Autophagy mediates the clearance of oligodendroglial SNCA/alpha-synuclein and TPPP/p25A in multiple system atrophy models

ABSTRACT

Accumulation of the neuronal protein SNCA/alpha-synuclein and of the oligodendroglial phosphoprotein TPPP/p25A within the glial cytoplasmic inclusions (GCIs) represents the key histophathological hallmark of multiple system atrophy (MSA). Even though the levels/distribution of both oligodendroglial SNCA and TPPP/p25A proteins are critical for disease pathogenesis, the proteolytic mechanisms involved in their turnover in health and disease remain poorly understood. Herein, by pharmacological and molecular modulation of the autophagy-lysosome pathway (ALP) and the proteasome we demonstrate that the endogenous oligodendroglial SNCA and TPPP/p25A are degraded mainly by the ALP in murine primary oligodendrocytes and oligodendroglial cell lines under basal conditions. We also identify a KFERQ-like motif in the TPPP/p25A sequence that enables its effective degradation via chaperone-mediated autophagy (CMA) in an in vitro system of rat brain lysosomes. Furthermore, in a MSA-like setting established by addition of human recombinant SNCA pre-formed fibrils (PFFs) as seeds of pathological SNCA, we thoroughly characterize the contribution of CMA and macroautophagy in particular, in the removal of the exogenously added and the seeded oligodendroglial SNCA pathological assemblies. We also show that PFF treatment impairs autophagic flux and that TPPP/p25A exerts an inhibitory effect on macroautophagy, while at the same time CMA is upregulated to remove the pathological SNCA species formed within oligodendrocytes. Finally, augmentation of CMA or macroautophagy accelerates the removal of the engendered pathological SNCA conformations further suggesting that autophagy targeting may represent a successful approach for the clearance of pathological SNCA and/or TPPP/p25A in the context of MSA.

Abbreviations: 3MA: 3-methyladenine; ACTB: actin, beta; ALP: autophagy-lysosome pathway; ATG5: autophagy related 5; AR7: atypical retinoid 7; CMA: chaperone-mediated autophagy; CMV: cytomegalovirus; CTSD: cathepsin D; DAPI: 4′,6-diamidino-2-phenylindole; DMEM: Dulbecco’s modified Eagle’s medium; Epox: epoxomicin; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GCIs: glial cytoplasmic inclusions; GFP: green fluorescent protein; HMW: high molecular weight; h: hours; HSPA8/HSC70: heat shock protein 8; LAMP1: lysosomal-associated membrane protein 1; LAMP2A: lysosomal-associated membrane protein 2A; MAP1LC3/LC3: microtubule-associated protein 1 light chain 3; mcherry: monomeric cherry; MFI: mean fluorescence intensity; mRFP: monomeric red fluorescent protein; MSA: multiple system atrophy; OLN: oligodendrocytes; OPCs: oligodendroglial progenitor cells; PBS: phosphate-buffered saline; PC12: pheochromocytoma cell line; PD: Parkinson disease; PFFs: pre-formed fibrils; PIs: protease inhibitors; PSMB5: proteasome (prosome, macropain) subunit, beta type 5; Rap: rapamycin; RFP: red fluorescent protein; Scr: scrambled; SDS: sodium dodecyl sulfate; SE: standard error; siRNAs: small interfering RNAs; SNCA: synuclein, alpha; SQSTM1: sequestosome 1; TPPP: tubulin polymerization promoting protein; TUBA: tubulin, alpha; UPS: ubiquitin-proteasome system; WT: wild type

KEYWORDS:

• Chaperone-mediated autophagy
• fibrils
• inclusions
• macroautophagy
• oligodendrocytes
• proteasome
• seeding

Introduction

Multiple system atrophy (MSA) is an adult-onset, devastating and relentlessly progressive neurodegenerative disorder of uncertain etiology, mainly characterized by autonomic failure, parkinsonian features and ataxia [1]. The neuropathological hallmark of the disease is the presence of glial cytoplasmic inclusions (GCIs) within the cytoplasm of oligodendrocytes. The main components of GCIs are the neuronal protein SNCA/alpha-synuclein and the oligodendroglial-specific phosphoprotein TPPP/p25A [2–6]. The origin of SNCA found in oligodendrocytes still remains enigmatic, taking into account that mature oligodendrocytes do not normally express the protein and there is a controversy regarding the presence of SNCA mRNA in human oligodendrocytes [7–10]. Nonetheless, others and we have recently reported that the presence of both the endogenous oligodendroglial SNCA and TPPP/p25A [11–14] is critical for the development and spread of disease pathology.

The identification of endogenous oligodendroglial SNCA as a major culprit for the development of MSA-like pathology [11] suggests that manipulation of the expression of SNCA and/or TPPP/p25A in oligodendrocytes may provide a rational approach to combat the accumulation of SNCA in GCIs and the progression of MSA. In contrast to the plethora of studies investigating the degradation of neuronal SNCA (reviewed in [15–17]), there is a paucity of data regarding the proteolytic machineries responsible for the clearance of endogenous oligodendroglial SNCA, under physiological and pathological conditions. For TPPP/p25A degradation, two reports showed that proteasomal inhibition resulted in accumulation of the ectopically expressed protein in cell lines [18,19], whereas no data exist regarding the role of autophagy in TPPP/p25A handling.

Furthermore, dysregulation of both the ubiquitin-proteasome system (UPS) and the autophagy-lysosome pathway (ALP) has been proposed to contribute to the aggregate formation within oligodendrocytes of MSA brains. Specifically, it has been reported that macroautophagy is upregulated during MSA pathogenesis and may participate in the removal of protein aggregates [20]. The presence of autophagy-related proteins within GCIs of MSA, such as MAP1LC3/LC3 and SQSTM1/p62, further supports a potential role of the ALP in disease progression [21–25].

Herein, by utilizing pharmacological and molecular inhibition of both autophagic and proteasomal pathways, we provide evidence that the endogenous oligodendroglial SNCA and TPPP/p25A are degraded mainly by the ALP (with the proteasome participating only in the clearance of SNCA) in rat oligodendroglial cell lines and mouse primary oligodendrocytes. Moreover, our data show that TPPP/p25A bears a KFERQ-like motif and is effectively cleared via chaperone-mediated autophagy (CMA) in an in vitro system of isolated rat brain lysosomes. Interestingly, by using human recombinant SNCA pre-formed fibrils (PFFs) as seeds of pathological SNCA, we thoroughly characterize the contribution of CMA and macroautophagy in particular, in the removal of the human exogenously added and the seeded rodent oligodendroglial SNCA pathological assemblies. Finally, we demonstrate that upon PFF treatment, autophagic flux is blocked and that TPPP/p25A exerts an inhibitory effect on macroautophagy, while at the same time CMA is upregulated as a compensatory mechanism to effectively remove the pathological SNCA species generated within oligodendrocytes. Most importantly, enhancement of CMA or macroautophagy accelerates the clearance of the engendered pathological SNCA conformations further suggesting that autophagy augmentation may represent a successful approach for the clearance of both SNCA and/or TPPP/p25A in the context of MSA.

Results

Both the proteasome and the ALP contribute to the degradation of the endogenous rat oligodendroglial SNCA

In order to study the degradation pathways responsible for the proteolysis of the endogenous rat oligodendroglial SNCA, we pharmacologically and molecularly inhibited the lysosome (CMA and macroautophagy) and the proteasome. Specifically, we utilized the rat oligodendroglial OLN-93 (control), OLN-AS7 (overexpressing human SNCA) and OLN-p25α (overexpressing human TPPP/p25A) cell lines, which express very low to non-detectable levels of the endogenous SNCA and we treated cells with 20 mM NH₄Cl (as total lysosomal inhibitor), 10 mM 3 MA (as macroautophagy inhibitor) or 15 nM epoxomicin (epox, as proteasomal inhibitor) for 48 h. Confocal microscopy analysis revealed that inhibition of the proteasome or the lysosome increased the oligodendroglial SNCA signal in all lines (expressed as μm² area/cell) as detected with the rodent-specific D37A6 antibody (Figure 1A-D), however to a different extent. These data complement recent reported data from our lab [11] supporting both the lysosomal (NH₄Cl- and 3 MA-dependent) and the proteasomal contribution to the oligodendroglial SNCA protein turnover. Moreover, immunoblot analysis in the control OLN-93 cells further confirmed the notion that the oligodendroglial SNCA is mostly degraded by the lysosome (Figure 1Ei, ii).

Figure 1. The endogenous rat oligodendroglial SNCA is degraded via both autophagy and the proteasome. (A-C) Confocal microscopy with endogenous rat-specific SNCA and TUBA antibodies reveals the cytoplasmic accumulation of the rodent oligodendroglial SNCA in OLN-93 (A), OLN-AS7 (B) and OLN-p25α cells (C) upon treatment with: proteasomal (epoxomicin, 15 nM) or lysosome/autophagy (NH₄Cl 20 mM, 3 MA 10 mM) inhibitors for 48 h. Representative immunofluorescence images with antibodies against TUBA (green) and rat SNCA (red, D37A6 antibody) and DAPI staining are shown. Scale bar: 25 μm. (D) Quantification of the endogenous rat SNCA protein levels in OLN cells measured as μm² area surface/cell following treatment with proteasomal (epox) or lysosomal (NH₄Cl, 3 MA)inhibitors for 48 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test (to compare between inhibitor-treated and untreated cells) or #p < 0.05; ###p < 0.001 by two-way ANOVA with Bonferroni’s correction (to compare between the different treated cell cultures). (Ei) Representative immunoblots for rodent SNCA (D37A6 antibody), poly-ubiquitinated proteins, LC3-I and LC3-II, and SQSTM1 (as macroautophagy markers), and ACTB (as loading control), verifying the increase of endogenous oligodendroglial SNCA protein levels in OLN-93 cells following lysosomal inhibition (NH₄Cl, 48 h). (Eii) Quantification of the endogenous rat SNCA protein levels in OLN-93 cells treated with proteasomal (epox) or lysosome/autophagy (NH₄Cl, 3 MA) inhibitors for 48 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05, by one-way ANOVA with Tukey’s post hoc test.

In order to dissect further the partitioning of CMA and macroautophagy lysosomal pathways in the rodent SNCA degradation under basal conditions, we treated all OLN cell lines with Lamp2a- (Lsi1 and Lsi2) or Atg5- (Atg5 si) specific siRNAs, targeting the Lamp2a receptor gene that acts as the rate-limiting step of the CMA pathway [26], or the core autophagy gene Atg5, encoding a component of the ATG12 conjugation system, that is indispensable for phagophore formation and its ablation abrogates macroautophagy [27,28]. The efficacy of these siRNAs on Lamp2a or Atg5 gene silencing was verified 72 h later by confocal microscopy that revealed an almost 3-fold decrease in the respective protein levels (Figure S1A-D), measured as mean fluorescence intensity (M.F.I.)/cell. The downregulation of ATG5 was accompanied by the detection of decreased levels of the autophagosome marker LC3B, further verifying the efficient inhibition of macroautophagy (Figure S1E-F). Interestingly, treatment of all OLN cells with Lsi1 and Lsi2 or Atg5 si RNAs significantly increased the protein levels of the endogenous rat oligodendroglial SNCA (Figure 2).

Figure 2. Downregulation of the CMA- and macroautophagy-related genes, Lamp2a and Atg5 leads to the accumulation of the endogenous rat oligodendroglial SNCA. (A-C) Representative immunofluorescence images of OLN-93 (A), OLN-AS7 (B) and OLN-p25α cells (C) treated with two different siRNAs targeting the rat Lamp2a receptor (Lsi1 and Lsi2, 60 nM) for 72 h. Scrambled RNA sequences (scr) were used as negative control. The antibodies utilized were against TUBA (green) and rat SNCA (red, D37A6 antibody). DAPI was used as a nuclear marker. Scale bar: 25 μm. (D) Quantification of the endogenous rat SNCA protein levels in OLN cells measured as μm² area surface/cell following treatment with Lamp2a-siRNAs (Lsi1 and Lsi2) for 72 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; **<0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test (to compare between siRNA-treated and untreated cells) or ##p < 0.01 by two-way ANOVA with Bonferroni’s correction (to compare between the different treated cell cultures). (E-G) Confocal microscopy images depicting the cytoplasmic accumulation of the rodent oligodendroglial SNCA (D37A6 ab, red) in OLN-93 (E), OLN-AS7 (F) and OLN-p25α cells (G) upon treatment with Atg5-siRNA (Atg5 si, 10 nM) for 72 h. Scrambled RNA sequences (scr) were used as negative control. TUBA was used as a cytoskeletal marker and DAPI as a nuclear marker. Scale bar: 25 μm. (H) Quantification of the endogenous rat SNCA protein levels in OLN cells measured as μm² area surface/cell following treatment with Atg5-siRNA (Atg5 si, 10 nM) for 72 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; **<0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test.

