- Email: [email protected]
Expression of GALNT2 in human extravillous trophoblasts and its suppressive role in trophoblast invasion
Placenta 33 (2012) 1005e1011
Contents lists available at SciVerse ScienceDirect
Placenta journal homepage: www.elsevier.com/locate/placenta
Expression of GALNT2 in human extravillous trophoblasts and its suppressive role in trophoblast invasion W.-C. Liao a,1, C.-H. Chen b, c,1, C.-H. Liu c, d, M.-J. Huang c, d, C.-W. Chen c, J.-S. Hung g, C.-H. Chou c, C.-H. Chen c, M.-I. Che c, H.-M. Chang a, C.-T. Lan a, H.-C. Huang e, G.-F. Tseng f, M.-K. Shyu b, *, M.-C. Huang c, d, ** a
Department of Anatomy, Faculty of Medicine, Chung Shan Medical University, Taichung, Taiwan Department of Obstetrics and Gynecology, National Taiwan University Hospital, 7 Chung-Shan South Road, Taipei 100, Taiwan Graduate Institute of Anatomy and Cell Biology, National Taiwan University College of Medicine, Taipei 100, Taiwan d Research Center for Developmental Biology and Regenerative Medicine, National Taiwan University, Taipei 100, Taiwan e Animal Technology Institute Taiwan, Miaoli, Taiwan f Department of Anatomy, College of Medicine, Tzu-Chi University, Hualien, Taiwan g Department of Surgery, National Taiwan University Hospital, Taipei 100, Taiwan b c
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 30 August 2012
Extravillus trophoblast (EVT) invasion plays a critical role in placental development. Integrins bind to extracellular matrix (ECM) proteins to mediate EVT cell adhesion, migration, and invasion. Changes in Oglycans on b1-integrin have been found to regulate cancer cell behavior. We hypothesize that O-glycosyltransferases can regulate EVT invasion through modulating the glycosylation and function of b1integrin. Here, we found that the GALNT1 and GALNT2 mRNA were highly expressed in HTR8/SVneo and first trimester EVT cells. Immunohistochemstry and immunofluorescence staining showed that GALNT2 was expressed in subpopulations of EVT cells in deciduas, but not in syncytiotrophoblasts and cytotrophoblasts of placental villi. The percentage of GALNT2-positive EVT cells increased with gestational ages. Overexpression of GALNT2 in HTR8/SVneo cells significantly enhanced cell-collagen IV adhesion, but suppressed cell migration and invasion. Notably, we found that GALNT2 increased the expression of Tn antigen (GalNAc-Ser/Thr) on b1-integrin as revealed by Vicia Villosa agglutinin (VVA) binding. Furthermore, GALNT2 suppressed the phosphorylation of focal adhesion kinase (FAK), a crucial downstream signaling molecule of b1-integrin. Our findings suggest that GALNT2 is a critical initiating enzyme of O-glycosylation for regulating EVT invasion. Ó 2012 Published by Elsevier Ltd.
Keywords: GALNT Mucin Integrin O-glycans EVT Invasion
1. Introduction During placental development, cytotrophoblasts in placental villi differentiate into syncytiotrophoblast and extravillous trophoblast (EVT). Interstitial invasion of EVT through endometrial decidua is crucial for successful pregnancy. A stringent regulation of EVT invasion is necessary for proper blood vessel remodeling at the maternalefetal interface and is crucial for maintaining a normal pregnancy. Failure of this remodeling process is associated with
* Corresponding author. Tel.: þ886 2 23123456x71537. ** Corresponding author. Graduate Institute of Anatomy and Cell Biology, National Taiwan University College of Medicine, No. 1, Sec. 1 Jen-Ai Road, Taipei 100, Taiwan. Tel.: þ886 2 23123456x88177; fax: þ886 2 23915292. E-mail addresses: [email protected] (M.-K. Shyu), [email protected] (M.-C. Huang). 1 These authors contributed equally to this article. 0143-4004/$ e see front matter Ó 2012 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.placenta.2012.08.007
pre-eclampsia [1], intrauterine growth restriction [2], late miscarriage [3], and preterm delivery [4]. Glycosylation is the most common post-translational modification of proteins. Aberrant glycosylation affects many cellular properties, including, cell proliferation, differentiation, transformation, migration, invasion, apoptosis, and immune responses [5]. N-linked and O-linked glycans are the two major types of carbohydrates in mammalian cells. The mucin-type O-glycosylation is initiated by the transfer of UDP-N-acetylgalactosamine (UDPGalNAc) to a serine (S) or threonine (T) residue to form Tn antigen (GalNAca-S/T) [6]. This reaction is catalyzed by a family of polypeptide GalNAc transferases (GALNTs), consisting of 20 members in humans, namely, GALNT1 to 14 and GALNTL1 to L6 [7,8]. O-linked glycans have been shown to regulate a variety of biological functions and contribute to the development of human diseases. Loss of GALNT1 activity in mice results in bleeding disorder [9]. GALNT2 can affect lipid metabolism through apoC-III glycosylation [10]. In
1006
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
addition, GALNT2 has been shown to regulate EGF-induced malignant phenotypes in hepatocellular carcinoma [11]. O-glycosylation of fibronectin regulated by GALNT3 and GALNT6 modulates epithelialemesenchymal transition [12]. GALNT6 modifies mucin 1 glycosylation and regulates proliferation of breast cancer cells [13]. Death-receptor O-glycosylation controls tumor-cell sensitivity to the proapoptotic ligand Apo2L/TRAIL [14]. Integrins, the major cell surface receptors mediating cellextracellular matrix (ECM) interactions, consist of 18 a subunits and 8 b subunits, which assemble into at least 24 ab heterodimers. More than half of these heterodimers contain b1 integrins. Functional studies have demonstrated the importance of a1b1 integrins in the regulation of EVT invasion into the uterine wall [15e17]. The b1 integrins have also been reported to modulate trophoblast motility during EVT invasion [18,19]. Both N-glycans and O-glycans are present on b1 integrins [20,21]. Changes in these carbohydrates can modulate b1 integrin activity and cell phenotypes including, cell growth, adhesion, differentiation, migration and invasion [20,21]. Identification of the glycosyltransferases responsible for the glycosylation of b1 integrins during EVT invasion will therefore contribute to our understanding of the development of EVT invasion. Although O-glycosylation has been demonstrated to regulate a variety of cellular behaviors, the expression and role of O-glycan initiating enzymes in human placenta have never been studied. Here we report that GALNT2 is highly expressed in subpopulations of EVT cells in human placenta. Moreover, GALNT2 overexpression modulates the O-glycosylation of b1 integrins and suppresses integrin-mediated signaling and invasion in HTR8/SVneo EVT cells. 2. Materials and methods 2.1. Clinical tissue collection Human placental tissues of the first trimester (7e12 wk gestation; n ¼ 10), second trimester (17e25 wk gestation; n ¼ 6), and third trimester (36e39 wk gestation; n ¼ 13) were obtained from the Department of Obstetrics and Gynecology, National Taiwan University Hospital, as previously described [22]. All the pregnancies were confirmed by ultrasound. The gestational age is dated by crownrump length before 14 weeks of gestation and bi-parietal diameter after 14 weeks. The placentas of first trimester were taken from the samples received by dilatation and curettage of viable pregnancies, which were decided to terminate due to unwillingness to keep the pregnancy further. The placentas of second trimester were obtained from preterm labors, where equal in number of vaginal delivery or cesarean section. The placentas of third trimester were all received from cesarean sections of normal pregnancies which were scheduled for section due to obstetric indications before labor pain onset. No major complications of these cases were noted during labor or postpartum. The use of human placentas for this study was approved by the local hospital ethics committee, and written consent was obtained from patients before sample collection. Specimens were fixed in 4% (w/v) paraformaldehyde/PBS. Serial 5-mm of paraffin-embedded placental sections were used for immunohistochemistry. Serial 14-mm cryostat sections were used for immunofluorescence microscopy. For RNA extraction, specimens were soaked in RNA later (Qiagen, Valencia, CA) at 4 C overnight and then stored at 20 C. For Western blot analysis, the samples were stored at 80 C until use. 2.2. Isolation of primary extravillous trophoblast cells Placenta collection and primary EVT cells were isolated as described by Kliman et al. [23]. Primary EVT cells were isolated from first trimester placentas (5e12 weeks, n ¼ 4) and third trimester placentas (36e38 weeks, n ¼ 3). Briefly, minced fragments of placentas were extensively washed and then subjected to three cycles of digestion with 1.25 mg/ml of trypsin (Life Technologies, Burlington, Canada) and 100 mg/ml of DNase (Sigma, St. Louis, MO), followed by a discontinuous 5% (5%e70%) Percoll (Sigma) gradient centrifugation. EVT-containing cell suspensions were recovered from 35 to 50% gradient layers and the purity of EVT was higher than 90%, analyzed by immunofluorescence staining using anti-cytokeratin (CK)7 antibody (Santa Cruz, CA).
