BMP4 induction of trophoblast from mouse embryonic stem cells in defined culture conditions on laminin.

Because mouse embryonic stem cells (mESCs) do not contribute to the formation of extraembryonic placenta when they are injected into blastocysts, it is believed that mESCs do not differentiate into trophoblast whereas human embryonic stem cells (hESCs) can express trophoblast markers when exposed to bone morphogenetic protein 4 (BMP4) in vitro. To test whether mESCs have the potential to differentiate into trophoblast, we assessed the effect of BMP4 on mESCs in a defined monolayer culture condition. The expression of trophoblast-specific transcription factors such as Cdx2, Dlx3, Esx1, Gata3, Hand1, Mash2, and Plx1 was specifically upregulated in the BMP4-treated differentiated cells, and these cells expressed trophoblast markers. These results suggest that BMP4 treatment in defined culture conditions enabled mESCs to differentiate into trophoblast. This differentiation was inhibited by serum or leukemia inhibitory factor, which are generally used for mESC culture. In addition, we studied the mechanism underlying BMP4-directed mESC differentiation into trophoblast. Our results showed that BMP4 activates the Smad pathway in mESCs inducing Cdx2 expression, which plays a crucial role in trophoblast differentiation, through the binding of Smad protein to the Cdx2 genomic enhancer sequence. Our findings imply that there is a common molecular mechanism underlying hESC and mESC differentiation into trophoblast.

Introduction

Mouse embryonic stem cells (mESCs) are pluripotent cells derived from the inner cell mass (ICM) of the blastocyst, which differentiate into all the three germ layers in vitro and in vivo (Evans and Kaufman 1981; Martin 1981). However, mESCs are thought to be incapable of differentiating into trophoblast because they do not contribute to placenta in chimeric mouse (Beddington and Robertson 1989). To obtain placental trophoblast from mESCs, genetic manipulations of transcription factors or signaling molecules have been reported, such as the decreased expression of Oct3/4 (Niwa et al. 2000) or Sox2 (Masui et al. 2007) or overexpression of Cdx2 (Niwa et al. 2005; Tolkunova et al. 2006), Eomes (Niwa et al. 2005), Ras (Lu et al. 2008), or Tead4 (Nishioka et al. 2009). However, the fundamental molecular mechanisms regulating the differentiation of ESCs into trophoblast have not yet been elucidated. Human embryonic stem cells (hESCs) have been reported to express trophoblast markers after treatment with bone morphogenetic protein 4 (BMP4) in vitro (Xu et al. 2002). By contrast, the effect of BMP4 on mESCs is still unclear (Kunath et al. 2007). BMP4 is thought to be involved in activin- or Wnt-induced mesoderm induction or mesodermal tissue specification from ESCs (Johansson and Wiles 1995; Wiles and Johansson 1999; Nostro et al. 2008; Sumi et al. 2008). Coordinating with leukemia inhibitory factor (LIF), BMP4 also supports the mESC self-renewal in defined culture conditions (Ying et al. 2003; Qi et al. 2004).

Recently, Smith and his colleagues have suggested that to elucidate physiologically relevant molecular signals in mESCs, culture conditions with fewer extrinsic stimulators are beneficial (Ying et al. 2008). We previously developed a chemically defined simple serum-free culture condition for mESCs (Furue et al. 2005). Under these culture conditions, the effects of extracellular matrices (ECM) on mESCs were studied, and the results revealed that laminin promoted differentiation into epiblast-like cells (Hayashi et al. 2007). In this study, we assessed the effect of BMP4 on mESC in the defined culture conditions with fewer extrinsic stimulators. Our results show that mESCs can be induced to differentiate into trophoblast by BMP4 in vitro. This differentiation was inhibited by serum or LIF. Furthermore, we also found that BMP4 activates the Smad pathway in mESCs, and in turn, the BMP–Smad pathway directly induces Cdx2 expression, which plays a crucial role in trophoblast differentiation.

Materials and Methods

Cell culture.

The mESC D3 line (CRL-1934, ATCC, Manassas, VA), B6G-2 line (AES0003, RIKEN Cell Bank, Ibaraki, Japan), E14 line, or EB3 line was routinely cultured in 75-cm2 plastic flasks (Corning, Corning, NY), coated with 15 µg/ml of type I collagen (Nitta gelatin) in a humidified atmosphere of 5% CO2 at 37°C in a defined ESF7 medium. The ESF7 consisted of ESF basal medium (Cell Science & Technology Institute, Sendai, Japan) supplemented with 10 µg/ml of insulin, 5 µg/ml transferrin, 10 µM 2-mercaptoethanol, 10 µM 2-ethanolamine, 20 nM sodium selenite, 9.4 µg/ml of oleic acid conjugated with 2 mg/ml of fatty-acid-free bovine serum albumin (FAF-BSA; Sigma, St. Louis, MO), and 10 ng/ml of LIF (Chemicon, Billerica, MA), as described previously (Furue et al. 2005; Hayashi et al. 2007). For the differentiation experiments, mESCs were seeded at a density of 1 × 104 cells per square centimeter in the ESF5 medium comprised of ESF basal medium supplemented with 10 µg/ml of insulin, 5 µg/ml transferrin, 10 µM 2-mercaptoethanol, 10 µM 2-ethanolamine, 20 nM sodium selenite added with 0.5 mg/ml FAF-BSA, and 10 ng/ml rhBMP4 (R&D Systems, Minneapolis, MN) on 2 µg/cm2 laminin-coated (Sigma) dishes. The medium was changed every 2 d. After 4 d of culture under differentiating conditions, the cells reached confluency. Then, in all experiments where cells were differentiated for more than 4 d, the cells cultured for 4 d in BMP4-supplemented ESF5 medium were harvested with 0.2 mg/ml ethylenediaminetetraacetic acid (EDTA)-4Na (Sigma) for 5 min at room temperature and subcultured into BMP4-supplemented ESF5 medium at a density of about 2 × 104 cells per square centimeter. For the examination of the effects of activin and FGF4 on the differentiation of mESCs into trophoblast, the mESCs were cultured in ESF5 medium supplemented with 25 ng/ml of FGF4 (Sigma) or 10 ng/ml of activin A (Ajinomoto, Kawasaki, Japan), respectively. When examining the effect of fetal calf serum (FCS), LIF, and Noggin on the differentiation of mESCs into trophoblast, the mESCs were cultured in ESF5 medium supplemented with 10 ng/ml of LIF (Chemicon), 10% FCS (ES qualified; Gibco, Grand Island, NY), or 300 ng/ml Noggin (R&D Systems), respectively, along with BMP4.

