Quantitative evaluation of the influence of ovarian steroids on plasminogen activators and inhibitors in human endometrial cells and trophoblasts

Thrombosis Research 108 (2003) 235 – 244

Regular Article

Quantitative evaluation of the influence of ovarian steroids on plasminogen activators and inhibitors in human endometrial cells and trophoblasts Mamoru Ueyama a,*, Noriko Kasatori b, Tsutomu Urayama b, Toshimitsu Maemura a, Yoichiro Yao a, Toshiko Shiraishi a, Sinichi Saito a, Harumi Kubo a a

1st Department of Obstetrics and Gynecology, Toho University School of Medicine, 6-11-1 Omori-Nishi, Ota, Tokyo 1430015, Japan b Department of Laboratory Medicine, Toho University School of Medicine, 6-11-1 Omori-Nishi, Ota, Tokyo 1430015, Japan Received 17 June 2002; received in revised form 7 November 2002; accepted 4 January 2003

Abstract Introduction: Plasminogen activators and inhibitors were quantitated in cultured human endometrial and trophoblast cells under the influence of ovarian steroids in order to investigate the role of the fibrinolytic system for trophoblast invasion and anchorage. Materials and methods: Plasminogen activators (t-PA and u-PA) and their inhibitors (PAI-1 and PAI-2) secretions were assayed in cultures of epithelial, stromal, and trophoblast cells. These cells were also cultured on a fibrin substrate for microscopic examination of the fibrinolytic degradation. Results: The u-PA from epithelial cells was predominant among PAs and PAI-1 in endometrial cells. Estradiol (E2) enhanced t-PA production in stromal cells and PAI-1 production in epithelial cells. Progesterone (P4) suppressed u-PA production in epithelial cells and enhanced PAI-1 production in both epithelial and stromal cells. Trophoblasts produced PAI-1, PAI-2, and small quantities of t-PA and u-PA, none of which were notably influenced by E2 or P4. The PAI-1 production in trophoblasts was more than four-fold greater than the u-PA production in epithelial cells. Epithelial and stromal cells initially grew on fibrin substrate but were gradually detached from the substrate with fibrinolytic degradation, with the exception of the stromal cells grown in the presence of P4 (or E2 + P4). Trophoblasts grew well on fibrin substrate without fibrinolytic degradation both in the presence and absence of the steroids tested. Conclusions: Fibrinolytic balance seemed to be basically maintained between the endometrial PAs and the relative excess of trophoblasts-derived PAI-1. This balance might be regulated principally by P4 and focally by E2 in the endometrial tissue for placental implantation. D 2003 Elsevier Science Ltd. All rights reserved. Keywords: Endometrial cell; Trophoblast; Plasminogen activator; Plasminogen activator inhibitor; Fibrinolysis; Ovarian steroids

1. Introduction Trophoblasts invade the extracellular matrix (ECM) of the interepithelial cells, basement membrane, stroma, and vasculature in the endometrium. This invasive activity is strictly controlled for the implantation in the endometrium tissue where abundant maternal blood is supplied to the stromal Abbreviations: PA, plasminogen activator; u-PA, urokinase type plasminogen activator; t-PA, tissue type plasminogen activator; PAI, plasminogen activator inhibitor; PAI-1, plasminogen activator inhibitor-1; PAI-2, plasminogen activator inhibitor-2; E, estrogen; E1, estrone; E2, estradiol; E3, estriol; P4, progesterone; EIA, enzyme immunoassay; CEA, carcino embryonic antigen; hCG, human chorionic gonadotropin; hPL, human placental lactogen. * Corresponding author. Tel.: +81-3-3762-4151; fax: +81-3-3765-7671. E-mail address: [email protected] (M. Ueyama).

tissue without hemorrhage. The interaction of trophoblasts and the endometrium during invasion and implantation must be kept under hormonal control, especially ovarian steroids that regulate plasmin and other local proteases. Tissue type plasminogen activator (t-PA) and urokinase type PA (u-PA) mediate fibrinolysis and ECM degradation, respectively, by converting inactive plasminogen to active plasmin [1]. The t-PA activates fibrin-linked plasminogen and exerts a strong fibrinolytic activity. The u-PA bound to the specific membrane receptors concentrated at the leading edges of the invading cells [2] activates plasminogen in the vicinity of the cells [3], thereby facilitating the enhancement of ECM degradation by the plasmin [4] and the plasmin-catalyzed activation of the latent matrix metalloproteinase [5]. In studies on the trophoblast invasion into the endometrium using mouse embryos, it has been confirmed that PA

0049-3848/03/$ - see front matter D 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0049-3848(03)00029-X

236

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

production from the trophoblasts involves the invasion of blastocyst [6] and is decreased in embryos that have not been implanted [7]. In studies on the roles of PA in human embryo implantation, Martin and Arias [8], and Queenan et al. [9] found that cultured trophoblasts secreted PA, and Yagel et al. [10] observed that anti-PA antibody inhibited the trophoblast invasiveness in vitro. These findings suggest that the PA activity derived from the trophoblast is closely associated with cell invasion. This PA activity is inhibited by PAI through the formation of a complex with PA molecule [11]. It could be conjectured that endometrial hemostasis, inhibition of trophoblast invasiveness, and formation of the chorionic placenta are facilitated via PAI production from decidual cells as well as from trophoblasts. On the other hand, Feinberg et al. [12] reported that PAI-1 is an immunocytochemical marker of invading trophoblasts. PA and PAI secretions have been known to be under steroid control [13 – 22,34]. In the present study, we investigated the influences of ovarian steroids on the stoichiometries in PA and PAI productivities based on these quantitative assays and our observation of the actual fibrin-degrading activities in endometrial cells and trophoblasts cultures in vitro in order to discuss the participation of ovarian steroids in the conditions for trophoblast invasion or quiescence in the endometrium.

