Wdr4 promotes cerebellar development and locomotion through Arhgap17-mediated Rac1 activation

Abstract

Patients with mutations of WDR4, a substrate adaptor of the CUL4 E3 ligase complex, develop cerebellar atrophy and gait phenotypes. However, the underlying mechanisms remain unexplored. Here, we identify a crucial role of Wdr4 in cerebellar development. Wdr4 deficiency in granule neuron progenitors (GNPs) not only reduces foliation and the sizes of external and internal granular layers but also compromises Purkinje neuron organization and the size of the molecular layer, leading to locomotion defects. Mechanistically, Wdr4 supports the proliferation of GNPs by preventing their cell cycle exit. This effect is mediated by Wdr4-induced ubiquitination and degradation of Arhgap17, thereby activating Rac1 to facilitate cell cycle progression. Disease-associated Wdr4 variants, however, cannot provide GNP cell cycle maintenance. Our study identifies Wdr4 as a previously unappreciated participant in cerebellar development and locomotion, providing potential insights into treatment strategies for diseases with WDR4 mutations, such as primordial dwarfism and Galloway-Mowat syndrome.

Fig. 1. Nervous system-specific knockout of Wdr4 impairs cerebellar development.

A–F Representative Nissl staining images (A, D) and quantitative data (B, C, E, F) for cerebellar size, foliation, and EGL size in P0 (A–C) and P7 (D–F) Wdr4 N-cKO and control mice. See also Fig. S1 and Video S1. G–K Representative confocal images (G) and quantitative data (H–K) showing cerebellum size (G), Calbindin⁺ Purkinje neurons (G–I), Ki67⁺ proliferating GNPs (G, J), and TUNEL⁺ apoptotic cells (G, K) in the P7 Wdr4 N-cKO and control cerebella. The most left panel in (G) was stitched with several images to show the whole cerebellum. The right panels in (G) are higher magnification images of the boxed regions in the left panels. Arrows indicate TUNEL⁺ apoptotic cells. The data in (H–I) were from the whole cerebellar section, and in (J–K) were from the EGL of lobule V–VI. Scale bars are 150 μm in (A), 250 μm in (D), (G, left), and 50 μm in (G, right). Data were from 3 (B, C, E, F) or 4 (H–K) cerebella in each group and analyzed using two-tailed unpaired Student’s t-test without Welch’s correction (equal variances, B, C, E, F, H, J, K) or with Welch’s correction (unequal variances, I), p = 0.0214 in (B), 0.0429 in (C), 0.0012 in (E), 0.0007 in (F), <0.0001 in (H), 0.0338 in (I), <0.0001 in (J), and 0.4951 in (K). Data are represented as individual points and mean; *p < 0.05, **p < 0.005, ***p < 0.0005; n.s., non-significant.

Fig. 2. Wdr4 ablation impairs GNP proliferation in a cell-autonomous manner.

A–P Representative confocal images and quantitative data from the whole cerebellar section (A, B, E, F, I, J, M, N), and the EGL of lobule V–VI (C, D, G, H, K, L, O, P) showing EGL size and Ki67⁺ proliferating GNPs in the indicated Wdr4 conditional knockout and control mice at P7 (scale bars, 250 μm in the left and 50 μm in the right). Data were from 3 cerebella in each group and analyzed using a two-tailed unpaired Student’s t-test without Welch’s correction (equal variances, B, D, F, H, J, L, P) or with Welch’s correction (unequal variances, N), p = 0.0048 in (B), 0.0300 in (D), 0.0174 in (F), 0.0437 in (H), 0.9848 in (J), 0.6309 in (L), 0.2248 in (N), and 0.5906 in (P). Data are represented as individual points and mean; *p < 0.05, **p < 0.005; n.s., non-significant. See also Figs. S2–4.

Fig. 3. Wdr4 deficiency in cerebellar GNPs accelerates cell cycle exit.

A–F Representative confocal images (A, C, E) and quantitative data (B, D, F) from lobule V for examining GNP cell cycle exit in the Wdr4 A-cKO and control cerebella labeled by EdU at P1, P2, or P3 and harvested 2 days later for staining. The Ki67⁺ proliferating GNPs mainly located in the outer EGL (oEGL), while the cell cycle exit population (EdU⁺Ki67^-) mostly located in the inner EGL (iEGL), ML, and IGL. G–H Representative confocal images (G) and quantitative data (H) from lobule V for measuring GNP differentiation into NeuN⁺ granule neurons in the Wdr4 A-cKO and control cerebella labeled by EdU at P2 and harvested 2 days later for staining. I Quantitative data measuring proliferation in GNPs isolated from P7 Wdr4 A-cKO and control mice, cultured for 3 days. See also Fig. S5. J–K Representative confocal images (J) and quantitative data (K) for measuring cell cycle exit in P7 Wdr4 A-cKO; Ai14 and control GNP cultures stained at DIV3. L–M Representative confocal images (L) and quantitative data (M) for measuring differentiation into NeuN⁺ granule neurons in P7 Wdr4 A-cKO; Ai14 GNP cultures and controls stained at DIV3. Scale bars are 50 μm in (A, C, E, G) and 25 μm in (J, L). Data were from 3 (A–H, L–M) or 4 (I–K) cerebella in each group and analyzed using two-tailed unpaired Student’s t-test without Welch’s correction (equal variances), p = 0.3160 in (B), 0.0028 in (D), 0.0219 in (F), 0.0028 in (H), 0.0006 in (I), 0.0013 in (K), and 0.0078 in (M). Data are represented as individual points and mean, or mean ± S.E.M.; *p < 0.05, **p < 0.005; n.s., non-significant.

