Syncytialization alters the extracellular matrix and barrier function of placental trophoblasts

Abstract

The human placenta is of vital importance for proper nutrient and waste exchange, immune regulation, and overall fetal health and growth. Specifically, the extracellular matrix (ECM) of placental syncytiotrophoblasts, which extends outward from the placental chorionic villi into maternal blood, acts on a molecular level to regulate and maintain this barrier. Importantly, placental barrier dysfunction has been linked to diseases of pregnancy such as preeclampsia and intrauterine growth restriction. To help facilitate our understanding of the interface and develop therapeutics to repair or prevent dysfunction of the placental barrier, in vitro models of the placental ECM would be of great value. In this study, we aimed to characterize the ECM of an in vitro model of the placental barrier using syncytialized BeWo choriocarcinoma cells. Syncytialization caused a marked change in syndecans, integral proteoglycans of the ECM, which matched observations of in vivo placental ECM. Syndecan-1 expression increased greatly and predominated the other variants. Barrier function of the ECM, as measured by electric cell-substrate impedance sensing (ECIS), increased significantly during and after syncytialization, whereas the ability of THP-1 monocytes to adhere to syncytialized BeWos was greatly reduced compared with nonsyncytialized controls. Furthermore, ECIS measurements indicated that ECM degradation with matrix metalloproteinase-9 (MMP-9), but not heparanase, decreased barrier function. This decrease in ECIS-measured barrier function was not associated with any changes in THP-1 adherence to syncytialized BeWos treated with heparanase or MMP-9. Thus, syncytialization of BeWos provides a physiologically accurate placental ECM with a barrier function matching that seen in vivo.

INTRODUCTION

The human placenta functions in many capacities during pregnancy, including acting as the primary barrier between maternal and fetal circulatory systems. Broad characteristics of the placental barrier have been defined, including, but not limited to, regulating cellular interactions, nutrient transport, hormone passage, and waste product removal (1, 2). Although our understanding of exactly how the placenta regulates the fetoplacental environment to maintain health is still limited, it is clear that placental barrier dysfunction can lead to disease (3). In addition to this, other complications of pregnancy, such as preeclampsia and intrauterine growth restriction, are often associated with placental dysfunction, specifically with regard to syncytiotrophoblasts, which are responsible for maintaining the placental barrier (4). Despite research being unclear on whether placental barrier dysfunction is the cause or effect of these diseases (5), it is evident that placental barrier dysfunction poses a significant problem to both maternal and fetal health. Thus, better understanding of placental dysfunction could lead to the development of therapeutics aimed at maintaining the placental barrier and improving maternal and fetal health in diseases of pregnancy.

At the cellular level, fetal-derived syncytiotrophoblasts compose the outermost layer of the fetal chorion, which projects into and lies in direct contact with maternal blood. Because of their location, syncytiotrophoblasts are inherently tasked with maintaining the physical and molecular barrier and regulating cellular interactions with the placental unit. These actions are largely achieved through the placental extracellular matrix (ECM), which projects from the apical side of syncytiotrophoblasts into the lumen of the maternal side of the placenta. The placental ECM is structurally very similar to the vascular ECM (6–8), often referred to as the glycocalyx, which has been described as contributing greatly to the blood-brain barrier and the vascular barrier properties (9, 10). It is therefore reasonable to assume the placental ECM is very important to placental barrier function, and it should be investigated thoroughly. Syndecans are of particular interest, as they are one of the most abundant proteoglycans in the ECM and participate in a variety of cellular processes including ligand binding, mechanical signal transduction, and leukocyte extravasation, among others (11, 12).

Despite its importance, the placental ECM has not been extensively studied and models of the placental barrier, although they exist, are not fully characterized. For example, Wong et al. (13) recently established an excellent in vitro coculture model of the placental barrier using forskolin-induced syncytialized BeWos and human umbilical vein endothelial cells (HUVECs). This model provides a physiologically relevant placental barrier with syncytiotrophoblast-like and endothelial cell monolayers; however, characterization of the syncytiotrophoblast ECM did not fall under the scope of their study. The use of forskolin to induce syncytialization in BeWos is well characterized, and syncytialized BeWos have a morphological and transcriptomic profile that matches those of syncytiotrophoblasts (14–16). Although many syncytium-specific proteins have been described in forskolin-induced syncytialized BeWos, the syncytialized BeWo ECM has not been described. In the current study, we sought to characterize the molecular composition and assess the barrier function of an in vitro model of the placental ECM. We aimed to test the hypothesis that syncytialization of trophoblasts produces a physiologically accurate ECM that regulates their barrier function. To achieve this, we induced syncytialization in the immortalized BeWo choriocarcinoma cell line. After confirming syncytialization both visually and through a migration assay, we then characterized changes in ECM composition. Furthermore, barrier function was measured via an electrical cell-substrate impedance sensing (ECIS) array and by the ability of syncytialized BeWos to exclude adherence of THP-1 monocytes. Finally, we examined the effects on the barrier function of these syncytialized cells of some ECM remodeling enzymes [matrix metalloproteinase-9 (MMP-9) and heparanase] known to be present in the placenta.

