The canonical Wnt/Wingless pathway is implicated in regulating cell proliferation and cell differentiation of neural stem/progenitor cells. Depending on the context, β-Catenin, a key mediator of the Wnt signaling pathway, may regulate either cell proliferation or differentiation. Here, we show that β-Catenin signaling regulates the differentiation of neural stem/progenitor cells in the presence of the β-Catenin interactor Homeodomain interacting protein kinase-1 gene (Hipk1). On one hand, Hipk1 is expressed at low levels during the entire embryonic forebrain development, allowing β-Catenin to foster proliferation and to inhibit differentiation of neural stem/progenitor cells. On the other hand, Hipk1 expression dramatically increases in neural stem/progenitor cells, residing within the subventricular zone (SVZ), at the time when the canonical Wnt signaling induces cell differentiation. Analysis of mouse brains electroporated with Hipk1, and the active form of β-Catenin reveals that coexpression of both genes induces proliferating neural stem/progenitor cells to escape the cell cycle. Moreover, in SVZ derive neurospheres cultures, the overexpression of both genes increases the expression of the cell-cycle inhibitor P16Ink4. Therefore, our data confirm that the β-Catenin signaling plays a dual role in controlling cell proliferation/differentiation in the brain and indicate that Hipk1 is the crucial interactor able to revert the outcome of β-Catenin signaling in neural stem/progenitor cells of adult germinal niches.
During early forebrain development, cortical radial glia (RG) preferentially undergo symmetric cell divisions (Haydar et al. 2003) to tangentially expand the cortical field (Chenn and Walsh 2003). At later time points, RG switch to asymmetric cell divisions, diminish their number into germinal niches, and migrate as differentiated cells toward the cortical plate (Haydar et al. 2003). Nevertheless, a subset of RG persist in germinal neurogenic areas of the mature brain as adult neural stem/precursor cells (aNPCs) (Gotz and Huttner 2005) where they continue to generate new neurons fated to become olfactory bulbs neurons. Several molecular machineries, acting in proliferating RG of developing forebrain, are also operating in adult germinal neurogenic areas (Doetsch and Alvarez-Buylla 1996; Brill et al. 2009). Among them, the canonical Wnt signaling exerts a critical role (Dickinson et al. 1994; Lee et al. 2000; Backman et al. 2005; Subramanian et al. 2009). Wnts are secreted glycoproteins (Galceran et al. 2000; Lee et al. 2000), bind to specific cell-surface receptor belonging to the Frizzled protein family, and activate the protein Disheveled (Dsh). This latter protein, in turn, induces the reduction of Glycogen Synthase Kinase (GSK-3β) activity (Cook et al. 1996). This event leads to the accumulation of β-Catenin in the cytoplasm as well as its nuclear migration and target gene regulation (Willert and Nusse 1998). On the other hand, in the absence of the Wnt intracellular signaling cascade, GSK-3β induces β-Catenin degradation throughout its phosphorylation.
There is a large body of evidence indicating several different roles exerted by Wnt signaling in both developing and aNPCs. Wnts transiently induce RG to proliferate during early neurogenesis (Logan and Nusse 2004), induce midend gestational RG to differentiate by promoting the expression of proneural genes, such as Ngn1 (Hirabayashi et al. 2004; Israsena et al. 2004; Guillemot 2007), force aNPCs of the adult dentate gyrus to exit the cell cycle by modulating the expression of NeuroD1 (Kuwabara et al. 2009), and increase proliferating rates of Mash1+ progenitor cells belonging to the adult subventricular zone (SVZ) (Adachi et al. 2007). The effects exerted by Wnt signaling on RG versus aNPCs are largely depending on the context in which such pathway operates and, above all, on the presence of specific interactor proteins. For instance, secreted Frizzled–related proteins (Sfrps) and Dickkopf (Dkk1) are extracellular Wnt’s inhibitors (Semenov et al. 2001; Mao et al. 2002; Bovolenta et al. 2008), while Groucho proteins repress the activation of Wnt/β-Catenin downstream genes by acting in the nucleus on TCF/Lef-dependent targets (Daniels and Weis 2005).
Here, we show that the opposite role exerted by Wnts on proliferating RG versus aNPCs is largely depending on the presence of the Homeobox-interacting protein kinase1 (Hipk1) gene. Hipk1 is scarcely expressed in RG forced by Wnts to proliferate while it raises in SVZ aNPCs forced by Wnts to differentiate. We also found that Hipk1 operates within SVZ aNPCs owing to the fact that it induces the expression of p16INK4a by physically interacting with β-Catenin.
Materials and Methods
Mouse Strains and Husbandry
β-Cateninflox/flox transgenic mouse line was provided by Jackson laboratories (Brault et al. 2001), β-CateninEx3 mouse line contains 2 LoxP sequences upstream and downstream the β-Catenin third exon (Harada et al. 1999), BAT-Gal transgenic mouse line was provided by Dr Piccolo (Maretto et al. 2003), and GlastCreERT2 transgenic mouse line was provided by Dr Götz (Mori et al. 2006). GfapCre mouse line (Zhuo et al. 2001) was obtained from Jackson laboratory, and NestinCreERT2 mouse line was provided by Dr Kageyama (Imayoshi et al. 2008). Transgenic mice were repeatedly backcrossed onto C57BL/6 (Charles River) mice and were genotyped by polymerase chain reaction (PCR) as previously described (Harada et al. 1999; Brault et al. 2001; Zhuo et al. 2001; Maretto et al. 2003; Mori et al. 2006). Postnatal mice were injected with bromodeoxyuridine (BrdU) (100 mg/kg) for 9 h before the sacrifice. They were killed by anesthetic overdose and transcardially perfused with 4% paraformaldehyde in phosphate-buffered saline (PBS) pH 7.2. Brains were then fixed 4% paraformaldehyde in PBS pH 7.2 for 12 h at +4 °C and cryoprotected for 24 h in 30% Sucrose (Sigma) in PBS at +4 °C. Brains were subsequently sectioned at 10 μm. β-Galactosidase assay was performed on 200-μm thick slices adult BAT-Gal slices as previously described (Mallamaci et al. 2000). Pregnant female was killed by cervical dislocation at appropriate time points, and brains were collected in ice-cold PBS as previously described (Muzio et al. 2002). Before the sacrifice, pregnant dams received the injections of the S-phase tracer EdU (5-ethynyl-2′-deoxyuridine; Invitrogen) at the concentration of 100 mg/kg. All efforts were made to minimize animal suffering and to reduce the number of mice used, in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC). All procedures involving animals were performed according to the guidelines of the Institutional Animal Care and Use Committee of the San Raffaele Scientific Institute (IACUC number 429). GlastCreERT2 transgenic mice were crossed with β-Cateninflox/flox and Rosa26YFP mice. Starting from P30, mice received 2 Tam treatments (Tam, 2 mg/day, LASvendi) of 1 week each with a washing out period of 1 week in between. Long-term effects of β-Catenin deprivation were analyzed 2 months after the last injection. NestinCreERT2 mice were crossed with β-CateninEx3 mice, and starting from P30, double transgenic mice were injected with Tam (5 mg/day, Sigma) for 5 consecutive days. A washing-out period of time of 1 week was applied and then mice were injected with EdU for 9 h before the sacrifice.
