SEARCH

SEARCH BY CITATION

Keywords:

  • chick embryo;
  • proepicardium;
  • fibroblast growth factor;
  • cell proliferation;
  • apoptosis;
  • gene expression;
  • Tbx18;
  • Wt1;
  • Tbx5

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The proepicardium forms at the venous pole of the embryonic heart and gives rise to several cell types of the mature heart. We investigated the role of fibroblast growth factors (FGFs) during proepicardium formation in the chick embryo. Several FGF ligands (Fgf2, Fgf10, and Fgf12) and receptors (Fgfr1, Fgfr2, and Fgfr4) are expressed in the proepicardium. Experimental modulation of FGF signaling in explant cultures affected cell proliferation and survival. In contrast, expression of Tbx18, Wt1, or Tbx5 were unaffected by FGF inhibition. In agreement with the explant data, villous outgrowth of the proepicardium was strongly impaired by FGF inhibition in vivo, however Tbx18 expression was maintained. These data suggest that during proepicardium formation, FGF ligands act as autocrine or paracrine growth factors to prevent apoptosis, maintain proliferation, and to promote villous outgrowth of the proepicardium. However, FGF is not involved in the induction or maintenance of proepicardium-specific marker gene expression. Developmental Dynamics 239:2393–2403, 2010. © 2010 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Cells of the epicardium arise from a mesodermal cell population that forms a protrusion, called the proepicardium (PE) immediately posterior to the presumptive atrium on the ventral surface of the sinus horn (Männer et al.,2001). The PE is composed of at least two morphologically distinct cell types, an external mesothelial epithelium and a mesenchymal core, which is rich in extracellular matrix (Nahirney et al.,2003). As soon as cell fate is specified at the sinus horn region, PE cells undergo a dramatic change in morphology and growth phase. In the avian embryo, proepicardial cells develop multiple finger-like protrusions or villi into the pericardial coelomic cavity (Ho and Shimada,1978; Viragh et al.,1993). The villi target the dorsal surface of the atrioventricular junction forming a tissue bridge that extends across the coelomic cavity. After reaching the cardiac surface, PE-derived mesothelial cells spread over the myocardial surface forming the epicardium (Männer,1992; Viragh et al.,1993). Epicardium-derived cells differentiate into multiple cell lineages such as fibroblasts, smooth muscle, and endothelial cells, but no significant contribution to the myocardial lineage have been reported for the chick embryo (Mikawa and Gourdie,1996; Männer et al.,2001).

With the help of marker genes, PE formation can first be visualized at Hamburger and Hamilton (HH) stage 11–12 (Hamburger and Hamilton,1951; Schlueter et al.,2006). Morphological signs of PE formation as a unilateral outgrowth of the pericardial serosa covering the right sinus horn appear at HH stage 14 (Männer et al.,2001). On the left body side, a mesothelial PE-like outgrowth develops; however, its formation is temporally retarded and subsequently the vestigial left-sided PE is eradicated by apoptosis (Schulte et al.,2007). Left–right asymmetry is also reflected by marker gene expression, which during early stages of proepicardial development is only seen on the right side (Schlueter et al.,2006; Schlueter and Brand,2009). Proepicardial asymmetry is controlled by a right-sided signaling pathway, which includes Fgf8 and Snai1, while the left-sided Nodal-Pitx2 pathway is dispensable for PE formation (Schlueter and Brand,2009).

The PE develops adjacent to the sinus myocardium, and there is the possibility that it arises from the lateral margins of the heart fields (Moorman et al.,2007). Nascent myocardial and proepicardial cells co-express marker genes of mesothelial cells and cardiac myocytes (Kruithof et al.,2006; Schlueter et al.,2006). Moreover, DiI (1,1′, di-octadecyl-3,3,3′,3′,-tetramethylindo-carbocyanine perchlorate) labeling experiments are consistent with a common origin of both cell populations from the lateral plate mesoderm (van Wijk et al.,2009). A common origin is also suggested by the ability of proepicardial explants to undergo cardiac myocyte differentiation in vitro (Kruithof et al.,2006; Schlueter et al.,2006). The exact mechanism that induces PE formation at the venous pole region is currently unknown. Because the PE is closely apposed to the liver bud and sinoatrial myocardium, both of these tissues might play a role in the induction process (Männer,2006; Ishii et al.,2007).