SVneo cells were grown in RPMI 1640 (HyClone Laboratories, Logan, UT) supplemented with 5% fetal bovine serum (FBS) (PAA Laboratories, Pasching, Austria) in a humidified tissue culture incubator at 37 C in 5% CO2 atmosphere. For transfection, 6 105 of HTR8/SVneo cells were transiently transfected with 10 mg of empty vector (Mock) or GALNT2/pcDNA3.1/myc-His [11] with Lipofectamine 2000 (Invitrogen). After 72 h of transfection, cells were harvested for experiments. 2.4. RNA extraction and cDNA synthesis Trizol reagent (Invitrogen, Carlsbad, CA) was used to extract total RNA according to the manufacturer’s protocol. StrataScript reverse transcriptase (Stratagene, La Jolla, CA) was used for RT-PCR. 2.5. Quantitative real-time PCR Quantitative PCR system Mx3000P (Stratagene) and Brilliant SYBR Green QPCR Master Mix (Stratagene) were used to analyze gene expression in human placentas according to the manufacturer’s protocol. PCR primers used in this study were as previously described [11]. Samples were analyzed in triplicate, and product purity was checked through dissociation curve at the end of real-time PCR cycles. Relative quantity of specific gene expression normalized to b-actin was analyzed with MxPro Software (Stratagene). 2.6. Immunohistochemistry and immunofluorescence staining 2.6.1. Paraffin-embedded human placental sections were de-paraffinized in xylene and re-hydrated in a series of graded alcohol. The sections were incubated with primary antibodies (1:100 for anti-GALNT2, Sigma; 1:100 for anti-cytokeratin (CK) 7 mAb, Santa Cruz, CA) diluted in 5% (w/v) non-fat milk/PBS for 16 h at 4 C. Negative controls were performed by replacing primary antibodies with an isotype-matched control IgG at the same concentration. Super SensitiveÔ Link-Label IHC detection System (BioGenex, San Ramon, CA) was used and the specific immunostaining was visualized with 3,3-diaminobenzidine (DAB) liquid substrate system (Sigma). All sections were counterstained with hematoxylin for mounted with UltraKitt (J.T. Baker, Deventer, Holland). For immunofluorescence staining, cryo-sections were incubated with anti-GALNT2 (1:100) and anti-CK7 (1:100) antibodies for 1 h at 4 C, followed by incubation with FITC-conjugated anti-rabbit IgG (1:200; Vector, Burlingame, CA) and Cy3-conjugated anti-rabbit IgG (1:400; Jackson ImmunoResearch, West Grove, PA). For immunofluorescence microscopy of HTR8/SVneo cells, primary antibodies anti-GALNT2 and anti-GM130 (1:100; BD Biosciences, San Jose, CA) and secondary antibodies FITC- or Cy3-conjugated anti-rabbit IgG were used. Cell nuclei were counterstained by DAPI (Sigma). Immunofluorescence micrographs were obtained at 400 magnification using Leica TCS SP5 Spectral Confocal System (Leica, Wetzlar, Germany). 2.7. Cell growth analysis Cells (4 104) were seeded in 6-well plates with RPMI 1640 containing 10% FBS (PAA Laboratories). Viable cells were determined at 24 h intervals for 72 h using hemocytometer with trypan blue exclusion. 2.8. Cell adhesion assay Ninety-six-well plates were coated with 5 mg/ml of BSA, human collagen IV (Sigma), human fibronectin (Sigma), or murine laminin (Sigma) at 4 C for overnight. Cells (2 104) were seeded and allowed to attach for 30 min at 37 C in a humidified 5% CO2 incubator. For functional blocking of cell adhesion, cells (2 104) were preincubated with b1-integrin blocking antibody (1:200; clone P4C10, Millipore, Temecula, CA) or control IgG for 30 min and then seeded into 96-well plates coated with collagen IV. Numbers of attached cells from four wells were determined under an inverted microscope. 2.9. Migration and invasion assays For migration assays, chambers with 8-mm pore size membrane (Corning, Lowell, MA) were used. Cells (3 105) in 500 ml serum free DMEM were seeded to the upper part of the chamber and 10% FBS was used as chemoattractant in the lower part. Cells were allowed to migrate for 30 h in a humidified tissue culture incubator at 37 C, 5% CO2 atmosphere. The migrated cells were fixed with 100% methanol and stained with 0.5% crystal violet. The numbers of migrated cells in each well were counted under a phase contrast microscope. Cell invasion assays were performed in BioCoatÔ MatrigelÔ Invasion Chambers (BectoneDickinson, Bedford, MA) according to the manufacturer’s protocol with similar conditions.
2.3. Cell culture and transfection 2.10. Western blot analysis and lectin pull down assay Human extravillous trophoblast cell line HTR8/SVneo derived from the first trimester placenta [24] was a gift from Dr. Charles H. Graham (Department of Anatomy and Cell Biology, Queen’s University at Kingston, Ontario, Canada). HTR8/
HTR8/SVneo cells were lyzed in lysis buffer containing 1% (v/v) Triton X-100, 20 mM TriseHCl (pH 8.0), 160 mM NaCl, 1 mM CaCl2, and 1 mM PMSF, followed by
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
1007
ultrasonication (3 5-s burst). Forty micrograms of total proteins were separated on an 8% SDS-PAGE and transferred to a Hybond enhanced chemiluminescence nitrocellulose membrane (GE Healthcare, Chalfont St Giles, UK). The membrane was incubated with rabbit anti-GALNT2 antibody (1:100, Sigma), rabbit anti-FAK polyclonal antibody (C-20, Santa Cruz), rabbit anti-FAK pY397 mAb (1:1000; Biosouce, Nivelles, Belgium), or mouse anti-GAPDH mAb (BD Pharmingen). Bands were visualized by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies (1:10,000 dilution; Vector laboratories) and enhanced chemiluminescence reagents (GE Healthcare, UK). The band signals were quantified and normalized to their control signals with Image Quant 5.1 software (Molecular Dynamics, Sunnyvale, CA, USA). For detection of glycoproteins decorated with the Tn antigen, biotinylated Vicia villosa agglutinin (VVA) (Vector Laboratories, Burlingame, CA) was used. For lectin pull down assays, 200 mg of cell lysates were incubated with VVA-agarose beads (Vector Laboratories, Burlingame, CA) overnight at 4 C. The pulled down proteins were then subjected to Western blotting. 2.11. Statistical analysis Student’s t-test was used for statistical analyses. Data are presented as means SD. P < 0.05 is considered statistically significant.
3. Results 3.1. GALNT2 mRNA is expressed in first trimester EVTs and is increased in third trimester EVTs
Fig. 1. mRNA expression of GALNT family genes in HTR8/SVneo and primary EVT cells. (A) mRNA expression levels of 20 GALNT family genes in HTR8/SVneo and first trimester EVT cells. Primary EVT cells were isolated from first trimester placentas (n ¼ 4). The mRNA levels were analyzed by real-time RT-PCR. The signals of GALNT genes were normalized to b-actin. (B) Expression levels of GALNT2 mRNA are increased in late pregnancy. Expression levels of GALNT2 in primary EVT cells from first (n ¼ 4) and third (n ¼ 3) trimester placentas were compared. Results are presented as the mean SD from three independent experiments. *P < 0.05, compared with the first trimester EVT.
O-glycosylation regulates a wide range of cellular behaviors. To investigate the roles of O-glycosylation in EVT cells, we first analyzed the mRNA expression of 20 GALNT family genes in the first trimester EVT cell line HTR8/SVneo and primary EVT cells isolated from first trimester placentas (n ¼ 4). Our results showed that GALNT1 and GALNT2 were expressed at the highest levels in both HTR8/SVneo and first trimester EVT cells, compared with other GALNT members (Fig. 1A). Since we previously found that GALNT2 is able to regulate cancer cell invasion [11], we first investigated the expression and role of GALNT2 in EVT cells. To know whether GALNT2 was differentially expressed in EVT cells from different
Fig. 2. Immunohistochemistry of GALNT2 in the human placenta. Human placentas from the first, second, and third trimester were immunostained with anti-CK7 mAb (left panel) or anti-GALNT2 polyclonal antibody (middle panel). Amplified images in rectangles of GALNT2 staining were shown in the right panel. CK7 is an EVT marker. The percentage of GALNT2-positive EVT cells (GALNT2/CK7 double positive cells) were shown in the middle panel. Results are presented as the mean SD from five different fields of at least three placentas for each trimester. *P < 0.05, compared with the first trimester. Placental villi did not show GALNT2 staining under our experimental conditions (arrows). The negative control did not show any signals (lower left of the left panel, 1st trimester). Cell nuclei were counterstained with hematoxylin. Scale bars ¼ 25 mm.
1008
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
Fig. 3. Immunofluorescence staining of GALNT2 and CK7 in placentas. (A) GALNT2 expression in EVT cells. The first (10 weeks) and third trimester (37 weeks) deciduas were immunostained with anti-CK7 mAb (green) and anti-GALNT2 polyclonal antibody (red). Representative images were shown. Cell nuclei were counterstained with DAPI (blue). Merged images reveal co-localization of GALNT2 and CK7. Scale bar ¼ 25 mm. Negative controls did not show signals (data not shown). (B) Percentage of GALNT2-positive EVT in deciduas. The percentage of GALNT2-positive EVT cells were calculated from the merged images at five different fields. Results are presented as the mean SD from three independent experiments. *P < 0.05, compared with the first trimester.
Fig. 4. Overexpression of GALNT2 in HTR8/SVneo cells. (A) Overexpression of GALNT2in HTR8/SVneo cells analyzed by Western blotting. GAPDH is an internal control. The relative intensity of signals is presented as the mean SD (n ¼ 3). *P < 0.05. (B) Immunofluorescence microscopy showed overexpression of GALNT2 (green) in >80% of the HTR8/SVneo cells transiently transfected with GALNT2 plasmids for 72 h. Nuclei were counterstained with DAPI (blue). The Golgi apparatus was stained with anti-GM130 antibody (red). Merged images reveal co-localization (yellow) of GALNT2 and the Golgi apparatus. Amplified images in rectangles were shown in the lower right. Scale bar ¼ 25 mm (C) GALNT2 overexpression enhances Vicia Villosa agglutinin (VVA) binding to glycoproteins. Equal amounts of cell lysates from Mock and GALNT2 transfectants of HTR8/SVneo cells were blotted with biotinylated VVA. GAPDH is a loading control.