Immunofluorescence and In Situ Hybridization
Immunofluorescence was performed as previously described (Centonze et al.). Sections were washed for 5 min 3 times in PBS and then incubated in the blocking mix (PBS 1×/FBS 10%/BSA 1 mg/mL/Triton X 100 0.1%), for 1 h at room temperature. Antibodies were diluted in blocking mix and incubated at + 4°C overnight as suggested by manufacturer’s instructions. The following day, sections were washed in PBS for 5 min, 3 times, and fluorescent secondary antibodies diluted in blocking mix (concentration suggested by the manufacturer’s instructions) were applied. Slides were washed 3 times in PBS for 5 min and incubated in Dapi solution for nuclei counterstaining. When necessary, antigens were unmasked by boiling samples in 10 mM sodium citrate (pH 6) for 5 min.
The following antibodies were used: α-mouse TuJ1 (Babco), rabbit α-phospho H3 (Upstate), rabbit α-glial fibrillary acidic protein (GFAP; Dako), mouse α-GFP (Abcam), chicken α-GFP (Chemicon), rabbit α-GFP (Molecular probes), rat α-Dcx (SantaCruz), rat α-BrdU (Abcam), mouse α-BrdU (BD), EdU detection kit (Invitrogen), rabbit α-Tbr2 (Millipore), mouse α-RC2 (Millipore), rabbit α-LacZ (Abcam), rabbit α-Olig2 (Millipore) and rabbit α-Id-1 (Bio Check Ink.), mouse α-Cre (Chemicon), rabbit α-Iba-1 (Wako), mouse α-PSA-NCAM (Santa Cruz), and rabbit α-NG2 (Chemicon). Appropriate fluorophore-conjugated secondary antibodies (Alexa-fluor 488, 546, and 633, Molecular Probes) were used. Nuclei were stained with 4′-6-Diamidino-2-phenylindole (DAPI; Roche). Id-1 was detected by using the TSA System (Perkin Elmer). Light (Olympus, BX51 with ×4 and ×20 objectives) and confocal (Leica, SP5 with ×40 objective) microscopy was performed to analyze tissue and cell staining. Analysis was performed by using Leica LCS lite software and Adobe Photoshop CS software.
In situ hybridization was performed as previously described (Muzio et al. 2002; Centonze et al. 2009). Briefly, Ten-micrometer thick brain sections were postfixed 15 min in 4% paraformaldehyde then washed 3 times in PBS. Slides were incubated in 0.5 mg/mL of Proteinase K in 100 mM Tris-HCl (pH 8) and 50 mM Ethylenediaminetetraacetic acid (EDTA) for 10 min at 30 °C. This was followed by 15 min in 4% paraformaldehyde. Slices were then washed 3 times in PBS then washed in H2O. Sections were incubated in Triethanolamine 0.1 M (pH 8) for 5 min, then 400 μL of acetic anhydride was added 2 times for 5 min each. Finally, sections were rinsed in H2O for 2 min and air-dried. Hybridization was performed overnight at 60 °C with P33 riboprobes at a concentration ranging from 106 to 107 counts per minute (cpm). The following day, sections were rinsed in SSC 5× for 5 min then washed in Formamide 50% SSC 2 × for 30 min at 60 °C. Then slides were incubated in Ribonuclease-A (Roche) 20 mg/mL in 0.5 M NaCl, 10 mM Tris-HCl (pH 8), and 5 mM EDTA 30 min at 37 °C. Sections were washed in Formamide 50% SSC 2× for 30 min at 60 °C then slides were rinsed 2 times in SSC 2×. Finally, slides were dried by using Ethanol series. Lm1 (Amersham) emulsion was applied in dark room, according to manufacturer instructions. After 1 week, sections were developed in dark room, counterstained with Dapi, and mounted with DPX (BDH) mounting solution. Mouse Hipk1 probe was generated by PCR on the basis of the National Center for Biotechnology Information (NCBI) reference (from nt735 to nt2160 of NCBI Reference Sequence NM_010432.2), and LacZ riboprobes correspond to the entire complementary deoxyribonucleic acid (cDNA) obtained from the pcDNA 3.1-LacZ plasmid.
Ten-micrometer thick sections from E16.5 forebrains were postfixed in 4% paraformaldehyde, then incubated in Proteinase K buffer for 5′ at room temperature, and then exposed to 10 mg/mL Proteinase K for 15′ at room temperature. Reaction was stopped by washing samples 3 times in PBS 1×. Slices were then incubated with terminal transferase buffer for 15′ before adding the following reagents: 10 μg/mL Biotin 16-dUTP, 1 mm CoCl2, and 10 U/mL of terminal transferase (Roche). Reaction was incubated 1 h at 37 °C and stopped with H2O. Endogenous peroxidase quenching was achieved by incubating slices in 0.1% H2O2 for 15′. Slices were incubated in blocking buffer (10% FBS, PBS 1×) for 15′ and then incubated in Streptavidin/Biotin amplification kit (Vector) for 2 h. Reaction product was visualized with 0.05% diaminobenzidine and 0.005% H2O2. Positive controls were obtained by incubating slices in 3 U/mL DNase for 15′ at room temperature.
