Ryk regulates Wnt5a repulsion of mouse corticospinal tract through modulating planar cell polarity signaling

Abstract

It was previously reported a role for Ryk in mediating Wnt5a repulsion of the corticospinal tract (CST) in mice. Recent evidence has shown that Ryk regulates planar cell polarity (PCP) signaling through interacting with Vangl2. Here, in vivo, in vitro and biochemical analyses were applied to investigate the molecular cross-talk between the Ryk and PCP signaling pathways, revealing that PCP pathway components play important roles in CST anterior-posterior guidance. Ryk-Vangl2 interactions are crucial for PCP signaling to mediate Wnt5a repulsion of CST axons. Cytoplasmic distribution of Ryk is increased under high concentrations of Wnt5a and facilitates the cytoplasmic distribution of Vangl2, leading to inhibition of Frizzled3 translocation to cytoplasm. Alternatively, Ryk stabilizes Vangl2 in the plasma membrane under low Wnt5a concentrations, which promotes cytoplasmic translocation of Frizzled3. We propose that Ryk regulates PCP signaling through asymmetric modulation of Vangl2 distribution in the cytoplasm and plasma membrane, which leads to repulsion of CST axons in response to the Wnt gradient.

Keywords: Ryk; Wnt5a; corticospinal tract; planar cell polarity signaling; repulsion.

Figure 1

Upregulation of Ryk, Fzd3, Vangl2 and Dvl1 in developing corticospinal neurons and axons. (a) Western blot analysis of Ryk and core PCP pathway components Fzd3, Vangl2 and Dvl1 in E18.5 and P0 cortex. (b) Quantification of Ryk, Fzd3, Vangl2 and Dvl1 immunoblotting intensity in E18.5 and P0 cortex. Data are represented as the mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001 (100% of E18.5, P0 Ryk=210±25.3%, t=4.34, P=0.0123; P0 Fzd3=216±15.4%, t=7.53, P=0.0017; P0 Vangl2=178±17.0%, t=4.63, P=0.0098; P0 Dvl1=197±8.36%, t=11.7, P=0.0003. Student’s t-test). Data were analyzed from three independent experiments, each including three mice per group. (c–f) Expression of Ryk (c) and core PCP pathway components Fzd3 (d), Vangl2 (e) and Dvl1 (f) in E18.5 corticospinal neurons. (g–j) Expression of Ryk (g) and core PCP pathway components Fzd3 (h), Vangl2 (i) and Dvl1 (j) in P0 corticospinal neurons. Co-staining was carried out with Ctip2 (blue) and E-cadherin (green) antibodies and Ryk, Fzd3, Vangl2 or Dvl1 antibodies (red). Scale bar, 50 μm. (c1–f1) Higher magnification images of the boxed areas in c–f. Scale bar, 10 μm. (k–n) Expression of Ryk (k) and core PCP pathway components Fzd3 (l), Vangl2 (m) and Dvl1 (n) in E18.5 corticospinal axons. (o–r) Expression of Ryk (o) and core PCP pathway components Fzd3 (p), Vangl2 (q) and Dvl1 (r) in P0 corticospinal axons. Co-staining was carried out with Ctip2 (blue), Tuj1 (red) antibodies and Ryk, Fzd3, Vangl2 or Dvl1 antibodies (green). Scale bar, 5 μm. (s) Quantification of Ryk, Fzd3, Vangl2 and Dvl1 staining intensity in E18.5 and P0 corticospinal neurons. Data are represented as the mean±s.e.m. ***P<0.001 (100% of E18.5, P0 Ryk=174±0.83%, t=88.5; P0 Fzd3=181±1.37%, t=58.8; P0 Vangl2=172±3.99%, t=18.0; P0 Dvl1=167±2.66%, t=25.0. Student’s t-test). Data were analyzed from three independent experiments, each including 20 corticospinal neurons per group.

