The data show that the chorion layer expresses significantly more EGVEGF mRNA compared to the placenta and to the amnion.

4

Article 73

Downloaded from www.biolreprod.org.

FIG. 2. EG-VEGF secretion from total fetal membrane (FM), placenta, chorion, and amnion. A) A comparison of EG-VEGF secreted levels between FM explants and their age-matched placentas from 1 to 24 h of culture. EG-VEGF levels were measured by ELISA. One-way ANOVA was used followed by the Bonferroni test (*P , 0.05; **P , 0.001). B) A comparison of EG-VEGF secreted levels between the chorion and amnion of same placentas. EG-VEGF levels were measured by ELISA. One-way ANOVA was used followed by the Bonferroni test (*P , 0.05). C) A comparison of EG-VEGF mRNA levels in placenta (n ¼ 5), FM (n ¼ 3), chorion (n ¼ 3), and amnion (n ¼ 3) samples collected from nonlaboring pregnant women at term of their pregnancy. Student ttest was used to compare values (*P , 0.05). Different letters are significantly different from each other.

EG-VEGF CONTROLS HUMAN PARTURITION

layer of the chorion (Fig. 3Bd). Figure 3, Bb and Be, show cytokeratin staining in the same sections. An obvious colocalization of cytokeratin and PROKR2 could be observed in the merged photographs shown in Figure 3Bf. A faint colocalization of PROKR1 and cytokeratin was observed in the chorion trophoblast, confirming that in the chorion, the mesenchymal layer is the major site for PROKR1 expression. We then compared EG-VEGF, PROKR1, and PROKR2 expression in the second- and third trimester FMs collected from nonlaboring women. Figure 4A–C show a significant increase in the expression of EG-VEGF and PROKR1. The increase in PROKR2 expression did not reach significance.

nonlaboring and laboring women. There was an important decrease in the EG-VEGF protein expression with labor. Figure 5B shows a representative Western blot analysis that compared EG-VEGF protein levels in both conditions in the FMs, and Figure 5C shows the quantification of EG-VEGF protein levels. There was a significant decrease in EG-VEGF protein levels with the labor process. We then compared circulating levels of EG-VEGF between laboring and nonlaboring sera. The data show that EG-VEGF circulating levels significantly decreased with the labor process (Fig. 5D). Using our explant culture model, we conducted a time course comparison of EG-VEGF production in nonlaboring versus laboring FMs. Figure 5E shows that EG-VEGF production from laboring FM explants was significantly lower compared to the production by FM explants collected from nonlaboring women. To further characterize the role of EG-VEGF/PROKRs system in the labor process, we compared the levels of expression of PROKR1 and PROKR2 in the same conditions.

EG-VEGF, PROKR1, and PROKR2 Expression Decreased with the Onset of Labor The increase in EG-VEGF levels during late pregnancy suggested a potential role of the EG-VEGF/PROKRs system in the mechanism of parturition. Figure 5A shows a comparison of EG-VEGF local expression in the FMs collected from 5

Article 73

Downloaded from www.biolreprod.org.

FIG. 3. EG-VEGF, PROKR1, and PROKR2 expression in human fetal membranes (FMs). FMs were collected from nonlaboring women during the third trimester of pregnancy. A) Immunohistochemistry of term FM immunostained (Brown staining) with anti-EG-VEGF (a); anti-PROKR1 (b), and anti-PROKR2 (c); d is a negative control of the staining. All the sections were counterstained with hematoxylin. Photographs in a 0 , b 0 , and c 0 are higher magnifications of a, b, and c, respectively. Bars ¼ 50 lm. B) Immunofluorescence of term FMs immunostained with anti-PROKR1 (a); anti-cytokeratin (b and e), and antiPROKR2 (d). Colocalization of PROKR1 and cytokeratin is shown in c and colocalization of PROKR2 and cytokeratin is shown in d. Amnion epithelial cells (Ae), amnion mesenchymal cells (Am), decidua (De), and cytotrophoblast (Ct). Bars ¼ 50 lm.

DUNAND ET AL.

