REST corepressors RCOR1 and RCOR2 and the repressor INSM1 regulate the proliferation-differentiation balance in the developing brain

Abstract

The transcriptional events that lead to the cessation of neural proliferation, and therefore enable the production of proper numbers of differentiated neurons and glia, are still largely uncharacterized. Here, we report that the transcription factor Insulinoma-associated 1 (INSM1) forms complexes with RE1 Silencing Transcription factor (REST) corepressors RCOR1 and RCOR2 in progenitors in embryonic mouse brain. Mice lacking both RCOR1 and RCOR2 in developing brain die perinatally and generate an abnormally high number of neural progenitors at the expense of differentiated neurons and oligodendrocyte precursor cells. In addition, Rcor1/2 deletion detrimentally affects complex morphological processes such as closure of the interganglionic sulcus. We find that INSM1, a transcription factor that induces cell-cycle arrest, is coexpressed with RCOR1/2 in a subset of neural progenitors and forms complexes with RCOR1/2 in embryonic brain. Further, the Insm1^-/- mouse phenocopies predominant brain phenotypes of the Rcor1/2 knockout. A large number of genes are concordantly misregulated in both knockout genotypes, and a majority of the down-regulated genes are targets of REST. Rest transcripts are up-regulated in both knockouts, and reducing transcripts to control levels in the Rcor1/2 knockout partially rescues the defect in interganglionic sulcus closure. Our findings indicate that an INSM1/RCOR1/2 complex controls the balance of proliferation and differentiation during brain development.

Keywords: INSM1; RCOR1; RCOR2; REST; neurogenesis.

Fig. S1.

Schematic diagram of Rcor2 gene targeting strategy. Mice with conditional Rcor2 alleles (floxed) were generated by inserting LoxP sites flanking exons 5–9. Exons are depicted as black rectangles and LoxP sites as white diamonds. HSVTK, herpes simplex virus thymidine kinase; PGKNEO, neomycin resistance gene driven by the phosphoglycerine kinase promoter.

Fig. S2.

Brain phenotypes of Rcor1 and Rcor2 knockouts at E18.5. Shown are H&E-stained coronal hemisections of telencephalon. Nes-Cre and Rcor2 KO hemisections are representative. Three of six Rcor1 KOs, including the one depicted here, had a cleft (arrowhead) at the surface of the ventral VZ/SVZ of the lateral subpallium. In other respects, single Rcor KOs were indistinguishable from Nes-Cre controls. Boxes in Top indicate Insets shown below at higher magnification. [Scale bar, 500 μm (Top) and 50 μm (Bottom).]

Fig. 1.

Analysis of RCOR1 and RCOR2 expression at E13.5 in the Rcor1/2 fl and Rcor1/2 knockout mice. (A) Western blot analysis. RCOR1 and RCOR2 levels were normalized to histone H3 levels. The means and SDs are indicated. Statistical significance was assessed by t test (n = 4 mice). ***P < 0.001. (B) Representative immunohistochemistry on coronal hemisections of control (Rcor1/2 fl) and Rcor1/2 KO telencephalon labeled with the indicated antibodies and DAPI to mark nuclei. Boxes indicate regions shown at higher magnification to the right. (Scale bars, 200 μm and 20 μm.) LGE, lateral ganglionic eminence; MGE, medial ganglionic eminence.

Fig. S3.

Western blot from which the bands from “Rcor1/2 fl #2” and “Rcor1/2 KO #3” were shown in Fig. 1.

Fig. S4.

Representative immunohistochemistry on coronal hemisections of E18.5 control (Rcor1/2 fl) and Rcor1/2 knockout brains labeled with the indicated antibodies and DAPI to mark nuclei. (Scale bar, 500 μm.)

Fig. 2.

Brain phenotypes of Rcor1/2 knockouts at E18.5. (A) Brain weights of control (Rcor1/2 fl and Nes-Cre) and Rcor1/2 KO mice. The results represent, for each genotype, a minimum of 11 brains from five litters. Statistical significance was assessed using ANOVA followed by Tukey’s multiple comparison test. ***P < 0.001; ns, not significant. (B) Representative H&E-stained coronal hemisections of Rcor1/2 fl and Rcor1/2 KO telencephalon. Note deep IGS (arrowhead), enlarged VZ/SVZs (*), and diminished corpus callosum (arrow) and axonal fasciculation in the striatum (str) in the KO compared with the control.

Fig. S5.

Representative H&E-stained coronal sections from the E18.5 Rcor1/2 fl and Rcor1/2 knockout. (Top) The anterior commissure (arrowhead) is abnormally thin in the Rcor1/2 KO. (Bottom) In the Rcor1/2 KO, CA3 (arrow) and the dentate gyrus (asterisk) of the hippocampus fail to resolve into distinct structures. The thalamus (th) and hypothalamus (hyp) of the Rcor1/2 KO are also hypoplastic.

Fig. 3.

Increased numbers of proliferating progenitors in E18.5 brains of Rcor1/2 knockouts compared with controls. Sections labeled with antibodies were counterstained with DAPI or DRAQ5 to mark nuclei. (A, Left) Representative images of MKI67 immunohistochemistry. Boxes indicate Insets shown below at higher magnification. [Scale bars, 500 μm (Top) and 100 μm (Bottom).] (Right) Quantification of the width of the MKI67+ region in indicated controls and the Rcor1/2 KO. Measurements were taken from images comparable to Insets. The means and SDs are indicated. Statistical significance was assessed by t test (n = 5 mice). ***P < 0.001. (B) Representative in situ hybridization histochemistry for Dlx2 and immunohistochemistry for OLIG2, NKX2-1, and ASCL1. (Scale bar, 200 μm.)

