Wnt Signaling Regulates Symmetry of Division of Neural Stem Cells in the Adult Brain and in Response to Injury§

Authors

  • David Piccin,

    1. Department of Surgery, Institute of Medical Science, Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author
  • Cindi M. Morshead

    Corresponding author
    1. Department of Surgery, Institute of Medical Science, Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada
    • Department of Surgery, Institute of Medical Science, Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada MS5 3E1
    Search for more papers by this author
    • Telephone: 416-946-5575; Fax: 416-978-8287


  • Disclosure of potential conflicts of interest is found at the end of this article.

  • Author contributions: D.P.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; C.M.M.: conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.

  • §

    First published online in STEM CELLSEXPRESS December 29, 2010.

Abstract

Neural stem cells comprise a small population of subependymal cells in the adult brain that divide asymmetrically under baseline conditions to maintain the stem cell pool and divide symmetrically in response to injury to increase their numbers. Using in vivo and in vitro models, we demonstrate that Wnt signaling plays a role in regulating the symmetric divisions of adult neural stem cells with no change in the proliferation kinetics of the progenitor population. Using BAT-gal transgenic reporter mice to identify cells with active Wnt signaling, we demonstrate that Wnt signaling is absent in stem cells in conditions where they are dividing asymmetrically and that it is upregulated when stem cells are dividing symmetrically, such as (a) during subependymal regeneration in vivo, (b) in response to stroke, and (c) during colony formation in vitro. Moreover, we demonstrate that blocking Wnt signaling in conditions where neural stem cells are dividing symmetrically inhibits neural stem cell expansion both in vivo and in vitro. Together, these findings reveal that the mechanism by which Wnt signaling modulates the size of the stem cell pool is by regulating the symmetry of stem cell division. STEM CELLS 2011;528–538

INTRODUCTION

One of the fundamental questions in stem cell biology is how cell fate decisions are regulated. To date, little is known about the mechanisms that regulate stem cell divisions. Stem cells can divide symmetrically to generate two stem cells, thereby expanding the stem cell pool, or asymmetrically to give rise to a stem cell and a more lineage restricted progenitor cell, thereby permitting maintenance of the stem cell pool and the generation of differentiated cell types [1]. The canonical Wnt signaling pathway (Wnt/β-catenin pathway) has been linked to stem cell expansion in several systems, including the hematopoietic system [2] and the embryonic nervous system [3]. The mechanism that underlies the increase in stem cell numbers has been difficult to ascertain as a change in the symmetry of division, cell cycle time, or cell survival, could all account for the expansion.

The adult brain contains a small pool of multipotential, self-renewing neural stem cells that reside in a specialized niche called the subependymal zone (SEZ). Under baseline conditions, adult neural stem cells divide asymmetrically and give rise to progenitor cells that proliferate and migrate to the olfactory bulb, where they differentiate into interneurons [4, 5]. As with most mammalian stem cell systems, no direct measure of symmetric division or link between orientation of cleave plane and symmetry of division has been established, in part due to the fact that no specific stem cell markers are known. The adult brain is an ideal model to examine the hypothesis that Wnt signaling plays a role in symmetry of division because the resident stem cells divide exclusively asymmetrically under baseline conditions, and therefore, an increase in stem cell numbers in vivo must necessarily indicate a change from asymmetric to symmetric divisions. Moreover, there are established in vivo manipulations of the stem cell niche, such as infusion of a mitotic inhibitor to eliminate the rapidly dividing stem cell progeny, that have been shown to promote symmetrical stem cell division leading to SEZ regeneration [6].

We have demonstrated that under baseline conditions in vivo (when stem cells are dividing asymmetrically), active Wnt signaling is rarely observed within the SEZ stem cells. However, when stem cells are dividing symmetrically, that is, during SEZ regeneration, following stroke, or during neurosphere formation, the proportion of stem cells with Wnt signaling is significantly increased. Moreover, blocking this Wnt signaling significantly reduces the number of stem cells. We have shown that increased Wnt signaling both in vitro and in vivo increased the numbers of neural stem cell-derived colonies (neurospheres) with no concomitant change in the proliferation kinetics of the progenitor population. Taken together, our findings reveal that Wnt signaling regulates the mode of division of stem cells and that activation of Wnt signaling is sufficient to promote symmetric division and expansion of the stem cell pool.

MATERIALS AND METHODS

Animals

BAT-gal mice [7], C57Bl/6, and CD1 mice (8–12 weeks old; Charles River, Wilmington, MA, http://www.criver.com) were housed at the University of Toronto animal facilities in accordance with the Animal Care Committee and institutional guidelines.

Infusions

Animals were anesthetized with isofluorane (1%–5%). SB216763 (20 μM, 4 days), secreted frizzled-related protein 2 (sFRP2; 4 μg/ml, 4 days), or cytosine-β-D-arabinofuranoside (Ara-C; 2%, 6 days) was intraventricularly infused unilaterally using an osmotic minipump (1007D, Alzet, Cupertino, CA, http://www.alzet.com) implanted onto the surface of the brain at A/P 0 mm, L 0.7 mm relative to bregma, and 2.5 mm below the surface of the skull. Vehicle control (dimethyl sulfoxide [DMSO] for SB216763 and sterile physiological saline for sFRP2 and Ara-C) infusions served as controls. Only the ipsilateral hemisphere was used for subsequent analysis. Some animals received i.p. injections of bromodeoxyuridine (BrdU; 65 mg/kg in saline; Sigma-Aldrich Canada, Oakville, Canada, http://www.sigmaaldrich.com) as described in the text.

