α-Synuclein strains cause distinct synucleinopathies after local and systemic administration

Brain α-synuclein deposits are the hallmark of various distinct neurodegenerative diseases, and it is proposed that α-synuclein assemblies with different structural characteristics or 'strains' (ribbons or fibrils) could account for pathological differences between these diseases; here different human α-synuclein strains are injected into rat brain, and are shown to propagate in a strain-dependent manner and cause different pathological and neurotoxic phenotypes. Synucleinopathies are neurodegenerative disorders characterized by α-synuclein-rich protein deposits which include Parkinson's disease, dementia with Lewy bodies and multiple system atrophy. The discovery of α-synuclein assemblies with different structural characteristics has led to the hypothesis that different 'strains' could account for pathological differences between these different neurodegenerative diseases. This study reports that when different human α-synuclein strains — oligomers, ribbons or fibrils — are injected into rat brain in vivo, they propagate in a strain-dependent manner and cause different pathological and neurotoxic phenotypes. This work has implications for disease diagnosis and prognosis and for the prospects of developing therapeutic strategies tailored for specific synucleinopathies. Misfolded protein aggregates represent a continuum with overlapping features in neurodegenerative diseases, but differences in protein components and affected brain regions1. The molecular hallmark of synucleinopathies such as Parkinson’s disease, dementia with Lewy bodies and multiple system atrophy are megadalton α-synuclein-rich deposits suggestive of one molecular event causing distinct disease phenotypes. Glial α-synuclein (α-SYN) filamentous deposits are prominent in multiple system atrophy and neuronal α-SYN inclusions are found in Parkinson’s disease and dementia with Lewy bodies2. The discovery of α-SYN assemblies with different structural characteristics or ‘strains’ has led to the hypothesis that strains could account for the different clinico-pathological traits within synucleinopathies3,4. In this study we show that α-SYN strain conformation and seeding propensity lead to distinct histopathological and behavioural phenotypes. We assess the properties of structurally well-defined α-SYN assemblies (oligomers, ribbons and fibrils) after injection in rat brain. We prove that α-SYN strains amplify in vivo. Fibrils seem to be the major toxic strain, resulting in progressive motor impairment and cell death, whereas ribbons cause a distinct histopathological phenotype displaying Parkinson’s disease and multiple system atrophy traits. Additionally, we show that α-SYN assemblies cross the blood–brain barrier and distribute to the central nervous system after intravenous injection. Our results demonstrate that distinct α-SYN strains display differential seeding capacities, inducing strain-specific pathology and neurotoxic phenotypes.

Misfolded protein aggregates represent a continuum with overlapping features in neurodegenerative diseases, but differences in protein components and affected brain regions 1 . The molecular hallmark of synucleinopathies such as Parkinson's disease, dementia with Lewy bodies and multiple system atrophy are megadalton a-synuclein-rich deposits suggestive of one molecular event causing distinct disease phenotypes. Glial a-synuclein (a-SYN) filamentous deposits are prominent in multiple system atrophy and neuronal a-SYN inclusions are found in Parkinson's disease and dementia with Lewy bodies 2 . The discovery of a-SYN assemblies with different structural characteristics or 'strains' has led to the hypothesis that strains could account for the different clinicopathological traits within synucleinopathies 3,4 . In this study we show that a-SYN strain conformation and seeding propensity lead to distinct histopathological and behavioural phenotypes. We assess the properties of structurally well-defined a-SYN assemblies (oligomers, ribbons and fibrils) after injection in rat brain. We prove that a-SYN strains amplify in vivo. Fibrils seem to be the major toxic strain, resulting in progressive motor impairment and cell death, whereas ribbons cause a distinct histopathological phenotype displaying Parkinson's disease and multiple system atrophy traits. Additionally, we show that a-SYN assemblies cross the blood-brain barrier and distribute to the central nervous system after intravenous injection. Our results demonstrate that distinct a-SYN strains display differential seeding capacities, inducing strain-specific pathology and neurotoxic phenotypes.