Induction of cardiogenic mesoderm depends upon paracrine signals from the underlying foregut endoderm including bone morphogenetic protein 2 (BMP2), fibroblast growth factor 8 (FGF8) and antagonists of canonical Wnt signaling (Schultheiss et al.,1995). A similar set of signals might also be involved in PE induction. We previously demonstrated, that BMP signaling is required to maintain PE marker gene expression (Schlueter et al.,2006). In this study, we have studied the role of FGF signaling for PE-specific gene expression, cell proliferation, and survival. Several FGF genes and FGF receptors are expressed in the PE. Gain- and loss-of-function experiments using cultured PE explants and cell implantation experiments in vivo established an important role of FGF to control cell proliferation and survival. We show here, that FGF, in contrast to BMP, is not required for the induction or maintenance of PE-specific marker gene expression. Thus, during PE formation, independent functions can be defined for these two classes of signaling molecules.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Expression Analysis of FGF Ligands and Receptor Genes During PE Formation

We analyzed the expression of several FGF ligands (Fgf2, -3, -4, -8, -10, -12, -13, -14, -18, -19) and FGF receptors (Fgfr1-4) during PE formation. The pattern of expression for FGF10 and FGF12 was similar to Tbx18 and expression was predominantly present in the epithelial layer of the PE, while FGF2 was found in the mesenchymal layer (Fig. 1A–D,I). Some FGF ligands were expressed at the base of the sinus horn myocardium including Fgf8, while others (Fgf10 and Fgf12) were present in the pericardium adjacent to the PE (Fig. 1B,D,I and data not shown). We observed FGF receptor expression in the PE, which included Fgfr1, Fgfr2, and Fgfr4, while Fgfr3 was not detected (Fig. 1E–H,I). Expression of FGF receptors was largely confined to the epithelial layer of the PE. For both sets of genes, proepicardial expression was confirmed by reverse transcriptase-polymerase chain reaction (RT-PCR) analysis (Fig. 1J).

thumbnail image

Figure 1. Fibroblast growth factor (FGF) ligand and receptor expression in the proepicardium (PE). A–H: Nonradioactive in situ hybridization of sagittal (A–D) or transverse (E–H) sections of the inflow tract region of Hamburger and Hamilton (HH) stage 17/18 chicken embryos using probes for Tbx18 (A), Fgf10 (B), Fgf2 (C), Fgf12 (D), Fgfr1 (E), Fgfr2 (F), Fgfr3 (G), Fgfr4 (H). I: Schematic depiction of the various expression domains of FGF ligands and receptors. J: Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of Fgf2, Fgf10, Fgf12, Fgfr1, Fgfr2, and Fgfr4. For control purposes, the expression of glyceraldehyde dehydrogenase (Gapdh) was analyzed. The arrow in H points to Fgfr4 expression in the nascent epicardium. A, atrium; PE, proepicardium; V, ventricle; SV, sinus venosus; E, epicardium.

Download figure to PowerPoint

Functional Analysis of FGF Signaling in PE Explant Cultures

The role of FGF for proepicardial growth was studied using PE explants that were cultured under serum-free conditions. Between 16 and 64 hr after explantation, the diameter of the explants increased by approximately 250% (Fig. 2A–C). Addition of the FGF receptor antagonist SU5402 to the culture medium resulted in a dose-dependent inhibition of epithelial outgrowth (Fig. 2A,D–G). A similar effect was observed after culturing PE explants in the presence of the ligand-binding domain of Fgfr1 (Fig. 2A,H,I). Two well-established signaling pathways that are downstream of FGF receptor activation are the Ras/MAPK and PI-3 kinase/Akt signaling pathways (Eswarakumar et al.,2005). To determine, whether outgrowth of PE explants was controlled by either pathway, specific antagonists of MEK1 (U0126) and PI-3 kinase (LY294002) were used, and both inhibitors used at concentrations that previously have been reported to interfere with kinase activity (Kang and Sucov,2005), negatively affected proepicardial outgrowth (Fig. 2A,J–M). We also analyzed the effect of exogenous addition of FGF2 to PE explant cultures and observed a strong growth-inducing effect. Addition of 2 μg/ml FGF2 led to a 10-fold increase in explant diameter and also induced epithelial–mesenchymal transition (Fig. 2A,N–Q). This was particularly apparent along the edges of the explant, an effect that has been previously also observed in the case of serum and platelet-derived growth factor (PDGF) treatment (Lu et al.,2001). In summary, modulation of FGF signaling in PE explant cultures revealed an important role as growth regulator.

