Articles, Development/Plasticity/Repair

Suppressor of Fused Controls Mid-Hindbrain Patterning and Cerebellar Morphogenesis via GLI3 Repressor

Jinny J. Kim, Paul S. Gill, Lianne Rotin, Matthijs van Eede, R. Mark Henkelman, Chi-Chung Hui and Norman D. Rosenblum
Journal of Neuroscience 2 February 2011, 31 (5) 1825-1836; https://doi.org/10.1523/JNEUROSCI.2166-10.2011

Abstract

Sonic Hedgehog and its GLI transcriptional effectors control foliation complexity during cerebellar morphogenesis by promoting granule cell precursor proliferation. Here, we reveal a novel contribution of Sonic Hedgehog-GLI signaling to cerebellar patterning and cell differentiation by generating mice with targeted deletion of Suppressor of Fused (SuFu), a regulator of Sonic Hedgehog signaling, in the mid-hindbrain. Postnatal SuFu-deficient mice exhibit impaired motor coordination and severe cerebellar mispatterning. SuFu conditional knock-out embryos display abnormal mid-hindbrain morphology associated with misexpression of Fgf8, and delayed differentiation and abnormal migration of major cerebellar cell types. Sonic Hedgehog is ectopically expressed in the external granule layer and Hedgehog signaling is upregulated. While expression of full-length GLI transcriptional activators downstream of Hedgehogs is markedly reduced, a processed form of GLI3, a transcriptional repressor, is essentially lost. Genetic expression of a Gli3 allele encoding GLI3 repressor in SuFu-deficient mice largely rescues abnormal cerebellar patterning and cell differentiation observed in mice with SuFu deficiency alone. Together, our data demonstrate that SuFu controls cerebellar patterning and cell differentiation in a GLI3 repressor-dependent manner.

Introduction

Cerebellar development is controlled by the state of Sonic Hedgehog (SHH) signaling. Binding of SHH ligand to the transmembrane receptor PATCHED results in generation of full-length GLI transcriptional activators (GLI1, GLI2, GLI3) and inhibition of proteolytic processing of full-length GLI3 protein to a shorter form that acts as a transcriptional repressor. Absence of SHH signaling promotes formation of GLI3 repressor (GLI3R). The expression of SHH is low in the dorsal mid-hindbrain, from which the cerebellum is derived (Fuccillo et al., 2006). Accordingly, GLI3R levels are predicted to be relatively high. Indeed, GLI3R controls the domain of Fgf8 expression, which regulates both mid-hindbrain specification and the size of the mid-hindbrain (Aoto et al., 2002; Blaess et al., 2006, 2008). As the cerebellum develops, foliation is established by SHH-dependent proliferation of granule cell precursors. High levels of SHH signaling are achieved at embryonic day 17 (E17) as Purkinje cells begin to secrete SHH (Corrales et al., 2006). Persistent SHH signaling results in uncontrolled proliferation of granule cell precursors and causes medulloblastoma, a cerebellar pediatric cancer (Dahmane and Ruiz i Altaba, 1999; Corrales et al., 2004, 2006; Lewis et al., 2004).

Suppressor of Fused (SUFU), an intracellular PEST-domain-containing protein, was initially identified as a suppressor of Hedgehog (HH) signaling in Drosophila (Préat, 1992). Murine SuFu inactivation causes early embryonic lethality at E9.5 with cephalic and neural defects and ligand-independent activation of HH signaling similar to that observed in mice deficient in Patched1 (Ptc1), a negative regulator of HH signaling (Svärd et al., 2006). Moreover, SuFu functions as a tumor suppressor in medulloblastoma (Taylor et al., 2002; Lee et al., 2007). Biochemical analyses in mammalian cells have provided insight into mechanisms by which SUFU exerts its effects. SUFU prevents nuclear translocation of GLI1 and GLI2 and recruits GSK3β to promote proteolytic degradation of full-length GLI3 (Kogerman et al., 1999; Barnfield et al., 2005; Kise et al., 2009). SUFU also recruits the SAP18-mSin3 histone deacetylase corepressor complex to GLI-binding regions within HH target promoters and represses transcription (Cheng and Bishop, 2002). Interestingly, recent data suggest that SUFU can also function to stabilize full-length GLI proteins by inhibiting SPOP-dependent proteolytic degradation of full-length GLI proteins (Chen et al., 2009; Wang et al., 2010). Together, these observations indicate that SUFU may function in a dual capacity during SHH signal transduction to control both GLI activator and repressor function.

The early embryonic lethality observed in mice with germline inactivation of SuFu has precluded analysis of SuFu functions during cerebellar morphogenesis. To overcome this limitation, we used Cre recombinase (Cre)-mediated SuFu inactivation targeted to the cerebellum. Here, we demonstrate a novel role for SUFU in mid-hindbrain patterning, cell differentiation and regulation of SHH-GLI signaling during cerebellar development. Remarkably, cerebellar dysplasia observed in SuFu-deficient mice is largely dependent on expression of GLI3R.

Materials and Methods

Mice.

SuFuloxP (Chen et al., 2009) and Gli3Δ699 (Böse et al., 2002) mice were maintained on a mixed background, and SuFu+/− (kindly provided by C.-C. Hui), Hoxb.7-Cre (Zhao et al., 2004), Ptc1+/− (Goodrich et al., 1997) and Ptc1loxP (Ellis et al., 2003) mice were maintained on a CD1/129 background. ROSAlacZ/lacZ mice (Soriano, 1999) were maintained on a C57BL/6 background. All animals were housed in the Animal Facility of the Hospital for Sick Children (Toronto, ON, Canada) and were genotyped by PCR as previously described (Böse et al., 2002; Zhao et al., 2004; Lee et al., 2007). Animal experiments were approved by the ethics committee at the Hospital for Sick Children.

Histology and immunohistochemistry.

The noon of the day of a vaginal plug was used as E0.5 after fertilization. Embryos at various developmental stages were harvested from killed pregnant females, decapitated and fixed in 4% paraformaldehyde (PFA) in PBS. Postnatal day 2 (P2)–P30 mice were killed by CO2 asphyxiation, and brains were dissected out and fixed in 4% PFA in PBS. Fixed tissues were embedded in paraffin for sectioning to prepare 5 μm parasagittal sections on glass slides. Hematoxylin and eosin staining was performed to visualize cerebellum morphology. Immunofluorescence was performed on PFA- or formalin-fixed, paraffin-embedded tissue sections using the following primary antibodies: mouse anti-calbindin D28K (1:200, Sigma), rabbit anti-SUFU (1:1500, kindly provided by C.-C. Hui), rabbit anti-BLBP (brain lipid-binding protein) (1:200, Abcam), mouse anti-GFAP (1:200, Millipore), mouse anti-NeuN (1:100, Millipore), rabbit anti-Pax2 (1:200, Covance), rabbit anti-Pax6 (1:300, Covance), mouse anti-PCNA (proliferating cell nuclear antigen) (1:500, Cell Signaling Technology), rabbit anti-Nestin (1:500, Abcam), mouse anti-Ki67 (1:200, BD Biosciences), and mouse anti-βIII tubulin (1:400, Covance). Fluorescence detection was achieved by staining tissue sections with species-specific Alexa Fluor 488 and Alexa Fluor 568 secondary antibodies (1:500, Invitrogen) and counterstaining with 4′,6′-diamidino-2-phenylindole dihydrochloride (1:1000, Sigma). Immunohistochemistry experiments were performed using rabbit anti-calbindin D28K (1:1000, Millipore) and biotinylated secondary antibodies (ABC Kit, Vector Laboratories).

β-Galactosidase reporter assay.

X-gal staining of whole embryos and whole brains was performed according to published methods (Cain et al., 2009). After staining, embryos were postfixed in 10% formalin for paraffin embedding in paraffin, and whole embryos were photographed using a Leica EZ4D dissecting microscope. X-gal stained tissue sections were then counterstained with neutral red for nuclear staining.