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
1009
gestational stage, primary EVT cells were also isolated from third trimester placentas (n ¼ 3). We found that GALNT2 mRNA expression was significantly increased in third trimester EVT cells compared with that in first trimester (Fig. 1B). These findings suggest that GALNT2 is expressed EVT cells and may play a role in regulating EVT behavior.
performed. Consistent with the results from immunohistochemistry, immunofluorescence staining revealed a higher percentage of GALNT2-positive EVT cells in the third trimester placenta compared with the first trimester placenta (Fig. 3).
3.2. Immunohistochemistry and immunofluorescence of GALNT2 in human placentas
We observed that a higher percentage of EVT cells expressed GALNT2 in the third trimester placenta than those in the first trimester. To investigate effects of GALNT2 expression on EVT cells, we first overexpressed GALNT2 in the human first trimester EVT cell line, HTR8/SVneo. HTR8/SVneo cells were transiently transfected with GALNT2 for 72 h and its overexpression was verified by Western blotting (Fig. 4A). Immunofluorescence microscopy showed that GALNT2 was overexpressed in most HTR8/SVneo EVT cells. GALNT2 subcellular localization was found in the Golgi, as expected for glycosyltransferases (Fig. 4B). To confirm that GALNT2 exhibited enzymatic activity of GalNAc transferase, biotinylated VVA was used to detect GalNAc on glycoproteins. Our data showed increased GalNAc expression on cellular proteins in GALNT2overexpressing HTR8/SVneo cells compared with mocktransfected cells (Fig. 4C).
To examine the expression pattern of GALNT2 in the human placenta, immunohistochemistry of placentas from different gestational stages was performed. CK7 is a marker for EVT (Fig. 2, left panel). In deciduas of the first trimester placenta, GALNT2 was expressed in 30% of EVT cells (Fig. 2, upper panel), which is calculated as GALNT2-positive cells/CK7-positive cells 100%. In deciduas of the second (Fig. 2, middle panel) and third (Fig. 2, lower panel) trimester placenta, GALNT2 was expressed in 72% and 73% of EVT cells, respectively. By contrast, GALNT2 could not be detected in placental villi throughout pregnancy (Fig. 2, middle column, arrowed; data not shown for the first trimester). Under the highpower magnification GALNT2 was shown to be detected in the Golgi apparatus of EVT (Fig. 2, right column), as expected. These data suggest that GALNT2 is expressed in subpopulations of EVT cells and the percentage of GALNT2-positive EVT cells is significantly increased in the second and third trimester placenta compared with the first trimester. To further verify the increase in the percentage of GALNT2positive EVT cells in late pregnancy, double staining of the first and third trimester deciduas with GALNT2 and CK7 antibodies was
3.3. Overexpression of GALNT2 in human EVT HTR8/SVneo cells
3.4. GALNT2 overexpression suppresses EVT migration and invasion To investigate effects of GALNT2 on EVT cells, we analyzed cell growth, adhesion, migration, and invasion in mock and GALNT2transfected HTR8/SVneo cells. Our data showed that GALNT2 did not affect cell growth observed for 3 days (Fig. 5A). Interestingly, we found that GALNT2 overexpression increased cell adhesion to
Fig. 5. Overexpression of GALNT2 inhibits EVT migration and invasion. (A) Effect of GALNT2 overexpression on EVT cell growth. Cell growth of Mock and GALNT2-overexpressing HTR8/SVneo cells was analyzed by trypan blue exclusion assays for 72 h (B) GALNT2 overexpression enhances cell-ECM adhesion of HTR8/SVneo cells. Left panel, cells adhered to 96-well plates coated with 5 mg/ml of collagen type IV (Col IV), fibronectin (Fn), and laminin (Ln) were shown. Right panel, GALNT2-enhanced cell adhesion to collagen IV is suppressed by b1-integrin blocking antibody, but not control IgG. (C) GALNT2 inhibits EVT cell migration. Cell migration of mock and GALNT2 transfectants was analyzed by transwell migration assays. Representative images are shown. Magnification is 200. (D) GALNT2 inhibits EVT cell invasion. Cell invasion of mock and GALNT2 transfectants was analyzed by matrigel invasion assays. Representative images are shown. For migration and invasion assays, 10% FBS were used as chemoattractants. Results are presented as means SD from three independent experiments. *P < 0.05, compared with mock.
1010
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
Fig. 6. Effects of GALNT2 on glycosylation and signaling of b1 integrin. (A) GALNT2 modifies O-linked glycans on b1 integrins. b1 integrins in HTR8/SVneo transfectants were pulled down by VVA-agarose beads and then immunoblotted with anti-b1 integrin polyclonal antibody. The b1 integrins pulled down by VVA were quantified and normalized to total b1 integrins. (B) GALNT2 inhibits the phosphorylation of FAK. Cells were seeded on collagen IV (5 mg/ml)-coated plates for 30 min and then harvested for Western blotting. Cell lysates were immunoblotted with anti-p-FAK (pY397) antibody or anti-total FAK antibody. Relative ratios of p-FAK/FAK were quantified and shown. Results are presented as means SD from three independent experiments. *P < 0.05.
collagen IV, but not fibronectin and laminin under serum free conditions (Fig. 5B, left panel). Moreover, GALNT2-increased cell adhesion to collagen IV was significantly suppressed by b1-integrin functional blocking antibody (Fig. 5B, right panel), suggesting that b1-integrin is involved in this process. In addition, we found that GALNT2 significantly suppressed cell migration (Fig. 5C) and invasion (Fig. 5D). These findings suggest that GALNT2 overexpression suppresses EVT migration and invasion. 3.5. GALNT2 regulates O-glycosylation of b1 integrin in EVT cells We found that GALNT2 was able to regulate EVT cell-ECM adhesion as well as cell migration and invasion. In addition, it has been demonstrated that the ECM receptor b1 integrin plays an important role in cell migration and invasion. We therefore examined whether O-glycans on the b1 integrin and its downstream signaling were modulated by GALNT2. We found that GALNT2 overexpression increased the amount of b1 integrins pulled down by Vicia Villosa agglutinin (VVA) lectins, which recognize the GalNAc attached to a serine or threonine residue (Fig. 6). To examine whether changes in the O-glycosylation of b1 integrin by GALNT2 could affect its downstream signaling, we analyzed the phosphorylation status of FAK, a critical downstream signaling molecule of b1 integrin. We found that GALNT2 overexpression significantly suppressed the phosphorylation of FAK (Fig. 6B). These findings suggest that GALNT2 can modulate Oglycans on b1 integrin and affect its downstream signaling. 4. Discussion The ECM receptor b1 integrin has been shown to regulate EVT invasion [19]. We previously showed that b1 integrin plays a crucial
role in mediating invasion of HTR8/SVneo cells [25]. Here we found that GALNT2 inhibits HTR8/SVneo cell invasion and suppresses phosphorylation of FAK, which is a critical downstream signaling molecule of b1 integrin and is able to mediate EVT cell migration and invasion [26]. In addition, GALNT2-enhanced cell-collagen IV adhesion under serum free conditions, which was significantly blocked by b1 integrin blocking antibody. These results suggest that GALNT2 inhibits trophoblast migration and invasion through increased cell-ECM adhesion via b1 integrin. Changes in the terminal sialic acids of O-glycans on b1 integrin by ST6GalNAc I can regulate b1 integrin activity and cell behaviors [20]. However, it is still unclear which GALNT enzymes initiate the synthesis of the O-glycans. In this study, we are the first to show that GALNT2 can enhance the expression of Tn antigen, a short Oglycan, on b1 integrin and modulates its downstream signaling. The mechanism we propose here is that the glycosylation change in b1 integrin caused by GALNT2 could induce conformational change of b1 integrin to enhance ECM binding, which impairs focal adhesion turnover and suppresses phosphorylation of FAK, and thereby inhibits trophoblast migration and invasion. GALNT2 may have other acceptor substrates, such as mucins and EGFR. Thus, it remains possible that GALNT2 may inhibit EVT invasion via multiple pathways in vivo. In addition to trophoblasts, effects of glycosylation on integrin functions have also been reported in cancer cells. Overexpression of N-acetylglucosaminyltransferase III (GnT-III) increases the bisected N-glycans on integrins and results in an inhibition of integrin-mediated cell migration and the cellular phosphorylation levels [21]. We used a transformed trophoblast cell line, HTR8/SVneo, derived from first trimester trophoblast rather than primary first trimester trophoblasts. Despite the widespread use of this cell line for migration and invasion assays in vitro, it still has to be noted that its gene expression patterns, such as GALNT family genes, and biological properties may be somewhat different from first trimester trophoblasts in vivo. Carbohydrates play critical roles in biological functions mediated through highly O-glycosylated