SEARCH

SEARCH BY CITATION

Keywords:

  • astrocytes;
  • cilium;
  • EGFR;
  • PDGFRα;
  • mitosis;
  • neurogenesis;
  • electron microscopy

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Neural stem cells (NSCs) in the subventricular zone (SVZ) continuously generate olfactory bulb interneurons in the adult rodent brain. Based on their ultrastructural and antigenic properties, NSCs, transient amplifying precursor cells, and neuroblasts (B, C, and A cells, respectively) have been distinguished in mouse SVZ. Here, we aimed to identify these cell types in rat SVZ ultrastructurally and at the light microscopy level, and to determine the antigenic properties of each cell type using gold and fluorescence immunolabeling. We found astrocytes with single cilia (NSCs, correspond to B cells) and neuroblasts (A cells). We also observed mitotic cells, ependymal cells, displaced ependymal cells, and mature astrocytes. In contrast, transient amplifying precursor cells (C cells) were not detected. The NSCs and neuroblasts had epidermal growth factor receptor (EGFR) and platelet-derived growth factor receptor alpha (PDGFRα) expressed on the ciliary apparatus and were the only cell types incorporating the proliferation marker BrdU. Throughout mitosis, EGFR and PDGFRα were associated with the microtubule of the mitotic spindle. Ependymal and displaced ependymal cells also expressed EGFR and PDGFRα on their cilia but did not incorporate BrdU. Our findings indicate that the NSCs in adult rat SVZ give rise directly to neuroblasts. During mitosis, the NSCs disassemble the primary cilium and symmetrically distribute EGFR and PDGFRα among their progeny. © 2008 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The adult subventricular zone (SVZ) contains neural stem cells (NSCs) (Reynolds and Weiss, 1992), which give rise to neurons (Lois et al., 1996) and oligodendrocytes (Menn et al., 2006). In the intact rodent and human brain, neuroblasts migrate via the rostral migratory stream (RMS) to the olfactory bulb, acquiring the phenotype of mature interneurons (Curtis et al., 2007; Lois et al., 1996). Following ischemic injury, there is a long-lasting migration of neuroblasts from the rodent SVZ into the damaged striatum, where they differentiate into mature striatal neurons (Arvidsson et al., 2002; Parent et al., 2002; Thored et al., 2006). Stroke-induced neurogenesis has also been reported in humans (Jin et al., 2006; Minger et al., 2007). The regulation and functional significance of this potential self-repair mechanism are still poorly understood (Lindvall and Kokaia, 2008), and it is virtually unknown how injury affects the cellular composition of the SVZ. Such studies will necessitate detailed morphological and immunohistochemical characterization of the various stem cell-related cell populations in the intact brain.

The ultrastructural composition and antigenic properties of NSCs and their progeny in SVZ have been described in birds (Alvarez-Buylla et al., 1998), mice (Doetsch et al., 1997), humans (Quinones-Hinojosa et al., 2006), and rabbits (Ponti et al., 2006). In adult canaries, Type-B cells (i.e., the NSCs) with single nonmotile cilia, Type-A cells (neuroblasts), and ependymal cells were identified and characterized using electron microscopy and serial section reconstruction. Type-B cells incorporated radioactive thymidine, underwent interkinetic migration and directly gave rise to neuroblasts (Alvarez-Buylla et al., 1998). In mice, the ultrastructural characteristics of the main cell types were complemented by description of their antigenic repertoire (Doetsch et al., 1997). The B cells, which expressed glial fibrillary acidic protein (GFAP), vimentin, and platelet-derived growth factor receptor alpha (PDGFRα), were shown to be the NSCs and to generate transient amplifying precursor cells. These so-called C cells, which were not detected in birds (Alvarez-Buylla et al., 1998), stained positively for epidermal growth factor receptor (EGFR) and Dlx2 and gave rise to doublecortin (Dcx)/PSA-NCAM-positive neuroblasts (A cells) (Doetsch et al., 1997). Infusion of EGF and PDGF increased proliferation of NSCs and arrested neuroblast production in SVZ (Doetsch et al., 2002; Jackson et al., 2006).

In both birds and mice, the NSCs have a single cilium coupled to the centriole (Alvarez-Buylla et al., 1998; Doetsch et al., 1999b). Disassembling of the primary cilium and duplication of the existing centriole represent the initial stages of mitosis in vertebrates, which suggests a pivotal role of these organelles in cell cycle regulation (Pugacheva et al., 2007). Indeed, “cilium–centriole–centrosome” cycle during mitosis has been extensively documented (Bernhard and De Harven, 1956; Rieder et al., 1979; Sorokin, 1962; Tucker et al., 1979), but the molecular mechanisms are poorly understood (Michaud and Yoder, 2006; Pan and Snell, 2007). PDGFRα-dependent regulation of the cell cycle occurs through the primary cilium in fibroblasts, where downregulation of PDGFRα triggers proliferation (Schneider et al., 2005; Tucker et al., 1979). Whether EGFR and PDGFR are expressed on the primary cilium in NSCs and their progeny and involved in stem cell proliferation is not known. Notably, multiciliated ependymal cells have been suggested to function as stem cells in SVZ (Coskun et al., 2008; Johansson et al., 1999), but their receptor repertoire is poorly known.

The human SVZ is anatomically different from that in mice. It is organized in four layers, where the ependymal cells are separated from an astrocytic ribbon by a hypocellular gap (Sanai et al., 2004). Displaced ependymal cells, three types of GFAP-positive astrocytes, and paucity of neuroblasts are the key features (Quinones-Hinojosa et al., 2006). Notably, C cells have not been identified.

Rats are the most commonly used species in studies of adult neurogenesis in pathological conditions, e.g., in models of stroke (Arvidsson et al., 2002), experimental autoimmune encephalomyelitis (Danilov et al., 2006), Huntington's disease (Tattersfield et al., 2004), and epilepsy (Bengzon et al., 1997; Parent et al., 1997). It has been assumed that the ultrastructural and antigenic properties of NSCs and their progeny in rat SVZ are identical to those of mouse SVZ, but these issues have not been investigated. The SVZ in rats was initially described as containing ependymal cells and “the cells of subependymal plate” (Blakemore, 1969; Privat and Leblond, 1972), which consist of neuroblasts (corresponding to A cells), migrating in chains inside glial tubes toward RMS (Peretto et al., 1997; Peretto et al., 1999; Peretto et al., 2005; Privat and Leblond, 1972). Mitotic cells (Altman, 1963) and cells expressing both EGFR and p75 neurotrophin receptors are also present in rat SVZ (Giuliani et al., 2004; Hoglinger et al., 2004; Young et al., 2007).

