Endogenous oligodendroglial alpha-synuclein and TPPP/p25α orchestrate alpha-synuclein pathology in experimental multiple system atrophy models

Cell culture and treatments

The stable cell line OLN-p25α was established essentially as the OLN-AS7 line [51] by transfecting OLN-93 oligodendroglial cells with a pcDNA3.1 zeo(−)human p25α vector. The parental immortalized OLN-93 cell line (control cells) was derived from primary Wistar rat brain glial cultures [57]. All cells were grown in Dulbecco’s modified Eagle’s medium (D6429; Gibco, Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (10,270; Gibco, Invitrogen, Carlsbad, CA, USA), 50 U/mL penicillin, and 50 µg/mL streptomycin. OLN-AS7 and OLN-p25α cells were maintained in 50 µg/mL Zeocin (R25001; Thermo Fisher Scientific, Waltham, MA, USA). The pcDNA3.1 zeo(−) plasmid expressing human p25α was generated as described previously [68]. Primary oligodendroglial progenitor cells derived from mixed glial cultures were prepared from P0 to P3 neonatal knockout (KO)-aSyn, wild-type (WT)-aSyn or PLP-human aSyn (haSyn) transgenic mice, as described previously [32], and cultured in SATO medium [5]. Cultures of rat (embryonic day 18) cortical neurons were prepared as described previously [78] and maintained in Neurobasal medium (Invitrogen, Carlsbad, CA, USA), supplemented with B27 supplement (Invitrogen, Carlsbad, CA, USA), l-glutamine (0.5 mM), and penicillin/streptomycin (1%). A detailed description of the conditions utilized for plating the different cell types, transfections, and treatments are provided in Supplementary Methods (Online Resource 16).

Preparation of haSyn PFFs

HaSyn fibrils were generated as described [6, 22], with several modifications. Shortly, before aggregation, monomeric aSyn in 50 mM HEPES (pH 7.4,) 100 mM NaCl, and 0.02% NaN₃ was centrifuged for 1 h at 84.000×g. The supernatant was ultra-filtrated (0.2 µm) and adjusted to 0.27 mM. Aggregation was performed for 12 days at 37 °C. Fibrils were collected by ultracentrifugation, washed twice with phosphate-buffered saline (PBS), pH 7.3, at room temperature, and quantified by subtracting the amount of monomeric aSyn protein in the supernatant from the total protein used for aggregation. PFFs were resuspended in PBS (pH 7.3) at a final concentration of 4.5 mg/mL and a working stock solution was prepared at 1 mg/mL. The cells were incubated with 0.5–3 μg non-sonicated haSyn PFFs (as described above) or PBS as a control for the indicated time points and were either processed for confocal microscopy or lysed and collected for western blot analysis as described below.

Subcellular fractionation and western immunoblotting

OLN cells incubated with haSyn PFFs for various time points were washed twice in ice-cold PBS and sequentially fractionated using buffers with increasing extraction strength, as described in Supplementary Methods (Online Resource 16). Primary oligodendroglial cells were lysed using RIPA buffer (150 mM NaCl, 25 mM Tris pH 7.6, 0.5% sodium deoxycholate, 0.1% SDS, 1% NP-40), left on ice for ~ 30 min, and centrifuged at 10,000×g for 15 min at 4 °C. Western blot analysis of protein extracts was performed as described previously [70] and the antibodies used are shown in Supplementary Table 2 (Online Resource 11). The intensity of each immunoreactive band was estimated by densitometric quantification using ImageJ software.

Immunocytochemistry and confocal microscopy

OLN cell lines and primary mouse oligodendrocyte cultures cultured on poly-d-lysine-coated glass coverslips were treated with haSyn PFFs at the indicated time points. The cells were fixed with 4% paraformaldehyde for 40 min, blocked in 10% normal goat serum containing 0.4% Triton X-100 for 1 h at room temperature, and incubated with primary antibodies overnight at 4 °C. The primary and fluorescent secondary antibodies used are shown in Supplementary Table 2 (Online Resource 11). Images were obtained using a Leica TCS SP5 confocal microscope combined with a dual (tandem) scanner. All confocal images were obtained under equivalent conditions of laser power, pinhole size, gain, and offset settings between the groups. For the transient transfections, quantifications were performed with the Imaris software suite (v7.7.2, Bitplane AG, Zurich, Switzerland), using a static set of parameters to isolate either GFP⁺ or p25α⁺ cell profiles, and then masking the channel containing the human aSyn (LB509 antibody), oxidized/nitrated aSyn (SYN-303 antibody) or the rodent aSyn (D37A6 antibody) signal using said profiles. The masking of the GFP⁺ or p25α⁺ signal was based on fluorescent intensity. ImageJ (v2.0.0) software was used to quantify relative protein levels expressed as mean fluorescence intensity or % area coverage, normalized to the total number of cells/field (the number of DAPI-stained nuclei).

