- Top of page
- Materials and Methods
- Disclosure of Potential Conflicts of Interest
The aim of the present study was to investigate the putative cooperative effects of transforming growth factor β (TGF-β) and glial cell line-derived neurotrophic factor (GDNF) family ligands in the differentiation of midbrain progenitors toward a dopaminergic phenotype. Therefore, a mouse midbrain embryonic day (E) 12 neurospheres culture was used as an experimental model. We show that neurturin and persephin (PSPN), but not GDNF, are capable of transient induction of dopaminergic neurons in vitro. This process, however, requires the presence of endogenous TGF-β. In contrast, after 8 days in vitro GDNF rescued the TGF-β neutralization-dependent loss of the TH-positive cells. In vivo, at E14.5, no apparent phenotype concerning dopaminergic neurons was observed in Tgf-β2−/−/gdnf−/− double mutant mice. In vitro, combined TGF-β/PSPN treatment achieved a yield of approximately 20% TH-positive cells that were less vulnerable against 1-methyl-4-phenyl pyridinium ion toxicity. The underlying TGF-β/PSPN differentiation signaling is receptor-mediated, involving p38 mitogen-activated protein kinase and phosphatidylinositol 3-kinase pathways. These results indicate that phenotype induction and survival of fully differentiated neurons are accomplished through distinct pathways and individual factor requirement. TGF-β is required for the induction of dopaminergic neurons, whereas GDNF is required for regulating and/or maintaining a differentiated neuronal phenotype. Moreover, this study suggests that the combination of TGF-β with PSPN is a potent inductive cocktail for the generation of dopaminergic neurons that should be considered in tissue engineering and cell replacement therapies for Parkinson's disease.
Disclosure of potential conflicts of interest is found at the end of this article.
- Top of page
- Materials and Methods
- Disclosure of Potential Conflicts of Interest
Author contributions: E.R.: conception and design, financial support, collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; O.O.: collection and assembly of data, data analysis and interpretation; B.R., S.Hee., and S.Hei.: collection and assembly of data; M.W.: collection and assembly of data, data analysis and interpretation; K.K.: conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.
Stem cells have been considered as potent candidates in cell replacement therapies for neurodegenerative disorders, such as Parkinson's disease . Consequently, the signals and molecular mechanisms that mediate and regulate the differentiation of stem/progenitor cells into dopaminergic neurons have been extensively investigated. Thus, many studies have focused on elucidating the transcriptional code specific for the induction and specification of mesencephalic dopaminergic neurons. Analysis of deficient mice has identified transcription factors, such as Nurr1, Pitx3, En1, Lmx1b, and Lmx1a, as key components in the model of dopaminergic neuron development [2, , , , –7]. Together with the signaling molecules sonic hedgehog and fibroblast growth factor (FGF) 8, the role of transforming growth factor β (TGF-β) has also been highlighted [8, , , –12].
Glial cell line-derived neurotrophic factor (GDNF), neurturin (NRTN), artemin (ARTN), and persephin (PSPN), also named the GDNF family ligands (GFLs), are distant members of the TGF-β superfamily. GFLs signal through a multicomponent receptor system comprised of a high-affinity binding component, a GPI-linked GDNF family receptor (GFR) α subunit, and a common signaling component, the transmembrane tyrosine kinase Ret (for review, see ). GFLs differentially promote survival and regulate differentiation in peripheral and central neuronal populations. Because of their survival-supporting effect on midbrain dopaminergic neurons, GFLs have been investigated in several animal models of Parkinson's disease. It appears that, besides the well-documented neuroprotective effects of GDNF on midbrain dopaminergic neurons in vitro and in vivo, NRTN prevented degeneration of mature dopaminergic neurons in the 6-hydroxydopamine model [14, –16]. Moreover, ARTN and PSPN have been shown to have effects equivalent to those of GDNF and NRTN and prevented the loss of dopaminergic neurons in lesion models of Parkinson's disease as well [17, 18].
