Glycodelin-a as a Modulator of Trophoblast Invasion

background: Trophoblast invasion is crucial to placentation. The relationship between decidual glycodelin-A and trophoblast invasion is not known. methods: Invasiveness of First trimester extravillous cytotrophoblast-1 (TEV-1) cell line, TEV-1, cells was determined by trans-well invasion assay. The gene expression, protein secretion and activities of the matrix metalloproteinase (MMP)-2 and-9, urokinase plasminogen activator (uPA), tissue inhibitor of metalloproteinase (TIMP)-1 and-2 and plasminogen activator inhibitor (PAI-1) of glycodelin-A-treated cells were measured by quantitative PCR, ELISA and gel zymography, respectively. results: Glycodelin-A bound to TEV-1 cells. At a concentration of 1 mg/ml, glycodelin-A, but not other glycodelin isoforms, suppressed the invasion of TEV-1 cells. The effect was glycosylation-dependent and was associated with reduction (P , 0.05) of MMP2, MMP9 and uPA activities in the conditioned medium from the treated culture. Glycodelin-A treatment suppressed the amount of MMP2 protein in the conditioned medium (P , 0.05) and MMP2 mRNA in the cells (P , 0.05), but did not affect that of MMP9. Glycodelin-A also significantly reduced the expression, secretion and activity of uPA (P , 0.05). The treatment did not affect the expression of TIMP-1, TIMP-2 or PAI-1, cell proliferation or survival of the cells. conclusions: Glycodelin-A inhibits the invasion of extravillous cytotrophoblasts mainly by suppressing the activity of MMP2 and MMP9 in a glycosylation-dependent fashion.


Introduction
The invasion of the trophoblasts into the endometrium is crucial to placentation (Staun-Ram and Shalev, 2005;Lunghi et al., 2007).The key cells in placentation are the cytotrophoblasts (Staun-Ram and Shalev, 2005).After implantation, the cytotrophoblasts differentiate into the villous cytotrophoblasts and the extravillous cytotrophoblasts (EVCTs).The villous cytotrophoblasts fuse to form the terminally differentiated multinucleated syncytiotrophoblasts responsible for fetomaternal exchange and production of hormones such as steroids and human chorionic gonadotrophin.The invasive EVCTs form migratory cell columns invading the endometrium (Lunghi et al., 2007).Although some of the EVCTs invade the decidual stroma and the superficial myometrium to become the interstitial EVCTs, others invade the maternal spiral arteries to form the endovascular EVCTs.
Trophoblast invasion is associated with remodeling and degradation of the extracellular matrix (ECM) of the endometrial tissue.Members of the matrix metalloproteinase (MMP) family and the urokinase plasminogen activator (uPA) system are the key molecules involved in matrix degradation (Lala and Chakraborty, 2003;Cohen et al., 2006).The MMPs are zinc-dependent proteolytic enzymes synthesized as secreted or transmembrane proenzymes (Cohen et al., 2006), converted to active enzymes by removal of the amino-terminal propeptide.The synthesis and activation of two of the family members, matrix metalloproteinase-2 (MMP2) and matrix metalloproteinase-9 (MMP9), are required for trophoblast invasion (Isaka et al., 2003).MMP2 and MMP9 are 72 and 92-kDa gelatinases, respectively, produced by the cytotrophoblasts.The expression and activation of these enzymes are essential to invasion by EVCTs (Staun-Ram et al., 2004;Ferretti et al., 2007).The activities of MMPs are inhibited by † Contribute equally to the article.
& The Author 2009.Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology.All rights reserved.
For Permissions, please email: journals.permissions@oxfordjournals.org the tissue inhibitors of metalloproteinases (TIMPs) that bind specifically to the conserved zinc-binding site of MMPs; thus TIMP-1 preferentially binds to MMP9 and TIMP-2 to MMP2 (Cohen et al., 2006).Human trophoblast cells also produce urokinase plasminogen activator (uPA) (Staun-Ram and Shalev, 2005), which is a serine proteinase converting plasminogen into active plasmin.Plasmin plays an essential role in trophoblast invasion through degradation of the ECM and activation of the MMPs proenzymes (Staun-Ram and Shalev, 2005).The activity of uPA is balanced by the plasminogen activator inhibitor-1 (PAI-1) (Vassalli et al., 1991).The simultaneous expression of matrix degrading enzymes and their inhibitors in human trophoblasts suggests that the balance between these components regulates the invasive activity of the trophoblast cells (Cohen et al., 2006).
Trophoblast invasion is similar to tumor cell invasion in many aspects (Ferretti et al., 2007).Both processes involve attachment of the cells to, and degradation of, ECM components; cellular proliferation; and migration through connective tissue.However, unlike tumor invasion, trophoblast invasion is spatially and temporally controlled, and confined to the inner third of the myometrium (Ferretti et al., 2007).Dysregulation of trophoblast invasion is associated with various pathologies, including gestational trophoblastic diseases (Shih Ie and Kurman, 2002) and pre-eclampsia (Goldman-Wohl and Yagel, 2002), which are associated with significant maternal and fetal death (Norwitz et al., 2001).Despite intensive research, our understanding of the mechanisms controlling trophoblast invasion in normal and abnormal pregnancy, are still poorly understood.
Glycodelin is a glycoprotein with four isoforms having different glycosylation patterns (Seppala et al., 2002(Seppala et al., , 2007)).The four isoforms are glycodelin-A (amniotic fluid isoform), glycodelin-S (seminal plasma isoform), glycodelin-F (follicular fluid isoform) and glycodelin-C (cumulus matrix isoform) (Chiu et al., 2007a, b;Seppala et al., 2007).In the uterus, glycodelin-A is mainly synthesized in the decidua and the secretory endometrial glands.During early pregnancy, the concentration of glycodelin-A rises in the endometrium and peaks in the decidua from the 6th to 12th week of gestation (Julkunen et al., 1985).Glycodelin-A has been related to fetomaternal defense and endometrial receptivity (Seppala et al., 2002;Yeung et al., 2006).However, its role in trophoblast invasion is unknown.Decrease in maternal serum glycodelin is associated with early spontaneous miscarriage and recurrent miscarriage (Seppala et al., 2002).
Trophoblast invasion peaks in the first trimester of pregnancy (Bischof et al., 2000).The coincidental rise of glycodelin-A concentration in the endometrium with trophoblast invasion within this period suggests a role of glycodelin-A in this important process of placentation.We hypothesize that glycodelin-A modulates the invasion of EVCTs.The first objective of this study was to investigate the role of glycodelin-A on the invasiveness of an EVCT cell line, TEV-1.The second objective was to explore the possible regulatory mechanisms of glycodelin-A on trophoblast invasion.These mechanisms included action of glycodelin-A on MMPs and uPA production of EVCTs.