The objectives of this study were twofold. First, to identify and ultrastructurally characterize B, C, and A cells in the rat SVZ. Second, to investigate the antigenic properties of NSCs and their progeny with special attention to the cellular and subcellular distribution of EGFR and PDGFRα. We identified B cells (astrocytes with single cilia), mitotic cells, and A cells, whereas C cells were not distinguished. The NSCs and neuroblasts expressed EGFR and PDGFRα on the ciliary apparatus and incorporated BrdU. In mitotic cells, EGFR and PDGFRα were associated with the mitotic spindle. Our data indicate that in adult rat SVZ during mitosis, NSCs disassemble primary cilium, symmetrically distribute EGFR and PDGFRα in the progeny, and directly give rise to neuroblasts.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Animals and BrdU Injections

Male Wistar rats (280–320 g; B&K Universal, Sollentuna, Sweden) were housed under 12 h light/12 h dark cycle with ad libitum access to food and water. Experimental procedures followed guidelines established by the Malmö-Lund Ethics Committee. Rats received injections with BrdU (50 mg/kg i.p.) either twice daily for 2 weeks or three times with 2-h interval and were perfused 2 h thereafter (Table 1).

Table 1. Experimental Design
 Electron microscopyLight microscopy
Ultrastructural analysis and cell countingImmunogold labeling and quantification of GPAnalysis of BrdU-incorporating cellsCell identity and analysis of proliferating cells
  1. For ultrastructural analysis, each 2 μm thick section was followed by six ultrathin sections and constituted one level of observation. For immunogold labeling, each 2 μm thick section was followed by three ultrathin sections and constituted one level of observation. Light microscopy was performed on 30 μm coronal sections. Abbreviations: PFA, paraformaldehyde; GA, glutaraldehyde.

Tissue fixation and embedding2% PFA, 1.5% GA “Epon”2% PFA, 0.5% GA “Lewicryl”2% PFA, 0.5% GA “Lewicryl”4% PFA “Epon”4% PFA
Number of rats11113
Number of levels/sections525234
Number of ultrathin sections3075615

Immunohistochemistry

Animals were transcardially perfused with ice-cold saline followed by 4% paraformaldehyde (PFA). Brains were removed, postfixed in 4% PFA overnight, cryo-protected in 30% sucrose, and cut on a freezing microtome in 30 μm coronal sections. Sections were stored in cryo-protective solution at −70°C. Four coronal sections from each rat were analyzed. Immunolabeling was performed on free-floating sections according to common protocols. Titration of optimal primary antibody dilution accompanied with antigen retrieval protocols using two different pH was performed to achieve optimal labeling. Briefly, sections were washed three times for 10 min with potassium phosphate-buffered saline solution (KPBS). Antibodies were: rabbit polyclonal anti-GFAP (1:500, DAKO), mouse monoclonal anti-vimentin (1:500, DAKO), rabbit polyclonal anti-PDGFRα (1:200, Santa Cruz), mouse monoclonal anti-PDGFRα (1:100, Abcam), sheep polyclonal anti-EGFR (1:100, Upstate), goat polyclonal anti-doublecortin (Dcx) (1:200, Santa Cruz), mouse monoclonal anti-PSA-NCAM (1:500, Chemicon), rabbit polyclonal anti-human p75 (1:200, Promega), rabbit polyclonal anti-Ki67 (1:50, Abcam), rabbit polyclonal anti-Dlx2 (1:100, Abcam), and mouse monoclonal anti-nestin (Chemicon). For PDGFRα staining, sections were incubated in 10 mM citric acid solution (pH 9.0) in a water bath at +80°C for 20 min to retrieve the antigens. Sections were then washed in KPBS and incubated with freshly made blocking solution, containing 5% serum and 0.25% Triton in KPBS. Sections were incubated overnight at +4°C with primary antibodies, diluted in blocking solution. Cy3 or biotinylated Alexa 488 was used for visualization of specific labeling. Omission of primary antibodies, or incubation with appropriate blocking peptides (for anti-EGFR and anti-PDGFRα antibodies) resulted in absence of staining. Co-labeling was validated in a laser-scanning confocal microscope (Leica TSP). Images were scanned at 1024 × 1024 resolution and assembled using Corel Paint Shop Pro X software. The depth of penetration of each antibody in the specimen was verified by examining 1-μm thick optical sections. Only optical layers in which penetration was sufficient for both antibodies were analyzed.

Electron Microscopy

For ultrastructural examination, rats were perfused with 2% PFA and 1.5% glutaraldehyde, and brains were cut coronally in 1 mm slices using a brain matrix. A slice was taken +0.2 mm from bregma, and part of the lateral wall of the SVZ (Fig. 1A, insert) was dissected under the microscope. The specimen was postfixed in perfusion solution, washed in phosphate buffer, dehydrated in alcohol, postfixed with osmium tetroxide, and embedded in Epon. Semithin sections were cut with Leica ultramicrotome, stained with Toloidine blue, and the correct orientation of the specimen was verified. Each 2-μm thick section was followed by six ultrathin sections and constituted one level of observation. Along the length of the SVZ, 500 μm were available for ultrastructural analysis. Five levels from one animal from the area depicted in Fig. 1A were analyzed. Ultrathin sections (70 nm) were cut with a diamond knife, collected on grids, postcontrasted with uranyl acetate and examined using a CM10 electron microscope (Philips). For ultrastructural analysis of individual cells, each cell was photographed on at least three magnifications (900×, 5200×, and 8900×), and more than 50 cells of each type were analyzed. Photographs were developed, negatives were scanned manually at 1200 or 2400 dpi and images were stored in digital form. Images were analyzed using Corel Paint Shop Pro X software. The main ultrastructural components in each cell were distinguished (Ghadially, 1982) and the main cell types were identified (Peters et al., 1991). Cells were counted using “Image J” software in SVZ photographed at low magnification (900×). Cell profiles in each ultrathin section within one level were quantified, and average numbers were considered to be representative for a 2 μm section. The total number of cells from five levels was then summed up, and the percentage of different cell types was calculated (Table 2).

thumbnail image

Figure 1. Ultrastructural and antigenic properties of astrocytes with single cilia. A: Analyzed part of SVZ as depicted in a coronal section (+2.0 mm from bregma) B: Cell with immature astrocytic morphology (arrow) contacting ependymal cells (e) and a neuroblast (n). C: Same cell as in B. Note paucity of organelles and lack of intermediate filaments. D: Astrocyte with single cilium residing between two ependymal cells (e). E: Basal body and cilia root from the same cell. EGFR-linked GP are localized on the cilia (arrowheads). F: Astrocyte with single cilium located inside the SVZ (arrow) neighbouring two mature astrocytes (as). G: Cytoplasm of cell in E. EGFR-linked GP (arrowheads) are associated with single cilium (arrow). H: Centriole and primary cilium are localized in the process (arrows) of astrocyte with single cilia, located on the striatal edge of SVZ. I: Centriole and primary cilium from the same cell. The cell contacts displaced ependymal cell via tight junctions (arrows). EGFR-linked GP associated with folding of the cell membrane (arrowheads). J: PDGFRα-linked GP associated with primary cilium and centriole in astrocyte, located inside the SVZ (highlighted area, arrowheads). K: Astrocyte with single cilium with 9 + 0 structure (arrow) contacting the ventricle. L: PDGFRα-linked GP localized on the cilium (arrowheads). M: Subcellular distribution of EGFR in the reticulated nucleoli of astrocytes (arrow). N: PDGFRα-linked GP associated with fibrous centres in the reticulated nucleolus (arrowheads and insert). O: Nucleus of SVZ astrocyte containing reticulated nucleolus (arrow). P: EGFR-linked GP localized on fibrous centres (arrowheads and insert). Scale bars: A, 10 μm, B, 3 μm, C, 5 μm, D, 2 μm, E, 200 nm, F, 2 μm, G, 200 nm, H, 1 μm, I, 250 nm, J, 500 nm, K, 2 μm, L, 500 nm, M, 2 μm, N, 200 nm, O, 1 μm, P, 250 nm.