Flow cytometry.

Flow cytometry was performed with EPICS ALTRA system (Beckman Coulter) as described previously (Furue et al. 2005; Hayashi et al. 2007). In this study, goat anti-Cdh3 antibody (R&D systems) was visualized with AlexaFluor-488-conjugated mouse antigoat IgG (Invitrogen, Carlsbad, CA). For DNA content analysis, ethanol-fixed samples were stained with propidium iodide (PI).

Immunocytochemistry.

Immunocytochemistry was performed as described previously (Furue et al. 2005; Hayashi et al. 2007). Briefly, mESCs were fixed in 4% (w/v) paraformaldehyde or ice-cold acetone, permeabilized with 0.1% Triton X-100, and reacted with primary antibodies. The primary antibodies were visualized with AlexaFluor-488-conjugated antirabbit, antimouse, or antigoat IgG or AlexaFluor-594-conjugated donkey antimouse, antirabbit, or antigoat IgG (Invitrogen). The primary antibodies used are as follows: anti-Cdx2 antibody (Biogenex, San Ramon, CA; 1:100), anti-Cdh3 antibody (R&D systems; 1:200), anti-CK7 antibody (Chemicon; 1:100), anti-Cx31 antibody (Chemicon; 1:100), anti-Nanog antibody (ReproCell, Tokyo, Japan; 1:200), and anti-SSEA1 antibody (Kyowa, Tokyo, Japan; 1:100).

Transfection.

mESCs were seeded at a density of 5 × 105 cells per well in a six-well plate coated with type I collagen in ESF7. The mESCs were transfected with plasmid DNA using Lipofectamine 2000 (Invitrogen), according to the supplier’s instructions. For inhibitory Smad overexpression experiment, Smad6 expression vector comprising whole Smad6 cDNA under CAG promoter in pCAG-IRES-PURO plasmid (a gift of Dr. Imamura) was used. The pCAG-IRES-PURO-FLAG was used as a mock. Transfected cells were reseeded in ESF5 with 10 ng/ml of BMP4 and 1 µg/ml of Puromycin, 24 h after transfection. These cells were used for immunocytochemistry and reverse transcription–polymerase chain reaction (RT–PCR), 96 h after transfection. For Cdx2-knockdown experiment, the shRNA for Cdx2 expression vector consisted of 29-mer shRNA constructs against Mouse Cdx2 under U6 promoter in pRS plasmid (purchased from OriGene, Rockville, MD). The pRS plasmid was used as a mock. Transfected cells were reseeded in ESF7 with 0.5 µg/ml of Puromycin, 24 h after transfection. After two passages on this culture conditions, these cells were used for immunofluorescence detection and RT–PCR.

Chromatin immunoprecipitation assay.

For chromatin immunoprecipitation (ChIP) assay of endogenous proteins, the mESCs (D3 line) were cultured in ESF5 medium for 48 h on laminin and treated with 100 ng/ml of BMP4 for 1 h. The cells were cross-linked with 1% formaldehyde for 10 min at 37°C. To stop the cross-linking, the samples were washed twice with ice-cold phosphate-buffered saline (PBS) with complete protease inhibitors mixture (Roche). After gentle washing with ChIP lysis buffer [1% sodium dodecyl sulfate (SDS), 10 mM EDTA, 50 mM Tris–HCl (pH 8.0), complete protease inhibitor mixture], the samples were lysed with 200 µl of ChIP lysis buffer. The lysates were then mixed with 800 µl of ChIP dilution buffer [0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris–HCl (pH 8.0), 167 mM NaCl, complete protease inhibitor mixture] and sonicated four times for 10 s each at the maximum setting (Sonifier 150; Branson, Danbury, CT). Then, 1 ml of ChIP dilution buffer was added, and the samples were centrifuged at 17,000×g for 10 min. The supernatants were transferred to a fresh centrifuge tube. An aliquot of 200 µl of the supernatant was stored at 4°C as the input DNA sample. These supernatants were conjugated with anti-phospho Smad1/5/8 antibody (Chemicon) or goat normal IgG (Upstate Biotechnology) bound with Dynabeads Protein G (Invitrogen) through overnight incubation at 4°C with rotation. These antibodies were found to adhere to the Dynabeads Protein G when 5 µg of the antibodies was incubated in 0.1 mg/ml of BSA in PBS for 1 h at room temperature with rotation. The beads were collected with Dynal MPC-S (Invitrogen) and were washed sequentially for 5 min each in low-salt buffer [0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris–HCl (pH 8.0), 150 mM NaCl], high-salt buffer [0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris–HCl (pH 8.0), 500 mM NaCl], and LiCl buffer [0.25 M LiCl, 1% Nonidet P-40, 1% deoxycholic acid, 1 mM EDTA, 10 mM Tris–HCl (pH 8.0)]. The precipitates were then washed twice with ChIP TE buffer [10 mM Tris–HCl (pH 8.0), 1 mM EDTA] for 5 min. The immunocomplex was extracted twice by incubation for 30 min at room temperature with 200 µl of ChIP elution buffer (1% SDS, 0.1 M NaHCO3, 10 mM DTT). The eluates and input DNA solutions were supplemented with 5 M NaCl to a final concentration of 200 mM and heated at 65°C for 8–12 h to reverse the formaldehyde cross-linking. All the samples were sequentially treated for 30 min with RNase I (Wako) at 37°C and for 1 h with proteinase K (Takara Bio) at 55°C. The DNA fragments were purified using the QIAquick PCR purification Kit (Qiagen, Hilden, Germany) and analyzed by PCR. The primers, annealing temperature, and number of cycles in the PCR analysis are as follows: for Cdx2 Intronic Conserved Sequence 1 (CICS1), forward, 5′-GGGCCACAGCTTCCCTACAT-3′ and reverse, 5′-TGGGTGGTCCGAGACTAGGG-3′, 60°C, 31 cycles; for DS 4 kb, forward, 5′-ATGCCAGAGCCAACCTGGAC-3′ and reverse, 5′-CTCCCGACTTCCCTTCACCA-3′, 60°C, 32 cycles; and for US 4 kb, forward, 5′-AGCCAAGGACCCTTGTTGCT-3′ and reverse, 5′–GGGGACTTGAACACCCTTCC-3′, 60°C, 32 cycles.

