Decidual Cells Block Inflammation-Mediated Inhibition of 15-Hydroxyprostaglandin Dehydrogenase in Trophoblasts

Abstract

Chorioamnionitis generates prostaglandin (PG) E₂ and F_2α, promoting fetal membrane rupture, cervical ripening, and uterine contractions. 15-Hydroxyprostaglandin dehydrogenase (HPGD) contributes to pregnancy maintenance by inactivating PGs. Herein, the role of decidual cells in the regulation of HPGD expression at the maternal-fetal interface was investigated. HPGD immunostaining was primarily detected in anchoring villi and choriodecidual extravillous trophoblasts (EVTs) during pregnancy. Chorionic EVTs adjacent to the decidua parietalis exhibited significantly higher HPGD levels than those adjacent to the amnion. HPGD histologic score levels were significantly lower in choriodecidua from chorioamnionitis versus gestational age-matched controls (means ± SEM, 132.6 ± 3.8 versus 31.2 ± 7.9; P < 0.05). Conditioned media supernatant (CMS) from in vitro decidualized term decidual cells (TDCs) up-regulated HPGD levels in differentiated EVTs, primary trophoblasts, and HTR8/SV^neo cells. However, CMS from 5 μg/mL lipopolysaccharide or 10 ng/mL IL-1β pretreated TDC cultures down-regulated HPGD levels in HTR8/SV^neo cultures. Similarly, direct treatment of HTR8/SV^neo with lipopolysaccharide or IL-1β significantly reduced HPGD levels versus control (P < 0.05) but not in TDC-CMS pretreated HTR8/SV^neo cultures. Collectively, these results uncover a novel decidual cell–mediated paracrine mechanism that stimulates levels of trophoblastic HPGD, whose function is to inactivate labor-inducing PGs, thereby promoting uterine quiescence during pregnancy. However, infectious/inflammatory stimuli in decidual cells cause a paracrine inhibition of trophoblastic HPGD expression, increasing PGE₂/PGF_2α levels, thereby contributing to preterm birth.

Graphical abstract

Preterm birth (PTB) is the leading contributor to perinatal mortality and morbidity worldwide.¹ The average frequency of PTB is approximately 10.4% in the United States, with the rate being 50% higher in African American women (US Centers for Disease Control and Prevention, https://www.cdc.gov/maternal-infant-health/preterm-birth/?CDC_AAref_ValZhttps://www.cdc.gov/reproductivehealth/maternalinfanthealth/pretermbirth.htm#, last accessed November 16, 2023).2, 3 PTB accounts for 18% of all deaths among children aged <5 years and as much as 35% of all deaths among newborns (aged <28 days).⁴ Chorioamnionitis (CAM) contributes to >50% of early PTB cases, and is characterized by infection-induced inflammation of the fetal membranes, chorion of the placenta, and/or decidua.5, 6, 7 PTB before 28 weeks’ gestation is linked to infections of the cervix, decidua, amniochorion, and amniotic cavity by vaginal microorganisms.⁸ At the maternal-fetal interface, bacteria or bacteria-derived products such as lipopolysaccharide (LPS) elicit a signaling cascade that disrupts maternal-immune tolerance and establishes an inflammatory response that impairs chorioamniotic membrane integrity and tensile strength.⁹ The amniotic fluid of patients with PTB and coexisting CAM contains elevated levels of inflammatory mediators such as IL-6, IL-8, colony-stimulating factors, and the potent proinflammatory cytokines IL-1β and tumor necrosis factor-α.¹⁰ Additionally, prostaglandin (PG) E₂ and F_2α are well-known mediators of term and preterm birth.¹¹ They are significantly elevated at the maternal-fetal interface during infection of the chorioamniotic membranes.¹² Moreover, CAM promotes fetal membrane rupture, cervical ripening, and uterine contractions.¹³

The enzyme 15-hydroxyprostaglandin dehydrogenase (HPGD) is primarily responsible for the metabolism of PGD₂, PGE₂, and PGF_2α through cytoplasmic oxidation into 15-keto-PGs, which are biologically inert.¹⁴ Production and metabolism of PGs is compartmentalized, with biosynthetic activity generated in the decidua and amnion, and metabolism occurring in the chorion.^15,16 Thus, the chorion is a bridge between the maternal and fetal tissues, and serves as a metabolic barrier to prevent transportation of PGs into the amnion, decidua, and myometrium, thereby contributing to uterine quiescence.¹⁷ The chorion plays a crucial role in inactivating PGs via HPGD.15, 16, 17 Furthermore, chorionic trophoblast is the primary cell type expressing HPGD, whose expression is reduced in placentas obtained from patients affected by PTB.¹⁸ Additionally, Hpgd knockout mice display pregnancy loss at embryonic day 8.5,¹⁹ whereas mice with hypomorphic expression of Hpgd deliver almost 24 hours early.²⁰ Kishore et al²¹ treated pregnant mice with a combination of PGE₂ and an HPGD inhibitor to induce early cervical ripening and PTB. These findings indicate an essential role for HPGD in pregnancy maintenance.

We hypothesized that decidual cell–derived paracrine factors induce expression of HPGD in chorionic extravillous trophoblasts (EVTs) by contributing to maintenance of uterine quiescence, whereas mediators of inflammation block this action to trigger PTB. Therefore, the aim of the study was to determine the primary cell types expressing HPGD at the maternal-fetal interface throughout pregnancy, to study decidual cell–mediated regulation of HPGD expression in different trophoblast types, and the involvement of HPGD levels in CAM as one of the pathways associated with PTB.