Downloaded from www.biolreprod.org. FIG. 4. EG-VEGF, PROKR1, and PROKR2 expression during late pregnancy. A) A comparison of the levels of EG-VEGF mRNA expression in total FM collected from the second and third trimesters of pregnancy. B and C) Comparisons of PROKR1 and PROKR2 protein levels in total FM collected from the second and third trimesters of pregnancy. Student t-test was used to compare the value in each graph (*P , 0.05); ns, not significant.

(Supplemental Fig. S1Aa). Vimentin-positive cells could also be observed as seen by the presence of some chorion mesenchymal and/or decidual cells (Supplementary Figure S1Ab). Confirmation of EG-VEGF, PROKR1, and PROKR2 expression by isolated chorion trophoblast cells is reported in photographs (Supplemental Fig. S1, Bd–Bf). Isolated chorion cells were used to assess EG-VEGF effect on PGDH expression. Figure 7A shows that treatment of chorion cells with EG-VEGF (25 and 50 ng/ml) significantly increased the expression of PGDH. These data were substantiated by Western blot analysis (Fig. 7B). Besides their role in the regulation of PG metabolism, chorion cells are also involved in the control of membrane rupture in late pregnancy via fine regulation of the enzymes MMP-2 and -9, key MMPs that are abundantly present in the chorion layer. MMP-2 and -9 expression and activity have been shown to increase with the onset of labor [23]. Using zymography, we determined the effect of EG-VEGF on the activity of MMP-2 and -9 in the FMs. Figure 7C reports a

Figure 6A shows that PROKR1 and PROKR2 expression significantly decreased in the FM collected from laboring women compared to the nonlaboring ones. These observations were substantiated using Western blot analyses (Fig. 6, B and C). Evidence for EG-VEGF Direct Involvement in FM Protection in Late Pregnancy Chorion trophoblast cells are the main site of PG metabolism, playing a major role in the quiescence of the intrauterine tissues during late pregnancy. This role is due to the high level of expression of the PG-metabolizing enzyme, PGDH. To determine the effect of EG-VEGF on the expression of PGDH, we isolated primary chorion trophoblast cells. Supplemental Figure S1A (available online at www.biolreprod. org) shows representative photographs of isolated primary chorion trophoblast cells. The preparation contained mainly trophoblast cells, as illustrated by the cytokeratin 7 staining 6

Article 73

EG-VEGF CONTROLS HUMAN PARTURITION

Downloaded from www.biolreprod.org. FIG. 5. Effect of the labor process on EG-VEGF circulating and FM levels. A) Representative photographs of EG-VEGF in term FMs collected from nonlaboring (n ¼ 6) and laboring women (n ¼ 6). B) A Western blot analysis of EG-VEGF protein expression in nonlaboring versus laboring samples. C) The quantification of the Western blot analysis. Student t-test was used to compare the values in the graph (*P , 0.05). D) A comparison of circulating EGVEGF levels in nonlaboring versus laboring women. A total of 23 serum samples were analyzed (labor samples, n ¼ 12; nonlabor samples, n ¼ 11). EGVEGF levels were measured by ELISA. One-way ANOVA test was used followed by Kruskal-Wallis analysis of variance on ranks (*P , 0.05). Amnion epithelial cells (Ae), amnion mesenchymal cells (Am), decidua (De), cytotrophoblast (Ct), Bars ¼ 50 lm. E) A time course comparison of EG-VEGF secretion from FM explants collected from nonlaboring and laboring women. Data were obtained from four independent experiments performed in triplicate and represent the mean 6 SEM. The Student t-test was used to compare the values for each time point (**P , 0.001).

