PMC

Disrupted radial glial development in Gsk3a^–/–;Gsk3b^lox/lox;Nestin-Cre cerebral cortex. (A-D) Inactivation of both Gsk3α and β in radial glial progenitors disrupts radial glial scaffolding. E16 cerebral cortices from control (A,C) and Gsk3-deficient (B,D) mice were labeled with radial glia-specific RC2 antibodies and deeper layer neuron-specific anti-Tbr1 antibodies. Instead of the polarized radial organization of the normally developing cerebral cortex (A,C), Gsk3α- and β-deficient radial glia are wavy (arrow and inset, D), shorter, and often have `bead'-like thickenings (D, arrowhead). Labeling with Tbr1 antibodies indicates a diffuse placement of neurons within the Gsk3α- and β-deficient cortex. Scale bar: 200 μm in A,B; 100 μm in C,D.

Dynamics of radial glial spines and growth cones. (A,B) All along the radial processes, spine-like protrusions are evident (arrowheads). (C-E) Time-lapse observations indicate dynamic activity and rearrangements of these protrusions (see changes in protrusions adjacent to arrowhead). Time elapsed is indicated in minutes. (F-I) Radial glial endfeet are dynamic and vary in shape from club-like (F,G) to highly branched (H,I). (J) As cortical development proceeds, proportionately more radial glia exhibit club-like endfeet. Data shown are mean±s.e.m. ^*P<0.001 (Student's t-test). Also see Movies 1-4 in the supplementary material. Scale bar: 100 μm in A; 28 μm in B-F; 66 μm in G-I.

Expression of Cdc42 in the leading edges of polarized radial glia. Mouse E15 cerebral cortex was electroporated with Cdc42-GFP or GFP plasmids, sliced, and imaged within 8 hours. (A) Low-magnification view of an entire radial glial cell indicates preferential localization of Cdc42-GFP at the leading tip (asterisk). Arrow, cell soma; arrowheads, polarized process. By contrast, control GFP is localized uniformly throughout radial glia (B). (C) Higher magnification view of the radial process shows Cdc42 accumulation at the leading tip (asterisk). (D) Double labeling with radial glia-specific RC2 antibody (red) confirms preferential accumulation of Cdc42 (green) in radial process tips. Also see Movies 11-14 in the supplementary material. Scale bar: 35 μm in A,B; 90 μm in C,D.

Assay for radial glial development in mouse embryonic cerebral cortex. (A,B) Large cohorts of radial glial cells (green) spanning the entire width of the developing embryonic cerebral cortices were labeled with GFP by electroporation with pBLBP promoter-GFP plasmids. Labeling of multiple, definable, adjacent radial glia (arrowheads, B) is essential for the examination of inter-radial glial interactions and developmental potential of these cells. (C-E) Higher magnification views of labeled radial glial cells, illustrating the polarized morphology of radial glial cell soma in the ventricular zone (VZ), with a long, polarized basal process oriented towards the pia. (F-H) pBLBP-EGFP-labeled radial glial cells (arrowheads) immunolabeled with radial glial-specific RC2 antibody (red). P, pial surface; V, ventricular surface. Scale bar: 90 μm in A; 70 μm in B; 30 μm in C-H.

Disrupted radial glial development in Cdc42^lox/lox;hGFAP-Cre cerebral cortex. hGFAP-Cre-mediated recombination was used to inactivate Cdc42 in radial glia. (A-C) Conditional deletion of Cdc42 in radial progenitors leads to branching defects in radial glial endfeet. Although the radial glia scaffold formed in Cdc42-deficient cortex, radial glial endfeet appear to be more branched when compared with controls. Arrowheads point to aberrantly branched endfeet of Cdc42-deficient radial glia (B,C). See Fig. S4 in the supplementary material for a higher magnification illustration. In addition, radial glial processes are less fasciculated (see Fig. S6 in the supplementary material) and their soma appear to be misplaced and less densely packed in Cdc42-deficient cortex (asterisk, B,C). See Fig. S5 in the supplementary material for higher magnification illustration. (D-F) Labeling with anti-Tbr1 antibodies (green) indicates normal placement of deeper layer neurons in control (D), but not in Cdc42-deficient (E,F), cortex. Moderate to severe misplacement of neurons was noted within the disrupted radial glia scaffolding of Cdc42 conditional null mice. Scale bar: 225 μm.