Materials and Methods

Placental Specimens

Previously deidentified banked paraffin human placental specimens from first-trimester pregnancies (at 7 to 8 weeks’ gestation; n = 2), mid pregnancies (at 20 weeks’ gestation; n = 2), and term pregnancies (n = 5) were obtained after approval by the Yale University Human Investigation Committee and the University of South Florida Institutional Review Board (number Pro00019472). Additional placental specimens were obtained from patients delivered preterm for CAM [n = 6; average gestational age (GA), 26 weeks] and GA-matched controls (n = 3; average GA, 27 weeks) from the Department of Clinical Pathology at Tampa General Hospital (Tampa, FL). The histologic diagnosis of CAM was based on the presence of a rich inflammatory infiltrate of neutrophils in the membranes, decidua, and cord in the setting of clinical infections. GA-matched controls were obtained from patients who had a medically indicated delivery for either maternal conditions, such as severe maternal cardiac or pulmonary disease; or fetal conditions requiring immediate delivery, unrelated to preterm labor or infection or rupture of membranes. Both GA-matched controls and patients with CAM were obtained from primary or repeat cesarean deliveries, who did not labor or have preterm premature rupture of membranes and did not receive corticosteroids within 7 days from delivery. For in vitro studies, placental specimens from early and term pregnancy were obtained from voluntary termination of uncomplicated pregnancies and from scheduled repeat cesarean deliveries, respectively. All procedures were performed following approval by the University of South Florida Institutional Review Board (19472).

Immunohistochemistry

Paraformaldehyde-fixed, paraffin-embedded placental sections (5 μm thick) were processed for immunohistochemistry, as previously described.^22,23 Briefly, the slides were deparaffinized and rehydrated, then boiled in citric acid pH of 6.0 for 20 minutes for antigen retrieval, followed by several washes with tris-buffered saline containing 0.1% Tween-20 (TBST) and endogenous peroxidase quenching in 3% hydrogen peroxidase for 10 minutes. Sections were then washed with TBST and blocked with 5% normal goat serum (NGS; Vector Labs, Burlingame, CA) at room temperature for 30 minutes and incubated overnight with rabbit monoclonal HPGD (HPH005679; Sigma-Aldrich, St. Louis, MO) antibody in a 1:500 dilution in 2% NGS in a cold room. For negative control, appropriate nonimmune rabbit IgG isotype was used at the same concentration of primary antibody. The next day, sections were washed with TBST and then incubated with biotinylated anti-rabbit IgG secondary antibody in a 1:400 (Vector Labs) dilution in 2% NGS for 30 minutes at room temperature. After several rinses with TBST, the sections were incubated with streptavidin peroxidase complex (Elite ABC Kit; Vector Labs) for 30 minutes. After multiple washes, the slides were exposed to diaminobenzidine tetrahydrochloride dehydrate (Vector Labs) as a chromogen for 2 minutes and counterstained with hematoxylin before permanent mounting. Immunoreactivity of HPGD was assessed by histologic score analysis, a semiquantitative method that evaluates the intensity and the number of immune-stained cells, by two blinded investigators (V.d.A. and N.S.) as described,^24,25 using the Axio Imager-A2 light microscope (Zeiss, Oberkochen, Germany) and the ZEN imaging system (Zeiss).

Decidual Cell Cultures

Previously isolated and cultured frozen leukocyte-free first-trimester decidual cells (FTDCs; n = 3) and term decidual cells (TDCs; n = 3), prepared as previously described,^22,26,27 were grown in basal medium containing phenol red–free Dulbecco's modified Eagle's medium/Ham's F12 (1:1; Life Technology, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS; GeminiBio, West Sacramento, CA) and 1% antibiotic and antimycotics (Life Technology) until reaching 70% confluency. These FTDC or TDC cultures were then incubated with either control or 10⁻⁸ mol/L estradiol (E₂; as control in HTR8/SV^neo and trophoblast stem cell experiments; Sigma-Aldrich) or decidualization media named EMC, containing 10⁻⁸ mol/L E₂, 10⁻⁷ mol/L medroxyprogesterone acetate (M; Sigma-Aldrich), and 5 × 10⁻⁵ mol/L 8-bromo 3′,5′-cAMP (C; Sigma-Aldrich) for 7 days (Figure 1A). Medium was refreshed every 48 hours, at the end of each treatment period, and cells were washed twice with phosphate-buffered saline (PBS) and switched to defined medium (DM) consisting of basal media (with 1% antibiotic and antimycotics without FBS) with 0.5 μmol/L ZnSO₄, 5 μmol/L FeSO₄, 1 nmol/L CuSO₄, 50 μg/mL ascorbic acid (Sigma Aldrich), and 1% ITS Premix (insulin, transferrin, and selenous acid; BD Biosciences, Bedford, MA) with or without 5% charcoal-stripped bovine calf serum (BCS) for 96 hours. At the end of the incubation time, the culture media supernatant (CMS) was collected at 24-, 48-, and 96-hour time points (Figure 1A) and centrifuged for 5 minutes at 500 × g and stored at −80°C. This CMS was used to mimic physiological conditions for treatment of trophoblast cell cultures in later experiments (Figure 1A).

Figure 1.

Description of methods for cell culture. A: Collection of term decidual cell (TDC) conditioned media supernatant (CMS) to treat trophoblast cells: (1) TDCs were incubated for 7 days with either 10⁻⁸ mol/L estradiol (E₂) as control or decidualization media (EMC) containing 10⁻⁸ mol/L E₂ + 10⁻⁷ mol/L medroxyprogesterone acetate (M) + 5 × 10⁻⁵ mol/L 8-bromo 3′,5′-cAMP (C) to induce decidualization. (2) TDCs were washed and switched to defined medium (DM) with/without 5% bovine calf serum (BCS). (3) CMS was collected at 24, 48, and 96 hours. (4) HTR8 and extravillous trophoblast–trophoblast stem cell (EVT-TSC) cultures were treated with 1:10 dilution of either E₂ or EMC CMS. B: Collection of CMS from TDCs treated with IL-1ß or lipopolysaccharide (LPS): (1) TDCs were treated with EMC for 7 days for in vitro decidualization. (2) Then, TDCs were incubated with 10 ng/mL IL-1ß or 5 μg/mL LPS for 6 hours. (3) Cells were washed with phosphate-buffered saline ×2 and incubated with DM with 5% BCS for 48 hours. (4) CMS was collected at 48 hours. (5) HTR8 cell cultures were treated with a 1:10 dilution from the collected CMS from LPS- or IL-1ß–treated term decidual cells and incubated for 48 hours. C: Direct treatment of HTR8 cells with inflammatory markers: (1) HTR8 cells were plated to 60% confluency; and (2) HTR8 cells were treated with control, 10 ng/mL IL-1ß, or 5 μg/mL LPS for 6 hours. D: TDC CMS pretreated HTR8 cells were incubated with inflammatory markers: (1) HTR8 cells were plated to 60% confluency; (2) HTR8 cells were preincubated for 48 hours with a 1:10 dilution of EMC CMS obtained from TDC; and (3) HTR8 cells were treated with 10 ng/mL IL-1ß or 5 μg/mL LPS for 6 hours.