Download figure to PowerPoint

Table 2. Antigenic Properties of Stem Cell-Related Cell Types in Rat Subventricular Zone
inline image

For ultrastructural identification of BrdU-positive cells in SVZ, sections from a rat, which had received BrdU for 2 weeks were stained with anti-BrdU antibodies using DAB-based protocol. Briefly, sections were rinsed in KPBS, quenched in 3% H2O2 (1:10) and 10% methanol (1:10) in KPBS (8:10) solution, rinsed in KPBS, incubated in 1 M HCl for 10 min at +65°C, incubated in blocking solution, and then stained with 1:100 rat anti-BrdU antibodies (Harlan-SeraLab) overnight at +4°C. Staining was visualized with standard ABC technique. Part of the SVZ was dissected, dehydrated, and embedded in Epon. Ultrathin sections (70 nm) were cut with a diamond knife and collected on grids. For better visualization of DAB-positive cells, poststaining with uranyl acetate was omitted. Sections were examined with CM10 electron microscope under lower voltage (40 kv). The pattern of DAB labeling on ultrathin sections was compared with control sections, where the specimen was processed in a similar manner, but no DAB labeling was performed.

Gold Immunolabeling

Intact rats and the rat which received three BrdU injections were perfused with 2% PFA and 0.5% glutaraldehyde, and the SVZ was dissected, postfixed in the perfusion solution, washed in phosphate buffer, processed via lower temperature dehydration technique, and embedded in Lewicryl resin (Armbruster et al., 1982). This procedure preserves the antigenic structure of the specimens and makes possible the detection of several antigens present on the surface of the ultrathin sections. Secondary antibodies are coupled to gold particles, which are 10 or 15 nm in diameter. Antibodies were the same as used for fluorescence immunolabeling (see earlier). Grids with ultrathin sections were washed in distilled water, blocked with 0.5% bovine serum albumin (BSA) for 30 min, incubated with primary antibodies at +4°C overnight, washed in distilled water, incubated with secondary antibodies for 30 min, washed in distilled water, and poststained with lead citrate. Identification of gold-labeled structures was performed at 6,600× (or higher) magnification. To investigate the subcellular antigen distribution, density of golden particles (GP) per 1 μm2 was quantified using ImageJ software. The GP were counted manually on digital images in ultrastructurally clearly defined structures, such as cytoplasm, nucleus, reticulated nucleolus, cilium, and centriole on a cross cut, and microtubules of the mitotic spindle, which were cut longitudinally. The brightness and contrast were adjusted when necessary. Quantification of GP in at least five cells of each type was used for statistical analysis. Background levels were calculated by counting the average density of GP in three section-free areas for each antigen. Comparisons were made after subtraction of background values. Each 2-μm thick SVZ section was followed by three ultrathin sections and constituted one level of observation. To verify the pattern of gold immunolabeling, each cell was examined on several neighboring ultrathin sections. In total 25 levels from the area depicted in Fig. 1A were analyzed. Table 1 summarizes experimental design.

Fibroblast Cultures

To validate the specificity of the PDGFRα antibody, we used fibroblast cultures (Bostrom et al., 1996). NIH 3T3 murine fibroblast cells were cultured in dulbeccos minimal essential medium (DMEM) supplemented with 10% fetal calf serum (FCS), 2 mM L-glutamine, 100 U/mL Penicillin, and 100 mg/mL Streptomycin (GIBCO). Cells were grown on chamber slides (Nunc), fixed with 4% PFA for 15 min, washed, and incubated in blocking solution, containing 0.025% TritonX-100 and 5% goat serum in KPBS. Slides were stained with primary antibodies against PDGFRα (1:400, Santa Cruz) and Ki67 (1:100, Nova Castro). Cell nuclei were counterstained using Hoechst (Molecular Probes).

Statistics

Nonparametric ANOVA with Kruskal-Wallis post hoc test or Mann-Whitney U test was used for comparisons of GP densities in different subcellular compartments (GraphPad Prism 5.0 software). Values are mean ± SD. Differences were considered significant at P < 0.05. For the exact P-values, see respective figures.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Astrocytes with Single Cilia

The rat SVZ is classically described as containing several layers of “light and dark cells” (Blakemore, 1969), where astrocytes are found among the pool of light cells (Fig. 1A, insert). Because astrocytes with single cilia (B cells) were identified as NSCs in birds and mice (Alvarez-Buylla et al., 1998; Doetsch et al., 1999b), we investigated the ultrastructural and antigenic properties of this cell type in rat SVZ. We found astrocytes with single cilia between or adjacent to ependymal cells (Fig. 1B–D,K) or inside the SVZ (Figs. 1F–I, 7). Some of these cells had a paucity of cell organelles and intermediate filaments, (Fig. 1B,C) whereas others were more mature (Fig. 1D,H). The cells had irregular contour, light cytoplasm, light ovoid nuclei, sometimes with one or two nucleoli, dispersed chromatin, sometimes with clumps, few free ribosomes, small rough endoplasmic reticulum, and Golgi complex. These astrocytes always had a centriole coupled to specialized cilium without rootlets, which contacted the ventricle (Fig. 1D,K) or was internalized (Fig. 1I,J). Membrane folding surrounded the root of the single cilium (Fig. 1I). Astrocytes with single cilia contacted ependymal cells or displaced ependymal cells via tight junctions (Fig. 1I).

PDGFRα and EGFR expression was not detected by light microscopy. At the ultrastructural level, although both PDGFRα- (Figs. 1K,L) and EGFR-linked GP (Fig. 1E, Table 2) were associated at high density with basal body and single cilium. The densities were comparable with those found on the cilia of ependymal cells (see later), and were higher on the centriole and primary cilium than in the nuclei and cytoplasm (see Fig. 5). The folding of the cell membrane surrounding the single cilium root also contained EGFR-linked GP (Fig. 1I). In the nuclei of interphase cells, which had reticulated nucleolus, PDGFRα-, and EGFR-linked GP were evenly distributed. The density of GP in the reticulated nucleolus was twice as high as in the whole nucleus for both EGFR and PDGFRα (Figs. 1M–P, 5). Notably, associations of EGFR-linked GP were found on the rough endoplasmic reticulum in some of the astrocytes with single cilium. No immunolabeling for the B cell marker vimentin (Doetsch et al., 1997) was detected in SVZ astrocytes either at the light or electron microscopical (EM) level.