Neurospheres cultures were derived from the SVZ of P30 BAT-Gal brains as previously described (Pluchino et al. 2008). Briefly, brains were cut in coronal sections from anterior pole of the brain. The dorsal SVZ were then microdissected from section approximately 2 mm far from the anterior pole, and tissues were subsequently incubated 30 min at 37 °C in Earle’s Balanced Salt solution supplemented with 1 mg/mL of Papain (27 U/mg, Sigma), 0.2 mg/mL Cysteine (Sigma), and 0.2 mg/mL EDTA (Sigma). Cells were then transferred in culture media Neurocult NSC proliferation medium (Stem Cell technology) supplemented with 20 ng/mL Epidermal growth factor (Provitro), 10 ng/mL Basic fibroblast growth factor (Provitro), and 2 μg/mL Heparin (Sigma). Cells were plated at the density of 8000/cm2 and maintained for more than 10 in vitro passages. Differentiation of aNPCs (n = 3 independent cultures) was obtained by using the Neurocult NSC differentiation medium (Stem Cell technology). Briefly, single suspension of cells was plated on Matrigel (BD) coated 10 mm dishes at the density of 30 000/cm2. Cells were collected every 2 days starting from the day of plating for real-time (RT) PCR assays. β-Cateninflox/flox neurospheres cultures were generated as above and infected with lentiviruses (LVs) expressing the Cre recombinase or the green fluorescent protein (GFP) (n = 3 independent experiments) as previously described (Muzio et al. 2009). Vesicular stomatitis virus-pseudotyped lentivirus generation: pRRLsin.PPT.hCMV iCre-ires-GFP, Pre plasmid (Borello et al. 2006) was transfected with the following plasmids: pMDLg/pRRE, pRSV-REV, and pMD2.VSVG into 293T cells. Then, 14–16 h after DNA transfection, the medium was substituted with fresh culture medium. Thirty-six hours later, cells supernatants were collected and filtered through a 0.22-μm pore nitrocellulose filter. To obtain high-titer vector stocks, the cells supernatants were concentrated by ultracentrifugation (55 kg, 140 min, 20 °C). Then, supernatants were discarded, and the pellets were suspended in 100 μL of PBS 1×, split in 20 μL aliquots, and stored at −80 °C. LV stocks were titrated by infecting HeLa cells with serial dilution of the viral stocks. Then, after 5 days, cells were harvested and titer was calculated on the basis of GFP expression by flow cytometry. β-Cateninflox/flox neurospheres were dissociated in single cells and plated at the concentration of 50 000/well. Soon after plating cells received iCre- or GFP-expressing lentiviruses, and cells were collected at days 8 and 9 after lentivirus delivery for western blot and RT analysis.
X-Gal staining on BAT-Gal derived aNPCs (n = 3 independent neurospheres cultures) was performed as follow. Briefly, upon fixation in 0.2% glutaraldehyde, 2% formaldehyde in 0.1 M phosphate buffer pH 7.4, supplemented with 2 mM MgCl2, and 5 mM ethyleneglycol-bis (2-aminoethylether)-N,N,N′,N′-tetra acetic acid for 5 min, cells were washed 3 times in PBS 1× and incubated in the staining solution (PBS 1×, MgCl2 2 mM, Na-Deoxycholate 0.01%, Nonidet P.40 0.02% supplemented with potassium ferricyanide 5 mM, potassium ferrocyanide 5 mM, and 1 mg/mL X-gal) overnight. Cells were then washed 3 times in PBS 1× and collected on cover slip for the analysis.
Neurospheres cultures were also nucleofected by using Amaxa Mouse NSC Nucleofector kit. Briefly, 6 × 106 single cells derived from established BAT-Gal neurospheres cultures (n = 3 independent cultures) were combined with 1 μg of Nestin-Δ90-β-Catenin-GFP with or without increasing amounts of CMV-fHIPK1 (1 or 2 μg) and then transferred in a single cuvette for nucleofection (Program A-033) accordingly with manufacturer’s instructions. After nucleofections, cells were grown in standard culture medium for 24 h before total RNA extraction.
Chromatin Immune Precipitation
Neurospheres cultures from the SVZ of P30 BAT-Gal mice2 Q3 were obtained as above described and then used for chromatin immune precipitation (ChiP) assay. Undifferentiated and differentiated neural stem cells (2 × 106) were collected and fixed in 1% formaldehyde. Extracts were incubated in cytoplasmic and nuclear lysis buffers. Sonication was performed in ice (Branson sonicator 450, 8 × 10″), and extracts were subsequently precleaned in Protein-A sepharose (Sigma) (20% of supernatants were collected as inputs). Each sample was then probed overnight with following antibodies: rabbit α-β-Catenin (cell signaling), rabbit α-GFP (molecular probes), or vehicle. Extracts were then incubated with Protein-A sepharose (Sigma) for 3 h and washed with low salt buffer, high salt buffer, LiCl buffer, and 10 mM Tris-HCl pH8, 1 mM EDTA pH8.5 buffer. Elution of complexes was performed by adding the elution buffer (1% Sodium dodecyl sulfate [SDS], 0.1%NaHCO3). Reverse cross-linking reaction was performed by incubating extracts in RNase A (0.02 mg/mL), NaCl 0.3 M at 65 °C for 4 h. Samples were then extracted with Phenol/Chloroform extraction buffer accordingly to standard laboratory procedures and resuspended in 50 μL of H2O. One microliter from each sample was used for the detection of p16INK4a promoter region by using PCR and the following primers: p16INK4a ChiP F: GTTGCACTGGGGAGGAAGGAGAGA and p16INK4a ChiP R: CCTGCTACCCACGCTAACACC. Amplicons were detected by standard agarose electrophoresis.
