LSD1 co-repressor Rcor2 orchestrates neurogenesis in the developing mouse brain

Abstract

Epigenetic regulatory complexes play key roles in the modulation of transcriptional regulation underlying neural stem cell (NSC) proliferation and progeny specification. How specific cofactors guide histone demethylase LSD1/KDM1A complex to regulate distinct NSC-related gene activation and repression in cortical neurogenesis remains unclear. Here we demonstrate that Rcor2, a co-repressor of LSD1, is mainly expressed in the central nervous system (CNS) and plays a key role in epigenetic regulation of cortical development. Depletion of Rcor2 results in reduced NPC proliferation, neuron population, neocortex thickness and brain size. We find that Rcor2 directly targets Dlx2 and Shh, and represses their expressions in developing neocortex. In addition, inhibition of Shh signals rescues the neurogenesis defects caused by Rcor2 depletion both in vivo and in vitro. Hence, our findings suggest that co-repressor Rcor2 is critical for cortical development by repressing Shh signalling pathway in dorsal telencephalon.

Figure 1. Rcor2 expresses in the CNS and regulates cortical development.

(a) X-gal staining to detect Rcor2 expression patterns. Whole-embryo staining at E11.5 stage shows Rcor2 is mainly expressed in the CNS. Scale bar, 1 mm. (b) Western blot analysis of Rcor2 expression levels during brain development. The decreased expression of Rcor2 with embryonic development is noteworthy. β-Actin is used as an endogenous control. (c) In situ hybridization to detect endogenous Rcor2 mRNA expression patterns in cortical development at E11.5, E13.5, E15.5 and E17.5. Insets show high-magnification image of Rcor2 expression in the neocortex at E13.5. VZ, ventricular zone; SVZ, subventricular zone; CP, cortical plate. Scale bar, 100 μm. (d) Confocal images of immunofluorescence to detect cellular localization of Rcor2 in the neocortex at E13.5. Rcor2 localized mainly in the nucleus at interphase and metaphase, and localized between separated chromosomes in anaphase of dividing cells in VZ. Dotted lines circle the shape of nuclei. Scale bar, 5 μm. (e) Western blot analysis of Rcor2 expression level in Rcor2^fl/fl and Rcor2^cko brains at E13.5 and E15.5, respectively. Rcor2 expression was depleted in Rcor2^cko brains. β-Actin is used as an endogenous control. (f) Representative images of Rcor2^fl/fl and Rcor2^cko brain size at different stages of development. Rcor2^cko mice show severe brain growth retardation at E13.5 and E15.5. Scale bar, 1 mm. (g) Representative images of Rcor2^fl/fl and Rcor2^cko cortex at E15.5 by Nissl staining. Structural abnormalities of lamination with reduced cortical thickness are observed in Rcor2^cko cortex. Scale bar, 200 μm. (h) RT–qPCR analysis of knockdown efficiencies of the two shRNAs targeting Rcor2. Transcripts were normalized to the control group. Data are shown as mean±s.e.m., t-test, ****P<0.0001, n=3. (i) Confocal images of E16.5 cortical sections electroporated with shControl (red), shRcor2-a (red) and shRcor2-b (red) plasmids at E13.5. Knockdown of Rcor2 results in impaired cortical development. IZ, intermediate zone. Scale bar, 20 μm. (j) Quantification of the percentage of RFP⁺ cells in different regions of the developing neocortex after electroporation as shown in i. proportion of RFP⁺ cells in different zones (y axis). Data are shown as mean±s.e.m., t-test, *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001, n=3 individual experiments.

Figure 2. Rcor2 regulates cortical NSC/NPC population and proliferation during development.

