Abstract
We examined the cell proliferative, neurogenic, and behavioral effects of transforming growth factor α (TGFα) in a 6-OHDA Parkinson's disease model when compared with naive rats. Intrastriatal TGFα infusion induced significant proliferation, hyperplastic nodules, and substantial migratory waves of nestin-positive progenitor cells from the adult subventricular zone (SVZ) of dopamine-denervated rats. Interestingly, SVZ cells in naive rats displayed proliferation but minimal migration in response to the TGFα infusion. The cells in the expanded SVZ accumulated cytoplasmic β-catenin, indicating activation of classical Wnt signaling. However, no evidence of any neuronal differentiation was found of these recruited progenitor cells anywhere examined in the brain. Consequently, no evidence of dopaminergic (DA) neurogenesis was found in the striatum or substantia nigra in any experimental group, and amphetamine-induced behavioral rotations did not improve. In summary, the cells in the TGFα-induced migratory cellular wave remain undifferentiated and do not differentiate into midbrain-like DA neurons.
Introduction
Recent studies of adult neurogenesis have raised the awareness of potential intrinsic mechanisms to regenerate lost neurons caused by trauma or neurodegenerative disease. Remarkably, one study (Fallon et al., 2000) claimed that intrastriatal infusion of transforming growth factor α (TGFα) enables endogenous adult stem cells to generate dopaminergic (DA) neurons in the striatum of a DA denervation rat model of Parkinson's disease (PD). These morphological findings were associated with a reduced behavioral response to the dopamine agonist apomorphine, potentially reflecting reduced dopamine receptor supersensitivity (Fallon et al., 2000). TGFα is a member of the epidermal growth factor (EGF) family of proteins and is a ligand for the EGF receptor (EGFr) family of tyrosine kinase receptors (Todaro et al., 1980). The expression of TGFα is temporally and spatially coordinated with EGFr expression, which is detected in neurogenic regions such as the adult subventricular zone (SVZ) and the prenatal ventral midbrain (Seroogy et al., 1995; Kornblum et al., 1997; Doetsch et al., 2002). TGFα null mice have fewer DA neurons in the substantia nigra pars compacta (SNc) (A9) region of the ventral midbrain prenatally, indicating a possible mitogenic effect of TGFα on the cells derived from the neuroectoderm (Blum, 1998). Adult neurogenesis of DA neurons has also recently been claimed in the substantia nigra in a study of 6-hydroxydopamine-denervated and naive rats (Zhao et al., 2003). Both Fallon et al. (2000) and Zhao et al. (2003) propose that their findings could be relevant to cell therapy for PD based on endogenous cell proliferation. The multipotent stem cell within the SVZ expresses glial fibrillary acidic protein (GFAP), a typical marker of an astrocyte (Doetsch et al., 1999a,b; Laywell et al., 2000). The transit amplifying progenitor (TAP) cells generated from these SVZ astrocytes respond to EGF, which induces their dedifferentiation into a GFAP-expressing multipotent stem cell (Doetsch et al., 2002). In turn, TAP cells generate neuroblasts that form rostrally migrating chains within the SVZ surrounded by astrocytes and TAP cells (Wichterle et al., 1997). In rodents, but apparently not patients (Sanai et al., 2004), these migrating chains converge at the anterior tip of the SVZ, forming the rostral migratory stream. The regulation of neurogenesis within the SVZ has been linked to several conditions and molecules. For example, cell-to-cell contact mechanisms via ephrin-Eph interactions regulate SVZ neurogenesis and neuroblast chain migration, whereas TGFα null mice exhibit decreased proliferation within the SVZ (Tropepe et al., 1997; Conover et al., 2000). In vitro experiments have shown that SVZ-derived progenitor cells can be expanded by EGF and fibroblast growth factor (FGF) administration (Reynolds and Weiss, 1992). Similarly, the in vivo administration of EGF, FGF2, and TGFα via intraventricular cannulas from osmotic pumps has altered neurogenesis in the adult brain (Craig et al., 1996; Kuhn et al., 1997; Alonso et al., 1999; Wagner et al., 1999; Fallon et al., 2000; Decker et al., 2002). In the present experiments, we examine the potential for DA neurogenesis after TGFα infusion.
