Platelet-derived factors dysregulate placental sphingosine-1-phosphate receptor 2 in human trophoblasts

,


S
phingosine-1-phosphate (S1P) is an essential and bioactive sphingolipid associated with the regulation of various cellular functions such as differentiation, proliferation, apoptosis, cell motility and calcium signalling (Bryan and Del Poeta, 2018;Maceyka et al., 2012).Moreover, S1P regulates vascular development during embryogenesis and has differential effects on motility, contraction, angiogenesis and vascular permeability, all of which are important to both maintenance and termination of pregnancy (Spiegel and Milstien, 2003;Yamamoto et al., 2010).The structure of S1P consists of a polar head group, a phosphate group and a longchain sphingoid base backbone, the sphingosine group (Ksi ą _ zek et al., 2015).It originates from phosphorylation of sphingosine by either sphingosine kinase 1 (SPHK1) or sphingosine kinase 2 (SPHK2).The concentration of S1P is controlled by two specific S1P phosphatases or an S1P lyase, which either remove the phosphate group or irreversibly degrade S1P (Maceyka et al., 2012;Tukijan et al., 2018).The major source of plasma S1P are red blood cells, vascular endothelial cells and activated platelets, making S1P a bloodborne lipid mediator, found in association with albumin and lipoproteins such as highdensity lipoprotein (HDL) (Patanapirunhakit et al., 2021).Although studies have shown that erythrocytes are the major source of S1P in blood plasma, platelets are also found to generate and store large quantities of S1P.Recent studies suggest that S1P is essential in thrombus formation due to its increased release during blood clot formation.The augmentation of S1P plasma concentrations by activated platelets is also assumed to play an important role in the repair of injured endothelial vessels (Tukijan et al., 2018).S1P signalling is mediated by five different G-proteincoupled receptors (S1PR1À5), which were originally referred to as Edg for endothelial differentiation gene (S1PR1/Edg-1, S1PR2/ Edg-5, S1PR3/Edg-3, S1PR4/Edg-6 and S1PR5/Edg-8) (Kerage et al., 2014).The expression of these receptors is tissue-and cell type-specific.While S1PR1 is ubiquitously expressed, S1PR2 and S1PR3 are more restricted in expression and are mainly found in the gastrointestinal, vascular and central nervous systems.S1PR4 is prevailingly expressed in lymphoid tissues, whereas S1PR5 is primarily found in the nervous system (Takabe et al., 2008).
Various cell types express different combinations of these receptors and the distribution of the receptors on different cell types and the coupling of receptors to different G-subunits allows S1P to influence numerous pathways in different ways.Because the commercial availability of antibodies for S1PR4 and S1PR5 is limited and their expression has been reported to be restricted to lymphoid tissue and the central nervous system, the main focus in this study was on S1PR1À3 (Obinata and Hla, 2019).
In fact, S1P signalling is increasingly becoming the focus of pregnancy research (see Fakhr et al., 2021, for a comprehensive review).Its functions have an impact on female reproductive organs before and during pregnancy, including progesterone production in the corpus luteum, differentiation of the endometrium (i.e.decidualization) and placenta development.In human pregnancy, trophoblast invasion into the maternal decidua, as well as fusion of mononucleated cytotrophoblasts into the multinucleated syncytiotrophoblast, are mandatory steps in placentation, both of which are described to be regulated by S1P.Moreover, S1P is suggested to regulate vascular function and immune cell trafficking at the maternalÀfetal interface and to play a powerful role in establishing an adequate uteroplacental microvasculature and development of immune tolerance of the mother towards the fetus (Nagamatsu et al., 2014).However, an increasing body of evidence suggests that S1P signalling pathways are dysregulated in pregnancy pathologies such as pre-eclampsia and intrauterine growth restriction (IUGR).Accordingly, circulating S1P plasma concentrations were significantly higher in women diagnosed with pre-eclampsia when compared with controls (Charkiewicz et al., 2017).Besides, placental S1P receptor subtypes are dysregulated in placental tissues from pregnancies complicated by severe pre-eclampsia (Dobierzewska et al., 2016;Li et al., 2014).Both aberrant trophoblast differentiation and turbulent blood flow (high-velocity jets and vortices), combined with elevated blood pressure in the intervillous space and increased wall shear stress at the villous syncytiotrophoblast surface, are suggested to be involved in the pathophysiology of pre-eclampsia (Drewlo et al., 2020;Roth et al., 2017).Notably, shear stress has been shown to induce S1PR1 expression in endothelial cells (Aoki et al., 2007;Josipovic et al., 2018), tempting speculation on the impact of onset of uteroplacental blood perfusion on placental S1P receptor subtype expression.
The aim of the present study was to analyse the expression dynamics of placental S1PR across different stages of gestation and to determine their expression in various placental cell types.Because platelets might be the first amongst maternal blood cells able to enter the early intervillous space through interstices of extravillous trophoblast columns (Guettler et al., 2021;Sato et al., 2005), this study investigated whether or not the S1PR subtype repertoire is dysregulated in human trophoblasts in response to plateletderived factors.Also tested was whether onset of uteroplacental perfusion with maternal blood could affect the placental S1PR repertoire by analysing the receptor expression in regard to different flow rates and oxygen concentrations.

