Placental Trophoblast Derived Exosomes Regulate Endometrial Epithelial Receptivity in Dairy Cows During Pregnancy

Inadequate feto-maternal interaction will directly lead to the failures of pregnancy and bring serious damage to dairy cows. Exosomes are widely involved in endometrial matrix remodeling, immune function changes, placental development, and other processes of embryo implantation and pregnancy of dairy cows. However, the role of placental trophoblast cells derived exosomes is still unclear in regulating the receptivity of endometrial cells and facilitating the interaction between mother and fetus. In this study, bovine trophoblast cells (BTCs) were obtained from bovine placenta and immortalized through the transfection of telomerase reverse transcriptase (TERT) gene. After that, the effect of trophoblast derived exosomes (TDEs) on endometrial receptivity in endometrial epithelial cells (EECs) was detected and the mechanism explored that TDEs and their proteins participated in feto-maternal interaction during bovine pregnancy. EECs were co-cultured with the exosomes derived from progesterone (P4) and treated with BTCs.


Introduction
Early pregnancy failures are as high as 50% in high-producing dairy cows and lead to the low reproductive e ciency and poor pro tability of the cattle production system, and such high pregnancy failure rate is generally considered to be the result of insu cient communication between the fetus and the environment [1] . In the mammalian fetus, placentae, a transient organ, play a critical role in fetomaternal interaction by ensuring adequate fetal oxygen and nutrient supply throughout the entire pregnancy, and determining the success of embryo implantation and pregnancy outcomes [2][3][4] . In bovine placentae, the barrier between mother and fetus is formed by trophoblast [5,6] . Thus, studying the function of trophoblast is vital to understand how to avoid obstetric diseases and increase pregnancy rate during rst trimester and pregnancy. There were only three bovine trophoblast cell lines, namely, F3 [7] , CT-1 [8] and BT-1 [9] . F3 is separated from bovine rst trimester placentae, but there is no report on whether F3 are immortal while isolated primary rst trimester trophoblast cells are known to have a nite lifespan [7,10] .
BT-1 and CT-1 can be cultured for more than 75 passages, but they are trophoblaststem cells derived from blastocysts which are different from trophoblast cells derived from rst trimester placentae, so there is no a model to research the function of placentae and it's di cult to collect samples [8,9] . Therefore, the establishment of a stable bovine trophoblast model derived from rst trimester placenta is helpful for better revealing the new mechanism of mother-fetus dialogue.
Feto-maternal interactions are comprised of the signal transmission between placental trophoblast cells and EECs to jointly advance embryonic development. In the maternal uterus, the success of feto-maternal interactions is determined by endometrial receptivity in EECs which is related to uterine factors such as integrin αv, integrin β3, wnt7a and mucin 1 (MUC1). Among them, integrin αv and β3 are transmembrane receptors which can bind to speci c ligands of the extracellular matrix. They can promote early embryomaternal interactions by impacting the opening of the endometrium implantation window and invasion process of trophoblast into EECs [11][12][13] . MUC1, an anti-adhesive, plays a negative role during early pregnancy by repressing the adhesion between trophoblast and uterine epithelial cells [14] . Wnt7a can establish signal connection between endometrial epithelial cells and mesenchyme and mediate the initial adhesion process between embryo and luminal epithelial cells by activating Wnt / β-catenin signaling pathway [15,16] .
Exosomes are double-layered extracellular vesicles with a diameter of 30-150 nm (extracellular vesicles, EVs) which release vesicles into the extracellular space to mediate the information exchange between cells by fusing with the plasma membrane [17,18] . Trophoblast cells can affect the maternal physiological function by sending speci c biologically active molecules to the maternal blood circulation through the villi space by packing them into exosomes [19,20] . The orderly expression of exosomes in time and space will enable the entire pregnancy to complete smoothly [21] . There have been many studies on the role of endometrium and exosomes secreted by embryos in the attachment of mammalian embryos, but there are few studies on the attachment of embryos by TDEs. First trimester trophoblast cells have been successfully immortalized by importing exogenous TERT gene in goat [22] , porcine [23] and human [10] , but not bovine. Therefore, it is still unclear what's the role of placental trophoblast cells derived exosomes in regulating the receptivity of endometrial cells and facilitating the interaction between mother and fetus. The aim of this study was to detect the receptivity changes of EECs resulting from taking in the exosomes secreted by the bovine trophoblast cell line which is established by introducing exogenous human telomerase during pregnancy period.