Materials and Methods
The study protocol was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster.

Cell culture
The immortalized TEV-1 cell line has been developed by transfection of human papilloma virus pLXSN-E6/E7 open-reading frames into primary culture of first trimester normal extravillous trophoblast cells from human placenta (Feng et al., 2005).TEV-1 cells retain most of the characteristics of EVCT (Feng et al., 2005), including the expression of immunoreactivities of two EVCT-specific markers, CD9 and human leukocyte antigen G1 (Hirano et al., 1999;Shiverick et al., 2001).Cells at passage 30 -38 were used in this study.They were cultured in DMEM/F12 medium (Gibco, Carlsbad, CA, USA) containing 10% heat-inactivated fetal bovine serum (Gibco), 1% penicillin and streptomycin in a humidified atmosphere of 5% CO 2 in air at 378C.

Purification of glycodelin isoforms
Glycodelin-A, -S, -F and -C were purified from first trimester amniotic fluid, seminal plasma, follicular fluid and cumulus matrix of human, respectively, as described (Chiu et al., 2007a).Briefly, amniotic fluid, seminal plasma and cumulus matrix were diluted with tris-buffered saline (TBS) containing 0.1% triton X-100 in a ratio of 1:3 -1:5 before loading onto a monoclonal anti-glycodelin antibody (clone F43-7F9) Sepharose column, which was then washed successively by TBS, 1 M NaCl with 1% isopropanol, 10 mM ammonium acetate with 0.1% isoproponal, pH 5 and TBS.The bound glycodelin was eluted with 0.1% trifluoroacetic acid and dialyzed against 100 mM sodium phosphate buffer (pH 7.2).Anion exchange chromatography was used to further purify glycodelin-S and -C (Chiu et al., 2007a).Glycodelin-F from human follicular fluid was purified by a series of chromatographic steps involving the use of Hi-Trap blue, protein-G, Con-A sepharose columns (GE Healthcare, Uppsala, Sweden), Amicon-10 concentrator (Amicon Inc.Beverly, CA, USA), Mono-Q and Superose columns (Chiu et al., 2003b).Deglycosylated glycodelin was prepared by denaturation of glycodelin-A in 0.1% b-mecaptoethanol before incubation with 0.5 mU PNGase F at 378C for 24 h.The deglycosylated protein was precipitated, redissolved in phosphate buffered saline, and purified by gel filtration chromatography as described (Chiu et al., 2003a).The purities of the native and deglycosylated glycodelins were confirmed by sodium dodecyl sulfatepolyacrylamide gel electrophoresis showing single band with the expected molecular size (Chiu et al., 2003a).The identity of the glycodelin was further confirmed by mass spectrometry as described (Lee et al., 2009).The concentrations of the purified proteins were determined by a commercial protein assay kit (Bio-Rad, Hercules).