Human placental tissue samples
The study was approved by the ethical committee of the Medical University of Graz (31-019 ex 18/19, approved 2 December 2022).First trimester placental tissues (n = 10) were obtained with written informed consent from women undergoing legal elective surgical pregnancy terminations between weeks 6 and 12 of gestation.Pre-term (n = 9) and term (n = 10) placental tissue was collected with written informed consent.Pre-term placental tissues were obtained mostly by Caesarean section because of placenta praevia with or without velamentous insertion of the cord and vaginal bleeding or cervical insufficiency from women who had no clinical evidence of infection.Villous tissue from all placentas was fixed in formalin and embedded in paraffin.Furthermore, additional tissue pieces were thoroughly rinsed in phosphate-buffered saline and dissected into small pieces of approximately 5 £ 5 mm, before they were snap-frozen in liquid nitrogen and stored at À80°C until further analyses.Characteristics of the study population are shown in TABLE 1.

Culture of BeWo cell line
Trophoblast cell line BeWo was purchased from the European Collection of Cell Cultures (ECACC) and was cultured in DMEM/F12 (1:1, Gibco TM , Life Technologies, Paisley, UK) supplemented with 10% FCS (Gibco), 0.1 U/ml penicillin and 0.1 mg/ml streptomycin (Gibco) and 1% (v/v) L-glutamine (Gibco; 200 mmol/l 100X) in a humidified atmosphere of 5% CO 2 at 37°C (Forstner et al., 2020).Cells between passage 10 and 20 were used for in-vitro experiments.For differentiation, 2 £ 10 5 cells per well were plated in 12well dishes (Nunc TM Lab-Tek TM , Thermo Scientific, NY, USA) 1 day prior to the start of the experiment in the above-described culture medium.Next day, culture medium was exchanged with medium including 20 mmol/l forskolin (Bio-Techne, Tocris, Abingdon, UK) and cells were incubated for another 48 h.Dimethylsulphoxide (DMSO) served as solvent control for forskolin and was applied at a final concentration of 0.1% (v/ v).Thereafter, cells were incubated under hypoxic conditions with 5% CO 2 and either 2.5%, 12% or 21% O 2 concentrations at 37°C for 24 h in a hypoxic workstation (BioSpherix; Redfield, NY, USA).After culture, cell lysates and supernatants were collected for further analysis.
Fluidic flow culture of BeWo cells BeWo cells were seeded on coverslips in a 12-well dish (Nunc Lab-Tek, Thermo Scientific) with a density of 2 £ 10 5 cells per well in DMEM/F-12 (1:1, Gibco) 1 day prior to the start of the experiment.Next day, cells were treated with forskolin (Bio-Techne, Tocris) with a final concentration of 20 mmol/l for 48 h.Thereafter, cells on coverslips were transferred to flow chambers (Kirkstall; QV500), which were connected to a tubing system (Kirkstall; 2 £ 1/16' diameter 22 cm length, 1 £ 3/ 32' diameter 22 cm length) and a 30 ml reservoir bottle.The flow system was placed in a TEB500 flow bioreactor (Ebers Medical Technology, Zaragoza, Spain).Each flow cycle included two chambers, which were perfused with 12 ml fresh media from the reservoir.Pumps were set to a perfusion rate of 1 ml/min and 3 ml/ min and culture conditions were set to 2.5% O 2 and 5% CO 2 at 37°C for an experiment duration of 24 h.For static controls, forskolin-differentiated BeWo cells, grown on coverslips, were transferred to 6-well dishes (Nunc Lab-Tek, Thermo Scientific) containing 6 ml fresh culture media.Static controls were cultured in parallel to flow cultures in the TEB500 flow bioreactor.The experiments were performed four times with different cell passages.