thumbnail image

Figure 2. Modulation of fibroblast growth factor (FGF) signaling in explant cultures reveals an important role for outgrowth control in vitro. A–O: Quantification of the surface area and phase contrast images of proepicardium (PE) explants after 16 hr (B,D,F,H,J,L,N) and 64 hr (C,E,G,I,K,M,O) of culture in serum-free medium containing (B,C), or 10 (D,E) and 100 μM (F,G) SU5402, 30 μg/μl recombinant ligand binding domain of FGFR1 (H,I), 10 μM LY294002 (J,K), 5 μM U0126 (L,M), or 2 μg/ml FGF2 (N,O). P,Q: Magnification of control PE cells (P) and PE cells conditioned with 2 μg/ml FGF2 (Q). Dashed lines in B–O demarcate the explant border. *P ≤ 0.0001; #P ≤ 0.001; §≤ 0.05. Scale bars = 500 μm in B–O, 250 μm in P–Q.

Download figure to PowerPoint

Role of FGF Signaling for Cell Proliferation and Apoptosis in PE Explants

To further analyze the effects of FGF inhibition on cell proliferation and apoptosis, BrdU incorporation and TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling) staining were used. Explants were cultured in the presence of 10 μM SU5402 or dimethyl sulfoxide (DMSO) for 64 hr. While under control conditions an extensive amount of cells showed BrdU incorporation (n = 5; Fig. 3A); this was largely blocked in the presence of SU5402 (n = 5; Fig. 3B). Moreover, apoptosis increased to a three-fold higher level than under control culture conditions (n = 5; Fig. 3C–E). Thus, the reduction in cellular outgrowth in response to FGF inhibition either by SU5402 or soluble Fgfr1 is probably mediated to a large extent by a loss of cell proliferation and simultaneously by the induction of cell apoptosis. We also quantified the effects of U0126 and LY294002. Both drugs increased the level of apoptosis and interfered with cell proliferation, suggesting that both pathways are involved in controlling cell survival and proliferation in PE cells (Fig. 3E)

thumbnail image

Figure 3. Reduction in cell proliferation and induction of apoptosis in explant cultures after inhibition of fibroblast growth factor (FGF) signaling. A–D: Immunofluorescent detection of bromodeoxyuridine (BrdU) incorporation (A,B) and TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling) staining (C,D) of control and SU5402-treated PE explants after 64 hr of culture. Dotted lines in (A–D) demarcate the border of explant outgrowth and the dashed lines the mesenchymal core region, which due to the high cell density was not considered for quantification of total cell number, apoptosis, or BrdU incorporation. E: Quantification of BrdU incorporation (red column) and TUNEL staining (green column) in control, SU5402, LY294002, and U0126-treated PE explants. The data are expressed as percentage of labeled cells relative to the total cell number. *P = 0.0001 compared with control. Scale bars = 500 μm.

Download figure to PowerPoint

FGF Signaling Does Not Control PE Marker Gene Expression

FGF signaling might also be involved in the control of proepicardial gene expression as has been previously demonstrated for BMP2 (Schlueter et al.,2006). The expression of three marker genes that previously have been implicated into PE formation and growth control, i.e., Tbx18, Wt1, and Tbx5 (Hatcher et al.,2004; Schlueter et al.,2006) were analyzed by quantitative real-time PCR. However, expression of neither gene was affected by FGF inhibition (Fig. 4A). Similarly, Wt1 and Tbx5 antibody staining revealed no alterations in the expression level or nuclear localization (Fig. 4B–E), suggesting that FGF, unlike BMP (Hatcher et al.,2004; Schlueter et al.,2006), does not control PE-specific marker gene expression.

thumbnail image

Figure 4. Altered fibroblast growth factor (FGF) signaling does not affect marker gene expression in explant cultures. A: Quantitative polymerase chain reaction (qPCR) analysis of Tbx18, Wt1, and Tbx5 expression in proepicardium (PE) cells cultured in the presence (white columns) or absence (blue columns) of 10 μM SU5402. Data are represented as relative levels compared with the level of cDNA samples of control cultures, which were set to 1. B–E: Immunofluorescent detection of Wt1 (B,C) and Tbx5 (D,E) in PE cells cultured in the absence (B,D) or presence (C,E) of SU5402.

