OUP user menu

Embryonic stem cell-derived L1 overexpressing neural aggregates enhance recovery in Parkinsonian mice

Yi-Fang Cui, Gunnar Hargus, Jin-Chong Xu, Janinne Sylvie Schmid, Yan-Qin Shen, Markus Glatzel, Melitta Schachner, Christian Bernreuther
DOI: http://dx.doi.org/10.1093/brain/awp290 189-204 First published online: 7 December 2009

Summary

Parkinson's disease is the second most common neurodegenerative disease, after Alzheimer's disease, and the most common movement disorder. Drug treatment and deep brain stimulation can ameliorate symptoms, but the progressive degeneration of dopaminergic neurons in the substantia nigra eventually leads to severe motor dysfunction. The transplantation of stem cells has emerged as a promising approach to replace lost neurons in order to restore dopamine levels in the striatum and reactivate functional circuits. We have generated substrate-adherent embryonic stem cell-derived neural aggregates overexpressing the neural cell adhesion molecule L1, because it has shown beneficial functions after central nervous system injury. L1 enhances neurite outgrowth and neuronal migration, differentiation and survival as well as myelination. In a previous study, L1 was shown to enhance functional recovery in a mouse model of Huntington's disease. In another study, a new differentiation protocol for murine embryonic stem cells was established allowing the transplantation of stem cell-derived neural aggregates consisting of differentiated neurons and radial glial cells into the lesioned brain. In the present study, this embryonic stem cell line was engineered to overexpress L1 constitutively at all stages of differentiation and used to generate stem cell-derived neural aggregates. These were monitored in their effects on stem cell survival and differentiation, rescue of endogenous dopaminergic neurons and ability to influence functional recovery after transplantation in an animal model of Parkinson's disease. Female C57BL/6J mice (2 months old) were treated with the mitochondrial toxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine intraperitoneally to deplete dopaminergic neurons selectively, followed by unilateral transplantation of stem cell-derived neural aggregates into the striatum. Mice grafted with L1 overexpressing stem cell-derived neural aggregates showed better functional recovery when compared to mice transplanted with wild-type stem cell-derived neural aggregates and vehicle-injected mice. Morphological analysis revealed increased numbers and migration of surviving transplanted cells, as well as increased numbers of dopaminergic neurons, leading to enhanced levels of dopamine in the striatum ipsilateral to the grafted side in L1 overexpressing stem cell-derived neural aggregates, when compared to wild-type stem cell-derived neural aggregates. The striatal levels of gamma-aminobutyric acid were not affected by L1 overexpressing stem cell-derived neural aggregates. Furthermore, L1 overexpressing, but not wild-type stem cell-derived neural aggregates, enhanced survival of endogenous host dopaminergic neurons after transplantation adjacent to the substantia nigra pars compacta. Thus, L1 overexpressing stem cell-derived neural aggregates enhance survival and migration of transplanted cells, differentiation into dopaminergic neurons, survival of endogenous dopaminergic neurons, and functional recovery after syngeneic transplantation in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine mouse model of Parkinson's disease.

  • L1
  • embryonic stem cells
  • neural aggregates
  • MPTP
  • dopaminergic neurons

Introduction

Parkinson's disease is the most common movement disorder characterized by a progressive loss of dopaminergic neurons in the substantia nigra pars compacta with decreased levels of dopamine in the striatum leading to bradykinesia, rigidity, tremor and postural instability in affected patients (Samii et al., 2004). Current established therapeutic approaches involving drug treatment and deep brain stimulation can temporarily ameliorate symptoms but cannot cure the disease. Transplantation of human foetal neural (ventral mesencephalic) tissue into the putamen or caudate nucleus has evolved as a potentially curative cell replacement therapy with long term survival of grafted cells (Kordower et al., 2008; Li et al., 2008; Mendez et al., 2008). Variable success and graft-induced dyskinesias (Freed et al., 2001; Olanow et al., 2003) are currently addressed in clinical studies by careful selection of patients and modification of cell preparation and transplantation. Yet, the low availability of foetal tissue remains a major disadvantage of ventral mesencephalic grafts. In contrast, neural stem cells and especially embryonic stem cells, are a more abundant source since they proliferate under appropriate culture conditions in vitro representing a replenishable mode for cell replacement therapy. These cells can be differentiated into dopaminergic neurons by applying exogenous factors or by genetic manipulation, and have shown enhanced functional recovery in rat models of Parkinson's disease (Studer et al., 1998; Lee et al., 2000; Kim et al., 2002, 2006; Park et al., 2005; Hermann et al., 2006). A major problem for cell replacement therapy in Parkinson's disease animal models is impaired survival of transplanted cells (Grothe et al., 2004; Park et al., 2005; Suon et al., 2006). Teratoma or tumour formation remains another obstacle concerning the transplantation of embryonic stem cells that has to be addressed before human embryonic stem cells can be safely applied in Parkinson's disease (Nishimura et al., 2003; Brederlau et al., 2006). In a previous study we developed a protocol for the transplantation of substrate adherent embryonic stem cell-derived neural aggregates (SENAs), cultures enriched in neurons and radial glial cells, that showed enhanced survival of transplanted cells with reduced teratoma formation after syngeneic transplantation in a mouse model of Huntington's disease (Dihne et al., 2006), when compared to cells differentiated by the five-stage differentiation protocol (Lee et al., 2000). Furthermore, we could show that L1 overexpressing embryonic stem cells pre-differentiated by the five-stage protocol increased neuronal differentiation of transplanted cells and enhanced functional recovery in a mouse model of Huntington's disease (Bernreuther et al., 2006). It had been shown previously that L1 promotes survival of foetal dopaminergic neurons in vitro (Hulley et al., 1998).

In the present study, we were thus interested to investigate whether L1 overexpressing SENAs, when transplanted into the striatum or substantia nigra of 1-methyl-4-phenyl-1,2,3,6 tetrahydropyridine (MPTP)-lesioned young adult mice, would enhance functional regeneration. Here we report that L1 overexpressing SENAs show increased numbers of dopaminergic neurons leading to enhanced striatal dopamine levels and functional recovery after transplantation into the striatum in MPTP-lesioned mice when compared to wild-type SENAs. Furthermore, L1 overexpressing SENAs rescued imperilled endogenous dopaminergic neurons more so than wild-type SENAs when transplanted adjacent to the lesioned substantia nigra.

Materials and methods

Generation and isolation of L1 overexpressing SENAs

In this study, we used embryonic stem cells derived from C57BL/6J mice expressing enhanced green fluorescent protein (EGFP) under the control of the chicken beta-actin promoter (Okabe et al., 1997) that had been transfected to overexpress the neural cell adhesion molecule L1 (Lindner et al., 1983; Rathjen and Schachner, 1984; for a recent review see Maness and Schachner, 2007) under the control of the promoter of isoform 1 of the 3-phosphoglycerokinase at all stages of differentiation (Bernreuther et al., 2006). The protocol described by Dihne et al. (2006) was used to differentiate L1 overexpressing and sham-transfected, wild-type EGFP-positive embryonic stem cells into SENAs. This protocol is a modification of the five-stage protocol described by Lee et al. (2000). Briefly, after propagation of undifferentiated stem cells on mitomycin C-inactivated (Roth, Karlsruhe, Germany) embryonic fibroblasts (stage 1), embryoid bodies were formed in hanging drops consisting of 800 cells per 20 µl drop for 2 days and kept for an additional 2 days in bacterial Petri dishes (stage 2). Selection of nestin-positive cells was performed for 8 days in serum-free medium containing 5 μg/ml insulin, 50 μg/ml transferrin, 30 nM selenium chloride and 5 μg/ml fibronectin (stage 3). Surviving cells were seeded onto poly-l-ornithine-coated cell culture dishes and maintained for 4 weeks under the influence of fibroblast growth factor-2 (prolonged stage 4) followed by fibroblast growth factor-2 withdrawal for 7 days to induce terminal differentiation of SENA-derived immature neurons (stage 5).

