As a first step in studying whether Tcf4 plays a role in cortical development, we examined the expression profile of Tcf4 in the developing mouse brain. Using in situ hybridization we found that Tcf4 mRNA was present in the subplate (SP) and proliferative zones of the cortical wall [the ventricular zone (VZ) and the subventricular zone (SVZ)] at E14 and in the cortical plate (CP) of the cerebral cortex and hippocampus at embryonic (E) day 17.5 and postnatal (P) day 0 (Figure 1A and Supplementary Figure S1). Immunofluorescent staining also shows that Tcf4 is highly expressed in the E17.5 mouse cortex (Figure 1B). Furthermore, we measured the abundance of Tcf4 in the developing cerebral cortex and found that high levels of Tcf4 protein were present in the neocortex from E14.5 to P0 (Figures 1C,D).

Because Tcf4 is expressed in the developing cerebral cortex, we wondered whether Tcf4 may regulate neuronal migration. To test this possibility, in utero electroporation was used to cotransfect radial glial progenitors in the cerebral cortex of E14.5 mice with plasmids expressing Tcf4 shRNA and GFP as a marker. The distribution of GFP-positive cells was examined at P0. Approximately 80% of the Tcf4 shRNA-expressing neurons were in the IZ and VZ, whereas the control cortical neurons migrated into CP (Figures 2A,B). The shTcf4 1# specificity was demonstrated by the rescue of neuronal migration following co-electroporation of a plasmid-expressing Tcf4 mutant resistant to the shTcf4 1# sequences (Tcf4^R) (Figures 2A–C). Furthermore, the western blot results showed that shTcf4 1# caused an 80% reduction of Tcf4 expression, whereas 2# caused a 75% reduction (Figures 2C,D). Thus, Tcf4 is essential for neuronal migration.

To explore the underlying mechanism by which Tcf4 controls radial migration, we analyzed the morphology of neurons in the IZ in detail. E14.5 mice were transfected with pSUPER and the short hairpin RNA against Tcf4 (shTcf4 1#) constructs into the developing mouse brain using in utero electroporation. We noticed that Tcf4 knockdown neurons exhibited impairments in leading processes formation at E17.5 with a large portion of cells (50.3%, compared to 12%in the control) lacking a healthy leading process (Figures 3A–C).

We also investigate whether Tcf4 knockdown might arrest neuronal migration by altering radial glia organization, cell division or cell survival. After 3 days of electroporation, we stained slices with Nestin to indicate radial glia organization, Ki67 to show cell division, and cleaved caspas3 to mark cell survival. There were no significant differences between control and Tcf4 knockdown slices (Supplementary Figures S2A–E). Tcf4 knockdown thus does not affect radial glia organization, cell division or cell survival.

Knockdown of human TCF4 is known to affect Bmp7 (Forrest et al., 2013). Here, we found that knockdown of Tcf4 upregulated Bmp7 expression in mouse progenitor cells (Figures 4C,D and Supplementary Figure S3). We also found that the Bmp7 expression profile in the developing mouse brain negatively correlates with Tcf4 (Figures 1B and 4A,B). We next identified a putative regulatory region of 1 kb upstream of the transcriptional start site of Bmp7 which contained six E-box (5′-CANNTG-3′) motifs that are known binding sites of Tcf4. Binding of Tcf4 to these motifs was tested by chromatin immunoprecipitation (ChIP), followed by quantitative real-time PCR using primer pairs (R1–R5) that specifically detected binding to these six motifs (R5 contains two E-box motifs). Figure 4F shows that Tcf4 binds on -201–49 bp of the Bmp7 promoter (R4 and R5). In addition, we found an enrichment of precipitated DNA of more than 5-fold using a Tcf4-specific antibody compared with an immunoglobulin G (IgG) control antibody (Figure 4E). Furthermore, this 1 kb regulatory sequence was tested for transcriptional activity by luciferase assays, yielding a repression of 50.5% (Figure 4G). This indicates that this sequence conveys functional repression through Tcf4 binding. Thus, Tcf4 binds to the Bmp7 promoter and negatively regulates Bmp7.