The discovery that Lewy bodies spread within the nervous system 5,6 led to the demonstration that fibrillar assemblies seed soluble a-SYN and spread between cells in cellular and animal models 7 . The most widely accepted paradigm postulates that pre-fibrillar oligomers, as opposed to mature fibrils, represent the neurotoxic entities in Parkinson's disease 8,9 . Although a direct quantitative in vivo comparison of the neurotoxic potential of different a-SYN assemblies is lacking, recent evidence indicates that fibrillar a-SYN toxicity is significantly greater than that of pre-fibrillar precursors 4,10 . Studies with inoculation of recombinant fibrils or pathogenic brain lysates in rodent brain triggered a-SYN pathology with varying efficiencies and moderate neuronal loss 11,12 . In contrast, recombinant adeno-associated viral vector-mediated (rAAV) overexpression of a-SYN in rodent and primate brain induces time-dependent robust behavioural impairment and dopaminergic neurodegeneration 13,14 . In this study, we compared the in vivo properties of a-SYN oligomers and two distinct a-SYN strains ('fibrils' and 'ribbons') (Extended Data Fig. 1a-e) 4 , by inoculation of highly purified and structurally characterized recombinant human a-SYN preparations in rat substantia nigra in absence and presence of rAAV-mediated human a-SYN overexpression.
To assess a-SYN uptake and spreading in a detailed and quantitative way, we fluorescently labelled all a-SYN assemblies (Extended Data Fig. 1f-h). First, we applied ex vivo fluorescent tomography to visualize spreading of a-SYN in whole brain after inoculating 10 mg of a-SYN assemblies ( Fig. 1a and Supplementary Table 1). Reconstructed images of the fluorescence volume after 20 min and 7 days showed that a-SYN oligomers spread more efficiently than fibrils, ribbons and free fluorophore (Fig. 1b). Subsequent histological analysis revealed that all assemblies were taken up in dopaminergic neurons, spread over the arborizing striatal dopaminergic axons and were transmitted transsynaptically in a time-dependent manner (Extended Data Fig. 2a-g). These results build upon previous findings that the a-SYN oligomers disseminate more efficiently than higher molecular weight assemblies after intracerebral inoculation 15,16 .
The reason why certain Lewy body-bearing regions are subject to neurodegeneration and others not, remains incompletely understood. The fact that patients with Parkinson's disease, dementia with Lewy bodies, and multiple system atrophy present variable striatonigral pathology is also confounding. We therefore assessed whether a-SYN oligomers, fibrils and ribbons translate their distinctive structural and biochemical properties into motor behavioural and neurotoxic phenotypes. Four months after nigral injection of ribbons or fibrils, Lewy body-and Lewy neurite-like inclusions formed essentially in dopaminergic neurons. These juxta-, peri-, nuclear and filamentous inclusions stained positive for a-SYN phosphorylated at Ser129 (Pa-SYN), resembling the situation in Lewy body/Lewy neurite pathology-associated inclusions (Fig. 1c-f and Extended Data Fig. 3). Pa-SYN inclusions were considerably more abundant for a-SYN ribbons compared to fibrils with a predominant Lewy neurite-like filamentous phenotype staining positive for the aggresome marker SQSTM1 (also known as p62; Fig. 1e-g). Injection of a-SYN oligomers and brain homogenate from aged 18-month transgenic mouse expressing mutant human a-SYN(A30P) did not result in detectable Pa-SYN accumulation. Combined rAAV-mediated a-SYN overexpression with ribbons and fibrils inoculation lead to increased Pa-SYN 1 cells (Extended Data Fig. 4). Remarkably, and only after combined rAAVmediated a-SYN overexpression with a-SYN ribbons inoculation, a second but sparse a-SYN phosphorylation pattern was observed in oligodendroglial cells (Fig. 1h, i).
In terms of neurotoxicity, nigral inoculation of brain homogenate, oligomers, fibrils and ribbons did not induce detectable dopaminergic cell death until the final time point of 120 days (Fig. 2a, b). In contrast, sole rAAV-mediated a-SYN overexpression (control) led to 27% cell loss of tyrosine hydroxylase-positive neurons and 32% striatal dopaminergic axonal loss (Fig. 2c, d). Combined rAAV-a-SYN overexpression with inoculation of brain homogenate, a-SYN oligomers, fibrils or ribbons accelerated tyrosine hydroxylase-positive cell death (Fig. 2c-e) and severely reduced striatal dopaminergic nerve terminal volume by 61% and 43% for a-SYN fibrils and brain homogenate, respectively. Combined rAAV-a-SYN overexpression and inoculation of a-SYN oligomers and ribbons did not yield additional tyrosine hydroxylase fibre loss (Fig. 2d). Assessment of spontaneous forepaw use with the cylinder test revealed a consistent significant motor deficit from 60 days onwards upon injection of a-SYN fibrils (Fig. 2f). Combined rAAV-mediated a-SYN overexpression with a-SYN assemblies inoculation further aggravated motor deficits (Fig. 2g). These findings demonstrate that introduction of seeds under conditions where a-SYN is expressed at higher levels greatly enhances neurodegeneration in a strain-dependent manner. Thus, although ribbon inoculation induced more pronounced Lewy body/Lewy neurite-like inclusions ( Fig. 1c-f), fibrils imposed the largest neurotoxic burden on the striatonigral pathway (Fig. 2h).
