Four genes code for FGF receptors (FGFRs), which are receptor tyrosine kinases (RTKs) and are activated by ligand binding. FGFRs have three immunoglobulin-like domains Ig1, Ig2, and Ig3 that bind ligand, HS-GAG, and are important in receptor dimerisation. Alternative splicing of part of the Ig3 domain generates different isoforms of FGFR1–3 with distinct binding specificities (Johnson et al., 1991) that are expressed in different tissues (Orr-Urtreger et al., 1993). Additionally, an individual FGF ligand may have a preference for a particular receptor. Using a mitogenic assay in cultured cells expressing distinct FGFRs, only FGF1 was shown to promiscuously activate all four FGFRs while the other ligands were more effective on cells expressing different receptors. For instance, the FGF4 subfamily is more active on cells expressing FGFR1c or 2c, while the FGF7 (see Table 1) subfamily is more active on cells expressing FGFR1b or 2b (Zhang et al., 2006). Another splice variant in FGFR1 has two amino-acid deletion of Val(423)–Thr(424) in the juxtamembrane region and shows different expression and activity than the isoform that includes these two amino acids (Paterno et al., 2000).
An in vivo study has shown that overexpressing constitutively active or dominant negative forms of different FGFRs in zebrafish embryos will generate different phenotypes (Ota et al., 2009). Overall, it was found that all of the constitutively active FGFRs tested gave rise to posteriorised/dorsalised embryos that lack forebrain structures and all of the dominant negative forms cause posterior truncations in zebrafish as described previously for Xenopus (Amaya et al., 1991). However, the potency of different receptors and their ability to affect the activity of individual FGF ligands varied. The dominant negative FGFRs (dnFGFR) were constructed by deleting the cytoplasmic kinase domains. When the dnFGFR are overexpressed in the embryo, the mutant receptors partner with endogenous wild-type FGFRs, but because the activation of signal transduction relies on cross-phosphorylation of the intracellular kinase domains, these mutant/wt FGFR dimers will be nonfunctional. A dominant negative form of FGFR1 (dnFGFR1) has been widely used in developmental biology (Amaya et al., 1991; Amaya et al., 1993; Branney et al., 2009; Griffin et al., 1995; Isaacs et al., 1994; Pownall et al., 1998) and has been shown to completely block the activation of ERK in early frog embryos (Christen and Slack, 1999). Although overexpressing a dominant negative form of FGFR4 (dnFGFR4) was found to give rise to a different phenotype than dnFGFR1 (Hardcastle and Papalopulu, 2000; Hongo et al., 1999), it is known that these receptors will heterodimerise promiscuously (Ueno et al., 1992) and therefore the specificity of the assay is lost. This has been shown recently for dnFGFR1 and dnFGFR4 in a microarray study (Branney et al., 2009).
In addition to the four RTKs, another member of the FGFR family has been identified in a number of species (Sleeman et al., 2001). The extracellular domain of FGFRL1 shares about 50% amino acid identity with the other FGFRs. As with FGFR1–4, FGFRL1 has been shown to have differing affinity for the various FGF ligands (Bertrand et al., 2009; Trueb et al., 2005); however, FGFRL1 lacks any intracellular tyrosine kinase activity. It has recently been shown that FGFRL1 is shed from cells and when overexpressed in Xenopus embryos can mimic the morphological effects of FGF inhibition (Steinberg et al., 2010).