Induction of initial heart α-actin, smooth muscle α-actin, in chick pregastrula epiblast: The role of hypoblast and fibroblast growth factor-8

Authors

  • Hiroko Matsui,

    Search for more papers by this author
    • These authors contributed equally to this work (Hiroko Matsui and Masahide Sakabe).

  • Masahide Sakabe,

    Search for more papers by this author
    • These authors contributed equally to this work (Hiroko Matsui and Masahide Sakabe).

  • Hirokazu Sakata,

    1. Department of Anatomy and Cell Biology, Graduate School of Medicine, Osaka City University, Osaka 545-8585, Japan, Department of Anatomy, School of Medicine, Saitama Medical University, Saitama 350-0495, Japan
    Search for more papers by this author
  • Nariaki Yanagawa,

    1. Department of Anatomy and Cell Biology, Graduate School of Medicine, Osaka City University, Osaka 545-8585, Japan, Department of Anatomy, School of Medicine, Saitama Medical University, Saitama 350-0495, Japan
    Search for more papers by this author
  • Kazuo Ikeda,

    1. Department of Anatomy and Cell Biology, Graduate School of Medicine, Osaka City University, Osaka 545-8585, Japan, Department of Anatomy, School of Medicine, Saitama Medical University, Saitama 350-0495, Japan
    Search for more papers by this author
  • Toshiyuki Yamagishi,

    1. Department of Anatomy and Cell Biology, Graduate School of Medicine, Osaka City University, Osaka 545-8585, Japan, Department of Anatomy, School of Medicine, Saitama Medical University, Saitama 350-0495, Japan
    Search for more papers by this author
  • Yuji Nakajima

    Corresponding authorSearch for more papers by this author

*Author to whom all correspondence should be addressed.
Email: yuji@med.osaka-cu.ac.jp

Abstract

During heart development at the gastrula stage, inhibition of bone morphogenetic protein (BMP) activity affects the heart specification but does not impair the expression of smooth muscle α-actin (SMA), which is first expressed in the heart mesoderm and recruited into initial heart myofibrils. Interaction of tissues between posterior epiblast and hypoblast at the early blastula stage is necessary to induce the expression of SMA, in which Nodal and Chordin are thought to be involved. Here we investigated the role of fibroblast growth factor-8 (FGF8) in the expression of SMA. In situ hybridization and reverse transcription–polymerase chain reaction showed that Fgf8b is expressed predominantly in the nascent hypoblast. Anti-FGF8b antibody inhibited the expression of SMA, cTNT, and Tbx5, which are BMP-independent heart mesoderm/early cardiomyocyte genes, but not Brachyury in cultured posterior blastoderm, and combined FGF8b and Nodal, but neither factor alone induced the expression of SMA in association with heart specific markers in cultured epiblast. Although FGF8b did not induce the upregulation of phospho-Smad2, anti-FGF8b properties suppressed phospho-Smad2 in cultured blastoderm. FGF8b was able to reverse the BMP-induced inhibition of cardiomyogenesis. The results suggest that FGF8b acts on the epiblast synergistically with Nodal at the pregastrula stage and may play a role in the expression of SMA during early cardiogenesis.

Introduction

In the early avian embryo at the blastula stage, presumptive heart cells reside in the posterior lateral region of the epiblast layer (Hatada & Stern 1994; Yatskievych et al. 1997). During gastrulation, prospective mesendoderm cells containing future heart mesoderm migrate into the midline region of the posterior epiblast and ingress from the primitive streak. At this stage, cells for the future heart mesoderm are located in the primitive streak, excluding the most anterior and posterior regions (Garcia-Martinez & Schoenwolf 1993). As gastrulation proceeds, the prospective heart cells leave the primitive streak, migrate anterolaterally, and generate a left and right anterior lateral plate mesoderm, called the precardiac mesoderm. During gastrulation, prospective heart cells make contact with the anterior endoderm and are specified to the heart lineage (Orts-Llorca 1963; Antin et al. 1994; for reviews, Lough & Sugi 2000; Brand 2003). After the mesoderm is committed to the heart lineage, embryonic folding occurs and the left and right visceral mesoderm move to the ventral midline of the intestinal portal and fuse to each other, resulting in the formation of the primitive heart tube.

During early heart development at the blastula to gastrula stage, two distinct tissue interactions are evident; one between the anterior lateral plate mesoderm and its subjacent endoderm at the gastrula stage, the other between the epiblast and hypoblast at the early blastula stage (Orts-Llorca 1963; Yatskievych et al. 1997). At the gastrula stage, endoderm-derived bone morphogenetic protein (BMP)-2/4 and fibroblast growth factor (FGF)-8 are thought to be required to regulate the heart specification process as well as terminal differentiation; therefore, these factors regulate the expression of not only heart-specific transcription factors (Nkx2.5 GATA4 and Mef2c) but also contractile proteins (such as sarcomeric α-actinin, sarcomeric myosin, and titin) (Schultheiss et al. 1995, 1997; Lough et al. 1996; Alsan & Schultheiss 2002; Nakajima et al. 2002). During heart specification and terminal differentiation at the gastrula stage, BMP antagonist inhibits the expression of Nkx2.5, GATA4, sarcomeric α-actinin, sarcomeric myosin and titin, but not smooth muscle α-actin (SMA), which is initially expressed in the developing precardiac mesoderm and then recruited into the initial I-band of the nascent myofibrils (Sugi & Lough 1992; Nakajima et al. 2002). Anti-FGF8 properties fail to inhibit cardiogenesis in cultured gastrula embryo (Alsan & Schultheiss 2002). Interaction of tissues between the posterior epiblast and hypoblast at the early blastula stage (Stage X–XI, Eyal-Giladi & Kochav 1976) is necessary to induce the expression of SMA in cultured posterior epiblast (i.e. posterior epiblast of early blastula, but not late blastula) is not capable of expressing SMA without cocultivation of the associated hypoblast. Nodal, which is first expressed in the posterior region of the blastoderm (Lawson et al. 2001; Chapman et al. 2002), is necessary, but not sufficient, for the expression of SMA in cultured epiblast (Matsui et al. 2005); however, the hypoblast-derived inductive signal(s), which acts on the posterior epiblast synergistically with Nodal to induce the expression of SMA, remains uncertain.