METHODS

BeWo Cell Culture Maintenance

BeWo cells (ATCC) were maintained in complete media, DMEM/Hams F12 50/50 mix (Corning) supplemented with 10% fetal bovine serum (ATCC) and 1% penicillin-streptomycin (GE Healthcare Life Sciences). THP-1 human monocytes were maintained in complete media, RPMI-1640 medium (ATCC) supplemented with 0.05 mM 2-Mercaptoethanol, 10% fetal bovine serum (ATCC), and 1% penicillin-streptomycin (GE Healthcare Life Sciences). Cells were kept in incubators at 37°C, 5% CO₂, and 21% O₂.

Phase Contrast Imaging

BeWos were plated in 24-well plates (Corning) at a density of 1.75 × 10⁵ cells per well in 1 mL of media. The forskolin-treated group was treated with 25 µM forskolin at the time of seeding, whereas the vehicle control group received an equal volume of DMSO. After 48 h of treatment, phase contrast images were obtained with a phase contrast microscope (Fisher Scientific) using a ×10 objective.

Immunofluorescent Imaging

For fluorescent imaging, BeWos were plated in two-well chamber slides (Lab-Tek II, ThermoFisher) at a seeding density of 2.5 × 10⁵/mL/well. Cells were plated and maintained in the conditions stated earlier, except for the forskolin group being treated with 25 µM forskolin (Alfa Aesar) and the vehicle control group being treated with an equal volume of DMSO at the time of plating. After 48 h of treatment, the cells were stained for immunofluorescent imaging in the following manner: media were aspirated off and wells were gently washed in iced PBS before being aspirated again. Then, 100 µL of cooled methanol was added to each well and the chamber slides were incubated at 37°C for 5 min. The methanol was then aspirated off and the chambers were removed from the slides. The slides were then washed with iced PBS and aspirated again. This wash was repeated three times. Then, 150 µL of Hoechst (Thermo Fisher) nuclear stain (1 µg/mL) was applied to each slide for 5 min and the wash step was repeated three times. Next, Cytopainter Deep Red membrane stain (Abcam) was applied for 15 min before aspiration and three additional wash steps. A drop of antifade was applied and coverslips were placed on top of each slide before imaging on an LSM confocal fluorescent microscope with the ×60 objective (Nikon Eclipse Ti2 with a Niko C2 Imaging System and associated software). Hoechst was excited at 405 nm and viewed in the 420–470 nm range (blue), whereas Cytopainter Deep Red was excited at 561 nm and viewed in the 575–625 nm range (red).

For imaging of syndecan-1, the aforementioned experimental protocol was followed, with the following exceptions: after the cells were fixed, cells were probed with an antibody specific to human syndecan-1 (H-174, Santa Cruz) diluted 1:250 for 1 h, then washed three times. A goat anti-rabbit secondary antibody (ab6717) diluted 1:1,000 (excitation 493, emission 528, green) was then added for 1 h before the final three washes were performed. Images were taken with an EVOS FL (Life Technologies, Eugene, OR). Hoechst staining was also performed in the same manner as mentioned earlier. Validation of the syndecan-1 antibody was performed originally by use of positive controls and siRNA-mediated knockdown cell lines.

BeWo Migration

BeWo migration was evaluated using the CytoSelect 24-Well Cell Migration Assay Kit (Cell Biolabs Inc.) and manufacturer protocol was followed. Briefly, a cell suspension of 3.0 × 10⁵ cells in 0.3 mL of serum media was prepared in 24-well inserts (Costar) with 8-µm pores. Initially, no media were added to the wells below the inserts. One group of BeWos was syncytialized with 25 µM forskolin, whereas the other group was treated with an equal volume of DMSO as a control. After 48 h, the media were aspirated and replaced with 0.3 mL of serumless media, whereas the wells below the inserts were filled with 0.5 mL of serum media so that the bottom of the inserts were covered. After an additional 24 h, the media were aspirated again. The insides of the inserts were swabbed with wet cotton swabs to remove nonmigratory cells. Then, 0.4 mL of cell stain solution was added to each insert for a 10-min incubation. The inserts were then washed in water and air-dried. Next, 0.2 mL of extraction solution was added to each insert for a 10-min incubation. Finally, 100µL of solution was taken from each insert and added to a 96-well plate and absorbance was measured at 560 nm. A sample size of six was used for each group.