strated its roles in 1) placental angiogenesis, 2) trophoblast proliferation and survival, and 3) the control of premature trophoblast invasion of the maternal decidua [3, 10]. Furthermore, we showed that EG-VEGF circulating levels were elevated in preeclampsia and in intrauterine growth restriction, raising the possibility that EG-VEGF might contribute to these disorders [3, 10]. In term pregnancy, we and others have shown that EGVEGF expression decreased locally in the placenta with stable circulating levels found toward term (i.e., measuring both laboring and nonlaboring specimens). However, when considering only nonlaboring second- and third-trimester samples, we

representative zymogram that shows that EG-VEGF decreased the activities of MMP-2 and -9 in chorion cells isolated from nonlaboring women. Quantification of three independent experiments showed a significant decrease in MMP-2 and -9 activities upon EG-VEGF treatment (Fig. 7D). DISCUSSION An increasing number of reports in the literature suggest that EG-VEGF is involved in the reproduction processes, particularly those associated with pregnancy [3–5, 8, 10]. Our group has shown that EG-VEGF is highly expressed in human placenta during the first trimester of pregnancy and demon7

Article 73

DUNAND ET AL.

Downloaded from www.biolreprod.org. FIG. 6. Effect of labor on PROKR1 and PROKR2 expression. A) Representative photographs PROKR1 (a and b) and PROKR2 (c and d) expression in the FMs of nonlaboring and laboring women at term. Photographs in e and f are negative controls. Amnion epithelial cells (Ae), amnion mesenchymal cells (Am), decidua (De), and cytotrophoblast (Ct). Bars ¼ 50 lm. B and C) Western blot analysis of PROKR1 and PROKR2 expression in the FMs of nonlaboring and laboring women at term. Data represent the mean 6 SEM of at least four independent FMs collected from different women. The Student t-test was used to compare the values for each time point (*P , 0.05).

which might allow parturition processes to occur; and 3) EGVEGF directly controls the levels of expression of three key enzymes known to govern the mechanisms of parturition and are reportedly strongly associated with premature membrane rupture [12, 24]. More importantly, the observed local profile of expression of EG-VEGF during late pregnancy was correlated with its circulating levels in agreement with our observation of a significant increase throughout late pregnancy and an abrupt decrease with the onset of labor. These data suggest that deregulations in the levels of angiogenic factors such as EG-VEGF might well be associated with preterm deliveries during human pregnancy. An imbalance between angiogenic and antiangiogenic factors is wellknown to be associated with pregnancy pathologies such as preeclampsia [25–27], fetal growth restriction [28, 29],

found a significant increase in EG-VEGF levels, suggesting a role for this factor in late pregnancy. Based on this observation, we demonstrated a new source for EG-VEGF in the chorion layer of the FMs as well as a local role for this cytokine during late pregnancy. The present data demonstrate a significant fall in EG-VEGF expression in FM that correlates with the initiation of labor. Hence, we propose that EG-VEGF is a local intrauterine factor that is highly produced by the chorion trophoblast cells and acts in an autocrine and paracrine manner to ensure the quiescence of the intrauterine system during late pregnancy. Our assumptions are based on three main observations: 1) EG-VEGF and its receptors levels increased toward the end of pregnancy in a site where quiescence is required during this gestational period; 2) their levels significantly decreased in the FMs with the onset of labor, 8

Article 73

EG-VEGF CONTROLS HUMAN PARTURITION

placental abruption [30], and unexplained fetal death [31]. However, there is a paucity of information regarding the changes of these factors in patients destined to develop spontaneous preterm labor. Further studies are needed to determine whether the angiogenic factor EG-VEGF could be used as a biomarker of prematurity. EG-VEGF and its receptors were mainly localized to the chorion layer with a privileged expression of EG-VEGF and PROKR2 to the trophoblast layer and PROKR1 to the mesenchymal layer. These findings are in line with our first publication reporting the same localization of EG-VEGF and PROKR2 to the same cell types, that is, the syncytiotrophoblast in the placenta, and PROKR1 to another cell type, the cytotrophoblast [6]. Nevertheless, the immunohistochemistry staining showed that EG-VEGF is also present in the juxtaposed decidua. The presence of EG-VEGF in the decidua is not surprising because this was previously described at the mRNA level, especially during early pregnancy [4, 5, 8]. In the present study, we do not exclude the presence of EG-VEGF in the decidua because EG-VEGF could be both produced by this tissue and accumulated within it before getting into the