The observation of a behavioural phenotype after inoculation of pure a-SYN strains in the absence of dopaminergic degeneration indicated that this functional deficit was not a downstream effect of dopaminergic cell death. Recent studies in cellular models, a-SYN transgenic animals and patients with incidental Lewy body disease, Parkinson's disease and dementia with Lewy bodies have shown neuronal dysfunction and motor abnormalities before overt neurodegeneration and Lewy body/Lewy neurite pathology [17][18][19] . We therefore employed cellular electrophysiology to study the impact of a-SYN assemblies on neuronal and synaptic physiology. Endogenous a/b-SYN is calculated to be over 40 mM within the rat synaptosome 20 . Four days exposure of rat primary cortical cultures to 1 mM of a-SYN assemblies left passive and excitable electrical membrane properties unaltered in single neurons, as recorded intracellularly by patch-clamp (Extended Data Fig. 5a-e). Instead, significant effects were observed on synaptic physiology. The frequency of occurrence of spontaneous bursts of action potentials revealed a direct substantial decrease after exposure to a-SYN assemblies (Fig. 2i, j). Further evidence was obtained via substrate-integrated microelectrode arrays (MEAs) allowing simultaneous extracellular monitoring of the spontaneous neuronal network electrical activity from 59 independent microelectrodes. The spontaneous firing of cortical neurons self-organizes ex vivo as episodic network-wide epochs of action potential synchrony, offering an indirect indicator of (dys)functional glutamatergic synaptic transmission. Exposure to low concentrations of different a-SYN assemblies revealed an increase in the average inter-spike intervals, which was significant for fibrils and oligomers (Fig. 2k, l).
Taken together, these data indicate that different a-SYN assemblies can affect neurotransmission after acute exposure, but only fibrillar a-SYN exhibited perpetual behavioural and aggravated neurotoxic phenotypes in vivo. To further investigate this, we assessed the capacities of a-SYN oligomers, ribbons and fibrils to seed endogenous a-SYN and to propagate, leading to long-term functional effects. We inoculated 40 mg of a-SYN assemblies divided over 3 injection sites in rat striatum in combination with rAAV-a-SYN overexpression (Fig. 3a) or not (Extended Data Fig. 6). Large amounts of sarkosylinsoluble rodent a-SYN were detected after 60 days throughout the striatonigral pathway for ribbons and fibrils ( Fig. 3a and Extended Data Fig. 6), although the inoculated human a-SYN was undetected after sarkosyl extraction in animals injected only with a-SYN assemblies (Extended Data Fig. 7). This implies that exogenous a-SYN strains seed the assembly of endogenous a-SYN, overcoming the species barrier. In addition, the proteinase K resistance of sarkosylinsoluble a-SYN extracted from the brains of animals overexpressing a-SYN inoculated with either ribbons or fibrils differed (Fig. 3b). Fibrils exhibited a higher resistance compared to ribbons, reflecting the capacity of each strain to amplify in a specific manner in vivo ( Fig. 3b and Extended Data Figs 6 and 7). Hence, fibrils possess the highest capacity to translate their neuropathological potential from a defined structural to a wide spatiotemporal level.