PCR Experiments

BeWos were plated in 24-well plates (Corning) at a density of 1.75 × 10⁵ cells per well in 1 mL of media. The forskolin-treated group was treated with 25 µM forskolin at the time of seeding, whereas the vehicle control group received an equal volume of DMSO. After 48 h of treatment, media were collected and frozen for later analysis and the cells were trypsinized, spun down, placed in RNAlater (ThermoFisher), and stored in a refrigerator until RNA isolation. RNA was isolated using a PureLink RNA Mini Kit (Ambion) and the associated protocol. RNA concentrations were determined using a Nanodrop 2000c (ThermoFisher). cDNA was synthesized using 300 ng of RNA and a RevertAid First Strand cDNA Synthesis Kit (ThermoFisher) and the associated protocol. Gene expression was carried out by digital droplet PCR. The PCR reaction used digital droplet PCR probes Supermix, 1 µL of TaqMan probes, and a total of 15 ng of RNA per sample. The reaction mixture was separated into nanodroplets using an automated droplet generator (BioRad). PCR was carried out for 40 cycles in the following thermal cycler steps: 1) 95°C for 10 min, 2) 94°C for 30 s, 3) 60°C for 1 min, 4) repeat steps 2–3 39 times, 5) 98°C for 10 min, and 6) 4°C hold. Droplets were counted using the QX200 Droplet Reader. Data were analyzed and copy count was calculated using QuantaSoft software. Samples were run in duplicate. The following TaqMan probes were used for the digital droplet PCR: Human SDC-1 (Hs00896423_m1), Human SDC-2 (Hs00299807_m1), Human SDC-3 (Hs01568665_m1), and Human SDC-4 (Hs00161617_m1), all procured from Invitrogen. A sample size of six was used for each group.

Protein Experiments

Two separate protein isolation experiments were performed. Cells were seeded at a density of 6.0 × 10⁵ cells per well in 3 mL of media. For the first experiment, there were two groups: one treated with 25 µM forskolin and the other treated with an equal volume of DMSO. After 48 h, the media were collected and frozen for later analysis while cells were lysed with RIPA Lysis Buffer (Santa Cruz) following the manufacturer’s protocol. Isolated protein was frozen for later analysis. For the second experiment, cells were organized into four groups. Initially, all groups were treated with 25 µM forskolin for 48 h. The media were then changed and replaced with media containing the following: forskolin group received 25 µM forskolin again, forskolin + heparanase group received 25 µM forskolin and 50 ng/mL heparanase, forskolin + MMP-9 group received 25 µM forskolin and 50 ng/mL activated MMP-9, and forskolin + combo group received 25 µM forskolin, 50 ng/mL heparanase, and 50 ng/mL activated MMP-9. A sample size of six was used for all groups.

Media and cell lysate levels of syndecan-1 and syndecan-4 for the aforementioned experiments were determined using DuoSet ELISA kits (DY2780 and DY2918, respectively, R&D Systems, Minneapolis, MN), which are specific to human syndecan-1 and human syndecan-4. Briefly, 96-well plates were treated with their respective capture antibodies for 24 h. The plates were washed with buffer and blocked with reagent diluent for 1 h. The plates were washed again, aspirated, and the respective standards and samples (all in duplicate) were plated and incubated for 2 h. The plates were washed again, and the respective detection antibodies were added to the plates for 2 h. The plates were once again washed, followed by a 20-min incubation with streptavidin-horseradish peroxidase in a black box. The plates were washed again before the addition of color reagent for 20 min, during which the plate was kept in a black box. After the 20-min incubation, stop solution was added, and the plates were read using the Infinite M200 Pro plate reader and Magellan software (Tecan, Grodig, Austria). For cell lysate levels of syndecan-1 and syndecan-4, measurements were normalized to the concentration of protein for each individual sample as determined by the Pierce bicinchoninic acid assay (Thermo Fisher) (expressed as pg of syndecan-1 or syndecan-4 per µg of total protein).