To obtain highly enriched preparations of SENAs, cultures were treated with 0.3 μg/ml collagenase XI (Sigma, St. Louis, MO, USA) at 37°C for 10 min, followed by gentle shaking of the culture dish for 10 s to detach entire SENAs and neural precursor cells growing as monolayers from the substrate. Detached cells were carefully pipetted up and down 10 times to allow SENAs to remain intact during this procedure. SENAs were harvested and separated from the singly dispersed monolayer cells by gravity-induced sedimentation for 2 min and resuspended in phosphate-buffered saline (PBS), pH 7.4, after removal of the supernatant containing single cells from the monolayer. The SENAs were picked individually using a 10 μl pipette tip. For transplantation, harvested SENAs were centrifuged at 100×g for 5 min and resuspended in PBS at a concentration of 10 SENAs/μl. Cell number per SENA was estimated by dissociation with trypsin, showing that, on average, one SENA consisted of about 10 000 cells.

Immunoblot analysis

L1 overexpressing and wild-type SENAs were taken from day 28 of stage 4 and day 7 of stage 5 for immunoblot analysis of L1 expression. Protein concentration was measured using the BCA Protein Assay Kit (Pierce Biotechnology, Rockford, IL, USA) and immunoblotting was carried out as described (Dihne et al., 2003) using monoclonal rat anti-mouse L1 antibody 555 (1:1000; Appel et al., 1995). Mouse monoclonal antibodies against glyceraldehyde 3-phosphate dehydrogenase (1:3000; Chemicon, Temecula, CA) were used to control for protein loading.

Immunocytochemistry and immunohistochemistry

For immunocytochemistry, cultured cells were fixed in 4% formaldehyde in PBS, washed with PBS and incubated in 0.1% bovine serum albumin (Sigma) for 40 min. For immunohistochemistry, mice were perfused with 4% formaldehyde in 0.1 M PBS. Perfused brain tissue was soaked in 20% sucrose overnight at 4°C, frozen in liquid nitrogen-cooled 2-methyl-butane and cut on a cryostat to obtain 25 μm thick sections. Primary antibodies incubated with the sections at 4°C overnight were monoclonal mouse antibodies against β-tubulin III (1:400; Sigma), CNPase (1:1000; Sigma), NeuN (1:1000; Chemicon), synaptic vesicle protein 2 (1:100; Developmental Studies Hybridoma Bank, Iowa City, IA, USA); polyclonal rabbit antibodies against glial fibrillary acidic protein (GFAP; 1:1000; Dako, Carpinteria, CA), glutamate decarboxylase (GAD-65/67, 1:500; Sigma), tyrosine hydroxylase (1:800; Chemicon), caspase-3 (1:1000, R&D Systems, Minneapolis, MN, USA), Ki-67 (1:500, Abcam, UK), ionized binding calcium adapter molecule 1 (1:1500; Wako Chemicals, Richmond, VA); and monoclonal rat antibody against CD11b (1:100, Serotec, Kidlington, UK). For detection of first antibodies, appropriate secondary antibodies, coupled to Cy2 or Cy3 (all from Dianova, Hamburg, Germany), were used. Specimens were examined with a confocal laser-scanning microscope (LSM510; Carl Zeiss Microimaging).

Differentiation of SENAs in vitro

To determine the percentages of marker-positive cells among all cells, cells were counterstained with 4-,6-diamidino-2-phenylindole (DAPI, Sigma) and the ratio of cell type-specific, marker-positive cells to all DAPI-positive cells was calculated. At least three independent experiments in duplicates and at least 1000 cells per marker and experiment were analysed. Percentages of double-labelled cells were determined and mean values ± standard error of the mean (SEM) were calculated.

Transplantation

Four days prior to transplantation, 2 month old female C57BL/6J mice were lesioned by four intraperitoneal injections of 15 mg/kg body weight of MPTP (Sigma) dissolved in saline at 2 h intervals. On the day of transplantation, SENAs were harvested and resuspended in PBS at a concentration of 10 SENAs per μl. One microlitre of L1 overexpressing SENAs, wild-type SENAs, or PBS only were unilaterally injected into the striatum using the following coordinates in relation to bregma: 0.1 mm posterior, 2.4 mm mediolateral, and 3.6 mm dorsal. The number of animals studied with striatal transplantation was: L1 overexpressing SENAs (n = 29), wild-type SENAs (n = 29) and PBS (n = 22). In another group of animals, L1 overexpressing SENAs (n = 7), wild-type SENAs (n = 7), or PBS (n = 5) were unilaterally injected adjacent to the substantia nigra at the coordinates: 3.1 mm posterior, 1.3 mm mediolateral and 4.0 mm dorsal in relation to bregma. Grafts were analysed 4 and 12 weeks after transplantation. All animal experiments were approved by the University and State of Hamburg Animal Care Committees.

Graft volume and cell density in the graft

Unbiased estimates of the total number of grafted cells and graft volume per animal, 4 and 12 weeks after transplantation, were calculated according to the optical disector and Cavalieri methods (Howard and Reed, 1998). An Axioskop microscope (Carl Zeiss Microimaging) equipped with a motorized stage and Neurolucida software-controlled computer system (MicroBrightField Europe, Magdeburg, Germany) was used for quantitative analysis. Graft volume and cell density in the graft were determined by measuring every 10th section of the graft. Transplanted cells were identified by their EGFP expression. Graft areas were outlined on digitized images to calculate volumes. Using random sampling in the graft core and in the periphery of the graft, cell counts were performed according to the optical disector principle at a magnification of ×40. All counts were performed in a double-blinded manner.

Astroglial and microglial graft-induced response in the host tissue

To determine the reaction of the host tissue to the grafted cells, immunostaining with antibodies against GFAP to analyse astrogliosis and also against ionized binding calcium adapter molecule 1 and CD11b to estimate numbers and density of microglial cells was performed. Confocal images were taken in the vicinity of the edge of the graft in mice transplanted with L1 overexpressing or wild-type stem cells, and corresponding areas in mice sham-injected with PBS. At least 20 images were analysed in each mouse. The software Image J (http://rsbweb.nih.gov/ij/index.html) was used to measure fluorescence intensities from all three groups. Fluorescent intensities from the L1 overexpressing group and wild-type group were normalized to the PBS injected group. For statistical evaluation one-way ANOVA followed by Tukey's post hoc test was used.

