Address correspondence and reprint requests to Karin Danzer, Department of Neurology, MassGeneral Institute for Neurodegenerative Disease, Massachusetts General Hospital and Harvard Medical School, 114 16th Street, Charlestown, MA 02129, USA. E-mail: firstname.lastname@example.org
Lewy bodies, α-synuclein (α-syn) immunopositive intracellular deposits, are the pathological hallmark of Parkinson’s disease (PD). Interestingly, Lewybody-like structures have been identified in fetal tissue grafts about one decade after transplantation into the striatum of PD patients. One possible explanation for the accelerated deposition of α-syn in the graft is that the aggregation of α-syn from the host tissue to the graft is spread by a prion disease-like mechanism. We discuss here an in vitro model which might recapitulate some aspects of disease propagation in PD. We found here that in vitro-generated α-syn oligomers induce transmembrane seeding of α-syn aggregation in a dose- and time-dependent manner. This effect was observed in primary neuronal cultures as well as in neuronal cell lines. The seeding oligomers were characterized by a distinctive lithium dodecyl sulfate-stable oligomer pattern and could be generated in a dynamic process out of pore-forming oligomers. We propose that α-syn oligomers form as a dynamic mixture of oligomer types with different properties and that α-syn oligomers can be converted into different types depending on the brain milieu conditions. Our data indicate that extracellular α-syn oligomers can induce intracellular α-syn aggregation, therefore we hypothesize that a similar mechanism might lead to α-syn pathology propagation.
The neuropathology of Parkinson’s disease (PD) is characterized by the presence of α-synuclein (α-syn) inclusions. In affected neurons, the inclusions appear as thread-like Lewy neurites or globular Lewy bodies. Vulnerable neurons with Lewy pathology all belong to the class of projection neurons with long unmyelinated or poorly myelinated neurites (Braak et al. 2006).
The degeneration of neuromelanin-containing dopaminergic neurons in the substantia nigra is regarded as the most important hallmark of PD and to be the primary cause of the motor deficits in PD (Dauer and Przedborski 2003). However, the neuropathology is not restricted to the substantia nigra and includes various specific, extra-nigral brain regions such as the dorsal IX/X motor nucleus, the reticulate zone, subnuclei of the Raphe system, the thalamus or amygdala and the cortex (Braak et al. 1998, 2003a,b; Braak and Braak 2000). The temporal and topographical order of the histopathological lesions in sporadic PD has been described in detail (Braak et al. 2003a,b). Braak et al. suggested a gradually progressing, ascending course with little interindividual deviation in the development of the PD neuropathology. According to this hypothesis it is conceivable that after an initial event, the disease pathology progresses without remission until reaching a terminal phase (Braak et al. 2003a, 2006). Recent studies have highlighted the possibility that a seeding-nucleation mechanism may exist by studying the fate of neurons grafted into the brains of patients with PD. In these studies, grafted healthy neurons gradually developed the same pathology as the host neurons in the diseased brains (Kordower et al. 2008; Li et al. 2008; Mendez et al. 2008). The presence of Lewy bodies in previously transplanted neurons, versus the lack of Lewy bodies in recently transplanted neurons suggests that aging of transplanted cells promotes the propagation of α-syn aggregation from host to graft. A prion-like mechanism might explain how PD pathology can transfer from the host to the graft (Brundin et al. 2008). A specific conformation of α-syn in host cells might promote misfolding and aggregation of α-syn in adjacent grafted neurons. The family of synucleins came into spotlight when genetic point mutations in the α-syn gene were found to be linked to the development of PD (Polymeropoulos et al. 1997; Krueger et al. 1998) and fibrillar aggregates of α-syn were found to be the main component of Lewy bodies and Lewy neurites (Goedert 2001). Since then a number of α-syn studies were conducted to understand the pathogenesis of PD and other synucleinopathies (reviewed in Surguchov 2008). A growing body of evidence suggests that pre-fibrillar oligomers are the key contributors to the development of PD (Auluck et al. 2002; Kayed et al. 2003; Park and Lansbury 2003; Bucciantini et al. 2004; Bodner et al. 2006). In a previous study we have shown that depending on in vitro aggregation conditions α-syn oligomers form distinct populations that differ in their biophysical properties and cellular effects (Danzer et al. 2007). To further understand the dynamic properties of α-syn oligomers, we have examined whether the different oligomer species can be converted into each other. Moreover, we investigated here the transmembrane seeding effect of distinct α-syn oligomers in various cellular systems to advance the understanding of the progression of α-syn pathology. These results provide the first preliminary evidence that transmembrane seeding might be a key mechanism in the ascending course of α-syn pathology.
Materials and methods
All chemicals used were purchased from Sigma Aldrich, Inc. (Munich, Germany) unless stated otherwise.