a-SYN knockout animals show no robust pathology, possibly because of functional redundancy (b-and c-SYN isoforms) 21 . The remarkable capacity of a-SYN strains to persist and act as seeds by imprinting their intrinsic structures, causing synaptic defects, is a, Sarkosyl-insoluble a-SYN is detected in striatum and substantia nigra after striatal inoculation of brain homogenate, oligomers, ribbons and fibrils in combination with rAAV-driven a-SYN overexpression. b, Degradation of sarkosyl-insoluble a-SYN in brain extracts from fibrils and ribbons inoculated animals with rAAV-driven a-SYN overexpression by increasing proteinase K concentrations. The results reveal that each strain imprints its intrinsic structure to endogenous a-SYN. Western blot analysis was performed using the monoclonal anti-a-SYN antibody clone 42. The data correspond to the mean and associated standard error calculated from 3 independent measurements from 2 biological replicates. evidence for a gain-of-toxic function. a-SYN fibrils can multiply by simple elongation with addition of monomers at their ends or by breakage creating additional nuclei 22,23 . This growth process appears strain-dependent, making strains the crucial pathogenic factor. Indeed, several months after nigral injection and in sharp contrast to all other assemblies, a-SYN fibrils were detectable in striatal dopaminergic axons (Extended Data Fig. 8). Therefore, and although all a-SYN assemblies might have a role during pathogenesis, fibrillar a-SYN exhibited the largest neurotoxic potential accompanied by disease-specific hallmarks at long-term (overview in Supplementary Table 2). In early stages of neurodegeneration pathological changes are often localized in defined brain areas, whereas at later stages neuronal dysfunction and associated protein aggregates are widespread. Oligomeric prion proteins invade the central nervous system acutely after crossing the blood-brain barrier or chronically via the splenic-nerve route 24 . To assess whether a-SYN assemblies cross the blood-brain barrier, or behave in a similar manner, we repeatedly injected a-SYN oligomers, ribbons and fibrils intravenously on a two-weekly basis during 4 months using two fluorescent labels, providing additional temporal resolution (Fig. 4a). No rAAV-a-SYN was used in these experiments. The red-shifted fluorescent label allowed epifluorescence imaging of the injected a-SYN strains in live animals. At the final time point we observed a striking, localized fluorescent signal originating from the brain for a-SYN oligomers, ribbons and fibrils, but not for control animals injected with saline or the free atto-647 label (Fig. 4b, c). Immunohistochemical staining revealed a-SYN deposits in the central nervous system for ribbons and fibrils (Fig. 4d, e and Extended Data Fig. 9) but only diffuse staining for oligomers (Fig. 4d, Extended Data  Fig. 10 a-f). This could indicate that a-SYN oligomers might either diffuse to a higher extent because of their molecular mass or are partly washed out during perfusion, while a-SYN fibrils and ribbons remain in place after crossing the blood-brain barrier. Detailed histological analysis showed an accumulation of atto-550-and atto-647-labelled human a-SYN fibrils in cortical neurons (Fig. 4f) and in spinal cord (Fig. 4g). This pattern was accompanied with microglial activation, which was most pronounced for fibrils throughout the spinal cord, and with redistribution of the early neuronal injury marker GAP43 (Fig. 4d, e and Extended Data Fig. 10g). These observations demonstrate that, in this experimental setup, the spread of a-SYN assemblies is not restricted to the intracerebral route but extends beyond the central nervous system.
The findings that a-SYN spread after intramuscular and gastric injection 25,26 , and the demonstration of systemic spread we make here are striking. The contribution of these routes to a-SYN assemblies spread and associated pathology deserves further investigation. In contrast to rapidly transmitted oligomers, the in vivo behaviour of slowly amplifying a-SYN strains may provide a basis for the observed heterogeneity in synucleinopathies and its slow but relentless pattern of spreading. This underlines the need for therapeutic strategies aimed at changing the surface properties of a-SYN assemblies either using epitope-or strain-specific antibodies or molecular chaperone-derived peptides 27 to interfere with the propagation of disease-associated a-SYN assemblies.
Online Content Methods, along with any additional Extended Data display items and Source Data, are available in the online version of the paper; references unique to these sections appear only in the online paper.

Production and purification of recombinant a-SYN and rAAV-a-SYN.
Recombinant a-SYN production was performed as previously described 4 . Vector production and purification was performed as previously described 28 .
