Summary
Patterning of the dorsal-ventral (D-V) axis of the mammalian telencephalon is fundamental to the formation of distinct functional regions including the neocortex and ganglionic eminences. Morphogenetic signaling by bone morphogenetic protein (BMP), Wnt, Sonic hedgehog (Shh), and fibroblast growth factor (FGF) pathways determines regional identity along this axis. It has remained unclear, however, how region-specific expression patterns of these morphogens along the D-V axis are established, especially at the level of epigenetic (chromatin) regulation. Here we show that epigenetic regulation by Ring1, an essential Polycomb group (PcG) protein, plays a key role in formation of ventral identity in the mouse telencephalon. Deletion of the Ring1b or both Ring1a and Ring1b genes in neuroepithelial cells of the mouse embryo attenuated expression of the gene for Shh, a key morphogen for induction of ventral identity, and induced misexpression of dorsal marker genes including those for BMP and Wnt ligands in the ventral telencephalon. PcG protein–mediated trimethylation of histone H3 on lysine-27 (H3K27me3) was also apparent at BMP and Wnt ligand genes in wild-type embryos. Importantly, forced activation of Wnt or BMP signaling repressed the expression of Shh in organotypic and dissociated cultures of the early-stage telencephalon. Our results thus indicate that epigenetic regulation by PcG proteins—and, in particular, that by Ring1— confers a permissive state for the induction of Shh expression through suppression of BMP and Wnt signaling pathways, which in turn allows the development of ventral identity in the telencephalon.
Introduction
In vertebrate embryos, the telencephalon is formed at the most anterior portion of the developing central nervous system (CNS). The cerebral cortex (CTX) and ganglionic eminences (GEs) are derived from the dorsal telencephalon (the pallium) and the ventral telencephalon (the subpallium), respectively (Campbell, 2003; Hebert and Fishell, 2008). Whereas neural stem-progenitor cells (NPCs) in the CTX produce excitatory cortical neurons, astrocytes, and some late-born oligodendrocytes, those in the GE produce local neurons and glial cells that constitute the basal ganglia as well as inhibitory neurons and early-born oligodendrocytes that migrate tangentially to the CTX. Regulation of dorsal-ventral (D-V) patterning is thus fundamental to development of the telencephalon.
The regional identity of telencephalic NPCs along the D-V axis is determined by region-specific transcription factors such as Pax6 and Emx1/2 in the CTX, Gsx2 in the lateral and medial GE (LGE and MGE, respectively), and Nkx2.1 in the MGE (Corbin et al., 2003; Kroll and O’Leary, 2005; Simeone et al., 1992; Sussel et al., 1999; Wigle and Eisenstat, 2008). Mutual gene repression by Pax6 and Gsx2 contributes to establishment of the D-V boundary (Corbin et al., 2003). Neurog1 and Neurog2, proneural genes in the CTX, and Ascl1, a proneural (and oligodendrogenic) gene in the GE, are expressed according to the regional identity of NPCs in mouse embryos (Casarosa et al., 1999; Fode et al., 2000; Toresson et al., 2000).
Telencephalic regionalization along the D-V axis begins before closure of the neural tube, which occurs around embryonic day (E) 9.0 in mice, and is established before the onset of neurogenesis at ∼E10. The initial stages of D-V patterning are controlled by secreted morphogenetic signals (morphogens) that spread over various distances. The combination of the activities of different morphogens gives rise to distinct expression patterns of region-specific transcription factors in the telencephalon (Gupta and Sen, 2016; Harrison-Uy and Pleasure, 2012; Lupo et al., 2006; Sur and Rubenstein, 2005), as is also the case in the vertebrate spinal cord (Andrews et al., 2019) and in invertebrate embryos (Briscoe and Small, 2015). Bone morphogenetic proteins (BMPs), Wnt ligands, Sonic hedgehog (Shh), and fibroblast growth factor 8 (FGF8) are among the morphogens involved in D-V patterning in the mammalian telencephalon.
The dorsal midline regulates dorsal patterning of the telencephalon (Monuki et al., 2001). BMPs (BMP4, −5, −6, and −7) are secreted from the dorsal midline and paramedial neuroectoderm in the prospective forebrain (Furuta et al., 1997) and play pivotal roles in such patterning through induction of target genes including Msx1, Lmx1a, and Wnt3a (Cheng et al., 2006; Currle et al., 2005; Fernandes et al., 2007; Furuta et al., 1997; J. A. Golden et al., 1999; Hebert et al., 2002; Panchision et al., 2001). Knockout of BMP receptors thus results in loss of the dorsal-most structures of the telencephalon including the cortical hem and choroid plexus (Fernandes et al., 2007). Wnt ligands are also expressed in the dorsal region (Wnt1, −3, −3a, and −7b in the dorsal telencephalic roof plate at E9.5; Wnt2b, −3a, −5a, −7b, and 8b in the cortical hem and Wnt7a and −7b in the CTX at later stages) and contribute to aspects of dorsal patterning such as formation of the cortical hem and the CTX through induction of various transcription factors including Lef1 as well as Emx1/2, Pax6, and Gli3, respectively (Backman et al., 2005; Galceran et al., 2000; Harrison-Uy and Pleasure, 2012; Hasenpusch-Theil et al., 2012). In addition, Wnt signaling increases the activity of Lhx2, a selector gene for the CTX (Chou and Tole, 2019; Hsu et al., 2015). Suppression of the Wnt signaling pathway increases expression of ventral-specific genes throughout the dorsal pallium, indicating the importance of such signaling in dorsal patterning (Backman et al., 2005).
