Hedgehog signaling is involved in development of the neocortex

The secreted protein sonic hedgehog (Shh) is involved in various aspects of embryogenesis. In the mammalian central nervous system, Shh plays an important role in ventral patterning along the entire neuraxis, including the telencephalon (Agarwala et al.,2001; Echelard et al.,1993; Ericson et al.,1995; Roelink et al.,1994; Ruiz i Altaba et al.,1995; Watanabe and Nakamura,2000). GABAergic interneurons, which comprise 20-30% of cortical neurons, are important for both development and function of the cerebral neocortex (Anderson et al.,2002; Fuccillo et al.,2004; Gulacsi and Anderson,2006; Xu et al.,2005). GABAergic interneurons have been shown to migrate tangentially into the developing dorsal pallium from medial ganglionic eminence (MGE) of the ventral regions(Anderson et al., 2002; Anderson et al., 2001; Lavdas et al., 1999; Wichterle et al., 1999; Wichterle et al., 2001).

Previous studies have also illustrated that Shh is expressed in dorsal parts of three major brain structures, i.e. the neocortex of the telencephalon, tectum and cerebellum, at perinatal stages(Dahmane et al., 2001; Ruiz i Altaba et al., 2002). The common feature of these three structures is layer formation parallel to the pial surface. In the developing cerebellum, Shh produced by Purkinje neurons promotes proliferation of granule neuron progenitors and induces differentiation of Bergmann glia. Shh signaling coordinates development of cortical cell types in growth and patterning of the cerebellum(Dahmane and Ruiz i Altaba,1999; Wallace,1999; Wechsler-Reya and Scott,1999).

Similar molecular mechanisms might be applied to development of the cerebral neocortex. Indeed, it has also been reported that Shh is a mitogen for the neocortex and tectal progenitors, and that it modulates cell proliferation in vitro (Dahmane et al.,2001; Palma and Ruiz i Altaba,2004). In addition, Machold, Lai and Palma et al. have reported that Shh signaling controls stem cell behavior in adult brains(Machold et al., 2003; Lai et al., 2003; Palma et al., 2005). In Shh-null mutants, specification of ventral cell populations and general brain growth are all affected(Chiang et al., 1996). The telencephalon of Shh mutants forms as a single fused vesicle(holoprosencephaly) that is strongly dorsalized(Chiang et al., 1996; Rallu et al., 2002). Therefore, the roles of Shh in corticogenesis cannot be elucidated by using the conventional Shh knockout mice.

Recently, the roles of Shh signaling in initial patterning of the telencephalon have been analyzed by using Smoothened (Smo)conditional knockout mice (Fuccillo et al., 2004). Smo encodes a membrane protein that transduces Hh signaling into the cytoplasm. For example, Cre recombinase under control of FoxG1 regulatory sequences (in telencephalic neural precursors) was used to inactivate Smo conditionally by embryonic day(E) 9.5 in the whole telencephalon and initial patterning of the ventral telencephalon failed. However, corticogenesis in the dorsal pallium appeared normal (Fuccillo et al.,2004). Nestin-Cre;Smo conditional knockouts lost Shh signaling by E12.5(Machold et al., 2003) and showed grossly normal patterning in contrast to FoxG1-Cre;Smo(Fuccillo et al., 2004; Machold et al., 2003). These contrasting phenotypes imply that the time period between E9.5 and E12.5 is crucial for specification of ventral characters in the telencephalon, in keeping with observation that cellular response to Shh varies over time(Fuccillo et al., 2004; Machold et al., 2003). Furthermore, Xu et al. (Xu et al.,2005) reported that the cortical thickness was decreased in Nestin-Cre; Shh mice. Lien et al.(Lien et al., 2006) reported that the cortical hyperplasia was caused by abnormal activation of hedgehog signaling in Nestin-Cre; αE-catenin mice. Taken together, these findings suggest the possibility that Shh signaling controls the progenitor cell number in the developing dorsal pallium. However, one important caveat is that Shh expression in this region has not been well established. In the embryonic dorsal pallium, Shh mRNA expression level was reported to be too `low' to be detected by in situ hybridization. Shh was detectable by only more-sensitive RT-PCR in this region from E14.5 to postnatal day 3(Dahmane et al., 2001).