Binding of glycodelin isoforms to human trophoblasts
Glycodelin-A, -F, -S, and -C were labeled with Alexa-594 using the Alexa Fluor w 594 Microscale Protein Labeling Kit (Molecular Probes, Carlsbad, CA, USA) according to manufacturer's instruction.TEV-1 cells were grown to semi-confluence in 24-well culture plates.They were then incubated with Alexa-594 conjugated glycodelin (1 mg/ml) for 3 h at 378C in an atmosphere of 5% CO 2 in air, gently washed thrice and observed under a fluorescence microscope.Cells incubated with labeled deglycosylated glycodelin or labeled ovalbumin was used as negative controls.

Effect of glycodelin-A and -F on trophoblast invasiveness
The effects of glycodelin-A and -F on TEV-1 and JAR invasiveness were tested using the quantitative CytoSelect TM 96-well Cell Invasion Assay (Cell Biolabs, San Diego, CA, USA) according to the manufacturer's instructions.In brief, the invasion assay was performed in an invasion chamber consisting of a cell culture insert placed in a tissue culture well.The insert had a polycarbonate membrane of pore sizes 8 mm coated with a thin layer of reconstituted basement membrane matrix of proteins.TEV-1 cells suspended in serum free medium at a density of 5 Â 10 5 cells/ ml were seeded into the culture inserts and were treated with 0.01, 0.1 or 1 mg/ml of glycodelin-A or -F.TEV-1 cells without glycodelin treatment were used as a control.The cells were allowed to invade the reconstituted basement membrane matrix for 24 h.Those cells that had invaded through the membrane were stained with Cell Stain (Millipore, Billerica, CA, USA) and viewed under a microscope.To quantify the invasiveness of the cells with or without glycodelin treatment, the cells on the upper surface of the membrane were removed by swabbing.The invasive cells at the bottom surface of the membrane were dislodged by the cell detachment solution provided with the assay kit, lyzed and quantified with CyQuant w GR Dye using a fluorescence plate reader with excitation at 480 nm and emission at 520 nm.Results were expressed as percentage of fluorescence intensity relative to the no treatment control.

Effect of glycodelin-A on TEV-1 cell viability
The XTT cell viability assay (Roche Diagnostics Co., USA) was used to investigate the effect of glycodelin-A (0.01, 0.1 or 1 mg/ml) on cell proliferation, according to the manufacturer's instructions.The colorimetric immunoassay measures the cleavage of a tetrazolium salt XTT into a detectable formazan end-product, the amount of which is proportional to the population size of the metabolically active cells.Briefly, 3 Â 10 4 cells/100 ml were incubated with different concentrations of glycodelin-A for 12 or 24 h.XTT labeling mixture (0.3 mg/ml) was freshly prepared by mixing XTT labeling reagent with electron coupling reagent in a ratio of 50:1.Fifty micro-liter of the labeling mixture was added to the culture 12 h before the end of the incubation period.The absorbance at the end of the incubation was measured at 450 nm with l correction at 575 nm.
The data from the XTT assay was confirmed by fluorometric Cyquant NF Cell Proliferation Assay Kit (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's instructions.This assay measured cellular DNA content through fluorescence dye binding after plasma membrane permeabilization.Briefly, 1 Â 10 4 cells/100 ml were incubated with different concentrations of glycodelin-A for 24 h in a microplate.The spent culture media were aspirated and 100 ml of dye binding solution was added to the cells.The fluorescence intensity at the end of the incubation was measured at excitation of 485 nm and emission of 530 nm.Cell viability was presented as percentage suppression using the following equation:

Reverse transcription and quantitative PCR analysis of mRNA expression
TEV-1 cells were cultured with or without glycodelin-A (0.01, 0.1 or 1 mg/ ml) treatment as above.The QuickPrep RNA extraction kit (GE healthcare) was used to extract total RNA from the cells, which was reverse transcribed with the use of the TaqMan reverse transcription reagent kit (Applied Biosystems, Foster City, CA, USA) and multiscript reverse transcriptase.The resulting cDNA were subjected to quantitative PCR analysis of MMP2 (Hs01548724_m1), MMP9 (Hs00957555_m1), TIMP-1 (Hs00171558_m1), TIMP-2 (Hs00234278_m1), uPA (Hs00170182_m1) and PAI-1 (Hs01126606_m1) using an ABI 7500 Sequence Detector (Applied Biosystems).Multiplex PCR was performed in a 20 ml reaction mixture containing 5 ml of sample DNA; 10 ml of 2X TaqMan Universal PCR Master Mix; 1 ml of eukaryotic 18S rRNA internal control (Hs99999901_s1) and 1 ml of 20X Gene Expression Assay for the targets (Applied Biosystems).18S was used as internal control for sample loading.Water was used as the no template control.The thermal cycling condition was as follow: an activation step at 958C for 10 min, followed by 45 cycles of denaturation (948C for 15 s), annealing (608C for 30 s) and amplification (768C for 30 s).Fluorescence data was collected during the annealing step.The relative gene expression levels were determined using the threshold cycle (C T ) method (2 24 4CT method) with reference to the endogenous 18S control.

Effect of glycodelin-A on MMP2, MMP9 and uPA secretion
To investigate the effect of glycodelin-A (0.01, 0.1 or 1 mg/ml) on protein secretion of MMP2, MMP9 and uPA by the TEV-1 cells, MMP2 and MMP9 ELISA detection kit (Calbiochem, San Diego, CA, USA) and uPA ELISA detection kit (American Diagnostica, Stamford, CT, USA) were used.The MMP kits detected free MMPs as well as TIMP bound MMPs.The uPA kit also detected total uPA, including pro-uPA, receptor bound uPA and uPA complexed with PAI-1.In brief, the cells were cultured with or without glycodelin-A treatment for 24 h.Duplicated samples of the conditioned media collected and biotinylated detector monoclonal antibody were added into the 96-well assay plates coated with capturing antibody.After incubation for 3 h for MMPs or overnight for uPA, unbound material was washed away.Horseradish peroxidase-conjugated streptavidin was then added.Color development was performed using tetra-methylbenzidine as the chromogen.The reaction was stopped by the addition of 2 M sulfuric acid (50 ml/well) and the absorbance was measured at 492 nm in an ELISA plate reader (Infinite F200; TECAN, Ma ¨nnedorf, Switzerland).Results were expressed as absorbance per sample.

Effect of glycodelin-A on MMPs and TIMPs activity
Gelatin zymography was performed to quantify the biological activity of MMP2 and MMP9.The conditioned culture media from TEV-1 cells treated with different concentrations of glycodelin-A (0.01, 0.1 and 1 mg/ml) were collected.The components in the medium were resolved in 8% polyacrylamide gel containing 0.1 mg/ml gelatin (Sigma-Aldrich) to determine MMP2 and MMP9 activities under non-reducing conditions.The loadings were normalized by measuring the total protein concentration of the cell pellets.After electrophoresis, the gel was rinsed twice in the renaturing buffer (50 mM Tris-HCl with 2.5% Triton X-100), incubated in the developing buffer (50 mM Tris -HCl, 5 mM CaCl 2 , at pH 7.5) at 378C for 24 h, stained with 0.2% (w/v) Coomassie Brilliant Blue R-250 (Sigma-Aldrich) for 30 min and destained in methanol: acetic acid: water (3:1:6).Proteolytic activities were seen as area with reduced protein staining.Identification of gelatinase band was based on the molecular sizes.Reverse zymography in 10% polyacrylamide gel containing 0.1 mg/ml of gelatin and 30 ng/ml of MMP2 or MMP9 (Calbiochem) was used to quantify the biological activity of TIMP-2 and TIMP-1, respectively.Gelatinase inhibitory activity of TIMP-1 and TIMP-2 appeared dark blue bands against a clear digested background.The bands were quantified using the AlphaImager HP gel documentation system installed with the AlphaEase FC software.The activities were expressed in relation to the control cells without treatment.

Effect of glycodelin-A on uPA activities
The effects of glycodelin-A on uPA activity was tested using the quantitative urokinase activity assay kit (Calbiochem) according to the manufacturer's instructions with a slight modification.In brief, the kit utilizes the ability of urokinase to digest the synthetic substrate Glutaryl-Gly-Arg-7-amino-4-methylcoumarin (AMC).The amount of digested AMC is then determined fluorometrically.Ten micro-liter of substrate (10Â) was added to the TEV-1 cells that had been treated with different concentrations of glycodelin-A (0.01, 0.1 and 1 mg/ml) for 24 h.After 180 min of incubation with the substrate at 378C, the fluorescence was measured with an excitation wavelength of 380 nm and an emission wavelength of 505 nm in an ELISA plate reader.The activities were expressed in relation to the control cells without treatment.