Total RNA was extracted by using RNeasy Mini Kit (Qiagen) according to manufacturer’s recommendations including DNase (Promega) digestion. cDNA synthesis was performed by using ThermoScript RT-PCR System (Invitrogen) and Random Hexamer (Invitrogen) according to the manufacturer’s instructions in final a volume of 20 μL. The LightCycler 480 System (Roche) and SYBR Green JumpStart Taq ReadyMix for High Throughput Q-PCR (Sigma). Depending from experiments, samples were normalized by using the following housekeeping genes: Histone H3 and β-actin. β-actin F: GACTCCTATGTGGGTGACGAGG; β-actin R: CATGGCTGGGGTGTTGAAGGTC; H3 F: GGTGAAGAAACCTCATCGTTACAGG CCTGGTAC; H3 R: CTGCAAAGCACCAATAGCTGCAC TCTGGAAGC. Gene specific primers: p16INK4a F: CCTGGAACTTCGCGGCCAATCCC; p16INK4a R: GCTCCCTCCCTCTGCTCCCTCC; p19Arf F: CTGGGGGCGGCGCTTCTCACC; p19Arf R: TCTAGCCTCAACAACATGTTCACG; Human p16INK4a F: TTCCTGGACACGCTGGTGGTG; Human p16INK4a R: ATCGGGGATGTCTGAGGGACC; Hipk1 F: GCTAGCTGACTGGAGGAATGCC; Hipk1 R: TGGTCTTGGACAGGAACTAGGG; Ki67 F: CCGAACAGACTTGCTCTGGCCTAC; Ki67 R: CTGGGCTGTGAGTGCCAAGAGAC; Lef1 F: CCGTGGTGCAGCCCTCTCACGC; Lef1 R: ATTTCAGGAGCTGGAGGGTGTCTGG.
TOP-Flash luciferase reporter plasmid was a gift from R. T. Moon (University of Washington, Seattle), while Fop-Flash plasmid was purchased by Chemicon. Luciferase assays were performed by using the dual-luciferase assay system (Promega, Madison, WI) accordingly with manufacturer’ instructions. Hek293T cells (4.5 × 104) were seeded in 12-well plates and transfected with TOP-Flash (0.5 μg), prlTK (0.1 μg), Δ90-βCatenin/GFP (0.5 μg), and increasing concentration of fHipk1 (0.1, 0.5, 0.7, 1, and 2 μg). Control experiments were done by transfecting cells with Fop-Flash plasmid (0.5 μg), prlTK (0.1 μg), Δ90-βCatenin/GFP (0.5 μg), and fHipk1 (2μg). Measurement of luciferase activity was done by using the dual-luciferase system (Promega) on a luminometer (GloMax 20/20 Luminometer; Promega). Relative luciferase activity was reported as a ratio of firefly over Renilla luciferase readouts.
Western Blot and Protein Coimmunoprecipitation
Western blot detection of antigens was performed as previously described (Centonze et al. 2009). Briefly, 106β-Cateninflox/flox cells were infected with Cre-ires-GFP or GFP lentiviruses at Multiplicity Of Infection of 10. Upon infection, cells were collected, respectively, at days 8 and 9. Total cell extracts were obtained by incubating cells in lyses buffer (Sucrose 320 mM [Sigma], 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid 1 mM pH 7.4 [Sigma], and MgCl2 1 mM (Sigma) supplemented with Protease inhibitor cocktail, Sigma). Ten micrograms of each protein extract was run on SDS–polyacrylamide gel electrophoresis (PAGE) electrophoresis and subsequently blotted on nitrocellulose (Millipore). The following primary antibodies were used: rabbit α-β-Catenin (Cell signaling), rabbit α-GFP (Chemicon), mouse α-Active Caspase3 (Cell Signaling), and mouse α-β-Actin (Sigma). Secondary antibodies (Biorad) were incubated for 2 h at RT, and signals were revealed by using Millipore ECL kit.
Coimmunoprecipitation was performed on HEK-293T cells transfected with the following plasmids: TOP-Flash luciferase reporter, Δ90-β-Catenin/GFP, and pCMV-fHipk1. Briefly, 106 HEK-293T were transfected with 2 μg of TOP-Flash plasmid, 4 μg of Δ90-β-Catenin/GFP with or without 4 μg of pCMV-fHipk1. Forty-eight hours later, cells were incubated in lyses buffer (Tris-HCl pH 7.5 20 mM, NaCl 100 mM, EDTA 5 mM, Triton 1%, Glycerol 10%, Na3VO4 1 mM, Na4P2O7 40 mM supplemented with Protease inhibitor cocktail, Sigma). Then, 50 μL of Protein-A (Sigma) sepharose beads were incubated with lysate for 1 h at +4 °C. Lysates were then incubated with mouse α-Flag (2 μg/mL, Sigma) for 2 h and then 100 μL of Protein A sepharose were added to the lysates. Lysates were further incubated for 2 h with Protein A beads, and subsequently, beads were separated by centrifugation. Protein A beads were subsequently washed 6 times with lyses buffer, and finally, elution was performed by adding SDS loading buffer directly to beads. Co-IP samples, inputs, and washes were run on SDS–PAGE and then blotted on nitrocellulose. Rabbit α-GFP (Abcam) primary antibody was used to detect Δ90-β-Catenin-GFP protein. Western blot represents 1 over 3 independent experiments. Total extract from parallel transfections was used for RT-PCR detection of human p16INK4a expression as above described.
