Background
Parkinson’s disease (PD) is a common neurodegenerative movement disorder that affects 1% of the population aged over 65. Clinically, PD is characterized by major motor symptoms consisting of bradykinesia, cogwheel rigidity, resting tremor and postural instability. The neuropathological cardinal features of the disease comprise both the preferential loss of the dopaminergic neurons of the substantia nigra (SN) and the prevalence of large cytoplasmic inclusions rich in alpha-synuclein (α-synuclein) protein, named Lewy bodies (LBs), which are present in the remaining cells [
1]. In addition, whole locus duplications and triplications and genetic single mutations (A53T, A30P, E46K, H50Q, G51D) in the
SNCA gene coding for α-synuclein result in autosomal-dominant familial forms of PD [
2‐
8]. The severity of the phenotype correlates with the
SNCA genomic dosage implying that high levels of α-synuclein can in fact trigger the neurodegenerative process [
9]. Finally, genome-wide association studies also support this critical role for α-synuclein in the pathogenesis of PD [
10,
11]. The discovery of α-synuclein both as a genetic cause for the disease and as the major component of the LBs in sporadic and familial cases of PD has strengthened the link between sporadic and hereditary PD forms and the possibility of an underlying common mechanism at the origin of the disease.
Since α-synuclein is so tightly linked to the pathogenesis of PD, numerous transgenic mouse lines overexpressing wild-type (WT) or mutant α-synuclein were developed over the past decade, aiming at replicating the neuropathology seen in patients [
12,
13]. Despite the fact that these mice have proven useful in modelling some features of the disease such as the abnormal accumulation of α-synuclein in cells, they do not display a convincing progressive degeneration of the nigral dopaminergic neurons nor a dopamine-dependent motor phenotype. Conversely, the overexpression of α-synuclein by means of viral vectors in rodents and non-human primates induces a progressive dopaminergic cell loss over time as the most prominent feature (see [
14] for a review). We have previously reported the overexpression of α-synuclein in mouse and rat brain using lentiviral vectors [
15‐
17], but recombinant adeno-associated viral (rAAV) vectors have steadily gained more interest to express a gene of interest in the brain in a controlled spatio-temporal manner because of their high titers and high tropism for dopaminergic cells, especially for the newer serotypes [
18‐
22].
Viral vector-based α-synuclein rat and non-human primate models display several cardinal neuropathological features of PD. Indeed, rAAV vector-mediated overexpression of either WT or mutant A53T α-synuclein induces a 20 to 80% dopaminergic cell loss in 3 to 8 weeks in rats, depending on the study and the serotype used [
23‐
28], and a 30 to 60% dopaminergic neuron loss at 4 months after injection in marmosets [
29]. Beside the dopaminergic cell death, rat and non-human primate models exhibit other hallmarks of PD such as the prevalence of α-synuclein-positive inclusions in the SN or the presence of a phosphorylated form of α-synuclein at serine 129 (P-S129), which is largely abundant in the LBs of PD patients [
30,
31]. Although promising, these viral vector-based α-synuclein models still suffer from a certain degree of variability and slow progression of the phenotype, hindering their value for testing novel therapeutics. To address this, we have taken a novel vector, rAAV2/7, which can be prepared at high titer and for which we have shown its efficient transduction of the dopaminergic neurons of the SN [
32]. We have developed an improved rat model for PD by means of rAAV2/7 vector-mediated overexpression of A53T α-synuclein in the SN, which displays progressive and robust neurodegeneration (Van der Perren et al., submitted). Intriguingly, the few studies with α-synuclein rAAV vectors performed in mice until now report either no or very limited dopaminergic neuron loss, with a maximum of 20 to 35% at 2 to 6 months after injection [
33‐
38]. Furthermore, no obvious α-synuclein-positive inclusions were detected in the surviving cell bodies [
35].
In view of the increasing number of transgenic and knockout mice, the availability of a viral vector-based α-synuclein mouse model would offer several interesting scientific possibilities. Therefore, we decided to explore whether it is possible to obtain a robust parkinsonian phenotype in mice with α-synuclein-encoding rAAV vectors. We performed a comprehensive comparative study of 3 different vector doses of rAAV2/7-α-synuclein WT in mouse SN and analyzed them at separate time points up to 2 months after surgery. In addition, we compared the overexpression of human WT and mutant A53T α-synuclein in mouse SN delivered at the same vector dose. We assessed the changes induced by the sustained localized overexpression of α-synuclein in mouse SN by means of motor behavioural tests, immunohistochemical stainings followed by stereological quantifications, and protein solubility analysis. We succeeded in developing a mouse model based on rAAV2/7 vector-mediated overexpression of α-synuclein with robust dopaminergic cell death and α-synuclein inclusions in the SN.
Methods
rAAV vector design and production
The plasmids for the rAAV vector production were previously described [
39]. These plasmids include the construct for the AAV2/7 serotype, the AAV transfer plasmid and the pAdvDeltaF6 adenoviral helper plasmid. The transfer plasmid encoded eGFP or the human WT or A53T α-synuclein transgene, followed by the woodchuck hepatitis posttranscriptional regulatory element, under the control of the CMVie-enhanced-synapsin1 promoter.
The viral vectors were produced at the Leuven Viral Vector Core (LVVC,
http://www.kuleuven.be/molmed/research/research_viral_vector_technology.html) from the KU Leuven (Leuven, Belgium). Vector production and purification were performed as described in detail elsewhere [
19]. The final titers for the concentrated vector stocks ranged between 8,0E + 11 and 2,7E + 12 genome copies per millilitre (GC/ml) as determined by quantitative PCR.
Stereotactic surgery
Housing and handling of mice was done in compliance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and approved by the Bioethical Committee of the KU Leuven. All animals were housed under 12/12 h light/dark cycle with access to food and water
ad libitum. Female, 8 weeks old, C57BL/6 mice received the human WT α-synuclein rAAV2/7 vector at different titers: 2,6E + 11 GC/ml; 4,0E + 11 GC/ml or 8, 0E + 11 GC/ml. The rAAV2/7-eGFP vector was injected at 4,0E + 11 GC/ml or 8,0E + 11 GC/ml. The human A53T α-synuclein rAAV2/7 vector was tested at a single dose: 4,0E + 11 GC/ml. The animals were anaesthetized and placed in a stereotactic head frame (Stoelting, IL, USA). After making a midline incision of the scalp, a burr hole was drilled in the appropriate location for the SN at the right site of the skull. 2 μl of viral vector were injected at a rate of 0.25 μl/min with a 30-gauge needle on a 10 μl Hamilton syringe. The needle was left in place for an additional 5 minutes before being slowly withdrawn from the brain. Coordinates for mouse SN were anteroposterior (AP) -3.1 mm and mediolateral (ML) -1.2 mm relative to bregma, and dorsoventral (DV) -4.0 mm from the dural surface [
40].
Motor behaviour
Three groups of mice (n = 10) were injected with rAAV2/7-eGFP (8,0E + 11 GC/ml) or human WT α-synuclein rAAV2/7 vector (8,0E + 11 GC/ml and 4,0E + 11 GC/ml) for behavioural evaluation in the cylinder test, the rotarod test and the open field test.
The cylinder test, which measures asymmetry in spontaneous forelimb use, was performed at 1, 4, 8 and 12 weeks after injection. Mice were placed individually inside a glass cylinder (12 cm diameter, 22 cm height) positioned in front of two vertical mirrors in order to be able to view the mouse from all angles. No habituation of the animals to the testing cylinder was allowed before video-recording. Video-recordings were examined by an observer blinded to the animal’s identity. Between 20 and 30 wall touches per animal (contacts with fully extended digits executed with the forelimb ipsilateral and contralateral to the lesion) were counted.
In the rotarod test, animals were trained for 2 minutes at a speed of 4 rpm. After this initial training, mice performed 8 trials of maximum 5 minutes with increasing speed starting from 4 rpm up to 40 rpm. This test was performed at 14 weeks after injection and fall off time was recorded.
In the open field test, animals were placed for 30 minutes in the dark prior to the open field cage (50×50 cm). After 1 minute of adaptation, the next 10 minutes were video-recorded for analysis. This test was performed at 15 weeks after injection.
Histology
At selected time points, mice were deeply anesthetized using an overdose of pentobarbital (Nembutal, CEVA Santé, Belgium) and transcardially perfused with saline followed by ice-cold 4% paraformaldehyde in phosphate-buffered saline (PBS). 50-μm-thick coronal brain sections were cut with a vibrating microtome (HM650V, Microm, Germany) and stored at 4°C in 0.1% sodium azide in PBS. Immunohistochemistry was performed under uniform conditions on free-floating sections using antibodies raised against α-synuclein (rabbit polyclonal, 1:5000, Millipore 5038), tyrosine hydroxylase (TH, rabbit polyclonal, 1:5000, Millipore 152) and P-S129 α-synuclein (mouse monoclonal, 1:5000, Elan Pharmaceuticals, [
31,
41]). Sections were pretreated with 3% hydrogen peroxide for 10 minutes to quench endogenous peroxidase activity and then incubated overnight at room temperature in primary antibody solution with 10% normal goat serum (DakoCytomation, Belgium). As secondary antibody, we used biotinylated anti-rabbit or anti-mouse IgG (1:600, DakoCytomation), followed by incubation with streptavidin-horseradish peroxidase complex (1:1000, DakoCytomation). TH immunoreactivity was visualized using Vector SG (SK-4700, Vector Laboratories) as chromogen; and (P-S129) α-synuclein was visualized with 3,3-diaminobenzidine (0,4 mg/ml, Sigma-Aldrich). After being rinsed and mounted, the sections were cover-slipped using DPX mounting medium (Sigma).
For fluorescent double staining, sections were rinsed three times in PBS and incubated overnight in the dark in PBS-0.1% Triton X-100 and 0.1% sodium azide, 10% donkey serum and the following antibodies: chicken anti-GFP (1:500, Aves Labs 1020), rabbit anti-α-synuclein (1:1000, Millipore 5038), mouse anti-TH (1:1000, Millipore MAB318), mouse anti-Neuronal Nuclei (NeuN, 1:500, Chemicon MAB377), mouse anti-α-synuclein (1:100, Invitrogen LB309) and rabbit anti-Glial Fibrillary Acidic Protein (GFAP, 1:100, Dako Z0334). After rinsing, the sections were incubated in the dark for 1 h in fluorochrome-conjugated secondary antibodies: donkey anti-chicken FITC (1:400, Lucron), donkey anti-rabbit Alexa 647 and donkey anti-mouse Alexa 555 (1:400, Molecular Probes, Invitrogen). After being rinsed and mounted, the sections were cover-slipped with Mowiol (Sigma).
Stereological quantification and confocal microscopy
The volume of α-synuclein expression in the SN and of TH-immunoreactive fibbers in the striatum (STR) was determined by stereological volume measurements based on the Cavalieri method as described before [
42,
43]. The number of TH-, α-synuclein- and P-S129 α-synuclein-positive cells in the SN was estimated using a random sampling stereological counting method, the optical fractionator [
44], in a computerized system (StereoInvestigator; MicroBright-Field, Magdeburg, Germany). Coefficient of error attributable to the sampling was calculated according to Gundersen and Jensen [
45]. Errors ≤0.10 were accepted. The SN was delineated from -2.54 to -3.88 mm posterior to bregma based on the Paxinos atlas for the mouse brain [
40]. For each animal, every third section throughout the rostro-caudal extent of the SN and every fourth section covering the entire extent of the STR were incorporated to the counting procedure. The conditions of the experiment were blinded to the investigator.
Fluorescent double immunostainings for α-synuclein or GFP and TH, and for α-synuclein and NeuN or GFAP, were visualized by confocal laser scanning microscopy (Fluoview 1000, Olympus). Estimation of co-transduction was performed based on quantifications of double-stained cells on 4 different sections.
After cervical dislocation, the SN were rapidly dissected from one to two 1 mm-thick coronal sections using a mouse brain slicer, and snap frozen in dry ice. Frozen nigral tissues were homogenized in 5 volumes of Tris-buffered saline (TBS, pH 7.5, 10 mM Tris, 0.15 M NaCl) plus complete protease inhibitor (TBS+, Roche Applied Science) using a tissue homogenizer (Ultra-Turrax; IKA Werke, Staufen, Germany). Next, proteins were sequentially extracted, slightly modified from previously published procedures [
46,
47]. Shortly, samples were spun at 120,000 × g for 30 minutes at 4°C. The resulting supernatant represented the TBS soluble fraction. The pellet was then resuspended in TBS+ containing 1% Triton X-100 (Sigma-Aldrich), TBS+ containing 1 M sucrose and RIPA buffer (50 mM Tris–HCl, pH 7.4, 175 mM NaCl, 5 mM ethylenediaminetetraacetic acid [EDTA], 1% NP-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate [SDS]). After a last ultra-centrifugation step, the detergent-insoluble pellet was solubilised in 2 M urea/5% SDS to preserve α-synuclein structures [
48]. Samples were supplemented with 10% glycerol and stored at -80°C.
Western blot analysis
Protein concentration was measured using the Bradford method (Protein Assay Dye Reagent Concentrate, Biorad, Munich, Germany). For western blotting, 30 μg of fractionated protein extracts were loaded on 12% acrylamide gel (Serva, Heidelberg, Germany) and blotted onto a methanol-activated PVDF membrane (Immobilon P, Millipore Corporation, Billerica, MA, USA). Immunoblots were blocked in 5% dry milk in TBST buffer (TBS plus 0.1% Tween 20) and subsequently probed with: human specific anti-α-synuclein 15G7 (1:10, AG Scientific), anti-α-synuclein Syn1 detecting both mouse and human α-synuclein (1:1000, BD Bioscience), anti-P-S129 α-synuclein (1:1000, Epitomics, 2014–1). Anti-GAPDH (1:1000, Abcam 125247) or anti-β-actin (1:10000, Sigma A5441) was used as internal loading control. Bound antibodies were detected with horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence (ECL or ECLplus, GE Healthcare) followed by exposure to hyperfilm (GE Healthcare). ImageJ software was used to determine the optical density of protein bands and all data were analyzed for each viral vector group (n = 3) based on 3 independent blots and normalized to the expression of the corresponding GAPDH or β-actin loading control within line.
Statistical analysis
Statistical analysis was performed using the GraphPad Prism 5.0 software package. Results are expressed as means ± standard error of the mean. Statistical significance was assessed using two-way ANOVA followed by Bonferroni post-hoc test for intergroup comparisons or one-way ANOVA for the behavioural tests and the western blotting analysis. Statistical significance level was set as follows: * if P < 0.05, ** if P < 0.01, *** if P < 0.001.
Discussion
The goal of the current study was to develop a robust mouse model that displays the most relevant neuropathological hallmarks of PD based on the targeted viral vector delivery of α-synuclein to the SN. We showed that rAAV2/7 vector-mediated overexpression of either WT or mutant A53T α-synuclein results in an overt degeneration of the nigral dopaminergic neurons concomitant with the development of motor impairments and the formation of cytoplasmic α-synuclein inclusion bodies in a particularly short time frame.
The identification of α-synuclein as a key factor in idiopathic and inherited PD has impelled the development of new mouse models overexpressing α-synuclein over the last decade. However, transgenic mouse lines have failed to reproduce convincingly a meaningful progressive dopaminergic neurodegeneration in the SN [
12]. In contrast, local delivery of viral vectors overexpressing α-synuclein in rat and non-human primate brains leads to gradual nigral dopaminergic neuron loss, which is in general more prominent with rAAV vectors compared to lentiviral vectors [
50]. However, the rAAV vector-based mouse models published to date exhibit no or limited dopaminergic cell loss with a maximum of 20 to 35% between 2 and 6 months post-injection [
33‐
38]. Moreover, these mouse models do not display striatal dopaminergic fiber loss [
35,
36]. We have previously reported that lentiviral vector-mediated overexpression of WT or mutant A30P α-synuclein in mouse SN induces up to 25% dopaminergic neurodegeneration at 10 to 12 months after injection [
16]. The long time scope needed to achieve a substantial neurodegeneration differs from our new mouse model developed in the present study in the type of viral vector used. rAAV vectors are more efficient than lentiviral vectors in transducing the dopaminergic neurons of the SN because of their high titers and neuronal tropism. In this regard, it is remarkable that a very recent study using lentiviral vectors for the overexpression of α-synuclein in mouse SN has reported around 60% dopaminergic cell loss after 45 days for the A30P, A53T and E46K mutant and about 25% for WT α-synuclein [
51].
The experience with viral vector-based rat models has proved that transduction efficiency and transgene expression levels greatly determine the degree of nigral neuron degeneration and the time span over which this occurs [
52,
53]. Consequently, we have previously optimized large-scale and flexible manufacturing processes for rAAV vector production [
54]. In addition, we have shown that our improved rAAV vectors efficiently transduce nigral dopaminergic neurons [
19]. Based on this comparative study, we selected the rAAV2/7 serotype for overexpression of α-synuclein because it combines high transgene expression levels and high transduction efficiency of the dopaminergic neurons
in vivo. We optimized a rat model of PD based on rAAV2/7 vector-mediated overexpression of A53T α-synuclein in the SN. Rats receiving a vector dose of 3,0E + 11 GC/ml displayed 80% loss of the nigral dopaminergic neurons at 4 weeks post-injection (Van der Perren et al., submitted). In parallel to that study, we have shown here that targeted nigral overexpression of WT α-synuclein by means of a rAAV2/7 vector induces up to 82% dopaminergic cell death in mouse SN within 2 months. The extent of cell loss for the whole population of α-synuclein overexpressing cells (maximum of about 50% at 8 weeks) is lower than the loss of TH-positive neurons (82% at 8 weeks with the highest titer), which suggests that the dopaminergic neurons are more vulnerable to α-synuclein neurotoxicity.
α-Synuclein pathology is the second major neuropathological feature of PD after the progressive dopaminergic cell loss. Therefore, it is an essential characteristic to replicate in a relevant animal model of PD. However, the rAAV vector-based mouse models for α-synuclein reported in literature until now do not present detectable α-synuclein-positive aggregates as far as 6 months post-injection [
35]. Other studies with rAAV overexpression of α-synuclein in rats and non-human primates demonstrate the presence of α-synuclein-rich inclusions [
23‐
25,
28,
29,
55]. Both our new rAAV2/7-based and previous lentiviral vector-based mouse models for α-synuclein exhibited visible inclusions abundant in α-synuclein protein in the surviving neurons [
16,
17]. In addition, we here demonstrated increasing levels of insoluble transgenic α-synuclein over time. We also observed a gradual increase in the percentage of P-S129 α-synuclein-positive cells in the SN, supporting the progressive development of α-synuclein neuropathology and the presence of LB-like α-synuclein-rich inclusions.
Another objective of the current study was to test a dose dependency of viral vector-mediated overexpression of α-synuclein. For this approach, we used 3 different doses of WT α-synuclein vector in mouse SN that led to a dose-dependent and gradual neurodegeneration over time. With our lowest vector dose, we achieved a similar degree of dopaminergic cell death as previous studies in mice [
33,
35,
37], while the highest dose tested induced extensive (57%) nigral dopaminergic neurodegeneration in just 4 weeks. To our knowledge, this is the first dose dependence study of viral vector overexpression of α-synuclein carried out in mouse brain. Similar comparative studies have been performed in rats, which also reported differences in rates of neurodegeneration in the SN and in behavioural performance based on the viral vector titer used [
28,
52,
53,
56].
In the present study, we investigated the behavioural performance of the mice at two different doses of WT α-synuclein viral vector. Somewhat surprisingly, the cylinder test showed only a minor asymmetric forepaw use at 12 weeks after injection compared to the eGFP control group, while all previous time points did not reveal any impairment. In an independent study, we also did not observe impairments in spontaneous forelimb use in mice with ~50% TH cell loss achieved at 4 weeks post-injection by overexpression of rAAV2/7-α-synuclein WT in mouse SN (Oliveras-Salvá et al., submitted). These observations are in contrast with own previous results obtained in rats where the same level of dopaminergic neuron loss resulted in a clear decrease in contralateral forepaw use (Van der Perren et al., submitted). However, careful observation of the mice in the cylinder test revealed a general behavioural impairment and slowness of movement. Therefore, we also performed the rotarod and open field tests which showed pronounced motor behaviour deficits in both α-synuclein groups at a later time point. These impairments can be ascribed to the extensive dopaminergic cell death obtained in mouse SN after the injection of the rAAV2/7-α-synuclein vector.
Finally, we evaluated whether the overexpression of the A53T clinical mutant triggered a substantially different degeneration of the dopaminergic cells in the SN compared to the WT α-synuclein. Overall, the neurodegeneration induced over time by the A53T α-synuclein viral vector in the SN and the STR was slightly enhanced when compared to the WT α-synuclein vector, although the difference was never significant. To date, there is no consensus in literature about the neurotoxicity of mutant A53T α-synuclein in comparison to WT α-synuclein
in vivo. This is partly due to the scarcity of parallel comparative studies in which WT and A53T α-synuclein are investigated under the same conditions. Other studies have compared human WT and mutant A53T α-synuclein encoded by viral vectors in other species. On the one hand, Eslamboli and colleagues reported that the A53T variant induces a significantly enhanced dopaminergic neuron loss when compared to WT at one year after rAAV nigral injection in non-human primate brain and, correlating with the degree of induced cell death, two different microglia activation profiles were reported [
55,
57]. On the other hand, other studies performed in rats and non-human primates showed that rAAV vector-based delivery of WT and A53T α-synuclein induces a comparable neurodegenerative process, as well as aggregate formation [
23,
29,
56]. Our findings in the present study are in agreement with these latter publications. In mice, we have previously reported that overexpression of WT and A30P α-synuclein by means of lentiviral vectors induces a similar rate in dopaminergic cell death and comparable α-synuclein pathology in a long-term study of 12 months [
16]. In our hands, the overexpression levels rather than the clinical mutation determine the neuropathological changes in our model.
Acknowledgements
The authors thank Dr Annelies Michiels, Sylvie De Swaef and Nam-Joo Van der Veken of the Leuven Viral Vector Core (LVVC,
http://www.kuleuven.be/molmed/research/research_viral_vector_technology.html) for the rAAV vector productions; and Marly Balcer, Joris Van Asselberghs and Caroline van Heijningen for excellent technical assistance. We thank Professor Johan Hofkens and Charlotte David for the use of the confocal laser scanning microscope (Molecular Imaging and Photonics, KU Leuven).
This work was supported by the European FP7 ITN NEUROMODEL (PITN-GA-2008-215618), the FP7 RTD MEFOPA (HEALTH-2009-241791), the FP7 project ‘INMiND’ (HEALTH-F2-2011-278850), the FWO Vlaanderen (G.0768.10), the MJFox Foundation (Target validation 2010), the Queen Elisabeth Medical Foundation the KU Leuven PF/10/017 IMIR and the IWT-Vlaanderen (IWT/SBO 80020).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MOS made substantial contribution to conception and design of the study, performed all in vivo experiments and wrote the manuscript. AVdP participated in the viral vector constructions. NC participated in the design of the western blotting study and carried out its analysis. SN participated in the analysis of the western blotting study. SS and RD participated in the design of the motor tests. CVdH participated in the viral vector constructions and productions, and in the design of the in vivo experiments. VB conceived and designed the study, and helped to write the manuscript. All authors read and approved the final manuscript.