Shh, on the other hand, plays a major role in ventral patterning of the telencephalon (Blaess et al., 2014). Shh is secreted initially from the anterior mesendoderm or the prechordal plate (Aoto et al., 2009), then from the ventral hypothalamus, and finally from the rostroventral telencephalon including the preoptic area and MGE (Ericson et al., 1995; Fuccillo et al., 2004; Mathieu et al., 2002; Rohr et al., 2001; Shimamura et al., 1995). Shh signaling induces the expression of ventral transcription factors such as Foxa2, Nkx2.1, and Gsx2 and suppresses the repressor activity of Gli3, which is crucial for development of the dorsal telencephalon (Jeong and Epstein, 2003; Kuschel et al., 2003; Rallu et al., 2002; Shimamura and Rubenstein, 1997). FGF8 expressed in the anterior neural ridge also contributes to formation of the ventral telencephalon. FGF signaling is required for ventral expression of Shh and Nkx2.1 (Gutin et al., 2006; Shinya et al., 2001; Storm et al., 2006), and, conversely, Shh is required for maintenance of FGF8 expression at the anterior neural ridge (Hayhurst et al., 2008; Ohkubo et al., 2002; Rash and Grove, 2007).
Expression of BMP ligands and activation of BMP signaling are confined to the dorsal midline, with this confinement being critical for development of the ventral telencephalon, given that ectopic BMP signaling can suppress the expression of ventral morphogenetic factors such as Shh and FGF8 as well as that of the ventral transcription factor Nkx2.1 in the chick forebrain (J. A. Golden et al., 1999; Ohkubo et al., 2002). It is also important that Wnt signaling be confined to the dorsal pallium, given that ectopic activation of such signaling suppresses ventral specification in the developing mouse telencephalon (Backman et al., 2005). Mutual inhibition between dorsal and ventral morphogenetic factors explains in part the regional confinement of BMP and Wnt signaling (Huang et al., 2007; Storm et al., 2003). However, it has remained unclear whether any epigenetic factors (histone modifiers) participate in the establishment and maintenance of regional identity along the D-V axis of the developing telencephalon.
Polycomb group (PcG) proteins are repressive epigenetic factors that consist of two complexes, PRC1 and PRC2. These complexes catalyze the ubiquitylation of histone H2A at lysine-119 (H2AK119ub) and the trimethylation of histone H3 at lysine-27 (H3K27me3), respectively (Di Croce and Helin, 2013; Simon and Kingston, 2013). PcG proteins were first identified as transcriptional repressors of Hox genes in Drosophila melanogaster. These genes maintain regional identity along the anterior-posterior (A-P) axis in the fly embryo (Maeda and Karch, 2009). In mammals, PcG proteins also contribute to maintenance of the A-P axis during embryogenesis through repression of Hox genes in the neural tube (Chambeyron et al., 2005) as well as through that of forebrain-related genes in the midbrain (Zemke et al., 2015). Moreover, PcG proteins participate in cell subtype specification in the spinal cord (M. G. Golden and Dasen, 2012) and in the CTX in a manner dependent on temporal codes (Hirabayashi et al., 2009; Morimoto-Suzki et al., 2014; Pereira et al., 2010; Sparmann et al., 2013; Tsuboi et al., 2018). However, it has remained unknown whether PcG proteins regulate D-V patterning of the mammalian CNS including the telencephalon.
We now show that Ring1, an E3 ubiquitin ligase and essential component of PRC1 (de Napoles et al., 2004; Wang et al., 2004), is required for formation of the ventral telencephalon. Neural-specific ablation of Ring1B or of both Ring1A and Ring1B thus attenuated expression of ventral-specific genes such as Gsx2, Nkx2.1, and Ascl1 as well as increased that of dorsal-specific genes such as Pax6, Emx1, and Neurog1 in the telencephalon of mouse embryos. We found that Shh expression was markedly reduced, whereas BMP and Wnt signaling pathways were activated, in the ventral telencephalon of such Ring1B knockout (KO) or Ring1A/B double knockout (dKO) embryos. Moreover, Ring1B and H3K27me3 were found to be enriched at the promoters of several BMP and Wnt ligand genes in the telencephalon of wild-type (WT) embryos. Consistent with these results, forced activation of BMP or Wnt signaling suppressed Shh expression in explant cultures prepared from the embryonic telencephalon. Overall, our findings indicate that Ring1 establishes a permissive state for Shh expression in the ventral region of the telencephalon through suppression of BMP and Wnt signaling in this region.
Results
Deletion of Ring1 in the neuroepithelium results in morphological defects in the telencephalon
To investigate the role of PcG proteins in the early stage of mouse telencephalic development, we deleted Ring1b with the use of the Sox1-Cre transgene, which confers expression of Cre recombinase in the neuroepithelium from before E8.5 (Takashima, 2007). We confirmed that expression of the Ring1B protein in the telencephalic wall was greatly reduced in Ring1bflox/flox;Sox1-Cre (Ring1B KO) mice at E10 compared with that in Ring1bflox/flox or Ring1bflox/+ (control) mice, whereas the abundance of Ring1B in tissues outside of the telencephalic wall appeared unchanged in the Ring1B KO embryos (Figure S1A). The expression of Ring1B in the telencephalic wall was also greatly reduced in mice lacking both Ring1B and its homolog Ring1A (Ring1a−/−;Ring1bflox/flox;Sox1-Cre, or Ring1A/B dKO, mice) compared with that in Ring1a−/−;Ring1bflox/flox or Ring1a−/−;Ring1bflox/+ (Ring1A KO) mice (Figure S1B). The level of H2AK119ub (a histone modification catalyzed by Ring1) in the telencephalic wall was reduced in Ring1B KO mice and, to a greater extent, in Ring1A/B dKO mice at E10 (Figure 1A–D). These results were consistent with the notion that Ring1A and Ring1B have overlapping roles in H2A ubiquitylation and that Ring1B makes a greater contribution to this modification than does Ring1A (Simon and Kingston, 2013).