The plasmids include the constructs for the AAV2/7 serotype, the AAV transfer plasmid encoding the human A53T mutant a-synuclein transgene under the control of the ubiquitous CMVie enhanced synapsin1 promoter and the pAdvDeltaF6 adenoviral helper plasmid. Real-time PCR analysis was used for genomic copy determination. Endotoxin detection. Protein endotoxin detection was performed using LAL Chromogenic Endotoxin Quantitation Kit (Thermofisher catalogue no. 88282) following the manufacturer's instruction. Briefly, 10 mg of a-SYN monomers, fibrils, ribbons and unbound Atto-550 were dissolved in 50 ml of endotoxin-free water provided in the kit and distributed in a 96-well microplate. The LAL reagent was added next and the plate was incubated 10 min at 37 uC in a thermomixer (Eppendorf, Germany). The chromogenic reagent was next added and the incubation further extended 6 min. The reaction was arrested with 25% acetic acid. Absorbance at 405 nm was measured using a Flexstation microplate reader (FlexStation, Molecular Devices). Protein endotoxin units are presented in Extended Data Fig. 1. rAAV-a-SYN vector endotoxin detection was performed using LAL bacterial endotoxin test (Endosafe-PTSTM, Charles River) following the manufacturer instruction. Briefly, rAAV a-SYN vector was loaded in the sample reservoir, mixed with the LAL reagent and combined with the chromogenic substrate. The Optical density was measured and analysed against an internally achieved standard curve. The measurements were performed in quadruple. Endotoxin units were determined to be 0.0000114 endotoxin unit per injection. Generation and labelling of a-SYN assemblies. a-SYN fibrils and ribbons were generated as previously described 4 . Oligomeric a-SYN was generated by incubating monomeric a-SYN 4 uC for 7 days. Oligomeric a-SYN was separated from monomeric a-SYN by size exclusion chromatography using a Superose6 HR10/30 column (GE Healthcare) equilibrated in phosphate buffered saline (PBS) buffer.
a-SYN assemblies in PBS were labelled by addition of 2 molar excess of the aminoreactive fluorescent dye atto-550 or atto-647 (ATTO-Tech GmbH). Labelling was performed following the manufacturer's recommendations. Unreacted dye was removed by size exclusion chromatography or three cycles of sedimentation and suspension in PBS for oligomers or fibrils and ribbons, respectively. The amount of incorporated dye was assessed by mass spectrometry. The samples were de-salted (with 5% acetonitrile, 0.1% trifluoroacetic acid (TFA)) and eluted from a C18 reversed-phase Zip-Tip (Millipore, Billerica, MA, USA) in 50% acetonitrile, 0.1% TFA. Peptide samples were mixed in a ratio of 1:5 to 1:20 (v/v) with sinapinic acid (10 mg ml 21 ) in 50% acetonitrile and 0.1% TFA) and spotted (0.5 ml) on a stainless steel MALDI target (Opti-TOF; Applied BioSystems). MALDI-TOF-TOF MS spectra were acquired with a MALDI-TOF-TOF 5800 mass spectrometer (Applied Biosystems) using linear mode acquisition. External calibration was performed using unmodified WT a-SYN. Acquisition and data analysis were performed using the Data Explorer software from Applied Biosystems.
The nature of all a-SYN assemblies used was routinely assessed using a Jeol 1400 (Jeol Ltd, Peabody, MA) Transmission Electron Microscope (TEM) after adsorption of the samples onto carbon-coated 200-mesh grids and negative staining with 1% uranyl acetate. The images were acquired with a Gatan Orius CCD camera (Gatan). Stereotactic injections. All animal experiments were carried out in accordance with the European Communities Council Directive of 24 November 1986 (86/609/ EEC) and approved by the Bioethical Committee of the KU Leuven (Belgium). Young adult female Wistar rats (Janvier, France) weighing about 200-250 g were housed under a normal 12 h light/dark cycle with free access to pelleted food and tap water. All surgical procedures were performed using aseptic techniques and ketamine (60 mg per kg intraperitoneally, Ketalar, Pfizer, Belgium) and medetomidine (0.4 mg per kg, Dormitor, Pfizer) anaesthesia. Following anaesthesia the rodents were placed in a stereotactic head frame (Stoelting, IL, USA). Injections were performed with a 30-gauge needle and a 10-ml Hamilton syringe. After making a midline incision of the scalp, a burr hole was drilled in the appropriate location for the substantia nigra (single injections) or striatum (triple injections) at the right site of the skull. Stereotactic coordinates, dose and concentration of a-SYN protein and rAAV-a-SYN vectors are described in Supplementary Table 1. Brain homogenate was extracted from 18 month old Thy-1 A30P transgenic mice 29 as described previously 30 adjusted to a total amount of 10 mg. Supplemented viral titres were diluted to 3.0 3 10 10 genome copies per ml A53T a-SYN rAAV2/7. Fluorescent imaging. Epifluorescent images and fluorescent tomography were performed using the IVIS Spectrum in vivo imaging system (Caliper, PerkinElmer, version 4.3.1). Anaesthesia was induced in an induction chamber with 3% isoflurane in 100% oxygen at a flow rate of 1 l min 21 and maintained in the IVIS with a 1.5% mixture at 0.5 l min 21 . Rats were injected with 500 ml saline supplemented with 10 mg of fluorescently labelled a-SYN protein intravenously in the tail vein. Subsequently, they were placed in the prone position in the IVIS and epifluorescent images of atto-647 a-SYN labelled proteins were acquired at the final time point. Atto-550 a-SYN labelled proteins could not be imaged due to high autofluorescent background levels in live animals. Each frame depicts the fluorescent image as a pseudocolour image superimposed on the grey-scale photographic image. The data are reported on a logarithmic scale as the fluorescent yield or fluorescence emission radiance per incident excitation intensity (p s 21 cm 22 sr 21 mW 21 cm 22 ) from a 2 cm 2 circular region of interest around the head.