Co-incubation of platelet-derived factors with BeWo cells
BeWo cells were seeded in a 12-well dish (Nunc Lab-Tek, Thermo Scientific) with a density of 2 £ 10 5 cells per well in DMEM/ F-12 (1:1, Gibco) 1 day prior to the start of the experimental.Next day, cells were treated with forskolin (Bio-Techne; Tocris) to undergo differentiation into multinucleated syncytia at a final concentration of 20 mmol/l for 48 h.After differentiation three different treatments were started for 24 h.For treatment with platelet-derived factors, platelets were first isolated from whole blood of pregnant donors (see Platelet isolation below); 2.25 £ 10 5 to 3.13 £ 10 5 platelets per ml in a total of 500 ml were transferred into a polycarbonate cell culture insert (Nunc Lab-Tek, Thermo Scientific) located above the well containing the differentiated BeWo cells.The insert prevented direct cell contact of platelets and trophoblasts, as platelets also express the S1P receptors and therefore the results would have been compromised by direct contact.Platelets in the insert were activated with 1 IU/ml thrombin directly administered into the culture medium.Control cells cultured in transwell dishes were subjected to either control medium DMEM/F-12, or control medium supplemented with 1 IU/ml thrombin (Sigma-Aldrich, St. Louis, MO, USA).After 24 h culture, cell lysates and supernatants were collected for further analysis.

Isolation of placental primary cells
Primary endothelial cells were isolated from term placentas of healthy pregnancies as previously described by Lang et al. (2008).In brief, arterial and venous chorionic blood vessels were resected and washed with HBSS (Gibco).Arterial endothelial cells (ECA) and venous endothelial cells (ECV) were isolated by separate perfusion of chorionic arteries and veins with collagenase/dispase (0.1/ 0.8 U/ml, Roche, Vienna, Austria) dissolved in HBSS, pre-warmed to 37°C.The obtained cell suspension was centrifuged and the cell pellet was resuspended in EGM-MV medium (Lonza, Verviers, Belgium) and plated on culture plates.ECA were incubated at 37°C with 12% O 2 and 5% CO 2 .ECV were incubated at 37°C, 21% O 2 and 5% CO 2 .
Primary trophoblasts were isolated from term placentas of healthy pregnancies as previously described by Loegl et al. (2017).In brief, placental villous tissue was minced and digested with trypsin/dispase/DNase (Gibco/Roche/Sigma) solution for 90 min.Cell suspension was centrifuged on a Percoll gradient (Sigma) at 4°C for 30 min at 300g.Trophoblast-enriched layers were purified by immunodepletion of contaminating cells using beads conjugated to MCA-81 antibody (Serotec, Puchheim, Germany) against HLA-A, B and C. Trophoblasts were then seeded in each well of a 6-well plate (Nunc Lab-Tek, Thermo Scientific) in 2 ml DMEM (Gibco) containing 10% FCS (Gibco), 0.1 U/ml penicillin and 0.1 mg/ml streptomycin (Gibco) at 37°C and 21% O 2 and 5% CO 2 .

Platelet isolation
Citrated whole blood (5 £ 3.Only cells with more than 300 detected genes and less than 20% total mitochondrial gene expression were used for further analysis (Vento-Tormo et al., 2018).Further bioinformatical analysis was done using R package Seurat (Butler et al., 2018) and RStudio to visualize gene expression data in a dot plot.

Transmission electron microscopy
The membrane was cut from the insert after termination of the cell culture Next, mRNA as well as protein expression levels of the three receptors in primary placental endothelial cells and trophoblasts were analysed (FIGURE 3).S1PR1 was significantly higher expressed in arterial as well as venous endothelial cells compared with primary trophoblasts (P < 0.0001) (FIGURE 3AÀC).In contrast, S1PR2 expression was significantly increased in primary trophoblast cells compared with endothelial cells (P = 0.0262) (FIGURE 3DÀF).S1PR3 showed somewhat adverse expression levels between qPCR, where it was expressed in all three cell types, and immunoblotting, where it was present in trophoblast cells and to some extent in endothelial cells, especially venous endothelial cells (P = 0.0182, P = 0.0064, respectively) (FIGURE 3GÀI).

