Syncytiotrophoblast‐derived extracellular vesicles carry apolipoprotein‐E and affect lipid synthesis of liver cells in vitro

Abstract In normal pregnancy, hepatic metabolism adaptation occurs with an increase in lipid biosynthesis. Placental shedding of syncytiotrophoblast‐derived extracellular vesicles (STBEVs) into the maternal circulation constitutes a major signalling mechanism between foetus and mother. We investigated whether STBEVs from normal pregnant women might target liver cells in vitro and induce changes in lipid synthesis. This study was performed at the Nuffield Department of Women's & Reproductive Health, Oxford, UK. STBEVs were obtained by dual‐lobe placental perfusion from 11 normal pregnancies at term. Medium/large and small STBEVs were collected by ultracentrifugation at 10,000g and 150,000g, respectively. STBEVs were analysed by Western blot analysis and flow cytometry for co‐expression of apolipoprotein‐E (apoE) and placental alkaline phosphatase (PLAP). The uptake of STBEVs by liver cells and the effect on lipid metabolism was evaluated using a hepatocarcinoma cell line (HepG2 cells). Data were analysed by one‐way ANOVA and Student's t test. We demonstrated that: (a) STBEVs carry apoE; (b) HepG2 cells take up STBEVs through an apoE‐LDL receptor interaction; (c) STBEV incorporation into HepG2 cells resulted in (i) increased cholesterol release (ELISA); (ii) increased expression of the genes SQLE and FDPS (microarray) involved in cholesterol biosynthesis; (iii) downregulation of the CLOCK gene (microarray and PCR), involved in the circadian negative control of lipid synthesis in liver cells. In conclusion, the placenta may orchestrate the metabolic adaptation of the maternal liver through release of apoE‐positive STBEVs, by increasing lipid synthesis in a circadian‐independent fashion, meeting the nutritional needs of the growing foetus.


| INTRODUC TI ON
Extracellular vesicles (EV) are produced by most cells and organs including the human placenta. During pregnancy, EVs are shed from the syncytiotrophoblast (called STBEVs), the placental layer in direct contact with maternal blood, into the maternal circulation. 1 The release of STBEVs increases throughout gestation (and is estimated to be in excess of 3g every 24 hours at term. 2 STBEVs are heterogeneous and include small EVs (~100 nm in diameter), medium/large EVs (0. 1-1 μm) and apoptotic bodies (0.5-5 μm)-which are a terminal cellular events and not the focus of this work. According to their biogenesis, STBEV can be separated into exosomes and microvesicles that are released constitutively or after activation in relation to pathologies, such as cancer, cardiovascular disease or cellular stress. EVs have signalling functions and the ability to dock with other cells 3 to which they can carry and donate cargo, 4 to be taken up and used by the recipient cell.
The cargoes include regulatory molecules such as miRNAs, lipids and DNA. Uniquely, STBEVs carry placental alkaline phosphatase (PLAP), which allows identification of STBEVs as being of placental origin. 5 Circulating extracellular vesicles are, therefore, a form of complex systemic signalling. Thus, syncytiotrophoblast (the placental epithelium bathed in maternal blood) releases STBEV, which extend the potential for feto-placental communication in ways that have not been clearly undefined.
During pregnancy, levels of circulating triglycerides and cholesterol rise progressively with gestational age, due to the increase of maternal hepatic biosynthesis and peripheral lipolysis, and revert to normal levels postpartum. 6,7 The mechanisms by which the fetoplacental unit delivers its message of increasing energy requirements to maternal tissues are still matter of investigation.
A proteomic analysis of STBEVs obtained by dual-lobe placental perfusion of normal term placentae revealed the presence of apolipoprotein-E (ApoE) as a protein carried on them (Tannetta et al. unpublished). ApoE is a 299-residue protein carried by circulating lipoproteins, mainly chylomicrons and Intermediate Density Lipoproteins (IDL). 8 It is a key regulator of plasma lipid levels, promoting clearance of triglyceride-rich lipoproteins through its lipidbinding ability. In particular, the interaction between ApoE and Low-Density Lipoproteins Receptor (LDLR) in hepatocytes mediates endocytosis of lipoprotein remnant particles and lipolysis. 8 Since ApoE has a central role in mediating lipoprotein uptake by liver cells, we hypothesized that the expression of ApoE on STBEVs might have a role in mediating uptake of STBEVs by maternal liver cells in pregnancy, inducing metabolic changes.

