OUP user menu

Congenital heart defects in Fgfr2-IIIb and Fgf10 mutant mice

Anita Marguerie, Fanny Bajolle, Stephane Zaffran, Nigel A. Brown, Clive Dickson, Margaret E. Buckingham, Robert G. Kelly
DOI: http://dx.doi.org/10.1016/j.cardiores.2006.03.021 50-60 First published online: 1 July 2006

Abstract

Objective Myocardial progenitor cells expressing Fgf10 give rise to the outflow tract and right ventricle of the mammalian heart. In order to define the role of fibroblast growth factor (FGF) signaling in this process we investigated whether Fgf10 or the major Fgf10 receptor Fgfr2-IIIb are required for normal heart development.

Methods The cardiac phenotype of Fgf10 and Fgfr2-IIIb mutant mice was analysed by histology, scanning electron microscopy and gene and transgene expression studies.

Results Outflow tract formation from Fgf10 expressing progenitor cells occurs normally in Fgf10 mutant embryos and in the majority of Fgfr2-IIIb mutant embryos; a proportion of Fgfr2-IIIb mutant embryos, however, display outflow tract and right ventricular hypoplasia. The predominant cardiac defects in Fgfr2-IIIb mutant embryos are ventricular septal defects associated with overriding aorta or double outlet right ventricle. In addition, loss of Fgfr2-IIIb is associated with ventricular anomalies including a thin myocardial wall, abnormal trabeculation and muscular ventricular septal defects. In contrast, Fgf10 is required to correctly position the heart in the thoracic cavity but not for outflow tract septation. Both Fgf10 and Fgfr2-IIIb mutant embryos lack pulmonary arteries and veins.

Conclusions Fgfr2-IIIb and Fgf10 mutant mice have distinct roles during cardiac morphogenesis, although neither gene is essential for outflow tract elongation from Fgf10 expressing progenitor cells. Fgfr2-IIIb and Fgf10 mutant mice provide new models for common components of congenital heart disease.

Keywords
  • Developmental biology
  • Fibroblast growth factors
  • Congenital defects
  • Transgenic animal models

This article is referred to in the Editorial by D. Franco (pages 1–3) in this issue.

1 Introduction

The embryonic mammalian heart forms as a tubular structure from splanchnic mesoderm. Subsequently, rightward looping and heart tube elongation take place, during which myocardial differentiation continues at the venous pole to give rise to atrial and inflow structures and at the arterial pole to give rise to the outflow tract [1]. Growth of the heart during foetal development is accompanied by remodeling of the atrioventricular and ventriculoarterial connections, and by atrial, ventricular and outflow tract septation, generating parallel systemic and pulmonary circulatory systems. The arterial pole of the heart gives rise to the aorta and pulmonary trunk connected to the left and right ventricles, respectively. This complex morphogenetic process requires the interaction of multiple cell types, including endocardial, myocardial and cardiac neural crest cells, and the function of numerous regulatory and structural genes. Perturbations of this process result in congenital heart defects, which affect 8 in 1000 live births [2].

Fibroblast growth factor (FGF) signaling regulates multiple facets of embryonic morphogenesis, including cell proliferation, differentiation and migration. FGFs are a large family of intercellular signaling molecules that mediate their biological responses by binding and activating high-affinity cell surface receptors (FGFRs) [reviewed in 3]). In mammals, four genes encode FGFRs with intrinsic tyrosine kinase activity (Fgfr1 to Fgfr4), with additional complexity achieved by alternative splicing [4]. Fgfr2 encodes two isoforms, Fgfr2-IIIb and Fgfr2-IIIc, that have different FGF-binding specificities and are expressed in different cell lineages. Fgfr2-IIIb is located on the epithelia of ectodermal and endodermal organs and is activated by four known ligands (Fgf1, Fgf3, Fgf7 and Fgf10); the latter two FGFs are expressed predominantly in mesenchyme adjacent to epithelia expressing Fgfr2-IIIb [5–9]. FGF signaling has been implicated in the early steps of heart development in flies, zebrafish, chick and mice [10–16]. However, the role of different FGF ligands and receptors involved in different stages of mammalian heart development is currently unclear. Fgf10 is expressed in pharyngeal mesoderm which contributes to the arterial pole of the heart, as first shown in the mouse embryo by an Fgf10-enhancer trap transgene [17]. This myocardial progenitor cell population is referred to as the anterior or secondary heart field (AHF) and is part of a larger population of Isl1-expressing myocardial progenitor cells [18–20]. The AHF gives rise to right ventricular and outflow tract myocardium [21]. Cardiac neural crest cells play an important role in outflow tract septation [22] and are required for normal addition of myocardial progenitor cells to the elongating heart tube [23]. Despite the early expression pattern of Fgf10 in the AHF, mice lacking Fgf10 or the major Fgf10 receptor Fgfr2-IIIb survive until birth with no reported heart defects, although both strains of mutant mice lack limbs and lungs and display agenesis or hypoplasia of multiple organs [24–26].

