SCIENCE ADVANCES | RESEARCH ARTICLE NEUROSCIENCE Copyright © 2020 The Authors, some Identification of distinct pathological signatures rights reserved; exclusive licensee induced by patient-derived -synuclein structures American Association for the Advancement in nonhuman primates of Science. No claim to original U.S. Government Works. Distributed M. Bourdenx1,2*†, A. Nioche1,2,3,4*‡, S. Dovero1,2*, M.-L. Arotcarena1,2*, S. Camus1,2, G. Porras1,2, under a Creative M.-L. Thiolat1,2, N. P. Rougier1,2,5, A. Prigent6,7,8, P. Aubert6,7,8, S. Bohic9, C. Sandt10, Commons Attribution F. Laferrière1,2, E. Doudnikoff1,2, N. Kruse11,12, B. Mollenhauer11,12, S. Novello13,14, M. Morari13,14, NonCommercial T. Leste-Lasserre15, I. Trigo Damas16,17, M. Goillandeau1,2, C. Perier17,18, C. Estrada19,20, License 4.0 (CC BY-NC). N. Garcia-Carrillo21, A. Recasens17,18, N. N. Vaikath22, O. M. A. El-Agnaf22, M. T. Herrero19,20, P. Derkinderen6,7,8, M. Vila17,18,23,24, J. A. Obeso16,17, B. Dehay1,2§||, E. Bezard1,2§|| Dopaminergic neuronal cell death, associated with intracellular -synuclein (-syn)–rich protein aggregates [termed “Lewy bodies” (LBs)], is a well-established characteristic of Parkinson’s disease (PD). Much evidence, accu- mulated from multiple experimental models, has suggested that -syn plays a role in PD pathogenesis, not only as a trigger of pathology but also as a mediator of disease progression through pathological spreading. Here, we have used a machine learning–based approach to identify unique signatures of neurodegeneration in monkeys induced by distinct -syn pathogenic structures derived from patients with PD. Unexpectedly, our results show that, in nonhuman primates, a small amount of singular -syn aggregates is as toxic as larger amyloid fibrils present in the LBs, thus reinforcing the need for preclinical research in this species. Furthermore, our results provide evidence supporting the true multifactorial nature of PD, as multiple causes can induce a similar outcome regard- ing dopaminergic neurodegeneration. INTRODUCTION overexpressing either mutant or wild-type -syn (7) and recently in The seminal work of Braak and colleagues (1) suggesting that Lewy macaque monkeys (8). In agreement with these findings, we have body (LB) pathology follows a predictable pattern of progression shown that intracerebral injection of low doses of -syn–containing within the brain in Parkinson’s disease (PD) and the “host-to-graft” LB extracts, purified from the substantia nigra pars compacta observation (2–4) led to the development of experimental models (SNpc) of postmortem PD brains, promotes -syn pathology and based on injection with -synuclein (-syn; the primary protein dopaminergic neurodegeneration in wild-type mice and nonhuman component of LB) assemblies (5). These experimental models suggest primates (9). This neuropathological effect was directly linked to that -syn, in pathological conformations such as the one found in the presence of -syn in LB extracts, since immunodepletion of LBs, initiates a cascade of events leading to dopaminergic neuron -syn from the LB fractions prevented the development of pathology degeneration and cell-to-cell propagation of -syn pathology through following injection into wild-type mice. a self-templating mechanism. In this study, we aimed at characterizing this experimental model Several studies have suggested that prefibrillar oligomers may of synucleinopathy in nonhuman primates. The initial study design represent one of the major neurotoxic entities in PD (6). This was to inject fractions derived from the same PD patients containing notion has been derived primarily from studies using large doses of either soluble and small -syn aggregates (hereafter named noLB) recombinant -syn applied to cell cultures or injected in adult mice or LB-type aggregates (hereafter named LB). However, because of 1 Univ. de Bordeaux, Institut des Maladies Neurodégénératives, UMR 5293, F-33000 Bordeaux, France. 2CNRS, Institut des Maladies Neurodégénératives, UMR 5293, F-33000 Bordeaux, France. 3Institut Jean Nicod, Département d’études cognitives, ENS, EHESS, PSL Research University, 75005 Paris, France. 4Institut Jean Nicod, Département d’études cognitives, CNRS, UMR 8129, Paris, France. 5INRIA Bordeaux Sud-Ouest, 33405 Talence, France. 6INSERM, U1235, Nantes F-44035, France. 7Nantes University, Nantes F-44035, France. 8CHU Nantes, Department of Neurology, Nantes F-44093, France. 9EA-7442 Rayonnement Synchrotron et Recherche Medicale, RSRM, University of Grenoble Alpes, 38000 Grenoble, France. 