S1PR2 and S1PR3 are dysregulated by platelet-derived factors
The next aim was to determine potential effects of fluidic flow on S1PR expression in differentiated BeWo cells, mimicking placental syncytiotrophoblasts exposed to maternal blood flow.For this purpose, BeWo cells were first stimulated with forskolin to undergo differentiation and syncytialization, and were afterwards cultured under fluidic flow.qPCR analysis showed no influence on the expression level of S1PR2 and S1PR3 in response to flow rates of 1 ml/min and 3 ml/min, respectively (FIGURE 4A and 4B).
To determine whether the physiological rise in intervillous oxygen tension during the course of pregnancy may cause the observed longitudinal expression changes of placental S1P receptors, the next test was whether different in-vitro oxygen concentrations can affect S1PR in differentiated BeWo cells.During the first trimester oxygen concentrations range between 2 and 3% in the intervillous space and increase during the ongoing development to about 8À10% during the second trimester.In order to determine oxygen effects, a rather broad range was tested, including 2.5% O 2 during the first trimester, 12% O 2 to mirror the rise in oxygen concentration and lastly 21% O 2 as a control to depict an atmospheric state.S1PR2 did not show significant differences in expression at 2.5%, 12% and 21% oxygen (FIGURE 4c); S1PR3 on the other hand showed no significant change from 2.5% to 12% O 2 but from there a significant decrease when cultured under 21% O 2 (P = 0.0240) (FIGURE 4D).
As platelets are the first maternal blood cells that establish contact to fetal-derived trophoblast cells, the effect of coincubation of platelets and trophoblast cells was examined.As platelets express various S1P receptors (Supplementary Figure 1), experiments with direct coincubation of platelets and trophoblast cells would be difficult to analyse because expression of the cellular origin of a certain receptor cannot be verified.To exclude this possibility, co-incubation experiments were conducted with an insert to keep the trophoblast samples free of platelet RNA (FIGURE 4EÀG).However, the platelet-derived factors may well diffuse through the 0.4 mm membrane of the insert and act on underlying BeWo cells.qPCR analysis after co-cultures revealed that S1PR2 was significantly down-regulated in BeWo cells upon treatment with platelet-derived factors when compared with controls (P = 0.0055) (FIGURE 4H).It is worth noting that thrombin itself, used as an agonist to activate platelets, had an impact on the S1PR2 expression level (P = 0.0461).Contrary effects, although not significant, could be seen in the expression of S1PR3,  which increased in the presence of both thrombin alone and platelet-derived factors (FIGURE 4I).

