Activin A and Inhibin A Differentially Regulate Human Uterine Matrix Metalloproteinases: Potential Interactions during Decidualiza
http://www.100md.com
《内分泌学杂志》
Prince Henry’s Institute of Medical Research (R.L.J., J.K.F., P.G.F., D.M.R., L.A.S.)
Monash University Department of Obstetrics & Gynaecology (E.W.), Clayton, Victoria 3168, Australia
Abstract
Embryo implantation and trophoblast invasion are tightly regulated processes, involving sophisticated communication between maternal decidual and fetal trophoblast cells. Decidualization is a prerequisite for successful implantation and is promoted by a number of paracrine agents, including activin A. To understand the downstream mechanisms of activin-promoted decidualization, the effects of activin on matrix metalloproteinases (MMPs) (important mediators of decidualization) were investigated. Activin A stimulated endometrial production of proMMPs-2, -3, -7, -9, and active MMP-2. In contrast, inhibin A was a potent inhibitor of proMMP-2, and antagonized the effect of activin on MMPs. Activin is up-regulated with decidualization, and MMPs-2, -3, and -9 increase in parallel. Furthermore, proMMP-2 production is stimulated when decidualization is accelerated with activin, and suppressed when activin is neutralized, attenuating decidualization. These data support that activin A promotes decidualization through up-regulating MMPs. Previous in vitro evidence proposes further roles for activin and MMPs in promoting trophoblast invasion; therefore, we examined their interrelationships in early human implantation sites. MMPs-7 and -9 were produced by static cytotrophoblast subpopulations, whereas MMP-2 was strikingly up-regulated in invasive extravillous cytotrophoblasts (EVT). Maternal decidua is the primary source of activin, where a role in stimulating MMP-2 in iEVTs can be envisaged. Inhibin was absent from cytotrophoblast populations, except for a dramatic up-regulation in endovascular EVT plugs, coinciding with a down-regulation of MMP-2. This suggests that inhibin may have a role in the cessation of vascular invasion. These data support that activin, via effects on MMPs, is an important factor in the maternal-fetal dialog regulating implantation.
Introduction
SUCCESSFUL IMPLANTATION is critical for the establishment of pregnancy. Failure of embryo implantation, a leading cause of in vitro fertilization failure, is commonly due to endometrial defects (1), whereas spontaneous abortion, intrauterine growth restriction, and preeclampsia are possible consequences of suboptimal trophoblast invasion. Understanding local regulation of endometrial preparation for implantation and trophoblast invasion is critical for the development of novel monitoring and therapeutic strategies for pathologies originating in early pregnancy.
Human trophoblast invasion is highly aggressive, and the endometrium undergoes extensive preparation in anticipation of implantation. In particular, endometrial stroma differentiates spontaneously every cycle to produce the decidua. A number of paracrine factors promote decidualization, including IL-11 and corticosteroid-releasing hormone (2, 3). Earlier studies in our laboratory demonstrated that activin A (a member of the TGF- superfamily) also accelerates decidualization, and that decidual cells secrete high concentrations of activin A (4).
A critical element of decidualization is remodeling of the extracellular environment. Matrix metalloproteinase (MMP) activity is essential during decidualization: MMP inhibition during early pregnancy in rat results in a severely diminished decidua (5, 6) and limits decidualization in primate stromal cells in culture (7). In addition to cleavage of extracellular matrix (ECM) components, MMPs play a broader role in the processing and liberation of ECM-tethered growth factors required for tissue remodeling (8). TGFs are important inhibitors of endometrial MMPs, transducing the suppressive effects of progesterone (9). Furthermore, activin A promotes MMP-2 production in monocytes and placenta (10, 11). We therefore hypothesized that activin stimulates MMP production in human endometrial cells, providing a potential downstream mechanism through which activin promotes decidualization.
Decidual activin A may also regulate trophoblast invasion via effects on cytotrophoblast MMPs. Invasive extravillous cytotrophoblasts (EVT) differentiate from progenitor cytotrophoblasts in anchoring cell columns, upon contact with maternal decidua. These either invade the decidua as interstitial (i) EVTs, or specifically target and enter maternal arteries as endovascular (v) EVTs. These form aggregates plug, and subsequently remodel, the maternal vessels (12). Focal digestion of decidual ECM by iEVTs is critical for their invasion, and cytotrophoblast cells secrete matrix degrading proteases (13). Factors secreted by decidual cells potently stimulate MMP production by cytotrophoblasts, up-regulating MMP-9 and -2 by 200 and 600%, respectively (14). Whereas the active constituents have not been identified, a separate study demonstrated that activin A stimulates MMP-2 secretion by cytotrophoblasts, and promotes cytotrophoblast outgrowth from placental villous tips in a model of trophoblast invasion (10). Because decidual cells secrete high levels of activin A (4), we propose that maternally derived activin significantly contributes to the promotion of trophoblast invasion.
The current study examined the interactions between activins and MMPs during the establishment of pregnancy in the human—both in the maternal decidua during preparation for implantation and in subsequent decidual invasion by fetal trophoblast cells. The effects of activin A on endometrial MMP production, in the presence and absence of antagonists inhibin and follistatin, are demonstrated, and supported by coexpression of activin and MMPs during decidualization. Furthermore, activin treatment and blockade during decidualization in vitro has downstream effects on MMP production. Finally, the interrelationships among activin, inhibin, and MMPs in decidua and distinct trophoblast subpopulations in early human implantation sites were investigated.
Materials and Methods
Tissues
Endometrial biopsies from the proliferative (d5–13; n = 5) and secretory (d15–14; n = 5) phases were collected by curettage from women (ages 25–40) undergoing minor gynecological investigations unrelated to endometrial pathology. Samples were collected in a 1:1 mixture of DMEM and Ham’s F12 medium (DMEM/F12; Trace Biosciences, Sydney, Australia) supplemented with 1% penicillin, streptomycin, and fungizone (Commonwealth Serum Laboratories, Melbourne, Australia) and L-glutamine (Sigma Diagnostics, St. Louis, MO). First trimester implantation sites (n = 6) were collected after surgical termination of pregnancy (6- to 8-wk amenorrhea), fixed in 10% buffered formalin, and paraffin embedded. Written informed consent was obtained from patients, and ethics approval was obtained from institutional ethic committees.
Cell isolation
Endometrial cells were isolated from individual biopsies, yielding stromal and epithelial cell preparations (n = 10) for subsequent hormone treatments. Endometrial tissue was enzymatically digested, and after separation by filtration (4, 15), stromal and epithelial cell-enriched fractions were plated overnight in DMEM/F12 with 10% charcoal-stripped fetal calf serum (Trace Biosciences), 1% penicillin, streptomycin, and fungizone, and L-glutamine.
Cells were further purified by selective adherence: stromal cells were detached with trypsin, plated in 24-well plates, and washed after 45 min to remove nonadherent cells. Epithelial cells were allowed to grow out from glandular structures for 48 h, then detached with trypsin, and serially replated (three times) in plastic culture dishes for 30 min, to allow adherence of contaminating stromal cells. Nonadherent cells were transferred to 24-well plates. This method achieves cultures of 97–99% (stromal) and 85–90% (epithelial) purity, as judged by morphological and immunohistochemical criteria (15). Cells were grown in DMEM/F12/ charcoal-stripped fetal calf serum for 48 h until 80% confluent, then placed in serum-free DMEM/F12 for 24–48 h before treatment. After 24 h of treatment, medium was collected, centrifuged, and snap-frozen in multiple aliquots. Cell number and viability were assessed by trypan blue exclusion.
Treatments
Epithelial and stromal cells were treated in DMEM/F12 containing: recombinant human (rh) activin A (R&D Systems Inc., Minneapolis, MN), rh-inhibin A (Diagnostic Systems Laboratories, Webster, TX), rh-follistatin-288 (Biotech Australia, Sydney, Australia) all between 1–4 nM. TGF-1 (R&D Systems) was used at 0.2–2 nM.
Zymography
Gelatin and casein zymography were performed to detect and quantitate secretion of latent and active forms of MMPs by endometrial cells (16). Conditioned media from baby hamster kidney cells stably transfected with human MMP-9 (provided by Dr. D. Edwards, University of East Anglia, Norwich, UK) was used as standards for gelatin zymography, whereas purified recombinant proMMP-1, -3, and -7 were used for casein zymography (Chemicon Australia Pty. Ltd., Boronia, Victoria, Australia). MMP levels were semiquantitated by densitometric analysis (size and intensity of band) using Gel Doc (Bio-Rad Laboratories, Regent’s Park, Australia). Analyses were performed before saturation of gel digestion (hence samples were assayed at a range of dilutions or concentrations), and comparisons were made between bands on the same gel.
ELISA
Prolactin production by endometrial stromal cells was assayed by ELISA (Bioclone Australia Pty. Ltd., Sydney, Australia) as a measure of extent of decidualization (4). The lower detection limit was 10 mIU/liter; inter- and intraassay variabilities were 6.2% and 3.6%. Activin A and inhibin A were measured by ELISA (Oxford-Bioinnovation Ltd., Upper Heyford, UK) (4, 17), using as standards: rhActivin A (National Institute of Biological Standards and Control, Potters Bar, UK) and rhInhibin A (Diagnostic Systems Laboratories). The ELISA sensitivities were: activin 35 pg/ml, with inter- and intraassay variabilities of 6.8% and 3.1%, and inhibin 6 pg/ml, with inter- and intraassay variabilities below 10%.
In vitro decidualization
Endometrial stromal cells were decidualized in vitro (4), with either cAMP (0.5 mM, Sigma) for 8 d, or 17-estradiol (E; 10–8 M, Sigma) and medroxyprogesterone acetate (MPA, 10–7 M, Sigma) for 12 d. During steroid-induced decidualization, cells were treated with rh-activin A (0–4 nM) or follistatin-288 (0–3 nM), the endogenous antagonist of activin bioactivity (18), throughout the experiment. Medium was collected every 48 h. cAMP-treated cells were snap-frozen for RNA extraction.