Download figure to PowerPoint

In Vivo Analysis of FGF-Dependent Villous Outgrowth During PE Formation

To analyze the role of FGF during PE formation in vivo, DF1 cells were used that were co-transfected with RCAS viruses producing the ligand-binding domains of Fgfr1, Fgfr2, and Fgfr4 (FGFR cells). For control purpose, alkaline phosphatase (AP) expressing cells were used. The cell aggregates were DiI-labeled and implanted into the sinus horn of HH stage 12 embryos (Fig. 5A). The embryos were cultured until they had reached HH stage 16 (Fig. 5B,C) or HH stage 18 (Fig. 5D,E). In the presence of control cell aggregates, the majority of embryos (9/11 at HH stage 16; 9/10 at HH stage 18) displayed normal villous outgrowth of the PE. However, in the presence of FGFR cells, the majority of embryos (7/9 at HH stage 16; 7/10 at HH stage 18) displayed severely reduced villous outgrowth of the PE.

thumbnail image

Figure 5. Implantation of fibroblast growth factor receptor (FGFR) -expressing cells in the right sinus horn interferes with proepicardial villi formation. A–E: The fluorescent DiI (1,1′, di-octadecyl-3,3,3′,3′,-tetramethylindo-carbocyanine perchlorate) signal was merged with the phase contrast image to visualize both the cell implant and the embryonic heart (A), or the proepicardium (PE) (B–E). A: Ventral view of the heart region of a Hamburger and Hamilton (HH) stage 12 chicken embryo, with a DiI-labeled cell aggregate, which was implanted into the right sinus horn. B–E: Left ventrolateral view of the PE outgrowth at HH stage 16 (B,C) or HH stage 18 (D,E) of chicken embryos which were implanted with control (B,D) or FGFR-expressing (C,E) cell aggregates. The arrow in A points to the cell implant in the right sinus horn. To allow better visualization of the PE, the tubular heart distal of the inflow tract was removed in the case of the embryos depicted in B–E. Arrowheads in B–E point to the PE outgrowth at the venous pole of the heart. RS, right sinus; LF, left sinus.

Download figure to PowerPoint

To further study the effect of FGF inhibition on PE growth, implanted embryos were subjected to fluorescent whole mount in situ hybridization using a Tbx18 probe. The expression domain was visualized and the volume quantified after three-dimensional (3D) reconstruction of consecutive sections. At HH stage 16, the proepicardial Tbx18 expression domain had a volume of 6 × 105 μm3 (n = 3; Fig. 6A,B,I). This value increased to 8.3 × 105 at HH stage 18 (n = 3; Fig. 6E,F,I). However, implantation of FGFR cell aggregates severely impaired PE growth. At HH stage 16, the mean volume of the PE was approximately 50% of that of a control implanted embryo, i.e., 2.9 × 105 μm3 (n = 3; P = 0.0001; Fig. 6C,D,I). At HH stage 18, the volume of the PE of FGFR-implanted embryos grew in size but the volume still was 50% less than that of a PE in control embryos, i.e., 4.2 × 105 μm3 (n = 3; P = 0.0001; Fig. 6G,H,I). Moreover, in contrast to the extensive outgrowth that was present in control embryos, in FGFR-treated embryos, the PE had a stump-like appearance. Significantly however, despite the severe reduction in volume and shape of the PE, the remaining proepicardial cells retained Tbx18 expression, suggesting that cell identity is not under the control of FGF signaling.

thumbnail image

Figure 6. Inhibition of fibroblast growth factor (FGF) signaling reduces proepicardium (PE) growth. A–H: Fluorescent in situ hybridization of sagittal sections through the PE (A,C,E,G) and three-dimensional (3D) reconstructions (B,D,F,H) of Tbx18 expression in the right inflow tract of embryos implanted at HH stage 12 with either control (A,B,E,F) or FGF receptor (FGFR) -expressing (C,D,G,H) cells. The embryos were cultured until HH stage (A–D) 16, or (E–H) 18. I: Volume measurement of 3D reconstructed proepicardia of embryos that were implanted with control cells (blue columns) or FGFR cells (white columns). *P = 0.0001 compared with control.

Download figure to PowerPoint

We were concerned that the growth inhibitory effect of the FGFR implant might also interfere with the development of other cell types within or adjacent to the inflow tract. Sinoatrial myocardium and the liver primordium have both been implicated in PE induction (Ishii et al.,2007; Mikawa and Brand,2010); hence, we focused our attention on these tissues. We analyzed AMHC, myosin heavy chain, and smooth muscle α-actin expression. However, none of these markers was affected by the FGFR cell implant (Fig. 7A,B,E,F,I–L). Similarly the expression domain of Hex, a marker for the liver primordium (Ishii et al.,2007), was unaffected by the FGFR implant (Fig. 7C,D,G,H). Thus, based on this analysis, we can conclude that FGF inhibition specifically affected villous outgrowth of the PE.