Transmission electron microscopy (EM)

RNA extraction, cDNA synthesis and real-time PCR

Total RNA was extracted from OLN-93 cells treated with haSyn PFFs for 2, 12, and 48 h (or PBS) using TRIzol^® (Ambion, Thermo Fisher Scientific, Waltham, MA, USA). Following digestion with 1 U/μg DNase I (Promega, Madison, WI, USA), 1 μg total RNA was used for first-strand cDNA synthesis with the Moloney murine leukemia virus reverse transcription system (Promega, Madison, WI, USA) and utilized for real-time PCR, according to the manufacturer’s instructions, as described in Supplementary Methods (Online Resource 16).

Autopsy case material

Human brain tissue was obtained through the University of Florida Neuromedicine Human Brain Tissue Bank according to protocols approved by the Institutional Review Board. Post-mortem pathological diagnoses were made according to standard neuropathological criteria [33, 66]. Cerebellar samples containing GCIs from three MSA cases were studied including the two major pathological subtypes: striatonigral degeneration and olivopontocerebellar atrophy (Online Resource 10). The cingulate cortex and midbrain containing cortical and brainstem Lewy Bodies (LBs), respectively, from three dementia with Lewy bodies (DLB) cases with concurrent Alzheimer’s disease pathology were also utilized.

Immunohistochemistry and semi-quantitative scoring

Immunostaining of the sections was performed using standard methods [11]. The sections were rehydrated and subsequent antigen retrieval was performed in a steam bath for 30 min in a 0.2% Tween-20 solution for all antibodies utilized. Non-specific antibody binding was minimized with a 2% fetal bovine serum/0.1 M Tris blocking solution; primary antibodies were diluted in the blocking solution and applied to tissue sections at 4 °C overnight. Biotinylated secondary antibodies (Vector Laboratories, Burlingame, CA, USA) were diluted in the blocking solution and applied to the sections for 1 h at room temperature. An avidin–biotin complex system (VECTASTAIN ABC Elite Kit; Vector Laboratories, Burlingame, CA, USA) was used to enhance detection of the immunocomplexes, which were visualized using the chromogen 3,3′-diaminobenzidine (DAB Kit; KPL, Gaithersburg, MD, USA). For comparison, all MSA and DLB sections were exposed to DAB for the same amount of time for a given antibody. Tissue sections were counterstained with hematoxylin. All slides were digitally scanned using an Aperio ScanScope CS instrument (40× magnification; Aperio Technologies Inc., Vista, CA, USA), and images of representative areas of pathology were captured using ImageScope software (40× magnification; Aperio Technologies, Inc., Vista, CA, USA). Inclusion pathology was observed and assessed qualitatively and confirmed independently by a second observer.

Animals

Eight-week-old male WT C57BL6/C3H (WT-aSyn), C57BL6/JOlaHsd aSyn null (KO-aSyn) (purchased from Harlan Laboratories), or C57BL6 PLP-haSyn transgenic (kindly provided by Dr. Nadia Stefanova, Innsbruck University) mice were housed (6–8 animals/cage) with free access to food and water under a 12-h light/dark cycle. All experimental procedures were approved by the Ethics Committee for the Use of Laboratory Animals in the Biomedical Research Foundation of Athens.

Surgical procedures

All surgical procedures were performed under general isoflurane (B506; Abbott, Chicago, USA) anesthesia. Eight-week-old male WT or KO aSyn mice were placed in a stereotaxic frame (Kopf Instruments, Tujunga, CA, USA) and the right dorsal striatum was targeted using the following coordinates from the bregma: + 0.5 mm anteroposterior, − 1.6 mm mediolateral, and − 3.4 mm dorsoventral according to the Mouse Brain Atlas [80]. A total of 2 μL haSyn PFF solution (final concentration, 2 μg/μL) or PBS was injected unilaterally at a constant flow rate of 0.1 μL/15 s using a pulled glass capillary (diameter, 60–80 mm) attached to a Hamilton syringe with a 22-gauge needle. Following delivery of haSyn PFFs or PBS, the capillary was held in place for 5 min and removed slowly from the brain.