circulation. The decidua is a vascularized tissue and the FMs are not [32]. Hence, even if the FMs and especially the chorion constitute an important source of EG-VEGF, this molecule must accumulate in the decidua before getting into the circulation. The decidua might then be both a source and a niche for EG-VEGF protein. Previous established studies have demonstrated that PGDH plays a major role in the control of active PGs that reach to the decidua to trigger contractions. A decrease in PGDH expression is associated with premature rupture of membranes [33]. It has also been demonstrated that the activity of this enzyme is controlled by glucocorticoids and by inflammatory cytokines such as IL8 and IL10 [34]. Nevertheless, while these factors strongly regulate PGDH expression and activity, they act in paracrine manner because they are both produced by the placenta. In contrast, EG-VEGF appears to be a new local regulator of PGDH that is produced by the same cell type. More importantly, this is the first time where a regulation of PGDH by a placental angiogenic factor is reported and that the onset of labor is controlled by an angiogenic factor. 9

Article 73

Downloaded from www.biolreprod.org.

FIG. 7. Effects of EG-VEGF on PGDH expression and on MMP-2 and -9 activities in primary chorion trophoblast cultures. A) Immunoreactive PGDH in chorion trophoblast culture incubated in the absence or presence of EG-VEGF (25 and 50 ng/ml) for 24 h. Bars ¼ 20 lm. B) Representative Western blot analysis of EG-VEGF effects on PGDH expression in chorion trophoblast cells for 24 h. The histogram represents the mean 6 SEM of PGDH levels in the control and 50 ng/ml EG-VEGF. Three independent experiments were performed. The histogram represents the mean 6 SEM of PGDH levels in the control and 25 and 50 ng/ml EG-VEGF. Three independent experiments were performed (*P , 0.05). C) The effect of EG-VEGF on MMP-2 and -9 activities in chorion trophoblast cells. Student t-test was used to compare the treated conditions to the control (*P , 0.05). Upper zymogram shows a representative zymography of cells treated or not with EG-VEGF for 24 h at indicated concentrations. D) The lower histogram shows densitometric analysis of zymographs of three independent experiments. Each bar represents the amount of MMP-2 or -9 as arbitrary units. Data are presented as the mean 6 SEM. Student t-test was used to compare the values for each time point (*P , 0.05 and **P , 0.001). CTL, control.

DUNAND ET AL.

A confirmation of the proposed role for EG-VEGF in FM protection is its ability to inhibit the activities of MMP-2 and -9 in the FM explants. The inhibition of MMPs activity in the FM substantiates our previous observation of an inhibition of these enzymes by EG-VEGF in the placenta during early pregnancy [10]. Because EG-VEGF belongs to the prokineticin family, with a described role in intestinal contraction [2], it was suggested that EG-VEGF might play a role in the contraction of the decidua at the time of labor. However, repeated injections of EG-VEGF to gravid mice during late gestation did not trigger any contractions, and mice proceeded normally to term [13]. These findings confirm the notion that EG-VEGF is rather a protecting than a constrictor factor during late pregnancy. In conclusion, we report for the first time a specific and local expression of EG-VEGF and its receptors in human FM, with a local increase toward the end of gestation. We have also reported an increase of its circulating levels toward term and proposed a role and a mechanism of action for this protein within the FM. These data are of enormous clinical importance because treatments using recombinant EG-VEGF during late pregnancy could be proposed to maintain suspected premature deliveries.

14.

15.

16.

17. 18.

19. 20.

ACKNOWLEDGMENT We thank Ms. Vanessa Garnier for her technical help.

22.

REFERENCES

23.