2. Materials and methods 2.1. Isolation of endometrial epithelial and stromal cells Human endometrial specimens were obtained from seven cycling patients (35 – 42 years old) who had undergone hysterectomies due to uterine myoma. Informed consent was obtained from all the patients. Specimens were classified as secretory endometria. Each specimen was rinsed with a large volume of calcium- and magnesium-free phosphate-buffered saline solution [PBS( )] containing 50 Ag/ml penicillin and 50 Ag/ml kanamycin. The specimens were minced into small pieces in 1 ml of 0.2% collagenase (Boehringer Mannheim)/medium [Dulbecco’s modified Eagle’s medium (D-MEM) and Ham F-12 mixture (1:1) (GIBCO)], placed in an Erlenmeyer flask containing 20 ml of the same collagenase solution, and stirred for 30 min at 37 jC. At 20 min of stirring, the contents were pipetted several times using a Pasteur pipette to disperse cell clusters. The cell suspension was centrifuged at 1200 rpm for 5 min, and the pellet obtained was washed in PBS( ). After centrifugation in the same manner, the cell pellets were resuspended in the medium containing l0% fetal calf serum (FCS) (GIBCO), dispensed into 6 cm diameter culture dishes, and incubated at 37 jC, in air with 5% CO2 to separate epithelial cells from stromal cells. Epithelial cells were isolated according to the method of Vigano et al. [23]. Rod-like clusters of cells not adhering to the dish from the mixed culture of epithelial and stromal

cells were collected under microscopy at l h intervals and incubated in separate dishes. After repeating this procedure several times, the adhering cells markedly decreased in number, and a great number of rod-like cell clusters were obtained. These cell clusters were collected by centrifugation, and the cells were dispersed in 0.25% trypsin (Difco)/ PBS( ) and centrifuged at 800 rpm for 5 min. The resulting pellet was dispersed to a homogeneous epithelial cell suspension in the medium. Primary stromal cells adhering to the dish were continuously incubated until reaching confluence, then dispersed to the cell suspension in the medium by trypsin digestion as described above. The living cells in both cell suspensions were counted with 0.5% trypan blue solution on a modified Neubauer’s hemocytometer. 2.2. Isolation and culture of trophoblast cells from chorionic villi in the late implantation phase Human early placenta1 tissue was obtained under sterile conditions by artificial abortion at 7 weeks of gestation from seven healthy women (23 –29 years) who had no complications requiring steroid therapy for autoimmune disease or allergic disease. The tissues were used as experimental materials after consent was obtained. Each specimen was thoroughly washed in 50 ml of Hanks’ solution containing 50 Ag/ml penicillin, 100 Ag/ml kanamycin, 20 mM HEPES, and 25 mM glucose without calcium or magnesium (H-G-Hanks; pH 7.2) and transferred into a dish containing 10 ml of new H-G-Hanks. Chorionic villi separated from the surrounding connective tissue under stereoscopic microscopy were minced into small pieces in 1 ml of medium containing 2000 U/ml dispase (Goudoushusei), placed in a centrifuge tube containing 50 ml of the same dispase solution supplemented with 80 kU/ml deoxyribonucleosidase 1 (DNase 1, Sigma), divided into two centrifuge tubes containing 25 ml of the same dispase solution, and shaken at 120 cycles/min in a shaking water bath at 37 jC for 20 min to disperse the cells. Next, after passing the cell suspension through a 38 Am pore-size mesh, the suspension was centrifuged at 1200 rpm for 5 min and the pellet obtained was resuspended in 20 ml of medium [Dulbecco’s modified Eagle’s medium (D-MEM) and Ham F-12 mixture (1:1) (GIBCO)] supplemented with 20 mM HEPES and 25 mM glucose at pH 7.2 to obtain a homogeneous cell suspension. Trophoblast cells were then isolated from this cell suspension according to the method of Kliman et al. [24]. The trophoblast cell fraction collected by Percol gradient centrifugation was transferred into a 10-ml centrifuge tube, into which medium was added to a final volume of 8 ml, and the solution was centrifuged at 1000 rpm for 10 min. After the cells were washed twice in 8 ml of medium, the pellet was suspended in 1.5 ml of medium to obtain a homogeneous trophoblast cell suspension. Viable cells in this cell

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

suspension were counted with 0.5% trypan blue solution in a modified Neubauer’s counting chamber. 2.3. Morphological observation and immunofluorescent staining of cells Two hundred fifty microliters suspension of endometrial epithelial cells, stromal cells, and trophoblast cells from five specimens at 3.0  105 cells/ml (7.5  l04 cells/well) were gently transplanted onto 12 mm diameter coverslips coated with collagen type 1 (Iwaki) in 24-well multiplates (Nunc), and preincubated at 37 jC in an atmosphere of 5% CO2 in air for 2 h. After preincubation, 2 ml of medium containing 10% FCS was gently poured into the plates and incubated for 72 h, with exchange of the medium after the first 12-h incubation. The morphology of the cultured cells was observed by inverted microscopy and photographed after 12 and 72 h of incubation. To observe the trophoblasts, the medium was gently exchanged with 20% formalin solution after 12 and 72 h of incubation, and the cells were fixed for 30 min and stained with hematoxylin and eosin (HE). To determine the ratio of syncytial cells to all cultured cells, three visual fields near the middle of each coverslip were randomly observed for each stained sample at l00 magnifications, and the mean percent of nuclei present in the apparent syncytial cells per 100 nuclei was calculated [24]. Epithelial cells and trophoblasts cultivated on coverslips were fixed with 98% ethanol and stained for the immunofluorescent studies of CEA, hCG, and hPL. 2.4. Measurement of ovarian steroids secreted from cultured trophoblasts The isolated trophoblast cells were suspended in a medium containing 10% FCS at a concentration of 6.0  104 cells/ml, and 250 Al volumes of this cell suspension (1.5  104 cells/well) were transferred into 10 wells of 96-well multiplates coated with collagen type 1 and then preincubated at 37 jC in an atmosphere of 5% CO2 in air for 72 h. Confluent layers of trophoblast cells in each well were washed three times with 250 Al of the medium containing 0.5% bovine serum albumin (BSA, Sigma), and the cells were subcultured with 200 Al of the medium containing 0.25% BSA. After incubation for 72 h, the conditioned medium (total volume 2.0 ml) was collected from each well and stored at 80 jC until measurement. E1, E2, E3, and P4 in conditioned medium obtained from cultivation performed twice under the same culture conditions were quantified in duplicate by RIA, and the mean values were obtained. 2.5. Preparation of specimens for the quantification of PA and PAI antigens Cell suspensions of the first generation of epithelial cells, the second generation of stromal cells, and the first