Electrical Cell-Substrate Impedance Sensing

Two electrical cell-substrate impedance sensing (ECIS) experiments were carried out using an ECISz instrument (Applied BioPhysics). Before cell seeding, 200 µL of media was added to each well of the array, and the arrays were “stabilized” using the provided ECIS software. This procedure cleans the electrodes to provide more accurate measurements of impedance. BeWos were plated in eight-well arrays (8W10E+ PET array, Applied BioPhysics) at a density of 8.0 × 10⁴ per well in a total of 500 µL of media. For the first experiment, wells were divided into two groups and treated with 25 µM forskolin or an equal volume of DMSO for 48 h. For the second experiment, wells were divided into four groups, with three being initially treated with 25 µM forskolin and one being treated with an equal volume of DMSO at the time of plating. After 48 h, media were replaced with the following: forskolin group received 25 µM forskolin again, forskolin + heparanase group received 25 µM forskolin and 50 ng/mL heparanase, forskolin + MMP-9 group received 25 µM forskolin and 50 ng/mL activated MMP-9, and forskolin + combo group received 25 µM forskolin, 50 ng/mL heparanase, and 50 ng/mL activated MMP-9 for 24 additional hours. A sample size of three was used for all experimental groups. An additional well with just media was used to ensure there was no drift in instrument readings over time. Electrical impedance was measured every 60 s at all available frequencies throughout the entire experiment, and measurements were only paused briefly during the media change. The 4,000 Hz AC frequency was used to measure impedance as a form of barrier function, owing to the fact that at 4,000 Hz, current cannot move through cell membranes and thus must move through cells’ ECM, which allows for a specific measurement of the ECM’s resistance to electrical flow without confounding cell membrane measurements. Data are presented as normalized impedance, to account for differences in each individual well’s initial impedance levels.

Monocyte Adhesion

Cell adhesion was assessed using a Vybrant Cell Adhesion Assay kit (Thermo Fisher). BeWos were plated in 0.3 mL of BeWo media at a density of 50,000 cells per well. Groups were treated with either 25 µM forskolin or an equal volume of DMSO for 48 h. A sample size of five was used for each group. For the heparanase and MMP-9 experiment, 1 h before addition of THP-1s, groups were treated with heparanase (HPSE), MMP-9, or a combination, and one group was left as a forskolin-treated control. A sample size of six was used for each group. A THP-1 cell suspension was prepared at 5.0 × 10⁶ cells per mL THP-1 media and treated with 5 µL of Calcein AM per mL of media. The suspension was incubated for 30 min at 37°C. Cells were centrifuged for 5 min at 300 g, the media were decanted, and the cells were washed in media. This process was repeated, and the cells were resuspended at the same density. BeWo media were aspirated and replaced with 100 µL of the THP-1 cell suspension and incubated for 2 h to allow for adhesion. After 2 h, 100 µL of media was added to each well, and they were blotted dry on a paper towel. This was repeated for a total of four times to remove all nonadherent THP-1s. After the last blot, 200 µL of PBS was added to each well. Fluorescent intensity was measured with an absorbance of 494 and emission of 517. A well of just 100 µL of Calcein AM-treated THP-1s + 100 µL of PBS was used as a 100% control, and a well of media + PBS was used as a negative control.

Statistical Analysis

All data are represented as means ± standard error. Comparisons between two groups were performed by unpaired two-tailed Student’s t test. Comparisons between other groups larger than two were performed using an ordinary one-way ANOVA with a Dunnett’s multiple-comparisons test. The significance value cutoff was P < 0.05, and statistical significance is represented by a bar connecting two groups (*P < 0.05, **P < 0.01, ***P < 0.001).

RESULTS

Syncytialization Alters ECM Protein Expression

First, we aimed to visually confirm syncytialization of BeWos after 48 h of treatment with 25 µM forskolin using cell membrane and nuclear stains. Using phase contrast microscopy, a cobblestone-like appearance was observed in the vehicle-treated cells where nuclei were clearly separated by independent cell membranes (Fig. 1A). In the forskolin-treated group, cell membranes were much larger and contained multiple nuclei within each cell membrane (Fig. 1B). Similarly, upon fluorescence visualization, we observed large groups of multinucleated, fused, cells with numerous nuclei (blue) in proximity without distinguishable cell membranes (red) separating nuclei in the forskolin-treated group (Fig. 1D). In contrast, the cells in the vehicle-treated group had nuclei clearly separated by cell membranes (Fig. 1C). In addition, cell migration in the forskolin-treated group was greatly reduced compared with controls (0.332 ± 0.03 AU vs. 1.000 ± 0.10 AU, respectively, P = 0.0001) (Fig. 1E).

Figure 1.

Representative phase contrast imaging of BeWo trophoblasts treated with DMSO as a control for 48 h (A) and of BeWo trophoblasts treated with 25 µM forskolin for 48 h (B). Representative confocal imaging of BeWo trophoblasts treated with DMSO as a control for 48 h (C) and of BeWo trophoblasts treated with 25 µM forskolin for 48 h (D). In C and D, red = cell membrane and blue = nuclei. E: relative migration of BeWos after 48 h of treatment with 25 µM forskolin or DMSO as a control. N = 6 for migration assay groups. Statistical analysis performed using unpaired Student’s t test. Significance is designated by a bar connecting groups (***P < 0.001).