Fig. 4. Wdr4 deficiency in cerebellar GNPs leads to IGL and ML size reduction and Purkinje neuron disorganization.

A–F Representative confocal images (A) and quantitative data (B–F) showing the sizes of cerebellum (A, B), IGL (A, C), ML (A, D), and the number (A, E) as well as the density (A, F) of Purkinje neurons, in the P30 Wdr4 A-cKO and control mice. The upper-left panel in (A) was stitched with several images to display the whole cerebellum. The lower panels in (A) are higher magnification images of the boxed regions in the upper panels, showing disorganized Purkinje neurons in P30 Wdr4 A-cKO cerebella (scale bars, 500 μm in the upper panels and 50 μm in the lower panels). Data were from 3 cerebella in each group and analyzed using two-tailed unpaired Student’s t-test without Welch’s correction (equal variances), p = 0.0039 in (B), < 0.0001 in (C), 0.0027 in (D), 0.4276 in (E), and 0.0002 in (F). Data are represented as individual points and mean; **p < 0.005, ***p < 0.0005; n.s., non-significant.

Fig. 5. Wdr4 ablation in cerebellar GNPs impairs locomotion and coordination.

A–C Representative moving traces (A) and quantitative data (B, C) for the moving distance (B) and time spent in the inner zone (C) in an open field test for Wdr4 A-cKO and control mice. D Quantitative data for latency to fall in rotarod test for Wdr4 A-cKO and control mice. E Quantitative data for grip strength in Wdr4 A-cKO and control mice. F Quantitative data for foot faults in grid-walking test for Wdr4 A-cKO and control mice. G–Q Quantitative data for walking speed (G), step length (K), stride length (L), stance phase time (I), double support time (J), base of support (M), foot angle (N), print length (O), intermediary toe spread (P), and toe spread (Q) in gait analyses for Wdr4 A-cKO and control mice. R Schematic representation of the parameters measured in gait analysis. Data were from 10 adult male mice (2–3 months old) in each group. Data following a normal distribution were analyzed using two-tailed unpaired Student’s t-test with Welch’s correction (unequal variances, I, J, M, O) or without Welch’s correction (equal variances, B–G, K, L, N, P), depending on the F-test results. Data not following a normal distribution were analyzed using Mann–Whitney test (H, Q). p = 0.0058 in (B), 0.0215 in (C), 0.0232 in (D), 0.2493 in (E), <0.0001 in (F), < 0.0001 in (G), 0.0605 in (H), < 0.0001 in (I), < 0.0001 in (J), 0.0008 in (K), 0.0007 in (L), 0.0003 in (M), 0.0018 in (N), 0.0005 in (O), 0.0003 in (P), and 0.0007 in (Q). Data are represented as individual points and mean ± S.E.M.; *p < 0.05, **p < 0.005, ***p < 0.0005; n.s., non-significant. See also Videos S2A and S2B.

Fig. 6. Wdr4 promotes the ubiquitination and degradation of Arhgap17 to activate Rac1.

A The Volcano plot of proteins identified by LC-MS/MS from GNPs isolated from the P7 Wdr4 A-cKO; Ai14 and control cerebella. Differentially expressed proteins (p < 0.05 and fold change > 1.5) are marked in blue (for up-regulated) and green (for down-regulated). Among the upregulated proteins, those with a known anti-proliferative function are marked in red, orange or yellow with their names. Data were from 3 cerebella in each group and were analyzed using pairwise-ratio-based, Student’s t-test by the Proteome Discoverer 2.3 software. B, C Immunoprecipitation analysis for the interaction between exogenous (B) or endogenous (C) Wdr4 with indicated proteins. D–E Western blot (D) and RT-qPCR (E) analyses for the expression of Arhgap17 protein or Arhgap17 mRNA in N2a cells expressing control or Wdr4 shRNAs. Data in (E) were from four repeats in each group and analyzed using one-way ANOVA post hoc Dunnett’s test, p = 0.3302 in shWdr4#1 v.s. shCtrl, and 0.1855 in shWdr4#2 v.s. shCtrl. Data are represented as individual points and mean. F Western blot analysis for Arhgap17 and Gli2 expression in the purified GNPs from the P7 Wdr4 A-cKO and control cerebella. G, H Western blot analysis for Arhgap17 expression in N2a cells stably expressing Wdr4 (G) or Wdr4 shRNAs (H) and treated with 1 μM MG132 for 16 h. I Western blot analysis for Arhgap17 expression in N2a cells stably expressing Wdr4 and treated with 100 μg/ml CHX for indicated time points. The levels of Arhgap17 were normalized to the 0 h time point in the control or Wdr4-expressing group, respectively, and indicated at the bottom. J, K In vivo ubiquitination assay using N2a cells (J) or N2a cells expressing Wdr4 shRNAs (K) and transfected with indicated constructs. L Rac1 activity assay using N2a cells expressing Wdr4 shRNAs or transfected with Flag-Arhgap17. The western blot results were quantified using ImageJ software. The protein levels were normalized first to the Gapdh protein level in each group and then to the corresponding control groups, and expressed as fold changes at the bottom. All western blot analyses were done at least twice.