Differentiation of SENAs in vivo

To determine differentiation of SENAs after transplantation in vivo, the ratio of double-labelled marker-positive and EGFP-positive cells among all EGFP-positive cells was determined by confocal laser scanning microscopy in serial sections and mean values ± SEM were calculated. At least 1000 cells per marker and experiment were analysed in a double-blinded manner.

Migration of grafted cells

The graft edge was outlined at low magnification (×5) in digitized images. The shortest distances of at least 400 individual cells from the graft edge of recipient animals were determined at higher magnification (×40).

Striatal dopamine levels

A commercial radioimmunoassay kit (Dopamine Research RIA, Labor Diagnostika Nord, Nordhorn, Germany) was used to measure the dopamine levels in the striatum ipsilateral and contralateral to the grafted side. One month after transplantation, mice were sacrificed, striata were dissected and kept in −80°C until analysis. On the day of analysis, striata were homogenized in 0.01 N hydrochloric acid containing 1 mM ethylenediaminetetraacetic acid and 4 mM sodium metabisulfite. Protein concentrations were measured using BCA Protein Assay Kit (Pierce). Samples were processed according to the manufacturer's instructions. For statistical evaluation one-way ANOVA followed by Tukey's post hoc test was used.

Striatal gamma-aminobutyric acid levels

A commercial enzyme-linked immunosorbent assay kit (GABA Research ELISA, Labor Diagnostika Nord, Nordhorn, Germany) was used to measure the gamma-aminobutyric acid (GABA) level in the striatum ipsilateral and contralateral to the grafted side. One month after transplantation, mice were sacrificed, striata were dissected and kept at 80°C until analysis. On the day of analysis, striata were homogenized in 0.01 N hydrochloric acid containing 1 mM ethylenediaminetetraacetic acid and 4 mM sodium metabisulfite. Protein concentrations were measured using BCA Protein Assay Kit (Pierce). Samples were processed according to the manufacturer's instructions. For statistical evaluation one-way ANOVA followed by Tukey's post hoc test was used.

Determination of tyrosine hydroxylase-positive cells in the substantia nigra

In the MPTP lesion paradigm, tyrosine hydroxylase-positive neurons were counted in the right and left substantia nigra, pars compacta of every 10th section throughout the entire extent of the substantia nigra pars compacta. An Axioskop microscope (Carl Zeiss) equipped with a motorized stage and a Neurolucida software-controlled computer system was used for quantitative analysis (MicroBrightField Europe, Magdeburg, Germany). The border of substantia nigra pars compacta and ventral tegmental area was delineated at lower magnification observing tyrosine hydroxylase immunostaining. Using random sampling in the substantia nigra pars compacta, cell counts were performed according to the optical disector principle at a magnification of ×40. All counts were performed in a double-blinded manner. As previous studies showed that cell death in the acute MPTP lesion model used in this study is confined to the first 7 days after MPTP application (Jackson-Lewis et al., 1995; Jakowec et al., 2004) the numbers of endogenous tyrosine hydroxylase-positive neurons in the substantia nigra were only analysed 1 month but not 3 months after transplantation in our study.

Determination of the density of tyrosine hydroxylase-positive axons in the striatum

To determine the striatal tyrosine hydroxylase-positive fibre density, immunostaining with antibodies against tyrosine hydroxylase was performed. Confocal images were taken in the vicinity of the edge of the graft (0–200 μm) and at 2 mm distance from the graft edge in the L1 overexpressing, wild-type and PBS groups. At least 20 images were analysed in each mouse. The software Image J (http://rsbweb.nih.gov/ij/index.html) was used to measure fluorescence intensities in all three groups. For statistical evaluation, fluorescence intensities from L1 overexpressing and wild-type groups were normalized to the PBS group and one-way ANOVA followed by Tukey's post hoc test was used.

Locomotor behaviour

Apomorphine-induced rotations were analysed 1 day before and 1, 3, 4, 6, 8, 10 and 12 weeks after transplantation, to evaluate the effects of transplantation on symmetry of motor function (Hudson et al., 1993; Da Cunha et al., 2008). Mice having received systemic MPTP-injections were randomly assigned to three groups that received a unilateral striatal transplantation of L1 overexpressing SENAs (n = 10), wild-type SENAs (n = 10), or a PBS injection only (n = 10) (sham-injected group). As nigral transplantation was only performed to monitor the effect of L1 overexpressing SENAs on the survival of endogenous tyrosine hydroxylase-positive neurons, no behavioural analysis was performed in mice receiving nigral grafts. Mice were tested for rotation in response to an intraperitoneal injection of 1 mg/kg apomorphine in PBS. All behavioural tests were performed at the beginning of the animals' dark phase cycle. Rotation was measured in an open field box for 30 min at 50 lux. Ethovision software (Noldus, Wageningen, The Netherlands) was used for processing data. Relative meander, which is net ipsilateral turning angle divided by distance, was calculated. For statistical evaluation one-way ANOVA followed by Tukey's post hoc test was used.

Results

Characterization of L1 overexpressing SENAs in vitro

In this study, we used embryonic stem cells expressing EGFP under the control of the chicken beta-actin promoter and the neural cell adhesion molecule L1 under the control of the 3-phosphoglycerokinase promoter at all stages of differentiation as described (Bernreuther et al., 2006). L1 overexpressing EGFP-positive embryonic stem cells and EGFP-positive wild-type cells were differentiated by the five-stage protocol (Lee et al., 2000) modified as described (Dihne et al., 2006). Briefly, proliferation of neural precursor cells under the influence of fibroblast growth factor-2 was prolonged in stage 4 giving rise to SENAs, neural aggregates consisting mainly of neurons and radial glial cells. Throughout differentiation (stages 4 and 5), L1 overexpressing cells showed enhanced levels of L1 in vitro when compared to wild-type cells as determined by western blot analysis (Fig. 1A).

Figure 1

L1 overexpression does not affect neuronal differentiation of SENAs in vitro. (A) Immunoblot analysis of L1 expression in SENAs generated from transfected L1 overexpressing (L1+) and mock-transfected wild-type (WT) embryonic stem cells expressing EGFP at all stages of differentiation at day 28 of prolonged stage 4 and day 7 of stage 5. Quantification of the immunoblot analysis of L1 expression indicates that L1+ SENAs showed enhanced L1 expression when compared with wild-type SENAs throughout differentiation. Normalized mean values ± SEM of L1 expression are shown. Student's t-test was performed for statistical analysis (*P < 0.05). The expression of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was measured to control for protein loading. (B) Generation of neurons from L1+ and wild-type SENAs was determined at day 7 of stage 5 by immunostaining and fluorescence light microscopy for β-tublin III (β-tub+ cells, red). EGFP+ cells (green), β-tub+/EGFP+ cells (yellow). Percentages of β-tub+ cells of all EGFP+ cells are shown at day 7 of stage 5 (mean ± SEM). No difference in neuronal differentiation was observed between L1+ and wild-type SENAs with a percentage of 90% neurons in both groups. Scale bar = 100 μm. (C) The proportion of tyrosin hydroxylase-positive (TH+) cells was determined at day 7 of stage 5 by immunohistochemistry. Tyrosine hydroxylase+ (red), EGFP+ cells (green), tyrosine hydroxylase+/EGFP+ cells (yellow). Scale bar = 100 μm. Percentages of tyrosine hydroxylase+ cells of all EGFP+ cells are shown at day 7 of stage 5 (mean ± SEM) Student's t-test was performed for statistical analysis.