237

generation of trophoblasts were prepared from five specimens at a concentration of 6.0  104 cells/ml supplemented with 10% FCS. Each of these suspensions was inoculated at a volume of 250 Al (1.5  104 cells) to 10 wells in 96-well titer plates, 2 wells for control and 8 wells for four different conditions with the following steroids. The titer plates transplanted with the epithelial, stromal, and trophoblast cells were preincubated at 37 jC in an atmosphere of 5% CO2 in air for 72 h, with exchange of the medium after the first 24 h of cultivation. The cells reaching the confluent state were gently washed three times with 300 Al of serum-free medium containing 0.5% BSA, and the medium was replaced with the fresh serum-free medium supplemented with 0.25% BSA and 0.1 AM E2, 0.05 AM E2 + 0.1 AM P4, or 0.1 AM P4 into the four respective pairs of wells. After continuous incubation for 72 h, the t-PA, u-PA, PAI-l, and PAI-2 produced by the cells in the media were quantified by enzyme immunoassay. 2.6. Preparation of fibrin substrate Fibrin monomer was prepared from bovine fibrinogen according to Donelly et al. [25] and adjusted to l% protein concentration. To prepare the fibrin gel substrates, 10 Al of monomer solution was poured gently into the bottom of each well of 96-well titer plates containing 300 Al of 3 mM CaCl2 in 20 mM HEPES-saline at pH 7.2 (HEPES-Ca2 +-saline) per well and allowed to stand for 3 h at 37 jC. After washing each well three times with the HEPES-Ca2 +-saline, the crosslinked fibrin gel substrates were prepared by adding 20 U/ml of blood coagulation factor XIII (Fibrogamin, Hoechst) activated with 1 U/ml thrombin (in the same buffer) onto the fibrin gels and incubating the plates for 3 h at 37 jC. The prepared plates were washed in 5 l of HEPES-Ca2 +-saline for 12 h with stirring [16]. 2.7. Cell culture on fibrin substrate For the control group, 100 Al of medium containing 0.05 cU/ml bovine plasminogen and ethanol (0.01%, about 2-fold final concentration of solvent for steroids) was added onto the prepared fibrin substrate. For the experimental group, the same volume of medium containing 0.05 cU/ml of bovine plasminogen and 0.2 AM E2, 0.1 AM E2, 0.2 AM P4, or 0.2 AM P4, respectively, was added onto fibrin substrate. The suspensions of trophoblast cells and endometrial cells (0.1 ml of 1.5  105 cells/ml) in the medium supplemented with 20% FCS were poured into each fibrin substrate wells and incubated at 37 jC in 5% CO2 in air for 72 h. FCS used in this experiment was pretreated through a lysine sepharose column to remove endogenous plasminogen. Fibrinolytic indications were discriminated by morphological changes of the cells with detachment from degraded fibrin substrate.

238

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

3. Statistical analysis Data represent the mean value F S.D. of replicate determinations from two separate cultures in five independent experiments. Differences were considered significant (*) when the p values were below 5% in the Student’s t-test.

nofluorescent staining. On the other hand, 64 of 100 cells in 72 h cultures of trophoblasts were observed to differentiate to giant syncytial cells having more than two nuclei in the HE-stained specimens, and CEA, hCG, and hPL production were confirmed by immunofluorescent staining (Fig. 1). 4.2. PA and PAI-1 antigens secreted from endometrial epithelial cells

4. Results 4.1. Properties of the isolated cells cultured on glass coverslips The endometrial epithelial cells were polymorphic in shape and were found to produce CEA in immunofluorescent staining. The stromal cells were spindle-shaped. The trophoblasts cultured for 12 h were all mononuclear and polygonal on HE staining, and none of them showed hormone production on enzyme immunostaining or immu-

No significant difference was noted in the t-PA level among the groups supplemented with any of the steroids, or between any of the groups and the control (Table 1). The epithelial cells secreted u-PA in far greater amounts than t-PA or PAI-1. Only P4 significantly suppressed the level of u-PA secretion ( p < 0.018), while no significant differences were found between the secretions in the group supplemented with E2 and those supplemented with other steroids (E2 + P4 or P4). Supplementation with E2, E2 + P4, or P4 significantly increased the level of PAI-1 secretion compared

Fig. 1. Cultured endometrial cells and trophoblast cells. Endometrial epithelial cells and stromal cells cultured on a plastic dish were photographed under inverted phase-contrast microscopy (A and B, respectively,  100). Epithelial cells cultured on coverslips were confirmed by the immunological staining with a specific antibody for carcinoembryonic antigen (a,  400). Trophoblasts cultured on a dish coated with type 1 collagen for 12 h were all mononuclear and polygonal in shape (C). At 72 h of cultivation (D), the trophoblasts were well developed. Expressions of carcinoembryonic antigen (1), human chorionic gonadotropin (2), and human placental lactogen (3) on giant syncytial cells transformed from trophoblasts cultured for 72 h on coverslip-coated with type 1 collagen were confirmed by immunofluorescent staining ( 400).

