Increased Expression of Prostasin Contributes to Early-Onset Severe Preeclampsia Through Inhibiting Trophoblast Invasion

ORIGINAL ARTICLE Increased expression of prostasin contributes to early-onset severe preeclampsia through inhibiting trophoblast invasion

Y Yang1,3, J Zhang1,3, Y Gong1, X Liu1, Y Bai1,WXu1,2 and R Zhou1

OBJECTIVE: To investigate the potential role of prostasin, as an invasion suppressor, in the process of trophoblast invasion in preeclampsia. STUDY DESIGN: This case-control study included 19 early-onset severe preeclampsia (⩽34 weeks), 20 late-onset severe preeclampsia (434 weeks) and 20 normal term pregnant women. Immunohistochemistry was conducted to identify the cellular localization of prostasin, as well as the matrix metalloproteinase 2 (MMP2) and MMP9 in the placenta tissues. Enzyme-linked immunosorbent assay was performed to analyze the expression of these three proteins in placental homogenates. The effect of prostasin on the invasive and migratory ability of trophoblast cells was detected by transwell assays. We also examined the regulation of the prostasin antibody in the MMP2 and MMP9 secretion by HTR-8/SVneo cells via blocking the prostasin activity. RESULT: This study demonstrated that the prostasin, MMP2 and MMP9 were all expressed in the placental syncytiotrophoblasts. Increased expression of prostasin was detected in cases with early-onset severe preeclampsia compared with the late-onset and control groups (Po0.05), whereas the expression patterns of MMP2 and MMP9 in placental homogenates were opposite to that of prostasin (Po0.05). Recombinant prostasin inhibited the invasion and migration of trophoblast cells, whereas prostasin antibody enhanced the MMP2 and MMP9 secretion in a dose- and time-dependent manner. CONCLUSION: These ﬁndings suggest that prostasin may suppress the invasion process in preeclampsia by attenuating MMP2 and MMP9 secretion. Journal of Perinatology (2015) 35, 16–22; doi:10.1038/jp.2014.136; published online 31 July 2014

INTRODUCTION ‘gelatinase A and B’, are rate-limiting enzymes in invasion pro- Preeclampsia, especially early-onset severe preeclampsia, con- cesses. It has been demonstrated that patients with preeclampsia 13,14 tinues to be a leading cause of maternal and perinatal morbidity exhibit high levels of placental MMP2 and MMP9, whereas and mortality.1,2 Although the exact etiology remains elusive, it is conflicting results were also observed, highlighting the signifi- 15,16 commonly believed that inadequate trophoblast invasion of the cance in invasion-regulatory function. When treated with maternal spiral arteries represents one characteristic pathologic functional antibody against prostasin, human choriocarcinoma factor.3 In the development of placenta, reduced invasion ability JEG-3 cells exhibit an increase in the production of MMP2 and 17 results in the trophoblasts failing to invade the myometrium MMP26, as well as enhanced invasion ability. This finding raises deeply and to remodel the uterine spiral arteries appropriately.4 the possibility that prostasin may participate in the regulation of Consequently, the transformation of spiral arteries is suppressed trophoblast invasion in preeclampsia via the changes of MMP2 from high-resistance, low-flow vessels into dilated vessels.5 The and MMP9 expression. However, the role of prostasin as mediators decreased blood flow and fetoplacental perfusion induce placen- of trophoblast invasion in preeclampsia is still unclear. tal ischemia, subsequent system endothelial dysfunction and Given the presence of impaired trophoblast invasion in preeclampsia.6,7 Furthermore, studies have indicated that the preeclampsia, we hypothesized that abnormal high expression interaction of trophoblast with the extracellular matrix is one of of prostasin in preeclamptic placental tissues may impede tro- the most important factors in trophoblastic invasion, through phoblast invasion through altering MMP2 and MMP9 secretion. To providing a substrate for attachment, growth and migration.4 test this hypothesis, we investigated the localization and expres- However, the exact mechanism that underlies incomplete tro- sion of prostasin, MMP2 and MMP9 in preeclamptic placental phoblastic invasion in preeclampsia remains to be elucidated. tissues, and determined the effects of recombinant human Prostasin is a glycosylphosphatidyl inositol-anchored active prostasin on the invasion and migration of HTR-8/SVneo cells, a serine protease, displaying trypsin-like enzymatic activities.8 It is human trophoblast-derived cell line. widely distributed in the prostate, ovaries, mammary glands, lungs, placenta and other tissues.9 Prostasin has been identified as an invasion suppressor for human prostate and breast cancers.10,11 METHODS In rhesus monkeys, prostasin was also shown to be involved in Study subjects placental trophoblastic invasion and endometrial tissue remodel- The study included 19 early-onset severe preeclampsia (⩽34 weeks, early- ing.12 Matrix metalloproteinase 2 (MMP2) and MMP9, also called onset group), 20 late-onset severe preeclampsia (434 weeks, late-onset