mucins. We found that GALNT2 is the major initiating enzyme for the synthesis of mucin-type O-glycans in the first trimester HTR8/SVneo EVT cells. In addition, GALNT2 is expressed by subpopulations of EVT cells in human placental deciduas and its expression increased with gestational ages. This observation is consistent with our previous finding in that MUC1positive EVT cells also increased during placental development [22]. GALNT2 has been demonstrated to glycosylate MUC1-derived peptides [27]. These data therefore suggest that GALNT2 could be one of the enzymes that initiate O-glycan synthesis on MUC1 in EVT. The suppressive role of GALNT2 and MUC1 [22] in invasion and a higher percentage of EVT cells expressing GALNT2 and MUC1 [22] in later pregnancy suggest that they cooperate to regulate EVT behaviors during placental development. In addition to mucins, we previously found that GALNT2 can modify O-glycans on EGFR in hepatocellular carcinoma [11]. Here we identify b1 integrin as a novel acceptor substrate for GALNT2, suggesting that GALNT2 can modify O-glycans on multiple targets to control EVT invasion. In syncytiotrophoblasts, GALNT2 expression could not be detected throughout pregnancy. In sharp contrast, MUC1 and MUC15 are highly expressed in late pregnancy [22,28]. Their molecular masses indicate that MUC1 and MUC15 are highly O-glycosylated [22,28], suggesting that they are glycosylated by other GALNT family enzymes instead of GALNT2 in syncytiotrophoblasts of late pregnancy. Although it is expected that expressions of GALNT family genes are strictly regulated in a spatiotemporal manner and play important roles in placental development and functions, their expression patterns during placental development remain largely unknown. This study has shed light on understanding the expression and function of O-glycosylation and GALNT genes in human
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
placenta. However, we still need to mention that the second trimester placentas used in this study were obtained from pregnancies complicated by preterm labor. The influence of preterm labor on GALNT2 expression remains unclear. Here, we identified a role for GALNT2 in regulating invasive behavior of EVT. We showed that amongst the 20 GALNT family genes, GALNT1 and GALNT2 mRNA were expressed at high levels in HTR8/SVneo and first trimester EVT cells. Immunohistochemistry revealed that GALNT2 was highly expressed by subpopulations of EVT cells in deciduas, and the percentage of GALNT2-positive EVT cells was higher in later pregnancy. GALNT2 overexpression affected cell adhesion, migration, and invasion of HTR8/SVneo cells. The mechanistic investigation showed that GALNT2 can modulate O-glycans on b1 integrin and suppress phosphorylation of FAK. These findings highlight an attractive mechanism by which the Oglycan initiating enzyme GALNT2 modulates the glycosylation and signaling of b1 integrin, which in turn regulates EVT behaviors. Our findings open a novel insight into the molecular mechanism of the role of O-glycosylation in EVT invasion. It will be of great interest to further investigate the physiopathologic roles of GALNT family genes during placental development. Conflict of interest No potential conflicts of interest were disclosed. Acknowledgments This study was supported by the grants from the National Science Council NSC101-2314-B-002-058 (to Dr. Hung JS), NSC982320-B-002-032-MY3 and NSC101-2320-B-002-007-MY3 (to Dr. Huang MC), NSC101-2320-B-040-003 (to Dr. Liao WC) and NSC962314-B-002-089-MY3 (to Dr. Shyu MK), as well as from National Taiwan University Hospital NTUH100S-1625 (to Dr. Shyu MK). References [1] Zhou Y, Damsky CH, Chiu K, Roberts JM, Fisher SJ. Preeclampsia is associated with abnormal expression of adhesion molecules by invasive cytotrophoblasts. J Clin Invest 1993;91(3):950e60. [2] Gerretsen G, Huisjes HJ, Elema JD. Morphological changes of the spiral arteries in the placental bed in relation to pre-eclampsia and fetal growth retardation. Br J Obstet Gynaecol 1981;88(9):876e81. [3] Ball E, Bulmer JN, Ayis S, Lyall F, Robson SC. Late sporadic miscarriage is associated with abnormalities in spiral artery transformation and trophoblast invasion. J Pathol 2006;208(4):535e42. [4] Kim YM, Bujold E, Chaiworapongsa T, Gomez R, Yoon BH, Thaler HT, et al. Failure of physiologic transformation of the spiral arteries in patients with preterm labor and intact membranes. Am J Obstet Gynecol 2003;189(4):1063e9. [5] Hakomori S. Glycosylation defining cancer malignancy: new wine in an old bottle. Proc Natl Acad Sci U S A 2002;99(16):10231e3. [6] Tian E, Ten Hagen KG. Recent insights into the biological roles of mucin-type O-glycosylation. Glycoconj J 2009;26(3):325e34. [7] Tarp MA, Clausen H. Mucin-type O-glycosylation and its potential use in drug and vaccine development. Biochim Biophys Acta 2008;1780(3):546e63.
1011
[8] Ten Hagen KG, Fritz TA, Tabak LA. All in the family: the UDP-GalNAc: polypeptide N-acetylgalactosaminyltransferases. Glycobiology 2003;13(1): 1Re16R. [9] Tenno M, Ohtsubo K, Hagen FK, Ditto D, Zarbock A, Schaerli P, et al. Initiation of protein O glycosylation by the polypeptide GalNAcT-1 in vascular biology and humoral immunity. Mol Cell Biol 2007;27(24):8783e96. [10] Holleboom AG, Karlsson H, Lin RS, Beres TM, Sierts JA, Herman DS, et al. Heterozygosity for a loss-of-function mutation in GALNT2 improves plasma triglyceride clearance in man. Cell Metab 2011;14(6):811e8. [11] Wu YM, Liu CH, Hu RH, Huang MJ, Lee JJ, Chen CH, et al. Mucin glycosylating enzyme GALNT2 regulates the malignant character of hepatocellular carcinoma by modifying the EGF receptor. Cancer Res 2011;71(23):7270e9. [12] Freire-de-Lima L, Gelfenbeyn K, Ding Y, Mandel U, Clausen H, Handa K, et al. Involvement of O-glycosylation defining oncofetal fibronectin in epithelialmesenchymal transition process. Proc Natl Acad Sci U S A 2011;108(43): 17690e5. [13] Park JH, Nishidate T, Kijima K, Ohashi T, Takegawa K, Fujikane T, et al. Critical roles of mucin 1 glycosylation by transactivated polypeptide N-acetylgalactosaminyltransferase 6 in mammary carcinogenesis. Cancer Res 2010; 70(7):2759e69. [14] Wagner KW, Punnoose EA, Januario T, Lawrence DA, Pitti RM, Lancaster K, et al. Death-receptor O-glycosylation controls tumor-cell sensitivity to the proapoptotic ligand Apo2L/TRAIL. Nat Med 2007;13(9):1070e7. [15] Fujiwara H, Higuchi T, Sato Y, Nishioka Y, Zeng BX, Yoshioka S, et al. Regulation of human extravillous trophoblast function by membrane-bound peptidases. Biochim Biophys Acta 2005;1751(1):26e32. [16] Burrows TD, King A, Loke YW. Trophoblast migration during human placental implantation. Hum Reprod Update 1996;2(4):307e21. [17] Damsky CH, Librach C, Lim KH, Fitzgerald ML, McMaster MT, Janatpour M, et al. Integrin switching regulates normal trophoblast invasion. Development 1994;120(12):3657e66. [18] Aplin JD, Haigh T, Jones CJ, Church HJ, Vicovac L. Development of cytotrophoblast columns from explanted first-trimester human placental villi: role of fibronectin and integrin alpha5beta1. Biol Reprod 1999;60(4):828e38. [19] Harris LK, Jones CJ, Aplin JD. Adhesion molecules in human trophoblast e a review. II. Extravillous trophoblast. Placenta 2009;30(4):299e304. [20] Clement M, Rocher J, Loirand G, Le Pendu J. Expression of sialyl-Tn epitopes on beta1 integrin alters epithelial cell phenotype, proliferation and haptotaxis. J Cell Sci 2004;117(Pt 21):5059e69. [21] Isaji T, Kariya Y, Xu Q, Fukuda T, Taniguchi N, Gu J. Functional roles of the bisecting GlcNAc in integrin-mediated cell adhesion. Methods Enzymol 2010; 480:445e59. [22] Shyu MK, Lin MC, Liu CH, Fu YR, Shih JC, Lee CN, et al. MUC1 expression is increased during human placental development and suppresses trophoblastlike cell invasion in vitro. Biol Reprod 2008;79(2):233e9. [23] Kliman HJ, Nestler JE, Sermasi E, Sanger JM, Strauss 3rd JF. Purification, characterization, and in vitro differentiation of cytotrophoblasts from human term placentae. Endocrinology 1986;118(4):1567e82. [24] Graham CH, Hawley TS, Hawley RG, MacDougall JR, Kerbel RS, Khoo N, et al. Establishment and characterization of first trimester human trophoblast cells with extended lifespan. Exp Cell Res 1993;206(2):204e11. [25] Shyu MK, Chen CW, Lin NY, Liao WC, Chen CH, Lin CJ, et al. MUC1 expression is elevated in severe preeclamptic placentas and suppresses trophoblast cell invasion via beta1-integrin signaling. J Clin Endocrinol Metab 2011;96(12): 3759e67. [26] Chakraborty C, Gleeson LM, McKinnon T, Lala PK. Regulation of human trophoblast migration and invasiveness. Can J Physiol Pharmacol 2002;80(2): 116e24. [27] Wandall HH, Hassan H, Mirgorodskaya E, Kristensen AK, Roepstorff P, Bennett EP, et al. Substrate specificities of three members of the human UDPN-acetyl-alpha-D-galactosamine:Polypeptide N-acetylgalactosaminyltransferase family, GalNAc-T1, -T2, and -T3. J Biol Chem 1997;272(38): 23503e14. [28] Shyu MK, Lin MC, Shih JC, Lee CN, Huang J, Liao CH, et al. Mucin 15 is expressed in human placenta and suppresses invasion of trophoblast-like cells in vitro. Hum Reprod 2007;22(10):2723e32.