During the last decade, an increasing number of studies have provided convincing evidence that multiple ways of interactions between members of the superfamily, as well as between families of growth factors and neurotrophins, may occur, by means of receptor cross-talk, cooperativity, or dependence of one factor on the presence of a cofactor. Although TGF-β, GDNF, NRTN, ARTN, and PSPN themselves are able to support survival of mesencephalic dopaminergic neurons, GDNF requires TGF-β to exert its entire neurotrophic potential in vitro and in vivo [19, 20]. TGF-β is apparently required for recruitment of the GFRα1 receptor in the plasma membrane . Moreover, although NRTN, ARTN, and PSPN promote the morphological differentiation of mesencephalic dopaminergic neurons , midbrain neural progenitors require TGF-βs as an inductive signal to acquire a dopaminergic cell fate . In the developing midbrain GDNF, NRTN, ARTN, and PSPN are spatially and temporally differentially expressed. Whereas no GDNF expression could be detected in the ventral midbrain during mouse embryonic development, NRTN expression was found at embryonic days (E) 10–14 but was absent at E16 and E18 . Moreover, although in the rat ventral mesencephalon PSPN mRNA was highly expressed prenatally , the ventral mesencephalon of E14 rats lacks ARTN mRNA expression .
Surprisingly, gdnf−/− [26, –28], nrtn−/− , artn−/− , and pspn−/−  mice lack severe deficits in dopaminergic neurons. In contrast, in Tgf-β2−/−/Tgf-β3+/− and in Tgf-β2+/−/Tgf-β3−/− mice the number of TH-immunopositive cells at E14.5 was significantly reduced compared with wild-type mice .
In the present study we investigated putative cooperative effects of TGF-β with GFLs on midbrain progenitor cell differentiation toward dopaminergic cell fate. The results demonstrate that phenotype induction and survival of fully differentiated neurons are accomplished through distinct pathways and requirements for individual factors. TGF-β is required for the induction of dopaminergic neurons, whereas GDNF is required for regulating and/or maintaining a differentiated neuronal phenotype. Moreover, this study suggests that TGF-β combined with PSPN is a potent inductive cocktail for the generation of dopaminergic neurons that should be considered in tissue engineering.
Materials and Methods
- Top of page
- Materials and Methods
- Disclosure of Potential Conflicts of Interest
TGF-β, GDNF, NRTN, PSPN, and FGF2 were obtained from tebu-bio (Offenbach, Germany, http://www.tebu-bio.com). Lyophilized factors were resuspended in Dulbecco's modified Eagle's medium (DMEM) (Gibco, Grand Island, NY, http://www.invitrogen.com) with 0.25% bovine serum albumin (BSA) (Sigma-Aldrich, Taufkirchen, Germany, http://www.sigmaaldrich.com), 100 U/ml penicillin, 0.5 μg/ml streptomycin, and 100 μg/ml neomycin (all from Gibco) to give a final concentration of 1 μg/ml and stored in aliquots of 50–100 μl at −70°C until use. Lyophilized TGF-β1 was resuspended in 1 mM HCl containing 0.25% BSA to give a final concentration of 0.5 μg/ml and stored in aliquots of 50 μl at −70°C until use. The neutralizing monoclonal mouse anti-TGF-β antibody (MAB1835) recognizing all three isoforms was obtained from R&D Systems Inc. (Wiesbaden, Germany, http://www.rndsystems.com). Activity of this antibody was determined using the Mv1Lu mink lung epithelial cell system [32, 33]. SB431542 was purchased from Tocris Bioscience (Bristol, U.K., http://www.tocris.com), SB203580 and LY294002 were obtained from Calbiochem (Schwalbach am Taunus, Germany, http://www.emdbiosciences.com). 1-Methyl-4-phenyl pyridinium ion (MPP+) was purchased from Sigma-Aldrich.