Electrophoretic mobility shift assay.

The proteins used in the electrophoretic mobility shift assay (EMSA) were transcribed and translated from the expression vectors pCS2-GST-Smad1 and pCS2-GST-Smad4 in BL21 E. coli strain. The probes generated from the sense and the antisense oligonucleotides were labeled with Cy5.5 (Sigma-Aldrich), mixed, and annealed. The DNA–protein binding reaction was performed in the binding buffer [20 mM HEPES (pH 7.8), 45 mM KCl, 10 mM NaCl, 1 mM EDTA, 10% (vol/vol) glycerol, 0.1% Nonidet P-40, 0.2 mg/ml BSA, 1 mM DTT] at 4°C for 1 h. After electrophoresis, the binding reactions were analyzed by using the Odyssey image reader (Li-Cor, Lincoln, NE; ALOKA, Tokyo, Japan) for the Cy5.5-labeled probe. The sense strand sequences of the probes used in EMSA are as follows: 5′-ACAAGGGCGCCCGGCGCCGACAGCGGTCTTGCCACCTCGGCGCGGGACTT-3 ′.

Luciferase assay.

The pGL4.74 (Promega) plasmid was used as an internal control. Cultured cells were transfected with the reporter vectors (pGL4.23 with Cdx2-intron1 firefly, 4 µg; Renilla, 100 ng) and were harvested 48 h after transfection. Reporter activities were measured by using the dual-luciferase reporter assay system (Promega). Each assay was performed in duplicate, and all the results presented the mean values (n = 4).

RT–PCR.

RT–PCR was performed as described previously (Furue et al. 2005; Hayashi et al. 2007). Briefly, total RNA was extracted from the cultured cells using the total RNA extraction kit (Agilent, Wilmington, DE) and reverse-transcribed using Quantitect RT kit (Qiagen). Quantitative real-time PCR was performed with SYBR Green PCR Master Mix according to the supplier’s directions (Applied Biosystems, Foster City, CA) in ABI PRISM 7700 sequence detector or Step One Plus sequence detector (Applied Biosystems). Relative expression of mRNA was calculated and compared with the expression in mouse whole-day 10.5 embryos or 12.5 placenta. Conventional PCR for the detection of trophoblast marker expression or ChIP assay was performed with SYBR Green PCR Master Mix according to the supplier’s directions. The sequences of the primers are listed in the Table 1. All the results are given as the mean values (n = 4).

Table 1.

Primer pairs used in RT–PCR

Names	Sequences	Product size	Cycles
Cdx2	5′-CTGCTGTAGGCGGAATGTATGTCT-3′	146	–
	5′-AAGGCTTGTTTGGCTCGTTACAC-3′
Dlx3	5′-TACTCGCCCAAGTCGGAATA-3′	174	–
	5′-AGTAGATCGTTCGCGGCTTT-3′
Eomes	5′-CGGCAAAGCGGACAATAACA-3′	195	–
	5′-ATGTGCAGCCTCGGTTGGTA-3′
Errb	5′-GCTGTATGCTATGCCTCCCAACG-3′	166	–
	5′-ACTCTGCAGCAGGCTCATCTGGT-3′
Esx1	5′-GAGCTGGAGGCCTTTTTCCA-3′	194	–
	5′-ACACCCACAGGGGGACTCAT-3′
Ets2	5′-CTCGGCTCAACACCGTCAAT-3′	132	–
	5′-AGCTGTCCCCACCGTTCTCT-3′
Flk1	5′-TCCTACAGACCCGGCCAAAC-3′	163	–
	5′-ACACGTTGGCAGCTTGGATG-3′
Gapdh	5′-ACCCAGAAGACTGTGGATGG-3′	173	–
	5′-CACATTGGGGGTAGGAACAC-3′
Gata3	5′-GGGCTACGGTGCAGAGGTAT-3′	163	–
	5′-TGGATGGACGTCTTGGAGAA-3′
Gcm1	5′-TACCTGAGACCCGCCATCTG-3′	152	35
	5′-AAGATGAAGCGTCCGTCGTG-3′
Hand1	5′-TCGCCGAGCTAAATGGAGAA-3′	124	–
	5′-TGCTGAGGCAACTCCCTTTT-3′
Mash2	5′-CGGGATCTGCACTCGAGGAT-3′	183	–
	5′-GGTGGGAAGTGGACGTTTGC-3′
Mixl1	5′-AAGTTGGGGAGTACACAATG-3′	195	–
	5′-CACCATACCACACATATGGA-3′
Pl1	5′-CATTTGGCTGAACTGTCTCA-3′	111	35
	5′-GACTTCCTCTCGATTCTCTG-3′
Plf	5′-AGGAACAAGCCAGGCTCACA-3′	178	35
	5′-TTCCGGACTGCGTTGATCTT-3′
Psx1	5′-CGATGGATGGGTGTGGATGA-3′	165	–
	5′-TGACAGGGCTGGCACTCAAG-3′
Sox1	5′-GTCATGTCCGAGGCCGAGAA-3′	118	–
	5′-AGCAGCGTCTTGGTCTTGCG-3′
Tpbpa	5′-AGTCCCTGAAGCGCAGTTGG-3′	138	35
	5′-TTGGAGCCTTCCGTCTCCTG-3′
Tpbpb	5′-GTCATCCTGTGCCTGGGTGT-3′	163	35
	5′-TGCCATCCTTCTCCTGGTCA-3′