To determine whether L1 influences neuronal differentiation of SENAs in vitro, cell type-specific expression of markers was determined by indirect immunofluorescence at day 7 of stage 5. SENAs contained a high percentage of β-tubulin III-positive neurons among all EGFP-positive cells (Fig. 1B) that did not differ between L1 overexpressing and wild-type group (L1: 93 ± 1%, control: 89 ± 1%). However we observed a tendency towards higher numbers of tyrosine hydroxylase-positive neurons in L1 overexpressing versus wild-type SENAs (L1: 4.5 ± 0.5%, control: 3.3 ± 0.5%) that was, however, statistically not significant (Fig. 1C).

Thus, in L1 overexpressing SENAs overall neuronal differentiation was not enhanced in vitro but a tendency towards increased dopaminergic differentiation was observed.

L1 overexpression in SENAs enhances graft size, number of surviving cells in the graft and migration from the graft edge after transplantation into the MPTP-lesioned mouse brain

Intraperitoneal injection of MPTP into young adult C57BL/6J female mice resulted in a reproducible reduction in the number of surviving tyrosine hydroxylase-positive cells in the substantia nigra (Supplementary Fig. 1) when compared to animals sham-injected with PBS 7 days after injection (MPTP: 5400 ± 500 cells; PBS: 10 600 ± 400 cells). To determine the effects of L1 overexpression on survival of transplanted SENAs and migration of cells from the graft into the host tissue, SENAs differentiated to day 7 of stage 5 were injected unilaterally into the striatum (Fig. 2A) or adjacent to the substantia nigra (Fig. 2B) 4 days after intraperitoneal injection of MPTP. One month after grafting L1 overexpressing SENAs showed increased graft size (Fig. 2C and D) and higher numbers of surviving cells (Fig. 2C and E) when compared to wild-type SENAs both after transplantation into the striatum (graft size—L1: 0.24 ± 0.03 mm3, wild-type: 0.11 ± 0.01 mm3; cell number—L1: 9300 ± 700, wild-type: 4600 ± 700) and adjacent to the substantia nigra (graft size—L1: 0.21 ± 0.03 mm3, wild-type: 0.1 ± 0.01 mm3; cell number—L1: 7100 ± 500, wild-type: 4100 ± 100). This effect was confirmed 3 months after transplantation of SENAs into the striatum (Fig. 2D and E). Furthermore, we observed increased migration of grafted cells from the graft edge of transplanted L1 overexpressing SENAs versus wild-type SENAs into the host striatum (Fig. 2F and G) and into the host substantia nigra (Fig. 2G) 1 month after grafting (striatum—L1: 126 ± 10 μm, wild-type: 52 ± 2 μm, substantia nigra—L1: 88 ± 4 μm, wild-type: 49 ± 4 μm). Occasionally, synaptic vesicle protein 2-positive synapses were observed abutting onto grafted cells that had migrated out of the graft edge, indicating that grafted cells had been contacted by host cells (Supplementary Fig. 2). To elucidate whether the increased numbers of surviving L1 overexpressing grafted cells were due to increased proliferation or decreased apoptosis of the grafted cells, numbers of Ki-67-positive and caspase-3-positive grafted cells were analysed 1 month after transplantation. No significant differences were observed in Ki-67-positive cells in L1-overexpressing versus wild-type grafts (Fig. 3A). In contrast, decreased numbers of caspase-3-positive cells were observed in L1 overexpressing versus wild-type grafts 1 month after transplantation indicating that L1 might exert its beneficial effects on the number of grafted cells by inhibiting apoptosis (Fig. 3B). No endogenous caspase-3-positive cells and Ki-67-positive cells were observed in the substantia nigra 1 month after transplantation of L1 overexpressing or wild-type SENAs (data not shown).

Figure 2

L1 overexpression in SENAs enhances graft size, number of surviving cells in the graft, and migration from the graft edge. SENAs differentiated to day 7 of stage 5 were transplanted into the mouse striatum (A) or the substantia nigra (B) 4 days after MPTP lesion. (C) Laser scanning microscopy of an L1 overexpressing (L1+) and a wild-type (WT) SENA graft 1 month after transplantation into the host striatum. Grafts were detected by green fluorescence of transplanted cells. Scale bar = 100 μm. (D) Graft volume 1 month and 3 months after transplantation of L1+ (n = 6) and wild-type (n = 6) SENAs into the striatum and 1 month after transplantation into the substantia nigra (mean ± SEM). Student's t-test was performed for statistical analysis (*P < 0.05, **P < 0.01). (E) Number of EGFP+ cells 1 month and 3 months after transplantation of L1+ (n = 6) and wild-type (n = 6) SENAs into the striatum and 1 month after transplantation into the substantia nigra (mean ± SEM). Student's t-test was performed for statistical analysis (**P < 0.01, ***P < 0.001). (F) Laser scanning microscopy of the periphery of L1+ (n = 6) and wild-type SENA grafts 1 month after transplantation. White line indicates graft edges. Note the enhanced number of EGFP+ cells migrated from the L1+ graft. Scale bar = 100 μm (G). Migration distance from the edge of L1+ (n = 6) and wild-type (n = 6) SENA grafts 1 month after transplantation into the striatum and 1 month after transplantation into the substantia nigra is shown (mean ± SEM). Student's t-test was performed for statistical analysis. (***P < 0.001).

Figure 3

L1 overexpression decreases apoptosis in grafted cells but does not influence proliferation of grafted cells. (A) Confocal images of L1 overexpressing (L1+) and wild-type (WT) SENA grafts (green) immunostained with antibodies against Ki-67 (red), 1 month after transplantation into the striatum. Scale bar = 50 μm. Percentages of Ki-67+ cells of all EGFP+ cells 1 month after transplantation of L1+ (n = 6) and wild-type (n = 6) SENAs into the striatum or substantia nigra are shown (mean ± SEM). Student's t-test was performed for statistical analysis. (B) Confocal images of L1+ and wild-type SENA grafts (green) immunostained with antibodies against caspase-3 (casp-3, red), 1 month after transplantation into the striatum. Scale bar = 50 μm. Percentages of caspase-3+ cells of all EGFP+ cells 1 month after transplantation of L1+ (n = 6) and wild-type (n = 6) SENAs into the striatum are shown (mean ± SEM), *P < 0.05.

To analyse whether the increased numbers of surviving cells in L1 overexpressing SENAs were due to differences in the reaction of the host tissue to the graft, immunohistological analysis of GFAP-positive astrocytes was performed to assess reactive astrogliosis in the host tissue surrounding the graft. No differences were observed between mice transplanted with L1 overexpressing SENAs or wild-type SENAs (Supplementary Fig. 3A, B). Furthermore, the microglial reaction of the host tissue was analysed by immunohistochemical analysis with antibodies against ionized binding calcium adapter molecule 1 (Supplementary Fig. 3C, D) and CD11b (Supplementary Fig. 3E, F) revealing a slightly, but significantly decreased density of ionized binding calcium adapter molecule 1-positive and CD11b-positive cells in the tissue surrounding L1 overexpressing grafts when compared to wild-type grafts and sham-injected PBS (Supplementary Fig. 3C–F).