1Department of Obstetrics and Gynecology, West China Second University Hospital, Sichuan University, Chengdu, China and 2Sichuan University-The Chinese University of Hong Kong Joint Laboratory for Reproductive Medicine, West China Institute of Women and Children’s Health, West China Second University Hospital, Sichuan University, Chengdu, China. Correspondence: Dr R Zhou, Department of Obstetrics and Gynecology, West China Second University Hospital, Sichuan University, Chengdu 610041, China. E-mail: [email protected] 3These authors contributed equally to this work. Received 16 March 2014; revised 13 June 2014; accepted 23 June 2014; published online 31 July 2014 Prostasin inhibits trophoblast invasion in preeclampsia Y Yang et al 17 group) and 20 normal term pregnant women (as the control group). All the The concentrations of MMP2 and MMP9 in supernatants were determined participants were enrolled at the time of diagnosis from the Department of using ELISA kits purchased form R&D Systems (Minneapolis, MN, USA) Obstetrics & Gynecology, West China Second University Hospital, Sichuan University during the period from August 2011 to October 2012, and they were all Han population living in Sichuan Province, southwest China. In the IHC early-onset group, there were all preterm deliveries; in the late-onset IHC was conducted with a biotin-streptavidin-peroxidase kit (Gene Tech, group, there were 10 term deliveries and 10 preterm deliveries; and in the Shanghai, China) according to the manufacturer’s instructions. The sections normal group, there were all term deliveries. Exclusion criteria include of placental samples were deparaffinized completely, rehydrated in vaginal delivery (avoiding potential influence on protein levels), multiple decreasing series of alcohol and subjected to antigen retrieval in citrate pregnancy, diabetes, heart diseases, chronic hypertension, fetal malforma- buffer (10 mM citrate sodium, 10 mM citric acid, pH 6.0) in a microwave tion, chronic nephritis and HELLP syndrome (hemolysis, elevated liver oven at 92 to 98 °C for 15 min followed by washing with 1 × phosphate- enzymes and low platelet count). The study was approved by the ethics buffered saline three times at room temperature for 5 min each. Sections committee of the West China Second University Hospital, Sichuan were then incubated with rabbit anti-human prostasin polyclonal antibody University. All participants gave written informed consent form at entry (1:50), mouse anti-human MMP2 monoclonal antibody (Cat. Number: of the study. Preeclampsia was defined as blood pressure4140/90 mm Hg ab86607, 1: 150, Abcam) and mouse anti-human MMP9 monoclonal on two separate occasions 6 h apart or a single recording of a diastolic antibody (Cat. Number: ab119906, 1:400, Abcam) at 4 °C overnight. The pressure of ⩾ 110 mm Hg, in association with proteinuria ⩾ 1+ on dipstick tissue sections were further incubated with biotinylated antibodies and testing or proteinuria ⩾ 300 mg per 24 h after 20 weeks’ gestation.18 horseradish peroxidase-labeled streptavidin. The substrate was stained Patients with severe preeclampsia were further classified as either having with diaminobenzidine, and counterstained with hematoxylin. early-onset (⩽34 weeks) or late-onset (4 34 weeks) disease according to 19 the gestational age at which onset preeclampsia was diagnosed. ELISA Gestational age was based on the last menstrual period and/or was confirmed by ultrasound examination conducted at the early stage of The placental homogenates of early-onset, late-onset and control groups, pregnancy. Blood pressure was measured on the right arm with a mercury and the cell culture supernatants collected at different end points (24, 48 sphygmomanometer with the subjects seated and after a 5-min rest. Three and 72 h) were analyzed by ELISA with commercially available ELISA kits. times of measurement were conducted at 1-min intervals, and the values (Uscn Life Science (Wuhan, China) for placental homogenates; R&D were averaged. Systolic blood pressure (SBP) and diastolic blood pressure Systems for cell supernatants). The concentrations of prostasin, MMP2 were recorded as phase I and V Korotkoff sounds. All the patients were and MMP9 in the culture supernatants were determined by comparison delivered by undergoing elective cesarean section, with the indications of with the standard curve using a spectrophotometer, which was set at a previous cesarean section, breech presentation and social indications. wavelength of 450 nm (TECAN, Männedorf, Switzerland).