Fibroblast growth factor has been reported as a potent inductive signaling that induces mesoderm induction in the Xenopus animal cap (Kimelman & Kirschner 1987; Slack et al. 1987). Experiments using a dominant negative form of the type I FGF receptor showed that the FGF signal is required for mesoderm induction by Activin (Cornell & Kimelman 1994). In chick gastrula embryos, several FGF isoforms, such as Fgf3, –4, –8, –13 and –18, are expressed at blastula to gastrula stage (Streit et al. 2000; Lawson et al. 2001; Chapman et al. 2002; Karabagli et al. 2002). During mouse development, Fgf3, –4, –5, –8 and –17 are expressed at prestreak to streak stage. The null mutant embryo for either Fgf3, –5 or –17 is viable and the Fgf4-null mutant dies shortly after implantation; therefore, these FGF isoforms are not individually required for gastrulation in mouse embryos (Sun et al. 1999). In contrast, targeted disruption of mouse Fgf8 showed that epiblast cells fail to migrate away from the primitive streak, thereby impairing the formation of the mesoderm and endoderm (Sun et al. 1999). In Xenopus, FGF8b, an FGF8 splice form, is a potent mesoderm inducer, while FGF8a has little effect on the development of mesoderm (Fletcher et al. 2006). In the present study, we examined spatiotemporal expression patterns for FGF8 in pregastrula chick embryos and carried out explantation experiments in serum-free defined medium to investigate the possible roles of FGF8b in the expression of SMA. The results showed that FGF8b is likely to be one of the inductive molecules secreted by the hypoblast and acting on the posterior epiblast synergistically with Nodal to induce the expression of SMA in cultured epiblast.

Materials and methods

Culture procedures

Stage X–XI pregastrula blastoderms of chick embryos were staged according to Eyal-Giladi & Kochav (1976; incubation = 0 h) and collected on ice-cooled phosphate-buffered saline (PBS). Posterior regions of the area pellucida (containing sickle), in which prospective heart cells reside, or anterior regions (non-cardiogenic) were cut with a sharp tungsten needle. Epiblast explants were prepared using a thin tungsten needle or eyebrow hair. The resulting explants were explanted onto chamber slides (Nunc, NY, USA) and cultured in serum-free defined medium (75% DMEM, 25% McCoy's medium, supplemented with 10−7 M dexamethasone and penicillin-streptomycin; Ladd et al. 1998) or various test conditions, such as medium containing SU5402 (CALBIOCHEM, CA, USA), anti-FGF8b antibody, anti-FGF2 antibody anti-FGF4 antibody, recombinant FGF8b, Nodal, Lefty1, and BMP2 (R&D, MN, USA).

Indirect immunofluorescence microscopy

Immunohistochemistry was carried out as described by Nakajima et al. (2002). Cultures were drained of medium, rinsed with PBS, fixed with 4% paraformaldehyde/PBS for 1 h at room temperature, then rinsed with PBS. Specimens were blocked for 1 h with 1% bovine serum albumin/PBS containing 0.1% Triton X-100, incubated with a primary antibody mixture (antismooth muscle α-actin (immunoglobulin G2a [IgG2a]) and anti-sarcomeric α-actinin (IgG1), Sigma, MO, USA) at 4°C overnight, rinsed with PBS, then incubated with a secondary antibody mixture (fluorescein isothiocyanate [FITC]-conjugated goat anti-mouse IgG1 and rhodamine isothiocyanate (RITC)-conjugated goat anti-mouse IgG2a, Southern Biotechnology, AL, USA) for 1 h at room temperature. The nuclei were stained with 4′6′-diamidino-2-phenylindole dihydrochloride (DAPI) for 20 min, rinsed with PBS, and mounted. Samples were observed under a laser confocal microscope (Zeiss, Tokyo, Japan). The number of explants that expressed SMA or sarcomeric α-actinin was counted under the microscope. The percentage incidence of sarcomeric protein-positive explants was calculated. Statistical analyses were carried out using Fisher's exact test, with significance being at P < 0.01.

Reverse transcription–polymerase chain reaction

RNA was extracted from cultured explants as described previously (Matsui et al. 2005). cDNAs were synthesized from 0.2 µg of total RNA, and polymerase chain reaction (PCR) was carried out in 10 µL of reaction buffer (QIAGEN, Hilden, Germany). The primers of Nkx-2.5 (Schultheiss et al. 1995), GATA4 (Schultheiss et al. 1997), sarcomeric α-actinin (Nakajima et al. 2002), Nodal (Levin et al. 1995), Tbx5 (Golz et al. 2004), FGF8s (Sato et al. 2001), SMA (Matsui et al. 2005), Brachyury (Muhr et al. 1999) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Yamagishi et al. 1999) have been described elsewhere. We also used the following primers: CFC, 5′-TATGGTGCGAGAACAACTGCA-3′ (forward) and 3′-TTCATCACCCAGTCGCCAT-5′ (reverse) [accession No. AF282984]; Chordin, 5′-TGCCATCCTGACCTCATCAGA-3′ (forward) and 3′-CGGTTTGGTCTCCAGTGTGA-5′ (reverse) [NM204980]; cardiac troponin T (cTNT), 5′-CAAGCATGTCGGACTCTGAAG-3′ (forward) and 3′-TCTTCCTCTTTGCGAGCTCT-5′ (reverse) [M10013]. Samples were cycled at 94°C for 30 s at an annealing temperature of 55°C except for Nodal (60°C), and at 72°C for 1 min, with a final extension at 72°C for 10 min. The number of cycles for the various primers was as follows: SMA, 23; cTNT, 23; Tbx-5, 30; GATA-4, 28; Nkx-2.5, 26; sarcomeric α-actinin, 30; Nodal, 26; CFC, 24; Chordin, 28; FGF8, 28; Brachyury, 26 and GAPDH, 28. PCR products were electrophoresed and stained with ethidium bromide.

Western blotting

Explants were homogenized in sodium dodecyl sulfate (SDS) sample buffer (2 µL/explant) (62.5 mm Tris-HCl, 10% glycerol, 2% SDS, 5% 2-β mercaptoethanol, and 1 mm Na3VO4, pH 6.8). After heat denaturation at 95°C for 5 min, samples (5 µg of protein) were subjected to 10% SDS–polyacrylamide gel electrophoresis (PAGE) and then transferred onto Immobilon-P membranes (Millipore, MA, USA). After blocking with 5% non-fat dry milk, membranes were incubated with the primary antibody against SMA (Sigma), GAPDH, phospho-Smad2 (C-terminal region, CHEMICON, CA, USA), Erk (Santa Cruz Biotechnology, Inc., CA, USA), Smad2, phospho-Erk, or phospho-Smad1/5/8 (C-terminal region, Cell Signaling) for 2 h at room temperature. After washing, they were incubated with horseradish peroxidase-conjugated secondary antibodies. Immunoreactive bands were visualized using ECL™ detection reagent (Amersham, Buckinghamshire, UK).