Figure 2

PCP pathway components mediate Wnt5a repulsion of corticospinal axon growth cones. (a) Analysis of the knockdown efficiency of Ryk, Vangl2, Fzd3 and Dvl1 shRNA by western blotting. GAPDH was used as an internal control. (b) Quantification of the knockdown efficiency of Ryk, Vangl2, Fzd3 and Dvl1 shRNA. Data are represented as the mean±s.e.m. **P<0.01, ***P<0.001 (100% of control shRNA, Ryk shRNA=28.4±9.3%, t=7.70, P=0.0015; Vangl2 shRNA=25.6±12.8%, t=5.82, P=0.0011; Fzd3 shRNA=5.3±1.1%, t=84.30, P<0.0001; Dvl1 shRNA=21.0±8.4%, t=9.40. Student’s t-test). Data were analyzed from three independent experiments per group. (c) Schematic diagram of in vivo experimental design to expose corticospinal axon growth cones to Wnt5a gradient using the Dunn chamber. Asterisk indicates the turning angel between midline (black dotted line) and the final position of the growth cone (red dotted line). (d) Growth cones of corticospinal axons expressing control shRNA or Ryk, Vangl2, Fzd3 or Dvl1 shRNA responding to Wnt5a gradient. Lentivirus-infected growth cones were identified with EGFP expression (green), and immunostaining was carried out with Ryk, Vangl2, Fzd3 or Dvl1 antibodies (red). Dashed lines outline the growth cones and indicate the midline of the growth cones. Scale bar, 2 μm. (e) Quantification of growth cone-turning ratio in the different groups. Data are represented as the mean±s.e.m. ***P<0.001 (control shRNA=82.0±0.67%, Ryk shRNA=13.0±1.03%, Vangl2 shRNA=14.6±1.89%, Fzd3 shRNA=8.6±2.46%, Dvl1 shRNA=16.7±2.21%. One-way ANOVA). Data were analyzed from three independent experiments. Each independent experiment included at least 18 neurons per group.

Figure 3

In vivo knockdown of Ryk and PCP pathway components results in A–P guidance defects of corticospinal axons. (a) Control-shRNA-expressing CST in developing spinal cord. The single green arrow indicates a CST projection initiation site in the spinal cord at first cervical (C1) segment. The double green arrows indicate control-shRNA-expressing CST continuing to descend at the fifth thoracic (T5) segment. (b, c) Higher magnification images of the boxed areas in a. (d) Ryk shRNA-expressing CST in developing spinal cord. The white arrow indicates where Ryk shRNA-expressing CST axons initiated randomized extension at the third cervical (C3) segment. The red arrow indicates where Ryk shRNA-expressing CST aberrantly ends at the sixth cervical (C6) segment. (e) Higher magnification image of boxed area in d. (f) Fzd3 shRNA-expressing CST in developing spinal cord. The white arrow indicates where Fzd3 shRNA-expressing CST axons initiated randomized extension at the second cervical (C2) segment. The red arrow indicates where Fzd3 shRNA-expressing CST aberrantly ends at the sixth cervical (C6) segment. (g) Higher magnification image of boxed area in f. (h) Vangl2 shRNA-expressing CST in developing spinal cord. The white arrow indicates where Vangl2 shRNA-expressing CST axons initiated randomized extension at the third cervical (C3) segment. The red arrow indicates where Vangl2 shRNA-expressing CST aberrantly ends at the fifth cervical (C5) segment. (i) Higher magnification image of the boxed area in h. (j) Dvl1 shRNA-expressing CST in developing spinal cord. The white arrow indicates where Dvl1 shRNA-expressing CST axons initiated randomized extension at the fifth cervical (C5) segment. The red arrow indicates where Dvl1 shRNA-expressing CST aberrantly ends at the sixth cervical (C6) segment. (k) Higher magnification image of the boxed area in j. Axons were identified with EGFP expression (green). Scale bar, 500 or 100 μm (higher magnification images). (l, m) Quantification of CST area and CST bundle length in each shRNA-expressing group. Data are represented as the mean±s.e.m. **P<0.01, ***P<0.001. Data of CST area and CST bundle length were analyzed from at least seven mice in each group using two-way ANOVA (l) or one-way ANOVA (m, 100% of control shRNA, Ryk shRNA=31.14±3.526%; Fzd3 shRNA=36.37±7.392%; Vangl2 shRNA=17.95±7.634%; Dvl1 shRNA=35.11±7.635%; P<0.001), respectively. NS, not significant.