We subsequently targeted the Psmb5 (proteasome (prosome, macropain) subunit, beta type 5) gene, encoding a component of the 20S proteasome, by treating all OLN cell lines with Psmb5-specific siRNAs (or scr sequences as control) for 72 h. The efficient inhibition of the proteasome was confirmed by detecting elevated levels of (poly)-ubiquitinated proteins (Figure 3A-D) and decreased chymotrypsin (CT)-like proteasomal activity (Figure 3E) in Psmb5 siRNA-treated cells. Similar to the effect of epoxomicin (Figure 1D), molecular inhibition of the 20S proteasomal activity resulted in elevated levels of the endogenous oligodendroglial SNCA in all OLN cells (Figure 3F). It is worth mentioning that upon proteasomal inhibition, the endogenous rat SNCA accumulates to a greater extent in human SNCA- or TPPP/p25A- overexpressing OLN cells, as compared to control OLN-93 cells (Figure 1 and 3Fi-Fii). This could be attributed to a clonal variability between the three lines or to the presence of more active proteasomes in OLN-AS7 and OLN-p25α cells. In addition, treatment of OLN-AS7 cells with the Psmb5 siRNA (or scr as control) resulted in a statistically significant increase in human SNCA levels in the Psmb5 siRNA-treated cells, thus supporting a role of the proteasome in the degradation of the protein under physiological conditions (Figure 3Gi-ii). Conversely, TPPP/p25A protein did not accumulate upon proteasomal inhibition under physiological conditions in OLN-p25α cells (Figure 3 Hi-ii).

Figure 3. Downregulation of the Psmb5 (proteasome (prosome, macropain) subunit, beta type 5) gene increases both the endogenous rat oligodendroglial and overexpressed human SNCA protein levels, whereas overexpressed TPPP/p25A levels remain unaltered. (A-D) The levels of ubiquitinated proteins (UBB) are increased upon Psmb5 downregulation. (A) Representative immunofluorescence images of OLN-93 cells treated with siRNAs targeting the rat Psmb5 gene (Psmb5 si, 10 nM) for 72 h. Scrambled RNA sequences (scr) were used as negative control. The antibodies utilized were against TUBA (green) and UBB (red). DAPI was used as a nuclear marker. Scale bar: 5 μm. (B) Quantification of UBB protein levels in OLN-93 cells, measured as M.F.I./cell. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05, by Student’s unpaired t test. (C) Representative immunoblots of RIPA-soluble protein cell extracts probed for poly-ubiquitinated (poly-UBB) proteins and ACTB (as loading control) antibodies, demonstrating the accumulation of poly-UBB proteins following Psmb5 silencing. (D) Quantification of the poly-UBB protein levels in OLN-93 cells treated with scr or Psmb5 siRNAs for 72 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05, by Student’s unpaired t test. (E) Quantification of the CT-like proteasomal activity in OLN-93 cells treated with Psmb5 or scr siRNAs for 72 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test (to compare between the epox-treated and the scr- or Psmb5 siRNAs-treated cells), ###p < 0.001, by one-way ANOVA with Tukey’s post hoc test (to compare between the scr- and Psmb5 siRNAs-treated cells). (Fi) Representative immunofluorescence images of OLN-93, OLN-AS7 and OLN-p25α cells treated with scr or Psmb5 siRNAs (10 nM) for 72 h. The antibodies utilized were against TUBA (green) and rat SNCA (red, D37A6 antibody). DAPI was used as a nuclear marker. Scale bar: 25 μm. (Fii) Quantification of the endogenous rat SNCA protein levels measured as μm² area surface/cell following treatment of OLN cells with Psmb5 or scr siRNAs for 72 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; p**<0.01; by one-way ANOVA with Tukey’s post hoc test (to compare between siRNA-treated and untreated cells) or #p < 0.05; ##p < 0.01 by two-way ANOVA with Bonferroni’s correction (to compare between the different cell lines). (Gi) Representative immunofluorescence images of OLN-AS7 cells showing the increased protein levels of human SNCA following siRNA delivery targeting Psmb5 using antibodies against human SNCA (red, LB509 antibody) and TUBA (green) and DAPI staining. Scale bar: 25 μm. (Gii) Quantification of human SNCA protein levels in scr- or Psmb5 siRNA (10 nM, for 72 h) transfected OLN-AS7 cells measured as M.F.I./cell. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; by Student’s unpaired t test. (Hi) Representative immunofluorescence images of OLN-p25α cells showing the protein levels of human TPPP/p25A upon molecular inhibition of the proteasome using antibodies against human TPPP/p25A (red) and TUBA (green) and DAPI staining. Scale bar: 25 μm. (Hii) Quantification of human TPPP/p25A protein levels in scr- or Psmb5 siRNA (10 nM, for 72 h) transfected OLN-p25α cells measured as M.F.I./cell. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment.

It is important to mention that the D37A6 antibody does not produce a specific immunofluorescence signal in control conditions (phosphate-buffered saline [PBS]- or scr siRNA-treated cells, Figures 1–3), whereas in immunoblot analysis, film overexposure enables the detection of the rodent SNCA signal at minute amounts even at basal conditions (Figure 1Ei, ii). In all cases, TUBA was used as a cytoskeletal marker, in order to verify the cytoplasmic distribution of the endogenous oligodendroglial SNCA protein.

The overexpressed human SNCA accumulates upon lysosomal or proteasomal inhibition in OLN-AS7 cells

Treatment of rat OLN-AS7 cells with NH₄Cl or 3 MA for 48 h increased human SNCA protein levels as verified by both immunofluorescence and Western blot (Figure 4A-B) analyses, suggesting that the over-expressed SNCA is degraded, at least partly, via the ALP. Interestingly, epox treatment of OLN-AS7 cells for 24 h evoked a robust increase of human SNCA mRNA levels (Figure S2A), accompanied by accumulation of human SNCA protein levels (data not shown) due to the epox-mediated nonspecific activation of the CMV promoter (which drives the expression of human SNCA in OLN-AS7 cells), as has been previously suggested by Biasini et al., 2004 [29]. Similar effect has been reported using lactacystin or MG132 [29], rendering difficult the use of pharmacological modulators of the proteasome in OLN-AS7 cells. On the contrary, treatment of OLN-AS7 cells with NH₄Cl for 24 h or 48 h did not alter human SNCA mRNA levels, suggesting that the observed SNCA protein accumulation upon lysosomal inhibition was due to impaired degradation (Figure S2B).

Figure 4. The overexpressed human SNCA accumulates upon pharmacological or molecular inhibition of the autophagy-lysosome pathway in OLN-AS7 cells. Representative immunofluorescence images of OLN-AS7 cells treated with NH₄Cl (20 mM) or with the macroautophagy inhibitor 3 MA (10 mM) (Ai) or following siRNA delivery targeting Lamp2a (Ci) or the autophagy-related Atg5 gene (Ei) using antibodies against human SNCA (red, LB509 antibody) and TUBA (green) and DAPI staining. Scale bar: 25 μm. Quantification of human SNCA protein levels in OLN-AS7 cells measured as M.F.I./cell following treatment with NH₄Cl, 3 MA for 48 h (Aii), or transfection with Lsi1 and Lsi2 (60 nM) (Cii), or Atg5 siRNAs (10 nM) (Eii) for 72 h. Scrambled RNA sequences (scr) were used as negative control. Data are expressed as the mean ± SE of four independent experiments with duplicate samples/condition within each experiment; *p < 0.05; **p < 0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test. Representative immunoblots of human SNCA protein levels (using the human SNCA-specific 4B12 antibody) derived from OLN-AS7 cells following total lysosomal (NH₄Cl) or macroautophagy inhibition (3 MA) (Bi), or CMA-dependent (Di) or macroautophagy-related (Fi) gene silencing using Lamp2a- (Lsi1 and Lsi2) and Atg5-siRNAs (or scr as control), respectively. Gene silencing was verified by the detection of LAMP2A or ATG5 protein levels. Antibodies against LC3-I and -II, and SQSTM1 were used as macroautophagy markers (Bi), and ACTB as a loading control. Quantification of the human SNCA protein levels in OLN-AS7 cells treated with NH₄Cl or 3 MA for 48 h (Bii), Lsi1 and Lsi2 siRNAs for 72 h (Dii) and Atg5 siRNAs for 72 h (Fii). Data are expressed as the mean ± SE of four independent experiments with duplicate samples/condition within each experiment; *p < 0.05; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test.

Likewise, transfection of OLN-AS7 cells with Lsi1 and Lsi2 or Atg5 siRNAs for 72 h also increased protein levels of human SNCA, as detected with confocal microscopy and immunoblot analyses (Figure 4C-F). Given that the LB509 antibody was raised against SNCA in Lewy bodies [30] and hence may prefer binding to certain SNCA species, we have also utilized another human-specific antibody, the SYN211 monoclonal antibody suggested to recognize all SNCA conformations (monomers, oligomers and fibrils) [31]. Using the SYN211 instead of the LB509 antibody we obtained similar results regarding the mode of human SNCA clearance (Figure S3). Overall, our data compellingly demonstrate that both the ALP and the proteasome participate to the degradation of the overexpressed human SNCA within oligodendrocytes.

The autophagy-lysosome pathway and not the proteasome is responsible for the clearance of the overexpressed human TPPP/p25A in OLN-p25α cells

To decipher the role of the ALP on TPPP/p25A proteolysis, OLN-p25α cells were incubated with 20 mM NH₄Cl or 10 mM 3 MA (48 h) for total lysosomal or macroautophagy-dependent inhibition, respectively, and protein levels of TPPP/p25A were assessed by immunofluorescence analysis (as M.F.I./cell) and Western blotting. According to the data presented in Figure 5A and B, treatment of OLN-p25α cells with NH₄Cl led to TPPP/p25A protein accumulation, suggesting that the lysosome is responsible, at least partly, for TPPP/p25A proteolysis under basal conditions. Inhibition of macroautophagy with 3 MA did not cause a significant elevation of TPPP/p25A albeit a small increase was observed (Figure 5A and B). As in OLN-AS7 cells, epoxomicin, but not NH₄Cl treatment, evoked a nonspecific upregulation of human TPPP/p25α mRNA levels due to the enhanced CMV-driven promoter gene transcription (Figure S2C, D).

Figure 5. The overexpressed human TPPP/p25A is increased following pharmacological or molecular inhibition of the autophagy-lysosome pathway in OLN-p25α cells. Representative immunofluorescence images of OLN-p25α cells showing the increased protein levels of human TPPP/p25A upon pharmacological (Ai) or molecular (Ci, Ei) inhibition of lysosomal pathways (total, CMA, macroautophagy) using antibodies against human TPPP/p25A (red) and TUBA (green) and DAPI staining. Scale bar: 25 μm. Quantification of human TPPP/p25A protein levels in OLN-p25α cells measured as M.F.I./cell following treatment with NH₄Cl, 3 MA for 48 h (Aii) or with Lamp2a-or Atg5-siRNAs for 72 h (Cii and Eii). Scrambled RNA sequences (scr) were used as negative control. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; by one-way ANOVA with Tukey’s post hoc test. Representative immunoblots of OLN-p25α cell lysates verifying the accumulation of TPPP/p25A protein levels upon lysosomal inhibition using either pharmacological inhibitors (Bi) or gene silencing methods (Di and Fi). Lamp2a and Atg5 downregulation was verified using antibodies against LAMP2A or ATG5 proteins (Di and Fi). Antibodies against LC3-I and -II, and SQSTM1 were used as macroautophagy markers (Di), and ACTB as a loading control. Quantification of the human TPPP/p25A protein levels in OLN-p25α cells treated with (Bii) NH₄Cl or 3 MA for 48 h, (Dii) Lsi1 and Lsi2 (60 nM) for 72 h or (Fii) Atg5 siRNAs (10 nM) for 72 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; **p < 0.01, ***p < 0.001 by one-way ANOVA with Tukey’s post hoc test.