Neuroblasts

The neuroblasts were found among the population of “dark cells” (Figs. 1A,C) (Blakemore, 1969) with elongated shape, located in clusters (Fig. 2A). Similar to what has been described in birds (Alvarez-Buylla et al., 1998), these cells had dark nucleus with clumps of chromatin surrounded by a thin rim of dark cytoplasm, which contained many ribosomes, few intermediate filaments but many microtubules mostly present in the processes (Fig. 2A–D). Primary cilium coupled to a centriole and small, rough endoplasmatic reticulum, and Golgi complex were also found (Fig. 2E,G). Neuroblasts formed tight junctions with each other inside the cluster and often directly contacted astrocytes with single cilia and ependymal cells (Fig. 2A). Neuroblasts localized close to the striatal border of the SVZ were in direct contact with nonmyelinated nerve terminals and myelinated nerves (not shown).

thumbnail image

Figure 2. Ultrastructural and antigenic properties of neuroblasts and mature protoplasmic astrocytes. A: Cluster of neuroblasts (n) with dark nucleus and dark cytoplasm, contacting ependymal cell (e). B: Cytoplasm of neuroblasts (boxed area from A) filled with ribosomes (red asterisk) and processes filled with microtubules (arrowheads). C: PDGFRα-linked GP localized on microtubules of neuroblasts (arrowheads). D: Part of processes from cell in C at higher magnification (15000×). E: PDGFRα-linked GP (arrowheads) associated with primary cilium (arrow) in neuroblast. F: Part of cytoplasm of neuroblast containing microtubules but no EGFR-linked GP. G: EGFR-linked GP (arrowhead) associated with neuroblast centriole. H: Co-expression of Dcx (Cy3, red) and PSA-NCAM (Alexa488, green) in all neuroblasts. I: No co-staining of PSA-NCAM-positive neuroblasts (green) with Dlx2 (red), but nuclei of all cells lining the ventricular wall are Dlx2-positive. J: Mature protoplasmic astrocyte (outlined black). Part of process with filament bundles from boxed area is shown in insert. K: Reticulated nucleoli (arrows) in the nucleus of mature astrocyte L: GFAP-positive “glial tubes” (confocal microscopy, average projection of 15 optical layers). M: Mature protoplasmic astrocyte (outlined black) on the striatal edge of SVZ. N: Process from the same cell containing intermediate filaments covered by GFAP-linked GP. O: Processes of mature astrocyte contacting endotheliocyte (En), located in the wall of a blood vessel. GFAP-linked GP (arrowheads) are localized on intermediate filaments. P: Density of GFAP-linked GP on bundles of intermediate filaments and rest of cytoplasm (values are means ± SD, n = 3). Q: Mature astrocyte containing two centrioles (arrows), one connected to primary cilium R: EGFR-linked GP (arrowheads) associated with primary cilium (arrow) from the cell in Q. S: PDGFRα-linked GP (arrowheads) localized on centriole and primary cilium of mature astrocyte. Scale bars: A, 3 μm; B, 500 nm; B,C, 500 nm; D, 250 nm; E, 100 nm; F,G, 250 nm; H, 55 μm; I, 22 μm; J, 5 μm, insert in J, 250 nm; K, 1 μm; L, 60 μm; M, 1 μm; N, 500 nm; O,Q, 1 μm; R,S, 250 nm.

Download figure to PowerPoint

EGFR and PDGFRα expression was only detected at EM level. PDGFRα-linked GP were preferentially located on microtubules in the neuroblast processes (Fig. 2C,D). PDGFRα-linked GP were also found on the primary cilium (Fig. 2E). EGFR-linked GP were not associated with microtubules (Fig. 2F) but with centriole and primary cilium (Fig. 2G). Dcx-linked GP were found exclusively in the cells fulfilling the ultrastructural criteria of neuroblasts (not shown). Moreover, at the light microscopy level, we found almost complete overlap between PSA-NCAM and Dcx immunolabeling (Fig. 2H, Table 2). Dlx2 expression was not detectable in the PSA-NCAM-positive neuroblasts (Fig. 2I). Occasionally, some of the Dcx-positive neuroblasts co-expressed the p75 receptor (not shown), although the level of p75 expression was lower when compared with that in EGFR-positive mitotic cells (see later).

Mature Protoplasmic Astrocytes

On the striatal edge of the SVZ, cells with the ultrastructural characteristics of mature protoplasmic astrocytes (Peters et al., 1991) were found (Fig. 2J). These cells had irregular contour, large light nucleus with dispersed chromatin, sometimes with reticulated nucleoli (Fig. 2K), light cytoplasm, and several thick processes, branching to the ventricular side and/or to the striatal edge of SVZ. The perinuclear cytoplasm and processes contained abundant intermediate filaments organized in bundles (Fig. 2J). The processes contacted ependymal cells, astrocytes with single cilia and neuroblasts, adjacent to the ependyma. Mature astrocytes formed feet-like processes, contacting blood vessels (Fig. 2O), and contained centriole and single nonmotile cilia (Fig. 2Q–S).

Strong GFAP immunoreactivity was observed in the SVZ, corresponding to “glial tubes,” created by the network of astrocytes (Fig. 2L) (Peretto et al., 1999, 2005). At the EM level, the cells intensely labeled with GFAP-linked GP fulfilled the ultrastructural characteristics of mature protoplasmic astrocytes (Fig. 2M,N). The highest density was observed in the processes, where GP followed the wave-like structure of intermediate filaments (Fig. 2N–P). The feet-like processes were stained similarly (Fig. 2O). Ultrastructurally, EGFR- (Fig. 2Q,R) and PDGFRα-linked GP (Fig. 2S) were associated with centriole and primary cilia in mature astrocytes (Table 2).