thumbnail image

Figure 7. Inhibition of fibroblast growth factor (FGF) signaling does not affect cardiac or liver marker gene expression. A–D: Whole mount in situ hybridization of control (A,C) and FGF receptor (FGFR) -implanted (B,D) embryos using AMHC1 as a myocardial (A,B) and Hex as a liver-specific (C,D) marker. E–H: Parasagittal sections through the inflow tract region of the embryos depicted in A–D. I–L: Immunohistochemical staining of transverse sections through the inflow tract region of control (I,K) and FGFR-implanted (J,L) embryos using antibodies directed against smooth muscle α-actin (SMA; I,J) and myosin heavy chain (K,L). Asterisks demarcate the location of the cell implant. Black arrowheads in A,C,E,F and white arrowheads in I,K point to the fully grown proepicardium (PE). Red arrowheads in B,D,F,H,J,L point to the growth-retarded PE after FGFR implantation. Arrows in I,J point to the SMA-positive sinus myocardium in the direct vicinity of the cell implant. A, atrium; PE, proepicardium; V, ventricle; SV, sinus venosus.

Download figure to PowerPoint

In Vivo Analysis of FGF-Dependent Growth Control of PE Formation

To find out whether the observed changes in shape and volume of the PE were due to a loss of PE cells through apoptosis, sections of embryos, which have been implanted with aggregates of FGFR- or control cells, were stained using an antibody directed against active Caspase 3. The right sinus horn and the PE of control embryos at HH stage 16 or 18 showed only a few apoptotic cells (n = 3; Fig. 8A,C,E). In contrast, the left sinus horn was strongly labeled and thus contained large amounts of apoptotic cells (Fig. 8A, C,E). In FGFR-implanted embryos both at HH stage 16 and 18, the right sinus horn including the PE displayed a significantly increased level of apoptosis (n = 3; Fig. 8B,D,E). At HH stage 18 but not HH stage 16, also the labeling index in the left sinus horn of FGFR implanted embryos was significantly increased (Fig. 8B,D,E). In addition, cell proliferation was also severely affected by FGFR cell implants (Fig. 8F–J).

thumbnail image

Figure 8. Inhibition of fibroblast growth factor (FGF) signaling in vivo results in an increased level of apoptosis and a reduction in proepicardial cell proliferation. A–D,F–I: Immunofluorescent detection of active Caspase 3 (A–D) and 5-ethynyl-2′-deoxyuridine (EdU; F–I) incorporation on transversal sections through the inflow tract region of embryos that were implanted with aggregates of (A,C,F,H) control, or (B,D,G,I) FGFR-expressing cells and cultured until Hamburger and Hamilton (HH) stage (A,B,F,G) 16 or (C,D,H,I) 18. Red arrowheads in (A–D) point to apoptotic areas in the right (RS) and left (LS) sinus horns. Asterisks mark the implanted cell aggregates. Sections were counterstained with DAPI (4,6-diamino-2-phenylindole). E,J: Quantification of active Caspase 3-labeled (E), and EdU-labeled (J) cells in the right (blue columns) vs. left sinus horn (white columns) of embryos implanted with control or FGFR-expressing cells, respectively. *P = 0.0001 compared with control right sinus horn; #P < 0.05 compared with control left sinus horn; n.s., nonsignificant difference.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

FGF signaling is involved in many important steps during embryogenesis (Mariani et al.,2008). We studied here the role of FGF signaling during PE formation, a transient embryonic structure at the venous pole that generates many essential cell types of the embryonic heart (Mikawa and Brand,2010). While progress has been made with regard to defining a role of FGF signaling during early cardiac specification (Alsan and Schultheiss,2002), its role during PE formation is not fully understood. We found that several FGF ligands are expressed in the proepicardium, including Fgf2, Fgf10, and Fgf12, corroborating the results of others (Kruithof et al.,2006). It is not known whether Fgf2 and Fgf10 do activate different signal transduction pathways; however, it is well known that FGF ligands differ in their ability to activate different FGF receptor (Itoh and Ornitz,2004). Significantly and corroborating the proposed function of FGF during PE formation, three FGF receptors were found to be expressed in the PE and here mostly in the epithelial compartment. This expression pattern fits well with the proposed function of FGF as a regulator of epithelial outgrowth during PE formation.