In this study, we clarify the roles of Shh signaling in dorsal pallium development by using Emx1-Cre;Shh^fl/- and Emx1-Cre;Smo^fl/- conditional knockout mice. Emx1 is expressed in the dorsal pallium from embryonic stages to adulthood (Iwasato et al.,2004). The conditional knockout mice lost Shh and Smo expression very early by about E10.5 in the dorsal telencephalon. Analyses of the mutant embryos revealed decreased proliferation and neuronal differentiation and increased cell death at neurogenesis stages. The positioning of neurons in the neocortex was also affected. Our data suggest that Shh signaling controls cell cycle kinetics of cortical neural progenitor/stem cells and positioning of cortical neurons in the dorsal pallium during corticogenesis. We also report detailed localization of Shh protein in the developing dorsal pallium by immunohistochemistry. Shh was detected in neural progenitor/stem cells, Cajal-Retzius (CR) neurons, cortical plate (CP) neurons and GABAergic interneurons. The immunostaining data support the conditional knockout phenotypes.

Emx1^Cre mice(Iwasato et al., 2004) were crossed with Shh heterozygous (a gift from Dr McMahon)(St-Jacques et al., 1998) and Smo heterozygous (The Jackson Laboratory, Maine, USA)(Zhang et al., 2001) mutants to generate Emx1^Cre;Shh^+/- and Emx1^Cre;Smo^+/- males. These mice were mated with Shh^fl/fl (a gift from Dr McMahon)(Lewis et al., 2001) and Smo^fl/fl (The Jackson Laboratory, Maine, USA)(Long et al., 2001) females to generate conditional knockout mice of the informative genotypes Emx1^Cre;Shh^fl/- (Shh-CKO) or Emx1^Cre;Smo^fl/- (Smo-CKO). The Emx1^Cre allele induces recombination in neural progenitor/stem cells and neurons of the dorsal telencephalon by E10,consistent with the known properties of the dorsal telencephalon-specific enhancer of Emx1 (Iwasato et al.,2004). Animals were maintained according to the guidelines from Animal Research Committee of Kyoto University.

For frozen sections, embryos were fixed for 30 minutes for Shh immunostaining or overnight for other staining in 4% paraformaldehyde at 4°C. These embryos were then washed in phosphate-buffered saline (PBS). Frozen samples were serially sectioned at 5 μm for in situ hybridization and immunohistochemistry. For paraffin sections, embryos were fixed for 3 hours in PLP fixative at 4°C. Tissues embedded in paraffin were sectioned at 5 μm for immunohistochemistry and histological analysis.

For in vivo labeling of S-phase cells (BrdU incorporation), a single intraperitoneal injection of 50 mg/kg BrdU (Sigma B5002-1G) was made 1 hour prior to sacrificing. For birth-date analysis, the timed-mated pregnant females received a single intraperitoneal injection of CldU (MP Biomedicals 105478) and IdU (Sigma 17125-5G) at E13.5 and E15.5, respectively. Pups were allowed to develop to E18.5 and processed for CldU and IdU immunohistochemistry. The quantification of young neurons and their distribution within cortical layers was analyzed according to Molyneaux et al.(Molyneaux et al., 2005) on anatomically matched sections from each mouse (three wild-type and three mutant mice) (details can be provided on request).

Standard immunostaining procedures were used. To analyze Shh localization in the dorsal telencephalon, samples must be kept at 4°C during all procedures. Details of antibodies used in this study can be provided on request.

Section RNA in situ hybridization was performed as previously described(Ishii et al., 1997). cDNA probes used were Smo (Akiyama et al., 1997), Wnt7b(Parr et al., 1993) and Tbr1 (Bulfone et al.,1995). Digoxigenin-labeled probes were synthesized using a digoxigenin RNA labeling kit (Roche 1362372).

For introduction of pCIG-Shh expression vector(Sun et al., 2008), pregnant mice were deeply anesthetized and a ventral midline incision was made to perform in utero manipulation. The expression vector (1 μg/μl) was injected through the uterine wall into the telencephalic vesicle of each embryo (E12.5). Electroporation was performed as described previously(Saito and Nakatsuji, 2001). The electroporated embryos were analyzed at E18.5.

Using RT-PCR, it has been reported that Shh mRNA is expressed in the developing dorsal pallium, the neocortex anlage, from E14.5. However, the expression level is too low to be detected by in situ hybridization(Dahmane et al., 2001). Therefore, functions of Shh in this region were unclear.