We found that Ring1b deletion with the use of Sox1-Cre resulted in a significant reduction in the size of the telencephalon at E11 (Figure 1E, F). Deletion of Ring1b at E13.5 with the use of the Nestin-CreERT2 transgene was previously found not to substantially affect the morphology or size of the telencephalon (Hirabayashi et al., 2009), indicating that Ring1B functions during the early stage of telencephalon development. The reduction in telencephalon size was also pronounced in Ring1A/B dKO mice compared with Ring1A KO mice (Figure 1G, H). The number of cells positive for the cleaved form of caspase-3 (a marker for apoptosis) in the telencephalic wall was increased in Ring1B KO mice and, to a greater extent, in Ring1A/B dKO mice at E10 (Figure 1I–K), suggesting that Ring1 is necessary for the survival of telencephalic cells during the early stage of development and that the reduction in telencephalic size induced by Ring1 deletion is due, at least in part, to the aberrant induction of apoptosis.
Deletion of Ring1 attenuates the expression of ventral-specific transcription factors in NPCs of the ventral telencephalon
We next investigated whether Ring1 deletion might affect the D-V axis of the telencephalon by determining the expression of region-specific transcription factors. We examined embryos mostly at E10 given the apparently normal telencephalic size in Ring1B KO mice at this stage. Nkx2.1 is a transcription factor specifically expressed in the MGE (Sussel et al., 1999), and we found that the expression of this protein was greatly diminished in Ring1B KO mice at E10 (Figure 2A). The dorsal border of the Nkx2.1 expression domain was thus shifted ventrally and the abundance of Nkx2.1 within this domain was also reduced by Ring1b deletion (Figure 2B, C). The extent of the loss of Nkx2.1 expression appeared greater in Ring1A/B dKO mice than in Ring1B KO mice (Figure 2D–F). In addition to its effect on Nkx2.1 expression at the protein level, Ring1b deletion resulted in marked down-regulation of Nkx2.1 mRNA in CD133+ NPCs isolated from the GE of the telencephalon at E11 (Figure 2I). These results thus indicated that Ring1 is necessary for ventral expression of the MGE marker Nkx2.1.
We also examined the expression of Gsx2, which is highly enriched in the LGE and whose mRNA is present in both the LGE and MGE (Toresson et al., 2000). Immunostaining indeed revealed the expression of Gsx2 protein within nuclei of NPCs in the LGE of control mice at E10, whereas such expression was markedly attenuated in Ring1A/B dKO mice (Figure 2G, H). Moreover, Ring1b deletion significantly reduced the level of Gsx2 mRNA in CD133+ NPCs isolated from the GE of the telencephalon at E11 (Figure 2I). These results together indicated that Ring1 plays a role in expression of ventral transcription factors in the ventral region of the telencephalon.
Deletion of Ring1 increases the expression of CTX-specific transcription factors in NPCs of the ventral telencephalon
We then investigated whether deletion of Ring1 affects the expression of dorsal-specific transcription factors in the developing telencephalon. Pax6 contributes to development of the CTX, with its expression being restricted to the dorsal pallium in mice (Corbin et al., 2003). However, we found that the expression of Pax6 extended to the ventral region of the telencephalon in Ring1B KO mice as well as in Ring1A/B dKO mice (Figure 3A, B, data not shown). Indeed, the dorsoventral gradient of Pax6 expression was shallower in Ring1B KO mice than in control mice (Figure 3C). We also examined the role of Ring1 in regulation of Emx1, another CTX-specific transcription factor (Simeone et al., 1992). Deletion of Ring1b increased the amount of Emx1 mRNA in CD133+ NPCs isolated from the GE of the telencephalon at E11 (Figure 3D). These results together indicated that Ring1 suppresses the expression of dorsal transcription factors in the ventral region of the telencephalon and thus prevents “dorsalization” of this region during the early stage of development. Of note, deletion of Ring1b with the use of the Foxg1-IRES-Cre transgene (that is, from ∼E9.0) did not appear to promote dorsalization of the ventral telencephalon (data not shown), revealing a time window for sensitivity to Ring1-dependent D-V regionalization.
Deletion of Ring1 confers dorsalized expression patterns of proneural genes in the ventral telencephalon
Given that Ring1 deletion appeared to induce dorsalization of the expression patterns of transcription factors related to NPC specification along the D-V axis, we examined the expression of proneural genes that contribute to region-specific neuronal differentiation. The basic helix-loop-helix proteins Neurog1 and Ascl1 are pallium- and subpallium-specific proneural factors, respectively (Casarosa et al., 1999; Fode et al., 2000). There was thus little overlap of Neurog1 and Ascl1 expression at the pallium-subpallium boundary of control mice at E10 (Figure 4A). However, deletion of Ring1b resulted in a ventral shift of the ventral border of Neurog1 expression and a marked overlap of Neurog1 expression with Ascl1 expression in the ventral region (Figure 4A–C, Figure S2), again suggesting that loss of Ring1 induces dorsalization of the early-stage telencephalon. Deletion of Ring1b did not obviously shift the dorsal border of Ascl1 expression (Figure 4C) but significantly reduced the level of Ascl1 within the LGE (Figure 4D). In Ring1A/B dKO mice, the border of the Neurog1+ region was also shifted ventrally (Figure 4E), and the level of Ascl1 protein in the ventral region appeared to be reduced to a greater extent than in Ring1A KO or Ring1B KO mice (Figure 4F). These results supported the notion that Ring1 plays a pivotal role in the establishment of ventral identity in the early stage of telencephalic development.