In Utero Electroporation
The following plasmids were introduced into progenitor cells of E13.5 (CD1, Charles River) embryos: pCAG-GFP (0.1 μg/μL), Nestin-Δ90-β-Catenin-GFP (1 μg/μL), pCMV-fHipk1 (1 μg/μL). Plasmids were mixed with 0.01% Fast green (Sigma), and 1–2 μL of each DNA mix was injected into the ventricle through a fine glass capillary. Electrodes (Tweezertrodes, BTX Harvard Apparatus) were placed flanking the ventricular region of each embryo covered by a drop of PBS and pulsed 4 times at 40 V for 50 ms separated by intervals of 950 ms with a square wave electroporator (ECM 830, BTX Harvard Apparatus). Then, the uterine horn was placed back into the abdominal cavity and filled with warm PBS 1×. Accordingly to the experimental plan, brains were collected 3 days or 40 h after electroporations. BLOCK-it miR RNAi selected oligos for Hipk1 interference (Mmi-51128; Mmi-51129; Mmi-51130; Mmi-51131, efficient functioning was tested by transfecting 293T cells) were purchased by Invitrogen, cloned into pcDNA6.2-GW/miR vector, and electroporated at the concentration of 250 ng/μL each into E13.5 embryos. Controls were electroporated with pcDNA6.2-GW/miR LacZ at the concentration of 1 μg/μL. Embryos were collected at E16.5 and pulsed with EdU 2 h before the sacrifice. Electroporation was also done at E16.5 (CD1, Charles River). Embryos received pCAG-GFP (0.1 μg/μL) with or without the Nestin-Δ90-β-Catenin-GFP (1 μg/μL) plasmids. Electroporation was done as above described, and brains were collected at P10. Six hours before the sacrifice mice were repeatedly injected with EdU (100 mg/kg every 2 h). GFP-expressing brains were fixed in 4% paraformaldehyde and processed for immune fluorescence. Only embryos showing comparable electroporated patches were included in our analysis.
Bar graphs represent the mean value ± standard error of the mean or the mean value ± standard deviation. Data were analyzed as appropriate by Student’s t-test or by one-way analysis of variance (ANOVA) using Graph-Pad Prism version 4. A Significance was accepted when P < 0.05.
β-Catenin Stimulates Cell Proliferation and Inhibits the Expression of Tbr2 in Progenitors of the Developing Forebrain
We generated embryos expressing the constitutively active form of β-Catenin by crossing the β-CateninEx3 transgenic mouse line, which contains 2 loxP sequences upstream and downstream the third exon of β-Catenin (Harada et al. 1999) with GfapCre mice. In this new transgenic mouse line, recombined β-Catenin becomes stabilized and constitutively active in the nucleus in virtually all RG of E13.5 cerebral cortex and the hippocampus (Fig. 1A) (Zhuo et al. 2001). GfapCre/β-CateninEx3 brains collected at P21 and P30 displayed a significant expansion of the tangential surface of the brain (Fig. 1B and not shown) and a severe reduction of the cortical thickness (Fig. 1C and not shown). We further characterized double transgenic mice at E14.5, that is, 1 day after the activation of the constitutively active form of β-Catenin; they showed a significant increase of the cerebral cortex surface (Fig. 1D–F). At this stage, we also analyzed the effects of β-Catenin stabilization on RG by staining sections for the RG marker RC2 (Hartfuss et al. 2001). In wild-type (WT) mice, RG showed the classical palisade organization, while they were severely disorganized in GfapCre/β-CateninEx3 brain, displaying distortion of cytoplasmic bundles at the level of the ventricular zone (VZ)/SVZ (Fig. 1G,I). We next measured the thickness of the neural layer (NL) by staining E14.5 sections for TuJ1. NL layer was significantly reduced in mice overexpressing stabilized β-Catenin (Fig. 1K,L). In contrast, the thickness of the germinal layer did not show any alteration (Fig. 1M). The distribution of pH3+ proliferating cells in the germinal layer of double transgenic mice was also severely altered. In WT embryos, pH3+ proliferating cells were placed within the ventricular lining and the SVZ (Carney et al. 2007), while in GfapCre/β-CateninEx3 mice, they were scattered throughout the entire germinal layer (Fig. 1N–P). These scattered proliferating cells may represent abventricular mitoses of Tbr2+ basal progenitor (BP) cells (Sessa et al. 2008). However, in GfapCre/β-CateninEx3 mice, the total number of Tbr2+ cells was significantly reduced (Fig. 1Q–S), confirming previous results that showed a delay of maturation of BP cells in mice overexpressing the Wnt signaling (Wrobel et al. 2007; Mutch et al.). To study proliferating RG and Tbr2+ cells and the contribution of BP cells to the total proliferating pools of the forebrain, embryos were pulsed with the s-phase tracer EdU. Proliferating Tbr2+ cells accounted for a small fraction of the brain total proliferating pool (Fig. 1T). Nevertheless, the fraction of Tbr2+ cells, incorporating the EdU tracer, over the total number Tbr2-expressing cells was similar in both genotypes (Fig. 1T), suggesting that their limited number derives from a derangement of their genesis rather than from alterations of their progression along the cell cycle. Thus, the persistent activation of the canonical Wnt signaling forces RG to maintain their cell identity and to reenter the cell cycle, inducing a tangential expansion of the cortical field. Similar results have been reported in previously published works (Wrobel et al. 2007; Mutch et al.), however, very few data are available on the role of β-Catenin at later time points of forebrain maturation. In order to fill this gap, we overexpressed the stabilized form of β-Catenin (Nes-Δ90-β-Catenin-GFP) (Chenn and Walsh 2002, 2003) at the end of neurogenesis by in utero electroporation. E16.5 embryos received Nes-Δ90-β-Catenin-GFP along with the GFP-expressing plasmid, while control embryos were injected with the GFP-expressing plasmid alone. Brains were then collected at P10, and GFP+ cells were scored in the SVZ (Supplementary Fig. 1A,B) and in the cortical plate (Supplementary Fig. 1D,E). The active form of β-Catenin significantly increased the number of GFP+ cells that were retained in the SVZ (Supplementary Fig. 1C). Since mice were pulsed with the EdU at the day of the sacrifice, we also established the fraction of proliferating cells in the SVZ. The number of GFP/EdU double positive was significantly increased in mice overexpressing Nes-Δ90-β-Catenin-GFP plasmids (Supplementary Fig. 1F,G). Thus, unlike previous published results (Hirabayashi et al. 2004), our data suggest that the activation of Wnt/β-Catenin signaling induces RG to proliferate also at late gestational time points similarly to earlier embryonic stages (Chenn and Walsh 2002; Wrobel et al. 2007; Mutch et al.).
