Author contributions: Y.H. and M. Sawada: conception and design, collection and assembly of data, data analysis and interpretation, and manuscript writing; Y.S.K.: conception and design, collection and assembly of data, data analysis and interpretation, and provision of study material or patients; S.H.: collection and assembly of data and data analysis and interpretation; M.Sakaguchi and O.Y.: conception and design and collection and assembly of data; T.O.: conception and design, provision of study material or patients, financial support, and administrative support; H.O.: conception and design, data analysis and interpretation, financial support, and administrative support; K.S.: conception and design, data analysis and interpretation, financial support, manuscript writing, and administrative support.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS May 24, 2012.
Neuronal precursor cells generated by stem cells in the subventricular zone (SVZ) migrate and differentiate into mature interneurons in the olfactory bulb (OB). The mechanisms responsible for the dynamic morphological changes in cells during this process are largely unknown. Wnt/planar cell polarity (PCP) signaling regulates various developmental events, including neuronal migration and neurite formation. Here, we studied the function of two components of the PCP pathway, Dishevelled2 and Van Gogh like-2, in the newborn neurons in the postnatal mouse OB. Electroporation- or lentivirus-mediated introduction of vectors carrying a knockdown or dominant-negative construct of these genes into the SVZ altered the distribution and dendrite formation of newborn neurons in the OB, suggesting that PCP signaling is involved in regulating the maturation of new neurons in the OB. STEM CELLS 2012;30:1726-1733
Neurogenesis is continuous throughout the life span. In the adult rodent brain, the largest germinal region for new neurons is the subventricular zone (SVZ) of the lateral ventricle . Within the SVZ, astrocytes (also referred to as type B cells) function as neural stem cells (NSCs) and give rise to rapidly dividing transit-amplifying progenitor cells (type C cells) . The type C cells generate immature neuronal precursor cells (NPCs) (type A cells or neuroblasts) that migrate to the olfactory bulb (OB) through the rostral migratory stream (RMS). Upon reaching the OB, NPCs migrate radially and differentiate into local interneurons, that is, granule cells (GCs) or periglomerular cells (PGCs), in the granular and periglomerular layers (PGLs), respectively [3, 4]. Neurogenesis in this region is regulated by complex molecular mechanisms mediated by both intrinsic and extrinsic factors; these factors control cellular events that include the self-renewal of NSCs, the migration of NPCs, and the differentiation, survival, and maturation of new neurons [5–9].
The signaling pathways activated by the secreted Wnt proteins control a number of important biological events in the developing and mature nervous system . Wnt-signaling components, including ligands (Wnt1, Wnt5a, and Wnt7a), receptors (Fz3, Fz7, Fz10 sFrp1, and sFrp2), and down-stream effector molecules (Dvl1, Dvl2, Celsr2, Celsr3, Tcf3, Vangl2, and Diversin), are broadly expressed in the postnatal forebrain including the SVZ, RMS, and OB [11–14]. There are two major Wnt-signaling pathways, the β-catenin-dependent (canonical) Wnt/β-catenin pathway, which regulates embryonic patterning, tumorigenesis, cell proliferation, and the self-renewal of stem cells, and the β-catenin-independent (noncanonical) planar cell polarity (PCP) pathway, which controls the cell movement and polarity of various tissues, mainly via c-Jun N-terminal kinase (JNK) and Rho GTPase, during the development of many animal species, including mouse, frog, and zebrafish [15, 16].
The proliferation and differentiation of NSCs and NPCs in both the developing and adult brain are regulated by the Wnt/β-catenin [17–22] and Wnt/PCP [11, 23] pathways. Furthermore, the migration of facial branchiomotor (FBM) neurons [24–31] and neural crest cells  depends on the function of the PCP genes, as does neuronal morphogenesis in the postnatal hippocampus [33, 34] and OB , indicating that PCP signaling is involved in the migration and differentiation of many types of neural cells [36–38]. However, the function of this signaling pathway in the postnatal migration and differentiation of neurons in the OB has not been clearly demonstrated. In this study, we investigated the in vivo effects of inhibiting two core components of the Wnt/PCP pathway, the Fz-binding intracellular protein Dishevelled2 (Dvl2) and the four-pass transmembrane protein Van Gogh like-2 (Vangl2), on the newborn OB neurons.
