Novel function of pregnancy-associated plasma protein A: promotes endometrium receptivity by up-regulating N-fucosylation

Glycosylation of uterine endometrial cells plays important roles to determine their receptive function to blastocysts. Trophoblast-derived pregnancy-associated plasma protein A (PAPPA) is specifically elevated in pregnant women serum, and is known to promote trophoblast cell proliferation and adhesion. However, the relationship between PAPPA and endometrium receptivity, as well as the regulation of N-fucosylation remains unclear. We found that rhPAPPA and PAPPA in the serum samples from pregnant women or conditioned medium of trophoblast cells promoted endometrium receptivity in vitro. Moreover, rhPAPPA increased α1,2-, α1,3- and α1,6-fucosylation levels by up-regulating N-fucosyltransferases FUT1, FUT4 and FUT8 expression, respectively, through IGF-1R/PI3K/Akt signaling pathway in human endometrial cells. Additionally, α1,2-, α1,3- and α1,6-fucosylation of integrin αVβ3, a critical endometrium receptivity biomarker, was up-regulated by PAPPA, thereby enhanced its adhesive functions. Furthermore, PAPPA blockage with antibody inhibited embryo implantation in vivo, mouse embryo adhesion and spreading in vitro, as well as N-fucosylation level of the endometrium in pregnant mice. In summary, this study suggests that PAPPA is essential to maintain a receptive endometrium by up-regulating N-fucosylation, which is a potential useful biomarker to evaluate the receptive functions of the endometrium.

to analyze the blocking effects on the endometrial cell adhesive capacity to JAR cells. Control or rhPAPPA pre-treated HEC-1A and Ishikawa cell monolayers were incubated with UEA-1, LTL or LCA (5 μg/ml) for 4 h before JAR cells were added. The analysis results showed that the adhesion rate enhanced by rhPAPPA could be inhibited by UEA-1, LTL or LCA ( Fig. 2A and B), indicating that N-fucosylated oligosaccharide chains were involved in maintaining endometrial cell adhesive functions. Using Lectin flourescent staining (Fig. 2C) and Lectin blotting (Fig. 2D), we confirmed that rhPAPPA increased the level of α1,2-, α1,3and α1,6-fucosylation in HEC-1A and Ishikawa cells. These results suggest that PAPPA promotes endometrium receptivity by increasing N-fucosylation. Lectin blotting detected the levels of α1,2-, α1,3and α1,6-fucosylation in HEC-1A and Ishikawa cells after rhPAPPA treatment. The bar represents 50 μm. *p < 0.05, **p < 0.01, ***p < 0.001. The data were presented as the means ± SEM of three independent experiments.
Scientific RepoRts | 7: 5315 | DOI:10.1038/s41598-017-04735-0 rhPAPPA increases the expression of FUT1,FUT4 and FUT8 in HEC-1A cells. Different subtypes of N-fucosylation were catalyzed by specific FUTs. Based on the results above, we sought to explore whether rhPAPPA could increase FUT1, FUT4 and FUT8 expression. HEC-1A cells were treated with different doses of rhPAPPA (1 ng/ml, 10 ng/ml) for 48 h, or rhPAPPA (10 ng/ml) for 24 h, 48 h and 72 h. RNA and total protein were collected. The results of q-PCR ( Fig. 3A and B), immunofluorescent staining ( Fig. 3C and D) and Western blotting ( Fig. 3E and F) showed that rhPAPPA up-regulated the mRNA and protein expression levels of FUT1, FUT4 and FUT8. rhPAPPA rescues the impaired endometrium receptivity through up-regulating the specific N-fucosylation level. To explore the roles of specific FUTs in regulating endometrium receptivity, FUT1, FUT4, or FUT8 siRNAs were transiently transfected into Ishikawa cells to knockdown their expression, respectively. We found that rhPAPPA recovered the reduced FUT1, FUT4 and FUT8 expression to the level nearly similar to scrambled siRNA treated cells (Fig. 4A). Moreover, Lectin blotting (Fig. 4B) and Lectin flourescent staining (Fig. 4C) results also displayed that the reduced α1,2-, α1,3and α1,6-fucosylation levels were restored by rhPAPPA. The adhesive statistics results showed that each FUT siRNA transfection impaired Ishikawa cell receptivity, while rhPAPPA partially rescued the receptivity. However, after incubation with UEA-1, LTL or LCA, respectively, the rescued receptivity was decreased (Fig. 4D). These results suggest that rhPAPPA rescues the impaired receptivity through up-regulating the specific N-fucosylation level.  