Specimen collection Transwell invasion assay Samples of venous blood were collected from each subject in sterile tubes Cellular invasion assay was performed by determining the ability of cells to μ with ethylene diamine tetraacetic acid anticoagulants. Serum was invade the 8- m pores of polycarbonate membranes (Cat. Number: 3422, extracted by centrifugation. Serum was extracted by centrifugation for Corning-Costar, Cambridge, UK), which were placed at the bottom of measurement of albumin, creatinine, urea nitrogen and lactate dehydro- transwell chambers. The membranes were pre-coated with 30 μl Matrigel − 1 genase. Samples were stored at − 20 °C until batch was analyzed. All the (1 mg ml ; Cat. Number: 356237, BD Biosciences, Bedford, MA, USA). 4 values were determined by the clinical laboratory of West China Second HTR-8/SVneo cells (2 × 10 ) in 200 μl serum-free medium, in the presence University Hospital. The clinical laboratory was College of American of recombinant human prostasin or not (Cat.Number: 4599-SE-010, R&D Pathologists certified for all the items used in this project. Systems), were plated on the top chamber and incubated with 600 μl Tissue of human placenta was collected from patients in each group as culture medium containing 20% fetal bovine serum in the bottom described previously.20 Approximately 1.0 cm3 biopsy tissues in the center chamber. After incubation at 37 °C for 48 h, the filters were scrubbed with of the placenta (avoiding the vessels and/or calcium deposits) was cotton swabs in the upper side of the membrane to remove the non- obtained from both the fetal and the maternal side after delivery as invading cells, subsequently fixed in 4% paraformaldehyde and stained immediately as possible. The tissue specimens were washed three times with 0.2% crystal violet. The invading cells on the underside of the in normal saline, and prepared for the placental homogenization and membrane were enumerated by using an inverted microscope in 10 immunohistochemistry (IHC) assays. random fields at a magnification of × 200 (IX-70; Olympus, Melville, NY, USA). Experiments were performed in triplicate and the results are represented as means of invasion percentage (%) ± s.d. compared with control. Preparation of placental homogenates Five hundred micrograms of the tissue samples were transferred to a 5-ml polypropylene tube containing 500 μl ice-cold homogenizing buffer with Transwell migration assay protease inhibitor added (250 mM sucrose, 0.7 μM pepstatin A, 1.6 μM The migration ability of HTR-8/SVneo cells was detected by their ability to μ antipain, 80 μM aprotinin, 1 mM EDTA, 10 mM HEPES-Tris (pH 7.4)). The cross the 8- m pores of migration chambers. The method used in the cell placental samples were then homogenized with a Powergen homogenizer migration assay was similar to the Matrigel invasion assay, except that the (Fisher Scientific, Pittsburg, PA, USA) for 30 s on ice. Then the homogenates inserts were not coated with Matrigel. were centrifuged at 5000g for 20 min at 4 °C, and the collected − supernatants were stored at 80 °C until enzyme-linked immunosorbent Statistical analysis assay (ELISA). Data were processed by using the SPSS 16.0 software package (IBM, Armonk, NY, USA). Continuous variables were recorded as mean ± s.d. Cell culture and treatment Clinical and demographic characteristics of women were compared by The human first-trimester extravillous trophoblast cell line HTR-8/SVneo one-way analysis of variance or the Wilcoxon Test. The correlation analyses cells, which were established by Graham et al.21 and characterized with were evaluated by Pearson and Spearman’s tests of correlation with high capacity of migration and invasion (kindly provided by Dr Yali Hu, Bonferroni correction. Po0.05 (bilateral) was regarded as statistically Affiliated Drum Tower Hospital, Medical College of Nanjing University, significant. Nanjing, China), were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum in 5% CO2 at 37 °C. Equal numbers of HTR-8/SVneo cells were seeded in 12-well plates RESULTS 5 (1 × 10 cells per well). The cells were then treated with the prostasin Demographic characteristics of the subjects antibody (Cat. Number: 15527-1-AP, Proteintech, Chicago, IL, USA), as well as the non-immune rabbit immunoglobulin G (Cat. Number: ab125938, Demographic characteristics of the subjects are summarized in Abcam, Cambridge, UK) used as the control, at 0, 3.5, 7 and 70 μgml− 1 for Table 1. As shown, risk factors such as the pregnancy body mass 24, 48 and 72 h. At each time point, cell-free supernatants were collected index, gestational age, SBP, diastolic blood pressure, serum by centrifugation at 1000 g for 20 min and stored at − 80 °C until analysis. albumin, serum creatinine, serum urea nitrogen, lactate

Table 1. Demographic characteristics of the subjects

Characteristic Early-onset group Late-onset group Controls

n 19 20 20 Age (years) 32.26 ± 5.16 31.15 ± 4.80 30.95 ± 3.98 BMI before pregnancy (Kg m −2) 24.53 ± 6.59a,b 21.81 ± 1.79 20.80 ± 2.90 Gestational age (days) 232.11 ± 12.19a,b 259.75 ± 2.02a 272.20 ± 5.29 SBP (mm Hg) 176.74 ± 25.97a,b 154.10 ± 10.81a 114.20 ± 9.46 DBP (mm Hg) 109.53 ± 12.25a,b 100.70 ± 9.52a 70.70 ± 8.33 Serum albumin (g l − 1) 27.66 ± 2.51a 29.26 ± 7.64a 37.26 ± 3.65 LDH (U l − 1) 517.58 ± 261.47a 459.55 ± 241.10a 202.65 ± 71.10 Serum Cr (μmol l − 1) 72.47 ± 18.24a 65.70 ± 40.66a 45.30 ± 4.87 Serum BU (mmol l − 1) 6.82 ± 2.89a,b 4.37 ± 3.42 3.22 ± 0.46 Birth weight (g) 1529.47 ± 364.32a,b 2758.75 ± 764.89b 3261.30 ± 423.12 Birth length (cm) 43.51 ± 1.97b 48.37 ± 1.96 49.35 ± 1.59 Urine protein (g per 24 h) 6.84 ± 4.30b 3.83 ± 2.42 Abbreviations: BMI, body mass index; DBP, diastolic blood pressure before delivery; LDH, lactate dehydrogenase; SBP, systolic blood pressure before delivery; serum BU, serum urea nitrogen; serum Cr, serum creatinine. Normal range: serum albumin (34.0 to 48.0) g l − 1, serum Cr (30.4 to 97.4) μmol l − 1, serum BU (3.2 to 8.2) mmol l − 1, LDH (120 to 246) U l − 1. Data are presented as mean ± s.d. One-way analysis of variance was used to examine differences in continuous variables except LDH. The Wilcoxon test was used to examine differences in LDH with Bonferroni correction. aPo0.01, when compared with controls. bPo0.01, when compared with the late-onset group.