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

239

Table 1 Amounts of t-PA, u-PA and PAI-1 antigens secreted in the medium from endometrial epithelial cells incubated for 72 h t-PA antigen (ng/ml)/ 1.5  104 cells/well control E2 E2 + P4 P4

p value *

24.8 F 2.2 24.3 F 3.5 24.8 F 2.5 22.8 F 1.2

NS NS NS

u-PA antigen (ng/ml)/ 1.5  104 cells/well

**

p value *

1038.2 F 164.6 893.2 F 240.8 794.8 F 195.5 768.2 F 87.0

NS NS

**

NS NS < 0.018

NS NS

PAI-1 antigen (ng/ml)/ 1.5  104 cells/well

p value

28.6 F 2.9 36.1 F 4.0 38.1 F 2.5 40.9 F 4.9

*

**

< 0.028 < 0.004 < 0.012

NS NS

*; comparison to control. **: comparison to the group with E2. NS; not significant.

Table 2 Amounts of t-PA, u-PA and PAI-1 antigens secreted in the medium from endometrial stromal cells incubated for 72 h t-PA antigen (ng/ml)/ 1.5  104 cells/well control E2 E2 + P4 P4

25.8 F 3.8 37.0 F 3.6 21.1 F 3.0 28.3 F 3.1

p value * < 0.001 NS NS

**

u-PA antigen (ng/ml)/ 1.5  104 cells/well

< 0.001 < 0.003

1.0 F 0.3 0.7 F 0.1 0.4 F 0.1 0.6 F 0.0

p value * NS < 0.008 < 0.025

**

PAI-1 antigen (ng/ml)/ 1.5  104 cells/well

p value *

**

< 0.004 < 0.037

28.0 F 2.6 28.5 F 5.2 40.0 F 3.1 37.0 F 2.9

NS < 0.001 < 0.001

< 0.003 < 0.013

*: comparison to control; **: comparison to the group with E2; NS: not significant.

levels observed in the control ( p < 0.008, p < 0.025, respectively) and E2-supplemented group ( p < 0.004, p < 0.037, respectively). The addition of E2 + P4 or P4 significantly increased the level of PAI-1 secretion compared with the levels measured in the control ( p < 0.001) and E2 group ( p < 0.003, p < 0.013, respectively).

with that in the control group ( p < 0.028, p < 0.004, p < 0.012, respectively), whereas no differences in the secretion of this inhibitor were noted between the groups supplemented with E2 and the other steroids. 4.3. PA and PAI-1 antigens secreted from endometrial stromal cells

4.4. PA and PAI antigens secreted from trophoblasts The addition of E2 induced significantly higher level of t-PA secretion compared to the levels measured in the control ( p < 0.001), E2 + P4 ( p < 0.001), and the P4-supplemented group ( p < 0.003) (Table 2). The u-PA level was very low to begin with, and the addition of E2 + P4 or P4 decreased it even further to levels significantly below the

The level of t-PA secretion was lower than that from endometrial cells, and the addition of E2 or P4 increased it only slightly ( p < 0.05). The u-PA level was also relatively low, and the addition of any steroids increased it slightly ( p < 0.001 – 0.022). These little variations of the PA con-

Table 3 Amounts of t-PA and u-PA antigens secreted in the medium from trophoblasts incubated for 72 h t-PA antigen (ng/ml)/ 1.5  104 cells/well control E2 E2 + P4 P4

5.4 F 0.3 5.6 F 0.1 5.6 F 0.2 5.8 F 0.2

p value * < 0.049 NS < 0.011

**

u-PA antigen (ng/ml)/ 1.5  104 cells/well

p value *

**

NS NS

1.4 F 0.5 2.4 F 0.5 2.0 F 0.4 2.2 F 0.7

< 0.001 < 0.017 < 0.022

NS NS

*: comparison to control; **: comparison to the group with E2; NS: not significant.

Table 4 Amounts of PAI-1 and PAI-2 antigens secreted in the medium from trophoblasts incubated for 72 h PAI-1 antigen (ng/ml)/ 1.5  104 cells/well control E2 E2 + P4 P4

3957 F 1353 4864 F 492 4677 F 727 4421 F 782

p value * NS NS NS

**

PAI-2 antigen (ng/ml)/ 1.5  104 cells/well

*

**

NS NS

43.1 F 10.5 37.5 F 6.3 39.5 F 4.2 34.7 F 5.1

NS NS NS

NS NS

*: comparison to control; **: comparison to the group with E2; NS: not significant.

p value

240

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

Table 5 Amount of hormones secreted in the medium from trophoblasts (1.5  104 cells per well) cultured for 72 h

estrogen (pg/ml)

P4 (ng/ml) hCG (mIU/ml) hPL (Ag/ml)

E1 E2 E3

exp.1

exp.2

542 5000 27.8 130 440 0.13

444 4720 19.8 110 240 0.07

Data represents the mean values of duplicate determinations from two separate cultures in two independent experiments. (n = 2).

centrations in a very low level were not likely to influence the overall protease system except for the local activity on the cell membrane (Table 3). A large amount of PAI was produced from trophoblast cells. No significant difference was found in this PAI level among the groups supplemented with any of the steroids, or between any of the groups and the control (Table 4). 4.5. Steroids produced from cultured trophoblast cells Ovarian steroids of El, E2, E3, P4, and protein hormones of hCG and hPL were secreted from the cultured trophoblasts (Table 5). It indicated that the cytotrophoblasts in this experiment could be differentiated from those hormoneproducing giant cyntitial cells. 4.6. Morphological changes of the cells cultured on the fibrin substrate by fibrinolysis In every specimen tested with steroids, the cell aggregation and detachment of endometrial epithelial cells were similar to those observed in the control (Fig. 2). Stromal cells were aggregated by the addition of E2, while well adhesion, spreading, and proliferation were observed in the presence of E2 + P4. At 72 h of incubation, the morphology of the trophoblasts exhibited natural, uninterrupted development without cell aggregation, regardless of the presence or absence of E2, E2 + P4, or P4.