Ring1 suppresses BMP and Wnt signaling pathways in the early-stage ventral telencephalon
We next investigated whether Ring1 deletion affects the gene expression profile of NPCs in the ventral telencephalon. Transcripts isolated from CD133+ NPCs derived from the GE of the control or Ring1B KO telencephalon at E11 were subjected to RT, and the resulting cDNA was amplified with the use of the Quartz protocol (Sasagawa et al., 2013) and subjected to high-throughput sequencing analysis (Figure 5A). Differentially expressed genes were determined with the use of edgeR of the R package (McCarthy et al., 2012; Robinson et al., 2010). We identified more up-regulated genes (953) than down-regulated genes (238) in ventral NPCs from Ring1B KO mice compared with those from control mice (Figure 5B, C), consistent with the general role of PcG proteins in gene repression. Importantly, the expression of genes for dorsal-specific transcription factors such as Emx1, Emx2, and Msx1 was up-regulated, whereas that of genes for ventral-specific transcription factors such as Nkx2.1 and Olig2 (Lu et al., 2000; Takebayashi et al., 2000) was down-regulated, in the NPCs from Ring1B KO mice (Figure 5B, C, Supplementary Table 1). These results thus confirmed the role of Ring1 in suppression of the dorsalization of ventral NPCs.
To shed light on the mechanism by which Ring1 establishes (or maintains) ventral identity in ventral telencephalic NPCs, we performed KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis for both the up-regulated and down-regulated gene sets (Figure 5D, E). Among the genes whose expression was up-regulated by Ring1b deletion, pathway analysis revealed an enrichment of categories such as pathways in cancer, extracellular matrix (ECM)–receptor interaction and focal adhesion (Figure 5D). This enrichment may be due in part to derepression of protocadherin-γ and collagen family genes in Ring1B-deficient NPCs compared with control NPCs (Supplementary Table 1). Furthermore, we found that Hippo, Wnt, and transforming growth factor–β (TGF-β) signaling pathways were also enriched among the genes whose expression was up-regulated by Ring1b deletion (Figure 5D). The up-regulated genes categorized in the Hippo signaling pathway included genes related to BMP and Wnt signaling pathways (Figure 5F). Of interest, the BMP inhibitor gene Bmper was among the top 10 down-regulated genes (Figure 5C). RT-qPCR analysis also showed that the expression of Bmp4 and Id1, which are major components of BMP signaling, as well as that of Wnt7b, Wnt8b, and Axin2, which are major components of Wnt signaling, were increased in Ring1B-null NPCs compared with control NPCs (Figure 5G). These results together indicated that Ring1B suppresses BMP and Wnt signaling pathways in NPCs of the ventral telencephalon at E11.
We also monitored the activity of the Wnt signaling pathway by examining the distribution of Axin2 mRNA with in situ hybridization analysis. The highest level of Axin2 expression is normally confined to the dorsal midline (presumptive cortical hem), with a lower level of expression also occurring in the pallium (presumptive neocortex) of the early-stage telencephalon. However, the region showing the greatest abundance of Axin2 mRNA at E10 was expanded in Ring1A/B dKO mice compared with Ring1A KO mice (Figure 6A, B), suggesting that Ring1 suppresses the activity of Wnt signaling outside of the dorsal midline in the telencephalon, with the exception of the most ventral region. We also examined the distribution of Id1 protein as a marker of the activity of the BMP signaling pathway. The highest level of Id1 expression is normally confined to the dorsal midline in the early-stage telencephalon, and again deletion of Ring1a and Ring1b resulted in expansion of this region to the entire ventricular wall of the telencephalon at E10 (Figure 6C–E). These results suggested that Ring1 suppresses the expression of Id1, and therefore possibly the activity of BMP signaling, outside of the dorsal midline at the early stage of telencephalic development.
Ring1 promotes Shh expression and activates the Shh signaling pathway in the early-stage telencephalon
In contrast to the genes whose expression was up-regulated in Ring1B-null NPCs, KEGG pathway analysis revealed an enrichment of hedgehog signaling pathway among the down-regulated genes (Figure 5E). Consistent with this observation, we found that the expression levels of genes related to Shh signaling—such as Gli1, Gli2, Ptch1, and Ptch2—were significantly lower in ventral NPCs from Ring1B KO mice compared with those from control mice at E11 (Figure 7A, B), suggesting that Ring1B is essential for activation of Shh signaling in ventral NPCs at this early stage of telencephalic development.
Given the attenuated expression of Shh target genes in Ring1B-deficient ventral NPCs, we examined whether deletion of Ring1b might affect the expression of Shh. In situ hybridization analysis of mice at E10 revealed that Ring1b deletion markedly reduced the abundance of Shh mRNA (Figure 7C, D), which is normally found at the ventral midline (presumptive preoptic area) of the developing telencephalon at this stage (Figure 7C). Furthermore, Shh expression was not apparent in the telencephalon of Ring1A/B dKO embryos (Figure 7E). These results together indicated that Ring1 is required for expression of Shh, the major ventral morphogen, which might explain the overall dorsalization phenotype of the Ring1-deficient telencephalon.
It remained unclear, however, whether the down-regulation of Shh target gene expression induced by Ring1 deletion was due simply to the attenuation of Shh expression or was also due to an inability of NPCs to express these genes in response to Shh signaling. We therefore prepared in vitro cultures of telencephalic NPCs at E10 and examined their responsiveness to Shh signaling by the addition of a Smoothened agonist (SAG) for 24 h. The extent of the induction of the Shh target genes Gli1 and Ptch1 was similar for NPCs isolated from Ring1B KO mice and from control mice (Figure 8A–C), suggesting that Ring1b deletion did not substantially affect the regulation of these Shh target genes and that the regulation of Shh expression itself is crucial for Ring1-dependent ventral identity.