BMP signaling is essential for PE marker gene expression but is also required for cardiac myocyte formation (Schultheiss et al.,1997; Andrée et al.,1998; Schlueter et al.,2006). We previously demonstrated that different thresholds of BMP signaling are involved in myocyte formation and PE cell specification (Schlueter et al.,2006). Others have proposed that FGF might act together with BMP in cell fate allocation at the venous pole (Kruithof et al.,2006). In this study, we found no evidence for a role of FGF as a modulator cell fate allocation in case of PE development. Loss of function experiments in explants and in vivo had no discernible effect on PE cell recruitment because proepicardial marker gene expression was not affected by interfering with FGF signaling. This is in contrast to the effects of blocking BMP signaling, which resulted in a complete loss of marker gene expression (Schlueter et al.,2006). Possibly, the level of inhibition that was achieved in this study by SU5402 or soluble FGF receptors was not sufficient to fully block FGF signaling. However, we consider this possibility less likely given the profound effects of blocking FGF signaling on proepicardial cell proliferation and the induction of apoptosis. We, therefore, conclude that FGF is not important for the recruitment of cells to the PE cell lineage, but is an essential factor of proepicardial cell expansion and survival. The results recently published by van Wijk et al. (2009) are based on data derived from proepicardial explants, which were cultured in the presence of serum. However, serum might alter the cellular responses to growth factor signals. In addition, the in vivo experiments, which were presented in that study, used globally applied antagonists in ovo for 48 hours to interfere with FGF signaling. In contrast in this study, FGF signaling was blocked specifically and locally on a short-term basis in direct vicinity to the PE anlage to minimize any unspecific interference with other organ formation processes. Under these experimental conditions, we obtained no evidence for a role of FGF in the recruitment of cells to the PE cell lineage.

Because PE explants showed a strong responsiveness toward exogenous FGFs, this finding suggests that FGF ligands are candidates for a heart-derived signal that is involved in inducing secondary bridge formation essential for epicardialization in the chick and frog embryo (Nahirney et al.,2003). FGF has also numerous other roles during epicardium formation: (1) FGF signaling has been implicated in the control of EMT and tissue invasion (Morabito et al.,2001; Pennisi and Mikawa,2009); (2) both Fgfr1 and −2 are required for coronary vasculogenesis (Lavine et al.,2006); and (3) FGF modulates the amount of endothelial and smooth muscle cells in PE cultures and in vivo (Pennisi and Mikawa,2009). Recent studies in zebrafish also demonstrated an important function of FGF signaling for epicardial EMT and tissue invasion after cardiac wounding and tissue homeostasis (Lepilina et al.,2006; Wills et al.,2008).

In many organ systems, FGFs are downstream targets through which T-box genes exert growth control (Takeuchi et al.,2003; Hu et al.,2004). It is, therefore, possible that such an epistatic relationship also exists during PE development. Three Tbx genes are expressed in the PE and in the epicardium, i.e., Tbx5, Tbx18, and Tbx20 (Hatcher et al.,2004; Schlueter et al.,2006; Shelton and Yutzey,2008). It, therefore, will be interesting to analyze, whether Tbx genes act through FGF genes in the PE. Interfering with Tbx5 for example blocks proepicardial cell migration but has no effect on cell proliferation (Hatcher et al.,2004). At present, no functional data are available for the role of Tbx18 and Tbx20 during PE formation or epicardium formation.

The data presented in this study and previous results can be summarized to formulate a working model of PE development (Fig. 9). PE development is bilateral asymmetric and under control of the left–right pathway. A mature PE develops only on the right body side, whereas a vestigial PE-like structure is formed contralaterally on the left side (Schlueter et al.,2006; Schulte et al.,2007). Bilateral asymmetric PE development is controlled by FGF8 and Snai1 (Schlueter and Brand,2009). The PE develops from mesodermal cells adjacent to the liver bud, which might produce some unknown inducing signaling factor (Männer et al.,2001; Ishii et al.,2007). BMP2 and -4 are expressed in the PE and the adjacent sinus horn myocardium, and BMP signaling is required to maintain PE marker gene expression (Schlueter et al.,2006). FGF is also involved in PE formation; however, FGF controls cell expansion and villous outgrowth but appears to be dispensable for specification.

thumbnail image

Figure 9. Model of proepicardium (PE) formation in the chick embryo. PE formation in the chick embryo is bilateral asymmetric and develops only on the right side. On the left side mesothelial cells form a vestigial PE-like structure, which ultimately gets lost by apoptosis. Right-sided PE formation is under control of the left–right axis and involves FGF8 and Snai1. Bone morphogenetic protein 2 (BMP2), present in both sinus horns, and BMP4, which is expressed in the PE are required for PE determination. In addition, PE induction might involve unknown signals from the liver bud. In this study, we show that cellular outgrowth and survival of the PE depends on fibroblast growth factor (FGF) signals including FGF2 and FGF10 acting in an autocrine or paracrine manner.