Statistical analyses
All the data were expressed as mean and standard error of the mean (SEM), analyzed by statistical software (SigmaPlot 10.0 and SigmaStat 2.03; Jandel Scientific, San Rafael, CA, USA).For all experiments, the nonparametric analysis of variance on rank test for multiple comparisons followed by the Mann -Whitney U-test was used.A probability value ,0.05 was considered to be statistically significant.

Binding of glycodelin isoforms to TEV-1
Using the present labeling protocol, each mole of glycodelin was labeled with 2.78 moles of Alexafluor-594 dye.Fluorescent labeled glycodelin-A and -F bound to the TEV-1 cells (Fig. 1).Glycodelin-A treatment produced a stronger signal when compared with that of glycodelin-F at the same concentration.On the other hand, fluorescent-labeled glycodelin-S and -C, deglycosylated glycodelin and ovalbumin did not bind to the TEV-1 cells (Fig. 1).Glycodelin-A suppressed invasion of the TEV-1 cells (see below).The Alexa-labeled glycodelin-A had similar invasion suppressive activity on TEV-1 as their native counterpart (data not shown).

Effects of glycodelin-A and -F on invasiveness of TEV-1 cells
Glycodelin-A at concentrations of 0.1 and 1 mg/ml significantly (P , 0.05) reduced the invasiveness of TEV-1 cells when compared with the no treatment control (Fig. 2).At 1 mg/ml, it inhibited 47.7 + 8.7% of the TEV-1 invasion.This effect was not due to actions of glycodelin-A on viability or proliferation of TEV-1 cells, as the treatment with glycodelin-A did not affect these growth parameters at any tested concentrations, when compared with the control (Table I).Glycodelin-F or deglycosylated glycodelin treatment did not affect the invasion of TEV-1 cells (Fig. 2).The suppressive effect of glycodelin-A on trophoblast invasion was further confirmed by repeating the experiment using JAR cell line.The data showed that glycodelin-A inhibited the invasion of JAR cells (Fig. 3).

Effects of glycodelin-A on production of MMPs and TIMPs
Glycodelin-A did not affect the mRNA expression of MMP9, but significantly reduced that of MMP2 (Fig. 4A) in TEV-1 cells.The mRNA expression of MMP2 was suppressed by 25.0 + 4.7% after treatment with glycodelin-A at a concentration of 1 mg/ml.Glycodelin-A suppressed the MMP2 and MMP9 activities in the conditioned media in a dose-dependent manner (Fig. 4B).At 1 mg/ml of glycodelin-A, the MMP2 and MMP9 activities were significantly inhibited by 48.2 + 11.2% and 76.0 + 2.0%, respectively, when compared with the control (P , 0.05).The concentration of MMP2 protein, but not MMP9, in the culture media was decreased by glycodelin-A treatment (Fig. 4C).At 1 mg/ml of glycodelin-A, the secretion of MMP2 protein was reduced by 37.6 + 13.1% (P , 0.05).

Effects of glycodelin-A on the production of uPA and PAI-1
Glycodelin-A significantly suppressed the mRNA level of uPA, but not that of PAI-1 (Fig. 6A).The concentration of uPA in the conditioned media was significantly (P , 0.05) decreased from 59.9 + 10.9 to 38.9 + 11.9 pg/ml (36.13 + 8.78%) after treatment with glycodelin-A at a concentration of 1 mg/ml (Fig. 6B).uPA activity was also inhibited by glycodelin-A.At a concentration of 1 mg/ml, glycodelin-A suppressed uPA activity by 36.6 + 7.0% (Fig. 6B).