Ependymal and Displaced Ependymal Cells

Ependymal cells are found among the “light cells” in SVZ and constitute the major cell type contacting the ventricles (Fig. 3A). We found that the perinuclear cytoplasm of ependymal cells contained 10 nm thick intermediate filaments organized in bundles (Fig. 3J), on which both GFAP- and vimentin-linked GP were preferentially located (Fig. 3K,M, Table 2). The nuclei of some of the cells also had short bundles of intermediate filaments (Hirano and Zimmerman, 1967). Mitochondria, basal bodies, rootlets, and multiple cilia directed toward the ventricular cavity were present in the apical cytoplasm. Lamella anulatae, a stack of narrow membrane cisternae, was found in some of the cells (not shown). Ependymal cells formed tight junctions with each other and contacted neuroblasts (Fig. 3A) and astrocytes with single cilia (Fig. 1D). Tanycytes were rarely found between the ependymal cells (not shown).

thumbnail image

Figure 3. Ultrastructural and antigenic properties of ependymal and displaced ependymal cells. A: Ependymal cell (e) in direct contact with a cluster of neuroblasts (n), surrounded by mature protoplasmic astrocytes (as). B: PDGFRα (green) expressed in all ependymal cells. C: Ependymal cell on ultrathin section stained for PDGFRα. D: Part of apical cytoplasm from the same cell with PDGFRα expressed on basal bodies, cilia roots and cilia (arrowheads). E: Crosscut showing PDGRα-linked GP (arrowheads) on outer membrane and inside the cilia. F: Apical cytoplasm of ependymal cell from ultrathin section stained for EGFR. G,H: EGFR-linked GP (arrowheads) localized on basal bodies (G) and cilia (H) of ependymal cells. I: Co-expression of vimentin (green) and Dlx2 (red) in all cells lining ventricular surface (confocal microscopy). J–L: Ependymal cells with perinuclear bundles of intermediate filaments (arrows), stained with immunogold for GFAP (K) and vimentin (L) (arrowheads). M: Displaced ependymal cell (arrow) located on striatal edge of SVZ. N: Cell in M with morphology of multiciliated cells and contacting astrocyte with single cilium (as). O: Several basal bodies (arrows) in the cytoplasm of the cell in N, contacting astrocyte with single cilium (as in N) via tight junctions. Cytoplasm of astrocyte has centriole and internalized cilium (dashed arrows). EGFR-linked GP (arrowheads) are associated with basal bodies of displaced ependymal cell and with folding of astrocyte membrane. P: Apical cytoplasm of displaced ependymal cell containing lamella anulatae (arrow). Note the tight junction's area. Q: Part of cytoplasm from P with Lamella anulatae and three basal bodies, consisting of nine triplets of microtubules (arrows). EGFR-linked GP are associated with microtubule of the triplet (arrowheads). R: Displaced ependymal cell located at striatal edge (arrow). S: Apical cytoplasm of same cell containing several basal bodies and internalized cilia (arrows). T: PDGFRα-linked GP associated with basal bodies and part of the membrane of the same cell (arrowheads). Scale bars: A, 5 μm; B, 50 μm; C, 1 μm; D, 250 nm; E, 100 nm; F, 500 nm; G, 250 nm; H, 100 nm; I, 13.5 μm; J, 500 nm; K, 250 nm; L, 500 nm; M, 5 μm; N, 1 μm; O, 500 nm; P, 1 μm; Q, 250 nm; R, 5 μm; S, 2 μm; T, 150 nm.

Download figure to PowerPoint

All ependymal cells were PDGFRα-positive (Fig. 3B–E) with the highest density of PDGFRα-linked GP on the basal bodies, roots, and motile cilia (Fig. 3C–E). To validate the specificity of the staining pattern with this antibody, we used murine fibroblasts, which exhibit a classical example of PDGFRα signaling (Ohba et al., 1994; Schneider et al., 2005). Primary cilium, cytoplasm, and nucleoli of cultured fibroblasts were stained identically to what has been described previously (not shown).

The subcellular distribution of EGFR-linked GP resembled that of PDGFRα expression (Fig. 3F–H), being found in the apical cytoplasm, on the basal bodies (Fig. 3G) and on the cilia (Fig. 3H). However, the density of EGFR-linked GP was lower compared with that of PDGFRα-linked GP (Fig. 5D), consistent with the lack of detectable fluorescence at the light microscopical level. PDGFRα- and EGFR-linked GP were also detected in the reticulated nucleoli (not shown). The nuclei of all cells within the ependymal layer were Dlx2-positive both using fluorescence (Fig. 3I, red fluorescence) and immunogold labeling (not shown). All ependymal cells were also double-labeled with vimentin/Dlx2 antibodies (Fig. 3I, Table 2).

Cells with ultrastructural morphology identical to that of mature ependymal cells were also observed deeply inside the SVZ (Fig. 3M–O). These cells did not contact the ventricle. The nuclei of the displaced ependymal cells were located as far as 22.7 ± 2.2 μm (n = 3) from the ventricular border of the SVZ, whereas the nuclei of most ependymal cells were found at a distance of 4.7 ± 0.9 μm (n = 5). The displaced cells had light nuclei with chromatin aggregates, light cytoplasm, few mitochondria, intermediate filaments, no microtubules, several centrioles, and roots of internalized cilia (Fig. 3O–Q). They formed tight junctions with neighboring cells similar to those between ependymal cells (Fig. 3O). Their cytoplasm contained Lamella anulatae (Fig. 3P,Q). The displaced ependymal cells contacted blood vessels in an astrocyte-like fashion (Fig. 3M), and occasionally were in direct contact with striatal neurons (not shown). Displaced ependymal cells were always found in contact with SVZ astrocytes, which had single cilia (Fig. 3O). The subcellular distribution of the EGFR- and PDGFRα-linked GP in displaced ependymal cells was identical to that in normal ependymal cells (Table 2). EGFR- (Fig. 3O–Q) and PDGFRα-linked GP (Fig. 3S,T) were found on the cilia, cilia roots, and basal bodies.

Transient Amplifying Precursor Cells

Rapidly dividing precursor cells (C cells) represent about 10% of all cells in mouse SVZ (Doetsch et al., 1997). Ultrastructurally, these cells are characterized by deeply invaginated nucleus, reticulated nucleolus, chromatin clumps, and lack of intermediate filaments (Doetsch et al., 1997). We could not identify the corresponding cell type in the rat SVZ based on these criteria. Invaginations of the nucleus were rarely seen in SVZ astrocytes, neuroblasts, microglia, or ependymal cells. Cells with invaginated nuclei did not have a reticulated nucleolus. Clearly distinguishable reticulated nucleolus was found in astrocytes with single cilia, neuroblasts, ependymal cells, and microglia.

Mitotic Cells

Because transient amplifying precursor cells were not identified, we investigated the properties of mitotic cells, being the only cell type in the SVZ with ultrastructural characteristics of proliferating cells. Mitotic cells represented less than 0.5% of the total cell population and were located close to the ependymal layer or on the striatal edge of SVZ. These cells had round contour, dark or light cytoplasm, and variable ultrastructural morphology depending on the cell cycle phase. Mitotic cells did not contain bundles of intermediate filaments. As EGF and PDGF are potent inducers of stem cell proliferation in the SVZ (Doetsch et al., 2002; Jackson et al., 2006), we investigated the subcellular distribution of EGFR and PDGFRα during different phases of mitotis.