DISCUSSION
This study demonstrates the spatiotemporal expression of different S1PR in the human placenta across gestation.While placental S1PR subtypes 1 and 3 increase with gestational age, S1PR2 is predominantly expressed in the first trimester villous trophoblast compartment and sharply declines towards term.In human trophoblasts, S1P action is suggested to inhibit differentiation through G(i)-coupled S1P receptor interaction, leading to the inhibition of adenylate cyclase and reduced production of intracellular cAMP (Johnstone et al., 2005;Singh et al., 2012).However, whether or not this inhibitory effect is mediated through a particular S1PR, e.g. through the prevailing S1PR2 in first trimester placenta, still remains to be determined.Of note, S1P acts as a potent inhibitor of extravillous trophoblast migration, predominantly through S1PR2/Ga12/13 and Rho activation (Westwood et al., 2017).Moreover, S1PR2 is involved in S1P-induced expression and secretion of IL6 in BeWo cells (Goyal et al., 2013).Based on these previous and our own studies, it is tempting to speculate that the predominant S1PR2 expression in human first trimester placenta is involved in regulating trophoblast differentiation and migration/invasion processes, both essential for early placentation.Later on in pregnancy, when the initial phase of rapid cell mass expansion and invasion is replaced by metabolic processes, expression of the placental S1PR repertoire may shift to meet these demands.Significantly reduced levels of S1PR1 and S1PR3 in non-severe as well as severe preeclampsia cases are suggested to be responsible for impaired vasculogenesis/ angiogenesis observed during this syndrome (Dobierzewska et al., 2016), and may be a sign for an insufficient shift in the placental S1PR repertoire.
This study confirms the expression of S1PR1 in endothelial cells, which is in agreement with previous studies showing the expression of this receptor subtype in the endothelium of various other organs (Panetti, 2002;Wang et al., 1999) and to be strongly induced by shear stress in the vascular network of flow-positive regions (Josipovic et al., 2018;Jung et al., 2012).S1PR2 and S1PR3 were mainly found in trophoblasts, which is in accordance with the speculations that these two receptor subtypes play a significant role in trophoblast development and remodelling of spiral arteries.This assumption is substantiated by recent studies highlighting the involvement of S1PR2 in trophoblast mobility (Liao et al., 2022).
S1PR are increasingly acknowledged for their role as cellular mechanosensors that convert various physical signals, such as laminar and turbulent flow, into biochemical signals, coupling them with downstream signalling pathways (Hu et al., 2022).In vascular endothelial cells, S1PR1 is described as one of the predominant mechanosensors, which is up-regulated by laminar shear stress itself, as shown in HUVEC cultured under laminar shear stress of 20 dyne/cm 2 (Aoki et al., 2007;Josipovic et al., 2018).During human gestation, uteroplacental blood flow is fully established with transition to the second trimester, causing fluid shear stress that is produced by haemodynamic forces across different trophoblast subtypes (Brugger et al., 2020).Whether onset of maternal blood flow is triggering the observed (up) regulation of placental S1PR towards term remains speculative.Our data from flow culture experiments rather argue against this speculation, because neither S1PR2 nor S1PR3 were up-regulated in differentiated BeWo cells in response to fluid flow.
Low oxygen concentrations of about 2À3% in the early developing placenta are a special attribute described for the first couple of weeks during pregnancy, while with ongoing placentation and remodelling of the spiral arteries the oxygen concentration increases to about 8À10% (Zhao et al., 2021).Previous studies have shown a significant up-regulation of SPHK1 under low oxygen tension, but a decrease in expression levels of receptors S1PR2 and S1PR3 in pre-eclamptic placentas (Dobierzewska et al., 2016).Because this study found no significant difference in the expression level of S1PR2 or S1PR3 when cells were cultured under 2.5% or 12% oxygen, the tremendous difference in the expression pattern of the receptors from first trimester to term is most likely not due to the rise in oxygen concentrations over gestation.The controls incubated at atmospheric conditions (21% oxygen) showed a slight decrease in the expression level of S1PR2 and even a significant difference in the expression of S1PR3, but it can be assumed that this condition is considered rather hyperoxic and therefore not physiologically present in the human placenta (Burton et al., 2021).
During uncomplicated pregnancy, the count of maternal platelets decreases gradually over time of gestation, reaching an overall decrease of about 10% by the end of pregnancy (Michelson et al., 2019).This decrease in platelet count can be explained mainly by haemodilution, and to some extent by accelerated sequestration and consumption of platelets in the uteroplacental circulation.Increased mean platelet volume (MPV), as a measure of platelet activation, has been associated with inflammatory pregnancy pathologies, including pre-eclampsia and IUGR (Guettler et al., 2022;Moser et al., 2019).Before haemochorial placentation is fully established, extravillous trophoblasts form plugs within the lumen of spiral arteries to prevent maternal blood cells from entering the intervillous space.However, by the middle of the first trimester these trophoblast plugs become loosely cohesive, forming narrow capillary-shaped channels (Roberts et al., 2017).We and others have found platelets within trophoblast columns and assume that they can enter the channels due to their small diameter of 2À3 mm (Forstner et al., 2020;Sato et al., 2005).Electron microscopy confirmed that these platelets are in an activated state, tempting speculation as to whether released factors can impact the expression of the S1P receptors (Guettler et al., 2021).The extensive reports of the involvement of S1P in inflammatory processes and the fact that a link between a dysfunction in the sphingolipid signature and pre-eclampsia has been discussed before, makes it an obligation to investigate the dysregulation of the receptors upon contact with platelet-derived factors (Del Gaudio et al., 2020;Dobierzewska et al., 2016;Spiegel and Milstien, 2011).The approach of incubating platelets from pregnant women in cell culture inserts above trophoblasts without any direct contact provides a perfect model to study the effect of platelet-derived factors without having to deal with the bias that platelets themselves express S1PR.
The effect observed on S1PR2 and S1PR3 subtypes upon treatment with thrombin alone can be drawn back to the fact that thrombin is a versatile enzyme, thereby able to bind to numerous substrates and thus affecting the molecular basis of cells in many ways (Davie and Kulman, 2006).However, because the difference in expression of S1PR2 upon treatment with platelet-derived factors is even higher compared with the difference seen with cells treated with thrombin alone, it can be argued that a considerable part of the effect is solely due to factors released from platelets.Interestingly, platelet-derived factors seem to have contrary effects on the expression levels of S1PR2 and S1PR3.The significant down-regulation of S1PR2 and the tendency of up-regulation of S1PR3 demonstrated here could be speculated to be in concordance with the opening of the spiral arteries and the influx of maternal blood into the intervillous space at the end of the first trimester of pregnancy.Here, one could speculate that an earlier influx of maternal blood, and thus potential activation of maternal platelets and release of their cargo, would dysregulate the expression pattern of the S1PR too soon and therefore could take part in pregnancy complications with insufficient infiltration and migration of extravillous trophoblasts into the maternal decidua.Further research is necessary to evaluate the timeline of S1PR pattern change and the influence this could have on the developing human placenta.Although the underlying mechanisms of S1PR2 and S1PR3 dysregulation are not widely discussed in the literature, it has been shown that S1PR1 expression on the cell surface is in fact down-regulated by its activation of agonists, like small-molecule FTY720-P, HDL-S1P or albumin-S1P (Cartier and Hla, 2019).We suggest a very similar response of S1PR2 due to the fact that it is down-regulated upon excessive release of platelet cargo, which includes S1P.
The lack of data for placental S1PR subtypes 4 and 5 at various stages of pregnancy is a clear limitation of the current study, but may be justified by the fact that accurate detection of endogenous S1PR proteins is per se a challenge and specific antibodies have either been discontinued or simply were not available (Talmont et al., 2019).However, comparative S1PR expression data from first trimester, pre-term and term placenta samples, and furthermore the use of platelets from pregnant women, can be considered a strength of this study.This is the first report on S1PR expression analysis in a trophoblast cell line cultured under fluidic flow in combination with low oxygen conditions.Moreover, to our knowledge this is the first paper to discuss the impact of platelet-derived factors on the expression pattern of the S1P receptors in the human trophoblast.
In summary, this study suggests that the placental S1PR repertoire is differently expressed across gestation.The trophoblastic S1PR expression pattern depends on differentiation status of the cells and may be dysregulated in response to platelet-derived factors, including S1P.