Expression and purification of recombinant wild-type α-syn
Expression and purification was performed as described (Nuscher et al. 2004). Briefly, pET-5a/α-syn wild-type plasmid was used to transform Escherichia coli BL21(DE3) pLysS (Novagen, Madison, WI, USA). Expression was induced with isopropyl-ß-d-thiogalactopyranose (Promega, Mannheim, Germany) for 4 h. Cells were harvested, resuspended in 20 mM Tris and 25 mM NaCl, pH 8.0 and lysed by freezing in liquid nitrogen followed by thawing. After boiling for 30 min, the lysate was centrifuged at 17 600 g for 15 min at 4°C. Supernatant was filtered through 0.22 μm filter (Millex-GV, Millipore Corp., Bedford, MA, USA) before loading onto a HiTrap Q HP column (5 mL, Amersham Biosciences, Munich, Germany) and eluting with a 25–500 mM NaCl salt gradient. The pooled α-syn peak was passed over a Superdex 200 HR10/30 size exclusion column (Amersham Biosciences) using 20 mM Tris, 25 mM NaCl, pH 8.0 as running buffer. The pooled α-syn peak was concentrated using Vivaspin columns MWCO 5 kD (Vivascience, Stonehouse, UK) and equilibrated with water. The protein concentration was determined using a BCA protein quantification kit (Pierce, Rockford, IL, USA). Aliquots were lyophilized and stored at −80°C.
Fluorescent labeling of α-syn
Protein labeling with the amino-reactive fluorescent dye Alexa Fluor-488-O-succinimidylester (Alexa488; Invitrogen, Eugene, OR, USA) was performed according to the manufacturer’s manual. Unbound fluorophores were separated by filtration in PD10 columns (Sephadex G25, Amersham Biosciences) equilibrated with 50 mM sodium phosphate pH 7.0. Quality control of labeled α-syn was performed by mass spectrometry and by Fluorescence Correlation Spectroscopy measurements on an Insight Reader (Evotec-Technologies, Hamburg, Germany). The typical labeling ratio was approximately two dye molecules per α-syn molecule. In order to remove preformed aggregates, the stock solution of labeled α-syn was subjected to size exclusion chromatography (Sephadex 200, Amersham Biosciences).
Preparation of α-syn oligomers
Pore-forming oligomer type A
Oligomer type A were prepared as described previously (Danzer et al. 2007). In short, lyophilized recombinant protein was dissolved to a final concentration of 70 μM in cell culture grade water (Sigma Aldrich, Inc.) and sonicated for 30 s. 7 μM α-syn in 50 mM Na2HPO4 buffer (pH 7.0) containing 20% ethanol was prepared by pipetting the reagents with the following order: water, Na2HPO4 buffer (pH 7.0, 250 mM), 100% Ethanol and α-syn (70 μM). After pipetting into an Eppendorf tube, all reagents were mixed by thoroughly vortexing. In this study we focused on oligomers prepared in the presence of 10 mM FeCl3 (J.T. Baker, Griesheim, Germany). After shaking for 4 h with a horizontal shaker (GFL GmbH, Burgwedel, Germany), oligomers were re-lyophilized and resuspended in 50 mM Na2HPO4 buffer (pH 7.0) containing 10% ethanol. This was followed by shaking for 24 h at 20°C in Eppendorf cups on an orbital shaker (setting 5, Eppendorf Thermomixer 5436, Wesseling-Berzdorf, Germany) at 21°C with open lids to evaporate residual ethanol. After 6 days incubation at 20°C without shaking and with closed lids, α-syn oligomers were used for experiments. Alexa Fluor-488-O-succinimidylester labeled oligomers were prepared in the same manner as non-labeled oligomers by using Alexa-488 conjugated monomers.
Seeding oligomer type C
Oligomer type C was as described previously (Danzer et al. 2007). Briefly, lyophilized protein was dissolved in 50 mM sodium phosphate buffer (pH 7.0) containing 20% ethanol to a final concentration of 7 μM in the same order as described for oligomer type A. All oligomer preparations were performed in the presence of 10 μM FeCl3. After overnight incubation at 21°C under slowly but continuously shaking (horizontal shaker, GFL GmbH), oligomers were concentrated 1 : 14 using ultracentrifugation (VivaSpin 500 columns, MWCO 30 kDa, Vivascience, Hannover, Germany). Alexa-488 labeled oligomers were prepared in the same manner as non-labeled oligomers by using Alexa-488 conjugated monomers.
The quality of each oligomer batch has been confirmed by their characteristic band pattern in a lithium dodecyl sulfate (LDS) gel, as irregularly, no soluble oligomers formed. These batches were discarded.
SH-SY5Y human dopaminergic neuroblastoma cells were maintained at 37°C in 5% CO2 in high glucose Dulbecco’s modified Eagle’s medium (PAA Laboratories GmbH, Pasching, Austria) supplemented with 15% fetal bovine serum (Invitrogen GmbH, Karlsruhe, Germany) and 4 mM glutamine (Invitrogen GmbH).
To generate stable cell lines, SH-SY5Y cells were transfected using Metafectene (Cambio Ltd, Cambridge, UK) with pcDNA 3.1neo encoding α-syn[A53T] (plasmid was a kind gift of C. Haas, LMU Munich, Germany). As mock vector controls, SH-SY5Y cells were transfected with plasmid pUHD15.1 encoding the transactivator protein for the tetracycline inducible expression system (Clontech, Saint-Germain-en-Laye, France). Mock transfected cells served as cells with endogenous level of α-syn.