To clarify directly the roles of Shh signaling in the development of the neocortex, we knocked out Shh and Smo genes specifically in the developing dorsal telencephalon using Shh and Smoconditional alleles (Lewis et al.,2001; Long et al.,2001) and Emx1^Cre knock-in mice. The Emx1 promoter directs expression of Cre recombinase specifically in the developing dorsal telencephalon including the dorsal pallium(Iwasato et al., 2004) (see Fig. S1D in the supplementary material). We confirmed Smo mRNA expression. Although Smo mRNA was observed in the subventricular zone(SVZ)/the ventricular zone (VZ) of wild-type brains, it was not detectable specifically in the dorsal pallium of Smo-CKO embryos(Fig. 1C,C′). We also examined Shh immunoreactivity (see below).

In conventional Shh-null mutants, which show rhinocephaly(Chiang et al., 1996), most aspects of ventral and dorsal patterning in the telencephalon are abrogated,leading to difficulty in analysis of corticogenesis. By contrast, Shh-CKO and Smo-CKO embryos showed almost normal appearance(data not shown). When the brains were excised at E18.5, their morphology seemed to be grossly normal, although they were significantly smaller than wild-type brains (Fig. 1A-A″), based on measures of the rostral-caudal length, the width and the area (measured on the photos) of the telencephalon(Table 1). Furthermore, the coronal and parasagittal sections revealed the decreased size of the dorsal telencephalon, while the morphology of the ventral telencephalon was not significantly changed (Fig. 1B-B″). In addition, we performed immunostaining with anti-neuronal class III β-tubulin (Tuj1), a pan-neuronal marker, and quantified the total thickness of the dorsal pallium and the thickness of CP(Tuj1-positive layer). There was a significant reduction of the thickness of both of them in Shh- and Smo-CKO mutants at E15.5, and in Smo-CKO mutants at E18.5 (see Fig. S2 in the supplementary material). These results indicate that Shh signaling regulates neurogenesis in the dorsal telencephalon and that its absence results in a smaller dorsal pallium.

Possible mechanisms that might account for the size reduction of the dorsal telencephalon are decreased proliferation and increased cell death. To examine whether cell proliferation is decreased in the mutants, we performed an analysis of BrdU incorporation (details can be provided on request)(Fig. 2A-D). At E13.5, the BrdU-incorporation index was significantly decreased in the dorsal pallium of both Shh-CKO and Smo-CKO embryos(Fig. 2D). Interestingly, in Shh-CKO embryos at E15.5, there was no significant change in the BrdU-incorporation index (Fig. 2D), suggesting that proliferation defects were somehow rescued at this stage. However, the BrdU-incorporation index in Smo-CKO embryos was significantly reduced at E15.5 compared with wild type(Fig. 2D), consistent with inability of Smo mutant cells to respond to Shh.

To determine weather the size reduction of the mutant dorsal pallium resulted from increased cell death, we performed immunostaining for cleaved caspase 3 to identify apoptotic cells (Fig. 2E-G). In comparing with wild type, Shh-CKO and Smo-CKO embryos showed significantly increased cell death at E13.5(Fig. 2H). These results indicate that the size reduction of the dorsal pallium was caused, at least partly, by proliferation defects and increased cell death in the absence of Shh signaling.

To investigate the cause of proliferation defects, we examined cell cycle kinetics (cell cycle length and cell cycle exit) of neural progenitor/stem cells (details can be provided on request)(Chenn and Walsh, 2002). A smaller proportion of BrdU-labeled cells in Ki67-positive cells indicates a longer cell cycle length. In Shh-CKO embryos, length of cell cycle was significantly prolonged at E13.5, but not significantly changed at E15.5(Fig. 3A-B′,G) compared with wild type, consistent with the above results of proliferation analyses. However, Smo-CKO neural progenitor/stem cells divided significantly more slowly than wild-type cells at both E13.5 and E15.5(Fig. 3A,A′,C,C′,G).

As it has been reported that Hh signaling also accelerates cell cycle exit in the amphibian and fish developing retina(Locker et al., 2006), we examined how cell cycle exit was affected in the mutants (methods can be provided on request) (Chenn and Walsh,2002). The quantification of these experiments showed the significantly decreased cell cycle exit in Shh-CKO at E13.5 and in Smo-CKO at E13.5 and E15.5 (Fig. 3D-F′,H).