Fig. 4.

Rcor1/2 knockouts have fewer neurons and OPCs than controls. All analyses were performed on DAPI-labeled coronal sections from E18.5 brain. (A) Representative images of MAP2 immunohistochemistry. (Scale bar, 500 μm.) (B) Quantification of MAP2 immunolabeling, showing the percentage of each coronal hemisection occupied by the MAP2+ domain. The total area of the hemisection was determined on the basis of DAPI labeling. The means and SDs are indicated. Statistical significance was assessed with a t test (n = 5 mice). (C) Representative image of OLIG2 immunohistochemistry. Boxes in Top indicate regions shown below at higher magnification. Boxes in Bottom show the IZ of the neocortex. [Scale bar, 500 μm (Top) and 100 μm (Bottom).] (D) Quantification of OLIG2+ immunolabeling in the IZ of the neocortex. For each mouse, one region of 3 × 10⁴ μm² was selected from each of six hemisections, and the numbers of OLIG2+ cells from all regions were added. Each mouse is represented by one dot. The means and SDs are indicated. Statistical significance was assessed by t test (n = 5 mice). ****P < 0.0001.

Fig. 5.

RCOR1 and RCOR2 interact with INSM1 in subpallial neural progenitors. (A) Representative immunohistochemistry on a coronal hemisection of E13.5 telencephalon shows a subset of cells expressing both RCOR2 and INSM1. Dashed and dotted lines indicate the boundaries of the subpallial progenitor zone, determined by alignment of the section to an adjacent section immunolabeled with the subpallial progenitor marker ASCL1. Dashed lines indicate the borders between the subpallial progenitor zone and the lateral ventricle. Dotted lines indicate the boundaries between the subpallial progenitor zone and other brain regions. The box indicates the region shown to the right at higher magnification. [Scale bars, 200 μm (Left) and 20 μm (Right).] (B–D) Co-IP analysis for INSM1–RCOR1/2 complexes. IPs were performed on nuclear extracts prepared from E13.5 brain. Labels to the left of each blot indicate the antibodies used for Western blotting. IgG is rabbit IgG. Normal serum is rabbit serum in C and guinea pig serum in D. ^+/+, Insm1^+/+ nuclear extracts; ^−/−, Insm1^−/− nuclear extracts.

Fig. S6.

Western blots corresponding to coimmunoprecipitation analysis in Fig. 5 B–D. Arrowheads indicate Western blot proteins of interest. (A) RCOR1 co-IP. (B) RCOR2 co-IP. (C) INSM1 co-IP. Note that the RCOR1 blot of the INSM1 co-IP (C, Middle) shows IP of RCOR1 by the INSM1 antibody in Insm1^−/− extracts, which precludes interpretation of this experiment.

Fig. 6.

Insm1^−/− mice phenocopy several Rcor1/2 knockout phenotypes in E18.5 brain. (A) H&E-stained E18.5 coronal hemisections. Asterisks indicate the VZ/SVZ; arrowheads indicate interganglionic sulci. (B) Representative hemisections immunolabeled for the proliferation marker MKI67 and counterstained with DAPI. Boxes indicate Insets shown below at higher magnification. [Scale bar, 500 μm (Top) and 100 μm (Bottom).] (C–E) Immunohistochemical analysis of coronal hemisections of Insm1^+/+ and Insm1^−/− forebrain. Statistical significance was assessed by t tests. The means and SDs are indicated. (C) Quantification of the width of the MKI67+ zone. Measurements were made from areas comparable to those depicted in the Insets in B. (D) Quantification of MAP2 immunolabeling, showing the percentage of each E18.5 coronal hemisection occupied by the MAP2+ domain. (E) Quantification of the number of OLIG2+ cells in the IZ of the neocortex. For each mouse, one region of 3 × 10⁴ μm² was selected from each of six hemisections, and the numbers of OLIG2+ cells from all regions were added. Each mouse is represented by one dot. Exploratory data analysis identified one wild-type count (107, indicated in red) as an outlier, so it was omitted from the statistical analysis. **P < 0.01; ****P < 0.0001.

Fig. 7.

Differential gene expression in Rcor1/2 knockout and Insm1^−/− mice at E13.5. (A) Venn diagrams comparing the genes misregulated in Rcor1/2 KO and Insm1^−/− mice (relative to Rcor1/2 fl and Insm1^+/+ mice, respectively). (B) Proportions of up- and down-regulated genes that are REST targets. (C, Top) RT-qPCR analysis of cDNA prepared from E13.5 MGE. Each transcript quantity was normalized to the geometric mean of the quantities of four reference genes: Aip, Cxxc1, Rn45s, and Rps20. The means and SDs are indicated (n = 6 mice). (Bottom) Significance was determined by ANOVA with Tukey’s multiple comparison tests on log-transformed data. (D) RT-qPCR analysis as described in C. Statistical significance was assessed using t tests with Welch’s correction on log-transformed data. ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 8.

Reducing Rest transcript levels to control levels in the Rcor1/2 knockout partly rescues the IGS phenotype (arrowhead). Shown are representative H&E-stained coronal hemisections of telencephalon at E18.5.

Fig. S7.

Reducing Rest transcript levels to control levels in the Rcor1/2 KO partly rescues the IGS phenotype (arrowheads in KOs). Shown are H&E-stained coronal sections of E18.5 telencephalon from five mice of each genotype. Images were correctly sorted into groups by morphology by a person blind to the genotypes. Sections shown in Fig. 8 are on the left. (Scale bar, 500 μm.)