Fluorescent tomography ex vivo images of a-SYN spreading were acquired by merging 11 distinct and predefined transillumination excitation sources of perfused and fixed rat brain. A small illumination spot is scanned over a predefined grid resulting in distinct illumination foci. The excitation source was positioned below the brain with the delivery of a concentrated excitation light beam and observed by a detector located above the object. Superimposing surface topography and fluorescent signal results in fluorescent tomography providing a 3D reconstruction of atto-550-tagged a-SYN fluorescent volume. Emitted fluorescent light propagation is calculated through diffusion approximation allowing the reconstruction of the fluorescent source in 3D. 3D reconstruction is presented as voxels with fluorescent intensity profile for each condition and is mapped to a 3D brain atlas. Fluorescence yield is expressed in units of pmol M 21 cm 21 . A threshold % value was set to keep the minimum count of pixels above 200 to exclude background pixels from the reconstruction. The accuracy of reconstruction was checked for all samples by comparing the measured light diffusion pattern with the simulated light diffusion pattern. If the % error between these two has a near-zero % difference, the reconstruction was considered successful and used for spatial measurements. Behavioural testing. The cylinder test was used to measure asymmetry in spontaneous forelimb use. Contacts made by each forepaw with the wall of 20 cm-wide clear glass cylinder were scored from the videotapes by an observer blinded to the animal's identity. Between 20 and 30 wall touches per animal (contacts with fully extended digits executed with the forelimb ipsilateral and contralateral to the lesion) were counted. The number of impaired forelimb contacts was expressed as a percentage of total forelimb contacts. Non-lesioned control rats should score around 50% in this test. No habituation of the animals to the testing cylinder was allowed before video recording. An investigator blind to different groups performed all the analyses. Histology. Rats were euthanized with an overdose of sodium pentobarbital (60 mg per kg intraperitoneally, Nembutal, Ceva Santé, Belgium) followed by intracardial perfusion with 4% paraformaldehyde in PBS. After postfixation overnight, 50 mm thick coronal brain sections were made with a vibrating microtome (HM 650V, Microm, Germany). IHC was performed as previously described 31 using antibodies summarized in Supplementary Table 3. For spinal cord staining, commercial TNB blocking buffer (Perkin Elmer) was added to all incubation steps to minimize background staining in 50 mm vibratome sections.
For fluorescent double or triple staining, sections were rinsed three times in PBS and incubated overnight in the dark in PBS-0.1% Triton X-100 and 0.1% sodium azide, 10% donkey or goat serum and antibodies summarized in Supplementary  Table 3. After rinsing, the sections were incubated in the dark for 2h in fluorochrome-conjugated secondary antibodies. After being rinsed and mounted, the sections were coverslipped with Mowiol. Fluorescent staining was visualized by confocal microscopy (Fluoview 1000, Olympus). Reconstructed 3D images from z-stack tiles were imaged using an Olympus confocal laser-scanning microscope and rendered into the complete neuronal reconstructions by tiling the entire set of z-stacks using Imaris software. Stereological quantification. The number of TH-positive cells and a-SYN-positive cells in the substantia nigra was determined by stereological measurements using the optical fractionator method in a computerized system as described before 32 (StereoInvestigator; MicroBright-Field, Magdeburg, Germany). Every fifth section throughout the entire substantia nigra was analysed, with a total of seven sections for each animal. The coefficients of error, calculated according to the procedure of Schmitz and Hof as estimates of precision 33 varied between 0.05 and 0.10. The volume of a-SYN expression in the brain and the substantia nigra was determined by the Cavalieri method. Every fifth section covering the entire extent of the substantia nigra, with a total of seven sections for each animal, was included in the counting procedure. The coefficients of error varied between 0.05 and 0.14. We quantified both the injected and non-injected substantia nigra (internal control), no cell loss was observed in the non-injected side. An investigator blind to different groups performed all the analyses.

LETTER RESEARCH
Sarkosyl extraction. Unfixed brains were dissected to isolate substantia nigra and striatum from both hemispheres and frozen in liquid nitrogen. Brain samples were weighted and homogenized at 10% (w/v) in PBS buffer supplemented with phosphatase inhibitor cocktail (sigma P5726 and P0044) and protease inhibitors (Roche Complete EDTA free). Sarkosyl was added to 1%. 250 ml were centrifuged at 1,000g for 30 min to remove cell debris. The supernatants (200 ml) were further centrifuged at 200,000g for 60 min at 20 uC in an Optima MAX-XP Ultracentrifuge (Beckman). The pellets were washed with PBS-1% Sarkosyl, and resuspended in 200 ml of PBS buffer. Proteinase K digestion, SDS-, native-PAGE and western blot analysis. Sarkosyl insoluble fractions containing a-SYN were incubated at 25 uC with proteinase K (0.001, 0.01, 0.1, 1 mg ml 21 ) (Roche) for 30 min. Aliquots were mixed v/v with denaturing buffer (50 mM Tris-HCl, pH 6.8, 4% SDS, 2% b-mercaptoethanol, 12% glycerol and 0.01% bromophenol blue) and heated at 90 uC to arrest immediately the cleavage reaction. After incubation for 10 min at 90 uC, the samples were processed to monitor the time course of a-SYN cleavage by SDS-PAGE (15% polyacrylamide gels) followed by western blotting and detection with Clone42 monoclonal antibody (BD Bioscience ref. 610787) or anti-phospho-Ser 129 a-SYN antibody. For native-PAGE, sarkosyl insoluble fractions subjected or not to PK treatment were supplemented with Triton X-100 (0.2%, v/v) b-mercaptoethanol (2%), glycerol (12%) and bromophenol blue (0.01%) and loaded on 15% native-PAGE. The samples were transferred to nitrocellulose membranes and probed with D37A6 or Syn 211 anti-rodent and human a-SYN antibodies (Supplementary Table 3). Electrophysiology and MEA recordings. Primary cortical cultures were obtained from postnatal Wistar rats, employing standard methods 34 . Briefly, after mechanical and enzymatic dissociation of the cerebral cortices of postnatal day 0 (P0) pups, cells were plated at a final density of ,1.500 mm 22 on glass coverslips and of 6.500 mm 22 on commercial microelectrode arrays (MEAs; Multichannel Systems GmBH, Reutlingen, Germany). Individual MEAs incorporated 59 substrate microelectrodes made of indium titanium nitrate, featuring a diameter of 30 mm and an electrical impedance of ,100 kV at 1 kHz. The inter-electrode spacing was 200 mm, and the microelectrodes arranged in a regular 8 3 8 layout that covered an area of ,2.56 mm 2 . MEAs and coverslips were coated overnight with polyethyleneimine 0.1% (w/v) in millQ water at room temperature, before cell seeding. Seeded MEAs and coverslips were maintained under sterile conditions at 37 uC under 5% CO2 atmosphere with 100% relative humidity for several weeks.
