Background
Parkinson’s disease (PD) is the second most common neurodegenerative disorder. Neuropathologically, it is characterized by the progressive loss of dopaminergic neurons in the substantia nigra (SN) and the presence of proteinaceous intracellular inclusions called Lewy bodies (LBs) and Lewy neurites in the surviving neurons [
1]. Although the etiology of sporadic PD remains still unclear, the discovery of genes linked to familial forms of the disease has improved our understanding of the pathogenic mechanisms leading to PD.
Point mutations and multiplications of the α-synuclein (α-SYN) gene, SNCA, cause a rare familial autosomal dominant form of PD [
2]. α-SYN is a small protein of 140 amino acids that is widely expressed in the brain and localizes predominantly to presynaptic terminals [
3]. The biological function of α-SYN remains unknown, although it’s involvement in dopamine transmission and biosynthesis [
4,
5], synaptic plasticity [
6] and turnover of synaptic vesicles has been suggested [
7]. Under pathological conditions, including mutations and increased expression levels, α-SYN has the propensity to adopt a β-sheet rich conformation which leads to the formation of oligomers and fibrillar aggregates [
8]. This fibrillar form of α-SYN is the main protein component of LBs and Lewy neurites, indicating that α-SYN plays a crucial role in the pathogenesis of PD [
9]. Moreover, animal models based on overexpression of wild type (WT) or mutant α-SYN, recapitulate some of the main hallmarks of PD, including neurodegeneration, motor dysfunction and inclusion formation [
10].
Mutations in the parkin gene are the major cause of early-onset autosomal recessive PD [
11]. Parkin has been identified as an E3 ubiquitin-ligating enzyme, catalyzing the attachment of ubiquitin to substrate proteins, which consequently leads to their proteasomal degradation [
12-
14]. It was therefore suggested that the loss of parkin function due to disease-causing mutations might damage neurons through the accumulation of toxic proteins. On the other hand, evidence is accumulating that parkin ubiquitin ligase activity also contributes to non-degradative poly- and mono-ubiquitination, which are involved in alternative cellular pathways, including mitochondrial quality control [
15] and signal transduction cascade activation [
16,
17]. However, the molecular mechanism by which loss of parkin function leads to the development of PD is still not completely elucidated. More specifically, the role of α-SYN in parkin-associated PD remains unclear, although there are lines of evidence suggesting a potential link between parkin and α-SYN. On one hand it has been reported that the unmodified form of α-SYN does not interact with parkin [
18] and no accumulation of α-SYN was detected in the brain of parkin knockout (parkin
−/−) mice [
19-
21]. On the other hand, in about two thirds of the seventeen neuropathological reports of patients with parkin mutations published to date, no LBs were found, indicating that parkin might play a role in LB formation[
22]. Furthermore, overexpression of parkin provides protection against α-SYN toxicity in a variety of cellular and animal models [
23-
29]. Conflicting results were published in a number of studies using a transgenic approach to investigate the effect of parkin on α-SYN-induced neurotoxicity. The neurodegenerative phenotype was unexpectedly delayed in the absence of parkin in A30P α-SYN transgenic mice [
30], while no effect of loss of parkin on neuropathology was found in A53T α-SYN transgenic mice [
31]. In a third publication, an increase in damaged mitochondria in neurons of the SN and a reduction of complex I activity was reported in old double mutant mice generated by the crossing of parkin
−/− mice with double mutated A53T and A30P α-SYN overexpressing mice [
32]. However, these results in transgenic mice showed that the absence of parkin did not clearly affect α-SYN-induced neurodegeneration, except for some minor changes.
In the present study, we chose an alternative approach to investigate the effect of the absence of parkin on α-SYN-induced cell death in adult rodent brain using viral vector technology. Therefore, we overexpressed human WT α-SYN in the SN of parkin−/− and wild type (parkin+/+) mice with recombinant adeno-associated viral (rAAV) vectors. Subsequently, we analyzed the degree of neurodegeneration and synucleinopathy in the SN of those mice.
Discussion
This study was designed to investigate the effect of the loss of parkin on α-SYN induced neurotoxicity in rodent brain. We found that the absence of parkin did not alter the vulnerability of dopaminergic neurons to WT α-SYN induced neurodegeneration. However, the number of P-S129 α-SYN positive cells in the SN of parkin−/− mice was increased compared to parkin+/+ mice. This increase in the number of P-S129 α-SYN positive cells in the parkin−/− mice was not due to differences in expression level of α-SYN, since the total number of α-SYN-positive cells was similar in both groups, These results were reproduced in a second, independent experiment performed with a 4 times lower titer of rAAV2/7-WT α-SYN.
A number of previous studies already addressed the question if the absence of parkin affects the development of α-synucleinopathy using a transgenic strategy. However, inconsistent results were reported, since in one study no effect was found in A53T α-SYN transgenic mice [
31], whereas in a second study the loss of parkin unexpectedly mitigated the α-SYN phenotype in A30P α-SYN transgenic mice [
30]. In the present study, we opted for an alternative approach with viral vector-mediated overexpression of WT α-SYN in the SN of parkin
−/− and parkin
+/+ mice. rAAV vectors are an attractive tool for gene delivery in the brain, since they provide several advantages: specific brain regions can be targeted, the transduction efficiency in dopaminergic neurons is high, and a long-lasting and stable expression of the transgene at different doses can be achieved with a single delivery [
38]. In addition, in a previous study in our own research group, we showed that rAAV2/7-mediated WT α-SYN overexpression in mouse SN leads to a dose-dependent, progressive dopaminergic cell death [
33].