FIGURE 1
FIGURE 1 Differential placental S1PR expression across human gestation.Human placental tissue was subjected to qPCR analysis of three different S1PR subtypes (S1PR1ÀS1PR3) in first trimester (a, n = 10), pre-term (b, n = 9) and term (c, n = 10) samples.Expression levels of S1PR1 (d), S1PR2 (e) and S1PR3 (f) were compared between first trimester, pre-term and term cases using one-way ANOVA.Data are presented as mean DDCq § SEM, with GAPDH/ TBP used as reference.ANOVA = analysis of variance; FT = first trimester; qPCR = quantitative polymerase chain reaction.

FIGURE 2
FIGURE 2 Differential expression of S1PR1, S1PR2 and S1PR3 across different cell types of the human placenta.First trimester placenta (aÀc) and term placenta (dÀf) were stained for antibodies S1PR1 (a and d), S1PR2 (b and e) and S1PR3 (c and f).FetalÀplacental endothelial cells: black arrow, cytotrophoblast: open arrowhead, syncytiotrophoblast: black arrowhead.Negative controls rabbit IgG in term placenta for both antigen retrieval protocols (pH6 in g and pH9 in h).Scale bars represent 50 mm.Single-cell RNA-Seq expression levels of S1P receptor subtypes (i).Published data for first trimester placenta from Vento-Tormo et al. (2018) were analysed.Intensity and colour scale (raw counts) show the gene expression level and dot circumference shows % of cells within a cell type group expressing a gene.DC = dendritic cells; dM = decidual macrophages; dNK = decidual natural killer cells; dS = decidual stromal cells; Endo = endothelial cells; Epi = epithelial glandular cells; EVT = extravillous trophoblast; f = fetal; FB = fibroblasts; HB = Hofbauer cells; ILC = innate lymphocyte cells; L = lymphatic; m = maternal; p = proliferative; SCT = syncytiotrophoblast; VCT = villous cytotrophoblast.First trimester placentas were stained with immunofluorescence double staining for HLA-G (jÀm) (FITC À green) and S1PR2 (j and k) or S1PR3 (l and m) (Cy3 À red).Villous syncytiotrophoblast (*), lining at the decidual border (white open arrowhead), cytotrophoblasts (white arrow), the first proximal parts of the trophoblast cell column (#), decidual stroma cells (white arrowhead).Negative control rabbit and mouse IgG on first trimester placenta (mouse IgG À FITC in n and rabbit IgG À Cy3 in o).Scale bars represent 100 mm.