How does Ring1 promote Shh expression? Given the general role of PcG proteins in gene repression, it was plausible that Ring1 indirectly increases Shh expression through repression of genes whose products inhibit Shh expression. Implantation of beads soaked with recombinant BMP in the anterior neuropore of chick embryos was previously shown to inhibit Shh and Nkx2.1 expression (J. A. Golden et al., 1999; Ohkubo et al., 2002). Furthermore, forced activation of canonical Wnt signaling by expression of a stabilized form of β-catenin was found to result in repression of ventral marker genes such as Nkx2.1 in the mouse subpallium (Backman et al., 2005). It was therefore possible that activation of BMP and Wnt signaling pathways might account for the down-regulation of Shh expression in Ring1-deficient mice, although it remained unclear whether Wnt signaling alone is able to regulate the level of Shh expression. We therefore examined whether activation of Wnt signaling can reduce the level of Shh mRNA in collaboration with BMP signaling in dissociated (monolayer) cultures and explant cultures prepared from the telencephalon of WT mice at E9. The addition of an activator of canonical Wnt signaling (the glycogen synthase kinase 3 inhibitor CHIR-99021) indeed significantly reduced the abundance of Shh mRNA in the dissociated telencephalic culture and, to a greater extent, in the explant culture (Figure 8D–G) under the condition. Moreover, exposure to both BMP4 and CHIR-99021 tended to have a greater effect on the amount of Shh mRNA in the dissociated culture than did either agent alone (Figure 8E, G). The activation of BMP and Wnt signaling pathways may therefore cooperate to suppress the expression of Shh in the telencephalon at this early stage of development, consistent with the notion that the Ring1-dependent establishment of ventral identity is mediated by suppression of these signaling pathways.
BMP and Wnt ligand genes are direct targets of PcG proteins in early-stage telencephalic NPCs
Given the dysregulation of BMP and Wnt ligand gene expression induced by Ring1 deletion, we next examined whether these genes are direct targets of PcG proteins by performing chromatin immunoprecipitation (ChIP)–qPCR assays for H3K27me3, a histone modification catalyzed by PRC2, as well as for Ring1B with the telencephalon of WT mice at E9. We indeed detected significant or nearly significant deposition of H3K27me3 at the promoters of Bmp4, Bmp7, Wnt7b, and Wnt8b at levels similar to those apparent at the promoters of Hoxa1 and Hoxd3, which were examined as positive controls (Figure 9B). Ring1B was found to be enriched at the promoters of Bmp4 and Wnt8b, but not at those of Bmp7 and Wnt7b (Figure 9C), suggesting that PcG proteins directly regulate the expression of at least Bmp4 and Wnt8b in the early-stage telencephalon.
Discussion
Morphogenetic signals and their downstream transcription factors determine regional identity along the D-V axis in the developing telencephalon. Mutual inhibition between such signaling plays a pivotal role in segregation of regional identity, but the contribution of epigenetic mechanisms that control the permissiveness for transcriptional activation to this telencephalic D-V regionalization have remained largely unknown. We have now found that Ring1A and Ring1B, core components of PRC1, play an essential role in establishment of the spatial expression patterns of morphogenetic signals and transcription factors along the D-V axis and in consequent regionalization of the telencephalon at the early stage of development. Our results thus indicate that Ring1 is required for expression of Shh in the ventral telencephalon and that the ablation of Ring1 results in dorsalization of the ventral telencephalon. This dorsalization phenotype of the Ring1-deficient telencephalon is likely due in part to down-regulation of the ventralizing morphogen Shh, given that the inactivation of Shh gives rise to morphological and molecular phenotypes similar to those associated with Ring1 deletion, including the induction of a rounder shape and malformation of the dorsal midline in the telencephalic wall (Chiang et al., 1996; Rallu et al., 2002), increased apoptosis (Aoto et al., 2009), and attenuated expression of ventral transcription factors such as Nkx2.1, Gsx2, and Ascl1 (Blaess et al., 2014). Moreover, our results indicate that Ring1 is required for suppression of BMP and Wnt signaling pathways and that genes for BMP and Wnt ligands are direct targets of Ring1 and other PcG proteins. Together with the observations that ectopic activation of BMP signaling (J. A. Golden et al., 1999; Ohkubo et al., 2002) or Wnt signaling (this study) is able to suppress the transcription of Shh in the developing telencephalon, these results suggest that Ring1 suppresses BMP and Wnt signaling in the telencephalon (outside of the dorsal midline) and thereby generates a permissive state for Shh expression, which is essential for establishment of ventral identity (Figure 10). In addition to the activation of Shh expression, suppression of BMP and Wnt signaling pathways per se may contribute to the Ring1-mediated establishment of ventral identity by a Shh-independent mechanism (Figure 10).
Deletion of Ring1 not only induced dorsalization of NPC identity in the telencephalon but also resulted in aberrant expression patterns of proneural genes. Ascl1 and Neurog1 are expressed in a mutually exclusive manner in WT embryos, in part as a result of the repression of Ascl1 expression by Neurog1 and Neurog2 (Fode et al., 2000). However, we found that Ring1b deletion resulted in a marked increase in the number of cells positive for both Neurog1 and Ascl1. Given that H3K27me3 and H2AK119ub were previously shown to be deposited at the promoters of Neurog1 and Ascl1 in early-stage NPCs (Hirabayashi et al., 2009; Tsuboi et al., 2018; data not shown), PcG proteins may participate in the mutually exclusive inhibition of Neurog1 and Ascl1 expression and thereby regulate the segregation of neurogenic properties between NPCs.