In situ hybridization

Digoxigenin-labeled single-stranded RNA was prepared using a digoxygenin (DIG) RNA labeling kit (Roche Diagnostics, Penzberg, Germany) according to the manufacturer's instructions. Full-length chick FGF8b cDNA (NM_001012767), Nodal (nucleotide position 172–723, XM_424385.2) and Chordin (664–1389, NM_204980.1) subcloned into pGEM7 Zf (+) or pGEM-T vector (Promega, WI, USA) were linearized using BamHI or Nco1, and transcribed using T7 RNA polymerase. Single-stranded RNA probe for FGF8b can hybridize to FGF8a or 8b mRNA. Whole-mount in situ hybridization was carried out as described by Nieto et al. (1996). Embryos were fixed in 4% paraformaldehyde in PBS for 2 h, washed with PBT (PBS containing 0.1% Triton X-100), dehydrated and rehydrated through a graded series of methanol in PBT, and then fixed with 0.2% glutaraldehyde/4% paraformaldehyde in PBT for 20 min. After rinsing with PBT, samples were prehybridized with hybridization buffer (50% formamide, 5 × standard saline citrate [SSC][pH 5.0], 50 µg/mL yeast tRNA, 1% SDS, and 50 µg/mL heparin) for 2 h at 60°C and then hybridized with a DIG-labeled probe (0.5 µg/mL in hybridization buffer) for 12 h at 60°C. After hybridization, samples were washed in 2 × SSC for 1 h at 60°C, then in 0.2 × SSC for 1 h at 60°C. After rinsing with KTBT (Tris-buffered saline containing 1% Triton X-100), embryos were blocked with 20% sheep serum in KTBT for 1 h, then incubated with alkaline phosphatase-conjugated anti-digoxigenin antibody for 12 h at 4°C. Hybridization was detected using 5-bromo-4-chloro-3-indolyl-phosphate/4-nitroblue tetrazolium chloride (BCIP/NBT).

Results

FGF8b is expressed in the hypoblast at early blastula stage

Spatiotemporal expression patterns of Fgf8 in pregastrula to early gastrula chick embryos (stage X to stage 3) were examined by in situ hybridization. At stage X–XI, the blastoderm consists of the upper epiblast, future embryonic body, and the lower nascent hypoblast, future extra-embryonic tissue. At this stage, mRNA for Fgf8 was detectable predominantly in the developing hypoblast (Fig. 1a; arrow in Fig. 1a1). At later pregastrula stages (stage XII–XIII), mRNA for Fgf8 was detectable predominantly in the hypoblast (Fig. 1b); and the Fgf8 transcript was predominant in the posterior half of the embryo. At the early gastrula stages, some posterior epiblast cells migrate into the mid-line region of the embryo and invade the subepiblast space as a mesendoderm mesenchyme (primitive streak). At this stage, Fgf8 was expressed predominantly in the mesendoderm cells beneath the primitive streak (Fig. 1c). A weak signal for Fgf8 was detectable in epiblast cells in the anterior region of the primitive streak (Fig. 1c). No detectable signal was observed in embryos incubated with sense probe (data not shown). RT–PCR was carried out to identify the splicing isoforms of Fgf8 expressed in the blastula to gastrula embryos. Embryonic regions, in which Fgf8 mRNA was detected (Fig. 1a–c), were obtained from stage X to stage 4 embryos and subjected to RT–PCR to detect FGF8 isoforms as described in Sato et al. (2001). As shown in Figure 1d, both Fgf8a and Fgf8b were detected, and the PCR products for Fgf8b were much higher than those of Fgf8a. Other isoforms were not detected. We also examined the expression patterns of Nodal and Chordin in blastula to gastrula embryos (supplementary Fig. 10). There was no detectable signal of Nodal or Chordin in early blastoderm at stage X–XI. Nodal and Chordin were expressed in the posterior epiblast at stage XII–XIV. The results indicated that Fgf8 was expressed in the nascent hypoblast and the most predominant isoform of Fgf8 expressed in the blastula to early gastrula embryos appeared to be Fgf8b.

Figure 1.

Fgf8 expression in pregastrula to gastrula embryos. At stage X–XI, Fgf8 was detectable predominantly in the developing hypoblast (arrow in a1). At stage XII–XIII, Fgf8 was detectable predominantly in the hypoblast and the transcript was predominant in the posterior half of the embryo (b). At stage 3, mesendoderm cells beneath the anterior primitive streak (PS) expressed Fgf8 predominantly (c1–3); and weak signal was detectable in epiblast cells in the anterior PS (c3). Fgf8-positive regions were subjected to reverse transcription–polymerase chain reaction (RT–PCR) and the results showed that Fgf8a and Fgf8b were expressed in these regions, and the expression of Fgf8b was predominant (d). Note that arrowhead (a2, b2) indicates boundaries between the area pellucida and area opaca. epi, epiblast; hypo, hypoblast. Red lines (a1–c1) indicate where histological sections were obtained. Bar, 1 mm (a1, b1, c1); 200 µm (a2, b2, c2); 100 µm (a3, b3, c3).

Anti-FGF8b antibody or SU5402 affects the expression of SMA in cultured posterior blastoderm

We previously reported that interaction between the epiblast and hypoblast at the early blastula stage (stage X–XI) is necessary to express SMA in cultured posterior epiblast; however, we do not know whether FGF8b is a key to regulating tissue interaction involving the expression of SMA. We next examined whether a neutralizing antibody against FGF8b or a FGF-receptor antagonist, SU5402 (Mohammadi et al. 1997), affected the expression of SMA. Stage X–XI posterior blastoderm was cultured in serum-free defined medium with or without the anti-FGF8b antibody or SU5402. After 48 h, explants were doubly stained with antibodies against SMA and sarcomeric α-actinin, a heart mesoderm marker and an early cardiomyocyte marker, respectively (Nakajima et al. 2002). As shown in Figure 2A,B, more than 90% (68/73) of explants cultured in control medium expressed SMA and 62% (45/73) expressed sarcomeric α-actinin with developing sarcomere. In explants treated with anti-FGF8b antibody or SU5402, the number of explants expressing SMA or sarcomeric α-actinin was decreased significantly (Fig. 2A,B; P < 0.01, Fisher's exact test). Western blot showed that the expression of SMA was downregulated in explants treated with anti-FGF8b antibody or SU5402 (Fig. 2C). Our preliminary experiment showed that FGF8b-dsRNA (target sequence was described in Sato & Nakamura 2003) introduced into the posterior hypoblast of stage X–XI embryos by electroporation inhibited the formation of SMA-positive heart mesoderm in cultured embryos (data not shown). PCR analysis showed that the PCR products for heart mesoderm/early cardiomyocyte marker genes, SMA, cTNT, and Tbx5 were not detectable in explants treated with 50 µg/mL of anti-FGF8b antibody. The expression of Nkx2.5 and sarcomeric α-actinin was reduced in explants treated with 50 µg/mL of anti-FGF8b antibody. The expression of pan-mesoderm marker, Brachyury, was not downregulated significantly (Fig. 2D). Either neutralizing anti-FGF2 antibody or anti-FGF4 antibody did not suppress the expression of heart mesoderm/early cardiomyocyte marker genes and Brachyury (data not shown). In explants treated with SU5402, there was no detectable PCR product for these cardiomyocyte marker genes and Brachyury (right lane in Fig. 3B). These results indicated that the perturbation of FGF8b function in cultured posterior blastoderm impaired the expression of heart mesoderm/early cardiomyocyte marker genes including SMA.

