Endosomal trafficking defects alter neural progenitor proliferation and cause microcephaly

Abstract

Primary microcephaly and megalencephaly are severe brain malformations defined by reduced and increased brain size, respectively. Whether these two pathologies arise from related alterations at the molecular level is unclear. Microcephaly has been largely associated with centrosomal defects, leading to cell death. Here, we investigate the consequences of WDR81 loss of function, which causes severe microcephaly in patients. We show that WDR81 regulates endosomal trafficking of EGFR and that loss of function leads to reduced MAP kinase pathway activation. Mouse radial glial progenitor cells knocked-out for WDR81 exhibit reduced proliferation rate, subsequently leading to reduced brain size. These proliferation defects are rescued in vivo by expressing a megalencephaly-causing mutant form of Cyclin D2. Our results identify the endosomal machinery as an important regulator of proliferation rates and brain growth, demonstrating that microcephaly and megalencephaly can be caused by opposite effects on the proliferation rate of radial glial progenitors.

Fig. 1. WDR81 KO mice display reduced brain size and altered neuronal positioning.

a Schematic representation of mouse WDR81 isoforms and predicted structure. b Quantification of WDR81 isoforms 1 and 2 mRNA levels in WT (p < 0.0001) and WDR81^−/− (p = 0.0002) E14.5 cortices (n = 3 independent brains for each genotype). Isoform 1 (ISO1) is the dominant isoform and its levels are strongly reduced in WDR81^−/− cortices. c WDR81^−/− postnatal day 7 brains are microcephalic and display reduced cortical surface area as compared to WT brains. d Quantification of hemisphere area at P7 in WT and WDR81 KO1 brains (p < 0.0001) (n = 4 independent brains for each genotype). e DAPI staining of P7 WT and WDR81^−/− cross sections reveals reduced cortical thickness in mutants. f Quantification of cortical thickness in WT, KO1 and KO2 brains at P0 and P7 (At P0, WT vs KO1 p = 0.0112; WT vs KO2 p = 0.0158. At P7, WT vs KO1 p < 0.0001; WT vs KO1 p < 0.0001) (n = 4 independent brains for each genotype and stage). g NeuN staining of WT and WDR81^−/− cortical plates (CP) at P0. h Quantification of NEUN+ cells in WT and WDR81 KO1 cortical plates at P0 in 600 × 300 μm crops reveals reduced number of neurons at birth (p = 0.0002) (n = 3 independent brains for each genotype). i CUX1 staining in P7 WT and WDR81^−/− cortices. Quantification of CUX1+ neuronal positioning reveals dispersion throughout the thickness of the neocortex (Bin 5, p = 0.0003) (n = 5 independent brains for each genotype). j CTIP2 staining in P7 WT and WDR81^−/− cortices. Quantification does not neuronal positioning defects, with CTIP2+ neurons still concentrated in the third bin (n = 5 independent brains for each genotype). i, j WT and mutant cortices were divided into 5 bins of equal size to measure neuronal relative positioning, independently of cortical thickness. All data are expressed as mean ± standard deviation (SD). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by two-tailed unpaired t tests.

Fig. 2. WDR81 KO alters radial glial progenitor proliferation.