Western blot.

Western blot was performed as described previously (Furue et al. 2005; Hayashi et al. 2007). Briefly, to detect the phosphorylation of Smads, mESCs were seeded at a density of 6 × 105 cells per square centimeter on six-well plates in the ESF basal medium and were stimulated using 100 ng/ml of BMP4 for 0, 15, or 60 min. The cells were lysed in 200 µl of ice-cold lysis buffer (50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 0.1% SDS, 1 mM Na3Vo4, 0.5% sodium deoxycholate, 5 mM EDTA, 1% NP-40) and 250 µl of PBS. Protein samples (25 or 50 µg) were separated in a 12.5% SDS–polyacrylamide gel and electroblotted to a polyvinylidene fluoride membrane (Amersham, Piscataway, NJ). After incubating in the blocking buffer for 30 min at room temperature, the membrane was incubated with primary antibodies overnight at 4°C. The primary antibodies used are as follows: anti-α-tubulin (Sigma, 1:1,000), anti-Smad1/5/8 (Santa Cruz Biotechnology, Santa Cruz, CA, 1:1,000), anti-phospho Smad1/5/8 (Cell Signaling Technology, Beverly, MA, 1:1,000), and anti-Smad6 (Abcam, Cambridge, UK, 1:1,000). The membranes were then reacted with secondary antibodies followed by horseradish peroxidase substrate, according to the supplier’s protocol (Pierce Biotechnology, Rockford, IL). Protein bands on the membranes were visualized with LAS-1000 and PRO-LAS 1000 software (Fujifilm, Tokyo, Japan).

Results

The characterization of BMP4-treated cells in defined conditions.

We observed morphological changes of the cells when undifferentiated mES cells at a density of 5 × 103 cells per square centimeter were cultured in ESF5 medium supplemented with 10 ng/ml BMP4 on laminin-coated dishes for 4 d. The morphology of the majority of cells became cobblestone-shaped (Fig. 1 A, left). Before the cells reached to confluent, we subcultured the cells into the same culture conditions at a density of 2 × 104 cells per square centimeter on culture day 4. After four additional days in culture, multinuclear cells appeared (Fig. 1 A, right). Trophoblast stem cells (TSCs) exhibit a cobblestone morphology (Tanaka et al. 1998) and subsequently differentiate into multinuclear trophoblast (Simmons and Cross 2005). These findings suggested the possibility that mESCs cultured with BMP4 in ESF5 might differentiate into trophoblast lineages. To explore this possibility further, we examined the gene expression of transcription factors which are expressed in trophoblast, Cdx2 (Strumpf et al. 2005), Dlx3 (Morasso et al. 1999), Eomes (Russ et al. 2000), Errb (Luo et al. 1997), Esx1 (Li and Behringer 1998), Ets2 (Yamamoto et al. 1998), Gata3 (Ng et al. 1994), Hand1 (Cross et al. 1995), Mash2 (Guillemot et al. 1995), and Psx1 (Chun et al. 1999) in the differentiated cells by quantitative real-time RT–PCR. The relative mRNA levels of the majority of these transcription factors were increased over tenfold compared with those of the undifferentiated mESCs (Fig. 1 B). The mRNA levels of Eomes, Errb, and Ets2, which were involved in self-renewal of undifferentiated TSCs (Luo et al. 1997; Russ et al. 2000; Wen et al. 2007), were not increased. To confirm whether BMP4 specifically activates the expression of trophoblast markers, we examined the transcription of other cell lineage markers in early mammalian development, Sox1 (ectoderm), Flk1 (mesoderm), Mixl1 (definitive endoderm), and Gata6 (primitive extraembryonic endoderm) in the cells cultured with BMP4 in ESF5 medium on culture day 4 by quantitative real-time RT–PCR. The gene expression of the ectoderm, mesoderm, definitive endoderm, or extraembryonic endoderm cell lineage markers examined here was no higher than that in undifferentiated mESCs (Fig. 1 C). These results indicate that the differentiated cells specifically upregulated transcription factors for trophoblast.

Figure 1.