Deletion of Ring1b with the use of the Nestin-CreERT2 transgene, which confers Cre expression in the entire CNS at E13.5 (Hirabayashi et al., 2009), or deletion of the gene for the histone methyltransferase Ezh2 with the Emx1-Cre transgene, which is expressed in the dorsal telencephalon from E10.5 (Gorski et al., 2002; Pereira et al., 2010), has been shown to induce neurogenesis through derepression of neurogenic genes (Tsuboi et al., 2018). With the use of the Sox1-Cre transgene, which is expressed in the neuroepithelium from before E8.5 (Takashima et al., 2007), we have now examined the role of Ring1 in the early stage of telencephalic development, before the onset of the neurogenic phase. During this early stage (for example, at E9), we did not detect promotion of neurogenesis (data not shown), suggesting that a Ring1-indpenedent mechanism is responsible for the suppression of neurogenesis at this time. Of interest, deletion of Ring1b with the use of the Foxg1-IRES-Cre transgene, which is expressed in the entire telencephalon from ∼E9.0 (Kawaguchi et al., 2016), did not appear to induce dorsalization of the ventral telencephalon (data not shown), suggesting that Ring1-mediated D-V patterning of the telencephalon takes place only during the early stage of development, although the mechanism underlying this temporal restriction remains unclear.
A key related question is how PcG proteins are recruited to specific genes in specific regions of the telencephalon at specific times. Deletion of Ezh2 in the dorsal midbrain with the use of the Wnt1-Cre transgene was previously shown to result in inhibition of Wnt signaling and to promote telencephalic identity at ∼E11.5 (that is, rostralization) (Zemke et al., 2015), in contrast to our finding that Ring1 deletion activates Wnt signaling in the early-stage telencephalon. This previous study also showed that Ezh2 deletion increased the expression of Wif1 and Dkk2, both of which encode inhibitors of the Wnt signaling pathway, and that H3K27me3 was deposited at these gene loci in WT embryos, suggesting that PcG proteins contribute to Wnt activation by repressing these Wnt inhibitor genes in the dorsal midbrain. Mechanisms by which recruitment of PcG proteins is regulated in a tissue-, cell type–, or stage-specific manner warrant clarification in future studies.
BMP signaling has been shown to be important for establishment of the dorsal midline and its activity to be confined to this region (Hebert et al., 2002; Panchision et al., 2001; Roy et al., 2014). The relevance of the absence of BMP signaling outside of the dorsal midline has not been known, however. Our results now suggest that suppression of BMP signaling outside of the dorsal midline is required for the expression of Shh at the ventral midline and that Ring1 mediates this suppression and thereby sets up a permissive state for Shh expression. Given that the suppression of BMP signaling is necessary for neural induction of ectoderm (Wilson and Hemmati-Brivanlou, 1995), its onset may occur before formation of the prospective forebrain. BMP signaling–related targets of PcG proteins identified in embryonic stem cells may be involved in this early process (Shan et al., 2017). The observed increase in Id1 expression in the telencephalon (outside of the dorsal midline) in response to Ring1 deletion from before E10 suggests that BMP signaling remains repressed in this region but becomes derepressed at the dorsal midline, although the mechanisms underlying this difference remain unknown.
We found that the targets of PcG proteins in the early-stage telencephalon include the genes for BMP and Wnt ligands. Of interest, we detected Ring1B binding to BMP and Wnt ligand gene loci in the telencephalon at E9, but the extent of this binding appeared less than that evident at Hox gene loci. In contrast, the levels of H3K27me3 deposition were similar for these two sets of loci. This difference may be due to the operation of different modes of PcG-mediated repression (Tsuboi et al., 2018). Future studies are required to reveal which PcG complexes are responsible for the repression of BMP and Wnt ligand genes.
The robust maintenance of the A-P axis through suppression of Hox gene expression by PcG proteins has been well established from flies to mammals (Montavon and Soshnikova, 2014). We now propose that PcG proteins also play an essential role in formation of the D-V axis in the early stage of mouse telencephalic development. Our study thus sheds light on the role of chromatin-level regulation in regionalization of the brain that is dependent on developmental genes that are not necessarily clustered like Hox genes.
Author Contributions
H.E., Y.K., and Y.G. designed the study and wrote the manuscript. H.E. and Y.K. performed the experiments and analyzed the data. H.K. generated Ring1a−/−;Ring1bflox/flox mice. Y.K. and Y.G. supervised the study.
Declaration of Interests
The authors declare no competing interests.
Materials and Methods
Animals
Ring1bflox/flox or Ring1a−/−;Ring1bflox/flox mice (Calés et al., 2008; Endoh et al., 2008) were crossed with Sox1-Cre transgenic mice (Takashima et al., 2007). Jcl:ICR (CLEA Japan) or Slc:ICR (SLC Japan) mice were studied as WT animals. All mice were maintained in a temperature- and relative humidity–controlled (23° ± 3°C and 50 ± 15%, respectively) environment with a normal 12-h-light, 12-h-dark cycle. They were housed two to six per sterile cage (Innocage, Innovive; or Micro BARRIER Systems, Edstrom Japan) with chips (PALSOFT, Oriental Yeast; or PaperClean, SLC Japan), irradiated food (CE-2, CLEA Japan), and filtered water available ad libitum. Mouse embryos were isolated at various ages, with E0.5 being considered the time of vaginal plug appearance. All animals were maintained and studied according to protocols approved by the Animal Care and Use Committee of The University of Tokyo.
Plasmid constructs
A pBluescript SK(-) vector encoding mouse Shh was kindly provided by D. Kawaguchi (The University of Tokyo). A portion of the Axin2 cDNA was cloned by PCR from cDNA derived from the mouse telencephalon and was subcloned into pBluescript SK(-). Amplified sequences are presented in Supplementary table 2.
