Histochemistry and Cell Biology

, 136:515

Characterization and identification of Sox2+ radial glia cells derived from rat embryonic cerebral cortex

Authors

  • Haoming Li
    • Department of Anatomy and CytoneurobiologyMedical College of Soochow University
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
    • Department of Anatomy and CytoneurobiologyMedical College of Soochow University
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
  • Jianbing Qin
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
  • Meiling Tian
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
  • Jinhong Shi
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
  • Weiwei Yang
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
  • Xuefeng Tan
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
  • Xinhua Zhang
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
  • Linqing Zou
    • Department of Anatomy and Neurobiology, The Jiangsu Key Laboratory of NeuroregenerationNantong University
Original Paper

DOI: 10.1007/s00418-011-0864-5

Cite this article as:
Li, H., Jin, G., Qin, J. et al. Histochem Cell Biol (2011) 136: 515. doi:10.1007/s00418-011-0864-5

Abstract

During the central nervous system (CNS) development, radial glia cells (RGCs) play at least two essential roles, they contribute to neuronal production and the subsequent guidance of neuronal migration, whereas its precise distribution and contribution to cerebral cortex remains less understood. In this research, we used Vimentin as an astroglial marker and Sox2 as a neural progenitor marker to identify and investigate RGCs in rat cerebral cortex at embryonic day (E) 16.5. We found that the Sox2+ progenitor cells localized in the germinal zone (GZ) of E16.5 cerebral cortex, ~95% Sox2+ cells co-localized with Vimentin+ or Nestin+ radial processes which extended to the pial surface across the cortical plate (CP). In vitro, we obtained RG-like cells from E16.5 cerebral cortex on adherent conditions, these Sox2+ Radial glia (RG)-like cells shared some properties with RGCs in vivo, and these Sox2+ RG-like cells could differentiate into astrocytes, oligodendrocytes and presented the radial glia—neuron lineage differentiation ability. Taken together, we identified and investigated some characterizations and properties of Sox2+ RGCs derived from E16.5 cerebral cortex, we suggested that the embryonic Sox2+ progenitor cells which located in the cortical GZ were mainly composed of Sox2+ RGCs, and the cortex-derived Sox2+ RG-like cells displayed the radial glia—neuron lineage differentiation ability as neuronal progenitors in vitro.

Keywords

Radial glia cellsCerebral cortexVimentinSox2GFAP

Introduction

Radial glia cells (RGCs) derive from neuroepithlium and exhibit a characteristic bipolar morphology, the longer radial processes of these cells connect the ventricular lumen with the external pial surface and support the migrating neurons into the final positions to become mature cells (Brunne et al. 2010; Goldman and Vaysse 1991; Schmechel and Rakic 1979; Voigt 1989). In addition to their role in radial migration, RGCs are putative neural stem (NS) cells in adult central nervous system (CNS), they are self-renewing and capable of differentiating into neurons (Anthony et al. 2004; Doetsch 2003a; Kriegstein and Alvarez-Buylla 2009; Noctor et al. 2001; Seri et al. 2004). RGCs exhibit neuroepithelial and astroglial properties, they express stem/progenitor markers such as the intermediate filament protein Nestin and pluripotent stem cells transcription factor Sox2, and also show several astroglial markers such as astrocyte specific glutamate transporter (GLAST), glial fibrillary acidic protein (GFAP), RC2, Vimentin, the Ca+ binding protein S100β and brain lipid binding protein (BLBP) (Anthony et al. 2004; Gubert et al. 2009; Kriegstein and Gotz 2003; Merkle et al. 2004; Suh et al. 2007). However, the expression of these markers in RGCs was variable and inconsistent during the CNS development (Brunne et al. 2010; Liu et al. 2010). GFAP was linked to astrocytes traditionally, recent studies demonstrated the postnatal neural progenitor cells also expressed GFAP (Liu et al. 2010; Seri et al. 2004). Human or rat-derived radial glia (RG)-like cells expressed GFAP in vitro, while the expression of GFAP in mouse-derived cells was negligible (Conti et al. 2005; Li et al. 2011). Several studies used a single astroglial marker to identified RGCs (Culican et al. 1990; Hartfuss et al. 2001; Hunter and Hatten 1995; Poluch and Juliano 2010; Sancho-Tello et al. 1995). However, morphological feature combined with stem/progenitor and astroglial markers are more accurate and reliable to identify RGCs (Conti et al. 2005; Gubert et al. 2009; Li et al. 2011; Mo et al. 2007; Sun et al. 2008; Zhang et al. 2010).

After the period of neuronal migration, RGCs retract their processes and differentiate into mature astrocytes. In the subventricular zone (SVZ), RGCs are abundant around birth and almost disappear during the first two postnatal weeks, which are paralleled by its transformation into mature astrocytes (Tramontin et al. 2003). Similarly, the RGCs confined to an astrocytic fate and differentiated into astroglial cells in the developing hippocampal subgranular zone (SGZ) (Brunne et al. 2010), and the astrogliogenesis ability of RG-like cells from neonatal hippocampus far exceeded the neurogenesis ability in vitro (Li et al. 2011).

In this research, we used Vimentin as an astroglial marker and Sox2 as a neural progenitor marker to investigate the distribution of RGCs in embryonic cerebral cortex at embryonic day (E) 16.5. Then, we obtained RG-like cells from E16.5 embryonic cortex on adherent conditions and identified some properties of these cells. Our findings suggested that the Sox2+ RGCs were the main type of neural progenitor cells in E16.5 cerebral cortex, the RG-like cells presented the radial glia—neuron lineage differentiation ability which was different from the neonatal hippocampus-derived RG-like cells in vitro (Li et al. 2011).

Materials and methods

Immunohistochemistry of embryonic or postnatal brain tissue sections

Animal experiments in this research were conducted according to protocols approved by the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals. Immunohistochemical analysis on embryonic and postnatal brain tissue sections were performed as described by Gregg et al (2003). Briefly, timed-pregnant (E16.5) and postnatal Sprague Dawley (SD) rats were killed, and the brains were removed from E16.5 embryos and postnatal (P) three rats. The brain tissues were fixed in 4% paraformaldehyde overnight at 4°C and then cryoprotected overnight in 20%, followed by 30% sucrose. Frozen brain sections were cut frontally at 40 μm thickness. Then coronal brain sections were rinsed in 0.01 M PBS and pre-incubated in blocking solution (10% normal goat serum and 1% BSA) for 2 h at room temperature. As primary antibodies, mouse anti-Vimentin (1:200; Abcam, Cambridge, UK), mouse anti-Nestin (1:200; Abcam), mouse anti-GFAP (1:500; Millipore; Billerica, MA, USA) or rabbit anti-GFAP (1:1,000; Abcam) and rabbit anti-Sox2 (1:1,000; Abcam) or mouse anti-Sox2 (1:200; Abcam) were incubated at 4°C overnight, then after multiple washes the sections were incubated with secondary antibodies conjugated to fluorescein 568 or FITC at room temperature for 2 h. Sections were counterstained by Hoechst (Sigma, St.Louis, MO, USA) for 15 min after immunolabeling. When the primary antibodies were omitted in immunofluorescent staining, no immunoreactivity was detected. All brain sections were photographed using an Olympus laser confocal microscope (FV10i, Olympus, Japan).

Protein isolation and Western blot analysis of GFAP at E16.5 and P3 cerebral cortex

For western blot analysis, protein homogenized and extracted from cerebral cortex at two time-points (E16.5 and P3; n = 3) using Mammalian Cell Lysis Kit (BBI, Markham, ON, Canada), and then protein concentration was determined using RC DC™ Protein Assay Kit (BioRad, Hercules, CA, USA). Equal amounts of protein resolved on SDS-polyasrylamide gel electrophoresis (PAGE) using 10% separation gel. Gels were transferred to a PVDF membrane using BioRad Semi-Dry Transfer Cell at 15 V for 30 min, and then blocked with 5% milk in Tris-buffered saline Tween (TBST) buffer and incubated with primary antibody mouse anti-GFAP (1:500; Millipore) and mouse anti-β-actin (1:3,000; Abcam) overnight at 4°C. After incubation with conjugated affinity-purified goat anti-mouse secondary antibody (1:5,000) labeled with IRDye 800, blots were washed and immunoreactive proteins were scanned on an Odyssey imager (Li-COR, Lincoln, NE, USA). Optical density on the membrane was measured and the relative expression of GFAP protein (GFAP/β-actin) of two groups was calculated.

Acutely dissociated cell culture

Acutely dissociated cell culture was performed as previously described (Hartfuss et al. 2001; Mo et al. 2007). Briefly, embryonic brains (E16.5) were isolated and the meninges were removed, the cerebral cortex were dissected and collected in separate vials. Tissues were dissociated in accutase (Sigma) for 15 min at 37°C and triturated into single cells through a fire-polish pipette. Cells were resuspended in Dulbecco’s modified Eagle’s medium/Ham’s F12 (DMEM/F12, 1:1) (GIBCO, Invitrogen, USA) containing 10% fetal bovine serum (FBS) supplemented with B27 (GIBCO), then seeded at a density of 1 × 105 into poly-l-lysine(Sigma)-coated coverslips in 24-well plates (BD Falcon, San Jose, CA, USA). Acutely dissociated cells were incubated at 37°C with 5% CO2 and 95% O2 for 4 h, then fixed in 4% paraformaldehyde and processed for immunocytochemistry.

Radial glia (RG)-like cells culture

The primary neurospheres were easily generated using a standard protocol (Brannvall et al. 2005; Li et al. 2011). Briefly, embryos were taken from the pregnant rats at E16.5 and embryonic cerebral cortex were dissected and isolated immediately. After removal of meningeal membranes, the tissue was incubated in accuse (Sigma) at 37°C for 20 min, then gently triturated into single cell suspensions using a fire polished pipette. The cell suspension was maintained at a density of 1 × 105 in the presence of 10 ml DMEM/F12 containing 2% B27 with 20 ng/ml EGF and FGF-2 (Sigma) as neurospheres expansion medium. The new formed neurospheres were passsaged every week by dissociation of bulk neurospheres using accutase. After expansion for three passages, neuroshperes were incubated in accutase about 20 min and triturated into single cell suspensions, then re-plated at a density of 1.5 × 104 cells per milliliter on poly-l-lysine-coated coverslips in 24-well plates (adherent conditions) which contained expansion medium (500 μl/well). 3 days later, these cells presented RG-like morphological feature and then processed for immunocytochemistry or extracted the total RNA.

To assess the differentiation ability of these RG-like cells, the RG-like cells were transferred into the DMEM/F12 medium supplemented with 2% FBS without B27 as basal differentiation medium. To promote neuronal differentiation, the RG-like cells were stimulated by basal differentiation medium supplemented with 10 ng/ml BDNF and 10 ng/ml NGF. Half the medium was exchanged every 3 days. 10 days later, immunocytochemistry were performed.

Immunocytochemistry of acutely dissociated cells and RG-like cells

Cells were fixed in 4% paraformaldehyde for 30 min at room temperature and then blocked using 10% normal goat serum with 1% BSA for 2 h. Then incubated with primary antibodies overnight at 4°C. Primary antibodies used to analysis the acutely dissociated cells and identify RG-like cells were as follows: mouse anti-Vimentin (1:200, Abcam), mouse anti-Nestin (1:200, Abcam), mouse anti-GFAP (1:500, Millipore), rabbit anti-Sox2 (1:1000, Abcam). To identify the neuronal differentiation cells, mouse anti-Tuj1 (1:1,000, Millipore) and rabbit anti-MAP2 (1:1,000, Abcam) were used; rabbit anti-GFAP (1:1,000, Abcam) and mouse anti-Sox2 (1:1,000, Abcam) were used to investigate astrocytes; mouse anti-CNPase (1:200; Millipore) was used to detect oligodendrocytes. Then the cells were incubated with secondary antibodies conjugated to fluorescein 568 or FITC at room temperature for 2 h. Cell nuclei were counterstained with Hoechst (Sigma). When the primary antibodies were omitted in immunocytochemistry, no immunoreactivity was detected.

PCR amplification of astroglial and progenitor genes in E16.5 cerebral cortex and RG-like cells

Total RNA from E16.5 fetal cortex and RG-like cells were isolated and purified using UNIQ-10 Spin Column RNA Purified Kit (Sangon, Shanghai, China). The first strand cDNA was synthesized using RevertAid™ First Stand cDNA Synthesis Kit (Fermentas), and the PCR amplification was performed using the Taq DNA Polymerase system (Fermentas, Burlington, Canada) as our previously described (Li et al. 2011). The sense and antisense primers were synthesized as follows: GAPDH 5′-GCAAGTTCAACGGCACAG-3′, 5′-GCCAGTAGACTCCACGACAT-3′; Vimentin 5′-AGGGATCCATGAGACACAGACA-3′, 5′-CGGATCCTTTACAGAGGTTTGG-3′; Nestin 5′-TACACCAGACCAGACCTTGTGC-3′, 5′-TTGACCCATGGTATTTGTCCTG-3′; GFAP 5′-GATGCAGAATAGCCAGGAAGAGA-3′, 5′-AACCTCTCTGGCCATTCATAGTG-3′; Sox2 5′-AGAACCCCAAGATGCACAAC-3′, 5′-CTCCGGGAAGCGTGTACTTA-3′, the reactions of PCR were optimized by varying the annealing temperatures from 48 to 53°C, the PCR products were visualized by UV irradiation.

Statistics analysis

We investigated positive stained cells and radial processes in the cerebral cortex from five coronal sections in each animal (E16.5 and P3; n = 3) using an Olympus laser confocal microscope (Olympus) with 42× objective (30 stacks in the z-axis), the double-labeled cells in the GZ were photographed using a 102× or 360× oil objective and processed using an Olympus Fluoview-ASW 2.1 software (Olympus).

In acutely dissociated cell culture, positive cells were detected using Olympus laser confocal microscope (120 × oil objective). In RG-like cells culture, labeled cells were visualized using Olympus laser confocal microscope (40×) or Leica DMR microscopy (20×) (Leica, Germany). Positive cells were counted in five randomly selected microscopic visual fields per well using Leica DMR microscopy (20×) (Leica).

Data from our experiments were subjected to Student’s t tests or One-way analysis of variance (ANOVA) using SPSS 11.5 statistical software, all data were represented as the mean ± SEM and all experimental data were obtained from a minimum of three independent experiments.

Results

RGCs in the embryonic cerebral cortex

Typical RGCs derive from neuroepithlium in vivo, their longer radial processes contribute to the neuronal migration. They can express neural stem/progenitor markers such as Sox2 or Nestin, and they also can be labeled by astroglia markers such Vimentin. They play a pivotal role in establishing the laminar structure of the cerebral cortex. In this report, we used neural progenitor marker Sox2 combined with astroglial marker Vimentin to investigate the distribution of RGCs in E16.5 cerebral cortex. We found that in the E16.5 cerebral cortex, the Sox2+ cells localized in the germinal zone (GZ) (Fig. 1a, f, k). Vimentin+ (Fig. 1b, d) or Nestin+ (Fig. 1g, i) radial processes co-localized with Sox2+ cells in the GZ (Fig. 1c, e, h, j), the longer radial processes of these cells projected into the cortical plate (CP) and reached into the pial surface (P) (Fig. 1b, g). High magnification (360×) of VZ and SVZ indicated the Sox2+ cells (Fig. 2b, f and n, r) co-localized with Vimentin (Fig. 2a, d and m, p) and Nestin (Fig. 2e, h and q, t). The radial morphological feature combined with the astroglial (expressed Vimentin) and neural progenitor (expressed Sox2 or Nestin) properties indicated that these Sox2+ cells were RGCs. We found 95.68 ± 0.59% and 96.98 ± 0.38% Sox2+ cells co-localized with Vimentin+ (Figs. 1e, 2d, p) or Nestin+ (Figs. 1j, 2h, t) radial processes in GZ, these results suggested that Sox2+ RGCs were the main type of neural progenitor cells in the E16.5 cerebral cortex.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0864-5/MediaObjects/418_2011_864_Fig1_HTML.gif
Fig. 1

Staining of E16.5 frontal cortex sections with astroglial and neural progenitor markers indicated in the figure. The left panel showed the Sox2+ cells localized in the germinal zone of cerebral cortex at embryonic (e) 16.5 (a, f, k), scale bar 500 μm; the middle panel (the higher magnification of the white boxes in the left panel) showed the Vimentin+ (b, c) or Nestin+ (g, h) radial processes crossed cortical plate into the pial surface, but the GFAP+ processes were undetectable (l, m), the Sox2+ cells could be double-labeled by Vimentin (c) or Nestin (h) but not GFAP (m), scale bar: 50 μm; the right panel showed the higher magnification (×102) of white boxes in (c), (h), (m) germinal zone (including VZ and SVZ). The 95.68 ± 0.59% and 96.98 ± 0.38% Sox2+ cells co-localized with Vimentin+ (d, e) or Nestin+ (i, j) radial processes, respectively, but the GFAP was negligible (o, p), scale bar 20 μm. GZ germinal zone, IZ intermediate zone, CP corical plate, P pial surface, VZ ventricular zone, SVZ subventricular zone

https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0864-5/MediaObjects/418_2011_864_Fig2_HTML.gif
Fig. 2

High magnification of the germinal zone in E16.5 cerebral cortex. Using a laser confocal microscope with ×360 oil objective, the Sox2+ cells (green) in the ventricular zone (VZ) (b, f, j) could co-localize with Vimentin (red) (a, d) and Nestin (red) (e, h), but not GFAP (i, l); in the subventricular zone (SVZ), the Sox2+ cells (green) (n, r, v) also co-localized with Vimentin (red) (m, p) and Nestin (red) (q, t), but not GFAP (u, x). Cell nuclei were stained by Hoechst (blue) (upper panelc, g, k; lower panel:o, s, w), and the typical co-localized cells were indicated by asterisks (*) in each panel. scale bars: 10 μm

It was reported that the GFAP mRNA and protein could be detected after E15 in the murine forebrain (Fox et al. 2004; Sancho-Tello et al. 1995). We found that the expression of GFAP protein in RGCs was nearly undetectable (Figs. 1K, L, O, 2i, u) and the Sox2+/GFAP+ double-labeled cells were negligible in the E16.5 cortex (Figs. 1m, p, 2l, x). Consistent with these findings, the acutely dissociated neural progenitor cells from E16.5 cerebral cortex (Fig. 3a, f) co-expressed Sox2 (Fig. 3c, h) with Vimentin (Fig. 3b, e) and Nestin (Fig. 3g, j). And the expression of GFAP in the acutely dissociated Sox2+ cells (Fig. 3m) was negligible (Fig. 3l, o). The ratio of Vimentin+/Sox+ (Fig. 3e) and Nestin+/Sox2+ (Fig. 3j) double-labeled cells to all acutely dissociated cells which labeled by Hoechst (blue) (Fig. 3d, i) were 16.31 ± 1.43% and 17.45 ± 1.81%, respectively.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0864-5/MediaObjects/418_2011_864_Fig3_HTML.jpg
Fig. 3

Absence of GFAP in acutely dissociated cells of E16.5 cerebral cortex. Some acutely dissociated cells (a, f, k) expressed Sox2 (c, h, m, arrows), and these Sox2+ cells could be double-labeled by Vimentin (b, e) and Nestin (g, j). However, the expression of GFAP (l, o) in the acutely dissociated Sox2+ cells was undetectable. The ratio of Vimentin+/Sox+ (E) and Nestin+/ Sox2+ (J) double-labeled cells to all acutely dissociated cells which marked by Hoechst (blue) (d, i) were 16.31 ± 1.43% and 17.45 ± 1.81%, respectively. Scale bar 200 μm

However at the P3 cerebral cortex, the GFAP+ radial processes could be found in the cortex (Fig. 4a, b), and the GFAP+/Sox2+ cells could be detected in the GZ (Fig. 4b, c, h, l) with the longer GFAP+ radial projections (Fig. 4e, i, arrows) crossing the CP (Fig. 4d) into the pial surface (P) (Fig. 4b), and these GFAP+ radial processes (Fig. 4m, p) could co-express with Vimentin (Fig. 4n, o) and Nestin (Fig. 4q, r). The relative expression of GFAP protein at P3 was about tenfold higher than the E16.5 (relative expression: 0.076 ± 0.008 at E16.5 vs. 0.899 ± 0.039 at P3; P < 0.01) (Fig. 4s, t).
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0864-5/MediaObjects/418_2011_864_Fig4_HTML.gif
Fig. 4

Expression of GFAP in P3 cerebral cortex. a,b In the frontal cortex sections, the Sox2+ cells (Green, b) in the germinal zone projected GFAP+ processes (red, a,b) across the cortical plate into the pial surface (×10); c,d The higher magnification of white and red box in b, respectively (×40), the GFAP+/Sox2+ cells localized in the germinal zone (c) and the GFAP+ processes radiated into the pial surface across the cortical plate (d). Note the radial fibers were long, thin and some Sox2+ cells located in the cortical plate (five cells were indicated randomly by arrows ↑), while the Sox2+ cells in E16.5 cortical plate were hardly detected (Fig. 1c, h, m). Scale bars 200 μm (a, b), 50 μm (c, d). el The Sox2+ cells (green) in the ventricular zone (f) and subventricular zone (j) co-localized with GFAP (e, i), the Sox2+/GFAP+ cells (h, l) were marked by asterisks (*), the GFAP+ radial processes were indicated by arrows (↑), cell nuclei were stained by Hoechst (blue) (g, k). mr The radial processes in P3 cerebral cortex could be labeled by GFAP (m, p), Vimentin (n) and Nestin (q), the merged images (o, r) indicated the GFAP+ radial processes co-expressed with Vimentin (o) and Nestin (r) (arrows, ↑), Scale bar: 5 μm. s,t Western blot analysis of GFAP from cerebral cortex at E16.5 and P3, the molecular weight of GFAP and β-actin was 51 and 43kD, respectively (s); the relative expression of GFAP levels at E16.5 and P3 cortex (t) was 0.076 ± 0.008 and 0.899 ± 0.039, respectively, there was significant difference between E16.5 and P3 (P < 0.01). GZ germinal zone, IZ intermediate zone, CP corical plate, P pial surface

In addition, we found that some Sox2+ cells in the CP at P3 which were adjacent to the radial processes (five Sox2+ cells were indicated randomly by arrows in Fig. 4d). However in the E16.5 cortex, the Sox2+ cells were hardly investigated in the CP and the Sox+ cells mainly aggregated in the GZ around the lateral ventricle (Fig. 1c, h, m). These results indicated that although the radial scaffold for the cell migration was established at E16.5 cortex, the Sox2+ progenitor cells still persisted quiescent and did not begin to migrate along these radial processes.

Identification of RG-like cells in vitro

The third-passage neurospheres were dissociated into single cells and re-plated into poly-l-lysine-coated coverslipes. 3 days later, these cells outgrew long, thin and unbranched processes from both ends of the cell bodies, and they displayed a typical bipolar morphological feature of RGCs (Fig. 5a, e, i). To identify these RG-like cells, we used neural progenitor marker Sox2 and Vimentin, Nestin, GFAP to label the radial processes of these cells. On the adherent conditions, these RG-like cells expressed the neural progenitor marker Sox2 (Fig. 5b, f, j), and these Sox+ RG-like cells elongated long and unbranched Vimentin+ (Fig. 5a, d), Nestin+ (Fig. 5e, h) or GFAP+ (Fig. 5i, l) radial processes. Then the number of double-labeled bipolar cells (Fig. 5d, h, l) versus the total cell number labeled by Hoechst (Fig. 5c, g, k) were calculated, the ratio of Vimentin+/ Sox2+ cells was 94.07 ± 0.76%, Nestin+/Sox2+ cells was 95.51 ± 0.79%, GFAP+/Sox2+ cells was 94.57 ± 0.56%, there was no significant difference in three groups (P > 0.05) (Fig. 5p). We did not find newborn neurons when used Tuj-1 (Fig. 5m) or DCX (Fig. 5n) to label cells. The astroglial and progenitor genes of Vimentin, GFAP, Nestin and Sox2 also could be detected in RG-like cells (RG) in vitro or cerebral cortex (Cor) in vivo (Fig. 5o).
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0864-5/MediaObjects/418_2011_864_Fig5_HTML.gif
Fig. 5

Identification of Sox2+ RG-like cells in vitro. The Sox2+ cells (green) (b, f, j) co-localized with Vimentin+ (a), Nestin+ (e) and GFAP+ (i) radial processes, respectively, cell nuclei were stained by Hoechst (blue) (c, g, k). The merged images (d, h, l) indicated the RG-like cells presented astroglial and progenitor characteristics. m and n showed no Tuj-1+ or DCX+ immature neurons could be detected on adherent conditions, scale bar 50 μm. o Some astoglial (Vimentin, GFAP) and progenitor (Nestin, Sox2) genes expressed in RG-like cells (RG) and E16.5 cerebral cortex (Cor). p The proportion of double-labeled cells to all cells in three groups showed no significant difference (P > 0.05)

Differentiation ability of E16.5 RG-like cells in vitro

To assess the differentiation ability of E16.5 RG-like cells in vitro, the RG-like cells were transferred into the basal differentiation medium (DMEM/F12 medium supplemented with 2% FBS). 10 days later, these RG-like cells changed their morphology, 95.13 ± 1.17% RG-like cells differentiated into GFAP+ astrocytes (Fig. 6a, c) but loss of Sox2 (Fig. 6b), 1.53 ± 0.52% oligodendrocytes were labeled by CNPase (Fig. 6f). And these RG-like cells also differentiated into neurons, we found 3.57 ± 0.58% cells could be labeled by the immature neurons marker Tuj1 (Fig. 6d) and 3.69 ± 0.75% cells were labeled by the mature neurons marker MAP2 (Fig. 6e).
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-011-0864-5/MediaObjects/418_2011_864_Fig6_HTML.gif
Fig. 6

Differentiation potentialities of RG-like cells in vitro. In 2% FBS basal differentiation medium, the RG-like cells differentiated into GFAP+ astrocytes (a, c) but loss of Sox2 (b), Tuj1+ immature neurons (d) or MAP2+ mature neurons (e), and CNPase+ oligodendrocytes (f). The proportion of differentiated cells was summarized in (g). Treatment of NGF and BDNF, the proportion of Tuj1+ (h) or MAP+ (i) neurons increased significantly respectively (P < 0.05) (j, k). Nuclei were stained by Hoechst. Neurons and oligodendrocytes were indicated by arrows (↑), scale bars 50 μm

Neurogenic microenvironment is a pivotal factor of neurogenesis in CNS, and the neurotrophic factors play important roles in both the developing and adult CNS microenvironment (Balu and Lucki 2009; Doetsch 2003b). NGF has been proved with potential of promoting neurogenesis in adult rats with focal cerebral ischemia (Zhu et al. 2011), and increasing concentration of NGF promoted the proportion of neurons in vitro (Wang et al. 2009). Administration of BDNF increased neurogenesis (Scharfman et al. 2005), BDNF enhanced cortical NS cells proliferation and directed cell fate toward neuronal differentiation (Horne et al. 2010). NGF combined with BDNF was used to promote neuronal differentiation in vitro (Conti et al. 2005). In this research, we used NGF combined with BDNF to simulate the changes of microenvironment and to promote neuronal differentiation. We detected 7.07 ± 1.24% Tuj1+ cells (Fig. 6h, j) or 8.45 ± 1.36% MAP2+ cells (Fig. 6i, k), the proportion of neurons increased significantly (P < 0.05) (Fig. 6j, k), these results indicated the RG-like cells could differentiate into more neurons when the microenvironment changed in vitro.

Discussion

During the development of CNS, RGCs appeared at approximately E13 and served as an important structural role by supporting the migration of neurons born in the germinal zone to their appropriate location and also function as neuronal and glial precursor cells (Anthony et al. 2004; Gregg and Weiss 2003; Kriegstein and Gotz 2003; Malatesta et al. 2003), RGCs in different brain regions showed regional heterogeneity (Malatesta et al. 2003), and the neurogenic capabilities of RGCs were both region and stage dependent (Mo et al. 2007). At the end of the developmental period, the RGCs retracted their processes and differentiated into astrocytes (Brunne et al. 2010; Voigt 1989). In many previous studies, a single astroglial marker was applied in identifying RGCs. Indeed morphological feature combined with stem/progenitor and astroglial markers provided a more reliable and effective method to identify RGCs (Conti et al. 2005; Liu et al. 2010, 2011; Pollard et al. 2006). Sox2 is a transcription factor of SRY-related HMG-box family, this factor is essential to maintain self-renewal of embryonic stem (ES) cells and is one of the key transcription factors required in induced pluripotent stem cells (ips) (Zhao and Daley 2008). In CNS, Sox2 is a persistent marker for multipotential neural stem/progenitor cells, it is expressed in functionally defined NS cells and also maintains NS cells properties, and Sox2 is required for the maintenance of neural progenitor cells and for their differentiation into neurons (Ellis et al. 2004; Pevny and Nicolis 2010). Evidences showed that RGCs maintain a Vimentin+ radial fiber throughout each stage of cell division (Weissman et al. 2003) and the expression of Vimentin could be detected from E13 to P21 (Sancho-Tello et al. 1995). Therefore, in this research, we used Sox2 as a neural progenitor marker and Vimentin as an astroglial marker to investigate the distribution of RGCs in E16.5 cerebral cortex. We observed that ~95% Sox2+ cells co-localized with Vimentin+ (Figs. 1c, e, 2d, p) or Nestin+ (Figs. 1h, j, 2h, t) radial processes in the GZ, these cells displayed a radial morphology with an elongated radial processes that extended to the pial surface (Fig. 1c, h), these results indicated that most neural progenitor cells in cerebral cortex were Sox2+ RGCs at E16.5. Sox2+ cells were undetectable in the CP at E16.5 (Fig. 1c, h, m), however the Sox2+ cells could be detected at P3 which adjacent to the radial processes (Fig. 4d), and these processes were longer than the E16.5 (Fig. 1b, g). We suggested that the longer processes contributed to the migration and differentiation of Sox2+ progenitor cells in the developing cerebral cortex.

Many strategies were applied in isolating RGCs. Transgenic reporters combined with fluorescent activated cell sorting (FACS) provided an effective method to isolate RGCs from primary cell cultures (Malatesta et al. 2000), and RGCs also could be induced by ES (Bibel et al. 2004; Glaser and Brustle 2005) or NS cells (Conti et al. 2005; Li et al. 2011; Pollard and Conti 2007; Pollard et al. 2006) on adherent conditions in vitro. When cultured on adherent conditions, RG-like cells exhibited bipolar or multipolar processes and expressed astroglial and neural progenitor markers (Conti et al. 2005; Hartfuss et al. 2001; Li et al. 2011; Pollard et al. 2006). However, the expression of GFAP in the RG-like cells from mouse fetal forebrain was negligible, while the GFAP could be detected in the RG-like cells from human fetal forebrain (Conti et al. 2005; Pollard and Conti 2007). And the gliogenesis ability of P1 rat hippocampus-derived RG-like cells far exceeded the neurogenesis ability and continued the radial glia—astrocyte lineage differentiation in vitro (Li et al. 2011). In this research, we obtained a ~95% pure RG-like cells which exhibited bipolar morphological feature, and these cells co-expressed Vimentin (Fig. 5a, d), Nestin (Fig. 5e, h) with Sox2 (Fig. 5b, f), these Sox2+ RG-like cells also expressed GFAP (Fig. 5i, l) which was consistent with previous research (Li et al. 2011), and suggested that the rat and human-derived RG-like cells expressed GFAP which was different from the mouse-derived cells.

In early postnatal hippocampal dentate gyrus, the RGCs confined to an astrocytic fate (Brunne et al. 2010), and the neonatal hippocampus-derived RG-like cells mainly differentiated into astrocytes in vitro (Li et al. 2011). To assess the neuronal differentiation ability of embryonic cortex-derived RG-like cells in vitro, the RG-like cells were transferred into basal differentiation medium which contained 2% FBS. 10 days later, we found 95.13 ± 1.17% astrocytes (Fig. 6g) labeled by GFAP (Fig. 6a, c) but loss of Sox2 (Fig. 6b), 1.53 ± 0.52% oligodendrocytes (Fig. 6g) labeled by CNPase (Fig. 6f), and a ~3% cells (Fig. 6g) expressed immature neurons marker Tuj1 (Fig. 6d) or mature neurons marker MAP2 (Fig. 6e). In our previous research, less than 0.5% neurons could be detected in the neonatal hippocampal RG-like cells under basal differentiation medium (Li et al. 2011), the difference of neuronal differentiation ability between E16.5 cortex and neonatal hippocampus suggested E16.5 RG-like cells conveniently differentiated into neurons and displayed neurogenic ability. To test the hypothesis that the changes of microenvironment played a role in determining the differentiation of E16.5 RG-like cells in vitro, neurotrophic factors were used to stimulate the neuronal differentiation of these cells. We found the treatment of basal differentiation medium with BDNF and NGF increased the proportion of neurons (~8%) in vitro. It was reported that in the early developing cortex harbored two main progenitor cells: neuron-restricted and bipotent (neuron or glial) progenitors, the latter were responsible for the generation of glial progenitors at mid or late development, and the early developing cortex was devoid of glial-restricted progenitors (Costa et al. 2009). In this research, the E16.5 RG-like cells were similar to the bipotent progenitors, they differentiated into neurons and glial cells, and the changes of neurogenic microenvironment appeared to promote the differentiation of E16.5 RG-like cells into more neurons, the E16.5 cerebral cortex-derived RG-like cells displayed a radial glia—neuron lineage differentiation as neuronal progenitors which was different from neonatal hippocampus-derived RG-like cells.

Evidence suggested astroglial cells could be directly converted into neurons by specific proteins, non-neurogenic astroglial from the cerebral cortex could be reprogrammed in vitro using a single transcription factor to yield functional neurons (Heinrich et al. 2010), the astroglial property was an important characterization of RGCs both in vivo and in vitro. However, the expression of astroglial markers in RGCs was not consistent during the development (Brunne et al. 2010). In this research, the expression of GFAP was negligible in E16.5 cerebral cortex (Figs. 1k–p, 2i, l, u, x) and absent in the acutely dissociated cerebral cortical cells (Fig. 3l, o). However, the expression of GFAP could be detected in P3 rat cortex (Fig. 4a, e, i). Combined with others results (Brunne et al. 2010; Fox et al. 2004; Liu et al. 2010; Sancho-Tello et al. 1995), it indicated that the temporal sequence of GFAP expression in the developing cortex, and the expression of GFAP was later than the other astroglial markers such as Vimentin or GLAST (Hartfuss et al. 2001; Sancho-Tello et al. 1995). When cultured on adherent conditions, the rat cortex-derived RG-like cells expressed GFAP (Fig. 5i, l), which was consistent with the human-derived but different from mouse-derived RG-like cells (Conti et al. 2005; Pollard and Conti 2007). So it indicted that the expression of GFAP in RGCs was both “timing-specific” in vivo (negligible in E16.5 and detectable in P3) and “species-specific” in vitro (negligible in mouse but detectable in human and rat), the dynamic expression of GFAP and the proportion of Sox2+ RGCs in cerebral cortex at different developing stages need to be further studied.

In conclusion, the Sox2+ neural progenitor cells in E16.5 cerebral cortex co-localized with Vimentin+ or Nestin+ radial processes which extended to the pial surface through the CP. These Sox2+ RGCs could be obtained on adherent conditions in vitro, the cultured RG-like cells presented some astroglial and progenitor properties just as RGCs in vivo. The RG-like cells could differentiate into astrocytes and oligodendrocytes, they also presented the radial glia—neuron lineage differentiation as neuronal progenitors which was different from P1 hippocampus-derived RG-like cells, and the expression of GFAP in cortex-derived Sox2+ RGCs was both “timing-specific” in vivo and “species-specific” in vitro.

Acknowledgments

This study was supported by grants from the National Natural Science Foundation of China (Grant Nos. 31070937), the Application Research Project of Nantong City (Grant Nos. K2009025), and a Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).

Copyright information

© Springer-Verlag 2011