We next performed digital droplet PCR (ddPCR) on all four syndecan family members to observe changes in ECM protein expression after induced syncytialization. There was a significant increase in the expression of syndecan-1 mRNA (SDC-1) in the forskolin-treated group compared with controls, as measured by ddPCR (5,700 ± 493 copies/ng RNA vs. 2,255 ± 119 copies/ng RNA, respectively, P < 0.0001) (Fig. 2A). There was also a significant increase in syndecan-2 mRNA (SDC-2) expression in the forskolin-treated group compared with controls (49.78 ± 4.29 copies/ng RNA vs. 15.16 ± 0.722 copies/ng RNA, respectively, P < 0.0001) (Fig. 2B) and in syndecan-3 mRNA (SDC-3) expression in the forskolin-treated group compared with controls (239.8 ± 18.1 copies/ng RNA vs. 124.4 ± 5.42 copies/ng RNA, respectively, P = 0.0001) (Fig. 2C). There was a significant decrease in the expression of syndecan-4 mRNA (SDC-4) in the forskolin-treated group compared with controls (1,206 ± 106 copies/ng RNA vs. 2,881 ± 137 copies/ng RNA, respectively, P < 0.0001) (Fig. 2D).

Figure 2.

Digital droplet PCR of syndecan-1 (SDC-1) (A), syndecan-2 (SDC-2) (B), syndecan-3 (SDC-3) (C), and syndecan-4 (SDC-4) (D) expression levels, expressed as copies of gene per nanogram (ng) of RNA, in BeWos treated with DMSO or 25 µM forskolin for 48 h. N = 6 for all groups. Statistical analysis performed using unpaired Student’s t test. Significance is designated by a bar connecting groups (***P < 0.001).

Syncytialization Alters ECM Composition

We next measured cell lysate levels of syndecan-1 and syndecan-4, as they were the most abundantly expressed syndecans. As determined by ELISA, cell lysate syndecan-1 was significantly higher in the forskolin-treated group compared with controls (24.7 ± 2.0 pg/µg vs. 13.4 ± 0.5 pg/µg, respectively, P = 0.0003) (Fig. 3A). Cell lysate syndecan-4 levels were significantly decreased in the forskolin-treated group compared with controls, as determined by ELISA (2.41 ± 0.03 pg/µg vs. 7.19 ± 0.13 pg/µg, respectively, P < 0.0001) (Fig. 3B). Immunofluorescence imaging for syndecan-1 in cells treated with vehicle control (Fig. 3C) and forskolin (Fig. 3D) showed no visual differences in localization.

Figure 3.

Syndecan-1 protein levels in the cell lysates of BeWos treated with DMSO or 25 µM forskolin for 48 h (A), as measured by ELISA. Syndecan-4 protein levels in the cell lysates of BeWos treated with DMSO or 25 µM forskolin for 48 h (B), as measured by ELISA. Immunofluorescent imaging for syndecan-1 (green) in cells treated with DMSO (C) and forskolin (D). Lysate protein levels were normalized to lysate total protein levels and are expressed in picograms (pg) of syndecan-1 or syndecan-4 per microgram (µg) of total protein. N = 6 for all groups. Statistical analysis performed using unpaired Student’s t test. Significance is designated by a bar connecting groups ( ***P < 0.001).

Syncytialization Alters Trophoblast Barrier Function

Barrier function was assessed using an ECIS-z instrument (Applied BioPhysics) measuring impedance at the 4,000 Hz AC frequency, which provides measurements of the ECM’s barrier function. Forskolin-treated BeWos had a noticeable increase in barrier function, as measured by impedance, compared with controls, indicated here by area under the curve (AUC) (29.1 ± 0.2 AU vs. 10.3 ± 0.1 AU, respectively, P < 0.0001) (Fig. 4, A and B). Although ECIS measurements are excellent for general quantitative barrier function measurements, we wished to determine if specific physiological aspects of barrier function were also altered. Therefore, we measured monocyte adherence to BeWos and syncytialized BeWos to determine if this specific barrier property was altered by syncytialization. THP-1 adherence to forskolin-treated BeWos was also reduced compared with controls (31.2 ± 7.9% vs. 55.1 ± 4.8%, respectively, P = 0.0329) (Fig. 4C).

Figure 4.

A: normalized impedance of BeWos treated with DMSO or 25 µM forskolin for 48 h, as well as media alone. Impedance was continuously measured at 4,000 Hz frequency. N = 3 for each group. B: area under the curve (AUC) of 48 h of impedance measurement. N = 3 for each group. C: THP-1 adhesion to BeWos treated with DMSO or 25 µM forskolin for 48 h, expressed as the percentage of total THP-1s added to each well. N = 5 for each group. Statistical analysis performed using unpaired Student’s t test. Significance is designated by a bar connecting groups (*P < 0.05, **P < 0.01).