In parallel experiments, in vitro decidualized TDCs (n = 3) were washed twice with PBS and treated with vehicle control, 5 μg/mL LPS (L2630; Sigma/Aldrich), or 10 ng/mL recombinant IL-1ß (201-LB-00/CF; R&D Systems, Minneapolis, MN) for 6 hours in DM, as shown in Figure 1B. Then, the medium was aspirated to eliminate the direct effect of LPS and IL-1ß stimuli. The cells were washed with PBS, and then incubated in fresh DM with 5% BCS for 48 hours (Figure 1B). This CMS, which mimics infectious-inflammatory conditions occurring in CAM in later experiments with trophoblast cell cultures, was collected at 48 hours (Figure 1B), centrifuged, and stored at −80°C.

Trophoblast Stem Cell Culture

Human trophoblast stem cells (TSCs) derived from first-trimester placentas were generated by Drs. Okae and Arima²⁸ and were provided by RIKEN BRC through the National Bio-resource Project of the Mext/AMED (Kyoto, Japan). TSCs were cultured and induced to EVT differentiation as described,^28,29 confirmed for human leukocyte antigen-G expression, a well-known EVT marker,^29,30 and incubated 48 hours with either 1:10 diluted control E₂ or EMC-treated TDC-derived CMS (Figure 1A).

Primary Trophoblast Cell Culture

Primary trophoblasts from uncomplicated pregnancies and repeat cesarean deliveries at term were isolated and cultured, as previously described,^31,32 in 6-well plates with Dulbecco's modified Eagle's medium/Ham's F12 (Life Technology) containing 10% FBS (GeminiBio) and 1% antibiotic and antimycotics (Life Technology). The cells were subsequently treated for 96 hours with control media without E₂ or media containing 1:10 dilution CMS obtained from TDC or FTDC cultures to measure HPGD levels.

Trophoblast Cell Line Culture

To compare endogenous expression of HPGD levels in different trophoblastic cell lines, HTR8/SV^neo (a first-trimester EVT cell line), Swan 71 (telomerase-immortalized first-trimester EVT cell line; gift from Dr. Gil Mor, Wayne State University, Detroit, MI), JEG3 (a choriocarcinoma cell line; ATCC, Manassas, VA), and BeWo cells (choriocarcinoma cell line; ATCC) seeded in 6-well plates were grown in Dulbecco's modified Eagle's medium/Ham's F12 (1:1; Life Technology) supplemented with 10% FBS (GeminiBio) and 1% antibiotic and antimycotics (Life Technology). Confluent cells were washed with ice-cold PBS and stored at −80°C for RNA isolation.

Treatment of HTR8/SV^neo Cell Line with CMS from TDCs

HTR8/SV^neo cell line cultures plated onto 6-well plates were treated in DM with 2.5% FBS containing either 1:10 dilution of E₂ or EMC ±5% BCS CMS from TDCs (Figure 1A). After a 48-hour incubation period, the cells were washed with cold PBS twice, and the plates were frozen at −80°C until ready for RNA extraction.

Direct Treatment of HTR8/SV^neo Cells with Inflammatory Mediators

The 60% confluent HTR8/SV^neo cells were directly treated with placebo or 10 ng/mL IL-1ß (R&D Systems) or 5 μg/mL LPS (Sigma-Aldrich) for 6 hours. The cells were washed twice with cold PBS, the PBS was removed entirely, and the plates were stored at −80°C until RNA extraction was performed (Figure 1C).

Treatment of HTR8/SV^neo Cells with CMS from LPS- or IL-1β–Treated TDCs

The 60% confluent HTR8/SV^neo cells were incubated for 48 hours in DM containing 10% CMS from TDC treated with vehicle control or 10 ng/mL IL-1ß (R&D Systems) or 5 μg/mL LPS (Sigma-Aldrich) (Figure 1B). The cells were washed twice with cold PBS, the PBS was aspirated, and the plates were stored at −80°C until RNA extraction was performed.

Treatment of LPS or IL-1β of HTR8/SV^neo Cells Preincubated with TDC CMS

HTR8/SV^neo cells were preincubated for 48 hours with in vitro decidualized CMS of TDCs (Figure 1A) in a 1:10 dilution, followed by treatment with 5 μg/mL LPS or 10 ng/mL IL-1ß for 6 hours (Figure 1D). The cells were washed twice with cold PBS, the PBS was aspirated, and the plates were stored at −80°C until RNA extraction was performed.

Immunocytochemistry

HTR8/SV^neo cells treated for 48 hours with E_2, or EMC CMS derived from TDCs was fixed with 4% paraformaldehyde solution for 20 minutes at 4°C. Thereafter, the slides were washed with TBST, followed by incubation in 3% hydrogen peroxidase for 10 minutes at room temperature. After several washings with TBST, the slides were incubated with 5% NGS for 30 minutes, followed by incubation with rabbit monoclonal HPGD primary antibody at 1:500 in 2% NGS at 4°C overnight. The slides were washed with TBST and incubated with biotinylated anti-rabbit IgG secondary antibody at 1:400 in 2% NGS for 30 minutes at room temperature in a humidified chamber. After several rinses with TBST, incubation with streptavidin peroxidase complex for 30 minutes followed. After washing again with TBST, the slides were exposed to diaminobenzidine tetrahydrochloride dehydrate as a chromogen for 45 seconds and counterstained with hematoxylin before permanent mounting.

