Overexpression of the Epidermal Growth Factor Receptor Confers Migratory Properties to Nonmigratory Postnatal Neural Progenitors

Introduction

The use of cell transplantation for brain repair depends on neural progenitors that retain their ability to proliferate, migrate, and differentiate to repopulate areas of lesions. A first step in designing cell repair strategies is to analyze the developmental properties of specific neural progenitor populations, to identify intrinsic mechanisms that could be exploited to promote cell regeneration by using endogenous or grafted cells.

The postnatal brain contains populations of progenitors that display different proliferative and migratory potentials. The most extensively characterized are those found in the subventricular zone (SVZ) (Doetsch et al., 1999, 2002) and in the dentate gyrus (Gage, 2000). One class of progenitors in these areas expresses the proteoglycan NG2. NG2⁺ cells are the most numerous progenitor population of the postnatal brain (Dawson et al., 2000) and can be either highly proliferative and migratory, such as in the SVZ, or slowly dividing and nonmigratory, such as in the cerebral cortex and olfactory bulb (OB) (Belachew et al., 2002; Aguirre and Gallo, 2004; Aguirre et al., 2004). We have demonstrated previously that SVZ NG2⁺ cells isochronically transplanted into the lateral ventricle (LV) proliferate and migrate caudally to the hippocampus and subcortical white matter (SCWM) and to the OB via the rostral migratory stream (RMS) (Aguirre and Gallo, 2004). Conversely, grafted cortical NG2⁺ cells neither proliferate nor migrate out of the LV (Aguirre and Gallo, 2004). These findings raise important questions about the intrinsic properties that render a progenitor cell proliferative and migratory and provide the opportunity to define signaling molecules that may promote proliferation and migration in quiescent neural progenitors.

Several molecular signals have been implicated in neuroblast tangential migration, including the following: (1) the polysialylated form of the neural cell-adhesion molecule, which confers a migratory phenotype to neuroblasts (Pencea and Luskin, 2003); (2) deleted in colorectal carcinoma and integrins, which mediate directional migration of neural precursors (Murase and Horwitz, 2002); (3) the soluble ligands Slit1 and Slit2, as guidance signals (Bagri et al., 2002), and (4) members of the receptor tyrosine kinase family (RTK), which includes the Eph family and their transmembrane-associated ephrin ligands (Conover et al., 2000) and members of the Erb RTK family, particularly Erb4 and its ligand neuregulin (Anton et al., 2004; Corfas et al., 2004; Flames et al., 2004). The Erb RTK family also includes the epidermal growth factor receptor (EGFR) (Erb1) and its ligand EGF, which play an important role in the proliferation and maintenance of embryonic and adult neural stem cells (Doetsch et al., 1999, 2002; Gage, 2000; Junier, 2000; Yarden and Sliwkowski, 2001). The EGFR is directly involved in the radial migration and maturation of neural precursors during embryonic cortical development (Burrows et al., 1997, 2000; Kornblum et al., 1997; Caric et al., 2001; Ciccolini et al., 2005), but its functional role in postnatal neural progenitor is not completely defined.

In the present study, we demonstrate that EGFR overexpression in nonmigratory cortical NG2⁺ cells induces these cells to migrate from the SVZ into the SCWM, the hippocampus, and the RMS/OB after grafting. Our results indicate that increasing EGFR expression in neural progenitor cells to enhance their migratory potential may help to develop new transplantation therapies by optimizing cell engraftment in different brain regions.

Materials and Methods

Transgenic mice. The CNP-enhanced green fluorescent protein (EGFP) and β-actin-GFP transgenic mice (Tg ActbGFP) (stock number 003772; The Jackson Laboratory, Bar Harbor, ME) have been described previously (Belachew et al., 2002; Yuan et al., 2002; Aguirre and Gallo, 2004). Details on the generation and characterization of the CNPase-hEGFR transgenic mice have been reported previously (Ling et al., 2005). Genotyping of these transgenic mice was performed by PCR (Ling et al., 2005). Transgenic mice were backcrossed more than four generations onto C57BL/6 mice. In two lines of CNPase-hEGFR mice, including the one used here, no obvious defects in brain oligodendrocytes were detected. For example, there was no change in width of the adult corpus callosum. Consistent with CNP driving expression in oligodendrocytes, robust hEGFR expression in total brain and spinal cord lysates from adult brain was detected previously by using monoclonal anti-human EGFR antibody to probe Western blots after immunoprecipitation with a polyclonal anti-EGFR antibody (Ling et al., 2005). In addition, hEGFR expression was detected in oligodendrocyte lineage cells of the white matter in both postnatal day 15 (P15) and adult CNPase-hEGFR mice (A. Aguirre and V. Gallo, unpublished observation). All animal procedures were performed according to the Institutional Animal Care and Use Committee of Children's National Medical Center and the National Institutes of Health guidelines.

Real-time PCR. Messenger RNA was extracted from postnatal day 8 whole brain and SVZ using the Micro-FastTrack kit for isolation of mRNA from small samples (Invitrogen, Carlsbad, CA). The mRNA was DNase treated (New England Biolabs, Ipswich, MA) according to the instructions of the manufacturer and purified with phenol chloroform. cDNA was generated as described previously (Ling et al., 2005), and duplicate samples lacking reverse transcriptase were conducted to control for genomic DNA contamination. Mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primers (sense, 5′-ACCCAGAAGACTGTGGATGG; antisense, 5′-GGAGACAACCTGGTCCTCAG; expected product size, 300 bp) were included in each reaction as a positive control for cDNA. For quantitative real-time PCR experiments, cDNA was used (as generated above), as were primers for known EGFR ligands: amphiregulin, betacellulin EGF, epiregulin, heparin-binding EGF, and transforming growth factor α (TGFα) (sequences available on request). Duplicate experiments with triplicate reactions were performed in an ABI Prism 7700 Sequence Detection System Cycler (Applied Biosystems, Foster City, CA) according to the instructions of the manufacturer. Briefly, 1 μl of cDNA or water control was placed into a 50 μl reaction volume containing 25 μl of SYBR Green Master Mix (2× concentration; Applied Biosystems) and volumes of primers that ranged between 0.5 and 9 μl, depending on the optimal conditions for each primer set. The remaining volume comprised water. The thermal cycling conditions comprised an initial equilibration step at 60°C (2 min), a denaturation step at 95°C (10 min), followed by 40 cycles of 95°C (15 s) and 60°C (1 min). When applicable, cycle threshold (C_t) values were obtained from the point during amplification at which the fluorescent intensity was in the geometric phase, as determined by PE Applied Biosystems (Foster City, CA) analysis software. All PCR products were analyzed on a 2% agarose gel. When applicable, the ΔC_t values were determined for whole brain and SVZ, and relative ligand expression was calibrated to GAPDH expression (primers as above). Fold change of ligands in SVZ compared with whole brain were calculated using the established equation (K. Luvak, PE ABI Sequence Detector User Bulletin 2): 2^{-ΔΔ Ct}, where C_t is the cycle number at the chosen amplification threshold, ΔC_t = C_t(ligand) - C_t(GAPDH), and ΔΔC_t = ΔC_t(+/-) - ΔC_t(+/+).

SVZ explants migration assays. P8 brains from CNP-EGFP mice were dissected out and transferred to ice-cold HBSS medium (Invitrogen). The brains were cut in 200 μm-thick sections, and only those sections containing the SVZ were selected for additional manipulation. The SVZ was dissected along the lateral wall of the lateral ventricles under a surgical microscope, and then cut into small pieces (50-100 μm in size) that were placed on top of 50% BD PuraMatrix Peptide Hydrogel (BD Biosciences, Franklin Lakes, NJ). Peptide Hydrogel was prepared following the manufacturer's recommendations. Briefly, 35 mm diameter cell inserts (0.4 μm pore size; Corning, Corning, NY) were coated with Peptide Hydrogel diluted 1:1 with 1xPBS (pH7.4). After polymerization, the gel was equilibrated three times with neurobasal medium containing B27, N2 supplements, l-glutamine (0.5 mm) and penicillin-streptomycin (all from Invitrogen). Cultures were maintained in a humidified incubator at 5% CO₂ and 37°C for 4 d in vitro. SVZ explants were cocultured with heparin bead (Sigma, St Louis, MO) loaded with EGF (2 μg/ml) and control (0.1% BSA; Sigma) proteins. Explants were symmetrically positioned in between EGF- and BSA-loaded beads (300-400 μm from each set of beads), using a graticulate with scale division. This configuration allowed the simultaneous analysis of EGFP⁺ cell migration with the two sets of beads. For direct quantification of cell migration, SVZ explants were subdivided into four quadrants, and the areas occupied by migrating cells in two of these quadrants (facing either BSA- or EGF-coated beads) were quantified by image analysis as previously described (Pozas et al., 2001). The same results were also obtained by separately testing individual explants with EGF- and BSA-loaded heparin beads.

In a separate set of experiments, the original SVZ explants were scraped off the culture dish at 72 h days after plating, and CNP-EGFP⁺ cells that had migrated out of the explants and had remained attached to the dish were FACS-purified. These cells were then used to grow clonal neurospheres (5 cells/μl) (Aguirre and Gallo, 2004) for 6-7 d, and then plated in differentiation medium to determine their lineage potential (Aguirre and Gallo, 2004; Aguirre et al., 2004).

Fluorescence activated cell sorting (FACS). FACS-purification of NG2⁺/EGFP⁺ and NG2⁺/GFP⁺ cells has been previously described in detail (Belachew et al., 2002; Aguirre and Gallo, 2004). Tissue from P8 CNP-EGFP, β-actin-GFP, CNP-hEGFR, or CNP-hEGFR/Tg ActbGFP mice was dissociated to a single cell suspension, followed by immunostaining for NG2 (Aguirre et al., 2004). To fluorescence-activated cell sorter (FACS) purify NG2⁺ cortical cells transfected with an hEGFR or a constitutively active human EGFR (hEGFRviii) construct (see below), anti-NG2 (Chemicon, Temecula, CA), and anti-hEGFR (Biofluids, Camarillo, CA) antibodies were used in combination with appropriate R-phycoerytrin (PE)-conjugated and PE- cyanine 5.5 (Cy5.5)-conjugated secondary antibodies, respectively (Caltag, Burlingame, CA). To FACS purify NG2⁺ cortical cells transfected with phosphorylated cytomegalovirus (pCMV)-internal ribosomal entry site (IRES)-GFP (mock-GFP), anti-NG2 antibodies and GFP expression were used (Belachew et al., 2002; Aguirre and Gallo, 2004). Cell suspensions were analyzed for light forward and side scatter using an FACStar plus instrument (Becton Dickinson, Franklin Lakes, NJ).

Chemotaxis assays in Boyden chambers. FACS-purified NG2⁺ cells were obtained from the SVZ or the cerebral cortex of P8 CNP-EGFP mice (Aguirre and Gallo, 2004). Cortical transfected NG2⁺/GFP⁺/hEGFR⁺, and NG2⁺/GFP⁺/hEGFR⁺viii and NG2⁺/mock-GFP⁺ cells were FACS purified as described above. All NG2⁺ cells were resuspended at 1 × 10⁶ cells/ml in stem cell medium (SCM) (serum-free DMEM/F-12) supplemented with N2 and B27 (Invitrogen) for chemotaxis assays. In vitro migration of NG2⁺ cells was assessed using poly-l-lysine-, laminin-, vitronectin-, or fibronectin-coated (used at 5, 2, and 5 μg/ml, respectively) polyvinylcarbonate-free membranes with 8 μm pore size (Neuroprobe, Gaithersburg, MD). Assays were performed in modified Boyden chambers. Briefly, 25 μl of a solution containing 1 × 10⁶ cells/ml in SCM was placed in the upper chamber. Growth factors, including EGF (10 ng/ml), TGFα (10 μg/ml), basic FGF (bFGF) (10 μg/ml), or platelet-derived growth factor (PDGF) (10 μg/ml) were all diluted in DMEM/F-12 and added to the lower chamber in DMEM/F-12 or to both chambers when indicated. In the experiments in which EGFR blockers were used, cells were preincubated with the inhibitors for 30 min at 37°C in 5% CO₂ before loading in the top chamber. Blockers included the following: (1) the cell-permeable, irreversible and selective inhibitor of EGFR tyrosine kinase activity PD168393 (4-[3(bromophenyl)-amino]-6-acrylamidoquinazoline), which does not inhibit other kinases (Calbiochem, La Jolla, CA); (2) selective anti-hEGFR (clone LA1, catalog #05-101) or anti-EGF-neutralizing antibodies (catalog #06-102) (both from Upstate, Charlottesville, VA), and (3) the arginine-glycine-aspartate (RGD) peptide inhibitor of integrin-ligand interactions (Calbiochem). After overnight incubation (12 h) at 37°C in 5% CO₂, the upper surface of membranes was scraped free of cells and debris, and membranes were air dried, fixed, and mounted. Cells that had migrated through the pores and adhered to the membrane were analyzed under high-power light microscopy and counted in five to six adjacent high-power fields (HPFs) (area of HPF, 0.78 mm²; area of each filter, 7.07 mm²) in triplicate (three different membranes for each experimental condition). Experiments were repeated three times, and data were analyzed from a total of 45-54 fields from nine membranes for each experimental condition. Data were expressed as numbers of cells per HPF ± SEM. Data were analyzed for statistical significance between groups by using the unpaired t test.

Bromodeoxyuridine pulse and terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling assays. FACS-purified NG2⁺ cells were seeded (25,000 cells per well) in 24-well culture plates precoated with laminin. Cells were cultured in SCM for 12 h in the presence or absence of EGF (10 ng/ml). Cells were pulsed with 10 μm bromodeoxyuridine (BrdU) for 12 h before fixation. BrdU immunostaining and terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling (TUNEL) assay were performed as described previously (Aguirre et al., 2004).

Antibodies and immunohistochemistry. Freshly cut, floating tissue sections (50 μm) from P8 mice were prepared as described previously (Aguirre et al., 2004). Primary antibody dilutions were as follows: 0.5 μg/ml for the specific anti-hEGFR (Biofluids); 1:50 for the anti-EGFR (Upstate) and anti-BrdU (BD Biosciences); 1:100 for the anti-caveolin (Santa Cruz Biotechnology, Santa Cruz, CA); 1:200 for anti-EGF (Chemicon); and 1:500 for anti-NG2 antibody (Chemicon), anti-Ki-67 (Novocastra, New-castle, UK), anti-β-catenin (BD Biosciences), anti-N-cadherin (BD Biosciences), anti-caspase 3 (Cell Signaling Technology, Beverly, MA), anti-GFP (Clontech, Palo Alto, CA), and anti-GFAP (Sigma, St. Louis, MO). Nuclear staining was performed using TO-PRO (Invitrogen).

Western blot analysis. FACS-purified NG2⁺ cells were used for protein extraction using lysis buffer [50 mm Tris-HCl, pH 7.5, 1 mm EDTA, 1 mm EGTA, 1 mm sodium orthovanadate, 50 mm sodium fluoride, 0.1% 2-mercaptoethanol, and 1% Triton X-100, plus proteases inhibitor cocktail (Sigma)]. Protein samples (20 μg) were separated on GENE Mate express Gels 4-20% (ISC BioExpress, Kaysville, UT) and transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA). hEGFR, EGFR, and EGF bands were detected using an enhanced chemiluminescence substrate mixture (ECL Plus; Amersham Biosciences, Buckinghamshire, UK). A selective primary anti-human EGFR (Biofluids), anti-EGFR (Upstate), anti-EGF (Chemicon), and anti-phosphorylated extracellular signal-regulated kinase (pERK) (Cell Signaling Technology) antibodies were used in combination with a secondary horseradish peroxidase conjugate (Jackson ImmunoResearch, West Grove, PA).

DNA constructs and transfection procedure. Cortical cell cultures were prepared from P8 Tg ActbGFP or from FVB/NxCB6 [wild type (wt)] mouse brains. Cortical tissue was dissociated to a single cell suspension (Aguirre et al., 2004) and plated onto poly-l-lysine precoated 60 mm dishes at a density of 1.5 × 10⁶ cell/ml in DMEM/10% FCS medium for 12 h before transfection. For cell transfection, CMV promoter vectors containing the cDNA of the wild-type hEGFR (Upstate) and of the constitutive active EGFR (pLTR-EGFRviii) (Schmidt et al., 1998) were used. A pCMV-IRES-GFP (Clontech) vector was used as a mock plasmid in controls. All plasmid constructs were introduced into cultured cortical cells by Lipofectamine-mediated transfection (GTS, San Diego, CA), using 3 μg of DNA and 10 μl of liposomes in OptiMEM (Invitrogen). The duration of the transfection was 8 h. After transfection, the culture medium was changed to DMEM/10% FBS for 36 h, and cells were harvested and then processed for double immunostaining with anti-NG2 and anti-hEGFR antibodies. The anti-hEGFR antibody that was used specifically recognizes the wt and the constitutively active hEGFRs and does not react with the murine EGFR (Biofluids). Cells were then FACS purified as described above.

Transplantation and analysis of grafted cells. Wild-type P4 FVB/NxCB6 pups were used as recipients. Donor cells included the following: (1) FACS-purified P8 NG2⁺/GFP⁺ cells from the β-actin-GFP mouse strain (Aguirre and Gallo, 2004); (2) FACS-purified P8 SVZ or cortical NG2⁺ cells from the CNP-hEGFR mouse strain (Ling et al., 2005); (3) FACS-purified P8 cortical NG2⁺/GFP⁺ cells from the CNP-hEGFR mouse strain crossed with the β-actin-GFP mouse strain (CNP-hEGFR/Tg ActbGFP mouse); (4) cultured cortical NG2⁺/GFP⁺ cells from β-actin-GFP mice transfected with the hEGFR or the hEGFRviii construct (NG2⁺/GFP⁺/hEGFR⁺ and NG2⁺/GFP⁺/hEGFR⁺viii, respectively); and (5) cultured cortical NG2⁺ cells from FVB/NxCB6 mice transfected with the pCMV-IRES-GFP construct (NG2⁺/mock-GFP⁺). Immediately after FACS, cells were washed with DMEM plus 10%FCS and resuspended in DMEM at 20,000 cells/μl. Microinjection was performed under deep hypothermia, and 1 μl of the cell suspension was injected directly into the ventricles. In some experiments, trypan blue (0.1%) was added to the cell suspension to confirm targeting of injections. The injection was performed with a glass capillary pipette (100-150 μm outer diameter with beveled tip) driven by a Harvard Instruments (Murrieta, CA) micromanipulator. In the grafting experiments using donor cortical cells from the CNP-hEGFR mouse, cells were also stained with chloromethylbenzamido-DiI (Invitrogen), as recommended. Briefly, NG2⁺/CNP-hEGFR⁺ cells were incubated for 5 min with DiI (1 μg/ml; Invitrogen) at 37°C and then placed on ice for 15 min. The reaction was stopped with basal DMEM medium, and cells were washed twice with DMEM. For migration and anatomical distribution studies, mice were maintained for 7-10 d after the cell graft and processed for immunohistochemistry as described previously (Aguirre et al., 2004). Cells derived from grafted NG2⁺/GFP⁺ cells or from transfected cells were readily visible under confocal microscopy in unstained tissue, using a 488 nm laser-line excitation. Cells derived from wt/NG2⁺/mock-GFP⁺ cells were visualized after anti-GFP immunostaining with an anti-GFP rabbit polyclonal antibody (IgG, 1:500; Clontech), as described previously (Aguirre et al., 2004). The percentages of total transplanted GFP⁺ or DiI⁺ cells found in striatum, SCWM, wall of the LV, SVZ, RMS, fimbria, OB, and hippocampus were quantified in three separate experiments as described below.

For analysis of cell differentiation 4 weeks after transplantation, neuronal-specific nuclear protein (NeuN), CNP, and GFAP antibodies were used as cell markers to identify neurons, oligodendrocytes, and astrocytes, respectively. For counting of cells derived from transplanted cortical NG2⁺/hEGFR⁺/EGFP⁺ cells, EGFP⁺ cells were counted in different brain regions. Cell numbers were obtained by analyzing a total of 1850 cells in 300 sections (40 μm thickness) from three different experiments (three different brains). Cell counting data in tissue sections were expressed as averages ± SEM. Statistical analysis was performed by paired t test.

Microscopy and cell counting. An Olympus Optical (Melville, NY) BX60 fluorescence inverted microscope was used to visualize immunofluorescence in cultured cells. Images were acquired using a 40× objective. For analysis in tissue sections, a Bio-Rad (Hercules, CA) MRC 1024 confocal laser-scanning microscope equipped with a krypton-argon laser and an Olympus Optical IX-70 inverted microscope were used to image localization of FITC (488 nm laser-line excitation; 522/35 emission filter), Texas Red (568 nm excitation; 605/32 emission filter), and Cy5 (647 excitation; 680/32 emission filter). Optical sections (Z = 0.5 μm) of confocal epifluorescence images were sequentially acquired using a 40× [numerical aperture (NA), 1.35], a 60× (NA, 1.40), or 100× oil objective (NA, 1.35) with Bio-Rad LaserSharp version 3.2 software. Confocal Assistant 4.02 was used to merge images. Merged images were processed in Photoshop 7.0 (Adobe Systems, San Jose, CA) with minimal manipulations of contrast.

For cell counting in cultured cells, averages were obtained from three separate sets of cultures. In each culture, at least 10 separate microscopic fields (248 μm²) were analyzed. For GFP⁺ and DiI⁺ cells in tissue sections, cells were counted in the striatum, SCWM, wall of the LV, SVZ, RMS, fimbria, OB, and hippocampus 1 week after grafting. The anatomical distribution of GFP⁺ and DiI⁺ cells was analyzed in Z-series confocal scanning images [20-30 μm thickness; step size, 0.5 μm between successive images of the same field (228 μm²)]. Cell numbers were obtained by analyzing the entire brain after grafting, i.e., an average of 144 total tissue sections per brain (40 μm thickness) from at least three to four independent grafting experiments (three to four different brains total). The total number of graft-derived cells per brain ranged between 1400 and 1600 for any of the transplanted cell populations (Aguirre and Gallo, 2004). The number of cells found in each region was expressed as percentage of the total number of graft-derived cells found in each brain. Cell counting data in tissue sections are expressed as averages ± SEM. Statistical analysis was performed by unpaired t test.

Results

Discussion

The identification of cellular and molecular signals that promote proliferation, migration, and survival of postnatal neural progenitors is a priority in designing approaches for cell transplantation-based therapies (Bjorklund and Lindvall, 2000; Lindvall et al., 2004). The present study shows that multipotential NG2⁺ progenitor cells offer a unique opportunity to devise cell transplantation strategies in the CNS, because these cells are found in different areas of the brain, in which they display distinct migratory and lineage potentials. By combining in vitro and in vivo studies, we demonstrate that endogenous EGFR expression is causally related to the different migratory potential that NG2⁺ progenitor cells display in distinct brain regions. We also show that enhancing EGFR expression in nonmigratory NG2⁺ cells that otherwise express lower levels of this receptor causes a conversion to a migratory phenotype and promotes their dispersal into different regions of the brain after transplantation.

Previous studies have demonstrated that neural progenitor cells from the postnatal/adult SVZ can migrate to different brain regions, i.e., rostrally through the RMS to the OB (Law et al., 1999; Alvarez-Buylla et al., 2000), and radially/caudally to the SCWM (Kakita and Goldman, 1999; Suzuki and Goldman, 2003). Under pathological conditions, SVZ progenitors display additional migration patterns. In the ischemia brain injury model, after exogenous EGF and FGF-2 application, Mash1⁺ multipotential neural progenitors migrate from the SVZ to the hippocampus, in which they terminally differentiate to interneurons (Nakatomi et al., 2002). NG2⁺ progenitor cells in the early postnatal SVZ are highly proliferative and express cellular properties that correspond to the adult type C cells, including Mash1 (Aguirre et al., 2004). When these cells were transplanted into the LV, they migrated rostrally and caudally to the OB and SCWM, respectively, but also to the hippocampus, in which they generated inhibitory interneurons (Aguirre and Gallo, 2004; Aguirre et al., 2004). Altogether, these results indicate that different SVZ progenitors are likely to use distinct endogenous migratory routes under physiological or pathological conditions.

EGF and EGFRs are expressed in the SVZ, RMS, and hippocampus (Seroogy et al., 1995; Anton et al., 2004; Fox and Kornblum, 2005). EGF is the best characterized mitogen for SVZ neural progenitor cells (Kuhn et al., 1997; Doetsch et al., 1999, 2002; Gritti et al., 1999). Intraventricular injection of EGF into the adult SVZ promotes progenitor proliferation as well as migration out of the lateral wall (Craig et al., 1996), indicating that EGF and EGFRs may play a dual role in SVZ progenitor development. Our results indicate that both EGFRs and EGF are expressed at higher levels in NG2⁺ progenitor cells of the SVZ than in cortical NG2⁺ cells, indicating that autocrine activation of EGFRs is likely to occur in SVZ NG2⁺ cells (Junier, 2000). In the rodent brain, mobilization and migration of SVZ progenitors into the SCWM is still occurring during the first postnatal week (LeVine and Goldman, 1988; Marshall and Goldman, 2002; Kakita et al., 2003; Zerlin et al., 2004). Therefore, the finding that EGFRs and EGF are highly expressed in NG2⁺ cells that display a migratory phenotype are consistent with the idea that EGFRs could regulate migration of these cells out of the postnatal SVZ.

In the present study, we demonstrate marked differences in the migratory potential of SVZ versus cortical NG2⁺ cells and a functional role for EGFRs in SVZ NG2⁺ cell migration. EGF stimulates SVZ NG2⁺ cell migration by a concentration-dependent gradient mechanism but is ineffective on cortical cells. Inhibition of EGFR activation by using either a receptor blocker or an EGF-neutralizing antibody prevents migration of SVZ NG2⁺ cells. Finally, EGF-induced migration of NG2⁺ cells is likely to involve integrin-ECM protein interactions (Lin and Bertics, 1995; Hintermann et al., 2001) because (1) it only occurs in the presence of laminin, fibronectin, or vitronectin and (2) is prevented by a peptide inhibitor of the integrin-ECM binding site. These results are consistent with previous findings showing that interactions between EGF and integrin signaling play an important role in supporting cell migration (Lin and Bertics, 1995; Moro et al., 1998; Xie et al., 1998).

Previous studies revealed a role for the EGFR in radial neural progenitor migration in the developing telencephalon (Burrows et al., 1997; Caric et al., 2001; Ciccolini et al., 2005). To demonstrate a function for EGFRs in SVZ tangential and radial/caudal migration in vivo, we used a previously established transplantation paradigm of early postnatal NG2⁺ cells FACS purified from the SVZ and the cerebral cortex. Grafted SVZ NG2⁺ progenitor cells migrate extensively into the postnatal RMS, OB, SCWM, and hippocampus, whereas cortical NG2⁺ cells display very restricted migration (Aguirre and Gallo, 2004). Here we show that SVZ NG2⁺ progenitor cells preincubated with EGF signaling-neutralizing antibodies before transplantation did not migrate out of the wall of the LV and the fimbria.

In additional support of a role of EGFR in promoting migration of NG2⁺ cells, we overexpressed the hEGFR in nonmigratory cortical NG2⁺ cells and found that cortical NG2⁺/hEGFR⁺ cells migrate in the presence of EGF in vitro. This process is (1) dependent on an EGF concentration gradient, (2) ECM/integrin-dependent, and (3) prevented by blockage of the EGFR. We then used two different approaches to demonstrate that enhancement of EGFR expression increased the migratory potential of cortical NG2⁺ cells in vivo. First, we transplanted transfected cortical NG2⁺/hEGFR⁺ cells into the early postnatal LV, and, second, we grafted cortical NG2⁺ cells FACS purified from a transgenic mouse overexpressing the hEGFR under the CNP promoter (Ling et al., 2005). Previous analysis from our laboratory demonstrated that all NG2⁺ cells express the CNP gene (Yuan et al., 2002; Aguirre et al., 2004; Chittajallu et al., 2004). Both approaches used in the present study gave very similar results, i.e., a few days after transplantation, graft-derived cortical cells were found in the SCWM and hippocampus and in numbers similar to those observed after transplantation of SVZ NG2⁺ cells. Furthermore, as demonstrated previously with SVZ NG2⁺ cells (Aguirre and Gallo, 2004), a large percentage of graft-derived cells were also found migrating through the RMS into the OB. The increased percentages of cells found in these regions was not attributable to an increased rate of cell proliferation of the grafted NG2⁺/hEGFR⁺ cells, as demonstrated by (1) similar anti-Ki-67 staining in graft-derived control and hEGFR⁺ cells, and (2) the finding that increased percentages of graft-derived hEGFR⁺ cells in specific brain regions (i.e., OB, RMS, SCWM, and hippocampus) consistently corresponded to a proportional decrease in those regions from which these cells were migrating (i.e., SVZ, fimbria, and lateral wall). These results are consistent with the interpretation that, in our experimental model, EGFRs are necessary for NG2⁺ cell migration in the postnatal brain. It appears that unique interactions between cellular EGFRs and the tissue environment play a crucial role in directing NG2⁺ cell migration to different brain regions.

Cortical NG2⁺ cells can generate neurons, oligodendrocytes, and astrocytes in vitro (Chittajallu et al., 2004), and a small percentage of cortical NG2⁺ cells express NeuN and other neuronal markers in situ (Dayer et al., 2005). In the present study, analysis of graft-derived cells demonstrated that, although cortical NG2⁺/CNP-hEGFR⁺ cells migrate similarly to wild-type SVZ NG2⁺ cells (Aguirre and Gallo, 2004), they only generate astrocytes and oligodendrocytes in OB, SCWM, and hippocampus but do not generate neurons. This finding is consistent with previous studies showing that overexpression of the EGFR promotes a gliogenic fate in neural progenitor cells (Burrows et al., 1997, 2000; Caric et al., 2001; Sun et al., 2005).

EGFR overexpression in SVZ progenitor cells might cause brain tumors, which may derive from brain progenitor/stem cells (for review, see Fomchenko and Holland, 2005). Oligodendrogliomas, which constitute 20% of glial tumors, and astrocytomas are correlated with increased EGFR signaling. In oligodendrogliomas, EGFR is frequently expressed at high levels; in astrocytomas, EGFR expression is associated with high-grade tumors and is associated with EGFR amplification or mutation (Libermann et al., 1985; Wong et al., 1992; Smith et al., 2001). When modeled in mice, only v-erb, a constitutively active allele, is transforming, whereas EGFR is transforming only in the presence of ligand (Nishikawa et al., 1994; Holland et al., 1998; Weiss et al., 2003). In our model, a wild-type EGFR allele unlikely to cause transformation is expressed, and ligand availability should determine when and where the receptor is activated (Ling et al., 2005). Supporting the lack of transformation by the wild-type EGFR allele, when we aged a cohort of 30 CNP-hEGFR mice, none developed brain tumors, and mice died with the same kinetics as wild-type mice (Ling et al., 2005). Furthermore, we did not observe tumor formation in host mouse brains up to 3-4 months after grafting NG2⁺/CNP-hEGFR⁺ cells.

In conclusion, our findings indicate that EGFR expression plays an important role in promoting migration of SVZ neural progenitors and that manipulation of EGFR expression can convert neural progenitors from a nonmigratory to a migratory phenotype in the postnatal brain. Our studies will not only help understanding how intrinsic properties and specific cellular environments cooperate to promote cell migration and differentiation, but will also lead to optimizing the use of neural progenitors for therapeutic approaches in the damaged brain.