Furthermore, Lamp2a gene silencing revealed that TPPP/p25A protein is efficiently degraded via the CMA pathway (Figure 5C and D), since both the Lsi1 and Lsi2 siRNA sequence targeting the rat Lamp2a evoked a statistically significant accumulation of TPPP/p25A protein levels (Figure 5Cii, Dii). Similar to the 3 MA treatment (Figure 5A, B), Atg5 downregulation did not yield a statistically significant increase of human TPPP/p25A protein levels (Figure 5E and F). It has to be noted that treatment of all OLN cells with the pharmacological inhibitors NH₄Cl and epox for 16 or 48 h, had no significant impact on cell survival (Figure S4A-F). However, only in the case of 3 MA-treated OLN-AS7 and OLN-p25α cells for 48 h, a 20% decrease in cell survival was observed (Figure S4D and F).

The oligodendroglial-specific protein TPPP/p25A is a putative CMA substrate

In order for a protein to be a CMA substrate, it has to bear a KFERQ-like CMA-targeting motif that guides the delivery of the protein to the level of the lysosomal membrane, where binding with the CMA-specific receptor, LAMP2A, takes place. We therefore examined whether this is the case for TPPP/p25A and found that it contains a pentapeptide sequence (KKRFK), consistent with a CMA targeting motif. According to the 3D structure of the mouse CGI38 protein (PDB id: 1WLM) (homologous to the human TPPP/p25A [NM_001108461]), the KKRFK motif is located in a loop, flanked by two adjacent α-helices, both fully exposed to the solvent. The superposed structures of the human, mouse and the 3D model of the TPPP/p25A bearing the KKRFK motif are shown in Figure 6A. Given that the presence of the CMA motif does not guarantee that the protein is actually degraded via this pathway [32], we incubated human recombinant TPPP/p25A with isolated rat brain lysosomes and assessed the in vitro degradation of the protein, in the presence or absence of Protease Inhibitors (PIs). Under these conditions, we found that purified recombinant TPPP/p25A added to the incubation medium was translocated into and efficiently degraded by intact brain lysosomes, since lysosomal protease inhibitors increased the levels of the lysosome-associated protein (Figure 6B-C). Interestingly, a competition assay utilizing 3x (0.6 μg) and 6x (1.2 μg) amount of human recombinant SNCA, a well-established CMA substrate [33] effectively inhibited the degradation of TPPP/p25A by intact lysosomes, in a dose-dependent manner (Figure 6B-C). These data, in combination with the increased protein levels of TPPP/p25A following LAMP2A downregulation (Figure 5C-D), further confirm that the oligodendroglial TPPP/p25A is a putative CMA substrate and is degraded via this pathway in OLN cells.

Figure 6. The oligodendroglial-specific protein TPPP/p25A is a putative CMA substrate, containing a KFERQ-like motif necessary for CMA targeting. (A) X-ray crystallography revealed the 3D structure of the rat TPPP3/Cgi-38 protein (PDB id: 1WLM) [homologous to the human TPPP/p25A (NM_001108461)]. The KFERQ-like motif region is in a loop, surrounded by two α-helices, fully exposed to the solvent. (B) Representative immunoblot of recombinant TPPP-p25A in vitro degradation via isolated rat brain-derived lysosomes. Lysosomes were incubated with 0.2 μg of recombinant human TPPP/p25A protein in the absence or the presence (as negative control) of protease inhibitors (PIs). The addition of increasing amounts of recombinant human monomeric SNCA (0.6 μg or 1.2 μg) reveals the competition of the two substrate proteins (SNCA and TPPP/p25A) for CMA targeting. Recombinant HSPA8/HSC70 was used for the substrate translocation to the lysosomes. The presence of purified lysosomes was verified by the detection of LAMP2A, LAMP1, ACTB and CTSD (cathepsin D; loading control) protein levels. (C) Quantification of the levels of human recombinant TPPP/p25A protein upon its degradation via purified rat brain-derived lysosomes in the presence or the absence of PIs or recombinant human monomeric SNCA. Data are expressed as the mean ± SE of three independent experiments; *p < 0.05, by one-way ANOVA with Tukey’s post hoc test.

Both the exogenously added human SNCA (HsSNCA PFFs) and the recruited endogenous rat oligodendroglial SNCA are partly degraded via the autophagy-lysosome pathway, without impairing lysosomal function

We subsequently assessed the proteolytic pathways responsible for SNCA and TPPP/p25A clearance, as well as the role of the overexpressed human SNCA and TPPP/p25A on the proteasomal and lysosomal function, under conditions of increased SNCA protein burden, thus mimicking the human MSA. To this end, OLN cells were inoculated with 1 μg/ml HsSNCA PFFs for 48 h and 16 h prior to cell-fixation, lysosomal or proteasomal inhibitors were added to the medium and then fixed cells were stained for immunofluorescence analysis. We first verified that incubation of all OLN cells with 1 μg/ml HsSNCA PFFs for 48 h or 10 days does not affect cell survival (Figure S4G-H). The time points for proteasomal or lysosomal inhibition were selected according to the data obtained from a thorough time-course analysis (0, 6, 16, 24, 48 h) of PFF-treated OLN cells (for 48 h or 10 days) with the pharmacological inhibitors NH₄Cl, 3 MA or Epox (Figure S5). Epoxomicin was used only for the assessment of the endogenous rodent SNCA degradation, since, as already mentioned, it evokes a CMV-dependent upregulation of the human SNCA and TPPP/p25α mRNA levels. Confocal microscopy analysis revealed that in PFF-treated OLN-93 cells, inhibition of total lysosomal (NH₄Cl) or proteasomal (epox) function leads to a significant increase of both the exogenously added human (green, LB509 antibody) and the recruited endogenous rodent (red, D37A6 antibody) SNCA (Figure 7A). Interestingly, macroautophagy (inhibited by 3 MA) seems to participate, albeit to a lesser extent, only in the degradation of HsSNCA PFFs and not to the seeded rodent SNCA. However, it should be stressed that if the inhibitors have an effect on the exogenously added human SNCA, this could alter the levels of the endogenous seeded protein through an indirect effect and, vice-versa, alterations in the seeded material may serve to stabilize the human PFFs.

Figure 7. Both the exogenously added human SNCA (HsSNCA PFFs) and the recruited endogenous rat oligodendroglial SNCA are partly degraded via the autophagy-lysosome pathway, without impairing lysosomal function. (A-C) Representative immunofluorescence images of OLN-93 (A), OLN-AS7 (B) and OLN-p25α (C) cells treated with NH₄Cl (20 mM), 3 MA (10 mM) or epoxomicin (epox, 15 nM) for 16 h following their incubation with 1 μg/ml HsSNCA PFFs for 32 h. (D-F) Quantification of the endogenous rodent SNCA (D), human SNCA (E) or TPPP/p25A (F) protein levels in OLN-93, OLN-AS7 and OLN-p25α cells measured as μm² area surface/cell following their treatment with 1 μg/ml HsSNCA PFFs (32 h) and the proteasome or lysosome inhibitors (16 h). Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; **p < 0.01, by one-way ANOVA with Tukey’s post hoc test (to compare between inhibitor-treated and untreated cells) or #p < 0.05; ##p < 0.01 by two-way ANOVA with Bonferroni’s correction (to compare between the different PFF-treated cell cultures). (G) Representative immunoblots of the UREA-soluble protein lysates of OLN-93 cells treated with 1 μg/ml HsSNCA PFFs (32 h) followed by their incubation with the pharmacological inhibitors epox or NH₄Cl for 16 h. Both human (4B12 antibody) and total (endogenous + human) SNCA (C20 antibody) seem to accumulate in the UREA-soluble fraction of OLN-93 cells treated with epox or NH₄Cl following incubation with 1 μg/ml HsSNCA PFFs. Antibodies against LC3-I and -II, and SQSTM1 were used as macroautophagy markers and ACTB as a loading control. (H) Quantification of monomeric and High Molecular Weight (HMW) species of human (Hi and Hii) and total (Hiii and Hiv) SNCA levels detected in the UREA-soluble fraction of OLN-93 cells treated with 1 μg/ml PFFs and epox or NH₄Cl for a total of 48 h. Data are expressed as the mean ± SE of four independent experiments; *p < 0.05; **p < 0.01, by one-way ANOVA with Tukey’s post hoc test. (I) Alterations in lysosomal degradation pathways of OLN cell lines following their treatment with 1 μg/ml HsSNCA PFFs for a total of 96 h. OLN cells incubated with HsSNCA PFFs for 48 h were labeled with [³H] leucine for another 48 h (2 µCi/ml). 16 h prior to media collection, cells were treated with or without NH₄Cl (20 mM) or 3 MA (10 mM) and degraded proteins were assayed. The rate of total (Ii) (inhibitable by NH₄Cl), of macroautophagic (Iii) (inhibitable by 3 MA) and of CMA-dependent (Iiii) long-lived protein degradation in OLN cells is shown. Data are expressed as the mean ± SE of three independent experiments; *p < 0.05; **p < 0.01, by one-way ANOVA with Tukey’s post hoc test (to compare between PBS- and HsSNCA PFFs-treated cells) or #p < 0.05; ##p < 0.01 by two-way ANOVA with Bonferroni’s correction (to compare between the different OLN cells).

Similarly, in PFF-treated OLN-AS7 cells, treatment with epox significantly increased the levels of the recruited endogenous SNCA, whereas total lysosomal inhibition resulted in the accumulation of both rodent and exogenously added fibrillar human SNCA (Figure 7B). Interestingly, in PFF-treated OLN-p25α cells levels of both rodent and human SNCA did not significantly change upon NH₄Cl or 3 MA addition, however when the proteasome was inhibited, the levels of the endogenous SNCA were found elevated (Figure 7C). This could indicate a role of the TPPP/p25A in the formation of highly insoluble aberrant SNCA species, as we have previously reported [11] that probably are degradation-resistant and/or impair lysosomal activity. Quantification of the rodent (Figure 7D) and human (Figure 7E) SNCA-positive signal measured as μm² area/cell in all PFF-treated OLN cells cemented the contribution of the ALP in particular in SNCA clearance. Interestingly, TPPP/p25A protein levels (gray) seem to be slightly, but not significantly, increased upon lysosomal inhibition (Figure 7F), which could again presumably be attributed to a potential inhibitory effect on the lysosome due to the concurrent increased SNCA load and TPPP/p25A overexpression or to a possible degradation of TPPP/p25A via the proteasome.

To further verify the contribution of the proteasome and the lysosome in the clearance of aberrant SNCA species engendered in the PFF-treated OLN cells, we used fractionated Western immunoblotting; we treated OLN-93 cells with 1 μg/ml HsSNCA PFFs followed by addition of NH₄Cl or epox as above. Cells were collected at 48 h and sequentially fractionated using buffers with increasing extraction strength. Immunoblot analysis of the SDS-soluble protein fraction did not reveal any differences in the protein levels of human or total (rodent + human) SNCA upon lysosomal or proteasomal inhibition (data not shown). However, addition of NH₄Cl or epox in PFF-treated OLN-93 cells, led to the detection of increased protein levels of both human (4B12 antibody) and total (sc7011, C20 antibody) monomeric and high molecular weight (HMW) SNCA species in the UREA-soluble fraction (Figure 7G-H). Interestingly, HsSNCA monomeric and HMW SNCA protein levels accumulated more upon epox treatment, whereas the respective total SNCA levels, containing also the endogenous rodent protein, accrued mainly upon NH₄Cl treatment, suggesting perhaps a distinct conformation of the engendered rodent SNCA seeded material that resists proteasomal degradation. Finally, the proteasome seems to also contribute to the clearance of the recruited endogenous rodent oligodendroglial SNCA and TPPP/p25A proteins upon PFF-addition, albeit to a different extent, as detected following molecular inhibition of the proteasome using the Psmb5 siRNAs (Figure S6A-F).