Here we demonstrate that Fgfr2-IIIb and Fgf10 are required for normal heart development. However, despite the expression of Fgf10 in the AHF, neither Fgf10 nor Fgfr2-IIIb is essential for initial formation of the outflow tract. Loss of Fgfr2-IIIb, but not Fgf10, results in severe defects in outflow tract septation and ventricular morphogenesis, and loss of Fgf10 is associated with abnormal positioning of the heart in the thoracic cavity. Furthermore, neither pulmonary veins nor arteries are present in either mutant. Our results provide novel insights into the role of FGF signaling during cardiac development and suggest that defects in FGF signaling may contribute to the etiology of human congenital heart defects.

2 Methods

2.1 Mouse strains and tissue collection

Fgfr2-IIIb null mice have been described previously [26]. Isoform specific Fgfr2-IIIb null mice were generated by placing translational stop codons in all three reading frames within exon IIIb, followed by an IRES-LacZ cassette. Fgf10− / − mice [25] and Fgf10-nlacZ enhancer trap mice (Mlc1v-nlacZ-24, [17]) have been described Fgf10+/ − mice were maintained on a C57BL/6J background. Embryos were genotyped by PCR using genomic DNA isolated from yolk sacs as described by Revest et al. [26] and Sekine et al. [25]. Hearts or trunks were dissected and fixed in 4% paraformaldehyde (PFA) in PBS overnight at 4 °C, dehydrated and embedded in paraffin for sectioning at 5 μm; for early developmental stages, embryos were fixed and embedded in wax for coronal or sagittal sections. Alternatively, embryos and hearts were fixed, transferred to 15% sucrose in PBS, followed by 15% sucrose, 7% gelatin, and frozen in liquid nitrogen prior to cryosectioning at 10 μm. Sections were counterstained in 1% aqueous eosin. Quantitative histomorphometry: the mean thickness of the compact region of the ventricular myocardium was determined by measuring the length and area in multiple sections through the four chambers of the heart using NIH Image 1.61 on 6 different samples. Mean values are presented ±S.E.M. Statistical significance was evaluated using Student's t-test. The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).

2.2 Histology, in situ hybridisation, immunohistochemistry

Histology and in situ hybridisation were carried out using standard procedures as described [17,26]. Probes detecting Fgf8, PlexinA2, Sox10, Crabp1 and Hand2 transcripts were synthesised as described [17,27–30].

Whole-mount immunohistochemistry was performed by bleaching fixed embryos in 6% H2O2 in PBT for 30 min. Embryos were rinsed and washed three times in PBT for 10 min and incubated with αsm-actin antibody overnight at 4 °C. After four washes in PBT for 15 min, detection was revealed using the Vector SG peroxidase substrate kit (Vector Labs, USA).

Expression of β-galactosidase in the Fgf10-nlacZ enhancer trap transgene was analysed in embryos fixed in 4% PFA for 40 min, using standard X-Gal revelation [17], for 2–3 h at 37 °C. Samples were then rinsed in PBS.

2.3 Scanning electron microscopy

Embryos were fixed in 2.5% glutaraldehyde in phosphate buffer (pH 7.4) for 24 h. Microdissection was performed under a stereomicroscope using iridectomy scissors. Samples were postfixed in 1% osmium tetroxide, dehydrated through a graded alcohol series, critical point-dried using liquid CO2, mounted on stubs, and sputter-coated with gold. Samples were viewed on a Zeiss SM940 scanning electron microscope.