1. LeCouter J, Kowalski J, Foster J, Hass P, Zhang Z, Dillard-Telm L, Frantz G, Rangell L, DeGuzman L, Keller GA, Peale F, Gurney A, et al. Identification of an angiogenic mitogen selective for endocrine gland endothelium. Nature 2001; 412:877–884. 2. Li M, Bullock CM, Knauer DJ, Ehlert FJ, Zhou QY. Identification of two prokineticin cDNAs: recombinant proteins potently contract gastrointestinal smooth muscle. Mol Pharmacol 2001; 59:692–698. 3. Brouillet S, Hoffmann P, Feige JJ, Alfaidy N. EG-VEGF: a key endocrine factor in placental development. Trends Endocrinol Metab 2012; 23: 501–508. 4. Evans J, Catalano RD, Brown P, Sherwin R, Critchley HO, Fazleabas AT, Jabbour HN. Prokineticin 1 mediates fetal-maternal dialogue regulating endometrial leukemia inhibitory factor. FASEB J 2009; 23:2165–2175. 5. Evans J, Catalano RD, Morgan K, Critchley HO, Millar RP, Jabbour HN. Prokineticin 1 signaling and gene regulation in early human pregnancy. Endocrinology 2008; 149:2877–2887. 6. Hoffmann P, Feige JJ, Alfaidy N. Expression and oxygen regulation of endocrine gland-derived vascular endothelial growth factor/prokineticin-1 and its receptors in human placenta during early pregnancy. Endocrinology 2006; 147:1675–1684. 7. Hoffmann P, Feige JJ, Alfaidy N. Placental expression of EG-VEGF and its receptors PKR1 (prokineticin receptor-1) and PKR2 throughout mouse gestation. Placenta 2007; 28:1049–1058. 8. Maldonado-Perez D, Evans J, Denison F, Millar RP, Jabbour HN. Potential roles of the prokineticins in reproduction. Trends Endocrinol Metab 2007; 18:66–72. 9. Lin DC, Bullock CM, Ehlert FJ, Chen JL, Tian H, Zhou QY. Identification and molecular characterization of two closely related G protein-coupled receptors activated by prokineticins/endocrine gland vascular endothelial growth factor. J Biol Chem 2002; 277:19276–19280. 10. Hoffmann P, Saoudi Y, Benharouga M, Graham CH, Schaal JP, Mazouni C, Feige JJ, Alfaidy N. Role of EG-VEGF in human placentation: physiological and pathological implications. J Cell Mol Med 2009; 13: 2224–2235. 11. Brouillet S, Murthi P, Hoffmann P, Salomon A, Sergent F, De Mazancourt P, Dakouane-Giudicelli M, Dieudonne MN, Rozenberg P, Vaiman D, Barbaux S, Benharouga M, et al. EG-VEGF controls placental growth and survival in normal and pathological pregnancies: case of fetal growth restriction (FGR). Cell Mol Life Sci 2013; 70:511–525. 12. Challis JRG, Matthews SG, Gibb W, Lye SJ. Endocrine and paracrine regulation of birth at term and preterm. Endocr Rev 2000; 21:514–550. 13. Lannagan TR, Wilson MR, Denison F, Norman JE, Catalano RD, Jabbour HN. Prokineticin 1 induces a pro-inflammatory response in murine fetal

24. 25.

26.

27.

28.

29.

30.

31.

32. 33.

34.

10

Article 73

Downloaded from www.biolreprod.org.

21.