p-IGF-1R (Tyr 1131 ), p-AKT (Tyr 308 ) and p-AKT (Ser 473 ) were increased by rhPAPPA in a dose-dependent manner. To further study the regulation of FUT1, FUT4 and FUT8 by PAPPA through the IGF-1R/PI3K/Akt axis, AG1024 (IGFR inhibitor) and LY294002 (PI3K inhibitor) were used. Western blotting results showed that AG1024 and rhPAPPA promotes endometrium receptivity by up-regulating the specific N-fucosylation of integrin αVβ3. Integrin αV, β3 and αVβ3 antibodies or three Lectins were used to treat Ishikawa cells for 4 h before JAR cells were added (Fig. 6A). The analysis results showed that anti-αVβ3 antibody significantly reduced the receptivity, suggesting that αVβ3 plays an important role in the interaction between Ishikawa cells and JAR cells. After Ishikawa cells were treated with rhPAPPA, the adhesion rates in the Lectin (UEA-1, LTL or LCA) + rhPAPPA + anti-αVβ3 groups were all suppressed compared with those in the rhPAPPA + anti-αVβ3 groups (p < 0.05) (Fig. 6B). To further evaluate whether rhPAPPA could up-regulates the three subtypes of N-fucosylation of αVβ3, integrin αVβ3 was immunoprecipitated from the whole protein lysates of rhPAPPA-treated Ishikawa cells, and was probed with biotin labeled-UEA-1, LTL and LCA. Western blotting showed that rhPAPPA up-regulated the level of α1,2-, α1,3-, α1,6-fucosylation of αVβ3 (Fig. 6C). The expression of αV subunit and β3 subunit was also examined to verify that rhPAPPA treatment did not change their expression. Similar results were observed in immunofluorescent staining (Fig. 6D). These results demonstrate that rhPAPPA promotes endometrium receptivity by up-regulating the α1,2-, α1,3-, α1,6-fucosylation level of integrin αVβ3.
Adhesion rate of JAR cells to pre-treated Ishikawa as indicated is shown in the histogram. (C) Integrin αVβ3 was immunoprecipitated from the whole-protein lysate of untreated control and rhPAPPA-treated Ishikawa cells. The specific N-fucosylation of αVβ3 was detected by Western blotting. αV and β3 were detected to show the loading protein amount. Input showed the efficiency of immunoprecipitation. (D) Immunofluorescence and Lectin staining detected the expression and cellular localization of the specific fucosylation and αVβ3 after rhPAPPA treatment. Green, α1,2-, α1,3-, or α1,6-fucosylation; red, αVβ3; yellow (overlay), co-staining of α1,2-, α1,3-, or α1,6-fucosylation with αVβ3. DAPI (blue) was used for nuclear staining. The bar represents 50 μm. **p < 0.01, ***p < 0.001. The data were presented as the means ± SEM of three independent experiments. PAPPA blockade inhibits embryo implantation in vivo, mouse embryo adhesion and spreading in vitro, as well as specific N-fucosylation of pregnant mouse endometrium. To further investigate the effects of PAPPA on embryo implantation and endometrium receptivity in vivo, a pregnant mouse model was employed. PAPPA antibody was injected into the pregnant mouse uterus cavity at pregnant day 3 (PD3), and the mice were sacrificed at PD8 to analyze the embryo implantation rate. The statistical results showed that PAPPA blockade suppressed the embryo implantation rate compared with the IgG injection control groups (p < 0.05) ( Fig. 7A and B). Mouse embryos were collected at PD4, and were co-cultured with mouse primary endometrial cells in the presence of IgG or anti-PAPPA. After co-culturing for 24 h, the adhesion status of embryos in each groups was