In situ mRNA hybridization. In situ hybridization was performed on PFA-fixed and paraffin-embedded tissue sections as described previously (Mo et al., 1997) using digoxigenin-labeled RNA probes encoding Fgf8 (kindly provided by Lijun Chi, Hospital for Sick Children, Toronto, ON, Canada), Ptc1 (kindly provided by Andrew McMahon, Harvard University, Cambridge, MA), and Shh and Otx2 (kindly provided by C.-C. Hui).

Magnetic resonance imaging.

P21 and P30 mice were anesthetized by injecting Rompun (20 mg/kg) and ketamine (100 mg/kg) intraperitoneally and underwent transcardiac perfusion with 4% PFA. They were subsequently decapitated. Magnetic resonance images were taken at 32 μm isotropic resolution and analyzed.

Western blot.

Whole-cell protein lysates were prepared from isolated postnatal day 7 mouse cerebellums by Dounce A homogenization in ice-cold radioimmunoprecipitation assay buffer (50 mm Tris-HCl pH 7.4, 1% Nonidet P-40, 0.5% Na-deoxycholate, 150 mm NaCl, 1 mm EGTA, and 5 mm EDTA) containing Protease Inhibitor Cocktail P8340 (Sigma). Protein lysates were separated by SDS-PAGE and transferred onto a polyvinylidene difluoride membrane. Following membrane blocking in 5% nonfat dry milk, membranes were exposed to one of the following primary antibodies overnight at 4°C: rabbit α-SUFU (1:5000 dilution; donated by C.-C. Hui), mouse α-GLI1 (1:3300 dilution; Abcam), rabbit α-GLI2 (1:250 dilution; donated by C.-C. Hui), or rabbit α-GLI3 H280 (1:250 dilution; Santa Cruz Biotechnology) antibody. Signal was detected using chemiluminescence (ECL Kit; GE Healthcare).

Cerebellar precursor isolation and culture.

E12 wild-type and SuFu-deficient cerebella were microdissected in HBSS (Invitrogen) and finely minced in Neurobasal media containing 1% v/v penicillin/streptomycin, 2% B27 supplements, 500 μm l-glutamine (all from Invitrogen) and 40 ng/ml FGF2 (Collaborative Biomedical Products). Cerebella were triturated to single cell suspension and plated on chamber slides coated with 1% poly-d-lysine and 2% laminin (both from Collaborative Biomedical Products). Dissociated cells were grown at 37°C and 5% CO2 for 3 d before immunostaining analysis according to standard protocols using antibodies listed above.

Statistics.

Student's t tests were applied to calculate SD and SE. Findings were determined to be significant if the p value was <0.05. All data reported here are based on representative samples and an experimental n ≥ 3.

Results

Targeted deletion of SuFu in the mid-hindbrain

Since SuFu-null mutants exhibit early embryonic lethality (E9.5) (Svärd et al., 2006), we generated conditional knock-out mice (Hoxb.7-Cre;SuFu−/loxP) using Cre recombinase under the control of the Hoxb.7 promoter. We identified the spatial location of Hoxb.7-Cre expression at the onset of cerebellar morphogenesis (between E9.5 and E10.5) using the ROSA reporter gene. Cre expression, assayed by β-galactosidase activity, was detected as early as E9.5 in the mid-hindbrain region including rhombomere 1 and was also observed in the Wolffian duct and the dorsal root ganglia, as previously reported (Zhao et al., 2004) (Fig. 1A,B). Cre recombinase activity was detected in all cells of the developing cerebellum (Fig. 1C). SUFU expression during normal cerebellar development was defined by immunostaining using specific anti-SUFU antibodies. SUFU was strongly expressed in ventricular zone precursors at E13.5 (Fig. 1D,D′). The spatial expression pattern of SUFU was comparable to that of Nestin, a marker of radial glia precursors (Fig. 1D″,D‴). Both Nestin and SUFU are expressed in fibers throughout the ventricular zone. At E18.5, SUFU expression was observed in the location of the Purkinje cell layer, Bergmann glia fibers and deep cerebellar nuclei (Fig. 1E,E′,F,F′). Moreover, SUFU is expressed in a pattern highly similar to that of Bergmann glia, marked by BLBP (Fig. 1E″,E‴) consistent with its expression in radial glia precursors at E13.5. At P6, during the period of granule cell precursor proliferation, the outer external granule layer (EGL) lacked SUFU expression. In contrast, SUFU was expressed in both the inner EGL and Purkinje cell layer (Fig. 1G,G′). In the adult cerebellum (P21), SUFU expression was restricted to Purkinje cells (Fig. 1H,H′). SUFU deficiency was confirmed in Hoxb.7-Cre;SuFu−/loxP mice using Western analysis and cerebellar protein lysates (Fig. 1I). In contrast to wild-type controls (Fig. 1I, lanes 1–4), SUFU protein was barely detectable in cerebellar lysates from Hoxb.7-Cre;SuFu−/loxP mice (Fig. 1I, lanes 5–7) and was only slightly greater in quantity than that observed in SuFu germline knock-out embryo lysates (Fig. 1I, lane 8). Residual amounts of SUFU in Hoxb.7-Cre;SuFu−/loxP mice is likely due to incomplete deletion of SuFu in some cells.

Figure 1.
Figure 1.

Hoxb.7-Cre is expressed in the mid-hindbrain and deletes SuFu in the cerebellum. A, B, Hoxb.7-Cre mice were crossed to ROSA26 mice to examine Cre expression. Hoxb.7 activates Cre in the mid-hindbrain at E9.5, before the onset of cerebellar morphogenesis (A). β-Galactosidase activity indicates Cre expression in the mid-hindbrain (arrows) and the Wolffian duct (B, arrowhead). C, Strong Cre activity is detected in the cerebellum at E18.5. LacZ stain is denoted by blue stain and neutral red stain is denoted by pink stain. D–D‴, SUFU and Nestin, a specific marker of radial glia precursors, are expressed in an overlapping pattern in the ventricular zone of E13.5 wild-type cerebella (arrows, D″, D‴). E–E‴, At E18.5, SUFU is detected in the Purkinje cell layer (PCL) and is expressed in a pattern similar to that of Bergmann glia radial fibers, identified by BLBP expression (E″, E‴, arrows). F, F′, SUFU is also observed in the presumptive deep cerebellar nuclei (DCN). G, G′, The inner external granule layer (iEGL) and PCL express SUFU, whereas the outer EGL (oEGL) lacks SUFU expression at P6. H, H′, SUFU is specifically expressed in Purkinje cells at P21. I, Western blot analysis reveals a marked decrease in SUFU protein levels in P7 cerebellar protein lysates. 1, SuFu+/loxP; 2–4, SuFu−/loxP; 5–7, Hoxb.7-Cre;SuFu−/loxP; 8, SuFu−/−. Scale bars: (in C) A, B, 1.1 mm; C, 200 μm; (in D) D–H, 200 μm; D′–H′, 100 μm; (in D″) D″, D‴, E″, E‴, 30 μm.

Loss of SUFU results in cerebellar dysfunction and severe patterning defects

Hoxb.7-Cre;SuFu−/loxP pups were comparable to control littermates at P0 in both gross anatomical appearance and body size (Fig. 2A). However, by P21 mutant mice demonstrated severe runting and most mutants failed to survive the first month of life (Fig. 2B). The presence of milk in the stomach of mutant pups and their capacity to consume food left on the bottom of their cage postweaning (data not shown) strongly suggested that runting was not due to decreased food intake. Interestingly, Hoxb.7-Cre;SuFu−/loxP mutants exhibited significantly impaired motor coordination that resulted in an abnormal gait and a failure to walk forward (Fig. 2C,D; supplemental Movie S1, available at www.jneurosci.org as supplemental material). Consistent with these findings, magnetic resonance imaging (MRI) demonstrated that Hoxb.7-Cre;SuFu−/loxP cerebella displayed abnormal gross anatomical morphology characterized by an absent or hypoplastic vermis and lack of normal foliation (Fig. 2E,F). Gross morphological analysis of Hoxb.7-Cre;SuFu−/loxP mutant cerebella revealed lack of normal foliation pattern and clear demarcation between the vermis and the lateral hemispheres (Fig. 2G,H). Furthermore, compared with control brains, the superior and inferior colliculi of Hoxb.7-Cre;SuFu−/loxP mutants appeared hyperplastic. Histological analysis demonstrated lack of folia and distinct cellular layers that characterize control cerebella (Fig. 2I,J). Moreover, in contrast to the normal cerebellum which consists of a Purkinje cell monolayer and an inner granule layer (Fig. 2K,M), in mutant mice both granule cells and Purkinje cells were randomly localized in cellular aggregates (Fig. 2L,N). Together, these data demonstrate a requirement for SuFu during cerebellar morphogenesis.