Figure 4

Ryk and Vangl2 interact in corticospinal neurons. (a) Co-localization of Ryk and Vangl2 in E18.5 and P0 corticospinal neurons. Co-staining with Ryk (blue), Vangl2 (red) and E-cadherin (E-cad, green) antibodies was carried out on sections from E18.5 and P0 cortex. Anti-E-cadherin staining was used to visualize the cell membrane. Scale bar, 5 μm. (b) Quantification of the co-localization of Ryk and Vangl2 from (a). Data are represented as the mean±s.e.m. of four independent experiments. ***P<0.001 (100% of E18.5 Ryk and Vangl2 co-localization, P0 Ryk and Vangl2 co-localization=242±11.68%, t=12.16, P=0.0003, Student’s t-test). (c) Co-immunoprecipitation (Co-IP) of Ryk and Vangl2 expressed in corticospinal neurons. The whole-cell lysate, membrane and cytoplasm extracts from E18.5 and P0 cortex samples were subjected to co-immunoprecipitation followed by western blot. GAPDH was used as an internal control, and confirmed the separation of the cytosolic and membrane fractions. (d) Quantification of co-immunoprecipitated Ryk and Vangl2 in whole-cell lysate, cytoplasm and membrane extracts. Data are represented as the mean±s.e.m. of four independent experiments. **P<0.01, ***P<0.001 (100% of E18.5 whole lysate. first and second columns, 100% of E18.5 whole lysate, P0 whole lysate=209.4±32.19%, t=4.74; third and fourth columns, E18.5 cytoplasm=85.794±7.685%, P0 cytoplasm=166.193±20.865%, t=3.483; fifth and sixth columns; 100% of E18.5 whole lysate, E18.5 membrane=34.324±12.121%, P0 membrane=27.46±9.411%, t=0.2974. Student’s t-test). Each independent experiment included six embryos per group. NS, not significant.

Figure 5

Cytoplasmic translocation of Ryk facilitates cytoplasmic translocation of Vangl2. (a) Schematic diagram of Ryk-FLAG and Vanlg2-HA expression constructs. (b–d) In vitro analysis of the distribution of Ryk and Vangl2 in response to Wnt5a gradient. Vangl2-HA and Ryk-FLAG were transfected into HEK293T cells. Vangl2-HA- and Ryk-FLAG-expressing cells were separately treated with BSA (b), 100 ng ml⁻¹ wnt5a (c) or 200 ng ml⁻¹ Wnt5a (d). DAPI staining (blue) labeled the nucleus. HA (red) and FLAG (green) staining show the expressed Vangl2 and Ryk, respectively. Scale bar, 5 μm. (e, g) Western blot analysis of whole-cell lysate, cytoplasm and membrane extracts from Vangl2-HA-expressing HEK293T cells with (g) or without (e) Ryk-FLAG transfection. Antibodies against GAPDH and E-cadherin were used to confirm the purity of the cytoplasm and the membrane fractions, respectively. (f, h) Quantification of the percentage of Vangl2 in the different fractions in e, g, respectively (f, BSA and 100 ng ml⁻¹, P=0.9970; BSA and 200 ng ml⁻¹, P=0.8703; 100 and 200 ng ml⁻¹, P=0.8345; h, BSA and 100 ng ml⁻¹, P=0.6743). Data are represented as the mean±s.e.m. of at least three independent experiments per group. ***P<0.001, (one-way ANOVA). NS, not significant.

Figure 6

Disruption of Ryk translocation leads to reduction in cytoplasmic distribution of Vangl2. (a) Schematic diagram of the Ryk-FLAG and RykΔPDZ-FLAG constructs used in co-immunoprecipitation experiments. (b) Co-immunoprecipitation (Co-IP) of Ryk and Vangl2 from lysates of HEK293T cells transfected with Ryk, Vanlg2, Ryk and Vangl2, RykΔPDZ and Vangl2, or Ryk and lp. (c) Quantification of co-immunoprecipitation ratio in each group. The ratio was calculated and normalized to Ryk and Vangl2 group. Data are represented as the mean±s.e.m. of three independent experiments per group. ***P<0.001 (immunoprecipitated Ryk, F=209; immunoprecipitated Vangl2, F=672, one-way ANOVA). (d–f) Vangl2 distribution when co-expressed with RykΔPDZ. Vangl2-HA- and RykΔPDZ-FLAG-expressing cells were treated with BSA (d), 100 ng ml⁻¹ wnt5a (e) or 200 ng ml⁻¹ Wnt5a (f). DAPI staining (blue) labeled the nucleus. HA (red) and FLAG (green) staining show the expressed Vangl2 and RykΔPDZ, respectively. Scale bar, 5 μm. (g, i) Western blot analysis of whole-cell lysate, cytoplasm and membrane extracts from Vangl2-HA and Ryk-FLAG (g), or Vangl2-HA and RykΔPDZ-FLAG (i) co-expressing HEK293T cells. Cells were treated with BSA, 100 ng ml⁻¹ Wnt5a or 200 ng ml⁻¹ Wnt5a. (k) Western blot analysis of whole-cell lysate, cytoplasm and membrane extracts from Vangl2-HA and Ryk-FLAG co-expressing HEK293T cells treated with DAPT. Antibodies against GAPDH and E-cadherin were used to confirm the purity of the cytoplasm and the membrane fractions, respectively. (h, j, l) Quantification of the percentage of Vangl2 in the different fractions in (g, i, k), respectively (h, BSA and 200, 100 and 200 ng ml⁻¹, P<0.0001; j, BSA and 100 ng ml⁻¹, P=0.8402; BSA and 200 ng ml⁻¹, P=0.7311; 100 and 200 ng ml⁻¹, P=0.4239; l, BSA and 100 ng ml⁻¹, P=0.9841; BSA and 200 ng ml⁻¹, P=0.9920; 100 and 200 ng ml⁻¹, P=0.9986). Data are represented as the mean±s.e.m. of at least three independent experiments per group. ***P<0.001, (one-way ANOVA). PTK, protein tyrosine kinase domain; NS, not significant; TMD, transmembrane domain; WIF, Wnt inhibitory factor motif.

