SEARCH

SEARCH BY CITATION

Keywords:

  • Adult neural stem cell;
  • Astroglial niche;
  • Olfactory bulb;
  • Self-renewal;
  • Symmetric division;
  • WNT signaling

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Adult neural stem cells (NSCs) located in the subventricular zone (SVZ) persistently produce new neurons destined to the olfactory bulb (OB). Recent research suggests that the OB is also a source of NSCs that remains largely unexplored. Using single/dual-labeling procedures, we address the existence of NSCs in the innermost layers of the OB. In vivo, these cells are more quiescent that their SVZ counterparts, but after in vitro expansion, they behave similarly. Self-renewal and proliferation assays in co-culture with niche astrocytes indicate that OB-glia restricts NSC activity whereas SVZ-glia has the opposite effect. Gene expression profiling identifies WNT7A as a key SVZ-glial factor lacking in OB-glia that enhances self-renewal, thereby improving the propagation of OB-NSC cultures. These data demonstrate that region-specific glial factors account for in vivo differences in NSC activity and point to WNT7A as a tool that may be instrumental for the NSC expansion phase that precedes grafting. STEM CELLS 2012;30:2796–2809


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Neural stem cells (NSCs) persistently produce new neurons in two defined locations of the postnatal and mature mammalian brain: the subventricular zone (SVZ) that lines the lateral ventricles and the subgranular zone of the hippocampal dentate gyrus [1, 2]. Adult NSCs in these locations can be identified through the coexpression of several markers, including the glial fibrillary acidic protein (GFAP) that also labels mature astrocytes, the intermediate filament NESTIN, the stem cell transcription factor SOX2, and the astrocyte-specific glutamate transporter (GLAST) [3–7]. The latter markers are expressed during development by radial glia, the central nervous system fetal stem cells that give rise to the adult NSC populations [8]. In the rodent brain, the SVZ comprises the most active adult stem cell pool, constantly generating hundreds of immature neurons that travel in chains through the rostral migratory stream (RMS) to the olfactory bulb (OB), where they terminally differentiate into interneurons [9, 10]. In the human brain, NSCs in the SVZ contribute new neurons to the RMS-OB and to the prefrontal cortex, although neurogenesis drastically decays beyond infancy [11].

Several lines of evidence suggest that the adult RMS-OB continuum is not only a conduit for the SVZ-derived neuronal progeny but also a source of local NSCs that has remained largely unexplored. Viral tracing experiments recently demonstrated that GFAP-expressing cells located along the adult RMS behave as stem cells and contribute functional neurons to the OB circuitry [12]. Moreover, NSC cultures have been successfully derived from both the adult rodent and human OB [13–17]. Provided the OB is easily accessible by surgery and given OB-derived NSCs generate neuronal progeny after grafting back into the brain [16, 17], OB-NSCs have been proposed in the last decade as a valuable source for autologous transplantation in neurodegenerative disorders [14, 18]. Nevertheless, the developmental origin of adult OB-NSCs, their precise location and proliferative dynamics in vivo, and their contribution to adult neurogenesis have been poorly characterized.

In the adult brain niches, NSCs are in close contact with astrocytes and endothelial cells [19–21], which participate in creating an environment that maintains stem cell self-renewal and neuronal birth throughout life. The identity of the molecules secreted by niche cells is not completely defined yet. Interestingly, astrocytes are regionally specified [19, 22–24] and only those from the postnatal and adult niches are capable of instructing stem cells to produce neuronal progeny [25, 26]. It remains to be explored whether astrocytes from different brain areas are also heterogeneous regarding the ability to induce adult NSC proliferation and self-renewal.

In this study, we characterize a population of slowly dividing cells expressing NSC markers that is located in the adult OB core and we compare the properties of OB-NSCs side-by-side with those of stem cells in the SVZ. We have found that OB-NSCs are more quiescent than SVZ-NSCs in vivo, but after in vitro expansion, they behave similarly. We also analyze the self-renewal and proliferation of OB-NSCs in coculture with astrocytes derived from several postnatal brain regions, including the SVZ and the OB. Following gene expression profiling of the glial cells and a series of gain and loss of function experiments, we identify WNT7A as an astrocyte-derived niche signal that promotes symmetric self-renewing divisions through the noncanonical WNT signaling pathway, thereby facilitating the expansion of OB-NSC cultures.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Animals and In Vivo Manipulations

Housing of CD-1 mice and all experiments were carried out according to European Union 86/609/EEC and protocols approved by the Instituto de Salud Carlos III animal care and use committee. Adult animals were injected intraperitoneally with bromodeoxyuridine (BrdU; 50.0 mg/kg b.wt.) three times every 2 hours and were sacrificed 1 hour after the last injection. Chlorodeoxyuridine (CldU; 42.5 mg/kg) was injected three times every 2 hours and after 18 days iododeoxyuridine (IdU; 57.5 mg/kg) was injected over 3 days, six times every 12 hours, and mice were sacrificed 12 hours after the last injection [27, 28].

Tissue Processing, Immunostaining, and Quantification

Animals were fixed with paraformaldehyde (PFA) and immunostained as described in [29] (Supporting Information Table S1 for antibody details). For immunohistochemistry of NSC markers, brains were coronally processed in 30-μm-thick cryostat sections that were collected in series of 10 slides. Each slide contained an anterior-posterior (AP) reconstruction of six brain sections separated 300 μm between them. The quantification of dividing GFAP+SOX2+ cells was performed in every second section of the reconstruction. Thus, the distance between the sections analyzed was 600 μm. At least three OB and three SVZ serial sections per animal were included in the analysis, which spanned AP from Bregma approximately 5.1–3.9 mm for the OB and 1.1 to −0.1 mm for the SVZ. The distribution of GFAP+SOX2+ cells in the subependymal, granule, and plexiform OB layers was calculated as the average number of double-positive cells in five random fields (100 μm2) per layer per animal. For CldU and IdU immunohistochemistry, 40-μm-thick vibratome sections were collected in series of 24 wells. Quantification of the CldU+ and IdU+ populations was analyzed in three consecutive OB sections that spanned AP from Bregma approximately 4.5–4.4 mm. For the SVZ analysis, three serial sections separated 960 μm between them, spanning AP from Bregma approximately 5.4–3.5 mm, per animal were counted. Tissue sections were captured in a Leica Spectral SP5 confocal microscope and analyzed with LAS AF Lite Leica confocal imaging software (Leica, Heidelberg, Germany). To delineate the OB cell layers, the image of the OB structure as stained with 4′-6-Diamidino-2-phenylindole (DAPI) was used. The inner granule cell layer (iGCL) was taken as the most internal 1/9 of the total GCL. The outer granule cell layer (oGCL) was taken as the most external 8/9 of the total GCL, up to the limit of the internal plexiform layer. The average area of the GCL per section was 990,563 ± 58,323 μm2 (n = 4 animals). The iGCL boundary and the 100 μm2 random fields were traced using the LAS AF Lite Leica confocal imaging software or ImageJ. Immunostained neurospheres were captured in a Leica Spectral SP5 confocal microscope. Phase-contrast micrographs of floating neurospheres and immunofluorescent images of cell pairs and glial cultures were captured in an Axiovert 40 CFL Zeiss inverted microscope or in an Axio Imager A1 Zeiss upright microscope. Images were analyzed with Axio Vision 4.8 Zeiss software.

Neurosphere Culture

The method for NSC culture has been mainly adapted from [14, 29]. OB and SVZ tissue was dissected from 2-month-old adult mouse brain. Single cells were seeded and expanded in NSC growth medium (Dulbecco's modified Eagle's medium [DMEM]:F12-GlutaMAX supplemented with 0.6% glucose, 0.1% NaHCO3, 5 mM HEPES, hormone mix [14], 0.4 mg/ml bovine serum albumin (BSA), 0.72 U/ml heparin, 20 ng/ml epidermal growth factor [EGF], and 10 ng/ml fibroblast growth factor 2 [FGF-2]).