RNA Isolation, Reverse Transcription, and qPCR

Total RNA from primary trophoblast cells or HTR8/SV^neo or TSC cultures were isolated using the RNeasy Mini Kit (Qiagen, Germantown, MD), followed by DNase I treatment (Qiagen) to eliminate genomic DNA contamination. Reverse transcription using RETROscript kit (Ambion, Austin, TX) was performed as described,³³ and real-time quantitative PCR (qPCR) was performed using TaqMan gene expression assays to detect gene expression levels of HPGD (probe identifier: Hs00960586_g). ß-Actin (probe identifier: Hs99999903_m1) was used for normalization of the results, and the 2^−ΔΔC_T method was used to calculate the relative expression levels. All reactions were performed in duplicates.

Liquid Chromatography–Tandem Mass Spectrometry

The levels of PGE₂ and its metabolite 15-keto-PGE₂ in CMS from HTR8 (n = 3) treated with control or EMC-treated CMS from TDCs (n = 3) were measured by liquid chromatography–mass spectrometry, as described.^34,35 The liquid chromatography–tandem mass spectrometry standards were purchased from Cayman Chemical (Ann Arbor, MI). Briefly, Analyst software 1.7 (Sciex, Framingham, MA), connected to process the data from Sciex QTRAP 6500⁺ Triple quadrupole mass spectrometer, was used to analyze the extracted lipids. Then, 0.1% formic acid in water and methanol was used as mobile phases A and B, respectively. The Kinetex Polar C-18 column (100 mm × 3 mm × 2.6 μm; Phenomenex, Torrance, CA) was used to separate analytes in a 40-μL sample that was injected into the liquid chromatography–tandem mass spectrometry system. The temperatures were maintained at 50°C for the column and 5°C for the sample. The analytes were separated by graded elution of the mobile phases; phase B was set at 10% initially, then after 0.1 to 2 minutes, 45% to 80% until 16.5 minutes, 98% at 16.6 minutes maintained until 18.5 minutes, and 98%, before going back to 10%. Run time was 20.5 minutes, and flow was 0.5 mL/minute. Sciex OS software version 2.0 was used for data integration.^34,35

Statistical Analysis

Data were analyzed by SigmaStat software version 3.0 (Systat Software, San Jose, CA), and Kolmogorov-Smirnov test (with Lilliefors correction) assessed data for normality. Pairwise multiple comparisons by one-way analysis of variance were followed by the post hoc Student-Newman-Keuls test because data were nonparametrically distributed. Comparison of two groups used a t-test or a U-test for parametric or nonparametric distribution, respectively.

Results

HPGD Is Primarily Expressed by EVTs of Anchoring Villi in the Human Placenta

HPGD immunostaining was performed in first-, second-, (n = 2 each) and third-trimester (n = 5) placental sections to identify cell type–specific expression profile and its gestational age–related expression pattern of HPGD (Figure 2). HPGD immunoreactivity was primarily found in EVTs in the first trimester (Figure 2, A–C), second trimester (Figure 2, D–F), and third trimester (Figure 2, G–I). Increasing HPGD expression was observed from the proximal to distal trophoblastic cell columns in the anchoring villi (Figure 2F). Also, HPGD expression was detected in choriodecidual EVTs in the decidua parietalis (Figure 2, D and G) throughout gestation. However, syncytiotrophoblasts displayed weak HPGD immunoreactivity, and no expression was observed in villous cytotrophoblasts (Figure 2), amnion epithelium (Figure 2, D and G), or decidual stromal cells (Figure 2F), indicating that EVTs are the primary source of HPGD.

Figure 2.

In situ 15-hydroxyprostaglandin dehydrogenase (HPGD) expression in human placental sections throughout pregnancy. A–C: HPGD expression was predominantly detected in extravillous trophoblasts in the first trimester. D–I: In the second and third trimesters, in addition to a strong HPGD expression in extravillous trophoblasts of anchoring villi (F and I) and choriodecidual extravillous trophoblasts (D and G), syncytiotrophoblasts (E and H) displayed a weak HPGD immunoreactivity in the first and second trimesters (n = 2;A–I); in the third trimester (n = 5; D–I). Scale bar = 50 μm (A–I). Original magnification: ×40 (A–C, E, F, H, and I); ×20 (D and G).

TSC Differentiation toward EVTs Enhances HPGD Expression

To confirm the in situ findings, HPGD expression was evaluated by qPCR in previously established TSC cultures and compared with TSC-differentiated EVT-like cells.²⁹ This comparison revealed approximately 100-fold higher HPGD expression in TSC-differentiated EVT-like cells compared with that in TSCs (P < 0.001) (Figure 3A).

Figure 3.

In vitro HPGD mRNA expression in various trophoblastic cell types. A:In vitro differentiation of cultured trophoblastic stem cells (TSCs) into extravillous trophoblasts (EVTs) significantly enhanced HPGD mRNA levels. B: Conditioned media supernatant (CMS) obtained from EMC [containing 10⁻⁸ mol/L estradiol (E₂) + 10⁻⁷ mol/L medroxyprogesterone acetate (M) + 5 × 10⁻⁵ mol/L 8-bromo 3′,5′-cAMP (C)]–treated term decidual cell (TDC) cultures and diluted at 1:10 induced HPGD mRNA levels in vitro differentiated EVT versus CMS from E₂-treated TDCs. C: HPGD mRNA levels in primary cultures of cytotrophoblasts. Compared with control media (Cont), a 1:10 dilution of CMS obtained from first-trimester decidual cell (FTDC) or TDC cultures induced HPGD mRNA levels. D: HTR8 cells displayed significantly higher HPGD mRNA levels compared with other trophoblastic cell lines. Data represent means ± SEM (A–D). n = 3 (A and B); n = 4 (C); n = 5 (D). ∗P < 0.05 Swan, JEG3, or BeWo versus HTR8; ∗∗P < 0.01 TDC or FTDC CMS versus Cont; ∗∗∗P < 0.001 EVT versus TSC (A); ∗∗∗P < 0.001 EVT + EMC CMS versus EVT + E₂ CMS (B).