To distinguish which cell population (neural stem cells versus basal progenitors) was affected, we performed double-immunostaining with anti-Pax6 and anti-Ki67 antibodies at E15.5 (methods can be provided on request). Pax6-positive cells represent neural stem cells in VZ(Englund et al., 2005) and Pax6-negative/Ki67-positive cells are basal progenitors in SVZ. There was no significant difference in the Pax6-positive cell number among wild-type, Shh-CKO and Smo-CKO mice (see Fig. S3A-D in the supplementary material). In addition, no significant difference was observed in the number of Pax6-negative/Ki67-positive basal progenitors, although there was a tendency towards a decrease in both mutants (see Fig. S3E in the supplementary material). Next, we used anti-Phospho histone H3 (PHH3) antibody to count the number of M-phase cells among Pax6-positive and Pax6-negative cells. The number of Pax6-positive/Phh3-positive cells was not significantly different among wild type, Shh-CKO and Smo-CKO mutants (see Fig. S3F-I in the supplementary material). By contrast, Phh3-positive basal progenitors (Pax6 negative) showed a tendency to decrease in number in Shh-CKO and were significantly decreased in number in Smo-CKO (see Fig. S3F-H, J in the supplementary material). Altogether, the basal progenitors in both mutants seemed to be more severely reduced in number.

To further examine how loss of Shh signaling affects the character of the basal progenitors, Tbr2 immunostaining was employed (methods can be provided on request). Tbr2, another T-domain transcription factor in the developing brain, is expressed at high levels in the basal progenitors in SVZ/VZ of the dorsal pallium (Bulfone et al.,1995; Englund et al.,2005). In Shh-CKO and Smo-CKO embryos, the number of Tbr2-positive cells was significantly decreased (see Fig. S4 in the supplementary material), suggesting that the basal progenitors were decreased in SVZ/VZ of the dorsal pallium.

To examine possible defects associated with neuronal positioning and differentiation, neuronal birth-date analysis was performed (methods can be provided on request). In wild-type embryos, neurons born at E13.5(CldU-positive) settled down mainly in the deeper layers (Bin3 and Bin4 in Fig. 4D) with the peak in Bin4. Neurons born at E15.5 (IdU-positive) migrated mostly to the superficial layer of the CP (Bin5 in Fig. 4G) at E18.5. Because migration of E15.5-born neurons was not completed at E18.5, a substantial number of labeled cells remained in Bin1-3(Fig. 4G).

In Shh-CKO embryos, neurons born at E13.5 distributed mainly in Bin3 and Bin4, as in wild-type embryos(Fig. 4D,E). However, the labeled neurons in Bin4 and Bin5 (^* and ^** in Fig. 4D,E) were significantly decreased in number, implying differentiation and/or positioning defects of E13.5-born neurons. Distribution of E15.5-born neurons in Shh-CKO mice showed a similar pattern to that in wild type(Fig. 4G,H). Here again, like proliferation defects, the phenotype might be consistently rescued.

Smo-CKO embryos showed a severely impaired neuronal positioning pattern at both stages. As shown in Fig. 4F, E13.5-born neurons were scattered throughout the dorsal pallium. We observed an increased proportion of E13.5-born neurons in Bin5(^* in Fig. 4F when compared with D), suggesting reduced numbers of late-born neurons with relatively superficial position. The increased CldU index in Bin1(^**** in Fig. 4)might reflect slower cycling cells stuck in SVZ/VZ or E15.5-born neurons that failed to migrate to the superficial layer (Bin5, # in Fig. 4G,I) of CP through the early-born neurons and remained in SVZ/VZ (Bin1, Bin2 in Fig. 4I). These data suggest that Shh signaling controls the timing of neuronal differentiation and the positioning of neurons during corticogenesis.