GLAST Microbead Purification

OB and SVZ tissue from 2-month-old mice was dissociated into a single-cell suspension and was incubated with anti-GLAST (ACSA-1) microbeads for magnetic cell separation according to manufacturer's instructions (Miltenyi Biotec Inc., Auburn, CA). Cell fractions were analyzed by flow cytometry (FACScalibur, Becton Dickinson, Franklin Lanes, NJ) or were plated in NSC growth medium.

Astrocyte Culture

Primary astrocyte cultures were established as described [19, 25, 30]. Briefly, tissue from OB, SVZ, and ventral mesencephalon (VM) of postnatal day 3 pups or 1-month-old animals was mechanically dissociated and plated in glial growth medium (DMEM:F12 with 10% Fetal Bovine Serum (FBS)). Astrocyte enrichment was performed by shaking for ≥2 hours to remove less adhesive cells.

In Vitro Assays

For all the in vitro assays, OB-NSCs were seeded in NSC growth medium at low density (2.5 cells per microliter). OB-neurospheres were counted and their diameter was measured after 5–6 days in vitro (div) [29]. For the indirect coculture, OB-NSCs were seeded in a 0.4 μm transwell placed on top of the glial monolayers, allowing the exchange of soluble factors. For the glia-conditioned medium (CM), NSC growth medium was conditioned during 72 hours by the glial feeder cultures and was filtered. OB-NSCs were plated in the glia-CM. For the loss-of-function assay, astrocytes were transfected with siRNA-Wnt7a SMARTpool (Dharmacon-Thermo Fischer Scientific, Lafayette, CO) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). After 4 days, OB-NSCs were plated in CM. For the gain-of-function assay, OB-neurospheres were seeded in the presence of human recombinant WNT7A (R&D Systems, Minneapolis, MN) or LiCl (Sigma-Aldrich Inc., St Louis, MO). For in vitro BrdU proliferation assay, spheres were attached to a coverslip coated with Matrigel (Becton Dickinson, Franklin Lanes, NJ), pulsed with 5 μM BrdU, and fixed with 2% PFA. For the self-renewal assays, spheres formed in presence or absence of WNT7A were mechanically dissociated to a single-cell suspension, replated in growth medium without WNT7A, and secondary spheres were counted. For the cell-pair assay, single cells were treated with WNT7A and fixed after 24 hours with 2% PFA as described [31].

NSC Nucleofection and Luciferase Assays

Neurospheres grown for 2 days were nucleofected with Topflash β-catenin-responsive firefly luciferase reporter and Renilla luciferase construct as an internal control [25]; or empty/β-cateninCA constructs [32] using the Mouse NSC Nucleofector Kit (Amaxa Biosystems, Lonza Cologne GmbH, Köln, Germany). Luciferase activity was measured in cell lysates 24 hours after, using the Dual Luciferase Assay System (Promega, Madison, WI) [33]. For gene expression assays, nucleofected cells were plated in the presence or absence of WNT7A and cultured for 6 or 24 hours before being harvested.

RNA Isolation, Reverse Transcription, and PCR Analysis

Total RNA was isolated using RNeasy micro/miniKit (Qiagen Inc., Valencia, CA) and 100–500 ng was used to synthesize cDNA using PrimeScript Reverse Transcriptase (Takara Bio Inc., Otsu, Japan). qRT-PCR was performed with Sybr Green in a Light Cycler 2.0 or 480 instrument (Roche Diagnostics GmbH, Mannheim, Germany) and analyzed by Pfaffl method. Primer sequences are available in Supporting Information Table S2.

Arrays

For the glial transcriptome analysis, generation of double-stranded cDNA, preparation and labeling of cRNA, hybridization to Mouse Genome 430 2.0 Array, and data analysis were performed according to Affymetrix (Santa Clara, CA). Data have been deposited in the gene expression omnibus (accession number GSE36456). OB-NSC expression was analyzed with GEArray S Series Mouse Stem Cell Gene Arrays MM-601.2 (SuperArray Bioscience Corporation, Frederick, MD) according to the manufacturer.

Western Blot

Astrocytes were lysed in 50 mM Tris-HCl pH 7.5, 137 mM NaCl, 2 mM EDTA, 0.2% Triton X-100, protease Inhibitor Cocktail (Roche Diagnostics GmbH, Mannheim, Germany). Total protein extracts (20 μg) were resolved in SDS-PAGE. Polyvinyl difluoride (PVDF) membrane was incubated with goat anti-WNT7A (Santa Cruz Biotechnology Inc., Santa Cruz, CA) and mouse anti-β-actin (Sigma-Aldrich Inc., St Louis, MO). Immunodetection was performed with enhanced chemiluminescence (ECL) reagent (GE Healthcare, Buckinghamshire, UK).

Statistical Analysis

Analysis of significant differences between means was performed using two-tailed Student's t test (paired or unpaired). When comparing relative values and percentages, data were first normalized using log or arcsen transformation. Data are presented as mean ± SEM and n indicates the number of independent cultures or mice used. A probability value of less than 5% was considered significant: ***, p < .001; **, p < .01; and *, p < .05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Slowly Dividing Cells Expressing NSC Markers Are Located in the Adult OB

We sought to investigate the existence of candidate stem cells in the adult mouse OB by double immunofluorescence for GFAP and SOX2, a combination of markers that identifies NSCs in the adult SVZ (Supporting Information Fig. S1). Confocal microscopy analysis of OB sections revealed the presence of GFAP+SOX2+ cells predominantly in the OB core region, including the subependymal layer (SEL) and the iGCL of the OB (Fig. 1A). To a lesser extent, double-positive cells were also found in the periglomerular layer, yet very few GFAP+SOX2+ cells were located in other outer bulbar regions (Fig. 1A, 1B). Given the staining was more prominent in the iGCL and SEL, which corresponds to the rostral extension of the germinal ventricle during development [14], we focused our analysis on these areas. Cells coexpressing GFAP, SOX2, and the cell cycle protein Ki67 were visualized in the iGCL-SEL, suggesting that part of the putative stem cells were dividing (Fig. 1C). To determine their proliferative activity, we injected adult mice with BrdU using a short administration paradigm. A small fraction of the GFAP+SOX2+ cells in the iGCL-SEL were undergoing S-phase and thus incorporated BrdU (2.0% ± 0.7%, n = 3; Fig. 1D). In some sections, we found SOX2+GFAP+ cells that were BrdU+ at the margin of an open cavity located in the central axis of the OB that could be a remnant of the collapsed olfactory ventricle (Fig. 1E). In contrast to the OB, quantitative analysis of the percentage of GFAP+SOX2+ cells that were proliferating in the SVZ showed that the stem cell pool next to the ventricles is at least 10 times more active (20.0% ± 1.7%, n = 3, p < .001; Fig. 1D, 1F).