Decidual Cell–Specific Up-Regulation of HPGD Expression in Trophoblasts

CMS from FTDCs and TDCs promotes Zika virus infection of EVTs in culture.³⁶ These results indicate the presence of a paracrine interaction between decidual cells and EVTs, because of their proximity in the decidua and fetal membranes. Therefore, the possible paracrine impact of decidual cells on regulation of HPGD in EVTs was evaluated by exposing in vitro differentiated TSC-differentiated EVT-like cell cultures to CMS obtained from EMC-treated TDCs for 48 hours. Compared with control CMS, EMC CMS significantly increased HPGD expression in TSC-differentiated EVT-like cells (6.86 ± 0.1 versus 1.001 ± 0.03; P < 0.001) (Figure 3B).

Decidual Cell–Induced HPGD Expression in Trophoblasts Is Independent of Gestational Age

To determine whether decidual cell induction of HPGD expression in EVTs is related to gestational age of decidual cells, primary trophoblast cultures isolated from term placentas were exposed for 48 hours to CMS collected from decidualized primary cultures of FTDCs and TDCs. qPCR results revealed significantly higher HPGD levels (approximately 2.5-fold increase) detected in trophoblasts exposed to CMS obtained from both FTCDs and TDCs compared with those in control (P < 0.01) (Figure 3C), indicating that decidual cell–mediated effect is independent of gestational age.

Altered HPGD Expression in Different Trophoblastic Cell Lines

To determine the optimal trophoblastic cell line for further experiments, endogenous HPGD expression was first examined in four trophoblastic cell lines: HTR8/SV^neo, Swan 71 (both EVT cell lines isolated from first-trimester placenta), JEG3, and BeWo. HTR8/SV^neo displayed significantly higher HPGD levels (P < 0.05) compared with those in Swan 71 and JEG3, whereas no expression was detected in choriocarcinoma BeWo cells (Figure 3D). Therefore, HTR8/SV^neo cells were chosen for further experiments as an in vitro representation of EVTs.

Paracrine Induction of HPGD Expression in EVTs Is Decidualization-Specific

To determine whether decidual cell–mediated increases in HPGD levels were associated with decidualization, HPGD levels were examined in HTR8/SV^neo cells exposed for 48 hours to CMS obtained from either E₂-treated (nondecidualized) or EMC-treated (decidualized) TDCs. This experiment was performed in the presence or absence of 5% BCS to investigate whether serum depletion had an additional effect on HPGD levels. HPGD mRNA levels were significantly higher in HTR8/SV^neo cells exposed to EMC-CMS compared with those in E₂-CMS, in the absence of BCS (P < 0.05) (Figure 4A), indicating a decidualization-specific effect. Additionally, HPGD levels were significantly higher in cells treated with serum-rich E₂-exposed TDCs when compared with those in serum-free E₂-exposed TDCs (means ± SEM, 1.3 ± 0.1 versus 1.02 ± 0.02; P < 0.05) (Figure 4A). Moreover, HTR8/SV^neo cells treated with CMS obtained from serum-rich EMC-exposed TDCs displayed an approximately 1.6-fold higher HPGD level compared with serum-free EMC-exposed TDCs (2.2 ± 0.2 versus 1.4 ± 0.1; P < 0.05) (Figure 4A), implying enhancement of decidual cell–mediated paracrine effect by serum factor(s). Decidual cell–mediated regulation of HPGD protein expression was also confirmed using immunocytochemistry in HTR8/SV^neo cells, showing enhanced HPDG immunoreactivity in trophoblasts exposed to EMC-derived CMS versus E₂-derived CMS (Figure 4B).

Figure 4.

Decidualization-specific induction of 15-hydroxyprostaglandin dehydrogenase (HPGD) mRNA and protein levels in HTR8 cells. A: HPGD expression in HTR8 cells treated with 1:10 dilution of conditioned media supernatant (CMS) obtained from decidualized [EMC, containing 10⁻⁸ mol/L estradiol (E₂) + 10⁻⁷ mol/L medroxyprogesterone acetate (M) + 5 × 10⁻⁵ mol/L 8-bromo 3′,5′-cAMP (C), exposed] or control (E₂-exposed) term decidual cells ± 5% bovine calf serum (BCS). Compared with E₂ without BCS, HPGD mRNA levels were significantly increased in HTR8 treated with EMC without BCS or E₂ with BCS, indicating both EMC and serum effects. However, EMC-induced increase was significantly higher in the presence of BCS. B: EMC-induced HPDG immunoreactivity in HTR8 cells. Decidual cell–mediated regulation of HPGD expression in HTR8 cells. HPGD expression in HTR8 cell lines treated with a 1:10 dilution of conditioned media supernatant obtained from third-trimester decidual cell cultures treated with E₂ versus EMC. Data represent means ± SEM (A). n = 3 from different TDC cultures (A). ∗P < 0.05 EMC – BCS or E₂ + 5% BCS versus E₂ – BCS; +P < 0.05 EMC + 5% BCS versus EMC – BCS (E₂ – BCS used as control). Scale bars = 25 μm (B). Original magnification: ×40 (B, left panels); ×100 (B, right panels).

To investigate the time-dependent effect of CMS on HPGD levels, HTR8/SV^neo cells were treated with EMC-CMS of TDCs obtained at 24, 48, and 96 hours. Subsequent qPCR analysis revealed that HPGD levels were highest at 48 hours, followed by those at 96 hours, versus 24-hour CMS treatment (2.8 ± 0.4 or 1.9 ± 0.2, respectively, versus 1.03 ± 0.03; P < 0.01) (Supplemental Figure S1A). Subsequently, the dose-dependent effect of EMC-CMS from TDCs on HPDG levels was investigated in HTR8/SV^neo cells. The cells were treated for 48 hours with either control medium (DM) or medium containing 10% or 25% EMC-CMS that was diluted in DM. Both 10% and 25% EMC-CMS significantly increased HPDG levels compared with control (2.6 ± 0.2 or 2.5 ± 0.2, respectively, versus 1.03 ± 0.04; P < 0.001) (Supplemental Figure S1B).

