Ephrin-A1-Mediated Dopaminergic Neurogenesis and Angiogenesis in a Rat Model of Parkinson's Disease

Cells of the neural stem cell lineage in the adult subventricular zone (SVZ) respond to brain insult by increasing their numbers and migrating through the rostral migratory stream. However, in most areas of the brain other than the SVZ and the subgranular zone of the dentate gyrus, such a regenerative response is extremely weak. Even these two neurogenic regions do not show extensive regenerative responses to repair tissue damage, suggesting the presence of an intrinsic inhibitory microenvironment (niche) for stem cells. In the present study, we assessed the effects of injection of clustered ephrin-A1-Fc into the lateral ventricle of rats with unilateral nigrostriatal dopamine depletion. Ephrin-A1-Fc clustered by anti-IgG(Fc) antibody was injected stereotaxically into the ipsilateral lateral ventricle of rats with unilateral nigrostriatal lesions induced by 6-hydroxydopamine, and histologic analysis and behavioral tests were performed. Clustered ephrin-A1-Fc transformed the subventricular niche, increasing bromodeoxyuridine-positive cells in the subventricular area, and the cells then migrated to the striatum and differentiated to dopaminergic neurons and astrocytes. In addition, clustered ephrin-A1-Fc enhanced angiogenesis in the striatum on the injected side. Along with histologic improvements, behavioral derangement improved dramatically. These findings indicate that the subventricular niche possesses a mechanism for regulating both stem cell and angiogenic responses via an EphA–mediated signal. We conclude that activation of EphA receptor–mediated signaling by clustered ephrin-A1-Fc from within the lateral ventricle could potentially be utilized in the treatment of neurodegenerative diseases such as Parkinson's disease.


Introduction
Self-renewal and differentiation of somatic stem cells are regulated by the stem cell environment (niche) via as yet undefined mechanisms [1]. Glial fibrillary acidic protein (GFAP)-positive cells give rise to neurons and interact directly with other cells in the subventricular zone (SVZ) and subgranular zone (SGZ). They are, therefore, regarded as neural stem cells [2,3]. Ependymal cells surrounding the lateral ventricle are in close proximity with these cells in the SVZ and produce stem cell-activating factors [4,5]. Blood vessel endothelial cells are another critical component of the SVZ [6]; dividing neural stem/progenitor cells (NSPCs) are tightly apposed to SVZ capillary vessels [7]. These findings suggest that GFAP-positive cells, ependymal cells, and capillary endothelial cells may function as niche cells.
Lesions to the brain initiate the proliferation of NSPCs and neurogenesis in the SVZ [8,9]. A major proportion of proliferating neuroblasts in the SVZ migrate to the olfactory bulb and become interneurons [10]. Because the striatum is closely associated with the SVZ, NSPCs there would be an ideal source of cellular replacements for the damaged striatum. In rats with lesions of the nigrostriatal pathway, the natural response of cellular proliferation in the SVZ is weak and disappears within a few weeks [9]. In this case, some NSPCs originating in the SVZ can migrate to the adjacent striatum, but few differentiate to neurons [9,10].
Ephrins signal via EphA and EphB receptor tyrosine kinases (forward signaling), and Eph receptors also transmit signals via ephrins (reverse signaling) [11]. EphAs bind to ephrin-As anchored to the cell membrane via a glycosylphosphatidylinositol linkage. EphBs bind to ephrin-Bs, which have a transmembrane domain and a short cytoplasmic domain. Eph/ephrin signals play important stimulatory and inhibitory roles in boundary formation, cell migration, repulsive axon guidance [12], and regulation of neuronal growth cone development [13]. They also regulate cellmatrix interactions [14,15,16] and cell proliferation [17,18]. Recent reports suggest that Eph receptors regulate angiogenesis in embryonic and adult tissues [19].
Neurogenesis is regulated by many factors, among which several ephrins and their Eph receptors play important roles. They are differentially expressed on distinct cell types of the neurogenic niche, and also have differential functions on stem cell proliferation, survival and differentiation. EphA2, EphA3 and EphA4 receptors are expressed in the SVZ, and involved in NSPC differentiation towards neuronal lineage in vitro [20]. EphA4 appears to be expressed in the adult neurogenic niches exclusively by neural stem cells, and function to maintain neural stem cell proliferation [21]. Complimentary expression of ephrin-A2 in transit-amplifying cells and neuroblasts and that of EphA7 on ependymal cells and neural stem cells (NSCs) inhibits neural progenitor proliferation through reverse signaling [22]. EphA7 is shown to be involved in the modulation of apoptosis of neural progenitors during embryonic development [23]. Infusion of the ectodomain of either EphB2 or ephrin-B2 into the lateral ventricle disrupts migration of neuroblasts and increase cell proliferation [24]. Ephrins B2 and B3 and their receptor EphB1 suppress proliferation and survival of NSPCs and migration of neuroblasts in the SVZ, rostral migratory stream (RMS) and SGZ [25,26]. EphB2 induces proliferation of SVZ cells in vitro [27]. EphB3 signal suppresses NSPC proliferation in a p53dependent manner [28]. Ephrin-B1 is shown to be critical for maintenance of NSPCs [29].
Fibroblast growth factor receptors (FGFRs) are also tyrosine kinases that attract and phosphorylate a variety of signaling proteins when activated, forming an assembly of signaling complexes that activates multiple signaling pathways [30]. FGF promotes proliferation of neural stem cells [31,32], and FGF activity has been implicated in the maintenance of stem cell niches in vivo [33]. We recently reported that interaction of activated EphA4 with FGFRs augments proliferative signaling via fibroblast growth factor receptor substrate 2a (FRS2a) [34,35] and that coactivated FGFRs stimulate migration-related signal through RhoA by phosphorylating Ephexin1 [36]. These findings suggested to us that there might be an intrinsic regulatory mechanism of ephrin-A/EphA-mediated signaling in the neural stem cell niche. In the present study, we show that ephrin-A/EphA-mediated signaling plays a role in neuronal regeneration in a rat model of Parkinson's disease. Preparation of Clustered Ephrin-A1-Fc, Clustered IgG(Fc), and FGF2