Real-time RT-PCR
Total RNA was extracted from cell pellets using RNeasy spin columns (Qiagen Pty. Ltd., Doncaster, Victoria, Australia), according to manufacturer’s instructions, and treated with ribonuclease-free deoxyribonuclease I (Ambion, Austin, TX). RNA samples were assessed for purity (A260:A280 1.8–2.0; A230:A280 >1) by spectrophotometry, and quantified using Ribogreen fluorescence RNA assay (Molecular Probes, Eugene, OR) (19).
One microgram of RNA was reverse transcribed using avian myeloblastosis virus reverse transcriptase (Promega, Annandale, Australia) and 100 ng random hexanucleotide primers (Amersham Biosciences, Piscataway, NJ) at 46 C for 90 min. All reactions were performed in triplicate and variability between triplicates assessed by quantitation of 18S by real-time PCR using a Roche Lightcycler (Roche, Castle Hill, New South Wales, Australia) (19). Triplicate samples within 15% coefficient of variation were pooled, to overcome the inherent variability of RT. Pooled cDNA samples, diluted 1:10, were analyzed for inhibin/activin subunits and MMP mRNA expression, using a protocol described previously (19), with annealing temperatures between 60 and 68 C, extension times between 10 and 20 sec, and MgCl2 between 3 and 5 mM. Primers used were MMP-2 (5'-TGGGAGCATGGCGATGGATA-3', 5'-ACAGTGGACATGGCGGTCTCA-3'), MMP-3 (5'-AGTCTTCCAATCCTACTGTTGCT-3', 5'-TCCCCGTCACCTCCAATCC-3'), MMP-7 (5'-GTTTAGAAGCCAAACTCAAGG-3', 5'-CTTTGACACTAATCGATCCAC-3'), MMP-9 (5'-GTATTTGTTCAAGGATGGGAAGTAC-3', 5'-GCAGGATGTCATAGGTCACGTAG-3'), inhibin (5'-ACGCTCAACTCCCCTGATG-3', 5'-ACCACCATGACAGTAGTGGAA-3') and activin A (5'-GGCTTGGAGTGCGACGGC-3', 5'-GCAGCCACACTCCTCCACAAT-3'). Standards were generated by conventional PCR (Hybaid PCR Express; Hybaid Instruments, Waltham, MA) and identity confirmed by DNA sequencing. PCR runs were repeated with inclusion of a quality control.
Immunohistochemistry
Inhibin/activin.
And A subunits were immunolocalized using affinity-purified rabbit polyclonal antibodies (kind gift from Prof. W. Vale, The Salk Institute for Biological Sciences, La Jolla, CA), as previously described (20) with modifications. In brief, microwave antigen-retrieval was performed, and antibodies applied at 2 μg/ml in nonimmune block containing 10% goat serum (Sigma), 2% human serum, and 0.1% Tween 20 (Bio-Rad) in Tris-buffered saline. Rabbit IgG (Dako, Glostrup, Denmark) at 2 μg/ml served as a negative control. Antibody binding was detected by sequential application of biotinylated goat antirabbit IgG (Dako) and avidin-biotin-peroxidase conjugate (Dako), with chromogen diaminobenzidine (Dako). Tissue sections were counterstained with Harris’s hematoxylin.
MMPs.
A similar protocol was used, with the following adaptations. Rabbit polyclonal antihuman MMP-7 antibody (kind gift from Prof. F. Woessner, Department of Biochemistry and Molecular Biology, University of Miami, Miami, FL) was applied at 0.5 μg/ml. MMP-2, -3, and -9 were detected using monoclonal antibodies (Merck Biosciences, Kilsyth, Victoria, Australia), diluted between 2 and 4 μg/ml (21, 22, 23). No microwaving was performed for MMP-9. Horse serum (Sigma) was used in the nonimmune block, and for MMP-2, no Tween 20 was included. The secondary antibody was biotinylated horse antimouse IgG (Vector Laboratories, Burlingame, CA). Rabbit/mouse IgG at matching concentrations to primary antibodies were used as negative controls
Cytokeratin.
An identical protocol was used as for MMP-9, with a pan-cytokeratin antibody (Dako) diluted 1:4 in nonimmune block after enzymatic antigen retrieval (0.1% trypsin in 0.1% CaCl2 for 15 min at 37 C).
Statistics
Data were analyzed by ANOVA with Tukey’s post hoc test (activin, inhibin, and follistatin treatments) or by paired Student’s t test (decidualization experiments).
Results
Initial experiments were performed to verify that expression of MMPs and inhibin/activin subunits was not disrupted by cell isolation and culture. Culture conditions were optimized to exclude all agents (e.g. fetal calf serum, BSA) known to artificially stimulate activins/MMPs. This resulted in low basal expression of inhibins, activins, and MMPs, resembling the in vivo expression of immunoreactive protein (20, 24).
Effect of activin on MMP production
Stromal cells expressed MMP-2 and -3 mRNA (data not shown), and protein production of proMMP-2 and -3 only was confirmed by zymography (Fig. 1A). Treatment of stromal cells with activin A stimulated proMMP-2 secretion (Fig. 1A). At 2 nM, activin A induced a 2.2-fold increase in proMMP-2 (Fig. 1B, n = 4 cell preparations, P < 0.05 vs. control). This dose was thereafter used for all analyses. Inhibin A (2 nM) significantly reduced proMMP-2 (P < 0.05 vs. control), and cotreatment of inhibin A with activin A at equimolar concentrations similarly reduced proMMP-2 (P < 0.05 vs. activin alone; Fig. 1B). Follistatin alone (2 nM) reduced proMMP-2 levels (not significant). Active MMP-2 was detected in only one of four stromal cell preparations, where it followed the same pattern as proMMP-2. No significant differences in proMMP-3 levels were detected after treatments (Fig. 1A).
The effect of activin A was compared with that of TGF-1: proMMP-2 and proMMP-3 were stimulated by 1 and 2 nM of activin A, but reduced by similar concentrations of TGF-1. A representative gelatin zymogram is shown in Fig. 1C. Due to considerable variation in the degree of response between cell preparations, representative data of two experiments for both proMMP-2 and proMMP-3 are shown in Fig. 1D. As before, active MMP-2 was detected in one of two cell preparations, and followed the same pattern as proMMP-2 (Fig. 1C).
Epithelial cells expressed mRNA for MMPs-2, -3, -7, and -9 (not shown), and these were all detectable at the protein level by zymography (Fig. 2, A and C). Epithelial cells (n = 4–6 cell preparations) were treated with activin A, inhibin A, both together and follistatin (2 nM). Activin A significantly stimulated production of proMMP-2, -3, and -7 (P < 0.05 vs. control), and nonsignificantly stimulated proMMP-9 and active MMP-2 (Fig. 2, B and D). Inhibin A alone potently inhibited production and activation of proMMP-2 only, but antagonized exogenous activin-stimulation of all MMPs (significant for MMPs-2, -3, -7; P < 0.05 vs. activin alone). Follistatin did not significantly affect MMP levels in epithelial cells, in the absence of exogenous activin.
Endogenous expression of inhibin and activin mRNAs were assessed using real-time PCR: inhibin mRNA was virtually absent in epithelial and stromal cells (<0.05 fg/pg 18S), whereas activin A mRNA was abundant in epithelial cells (150 fg/pg 18S vs. 3 fg/pg 18S in stromal cells) (not shown). This was verified by measurement of dimeric inhibin A and activin A secretion by ELISA: inhibin A levels were 5–32 pg/105 cells from endometrial cells, whereas activin A was secreted at levels of 5000 pg/105 cells from epithelial and 1000 pg/105 cells from stromal cells (not shown).
Activin and MMPs in decidualized stromal cells
To investigate the relationship among inhibin, activin, and MMPs after decidualization, highly decidualized cells, as determined by prolactin secretion, were generated by treatment with cAMP for 8d (n = 3; Fig. 3A). Expression levels of inhibin /activin A subunits and MMP-2, -3, -9 were quantitated by real-time PCR on d 8. Inhibin subunit mRNA expression was very low in decidual cells (0.2 fg/pg 18S, not shown), whereas activin A subunit mRNA was up-regulated more than 7-fold after cAMP-induced decidualization (P < 0.05; Fig. 3B). MMP-2, -3, -9 mRNAs were similarly up-regulated in decidualized cells, but this failed to reach significance due to variability between individual cell preparations (Fig. 3C). This pattern was confirmed at the protein level: inhibin A secretion was minimal from decidual cells (32 pg/105 cells, not shown), activin A has previously been shown to be up-regulated approximately 5-fold to 5000 pg/105 cells (4). ProMMP-2, -3, and -9, and active MMP-2 were up-regulated with decidualization (representative zymograms, Fig. 3D). ProMMP-1 was detected at low levels and was unchanged with decidualization. No MMP-7 mRNA or protein was detected, consistent with it being epithelial specific. Moreover, activin A subunit and MMP-2, -3, -9 immunolocalized to decidual cells of early pregnancy (Fig. 4, A–F). No immunoreactive inhibin -subunit or MMP-7 were detected.
To investigate whether MMPs were potential downstream mediators of activin-promoted decidualization, MMP levels were assessed where progesterone-induced decidualization was accelerated (2- to 3-fold) by addition of activin A (n = 2), or attenuated (25%) by neutralization of activin bioactivity with follistatin (n = 2) (4). Activin A (0.4–4 nM) stimulated proMMP-2 secretion and activation on d 8 of decidualization (Fig. 3E). Furthermore, proMMP-2 secretion was reduced on d 8 of decidualization in the presence of follistatin (Fig. 3F).
Activin, inhibin, and MMPs during trophoblast invasion
Spatial expression of MMP-2, -7, -9, and inhibin /activin A subunits were assessed in trophoblast subtypes in serial sections of implantation sites from normal human pregnancies. Trophoblast cells were identified by immunostaining for cytokeratin.