An aliquot of 40 ng of TGF-β1, GDNF, NRTN, or PSPN was loaded onto a nitrocellulose membrane. The membrane was blocked with 3% low-fat milk powder in Tris-Tween 20-buffered saline (TTBS) and incubated with primary anti-TGF-β1,2,3 antibody (1:200) overnight at 4°C. The following day the membrane was incubated with anti-mouse IgG coupled to peroxidase (1:10,000) and was developed using the Amersham enhanced chemiluminescence detection system. Signals were visualized on x-ray film.
Cell Culture of E12 Ventral Mesencephalon
Ventral mesencephalon from mouse E12 was isolated as described earlier [11, 12, 34]. Briefly, pregnant NMRI mice were sacrificed by cervical dislocation, and E12 embryos were collected in Ca2+- and Mg2+-free Hanks' balanced salt solution (Sigma-Aldrich). The day of vaginal plug identification was designated E1. Ventral mesencephalon was dissected, freed from meninges, and incubated in 0.15% trypsin for 15 minutes at 37°C. Small pieces of ventral mesencephalon were subsequently dissociated by gentle trituration using fire-polished Pasteur pipettes. Dissociated cells were resuspended in high-glucose DMEM-Ham's F-12 medium supplemented with 0.25% BSA, N1 additives, and 100 U/ml penicillin, 0.5 μg/ml streptomycin, and 100 μg/ml neomycin (Gibco). Cells in suspension were plated in noncoated culture dishes and cultured in the presence of 20 ng/ml FGF2. After 3 days in culture, formation of neurospheres was evident. Cells were allowed to expand for 7 days before the neurospheres were dissociated, and cells were plated onto polyornithin- and laminin-coated 12 mm2 glass coverslips in 24-well plates at a density of 200,000 cells/coverslip. FGF2 was removed from the expansion medium, and cells were incubated in serum-free medium in a 95% air/5% CO2 atmosphere at 37°C.
One, 4, and 7 days after plating (day in vitro [DIV] 1, 4, and 7) factors were applied at a final volume of 750 μl of medium at the following concentrations: 1 ng/ml TGF-β1, 10 μg/ml anti-TGF-β antibody, 10 ng/ml GDNF, 10 ng/ml NRTN, and 10 ng/ml PSPN. LY294002, SB431542, and SB203580 were applied at 20 nM, 10 μM, and 20 μM, respectively. At DIV3, DIV5, and DIV8 cells were fixed and processed for immunocytochemistry.
Immunocytochemistry on cultured cells was performed essentially as described earlier [12, 19]. Cultures were fixed in 4% paraformaldehyde for 30 minutes at room temperature, permeabilized with acetone for 10 minutes at −20°C, and blocked with 10% normal goat serum (NGS) in phosphate-buffered saline (PBS). Subsequently cells were incubated overnight at 4°C with primary antibodies diluted in PBS containing 10% NGS and 0.1% Triton X-100. Rabbit polyclonal anti-TH (1:1,000) (Chemicon, Hofheim, Germany, http://www.chemicon.com) and anti-Pitx3 (1:400) (Chemicon), mouse monoclonal anti-Nurr1 (1:200) (Santa Cruz Biotechnology; Santa Cruz, CA, http://www.scbt.com), anti-β-III-tubulin (1:25) (Developmental Studies Hybridoma Bank, Iowa City, IA, http://www.dshb.biology.uiowa.edu), and anti-GABA (1:50) (Sigma-Aldrich) and goat polyclonal anti-Pitx3 (1:50) (Santa Cruz Biotechnology) were used as primary antibodies. Donkey anti-rabbit, anti-mouse, or anti-goat IgGs coupled to either indocarbocyanin (CY3) or to fluorescein isothiocyanate (FITC) were used as secondary antibodies. Cells were viewed with a fluorescence microscope (Carl Zeiss, Jena, Germany, http://www.zeiss.com).
In control experiments performed by omitting the primary antibody, immunostaining was abolished. In addition, the specificity of Pitx3 labeling was assessed by preincubating the goat polyclonal antibody (1:50) with the respective blocking peptide (1:10) (Santa Cruz Biotechnology).
For immunodetection of c-Ret, GFRα1, GFRα2, and GFRα4, the following protocol was applied: fixed cells were incubated with 1% SDS diluted in PBS for 5 minutes and subsequently washed with PBS. After blocking with 1% BSA diluted in PBS for 15 minutes, cells were incubated with primary antibodies diluted in PBS containing either 10% NGS (for c-Ret and GFRα2) or 10% horse serum (for GFRα1 and GFRα4) and 0.1% Triton X-100 overnight at 4°C. Rabbit polyclonal anti-c-Ret (1:25) and anti-GFRα2 (1:25) (Santa Cruz Biotechnolgy), and goat polyclonal anti-GFRα1 (1:25) and -GFRα4 (1:25) (Santa Cruz Biotechnology) were used as primary antibodies. Donkey anti-rabbit or anti-goat IgGs coupled to CY3 were used as secondary antibodies.
Reverse Transcriptase-Polymerase Chain Reaction
Total RNA was isolated from mouse ventral mesencephalic primary dissected tissue, using the Qiagen RNeasy kit and following the manufacturer's instructions (Qiagen, Hilden, Germany, http://www.qiagen.com). Total RNA (1.0 μg) was reverse transcribed from an oligo(dT) primer using a Omniscript kit (Qiagen). Samples (3 μl) of the reverse transcriptase (RT) reaction were used in a polymerase chain reaction (PCR) containing 2.5 U of Taq DNA polymerase, 0.2 μM concentrations of specific primers, 1× PCR buffer (Qiagen), 200 μM deoxynucleoside triphosphates, and 2 mM MgCl2. For amplification of cDNA encoding receptors of the GDNF family members the following forward (F) and reverse (R) primer sequences were used:
C-Ret F: 5′-TGGCACACCTCTGCTCTATG-3′ (corresponding to nucleotides [nt] 749–768; GenBank Accession number NM 001080780)
C-Ret R: 5′-TGTTCCCAGGAACTGTGGTC-3′ (corresponding to nt 934-915; GenBank accession number NM 001080780)
GFRα1 F: 5′-GGGCACATGGAAGTTTCTTT-3′ (corresponding to nt 2,715-2,734; GenBank accession number NM010279.2)
GFRα1 R: 5′-TATCAGGAGACCGTGCACCA-3′ (corresponding to nt 3,170-3,151; GenBank accession number: NM 010279.2)
for full-length GFRα2 (GenBank accession number NM 008115.2):
GFRα2 F: 5′-CCACCATGATCTTGGCAAACGCCTTC-3′
GFRα2 R: 5′-CTAGGCCAAGGTCACCATCAG-3′
for full-length GFRα4:
GFRα4 F: 5′-CCACCATGGCCCACTGCATGGAGTC-3′
GFRα4 R: 5′-TTCAGCTCAGTGAGCAGTCATCG-3′
For detection of cDNAs encoding the c-Ret and GFRα1 proteins, the protocol was denaturation at 95°C for 15 minutes, for the optimum number of cycles (depending on primer pair) of PCR amplification under the following conditions: denaturation at 94°C for 1 minutes, annealing at the appropriate temperature (depending on primer pair) for 1minute, and elongation at 72°C for 1 minutes. Final extension at 72°C for 10 minutes was terminated by rapid cooling at 4°C.
For amplification of GFRα2 and GFRα4, the following protocol was used. After denaturation at 95°C for 2 minutes, nine cycles were performed at 95°C for 40 seconds, 55°C for 40 seconds, and 72°C for 1 minute 20 seconds, followed by 26 cycles at 95°C for 40 minutes, 55°C for 40 minutes, and 72°C for 1.20 minutes, increasing by 5 seconds for each cycle. Final extension at 72°C for 7 minutes was terminated by rapid cooling at 4°C.
PCR products were analyzed by agarose gel electrophoresis. The size of the reaction products was determined after ethidium bromide staining.