Contents lists available at SciVerse ScienceDirect
Placenta journal homepage: www.elsevier.com/locate/placenta
Expression of GALNT2 in human extravillous trophoblasts and its suppressive role in trophoblast invasion W.-C. Liao a,1, C.-H. Chen b, c,1, C.-H. Liu c, d, M.-J. Huang c, d, C.-W. Chen c, J.-S. Hung g, C.-H. Chou c, C.-H. Chen c, M.-I. Che c, H.-M. Chang a, C.-T. Lan a, H.-C. Huang e, G.-F. Tseng f, M.-K. Shyu b, *, M.-C. Huang c, d, ** a
Department of Anatomy, Faculty of Medicine, Chung Shan Medical University, Taichung, Taiwan Department of Obstetrics and Gynecology, National Taiwan University Hospital, 7 Chung-Shan South Road, Taipei 100, Taiwan Graduate Institute of Anatomy and Cell Biology, National Taiwan University College of Medicine, Taipei 100, Taiwan d Research Center for Developmental Biology and Regenerative Medicine, National Taiwan University, Taipei 100, Taiwan e Animal Technology Institute Taiwan, Miaoli, Taiwan f Department of Anatomy, College of Medicine, Tzu-Chi University, Hualien, Taiwan g Department of Surgery, National Taiwan University Hospital, Taipei 100, Taiwan b c
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 30 August 2012
Extravillus trophoblast (EVT) invasion plays a critical role in placental development. Integrins bind to extracellular matrix (ECM) proteins to mediate EVT cell adhesion, migration, and invasion. Changes in Oglycans on b1-integrin have been found to regulate cancer cell behavior. We hypothesize that O-glycosyltransferases can regulate EVT invasion through modulating the glycosylation and function of b1integrin. Here, we found that the GALNT1 and GALNT2 mRNA were highly expressed in HTR8/SVneo and first trimester EVT cells. Immunohistochemstry and immunofluorescence staining showed that GALNT2 was expressed in subpopulations of EVT cells in deciduas, but not in syncytiotrophoblasts and cytotrophoblasts of placental villi. The percentage of GALNT2-positive EVT cells increased with gestational ages. Overexpression of GALNT2 in HTR8/SVneo cells significantly enhanced cell-collagen IV adhesion, but suppressed cell migration and invasion. Notably, we found that GALNT2 increased the expression of Tn antigen (GalNAc-Ser/Thr) on b1-integrin as revealed by Vicia Villosa agglutinin (VVA) binding. Furthermore, GALNT2 suppressed the phosphorylation of focal adhesion kinase (FAK), a crucial downstream signaling molecule of b1-integrin. Our findings suggest that GALNT2 is a critical initiating enzyme of O-glycosylation for regulating EVT invasion. Ó 2012 Published by Elsevier Ltd.
Keywords: GALNT Mucin Integrin O-glycans EVT Invasion
1. Introduction During placental development, cytotrophoblasts in placental villi differentiate into syncytiotrophoblast and extravillous trophoblast (EVT). Interstitial invasion of EVT through endometrial decidua is crucial for successful pregnancy. A stringent regulation of EVT invasion is necessary for proper blood vessel remodeling at the maternalefetal interface and is crucial for maintaining a normal pregnancy. Failure of this remodeling process is associated with
* Corresponding author. Tel.: þ886 2 23123456x71537. ** Corresponding author. Graduate Institute of Anatomy and Cell Biology, National Taiwan University College of Medicine, No. 1, Sec. 1 Jen-Ai Road, Taipei 100, Taiwan. Tel.: þ886 2 23123456x88177; fax: þ886 2 23915292. E-mail addresses: [email protected] (M.-K. Shyu), [email protected] (M.-C. Huang). 1 These authors contributed equally to this article. 0143-4004/$ e see front matter Ó 2012 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.placenta.2012.08.007
pre-eclampsia [1], intrauterine growth restriction [2], late miscarriage [3], and preterm delivery [4]. Glycosylation is the most common post-translational modification of proteins. Aberrant glycosylation affects many cellular properties, including, cell proliferation, differentiation, transformation, migration, invasion, apoptosis, and immune responses [5]. N-linked and O-linked glycans are the two major types of carbohydrates in mammalian cells. The mucin-type O-glycosylation is initiated by the transfer of UDP-N-acetylgalactosamine (UDPGalNAc) to a serine (S) or threonine (T) residue to form Tn antigen (GalNAca-S/T) [6]. This reaction is catalyzed by a family of polypeptide GalNAc transferases (GALNTs), consisting of 20 members in humans, namely, GALNT1 to 14 and GALNTL1 to L6 [7,8]. O-linked glycans have been shown to regulate a variety of biological functions and contribute to the development of human diseases. Loss of GALNT1 activity in mice results in bleeding disorder [9]. GALNT2 can affect lipid metabolism through apoC-III glycosylation [10]. In
1006
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
addition, GALNT2 has been shown to regulate EGF-induced malignant phenotypes in hepatocellular carcinoma [11]. O-glycosylation of fibronectin regulated by GALNT3 and GALNT6 modulates epithelialemesenchymal transition [12]. GALNT6 modifies mucin 1 glycosylation and regulates proliferation of breast cancer cells [13]. Death-receptor O-glycosylation controls tumor-cell sensitivity to the proapoptotic ligand Apo2L/TRAIL [14]. Integrins, the major cell surface receptors mediating cellextracellular matrix (ECM) interactions, consist of 18 a subunits and 8 b subunits, which assemble into at least 24 ab heterodimers. More than half of these heterodimers contain b1 integrins. Functional studies have demonstrated the importance of a1b1 integrins in the regulation of EVT invasion into the uterine wall [15e17]. The b1 integrins have also been reported to modulate trophoblast motility during EVT invasion [18,19]. Both N-glycans and O-glycans are present on b1 integrins [20,21]. Changes in these carbohydrates can modulate b1 integrin activity and cell phenotypes including, cell growth, adhesion, differentiation, migration and invasion [20,21]. Identification of the glycosyltransferases responsible for the glycosylation of b1 integrins during EVT invasion will therefore contribute to our understanding of the development of EVT invasion. Although O-glycosylation has been demonstrated to regulate a variety of cellular behaviors, the expression and role of O-glycan initiating enzymes in human placenta have never been studied. Here we report that GALNT2 is highly expressed in subpopulations of EVT cells in human placenta. Moreover, GALNT2 overexpression modulates the O-glycosylation of b1 integrins and suppresses integrin-mediated signaling and invasion in HTR8/SVneo EVT cells. 2. Materials and methods 2.1. Clinical tissue collection Human placental tissues of the first trimester (7e12 wk gestation; n ¼ 10), second trimester (17e25 wk gestation; n ¼ 6), and third trimester (36e39 wk gestation; n ¼ 13) were obtained from the Department of Obstetrics and Gynecology, National Taiwan University Hospital, as previously described [22]. All the pregnancies were confirmed by ultrasound. The gestational age is dated by crownrump length before 14 weeks of gestation and bi-parietal diameter after 14 weeks. The placentas of first trimester were taken from the samples received by dilatation and curettage of viable pregnancies, which were decided to terminate due to unwillingness to keep the pregnancy further. The placentas of second trimester were obtained from preterm labors, where equal in number of vaginal delivery or cesarean section. The placentas of third trimester were all received from cesarean sections of normal pregnancies which were scheduled for section due to obstetric indications before labor pain onset. No major complications of these cases were noted during labor or postpartum. The use of human placentas for this study was approved by the local hospital ethics committee, and written consent was obtained from patients before sample collection. Specimens were fixed in 4% (w/v) paraformaldehyde/PBS. Serial 5-mm of paraffin-embedded placental sections were used for immunohistochemistry. Serial 14-mm cryostat sections were used for immunofluorescence microscopy. For RNA extraction, specimens were soaked in RNA later (Qiagen, Valencia, CA) at 4 C overnight and then stored at 20 C. For Western blot analysis, the samples were stored at 80 C until use. 2.2. Isolation of primary extravillous trophoblast cells Placenta collection and primary EVT cells were isolated as described by Kliman et al. [23]. Primary EVT cells were isolated from first trimester placentas (5e12 weeks, n ¼ 4) and third trimester placentas (36e38 weeks, n ¼ 3). Briefly, minced fragments of placentas were extensively washed and then subjected to three cycles of digestion with 1.25 mg/ml of trypsin (Life Technologies, Burlington, Canada) and 100 mg/ml of DNase (Sigma, St. Louis, MO), followed by a discontinuous 5% (5%e70%) Percoll (Sigma) gradient centrifugation. EVT-containing cell suspensions were recovered from 35 to 50% gradient layers and the purity of EVT was higher than 90%, analyzed by immunofluorescence staining using anti-cytokeratin (CK)7 antibody (Santa Cruz, CA).