Gsk3 inhibition disrupts radial glial development and organization. (A-C) Cerebral cortices of E15.5 mouse embryos were electroporated with control or Gsk3 shRNA and pBLBP-IRES-EGFP plasmids and radial glia were analyzed 2 days later. (A) Radial glia expressing control shRNA displayed the characteristic polarized morphology. (B) By contrast, Gsk3 shRNA expression resulted in significant disruption to the radial glia scaffold, with many radial glia displaying wavy radial processes (arrowhead, B); compare the radial processes within the boxed areas of control (A) and Gsk3-deficient (B) cortex. (C) This deficit was rescued by co-electroporation of mutant Gsk3β resistant to shRNA-mediated degradation. (D,E) Inhibition of Gsk3 with a cell-permeable myristylated form of Gsk3 peptide inhibitor (D) or 6-bromoindirubin (E) also resulted in strong defects in the polarized morphology of radial glia. Arrow and arrowhead in D indicate an aberrantly branched and misoriented radial glia cell, respectively. (F) Quantification of disrupted radial glial polarity following the different methods of Gsk3 inhibition. Two dominant-negative Gsk3 constructs, Gsk3KM and Gsk3 R96A, which are known to inhibit all Gsk3 kinase activity and the activity of Gsk3 towards previously primed substrates, respectively, also disrupted radial glial polarity. Data shown are mean±s.e.m. ^*, P<0.001 versus controls (Student's t-test). P, pial surface; V, ventricular surface. Scale bar: 130 μm in A-C; 120 μm in D,E.

Cdc42 activity regulates distinct aspects of radial glial morphology. (A,B) Radial glia in mouse E15 cerebral cortices were electroporated with BLBP promoter-driven dominant-negative Cdc42-GFP or GFP (pBLBP-DN Cdc42-GFP or pBLBP-GFP) plasmids and imaged after 24 hours. (A) Individual radial glia cells expressing control GFP display the characteristic elongated, polarized basal process oriented towards pia. (B) By contrast, DN-Cdc42-expressing radial glia show highly branched basal process ends (arrow). (C) Quantification of radial glial endfeet types indicates a significant increase in branched endfeet following DN-Cdc42 expression. (D) Quantification of the effect of DN-Cdc42 on radial glial morphology. Radial glia spanning the full length of the cerebral wall or with shorter processes were counted. DN-Cdc42 expression leads to an increase in the number of radial glia with shorter processes. GFP⁺ cells without processes and multipolar cells were included in the `other' group. (E) Live imaging of Cdc42-GFP expression in radial progenitors in VZ indicates that Cdc42 preferentially accumulates at the apical or basal poles of radial glial cell soma (arrows), corresponding to the direction of interkinetic cell movement in the VZ. Cdc42-GFP images were pseudocolored (red, high Cdc42-GFP; blue, low Cdc42-GFP). Time elapsed is indicated in minutes. (F) Measurement of relative fluorescence intensities of Cdc42-GFP in defined areas within the apical or basal poles of progenitors undergoing cell movement confirms the differential distribution of Cdc42-GFP towards the direction of cell movement. A basal/apical Cdc42 fluorescence intensity ratio above 1 is indicative of higher localization at the basal pole, whereas values below 1 indicate accumulation at the apical pole. Arrows indicate the direction of cell movement. Data shown are mean±s.e.m. ^*, P<0.001 versus controls (Student's t-test). VZ, ventricular zone. Scale bar: 100 μm in A,B; 10 μm in E.