Figure 2.

Anti-fibroblast growth factor-8b (FGF8b) antibody or SU5402 inhibited the expression of smooth muscle α-actin (SMA) in cultured posterior blastoderm. Stage X–XI posterior blastoderm was cultured with or without anti-FGF8b antibody or SU5402. After 48 h, cultures were subjected to immunological detection for SMA or sarcomeric α-actinin. (A, B) More than 90% of explants cultured in control medium (medium alone or medium supplemented with normal goat immunoglobulin G [IgG]) expressed SMA and some explants expressed sarcomeric α-actinin (A, Ba1–3). In explants treated with anti-FGF8b antibody or SU5402, the incidence of the expression of SMA or sarcomeric α-actinin was decreased significantly (P < 0.01, Fisher's exact test; A, Bb1–3, Bc1–3). (C) Similarly cultured explants were subjected to Western blotting to detect the expression of SMA and the result showed that the expression of SMA was downregulated when explants were treated with anti-FGF8 antibody or SU5402. (D) Reverse transcription–polymerase chain reaction (RT–PCR) showed that there was no detectable PCR product for SMA, cTNT, and Tbx5 in explants treated with 50 µg of anti-FGF8b antibody. The expression of Nkx2.5 and sarcomeric α-actinin was reduced in explants treated with 50 µg of anti-FGF8b antibody. GAPDH (glyceraldehyde-3-phosphate dehydrogenase) is shown as a normalization control. n, number of explants tested. Bar, 50 µm.

Figure 3.

Time window that anti-fibroblast growth factor-8b (FGF8b) antibody or SU5402 affects the expression of early cardiomyocyte marker genes. To examine whether FGF signal at blastula stage was required for the expression of cardiomyocyte marker genes, stage X–XI posterior blastoderm was cultured and treated with anti-FGF8b antibody (A) or SU5402 (B) at the beginning of cultivation for 5, 10, and 20 h. After additional cultivation with control medium (total incubation time = 48 h), explants were subjected to reverse transcription–polymerase chain reaction (RT–PCR). The expression of SMA, cTNT, and Tbx5 was downregulated in explants treated with anti-FGF8b antibody for 10 h or more (A). In explants treated with SU5402 for 10 h or more, the expression of Brachyury, SMA, cTNT, Tbx5, Nkx2.5 and sarcomeric α-actinin was downregulated (B).

To examine whether an FGF signal at blastula stage was required for the expression of cardiomyocyte marker genes, stage X–XI posterior blastoderm was cultured and treated with anti-FGF8b antibody or SU5402 at the beginning of cultivation for 5, 10, and 20 h. After additional cultivation with control medium (total incubation time = 48 h), explants were subjected to RT–PCR and examined the expression of early cardiomyocyte marker genes and Brachyury. As shown in Figure 3A, the expression of SMA, cTNT, and Tbx5 was downregulated when the explants were treated with anti-FGF8b antibody for 10 h or more (Fig. 3A). The expression of Nkx2.5, sarcomeric α-actinin, and Brachyury was not downregulated significantly even if the explants were treated with anti-FGF8b antibody for 20 h. In explants treated with SU5402 for 10 h, the expression of Brachyury, SMA, cTNT, Tbx5, Nkx2.5 and sarcomeric α-actinin was downregulated (Fig. 3B). Our preliminary RT–PCR showed that the mesoderm marker, Brachyury, was not detectable even if the explants were cultured for 10 h in control medium; and the expression was observed after 15 h in culture (not shown). In situ hybridization showed that the marked expression of Brachyury is observed in primitive streak at stage 3 (Lawson et al. 2001). These results indicated that the inhibition of FGF8b function at the beginning of cultivation in cultured posterior blastoderm was capable of impairing the expression of SMA, cTNT and Tbx5.

Combined FGF8b and Nodal, but neither factor alone, induces the expression of SMA in cultured epiblast

We next investigated whether FGF8b could induce the expression of SMA in cultured posterior epiblast. Posterior epiblast was cultured in medium supplemented with FGF8b or medium with FGF8b plus Nodal. As shown in Figure 4A,B, when the posterior epiblast was cultured in control medium, only 7% (3/43) of explants expressed SMA. In explants cultured with medium containing FGF8b alone, only 24% (6/25) or less expressed SMA. By contrast, in posterior epiblast explants cultured with FGF8b plus Nodal, the number of explants expressing SMA and sarcomeric α-actinin was increased significantly (Fig. 4A,B; P < 0.001, Fisher's exact test). Western blot showed no detectable or faint band of SMA in explants cultured in control medium or medium containing FGF8b alone. Posterior epiblast explants treated with Nodal alone did not express SMA (fig. 7 in Matsui et al. 2005). An intense band was detected in explants treated with 1 µg/mL of FGF8b plus Nodal (Fig. 4C). RT–PCR was carried out to detect pan-mesoderm marker, Brachyury, and several early cardiomyocyte markers (Fig. 4D). A faint band of Brachyury was detectable in explants treated with Nodal alone, but not in explants treated with FGF8b. Posterior epiblasts treated with FGF8b plus Nodal expressed SMA, cTNT, Tbx5, Nkx2.5, and sarcomeric α-actinin. Explants treated with 1 µg/mL of FGF8b plus Nodal expressed extensively these cardiac marker genes (Fig. 4D).

Figure 4.