Immunostaining for activin A subunit was absent in anchoring cell columns (Fig. 4H) but was strongly expressed, together with -subunit, in trophoblast cells undergoing fusion at the tips and edges of the columns (Fig. 4, H and I). Cytotrophoblast cells in the upper cell column were negative for MMP-2, but it was strikingly up-regulated in distal cells and EVTs breaking-off and invading decidua (Fig. 4, J and K). MMP-7 (not shown) and -9 (Fig. 4L) were strongly expressed by all column cytotrophoblast cells.
iEVTs invading decidua did not express inhibin -subunit, and only weakly expressed activin A (Fig. 4, N and P). However, iEVTs were intensely stained for MMP-2 (Fig. 4, Q and S), whereas staining for MMP-7 (not shown) and MMP-9 was faint or absent (Fig. 4U). Conversely, strong immunostaining for activin A and inhibin was detected in vEVT plugs (Fig. 5, B and C). MMP-2 expression was strikingly diminished in vEVTs (Fig. 5D), whereas MMPs-7 and 9 were maintained (Fig. 5, E and F).
In addition, inhibin and activin A subunits immunostained enlarged mono- and multinucleated trophoblast cells, identified as cytotrophoblast giant cells (Fig. 5, G and H) and syncytial fragments (Fig. 5I). No staining was observed for MMPs (not shown). Leukocytes within villi and decidua intensely stained for MMPs and activin A-subunit. Specificity of immunostaining for inhibin subunits was confirmed by their selective staining in syncytio- but not villous cytotrophoblast (25) (Fig. 5, K and L). MMPs were not produced by syncytiotrophoblast; however, villous cytotrophoblasts were strongly positive for MMP-7 (not shown) and -9 (Fig. 5N). These data are summarized in Fig. 6A.
Discussion
This study demonstrates that activin A stimulates MMP production (proMMP-2, -3, -7, -9) by endometrial cells. MMP-2 activation was also enhanced by activin, in parallel to increased availability of latent form. Furthermore, inhibin A antagonizes activin-induction of MMPs and is a potent inhibitor of proMMP-2. MMP up-regulation by activin A is consistent with actions in other cell types (10, 11), suggesting modulation of MMPs is a prime downstream action of activin during tissue remodeling, e.g. wound healing, embryogenesis, and inflammation. Activin and TGFs (9) therefore appear to be opposing regulators of MMPs in endometrial cells, consistent with their contrasting effects on cytotrophoblast invasion (10, 26) and placental steroidogenesis (27). This indicates divergent signaling pathways for activins and TGFs in endometrium and placenta, either through differential phosphorylation of Smads, as in pituitary cells (28), or through use of specific downstream cofactors.
Inhibin was a potent antagonist of activin in endometrial cells, overriding the stimulatory effect of activin on MMPs at equimolar concentrations. Both endometrium and placenta express activin type II receptors and betaglycan (29, 30), conferring sensitivity to inhibin-mediated antagonism of activin action (31). However, this is the first demonstration of a function for inhibin within the uterus. A role for auto/paracrine inhibin is doubtful because endometrial inhibin levels are limited. However, inhibin is a potent endocrine hormone, and circulates at high levels throughout the menstrual cycle when MMP production is suppressed (32). Progesterone withdrawal premenstrually releases MMP suppression in the endometrium, and triggers MMP-mediated endometrial breakdown (16). Because inhibin A levels mirror those of progesterone in the luteal phase, a parallel role may be anticipated. Moreover, the sharp increase in inhibin B (believed to have similar actions to inhibin A) in the late menstrual phase (33) potentially represents a previously unrecognized endocrine mechanism for the cessation of menstruation-associated MMP production.
Although all MMPs were stimulated by activin, MMP-2 was more sensitive to inhibition by inhibin than other MMPs. That MMPs are differentially regulated is well established (reviewed in Ref.34). However, MMP-2 is unusual in that its promoter does not contain the activator protein-1 or polyoma-enhancing activity-3/E26 virus elements found in most other MMPs expressed in the endometrium, although it does have two SP1-binding sites (35). In some cell types, MMP-2 expression can be induced by TGF (36), but in general expression levels are little affected by the action of growth factors and cytokines. Clearly, the balance of stimulatory and inhibitory factors in a particular microenvironment will determine MMP output. This is undoubtedly related to distinct functions; for example in the placenta it appears that MMP-2 expression corresponds to cytotrophoblast invasive potential, whereas MMP-7 and -9 are maximal in static but metabolically active cells (e.g. villous cytotrophoblasts, vEVTs).
Inhibin A was a more potent antagonist in stromal than epithelial cells; no inhibitory effect was observed on epithelial MMP-3, -7, -9 in the absence of exogenous activin. This probably relates to differences in endogenous activin levels. Although activin production by epithelial cells is approximately 5 times higher than by stromal cells, follistatin is coexpressed in endometrial epithelial, but not stromal, cells (29). Thus, bioactive activin may be negligible in epithelial-conditioned media. This concept is also supported by the failure of follistatin treatment to suppress epithelial MMPs.
Activin A is dramatically up-regulated in decidualizing cells and promotes decidual progression (4). Inhibin expression is limited in the decidua, suggesting an activin-dominated environment. Coincidental to activin up-regulation, we observe an elevation of decidual MMP-2, -3, -9, and indeed MMP-2 production and activation is further increased by treatment of decidualizing cells with activin. Inhibition of MMPs impedes the decidualization process (5, 6, 7). A similar attenuation of decidualization occurs with activin neutralization (4, 37), which we demonstrate here results in diminished proMMP-2. These data reinforce a physiological interaction between activins and MMPs during decidualization and support the hypothesis that activin promotes decidual progression, at least in part, through elevating MMP production.
The placenta is a significant source of inhibins/activins, primarily from syncytiotrophoblast (25), and roles in regulating placental steroidogenesis have been identified (38). Furthermore, exogenous activin promotes differentiation of EVTs and MMP-2 secretion, in an in vitro model of trophoblast invasion (10). Our in vivo data support this, demonstrating that maternal decidua is the prime source of activin, and that MMP-2 is strikingly up-regulated in cytotrophoblast cells as they contact and invade through the activin-rich decidua. We therefore propose that activin is an important constituent of decidual secretions, stimulating trophoblast MMPs (14).
Immunohistochemistry is invaluable for determining protein expression by individual cell populations in vivo. In early implantation sites, distinct trophoblast populations clearly expressed specific repertoires of inhibin, activin, and MMPs. MMP-2 appeared to be associated with cytotrophoblast migration, consistent with its abundance in iEVTs in ectopic pregnancies (39). Conversely, MMP-7 and -9 were expressed by nonmotile cytotrophoblasts, suggesting roles in processing, rather than invasion, and indicating differential regulation of MMPs by the microenvironmental milieu of cytokines/growth factors. These data contrast with studies using primary cytotrophoblast cells, which attribute MMP-9 with invasive potential (13, 40). However, isolated cytotrophoblast cells are likely to be a mixed population, and we demonstrate that MMP-9 is abundant in all cytotrophoblast subtypes, except for the minor iEVT population. MMP-9 levels are reduced in isolated cytotrophoblasts with advancing gestation: however, a variety of other MMPs (including MMP-2) are strictly down-regulated after the first trimester (41). Moreover trophoblast invasion is limited in vitro by neutralization of MMP-2 (42). This indicates species-specific differences in placentation because MMP-9 is the predominant MMP in murine invasive giant cells (6).
Absence of MMP-2, but maintenance of MMP-7 and -9, was observed in endovascular aggregates of vEVTs, consistent with their cessation of migration. Interestingly, strong immunoreactivity for inhibin/activin subunits was simultaneously observed in vEVTs. A similar expression profile is seen in human giant cells—the terminal differentiation state of iEVTs—again associated with loss of migratory potential (12). EVT invasion must be tightly limited, to prevent overinvasion into myometrium and vasculature. Given the increased production of inhibin subunits, with a paralleled decrease in MMP-2 production in immobilized EVT populations, it is tempting to hypothesize that inhibin contributes to diminished trophoblast invasiveness.
The trigger for inhibin subunit expression in cytotrophoblasts is unknown. However, the implantation site is maintained in a hypoxic state, and this is believed to be critical for regulating trophoblast differentiation (43). Inducing placental hypoxia in vitro significantly suppresses inhibin and activin production (44). This suggests that inhibin/activin may be up-regulated as vEVTs invade maternal vessels and encounter an oxygen-rich environment. Giant cells also produce inhibin/activin subunits. These enlarged cells are multinucleated, and a similar up-regulation of inhibin/activin subunits occurs with syncytialization in vivo and in vitro (45). Whether this is cause or effect is not known, as syncytial formation has not been clearly defined.
In summary, we show that activin A stimulates endometrial MMPs, and inhibin potently antagonizes activin action. The relationship between activin A and MMP-2 suggests MMP up-regulation is a downstream event during activin-stimulated decidualization. Spatial expression patterns of inhibin/activin and MMPs in intrauterine implantation sites demonstrate that cytotrophoblast MMP-2 is up-regulated by contact with decidua; thus, we propose that decidual-derived activin is important for promoting iEVT invasion (Fig. 6B). The physiological significance of the inhibition of MMP-2 by inhibin is reinforced by their lack of colocalization in cytotrophoblast populations, indicating that autocrine inhibin production may be a critical switch during trophoblast differentiation. Thus, these data highlight a number of mechanisms through which activins and inhibins may contribute to the bidirectional communication between maternal and fetal cells during the establishment of pregnancy.
Acknowledgments
We acknowledge the technical assistance of Tu’uhevaha Kaitu’u, Premila Paiva, Natalie Hannan, Ashwini Chand, and Sue Panckridge.
Footnotes
This work was supported by National Health & Medical Research Council (NH&MRC) of Australia, Royal Australian and New Zealand College of Obstetrics and Gynaecology). Studies were funded by NH&MRC Program Grant (No. 241000). R.L.J. was the recipient of the Ella McKnight Scholarship (Royal Australian and New Zealand College of Obstetrics & Gynaecology). L.A.S., D.M.R., and J.F.K. are supported by Fellowship Nos. 143798, 198707 and 198705 from NH&MRC.
Disclosure: R.L.J., J.K.F., P.G.F., D.M.R., E.W., and L.A.S. have nothing to declare.
First Published Online November 10, 2005
Abbreviations: ECM, Extracellular matrix; EVT, extravillous cytotrophoblast; iEVT, interstitial; MMP, matrix metalloproteinase; rh, recombinant human; vEVT, endovascular.
Accepted for publication October 31, 2005.
