An autocrine Wnt5a-Ror signaling loop mediates sympathetic target innervation

Abstract

During nervous system development, axon branching at nerve terminals is an essential step in the formation of functional connections between neurons and target cells. It is known that target tissues exert control of terminal arborization through secretion of trophic factors. However, whether the in-growing axons themselves produce diffusible cues to instruct target innervation remains unclear. Here, we use conditional mutant mice to show that Wnt5a derived from sympathetic neurons is required for their target innervation in vivo. Conditional deletion of Wnt5a resulted in specific deficits in the extension and arborization of sympathetic fibers in their final target fields, while no defects were observed in the overall tissue patterning, proliferation, migration or differentiation of neuronal progenitors. Using compartmentalized neuronal cultures, we further demonstrate that the Ror receptor tyrosine kinases are required locally in sympathetic axons to mediate Wnt5a-dependent branching. Thus, our study suggests an autocrine Wnt5a-Ror signaling pathway that directs sympathetic axon branching during target innervation.

Fig. 1

Generation of conditional Wnt5a mutant mice. (A) Strategy for generating the conditional Wnt5a allele. Exons 2 and 3 in the Wnt5a gene are indicated in each line. (B) Genotyping results from Wnt5a^+/+, Wnt5a^fl/+ and Wnt5a^fl/fl mice using primers (purple arrows in A) targeting the first loxP site. (C) An E16.5 Wnt1::Cre:Wnt5a^fl/− embryo shows normal anterior-posterior length, limbs and tail but has craniofacial abnormalities as compared to a wild-type embryo. The craniofacial defects in the Wnt1::Cre:Wnt5a^fl/− embryo are similar to that seen in a constitutive Wnt5a null embryo obtained from the same litter due to occasional germline expression of Cre. (D–E) In situ hybridization using a probe that spans the targeted exon 2 of Wnt5a shows decreased Wnt5a transcript in E16.5 Wnt1::Cre;Wnt5a^fl/− mice in the superior cervical ganglia (SCG) (D) but the Wnt5a signal can be detected in salivary glands (E). Scale bar, 1 cm for C and 100 μm for D, E. For all experiments, 3 embryos for each genotype were analyzed.

Fig. 2

Neuronal Wnt5a is dispensable for sympathetic neuron differentiation, proliferation and survival. (A) Normal expression of the sympathetic lineage marker, Phox2b, in E13.5 control and Wnt1::Cre:Wnt5a^fl/− SCGs. Scale bar, 50 μm. (B) A 24-hr pulse of EdU reveals similar levels of proliferating cells in Wnt1::Cre:Wnt5a^fl/− and control SCGs at E13.5. Scale bar, 100μm. Dashed lines outline the SCG. (C) Quantification of the total number of EdU-positive cells per SCG. Values are the mean ± s.e.m., n=3 embryos. (D) The general architecture of the sympathetic chain is normal in E16.5 Wnt1::Cre:Wnt5a^fl/− mice, with ganglia size comparable to that in wild-type controls. Scale bar, 250 μm. (E, F) Immunostaining against TrkA (E) and brain-lipid-binding protein (BLBP) (F) at P0.5 show no changes in the expression of mature neuronal and glial markers upon conditional deletion of Wnt5a. Scale bar, 100 μm. (G) Quantification of neuronal numbers show that there is no significant cell loss in the developing SCG at E16.5 and P0.5 in Wnt1::Cre:Wnt5a^fl/− mice compared to controls. Values are the mean ± s.e.m., n=3 animals for each genotype at each developmental stage.

Fig. 3

Wnt5a is required for terminal arborization of sympathetic axons. Wnt1::Cre;Wnt5a^f/− mice show reduced sympathetic arborization in several peripheral targets. Major axon bundles reach the final targets in the mutants, but they fail to extend and branch within final target fields as compared to control litter-mates. (A–J) Whole-mount TH staining of salivary glands (A), thymus (C), heart (E), spleen (G) and bladder (I) in E16.5 WT and Wnt1::Cre;Wnt5a^fl/− mice. Quantification of branching in the target tissues are shown in (B, salivary glands), (D, thymus), (F, heart), (H, spleen) and (J, bladder). Scale bar, 200 μm. n=5 embryos for each genotype. Values are the mean ± s.e.m. * p<0.05.

Fig. 4

Sympathetic innervation defects in TH::Cre;Wnt5a^fl/fl embryos. (A, B) In situ hybridization shows decreased Wnt5a transcript in E16.5 TH::Cre;Wnt5a^fl/fl mice in the superior cervical ganglia (SCG) (A) but the Wnt5a signal can be detected in salivary glands (B). Scale bar, 100 μm. (C) Q-PCR analyses show that Wnt5a transcript levels are significantly reduced in sympathetic ganglia, but not salivary glands in TH::Cre;Wnt5a^fl/fl mice at P0.5. SCGs were harvested from n=5 Wnt5a^fl/fl and n=8 TH::Cre;Wnt5a^fl/fl mice. Salivary glands were dissected from n=5 Wnt5a^fl/fl and n=9 TH-Cre;Wnt5a^fl/fl mice. Values are the mean ± s.e.m. * p<0.05. (D–G) Whole mount TH immunostaining of target tissues shows reduced sympathetic innervation of the salivary glands (D) and heart (F) in E16.5 TH::Cre;Wnt5a^fl/fl embryos compared to control Wnt5a^fl/fl mice. Scale bar: 200 μm. Quantification of the branching are indicated in (E, salivary glands) and (G, heart). Salivary glands were harvested from n=5 animals and hearts from n=4 animals, for each genotype. * p<0.05, ** p<0.01. (H) TH immunostaining of tissue sections show a reduction in superior cervical ganglia (SCG) in TH::Cre;Wnt5a^fl/fl mice compared to control Wnt5a^fl/fl mice at one week after birth (P7), Scale bar, 200μm. (I) Quantification of neuronal numbers show no change in SCG cell numbers at P0.5 but significant cell loss by P7 in TH::Cre:Wnt5a^fl/fl mice compared to Wnt5a^fl/fl mice. n=4 animals for each genotype at P0 and n=3 animals at P7. Values are the mean ± s.e.m.,*p<0.05.

Fig. 5

Wnt5a-dependent branching is mediated by local actions of Ror receptors in sympathetic axons. (A–D) In compartmentalized cultures, Wnt5a-treated axons exhibit robust axon growth and branching (B), which is eliminated by the addition of Ror1/2 neutralizing antibodies to distal axons (da) (C), but not cell bodies (cb) (D). Compartmentalized cultures were initially maintained in NGF-containing media for 5–7 days to allow robust growth into axon compartments. NGF was then withdrawn from the culture media, and Ror1/2 inhibitory antibodies added either to the axon or cell body compartments in the presence of either control- or Wnt5a-conditioned medium for 24 hr. The broad-spectrum caspase inhibitor, BAF, was added to the cell body compartments to prevent neuronal apoptosis in the absence of NGF. Axons were stained with β-III-tubulin for visualization. Scale bar, 100 μm. Quantification of Wnt5a-mediated axon branching (E) and growth (F) in compartmentalized cultures over 24 hr. n = 4 independent experiments, **p < 0.01.