Immunohistochemistry and confocal microscopy

At 1 month post-injection, the animals were anesthetized with isoflurane followed by intracardial perfusion with ice-cold PBS and then ice-cold 4% paraformaldehyde solution (n = 4–5 animals/condition/genotype). The brains were removed and post-fixed for 24 h in the same preparation of paraformaldehyde and transferred to 15% sucrose for 24 h and finally in 30% sucrose until sectioning. The brains were sectioned through the coronal plane at 30 μm increments, and every section throughout the striatum and ventral midbrain was collected. Immunohistochemical staining was carried out in free-floating sections as described previously [79] using the primary and secondary antibodies described in the Immunocytochemistry and Confocal Microscopy section.

Statistical analysis

The data are shown as the mean ± standard error (SE) of 3 or 5 independent experiments with triplicate samples/condition within each experiment. The mean values ± standard deviation (SD) of each experiment are also provided in Supplementary Table 3 (Online Resource 12). Image acquisition and analysis were performed by two independent examiners blinded to the true identity of the samples. Statistical analysis was carried out with GraphPad Prism 5 using one-way analysis of variance (ANOVA) with Tukey’s post hoc test or two-way ANOVA with Bonferroni’s correction. Differences were considered significant for p < 0.05.

Internalization of haSyn PFFs by oligodendroglial cell lines leads to the formation of Triton-insoluble aSyn species

In order to assess the uptake and seeding capacity of haSyn PFFs in oligodendroglial cells, we initially utilized the immortalized rat OLN-93 cell line, which expresses very low to non-detectable levels of endogenous aSyn and p25α, and the stable cell lines OLN-AS7 and OLN-p25α, which overexpress human WT-aSyn and p25α, respectively (Online Resource 1a). Stable or transient overexpression of haSyn and p25α in OLN cells was not accompanied by changes in cell viability (Online Resource 1b). EM used to monitor fibril morphology and to confirm the presence of fibrils (Online Resource 1c). Treatment of OLN-93 cells with 0.2 and 0.5 μg haSyn PFFs for 48 h, led to the dose-dependent accumulation of SDS- and urea-soluble high molecular weight (HMW) aSyn species (Online Resource 1d). We used 0.5 μg haSyn PFFs in subsequent experiments, and we investigated the solubility and aggregation of internalized PFFs in the OLN cell lines over time (24–96 h). The Triton-soluble fraction of all cell lines contained only monomeric aSyn (detected with antibodies against human and human + rodent aSyn) that showed the highest levels at 24 h post-addition, which was followed by a time-dependent decrease to baseline levels (PBS-treated cells), evident mostly in the control OLN-93 cells (Fig. 1a and Online Resource 2a, b). This could be attributed either to the effective clearance of the formed aSyn species or to the dilution of exogenous aSyn due to cell proliferation. Notably, monomeric aSyn levels detected in the Triton-soluble fraction of OLN-AS7 cells remained high over time, possibly due to the stable overexpression of the human protein in these cells, whereas in the OLN-p25α cells both exogenously added human and total aSyn remained at higher levels than in the OLN-93 cells. HMW-aSyn species were detected mostly in the SDS- and urea- soluble fractions (Fig. 1b, c), thus representing rather insoluble aSyn species. Similar to the Triton-soluble fraction, in the SDS-soluble fraction, the peak levels of monomeric and HMW aSyn were observed at 24 h in all cell lines, followed by a gradual decrease over time (48–96 h); however, they remained at higher levels than in the PBS-treated cells (Fig. 1b and Online Resource 2c–f). Interestingly, in the urea-soluble fraction of OLN-p25α cells, aSyn monomeric and oligomeric forms (both human and total) were relatively resistant to degradation, since their levels remained high compared to control OLN-93 cells, in which their levels decreased over time (Fig. 1c–e). This could be indicative of a role for p25α in promoting aSyn aggregate formation within oligodendrocytes or in inhibiting the clearance of aSyn aggregates.