Figure 2.
Figure 2.

SUFU deficiency causes a severe reduction in body size, defective motor coordination and cerebellar mispatterning. A, B, Hoxb.7-Cre;SuFu−/loxP mutants display a reduction in body size, abnormal cerebellar morphology and defective patterning. Newborn Hoxb.7-Cre;SuFu−/loxP mutants are comparable to their control littermates in body size (A). Adult mutants display severe body size reduction (B). C, D, While P21 control mice display a normal gait marked by forefoot (red)-hindfoot (blue) coordination (C), Hoxb.7-Cre;SuFu−/loxP mutants display a severely ataxic gait, a sign of impaired motor coordination, characterized by an inability to walk forward (D). Arrows indicate the direction in which mice faced. E, F, MRI three-dimensional volume renderings of Hoxb.7-Cre;SuFu−/loxP mutant brains (F) indicate severe abnormalities in cerebellar gross morphology (red) compared with control cerebella (E). Note the hypoplastic or absent vermis in mutant cerebella. G, H, Hoxb.7-Cre;SuFu−/loxP cerebella (H) are characterized by a lack of demarcation between the vermis and hemispheres and a lack of foliation. I, J, Hematoxylin and eosin staining of control (I) and mutant cerebella (J) reveals severe cerebellar mispatterning marked by absence of distinct folia and fissures in mutant tissue at P21. K–N, Higher resolution imaging of control (K, M) and mutant cerebella (L, N) indicates cellular disorganization and abnormal cytoarchitecture in mutant tissue. MB, Midbrain; CH, cerebellar hemisphere; CV, cerebellar vermis; ML, molecular layer; PC, Purkinje cell; IGL, inner granule layer; WM, white matter (see also supplemental Movie S1, available at www.jneurosci.org as supplemental material).

SUFU deficiency causes abnormal mid-hindbrain patterning and delayed cerebellar formation

To determine the embryonic origins of the postnatal cerebellar phenotype, we examined control and Hoxb.7-Cre;SuFu−/loxP embryos at various stages of cerebellar morphogenesis. At E13.5, when a thickened cerebellar primordium is clearly visible in control embryos, no such structure was detected in Hoxb.7-Cre;SuFu−/loxP mutants (Fig. 3A,B). A significant overgrowth and expansion of the isthmic tissue was observed in mutant embryos, which led to truncation of the midbrain. This tissue overgrowth between the cerebellar primordium and inferior colliculus was determined to be isthmic based on the lack of cerebellum-specific cell marker expression, such as calbindin and PAX6, and the failure of the EGL to extend into the isthmic region (see Fig. 5). Interestingly, conditional knock-out of Ptc1 using Hoxb.7-Cre mice at E13.5 resulted in severe overgrowth of the mid-hindbrain without delayed cerebellar formation or apparent defects in anterior–posterior patterning in 67% of all mutants examined (4 of 6) at E13.5 (supplemental Fig. S1A–C, available at www.jneurosci.org as supplemental material). In 33% of mutants (2 of 6), exencephaly was detected caused by an open neural tube defect in the region of dorsal mid-hindbrain. Thus, in contrast to Ptc1, SuFu controls anterior–posterior patterning of the mid-hindbrain.

Figure 3.
Figure 3.

Histologic analysis of the embryonic SuFu-deficient phenotype. A, B, Hoxb.7-Cre;SuFu−/loxP embryos display abnormal midbrain patterning and severely dysplastic cerebella. Mutant embryos lack a cerebellar primordium at E13.5 (arrowhead) and have a truncated dorsal midbrain (MB). Note the presence of hyperplastic isthmic tissue (IS) at the midbrain-cerebellum boundary (arrows). C, D, Abnormal-looking cerebellum is present in Hoxb.7-Cre;SuFu−/loxP mutants (inset) and the IS is significantly expanded at E15.5 (arrow). E, F, Mutant cerebella lack foliation and distinct cell layers and display defective patterning at E18.5. Scale bar: (in E) A–D, 800 μm; E, F, 200 μm (see also supplemental Fig. S1, available at www.jneurosci.org as supplemental material).

An EGL was detected in E15.5 control embryos as granule cell precursors migrate from the rhombic lip over the dorsal surface of the cerebellar primordium (Fig. 3C). Although an EGL was detected in Hoxb.7-Cre;SuFu−/loxP mutants, it failed to extend from the rhombic to the isthmus-cerebellum junction. Further, the rhombic lip was larger compared with that of control embryos, indicating delayed granule cell migration from the rhombic lip (Fig. 3D). In addition, the isthmic overgrowth was dramatically increased in size, forming large folds of neural tissue between the midbrain and the cerebellar primordium and resulting in persistent truncation of the midbrain. By E18.5, distinct cell layers and foliation were observed in control cerebella as Purkinje cells formed a monolayer beneath the EGL (Fig. 3E). However, in Hoxb.7-Cre;SuFu−/loxP cerebella, there was a lack of distinct cell layers and foliation (Fig. 3F). Together, these observations demonstrate that abnormalities caused by SUFU deficiency arise early during mid-hindbrain patterning and delay the onset of cerebellar formation.

Fgf8 expression is abnormal in the absence of SuFu

Mid-hindbrain structures are characterized by the expression of transcription factors and secreted factors that are generated by the isthmic organizer. To elucidate the mechanism by which abnormal mid-hindbrain morphology is generated in SuFu-deficient mice, we performed in situ hybridization to investigate specification of specific regions of the mid-hindbrain. In both control and Hoxb.7-Cre;SuFu−/loxP embryos, Otx2 marked the entire midbrain with a sharp boundary at the isthmus (Fig. 4A,A′,B,B′). While the Fgf8 expression domain in control embryos marked the junction between the midbrain and the cerebellar primordium with sharp boundaries, Fgf8 expression was shifted in an anterior manner in Hoxb.7-Cre;SuFu−/loxP mutants (Fig. 4C,D). As a result, although Otx2 and Fgf8 expression domains do not overlap in wild-type embryos, a significant overlap of the two expression domains occurred in Hoxb.7-Cre;SuFu−/loxP mutants (Fig. 4C′,D′). Furthermore, Fgf8 expression detected in the nonoverlapping region was markedly reduced and lacked clear boundaries that demarcate its expression domain. This abnormal expression of mid-hindbrain markers suggested that SuFu controls proper specification of distinct neural compartments during mid-hindbrain patterning.

Figure 4.
Figure 4.

Loss of SUFU causes abnormal mid-hindbrain patterning. A, B, Otx2 marks the midbrain (MB) in control and Hoxb.7-Cre;SuFu−/loxP embryos at E12.5. Note the anterior shift of the posterior limit of Otx2 expression in Hoxb.7-Cre;SuFu−/loxP mutants. Arrows indicate the presumptive cerebellar primordium. Areas demarcated by rectangles are shown at a higher magnification in A′–D′. C, D, Fgf8 expression is detected in a narrow band at the boundary of the midbrain and the cerebellar primordium. In Hoxb.7-Cre;SuFu−/loxP mutants, Fgf8 expression is shifted in an anterior manner. A′–D′, While Otx2 and Fgf8 expression domains do not overlap and have sharp boundaries in control embryos, a significant overlap (red arrowhead) is observed due to the anterior shift of Fgf8 expression in Hoxb.7-Cre;SuFu−/loxP mutants. Fgf8 expression in the nonoverlapping region (black arrowhead) is weak and lacks sharp boundaries. Dashed line outlines the expression domain of Otx2. Scale bars: (in A′) A–D, 800 μm; A′–D′, 200 μm.