Isolation and Culture of SEZ Cells

Mice were killed by cervical dislocation, and SEZ cells were isolated as described previously [8, 9]. Briefly, the lateral ventricle subependyma was dissected and placed in artificial cerebrospinal fluid. Tissue was digested with 1.33 mg/ml trypsin (Sigma), 0.67 mg/ml hyaluronidase (Sigma), and 0.2 mg/ml kynurenic acid (Sigma) at 37°C, then transferred to serum-free medium (SFM) containing 0.7 mg/ml trypsin inhibitor (Roche Applied Sciences, Laval, Canada, www.roche-applied-science.com), and triturated with a fire-polished Pasteur pipette. Live cells were counted by trypan blue exclusion (Riechert Brigth Line Hemacytometer) and were cultured at 10 cells per milliliter (unless otherwise stated) in 24-well (0.5 milliliter per well) uncoated plates (Nunc, Rochester, NY, http://www.nuncbrand.com) in SFM [10] containing epidermal growth factor (EGF; 20 ng/ml; Upstate - Millipore, Billerica, MA, http://www.millipore.com), basic fibroblast growth factor (10 ng/ml; Sigma), and heparin (2 mg/ml; Upstate). For single-neurosphere passaging, each neurosphere was transferred to a 500-ml Eppendorf tube containing 200 ml plating medium, and then triturated and plated in a final volume of 500 ml medium in a 24-well plate. For differentiation, individual neurospheres were lightly titrated and equally divided into three separate Matrigel-coated wells (24-well plate) in SFM containing 5% fetal bovine serum. BrdU (0.6 μM; Sigma) was added to the media wherever indicated in the text.

Flow Cytometry

Fluorescence-activated cell sorting (FACS) was conducted using a FACSaria (BD Biosciences, Mississauga, Canada, http://www.bdbiosciences.com). lacZ-expressing cells were visualized using the FluoReporter lacZ Flow Cytometry Kit (Invitrogen, Carlsbad, California, http://www.invitrogen.com). Gates were set using C57Bl/6 neurosphere-derived cells that were exposed to the identical fluorescein-di-β-D-galactopyranoside (FDG) substrate (negative control). C57BL/6 and BAT-gal cells were sorted directly into wells containing neurosphere plating media (as described above) at 200 cells per well.

Immunohistochemistry

Animals were overdosed with ketamine-xylene mixture and perfused transcardially with 2% paraformaldehyde (Sigma), and then the brains were cryoprotected in 20% sucrose overnight before sectioning (14-mm-thick coronal sections). For X-gal staining, the sections were washed and incubated in X-gal (Novagen, EMB Chemicals, Darmstadt, Germany, http://www.emdchemicals.com) solution for 1 or 6 hours as described previously [11]. For lacZ and glial fibrillary acid protein (GFAP) staining, sections were postfixed with 4% paraformaldehyde, then sequentially stained for lacZ (β-galactosidase) with rabbit anti-β-gal (1:3,000; ab616, Abcam, Cambridge, MA, http://www.abcam.com) and for GFAP with rabbit anti-GFAP (1:400; G9269, Sigma) with Alexa Fluor (1:400; Invitrogen) secondary antibodies. For BrdU, sections were incubated in 1 N HCl at 60°C for 30 minutes before incubation in rat anti-BrdU antibody (1:100; Seralabs, Haywards Heath, UK, http://www.seralab.co.uk) at 4°C, followed either by tetramethyl rhodamine iso-thiocyanate (TRITC) donkey anti-rat antibody (1:200; Jackson Immunoresearch, West Grove, PA, http://www.jacksonimmuno.com) or by biotinylated anti-rat antibody (1:100; Vector Labs, Burlingame, CA, http://www.vectorlabs.com). Biotinylated antibodies were visualized using the Vectastain Elite ABC Kit and the DAB Substrate Kit (Vector Labs). The total numbers of BrdU-labeled cells were counted surrounding the lateral ventricles between the genu of the corpus callosum anteriorally and the crossing of the anterior commissure. Fluorescent sections were examined with the Zeiss Axiovert 200M or Zeiss LSM510. For cell culture, cells were fixed with 4% paraformaldehyde for 20 minutes at room temperature, and then stained with mouse anti-O4 monoclonal (1:200; MAB345, Millipore, Billerica, MA, http://www.millipore.com), rabbit anti-GFAP (1:400; G9269, Sigma), and mouse anti-βIII tubulin monoclonal (1:400; T8660, Sigma) primary antibodies, and anti-mouse, anti-rabbit and anti-mouse Alexa Fluor secondary antibodies, respectively (1:400; Invitrogen).

Stroke

Animals were anesthetized with isofluorane (1%–5%). A drill was used to cut a rectangular hole in the frontal and parietal bones running from 0.5 mm posterior to 2.5 anterior to the bregma and running from 0.5 to 3.5 mm lateral from the midline. The dura was removed, and the pia and the attached blood vessels were wiped from the cortical surface using a sterile saline-soaked cotton swab. At day 7 poststroke, the brains were perfused as above.

