Celsr1 suppresses Wnt5a-mediated chemoattraction to prevent incorrect rostral migration of facial branchiomotor neurons

ABSTRACT In the developing hindbrain, facial branchiomotor (FBM) neurons migrate caudally from rhombomere 4 (r4) to r6 to establish the circuit that drives jaw movements. Although the mechanisms regulating initiation of FBM neuron migration are well defined, those regulating directionality are not. In mutants lacking the Wnt/planar cell polarity (PCP) component Celsr1, many FBM neurons inappropriately migrate rostrally into r3. We hypothesized that Celsr1 normally blocks inappropriate rostral migration of FBM neurons by suppressing chemoattraction towards Wnt5a in r3 and successfully tested this model. First, FBM neurons in Celsr1; Wnt5a double mutant embryos never migrated rostrally, indicating that inappropriate rostral migration in Celsr1 mutants results from Wnt5a-mediated chemoattraction, which is suppressed in wild-type embryos. Second, FBM neurons migrated rostrally toward Wnt5a-coated beads placed in r3 of wild-type hindbrain explants, suggesting that excess Wnt5a chemoattractant can overcome endogenous Celsr1-mediated suppression. Third, rostral migration of FBM neurons was greatly enhanced in Celsr1 mutants overexpressing Wnt5a in r3. These results reveal a novel role for a Wnt/PCP component in regulating neuronal migration through suppression of chemoattraction.

Each hindbrain was sectioned by hand to isolate rhombomeres r2 to r5, which were mounted and imaged as cross-sections shown in the lower panels.Asterisks in E10.5 hindbrains indicate non-specific trapping of probe in the otic vesicles.(A, C) Celsr1 is expressed in midline tissues (arrowheads) at all axial levels (rhombomeres) at both ages, with stronger expression evident in r2 and r3.(B, D) Wnt5a is expressed (arrowheads) at a low level in rostral midline tissues of the E10.5 hindbrain, and at a higher level up to the r3/r4 boundary in the E11.5 hindbrain.At E11.5, Wnt5a is not expressed in midline tissues in r4 and r5 but is expressed strongly in the ventricular zone in r5 and more posterior rhombomeres.Scale bars: (in A) for A-D whole-mount hindbrain panels, 1000 µm; (in A, r2) for all rhombomere cross-section panels, 500 µm.

Fig. S1 .
Fig. S1.Celsr1 and Wnt5a are expressed in midline tissues at the onset of migration.Dorsal views of E10.5 (A, B) and E11.5 (C, D) hindbrains processed for Celsr1 (A, C) and Wnt5a (B, D) in situ hybridization.The whole-mounted hindbrains are shown in top panels.

Fig. S2 .
Fig. S2.Breeding scheme used to generate Celsr1; Wnt5a double mutants.The observed numbers of embryos obtained for various genotypes (pooled for all ages) approximate Mendelian ratios ruling out developmental lethality or arrest in double mutant embryos.

Fig. S3 .
Fig. S3.FBM migration phenotypes in Celsr1 and Wnt5a mutants.Dorsal views of E12.5 hindbrains processed for Tbx20 in situ hybridization.Tbx20 staining marks FBM neurons and trigeminal neurons (asterisk).For all samples, the extent of rostral migration was scored and subsequently quantified (G).While varying numbers of FBM neurons migrated rostrally in nearly all Celsr1 mutants (C), these neurons never migrated rostrally in Wnt5a mutants (F).Rostral migration was scored separately for the left and right sides of the hindbrain due to variable expressivity of the phenotype.Number of embryo sides in parentheses (double the number of embryos).Scale bar (in A) for A-F, 400 µm.

Fig. S6 .
Fig. S6.Hindbrain size and rhombomere dimensions in Wnt5a GOF embryos.(A, C) Dorsal views of E12.5 hindbrains processed for Tbx20 (A) and Wnt5a (C) in situs.(A)Length and width were defined as shown, using as landmarks the outlined edges of the trigeminal motor nuclei in r2 (circled, asterisk) and the caudal edge of the stream of migrating FBM neurons in r6.The boundaries of r3 and r4 were defined for control embryos from the location of the Tbx20-expressing trigeminal and FBM neurons and used to establish the corresponding boundaries (and rhombomere lengths) in Wnt5a GOF embryos.(B) Hindbrain

Fig. S7 .
Fig. S7.Quantifying rostral migration of FBM neurons using the Fiji (ImageJ) program.(A-D) Dorsal views of E12.5 hindbrains processed for Tbx20 in situs.Using the caudal margin of the trigeminal motor nuclei (asterisks), the putative r3/4 boundary was drawn using the