| Human subjects
This study was approved by the Central Oxford Research Ethics Committee (07/H0607/74 and 07/H0606/148). Healthy pregnant women, undergoing elective caesarean section at term (indication for caesarean section was breech presentation or previous caesarean delivery), provided written informed consent for the use of the placental tissues and blood. Placentae were collected within 10 minutes of delivery. Normal pregnancy (NP) included healthy women with no history of hypertension or chronic illness, a singleton pregnancy without known foetal abnormality, and natural conception.

| Tissue immunofluorescence
Placental sections were deparaffinised by successive washings in Histo-clear 1 and 2 solutions (Dako, Glostrup, Denmark) before being rehydrated in alcohol solutions. Slides were dipped in sodium citrate buffer (10 mM, pH 6.0) (VWR, UK) for 10 minutes in a microwave oven, cooled at room temperature (R/T) and washed in PBS, before being incubated in a blocking solution (10% v/v FCS in PBS) for 1 hour. Sections were incubated overnight with 20 µg/ ml of apoE primary rabbit monoclonal antibody (Abcam, Cambridge, UK) and washed before and after staining with Alexa Fluor donkey anti-rabbit IgG secondary antibody (Life Technologies, Carlsbad, CA, USA) at 1:400 for 1 hour. Vectashield mounting medium containing DAPI (Vector Laboratories, USA) was used to counterstain the nuclei. Sections were visualized using a Leica DMIRE2 inverted fluorescence microscope. Images were taken using a Hamamatsu Orca digital camera with Simple PCI software.

| Isolation and characterization of STBEVs
Syncytiotrophoblast-derived extracellular vesicles were prepared using a dual-lobe placental perfusion system and serial centrifugation protocol as we have previously described in Dragovic et al. 9 Maternal placental perfusate was centrifuged at 10,000 × g to pellet the fractions enriched with the medium/large STBEVs, or microvesicles (MV), and 150,000 × g to pellet the small STBEVs, or nanovesicles (NV) or exosomes. Medium/large STBEVs were phenotyped using a BD LSRII flow cytometer (BD Biosciences) as described below. STBEVs' size and concentration were further characterized using a NanoSight NS500 (Malvern, UK). Western blotting was used to confirm syncytiotrophoblast origin of STBEVs, using a specific placental marker, placental alkaline phosphatase (PLAP) and the exosomal markers (Alix, Syntenin and CD9) for the small STBEVs. PLAP is exclusively expressed in placental tissues and is commonly used to distinguish vesicles released by the placenta from those coming from other cell types. The protein concentration of STBEVs was determined by bicinchoninic acid protein assay (BCA), prior to storage at −80°C.

| Nanoparticle tracking analysis
Syncytiotrophoblast-derived extracellular vesicles diameter and concentration were measured using a NanoSight NS500 (Malvern Instruments, Malvern, UK) equipped with an sCMOS camera and the nanoparticle tracking analysis software version 2.3, Build 0033 (Malvern Instruments, Malvern, UK). The instrument was calibrated prior to our measurements, using silica 100 nm microspheres (Polysciences, Warrington, UK). Size distribution profiles and concentration of STBEVs were measured using a protocol previously described. 9 International UK) in a humidified atmosphere containing 5% CO 2 in air at 37°C.

| Incubation of HepG2 with STBEVs
HepG2 cells were seeded at concentration of 2 × 10 6 cells/ml in 6 well plates (Merck Millipore, Germany). HepG2 cells were incubated for 24 hours with 2 ml of culture medium only or 2 ml of medium containing: (a) medium/large STBEVs, (b) small STBEVs, (c) medium/large STBEVs depleted of apoE, (d) small STBEVs depleted of apoE at a concentration of 50 µg/ml. After washing, cells were treated with 250 µl ice-cold lysis buffer (10 mM Tris HCl pH 7.4) for 1 hour. Supernatants were collected after centrifugation at 15,000 × g at 4°C for 10 min. Protein content was measured by BCA prior to storage at −80°C until use. For gene expression array or qPCR, cells were washed 3 times with PBS, collected and stored in PBS or RNA later (Qiagen, Hilden, Germany) at −20°C until use. Images were recorded using Zen software (Rochdale, UK).

| Flow cytometry analysis
Analyses of medium/large STBEVs and HepG2 cells were carried out by multi-colour flow cytometry, using a BD LSRII Flow Cytometer (BD Biosciences, Oxford, UK) equipped with a 488 nm (blue) and 633 nm (red) laser. All data were analysed using FACS DIVA software (Becton Dickinson, Oxford, UK).

