Removal of prolyl oligopeptidase reduces alpha-synuclein toxicity in cells and in vivo

Prolyl oligopeptidase (PREP) inhibition by small-molecule inhibitors can reduce alpha-synuclein (aSyn) aggregation, a key player in Parkinson’s disease pathology. However, the significance of PREP protein for aSyn aggregation and toxicity is not known. We studied this in vivo by using PREP knock-out mice with viral vector injections of aSyn and PREP. Animal behavior was studied by locomotor activity and cylinder tests, microdialysis and HPLC were used to analyze dopamine levels, and different aSyn forms and loss of dopaminergic neurons were studied by immunostainings. Additionally, PREP knock-out cells were used to characterize the impact of PREP and aSyn on autophagy, proteasomal system and aSyn secretion. PREP knock-out animals were nonresponsive to aSyn-induced unilateral toxicity but combination of PREP and aSyn injections increased aSyn toxicity. Phosphorylated p129, proteinase K resistant aSyn levels and tyrosine hydroxylase positive cells were decreased in aSyn and PREP injected knock-out animals. These changes were accompanied by altered dopamine metabolite levels. PREP knock-out cells showed reduced response to aSyn, while cells were restored to wild-type cell levels after PREP overexpression. Taken together, our data suggests that PREP can enhance aSyn toxicity in vivo.


Results
Locomotor activity in PREP ko animals is restored to the wt animal levels after PREP and aSyn viral vector co-injection. There was a statistically significant interaction between the aSyn and aSyn + PREP viral vector injections and time on total traveled distance in the PREPko animal groups ( Fig. 1A; F (5,75) = 4.174, p = 0.002, 2-way ANOVA). Traveled distance was decreased in the PREPko animal group that received aSyn + PREP viral injection at 5-week time point (F (1,15) = 5.612, p = 0.032, Univariate analyses) and viral vector effect extended until the end of the experiment at 13-week time point (F (1,15) = 7.642, p = 0.014). A similar effect was not observed in wt littermates ( Fig. 1A; F (5,70) = 1.002, p = 0.395, 2-way ANOVA). All animal groups exhibited decreased locomotor activity when compared to baseline (BL) levels from 5-week time point onwards (locomotor activity vs. BL; wt p = 0.001; PREPko animals p < 0.0005). In this experimental setting, we wanted to assess the effect of PREP on aSyn overexpression and therefore the green fluorescent protein (GFP) injected animal groups were deemed redundant. Additionally, it has been previously reported that aSyn can decrease locomotor activity comparatively to GFP viral vector injections 22 .
Scientific REPORTS | (2018) 8:1552 | DOI: 10.1038/s41598-018-19823-y staining was performed. Statistical differences were not observed in total aSyn oligomer specific staining using stereological investigation (Fig. 2B). We have previously reported that PREP inhibition by a small-molecule compound can reduce soluble oligomer amounts 15,17,18 but the same observation did not hold true when PREP was knocked down. Proteinase K (PK) treatment was performed followed by aSyn oligomer specific staining and stereology. Statistical differences were observed between phenotype and viral vector injections ( Fig. 2C; F (1, 16) = 6.413, p = 0.022, 2-way ANOVA). Interestingly, aSyn + PREP injected PREPko mice had less PK resistant aSyn oligomers than aSyn injected PREPko animals, while the effect was not observed in wt animals (Fig. 2C). Visually, aSyn + PREP injected PREPko animals had lighter and more diffuse PK resistant aSyn oligomers in Figure 2. TH+ cells, aSyn oligomer particles and of p129-aSyn OD measurements showed alteration in PREPko animals. (A) Significant TH + cell decrease was observed between aSyn injected wt and PREPko animal groups and aSyn+ PREP injected PREPko animal group (n = 6-7). (B) Oligomer specific particle stereological counts were not different between the animal groups (n = 7-8) while (C), PK resistant aSyn oligomer stereology showed a statistically decreased number of PK resistant oligomers in the PREPko aSyn + PREP animal group (n = 4-6). (D) p129-aSyn phosphostaining showed a significant interaction and decreased immunostained area in the aSyn + PREP injected PREPko group (n = 5). (E,F) Ratio between PK resistant oligomers or p129-aSyn stained particles and total aSyn oligomer count showed reduction in aSyn + PREP PREPko animal group (n = 4-6). (G) Representative images of aSyn and PK resistant oligomers and p129-aSyn staining in SN brain sections. Particularly in PREPko animals with aSyn + PREP injection, more diffuse staining in total aSyn, PK resistant aSyn oligomers and p129-aSyn is seen. Scale bar 5 μm. Bars represent mean ± SEM, 2-way ANOVA.