Thus, graft size, cell number and migration into the host tissue are enhanced in L1 overexpressing versus wild-type SENAs after transplantation into the striatum or adjacent to the substantia nigra of MPTP-lesioned mice. Furthermore, apoptosis of grafted cells as determined by caspase-3 immunohistochemistry was decreased in L1 overexpressing versus wild-type grafts, while no difference in proliferation of grafted cells as determined by Ki-67 immunohistochemistry and astroglial reaction of the host tissue to the grafted cells was observed.

Enhanced numbers of tyrosine hydroxylase-positive cells in L1 overexpressing SENAs after transplantation into the MPTP-lesioned mouse brain

Percentages of NeuN-positive neurons among all EGFP-positive cells (Fig. 4A) were slightly but significantly enhanced in L1 overexpressing SENAs when compared to wild-type SENAs 1 month and 3 months after transplantation into the striatum (1 month—L1: 46 ± 2%, control: 39 ± 1%; 3 months—L1: 40 ± 2%, wild-type: 33 ± 1%) and 1 month after transplantation adjacent to the substantia nigra (L1: 42 ± 1%, control: 37 ± 1%). GFAP-positive astrocytes were less abundant in L1 overexpressing SENAs (Fig. 4B) 1 month and 3 months after transplantation into the striatum (1 month—L1: 41 ± 2%, wild-type: 53 ± 2%; 3 months—L1: 49 ± 2%, wild-type: 57 ± 2%). The percentage of oligodendrocytes was negligible in both groups and amounted to <1% of all EGFP-positive cells (data not shown).

Figure 4

L1 overexpressing SENAs show increased neuronal differentiation and decreased astrocytic differentiation after transplantation into the mouse striatum or adjacent to the substantia nigra after MPTP lesion. (A) Confocal images of L1 overexpressing (L1+) and wild-type (WT) SENA grafts (green) immunostained for the neuronal marker neuronal nuclear antigen (NeuN, red) 1 month after transplantation into the striatum. Scale bar = 50 μm. In the upper right, a Z-stack of 15 images of 1 µm thickness of the area outlined by a square in the merged image of the L1+ graft is shown with orthogonal views of the xz- and yz-planes showing EGFP+/NeuN+ neurons. Scale bar = 50 µm. Percentages of NeuN+ cells of all EGFP+ cells 1 month and 3 months after transplantation of L1+ (n = 6) and wild-type (n = 6) SENAs into the striatum and 1 month after transplantation into the substantia nigra are shown (mean ± SEM). Student's t-test was performed for statistical analysis (*P < 0.05). (B) Confocal images of L1+ and wild-type SENAs (green) immunostained with an antibody against GFAP (red) 4 weeks after transplantation. Scale bar = 50 µm. In the upper right, a Z-stack of 15 images of 1 µm thickness of the area outlined by a square in the merged image of the L1+ graft is shown with orthogonal views of the xz- and yz-planes showing EGFP+/GFAP+ astrocytes. Scale bar = 50 µm. Percentages of GFAP+ cells of all EGFP+ cells 1 month and 3 months after transplantation of L1+ (n = 6) and wild-type (n = 6) SENAs into the striatum and 1 month after transplantation into the substantia nigra are shown (mean ± SEM). Student's t-test was performed for statistical analysis (*P < 0.05, **P < 0.01).

The percentage of GABAergic neurons as identified by immunohistochemical labelling with glutamate decarboxylase antibodies (Fig. 5A) did not differ between the groups at all investigated time points after transplantation (Fig. 5A). In contrast, the percentage of dopaminergic neurons of all grafted EGFP-positive cells as determined by immunohistochemical labelling with tyrosine hydroxylase antibodies was enhanced in L1 overexpressing SENAs versus wild-type SENAs 1 month and 3 months after transplantation into the striatum (Fig. 5B; 1 month: L1: 5.5 ± 0.5%, wild-type: 3.5 ± 0.4%; 3 months: L1: 2.4 ± 0.6%, wild-type: 1.0 ± 0.1%) and 1 month after transplantation adjacent to the substantia nigra (Fig. 5B; L1: 5.3 ± 0.6%, wild-type: 3.6 ± 0.2%). To monitor the effects of L1 overexpressing SENAs on endogenous tyrosine hydroxylase-positive axons, the immunofluorescence intensity of axons labelled with antibodies against tyrosine hydroxylase was measured in the host tissue in close (up to 200 μm) and remote (2 mm) distance to the graft. The density of endogenous tyrosine hydroxylase-positive axons was increased close to the graft in L1 overexpressing versus wild-type SENAs while no differences were detected in remote distance to the graft (Supplementary Fig. 4).

Figure 5

L1 overexpressing SENAs show increased differentiation into tyrosine hydroxylase-positive neurons with unchanged GABAergic differentiation after transplantation into MPTP-lesioned mouse striatum or substantia nigra. (A) Confocal images of L1 overexpressing (L1+) and wild-type (WT) SENA grafts (green) immunostained for glutamate decarboxylase (GAD, red), a marker for GABAergic neurons, 1 month after transplantation into the striatum. Scale bar = 50 μm. In the upper right, a Z-stack of 15 images of 1 µm thickness of the area outlined by a square in the merged image of the L1+ graft is shown with orthogonal views of the xz- and yz-planes showing EGFP+/glutamate decarboxylase+ neurons. Scale bar = 50 µm. Percentages of glutamate decarboxylase+ cells of all EGFP+ cells 1 month and 3 months after transplantation of L1+ (n = 6) and wild-type (n = 6) SENAs into the striatum and 1 month after transplantation into the substantia nigra are shown (mean ± SEM). (B) Confocal images of L1+ and wild-type SENAs (green) immunostained with an antibody against tyrosin hydroxylase (red) 1 month after transplantation. Scale bar = 50 µm. In the upper right, a Z-stack of 15 images of 1 µm thickness of the area outlined by a square in the merged image of the L1+ graft is shown with orthogonal views of the xz- and yz-planes showing EGFP+/tyrosine hydroxylase+ cells. Scale bar = 50 µm. Percentages of tyrosine hydroxylase+ cells of all EGFP+ cells 1 month and 3 months after transplantation of L1+ (n = 6) and wild-type (n = 6) SENAs into the striatum and 1 month after transplantation after transplantation into the substantia nigra are shown (mean ± SEM). Student's t-test was performed for statistical analysis (*P < 0.05).

Thus, L1 overexpressing SENAs favour neuronal versus astrocytic differentiation and show a higher percentage of graft-derived dopaminergic neurons and an enhanced density of endogenous tyrosine hydroxylase-positive axons in the striatum close to the graft after transplantation into the MPTP-lesioned mouse brain.