Transfected cells were selected with 1000 μg/mL G418 (PAA Laboratories GmbH) for 2–3 weeks until colonies emerged. Stable transfectants established from these colonies were tested for their α-syn expression levels using immunofluorescence and western blot analyses.
Cortical cell culture
Neuron-enriched cortical cells were prepared from embryonic mouse brains (E14). Cortices were dissected from embryonic brain and the meninges were removed. The cells were dissociated by trypsinization and tituration. The dissociated cells were resuspended in serum free B27/neurobasal medium (Invitrogen) and plated at a density of 1.25 × 105 cells/cm2 on dishes pre-coated with poly-d-lysine/laminin. Cells were maintained at 37°C in 10% CO2 in a humidified incubator. Medium was changed every third day.
Measurement of intracellular Ca2+
Fluorescence imaging with fluorescent imaging plate reader
For intracellular Ca2+ measurements using fluorescent imaging plate reader (FLIPR), SH-SY5Y cells were seeded into Collagen I coated 384-well black clear bottom microtiter plates (BD Biosciences, Heidelberg, Germany) at a density of 6000 cells/well and cultured overnight. Cells were washed with Ringer buffer (130 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 2 mM KH2PO4, 5 mM glucose, 20 mM HEPES) and loaded for 60 min with 2 μM cell-permeable Fluo-4 AM (Teflabs, Austin, TX, USA) in Ringer buffer containing 0.1% (w/v) pluronic acid F127 (Invitrogen). After removal of the fluorophore loading solution, cells were washed with Ringer buffer, incubated at 21°C for 30 min and washed again. The cell plates were then loaded into a fluorescence imaging plate reader (FLIPR TETRA TM, Molecular Devices, Wokingham, UK) together with a separate 384-well plate containing oligomers and controls. Treatment of cells was carried out with 7 μM oligomer type A (referring to moles of monomer starting concentration). The FLIPR was programmed to transfer the oligomer preparations and solvents simultaneously to all 384 wells 10 s after starting the recording of fluorescence (expressed as relative fluorescence units). Fluorescence was excited at 488 nm and emission was measured at 510–560 nm. Duration of recording was 10 min. Data are displayed as negative control corrected values, meaning signal response to oligomers minus the corresponding solvent controls.
Cell treatment and immunofluorescence staining
To investigate the seeding properties of different oligomer preparations, primary cortical neurons were seeded at a density of 1.25 × 105 cells/cm2 on 384-well black clear bottom microtiter plates (BD Biosciences) pre-coated with poly-d-lysine/laminin and cultured as described above. After 7 days, cortical neurons were used for the seeding assays. SH-SY5Y over-expressing mutant α-syn[A53T] were seeded at a density of 5000 cells/well using Collagen I coated 384-well black clear bottom microtiter plates (BD Biosciences) and cultured overnight.
On the day of the experiment, SH-SY5Y or cortical neurons were treated with 10 μM (referring to moles of monomer) of different oligomers unless otherwise stated and the same volume of the corresponding solvent controls. After 2 h treatment, oligomers were diluted 1 : 2 in culture medium for subsequent overnight treatment.
After overnight incubation, SH-SY5Y cells or cortical neurons were washed three times with phosphate-buffered saline (PBS) following a 30 min incubation in a fixation solution containing 2% formaldehyde and 1 μM Hoechst 33342TM (Invitrogen) in PBS. After washing, the cells were permeabilized and unspecific binding sites were blocked using 0.05% Saponin and 1% bovine serum albumin in PBS followed by another washing step. The primary rabbit antibody against α-syn (ASY-1, kind gift of Poul Henning Jensen, University of Aarhus, Denmark; described in Jensen et al. 2000) was added for 1 h at 21°C, followed by another washing step and incubation with the secondary antibody (anti-rabbit antibody labeled with Alexa-Fluor 647; Invitrogen) for 1 h at 21°C.
Western blotting of SH-SY5Y cell extracts
SH-SY5Y over-expressing mutant α-syn[A53T] were seeded at 70% confluence using Collagen I coated cell culture dish 60 × 15 mm (BD Biosciences Heidelberg, Germany) and cultured overnight. 10 μM (referring to moles of monomer) of oligomer type C or the same volume of the corresponding solvent controls were added to cell culture media and incubated at 37°C for the indicated times. After the indicated incubation periods, SH-SY5Y α-syn[A53T] cells were washed three times with cold PBS. Cell lysis was performed by addition of 400 μL ristocetin-induced platelet agglutination buffer containing complete protease inhibitor cocktail (Roche Applied Science, Mannheim, Germany). After 20 min incubation on ice, cells were scraped from 60 × 15 mm dishes and centrifuged at 4°C for 20 min at 12 500 g. The pellets, containing the insoluble fraction of the lysate, were resuspended with 40 μL ristocetin-induced platelet agglutination buffer containing 2 μL Benzonase (25 U/μL, Roche). To digest genomic DNA, resuspended pellets were incubated for 30 min at 37°C. To facilitate extraction, samples were sonified for 1 min. Protein concentration from the resulting extracts was determined using the BCA assay (Pierce). Lysates (30 μg of protein) were resolved by electrophoresis on a 4–12% Bis-Tris gradient gel (NuPAGE Novex Bis-Tris Gel, Invitrogen, Carlsbad, CA, USA) according to manufacturer’s instructions using NuPAGE MOPS buffer. After transfer to nitrocellulose membrane (Protran; Whatman, Dassel, Germany), the blot was blocked for 30 min at 21°C with Odyssey blocking buffer (Li-Cor Biosciences GmbH, Bad Homburg, Germany). The blot was probed with primary ASY-1 antibody (1 : 500, kind gift of Poul Henning Jensen, University of Aarhus, Denmark) overnight at 4°C. Primary antibodies were detected by 1 h incubation with IRDye 680 goat anti-rabbit IgG (1 : 15 000; Li-Cor) conjugated secondary antibodies in Odyssey blocking buffer and scanned at intensity level 3 with Odyssey Infrared Imaging System (Li-Cor).