Scientific REPORTS | (2018) 8:1552 | DOI:10.1038/s41598-018-19823-y stereo-investigated SN brain sections (Fig. 2G). The ratio between the number of PK resistant aSyn particles and total amount of aSyn oligomer specific particles was matched for individual animals. Statistical interaction was observed between viral vectors and animal phenotype ( Fig. 2E; F (1,16) = 5.847, p = 0.028, 2-way ANOVA). PK resistant and total aSyn oligomer ratio in the PREPko animal group with aSyn + PREP injection was decreased while the opposite effect was seen between wt animal groups. Additionally, differences were not seen after optical density (OD) analyses of total aSyn IHC staining in striatum, substantia nigra pars compacta (SNpc) or substantia nigra pars reticulate (SNpr) (Fig. 3A-C,G). aSyn phosphorylation at serine 129 in SN. aSyn phospho S129 (p129-aSyn) was quantified to assess the distribution pattern differences between the phenotypes, additionally p129-aSyn particles were correlated with the total aSyn oligomer numbers in SN. We observed an interaction between the phenotype and viral vector injections for p129-aSyn phosphorylation (Fig. 2D,G; F (1,16) = 8.657, p = 0.0096, 2-way ANOVA). The PREPko animal group with aSyn + PREP injections had the smallest area of the p129-aSyn staining (Fig. 2D) and it was visually more diffuse (Fig. 2G). The ratio between the number of particles and total amount of aSyn oligomer specific particles was matched for individual animals. No differences were observed between the groups (Fig. 2F). aSyn viral vector causes mild loss of nigrostriatal tyrosine hydroxylase. To study whether overexpression of aSyn + PREP affects DAergic neuron loss in mice differently from single aSyn microinjection, we performed stereological tyrosine hydroxylase positive (TH+) neuron quantification in SNpc and OD for TH+ fibers in SNpc, SNpr and striatum by IHC. The interaction effect between viral vectors and animal phenotypes was not statistically significant. Therefore, an analysis of the main effect for viral vector injection was performed, which indicated that the main effect between aSyn and aSyn+ PREP injection conditions was statistically significant ( Fig. 2A; F (1,23) = 7.965, p = 0.0097, 2-way ANOVA). TH+ cell loss was more pronounced in both animal groups that were injected with aSyn+ PREP. Our previous data has showed that PREP inhibition is partially protective against aSyn-caused toxicity 18 . Tyrosine hydroxylase (TH) OD analyses did not show a clear loss of TH+ fibers in either the striatum or the SNpc ( Fig. 3D-E,G). However, in SNpr main effects for viral vector injection  Fig. S4) in the striatal tissue were similar in all groups. Striatal DA was significantly decreased compared to the intact side of the brain in all other groups (F (7,56) = 9.403, wt-aSyn p = 0.0018, wt-aSyn + PREP p = 0.0082, PREPko-aSyn p = 0.0071, 1-way ANOVA with Tukey's post hoc comparison) except in PREPko mice that received aSyn + PREP injection (F (7,56) = 9.403, p = 0.2176), but DOPAC was increased significantly only in this group (F (7,56)  Changes in autophagy markers were measured from the soluble fraction ( Fig. 7) to establish if redistribution of aSyn and p129-aSyn is mediated via an autophagy pathway. Levels of p62, a protein accumulation marker, were significantly elevated by stress ( Fig. 7B; F (1,30) = 5.117, p = 0.0311, 2-way ANOVA) and transfections ( Fig. 7B; F (1, 30) = 5.203, p = 0.0115) in PREPko cells. However, the BL levels of p62 were approximately 50% lower in PREPko cells compared to HEK-293 cells (Fig. 7B). aSyn caused an increased accumulation of p62 compared to control treatment while the level of p62 was even greater than in the aforementioned groups, both non-stressed and stressed groups (Fig. 7B). Changes in beclin1 levels were not seen in either HEK-293 or PREPko groups (Fig. 7A). LC3BII levels were only mildly affected by aSyn and aSyn + PREP transfections in HEK-293 cells (     We did not observe changes in catalase activity between groups (Fig. 9A,B,E) while thioredoxin levels were not analyzed as marker was visible only in the HEK-293 oxidative stress group, indicating that PREPko cells do not upregulate enzyme in particular oxidative stress condition (Fig. 9E).