Heparanase and MMP-9 Alter ECM Composition

Next, two common ECM-degrading enzymes found in the placenta, HPSE and MMP-9, were administered, and their effects on syncytialized trophoblasts’ ECM composition were measured. Compared with the forskolin-treated group, the forskolin + HPSE and MMP-9 group had a small but significant reduction in syndecan-1 protein levels in the cell lysate, as determined by ELISA (10.7 ± 0.47 pg/µg vs. 12.8 ± 0.33 pg/µg, respectively, P = 0.0097) (Fig. 5A). There were no significant differences between the forskolin-treated group and groups treated with HPSE or MMP-9 alone. In addition, there was a significant increase in media syndecan-1 levels in the forskolin + HPSE and MMP-9 group compared with the forskolin-treated group (1,066 ± 55.7 pg/mL vs. 766.6 ± 122 pg/mL, respectively, P = 0.0267) (Fig. 5B). There was also a significant increase in media syndecan-1 levels in the forskolin + HPSE group compared with the forskolin-treated group (1,257 ± 59.2 pg/mL vs. 766.6 ± 122 pg/mL, respectively, P = 0.0005) (Fig. 5B). Cell lysate levels of syndecan-4 were significantly increased in the forskolin + MMP-9 group (17.2 ± 0.74 pg/µg, P < 0.0001) and the forskolin + HPSE and MMP-9 group (13.7 ± 0.40 pg/µg, P < 0.0001) compared with the forskolin-treated group (2.57 ± 0.04 pg/µg) (Fig. 5C). There was no significant difference between the cell lysate levels of syndecan-4 in the forskolin + HPSE group compared with the forskolin-treated group. Media levels of syndecan-4 were significantly increased in the forskolin + HPSE-treated group compared with the forskolin-treated group (244 ± 12.7 pg/mL vs. 192 ± 8.4 pg/mL, respectively, P = 0.0038) (Fig. 5D).

Figure 5.

Syndecan-1 protein levels in the cell lysates (A) and media (B) of BeWos first treated with 25 µM forskolin for 48 h and then treated with forskolin, forskolin (For) + heparanase (HPSE), For + MMP-9, or For + HSPE + MMP-9 (Combo) for 24 h, as measured by ELISA. Syndecan-4 protein levels in the cell lysates (C) and media (D) of BeWos first treated with 25 µM forskolin for 48 h and then treated with forskolin, forskolin (For) + heparanase (HPSE), For + MMP-9, or For + HSPE + MMP-9 (Combo), as measured by ELISA. Lysate protein levels were normalized to lysate total protein levels and are expressed in picograms (pg) of syndecan-1 or syndecan-4 per microgram (µg) of total protein. N = 6 for all groups. Statistical analysis performed using one-way ANOVA with post hoc Dunnett’s multiple-comparisons test. Significance is designated by a bar connecting groups (*P < 0.05, **P < 0.01, ***P < 0.001). MMP-9, matrix metalloproteinase-9.

ECM-Degrading Enzymes Alter Trophoblast Barrier Function

Finally, we wished to examine the effects of syncytialization and ECM-degrading enzymes on barrier function. BeWos were cultured in ECIS arrays and treated with either forskolin or vehicle control for 48 h, after which three of the forskolin groups were also treated with either heparanase, MMP-9, or a combination of the two for an additional 24 h. Figure 6A displays data from syncytialized BeWos that were treated with either forskolin + HPSE, forskolin + MMP-9, forskolin + a combination of the two, or only forskolin for a control. All BeWos were syncytialized with forskolin for 48 h before addition of the abovementioned treatments, but data shown begin at treatment time. Barrier function, as measured by the AUC of impedance plotted against time, decreased markedly in the MMP-9- (15.2 ± 0.03 AU; P < 0.0001) and combination-treated (17.0 ± 0.09 AU; P < 0.0001) groups compared with the forskolin control group (19.2 ± 0.04 AU) (Fig. 6B). THP-1 adherence was not statistically different between any treatment groups compared with the forskolin control group (Fig. 6C).

Figure 6.

A: normalized impedance of BeWos first treated with 25 µM forskolin for 48 h and then treated with forskolin, forskolin (For) + heparanase (HPSE), For + MMP-9, or For + HSPE + MMP-9 (Combo) for 24 h, as well as media alone. Continuous impedance measurements throughout the 24 h of treatment with enzymes is shown, which was measured at 4,000 Hz frequency. N = 3 for each group. B: area under the curve (AUC) of 24 h of impedance measurement. N = 3 for each group. C: THP-1 adhesion to BeWos first treated with 25 µM forskolin for 48 h and then treated with forskolin, forskolin (For) + heparanase (HPSE), For + MMP-9, or For + HSPE + MMP-9 (Combo) for 1 h, expressed as the percentage of total THP-1s added to each well. MMP-9, matrix metalloproteinase-9. Significance is designated by a bar connecting groups (**P < 0.01).