Cells were harvested in 10 ml of ice-cold homogenizing buffer containing 280 mM mannitol, 10 mM HEPES, 10 mM KCl, 1 mM MgCl2, adjusted to pH 7.0, and a protease inhibitor “cocktail” (10 μM leupeptin, 2 mM benzamidine, and 0.1 mM Pefabloc SC), scraped off the culture flasks with a rubber policeman, pelleted by centrifugation at 250g for 5 min, and resuspended in homogenization buffer. Homogenization was performed by sonication. Protein concentration was determined according to Bradford . Electrophoresis and blot procedures were performed as described . Proteins were separated by SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. Blots were blocked with 3% nonfat dry milk in TTBS for 8 hours and incubated with primary antibody overnight at a dilution of 1:500 for phosphorylated extracellular signal-regulated kinase (pERK) (rabbit polyclonal, obtained from Santa Cruz Biotechnology), phosphorylated Akt (pAkt) (rabbit polyclonal purchased from Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com; New England Biolabs, Frankfurt, Germany, http://www.neb.com), and anti-GFRα4 (goat polyclonal; Santa Cruz Biotechnology) and 1:10,000 for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (mouse monoclonal purchased from Abcam, Cambridge, U.K., http://www.abcam.com), respectively. After incubation with goat anti-mouse, goat anti-rabbit, or donkey anti-goat IgG coupled to horseradish peroxidase secondary antibodies (1:10,000 dilution), blots were developed in enhanced chemiluminescence reagents, and signals were visualized on x-ray film. Subsequently, films were scanned using a flat-bed scanner and the signal ratios pERK/GAPDH, pAkt/GAPDH, and GFRα4/GAPDH were quantified densitometrically. Differences in signal ratios were tested for significance using Student's t test. Results with levels of p < .05 were considered significant.
Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labeling Assay
Dissociated mouse E12 ventral mesencephalic neurospheres were plated onto 12 mm2 coverslips at a density of 200,000 cells/well. Cultures were treated with TGF-β alone or in combination with PSPN at DIV1. At DIV2 and DIV3 cells were fixed in 4% paraformaldehyde and subsequently washed. For detection of apoptotic cells, a terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay kit from Roche Diagnostics (Basel, Switzerland, http://www.roche-applied-science.com) was used following the manufacturer's instructions. Apoptotic cells were then visualized by fluorescence microscopy.
Dissociated mouse E12 ventral mesencephalic neurospheres were plated onto 12 mm2 coverslips at a density of 200,000 cells/well. For 5-bromo-2′-deoxyuridine (BrdU) labeling and detection, a detection kit from Boehringer/Mannheim (Mannheim, Germany, http://www.boehringer.com) was used. In brief, dissociated cells were incubated at DIV1, DIV2, and DIV3 with BrdU for 1 h in a 95% air/5% CO2 atmosphere at 37°C, fixed with 70% ethanol in 50 mM glycine buffer (pH 2.0) for 20 minutes at −20°C, and incubated with anti-BrdU antibody for 30 minutes at 37°C. Subsequently, coverslips were washed with buffer, incubated with mouse IgG coupled to FITC for 30 minutes at 37°C, washed again with buffer, and mounted with Vectashield (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). Finally, coverslips were viewed with an epifluorescence microscope.
Treatment of the Cultures with MPP+
MPP+ at a final concentration of 2 μM was added at DIV5 in dissociated neurospheres without treatment (controls) and in dissociated neurospheres treated with TGF-β in combination with PSPN at DIV1 and DIV4. One and 3 days after toxin application, that is, DIV6 and DIV8 after plating of dissociated neurospheres, respectively, controls and treated cells were fixed in 4% paraformaldehyde and processed for immunocytochemistry, as described above. For comparison between controls and treated cells, the cell number of TH-immunopositive cells at the day of MPP+ application, that is, at DIV5 after plating of dissociated neurospheres, was set to 100%. Statistical analysis was performed as described below.
Tgf-β2+/− heterozygous mice were offspring from breeding pairs kindly provided by T. Doetschman, University of Cincinnati (Cincinnati, OH). The generation of these strains has been described elsewhere [37, 38]. gdnf+/− heterozygous mice were kindly provided by M. Saarma, Institute for Biotechnology, University of Helsinki (Helsinki, Finland). Tgf-β2+/−/gdnf−/−, Tgf-β2−/−/gdnf+/−, and double mutant Tgf-β2−/−/gdnf−/− mice were generated by cross-breading the two heterozygous strains and setting up matings between these double heterozygous mice (Tgf-β2+/−gdnf+/− × Tgf-β2+/−gdnf+/−). For morphological comparison Tgf-β2+/+gdnf+/+ littermates were used as wild-type controls. The morning of the day on which a vaginal plug was detected in mating was designated gestation day 0.5. Analysis was performed on 14.5-day-old embryos. Heads were fixed in Bouin's fixative (75% picric acid, 20% formaldehyde, and 5% glacial acetic acid) for several hours, dehydrated in a graded series of ethanol, and embedded in paraffin wax. Paraffin blocks were cut into 10-μm serial sections.
Paraffin sections (10 μm) were deparaffinized and heated for 5 minutes in citrate buffer (pH 6.0) in a microwave oven at 600 W to improve antigen retrieval. After blockade of endogenous peroxidase activity by 5 minutes of treatment with 3% H2O2 in H2O, sections were preblocked with a Vector blocking kit (Linaris GmbH, Bettinen, Germany, http://www.linaris.com) to avoid nonspecific binding of the biotin/avidin system used for immunodetection (see below). Immunostaining was performed using a specific monoclonal mouse anti-TH antibody at a dilution of 1:100 following the manufacturer's instructions for the Vector M.O.M. peroxidase immunodetection kit (Linaris GmbH). After a 1-hour incubation at room temperature the reaction was visualized by a nickel-intensified 3,3′-diaminobenzidine (Kem-En-Tec Diagnostics A/S, Taastrup, Denmark, http://www.kem-en-tec.com) reaction. For controls, PBS was substituted for the primary antisera to test for nonspecific labeling. No specific cellular staining was observed when the primary antiserum was omitted.
Numbers of TH-labeled neurons on the complete series of 10 μm transverse sections were counted. A neuron was designated as TH-positive if it showed a darkly labeled cytoplasm and a clearly visible, unstained nucleus. Only cells fulfilling these criteria were included in the cell counts. To avoid double counting the same cell on two sequential sections, only every fifth section was counted.
Data are presented as the mean ± SEM. Statistical analysis was performed using the Student's double t test when one treated group was compared with the control. For multiple comparisons between treated groups statistical differences were compared using a one-way analysis of variance and Dunnett's post hoc test for multiple comparisons. Differences were considered statistically significant at p < .05, p < .01, and p < .001.
- Top of page
- Materials and Methods
- Disclosure of Potential Conflicts of Interest
The TGF-β superfamily consists of more than 40 members, including TGF-βs, activin, nodal, the GDNF family, bone morphogenic proteins, and GDF [10, 39]. In our efforts to elucidate the mechanisms that dictate induction and differentiation of mesencephalic dopaminergic neurons, we have previously shown that TGF-βs are required for the induction, differentiation, and survival of midbrain dopaminergic neurons in vivo and in vitro [8, 9, 12]. TGF-β apparently triggers midbrain progenitors to acquire a dopaminergic phenotype. Although GDNF itself has no differentiation-promoting effects on midbrain dopaminergic precursor cells , GDNF and other GFLs, such as NRTN, ARTN, and PSPN, have been identified as potent neurotrophic factors for several neuronal populations, including mesencephalic dopaminergic neurons. In addition, NRTN is an indispensable component in standard protocols that drive differentiation of embryonic stem cells toward dopaminergic neurons, reflecting the beneficial effects of NRTN in this process . Considering this background together with the observation that GDNF requires TGF-β to exert its survival-promoting effects [19, –21], the aim of the present study was to investigate putative cooperative effects of TGF-β and GFLs in the differentiation of midbrain progenitors toward dopaminergic neurons. Therefore, the well-characterized neurosphere culture system from mouse E12 ventral mesencephalon has been used as an experimental model in vitro [12, 34].
GFLs commonly bind to their specific GFRα receptor and subsequently activate Ret transmembrane receptor kinase . The neurospheres express c-Ret, as well as GFRα1, GFRα2, and GFRα4, the preferential receptors for GDNF, NRTN, and PSPN, respectively (Fig. 1), demonstrating that this cell system is suitable for investigating GFL effects. GFRα2 expression has been reported in ventral midbrain of mouse at E10, E12, and E14  but not in prenatal (E14) and postnatal day 7 rat midbrain . During amplification of GFRα2 cDNA additional bands that probably correspond to the GFRα2a, GFRα2b, and GFRα2c splice isoforms were also detected . The band detected during amplification of GFRα4 cDNA probably represents GFRα4 splicing. Alternative splicing of the mouse gfra4 gene has been shown to give rise to three variants, the putative GPI-anchored, the transmembrane, and the soluble form  in a tissue-specific manner. GFRα4 was found to be weakly expressed in the neurospheres (Fig. 1), consistent with previous data showing GFRα4 expression in mouse E12 ventral midbrain and in adult rat midbrain [43, 44]. The GFRα4 fragment observed presumably represents the soluble form of GFRα4. Indeed, immunoblotting of dissociated control neurospheres showed an immunoreactive band at ∼20 kDa (Fig. 5D) consistent with the predicted molecular mass of soluble GFRα4. Protein abundance for GFL receptors was further confirmed by immunofluorescence (Fig. 1B–1E). GFRα3 is also expressed in neurospheres (data not shown); however, because ARTN mRNA could not be detected in the ventral midbrain of E14 rats , the effects of ARTN in the differentiation potential of the neuropheres has not been considered in the present study. In addition, it should be noted that GFLs can also signal via the neural cell adhesion molecule NCAM . A potential functional implication of this observation in the ventral mesencephalon is not clear yet.
Treatment of dissociated neurospheres with GDNF together with TGF-β showed that GDNF potentiated the TGF-β effects on the number of cells expressing the early dopaminergic marker Nurr1 at DIV3 but was not able to rescue the decreased number of Nurr1-immunoreactive cells after neutralization of endogenous TGF-β (Fig. 2A). Interestingly, GDNF rescued the TGF-β neutralization-dependent loss of the TH-positive cells at DIV8 (Fig. 4A) in vitro. In vivo, no apparent phenotype with regard to dopaminergic neurons could be observed in Tgf-β2−/−/gdnf−/− mice at E14.5 (Fig. 3), matching the data observed in gdnf−/− mice, in which only the neurons of the enteric nervous system, but not the midbrain dopaminergic neurons are lost [26, –28]. The most likely explanation for these observations is that TGF-β isoforms may compensate for the loss of each other. Indeed, in Tgf-β2−/−/Tgf-β3+/− or Tgf-β2+/−/Tgf-β3−/− mice, the number of mesencephalic TH-positive cells was significantly reduced compared with wild-type mice at E14.5 . Taking the in vitro and in vivo data together, these results also suggest that phenotype induction and survival of fully differentiated neurons are accomplished through distinct pathways and requirements for individual factors. TGF-β is required for the induction of dopaminergic neurons, whereas GDNF is required for differentiated dopaminergic neurons rather than developing ones, suggesting that GDNF may play an essential role in regulating and/or maintaining a differentiated neuronal phenotype. For accomplishing this function, however, GDNF indeed requires TGF-β, as shown in vivo for dopaminergic and sympathetic neurons [20, 45].
Like GDNF, NRTN and PSPN have been shown to exhibit neurotrophic activity on mesencephalic dopaminergic neurons as well. Together with the lack of apparent developmental deficits in dopaminergic neurons in gdnf−/− [26, –28], pspn−/− , nrtn−/− , and artn−/−  mice, it appears that considerable redundancy of factors with the same actions is present in this central neuronal population. In vitro, treatment of dissociated neurospheres with NRTN or PSPN caused a significant but transient increase in the number of Nurr1-, Pitx3-, and TH-positive cells at DIV3 compared with the untreated controls (Fig. 2). In addition, the effects of individual GFLs were diminished after neutralization of endogenous TGF-β, demonstrating dependence of GFLs on TGF-β. With one exception, the combination of exogenous TGF-β with GFLs had no further beneficial effects on the differentiation of midbrain precursors toward dopaminergic neurons compared with single factor treatment: at DIV8, the combined application of TGF-β with PSPN impressively increased the number of TH-positive cells to approximately 18% (Fig. 4A). The TH-immunopositive neurons generated revealed Pitx3 immunoreactivity, were devoid of GABA expression, and exhibited a mature neuronal morphology (Figs. 2 and 4), suggesting that they are true mesencephalic dopaminergic neurons. A survival-promoting or proliferative effect of combined treatment could be excluded using TUNEL and BrdU assays (Figs. 4H, 4J), suggesting that TGF-β/PSPN has been acting as an inductive signal on precursor cells towards a dopaminergic cell fate.
An open issue concerning ventral mesencephalic precursors remains: does the loss of ability to generate dopaminergic neurons after extensive expansion of ventral mesencephalic precursors and repeated passaging of neurospheres [46, 47] argue against the presence of true dopaminergic precursors in unpassaged neurospheres, as well as in the mesencephalic neural tube? To answer this question we should consider the fact that neural stem cells in their biological environment in the neural tube during development become more committed  with each round of proliferation. In addition, proliferation and commitment of neural stem cells are temporally and spatially strictly regulated. Therefore, a consequence of extensive expansion of precursors and multipassaging of the neurospheres may result in a complete loss of TH-positive cells, reflecting a change in the developmental properties of mesencephalic progenitors.
The effect of TGF-β/PSPN treatment on mesencephalic progenitors toward the dopaminergic cell fate observed in the present study is mediated at least by converging p38 MAPK and PI-3K signaling pathways. Several lines of evidence support this view. Application of SB203580, a specific p38 MAPK inhibitor, or SB431542, a potent ALK4, ALK5, and ALK7 inhibitor, together with LY294002, a PI-3K inhibitor, caused a decrease in the number of TH-positive cells additional to that observed after single inhibitor application (Fig. 5A). Moreover, combined TGF-β/PSPN treatment of the cultures upregulated pERK and pAkt proteins (Fig. 5B). These results are consistent with previous observations documenting that TGF-β-induced dopaminergic differentiation depends on the TGF-β receptor complex and is mediated via the Smad and p38 MAPK pathways . In contrast, the TGF-β/GDNF cooperative effect on the survival of many neuronal subpopulations is PI-3K-independent . These data further support the hypothesis that differentiation and survival of neurons may require the same key players, but distinct underlying mechanisms. The signaling pathways underlying PSPN signaling are poorly understood, but in vitro and in vivo studies have demonstrated that PSPN can promote the survival of many neuronal subpopulation except the peripheral neurons [13, 17, 49], via a PI-3K/AKT pathway. Interestingly, treatment of the cells with TGF-β in combination with PSPN upregulated the expression of GFRα2a and GFRα4 (Fig. 5C, 5D).
Finally, the TGF-β/PSPN-induced TH-positive neurons were less vulnerable against MPP+ toxicity (Fig. 6). Three days after MPP+ application about 64% of the generated neurons were still alive compared with 39% of the controls. Intrastriatal grafting of an engineered neural stem cell line to overexpress PSPN in rats promoted survival of dopaminergic neurons, but these cells preferentially differentiated into oligodendrocytes . It would be interesting to show the differentiation potential of neural stem cells after overexpression of PSPN together with TGF-β. The results of the present study suggest that the factor combination TGF-β/PSPN is a potent inductor of midbrain dopaminergic neurons and suggest its application in protocols for the generation of dopaminergic neurons from embryonic stem cells. Moreover, these data provide new insights for tissue engineering using neural stem cells.