5. Discussion Implanted human trophoblasts progressively invade the epithelial ECM, basement membrane ECM, and stromal ECM into deeper layers of the endometrium. This process is partly mediated with proteolytic enzymes such as collagenase and plasmin. Since the fibrinolytic activity is increased in the early implantation phase [26] and markedly decreased in the late phase, we can speculate that the increased activity of PA may be involved in the invasion of the embryo into the endometrium [8,9]. The development of the decidualization reaction in the endometrium is closely related to trophoblast invasiveness [8]. If the PA activity in tropho-

blast or endometrium is continuously increased, local fibrinolysis progresses, and the endometrial ECM and vascular wall ECM can be expected to undergo a process of continuous degradation, eventually resulting in endometrial bleeding or ectopic implantation such as placenta accreta and placenta increta. It had been difficult to determine the biochemical characteristics of the trophoblast in vitro. Under conventional conditions, trophoblasts do not readily proliferate and it is difficult to separate single trophoblasts and keep them alive for long periods of time. In 1986, Kliman et a1. [24] isolated trophoblasts from term placentas and obtained more than 90% of syncytial cells from the trophoblasts after incubation for as long as 120 h. Since then, many investigators have cultured trophoblasts to study the trophoblast invasiveness and resting in the endometrium [10,27,28]. Studies have shown that the invasive growth of trophoblasts is inhibited by the conditioned medium of the decidual cells culture [29], and that the trophoblast-derived hCG brings about an inhibitory effect on the PA activity [27,30]. The proteins and steroids secreted from the trophoblasts in the present experiment seemed to be similar to those derived in the late implantation phase. The trophoblasts were derived from the chorionic villi at 7 weeks gestation, and 56% of those were functional syncytial cells differentiated in the 72-h subculture. Yet the prepared cytotrophoblasts were probably still in a proliferating phase, since the completion of placental construct is generally known to take about 12 weeks of gestation. The level of secreted PAI-1 was much higher than that of PAI-2 in these trophoblasts, principally because the PAI-1 is secreted extracellularly while PAI-2 is accumulated intracellularly within the trophoblast [31]. This strong antifibrinolytic environment was confirmed by the morphological stability of spreading trophoblasts in the long-term culture on fibrin substrates without degradation under all conditions described in the present study. This differs from the fibrinolysis-enhancing environments that induced cell aggregation and detachment from fibrin substrates in cultures of endometrial epithelial and stromal cells in the presence of estrogen [16]. Notwithstanding the source of these results in the culture of trophoblasts without embryonic components, this apparently strong anti-fibrinolytic capability mediated by massive secretion of PAI-1 appeared to be an independent function of the trophoblasts without influences by embryonic components during this period of late implantation phase. This could be supported by the experiments in blastulae by Kubo et al. [26], which showed that the fibrinolytic activity of trophoblast cells decreased in the late implantation phase regardless of the presence or absence of inner cell mass. This anti-fibrinolytic activity must be an optimal condition requisite for the stabilization of trophoblasts and maintenance of the placental tissue. Trophoblasts invading stroma in vivo should stop growing when sufficient levels of hCG are produced from enough proliferated trophoblasts in the decidua by the actions of

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

241

Fig. 2. Endometrial cells and trophoblasts cultured for 72 h on the fibrin substrates in the media supplemented with FCS containing 0.025 cU/ml of plasminogen ( 100). Endometrial epithelial cells were aggregated and detached from the fibrin substrate through fibrinolysis in the presence of every steroid and steroid combination examined (A-1) and under the control condition (A). Stromal cells were aggregated on fibrin substrate by fibrinolysis in the presence of E2 (B-1), and well proliferated in the presence of E2 + P4 (B-2) and under the control condition (B). Trophoblasts were well proliferated in the presence of every steroid and steroid combination examined (C-1) and under the control condition (C).

ovarian P4 and E2. This hormonal condition in vivo might correspond to the experimental condition supplemented with E2 + P4 in this study. With the addition of E2 and P4, the productions of t-PA in stromal cells were significantly enhanced, whereas t-PA and u-PA in trophoblasts were enhanced only slightly to a less than significant extent. Those PAs were, however, produced in extremely low amount compared with the level of PAI-1 secreted from

trophoblasts and the secretion of trophoblastic PAI-1 and PAI-2 was not affected by exogenous steroids. Those results as mentioned above were reconfirmed by another experiment using extravillous trophoblasts prepared by growing chorionic villous explants without enzymatic digestion (data not shown). Wilson et al. [32] observed an inhibition of hCG production in an investigation of the changes in the protein