10SMIS beamline, Synchrotron SOLEIL, l’orme des merisiers, 91192 Gif sur Yvette, France. 11Paracelsus-Elena-Klinik, Kassel, Germa- ny. 12University Medical Center Goettingen, Institute of Neuropathology, Goettingen, Germany. 13Department of Medical Sciences, Section of Pharmacology, University of Ferrara, via Fossato di Mortara 17-19, 44121 Ferrara, Italy. 14Neuroscience Center and National Institute of Neuroscience, University of Ferrara, via Fossato di Mortara 17-19, 44121 Ferrara, Italy. 15INSERM, Neurocentre Magendie, U1215, Physiopathologie de la Plasticité Neuronale, F-33000 Bordeaux, France. 16HM CINAC, HM Puerta del Sur and CEU–San Pablo University Madrid, E-28938 Mostoles, Spain. 17Center for Networked Biomedical Research on Neurodegenerative Diseases (CIBERNED), Instituto Carlos III, Madrid, Spain. 18Neurodegenerative Diseases Research Group, Vall d’Hebron Research Institute (VHIR)–Center for Networked Biomedical Research on Neurodegenerative Diseases (CIBERNED), Barcelona, Spain. 19Clinical and Experimental Neuroscience Unit, School of Medicine, Biomedical Research Institute of Murcia (IMIB), University of Murcia, Campus Mare Nostrum, 30100 Murcia, Spain. 20Institute of Research on Aging (IUIE), School of Medicine, University of Murcia, 30100 Murcia, Spain. 21Centro Experimental en Investigaciones Biomédica (CEIB), Universidad de Murcia, Murcia, Spain. 22Neurological Disorders Research Center, Qatar Biomedical Research Institute (QBRI), Hamad Bin Khalifa University (HBKU), Education City, Qatar. 23Department of Biochemistry and Molecular Biology, Autonomous University of Barcelona (UAB), Barcelona, Spain. 24Catalan Institution for Research and Advanced Studies (ICREA), Barcelona, Spain. *These authors contributed equally to this work. †Present address: Albert Einstein College of Medicine, Department of Developmental and Molecular Biology, 1300 Morris Park Ave., Bronx, NY 10461, USA. ‡Present address: Aalto University, School of Electrical Engineering, Department of Communications and Networking, 02150 Espoo, Finland. §Corresponding author. Email: benjamin.dehay@u-bordeaux.fr (B.D.); erwan.bezard@u-bordeaux.fr (E.B.) ∥These authors contributed equally to this work as co-last authors. Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 1 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE the unexpected finding that nonhuman primates, unlike mice, are that, in the range of 1590 to 1700 cm−1, the LB group contained a susceptible to soluble or finely granular -syn, we sought to elucidate fraction of amyloid aggregates with different amyloid structures from the response characteristics induced by either LB or noLB fractions. those in the noLB group, as PCA clearly segregated them in two To achieve a thorough analysis of these -syn–related characteristics, clusters (fig. S2, D and E). Together, these results suggest that while we took advantage of the strength of machine learning algorithms LB fractions primarily contained large aggregated -syn fibrils, noLB for finding fine patterns among complex sets of data and developed fractions contained soluble -syn and a smaller enrichment of -syn a new method compatible with the constraints of experimental biology. aggregates featuring a specific amyloid structure not found in LB fractions. Here, we report on the identification of primate-specific responses Data from several studies suggest that both recombinant -syn to selected -syn assemblies and their accompanying species-specific preformed fibrils (10, 11) and patient-derived -syn (9) can promote pathogenic mechanisms. Overall, our results support the concept of pathogenic templating of endogenous -syn, ultimately leading the multifactorial nature of synucleinopathies. to dopaminergic neurodegeneration in SNpc. The -syn content quantified by ELISA, as previously reported, was (9) noLB (41.33 pg/l) and LB (24.52 pg/l). To allow comparison between experimental RESULTS groups, we diluted both pools of fractions to the least concentrated Purification and characterization of -syn extracts pool (24 pg/l). Then, those fractions were tested for their patho- from patients with PD genic effects on tyrosine hydroxylase (TH)–positive dopaminergic noLB and LB fractions were obtained from the SNpc of five sporadic neurons in primary mesencephalic cultures (fig. S3A) and in wild- PD brains exhibiting conspicuous LB pathology. The samples were type mice. Four months after supranigral injection, LB-injected