MATERIALS AND METHODS
Wild-type institute of cancer research (ICR) mice were purchased from SLC (Shizuoka, Japan, /jslc.co.jp/). All animal experimental procedures were in compliance with National Regulations and Guidelines, reviewed by the Institutional Laboratory Animal Care and Use Committee, and approved by the President of Nagoya City University.
Brains were perfusion-fixed with 4% paraformaldehyde, postfixed in the same fixative overnight, and cut into 50-μm sections on a Vibratome sectioning system (VT1200S; Leica, Heidelberg, Germany, www.leica-microsystems.com/). For immunostaining, after being rinsed three times in phosphate buffered saline (PBS), the sections were incubated for 1 hour in blocking solution (10% donkey serum and 0.5% Triton X-100 in PBS), then overnight with the primary antibodies, and then for 2 hours at room temperature with a biotinylated secondary antibody (1:500; Jackson, West Grove, PA, www. jacksonimmuno.com) or an Alexa Fluor-conjugated secondary antibody (1:300; Invitrogen, Carlsbad, CA, www.invitrogen.com). The biotinylated antibodies were visualized using streptavidin-Alexa 568 (1:500; Invitrogen). The following primary antibodies were used: rabbit anti-Dvl2 (1:300; Santa Cruz Biotechnology, Santa Cruz, CA, www.scbt.com/, sc-13974); rabbit anti-Vangl2 (1:300; MBL, Aichi, Japan, www.mbl.co.jp, a custom purified antibody against the peptide corresponding to residues 16-43 of Vangl2); chick anti-green fluorescent protein (GFP) (1:500; Aves Labs, Tigard, OR, www.aveslab.com, 2-P004-07D); rat anti-GFP (1:500; Nacalai, Kyoto, Japan, www.nacalai.co.jp/, GF090R); monoclonal anti-Mash1 (1:100; BD Pharmingen, San Diego, CA, www.bdbiosciences.com/, 556604); goat antidoublecortin (Dcx) (1:200; Santa Cruz Biotechnology; sc-8066); and mouse anti-NeuN (1:100; Millipore, Billerica, MA, www.millipore.com, MAB377).
Confocal images were obtained using an LSM5-Pascal or LSM710 laser-scanning microscope (Zeiss, Jena, Germany, www.zeiss.com). They were analyzed using NIH Fiji/ImageJ version 1.36 and Photoshop (Adobe Systems, San Jose, CA, www.adobe.com/). For quantitative analyses of NPC distribution in the OB, the GC layer (GCL) was divided into the inner and outer regions, based on the nuclear staining pattern. For quantitative analyses of neuronal morphology, reconstructions were performed from confocal z-series stacks (20–35 μm). The total dendrite length, primary dendrite length, and number of dendritic branch points were analyzed using the “simple neurite tracer” plugin of Fiji/ImageJ software.
The mouse Dvl2ΔPDZ construct was described previously . The C terminus of Dvl2 (Dvl2-C), spanning amino acids 508–736, was amplified by PCR. Plasmids encoding GFP fusion proteins (Dvl2ΔPDZ-GFP and Dvl2-C-GFP) were described previously . pEGFP-N1 (Clontech, Mountain View, CA, www.clontech.com/) was used as the control vector. Self-inactivating lentivirus vectors (CSII-EF-MCS), the Rev-expressing plasmid pCMV-VSV-G-RSV-Rev, and the packaging construct pCAG-HIVgp were provided by Dr. H. Miyoshi, Riken Tsukuba BioResource Center . Dvl2ΔPDZ-GFP and Dvl2-C-GFP were subcloned into CSII-EF-MCS to produce CSII-EF-Dvl2ΔPDZ-GFP and CSII-EF-Dvl2ΔPDZ-GFP, respectively. Vangl2 and Vangl2ΔC were subcloned into CSII-EF-MCS-IRES-GFP to produce CSII-EF-Vangl2-IRES-GFP and CSII-EF-Vangl2ΔC-IRES-GFP, respectively. For the knockdown (KD) of mouse Vangl2 using a miRNA system (Invitrogen), the targeting sequence of Vangl2 (antisense target sequence: AAGACAAGCACCATGAGCAGA, miR RNAi Select Mmi581648, Invitrogen) and lacZ (control) were cloned into a modified Block-iT Pol II miR RNAi expression vector (Invitrogen). DNA cassettes containing GFP or DsRed-Express and the miRNA sequence were cloned into pCAGGS vectors  using the Gateway system (Invitrogen).
HEK293T cells were cotransfected with pCAGGS-HA-Vangl2-GFP and the shRNA expression vector against Vangl2 (pCAGGS-DsRed-miVangl2) or lacZ (pCAGGS-DsRed-milacZ, as a control) using the Fugene transfection reagent (Roche Diagnostics, Basel, Switzerland, www.roche-applied-science.com). Two days later, cell lysates were separated by SDS-PAGE and then immunoblotted with anti-Flag (1:500; Sigma, St. Louis, MO, www.sigmaaldrich.com F7425) and rabbit anti-β-actin (1:2,000; Abcam, Cambridge, U.K., www.abcam.com, ab8227) antibodies.
Preparation and Injection of Lentivirus
Lentiviruses were produced by the transient transfection of 293T cells with the lentiviral constructs, pCMV-VSV-G-RSV-Rev and pCAG-HIVgp, using the GeneJuice transfection reagent (Novagen, Madison, WI, www.merckgroup.com). High-titer, concentrated stocks prepared by ultracentrifugation and resuspension in PBS were used to achieve efficient infection. For the injection of lentivirus, mice were anesthetized by hypothermia and placed on the platform of a stereotaxic injection apparatus (David Kopf Instruments, Tujunga, CA, www.kopfinstruments.com) using vinyl tape. For the injection of lentivirus, 1 μl of the lentivirus-containing solution was stereotaxically injected into the lateral ventricle of P0-1 mice ([relative to lambda]: anterior, lateral, and depth [in mm]: 1.8, 1.0, and 1.8).
Electroporations of P0-1 mice were performed as described previously  with some modifications. Briefly, P0-1 mice were anesthetized by hypothermia and fixed to the platform of a stereotaxic injection apparatus (David Kopf Instruments) by vinyl tape. Approximately, 2 μl of the plasmid solution (3 mg/ml) containing 0.01% fast green solution was injected into the lateral ventricle of the P0-1 mice, and electronic pulses (90 V, 50 ms, four times) were applied using an electroporator (CUY-21SC; Nepagene, Chiba, Japan, www.nepagene.jp) with a forceps-type electrode (CUY650P7). For the experiments involving sequential electroporation, plasmids were introduced into the same ventricle at P0 and again, after a 24-hour interval, at P1. Organotypic brain slices were prepared from P6-7 ICR mice after the sequential electroporations, essentially as reported previously .
SVZ Monolayer Culture and Morphological Analysis
The SVZ cells from neonatal mice (postnatal day 1) were dissociated with trypsin-EDTA and transfected with plasmids by electroporation using an Amaxa Nucleofector device (Lonza, Switzerland, www.lonza.com). The cells were then cultured in neurobasal medium containing 10% fetal bovine serum (FBS), 50 U/ml penicillin-streptomycin, 2 mM L-glutamine, and B27 supplement on coverglasses coated with poly(L-lysine) and laminin. Six days after seeding, the cells were fixed with 4% paraformaldehyde. To detect GFP+ neurons, double immunostaining using anti-GFP and anti-Dcx antibodies (see Histological Analysis) was performed. To analyze the morphology of cultured neurons quantitatively, Neurolucida software (MBF Biosciences, Williston, VT, www.mbfbioscience.com) was used. The Dcx-expressing GFP+ neurons were traced and analyzed for the number of branch points, number of tips, and total dendritic length.
All data were expressed as the mean ± SEM. Differences between means were determined by unpaired two-tailed Student's t tests. A p-value of <.05 was considered significant.