Download figure to PowerPoint

FGF signaling, therefore, has two roles in PE formation: an early patterning function at gastrulation to specify the right side of the embryo (Schlueter and Brand,2009), and a later function as a regulator of villous outgrowth (this study). FGF could in principle be a limiting factor for PE maturation on the left body side. However, implantation of FGF2 beads into the left sinus horn was not sufficient to induce villous outgrowth (data not shown). This is in contrast to the experimental outcome of ectopic expression of FGF8 on the left side of Hensen's node at gastrula stage, which induces ectopic PE formation on the left side (Schlueter and Brand,2009).

In this report, we show that several FGF genes and FGF receptors are expressed in the PE. The utilization of pharmacological reagents to interfere with FGF signaling in vivo and in explant cultures established that FGF signaling is important for proepicardial growth and survival, but appears to be dispensable for the induction and maintenance of proepicardial cell identity, because marker gene expression was maintained in the absence of FGF signaling. This is in contrast to BMP signaling, which previously was shown to be required for proepicardial gene expression. Thus, both signals control different aspects of proepicardium formation. It is likely that additional signals are required for cell specification and pattern formation at the venous pole.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Chick Embryos

Fertilized White Leghorn eggs (Lohmann, Cuxhaven, Germany) were incubated at 38°C in a moist atmosphere until the desired stage was reached. Embryos were staged according to Hamburger and Hamilton (Chapman et al.,2001), dissected in phosphate buffered saline (PBS), and fixed in 4% (w/v) paraformaldehyde (PFA) in PBS at 4°C. For sectioning, embryos were dehydrated in a graded alcohol series and embedded in paraffin.

Explant Cultures

For explant cultures, eggs were incubated until the embryos had reached HH stage 17. PE explants were isolated using a micro-dissecting knife and cultured in serum-free M199 (Invitrogen) medium on fibronectin-coated culture slides (BD Biosciences) or uncoated plastic culture dishes (Nunclon) as previously described (Schlueter et al.,2006). After 16 hr of culture in serum free medium, the explants were photographed. Any explant that showed beating areas (myocyte contamination, or was nonadherent at this time point were discarded. Explants, which were unusually big or small, were also excluded. Final concentrations of 10 μM or 100 μM SU5402 (Calbiochem), 30 μg/ml recombinant human FGFR1β(IIIc)/Fc protein (R&D Systems), 5 μM U0126 (Calbiochem), 10 μM LY294002 (Cell Signaling Technology) or 2 μg/ml recombinant mouse basic FGF protein (R&D Systems) were added to the culture medium. Control explants were incubated with the addition of an equal volume of vehicle (1% (v/v) DMSO or 0.2% (w/v) bovine serum albumin in PBS). The PE explants were cultured for a total period of 64 hr.

Embryo Manipulation

Embryos were explanted at HH stage 12 and cultured according to the EC culture method (Chapman et al.,2001). Cells of the chicken embryonic fibroblast cell line DF1 (ATCC) were co-transfected with RCAS-BP(A) constructs encoding FGFR1-FC, FGFR2-FC, FGFR4-FC, or AP (Fekete and Cepko,1993; Marics et al.,2002; Mandler and Neubuser,2004). Cell aggregates were implanted into the right sinus horn and embryos were cultured until HH stage 16 or 18.

In Situ Hybridization

Detection of mRNA in tissue sections was carried out as previously described (Moorman et al.,2001). For expression analysis cRNA probes for AMHC, Hex, Tbx18, Fgf2, Fgf3, Fgf4, Fgf8, Fgf10, Fgf12, Fgf13, Fgf14, Fgf18, Fgf19, Fgfr1, Fgfr2, Fgfr3, and Fgfr4 were used (Bachler and Neubuser,2001; Karabagli et al.,2002; Schlueter et al.,2006; Lunn et al.,2007). Fluorescent in situ hybridization (FISH) analysis of Tbx18 expression was performed as previously described (Schlueter and Brand,2009). For volume rendering and three-dimensional reconstruction of the Tbx18 expression domain, whole mount-stained embryos subjected to FISH were mounted in Moviol on glass slides and the venous pole area was scanned with the confocal microscope. The obtained images of the Z-stack were loaded in AMIRA and Label Fields of individual images were generated, which were subsequently used to generate a 3D reconstruction of the Z-stack allowing volume quantification as previously described (Soufan et al.,2007).