Antibodies
Antibodies for immunofluorescence and ChIP analyses included mouse antibodies to Ascl1 (Mash1, BD Pharmingen, 556604, 1:500) and H3K27me3 (MBL, MABI0323, 2 μg/sample for ChIP), goat antibodies to Neurog1 (Santa Cruz, sc-19231, 1:200) and rabbit antibodies to H2AK119ub (Cell Signaling Technology, 8240S, 1:1000), Ring1B (Cell Signaling Technology, 5694S, 1:200, 3 μg/sample for ChIP), Cleaved caspase-3 (Cell signaling Technology, 9664S, 1:1000), Nkx2.1 (TTF1, Abcam, ab76013, 1:1000), Gsx2 (Gsh2, Millipore, ABN162, 1:200), Pax6 (Millipore, AB2237, 1:500) and Id1 (Biocheck, BCH-1/37-2, 1:200). Alexa-labeled secondary antibodies and Hoechst 33342 (for nuclear staining) were obtained from Molecular Probes.
Immunohistofluorescence analysis
Immunohistofluorescence staining was performed as previously described (Morimoto-Suzki et al., 2014), with minor modifications. In brief, embryos were fixed for 3 h with 4% paraformaldehyde in phosphate-buffered saline (PBS), incubated overnight at 4°C with 30% sucrose in PBS, embedded in OCT compound (Sakura Finetek), and sectioned with a cryostat at a thickness of 10 µm. The sections were exposed to 0.1% Triton X-100 and 3% bovine serum albumin in Tris-buffered saline (blocking solution) for 1 h at room temperature before incubation first overnight at 4°C with primary antibodies diluted in blocking solution and then for 1 h at room temperature with fluorophore-labeled secondary antibodies also diluted in blocking solution. They were finally mounted in Mowiol (Calbiochem) for imaging with a laser-scanning confocal microscope (TSC-SP5, Leica) and ImageJ software (NIH).
Isolation of ventral NPCs by FACS
The ventral telencephalon was dissected and subjected to enzymatic digestion with a papain-based solution (Sumitomo Bakelite). Cell suspensions were stained with allophycocyanin-conjugated antibodies to CD133 (141210, BioLegend) at a dilution of 1:400 and were then subjected to fluorescence-activated cell sorting (FACS) with a FACSAria instrument (Becton Dickinson). CD133+ NPCs were isolated as the top 50% of allophycocyanin-positive cells.
Quartz-seq analysis
Both cDNA synthesis and amplification were performed with total RNA from 2000 cells as described previously (Sasagawa et al., 2013). In brief, total RNA was purified from cells with the use of Ampure XP RNA (Beckman) and subjected to RT with Super Script III (Thermo Scientific), and the resulting cDNA was purified with Ampure XP (Beckman) and treated with ExoI (Takara) for primer digestion. After addition of a poly(A) tail with terminal deoxynucleotidyl transferase (Roche), the cDNA was subjected to second-strand synthesis and the resulting double-stranded cDNA was amplified with the use of MightyAmp DNA polymerase (Takara). The amplified cDNA was prepared for sequencing with the use of a Nextera XT DNA Sample Prep Kit (Illumina) and subjected to deep sequencing analysis on the Illumina HiSeq2500 platform to yield 36-base single-end reads. Approximately 20 million sequences were obtained from each sample. Sequences were mapped to the reference mouse genome (mm9) with ELAND v2 (Illumina). Only uniquely mapped tags with no base mismatches were used for the analysis. Gene expression was quantitated as reads per kilobase of mRNA model per million total reads (RPKM) on the basis of RefSeq gene models (mm9).
RT-qPCR analysis
Total RNA was isolated from cells with the use of RNAiso plus (Takara), and up to 0.5 µg of the RNA was subjected to RT with the use of ReverTra Ace qPCR RT Master Mix with gDNA remover (Toyobo). The resulting cDNA was subjected to real-time PCR analysis in a LightCycler 480 instrument (Roche) with either KAPA SYBR FAST for LightCycler 480 (Kapa Biosystems) or Thunderbird SYBR qPCR mix (Toyobo). The amount of each target mRNA was normalized by that of β-actin mRNA. Primer sequences are presented in Supplementary table 2.
In situ hybridization analysis
For preparation of digoxigenin-labeled riboprobes, linearized plasmids containing probe sequences were incubated for 3 h at 37°C with DIG RNA Labeling Mix, Transcription Buffer, and RNA polymerase (Roche) as well as RNase inhibitor (Toyobo). The plasmids were then digested with DNaseI (Takara) for 30 min at 37°C, after which the DNase reaction was stopped by the addition of Stop Solution (Promega). Synthesized riboprobes were purified with the use of a ProbeQuant G-50 column (GE Healthcare) and diluted with hybridization buffer (5× Denhardt’s solution, 5× standard saline citrate, 50% formamide, tRNA at 250 µg/ml, salmon testis DNA at 200 µg/ml, heparin at 100 µg/ml, and 0.1% Tween 20). The riboprobes (0.5 µg/ml) were denatured at 85°C for 5 min, placed on ice for 2 min, and then maintained at 65°C before in situ hybridization. Embryos were fixed for 3 h (Shh) or overnight (Axin2) with 4% paraformaldehyde in PBS and then incubated with 30% sucrose in PBS, embedded, and sectioned as described for immunohistofluorescence analysis. Sections were fixed for 10 min with 4% paraformaldehyde in PBS, washed with 0.1% Tween 20 in PBS, and incubated at room temperature first with 0.1 M triethanolamine for 3 min and then with the same solution containing 0.1% acetic anhydride for 10 min. They were washed again with 0.1% Tween 20 in PBS before incubation at 65°C first for 1 h with hybridization buffer and then overnight with denatured RNA probes within a humidified box with 50% formamide. The sections were washed twice for 30 min at 65°C with 2× standard saline citrate, twice for 30 min at 65°C with the same solution containing 50% formamide, and three times for 5 min at room temperature with 0.1% Tween 20 in MAB buffer (MABT). After exposure for 1 h at room temperature to 10% fetal bovine serum in MABT, the sections were incubated overnight at 4°C with alkaline phosphatase–conjugated antibodies to digoxigenin (Roche) at a dilution of 1:2000 in the same solution, washed twice for 10 min at room temperature with MABT and twice for 10 min at room temperature with a solution containing 100 mM Tris-HCl (pH 9.5), 100 mM NaCl, 50 mM MgCl2, and 0.02% Tween 20, and then incubated at room temperature in the same solution containing NBT-BCIP (nitrotetrazolium blue chloride at 350 µg/ml and 5-bromo-4-chloro-3-indolyl phosphate p-toluidine salt at 175 µg/ml) (Roche) until the color appeared. They were finally washed with 0.1% Tween 20 in PBS and mounted in Mowiol (Calbiochem). Images were acquired with an Axiovert 200M microscope fitted with an Axiocam or Axiocam 305 camera (Carl Zeiss) and were processed with ImageJ (NIH).