LPS- or IL-1β–Mediated Regulation of HPGD Levels

To evaluate the effect of proinflammatory mediators on HPGD levels, HTR8/SV^neo cells were treated with vehicle control, 10 ng/mL IL-1β, or 5 μg/mL LPS for 6 hours. qPCR results indicated that both IL-1β and LPS treatment significantly inhibited HPGD levels versus control (0.47 ± 0.1 or 0.57 ± 0.1, respectively, versus 1.03 ± 0.03; P < 0.01) (Figure 5A). Whether EMC-CMS derived from TDC cultures treated with IL-1β or LPS (mimicking conditions of deciduitis) maintained the ability to induce HPGD levels was examined next. HTR8/SV^neo cells were incubated for 48 hours with EMC-CMS collected from TDC cultures, which were pretreated for 6 hours with 10 ng/mL IL-1β or 5 μg/mL LPS. Pretreatment of TDCs with IL-1β or LPS resulted in significant inhibition of HPGD levels in HTR8/SV^neo cells compared with that in EMC-CMS without proinflammatory mediators (0.7 ± 0.1 or 0.63 ± 0.1, respectively, versus 1.03 ± 0.03; P < 0.05) (Figure 5B).

Figure 5.

Decidual cell conditioned media supernatant (CMS) prevents proinflammatory-mediated suppression of HPGD expression in HTR8 cells. A: HPGD expression in HTR8 cells treated with control (Cont) or 10 ng/mL IL-1β or 5 μg/mL lipopolysaccharide (LPS). B: HPGD expression in HTR8 cells incubated for 48 hours with a 1:10 dilution of CMS collected from Cont or 10 ng/mL IL-1β or 5 μg/mL LPS treated term decidual cell (TDC) cultures. C: A 1:10 dilution of EMC [containing 10⁻⁸ mol/L estradiol (E) + 10⁻⁷ mol/L medroxyprogesterone acetate (M) + 5 × 10⁻⁵ mol/L 8-bromo 3′,5′-cAMP (C)]–CMS pretreatment blocks inhibitory effect of IL-1β or LPS on HPGD levels in HTR8 cells. P = 0.47. Data represent means ± SEM (A–C). n = 3 (A–C). ∗P < 0.05 IL-1β or LPS CMS versus Cont CMS; ∗∗P < 0.01 IL-1β or LPS versus Cont.

Decidual Cells Prevent Infectious/Inflammatory-Mediated Inhibition of HPGD Levels in Trophoblasts

To determine whether TDCs display a protective role in trophoblasts against inflammatory or infectious stimuli, HTR8/SV^neo cells were pretreated for 48 hours with CMS from decidualized TDCs, followed by exposure to 10 ng/mL IL-1β or 5 μg/mL LPS for 6 hours (mimicking conditions of infection in the fetus, such as amnionitis). Interestingly, neither IL-1β nor LPS significantly suppressed HPGD mRNA levels versus control CMS (0.8 ± 0.1 or 1.11 ± 0.2, respectively, versus 1.03 ± 0.04; P = 0.47) (Figure 5C).

Prostaglandin-endoperoxide synthase 2 (PTGS2), also known as cyclooxygenase 2, is the key enzyme in prostaglandin biosynthesis, and acts as both a dioxygenase and a peroxidase.37, 38, 39 Therefore, to confirm the inflammatory effects of IL-1β and LPS in HTR8/SV^neo cells, PTGS2 mRNA levels were measured by qPCR. IL-1β and LPS significantly increased PTGS2 levels versus control (4.02 ± 0.03 or 1.69 ± 0.2, respectively, versus 1.04 ± 0.04; P < 0.001) (Supplemental Figure S2A) in HTR8/SV^neo cells. PTGS2 levels were also significantly increased by IL-1β and LPS in HTR8/SV^neo cells pretreated with EMC-CMS versus control CMS (4.8 ± 0.6 or 2.3 ± 0.3, respectively, versus 1.04 ± 0.02; P < 0.01) (Supplemental Figure S2B). To investigate the functional role of elevated HPGD levels in HTR8 cells treated with ECM-CMS obtained from TDCs, PGE₂- and HPGD-mediated metabolite 15-keto PGE₂ levels were measured by liquid chromatography–tandem mass spectrometry. Compared with the control group, the ratio of 15-keto-PGE₂/total PGE₂ levels was significantly increased in HTR8 cells treated with TDC CMS (Supplemental Figure S3A), whereas 15-keto-PGE₂ or total PGE₂ levels did not reach significance (Supplemental Figure S3, B and C).

EVTs Exhibit Lower HPGD Levels in High Leukocyte-Infiltrated Areas of CAM-Complicated Placentas

To confirm the protective effect of TDCs on trophoblasts observed in in vitro results, HPGD immunohistochemistry was performed in placental specimens of patients affected with CAM versus GA-matched controls. HPGD immunoreactivity was evaluated by histologic score in chorioamniotic membranes with high and low leukocyte-infiltrated areas from CAM specimens and compared with specimens from GA-matched controls. In serial sections, EVTs and decidual cells were distinguished by differential staining for cytokeratin 7 (brown) and vimentin (red) double immunostaining, respectively (Figure 6, A–C). Compared with control specimens (Figure 6D), areas with high leukocyte infiltration displayed lower levels of HPGD immunostaining (Figure 6E); however, the same was not observed in areas with low leukocyte infiltration (Figure 6F). Moreover, HPDG levels were higher in fetal membrane EVTs adjacent to decidua parietalis versus those adjacent to amnion (Figure 6F). Histologic score analysis confirmed that HPGD immunoreactivity is significantly reduced in EVTs from CAM with high leukocyte infiltrates versus control EVTs or EVTs from CAM with low leukocyte infiltrates (31.2 ± 7.9 versus 121.4 ± 10.6 or 160.6 ± 18.6, respectively; P < 0.001).