Discussion
Human trophoblast invasion is believed to be controlled by a balance between the actions of the stimulatory and inhibitory factors of both trophoblastic and uterine origins (Bischof et al., 2000).A number of cytokines and growth factors present in the peri-implantation uterine milieu, such as interleukin-12 and transforming growth factor b, suppress trophoblast invasion through inhibition of matrixdegrading enzymes (Graham et al., 1994;Karmakar et al., 2004).Glycodelin-A is abundantly expressed in the decidua but not in the trophoblast (Seppala et al., 2002).Studies on the role of glycodelin-A in trophoblast function are limited.This is the first report demonstrating a suppressive action of glycodelin-A on trophoblast invasion.This was observed at the concentrations of glycodelin-A similar to those present in late secretory phase endometrium (Julkunen et al., 1986) and early pregnancy deciduas (Julkunen et al., 1985).Results from XTT and CyQuant w NF cell proliferation assay suggest that the effect was not due to actions of glycodelin-A on the viability or proliferation of TEV-1 cells.This suggestion needs to be confirmed with thymidine or bromouridine uptake experiments because the XTT assay determines only the steady state metabolic content whereas the latter measures only the DNA content of the cell population.The data add glycodelin-A to the group of molecules that regulates the expression/activity of matrix-degrading enzymes.
Trophoblast-derived proteinases, including the MMPs and the serine proteases, are essential for placental invasion.MMPs are believed to be the dominant system in trophoblast invasion (Librach TEV-1 cells were incubated with 1 mg/ml Alexa-594 labeled glycodelin-A, -F, -S, -C, deglycosylated glycodelin-A or ovalbumin.The results shown are representative of three replicate experiments.et al., 1991).Their production and activity in trophoblasts peak in the first trimester (Xu et al., 2000;Staun-Ram et al., 2004) coinciding with the maximal invasive behavior of trophoblast in vivo.Thus, regulation of MMPs is crucial to the invasion process.Our results showed that maternally-derived glycodelin-A bound to the TEV-1 cells, and suppressed their invasiveness significantly.This inhibitory activity is likely mediated in part through modulating the activity of trophoblastderived MMP2 and MMP9.
In this study, glycodelin-A suppressed MMP2 and MMP9 activities of TEV-1 cells.The mRNA expression and protein of MMP2 were decreased after glycodelin-A treatment, indicating that glycodelin-A inhibits the transcription and secretion of MMP2.The same treatment did not affect the mRNA expression or secretion of MMP9, though the activity of the enzyme in the conditioned media was decreased.The ELISA assay used in this study measured total MMP9 i.e. proenzyme as well as active enzyme.Thus the action of glycodelin-A on MMP9 bioactivity is likely mediated via inhibition of the conversion of pro-MMP9 to the active form.MMPs are secreted as zymogen (Cohen et al., 2006), subsequently activated through cleavage and removal of part of the molecule by other proteinases.Pro-MMP9