References
Sharkey AM, Smith SK 2003 The endometrium as a cause of implantation failure. Best Pract Res Clin Obstet Gynaecol 17:289–307
Dimitriadis E, Robb L, Salamonsen LA 2002 Interleukin 11 advances progesterone-induced decidualization of human endometrial stromal cells. Mol Hum Reprod 8:636–643
Zoumakis E, Margioris AN, Stournaras C, Dermitzaki E, Angelakis E, Makrigiannakis A, Koumantakis E, Gravanis A 2000 Corticotrophin-releasing hormone (CRH) interacts with inflammatory prostaglandins and interleukins and affects the decidualization of human endometrial stroma. Mol Hum Reprod 6:344–351
Jones RL, Salamonsen LA, Findlay JK 2002 Activin A promotes human endometrial stromal cell decidualization in vitro. J Clin Endocrinol Metab 87:4001–4004
Rechtman MP, Zhang J, Salamonsen LA 1999 Effect of inhibition of matrix metalloproteinases on endometrial decidualization and implantation in mated rats. J Reprod Fertil 117:169–177
Alexander CM, Hansell EJ, Behrendtsen O, Flannery ML, Kishnani NS, Hawkes SP, Werb Z 1996 Expression and function of matrix metalloproteinases and their inhibitors at the maternal-embryonic boundary during mouse embryo implantation. Development 122:1723–1736
Strakova Z, Szmidt M, Srisuparp S, Fazleabas AT 2003 Inhibition of matrix metalloproteinases prevents the synthesis of insulin-like growth factor binding protein-1 during decidualization in the baboon. Endocrinology 144:5339–5346
McCawley LJ, Matrisian LM 2001 Matrix metalloproteinases: they’re not just for matrix anymore! Curr Opin Cell Biol 13:534–540
Osteen KG, Igarashi TM, Bruner-Tran KL 2003 Progesterone action in the human endometrium: induction of a unique tissue environment which limits matrix metalloproteinase (MMP) expression. Front Biosci 8:d78–d86
Caniggia I, Lye SJ, Cross JC 1997 Activin is a local regulator of human cytotrophoblast cell differentiation. Endocrinology 138:3976–3986
Ogawa K, Funaba M, Mathews LS, Mizutani T 2000 Activin A stimulates type IV collagenase (matrix metalloproteinase-2) production in mouse peritoneal macrophages. J Immunol 165:2997–3003
Loke YW, King A 1995 Human implantation. Cell biology and immunology. Cambridge University Press, Cambridge
Fisher SJ, Leitch MS, Kantor MS, Basbaum CB, Kramer RH 1985 Degradation of extracellular matrix by the trophoblastic cells of first-trimester human placentas. J Cell Biochem 27:31–41
Bischof P, Meisser A, Campana A, Tseng L 1998 Effects of decidua-conditioned medium and insulin-like growth factor binding protein-1 on trophoblastic matrix metalloproteinases and their inhibitors. Placenta 19:457–464
Marsh MM, Hampton AL, Riley SC, Findlay JK, Salamonsen LA 1994 Production and characterization of endothelin released by human endometrial epithelial cells in culture. J Clin Endocrinol Metab 79:1625–1631
Salamonsen LA, Butt AR, Hammond FR, Garcia S, Zhang J 1997 Production of endometrial matrix metalloproteinases, but not their tissue inhibitors, is modulated by progesterone withdrawal in an in vitro model for menstruation. J Clin Endocrinol Metab 82:1409–1415
Landgren BM, Collins A, Csemiczky G, Burger HG, Baksheev L, Robertson DM 2004 Menopause transition: annual changes in serum hormonal patterns over the menstrual cycle in women during a nine-year period prior to menopause. J Clin Endocrinol Metab 89:2763–2769
de Winter JP, ten Dijke P, de Vries CJ, van Achterberg TA, Sugino H, de Waele P, Huylebroeck D, Verschueren K, van den Eijnden-van Raaij AJ 1996 Follistatins neutralize activin bioactivity by inhibition of activin binding to its type II receptors. Mol Cell Endocrinol 116:105–114
Jones RL, Hannan NJ, Kaitu’u TJ, Zhang J, Salamonsen LA 2004 Identification of chemokines important for leukocyte recruitment to the human endometrium at the times of embryo implantation and menstruation. J Clin Endocrinol Metab 89:6155–6167
Jones RL, Salamonsen LA, Critchley HO, Rogers PA, Affandi B, Findlay JK 2000 Inhibin and activin subunits are differentially expressed in endometrial cells and leukocytes during the menstrual cycle, in early pregnancy and in women using progestin-only contraception. Mol Hum Reprod 6:1107–1117
Vincent AJ, Malakooti N, Zhang J, Rogers PA, Affandi B, Salamonsen LA 1999 Endometrial breakdown in women using Norplant is associated with migratory cells expressing matrix metalloproteinase-9 (gelatinase B). Hum Reprod 14:807–815
Vincent AJ, Zhang J, Ostor A, Rogers PA, Affandi B, Kovacs G, Salamonsen LA 2000 Matrix metalloproteinase-1 and -3 and mast cells are present in the endometrium of women using progestin-only contraceptives. Hum Reprod 15:123–130
Zhang J, Hampton AL, Nie G, Salamonsen LA 2000 Progesterone inhibits activation of latent matrix metalloproteinase (MMP)-2 by membrane-type 1 MMP: enzymes coordinately expressed in human endometrium. Biol Reprod 62:85–94
Salamonsen LA, Woolley DE 1996 Matrix metalloproteinases in normal menstruation. Hum Reprod 11(Suppl 2):124–133
Minami S, Yamoto M, Nakano R 1992 Immunohistochemical localization of inhibin/activin subunits in human placenta. Obstet Gynecol 80:410–414
Chakraborty C, Gleeson LM, McKinnon T, Lala PK 2002 Regulation of human trophoblast migration and invasiveness. Can J Physiol Pharmacol 80:116–124
Song Y, Keelan J, France JT 1996 Activin-A stimulates, while transforming growth factor 1 inhibits, chorionic gonadotrophin production and aromatase activity in cultured human placental trophoblasts. Placenta 17:603–610
Suszko MI, Balkin DM, Chen Y, Woodruff TK 2005 Smad3 mediates activin-induced transcription of follicle-stimulating hormone -subunit gene. Mol Endocrinol 19:1849–1858
Jones RL, Salamonsen LA, Zhao YC, Ethier JF, Drummond AE, Findlay JK 2002 Expression of activin receptors, follistatin and betaglycan by human endometrial stromal cells; consistent with a role for activins during decidualization. Mol Hum Reprod 8:363–374
Ciarmela P, Florio P, Toti P, Grasso D, Santopietro R, Tosi P, Petraglia F 2003 Expression of betaglycan in pregnant tissues throughout gestation. Eur J Endocrinol 149:433–437
Lewis KA, Gray PC, Blount AL, MacConell LA, Wiater E, Bilezikjian LM, Vale W 2000 Betaglycan binds inhibin and can mediate functional antagonism of activin signalling. Nature 404:411–414
Muttukrishna S, Fowler PA, Groome NP, Mitchell GG, Robertson WR, Knight PG 1994 Serum concentrations of dimeric inhibin during the spontaneous human menstrual cycle and after treatment with exogenous gonadotrophin. Hum Reprod 9:1634–1642
Groome NP, Illingworth PJ, O’Brien M, Pai R, Rodger FE, Mather JP, McNeilly AS 1996 Measurement of dimeric inhibin B throughout the human menstrual cycle. J Clin Endocrinol Metab 81:1401–1405
Curry Jr TE, Osteen KG 2003 The matrix metalloproteinase system: changes, regulation, and impact throughout the ovarian and uterine reproductive cycle. Endocr Rev 24:428–465
Murphy G 2004 Gelatinase A. London: Elsevier Academic Press
Overall CM, Wrana JL, Sodek J 1991 Transcriptional and post-transcriptional regulation of 72-kDa gelatinase/type IV collagenase by transforming growth factor-1 in human fibroblasts. Comparisons with collagenase and tissue inhibitor of matrix metalloproteinase gene expression. J Biol Chem 266:14064–14071
Tierney EP, Giudice LC 2004 Role of activin A as a mediator of in vitro endometrial stromal cell decidualization via the cyclic adenosine monophosphate pathway. Fertil Steril 81(Suppl 1):899–903
Ni X, Luo S, Minegishi T, Peng C 2000 Activin A in JEG-3 cells: potential role as an autocrine regulator of steroidogenesis in humans. Biol Reprod 62:1224–1230
Bai SX, Wang YL, Qin L, Xiao ZJ, Herva R, Piao YS 2005 Dynamic expression of matrix metalloproteinases (MMP-2, -9 and -14) and the tissue inhibitors of MMPs (TIMP-1, -2 and -3) at the implantation site during tubal pregnancy. Reproduction 129:103–113
Librach CL, Werb Z, Fitzgerald ML, Chiu K, Corwin NM, Esteves RA, Grobelny D, Galardy R, Damsky CH, Fisher SJ 1991 92-kD type IV collagenase mediates invasion of human cytotrophoblasts. J Cell Biol 113:437–449
Fisher SJ, Cui TY, Zhang L, Hartman L, Grahl K, Zhang GY, Tarpey J, Damsky CH 1989 Adhesive and degradative properties of human placental cytotrophoblast cells in vitro. J Cell Biol 109:891–902
Isaka K, Usuda S, Ito H, Sagawa Y, Nakamura H, Nishi H, Suzuki Y, Li YF, Takayama M 2003 Expression and activity of matrix metalloproteinase 2 and 9 in human trophoblasts. Placenta 24:53–64
Adelman DM, Gertsenstein M, Nagy A, Simon MC, Maltepe E 2000 Placental cell fates are regulated in vivo by HIF-mediated hypoxia responses. Genes Dev 14:3191–3203
Manuelpillai U, Schneider-Kolsky M, Thirunavukarasu P, Dole A, Waldron K, Wallace EM 2003 Effect of hypoxia on placental activin A, inhibin A and follistatin synthesis. Placenta 24:77–83
Debieve F, Pampfer S, Thomas K 2000 Inhibin and activin production and subunit expression in human placental cells cultured in vitro. Mol Hum Reprod 6:743–749(Rebecca L. Jones, Jock K. Findlay, Paul )
Monash University Department of Obstetrics & Gynaecology (E.W.), Clayton, Victoria 3168, Australia
Abstract
Embryo implantation and trophoblast invasion are tightly regulated processes, involving sophisticated communication between maternal decidual and fetal trophoblast cells. Decidualization is a prerequisite for successful implantation and is promoted by a number of paracrine agents, including activin A. To understand the downstream mechanisms of activin-promoted decidualization, the effects of activin on matrix metalloproteinases (MMPs) (important mediators of decidualization) were investigated. Activin A stimulated endometrial production of proMMPs-2, -3, -7, -9, and active MMP-2. In contrast, inhibin A was a potent inhibitor of proMMP-2, and antagonized the effect of activin on MMPs. Activin is up-regulated with decidualization, and MMPs-2, -3, and -9 increase in parallel. Furthermore, proMMP-2 production is stimulated when decidualization is accelerated with activin, and suppressed when activin is neutralized, attenuating decidualization. These data support that activin A promotes decidualization through up-regulating MMPs. Previous in vitro evidence proposes further roles for activin and MMPs in promoting trophoblast invasion; therefore, we examined their interrelationships in early human implantation sites. MMPs-7 and -9 were produced by static cytotrophoblast subpopulations, whereas MMP-2 was strikingly up-regulated in invasive extravillous cytotrophoblasts (EVT). Maternal decidua is the primary source of activin, where a role in stimulating MMP-2 in iEVTs can be envisaged. Inhibin was absent from cytotrophoblast populations, except for a dramatic up-regulation in endovascular EVT plugs, coinciding with a down-regulation of MMP-2. This suggests that inhibin may have a role in the cessation of vascular invasion. These data support that activin, via effects on MMPs, is an important factor in the maternal-fetal dialog regulating implantation.