ChIP-qPCR analysis
ChIP for Ring1B and H3K27me3 was performed as previously described (Tsuboi et al., 2018), with minor modifications. Cells were fixed with 1% formaldehyde and then suspended in radioimmunoprecipitation (RIPA) buffer for sonication (10 mM Tris-HCl at pH 8.0, 1 mM EDTA, 140 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.1% sodium deoxycholate). They were subjected to ultrasonic treatment to shear genomic chromatin into DNA fragments, and the cell lysates were then diluted with RIPA buffer for immunoprecipitation (50 mM Tris-HCl at pH 8.0, 150 mM NaCl, 2 mM EDTA, 1% Nonidet P-40, 0.1% SDS, 0.5% sodium deoxycholate) and incubated for 1 h at 4°C with ProteinA/G Magnetic Beads (Pierce) to clear nonspecific reactivity. They were then incubated overnight at 4°C with ProteinA/G Magnetic Beads that had previously been incubated overnight at 4°C with antibodies to Ring1B or to H3K27me3. The beads were then isolated and washed three times with wash buffer (2 mM EDTA, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, and 20 mM Tris-HCl at pH 8.0) and then once with wash buffer containing 500 mM NaCl. Immune complexes were eluted from the beads with a solution containing 10 mM Tris-HCl (pH 8.0), 5 mM EDTA, 300 mM NaCl, and 0.5% SDS at 65°C for 15 min, and they were then subjected to digestion with proteinase K (Nakarai) at 37°C for more than 6 h, removal of cross links by incubation at 65°C for more than 6 h, and extraction of the remaining DNA with phenol–chloroform–isoamyl alcohol and ethanol. The DNA was washed with 70% ethanol, suspended in water, and subjected to real-time PCR analysis in a LightCycler 480 instrument (Roche) with Thunderbird SYBR qPCR Mix (Toyobo). Primer sequences are presented in Supplementary table 2.
Primary culture of the telencephalon and treatment with pharmacological agents
For monolayer culture, primary NPCs were isolated from the indicated regions of the telencephalon of ICR or of Ring1B KO or control mouse embryos. The dissected tissue was thus subjected to digestion with a papain-based solution (Sumitomo Bakelite), and the dissociated cells were cultured in dishes coated with poly-D-lysine (Sigma) and containing Dulbecco’s modified Eagle’s medium (DMEM)–F12 (1:1, v/v) supplemented with B27 (Invitrogen) and recombinant human FGF2 (Invitrogen) at 20 ng/ml. For explant culture, the dissected telencephalon of ICR mouse embryos was cultured in DMEM-F12 supplemented with B27 and recombinant human FGF2. After culture of cells or explant tissue for 6 h, half of the medium was removed and replaced with medium supplemented with B27, human FGF2, and either Smoothened agonist (SIGMA-ALDRICH), recombinant human BMP4 (R&D Systems) or CHIR-99201 (Wako). Smoothened agonist was dissolved in dimethyl sulfoxide at a concentration of 5 mM and was added to culture medium at a final concentration of 2 μM. The recombinant BMP4 protein was dissolved at a concentration of 100 µg/ml in sterile 4 mM HCl containing 0.1% bovine serum albumin and was added to the culture medium at a final concentration of 50 ng/ml. CHIR-99201 was dissolved in dimethyl sulfoxide at a concentration of 1 mM and was added to culture medium at a final concentration of 5 µM. The cells or tissue were cultured for an additional 24 h and then frozen in liquid nitrogen before analysis.
Statistical analysis
Data are presented as means ± s.e.m. and were compared with two-tailed Student’s paired or unpaired t test or by analysis of variance (ANOVA) followed by Bonferroni’s multiple-comparison test.
Acknowledgments
We thank S. Nishikawa (RIKEN) for providing Sox1-Cre mice; I. Nikaido (RIKEN) for providing tips for the Quartz method; K. Imamura, T. Horiuchi, S. Sugano and Y. Suzuki (The University of Tokyo) for performing high-throughput-sequencing analysis; M. Endoh (National University of Singapore) for providing tips for the Ring1B ChIP assay; Y. Koseki (RIKEN) for providing Ring1 mutant mice; Y. Maeda, R. Nagayoshi, and Y. Kakeya for technical assistance; and members of the Gotoh laboratory for discussion. This research was supported by AMED-CREST of the Japan Science and Technology Agency (grant JP18gm0610013) and by the Ministry of Education, Culture, Sports, Science, and Technology of Japan (JSPS KAKENHI grants JP15H05773, JP16H06481, JP16H06479, and JP16H06279 to Y.G., and JP18K14622, JP19H05253, and JP16H06279 to Y.K.).