L1 overexpressing SENAs, but not wild-type SENAs, enhance striatal dopamine levels without influencing striatal GABA levels leading to increased apomorphine-induced rotation after transplantation into the striatum of MPTP-lesioned mice

To test the effect of L1 overexpressing SENAs on motor behaviour, we analysed apomorphine-induced rotation in MPTP-lesioned mice with unilateral grafts of L1 overexpressing SENAs, wild-type SENAs or mice sham-injected with PBS. As expected, the animals showed no rotational bias after lesioning 1 day before transplantation, since systemic MPTP application induces symmetric cell death of the dopaminergic neurons in the substantia nigra. Animals transplanted with wild-type SENAs showed a tendency towards an ipsilateral rotation bias in apomorphine-induced rotation behaviour one to four weeks after transplantation when compared to the PBS sham-injected group; a difference that was, however, not statistically significant at any time point. In contrast, transplantation of L1 overexpressing SENAs led to a prominent ipsilateral rotation bias in apomorphine-induced rotation behaviour three to ten weeks after transplantation, when compared to the group transplanted with wild-type SENAs and PBS-injected control animals (Fig. 6A).

Figure 6

Behavioural analysis of apomorphine-induced rotation in MPTP-lesioned mice. Mice with unilateral grafts of L1 overexpressing (L1+) (n = 10) or wild-type (WT) SENAs (n = 10), or sham-injected with PBS (n = 10) were analysed. Relative meander was calculated as turning angle divided by the distance moved (mean ± SEM). Tukey's one-way ANOVA was performed for statistical analysis. (*,**P < 0.05 and 0.01, compared with PBS group, #P < 0.05, compared with wild-type group). (B) Striatal dopamine expression level determined by radioimmunoassay specific for dopamine 1 month after unilateral striatal transplantation of L1+ SENAs (n = 7), wild-type SENAs (n = 5), or sham-injection of PBS (n = 5) into MPTP-lesioned mice. Mean values ± SEM are shown. Tukey's one-way ANOVA was performed for statistical analysis (*P < 0.05). (C) Striatal expression level of GABA as determined by enzyme-linked immunosorbent assay specific for GABA 1 month after unilateral striatal transplantation of L1+ SENAs (n = 7), wild-type SENAs (n = 5), or sham-injection of PBS (n = 5) into MPTP-lesioned mice. Mean values ± SEM are shown. Tukey's one-way ANOVA was performed for statistical analysis.

As apomorphine-induced rotations correlate with the underlying degree of loss of nigrostriatal projections and dopaminergic depletion (Hudson et al., 1993), we analysed dopamine and GABA levels in the striatum 1 month after unilateral transplantation of SENAs into the striatum (Fig. 6B and C). L1 overexpressing but not wild-type SENAs enhanced the level of dopamine in the striatum ipsilateral to the grafted side, when compared to the contralateral side (Fig. 6B), but did not influence striatal GABA levels (Fig. 6C). Thus, L1 overexpressing SENAs, but not wild-type SENAs, influence rotation behaviour in MPTP-lesioned mice by selectively enhancing striatal dopamine levels ipsilateral to the grafted side.

L1 overexpressing SENAs rescue imperilled host dopaminergic neurons after transplantation adjacent to the substantia nigra, but not after transplantation into the striatum of MPTP-lesioned mice

To monitor the effects of L1 overexpressing SENAs on the survival of endogenous, host dopaminergic neurons in the substantia nigra, the percentage of tyrosine hydroxylase-positive cells in the substantia nigra was determined ipsilateral and contralateral to the grafted side 1 month after transplantation of L1 overexpressing or wild-type SENAs into the striatum or adjacent to the substantia nigra, or after sham-injection of PBS (Fig. 7). While no differences in the percentage of tyrosine hydroxylase-positive cells were observed between the ipsilateral and contralateral sides after transplantation of SENAs into the striatum, L1 overexpressing but not wild-type SENAs enhanced the numbers of endogenous tyrosine hydroxylase-positive cells in the substantia nigra ipsilateral to the grafted side, when compared to the contralateral side.

Figure 7

Transplantation of L1 overexpressing SENAs into the substantia nigra but not into the striatum rescues host dopaminergic neurons after MPTP lesioning. Immunohistochemical analysis of endogenous tyrosine hydroxylase-positive (TH+) neurons (red) 1 month after transplantation of L1 overexpressing (L1+, n = 7) and wild-type (WT, n = 7) SENA grafts or sham-injection of PBS (n = 5) into the substantia nigra or striatum of MPTP-lesioned mice. Scale bar = 100 µm. Absolute numbers of endogenous tyrosine hydroxylase+ neurons ipsilateral and contralateral to the grafted side are shown (mean ± SEM). Note the increased number of endogenous tyrosine hydroxylase+ neurons ipsilateral to the grafted side in mice transplanted with L1 overexpressing SENAs. Student's t-test was performed for statistical analysis (***P < 0.001).

Discussion

We have shown in this study that overexpression of the neural cell adhesion molecule L1 by grafted SENAs improved functional recovery after syngeneic intrastriatal transplantation into the MPTP-lesioned mouse brain. L1 beneficially influenced survival and neuronal differentiation of grafted cells and the percentage of graft-derived tyrosine hydroxylase-positive neurons after transplantation into the striatum or adjacent to the substantia nigra. Furthermore, intrastriatal transplantation of L1 overexpressing SENAs increased the density of host tyrosine hydroxylase-positive fibres close to the graft but did not enhance the survival of endogenous nigral tyrosine hydroxylase-positive neurons. In contrast, transplantation of L1 overexpressing SENAs adjacent to the substantia nigra rescued imperilled endogenous nigral dopaminergic neurons when compared to wild-type SENAs.

Previous studies had shown that L1 increased neuronal differentiation of embryonic and neural stem cells in vitro (Dihne et al., 2003; Bernreuther et al., 2006). Furthermore, neural pre-differentiation of embryonic stem cells into SENAs enhanced neuronal differentiation of embryonic stem cells in vitro (Dihne et al., 2006). Not surprisingly, L1 did not affect overall neuronal differentiation of SENAs in vitro in our study as wild-type SENAs already consisted of 90% β-tubulin III-positive neurons. In contrast, we observed a tendency towards an increased percentage of tyrosine hydroxylase-positive cells in L1 overexpressing SENAs in vitro indicating a potential effect of L1 on dopaminergic differentiation. Previous studies showed that the addition of sonic hedgehog, fibroblast growth factor-8, and ascorbic acid or the overexpression of the transcription factor Nurr1 leads to increased percentages of tyrosine hydroxylase-positive neurons derived from murine embryonic stem cells in vitro ranging from 30% to 90% (Kawasaki et al., 2000; Lee et al., 2000; Kim et al., 2002, 2006; Nishimura et al., 2003). Thus, the application of these protocols or the combined overexpression of L1 and Nurr1 might further enhance dopaminergic differentiation in vitro and in vivo.

In our study, L1 overexpressing and wild-type SENAs were first transplanted into the striatum of MPTP-lesioned mice to investigate their effects on locomotor recovery. In a second experiment, L1 overexpressing and wild-type SENAs were transplanted adjacent to the substantia nigra of MPTP-lesioned mice to investigate further the effects of L1 overexpressing SENAs on host tyrosine hydroxylase-positive neurons in the substantia nigra without analysing locomotor recovery. Interestingly, the influence of L1 on survival, migration and differentiation of grafted cells was independent of the site of transplantation suggesting that L1 overexpessing SENAs provide a microenvironment promoting cell-autonomous functions of L1 in grafted cells irrespective of the most likely distinct environmental cues in the two different brain regions. In contrast, significant differences were detected in the effect of L1 overexpressing SENAs on endogenous nigral tyrosine hydroxylase-positive neurons depending on the site of transplantation. When transplanted close to the substantia nigra, L1 overexpressing SENAs rescued imperilled tyrosine hydroxylase-positive neurons when compared to wild-type SENAs, while this was not seen after transplantation into the striatum where no differences were observed in survival of these neurons after transplantation.

L1 overexpressing SENA grafts showed increased numbers of cells four and twelve weeks after transplantation when compared to wild-type SENAs both after transplantation into the striatum and the substantia nigra. This is an important finding in view of the fact that several studies describe poor survival of human neural and embryonic stem cell-derived tyrosine hydroxylase-positive neurons after transplantation in rat models of Parkinson's disease when injected into the striatum (Ostenfeld et al., 2000; Schulz et al., 2004; Park et al., 2005; Brederlau et al., 2006; Martinat et al., 2006). Analysis of proliferation, as determined by the percentage of Ki-67-positive grafted cells, did not reveal a difference between the L1 overexpressing and wild-type group while the decreased percentage of caspase-3-positive cells in L1 overexpressing grafts indicated that the increased survival of L1 overexpressing grafted cells was at least to some extent mediated by decreased apoptosis. It is noteworthy in this respect that inhibition of caspases as well as an involvement of phosphoinositide-3 kinase, Src family kinases, mitogen-activated protein kinase kinase and the extracellular signal-regulated kinases 1/2, Akt and Bad, have been shown in the context of improved cell survival mediated by L1 (Loers et al., 2005).

In our study, overexpression of L1 enhanced migration of transplanted cells from the graft edge both after transplantation into the striatum and the substantia nigra. A previous study had shown that L1 increased migration of transplanted embryonic stem cell-derived neural precursor cells in the quinolinic acid-lesion paradigm of Huntington's disease (Bernreuther et al., 2006). Furthermore, neural pre-differentiation of embryonic stem cells into SENAs enhanced migration of transplanted cells in the quinolinic acid-lesion model (Dihne et al., 2006). The reason for this effect may be two-fold: L1 overexpressing cells may interact with L1 in the host environment in a homophilic trans-interaction mechanism as shown by the beneficial effects of ectopic L1 expression by astrocytes in a transgenic mouse, where L1 overexpressing neural stem cells migrated more from the site of injection than wild-type cells in the L1-enriched versus the wild-type environment (Ourednik et al., 2009). However, heterophilic interactions of the L1 overexpressing SENAs with the host environment are also conceivable. In the present study, the extent of migration was not substantial when compared to the size of the striatum and the substantia nigra and thus most likely did not contribute to enhanced functional recovery.

L1 overexpressing SENAs contained a higher proportion of neurons after transplantation into the MPTP-lesioned striatum or substantia nigra than wild-type SENAs, demonstrating that L1 can further improve the previously shown enhanced neuronal differentiation of grafted cells pre-differentiated by the SENA protocol (Dihne et al., 2006). In the present study, overexpression of L1 by SENAs increased the percentage of graft-derived tyrosine hydroxylase-positive neurons both after transplantation into the striatum and the substantia nigra. The higher percentage of graft-derived tyrosine hydroxylase-positive neurons may be attributed either to improved survival of graft-derived tyrosine hydroxylase-positive cells versus other neuronal cell types mediated by increased expression of L1 in grafted cells or to a direct influence of L1 on dopaminergic differentiation as seen in L1 overexpressing neural stem cells (Ourednik et al., 2009). A direct influence of L1 on dopaminergic differentiation is supported by the fact that L1 overexpressing SENAs tended to increase dopaminergic differentiation in vitro, although cells may react differently to similar cues in vitro and in the complex brain tissue. Furthermore, a previous study showed that overexpression of L1 increased the percentage of dopaminergic neurons differentiated from embryonic stem cells in vitro (Bernreuther et al., 2006). In addition, previous findings showed that L1 is a survival factor for dopaminergic neurons in vitro (Hulley et al., 1998). In the present study, L1 overexpressing grafts revealed a decreased percentage of caspase-3-positive cells indicating a potentially beneficial effect of L1 counteracting apoptosis of grafted cells. It is thus conceivable that the enhanced percentage of tyrosine hydroxylase-positive neurons within all grafted neurons in vivo is due either to a direct effect of L1 on dopaminergic differentiation or to enhanced survival of this highly vulnerable cell type or a combination of both mechanisms. GABAergic differentation of grafted SENAs was not affected by L1 overexpression. This is in line with previous results showing that L1 enhanced overall neuronal differentiation of transplanted embryonic stem cell-derived neural precursor cells after transplantation into the quinolinic acid-lesioned mouse striatum without enhancing the percentage of graft-derived GABAergic neurons of all graft-derived neurons (Bernreuther et al., 2006). Interestingly, the percentage of graft-derived tyrosine hydroxylase-positive neurons was negligible after transplantation of embryonic stem cell-derived neural precursor cells as a single cell suspension into the quinolinic acid-lesioned mouse striatum in the previous study (Bernreuther et al., 2006). The enhanced percentage of tyrosine hydroxylase-positive cells in transplanted L1 overexpressing and wild-type SENAs in the present study when compared to the previous study indicates that different microenvironments of specific lesion paradigms critically influence differentiation of transplanted stem cells. Furthermore, single cell suspensions transplanted in the previous study (Bernreuther et al., 2006) may be more prone to host cues when compared to the microarchitecture of SENA grafts used in the present study. Thus, the enhanced percentage of tyrosine hydroxylase-positive cells in transplanted L1 overexpressing and wild-type SENAs could be attributed either to the microenvironment generated by the lesioned host tissue or the local microenvironment generated by transplanted cells.

Moreover, we could show that L1 overexpressing, but not wild-type SENAs, increased the number of imperilled endogenous dopaminergic neurons in the ipsilateral, but not contralateral substantia nigra pars compacta after unilateral transplantation adjacent to the substantia nigra and not after transplantation into the striatum in MPTP-lesioned animals, indicating that L1 overexpressing SENAs reduce the progressive loss of dopaminergic neurons in the substantia nigra, a major feature of Parkinson's disease. This effect appears to be spatially restricted as transplantation of L1 overexpressing SENAs did not influence the survival of endogenous dopaminergic neurons in the contralateral substantia nigra, nor did transplantation into the striatum affect the survival of endogenous dopaminergic neurons. Interestingly however, striatal transplantation of L1 overexpressing SENAs enhanced the density of host tyrosine hydroxylase-positive axons in the vicinity of but not in remote distance to the graft when compared to wild-type SENAs or sham-injected PBS. These findings indicate that L1 exerts beneficial effects on endogenous dopaminergic axons in proximity to the graft potentially contributing to functional recovery. The lack of rescue of endogenous tyrosine hydroxylase-positive neurons in the substantia nigra when SENAs have been engrafted in the striatum (in contrast to transplantation close to the substantia nigra) is most probably due to the longer distance between grafted cells and endogenous tyrosine hydroxylase-positive neurons. While L1 overexpressing SENAs are located in proximity to most of the ipsilateral endogenous tyrosine hydroxylase-positive neurons after nigral transplantation, SENAs grafted into the striatum are located in the proximity of only a minor percentage of tyrosine hydroxylase-positive axons projecting from the substantia nigra due to the small graft size compared to the volume of the striatum. Thus, the fact that L1 exerts beneficial effects on endogenous tyrosine hydroxylase-positive axons close to the graft speaks in favour of a survival promoting effect of L1 in proximity to endogenous dopaminergic neurons, possibly due to direct cell contact or diffusion of a proteolytically cleaved and diffusing functional extracellular domain of L1. Few transplantation studies describe the rescue of endogenous dopaminergic neurons in animal models of Parkinson's disease. Ourednik et al. (2002, 2009) and Yasuhara et al. (2006) observed a neuroprotective effect of grafted neural precursor cells on host cells in the murine MPTP model and the rat 6-hydroxy-dopamine model of Parkinson's disease, respectively. These authors propose secretion of neuroprotective agents or overexpression of L1 by the neural precursor cells as the underlying cause. We did not observe a similar neuroprotective effect after transplantation of embryonic stem cell-derived wild-type SENAs. This could be due to the decreased numbers of neural precursor cells in the SENA grafts that consist almost exclusively of immature neurons and astrocytes with decreased expression of neuroprotective factors when compared to neural precursor cells. Notably, Ebert et al. (2008) described a neuroprotective effect of glial cell line-derived neurotrophic factor overexpressing transplanted neural stem cells on endogenous dopaminergic neurons in a rat model of Parkinson's disease, while wild-type neural stem cells did not show an effect. Other studies described protection of nigral tyrosine hydroxylase-positive cells by astrocytes transduced to overexpress glial cell line-derived neurotrophic factor or fibroblasts expressing brain-derived neurotrophic factor (Lucidi-Phillipi et al., 1995; Ericson et al., 2005). Redmond et al. (2007) showed a neuroprotective effect of human neural stem cells in a primate model of Parkinson's disease. Enhanced rescue of endogenous dopaminergic neurons after transplantation of mesenchymal stem cells was attributed to either the expression of trophic or immunomodulatory factors (Keshet et al., 2007; Park et al., 2008). In the present study, we could show that a different type of neuroprotective mechanism as exerted by a cell adhesion molecule can rescue imperilled endogenous dopaminergic neurons in an animal model of Parkinson's disease.

We could also show in the present study that L1 overexpressing SENAs, in comparison to the wild-type SENAs, decreased the number of microglial cells in the host tissue. The reason why the number of microglial cells was decreased in the vicinity of L1 overexpressing versus wild-type grafts is currently not understood but encourages further investigation of the role of L1 on immune system cells in the brain as the immune response of the brain to injury and grafted tissue is an important issue in neurodegenerative diseases and in stem cell biology (Barker and Widner, 2004; Chen and Palmer, 2008; Ideguchi et al., 2008). Although this effect was not pronounced, it may contribute to the enhanced number of surviving cells in L1 overexpressing SENAs although previous studies showed that microglial cells have the potential to both increase and decrease neuronal survival (Bessis et al., 2007).

L1 overexpressing SENAs, but not wild-type SENAs or sham-injected PBS, led to a prominent ipsilateral bias in apomorphine-induced rotation behaviour after intrastriatal SENA transplantation into the MPTP-lesioned brain indicating improved motor behaviour. Most stem cell transplantation studies used unilateral lesioning of the rat striatum applying 6-hydroxy-dopamine or bilateral lesioning via MPTP. In these experiments, decrease of apomorphine- or amphetamine-induced rotation behaviour was used to measure functional recovery (Kim et al., 2002; Dezawa et al., 2004; Park et al., 2005; Sanchez-Pernaute et al., 2005; Fu et al., 2006). In our study, we syngeneically transplanted stem cells into mice that had received intraperitoneal injections of MPTP and thus showed a bilateral loss of dopaminergic neurons mimicking disease progression in Parkinson's disease. Consequently, no rotation bias was observed in apomorphine-induced rotation before transplantation. However, unilateral transplantation of L1 overexpressing SENAs led to an ipsilateral rotation bias after stimulation with apomorphine that was not observed in the wild-type SENA and PBS groups. We interpret this as a sign of functional recovery in view of previous studies describing an apomorphin-induced rotation to the contralateral (unlesioned) side after unilateral lesioning with 6-hydroxydopamine (Hudson et al., 1993; Da Cunha et al., 2008) although we are aware that the interpretation of a singular behavioural analysis is limited and other behavioural experiments such as paw reaching test, stepping test, rotarod test, pole test and assessment of tremor as well as sensory deficits (Sedelis et al., 2001) could be used to substantiate these findings. This needs to be taken into consideration, since a limited correlation between improvement in rotation behaviour and improvement in motor disability has been suggested (Lundblad et al., 2002; Metz and Whishaw, 2002). The effect in the present study is most probably due to the increased percentage of tyrosine hydroxylase-positive cells in L1 overexpressing versus wild-type SENAs and potentially to the enhanced density of endogenous tyrosine hydroxylase-positive axons; and most likely not due to unspecific effects mediated by the increased graft size in L1 overexpressing SENAs, as striatal dopamine levels ipsilateral to the graft were enhanced in mice transplanted with L1 overexpressing SENAs, but not wild-type SENAs, while striatal GABA levels were unchanged. Thus, L1 overexpressing SENAs enhanced functional recovery after transplantation into the MPTP-lesioned striatum.

In summary, the combination of the SENA differentiation protocol leading to enhanced neuronal differentiation, migration and decreased tumour formation of embryonic stem cells, with overexpression of the neural cell adhesion molecule L1 leading to enhanced dopaminergic differentiation and migration results in increased survival of grafted cells after intrastriatal transplantation in the MPTP-lesioned mouse brain with an increased percentage of graft-derived tyrosine hydroxylase-positive neurons in the grafted population of neurons, and enhanced the density of endogenous tyrosine hydroxylase-positive fibres close to the graft. This therapeutical intervention leads to elevated striatal dopamine levels that promote functional recovery. This is an important finding in view of the fact that human embryonic stem cell-derived tyrosine hydroxylase-positive neurons show poor survival after transplantation in rodent models of Parkinson's disease. We also show that nigral transplantation of stem cells overexpressing a neural cell adhesion molecule enhances the survival of imperilled endogenous dopaminergic neurons in the substantia nigra, pointing to the potential of L1 overexpression in stem cells for the therapy of Parkinson's disease.

Funding

This work was supported by the New Jersey Commission for Spinal Cord Research and the University Hospital Hamburg-Eppendorf.

Supplementary material

Supplementary material is available at Brain online.

Footnotes

  • Abbreviations:
    Abbreviations
    EGFP
    enhanced green fluorescent protein
    GABA
    gamma-aminobutyric acid
    GFAP
    glial fibrillary acidic protein
    MPTP
    1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine
    PBS
    phosphate-buffered saline
    SENAs
    substrate-adherent embryonic stem cell-derived neural aggregates

References

View Abstract