Proteasomal activity alterations in PREPko cells. Proteasomal activity was measured in HEK-293 and
PREPko cells in the presence and absence of oxidative stress. In HEK-293 cells, both oxidative stress and plasmids transfections reduced the proteasomal activity ( Fig. 10A; F (4,36) = 16.58, p < 0.0001, 2-way ANOVA with Tukey's post hoc comparison). Non-stressed control group's S20 proteasomal activity was statistically higher compared to the rest of the groups (see Fig. 10A). The oxidative stress conditioned aSyn + PREP group had the most decreased proteasomal S20 activity in HEK-293 cells (Fig. 10A), where it was lower compared to the rest of the oxidative stress groups (vs. control group (p = 0.0029) and GFP group (p = 0.008); vs. aSyn group (p = 0.013)). While in non-stressed conditions, the aSyn and aSyn + PREP group proteasomal activity was not statistically different (Fig. 10A). An important observation is the statistical difference between HEK-293 cell aSyn and PREP transfected group in non-stressed conditions (p = 0.037) while a difference is not observed with oxidative conditions. PREPko cells have decreased proteasomal activity compared to the HEK-293 cells at the BL level (p < 0.0005) while differences between cell types were not observed in the presence of oxidative stress ( Fig. 10C; F (1,28) = 12.842, p < 0.001, 2-way ANOVA), which could be explained partially due to increased autophagic flux in PREPko cells (See Fig. 10D). Nevertheless, similar to HEK-293 cells, PREPko cells had a significant 2-way interaction between stress and plasmid transfection conditions ( Fig. 10B; F (4,36) = 5.480, p = 0.0015, 2-way ANOVA with Tukey's post hoc comparison). When post hoc comparison was done for PREPko cell 20S proteasomal activity, the non-stressed control group had statistically higher 20S proteasomal activity compared to all the groups except the PREP transfected non-stressed and oxidative stress conditions (Fig. 10B). The aSyn + PREP oxidative stress groups were statistically different only to non-stressed control group and PREP transfected group (p < 0.0005) and oxidative stress control (p = 0.005) and PREP transfected group (p < 0.0005). Moreover, only PREP transfected groups did not show statistical differences to both control groups (Fig. 10B). When aSyn and PREP transfected groups were compared, a statistical difference was observed between the non-stressed aSyn and PREP groups (p = 0.015) and oxidative stress aSyn and PREP group (p = 0.0001), that indicated that in PREPko cells overexpression of PREP alone does not significantly decrease 20S proteasomal activity.