Finally, to elucidate the impact of HsSNCA PFFs on the proteasomal and lysosomal function of all OLN cell lines and the effect, if any, of the overexpressed SNCA or TPPP/p25A on these proteolytic systems, we measured the CT proteasomal activity and the degradation rate of long-lived proteins (total lysosomal, macroautophagy- and CMA- dependent), using well-established methods to selectively monitor the activity of the proteasome [34] and the lysosome [35], respectively (see Materials and Methods section for a detailed description of these assays). Strikingly, the CT-like proteasomal activity remained unchanged in all OLN cells upon treatment with 1 μg/ml HsSNCA PFFs for 48 h (Figure S6G). On the other hand, long-lived protein degradation assay revealed a significant increase of the lysosomal activity upon inoculation of OLN cells with HsSNCA PFFs, which however varied between the different OLN lines. Specifically, in PFF-treated OLN-93 cells, macroautophagy (3 MA-inhibitable) seems to increase in response to the treatment, whereas in PFF-treated OLN-AS7 and OLN-p25α cells, CMA-dependent (NH₄Cl-3 MA inhibitable) degradation is significantly enhanced (Figure 7Ii-iii). It has to be noted that with this assay we estimate CMA activity by subtracting NH₄Cl-3 MA–dependent proteolysis that also contains the contribution of microautophagy, which however is considered to be relatively small. With LysoTracker red staining we observed an increase in the perinuclear re-localization of lysosomes upon addition of HsSNCA PFFs similar to the effect evoked by the lysosomal inhibitor NH₄Cl (Figure S7), which may represent a compensatory response of the cell to counteract the impaired autophagic flux (see below). Importantly, the alterations in the activity of macroautophagy upon PFF-treatment did not seem to be attributed to alterations in the expression levels of macroautophagy-related proteins involved in the early stages of macroautophagy, such as ULK1, BECN1, MTOR or p-RPS6 (Figure S8A-B). Moreover, the observed induction of the CMA activity measured by the long-lived degradation assay in the PFF-treated OLN-AS7 and OLN-p25α cells (Figure 7Iiii) was further supported by the increased LAMP2A protein levels detected in these cells, as compared to the OLN-93 cells (Figure S8C-D). Interestingly, the increase of CMA-dependent proteolysis in OLN-p25α cells was accompanied by a significant decrease of macroautophagic activity, thus leading to overall unchanged total lysosomal activity levels, compared to OLN-93 and OLN-AS7 cells. It is important, though, to note that all the above differential responses could also be attributed to clonal variability between the three OLN cell lines.

TPPP/p25A overexpression favors the degradation of both exogenously added (HsSNCA PFFs) and recruited endogenous oligodendroglial SNCA via CMA and not via macroautophagy

To further illuminate the contribution of macroautophagy and/or CMA in the degradation of rodent SNCA, human SNCA and TPPP/p25A in the context of MSA, we transfected all OLN cells with Lsi1 and Lsi2 (targeting Lamp2a) or Atg5 (targeting Atg5) siRNAs followed by incubation with 1 μg/ml HsSNCA PFFs for 48 h. According to the data shown in Figure 8, LAMP2A downregulation increased protein levels of the recruited rodent SNCA in all PFF-treated OLN cells; however, the levels of the human SNCA were found elevated only in PFF-treated OLN-p25α cells transfected with Lsi1 and Lsi2 siRNAs (Figure 8A-C, Di, Ei). On the other hand, Atg5 gene silencing led to the accumulation of the rodent and the human SNCA only in PFF-treated OLN-93 and OLN-AS7 cells (Figure 8A-C, Dii, Eii). Moreover, protein levels of TPPP/p25A were significantly accumulated upon LAMP2A downregulation, whereas transfection of OLN-p25α cells with Atg5 siRNA led to a slight, non-significant, increase of TPPP/p25A protein levels (Figure 8C, Fi, Fii). These data, when combined to those presented in Figure 7A-F may lead to the hypothesis that under pathological conditions (addition of HsSNCA PFFs), CMA seems to be the main pathway responsible for the clearance of SNCA (rodent and human) and TPPP/p25A in OLN-p25α cells, whereas both CMA and macroautophagy contribute to the degradation of SNCA (rodent and human) in PFF-treated OLN-93 and OLN-AS7 cells.

Figure 8. TPPP/p25A overexpression favors the degradation of both exogenously added (HsSNCA PFFs) and recruited endogenous oligodendroglial SNCA via CMA and not via macroautophagy. (A-C) Representative immunofluorescence images of OLN-93, OLN-AS7 and OLN-p25α cells incubated with 1 μg/ml HsSNCA PFFs added at 24 h following transfection with scrambled RNA (as negative control) and Lamp2a- (Lsi1 and Lsi2, 60 nM) or Atg5-siRNAs (Atg5 si, 10 nM) for another 48 h, using antibodies against the endogenous rodent SNCA (red, D37A6 antibody), human SNCA (green, LB509 antibody) and TPPP/p25A (gray) and DAPI staining. Scale bar: 25 μm. (D-F) Quantification of the endogenous rodent SNCA (Di, Dii), human SNCA (Ei-Eii) or TPPP/p25A (Fi-Fii) protein levels in OLN-93, OLN-AS7 and OLN-p25α cells measured as μm² area surface/cell following transfection with Lsi1 and Lsi2 or Atg5 siRNAs and incubation with HsSNCA PFFs added at 24 h post-transfection for an additional 48 h treatment. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; **p < 0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test.

Addition of HsSNCA PFFs in OLN cells does not impair autophagosome formation, but seems to interfere with the fusion of autophagosomes with the lysosome

Autophagy is a dynamic process that includes the autophagosome formation, maturation and fusion with the lysosomes. In order to monitor the autophagic flux in PFF-treated OLN cells we utilized the GFP/RFP-LC3 and GFP/mcherry-SQSTM1 tandem fluorescent-tagged LC3 and SQSTM1/p62, respectively, which allows the dynamic visualization of the formation of autophagosomes and/or autolysosomes. Specifically, OLN cells were transfected with GFP/RFP-LC3 or GFP/mcherry-SQSTM1 cDNAs and 6 h later, 1 μg/ml HsSNCA PFFs were added to the medium for 48 h. Based on the fact that the fluorescence of GFP, contrarily to the mRFP or mcherry, is quenched in an acidic environment [36,37] autophagy inhibition results in a decrease of red puncta followed by an increase of green puncta, indicative of the low autolysosome formation. Incubation of all OLN cells with HsSNCA PFFs impaired autophagic flux, due to the detection of a lower number of red puncta/cell (in both GFP/RFP-LC3 and GFP/mcherry-SQSTM1), as presented in Figure 9A-B and Figure S9A-D. The measurement of the integrated intensity of puncta in different channels (GFP and RFP or mcherry) further supported an impairment in the autophagic flux upon addition of PFFs in all OLN cells, where the high intensity values from the GFP signal were accompanied by a decrease in the RFP or mcherry fluorescence for both GFP/RFP-LC3 and GFP/mcherry-SQSTM1, respectively (Figure S9E-H). Moreover, calculation of GFP:RFP (Lc3 cDNA) or GFP:mcherry (Sqstm1 cDNA) ratio is indicative for the autophagic flux process; the ratio increases when autophagic flux is low and decreases when autophagic flux is enhanced. According to our results, autophagic flux was inhibited in all PFF-treated OLN cells, with no differences detected amongst the different cell lines (Figure 9C-D). Immunoblot analyses of the SQSTM1 and LC3-II:I protein levels in control or PFF-treated OLN cells in the presence/absence of NH₄Cl revealed a statistically significant increase of LC3-II levels in OLN-AS7 and OLN-p25α PFF-NH₄Cl-treated cells and a non-significant induction of SQSTM1 levels, indicative of a late stage autophagy inhibition resulting in an overall decreased autophagic flux (Figure S10).

Figure 9. Addition of HsSNCA PFFs does not impair autophagosome formation, but seems to interfere with the fusion of autophagosomes with the lysosome in all OLN cell lines. (A-B) Representative immunofluorescence images of OLN-93, OLN-AS7 and OLN-p25α cells transfected with GFP/RFP-LC3 (A) or GFP/mcherry-SQSTM1 (B) constructs for 48 h. Cells were incubated with PBS (as control) or with 1 μg/ml HsSNCA PFFs 6 h post-transfection with the fluorescent constructs and the autophagic flux was assessed via confocal microscopy. Treatment of OLN cells with NH₄Cl (20 mM) for 16 h was used as a positive control for the inhibition of lysosomal function. DAPI is used as a nuclear marker. Scale bar: 25 μm. (C-D) Calculation of the GFP:RFP (in case of LC3 cDNA) or GFP:mcherry (in case of SQSTM1 cDNA) fluorescence ratio as an estimation of autophagic flux. The ratio increases when autophagic flux is low, as presented in the graphs. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test.

Pharmacological enhancement of CMA (AR7) or macroautophagy (rapamycin) decreases overexpressed human SNCA (OLN-AS7 cells) and TPPP/p25A (OLN-p25α cells) protein levels under basal conditions

Having established the lysosomal contribution to the clearance of both SNCA and TPPP/p25A proteins, we investigated the potential therapeutic potential of enhancing macroautophagy (rapamycin, 1 μM) or CMA (AR7, 40 μM), under basal conditions and upon HsSNCA PFF treatment (see below). We initially verified the induction of CMA activity upon AR7 addition (40 μM, 16 h) with confocal microscopy, where we detected increased LAMP2A-positive lysosomes to the perinuclear region of OLN-93 cells, an indirect indicator of increased CMA activity [38] (Figure S11A-B). Incubation of OLN-AS7 and OLN-p25α cells with either one of the two enhancers for 48 h under basal conditions led to a significant decrease of human SNCA (human-specific LB509 or 4B12 antibodies), and TPPP/p25A levels respectively, verified by both confocal microscopy imaging and Western immunoblotting analyses (Figure 10).

Figure 10. Pharmacological enhancement of CMA (AR7) or macroautophagy (Rapamycin) decreases overexpressed human SNCA (OLN-AS7 cells) and TPPP/p25A (OLN-p25α cells) protein levels under basal conditions. (A-B) Confocal microscopy with antibodies against human SNCA (A, red, LB509 antibody) and TPPP/p25A (B, red) reveals the enhanced degradation of these proteins upon treatment of OLN-AS7 and OLN-p25α cells with 40 μM AR7 (CMA enhancer) or 1 μM rapamycin (rap, macroautophagy enhancer) for 48 h. TUBA is used as a cytoskeletal marker (green) and DAPI as a nuclear marker. Scale bar: 25 μm. (C-D) Quantification of human SNCA (C) and TPPP/p25A (D) protein levels in OLN-AS7 and OLN-p25α cells respectively, measured as M.F.I./cell following treatment with AR7 or rap for 48 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; **p < 0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test. (Ei, Fi) Representative immunoblots of protein cell lysates derived from OLN-AS7 (Ei) and OLN-p25α cells (Eii) demonstrating the decline of human SNCA (4B12 antibody) and TPPP/p25A protein levels upon treatment of cells with AR7 or rap for 48 h. Antibodies against LC3-I and -II, and SQSTM1 were used as macroautophagy markers and ACTB as a loading control. (Eii, Fii) Quantification of human SNCA (Eii) and TPPP/p25A (Fii) protein levels vs ACTB in OLN-AS7 and OLN-p25α cells, respectively, treated with AR7 or rap for 48 h. Data are expressed as the mean ± SE of three independent experiments; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test.

Pharmacological augmentation of CMA or macroautophagy accelerates the clearance of aberrant SNCA conformations formed upon treatment of OLN cells with HsSNCA PFFs

To assess the therapeutic potential of macroautophagy or CMA enhancement on the clearance of pathological SNCA assemblies and/or TPPP/p25A following addition of HsSNCA PFFs, OLN cells were incubated with 1 μg/ml HsSNCA PFFs for 48 h or 10 days. Following PFF-addition and 16 or 48 h prior to cell fixation, rapamycin (1 μM) or AR7 (40 μM) were added to the cell medium and then cells were processed for confocal microscopy imaging. Immunofluorescence analysis revealed that addition of rap or AR7 in OLN cells treated for 16 h with PFFs evoked a reduction of the endogenous seeded rodent and human SNCA, as well as of the overexpressed TPPP/p25A (in OLN-p25α cells) protein levels, suggesting that these proteins/conformations can be efficiently cleared via macroautophagy and/or CMA pathways (Figure 11A-Ei). Interestingly, at 48 h post-PFF treatment the levels of oxidized/nitrated SNCA detected by the specific Syn303 antibody [39] were significantly decreased upon macroautophagy induction with rapamycin only in OLN-93 cells (Figure 11Eii and Figure S12A-C). Likewise, CMA induction via AR7 also decreased oxidized/nitrated SNCA protein species only in OLN-93 cells, although this drop did not reach statistical significance. Furthermore, protein levels of aggregated (detected by the conformation-specific antibody MJFR-14 [40],) and total SNCA were also decreased upon rap or AR7 (to a lesser extent) treatment in all OLN cells, following short-term incubation with PFFs, suggesting that in this early stage of aggregate formation there is no significant effect of the overexpressed human SNCA or TPPP/p25A in the clearance of aggregated SNCA species (Figure 11Fi-Fii and Figure S12D-F).