The effect of BMP4 on mESCs in defined culture conditions. (A) phase contrast photomicrographs of differentiated mESCs cultured in BMP4-supplemented ESF5 medium for 4 d (left) and 8 d (right). Scale bars are 100 µm. (B) quantitative real-time RT–PCR analysis of trophoblast-specific transcription factor expressions. The mESCs cultured in BMP4-supplemented ESF5 medium for 0, 2, 4, 6, and 8 d were analyzed. (C) Quantitative RT–PCR analysis of differentiation markers of mESCs. The cells cultured in BMP4-supplemented ESF5 medium for 4 d were analyzed. The amount of undifferentiated mESCs is indicated as 1. (D) Proliferation of differentiating mESCs on various BMP4 concentrations. mESCs were seeded in a 24-well dish at 5 × 103 cells per well on each BMP4 concentration in ESF5. Cells were counted every 24 h. The values are the mean ± SEM (n = 4). (E) quantitative RT–PCR analysis of BMP4-induced trophoblast transcription factors in mESCs cultured in BMP4, activin A (10 ng/ml), or FGF4-supplemented (25 ng/ml) ESF5 medium on laminin for 4 d was analyzed. The gene expressions were normalized by the amount of Gapdh. The amount of the undifferentiated mES is indicated as 1. The values are the mean ± SEM (n = 4).

As the mESCs proliferated in a dose-dependent manner in response to BMP4 concentration in ESF5 medium (Fig. 1 D) and did not proliferate without addition of BMP4, we could not compare the gene expression with those in the BMP4-untreated cells. To examine whether these genes were specifically induced in BMP4-treated cells or not, we compared the effect of FGF4 or activin with that of BMP4 on mESC gene expression. Gene expression profiles in the cells cultured with BMP4, activin, or FGF4 in ESF5 on laminin for 4 d were analyzed by quantitative RT–PCR. FGF-4 increased the Cdx2 expression in the cells as BMP4 did, but activin did not increase Cdx2 expression. Neither FGF-4 nor activin increased the expression of Dlx3, Esx1, Gata3, Hand1, Mash2, or Psx1 (Fig. 1 E). These results indicated that BMP4 specifically upregulated the expression of these trophoblast-specific transcription factors.

To characterize these putative trophoblast cells, we further analyzed the expression of the trophoblast marker proteins Cdh3 (placental cadherin; Nose and Takeichi 1986), Cdx2, cytokeratin (CK) 7 (Lu et al. 2005), and connexin (Cx) 31 (Zheng-Fischhofer et al. 2007), by immunocytochemistry. The differentiated cells expressed Cdh3, Cdx2, CK7, and Cx31 proteins (Fig. 2 A). Furthermore, we examined the expression of additional trophoblast-specific genes: Gcm1 (Anson-Cartwright et al. 2000), Placental lactogen (Pl)-1 (Faria et al. 1991), Plf (Simmons et al. 2008), Tpbpa, and Tpbpb (Lescisin et al. 1988) by RT–PCR. These trophoblast markers are expressed at much higher levels in placenta than in whole embryos. The expression of trophoblast marker genes was prominent in the cells cultured with BMP4 in ESF5 for 8 d (Fig. 2 B). By flow cytometry analysis, 83.2% of cells cultured with BMP4 in ESF5 were positive for Cdh3 expression (Fig. 2 C). These results indicate that the majority of the differentiated cells exhibit trophoblast characteristics. We observed multinuclear cells surrounded by Cdh3 in this culture conditions, which is a characteristic of differentiated trophoblast (Simmons and Cross 2005; Fig. 2 D). To examine the frequencies of multinuclear cells in the culture, the intensity of PI in the cells cultured with BMP4 in ESF5 for 8 d was examined. A small population of hyperploid cells (>4 N) was observed in the differentiated cell cultures (Fig. 2 E). The percentage of hyperploid cells (>4 N) was 37% in the cells, suggesting the presence of multinuclear differentiated trophoblast. We have confirmed that other mESC lines, B6G-2, E14, and EB3, also differentiated into trophoblast-like cells under these conditions (data not shown).

Figure 2.

Differentiation into trophoblast from mESCs in vivo. (A) Immunocytostaining with trophoblast marker antibodies of the cells cultured in BMP4-supplemented ESF5 medium for 4 d (right) or undifferentiated mESCs (left). Immunoreactivity of Cdh3 (upper left), Cdx2 (upper right), CK7 (lower left), and Cx31 (lower right) was visualized with AlexaFluor-488-conjugated secondary antibodies (green). Nuclei were stained with DAPI (blue). Scale bars were 50 µm. (B) RT–PCR analysis of the expression of placental markers in differentiated mESCs. (C) Flow cytometric analysis of mESCs (blue) and differentiated cells at eight culture days (red). (D) Immunocytostaining with anti-Cdh3 antibodies of the cells cultured in BMP4-supplemented ESF5 medium for 8 d. Arrows indicate the nuclei of hyperploid cells. Scale bars were 50 µm. (E) Flow cytometric analysis of mESCs (blue) and differentiated cells at eight culture days (red). The DNA contents were visualized with propidium iodide. Diploid (2N), tetraploid (4N), and octaploid (8N) DNA contents are indicated in the DNA content graph.

We examined the effect of extracellular matrix on the differentiation into trophoblast cells from mESCs. We found that the cells steadily proliferated and expressed CK7 when mESC were cultured with BMP4 in ESF5 on fibronectin or laminin (Fig. 3). In contrast, the cells died and failed to express CK7 on type I collagen or PDL. These results indicate that fibronectin and laminin promoted mESC differentiation into trophoblast lineages.

Figure 3.