thumbnail image

Figure 1. In vivo distribution and proliferative activity of candidate OB-neural stem cells (NSCs) coexpressing GFAP and SOX2. (A): Upper panel, sagittal diagram of the mouse brain showing a coronal plane that indicates how the tissue was processed. A coronal view of the OB structure (as it appears in all coronal sections analyzed) is also provided on the left. The RMS is shown in red and the SVZ in gray. Lower panel, confocal microscopy image of a coronal OB section (corresponding to the area boxed in the upper panel) stained for GFAP (red) and SOX2 (green). Numerous cells coexpressing SOX2 and GFAP are seen in the inner region. (B): Quantitative analysis of GFAP+ SOX2+ cell distribution in the different OB regions. Double-positive cells are more abundant in the SEL and iGCL than in peripheral layers. (C): Confocal images showing a subset of GFAP+SOX2+ (red, magenta) candidate NSCs in the iGCL-SEL region of the OB that are cycling (Ki67+ cells, green). (D): Top, diagram representing the BrdU injection procedure used. Bottom, quantitative analysis of the percentage of GFAP+SOX2+ NSCs that are also BrdU+ in the SVZ and in the OB of 2-month-old wild-type (CD-1) mice (n = 3 animals). (E): Confocal micrographs showing a GFAP+SOX2+ (magenta, red) candidate OB-NSC that incorporated BrdU (green) lining a central cavity that may be a remnant of the OV. The margin of the cavity is indicated by the dotted line. (F): Confocal micrographs showing GFAP+SOX2+ (magenta, red) SVZ stem cells that incorporated BrdU (green). The lateral ventricle wall is indicated by the dotted line. Nuclear labeling with DAPI is shown in blue. Anterior-posterior coordinates from Bregma are approximately + 4.5 mm for the sections displayed in (A), (C), and (E), and + 0.5 mm for the section displayed in (F). Unpaired t test: **, p < .01; ***, p < .001. Scale bars = 100 μm in (A), 25 μm in (C), 10 μm in (E) and (F). Abbreviations: BrdU, bromodeoxyuridine; DAPI, 4′-6-diamidino-2-phenylindole; EPL, external plexiform layer; GFAP, glial fibrillary acidic protein; GL, glomerular layer; iGCL, inner granule cell layer; IPL, internal plexiform layer; MCL, mitral cell layer; OB, olfactory bulb; oGCL, outer granule cell layer; OV, olfactory ventricle; RMS, rostral migratory stream; SEL, subependymal layer; St, striatum; SVZ, subventricular zone.

Download figure to PowerPoint

In the SVZ, the slowly dividing label-retaining cells (LRCs) correspond to a population of resident stem cells [3, 34, 35]. To assess the existence of LRCs in the OB core and to determine their proliferative activity, we next injected adult mice with two different thymidine analogs. Initially, CldU was administered and after 3 weeks, IdU was injected (Fig. 2A). This delay allows intermediate progenitors to dilute out the CldU label and immature neurons to exit the cell cycle, migrate out of the RMS, and reach their final location at the OB superficial layers [36]. Double immunofluorescence analysis showed that most CldU+ cells were located in the outer GCL (oGCL, Fig. 2B, 2D) and corresponded to newly generated neurons that incorporated CldU during their last S-phase (NeuN+CldU+ neurons, Supporting Information Fig. S2A). However, a small percentage of the CldU+ cells were found scattered throughout the innermost GCL and SEL (Fig. 2B). These CldU+ cells coexpressed the stem cell markers SOX2 and GFAP (Fig. 2C) and were considered as slowly dividing LRCs. In marked contrast to the CldU staining, the IdU signal was predominantly detected in the iGCL-SEL, with little signal in the oGCL (Fig. 2B, 2D). The majority of the IdU+ cells were CldU and corresponded to immature neurons expressing doublecortin that reached the OB through the RMS (Supporting Information Fig. S2B). However, we also detected LRCs in the iGCL-SEL that incorporated IdU (Fig. 2F). Quantitative analysis indicated that 9.0% ± 5.6% (n = 4) of the CldU+ LRCs cells in the iGCL-SEL colabeled for IdU, which is approximately six times less than the percentage of CldU+ LRCs in the SVZ that incorporated IdU (56.8% ± 1.4%, n = 4; p < .001; Fig. 2A, 2E). Together, our results indicate that slowly dividing cells expressing NSC markers are found in the adult OB core, and that these cells show a marked reduction in proliferative activity compared to their SVZ counterparts.

thumbnail image

Figure 2. LRCs in the inner layers of the OB undergo proliferation. (A): Diagram depicting the dual labeling procedure used to analyze the activity of LRCs. The thymidine analogs CldU and IdU were administered sequentially to 2-month-old wild-type (CD-1) mice, at equimolar concentrations. The percentage of LRCs (CldU+ cells) that incorporated IdU is shown (n = 4). The double labeling was analyzed by confocal microscopy in approximately 20 CldU+ cells from the SEL-iGCL per animal. (B): Coronal section showing the distribution of CldU+ (green) and IdU+ (red) cells in the OB by confocal microscopy. Note that IdU+ cells were predominantly located around the SEL, while CldU+ cells were scattered along the GCL. (C): Confocal micrograph showing a GFAP+SOX2+ (white, red) LRC in the SEL-iGCL that stained for CldU+ (green). (D): Quantitative analysis of the CldU+ and IdU+ cell distribution in the OB layers. The average number ± SEM of CldU+ cells per OB section was: 14 ± 3 in the SEL+iGCL and 232 ± 75 in the oGCL. The average number ± SEM of IdU+ cells per OB section was: 199 ± 44 in the SEL+iGCL and 22 ± 4 in the oGCL. Data in the graph are shown as average percentage ± SEM (n = 4 animals). (E): Confocal micrograph showing a CldU+ LRC (green) in the SVZ that incorporated IdU (red, arrowhead) and a CldU+ LRC that did not incorporate IdU (arrow). (F): Confocal micrograph showing a CldU+ LRC (green) in the OB that incorporated IdU (red, arrowhead). Nuclear labeling with DAPI is shown in blue. Unpaired t test: ***, p < .001. Scale bars = 20 μm in (B), 10 μm in (C), (E), and (F). Abbreviations: CldU, chlorodeoxyuridine; DAPI, 4′-6-diamidino-2-phenylindole; IdU, iododeoxyuridine; GFAP, glial fibrillary acidic protein; iGCL, inner granule cell layer; LRCs, label-retaining cells; OB, olfactory bulb; oGCL, outer granule cell layer; SEL, subependymal layer; St, striatum; SVZ, subventricular zone.

Download figure to PowerPoint

GLAST-Positive Cells Comprise All OB Cells Forming Self-Renewing Neurospheres

Previous studies demonstrated that cells dissociated from the adult OB tissue can give rise to neurosphere cultures [14, 16, 17]. However, the cell-of-origin of the primary spheres has not been established. To define the identity of the OB cells with neurosphere-forming capacity, we next turned to a microbead assay that allows for the separation of cells based on the expression of extracellular epitopes. We hypothesized that antibodies raised against the transmembrane glycoprotein GLAST may allow for the prospective purification of OB-NSCs, since GLAST is expressed by radial glia-derived GFAP+ stem cells in the adult SVZ [6].