Electroporation of Bmp7 cDNA at E14.5 resulted in migration defects of cortical neurons similar to those observed in Tcf4-depleted neurons at E17.5 (Figures 5A,B). In addition, the proportion of neurons without leading processes was significantly increased upon overexpression of Bmp7 in comparison to a control vector in IZ (Figures 5C,D). Finally, we performed rescue experiments by introducing Bmp7 shRNA constructs into Tcf4-deficient cells. Figures 6A,B show that knockdown of Bmp7 partially rescued neuronal migration defects in Tcf4-deficient neurons. Altogether, our data provide evidence that Bmp7 is an important downstream effector of Tcf4, regulating radial migration of cortical neurons.

The shRNA target sequence of shTcf41# was 5′-GAACGGAGGATGGCCAATAAT-3′. shTcf42# was GGTCAAGATCTAGCAATAACG. α2-shRNA-2 was 5′-ACATATGCCAGTCCTGAAA-3′. Bmp7 was 5′-TCCATCTCCGTAGTATCCG-3′. shRNA sequences were designed and subcloned into pSUPER plasmid. The cDNA of mouse Tcf4 and shRNA-resistant constructs of mutants of Tcf4 were generated with the QuikChange mutagenesis kit (Stratagene) were cloned into pCAG-IRES-EGFP plasmid.

For neurosphere cultures, single dissociated cortical progenitor cells (E12.5) were cultured in serum-free DMEM medium with 20 ng/mL FGF2 and EGF (Invitrogen) for 7 DIV. To subclone, neurospheres were collected and gently dissociated using Accutase (Gibco) for 20 min at 37°C. Cells were replated at 100 cells per ul for each condition. The differentiation media was made up of low glucose DMEM with penicillin-streptomycin-glutamine, 2% B27 supplement, and 1% fetal bovine serum (Invitrogen).

Total mRNAs of the neocortex of E14.5, E17.5, P0, P7, P14, P28, and P60 mice were extracted with TRIzol reagent(Invitrogen). Super-Script II reverse transcriptase (Invitrogen) was used for reverse transcription to produce complementary DNAs (cDNAs). Real-time PCR was performed with the KAPA SYBR FAST qPCR Kits (Kapa Biosystems) and on a Roche LC96 apparatus. Primer pairs are, forward, 5′-TCTTCTCTCAGCCAACAGACAC-3′ and reverse, 5′-TTCAAGTCAGGGGAAGTTGC-3′.

In situ hybridization on the brain sections was performed with digoxigenin-labeled antisense riboprobes. Full length cDNA of Tcf4 was amplified with PCR primers and cloned into pGEM-T vector (Promega) to generate antisense and sense probes for Tcf4. The digoxigenin-labeled antisense and sense riboprobes were synthesized by in vitro transcription with DIG RNA Labeling Mix (Roche). Mice were perfused with 4% paraformaldehyde (PFA) and fixed overnight in 4% PFA at 4°C. Fixed brains were cryoprotected overnight in 30% sucrose/ phosphate buffered saline (PBS) at 4°C and mounted in OCT compound and sectioned coronally (20 mm) with a cryostat (Leica). In situ hybridization was performed as described previously (Wu et al., 2012). Briefly, brain sections were hybridized for 18 h at 65°C. The hybridization signal was detected with an alkaline phosphatase-coupled antibody (1:2,000) against digoxigenin, as well as nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate as color reaction substrates.

For immunohistochemistry, brain sections were washed in PBS for three times, and blocked in PBS supplemented with 5% BSA and 0.1% Triton X-100 for 30 min at room temperature. Thereafter, the brain sections were incubated with primary antibodies against Cux1 (1:50, Santa Cruz), Ki67 (1:1,000, Invitrogen), Cleaved Caspase-3 (1:1,000, Cell Signaling Technology), Nestin (1:200, Sigma), Pax6(1:100, Millipore), and GFP(1:1000, Molecular Probes), overnight at 4°C. After washing, sections were incubated with the correspondent secondary antibodies for 2 h at room temperature and then counterstained with DAPI (1:2,000, Sigma) for 15 min at room temperature before coverslipping.