dehydrogenase and birth weight in early-onset group were the comparisons between MMP2, as well as MMP9, and other significantly different from the controls; the body mass index, variables as required by the Bonferroni correction. After correction, gestational age, SBP, diastolic blood pressure, serum creatinine, the level of MMP9 in the early-onset group was negatively asso- birth weight, birth length and 24-h urine protein in the early-onset ciated with SBP before delivery (r = − 0.734, P = 0.000). However, group were also significantly different from the late-onset group there were no significant correlations between clinical parameters (all P-valueso0.05). and the levels of prostasin, MMP2 and MMP9 in the late-onset and control groups (Table 3). Localization of prostasin, MMP2 and MMP9 in placenta The localization of prostasin, MMP2 and MMP9 in placenta was Prostasin inhibits the invasion and migration of extravillous detected by IHC. The immunoreactivity of prostasin, MMP2 and trophoblast cell line HTR-8/SVneo MMP9 was observed in the cytoplasm of placental syncytiotro- In our preliminary experiments, we found that the basal prostasin phoblast in the control, early-onset and late-onset groups. concentrations secreted by HTR-8/SVneo was 75 ng ml − 1 deter- Positive stainings for these three proteins were also observed mined by ELISA (data not shown). Hence, for the transwell in some decidual cells and villous mesenchymal cells, but invasion and migration assays, prostasin was used at the level of were relatively weaker than those in trophoblasts. There were − 750 ng ml 1. no significant differences in the expression of these three proteins As shown in Figure 2, when the prostasin was added to the between the maternal and fetal compartments (data not shown) fi (Figure 1). upper chamber, HTR-8/SVneo cells displayed signi cantly decreased invasion into Matrigel compared with that observed in the blank control group. Similarly, the migration ability of Expression of prostasin, MMP2 and MMP9 in placental tissue HTR-8/SVneo cells was also significantly inhibited in the presence homogenates of prostasin, compared with that found in the absence of Table 2 depicts the levels of prostasin, MMP2 and MMP9 in prostasin. placental homogenates from the early-onset, late-onset and control groups. The placental prostasin levels were significantly higher in the early-onset group than those in the late-onset and Secretion of MMP2 and MMP9 from HTR-8/SVneo cells is control groups (Po0.05). There were also significant differences in enhanced by prostasin antibody the levels of MMP2 in placentas from the control group compared To understand the potential mechanisms by which prostasin with the early-onset group (Po0.05). The MMP9 levels in the inhibits the HTR-8/SVneo cells invasion and migration, we examined early-onset group were significantly higher than those in the late- the effect of prostasin antibody on the secretion of MMP2 onset and control groups (Po0.05). Moreover, the levels of and MMP9. prostasin and MMP9 in the late-onset group were significantly As measured by ELISA, the levels of MMP2 and MMP9 in the increased when compared with controls, respectively. However, culture supernatants from HTR-8/SVneo cells at basal state there were no statistically significant differences in the MMP2 increased in a time-dependent manner. The basal MMP2 and levels between the late-onset and control groups (Table 2). MMP9 concentrations were increased from 45.17 to 75.80 pg ml − 1 and from 117.86 to 247.88 pg ml − 1, respectively. Correlations between the placental levels of prostasin, MMP2 and The data also indicated that treatment with the prostasin MMP9 and clinical parameters in each group antibody enhanced the MMP2 and MMP9 secretion in a An alpha level of 0.004 was used for the comparisons between concentration–time-dependent manner at the higher concentra- − 1 − 1 prostasin and other variables as required by the Bonferroni tions (7, and 70 μgml , but not 3.5 μgml ; Po0.05). However, correction. After Bonferroni correction, the level of prostasin in the there were no significant differences in the MMP2 and early-onset group was positively correlated with SBP before MMP9 secretion when the cells were treated with non-immune delivery (r = 0.891, P = 0.000); an alpha level of 0.005 was used for rabbit immunoglobulin G (Table 4).

Figure 1. Expression of prostasin, MMP2 and MMP9 in human placenta. Immunoreactivity for prostasin, MMP2 and MMP9 in placental tissues from maternal surface (a–c) for the early-onset group; (d–f) for the late-onset group; (g–i) for the control group; prostasin, MMP2 and MMP9 are mainly localized in syncytiotrophoblasts (×200). Prostate cancer, breast cancer and gastric cancer tissue samples were used as positive controls for prostasin, MMP2 and MMP9, and the results showed strong expression (j–l; × 200); negative controls with non-immune rabbit or mouse immunoglobulin G produced no immunostainings (m and n, respectively; × 200). STB, syncytiotrophoblast; MMP2, matrix metalloproteinase 2. Bar, 100 μm.