Figure 11. Pharmacological augmentation of CMA or macroautophagy accelerate the clearance of aberrant SNCA conformations formed upon treatment of OLN cells with HsSNCA PFFs. (A-C) Representative immunofluorescence images using antibodies against the endogenous rodent SNCA (red, D37A6 antibody), the human SNCA (green, LB509 antibody) and TPPP/p25A (gray) are shown. DAPI is used as nuclear marker. Scale bar: 25 μm. (D-F) Quantification of the recruited endogenous rodent (Di) or the exogenously added human (Dii) SNCA, the overexpressed human TPPP/p25A (in OLN-p25α cells) (Ei), as well as the pathological conformations of oxidized/nitrated (Eii) and aggregated (Fi) SNCA and the total (rodent + human) (Fii) SNCA protein levels measured as μm² area surface/cell in OLN cells treated with 40 μM AR7 or 1 μM rap (16 h) following the addition of 1 μg/ml HsSNCA PFFs for a total of 48 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; **p < 0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test (to compare between enhancer-treated and untreated cells) or #p < 0.05 by two-way ANOVA with Bonferroni’s correction (to compare between the different treated cells). (G) Representative immunoblots of the SDS-soluble SNCA species of control OLN-93 cells incubated with 1 μg/ml HsSNCA PFFs for 32 h followed by addition of AR7 or rap for 16 h using antibodies against HsSNCA (4B12 antibody) or total (rodent + human) SNCA (SYN1 antibody). Equal loading was verified by the detection of ACTB levels. (H) Quantification of monomeric and HMW species of human (Hi and Hii) and total (rodent + human) (Hiii and Hiv) SNCA levels detected in the SDS-soluble fraction of OLN-93 shown in G. Data are expressed as the mean ± SE of four independent experiments; *p < 0.05, by one-way ANOVA with Tukey’s post hoc test. (Ii) Representative immunoblots for TPPP/p25A and ACTB (as loading control) protein levels of total protein lysates of OLN-p25α cells incubated with AR7 or Rap (16 h) following their treatment with 1 μg/ml HsSNCA PFFs (32 h). (Iii) Quantification of the protein levels of TPPP/p25A in PFF-treated OLN-p25α cells incubated with AR7 of Rap. Data are expressed as the mean ± SE of three independent experiments; *p < 0.05, **p < 0.01 by one-way ANOVA with Tukey’s post hoc test. (J) Representative immunoblots of the UREA-soluble protein fraction derived from OLN-93 cells treated with AR7 or rap following HsSNCA PFFs addition. Antibodies against rodent SNCA (D37A6 antibody), HsSNCA (4B12 antibody) or total (rodent + human) SNCA (SYN1 antibody) verified the decrease of the protein levels detected in this fraction upon addition of CMA (AR7) or macroautophagy (rap) enhancers. An antibody against ACTB was used as loading control. (K-M) Quantification of monomeric and HMW species of rodent (Ki and Kii), human (Li and Lii) and total (endogenous + human) (Mi and Mii) SNCA levels detected in the UREA-soluble fraction of OLN-93 cells treated with 1 μg/ml HsSNCA PFFs and the autophagy modulators. Data are expressed as the mean ± SE of four independent experiments; *p < 0.05; **p < 0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test.

In agreement with the confocal imaging data, we found that the SDS-soluble HMW species of human and total SNCA protein were decreased mainly upon macroautophagy enhancement (addition of 1 μM rap for 16 h), although the levels of monomeric SNCA species did not seem to be significantly reduced following incubation of PFF-treated OLN-93 cells either with rap or with AR7 (Figure 11G-H). Interestingly, the monomeric and HMW SNCA species of the seeded endogenous rodent protein, which are detectable only in the UREA-soluble fraction that contains the most aggregated protein species, were significantly reduced following addition of rap or AR7 in PFF-treated OLN-93 cells (Figure 11J-K). Similarly, human and total SNCA protein levels in this UREA-soluble fraction displayed significant reduction upon macroautophagy or CMA enhancement, as presented in Figure 11K, L, M. Finally, addition of AR7 or rap seems to lead to an efficient reduction in the protein levels of TPPP/p25A in PFF-treated OLN-p25α cells (Figure 11Ii-ii).

Upon prolonged incubation (10 days) with PFFs, endogenous and human SNCA and TPPP/p25A (in OLN-p25α cells), as well as oxidized/nitrated SNCA protein levels were efficiently removed upon induction of either macroautophagy or CMA in all OLN cells (Figure S13). Strikingly, at this time point, aggregated and total SNCA could be effectively removed in OLN-93 and OLN-AS7, but not in OLN-p25α cells, thus implying a potential role of the overexpressed TPPP/p25α in the observed resistance to the degradation of aggregated SNCA assemblies via macroautophagy or CMA (Figure S13Dii, Eii, Fii, Ii–Iii).

All the above results unambiguously suggest that SNCA (endogenous, human, total and pathological species) and TPPP/p25A proteins can be effectively cleared via the ALP in an MSA cell context.

The MSA-related proteins TPPP/p25A and SNCA are mainly degraded via the ALP in murine primary oligodendroglial cultures, under physiological and pathological conditions

To delineate the clearance processes responsible for the removal of SNCA (endogenous and exogenously added) and TPPP/p25A, in health and disease conditions we cultivated and differentiated primary oligodendrocytes derived from P0 to P3 mouse brains in order to resemble a cellular setting closer to the oligodendrocytes of the central nervous system. To this end, primary oligodendrocytes were cultivated in the presence of PBS (as control; Gibco, 10,010–015) or 1 μg/ml HsSNCA PFFs for 24 h, followed by addition of lysosome/macroautophagy (20 mM NH₄Cl, 10 mM 3 MA) or proteasome (15 nM epox) inhibitors for an additional 48 h. According to the results presented in Figure 12A-C, TPPP/p25A seems to be mainly degraded via the ALP both under baseline and upon PFF-treatment, since its levels were found significantly elevated upon NH₄Cl (both in PBS and PFF-treated cultures) and/or 3 MA (PFF-treatment) addition (Figure 12Ci-Cii).

Figure 12. The MSA-related proteins TPPP/p25A and SNCA are mainly degraded via the autophagy-lysosome pathway in murine primary oligodendroglial cultures, under physiological and pathological conditions. Mouse primary oligodendroglial cultures were incubated with PBS (as control) or HsSNCA PFFs for 24 h followed by their treatment with lysosomal or proteasomal inhibitors for another 48 h. (A-B) Representative confocal microscopy images depicting the protein levels of TPPP/p25A (red) upon total lysosomal (NH₄Cl, 20 mM), macroautophagic (3 MA, 10 mM) or proteasomal (epox, 15 nM) inhibition (for 48 h) in the absence (A) or presence (B) of 1 μg/ml HsSNCA PFFs for a total of 72 h. The recruited endogenous rodent SNCA (green, D37A6 antibody) and the exogenously added HsSNCA PFFs (gray, LB509 antibody) also seem to be preferentially degraded via the autophagy-lysosome pathway under pathological (B, PFFs-treated) conditions. DAPI staining is used as a nuclear marker. Scale bar: 25 μm. (C) Quantification of TPPP/p25A (Ci, Cii), rodent SNCA (Di, Dii) and human SNCA (E) protein levels in primary oligodendrocytes, measured as μm²/cell following treatment with PBS, as control, or with 1 μg/ml HsSNCA PFFs and in the presence or absence of pharmacological inhibitors (NH₄Cl, 3 MA or epox) for a total of 72 h. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; **p < 0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test. (F-G) Representative immunofluorescence images of TPPP/p25A (red), rodent SNCA (green, D37A6 antibody) and HsSNCA (gray, LB509 antibody) protein levels upon treatment of mouse oligodendrocytes with AR7 (40 μΜ) or rap (1 μΜ) for 48 h, which were added to cells 24 h following addition of PBS, as control, (F) or 1 μg/ml HsSNCA PFFs (G). DAPI is used as a nuclear marker. Scale bar: 25 μm. (H) Quantification of TPPP/p25A (Hi, Hii), rodent SNCA (Hiii) and human SNCA (Hiv) protein levels in primary oligodendrocytes, measured as μm²/cell when cells were treated with PBS, as control, or with 1 μg/ml HsSNCA PFFs for 24 h, followed by addition of CMA or macroautophagy enhancers for an additional 48 h treatment. Data are expressed as the mean ± SE of three independent experiments with duplicate samples/condition within each experiment; *p < 0.05; **p < 0.01; ***p < 0.001, by one-way ANOVA with Tukey’s post hoc test.

Furthermore, the endogenous rodent SNCA accumulated when either the lysosome or the proteasome (to a lesser extent) was inhibited (Figure 12Di-Dii), whereas the levels of human SNCA were increased upon total lysosomal and macroautophagy impairment (Figure 12E). No signal for the endogenous SNCA could be detected in PBS-treated primary oligodendrocytes using the rodent SNCA-specific D73A6 antibody, as shown in Figure 12A and Di.

To elucidate further the potential beneficial role of autophagy augmentation in the removal of SNCA and TPPP/p25A protein levels in primary oligodendrocytes, PBS- or PFF-treated oligodendroglial cultures were incubated with AR7 (40 μM) or rap (1 μM) for 48 h. Confocal microscopy analysis and quantification of SNCA and TPPP/p25A protein levels revealed that in all conditions tested, incubation with either AR7 or rap led to the effective clearance of TPPP/p25A (Figure 12F-G, Hi -Hii), the endogenous rodent (Figure 12G, Hiii) and the human SNCA (Figure 12G, Hiv) proteins, thus suggesting that enhancement of macroautophagy or the CMA pathway could potentially mitigate SNCA pathology in MSA and MSA-like conditions.

Discussion

Our previous work uncovered the endogenous oligodendroglial SNCA (expressed in minute amounts) and TPPP/p25A as major culprits for the formation of pathological SNCA aggregates in MSA-like experimental models [11]. Such data suggest that manipulation of the expression of SNCA and/or TPPP/p25A in oligodendrocytes may provide a rational approach to combat the accumulation of SNCA in GCIs and the progression of MSA. However, the degradation pathways responsible for the clearance of the endogenous oligodendroglial SNCA and TPPP/p25A in health and disease and the role of ALP manipulation in the context of MSA remained unexplored.

Herein, by utilizing naïve (low-undetectable levels of endogenous SNCA and TPPP/p25A) oligodendroglial cell lines, and cells stably overexpressing human SNCA (OLN-AS7) or TPPP/p25A (OLN-p25α) we report that both the ALP and the UPS contribute to the clearance of endogenous oligodendroglial SNCA under basal conditions and upon treatment with HsSNCA PFFs. Both CMA and macroautophagy are responsible for the clearance of the various SNCA conformations (rodent endogenous and seeded, human exogenously added, oxidized/nitrated, aggregated) engendered in PFF-treated OLN cell lines, although to a different extent (Figure 13). We also demonstrate that TPPP/p25A bears a CMA-targeting motif and is efficiently cleared via CMA in a cell-free in vitro system of isolated brain lysosomes and in the OLN-p25α cells. Finally, we provide proof-of-concept experiments in murine primary oligodendrocytes further corroborating the contribution of the ALP in the efficient removal of oligodendroglial SNCA and TPPP/p25A both under baseline and following inoculation with HsSNCA PFFs as seeds of pathology. Such data indicate that augmentation of CMA or macroautophagy represents an attractive therapeutic approach to counteract the accumulation of SNCA and/or TPPP/p25A in MSA-like conditions.

Figure 13. A schematic representation depicting the proteolytic pathways responsible for the clearance of SNCA and TPPP/p25A proteins within oligodendrocytes, under physiological (PBS) and pathological (HsSNCA PFFs) conditions. (A) In control OLN-93 cells both the soluble monomeric and the HsSNCA PFF-recruited endogenous oligodendroglial SNCA (shown in purple) are degraded via all proteolytic pathways (CMA, macroautophagy, proteasome), whereas the human exogenously added SNCA (PFFs, shown in brown) are preferably cleared via macroautophagy and the proteasome. (B) In OLN-AS7 cells, all proteolytic pathways participate in the clearance of both the overexpressed human and the endogenous SNCA (which is degraded via the proteasome to a higher extent) under physiological conditions. Similarly, in the presence of HsSNCA PFFs, the recruited endogenous oligodendroglial SNCA is cleared via both the lysosome (CMA and macroautophagy) and the proteasome, whereas the exogenously added PFFs of human origin are targeted to macroautophagy for degradation. Moreover, the addition of PFFs increases CMA activity, probably due to the increased protein levels of the CMA’s rate limiting step, the LAMP2A receptor, to the lysosomal membrane. (C) In OLN-p25α cells, CMA is mainly responsible for the clearance of all proteins of interest (endogenous SNCA, HsSNCA PFFs and TPPP/p25A) under both physiological and pathological conditions. The elevated levels of the LAMP2A receptor detected in PFF-treated OLN-p25α cells may underlay the accelerated CMA-dependent clearance of these proteins. Moreover, TPPP/p25A (shown in green) and the recruited endogenous SNCA (upon PFF-treatment), as well as the monomeric oligodendroglial SNCA (in PBS conditions) are also targeted to the proteasome for their clearance. The addition of PFFs seems to impair macroautophagic-related activity in OLN-p25α cells, which is otherwise responsible only for the clearance of the soluble oligodendroglial SNCA in basal conditions. (D) A table summarizing the degradation pathways that participate in the clearance of the endogenous oligodendroglial SNCA, the human SNCA (both stably overexpressed and HsSNCA PFFs) and TPPP/p25A proteins in physiology and pathology.