The effect of ECM components on the differentiation into trophoblast from mESCs. (A) Proliferation of differentiating mESCs on various ECM components. mESCs were seeded in a 24-well dish at 5 × 103 cells per well on each ECM component in ESF5 medium supplemented with BMP4. Cells were counted every 48 h. The values are the mean ± SEM (n = 3). (B) Immunocytostaining and TUNEL assay of the differentiated mESCs cultured for 4 d in BMP4-supplemented medium. As trophoblast markers, CK7 was detected with specific antibodies (red). TUNEL assay (Chemicon) was performed according to the manufacturers’ direction and labeled fixed cells with fluorescence (green). Nuclei were stained with DAPI (blue). Scale bars are 50 µm. (C) Percentages of CK7-positive cells. Percentages are calculated from the observation of more than 500 cells for each sample. The values are the mean + SEM (n = 3). (D) Percentages of TUNEL-positive cells. Percentages are calculated from the observation of more than 500 cells for each sample. The values are the mean + SEM (n = 3).

The effect of serum and LIF on mESC differentiation into trophoblast cells.

Although it is generally accepted that few mESCs differentiate into trophoblast lineages, we successfully directed differentiation of several mESC lines into trophoblast cells. We presume that this phenotype stems from our use of serum-free culture conditions. To elucidate the effects of serum on mESC differentiation into trophoblast, we compared trophoblast marker expression in the cells cultured with BMP4 in ESF5 medium in the presence or absence of 10% FCS. The addition of FCS decreased BMP4-induced expression of trophoblast transcription factors (Fig. 4 A) and Cdh3 protein expression (Fig. 4 B). These results indicate that FCS inhibited the differentiation of mESCs into trophoblast. Another candidate inhibitor of mESC differentiation into trophoblast is LIF, which is known to maintain mESCs in the undifferentiated state (Smith et al. 1988; Williams et al. 1988). To elucidate the effects of LIF on trophoblast differentiation, we examined trophoblast gene expression in cells cultured with BMP4 in the presence or absence of 10 ng/ml of LIF in ESF5. Addition of LIF decreased the BMP4-induced expression of trophoblast-specific transcription factors (Fig. 4 C) and Cdh3 protein expression (Fig. 4 D) and enhanced the expression of undifferentiated pluripotent stem cell markers, Nanog and SSEA1 (Fig. 4 E). These results suggested that LIF inhibited differentiation of mESCs into trophoblast and confirmed that BMP4 in conjunction with LIF functions to maintain the undifferentiated state of mESCs.

Figure 4.

Effect of FCS and LIF on the differentiation into trophoblast. The effect of FCS: the cells were cultured in BMP4-supplemented ESF5 medium with 10% FCS (indicating as “BMP4 with FCS”) or without FCS (indicating as “BMP4 only”) for 4 d. (A) Quantitative real-time RT–PCR analysis of the expression of trophoblast-specific transcription factors. The gene expressions were normalized by the amount of Gapdh. The values are the mean ± SEM (n = 4). (B) Immunocytochemistry with Cdh3 antibodies. Immunopositive reaction of Cdh3 antibody was visualized with AlexaFluor-488-conjugated secondary antibodies (green). Nuclei were stained with DAPI (blue). Scale bars are 50 µm. The effect of LIF: the cells were cultured in BMP4-supplemented ESF5 medium with 10 ng/ml of LIF (indicating as “BMP4 with LIF”) or without LIF (indicating as “BMP4 only”) for 4 d. (C) Quantitative real-time RT–PCR analysis of the expression of trophoblast-specific transcription factors. (D) Immunocytochemistry with Cdh3 antibodies. (E) Immunocytochemistry with anti-Nanog or anti-SSEA1 antibodies. Immunopositive reaction of anti-Nanog or anti-SSEA1 antibody was visualized with AlexaFluor-488-conjugated secondary antibodies (green).

Involvement of BMP–Smad pathway in mESC differentiation into trophoblast.

To examine whether BMP4 itself may promote the differentiation of mESCs into trophoblast, we tested the effects of a BMP antagonist, Noggin (300 ng/ml), on the response of mESCs to BMP4. Addition of Noggin decreased the BMP4-induced expression of trophoblast transcription factors (Fig. 5 A) as well as Cdh3 (Fig. 5 B) in the differentiated cells. This result indicated that BMP4 itself promoted the mESC differentiation into trophoblast. Next, we examined whether the BMP–Smad pathway was involved in the BMP4-induced differentiation of mESCs into trophoblast. Western blotting analysis showed that the activation by phosphorylation of Smad1/5/8 was observed in cells treated with BMP4 (Fig. 5 C). Furthermore, the overexpression of Smad6 (Fig. 5 D), which inhibits the transcriptional activity of Smad proteins, decreased the BMP4-induced expression of trophoblast transcription factors (Fig. 5 E) as well as Cdh3 (Fig. 5 F) in the differentiated cells. These results indicated that BMP4-stimulated differentiation of mESCs into trophoblast is mediated through by the Smad pathway.

Figure 5.

The effects of Noggin, inhibitory Smad6 on the BMP4-induced differentiation of mESCs into trophoblast. The effect of Noggin: the cells were cultured in BMP4-supplemented ESF5 medium with 300 ng/ml of Noggin (indicating as “BMP4 with Noggin”) or without Noggin (indicating as “BMP4 only”) for 4 d. (A) Quantitative real-time RT–PCR analysis of the BMP4-induced expression of trophoblast-specific transcription factors. These gene expressions were normalized by the amount of Gapdh. The values are the mean ± SEM (n = 4). (B) Immunocytostaining with Cdh3 antibodies. Immunoreactivity of Cdh3 antibody was visualized with AlexaFluor-488-conjugated secondary antibodies (green). Nuclei were stained with DAPI (blue). Scale bars were 50 µm. The effect of inhibitory Smad; (C) protein samples were lysed from the mESCs stimulated by 100 ng/ml of BMP4 for 0, 15, and 60 min. Phosphorylation level of Smad1/5/8 in mESCs was analyzed by Western blotting using antibodies to phospho Smad1/5/8 proteins or total Smad1/5/8 proteins. Alpha-tubulin was used as the loading control. (D) Protein samples were lysed from mESCs transfected with Smad6 plasmids (indicating as “Smad6 OE”) or mock plasmids (indicating as “Mock”) for 48 h. Protein content of Smad6 in mESCs was analyzed by Western blotting using antibodies to Smad6 proteins. Alpha-tubulin was used as the loading control. (E) The cells transfected with Smad6-expressing vectors (indicating as “Smad6 OE”) or mock vectors (indicating as “Mock”) were cultured in BMP4-supplemented ESF5 medium for 72 h. Quantitative real-time RT–PCR analysis of the BMP4-induced expression of trophoblast-specific transcription factors. (F) Immunocytostaining with Cdh3 antibodies.