(a) Immunostaining images of Sox2 at E13.5. Sox2 is dramatically reduced on Rcor2 depletion. VZ, ventricular zone; SVZ, subventricular zone. Scale bar, 20 μm. (b) Quantification of Sox2⁺ cell ratios in VZ/SVZ regions shown in a. Data are shown as mean±s.e.m., t-test, ** P<0.01, n=3 individual experiments. (c) Confocal images of Tbr2 expression at E13.5. Tbr2 is dramatically reduced on Rcor2 depletion. Scale bar, 20 μm. (d) Quantification of Tbr2⁺ cell ratios in VZ/SVZ regions shown in c, respectively. Data are shown as mean±s.e.m., t-test, *P<0.05, n=3 individual experiments. (e) Immunostaining images of Nestin, Sox2 and Tbr2 in cultured Rcor2^fl/fl and Rcor2^cko NPCs, all of which exhibit significantly reduced expression in the Rcor2^cko NPCs. Scale bar, 20 μm. (f) Confocal images of immunofluorescence for Ki67 and PHH3 in Rcor2^fl/fl and Rcor2^cko cortex at E13.5 and E15.5. Ki67 signals (red), but not PHH3 signals (green), are dramatically reduced in Rcor2^cko developing brains. Scale bar, 20 μm. (g) Quantification of Ki67⁺ cells in the VZ/SVZ regions of the developing neocortex as shown in f. Data are shown as mean±s.e.m., t-test, **P<0.01, n=3 separate stainings from three independent brains. (h) Confocal images of BrdU (green) and Ki67 (red) staining in Rcor2^fl/fl and Rcor2^cko cortex 24 and 48 h after BrdU incorporation. Scale bar, 100 μm. (i) Quantification of the cell cycle exit by percentage of BrdU⁺ and Ki67⁺ NPCs divided by BrdU⁺ cells shown in h. Data are shown as mean±s.e.m., t-test, ***P<0.001 and ****P<0.0001, n=3 individual experiments. (j,k) Representative images (j) and quantification of (k) of Rcor2^fl/fl and Rcor2^cko neurosphere sizes. The neurospheres'radius of Rcor2^cko are much smaller than those of Rcor2^fl/fl, t-test, ****P<0.0001, n=12. Scale bar, 50 μm. (l) Representative time-lapse imaging of the RGC dividing process in the sections of the cerebral cortex electroporated with RFP-shControl (upper panels) and RFP-shRcor2 (lower panels). The radial glial dividing pattern is abnormal on Rcor2 knockdown, resulting in cell death. Arrows, mother RGCs. Arrowheads, two daughter cells. Scale bar, 50 μm. (m) Representative time-lapse images of Rcor2^fl/fl cortex sections electroporated with EGFP-Control (upper panels) and EGFP-Cre (lower panels). Loss of cells is observed with Rcor2 knockout by Cre recombinase electroporation. Scale bar, 50 μm.

Figure 3. Rcor2 regulates cortical neurogenesis during development.

(a) Confocal images of Satb2 and Tbr1 expressions in Rcor2^fl/fl and Rcor2^cko cortex at E15.5, which exhibit significant reduction on Rcor2 knockout. Scale bar, 50 μm. (b) Quantification of Satb2⁺ and Tbr1⁺ cells in Rcor2^fl/fl and Rcor2^cko cortex at E15.5 in a indicates Satb2 and Tbr1 expressions are decreased on Rcor2 depletion during development. Data are shown as mean±s.e.m., t-test, **P<0.01, n=3 individual experiments. (c) Western blot to analyse Dcx, Satb2 and Tbr1 expressions in Rcor2^fl/fl and Rcor2^cko cortex at E15.5. β-Actin is used as an endogenous control. (d) Representative images of Map2 and Tuj1 immunostaining in cultured neurons directly dissociated from Rcor2^fl/fl and Rcor2^cko brain cortex at E15.5. Decreased expression of both markers and reduced neurofilaments can be observed in Rcor2^cko cultured neurons. Scale bar, 20 μm. (e) Confocal images of in-vitro cultured Rcor2^fl/fl and Rcor2^cko NPCs 5 days post spontaneous differentiation using neuronal marker Map2 and Tuj1 antibodies, both of which are significantly reduced in the differentiated Rcor2^cko NPCs. Scale bar, 20 μm. (f,g) RT–qPCR analysis of neuronal markers expression in both Rcor2^fl/fl and Rcor2^cko neocortex at E15.5 stage (f) and in-vitro-cultured Rcor2^fl/fl and Rcor2^cko NPCs 5 days post spontaneous differentiation (g). Transcripts were normalized to Rcor2^fl/fl group. Data are shown as mean±s.d., t-test, *P<0.05, ** P<0.01 and ***P<0.001, n=3.