Because timing of differentiation and migration are closely linked to the layer identities during corticogenesis, the above results prompted us to investigate specification defects in the dorsal pallium. The marginal zone(MZ), CP, IZ and SVZ/VZ were histologically distinct in the E18.5 dorsal pallium (Fig. 5A). The mutants showed a comparable structure (Fig. 5B,C). To investigate specification defects, we analyzed the expression of Wnt7b, a marker of layer VI in the dorsomedial cortex(Rubenstein et al., 1999)(Fig. 5D). In wild-type embryos, Wnt7b was expressed weakly in SVZ/VZ and intensely in the deep layer of CP and SP at E18.5, but not in IZ, making an unstained gap between CP and SVZ/VZ (Fig. 5D). In Shh-CKO embryos, this unstained gap was lost at E18.5 (Fig. 5E), suggesting that IZ cells have been changed in character and/or reduced to almost none. In Smo-CKO embryos, discrete expression in SP and weak expression in SVZ/VZ were reduced to the background level, while the expression in the deep layer of CP was comparable with wild type(Fig. 5D-F). The results imply that the character of the neural progenitor/stem cells in SVZ/VZ was affected in Smo-CKO mutants. In addition, the Wnt7b-negative upper layers of CP in both CKO brains seemed to be thinner than those in wild-type brains, consistent with the above birth-date analyses.

Tbr1 is a transcription factor gene of the T-box family that is expressed soon after cortical progenitors begin to differentiate. It is highly expressed in early-born postmitotic neurons of MZ, layer VI in CP and SP(Bulfone et al., 1995). SVZ/VZ and superficial layers of CP were weakly stained as well.(Fig. 5G). In both Shh-CKO and Smo-CKO embryos, there was a tendency towards reduction of the Tbr1-expressing layer VI relative to wild type(Fig. 5G-I). In addition, Tbr1 expression in SP (Fig. 5G,H) was not detected in Smo-CKO embryos(Fig. 5I). Weak expression in SVZ/VZ and superficial layers of CP was reduced to almost the background level, suggesting that characters of cells in these regions, especially neural progenitor/stem cells in SVZ/VZ were also affected.

To examine whether SP was correctly formed in the mutants, we performed the immunostaining using anti-Map2 antibody, which can distinguish SP neurons from CP neurons (Shinozaki et al.,2002) and anti-chondroitin sulfate (CS56), a SP marker(Gilmore et al., 1998). At E18.5, in Smo-CKO mutants, fewer Map2-positive neurons were observed in SP than in wild type (Fig. 5J,L), while Shh-CKO showed comparable Map2-positive neurons in SP with wild type (Fig. 5J,K). At E15.5, CS56 immunoreactivity of Smo-CKO was substantially decreased in SP, when compared with wild type and Shh-CKO mutants (Fig. 5M-O, arrow). These results indicate that SP was underdeveloped in the absence of Shh signaling.

The above results are consistent with the previous report about the developing retina (Locker et al.,2006), suggesting that Shh signaling promotes: (1) transition from slow cycling stem cells to fast cycling progenitors; as well as (2) cell cycle exit (neuronal differentiation) in the dorsal pallium. To confirm these roles of Shh signaling, we performed a gain-of-function experiment by in utero electroporation. The pCIG-Shh expressing vector, which expresses both Shh and EGFP, was electroporated into the dorsal pallium at E12.5. At E18.5, these transfected embryos were subjected to analyses. Both control (EGFP-expression vector only) and pCIG-Shh transfected embryos, most of EGFP-positive cells were present in CP at E18.5 (data not shown and Fig. 6A). In pCIG-Shh transfected regions, CP was expanded in a planar direction and the thickness of SVZ/VZ was extremely increased (Fig. 6B). The number of the proliferative cells with Ki67 immunoreactivity was increased on the pCIG-Shh transfected side(Fig. 6C,C′). To investigate whether cell characters are altered in each layer, we examined expression of Wnt7b and Tbr1. On the pCIG-Shh transfected side, SVZ/VZ where Wnt7b and Tbr1 were not expressed, were greatly expanded (Fig. 6D-E′). TBR2-positive basal progenitors were increased and some of them were scattered distantly from the ventricular surface(Fig. 6F′, arrows). These results demonstrated that CP-derived Shh can accelerate proliferation of neural progenitor/stem cells in SVZ/VZ and neuronal differentiation.