Identification of Cdx2 as a critical and direct target of BMP–Smad pathway in mESC differentiation into trophoblast.

Next, we searched for a crucial gene involved in trophoblast differentiation induced by BMP4. Previous studies have demonstrated that Cdx2 regulates the induction of trophoblast from mESCs (Niwa et al. 2005; Tolkunova et al. 2006). Thus, we hypothesized that Cdx2 could be a crucial gene involved in trophoblast differentiation by BMP4. We generated Cdx2-knockdown mESCs carrying an shRNA plasmid against Cdx2 with a puromycin-resistance gene. The expression of Esx1 and Psx1 (Fig. 6 A) and the expression of Cdh3 protein were decreased in the Cdx2-knockdown mESCs cultured with BMP4 in ESF5 (Fig. 6 B). These results suggested that Cdx2 was crucial for the BMP4-induced differentiation of mESCs into trophoblast through the regulation of the expression of Esx1 and Psx1.

Figure 6.

The effects of shRNA against Cdx2 on the BMP4-induced differentiation of mESCs into trophoblast. The cells carrying shRNA-expressing vectors against Cdx2 (indicating as “Cdx2 shRNA”) or control shRNA-expressing vectors (indicating as “Control shRNA”) were cultured in BMP4-supplemented ESF5 medium for 48 h. (A) Quantitative real-time RT–PCR analysis of the BMP4-induced expression of trophoblast-specific transcription factors. These gene expressions were normalized by the amount of Gapdh. The values are the mean ± SEM (n = 4). Asterisks indicate p < 0.05 by Student’s t test. (B) Immunocytostaining with Cdh3 antibodies. Immunoreactivity of Cdh3 antibody was visualized with AlexaFluor-488-conjugated secondary antibodies (green). Nuclei were stained with DAPI (blue). Scale bars were 50 µm.

Finally, we explored the possibility that BMP–Smad pathway could directly regulate the Cdx2 transcription. We first searched for the putative Smad1/5/8-binding sequence (GCCG) in the noncoding conserved sequences in the genomic region of Cdx2 using the Vista comparative genomics tool (Frazer et al. 2004; Danno et al. 2008). We found a set of putative Smad1-binding sequences (GCCG) highly conserved among mammals in intron 1 of the Cdx2 gene, designated as CICS1 (Fig. 7 A). We cloned a 350-bp sequence containing CICS1 and inserted it into the pGL4.23 luciferase reporter plasmid. To confirm that this sequence has enhancer activity in response to BMP4, a luciferase reporter assay was performed in mESCs transfected with the reporter plasmid or empty plasmid. Transcriptional activity was increased in a BMP4 dose-dependent manner in transfected mESCs, whereas Noggin (300 ng/ml) decreased BMP4-induced transcriptional activity (Fig. 7 B). These results indicate that CICS1 has BMP4-dependent transcriptional activity. To examine whether Smad proteins can bind to CICS1, we performed EMSA. EMSA revealed that the Smad proteins were bound to CICS1 in vitro (Fig. 7 C). To confirm whether endogenous Smad proteins were bound to CICS1 in the BMP4-induced trophoblast cells, we performed ChIP assays (Fig. 7 D, E). The CICS1 sequence was specifically contained in the DNA–protein complex immunoprecipitated by the anti-pSmad1 antibodies. These results demonstrate specific binding of the Smad1 proteins to CICS1 in BMP4-treated mESCs. Based on the experimental results obtained here, we conclude that Cdx2 transcription is directly regulated by the BMP–Smad pathway during BMP4-induced mESC differentiation.

Figure 7.

Direct activation of Cdx2 by BMP–Smad pathway. (A) Conserved sequence domain in the genomic region of Cdx2 gene in vertebrates. Colored peaks (purple, coding; pink, non-coding) indicate the regions of at least 100 bp and 60% homology. Magnified sequence of CICS1 compared between human and mouse is shown in the bottom part. Putative Smad1 protein binding sequences (GCCG or CGGC) are enclosed in green box. (B) Dual-luciferase assay of CICS1 transcriptional activity by BMP4. The cells carrying the reporter vectors constructed with the sequence containing CICS1 were cultured in BMP4-supplemented ESF5 medium for 48 h. N.C. indicates empty vectors. The values are the mean ± SEM (n = 4). (C) EMSA of CICS1 bound by Smad1 and the cofactor, Smad4. (D) ChIP assays of CICS1 bound by pSmad1. Genomic fragments of CICS1, 4 kb upstream or 4 kb downstream regions bound by pSmad1/5/8 protein, were immunoprecipitated with anti-pSmad1/5/8 antibody and were analyzed by PCR. (E) Quantitative graphs of ChIP assays calculated from the ratio of the amount of bound anti-pSmad1 antibodies to the amount of control IgG. The values are the mean ± SEM (n = 4). * indicates p < 0.05 by Student’s t test.