Cells in the early prophase had dark nucleus, completely covered with heterochromatin aggregates, disrupted nuclear envelope, and light cytoplasm. The cytoplasm of these cells did not contain microtubules and had very low density of EGFR- and PDGFRα-linked GP (Fig. 4A,E). During anaphase, the nuclear envelope disintegrated, chromatin became separated, and EGFR- and PDGFRα-linked GP were associated with the microtubules of the mitotic spindle (Fig. 4B,F respectively). Here, the density of GP in these cells was several 100-fold higher than in the rest of the cytoplasm (Fig. 5B). Chromatin in mitotic cells during telophase was completely separated in two daughter nuclei. Microtubules of mitotic spindle were directed toward the polar region outside the nucleus (Fig. 4C,D for EGFR and Fig. 4G,H for PDGFRα). The density of EGFR- and PDGFRα-linked GP on microtubules of mitotic spindle was still high and comparable with that found during anaphase. During cytokinesis, midbody was formed between mother and daughter cells. PDGFRα-linked GP were associated with midbody in a manner similar to that observed in nonmotile cilium of interphase astrocytes (Fig. 4I).

thumbnail image

Figure 4. Subcellular distribution of EGFR and PDGFRα in mitotic cells. A: Mitotic cell in early prophase. Chromatin is condensed, nuclear envelope is partly preserved but mitotic spindle is not yet formed. Note absence of EGFR-linked GP in cytoplasm. B: Part of cytoplasm of mitotic cells in anaphase. EGFR-linked GP are associated with microtubules of mitotic spindle (arrows) at high density (arrowheads). C: Mitotic cells in telophase. Chromatin becomes separated between daughter cells and nuclear envelope is formed. D: EGFR-linked GP (arrowheads) associated with remaining microtubuli of spindle (arrows). E: Absence of PDGFRα-linked GP in cytoplasm of mitotic cell in early prophase. F: Part of cytoplasm of mitotic cells in anaphase, with PDGFRα-linked GP localized on microtubules of mitotic spindle (arrows) at high density (arrowheads). G: Mitotic cells in early telophase. H: Part of perinuclear cytoplasm from cell in G. PDGFRα-linked GP (arrowheads) associated with disassembling spindle apparatus (arrows) and with mother centriole (dashed arrow). I: PDGFRα (arrowheads) localized on mid-body. Scale bars: A, 2 μm; B, 500 nm; C, 1 μm; D, 250 nm; E, 500 nm; F, 250 nm; G, 500 nm; H, 250 nm; I, 200 nm.

Download figure to PowerPoint

thumbnail image

Figure 5. Quantitative analysis of subcellular distribution of EGFR and PDGFRα in stem cell-related cell types in SVZ. The density of GP per 1 μm2 is plotted on left Y-axis for EGFR, and on a right Y-axis for PDGFRα in astrocytes with single cilia (A, n = 6), mitotic cells during anaphase (B, n = 4), neuroblasts (C, n = 4) and ependymal cells (D, n = 8). Values are presented as means ± SD and compared using nonparametric ANOVA with Kruskal Wallis post hoc comparison. Asterisk (*) denotes values significantly different from GP densities in cytoplasm and nuclei in each cell type. P < 0.05, **P < 0.01, ***P < 0.001.

Download figure to PowerPoint

Transient amplifying precursor cells, which express EGFR in mice, were not identified ultrastructurally in our experiments, whereas the highest density of EGFR-linked GP was found on the spindle apparatus in the mitotic cells (Fig. 5B). We therefore wanted to confirm the mitotic identity of EGFR-positive cells using fluorescence double-labeling. EGFR-positive cells in SVZ were found in a location similar to that of mitotic cells, namely on the striatal edge of the SVZ or close to the ependymal layer. EGFR-positive cells did not co-express markers for ependymal or displaced ependymal cells (vimentin, PDGFRα, Dlx2), neuroblasts (PSA-NCAM, Dcx) or mature astrocytes (GFAP) but were nestin-positive (Fig. 6A). All EGFR-positive cells expressed the proliferation marker Ki67 (Fig. 6B). The pattern of EGFR fluorescence staining was similar to that observed with immunogold labeling. Tubular structures, surrounding a Ki67-positive nucleus (Fig. 6B) were stained, indicating that the mitotic cells corresponded to the EGFR-positive cell population in rat SVZ (Table 2). A small fraction of the EGFR/Ki67-positive cells expressed the p75 neurotrophin receptor (Fig. 6C), and all p75-positive cells co-expressed Ki67 (Fig. 6D), which is in agreement with previous observations (Giuliani et al., 2004).

thumbnail image

Figure 6. Antigenic properties of mitotic and BrdU-incorporating cells in SVZ. A: EGFR-positive cells (green, arrows) co-expressing nestin (red) (Confocal microscopy, V, ventricular cavity). B: EGFR-positive cells (green) co-expressing Ki67 (red). Note tubular structures stained with anti-EGFR antibodies, resembling microtubule of mitotic spindle (dashed arrows), (average projection of 15 optical sections). C: EGFR-positive cells (green) co-expressing p75 receptor (red) (confocal microscopy, 1 μm optical section). D: p75-positive cells (green) co-expressing Ki67 (red) (average projection of 15 optical sections). E: PDGFRα-positive ependymal cells (green) do not incorporate BrdU (red), whereas PDGFRα-positive cell outside ependymal layer is BrdU-positive (arrowhead). F,G: BrdU-linked GP (arrowheads) present in nucleus of SVZ astrocyte (as) and neuroblast (n in G) after short-term BrdU administration. Note, that nucleus of mature astrocyte (as) on G is free from BrdU-linked GP. H: Nucleus of ependymal cell lacking BrdU-linked GP. I: Astrocyte contacting ventricle (*) incorporated BrdU, administered during 2 weeks and visualized by DAB staining (arrowheads). Nucleus of adjacent ependymal cell (e) does not contain DAB aggregates. J: Two astrocytes inside SVZ incorporated BrdU after 2 weeks of administration. DAB aggregates are clearly identified in the nuclei (arrowheads). K: Cytoplasm of one of the cells containing primary cilium and centriole (arrows). L: Nuclei of neuroblasts (n) and astrocytes (as) on striatal edge of SVZ containing BrdU. Nucleus of adjacent striatal neuron (*) exhibit no DAB deposits. Scale bars: A, 150 μm; B, 10 μm; C, 50 μm; D, 16 μm; E, 10 μm; F,G,H, 500 nm; I, 2 μm; J, 1 μm; K, 500 nm; L, 2 μm.

Download figure to PowerPoint

Taken together the ultrastructural and light microscopical data indicate that the mitotic cells constitute the EGFR-positive cell population in rat the SVZ.

Growth Factor Receptors in Stem Cell-Related Cell Types

We observed that at EM level, the subcellular distribution of PDGFRα and EGFR in NSCs and their progeny is broader than previously anticipated and appears to follow a specific pattern. To confirm that, we performed systematic quantitative analysis of the receptor-linked GP densities in subcellular compartments in stem cell-related cell types (see Fig. 5). In agreement with our qualitative observations (see earlier), we found that EGFR and PDGFRα were preferentially associated with centriole and cilia at highest density in astrocytes with single cilia (i.e. NSCs), neuroblasts, and ependymal cells, whereas in mitotic cells both receptors were found on the microtubules of the mitotic spindle (Fig. 5, Table 2). Moreover, in neuroblasts only PDGFRα and not EGFR was associated with microtubules (Fig. 2C,D,F).