Table 2. Comparisons of prostasin, MMP2 and MMP9 levels of the placental homogenates among the three groups

Group n Prostasin (ng ml − 1) MMP2 (pg ml − 1) MMP9 (ng ml − 1)

Early-onset group 19 1466.23 ± 327.44a 326.42 ± 198.35a 16.63 ± 5.44a Late-onset group 20 1236.83 ± 121.53b 368.25 ± 204.61 22.38 ± 3.39b Controls 20 1208.63 ± 188.86 488.89 ± 291.68 21.15 ± 7.44 Abbreviation: MMP, matrix metalloproteinase. Data are presented as mean ± s.d. One-way analysis of variance was used to examine differences in continuous variables. aPo0.05, when compared with controls. bPo0.05, when compared with the early-onset group.

DISCUSSION early-onset and late-onset severe preeclampsia, which were 22 In the present study, IHC analysis of human placental partly in agreement with a previous study. Our study also tissue demonstrated that prostasin, MMP2 and MMP9 were indicated that in the early-onset group, placental prostasin expressed in the syneytiotrophoblasts in normal pregnancy, levels were signiﬁcantly higher, whereas the MMP2 and

© 2015 Nature America, Inc. Journal of Perinatology (2015), 16 – 22 Prostasin inhibits trophoblast invasion in preeclampsia Y Yang et al 20 MMP9 levels were signiﬁcantly lower compared with the normal group. Prostasin is a serine protease with a molecular weight of 40 kDa. In humans, this protein is encoded by a single-copy gene that is 8

—— located on chromosome 16p11.2. Prostasin expression has been 0.159 0.503 0.070 0.768 0.176 0.459 0.203 0.391 0.266 0.257 0.168 0.480 0.306 0.190 − − − − − − − detected in villi, trophoblastic columns and trophoblastic shells of rhesus placentas, suggesting that prostasin may have an important role in endometrial epithelial morphology establishment, tissue remodeling and trophoblastic invasion during early pregnancy (at 12, 18 and 26 days).12 Moreover, prostasin expression has been found in villous trophoblasts and cytotro- 0.006 0.979 0.168 0.478 0.225 0.339 0.023 0.923 0.267 0.255 0.051 0.108 0.027 0.912 0.258 0.272 phoblastic columns in human placenta.17 High prostasin expres- − − − − − sion was observed in villous cytotrophoblasts, syncytio- a trophoblasts and cytotrophoblastic columns between 6 and 9 weeks of gestation.17 To our knowledge, this study is the first report showing abnormal expression of placental prostasin in preeclampsia. Our results showed that treatment with recombinant human 0.593 0.017 0.557 0.137 0.097 0.720 0.462 0.139 0.116 0.625 0.311 0.187 0.068 0.774 0.734 0.000 0.395 0.079 0.124 0.603 arisons between MMP2 and other variables and 0.005 for − − − − − − − prostasin significantly decreased the invasion and migration ability of HTR-8/SVneo cells compared with the controls. As MMP2 and MMP9 have been identified as the rate-limiting

; MMP, matrix metalloproteinase; SBP, systolic blood pressure before enzymes for cell invasion, we speculated that prostasin may exert the inhibition function via regulation of the MMP2 and MMP9 secretion. Therefore, we investigated the secretion of —— 0.068 0.777 0.057 0.810 MMP2 and MMP9 by attenuating the activity of prostasin with a − − specific prostasin antibody. The current study revealed that HTR-8/ SVneo cells secreted MMP2 and MMP9, and the secretion of MMP2 and MMP9 increased over time. Compared with the control group, the MMP2 and MMP9 levels increased in a dose- and time- dependent manner in the culture supernatants from HTR-8/SVneo 0.063 0.793 0.186 0.490 0.216 0.360 0.032 0.480 0.125 0.608 0.260 0.269 0.052 0.829 0.157 0.405 0.051 0.831 0.334 0.162 0.160 0.502 0.435 0.055 0.124 0.602 0.197 0.406 0.275 0.240 0.090 0.707 0.223 0.349 0.205 0.387 cells when treated with different concentrations of prostasin − − − − − − − − antibody. These results, to some extent, support the possibility that the addition of prostasin antibody to the culture medium in different concentrations partly blocked the complex formation of prostasin and the prostasin-binding protein, protease nexin-1.23 As a result, the serine protease activation of prostasin was attenuated,23 in company with a significant increase in MMP2 and 0.428 0.055 0.858 0.136 0.049 0.836 0.129 0.669 0.412 0.079 0.310 0.183 0.538 0.012 − − − − − − MMP9 secretion. Our results also showed that prostasin and MMP9 expression in human placental homogenates was associated with SBP before delivery, suggesting that these proteins may be coordinately regulated in the pathogenesis of preeclampsia. Because shallow trophoblastic invasion contributes to preeclampsia, studies have also examined the levels of MMP2 and — 0.168 0.478 0.219 0.352 0.553 0.007 0.134 0.572 0.278 0.219 MMP9 in preeclampsia. Patients with preeclampsia exhibit − − − significantly higher serum levels of MMP2, as well as placental expression in some studies,13,24–26 whereas another study did not find significant differences in the placental MMP2 expression and