Table I Effect of glycodelin-A on proliferation of TEV-1
Glycodelin-A mg/ml Suppression Index (%) Cell proliferation was determined by the XTT and Cyquant NF cell proliferation assay on TEV-1 cells incubated with different concentrations of glycodelin-A for 12 or 24 h.Data are expressed as percentage of control without glycodelin-A treatment (mean + SEM, N = 5).Suppression index (%) = (Absorbance of glycodelin treated cells 2 Absorbance of blank)/(Absorbance of control 2 Absorbance of blank) Â 100%.
can be activated by uPA-mediated activation of MMP3 (stromelysin-1); uPA converts plasminogen to plasmin, an efficient activator of pro-MMP3, and activated MMP3 is a documented activator of pro-MMP9 (Hahn-Dantona et al., 1999;Ramos-DeSimone et al., 1999).MMP3 expression has been detected in extravillous trophoblast cells and is reduced in preeclamptic women (Reister et al., 2006).The actions of glycodelin-A on expression of MMP3 in TEV-1 cells remain to be investigated.MMP2 and MMP9 act on native collagen IV to facilitate cell invasion (Cohen et al., 2006) and are believed to be the key enzymes participating in trophoblast invasion (Isaka et al., 2003;Staun-Ram et al., 2004).Which of these two MMPs is the dominant gelatinase in the process is under debate.Some reports suggested that MMP9 plays a more important role in trophoblast invasion (Librach et al., 1991;Bischof et al., 2003), although other reported suggest that the action of MMP2 is more pronounced (Bjorn et al., 2000;Isaka et al., 2003).The discrepancy may be due to the use of trophoblast tissue from different weeks of gestation, as gelatinases are differentially expressed in the first trimester trophoblast cells depending on the gestational age (Xu et al., 2000).MMP2 was the main gelatinase in trophoblasts before 9 weeks of gestation whereas MMP9 was more pronounced thereafter.Similar differential expression patterns of MMP9 and MMP2 have also been reported to be involved in trophoblast invasion after 9 weeks of gestation (Staun-Ram et al., 2004).A recent report showed that the expression of MMP9, but not MMP2, was higher in the third trimester cytotrophoblast than in the first trimester trophoblast (Pang et al., 2008).Trophoblast cells express uPA and its receptor, uPAR (Lala and Chakraborty, 2003).Like MMPs, uPA is secreted as an inactive precursor activated through cleavage by plasmin and by binding to membrane-bound uPAR.Activated uPA cleaves plasminogen into plasmin, thus creating a positive feedback loop for its own activation.Unlike uPA, plasmin has a broad spectrum of substrates (Vassalli et al., 1991).Apart from degrading ECM, plasmin also activates MMPs from their proenzymes.Glycodelin-A treatment reduced the mRNA expression and protein secretion/activity of uPA in TEV-1 cells.Zini and co-workers has demonstrated that the uPA proteolytic activities in first trimester trophoblast are greater than that in term trophoblast (Zini et al., 1992).It is inconclusive to say whether the MMPs or the uPA system is dominant in our TEV-1 cells, but the literature provides supportive evidence that the MMPs system is dominant.Librach and co-workers demonstrated that inhibitors of, and function-perturbing antibody against, metalloproteinase could completely inhibited cytotrophoblast invasion, whereas inhibitors of uPA system could only produced a partial inhibitory effect (Librach et al., 1991).
Glycodelin-A had no effect on mRNA expression and biological activity of TIMP-1 and TIMP-2.The mRNA expression of PAI-1 was not affected by glycodelin-A treatment either.Although modulation of the enzyme inhibitors is a possible strategy of regulating the activities of matrix degrading enzymes, this and other studies (Karmakar et al., 2004;Lash et al., 2005) suggest that a direct action on the enzymes is favored over influencing the inhibitors in the trophoblast system.
Trophoblast invasion is governed not only by the intrinsic properties of the EVCTs but also by interaction with the decidual cells (Kearns and Lala, 1983;Lockwood et al., 1999).Conditioned medium from human decidual cells suppresses invasion and MMP expression of trophoblasts (Graham and Lala, 1991;Zhang et al., 2002).Decidua  secretes glycodelin-A which stimulates the secretion of progesterone from primary cytotrophoblasts (Jeschke et al., 2004(Jeschke et al., , 2005)).Progesterone receptors are present in EVCTs (Shi et al., 1993;Wang et al., 1996).Progesterone has been suggested to affect trophoblast invasion through down-regulation of MMP-2 and -9 activities and gene expression (Shimonovitz et al., 1998;Goldman and Shalev, 2006), and up-regulation of TIMP-1 activities (Goldman and Shalev, 2006).The lack of action on TIMP-1 expression by glycodelin-A shown in this report suggests that glycodelin-A and progesterone are independent modulators of trophoblast invasion.
The inhibitory activity of glycodelin on trophoblast invasion was found to be glycosylation-dependent.Glycosylation is important for other biological activities of glycodelin as well (Seppala et al., 2002(Seppala et al., , 2007;;Yeung et al., 2006).It is involved in the binding of glycodelin to T cell receptors (Rachmilewitz et al., 2003) and human spermatozoa (Yeung et al., 2006).The sialic acid of glycodelin-A contributes to the apoptotic activity of the molecule on T-cells (Mukhopadhyay et al., 2004;Jayachandran et al., 2006).Glycosylation affects the action of glycodelin isoforms on spermatozoa-zona pellucida binding; glycodelin-A and -F inhibit, although glycodelin-C stimulate the binding (Chiu et al., 2007a, b).Consistent with these findings, the present data shows that the biological activity of glycodelin on trophoblast is also glycosylationdependent.Glycodelin-A, but not the other glycodelin isoforms or deglycosylated glycodelin, suppressed the invasiveness of TEV-1.Although glycodelin-F bound to TEV-1 cells, unlike glycodelin-A it did not suppress trophoblast invasion.One possible explanation for this is the presence of different receptors for glycodelin-A and -F on the TEV-1 cells.In human spermatozoa, there are different receptors for these isoforms (Chiu et al., 2003b).This possibility is supported by the difference in intensity of the fluorescent signal in TEV-1 cells treated with glycodelin-A and -F at the same concentration.It is also possible that different glycoforms bind to the same receptor, but only certain glycoforms can activate the signaling pathway, as reported for the glycoforms of human chorionic gonadotrophin (Wheatley and Hawtin, 1999;Fares, 2006).The action(s) of glycodelin-F on trophoblast function remain to be elucidated.
The receptor of glycodelin-A on human trophoblast is not known.A surface tyrosine phosphatase, CD45, has been suggested to be the receptor of glycodelin-A in T-cells (Rachmilewitz et al., 2003).However, CD45 is absent in EVCTs (Tarrade et al., 2001).We have identified fucosyltransferase-5 as a glycodelin-A binding protein on human spermatozoa (Chiu et al., 2007b).Fucosyltransferase-5 is unlikely to be involved in the invasion suppressive action of glycodelin-A on TEV-1 as its acceptors do not reduce the suppressive activity of glycodelin-A (Chiu and Yeung, unpublished observation) though they can compete for the glycodelin-A binding sites on human spermatozoa.The responsible receptor of glycodelin-A on EVCTs has yet to be identified.
The concentration of glycodelin-A increases from 6.3 to 7.8 mg/g protein in the secretory endometrium (Julkunen et al., 1986) to 101-103 mg/g protein in the deciduas (Julkunen et al., 1985).This rapid rise of concentration supports a role of glycodelin-A during pregnancy and implantation.The inhibitory activity of glycodelin-A is a potential mechanism in limiting the extent of trophoblast invasion, and may have implications in pregnancy complications involving defective invasion such as pre-eclampsia, which is characterized by shallow implantation.Although the precise mechanism leading to the pathogenesis of pre-eclampsia is still unclear, Lindoff and Astedt (1994) have demonstrated that low serum uPA level is a marker of preeclasmpsia.Furthermore, preeclamptic trophoblast cells have reduced cell surface uPA activities and secrete less active MMP-9 (Graham and McCrae, 1996) similar to glycodelin-A-treated TEV-1 cells.Whether glycodelin-A is involved in these pregnancies complications is unknown and is worth investigation.
Trophoblast invasion is tightly regulated.Decidua produces molecules that limit the invasion process.This study provides evidence that glycodelin-A is one of the decidual factors that restrains invasion of human trophoblasts into the uterus.Importantly, like the other biological actions of glycodelin-A, inhibition of trophoblast invasion is strictly glycosylation-dependent because both declycosylation and even subtle changes in glycosylation as present in various glycoforms is mitigates this activity.In the future, the role of differential glycosylation may turn out more important for biological actions than is currently understood.