mice processed through differential ultracentrifugation in a sucrose displayed, as expected, significant dopaminergic degeneration, while gradient and analyzed for the presence of -syn aggregates by filter noLB-injected mice showed no dopamine neuron loss (Fig. 1, G and H). retardation assay (Fig. 1A) (9). Further characterization of noLB This confirms previously reported toxicity of other SNpc-derived LB and LB fractions was performed by colocalization of -syn and the fractions (9), validating the toxicity of the preparation to further injec- amyloid dye Thioflavin S (Fig. 1B) as well as ultrastructural exam- tion into nonhuman primates. ination by electron microscopy (Fig. 1C). These assays confirmed the presence of misfolded -syn in both fractions. We also performed Intrastriatal injection of LB and noLB fractions from biochemical characterization of the stability of assemblies after patients with PD induces nigrostriatal neurodegeneration proteinase K digestion (Fig. 1D) and detergent treatments (Fig. 1E), in baboon monkeys followed by -syn dot-blot assays. While total -syn content was To determine the mechanisms of -syn aggregate toxicity in a species comparable between selected fractions [as measured by -syn closer to humans, adult baboon monkeys (n = 4 to 7 per experimental enzyme-linked immunosorbent assay (ELISA)], LB fractions showed group) received bilateral stereotactic injections (100 l) of either LB higher resistance to proteinase K treatment (noLB t1/2 = 15.23 min or noLB fractions into the putamen. Since our previous studies in- versus LB t1/2 > 60 min) (Fig. 1D) and greater resistance to multiple dicated that ongoing pathogenic effects can already be measured detergents, including 8 M urea (Fig. 1E). We then measured the at 14 months after LB injection, we decided to further extend this content of -syn aggregates using human -syn aggregation time-­ time frame to 24 months in the attempt to potentially observe resolved Förster’s resonance energy transfer (TR-FRET)–based disease-relevant lesions. Two years after administration, LB-injected immunoassay, which revealed a significantly higher amount of monkeys displayed significant striatal dopaminergic terminal loss aggregated -syn in LB fractions (Fig. 1F). To obtain insight into both in the putamen and in the caudate nucleus, which was accom- the content of monomeric and aggregated -syn within noLB and panied by a significant decrease in TH immunoreactivity in the LB fractions of patients with PD, we applied sarkosyl treatment SNpc (Fig. 2). Stereological counts showed that LB-injected animals to both fractions to induce physical separation and then performed exhibited TH- and Nissl-positive cell loss in the SNpc (16 and 23%, velocity sedimentation and density floatation gradients to quantify respectively; control animals, 176,719 ± 13,110 TH-positive neurons; these two populations and determine their relative abundance in LB-injected animals, 139,104 ± 13,256 TH-positive neurons; noLB-­ each fraction (fig. S1, A to H). Notably, while LB fractions contained injected animals, 138,046 ± 16,615 TH-positive neurons). No overt ~90% of aggregated -syn, noLB fractions were composed of ~10% of parkinsonism was observed, however, since the extent of the lesion this pathological form of the protein (fig. S1I). In addition, to confirm remained below the threshold for motor symptom appearance, i.e., the quality of the LB extraction, we performed a filter retardation assay, 45% of cell loss (12). However, using a validated ethological evalua- which showed that LB, but not noLB, fractions were highly enriched in tion, we were able to unravel changes in the behavioral profiles known components of LBs, such as phosphorylated S129 -syn, ubiquitin, characterized by a decrease in spontaneous locomotion associated p62, hyperphosphorylated tau, and -amyloid (A) (fig. S2A). with social isolation, following LB and noLB injections (fig. S4). Micro-infrared spectroscopy of LB and noLB fractions was per- These two phenotypes could be reminiscent of early parkinsonism, formed to show conformational changes in amyloid structures at notably the loss of interest for reinforcing situation (i.e., social the molecular level (fig. S2, B to E), and this confirmed the presence interaction), and compatible with the level of nigrostriatal neuro- of  sheet structures in both assemblies (fig. S2, B and C). Although degeneration following LB and noLB injections. their velocity of sedimentation and density floatation characteristics