BDNF promotes target innervation of Xenopus mandibular trigeminal axons in vivo.
Trigeminal nerves consist of ophthalmic, maxillary, and mandibular branches that project to distinct regions of the facial epidermis. In Xenopus embryos, the mandibular branch of the trigeminal nerve extends toward and innervates the cement gland in the anterior facial epithelium. The cement gland has previously been proposed to provide a short-range chemoattractive signal to promote target innervation by mandibular trigeminal axons. Brain derived neurotrophic factor, BDNF is known to stimulate axon outgrowth and branching. The goal of this study is to determine whether BDNF functions as the proposed target recognition signal in the Xenopus cement gland. We found that the cement gland is enriched in BDNF mRNA transcripts compared to the other neurotrophins NT3 and NT4 during mandibular trigeminal nerve innervation. BDNF knockdown in Xenopus embryos or specifically in cement glands resulted in the failure of mandibular trigeminal axons to arborise or grow into the cement gland. BDNF expressed ectodermal grafts, when positioned in place of the cement gland, promoted local trigeminal axon arborisation in vivo. BDNF is necessary locally to promote end stage target innervation of trigeminal axons in vivo, suggesting that BDNF functions as a short-range signal that stimulates mandibular trigeminal axon arborisation and growth into the cement gland.
PubMed ID: 17540021
PMC ID: PMC1899173
Article link: BMC Dev Biol.
Grant support: RR13221 NCRR NIH HHS , Wellcome Trust
Genes referenced: bdnf ntf3 ntf4 odc1 plaa
Antibodies referenced: Tuba4b Ab5 Tubb2b Ab3
Morpholinos referenced: bdnf MO2
Article Images: [+] show captions
|Figure 1. Xenopus mandibular trigeminal nerve innervates the cement gland. (A) Whole mount immunostaining of a Xenopus embryo (St. 29) labeled with β-tubulin antibody and cleared in Murray's clearing solution reveals the embryonic axonal network, including the prominent mandibular trigeminal nerve. (B) The mandibular trigeminal nerve (arrow) extends between the eye vesicle (e) and otic vesicle (ov), and then turns ventrally before terminating at the ventral region of the cement gland (asterisk). Deletion of cement glands results in the loss of trigeminal targeting, where the axons either (C) stop growth, or extended (D) dorsally or (E) ventrally. Because the embryos are transparent, trigeminal axons on the opposite side could also be detected. In the cement gland null embryos, trigeminal axons on either side of the face do not necessarily project erroneously in the same directions.|
|Figure 2. Detection of BDNF transcripts in the Xenopus cement gland. (A) RT-PCR shows that BDNF, NT3 and NT4 were detected throughout all stages of embryonic development. Expression of these genes increases during neurulation (St. 18). The BDNF transcript was particularly enriched in the cement gland compared to the other neurotrophin transcripts at tailbud stage (St. 24). The amount of input cDNA was confirmed with primers generated against the housekeeping gene, ornithine decarboxylase (ODC) to ensure equal loading. (B) Whole mount in situ hybridisation shows that BDNF mRNA are initially detected in the neural plate at St. 18. (C) At St. 24, BDNF is detected at the posterior region of the cement gland (arrowhead), which corresponds to the time of trigeminal nerve innervation. (D) At St. 29, BDNF expression remains in the cement gland and can also be detected in the eye and otic vesicles. (E) Inset from (C) shows the detection of BDNF mRNA in the cement gland at St. 24. we = whole embryo, cg = cement gland, we-cg = whole embryo without cement gland.|
|Figure 3. Effect of BDNF antisense morpholino oligonucleotides (MO) on Xenopus morphology and behaviour. (A) BDNF MO was designed against the 5' UTR and start ATG region (underlined). This region was 100% conserved between X. laevis and X. tropicalis. Asterisks indicate conserved bases. (B) Embryos were injected with 250 pg of XtBDNF-HA mRNA alone or with 20 ng of MO BDNFatg or the control MO (MOC). Embryos were harvested at St 12 and proteins extracts analysed by Western blotting. Xt BDNF was detected using anti-HA antibodies (arrowhead). The asterisk indicates a non-specific band from the same blot, which ensures equal loading in each lane. (C) Uninjected X. laevis embryos at St. 29. (D) Control morpholino (MOC) injected embryos at St. 29 developed normally. (E) BDNF morphants appeared slightly truncated at St. 29. (F) An uninjected tadpole with normal mechanosensory response, which escaped when probed with forceps. (G) BDNF morphants exhibited impaired mechanosensory response and did not escape when probed.|
|Figure 4. MO BDNFatg injected into N-PLAP transgenic X. tropicalis embryos exhibit possible disruption in mandibular trigeminal axon targeting. (A) A transgenic X. laevis embryo (St. 29) expressing N-PLAP following alkaline phosphatase reaction demonstrates the neurospecific labeling of CNS axons. The mandibular trigeminal nerve can be observed to project to the cement gland where terminal arborisation occurs. (B) The detection of cranial nerves in a transgenic tadpole (St. 47). (C) An uninjected X. tropicalis transgenic embryo (WT) displaying normal trigeminal nerve projection. (D) A MOC injected transgenic embryo and (E) a MO NT3 injected embryo show normal trigeminal nerve projection. (F) A MO BDNFatg injected transgenic embryo shows that the trigeminal axons appeared thinner and possibly less fasciculated. Some of the axons did not extend to the posterior cement gland where they normally innervate.|
|Figure 5. Decreased axon terminal arborisation in cement glands of BDNF morphants. Whole mount immunofluorescent staining of St. 29 X. laevis embryos with anti-acetylated α-tubulin shows trigeminal axon arborisation and growth into the cement gland in (A), (B) and (C) uninjected, and (D), (E) and (F) MOC injected embryos. (G), (H) and (I) In MO BDNFatg injected embryos trigeminal axons extended to the cement gland but did not arborise or grow into the cement gland.|
|Figure 6. Increased apoptosis observed in the head region of BDNF morphants. Whole mount TUNEL staining of uninjected embryos at (A) St. 27 and (B) 32. In MO BDNFatg injected morphants, cell death activity in the head region was comparable to those in uninjected embryos at (C) St. 27, but number of cells undergoing apoptosis increased significantly at (D) St. 32. (E) Quantitation of cells undergoing apoptosis shows a 3-fold increase of cell death activity in St. 32 BDNF morphants.|
|Figure 7. BDNF expressed ectodermal animal cap grafts stimulate mandibular trigeminal axon arborisation. (A) Diagram of the in vivo cement gland substitution assay. (B) Successfully grafted embryos were identified by GFP expression. (C) Neither GFP+ nor (D) NT3+ grafts stimulated trigeminal axon arborisation or growth onto the grafts. (E) BDNF+ graft stimulated axon terminal arborisation onto the graft. AC = animal cap, CG = cement gland, Asterisk = graft.|
|Figure 8. Wildtype cement gland grafts rescued mandibular trigeminal axon target innervation in MO BDNFatg injected embryos. In vivo cement gland swap experiments were performed between MOC injected embryos and uninjected embryos (WT). (A) MOC cement gland placed onto an uninjected embryo exhibits trigeminal axon arborisation and growth into the cement gland. (B) Similar observation was made when wildtype cement gland was placed onto a MOC injected embryo. In vivo cement gland swaps were also performed between BDNF morphants and uninjected embryos. (C) MO BDNFatg cement gland placed onto an uninjected embryo shows trigeminal arborisation at the cement gland was barely detectable. (D) Wildtype cement gland placed onto a MO BDNFatg injected embryo shows that trigeminal axons were able to arborise and enter the cement gland.|