Combined fibroblast growth factor-8b (FGF8b) and Nodal induced the expression of smooth muscle α-actin (SMA) in cultured posterior epiblast. Stage X–XI posterior epiblast was cultured in medium supplemented with FGF8b or FGF8b plus Nodal protein. After 48 h in culture, explants were subjected to immunological detection for SMA and sarcomeric α-actinin (A, B). When the posterior epiblast was cultured in medium alone, only 7% (3/43) of explants expressed SMA (A, Ba1–3). In explants cultured with medium containing FGF8b alone, only 24% (6/25) or less expressed SMA (A, Bb1–3). By contrast, when the posterior epiblast was cultured in FGF8b plus Nodal protein, 50 (4/8) ~ 84% (11/14) of explants expressed SMA (P < 0.001, Fisher's exact test; A, Bc1–3). Some explants treated with FGF8b plus Nodal expressed sarcomeric α-actinin (A, Bc1–3). (C) Western blotting showed no detectable band of SMA in explants cultured in control medium, while a very faint band was detectable in explants treated with 1 µg/mL of FGF8b. In contrast, an intense band was apparent in explants treated with 1 µg/mL of FGF8b plus Nodal. (D) Reverse transcription–polymerase chain reaction (RT–PCR) showed that a faint band of Brachyury was detectable in explants treated with Nodal alone, but not in explants treated with FGF8b. Posterior epiblasts treated with FGF8b and Nodal expressed SMA, cTNT, Tbx5, Nkx2.5 and sarcomeric α-actinin as well as Brachyury. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; n, number of explants tested; Bar, 50 µm.

We next examined whether FGF8b plus Nodal could induce the expression of SMA and heart-specific markers in cultured anterior epiblast, from which cardiomyocyte will not develop normally (Yatskievych et al. 1997). Anterior epiblast (non-cardiogenic) was cultured in medium with or without FGF8b plus Nodal. As shown in Figure 5, there was no detectable staining for SMA in explants cultured in control medium after 48 h in incubation (0/20; Fig. 5A,B). By contrast, 70% (12/17) of explants from anterior epiblast treated with 1 µg/mL of FGF8b plus Nodal expressed SMA (Fig. 5A,B; P < 0.001, Fisher's exact test). Some explants expressed sarcomeric α-actinin that was incorporated into SMA-positive nascent myofibrils. Western blot showed that explants treated with both factors expressed SMA (Fig. 5C). Similarly cultured anterior epiblasts were subjected to RT–PCR to detect Brachyury and early cardiomyocyte markers (Fig. 5D). Results showed that a faint band of Brachyury was detectable in explants treated with Nodal alone, but not in explants treated with FGF8b alone. Anterior epiblast treated with 100 or 1000 ng/mL of FGF8b plus Nodal expressed Brachyury and cardiac marker genes in a dose-dependent manner (Fig. 5D). There was no detectable band for cardiomyocyte markers in explants treated with FGF8b or Nodal alone (Fig. 5D). PCR products for GATA4 were detected even if explants were cultured in control medium alone (data not shown). These results indicated that combined addition of FGF8b and Nodal, but neither factor alone successfully induced the expression of SMA as well as cardiomyogenesis in not only the posterior epiblast but also the anterior epiblast.

Figure 5.

Fibroblast growth factor-8b (FGF8b) and Nodal induced the expression of smooth muscle α-actin (SMA) in cultured anterior epiblast. Anterior epiblast (non-cardiogenic) was cultured in medium with or without FGF8b plus Nodal protein. After 48 h, explants were subjected to immunological detection for SMA and sarcomeric α-actinin. (A, B) Anterior epiblast explants cultured in control medium did not show any apparent expression of SMA and sarcomeric α-actinin (0/20; A, Ba1–3). By contrast, 70% (12/17) of explants from anterior epiblast treated with 1 µg/mL of FGF8b plus Nodal expressed SMA (A, Bb1; P < 0.001, Fisher's exact test). Some explants treated with both factors expressed sarcomeric α-actinin (A, Bb1–3). (C) Western blot analysis showed that explants treated with 1 µg/mL of FGF8b plus Nodal expressed SMA. (D) Similarly cultured anterior epiblast was subjected to reverse transcription–polymerase chain reaction (RT–PCR) to detect Brachyury and heart mesoderm/heart marker genes. Results showed that a faint band of Brachyury was detectable in explants treated with Nodal alone, but not in explants treated with FGF8b alone. Explants treated with 100 ng or 1000 ng/mL of FGF8b plus Nodal expressed Bracyury and early cardiomyocyte marker genes in a dose dependent manner. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; n, number of explants tested; Bar, 50 µm.

Anti-FGF8b antibody inhibits the expression of Chordin in cultured blastoderm

We previously reported that the addition of Nodal plus BMP-antagonist induces the expression of SMA in cultured posterior epiblast (Matsui et al. 2005). Therefore, we next investigated whether FGF8b was involved in the expression of the BMP-antagonist Chordin in cultured posterior blastoderm. Posterior blastoderm was cultured in medium with or without anti-FGF8b antibody. After 3, 6, and 24 h in culture, explants were subjected to RT–PCR to detect mRNAs for Nodal, Nodal cofactor CFC, and Chordin. As shown in Figure 6A, CFC and Nodal were expressed in explants after 3 h of incubation and the expression of these factors was not affected markedly in cultures treated with anti-FGF8b antibody. In control cultures, PCR products for Chordin were detectable after 6 h of incubation and the expression was suppressed in cultures treated with anti-FGF8b antibody. We next examined whether FGF8b could induce the expression of Chordin in cultured epiblast. Stage X–XI anterior epiblast was cultured in medium with FGF8b, Nodal, or both factors. After 24 h in culture, explants were subjected to RT–PCR to detect mRNAs for Nodal, CFC, and Chordin. Anterior epiblast treated with FGF8b plus Nodal was stimulated to express Chordin, Nodal and Nodal cofactor, CFC (Fig. 6B). There were no detectable PCR products for Chordin and CFC in either control or explants treated with either factor alone (Fig. 6B). The results suggested that FGF8b was likely to be necessary, but not sufficient, to induce the expression of Chordin and that combined FGF8b and Nodal were able to induce the expression of Chordin in cultured epiblast.

Figure 6.

Fibroblast growth factor-8b (FGF8b) combined with Nodal induced the expression of Chordin in cultured epiblast. (A) Posterior blastoderm was cultured in medium with or without anti-FGF8b antibody. After 3, 6, and 24 h in culture, explants were subjected to reverse transcription–polymerase chain reaction (RT–PCR) to detect mRNAs for Nodal, the Nodal cofactor CFC, and Chordin. Results showed that Nodal and CFC were expressed in explants after 3 h, and the expression of these factors was not affected markedly by anti-FGF8b antibody. PCR products for Chordin were detectable after 6 h of incubation and the expression was suppressed in cultures treated with anti-FGF8b antibody. (B) Stage X–XI anterior epiblast was cultured in medium supplemented with FGF8b, Nodal or both factors (1 µg/mL per factor). After 24 h in culture, explants were subjected to RT–PCR to detect mRNAs for Nodal, CFC, and Chordin. Results showed that anterior epiblast treated with FGF8b plus Nodal expressed Nodal, CFC and Chordin. A faint band for Nodal was amplified in explants treated with Nodal. There were no detectable PCR products for Chordin in either control explants (medium alone) or explants treated with Nodal or FGF8b alone.