Unilateral Lesioning of the Nigrostriatal Dopaminergic Pathway
Male Sprague Dawley rats (7-9 weeks old) were injected stereotaxically with 2 ml saline containing 6 mg/ml 6-hydroxydompamine (6OHDA) and 0.2% ascorbic acid into the right medial forebrain bundle (coordinates: anterior/posterior 22.2 mm, lateral [right] 1.5 mm relative to bregma, dorsoventral 28.0 mm from the dural surface) under anesthesia. Rats were examined for typical rotational behavior after intraperitoneal apomorphine injection (0.5 mg/kg) 6-8 weeks after injection of 6OHDA. Rats that rotated on average more than 100 times in 20 minutes when measured in 3 successive 20-minute periods were regarded as having severe dopamine deficiency (unilaterally lesioned rats) [37]. These unilaterally lesioned rats were always used 6-8 weeks after lesioning with 6OHDA.

RNA Extraction and RT-PCR
Tissue facing the ventricle (,100 mm in depth) was microdissected under a stereomicroscope. RNA extraction was performed with TRI reagent (Sigma-Aldrich). Reverse transcription (RT) was performed with MuLV reverse transcriptase and an oligo d(T) 16 reverse transcription primer in 20 ml (Gene Amp RNA PCR kit; Roche, Indianapolis, IN). Polymerase chain reaction (PCR) was performed with 35 2-step cycles of 30 seconds at 94uC and 45 seconds at 58uC using 3 ml of the RT product and GoTaq polymerase (Promega, Madison, WI). The PCR products were fractionated on 2% agarose gels, and bands were stained with ethidium bromide and detected under ultraviolet irradiation. Amplified PCR products were confirmed to have the expected base sequence by DNA sequencing from both ends. Primers for reverse transcription polymerase chain reaction (RT-PCR) are listed in Table S1.

Immunohistochemistry
Rats were anesthetized and perfused with ice-cold PBS followed by 4% paraformaldehyde, and brains were removed and fixed in 4% paraformaldehyde overnight at 4uC. Brains were cryoprotected in a PBS series containing increasing concentrations (10%, 20%, 30%) of sucrose. Indirect immunohistochemistry was performed with 40-mm-thick, free-floating, coronal sections. For detection of bromodeoxyuridine (BrdU), sections were rinsed in PBS and incubated in formamide in 26 saline-sodium citrate (SSC) (50%, v/v) for 2 hours at 65uC followed by a 15-minute rinse in 26 SSC at 37uC. The DNA was then treated with 2 N HCl for 30 minutes at 37uC followed by a 10-minute rinse in 0.1 M boric acid buffer, pH 8.5. Then, the following steps were performed as routine immunohistochemical staining. Tissue sections were permeabilized with 0.25% Triton X-100 in PBS mixed with 1% normal rabbit serum and incubated with primary antibodies at 4uC for 48 hours. After a wash in PBS, they were incubated with second antibodies (DAKO) conjugated to horseradish peroxidase (for 3,39-diaminobenzidine staining) or Alexa Fluor 488 and Alexa Fluor 568 (Molecular Probes) at 4uC for 2 hours, followed by a wash in PBS. Sections were counterstained with 49,6-diamino-2-phenylindole (DAPI; Molecular Probes) or with Wheat Germ Agglutinin-Alexa Fluor 488 Conjugate (Molecular Probes) as necessary, and mounted with Gel/Mount (Biomeda Corp, Foster City, CA). Staining with 3,39-diaminobenzidine was detected with a Nikon Eclipse E800 microscope or a Keyence BZ-9000 microscope. Immunofluorescence was detected with a Keyence BZ-9000 microscope, Nikon Eclipse E800 fluorescence microscope, Nikon Eclipse TE300 inverted microscope, BioRad Radiance 2000 confocal microscope, or a Zeiss LSM5 Pascal confocal microscope. Antibodies used in this study were listed in Table S2.
When quantification is needed, we counted the number or measured the area of the marker-labeled cells in 8 areas (9610 4 mm 2 /area and 10 mm thickness) per animal from 4 equally spaced coronal sections of striatum or sagittal sections of the granule cell layer of the olfactory bulb, and the calculated average value was taken as representing the animal. The selected areas in the striatum were 500-800 mm lateral from the lateral wall of the lateral ventricle and around the center between the dorsal and ventral edge of the striatum in each section.

Immunoprecipitation and immunoblotting
Immunoblotting of Ephs was performed with brain tissue lysate that was taken from the brain region (100 mm thickness) surrounding the ventricles, and lysate containing 200-300 mg total protein was incubated with unclustered ephrin-A1-Fc or specific EphA antibodies followed by immunoprecipitation with protein A agarose. Lysis buffer contained 50 mM HEPES, 1% Triton X-100, 5 mM ethylenediaminetetraacetic acid, 50 mM NaCl, 10 mM sodium pyrophosphate, 50 mM sodium fluoride, 1 mM sodium orthovanadate, and protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 mM aprotinin, 1 mM leupeptin, and 1 mM pepstatin A). Immunoprecipitated proteins were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, blotted onto polyvinylidene fluoride membranes (Millipore, Billerica, MA), and incubated with EphAspecific antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) or mouse monoclonal antiphosphotyrosine antibody (clone 4G10; Millipore, Billerica, MA). Immunodetection was performed with an Immobilon Western Blotting Detection System (Millipore, Billerica, MA). Antibodies used in this study were listed in Table S2.

BrdU Labeling
BrdU (20 mg/ml) dissolved in PBS was injected into the peritoneal cavity (80 mg/kg) at intervals of several hours, as indicated.

Stereologic Cell Counting
Stereologic analysis was performed under blinded conditions on coded slides. For each rat, we analyzed the entire striatum on the lesioned and reagent (ephrin-A1-Fc or IgG[Fc])-infused side. Numbers of BrdU(+) cells were determined in every tenth section in a series of 40 mm coronal sections throughout the rostrocaudal extent of the striatum with a semiautomatic stereology system (Stereoinvestigator; MicroBrightField, Williston, VT) and a 56 objective to trace the striatum. Volume was determined by summing traced areas of each brain and multiplying by intersection distance (400 mm). Our criterion for selecting individual BrdU(+) nuclei was the presence within the counting frame or touching the right or top frame lines but not touching the left or bottom lines.

Statistical Analysis
All the values were expressed as mean 6 SD. Comparison of 2 values was performed with Student t-test, or with a nonparametric analysis, Mann Whitney test. Comparison of $3 values was performed by one-way analysis of variance followed by Tukey multiple comparison test.

Expression of EphA Receptors, FGFRs and their downstream molecules in Adult Subventricular Tissue and Activation of EphA4 by Clustered Ephrin-A1-Fc
To examine expression of the molecules related to the ephrin/ Eph signal transduction pathway, we extracted RNA from the microdissected subventricular area (100 mm thickness from the surface of the ventricles), and performed RT-PCR using the primers shown in Table S1 (Fig. 1A). All Epha mRNAs and those for ephrin-A2, -A3, and -A5 (Efna2, Efna3, and Efna5) were expressed in the SVZ, but those for ephrin-A1 and -A4 (Efna1 and Efna4) were barely detectable or undetectable. mRNAs for Fgfrs, Frs2a, and Ephexin1 were also expressed.
To investigate which EphA receptors possibly bind to the recombinant ephrin-A1-Fc, we incubated the tissue lysate from the subventricular area with unclustered ephrin-A1-Fc, followed by immunoprecipitation with protein A-agarose. Only EphA4 was co-immunoprecipitated and detectable (Fig. 1B). These findings suggest that the clustered ephrin-A1-Fc used in the current study predominantly binds to EphA4 in the SVZ. We obtained the same results using the microdissected subventricular tissue from normal rats and rats with unilaterally lesioned nigrostriatal dopaminergic pathway described below. Studies were repeated 3 times in both normal and lesioned rats, and showed the same results.
To examine the effect of ephrin-A1-Fc on the cells of SVZ, we used rats with a unilaterally lesioned nigrostriatal dopaminergic pathway because the evaluation of striatal dopamine depletion and repletion is standardized. The nigrostriatal dopaminergic pathway of rats was lesioned 6-8 weeks before intraventricular injection of clustered ephrin-A1-Fc or other reagents. Clustered or unclustered ephrin-A1-Fc, clustered IgG(Fc), or FGF2 was injected stereotaxically into the lateral ventricle on the lesioned side. Immunoblotting with anti-phosphtyrosine antibody 18 hours after injection showed that clustered ephrin-A1-Fc phosphorylated EphA4, however, FGF2, unclustered ephrin-A1-Fc or clustered IgG(Fc) did not (Fig. 1C). The effect was limited to the lesioned side where clustered ephrin-A1-Fc was injected.

Effect of Clustered Ephrin-A1-Fc in Rats with Unilateral Nigrostriatal Pathway Lesions
To roughly locate the area of brain that responds to ephrin-A1-Fc treatment, we injected clustered ephrin-A1-Fc into the striatum close to the SVZ. Injection of clustered ephrin-A1-Fc (3 mg) induced the appearance of tyrosine hydroxylase (TH)-positive (+) neuronal fibers on the side of the ventricle 4-12 weeks after injection, but not on the striatal side ( Fig. 2A). Intraventricular injection of clustered ephrin-A1-Fc ipsilateral to the lesion increased the number of TH(+) neurons on the lesioned side of the striatum. These were clearly localized as multiple islands in the ephrin-A1-Fc-injected, lesioned side of the striatum 4-12 weeks after injection, although they were not yet distributed as densely as on the normal side ( Fig. 2B; Fig S1). We confirmed similar intrastriatal and intraventricular effects of clustered ephrin-A1-Fc in 2-3 rats at each time point after injection (4, 6, 8, and 12 weeks).

Behavior of Rats with Unilateral Dopamine Depletion before and after Treatment with Clustered Ephrin-A1-Fc
Rotation frequency after injection of clustered IgG(Fc) or ephrin-A1-Fc was analyzed before and 6 weeks after treatment in the rats treated with a single injection, and before and 4-12 weeks after treatment in the rats treated with a 7-day infusion. A single intraventricular injection of clustered ephrin-A1-Fc (3 mg) significantly decreased rotation frequency to about 60% of that of IgG(Fc) injection at 6 weeks after treatment, (p,0.01, n = 7) (Fig. 3A). A 7-day continuous ventricular infusion of clustered ephrin-A1-Fc (2 mg/day) also decreased rotation frequency at 8 and 12 week points, to 53% and 40%, respectively, as compared to the value before infusion (p,0.01, n = 5) (Fig. 3B).

Distribution of BrdU(+) Cells from the Ventricular Side to the Striatum after Intraventricular Injection of Clustered Ephrin-A1-Fc
To study possible distribution of BrdU(+) cells along the SVZstriatum axis, rats were treated with a single intraventricular injection of clustered ephrin-A1-Fc (3 mg) to the lesioned side of the lateral ventricle, followed by 3 intraperitoneal injections of BrdU (80 mg/kg) at 6-hour intervals, and killed 1, 7, or 14 days after the ephrin injection. Coronal brain sections were stained for nuclear BrdU incorporation. Cells on both sides of the SVZ showed extensive BrdU incorporation 1 day after ephrin injection (6 h after the last BrdU injection). The BrdU(+) cells ipsilateral to the ephrin injection side, which is also the lesioned side, showed dense BrdU(+) cell populations moving toward the periphery of the lesioned striatum from the SVZ over 14 days (moving front shown by vertical lines), whereas BrdU(+) cells did not migrate to the striatum on the nonlesioned and non-ephrin-injected sides (Fig. 4A, B). Unilaterally lesioned rats injected with clustered IgG(Fc) followed by intraperitoneal BrdU injection did not show such effect (Fig. 4C, D). These studies were repeated in 3 rats for each treatment and each time point, and showed similar results. A possible diffusion of ephrin-A1-Fc into the striatum was also examined immunohistochemically in the brain sections using fluorescence-labeled anti-human IgG(Fc) antibody. The results showed no detectable staining (data not shown).

Tracking of Migrating BrdU(+) Cells with CM-DiI from the Subventricular Zone to the Striatum and Olfactory Bulb
To study whether clustered ephrin-A1-Fc increased the number of the BrdU(+) cells in the striatum, and to compare the effect with that of unclustered ephrin-A1-Fc or FGF2, the unilaterally lesioned rats were infused with clustered IgG(Fc) (3 mg/day), unclustered ephrin-A1-Fc (3 mg/day), clustered ephrin-A1-Fc (3 mg/day), or FGF2 (50 ng/day) into the lateral ventricle of the lesioned side continuously for 1 week with simultaneous intraperitoneal injection of BrdU (80 mg/kg) twice a day (every 12 hours). Brains were removed 6 weeks after the start of infusion, sliced into 40-mm thick coronal sections, and stained for BrdU. We counted the number of BrdU(+) cells in 8 areas (9610 4 mm 2 ) per animal in the striatum and in the granule cell layer of the olfactory bulb as described in the Materials and Methods section ( Fig S2). In the striatum, clustered ephrin-A1-Fc, but not unclustered ephrin-A1-Fc or FGF2, increased the number of BrdU(+) cells by ,4-fold over clustered IgG(Fc) treatment. In the olfactory bulb, the normal target of subventricular neuroblasts migration [38], clustered ephrin-A1-Fc and FGF2, but not unclustered ephrin-A1-Fc, increased the number of BrdU(+) cells significantly over clustered IgG(Fc) treatment. To clarify if the increase of BrdU(+) cells was extended to all over the striatum, we performed stereologic counting to quantify precisely the number of BrdU(+) cells throughout the striatum on the treated side after infusion of clustered ephrin-A1-Fc or IgG(Fc) and found that clustered ephrin-A1-Fc increased the number significantly (,2.5-fold) over IgG(Fc) treatment (Fig. 5A).
Then, to study whether the ephrin-A1-Fc-induced distribution of BrdU(+) cells in the striatum was caused by migration of NSPCs from the SVZ, we injected CM-DiI into the lesioned side of the lateral ventricle right before infusion of clustered ephrin-A1-Fc and covalently labeled the cells exposed to the ventricular surface with the fluorescent dye. Brains were removed 6 weeks after the start of ephrin-A1-Fc infusion, sliced into 40-mm thick coronal sections, and stained for BrdU. Brain slices were stained for BrdU and with Wheat Germ Agglutinin conjugated with Alexa Fluor 488 (Molecular Probes), and confocal 3D micrographs were taken at 1-mm intervals. Then, ten serial confocal micrographs were compiled for one all-in-focus micrograph, and the number of cells labeled with BrdU or co-labeled with both BrdU and CM-DiI was counted in selected areas. When cells 500-800 mm lateral from the ventricular surface were examined 6 weeks after the start of clustered ephrin-A1-Fc infusion, the ephrin infusion enhanced almost 4 times the number of double-labeled cells over infusion of clustered IgG(Fc) (Fig. 5B). Unclustered ephrin-A1-Fc did not increase the number of double-labeled cells beyond that induced by IgG(Fc) (data not shown). In the granule layer of the olfactory bulb, we also found that the number of double-labeled cells significantly increased after treatment with clustered ephrin-A1-Fc (Fig. 5C). However, the ratio of BrdU(+)CM-DiI(+) cells over BrdU(+) cells, which is around 1/3, stayed almost same in treatment with clustered IgG(FC) or ephrin-A1-Fc in both striatum and olfactory bulb. The relative localization of CM-DiI and BrdU in a cell level was clearly shown by staining with labeled Wheat Germ Agglutinin that binds to cell surface carbohydrates (Fig. 5D). The results indicate that many BrdU(+) cells were labeled with CM-DiI at their surface.

Transformation of Subventricular Niche Cells
In rats that received 7-day continuous infusion of clustered ephrin-A1-Fc (3 mg/day) into the lesioned side of the lateral ventricle together with intraperitoneal 7-day BrdU injection (80 mg/kg twice a day), the infused side showed many BrdU(+) cells, densely distributed throughout the striatum and with high localization in the SVZ (Fig S3). In the ephrin-A1-Fc infused side of the brain the SVZ became thicker, and many BrdU(+) cells in the SVZ and the adjacent area were also positive for GFAP (Fig. 6A), suggesting that they are possibly adult neural stem cells capable of further neuronal differentiation. However, it cannot be excluded that GFAP (+) cells are astrocytes [2].
We counted the number of GFAP(+) cells in the striatum 500-800 mm lateral from the ventricular surface in 8 microscopic areas after treatment. Clustered ephrin-A1-Fc increased the number of BrdU(+) cells as well as the percentage of GFAP(+)BrdU(+) cells over control (IgG[Fc]) (Fig. 6B). In the same study, unclustered ephrin-A1-Fc or FGF2 did not increase the BrdU(+) or GFAP(+) cells. Microglial cells were also examined immunohistochemically by staining for ionized calcium-binding adapter molecule 1 (Iba1). These consistently represented 32% to 35% of BrdU(+) cells in response to any treatment (data not shown), suggesting that they presumably appeared in response to 6OHDA damage [39].
The same SVZ cells on the lesioned side were examined by triple-staining for GFAP, CD24, and BrdU (Fig. 6C). When clustered IgG(Fc) was infused, the lesioned side of the SVZ showed a thin layer of GFAP(+) cells beneath a monolayer of CD24positive ependymal cells facing the lateral ventricle. However, clustered ephrin-A1-Fc induced thickening of the layer of GFAP(+) cells without affecting the monolayer of ependymal cells. We defined the width of SVZ as the thickness of GFAP staining beneath the CD24 positive ependymal cells on the striatal side, and measured the width in coronal sections at the middle of antero-posterior and dorso-ventral axes. Rats infused with clustered IgG(Fc) and ephrin-A1-Fc exhibited the width of 13.363.1 and 3564.5 mm, respectively (mean 6 SD; n = 6; p,0.001).
We also characterized the BrdU(+) cells in and around the SVZ using neural progenitor cell markers. As shown in Fig. 6D, E, and F, many BrdU(+) cells stained positive for MASH1, a marker for transit-amplifying progenitor cells [40,41], and for Doublecortin (DCX), a marker for neuroblasts, or Nestin, a marker for neural stem cells and transit-amplifying progenitor cells. They are present not only in the SVZ but also in the adjacent striatum, supporting the migration scheme of BrdU(+) cells from the SVZ to the striatum.

Differentiation of BrdU(+) cells to Neurons after Intraventricular Infusion of Clustered Ephrin-A1-Fc
To study whether BrdU(+) cells differentiate to neurons, we counted the numbers of BrdU(+) cells and those of cells doublelabeled for NeuN (neuronal nuclei) and BrdU 4 weeks after a 1week infusion of clustered ephrin-A1-Fc in striatal sections of the unilaterally lesioned rats (Fig. 7A; Fig S4). Clustered ephrin-A1-Fc increased the number of BrdU(+) cells on the lesioned, infused side of the striatum significantly more than 2-fold compared to the lesioned, clustered IgG(Fc)-infused side (p,0.01; 444.8683.5 vs. 186.8648.9; n = 6). Clustered ephrin-A1-Fc did not affect the number of BrdU(+) cells on the contralateral side of the striatum in the same rats. The Importantly, some of the BrdU(+) cells were also TH(+) (Fig. 7B). These were distributed most densely in the striatum close to the anteroventral region of the clustered ephrin-A1-Fc-infused lateral ventricle. Three dimensional confocal imaging of 6 areas

Vascular Development after Intraventricular Infusion of Clustered Ephrin-A1-Fc
Ephrin-A1 induces angiogenesis in tumors as well as in embryonic and adult tissues [19]. To study whether the same effect of ephrin-A1 can be detected in the brain, clustered ephrin-A1-Fc (3 mg/day) was infused into the lesioned side of the lateral ventricle in the unilaterally lesioned rats. Brain taken from the rats 6 weeks after the start of infusion were sectioned coronally and stained for BrdU and Rat Endothelial Cell Antigen-1 (RECA-1). Treatment with clustered ephrin-A1-Fc increased BrdU(+) cells and enhanced the percentage of BrdU(+) RECA-1(+) cells (Fig. 8A). In the striatum of clustered ephrin-A1-Fc-infused rats, GFAP(+) astrocytic cells were juxtaposed to RECA-1(+) endothelial cells (Fig. 8B; Fig S5), and the endothelial cell area measured as the area of RECA-1 staining was almost twice that of clustered IgG(Fc)-injected rats (Fig. 8C). In the experiment using intraventricular injection of CM-DiI right before infusion of clustered ephrin-A1-Fc, some endothelial cells were double-labeled with CM-DiI and BrdU, suggesting that they were derived from the cells once facing the lateral ventricle, and presumably migrated to the striatum (Fig. 8D).

Discussion
The present results indicate that subventricular cells are induced to increase their number, migrate to the striatum, and differentiate in the striatum after injection of clustered ephrin-A1-Fc into the lateral ventricle. A recent report indicates that normal neural stem cells (NSCs, B1 cells) in the SVZ possess a monociliated process that directly contacts the ventricular lumen [42]. A cluster of these apical processes is surrounded by ependymal cells. The cilia of the ependymal cells (E1 and E2) encircling the B1 stem cell processes may serve to concentrate soluble factors in the cerebrospinal fluid and induce contact of these factors with the apical processes of B1 cells. Similar signaling via primary cilia of stem cells is reported with hedgehog signals in the SGZ [43]. Alternatively, ependymal cells surrounding B1 cells might mediate signal to NSCs. However, it is less likely that clustered ephrin-A1-Fc diffuses through the ependymal layer into the brain tissue as it has a large mass, even if ependymal cells do not have tight junctions [44]. In either case, intraventricular infusion of clustered ephrin-A1-Fc appears to transform the subventricular stem cell niche, enlarging the width of the subventricular zone without affecting the ependymal layer. We have also shown that cells sensitive to clustered ephrin-A1-Fc are localized to the SVZ; the recombinant ephrin-A1-Fc affected only the ventricular side when injected into the striatum close to the ventricle. These findings strongly support that the subventricular NSPCs are the major cells affected by injection of clustered ephrin-A1-Fc into the lateral ventricle.
Recent studies show complicated results on the function of ephrins and Ephs in neurogenesis. Ephrins A2 and A5, and their receptor EphA7, regulate neural stem cell survival and proliferation in the embryonic telencephalon [23]. Ephrin-A2 reverse signaling mediates antiproliferative effects in the adult SVZ [22].
Endogenous ephrin-A2 and -A3 expressed in astrocytes appear to form an inhibitory niche that negatively regulates neural progenitor cell proliferation in adult mammalian brain areas other than the SGZ of the hippocampus and the SVZ [45]. On the other hand, EphA4 is shown to be important to maintain postnatal and adult neural stem cells in vivo [21], and EphA2, A3 and A4 appears to be important for neuronal differentiation [20]. Some ephrin-Bs and EphBs appear to be involved in suppression of NSPC proliferation in the neurogenic niche [25,26,28], whereas others are involved in proliferation or at least maintenance of NSPCs [24,27,29]. Of course, different ephrins might have different functions in vivo. However, at least a part of these complicating effects could be explained by the experimental system used for each study. We suspect that our strategy of ectopic application of clustered ephrin-A1-Fc from inside the ventricle, in which clustered ephrin stays out of the tissue and affects only from the side of cerebral fluid, might differ from endogenous expression or deletion via genetic engineering in that it elicits forward signaling with barely affecting the reverse signaling. However, we cannot completely exclude the possibility that some clustered ephrin diffuses into the SVZ tissue, and stimulates the forward signaling and blocks the reverse signaling.
Pulse-chase experiments and CM-DiI labeling from the ventricular side demonstrated that BrdU(+) cells migrated extensively from the SVZ to the striatum after injection of clustered ephrin-A1-Fc into the lesioned side of the lateral ventricle. Staining of the cells with several cell markers, Nestin  [46], MASH1 [40], and DCX [47], also support that the migrating cells in and around the SVZ are neural stem cells, transit-amplifying cells and neuroblasts. Migration occurred on the lesioned side of the striatum only after ipsilateral intraventricular injection. The normal side was barely affected unless ephrin was injected into this side of the lateral ventricle. In addition to the above-mentioned signaling mechanisms via stem cell cilia or ependymal cells [42], increased sensitivity of cells residing on the damaged side might play a role in this effect [8,9]. Migration of BrdU(+) cells via the default rostral migratory stream, as shown in our control studies, was also augmented heavily by clustered ephrin-A1-Fc and moderately by FGF2, suggesting that the effect of clustered ephrin-A1-Fc on striatal migration is not a nonspecific phenomenon. The molecular mechanisms of cellular migration involve cytoskeletal changes induced by a dynamic structural shift of actin microfilaments and microtubules in response to external signals [48]. We suspect that reorganization of microfilaments mediated by RhoA activation following complex formation of FGFR, EphA4 and ephexin1, which depends on EphA4 activation, is at least partially involved in the mechanism of stem cell migration into the striatum [13,36]. In this case, FGFR is responsible for activation of the Rho family guanine nucleotide exchange factor, ephexin1. We have shown that all four classes of FGFRs and ephexin1 are also expressed in the SVZ, and EphA4, a predominant Eph that binds to recombinant ephrin-A1-Fc in the SVZ, is phosphorylated in the SVZ cells of rats intraventricularly infused with clustered ephrin-A1-Fc. Examining this way of signaling, we performed injection of SU5402, an FGFR inhibitor, into the lateral ventricle to test if the function of clustered ephrin-A1-Fc can be blocked. We have preliminary data to be completed showing that the inhibitor suppresses not only migration of BrdU(+) cells into the striatum but also recovery from the typical rotational behavior induced by apomorphine injection in the rats with a unilaterally lesioned nigrostriatal dopaminergic pathway.
Injection of clustered ephrin-A1-Fc into the lateral ventricle induced histologic and behavioral improvements in the unilaterally lesioned rats. A single injection (3 mg) into the lesioned side of the lateral ventricle improved apomorphine-induced rotating behavior significantly. Continuous infusion for 1 week with a dose of 2 mg/ day produced even greater improvements. The effect of ephrin injection on these histologic and behavioral improvements was evident 12 weeks after ephrin injection. This behavioral recovery is most likely to be caused by differentiation of NSPCs to astrocytes and dopaminergic neurons. Regeneration of damaged but surviving residual dopaminergic fibers might be induced by infusion of clustered ephrin-A1-Fc into the lateral ventricle, and had some functions in the behavioral improvement. However, we waited for 6 weeks after 6OHDA injection, and confirmed almost complete depletion of dopaminergic neural fibers in the striatum right before ephrin-A1-Fc infusion. We found that a major proportion of BrdU(+) cells migrating through the striatum differentiated to astrocytes when stimulated by clustered ephrin-A1-Fc from inside the lateral ventricle, and that differentiation of BrdU(+) cells to dopaminergic neurons also occurred in the striatum. Differentiation of subventricular NSPCs to dopaminergic neurons by ephrin-A ligands has been reported in an in vitro study, and appears to be mediated by MAP kinase activation [20]. Astrocytes generally interact with neurons directly [49] and play an important role in supporting neurons. In the RMS, chains of migrating adult neuroblasts are ensheathed by astrocytes [50]. They express cytokines and cytokine receptors; other factors that act as neuroprotective agents of damaged neurons [51,52]; constituents of the niche microenvironment [53]; differentiation/maturation factors [54]; and synaptogenic factors [54]. Thus, newly differentiated astrocytes in response to clustered ephrin-A1-Fc injection are suspected to play a critical role in differentiation of neural stem cells to dopaminergic neurons and their maintenance in the striatum.
Another finding is that clustered ephrin-A1-Fc increased angiogenesis in the striatum on the injected side. Ephrins and their Eph receptors are known to regulate angiogenesis [19]. Originally, B-class receptors and ligands were considered important players in endothelial cell migration and differentiation, leading, in concert with other growth factors, to capillary formation. Recent studies demonstrate the additional involvement of ephrin-A1 and EphA2 in postnatal and tumor angiogenesis [55]. Intraventricularly injected clustered ephrin-A1-Fc likely increased capillary formation via binding to receptors on ependymal cells or B1 neural stem cells that directly contact the ventricular fluid, as subventricular capillary vessels locate far from the ventricular surface [7,42]. In either case, signaling for endothelial cell proliferation appears to be indirect. However, the finding of endothelial cells double-labeled with BrdU and CM-DiI suggests a possibility that neural stem cells differentiate to endothelial cells. This kind of transdifferentiation has been reported in neural stem cells in vivo [56]. Subventricular capillary vessels, once formed, are permeable to small molecules; this facilitates access of stem cells and transit-amplifying cells to molecules in the bloodstream such as growth factors, hormones, and nutrients [7]. Ephrin-A1 further increases this vascular permeability upon stimulation of its receptor EphAs as demonstrated in the lung [57]. Thus, clustered ephrin-A1-Fc appears to stimulate angiogenesis by inducing stem cell proliferation in the SVZ. In turn, the enhanced angiogenesis would help increase the stem cell proliferation as well.
In conclusion, injection of clustered ephrin-A1-Fc into the lateral ventricle induced transformation of the subventricular niche, resulting in increase of BrdU(+) NSPCs in and around the SVZ, their migration to the striatum, their differentiation to astrocytes and neuronal cells, including dopaminergic cells, and angiogenesis. Owing to this dopaminergic regeneration supported by increasing numbers of astrocytes and capillary vessels in the striatum, behavioral abnormalities caused by lesion to the nigrostriatal pathway decreased dramatically. These findings may lead to the development of new therapeutic approaches for neurodegenerative diseases such as Parkinson's disease and related disorders in humans.  Figure S5 Effect of clustered ephrin-A1-Fc on vascular formation in the rat striatum. Clustered ephrin-A1-Fc was injected into the lesioned side of the lateral ventricle in the unilaterally lesioned rats. Brains taken 6 weeks after injection were sectioned coronally and stained for GFAP (green) and RECA-1 (red) and with DAPI (nuclei; blue). The rectangular insets are shown in Fig. 8B. Scale bar: 100 mm.