Western blotting of different oligomers
Different oligomers (400 ng of protein) were resolved by electrophoresis on a 4–12% Bis-Tris gradient gel (NuPAGE Novex Bis-Tris Gel, Invitrogen) according to manufacturer’s instructions using NuPAGE MOPS buffer. After transfer to nitrocellulose membrane (Protran; Whatman), the blot was blocked for 30 min at 21°C with Odyssey blocking buffer (Li-Cor Biosciences GmbH). The blot was probed with primary ASY-1 antibody (1 : 500, kind gift of Poul Henning Jensen, University of Aarhus, Denmark) overnight at 4°C. Primary antibodies were detected by 1 h incubation with IRDye 680 goat anti-rabbit IgG (1 : 15 000; Li-Cor) conjugated secondary antibodies in Odyssey blocking buffer and scanned at intensity level 3 with Odyssey Infrared Imaging System (Li-Cor).
Thioflavin T fluorescence assays
Thioflavin T assay was modified from Munishkina et al. (2003). Briefly, α-syn (1.25 mg/mL) was incubated in 20 mM Tris and 0.1 M NaCl, pH 7.4, 75 mg/mL PEG 3350, 5 mM dithiothreitol containing 25 μM Thioflavin T. The sample volume of 45 μL was pipetted into a 384-well-plate (Corning 3575) and three steel beads with a diameter of 1 mm were added to each well. Seeding was performed by addition of 1 μL of 0.2 mg/mL oligomer type A or type C solution to each well. To avoid evaporation, 10 μL of mineral oil was added to each well. The sample plate was covered with adhesive aluminum plate sealer and incubated at 37°C with shaking at 700 rpm. The fluorescence was measured at indicated points in time in the fluorescent plate reader (SpectraMax, Molecular Devices, Ismaning Munich, Germany) in top read mode with excitation at 450 nm and emission at 490 nm. Data were measured with n = 24 and averaged for evaluations.
Measurement of the assay plates in the IN Cell Analyzer 3000TM
Automated confocal fluorescence microscopy using the IN Cell Analyzer 3000TM (GE Healthcare Bio-Sciences, Little Chalfont, UK) has been described in detail (Haasen et al. 2006; Wolff et al. 2006). For our experiments, we employed the 364 nm laser line combined with a 450BP25 emission filter for Hoechst 33342TM, the 647 nm laser line combined with a 695BP55 emission filter for Alexa Fluor 647, and the 488 nm laser line with a 535BP45 emission filter for Alexa Fluor 488. Fluorescence emission was recorded separately in the blue, red and green channel, applying flat field correction for inhomogeneous illumination of the scanned area for each of the three channels.
Measurement of the assay plates in the Opera™ QEHS device
Automated confocal fluorescence microscopy was performed on the Opera™ QEHS device (PerkinElmer, Waltham, MA, USA). For our experiments, we employed the 405 nm laser line combined with a 450BP50 emission filter for Hoechst 33342TM, the 635 nm laser line combined with a 690BP50 emission filter for Alexa Fluor 647, and the 488 nm laser line with a 540BP75 emission filter for Alexa Fluor 488. Fluorescence emission was recorded separately in the blue, red and green channel.
Fluorescence image capture with confocal microscopy
For confocal microscopy, SH-SY5Y α-syn[A53T] were seeded into collagen I coated 96-well black μ-clear bottom microtiter plates (Greiner bio-one, Frickenhausen, Germany) at a density of 6000 cells/well and treated with oligomer type C and stained as described above.
The imaging system consisted of an inverted confocal microscope (Leica DM IRBE, Leica Microsystems GmbH, Wetzlar, Germany) equipped with a Leica 63× objective and scan head (Leica TCS SP, Leica Microsystems GmbH). Images were captured using the 488 nm Argon laser line and the 543 nm Helium/Neon laser line, respectively. Single z-sections were collected from 25 to 30 (0.8 μm thick) z-sections using Leica Confocal software package (Leica Microsystems GmbH).