membranes but does not induce preterm delivery. Reproduction 2013; 146:581–591. McLaren J, Taylor DJ, Bell SC. Increased concentration of pro-matrix metalloproteinase 9 in term fetal membranes overlying the cervix before labor: implications for membrane remodeling and rupture. Am J Obstet Gynecol 2000; 182:409–416. Vadillo-Ortega F, Gonzalez-Avila G, Furth EE, Lei H, Muschel RJ, Stetler-Stevenson WG, Strauss JF III. 92-kd type IV collagenase (matrix metalloproteinase-9) activity in human amniochorion increases with labor. Am J Pathol 1995; 146:148–156. Pomini F, Patel FA, Mancuso S, Challis JR. Activity and expression of 15hydroxyprostaglandin dehydrogenase in cultured chorionic trophoblast and villous trophoblast cells and in chorionic explants at term with and without spontaneous labor. Am J Obstet Gynecol 2000; 182:221–226. Challis JR, Patel FA, Pomini F. Prostaglandin dehydrogenase and the initiation of labor. J Perinat Med 1999; 27:26–34. Patel FA, Challis JR. Cortisol/progesterone antagonism in regulation of 15-hydroxysteroid dehydrogenase activity and mRNA levels in human chorion and placental trophoblast cells at term. J Clin Endocrinol Metab 2002; 87:700–708. Challis JR, Lye SJ, Gibb W. Prostaglandins and parturition. Ann N Y Acad Sci 1997; 828:254–267. Alfaidy N, Li W, MacIntosh T, Yang K, Challis J. Late gestation increase in 11beta-hydroxysteroid dehydrogenase 1 expression in human fetal membranes: a novel intrauterine source of cortisol. J Clin Endocrinol Metab 2003; 88:5033–5038. Alfaidy N, Xiong ZG, Myatt L, Lye SJ, MacDonald JF, Challis JR. Prostaglandin F2alpha potentiates cortisol production by stimulating 11beta-hydroxysteroid dehydrogenase 1: a novel feedback loop that may contribute to human labor. J Clin Endocrinol Metab 2001; 86:5585–5592. Kliman HJ, Nestler JE, Sermasi E, Sanger JM, Strauss JF III. Purification, characterization, and in vitro differentiation of cytotrophoblasts from human term placentae. Endocrinology 1986; 118:1567–1582. Xu P, Alfaidy N, Challis JR. Expression of matrix metalloproteinase (MMP)-2 and MMP-9 in human placenta and fetal membranes in relation to preterm and term labor. J Clin Endocrinol Metab 2002; 87:1353–1361. Challis JR, Lye SJ, Gibb W, Whittle W, Patel F, Alfaidy N. Understanding preterm labor. Ann N Y Acad Sci 2001; 943:225–234. Levine RJ, Maynard SE, Qian C, Lim KH, England LJ, Yu KF, Schisterman EF, Thadhani R, Sachs BP, Epstein FH, Sibai BM, Sukhatme VP, et al. Circulating angiogenic factors and the risk of preeclampsia. N Engl J Med 2004; 350:672–683. Torry DS, Wang HS, Wang TH, Caudle MR, Torry RJ. Preeclampsia is associated with reduced serum levels of placenta growth factor. Am J Obstet Gynecol 1998; 179:1539–1544. Venkatesha S, Toporsian M, Lam C, Hanai J, Mammoto T, Kim YM, Bdolah Y, Lim KH, Yuan HT, Libermann TA, Stillman IE, Roberts D, et al. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med 2006; 12:642–649. Crispi F, Llurba E, Dominguez C, Martin-Gallan P, Cabero L, Gratacos E. Predictive value of angiogenic factors and uterine artery Doppler for earlyversus late-onset pre-eclampsia and intrauterine growth restriction. Ultrasound Obstet Gynecol 2008; 31:303–309. Savvidou MD, Noori M, Anderson JM, Hingorani AD, Nicolaides KH. Maternal endothelial function and serum concentrations of placental growth factor and soluble endoglin in women with abnormal placentation. Ultrasound Obstet Gynecol 2008; 32:871–876. Signore C, Mills JL, Qian C, Yu K, Lam C, Epstein FH, Karumanchi SA, Levine RJ. Circulating angiogenic factors and placental abruption. Obstet Gynecol 2006; 108:338–344. Espinoza J, Chaiworapongsa T, Romero R, Kim YM, Kim GJ, Nien JK, Kusanovic JP, Erez O, Bujold E, Goncalves LF, Gomez R, Edwin S. Unexplained fetal death: another anti-angiogenic state. J Matern Fetal Neonatal Med 2007; 20:495–507. Jones CJ, Jauniaux E. Ultrastructure of the materno-embryonic interface in the first trimester of pregnancy. Micron 1995; 26:145–173. Rizek RM, Watson CS, Keating S, Tai HH, Challis JR, Bocking AD. 15Hydroxyprostaglandin dehydrogenase protein expression in human fetal membranes with and without subclinical inflammation. Reprod Sci 2007; 14:260–269. Pomini F, Caruso A, Challis JR. Interleukin-10 modifies the effects of interleukin-1beta and tumor necrosis factor-alpha on the activity and expression of prostaglandin H synthase-2 and the NAD þ-dependent 15hydroxyprostaglandin dehydrogenase in cultured term human villous trophoblast and chorion trophoblast cells. J Clin Endocrinol Metab 1999; 84:4645–4651.