At odds with mice either generated for the purpose of this study were similar, the aggregates present in the LB and noLB fractions (Fig. 1, G and H) and previously characterized (9) or produced in were different in nature on the basis of the evidence of micro-infrared the context of other in-house studies, noLB-injected monkeys showed spectroscopy. Principal components analysis (PCA) showed that, in degeneration of the nigrostriatal pathway including dopaminergic LB fractions, large aggregates corresponding to the major pieces of cell loss (i.e., 16% of TH-positive neurons and 28% of Nissl-positive LB were present (fig. S2D, cluster on the right). PCA further showed neurons quantified by stereology), similar to that observed in Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 2 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE Fig. 1. Purification and characterization of LBs and noLB inocula from PD brains. (A) Left: Immunohistochemistry image of -syn–positive LB (arrows) in nigral postmortem brain samples (PD #1; -syn in brown and neuromelanin in dark brown) before sucrose gradient purification. The pie chart indicates the relative contribution of the five patients to the final pool of LB and noLB inocula. Middle: Schematic representation of the sucrose gradient fractionation procedure used to purify LB/noLB-containing fractions from freshly frozen postmortem nigral brain tissue of five sporadic patients with PD. Right: Filter retardation assay probed with a human -syn antibody to assess the presence of -syn aggregates in the different fractions obtained by sucrose gradient fractionation from freshly frozen postmortem nigral brain tissue from sporadic patients with PD (PD #1). Green square indicates noLB-containing fraction, and blue rectangle highlights LB-containing fraction selected to prepare the mixture used for injections. (B) Confocal exam- ination of purified noLB and LB fractions with -syn immunofluorescence (red) and thioflavin S (green) staining. Both LB and noLB present thioflavin S–positive aggregates but much smaller in noLB fractions. Scale bar, 10 m. (C) Ultrastructural examination of noLB and LB fractions by electron microscopy showing massive fibrils in LB fractions, while noLB fractions contain, besides soluble -syn, some punctiform small-size aggregates. (D) noLB and LB fractions derived from PD brains (left) were treated with proteinase K (PK) (1 g/ml) for 0, 15, 30, 45, and 60 min and analyzed by immunoblotting with syn211 antibody. The median effective concentration (EC50) value was determined as the concentration at which this ratio is decreased by 50%. The corresponding EC50 value for LB (>60 min) was approximately fourfold greater than with noLB (15.23 min). (E) noLB and LB fractions were treated for 6 hours with increasing concentrations of either urea or SDS or buffer as control. Syn211 was used to detect the forms of -syn. The LB fractions appear to be more resistant to breakdown compared with noLB fractions in both urea (F1,8 = 6.063, P = 0.0392) and SDS treatments (F1,12 = 17.41, P = 0.0013). The dotted lines show levels of control fractions. Comparison was made using two-way analysis of variance (ANOVA). (F) TR-FRET immunoassay analysis of noLB and LB fractions. Fluorescence measurements were taken 20 hours after antibody. Analysis by unpaired Student’s t test (t7 = 2623, P = 0.0343). *P < 0.05. Means ± SEM, n = 4 to 5. (G) Representative pictures of tyrosine hydroxylase (TH)–positive SNpc neurons (brown; Nissl staining in purple) in noninjected and noLB- or LB-injected mice at 4 months after injections. Scale bar, 500 m. (H) Quantification of TH-positive SNpc neurons by stereology in control and LB- and noLB-injected mice. Control mice, n = 10; LB-injected mice at 4 months, n = 10; noLB-injected mice at 4 months, n = 10. Comparison was made using one-way ANOVA followed by Tukey test for multiple comparisons. *P < 0.05 compared with control and noLB-injected mice at 4 months. Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 3 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE Fig. 2. Intrastriatal injection of LB and noLB fractions from patients with PD induces nigrostriatal neurodegeneration in baboon monkeys. (A) TH staining at striatum and SNpc levels. A green fire blue LUT (lookup table) was used to enhance contrast and highlight the difference between noninjected and LB- and noLB-injected baboon monkeys at the striatum level. Scale bars: 5 mm (striatum) and 10 m (SNpc). (B) Scatter plot of TH immunostaining in SNpc (F2,14 = 9.439, P = 0.0025; control versus LB-injected, P = 0.0029; control versus noLB-injected, P = 0.0248). (C and D) Scatter plots of mean gray values of striatal TH immunoreactivity in the putamen (F2,14 = 7.313, P = 0.0067; control versus LB-injected, P = 0.0059) (C) and in the caudate (F2,14 = 16.25, P = 0.0002; control versus LB-injected, P = 0.0008; control versus noLB-injected, P = 0.0008) (D) in noninjected and LB- and noLB-injected baboon monkeys. The horizontal line indicates the average value per group ± SEM (n = 7 from control animals; n = 6 for LB-injected animals; n = 4 for noLB-injected animals). Comparison was made using one-way ANOVA and Tukey’s correction for multiple comparisons. *P < 0.05 compared with control animals. LB-injected monkeys (Fig. 2). Facing such an unexpected finding, is able to approximate virtually any functions (14). This algorithm we aimed to identify specific characteristics of the pathological was given, as input, the data obtained for each animal for the 163 mechanisms involved in -syn toxicity induced by each fraction inde- aforementioned variables, and its output is a rank of these variables pendently using a large-scale approach in combination with ma- regarding their ability to predict three indicators of dopaminergic chine learning for pattern identification. tract integrity, which were levels of TH staining in (i) the SNpc, (ii) the putamen, and (iii) the caudate nucleus. Machine learning algorithm predicts The main difficulty was to overcome the large number of input nigrostriatal degeneration variables (163) compared to the sample size (n = 4 to 7 per group), We performed an exploratory approach and aimed to distinguish which can induce a selection and reporting bias (15). To tackle this relevant variables allowing accurate prediction of neurodegeneration “p > n” problem, instead of using a single network that could be (i.e., to operate a feature selection). Overall, we investigated a large prone to overfitting, we put in competition several networks. number of variables tapping on behavioral, histological, biochemical, Each MLP was composed of a single hidden layer of three neurons transcriptional, and biophysical approaches (Fig. 3A) applied to (Fig. 3C). It has as input a subset of three variables (of the 163) and several brain areas (n = 40; Fig. 3B), totalizing 180 variables measured as output the three indicators of dopaminergic tract integrity. In for each individual (fig. S5A for variable abbreviation nomenclature, total, we used 708,561 sets of three input variables. Every instance of table S1 for exhaustive list of variables, and table S2 for all raw data). MLP was trained with 80% of our sample (always a combination of We first extracted from this dataset every variable that actually control and injected animals) and tested on the remaining 20%. The quantified neurodegeneration (i.e., dopaminergic markers such as performance of each set of three input variables was evaluated TH or dopamine transporter evaluated by immunohistochemistry), according to the difference between the predicted values of TH ending up with 163 variables per animal. staining and the actual ones. Then, to operate feature selection, we designed a distributed We focused on the top 1% of the best networks and counted the algorithm using multiple-layer perceptron (MLP) (13), a classic occurrence of each of the 163 variables in the subset of three variables machine learning algorithm based on artificial neural network that used by these best networks (Fig. 3C). We ranked each variable Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 4 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE Fig. 3. MLP-based identification of specific signature. (A) Several end points (n = 180) were measured using multiple methods (colors). End points can be grouped as clusters: (1) dopaminergic degeneration, (2) behavior, (3) -syn–related pathology, (4) non–-syn–related pathology, and (5) putative biomarkers. PTM, posttranslational modification; CSF, cerebrospinal fluid; qRT-PCR, quantitative reverse transcription polymerase chain reaction; DA, dopamine. (B) Multiple brain regions (n = 40) were in- vestigated from coronal sections at two levels: anterior commissure (ac.), −3 mm (striatum and entorhinal cortex) and −7 mm (SNpc and hippocampus). SXRF, nano-­synchrotron X-Ray fluorescence. (C) Detailed methodology. (1) Representative scheme of one MLP predicting three neurodegeneration-related variables (ND1, ND2, and