Introduction
SUCCESSFUL IMPLANTATION is critical for the establishment of pregnancy. Failure of embryo implantation, a leading cause of in vitro fertilization failure, is commonly due to endometrial defects (1), whereas spontaneous abortion, intrauterine growth restriction, and preeclampsia are possible consequences of suboptimal trophoblast invasion. Understanding local regulation of endometrial preparation for implantation and trophoblast invasion is critical for the development of novel monitoring and therapeutic strategies for pathologies originating in early pregnancy.
Human trophoblast invasion is highly aggressive, and the endometrium undergoes extensive preparation in anticipation of implantation. In particular, endometrial stroma differentiates spontaneously every cycle to produce the decidua. A number of paracrine factors promote decidualization, including IL-11 and corticosteroid-releasing hormone (2, 3). Earlier studies in our laboratory demonstrated that activin A (a member of the TGF- superfamily) also accelerates decidualization, and that decidual cells secrete high concentrations of activin A (4).
A critical element of decidualization is remodeling of the extracellular environment. Matrix metalloproteinase (MMP) activity is essential during decidualization: MMP inhibition during early pregnancy in rat results in a severely diminished decidua (5, 6) and limits decidualization in primate stromal cells in culture (7). In addition to cleavage of extracellular matrix (ECM) components, MMPs play a broader role in the processing and liberation of ECM-tethered growth factors required for tissue remodeling (8). TGFs are important inhibitors of endometrial MMPs, transducing the suppressive effects of progesterone (9). Furthermore, activin A promotes MMP-2 production in monocytes and placenta (10, 11). We therefore hypothesized that activin stimulates MMP production in human endometrial cells, providing a potential downstream mechanism through which activin promotes decidualization.
Decidual activin A may also regulate trophoblast invasion via effects on cytotrophoblast MMPs. Invasive extravillous cytotrophoblasts (EVT) differentiate from progenitor cytotrophoblasts in anchoring cell columns, upon contact with maternal decidua. These either invade the decidua as interstitial (i) EVTs, or specifically target and enter maternal arteries as endovascular (v) EVTs. These form aggregates plug, and subsequently remodel, the maternal vessels (12). Focal digestion of decidual ECM by iEVTs is critical for their invasion, and cytotrophoblast cells secrete matrix degrading proteases (13). Factors secreted by decidual cells potently stimulate MMP production by cytotrophoblasts, up-regulating MMP-9 and -2 by 200 and 600%, respectively (14). Whereas the active constituents have not been identified, a separate study demonstrated that activin A stimulates MMP-2 secretion by cytotrophoblasts, and promotes cytotrophoblast outgrowth from placental villous tips in a model of trophoblast invasion (10). Because decidual cells secrete high levels of activin A (4), we propose that maternally derived activin significantly contributes to the promotion of trophoblast invasion.
The current study examined the interactions between activins and MMPs during the establishment of pregnancy in the human—both in the maternal decidua during preparation for implantation and in subsequent decidual invasion by fetal trophoblast cells. The effects of activin A on endometrial MMP production, in the presence and absence of antagonists inhibin and follistatin, are demonstrated, and supported by coexpression of activin and MMPs during decidualization. Furthermore, activin treatment and blockade during decidualization in vitro has downstream effects on MMP production. Finally, the interrelationships among activin, inhibin, and MMPs in decidua and distinct trophoblast subpopulations in early human implantation sites were investigated.
Materials and Methods
Tissues
Endometrial biopsies from the proliferative (d5–13; n = 5) and secretory (d15–14; n = 5) phases were collected by curettage from women (ages 25–40) undergoing minor gynecological investigations unrelated to endometrial pathology. Samples were collected in a 1:1 mixture of DMEM and Ham’s F12 medium (DMEM/F12; Trace Biosciences, Sydney, Australia) supplemented with 1% penicillin, streptomycin, and fungizone (Commonwealth Serum Laboratories, Melbourne, Australia) and L-glutamine (Sigma Diagnostics, St. Louis, MO). First trimester implantation sites (n = 6) were collected after surgical termination of pregnancy (6- to 8-wk amenorrhea), fixed in 10% buffered formalin, and paraffin embedded. Written informed consent was obtained from patients, and ethics approval was obtained from institutional ethic committees.
Cell isolation
Endometrial cells were isolated from individual biopsies, yielding stromal and epithelial cell preparations (n = 10) for subsequent hormone treatments. Endometrial tissue was enzymatically digested, and after separation by filtration (4, 15), stromal and epithelial cell-enriched fractions were plated overnight in DMEM/F12 with 10% charcoal-stripped fetal calf serum (Trace Biosciences), 1% penicillin, streptomycin, and fungizone, and L-glutamine.
Cells were further purified by selective adherence: stromal cells were detached with trypsin, plated in 24-well plates, and washed after 45 min to remove nonadherent cells. Epithelial cells were allowed to grow out from glandular structures for 48 h, then detached with trypsin, and serially replated (three times) in plastic culture dishes for 30 min, to allow adherence of contaminating stromal cells. Nonadherent cells were transferred to 24-well plates. This method achieves cultures of 97–99% (stromal) and 85–90% (epithelial) purity, as judged by morphological and immunohistochemical criteria (15). Cells were grown in DMEM/F12/ charcoal-stripped fetal calf serum for 48 h until 80% confluent, then placed in serum-free DMEM/F12 for 24–48 h before treatment. After 24 h of treatment, medium was collected, centrifuged, and snap-frozen in multiple aliquots. Cell number and viability were assessed by trypan blue exclusion.
Treatments
Epithelial and stromal cells were treated in DMEM/F12 containing: recombinant human (rh) activin A (R&D Systems Inc., Minneapolis, MN), rh-inhibin A (Diagnostic Systems Laboratories, Webster, TX), rh-follistatin-288 (Biotech Australia, Sydney, Australia) all between 1–4 nM. TGF-1 (R&D Systems) was used at 0.2–2 nM.
Zymography
Gelatin and casein zymography were performed to detect and quantitate secretion of latent and active forms of MMPs by endometrial cells (16). Conditioned media from baby hamster kidney cells stably transfected with human MMP-9 (provided by Dr. D. Edwards, University of East Anglia, Norwich, UK) was used as standards for gelatin zymography, whereas purified recombinant proMMP-1, -3, and -7 were used for casein zymography (Chemicon Australia Pty. Ltd., Boronia, Victoria, Australia). MMP levels were semiquantitated by densitometric analysis (size and intensity of band) using Gel Doc (Bio-Rad Laboratories, Regent’s Park, Australia). Analyses were performed before saturation of gel digestion (hence samples were assayed at a range of dilutions or concentrations), and comparisons were made between bands on the same gel.
ELISA
Prolactin production by endometrial stromal cells was assayed by ELISA (Bioclone Australia Pty. Ltd., Sydney, Australia) as a measure of extent of decidualization (4). The lower detection limit was 10 mIU/liter; inter- and intraassay variabilities were 6.2% and 3.6%. Activin A and inhibin A were measured by ELISA (Oxford-Bioinnovation Ltd., Upper Heyford, UK) (4, 17), using as standards: rhActivin A (National Institute of Biological Standards and Control, Potters Bar, UK) and rhInhibin A (Diagnostic Systems Laboratories). The ELISA sensitivities were: activin 35 pg/ml, with inter- and intraassay variabilities of 6.8% and 3.1%, and inhibin 6 pg/ml, with inter- and intraassay variabilities below 10%.
In vitro decidualization
Endometrial stromal cells were decidualized in vitro (4), with either cAMP (0.5 mM, Sigma) for 8 d, or 17-estradiol (E; 10–8 M, Sigma) and medroxyprogesterone acetate (MPA, 10–7 M, Sigma) for 12 d. During steroid-induced decidualization, cells were treated with rh-activin A (0–4 nM) or follistatin-288 (0–3 nM), the endogenous antagonist of activin bioactivity (18), throughout the experiment. Medium was collected every 48 h. cAMP-treated cells were snap-frozen for RNA extraction.
Real-time RT-PCR
Total RNA was extracted from cell pellets using RNeasy spin columns (Qiagen Pty. Ltd., Doncaster, Victoria, Australia), according to manufacturer’s instructions, and treated with ribonuclease-free deoxyribonuclease I (Ambion, Austin, TX). RNA samples were assessed for purity (A260:A280 1.8–2.0; A230:A280 >1) by spectrophotometry, and quantified using Ribogreen fluorescence RNA assay (Molecular Probes, Eugene, OR) (19).