Materials And Methods
Isolation, puri cation and culture of bovine trophoblast cells Placentae of early pregnant Holsteins (45-60 days of pregnancy, with a fetal cow with a crown-to-rump length of about 7 cm) were obtained from the cows sacri ced by exsanguination. This study was performed in accordance with the guidelines of the Animal Ethic Committee of Beijing University of Agriculture (Permit No.: SYXK(JING)2015-0004). The isolation of cells was performed according to the previous study [6] . Brie y speaking, the fetal cotyledons were separated from the maternal caruncles aseptically and minced into 1mm 3 pieces, which were dispersed into a 10 cm 2 dish and cultured in a 37 o C incubator for 30 min till the pieces conglutinated on the dish sturdily. The pieces were rinsed and cultured with complete DMEM/F12 medium (Gibco, USA) supplemented with 10% exosome-depleted fetal bovine serum (FBS) (EXO-FBS-50A-1, SBI, USA) at 37 o C in an atmosphere of 5% CO 2 . The medium was changed every 3 days till the cells can be seen under a phase-contrast invert microscope. The cells were puri ed with the differential velocity adherent method [24] . Immuno uorescence detection was used to identify the primary bovine trophoblast cells (described below).

Giemsa stain assay
The cells were xed with 4% paraformaldehyde for 40 min and washed with PBS. After being dyed with Giemsa stain for 20 min, they were observed under a Nikon inverted optical microscope (Nikon, Japan).

Cell viability assay
According to the manufacturer's instructions (Solarbio, Beijing, China), the cells were seeded in a 96-well plate at the density of 10 4 cells per well, incubated with CCK-8 solution for 1 h, and then measured for the absorbance using a wavelength of 450 nm. The relative cell viability was calculated with the following formula: Relative Cell Viability = (The Absorbance of Test Group / The Absorbance of Control Group) * 100%.

Enzyme-linked immunosorbent assay (ELISA)
Analysis of the endocrine ability of hormone was performed according to the manufacturer's instructions (J L Biological, Shanghai, China).

Immuno uorescence assay
The cells were xed with 4% paraformaldehyde for 40 min and washed with PBS. The washed cells were immersed with permeabilization buffer for 30 min (PBS, 0.1% Triton X-100 and 1% BSA) and blocked with 1% BSA-PBS for 1h at room temperature. The cells then were incubated with primary antibodies including CK7 (1:100), vimentin (1:100), E-Cadherin (1:100), CD90 (1:100), CK8 (1:100) at 4 o C overnight, washed with PBS and incubated with uorescein isothiocyanate (FITC) -conjugated goat anti-rabbit secondary antibody (ab6881, 1:100, Abcam) in the dark for 1h at 37℃. The cells were washed with PBS three times and incubated with DAPI for 5min at room temperature. Next, the cells were observed with a laser scanning confocal microscopy (Olympus, Japan) after being washed with PBS three times.

Soft agar assay
Soft agar assay was performed according to the published protocol [25] . The 1.2% agar was lled into a six-well plate as a bottom layer which then was placed still at 4 o C till the agar solidi ed. 5×104 cells were suspended in 1 mL 0.6% noble agar in DMEM/F12 media and plated onto wells. The cells were cultured at 37 o C in an atmosphere of 5% CO 2 overnight and complete DMEM/F12 medium was added. They were fed with complete medium once a week and evaluated for the colony growth over the next two weeks. The cell clumps greater than 100 mm were considered as colonies and photographed with an inverted optical microscope.

