Journal of Neural Transmission

, Volume 117, Issue 5, pp 559–572

Effects of GDF5 overexpression on embryonic rat dopaminergic neurones in vitro and in vivo

Authors

  • David B. O’Sullivan
    • Department of Neuroscience/Anatomy, Biosciences Research InstituteUniversity College Cork
  • Patrick T. Harrison
    • Department of Physiology, Biosciences Research InstituteUniversity College Cork
    • Department of Neuroscience/Anatomy, Biosciences Research InstituteUniversity College Cork
Basic Neurosciences, Genetics and Immunology - Original Articles

DOI: 10.1007/s00702-010-0392-9

Cite this article as:
O’Sullivan, D.B., Harrison, P.T. & Sullivan, A.M. J Neural Transm (2010) 117: 559. doi:10.1007/s00702-010-0392-9

Abstract

Transplantation of embryonic dopaminergic neurones has shown promise for the treatment of Parkinson’s disease (PD), but this approach is limited by the poor survival of the transplanted cells. Exogenous dopaminergic neurotrophic factors such as growth/differentiation factor 5 (GDF5) have been found to enhance the survival of transplanted dopaminergic neurones. However, this approach is limited by the rapid degradation of such factors in vivo; thus, methods for long-term delivery of these factors are under investigation. The present study shows, using optimised lipid-mediated transfection procedures, that overexpression of GDF5 significantly improves the survival of dopaminergic neurones in cultures of embryonic day (E) 13 rat ventral mesencephalon (VM) and protects them against 6-hydroxydopamine (6-OHDA)-induced toxicity. In another experiment, E13 VM cells were transfected with GDF5 after 1 day in vitro (DIV), then transplanted into 6-OHDA-lesioned adult rat striata after 2 DIV. The survival of these E13 VM dopaminergic neurones after transfection and transplantation was as least as high as that of freshly dissected E14 VM dopaminergic neurones, demonstrating that transfection was not detrimental to these cells. Furthermore, GDF5-overexpressing E13 VM transplants significantly reduced amphetamine-induced rotational asymmetry in the lesioned rats. This study shows that lipid-mediated transfection in vitro prior to transplantation is a valid approach for the introduction of neurotrophic proteins such as GDF5, as well as lending further support to the potential use of GDF5 in neuroprotective therapy for PD.

Keywords

Growth/differentiation factor 5Parkinson’s diseaseTransplantationNeurotrophic factorDopaminergic neuroneRat

Abbreviations

6-OHDA

6-Hydroxydopamine

DIV

Day in vitro

E

Embryonic day

GDF5

Growth/differentiation factor 5

GDNF

Glial cell line-derived neurotrophic factor

GFAP

Glial fibrillary acidic protein

PD

Parkinson’s disease

PI

Propidium iodide

TGF-β

Transforming growth factor-β

TH

Tyrosine hydroxylase

VM

Ventral mesencephalon

Introduction

Parkinson’s disease (PD) is characterised by the progressive degeneration of midbrain dopaminergic neurones, resulting in a loss of dopamine neurotransmission in the striatum. The cause of this neurodegeneration is unknown. Current pharmacological treatments focus on restoring striatal dopamine levels, but they do not affect the underlying neurodegeneration and thus the pathological process continues unabated (for review, see Toulouse and Sullivan 2008). Over the past number of years, two promising therapies that address the neuronal cell loss associated with PD have been developed. The first involves transplantation of embryonic dopaminergic neurones to replace those that have been lost, while the second is aimed at preventing dopaminergic neuronal cell death through the administration of neurotrophic proteins.

Transplantation of dopaminergic neurones from the embryonic midbrain (the ventral mesencephalon, VM) into the striatum of PD patients has been reported to restore dopaminergic function and provide long-lasting relief of symptoms (for reviews, see Lindvall and Hagell 2000; Dunnett et al. 2001; Hagell and Brundin 2001; Bjorklund et al. 2003; Sayles et al. 2004). However, this field of study received a considerable setback when two NIH-funded, double-blind placebo-controlled trials failed to find significant improvements in transplanted patients compared to those who had received sham-surgery (Freed et al. 2001; Olanow et al. 2003). The value of trials involving sham-surgery has been questioned and the scientific community appears to be split on this issue. Nevertheless, it is clear that there is a need for optimisation of the methods used to harvest, prepare and deliver foetal dopaminergic cells for transplantation, as well as for standardisation of these procedures between centres. Three recent reports have documented the presence of Lewy-body-like cytoplasmic inclusions in grafted dopaminergic neurones at post-mortem examination in patients who had received transplants 11–16 years before death (Kordower et al. 2008a, b; Li et al. 2008). There has been much debate on the future of foetal neuronal cell transplantation for PD patients, in light of these neuropathological findings. Other problems associated with the use of cell transplants include variability in clinical outcome and the occurrence of graft-induced dyskinesias (for reviews, see Hagell and Cenci 2005; Piccini et al. 2005).

One major problem with the use of primary neuronal transplants is that the survival of the cells in the host brain is very poor; it is estimated that as few as 5–10% of the dopaminergic neurones survive the grafting procedure (for reviews, see Brundin and Hagell 2001; Bjorklund 2005). Several agents, including neurotrophic factors, are being evaluated for their abilities to improve dopaminergic neuronal survival after transplantation (for reviews, see Brundin et al. 2000; Sortwell 2003; O’Neill et al. 2007; Deierborg et al. 2008).