| Syncytiotrophoblast-derived extracellular vesicles
Medium/Large STBEVs were samples were diluted in filtered PBS to optimize an event rate of ~300 events/s in a final volume of 300 µl. Prior to staining, samples were incubated with 10 µl of Fc receptor blocker (Miltenyi, Woking, UK) for 10 min, at 4°C.
Samples were then labelled with anti-PLAP-PE (Biolegend UK Ltd., Cambridge, UK) or anti-apoE-FITC (Abcam, Cambridge, UK) for 15 min at R/T. Isotype controls were matched to their respective antibodies according to the concentration, fluorochrome type and heavy chain. Prior to data acquisition, final volumes were adjusted to 300 µl with PBS. A total of 100,000 events were collected for each sample. EV gates were set at <1 µm using fluorescent beads.
The negative gates for staining were determined using isotype control tests and set at 1%.

| Cells
1-2 × 10 6 HepG2 cells were seeded at high density in six well plates (Merck Millipore, Germany). 2 ml of culture medium containing medium/ large or small STBEVs (50 µg/ml), isolated from 3 normal placentae and pooled together, depleted or not of apoE, were added in each well in a humidified atmosphere containing 5% CO2 in air at 37°C for 1 hour.
Cells were then treated with 0.5 ml of fixation buffer (BioLegend) for 20 minutes before being incubated in 1.5 ml of Accutase ® solution (Sigma-Aldrich) to detach the cells. Cells were centrifuged at 350g for 3 minutes, resuspended in 1X intracellular staining permabilization wash buffer (BioLegend) and centrifuges at 350g for 10 minutes.

| RNA isolation and mRNA-to-cDNA reverse transcription
After incubating HepG2 cells for 24 hours with (a) medium only (b) medium/large STBEVs, (c) small STBEVs, (d) medium/large STBEVs depleted of apoE, (e) small STBEVs depleted of apoE at a concentration of 50 µg/ml, cells were washed with warm PBS prior to RNA extraction using the RNeasy micro kit (Qiagen, Hilden, Germany). Samples (N = 10) were then treated with DNA-free removal kit (Thermo Fisher, Waltham, Massachusetts, USA) to remove any contaminating DNA. The RNA concentration was measured using a NanoDrop ND-1000 Spectrophotometer (Thermo Fisher) at 260 nm absorbance, and RNA purity was assessed using the 260:280 nm absorbance ratio. Samples were diluted with RNase-free water and adjusted to 100 ng/μL concentration. RNA samples were converted into cDNA using a High capacity RNA-cDNA conversion kit (Applied Biosystems, Thermo Fisher).

| Microarray analysis
Microarray analysis of HepG2 cells was performed using RNA prepared and hybridized to Illumina BeadChip microarrays, following manufacturers' protocols. Raw data were imported into the R statistical software (http://www.R-proje ct.org) and processed using BioConductor packages. 11 Ingenuity Pathway Analysis (Qiagen) was used to analyse the end-results, according to manufacturers' instructions. The experiment has been repeated twice.

| Quantitative polymerase chain reaction
The nucleotide Basic Local Alignment Search Tool and Primer-BLAST were used to design and analyse the specific- City, USA). The experiment has been repeated twice.

| Statistical analysis
Data were analysed using GraphPad Prism 7 (GraphPad Software, CA). Differences between 2 treatments were tested for statistical significance (p < 0.05) using a Student's unpaired t-test. Values were expressed as Mean ±Standard Error of the Mean (SEM).

| Clinical characteristics of recruited patients
Experiments were performed on 11 placentae collected from normal pregnant (NP) women. Clinical characteristics of women enrolled are shown in Table 1.

| ApoE is expressed in the syncytiotrophoblast of the placenta and on STBEVs
STBEVs derived from NP placentae were characterised prior to downstream experiments by transmission electron microscopy ( Figure 1A).
Medium/large STBEVs ( Figure 1A left) revealed a characteristic heterogeneous morphology and size consistent with previously observed findings, 5 while small STBEVs ( Figure 1A Right) were more uniformed, smaller in size and with the characteristic cup shape associated with exosomes as previously reported. 5 Western blotting confirmed the expression of PLAP on placental lysates, with increased expression in medium/large STBEV and small STBEV confirming the previous findings. 5 Additionally, small STBEVs were enriched for the known exosomal markers syntenin and CD9, as previously reported. 5 Having confirmed that STBEVs that we had isolated had the correct size, morphology and vesicular markers, we went on to investigate apoE expression.
Immunofluorescence staining of placental sections displayed high expression of apoE in the syncytiotrophoblast layer ( Figure 2A).