Open image in new window

Addition of haSyn PFFs to OLN cell lines recruits endogenous rat oligodendroglial aSyn into the formation of highly insoluble aSyn conformations

In order to investigate the ability of haSyn PFFs to act as a template for endogenous oligodendroglial aSyn to form insoluble protein species, we utilized the rodent-specific D37A6 antibody in western blot and confocal microscopy analyses. This antibody does not give any immunofluorescence signal in PBS-treated OLN cells (Fig. 2a and Online Resource 3); the addition of 0.5 μg haSyn PFFs was accompanied by the detection of an elevated endogenous aSyn signal, which co-localized with the exogenously added human aSyn that persisted over time (48 h–10 days) and depended, in terms of its intensity, on the exogenously added human material (Online Resource 4). The uptake of haSyn PFFs was further verified by staining the lectins of the plasma membrane using wheat germ agglutinin (Online Resource 1e and 13) and ultrastructural EM analysis, which revealed the presence of filaments in the cell soma of OLN-93 cells treated with haSyn fibrils, which were absent in PBS-treated cells (Fig. 2b, upper panels). Double immuno-EM analysis with nanogold particles of different diameters, showed the presence of human and rat aSyn in the formed intracellular filamentous inclusions (Fig. 2b, bottom panels). Measurement of human or rat endogenous aSyn levels, expressed as % of surface area normalized to the total number of cells/field, revealed no significant differences among the different OLN cells, at 48 h and 10 days post-addition of 0.5 μg haSyn PFFs (Fig. 2c). However, and in agreement with the immunoblot data (Fig. 1c–e), OLN-p25α cells retained significantly more proteinase K-resistant human and rodent aSyn species compared to OLN-93 and OLN-AS7 cells (Online Resource 5).

Open image in new window

Similarly, in the urea-soluble fraction of OLN-p25α cells from 24 h to 10 days post-addition of haSyn PFFs, higher levels of monomeric and HMW forms of endogenous rodent aSyn were found compared to OLN-93 and OLN-AS7 cells (Fig. 2d–f). No rat-specific signal was detected in the Triton- or the SDS-soluble fractions at 24 h–10 days post-addition of haSyn PFFs (data not shown). The elevated levels of endogenous aSyn detected following treatment of OLN cells with haSyn PFFs were not attributed to increased transcription of the rat Snca gene, encoding endogenous oligodendroglial aSyn, as revealed by quantitative real-time-PCR analysis (Fig. 2g).

The specificity of the D37A6 aSyn antibody to recognize only the rodent protein was validated further by confocal microscopy and western immunoblotting, utilizing human SH-SY5Y dopaminergic cells inducibly overexpressing the WT-aSyn upon the removal of doxycycline (dox) [69] and extracts from WT-, KO-, and hA53T-aSyn transgenic mice (Online Resource 6 a, b). Importantly, treatment of OLN-93 cells with the proteasomal inhibitor epoxomicin (epox) or the general lysosomal inhibitor NH₄Cl alone or in combination (epox + NH₄Cl) for 48 h revealed an increased aSyn signal upon lysosomal inhibition, as detected with the D37A6 antibody in the absence of haSyn PFFs, further supporting the presence of miniscule amounts of endogenous aSyn in oligodendrocytes that are increased upon inhibition of protein degradation (Online Resource 6c, d).

Overexpression of human aSyn or p25α accelerates the recruitment of the endogenous rat oligodendroglial aSyn and the formation of intracellular oxidized/nitrated aSyn species following the addition of haSyn PFFs

To decipher the impact of human aSyn or p25α overexpression on the seeding of aSyn pathology in OLN cells, we performed time-course analysis from 30 min until 24 h post-treatment. Confocal microscopy analysis with antibodies against human and rodent aSyn (LB509 and D37A6 antibodies, respectively) revealed that haSyn PFFs were taken up effectively by all OLN cell lines within 1 h, whereas the recruitment of the endogenous rat aSyn represented a later event (Fig. 3). Interestingly, in control OLN-93 cells, the seeding process was initiated at 12 h post-treatment with the haSyn PFFs (Fig. 3a), whereas in OLN-AS7 and OLN-p25α cells, the endogenous rat aSyn signal was detected earlier, at 2 h post-addition (Fig. 3b, c). Importantly, the recruitment of endogenous rat aSyn evokes a dramatic change in the pattern of haSyn immunostaining and an increase in haSyn immunoreactivity, measured as % area surface/cell, at the onset of seeding (Fig. 3). This observation suggests a possible role for the seeded rat aSyn in the rapid formation of aSyn aggregates positive for both human + rodent aSyn or in the stabilization of haSyn aggregates, rendering them resistant to degradation.