Image analysis was carried out using a proprietary application based on a commercially available machine vision software library (Halcon 8.0.2, MVTec Software GmbH, Munich, Germany). A local threshold applied to the blue channel segmented blue stained nuclei. Post-processing separated clusters removed debris and determined parameter ‘number of nuclei’. A dilated version of the region covered by nuclei was generated to be used as ‘cytoplasmic mask’ region in a later step. Red stained areas for α-syn were segmented by setting a threshold in the smoothed red channel of the image. Parameter ‘red area’ was determined by restricting red fluorescence to ‘cytoplasmic mask’. Neurite structures stained for α-syn were found by applying the function ‘lines-gauss’ to red channel. The length of all lines longer than 7 pixels is added up to form parameter ‘neurite length’. Yellow spots were composed of exogenous Alexa 488 labeled α-syn oligomers and red staining for α-syn (both of endogenous and exogenous origin) in overlaid images. For the quantification of yellow spots, a change in hue, saturation, intensity color space was performed. Spots above a certain saturation, appropriate hue value for yellow and within a specified size range were counted as parameter ‘yellow spots’.
Comparisons between oligomer type C and solvent control treated wells concerning the parameters ‘red area per nucleus’ and ‘neurite length per nucleus’ were performed by Kruskal–Wallis tests followed by Dunn’s multiple comparison tests. Each period of oligomer type C treatment was compared to 240 min of treatment with solvent control. It is assumed that shorter incubation times with solvent controls have no effect.
Statistical analysis for yellow spot quantification was performed by Wilcoxon signed rank tests comparing median values versus the hypothetical value 0, because solvent controls did not contain any protein.
Statistics were performed using GraphPad Prism software version 5.01 (San Diego, CA, USA).
Seeding characteristics of α-syn oligomer type C on primary cortical neurons
We have shown previously that a distinct type of soluble α-syn oligomers (seeding oligomer type C) is capable of seeding cytosolic α-syn aggregation after exogenous application to the neuroblastoma cell line SH-SY5Y (Danzer et al. 2007). To further characterize transmembrane seeding of α-syn aggregation in cell culture, we first reconfirmed the α-syn seeding effect of oligomer type C in primary cortical neurons. As illustrated in Fig. 1(a), addition of Alexa-488 labeled oligomer type C to the cell culture medium of primary cortical neurons induced the intracellular formation of large aggregates. In confocal images the aggregates had a yellow hue, indicating a co-localization of green-labeled exogenous oligomer type C and red-labeled α-syn (of both endogenous and exogenous origin). Concurrently, the cytoplasmic staining of endogenous α-syn vanished, whereas neuronal morphology seen in bright field images remained unchanged (data not shown). Corresponding solvent-treated control neurons, however, maintained a homogeneous cytoplasmic α-syn staining. Additionally, neurons treated with soluble or fibrillar α-syn showed no detectable seeding effect (data not shown).
Thus, when exogenously added, α-syn oligomer type C have entered cortical neurons and seeded aggregation of cytosolic α-syn with a resulting increased protein aggregation localized to one area.
Time-dependent seeding effect of oligomer type C on primary cortical neurons
To evaluate whether the transmembrane seeding effect of oligomer type C evolves over time, we treated primary cortical neurons with Alexa-488 labeled oligomer type C for different time periods. Only a few yellow-tinged aggregates were detected after 30 min of oligomer type C incubation but this number increased with prolonged incubation time. Concurrently, we also found a reduced cytoplasmic staining after 30 min of oligomer type C treatment. This staining was further reduced over time. Solvent-treated control neurons, however, showed after 4 h a persistent homogenous cytoplasmic staining of α-syn (Fig. 1b).
To quantify the seeding effect of α-syn oligomer type C, we determined the red area per nucleus that is composed of α-syn from both endogenous and exogenous origins. Only values that exceeded a fixed threshold were included into the analysis. In contrast to solvent control, oligomer type C-treated neurons showed a dramatic (4- to 9-fold) decrease in cytoplasmic red area per nucleus over time (Fig. 1c). Simultaneously, the number of yellow-hued aggregates in merged images (named yellow spots) increased with prolonged incubation times (Fig. 1d).
These data indicate that transmembrane seeding of cytoplasmic α-syn induced by Alexa-488 labeled oligomer type C is a time-dependent effect.
Dose-dependent seeding effect of oligomer type C on primary cortical neurons
Next we examined whether the seeding ability of Alexa-488 labeled oligomer type C depends on the concentration of α-syn oligomers. Confocal images illustrated a dose-dependent decrease in cytoplasmic staining that correlated with an increased number of aggregates (Fig. 2a).
To quantitatively assess the seeding effect dependent on increasing oligomer concentrations, we determined the length of α-syn-stained neurites per nucleus. We found that quantification of neurite length per nucleus allowed a more reproducible determination of the seeding effect than by solely measuring α-syn staining. This latter method is confounded by increasing amounts of exogenous α-syn.
We found a pronounced decrease of the neurite length per nucleus after treatment with Alexa-labeled seeding oligomer type C (Fig. 2b). Concurrently, the number of yellow spots increased with ascending oligomer concentrations (Fig. 2c). Taken together, seeding of cytoplasmic α-syn induced by oligomer type C occurs in a dose-dependent manner.