aSyn levels in cell medium are increased in PREPko cells transfected with aSyn plasmid. We measured aSyn levels in cell medium after we noticed that some of the data from WB showed an apparent decrease of total aSyn in membrane bound and insoluble fractions. HEK-293 and PREPko cells were used in non-stressed and oxidative stressed conditions to measure amount of aSyn in cell medium. There was a significant increase in aSyn transfected PREPko cells after oxidative stress compared to HEK-293 cells ( Fig. 11; p < 0.0005). PREPko cells showed a significant oxidative stress-related increase of secreted aSyn between aSyn transfected PREPko cells (p < 0.0005) and aSyn + PREP transfected PREPko cells (Fig. 11; p = 0.021). A significant interaction effect was observed between cell phenotype, stress, and plasmid transfection conditions ( Fig. 11; There was a statistically significant simple 2-way interaction between stress conditions and cell phenotype for aSyn treated groups (F (1,16) = 27.577, p < 0.0005). Moreover, there was a statistically significant simple two-way interaction between cell phenotype and transfection conditions for oxidative stress groups (F (1,16) = 20.755, p < 0.0005) but not for non-stressed groups.
Additionally, a cut off column with 30-kDa exclusion size was used to measure the monomeric and dimeric aSyn in the cell medium. However, the monomeric and dimeric aSyn levels were too low in the medium to be detected with ELISA and this proposes that most of the secreted aSyn in cell medium was higher order aSyn forms.

Discussion
In this study, our aim was to clarify the significance of PREP for aSyn aggregation and toxicity in vivo and reveal the mechanisms by using cell culture models. We showed that PREPko animals were not sensitive for overexpression of aSyn by using unilateral viral vector injections in the nigrostriatal pathway and restoring PREP together with aSyn significantly reduced animal locomotor activity. Moreover, biochemical investigation revealed that aSyn modifications and TH+ cell loss in the PREPko animal phenotype injected with aSyn + PREP viral vectors was the most pronounced. Cell experiments corroborated the in vivo data since combination of aSyn + PREP transfection in PREPko cells was found to be the most toxic. Moreover, our cellular data explained in part the mechanisms behind the PREP mediated aSyn toxicity, notably the changes in aSyn clearance and phosphorylation.
PREPko animals that received aSyn + PREP injection exhibited significant changes in their locomotor behavior. Our group had previously tested the effects of PREP protein restoration in PREPko mice but significant changes were not observed 11 . Initial PREP protein detectability in striatum 11 coincides with the period when the first divergence in the motor behavior between PREPko mouse groups appear. Consequently, it indicates that locomotor activity decrease is due to the synergistic toxic effect of aSyn and PREP overexpression rather than rescue of the PREPko animal phenotype. Additionally, only wt animals exhibited changes in the paw preference, similar to our previous study showing that the unilateral aSyn viral vector overexpression produces changes in paw preference as early as 2 weeks post-injection 18 . Interestingly, PREPko animals did not exhibit unilateral paw misbalance but it could be attributed to the increased extracellular levels of DA in the naïve PREPko animal striatum as we have shown in our earlier study 11 .
Overexpression of aSyn decreased striatal tissue concentration of DA to an average of 60-70% of the concentration in the intact side of the brain, however striatal DA was decreased less in aSyn + PREP injected PREPko mice. Restoring PREP to PREPko mice elevated striatal DA level in our previous study 11 , and the same effect can be seen to some extent in this study. Overexpression of aSyn induced a decrease in striatal DA as reported  [22][23][24][25][26] . In our previous study 11 , naïve PREPko mice had a higher extracellular level of DA than wt littermates but in this study, we did not find this difference between the injected hemispheres. aSyn regulates extracellular DA under normal conditions by stabilizing DAT on the plasma membrane 27 and our previous study showed that PREPko mice have impaired DAT function and more internalized DAT 11 . Moreover, aSyn can alter the synaptic DA vesicle genesis and recycling that leads to decreased release of DA 19 . This indicates that the overexpression of aSyn regulates nigrostriatal DA and DAT function strongly enough to cover the effect of absence or restoration of PREP. Extracellular and tissue concentration of DA did not correlate with the behavioral changes most likely because both aSyn and PREP regulate the nigrostriatal DAergic system. Elevated DOPAC in the striatum of aSyn + PREP injected PREPko mice can indicate altered DA metabolism by monoamine oxidase and increase in more toxic metabolite DOPAL. Increased DOPAC and DOPAL could possibly increase oxidative stress, impair synaptic vesicle function and prevent aSyn fibrillation by stabilizing aSyn oligomers 28,29 .