SVneo cells were grown in RPMI 1640 (HyClone Laboratories, Logan, UT) supplemented with 5% fetal bovine serum (FBS) (PAA Laboratories, Pasching, Austria) in a humidified tissue culture incubator at 37 C in 5% CO2 atmosphere. For transfection, 6 105 of HTR8/SVneo cells were transiently transfected with 10 mg of empty vector (Mock) or GALNT2/pcDNA3.1/myc-His [11] with Lipofectamine 2000 (Invitrogen). After 72 h of transfection, cells were harvested for experiments. 2.4. RNA extraction and cDNA synthesis Trizol reagent (Invitrogen, Carlsbad, CA) was used to extract total RNA according to the manufacturer’s protocol. StrataScript reverse transcriptase (Stratagene, La Jolla, CA) was used for RT-PCR. 2.5. Quantitative real-time PCR Quantitative PCR system Mx3000P (Stratagene) and Brilliant SYBR Green QPCR Master Mix (Stratagene) were used to analyze gene expression in human placentas according to the manufacturer’s protocol. PCR primers used in this study were as previously described [11]. Samples were analyzed in triplicate, and product purity was checked through dissociation curve at the end of real-time PCR cycles. Relative quantity of specific gene expression normalized to b-actin was analyzed with MxPro Software (Stratagene). 2.6. Immunohistochemistry and immunofluorescence staining 2.6.1. Paraffin-embedded human placental sections were de-paraffinized in xylene and re-hydrated in a series of graded alcohol. The sections were incubated with primary antibodies (1:100 for anti-GALNT2, Sigma; 1:100 for anti-cytokeratin (CK) 7 mAb, Santa Cruz, CA) diluted in 5% (w/v) non-fat milk/PBS for 16 h at 4 C. Negative controls were performed by replacing primary antibodies with an isotype-matched control IgG at the same concentration. Super SensitiveÔ Link-Label IHC detection System (BioGenex, San Ramon, CA) was used and the specific immunostaining was visualized with 3,3-diaminobenzidine (DAB) liquid substrate system (Sigma). All sections were counterstained with hematoxylin for mounted with UltraKitt (J.T. Baker, Deventer, Holland). For immunofluorescence staining, cryo-sections were incubated with anti-GALNT2 (1:100) and anti-CK7 (1:100) antibodies for 1 h at 4 C, followed by incubation with FITC-conjugated anti-rabbit IgG (1:200; Vector, Burlingame, CA) and Cy3-conjugated anti-rabbit IgG (1:400; Jackson ImmunoResearch, West Grove, PA). For immunofluorescence microscopy of HTR8/SVneo cells, primary antibodies anti-GALNT2 and anti-GM130 (1:100; BD Biosciences, San Jose, CA) and secondary antibodies FITC- or Cy3-conjugated anti-rabbit IgG were used. Cell nuclei were counterstained by DAPI (Sigma). Immunofluorescence micrographs were obtained at 400 magnification using Leica TCS SP5 Spectral Confocal System (Leica, Wetzlar, Germany). 2.7. Cell growth analysis Cells (4 104) were seeded in 6-well plates with RPMI 1640 containing 10% FBS (PAA Laboratories). Viable cells were determined at 24 h intervals for 72 h using hemocytometer with trypan blue exclusion. 2.8. Cell adhesion assay Ninety-six-well plates were coated with 5 mg/ml of BSA, human collagen IV (Sigma), human fibronectin (Sigma), or murine laminin (Sigma) at 4 C for overnight. Cells (2 104) were seeded and allowed to attach for 30 min at 37 C in a humidified 5% CO2 incubator. For functional blocking of cell adhesion, cells (2 104) were preincubated with b1-integrin blocking antibody (1:200; clone P4C10, Millipore, Temecula, CA) or control IgG for 30 min and then seeded into 96-well plates coated with collagen IV. Numbers of attached cells from four wells were determined under an inverted microscope. 2.9. Migration and invasion assays For migration assays, chambers with 8-mm pore size membrane (Corning, Lowell, MA) were used. Cells (3 105) in 500 ml serum free DMEM were seeded to the upper part of the chamber and 10% FBS was used as chemoattractant in the lower part. Cells were allowed to migrate for 30 h in a humidified tissue culture incubator at 37 C, 5% CO2 atmosphere. The migrated cells were fixed with 100% methanol and stained with 0.5% crystal violet. The numbers of migrated cells in each well were counted under a phase contrast microscope. Cell invasion assays were performed in BioCoatÔ MatrigelÔ Invasion Chambers (BectoneDickinson, Bedford, MA) according to the manufacturer’s protocol with similar conditions.
2.3. Cell culture and transfection 2.10. Western blot analysis and lectin pull down assay Human extravillous trophoblast cell line HTR8/SVneo derived from the first trimester placenta [24] was a gift from Dr. Charles H. Graham (Department of Anatomy and Cell Biology, Queen’s University at Kingston, Ontario, Canada). HTR8/
HTR8/SVneo cells were lyzed in lysis buffer containing 1% (v/v) Triton X-100, 20 mM TriseHCl (pH 8.0), 160 mM NaCl, 1 mM CaCl2, and 1 mM PMSF, followed by
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
1007
ultrasonication (3 5-s burst). Forty micrograms of total proteins were separated on an 8% SDS-PAGE and transferred to a Hybond enhanced chemiluminescence nitrocellulose membrane (GE Healthcare, Chalfont St Giles, UK). The membrane was incubated with rabbit anti-GALNT2 antibody (1:100, Sigma), rabbit anti-FAK polyclonal antibody (C-20, Santa Cruz), rabbit anti-FAK pY397 mAb (1:1000; Biosouce, Nivelles, Belgium), or mouse anti-GAPDH mAb (BD Pharmingen). Bands were visualized by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies (1:10,000 dilution; Vector laboratories) and enhanced chemiluminescence reagents (GE Healthcare, UK). The band signals were quantified and normalized to their control signals with Image Quant 5.1 software (Molecular Dynamics, Sunnyvale, CA, USA). For detection of glycoproteins decorated with the Tn antigen, biotinylated Vicia villosa agglutinin (VVA) (Vector Laboratories, Burlingame, CA) was used. For lectin pull down assays, 200 mg of cell lysates were incubated with VVA-agarose beads (Vector Laboratories, Burlingame, CA) overnight at 4 C. The pulled down proteins were then subjected to Western blotting. 2.11. Statistical analysis Student’s t-test was used for statistical analyses. Data are presented as means SD. P < 0.05 is considered statistically significant.
3. Results 3.1. GALNT2 mRNA is expressed in first trimester EVTs and is increased in third trimester EVTs
Fig. 1. mRNA expression of GALNT family genes in HTR8/SVneo and primary EVT cells. (A) mRNA expression levels of 20 GALNT family genes in HTR8/SVneo and first trimester EVT cells. Primary EVT cells were isolated from first trimester placentas (n ¼ 4). The mRNA levels were analyzed by real-time RT-PCR. The signals of GALNT genes were normalized to b-actin. (B) Expression levels of GALNT2 mRNA are increased in late pregnancy. Expression levels of GALNT2 in primary EVT cells from first (n ¼ 4) and third (n ¼ 3) trimester placentas were compared. Results are presented as the mean SD from three independent experiments. *P < 0.05, compared with the first trimester EVT.
O-glycosylation regulates a wide range of cellular behaviors. To investigate the roles of O-glycosylation in EVT cells, we first analyzed the mRNA expression of 20 GALNT family genes in the first trimester EVT cell line HTR8/SVneo and primary EVT cells isolated from first trimester placentas (n ¼ 4). Our results showed that GALNT1 and GALNT2 were expressed at the highest levels in both HTR8/SVneo and first trimester EVT cells, compared with other GALNT members (Fig. 1A). Since we previously found that GALNT2 is able to regulate cancer cell invasion [11], we first investigated the expression and role of GALNT2 in EVT cells. To know whether GALNT2 was differentially expressed in EVT cells from different
Fig. 2. Immunohistochemistry of GALNT2 in the human placenta. Human placentas from the first, second, and third trimester were immunostained with anti-CK7 mAb (left panel) or anti-GALNT2 polyclonal antibody (middle panel). Amplified images in rectangles of GALNT2 staining were shown in the right panel. CK7 is an EVT marker. The percentage of GALNT2-positive EVT cells (GALNT2/CK7 double positive cells) were shown in the middle panel. Results are presented as the mean SD from five different fields of at least three placentas for each trimester. *P < 0.05, compared with the first trimester. Placental villi did not show GALNT2 staining under our experimental conditions (arrows). The negative control did not show any signals (lower left of the left panel, 1st trimester). Cell nuclei were counterstained with hematoxylin. Scale bars ¼ 25 mm.
1008
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
Fig. 3. Immunofluorescence staining of GALNT2 and CK7 in placentas. (A) GALNT2 expression in EVT cells. The first (10 weeks) and third trimester (37 weeks) deciduas were immunostained with anti-CK7 mAb (green) and anti-GALNT2 polyclonal antibody (red). Representative images were shown. Cell nuclei were counterstained with DAPI (blue). Merged images reveal co-localization of GALNT2 and CK7. Scale bar ¼ 25 mm. Negative controls did not show signals (data not shown). (B) Percentage of GALNT2-positive EVT in deciduas. The percentage of GALNT2-positive EVT cells were calculated from the merged images at five different fields. Results are presented as the mean SD from three independent experiments. *P < 0.05, compared with the first trimester.
Fig. 4. Overexpression of GALNT2 in HTR8/SVneo cells. (A) Overexpression of GALNT2in HTR8/SVneo cells analyzed by Western blotting. GAPDH is an internal control. The relative intensity of signals is presented as the mean SD (n ¼ 3). *P < 0.05. (B) Immunofluorescence microscopy showed overexpression of GALNT2 (green) in >80% of the HTR8/SVneo cells transiently transfected with GALNT2 plasmids for 72 h. Nuclei were counterstained with DAPI (blue). The Golgi apparatus was stained with anti-GM130 antibody (red). Merged images reveal co-localization (yellow) of GALNT2 and the Golgi apparatus. Amplified images in rectangles were shown in the lower right. Scale bar ¼ 25 mm (C) GALNT2 overexpression enhances Vicia Villosa agglutinin (VVA) binding to glycoproteins. Equal amounts of cell lysates from Mock and GALNT2 transfectants of HTR8/SVneo cells were blotted with biotinylated VVA. GAPDH is a loading control.
