Co-Transplantation of GDNF-Overexpressing Neural Stem Cells and Fetal Dopaminergic Neurons Mitigates Motor Symptoms in a Rat Model of Parkinson’s Disease

Striatal transplantation of dopaminergic (DA) neurons or neural stem cells (NSCs) has been reported to improve the symptoms of Parkinson’s disease (PD), but the low rate of cell survival, differentiation, and integration in the host brain limits the therapeutic efficacy. We investigated the therapeutic effects of intracranial co-transplantation of mesencephalic NSCs stably overexpressing human glial-derived neurotrophic factor (GDNF-mNSCs) together with fetal DA neurons in the 6-OHDA rat model of PD. Striatal injection of mNSCs labeled by the contrast enhancer superparamagnetic iron oxide (SPIO) resulted in a hypointense signal in the striatum on T2-weighted magnetic resonance images that lasted for at least 8 weeks post-injection, confirming the long-term survival of injected stem cells in vivo. Co-transplantation of GDNF-mNSCs with fetal DA neurons significantly reduced apomorphine-induced rotation, a behavioral endophenotype of PD, compared to sham-treated controls, rats injected with mNSCs expressing empty vector (control mNSCs) plus fetal DA neurons, or rats injected separately with either control mNSCs, GDNF-mNSCs, or fetal DA neurons. In addition, survival and differentiation of mNSCs into DA neurons was significantly greater following co-transplantation of GDNF-mNSCs plus fetal DA neurons compared to the other treatment groups as indicated by the greater number of cell expressing both the mNSCs lineage tracer enhanced green fluorescent protein (eGFP) and the DA neuron marker tyrosine hydroxylase. The success of cell-based therapies for PD may be greatly improved by co-transplantation of fetal DA neurons with mNSCs genetically modified to overexpress trophic factors such as GDNF that support differentiation into DA cells and their survival in vivo.


Introduction
Parkinson's disease (PD) is one of the most prevalent diseases of aging. Although it has been recognized for many years that PD results from the degeneration of dopaminergic (DA) neurons in the nigrostriatal system [1], current therapies such as dopamine replacement and stereotaxic surgery yield only modest improvements in symptoms [2] and there is no treatment to halt disease progression.
Cell-based therapies for PD aim to replace lost nigral DA neurons with functional DA neurons derived from autologous or homologous neural stem cells (NSCs) or fetal DA neurons [3], and indeed both animal studies and clinical reports have provided evidence for the potential of NSCs (especially those derived from ventral mesencephalon) to differentiate into DA neurons in the adult mammalian brain [4,5]. Thus, transplantation of NSCs [6] or fetal DA neurons [7][8][9] provide a potential therapeutic option for late-stage PD, and both striatal reinnervation and behavioral recovery have been observed in animal models and PD patients. However, the limited survival and poor function integration of transplanted cells remain major impediments to full clinical efficacy [10][11][12][13][14].
In an effort to improve the therapeutic effects of cell transplantation, we co-grafted mesencephalic neural stem cells (mNSCs) modified to express glial-derived neurotrophic factor (GDNF) together with fetal DA neurons into the 6-hydroxydopamine (6-OHDA) rat model of PD. This co-transplantation strategy reduced apomorphine-induced rotation, a Parkinsonian motor symptom in rats, improved NSCs survival in vivo, and promoted greater differentiation of NSCs into DA neurons compared to separate transplantation of NSCs or fetal DA neurons alone. This co-transplantation strategy may improve the clinical outcome of cell-based therapy for treatment of PD.

Cell Culture
The human astrocytoma cell line U251 was purchased from the Cell Bank of the Chinese Academy of Sciences, Shanghai, China. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies, USA) containing 10% heat-inactivated fetal bovine serum (FBS) (Life Technologies, Catalogue number [No.]: 12483020) at 37uC under a humidified 5% CO 2 and 95% air atmosphere.

