Figure 1.
hGFAP-Cre activity in the cerebellum.
Sagittal sections were shown. (A) Xgal staining in an E12.5 hGFAP-Cre, R26R mouse, showing the absence of Cre activity in the cerebellum. (B) Xgal staining of an E13.5 hGFAP-Cre, R26R, β-cateninfl/fl mouse. The rl (black arrowhead) showed Xgal labeling. (C) Xgal staining of an E14.5 hGFAP-Cre, R26R, β-cateninfl/wt mouse. The rl and egl showed extensive Xgal labeling. Xgal labeling also appeared in the mhb, vz (black arrowhead) and the interior of cerebellum (white arrowhead). The inset is a lateral section from the same mouse, showing more staining in the vz region. There is no staining in the pm (white arrow). (D) Xgal staining of an E16.5 hGFAP-Cre, R26R, β-cateninfl/wt mouse. The egl (arrow) and the interior of the cerebellum showed extensive Cre activity. (E) A P22 hGFAP-Cre, R26R mouse first stained with Xgal, and then stained with calbindin antibody using a fluorescent secondary antibody. Some cells were double-positive (arrows) and formed Xgal clusters. (F) Sections from E14.5, E17.5, and E18.5 hGFAP-Cre, β-cateninfl/fl mice and control littermates stained with anti-β-catenin antibody. The level of β-catenin was normal in the E14.5 mutant mouse, but clearly decreased in the EGL of E17.5 or E18.5 mutant mice. The white dashed line in E17.5 sections indicates the inner border of the EGL. β-catenin was still expressed in the pm of mutant mice. Abbreviations: cp, choroid plexus; egl, external granule cell layer; mhb, mid-hindbrain boundary; pm, pia mater; rl, rhombic lip; vz, ventricular zone. Scale bar: 100 μm (A–E) and 50 μm (F).
Figure 2.
Foliation defects in the hGFAP-Cre, β-cateninfl/fl mouse.
(A) The surface of a control P18 cerebellum was separated by several grooves (white arrowheads). The cerebella of its mutant littermates were smaller and had a smooth surface. (B) Midsagittal sections stained with hematoxylin and eosin. In the mutant mice recognizable fissures were designated by an asterisk after the corresponding name. Other fissures were assigned from rostral to caudal as ‘a1, a2...an’ so that the number and the sequence of these fissures are easy to be identified. At E14.5 and E17.5, the cerebella of mutant mice showed no evident difference from control mice. The pattern of fissures changed in the mutant cerebellum at P0 and P3, with more fissures in the rostral part and fewer fissures in the caudal part. The white dotted line indicates the position of the pr, recognized as the boundary between the anterior raised part and the posterior sloped part, or the deepest fissure in the anterior lobe (pcu in the control mice and dp in the mutant mice). The P0 mutant cerebellum had two ectopic fissures (a1 and a2) anterior to the dp and one ectopic fissure between the dp and pr (a4). As a result, the shapes of ABL and ADL were changed, respectively. The inset is immunostaining of calbindin, showing more clear structure of fissures a1 and a2. The P3 mutant cerebellum had three fissures anterior to the dp (a1, a2 and a3). The control cerebellum had only one (pct). At P8 and P18, the laminar structure was disordered and fissures were formed in the caudal part of the mutant cerebellum. Rostral is to the left and dorsal is to the top in each photograph. Abbreviations: ABL, anterobasal lobe; ADL, anterodorsal lobe; itc, intercrural; pct, precentral; pcu, preculminate; pl, posterolateral; ppy, prepyramidal; pr, primary; sec, secondary fissure. Scale bars: 100 μm in E14.5 and E17.5 sections, and 250 μm in P0-P18 sections.
Figure 3.
Variable foliation patterns among hGFAP-Cre, β-cateninfl/fl mice.
The white dotted line indicates the position of the pr, or the deepest fissure in the anterior lobe (pcu in the control mice and dp in the mutant mice). In the mutant mice recognizable fissures were designated by an asterisk after the corresponding name. Other fissures were assigned from rostral to caudal as ‘a1, a2...an’. (A) Example medial and lateral sections of four P4 littermates stained for calbindin and counterstained with DAPI. While the foliation patterns were similar in the two control cerebella, they were different in the two mutant cerebella. In both mutant 1# and mutant 2#, one ectopic fissure appeared between the two dotted lines (a4 and a5, respectively). Therefore the shape of Lobule IV/V was changed. From anterior to deepest fissure, there is only one fissure (pct) in the control mice, but there were two or more fissures in the mutant mice (a1 and a2 in mutant 1#; a1, a2 and a3 in mutant 2#). Lateral sections showed simpler folia than medial sections. Note that the mutant egl tended to be torn off (arrowhead). (B) Example medial sections of one control and three mutant cerebella at P7 or P8 stained with antibodies to calbindin and NeuN. In mutant 1# and mutant 2# one ectopic fissure (a4) appeared between the two dotted lines. There were more fissures anterior to dp in all the three mutant cerebella (a1 and a2 in mutant 1# and 2#; a1, a2 and a3 in mutant 3#). Arrows indicate the deeply-located lobules. Rostral is to the left and dorsal is to the top in each photograph. Abbreviations: itc, intercrural; pct, precentral; pcu, preculminate; pl, posterolateral; ppy, prepyramidal; pr, primary; sec, secondary fissure. Scale bars: 250 μm.
Figure 4.
Lamination defects in the hGFAP-Cre, β-cateninfl/fl mouse.
(A) Sagittal sections through the cerebellar cortex of P3 mice stained with antibody against laminin and with DAPI. The right panels are enlargements of the squares in the middle panels. Note that laminin was deposited around blood vessels in control and mutant mice (arrowheads). In control mice, laminin was incorporated into the meningeal basement membranes overlying the EGL, and penetrated into the folia at fissures (arrow in a). In mutant mice, laminin either crossed the base of granule cell accumulation (arrow in b), or did not penetrate into the granule cell accumulation (c and d). (B) Midsagittal sections stained for calbindin and DAPI. At P3, mutant cerebella had a relatively normal PCL and EGL. The PCL was disrupted at P7. The squares are enlarged in the lower panels (e, f, g and h). Green arrowheads in e and g indicate deeply-located PCL covered externally by granule cells. In f and h, white asterisks indicate fused sites of the PCL from adjacent lobules. The green asterisk indicates the PCL fused in the middle of an unsplit fissure. White arrows indicate delaminated Purkinje cells, and green arrows indicate gaps in the PCL. White arrowheads indicate granule cell ectopia at the fusion lines between adjacent lobules, and yellow arrowhead indicates granule cell ectopia below the pial surface. Scale bars: 100 μm in (A) and 250 μm in (B).
Figure 5.
Proliferation and differentiation of granule cells in the hGFAP-Cre, β-cateninfl/fl mouse.
(A) PH3 immunostaining showed that the proliferation zone in the EGL was largely maintained with no apparent ectopic proliferation at P0 and P4. At P8, some ectopic cell accumulations formed, more clearly seen in the enlargement of square a (brackets). PH3-positive cells formed small clusters at anterior fissures (arrows in b). The right panel is quantification of PH3-expressing cells in the EGL. Counted cells were divided by the length of the EGL (measured along the outer surface). The results were normalized to the mean of control cerebella and presented as mean ± SE. 3–5 slices from 2 or 3 mice were used for each group. Because the caudal part of the EGL in some P4 mutant mice tended to be torn off, these regions were excluded from the assay. Differences were statistically significant (p<0.05) as determined by Mann-Whitney two-sample rank sum test. (B) Immunohistochemistry for TAG1. At P3 in the inner EGL of both control and mutant cerebellum, an intact TAG1-expressing layer was seen, which became discontinuous and disordered in the P4 mutant cerebellum (arrowheads and arrows). (C) Immunohistochemistry for NeuN. Sections were also stained with anti-calbindin antibody and DAPI. In the control cerebellum, NeuN was expressed in the inner EGL and IGL. At P3, the mutant cerebellum still had this pattern. Later, ectopic NeuN was expressed by granule cells outside the deeply-located PCL (arrowheads), or by granule cells in an unseparated fissure (enlargements of squares j and k, compared with enlargement of square h in the bottom panel). P7 and P8 mutant cerebella are shown; NeuN+ granule cell ectopia were seen adjoining the pial surface of the P8 mutant cerebellum (arrow), but were rarely seen in the P7 mutant cerebellum. Despite these defects, the IGL did exist where a relatively normal monolayer of PCL was formed (white asterisks). Scale bars: 100 μm in (A), 20 μm in (B), and 100 μm in (C).
Figure 6.
Developmental stage-dependent glial defects in the hGFAP-Cre, β-cateninfl/fl mouse.
(A) Sagittal sections immunolabeled with anti-BLBP antibody. At E14.5, radial glial fibers were attached to the pial surface in both control and mutant mice. The right panels were acquired under high power (60×). At E17.5, in the caudal part of the mutant cerebellum, the density of radial fibers decreased, and some BLBP-positive cell bodies appeared in the EGL (arrows). At E18.5, more ectopic glial cells were seen in the EGL (arrows). (B) Section of a P0 mutant cerebellum stained with anti-BLBP antibody and DAPI. Colocalization indicates they were indeed cell bodies (arrows). (C) Sections from P0 control and mutant mice stained with antibodies against BLBP and nestin. Many double-positive cells were found in the mutant EGL (arrows), with a polarized appearance. (D) Sections of P8 control or mutant cerebellum double-stained with anti-BLBP and anti-GFAP antibodies. In the control mouse, GFAP was expressed in the white matter glia (arrowhead in the upper panel), Bergmann glia (arrows in a) and glia in-between (short arrows in b). In the mutant mouse, GFAP was only expressed by white matter glia (arrowhead in the upper panel) and ectopic glia in the EGL (arrows in c), but not by glial cells in-between (short arrows in d). The lower panels are enlargements of the squares in the upper panels (a–d). Scale bars: 20 μm.
Figure 7.
Non cell-autonomous role of β-catenin in major cerebellar cell types.
(A) Sections from P8 and P26 L7-Cre, β-cateninfl/wt or L7-Cre, β-cateninfl/fl cerebellum stained with anti-calbindin antibody and DAPI. Purkinje cells from the L7-Cre, β-cateninfl/fl mouse had normal morphology and well-organized lamination, just like those from control cerebellum. (B) Wild-type C57/BL6 mice electroporated with pSuper vectors carrying either control siRNA or β-catenin siRNA at P4 and analyzed at P14. Most siRNA-transfected cells were able to migrate into the IGL. There was no significant difference between the two groups. (C) E13.5 R26R, β-cateninfl/wt or R26R, β-cateninfl/fl mice injected with AdCreM2. The mice were annalyzed at P16. Sections were first stained with Xgal and then with anti-NeuN antibody using the DAB method. Most of NeuN-positive cells were located in the IGL (arrows), indicating that Cre-expressing granule cells had migrated into the IGL. (D and E) GFAP-CreERT2, R26R, β-cat fl/fl mice were induced twice between P0 and P3, and analyzed at P7 or P18, as indicated. Xgal staining showed extensive recombination in the BGL. The ml also had scattered Xgal+ cells, probably interneurons. The EGL had no Xgal+ cells. The insets are enlargements of the squares. (F and G) GFAPCreERT2, R26R, β-catfl/wt mice or GFAPCreERT2, R26R, β-catfl/fl mice were induced and analyzed on postnatal days as indicated, and immunostained with anti-GFAP (F) or anti-BLBP (G) antibody. In mutant cerebella, Bergmann glia also expressed BLBP and GFAP and extended fibers to the pial surface. Inset in G is the enlargement of the squared area showing glial cell bodies were seen in the BGL (arrows) and rarely seen in the ml. The slices in (E) and (G) were from the same mouse. Abbreviations: Bgl, Bergmann glia layer; egl, external granule cell layer; ml, molecular layer; pm, pia mater. Scale bars: 100 μm.