PREPko mice that received aSyn + PREP injection had lower levels of phosphorylated aSyn and the particles were more diffuse. There are reports that aSyn phosphorylation, especially p129-aSyn modification, is decreasing neurotoxicity in the drosophila 30 and rat models of PD 31 . Besides, p129-aSyn modification has been shown to facilitate degradation of aSyn via autophagic and proteasomal pathways 32 . The decreased levels of p129-aSyn that we observed in aSyn + PREP injected PREPko animals could be due to post-translational modifications that reduce aSyn degradation, especially as p129-aSyn modification is shown to decrease aSyn half-life 32 . Additionally, aSyn + PREP injected PREPko animals had a lower amount of PK resistant aSyn oligomers. It has been shown that PK resistant oligomers and Lewy body structures are part of the cell's coping mechanism that reduces cell exposure to the toxic aSyn species [33][34][35] . Besides, SNpc neurons that do not contain Lewy bodies could represent a neuronal survival strategy 36,37 . Nevertheless, TH+ stereology in SNpc indicated a slightly higher cell loss in the aSyn + PREP injected animal groups. All animal groups apart from the PREPko animals injected with aSyn + PREP showed clear aSyn inclusion bodies. PREP has been shown to increase aSyn dimerization in cell free conditions and via direct protein-protein interaction 16 while in mouse models, inhibition of PREP activity reduced aSyn amount in a transgenic mouse strain 17 and in the viral vector overexpression model of PD 18 . However, some of the observed IHC differences between animal phenotypes could be due to the PREPko animal signaling pathway alterations 11 . The ability of the PREP protein to increase aSyn dimerization could lead to a faster accumulation of soluble, more toxic aSyn species 33 that could not be detected with stereology.
Similar to in vivo results, PREPko cells were more resistant to the toxicity of aSyn plasmid transfection and oxidative stress, and in both cell lines, the combination of aSyn + PREP transfection was the most cytotoxic. Although oxidative stress is commonly used to initiate aSyn aggregation, we did not see a significant elevation of Triton X-100 or sodium dodecyl sulfate (SDS) soluble aSyn particles after oxidative stress apart from membrane-bound S129 phosphorylated aSyn, but aSyn particles were decreased after stress, and this was seen both in HEK-293 and PREPko cells and after aSyn + PREP transfections as well. Surprisingly, when we studied the oxidative stress response in cells, we found that ROS levels and ROS response protein SOD1 and thioredoxin levels were elevated in the HEK-293 cells after oxidative stress induction but no alterations were seen in PREPko cells. Catalase levels in HEK-293 and PREPko cells were similar, indicating that PREP role in some of the oxidative stress response is likely upstream of the hydrogen peroxide (H 2 O 2 ) reduction 38 . Additionally, the PREPko cell apparent lack of SOD1 upregulation could indicate that these cells does not produce as much superoxide 39 or that PREPko cells would reduce H 2 O 2 via thioredoxin system 40 . Another possibility is that deletion of PREP could  41 . However, further studies are required to establish mechanisms that govern the cell survival upon PREP deletion.
We found that cells with aSyn + PREP plasmid transfection inhibited the most proteasomal activity that is indicative of the decreased aSyn protein degradation 42,43 , and PREP inhibition has been shown to attenuate toxicity of proteasomal inhibition in aSyn overexpressing cells 44 . In this study, PREPko cells had lower basal 20S proteasomal activity in comparison to the control cells. However, PREPko cell proteasomal activity changes in the presence of oxidative stress and protein overexpression were lower. Some evidence suggests that low-level proteasomal inhibition could have a beneficial effect on neuroprotection 37,42,43,45 . The proteasomal system is the main aSyn degradation pathway during basal conditions but with increased stress the autophagic pathway is recruited 43,46 . PREP is thought to act as a negative autophagy regulator 17 , and we saw an increased basal autophagic flux in PREPko cells. After co-transfection of aSyn and PREP, PREPko cells showed decreased proteasomal activity and as PREP is known to inhibit autophagic flux, it adds a subsequent strain on the autophagic-lysosome systems ability to cope with the aSyn overexpression.