Open image in new window

We next investigated whether excess intracellular aSyn protein load (OLN-AS7 cells) or the presence of p25α (OLN-p25α cells) affects the onset of the appearance of aSyn pathology in oligodendrocytes by assessing the formation of oxidized/nitrated aSyn species over time (30 min–10 days) with confocal microscopy. Indeed, in OLN-AS7 and OLN-p25α cells, the appearance of oxidized/nitrated aSyn (detected with the Syn303 aSyn antibody) was observed within 8 h post-addition of 0.5 μg haSyn PFFs, whereas in the control OLN-93 cells, these pathological species were observed after 20 h (Fig. 4). Notably, co-staining with the rodent-specific aSyn antibody revealed that endogenous aSyn was a major component of the seeded time-resistant pathological aSyn assemblies. Importantly, the Syn303 aSyn antibody did not recognize exogenously added haSyn PFFs per se, since no signal was observed at 1 h post-addition, when haSyn PFFs were detected inside the cells (data not shown), further supporting a major role for oligodendroglial aSyn in the seeding of aSyn pathology in OLN cells.

Open image in new window

Transient overexpression of human p25α augments the formation of insoluble aberrant aSyn species, consisting of human and endogenous rat protein

In order to validate further the role of p25α in the seeding of aSyn pathology, we used another experimental approach. Specifically, OLN-93 and OLN-AS7 cells were transiently transfected (or co-transfected where indicated) with p25α or/and GFP (as a control) plasmids 6 h prior to the addition of 0.5 μg of haSyn PFFs. At 2 days post-transfection, the cells were fixed and processed for confocal microscopy using antibodies against human aSyn (LB509 antibody, Fig. 5a, left panel), oxidized/nitrated aSyn (Syn303 antibody, Fig. 5b, left panel), or rodent aSyn (D37A6 antibody, Fig. 5c, left panel). The transient overexpression of human p25α increased the accumulation of human (Fig. 5a) and endogenous rat (Fig. 5c) aSyn-positive signals and the generation of oxidized/nitrated aSyn species (Fig. 5b) in OLN-93 and OLN-AS7 cells. Additionally, p25α overexpression in OLN-93 cells augmented the formation of monomeric and HMW aSyn species extractable in urea buffer (Fig. 5d). Such data further support a critical role for p25α in the aggregation of pathological aSyn conformations within oligodendrocytes.

Open image in new window

Phosphorylation of aSyn at Ser129 depends on aSyn and p25α protein load and may involve different aSyn aggregate “strains” present in oligodendroglial and neuronal synucleinopathies

Given that the presence of aSyn phosphorylated at Ser129 (pSer129 aSyn) is an indicator of pathology in synucleinopathies, we assessed pSer129 aSyn levels in all OLN cell lines upon treatment with 0.5 or 3 μg haSyn PFFs for 48 h and 10 days using a panel of specific pSer129 aSyn antibodies. Surprisingly, we detected positive pSer129 aSyn signals only in PFFs-treated OLN cells overexpressing human aSyn or p25α (in the presence of 3 μg haSyn PFFs for 10 days) and not in the control OLN-93 cells, using the pSer129 aSyn-specific antibodies 11A5 (Fig. 6a, middle row) and EP1536Y (Fig. 6a, bottom row). Conversely, no positive signal was detected in any PFF-OLN-treated cell line using the pSer129 aSyn antibody LS4-2G12 (Fig. 6a, upper row), which could be attributed to the low affinity of this antibody. Interestingly, no positive pSer129 aSyn signal was detected in the Triton- or SDS-soluble fraction of all PFF-treated OLN cells (Fig. 6b), whereas pSer129 immunoreactivity was observed only in the urea-soluble fraction of OLN-AS7 cells and to a greater extent in OLN-p25α cells (Fig. 6c). Most importantly, endogenous rat aSyn co-localized with the 11A5 pSer129 aSyn antibody, suggesting that the seeded oligodendroglial aSyn is phosphorylated, upon the addition of haSyn PFFs (Fig. 6d).