We performed the same type of experiments as described above for primary cortical neurons with the neuroblastoma cell lines SH-SY5Y and BeM17. However, seeding oligomer type C used in these experiments was not labeled with a fluorescent dye. We confirmed by confocal microscopy that both in mutant α-syn A53T-over-expressing BeM17 and SH-SY5Y cells, oligomer type C added to the cell culture medium led to a significant decrease in cytoplasmic α-syn staining and the formation of α-syn aggregates. Solvent-treated control neuroblastoma cells maintained their homogenous cytoplasmic α-syn staining (Fig. S1). We observed the seeding effect with oligomer type C also in neuroblastoma cells over-expressing wild-type α-syn (data not shown). However, because of high over-expression levels of clones over-expressing mutant α-syn A53T, the seeding effect has here been most pronounced.
The cellular studies with cortical neurons, neuroblastoma cell lines SH-SY5Y and BeM17 described above have all been performed with confocal microscopy to minimize the confounding effect of larger α-syn aggregates which might have formed in the cell culture medium and attached to the outside of the cells. The intracellular localization of the α-syn aggregates, which were observed after the addition of oligomer type C to the cell culture medium, has been confirmed by analyzing the z-stacks of confocal images. Perinuclear localization of α-syn aggregates in confocal images taken at various z-levels demonstrates that the α-syn aggregates labeled with Alexa-488 are localized within SH-SY5Y cells and not at the outside of cells. Both small punctuate red aggregates and larger yellow-hued juxtanuclear inclusions were observed intracellularly (Fig. S2). Analysis of many different SH-SY5Y cell cultures demonstrated that the yellow spots localized mainly to the perinuclear space.
To exclude the possibility that the decrease in cytoplasmic α-syn staining and punctuate staining pattern were fixation artifacts and to address the issue of the aggregate localization we analyzed transmembrane seeding with a different, microscopy technique. In this set of experiments, we added seeding oligomer type C to the cell culture media of SH-SY5Y α-syn [A53T] cells as in the previous experiments. After different incubation times, cells were washed, lysed and the insoluble fractions of the extracts were subjected to western blot analysis using α-syn specific antibody ASY-1 (Fig. 2d). Lane 1 of Fig. 4 shows the starting material, oligomer type C, which was added to the cells. The insoluble lysate fraction of cells was treated for 5 min with oligomer type C and this fraction was loaded into lane 2 of the gel. The absence of any α-syn signal indicates that the oligomer type C did not enter the cells within the first 5 min after the addition to the culture medium. This is in contrast to lane 3, where the insoluble lysates of cells treated for 22 h with oligomer type C have been applied. In this lane, high molecular weight bands of aggregated, LDS-stable α-syn species appear. These bands are probably composed of both cellular α-syn and recombinant oligomer type C, which suggests that the attachment of cytosolic α-syn to recombinant seeding oligomer type C results in LDS-stable insoluble aggregates.
Conversion of oligomer type A into type C
In our previous study, we have shown that depending on the aggregation conditions different oligomer types are forming with distinct biological effects on cells (Danzer et al. 2007). However, little is known about the dynamics and origin of the different oligomer types. To test the hypothesis that oligomerization of α-syn is a dynamic process we proceeded to convert the preformed pore-forming oligomer type A into seeding oligomer type C. In addition to the desired oligomer, both preparations contain α-syn monomers and several oligomeric forms that can be separated on a LDS gel. Pore-forming oligomer type A is characterized by a prominent band at about 30 kDa, which is absent in seeding oligomer type C. Ultrafiltration is an essential step in oligomer type C generation (Danzer et al. 2007). For that reason we prepared pore-forming oligomer type A according to the standard protocol and combined it afterwards with an ultrafiltration step. After converting pore-forming oligomer type A into seeding type C by ultrafiltration, the characteristic 30 kDa band in the LDS gel of the type A preparation is not present anymore and a characteristic band slightly higher than 50 kDa may appears which may correspond to a tetramer (Fig. 3a). This result indicates that pore-forming oligomer type A is dynamic and that this type can be converted into other species, that differ from type A oligomers in their band pattern in a denaturing LDS gel.
To evaluate whether pore-forming oligomer type A converted into seeding type C could induce transmembrane seeding, we added unlabeled oligomer type A and converted oligomer (type A into type C) to the cell culture media of SH-SY5Y over-expressing α-syn [A53T]. Interestingly, treatment of SH-SY5Y α-syn [A53T] resulted in a significant decrease in cytoplasmic α-syn staining and an increase in a punctuate staining pattern, whereas oligomer type A treated cells maintained their homogenous cytoplasmic α-syn distribution in the surviving cells (Fig. 3b).
We have reported previously that pore-forming oligomer type A induces a calcium influx but has no seeding abilities (Danzer et al. 2007). We now asked whether the conversion of pore-forming oligomer type A into seeding oligomer type C has also an influence on the calcium influx effect. Indeed, after conversion of pore-forming oligomer type A into seeding type C, the increase in intracellular calcium was abolished, whereas of oligomer type A without conversion into type C evoked a clear increase in intracellular calcium (Fig. 4). These data demonstrate that conversion of pore-forming oligomer type A into seeding oligomer type C also correlates with their biological effects.