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
1009
gestational stage, primary EVT cells were also isolated from third trimester placentas (n ¼ 3). We found that GALNT2 mRNA expression was significantly increased in third trimester EVT cells compared with that in first trimester (Fig. 1B). These findings suggest that GALNT2 is expressed EVT cells and may play a role in regulating EVT behavior.
performed. Consistent with the results from immunohistochemistry, immunofluorescence staining revealed a higher percentage of GALNT2-positive EVT cells in the third trimester placenta compared with the first trimester placenta (Fig. 3).
3.2. Immunohistochemistry and immunofluorescence of GALNT2 in human placentas
We observed that a higher percentage of EVT cells expressed GALNT2 in the third trimester placenta than those in the first trimester. To investigate effects of GALNT2 expression on EVT cells, we first overexpressed GALNT2 in the human first trimester EVT cell line, HTR8/SVneo. HTR8/SVneo cells were transiently transfected with GALNT2 for 72 h and its overexpression was verified by Western blotting (Fig. 4A). Immunofluorescence microscopy showed that GALNT2 was overexpressed in most HTR8/SVneo EVT cells. GALNT2 subcellular localization was found in the Golgi, as expected for glycosyltransferases (Fig. 4B). To confirm that GALNT2 exhibited enzymatic activity of GalNAc transferase, biotinylated VVA was used to detect GalNAc on glycoproteins. Our data showed increased GalNAc expression on cellular proteins in GALNT2overexpressing HTR8/SVneo cells compared with mocktransfected cells (Fig. 4C).
To examine the expression pattern of GALNT2 in the human placenta, immunohistochemistry of placentas from different gestational stages was performed. CK7 is a marker for EVT (Fig. 2, left panel). In deciduas of the first trimester placenta, GALNT2 was expressed in 30% of EVT cells (Fig. 2, upper panel), which is calculated as GALNT2-positive cells/CK7-positive cells 100%. In deciduas of the second (Fig. 2, middle panel) and third (Fig. 2, lower panel) trimester placenta, GALNT2 was expressed in 72% and 73% of EVT cells, respectively. By contrast, GALNT2 could not be detected in placental villi throughout pregnancy (Fig. 2, middle column, arrowed; data not shown for the first trimester). Under the highpower magnification GALNT2 was shown to be detected in the Golgi apparatus of EVT (Fig. 2, right column), as expected. These data suggest that GALNT2 is expressed in subpopulations of EVT cells and the percentage of GALNT2-positive EVT cells is significantly increased in the second and third trimester placenta compared with the first trimester. To further verify the increase in the percentage of GALNT2positive EVT cells in late pregnancy, double staining of the first and third trimester deciduas with GALNT2 and CK7 antibodies was
3.3. Overexpression of GALNT2 in human EVT HTR8/SVneo cells
3.4. GALNT2 overexpression suppresses EVT migration and invasion To investigate effects of GALNT2 on EVT cells, we analyzed cell growth, adhesion, migration, and invasion in mock and GALNT2transfected HTR8/SVneo cells. Our data showed that GALNT2 did not affect cell growth observed for 3 days (Fig. 5A). Interestingly, we found that GALNT2 overexpression increased cell adhesion to
Fig. 5. Overexpression of GALNT2 inhibits EVT migration and invasion. (A) Effect of GALNT2 overexpression on EVT cell growth. Cell growth of Mock and GALNT2-overexpressing HTR8/SVneo cells was analyzed by trypan blue exclusion assays for 72 h (B) GALNT2 overexpression enhances cell-ECM adhesion of HTR8/SVneo cells. Left panel, cells adhered to 96-well plates coated with 5 mg/ml of collagen type IV (Col IV), fibronectin (Fn), and laminin (Ln) were shown. Right panel, GALNT2-enhanced cell adhesion to collagen IV is suppressed by b1-integrin blocking antibody, but not control IgG. (C) GALNT2 inhibits EVT cell migration. Cell migration of mock and GALNT2 transfectants was analyzed by transwell migration assays. Representative images are shown. Magnification is 200. (D) GALNT2 inhibits EVT cell invasion. Cell invasion of mock and GALNT2 transfectants was analyzed by matrigel invasion assays. Representative images are shown. For migration and invasion assays, 10% FBS were used as chemoattractants. Results are presented as means SD from three independent experiments. *P < 0.05, compared with mock.
1010
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
Fig. 6. Effects of GALNT2 on glycosylation and signaling of b1 integrin. (A) GALNT2 modifies O-linked glycans on b1 integrins. b1 integrins in HTR8/SVneo transfectants were pulled down by VVA-agarose beads and then immunoblotted with anti-b1 integrin polyclonal antibody. The b1 integrins pulled down by VVA were quantified and normalized to total b1 integrins. (B) GALNT2 inhibits the phosphorylation of FAK. Cells were seeded on collagen IV (5 mg/ml)-coated plates for 30 min and then harvested for Western blotting. Cell lysates were immunoblotted with anti-p-FAK (pY397) antibody or anti-total FAK antibody. Relative ratios of p-FAK/FAK were quantified and shown. Results are presented as means SD from three independent experiments. *P < 0.05.
collagen IV, but not fibronectin and laminin under serum free conditions (Fig. 5B, left panel). Moreover, GALNT2-increased cell adhesion to collagen IV was significantly suppressed by b1-integrin functional blocking antibody (Fig. 5B, right panel), suggesting that b1-integrin is involved in this process. In addition, we found that GALNT2 significantly suppressed cell migration (Fig. 5C) and invasion (Fig. 5D). These findings suggest that GALNT2 overexpression suppresses EVT migration and invasion. 3.5. GALNT2 regulates O-glycosylation of b1 integrin in EVT cells We found that GALNT2 was able to regulate EVT cell-ECM adhesion as well as cell migration and invasion. In addition, it has been demonstrated that the ECM receptor b1 integrin plays an important role in cell migration and invasion. We therefore examined whether O-glycans on the b1 integrin and its downstream signaling were modulated by GALNT2. We found that GALNT2 overexpression increased the amount of b1 integrins pulled down by Vicia Villosa agglutinin (VVA) lectins, which recognize the GalNAc attached to a serine or threonine residue (Fig. 6). To examine whether changes in the O-glycosylation of b1 integrin by GALNT2 could affect its downstream signaling, we analyzed the phosphorylation status of FAK, a critical downstream signaling molecule of b1 integrin. We found that GALNT2 overexpression significantly suppressed the phosphorylation of FAK (Fig. 6B). These findings suggest that GALNT2 can modulate Oglycans on b1 integrin and affect its downstream signaling. 4. Discussion The ECM receptor b1 integrin has been shown to regulate EVT invasion [19]. We previously showed that b1 integrin plays a crucial
role in mediating invasion of HTR8/SVneo cells [25]. Here we found that GALNT2 inhibits HTR8/SVneo cell invasion and suppresses phosphorylation of FAK, which is a critical downstream signaling molecule of b1 integrin and is able to mediate EVT cell migration and invasion [26]. In addition, GALNT2-enhanced cell-collagen IV adhesion under serum free conditions, which was significantly blocked by b1 integrin blocking antibody. These results suggest that GALNT2 inhibits trophoblast migration and invasion through increased cell-ECM adhesion via b1 integrin. Changes in the terminal sialic acids of O-glycans on b1 integrin by ST6GalNAc I can regulate b1 integrin activity and cell behaviors [20]. However, it is still unclear which GALNT enzymes initiate the synthesis of the O-glycans. In this study, we are the first to show that GALNT2 can enhance the expression of Tn antigen, a short Oglycan, on b1 integrin and modulates its downstream signaling. The mechanism we propose here is that the glycosylation change in b1 integrin caused by GALNT2 could induce conformational change of b1 integrin to enhance ECM binding, which impairs focal adhesion turnover and suppresses phosphorylation of FAK, and thereby inhibits trophoblast migration and invasion. GALNT2 may have other acceptor substrates, such as mucins and EGFR. Thus, it remains possible that GALNT2 may inhibit EVT invasion via multiple pathways in vivo. In addition to trophoblasts, effects of glycosylation on integrin functions have also been reported in cancer cells. Overexpression of N-acetylglucosaminyltransferase III (GnT-III) increases the bisected N-glycans on integrins and results in an inhibition of integrin-mediated cell migration and the cellular phosphorylation levels [21]. We used a transformed trophoblast cell line, HTR8/SVneo, derived from first trimester trophoblast rather than primary first trimester trophoblasts. Despite the widespread use of this cell line for migration and invasion assays in vitro, it still has to be noted that its gene expression patterns, such as GALNT family genes, and biological properties may be somewhat different from first trimester trophoblasts in vivo. Carbohydrates play critical roles in biological functions mediated through highly O-glycosylated