Since we saw reduced levels of different aSyn forms after oxidative stress, we wanted to measure the aSyn secretion. PREPko cells with aSyn transfection had significantly increased extracellular aSyn levels and showed the largest difference between stressed and non-stressed conditions in SDS soluble fractions. PREP restoration reduced aSyn level in cell medium to the levels of control cells. It seems that PREPko cells are able to remove excessive aSyn by facilitating cytosolic aSyn transport. Despite that, PREP's role in vesicular trafficking has not been shown apart from a report that PREP could be involved in axonal transport by an unknown mechanism 47 . It has been shown that aSyn can be secreted in the cell medium via a non-conventional vesicular pathway 48 and exosomes 49,50 . Moreover, aSyn can be released in the extracellular space upon autophagic failure 51 and lysosomal dysfunction 52 . Although extracellular aSyn can be potentially toxic, our results indicate that PREPko cells do not exhibit a strong cytotoxic response even with high amounts of aSyn in the extracellular space although most of the aSyn in cell medium were higher order aggregates.
Studies with PREP have shown beneficial effects on aSyn aggregation and clearance after PREP inhibitor treatment [16][17][18]44,53 . However, information about aSyn toxicity in the total absence of PREP has not been reported. We were able to show that co-overexpression of aSyn and PREP in PREPko animals and cells increases toxicity and ablates the ability of the proteasomal systems to process aSyn. Even though we could not identify the toxic aSyn species that presence of PREP causes, PREPko cells had several mechanisms to cope with the increased aSyn load, such as increased autophagy and increased aSyn oligomer transport into the cell medium. Taken together, although aSyn forms aggregates in the absence of PREP, our findings showed that PREP is important for aSyn-mediated toxicity and this further emphasizes the possibilities of PREP inhibitors as disease-modifying drugs for PD and other synucleinopathies.

Materials and Methods
Reagents. Reagents were purchased from Sigma-Aldrich if not otherwise specified. Ethanol was purchased from Altia (Helsinki, Finland). Adeno-associated virus (AAV) driven by chicken β-actin promoter (CBA) was acquired from the MichaelJ. Fox Foundation. AAV2-CBA-α-synuclein (AAV-aSyn; 1.5 × 10 13 vg/mL) viral vectors were constructed, produced, and titered by Vector Core at the University of North Carolina (Chapel Hill, USA). AAV1-EF1α-PREP was obtained from the National Institute of Drug Abuse (Dr. Brandon Harvey, Intramural Research Program, Baltimore MD, USA). Plasmids, pAAV1-EF1α-PREP (Addgene #59967), AAV1-EF1α-GFP (Addgene #60058) and AAV1-EF1a-V5-synuclein (Addgene #60057) were obtained from National Institute of Drug Abuse (Dr. Brandon Harvey, Intramural Research Program, Baltimore, MD, USA). Plasmid construction was described in 17 while PREP viral vector characterization can be found in 11 . Animals. PREPko mice (Deltagene Inc, CA, USA) and wt littermates were back crossed in C57BL/6JRccHsd genetic background (Envigo, The Netherlands; 5-10 back crossings). Generation of PREPko mice has been described in 54 , behavioral phenotyping was done by 13 . Mice (7-9 weeks old; Envigo, The Netherlands) were housed under standard laboratory conditions. The experiments were carried out according to the European Communities Council Directive 86/609/EEC and were approved by the Finnish National Animal Experiment Board.
Surgical procedures. Mice were anesthetized with isoflurane (4% induction, 1.5-2.0% maintenance) and the recombinant AAV vectors were injected above mouse SN in a stereotaxic operation. AAV vectors were tested earlier 11,18 . To target the SN, viral vectors were given as single injection (volume 1 µL or 2 µL co-injection, rate 0.2 µL/min) into the left hemisphere, 3.1 mm anterior and 1.2 mm lateral to bregma, and 4.2 mm below the dura 55 .
Availability of materials and data. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Behavioral experiments
Cylinder test. Cylinder test (height 15 cm, diameter 12 cm) was used to measure motor asymmetry in spontaneous forelimb preference after unilateral microinjections as described in 18 . Animals were grouped by their BL paw preference prior to the surgery. No habituation of the animals to the testing cylinder was allowed before video recording. GA, USA) and activity was recorded at the same time of the day for all of the animals. Total distance (locomotor activity) and vertical activity were analyzed. Data was collected in 5 min intervals and activity was recorded for 120 min (light intensity 150 lx). Data was analyzed with Activity Monitor v 7.06 software.
Tissue processing. At 13 weeks post-injection, mice intended for IHC analysis were deeply anesthetized with sodium pentobarbital (150 mg/kg) and transcardially perfused as described in 18 . Frozen brain sections were sectioned as 30 µm free-floating sections on a cryostat (Leica CM3050, Wetzlar, Germany) and kept in a cryoprotectant solution.
Proteinase K treatment. PK protocol to remove soluble proteins in sections was performed as described in 18 . Shortly, sections were digested with 10 μg/ml PK (#V3021, Promega, Madison, USA) in TBS-T for 10 min at 55 °C. The sections were post-fixed with 4% PFA for 10 min and processed for aSyn oligomer specific immunostaining with the same primary and secondary antibody concentrations as described above.
Microscopy and stereology. OD of TH, aSyn from ipsilateral and contralateral striatum and SN were determined. Digital images were scanned at 40× magnification with Pannoramic Flash II Scanner (Version 1.15.3, RRID:SCR_014424, 3DHISTECH) and three coronal sections from each mouse were processed for further analyses with Pannoramic Viewer (Version 1.15.3, 3DHISTECH). Images were converted to grayscale and inverted, line analyses tools for striatum or freehand for SN in ImageJ (1.48b; RRID:SCR_003070, NIH, USA) were used to measure the OD of immunoreactivity.
p129-aSyn immunohistochemical sections were imaged and average particle area and numbers per section were quantified using Image-Pro Plus software (Media Cybernetics, Inc., Rockville, MD, USA), four representative SN sections per brain were used.
The number of TH+ cells in SNpc was estimated using the optical fractionator method in combination with the dissector principle and unbiased counting rules 57 . The SNpc was analyzed with a Stereo Investigator platform (MicroBright-Field, RRID:SCR_002526, Magdeburg, Germany) attached to an Olympus BX51 microscope (Olympus Optical, Tokyo, Japan) as described previously 18 . From each animal, three sections from the central portion of the SNpc were selected for quantitative analysis. Grid size was 100 × 80 µm and the counting frames were 60 × 60 µm. The coefficient of error was between 0.05 and 0.10 58 .
The number of aSyn oligomer specific particles in SN was estimated using the optical fractionator method in combination with the dissector principle and unbiased counting rules 57 . From each animal, four representative sections from SN were selected for quantitative analysis as described previously 18 . Grid size was 120 × 120 µm and the counting frames were 60 × 60 µm large. The average coefficient of error for each region was in range of 0.05 to 0.1 58  No-net-flux microdialysis. Extracellular concentration of striatal DA was measured by no-net-flux microdialysis that was performed with PREPko mice and wt littermates 14-15 weeks after the injection of viral vectors as described in 11 . After the microdialysis experiment, the brains were removed and processed for tissue HPLC analysis.