Figure 2
Figure 2 Effect of 1 mg/ml glycodelin-A, -F or deglycosylated glycodelin on the invasion of TEV-1 cells.Above: Glycodelin-A, but not glycodelin-F and deglycosylated glycodelin, suppresses TEV-1 invasion in cell invasion transwell assay.The area covered by the cells (stained purple) on the membrane is reduced after treatment with glycodelin for 24 h.The results shown are representative of three replicate experiments.Below: Quantitative determination of TEV-1 invasion by CytoSelect TM 96-well Cell Invasion Assay (N = 5).*P , 0.05 when compared with the control without glycodelin treatment.

Figure 3
Figure 3 Effect of 1 mg/ml glycodelin-A on the invasion of JAR cells.Above: Glycodelin-A suppressed JAR invasion in cell invasion transwell assay.The area covered by the cells (stained purple) on the membrane is reduced after treatment with glycodelin-A for 24 h.The results shown are representative of three replicate experiments.Below: Quantitative determination of JAR invasion by CytoSelect TM 96-well Cell Invasion Assay (N = 5).*P , 0.05 when compared with the control without glycodelin-A treatment.

Figure 4
Figure 4 Effect of glycodelin-A on MMPs production in TEV-1 cells.(A) Real-time PCR analysis of the mRNA expression of MMP2 and MMP9 under glycodelin-A treatment.The values represent mean + SEM from five independent experiments.(B) Right: Determination of MMP2 and MMP9 activity in conditioned culture media of TEV-1 cells with or without glycodelin-A treatment by substrate gel zymography.Left: Bar chart showing the relative change of MMP2 and MMP9 activity.The values represent mean + SEM from three independent experiments.(C) Quantitative analysis of MMP2 and MMP9 secretion under glycodelin-A treatment using sandwich ELISA.The values represent mean + SEM from five independent experiments.*P , 0.05 when compared with the control without glycodelin-A treatment.

Figure 5
Figure 5 Effects of glycodelin-A on production of TIMPs in TEV-1 cells.(A) Real-time PCR analysis of the mRNA expression of TIMP-1, TIMP-2 and PAI-1 under glycodelin-A treatment.The values represent mean + SEM from five independent experiments.(B) Top: Determination of TIMP-1 and TIMP-2 activity in conditioned culture media of TEV-1 cells with or without glycodelin-A treatment by reverse zymography.Bottom: Bar chart showing the relative change of the TIMPs activity.The values represent mean + SEM from three independent experiments.

Figure 6
Figure 6 Effects of glycodelin-A on the uPA system in TEV-1 cells.(A) Real-time PCR analysis of the mRNA expression of uPA and PAI-1 under glycodelin-A treatment.(B) Quantitative analysis of uPA activity and secretion under glycodelin-A treatment using urokinase activity assay kit and sandwich ELISA, respectively.*P , 0.05 when compared with the control without glycodelin-A treatment.All values represent mean + SEM from five independent experiments.