cance after Bonferroni correction. 16

fi activity between preeclampsia and normal pregnancies. There were also conflicting studies showing significantly higher or lower Signi Prostasin MMP2 MMP9 a 0.222 0.344 0.135 0.569 0.115 0.669 0.234 0.320 0.156 0.508 0.204 0.386 0.081 0.733 0.265 0.257 0.348 0.131 0.549 0.012 0.110 0.644 MMP9 expression in the placental tissue from preeclampsia – − − − − − − compared with the normal controls.14 16,27 Thus, the expression

a of MMP2 and MMP9 in preeclampsia merits further investigation. In contrast with previously reported results,13,24,25 our study showed a consistent decrease of MMP2 and MMP9 in human

ELCELCELC preeclamptic placental tissue. Although we have no precise explanation for these conﬂicting results, differences in the 0.247 0.308 0.561 0.012 0.085 0.720 0.316 0.174 0.132 0.589 0.178 0.451 0.144 0.543 0.036 0.897 methods (ELISA vs zymography or western blot analysis) per- − − − formed to measure the levels of MMP2 and MMP9 in different samples (placenta vs plasma or cell culture models) may inﬂuence ) 0.329 0.169 0.032 0.890 0.111 0.639 ) 0.390 0.099 1 the results. In addition, it is possible that the variation in 1 − ) 0.329 0.169 − 1 gestational ages between studies may have an effect on the − ) 0.025 0.920 ) levels of MMP2 and MMP9. 1 1 − −

mol l Inadequate trophoblastic invasion leads to miscarriage or μ )

1 pregnancy-induced hypertension, whereas excessive invasion Correlations between clinical parameters and placental levels of prostasin, MMP2 and MMP9 among the three groups − leads to trophoblastic disease or placenta accrete.4,28 Interestingly, our study showed that the placental levels of prostasin increased, whereas MMP2 and MMP9 levels decreased signiﬁcantly in the MMP2(pg ml Urine protein (g per 24 h) 0.586 0.008 0.213 0.367 Serum Cr ( Serum BU (mmol l ParametersSBP (mm Hg) R 0.891 0.000 P R P R P R P R P R P R P R P R P Birth length (cm) 0.957 0.052 Birth Weight (g) 0.160 0.513 LDH(U l MMP9(ng ml Serum albumin (g l Gestational age DBP (mm Hg) 0.597 0.007 0.200 0.396 0.415 0.068 0.011 0.876 0.175 0.460 0.459 0.062 delivery; serum BU, serumAn urea alpha level nitrogen; of serum 0.004 Cr,the was used serum comparisons for creatinine. between the comparisons MMP9 between and prostasin other and other variables. variables as required by the Bonferroni correction, 0.005 for the comp Table 3. Abbreviations: C, controls; DBP, diastolic blood pressure before delivery; E, early-onset group; L, late-onset group; LDH, lactate dehydrogenase third trimester of pregnancy in patients with early-onset

Figure 2. Prostasin inhibits invasion and migration of the HTR-8/SVneo cells. (a) Invasion and migration analysis of HTR-8/SVneo cells. HTR-8/ SVneo cells were plated on a Boyden chamber coated with Matrigel and incubated for 48 h (A and C); migration analysis of HTR-8/SVneo cells (B and D). HTR-8/SVneo cells were plated on a chamber that was not coated with Matrigel and incubated for 48 h. Equal numbers (2 × 104)of HTR-8/SVneo cells were plated on the top chamber with serum-free medium in the presence of 750 ng ml − 1 prostasin (C and D) or not (A and B). The invading cells on the underside of the membrane were enumerated by using an inverted microscope in 10 random ﬁelds. Experiments were performed in triplicate and the results are represented as means of invasion or migration percentage (%) ± s.d. compared with control (b and c). Representative photographs are shown at a magniﬁcation of × 200. Bar, 100 μm.

Table 4. Expression of MMP2 and MMP9 in the supernatants of HTR-8/Svneo cells treated with prostasin antibody

MMP2 (pg ml − 1) MMP9 (pg ml − 1)