DISCUSSION

The placental barrier is a necessity for fetal health and growth. Placental barrier dysfunction has been described in multiple diseases of pregnancy and has been associated with maternal and fetal morbidity and mortality (4). Because of the detrimental effects of placental barrier dysfunction, basic research of the placental barrier is needed to make progress toward translational medicinal approaches. The ability of the placental barrier to exclude unwanted immune cells, hormones, drugs, and chemicals is derived primarily from syncytiotrophoblasts and their ECM. Therefore, proper characterization of an in vitro model of the placental ECM is necessary to better understand its composition and function, as well as any alterations during typical disturbances.

BeWos are an immortalized trophoblast-like choriocarcinoma cell line that is often used as an in vitro model of placental trophoblasts. However, with regard to modeling syncytiotrophoblasts, BeWos do not spontaneously syncytialize at a fast rate, and therefore do not ordinarily provide a physiologically relevant model of this cell type. Despite this, BeWos can and have been induced to syncytialize with the cAMP activator forskolin and have been described as one of the best trophoblast models for syncytial fusion and cellular barriers (15, 17). First, we confirmed syncytialization of BeWos after 48 h of treatment with 25 µM forskolin by phase contrast and confocal fluorescent microscopy, as well as through observation of an impaired migratory ability (indicating that they had not differentiated into invasive migratory cytotrophoblasts).

To the best of our knowledge, the ECM of syncytialized BeWos has not been characterized to date; thus, we aimed to evaluate changes that syncytialization exerts on key ECM components. We did observe a shift in the expression profile of syndecans in syncytialized BeWos compared with nonsyncytialized BeWos. Syndecan-4, which under control conditions was the predominant syndecan both in terms of RNA expression and protein levels, was drastically downregulated after syncytialization, whereas syndecan-1 oppositely increased. Visual localization of syndecan-1 did not change with syncytialization. Syndecan-1 has been described as a typical epithelial, endothelial, and fibroblast syndecan, whereas syndecan-4 is expressed in not only epithelia and fibroblasts but also neuronal cells, endothelial cells, and other cell types (18). The shift in syndecan expression to the primarily epithelial type agrees with the typical characterization of syncytiotrophoblasts as epithelial-like cells (19). Furthermore, in humans, syndecan-1 is restricted to syncytiotrophoblasts and is not detectable on the surface of cytotrophoblasts (20). Syndecan-2 is primarily found in endothelial and neuronal cells, and syndecan-3 is concentrated mostly in neuronal cells (18). Although their expression was increased during syncytialization, neither syndecan-2 nor syndecan-3 reached substantial levels of expression, as their mRNA count remained hundreds of times less than syndecan-1 and -4. Little is known about the presence and/or roles of syndecan-2 and syndecan-3 with regard to trophoblasts. However, it has been suggested that syndecan-2 plays a role in the remodeling of the syncytiotrophoblast ECM during chorionic villous tree development, which offers an explanation for its increased expression in our own findings (21). Syndecan-3 has been observed on syncytiotrophoblasts by one group, although no specific function was attributed (22). The reason for an increase in syndecan-3 expression in our syncytialized trophoblasts is unknown, although it could be related to the generalized actions of syndecans as co-receptors for growth factors (23). Our evidence, taken with consideration of the literature, strongly suggests that syncytialization of BeWos using forskolin produces a physiologically accurate ECM.

We observed a sustained increase in impedance in syncytialized BeWos at the 4,000 Hz range as measured by ECIS array. According to manufacturer details, low frequencies such as 4,000 Hz are indicative of ECM measurements, as at that frequency, current cannot pass through cell membranes but rather passes around and between cells through their ECM. This frequency is indicated by the manufacturer for the use of assessing barrier function and has been done so successfully by other groups (24, 25). In addition, all groups were seeded at confluency to avoid differences in the number of cells adhered to the wells. This increase in impedance can be described as an increase in the resistance of the ECM to electrical flow. ECIS measurements of impedance are often used as proxies to represent barrier function (26–28) and are excellent hypothesis-generating experiments. Our PCR and ECIS measurements indicate that there is an increase in ECM synthesis in syncytialized BeWos compared with nonsyncytialized BeWos, which would suggest an improved barrier function or ability to exclude unwanted cells or molecules from their cell surface. Further supporting this, we observed a decreased ability of THP-1 monocytes to adhere to the surface of syncytialized BeWos. Because the ability of the syncytial layer to exclude immune cells from adhesion and infiltration is a primary function of the placental barrier, we are confident that these syncytialized BeWos are behaving in a physiologically relevant manner.

