One of the most intriguing questions in developmental neuroscience is: how are growth cones transformed into presynaptic terminals? Retrograde signals from the postsynaptic cell are believed to be instrumental in this process, and a new study from the laboratory of Joshua Sanes has uncovered an important role for members of the fibroblast growth factor (FGF) family in presynaptic organization.

Umemori et al. prepared an extract from postnatal day 7 mouse forebrains, and screened this extract for molecules that could induce clustering of synaptic vesicles in the neurites of cultured chick motor neurons. They purified one of the main components of the extract that promoted vesicle clustering, and identified this component as FGF22. The FGF22-induced vesicle clusters underwent recycling in response to neuronal depolarization, implying that FGF22 was promoting the development of functional presynaptic structures.

The authors then went on to investigate whether FGF22 promotes presynaptic differentiation in vivo, and they obtained several lines of evidence to indicate that this is indeed the case. FGF22 is expressed in cerebellar granule cells at the same time that they are being innervated by mossy fibres from vestibular and pontine neurons, and, in turn, the vestibular and pontine neurons express the FGFR2 gene, which codes for the main FGF22 receptor FGFR2b. Mossy-fibre presynaptic differentiation was impaired both in vivo and in vitro when FGF22 activity was neutralized with a soluble recombinant form of FGFR2b (FGFR2bAP), and conditional deletion of the FGFR2 gene in newborn mice produced a similar result.

FGF22 belongs to an FGF subfamily that also includes FGF7 and FGF10. All three molecules were shown to promote vesicle clustering in vitro, and all three bind to FGFR2b and are neutralized by FGFR2bAP. Therefore, the authors could not rule out the possibility that FGF7 and FGF10 were responsible for the effects that they observed in vivo. However, FGF22 is expressed at a much higher level in the cerebellum than either FGF7 or FGF10, so it seems likely that it is the main presynaptic organizing molecule, at least in this part of the brain.

Through this work, Umemori et al. have not only defined a new role for postsynaptically derived FGFs in presynaptic assembly, but they have also assigned a developmental function to the FGF7/10/22 subfamily, which has not previously been implicated in neural development. The next step will be to show by more direct means that FGF22 induces presynaptic differentiation in vivo, and mice with mutations in the FGF22 gene are already being generated to test this hypothesis. It will also be interesting to find out whether FGF7 and FGF10 have similar roles at other types of synapse.