Animals
Fetal Sprague-Dawley (SD) rats at embryonic day 14 (E14) and E18 and adult male SD rats were obtained from Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China). Adult males (body weight: 250610 g; 5-6 weeks of age) were housed at 20225uC and 50%65% humidity with ad libitum access to food and water under a 12:12 h light/dark cycle. The protocols in this study were approved by institutional review board and the Animal Care and Use Committee of Kunming Medical University (2009KMUL029).

Isolation and Culture of Ventral Mesencephalic Neural Stem Cells (mNSCs) and Fetal DA Neurons
Ventral mesencephalic neural tissue was obtained from fetal SD rats at E14 for isolation of mNSCs and at E18 for isolation of fetal DA neurons. Single cell suspensions were prepared as described in a previous report [15].

Superparamagnetic Iron Oxide (SPIO) and Enhanced Green Fluorescent Protein (eGFP) Double-Labeling of GDNF-Overexpressing mNSCs
The cDNA sequence of human GDNF (GenBank BC069369) was amplified from the human astrocytoma cell line U251 using forward 59-CGAAGCTTGCCACCATGAAGTTATGG-GATGTCGTG-39 (containing a HindIII site, a Kozak sequence, and an ATG start codon) and reverse 59-GCGGATCCCAGATA-CATCCACACCTTTTAGC-39 (containing a BamHI site). The human GDNF PCR product was sequenced to verify accuracy, then digested with HindIII and BamHI and ligated into the plasmid vector peGFPN1 (Clontech, Germany) containing a reporter gene encoding enhanced green fluorescent protein (eGFP) to yield the  recombinant plasmid peGFPN1-GDNF. The sequence and restriction map of peGFPN1-GDNF were verified and the plasmid was then transfected into mNSCs using the Amaxa TM Basic Nucleofector Kit (Germany, VPI-1003) for primary mammalian neurons.

RT-PCR
After GDNF-mNSCs were selected in medium containing G418, total RNA was isolated using the RNeasy Plus Mini Kit (QIAGEN, 74134) and purity determined by absorbance at 260 and 280 nm (A260/280). The integrity of the RNA was verified by formaldehyde denaturing gel electrophoresis. cDNAs were synthesized using the Revert Aid TM First Strand cDNA Synthesis Kit (MBI Fermentas, K1611) according to the manufacturer's instructions. The cDNAs were amplified using the primers 59-AGTTATGGGATGTCGTGG-39 (forward) and 59-GAAG-CACTGCCATTTGTT-39 (reverse) to determine GDNF mRNA expression (product length: 255 bp). Expression was normalized to that of b-actin (amplified using forward 59-GGACATCCGTAAA-GACCTG-39 and reverse 59-AAAGGGTGTAACGCAACTAA-39, product length: 300 bp). All transcripts were amplified by initial denaturation at 94uC for 3 min, 25 cycles of 94uC for 45 s, 50uC for 1 min, and 72uC for 1 min, and a final extension at 72uC for 10 min. The PCR products were subjected to 1% agarose gel electrophoresis and the abundance of each mRNA was normalized to b-actin using ImageJ software (NIH, USA).

Immunofluorescence
For immunocytochemistry staining, mNSCs were cultured for 1 hour in DMEM/F-12 lacking bFGF and EGF, but containing 1% bovine serum. The cells were washed three times in PBS, fixed in 4% paraformaldehyde for 10 min at room temperature (RT), and permeabilized with PBS plus 0.25% Triton X-100 (PBST) for 10 min at RT (or 15 min at RT if the target antigen was a nucleoprotein). After three washes in PBS, the cells were blocked with 1% BSA for 30 min at RT and incubated with anti-nestin (mouse monoclonal, 1:1000, Abcam, No. ab6142) and anti-GDNF (rabbit polyclonal, 1:20, No. ab18956) overnight at 4uC. Immunolabeling was visualized by incubation (1 h at RT) with rhodamine-conjugated Affinipure goat anti-mouse or anti-rabbit IgG (1:100, Zymed, USA). After three washes in PBS and two in deionized water, immunolabeled cells were counterstained with 50 ml DAPI at 37uC for 10 min and washed in sequence with methanol and ethanol. Cells were observed under a fluorescence microscope (OLYMPUS, Japan).

Western Blotting
Cells or tissues were lysed in Neuronal Protein Extraction Reagent (N-PER, Pierce, USA) on ice for 20 min and centrifuged at 13000 rpm for 5 min at 4uC. The supermatant protein concentrations were measured by the Bradford assay (Pierce, USA). Equal amounts of protein per gel lane were separated by 15% SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (Whatman, Germany). Membranes were blocked with 5% dry milk/Tris-buffered saline (TBS) for 1 h and then incubated overnight at 4uC with a GDNF antibody (rabbit polyclonal, 1:50, Abcam, No. ab18956) and b-actin antibody (Abcam) in 5% dry milk/TBS. Following three washes with 0.1% Tween 20/TBS (TBST), blots were incubated with horseradish peroxidase (HRP)-conjugated anti-rabbit IgG (1:5000, Abcam) at room temperature for 1 h and developed using a chemiluminescence system (KPL, USA). The relative protein expression was normalized to b-actin using Image J software.  [7]. One week after surgery, apomorphine (APO, 0.25 mg/kg, Sigma,) was injected (i.p.) to induce rotational behavior, a behavioral endophenotype of this PD model. Beginning 10 minutes after APO administration, the absolute number of rotations over the next 30 min was counted, and the average rotations per minute (RPM) were calculated. The test was performed after experimental treatments (Table 1) as a measure of the severity of PD-like motor symptoms. Rats with an average RPM of less than six in the first 30 min after apomorphine administration were excluded from the transplantation experiments [23]. Thirty-six PD model rats were used for subsequent transplantation experiments. Approximately 2.5610 5 cells in 10 ml were injected into the ipsilateral striatum (right side) at A/P = +0.5 mm; L = +3.0 mm; V = 25.0 mm using a Hamilton syringe. Sham-treated rats (n = 6) were punctured at the same coordinates with the empty syringe. All other treatments were identical. Experimental animals were injected with fetal DA neurons, mNSCs expressing empty vector (control mNSCs), GDNF-mNSCs, control mNSCs+fetal DA neurons, and GDNF-mNSCs+fetal DA neurons (each n = 6). Apomorphine-induced rotational behavior was examined 2, 4, 6 and 8 weeks after cell transplantation. The ratio of the APOevoked rotation rate after transplantation to the rotation rate before transplantation was used as a measure of improved motor function [24].

Survival, Migration, and Differentiation of Transplanted Cells
Survival and migration of transplanted stem cells were determined by MRI using intracellular SPIO as a contrast agent. All animals received an MRI scan 1, 2, 4 and 8 weeks after cell transplantation. Animals were anesthetized with 2% sodium pentobarbital (i.p.), mounted onto the stereotaxic frame to hold the animal, and MRI scans acquired using a Philips Intera 1.5 T MRI system (Netherlands) with a 5-inch surface coil. A three planar positioning scan was followed by gradient-echo T2weighted (T2W/FFE) imaging using the following acquisition parameters: Repetition Time (TR) 388 ms, Echo Time (TE) 23 ms, flip angle 18u, field of view (FOV) 120 mm, matrix 2056256, slice thickness 2.0 mm, and scan time 2 min 8 s.

Statistical Analyses
The SPSS12.0 statistical software package was used for statistical analyses. Results are presented as means 6 standard deviation (SD). Statistical significance was evaluated by t-test or one-way analysis of variance (ANOVA) with least significant difference (LSD) test for post hoc analysis. A P,0.05 was considered statistically significant.

Identification of SPIO/eGFP Double-labeled GDNF-mNSCs and Fetal DA Neurons in vitro
Mesencephalic neural stem cells (mNSCs) were isolated from E14 rats as described (Figure 1). Three days after plating, most cells were solitary or in small aggregates of a few cells, but many exhibited morphological signs of mitosis ( Figure 1A). After 4 days in vitro (DIV), these proliferating cells formed larger spherical colonies (neurospheres) that grew in size over the next 325 days and eventually detached from the substrate ( Figure 1B, 1C). These neurospheres were strongly immunopositive for the stem cell marker nestin ( Figure 1D). Fetal dopaminergic (DA) neurons were isolated from the ventral midbrain of E18 rats and identified in culture by the elaboration of neurites ( Figure 2B2D) and immunoreactivity for the DA synthesis enzyme tyrosine hydroxylase ( Figure 2C, 2D) and the mature neuron marker b-III-tubulin ( Figure 2F).
Neurospheres from cells stably transfected with the peGFPN1-GDNF plasmid demonstrated much higher GDNF protein expression than mNSCs transfected with the empty peGFPN1 plasmid ( Figure 3A). These GDNF-mNSCs also expressing eGFP were then labeled with SPIO for in vivo detection by MRI. Prussian blue staining and transmission electron microscopy confirmed SPIO particles in the cytoplasm of mNSCs ( Figure 3B). Furthermore, RT-PCR indicated enhanced GDNF mRNA expression in GDNF-mNSCs (about 2.1-fold higher than that in mNSCs transfected with the empty peGFPN1 vector) ( Figure 3C), while Western blotting revealed approximately 2.2fold higher GDNF expression ( Figure 3D).

Success Rate of the PD Model
Rats were injected with the dopaminergic neurotoxin 6-OHDA as described and then examined for a PD-like behavioral phenotype using the apomorphine (APO)-induced rotation assay. Of the 70 rats injected with 6-OHDA, 39 exhibited a minimum of 6 rotations/min over the first 30 min post-injection (55.7%), the criterion for successful PD-model induction.

Reduction of APO-induced Rotational Behavior after Cell Transplantation
Parkinson's disease model rats (n = 36) were divided into 6 equal groups. One group was sham-treated while the other 5 groups were injected with fetal DA neurons (isolate and identified as in Figure 2), control mNSCs, GDNF-mNSCs, control mNSCs plus fetal DA neurons, or GDNF-mNSCs plus fetal DA neurons. All groups that received cell transplantation demonstrated a reduction in APO-induced rotation behavior as expressed by the after/ before transplant rotational ratio compared to the sham-treated group ( Table 1). The largest decrease in rotational ratio compared to the sham-operation group (1.27260.032) was observed in the GDNF-mNSCs+DA group (0.24260.039), (P,0.01) ( Table 1). Thus, transplantation of multipotent stem cells or fetal DA neurons improved a behavioral endophenotype of PD, but transplantation of both fetal DA neurons and GDNF-overexpressing mNSCs resulted in the largest degree of improvement.

Survival, Migration, and Differentiation of Transplanted Cells
Magnetic resonance images of anesthetized rats ( Figure 5) and immunohistochemical staining of brain sections through the striatum ( Figure 6) showed that almost all the injected cells were retained at the transplantation site.  Table 2). Indeed, the rate of mNSCs differentiation into DA neurons was markedly higher in the GDNF-mNSCs+DA group than in rats injected with control mNSCs (7.3%60.8%) or control mNSCs plus DA neurons (13.2% 61.0%). While GDNF-mNSCs also demonstrated a high differentiation rate when injected alone (18.3%61.9%), the total number of mNSCs-derived DA neurons was far lower than in the GDNF-mNSCs+DA group (86612 vs. 2063 TH + /eGFP + cells) due to the smaller population of viable injected mNSCs (381615 vs. 10867 eGFP+ cells). Moreover, while co-injection of fetal DA neurons and control mNSCs also resulted in a relatively large number of surviving stem cells (eGFP+ cells), the lower rate of differentiation (13.2% 61.0%) resulted in many fewer mNSCs-derived DA neurons. Thus, overexpression of GDNF and the presence of fetal DA neurons act synergistically to enhance stem cell survival and differentiation into DA neurons.

Discussion
Although numerous studies have shown that transplantation of fetal DA neurons or NSCs can improve the abnormal behavior of PD model animals [6][7][8][9], the therapeutic effect of cell transplantation is both limited and transient due to the low survival rate of transplanted DA neurons and the low efficiency of NSCs differentiation into DA neurons in the local microenvironment of the host [10][11][12][13][14]. Thus, in order to obtain a significant therapeutic effect, cell-based therapies must not only to replace lost DA neurons but must also modify the local host microenvironment to make it more conducive to the survival and differentiation of grafted DA neurons or NSCs.
Glial-derived neurotrophic factor is the most powerful trophic factor for DA neurons [6,7], suggesting that augmenting the local GDNF concentration may enhanced graft survival. However, GDNF alone is not easily delivered to the mesencephalon as it does not cross the blood-brain barrier [25][26][27] and has a limited half-life in vivo. Thus, alternative routes are needed [28][29][30], such as genetically modified grafted cells that stably overexpress GDNF as demonstrated in the current study.
With the advent of molecular imaging technologies, in vivo realtime monitoring of transplanted cell fate has become a reality. Cells labeled with MR contrast agents prior to transplantation can be detected by MRI [31]. In this study, we labeled neural stem cells with the superparamagnetic iron oxide (SPIO) MRI contrast agent Feridex using a liposomal transfection agent. Most mNSCs remained at the injection sites in the striatum and many expressed tyrosine hydroxylase, the rate limiting enzyme for DA synthesis. Local release of DA from these cells may account for the suppression of APO-induced rotational behavior observed in all transplantation groups, but especially in rats injected with both fetal DA neurons and mNSCs overexpressing GDNF.
In addition to a greater reduction in rotational behavior, cotransplantation of GDNF-mNSCs with fetal DA neurons improved grafted cell survival and promoted the differentiation of mNSCs into DA neurons at the transplantation site compared to transplantation of GDNF-mNSCs alone or to co-transplantation of control mNSCs with fetal DA neurons. Previous publications have reported that only 3%-20% of grafted DA neurons survived [14], and that only <10% of mNSCs could be induced to differentiate into DA neuron in vitro or in vivo [32,33]. Results of our study demonstrated the survival of a significantly greater proportion of grafted cells in the GDNF-mNSCs+DA group than in the other groups. About 22.5% of GDNF-overexpressing mNSCs differentiated into DA neurons when co-grafted with fetal mesencephalic DA neurons, a rate considerably higher than in some previous reports [32][33][34]. In contrast, only 7.3% of mNSCs differentiated into (TH + ) DA neurons when transplanted alone, more consistent with previous reports [32,33]. Nikkhah et al. suggested that 100 surviving TH+ neurons are sufficient to produce therapeutic effects in the APO-induced rotation test [15], and we surmise that the large number of surviving TH+ neurons in our study at least partially accounts for the significant (81%) decrease in APO-induced rotational behavior in the GDNF-mNSCs+DA group compared to the sham group 8 weeks after 6-OHDA injection.
Our study also showed that some autologous NSCs (Nestin + / eGFP-cells) migrated to the transplantation site, which might also contribute to the recovery. We propose that the better outcome obtained by co-transplantation of GDNF-overexpressing mNSCs with fetal DA neurons is due to both GDNF expression at the transplantation site and cooperative interactions between the grafted mNSCs and DA neurons.
We demonstrate that co-injection of fetal DA neurons with mesencephalic neural stems cells overexpressing GDNF (GDNF-mNSCs) substantially enhances graft survival, NSCs differentiation to DA neurons, and functional improvement in a rat model of Parkinson's disease. This strategy may improve the efficacy of cellbased therapies for PD.