Migration and invasion assay
The migration ability of BTCs was assessed using a 24-well plate in the BD Bio-Coat Matrigel Invasion Chamber (BD Bios-ciences). 10 4 trypsinized cells were added into 1 mL serum-free media. 500 µL complete media was added to the lower well while 100 µL cell suspension to the upper well. The plate was incubated at 37 o C in a 5% CO 2 atmosphere for 24 h. Those non-invading cells on the upper surface were removed using a cotton swabs and the lower surface of the basement membrane was xed in 4% paraformaldehyde for 30 min, and then the basement was washed with PBS three times and stained with crystal violet. The cells were observed and photographed with an inverted optical microscope.

Cell transfection
The plasmid was transfected into the cells using lipofectamine 2000 (Invitrogen, USA) according to the manufacturer's instructions. For transfection, the cells were seeded in 6-well culture plates and incubated with plasmid -lipofectamine complex at 50 nM in serum-free OPTI-MEM medium. The medium was replaced after 4 hours after transfection. The cells with 400mg / mL G418 (Solarbio, Beijing, China) were selected during two weeks. After selection, the transfected cells were subjected to the treatment for the follow-up experiment. pCI-neo-hTERT plasmid was identi ed and donated with Pro.Jin [22] .

Exosomes isolation
To imitate the early embryo culture conditions, the cells were treated with 2 ng/mL P4 when their density was about 80%. After 48h, the medium was collected and centrifuged at 4 o C (300×g, 10 min to remove the cells), and then the supernatant was transferred to a sterile vessel. Exosome isolation were performed from the supernatant according to the manufacturer's recommendations (BB-3901, Shanghai Bestbio Biotechnology Co., Ltd.).

Examination of exosomes morphology with transmission electron microscopy
The exosomes were resuspended in phosphate-buffered saline (PBS), and 15 µL samples were added to a 300-mesh copper grid for 90 s and dried. Each grid was washed ve times in distilled H 2 O, stained with 5µL of 3% uranyl acetate (phosphotungstic acid) for 1min and dried, and then placed on a lter paper for 1h at room temperature. The samples were visualized with a transmission electron microscope at 80 kV.

Sample preparation for proteomic analysis
Lysates from the exosomes were extracted with lysis buffer (500 mM Tris-HCl, 50 mM EDTA, 700 mM sucrose, 100 mM KCl, 2% β-mercaptoethanol and 1 mM phenylmethylsulfonyl uoride, pH 8.0), the protein was solidi ed by phenol-acetone assay and the precipitation was solubilized in 4% SDS. The protein concentrations were calculated by BCA assay and the enzymolysis protein concentrations tested by FASP assay.
Proteomic experiments were performed in biological triplicate. MS analyses were performed on an Agilent 1100 high-performance liquid chromatography (Agilent, USA). Peptides were loaded and separated over a 120 min gradient run using a Waters XBridge C18 analytical column (5µm 120 Å, 460µm × 250 mm, Waters, China). Trapping was conducted for 3 min at 5µL min − 1 , 97% buffer A (99% water, 2% ACN, pH 10.0), and 3% buffer B (98% ACN, 2% water, pH 10.0), before being eluted at 2-100% 0.1% FA in acetonitrile (2-40% from 0 to 100 min, 40-80% from 100 to 110 min ( ow rate, 250 nL min − 1 ). According to the building requirements of the DDA library, after the ow-through peaks were removed, cross-merging, vacuum concentration and mass spectrometry library building analysis were completed. Proteome Discoverer 2.1 (Thermo Fisher Scienti c, Rockford, IL, USA) was used to search and analyze the DDA data collected after classi cation. Biological samples were collected for DIA data, and quantitative analysis was performed with Skyline software (Department of Genome Sciences, University of Washington, Ave. NE, Seattle, WA). For pathway analyses, Kyoto Encyclopedia of Genes and Genomes (KEGG) resources were utilized with recommended analytical parameters. For gene ontology enrichment and network analyses, UniProt (www.uniprot.org) database resource (biological process, molecular function), Ingenuity Pathway Analysis, and Reactome knowledgebase were utilized. The heat map of proteins adopted gplots.