Anti-FGF8b antibody or SU5402 downregulates phosphorylation of Smad2 in cultured posterior blastoderm

The above experiments suggested that both FGF8b and Nodal were necessary for the expression of SMA as well as heart markers in cultured epiblast. In addition to FGF signaling (Cornell & Kimelman 1994; LaBonne & Whitman 1994), phosphorylation of the C-terminal region of Smad2, downstream of Activin/Nodal signaling, is required for dorsal mesoderm to form (Baker & Harland 1996); however, the interaction between FGF and Activin/Nodal pathways during mesoderm induction remains unclear. To address this issue, we examined the C-terminal phosphorylation of Smad2 (P-Smad2) in cultured posterior blastoderm with or without anti-FGF8b or anti-Nodal property. Posterior blastoderm was cultured in medium supplemented with Lefty1 (anti-Nodal activity), SU5402, or anti-FGF8b antibody. After 3 h of incubation, explants were subjected to Western blot to detect P-Smad2. As shown in Figure 7, P-Smad2 was detectable in posterior blastoderm cultured in control medium after 3 h of incubation. By contrast, the amount of P-Smad2 was reduced in explants treated with Lefty1. Furthermore, the amount of P-Smad2 was downregulated in explants treated with SU5402 or anti-FGF8b antibody (Fig. 7A). We next examined whether FGF8b could induce the phosphorylation of Smad2 without the addition of Nodal protein in cultured posterior epiblast. Posterior epiblast was cultured in medium supplemented with Nodal, FGF8b, or both factors. After 3 h in culture, explants were subjected to Western blot and the results showed no significant P-Smad2 in posterior epiblast treated with FGF8b (Fig. 7B). The results suggested that although FGF8b itself could not induce the phosphorylation of Smad2; FGF8b appeared to be necessary to enhance/maintain the phosphorylation of Smad2 in cultured blastoderm.

Figure 7.

Anti-fibroblast growth factor-8b (FGF8b) properties affected phosphorylation of Smad2. (A) Posterior blastoderm was cultured in medium supplemented with Lefty1 (1 µg/mL), SU5402 (5 µm) or anti-FGF8b antibody (50 µg/mL). After 3 h of incubation, explants were subjected to Western blotting to detect phospho-Smad2 (P-Smad2) and Smad2. Results showed that P-Smad2 was detectable in posterior blastoderm cultured in control medium after 3 h. By contrast, the amount of P-Smad2 was reduced in explants treated with Lefty1, SU5402 and anti-FGF8b antibody. P-Erk was downregulated in explants treated with SU5402 or anti-FGF8b antibody (not shown). (B) Posterior epiblast was cultured in medium containing Nodal, FGF8b, or both factors (1 µg/mL). After 3 h in culture, explants were subjected to Western blotting to detect P-Smad2 and Smad2. There was no significant or faint band of P-Smad2 in explants cultured in control medium or medium supplemented with FGF8b protein. P-Smad2 was apparent in explants treated with Nodal and Nodal plus FGF8b. P-Erk was upregulated significantly in cultures treated with FGF8b or both factors (not shown).

FGF8b is able to reverse the BMP-induced inhibition of cardiomyogenesis

It has been reported that BMP inhibits cardiomyogenesis in cultured posterior blastoderm (Ladd et al. 1998); however, neither transforming growth factor-β (TGFβ) nor Activin is capable of reversing this inhibitory effect (Matsui et al. 2005). We next examined whether FGF8b could reverse the inhibitory effect of BMP on cardiogenesis in cultured posterior blastoderm. Posterior blastoderm was cultured in medium containing BMP2, or BMP2 plus FGF8b. After 48 h, explants were subjected to immunohistochemical staining for SMA and sarcomeric α-actinin. As previously reported, more than 90% (17/18) of the posterior blastoderm explants cultured in control medium expressed SMA and some expressed sarcomeric α-actinin, whereas only 26% (6/23) of explants cultured in medium with BMP2 expressed SMA (P < 0.001). In explants cultured in medium containing BMP plus FGF8b, the number of explants expressing SMA and sarcomeric α-actinin was increased significantly (Fig. 8A,B, P < 0.001, Fisher's exact test). Western blot showed that the expression of SMA was significant in explants cultured in BMP plus FGF8b, while only a faint band for SMA was detectable in explants treated with BMP2 protein (Fig. 8C). RT–PCR showed that the amounts of PCR products for FGF8b, Nodal, CFC and Chordin were decreased in cultures treated with BMP2 after 24 h in culture (Fig. 8D). In explants cultured in medium containing BMP2 plus FGF8b, the amounts of PCR products for these factors were increased (Fig. 8D). We further examined the amounts of P-Smad2 and P-Smad1/5/8 in explants treated with BMP2 or BMP2 plus FGF8b. Western blot showed that the amount of P-Smad2 was reduced in explants treated with BMP2, whereas the amount of P-Smad2 was reversed in explants treated with both factors (Fig. 8E). P-Samd1/5/8 was upregulated in explants treated with BMP2 or BMP2 plus FGF8b (Fig. 8E). These results indicated that FGF8b reversed the BMP-induced inhibition of SMA expression and cardiomyogenesis, in which the expressions of Fgf8b, Nodal and Chordin as well as phosphorylation of Smad2 were affected.

Figure 8.

Fibroblast growth factor-8b (FGF8b) was able to reverse the bone morphogenetic protein (BMP)-induced inhibition of cardiomyogenesis. (A, B) Posterior blastoderm was cultured in various culture conditions (control medium alone, medium containing 100 ng/mL of BMP2, and medium containing BMP2 plus FGF8b). After 48 h in culture, explants were subjected to immunohistochemical staining for smooth muscle α-actin (SMA) and sarcomeric α-actinin. In control explants, 94% (17/18) of posterior blastoderm explants expressed SMA and some expressed sarcomeric α-actinin (A, Ba1–3), whereas only 26% (6/23) of explants cultured in medium supplemented with BMP2 expressed SMA (P < 0.001, Fisher's exact test; A, Bb1–3). By contrast, 84% (21/25) of explants cultured in medium containing BMP2 plus 1 µg/mL of FGF8b expressed SMA and some explants expressed sarcomeric α-actinin (A, Bc1–3). (C) Western blot analysis showed that only a faint band for SMA was detectable when explants were treated with BMP2 protein. By contrast, the expression of SMA was reversed in explants cultured in BMP plus FGF8b. (D) Posterior blastoderm was cultured in medium supplemented with recombinant BMP2 or BMP2 plus FGF8b. After 24 h in culture, explants were subjected to reverse transcription–polymerase chain reaction (RT–PCR) to detect mRNAs for FGF8b, Nodal, CFC, and Chordin. Results showed that the amounts of PCR products for these genes were decreased in cultures treated with recombinant BMP2. By contrast, when explants were cultured in medium containing BMP2 plus FGF8b, the amounts of PCR products were increased. (E) Similarly prepared explants cultured for 3 h were subjected to Western blot to examine C-terminal phosphorylation of Smad2 (P-Smad2) or Smad1/5/8 (PSmad1/5/8). Results showed that P-Smad2 was reduced in explants treated with BMP, while the phosphorylation of Smad2 was increased in explants treated with both FGF8b and BMP2. P-Smad1/5/8 was detectable in explants treated with BMP2 or BMP2 plus FGF8b. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; n, number of explants tested; Bar, 50 µm.