One microgram of RNA was reverse transcribed using avian myeloblastosis virus reverse transcriptase (Promega, Annandale, Australia) and 100 ng random hexanucleotide primers (Amersham Biosciences, Piscataway, NJ) at 46 C for 90 min. All reactions were performed in triplicate and variability between triplicates assessed by quantitation of 18S by real-time PCR using a Roche Lightcycler (Roche, Castle Hill, New South Wales, Australia) (19). Triplicate samples within 15% coefficient of variation were pooled, to overcome the inherent variability of RT. Pooled cDNA samples, diluted 1:10, were analyzed for inhibin/activin subunits and MMP mRNA expression, using a protocol described previously (19), with annealing temperatures between 60 and 68 C, extension times between 10 and 20 sec, and MgCl2 between 3 and 5 mM. Primers used were MMP-2 (5'-TGGGAGCATGGCGATGGATA-3', 5'-ACAGTGGACATGGCGGTCTCA-3'), MMP-3 (5'-AGTCTTCCAATCCTACTGTTGCT-3', 5'-TCCCCGTCACCTCCAATCC-3'), MMP-7 (5'-GTTTAGAAGCCAAACTCAAGG-3', 5'-CTTTGACACTAATCGATCCAC-3'), MMP-9 (5'-GTATTTGTTCAAGGATGGGAAGTAC-3', 5'-GCAGGATGTCATAGGTCACGTAG-3'), inhibin (5'-ACGCTCAACTCCCCTGATG-3', 5'-ACCACCATGACAGTAGTGGAA-3') and activin A (5'-GGCTTGGAGTGCGACGGC-3', 5'-GCAGCCACACTCCTCCACAAT-3'). Standards were generated by conventional PCR (Hybaid PCR Express; Hybaid Instruments, Waltham, MA) and identity confirmed by DNA sequencing. PCR runs were repeated with inclusion of a quality control.
Immunohistochemistry
Inhibin/activin.
And A subunits were immunolocalized using affinity-purified rabbit polyclonal antibodies (kind gift from Prof. W. Vale, The Salk Institute for Biological Sciences, La Jolla, CA), as previously described (20) with modifications. In brief, microwave antigen-retrieval was performed, and antibodies applied at 2 μg/ml in nonimmune block containing 10% goat serum (Sigma), 2% human serum, and 0.1% Tween 20 (Bio-Rad) in Tris-buffered saline. Rabbit IgG (Dako, Glostrup, Denmark) at 2 μg/ml served as a negative control. Antibody binding was detected by sequential application of biotinylated goat antirabbit IgG (Dako) and avidin-biotin-peroxidase conjugate (Dako), with chromogen diaminobenzidine (Dako). Tissue sections were counterstained with Harris’s hematoxylin.
MMPs.
A similar protocol was used, with the following adaptations. Rabbit polyclonal antihuman MMP-7 antibody (kind gift from Prof. F. Woessner, Department of Biochemistry and Molecular Biology, University of Miami, Miami, FL) was applied at 0.5 μg/ml. MMP-2, -3, and -9 were detected using monoclonal antibodies (Merck Biosciences, Kilsyth, Victoria, Australia), diluted between 2 and 4 μg/ml (21, 22, 23). No microwaving was performed for MMP-9. Horse serum (Sigma) was used in the nonimmune block, and for MMP-2, no Tween 20 was included. The secondary antibody was biotinylated horse antimouse IgG (Vector Laboratories, Burlingame, CA). Rabbit/mouse IgG at matching concentrations to primary antibodies were used as negative controls
Cytokeratin.
An identical protocol was used as for MMP-9, with a pan-cytokeratin antibody (Dako) diluted 1:4 in nonimmune block after enzymatic antigen retrieval (0.1% trypsin in 0.1% CaCl2 for 15 min at 37 C).
Statistics
Data were analyzed by ANOVA with Tukey’s post hoc test (activin, inhibin, and follistatin treatments) or by paired Student’s t test (decidualization experiments).
Results
Initial experiments were performed to verify that expression of MMPs and inhibin/activin subunits was not disrupted by cell isolation and culture. Culture conditions were optimized to exclude all agents (e.g. fetal calf serum, BSA) known to artificially stimulate activins/MMPs. This resulted in low basal expression of inhibins, activins, and MMPs, resembling the in vivo expression of immunoreactive protein (20, 24).
Effect of activin on MMP production
Stromal cells expressed MMP-2 and -3 mRNA (data not shown), and protein production of proMMP-2 and -3 only was confirmed by zymography (Fig. 1A). Treatment of stromal cells with activin A stimulated proMMP-2 secretion (Fig. 1A). At 2 nM, activin A induced a 2.2-fold increase in proMMP-2 (Fig. 1B, n = 4 cell preparations, P < 0.05 vs. control). This dose was thereafter used for all analyses. Inhibin A (2 nM) significantly reduced proMMP-2 (P < 0.05 vs. control), and cotreatment of inhibin A with activin A at equimolar concentrations similarly reduced proMMP-2 (P < 0.05 vs. activin alone; Fig. 1B). Follistatin alone (2 nM) reduced proMMP-2 levels (not significant). Active MMP-2 was detected in only one of four stromal cell preparations, where it followed the same pattern as proMMP-2. No significant differences in proMMP-3 levels were detected after treatments (Fig. 1A).
The effect of activin A was compared with that of TGF-1: proMMP-2 and proMMP-3 were stimulated by 1 and 2 nM of activin A, but reduced by similar concentrations of TGF-1. A representative gelatin zymogram is shown in Fig. 1C. Due to considerable variation in the degree of response between cell preparations, representative data of two experiments for both proMMP-2 and proMMP-3 are shown in Fig. 1D. As before, active MMP-2 was detected in one of two cell preparations, and followed the same pattern as proMMP-2 (Fig. 1C).
Epithelial cells expressed mRNA for MMPs-2, -3, -7, and -9 (not shown), and these were all detectable at the protein level by zymography (Fig. 2, A and C). Epithelial cells (n = 4–6 cell preparations) were treated with activin A, inhibin A, both together and follistatin (2 nM). Activin A significantly stimulated production of proMMP-2, -3, and -7 (P < 0.05 vs. control), and nonsignificantly stimulated proMMP-9 and active MMP-2 (Fig. 2, B and D). Inhibin A alone potently inhibited production and activation of proMMP-2 only, but antagonized exogenous activin-stimulation of all MMPs (significant for MMPs-2, -3, -7; P < 0.05 vs. activin alone). Follistatin did not significantly affect MMP levels in epithelial cells, in the absence of exogenous activin.
Endogenous expression of inhibin and activin mRNAs were assessed using real-time PCR: inhibin mRNA was virtually absent in epithelial and stromal cells (<0.05 fg/pg 18S), whereas activin A mRNA was abundant in epithelial cells (150 fg/pg 18S vs. 3 fg/pg 18S in stromal cells) (not shown). This was verified by measurement of dimeric inhibin A and activin A secretion by ELISA: inhibin A levels were 5–32 pg/105 cells from endometrial cells, whereas activin A was secreted at levels of 5000 pg/105 cells from epithelial and 1000 pg/105 cells from stromal cells (not shown).
Activin and MMPs in decidualized stromal cells
To investigate the relationship among inhibin, activin, and MMPs after decidualization, highly decidualized cells, as determined by prolactin secretion, were generated by treatment with cAMP for 8d (n = 3; Fig. 3A). Expression levels of inhibin /activin A subunits and MMP-2, -3, -9 were quantitated by real-time PCR on d 8. Inhibin subunit mRNA expression was very low in decidual cells (0.2 fg/pg 18S, not shown), whereas activin A subunit mRNA was up-regulated more than 7-fold after cAMP-induced decidualization (P < 0.05; Fig. 3B). MMP-2, -3, -9 mRNAs were similarly up-regulated in decidualized cells, but this failed to reach significance due to variability between individual cell preparations (Fig. 3C). This pattern was confirmed at the protein level: inhibin A secretion was minimal from decidual cells (32 pg/105 cells, not shown), activin A has previously been shown to be up-regulated approximately 5-fold to 5000 pg/105 cells (4). ProMMP-2, -3, and -9, and active MMP-2 were up-regulated with decidualization (representative zymograms, Fig. 3D). ProMMP-1 was detected at low levels and was unchanged with decidualization. No MMP-7 mRNA or protein was detected, consistent with it being epithelial specific. Moreover, activin A subunit and MMP-2, -3, -9 immunolocalized to decidual cells of early pregnancy (Fig. 4, A–F). No immunoreactive inhibin -subunit or MMP-7 were detected.
To investigate whether MMPs were potential downstream mediators of activin-promoted decidualization, MMP levels were assessed where progesterone-induced decidualization was accelerated (2- to 3-fold) by addition of activin A (n = 2), or attenuated (25%) by neutralization of activin bioactivity with follistatin (n = 2) (4). Activin A (0.4–4 nM) stimulated proMMP-2 secretion and activation on d 8 of decidualization (Fig. 3E). Furthermore, proMMP-2 secretion was reduced on d 8 of decidualization in the presence of follistatin (Fig. 3F).
Activin, inhibin, and MMPs during trophoblast invasion
Spatial expression of MMP-2, -7, -9, and inhibin /activin A subunits were assessed in trophoblast subtypes in serial sections of implantation sites from normal human pregnancies. Trophoblast cells were identified by immunostaining for cytokeratin.
Immunostaining for activin A subunit was absent in anchoring cell columns (Fig. 4H) but was strongly expressed, together with -subunit, in trophoblast cells undergoing fusion at the tips and edges of the columns (Fig. 4, H and I). Cytotrophoblast cells in the upper cell column were negative for MMP-2, but it was strikingly up-regulated in distal cells and EVTs breaking-off and invading decidua (Fig. 4, J and K). MMP-7 (not shown) and -9 (Fig. 4L) were strongly expressed by all column cytotrophoblast cells.
iEVTs invading decidua did not express inhibin -subunit, and only weakly expressed activin A (Fig. 4, N and P). However, iEVTs were intensely stained for MMP-2 (Fig. 4, Q and S), whereas staining for MMP-7 (not shown) and MMP-9 was faint or absent (Fig. 4U). Conversely, strong immunostaining for activin A and inhibin was detected in vEVT plugs (Fig. 5, B and C). MMP-2 expression was strikingly diminished in vEVTs (Fig. 5D), whereas MMPs-7 and 9 were maintained (Fig. 5, E and F).