242

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

synthesis of cells at different steroid concentrations in a culture system. The unphysiological high concentrations of steroids could possibly exert non-specific inhibitory effects on cell functions. Concentrations of ovarian steroids in the present study were set on the basis of data from other investigators or from our works described elsewhere [15,17 –20]. Activation of procollagenase by plasmin may be necessary for infiltrative invasion of trophoblasts to the adjacent tissue, and the collagenase activity in the endometrial stroma plays an important role in the development of placental formation through decomposition of interstitial collagen [33]. hCG was reported to inhibit the activity of trypsin-like proteolytic enzyme produced by cultured trophoblasts in vitro, and thereby inhibit the invasion of trophoblasts into the deep region of the endometrium by reducing the conversion from procollagenase to collagenase [27,30] (the direct inhibition of the protease enzyme by hCG was doubtful). In the present study, cultured trophoblasts produced 4.2 nM of E (0.9, 3.2, and 0.05 nM of E1, E2, and E3, respectively) and 20.0 nM of P4. These trophoblast-derived ovarian steroids seemed to be produced at an almost constant E/P4 ratio (1:5) under all experimental conditions. These cell-derived endogenous steroids might be regarded to immediately bind to the corresponding receptors on the trophoblast cell membrane. Given the influence from the conditions involving endogenously produced E and P4, final concentrations of steroids in the test samples of experimental culture systems will probably never be exactly the same as the concentrations expected based on the levels of exogenous steroids added, unless the effects of the endogenous steroids can be absolutely inhibited without disturbing all other cell functions through the use of antagonists such as tamoxifen and RU486 [18]. Orland and Fernando [17] proposed that trophoblastic PA secretion are not under exogenous E or P4 control and that PA is produced continuously in vitro as long as the trophoblast cell remains functionally active. However, the involvement of endogenously produced steroids has not been investigated. In the present study, concentrations of E2 and P4 produced in the medium after incubation for 72 h were 3.2% and 20% (molar concentration ratio) of the concentrations for exogenous application, respectively. E2, E2 + P4, and P4 exerted only a slight effect on the u-PA secretion enhancement in trophoblasts, and they exerted no significant effect on the relatively high level of PAI-1 productivity. While this may be the supplementary effect of the high-concentration of exogenous steroids or E2/P4 concentration balance on the endogenously hormone-producing cells, the variations in an extremely low level concentration of PAs from trophoblasts compared with those from epithelial or stromal cells keep us confirming this. Alternatively, It could be that sufficient PA function in endometrial epithelial cells is achieved in cooperation with t-PA and u-PA in an estrogen-dominant phase and is slightly depressed by enhanced PAI-1 in the P4-dominant environment (luteal phase).

In the early implantation phase, trophoblast invasiveness into the endometrial stroma is initiated by breakage into intercellular spaces and degradation of the ECM between epithelial cells through the local PA activity from trophoblasts [26] in cooperation with the strong PA produced from endometrial epithelial cells. The abundant trophoblastic PAI-1 in the later stage might sustain the adhesion among trophoblasts, epithelial cells, and basement membrane, thereby helping to prevent the proteolytic abrasion of those cells and the associated bleeding. Feng et al. [34] proposed that PAI-2 in unidentified cells along the maternofetal junction have a role as a protective curtain with anti-apoptotic function. The decidualization reaction was reportedly induced by P4 in cultured endometrial stromal cells, and the enzyme activity for synthesis of ECM protein and prolactin in the decidualized cells differed from the same enzymatic activity in stromal cells [18,35]. P4 was found to promote PAI production from decidual cells more than PAI production from stromal cells [21]. These reports suggested that a phenotype of the decidualized cells converted from stromal cells by P4 in vitro was similar to that decidual cells in pregnancy. P4 suppresses the production of t-PA and u-PA, and increases the production of PAI-1 in decidual cells. This P4elicited down-regulation of PA and up-regulation of PAI-1 production tends to be further accelerated by the co-administration of E2 to P4, indicating an increase of cell membrane P4 receptors in the presence of E2 [18]. Recent reports suggest that PAI-1 stimulates detachment and migration of certain tumor cells from the matrix, resulting in tumor growth promotion [36,37]. The protease system for trophoblast invasion, however, must be controlled by the enhancement of the levels of decidual PA1-1 and trophoblastic PAI-1 or PAI-2, promoted through the binding to the ECM or vitronectin in vascular ECM [38,39] or through the inducing of expression of the tissue factor of the decidual cell [40], and ultimate enhancement of the hemostatic effect that results. The enhancement of decidual cell PAI-1 expression by P4 during pregnancy has clinical importance. In preeclampsia, excessively produced placental P4 induces excessive PAI-1 expression from the trophoblast and deciduas, which, in turn, decreases trophoblast invasiveness into the stroma and spiral artery [41], subsequently leading to the formation of incomplete intervillous spaces. Moreover, the placental blood circulation is impaired by the enhanced fibrin deposition in those intervillous spaces [42]. In contrast, the decrease in decidual PAI-1 due to anaplasia or hypoplasia of the decidua might enhance trophoblast invasiveness, possibly causing massive hemorrhage such as that seen in ectopic pregnancy, atypical attachment of the placenta such as placenta increta, accrete, or premature rupture in the chorion avillosum [22,43 –45]. Floridon et al. [14] concluded that trophoblast invasion is primarily regulated by signals from decidual cells, not by PAI-1, since the

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

lack of maternal cellular PAI-1 in tubal pregnancies and excessive decidual necrosis in molar pregnancies indicate an uncontrolled placental invasion. Feng et al. [34] demonstrated that the distribution of u-PA and u-PAR mediate trophoblast invasion. Graham [46] reported that the deciduaderived transforming growth factor-h (TGF-h) inhibits trophoblast invasiveness via reduction of u-PA and increase in PAI-1 and tissue inhibitor of metalloproteinase-1 (TIMP1) in trophoblasts. Smith et al. [47] reported that TGF-h induces tissue inhibitor of metalloproteinase-1 and PAI-1, thereby inhibiting the cell growth of human trophoblastderived cell line. While E2 and P4 have practically no influence on the trophoblasts, our quantitative assay to examine the effect of these ovarian steroids on the protease system in the present study revealed that E2 was effective mostly for up-regulation of stromal t-PA and that P4 was effective for down-regulation of u-PA and up-regulation of PAI-1 in both epithelial and stromal cells. Hence, the proteolytic function varied only slightly in trophoblasts under the influence of ovarian steroids, whereas varied significantly in maternal tissues. Trophoblasts invasion probably requires PAs for penetrating the intercellular matrices or capillaries and for maintaining blood fluidity for the lowresistant blood flow in the placental intervillous space. Trophoblasts invasion also requires PAIs for stabilization and prevention of placenta from decidual hemorrhage to limit the risk of abruption or abortion. In conclusion, the results of the culture experiment with ovarian steroids seemed to agree with the earlier speculation [14,15] that trophoblasts invasion and local hemostasis are controlled by signals or expression from decidual cells, potentially via the steroids-induced response of epithelial cells and stromal cells in the decidualized endometrium tissues.