Blocking PCP Signaling Alters the Distribution of NPCs in the OB
First, we examined the effects of dominant-negative forms of Dvl2 on the migration and differentiation of newborn OB neurons in postnatal mice. We used two deletion constructs of Dvl2, both of which were previously reported to act as specific dominant-negative inhibitors of the PCP signal, but not of the canonical Wnt signal: Dvl2ΔPDZ, which lacks the PDZ domain , and Dvl2-C, which contains only the conserved C-terminus of Dvl2  (Supporting Information Fig. S2). We did not perform a KD of Dvl2, because doing so would inhibit the Wnt/β-catenin pathway  as well as the PCP pathway. Plasmids encoding a Dvl2-C-GFP or Dvl2ΔPDZ-GFP fusion protein were introduced into the lateral ventricle of P0-1 mice by electroporation. One day after the electroporation, most of the GFP-positive (GFP+) cells were observed at the walls of the lateral ventricles (data not shown). At P7, we quantified the number of GFP+ cells in the OB. The percentage of GFP+ cells observed in the OB, including the most anterior part of the RMS, was significantly decreased by the expression of the dominant-negative Dvl2 constructs (control, 71.2% ± 2.0%, n = 6,714 cells from three mice; Dvl2-C, 57.5% ± 2.2%, n = 294 cells from three mice, p = .0179 vs. control; Dvl2ΔPDZ, 36.9% ± 3.4% n = 254 cells from three mice; p = .0015 vs. control) (Fig. 1A–1C), suggesting that the inhibition of Dvl2 disturbed the migration of NPCs into the OB. Consistent with this, the KD of Vangl2 (Supporting Information Fig. S3), another core PCP component, led to a slight decrease in NPCs' migratory speed and altered their directionality in cultured brain slices (Supporting Information Movie 1 and Supporting Information Fig. S4). In contrast, the dominant-negative forms of Dvl2 did not affect the percentage of GFP+ cells that were also positive for Mash1 (a marker of type C cells) [47, 48] or Dcx (a marker of NPCs)  or undergoing mitosis in the SVZ (Supporting Information Fig. S5), suggesting that the decreased percentage of GFP+ cells in the OB induced by the dominant-negative forms of Dvl2 (Fig. 1) was not due to the altered differentiation or proliferation of SVZ cells. Taken together, these results suggest that blocking PCP signaling disturbs the migration of NPCs into the OB.
Second, to study the final distribution of newborn OB neurons, a lentivirus expression vector encoding GFP with or without a dominant-negative construct was injected into the lateral ventricle of neonatal mice. As a dominant-negative form of Vangl2, we used Vangl2ΔC (Supporting Information Fig. S2), which lacks the C-terminal region . Six to eight weeks later, numerous GFP+ cells were observed in the OB in all the experimental groups (data not shown), indicating that, although blocking the PCP signaling could transiently disturb the directional migration of NPCs in the RMS (Fig. 1), the NPCs eventually reached the OB. The differentiation of GFP+ cells in the OB was not disturbed by the blocking of PCP signaling (Supporting Information Fig. S6). Notably, a significantly smaller percentage of GFP+ cells expressing dominant-negative constructs was observed in the outer half of the GCL compared with that in the control group (control, 40.6% ± 2.8%, n = 795 cells from three mice; Dvl2-C, 22.9% ± 3.1%, n = 347 cells from three mice, p = .0425 vs. control; Dvl2ΔPDZ, 29.4% ± 3.1%, n = 327 cells from three mice, p = .0429 vs. control; Vangl2ΔC 32.3% ± 1.6%, n = 1,561 cells from three mice, p = .0398 vs. control; Vangl2, 47.9% ± 3.1%, n = 1,543 cells from three mice; p = .1601 vs. control) (Fig. 2A–2C), but the total percentage of GFP+ cells in the GCL and PGL was not altered. Full-length Vangl2 did not affect the distribution of infected cells in the OB (Vangl2, 47.9% ± 3.1%, n = 1,543 cells from three mice; p = .1601 vs. control) (Fig. 2C). These results suggest that long-term blocking of the PCP pathway decreases the distribution of newly generated neurons to the outer GCL of the OB.
Dvl2 and Vangl2 Act in a Cell-Autonomous Manner During the Dendritogenesis of OB Neurons
Next, we studied the morphology of OB neurons in which PCP signaling was continuously blocked in vitro. To examine whether PCP signaling acts in a cell-autonomous manner during the dendritogenesis of OB neurons, we introduced a dominant-negative Dvl2 (Dvl2-C) or Vangl2 KD (Vangl2 KD) vector into dissociated SVZ cells and cultured them for 6 days (Fig. 3). Compared with control cells expressing GFP (Fig. 3A, 3F), the neurons expressing Dvl2-C or Vangl2 KD showed impaired dendritogenesis (Fig. 3B, 3G, respectively). Quantitative analysis revealed that the expression of Dvl2-C or Vangl2 KD significantly decreased the number of branch points (control [pEGFP-N1], 1.40 ± 0.18, n = 20 cells; Dvl2-C, 0.25 ± 0.11, n = 16 cells, p = .000016; control [lacZ KD], 1.44 ± 0.35, n = 18 cells; Vangl2 KD, 0.64 ± 0.19, n = 25 cells, p = .038) (Fig. 3C, 3H), the number of tips (control [pEGFP-N1], 3.50 ± 0.21, n = 20 cells; Dvl2-C, 1.93 ± 0.19, n = 16 cells, p = .0000063; control [lacZ KD], 3.83 ± 0.43, n = 18 cells; Vangl2 KD, 2.84 ± 0.24, n = 25 cells, p = .038) (Fig. 3D, 3I), and the total dendrite length (control [pEGFP-N1], 137.4 ± 10.9 μm, n = 20 cells; Dvl2-C, 60.8 ± 9.3 μm, n = 16 cells, p = .0000096; control [lacZ KD], 137.1 ± 12.5 μm, n = 18 cells; Vangl2 KD, 102.2 ± 9.3 μm, n = 25 cells, p = .027) (Fig. 3E, 3J), suggesting that Dvl2 and Vangl2 act in a cell-autonomous manner during dendritogenesis.
Blocking PCP Signaling Alters the Morphology of Newborn OB Neurons In Vivo
Finally, we studied the effect of blocking the PCP signaling on the morphology of OB neurons in vivo. First, we traced and compared the dendritic pattern of GCs expressing GFP and a dominant-negative form of Dvl2 with that of control GCs in vivo (Fig. 4A–4D). Most of the GFP+ GCs had no axon and one long primary dendrite that formed a dendritic tree, regardless of whether the PCP signaling was blocked. However, quantitative analyses revealed that the expression of dominant-negative Dvl2 decreased the total length of dendrites (control, 608.5 ± 44.0 μm, n = 10 cells from three mice; Dvl2-C, 429.3 ± 44.0 μm, n = 6 cells from two mice; p = .0175), increased the length of the primary dendrite (control, 147.8 ± 24.3 μm, n = 10 cells from three mice; Dvl2-C, 236.9 ± 31.7 μm, n = 6 cells from two mice; p = .0420), and decreased the number of dendritic branch points (control, 4.00 ± 0.46, n = 10 cells from three mice; Dvl2-C, 2.14 ± 0.26, n = 6 cells from two mice; p = .0052) (Fig. 4E–4G), suggesting that dendritogenesis was impaired.
We also examined the effect of electroporating the Vangl2 KD vector in vivo, which labels the morphology of GCs and PGCs more clearly. Consistent with the effect of Dvl2-C, Vangl2 KD also caused a decreased total length of dendrites in GCs (control, 411.0 ± 28.6 μm, n = 55 cells from three mice; Vangl2 KD, 328.5 ± 30.5 μm, n = 52 cells from three mice; p = .0386) (Fig. 5A–5C). In addition, PGCs expressing Vangl2 KD had a significantly decreased total dendrite length (control, 108.00 ± 11.98 μm, n = 29 cells from three mice; Vangl2 KD, 66.59 ± 8.17 μm, n = 32 cells from three mice; p = .0036), a reduced number of dendritic branch points (control, 2.00 ± 0.31, n = 29 cells from three mice; Vangl2 KD, 0.56 ± 0.17, n = 32 cells from three mice; p < .001), and a reduced number of neurites extending from the soma (control, 2.52 ± 0.21, n = 29 cells from three mice; Vangl2 KD, 1.88 ± 0.17, n = 32 cells from three mice; p = .0013) (Fig. 5D–5H). Together, these results suggest that the location and dendritogenesis of newborn neurons in the OB are altered when the PCP signal is blocked.