24 h 48 h 72 h 24 h 48 h 72 h

TG (0 μgml− 1) 45.17 ± 3.73 70.42 ± 4.95 75.80 ± 8.33 117.86 ± 8.15 224.11 ± 3.4 247.88 ± 10.9 CG (70 μgml− 1) 30.54 ± 3.52 78.09 ± 1.77 69.34 ± 7.37 161.94 ± 1.82 171.30 ± 11.90 210.08 ± 6.58 TG (70 μgml− 1) 59.34 ± 3.63a 142.42 ± 3.07a 179.01 ± 3.37a 287.86 ± 9.09a 338.94 ± 7.99a 786.87 ± 7.87a CG (7 μgml− 1) 41.07 ± 5.22 77.62 ± 2.69 70.12 ± 4.24 119.96 ± 13.06 170.48 ± 3.79 227.58 ± 9.16 TG (7 μgml− 1) 65.13 ± 4.00a 132.65 ± 3.69a 159.54 ± 2.01a 252.24 ± 10.05a 340.61 ± 7.29a 444.45 ± 10.88a CG (3.5 μgml− 1) 43.89 ± 4.42 60.48 ± 3.37 76.95 ± 2.68 133.67 ± 10.85 142.47 ± 5.94 225.10 ± 7.85 TG (3.5 μgml− 1) 41.00 ± 3.81 54.24 ± 3.06 65.65 ± 3.24 132.13 ± 8.29 169.90 ± 11.29 226.21 ± 8.74 Abbreviations: CG, control group; MMP, matrix metalloproteinase; TG, treatment group. Data are presented as mean ± s.d. aPo0.05 when compared between the treatment group and control group at the same time corresponding to the same concentration, respectively. preeclampsia. Although it has been shown that MMPs are in early pregnancy. Therefore, the placental levels of prostasin, involved in the remodeling of uterine spiral arteries, and reduced MMP2 and MMP9 in the third trimester as shown in the present MMP activity may hamper trophoblastic invasion, which subse- study might be able to reﬂect the levels and invasion/migration quently induces abnormal placental development and the onset ability in early pregnancy to some extent. of preeclampsia during early pregnancy,29,30 Sawicki G et al.22 Our previous studies have demonstrated that the serum levels demonstrated that during the third trimester of gestation, of prostasin in severe preeclampsia were signiﬁcantly higher than placental villus and umbilical vasculature increase with respond those in normal pregnancies and mild cases (unpublished data; all to the accelerating fetal growth. The growth of villous requires a Po0.05). The serum prostasin levels were also statistically lager increase in the face area of the syncytiotrophoblast and correlated with some clinical parameters of severe preeclampsia, extensive remodeling of the underlying villous tissue. This process such as blood pressure and proteinuria (unpublished data). The is similar to the remodeling of uterine spiral arteries, which occurs current study further suggests that prostasin may inhibit the