Accelerated aggregation of monomeric α-syn by oligomer type A and C in vitro
As pore-forming oligomer type A can be converted into seeding oligomer type C, we tested whether both types of oligomers would accelerate aggregation of monomeric α-syn in a cell-free system. We assumed that that seeding should be independent of a cellular or recombinant origin of monomeric α-syn.
To test this assumption, we measured aggregation by a Thioflavin T assay after the addition of 2% (v/v) oligomer type A and type C to monomeric recombinant α-syn. As expected, addition of both oligomer types A and C resulted in a reduced lag time and an increased Thioflavin T fluorescence signal (Fig. 5). Therefore, we conclude that both oligomer types can seed amyloid formation also in a cell-free system and are along the pathway to amyloid fibrils.
Our study provides a more complete and better understanding of the α-syn seeding mechanism. We could conclusively demonstrate that a specific type of α-syn oligomers (oligomer type C) is capable to induce transmembrane α-syn seeding in a dose- and time-dependent manner. This study also indicates that transmembrane α-syn seeding is not restricted to one neuronal cell type, rather we observed seeded aggregation both in primary cortical neurons and neuronal cell lines such as SH-SY5Y and BeM17 (Fig. S1). This finding is in line with the studies of Braak et al. suggesting that the neuropathology of PD is not restricted to dopaminergic neurons in the substantia nigra and includes other non-dopaminergic neurons in various specific, extra-nigral brain regions (Braak et al. 2003a).
The seeding experiments from Murray (2003) revealed that carboxy-truncated α-syn was very efficient in seeding full length α-syn. These authors support the view that the middle region of α-syn forms the core of α-syn filaments and that negative charges in the carboxy-terminus counteract α-syn aggregation. Because of the high propensity of our oligomer type C to seed intracellular soluble α-syn, we speculate that middle region of α-syn in our seeding oligomer type C might be exposed and the C-terminal tails are hidden to facilitate α-syn aggregation.
A previous study reported that membrane-bound α-syn can seed intracellular α-syn (Lee et al. 2002). The underlying mechanism of the seeding effect described by Lee and co-workers might be related to the observed effects in our study. Seeding oligomer type C might also resemble the nucleating α-syn species as described in a cell-free system by Hoyer et al. (2002). Similar to this study, we observed that oligomer type C is a nucleating species in a cell-free system and is able to seed the aggregation of monomeric α-syn (Fig. 5). Seeding of α-syn aggregation in cell-free systems, however, seems to be a more general phenomenon, as it was also observed with pore-forming oligomer type A. Oligomer type A differs in its structure from seeding oligomer type C and does not seed α-syn aggregation in a cellular environment. This discrepancy could be explained by membrane attachment and pore formation of oligomer type A (Danzer et al. 2007), which might prevent the direct contact of the oligomer with cytoplasmic monomeric α-syn. In the cell-free system, however, pore-forming oligomer type A comes into close contact with monomeric α-syn over a long period of time.
Very little is known about the mechanism of α-syn oligomer generation in vivo. El-Agnaf and co-workers described oligomeric forms of α-syn in the plasma of PD patients (El-Agnaf et al. 2006) but the α-syn ELISA, which has been applied in this study, cannot differentiate between different oligomer types, which might be present in the samples. The same holds true for the elegant detection system of Outeiro et al. (2008), which visualizes intracellular α-syn oligomers. However, this method is capable of studying the dynamics of protein-protein interactions in living cells. In vitro, the two types of oligomers described here differ fundamentally in their properties. This observation allowed us to address the question whether the two oligomer types represent independent, static forms of α-syn oligomers or whether they are related in a more dynamic manner. Indeed, we could show that one type of oligomer can be converted into another type of α-syn oligomer. The converted oligomers differed from the starting material by the band pattern in denaturing protein gel and also by the biological effects on cells. Oligomer type A induced cell death via disruption of ion homeostasis, presumably by a pore-forming mechanism (Danzer et al. 2007). Interestingly, this pore-forming oligomer type A lacks any seeding properties in a cellular environment. After the in vitro conversion of oligomer type A into type C both oligomer types showed a similar band pattern after western blotting (Fig. 3a). Importantly, oligomer type A gained remarkable seeding properties after conversion into oligomer type C whereas the pore-forming properties, leading to an increase in intracellular calcium were completely abolished (Figs 3b and 4). Taken together, these data emphasize a clear correlation between structural properties of α-syn oligomers and their dynamic biological effects on cells.
In this study, we have focused on oligomers prepared in the presence of ferric chloride. The influence of iron on in vitro aggregation of α-syn has been reported previously by Uversky et al. (2001). The same biological effects were observed with oligomers prepared in the absence of ferric chloride; however, the biological effects were less pronounced (data not shown).