Cells Incorporating BrdU

As nonmitotic and mitotic cells had similar distribution of EGFR and PDGFRα at the ultrastructural level (see Fig. 5), we explored which cells are capable of entering the cell cycle in the intact adult rat SVZ. We used ultrastructural analysis of BrdU-positive cells and performed fluorescence double labeling for BrdU and markers of the main cell types in the SVZ (Table 2).

First, we analyzed which cells incorporated BrdU following short-term administration. After three BrdU injections, performed with 2-h interval and animals killed 2 h thereafter, astrocytes with single cilia and neuroblasts were the only cell types incorporating BrdU, as revealed by gold immunolabeling (Figs. 6F–H, 7).

thumbnail image

Figure 7. Schematic drawing illustrating the cellular composition of the rat SVZ. Cells incorporating BrdU are labeled with white asterisk. Tanycytes, microglia, mature neurons and components of basal lamina are not depicted.

Download figure to PowerPoint

Because NSCs are considered to divide rarely, we then administered BrdU for 2 weeks to label sufficient number of NSCs and their progeny. Astrocytes with single cilia and neuroblasts were still the only cell types incorporating BrdU, as revealed by EM analysis of DAB-labeled cells (Figs. 6I–L, 7). Ependymal cells have been suggested to function as stem cells in the SVZ (Coskun et al., 2008; Johansson et al., 1999), but none of 400 examined cells incorporated BrdU (Fig. 6G,D). The BrdU-positive cells residing in the ependymal layer always had the ultrastructural morphology of astrocytes with single cilia (Fig. 6I).

Our light microscopy observations were consistent with the data obtained in the ultrastructural analysis. After 2 weeks of BrdU administration, neither vimentin- and PDGFRα-positive ependymal cells (Fig. 6E) nor GFAP-positive astrocytes (not shown) incorporated BrdU. PDGFRα/BrdU double-labeled cells were rarely observed outside the ependymal layer (Fig. 6E). These cells probably corresponded to displaced ependymal cells, because the highest density of PDGFRα-linked GP in the cells residing outside the ependymal layer was detected on the ciliary apparatus of displaced ependymal cells (see Fig. 3T). Neuroblasts also incorporated BrdU (not shown).

As NSCs are the cells which give rise to neuroblasts (Doetsch et al., 1999) and both of these cell types incorporated BrdU. Our ultrastructural and light microscopy data indicate that astrocytes with single cilia (NSCs, B cells) give rise directly to neuroblasts (A cells) via mitosis in rat SVZ, without passing through the stage of transient amplifying precursor cells (C cells).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Here, we have characterized the ultrastructural and antigenic properties of the stem cell-related cell types in SVZ of intact rat brain. We identified astrocytes with single cilia (corresponding to B cells in mouse SVZ), mitotic cells, and neuroblasts (A cells), whereas transient amplifying precursor cells (C cells) were not found. Ependymal cells, displaced ependymal cells and mature protoplasmic astrocytes were also identified. Astrocytes with single cilia (NSCs) and neuroblasts expressed EGFR and PDGFRα on the cillary apparatus and incorporated BrdU after both short- and long-term administration. Ependymal and displaced ependymal cells had similar subcellular distribution of both receptors. Ependymal cells did not incorporate BrdU. We show here for the first time that during mitosis NSCs disassemble primary cilia, symmetrically distribute EGFR, and PDGFRα in their progeny, and give rise directly to neuroblasts in adult rat SVZ.

The subcellular distribution of growth factor receptors was determined using the immunogold labeling technique (Armbruster et al., 1983; Carlemalm et al., 1982; D'Amico and Skarmoutsou, 2008; Ferreira et al., 2007; Mengual et al., 2008; Renno, 2001; Roth, 1982; Roth et al., 1978). Several lines of evidence confirmed the specificity of the EGFR and PDGFRα immunolabeling in our study. First, omitting the primary antibodies did not result in nonspecific labeling. Second, incubation with blocking peptides completely abolished the specific staining. Moreover, the reproducibility of the labeling in each case was high and allowed us to assess quantitatively the distribution of EGFR and PDGFRα in different subcellular compartments (Blackstad et al., 1990; Ottersen, 1989; Shupliakov et al., 1992; Storm-Mathisen et al., 1992). We also found that the PDGFRα expression pattern in fibroblasts in our experiment was identical to what has been reported previously (Bostrom et al., 1996; Ohba et al., 1994; Schneider et al., 2005). Third, in several cases, high density of GP observed in the electron microscope corresponded to bright fluorescent signal at the light microscopy level (Griffiths, 1996). However, in some cases, molecules were successfully detected at EM level, but no corresponding fluorescent signal was detected by light microscopy. This discrepancy may be due to insufficient penetration of the antibodies and/or low availability of the epitopes (Mathiisen et al., 2006).

Two types of astrocytes were identified in rat SVZ: astrocytes with single cilia and mature protoplasmic astrocytes. The astrocytes with single cilia corresponded to B cells, i.e., the NSCs identified in birds (Alvarez-Buylla et al., 1998) and mice (Doetsch et al., 1999a). In the rat SVZ, these cells had a paucity of intermediate filaments, low density of GFAP-linked GP and were vimentin negative. These characteristics might explain the difficulties in identifying NSCs in rat SVZ based on GFAP expression. Administration of BrdU for 2 weeks resulted in labeling of the NSCs as well as their progeny throughout the SVZ, but we found no GFAP/BrdU double-labeled cells. In contrast, it is well established that NSCs in mice contain GFAP because virally labeled GFAP-positive cells gave rise to olfactory bulb neurons (Doetsch et al., 1999a; Imura et al.; 2003; Morshead et al., 2003). Our data indicate that markers other than GFAP should be used for identification of NSCs in rat SVZ, also because high density of GFAP-linked GP and strong immunofluorescence were found in the processes of mature protoplasmic astrocytes (Blakemore, 1969; Peretto et al., 1997; Privat and Leblond, 1972), which contained thick filament bundles.

The response to EGF and PDGF is a key feature of single-ciliated B cells in mice (Doetsch et al., 2002; Morshead et al., 1994). We found that in the rat SVZ, EGFR and PDGFRα were expressed on the basal bodies and single cilia of these cells. Astrocytes with single cilia incorporated BrdU after long- and short-term administration, consistent with the idea that in these cells, EGFR and PDGFRα signaling might be involved in cell cycle regulation, as occurs in fibroblasts (Schneider et al., 2005). The presence of EGFR and PDGFR in the reticulated nucleoli may reflect how mitogenic signals are transmitted in nonmitotic NSCs (Baldin et al., 1990; Bouche et al., 1987), as direct activation of cell cycle- regulating genes by EGFR has recently been discovered (Lin et al., 2001).