In this work, we studied the function of two major core components of the Wnt/PCP pathway, Dvl2 and Vangl2, in postnatal neuronal migration and differentiation in the OB. The functions of Vangl2 during embryonic development have been extensively investigated, mainly using a mutant mouse line, Loop-tail (Lp) [51, 52]. Since Vangl2Lp/Lp mice die as embryos, little is known about its function in the postnatal brain. Although adult Dvl2 mutant mice are viable, functional redundancy with Dvl1 makes it difficult to detect the specific role of Dvl2 in postnatal neurogenesis . Furthermore, it is possible that these mutations impair the development of the SVZ, RMS, and/or OB, which could indirectly affect postnatal neurogenesis. To address these issues, we used dominant-negative forms and a KD construct of these genes.
Roles of the Wnt/PCP pathway in cell migration have been reported in several different contexts [24–32]. In this work, blocking the Wnt/PCP pathway decreased the number of NPCs that reached the OB, without affecting their differentiation or proliferation (Figs. 1 and 2, Supporting Information Figs. S4–S6), suggesting that this pathway's function in NPC migration into the OB is cell-autonomous. On the other hand, NPC migration in the SVZ-RMS-OB path is known to be regulated by the surrounding cells, including ependymal cells, astrocytes, and blood vessels [54–57]. Since Dvl2 and Vangl2 are expressed broadly in this region (Supporting Information Fig. S1), it is also possible that these cells regulate NPC migration in the SVZ-RMS-OB path via PCP signaling in a non-cell-autonomous manner, as shown for the FBM neuronal migration in mice and zebrafish [25, 27, 29-31].
Previous studies identified several extracellular factors involved in the dendritogenesis of SVZ-derived OB interneurons, including GABA  and vascular endothelial growth factor (VEGF) . Wnt5a, which likely activates the PCP signaling pathway, increases the neurite complexity of OB interneurons derived from the SVZ . However, the cell-intrinsic mechanisms involved in dendrite formation in newborn NPCs in the postnatal OB remain largely elusive. We showed that blocking the PCP signal in the newborn neurons using nucleofection and lentiviral vectors impaired their dendritic growth and arborization in vitro (Fig. 3) and in the OB (Figs. 4 and 5), suggesting that PCP signaling's role in dendritogenesis is cell-autonomous. In addition, since impaired PCP signaling decreases NPC migration into the outer GCL (Fig. 2), it is possible that the altered morphology of the GCs, such as the longer primary dendrite, is partly due to their altered migration and localization . Roles for Wnt signaling in neurite formation have been reported in invertebrates and vertebrates. In cultured hippocampal neurons, Wnt7 and Dvl activate Rac and JNK, promoting dendrite development . Wnt3a stimulates the neurite outgrowth of Ewing tumor cells via JNK . Celsr2 and Celsr3 regulate dendrite branching patterns . Given that Wnt5a is expressed in the olfactory sensory axons, PGCs, and granule cells in the postnatal OB [35, 62], and given that it functions in the noncanonical pathway to increase neurite complexity , it is possible that Dvl2 and Vangl2 function downstream of Wnt5a in this system. Interestingly, the Wnt/β-catenin signal suppresses dendrite maturation in SVZ-derived neurons , suggesting that both the Wnt/β-catenin and Wnt/PCP pathways cooperatively regulate the dendritogenesis of NPCs in the adult OB.
In conclusion, this study suggests possible roles for PCP signaling in the migration and morphogenesis of new neurons in the OB and provides new insights into the mechanisms underlying postnatal neurogenesis.
We thank R. Habas, H. Miyoshi, and H. Naka for providing materials and members of the Sawamoto laboratory for useful discussions. M. Sawada is a Research Fellow of the Japan Society for the Promotion of Science. This work was supported by the Funding Program for Next Generation World-Leading Researchers (to K.S.), a Grant-in-Aid for Young Scientists (S) (to K.S.), the International Human Frontier Science Program Organization (to K.S.), and the Japan Brain Foundation (to Y.H.).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.