3 Results

3.1 Fgfr2-IIIb is required for outflow tract septation and ventricular development

The Fgfr2-IIIb mutant mice analysed here (Fgfr2-IIIb− / −) produce a truncated receptor that lacks part of the ligand binding domain and the entire transmembrane and tyrosine kinase domains [26]. Expression of the alternatively spliced isoform, Fgfr2-IIIc, is not noticeably affected. Fgfr2-IIIb− / − mice lack lungs and limbs and die at birth [26]. We analysed cardiac development in Fgfr2-IIIb− / − embryos and found severe defects in heart morphogenesis.

The most frequent malformations in Fgfr2-IIIb-deficient hearts affect outflow tract and ventricular septation (Fig. 1, Table 1). More than 90% of mutant embryos analysed between E14.5 and E18.5 display a perimembranous ventricular septal defect (13 out of 14; Fig. 1B, C). This defect is associated with an overriding aorta in 70% of mutant embryos (10 out of 14; Fig. 1B, C). Other mutant embryos displayed a ventricular septal defect with double outlet right ventricle, rather than over-riding aorta (Fig. 1D, E). In addition to defects in the perimembranous portion of the interventricular septum, a defect in the muscular component of the septum was observed in 40% of Fgfr2-IIIb− / − embryos (Fig. 1F, H). Additional defects in ventricular morphology were observed, including right ventricular hypoplasia (4 out of 14 mutant hearts; Fig. 1H) and thin ventricular walls (5 out of 14 Fgfr2-IIIb− / − hearts; Fig. 1D, J–L). Statistical evaluation revealed a significant reduction in the thickness of the left ventricular compact myocardial layer between wild-type hearts (63±2 μm) and Fgfr2-IIIb− / − hearts (33±11 μm) at E17.5 (p<0.05).

Fig. 1

Multiple requirements for Fgfr2-IIIb in ventricular development. Histological sections of hearts of littermate controls (A, G, J) or Fgfr2-IIIb− / − embryos (B–F, H, I, K) at E14.5 (A, B, D, E, G, H, J, K) and E17.5 (C, F, I). A perimembranous ventricular septal defect was detected in 90% of the Fgfr2-IIIb− / − embryos (arrowheads in B and C). The ventricular septal defect is associated with an overriding aorta (70% of Fgfr2-IIIb− / − embryos; B, C) or double outlet right ventricle (D, E). Ventricular septal defects were also observed in the muscular component of the septum (arrowheads in F and H). Other ventricular defects observed in Fgfr2-IIIb− / − embryos include a hypoplastic right ventricle (H) and thin ventricular wall (arrowhead in D; J, K). Measurement of left ventricular wall thickness at E17.5 (L) revealed a significant difference between mutant embryos and wild-type littermates. Ao, aorta; PT, pulmonary trunk; RV, right ventricle; LV, left ventricle. Scale bars: 200 μm (A, B, D, E, G, H); 100 μm (C, F, I); 50 μm (J, K).

View this table:
Table 1

Cardiac defects in Fgfr2-IIIb and Fgf10 mutant and control littermate hearts analysed by histological sections between E14.5 and E18.5

Fgfr2-IIIbFgf10
+/ − or +/+− / −+/ − or +/+− / −
Number of hearts analysed8141113
Perimembranous ventricular septal defect01300
Overriding aorta01000
Double outlet right ventricle0300
Muscular ventricular septal defect0600
Thin ventricular wall0500
Hypoplastic right ventricle0400
Abnormal direction of ventricular apex00010
Absent pulmonary arteries014013
Absent pulmonary veins014013

Defective alignment of the aorta and left ventricle is likely to arise as a result of earlier defects in outflow tract development and septation. Scanning electron microscopy (SEM) revealed that endocardial cushions in the distal outflow tract were fused by E11.5 in control embryos (Fig. 2A), whereas fusion had not occurred at this stage in Fgfr2-IIIb− / − littermates; moreover, both sinistroventral and dextrodorsal outflow tract cushions were hypoplastic (Fig. 2B). It is important to note that this may reflect a general developmental delay observed in Fgfr2-IIIb− / − embryos [31]. At E12.5, the pulmonary trunk and ascending aorta of control embryos were well separated and displayed a marked spiral (Fig. 2C), whereas analysis of Fgfr2-IIIb− / − embryos showed that the outflow tract cushions had only initiated fusion and showed reduced spiraling (Fig. 2D). The alignment defects observed at later stages are therefore likely to result from delayed outflow tract septation associated with reduced spiraling of the proximal great arteries. In addition to outflow tract defects, SEM revealed the absence of a pronounced interventricular groove in Fgfr2-IIIb− / − embryos and defective trabeculation (Fig. 2E, F). In addition to these ventricular defects, left atrial hypoplasia was noted at E10.5 in Fgfr2-IIIb− / − hearts (Fig. 2H).