Immunohistochemistry, Proliferation, and Apoptosis Assays

Immunohistochemical detection of Wt1 and Tbx5 in PE explant cultures was performed as previously described (Schlueter et al.,2006; Bimber et al.,2007). Smooth muscle α-actin and myosin heavy chain were detected using an sm-alpha actin antibody (SIGMA A7607) and a monoclonal antibody against sarcomeric myosin heavy chain, MF20 (Developmental Hybridoma Bank), respectively. Apoptosis was detected by TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling) assay (Roche) according to manufacturer's specifications. Active Caspase 3 antibody (BD Pharmingen) was detected with donkey anti-rabbit antibody coupled to Alexa 488 (Molecular Probes). Caspase 3 staining was quantified using three sections of the inflow tract region of three experimental and control embryos. Bromodeoxyuridine (BrdU; Roche) was applied to five drug-conditioned and control explants at a final concentration of 40 μM. BrdU incorporation was immunodetected by BrdU antibody (Roche) and a secondary antibody coupled to Alexa 555 (Molecular Probes). For in vivo labeling, the thymidine analogue 5-ethynyl-2′-deoxyuridine (EdU, Invitrogen) was used. Embryos were labeled for 6 hr and incorporation was detected on sections of three control and experimental embryos following manufacturer's specifications. Sections or explants were counterstained with 4,6-diamino-2-phenylindole (DAPI).

RNA Isolation, RT-PCR, and qRT-PCR

Total RNA of dissected tissues and cultured PE explants was isolated as previously described (Schlange et al.,2000) and cDNA was synthesized from DNAse treated total RNA using AMV reverse transcriptase. The following primer pairs were used for RT-PCR or qPCR: Fgf2fwd: 5′-GCACTTC AAGGACCCCAAGC-3′, Fgf2rev: 5′-AG CAATCTGCCATCCTCCTT-3′ (55°C, 200 bp); Fgf10fwd: 5′-CCCCGGAGGCCAC CAACT-3′, Fgf10rev: 5′-CCCCTTCCATTCAGAGCAACAAAC-3′ (57°C, 436 bp); Fgf12fwd: 5′-AGAAGCGGCAGGCGAGGGAGTC-3′, Fgf12rev: 5′-TGCCGGTACA GCGTGGAAGAATAA-3′ (60°C, 464 bp); Fgfr1fwd: 5′-TGACGTGCAGAGCATC AAC-3′, Fgfr1rev: 5′-GCAGCTTCTTCTCCATCTT-3′ (55°C, 317 bp); Fgfr2fwd: 5′-T GAACTCCAACACGCCTC-3′, Fgfr2rev: 5′-GG GACCCTGTTAATATCA-3′ (55°C, 500 bp); Fgfr4fwd: 5′-AGCCCGTCTACGTG CACA-3′, Fgfr4rev: 5′-GTAGTTGCCG CGGTCGGA-3′ (55°C, 400 bp); Tbx18fwd: 5′-CATATGTGCAGACACT-3′, Tbx18rev: 5′-CATATGTGCAGACACT-3′ (48°C, 227 bp); Wt1fwd: 5′-AGCCAGCA AGCCATTCGCAACC-3′, Wt1rev: 5′-TTCTCATTTTCATATCCTGTCC-3′ (58°C, 355bp); Tbx5fwd: 5′-CCGTCCTACAGCAGTTGCAC-3′, Tbx5rev: 5′-GCCGTG GGAATAGAGGAACT-3′ (55°C, 273 bp); Gapdhfwd: 5′-ACGCCATCACTATCT TCCAG-3′, Gapdhrev: 5′-CAGGC CTTCACTACCCTCTTG-3′ (52°C, 578 bp). A Bio-Rad iCycler and the QuantiTect SYBR Green PCR Kit (Qiagen) were used for qRT-PCR. Gene expression levels were quantified based on the threshold cycle (C(t)), normalized to Gapdh and corrected for efficiency of each primer pair (Pfaffl,2001).

Data Documentation and Statistical Analyses

Images were captured on a Leica MZ FLIII stereo microscope or a confocal laser scanning microscope (TCS SP2, Leica). Standard deviation from the mean was calculated for groups of quantitative data. Statistical significance was determined using an unpaired Student's t-test.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The donation of reagents by Constance Cepko, Christophe Marcelle, Annette Neubüser, Gary C. Schoenwolf, Hans-Georg Simon, and Kate Storey is gratefully acknowledged. We thank Anneliese Striewe-Conz and Ursula Herbort-Brand for excellent technical assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES