• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Pownall ME, Isaacs HV. FGF Signalling in Vertebrate Development. San Rafael (CA): Morgan & Claypool Life Sciences; 2010.

Cover of FGF Signalling in Vertebrate Development

FGF Signalling in Vertebrate Development.

Show details



In animal embryos, cells not only receive and interpret signals to commit to particular a lineage, but in addition, cells can also change their cytoskeletal structure and behaviour so that, in response to signals, cells move. The process by which cells move during development to change tissue shape and the relative positions of different cell types is known as morphogenesis. It can be a difficult challenge to dissect distinct effects of cell signals in determining cell fate versus their effects on morphogenesis. In Xenopus, early canonical Wnt signalling (that acts through the stabilisation of β-catenin) on the dorsal side of the embryo specifies the cells of the organiser (Tao et al., 2005). Subsequently, these cells drive the initiation of gastrulation movements, including a process called convergent extension (CE) (Figure 10 ). During CE, the coordinated polarisation and intercalation of dorsal mesoderm cells drives the lengthening of the anteroposterior axis. Interestingly, this cell behaviour is known to require gene products in the planar cell polarity (PCP) pathway, one of the noncanonical Wnt signalling pathways (Tada et al., 2002). Thus, both canonical and noncanonical Wnt signalling pathways are central in the specification and behaviour of organiser cells.

FIGURE 10. Convergent extension (CE).


Convergent extension (CE). Activation of the planar cell polarity pathway results in the membrane localisation of dishevelled (Dsh) and PKCδ and the reorganisation of the cytoskeleton such that cells become polarised along on axis. Polarised cells (more...)

In amniotes, a different cell behaviour underlies gastrulation: as cells enter the primitive streak, they undergo a epithelial to mesenchymal transformation (EMT); this is where a cell that is part of an epithelium loses contact with its neighbours, changes shape, and moves free. Snail codes for a transcriptional repressor that inhibits the expression of E-cadherin, a gene that codes for a cell adhesion protein. Activation of Snail down-regulates E-cadherin, which loosens cell–cell association in an epithelium so that a cell can migrate away. Importantly, this kind of cell behaviour is also characteristic of metastatic cancer cells and E-cadherin is known to act as a tumour suppressor (Jeanes et al., 2008). Analysis of chimeric mice mutant for FGFR1 has shown that cells require FGF signalling to down-regulate E-cadherin and undergo EMT; cells that fail to do this are trapped in the primitive streak (Ciruna and Rossant, 2001; Ciruna et al., 1997).

There are severe gastrulation defects in amniote and fish and frog embryos deficient in FGF signalling, including FGFR1 mutant mice (Ciruna and Rossant, 2001) and zebrafish and Xenopus embryos injected with mRNA coding for a dominant negative form of FGFR 1 (Amaya et al., 1991; Griffin et al., 1995; Isaacs et al., 1994). These data point to an important role for FGF signalling in regulating the diverse strategies for gastrulation in different vertebrates. In chicks, an in vivo study showed that cells migrating through the primitive streak during gastrulation will take different paths depending on their proximity to sources of FGF ligand. FGF4 beads implanted into gastrula-stage chick embryos were found to act as a chemoattractant and GFP-expressing mesodermal cells migrated towards the FGF4 beads. In contrast, FGF8 beads repelled labelled mesodermal cells. The authors’ interpretation of these findings are that cells close to the chick organiser (Hensen’s node) are attracted to the FGF4 present in the node and early notochord and so they migrate anteriorly, while FGF8, which is expressed more posteriorly in the late primitive streak, acts as a repulsive cue encouraging cells to migrate away from the streak (Yang et al., 2002).

During and subsequent to gastrulation, the anteroposterior axis of chordates lengthens by means of CE. The PCP pathway acts intracellularly to polarise a cell by driving proteins including Dishevelled (Dsh) to move to a particular part of the cell membrane. However, CE requires a field of cells to act together and to coordinately polarise and intercalate with each other. In C. intestinalis, FGF3 is expressed in the ventral midline of the nerve cord (neural tube), lying just dorsal to the forming notochord that expresses FGFR. Notochord cells expressing dnFGFR, or depleted of FGF3, fail to intercalate suggesting that FGF signalling provides an exogenous cue that gives directionality to the intracellular PCP pathway (Shi et al., 2009). Interestingly, dpERK is not active in the developing notochord in Ciona, indicating that the MAPK pathway is not important in this process. The dorsal mesoderm in Xenopus embryos overexpressing dnFGFR1 also fails to extend along the anteroposterior axis and the abnormal cell movements exhibited by the dorsal mesoderm in these embryos results in their characteristic open-blastopore phenotype (Isaacs et al., 1994). These data clearly demonstrate that the cells of the dorsal mesoderm, that is, the organiser, require FGF signalling for appropriate cell movement.

Sprouty was identified in Drosophila as an antagonist of FGF signalling (Hacohen et al., 1998) and in Xenopus, Xsprouty2 has been shown to inhibit morphogenesis while not effecting mesoderm specification (Nutt et al., 2001). A related protein called Spred inhibits MAPK activation by FGF signalling, so it has been suggested that these two proteins act on different intracellular pathways downstream of FGF (Sivak et al., 2005). This conclusion was based on the finding that Sprouty has inhibited FGF-mediated calcium release, membrane localisation of PKCδ, and morphogenetic movements, without affecting dpERK activity or the expression of Xbra. In contrast, Spred did not inhibit membrane localisation of PKCδ, but did abolish Xbra expression and the activation of dpERK. Another study has shown that Sprouty, and not Spred, physically associates with paraxial protocadherin (PAPC) (Wang et al., 2008), a protein known to be important for CE movements in fish and frogs (Kim et al., 1998; Yamamoto et al., 1998). Because Sprouty is known to inhibit PCP-directed morphogenesis, and PAPC is known to be a positive regulator of PCP, one model is that the physical association of PAPC and Sprouty weakens Sprouty’s inhibitory activity (Wang et al., 2008). This notion is supported by experiments that show membrane localisation of PKCδ, stimulated by expressing the Wnt receptor Frizzled-7 (Fz7), is inhibited by Sprouty and that this inhibition is prevented by PAPC. PAPC on its own does not promote membrane recruitment of PKCδ indicating that PAPC is acting by sequestering Sprouty. Interestingly, the expression of both Sprouty and PAPC in early Xenopus embryos requires FGF signalling (Branney et al., 2009; Nutt et al., 2001). The neutrophin receptor homologue is a known transcriptional target of FGF signalling, which is required for normal CE and provides another plausible node of interaction for the FGF and PCP pathways (Chung et al., 2005). These data weave together the FGF signalling pathway with the PCP pathway downstream of noncanonical Wnt signalling.


Studies in amniote embryos have shown that FGF signalling is required for the development of the inner ear (Ladher et al., 2005; Wright and Mansour, 2003). FGF has been found to play a role in the induction of the otic vesicle and the activation of genes required for inner ear development including Pax2 and Nkx5.1. In addition, FGF also drives some of the morphological changes underlying ear development. Early FGF signalling leads to the formation of a progenitor domain that will give rise both to the inner ear and to the epibranchial placode from which several different ganglia will eventually arise. Mesodermally expressed FGF3 and FGF 19 promote proliferation in an otic-epibranchial precursor domain, and it is the attenuation of FGF signalling together with canonical Wnt signals that specifies inner ear fate, while continued FGF signalling promotes epibranchial placode development (Freter et al., 2008). After the otic placode is induced, the ectoderm thickens and subsequently invaginates to form a hallow ball called the otocyst that lies within the head mesenchyme. Explant studies in chick showed that invagination of isolated otic ectoderm requires application of FGF to the basal side of the tissue (Sai and Ladher, 2008). FGF was found to induce basally localised PLCγ and myosin II, which causes a local depletion of actin fibres on the basal part of the cells and an enrichment of actin filaments apically. These effects of FGF were found not to be sensitive to protein synthesis inhibitors, indicating that in these cells, FGF can directly remodel the cytoskeleton to influence morphogenesis of a tissue. Later during ear development, a specialised region of the inner epithelium called the organ of Corti is patterned into a highly ordered array of sensory and nonsensory cells. The nonsensory cells are supporting cells, and the sensory cells are hair cells that act as mechanosensory receptors and are critical in translating the sound waves into electrical signals that are sent to the brain. In addition to the well-described role of Notch signalling in patterning inner ear development (Daudet and Lewis, 2005), FGFRs 1 and 3 have been found to be important in specifying cell types in the auditory sensory epithelium (Colvin et al., 1996; Pirvola et al., 2002). The precise level of FGF8 activity in the organ of Corti is modulated by the presence of Sprouty2, and the amount of FGF signalling is interpreted to produce support cells called pillar cells. In the absence of Sprouty2, higher levels of FGF result in three pillar cells instead of two forming, which leads to a deformed cochlea and a deaf mouse. These effects can be partially rescued by genetically reducing FGF8 levels in Sprouty2 mutant mice (Shim et al., 2005). Hair cells are marked by the expression of the proneural gene atonal homologue 1 (Atoh1), and treatment with the FGF inhibitor SU4502 in zebrafish has shown that the expression of atoh1 and hair cell development requires continuing FGF signalling (Millimaki et al., 2007).


Mechanosensory hair cells are also found in another organ that requires FGF during its morphogenesis. In zebrafish, the lateral line placode is a migrating epithelium that deposits precursors of hair cell organs, called neuromasts, at regular intervals in a line along each side of the embryonic trunk. As the lateral line placode migrates from head to tail, rosettes of cells bud off from the epithelium just behind the leading edge. These rosettes, or proneuromasts, will give rise to the neuromasts. The leading edge of this migrating epithelium undergoes a partial EMT, or a pseudo-EMT, where cells at the leading edge lose some apicobasal polarity and have increased numbers of filapodia but remain in contact with neighbouring cells. FGFR1, FGF3, and FGF10 are all expressed in the migrating lateral line primordium, where the ligands are specifically expressed in cells at the leading edge and the receptor in the trailing region. In zebrafish embryos mutant for FGF3 and FGF10, or in embryos treated with the FGF inhibitor SU5402, migration of the lateral line primordium ceases and rosettes are not formed (Lecaudey et al., 2008; Nechiporuk and Raible, 2008). Time lapse microscopy revealed a requirement for FGF3 and FGF10 for the formation of the proneuromasts and that the formation of these rosettes is a prerequisite for the migration of the lateral line primordial (Nechiporuk and Raible, 2008). Together, the findings point to a role for FGF signalling in regulating the pseudo-EMT that underlies the deposition of neuromasts along the trunk.


FGF plays a role in EMT during amniote gastrulation and lateral line migration in zebrafish; however, FGF is also known to be essential for mesenchymal to epithelial transformation (MET) during the morphogenesis of the Xenopus larval kidney, the pronephros. The pronephros is comparable physiologically, morphologically, and developmentally to the cortical nephron present in the adult kidney of the frog and the metanephric kidney in mammals. The pronephric mesoderm is derived from the intermediate mesoderm and the condensation of the pronephric mesenchyme can be inhibited with SU5402. Pronephric morphogenesis begins as this mesenchyme segregates away from the lateral plate and somites and begins to migrate posteriorly. The pronephric duct forms as the migrating pronephric mesenchyme undergo epithelialisation in an anterior to posterior wave to form a hollow tube that fuses with the cloaca. FGF8 is expressed transiently in the Xenopus pronephros during its morphogenesis, and morpholino knockdown of FGF8 was found to block the development of the pronephric duct (Urban et al., 2006). A key aspect of these effects was the loss of epithelialisation in the pronephric duct in embryos lacking FGF8. There are therefore two stages of Xenopus kidney development that require FGF signalling: the specification and condensation of the pronephric mesoderm and its epithelialisation to form the pronephric duct. Interestingly, FGF signalling is also known to be important in the branching, cell proliferation, and growth of the mammalian metanephric kidney (Qiao et al., 2001). A recent study indicates that FGFR1 is required for the development of the mouse metanephric kidney (Gerber et al., 2009).


There are branched tubular networks in many vertebrate organs such as the kidney, the lungs, and the vasculature. Most of these develop from a simple epithelial sac that undergoes reiterative budding to form a complex, tree-like array. A clear and important role for FGF signalling during branching morphogenesis has been established for the development of the trachea system in the Drosophila larva (Metzger and Krasnow, 1999). Branchless (an FGF ligand) is expressed in cells surrounding the epithelial trachea cells and is responsible for activating Breathless (an FGFR) in the trachea cells, which results in the expression or activation of downstream components of the FGF pathway. Sprouty was also first identified by its effects on branching morphogenesis in Drosophila (Hacohen et al., 1998). Spry is expressed in cells close to the source of FGF and is important for locally preventing secondary branching. Similarly, FGF signalling has been found to be a key regulator of branching morphogenesis in the mouse lung.

The developing mouse lung is composed of an epithelial layer, the surrounding mesenchyme, and the outer mesothelium. Signalling between these layers is critical for the coordinated mesenchymal growth and epithelial branching essential for normal lung development. The branching morphogenesis from which the lung evolves has recently been described in detail and found to be remarkably stereotypical (Metzger et al., 2008). There are several FGF ligands expressed in the developing lung. FGF9 is expressed in the outer mesothelium as well as in the epithelial layer, and in FGF9–/– knockout mouse, there is a loss of mesenchymal proliferation as well as reduced branching (Min et al., 1998), suggesting FGF9 is a part of the epithelial to mesenchyme signal. Using explant cultures of mouse lungs, FGF9 was found to stimulate mesenchyme proliferation when supplied on a heparin bead, while FGF10 did not have this activity. FGF9 and canonical Wnt signals promote the proliferation of lung mesenchyme. These effects could in part be attributed to the ability of epithelial FGF9 to activate mesenchymal expression of sonic hedgehog (Shh), a signal known to promote proliferation of undifferentiated cells (White et al., 2006; Yin et al., 2008). FGF10 is expressed in the lung mesenchyme and FGFR2 is expressed in the lung epithelium; loss of function of either of these genes results in a total loss of branching (Abler et al., 2009; Arman et al., 1999; Min et al., 1998). FGF10 signals from the mesenchyme act on the epithelium to induce the expression of target genes that include Spry and Shh. Spry feeds back to inhibit MAPK activity in the tip cells and restrict the formation of new branches (Figure 11); Shh also inhibits local FGF signalling by repressing FGF10 expression (reviewed by Affolter et al., 2009; Horowitz and Simons, 2008).

FIGURE 11. FGF10 signals to drive branching morphogenesis in the developing mouse lung.


FGF10 signals to drive branching morphogenesis in the developing mouse lung. FGF10 is present in the mesenchyme, while FGFR2 is expressed in the epithelium. In response to FGF signals, the epithelium forms a bud and also activates the expression of Sprouty2 (more...)


In the developing embryo, there are two distinct sources of cells that give rise to the heart: the first heart field (FHF), which forms early in the lateral plate mesoderm to form the cardiac crescent during primitive streak stages, and the secondary heart field (SHF), which is initially located anterior and dorsal to the cardiac crescent in the pharyngeal mesoderm (Buckingham et al., 2005; Srivastava, 2006). The FHF fuses at the midline to form a simple heart tube composed of inner endocardial cells and outer myocardial cells. The morphogenesis of the heart continues as it loops dramatically to the right and the posterior part of the tube shifts towards the anterior. The cells from the SHF remain as progenitors until they migrate and incorporate into the heart, using the FHF as a scaffold. Fate mapping studies in mice using FGF10:LacZ transgenics to mark cells in the pharyngeal mesoderm revealed that while SHF gives rise to cells that populate most of the fully developed heart, cells from the FHF are largely restricted to the left ventricle (Zaffran et al., 2004). FGF8 and FGF10 are expressed in the SHF, but only FGF8 has a mutant phenotype, where an FGF8 hypomorph displays septation defects in the outflow tract (OFT) (Abu-Issa et al., 2002). A malformed OFT underlies many congenital heart diseases; failure of OFT septation disrupts the correct partitioning of blood flow because the OFT is not divided into the pulmonary artery and the aorta.

Using conditional loss-of-function mutants for FGFR1 and FGFR2 and gain-of-function Sprouty, the effects of FGF signalling on OFT morphogenesis was investigated (Park et al., 2008). When these mutant constructs were expressed in the SHF, severe OFT defects were apparent where endothelial cells failed to migrate and undergo EMT, in addition to a fewer number of mesoderm cells in these mutants. In contrast, knocking out FGF signalling in the endothelial cells caused no OFT defects. The requirement for FGF was determined to be in the mesoderm, as endothelial defects could be rescued with grafts of wild-type mesoderm suggesting the effects of blocking FGF signalling was not cell-autonomous. Interestingly, a microarray analysis of OFT derived from FGFR mutant SHF revealed that FGF is required not only for known targets of the FGF pathway, but also for the activity of the BMP pathway. BMP signalling is known to be important for the recruitment and differentiation of myocardial cells during heart development.

In zebrafish, FGF signalling has been found to be important for establishing the correct heart chamber proportions (Marques et al., 2008) and FGF8 mutants have been found to have notably small ventricles (Reifers et al., 2000). Using a drug (SU5402) to inhibit FGF signalling and a zebrafish line transgenic for an inducible activated FGFR, the effects and requirement for FGF signalling in heart development were tested at different developmental time points (Marques et al., 2008). These studies found that FGF plays both an early role in establishing pools of cardiac progenitors and later in regulating the number of cells that give rise to the ventricular myocardium.

Copyright © 2010 by Morgan & Claypool Life Sciences.
Bookshelf ID: NBK53153
PubReader format: click here to try


  • PubReader
  • Print View
  • Cite this Page

Other titles in this collection