Effects of extracellular oligomers on cells might play an important role, also under pathophysiological conditions, as intriguing findings have recently shown that α-syn and α-syn aggregates could be secreted from cells and therefore possibly insult neighboring cells (Lee et al. 2005). El-Agnaf et al. detected elevated levels of soluble oligomeric forms of α-syn in plasma levels of PD patients compared to controls and suggested them as potential biomarker for PD (El-Agnaf et al. 2006). Earlier studies have shown that monomeric α-syn and pre-fibrillar aggregates of non-disease related proteins can translocate into cells, although the mechanism of aggregate internalization remains unclear to date (Bucciantini et al. 2004; Ahn et al. 2006). Lee et al. point to internalization of α-syn via endocytosis depending on the assembly of α-syn (Lee et al. 2008). A recent report provides evidence that extracellular α-syn could also be re-secreted out of neurons via a process modulated by a recycling endosome regulator rab11a (Liu et al. 2009). However, further studies are needed to fully understand the mechanism of α-syn translocation. In this study we were able to confirm that seeding oligomer type C could also be internalized by cells and enable them to develop their seeding properties. Our results also open the possibility that α-syn oligomers with the potential to induce α-syn aggregation might be present extracellularally. We hypothesize that these oligomers, in a specific conformation, might spread from one cellular system to another and induce then a seeding process that ultimately contributes to inclusion formation.
Whether these inclusions protect against or contribute to cytotoxicity is the subject of intense current debate. However, a recent report shows that cells containing α-syn inclusions are mostly present in surviving cells and less so in apoptotic cells, suggesting that these inclusions may play a protective role in cell death by sequestering toxic molecular species (Tanaka et al. 2004). Comparably, the intracellular aggregates, which form after seeding with oligomer type C, show no overt toxicity. After 24 h of treatment, the number of intact nuclei did not decline in various experiments with different cell lines. This is in sharp contrast to oligomer type A, which induced a > 70% decrease in intact nuclei after 24 h (Danzer et al. 2007). However, seeding oligomer type C might be the oligomer type responsible for propagation of α-syn pathology. Braak et al. suggested a gradually progressing, ascending course of α-syn pathology that worsens with disease duration and progresses in a predictable sequence in six stages. Although the temporal and topographical order of the histopathological lesions in sporadic PD has been described in detail, the underlying molecular mechanism of the hypothesized ordered transmission of the α-syn pathology has not been identified. Braak et al. hypothesized about an unidentified, causative pathogen which is transmitted via retrograde and transneuronal transport to the susceptible brain regions (Braak et al. 2003a,b). Similar mechanisms of transneuronal transmission have been proposed for prion diseases (Unterberger et al. 2005) as well as Alzheimer’s disease (Hardy and Selkoe 2002; Taylor et al. 2002). Based on our study we suggest that the proposed, unidentified pathogen might, in analogy to prion disease, be α-syn in a specific conformation, which induces aggregation without the propensity to self-aggregate. There are three reasons why our special oligomer type C might represent an in vitro model of the described pathogen. First, oligomer type C fosters the aggregation of soluble, homogeneously distributed cytoplasmic α-syn into insoluble inclusions. Second, induction of inclusion formation by oligomer type C proceeds in a dose- and time-dependent manner, suggesting that aging promotes the propagation of α-syn aggregation. Finally, we found the seeding ability of oligomer type C both in primary cortical neurons and in SH-SY5Y and BeM17 cell lines, indicating that the seeding effect is not restricted to a specific subset of neurons.
Further supporting evidence for the intriguing possibility that α-syn seeding might be a key player in the spreading of α-syn pathology comes from three publications in 2008 (Kordower et al. 2008; Li et al. 2008; Mendez et al. 2008). In these studies, α-syn and ubiquitin positive Lewy body-like inclusions were found in grafted embryonic neurons which have been transplanted into the striatum of PD patients more than a decade before. Interestingly, 18 months after transplantation no pathological changes were observed in grafted neurons (Kordower et al. 2008). It has been argued that either a pathogenic factor in the brain milieu exists there or there is a pathological process that can spread from one cellular system to another (Brundin et al. 2008). Our findings of the seeding abilities of oligomer type C are in line with these observations and lead us to speculate that oligomers in a specific conformation resemble the pathogenic factor in the brain milieu and that the seeding mechanism could be a pathological process that can spread from one cellular system to another.
In conclusion, we propose that α-syn oligomers form as a dynamic mixture of oligomer types with different properties and that α-syn oligomers can be converted into different types depending on the brain milieu conditions. Our pore-forming oligomer type A is similar to annular α-syn structures extracted from postmortem brain tissues from a multiple system atrophy patient (Pountney et al. 2005). Another α-syn oligomer type C has been identified with the propensity to induce α-syn aggregation in a cellular environment, which could be one possible explanation for the ascending spreading of the α-syn pathology in PD. The cellular effects of α-syn oligomers described here in cell culture could resemble events that take place in PD patients. However, further studies are needed to characterize the pathophysiologically relevant oligomeric forms in the brains of PD patients. Preventing the formation of specific types of α-syn oligomers could lead to new therapeutic strategies for the effective treatment of PD and other synucleinopathies.
The authors are grateful to S. Brezina and A. Bloching for excellent technical assistance. The authors also thank C. Ittrich for her help with statistical analysis and S.R. Edwards for carefully reading the manuscript.