Fig. 2

Abnormal outflow tract and ventricular septation in Fgfr2-IIIb− / − hearts. Scanning electron microscopy (SEM) at E11.5 (A–B) showing fusion of distal outflow tract cushions in control littermates (A) and hypoplastic sinistroventral (SV) and dextrodorsal (DD) outflow tract cushions in Fgfr2-IIIb− / − embryos (B). At E12.5 (C, D) the pulmonary trunk (PT) and ascending aorta (Ao) of control embryos were well separated with a marked spiral (C), while outflow tract cushions had just initiated fusion in Fgfr2-IIIb− / − embryos (D). SEM of bisected ventricles at E11.5 revealed that the interventricular groove is reduced in Fgfr2-IIIb− / − embryos (arrowhead in F). The main trabeculae in the right ventricle of control hearts (E) are radially orientated (broken lines) and thicker than trabeculae in the left ventricle; in contrast, trabeculae are randomly orientated in both right and left ventricles of Fgfr2-IIIb− / − embryos (F). SEM at E10.5 revealed a smaller left atrium in Fgfr2-IIIb− / − embryos compared to control littermates (G, H). LA, left atrium; RA, right atrium, LV, left ventricle; RV, right ventricle.

In order to investigate whether neural crest cell migration and patterning were affected in Fgfr2-IIIb− / − embryos, the expression of two neural crest genes, Sox10 and Crabp1, was analysed at E10.5. There were no detectable differences in Sox10 or Crabp1 expression patterns between control and Fgfr2-IIIb− / − embryos, suggesting that neural crest cell migration occurs normally in the absence of Fgfr2-IIIb (Fig. 3A–D). A third marker of neural crest cells, PlexinA2, expressed by migrating and postmigratory cardiac neural crest cells [27], also showed an indistinguishable pattern of expression in mutant and control embryos (data not shown). The transcription factor encoding gene Hand2 was also found to be expressed equivalently in post-migratory neural crest cells of the pharyngeal arches and outflow tract myocardium in mutant and control embryos (Fig. 3E, F).

Fig. 3

Contribution of neural crest cells and the anterior heart field to Fgfr2-IIIb− / − hearts. Whole mount in situ hybridization at E10.5 showing no detectable difference in cardiac neural crest streams (arrows) labeled by Sox10 (A, B), and Crabp1 (C, D) between control littermates (A, C) and Fgfr2-IIIb− / − embryos (B, D). Hand2 is expressed normally in control (E) and Fgfr2-IIIb− / − embryos (F). X-Gal stained embryos reveal an indistinguishable expression profile of the Fgf10 driven transgene Mlc1v-nlacZ-24 at E9.5 in control (G) and Fgfr2-IIIb− / − (H) embryos, including cells of the anterior heart field (asterisk). Histological sections showing similar morphology, outflow tract diameter and Mlc1v-nlacZ-24 transgene expression in control and mutant hearts at E9.5 (I, J) and E11.5 (K, L). Whole mount in situ hybridization at E9.5 showing reduction of Fgf8 expression specifically in ectoderm in the region of the second pharyngeal cleft of Fgfr2-IIIb− / − embryos (arrow; M, N). Anti-αsm-actin immunohistochemistry at E10.5 in control (O) and mutant (P) embryos, showing elevated labeling in the distal outflow tract of a mutant heart with right ventricular and outflow tract hypoplasia. Scale bars: 100 μm (I–L).

Since the myocardial wall of the outflow tract and right ventricle arises from the AHF [17,21], we analysed the expression profile of the Fgf10-nlacZ enhancer trap transgene, Mlc1v-nlacZ-24, in Fgfr2-IIIb− / − embryos. Transgene expression in the pharyngeal region, outflow tract and right ventricle was not detectably different in Fgfr2-IIIb− / − embryos compared to control embryos at E9.5 and E10.5 (Fig. 3G–L and data not shown). Normal outflow tract morphogenesis was apparent after histological analysis of transverse sections of mutant hearts (Fig. 3J and L). We also investigated the expression of Fgf8, required for normal arterial pole development [15,16], in Fgfr2-IIIb− / − embryos. At E9.5 Fgf8 was expressed equivalently to control embryos, apart from a reduction in ectoderm in the region of the second pharyngeal cleft in Fgfr2-IIIb− / − relative to control embryos (arrows in Fig. 3M and N). Although signaling through Fgfr2-IIIb is not essential for extension of the outflow tract from cells of the AHF, approximately 30% of Fgfr2-IIIb− / − embryos analysed between E14.5 and E18.5 display right ventricular hypoplasia (Table 1). Similarly, a fraction of embryos analysed at earlier developmental stages were found to have a hypoplastic right ventricle and outflow tract (Fig. 3P). Elevated αsm-actin levels were observed in the Fgfr2-IIIb− / − hypoplastic distal outflow tract compared to the non-mutant situation (Fig. 3O and P). While not essential for outflow tract elongation, loss of Fgfr2-IIIb can therefore impact on anterior heart field deployment in a subset of mutant embryos.

The cardiac phenotype in the mice analysed here was observed in lines generated from three independent targeted ES cell clones. However, Fgfr2-IIIb null mice derived by the alternative strategy of deleting exon IIIb [32] do not show a significant heart phenotype (AM and CD, unpublished data). RT-PCR analysis in Fgfr2-IIIb exon IIIb deleted embryos revealed low-level splicing between exons 7 and 9, yielding a viable Fgfr2-IIIc receptor in cells that would normally express Fgfr2-IIIb (AM and CD, unpublished data). In the heart, this would be expected to rescue the phenotype through signaling via alternative FGF ligands, such as Fgf8.

3.2 Fgf10 is required for normal positioning of the ventricular apex but not for outflow tract septation

Embryos lacking the Fgfr2-IIIb ligand Fgf10 die at birth with severe morphological abnormalities, including failure of lung and limb development and aplasia and hypoplasia of multiple organs, a phenotype similar to that of Fgfr2-IIIb− / − embryos [24–26,33]. We investigated the morphology of Fgf10− / − hearts by whole-mount and histological analysis (Fig. 4; Table 1).

Fig. 4

Fgf10 is required for normal positioning of the ventricular apex but not for outflow tract or ventricular septation. Ventral views of E17.5 hearts showing a leftward positioned apex in a control embryo (A) and leftward (B), rightward (C,D), or medially (E) positioned apexes in Fgf10− / − embryos. The apex of the embryo shown in (F) is oriented cranially and to the right (arrowhead), revealing the normally dorsal surface of the heart in this ventral view (asterisk). Note the abnormally angled pulmonary trunk (arrow in D) and position of the atria; the right atrium in (D) is positioned dorsally and obscured, as is the left atrium in (E). Cryostat sections showing normal ventricular septum formation and atrioventricular connections for leftward (G) and rightward (H) positioned Fgf10− / − hearts. Fgf10− / − mutant hearts also have normal venticuloarterial connections: the pulmonary trunk arises from the right ventricle and the aorta arises from the left ventricle in both leftward (I, K) and rightward (J, L) positioned Fgf10− / − hearts. RA, right atrium; LA, left atrium; RV, right ventricle; LV, left ventricle; Ao, aorta; PT, pulmonary trunk. Scale bars: 200 μm (G–L).

Cardiogenesis in Fgf10− / − embryos initiates normally with no apparent impairment in the contribution of Fgf10 expressing progenitor cells. In contrast to our observations with Fgfr2-IIIb− / − embryos, the major defect observed in Fgf10− / − hearts concerns the orientation of the heart in the thoracic cavity. Whereas the cardiac apex in wild-type and Fgf10+/ − mice points to the left (Fig. 4A), in Fgf10− / − hearts, the direction of the cardiac apex was frequently abnormal, pointing either to the left (14%; Fig. 4B), to the right (62%; Fig. 4C, D) or positioned medially (20%; Fig. 4E). In 4% of mutant hearts, an extreme anterior rightward positioned ventricular apex was observed, such that the normally dorsal ventricular surface was apparent in a ventral view (asterisk in Fig. 4F). Rightward positioning of the ventricular apex in Fgf10− / − mice is not associated with situs inversus, since all hearts analysed showed normal rightward looping with the pulmonary ventricle positioned to the right. Furthermore, no extra-cardiac laterality defects were observed; for example, the stomach was positioned to the left in all Fgf10− / − embryos analysed (n=5, data not shown). A rightward-pointing cardiac apex with situs solitus is characteristic of isolated dextrocardia in man [34]. Left and right atrial appendages are normally positioned laterally (Fig. 4A); abnormally positioned atrial appendages were associated with ventricular apex malpositioning in Fgf10− / − hearts. In mutant hearts with a rightward positioned ventricular apex, the right atrial appendage was positioned dorsally such that it was partially or totally obscured in a ventral view (Fig. 4C, D). A dorsally displaced left atrial appendage was observed in a subset of mutant hearts (Fig. 4E). Abnormal positioning of the ventricular apex and atrial appendages in Fgf10− / − embryos was not evident prior to E14.5 (data not shown). In contrast to this phenotype, neither the direction of the ventricular apex nor the position of the atrial appendages was affected in Fgfr2-IIIb/ − embryos.

Histological analysis revealed that despite abnormal positioning of the ventricular apex and atrial appendages, atrioventricular and ventriculoarterial connections were normal in Fgf10− / − embryos (Fig. 4G–L). Normal connections were observed between the right ventricle and pulmonary trunk which arose ventrally to the aorta (9/9 mutant hearts sectioned; Fig. 4I, J). Similarly, normal connections were observed between the left ventricle and the ascending aorta (Fig. 4K, L). In contrast to Fgfr2-IIIb− / − hearts, no muscular or perimembranous ventricular septal defects were observed in Fgf10− / − embryos (Fig. 4G, H). Neither right ventricular hypoplasia nor myocardial wall thinning were observed. However, the orientation of the pulmonary trunk was frequently abnormal in Fgf10− / − hearts, running transversely (arrow in Fig. 4D), orthogonal to the direction in control hearts (Fig. 4A). Analysis of earlier developmental stages revealed that ventricular and arterial pole septation occurred at the same stages in Fgf10− / − and control embryos (data not shown).

Fgf10 and Fgfr2-IIIb are both essential for lung morphogenesis [24–26]. We investigated whether pulmonary arteries and veins are present in Fgf10− / − or Fgfr2-IIIb− / − embryos. At E17.5, pulmonary arteries are normally connected with the dorsal wall of the pulmonary trunk and pulmonary veins with the dorsal wall of the left atrium (Fig. 5A, D). No pulmonary arteries were observed in Fgf10− / − (Fig. 5B, C) or Fgfr2-IIIb− / − mutant embryos. Similarly, pulmonary veins were not detected in Fgf10− / − (Fig. 5E, F) or Fgfr2-IIIb− / − hearts. Analysis of earlier developmental stages by histology and SEM revealed no evidence for the development of pulmonary circulation in Fgf10− / − or Fgfr2-IIIb− / − embryos (Fig. 5G–K). This result reveals that systemic and pulmonary chamber formation can proceed in utero in the complete absence of pulmonary vasculature.

Fig. 5

Pulmonary arteries and veins are absent from Fgf10− / − and Fgfr2-IIIb− / − hearts. Cryostat sections at E17.5 showing normal formation of the right and left pulmonary arteries (A) and pulmonary veins (D) in a control littermate and the absence of pulmonary arteries arising from the pulmonary trunk (B, C) and pulmonary veins associated with the left atrium (E, F) for leftward (B, E) and rightward (C, F) positioned Fgf10− / − hearts. At E14.5 pulmonary veins are absent from Fgf10− / − and Fgfr2-IIIb− / − hearts (G–I). SEM at E11.5 reveals pulmonary veins in littermate controls (box in J) which are absent from Fgfr2-IIIb− / − hearts (box in K). Ao, aorta; PT, pulmonary trunk; RPA, right pulmonary artery; LPA, left pulmonary artery; PV, pulmonary veins; Dao, descending aorta; LSCV, left superior caval vein; RB, right bronchus; LB, left bronchus. Scale bars: 200 μm (A–F); 100 μm (G–I).

4 Discussion

Analysis of Fgf10− / − and Fgfr2-IIIb− / − embryos has revealed a critical role for FGF signaling during cardiac morphogenesis. In the absence of Fgf10, we observed malpositioning of the ventricular apex in the thorax and lack of pulmonary circulation, while formation and septation of the outflow tract, a region that forms from Fgf10 expressing progenitor cells in pharyngeal mesoderm, is normal. In contrast, in Fgfr2-IIIb− / − mice, we observed ventricular septal defects and outflow tract anomalies including overriding aorta and double outlet right ventricle.

Fgf10 is expressed in the anterior heart field, a population of cells in pharyngeal mesoderm which give rise to myocardium of the right ventricle and outflow tract [17,21]. FGF and bone morphogenetic protein (BMP) signaling have been implicated in deployment of the AHF [15,16,35,36]. However, despite expression in the AHF, Fgf10 is dispensable for arterial pole extension and septation. Consistent with these observations, the major receptor for Fgf10, Fgfr2-IIIb, is not essential for outflow tract elongation, yet plays a critical role in subsequent outflow tract septation. Addition of cells from the AHF can therefore occur in the absence of both Fgf10 and Fgfr2-IIIb. However, a fraction of Fgfr2-IIIb mutant embryos display a hypoplastic right ventricle suggesting that AHF deployment is not completely independent of Fgfr2-IIIb. Future experiments will explore whether this partial dependence on Fgfr2-IIIb affects progenitor cell pool size or deployment and differentiation of the AHF. Elevated smooth muscle actin expression suggests that outflow tract development may be delayed in a fraction of Fgfr2-IIIb− / − hearts. Indeed, subtle anomalies in AHF development in the absence of Fgfr2-IIIb may contribute to the outflow tract septation defects observed at later stages. Consistent with such a hypothesis, ablation of the more restricted avian secondary heart field has recently been shown to result in congenital heart defects with overriding aorta [37].

Embryos carrying one null and one hypomorphic Fgf8 allele exhibit outflow tract defects consistent with a reduced contribution of pharyngeal mesodermal cells to the growing arterial pole [15,16]. Similarly, outflow tract defects in Tbx1− / − embryos are associated with a loss of Fgf8 expression [38]. Fgf8 is expressed in pharyngeal endoderm, ectoderm and mesoderm during arterial pole extension, overlapping in expression with Fgf10 in the AHF [17,39,40]. Fgf8 signals through IIIc FGF receptor isoforms and it is these ligand–receptor interactions which are likely to be critical for arterial pole extension in the mouse. The reduction of a specific domain of Fgf8 expression in pharyngeal ectoderm in Fgfr2-IIIb mutant embryos may affect this process either through direct signaling to cells of the AHF or indirectly by regulating neural crest cell properties in the second pharyngeal arch. Fgf8 expression is unaltered in the pharyngeal region in the absence of Fgf10, in contrast to the situation in the apical ectodermal ridge of the developing limb-bud [25,41]. Different regulatory hierarchies, therefore, control FGF signaling during limb and heart development. Analysis of Fgf10 function in the presence of hypomorphic or conditional Fgf8 alleles will reveal whether Fgf8 masks an overlapping role of Fgf10 during arterial pole development.

Ablation of the cardiac neural crest in the chick results in ventricular septal defects and outflow tract alignment anomalies similar to those observed in Fgfr2-IIIb− / − embryos [22]. It has recently been shown that interactions between the AHF and closely opposed cardiac neural crest are required for normal outflow tract elongation and septation [23]. Fgfr2 is expressed in neural crest cells, suggesting that it could mediate neural crest–AHF interactions, as has been shown for BMP signaling [36]. Our results indicate that neural crest cells migrate through the pharyngeal arches and reach their appropriate targets in Fgfr2-IIIb− / − mice; post-migratory neural crest cell function, however, might be impaired. Since ventriculoarterial connections are normal in Fgf10− / − embryos, it is likely that some other member(s) of the FGF family signals through Fgfr2-IIIb to regulate outflow tract septation. Fgf10 may either play no role during normal outflow tract septation or overlap in function with other ligands, as has been shown during otic placode induction for Fgf10 and Fgf3 [42,43] and in the regulation of late proliferation of the thymic epithelium for Fgf10 and Fgf7 [44]. Fgf7 is expressed in the developing heart from E8.5, and Fgf3 is expressed in caudal pharyngeal endoderm at E9.5; however, no cardiac defects have been observed in Fgf7− / − or Fgf3− / − mice [7,45,46]. Mice lacking Fgf15, which has been proposed to interact exclusively with Fgfr4 [47], have recently been shown to display outflow tract defects strikingly similar to those reported here in Fgfr2-IIIb− / − embryos [48]. Our results raise the possibility that Fgf15 functions in vivo through Fgfr2-IIIb. Future tissue-specific loss of function experiments will test the hypothesis that Fgfr2-IIIb in cardiac neural crest cells receives redundant ligand signals from surrounding pharyngeal mesoderm (Fgf10, Fgf7) and pharyngeal endoderm (Fgf3, Fgf15), and that transduction of such signals is essential for normal outflow tract septation.

Our study has uncovered additional functions for Fgf10 and Fgf2R-IIIb during heart development. Fgf10 is required for normal leftward positioning of the ventricular apex. Since abnormal cardiac positioning was not observed prior to E14.5, we conclude that late events during cardiac and thoracic development underlie this anomaly. Similar observations have been documented in an independent Fgf10 mutant (Erik Meyers, Duke University Medical Center, personal communication). Dorsal atrial positioning is associated with an aberrantly positioned ventricular apex. The mechanisms underlying these apparently linked malformations downstream of Fgf10 function are unknown. The lack of lungs in Fgf10− / − embryos may remove spatial constraints on the developing heart such that it is no longer consistently oriented with the apex towards the left. However, since Fgfr2-IIIb− / − embryos also lack lungs yet display correct positioning of the ventricular apex we can conclude that neither lack of lungs nor absence of pulmonary vasculature underlie this phenotype. Alternatively, mediastinal anomalies could confer aberrant constraints on the developing heart in the absence of Fgf10. The pulmonary trunk is also abnormally oriented in Fgf10− / − embryos; future work will analyse the fine structure of the mediastinum and great arteries in Fgf10− / − hearts to investigate the etiology of this anomaly. In addition, we have observed thinning of the ventricular wall, trabecular defects and muscular ventricular septal defects in Fgfr2-IIIb− / − hearts. This form of the receptor is therefore involved in multiple aspects of heart development in addition to outflow tract septation, including cardiomyocyte proliferation, maturation, and muscular septum morphogenesis. Consistent with these conclusions is the observation that FGF signaling plays a role in proliferation within the early myocardium [49], and the demonstration that epicardial and endocardial sources of FGF regulate receptor activity during subsequent ventricular growth [50].

Our study of Fgf10 and Fgfr2-IIIb mutant hearts provides insight into the multiple complex roles played by FGF signaling during cardiac morphogenesis. Although not essential for early heart tube elongation, conforming to an emerging model of Fgf8-driven anterior heart field deployment, this ligand and receptor play roles in outflow tract septation, positioning of the heart within the thorax, development of the pulmonary vasculature and ventricular myocardial proliferation and maturation. These roles, with the exception of the development of the pulmonary vasculature, involve Fgf10 acting through other receptors and additional ligands activating Fgfr2-IIIb. The defects we have observed in these mice correspond to common features of congenital heart disease in man, including components of Tetralogy of Fallot (ventricular septal defect with overriding aorta), isolated dextrocardia (rightward positioned ventricular apex with situs solitus) and muscular interventricular septal defects. Fgf10 and Fgfr2-IIIb mutant mice therefore provide new models with which to probe the etiology and implication of FGF signaling in the development of congenital heart defects. Furthermore, FGF ligand and receptor genes should be considered as important candidates in searches for genes underlying sporadic and familial congenital heart defects.

Acknowledgements

We thank Profs. Nobu Itoh and Shigeaki Kato for providing Fgf10 mutant mice. We thank George Elia (Cancer Research UK) and Catherine Bodin (Pasteur Institute) for histology, Janet MacDonald and Gillian Hutchinson (Cancer Research UK) for animal husbandry. We also thank Pooja Seedhar for technical assistance and Dr Diego Franco for helpful discussions. Work on cardiogenesis in MB's laboratory is supported by the Pasteur Institute, the CNRS and an ACI grant from the French Ministry of Science. The laboratory of NB is supported by the British Heart Foundation Program Grant RG/03012. RK is an INSERM research fellow and is supported by the INSERM Avenir programme, the Fondation de France and the Fondation pour la Recherche Medicale.

Footnotes

  • 1 These authors contributed equally to this work.

  • 2 Current address: DanioLabs Ltd., Unit 7330, Cambridge Research Park, Waterbeach, CB5 9TN, UK.

  • Time for primary review 36 days

References

  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
View Abstract