observed, and the results showed that anti-PAPPA inhibited mouse embryo adhesion to mouse primary endometrial cells (p < 0.01) (Fig. 7D). After co-culturing for 48 h, the embryos and endometrial cells were stained with CMFDA, and were photographed under a fluorescent microscope (Fig. 7E). The relative spreading area was analyzed, and the results showed that anti-PAPPA inhibited mouse embryo spreading on the endometrial cells (p < 0.05) (Fig. 7F). Pregnant mouse endometrium at PD4 exhibited advanced receptivity, which is the "implantation window" of mice. After injection with IgG or anti-PAPPA at PD3, uterus tissues and endometrial protein were collected at PD4. Using Lectin fluorescent staining (Fig. 7G) and Lectin blotting (Fig. 7H), we found that the endometrium at PD4 exhibited high level of α1,2-, α1,3and α1,6-fucosylation, whereas α1,2-, α1,3and α1,6-fucosylation levels were reduced by anti-PAPPA injection. Western blotting (Fig. 7I) confirmed that anti-PAPPA down-regulated the expression of FUT1, FUT4 and FUT8 at PD4. These findings suggest that PAPPA is essential to maintain the receptive functions of the endometrium and successful embryo implantation in mice.

Discussion
During the initial stages of pregnancy, the fertilized ovum develops to form a 2-, 4-, and 8-cell embryo and then the morula, finally becomes a mature blastocyst. The mature blastocyst enters into the uterine cavity and adheres to the endometrial luminal epithelium followed by the initiation of implantation and placentation 36 . At these stages, embryo-secreted factors not only regulate embryo development and implantation itself in an autocrine manner but also modulate the receptive functions of the endometrium in a paracrine manner 37 . For example, Paiva et al. reported that embryo-derived hCG enhanced endometrium receptivity through up-regulating the secretion of cytokines and growth factors (e.g., LIF and FGF2) from primary human endometrial epithelial cells (hEECs) 38 . Sakkas et al. also found that human blastocyst-released factors increased the expression of Hoxa10, which is a receptive endometrium marker in Ishikawa cells 39 . Evidence has also shown that abnormal or insufficient secreted factors, such as vascular endothelial growth factor (VEGF), placental growth factor (PlGF), epidermal growth factor (EGF), and PAPPA have been associated with an increased risk of preeclampsia 40 . In this study, we found that pregnant serum samples (high PAPPA level) dramatically enhanced HEC-1A receptivity compared with threatened abortion serum samples (low PAPPA level) (p < 0.01) (Fig. 1C). However, PAPPA antibody in human pregnant serum inhibited the adhesiveness of HEC-1A cells to JAR cells. Additionally, we further confirmed that PAPPA antibody caused the inhibition effects of embryo implantation in vivo, as well as mouse embryo adhesion and spreading in vitro. (Fig. 7A-F). These results indicate that a low PAPPA level around the endometrial microenvironment correlates with defective endometrium receptivity, and may directly lead to embryo implantation failure. Taken together, our results suggest that rhPAPPA facilitates trophoblast cell proliferation and adhesion 15 , as well as endometrium receptivity (Fig. 1A). We propose that exogenous rhPAPPA treatment in human embryo culture medium, or in the solutions when the embryo is transferred back into the uterus could be a novel and valuable approach to increase the successful rate of IVF-ET (in vitro fertilization and embryo transfer). For the first time, we demonstrate that PAPPA derived from trophoblast cells is an essential factor to maintain the condition of receptive endometrium. Therefore, we suggest that PAPPA could be considered a potential clinical biomarker to diagnose female infertility and a therapeutic target to improve embryo and endometrium dysfunction in implantation.
Glycans at the maternal-fetal interface are directly involved in regulating the interaction between the embryo and endometrium during the "window of implantation" 19 . For instance, Sd a antigen was found to be localized on the surface of blastocysts, and Lectin Dolichos biflous agglutinin (DBA) blockage inhibited the adhesion of mouse blastocysts to Ishikawa cells in vitro 41 . sLeX is stage-specifically expressed on both the endometrium and trophoblast cell surface, and is considered a functional biomarker of embryo implantation. Our previous study found that reduced sLeX level by FUT7 siRNA or sLeX antibody blockage inhibited the adhesive capacity of JAR cells to RL95-2 cells 42 . We also found that the LeY level was positively correlated with the receptive characteristics of HEC-1A and RL95-2 cells, and LeY antibody blockage prominently inhibited RL95-2 receptivity in vitro 43 . To systematically understand the effects of general N-fucosylation on endometrium receptivity, we detected all three types, α1,2-, α1,3and α1,6-fucosylation, and the correspondingly catalytic enzymes, FUT1, FUT4 and FUT8, respectively. The results showed that rhPAPPA enhanced HEC-1A and Ishikawa cell receptivity, while UEA-1, LTL, LCA incubation inhibited the receptive ability of endometrial cells to JAR cells ( Fig. 2A  and B). The results also showed that rhPAPPA up-regulated the expression of FUT1, FUT4 and FUT8 at both the gene and protein levels in HEC-1A cells (Fig. 3). Meanwhile, decreased α1,2-, α1,3and α1,6-fucosylation level by specific siRNA inhibited Ishikawa receptivity, whereas rhPAPPA partly recovered the N-fucosylation level and their adhesive capacity (Fig. 4). Additionally, after anti-PAPPA injection into the uterus cavity of pregnant mice at PD3, N-fucosylation and three N-fucosyltransferases were inhibited in the endometrium at PD4 (Fig. 7G-I). Other evidences also showed that the regulation of N-fucosyltransferases by different factors played crucial roles in maintaining endometrium receptivity. In LIF(−/−) mice, blastocysts do not attach normally to the maternal epithelium due to the down-regulated level of α1,2-fucosylation catalyzed by FUT1 in endometrial epithelial cells during the pre-implantation phase of pregnancy 44 . Nakamura et al. found that FUT1 expression was increased by cytokines secreted from macrophages in HEC-1A, Ishikawa, RL95-2 and primary endometrial epithelial cells 28 . Our previous study found that Baicalin promoted endometrium receptivity by up-regulating the expression of FUT4 in RL95-2 and mouse endometrial cells via Wnt/β-catenin signaling pathway 30 . Limited studies have reported the linkage between FUT8 and endometrium receptivity. However, FUT8 plays important roles in regulating cancer cell adhesion. For instance, Osumi D et al. found that FUT8 catalyzed α1,6-fucosylation of E-cadherin enhanced cell-cell adhesion in human colon carcinoma cells 45 . Taken together, our results demonstrate that each subtype of N-fucosylation participates in regulating a receptive functional endometrium, and PAPPA promotes endometrium receptivity through increasing the general N-fucosylation level.
An aberrant IGF-1 axis is implicated in many diseases, such as rheumatic diseases, cardiovascular diseases, diabetes and cancer, as well as infertility 46 . The studies also showed that an aberrant IGF-1 axis leads to insufficient endometrium functions. Baker et al. found that IGF1-deficient female mice were infertile, and exhibited uterine hypoplasia, suggesting that IGF-1 was crucial for uterine growth and receptive functions 47 . Kang YJ et al. also reported that the reduced expression of IGF-1R by miR-145 in endometrium inhibited embryo attachment 48 . PAPPA is an initiating regulator for the release of IGF-1 and activation of the IGF-1R signaling pathway. Recent studies have revealed that the PAPPA/IGF-1 axis is correlated with multiple reproduction processes. In PAPPA (−/−) mice, the IGF-1 axis was completely blocked, resulting in proportional dwarfism 49 . Nyegaard M et al. also found that a defective IGF-1 axis in PAPPA (−/−) mouse ovaries induced a decrease in the number of ovulated oocytes and serum hormone levels, and the reduced expression of ovarian steroidogenic enzyme genes 50 . In the current study, the results showed that the p-IGF-1R (Tyr 1131 ), p-AKT (Tyr 308 ) and p-AKT (Ser 473 ) expression levels were increased by rhPAPPA, indicating that PAPPA activated the IGF-1R/PI3K/Akt signaling pathway. Using the signaling pathway inhibitors AG1024 and LY294002, our results showed that AG1024 and LY294002 inhibited the expression of FUT1/4/8. Meanwhile, rhPAPPA slightly increased the expression of FUT1/4/8 in the presence of AG1024 and LY294002 (Fig. 5B-C). For the first time, we explored the mechanism of PAPPA enhanced endometrium receptivity by up-regulating the expression of FUT1, FUT4 and FUT8 via the PAPPA/IGFR/PI3K/Akt axis. This axis provided a new idea to elucidate the unexplained infertility due to an aberrant IGF-1 axis. PAPPA, the root of the IGF-1 axis, should attract sufficient attention.
N-linked glycans carried by integrins are involved in cell-cell and cell-extracellular matrix (ECM) interactions, thus modulating cell adhesion, proliferation, differentiation and migration by transferring signals from ECM to the cells 51 . For example, elevated α2,3-sialic acid levels of α2β1 by the overexpression of ST3Gal III promoted pancreatic cancer cell adhesion to type 1 collagen, and the activation of phosphorylated FAK 52 . Pocheć E et al. demonstrated that increased β1,6-branched N-glycan levels of αVβ3 enhanced melanoma cell migration on vitronectin and activated the FAK signaling pathway 53 . N-fucosylation of integrins also plays critical roles in regulating integrins-mediated adhesion. Li W et al. demonstrated that the loss of α1,6fucosylation on α4β1 led to a decreased binding between pre-B cells and stromal cells, which impaired pre-B cell generation in FUT8(−/−) mice 54 . In this study, the results showed that anti-αVβ3 dramatically inhibited. Ishikawa cell receptivity, suggesting that αVβ3 is an important molecule for trophoblast cell attachment. The adhesion rates of rhPAPPA and anti-αVβ3 combined with UEA-1, LTL, or LCA were all decreased compared with the rhPAPPA and anti-αVβ3 groups ( Fig. 6A and B). We further showed that rhPAPPA up-regulated the α1,2-, α1,3and α,1,6-fucosylation levels of αVβ3 by immunoprecipitation, immunofluorescent and Lectin fluorescent staining ( Fig. 6C and D). These results indicate that the general N-fucosylation level on αVβ3 is up-regulated by PAPPA, which promotes its adhesive functions. We suggest that exploring the specific N-fucosylation level or specific N-linked sugar chains on important adhesion molecule markers of receptive endometrium could be a novel approach for exploring the mechanism of unreceptive endometrium and unexplained infertility.
In summary, our study demonstrates that PAPPA derived from embryonic trophoblast cells at the maternal-fetal interface promotes endometrium receptivity by increasing α1,2-, α1,3and α1,6-fucosylation. PAPPA also up-regulates the expression of FUT1, FUT4 and FUT8 via IGFR/PI3K/Akt signaling pathway. Additionally, the three subtypes of N-fucosylation on integrin αVβ3 are enhanced by PAPPA. Furthermore, PAPPA antibody injection into the uterus cavity of pregnant mice inhibits embryo implantation, and the α1,2-, α1,3and α1,6-fucosylation levels of mouse endometrium. The findings of this work provide a novel glycobiological mechanism of PAPPA in regulating endometrium receptivity. Our study may help the development of a clinical diagnosis and potential therapeutics for unexplained infertility. Transient transfection. Ishikawa cells were seeded onto six-well plates or 96-well plates. When cells reached 70% confluence, scrambled siRNA, FUT1 siRNA, FUT4 siRNA and FUT8 siRNA (GenePharma, China) were transiently transfected into the cells using Lipofectamine 2000 reagent (Invitrogen, USA), respectively, following the manufacturer's instructions. The transfection reagent was removed after 6 h. Total protein was collected after 48 h for Western blotting. Transfected cells in 96-well plates were used to test the adhesive capacity with JAR cells using the adhesion assay.
Immunofluorescent and Lectin fluorescent staining. Cells pated on cover-slips or frozen slices (tissues) were fixed in 4% paraformaldehyde or cold acetone for 30 min, followed by blocking with 1% goat serum (Beyotime, China) for 2 h. Next, the cover-slips or slices were incubated with primary antibody or biotinylated Lectin at 4 °C overnight followed by incubation with FITC (green), TRITC (red)-conjugated second antibody or FITC or TRITC-conjugated streptavidin for 1 h. After incubation with DAPI (blue) for 5 min, anti-fade solution (Beyotime, China) was added to the cover-slips or slices, followed by photography under the fluorescent microscope (Olympus, Japan).

Immunoprecipitation.
Immunoprecipitation was performed with the Dynabeads ® Protein G Kit (Life technologies, USA) by following the standard procedure.

Animals and antibody injection.
All animal experiments performed in this study were approved by the Animal Ethics Committee of Dalian Medical University. The detail protocols and experimental processes conformed to the Experimental Animal Management Regulations of Dalian Medical University (Permit Number: #3555). Mice of the Kunming species (6-8 weeks) were from the Laboratory Animal Center of Dalian Medical University, China. Mice were maintained under controlled environmental conditions (temperature 22-25 °C; humidity: 60%; light-controlled 12-h light/12-h darkness). After mating, if the females mice were confirmed for the presence of a vaginal plug in the next morning, it was defined as pregnant day 1 (PD1). Twenty-four pregnant mice were randomly divided into 2 groups. On PD3 (8:30 AM), 12 mice were anesthetized with pentobarbital sodium (50 mg/kg); PAPPA antibody (10 μl, 200 μg/ml, Santa Cruz, USA) was injected into the right uterus horn, and IgG was injected into the left uterus horn as a control. In addition, 12 mice were injected with IgG into the left uterus horn, with no treatment of the right uterus horn. On PD4 (8:30 AM), the pregnant mice in each group (n = 6) were euthanized by cervical dislocation. The uteri were fixed in 4% (v/v) paraformaldehyde to prepare frozen tissue sections, and the endometrial tissues were carefully collected for protein extraction. On PD8, the pregnant mice were sacrificed, and the number of implanted embryos was counted and analyzed.
Culture of primary mouse endometrial cells. The uteri of pregnant mice at PD4 were split longitudinally and were washed with PBS (without Ca 2+ and Mg 2+ ) followed by digestion with 2% trypsin. Tissues were incubated at 4 °C for 2 h followed by another 30 min at room temperature. Tissues were gently shaken, and the endometrial cells were collected by centrifugation at 500 rpm for 10 min. Cells were washed three times with DMEM/F12. Next, the cell suspension was adjusted to 500 cells/μl and then was placed in 96-well plates, followed by culture according to standard procedures. The culture medium was changed the following day to remove unattached cells and cell debris.
Embryo collection. Mouse embryos were flushed from the uteri of pregnant mice at PD4 with DMEM/ F12 as mentioned above. Normally developed blastocysts were selected, and prepared for transferring to the co-culture medium with primary endometrial cells.
Embryos and endometrial cells co-culture. Harvested epithelial cells were placed in 96-well plates and cultured under the same condition as above. After endometrial cells formed a monolayer, embryos were transferred into the medium treated with IgG or anti-PAPPA, and their attachment and spreading condition were observed under the microscope. Statistical analysis. GraphPad Prism ® (GraphPad Software Inc., USA) was used for statistical analysis.
All experiments were performed at least 3 independent times, and the data were shown as means ± SEM. For the analysis of difference between groups, independent-samples t-test or one-way ANOVA was performed, and p < 0.05 was considered statistically significant.