Fig. 7. BDNF/TrkB signalling is required for terminal innervation and formation of functional sensory nerve endings. (A-H) Immunohistochemical double staining against NF200 to detect nerve fibres (red) and calretinin to detect calyces (green). (A,B) Whole-mount preparations from the utricular maculae in wild-type (A) and BDNF^NT3/NT3 mice (B). Scale bars: 30 µm in A and B. The analyses using confocal microscopy confirmed a sparse innervation of the epithelia and also revealed that the nerve fibres in the subepithelial layer were highly disorganised in BDNF^NT3/NT3 mice (B). (C-H) Co-cultures of wild-type vestibular ganglion neurons with wild-type hair cells (C-D) BDNF^–/– hair cells (E-F), or BDNF^NT3/NT3 hair cells (G-H). Scale bars: 20 µm in C-H. With wild-type hair cells the nerve endings branch extensively and wrap around the hair cells forming digitations similar of the process of calyx formation in vivo (C, nerve including calretinin-stained hair cells and D, nerve only). In only two out of 200 cases did wild-type neurons co-cultured with BDNF^–/– hair cells contact a hair cell (E-F). Note that the diameter of the nerve fibre is much thinner and it shows less branching on the surface of the BDNF^–/– hair cell. (G,H) In contrast to BDNF^–/– hair cells, nearly all BDNF^NT3/NT3 hair cell fragments were contacted by vestibular neurons (11 out of 12). However the contacting neurites remained undifferentiated on the surface of the epithelia and did not develop digitations.