Confined Expression of the Canonical Wnt Signaling in the Postnatal SVZ
Given the pivotal role of β-Catenin in regulating cell proliferation during forebrain development (Chenn and Walsh 2002; Wrobel et al. 2007; Mutch et al.), we turned our attention to the postnatal SVZ. We used the reporter BAT-Gal mice (Maretto et al. 2003) to trace cells in which the canonical Wnt/β-Catenin signaling is still active. These mice have multiple TCF/Lef-binding sites coupled to a minimal promoter that drives the expression of the LacZ reporter (Maretto et al. 2003). The analysis of P30 BAT-Gal brains showed the presence of many X-Gal+ cells that were placed in the dorsal but not in the ventrolateral SVZ (Fig. 2A and inset) (Doetsch et al. 1999; Temple and Alvarez-Buylla 1999). The presence of LacZ transcripts in the dorsal SVZ was confirmed by in situ hybridization (Fig. 2B). A subset of LacZ-expressing cells incorporated the BrdU tracer, suggesting that they belong to a proliferating SVZ aNPCs (Fig. 2C). In addition, many LacZ+ cells coexpressed Olig2, which labels parenchymal oligodendrocyte precursors and type-C aNPCs of the SVZ (Fig. 2D). Some LacZ+ cells coexpressed Dcx, which is a marker for type-A aNPCs (Lois et al. 1996; Hack et al. 2005) (Fig. 2E), while a relatively high number of them coexpressed GFAP, a marker of astrocytes in the CNS parenchyma and type-B aNPCs in the SVZ (Fig. 2F) (Doetsch et al. 1999; Ahn and Joyner 2005; Tavazoie et al. 2008). LacZ+ cells of the most dorsal SVZ were positive for Id-1 (Fig. 2G), a marker of SVZ-restricted slow dividing aNPCs (Nam and Benezra 2009). The SVZ, however, contains also microglia and oligodendrocytes (Menn et al. 2006). Thus, we tested whether some of LacZ+ cells might belong to these lineages by probing sections for Iba-1 or NG2. However, neither LacZ/Iba-1 nor LacZ/NG2 double positive cells were detected in the SVZ of BAT-Gal mice (Fig. 2H,I). Taken together, these results indicate, for the first time, that the canonical Wnt signaling is confined in adult brains to a subset of aNPCs of the dorsal SVZ.
The Activation of β-Catenin Signaling in the Adult SVZ inhibits Cell Proliferation and Activates the Expression of p16INK4a
In order to define the contribution of the canonical Wnt signaling to adult neurogenesis, we took advantage an inducible Cre allele. NestinCreERT2 transgenic mouse line expresses the Tamoxifen-dependent Cre recombinase in aNPCs of the SVZ (Imayoshi et al. 2008). Starting from P30, NestinCreERT2/β-CateninEx3 mice and their control litters—that is, mice containing only the NestinCreERT2 allele—were treated with Tamoxifen for 5 consecutive days. Brains were collected 1 week after the last Tamoxifen injection. At the day of the sacrifice, we labeled proliferating cells using EdU. Tamoxifen treatments did not alter gross morphology but significantly inhibited aNPC proliferation in the SVZ (Fig. 3A–C). On the other hand, the number of PSA-NCAM+ cells was significantly increased in the SVZ of NestinCreERT2/β-CateninEx3 mice (Fig. 3D–F), thus suggesting that the stabilization of β-Catenin in the adult SVZ inhibits cell proliferation, favoring cell differentiation. Accordingly, also the number of Id-1+ slow dividing progenitor cells was greatly diminished in double transgenic mice (Fig. 3G,H).
A large body of experimental evidence suggests that β-Catenin signaling, in certain circumstances, might positively or negatively regulate of the cell cycle–dependent kinase inhibitor (Cdki) p16INK4a (Saegusa et al. 2006; Delmas et al. 2007). For example, in tumor cells, the β-Catenin–mediated modulation of p16INK4a forces these cells to differentiate (Saegusa et al. 2006). To determine whether or not β-Catenin can modulate p16INK4a in aNPCs of the adult SVZ, we microdissected SVZs from NestinCreERT2/β-CateninEx3 mice administered with Tamoxifen as above (Fig. 3I). The stabilization of β-Catenin induced the expression of the target gene Lef1 (Fig. 3J) (Hovanes et al. 2001; Filali et al. 2002) as well as of p16INK4a (Fig. 3K), suggesting that the activation of ectopic β-Catenin is operating, and such manipulation induced the expression of p16INK4a, that in turn orchestrate the opposite role of β-Catenin functioning in adult germinal niche. In order to confirm these results, we raised neurospheres cultures (Reynolds and Weiss 1992) from the dorsal SVZ of P30 BAT-Gal mice (Maretto et al. 2003) to establish if the canonical Wnt signaling is still working in vitro. X-Gal staining of neurospheres revealed the presence of LacZ+ cells in approximately 40% of them (Supplementary Fig. 2A). However, upon serial in vitro passages (ivp), this number greatly diminished, possibly because the prolonged exposure to high levels of growth factors may hamper their capacity to express Wnt genes (not shown). Early generated neurospheres cultures were used for chromatin immunoprecipitation (ChIP) of β-Catenin targets. By using a specific antibody for β-Catenin, we demonstrated that the promoter region of p16INK4a is a direct target of β-Catenin (Supplementary Fig. 2B). In addition, we also demonstrated that p16INK4a messenger RNA (mRNA) levels were significantly increased when aNPCs were induced to differentiate by absence of growth factors (Supplementary Fig. 2C). We also raised neurospheres cultures form P30 β-Cateninflox/flox SVZs (a β-Catenin conditional knockout allele) (Brault et al. 2001). These cultures were infected with lentivirus expressing either the Cre-IRES-GFP cassette (Borello et al. 2006) or the GFP gene. Starting from day 8, after lentivirus delivery, the Cre-mediated recombination of the β-Cateninflox/flox locus abolished the expression of β-Catenin (Supplementary Fig. 2D). Because β-Catenin is a component of adherens junctions (Aberle et al. 1996), its depletion deranged cell morphology of targeted aNPCs and affected their long-term survival. Indeed, 9 days after lentivirus delivery, we detected a slight activation of Caspase-3 (Supplementary Fig. 2D), and starting from 10 days after viruses, delivery targeted cells did not adhere to each other (not shown) (Holowacz et al.). At day 8, a time when infected cells have already lost β-Catenin expression but the activation of Caspase-3 is undetectable (Supplementary Fig. 2D), we observed a significant reduction of p16INK4a expression (Supplementary Fig. 2E). Because the INK4a locus contains 2 distinct gene variants, the p16INK4a and the p19Arf, in addition to the short p16INK4a variant, we measured also the expression of the long p19Arf mRNA levels that did not show any alteration (Supplementary Fig. 2E). Thus, we demonstrated that β-Catenin can directly interact with the promoter region of p16INK4a in vivo and in aNPCs derived from the adult SVZ.