In addition, inhibin and activin A subunits immunostained enlarged mono- and multinucleated trophoblast cells, identified as cytotrophoblast giant cells (Fig. 5, G and H) and syncytial fragments (Fig. 5I). No staining was observed for MMPs (not shown). Leukocytes within villi and decidua intensely stained for MMPs and activin A-subunit. Specificity of immunostaining for inhibin subunits was confirmed by their selective staining in syncytio- but not villous cytotrophoblast (25) (Fig. 5, K and L). MMPs were not produced by syncytiotrophoblast; however, villous cytotrophoblasts were strongly positive for MMP-7 (not shown) and -9 (Fig. 5N). These data are summarized in Fig. 6A.
Discussion
This study demonstrates that activin A stimulates MMP production (proMMP-2, -3, -7, -9) by endometrial cells. MMP-2 activation was also enhanced by activin, in parallel to increased availability of latent form. Furthermore, inhibin A antagonizes activin-induction of MMPs and is a potent inhibitor of proMMP-2. MMP up-regulation by activin A is consistent with actions in other cell types (10, 11), suggesting modulation of MMPs is a prime downstream action of activin during tissue remodeling, e.g. wound healing, embryogenesis, and inflammation. Activin and TGFs (9) therefore appear to be opposing regulators of MMPs in endometrial cells, consistent with their contrasting effects on cytotrophoblast invasion (10, 26) and placental steroidogenesis (27). This indicates divergent signaling pathways for activins and TGFs in endometrium and placenta, either through differential phosphorylation of Smads, as in pituitary cells (28), or through use of specific downstream cofactors.
Inhibin was a potent antagonist of activin in endometrial cells, overriding the stimulatory effect of activin on MMPs at equimolar concentrations. Both endometrium and placenta express activin type II receptors and betaglycan (29, 30), conferring sensitivity to inhibin-mediated antagonism of activin action (31). However, this is the first demonstration of a function for inhibin within the uterus. A role for auto/paracrine inhibin is doubtful because endometrial inhibin levels are limited. However, inhibin is a potent endocrine hormone, and circulates at high levels throughout the menstrual cycle when MMP production is suppressed (32). Progesterone withdrawal premenstrually releases MMP suppression in the endometrium, and triggers MMP-mediated endometrial breakdown (16). Because inhibin A levels mirror those of progesterone in the luteal phase, a parallel role may be anticipated. Moreover, the sharp increase in inhibin B (believed to have similar actions to inhibin A) in the late menstrual phase (33) potentially represents a previously unrecognized endocrine mechanism for the cessation of menstruation-associated MMP production.
Although all MMPs were stimulated by activin, MMP-2 was more sensitive to inhibition by inhibin than other MMPs. That MMPs are differentially regulated is well established (reviewed in Ref.34). However, MMP-2 is unusual in that its promoter does not contain the activator protein-1 or polyoma-enhancing activity-3/E26 virus elements found in most other MMPs expressed in the endometrium, although it does have two SP1-binding sites (35). In some cell types, MMP-2 expression can be induced by TGF (36), but in general expression levels are little affected by the action of growth factors and cytokines. Clearly, the balance of stimulatory and inhibitory factors in a particular microenvironment will determine MMP output. This is undoubtedly related to distinct functions; for example in the placenta it appears that MMP-2 expression corresponds to cytotrophoblast invasive potential, whereas MMP-7 and -9 are maximal in static but metabolically active cells (e.g. villous cytotrophoblasts, vEVTs).
Inhibin A was a more potent antagonist in stromal than epithelial cells; no inhibitory effect was observed on epithelial MMP-3, -7, -9 in the absence of exogenous activin. This probably relates to differences in endogenous activin levels. Although activin production by epithelial cells is approximately 5 times higher than by stromal cells, follistatin is coexpressed in endometrial epithelial, but not stromal, cells (29). Thus, bioactive activin may be negligible in epithelial-conditioned media. This concept is also supported by the failure of follistatin treatment to suppress epithelial MMPs.
Activin A is dramatically up-regulated in decidualizing cells and promotes decidual progression (4). Inhibin expression is limited in the decidua, suggesting an activin-dominated environment. Coincidental to activin up-regulation, we observe an elevation of decidual MMP-2, -3, -9, and indeed MMP-2 production and activation is further increased by treatment of decidualizing cells with activin. Inhibition of MMPs impedes the decidualization process (5, 6, 7). A similar attenuation of decidualization occurs with activin neutralization (4, 37), which we demonstrate here results in diminished proMMP-2. These data reinforce a physiological interaction between activins and MMPs during decidualization and support the hypothesis that activin promotes decidual progression, at least in part, through elevating MMP production.
The placenta is a significant source of inhibins/activins, primarily from syncytiotrophoblast (25), and roles in regulating placental steroidogenesis have been identified (38). Furthermore, exogenous activin promotes differentiation of EVTs and MMP-2 secretion, in an in vitro model of trophoblast invasion (10). Our in vivo data support this, demonstrating that maternal decidua is the prime source of activin, and that MMP-2 is strikingly up-regulated in cytotrophoblast cells as they contact and invade through the activin-rich decidua. We therefore propose that activin is an important constituent of decidual secretions, stimulating trophoblast MMPs (14).
Immunohistochemistry is invaluable for determining protein expression by individual cell populations in vivo. In early implantation sites, distinct trophoblast populations clearly expressed specific repertoires of inhibin, activin, and MMPs. MMP-2 appeared to be associated with cytotrophoblast migration, consistent with its abundance in iEVTs in ectopic pregnancies (39). Conversely, MMP-7 and -9 were expressed by nonmotile cytotrophoblasts, suggesting roles in processing, rather than invasion, and indicating differential regulation of MMPs by the microenvironmental milieu of cytokines/growth factors. These data contrast with studies using primary cytotrophoblast cells, which attribute MMP-9 with invasive potential (13, 40). However, isolated cytotrophoblast cells are likely to be a mixed population, and we demonstrate that MMP-9 is abundant in all cytotrophoblast subtypes, except for the minor iEVT population. MMP-9 levels are reduced in isolated cytotrophoblasts with advancing gestation: however, a variety of other MMPs (including MMP-2) are strictly down-regulated after the first trimester (41). Moreover trophoblast invasion is limited in vitro by neutralization of MMP-2 (42). This indicates species-specific differences in placentation because MMP-9 is the predominant MMP in murine invasive giant cells (6).
Absence of MMP-2, but maintenance of MMP-7 and -9, was observed in endovascular aggregates of vEVTs, consistent with their cessation of migration. Interestingly, strong immunoreactivity for inhibin/activin subunits was simultaneously observed in vEVTs. A similar expression profile is seen in human giant cells—the terminal differentiation state of iEVTs—again associated with loss of migratory potential (12). EVT invasion must be tightly limited, to prevent overinvasion into myometrium and vasculature. Given the increased production of inhibin subunits, with a paralleled decrease in MMP-2 production in immobilized EVT populations, it is tempting to hypothesize that inhibin contributes to diminished trophoblast invasiveness.
The trigger for inhibin subunit expression in cytotrophoblasts is unknown. However, the implantation site is maintained in a hypoxic state, and this is believed to be critical for regulating trophoblast differentiation (43). Inducing placental hypoxia in vitro significantly suppresses inhibin and activin production (44). This suggests that inhibin/activin may be up-regulated as vEVTs invade maternal vessels and encounter an oxygen-rich environment. Giant cells also produce inhibin/activin subunits. These enlarged cells are multinucleated, and a similar up-regulation of inhibin/activin subunits occurs with syncytialization in vivo and in vitro (45). Whether this is cause or effect is not known, as syncytial formation has not been clearly defined.
In summary, we show that activin A stimulates endometrial MMPs, and inhibin potently antagonizes activin action. The relationship between activin A and MMP-2 suggests MMP up-regulation is a downstream event during activin-stimulated decidualization. Spatial expression patterns of inhibin/activin and MMPs in intrauterine implantation sites demonstrate that cytotrophoblast MMP-2 is up-regulated by contact with decidua; thus, we propose that decidual-derived activin is important for promoting iEVT invasion (Fig. 6B). The physiological significance of the inhibition of MMP-2 by inhibin is reinforced by their lack of colocalization in cytotrophoblast populations, indicating that autocrine inhibin production may be a critical switch during trophoblast differentiation. Thus, these data highlight a number of mechanisms through which activins and inhibins may contribute to the bidirectional communication between maternal and fetal cells during the establishment of pregnancy.
Acknowledgments
We acknowledge the technical assistance of Tu’uhevaha Kaitu’u, Premila Paiva, Natalie Hannan, Ashwini Chand, and Sue Panckridge.
Footnotes
This work was supported by National Health & Medical Research Council (NH&MRC) of Australia, Royal Australian and New Zealand College of Obstetrics and Gynaecology). Studies were funded by NH&MRC Program Grant (No. 241000). R.L.J. was the recipient of the Ella McKnight Scholarship (Royal Australian and New Zealand College of Obstetrics & Gynaecology). L.A.S., D.M.R., and J.F.K. are supported by Fellowship Nos. 143798, 198707 and 198705 from NH&MRC.
Disclosure: R.L.J., J.K.F., P.G.F., D.M.R., E.W., and L.A.S. have nothing to declare.
First Published Online November 10, 2005
Abbreviations: ECM, Extracellular matrix; EVT, extravillous cytotrophoblast; iEVT, interstitial; MMP, matrix metalloproteinase; rh, recombinant human; vEVT, endovascular.
Accepted for publication October 31, 2005.