FIGURE 3
FIGURE 3 Differential expression of S1PR1, S1PR2 and S1PR3 across isolated primary cells of the human placenta.mRNA was investigated in arterial endothelial cells (ECA), venous endothelial cells (ECV) and primary trophoblasts (PT) (n = 4 each group) for S1PR1 (a), S1PR2 (d), S1PR3 (g) and protein expression (n = 3 each group) for S1PR1 (43 kDa) (b and c), S1PR2 (50À60 kDa) (e and f) and S1PR3 (42 kDa) (h and i).Data are presented as mean § SEM.GAPDH/TBP was used as reference gene for quantitative polymerase chain reaction, and b-actin (42 kDa) as reference protein in densitometry analysis.Differences were assessed by one-way analysis of variance, with post-hoc differences by Tukey's multiple comparison test.

FIGURE 4
FIGURE 4Effect of fluidic flow, different oxygen concentrations and platelet-derived factors on S1PR2 and S1PR3 expression in trophoblast cell line BeWo.BeWo cells were stimulated with forskolin (20 mmol/l) for 48 h to undergo differentiation under static conditions.Thereafter, cells were cultured under fluidic flow culture with indicated flow rates of 1 ml/min and 3 ml/min, respectively, for 24 h.Expression levels of S1PR2 (a) and S1PR3 (b) were compared with quantitative polymerase chain reaction (qPCR) between cells cultured under static control conditions and different flow rates (n = 4 independent experiments).After 48 h of differentiation, BeWo cells were cultured under three different oxygen concentrations (2.5%, 12% and 21% O 2 ) for 24 h.Expression levels of S1PR2 (c) and S1PR3 (d) were compared with qPCR between the different oxygen concentrations (n = 3 independent experiments).After 48 h of differentiation BeWo cells were co-incubated with freshly isolated human platelets from pregnant donors.2.25 £ 10 5 to 3.13 £ 10 5 platelets per ml in a total of 500 ml were loaded into an insert and activated with thrombin (1 IU/ml).Released platelet cargo could penetrate through the 0.4 mm insert membrane to get in contact with the underlying layer of trophoblast cells (e).The insert membrane was embedded for immunohistochemistry after the experiment and stained for platelet marker CD42b (f) scale bar represents 50 mm.The membrane was also embedded for transmission electron microscopy to further visualize the membrane and the incubated platelets on one side of the membrane (g).Scale bar represents 2 mm.Expression of S1PR2 (h) and S1PR3 (i) were compared with qPCR between the control (CTR), thrombin control (THR) and plateletderived factors (PLT IN) (n = 3 independent experiments).All data are presented as mean DDCq § SD, GAPDH/TBP was used as reference gene for qPCR and differences assessed by one-way analysis of variance with post-hoc differences by Tukey's multiple comparison test.

TABLE 1
Data are presented as mean (SD) or number.a Differences between all groups were calculated using one-way ANOVA, with post-hoc differences by Tukey's multiple comparison test.(Maternal age: first trimester versus preterm P = 0.4027, first trimester versus term P = 0.9727, pre-term versus term P = 0.5107; BMI: first trimester versus pre-term P = 0.9845, first trimester versus term P = 0.8143, pre-term versus term P = 0.9042) b Differences pre-term versus term were assessed using the unpaired Student's t-test.
ANOVA = analysis of variance; BMI = body mass index; NS = not significant.