Figure 6.

Immunohistochemical comparison of placental specimens for 15-hydroxyprostaglandin dehydrogenase (HPGD) expression. A–C: Extravillous trophoblasts (EVTs) and decidual cells are distinguished by cytokeratin (CYT; brown) and vimentin (VIM; red) dual immunostaining, respectively, in control (A) and chorioamnionitis (CAM) complicated placentas with high (B) and low (C) leukocytic infiltration. D–F: HPGD immunoreactivity in control (D) and CAM complicated placentas with high (E) and low (F) leukocytic infiltration. Leukocyte numbers within the chorionic EVT layer between 25 and 50 accepted as low leukocytic infiltrates (C and F), and >50 considered high infiltrated areas (B and E). G: Bar graph represents means ± SEM HPGD histologic score (HSCORE). n = 3 (D); n = 6 (E and F). ∗∗∗P < 0.001 CAM with high leukocyte-infiltrated EVT versus control EVT or CAM with low leukocyte-infiltrated EVT. Scale bar = 50 μm (A–F). Original magnification: ×40 (A, C, D, and F); ×20 (B and E).

Discussion

The current study extends the knowledge about maternal and fetal cell crosstalk by identifying the role of decidual cells on HPGD expression in trophoblasts. The expression of HPGD in first-, second-, and third-trimester placental sections demonstrated that HPGD is primarily expressed by EVTs throughout gestation. Gradual increase in HPGD expression observed from the proximal to the distal trophoblastic cell columns in the anchoring villa, which are adjacent to decidual cells, along with the in vitro observation of elevated trophoblastic HPGD levels by decidual culture media supernatant, suggest a paracrine communication between maternal and fetal cells.

Higher HPGD levels in choriodecidual EVTs, lower levels in syncytiotrophoblast, and no immunostaining in villous cytotrophoblasts, amnion epithelium, and decidual cells further support dominant expression of HPGD in EVTs adjacent to decidual cells. These results are consistent with prior studies performed on term placental specimens.^40,41 Additionally, this study shows, for the first time, the expression of HPGD in EVTs in first-trimester placental specimens, suggesting that HPGD is involved in pregnancy maintenance as early as the first trimester. This premise is supported by a prior study showing early pregnancy loss in Hpgd knockout mice.¹⁹ Erwich and Keirse⁴² also found increased activity of HPGD in syncytiotrophoblasts and villous trophoblasts at term compared with the first trimester, suggesting that the fetus protects itself from preterm birth by degrading biologically active maternally derived PGs at the maternal-fetal interface. This is supported by a study in mice with hypomorphic Hpgd expression, which causes preterm birth.²⁰ Overall, these studies have implicated a central role of HPGD in pregnancy maintenance.

At the maternal-fetal interface, decidual cell and EVT crosstalk using hormones, growth factors, metabolites, and other secreted factors, appears to be required for maternal adaptation and appropriate placental development, helping to explain the dense infiltration of human decidua by EVTs.⁴³ This is the first report of a preferential increase in HPGD expression by EVTs differentiated from human TSCs. In situ observation of higher HPGD expression by EVTs adjacent to decidual cells suggests a decidual cell–mediated paracrine induction of HPGD expression in EVTs. This hypothesis is further supported by in vitro findings that TDC-derived CMS enhanced HPGD levels in both human TSC-derived EVT and primary trophoblast cultures. Moreover, the observation showing a similar increase in HPGD expression in primary trophoblast cultures treated with CMS from FTDCs or TDCs indicated that this paracrine stimulation occurs throughout gestation.

Decidualization is critical in trophoblast invasion and placentation needed for successful pregnancy. It is well known that decidualization induces decidual stromal cells to become endocrine cells secreting prolactin and insulin-like growth factor binding protein-1.⁴⁴ Herein, significant up-regulation of trophoblastic HPGD levels by CMS from decidualized (EMC-treated) versus undecidualized (E₂-treated) TDCs verified that the paracrine induction is specific to decidualization. Additionally, both 96- and 48-hour decidual cell CMS treatment had a significantly higher inductive effect on trophoblastic HPGD levels compared with 24-hour treatment, implying that decidual cell–derived factors can maintain the effect on trophoblasts at least up to 96 hours. These results expand on previous studies demonstrating that decidual secretome alters trophoblast physiology, including expression of matrix metalloprotein, growth factors, and cytokines.45, 46, 47

Although the study clearly demonstrated that decidual cells have an inductive effect on HPGD under physiological conditions to maintain pregnancy, this inductive effect may be impaired under infection-induced inflammatory conditions, such as CAM, leading to preterm birth.^6,12,13 Exposure of trophoblasts to CMS from decidual cells pretreated with LPS and IL-1β, mimicking deciduitis (decidual-specific infection), resulted in blunting of HPGD expression rather than induction, as detected in the absence of inflammatory conditions (Figures 3 and 4). This observation suggests that under infectious/inflammatory conditions, decidual cells down-regulate trophoblastic HPGD expression, likely enhancing prostaglandin bioavailability (Figure 5B), which contributes to preterm birth. This response is supported by previous findings that infectious or bioactive inflammatory mediators cause impaired decidualization,^48,49 which, in turn, may prevent the secretion of paracrine factor(s) that induce trophoblastic HPGD levels. This is supported by a previous report showing significantly reduced levels of progesterone receptor expression in decidual cells from CAM-complicated specimens and in IL-1β–treated TDC cultures,¹³ consistent with impaired decidualization observed in progesterone receptor silencing in human or knockdown in mice.50, 51, 52, 53

Infection-induced inflammatory activation results in mature IL-1β proinflammatory cytokine production in trophoblast and other cell types.54, 55, 56 Moreover, prior studies have reported inhibitory effects of direct LPS and IL-1β treatment on HPGD expression in both human and animal models.^18,41,57 Herein, HPGD expression was down-regulated following exposure of trophoblastic cells to LPS and IL-1β, which mimicked fetal site-specific infection (Figure 5A). However, this down-regulation was blunted in trophoblast pretreated with TDC-CMS (Figure 5C), indicating that decidual cells may be able to maintain their stimulatory effect on trophoblastic HPGD expression in cases where infection is present only on the fetal sides. It occurs with early stages of hematogenous spread pathogens, like listeria, although ascending infections leading to CAM are typically preceded by deciduitis.