The dopaminergic neurotrophic factor which has been most widely studied is glial cell line-derived neurotrophic factor (GDNF), a member of the transforming growth factor-β (TGF-β) superfamily. GDNF has been shown to promote dopaminergic neuronal survival in vitro (Lin et al. 1993; Choi-Lundberg and Bohn 1995; Krieglstein et al. 1995a; Hou et al. 1996) and in vivo (Hoffer et al. 1994; Kearns and Gash 1995; Sauer et al. 1995; Tomac et al. 1995; Sullivan et al. 1998a). GDNF was also shown to improve the survival and integration of grafted dopaminergic neurones in animal models of PD (Apostolides et al. 1993; Rosenblad et al. 1996; Sinclair et al. 1996; Wang et al. 1996; Sullivan et al. 1998b). As a result of these studies, GDNF was evaluated as a therapeutic agent for PD. Promising results emerged from two open-label clinical trials involving intraputamenal application of GDNF in PD, with improvements recorded in the patients’ motor symptoms, without any serious side-effects (Gill et al. 2003; Patel et al. 2005; Slevin et al. 2005). However, a recent double-blind trial reported no significant motor improvements in PD patients (Nutt et al. 2003; Lang et al. 2006). Furthermore, safety concerns were raised, as some of the patients developed antibodies to GDNF (Tatarewicz et al. 2007). It is clear that GDNF therapy for PD requires further development and it is probable that alternative methodologies may prove to be more effective for achieving long-term and targeted GDNF delivery (see Sherer et al. 2006; Morrison et al. 2007). Various delivery methods are being assessed in animal models of PD, such as grafting of genetically modified cells expressing GDNF (Zurn et al. 2001; Ericson et al. 2005) and viral-mediated GDNF delivery (for reviews, see Bjorklund et al. 2000; Kordower 2003; Deierborg et al. 2008).

Growth/differentiation factor 5 (GDF5) is another member of the TGF-β superfamily, which has comparable actions to GDNF on dopaminergic neurones in vitro (Krieglstein et al. 1995b; O’Keeffe et al. 2004a; Wood et al. 2005; Clayton and Sullivan 2007) and in in vivo rat models of PD (Sullivan et al. 1997, 1999; Hurley et al. 2004). GDF5 has also been reported to improve the survival and function of grafted dopaminergic neurones to the same extent as GDNF (Sullivan et al. 1998b). Each of these in vivo studies have involved infusion of recombinant GDF5 protein which can only be effective for a limited time due to its being metabolised in the brain. A better strategy may be the implantation of dopaminergic neurones which have been genetically modified to overexpress GDF5. This would provide dopaminergic neurones to replace those that degenerate in the disease, as well as a source of GDF5 to improve the survival of the implanted neurones and potentially to protect the remaining endogenous dopaminergic innervation of the striatum.

With the development of enhanced cationic lipid-based gene-transfer methods (for review, see Washbourne and McAllister 2002), the prospect of manipulating embryonic rat VM dopaminergic neurones to permit high-level expression of genes encoding neurotrophic factors may be achievable. Cationic lipids have proven to be one of the most efficient and versatile mediators of non-viral DNA transfer into cultured cells. The present study investigated the effects of overexpression of GDF5, using cationic lipid-mediated gene delivery, on the survival of embryonic rat dopaminergic neurones in vitro. The survival and functional effects of these GDF5-overexpressing neurones were then examined after their transplantation into an adult rat model of PD. Initially, the phenotypic composition of E13 rat VM cultures after 2 day in vitro (DIV) was compared to that of cultures of E14 rat VM (the standard tissue used for transplants) after 1 DIV. This period of 2 days in culture was necessary to allow time for transfection and subsequent removal of the cells for transplantation. Since there were no significant differences between the phenotypes or mitotic activity of cells in these two types of cultures, it was feasible to use E13 VM cultures for the subsequent transplantation study. Thus, after optimisation of the transfection protocols, GDF5-overexpressing E13 VM cells were transplanted into 6-hydroxydopamine (6-OHDA)-lesioned rat striata, a commonly used model of PD. The survival and functional effects of these transplants were compared with those of freshly dissected E14 VM transplants, using behavioural testing and post-mortem immunohistochemistry.

Materials and methods

Primary culture of E13 or E14 rat VM

Date-mated Sprague–Dawley rats (Biological Services Unit, University College Cork) at E13 or E14 were anaesthetised with halothane and killed by decapitation. The crown-rump lengths were 7.5–8.5 mm for E13 and 9.5–10.5 mm for E14 rats (see Torres et al. 2008). Embryos were transferred to ice-cold Hank’s Balanced Salt Solution (HBSS; Sigma). VMs were removed and incubated in 1 ml of 0.1% trypsin–EDTA solution (Sigma) for 5 min at 37°C. 100 μl of 0.5 mg/ml soybean trypsin inhibitor (Sigma) was added and the suspension was centrifuged at 1,000×g for 3 min and the supernatant was discarded. The cells were resuspended in 1 ml “growth medium” [DMEM/F12 (Sigma), 1% foetal calf serum (Sigma), 1% l-glutamine (Gibco), 33 mM d-glucose (Sigma), 2% B27 (Gibco), 1% Penicillin/Streptomycin (Gibco)] and seeded at 5 × 105 cells/cm2 on poly-d-lysine (Sigma)-coated 6- or 24-well plates and incubated at 37°C in 5% CO2. Medium was replaced after 1 DIV, and every 2 DIV thereafter.

Plasmids

Human GDF5 cDNA was expressed under the control of a β-globin promoter in the plasmid pSGDF5 (a gift from Biopharm GmbH) and green fluorescent protein (GFP) was expressed using the pEGFP-N1 vector (Clontech). pcDNA3.1 (Invitrogen) was used for mock transfection in all experiments. Plasmids were propagated in E. coli DH5α (Invitrogen) and purified using HiSpeed Plasmid Midi Kit (Qiagen).

Transfection in E13 or E14 rat VM cultures

E13 or E14 rat VM cultures were transfected after 1, 2 or 3 DIV. Either pEGFP-N1, pSGDF5 (“pSGDF5-transfected” cultures) or pcDNA3.1 (“mock-transfected” cultures) (1–4 μg in 300 μl DMEM–F12 or Opti-MEM®) was mixed with Lipofectamine™ or Lipofectamine 2000™ (1–4 μl in 300 μl DMEM–F12 or Opti-MEM®), vortexed, incubated for 5 min at room temperature, then added dropwise to the cultures. After 5 h at 37°C, medium was replaced with DMEM–F12, supplemented with 10% foetal calf serum and incubated at 37°C for 24 h. Transfections were performed in triplicate.