References [1] Vassalli JD, Sappino AP, Belin D. The plasminogen activator/plasmin system. J Clin Invest 1991;88:1067 – 72. [2] Estreicher A, Muhlhauser J, Carpentier JL, Orei L, Vassalli JD. The receptor for urokinase type plasminogen activator polarizes expression of the protease to the leading edge of migrating monocytes promotes degradation of enzyme inhibitor complexes. Cell Biol 1990; 111:783 – 92. [3] Kirchheimer JC, Binder BR. Function of receptor-bound urokinase. Seminar Thromb Haemost 1991;17:246 – 50. [4] Mullins DE, Rohrlich ST. The role of proteases in cellular invasiveness. Biochim Biophys Acta 1983;695:177 – 214. [5] O’Grady RL, Upfold LI, Stephens RW. Rat mammary carcinoma cells secrete active collagenase and active latent enzymes in the stroma via plasminogen activator. Int Cancer 1981;28:509 – 15. [6] Strickland S, Reich E, Sherman MI. Plasminogen activator in early embryogenesis: enzyme production by trophoblast and parietal endoderm. Cell 1976;9:231 – 40. [7] Axelrod HR. Altered trophoblast function in implantation defective mouse embryos. Dev Biol 1985;108:185 – 90. [8] Martin O, Arias F. Plasminogen activator production by trophoblast cells in vitro: effect of steroid hormones and protein synthesis inhibitors. Am J Obstet Gynecol 1982;142:402 – 9.

243

[9] Queenan Jr JT, Kao L-C, Arbolenda CE, Ulloa-Aguirre A, Golos TG, Cines DB, et al. Regulation of urokinase-type plasminogen activator production by cultured human cytotorophoblasts. J Biol Chem 1987;262:10903 – 6. [10] Yagel S, Parhar RS, Jeffrey JJ, Lala PK. Normal nonmetastatic human trophoblast cells share in vitro invasive properties of malignant cells. J Cell Physiol 1988;136:455 – 62. [11] Loskutoff DJ, Ny T, Sawdey M, Lawrence D. Fibrinolytic system of cultured endotherial cells: regulation by plasminogen activator inhibitor. J Cell Biochem 1986;32:273 – 80. [12] Feinberg RF, Kao LC, Haimowitz JE, Queenan Jr JT, Wun TC, Strauss III JF, et al. Plasminogen activator type 1 and 2 in human trophoblasts. PAI-1 is an immunocytochemical marker of invading trophoblasts. Lab Invest 1989;61:20 – 6. [13] Chen GA, Feng Q, Zhang LZ, Liu YX. Plasminogen activators and plasminogen activator inhibitor type-1 in human endometrium. She¯nglı˘ Xue´ba`o 1992;44:502 – 9. [14] Floridon C, Nieelsen O, Holund B, Sweep F, Sunde L, Thompsen SG, et al. Does plasminogen activator inhibitor-1 (PAI-1) control trophoblast invasion? A study of fetal and maternal tissue in intrauterine, tubal, and molar pregnancies. Placenta 2000;21:754 – 62. [15] Lochwood CJ, Kribun G, Schatz F. The deciduas regulate hemostasis in human endometrium. Semin Reprod Endocrinol 1999;17:45 – 51. [16] Asakawa M, Ueyama M, Hirakawa S, Momose K, Urayama T, Kasatori N. Role of endothelial and endometrial fibroblastic cells in placental fibrinolysis. Hormonal regulation of t-PA and PAI production in the cells culture. Connect Tissue 1991;22:125 – 35. [17] Orlando M, Fernando A. Plasminogen activator production by trophoblast cells in vitro: effect of steroid hormones and protein synthesis inhibitors. Am J Obstet Gynecol 1982;142:402 – 9. [18] Tseng L, Mazella J, Sun B. Modulation of aromatase activity in human endometrial stromal cells by steroid tamoxifen and RU486. Endocrinology 1986;118:1312 – 8. [19] Ueyama M, Furukawa T, Kasatori N, Urayama T, Saito S, Kubo H, et al. Effect of sex steroid hormones on the production of plasminogen activators and inhibitor in the mucosal cells of the human reproductive tract. In: Kuramoto H, Gurpide E, editors. In vitro biology of sex steroid hormone action. Tokyo: Churchill Livingstone; 1996. p. 41 – 52. [20] Ueyama M, Furukawa T, Saito S, Kubo H, Hirakawa S. Effect of sex steroid hormones and Saireito(TJ-114) on fibrinolysis and anti-fibrinolysis factors production in cultured human endosalpingian cells. Jpn J Obstet Gynecol Neonatal Hematol 1998;8:42 – 50. [21] Schatz F, Lockwood CJ. Progestin regulation of plasminogen activator inhibitor type 1 in primary cultures of endometrial stromal and decidual cells. J Clin Endocrinol Metab 1993;77:621 – 5. [22] Feng Q, Liu Y, Liu K, Byrne S, Liu G, Wang X, et al. Expression of urokinase, plasminogen activator inhibitors and urokinase receptor in pregnant monkey uterus during early placentation. Placenta 2000;21: 184 – 93. [23] Vigano P, Blasio AM, Antonio GD, Vignali M. Culture of human endometrial cells: a new simple technique to completely separate epithelial glands. Acta Obstet Gynecol Scand 1993;72:87 – 92. [24] Kliman HJ, Nestler JE, Sermasi E, Sanger JM, Strauss III JF. Purification, characterization, and in vitro differentiation of cytotrophoblasts from human term placentae. Endocrinology 1986;118:1567 – 82. [25] Donnely TH, Laskowski Jr M, Notley N, Scheraga HA. Equilibra in the fibrinogen-fibrin conversion: II. steps reversibility of the polymerization. Arch Biochem Biophys 1955;56:369 – 87. [26] Kubo H, Katayama S, Amano H, Spindle IA. Plasminogen activator activity in mouse embryos cultured on decidual cell monolayers. Acta Obstet Gynaecol Jpn 1982;34:801 – 8. [27] Yagel S, Geva TE, Solomon H, Shimonovitz S, Reich R, Finci-Yeheskel Z, et al. High levels of human chorionic gonadotropinretard first trimester trophoblast invasion in vitro by decreasing urokinase plasminogen activator and collagenase activities. J Clin Endocrinol Metab 1993;77:1506 – 11.