References
Sharkey AM, Smith SK 2003 The endometrium as a cause of implantation failure. Best Pract Res Clin Obstet Gynaecol 17:289–307
Dimitriadis E, Robb L, Salamonsen LA 2002 Interleukin 11 advances progesterone-induced decidualization of human endometrial stromal cells. Mol Hum Reprod 8:636–643
Zoumakis E, Margioris AN, Stournaras C, Dermitzaki E, Angelakis E, Makrigiannakis A, Koumantakis E, Gravanis A 2000 Corticotrophin-releasing hormone (CRH) interacts with inflammatory prostaglandins and interleukins and affects the decidualization of human endometrial stroma. Mol Hum Reprod 6:344–351
Jones RL, Salamonsen LA, Findlay JK 2002 Activin A promotes human endometrial stromal cell decidualization in vitro. J Clin Endocrinol Metab 87:4001–4004
Rechtman MP, Zhang J, Salamonsen LA 1999 Effect of inhibition of matrix metalloproteinases on endometrial decidualization and implantation in mated rats. J Reprod Fertil 117:169–177
Alexander CM, Hansell EJ, Behrendtsen O, Flannery ML, Kishnani NS, Hawkes SP, Werb Z 1996 Expression and function of matrix metalloproteinases and their inhibitors at the maternal-embryonic boundary during mouse embryo implantation. Development 122:1723–1736
Strakova Z, Szmidt M, Srisuparp S, Fazleabas AT 2003 Inhibition of matrix metalloproteinases prevents the synthesis of insulin-like growth factor binding protein-1 during decidualization in the baboon. Endocrinology 144:5339–5346
McCawley LJ, Matrisian LM 2001 Matrix metalloproteinases: they’re not just for matrix anymore! Curr Opin Cell Biol 13:534–540
Osteen KG, Igarashi TM, Bruner-Tran KL 2003 Progesterone action in the human endometrium: induction of a unique tissue environment which limits matrix metalloproteinase (MMP) expression. Front Biosci 8:d78–d86
Caniggia I, Lye SJ, Cross JC 1997 Activin is a local regulator of human cytotrophoblast cell differentiation. Endocrinology 138:3976–3986
Ogawa K, Funaba M, Mathews LS, Mizutani T 2000 Activin A stimulates type IV collagenase (matrix metalloproteinase-2) production in mouse peritoneal macrophages. J Immunol 165:2997–3003
Loke YW, King A 1995 Human implantation. Cell biology and immunology. Cambridge University Press, Cambridge
Fisher SJ, Leitch MS, Kantor MS, Basbaum CB, Kramer RH 1985 Degradation of extracellular matrix by the trophoblastic cells of first-trimester human placentas. J Cell Biochem 27:31–41
Bischof P, Meisser A, Campana A, Tseng L 1998 Effects of decidua-conditioned medium and insulin-like growth factor binding protein-1 on trophoblastic matrix metalloproteinases and their inhibitors. Placenta 19:457–464
Marsh MM, Hampton AL, Riley SC, Findlay JK, Salamonsen LA 1994 Production and characterization of endothelin released by human endometrial epithelial cells in culture. J Clin Endocrinol Metab 79:1625–1631
Salamonsen LA, Butt AR, Hammond FR, Garcia S, Zhang J 1997 Production of endometrial matrix metalloproteinases, but not their tissue inhibitors, is modulated by progesterone withdrawal in an in vitro model for menstruation. J Clin Endocrinol Metab 82:1409–1415
Landgren BM, Collins A, Csemiczky G, Burger HG, Baksheev L, Robertson DM 2004 Menopause transition: annual changes in serum hormonal patterns over the menstrual cycle in women during a nine-year period prior to menopause. J Clin Endocrinol Metab 89:2763–2769
de Winter JP, ten Dijke P, de Vries CJ, van Achterberg TA, Sugino H, de Waele P, Huylebroeck D, Verschueren K, van den Eijnden-van Raaij AJ 1996 Follistatins neutralize activin bioactivity by inhibition of activin binding to its type II receptors. Mol Cell Endocrinol 116:105–114
Jones RL, Hannan NJ, Kaitu’u TJ, Zhang J, Salamonsen LA 2004 Identification of chemokines important for leukocyte recruitment to the human endometrium at the times of embryo implantation and menstruation. J Clin Endocrinol Metab 89:6155–6167
Jones RL, Salamonsen LA, Critchley HO, Rogers PA, Affandi B, Findlay JK 2000 Inhibin and activin subunits are differentially expressed in endometrial cells and leukocytes during the menstrual cycle, in early pregnancy and in women using progestin-only contraception. Mol Hum Reprod 6:1107–1117
Vincent AJ, Malakooti N, Zhang J, Rogers PA, Affandi B, Salamonsen LA 1999 Endometrial breakdown in women using Norplant is associated with migratory cells expressing matrix metalloproteinase-9 (gelatinase B). Hum Reprod 14:807–815
Vincent AJ, Zhang J, Ostor A, Rogers PA, Affandi B, Kovacs G, Salamonsen LA 2000 Matrix metalloproteinase-1 and -3 and mast cells are present in the endometrium of women using progestin-only contraceptives. Hum Reprod 15:123–130
Zhang J, Hampton AL, Nie G, Salamonsen LA 2000 Progesterone inhibits activation of latent matrix metalloproteinase (MMP)-2 by membrane-type 1 MMP: enzymes coordinately expressed in human endometrium. Biol Reprod 62:85–94
Salamonsen LA, Woolley DE 1996 Matrix metalloproteinases in normal menstruation. Hum Reprod 11(Suppl 2):124–133
Minami S, Yamoto M, Nakano R 1992 Immunohistochemical localization of inhibin/activin subunits in human placenta. Obstet Gynecol 80:410–414
Chakraborty C, Gleeson LM, McKinnon T, Lala PK 2002 Regulation of human trophoblast migration and invasiveness. Can J Physiol Pharmacol 80:116–124
Song Y, Keelan J, France JT 1996 Activin-A stimulates, while transforming growth factor 1 inhibits, chorionic gonadotrophin production and aromatase activity in cultured human placental trophoblasts. Placenta 17:603–610
Suszko MI, Balkin DM, Chen Y, Woodruff TK 2005 Smad3 mediates activin-induced transcription of follicle-stimulating hormone -subunit gene. Mol Endocrinol 19:1849–1858
Jones RL, Salamonsen LA, Zhao YC, Ethier JF, Drummond AE, Findlay JK 2002 Expression of activin receptors, follistatin and betaglycan by human endometrial stromal cells; consistent with a role for activins during decidualization. Mol Hum Reprod 8:363–374
Ciarmela P, Florio P, Toti P, Grasso D, Santopietro R, Tosi P, Petraglia F 2003 Expression of betaglycan in pregnant tissues throughout gestation. Eur J Endocrinol 149:433–437
Lewis KA, Gray PC, Blount AL, MacConell LA, Wiater E, Bilezikjian LM, Vale W 2000 Betaglycan binds inhibin and can mediate functional antagonism of activin signalling. Nature 404:411–414
Muttukrishna S, Fowler PA, Groome NP, Mitchell GG, Robertson WR, Knight PG 1994 Serum concentrations of dimeric inhibin during the spontaneous human menstrual cycle and after treatment with exogenous gonadotrophin. Hum Reprod 9:1634–1642
Groome NP, Illingworth PJ, O’Brien M, Pai R, Rodger FE, Mather JP, McNeilly AS 1996 Measurement of dimeric inhibin B throughout the human menstrual cycle. J Clin Endocrinol Metab 81:1401–1405
Curry Jr TE, Osteen KG 2003 The matrix metalloproteinase system: changes, regulation, and impact throughout the ovarian and uterine reproductive cycle. Endocr Rev 24:428–465
Murphy G 2004 Gelatinase A. London: Elsevier Academic Press
Overall CM, Wrana JL, Sodek J 1991 Transcriptional and post-transcriptional regulation of 72-kDa gelatinase/type IV collagenase by transforming growth factor-1 in human fibroblasts. Comparisons with collagenase and tissue inhibitor of matrix metalloproteinase gene expression. J Biol Chem 266:14064–14071
Tierney EP, Giudice LC 2004 Role of activin A as a mediator of in vitro endometrial stromal cell decidualization via the cyclic adenosine monophosphate pathway. Fertil Steril 81(Suppl 1):899–903
Ni X, Luo S, Minegishi T, Peng C 2000 Activin A in JEG-3 cells: potential role as an autocrine regulator of steroidogenesis in humans. Biol Reprod 62:1224–1230
Bai SX, Wang YL, Qin L, Xiao ZJ, Herva R, Piao YS 2005 Dynamic expression of matrix metalloproteinases (MMP-2, -9 and -14) and the tissue inhibitors of MMPs (TIMP-1, -2 and -3) at the implantation site during tubal pregnancy. Reproduction 129:103–113
Librach CL, Werb Z, Fitzgerald ML, Chiu K, Corwin NM, Esteves RA, Grobelny D, Galardy R, Damsky CH, Fisher SJ 1991 92-kD type IV collagenase mediates invasion of human cytotrophoblasts. J Cell Biol 113:437–449
Fisher SJ, Cui TY, Zhang L, Hartman L, Grahl K, Zhang GY, Tarpey J, Damsky CH 1989 Adhesive and degradative properties of human placental cytotrophoblast cells in vitro. J Cell Biol 109:891–902
Isaka K, Usuda S, Ito H, Sagawa Y, Nakamura H, Nishi H, Suzuki Y, Li YF, Takayama M 2003 Expression and activity of matrix metalloproteinase 2 and 9 in human trophoblasts. Placenta 24:53–64
Adelman DM, Gertsenstein M, Nagy A, Simon MC, Maltepe E 2000 Placental cell fates are regulated in vivo by HIF-mediated hypoxia responses. Genes Dev 14:3191–3203
Manuelpillai U, Schneider-Kolsky M, Thirunavukarasu P, Dole A, Waldron K, Wallace EM 2003 Effect of hypoxia on placental activin A, inhibin A and follistatin synthesis. Placenta 24:77–83
Debieve F, Pampfer S, Thomas K 2000 Inhibin and activin production and subunit expression in human placental cells cultured in vitro. Mol Hum Reprod 6:743–749(Rebecca L. Jones, Jock K. Findlay, Paul )