Figure 4. ChIP-seq analysis of Rcor2 enrichments in genome-wide scale reveals direct binding of Rcor2 at regulatory regions of genes related to Shh signalling pathway.

(a) Schematic overview of strategy to generate an Rcor2^Flag knock-in allele by CRISPR/Cas9. The sgRNA sequence site is shown as a green arrowhead. The start codon of Rcor2 is indicated and capitalized. The oligo donor contained 50 bp homologies on both sides flanking the DSB, in which 3 × Flag sequences are labelled as a red box. (b) Western blot analysis to validate FLAG, RCOR2 and LSD1 expressions in Rcor2^Flag knock-in neocortex using Flag-M2 antibody. β-Actin was used as an endogenous control. (c) Pie chart depicts distribution of Rcor2 occupancies in genome-wide scale in FLAG ChIP-seq results using Rcor2^Flag knock-in neocortex at E13.5. (d) WebLogos of consensus binding motifs of Rcor2 generated by Multiple EM for Motif Elicitation (MEME) motif analysis tool. (e) GO analysis for Rcor2-binding regions in genome-wide scale revealed by Flag ChIP-seq results using Rcor2^Flag brain. (f) GO analysis for LSD1 occupancy in genome-wide scale revealed by LSD1 ChIP-seq results using Rcor2^Flag brain. (g) Density plots analysis of H3K4me1 signal change in promoter regions (−2- to ∼0.5 kb from TSS) on Rcor2 depletion. Compared with all genes, the change of H3K4me1 signal is significantly (P<0.0005, Kolmogorov–Smirnov test) enhanced in the promoter regions of Shh pathway-related genes on Rcor2 depletion. H3K4me1 signal change on Rcor2 depletion (x axis); H3K4me1 signal density (y axis). (h) Gene tracks of Rcor2, LSD1 and H3K4me1 enrichments by ChIP-seq analysis at core promoter regions of Dlx2 and upstream regulatory regions of Shh, which are closely related to Shh signalling. (i) ChIP–qPCR analysis of Rcor2^Flag and Rcor2^cko cortex at E13.5 using specific FLAG-M2 antibody. Significant enrichments of the Rcor2 at the regulatory regions of Dlx2 and Shh gene locus detected in g in the Rcor2^flag samples are worth noting. Fold enrichments of Rcor2 occupancy compared with input (y axis). Data are shown as mean±s.d., t-test, ***P<0.001, n=3.

Figure 5. Genome-wide expression changes on Rcor2 depletion in mouse developing cortex by RNA-seq analysis.

(a,b) Scatter plot analysis of genome-wide expression profiles of Rcor2^cko versus Rcor2^fl/fl samples at E13.5 (a) and E15.5 (b). Dots above or below the dash line indicate upregulated or downregulated genes on Rcor2 depletion, respectively. Red dots or green dots highlight the significantly differentially expressed genes on Rcor2 depletion. Raw counts (x axis); gene expression fold changes on Rcor2 depletion (y axis). (c) Venn diagrams of upregulated genes (left) and downregulated genes (right) in Rcor2^cko samples compared with Rcor2^fl/fl samples. (d) The profiles of Rcor2 and H3k4me1 enrichments analysed in ChIP-seq results shown in Fig. 3 in regulatory regions of genome-wide scale (red) and of the upregulated genes (purple) according to RNA-seq results. (e) Correlation network of overlapped upregulated genes in both E13.5 and E15.5 samples. Lines indicate the correlations between two connected genes with R>0.55. Genes were analysed by GO analysis and divided into different categories.

Figure 6. Rcor2 regulates Shh signalling pathway during cortical development.