RESULTS

A Subpopulation of Cells Have Active Wnt Signaling Under Baseline Conditions in the Adult Subependyma

In the adult brain, stem cells comprise a rare subpopulation (∼0.5%) of SEZ cells [5, 12]. They are relatively quiescent with a cell cycle time of ∼15 days, and they proliferate to give rise to rapidly dividing progenitor cells (∼13-hour cell cycle), which comprise 15% of all the SEZ cells and account for the vast majority of the dividing cells in the SEZ [12, 13]. We used BAT-gal mice, which express lacZ under the control of a T-cell factor (TCF)-responsive element to identify cells with active Wnt signaling. A previous report using BAT-gal mice revealed Wnt signaling in a variety of regions throughout the developing brain, including the germinal zones, and in the hippocampus (a neurogenic region) of the adult brain [14]. We observed a small subpopulation of cells (2.7% ± 0.5%) within the adult SEZ with active Wnt signaling (X-gal+) under baseline conditions (Fig. 1A–1C); the majority of which (89.8% ± 1.5%) were not colabeled with the proliferation marker BrdU administered 12 hours prior to sacrifice (five injections; one in every 3 hours), indicating that the cells with active Wnt signaling were not rapidly dividing progenitors (Fig. 1D).

Figure 1.

Wnt signaling in the subependymal zone of adult BAT-gal mice. (A–C): X-gal staining of BAT-gal brain sections reveals that a small subpopulation of cells have active Wnt signaling (blue) under baseline conditions. Sections were exposed to X-gal for 6 hours, resulting in deeply stained cells. Scale bar = 100 μm (A) and 25 μm (B, C). (D): The majority (89%) of the cells with Wnt signaling are not colabeled with bromodeoxyuridine (BrdU; arrow, lacZ+/BrdU− cell; arrowhead, lacZ+/BrdU+ cell). Sections were exposed to X-gal for 1 hour, resulting in more faintly stained cells, and thus ensuring visualization of double-labeled cells. Scale bar = 25 μm. (E): Only 3.7% of long-term label-retaining cells have active Wnt signaling under baseline conditions (arrow, lacZ+/BrdU− cell; arrowhead, lacZ+/BrdU+ cell). Scale bar = 25 μm. (F): >99% of lacZ+ cells (red) did not coexpress GFAP (green; insert arrowhead, lacZ+/GFAP+). Scale bar = 25 μm. Abbreviations: CC, corpus callosum; GFAP, glial fibrillary acid protein; LV, lateral ventricle; St, striatum.

To determine whether the cells with active Wnt signaling comprised the neural stem cell fraction, animals received a series of BrdU injections as described above and were sacrificed after 30 days. The rapidly dividing progenitor cells would have diluted out the BrdU over the 30-day period, whereas the slowly dividing stem cell population would retain the BrdU label (long-term label-retaining cells). We found that only 3.7% ± 1.4% of the long-term label-retaining (BrdU+) stem cells coexpressed lacZ, revealing that they do not have active Wnt signaling (Fig. 1E). Additionally, using confocal imaging, we found that only two of >300 lacZ+ cells examined expressed GFAP, which is known to be expressed in neural stem cells in the adult SEZ (Fig. 1F) [9, 13]. Thus, we have found that the vast majority of the SEZ stem cells do not have active Wnt signaling.

Enhancing Wnt Signaling In Vivo Increases Stem Cell Number

In the adult brain, where the primary mode of division is asymmetric (maintenance mode), enhancing the survival or shortening the cell cycle time of the stem cells would not lead to increased numbers of stem cells. Accordingly, any observed increase in the numbers of stem cells in vivo in response to enhanced Wnt signaling would necessarily indicate a change from asymmetric to symmetric division. We enhanced Wnt signaling in vivo by infusing SB216763, a GSK3β inhibitor that activates the Wnt signaling pathway [15], intraventricularly into adult mice for 4 days. To confirm that the SB216763 infusion increased Wnt signaling in vivo, we examined the numbers of lacZ+ (Wnt signaling) cells and found a 1.8-fold increase in the SEZ of BAT-gal mice, although interestingly, many SEZ cells remain refractory to the treatment (Fig. 2A–2C). We then used the standard neurosphere assay [16–18] to determine the number of stem cells present in the SEZ after treatment. Stem cells in low cell density cultures proliferate to form clonally derived colonies of cells termed neurospheres, which are composed of a mixed population of stem cells (<1%) and progenitor cells (>99%) [19, 20]. We observed a significant 1.7-fold increase in the numbers of clonally derived neurospheres from the SB216763-treated brains relative to vehicle-infused controls (Fig. 2D–2F). The treatment had no effect on the number of live cells that survived the dissection (48.7 × 103 ± 14.3 × 103 vs. 51.2 × 103 ± 7.4 × 103 cells per hemisphere, n = 14 mice, three independent experiments, p = .544), indicating that the enhanced cell survival following the dissection and plating did not account for the increased numbers of neurospheres observed with enhanced Wnt signaling. We also found no change in the size of the neurospheres that formed (144.8 ± 4.7 μm vs. 139.1 ± 6.4 μm, control vs. SB216763 infusion, n > 50 neurospheres per group, p = .505), suggesting that the proliferative potential of colony-forming cells is not changed in response to the treatment. We rigorously examined the neurospheres that were formed from both the SB216763-treated and control brains to confirm that they elicited the stem cell characteristics of self-renewal and multipotency. When single neurospheres collected from both SB216763- and control-treated groups were differentiated in the presence of fetal bovine serum for 7 days, 100% of the neurospheres were multipotential, giving rise to neurons, astrocytes, and oligodendrocytes (Fig. 2J–2O). Serial passaging revealed that >90% of the single neurospheres could be passaged five times (the longest time examined). Hence, neural stem cells underwent symmetric divisions in response to enhanced Wnt signaling in vivo.