Patch-clamp intracellular recordings were performed 22-23 days after plating, from neurons cultured under control conditions (that is, 1 mM bovine serum albumin) or incubated for 3-4 days with 1 mM a-SYN (that is, as ribbons, fibrils, oligomers). Patch-clamp recordings were made from the cell somata under the whole-cell configuration using an Axon Multiclamp 700B Amplifier (Molecular Devices LLC, Sunnyvale, CA, US), employing both current-clamp and voltageclamp modes. Patch electrodes were pulled from thick-walled borosilicate glass capillaries (1BF150, World Precision Instruments, Hitchin, UK) by a horizontal puller (P97, Sutter, Novato, US) to a resistance of 6-7 MV. Electrodes were filled with an intracellular solution containing (in mM): 135 K-gluconate, 10 KCl, 10 HEPES, 0.2 EGTA, 4 Mg-ATP, 0.4 Na 3 GTP, and 14 Na 2 -phosphocreatine (pH 7.3, titrated with KOH). All recordings were obtained at 34 uC and replacing the culture medium by an artificial cerebrospinal fluid, constantly perfused at a rate of 1 ml min 21 and containing (in mM): 145 NaCl, 4 KCl, 2 Na-pyruvate, 5 HEPES, 5 glucose, 2 CaCl 2 , and 1 MgCl 2 (pH adjusted to 7.4 with NaOH). Voltage and current traces were sampled at 15 kHz, analog to digital converted at 16 bits, and stored on a PC employing the software LCG 35 . Data were analysed off-line, employing custom scripts written in MATLAB (The MathWorks, Natick, US) or Excel (Microsoft, USA).
Extracellular recordings experiments were performed 15-20 days after plating, and continued until day 28, from neurons cultured over MEAs under control conditions (that is, 1 mM bovine serum albumin) or incubated for up to 10 days with 1 mM a-SYN (that is, as ribbons, fibrils, oligomers). Recordings were performed by a MEA1060BC amplifier, (Multichannel Systems GmBH, Reutlingen, Germany) inside an electronic-friendly incubator, maintaining 37 uC, 5% CO 2 , and 100% relative humidity. This allowed us to monitor simultaneously, chronically, and non-invasively the neuronal electrical activity from up to 60 locations within the very same culture. MEA recordings were analysed off-line by QSpikeTools 36 . Briefly, the extracellular electric fields were monitored from the 59 independent electrodes in each MEAs were sampled at 25 kHz per channel, 1,2003 amplified, bandpass-filtered (200-3,000 Hz), and digitally recorded for 30 min every day, up to 6 days following the bath application of a-SYN, dissolved in culture medium with distinct concentrations (that is, 0.1, 0.5, and 1 mM). Simple spike-sorting, based on spike-shape peak polarity, was performed. Epochs of spontaneous synchronized firing across the MEAs electrodes were identified over 25 ms bins by standard procedures 36 , and their features were extracted from each recording session. Statistical analysis. Results are expressed as means 6 standard error of the mean. Normality of data was assessed with the Lilliefors test. Statistical significance was assessed using one-way ANOVA followed by Dunnet's or Tukey's post-hoc test and column statistics followed by Bonferroni's post-hoc test. Null hypothesis was rejected if the P value was below 0.5. Statistical significance level was set as follows: * if P , 0.05, ** if P , 0.01, *** if P , 0.001. In case of improper fitting of data by a normal distribution, median and interquartile range were used for data representation, and Wilcoxon rank sum test was used for analysis.
No statistical methods were used to predetermine sample size. Randomization was not applied because all animals (acquired via commercial animal lab, Janvier) used in this study were similar for age, sex and strain background. Fibrils appear to persist to a much higher extent compared to ribbons ( ### P , 0.001, one-way ANOVA with Tukey's multiple comparison test, n 5 4). Oligomers and brain homogenate show very low immunofluorescent staining for human-specific a-SYN. d, Fibrils appear as axonal inclusions and are abundantly present in striatum in contrast to ribbons after nigral injection. Oligomers and brain homogenate are not detected using the same human-specific antibody Syn211. Scale bar indicates 40 mm.

LETTER RESEARCH
Extended Data Figure 10 | Intravenous injection of a-SYN strains leads to increased microglial response and distinct histopathological hallmarks in the spinal cord. Quantification of human a-SYN (hua-SYN) density and Mac1/CD11b microglial response following intravenous administration of a saline solution or a-SYN oligomers, ribbons or fibrils in a, b, cervical, c, d, thoracic and e, f, lumbosacral segments (**P , 0.05, one-way ANOVA with Dunnet's multiple comparison test versus control, ### P , 0.001, ## P , 0.01, # P , 0.05, one-way ANOVA with Tukey's multiple comparison test, n 5 4). Error bars indicate s.e.m., white boxes and grey boxes indicate 60 and 120 days after intravenous injections, respectively. g, Immunohistochemical staining for growth associated protein 43 (GAP43), an early injury marker, shows a filamentous distribution for a-SYN ribbons and cellular distribution for a-SYN fibrils (white arrow heads). Scale bars indicate 50 mm.