Wnt7a Regulates Multiple Steps of Neurogenesis

Immunostaining and quantification.

Immunostaining was carried out on 20- or 40-μm brain sections or cultured cells using antibodies, including rat antibromodeoxyuridine (anti-BrdU) (Accurate; 1:5,000), mouse anti-Ki67 (Lab Vision; 1:400), rabbit anti-Tuj1 (Covance; 1:5,000), mouse anti-glial fibrillary acidic protein (anti-GFAP) (Sigma; 1:1,000), mouse anti-O4 (Sigma; 1:400), mouse anti-Map2 (Sigma; 1:600), goat anti-Sox2 (Santa Cruz; 1:100), and goat antidoublecortin (anti-DCX) (Santa Cruz; 1:200). Images were visualized by a Zeiss confocal or inverted microscope. In vivo BrdU labeling was performed as described previously (24). For BrdU and DCX double labeling, mice were injected with BrdU intraperitoneally once daily at 50 mg/kg of body weight for 7 days. Brains were harvested 3 weeks later for BrdU and DCX costaining. For quantification, every eighth section of slides per animal (about 8 sections per brain), including the hippocampal dentate gyrus region, was selected. The dentate gyri were scanned using a confocal microscope. Z series of confocal images were captured. The number of single- or double-stained cells was counted.

For analysis of dendritic development in vivo, a three-dimensional (3D) reconstruction of entire dendritic processes of neurons was made from Z stacks of confocal images. The two-dimensional (2D) projection images were traced with NIH Image J. Dentate granule cells with largely intact dendritic trees were analyzed for total dendritic length as described previously (25). Dendritic complexity was measured using Sholl analysis plugged into NIH Image J by counting the number of dendrites that cross a series of concentric circles from the cell soma as described previously (25). A minimum of 10 neurons from randomly picked sections per animal and six animals was analyzed for neuronal morphology under each condition. Two-way analysis of variance (ANOVA) was used for statistical analysis.

Neural stem cell culture, proliferation, differentiation, and self-renewal analysis.

Mouse neural stem cells were isolated from 6-week-old WT or Wnt7a^−/− brains (Jackson Laboratory) using methods modified from published protocols (24, 26). Specifically, adult mouse forebrains containing the hippocampal dentate gyri and subventricular zones were sliced into 400-μm sections. Brain tissues were dissociated using the MACS neural dissociation kit (Miltenyi). Dissociated cells were centrifuged in a 10% Percoll solution in phosphate-buffered saline (PBS). Cells were collected and plated in neurobasal medium containing 2% B27, 1% glutamine, 20 ng/ml epithelial growth factor (EGF), and fibroblast growth factor 2 (FGF2). Newly formed neurospheres were dissociated into single cells using trypsin and transferred into Dulbecco's modified Eagle medium (DMEM)–F-12 medium supplemented with N2, 20 ng/ml EGF and FGF2, and 5 μg/ml heparin and cultured in this medium for further passages.

BrdU labeling was performed to monitor neural stem cell proliferation by treating neural stem cells with 10 μM BrdU for 1 h. The BrdU-treated cells were HCl treated, followed by BrdU antibody staining. Clonal analysis was performed by dissociating neurospheres into single cells and plating them at one cell per well in 96-well plates. These cells were cultured for 2 weeks to allow primary neurosphere formation. The primary neurospheres were dissociated into single cells and seeded at 50 cells per well in 96-well plates to allow secondary sphere formation. Data presented are secondary sphere size and formation rate.

For differentiation assays, neural stem cells were dissociated into single cells and cultured in DMEM–F-12 medium, supplemented with N2, 1 μM retinoic acid (RA), and 0.5% fetal bovine serum (FBS). Three days after differentiation, cells were immunostained for Tuj1, GFAP, and O4. The Tuj1-positive, GFAP-positive, or O4-positive cells were counted and calculated for the percentage of Tuj1-positive, GFAP-positive, or O4-positive cells. More than 5,000 cells were counted for each experiment. To measure dendritic length and complexity of differentiated neurons, neural stem cells were differentiated in the same differentiation medium for 5 days. Dendritic length was measured using NIH Image J. Dendritic complexity was measured using Sholl analysis plugged into NIH Image J.

In vivo cell cycle analysis.

Six-week-old WT and Wnt7a^−/− littermate mice were injected with BrdU intraperitoneally at 50 mg/kg of mouse body weight, and animals were allowed to survive for 30 min or 24 h. Brains of the treated animals were collected for cell cycle analysis as described previously (27). For cell cycle length analysis, mice were treated with BrdU for 30 min, followed by BrdU and Ki67 double staining of brain sections. The fraction of progenitor cells in S phase, a labeling index, was determined by scoring the percentage of BrdU-labeled neural progenitor cells (BrdU⁺ Ki67⁺/Ki67⁺). This labeling index provides an estimation of cell cycle length. Decreased labeling index suggests lengthened cell cycles. For cell cycle reentry analysis, mice were treated with BrdU for 24 h, followed by BrdU and Ki67 double staining of brain sections. Cells that left the cell cycle were identified as BrdU⁺ and Ki67⁻, and cells that reentered the cell cycle were identified as BrdU⁺ and Ki67⁺. Percentage of cell cycle reentry was expressed as BrdU⁺ Ki67⁺/BrdU⁺ cells.

Apoptosis assay.

We used the active form of caspase-3 as an indicator for apoptosis. Immunostaining was used to detect active caspase-3 in both WT and Wnt7a^−/− neural stem cells by incubation with an antibody specific for the cleaved, active caspase-3 (Cell Signaling; 1:100), followed by incubation with the fluorescein isothiocyanate (FITC)-conjugated secondary antibody.

Viral vector construction, viral preparation, and transduction.

The DNA sequence containing the Wnt7a small interfering RNA (siRNA) was inserted into the pHIV lentiviral vector with a puromycin resistance gene. The neurogenin 2 (Ngn2) lentivirus vector (Addgene plasmid 34999) was purchased from Addgene. The constitutively active β-catenin that lacks its amino-terminal 90 amino acids was cloned into the NIT retroviral vector as described previously (14). Viral preparation and transduction were performed as previously described (14).

Real-time RT-PCR analysis.

Total RNA was purified using TRIzol reagent (Invitrogen). Reverse transcription was performed with 1 μg of RNA using the Omniscript reverse transcription kit (Qiagen). Real-time reverse transcription-PCR (RT-PCR) was performed using iTaq SYBR green Supermix with 5-carboxy-X-rhodamine (ROX) (Bio-Rad) in an Applied Biosystems Step One Plus real-time PCR system. The primers for RT-PCR included the following: Wnt7a forward, 5′ CTG TGG CTG CGA CAA GGA GA 3′; Wnt7a reverse, 5′ TGG CAC GCA CAG GCT CCA CGT GG 3′; cyclin D1 forward, 5′ GAG ACC ACA GCC CTC CCC AGA CGG 3′; cyclin D1 reverse, 5′ AAG CGG TCC AGG TAG TTC ATG G 3′; Ngn2 forward, 5′ TCG TCA AAT CTG AGA CTC TGG AG 3′; Ngn2 reverse, 5′ CGG CGC AGC TCC TCG TCC T 3′; β-actin forward, 5′ CCG AGC GTG GCT ACA GCT TC 3′; and β-actin reverse, 5′ ACC TGG CCG TCA GGC AGC TC 3′.

ChIP assays.

We performed chromatin immunoprecipitation (ChIP) assays using the EZ-ChIP kit (Millipore) with precleaned chromatin from 2 × 10⁶ mouse neural stem cells and 5 μg antibody for each ChIP assay using antibodies for β-catenin (BD BioScience; 610153), acetylated histone H3 (AcH3) (Millipore; 06-599), and trimethylated histone H3 lysine 4 (H3K4me3) (Abcam; Ab8580). Mouse IgG (IgG1) and rabbit IgG (IgG2) were included as the negative controls. The primers for ChIP assays included the following: cyclin D1 F, 5′ AGG GAC TTC AAC ATA AAT CAT GC 3′; cyclin D1 R, 5′ GCA CCA CCA TAC CCA GCA TAT GC 3′; Ngn2 F, 5′ CTG ACA CAG CAG GTT CCT CTC AT 3′; and Ngn2 R, 5′ AGC ATT GAG ACC AAA CAC TTA AC 3′.

Golgi staining.

Golgi staining was performed as described previously (28). Briefly, brains were taken from freshly blood-drained animals, fixed in 4% paraformaldehyde (PFA) overnight, and impregnated in a 5% solution of potassium dichromate at room temperature for 3 days with gentle rotating. The brains were then stained in 2% silver nitrite for 3 to 5 days with daily solution change. The stained brains were embedded in a 3% low-melting-temperature agarose gel (Sigma) and cut into 60-μm coronal sections on a vibratome. Sections were rinsed in distilled water and mounted onto glass slides with Paramount.

Loss of Wnt7a expression affected neural progenitor cell cycle progression.

Since Wnt7a is expressed in the active adult neurogenic areas (14, 29), we tested whether loss of Wnt7a expression affects adult neural stem/progenitor cell proliferation and self-renewal. We isolated neural stem/progenitor cells from forebrains of 6-week-old WT and Wnt7a^−/− mice and cultured these cells in DMEM–F-12 medium supplemented with N2, epithelial growth factor (EGF), fibroblast growth factor 2 (FGF2), and heparin. Neural stem/progenitor cells were treated with BrdU for 1 h, and then cell proliferation was monitored by BrdU labeling. The number of BrdU-positive cells was dramatically reduced in Wnt7a^−/− neural stem cells compared to WT cells (Fig. 1a and b). Minimal levels of active caspase-3, an indicator of apoptosis, were detected in both WT and Wnt7a^−/− cells (Fig. 1c and d), suggesting that the decreased BrdU labeling in Wnt7a^−/− cells was not due to increased apoptosis.

The self-renewal capacity of WT and Wnt7a^−/− neural stem cells was determined by clonal analysis. Primary clonal populations were derived from single WT or Wnt7a^−/− neural stem cells. The primary neurospheres were dissociated into single cells and plated at 50 cells per well in 96-well plates to allow secondary sphere formation. The secondary sphere size and formation rate served as an index of self-renewal ability. Loss of Wnt7a expression led to a dramatically reduced self-renewal ability in neural stem cells, as revealed by decreases in both neurosphere size and neurosphere formation rate in Wnt7a^−/− neural stem cells, compared to WT neural stem cells (Fig. 1e to g). These results indicate that Wnt7a plays an important role in neural stem/progenitor cell proliferation and self-renewal.

In adult mammalian brains, both the GFAP-expressing radial glia-like cells and the Sox2-positive cells have been considered neural stem cells in the subgranular zone of the hippocampal dentate gyrus (30, 31). To determine if the neural stem cell population was affected by loss of Wnt7a expression, we immunostained hippocampal sections of 6-week-old WT and Wnt7a^−/− littermate mice for GFAP and Sox2. The number of both GFAP⁺ cells and Sox2⁺ cells was reduced in the subgranular zone of the hippocampal dentate gyrus of Wnt7a^−/− brains, compared to those in the corresponding region of the WT brains (Fig. 2). These results suggest that Wnt7a expression is important for maintaining neural stem cell population in the hippocampal dentate gyrus of adult mouse brains.

Next, we analyzed in vivo cell cycle progression of WT and Wnt7a^−/− neural precursors in the hippocampal dentate gyrus to compare their relative cell cycle lengths. Specifically, 6-week-old WT and Wnt7a^−/− mice were pulsed with BrdU for 30 min. Neural progenitor cells were identified by Ki67 immunoreactivity (32, 33), and cells in S phase were labeled by BrdU staining. The fraction of progenitor cells in S phase, a labeling index, was determined by calculating the percentage of BrdU-positive and Ki67-positive cells out of total Ki67-positive cells (BrdU⁺ Ki67⁺/Ki67⁺). Because the length of S phase remains relatively constant in mammalian cells, whereas the length of G₁ phase regulates proliferation (34), this labeling index provides an estimate of cell cycle length. A decreased labeling index suggests longer cell cycles (2). By examining hippocampal dentate gyrus sections from six pairs of brains, we showed that neural progenitors in the dentate gyrus of Wnt7a^−/− adult brains divided more slowly than those in the dentate gyrus of WT littermates, as indicated by the decreased labeling index in the Wnt7a^−/− cells (Fig. 3a and b).

In addition to cell cycle length, we also determined whether Wnt7a regulates cell cycle reentry and exit of neural progenitors in adult hippocampal dentate gyrus. Cell cycle reentry was determined by scoring the fraction of cells that reentered cell cycle after pulse-labeling with BrdU for 24 h before brain harvest. Cells that had left the cell cycle were stained as BrdU⁺ and Ki67⁻, whereas cells that reentered the cell cycle were stained as BrdU⁺ and Ki67⁺. The index of cell cycle reentry is defined as the percentage of BrdU⁺ Ki67⁺ cells in total BrdU⁺ cells (35). A significant decrease in the percentage of neural progenitors that reentered the cell cycle was observed in the hippocampal dentate gyrus of Wnt7a^−/− brains, compared to that in WT brains, indicating that loss of Wnt7a expression led to increased cell cycle exit (Fig. 3c and d). Together, these results suggest that Wnt7a is important for controlling cell cycle progression of neural progenitors in the adult hippocampus.

Loss of Wnt7a expression inhibited neuronal differentiation and maturation.

To test whether Wnt7a regulates neural differentiation, we treated neural stem/progenitor cells isolated from WT or Wnt7a^−/− adult brains with retinoic acid (RA) and fetal bovine serum (FBS) to induce differentiation. After 3 days of treatment, the differentiated cells were stained with Tuj1, a neuronal marker; GFAP, an astrocyte marker; and O4, an oligodendrocyte marker. Both WT and Wnt7a^−/− neural stem/progenitor cells were able to differentiate into all three lineages (Fig. 4a), suggesting that both are multipotent. Interestingly, there was a substantial decrease in the percentage of Tuj1⁺ neurons in Wnt7a^−/− cells, compared to WT cells (Fig. 4a and b). In contrast, the GFAP⁺ astrocytes were slightly increased in Wnt7a^−/− cells (Fig. 4c), and we saw no significant difference in the percentages of O4⁺ oligodendrocytes between the WT and Wnt7a^−/− cells (Fig. 4d). These results suggest that Wnt7a is important for neuronal differentiation.

To test whether Wnt7a regulates neurogenesis in vivo, we treated 6-week-old WT and Wnt7a^−/− mice with BrdU for 7 days and harvested the brains 3 weeks after BrdU treatment. Neurogenesis was monitored by BrdU and doublecortin (DCX) double staining. When quantified, the number of BrdU-positive and DCX-positive newborn neurons in the hippocampal dentate gyrus was substantially decreased in Wnt7a^−/− brains compared to WT brains (Fig. 5a and b). Moreover, the dendritic growth of Wnt7a^−/− neurons was dramatically impaired compared to that of WT neurons, as revealed by a considerable decrease in dendritic length (Fig. 5c and d). Sholl analysis showed substantially reduced dendritic complexity of the Wnt7a^−/− dentate granule neurons (Fig. 5e).

The role of Wnt7a in dendritic arborization in the hippocampal dentate gyrus was further investigated using Golgi staining, a method that has been widely used for studying neuronal morphology (36). Decreases in both dendritic length and complexity were observed in the dentate granule neurons in the hippocampus of adult Wnt7a^−/− mice, compared to those in the hippocampus of WT littermate mice (Fig. 5f, g, and h). Taken together, these results further support our hypothesis that Wnt7a plays an important role in neurogenesis by promoting both neuronal differentiation and maturation.

Wnt7a regulates neural stem/progenitor cell proliferation and neuronal differentiation by activating the expression of cyclin D1 and neurogenin 2 (Ngn2).

Wnt7a is expressed in both neural stem/progenitor cells and neurons, with a relatively higher expression in neural stem/progenitor cells (Fig. 6a). We transduced both neural stem/progenitor cells and neurons with a lentiviral vector expressing a small hairpin interference RNA (siRNA) for Wnt7a. Knockdown of Wnt7a expression was evident in both neural stem/progenitor cells and neurons (Fig. 6a). Next, we tested if knockdown of Wnt7a expression affects β-catenin downstream target gene expression in neural stem/progenitor cells and neurons. We observed dramatically reduced expression of cyclin D1, a downstream target gene of β-catenin (37), in the Wnt7a knockdown neural stem/progenitor cells (Fig. 6b). The expression of cyclin D1 was relatively low in neurons, and knockdown of Wnt7a exhibited no obvious effect on the neuronal expression of cyclin D1 (Fig. 6b). In contrast, we saw a dramatic reduction in the expression of Ngn2, a proneural basic helix-loop-helix transcription factor that has been shown to promote neuronal differentiation (38–40), in Wnt7a knockdown neurons (Fig. 6c). However, there was no detectable effect on Ngn2 expression in Wnt7a knockdown neural stem/progenitor cells (Fig. 6c).

We used chromatin immunoprecipitation (ChIP) analysis to show that the Wnt downstream effector β-catenin was bound to the lymphoid enhancer binding factor (LEF)/T cell transcription factor (TCF) binding site on the cyclin D1 promoter in neural stem/progenitor cells, and this site was associated with the active chromatin markers (Fig. 6d) acetylated histone H3 (AcH3) and trimethylated histone H3 lysine 4 (H3K4me3) (41, 42). We did not detect any β-catenin at the LEF/TCF binding site in the Ngn2 promoter in neural stem/progenitor cells (Fig. 6e). However, when neural stem/progenitor cells were induced to differentiate into neurons, β-catenin could no longer be detected at the cyclin D1 promoter (Fig. 6f). Instead, it was bound to the LEF/TCF binding site in the Ngn2 promoter, and this site was associated with the active chromatin markers AcH3 and H3K4me3 (Fig. 6g). The binding of β-catenin to the Ngn2 promoter, in combination with the previous observation that β-catenin activated a luciferase reporter gene under the control of the Ngn2 promoter (43), indicates that Ngn2 is a direct downstream target gene of β-catenin. Together, these results suggest that Wnt7a regulates neural stem/progenitor cell proliferation and neuronal differentiation by preferentially activating downstream target genes of β-catenin that are involved in cell cycle control under proliferating conditions and those involved in neuronal differentiation upon differentiation.

To further determine if Wnt7a regulates neural stem/progenitor cell proliferation by activating β-catenin and its downstream target genes, we transduced neural stem/progenitor cells with lentivirus expressing Wnt7a siRNA to knock down Wnt7a expression. Consistent with our observation that cell proliferation in neural stem/progenitor cells isolated from adult Wnt7a^−/− mouse brains was greatly reduced (Fig. 1a and b), a dramatic decrease in BrdU labeling was observed in Wnt7a knockdown neural stem/progenitor cells (Fig. 7a and b). Cotransduction of the Wnt7a siRNA with a constitutively active β-catenin that lacks its amino-terminal 90 amino acids containing the GSK3β phosphorylation site to mimic active Wnt signaling (14) substantially restored the cell proliferation that was inhibited by the knockdown of Wnt7a (Fig. 7a and b). When cyclin D1 was cotransduced with Wnt7a siRNA, there was a similar recovery of cell proliferation (Fig. 7a and b). Together, these results indicate that Wnt7a stimulates neural stem/progenitor cell proliferation by activating the β-catenin–cyclin D1 pathway.

To determine if Wnt7a regulates neuronal differentiation by activating β-catenin, we transduced neural stem/progenitor cells with Wnt7a siRNA alone, or together with the constitutively active β-catenin. The transduced cells were then induced to differentiate into neurons. Consistent with the observed decrease in neuronal differentiation of neural stem/progenitor cells isolated from the Wnt7a^−/− adult mouse brains (Fig. 4), the number of Tuj1⁺ neurons induced from Wnt7a knockdown neural stem/progenitor cells was dramatically reduced (Fig. 7c and d). Remarkably, cotransducing Wnt7a siRNA with the constitutively active β-catenin restored neuronal differentiation considerably. Similarly, when we cotransduced Wnt7a siRNA with Ngn2, a downstream target gene of β-catenin that is involved in neuronal differentiation, neuronal differentiation was substantially restored as revealed by the increased number of Tuj1⁺ cells (Fig. 7c and d). In addition to rescuing the reduced levels of neuronal differentiation induced by Wnt7a knockdown, both the constitutively active β-catenin and Ngn2 also restored the neuronal maturation defects induced by Wnt7a siRNA. Both neurite length and neuronal complexity recovered substantially in cells treated with the Wnt7a siRNA in combination with either the active β-catenin or Ngn2, compared to those in cells treated with Wnt7a siRNA alone (Fig. 7e, f, and g). These results together indicate that Wnt7a regulates neural stem/progenitor cell proliferation and differentiation by activating β-catenin and specific downstream target genes involved in cell cycle control and neuronal differentiation.