In situ findings showing significantly lower HPGD levels in chorionic EVTs with high leukocytic infiltration versus low leukocytic infiltration in CAM-complicated specimens were consistent with those of Challis et al.⁴¹ This also provided in situ evidence that under low leukocytic infiltration, an indicator of mild levels of inflammation, decidual cell–secreted molecule(s) sustain their stimulatory impact on HPGD levels in trophoblasts. However, under high leukocyte infiltration, indicative of severe inflammation, a disruption of decidual cell–mediated paracrine stimulation on trophoblastic HPGD expression occurs, possibly tilting the balance toward preterm birth as this enzyme functions to inactivate labor-inducing PGs.¹⁴

In conclusion, the results indicate that there is paracrine stimulation by decidual cells on EVTs to up-regulate expression of HPGD, which inactivates PGE₂ and PGF_2α under physiological conditions, supporting a pregnancy-maintaining role of decidual cells, as shown in Figure 7. However, infectious/inflammatory stimuli in decidual cells caused a paracrine inhibition of trophoblastic expression of PGE₂/PGF_2α-metabolizing enzyme HPGD, thereby contributing to CAM-associated PTB (Figure 7). Further studies are needed to identify decidual cell–secreted molecule(s) responsible for induction of trophoblastic HPGD expression, as they carry therapeutic potential in preventing CAM-associated PTB.

Figure 7.

Schematic representation of the regulation of 15-hydroxyprostaglandin dehydrogenase (HPGD) expression in trophoblasts by paracrine factors derived from decidual cells. Decidualization promotes secretion of factor(s) enhancing HPGD expression in extravillous trophoblasts (EVTs), which oxidize prostaglandin (PG) E₂/F_2α into the inactive 15-keto PGE₂/PGF_2α form, serving as a metabolic barrier to maintain pregnancy under physiological conditions (green arrows). Under pathologic conditions, such as infection, proinflammatory mediators in decidual cells reverse this paracrine stimulation to a paracrine inhibition of HPGD expression in trophoblasts, causing increased levels of PGE₂ and PGF_2α, thereby contributing to chorioamnionitis-associated preterm birth (red arrows). Moreover, proinflammatory mediators can also directly inhibit HPGD expression and enhance PGH₂ levels from arachidonic acid by increasing cyclooxygenase 2 (COX2), leading to overall increase in PGE₂/PGF_2α bioavailability. EMC, 10⁻⁸ mol/L estradiol (E) + 10⁻⁷ mol/L medroxyprogesterone acetate (M) + 5 × 10⁻⁵ mol/L 8-bromo 3′,5′-cAMP (C); LPS, lipopolysaccharide; TDC, term decidual cell.

Disclosure Statement

None declared.

Supplemental Data

Supplemental Figure 1.

Effect of EMC [containing 10⁻⁸ mol/L estradiol (E) + 10⁻⁷ mol/L medroxyprogesterone acetate (M) + 5 × 10⁻⁵ mol/L 8-bromo 3′,5′-cAMP (C)]–conditioned media supernatant (CMS) on 15-hydroxyprostaglandin dehydrogenase (HPGD) expression in HTR8/SV^neo cells at different time points and doses. A:HPGD mRNA expression in HTR8/SV^neo cells treated with EMC-CMS of term decidual cells obtained at 24, 48, and 96 hours. B: Dose-dependent effect of EMC-CMS on HPGD mRNA expression in HTR8/SV^neo cells. Both 25% and 10% EMC-CMS significantly increased HPDG levels. Data represent means ± SEM (A and B). n = 3 (A and B). ∗∗P < 0.01 versus 24 hours; ∗∗∗P < 0.001 versus control (Cont).

Supplemental Figure 2.

A: IL-1β or lipopolysaccharide (LPS) increases mRNA expression of PTGS2 in HTR8/SV^neo cells. IL-1β at 10 ng/mL or LPS at 5 μg/mL significantly increased PTGS2 levels in HTR8/SV^neo cells versus control (Cont; 4.02 ± 0.03 or 1.69 ± 0.24, respectively, versus 1.04 ± 0.04). B: IL-1β or LPS increases PTGS2 mRNA expression in conditioned media supernatant (CMS)–pretreated HTR8/SV^neo cells. IL-1β at 10 ng/mL or LPS at 5 μg/mL significantly increased PTGS2 levels in HTR8/SV^neo cells pretreated with CMS versus control (4.88 ± 0.65 or 2.37 ± 0.32, respectively, versus 1.03 ± 0.03). n = 3 (A and B). ∗∗P < 0.01, ∗∗∗P < 0.001. TDC, term decidual cell.

Supplemental Figure 3.

15-Keto (oxidized) and total prostaglandin (PG) E₂ levels in conditioned media supernatant (CMS) of HTR8/SV^neo cultures exposed to control (Cont) media or EMC [containing 10⁻⁸ mol/L estradiol (E) + 10⁻⁷ mol/L medroxyprogesterone acetate (M) + 5 × 10⁻⁵ mol/L 8-bromo 3′,5′-cAMP (C)] media obtained from term decidual cell (TDC) cultures. A: Ratio of 15-keto-PGE₂ levels/total PGE₂ levels, which are significantly higher in HTR8/SV^neo cells treated with EMC-media versus control media by liquid chromatography–tandem mass spectrometry (LC-MS/MS). B: Total PGE₂ levels by LC-MS/MS. C: 15-Keto-PGE₂ levels by LC-MS/MS. Data represent means ± SEM (A–C). n = 3 (A–C). ∗P < 0.05 versus Cont.