Statistical analysis
The research values were presented as mean ± SD. The data from three independent experiments were analyzed. One-way analysis of variance (ANOVA) was used to compare with the multiple time points and concentrations in the experiments. When the comparison between two groups means signi cant, * or + was used for P < 0.05, and ** or + + for P < 0.01.

Identi cation and telomerase transfection of BTCs
The cells derived from the bovine placenta were puri ed to Passage #5 with the differential velocity adherent method. The cells, growing as epithelial-like monolayers, showed a clear epithelioid cell morphology under a phase-contrast inverted microscope (Fig. 1A). After being dyed with Giemsa stain, they were observed under a microscope. There were minor binucleate cells among the uninucleate cells, which communicated with each other by the extended pseudopodia (Fig. 1B). BTCs were identi ed by detecting the biomarkers of extravillous trophoblast cells via immuno uorescence assay. Fetal bovine broblast cells (FBFs) were used as a negative control. As expected, all of the puri ed primary BTCs were CK7 positive while all of fetal bovine broblast cells were negative. The result that 100% of BTCs were negative of CD90 exactly veri ed there were no broblast mixing in the puri ed BTCs. Furthermore, the expression of two representative proteins in the epithelial cells, vimentin and E-cadherin suggested that BTCs were epithelial-origin cells (Fig. 1C). pCI-neo-hTERT plasmid was transfected in primary BTCs at Passage #5. Total RNA and protein were isolated from the primary BTCs, hTERT-BTCs (at Passages #30 and #50) and HeLa cells, and the expression of telomerase was detected on the protein level by western blot assay. The growth curve of BTCs, measured by CCK-8 assay, showed a basically similar trend in BTCs no matter whether they were transfected with plasmid or not (Fig. 1D). However, the proliferation activity of immortalized BTCs was higher than that of primary BTCs, and 50-passage hTERT-BTCs entered the exponential growth phase after a 3-day latent phase, while the primary BTCs showed a 5-day latent phase. Telomerase protein was expressed greatly and steadily in the hTERT-BTCs and HeLa cells while the expression level was signi cantly higher than that in the primary cells ( Fig. 1E-G, p < 0.01).

Immortalized BTCs maintain the functional characteristics of primary BTCs
Secretion of PL and CG is one of the most improtant feature of extravillous trophoblast cells [3] . Analyzed by ELISA assay, both primary BTCs and hTERT-BTCs manifest an ability of secreting PL and CG ( Fig. 2A).
Showing an invasive property con rmed by trans-well invasive assay, both primary BTCs and hTERT-BTCs have the same ability to migrate cells (Fig. 2B). In addition, hTERT-BTCs at Passage #50 showed the similar features of marker protein expression to those of primary BTCs (Fig. 2C, D). Soft agar assay was performed to con rm whether hTERT-BTCs have an ability of anchorage-independent growth. hTERT-BTCs were incapable of forming any colonies after 14-day incubation while HeLa cells formed, which indicated that hTERT-BTCs can't transform to a malignant phenotype (Fig. 2E).

Isolation and identi cation of exosomes derived from BTCs
Exosomes were extracted from the BTCs with an exosome extraction kit. Round vesicular membranes with a size of 30-150nm were observed under a transmission electron microscope (Fig. 3A). CD63 and CD9, the tetraspanin family members localized to the internal vesicles of the exosomes, are marker proteins of the exosomes. The expressions of CD63 and CD9 were detected in whole cell lysates (WCL) and exosomes derived from the BTCs by immunoblot assay. In the exosomes, the lack of the expression of calnexin, an endoplasmic reticulum protein marker, suggested that the exosomes we collected were free from cell debris contamination (Fig. 3B).

Effect of P4-treated BTCs derived exosomes on the endometrial receptivity of EECs
All of the bovine EECs we cultured showed a typical "paving stone"-like appearance with the positive expression of CK8, a marker protein of EECs (Fig. 4A). In this study, the EECs co-incubated with the TDE stained with PKH26 (Sigma Aldrich, USA) showed a signi cant red uorescence indicating that the TDE was fused with the EECs after 12 h of incubation (Fig. 4B). It has been reported that endometrial receptivity, as a key factor to determine the embryonic development process, can be affected by TDE and P4 stimulation during pregnancy. Exosomes derived from BTCs treated with or without 10 ng / mL P4 for 24 h were collected and co-cultured with EECs for 12 h. The expression changes of uterine receptivity factors related to embryo implantation were detected in the EECs by western blot assay. Compared with the BTCs which weren't treated, the P4-treated cells derived from exosomes increased the expression levels of uterine receptivity-related proteins in the EECs, such as integrin αv, integrin β3, and Wnt7a while decreasing the expression of MUC1 (Fig. 4C, D; p < 0.01).

Functional analysis of TDE proteome
Proteomic pro ling was performed on the exosomes derived from the BTCs cultured with / without P4 by nanoLC-MS/MS data-dependent acquisition. The samples were analyzed in biological triplicate, with technical replicates and a stringent metric for protein and peptide identi cation. The protein expression heatmap was shown for a total of 923 proteins identi ed in the proteomics analysis (Fig. S1). KEGG and GO classi cation showed that the proteins identi cated were concentrated on the immune system and endocrine system for antioxidant enzyme activity, binding, molecular function regulator, structural molecular activity, transporter activity, etc. (Fig. S1). Of these 923 proteins, 19 ones were differentially expressed in the cells under pregnancy situations, including 6 ones such as CTSD, PCOLCE and COPS6, that decreased expressions, and 13 ones such as GNPTG, EIF4A2 and FGB, that increased expression. (Fig. 5A, B). Functional annotation and pathway analysis of these differential expressed protein subsets associated with pregnancy group revealed the protein categories associated with collagen binding, catalytic activity and hydrolase activity (Fig. 5D, E).

Discussion
Success of pregnancy relies on the differentiation, proliferation, and invasion of placental trophoblast cells during early pregnancy [26,27] . The dysfunction of trophoblast may lead to severe complications threatening both the mother and the developing fetus and resulting in economic losses. Placenta trophoblast cells are consisted of two kinds, mononucleate trophoblast cells and binucleate trophoblast cells [28,29] . In this study, the cells we separated and puri ed from placentae contained mononucleated cells mixing with minority multinucleated cells which had similar features to those of the rst trimester trophoblast. The potential non-trophoblast components were mainly Hofbauer cells, blood leukocytes cells and broblast cells in this study. Hofbauer cells and blood leukocytes cells were easily deleted through the continuous passage [29] . We eliminated the broblast cells by continuous passage with the differential adhesion method. To con rm whether there was no broblast contamination, we detected the expression of CD90 / Thy1, the most iconic broblast biomarker, and found there were no broblast after the puri cation [10,30] . Cytokeratin 7, one intermediate protein, is a foremost marker of trophoblast cells [31,32] . The result that all of the cells were CK-7 positive con rmed that the cells we separated were derived by trophoblast and suitable for transfection.
To establish a normal bovine placental trophoblast cell line, we changed telomerase activity with the ectopic expression of exogenous hTERT gene. Based on western blot and RT-qPCR assay, the transfected hTERT-BTCs were positive for telomerase expression and have been cultured for over 50 passages without showing a sign of stopping the division. In addition, hTERT-BTCs showed a more aggressive growth rate compared to those without transfection, suggesting that exogenous telomerase expression accelerated the proliferation of BTCs. Thus, the cells can be considered to be immortalized. The secretory capability of pregnancy-related hormones such as chorionic gonadotropin (CG) and placental lactogen (PL) is one of the most critical features of trophoblast cells [10,22] . The immortalized BTCs keep the same characteristics of CG and PL secretion, and CK7, vimentin and E-cadherin expression till Passage #50. The trans-well Invasion assay indicated the immortalized BTCs inherited invasiveness. Thus, these results suggested that immortalized BTCs had epithelial-origin, endocrine and invasive features of extravillous trophoblast cells, which were maintained by these primary BTCs. Moreover, no CD90 expression in the BTCs suggested that there is no broblast contamination through the 50 passages of culturing process. In a word, the results showed that the immortalized bovine trophoblast cells possessed a prolonged lifespan and retained some properties of primary cells which are as same as the trophoblast cells previously reported [31] .
The increasing level of P4 was accompanied by early embryonic development process, so we suppose that P4 secreted by corpus luteum may stimulate trophoblast and change placenta derived exosomes to enhance the endometrial receptivity. In the present study, the exosomes derived from BTCs were stimulated with P4 in a concentration during the early pregnancy and co-cultured with EECs. We found these exosomes exert a positive effect on the expressions of endometrial receptivity-related proteins in EECs manifested by the upregulation of Integrin αv, β3 and Wnt7a while the downregulation of MUC1.
To detect if there were different proteins loaded in trophoblast derived exosomes, mass-spectrometrybased proteomics were performed and we found whether treated with P4, TDE proteins both focused on the immune system and endocrine system, while TDE could adjust the functions of EECs and regulate endometrial receptivity through immune and endocrine system. For example, TDE can give assistance to fetus escaping from maternal immune system and improve the secretory function of endometrium [16,33] . What's more, our results are also consistent with the existing literature that trophoblast can regulate endometrial receptivity [34] . From the further KEGG and GO classi cation research of differentially expressed proteins, we found that the expressions of these proteins including CTSD (cathepsin D), PCOLCE (procollagen C-endopeptidase enhancer), COPS6 (COP9 signalosome complex subunit 6) which can decompose extracellular matrix were all decreased with a P4 stimulation while the expressions of these antioxidant and hydrolase related proteins, such as ATIC (bifunctional purine biosynthesis protein ATIC), NANS (N-acetylneuraminate synthase) and ESD (S-formylglutathione hydrolase), were increased.
Dysregulation of endometrial extracellular matrix remodeling can impair receptivity which is linked to pregnancy outcomes [35] . Redox homeostasis and energy metabolism plays an important role on maternal receptivity by affecting endometrial response to conceptus-derived pregnancy signals [36] . In combination with the result that the cells derived from exosomes and treated with P4 can upregulated the expression of endometrial receptivity related proteins, the results of differentially expressed proteins between exosomes implied that endometrial receptivity was increased with a P4 stimulation under a pregnant situation through TDE by maintaining a steadier state of oxidation balance and extracellular environment. The expression levels of some transcription factors mediating cellular response, such as EIF4A2 (Eukaryotic initiation factor 4A-II), were also affected by P4 stimulation. In brief, TDE which aims on immune and endocrine ways, may regulate endometrial receptivity by ameliorating the extracellular environment, regulating cell redox balance and changing some of cell signal transduction during pregnancy.

Conclusion
In summary, we established an immortal placental trophoblast cell line by transferring exogenous hTERT. With this cell line, we found P4 regulates the function of exosomes derived from trophoblast cells, reversely regulates the receptivity of EEC, and ensures normal embryo development during pregnancy.

Declarations
Ethical Approval and Consent to participate Not applicable.

Consent for publication
Written informed consent for publication was obtained from all participants.    Expressions of integrin αv, integrin β3, Wnt7a and MUC1 were detected by western blot assay. D) The protein expression levels were measured by image J in representative Western blot analysis.

Figure 5
Differentially expressed protein landscape of TDE before and after P4 stimulation. A) Differential protein volcano map. B) Clustering heat map of THE differential proteins. Statistics of GO C) and KEGG D) enrichment classi cation of the differential proteins.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.