The above phenotypes in Shh- and Smo-CKOs demonstrated that Shh signaling plays physiological roles in the developing dorsal telencephalon. To examine Shh localization in detail, immunohistochemistry was performed. At E13.5, Shh was detected in almost all layers of the dorsal pallium, including VZ, SVZ, CP and MZ (Fig. 7A). The nuclei were not stained with anti-Shh antibody;consistent with this, Shh was in the cytoplasm and extracellular space. Although Shh was still detected in all layers at E15.5, the MZ and intermediate zone (IZ) were more abundantly stained than the other layers(Fig. 7B). To confirm that this abundantly stained region coincides with IZ, we performed double immunostaining with anti-Shh and anti-N-cadherin antibodies. N-cadherin is intensely expressed in IZ (Ajioka and Nakajima, 2005) (Fig. 7C). The Shh staining overlapped with N-cadherin in IZ at E15.5(Fig. 7C-E). As a whole, the intensity of Shh immunoreactivity in the dorsal pallium was much lower than that in other regions such as the floor plate of the telencephalon at E13.5 and E10.5 (arrow in Fig. S1A,C in the supplementary material) The intense specific staining was also observed in the floor plate and notochord at E8.5(data not shown).

To reveal which neuron cell types are positive for Shh in MZ and IZ, we performed double immunostaining using anti-Shh antibody and cell-type specific markers. One of the main populations of MZ is CR neurons, the first-born neurons that invade tangentially into MZ from the extra-cortical origins(Garcia-Moreno et al., 2007). GABAergic interneurons have been shown to migrate tangentially from MGE into the presumptive neocortex (Wonders and Anderson, 2006). The distribution of Shh-positive cells suggested that these cells include CR neurons and GABAergic interneurons. CR neurons express reelin, a large secreted protein that is associated with extracellular matrix (Ogawa et al., 1995; Schiffmann et al., 1997), and calretinin, a Ca²⁺-binding protein(Alcantara et al., 1998; del Rio et al., 1995). At E13.5, we performed double-immunostaining with anti-Shh and anti-reelin or anti-calretinin antibodies on parasagittal sections of the embryonic brains. Several Shh-immunoreactive cells in MZ were positive for reelin(Fig. 7F, see Fig. S5A-A″in the supplementary material) and calretinin(Fig. 7G, see Fig. S5B-B″in the supplementary material), suggesting that some of the Shh-positive cells in MZ are CR neurons.

GABAergic interneurons are observed in the dorsal pallium at E12-E13 in the primitive plexiform layer. During the following embryonic stages (E14-E19),most GABAergic interneurons are present in MZ, subplate (SP), IZ and SVZ(Del Rio et al., 1992). At E15.5, some of the GABA-positive cells were immunopositive for Shh in MZ of the dorsal pallium (Fig. 7H;see Fig. S5C-C″ in the supplementary material). These results indicate that Shh is produced by GABAergic interneurons. The rest of the Shh-positive cells were probably projection neurons in CP and neural progenitor/stem cells in SVZ/VZ.

In Shh-CKO, Shh was removed from most of the cells, but not completely abolished in the dorsal pallium at E13.5(Fig. 7A′; see Fig. 1SB,B′ in the supplementary material). In the ventral regions of the telencephalon including the floor plate, Shh immunoreactivity was not affected in the mutants (arrow, Fig. 1SB in the supplementary material). Therefore,conditional deletion of Shh was specific to the dorsal telencephalon. At E15.5, Shh was decreased in CP and SVZ/VZ, while abundantly observed in IZ of Shh-CKO embryos (Fig. 7B′). These Shh-positive cells in IZ are most probably interneurons derived from MGE because Shh immunoreactivity was not decreased in the ventral telencephalon of Shh-CKO embryos (see Fig. S1B in the supplementary material). Importantly, GABAergic interneurons, which were derived from MGE, were lacZ negative when the Emx1^Cre mice were crossed with ROSA26 reporter mice(Iwasato et al., 2004).

The involvement of Shh signaling for progenitor/stem cell proliferation has been demonstrated by in vitro assays(Palma et al., 2005; Palma and Ruiz i Altaba, 2004)and by analyses of the postnatal forebrain(Machold et al., 2003). Although Hh signaling has been proposed to be crucial for development of the neocortex (Ruiz i Altaba et al.,2002), this hypothesis has not been tested in vivo yet.