Now that we had a physiological model of the syncytiotrophoblast ECM as a placental barrier, we wished to observe alterations in its composition and function in response to typical disturbances of placental dysfunction. HPSE is an ECM-degrading enzyme that is highly expressed in placentas during pregnancy and has been implicated in placental dysfunction (29). Likewise, MMP-9 is very highly expressed in pregnancy during trophoblast invasion and vascular remodeling (30). Interestingly, the MMP-9 cleavage sites on syndecan-1 and syndecan-4 are below their heparan sulfate side chains and become more accessible to MMP-9 after side-chain cleavage by HPSE (31). Because of their reputations, we wished to evaluate HPSE’s and MMP-9’s effects on the syncytiotrophoblast ECM composition in vitro and also to understand if either of these enzymes impairs placental barrier function. Only combination treatment with HPSE and MMP-9 significantly reduced cellular levels of syndecan-1, and the magnitude of the decrease was small. Syndecan-1 has an extracellular cleavage site on a portion of the protein that resides near the cell membrane where MMP-9 can act. This site is normally masked by the large heparan sulfate side chains that protrude from the core of the protein. However, with the enzymatic activity of HPSE, heparan sulfate side chains are cleaved and the underlying MMP-9 cleavage site is exposed. Because of the structural location of the MMP-9 cleavage site, it reasons that the actions of HPSE would either be necessary or very inducing of subsequent MMP-9 cleavage. Interestingly, syndecan-4 was increased in the MMP-9- and combination-treated groups to levels exceeding syndecan-1 levels, although the reason for this is unknown. Therefore, HPSE and MMP-9, which have been implicated in placental pathologies, may exert negative effects on placental barrier dysfunction through their actions on syndecan-1. Increases in syndecan-4 levels could be a compensatory response to loss of syndecan-1 or a side effect of MMP-9 deactivating another enzyme that regularly cleaves syndecan-4.

Although MMP-9 and combination treatment with HPSE and MMP-9 reduced impedance of syncytialized BeWos considerably compared with syncytialized controls, HPSE treatment alone did not have such an effect. Although there was a decrease in syndecan-1 in the combination-treated group that could account for the reduced impedance, this does not account for the decrease in impedance in the MMP-9-treated group. Syndecan-4 on the other hand was increased in both the MMP-9- and combination-treated groups that had decreased impedance levels. These data suggest that syndecan-4 does not contribute significantly to the barrier function of syncytialized BeWos. Because impedance correlates with barrier function, we expected THP-1 adherence to MMP-9- and combination-treated syncytialized BeWos to be increased, but that was not the case. THP-1 adherence was unaltered in all treatment groups compared with untreated syncytialized BeWos, suggesting a disconnect, in this case, between ECIS measurements and immune cell adhesion. These data suggest that syndecan-4 likely does not contribute mechanistically to the exclusion of monocytes from the syncytiotrophoblast surface under these conditions. Syndecan-1 could potentially be responsible for the exclusion of monocytes from the cell surface because cellular levels, despite being decreased during combination treatment, remained relatively high in all groups. Syndecan-1 is not only notably larger than syndecan-4, which means it protrudes farther from the cell surface, and it also contains chondroitin chains not found in syndecan-4 (23). These structural differences may mean syndecan-1 promotes more steric hindrance of the molecular interactions needed for monocytes to bind to syncytialized BeWos, as adhesion molecules are significantly smaller and do not extend as far from the cell surface.

Syncytialization of BeWos produces a physiological ECM that can be utilized to study placental barrier function. There is a shift in syndecan expression toward what one would expect from an epithelial syncytial layer of cells. In addition, ECIS measurements and the improved exclusion of THP-1 monocytes from syncytialized BeWos indicate the production of a functional ECM barrier. HPSE and MMP-9 altered cellular syndecan levels but did not change THP-1 monocyte adhesion. However, because impedance levels were decreased so drastically, there are likely other properties of the syncytiotrophoblast barrier that are affected by HPSE and MMP-9. Here, we have shown that syncytialized BeWo trophoblasts are a more physiologically accurate model than nonsyncytialized BeWos and should be used in future research investigating the maternal-fetal interface.

GRANTS

This work was supported by NIH Grants P01HL51971, P20GM104357, T32HL105324, and R01HL137791.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.H.M. and E.M.G. conceived and designed research; K.H.M., H.A.M., and H.C. performed experiments; K.H.M. analyzed data; K.H.M. interpreted results of experiments; K.H.M. prepared figures; K.H.M. drafted manuscript; K.H.M., H.A.M., and E.M.G. edited and revised manuscript; K.H.M., H.A.M., H.C., and E.M.G. approved final version of manuscript.