Treatment of pSGDF5-transfected E13 rat VM cultures with 6-OHDA

At 1 DIV after transfection, E13 VM cultures were exposed to 50 μM 6-OHDA hydrobromide (Sigma) in growth medium for 30 min. Medium was replaced with fresh growth medium. After 3 DIV, cultures were stained with propidium iodide (PI), to assess cell viability. Cultures were incubated in 5 μg/ml PI (Sigma) in phosphate-buffered saline (PBS) for 5 min at RT, washed with PBS, fixed in 4% paraformaldehyde, washed (3× 5 min) in PBS and visualised. Separate cultures were immunocytochemically stained for tyrosine hydroxylase (TH) after 3 DIV, as described below.

Enzymatic removal of transfected E13 rat VM cultures

At 1 DIV after transfection, medium was removed and cultures were incubated in 2 ml Accutase solution (Chemicon) for 10 min at 37°C. Accutase solution was removed from the culture dishes by flushing with “transfection/transplantation medium” (DMEM–F12 with 1% l-glutamine, 30 mM d-glucose). The cell suspension was centrifuged at 1,000×g for 3 min, the supernatant was discarded and the pellet was resuspended in 100 μl transfection/transplantation medium. Cells were counted using a haematocytometer and maintained at 37°C until used for transplantation.

Immunocytochemistry on E13 or E14 rat VM cultures

For β-III-tubulin immunocytochemistry, cultures were incubated in 100% ice-cold methanol (Sigma) for 10 min at RT. For all other immunocytochemistries, cultures were fixed in 4% paraformaldehyde (Sigma) for 20 min at RT. Cultures were washed (3× 5 min) in 10 mM PBS containing 0.02% Triton X-100 (Sigma) then incubated in “blocking solution” (5% normal horse serum, 0.2% Triton X-100 in PBS) overnight at 4°C. Primary antibodies for β-III-tubulin (Promega), TH (Chemicon), GDF5 (aMP5; Biopharm GmbH), nestin (Chemicon) or glial fibrillary acidic protein (GFAP; Dako) were added (1:300 for all except aMP5, which was 1:50, in blocking solution), then cultures were incubated overnight at 4°C. Cultures were washed (3× 5 min) in PBS then incubated with secondary antibody (FITC- or TRITC-labelled anti-mouse or anti-rabbit IgG, 1:50 in PBS, as appropriate; Sigma) for 1 h at RT. Nuclear labelling was carried out by counterstaining for 4 min with DAPI (10 μg/ml in PBS). Fluorescence was visualised using an inverted microscope (Olympus IX 70). Each cell culture experiment was repeated three times. Immunopositive cells were counted in at least six fields in each of three wells for each of the three cultures. For estimation of TH-positive cell somal area, 50 neurones per treatment group from three culture experiments were analysed using established stereological methods (for review, see Mayhew 1992). The formula used to calculate cell somal area was nB, where n = the number of points overlying the cell body and B = the area associated with each point (taking the magnification into account).

Bromo-2-deoxyuridine (BrdU) labelling of E13 or E14 rat VM cultures

After 1 or 2 DIV, cultures were incubated in 10 μM BrdU (Sigma) for 5 h at 37°C, washed in PBS, fixed in ice-cold methanol, washed in PBS with 0.02% Triton X-100 and denaturated using 2 N HCl for 15 min. Cultures were washed in PBS and incubated for 3 h at RT in FITC-conjugated anti-BrdU goat polyclonal antibody (1:50 in PBS; BD Pharmingen), washed in PBS and visualised as above.

Western blotting on E13 rat VM cultures

After 3, 7 or 9 DIV, medium was removed from cultures and concentrated using Amicon ultra centrifuge tubes with a 5 kDa cut-off. Cells were removed from the culture wells and resuspended in ice-cold RIPA buffer (50 mM Tris pH 7.4, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM EDTA, 0.32 M sucrose, 100 μg/ml PMSF, 1 μg/ml bacitracin, 1 μg/ml pepstatin A, 1 μg/ml aprotinin, 1 μg/ml leupeptin, 1 μg/ml antipain), incubated on ice for 1 h and centrifuged at 10,000×g for 10 min at RT. Aliquots (30 μg) of cell or medium samples were diluted 1:1 (v:v) in non-denaturing (DTT-free) “sample buffer” [7 M urea, 0.1% sodium phosphate, 1% sodium dodecyl sulphate (SDS), 1% bromophenol blue], separated on a 12% SDS–polyacrylamide gels, transferred to PVDF membranes and incubated in blocking solution (5% BSA in PBS with 1% TWEEN-20; PBS-T) for 1 h at RT. Membranes were probed with mouse monoclonal antibody to GDF5 (aMP5; 1:1,000; Biopharm GmbH) in blocking solution overnight at 4°C, washed (6× 5 min) in PBS-T, incubated in secondary antibody solution (anti-mouse IgG conjugated to HRP; 1:10,000; Dako) in blocking solution for 1 h at RT and visualised using the ECL-Plus Western blot detection system (Amersham Biosciences). Membranes were stripped (Western blot recycling kit; Chemicon), incubated with goat polyclonal antibody to β-actin (1:1,000; Santa-Cruz) in blocking solution for 1 h at RT, washed, incubated for 1 h with anti-goat IgG conjugated to HRP (1:10,000; Dako) in blocking solution and visualised using the ECL-Plus Western blot detection system to confirm equal loading of gels. For concentrated medium samples, equal loading was confirmed by comparison of Coomassie blue-stained gels.

6-OHDA lesion in adult rats

Twenty adult male Sprague–Dawley rats (Biological Services Unit, UCC), weighing 220–270 g at the time of surgery, were used. These animals were maintained under veterinary supervision on a 12 h light:12 h dark cycle with food and water available ad libitum. All procedures were performed under licences issued by the Irish Government Department of Health and Children and the study was approved by the Animal Studies Ethics Committee, UCC. Sixteen rats received a unilateral 6-OHDA lesion of the striatum. Rats were anaesthetised using a mixture of ketamine (5 mg/kg, i.p.) and xylazine (5 mg/kg, i.p.) and placed in a Kopf® stereotaxic frame. Coordinates were taken from bregma and dura. A 10-μl microsyringe (Hamilton) fitted with a removable needle (diameter 0.13 mm) was used to inject 6-OHDA hydrobromide (20 μg as free base in 3 μl 0.9% saline with 0.1% ascorbic acid, Sigma) into the left striatum, at the following coordinates: AP +1.0, ML +3.0, DV −5.0, incisor bar at 0.0 mm. The solution was infused at a rate of 1 μl/min and the needle was left in place for a further 1 min before retraction. At 2 weeks after lesion surgery, 12 of the rats underwent further stereotaxic surgery for transplantation of either (i) E13 rat VM cells transfected with pSGDF5, (ii) E13 rat VM cells transfected with pcDNA3.1 or (iii) freshly dissected E14 rat VM cells (n = 4 per group). In each case, 2.5 × 105 cells in 7.5 μl transfection/transplantation medium were injected using a 50-μl Hamilton microsyringe into two sites in the left striatum at the following coordinates: AP +1.0, ML +3.0, DV −6.0 and AP +1.0, ML +3.0, DV −5.0; incisor bar at 0.0 mm.