With our strategy, we did not observe a difference in sensitivity to WT α-SYN induced dopaminergic cell death between parkin
−/− and parkin
+/+ mice. The strength of our study is that it allowed us to specifically investigate the sensitivity of the dopaminergic cell population to α-SYN toxicity in the absence of parkin. Since the majority of α-SYN transgenic mice do not develop a robust dopaminergic phenotype, they are less suitable to study the effect of parkin deficiency and α-synucleinopathy in nigral dopaminergic neurons [
39,
40]. On the other hand, our finding that loss of parkin does not exacerbate dopaminergic degeneration is partly consistent with the results in the transgenic mice, since in those studies, no difference in dopaminergic cell survival was found between the α-SYN overexpressing mice and the parkin
−/− - α-SYN double transgenic mice, indicating that the complete absence of parkin does not affect dopaminergic cell survival [
30,
31]. A possible explanation for this somewhat unexpected observation might be the presence of compensatory mechanisms in the parkin
−/− mice, which may counterbalance for the total loss of parkin protein occurring already during embryonic development. Reports of an increased sensitivity of striatal metabotropic glutamate receptors and elevated levels of reduced glutathione in parkin
−/− mice indicate that such compensatory adaptations exist [
41,
42]. Locoregional downregulation of parkin with viral vectors or the generation of conditional parkin knockout animals may be a valuable strategy to overcome this issue. Indeed, it was recently reported that adult depletion of parkin in the SN of conditional parkin
−/− mice resulted in a progressive loss of dopaminergic neurons up to 40%, a phenotype that has never been observed in the constitutive parkin
−/− mice [
21,
43].
The lack of increased vulnerability of parkin−/− mice to WT α-SYN induced dopaminergic cell death was observed with two different doses of WT α-SYN. In the second experiment, we opted for a lower dose, because we reasoned that the dramatic dopaminergic cell loss achieved with the highest dose of WT α-SYN might hide small differences in sensitivity between parkin−/− and parkin+/+ mice. Although the lower dose of WT α-SYN resulted in a milder dopaminergic cell death, the degree of degeneration was still considerable (approximately 40% at 4 weeks). Therefore, we cannot exclude that subtle differences in sensitivity might emerge when using even lower overexpression of α-SYN or at later time points than analyzed in this study.
In a next step, we wondered whether the absence of parkin would influence the phosphorylation of α-SYN at serine residue 129, since the level of P-S129 α-SYN is highly elevated in the brains of PD patients [
34,
35]. We found that the number of nigral cells positive for P-S129 α-SYN was significantly higher in the parkin
−/− mice compared to parkin
+/+ mice. This was not the case for the dopaminergic terminals in the striatum. This discrepancy might suggest that mainly the non-dopaminergic neurons show an increased phosphorylation in the parkin
−/− mice or that the dopaminergic neurons with increased P-S129 α-SYN have relatively lower terminal densities. This increased phosphorylation is in apparent contradiction with the findings of Lo Bianco
et al. who reported an increase in P-S129 α-SYN positive aggregates after overexpression of parkin together with A30P α-SYN in the rat SN using lentiviral vectors [
25]. That observation was associated with a protective effect of parkin overexpression on A30P α-SYN induced dopaminergic degeneration. Furthermore, no increase in P-S129 α-SYN abundance was noticed in A30P α-SYN transgenic mice on a parkin
−/− background [
30]. On the contrary, and in agreement with our results, lentiviral vector-mediated co-expression of parkin with WT α-SYN in the striatum of rats reduced the levels of P-S129 α-SYN [
24]. A similar result was found in the striatum of macaque monkeys when parkin and WT α-SYN were overexpressed by means of rAAV vectors, although in this study a decrease in total α-SYN was also reported [
29]. In the study with the A53T α-SYN transgenic mice crossed with parkin
−/− mice phosphorylation of α-SYN has not been investigated [
31]. So far, we cannot clearly explain the inconsistencies between these reports, although the most obvious explanation is differences in experimental conditions. Altogether, the studies performed with WT α-SYN, including ours, point towards a correlation between decreased levels of parkin and increased phosphorylation of α-SYN at S129, whereas in the studies with mutant A30P α-SYN either the reverse or no effect is suggested. Thus, it is possible that the influence of parkin on S129-phosphorylation is different for WT α-SYN than for A30P α-SYN, since the two forms also differ in other properties including aggregation [
44] and membrane-binding [
45].