Discussion

Fgf8b is expressed in regions from which heart cells will later develop

In the present study, we showed that Fgf8 was expressed in the nascent hypoblast, just beneath the posterior epiblast at blastula stage. At early gastrula stages, Fgf8 was expressed in the anterior region of the primitive streak and early ingressive mesendoderm. RT–PCR showed that PCR products for Fgf8a and Fgf8b were detectable in these regions and the expression of Fgf8b was predominant; therefore, observations suggest that Fgf8b was expressed in regions where prospective heart cells reside. In mice, Fgf8 is alternatively spliced to generate at least seven secreted isoforms that differ only at their mature amino terminus, and FGF8b and FGF8c, but not FGF8a, have potent biological activity in limb and craniofacial development (MacArthur et al. 1995). In zebrafish embryos, both Fgf8a and Fgf8b are expressed, but the expression of Fgf8b is predominant. Misexpression experiments of these two isoforms in zebrafish embryos following mRNA injection showed that both isoforms have dorsalizing activity with more potent FGF8b (Inoue et al. 2006). It is also reported that FGF8 is necessary for the proper gastrulation and formation of mesoderm and that FGF8b is the predominant FGF8 isoform involved in early mesoderm development in Xenopus (Fletcher et al. 2006). We showed that anti-FGF8b antibody suppressed the expression of SMA, cTNT and Tbx5, which are BMP-independent heart mesoderm/early cardiomyocyte genes (Yamada et al. 2000; Antin et al. 2002; Nakajima et al. 2002) in cultured posterior blastoderm, and that combined FGF8b and Nodal induced the expression of early cardiomyocyte marker genes as well as Brachyury in cultured non-cardiogenic epiblast. Taken together, these observations suggest that during early chick development at the blastula to gastrula stage, Fgf8b appears to be expressed in regions where presumptive heart cells reside, and that FGF8b may play a role in the regulation of heart mesoderm formation as well as early cardiomyogenesis.

Combined FGF8b and Nodal is necessary for the expression of SMA in cultured epiblast

It has been reported that signaling from the hypoblast is required as an upstream of endoderm-derived heart inducing signal during early blastoderm stages (Yatskievych et al. 1997). Nodal is an endogenous molecule implicated in the expression of SMA (early heart mesoderm marker) in cultured blastoderm and Nodal plus Chordin effectively induces the expression of SMA in cultured epiblast (Matsui et al. 2005); however, signaling molecule(s) emitted from the hypoblast that act on the epiblast to generate SMA-positive heart tissue, is uncertain, because Nodal is expressed in the posterior epiblast and Nodal itself is not able to induce SMA expression in cultured epiblast (Chapman et al. 2002; Matsui et al. 2005). In the present study, Fgf8b was transcribed in the nascent hypoblast at stage X–XI; 10 h-inhibition of FGF8b activity at the beginning of posterior blastoderm culture was able to suppress the expression of Chordin, SMA, cTNT and Tbx5; and FGF8b combined with Nodal, but neither factor alone, induced the expression of Chordin and SMA in association with heart-specific marker genes in the anterior epiblast (which normally does not give rise to cardiomyocyte). Therefore, it is likely that FGF8b is a candidate molecule secreted by hypoblast and regulates the heart mesoderm formation/early cardiomyogenesis.

A requirement of Activin/Nodal and FGF signaling as well as interaction between the two signaling pathways has long been thought to play an important role in mesoderm induction (Cornell & Kimelman 1994; Mathieu et al. 2004). Mouse null-mutants for Nodal do not develop a primitive streak and thus fail to form a mesoderm (Conlon et al. 1994). In Xenopus, inhibition of Nodal signaling by Lefty or Cerberus-S blocks formation of the mesendoderm (Agius et al. 2000; Cheng et al. 2000). In Fgfr1-null or Fgf8-null mutants, severe gastrulation-related defects are observed (Yamaguchi et al. 1994; Sun et al. 1999). In zebrafish, Nodal and Fgf8 act synergistically to maintain the mesoderm cell population (Mathieu et al. 2004). In our experiments, although anti-FGF8b antibody did not suppress the expression of pan-mesoderm marker, Brachyury, in cultured posterior blastoderm, it did downregulate the expression of BMP-independent heart mesoderm/early cardiomyocyte genes, SMA, cTNT, and Tbx5 (Yamada et al. 2000; Antin et al. 2002; Nakajima et al. 2002). SU5402 suppressed the expression of Brachyury and heart mesoderm/cardiomyocyte genes. In mice with a splice-site mutation abolishing Fgf8b expression shows that Fgf8a can compensate for the loss of Fgf8b in the expression of Brachyury during gastrulation (Guo & Li 2007). In chick cardiogenesis at the gastrula stage, endoderm-derived FGF signaling cooperates with BMP to regulate early cardiogenesis; however, anti-FGF8 properties, such as neutralizing antibody, SU5402, and truncated Fgf-receptor fail to inhibit cardiogenesis (Alsan & Schultheiss 2002). Taken together, it is suggested that in addition to mesoderm induction at the early blastula stage, FGF8b may play a specific role in the formation of heart mesoderm.

Another experiment showed that although the anti-FGF8b antibody did not suppress the expression of Nodal or Nodal cocofactor CFC in cultured posterior blastoderm, it did downregulate the expression of Chordin, with which Nodal acts cooperatively to induce the expression of SMA (Matsui et al. 2005). In late chick blastoderm, Nodal is expressed in the posterior epiblast confined to the region composing the middle two-thirds of the sickle, and Chordin is expressed in the mid-posterior epiblast just anterior to the sickle (Lawson et al. 2001); therefore, an organizer factor, Chordin, is expressed in the posterior region of the epiblast, where both FGF8b and Nodal signaling seem to act intensely (Fig. 9). Bertocchini et al. (2004) reported that FGF signaling is required for the primitive streak to form in cooperation with Nodal and Chordin during gastrulation in chick. Taking our observations together with others, it is suggested that hypoblast secreted FGF8b appears to act on the posterior epiblast synergistically with Nodal and plays a role in the expression of SMA, suggesting that combined FGF8 and Nodal activity at blastula stage are important for heart mesoderm formation (Fig. 9).