a PAX6 and PH3 double staining in E14.5 WT and WDR81^−/− brains. Quantification of the mitotic index of PAX6+ cells reveals decreased proliferation of WDR81^−/− radial glial progenitors at E14.5 (p = 0.0005) and E16.5 (p = 0.0003) (n = 5–8 independent brains for each genotype and stage). b TBR2 and PH3 double staining in E14.5 WT and WDR81^−/− brains. Quantification of the mitotic index of TBR2+ cells indicates that proliferation of WDR81^−/− intermediate progenitors in not affected (n = 3 independent brains for each genotype and stage). c Schematic representation of the BrdU labeling experimental approach and Pax6 and BrdU staining in E14.5 WT and WDR81^−/− brains. d Quantification of the percentage of BrdU+ PAX6+ out of total PAX6 cells reveals increased number of cells in S phase in WDR81^−/− radial glial progenitors at E14.5 (p = 0.0007) (n = 3 independent brains for each genotype). e Schematic representation of the BrdU/EdU double labeling experimental approach and PAX6, BrdU and EdU staining in E14.5 WT and WDR81^−/− brains. Arrowheads indicate EdU+ BrdU+ cells. f Quantification of the percentage of of BrdU+ EdU− PAX6+ out of the total BrdU+ PAX6+ cells reveals a decreased proportion of cells that exited S phase following BrdU injection in WDR81^−/− radial glial progenitors at E14.5 (p = 0.0174) (n = 3 independent brains for each genotype). g Staining for the cell fate markers PAX6 (radial glial progenitors), TBR2 (Intermediate progenitors) NEUN (Neurons) in E14.5 WT and WDR81^−/− brains, and quantification of cell fate distribution at E12.5, E14.5 (p = 0.071 for NeuN) and E16.5 (p = 0.0005 for NeuN). (n = 3–5 independent brains for each genotype, staining and stage) h Staining for Cleaved Caspase-3 (CC3) and DAPI in E14.5 WT and WDR81^−/− brains, showing an absence of apoptosis induction. All data are expressed as mean ± standard deviation (SD). *p < 0.05; **p < 0.01; ***p < 0.001 by two-tailed unpaired t tests.

Fig. 3. Reduced proliferation and EFGR signaling in WDR81 patient cells.

a PH3 and DAPI staining in control and WDR81 patient fibroblasts. b Quantification of the percentage of PH3+ cells reveals decreased mitotic index in patient cells (control1-patient1 p = 0.0051; control1-patient2 p = 0.0043; control2-patient1 p = 0.0015; control2-patient2 p = 0.0016) (n = 3 independent experiments). c Ki67 and DAPI staining in control and WDR81 patient fibroblasts. d Quantification of the percentage of Ki67+ cells shows decreased proliferation in patient cells (control1-patient1 p = 0.0207; control1-patient2 p = 0.0476; control2-patient1 p = 0.0245; control2-patient2 p = 0.0473) (n = 3 independent experiments). e Western Blot for EGFR in control and WDR81 patient fibroblasts (n = 3 independent experiments). f Quantification reveals a strong reduction of EGFR levels in patient cells (control1-patient1 p = 0.0017; control1-patient2 p = 0.0004) (n = 5 independent experiments). g Time course of EGFR and P-ERK levels in control1 and WDR81 patient1 fibroblasts following an EGF pulse. h Quantification of EGFR levels, normalized to control levels at T0 in control1 and patient1 (T0 p = p < 0.0001; T5 p = p < 0.0001; T15 p = 0.0005; T30 p < 0.0001; T60 p = 0.0075) (n = 5 independent experiments). i Quantification of P-ERK levels, normalized to control levels at T5 in control1 and patient1 (T5 p < 0.0001; T30 p < 0.0001) (n = 5 independent experiments). j Western Blot for EGFR in control and WDR81 patient fibroblasts at steady state (+EGF) and cultivated for 24H in the absence of EGF. k Quantification reveals a restoration of EGFR levels following starvation (at steady state, control1-patient1 p = 0.0161; control1-patient2 p = 0.0368; control2-patient1 p = 0.0075; control2-patient2 p = 0.0176) (n = 4 independent experiments). All data are expressed as mean ± standard deviation (SD). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by two-tailed unpaired t tests.

