Introduction
Parkinson’s disease (PD) is a progressive neurodegenerative disease characterized by a complex motor disorder known as Parkinsonism which manifests as resting tremor, bradykinesia, rigidity, and postural abnormalities. Additionally, there are many non-motor symptoms such as olfactory deficit, sleep abnormalities, depression, and cognitive disturbances that are increasingly recognized as being integral components of the disease [
7,
12]. Pathologically, PD is characterized by a selective loss of dopaminergic neurons in the substantia nigra (SN) and abnormal intracellular protein deposits called Lewy bodies (LB) that are mainly composed of fibrillary alpha-synuclein (αsyn) protein aggregates. Large bodies of evidence point to αsyn involvement in PD, since mutations and multiplications of the gene encoding αsyn have been linked to onset of familial forms of PD [
20,
40,
45] and aggregation of αsyn is causally linked to sporadic forms of the disease [
48]. In recent years these pathological hallmarks have also been observed in multiple system atrophy and dementia with LBs (DLB), which together with PD are referred to as αsynucleinopathies [
19,
47].
Modeling PD and its related disorders in animals to recreate specific pathogenic events and behavioral outcomes is a crucial step in basic research for mechanistic studies and therapeutic screening. The discovery of the involvement of αsyn in pathogenesis prompted the development of numerous animal models based on αsyn overexpression primarily through genetic manipulation methods in rodents [
28] and later using viral vector technology that can specifically target a region of interest [
22,
27]. To date, αsyn transgenic mice harboring various extents of pathological features of PD, decreased dopamine, and behavioral impairments but no significant dopaminergic cell loss [
14,
30] have been developed. In contrast, viral vector-mediated overexpression of αsyn models display a progressive pathology associated with clear dopaminergic neurodegeneration, replicating primary motor symptoms of PD [
50]. These models represent perhaps one of the most reproducible rodent models of PD, however, the targeted and local αsyn overexpression has its own limitations and these models do not possess non-motor symptoms observed in PD patients. αSyn overexpression studies have been useful in uncovering the relationship between αsyn protein expression and nigrostriatal neurodegeneration; however, PD is a complex syndrome not only associated with dopaminergic changes, and it is clear that multiple neurotransmitters and circuitries other than the basal ganglia are also affected [
2,
24]. Similarly, animal models must evolve and innovative strategies are needed to incorporate the new criteria of the neurological signs and symptoms of the disease.
Because duplication and triplication of the αsyn gene locus results in familial PD [
6,
45], there continues to be substantial interest in models that overexpress αsyn to model synucleinopathies. However, targeting a large brain region or even the whole rodent brain using viral vectors has remained challenging. Recently, it was demonstrated that gene transfer throughout the CNS (central nervous system) can be achieved via intracerebroventricular injection (ICV) of adeno-associated-virus (AAV) virus into the neonatal mouse brain [
39]. This method of viral vector delivery into the maturing brain allows a more efficient diffusion and infection compared to transduction of adult brain. Expression begins within days of injection and persists for the lifetime of the animal [
21]. This technique provides a fast and easy means of manipulating neuronal gene expression in vivo without complex surgical intervention, or time-consuming germline transgenesis. Moreover, it has proven to be effective for modeling other neurodegenerative disorders such as frontotemporal dementia [
9] and tauopathy [
10]. Both of these published models developed striking behavioral and neuropathological characteristics of their respective diseases in a short period of time.
Herein, we used somatic brain transgenesis aiming to engineer and characterize a novel synucleopathy model. To this end, we expressed human wild-type αsyn by injection of AAV serotype 2/1 into lateral ventricles of postnatal day 0 (P0) non-transgenic C57BL/6 mouse pups. Histology and behavioral analysis were conducted from 1 to 6 months of age. We were able to create a novel mouse model exhibiting widespread expression of αsyn as early as 1 month post injection. Interestingly, these animals displayed pathological forms of αsyn evidenced by the presence of phosphorylated αsyn and small aggregates resistant to mild proteinase K (PK) treatment in several brain regions. However, no progressive neurodegeneration and behavioral phenotype associated with the observed pathology indicate this model may represent a pre-symptomatic stage of synucleinopathy.
Materials and methods
Viral vector production
The viral vector construct rAAV- αsyn and rAAV-Venus were constructed as follows: The following expression components were inserted between two AAV2 inverted terminal repeats (ITRs): The SalI
-HindIII fragment of the pCAGGS vector (kindly provided by Mark Sands, University of Washington, St Louis) containing the hybrid CMV immediate-early enhancer/chicken β-actin promoter/exon1/intron and the poly (A) tail from rabbit beta-globin gene; full-length human wild-type αsyn cDNA(AAV-αsyn) [
32]; venusYFP cDNA (AAV-venus); and the woodchuck hepatitis virus post-transcriptional regulatory element (WPRE) (kindly provided by Dr. T. J. Hope, University of Illinois at Chicago, IL, USA). Briefly, adeno-associated virus (AAV) serotype 2/1 vectors expressing full length human αsyn or venus under the control of the CMV promoter were generated by plasmid triple transfection with helper plasmids in HEK293T cells 48 h later, cells were then harvested and lysed in the presence of 0.5% sodium deoxycholate and 50 U/ml Benzonase (Sigma-Aldrich, St. Louis, MO) by freeze-thawing, and the virus was isolated using a discontinuous iodixanol gradient. The genomic titer of each virus was determined by quantitative PCR.
Intracerebroventricular injections
All animal procedures were approved by the Mayo Clinic Institutional Animal Care and Use Committee and are in accordance with the NIH Guide for Care and Use of Laboratory Animals. Bilateral intracerebroventricular (ICV) injections were performed as previously described [
5] in C57BL/6 mouse pups on postnatal day 0. Briefly, Newborn C57BL/6 mice were cryoanesthetized and subsequently placed on a cold metal plate. A 30-gauge needle was used to pierce the skull just posterior to bregma and 2 mm lateral to the midline, and 2 μl of AAV (AAV-αsyn or AAV-venus) was injected into each cerebral ventricle (1.35E + 10gc/μl). Neonatal mice were kept with parent until weaned.
Mice were sacrificed at set time points as follow: 1 month (n = 8 per group), 3 months (n = 12–15 per group) and 6 months (n = 12–14 per group) post-injection. Of note, 3 and 6 month groups were behaviorally assessed before being euthanized for biochemical and histological analysis.
Behavioral analysis
At 3 and 6 months post injection mice underwent a battery of behavioral tests in collaboration with our mouse behavior core at Mayo Clinic Jacksonville. All mice were acclimated to the testing room for 1–2 h prior to testing. All behavioral equipment was cleaned with 30% ethanol prior to use with each animal. All mice were returned to their home cages and home room after each test. All behavioral tests were performed during the light cycle. To reduce experimental bias, all behavioral testing and scoring was done with the experimenter blind to the genotype of the mice.
Open-field assay
Mice were placed in the center of an open-field arena (40 W × 40 L × 30H cm) and allowed to roam freely for 15 min. An overhead camera was used to track movement with AnyMaze software (Stoelting Co., Wood Dale, IL), and mice were analyzed for multiple measures, including total distance traveled, average speed, time mobile, and distance traveled in an imaginary ‘center’ zone (20 × 20 cm).
Pole test
The pole test was performed according to Matsuura et al., 1997 [
31] with minor modifications. Animals were placed head-upward on the top of a vertical wooden pole (diameter 1 cm; height 50 cm). The animals orient themselves downward and descend the length of the pole. Each mouse received 2 days of training consisting of five trials for each session. On the test day, the total time until the mouse reached the floor with its four paws was recorded (T-total) as well as the time needed for the animal to turn downward (T-turn). The best performance over the five trials was used.
Rotarod test
Motor coordination and balance was measured using an automated rotarod system (Med Associates, Inc). Each mouse was placed on an accelerating spindle (4–40 rpm) for 5 min for three consecutive trials with at least 20 min of rest in between trials. The latency to fall time was recorded when the mouse fell off the spindle, triggering a sensor that automatically stops the timer located underneath the spindle. The animals were trained for 3 consecutive days and test on day 4.
Beam traversal
This protocol is based on Southwell et al., 2009 [
46]. Mice were trained for two sessions (1session/day) to walk across a beam toward their home cage (with gentle nudging, if necessary) until they were able to traverse the entire length of the beam unassisted. On the third day (test day), beam traversal was videotaped and each mouse had five trials. Videotapes were rated on slow motion by an experimenter blind to genotype and the total time taken to traverse the beam and the number of slip made during beam traversal were counted. The mean of the five trials was used as the score for each mouse.
Contextual and cued fear conditioning (CFC) test
This test was performed as previously described [
10]. Briefly, CFC was conducted in a sound attenuating chamber with a grid floor capable of delivering an electric shock, and freezing was measured with an overhead camera and FreezeFrame software (Actimetrics, Wilmette, IL). Mice were trained and tested on 2 consecutive days. On the training day, an 80-dB white noise served as the conditioned stimulus (CS) and was presented for 30 s. During the final 2 s of this noise, mice received a mild foot shock (0.5 mA), which served as the unconditioned stimulus (US). After 1 min, another CS–US pair was presented. The mouse was removed 30 s after the second CS–US pair and returned to its home cage. Twenty-four hours later, mice were tested by being returned to the conditioning chamber for 5 min without any shock, and freezing behavior was recorded. For the auditory CS test, environmental and contextual cues were changed by: wiping testing boxes with 30% isopropyl alcohol instead of 30% ethanol; replacing white house lights with red house lights; placing a colored plastic triangular insert in the chamber to alter its shape and spatial cues; covering the wire grid floor with opaque plastic and altering the smell in the chamber with vanilla extract. The animals were placed in the apparatus for 3 min, and then the auditory CS was presented and freezing was recorded for another 3 min (cued test). Baseline freezing behavior obtained during training was subtracted from the context or cued tests to control for animal variability.
Tissue processing
After behavioral analysis, all animals euthanized by deep anesthesia with sodium pentobarbital prior to transcardial perfusion with phosphate buffered saline (PBS). The brain was removed and bisected along the midline. Half brain was drop-fixed in 10% neutral buffered formalin (Fisher Scientific, Waltham, MA) overnight at 4 °C for histology, whereas the other half was frozen for biochemical studies. The half brain fixed in 10% formalin was embedded in paraffin wax, sectioned in a sagittal plane at 5 μm thickness and mounted on glass slides. The tissue sections were deparaffinized in xylene and rehydrated in a graded series of alcohols. Antigen retrieval was performed by steaming in distilled water for 30 min, and endogenous peroxidase activity was blocked by incubation in 0.03% hydrogen peroxide. Sections were then immunostained with anti αsyn (Covance, 4B12; BD biosciences, clone 42; Millipore, 5G4), αsyn pS129 (Wako, pSyn#64), glial fibrillar acidic protein (GFAP) (Biogenex, ARO20), ionized calcium-binding adaptor molecule 1 (iba1) (Wako, 019–19,741), and visualized using the Envision Plus system (DAKO, CA, USA). Slides were counterstained with hematoxylin, dehydrated in a graded series of alcohol and xylene, and coverslipped.To avoid undesired background staining the use of Vector MOM immunodetection KIT (Vector laboratories) was required for 4B12 and αsyn pS129 staining. According to the Kit instructions, 1 h of preincubation with blocking unspecific protein from MOM kit was followed by incubation with the primary antibody in the MOM protein concentrate at room temperature for 30 min. For double immunofluorescence sections were immunostained with primary antibody against αsyn (4B12) in combination with anti-tyrosine hydroxylase (TH; Thermos, OPA1–04050) overnight at 4 °C. For visualization fluorescent conjugated antibodies, Alexa 488-goat anti-mouse and Alexa 568-goat anti rabbit at 1:500 (Invitrogen) were used for 2 h at room temperature. For the detection of β-pleated sheets, some of the sections were incubated with 1% Thioflavin S (ThioS, Sigma) for 5 min, washed three times with 70% ethanol and two times with PBS. Sections were mounted with Vectashield mounting medium (Vector laboratories). Lastly, for proteinase K (PK) digestion, tissue sections were preincubated with proteinase K (DAKO) in PBS for 2 min at room temperature before performing regular immunostaining for αsyn using LB509 antibody (Thermo fisher).
Western blot
Frozen hemi brains were mechanically homogenized on ice in 10% (w/v) of cold lysis RIPA buffer (Millipore) containing protease inhibitor cocktail (Roche Diagnostics), and centrifuged at 100,000 x g for 20 min. The supernatant was saved as the soluble fraction. Triton X-100 was added to the pellet (final concentration 1%) and incubated for 20 min on ice followed by centrifugation (100,000 x g for 30 min). The supernatant was designated as Triton X-100 soluble fraction. The insoluble fraction was finally dissolved in lysis buffer containing 2% SDS (sodium dodecyl sulfate) and sonicated for 10s. Protein concentration of each lysate was determined by BCA. Equal amounts of soluble and insoluble fractions were analyzed by SDS protein electrophoresis and immunoblotted for total αsyn (BD biosciences, clone 42) and actin (Sigma, A5060). Immunoreactivity was visualized using an ECL chemiluminescent detection Kit (Thermo Fisher Sciences) and images were acquired with a CCD imaging system (LAS-4000, Fujifilm, Japan).
Taqman qRT-PCR analysis
Hemi brains were carefully dissected and snap-frozen on dry ice. Total RNA was extracted using TRIzol Reagent (Ambion Life Technology) followed by DNase RNA cleanup step using RNeasy (Qiagen, Germantown, MD). The quantity and quality of RNA samples were determined by the Agilent 2100 Bioanalyzer using an Agilent RNA 6000 Nano Chip. Complementary DNA (cDNA) synthesized from 500 ng of RNA with Applied Biosystems High-Capacity cDNA Archive Kit was used as a template for relative quantitative PCR using ABI Taqman chemistry (Applied Biosystems). mRNA expression was quantified using Hs00240906_m1 (human SNCA) and Mm01188700_m1 (mouse Snca). Mm00441941_m1 (tfrc, Transferrin receptor), Mm00497442 _m1 (txnl1, Thioredoxin-Related Protein 1) assays were used as endogenous controls for global normalization. Each sample was run in quadruple replicates on the QuantStudio 7 Real-Time PCR System (Thermo Fisher).
Image analysis
Brightfield images were captured using the Aperio slide scanner (Vista, CA, USA). Fluorescent images were taken with a 40 x Plan-Apochromat objective using a Zeiss AxioObserver equipped with an ApoTome Imaging System (Zeiss). Microglial and astroglial cell counts and morpholological analysis (process length and cell body size) were quantified using MetaMorph Image Analysis Software (Molecular Devices) with the neurite outgrowth application module [
4]. MetaMorph Software with the cell counting module was used to assess the burden of NeuN positive neurons. First, ImageScope® software (v12.1; Leica Biosystems) was used to annotate the cortex on mid-sagittal sections stained for NeuN for each mouse. Then, Positive Pixel Count Algorithm was established to recognize and quantify NeuN positive cells .The output parameter was the number of NeuN-positive neurons per given mm2 area annotated.
Statistics
Data were analyzed using GraphPad Prism 6 (San Diego, CA) and are presented as mean ± standard error of the mean (S.E.M.). Statistical significance was determined using a Student’s t-test or one-way analysis of variance with Tukey’s multiple comparison post-hoc. p < 0.05 was considered significant.
Discussion
Preclinical models to study αsyn-associated disease pathophysiology and to develop new therapeutic strategies are crucial tools in translational research. In the last two decades, viral vector gene delivery has offered great insights into the understanding of PD disease etiology, pathology, and molecular mechanisms. However, this method requires complex surgical intervention to locally deliver genes of interest. Herein, we generated a novel animal model, based on over-expression of human αsyn via neonatal injection, to express the transgene in the whole brain from birth to adulthood of mice. Our data demonstrate that following neonatal ICV injection, αsyn is distributed diffusely throughout the neuronal cytoplasm and also as puncta and inclusion bodies of varying size in several regions of the brain, from the most frontal part of the CNS (olfactory bulb) to the brainstem. The pathology observed in this model replicates a number of characteristics found in other in vivo models of αsyn-mediated toxicity [
18,
35,
42]. Most importantly, the presence of phosphorylated αsyn and small aggregates (PK resistant) in various brain regions clearly mimics what is observed in post mortem brain of PD patients. Interestingly, no locomotor dysfunction or cognitive deficits were observed at the latest end point of our study (6 months). Despite the lack of a behavioral phenotype this model may represent an attractive tool for understanding key mechanisms taking place in the presymptomatic phases of PD and synucleopathies.
Extensive diffusion of viral particles in the brain remains a challenge in vivo. Many AAV serotypes are available, each incorporating a different viral capsid protein and mediating different transduction characteristics within the brain [
38,
51]. In addition, the timing of AAV injection is a major factor that determines the overall biodistribution of the transgene. To overcome these issues, newer techniques have emerged and gene transfer throughout the CNS can now be achieved via ICV injection of AAV into the neonatal mouse brain [
21,
39]. Following ICV AAV administration it is hypothesized that the virus follows the flow of the CSF through the subarachnoid space. The absence of myelinated structures in the neonate brains allows for a better diffusion of the viral particles. Several groups have recently taken advantage of neonatal ICV AAV injections to develop novel, rapid, neurodegenerative disease animal models [
9,
10]. To our knowledge this is the first report of a successful transduction of human wild-type αsyn via somatic brain transgenesis. In the present study we used the serotype AAV2/1 at P0 (1–12 h postnatal) as described previously [
26]. In our hand this specific serotype results in widespread expression of αsyn but the timing of injection in this case is crucial; with the expression being limited if the injection is performed after P0. Chakrabarti and colleagues [
5] tested several serotypes at different time-points and reported that AAV2/8, and AAV2/9 have high biodistribution property independent of the timing of injection. Therefore a comparative study with the different serotypes may be of interest for this model. Lastly, it is worth mentioning that through our paradigm, AAV2/1 at P0, we achieved widespread transduction and pathology with a 2–3 fold increase level of total αsyn. This is very similar to level observed in well characterized transgenic αsyn mouse line, especially the line 61 [
8,
43]. Importantly, it reflects the level seen in patients with SNCA gene triplication of the Swedish-American kindred [
13].
Synucleinopathies including PD are a group of diseases linked by the abnormal accumulation of αsyn in various cells and brain regions, depending on the specific disease [
49]. The pattern of αsyn expression and accumulation in our neonatal AAV model resembles this aspect of PD pathology. Several other pathophysiologic characteristics could also be replicated in our AAV model. Indeed, at 3 months of age we observed a change in the solubility of αsyn and could even detect small aggregates that were proteinase-K resistant. The presence of insoluble and aggregated forms of αsyn correlates with LB pathology in post mortem brains and support their importance in any preclinical model. Also, αsyn aggregates are generally found as non-fibrillar, which is in line with our observations of aggregates that were not Thioflavin S positive. Accumulation of pS129 αsyn in PD brains, as well as in animal models, suggests that this post-translational modification plays an important role in the regulation of αsyn aggregation, LB formation, and neuronal degeneration [
37,
53]. Enrichment of pS129 αsyn is often observed in transgenic mouse brain [
35,
42]. In our model the detection of phosphorylated αsyn was pronounced in specific brain regions, suggesting that some neurons are more vulnerable to phosphorylation events than others. Interestingly, phospho-specific immunostaining is commonly associated with the nucleus of the cell. This could be of importance as nuclear localization may further aggravate neuronal toxicity [
44,
52]. In this model the presence of phosphorylated forms of αsyn do not appear to be sufficient to induce neurotoxicity as no cellular loss was observed, at least by 6 months of age. The exact implication of pS129 on αsyn aggregation and toxicity in vivo remains to be determined. Lastly, there is increasing evidence for the importance of neuroinflammation in PD pathogenesis. Our histological analysis reveals an increase of astroglial cells (GFAP) in the hippocampus of AAV-αsyn animals (Fig.
6a and
b) but no change in microglia profile or number. Again, this is consistent with what is has been described in PD post mortem brain, increased dystrophic astroglia cell density [
3,
11].
Traditionally, αsyn overexpressing mice develop pathology, loss of striatal dopamine, and present with locomotor dysfunction at various ages, but these cardinal features of PD are not often observed in the same lines and neurodegeneration is often small or non-existent [
27,
30]. Our AAV model did not develop locomotor dysfunction at an early age (3 and 6 months), despite the presence of αsyn accumulation in the dopaminergic system. Although this is in contrast with other mice over-expressing human αsyn [
15], we cannot rule out that behavioral changes may take longer to develop in this model (> 6 months) due to the neonatal administration paradigm. Also, αsyn levels are causative in PD pathogenesis and familial SNCA multiplication cases showed a dose-dependent correlation of αsyn load to the PD phenotype [
45]. Even though we observed 2–3 fold increase level of total αsyn in the whole brain, we did notice a variability of expression among the animals and different degree of transduction from region to region that may explain that in our model no abnormal behavior could be observed. Recently, new rodent models has emerged using αsyn overpexression in combination with other risk factors contributing to the disease. A dual exposure of αsyn, and the pesticide, rotenone, in rodent has been reported to to precipitate motor dysfunction and nigrostriatal neurodegeneration [
34]. Thus, using genetic component with environmental risk factors or other disease-causing factors may be another approach to provide relevant preclinical models that replicate motor dysfunctions.
Cognitive impairments have been rarely observed in other PD models, although less freezing behavior in the fear conditioning task was reported in the line 61 at 8 months of age [
41]. As with motor functions, we may anticipate learning disabilities at a later stage. Finally, the presence of early non-motor symptoms of PD were not evaluated in the present study however future studies addressing these symptoms in the neonatal AAV transduction model would be worth addressing in future study. It will be interesting to determine if impairments in gastrointestinal function, olfaction, or sleep cycle behavior can be recapitulated.
Conclusion
Taken together, we believe that this novel model may provide significant advantages over current transgenic models to investigate early pathogenesis. The ease of the gene delivery may offer a rapid and effective preclinical model. Although we used human wild-type αsyn in our study, the neonatal gene delivery approach described herein may be used to investigate mutated or modified forms of αsyn. Interestingly, somatic brain transgenesis may offer much more than a successful spread of the viral particles. Indeed, the possibility remains to inject two AAVs simultaneously to express multiple proteins in either overlapping or distinct neuronal populations without the constraints of complex genetic backgrounds. Furthermore one could induce αsyn overexpression in different background mice to study the interaction of αsyn and others PD related genes such as LRRK2 or even model other synucleopathies (ie DLB) given the fact that αsyn interacts with a numbers of other proteins related to neurodegeneration. For example, αsyn positive inclusions are described in tauopathies and vice versa, suggesting a co-existence of these proteinopathies [
33]. Therefore somatic brain transgenesis can serve as a tool to evaluate whether the pathogenicity of αsyn can be altered in the presence of an additional insult, such as tau protein.
In conclusion, this study paves the road to a new era of preclinical models in the field of synucleinopathies .and we believe that the present model is only a premise of the possibility that somatic brain transgenesis has to offer to elucidate pathological mechanisms in PD.