At this point, we can only speculate about the mechanism behind the increased phosphorylation of WT α-SYN in the absence of parkin. No direct binding of parkin to P-S129 α-SYN was observed in brain extracts of A30P α-SYN transgenic mice [
30]. We also quantified the percentage of ubiquitin and α-SYN double-positive cells but no difference was seen between parkin
+/+ and parkin
−/− mice (data not shown). This suggests that the increased phosphorylation of WT α-SYN is not a consequence of a difference in ubiquitination of α-SYN. It is possible that parkin modulates kinases and phosphatases regulating the phosphorylation of α-SYN at S129. In agreement with this hypothesis it was reported that overexpression of α-SYN in the brain of rats resulted in an increase in the level of Polo-Like-Kinase-2 (PLK2) and this increase was annihilated when parkin was co-expressed with α-SYN [
24]. Absence of parkin might then result in a more pronounced increase in PLK2 levels and therefore increased S129 phosphorylation of α-SYN, since PLK2 is known to phosphorylate α-SYN at S129 [
36]. We performed stainings for PLK2 on brain sections, but we failed to reliably detect endogenous expression levels of PLK2 (data not shown). Therefore we induced stable parkin knockdown in human SHSY5Y neuroblastoma cells overexpressing WT α-SYN. Interestingly, the increased S129 phosphorylation of α-SYN after parkin knockdown was replicated in this cell culture model, without effect on total α-SYN level. However, PLK2 protein levels were not altered after parkin knockdown in cell culture. Another potential mechanism involves protein phosphatase-2A (PP2A) that has been shown to dephosphorylate α-SYN at S129 [
37]. Indeed, the level of protein phosphatase-2A (PP2A) was reportedly increased when parkin was co-expressed with α-SYN compared to expression of α-SYN alone in rat striatal extracts [
24]. Absence of parkin might then decrease PP2A levels in these conditions, resulting in increased S129 phosphorylation of α-SYN. However, we failed to detect alterations in PP2A expression in the parkin knockdown cells. Thus, our data suggest that the mechanism behind the increased α-SYN phosphorylation might be independent from the PLK2 or PP2A pathway, although we cannot exclude changes in activity of either PLK2 or PP2A.
Furthermore, it has been demonstrated that 26S proteasomal activity is decreased in parkin
−/− mice and parkin null
Drosophila [
46]. Also proteasomal dysfunction and S129 phosphorylation of α-SYN have been linked before. First, two independent studies describe that proteasomal inhibition increases casein kinase 2 activity, another kinase mediating phosphorylation of α-SYN at S129 [
47], resulting in enhanced S129 phosphorylation of α-SYN [
48,
49]. Second, it was demonstrated that inhibition of the proteasome pathway resulted in the accumulation of P-S129 α-SYN without alteration in the total levels of α-SYN, suggesting that P-S129 α-SYN specifically undergoes degradation by the proteasome pathway [
50].
In the present study, the increased level of P-S129 α-SYN in the parkin
−/− mice was not associated with increased dopaminergic degeneration, which is intuitively in contradiction with the knowledge that ± 90% of α-SYN within LBs from PD patients is phosphorylated at S129 [
34,
35]. The toxicity of P-S129 α-SYN
in vivo has been studied extensively in the last years using mutated forms of α-SYN in which S129 is replaced either with an aspartate (S129D), to mimic phosphorylation or with an alanine (S129A), to block phosphorylation. However, conflicting results were reported. Expression of S129D α-SYN in
Drosophila resulted in an enhanced toxicity, whereas no dopaminergic cell loss was observed if S129A α-SYN was expressed [
51]. On the contrary,
Caenorhabditis elegans overexpressing S129A α-SYN showed severe motor dysfunction and synaptic abnormalities, unlike worms overexpressing S129D α-SYN that exhibited a nearly normal phenotype [
52]. Two studies using rAAV vectors in rats found expression of S129A α-SYN to be more toxic than S129D α-SYN [
53,
54]. In a third study, there was no difference between the two forms of α-SYN [
55]. In a recent study, S129D α-SYN expression in rat SN resulted in an accelerated striatal dopaminergic fiber loss and an earlier appearance of motor deficits compared to S129A α-SYN, although the nigral degeneration was similar [
56]. However, the S129D mutant might not completely mimick the constitutively phophorylated α-SYN. In another study phosphorylation of A53T α-SYN was induced by rAAV-mediated overexpression of G-protein-coupled receptor kinase 6 in rats, which resulted in accelerated A53T α-SYN induced neurodegeneration [
57], in contrast to our data. Species differences might be involved, since we observed that rats are more sensitive to rAAV-α-SYN-induced neurodegeneration compared to mice [
33,
58]. Altogether, the role of S129 phosphorylation of α-SYN in neurodegeneration still remains unclear.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
A-SVR designed, performed and analyzed all in vivo experiments and wrote the manuscript. MOS contributed to the design of the in vivo experiments and the optimization of the dose of rAAV2/7-WT α-SYN. AVdP cloned the rAAV transfer plasmids and assisted in the optimization of the viral vector production process. OC provided the parkin−/− and parkin+/+ mice, has been involved in the design of the study and revised critically the manuscript. CVdH participated in the design of the in vivo experiments, performed the cell culture experiments and has been involved in drafting the manuscript. VB conceived the study, participated in its design and helped to write the manuscript. All authors read and approved the final manuscript.