mucins. We found that GALNT2 is the major initiating enzyme for the synthesis of mucin-type O-glycans in the first trimester HTR8/SVneo EVT cells. In addition, GALNT2 is expressed by subpopulations of EVT cells in human placental deciduas and its expression increased with gestational ages. This observation is consistent with our previous finding in that MUC1positive EVT cells also increased during placental development [22]. GALNT2 has been demonstrated to glycosylate MUC1-derived peptides [27]. These data therefore suggest that GALNT2 could be one of the enzymes that initiate O-glycan synthesis on MUC1 in EVT. The suppressive role of GALNT2 and MUC1 [22] in invasion and a higher percentage of EVT cells expressing GALNT2 and MUC1 [22] in later pregnancy suggest that they cooperate to regulate EVT behaviors during placental development. In addition to mucins, we previously found that GALNT2 can modify O-glycans on EGFR in hepatocellular carcinoma [11]. Here we identify b1 integrin as a novel acceptor substrate for GALNT2, suggesting that GALNT2 can modify O-glycans on multiple targets to control EVT invasion. In syncytiotrophoblasts, GALNT2 expression could not be detected throughout pregnancy. In sharp contrast, MUC1 and MUC15 are highly expressed in late pregnancy [22,28]. Their molecular masses indicate that MUC1 and MUC15 are highly O-glycosylated [22,28], suggesting that they are glycosylated by other GALNT family enzymes instead of GALNT2 in syncytiotrophoblasts of late pregnancy. Although it is expected that expressions of GALNT family genes are strictly regulated in a spatiotemporal manner and play important roles in placental development and functions, their expression patterns during placental development remain largely unknown. This study has shed light on understanding the expression and function of O-glycosylation and GALNT genes in human
W.-C. Liao et al. / Placenta 33 (2012) 1005e1011
placenta. However, we still need to mention that the second trimester placentas used in this study were obtained from pregnancies complicated by preterm labor. The influence of preterm labor on GALNT2 expression remains unclear. Here, we identified a role for GALNT2 in regulating invasive behavior of EVT. We showed that amongst the 20 GALNT family genes, GALNT1 and GALNT2 mRNA were expressed at high levels in HTR8/SVneo and first trimester EVT cells. Immunohistochemistry revealed that GALNT2 was highly expressed by subpopulations of EVT cells in deciduas, and the percentage of GALNT2-positive EVT cells was higher in later pregnancy. GALNT2 overexpression affected cell adhesion, migration, and invasion of HTR8/SVneo cells. The mechanistic investigation showed that GALNT2 can modulate O-glycans on b1 integrin and suppress phosphorylation of FAK. These findings highlight an attractive mechanism by which the Oglycan initiating enzyme GALNT2 modulates the glycosylation and signaling of b1 integrin, which in turn regulates EVT behaviors. Our findings open a novel insight into the molecular mechanism of the role of O-glycosylation in EVT invasion. It will be of great interest to further investigate the physiopathologic roles of GALNT family genes during placental development. Conflict of interest No potential conflicts of interest were disclosed. Acknowledgments This study was supported by the grants from the National Science Council NSC101-2314-B-002-058 (to Dr. Hung JS), NSC982320-B-002-032-MY3 and NSC101-2320-B-002-007-MY3 (to Dr. Huang MC), NSC101-2320-B-040-003 (to Dr. Liao WC) and NSC962314-B-002-089-MY3 (to Dr. Shyu MK), as well as from National Taiwan University Hospital NTUH100S-1625 (to Dr. Shyu MK). References [1] Zhou Y, Damsky CH, Chiu K, Roberts JM, Fisher SJ. Preeclampsia is associated with abnormal expression of adhesion molecules by invasive cytotrophoblasts. J Clin Invest 1993;91(3):950e60. [2] Gerretsen G, Huisjes HJ, Elema JD. Morphological changes of the spiral arteries in the placental bed in relation to pre-eclampsia and fetal growth retardation. Br J Obstet Gynaecol 1981;88(9):876e81. [3] Ball E, Bulmer JN, Ayis S, Lyall F, Robson SC. Late sporadic miscarriage is associated with abnormalities in spiral artery transformation and trophoblast invasion. J Pathol 2006;208(4):535e42. [4] Kim YM, Bujold E, Chaiworapongsa T, Gomez R, Yoon BH, Thaler HT, et al. Failure of physiologic transformation of the spiral arteries in patients with preterm labor and intact membranes. Am J Obstet Gynecol 2003;189(4):1063e9. [5] Hakomori S. Glycosylation defining cancer malignancy: new wine in an old bottle. Proc Natl Acad Sci U S A 2002;99(16):10231e3. [6] Tian E, Ten Hagen KG. Recent insights into the biological roles of mucin-type O-glycosylation. Glycoconj J 2009;26(3):325e34. [7] Tarp MA, Clausen H. Mucin-type O-glycosylation and its potential use in drug and vaccine development. Biochim Biophys Acta 2008;1780(3):546e63.
1011
[8] Ten Hagen KG, Fritz TA, Tabak LA. All in the family: the UDP-GalNAc: polypeptide N-acetylgalactosaminyltransferases. Glycobiology 2003;13(1): 1Re16R. [9] Tenno M, Ohtsubo K, Hagen FK, Ditto D, Zarbock A, Schaerli P, et al. Initiation of protein O glycosylation by the polypeptide GalNAcT-1 in vascular biology and humoral immunity. Mol Cell Biol 2007;27(24):8783e96. [10] Holleboom AG, Karlsson H, Lin RS, Beres TM, Sierts JA, Herman DS, et al. Heterozygosity for a loss-of-function mutation in GALNT2 improves plasma triglyceride clearance in man. Cell Metab 2011;14(6):811e8. [11] Wu YM, Liu CH, Hu RH, Huang MJ, Lee JJ, Chen CH, et al. Mucin glycosylating enzyme GALNT2 regulates the malignant character of hepatocellular carcinoma by modifying the EGF receptor. Cancer Res 2011;71(23):7270e9. [12] Freire-de-Lima L, Gelfenbeyn K, Ding Y, Mandel U, Clausen H, Handa K, et al. Involvement of O-glycosylation defining oncofetal fibronectin in epithelialmesenchymal transition process. Proc Natl Acad Sci U S A 2011;108(43): 17690e5. [13] Park JH, Nishidate T, Kijima K, Ohashi T, Takegawa K, Fujikane T, et al. Critical roles of mucin 1 glycosylation by transactivated polypeptide N-acetylgalactosaminyltransferase 6 in mammary carcinogenesis. Cancer Res 2010; 70(7):2759e69. [14] Wagner KW, Punnoose EA, Januario T, Lawrence DA, Pitti RM, Lancaster K, et al. Death-receptor O-glycosylation controls tumor-cell sensitivity to the proapoptotic ligand Apo2L/TRAIL. Nat Med 2007;13(9):1070e7. [15] Fujiwara H, Higuchi T, Sato Y, Nishioka Y, Zeng BX, Yoshioka S, et al. Regulation of human extravillous trophoblast function by membrane-bound peptidases. Biochim Biophys Acta 2005;1751(1):26e32. [16] Burrows TD, King A, Loke YW. Trophoblast migration during human placental implantation. Hum Reprod Update 1996;2(4):307e21. [17] Damsky CH, Librach C, Lim KH, Fitzgerald ML, McMaster MT, Janatpour M, et al. Integrin switching regulates normal trophoblast invasion. Development 1994;120(12):3657e66. [18] Aplin JD, Haigh T, Jones CJ, Church HJ, Vicovac L. Development of cytotrophoblast columns from explanted first-trimester human placental villi: role of fibronectin and integrin alpha5beta1. Biol Reprod 1999;60(4):828e38. [19] Harris LK, Jones CJ, Aplin JD. Adhesion molecules in human trophoblast e a review. II. Extravillous trophoblast. Placenta 2009;30(4):299e304. [20] Clement M, Rocher J, Loirand G, Le Pendu J. Expression of sialyl-Tn epitopes on beta1 integrin alters epithelial cell phenotype, proliferation and haptotaxis. J Cell Sci 2004;117(Pt 21):5059e69. [21] Isaji T, Kariya Y, Xu Q, Fukuda T, Taniguchi N, Gu J. Functional roles of the bisecting GlcNAc in integrin-mediated cell adhesion. Methods Enzymol 2010; 480:445e59. [22] Shyu MK, Lin MC, Liu CH, Fu YR, Shih JC, Lee CN, et al. MUC1 expression is increased during human placental development and suppresses trophoblastlike cell invasion in vitro. Biol Reprod 2008;79(2):233e9. [23] Kliman HJ, Nestler JE, Sermasi E, Sanger JM, Strauss 3rd JF. Purification, characterization, and in vitro differentiation of cytotrophoblasts from human term placentae. Endocrinology 1986;118(4):1567e82. [24] Graham CH, Hawley TS, Hawley RG, MacDougall JR, Kerbel RS, Khoo N, et al. Establishment and characterization of first trimester human trophoblast cells with extended lifespan. Exp Cell Res 1993;206(2):204e11. [25] Shyu MK, Chen CW, Lin NY, Liao WC, Chen CH, Lin CJ, et al. MUC1 expression is elevated in severe preeclamptic placentas and suppresses trophoblast cell invasion via beta1-integrin signaling. J Clin Endocrinol Metab 2011;96(12): 3759e67. [26] Chakraborty C, Gleeson LM, McKinnon T, Lala PK. Regulation of human trophoblast migration and invasiveness. Can J Physiol Pharmacol 2002;80(2): 116e24. [27] Wandall HH, Hassan H, Mirgorodskaya E, Kristensen AK, Roepstorff P, Bennett EP, et al. Substrate specificities of three members of the human UDPN-acetyl-alpha-D-galactosamine:Polypeptide N-acetylgalactosaminyltransferase family, GalNAc-T1, -T2, and -T3. J Biol Chem 1997;272(38): 23503e14. [28] Shyu MK, Lin MC, Shih JC, Lee CN, Huang J, Liao CH, et al. Mucin 15 is expressed in human placenta and suppresses invasion of trophoblast-like cells in vitro. Hum Reprod 2007;22(10):2723e32.