EphA Receptors Direct the Differentiation of Mammalian Neural Precursor Cells through a Mitogen-activated Protein Kinase-dependent Pathway*

Ephrins are cell surface-associated ligands for Eph receptor tyrosine kinases and are implicated in repulsive axon guidance and cell migration. EphA2, 3, and 4 receptors and one of their cognate ligands, ephrin-A2, are expressed by cells in the subventricular zone and ganglionic eminence of the embryonic day 14.5 telencephalon and by neural precursor cells in vitro . Activation of EphA receptors in dissociated neural precursor cells in vitro facilitates the commitment to neuronal fates. The majority of ephrin-A1-induced neurons is immunoreactive for tyrosine hydroxylase. Blocking the signal by the extracellular domain of EphA in forebrain slices results in a decrease in neurogenesis. Extracellular signal-regulated kinase is activated by the ligand binding to EphA receptors and is involved in the neurogenesis through EphA receptors. Rap1, but not Ras, is activated in response to ephrin-A1. Our results identify EphA receptors as positive regulators of the mitogen-activated protein kinase pathway that exerts neurogenesis of neural precursor cells from the developing central nervous system. The mammalian central nervous system is derived from a monolayer of germinal neuroepithelial cells, which are composed of self-renewing multipotent precursor cells in the ventricular

The mammalian central nervous system is derived from a monolayer of germinal neuroepithelial cells, which are composed of self-renewing multipotent precursor cells in the ventricular zone. These most immature precursor cells generate mitotic and lineage-restricted intermediate progenitor cells. Fetal telencephalic neuroepithelial cells contain neural precursors that give rise to the neuronal lineage and the glial lineage. The fate of neural precursors in the developing brain is believed to be determined by intrinsic cellular programs and by external cues (1). Newly generated cortical neurons in the ventricular zone migrate along the surface of radial glial fibers and settle in the cortical plate to orderly layers of the cortex (2). After reaching its destination, each neuron develops a set of dendrites characteristic of its phenotype and a single long axon that extends along specific routes to reach prospective synaptic partners. The migration of either the entire neuron or its nerve growth cones is guided by the interaction between the neuron and its local environment. These navigations use similar guidance molecules.
Eph receptor tyrosine kinases/ephrins interactions are implicated in axon guidance, neural crest cell migration, establishment of segmental boundaries, and formation of angiogenic capillary plexi (3). Ephrins play important roles during axon guidance by providing a repulsive guidance signal to Eph receptors in cells. A migrating growth cone expressing a particular Eph receptors would turn away from a cell that presents the cognate ephrin ligand. Eph receptors and ephrins are divided into two subclasses, A and B, based on binding specificities. Ephrin subclasses are further distinguished by their mode of attachment to the plasma membrane; ephrin-A ligands bind EphA receptors and are anchored to the plasma membrane via a glycosylphosphatidylinositol linkage, whereas ephrin-B ligands bind EphB receptors and are anchored via a transmembrane domain. A recent study localized Eph receptors and ephrin ligands to the subventricular zone in the adult rat where neural stem cells reside (4). Activation of EphB by a 3-day infusion of the ectodomain of either EphB2 or ephrin-B2 into the lateral ventricle of the adult rat appeared to increase the number of neural stem cells, suggesting promotion of proliferation (5). On the other hand, in some non-neuronal cell lines, such as pRNS-1-1, PC-3, or MEF cells, stimulation with ephrin-A1-Fc inhibits Ras/mitogen-activated protein kinase (MAPK) 1 pathway, leading to cessation of proliferation (6). These findings prompted us to hypothesize that Eph/ephrin interactions are involved in the differentiation or maintenance of the neural stem cells in the developing central nervous system. In this manuscript, we show the presence of EphA receptors and its cognate ligand ephrin-A in the neural precursor cells in vitro and in vivo and that activation of EphA receptors alters the fate of neural precursor cells to a neuronal commitment.

EXPERIMENTAL PROCEDURES
Immunohistochemistry-Mouse embryos at embryonic day 14.5 were killed by anesthetic overdose and dipped in 4% paraformaldehyde in PBS (pH 7.2). Next, they were cryoprotected overnight in SPB composed of 0.1 M phosphate-buffered saline and 3% sucrose. The bodies were embedded in Tissue Tek O.C.T. compound (Sankura Finetek, Torrance, CA), followed by cryosection at 14 m. The following antibodies were used as primary antibodies: anti-EphA3 antibody (1:500; Santa Cruz), anti-EphA4 antibody (1:200; Santa Cruz), anti-ephrin-A1 antibody (1: 500; Santa Cruz), and anti-nestin antibody (Rat-401, 1:300; Developmental Studies Hybridoma Bank). Ephrin-A1-Fc or EphA2-Fc preincubated with the anti-human IgG-Fc antibody (Sigma) at a fixed ratio of 0.1ϫ indicated concentrations of Fc fusion proteins. Sections were permeabilized with 0.2% Triton X-100/PBS and then were incubated overnight with the primary antibodies, followed by incubation with corresponding secondary antibodies conjugated with FITC or Alexa fluor 568 for 1 h at room temperature. After rinsing with PBS, the sections were mounted with Immunon TM (Thermo Shandon) and viewed under a SM510-V2.01 laser confocal microscope (Carl Zeiss). The specificity of the antibodies was assessed with a competition study using a ratio of 10ϫ blocking peptide, EphA3P (Santa Cruz Biotechnology) or by leaving out the primary antibodies. Western blot analysis of cells expressing the proteins was also performed.
TUNEL Analysis-Analysis of apoptotic cells was carried out by employing DeadEnd TM Fluotometric TUNEL System (Promega). Briefly, the cells were immobilized to slides in freshly prepared 4% methanol-free formaldehyde solution in PBS (pH 7.4). The cells were Permeabilized by immersing the slides in 0.2% Triton X-100 solution in PBS for 5 min. After rinsing in PBS, excess liquid was removed by tapping the slides. The cells were covered with Equilibration Buffer and stayed for 5-10 min. Then the cells were added to TdT incubation buffer and incubated at 37°C for 60 min. The reaction was stopped by immersing slides in 2ϫ SSC for 15 min at room temperature. The cells were rinsed with PBS and observed under a microscope.
Slice Culture-The forebrains of mouse embryos at E14.5 were sagittally sliced manually on a glass plate. The slices were mounted in collagen gel as previously described (8) and placed in an incubator. The slices were kept in MHM on ice until mounting. The culture medium was modified neurosphere culture containing BrdUrd (10 g/ml) and 80 ng/ml EGF plus 80 ng/ml bFGF with or without soluble 100 g/ml EphA2-Fc or 100 g/ml EphB2-Fc (Genzyme Techne). After a 5-DIV culture in a well of a 6-well plate, the slices were embedded in Tissue Tek O.C.T. compound, followed by cryosection at 14 m, and sectioned with a virbratome slicing machine for immunohistochemistry.
The numbers of cells double-positive for BrdUrd and TuJ or microtubule-associated protein 2 were counted on sections taken from 80 ng/ml EGF plus 80 ng/ml bFGF with or without soluble 100 g/ml EphA2-Fc-treated slice culture (six different culture series of ϳ10 slice cultures each). Within each sample slice culture, several sections were analyzed. For each section at least three different grid-fields (10000 m 2 ) in SVZ were randomly selected for analysis.
Immunoprecipitation-The neurospheres cultured for 3 days were rinsed with PBS by centrifuging at 400 rpm for 5 min, and the medium was replaced with the same volume of MHM without mitogens. Twentyfour hours later, the neurospheres were exposed to 340 ng/ml human IgG-Fc or 500 ng/ml ephrin-A1-Fc for 0 min, 1 min, 5 min, 30 min, or 1 h. Then the cells were lysed in lysis buffer (1% Nonidet P-40, 10% glycerol, 50 mM Tris-HCl (pH 7.4), 200 mM NaCl, 2.5 mM MgCl 2 , 2 mM sodium orthovanadate, l g/ml leupeptin, 10 g/ml trypsin inhibitor). Protein concentrations between the cell lines were equalized. Either lysates were immunoprecipitated with anti-phosphotyrosine as previously described (23), or for cytoplasmic lysates, the extracts were mixed with an equal volume of 2ϫ SDS loading buffer.
Assay for ERK Activity-The neurospheres cultured for 3 days were rinsed with PBS by centrifuging at 400 rpm for 5 min, and the medium was replaced with the same volume of MHM without mitogens. Twentyfour hours later, the neurospheres were exposed to 340 ng/ml human IgG-Fc, 500 ng/ml ephrin-A1-Fc, or 500 ng/ml ephirn-B2-Fc for 0 min, 1 min, 5 min, 30 min, or 1 h. Then the cells were lysed in the lysis buffer. Western blot was performed with anti-MAPK (Sigma), anti-monophosphorylated ERK (Sigma), or anti-ERK (Sigma) antibodies.
Ras Activity Assay-Ras activity was assessed by employing the Ras activation assay kit (Upstate Biotechnology, Inc.). Briefly, treated cells were lysed in Mg 2ϩ lysis buffer (MLB buffer) containing 25 mM HEPES, pH 7.5, 150 mM NaCl, 1% Igepal CA-630, 10 mM MgCl 2 , 1 mM EDTA, 2% glycerol, 2 mM sodium orthovanadate, 1 mM ethylsulfonyl fluoride, and a protease inhibitor mixture tablet. The cell lysates were incubated with 20 l of Raf-1 Ras binding domain-agarose conjugate for 30 min at 4°C. The beads were washed with MLB buffer and resuspended in 2ϫ SDS sample buffer. Anti-Ras antibody (Upstate Biotechnology, Inc.) was used for the blotting.
Rap1 Activity Assay-Affinity precipitation of the activated Rap1 was performed with glutathione S-transferase-Ral binding domain fusion protein precoupled to glutathione-Sepharose (Amersham Biosciences). The Ral-binding domain of Ral GDS in pGEX vector was kindly provided by Dr. Johannes L. Bos. The activity of Rap1 was determined in immunoprecipitates by Western blot analysis using the anti-Rap1 antibody (BD Bioscience). To quantify the relative amounts of protein, we scanned x-ray film and analyzed the digital images using NIH Image.

EphA Receptors Are Expressed in Neuroepithelial Cells in the Embryonic Day 14.5 Telencephalon-To investigate whether
Ephrin/Eph receptors are associated with neurogenesis during the developmental stage, the expression patterns of EphA receptors and their cognate ligands were examined. The antibody RAT401, which recognizes nestin, an intermediate filament specific to undifferentiated neural precursor cells (7), was used to identify the neural precursor cells (8). At E14.5, immunoreactivities for EphA3 and 4 were observed in the ventricular zone and the adjacent mantle/intermediate zone (Fig. 1, A and  B). Double immunostaining revealed that EphA3 and 4 were colocalized with nestin, which was expressed on the fibers, in the ventricular zone of the telencephalon. Colocalization of other members, such as EphA2, with nestin was also found (data not shown). One of their cognate ligand, ephrin-A1, was weekly expressed in neuroepithelial cells at E14.5 (Fig. 1C). There is a little population that showed double-positive for ephrin-A1 (Fig. 1D, arrowheads) and EphA3 (Fig. 1D). Using Eph-A2-Fc or ephrin-A1-Fc fusion protein, the expression of ephrins and Eph receptors was verified (Fig. 1E). Binding of Eph-A2-Fc ( Fig. 1E-a) was weak, whereas that of ephrin-A1-Fc (Fig. 1E, panel b) was strong and widespread in the ventricular zone, suggesting predominant expression of Eph-A receptors. These results suggest that ephrin-A/EphA receptors signals have a role in neurogenesis in the developmental central nervous system.
Neural Precursor Cells from the E14.5 Telencephalon Express EphA Receptors-In vitro, single neural stem cells proliferate to form clonally derived floating sphere colonies, designated neurospheres, which contain cells that, upon dissociation into single cells, give rise to new sphere colonies. These cells that can differentiate into neurons and glia proliferate in the presence of bFGF and EGF (9). Sphere-producing cells derived from E14.5 mouse brain were isolated by primary neurosphere formation, and their self-renewing capacity was demonstrated by assessing the number of secondary neurospheres formed (8). Immunocytochemistry was performed using the primary spheres cultured for 7 days after dissociation. As expected, immunoreactivities for EphA3 and 4 were abundantly observed in the spheres and were colocalized with those for nestin ( Fig.  2A, panels b and c). Although neurospheres are heterogenous, and only a minority of the cells positive for these EphA immunoreactivities was nestin-positive, EphA2 was also found to be colocalized with nestin in the neurospheres (Fig. 2A, panel a). Using Eph-A2-Fc and ephrin-A1-Fc fusion proteins, the expression of ephrins and Eph receptors was verified (Fig. 2C). Binding of Eph-A2-Fc (Fig. 2C, panel a) or ephrin-A1-Fc (Fig. 2C, panel b) fusion protein to some of the nestin-positive cells was observed. One of the cognate ligands for EphA, ephrin-A1, was expressed weakly in a few cells in each sphere and was not colocalized with EphA2 (Fig. 2B, arrowheads). These results suggest the possibility that neural precursor cells expressing EphA receptor may be responsive to ephrin-A ligands.
Activation of EphA Receptors Did Not Cause Proliferation of Neural Precursor Cells-It was previously reported that infusion of EphB2 or ephrin-B2 into the lateral ventricle in the adult rat disrupted migration of neuroblasts and increased cell proliferation (5). The authors suggested that EphB/ephrin-B signaling is involved in the regulation of cell proliferation. In contrast, another paper showed nonmitogenic activity of ephrin-B (10). To determine whether EphA receptors maintained neural stem cells in an undifferentiated state, we stimulated EphA receptors by adding a fusion protein containing ephrin-A1 joined to human Fc to the cultures in vitro. Because the ephrin-Eph signaling system is able to elicit bi-directional signals, ephrin-A can have dual effects: the inhibition of endogenous ephrin-Eph interactions and the unidirectional activation of their respective binding partners. Dissociation of the neurospheres was done to eliminate the effects elicited by cell-cell interaction. Therefore, the addition of ephrin-A is expected to activate the EphA receptor while not inhibiting the ephrin-A reverse signal. These dissociated cells were transferred onto poly-L-ornithine-coated glass slides and cultured in the absence of EGF and bFGF to facilitate the differentiation. DAPI nuclear staining suggested that ephrin-A1-Fc did not increase the cell number and that cell death was apparently not induced by ephrin-A1-Fc (data not shown). Staining with the anti-nestin antibody revealed that the number of nestin-positive cells was gradually decreased between the 1-day and 5-day assay periods, demonstrating the differentiation of neural precursor cells under the condition we adopted (Fig. 3A). Cessation of cell proliferation was clear, from the total number of the cells stained by DAPI. The number of nestin-positive cells treated with ephrin-A1 for 1 or 5 days was not significantly changed compared with that without ephrin-A1 treatment (Fig. 3A), demonstrating that ephrin-A1 did not inhibit the differentiation of neural precursor cells under our experimental conditions. In contrast, ephrin-B treatment resulted in a significant increase in the number of nestin-positive cells, similar to a previous observation in vivo (5). On the other hand, ephrin-A or -B treatment had no effect on apoptosis (Fig. 3B). Thus, activation of EphA receptors may not contribute to proliferation or death of neural precursor cells derived from the E14.5 telencephalon.
Ephrin-A-induced Switch in the Fate of Neural Precursor Cells-Because activation of EphA receptors did not promote the differentiation of neural precursor cells in the undifferentiated condition that contained bFGF and EGF, we next determined whether ephrin-A has the potential to switch the fate of neural precursor cells. A thymidine analog, BrdUrd, was included in the culture medium, and the cells positive for Br-dUrd, TuJ1, or GFAP were counted in 1-and 5-day cultures of the precursor cells in the differentiated condition, which did not contain bFGF or EGF. This method allowed the direct examination of the possibility that ephrin-A ligands affect the cell fate by specific killing of the differentiated glia in the neurospheres. BrdUrd-unlabeled neurons or glia may be present from the beginning of the culture, and the number of these neurons or glia was not significantly altered by the addition of ephrin-A1 (data not shown). As shown in Fig. 4 (A and B), the ratio of the cells double-positive for TuJ1, a marker for immature neuron, and BrdUrd to cells positive for BrdUrd was significantly higher in ephrin-A1-treated cultures compared with untreated cultures. In addition, eprhin-A1 treatment increased the expression of neurofilament, a marker for mature neuron, positive cells. Next, we used immobilized ephrin-A1-Fc on poly-L-lysine-coated coverslips to mimic the cell-cell contact condition. Immobilized ephrin-A1-Fc induced neurogenesis to the same extent as soluble ephrin-A1-Fc (Fig. 4, A and B). As TuJ1-positive cells express the EphA3 receptor (Fig. 4C), these data suggest that ephrin-A promotes neurogenesis via an EphAdependent mechanism. In contrast, ephrin-A1 did not increase the number of GFAP-positive cells (Fig. 4, D and E). The ratio of the cells positive for both neuronal and BrdUrd markers to GFAP and BrdUrd markers was significantly increased with ephrin-A-Fc treatment for 5 days (Fig. 4E). The majority of the cells negative for TuJ1 as well as GFAP may be immature precursor cells, as much of these cells were also negative for oligodendrocyte markers. Our data show that ephrin-A1 appears to direct the neuronal fate of proliferating precursors.
Next, we examined whether ephrin-A induced specific neuronal lineages from the neuroepithelial cells contained in neurospheres. As in differentiation assay, immunocytochemical analysis of the characteristics of ephrin-A-induced neurons revealed that most of the neurons immunoreactive for TuJ1 express tyrosine hydroxylase (Fig. 5, A and B), although a significantly lower percentage of neurons in the control cultures were immunoreactive for tyrosine hydroxylase. We assessed expression of cholin acetyltransferase, dopamine ␤-hydroxylase, or glutamate decarboxylase and found that the values for these markers were much smaller in the ephrin-A-treated group than that for tyrosine hydroxylase (data not shown).
Reverse transcription-PCR analyses showed up-regulation of mRNAs for dopaminergic neurons, such as tyrosine hydroxylase, L-aromatic amino acid decarboxylase, and the midbrain dopaminergic neuron marker (Ptx-3) in 5-DIV ephrin-A1treated cells, whereas the band for vesicular monoamine transporter mRNA was not found (Fig. 5C). These results show that the generation of tyrosine hydroxylase-positive neurons was facilitated through EphA pathways.
Neurogenesis through an EphA-dependent Pathway in a Slice Culture of Telencephalon-We assessed the effects of endogenouse ephrin-A using forebrain slices from E14.5 mice. Mitotic cells were labeled by BrdUrd, and the neurogenesis was examined 5 days later. Extensive colocalization of BrdUrd with TuJ1, demonstrating the proper neurogenesis in the slices, was observed (Fig. 6A, arrowheads). Next, we employed the extracellular domain of EphA2 or EphB2 fused with Fc (EphA2-Fc and EphB2-Fc) to competitively block the endogenous interaction of EphA with ephrin-A or of EphB with ephrin-B. The number of the cells positive for TuJ1 was significantly decreased by the addition of excess EphA2-Fc (Fig. 6, arrowheads). However, EphA2-Fc had no effect on the differentiation to the GFAP-positive cells (Fig. 6B). In contrast, EphB2-Fc treatment did not result in decrease in the number of the cells positive for TuJ1 (Fig. 6A). These results also observed in clustered EphA2-Fc treatment (data not shown). Thus, EphA is required for only the differentiation of neural precursor cells into neurons in a slice preparation. These results strengthen our suggestion that is raised from the data using neurospheres.
Signaling Mechanisms of EphA-induced Neurogenesis-In some non-neuronal cell lines, such as pRNS-1-1, PC-3, or MEF cells, ephrin-A1-Fc stimulation inhibits the Ras/MAPK pathway, leading to cessation of proliferation (6). These reports prompted us to examine whether the biological activities of ephrin-A on neural precursor cells were attributable to Ras-MAPK activities, because we observed no growth inhibition by ephrin-A (data not shown), similar to the findings in fibroblast cells (6). To address this question, we examined ERK activation. At first, we confirmed that ephrin-A1 treatment caused tyrosine phosphorylation of EphA3 expressed on neural progenitor cells (Fig. 7A). Dissociated neuroepithelial cells were deprived of EGF and bFGF for 24 h, and then ephrin-A1-Fc, at a concentration of 500 ng/ml, was added. Rapid (1, 5 min) and sustained (until 60 min) activation of ERK was induced by ephrin-A1 (Fig. 7B), although no significant activation was seen in the human IgG-Fc-treated cells. In contrast, ephrin-B-Fc treatment abolished the basal ERK phosphorylation (Fig. 7B). Next, we asked whether MAPK activity was required for the biological effects of ephrin-A. PD098059, a specific inhibitor of MAPK kinase-1, completely blocked the ephrin-A effect, whereas PD098059 itself had no effect on the neurogenesis (Fig. 7C). Inhibition of the basal MAPK pathway increased the number of GFAP-positive cells (Fig. 7D) and did not affect the ratio of nestin-positive to DAPIpositive cells (data not shown). These results show that neurogenesis induced by ephrin-A is, at least partly, attributable to activation of the MAPK pathway.
We explored further the molecular mechanisms involved in the biological action. The activity of Ras, an upstream regulator of MAPK, was measured by a pull-down assay (10). Intriguingly, stimulation of neuroepithelial cells with ephrin-A1-Fc in vitro caused a mild decrease (5, 30, 60 min) in the amount of GTP-bound form of Ras compared with the basal activity (Fig.  8A). In contrast, ephrin-B-Fc as well as EGF activated Ras (data not shown). Our results are consistent with the previous report that shows the inactivation of Ras by ephrin-A in prostatic epithelial cells and endothelial cells (6).
Differentiation of PC12 cells in response to nerve growth factor is reported to be involved in two distinct pathways: Ras that induces the initial activation of MAPK and another small G protein Rap1 that induces sustained activation of MAPK (11). Because we observed sustained activation of ERK in the neuroepithelial cells, we hypothesized that Rap1 might be responsible for the activation of ERK. By employing a construct consisting of the Rap1-binding domain of Ral fused to glutathione-Sepharose, affinity precipitation was done to detect GTPbound active Rap1 (12). The amount of the GTP-bound form of Rap1 was increased after the addition of ephrin-A1-Fc (5, 30, and 60 min) (Fig. 8B). Because of the similarity of temporal activity change of ERK and Rap1, it is suggested that the Rap1/MAPK pathway contributes to the neurogenesis-inducing activity of EphA on the neural precursor cells. DISCUSSION In this manuscript, we show that the presence of EphA receptors and its cognate ligand ephrin-A in the neural precursor cells in vitro and in vivo and that activation of EphA receptors alters the fate of neural precursor cells to a neuronal commitment. Blocking the signal by the extracellular domain of EphA in forebrain slices results in a decrease in neurogenesis, suggesting that endogenous ephrin-A signal is required for the differentiation of the neural precursor cells. Extracellular signal-regulated kinase is activated by the ligand binding to EphA receptors and is involved in the neurogenesis through EphA receptors. Rap1, but not Ras, is activated in response to ephrin-A1.
Ephrin-A5-induced repulsive guidance was suggested to be mediated by activation of Rho (13), which plays major roles in neurite growth and growth cone guidance (14). In fact, the group led by Greenberg (15) identified a guanine nucleotide exchange factor for Rho, as an EphA-interacting protein, elucidating a molecular link between EphA receptors and reorganization of the actin cytoskeleton. Independently, others reported that inhibition of cell proliferation in prostatic epithelial cells and endothelial cells, but not fibroblasts, was mediated by ephrin-A1, which inhibits the Ras/MAPK pathway (6). In contrast, prolonged activation of ERK-1 and ERK-2 in response to ephrin-A5 was reported to contribute to the changes in cell morphology in NIH3T3 cells (16). In the neuroepithelial cells used in the current study, activation of ERK is necessary for the neurogenesis induced by the activated EphA. These seemingly contradictory findings suggest that the ligand binding to EphA receptors elicits bi-directional signals. The missing link between the intracellular domain of EphA receptors and MAPK should be elucidated to clearly explain the molecular events underlying these observations.
It should be noted that, whereas our data demonstrate that activation of MAPK is a necessary component of neurogenesis induced by EphA receptors, it remains to be determined whether activation of MAPK is sufficient for the effect. The rapid and potent activation of the MAPK cascade by EphA receptors is inconsistent with other receptor tyrosine kinases (17). Previous work has explored the influence of the diffusible factors that, by acting via receptor tyrosine kinases and recruiting the MAPK pathway, regulate the phenotypical potential of neural stem/precursor cells (1). For example, ciliary neurotrophic factor causes transient activation of MAPK that contributes to initiation of glial differentiation of the neural precursor cells from E14 rat (18). Taken together, EphA receptors might work on neural precursor cells through multiple signals that include MAPK to direct the differentiation of neural precursor cells.
Recently, the MAPK cascade was shown to play a role in axonal guidance. Netrin-1-mediated attraction of the growth cones requires activation of MAPK, which directly interacts with the receptor deleted in colorectal cancer (19). This finding may suggest a role for MAPK in growth cone guidance in general, because MAPK activation is known to stimulate cell motility by phosphorylating and activating myosin light chain kinase (20). In regard to the issue above mentioned, it would be interesting to assess whether netrin-1 regulates proliferation or differentiation of neural precursor cells. Although Ras seems not to be responsible for the activation of MAPK in ephrin-Atreated cells, Rap1 could be one of the factors that contribute to the activation of MAPK. It remains unclear yet whether Rap1 regulates ERKs (11) or not (21). Rap1 has been shown to be implicated in a particularly wide range of biological processes, from cell proliferation and differentiation to cell adhesion. During the course of the differentiation of PC12 cells induced by nerve growth factor, Rap1 as well as Ras is activated. Rap1 mediates B-Raf-mediated sustained ERK activation, whereas Ras elicits transient activation of ERK (11). Sustained activation of ERK, which is dependent on Rap, is reported to be the key for the very different outcomes: differentiation or proliferation. Therefore, the Rap1 signal may be an important factor for the new roles of EphA.
Recently, it was demonstrated that ephrin-A3 of astrocytes could induce spine retraction by stimulating EphA4 of neurons (22). Conover et al. (5) reported that ephrin-B of astrocytes stimulated proliferation of neuroblast. These findings suggest a possible involvement of glial cells in controlling the differenti-ation of neural precursor cells. Although EphA receptors are shown to be involved in the promotion of neurogenesis in the developmental stages, other numerous factors should affect the fate of the neural precursor cells. Uncovering the signaling mechanisms of these molecules will answer the question what determines the fate of the precursor cells such that some of them differentiate into neurons while others become glial cells.
Further work is required to address clearly the molecular mechanisms of ehrinA-EphA. Our work identifies ephrin-A/ EphA as possible regulators of the fate of neural precursor cells in the developing central nervous system.