We first checked for the expression of GLAST in the GFAP+SOX2+ cells of the OB. As in the SVZ (Fig. 3A), we found OB cells coexpressing the three markers in the iGCL-SEL (Fig. 3B). Quantitative analysis by confocal microscopy showed that 57.4% ± 4.5% of the GFAP+SOX2+ cells was GLAST+ (n = 4, Supporting Information Fig. S3). Next, we dissected adult OB tissue and adult SVZ tissue as a control, and we separated GLAST+ and GLAST living cells by means of the magnetic microbeads. As shown in Figure 3C and in Supporting Information Figure S3, only the GLAST+ cell fraction from the OB and SVZ was capable of generating primary spheres in serum-free defined medium supplemented with EGF and FGF-2. These data suggest that the OB stem cell population belongs to the radial glia GLAST+ lineage, which in our experimental conditions comprises all OB cells forming neurospheres in vitro. Interestingly, we found that primary spheres derived from OB GLAST+ cells were smaller in size compared to the spheres generated from SVZ GLAST+ cells, pointing to a reduced proliferative activity of the OB stem cells. This was also observed in primary sphere cultures that did not undergo the microbead purification step (Fig. 3D). Nevertheless, OB neurosphere cultures could be expanded in vitro over time at a stable rate (Fig. 3E). Actively proliferating OB cultures at late passages expressed the stem cell marker SOX2 and the immature filament NESTIN and were indistinguishable from SVZ cultures (Fig. 3F, 3G).

thumbnail image

Figure 3. OB-neural stem cells (NSCs) belong to the GLAST+ lineage, which comprises all spherogenic cells in vitro. (A): Confocal micrographs showing GLAST expression in GFAP+SOX2+ cells in the SVZ. (B): Confocal micrographs showing GLAST expression in GFAP+SOX2+ cells located in the inner granule cell layer-subependymal layer of the OB. (C): Left, microbead assay used to enrich for GLAST+ cells. Right, phase-contrast micrographs showing that only the GLAST+ cell fraction of the OB and SVZ tissue formed primary neurospheres. The images are representative of three independent experiments. (D): OB neurosphere cultures proliferate more slowly than SVZ cultures upon isolation, as indicated by the reduced primary sphere diameters (n = 7). (E): Long-term growth curve of OB neurosphere cultures expanded in vitro over 4 weeks (n = 6). (F): At late passages, OB-NSC cultures and SVZ-NSC cultures had a similar proliferative behavior, as indicated by the average sphere diameter (n ≥ 6). (G): Confocal micrograph of a representative OB neurosphere at late-passage, showing the widespread expression of SOX2 (red) and NESTIN (white) as well as the incorporation of BrdU (green). Nuclear labeling with DAPI is shown in blue. Paired t test: *, p < 0.05. Scale bars = 10 μm in (A) and (B), 100 μm in (C), 5 μm in (G). Abbreviations: BrdU, bromodeoxyuridine; DAPI, 4′-6-diamidino-2-phenylindole; GFAP, glial fibrillary acidic protein; GLAST, astrocyte-specific glutamate transporter; OB, olfactory bulb; St, striatum; SVZ, subventricular zone.

Download figure to PowerPoint

Diffusible Factors from OB and SVZ Astrocytes Have Opposing Effects on OB-NSC Activity

Our in vivo and in vitro results indicated that OB-NSCs are less proliferative than SVZ-NSCs, yet the lack of activity can be progressively surmounted in vitro after several passages in the presence of mitogenic stimulation. This raises the possibility that the activity of the OB stem cell pool is limited in vivo and during the initial phase of cell culture growth due either to the lack of stimulatory signals and/or to the presence of inhibitory molecules in the OB tissue relative to the SVZ tissue. Provided astrocytes are important components of adult NSC niches, endowed with region-specific properties, we reasoned that OB and SVZ astrocyte-derived signals could contribute to the behavior of adult NSCs. We also reasoned that identifying such signals would be instrumental to improve the early expansion of OB-NSC cultures.

To investigate the role of astrocytes, we first established primary astrocyte cultures from three postnatal brain regions: the OB, the SVZ, and the VM (an unrelated non-neurogenic region used as a control). The cultures were grown as monolayers in the presence of serum and were characterized by immunofluorescence as GFAP+A2B5 type 1 astrocytes with protoplasmic morphology (Supporting Information Fig. S4 and Fig. 4D). Next, we analyzed the influence of the astrocytes on the adult OB stem cells by means of indirect coculture performed in serum-free medium supplemented with EGF and FGF-2 (Fig. 4A). OB neurospheres were mechanically dissociated and seeded at low density (2.5 cells per microliter) in the upper transwell insert of the coculture device and astrocytes were grown as feeders in the lower compartment. The number of neurospheres formed after 6 days allowed to estimate the clonogenic capacity of the OB-NSCs exposed to the astrocytic signals. The diameter of the spherical clones was measured to determine the overall effect of the astrocytes on proliferation. As shown in Figure 4B–4D, diffusible factors from the postnatal SVZ astrocytes markedly promoted sphere formation (p < .01) and proliferation (p < .05) when compared with control conditions without glial feeders or with astrocytes from the postnatal VM. Instead, a significant although modest decrease in the number of spheres, but no effect on proliferation, was found in the coculture with postnatal OB astrocytes. Similar results were obtained when OB-NSCs were cocultured with astrocytes isolated from equivalent brain regions of adult animals (Supporting Information Fig. S5A). To further confirm these observations, we examined the effect elicited by CM from the different glial cultures. As shown in Figure 4E–4G, CM from the postnatal region-specific astrocytes recapitulated the effect found in the coculture system (p < .001). Similar results were also obtained when CM was assayed on adult NSCs derived from the SVZ (Supporting Information Fig. S5B). These data indicate that diffusible factors secreted by OB and SVZ astroglial cells have opposing effects on the spherogenic capacity of adult NSCs, while only the factors from SVZ astrocytes promote proliferation. Thus, the regionalization of OB and SVZ astrocytes may underlie, at least in part, the differences between OB and SVZ NSC activity detected in vivo.

thumbnail image

Figure 4. Diffusible factors from OB and SVZ astrocytes regulate adult NSC activity. (A): Schematic drawing illustrating the indirect coculture system. OB-NSCs were seeded in the upper transwell, while astrocytes were grown as feeder cells in the lower compartment. The two independent compartments were separated by a 0.4 μm pore semipermeable membrane that allows diffusion of soluble factors. (B): Number of OB neurospheres obtained by plating equal numbers of cells dissociated from OB-NSC cultures in the presence of astrocytes from different brain areas or in the absence of astrocytes as a control (n ≥ 4). (C): Average size of the OB neurospheres obtained in the coculture (n ≥ 4). (D): Top panel, phase-contrast micrographs of floating OB neurospheres in the cocultures. OB-NSCs are more spherogenic and proliferate more rapidly in the presence of SVZ astrocytes, as it can be seen by the number of spherical clones and their enlarged diameter. Bottom panel, region-specific astrocytes stained for GFAP (green). Nuclear label DAPI is shown in blue. (E): Schematic drawing illustrating the experimental setup for the conditioned medium (CM) assays. Adult OB-NSCs were treated with filtered serum-free medium from region-specific astrocytic cultures that had been grown for 3 days in vitro (div). The CM was supplemented with EGF and FGF-2. (F): Number of OB neurospheres obtained by plating equal numbers of cells dissociated from OB-NSC cultures in CM from the astrocytes (n = 6). (G): Average size of the OB neurospheres generated in the CM (n = 6). Paired t test: *, p < .05; **, p < .01; ***, p < .001. Scale bars = 100 μm in top panel and 30 μm in bottom panel of (D). Abbreviations: DAPI, 4′-6-diamidino-2-phenylindole; GFAP, glial fibrillary acidic protein; NSCs, neural stem cells; OB, olfactory bulb; SVZ, subventricular zone; VM, ventral mesencephalon.

Download figure to PowerPoint

Transcriptome Analysis Identifies WNT7A as a Niche Signal Secreted by SVZ Astrocytes

Since the most relevant effect in the coculture and CM experiments was the marked NSC stimulation caused by SVZ astrocytes, we next aimed at identifying the diffusible factors specifically secreted by SVZ astroglia. To this end, the transcriptome of the different glial cultures was examined using Affymetrix Mouse Genome 430A 2.0 arrays (Fig. 5A). We focused our analysis on the genes that were significantly overexpressed in SVZ relative to control VM astrocytes with a signal log ratio (the log 2 of the fold change) greater than two. According to our criteria, 267 genes were overexpressed in SVZ versus VM samples. Functional data mining by Gene Ontology identified 58 genes of the “extracellular space” class (GO: 0005615). Among these genes, we selected those that did not belong simultaneously to the “proteinaceous extracellular matrix” or the “integral/intrinsic to plasma membrane” classes (GO: 0005578; GO: 0005887; and GO: 0031226), which code for nondiffusible proteins. Next, we narrowed the list to the genes that were also overexpressed in SVZ relative to OB samples. After completing the analysis, we found 15 genes specifically overexpressed in SVZ astrocytes that encode secreted soluble proteins (Fig. 5B; Supporting Information Table S3), including a variety of factors that are related to neural development (Wnt7a, Metrn, Sema3f, and Sostdc1), hormones and growth factors (Pdgfb, Insl6, Igfbp3, and Rln1), as well as genes involved in the immune and inflammatory response (Defb1 and Hc), and chemokines and neuropeptides (Ccl20, Npvf, and Scg5). A similar analysis was carried out to identify the genes overexpressed in OB astrocytes. We found 21 genes specifically overexpressed in the OB glial cultures that encode secreted soluble proteins (Supporting Information Table S4), including complement components (C1qa, C1qb, and C1qc), several chemokines (Cxcl13, Ccl12, and Ccl3), peptidase-related genes (Egfl6, Ctss, Klk6, Prss35, and Serpind1), and other signaling genes, such as Bmp7 and the WNT antagonist Sfrp4.

thumbnail image

Figure 5. Comparative transcriptome analysis identifies WNT7A as an SVZ astrocytic niche factor. (A): Scatter graphs of the comparative global gene expression analysis between SVZ glia, OB glia, and VM glia. Red, genes that are overexpressed in the comparison. Blue, genes that are repressed in the comparison. Scatters graphs show a typical “rocket distribution” consistent with normal gene expression in the three comparisons. (B): Venn diagram illustrating the differential gene expression between the three glial populations. Fifteen genes that encode secreted soluble molecules were specifically overexpressed in SVZ astrocytes with regard to OB glia and VM glia. (C): qRT-PCR validation panel of 12 overexpressed SVZ glia genes. VM glia was used as the calibrator sample (n ≥ 3). (D): Quantification of WNT7A protein from OB, SVZ, and VM glia lysates after Western blot. Consistent with the mRNA results, SVZ glia expresses more WNT7A protein than OB glia (n = 3). (E): Western blot analysis of WNT7A and β-actin expression (used as the housekeeping control) in OB, SVZ, and VM glia lysates. (F): qRT-PCR analysis of Sfrp4 expression in OB and SVZ astrocytes. VM glia was used as the calibrator sample (n = 6). Unpaired t test: *, p < .05; **, p < .01; ***, p < .001. Abbreviations: OB, olfactory bulb; SVZ, subventricular zone; VM, ventral mesencephalon.

Download figure to PowerPoint

To validate the changes found in the array, the expression of a subset of 12 SVZ genes was determined by qRT-PCR analysis of independent glial cultures isolated from the same brain areas (Fig. 5C). The direction of gene expression was confirmed for all genes, and the most significant differences were found for Scg5, Metrn, Nvpf, and Wnt7a. Among the validated factors, we were interested in Wnt7a, provided it has been shown that WNT signaling regulates NSCs during embryonic development and in adulthood [25, 37–39]. Next, we confirmed that Wnt7a was overexpressed in SVZ astrocytes not only at the mRNA level (4.4 ± 1.5-fold increase in gene expression, n = 3, p < .01; Fig. 5C) but also most importantly at the protein level, as analyzed quantitatively by Western blot (3.2 ± 0.9-fold increase in WNT7A protein, n = 3, p < .05; Fig. 5D, 5E). Conversely, we confirmed that the gene encoding the WNT7A antagonist-secreted frizzled-related protein 4 (sFRP4) [40] was strongly expressed in OB astrocytes (Fig. 5F). Thus, we concluded that WNT7A is a niche signal produced by SVZ astrocytes that could potentially regulate NSCs in the coculture system and we propose that differences in OB/SVZ glial expression of WNT antagonists (Sfrp4 in OB glial cells) and ligands (Wnt7A in SVZ glial cells) may partly explain the differential activity of NSCs in these brain regions (Fig. 7J).

WNT7A Mediates the Spherogenic Effect of SVZ Astrocytes

To explore whether OB-NSCs express receptors for the WNT ligands and thereby are able to bind WNTs, we hybridized total cDNA from OB neurospheres to a low-density array containing genes from the WNT signaling pathway. We found that OB-NSC cultures express several WNT receptors including Frizzled-9 and Frizzled-7, which are known to interact with WNT7A [41, 42], (Supporting Information Fig. S6; Table S5). These results were confirmed by qRT-PCR (Supporting Information Fig. S7A). In addition, other crucial components of both canonical and noncanonical WNT signaling were expressed, indicating that OB-NSCs can respond to WNT proteins (Supporting Information Fig. S7B).

We next performed a loss of function experiment using RNA interference to determine whether Wnt7a expression is necessary for the effects of SVZ astrocytes on OB-NSCs. We transfected SVZ glial cultures with a pool of Wnt7a-siRNAs or Scramble-siRNA as a control (Fig. 6A). Silencing was specific, since Wnt7a-siRNAs did not decrease the expression of other WNT transcripts expressed by SVZ astrocytes such as Wnt5a (not shown). Four days after silencing, CM from the interfered SVZ astrocytes was collected and was used in a neurosphere formation assay. OB-NSCs exposed to Wnt7a-siRNA CM showed a 40% reduction in the number of spheres relative to control CM (p < .05, Fig. 6B, 6D), indicating that WNT7A is required for the spherogenic effect of SVZ astrocytes. A similar trend was observed upon the addition of the WNT antagonist sFRP2 to the CM (reduction in the number of spheres in SVZ-CM supplemented with 250 ng/ml sFRP2: 32.0 ± 4.0%, n = 2). However, as inferred from the size of the spherical clones, Wnt7a-siRNA had no effect on proliferation (Fig. 6C, 6D), suggesting that factors other than WNT7A account for the mitogenic action of SVZ astrocytes.

thumbnail image

Figure 6. WNT7A enhances the spherogenic capacity of olfactory bulb (OB)-neural stem cells (NSCs). (A): qRT-PCR analysis of Wnt7a expression in subventricular zone (SVZ) astrocytes that were interfered with Wnt7a-siRNA or Scrambled siRNA as a control (n = 3). (B): Fold decrease in the number of OB neurospheres obtained after treating equal numbers of cells from OB-NSC cultures in the presence of conditioned medium (CM) from Wnt7a-siRNA or control (scrambled siRNA) SVZ astrocytes (n = 7). (C): Average size of the OB neurospheres obtained in (B) (n = 4). (D): Representative phase-contrast micrographs of (B) and (C). There are less OB-neurospheres in the CM from Wnt7a interfered astrocytes, but no changes were observed in their diameter. (E): Dose-response effect of recombinant WNT7A on the number of spheres generated from OB-NSCs (n ≥ 5). (F): Average size of the OB neurospheres obtained in (D) (n = 6). (G): Immunofluorescence analysis of BrdU incorporation (green) in control or WNT7A-treated neurospheres. Data refer to average ± SEM of the percentage of BrdU+ cells. No significant differences in proliferation were found (p = .790, n = 5). Nuclear label DAPI is shown in blue. (H): Dose-response effect of LiCl on the number of spheres generated from OB-NSCs (n ≥ 3). (I): Average size of the OB neurospheres obtained in (H) (n ≥ 3). (J): Immunofluorescence analysis of BrdU incorporation (green) in control or LiCl-treated neurospheres. Data refer to average ± SEM of the percentage of BrdU+ cells. Significant differences in proliferation were found (p < .05, n = 5). Nuclear label DAPI is shown in blue. Paired t test: *, p < .05; **, p < .01; ***, p < .001. Scale bars = 100 μm in (D), 10 μm in (G) and (J). Abbreviations: BrdU, bromodeoxyuridine; DAPI, 4′-6-diamidino-2-phenylindole.

Download figure to PowerPoint

To investigate whether WNT7A alone is sufficient to stimulate sphere formation, dissociated OB-NSCs were treated with increasing doses of purified recombinant WNT7A protein. As shown in Figure 6E, the addition of 50–100 ng/ml WNT7A significantly enhanced sphere formation (fold increase at 50 ng/ml WNT7A: 1.4 ± 0.1, n = 10, p < .01). A similar result was obtained when SVZ-NSCs were assayed (Supporting Information Fig. S8). Interestingly, the diameter of the spheres remained unchanged at all doses (Fig. 6F), confirming that WNT7A does not act as a mitogen for OB-NSC cultures. The lack of activity on overall proliferation was verified by examining BrdU incorporation. As shown in Figure 6G, the proportion of BrdU+ cells did not change in the presence of WNT7A (n = 5, p = .790), pointing to a selective effect of WNT7A on the clonogenic index. Interestingly, when OB-NSCs were treated with LiCl, a GSK3β inhibitor that mimics canonical WNT signaling through β-catenin stabilization, we found a dose-dependent effect in the number of neurospheres as well as an increase in the diameter of the clones (Fig. 6H, 6I). Accordingly, LiCl promoted BrdU incorporation (Fig. 6J, p < .05), pointing to a generalized effect of canonical WNT signaling on proliferation.

WNT7A Promotes Self-Renewal of OB-NSCs Through the Noncanonical Signaling Pathway

In low-density cultures, spheres are clonally derived colonies that are composed of a minority of stem cells and a majority of progenitor cells, the founding cell of the colony being a NSC. Thus, changes in the spherogenic capacity of expanded cultures are commonly interpreted as alterations in stem cell behavior while changes in the average size of the spheres are interpreted as an effect on progenitor proliferation. Consequently, the difference between the LiCl and WNT7A treatments suggests that LiCl may be acting on both the founding NSCs and the progenitor fraction to increase proliferation, while WNT7A may be acting specifically on the stem cell population, pointing to a role in self-renewal.

To confirm that WNT7A regulates self-renewal, neurospheres that had been grown in WNT7A were dissociated and plated in the presence of mitogens but without the WNT ligand. As shown in Figure 7A, WNT7A pretreatment increased the formation of secondary spheres, indicating that WNT7A promotes symmetric self-renewing divisions of OB-NSCs. A similar result was obtained when SVZ-NSCs were assayed (Supporting Information Fig. S8). To further verify this role, we next performed a cell-pair assay [33]. The outcome of the first OB-NSC division was scored by staining the cell doublets for epidermal growth factor receptor (EGFR) (Fig. 7B), which allows to distinguish symmetric stem cell divisions (EGFRhigh/EGFRhigh pairs) from asymmetric ones (EGFRhigh/EGFRlow pairs). Importantly, WNT7A treatment significantly increased the percentage of symmetric divisions from 42.8% ± 5.1% to 69.9% ± 4.3% (n = 4, p < .01, Fig. 7C) while decreasing asymmetric pairs.

thumbnail image

Figure 7. WNT7A promotes self-renewal and improves the expansion efficiency of OB-NSCs through the noncanonical signaling pathway. (A): WNT7A pretreatment improved the self-renewal capacity of OB-NSCs. More secondary spheres were formed in pretreated cells (n = 4). (B): Immunofluorescence for EGFR (green) in cell pairs showing symmetric (S-EGFR) and asymmetric (A-EGFR) distribution of the receptor among daughter cells. Nuclear label DAPI is shown in blue. (C): Quantification of the percentage of cell pairs with asymmetrical (A) or symmetrical (S) distribution of EGFR. Treatment with WNT7A significantly increases the percentage of S-divisions at the expense of A-divisions (n = 4). (D): WNT canonical signaling reporter assay in OB-NSCs. Cells were cotransfected with Topflash β-catenin-responsive and Renilla (internal control) luciferase reporters and stimulated with WNT7A for 24 hours. As positive control for TCF-induced transcription, cells were cotransfected with the constitutive active β-catenin plasmid (β-cateninCA) (A.U., arbitrary units of Topflash luciferase activity relative to Renilla values). (E): qRT-PCR analysis of Axin2 expression in OB-NSCs that were transfected with empty or β-cateninCA and stimulated with WNT7A for 6 hours and 24 hours. WNT7A transiently inhibits Axin2 expression at 6 hours (n = 3). (F): qRT-PCR analysis of Vangl2 expression in OB-neurospheres 24 hours after WNT7A treatment (n = 4). (G): Diagram representing the strategy used to assess the effect of WNT7A on the expansion of primary OB-NSCs. (H): Quantification of primary OB-neurosphere yield from 2-month-old CD-1 mice untreated and treated with WNT7A. (I): Quantification of secondary neurospheres formed from OB-primary neurospheres (H) that were dissociated to single cells and replated again in the same conditions (n = 7, both in H and I). (J): Model describing the role of WNT/sFRP soluble factors secreted by astrocytes in the regulation of adult NSC activity. In the SVZ, SVZ-astrocytes in the niche secrete WNT7A that specifically promotes NSC self-renewal but does not affect transient amplifying cell proliferation. However, in the OB, OB-astrocytes secrete factors that negatively impact on NSCs such as sFRP4 (inhibitor of WNT signaling). The lack of a net stimulatory signal could explain the low activity of OB-NSCs with regard to SVZ-NSCs. Paired t test: *, p < .05; **, p < .01. Scale bar = 2 μm. Abbreviations: DAPI, 4′-6-diamidino-2-phenylindole; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; FGF-2, fibroblast growth factor 2; iGCL, inner granule cell layer; NSC, neural stem cell; OB, olfactory bulb; SEL, subependymal layer; SVZ, subventricular zone; TAP, transient amplifying cell proliferation.

Download figure to PowerPoint

WNT7A can trigger either the β-catenin-dependent canonical pathway [38, 43] or the β-catenin-independent noncanonical pathway [41, 44], depending on the cellular context. To distinguish the possible function of WNT7A as a canonical or a noncanonical WNT, OB-NSCs were stimulated with WNT7A protein or were nucleofected with a stabilized constitutive active (CA) form of β-catenin as a control. The expression of the β-catenin/T-cell factor (TCF) target gene Axin2 was then analyzed at 6 and 24 hours after treatment (Fig. 7E). We found that WNT7A transiently decreased Axin2 expression at 6 hours while β-cateninCA activated Axin2 at both time points. Consistently, WNT7A failed to induce the β-catenin-responsive luciferase reporter Topflash while β-cateninCA strongly increased its activity (Fig. 7D). Of note, WNT7A increased the expression of the noncanonical pathway gene Vangl2 (Fig. 7F). Taken together, these results suggest that WNT7A signals through a noncanonical β-catenin-independent pathway in OB-NSCs.

Finally, to explore the effects of exogenous WNT7A on the early expansion of primary cells, we challenged freshly isolated cells from the adult OB tissue with the purified protein. Cultures were maintained for 10 days in growth medium with EGF and FGF-2, supplemented or not with WNT7A. (Fig. 7G). There was no effect on the frequency of sphere-forming cells in the primary culture (Fig. 7H), suggesting no initial role of WNT7A on cell survival or on recruitment of the cells into the cell cycle; a similar result was obtained when freshly isolated cells from the adult SVZ tissue were challenged with the purified protein (Supporting Information Fig. S8). However, WNT7A treatment significantly increased the number of secondary spheres (Fig. 7I). Thus, we conclude that WNT7A improved self-renewal of the primary cells directly isolated from the OB tissue, and we propose it may be instrumental in the establishment and expansion of adult brain NSC cultures, including those derived from the OB.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Defining the dynamics of stem cell niches in the adult mammalian brain is critically important not only to improve our understanding of neurogenesis during adulthood but also to explore the potential of endogenous NSCs for repair. In this study, we described the proliferative activity of a candidate stem cell population located in the OB, the most anterior structure of the rodent brain. We showed that candidate NSCs colabeled for GFAP and SOX2 are predominantly found in the iGCL and SEL of the OB and that at least a fraction of these cells is cycling. We further proved the existence of slowly dividing LRCs in the central OB region, and using a dual labeling procedure with different thymidine analogs, we demonstrated that OB-LRCs can undergo at least two rounds of cell division. Moreover, by comparing side-by-side the OB and the SVZ niche, we showed those candidate NSCs in the OB are more quiescent than SVZ-NSCs in vivo.

The existence of resident NSCs in the adult OB was already suggested a decade ago after the derivation of stem cell cultures from the adult rodent and human OB tissue [13–17]. A recent report using Nestin::CreERT2 mice identified putative NSCs clustered in ependymal-like rosettes located in the OB core [16], a type of structure that may be a remnant of the olfactory ventricle after completion of development. As in our study, some of the clustered cells expressed SOX2 and GFAP and appeared to survive antimitotic Ara-C treatment, suggesting they were mostly quiescent. However, the Nestin::CreERT2-positive cells contributed <10% of the LRCs in the OB core, indicating that the putative NSCs identified in the transgenic strain represented a subpopulation of the total OB stem cell pool. We did not address the presence of displaced ependymal rosettes in our study, yet we found numerous SOX2+GFAP+ cells throughout the iGCL and the SEL, suggesting that OB-NSCs are not confined to rosette-related structures. We also noticed that SOX2+GFAP+ cells in the iGCL-SEL coexpressed GLAST, a marker that labels embryonic radial glia and stem cells in the SVZ [6]. Taking advantage of a magnetic bead-based isolation, we added new insight to the cell type-of-origin of the primary neurospheres, revealing it belongs to the GLAST+ lineage, which comprises all OB cells forming spheres in vitro in our experimental conditions. In addition, our data show that antibodies raised against the transmembrane glycoprotein GLAST coupled to microbeads are convenient for the prospective purification of OB-NSCs.

We consistently observed that primary neurospheres derived from the OB were smaller compared to those generated from the SVZ, pointing to a reduced proliferative activity or a reduced responsiveness to mitogens of OB cells. We interpret the differences in the proliferation of OB and SVZ primary cultures as a recapitulation of the endogenous NSC activity in its corresponding structure of origin. This in vivo activity may depend on the complex interplay of region-specific signals found in the OB and SVZ niches. Among the niche cells that provide local cues to NSCs, we focused on the astrocytes, since an emerging view suggests they are regionally specified cells [19, 22–24]. In addition to astrocytes, ependymal cells lining the ventricles are also major players of the SVZ niche, presenting transmembrane ligands and releasing soluble factors that influence NSCs [35, 45, 46]. Thus, it is possible that the absence of an ependymal wall in the OB as opposed to the SVZ also contributes to the decreased in vivo activity of OB-NSCs relative to SVZ-NSCs.

We hypothesized that the lack of adequate stimulatory signals such as those derived from SVZ astrocytes and/or the presence of inhibitory molecules from OB astrocytes could partly underlie the differences found between OB and SVZ stem cell activity. Glial cells may communicate in vivo with adjacent NSCs by releasing diffusible ligands. We demonstrated that soluble factors from astrocytes are crucial in the regulation of self-renewal and proliferation of adult NSCs, since signals from OB astrocytes had a mild negative impact on neurospheres in vitro while the most remarkable effect was the prominent stem cell stimulation caused by SVZ astrocytes. Our global transcriptome analysis aimed at identifying the genes specifically overexpressed in SVZ astrocytes or in OB astrocytes that encode secreted proteins. The expression profile of OB astrocytes identified the WNT inhibitor Sfrp4 and the ligand Bmp7 as candidate genes that may block the proliferative activity of the adult OB-NSCs. We validated the overexpression of these genes in OB glia compared to SVZ glia. Interestingly, our preliminary data point to a negative impact of sFRPs and BMP7 on sphere formation (M. Moreno-Estellés and H. Mira, unpublished observations). The role of the OB-derived signals in the differential activity of OB-NSCs and SVZ-NSCs will be explored in future studies.

In this work, we chose to focus on the diffusible factors specifically secreted by SVZ astroglia, since the most relevant effect in the coculture and CM experiments was the marked NSC stimulation caused by SVZ astrocytes. The microarray data pointed out several relevant candidates overexpressed in SVZ astrocytes, and among them, we chose to focus on WNT7A given the widespread role of WNT signaling in the self-renewal of adult stem cells from various tissues and lineages. In the adult periventricular region, several components of the WNT signaling pathway, including Wnt7a, Fzd, and Tcf3, are expressed [47] and the function of Wnt7a in adult NSC niches is beginning to unfold. A marked reduction in BrdU incorporation in the SVZ and hippocampus of Wnt7a knockout mice has been reported, indicating that WNT7A is important for adult NSC activity in vivo [38]. Previous data also suggest that WNT7A has a role as a NSC autocrine signal that stimulates proliferation and self-renewal, acting downstream of the nuclear receptor TLX [38]. We now add a paracrine function for WNT7A, as part of the signaling cocktail released by niche glial cells. Another factor acting on NSCs, the atypical NOTCH ligand delta-like homolog 1 (Dlk1), is also expressed both in SVZ-NSCs and in SVZ astrocytes and was recently highlighted as a single gene that functions co-ordinately in stem cells and niche glial cells [48]. We found that, in addition to Wnt7a and Dlk1, Metrn is also overexpressed in SVZ-NSCs (M. Moreno-Estellés and H. Mira, unpublished observations) and in SVZ astrocyte cells (this study), emphasizing the existence of a wide molecular relationship between NSCs and the local glial microenvironment.

Canonical WNT signaling has been related to the regulation of SVZ-NSCs after progenitor cell ablation with Ara-C and following stroke [39]. Despite the fact that canonical WNT signaling was absent in most adult SVZ-NSCs under noninjury conditions, enhancing WNT signaling with a GSK3β inhibitor increased the number of stem cells without affecting progenitor proliferation. Contrary to these observations, another study using the transgenic WNT reporter Axin2-d2EGFP demonstrated that canonical WNT signaling in the SVZ is primarily detected in transient amplifying progenitors [49]. Selective GSK3β inhibition mainly increased the proliferative activity and the number of progenitors in the adult SVZ, and the same effect was reported upon the injection of a retroviral vector expressing a CA form of β-catenin. These contradicting in vivo data remain to be reconciled.

We used the neurosphere assay to study the effects of WNT7A and the GSK3β inhibitor LiCl on OB-NSC cultures, which is an excellent system to simultaneously check for self-renewal of stem cells and proliferation of progenitors in vitro. Our results demonstrate that WNT7A protein is sufficient to trigger NSC self-renewal in the presence of growth factors yet it does not act as a general mitogen that promotes proliferation; instead, WNT7A controls the symmetric stem cell fate of the two daughter cells. In contrast, LiCl has an effect on proliferation of stem/progenitor cells. Consistently, WNT7A does not signal through β-catenin and even antagonizes this pathway at early time points, in line with published studies showing that noncanonical WNT signaling has the ability to represses canonical signaling [50]. Indeed, we found that WNT7A stimulation increased the expression of Vangl2, a component of the noncanonical planar cell polarity pathway. Interestingly, WNT7A in the muscle markedly stimulates the symmetric expansion of satellite stem cells acting through Vangl2 but does not affect the growth of myoblasts [44]. We speculate that differences in WNT7A responsiveness between stem and progenitor cells (i.e., differential expression of Fzd receptors or other pathway components) may underlie the discrepancy observed between the cellular response to WNT7A stimulation and GSK3β inhibition. Future studies will aim at defining the receptor(s) activated by WNT7A and at better understanding the molecular components of the cascade triggered by WNT7A, which will be fundamental for exploring the interplay with other relevant pathways already related to NSC self-renewal, such as NOTCH and EGFR [33].

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

In sum, our study confirms the existence of an endogenous stem cell population in the OB. The activity of OB-NSCs is hampered by the lack of stimulatory signals. Our findings also identify glial-derived WNT7A as a potential tool to enhance the self-renewal ability of OB-NSCs by promoting symmetry during stem cell division. Thus, lessons from glial signals may improve our ability to manipulate NSCs for therapeutic applications.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

We would like to thank Marçal Vilar for assistance with Western blots; Fernando Gonzalez for confocal microscope support; Edurne García and Raquel Pérez for technical assistance; Sergio Noriega for advice and help in graphic design; and Aixa Morales for providing reagents. This work was supported by grants from Ministerio de Sanidad y Consumo (Fondo de Investigación Sanitaria [FIS]-PI06/0754 and PI09/2254) to H.M.; from the Ministerio de Ciencia e Innovación (SAF program, CIBERNED, RETIC Tercel) and Generalitat Valenciana (Programa Prometeo) to I.F.; M.M.E. was supported by a PFIS fellowship from Ministerio de Sanidad, Política Social e Igualdad; P.G.G. was supported by a “Sara Borrell” Postdoctoral fellowship from Ministerio de Ciencia e Innovación (MICINN)-FIS; M.D.M. was a recipient of a predoctoral fellowship FPU from Ministerio de Educación.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
sc-12-0303_sm_SupplFigure1.tif17348KSupplementary Figure 1: In vivo distribution of SVZ-NSCs co-expressing GFAP and SOX2. Coronal section through the SVZ from 2-m old wild-type (CD-1) mice, stained for GFAP (green) and SOX2 (red), as seen by confocal microscopy. Nuclear label DAPI is shown in blue. The dashed line delineates the collapsed ventricular cavity. Sp, septum. St, striatum. Scale bar: 50 μm
sc-12-0303_sm_SupplFigure2.tif36597KSupplementary Figure 2: Identification of CldU+ and IdU+ cells in the olfactory bulb. (A) Coronal section inmunostained for CldU (green) and NeuN (red). CIdU+ cells in the oGCL colocalize with the nuclear mature neuronal marker NeuN. Nuclear label DAPI is shown in blue. Scale bar: 20 μm. (B) Coronal section inmunostained for IdU (red) and DCX (magenta). Most IdU+ cells located in the SEL (or very close to it) co-localize with the neuroblast marker DCX. Nuclear label DAPI is shown in blue. Scale bar: 50 μm
sc-12-0303_sm_SupplFigure3.tif2081KSupplementary Figure 3: The OB cells with spherogenic capacity are GLAST positive. (A) Percentage of GFAP+SOX2+ cells that are also GLAST+ in the SEL-iGCL of 2-m old wild-type (CD-1) mice (average ± sem, n = 4 animals), as determined by confocal microscopy. (B) After magnetic separation, positive and negative GLAST (ACSA-1) fractions derived from the OB tissue were stained with secondary Alexa 488 antibody and analyzed by flow cytometry. The result demonstrates that the GLAST+ fraction was highly enriched in cells expressing high levels of GLAST. (C) Summary table showing data of three independent experiments performed. The GLAST negative fraction did not form neurospheres under the experimental conditions that were used. The sphere formation efficiency of the GLAST positive fraction was 0.26 ± 0.05 % (average ± sem, n = 3 independent experiments).
sc-12-0303_sm_SupplFigure4.tif8734KSupplementary Figure 4: Characterisation of the glial cultures. Cultures from postnatal mice were highly enriched in Glial Fibrillary Acidic Protein (GFAP). Most cells displayed a protoplasmic morphology and were negative for the type C polyganglioside marker (A2B5), indicating they were type 1 astrocytes. Immunocytochemistry for GFAP is shown in red and for A2B5 is shown in green. Nuclear label DAPI is shown in blue. (A) Type 1 astrocyte, protoplasmic, GFAP positive and A2B5 negative. (B) Type 2 astrocyte, fibrous, GFAP positive and A2B5 positive. Scale bar: 20 μm
sc-12-0303_sm_SupplFigure5.tif2556KSupplementary Figure 5: Diffusible factors from adult OB and SVZ astrocytes regulate adult OB-NSC activity, while SVZ-NSCs are also regulated by astrocytic soluble factors. (A) Schematic drawing illustrating the indirect co-culture system. OB-NSCs were seeded in the upper transwell, while adult astrocytes were grown as feeder cells in the lower compartment. (B) Number of OB neurospheres obtained by plating equal numbers of cells dissociated from OB-NSC cultures in the presence of adult astrocytes from different brain areas or in the absence of astrocytes as a control (n = 3). (C) Average size of the OB neurospheres obtained in the adult glial co-culture (n = 3). (D) Schematic drawing illustrating the experimental setup for the conditioned medium assays. Adult SVZ-NSCs were treated with filtered serum-free medium from region-specific astrocytic cultures that had been grown for 3 days in vitro (DIV). The conditioned medium was supplemented with EGF and FGF-2. (E) Number of SVZ- neurospheres obtained by plating equal numbers of cells dissociated from SVZ-NSC cultures in conditioned medium from the astrocytes (n = 3). (F) Average size of the SVZ neurospheres generated in the conditioned medium (n = 3). Paired t-test: *p < 0.05, ** p < 0.01.
sc-12-0303_sm_SupplFigure6.tif6102KSupplementary Figure 6: OB-NSCs transcriptome expression analysis. Scanned images of the GEArray S Series Mouse Stem Cell Gene Arrays MM-601.2 membranes hybridised with biotinylated-dUTP cDNA probes prepared from RNA of OB-NSC cultures (n = 2). The results from the analysis are shown in Supplementary Table 5.
sc-12-0303_sm_SupplFigure7.tif4562KSupplementary Figure 7: OB-NSCs express Wnt signalling components. PCR validation for several Fzd genes, which code for Wnt receptors (A) and other essential components of both canonical and non-canonical WNT signalling (B) in three independent OB-neurosphere cultures.
sc-12-0303_sm_SupplFigure8.tif1519KSupplementary Figure 8: WNT7A increases the number of secondary SVZ-neurospheres. (A) Primary SVZ-neurosphere yield from 2-m old CD-1 mice untreated and treated with WNT7A 50ng/ml (n = 12). (B) Fold increase in the number of secondary SVZ-neurospheres grown in the presence of WNT7A 50 ng/ml (n = 7). (C) Self-renewal assay in which neurospheres formed in growth medium with or without WNT7A were dissociated to single cells and plated at low density in the absence of WNT7A. The WNT7A pre-treatment improved the self-renewal capacity of SVZNSCs. More secondary spheres were formed from the pre-treated cells (n = 2). Paired t-test: ** p < 0.01.
sc-12-0303_sm_SupplTable1.pdf28KSupplementary Table 1
sc-12-0303_sm_SupplTable2.pdf40KSupplementary Table 2
sc-12-0303_sm_SupplTable3.pdf20KSupplementary Table 3
sc-12-0303_sm_SupplTable4.pdf34KSupplementary Table 4
sc-12-0303_sm_SupplTable5.pdf30KSupplementary Table 5

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.