MSA is characterized by the accumulation of neuronal SNCA and oligodendroglial-specific TPPP/p25A proteins within the cytoplasm of oligodendrocytes, by a hitherto unknown mechanism [2–4,6]. According to the prevailing hypothesis, oligodendrocytes internalize the naturally secreted SNCA from the neighboring neurons, which is subsequently incorporated into pathological aggregates along with other proteins such as TPPP/p25A, ubiquitin, tubulin, HSP70, etc [41–43]. The detection of various aggresome-related proteins in GCIs of MSA brains has also exposed a crucial role of perturbed proteolysis in the formation of oligodendroglial proteinaceous inclusions [21,23–25,44]. Alterations in the levels of autophagic protein markers such as SQSTM1/p62 and LC3 detected in several MSA cases further denote a role of an ALP malfunction in disease pathogenesis [20,22,45–47]. Likewise, pathological and biochemical analyses using human brain MSA samples revealed that the Autophagy And Beclin 1 Regulator 1 is a component of the pathological inclusions of MSA and upstream proteins of autophagy are impaired in the MSA brain [48]. A recent comparative study in human postmortem material from MSA and PD brains concluded that the lysosomal response in relation to SNCA pathology differs between the two synucleinopathies [49]. By systematic comparisons of differently affected neuronal populations in PD, MSA, and non-diseased brains using morphometric immunohistochemistry (cathepsin D), double immunolabelling (cathepsin D/SNCA) laser confocal microscopy, and SNCA immunogold electron microscopy the authors concluded, amongst others, that lysosome-associated SNCA is observed in astroglia and rarely in oligodendroglia and in neurons in MSA, whereas cathepsin D immunoreactivity frequently colocalises with SNCA pre-aggregates in PD nigral neurons [49].

The role of the oligodendroglial SNCA in the formation of GCIs and the spread of pathology in MSA is still under debate. Our recently published data suggest that the presence of even minute amounts of the endogenous SNCA is a prerequisite for the seeding of SNCA-pathology, the re-distribution of TPPP/p25A and the collapse of the myelin network to occur, in PFF-treated murine oligodendrocytes [11]. Thus, the elucidation of the mechanisms governing the degradation of oligodendroglial SNCA and TPPP/p25A may represent an obvious target for therapy in MSA. To put this hypothesis under scrutiny, we treated oligodendroglial cell lines with pharmacological inhibitors or enhancers of the ALP and UPS, or with siRNAs targeting autophagy-and proteasome-related genes, under physiological (PBS) or pathological (HsSNCA PFFs) conditions. According to our results, SNCA (endogenous rodent, overexpressed human and pathology-related conformations) seem to be cleared by both the ALP and the proteasome; however to a different magnitude in the various cell lines (Figure 13). A role of macroautophagy in the clearance of SNCA within oligodendrocytes has been previously reported [20]; however this is the first study demonstrating a role of CMA in the removal of the endogenous and seeded SNCA protein levels and/or aberrant species in oligodendrocytes. This finding contradicts with a prior study that reported a lack of GCI-like formation in human oligodendroglia exposed to extracellular soluble/monomeric or fibrillar SNCA concurrently with pharmacological blocking of the fusion of the autophagosome with the lysosome with bafilomycin A₁, as well as following genetic knockdown of LC3B [50]. Such discrepancies may be attributed to the different oligodendroglial cells utilized in the two studies (rat OLN vs human MO 3.13 cells) or the timing of the lysosomal inhibitor application (prior or following to PFF addition).

Regarding macroautophagy, it has been previously shown that the endogenously stably overexpressed (OLN‐t40 transfected cells) or the exogenously added soluble or pre-aggregated SNCA is mainly removed via macroautophagy in oligodendrocytes [51]. Moreover, inhibition of the deubiquitylating enzyme Ubiquitin Carboxy-terminal Hydrolase L1 in oligodendroglial cells was shown to upregulate macroautophagy, thus resulting in the effective removal of SNCA aggregates [52]. Other proteolytic pathways may also be involved in SNCA clearance in oligodendrocytes. KLK6/Neurosin (kallikrein related-peptidase 6) appears to be efficient in reducing the levels of SNCA in oligodendrocytes both in vitro and in vivo [53–55]. Interestingly, treatment of transgenic human SNCA-PLP mice, a well-established mouse model for MSA, with the proteasome inhibitor I led to the detection of intracellular aggregates of both human and endogenous murine SNCA, three months after administration of the inhibitor, further supporting the role of the UPS in SNCA clearance [56]. Furthermore, in agreement with our data showing an impairment of the autophagic flux in PFF-treated OLN cells, prior studies also suggested that aggregated SNCA within oligodendrocytes is closely related to autophagy dysregulation [48,52], although such dysregulation was not identified in another study [48,52]. Interestingly, even though with the GFP/RFP-LC3 and GFP/mcherry-SQSTM1 constructs we detect a late-stage defect in the autophagic flux upon PFF treatment (Figure 9, Figures S9 and S10), this was not followed by a statistically significant alteration in the macroautophagic-dependent proteolysis (inhibited by 3 MA) in OLN-93 and OLN-AS7 cells (Figure 7Iii). A possible explanation for this discrepancy may be derived from the observation that upon PFF treatment we detect a change from a diffuse to a more punctuate pattern, indicative of LC3-II induction and autophagosome formation. If the rate that these autophagosomes are fused with the lysosome is reduced in these cells, then the overall process is as efficient as baseline. In addition, the different time points utilized for the assessment of the autophagic flux (48 h) and lysosomal activity (96 h) may also account for the observed differences.

The current study highlights for the first time a central role of CMA and macroautophagy in the clearance of TPPP/p25A protein in oligodendrocytes, both under physiological and pathological conditions. The contribution of CMA to TPPP/p25A protein degradation was verified by the identification of the KKRFK pentapeptide motif that meets the criteria for a KFERQ-like motif in TPPP/p25A amino acid sequence and its efficient degradation by the in vitro system of isolated rat brain lysosomes. In addition, TPPP/p25A displayed a canonical substrate/pathway relationship, since its levels were increased upon LAMP2A downregulation and by extension CMA inhibition and decreased following pharmacological induction of the CMA pathway with AR7. Induction of macroautophagy also reduced TPPP/p25A levels, both in OLN-p25α cells and in primary oligodendrocytes, but less efficiently. Previous studies proposed that TPPP/p25A is degraded via the proteasome [18,19]; however, our analysis in the Psmb5-siRNA treated OLN-p25α cells or the epoxomicin-treated primary oligodendrocytes did not yield a statistical contribution of the proteasome in TPPP/p25A clearance. Strikingly, the expression of TPPP/p25A protein (in OLN-p25α cells) seems to hinder the macroautophagic activity (inhibited by 3 MA) upon addition of HsSNCA PFFs thus favoring the CMA activity, probably as a counterpoise to the proteolytic dysregulation. These findings are in agreement with recent data obtained using recombinant TPPP/p25A and SNCA proteins published while this study was under review suggesting that TPPP/p25A counteracts SNCA degradation by hindering the autophagy maturation at the stage of LC3B-SQSTM1/p62-derived autophagosome formation and its fusion with lysosome [57]. Likewise, Ejlerskov and et al. have also demonstrated that TPPP/p25A inhibits the fusion of autophagosomes with the lysosome and leads to the secretion of monomeric and aggregated SNCA via exophagy in PC12 cells [58].

Finally, the pharmacological CMA- and macroautophagy-specific enhancers AR7 and rapamycin respectively, mitigated the levels of the endogenous and exogenously added SNCA and all its pathological conformations (oxidized/nitrated and aggregated SNCA), together with TPPP/p25A, both in oligodendroglial cell lines and primary oligodendrocytes. The beneficial effect of macroautophagy induction in the removal of SNCA species in oligodendrocytes has been previously reported. In particular, the geldanamycin analog 17-AAG [17-(Allylamino)-17-demethoxygeldanamycin] attenuated the formation of SNCA aggregates in OLN-93 cells stably overexpressing the human PD-linked A53T SNCA mutation by stimulating macroautophagy [59]. This effect on macroautophagy induction was also observed in cultured oligodendrocytes derived from the brains of newborn rats [59]. On the other hand, numerous studies have explored the use of various autophagy-enhancing agents in the context of PD pathology, concluding that they exert beneficial effects on neuronal cell survival and accelerate SNCA clearance [60–68]. Moreover, the post-translational regulation of autophagy via the use of micro-RNAs targeting LAMP2A and HSPA8/HSC70, has been shown to decrease SNCA aggregation in SH-SY5Y cells [69]. Similarly, we have previously shown that CMA induction via overexpression of the LAMP2A receptor in neuronal cellular and animal synucleinopathy models alleviated SNCA-induced neurotoxic effects [70].

Collectively, our study reveals that the endogenous oligodendroglial SNCA and TPPP/p25A, the two main GCI-components involved in MSA pathogenesis, are degraded via the ALP and that conversely, the presence of pathological SNCA (HsSNCA PFFs) decreases the autophagic flux of OLN cells, albeit without impairing the overall lysosomal activity. The data obtained from cell lines and primary cultures compellingly suggest that enhancement of CMA or macroautophagy prevents the accumulation/aggregation of SNCA and/or TPPP/p25A in oligodendrocytes. Further validation in pre-clinical models of the disease may pave the way for the use of autophagy modulators or gene-based approaches as therapeutic approaches to halt or attenuate disease progression in human MSA.

Materials and methods

Cell culture and treatments

Three different oligodendroglial (OLN) cell lines have been utilized: the immortalized control OLN-93 line originated from primary Wistar rat brain glial cultures [71], and the OLN-AS7 and OLN-p25α lines that were generated by transduction of the OLN-93 line with the human wild-type (WT) SNCA or human TPPP/p25A cDNA, respectively. All cells were cultured in Dulbecco’s modified Eagle’s medium (Gibco, Invitrogen, D6429) under conditions of 10% fetal bovine serum (Gibco, Invitrogen, 10,270), 50 U/mL penicillin, and 50 μg/mL streptomycin. For the selection of OLN-AS7 and OLN-p25α, cells were maintained in 50 μg/mL Zeocin (Thermo Fisher Scientific, R25001).

Primary oligodendroglial cultures

Mixed glial cultures generated from P0 to P3 neonatal wild-type (WT) mice were maintained in full DMEM for 10 to 14 days until a monolayer of astrocytes on the bottom and primary oligodendroglial progenitor cells (OPCs) with loosely attached microglia on the top, were apparent. The separation of OPCs was achieved initially with the removal of microglia, by shaking in 200 rpm for 1 h at 37°C and then with continuous shaking under the same conditions for 18 h [72]. Afterward, isolated cells were platted on poly-D-lysine-coated coverslips (Sigma-Aldrich, P7405) with a density of 80,000 cells/mm² and maintained in SATO medium [73] supplemented with Insulin-Transferrin-Selenium solution (Gibco, Invitrogen, 41,400,045), 1% penicillin-streptomycin and 1% horse serum (Sigma-Aldrich, H1138) for 4 days. Human SNCA PFFs (final concentration 1 µg/mL culture medium/well) were added to TPPP/p25A-positive mature differentiated OLNs for 24 h, followed by the addition of either inhibitors or enhancers of the protein degradation pathways for 48 h in the appropriate concentrations. Subsequently, cells were fixed and processed for immunofluorescence analysis using antibodies shown in Table 1.

Table 1. A table depicting the primary and secondary antibodies used either in Western blot or immunocytochemistry including species reactivity and working dilutions.

Primary Antibody against /relative reference	Clone	Catalog number	Working dilutions (WB)	Working dilutions (ICC)	Species Reactivity	Company
aggregated SNCA [40]	MJFR 14-6-4-2	ab209538	-	1:1000	rabbit	Abcam
human SNCA [87–89]	LB509	807,701	-	1:1000	mouse	Biolegend
	Syn211	36–008	-	1:2000	mouse	Millipore
	4B12		1:1000	-	mouse	GeneTex
oxidized/ nitrated SNCA [39]	SYN303	824,301	-	1:1000	mouse	Biolegend
rodent SNCA	D37A6	4179	1:1000	1:400	rabbit	Cell Signaling Technology
total SNCA	C20	sc7011R	1:1000	1:1000	rabbit	Santa Cruz Biotechnology
	SYN1	610,786	-	1:1000	mouse	BD Transduction Laboratories
TUBA		62,004	-	1:750	mouse	Invitrogen
ACTB		TA309077	1:2000	-	mouse	OriGene
LC3 LC3B		PM036 NB100-2220	1:1000	-1:200	rabbit rabbit	MBL Novus
TPPP/p25A			1:1000	1:400	rabbit/rat
SQSTM1/ p62		PM045	1:1000	-	rabbit	MBL
Ubiquitin		Z0458	1:1000	-	rabbit	DAKO
Phospho-MTOR	D9C2	5536	1:1000	-	rabbit	Cell Signaling Technology
Phospho-ULK	D1H4	5869	1:1000	-	rabbit	Cell Signaling Technology
Total ULK	D8H5	8054	1:1000	-	rabbit	Cell Signaling Technology
BECN1	H-300	sc-11,427	1:1000	-	rabbit	Santa Cruz Biotechnology
Phospho-RPS6	D57.2.2E	4858	1:1000	-	rabbit	Cell Signaling Technology
Total RPS6	5G10	2217	1:1000	-	rabbit	Cell Signaling Technology
LAMP2A		51–2200	1:1000	1:1000	rabbit	Invitrogen
HSPA8/HSC70		Ab1427	1:1000	-	rabbit	abcam
ATG5		10,181-2-AP	-	1:400	rabbit	proteintech
Secondary antibody		Catalog number		Working dilutions (ICC or WB)	Species Reactivity	Company
Goat anti-mouse IgG-HRP conjugated		AP124P		1:10,000	mouse	Sigma-Aldrich
Goat anti-rabbit IgG-HRP conjugated		AP132P		1:10,000	rabbit	Sigma-Aldrich
CF555 red		20,033		1:2000	rabbit	Biotium
		20,030		1:2000	mouse	Biotium
		20,096		1:2000	rat	Biotium
CF488A green		20,012		1:2000	rabbit	Biotium
		20,010		1:2000	mouse	Biotium
		115–175-146		1:400	mouse	Jackson Imm. Affinipure
Cy5		115–175-144		1:400	rabbit	Jackson Imm. Affinipure
Dyes		Catalog number		Working concentration (ICC)		Company
DAPI		10,236,276,001		1 μg/ml		Sigma- Aldrich
Wheat Germ Agglutinin (WGA)		W11262		1 μg/ml		Thermo Scientific
LysoTracker Red DND-99		L7528		75 nM		Invitrogen

Pharmacological reagents

The inhibition of the proteasome was achieved with epoxomicin (15 nM; Calbiochem, 324,800) that selectively blocks the chymotrypsin-like (CT-like) activity of the 20S catalytic subunit [74]. For the lysosomal pathway, the inhibitor of phosphatidylinositol-3-kinase, 3-methyladenine (3 MA, 10 mM; Sigma-Aldrich, M9281) and the general lysosomal inhibitor NH₄Cl (20 mM) were utilized. Because NH₄Cl is unstable, for the 24 and 48 h incubation a renewal of the inhibitor was required every 12 h. Induction of macroautophagy was accomplished with the use of rapamycin (1 μΜ; R0395, Sigma-Aldrich) that induces the MTOR-dependent macroautophagic pathway [75] and of the CMA pathway with the use of the RARA/RARα antagonist AR7 (40 μM; Sigma-Aldrich, SML0921) [76]. The reagents were applied for 16–48 h and the analysis of their effects was performed either by Western blotting or with immunocytochemistry.

Preparation of HsSNCA PFFs

Pre-formed human SNCA fibrils (HsSNCA PFFs) were generated as previously described [11]. PFFs were resuspended in PBS (pH 7.3) to obtain a concentration of 4.5 mg/mL and a working stock solution was prepared with a concentration of 1 mg/mL. In all experiments cells were incubated with 1 μg/ml HsSNCA PFFs or PBS as a control for 48 h or 10 days and afterward were either processed for immunocytochemistry and confocal microscopy or lysed and collected for Western blot analysis as described below.

Cell transfections

The autophagic flux of OLN cells upon treatment with HsSNCA PFFs (or PBS as control) was monitored with the use of RFP/GFP-LC3 [77], and mCherry/GFP-SQSTM1 [78] cDNA plasmids, upon transient transfection with polyethylenimine (PEI; Polysciences, 26,966). The RFP/GFP-LC3 plasmid was kindly provided by Dr Craig-Curtis Garner (DZNE, Germany) and the mCherry/GFP-SQSTM1 plasmid by Dr Terje Johansen (University of Tromso, Norway). Cells were cultured in 24-well dishes and PEI with cDNA were diluted in Opti-MEM (Gibco, 31,985–047) in a 1:3 ratio. After 4 h of incubation, the transfection medium was replaced by fresh DMEM 10% FBS containing either PBS or 1 μg/ml HsSNCA PFFs. The lysosomal inhibitor NH₄Cl was used as a positive control for autophagic flux impairment. Forty-eight (48) hours later, cells were fixed using 4% paraformaldehyde (Sigma-Aldrich, P6148) in PBS and processed for immunofluorescence analysis.

RNA interference

Small interfering RNAs (siRNAs) targeting the rat Lamp2a or the Atg5 gene were utilized to assess the contribution of CMA or macroautophagy, respectively, in the proteolysis of SNCA and TPPP/p25A in OLN cell lines. In particular, OLN cells were transfected with Lsi1 and Lsi2 (Lamp2a siRNAs, final concentration 60 nM), Atg5 siRNAs (rn.Ri.Atg5.13.1/2/3 TriFECTa DsiRNA kit, final concentration 10 nM), Psmb5 siRNAs (Psmb5 Rat siRNA Oligo Duplex, Origene, SR512544; final concentration 10 nM) or control scrambled siRNA (MISSION siRNA Universal Negative Control 1, SIC001, MERCK, final concentration 60 nM and SR30004, Universal scrambled negative control siRNA duplex, Origene, final concentration 10 nM; SR30004) in Lipofectamine 2000 (Invitrogen, 11,668,019)-containing solution for 72 h. The siRNAs sequences targeting the rat Lamp2a and Psmb5 are shown in Table 2. Analysis was performed with both Western blotting and immunocytochemistry.

Table 2. A table depicting the sequences of the siRNAs targeting the rat Lamp2a and the rat Psmb5, as well as the human SNCA, TPPP/p25α and rat Gapdh primer sequences utilized in the RT-PCR.

Target name	siRNA design
Rat Lamp2a siRNA sense (1) (Lsi1)	CCAUCAUACUGGAUAUGAGdTdT
Rat Lamp2a siRNA antisense (1) (Lsi1)	CUCAUAUCCAGUAUGAUGGdTdT
Rat Lamp2a siRNA sense (2) (Lsi2)	GGUCUCAAGCGCCAUCAUAdTdT
Rat Lamp2a siRNA antisense (2) (Lsi2)	UAUGAUGGCGCUUGAGACCdTdT
Rat Psmb5 (SR512544A/B/C)	1. rCrUrGrArCrUrUrArCrArUrGrArCrArArGrUrArUrArCrUTC 2. rCrArArCrArUrGrGrUrGrUrArUrCrArGrUrArCrArArArGGC 3. rArGrUrGrArUrArGrArGrArUrCrArArCrCrCrUrUrArCrCTT
Oligo name	Primer sequence (5΄→3΄)
huSNCA-Rev	CACCACACTGTCGTCGAATGG
huSNCA-For	CGCCTTGCCTTCAAGCCTTC
huTPPP/p25α-Rev (1)	CCGGACACATAGCCTGACTC
huTPPP/p25α-For (1)	GGGGTGACGAAAGCCATCTC
huTPPP/p25α-Rev (2)	CGAGATGGCTTTCGTCACCC
huTPPP/p25α-For (2)	AAGTCTTGCCGGACCATCAC
Rat Gapdh-Rev	TTCAGCTCTGGGATGACCTT
Rat Gapdh-For	TGCCACTCAGAAGACTGTGG

Subcellular fractionation and Western immunoblotting

The cell pellets after being washed with PBS were homogenized either in lysis buffers with progressively higher extraction strength or in RIPA buffer. All of them contained protease (Roche, 11,836,170,001) and phosphatase (Roche, 04406837001) inhibitors. For the subcellular fractionation, initially, cells were lysed with 1% Triton X-100 (Applichem, A1388)-containing buffer (150 mM NaCl, 50 mM Tris, pH 7.6, 2 mM EDTA), left on ice for 30 min and centrifuged at 13,400 × g for 30 min at 4°C. The supernatant was collected to obtain the Triton-soluble fraction and the pellet, after 2x washes with PBS, was resuspended in 1% SDS-containing buffer (150 mM NaCl, 50 mM Tris, pH 7.6, 2 mM EDTA), sonicated and centrifuged to acquire the SDS-soluble fraction. Finally, the remaining pellet, after 2x washes with PBS, was solubilized in 8 M urea-5% SDS-containing buffer. Then, samples of equal protein concentration were processed for Western blot analysis utilizing the utilized primary and secondary antibodies shown in Table 1. The immunoreactivity for protein band intensity within the linear range of detection was quantified with ImageJ software. All measurements were normalized to ACTB as a loading control.

Immunocytochemistry (ICC) and confocal microscopy

Cultured cells in poly-D-lysine-coated glass coverslips were fixed in 4% paraformaldehyde diluted in 1X PBS for 30 min at room temperature. Blocking was conducted utilizing 10% normal goat serum (Sigma-Aldrich, G6767) solution containing 0.4% Triton X-100 for 1 h at room temperature. To stain the lectins of the plasma membrane or the lysosomes, wheat germ agglutinin (Alexa Fluor™ 594-conjugated; Thermo Fisher Scientific, W11262; final concentration, 1 μg/mL) or LysoTracker Red DND-99 (Invitrogen, L7528; final concentration 75 nM) was added to OLN cells for 10 min at 37°C, respectively, prior to fixation for confocal microscopy analysis. Afterward, overnight incubation with primary antibodies at 4°C was performed followed by incubation with secondary antibodies, in dilutions shown in Table 1. The dye 4΄, 6-diamidine-2΄-phenylindole dihydrochloride (DAPI) was utilized for nuclei staining. Finally, the stained cells were visualized using a Leica TCS SP5 confocal microscope combined with dual (tandem) scanner. ImageJ (v2.0.0) software was used to quantify relative protein levels expressed as mean fluorescence intensity (M.F.I.) or area coverage (μm^2), normalized to the total number of cells/field (the number of DAPI-stained nuclei). M.F.I./cell or area/cell was used to express the protein levels of interest per cell in the absence or presence of HsSNCA PFFs respectively. In the case of primary oligodendroglial cell cultures the area of signal was normalized to the number of TPPP/p25A⁺ cells (marker for mature oligodendrocytes). For the estimation of autophagic flux (integrated intensity and number of puncta/cell) the images were analyzed using Imaris 9.1.2 (Bitplane, South Windsor, CT, USA). We used the surface module of Imaris for image segmentation. All measurements were made from the corresponding surface standard Imaris statistics. Different channels were processed independently.