Tissues dissected from the mouse brains and cultured cells were lysed in RIPA buffer [20 mM Tris-HCl (pH = 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% NP-40, 1% sodium deoxycholate, 1 mM PMSF, 10 mg/ml aprotinin, 1 mg/ml pepstatin A and 1 mg/ml leupeptin]. The equivalent denatured samples were subjected to SDS-PAGE, transferred and probed with antibodies against Tcf4 (1:1,000, proteintech), GAPDH (1:4,000, Abcam). Bmp7 (1:1000, Proteintech), Smad1 (1:1000, Proteintech), phosphorylated Smad1 (1:1000, Abcam), Neurogenin2 (1:1000, Abcam), and FLAG (1:1000, sigma).

E14.5 ICR-strain mice were used for in utero electroporation as described previously (Ip et al., 2012). Briefly, E14.5 pregnant mice were anesthetized with 0.7% pentobarbital sodium. The midline laparotomy was performed after the cleaning of the abdomen and the uterus were taken out. DNA plasmids (1–2 μl) of high concentration with 0.05% Fast Green (Sigma) were injected into the lateral ventricle through a polished micropipette. pSUPER shRNA was mixed with pCAG-IRES vector expressing GFP at 1:1 ratio (1 μg:1 μg) with a final concentration of 2 μg/μl. In rescue experiments, pSUPER shRNA was mixed with indicated pCAG-IRES expressing constructs at a 1:2 ratio (1 μg:2 μg) with a final concentration of 3 μg/μl. Square electric pulses were delivered at a rate of one pulse per second to embryos through the uterus by holding them with forceps-type electrodes, while the uterus was kept wet by dropping saline (prewarmed at 37°C) between the electrodes. Five electrical pulses (33V, 50 ms, 1s interval for mice) were applied across the uterine wall using electroporator (ECM-830 BTX). The uterine horns were then replaced in the abdominal cavity and the abdomen wall and skin were sutured using surgical needle and thread.

ChIPs were performed as previously described (Chen et al., 2012), using anti-H3K9Me3 (Abcam), anti-TCF4 (Abcam), anti-IgG (invitrogen), DNA released from the precipitated complexes was amplified by PCR using sequence-specific primers. Primer pairs used in Figure 5C are, Bmp7 forward GATCGGAAAGGGGTTTGTTG, reverse ACCCGAGGTCACTTGCTG; β-actin forward, AAATGCTGCACTGTGCGGCGAA, reverse TGCTCGCGGGCGGACGCGGTCTCGG.

The 1 kb region in the promoter of Bmp7 was cloned into the PGL-3 basic Vector (Promega). This construct was transfected into HEK293 cells with and without pCAG-Tcf4 using Lipofectamine 2000 in accordance with the manufacturer’s instructions (Invitrogen). Renilla was co-transfected in each well as a transfection control. Supernatant from transfected cells was analyzed 48 h after transfection. Luciferase assays were performed using the Dual Luminesence Reporter Assay system (Promega) in accordance with the manufacturer’s instructions and BioTek’s Synergy H1. The values are reported as the mean ratio of luminescence intensity (RLU) of firefly over Renilla. Values were collected from three independent experiments performed with at least three replicates perexperiment.

All date are presented as the mean ± SEM. Statistical significance was calculated using an unpaired Student’s t-test, one-way ANOVA, or two-way ANOVA. Differences were considered significant at p < 0.05. Quantification of neuronal migration was estimated by recording GFP positive neurons in distinct regions of the cerebral cortices (CP, IZ, and VZ). Two-way ANOVA followed by a Bonferroni post hoc test was used. More than 600 neurons GFP+ neurons from three brains were analyzed in each group. Quantification of morphology of neurons was estimated by recording percentage GFP posictive of neurons with or without leading process. Student’s t-test. More than 100 GFP+ neurons from three brains were examined in each group.

All methods were carried out in accordance with the approved guidelines.

All protocols was approved by the Animal Care and Use Committee of Peking University Health Science Center.