Fig. 4. WDR81 is required for endosomal trafficking of EGFR.

a Expression of control shRNA and shRNA-mediated knockdown constructs for WDR81, WDR91 and p62, together with pCAG-GFP. Plasmids were delivered by in utero electroporation at E13.5 and analysis was performed at E16.5. Ventricular Zone and Sub-Ventricular Zone (VZ + SVZ), Intermediate Zone (IZ) and Cortical Plate (CP) were identified based on DAPI staining. b Quantification of electroporated cell distribution reveals major accumulation in the IZ following WDR81 and WDR91 knockdown (VZ + SVZ, WDR91 shRNA p = 0.0025; IZ, WDR81 shRNA p = 0.0001, WDR91 shRNA p = 0.067; CP, WDR81 shRNA p = 0.0003, WDR91 shRNA p = 0.0234) (n = 3–4 independent brains per condition). c Ventricular zone of E14.5 WT and WDR81^−/− mice cortices stained for EEA1 and Actin. d Control and WDR81 patient fibroblasts stained for EEA1. e Quantification of individual EEA1+ early endosomes in WT and WDR81^−/− VZ in 200 × 100 μm crops reveals increased size in mutant brains (p < 0.0001) (n = 3 independent experiments). f Quantification of individual EEA1+ early endosomes in control and WDR81 patient fibroblasts reveals increased size in mutant cells (For all control-patient couples p < 0.0001) (n = 3 independent experiments). g EGF⁵⁵⁵ uptake assay in control and WDR81 patient fibroblasts, and stained for EEA1. h Quantification of EGF⁵⁵⁵ and EEA1 colocalization during EGF⁵⁵⁵ uptake reveals prolonged colocalization between EGF and early endosomes in WDR81 patient cells (At T15, C1 vs P2 p = 0.0158; C2 vs P1 p < 0.0001; C2 vs P2 p < 0.0001. At T30, C2 vs P1 p = 0.0149, C2 vs P2 p = 0.0033. At T60, C1 vs P1 p < 0.0001; C1 vs P2 p = 0.001; C2 vs P1 p = 0.0028. At T120, C1 vs P1 p = 0.0008; C1 vs P2 p = 0.0011; C2 vs P1 p = 0.0072; C2 vs P2 p = 0.0148. At T360, C1 vs P1 p = 0.0349; C2 vs P1 p < 0.0001. C2 vs P2 p = 0.0001) (n = 7 independent experiments). All experiments were performed at least three independent times. All data are expressed as mean ± standard deviation (SD). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by two-tailed unpaired t test (A & H) and Mann–Whitney tests (E & F).

Fig. 5. Undegradable Cyclin D2^Thr280Ala rescues WDR81^−/− proliferation index.

a Expression of Cyclin D2, Cyclin D2^Thr280Ala and empty vector in WT and WDR81^−/− brains. Constructs were in utero electroporated at E14.5, and brains were fixed at 16.5 and stained for PAX6 and PH3. b Quantification of the percentage of mitotic (PH3+) electroporated radial glial cells (PAX6+) out of total electroporated radial glial cells reveals rescue of mitotic index in WDR81^−/− cells expressing Cyclin D2^Thr280Ala (WT-empty vector vs WT-Cyclin D2^Thr280Ala p = 0.0382; WT-empty vector vs WDR81^−/− empty vector p = 0.0247; WDR81^−/− empty vector vs WDR81^−/− Cyclin D2 p = 0.031; WDR81^−/− empty vector vs WDR81^−/− Cyclin D2^Thr280Ala p = 0.0023) (n = 3 independent brains per genotype and condition). c Model. WDR81 loss of function leads to reduced activation of the MAPK signaling pathway downstream of EGFR, to reduced radial glial progenitor proliferation, and to microcephaly. Gain of function in the Pi3K-AKT pathway or stabilizing mutations in Cyclin D2 lead to increased radial glial progenitor proliferation, and to megalencephaly. Cyclin D2 mutants can rescue proliferation defects in WDR81^−/− brains, indicating that these two pathologies can arise from opposite effects on the proliferation rates of radial glial progenitor. All data are expressed as mean ± standard deviation (SD). *p < 0.05; **p < 0.01 by two-tailed unpaired t tests.