Fig. 7. Arhgap17 knockdown or Rac1 activation rescues the GNP proliferation defect caused by Wdr4 deletion.

A, D, I Quantitative data measuring proliferation in GNPs isolated from P7 Wdr4 A-cKO mice and infected with lentiviruses carrying GFP and shRNA against Arhgap17 or control (A), treated with Rac1 activator ML099 or DMSO (D), or infected with lentiviruses carrying GFP and Wdr4 WT or mutants (I) at DIV 0. The capability of Arhgap17 shRNAs to knockdown the expression of Arhgap17 is shown in Fig. S7. The total cell numbers at DIV 0 and DIV 3 were counted. Data were from 3 (A, I) or 6 (D) cerebella in each group and analyzed using one-way ANOVA post hoc Holm-Šídák’s test (A), two-tailed unpaired Student’s t-test with Welch’s correction (unequal variances, D), or one-way ANOVA post hoc Dunnett’s test (I). p = 0.012 in shArhgap17#1 v.s. shCtrl, and 0.0376 in shArhgap17#2 v.s. shCtrl (A), p = 0.0082 in (D), p = 0.0001 in Wdr4 WT, >0.9999 in Wdr4 D166A, and 0.5937 in Wdr4 R172Q, compared to the control group (I). B, C, J, K Representative confocal images (B, J) and quantitative data (C, K) for measuring cell cycle exit in P7 Wdr4 A-cKO GNP cultures with Arhgap17 knockdown (B, C) or overexpression of WT or mutant Wdr4 (J, K). The GNPs were isolated, treated with EdU for 2 h, infected with lentiviruses carrying GFP and shRNA or cDNA at DIV 0, and harvested for staining analysis at DIV 3. Data were from 3 cerebella in each group and analyzed using one-way ANOVA post hoc Dunnett’s test. p = 0.0055 in shArhgap17#1 v.s. shCtrl, and 0.0256 in shArhgap17#2 v.s. shCtrl (C), p < 0.0001 in Wdr4 WT, >0.9999 in Wdr4 D166A, and 0.3568 in Wdr4 R172, compared to the control group (K). E, F Representative confocal images (E) and quantitative data (F) for measuring cell cycle exit in P7 Wdr4 A-cKO; Ai14 GNP cultures treated with ML099 or DMSO at DIV 0 and harvested 3 days later. The GNP cultures were treated with EdU for 2 h before ML099 treatment. Data were from 3 cerebella in each group and analyzed using two-tailed unpaired Student’s t-test without Welch’s correction (equal variances), p = 0.0052. G, H Representative confocal images (G) and quantitative data (H) from the EGL of lobule V–VI measuring the number of Ki67⁺ proliferating GNPs in the P7 Wdr4 A-cKO or control cerebella. The Wdr4 A-cKO and control mice were injected once with DMSO or ML099 (20 mg/kg) at P3, and harvested 4 days later. Data were from 5 cerebella in each group and analyzed using one-way ANOVA post hoc Turkey’s test, p < 0.0001 in (H, A-cKO v.s. Ctrl with DMSO), and =0.0439 in (H, ML099 v.s. DMSO in A-cKO). See also Fig. S8. Scale bars are 25 μm in (B, E, J) and 50 μm in (G). Data are represented as individual points and mean, or mean ± S.E.M.; *p < 0.05, ***p < 0.0005; n.s., non-significant. L Western blot analysis for Wdr4 expression in GNPs infected with lentiviruses carrying Wdr4 WT or mutants at DIV 0 and harvested at DIV 3, showing comparable expression levels among WT and mutants. M Western blot analysis for Arhgap17 expression in N2a cells overexpressing Wdr4 WT or mutants. The western blot results were quantified using ImageJ software. The protein levels were normalized first to the Gapdh protein level in each group and then to the corresponding control groups, and expressed as fold changes at the bottom. All western blot analyses were done at least twice.