244

M. Ueyama et al. / Thrombosis Research 108 (2003) 235–244

[28] Monzon-Bordonaba F, Wang C-L, Feinberg RF. Fibronectinase activity in cultured human trophoblast is mediated by urokinase-type plasminogen activator. Am J Obstet Gynecol 1997;176:58 – 65. [29] Graham CH, Lala PK. Mechanism of control of trophoblast invasion in situ. Cell Physiol 1991;148:228 – 34. [30] Milwidsky A, Finci-Yeheskel Z, Yagel S, Mayer M. Gonadotropinmediated inhibition of proteolytic enzymes produced by human trophoblast in culture. J Clin Endocrinol Metab 1993;76:1101 – 5. [31] Astedt B, Hagerstrand I, Lacander I. Cellular localization in placenta of placental type plasminogen activator inhibitor. Thromb Haemost 1986;56:63 – 5. [32] Wilson EA, Jawad MJ, Dickson LR. Suppression of human chorionic gonadotropin by progestational steroids. Am J Obstet Gynecol 1980; 138:708 – 13. [33] Moll UM, Lane BL. Proteolytic activity of first trimester human placentae: localization of interstitial collagenase in villous and extravillous trophoblast. Histochemistry 1990;94:555 – 61. [34] Feng Q, Liu K, Liu YX, Byme S, Ockleford CD. Plasminogen activators and inhibitors are transcribed during early macaque implantation. Placenta 2001;22:186 – 99. [35] Irvin JC, Kirk D, King RJB, Quigley MM, Gwatkin RBL. Hormonal regulation of human endometrial stromal cells in culture: an in vitro model for decidualization. Fertil Steril 1989;52:758 – 61. [36] Deng G, Curriden SA, Wang S, Rosenberg S, Loskutoff DJ. Is plasminogen activator inhibitor-1 the molecular switch that governs urokinase receptor-mediated cell adhesion and release? J Cell Biol 1996;134:1563 – 71. [37] Waltz DA, Natkin LR, Fujita RM, Wei Y, Chapman HA. Plasmin and plasminogen activator inhibitor type 1 promote cellular motility by regulating the interaction between the urokinase receptor and vitronectin. J Clin Invest 1997;100:58 – 67. [38] Jensen PH, Lorand L, Jensen PH, Lorand L. Type-2 plasminogen-

[39] [40]

[41]

[42]

[43] [44]

[45]

[46]

[47]

activator inhibitor is a substrate for trophoblast transglutaminase and factor XIIIa. Transglutaminase-catalyzed cross-linking to cellular and extracellular structures. Eur J Biochem 1993;214:141 – 6. Preissner KT, Jenne D. Structure of vitronectin and its biological role in hemostasis. Thromb Haemost 1991;66:123 – 32. Lockwood CJ, Nemerson Y, Guller S, Krikun G, Alvarez M, Hausknecht V, et al. Progestational regulation of human endometrial stromal cell tissue factor expression during decidualization. J Clin Endocrinol Matab 1993;76:231 – 6. Walsh SW, Coulter S. Increased placental progesterone may cause decreased placental prostacyclin production in preeclampsia. Am J Obstet Gynecol 1989;161:1586 – 92. Graham CH, Mc Crae KR. Altered expression of gelatinase and surface-associated plasminogen activator activity by trophoblast cells isolated from placentas of preeclamptic patients. Am J Obstet Gynecol 1996;175:555 – 62. Billington WD. Biology of the trophoblast. Adv Reprod Physiol 1971;5:39 – 42. Jenkins DM, O’Neill M, Mattar M, France VM, Hsi BL, Faulk WP. Degenerative changes and detection of plasminogen in fetal membranes that rematurely. Br J Obstet Gynaecol 1983;90:841 – 6. Liu YX, Hu ZY, Liu K, Byrne S, Zou RJ, Ny T, et al. Localization and distribution of tissue type and urokinase type plasminogen activators and their inhibitors Type1 and 2 in human and rhesus monkey fetal membranes. Placenta 1998;19:171 – 80. Graham CH. Effect of transforming growth factor-beta on the plasminogen activator system in cultured fiest trimester human cytotrophoblasts. Placenta 1997;18:137 – 43. Smith AN, Carter QL, Kniss DA, Brown TL. Characterizaion of a TGFbeta-responsive human trophoblast-derived cell line. Placenta 2001;22:425 – 43.