SuFu deficiency delays cerebellar cell differentiation

To investigate the effect of SuFu deficiency on cell differentiation, we studied expression of markers of cerebellar cell types. Marked by PAX6, granule cell precursors arise from the rhombic lip at E13.5, and subsequently form the EGL over the surface of the cerebellum (Fig. 5A–C). Granule cell precursors were absent in Hoxb.7-Cre;SuFu−/loxP mutants at E13.5 (Fig. 5D). Although the EGL was detected at later embryonic stages in mutant mice, it failed to form a thin layer of cells over the dorsal surface of the cerebellar primordium (Fig. 5E,F). In addition, the rhombic lip was enlarged compared with control cerebella. BLBP-expressing Bergmann glia, specialized radial glia found in the cerebellum, normally arise from the ventricular zone at E13.5 and migrate toward the EGL to form a layer with Purkinje cells (Fig. 5G–I). However, Bergmann glia were not detected in Hoxb.7-Cre;SuFu−/loxP cerebella at E13.5 (Fig. 5J). Interestingly, Bergmann glia were localized in the ventricular zone at E16.5 when they normally migrate toward the EGL (Fig. 5K). These cells failed to form a layer and were severely disorganized and ectopic at E18.5 in mutant cerebella (Fig. 5L). In contrast, BLBP-expressing radial glia precursors were detected in Hoxb.7-Cre;Ptc1−/loxP embryos at E13.5, suggesting a role for SuFu in cell differentiation that is distinct from that of Ptc1 (supplemental Fig. S1D, available at www.jneurosci.org as supplemental material). Similar to Bergmann glia, Purkinje cells, marked by calbindin, were detected in the ventricular zone at E14.5 in control cerebella (Fig. 5M). Purkinje cell migration and the formation of a Purkinje cell layer beneath the EGL was observed by E18.5 (Fig. 5N,O). In contrast, Purkinje cells were absent at E14.5 and E16.5 in Hoxb.7-Cre;SuFu−/loxP mutants (Fig. 5P,Q). While Purkinje cells were detected in a region underlying the EGL at E18.5, the number of Purkinje cells was decreased compared with control cerebella (Fig. 5R). GABAergic interneuron precursors, marked by PAX2 expression, were detected as early as E13.5 in control cerebella and were located diffusely throughout the mantle zone of the cerebellum by E18.5 (Fig. 5S–U). In contrast, while PAX2 expression was undetectable at E13.5 in Hoxb.7-Cre;SuFu−/loxP mutants, the distribution and number of PAX2-positive cells was dramatically expanded at E16.5 and E18.5, extending rostrally into the isthmus (Fig. 5V–X). Quantitative analysis of PAX2-positive cells revealed an ∼3-fold increase in the number of GABAergic interneuron precursors in SuFu-deficient cerebella (*p = 0.01; supplemental Fig. S2A, available at www.jneurosci.org as supplemental material). Collectively, these findings suggest that SuFu plays an important role in the temporal regulation of cell differentiation.

Figure 5.
Figure 5.

Cerebellar cell differentiation and migration is abnormal in the absence of SuFu. A–F, PAX6 staining reveals that Hoxb.7-Cre;SuFu−/loxP granule cell precursors are delayed in cell specification and form a defective EGL. G–L, As marked by BLBP, Bergmann glia arise from the ventricular zone (VZ) late and display abnormal migration and disorganization in Hoxb.7-Cre;SuFu−/loxP mutants. M–R, Calbindin-expressing Purkinje cell precursors in Hoxb.7-Cre;SuFu−/loxP cerebella are delayed in specification and fail to form a thin layer underneath the granule cells. Note that the number of Purkinje cells is markedly reduced compared with that of control cerebella. S–X, PAX2 staining demonstrates that GABAergic interneuron precursors in mutant cerebella exhibit a delay in cell specification and are significantly increased in number compared with control cerebella. Also, ectopic localization of interneuron precursors is detected in the isthmus (IS) at E18.5. CB, cerebellum; PCL, Purkinje cell layer; RL, rhombic lip. Arrowheads indicate cell marker-positive precursor cells. Scale bar: (in D) A–F, M–X, 200 μm; G–L, 100 μm (see also supplemental Figs. S1, S2, available at www.jneurosci.org as supplemental material).

To determine whether SUFU controls precursor cell fate or the timing of terminal cell differentiation, we isolated cerebellar precursors from E12 embryos and cultured cells in vitro. Double staining with BLBP and βIII tubulin, markers of glial precursors and neurons, respectively, 3 d postplating, revealed a sixfold increase in the proportion of BLBP-positive cells versus βIII tubulin in cell cultures derived from Hoxb.7-Cre;SuFu−/loxP cerebellar tissue compared with wild-type cerebellar cultures (control: 0.53 ± 0.08; mutant: 3.00 ± 0.45, *p < 0.001, Fig. 6A,B). Consistent with the fact that BLBP marks radial glia precursor cells, both Nestin and Ki67 were significantly upregulated in Hoxb.7-Cre;SuFu−/loxP cultures, as indicated by a 2.3-fold increase in the proportion of Nestin- and Ki67-positive cells (control: 32.0 ± 3.47%; mutant: 74.4 ± 3.46%, *p < 0.0001, Fig. 6C,D). In parallel, neuronal cells, marked by calbindin and PAX6, were decreased in Hoxb.7-Cre;SuFu−/loxP cultures (Fig. 6E–H). Thus, loss of SUFU leads to enhanced gliogenesis at the expense of neurogenesis and causes a delay in neuronal differentiation.

Figure 6.
Figure 6.

SuFu-deficient cells undergo enhanced gliogenesis and are delayed in neuronal differentiation. A, B, Double staining of BLBP (green) and βIII tubulin (red) demonstrates that the proportion of BLBP-expressing cells versus βIII tubulin cells is significantly increased in Hoxb.7-Cre;SuFu−/loxP cerebellar cultures (control: 0.53 ± 0.08; mutant: 3.00 ± 0.45, p < 0.001). C, D, Nestin-positive cells (green) coexpress Ki67 (red), and the proportion of Nestin- and Ki67-positive cells is significantly upregulated in Hoxb.7-Cre;SuFu−/loxP cerebellar cultures (control: 32.0 ± 3.47%; mutant: 74.4 ± 3.46%, p < 0.0001). Note that the neuronal processes, as visualized by Nestin staining, are markedly longer in Hoxb.7-Cre;SuFu−/loxP cells (insets). E, F, Fewer calbindin-positive cells (arrowheads) are detected in Hoxb.7-Cre;SuFu−/loxP cerebellar cultures compared with control cultures. G, H, PAX6 expression (arrowheads) is decreased in Hoxb.7-Cre;SuFu−/loxP cultures. Scale bar, 200 μm.

SUFU deficiency results in aberrant SHH signaling and decreased GLI proteins

Since SuFu is implicated in HH signaling as a mediator of proteolytic processing of GLI activators into repressors, we investigated the state of SHH signaling activity in the absence of SUFU. At the onset of SHH signaling activation in the cerebellum (E17), weak Ptc1 expression, a marker of SHH signaling activity, was detected in the EGL (Fig. 7A). In Hoxb.7-Cre;SuFu−/loxP mutants, significant upregulation and ectopic expression of Ptc1 were observed in the EGL and in the ventricular zone, respectively, indicating an increase in the number of cells responsive to SHH signal (Fig. 7B). To determine the ability of mutant Purkinje cells to generate SHH, we examined expression of Shh mRNA in E18.5 cerebella. Interestingly, while Shh expression was observed in the Purkinje cell layer in control cerebella, Shh was misexpressed in the EGL and absent in the Purkinje cell layer in Hoxb.7-Cre;SuFu−/loxP cerebella (Fig. 7C,D). These findings indicate that granule cell precursors, in the absence of SUFU, aberrantly express Shh and upregulate their response to SHH signaling. Consistent with this overactivation of SHH signaling, cell proliferation, as assayed by PCNA staining, was markedly increased in Hoxb.7-Cre;SuFu−/loxP mutants at E18.5 (supplemental Fig. S3, available at www.jneurosci.org as supplemental material). Consistent with these results and contrasted with controls, expression of Gli3 was absent in the EGL in Hoxb.7-Cre;SuFu−/loxP cerebella at E18.5 (Fig. 7E,F). Interestingly, by P6, Gli3 expression in the outer EGL was similar in Hoxb.7-Cre;SuFu−/loxP cerebella (Fig. 7H) compared with controls (Fig. 7G). Next, we examined the effect of SUFU deficiency on GLI protein levels. Western blot analysis revealed that levels of GLI activators were significantly reduced in P7 Hoxb.7-Cre;SuFu−/loxP cerebellar lysates compared with control samples (Fig. 7I). Interestingly, the level of GLI3R was barely detectable in the absence SUFU. These results indicate that SUFU positively regulates the levels of GLI activators and the generation of GLI3R.

Figure 7.
Figure 7.

Loss of SUFU leads to upregulation and misexpression of SHH signaling activity and decreased levels of GLI proteins. A, B, mRNA in situ hybridization shows weak expression Ptc1, a readout of SHH signaling in the EGL in E17.5 control cerebella (A). In contrast, Hoxb.7-Cre;SuFu−/loxP cerebella display upregulation of Ptc1 and ectopic expression in the ventricular zone (VZ) (B). C, D, Shh is detected in the Purkinje cell layer (PCL) and weakly in the EGL in control cerebella at E18.5 (C). In the absence of SUFU, Shh is absent in the PCL (dashed line) and misexpressed by the EGL (D). E, F, Gli3 mRNA expression is absent in the EGL in Hoxb.7-Cre;SuFu−/loxP cerebella at E18.5. G, H, Gli3 expression is observed in all cell types and slightly weaker in the outer EGL (oEGL) compared with the inner EGL (iEGL) in control cerebellar at P6 (G). Gli3 is expressed ubiquitously in the cerebellum in Hoxb.7-Cre;SuFu−/loxP mice at P6 (H). I, Western blot analysis of P7 cerebellar lysates reveals reduced levels of GLI1 and GLI2 and loss of the full-length (GLI3190 kDa) and truncated (GLI383 kDa) forms of GLI3 in the absence of SUFU. 1–3, SuFu+/loxP; 4, 5, Hoxb.7-Cre;SuFu−/loxP. Asterisks indicate background bands. Scale bar: (in A) A–D, 200 μm (see also supplemental Fig. S3, available at www.jneurosci.org as supplemental material).

Restoring GLI3R rescues the SuFu-deficient phenotype

To investigate the functional contribution of GLI3R in mediating the effects of SuFu on cerebellar morphogenesis, we generated mice that express GLI3R in the absence of SuFu. Using the Gli3Δ699 knock-in allele, which results in a premature termination of Gli3 transcription and obligate expression of GLI3R (Böse et al., 2002), we increased the levels of GLI3R in Hoxb.7-Cre;SuFu−/loxP mutants (Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699). Remarkably, Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 mice displayed a body size comparable to that of control littermates and normal motor coordination (Fig. 8A,B). MRI analysis revealed that Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella were largely normal in gross morphology and foliation and consisted of a vermis flanked by two hemispheres as seen in control cerebella (Fig. 8D,E). Histological analysis revealed a significant restoration of normal cytoarchitecture (Fig. 8F–I). In contrast to Hoxb.7-Cre;SuFu−/loxP cerebella (Fig. 2I,J), Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella were characterized by foliation, albeit fewer folia, and the presence of distinct cell layers that were properly organized. However, rescue of patterning in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 mice was not complete and small dysplastic regions were observed in variable locations throughout the cerebellum. In addition, analysis of Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 embryos indicated a significant rescue of the mid-hindbrain patterning defects and delayed cerebellar formation during embryogenesis (supplemental Fig. S4A–F, available at www.jneurosci.org as supplemental material). These findings suggest that expression of GLI3R in the absence of SUFU can significantly rescue the morphological and patterning defects observed in Hoxb.7-Cre;SuFu−/loxP mutants.

Figure 8.
Figure 8.

GLI3R significantly rescues reduced body size, abnormal cerebellar gross morphology and patterning defects. A, Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 mutants (top) are comparable to control littermates (bottom two) in body size. B, Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 mutants display normal motor coordination (Fig. 2C,D). C, A whole-mount image of a Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 brain shows a rescue in cerebellar gross morphology, although distinct folia and demarcation between the vermis and hemispheres are not completely restored. D, E, MRI three-dimensional volume and slice images of control (left) and Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 (right) brains reveal significant restoration of normal cerebellar morphology by GLI3R. F–I, Severe patterning defects caused by loss of SUFU (Fig. 2I–N) are significantly rescued. All cell layers are clearly detected in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 mutants and laminar organization is comparable to control cerebella at P14 and P21 (G, I). Note that slightly abnormal cellular organization is detected in some regions. IGL, Inner granule layer; ML, molecular layer; PCL, Purkinje cell layer. Scale bars: (in F) C, 2.5 mm; F, H, 500 μm; G, I, 62.2 μm (see also supplemental Fig. S4, available at www.jneurosci.org as supplemental material).

To characterize the effect of GLI3R restoration on cell differentiation, we analyzed expression of cell-specific markers. While calbindin staining was barely detectable in Hoxb.7-Cre;SuFu−/loxP mutants and Purkinje cells lacked the characteristic dendritic morphology at P7, in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699, calbindin-positive Purkinje cells were clearly present at P7. However, at this stage, Purkinje cells exhibited dendrites that were slightly less elaborate than control cerebella (Fig. 9A–C, green). Interestingly, by P14, Purkinje cells were indistinguishable from those observed in control cerebella and indeed, had matured at an earlier stage than that observed in Hoxb.7-Cre;SuFu−/loxP mutants (Fig. 9D–I). Further, quantitative analysis demonstrated a significant decrease in the number of Purkinje cells in Hoxb.7-Cre;SuFu−/loxP cerebella compared with control cerebella (*p = 0.03) and restoration of normal Purkinje cell number in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella (supplemental Fig. S2B, available at www.jneurosci.org as supplemental material). In P7 Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella, expression of GFAP, a marker of Bergmann glia and astrocytes, was comparable to control cerebella whereas a significant downregulation was observed in Hoxb.7-Cre;SuFu−/loxP cerebella (Fig. 9A–C, red). By P14, properly aligned Bergmann glia cell somas and radial fibers were present in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 mutants, although ectopic Bergmann glia cell somas were located in the inner granule layer as revealed by BLBP staining (Fig. 9F,I,L). In contrast, Hoxb.7-Cre;SuFu−/loxP Bergmann glia exhibited abnormal cell body organization characterized by a failure to align with Purkinje cells, and defective radial fibers at P14 and P21, indicating a rescue of the defective cell differentiation by GLI3R (Fig. 9E,H,K). Double staining of PAX6 and NeuN, markers of immature migrating granule cells and mature granule cells, respectively, revealed that at P14, expression of both markers in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella was comparable to control cerebella (Fig. 9J,L). Expression of PAX6 was upregulated in the inner granule layer of Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 mutants compared with control cerebella, suggesting a slight increase in the proportion of cells in the inner granule layer that were not fully mature. In contrast, PAX6 expression was significantly enhanced in the presumptive EGL of Hoxb.7-Cre;SuFu−/loxP cerebella, which resulted in a marked proportional decrease in the number of NeuN-expressing mature granule cells (Fig. 9K). By P21, however, immature granule cells had eventually differentiated and all granule cells expressed NeuN in Hoxb.7-Cre;SuFu−/loxP cerebella (supplemental Fig. S5A–C, available at www.jneurosci.org as supplemental material). In addition, a significant rescue of organization of PAX6-expressing cells in the EGL and upregulation of PAX2-expressing GABAergic interneuron precursors was observed in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 mutants during embryogenesis (supplemental Fig. S4G,H, available at www.jneurosci.org as supplemental material). Together, these data indicate that restoration of GLI3R in the absence of SUFU significantly rescues the abnormal cell differentiation of all major cell types observed in SuFu-deficient cerebella.

Figure 9.
Figure 9.

GLI3R rescues delayed cell differentiation in SuFu-deficient cerebella. A–I, Calbindin staining (green) reveals a complete restoration of normal Purkinje cell morphology in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella by P14 (C, F). Hoxb.7-Cre;SuFu−/loxP Purkinje cells show decreased calbindin expression and are abnormal in dendritic morphology at P7 and P14 (B, E). By P21, Purkinje cells in Hoxb.7-Cre;SuFu−/loxP cerebella are comparable to those observed in control and Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella (G–I). GFAP expression (red), a marker of Bergmann glia, in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella is comparable to that observed in control cerebella as early as P7 (C, F, I). In Hoxb.7-Cre;SuFu−/loxP mutants, Bergmann glia are severely disorganized and radial fibers are misaligned (B, E, H). J–L, BLBP-positive Bergmann glia cell somas (white arrowheads) are severely disorganized and fail to alight with Purkinje cells (yellow asterisks) in Hoxb.7-Cre;SuFu−/loxP cerebella. Normal Bergmann glia organization is restored with the exception of some Bergmann glia cell somas that are detected in the inner granule layer (IGL) in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella. M–O, PAX6 and NeuN double staining demonstrates normal expression of the granule cell markers in Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella (L). Note that PAX6 expression is slightly upregulated in the IGL of Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella. In Hoxb.7-Cre;SuFu−/loxP cerebella, the proportion of PAX6-expressing cells in the presumptive EGL to NeuN-expressing cells in the IGL is significantly increased compared with control and Hoxb.7-Cre;SuFu−/loxP;Gli3Δ699 cerebella (K). ML, Molecular layer. Scale bars: (in G) A–I, 200 μm; J–L, 100 μm; (in M) M–O, 200 μm (see also supplemental Figs. S4, S5, available at www.jneurosci.org as supplemental material).

Discussion

While the essential role of SuFu during mammalian development has been clearly demonstrated in SuFu germline knock-out mice, the functions of SuFu in the development of specific tissues after E9.5, the time point at which SuFu-null embryos die, remain unknown. By generating conditional knock-out mice in which SuFu deficiency is targeted to the cerebellum and some parts of the midbrain, we have uncovered novel functions of SuFu and SHH signaling in cerebellar morphogenesis. Collectively, our findings suggest that (1) SuFu is required for proper mid-hindbrain patterning, (2) SuFu controls cerebellar patterning by regulating cell differentiation and migration, (3) SuFu regulates the localization and level of SHH signaling and the levels of GLIs, GLI3R in particular, and (4) GLI3R partially mediates SuFu functions during cerebellar morphogenesis.

Our analyses show that the striking cerebellar dysplasia observed in postnatal Hoxb.7-Cre;SuFu−/loxP cerebella is associated with defective mid-hindbrain patterning. The expression domain of Fgf8 is shifted anteriorly and overlaps with that of Otx2 in the absence of SuFu, whereas the two expression domains are mutually exclusive with clear boundaries in wild-type embryos. This finding suggests that while the specification of the midbrain remains intact in the absence of SuFu, the anterior shift of the isthmus is the probable cause of the aberrant mid-hindbrain patterning in mutant mice. Previous studies suggest that the dose of FGF8 released in the isthmus is critical to the mechanism by which different regions of the mid-hindbrain are induced. Two of the various FGF8 protein isoforms, FGF8a and FGF8b, are known to be expressed in the isthmus (Blunt et al., 1997; Sato et al., 2001). FGF8b induces cerebellar tissue whereas FGF8a induces midbrain structures (Lee et al., 1997; Liu et al., 1999, 2003; Sato et al., 2001). Moreover, mutants that express low levels of FGF8 only develop the superior colliculus and lateral cerebellum structures (Chi et al., 2003). Thus, in light of the finding that the source of FGF8 signaling is shifted further away from the presumptive cerebellar primordium in Hoxb.7-Cre;SuFu−/loxP mutants, it is likely that levels of FGF8 are reduced in comparison with wild-type tissue. This can result in a prolonged period of time for a sufficient amount of FGF8 protein to accumulate and diffuse posteriorly and induce cerebellum tissue, causing delayed formation of the cerebellar primordium. Also, an expanded domain of FGF8 signaling activity between the isthmus and the presumptive cerebellum likely results in an enlarged hindbrain, consistent with the expanded isthmic tissue between the midbrain and the cerebellum seen in Hoxb.7-Cre;SuFu−/loxP mutants. However, since we did not observe induction of cerebellum tissue in the region immediately adjacent to the shifted isthmus, it is unlikely that the level of FGF8 signaling is the only factor that determines specification of the cerebellum. Expression of specific genes before activation of FGF8 signaling in the isthmus may predetermine a cerebellar fate. Once the isthmus has been established, a proper level of FGF8 signaling may be required for the neuroepithelium of a cerebellar fate to be specified and differentiate into cerebellar precursors.

Although the role of SuFu in mid-hindbrain development has never been investigated previously, GLI3R has been implicated in regulation of Fgf8 expression. Also, numerous studies suggest an essential role for SUFU in generation and/or promotion of GLI3R (Cheng and Bishop, 2002; Svärd et al., 2006; Kise et al., 2009). GLI3R is required for proper restriction of Fgf8 expression and tissue growth in the mid-hindbrain (Aoto et al., 2002). Our data indicate that GLI3R is undetectable in the absence of SUFU and restoration of GLI3R can partially rescue the abnormal mid-hindbrain patterning observed in Hoxb.7-Cre;SuFu−/loxP embryos. This indicates that GLI3R mediates some of the SuFu functions during mid-hindbrain specification. However, we observe a shift in Fgf8 expression rather than an expansion as observed in Gli3-null mutants (Blaess et al., 2008). This novel finding suggests an additional function of SuFu distinct from GLI3R during mid-hindbrain patterning. Also, Hoxb.7-Cre;SuFu−/loxP mutants exhibit anterior–posterior patterning defects rather than the tissue overgrowth or exencephaly observed in Hoxb.7-Cre;Ptc1−/loxP mutants caused by ligand-independent activation of SHH signaling. Since Hoxb.7-Cre;Ptc1−/loxP mutants lack apparent mid-hindbrain defects and exhibit normal timing of cerebellar formation, the function of Ptc1 is likely limited to controlling cell proliferation by constitutively suppressing SHH activity in the low SHH region of the neural tube. Thus, SuFu likely plays a role that is distinct from that of GLI3R or Ptc1 by regulating the localization of Fgf8 expression during mid-hindbrain patterning. It is presently unknown whether SUFU has HH signaling-independent functions, if any, and to what extent GLI3R mediates SUFU functions during mid-hindbrain development.

A theme common to all cell lineages investigated here is a delay in specification and differentiation in the absence of SuFu. The fact that a delay in differentiation is observed regardless of the cell type indicates that SuFu plays a role in all cerebellar precursors in determining the timing of specification. Delayed cerebellar specification during mid-hindbrain patterning caused by aberrant Fgf8 expression in the absence of SuFu may in turn delay the transition of neuroepithelial cells to cerebellar neural precursors, resulting in delayed differentiation into lineage-specific precursors. Our data show that SUFU is normally expressed in radial glia precursors along the ventricular zone in vivo. In vitro data demonstrated enhanced gliogenesis versus neurogenesis in the absence of SUFU, suggesting that SUFU in radial glia precursors functions to promote neuronal differentiation during early stages of cerebellar morphogenesis. It is also possible that SuFu coordinates the timing of expression of genes that promote differentiation in a coordinate manner. Extensive coordination and interactions between cell types during cerebellar morphogenesis have been described in depth. In particular, it is known that migration of Purkinje cells and Bergmann glia are closely correlated in a spatial and temporal manner. Thus, defective timing of differentiation of one or more cell types can have a global negative impact on differentiation and migration of other cell types, and ultimately, on cerebellar patterning. Intriguingly, this also raises the question whether SuFu controls differentiation in all cerebellar cell types or the abnormal differentiation in all major cell types observed in Hoxb.7-Cre;SuFu−/loxP mutants is a result of non-cell-autonomous effects of SuFu deletion. Further investigation using cell lineage-specific deletion of SuFu is needed to determine the cell-autonomous effects of SuFu.

The phenotype we observe in Hoxb.7-Cre;SuFu−/loxP mutants is similar to that in mice with primary defects in radial glia. Pten conditional knock-out in the cerebellum leads to cerebellar mispatterning, severe disorganization and abnormal morphology of Bergmann glia and granule cell migration defects (Yue et al., 2005). Targeted deletion of Fgf9 results in abnormal Bergmann glia scaffold formation, impaired granule cell migration and defective Purkinje cell maturation, as observed in SuFu-deficient cerebella (Lin et al., 2009). The defects observed in Pten and Fgf9 mutant mice suggest that abnormal Bergmann glia differentiation may be a primary cause of the dysplasia observed in Hoxb.7-Cre;SuFu−/loxP mice.

SHH signaling has been extensively studied in the context of cerebellar development for its role in medulloblastoma as well as during normal cerebellar morphogenesis (Goodrich et al., 1997; Dahmane and Ruiz i Altaba, 1999; Kenney et al., 2003; Corrales et al., 2004, 2006; Lewis et al., 2004; Yang et al., 2008). In Hoxb.7-Cre;SuFu−/loxP cerebella, the misexpression and absence of Shh in the external granule layer and the Purkinje cell layer, respectively, suggests that abnormal or delayed differentiation of Purkinje cells likely leads to a failure of Purkinje cells to provide SHH signaling to granule cell precursor at the right time. The upregulation and misexpression of Shh in granule cell precursors in Hoxb.7-Cre;SuFu−/loxP mice may reflect the effect of ligand-independent activation of SHH signaling in the absence of SUFU-mediated negative regulation of the pathway. Further, the essential loss of GLI3R and the presence of full-length GLI activators in Hoxb.7-Cre;SuFu−/loxP cerebellar protein lysates suggests that GLI activator activity may be exaggerated as enhanced SHH signaling activity due to the increased GLI activator to repressor ratio. These results are consistent with a recent study that demonstrated a role for SUFU in stabilizing GLI activator proteins (Chen et al., 2009). Intriguingly, the rescue of the SuFu-deficient cerebellar phenotype by GLI3R indicates that GLI3R, unlike GLI activators, does not require SUFU for stabilization, consistent with a recent study that in which the SuFu−/− neural tube phenotype was rescued by homozygous expression of Gli3Δ699 (Wang et al., 2010). The apparent dual and complex functions of SUFU in the HH signaling pathway are likely context-dependent, and further investigation is required to gain better insights into the mechanism by which SUFU regulates HH signaling.

Our studies reveal novel and essential roles for SuFu during mid-hindbrain specification and cerebellar morphogenesis. Data shown here demonstrate that SuFu is a key regulator of mid-hindbrain patterning, cerebellar cell differentiation, and SHH signaling. We show that GLI3R acts as a downstream effector to mediate the functions of SuFu during cerebellar morphogenesis. These results provide a basis for further investigation of the autonomous functions of SuFu during mid-hindbrain patterning, lineage-specific cell differentiation, and cell–cell interactions during development of the cerebellum.

Footnotes

  • This work was supported by a Hospital for Sick Children Graduate Training Award (to J.J.K.) and a Canada Research Chair (to N.D.R.). We thank Andree Gauthier Fisher, David Tsui, Freda Miller, Jason Cain, and Lin Chen for technical assistance and helpful advice. We also thank Vijitha Thanabalasingam (Hui laboratory) for providing anti-SUFU antibody.

  • Correspondence should be addressed to Norman D. Rosenblum, Division of Nephrology, The Hospital for Sick Children, 555 University Avenue, Toronto, ON M5G 1X8, Canada. norman.rosenblum{at}sickkids.ca

References

  1. ↵
    1. Aoto K,
    2. Nishimura T,
    3. Eto K,
    4. Motoyama J
    (2002) Mouse GLI3 regulates Fgf8 expression and apoptosis in the developing neural tube, face, and limb bud. Dev Biol 251:320–332.
    OpenUrlCrossRefPubMed
  2. ↵
    1. Barnfield PC,
    2. Zhang X,
    3. Thanabalasingham V,
    4. Yoshida M,
    5. Hui CC
    (2005) Negative regulation of Gli1 and Gli2 activator function by Suppressor of fused through multiple mechanisms. Differentiation 73:397–405.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Blaess S,
    2. Corrales JD,
    3. Joyner AL
    (2006) Sonic hedgehog regulates Gli activator and repressor functions with spatial and temporal precision in the mid/hindbrain region. Development 133:1799–1809.
    OpenUrlAbstract/FREE Full Text
  4. ↵
    1. Blaess S,
    2. Stephen D,
    3. Joyner AL
    (2008) Gli3 coordinates three-dimensional patterning and growth of the tectum and cerebellum by integrating Shh and Fgf8 signaling. Development 135:2093–2103.
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Blunt AG,
    2. Lawshé A,
    3. Cunningham ML,
    4. Seto ML,
    5. Ornitz DM,
    6. MacArthur CA
    (1997) Overlapping expression and redundant activation of mesenchymal fibroblast growth factor (FGF) receptors by alternatively spliced FGF-8 ligands. J Biol Chem 272:3733–3738.
    OpenUrlAbstract/FREE Full Text
  6. ↵
    1. Böse J,
    2. Grotewold L,
    3. Rüther U
    (2002) Pallister-Hall syndrome phenotype in mice mutant for Gli3. Hum Mol Genet 11:1129–1135.
    OpenUrlAbstract/FREE Full Text
  7. ↵
    1. Cain JE,
    2. Islam E,
    3. Haxho F,
    4. Chen L,
    5. Bridgewater D,
    6. Nieuwenhuis E,
    7. Hui CC,
    8. Rosenblum ND
    (2009) GLI3 repressor controls nephron number via regulation of Wnt11 and Ret in ureteric tip cells. PLoS One 4:e7313.
    OpenUrlCrossRefPubMed
  8. ↵
    1. Chen MH,
    2. Wilson CW,
    3. Li YJ,
    4. Law KK,
    5. Lu CS,
    6. Gacayan R,
    7. Zhang X,
    8. Hui CC,
    9. Chuang PT
    (2009) Cilium-independent regulation of Gli protein function by Sufu in Hedgehog signaling is evolutionarily conserved. Genes Dev 23:1910–1928.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Cheng SY,
    2. Bishop JM
    (2002) Suppressor of Fused represses Gli-mediated transcription by recruiting the SAP18-mSin3 corepressor complex. Proc Natl Acad Sci U S A 99:5442–5447.
    OpenUrlAbstract/FREE Full Text
  10. ↵
    1. Chi CL,
    2. Martinez S,
    3. Wurst W,
    4. Martin GR
    (2003) The isthmic organizer signal FGF8 is required for cell survival in the prospective midbrain and cerebellum. Development 130:2633–2644.
    OpenUrlAbstract/FREE Full Text
  11. ↵
    1. Corrales JD,
    2. Rocco GL,
    3. Blaess S,
    4. Guo Q,
    5. Joyner AL
    (2004) Spatial pattern of sonic hedgehog signaling through Gli genes during cerebellum development. Development 131:5581–5590.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Corrales JD,
    2. Blaess S,
    3. Mahoney EM,
    4. Joyner AL
    (2006) The level of sonic hedgehog signaling regulates the complexity of cerebellar foliation. Development 133:1811–1821.
    OpenUrlAbstract/FREE Full Text
  13. ↵
    1. Dahmane N,
    2. Ruiz i Altaba A
    (1999) Sonic hedgehog regulates the growth and patterning of the cerebellum. Development 126:3089–3100.
    OpenUrlAbstract
  14. ↵
    1. Ellis T,
    2. Smyth I,
    3. Riley E,
    4. Graham S,
    5. Elliot K,
    6. Narang M,
    7. Kay GF,
    8. Wicking C,
    9. Wainwright B
    (2003) Patched 1 conditional null allele in mice. Genesis 36:158–161.
    OpenUrlCrossRefPubMed
  15. ↵
    1. Fuccillo M,
    2. Joyner AL,
    3. Fishell G
    (2006) Morphogen to mitogen: the multiple roles of hedgehog signalling in vertebrate neural development. Nat Rev Neurosci 7:772–783.
    OpenUrlCrossRefPubMed
  16. ↵
    1. Goodrich LV,
    2. Milenković L,
    3. Higgins KM,
    4. Scott MP
    (1997) Altered neural cell fates and medulloblastoma in mouse patched mutants. Science 277:1109–1113.
    OpenUrlAbstract/FREE Full Text
  17. ↵
    1. Kenney AM,
    2. Cole MD,
    3. Rowitch DH
    (2003) Nmyc upregulation by sonic hedgehog signaling promotes proliferation in developing cerebellar granule neuron precursors. Development 130:15–28.
    OpenUrlAbstract/FREE Full Text
  18. ↵
    1. Kise Y,
    2. Morinaka A,
    3. Teglund S,
    4. Miki H
    (2009) Sufu recruits GSK3beta for efficient processing of Gli3. Biochem Biophys Res Commun 387:569–574.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Kogerman P,
    2. Grimm T,
    3. Kogerman L,
    4. Krause D,
    5. Undén AB,
    6. Sandstedt B,
    7. Toftgård R,
    8. Zaphiropoulos PG
    (1999) Mammalian suppressor-of-fused modulates nuclear-cytoplasmic shuttling of Gli-1. Nat Cell Biol 1:312–319.
    OpenUrlCrossRefPubMed
  20. ↵
    1. Lee SM,
    2. Danielian PS,
    3. Fritzsch B,
    4. McMahon AP
    (1997) Evidence that FGF8 signalling from the midbrain-hindbrain junction regulates growth and polarity in the developing midbrain. Development 124:959–969.
    OpenUrlAbstract
  21. ↵
    1. Lee Y,
    2. Kawagoe R,
    3. Sasai K,
    4. Li Y,
    5. Russell HR,
    6. Curran T,
    7. McKinnon PJ
    (2007) Loss of suppressor-of-fused function promotes tumorigenesis. Oncogene 26:6442–6447.
    OpenUrlCrossRefPubMed
  22. ↵
    1. Lewis PM,
    2. Gritli-Linde A,
    3. Smeyne R,
    4. Kottmann A,
    5. McMahon AP
    (2004) Sonic hedgehog signaling is required for expansion of granule neuron precursors and patterning of the mouse cerebellum. Dev Biol 270:393–410.
    OpenUrlCrossRefPubMed
  23. ↵
    1. Lin Y,
    2. Chen L,
    3. Lin C,
    4. Luo Y,
    5. Tsai RY,
    6. Wang F
    (2009) Neuron-derived FGF9 is essential for scaffold formation of Bergmann radial fibers and migration of granule neurons in the cerebellum. Dev Biol 329:44–54.
    OpenUrlCrossRefPubMed
  24. ↵
    1. Liu A,
    2. Losos K,
    3. Joyner AL
    (1999) FGF8 can activate Gbx2 and transform regions of the rostral mouse brain into a hindbrain fate. Development 126:4827–4838.
    OpenUrlAbstract
  25. ↵
    1. Liu A,
    2. Li JY,
    3. Bromleigh C,
    4. Lao Z,
    5. Niswander LA,
    6. Joyner AL
    (2003) FGF17b and FGF18 have different midbrain regulatory properties from FGF8b or activated FGF receptors. Development 130:6175–6185.
    OpenUrlAbstract/FREE Full Text
  26. ↵
    1. Mo R,
    2. Freer AM,
    3. Zinyk DL,
    4. Crackower MA,
    5. Michaud J,
    6. Heng HH,
    7. Chik KW,
    8. Shi XM,
    9. Tsui LC,
    10. Cheng SH,
    11. Joyner AL,
    12. Hui C
    (1997) Specific and redundant functions of Gli2 and Gli3 zinc finger genes in skeletal patterning and development. Development 124:113–123.
    OpenUrlAbstract
  27. ↵
    1. Préat T
    (1992) Characterization of Suppressor of fused, a complete suppressor of the fused segment polarity gene of Drosophila melanogaster. Genetics 132:725–736.
    OpenUrlPubMed
  28. ↵
    1. Sato T,
    2. Araki I,
    3. Nakamura H
    (2001) Inductive signal and tissue responsiveness defining the tectum and the cerebellum. Development 128:2461–2469.
    OpenUrlAbstract/FREE Full Text
  29. ↵
    1. Soriano P
    (1999) Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat Genet 21:70–71.
    OpenUrlCrossRefPubMed
  30. ↵
    1. Svärd J,
    2. Heby-Henricson K,
    3. Persson-Lek M,
    4. Rozell B,
    5. Lauth M,
    6. Bergström A,
    7. Ericson J,
    8. Toftgård R,
    9. Teglund S
    (2006) Genetic elimination of Suppressor of fused reveals an essential repressor function in the mammalian Hedgehog signaling pathway. Dev Cell 10:187–197.
    OpenUrlCrossRefPubMed
  31. ↵
    1. Taylor MD,
    2. Liu L,
    3. Raffel C,
    4. Hui CC,
    5. Mainprize TG,
    6. Zhang X,
    7. Agatep R,
    8. Chiappa S,
    9. Gao L,
    10. Lowrance A,
    11. Hao A,
    12. Goldstein AM,
    13. Stavrou T,
    14. Scherer SW,
    15. Dura WT,
    16. Wainwright B,
    17. Squire JA,
    18. Rutka JT,
    19. Hogg D
    (2002) Mutations in SUFU predispose to medulloblastoma. Nat Genet 31:306–310.
    OpenUrlCrossRefPubMed
  32. ↵
    1. Wang C,
    2. Pan Y,
    3. Wang B
    (2010) Suppressor of fused and Spop regulate the stability, processing and function of Gli2 and Gli3 full-length activators but not their repressors. Development 137:2001–2009.
    OpenUrlAbstract/FREE Full Text
  33. ↵
    1. Yang ZJ,
    2. Ellis T,
    3. Markant SL,
    4. Read TA,
    5. Kessler JD,
    6. Bourboulas M,
    7. Schüller U,
    8. Machold R,
    9. Fishell G,
    10. Rowitch DH,
    11. Wainwright BJ,
    12. Wechsler-Reya RJ
    (2008) Medulloblastoma can be initiated by deletion of Patched in lineage-restricted progenitors or stem cells. Cancer Cell 14:135–145.
    OpenUrlCrossRefPubMed
  34. ↵
    1. Yue Q,
    2. Groszer M,
    3. Gil JS,
    4. Berk AJ,
    5. Messing A,
    6. Wu H,
    7. Liu X
    (2005) PTEN deletion in Bergmann glia leads to premature differentiation and affects laminar organization. Development 132:3281–3291.
    OpenUrlAbstract/FREE Full Text
  35. ↵
    1. Zhao H,
    2. Kegg H,
    3. Grady S,
    4. Truong HT,
    5. Robinson ML,
    6. Baum M,
    7. Bates CM
    (2004) Role of fibroblast growth factor receptors 1 and 2 in the ureteric bud. Dev Biol 276:403–415.
    OpenUrlCrossRefPubMed
Back to top

In this issue

Email
Print
View Full Page PDF
Citation Tools
Respond to this article
Suppressor of Fused Controls Mid-Hindbrain Patterning and Cerebellar Morphogenesis via GLI3 Repressor
Jinny J. Kim, Paul S. Gill, Lianne Rotin, Matthijs van Eede, R. Mark Henkelman, Chi-Chung Hui, Norman D. Rosenblum
Journal of Neuroscience 2 February 2011, 31 (5) 1825-1836; DOI: 10.1523/JNEUROSCI.2166-10.2011
Twitter logo Facebook logo Mendeley logo

Jump to section

Responses to this article

Respond to this article

Jump to comment:

No eLetters have been published for this article.