Molecular modeling

The amino acid sequence of the human TPPP/p25A was obtained from the UniProt database (accession no: O94811). Using the Gapped-BLAST through NCBI the homologous human protein with PDB id 2JRF was identified, which was used as template for the homology modeling. The homology modeling of the TPPP/p25A model was carried out using the RCSB entries 2JRF (human) and 1WLM (mouse) as template structures. The sequence alignment between the raw sequence of the all the above revealed more than 87% sequence identity, which allowed conventional homology modeling techniques to be applied. Electrostatic potential surfaces were calculated on grid points per side (65, 65, 65) and the grid fill by solute parameter was set to 80%. The dielectric constants of the solvent and the solute were set to 80.0 and 2.0, respectively. An ionic exclusion radius of 2.0 Å, a solvent radius of 1.4 Å and a solvent ionic strength of 0.145 M were applied. Amber99 charges and atomic radii were used for this calculation. Energy minimizations were used to remove any residual geometrical strain in each molecular system, using the Charmm27 forcefield as it is implemented into the MOE suite. Molecular systems were then subjected to unrestrained Molecular Dynamics Simulations (MDS) using the MOE suite. MDS took place in a SPC water-solvated, periodic environment. Water molecules were added using the truncated octahedron box extending 7 Å from each atom. Molecular systems were neutralized with counter-ions as required. For the purposes of this study all MDS were performed using the NVT ensemble in a canonical environment, at 300 K and a step size equal to 2 femtoseconds for a total 100 nanoseconds simulation time. An NVT ensemble requires that the Number of atoms, Volume and Temperature remain constant throughout the simulation.

KFERQ-like motif discovery

The developed scoring methodology that was used to identify hidden KFERQ-like motifs in TPPP/p25A is capable of reading the primary amino acid sequence both from the N to C termini and vice versa. The methodology is based on the fusion and analysis of amino acid physicochemical and structural properties that resemble the reference properties of the KFERQ motif. We calculated 435 molecular and physicochemical descriptors (dimensions) for each amino acid in an effort to be able to confidently scan the sequence of TPPP/p25A for KFERQ-like motifs. Only the KKRFK pentapeptide was discovered that meets the criteria for a KFERQ-like motif. The acceptable thresholds for each dimension used were calibrated and trained based on the properties of KFERQ and all known and established KREFQ-like motifs. The KFERQ-like motifs included from literature were: KFERQ, RKVEQ, QEKRV, QDLKF, QRFFE, DRIKQ, IRDLQ, QDIRR, QEFVR, QKIIE, DLLRQ, QKDFR, DFRKQ, KDLLQ [32]. Therefore, the KFERQ likeness of KKRFK was based on an algorithm that has been trained on KFERQ and the 13 established KFERQ-like motifs. Our analysis exploits the dynamic and static information and metrics calculated in an effort to discover patterns and associations among them that will lead to induction of new rules updating the knowledge base and affecting the reasoning process. The combination of the multimodal data collected was used as new knowledge in the form of rules. These rules were exploited to update the reasoning system in order to make adaptive the reasoning processes, i.e., tailored to unique information and characteristics of each motif. Efficient mining of association rules generated a scaled quantification of KFERQ likeness based on 435 physicochemical dimensions. Not only our multimodal methodology on determining KFERQ-likeness, confirms that KKRFK is KFERQ-like based on 435 fused and analyzed physicochemical properties, but we also propose that our approach can be used as a standard of quantification of KFERQ-likeness henceforward.

Isolation of lysosomes

To validate that TPPP/p25A is a CMA substrate, brain lysosomes were isolated from starved Sprague Dawley rats, as previously described [79,80]. Tissue was homogenized in 0.25 M sucrose (Applichem, A4734) and the lysosomes were removed from the light mitochondrial fraction using a Nycodenz (Ncd; Abott, AXS-1002424) density gradient after an ultracentrifugation at 141,000 x g for 1 hr. To examine the integrity of the isolated lysosomes, the enzymatic activity of HEXB/β-hexosaminidase was measured and only when less of 10% of the lysosomes were broken the process could proceed. In our experiments the ability of isolated lysosomes to degrade recombinant TPPP/p25A was explored via Western blot analysis using LAMP1, LAMP2A and CTSD (cathepsin D) as lysosomal markers.

CMA of recombinant TPPP/p25A by isolated lysosomes

Transport of human recombinant TPPP/p25A into isolated rat brain lysosomes was analyzed using an in vitro system as previously described [81]. Briefly, 0.2 μg of recombinant TPPP/p25A were incubated with freshly isolated rat brain lysosomes in MOPS buffer (10 mM 3-[N-morpholino] propanesulfonic acid, [Sigma-Aldrich, M1254] pH 7.3, 0.3 M sucrose [Applichem, A4734]), in the presence of 0.6 μg recombinant HSPA8/HSC70 (Enzo Life Sciences, AD1-SPP-751-D) for 20 min at 37°C. Where indicated, lysosomes were pre-incubated with a cocktail of proteinase inhibitors (Roche, 11,836,153,001) for 10 min at 0°C. A competition assay utilizing 3x (0.6 μg) and 6x (1.2 μg) amount of human recombinant SNCA, a well-established CMA substrate [33] was also performed to further verify the contribution of CMA to TPPP/p25A proteolysis. At the end of the incubation, lysosomes were collected by centrifugation, washed and subjected to SDS-PAGE and immunoblotted for TPPP/p25A, SNCA, LAMP1, LAMP2A, CTSD, HSPA8/HSC70 and ACTB.

Intracellular long-lived protein degradation assay

OLN cells were cultured to 60–70% confluence in 12-well plates in full DMEM and incubated with 1 μg/ml of HsSNCA PFFs for 48 h. Total lysosomal protein degradation was measured by pulse-chase experiments, by labeling PBS- or PFF-treated OLN cells with [³H] leucine (2 μCi/ml) (leucine, L-3,4,5; NEN-Perkin Elmer Life Sciences, NET460001MC) at 37°C for 24 h. The cultures were then extensively washed with medium and returned in complete growth medium containing 2 mM of unlabeled leucine for 6 h. This medium containing mainly short-lived proteins was removed and replaced with fresh medium (DMEM + 0,5% FBS) containing cold leucine (control conditions), medium containing 20 mM NH₄Cl (total lysosomal proteolysis) or medium containing 10 mM 3 MA (macroautophagic degradation). The lysosomal degradation that remains unaffected by the use of 3 MA was attributed to the CMA pathway, after the assumption that microautophagy can be considered minor [38].

Aliquots of the medium were taken at 16 h after labeling and proteins in the medium were precipitated with 20% thrichloroacetic acid for 20 min on ice and centrifuged (10,000 X g, 10 min, 4°C). Radioactivity in the supernatant (representing degraded proteins) and pellet (representing undegraded proteins) was measured in a liquid scintillation counter (Wallac T414, Perkin Elmer). At the last time point, cells were lysed a mild lysis buffer, containing 0.1 N NaOH and 0.1% sodium deoxycholate (Sigma-Aldrich, 30,970). Proteolysis was expressed as the percentage of the initial total acid-precipitable radioactivity (protein) in the cell lysates transformed to acid soluble radioactivity (amino acids and small peptides) in the medium during the incubation [38].

Proteasome activity assay

OLN cells were treated with 1 μg/ml of HsSNCA PFFs for 48 h or Psmb5 siRNA for 72 h and the measurement of the chymotrypsin-like (CT-like) enzymatic activity of the proteasome was accomplished with the use of the fluorescent properties of the synthetic peptide Suc-LLVY-AMC, as previously described [82]. Suc-LLVY-AMC (Calbiochem, 539,142) is a proteasomal substrate that releases a fluorescent molecule (AMC) after cleavage, with emission at 437 nm. The reaction was performed in 5 μg of total protein lysate, with the addition of a resuspension buffer (50 mM Tris-HCl, pH 7.6, 5 mM DTT, 10 mM ATP [Applichem, A1348], 50 mM MgCl₂) and 100 μΜ of the fluorogenic substrate, followed by an incubation at 37° for 10 min. Reaction was terminated by addition of 5% SDS and the released fluorescence was calculated with a PerkinElmer LS-55 luminescence spectrophotometer.

Assessment of cell survival

Viable cells were quantified by counting the number of intact nuclei in a hemacytometer, after lysing the cells in detergent- containing solution [83,84]. This method has been shown to be reproducible and accurate and to correlate well with other methods of assessing cell survival-death [85,86]. In more detail, cells were resuspended in a detergent-containing lysis buffer (0.1x PBS, 0.5% Triton X-100, 2 mM MgCl₂, 0.013 mM ethyl-hexadecyl-dimethyl-ammonium bromide [Sigma-Aldrich, C5337], 0.28% glacial acetic acid, 2.82 mM NaCl) and two independent examiners blinded to the identity of the samples counted the number of intact nuclei with the aid of a hemocytometer. Cell counts were performed in triplicate and the results were presented as means ± SE.

RNA extraction, cDNA synthesis and real-time PCR

Total RNA was extracted from OLN-AS7 and OLN-p25α cells treated with PBS or epoxomicin (15 nM) or NH₄Cl (20 mM) for 24 or 48 h using TRIzol® reagent (Ambion, Thermo Fisher Scientific, 15,596,018). Following digestion with 1 U/μg DNase I (Invitrogen, 18,047–019) 1 μg total RNA was used to synthesize the first strand cDNA according to the Moloney murine leukemia virus reverse transcription reaction system (Promega, M1708), which subsequently would be the template for the RT-PCR reaction. To this end, duplicates of each sample were analyzed using a Light Cycler 96 (Roche Applied Science, Mannheim, Germany) to determine the levels of human SNCA and TPPP/p25A mRNA and rat Gapdh was the reference gene for normalization. Primer sequences utilized in the study are shown in Table 2. Each cDNA sample was diluted 1:20 before use in the amplification assay. The utilized PCR conditions were 1× buffer (-Mg), 1.5 mm MgCl₂, 0.2 mm dNTPs, 0.2 μm primers, template < 500 ng, 2 U Platinum Taq and SYBR Green (Bio-Rad, 172–5120), whereas the PCR cycling conditions were 95°C for 180 s, 95°C for 10 s, 60°C for 15 s, 72°C for 15 s (45 cycles), 95°C for 60 s, 60°C for 60 s, 95°C for 10 s, 37°C for 30 s for human SNCA and 52°C for 120 s, 95°C for 120 s, 95°C for 15 s, 59°C for 40 s (50 cycles), 95°C for 10 s, 55°C for 60 s, 98°C for 1 s, 37°C for 30 s for human TPPP/p25A. No template samples served as negative controls. Data were analyzed automatically with a threshold set in the linear range of amplification. The cycle number at which any particular sample crossed that threshold (Ct) was used to determine fold difference, whereas the geometric mean of the control gene served as a reference for normalization. Fold difference was calculated with the 2− ΔΔCt method.

Statistical analysis

All statistical analysis was performed utilizing GraphPad Prism 5. To be more specific, differences between or within groups were assessed by unpaired t-test, one-way and two-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test and Bonferroni’s correction, respectively. Results are displayed as the mean ± standard error (SE), with a p value of < 0.05 defined as statistically significant. Results are based on the analysis of three-five independent experiments with at least duplicate samples/condition within each experiment.

Acknowledgments

The authors would like to thank the BRFAA biological imaging facility for their valuable contribution to confocal imaging and image analysis, especially Drs Tasos Dellis (Post-Doctoral Fellow) and Stamatis Pagkakis (Staff Research Scientist), the BRFAA Animal Facility, as well as Grigoria Tsaka, MSc, Anastasia Vamvaka-Iakovou, BSc, Evangelos Doukoumopoulos, BSc and Karin Giller for their technical assistance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed here

Additional information

Funding

This work was supported by an Multiple System Atrophy Trust grant (2019/MX60185), and partly by an Multiple System Atrophy Coalition grant and a Bodossaki grant to MX. We acknowledge support of this work by the action “Precision Medicine Hellenic Network in Genetic Neurodegenerative Diseases”, which is implemented under the Project 1. “GR Inherited Network in Cardiology: Network of Sudden Death Prevention in the Young and Precision Medicine in Cardiology – Precision Medicine Hellenic Network in Genetic Neurodegenerative Diseases”, funded by the Programme “Precision Medicine and Climate Change National Research Networks Infrastructures” (2018ΣΕ01300001), of the Hellenic Public Investments Programme of GSRT. This research is co-financed by Greece and the European Union (European Social Fund- ESF) through the Operational Programme «Human Resources Development, Education and Lifelong Learning» in the context of the project “Strengthening Human Resources Research Potential via Doctorate Research” (MIS-5000432), implemented by the State Scholarships Foundation (ΙΚΥ)», awarded to PM. The study was for PHJ supported by Lundbeck Foundation grants R223-2015-4222 & R248-2016-2518 for Danish Research Institute of Translational Neuroscience-DANDRITE, Nordic-EMBL Partnership for Molecular Medicine, Aarhus University, Denmark. Aarhus University. LS has been supported by a GSRT-HFRI grant for Faculty Members & Researchers (Foundation for Research and Technology-Hellas HFRI-FM17-3013);Bodossaki Foundation;Multiple System Atrophy Coalition (US);