Western blot analysis demonstrated expression of both PLAP
and apoE in plasma, medium/large and small STBEVs ( Figure 2B).
Particularly, apoE expression seemed higher and with less variable concentration in small than in medium/large vesicles. An alternate apolipoprotein apo-A1 (which was not detected in our proteomics analysis) was detected in plasma but was not detected in either medium/large or small STBEVs confirming that apoE expression in STBEV was specific, and not due to lipoproteins contamination of STBEVs during ultracentrifugation process.

| Immunoaffinity depletion confirms that PLAP and apoE are expressed on the same vesicles
Both PLAP and apoE were expressed in pregnant plasma, medium/large and small STBEVs (pooled from placental perfusions) ( Figure 3A). After PLAP magnetic bead immunoaffinity depletion, medium/large STBEVs demonstrated complete depletion of both PLAP and apoE. Correspondingly, after apoE magnetic bead immunoaffinity depletion, medium/large STBEVs showed complete depletion of apoE and virtually all PLAP signal. These results suggest that almost all apoE-positive medium/large STBEVs co-express PLAP, and vice versa. Small STBEVs showed similar results with no PLAP and no apoE signal when PLAP depletion was applied. ApoE magnetic bead depletion completely denuded the samples of apoE expression, but a small PLAP signal remained, suggesting that there are some PLAP-positive small STBEVs, which do not express apoE.
These data suggest that significant amounts of apoE on STBEV is co-expressed with PLAP on the same vesicles. Using NTA, a modal size of 139±6nm was shown in apoE-positive small STBEVs, with a similar size distribution (130 ± 4 nm) after depletion for apoE ( Figure 3C). Furthermore, after immunodepletion for apoE, we observed a reduction of STBEVs concentration by 55%, thus, indirectly demonstrating that the apoE-positive small STBEVs comprised 45% of the total, substantially higher than that observed for medium/large STBEVs (about 16%). These data are consistent with the results of apoE obtained by Western blot.

| Medium/large and small STBEVs enter HepG2 cells through apoE-LDLR interaction
Having confirmed that apoE and PLAP were co-expressed on STBEV, we next assessed uptake of apoE-containing STBEV by analysing the presence of PLAP in HepG2 cells treated with native and apoEdepleted vesicles (HepG2 cells do not express PLAP). Confocal microscopy demonstrated uptake of PLAP-positive medium/large and small STBEVs by HepG2 cells (Figure 4 Panel A-H). Cells incubated TA B L E 1 Clinical data of human subjects (n = 11) whose placentae were used for isolation of STBEV from normal pregnancy. Data are expressed ad mean ± SD or number and percentage (%) of the total population analysed (n = 11).
with apoE-depleted STBEVs showed no positivity for PLAP ( Figure 4C and G). ApoE usually enters hepatic cells via the LDL receptor. 12,13 In

| STBEVs increase synthesis of cholesterol in HepG2 cells
The biological effects of ApoE-positive STBEV on HepG2 cells was assessed by measurement of cholesterol synthesis. HepG2 cells analysed by ELISA showed significantly higher production of cholesterol when incubated with medium/large or small STBEVs (p < 0.001 versus untreated) ( Figure 5A). When HepG2 cells were treated with either medium/large or small STBEVs depleted of apoE, synthesis of cholesterol did not increase ( Figure 5A).

| ApoE-positive STBEVs downregulate CLOCK gene expression in HepG2 cells
We   The CLOCK gene is a member of the clock gene family, which is central to the cellular regulation of circadian rhythm homeostasis and orchestrates shifts in metabolic patterns to accompany changes in activity and food consumption. [16][17][18][19][20] Nightly induction of CLOCK gene expression in liver cells reduces glucose and lipid synthesis, corresponding to the reduced energy requirements of diurnal mammals. 16,17 However, in pregnancy there is significant derangement of lipid synthesis with circadian changes in liver permitting maintenance of nutrient availability in pregnancy. 21 To confirm a metabolic effect, cholesterol synthesis was investigated in HepG2 cells incubated for 24 hours with apoE-positive STBEVs and we observed a striking increase in cholesterol synthesis after the uptake of both medium/large and small STBEVs. No significant increase of lipid synthesis was seen after incubation of liver cell with apoE-negative STBEVs, confirming that apoE expression on STBEVs is required to target liver cells and induce metabolic modifications in pregnancy.

| CON CLUS IONS
Our study provides a novel model of placenta-liver communication in human pregnancy, mediated by placental extracellular vesicles, that may be, in part, responsible for inducing the circadian-independent metabolic lipid changes occurring in pregnancy. The results of this study come from in vitro experiments and validation of these observations in vivo is required.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.