Figure 2.

Enhancing Wnt signaling in vivo. (A–I): Wnt signaling was enhanced in vivo following intraventricular infusion of 20 μM SB216763 for 4 days. (A–C): SB216763 infusion increased the number of cells with active Wnt signaling (X-gal+) in vivo (n = 3 mice per group; *, p = .050). (D–F): SB216763 infusion increased the number of neurosphere-forming cells within the subependymal zone (SEZ; n = 15 mice per group from three independent experiments; *, p = .041). Dimethyl sulfoxide (DMSO) (vehicle control) infusion served as control. There was no difference in the numbers of neurospheres from saline-infused and DMSO-infused control brains (39.4% ± 11.1% vs. 37.6% ± 6.6%, saline vs. DMSO, n = 15 mice per group from three independent experiments). (G–I): There was no change in overall in vivo proliferation rates (number of BrdU+ cells) at the end of the SB216763 infusion. Animals received two injections of BrdU (one in every 2 hours) and were sacrificed 2 hours after the last injection (n = 3 mice per group). (J–O): Individual neurospheres grown from the vehicle control (J–L) or SB216763 (M, N) infused SEZ were lightly triturated, and the cells were equally divided into three wells under differentiation conditions. After 7 days, the cells were stained for GFAP (astrocytes) (J, M), βIII-tubulin (neurons) (K, N), and O4 (oligodendrocytes) (L, O). Scale bars = 100 μm. All data are represented as mean ± SEM. Abbreviations: BrdU, bromodeoxyuridine; GFAP, glial fibrillary acid protein; LV, lateral ventricle; St, striatum.

When infused mice were exposed to BrdU during the 4 hours prior to sacrifice (two injections, 2 hours apart), we observed no change in the numbers of BrdU+ cells in vivo in the SEZ of SB216763-infused animals (Fig. 2G–2I), indicating that enhanced Wnt signaling has no effect on overall proliferation rate of the progenitor cells. Taken together, these data reveal an increase in stem cell number, with no concomitant change in overall proliferation.

We predicted that if enhanced Wnt signaling increased stem cell number in the SEZ by enhancing symmetry of division rather than by changing the stem cell proliferation rate or survival, then blocking Wnt signaling in vivo would have no effect on the numbers of stem cells, as they are already dividing asymmetrically. We reduced Wnt signaling by infusing sFRP2, which directly binds Wnt protein thereby antagonizing the endogenous Wnt signaling [21]. Intraventricular infusion of sFRP2 into adult BAT-gal mice for 4 days resulted in a significant 4.5-fold reduction in the number of SEZ cells with active Wnt signaling (lacZ+ cells) relative to saline-infused controls (Fig. 3A–3C). However, we observed no change in the numbers of neurospheres (Fig. 3D–3F) or BrdU incorporation in vivo (Fig. 3G–3I). Hence, in conditions under which stem cells are already dividing asymmetrically, blocking Wnt signaling does not affect stem cell numbers.

Figure 3.

Blocking Wnt signaling in vivo. (A–I): Wnt signaling was blocked in vivo following intraventricular infusion of 4 μg/ml sFRP2 for 4 days. (A–C): sFRP2 infusion decreased the number of cells with active Wnt signaling (X-gal+) in vivo (n = 3 mice per group; *, p = .001). (D–F): The sFRP2 infusion had no effect on the number of neurosphere-forming cells in the subependymal zone (n = 14 mice per group, three independent experiments; *, p = .921). (G–I): There was no change in overall proliferation rates (number of BrdU+ cells) at the end of the sFRP2 infusion. Animals received two injections of BrdU (one in every 2 hours) and were sacrificed 2 hours after the last injection (n = 3 mice per group). Scale bars = 100 μm. All data are represented as mean ± SEM. Abbreviations: BrdU, bromodeoxyuridine; LV, lateral ventricle; sFRP2, secreted frizzled-related protein 2; St, striatum.

Enhanced Wnt Signaling In Vitro Increases the Number of Symmetric Stem Cell Divisions

Next, we used in vitro assays to investigate the role of Wnt signaling in adult neural stem and progenitor cells. The power of the in vitro assay is that the founding neural stem cell of an individual colony (neurosphere) must necessarily undergo symmetric divisions during neurosphere formation as a single neurosphere gives rise to multiple secondary neurospheres on passaging [22]. We performed histochemistry on passaged neurospheres from BAT-gal mice. Control (untreated) neurospheres contained a small subpopulation of cells that were lacZ+ (Fig. 4B), and the number of lacZ+ cells was greatly increased in neurosphere grown in the presence of factors to enhance Wnt signaling (SB216763-treated or Wnt3a-expressing retrovirus [Wnt3a-RV]-infected; Fig. 4C, 4D). Similar to our in vivo observations, we found that many of the neurosphere cells were refractory to the treatment, resulting in the appearance of mottled neurospheres.

Figure 4.

Modulating Wnt signaling in vitro changes stem cell number. (A–E): Neurospheres derived from C57Bl/6 (A) or BAT-gal (B–E) subependymal zone (SEZ)-derived cells were stained with X-gal. Increased numbers of X-gal positive (Wnt signaling) cells are present in neurospheres derived from cultures containing 1 μM SB216763 (C) or Wnt3a-RV (D) relative to untreated controls (B). (E): No X-gal expression is observed in neurospheres derived from sFRP2 culture conditions. Scale bars = 25 μm. (F): Primary neurosphere-derived cells plated at clonal density in varying concentrations of SB216763 reveal no effect on the numbers of neurospheres which formed after 7 day in vitro (n = 3 independent experiments). (G): Neurospheres derived in SB216763 conditions (1 μM) that were individually passaged (i.e., single-neurosphere passaged) to control media formed more secondary neurospheres when compared with controls (untreated; n = 72 neurospheres, three independent experiments; *, p = .032). (H): Primary adult SEZ neurosphere-derived cells were transfected with Wnt3a-expressing retrovirus or control GFP-RV revealing no effect of enhanced Wnt signaling on neurosphere number (n = 3 independent experiments). (I): However, passaging of individual transfected neurosphere into control media revealed a significant increase in secondary neurosphere numbers (n = 50 neurospheres, three independent experiments; *, p = .006). The relative change in neurosphere number from the SB216763 and Wnt3a-RV is not significantly different. (J): Primary neurosphere-derived cells plated at clonal density in varying concentrations of sFRP2 reveal no effect on the numbers of neurospheres which formed after 7 day in vitro (n = 3 independent experiments). (K): Neurospheres derived in sFRP2 conditions (200 ng/ml) that were individually passaged (i.e., single-neurosphere passaged) to control media formed fewer secondary neurospheres when compared with controls (n = 39 neurospheres, three independent experiments; *, p = .025). All data are represented as mean ± SEM. Abbreviations: GFP-RV, green fluorescent protein-expressing retrovirus; sFRP2, secreted frizzled-related protein 2; Wnt3a-RV, Wnt3a-expressing retrovirus.

In the presence of enhanced Wnt signaling, we observed no effect on the number (Fig. 4F) or size (75.7 ± 4.04 μm vs. 75.5 ± 5.81 μm, control vs. 1 mM SB216763, n = 100 neurospheres per group, p = .982) of neurospheres grown in the presence of SB216763 over a range of concentrations. These observations imply that enhanced Wnt signaling has no effect on stem cell survival or their ability to form neurospheres (no change in number) or on progenitor cell proliferation (no change in size). To confirm that there was no change in proliferation in the presence of the Wnt enhancer, cells were exposed to BrdU for 4 hours on day 4 in culture and then fixed immediately; there was no change in the numbers of BrdU+ cells (17.5% ± 2.9% vs. 22.7% ± 1.9%, control vs. SB216763-treated, n = 250 cells from two mice per group, p = .624). Most interesting, however, was the finding that individual SB216763-treated neurospheres passaged to control media (i.e., in the absence of the drug) gave rise to 1.7-fold more secondary neurospheres than control-treated neurospheres (Fig. 4G), indicating that enhanced Wnt signaling increased the number of stem cells present within an individual neurosphere. We found identical results using a Wnt3a-RV to enhance Wnt signaling (Fig. 4H, 4I), supporting the hypothesis that Wnt signaling promotes symmetry of division. To rule out the possibility that the increase in secondary neurosphere number is the result of changes in cell survival within the developing neurosphere or specifically during dissociation, we bulk passaged the treated cultures at a fixed plating density of seven cells per microliter, and we observed a comparable 60.3% ± 12.3% increase in the number of secondary neurospheres from the 1 μM SB216763-treated primary neurosphere cultures (42.1 ± 8.9 vs. 69.8 ± 4.3 neurospheres per 3,500 cells, control vs. SB216763-treated, n = 3 independent experiments, p = .018). These findings indicate that Wnt signaling increases the ratio of stem cells to progenitor cells in the neurospheres, which is consistent with an increase in symmetric stem cell divisions rather than increased overall survival. Hence, we have found that enhancing Wnt signaling increases stem cell number without changing stem cell survival or overall proliferation, consistent with the hypothesis that Wnt signaling enhances symmetry of division.

Stem Cells Are Enriched in the Fraction with Active Wnt Signaling

During neurosphere formation, neural stem cells undergo asymmetric divisions to generate the progenitor cells that comprise the majority of the cells within the neurosphere and must also necessarily undergo symmetric divisions to expand the population and generate multiple new neurospheres upon passaging. Accordingly, we predicted that many stem cells would have active Wnt signaling (lacZ+) during neurosphere formation. Primary neurosphere were dissociated into single cells and exposed to FDG, a fluorescent β-galactosidase substrate, and FACS sorted into lacZ+ and lacZ− fractions. Consistent with the X-gal staining, a subset of the cells within the neurospheres had active Wnt signaling under baseline conditions (∼2%; Fig. 5A, 5B). The positive and negative fractions were collected and plated in the neurosphere assay to determine which contained the stem cells. We observed five times more neurospheres from the lacZ+ fraction (0.29 ± 0.21 vs. 1.4 ± 0.24 neurosphere per 200 cells, n = 15 wells over three independent sorts, p = .002; Fig. 5C), with no difference in the gross morphology of the neurospheres (size and degree of lacZ expression) isolated from each group (data not shown). Thus, in conditions where neural stem cells are dividing symmetrically in vitro (during neurosphere formation), the majority (85%) of the stem cells have active Wnt signaling.

Figure 5.

Flow cytometric sorting reveals that 85% of neurosphere-derived neural stem cells have active Wnt signaling. Cells isolated from control (C57Bl/6 derived) neurospheres (A) and BAT-gal-derived neurospheres (B) were exposed to fluorescein di-β-D-galactopyranoside substrate and were sorted for fluorescence (lacZ activity). Cells were collected from the positive and negative fractions directly into wells (200 cells per well). (C): The numbers of neurospheres were significantly greater from the Wnt+ fraction, containing 85% of all the neurosphere-forming cells (n = 15 wells, three independent sorts; *, p = .002). All data are represented as mean ± SEM.

Blocking Wnt Signaling In Vitro Reduces Stem Cell Number

Based on our hypothesis that Wnt signaling enhances the numbers of symmetric divisions in the stem cell population, we predicted that blocking Wnt signaling would result in a reduction in the numbers of secondary neurospheres following single-neurosphere passaging. We cultured primary SEZ cells or pure populations of neural stem and progenitor cells (derived from primary neurospheres) in the presence of sFRP2. Again, we observed no effect on the numbers (Fig. 4J) or size of the primary neurospheres in sFRP2-treated cultures (76.75 ± 3.50 μm vs. 74.09 ± 4.72 μm, control vs. sFRP2-treated, n = 100 neurospheres per group, p = .708). BrdU incorporation during neurosphere formation also revealed no changes in progenitor cell proliferation (23.8% ± 7.3% vs. 19.6% ± 4.3% BrdU+ cells, control vs. sFRP2-treated, n = 250 cells from two mice per group, p = .552). However, when individual sFRP2-treated neurospheres were passaged into control media (in the absence of sFRP2), we observed a 69.6% ± 13.1% decrease in the numbers of secondary neurospheres relative to controls (untreated; Fig. 4K). Together, these data indicate that endogenous Wnt signaling does not play a role in stem cell survival or progenitor cell proliferation, but is important for regulating stem cell number by promoting symmetric divisions.

Wnt Signaling Is Required for Stem Cell Expansion During Regeneration In Vivo

As a further test of our hypothesis that Wnt signaling is active in adult neural stem cells undergoing symmetric divisions, we took advantage of an in vivo paradigm that induces symmetric divisions of stem cells following progenitor cell ablation. Previous work has shown that 6 days of infusion of a mitotic inhibitor kills the rapidly dividing progenitor cells and decreases the numbers of neural stem cells in the SEZ immediately following the infusion [6, 12]. Following Ara-C treatment, neural stem cells divide symmetrically to repopulate the SEZ, and electron microscopic analysis reveals that only the stem cells are dividing at 24 hours post-Ara-C infusion [6]. These same studies demonstrated that by 2 days following Ara-C treatment, the numbers of neurospheres derived from the SEZ returns to control values [23], which is consistent with stem cells undergoing symmetric divisions during SEZ regeneration. Accordingly, this SEZ regeneration paradigm allows us to make some strong predictions regarding the role of Wnt signaling and symmetry of division. First, we predicted that within the first 24 hours post-Ara-C infusion, the BrdU-labeled stem cells (the only cells proliferating in this time frame) would have active Wnt signaling as they divided symmetrically to repopulate the SEZ. BAT-gal animals received intraventricular infusions of Ara-C for 6 days followed by two injections of BrdU (one in every 2 hours) starting 12 hours post-Ara-C treatment and were sacrificed immediately. We found that 30.0% ± 6.7% of the BrdU+ cells had active Wnt signaling (lacZ+) compared with 3.2% ± 0.9% of BrdU+ cells in saline-infused mice (Fig. 6A). Given that only 3.7% of stem cells have active Wnt signaling under baseline conditions (Fig. 1D), this represents more than eightfold increase in the number of stem cells with active Wnt signaling when stem cells are dividing symmetrically. Importantly, the overall increase in number cells with Wnt signaling within the SEZ was only 1.3-fold, indicating that the large increase is specific to the stem cell pool.

Figure 6.

Wnt signaling is important for stem cell expansion in vivo in response to injury. (A): There were significantly more BrdU/X-gal double-labeled cells in the SEZ 12 hours after the end of 6-day intraventricular Ara-C infusion when compared with both the control-infused and the baseline long-term label-retaining cells (as determined in Fig. 1D; n = 3 mice per group; *, p = .001). (B): Intraventricular infusion of sFRP2 (4 mg/ml) immediately following a 6-day intraventricular Ara-C infusion decreased the number of neurosphere-forming cells in the SEZ, indicating a slower or impaired recovery of stem cell numbers (n = 14 mice per group, two independent experiments; *, p = .047). (C–F): Stroke-lesioned animals have (C) increased numbers of neurosphere-forming cells (n = 5 mice per group; *, p = .013) and (D–F) increased numbers of cells with active Wnt signaling (X-gal+, blue; n = 4 mice per group; *, p = .040) within the SEZ 7 days post-PVD when compared with sham surgery. Scale bars = 100 μm. All data are represented as mean ± SEM. Abbreviations: Ara-C, cytosine-β-D-arabinofuranoside; BrdU, bromodeoxyuridine; LV, lateral ventricle; sFRP2, secreted frizzled-related protein 2; St, striatum.

A second prediction is that blocking Wnt signaling immediately following Ara-C treatment will reduce the number of symmetric divisions, thereby inhibiting the return of neurospheres during this time. Mice received intraventricular infusions of sFRP2 or saline for 2 days following Ara-C infusion. In sharp contrast to the lack of effect of sFRP2 infusion in control animals where stem cells are already divided asymmetrically, we observed a striking 9.3-fold decrease in the number of neurospheres from Ara-C- + sFRP2-treated brains relative to Ara-C- + saline-treated controls (Fig. 6B). Hence, blocking Wnt signaling during SEZ regeneration impairs neural stem cell recovery by inhibiting symmetric divisions.

Stroke Results in an Increase in Wnt Signaling in the SEZ In Vivo

The absolute numbers of neural stem cells in the adult central nervous system increases in response to injury [24, 25], suggesting that stem cells are induced to divide symmetrically, similar to what is observed following the Ara-C treatment. We asked whether Wnt signaling plays a role in this observed increase following brain injury. We performed a cortical stroke (pial vessel disruption [PVD]) overlying the motor and sensory cortices to avoid any mechanical damage to the SEZ [26]. Consistent with other cortical injury models [24], we observed a 286.2% ± 31.4% increase in neurosphere numbers at 7 days post-PVD, indicating an increase in the number of symmetric stem cell divisions in the SEZ (Fig. 6C). We examined the numbers of lacZ+ cells in adult BAT-gal mice at 7 days post-PVD and found a 251.4% ± 42.5% increase in the number of X-gal+ cells in the ipsilateral SEZ (Fig. 6D–6F). This finding is consistent with our data, which indicates that Wnt signaling is enhanced in conditions that promote symmetric stem cell division and suggests that Wnt is important for the endogenous regeneration response after brain injury.

DISCUSSION

We have found that Wnt signaling has a role in regulating the symmetry of division of adult neural stem cells. Active Wnt signaling was absent in most of the adult neural stem cells in the SEZ, consistent with previous reports that the vast majority of neural stem cells divide asymmetrically in that region. Further, we found that enhancing Wnt signaling resulted in increased numbers of stem cells without affecting progenitor cell proliferation or overall survival. Conversely, blocking Wnt signaling in conditions where stem cells are dividing symmetrically, such as during neurosphere formation in vitro or during SEZ repopulation in vivo, leads to a decrease in neural stem cells. Also important is the finding that blocking Wnt signaling in conditions where stem cells are dividing asymmetrically has no effect on stem cell number. Increased Wnt signaling was also observed in a model of cortical ischemia known to lead to increased numbers of stem cells, consistent with our observations that enhanced Wnt signaling results in symmetric divisions of the stem cells. Hence, we have demonstrated a role for Wnt signaling in modulating symmetry of division of adult neural stem cells under baseline conditions, during SEZ regeneration, and following injury.

We demonstrated a change in stem cell number depending on Wnt signaling, with no concomitant change in stem cell survival or progenitor cell proliferation. Although we can rule out a change in the proliferation kinetics of the majority of the population (i.e., the progenitor pool), it remains possible that we were unable to detect a change in the proliferation kinetics of the stem cell population as they comprise such a small subpopulation of cells. Interestingly, two recent studies have reported seemingly contradictorily results in response to Wnt signaling. Kalani et al. [3] and Qu et al. [27] reported increased and decreased progenitor proliferation, respectively. We suggest that the differences between these findings likely reflect the source of progenitor cells examined (embryonic forebrain vs. adult forebrain, including the hippocampus and SEZ). Notably, although our results do not rule out a shortened cell cycle time specifically in the stem cells, to observe the increased numbers of neurospheres, this change in cell cycle time would necessarily be concomitant with an increase in the frequency of symmetric divisions.

It was interesting to find that neither the drug treatment nor Wnt3a-RV used to enhance Wnt signaling resulted in activation of Wnt signaling in all cells. Instead, we observed enhanced Wnt signaling in a subpopulation of cells in the adult SEZ and adult-derived neurospheres. This is different from what we have seen in embryonic day 13.5 BAT-gal-derived neurospheres, which displayed more uniform activation within the precursor pool following drug or Wnt3a-RV treatments (unpublished data). This selective responsiveness may play an important role in modulating the response to Wnt at various stages of development and may account for the differences observed by different groups examining Wnt signaling. Muroyama et al. [7] and Hirabayashi et al. [28] found that enhancing Wnt signaling in cells cultured from the early embryonic brain (E11.5 forebrain and cortex respectively) resulted in a decrease in the number of neurospheres. In these publications, the decrease in the number of neurospheres was not attributed directly to a decrease in stem cell number, but rather to an impaired ability to form neurospheres, as it was accompanied by an increase in adherent cell clones on the bottom of the wells. Studies using cells from later embryonic stages (E13.5 lateral and medical ganglionic eminence and E14.5 SVZ) [3, 29, 30] observed an increase in neurosphere number. Together, these studies suggest that the response to Wnt signaling in neural stem and progenitor cells varies temporally; however, the findings are consistent with Wnt signaling enhancing symmetric stem cell division in the embryo.

Our in vitro studies used the neurosphere assay, a relatively simple and commonly used assay to assess neural stem and progenitor cells. Keeping in mind the limitations of the assay [31], when used with rigor and an understanding of the limitations of the colony-forming assay, neurosphere cultures remain one of the best ways to examine neural stem and progenitor populations. More recently, Louis et al. [32] developed a modified version of the neurosphere assay that involves growing neural-derived colonies for 21 days in a semisolid media (the neural colony-forming cell assay [NCFCA]). The advantages of the NCFCA are based on the claims that the media inhibits the movement of cells thereby ensuring clonality. Furthermore, the assay correlates the size of individual colonies at 21 days with multipotentiality, confirming that large colonies are stem cell derived. In regard to clonality, we have demonstrated that suspension cultures with low plating densities (used in our study) and in the absence of movement during the culture period will result in clonal neurospheres [20]. Regarding the size of the clones being an indicator of multipotentiality, our findings are entirely consistent with this observation as we have demonstrated both here and in a previous report that >95% of the neurospheres ≥100 μm in diameter that were taken after 7 days in vitro are multipotent [33]. Hence, the neurosphere assay remains a powerful tool to explore neural stem and progenitor cell populations.

Throughout our study, Wnt signaling was investigated in conditions that favor proliferation. Several studies have shown that in conditions that favor differentiation, Wnt signaling promotes neuronal phenotypes [28, 30, 34]. This context-dependent role of Wnt signaling is well illustrated in studies by Woodhead et al., demonstrating that during cerebral cortex development, Wnt signaling was crucial to the continued proliferation of the progenitor cells in the ventricular region and that the loss of Wnt signaling coincided with cell migration to, and neuronal differentiation within, the cortical plate. Interestingly, once the cells reached the cortical plate, Wnt signaling was reactivated [35]. In the adult brain, Wnt signaling may also be important for the promotion of neurogenesis as we observed extensive Wnt signaling in the adult olfactory bulb, the ultimate destination of the SEZ progenitor cells (data not shown). We also observed a rare fraction of stem cells with active Wnt signaling under baseline condition in the SEZ. These may represent instances of symmetric stem cell division that were too infrequent to be observed in previous work performing clonal analysis in vivo [5].

Our neurosphere culture conditions contain the mitogen EGF, which has been shown to be important for neural stem cell expansion in culture [36, 37]. Interestingly, an unequal distribution of the EGF receptor (EGFR) has been linked to cell fate in the neural stem and progenitor pool in vivo [38, 39], and more recently, EGFR has been demonstrated to regulate Notch signaling [40], a signaling pathway which is important for stem cell self-renewal [41]. Furthermore, EGFR asymmetry in the neural stem cell pool has been shown to be regulated by Dyrk1A [42], a kinase shown to play a role in a range of signaling pathways including Notch [43]. Cross talk between the Notch and Wnt signaling pathways has been observed in a number of stem cell systems [44–46] and it will be interesting to determine the potential interaction between the Wnt, Notch, and EGFR signaling pathways in the regulation of stem cell symmetry of division.

We observed activation of Wnt signaling in response to two distinct injury paradigms previously shown to lead to increased symmetric stem cell division: Ara-C infusion and stroke. Wnt signaling has been shown to play a role in response to injury in other systems including a model of neurotoxic injury in mouse retina [47], hair follicle regeneration after wounding in mice [48], and in Xenopus tadpole tail regeneration [49]. These studies hint at the possibility that Wnt signaling activation is a generalized response to injury, and our work suggests that the mechanism of enhanced regeneration is in part due to enhanced symmetry of division of the cells contributing to tissue repair.

Overall, this study reveals that Wnt signaling regulates the symmetry of adult neural stem cell division, promoting symmetric division leading to stem cell expansion. This finding provides insight into the mechanisms that regulate the symmetry of stem cell division and has implications for the development of regenerative medicine strategies that would benefit from expansion of the neural stem cell pool.

CONCLUSION

Wnt signaling modulates the size of the adult neural stem cell pool by regulating the symmetry of stem cell division.

Acknowledgements

We thank the van der Kooy Laboratory for providing the Wn3a-RV. This work was supported by the Canadian Institutes of Health Research and the Heart and Stroke Foundation (operating grants, C.M.M.). D.P. is the recipient of a CIHR Training Program in Regenerative Medicine Award and a CIHR Fellowship.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

The authors indicate no potential conflicts of interest.

Ancillary