Dynamic Expression of Hipk1 during Forebrain Neurogenesis
The context-dependent role of β-Catenin pathway in promoting cell self-renewal or neuronal differentiation may rely on the presence of specific molecular interactors (Cavallo et al. 1998). Some of these molecules dictate the outcomes, in term of cell proliferation and cell differentiation, for the canonical Wnt/β-Catenin signaling (Sierra et al. 2006; Li and Wang 2008). Among them, Hipk genes have been recently indicated as modulators of β-Catenin signaling in the neural plate cells of Xenopus laevis embryos (Louie et al. 2009) and in stem cells of the skin (Wei et al. 2007). Hipk1 is expressed in germinal niches of E9.5, E14.5, E18.5, and P30 brains, as we showed by in situ hybridization assay (Fig. 4A–E) (Isono et al. 2006). However, starting from E14.5, also cells of the presumptive hippocampal plate and cells of the choroid plexus exhibited detectable levels of Hipk1 (Fig. 4B,C). Thus, the pattern of expression of Hipk1 overlaps the pattern of expression of the Wnt signaling (Maretto et al. 2003; Muzio et al. 2005). Hipk1 expression levels are very low during forebrain development but dramatically raise in germinal niches of the adult brain (Fig. 4D). Indeed, by coupling in situ hybridization for Hipk1 with immunofluorescence-mediated detection of GFAP, we were able to detect, in the adult SVZ, a large number of periventricular GFAP+ cells coexpressing at high levels Hipk1 mRNA (Fig. 4E). The expression of Hipk1 levels was also evaluated by RT-PCR on cortical microdissections of developing forebrains (E9.5, E11.5, E14.5, E18.5, and P4) and postnatal SVZs (P10 and P30). Hipk1 mRNA levels were normalized on the expression of the housekeeping gene β-Actin, and data from each time point were compared with the normalized expression of Hipk1 detected at E9.5. During forebrain development, Hipk1 expression levels were pretty low, if compared with levels measured in P30 SVZs (Fig. 4F). We also established that Hipk1 was also expressed in neurospheres cultures (Fig. 5F), and its expression increased during their differentiation (Fig. 4G).
A large number of genes implicated in the control of Wnt/β-Catenin signaling are often regulated by β-Catenin through positive or negative feedback loops (Kazanskaya et al. 2004; Logan and Nusse 2004; Chamorro et al. 2005; Khan et al. 2007), we asked whether Hipk1 expression might be modulated by the β-Catenin. We firstly induced the β-Catenin depletion by crossing β-Cateninflox/flox mice (Brault et al. 2001) with GlastCreERT2 transgenic mice (Mori et al. 2006). Cells expressing Glast, Gfap, and Nestin are presumably high hierarchical aNPCs of the adult SVZ that are capable to self-renew in vitro and to generate multipotent neurosphere cultures (Ninkovic et al. 2007; Beckervordersandforth et al.). Adult GlastCreERT2/β-Cateninflox/flox and their controls—that is, GlastCreERT2/β-Cateninflox/+—were fed with Tamoxifen and sacrificed 4 weeks after Tamoxifen treatment. Efficiency of recombination was tested by including the Rosa26YFP allele. Virtually, all cell or the periventricular lining expressed robust yellow fluorescent protein (YFP) levels (not shown). Robust Hipk1 expression levels were detected in controls (Supplementary Fig. 6A) but significantly reduced in conditional KO mice (Supplementary Fig. 6B). Accordingly, neurospheres deprived of β-Catenin significantly lowered Hipk1 expression levels (Supplementary Fig. 5C). We also examined Hipk1 in brains overexpressing the stabilized form of β-Catenin. NestinCreERT2/β-CateninEx3 mice administered with Tamoxifen for 5 consecutive days were collected 1 week after the last Tam injection and probed for Hipk1 detection. In parallel, the SVZ of some of them was microdissected for RT-PCR analysis. However, the stabilization of β-Catenin produced a slight, but not significant, increase of Hipk1 levels as measured by RT-PCR (Supplementary Fig. 6D–F). Thus, the inactivation of the β-Catenin signaling inhibited Hipk1 expression, while the upregulation of this molecule did not.
In the forebrain, early ectopic activation of the canonical Wnt signaling fosters RG to proliferate thus promoting an excessive tangential expansion of the cortical field (Chenn and Walsh 2002, 2003). By contrary, later time point activation of Wnt signaling—via stabilization of the active form of β-Catenin—fosters RG to differentiate (Hirabayashi et al. 2004). This latter effect is due to the positive regulation of the proneural gene Ngn1 expression (Hirabayashi et al. 2004). By studying cell proliferation of GfapCre/β-CateninEx3 brains, we found that the persistent activation of Wnt signaling promotes RG proliferation also at later time points of forebrain development. RG proliferation increased in E13.5 embryos once the expression of the activated form of β-Catenin was transgenically induced. Consistently with this finding, the electroporation of Nes-Δ90-β-Catenin-GFP plasmid at E16.5 induces targeted RG to proliferate and to inhibit the expression of TuJ1 and Tbr2. Interestingly, we also observed a severe alteration of RG morphology in our transgenic mice consisting with the presence of a significant high number of abventricular mitoses in the VZ/SVZ. At this stage, we can only speculate that the active form of β-Catenin might have directly influenced the positioning of pH3+ mitoses by regulating nuclear interkinesis.