Materials and Methods
Naive (n = 8) and adult male Sprague Dawley rats with medial forebrain bundle-administered 6-OHDA lesions previously tested by amphetamine rotation (Bjorklund et al., 2002) (n = 17) were obtained from Taconic Farms (Germantown, NY). The animals were maintained in accordance with current National Institutes of Health guidelines and McLean Hospital-Harvard University Institutional Animal Care and Use Committee protocols.
TGFα infusions. Rats received infusions for 1, 2, or 4 weeks. The infusions of TGFα were obtained through an intrastriatal cannula (Brain Infusion Kit II; Alzet, Cupertino, CA) (+1.2 anteroposterior; +2.7 mediolateral) from a surgically inserted interscapular osmotic pump (Osmotic Pump 2004; Alzet). The osmotic pumps delivered either 100 μg of TGFα (Stem Cell Pharmaceuticals, Seattle, WA) in 200 μl of 0.1 m PBS or only PBS vehicle, both at an infusion rate of 0.25 μl/hr. These parameters were identical to those used by Fallon et al. (2000) (J. Fallon, personal communication).
Behavioral testing. Unilaterally lesioned rats were tested for rotational behavior in response to amphetamine (4 mg/kg, i.p.) before infusion and 4 weeks postinfusion. Animals were placed (randomized) into automated rotometer bowls, and left and right full-body turns were monitored by a computerized activity monitor system, as described previously (Bjorklund et al., 2002).
Bromodeoxyuridine administration. Rats that received intrastriatal PBS or TGFα infusions for 2 weeks were administered bromodeoxyuridine (BrdU) (Sigma, St. Louis, MO) intraperitoneally for 3 d postimplantation at a dose of 50 mg/kg body mass three times per day. Rats that received intrastriatal PBS or TGFα infusions for 1 or 4 weeks were administered BrdU intraperitoneally three times a day for the day before implantation and 3 d postimplantation (4 d total) at a dose of 50 mg/kg body weight. In addition, the rats that received PBS or TGFα infusions for 1 or 4 weeks were concomitantly administered BrdU in the drinking water for 7 d postimplantation at a dose of 225 mg/kg body weight per day with the rat drinking on average 15 ml per day. At histological evaluation, there were no discernible differences between the two BrdU administration protocols used. These BrdU administration protocols were applied for between 3 and 7 d, and the protocol used by Fallon et al. (2000) was applied for 3 d. The cumulative BrdU doses used in these experiments were lower than those used by Fallon et al. (2000). The lower BrdU dosage reduced the likelihood of falsely labeled cells, but the resulting cell fates after infusion remained identifiable (Cooper-Kuhn and Kuhn, 2002; Rakic, 2002).
Histological procedures. Animals were terminally anesthetized by an intraperitoneal injection of sodium pentobarbital (100 mg/kg) and perfused with intracardial heparin saline (0.1% heparin in 0.9% saline; 100 ml/rat) followed by paraformaldehyde (4% in PBS; 200 ml/rat). The brains were removed and postfixed for 8 hr in 4% paraformaldehyde solution. After postfixation, the brains were equilibrated in 20% sucrose in PBS, sectioned (40 μm) on a freezing microtome, and collected in PBS, as described previously (Bjorklund et al., 2002).
Immunohistochemistry and cell counting. All immunohistochemistry was performed on randomly selected series of sections that represented 
th of the total brain. Routine indirect immunofluorescence was performed as described previously (Bjorklund et al., 2002) using the following primary antibodies raised against tyrosine hydroxylase (TH) (1:300; Pel-Freez Biologicals, Rogers, AK), dopamine transporter (DAT) (1:2000), polysialylated neuronal cell adhesion molecule (PSA-NCAM) (1:50), neuronal nuclei (NeuN) (1:100; all from Chemicon, Temecula, CA), β-tubulin (1:500 and 1:2000; Covance, Berkeley, CA), S100β (1:50; Swant, Bellinzona, Switzerland), GFAP (1:500; Dako, High Wycombe, UK), nestin (1 μg/ml), 3CB2 (1 μg/ml), 40E-C (1 μg/ml; all from Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA), proliferating cell nuclear antigen (PCNA) (1:200), doublecortin (1:400), and β-catenin (1:100; all from Santa Cruz Biotechnology, Santa Cruz, CA). Immunohistochemistry for PCNA required antigen retrieval. Antigen retrieval was performed by incubating the free-floating sections in antigen retrieval solution (10× Target Retrieval Solution; Dako) diluted to 1× in distilled water. The antigen retrieval solution was preheated in a waterbath at 80°C before the sections were added for 30 min. The sections were subsequently washed three times in PBS at room temperature before immunohistochemistry was performed. Visualization of incorporated BrdU requires DNA denaturation before performing immunohistochemistry, and this was performed in a similar manner, as described previously (Kuhn et al., 1997). Sections were incubated in 2N hydrochloric acid for 30 min at 37°C before incubating in 0.1 m sodium borate for 10 min. The sections were then rinsed three times in PBS before commencing immunohistochemistry, as above. Biotinylated sheep anti-BrdU (1:200; Novus Biologicals, Littleton, CO) was diluted with other primary antibodies derived from suitable species. The BrdU was visualized using Alexa Fluor 488-conjugated streptavidin (1:500; Molecular Probes, Eugene, OR). Control sections processed without primary antibody were performed on selected sections that verified the specificity of the technique. The sections were examined using a laser scanning confocal microscope (LSM510/Meta; Carl Zeiss, Thornwood, NY). PSA-NCAM and nestin-immunopositive cell counts and SVZ measurements were performed on 
th of the sections using a stereology workstation (Stereoinvestigator; Microbrightfield, Williston, VT) with an integrated epifluorescence microscope (Axioskop 2+; Carl Zeiss).
Statistical analysis. Data are presented as mean ± SEM. ANOVA and t tests were used to assess differences between data groups using Statview software (SAS Institute, Cary, NC). Differences were considered statistically significant when p < 0.05.
Results
Induction of SVZ proliferation and migration
Intrastriatal infusion of TGFα into unilateral 6-OHDA-lesioned rats expanded the ipsilateral and contralateral SVZ and induced bilateral cell migratory waves rostral to the infusion site (Fig. 1a). With in 7 d of TGFα infusion, hyperplastic nodules formed bilaterally in the medial, lateral, and dorsal walls of the SVZ (Fig. 1b). Radiating in an anterior direction from these expanded SVZs, cell migratory waves formed in the adjacent striatum and septum. The cells in these migratory waves expressed nestin, a marker for multipotent progenitor cells in the developing brain (Lendahl et al., 1990), and exhibited small somata and extended fine, branched processes (Fig. 1c,d). The intrastriatal TGFα infusion expanded a nestin+/GFAP+ cell population in the SVZ that exhibited an astrocytic morphology (Figs. 2a,b, 3b). The SVZ expansion and migration were not present in vehicle-administered rats (Figs. 2c,d, 3a). Interestingly, a general increase and change in localization of β-catenin expression was also observed in the expanded SVZ, such that the normal membranous localization was complemented by prominent cytoplasmic β-catenin (Fig. 2e,f). Throughout the infusion period, cells in the SVZ of vehicle-administered rats expressed the PCNA (Fig. 4a-d). Numerous cells in the SVZ of TGFα-treated rats also expressed PCNA (Fig. 4e-h). The contralateral SVZ-derived nodules continued to increase in size (>100 μm in diameter) throughout the 4 week infusion period (Fig. 4h) and occasionally bridged the lateral ventricle. After 4 weeks, the ipsilateral SVZ-derived nodules were shallower and less circumscribed, and an increase in the size of the adjacent parenchymal wave of cells was observed (Fig. 4g). Immunohistochemical analysis demonstrated an increase in the number of BrdU+ cells in the SVZ of TGFα-treated rats (Fig. 5a-d). The vast majority of BrdU+ SVZ cells expressed nestin (Fig. 5e,f), whereas very few BrdU+ SVZ cells expressed solely β-tubulin or GFAP (Fig. 5g,h). Interestingly, the majority of the BrdU-positive cells within the hyperplastic nodules resided at the SVZ-nodule interface (Fig. 5e).
Representative images from 6-OHDA-lesioned rats treated with TGFα for 2 weeks. Under low magnification, two distinct nestin+ (green) waves of cells were visible (a). Whereas one wave appears to arise from the ipsilateral (IL) SVZ and migrates through the IL striatum toward the TGFα infusion site, the other wave appears to arise from the contralateral (CL) SVZ and migrates through the septum toward the TGFα infusion site. Within the lateral ventricle SVZ of the TGFα-treated rats, ubiquitous hypercellular nodules were present that contained both GFAP+ (red) and nestin+/GFAP+ colocalized (yellow) cells (b). Under higher magnification, the striatal wave of cells is composed of nestin+ cells (c, d). The nestin+ cells exhibit a high nuclear-to-cytoplasmic ratio and possess numerous fine processes that extend considerable distances (d). *, Hypercellular nodule; CC, corpus callosum; LV, lateral ventricle; S, septum; Str, striatum. Scale bars: b, 50 μm; c, 20 μm; d, 10 μm.
Representative images of SVZ in 6-OHDA-lesioned rats treated with TGFα or vehicle for 2 weeks. On examination of the bilateral SVZ at low power (a-d), more nestin+ (green) cells are present in the lesioned TGFα-treated SVZ (a, b) than in the lesioned PBS-administered SVZ (c, d). Within the expanded SVZ of TGFα-treated rats (e),β-catenin accumulated compared with PBS-administered SVZ (f). *, Lateral ventricle; CC, corpus callosum; S, septum; Str, striatum. Scale bars: a-d, 200 μm; e, f, 50 μm.
Intrastriatal TGFα administration increases nestin+ SVZ cells. High-magnification images of the lesioned TGFα-treated SVZ (a) and lesioned PBS-administered SVZ (b) depict a thickened SVZ when TGFα was administered. The thickened SVZ contained numerous nestin+/GFAP+ (yellow) cells that exhibited an oval soma with elongated processes (1) accompanied by numerous nestin-/GFAP+ (red) cells (2). The lesioned PBS-administered SVZ is within normal limits of cellularity and contains both GFAP+ (red) cells (1) and nestin+ (green) cells (3). This difference in the number of nestin+ cells was quantified as cells per millimeter2 after 2 and 4 weeks of infusion (c). There were significantly more nestin+ cells in the SVZ of the TGFα-treated rats (white) relative to the SVZ in the vehicle control (black) at both 2 and 4 weeks (t test; p < 0.05). *, Lateral ventricle; IL, ipsilateral. Scale bar, 20 μm.
Representative images of SVZ in 6-OHDA-lesioned rats treated with TGFα or vehicle for 2 and 4 weeks. Vehicle-administered rats exhibited scattered nestin+ (red) cells among numerous GFAP+ (blue) cells in the SVZ with an expected number of PCNA+ (green) nuclei homogeneously distributed throughout the lateral wall (a-d). The ipsilateral (IL) and contralateral (CL) SVZ in vehicle-administered rats exhibited similar features at both 2 (a, b) and 4 (c, d) weeks of infusion. After 2 weeks of TGFα infusion, both ipsilateral (e) and contralateral SVZ (f) contained numerous nestin+ (red) cells with PCNA+ (green) nuclei. The hypercellular nodules present in both SVZ were composed entirely of PCNA+ nuclei. After 4 weeks of TGFα infusion, the ipsilateral SVZ (g) contained numerous nestin+ cells that were also present in the adjacent striatum. The ipsilateral hypercellular nodules appeared to become confluent, forming a thickened SVZ (g). The contralateral SVZ (h) was also thickened but lacked any nestin+ cells in the adjacent striatum. The hypercellular nodules in the contralateral SVZ remained prominent (h). *, Lateral ventricle; CC, corpus callosum; Str, striatum. Scale bar, 250 μm.
Representative images of BrdU-labeled cells in 6-OHDA-lesioned rats after 2 weeks infusion of TGFα or vehicle. Intrastriatal infusion of TGFα produced numerous BrdU+ cells in the ipsilateral SVZ (a, b) compared with vehicle-administered rats (c, d). The hypercellular SVZ nodules (e) that extend into the lateral ventricle (*) contain numerous nestin+ cells (red), but only a fraction of these cells are also BrdU+ (green). The majority of the intensely stained BrdU+ nuclei reside at the SVZ-nodule interface. At higher magnification (f), a proportion of the nestin+ cells (red) are BrdU+ (green). GFAP (blue) and β-tubulin (red) were also examined in the SVZ (g). Isolated GFAP+/BrdU+ and β-tubulin+/BrdU+ cells were present. At higher magnification (h), the isolated β-tubulin+ (red)/BrdU+ (green) cell was confirmed by three-dimensional confocal imaging. Str, Striatum. Scale bar, 50 μm.
The cell types that developed in the SVZ of TGFα-administered rats differed from vehicle-administered rats. A 128% increase in the numbers of SVZ cells expressing nestin was observed after TGFα administration (Fig. 3). A concomitant 53% depletion of the PSA-NCAM-expressing cell population was observed in the SVZ (Fig. 6), indicating a loss of neuroblasts. Doublecortin is a similar marker of migrating neuroblasts (Brown et al., 2003), and the number of doublecortin-expressing cells in the SVZ was similarly diminished after TGFα infusion, corroborating the neuroblast loss characterized by PSA-NCAM (data not shown). Nestin-expressing cells and PSA-NCAM-expressing cells in the SVZ of naive animals administered TGFα or vehicle infusions were also quantified (data not shown). Although there appeared to be a difference in SVZ cell populations between naive and 6-OHDA-lesioned rats, detailed quantification did not establish a significant difference between the groups (p > 0.05). As expected with increased EGF agonist availability, immunohistochemistry demonstrated a marked increase of EGFr expression in the SVZ of TGFα-treated rats compared with vehicle-administered rats after 28 d of infusion (data not shown).
Intrastriatal TGFα administration reduces PSA-NCAM+ SVZ cells. Rats that received an intrastriatal infusion of vehicle demonstrated numerous PSA-NCAM+ cells in the SVZ (a), whereas rats treated with TGFα possessed relatively few PSA-NCAM+ cells in the SVZ (b). This difference was quantified as cells per millimeter after 2 and 4 weeks of infusion (c). There were significantly fewer PSA-NCAM+ cells in the SVZ of the TGFα-treated rats (white) relative to the SVZ in the vehicle control (black) (t test; p < 0.05). IL, Ipsilateral. Scale bar, 20 μm.
Within the parenchyma, ipsilateral and contralateral cell migratory waves radiated from the SVZ (Fig. 7). The cell migratory waves only appeared anterior to the TGFα infusion site (Fig. 7a,b). The waves of cells coexpressed nestin and PCNA in the striatum (Fig. 7c,f), around the anterior commissure (Fig. 7d), and in the septum (Fig. 7e,g,h), indicating that these cells were still dividing in response to the TGFα-induced mitogenic effect, but very few of the cells in the migratory waves were BrdU immunopositive. At higher magnification, the cells in the migratory waves frequently formed chains (Fig. 7g,h). A gradient of nestin+ progenitors was observed with the majority of cells closest to the infusion site at 2 weeks. After 4 weeks, numerous nestin+ progenitors were observed in the rostral areas of the cell migratory waves (Fig. 7i,j). In addition, a large number of nestin+ progenitors was observed in the contralateral cell migratory wave that extended into the medial orbital cortex, prelimbic cortex, and medial olfactory bulb. In the ipsilateral hemisphere, the cell migratory wave focused into the rostral migratory stream and expanded the olfactory part of the lateral ventricle into the medial part of the anterior olfactory nucleus.
Characterization of the cell migratory waves. In the parenchyma (gray), ipsilateral (IL; dark green) and contralateral (CL; light green) cell, migratory waves radiated away from the respective SVZs (dark and light blue) during intrastriatal infusion of TGFα (orange) (a). After 2 weeks, the cell migratory waves were only present anterior (A) to the infusion site, and cell densities reduced at the anterior tip of the waves (b). The cell migratory waves were predominantly composed of nestin+/PCNA+ cells exhibiting a fusiform morphology (c-h). After 4 weeks of TGFα infusion, dopamine-denervated rats showed prominent nestin+ cellular waves in the contralateral prelimbic and orbital cortex and olfactory bulb with a hypercellular thickening of the ipsilateral olfactory ventricle lining (i, j). D, Dorsal; V, ventral; P, posterior; AC, anterocaudal. Scale bar, 25 μm.
Differentiation of nestin+ progenitors
Cells expressing PSA-NCAM, doublecortin, NeuN, β-tubulin, or GFAP were not present in the striatal or septal migratory wave regions throughout the 4 week TGFα infusion period. We also immunostained postmortem tissue for TH, the rate-limiting enzyme of dopamine synthesis (Fig. 8a). The majority of the striata from 6-OHDA-lesioned rats contained no TH-positive neurons. This was also true for brains analyzed for non-6-OHDA-lesioned animals. Only 3 of 16 and 1 of 3 rats in the 4 week 6-OHDA lesion group contained any TH-positive cells, and one rat received a TGFα infusion. A mean of 2079 cells that had incorporated BrdU was examined for TH and DAT colocalization in the ipsilateral striatum of each rat (n = 4). None of these striatal TH-positive neurons were DAT- or BrdU-positive. In the SNc, a few remaining DA neurons were present in the 6-OHDA-lesioned side, but neither these neurons nor the TH-positive neurons in the contralateral SNc had incorporated BrdU (Fig. 8a). To test whether intrastriatal TGFα administration can restore motor function to rats with a unilateral 6-OHDA lesion, amphetamine rotations were used before infusion and after 4 weeks of infusion (Moore et al., 2001). No significant reductions in rotation scores were recorded using this behavioral test (Fig. 8b).
Lack of new DA neurons. None of the TH (red)-immunopositive neurons in the SNc of any animal in any group had nuclear incorporation of BrdU (green), although small infrequent BrdU+ nuclei were found in the region (a). The lack of striatal or nigral DA neurogenesis was reflected by an absence of improvement in amphetamine-induced side bias in either TGFα- or vehicle-administered rats (b). Scale bar, 20 μm.
Discussion
In these experiments, we found a cellular proliferation in the SVZ in response to the TGFα infusion, with hyperplastic ventricular nodules and migration of neural progenitors into the adjoining striatal, septal, and cortical parenchyma in 6-OHDA-lesioned animals but not naive animals. However, there was no evidence in any of the locations examined that the mitotically active migratory cells differentiated into a mature neuronal or glial cell type. DA neurons with incorporated BrdU label were not present in the lesioned striatum or substantia nigra. This lack of DA neurogenesis in the lesioned striatum and substantia nigra was corroborated by an absence of functional recovery from the amphetamine-induced side bias in treated and control rats. Interestingly, an independent study published recently is consistent with our conclusion that there is no evidence of new dopaminergic neurons formed in the adult mammalian substantia nigra under normal conditions (Frielingsdorf et al., 2004).
In normal animals that did not have a previous dopamine denervation of the striatum, infusion of TGFα was insufficient to induce a migratory wave of nestin-expressing cells from the SVZ. However, in 6-OHDA-lesioned animals, we observed migratory waves of nestin-positive cells orientated from the SVZs of both hemispheres toward the infusion site. Although EGFr mediates chemotactic migration in the developing telencephalon, the complete set of molecular migratory cues in the dopamine-denervated striatum are unknown (Caric et al., 2001). One hypothesis is that dopamine receptors (D1) on the nestin-positive cells are involved in limiting the proliferative response (Ohtani et al., 2003); thus, DA denervation would release these cells from such inhibition. The nestin-positive cells in the cell migratory wave also expressed PCNA, indicating that the cells were a mitotically active progenitor cell type, which exhibited a fusiform morphology consistent with the process of migration. The absence of BrdU label in the nestin-positive waves of cells may reflect a dilution of BrdU retention with successive mitoses (Dayer et al., 2003). Nonetheless, there were no neuroblasts that expressed doublecortin or PSA-NCAM in the area of the migratory progenitor cell waves. This provided evidence, independent of BrdU label, that the cells within the migratory wave were not differentiating toward a neuronal phenotype. Immunohistochemistry using 3CB2 or 40E-C antibodies showed that the wave of cells did not have a scaffold or the presence of radial glia that provide mechanisms for migration in the neocortex and telencephalon (Alvarez-Buylla et al., 2001; Rakic, 2003). In addition, these nestin+/PCNA+ cells did not express mature cellular proteins, such as NeuN, β-tubulin, or GFAP. These data points suggest that all of the cells in the wave are immature and relatively undifferentiated. Whereas the ipsilateral migratory waves of cells were solely anterior of the infusion site, the contralateral migratory waves were orientated in an anterior but more lateral direction toward the ipsilateral hemisphere. This reproducible distribution of nestin+/PCNA+ cells in the parenchyma indicates a potential gradient of stimulus associated with striatal TGFα infusion. The TGFα stimulus may cause elevated expression of developmental guidance cues in the adult brain that are normally present in the developing brain (Seroogy et al., 1995; Kornblum et al., 1997). Furthermore, the effect of TGFα in the SVZ was to induce dedifferentiation of progenitor cells, as shown by the significant depletion of SVZ neuroblasts during TGFα infusion that is consistent with an EGF infusion study (Doetsch et al., 2002). This TGFα-induced dedifferentiation may create a cell population that is more sensitive to exogenous guidance cues (Doetsch et al., 2002). The combination of the characteristic changes in migration patterns according to the site of the infusion cannula and the consistent cell orientation indicates that the PCNA+/nestin+ cells in the parenchyma are derived from the SVZ, although non-neural cells can also express nestin in the CNS (Lardon et al., 2002).
The hypercellular SVZ-derived nodules protruding into the neighboring lateral ventricle were the result of this proliferative response. Interestingly, similar nodules have been demonstrated during intraventricular EGF infusions (Kuhn et al., 1997). In that study, the nodules resolved after cessation of infusion. However, an inherent risk of carcinogenesis may be associated with such repeated mitotic events, and high levels of EGFr expression have been linked to aggressive forms of ependymoma (Vogelstein et al., 1988; Gilbertson et al., 2002). The bilateral mitogenic effect of the TGFα infusion indicated that the protein was bioactive at comparatively long distances from the infusion site, although the proliferative effects were restricted to the SVZ of the lateral ventricles and much more prominently ipsilateral than contralateral. The increased β-catenin immunoreactivity we observed in the SVZ during TGFα administration is an interesting finding. β-Catenin was identified by its association with E-cadherin and is an integral component of classical Wnt signaling that facilitates proliferation and differentiation via cell-to-cell contact mechanisms (e.g., cadherin and Eph-Ephrin) in a variety of tissues (McCrea et al., 1991; Nelson and Nusse, 2004). β-Catenin signals are important during developmental cell growth and fate determination in the nervous system (Zechner et al., 2003). β-Catenin-mediated signaling is only active if the protein leaves the cell membrane and cytoplasmic localization is an indicator of Wnt signaling initiation leading to nuclear accumulation and target gene transcription (Huelsken and Birchmeier, 2001; Nelson and Nusse, 2004). Our identification of this component of the signaling pathway after TGFα stimulus suggests a role for Wnt signals in the proliferative regulation of the SVZ.
After 4 weeks, no new dopaminergic neurons were found in the lesioned striatum of TGFα- or vehicle-administered rats. This contrasts with a previous report (Fallon et al., 2000) using identical methods to our study that illustrated the presence of new DA neurons identified by tyrosine hydroxylase and dopamine transporter immunoreactivity in cells containing BrdU (Fallon et al., 2000). In addition, no new dopaminergic neurons were found in the lesioned substantia nigra pars compacta, which differs from a report that identified ongoing neurogenesis in that region (Zhao et al., 2003). This may be attributable to the challenges of BrdU incorporation and detection methods but still highlights the stringency required for identifying neurogenesis in novel regions of the adult brain (Kornack and Rakic, 2001; Cooper-Kuhn and Kuhn, 2002; Rakic, 2002).
Fallon et al. (2000) demonstrated a reduced apomorphine sensitivity measured by behavioral rotations after 4 weeks of intrastriatal TGFα infusion. This behavioral test relies on apomorphine stimulating supersensitive dopamine receptors in the dopamine-denervated striatum, which induces contralateral rotations in the rat. Structural perturbations in the striatum caused by cannulas or inflammation (e.g., post 6-OHDA denervation) (Dunnett et al., 1988; Jeyasingham et al., 2001) can make apomorphine behavior dopamine supersensitivity tests unreliable. The striatal structural perturbations may cause the supersensitive receptors to be lost or damaged, thus reducing apomorphine-induced side bias in a dopaminergic-independent manner (Borlongan et al., 1997). Therefore, given these well known limitations of the apomorphine test in cases of surgical perturbations of the striatum, we used amphetamine-induced side bias to investigate behavioral improvement. Amphetamine acts in a similar manner to cocaine by opening the dopamine transporter, causing a sharp elevation of DA in the synaptic cleft disrupting the normal feedback control of the transmitter level (Fischer and Cho, 1979; Pierce and Kalivas, 1997). This test provides robust ipsilateral side bias, even in perturbed striata, because it more realistically reflects dopamine release capacity (Dunnett et al., 1988). The lack of motor improvement after amphetamine administration in TGFα-infused 6-OHDA-lesioned rats is consistent with the observed absence of dopamine-producing neurons within the lesioned striatum. In addition, an olfactory discrimination test showed no difference between TGFα and vehicle infusions, indicating that the TGFα infusion also does not affect behavioral tests of olfactory function via altered neurogenesis (O. Cooper, unpublished observations).
Although we reproduced the findings of proliferation and migration of neural precursors, our study contrasts with a previous report of intrastriatal TGFα infusion leading to DA neurogenesis in the adult brain because there was an absence of differentiation of such neural precursors (Fallon et al., 2000). The amount of TGFα that we administered, at least 2 weeks postlesion, was shown by Fallon et al. to be sufficient to induce a migratory ridge and improve apomorphine-induced side bias after 2 weeks of infusion (Fallon et al., 2000). Furthermore, the development of the non-neurogenic hyperplastic nodules in the SVZ clearly demonstrates effective delivery of active TGFα. The lack of neural differentiation and DA neurogenesis was congruent with our negative behavioral data that realistically reflect the dopamine release capacity of the striatum.
In the SVZ, control of the neurogenic niche is created by the expression of the bone morphogenetic protein 4 (BMP4) and the BMP antagonist noggin by SVZ astrocytes and ependymal cells, respectively (Lim et al., 2000). These two proteins interact and control neuronal-glial fate determination within the SVZ, and manipulation of this interaction by exogenous noggin enhances neuronal fate determination (Lim et al., 2000; Chmielnicki et al., 2004). Interestingly, after DA denervation, expression of BMP, BMP receptor, follistatin, cerberus, and chordin mRNA decreases in the striatum and substantia nigra, indicating altered homeostasis in these nonneurogenic zones (Chen et al., 2003). Delivery of noggin protein to the denervated striatum may create a neurogenic niche of the SVZ and promote neuronal differentiation of the migrating nestin progenitors induced by intrastriatal TGFα infusion. During development, TGFα does control the number of midbrain DA neurons, because TGFα mutations reduce midbrain DA neurons (A9 group) in newborn mice (Blum, 1998). Presumably, this action of TGFα occurs within the neuroectoderm before the cells become postmitotic and migrate ventrally and laterally. Although intrastriatal TGFα infusion provides an external mechanism to induce proliferation of cells within the adult SVZ, perhaps reconstituting neuroectoderm-like qualities, our data indicate that neither TGFα administration nor a lesioned striatum will provide sufficient cues to affect the differentiation of nestin precursors into functional DA neurons (Cameron et al., 1998). Future directions for this work will focus on providing additional signals to the SVZ-derived cells in the dopamine-denervated striatum to differentiate into functional midbrain-like DA neurons.
Footnotes
This work was supported by the Harvard Center for Neurodegeneration and Repair.
Correspondence should be addressed to Dr. Ole Isacson, McLean Hospital-Harvard Medical School, Neuroregeneration Laboratories, Mailman Research Center, 115 Mill Street, Belmont, MA 02478. E-mail: isacson{at}hms.harvard.edu.
Copyright © 2004 Society for Neuroscience 0270-6474/04/248924-08$15.00/0