Open image in new window

Strikingly, in PFF-treated primary rat cortical cultures we detected a pSer129 aSyn signal with all antibodies utilized, including the LS4-2G12 antibody that did not give a signal in the OLN seeding model (Fig. 6e), suggesting that this antibody may recognize distinct pSer129 aSyn assemblies generated specifically in neurons and not in oligodendrocytes. Importantly, this hypothesis was further buttressed by data obtained from human post-mortem brain material derived from patients with neuronal synucleinopathy (DLB) and oligodendroglial synucleinopathy (MSA). As shown in Fig. 6f, and in agreement with our results from PFF-treated primary cortical cultures, all pSer129 antibodies recognized pSer129 aSyn in LBs in the substantia nigra pars compacta (left columns) and cortex (middle columns) of DLB brains to a similar extent. In contrast, and congruent with the data obtained from PFF-treated oligodendroglial cells, LS4-2G12 reactivity for MSA cerebellar GCIs was much weaker than the 11A5 and EP1536Y signals. Finally, the conformation-specific aSyn antibody MJFR-14 recognized better the aberrant aSyn species present in GCIs, compared to LBs (Fig. 6f, bottom row), supporting the notion that aSyn generated within oligodendrocytes (MSA) is incorporated into higher order pathological conformations, compared to those generated in neurons (DLB) [8]. In our cellular systems, the MJFR-14 antibody recognized aggregated aSyn in oligodendrocytes (Online Resource 7) and neurons (Fig. 6e, bottom row) to a similar extent, a finding that could be explained by the fact that this antibody also recognizes exogenously added haSyn PFFs.

Endogenous aSyn is a major component of the seeded pathological aSyn aggregates following the addition of haSyn PFFs

To delineate further the contribution of endogenous oligodendroglial aSyn and p25α in the spread of aSyn pathology, we treated OLN cells with 3 μg PFFs and analyzed aSyn levels at 48 h or 10 days (Fig. 7). With this higher amount of PFFs, over time, the equilibrium of intracellular aSyn protein load in OLN cells shifted toward the seeded endogenous rat protein and notably, this phenomenon was more robust in the presence of human aSyn or p25α (Fig. 7a, b). Similarly, oxidized/nitrated aSyn conformations were elevated in OLN-AS7 and OLN-p25α cells over time (Fig. 7c, d). Finally, using the conformation-specific aSyn antibody MJFR-14, which recognizes aggregated conformations exclusively [27], we detected the induction of MJFR-14-positive structures that colocalized from the earliest time point of 1 h with structures recognized with a pan-aSyn (human + rodent) antibody (D10); the MJFR-14 antibody showed no reactivity against overexpressed aSyn in the OLN-AS7 cell line (Online Resource 7). In order to quantitatively estimate the contribution of aSyn and p25α load to the formation and stability of inclusions, we utilized an indirect immunostaining strategy with antibodies against either total aSyn (D10 antibody) or human aSyn (211 antibody) and aggregated aSyn (MJFR-14 antibody), since both the rodent-specific (D37A6) and MJFR-14 antibodies are generated in rabbits (Fig. 7e). Total aSyn (human and endogenous rat) and aggregate-aSyn fluorescence signals exhibited a nearly complete co-localization (Fig. 7e and Online Resource 7b), whereas staining with the 211 human-specific and MJFR-14 aggregate-specific aSyn antibodies (Fig. 7e and Online Resource 7c) revealed partial co-localization, suggesting indirectly that endogenous rat aSyn is a major component of the formed aggregates. Consistent with this idea, MJFR-14-positive immunostaining increased following the addition of PFFs over a period of 10 days, matching the pattern of endogenous rat aSyn, whereas human aSyn-positive structures remained stable or decreased over time, with OLN-p25α cells again exhibiting higher levels of total and aggregated aSyn (Fig. 7e). Interestingly, the percentage of cells positive for the various aSyn conformations was similar among all OLN cell lines, with the exception of pSer129 aSyn (EP1536Y antibody) (Fig. 7f), suggesting that the observed differences shown in Fig. 7e are possibly attributed to the formation of larger aggregates in aSyn- and p25α-expressing cells, compared to OLN-93 cells. Finally, proof-of-concept experiments utilizing sonicated fibrils yielded similar results regarding the seeding of endogenous aSyn and the formation of pathological assemblies (Online Resource 8).

Open image in new window

Endogenous mouse oligodendroglial aSyn is incorporated into pathological aSyn assemblies in primary oligodendrocytes following the addition of haSyn PFFs

To investigate the intracellular processes governing the templating of aSyn pathology in a more physiological cellular setting with properties resembling more closely those of the oligodendrocytes in the central nervous system, we established a protocol to isolate and cultivate mature oligodendrocytes derived from P0 to P3 mouse brains. In order to decipher whether the presence of aSyn, which is small in amount, is a prerequisite for the seeding of exogenous added haSyn, we used primary oligodendrocytes prepared from WT- or KO-aSyn mice. In parallel experiments, we utilized primary cultures from PLP-haSyn mice, a well-established MSA mouse model in which haSyn is overexpressed only in oligodendrocytes, under the PLP promoter [20]. To differentiate the oligodendroglial progenitor cells into mature, myelin-producing (MBP⁺) oligodendrocytes, the cells were grown in SATO medium. The enrichment of the cultures with oligodendrocytes was verified utilizing markers for astrocytes (GFAP) and microglia (AIF1/IbaI), as well as markers for the oligodendroglial lineage such as Olig2, O4, and MBP (Fig. 8a, b). As observed in the OLN cell lines, the addition of haSyn PFFs (0.5 μg, 48 h) evoked a dramatic increase in the levels of endogenous mouse oligodendroglial aSyn only in the WT-aSyn and PLP-haSyn primary cultures (Fig. 8c). No fluorescence signal for the rodent-specific D37A6 aSyn antibody was detected in the KO-aSyn cultures or at baseline in the WT-aSyn and PLP-haSyn cultures in the absence of haSyn PFF treatment (Fig. 8c, f). The magnitude of the increase was dependent on the endogenous aSyn load, since the rodent-specific aSyn signal was more intense and more compact in the PLP-haSyn oligodendrocytes than in WT-aSyn oligodendrocytes (Fig. 8c). The formed aSyn species were aggregated, as verified by the strong immunoreactivity of the conformation-specific MJFR-14 antibody, which co-localized completely with the D10 antibody, recognizing the human + rodent protein (Fig. 8d). Interestingly, the expression pattern of aSyn in the haSyn-PLP cultures changed dramatically from diffuse (in PBS-treated cultures) to strongly punctated (in PFF-treated cultures), as detected by the LB509 and D10 antibodies (Fig. 8c, d). Finally, the rodent-specific positive puncta colocalized to a great extent with the oxidized/nitrated aSyn antibody in the WT-aSyn and haSyn-PLP cultures, but not in the KO-aSyn cultures (Fig. 8e, f), further indicating that the endogenous protein contributes to the formation of such pathological conformations.

Open image in new window

Endogenous oligodendroglial aSyn protein load is closely linked to the re-distribution of microtubule-associated proteins upon the addition of haSyn PFFs

Given the close link between aSyn levels and the redistribution p25α in disease pathogenesis, we assessed the expression pattern of p25α in PFF-treated primary mouse oligodendrocytes. As shown in Fig. 9a, d, there was no apparent change in the pattern of p25α expression in PFF-treated KO-aSyn cultures (upper rows), in contrast to WT-aSyn cultures, in which p25α was redistributed from the oligodendroglial processes to the cell body, although no obvious co-localization with haSyn-positive aggregates was observed (middle rows). Remarkably, p25α in the PFF-treated PLP-haSyn oligodendrocytes, beyond its cell body redistribution, accumulated clearly in haSyn-positive cytoplasmic aggregates (bottom rows). This change was related to haSyn PFFs, given that p25α protein levels were similar between the untreated oligodendroglial cultures (Fig. 9a, d, e).

Open image in new window

Since p25α is a tubulin polymerization-promoting protein that is involved in the organization of the microtubule network [65], its re-localization in the presence of haSyn PFFs might also be linked to the altered localization of other microtubule-associated proteins, such as tau, which has also been found to be present in GCIs [64]. To investigate whether the distribution pattern of tau was affected upon the addition of haSyn PFFs, we studied tau phosphorylated at residue Ser404 (pSer404-tau), which is one of the first sites to be hyperphosphorylated leading to the detachment of tau from microtubules, a phenomenon that is closely related to disease pathogenesis in tauopathies [1, 16, 36, 63]. Our data showed that pSer404-tau was redistributed from the processes to the cell soma upon the addition of haSyn PFFs in a manner dependent on the endogenous aSyn load, since this effect was more evident and robust in PLP-haSyn cultures, whereas the pSer404-tau immunostaining pattern remained unchanged only in the case of KO-aSyn primary oligodendrocytes (Fig. 9b, upper panel and d). However, pSer404-tau largely failed to be incorporated into haSyn aggregates, even in PLP-haSyn cultures, revealing a different behavior compared to p25α; this may have occured because this specific phosphorylated form, in contrast to p25α, is unable to bind to microtubules.

Finally, we assessed the impact of exogenously added haSyn PFFs on myelination, which is impaired in MSA brains [10, 12, 61]. The addition of haSyn PFFs to primary oligodendrocyte precursor cells reportedly leads to the disruption of MBP expression [21]. However, the role of endogenous aSyn in the levels and distribution of MBP remains unknown. In order to investigate whether alterations in the levels of endogenous oligodendroglial aSyn may be linked to alterations in myelin integrity, we assessed haSyn and MBP by confocal microscopy. Our data showed that the presence of endogenous mouse aSyn (in WT-aSyn cultures) or overexpressed human aSyn (in PLP-haSyn cultures) led to a dramatic alteration of MBP immunoreactivity following the addition of haSyn PFFs; in particular, MBP staining was drastically decreased and redistributed from the processes to the cell soma, largely colocalizing with aggregated haSyn (Fig. 9c, d). On the contrary, KO-aSyn mouse oligodendrocytes did not display any significant changes in MBP levels or distribution (Fig. 9c, upper panel), thus indicating a crucial role for seeded oligodendroglial aSyn aggregation in the alterations of MBP.

Delivery of haSyn PFFs into the mouse striatum is accompanied by the recruitment of endogenous oligodendroglial protein in MBP⁺ pathological aggregates and the impairment of myelin integrity

In order to evaluate the seeding capacity of haSyn within oligodendrocytes in vivo, we performed unilateral stereotactic injections of haSyn PFFs (4 μg/brain) (or PBS as a control) into the mouse striatum, to imitate the situation occurring in MSA brains. As shown in Fig. 10a (bottom panels), the delivery of haSyn PFFs into the mouse striatum seeded the recruitment of endogenous mouse aSyn to form MBP-positive inclusions, which were positive for human and rodent aSyn, detected at 1 month post-injection (Fig. 10a, IPSI, indicated by the asterisks). The uptake of haSyn PFFs by oligodendrocytes was verified further by co-staining with the p25α antibody (Fig. 10b and Online Resource 15), whose specificity for oligodendrocytes in the mouse striatum was confirmed by its complete co-localization with O4 and Olig2 antibodies (Fig. 10c).

Open image in new window

Consistent with the results in the primary oligodendroglial cultures, the MBP-positive signal changed dramatically upon the addition of haSyn PFFs, and MBP⁺ puncta colocalized with the human (LB-509), rodent (D37A6), total (SYN1) and aggregated aSyn antibodies (Fig. 10a, d and Online Resource 14). Importantly, such alterations were not observed when a similar amount of haSyn PFFs was injected into the striatum of KO-aSyn mice (Fig. 10a, upper panels). Staining with the Syn303 aSyn antibody, which recognizes oxidized/nitrated species, revealed the formation of pathological aggregated aSyn assemblies in MBP⁺ oligodendrocytes of the ipsilateral striatum within 1 month of injection (Fig. 10e). No Syn303 aSyn-positive signal was detected in haSyn PFF-injected KO-aSyn striatum (data not shown), and none of the aforementioned pathological alterations were observed in the ipsilateral striatum of WT-aSyn mice injected with PBS (Online Resource 9). Importantly, ultrastructural EM analysis revealed that haSyn PFFs evoked a clear rearrangement of the myelin sheath in the striatum of WT-aSyn mice, compared to PFF-treated KO-aSyn mice, as depicted by the increased number of axons with a decompacted myelin sheath and axons with myelin vesiculation (Fig. 10f, g). Such data support further the aforementioned results in PFF-treated primary oligodendrocytes (Fig. 9c, d) and in the PFF-treated ipsilateral WT-aSyn striatum (Fig. 10a, d).