Figure 9.

Schematic drawing to show the interaction of tissues between the posterior epiblast and hypoblast is necessary for cardiogenesis. At pregastrula stage, posterior hypoblast-derived fibroblast growth factor-8 (FGF8) acts on the posterior epiblast synergistically with epiblast-derived Nodal. Chordin is expressed in the posterior epiblast and affects bone morphogenetic protein (BMP) activity in the posterior region. Posterior epiblast, in which FGF8, Nodal and anti-BMP activity are predominant, develops anterior lateral mesoderm to express smooth muscle α-actin (SMA), cardiac troponin T (cTNT) and Tbx5 independently of endoderm-derived signals. Endoderm-derived signals including BMP2/4 and FGF8 act on the anterior lateral mesoderm and thereby regulate the heart specification as well as terminal differentiation.

FGF8b is necessary to maintain Nodal signaling in chick blastoderm

Numerous studies have shown that Activin/Nodal/Smad2/3 and FGF/mitogen-activated protein kinase (MAPK) signaling pathways act synergistically on the ectoderm/epiblast and thereby induce the formation of mesoderm during gastrulation (Gotoh et al. 1995; Umbhauer et al. 1995; Baker & Harland 1996; Nomura & Li 1998; Yao et al. 2003; Dunn et al. 2004; Stern 2004; Sivak et al. 2005); however, the mechanisms regulating the interaction between the two pathways are largely unknown. In the present study, we showed that inhibition of the endogenous FGF8b function downregulated P-Smad2 in cultured posterior blastoderm, but exogenously given FGF8b failed to upregulate P-Smad2 in preactivated epiblast obtained from stage X–XI blastoderm. Therefore, the results suggested that FGF8b appeared to maintain/enhance the phosphorylation of the C-terminal region of Smad2 activated by Nodal in the posterior blastoderm. Both positive and negative regulatory interaction between TGFβ/Smad2 and MAPK pathways has been reported (Feng & Derynck 2005; Javelaud & Mauviel 2005). For example, c-Jun N-terminal kinase (JNK), which is activated by mitogenic and stress signals, and phosphorylates Smad3, resulting in its nuclear transfer (Engel et al. 1999); and the activation of MAPK kinase-1, an activator of JNK and extracellular-signal-regulated kinase (Erk)/MAPK, leads to the phosphorylation and activation of Smad2 (De Caestecker et al. 1998; Brown et al. 1999). Erk/MAPK, which is activated in response to mitogenic growth factors or oncogenic Ras mutants, can phosphorylate the linker regions of Smad1 and Smad2/3, thereby inhibiting the ligand-induced nuclear translocation of Smads (Kretzschmar et al. 1997, 1999; Pera et al. 2003). In Xenopus development, FGF8, in combination with IGF2, induces MAPK-dependent inhibitory phosphorylation of the Smad1 linker region, which contributes to the induction of neural cell fate (Pera et al. 2003). Recently, it has been reported that receptor tyrosine kinase/Ras/MAPK activity induces p53 N-terminal phosphorylation, enabling the interaction of p53 with the TGFβ-activated Smads for mesoderm specification in Xenopus embryos (Cordenonsi et al. 2007). Our results suggest that Nodal and FGF8b stimulate distinct receptors to activate downstream signaling pathways, within which the FGF/MAPK pathway may maintain/enhance the phosphorylation of the C-terminal region of Smad2, thereby sustaining or increasing Nodal signaling. However, the mechanism(s) regulating the enhancement of the phosphorylation of Smad2 by FGF signal remains unknown.

FGF8b reverses the BMP-induced suppression of cardiomyogenesis

It has been reported that BMP inhibits cardiomyogenesis from the posterior blastoderm in culture (Ladd et al. 1998). In addition to cardiomyogenesis, BMP is well known as a potent inhibitor for neural induction, organizer formation, primitive streak formation and mesoderm patterning in various vertebrates (Streit et al. 1998; Streit & Stern 1999; Koshida et al. 2002; Marom et al. 2005). In the present study, we showed that exogenously added FGF8b was able to reverse the BMP-induced inhibition of cardiomyogenesis in cultured posterior blastoderm. BMP suppressed the expression of Fgf8b, Nodal, Chordin as well as the phosphorylation of Smad2 in cultured posterior blastoderm, and this inhibition was reversed by the addition of FGF8b. C-terminal phosphorylation of Smad1/5/8 was induced by BMP2, but maintained in cultures treated with BMP2 and FGF8b. Therefore, it is suggested that Nodal/Smad2-dependent mesoderm induction is capable to occur in vitro even if the BMP/Smad1 signaling is presenting. During chick gastrulation, the phosphorylation of Smad1 is extinguished at the site of the gastrulation/primitive streak, in which Fgf8, Nodal, and Chordin are expressed extensively. In contrast, phospho-Smad1 is ubiquitously distributed in other regions of the epiblast (Faure et al. 2002). In zebrafish, Fgf3 induces the expression of Chordin in the dorsolateral margin to induce neural induction at the blastula to gastrula stages (Koshida et al. 2002). Therefore, it seems likely that FGF is required to upregulate or maintain the expression of Chordin and thereby suppress BMP activity extracellularly at the site where gastrulation occurs. Furthermore, we showed that FGF signaling appeared necessary to maintain/enhance the phosphorylation of Smad2 in cultured blastoderm (Fig. 7). Taken together with these results, it is suggested that in addition to the induction of organizer factor Chordin, FGF8 may play a role in the maintenance of nuclear transport of Activin/Nodal/Smad2 signaling; thereby FGF8 is able to liberate the prospective dorsal mesoderm including heart from the BMP activity.

In conclusion, FGF8b is expressed in nascent hypoblast in early chick blastoderm; and FGF8b acts on the epiblast synergistically with Nodal at the pregastrula stage and may play a specific role in the expression of SMA during early cardiogenesis.

Acknowledgments

The authors thank Ms S. Uoya for technical assistance. This work was supported by Japan Society for Promotion of Science (JSPS) Grant-in-Aid for Scientific Research (#17590169 and #17-3456); The Takeda Science Foundation; Terumo Life Science Foundation; Miyata Heart Foundation; The Naito Foundation; Mitsubishi Pharma Research Foundation.

Ancillary