Of the five known high-affinity FGFRs, FGFR-4 is relatively little studied, especially within the central nervous system. In the mature brain, it is reported to be expressed principally within the median habenular nucleus,
40 whereas we have shown previously that it is expressed by adult retina,
20 22 predominantly within photoreceptors. It is more widely expressed during the development of the central nervous system
39 and has been reported to be a precocious marker of human macula, where it may play a role in cone differentiation.
41 42 Transgenic
Xenopus, carrying a dominant-negative FGFR-4 construct targeted to the retina, displays profound defects in retinal formation,
43 underscoring the importance of this receptor in retinal development. Although the activation and intracellular signaling of FGFR-4 are only poorly known, data indicate several notable differences with respect to the other FGFRs. When transfected into COS cells, FGFR-4 is the only FGFR that induces membrane ruffling, indicative of cytoskeletal reorganization.
44 Originally reported to constitute a specific high-affinity receptor for FGF-1,
45 FGFR-4 activation is prolonged compared with FGFR-1, and there is only weak tyrosine phosphorylation of intracellular proteins, including mitogen-activated protein kinase.
46 47 Ligand binding has been reported to activate a unique 85-kDa protein with serine/threonine phosphorylation activity.
48 Some of these characteristics were also seen in the studies performed here. FGF-19 binding led to delayed but prolonged receptor activation, with tyrosine phosphorylation still observed after 45 minutes. By comparison, FGFR-1, present in photoreceptors, is rapidly (within 30 seconds) activated by FGF-2 and deactivates within 5 to 10 minutes.
16 18 In the studies performed here, immunoprecipitation of FGF-19-stimulated cells and immunoblotting with phosphotyrosine antibody revealed a band of approximately 80 kDa, which may represent the associated intracellular signaling protein or an isoform of FGFR-4 itself. Interestingly, we observed coimmunoprecipitation of FGFR-1 and FGFR-4 after FGF-19 stimulation of cultured cells. Immunoblotting for FGFR-1 in FGFR-4-immunoprecipitated material detected a single band at approximately 110 kDa, in accordance with one of the FGFR-1 isoforms observed in previous studies of the retina.
18 The intensity of this band varied, depending on the duration of FGF-19 stimulation; it was maximal after 15 minutes, indicating that the formation of FGFR-1/FGFR-4 heterodimers increased after receptor activation. To the best of our knowledge, this is the first time dimerization between these two receptors has been identified. Our previous studies indicate that photoreceptors contain mainly FGFR-1 and FGFR-4 in approximately equal concentrations.
19 20 22 It is, hence, plausible that the two forms interact physically on ligand binding. However, we have no formal proof that FGF-19 uniquely activates FGFR-4, as has been reported under cell-free conditions.
26 Cellular context is very important in determining receptor-binding interactions, and these parameters have not been examined in detail within the retina. Finally, very recent studies demonstrate the existence of coreceptors that interact with FGFR-4, termed αKlotho and βKlotho.
49 50 Highly abundant in liver and fat cells, these single-span transmembrane proteins increase FGF-19/FGFR-4 interactions in a heparin-dependent manner, suggesting a complex receptor organization for this factor. It will be of great interest to see whether αKlotho and βKlotho are expressed in retina.