Behavioural analysis in adult rats

At 1 week after lesion surgery and at 2 weeks after transplantation surgery, all animals underwent testing of motor behaviour. Ipsilateral rotations were recorded in unrestrained animals over a 1 h period beginning 5 min after (+)-amphetamine sulphate (Sigma) administration (5 mg/kg, i.p.).

Post-mortem immunohistochemistry in adult rats

At 2 weeks after the last behavioural testing session (i.e. 6 weeks after lesion surgery), rats were killed by terminal anaesthesia (150 mg/kg sodium pentobarbitone, i.p.), perfused intracardially with 100 ml of PBS, then with 200 ml of freshly prepared 4% ice-cold paraformaldehyde in PBS. Brains were incubated in 4% paraformaldehyde in PBS overnight, cryoprotected in 30% sucrose in PBS, and snap-frozen in isopentane on liquid nitrogen. Three pairs of coronal cryosections (10 μm) were collected (Leica cryostat) at each of six levels though the striatum (+1.6, +1.0, −0.2, −0.3, −0.9, −1.4, −2.1 mm relative to bregma; Paxinos and Watson 1988). Three pairs of cryosections were also collected at each of three levels through the midbrain (−4.8, −5.6, −6.4 mm relative to bregma; Paxinos and Watson 1988). All sections were incubated in “blocking solution” (20% normal horse serum, 0.2% Triton X-100 in PBS) at 4°C overnight, then in rabbit polyclonal antibody to TH (Chemicon) solution (1:300 in 5% normal horse serum, 0.2% Triton X-100 in PBS) at 4°C overnight. Sections were washed in Triton X-100 in PBS and incubated in FITC-labelled anti-rabbit IgG (Sigma) solution (1:50 in PBS) at RT for 3 h. Sections were washed, mounted and cover-slipped (Sigma), then visualised using an upright fluorescence microscope (Olympus AX70) or an inverted confocal fluorescence microscope (Olympus IX81). Graft volume was estimated using the Cavalieri method. The number of dopaminergic neurones per graft was estimated using the Physical Dissector method (Mayhew 1992; Howard and Reed 1998). For the estimation of TH-positive cell somal area and neurite length, 50 neurones per treatment group were analysed using established stereological methods (for review, see Mayhew 1992). The formula used to calculate cell somal area was nB, where n = the number of points overlying the cell body and B = the area associated with each point (taking the magnification into account). The formula used to calculate neurite length was nTπ/2, where n = the number of times the neurites intersected the grid lines and T = the distance between the grid lines (taking the magnification into account).

Statistical analysis

Statistical analysis was performed on all data using either Students’s t test or ANOVA with post hoc Tukey’s test. Differences were considered to be significant at P < 0.05. All data are presented as mean ± SEM.

Results

Characterisation of cell phenotypes in E13 and E14 rat VM cultures

The transplantation study described below required VM cells to be cultured for a day before transfection, then cultured for a second day before removal for transplantation. Since the standard tissue used for transplantation studies in PD models is E14 VM, the current study used E13 VM, to allow for cell development during the necessary time spent in culture. Thus, the initial part of this study involved the comparison of the cellular compositions of E13 and E14 rat VM cultures.

The proportions of neurones, dopaminergic neurones, precursor cells and dividing cells in E13 and E14 rat VM cultures were characterised using immunocytochemistry to detect β-III-tubulin, TH, nestin and BrdU, respectively (Fig. 1a, b). E13 cultures after 1 DIV contained significantly lower (P < 0.001) numbers of neurones and of dopaminergic neurones, and significantly higher (P < 0.001) numbers of nestin-positive and of BrdU-positive cells than did E14 cultures after 1 DIV. In contrast, E13 cultures after 2 DIV were composed of cellular proportions that were not significantly different to those of E14 cultures after 1 DIV (Fig. 1a, b). These findings validate the use of E13 VM tissue in the subsequent experiments, which required 2 days in culture to allow time for transfection and removal from culture, prior to transplantation.
https://static-content.springer.com/image/art%3A10.1007%2Fs00702-010-0392-9/MediaObjects/702_2010_392_Fig1_HTML.gif
Fig. 1

a Percentages of cells that were immunopositive for β-III-tubulin, TH, nestin or BrdU in E13 or E14 rat VM cultures after 1 DIV, or in E13 rat VM cultures after 2 DIV. Data are expressed as mean ± SEM. ***P < 0.001 versus E14 1 DIV; ###P < 0.001 versus E13 2 DIV. b Representative photomicrographs showing E13 or E14 rat VM cultures after 1 DIV, or E13 rat VM cultures after 2 DIV, immunocytochemically stained for β-III-tubulin, TH, nestin or BrdU, as indicated (green). Cultures were counterstained with DAPI (blue)

Optimisation of transfection in E13 and E14 rat VM cultures

Transfection of E14 VM cultures after 1 DIV with pEGFP-N1 (2 μg) using Lipofectamine™ (1–4 μl) resulted in relatively low percentages of cells expressing GFP at 24 h post-transfection, ranging from 4.56 ± 0.41 (1 μl Lipofectamine™) to 6.50 ± 0.31 (4 μl Lipofectamine™). Use of Lipofectamine 2000™ resulted in a ~3-fold increase in the percentages of cells expressing GFP, ranging from 10.11 ± 1.25 (1 μl Lipofectamine 2000™) to 20.23 ± 1.19 (4 μl Lipofectamine 2000™). Consistently high transfection efficiencies (15–21%) were observed using 2 μg DNA and 4 μl Lipofectamine 2000™ in both E13 and E14 cultures at either 1, 2 or 3 DIV. Thus, all subsequent transfections were performed in this manner.

To determine if a particular subset of neurones were preferentially transfected, cells were immunostained for either β-III-tubulin or TH (Fig. 2b, c). The transfection efficiencies of β-III-tubulin-positive neurones (16.36 ± 1.03%) and TH-positive neurones (21.05 ± 1.68%) in E13 cultures at 2 DIV post-transfection were within the range observed in the above studies (15–21%).
https://static-content.springer.com/image/art%3A10.1007%2Fs00702-010-0392-9/MediaObjects/702_2010_392_Fig2_HTML.gif
Fig. 2

ac Representative photomicrographs showing EGFP expression after 2 DIV in E13 rat VM cultures transfected with 2 μg pEGFP-N1 using Lipofectamine 2000™ in Opti-MEM medium, immunocytochemically stained for b β-III-tubulin (red) or c TH (red) and counterstained with DAPI (blue). Representative immunoblots showing GDF5 expression in d cell lysates or e medium fractions from E13 rat VM cultures which had been pSGDF5-transfected at 1 DIV and analysed at 3, 7 and 9 DIV. The bands at ~55 and ~25 kDa correspond to the predicted molecular weights of GDF5 monomeric precursor and GDF5 active dimer, respectively. M mock-transfected, pS pSGDF5-transfected. Representative photomicrographs of f pSGDF5-transfected or g mock-transfected E13 rat VM cultures after 3 DIV, immunocytochemically stained for GDF5 (green). Cultures were counterstained with DAPI (blue)

Together, these findings define a robust and reliable method to achieve transfection efficiencies of ~20%, which are largely unaffected by the age of culture (E13 or 14), growth period in vitro (1–3 days) or cell type.

GDF5 expression and secretion in transfected E13 rat VM cultures

Having established suitable conditions for transfection using GFP, cells were transfected with the GDF5 expression vector pSGDF5 and the expression of recombinant GDF5 was measured by immunoblotting and immunocytochemistry. At each of the time-points examined (3, 7 and 9 DIV), cell lysates from pSGDF5-transfected E13 VM cultures displayed a GDF5-immunoreactive band at ~55 kDa, which was absent from mock-transfected E13 cultures (Fig. 2d).

The overexpression of GDF5 resulted in increased levels of a ~25 kDa GDF5-immunoreactive band in conditioned medium at each of the three time-points (Fig. 2e). Densitometric analysis indicated that the increase in the level of expression, relative to that of the mock-transfected cells, was 1.3-fold at 3 DIV, 1.5-fold at 7 DIV and 1.2-fold at 9 DIV.

Immunocytochemical analysis further confirmed GDF5 expression throughout the cell bodies of ~15–20% of cells in pSGDF5-transfected E13 VM cultures after 3 DIV, but not in mock-transfected cultures (Fig. 2f, g).

Effects of GDF5 overexpression in E13 rat VM cultures

Having shown GDF5 expression in and secretion from pSGDF5-transfected cells, E13 VM cultures were analysed for possible morphological changes caused by this overexpression of GDF5. E13 VM cultures which had been transfected with pSGDF5 contained a significantly higher number of TH-positive neurones at 3, 5 and 7 DIV than either control or mock-transfected cultures (P < 0.05; Fig. 3a, d). Morphological analysis revealed that TH-positive neurones in pSGDF5-transfected cultures displayed a significantly larger mean somal area than those in either control or mock-transfected cultures at 7 DIV (P < 0.05; Fig. 3c, d). pSGDF5-transfected cultures also contained significantly more GFAP-immunopositive cells than either control or mock-transfected cultures at 5 and 7 DIV (P < 0.001; Fig. 3b, d).
https://static-content.springer.com/image/art%3A10.1007%2Fs00702-010-0392-9/MediaObjects/702_2010_392_Fig3_HTML.gif
Fig. 3

Percentages of cells that were immunopositive for a TH or b GFAP in pSGDF5-transfected E13 rat VM cultures after 3, 5 or 7 DIV, and c somal area of TH-immunopositive cells after 7 DIV. Data are expressed as mean ± SEM. *P < 0.05, ***P < 0.001 versus control cultures; #P < 0.05, ##P < 0.01, ###P < 0.001 versus mock-transfected cultures. d Representative photomicrographs of control, mock- or pSGDF5-transfected E13 rat VM cultures as indicated, immunocytochemically stained for TH (left column) after 5 DIV (green) or GFAP (right column) after 7 DIV (green). Cultures were counterstained with DAPI (blue)

To determine if overexpression of GDF5 affected cell death induced by the dopaminergic neurotoxin 6-OHDA, total cell survival was determined using PI uptake and TH-positive neuronal survival was measured using immunocytochemistry. Treatment with 6-OHDA resulted in significant increases in the number of dead cells and significant decreases in the percentage of TH-positive cells in control, mock-transfected and pSGDF5-transfected cultures after 3 DIV (P < 0.01; Fig. 4a, b). However, both total cell death and TH-positive neuronal loss induced by 6-OHDA were significantly lower in pSGDF5-transfected cultures than in either control or mock-transfected cultures (P < 0.01; Fig. 4a, b). In cultures which had not been treated with 6-OHDA, transfection with pSGDF5 or mock transfection did not either increase the number of dead cells or decrease the percentage of TH-positive neurones compared to those in untreated control cultures (Fig. 4a, b), indicating that transfection per se had no effect on cell viability. pSGDF5 transfection significantly increased the percentage of TH-positive neurones compared to those in both untreated control and mock-transfected cultures, confirming its neurotrophic action on these cells (P < 0.05; Fig. 4b).
https://static-content.springer.com/image/art%3A10.1007%2Fs00702-010-0392-9/MediaObjects/702_2010_392_Fig4_HTML.gif
Fig. 4

a Numbers of cells that were PI-positive and b percentages of cells that were TH-immunopositive in control, mock-transfected or pSGDF5-transfected E13 rat VM cultures at 3 DIV after treatment with 50 μM 6-OHDA or no treatment. Data are expressed as mean ± SEM. ***P < 0.001 versus corresponding untreated cultures; ##P < 0.01 versus 6-OHDA-treated mock-transfected cultures; ++P < 0.01 versus 6-OHDA-treated control cultures; £P < 0.05 versus untreated control cultures; ΦP < 0.05 versus untreated mock-transfected cultures. c Total cell number per field, percentage of cells that displayed apoptotic nuclei and percentage of cells that were TH-immunopositive after 3 DIV in control E13 rat VM cultures or in E13 rat VM cultures which had been mock-transfected at 1 DIV, enzymatically removed from the culture plates and re-plated at 2 DIV. Data are expressed as mean ± SEM. ***P < 0.001 versus control cultures. Representative photomicrographs of d control E13 rat VM cultures or e E13 rat VM cultures which had been mock-transfected at 1 DIV, enzymatically removed from the culture plates and re-plated at 2 DIV. Cultures were immunocytochemically stained for TH (red) and counterstained with DAPI (blue) at 3 DIV

Effects of enzymatic removal on viability of transfected cells in E13 rat VM cultures

Since removal of the cells from the culture dishes is a necessary part of the transplantation procedure, the effect of this on cell viability was determined. This would enable any cell loss due to removal from culture to be accounted for, in the eventual calculation of cell survival post-transplantation. The effect of enzymatic removal followed by re-plating on cell viability was assessed using DAPI staining. E13 VM cultures were mock-transfected after 1 DIV, enzymatically removed after 2 DIV, re-plated, then analysed after 3 DIV. There was no significant difference between the number of cells per field in cultures which had been transfected and re-plated and that in control cultures (Fig. 4c–e). The number of DAPI-stained nuclei displaying apoptotic morphology was significantly lower in re-plated than in control cultures (P < 0.001; Fig. 4c–e). There was no significant difference between the percentage of dopaminergic neurones in re-plated cultures and that in controls (Fig. 4c–e).

Effects of transplantation of pSGDF5-transfected E13 VM cultures into the 6-OHDA-lesioned rat striatum

Following the in vitro tests on the effects of GDF5 transfection in E13 VM cells, an in vivo study was performed to evaluate the survival and functional effects of such cells after transplantation into the 6-OHDA-lesioned striatum. Behavioural and post-mortem analyses were performed on each of the following groups of rats: (i) “pSGDF5-transfected” group (lesion, then transplant of E13 rat VM cells transfected with pSGDF5), (ii) “mock-transfected” group (lesion, then transplant of E13 rat VM cells transfected with pcDNA3.1), (iii) “E14 VM transplant” group (lesion, then transplant of freshly dissected E14 rat VM tissue), (iv) “6-OHDA only” group (lesion, no transplant) and (v) “control” group (no lesion, no transplant).

At 1 week after 6-OHDA lesion, but before transplantation, each of the rats (apart from those of the control group) rotated ipsilaterally at a rate of at least 8 turns per minute in response to amphetamine (Fig. 5a). At 2 weeks after lesion surgery, significant decreases in rotation rate were observed in each of the three transplant groups (P < 0.01 for E14 VM transplant and mock-transfected groups; P < 0.001 for pSGDF5-transfected group), compared to the respective pre-grafting values (Fig. 5a). At the post-grafting test-point, each of the three transplant groups displayed rotation rates that were significantly lower than that of the 6-OHDA only group (P < 0.001; Fig. 5a). At this time-point, the rotation rate of the pSGDF5-transfected transplant group was significantly lower than that of both the E14 VM and mock-transfected transplant groups (P < 0.01; Fig. 5a).
https://static-content.springer.com/image/art%3A10.1007%2Fs00702-010-0392-9/MediaObjects/702_2010_392_Fig5_HTML.gif
Fig. 5

a Number of amphetamine-induced rotations in each treatment group at 1 and 4 weeks after lesion surgery (i.e. 1 week before and 2 weeks after grafting). Data are expressed as mean ± SEM calculated over a 60-min period beginning 5 min after amphetamine administration. N = 4 for each group at each time-point. **P < 0.01, ***P < 0.001 versus same treatment group at pre-grafting time-point; ###P < 0.001 versus 6-OHDA group at post-grafting time-point; ++P < 0.01 versus pSGDF5-transfected group at post-grafting time-point. b Somal area and c total neurite length of TH-immunopositive cells in freshly dissected E14 VM grafts, mock-transfected grafts and pSGDF5-transfected grafts at 3 weeks after intrastriatal grafting. Data are expressed as mean ± SEM. *P < 0.05, **P < 0.01 versus freshly dissected E14 VM grafts; #P < 0.05, ##P < 0.01 versus mock-transfected grafts. d Representative photomicrograph of a cryosection through a 6-OHDA-lesioned striatum 3 weeks after intrastriatal grafting of pSGDF5-transfected E13 rat VM cultures, immunohistochemically stained for TH (green) and GFAP (red), and counterstained with DAPI (blue). The boxed area indicated in (d) is shown at higher magnification in (e)

Immunohistochemical analysis at 4 weeks after transplantation showed the presence of TH-positive cells in the left striatum of each rat which had received a transplant (Fig. 5b, c). Comparable numbers of TH-positive cells were present in the striata of rats which had received pSGDF5-transfected VM cultures (5,861 ± 191) or mock-transfected cultures (4,922 ± 496) as in those which had received freshly dissected E14 VM tissue (5,239 ± 223). There were no significant differences between the treatment groups.

The mean somal area and total neurite length of TH-immunopositive cells were significantly higher in pSGDF5-transfected transplants than in either mock-transfected or freshly dissected transplants (P < 0.05; Fig. 5b, c).

Discussion

Two promising strategies for the optimisation of cell transplantation therapy for PD have emerged in the past several years. The first is the delivery of neurotrophic factors to improve the survival of transplanted dopaminergic neurones. The second involves the use of other types of cells, such as neural progenitor or stem cells, as an alternative to foetal dopaminergic neurones. The current study examined the effects of overexpression of the dopaminergic neurotrophic factor, GDF5, on the survival of embryonic rat dopaminergic neurones in vitro and after transplantation into an in vivo rat model of PD. Such an approach may also be applied to neural progenitor and stem cells, in an attempt to optimise the post-transplantation survival of dopaminergic neurones induced from these cells.

E14 VM is the standard tissue used for transplantation in animal models of PD. Since the current study required VM cells to be cultured for a day before transfection, then cultured for a second day before removal for transplantation, the cellular compositions of E13 and E14 rat VM cultures were initially compared. The cell phenotypes examined were neurones, dopaminergic neurones, glia and dividing cells. The composition of E13 cultures after 2 DIV was found to be practically identical to that of E14 cultures after 1 DIV. Thus, it is reasonable to compare the survival and functional efficacy of E13 cultures transplanted after 2 DIV to the standard E14 transplants in an in vivo study. In contrast, E13 cultures after 1 DIV contained significantly fewer dopaminergic neurones and total neurones, and significantly more nestin-positive cells and dividing cells, than either of the other two culture types, indicating their less mature nature.

Optimisation experiments using a GFP-expressing plasmid (pEGFP-N1) indicated that transfections performed using 2 μg DNA and 4 μl Lipofectamine 2000™ routinely yielded transfection efficiencies of 15–21% in E13 and E14 VM cultures. These amounts of DNA and Lipofectamine 2000™ were used in subsequent transfections with the GDF5-expressing plasmid (pSGDF5). The transfection efficiency was comparable to those in previously published studies using liposome transfection reagents in primary neuronal cultures. For example, transfection efficiencies of 20–30% have been achieved in E18 rat cortical and hippocampal neuronal cultures using Lipofectamine 2000™ (Ohki et al. 2001). Alberi et al. (2004) reported a transfection efficiency of ~25% (based on the rate of cell death induced by the transfected siRNA) using the Transmessenger™ reagent in E12.5 mouse VM primary cultures. Some studies have presented data in terms of absolute number of transfected cells rather than percentages (Bauer et al. 2006; Yu and Arumade 2008); therefore, it is not possible to compare transfection efficiencies between studies. It has been reported that Lipofectamine™ can induce cellular toxicity, demonstrated by an increase in PI uptake at 1 day post-transfection, in E15 rat VM explant cultures (Bauer et al. 2006). In contrast, the present study found that the numbers of dead (PI-positive) cells in pSGDF5-transfected and in mock-transfected cultures were not different from that in controls, indicating that the transfection protocol used here was not detrimental to the cells. The difference between the findings of these two studies may be due to the different types of cultures (E15 VM explants vs. E13 VM dissociated cultures) or to the variation in transfection reagent (Lipofectamine™ vs. Lipofectamine 2000™).

Intracellular expression of GDF5 protein in pSGDF5-transfected E13 VM cultures was demonstrated using Western blotting. Cell lysates from pSGDF5-transfected, but not those from mock-transfected, cultures displayed a GDF5-immunoreactive band at approximately 55 kDa. No GDF5-immunoreactive band at ~25 kDa was observed in these samples. This is in accordance with a previous study which showed that GDF5 exists intracellularly as an uncleaved precursor protein of 55 kDa, and not as the active dimeric (25 kDa) form (O’Keeffe et al. 2004b). In support of the Western blotting results, immunocytochemical analysis showed strong expression of GDF5 throughout the soma of numerous cells in pSGDF5-transfected cultures. Weak GDF5 expression was observed in some cells in the mock-transfected cultures, which probably reflects a low level of endogenous GDF5 expression in these cells, which was not detectable by Western blotting. Endogenous expression of GDF5 in the E13 rat VM has been previously reported (O’Keeffe et al. 2004b).

Secretion of the fully processed active 25 kDa GDF5 dimer (Hotten et al. 1994) into the culture medium of pSGDF5-transfected cultures was detectable by Western blotting for at least 9 DIV. GDF5 dimer was also detectable in the medium of mock-transfected cultures, but at a lower level. Again, this probably reflects a low level of endogenous GDF5 expression and release by E13 VM cells.

GDF5 secretion by transfected E13 VM cultures resulted in a significant increase in dopaminergic neuronal survival, compared to mock-transfected or control cultures. This survival-promoting effect was observed at 3 days after transfection, and larger effects were seen after 5 and 7 days. This agrees with previous studies which found that treatment with recombinant human GDF5 significantly increased dopaminergic neuronal survival in primary cultures of E14 rat VM for up to 8 DIV (O’Keeffe et al. 2004a; Wood et al. 2005). At 5 and 7 DIV, the number of glial cells was also significantly higher in pSGDF5-transfected cultures than in mock-transfected or control cultures, in agreement with a previous report using recombinant GDF5 (Wood et al. 2005). That study showed that GDF5 acts directly on dopaminergic neurones in vitro to confer its neurotrophic effect, rather than indirectly via an action on glial cells (Wood et al. 2005). Thus, it is unlikely that the survival-promoting effects of GDF5 released from the transfected cells in the present study are mediated through the increased astroglial cell population. GDF5 secretion by these cells also had effects on their morphology, with dopaminergic neurones in the transfected cultures displaying larger cell soma than those in the mock-transfected or control cultures. Treatment of E14 VM cultures with recombinant GDF5 has been previously shown to increase somal area of dopaminergic neurones in E14 rat VM primary cultures (O’Keeffe et al. 2004a).

This study found that overexpression of GDF5 prevented cell death induced by the dopaminergic neurotoxin, 6-OHDA, in E13 VM cultures. Recombinant human GDF5 has been previously shown to protect cultured embryonic rat dopaminergic neurones against another dopaminergic neurotoxin, 1-methyl-4-phenyl-pyridinium ion (MPP+; Krieglstein et al. 1995b). In the current study, pSGDF5 transfection significantly protected against the increase in dead cells and the decrease in the percentage of dopaminergic neurones which was induced by 6-OHDA. This experiment also showed that the transfection procedure itself did not have deleterious effects on cell survival, since pSGDF5 or mock transfection did not significantly increase the number of dead cells or decrease the percentage of TH-positive neurones compared to those in control cultures. In fact, GDF5 overexpression conferred neurotrophic effects on dopaminergic neurones in these cultures, indicated by an increased TH-positive neuronal percentage in pSGDF5-transfected cultures. This is in agreement with previous studies which showed neurotrophic effects of recombinant GDF5 in vitro (Krieglstein et al. 1995b; O’Keeffe et al. 2004a; Wood et al. 2005; Clayton and Sullivan 2007).

The effects on cell viability of enzymatic removal of the cells from culture were examined, to control for any detrimental effects that removal from culture may have on these cells after transplantation. The number of nuclei displaying apoptotic morphology in cells which had been mock-transfected at 1 DIV, then removed from culture and re-plated at 2 DIV, was significantly lower than that in control cultures. This may be because of the loss during the re-plating procedure of dead (or dying) cells due to their failure to adhere to the culture plates. Nevertheless, this finding indicates that physical removal per se did not result in an increase in the number of apoptotic cells compared to the control cultures. Dopaminergic neurones in the re-plated cultures displayed morphology typical of those in control cultures, and there was no significant difference in either percentage of dopaminergic neurones or total cell number between these two culture types. These data show that the transfection and enzymatic removal procedures in themselves did not aberrantly affect the viability or composition of the cultures and thus give support to the proposal of using transfected cells in transplantation studies.

In an in vivo study, the survival and functional effects of pSGDF5-transfected and mock-transfected E13 cultures were compared to those of the standard tissue which is used for transplantation in rat models of PD, freshly dissected E14 VM cells. The percentages of grafted TH-positive neurones surviving in the striatum at 4 weeks after transplantation in each of the three treatment groups were similar to those in other studies (for review, see Brundin et al. 2000). It has been reported that around 2,000 transplanted dopaminergic neurones are necessary to achieve a ~50% reduction in amphetamine-induced rotations at 2 weeks after transplantation (Nakao et al. 1994). In the present study, almost complete reversal of rotational asymmetry was achieved at 2 weeks after transplantation, with transplants consisting of 4,900–5,900 dopaminergic cells. The survival of the transplanted dopaminergic cells was ~20% in the freshly dissected and mock-transfected groups (assuming that around 10% of the transplanted cells were dopaminergic), while it was ~23% in the pSGDF5-transfected group. Although the difference between these groups was not statistically significant, the high survival rate of dopaminergic neurones in the pSGDF5-transfected grafts is very promising, considering that these cells had been transfected and then removed from culture prior to transplantation. These procedures may have been expected to result in a reduction in cell viability, but this study shows that this is not the case. The fact that the dopaminergic cell survival rate in the mock-transfected grafts was almost identical to that in the freshly dissected grafts indicates that the slight increase in dopaminergic cell survival in the pSGDF5-transfected grafts was due to the presence of GDF5. A previous study reported a 2.3-fold increase in dopaminergic neuronal survival in GDF5-treated E14 rat VM grafts over that in untreated grafts at 8 weeks after transplantation (Sullivan et al. 1998b). In that study, 500 μg of recombinant human GDF5 was added to the E14 VM cell suspension immediately prior to transplantation. It is likely that the amount of GDF5 produced by the transfected cells in the present study was substantially less. The production of GDF5 by the cells is likely to have declined during the 4-week period before killing the animals; this would be expected from the in vitro studies, which indicated a decrease in the level of GDF5 secretion by transfected cells after 9 DIV. Follow-on studies using viral vectors to achieve long-term delivery of GDF5 to the parkinsonian rat brain have been conducted (manuscript in preparation).

Although there was no significant difference in dopaminergic cell survival between the three transplant groups, the rotational rate in the pSGDF5-transfected graft group was significantly lower than that in either the mock-transfected or freshly dissected graft group. This may reflect a protective effect of GDF5 released from the overexpressing transplanted cells on the remaining dopaminergic terminals in the host striatum.

Analysis of the morphology of the transplanted dopaminergic neurones showed that the pSGDF5-transfected cells displayed significantly larger cell soma and significantly greater total neurite length than either mock-transfected or freshly dissected cells. This suggests that the GDF5-overexpressing cells are more mature than those in the other transplants and may explain the significantly lower rotational rates observed in the pSGDF5 transplant group. That is, the more mature GDF5-overexpressing cells possess more and/or longer neurites and thus have the ability to make more synaptic connections with the host striatum. A higher level of dopamine release from these transfected cells could account for the significantly greater reversal of rotational asymmetry observed in this group.

Since the decision to stop all clinical trials involving GDNF infusion, there has been much debate on the future of this type of approach to PD therapy. It is clear that several issues, in particular improval of delivery methods and minimisation of side-effects in patients, will need to be addressed before the use of neurotrophic factors is a viable therapeutic option for this disease (see Sherer et al. 2006; Deierborg et al. 2008). The field of neuronal cell transplantation also needs to be optimised, perhaps to consider the use of alternatives to foetal dopaminergic neurones (such as neural or embryonic stem cells), as well as improvements to the methodologies which are currently used.

The present study shows that overexpression of the neurotrophic factor GDF5 confers significant trophic and protective effects on dopaminergic neurones in cultures of E13 rat VM. Furthermore, the procedure of transfection and removal from culture did not have any detrimental effects on these cells after their subsequent transplantation into the striatum of an adult rat model of PD. This study combines the approaches of foetal neuronal transplantation and of neurotrophic factor administration, as therapeutic options for PD. It demonstrates that transfection with neurotrophic factors in vitro prior to transplantation is a viable option for improving graft survival and function. This approach could easily be applied to other cell types, such as stem cells, prior to transplantation. Indeed, the transfection procedure could be used to introduce various other types of factors, such as anti-apoptotic proteins or molecules that can induce a neuronal/dopaminergic neuronal phenotype in uncommitted stem or progenitor cells, prior to their transplantation.

Acknowledgments

This work was funded by the Health Research Board of Ireland. The authors wish to thank Dr Jens Pohl and Dr Michael Hanke of Biopharm GmbH, Germany, for the generous gifts of pSGDF5 plasmid and aMP5 antibody. We also wish to acknowledge Ms Bereniece Riedewald for her assistance with the preparation of the figures and Dr Gerard O’Keeffe for critical reading of the manuscript.

Copyright information

© Springer-Verlag 2010