Neuroblasts in rat SVZ (corresponding to A cells in mice) had ultrastructural characteristics resembling those of bird neuroblasts (Alvarez-Buylla et al., 1998), i.e., abundant microtubules and centriole coupled to single nonmotile cilium. In contrast to mouse SVZ, neuroblasts were frequently found in direct contact with myelinated and unmyelinated nerves or ependymal cells in rat SVZ. These findings raise the possibility that the properties of neuroblasts may be affected by signaling both from the ventricular cavity via ependymal cells and from the striatum. The antigenic properties of neuroblasts in rat SVZ were partly different from those in mice (Doetsch et al., 1997). EGFR and PDGFRα were localized on the centriole and primary cilium of neuroblasts, possibly mediating their retained proliferative capacity (Zhang et al., 2007). In accordance, neuroblasts incorporated BrdU after both short- and long-term administration. PDGFRα was also associated with microtubules in the processes, suggesting that neuroblasts in the adult rat SVZ recapitulate PDGFRα signaling similar to neuronal precursors during development (Andrae et al., 2001). Interestingly, a subpopulation of PSA-NCAM and Dcx-positive neuroblasts expressed the p75 receptor, consistent with previous observations that neurotrophin delivery to the SVZ can affect adult neurogenesis (Gascon et al., 2007; Kirschenbaum and Goldman, 1995).

In mice, EGFR-positive transient amplifying precursor cells (C cells) are the immediate precursors and source of neuroblast production (Doetsch et al., 1997, 1999b). The presence of EGFR-positive cells, co-expressing proliferation markers such as Ki67 and BrdU, has been reported also in rat SVZ but the ultrastructural properties of C cells have not been investigated, (Giuliani et al., 2004; Hoglinger et al., 2004; Young et al., 2007). In our experiments, all EGFR-positive cells co-expressed Ki67 in agreement with previous observations. To determine the ultrastructural properties of the EGFR-positive cells, we performed EM analysis of ultrathin sections stained for EGFR with immunogold. We could not ultrastructurally distinguish C cells as a separate population in rat SVZ, based on the criteria defined in mice (Doetsch et al., 1997), whereas all other cell types were identified. Mitotic cells were the only cell type, in which EGFR-linked GP were associated with microtubuli of the spindle apparatus at high density. In agreement, strong fluorescence was observed at light microscopy level, indicating that the mitotic cells constituted the EGFR-positive population in rat SVZ. Because NSCs (astrocytes with single cilia) and their progeny (neuroblasts) incorporated BrdU, and mitotic cells constituted the EGFR-positive cell population in rat SVZ, our data indicate that NSCs give rise directly to neuroblasts via mitosis. This is similar to what occurs in avian (Alvarez-Buylla et al., 1998) and human SVZ (Quinones-Hinojosa et al., 2006).

On the basis of these findings, we further explored the ultrastructural and antigenic properties of the mitotic cells in the rat SVZ. Mitotic cells were rare and had variable ultrastructure where reticulated nucleolus, primary cilium, and intermediate filaments (Pardal et al., 2007) were not present. As EGFR and PDGFRα were expressed in nonmitotic NSCs on their cilia and centrioles, which act as mitotic spindle pole organizers (Bernhard and De Harven, 1956; Rieder et al., 1979; Sorokin, 1962; Tucker et al., 1979), we investigated the subcellular localization of both receptors during different mitotic phases. We discovered that EGFR and PDGFRα were associated at high density with microtubules of the mitotic spindle.

EGFR and PDGFRα have traditionally been viewed as transmembrane receptors, which are phosphorylated and internalized upon activation (Mori et al., 1991; Sorkin et al., 1991). The subcellular trafficking of these particular receptors during mitosis is not known, although the phosphorylated proteins are abundant on the mitotic apparatus (Burry et al., 1992; Vandre and Burry, 1992). Hypothetically, EGFR and PDGFRα could interact with the mitotic apparatus in NSCs similarly to insulin-like growth factor binding protein-4, which associates with the microtubules of the mitotic spindle and centriole of astrocytes and mediates biological effects of insulin growth factor (Chesik et al., 2004, 2007).

In CNS development, cortical progenitor cells unequally distribute EGFR during asymmetric division (Sun et al., 2005). We found that both EGFR and PDGFRα were distributed symmetrically in mitotic cells. It is tempting to speculate that downregulation of PDGFRα or EGFR on the primary cilium might trigger NSC proliferation (Jackson et al., 2006; Kuhn et al., 1997), whereas association with the mitotic spindle ensures equal distribution of receptors between progeny. A similar mechanism has recently been discovered for R-Smad transcription factor, which is a key molecule mediating TGFβ effects (Bokel et al., 2006). Some of the EGFR-positive cells expressed p75, suggesting that neurotrophins may influence the differentiation of progenitor cells toward neuronal lineage already during mitosis (Giuliani et al., 2004; Young et al., 2007).

The ultrastructural properties of rat ependymal cells were similar to those described in other species (Alvarez-Buylla et al., 1998; Blakemore, 1969; Doetsch et al., 1997; Quinones-Hinojosa et al., 2006; Sanai et al., 2004). PDGFRα and EGFR receptors were expressed on the ciliary apparatus, consistent with previous findings that multiciliated cells can enter the cell cycle in vitro (Coskun et al., 2008; Gregg and Weiss, 2003; Johansson et al., 1999) or in vivo in response to injury (Danilov et al., 2006). The Dlx2 expression in ependymal cells further supported their retained potential to divide, as downregulation of Dlx2 occurs during mitosis in cortical progenitor cells (Petryniak et al., 2007) and in C cells in mouse SVZ (Doetsch et al., 2002). However, we found that ependymal cells did not incorporate BrdU in agreement with their postmitotic nature in the intact adult brain (Spassky et al., 2005).

Displaced ependymal cells have not previously been described in rat SVZ but have been found in humans (Quinones-Hinojosa et al., 2006). We found that these cells had ultrastructural and antigenic properties identical to normal ependymal cells. Notably, the cytoplasm of displaced ependymal cells contained lamellae annulate, which are present in NSCs in birds (Alvarez-Buylla et al., 1998) and in cells forming malignant tumours (Biernat et al., 2001; Jesionek-Kupnicka et al., 1999; Mirejovsky, 1990; Wang et al., 1992). Interestingly, PDGFRα-positive cells outside the ependyma, which might correspond to displaced ependymal cells, incorporated BrdU suggesting that they may enter the cell cycle, thus being an additional source of new cells.

We provide here the first comprehensive description of the ultrastructural and antigenic properties of NSCs and their progeny in the adult rat SVZ. Compared with the widely accepted cellular composition of the mouse SVZ, the rat SVZ exhibits several unique features, most importantly the lack of transient amplifying precursor cells and the direct conversion of NSCs to neuroblasts via mitosis. Our data provide the morphological basis for further studies on the cellular mechanisms involved in injury-induced neurogenesis in adult rat brain.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The authors are grateful to Lina Gefors for excellent technical assistance with electron microscopical preparations and to Ulrika Sparrhult-Björk for BrdU injections.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES