Prolyl oligopeptidase inhibition reduces alpha‐synuclein aggregation in a cellular model of multiple system atrophy

Abstract Multiple system atrophy (MSA) is a fatal neurodegenerative disease where the histopathological hallmark is glial cytoplasmic inclusions in oligodendrocytes, rich of aggregated alpha‐synuclein (aSyn). Therefore, therapies targeting aSyn aggregation and toxicity have been studied as a possible disease‐modifying therapy for MSA. Our earlier studies show that inhibition of prolyl oligopeptidase (PREP) with KYP‐2047 reduces aSyn aggregates in several models. Here, we tested the effects of KYP‐2047 on a MSA cellular models, using rat OLN‐AS7 and human MO3.13 oligodendrocyte cells. As translocation of p25α to cell cytosol has been identified as an inducer of aSyn aggregation in MSA models, the cells were transiently transfected with p25α. Similar to earlier studies, p25α increased aSyn phosphorylation and aggregation, and caused tubulin retraction and impaired autophagy in OLN‐AS7 cells. In both cellular models, p25α transfection increased significantly aSyn mRNA levels and also increased the levels of inactive protein phosphatase 2A (PP2A). However, aSyn or p25α did not cause any cellular death in MO3.13 cells, questioning their use as a MSA model. Simultaneous administration of 10 µM KYP‐2047 improved cell viability, decreased insoluble phosphorylated aSyn and normalized autophagy in OLN‐AS7 cells but similar impact was not seen in MO3.13 cells.


| INTRODUC TI ON
Multiple system atrophy (MSA) is a sporadic, adult-onset and progressive neurodegenerative disease with a prevalence of 0.1-3/100,000 persons/year that increases with ageing. 1 Since current therapies are only symptomatic, MSA is a fatal disease with survival duration of 6-9 years after the onset of symptoms. According to the predominant motor disorder, MSA is categorized as MSA-P or MSA-C. 1 Both categories have serious autonomic failure but MSA-P is described mostly by parkinsonism due to the prevailing striatonigral degeneration and MSA-C is characterized by cerebellar ataxia due to olivopontocerebellar atrophy. 1 The main histopathological hallmark of MSA is the prevalence of glial cytoplasmic inclusions (GCIs) in oligodendrocytes. 2  Unlike in PD and dementia with Lewy bodies, where aSyn is accumulated mostly in neurons, in MSA aSyn is accumulated in the cytoplasm of oligodendroglial cells. 3 aSyn in GCIs is mostly present in a fibrillary oligomeric form 4 and as serine 129 phosphorylated (pS129 aSyn) that is considered to destabilize aSyn and elevate its aggregation. 5 aSyn oligomerization has been shown to cause lossof-function toxicity, for example in synaptic vesicle regulation, and damage several cellular organelles, including endoplasmic reticulum, mitochondria and protein degradation pathways (for review see 6 ).
Animal models of MSA have shown that aSyn overexpression in oligodendrocytes causes cellular death, leading to demyelination, neuronal damages and premature death, 7,8 suggesting the involvement of aSyn in the pathophysiology of MSA.
It has been under debate if aSyn in GCIs is exogenous and transferred from neurons by cell-to-cell transfer since only low amounts of aSyn mRNA have been measured in healthy and MSA patientderived oligodendrocytes. 9 However, recent studies suggest that endogenous aSyn in oligodendrocytes is critical for aSyn aggregation in MSA after the exogenous aSyn has triggered the process. 10 Another important factor contributing to aSyn aggregation in MSA is p25α/TPPP protein (tubulin polymerization-promoting protein; p25α). p25α is a protein promoting tubulin assembly and maintaining myelin sheath 11 but in MSA it relocates from myelin sheaths and accumulates to oligodendroglial soma. 12 This redistribution is associated with demyelination of axons, and subsequently, phosphorylation and aggregation of aSyn, leading to the abovementioned aggregation process.
Compounds having an effect on aSyn aggregation or degradation, such as anle138b 13 or enhancement of aSyn degradation by autophagy activation, 14 respectively, have shown beneficial effects on experimental MSA models. Therefore, targeting aSyn could be a potential disease-modifying therapy for MSA. In our previous studies, we have shown that a serine protease, prolyl oligopeptidase (PREP), increases aSyn oligomerization via direct protein-protein interaction 15,16 and negatively regulates autophagy. 17 Importantly, small-molecular PREP inhibitors can interfere with this interaction, leading to decreased aSyn aggregation, 15,18,19 and PREP inhibition or deletion also stimulates autophagy and increases the degradation of aSyn oligomers [16][17][18] and fibrils. 20 In aSyn-based in vivo PD model, PREP inhibition has shown disease-modifying impact by restoring an aSyn virus vector-induced behavioural deficit. 21  Abuse, Baltimore, MD). aSyn plasmid construction was described in. 17 pAAV-EF1a control vector with 50 bp insert was created as described in. 25 aSyn plasmid concentration was kept similar when transfected with empty plasmid or in combination with p25α to avoid changes in protein expression, and empty plasmid was used also with p25α transfection. For PREP inhibitor experiments, KYP-2047 was diluted to cell culture medium from 100 mM stock in 100% DMSO to 1 and 10 µM, and corresponding concentration of DMSO was used as a vehicle control.

| PREP enzyme activity assay
For the PREP activity assay, fluorometric assay based on Suc-Gly-Pro-aminomethylcoumarin (AMC) substrate was used as in Myöhänen et al. (2012). 26 Briefly, cells were plated to 6-wellplate with density of 400,000 cells/well, transfected with p25α and after 24 h, lysed with a lysis buffer. PREP activity was measured from supernatants, and released AMC was correlated on total protein amounts measured by using bicinchoninic acid method (BCA; Pierce BCA Protein Assay Kit, #23225, Thermo Fisher Scientific). All activity measurements were made in triplicate with 2 biological repeats.

| Cell viability assay
OLN-AS7 and MO3.13 cells were plated with the density of 10,000 cell/well in 96-well plate and the next day transfected by p25α (OLN-AS7) or with p25α and aSyn (MO3.13). 24 h after transfection, the cells were incubated for 48 h in the presence of 1 μM or 10 µM KYP-2047 or DMSO vehicle (0.01% DMSO; 150 μl/well). LDH release assay was performed as previously described. 26

| Immunocytochemistry
Immunocytochemistry (ICC) was used to detect changes in total aSyn, pS129 aSyn and tubulin retraction after p25α transfection and PREP inhibition. Briefly, OLN-AS7 cells were plated over glass coverslips in a 12-well plate with a density of 60,000 cells/ well and allowed to attach overnight. 24 h after p25α transfection, cells were treated with 10 µM KYP-2047 for 48 h and finally fixed with 4% paraformaldehyde for staining. 10% normal goat serum (S-1000, Vector Laboratories) was used for blocking for 30 min and thereafter cells were incubated with primary antibodies against aSyn, pS129 aSyn, α-tubulin or p25α overnight at +4°C (details in Table 1). After washes, the following secondary antibodies were used to incubate cells 2 h in room temperature: for sheep aSyn, anti-sheep Alexa Fluor488 (1:400; ab150177, Abcam; RRID: AB_2801320); for mouse pS129 aSyn, anti-mouse Alexa Fluor488 (1:800; ab150113, Abcam; RRID: AB_2576208); for rabbit p25α, anti-rabbit Alexa Fluor 568 (1:800; ab175471, Abcam; RRID: AB_2576207). α-tubulin antibody was conjugated with Alexa Fluor 488, and no secondary antibody incubation was used. Cells were mounted with Vectashield containing DAPI to stain nuclei (H-1200, Vector Laboratories; RRID: AB_2336790). Imaging was performed using Leica TCS SP5 confocal microscope (Leica Microsystems). In tubulin retraction analysis, images were converted to 8-bit grayscale format, and the area of tubulin network in p25α transfected cells was analysed by ImageJ (version 1.51; National Institute of Health; RRID:SCR_002285) with a protocol described in. 23 For publication purposes, the images were converted from grayscale to RGB colour and recoloured with corresponding colours, and minor modifications to brightness and contrast were made. All modifications to brightness and contrast were done similarly to each image, and they were analysed prior to modifications. Thereafter, the membranes for aSyn and pS129 aSyn immunoblotting were incubated 30 min in 4% paraformaldehyde for fixing. The membranes were blocked with 5% skim milk and incubated in primary antibodies (Table 1)    Based on cell viability studies, 10 µM KYP-2047 was selected for further assays.

| PREP inhibition reduces p25α induced cell death and tubulin retraction in OLN-AS7 cells
Translocation of p25α to the cytosol leads to loss of its normal function in myelination and this causes microtubule retraction contributing to cellular toxicity. 10,23 Similarly to previous studies, our ICC assay showed that transfected p25α caused significant decrease in the tubulin area in 48 h compared to LFC control or 10 µM KYP-2047 ( Figure 1C-F; F 3,73 = 1.358, p < 0.05, 1-way ANOVA with Tukey's post-test). When p25α transfection was combined with 10 µM KYP-2047, the tubulin retraction was not significant compared to LFC but also not significantly different from p25α+vehicle ( Figure 1C).

| PREP inhibition reduces insoluble total and phosphorylated aSyn in OLN-AS7 cells
An earlier study with OLN-AS7 cells showed that aSyn phosphorylation contributes to microtubule retraction and aSyn toxicity. 23 PREP inhibition has reduced the number of aSyn oligomeric particles and particularly insoluble aggregates in our earlier studies, 17,21,26 and therefore, we characterized if KYP-2047 has an impact on aSyn phosphorylation and aggregation in OLN-AS7 cells. We first tested 24 h p25α transfection but the WB analysis revealed the increase in aSyn and pS129 aSyn levels particularly in SDS fraction were mild ( Figure S1). p25α toxicity has been shown to be the most evident at 48 h after p25α transfection in OLN-AS7 cells, 23 and therefore, we

| PREP inhibition induces autophagy and reactivates protein phosphatase 2A (PP2A) in OLN-AS7 cells
Our previous results show that PREP inhibition can activate autophagy to reduce aSyn aggregates, 16,17,27 and we recently showed that PREP inhibition induces autophagy by activating PP2A. 25 We studied the levels of the total catalytic subunit of PP2A (PP2Ac) and inactive Tyr307 phosphorylated PP2Ac (pPP2A; antibody specific for inactive PP2A

| DISCUSS ION
MSA is a devastating neurodegenerative disease that lacks diseasemodifying therapy. In this study, we have shown that KYP-2047, a PREP inhibitor that has shown beneficial effects in several aSynbased PD models, protects cells from aSyn toxicity in a cellular model of MSA. Our results demonstrate that in OLN-AS7 cells, KYP-2047 induces autophagy by re-activating PP2A and this leads to decreased levels of aSyn aggregates and elevated cell viability.
Alterations in PREP activity have been reported in several neurodegenerative diseases, including Parkinson's and Alzheimer's diseases, 29,30 but there are no reports about PREP in MSA. PREP activity has been reported in rat oligodendrocytes, 31 but the activity was lower than in neurons or astrocytes. However, in the current study, the PREP activity measured from OLN-AS7 cells was remarkably high when compared, for example to SH-SY5Y neuronal cell culture 26 and p25α transfection elevated PREP activity.
It has been shown that PREP activity and levels are increased in astrocytes and microglial cells when they are activated, suggesting that PREP is inducible in glial cell activation. 29,32 Therefore, it could be possible that PREP is also induced in MSA after p25α translocation but this requires further studies by using MSA patient material.
Our results revealed that p25α transfection in OLN-AS7 cells caused increased phosphorylation of PP2A at Tyr307 that is indicative for PP2A inactivation. 33 40,41 There are no reports about PP2A levels in MSA, or a connection between p25α and PP2A, but our finding suggests that this should be further studied and that PP2A inactivation could contribute to MSA as well. Monomeric aSyn is degraded via proteasomes and chaperonemediated autophagy but these systems are easily damaged by aSyn aggregates, leading to elevated aSyn concentrations and aggregation. 46,47 Macroautophagy (autophagy) is able to degrade aSyn oligomers and larger aggregates, 48 and the activation of autophagy has been suggested as a possible therapy against synucleinopathies and F I G U R E 5 p25α transfection elevates aSyn levels and phosphorylation in MO3.13 cells. p25α co-transfection with aSyn elevated aSyn levels both in soluble and SDS fraction (A, C). After 48h cotransfection, pS129 aSyn was increased significantly compared to empty control and aSyn+empty in both fractions (B, D). This was in correlation with mRNA levels where similar increase in aSyn mRNA expression was seen when co-transfected with p25α (E; SNCA). However, no changes in cell viability were seen by any transfections or treatments (F), and 10 µM KYP-2047 had no effect on aSyn levels or aSyn mRNA. Data are presented as mean ± SEM. p < 0.05; **, p < 0.01; ***, p < 0. Taken together, our data suggest that p25α transfection to cellular models of MSA induces not only aSyn phosphorylation and aggregation but also PP2A inactivation. Additionally, our results suggest that autophagy impairment is important step for aSyn-induced toxicity in cellular models of MSA. In OLN-AS7 cells, PREP inhibition after p25α transfection re-activated PP2A, reduced both total and pSer129 aSyn in the insoluble fraction by inducing autophagy and attenuated the toxicity of aSyn in in OLN-AS7 cellular model of MSA.
Our results suggest that PREP inhibition could be a possible diseasemodifying therapy that is applicable for several synucleinopathies.

CO N FLI C T S O F I NTE R E S T
The authors confirm that there are no conflicts of interest.
F I G U R E 6 p25α transfection increases PP2A phosphorylation in MO3.13 cells. Similar to OLN-AS7 cells, p25α, aSyn or p25α+aSyn transfection did not alter the levels of catalytic subunit of PP2A (A; PP2Ac). However, Tyr307 phosphorylated PP2Ac (B; pPP2Ac) was significantly increased by p25α transfection but not with other combinations. 10 µM KYP-2047 reduced pPP2Ac levels but not significantly (B). p62 or autophagosome marker LC3BII were not altered by p25α, aSyn or their combination in MO3.13 cells (C-D). Data are presented as mean ± SEM. *, p < 0.05; empty+veh vs. p25α+veh, 1-way ANOVA with Tukey's post hoc test. See figure S7 for uncut blots. Representative blots in Figure 6B and D are from the same membrane and therefore share the vinculin loading control Writing-review & editing (lead).

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this published article. Raw values are available from the corresponding author on reasonable request.