Our data suggest that Shh signaling regulates the cell cycle length and exit in the embryonic dorsal pallium (Fig. 3). Shh affects various checkpoints of the cell cycle by controlling activities of cyclins, cyclin-dependent kinases, etc.(Ruiz i Altaba et al., 2002). Because neuronal differentiation is closely linked to the cell cycle exit, the timing of differentiation also seemed to be regulated by Shh. In the amphibian and fish developing retina, Hh signaling speeds up the cell cycle by reducing the length of G1 and G2 phases. This leads to transition from slow cycling stem cells to fast cycling progenitors that are closer to cell cycle exit and differentiation. The cell cycle exit itself is also promoted by Hh signaling(Agathocleous et al., 2007; Locker et al., 2006). These previous reports are consistent with our results that proliferation of neural stem cells/basal progenitors and neurogenesis are decreased in our mutants(Figs 2, 3; see Figs S3, S4 in the supplementary material). In addition, the number of TBR2-positive basal progenitors was reduced in both mutants (Fig. S4 in the supplementary material), which is consistent with the fact that Shh signaling induces the transition from neural stem cells to basal progenitors. However, we cannot exclude the possibility that decreased basal progenitors and reduced neurogenesis are secondary effects of decreased proliferation and/or increased cell death.

It has also been reported that de-regulated Hh signaling in the dorsal pallium disturbed proliferation, differentiation and survival(Giros et al., 2007; Lien et al., 2006). The cortical thickness was decreased in Nestin-Cre; Shh mice(Xu et al., 2005). These reports are consistent with our results.

It has been reported that the timing of differentiation is closely related to layer identities. However, the heterochronic transplantation experiments demonstrated that early-born neurons (i.e. layers IV) could change their identity to that of late-born neurons (i.e. layers II/III) when they were transplanted into an older host (Desai and McConnell, 2000).

In Smo-CKO embryos, the distribution of cortical neurons was impaired in the dorsal pallium at E18.5(Fig. 4), implying differentiation defects. Although E13.5-born neurons resided in the more superficial layer in Smo-CKO CP, they did not seem to express the deep layer markers (Fig. 5). This result implies that E13.5-born neurons in the superficial layer changed their identity, which is consistent with the previous reports mentioned above.

Abnormal positioning of neurons in the mutants also implies migration defects. Loss of Shh signaling might lead to abnormal migrating behavior of young projection neurons in the mutants. Alternatively, Shh signaling may regulate the property of processes of radial glias. In Smo-CKO embryos, radially migrating neurons might not be able to hold the processes of neural progenitor/stem cells as appropriate scaffolds. Another possibility is that timing of migration might be retarded. In addition, selective cell death cannot be excluded as cell death was significantly increased in the CKOs(Fig. 2E-H). In either case,cell behaviors should be examined by time lapse video-microscopy and/or other methods in future studies.

In this study, we showed that Wnt7b was undetectable in neural progenitor/stem cells of Smo-CKO in SVZ/VZ at E18.5(Fig. 5). This result indicates that characters of neural progenitor/stem cells may be changed in SVZ/VZ of the dorsal pallium in this mutant. It is possible that characters of neural progenitor/stem cells were changed by disturbance of cell cycle length and timing of cell cycle exit.

Tbr1 and Wnt7b were expressed in SP of wild-type and Shh-CKO embryos (Fig. 5D,E,G,H). However, in Smo-CKO embryos, they were undetectable in SP (Fig. 5F,I). MAP2 and chondroitin sulfate expression was decreased in SP neurons(Fig. 5L,O). The SP, MZ and CP are formed by splitting of the preplate at E13.5(Allendoerfer and Shatz, 1994). Our data indicate that SP split from the preplate but was underdeveloped, and that SP neurons did not express Tbr1 and Wnt7b in the absence of Shh signaling. Perlecan is a proteoglycan that is localized in the basal lamina of the neuroepithelium during development, and is required to spread and concentrate Shh in the developing brain(Giros et al., 2007). In perlecan-null mutants, SP did not segregate from CP, suggesting roles of Shh signaling in SP formation. Therefore, our results suggest that Shh signaling is required to form the mature SP and induce Tbr1 expression in the early-born neurons of SP.

Recently, Shh mRNA expression in the dorsal pallium was demonstrated by RT-PCR and the effects of Shh on neural stem/progenitor cells from the dorsal pallium were analyzed using an in vitro assay(Dahmane et al., 2001; Palma and Ruiz i Altaba,2004). Accumulating data suggest that endogenous Shh signaling plays important roles in the neocortical development(Giros et al., 2007; Lien et al., 2006; Xu et al., 2005). However,precise localization of Shh in the dorsal pallium has not yet been revealed in vivo.