(a) qPCR analysis of the expression of genes related to the Shh signalling pathway in the cortex of Rcor2^fl/fl and Rcor2^cko brains during development. Significant upregulation of these genes on Rcor2 depletion is noteworthy. Transcripts were normalized to Rcor2^fl/fl group. Data are shown as mean±s.d., t-test, *P<0.05, **P<0.01 and ***P<0.001, n=3. (b) Confocal images of Shh and Ptch1 expressions in Rcor2^fl/fl and Rcor2^cko cortex. Enhanced Shh and Ptch1 signals are observed in Rcor2^cko compared with Rcor2^fl/fl neocortex at E13.5 and E15.5. Insets show high-magnification images of the outlined regions. Scale bars, 50 μm. (c) Dlx2 expression in Rcor2^fl/fl and Rcor2^cko cortex detected by immunofluorescence analysis at E15.5. Dlx2⁺ cells were observed in the neocortex on Rcor2 depletion. Scale bars, 50 μm. (d) Confocal images of Shh, Ptch1 and Dlx2 expressions in in-vitro-cultured Rcor2^cko NPCs. Scale bar, 20 μm. (e) Western blot analysis of expression levels of Dlx2, Shh and Ptch1 in Rcor2^fl/fl and Rcor2^cko cortex at E15.5. β-Actin is used as an endogenous control.

Figure 7. Rcor2 regulates cortical development by inhibition of Shh.

(a) Knockdown of Rcor2 impairs cortical neurogenesis, which can be partially rescued by knockdown of Shh during cortical development. In-utero electroporation with RFP-shControl (red)/GFP-shControl (green), RFP-shRcor2 (red)/GFP-shControl (green), RFP-shControl (red)/GFP-shShh (green) or RFP-shRcor2 (red)/GFP-shShh (green) plasmids was performed at E13.5. Cerebral sections were fixed and imaged at E16.5. VZ, ventricular zone; SVZ, subventricular zone; CP, cortical plate. Scale bar, 50 μm. (b) Quantification of the percentage of RFP⁺/GFP⁺ cells in different regions of the developing cortex after electroporation shown in a. Data are shown as mean±s.e.m., t-test, **P<0.01 and ***P<0.001, n=3 individual experiments. (c) Inhibition of Shh by Cyclopamine can partially rescue neurogenesis defects caused by Rcor2 downregulation during cortical development. Rcor2 was knocked down at the lateral ventricle in the brain by in-utero electroporation with RFP-shRcor2 plasmids at E13.5. Cerebral sections were collected at E14.5 and then treated with cyclopamine to inhibit Shh activity for 48 h. Scale bar, 50 μm. (d) Quantification of the percentage of RFP⁺ cells in different regions of the developing neocortex after knockdown of Rcor2 or inhibition of Shh shown in c. Data are shown as mean±s.e.m., t-test, *P<0.05, n=3 individual experiments. (e) Representative images depicting neurosphere size is partially rescued in the in-vitro-cultured Rcor2^cko NPCs after treatment with Cyclopamine. Scale bar, 50 μm. (f) Histogram depicting cell numbers of in-vitro-cultured Rcor2^fl/fl and Rcor2^cko NPCs with or without Cyclopamine treatment for 48 h. Cells (5 × 10⁵) are seeded initially. Data are shown as mean±s.e.m., t-test, **P<0.01, n=3. (g) Confocal images of Tuj1 expression in the differentiated cells from in-vitro-cultured Rcor2^fl/fl and Rcor2^cko NPCs with or without Cyclopamine treatment. Tuj1 expressions are partially restored in Cyclopamine-treated Rcor2^cko cells. Scale bar, 20 μm. (h) Model of Rcor2 function in neurogenesis in the developing neocortex. Rcor2 safeguards cortical neurogenesis by recruiting LSD1 complex to the regulatory regions of Dlx2 and Shh genes, to inhibit the Shh pathway activation during development. The absence of Rcor2 leads to inhibition release of these genes and thus ectopic activation of Shh signalling in the developing neocortex, resulting in cortical neurogenesis defects.