It is well known that Wnts are expressed by aNPCs residing within neurogenic niches (Lie et al. 2005; Adachi et al. 2007). However, their role is still partially unclear. As a matter of fact, the presence of a stabilized form of β-Catenin increases proliferation of Mash1 expressing aNPCs placed in both the ventral and the dorsal SVZ (Adachi et al. 2007). On the other hand, Wnt signaling induces differentiation of aNPCs residing within the dentate gyrus of the hippocampus (Lie et al. 2005) by activating the expression of NeuroD1 (Kuwabara et al. 2009). We found that Wnt signaling is activated only in the aNPCs residing within the dorsal SVZ by using BAT-Gal reporter mice. As such, BAT-Gal–expressing cells of the dorsal SVZ incorporate the S-phase tracer BrdU and a large part of them express molecular markers of SVZ-restricted aNPCs. Similarly to the results shown by Li and Wang (2008), we found that Wnt signaling significantly reduced the number of proliferating aNPCs in the SVZ: a 2-fold increase of type-A neuroblasts was measured in NestinCreERT2/β-CateninEx3 mice. Finally, we observed that the active form of β-Catenin induces the expression of the p16INK4a in aNPCs of the SVZ (Lukas et al. 1995). Reduced expression level of p16INK4a was observed upon inactivation of the expression of β-Catenin in SVZ-derived neurosphere cultures derived from β-Cateninflox/flox mice. This was attributed to the capability of β-Catenin to engage the promoter region of the mouse p16INK4a (Saegusa et al. 2006). While technical hurdles (i.e., retroviral vector-based cell tracing vs. inducible transgenes) might explain the different results we obtained from Adachi et al. (2007), it is reasonable to think that the activation of the Wnt signaling in different SVZ cell populations might lead to different phenotypes. Future studies must address the relative contribution of Wnts in each specific cell population belonging to the SVZ.
The opposite effects produced by β-Catenin signaling in aNPCs may be due to signals capable of positively or negatively regulate β-Catenin target genes. Among such proteins those coded by Hipk genes—Hipk1, Hipk2, Hipk3, and Hipk4—have been shown to be reliable modulators of the β-Catenin signaling (Rinaldo et al. 2007). Hipk1 can regulate the transcription of Wnt/β-Catenin targets during X. levis gastrulation (Louie et al. 2009). Hipk2 interacting with β-Catenin and repressing β-Catenin/Lef1 target genes impairs self-renewal of skin stem cells (Wei et al. 2007): This suppressive effect does not require the phosphorylation of β-Catenin critical residues (Wei et al. 2007). The constitutive inactivation of Hipk1 and Hipk2 induces early exencephaly, which is characterized by an exaggerate overgrowth of forebrain/midbrain neural tissue (Isono et al. 2006). While the expression pattern of Hipk1 has been deeply investigated in early embryos (Isono et al. 2006), still few data are available at later time points during forebrain maturation. We found that Hipk1 is expressed within developing and adult neurogenic germinal niches: Its expression levels are generally low during forebrain development but significantly increased in aNPCs of the adult SVZ, when Wnts induced aNPCs differentiation. We also found that overexpression of Hipk1 during embryonic development—when Hipk1 expression levels are low—slightly affected RG proliferation. However, the coelectroporation of Hipk1 with the constitutive active form of β-Catenin significantly reduced the number of proliferating cells and a significant number of RG receiving both constructs differentiated into TuJ1+ cells. While excluding that our findings could be due to the exaggerated rate of cells death induced by overexpression of Hipk1 (D'Orazi et al. 2002; Doxakis et al. 2004), we finally demonstrated that Hipk1 was able to physically interact with β-Catenin and that Hipk1 is able to regulate the β-Catenin–mediated expression of the reporter plasmid TOP-Flash.
Since we demonstrated that β-Catenin is capable to engage the promoter region of the mouse p16INK4a and that Hipk genes are able to physically interact with β-Catenin, we finally explore the possibility that Hipk1-β-Catenin interaction regulates p16INK4a expression. In neurosphere cultures, the overexpression of β-Catenin slightly modulated the expression of p16INK4a. However, increased expression of p16INK4a was obtained in the presence of both Hipk1 and the active form of β-Catenin. Our data, showing that Hipk1 might cooperate with the canonical Wnt signaling in the adult SVZ, parallel the above-mentioned previous results obtained showing that Hipk2 and β-Catenin act jointly as cell cycle repressors in skin stem cells (Wei et al. 2007). However, we cannot exclude that Hipk genes can interact with other molecules to regulate the cell cycle as Hipk2 can interact with Axin to stimulate the expression of P53 (Rui et al. 2004).
In conclusion, our data further reiterate the strong involvement of the Wnt signaling pathway in the regulation and orchestration of the cell cycle in embryonic and adult neural progenitors.
BMW Italy group and Italian Association for Multiple Sclerosis (Fism/Aism) grant number “Stem cells 2008.”
We deeply thank Dr Furlan and Dr P. Brown for their helpful discussions and support. We thank Dr Hiroyoshi Ariga for providing the Hipk1 expressing construct, Dr Stefano Piccolo for providing BAT-Gal transgenic mice, Dr R. Kageyama for providing NestinCreERT2 transgenic mice, Dr C. Walsh for providing the Nestin-Δ90-β-Catenin-GFP construct, and Dr U. Borello for the Cre-ires-GFP construct. This work is dedicated to the memory of Dr G. Corte. Conflict of Interest : None declared.