HPLC tissue analysis. Striatal tissue samples were punched below corpus callosum +0.74 mm from bregma to 2 mm depth by using sample corer (i.d. of 2 mm) with a plunger (Stoelting Co, Wood Dale, IL, USA) on a cryostat (Leica CM3050). Tissue processing was done as earlier described in 59 . The concentration of DA, its metabolites, DOPAC and HVA, 5-HT, its metabolite 5-HIAA and GABA in the tissue samples of striatum were analyzed with an HPLC as earlier described in 18 . Concentrations were calculated as nanograms per milligram of brain tissue.
Induction of aSyn aggregation. For WB, cells were seeded in a 6-well plate with the density 200,000 cells/ well or 400,000 cells/well for oxidative stress conditions and allowed to attach overnight. Thereafter, the cells were transfected with aSyn, GFP, aSyn + GFP or aSyn + PREP. Lipofectamine 3000 (L3000015; Thermo Fisher Scientific) was used as a control. Non-stressed cells were grown for 72 hours. In oxidative stress groups 24 hours after plasmid transfection, the aggregation process of aSyn was induced by adding 100 mM H 2 O 2 and 10 mM FeCl 2 in cell culturing medium, adapted from 15 . Cell lysis and fractionation were performed as described below.
Cell viability assay. Cells were plated with the density of 10,000 cell/well in 96-well plate, the next day transfected with aSyn, GFP and aSyn + PREP and thereafter incubated for 24 h. Lipofectamine 3000 was used as a control. The cells were exposed to oxidative stress for 48 hr as described above (See Materials and Methods). To assess cell viability standard MTT test was performed as previously described in 61 .

Reactive oxygen species detection. ROS in cells were measured using DCFDA Cellular ROS Detection
Assay Kit (ab113851, AbCam) after induction of oxidative stress as described above. After 48 hours, DCFDA treatment was performed according to the manufacturer's instructions. Fluorescence signal was adjusted to the total protein amount.
Proteasome activity assay. For determining chymotrypsin-like 20S proteasomal activity, the protocol based on Suc-Leu-Leu-Val-Tyr-AMC (#I-1395, Bachem, Bubendorf, Switzerland) substrate was used as described in 17 . In brief, cells were lysed in buffer for 20S activity. After 1 hr incubation at 37 °C with substrate, fluorescence was read at 355/460 nm with Victor 2 well-plate reader (PerkinElmer, Waltham, USA). Proteolytic activity was expressed as the amount of free AMC/min * mg protein.
Scientific REPORTS | (2018) 8:1552 | DOI:10.1038/s41598-018-19823-y Autophagic Flux measurements. To assess the autophagic flux PREPko and HEK-293 cells were treated with 10 and 50 nM concentration of bafilomycin A1 for 4 hrs. Dimethyl sulfoxide (DMSO) served as a control. Cells were lysed in ice cold modified RIPA buffer (50 mM Tris HCl pH 7.4, 1% NP-40, 0.25% sodium deoxycholate, 150 mM NaCl) and WB for LC3B was performed as describe above (see Materials and Methods). aSyn ELISA from cell culture supernatant. aSyn levels in cell culture medium of PREPko and HEK-293 cells were measured after cell transfection either with aSyn or aSyn + PREP plasmids and induction of aSyn aggregation as described above (see Materials and Methods). Human aSyn ELISA Kit (ab210973; AbCam) was used according to the manufacturer's instruction. Fluorescence was red at 450 nm with Victor 2 well-plate reader (PerkinElmer). aSyn amount in medium was adjusted to the cell protein concentration in cell lysates, pmol/ mg*protein.
Statistical Analysis. Statistical analyses were performed using either GraphPad Prism (version 6.07, GraphPad Software, Inc., San Diego, USA) or SPSS Statistics (Version 22.0.0.1 IBM Corporation, Armonk, USA) tools. Statistical tests that were used were Student's t-test and two-way ANOVA with Bonferroni's post hoc comparison for behavioral assessment and one-way, two-way and three-way ANOVA with Tukey's post hoc comparison for in vitro and cell data. Data are presented as mean ± SEM. Statistically significant differences were considered at p < 0.05.