© 2015 Nature America, Inc. Journal of Perinatology (2015), 16 – 22 Prostasin inhibits trophoblast invasion in preeclampsia Y Yang et al 22 invasion and migration ability of trophoblasts in preeclampsia by and placenta during early pregnancy. J Histochem Cytochem 2006; 54(10): decreasing the secretion of MMP2 and MMP9. Further studies 1139–1147. investigating the precise underlying mechanism by which 13 Shokry M, Omran OM, Hassan HI, Elsedfy GO, Hussein MR. Expression of matrix metalloproteinases 2 and 9 in human trophoblasts of normal and preeclamptic prostasin regulates the levels of MMP2 and MMP9 will lead to fi 87 – a better understanding of prostasin in the pathogenesis of placentas: preliminary ndings. Exp Mol Pathol 2009; (3): 219 225. 14 Wang Z, Lu S, Liu C, Zhao B, Pei K, Tian L et al. Expressional and epigenetic preeclampsia. alterations of placental matrix metalloproteinase 9 in preeclampsia. Gynecol Endocrinol 2010; 26(2): 96–102. 15 Kolben M, Lopens A, Blaser J, Ulm K, Schmitt M, Schneider KT et al. Proteases and CONFLICT OF INTEREST their inhibitors are indicative in gestational disease. Eur J Obstet Gynecol Reprod The authors declare no conflict of interest. Biol 1996; 68(1-2): 59–65. 16 Huisman MA, Timmer A, Zeinstra M, Serlier EK, Hanemaaijer R, Goor H et al. Matrix-metalloproteinase activity in first trimester placental bed biopsies in fur- ACKNOWLEDGEMENTS ther complicated and uncomplicated pregnancies. Placenta 2004; 25(4): 253–258. This study was supported by the National Science Foundation of China (Grant No. 17 Ma XJ, Fu YY, Li YX, Chen LM, Chai K, Wang YL. Prostasin inhibits cell invasion in human choriocarcinomal JEG-3 cells. Histochem Cell Biol 2009; 132(6): 639–646. 81300512), Doctoral Fund of Ministry of Education of China (Grant No. 18 ACOG Committee on Practice Bulletins—Obstetrics. ACOG practice bulletin. 20110181110010), Sichuan Provincial Science and technology support project (Grant Diagnosis and management of preeclampsia and eclampsia Number 33, No. 2013SZ0074) and the Office of Science & Technology of Chengdu (Grant No. January 2002. Obstet Gynecol 2002; 99(1): 159–167. 12PPYB097SF-002). 19 von Dadelszen P, Magee LA, Roberts JM. Subclassification of preeclampsia. Hypertens Pregnancy 2003; 22(2): 143–148. REFERENCES 20 Tie W, Yu H, Chen J, Wang X, Chen W, Zhou R. Expressions of adiponectin receptors in placenta and their correlation with preeclampsia. Reprod Sci 2009; 16 1 Ilekis JV, Reddy UM, Roberts JM. Preeclampsia—a pressing problem: an executive (7): 676–684. summary of a National Institute of Child Health and Human Development 21 Graham CH, Hawley TS, Hawley RG, MacDougall JR, Kerbel RS, Khoo N et al. workshop. Reprod Sci 2007; 14(6): 508–523. Establishment and characterization of first trimester human trophoblast cells with 2 Sibai B, Dekker G, Kupferminc M. Pre-eclampsia. Lancet 2005; 365(9461): 785–799. extended lifespan. Exp Cell Res 1993; 206(2): 204–211. 3 Redman CW, Sargent IL. Latest advances in understanding preeclampsia. Science 22 Sawicki G, Radomski MW, Winkler-Lowen B, Krzymien A, Guilbert LJ. Polarized 2005; 308(5728): 1592–1594. release of matrix metalloproteinase-2 and -9 from cultured human placental 4 Goldman-Wohl D, Yagel S. Regulation of trophoblast invasion: from normal syncytiotrophoblasts. Biol Reprod 2000; 63(5): 1390–1395. implantation to pre-eclampsia. Mol Cell Endocrinol 2002; 187(1-2): 233–238. 23 Chen LM, Zhang X, Chai KX. Regulation of prostasin expression and function in 5 Burton GJ, Woods AW, Jauniaux E, Kingdom JC. Rheological and physiological the prostate. Prostate 2004; 59(1): 1–12. consequences of conversion of the maternal spiral arteries for uteroplacental 24 Montagnana M, Lippi G, Albiero A, Scevarolli S, Salvagno GL, Franchi M et al. blood flow during human pregnancy. Placenta 2009; 30(6): 473–482. Evaluation of metalloproteinases 2 and 9 and their inhibitors in physiologic and 6 Chen LM, Liu B, Zhao HB, Stone P, Chen Q, Chamley L. IL-6 TNFalpha and TGFbeta pre-eclamptic pregnancy. J Clin Lab Anal 2009; 23(2): 88–92. promote nonapoptotic trophoblast deportation and subsequently causes endo- 25 Myers JE, Merchant SJ, Macleod M, Mires GJ, Baker PN, Davidge ST. MMP-2 levels thelial cell activation. Placenta 2010; 31(1): 75–80. are elevated in the plasma of women who subsequently develop preeclampsia. 7 Roberts JM, Taylor RN, Musci TJ, Rodgers GM, Hubel CA, McLaughlin MK. Hypertens Pregnancy 2005; 24(2): 103–115. Preeclampsia: an endothelial cell disorder. Am J Obstet Gynecol 1989; 161(5): 26 Narumiya H, Zhang Y, Fernandez-Patron C, Guilbert LJ, Davidge ST. Matrix 1200–1204. metalloproteinase-2 is elevated in the plasma of women with preeclampsia. 8 Yu JX, Chao L, Chao J. Prostasin is a novel human serine proteinase from seminal Hypertens Pregnancy 2001; 20(2): 185–194. fluid. Purification, tissue distribution, and localization in prostate gland. J Biol 27 Lockwood CJ, Oner C, Uz YH, Kayisli UA, Huang SJ, Buchwalder LF et al. Matrix Chem 1994; 269(29): 18843–18848. metalloproteinase 9 (MMP9) expression in preeclamptic decidua and MMP9 9 Yu JX, Chao L, Chao J. Molecular cloning, tissue-specific expression, and cellular induction by tumor necrosis factor alpha and interleukin 1 beta in human first localization of human prostasin mRNA. J Biol Chem 1995; 270(22): 13483–13489. trimester decidual cells. Biol Reprod 2008; 78(6): 1064–1072. 10 Chen LM, Hodge GB, Guarda LA, Welch JL, Greenberg NM, Chai KX. Down- 28 Hannon T, Innes BA, Lash GE, Bulmer JN, Robson SC. Effects of local decidua on regulation of prostasin serine protease: a potential invasion suppressor in trophoblast invasion and spiral artery remodeling in focal placenta creta - an prostate cancer. Prostate 2001; 48(2): 93–103. immunohistochemical study. Placenta 2012; 33(12): 998–1004. 11 Chen LM, Chai KX. Prostasin serine protease inhibits breast cancer invasiveness 29 Cockle JV, Gopichandran N, Walker JJ, Levene MI, Orsi NM. Matrix metallopro- and is transcriptionally regulated by promoter DNA methylation. Int J Cancer teinases and their tissue inhibitors in preterm perinatal complications. Reprod Sci 2002; 97(3): 323–329. 2007; 14(7): 629–645. 12 Lin HY, Zhang H, Yang Q, Wang HX, Wang HM, Chai KX et al. Expression of 30 Roth I, Fisher SJ. IL-10 is an autocrine inhibitor of human placental cytotropho- prostasin and protease nexin-1 in rhesus monkey (Macaca mulatta) endometrium blast MMP-9 production and invasion. Dev Biol 1999; 205(1): 194–204.

Supplementary Information accompanies the paper on the Journal of Perinatology website (http://www.nature.com/jp)