Figure 7

Cytoplasmic translocation of Ryk and Vangl2 are asymmetrically in growth cones of corticospinal axons under Wnt5a gradient. (a) Ryk and Vangl2 transport into corticospinal neurons in response to Wnt5a. E18.5 corticospinal neurons were co-transfected with Vangl2-HA and Myc-Ryk-FLAG. HA (red), Myc (green) or FLAG (green) antibodies were applied to label Vangl2-HA, Ryk ECD or Ryk ICD, respectively. Scale bar, 5 μm. (b) Quantification of expression ratios of Ryk ICD, Ryk ECD and Vangl2 in neurons. The expression intensity of Ryk ICD, Ryk ECD or Vangl2 in axons was measured and normalized to expression intensity of Ryk ICD, Ryk ECD or Vangl2 in entire neurons. Data are represented as the mean±s.e.m. of four independent experiments, each of which included at least 30 neurons per group. *P<0.05, **P<0.01, ***P<0.001 (Left, BSA and 100 ng ml⁻¹, P=0.0480; BSA and 200 ng ml⁻¹, P=0.0202; 100 and 200 ng ml⁻¹, P=0.0011; BSA and 400 ng ml⁻¹, P<0.0001; middle, BSA and 100 ng ml⁻¹, P=0.0013; BSA and 200 ng ml⁻¹, P<0.0001; 100 and 200 ng ml⁻¹, P<0.0001; BSA and 400 ng ml⁻¹, P<0.0001; right, BSA and 100 ng ml⁻¹, P=0.0019; BSA and 200 ng ml⁻¹, P<0.0001; 100 and 200 ng ml⁻¹, P<0.0001; BSA and 400 ng ml⁻¹, P<0.0001. One-way ANOVA). (c) Response of growth cones of corticospinal neurons co-transfected with Ryk and Vangl2 to Wnt5a and BSA gradients. Growth cones perpendicular to the gradient were selected, and divided into proximal and distal sides relative to the gradient. Immunostaining with HA (red) and FLAG (green) antibodies labeled Vangl2 and Ryk in the growth cones. (d) Quantification of the proximal/total (P/T) and distal/total (D/T) ratio of Ryk and Vangl2 intensity. Data are represented as the mean±s.e.m. of four independent experiments, with each experiment including at least 18 growth cones per group. ***P<0.001 (BSA, P=0.9653; Wnt5a, P<0.0001) and Vangl2 (d; BSA, P=0.9998; Wnt5a, P<0.0001. One-way ANOVA). (e) Response of growth cones of corticospinal neurons transfected with RykΔPDZ and Vangl2 to Wnt5a and BSA gradients. Immunostaining with HA (red) and FLAG (green) antibodies labeled Vangl2 and RykΔPDZ in the growth cones, respectively. (f) Response of Ryk- and Vangl2-expressing growth cones of dissociated corticospinal neurons to Wnt5a with either DAPT or DMSO treatment (1 μm). Immunostaining with HA (red) and FLAG (green) antibodies labeled Vangl2 and Ryk in the growth cones, respectively. (g) Quantification of growth cone-turning ratio in different groups. Data are represented as the mean±s.e.m. of three independent experiments. **P<0.01, ***P<0.001 (Ryk=81.4±6.56%, RykΔPDZ=15.1±4.19%, P=0.001. Ryk+DMSO=87.3±6.39%, Ryk+DAPT=8.9±4.85%, P<0.0001. One-way ANOVA). Each independent experiment included at least 20 neurons per group. (h, i) Response of growth cones of corticospinal neurons to Wnt5a or BSA gradient when transfected with Fzd3-FLAG. Immunostaining with FLAG (red) and α-adaptin (AP-2; green) antibodies labeled intracellular Fzd3 and α-adaptin, respectively. (j, k) Response of growth cones of corticospinal neurons to Wnt5a or BSA gradient when transfected with Vangl2-HA and Ryk-FLAG. Immunostaining with HA (green), FLAG (red) and α-adaptin (AP-2; blue) antibodies labeled Vangl2, intracellular Ryk and α-adaptin, respectively. The insets show the high-magnification images of boxed areas in (c, e, f, h–k), respectively. Scale bar, 2 μm (c, e, f, h–k). NS, not significant.