Discussion

Previous studies reported that isolated ICM could partially differentiate into trophoblast (Handyside 1978; Hogan and Tilly 1978) whereas it was reported that mESCs did not contribute to extraembryonic trophoblast in chimeric embryos (Beddington and Robertson 1989). From these findings, we inferred that mESCs under certain conditions have the potential to differentiate into trophoblast lineages as well as all three embryonic germ layers. Toumadje et al. observed spontaneous expression of the trophoblast marker cytokeratin endo-A (CK8) in D3 mESC embryoid bodies cultured in serum in the absence of LIF (Toumadje et al. 2003). A recent study reported that a small subset of trophoblast-like cells appeared when mESCs were cultured on type IV collagen (Schenke-Layland et al. 2007). Another study reported that the addition of Wnt3a induced a small subset of trophoblast lineages from mESCs (He et al. 2008). These findings suggested that mESCs may have the potential to differentiate into trophoblast lineages and that the modulation of culture conditions regulates mESCs to differentiate into trophoblast in vitro.

To induce trophoblast from mESCs, laminin proved to be a key component. We previously reported that laminin or fibronectin promotes cell differentiation in our defined culture conditions (Hayashi et al. 2007). The results in this study indicated that mESC differentiated into trophoblast lineages only on fibronectin or laminin. Thus, our findings suggested that laminin regulates mESCs differentiation into trophoblast in the defined culture condition. Previous studies using serum or serum replacement which contain a variety of ECM components most likely associated with an ill-defined cocktail of growth factors. We have suggested that serum hampers the elucidation of the biological mechanisms and gives rise to apparently different conclusions on the ability of ESCs to differentiate into trophoblast.

HESCs have been reported to express trophoblast markers after treatment with BMP4 in vitro (Xu et al. 2002). Previous studies reported that mouse epiblast stem cells or a subpopulation of epiblast-like mESCs express trophoblast markers in response to BMP4 (Brons et al. 2007; Hayashi et al. 2008). Mouse epiblast stem cells resemble hESCs more closely than mESCs (Brons et al. 2007; Tesar et al. 2007; Hayashi et al. 2008). We have previously found polarized cell morphology and the upregulation of Fgf5 expression and downregulation of some self-renewal marker gene expression in mESCs cultured on laminin or fibronectin (Hayashi et al. 2007). These are characteristics of epiblast cells. In this study, cells were treated with BMP4 on laminin in our defined culture conditions. Our findings suggest that the differentiation via an epiblast-like state is important for mESC differentiation into trophoblast, and BMP4 can induce from both human and mouse pluripotent stem cells to differentiate into trophoblast.

As described above, BMP4 is a key component promoting trophoblast differentiation from both mESCs and hESCs (Xu et al. 2002). These findings contradict previous reports in which BMP4 supported cell self-renewal. However, that was observed with BMP4 in conjunction with LIF (Ying et al. 2003; Qi et al. 2004). BMP4 supports mESC self-renewal by inhibiting neural differentiation which is not inhibited by LIF. We have shown in this study that LIF inhibits trophoblast differentiation by BMP4. These results suggest that LIF and BMP4 reciprocally inhibit differentiation of mESCs into neural or trophoblast lineages and lead to maintain self-renewal of mESCs. Previous studies have also reported that BMP4 is also thought to be involved in mesoderm differentiation from ESCs (Johansson and Wiles 1995; Wiles and Johansson 1999; Nostro et al. 2008; Sumi et al. 2008). These studies employed other signal activation by cytokines or tight cell–cell attachments. Our results suggest that BMP4 induces differentiation into trophoblast in defined conditions in the absence of other activating signals and that differentiation was perturbed by the addition of serum. Thus, BMP4 may function as trophoblast inducer when it acts alone and as mesoderm inducer when it is coordinated with other as yet unidentified signals.

We demonstrate that the BMP4–Smad pathway is involved in mESC differentiation into trophoblast. Components of BMP–Smad pathway are expressed in extraembryonic tissues and affect their development (Lechleider et al. 2001; Tremblay et al. 2001). Previous study reported that Nanog protein, which maintains mESC self-renewal, binds to Smad1 protein and inhibits the transcriptional activity of Smad1 protein in mESCs (Suzuki et al. 2006). Our results reveal that the expression of Nanog protein was absent from cells cultured with BMP4 in ESF5 but expressed in cells in the presence of LIF. Thus, Nanog expression induced by LIF might inhibit mESC differentiation into trophoblast.

Our knockdown experiment showed that the cells which have reduced Cdx2 expression did not express Esx1, Psx1, and Cdh3. Cdx2 is essential for trophoblast lineages development in vivo (Niwa et al. 2005; Strumpf et al. 2005). Thus, our data imply that these cells may differentiate into cell lineages other than trophoblast. Although the characterization of these cell lineages is interesting for further experiment, they may be induced by BMP4 and not express Cdx2.

We also demonstrate that BMP–Smad pathway activates the Cdx2 expression during the differentiation of mESCs into trophoblast and that the Cdx2 expression is directly regulated by the BMP–Smad pathway through binding of Smad1 protein with evolutionary conserved intronic enhancer in Cdx2 gene. Correlation between BMP–Smad pathway and CDX2 expression is reported in human gastric cancer cell lines (Barros et al. 2008). Together with the fact that BMP4 increases the trophoblast marker gene expression in hESCs (Xu et al. 2002), we conclude that common developmental and molecular mechanisms are involved in the differentiation of mammalian pluripotent stem cells into trophoblast. We believe that our methods and findings in this study provide a better understanding of the molecular mechanisms that regulate the differentiation of pluripotent stem cells into trophoblast and placenta during early mammalian development.