In this study, we clarified the Shh protein-positive region in the dorsal pallium by immunohistochemistry (Fig. 7). Shh was mainly localized in MZ, IZ and SVZ/VZ of the dorsal pallium from E13.5 (Fig. 7A). Our data revealed that the Shh-positive cells were neural progenitor/stem cells and CR neurons at early neurogenesis stages. At later neurogenesis stages, Shh was also detected in postmitotic neurons in CP (projection neurons) and in GABAergic interneurons in IZ(Fig. 7F-H; see Fig. S5 in the supplementary material). In the developing cerebellum, proliferation of granule cell progenitors is maintained by transient autocrine and paracrine Shh and then by Shh from Purkinje neurons. Purkinje neurons secrete Shh protein and play a central role in development of granule neurons(Dahmane and Ruiz i Altaba,1999). Our data imply that similar molecular mechanisms exist in the developing cerebral neocortex. Most probably neural progenitor/stem cells and CR neurons in the dorsal pallium secrete Shh protein to maintain the stem cell niche by autocrine and/or paracrine mechanisms during early neurogenesis. During later neurogenesis, in addition to neural progenitor/stem and CR neurons, postmitotic neurons (including projection neurons and GABAergic interneurons) secrete Shh protein, which may also regulate cell cycle kinetics.

As there were few interneurons that produce Shh at E13.5(Fig. 7A,A′), Shh-CKO showed the defects of proliferation and positioning at E13.5. By contrast, at E15.5, these phenotypes were rescued in Shh-CKO most probably by Shh secreted from many interneurons(Fig. 7B′), which were originated from MGE. This is consistent with the work of Xu et al.(Xu et al., 2005).

Interestingly, the expression level of Shh in the dorsal pallium is extremely low as mentioned above. There was in vitro evidence that `high level' Shh instead suppresses cell proliferation, when compared with relatively `low level' Shh. Palma and Ruiz i Altaba(Palma and Ruiz i Altaba,2004) have reported that proliferation of neocortical precursors showed a Shh concentration-dependent increase between 1 and 5 nM in neurosphere assay. Very intriguingly, the proliferation was significantly decreased with higher concentration of Shh (25 nM), supporting our observation.

In addition, Shh is obviously required for cell survival because cell death was significantly increased in the absence of Shh signaling(Fig. 2). The hypomorphic phenotype of the dorsal telencephalon is partly caused by increased cell death.

The previous studies have reported that Shh mutations are found in human telencephalic anomalies, including holoprocencephaly (HPE)(Wallis and Muenke, 2000). This is consistent with the HPE phenotype of Shh KO mice(Chiang et al., 1996). It is also possible that abnormal Shh signaling in the dorsal telencephalon leads to macrencephaly and microcephaly (Lien et al., 2006; Xu et al.,2005).

As neuronal differentiation is closely linked to the cell cycle exit during neurogenesis, the regulation of cell cycle kinetics is very important for proper corticogenesis. Taken together, Shh signaling might be important for the maintenance of stem cell niches in the embryonic dorsal telencephalon and allow an appropriate probability of neuronal differentiation to form the precise layered structure. In conclusion, our findings suggest that `low level' Shh signaling plays an important role in fine-tuning of cell cycle kinetics to form the proper neocortex.

We thank Dr Takuji Iwasato for Emx1-Cre knock-in mice, Dr Andrew P. McMahon for Shh mutant mice, Dr Hidenori Tabata for technical advice about birth-date analysis, Dr Yutaka Itokazu for technical advice about the BrdU incorporation assay, Dr Kazunori Nakajima for anti-Reelin antibody,Dr Yasuto Tanabe for Wnt7b and Tbr1 probes, Drs Yukiko Gotoh and Toshiyuki Ohtsuka for useful suggestions, Drs Wataru Ochiai and Takaki Miyata for technical assistance, Dr Jun Motoyama for Smo probe, and Drs David Rowitch and Masaaki Kitada for critical comments on the manuscript. This work was supported by The Japanese Ministry of Education, Culture,Sports, Science and Technology (Grant number 15689004, 16015264, 18590168) and by the Kato Memorial Bioscience Foundation.