Figure 8

Synchronous translocation of Ryk and Vangl2 to the cytoplasm is required for Wnt repulsion of CST in the spinal cord. (a, b) The expression pattern of Ryk in coronal sections of cortex tissues injected with DMSO or DAPT. Anti-Ryk labeled Ryk ICD (red), anti-Ctip2 labeled corticospinal neurons (blue) and anti-E-cadherin (E-cad) labeled the cell membrane (green). Scale bar, 20 μm. (a1, a2, b1, b2) Higher magnification images of the boxed areas in a, b. Scale bar, 5 μm. (c) Quantification of the ratio of cytoplasmic Ryk to total Ryk. The expression intensity of Ryk in the cytoplasm was normalized to the expression intensity of Ryk in the entire neuron. Data are represented as the mean±s.e.m. of data from 10 mice in each group. At least 15 neurons were measured in each mouse. ***P<0.001 (DAPT=0.1520±0.0076%, DMSO=0.5235±0.0452%, t=8.100, P<0.0001. Student’s t-test). (d–h) CST axonal pathfinding in spinal cord of mice with cortex injection of DMSO or DAPT visualized with EGFP immunostaining. CST axons in sagittal sections of spinal cord with DMSO (d) or DAPT (g) injection. The double green arrows indicate control EGFP-expressing CST continuing to descend at the third thoracic (T3) segment. The red arrow indicates where DAPT-treated CST aberrantly terminates at the third cervical (C3) segment. (e, f, h) Higher magnification images of the boxed areas in d, g, respectively. Scale bar, 500 or 100 μm (higher magnification images). (i, j) Quantification of CST area and CST bundle length in each group. Data are represented as the mean±s.e.m. *P<0.05, ***P<0.001. Data of CST bundle length and CST area were analyzed from eight mice in each group, using two-way ANOVA (i, C1, P>0.9999; C2, P=0.7363; C3, P=0.3244; C4, P=0.0004; C5–T2, P<0.0001; T3, P=0.0369) or Student’s t-test (j, 100% of DMSO, DAPT=23.9±7.80, t=9.75, P<0.0001), respectively. (k–o) CST axonal pathfinding in spinal cord of mice with control lentivirus or RykΔPDZ-overexpression lentivirus injected into the cortex. CST axons were identified with EGFP expression (green). The double green arrows indicate control EGFP-expressing CST continuing to descend at the fifth thoracic (T5) segment. The red arrow indicates where RykΔPDZ-expressing CST present random extension at the second cervical (C2) segment. The double red arrow indicates where RykΔPDZ-expressing CST aberrantly ends at the third cervical (C3) segment. (l, m, o) Higher magnification images of the boxed areas in k, n, respectively. Scale bar, 500 or 100 μm (higher magnification images). (p, q) Quantification of CST area and CST bundle length in each group. Data are represented as the mean±s.e.m. *P<0.05, ***P<0.001. Data of CST area and CST bundle length were analyzed from five mice in each group, using two-way ANOVA (p, C1, P=0.9995; C2, P=0.0939; C3, P=0.0357; C4–T4, P<0.0001; T5, P=0.0002) or Student’s t-test (q, 100% of control, RykΔPDZ=14.7±6.41, t=13.3, P<0.0001), respectively.