ND3) with three experimental variables as input (var1, var2, and var3). Of the 180 variables measured in total, 163 were used as inputs for the MLP. (2) One MLP was trained for every unique combination of three variables. (3) Combinations were ranked on the basis their prediction error, and top 1% was selected for further anal- ysis. (4) Combinations were deconvoluted to extract single variables and count occurrence of individual variables. (5) Variables were sorted on the basis of the number of occurrences in the top 1% of the best combination. (D) Raw ranking obtained for LB-injected animals. Color code highlights measurement methods as in (A). (E) Raw ranking obtained for noLB-injected animals. Color code highlights measurement methods as in (A). according to the number of occurrences (Fig. 3C) for LB-injected dependently for each network), while the 20% remaining were used (Fig. 3D) and noLB-injected (Fig. 3E) animals independently. to evaluate the networks. Then, to evaluate the robustness of the To avoid possible overfitting, we used several methods in combi- quality of prediction for a given set, we repeated this cross-validation nation. First, we performed cross-validation by splitting the dataset step 50 times for every set of three input variables (each network into two parts: a training and a testing set of data. Eighty percent of was trained and tested using a different partition of the dataset; total the data were randomly selected to train the networks (and in- number of networks, 35,428,050). Last, we generated random data Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 5 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE and used them as input for the MLP. As expected, performances were classically associated with parkinsonism (Fig. 4C and fig. S6) (16). significantly lower compared to our actual dataset (fig. S5, B and C). Conversely, noLB-exposed monkeys exhibited more diverse nigrostriatal-­ Overall, this unique approach allowed us to rank input variables centric characteristics with variables related to -syn aggregation, according to their explanatory power and therefore to extract the proteostasis, and Zn homeostasis (Fig. 4D and fig. S7). Together, we strongest predictors of neurodegeneration for each experimental identified specific properties for both groups with limited overlap group. Despite similar levels of nigrostriatal degeneration between (35%; 7 of 20 variables) for an identical level of degeneration. LB- and noLB-injected animals (Fig. 2B), the algorithm allowed us to identify differential variable sorting patterns (Fig. 3, D and E). Retrospective literature search validates MLP-derived signatures MLP-derived signatures can identify unique characteristics We next used a retrospective analysis to validate the relevance of the between experiment group MLP-derived signature in PD. Although some variables have never Next, we compared the LB and noLB characteristics using the rank- been investigated in the context of PD, others have been studied, rank hypergeometric overlap (RRHO) test (Fig. 4A). Low similarity and reports exist in the literature. For instance, the amount of phos- was observed for the highly ranked variables, suggesting specific phorylated Ser129 -syn in the entorhinal (h.psyn.ctx.er.ant) and differences in the biological response to the injection of LB or noLB parahippocampal (h.psyn.ctx.phipp) cortex—first and second best (Fig. 4B). Focusing on the 20 first variables that showed low similarity predictors for the LB group—has been already associated with PD between groups, we found that LB-exposed monkeys were charac- pathology. Studies of postmortem brains from patients with PD terized by both quantitative and qualitative changes in -syn levels revealed the presence of LB in these regions, which was correlated (i.e., phosphorylation at Ser129 and aggregation) especially in cortical with disease progression (17), and predicted cognitive deficit in areas corroborated by distinct methodologies and by a dysfunctional patients with PD (18). The anterior entorhinal cortex has also been equilibrium in neurochemistry of basal ganglia output structures shown to be affected by severe -syn pathology, related to olfactory Fig. 4. Direct comparison of MLP-derived signatures shows specific pattern between experiment groups. (A) RRHO test between variable sorting of LB- and noLB-injected animals. Highly enriched variables are in the lower left corner. Diagonal (highlighted by white dashed line) was extracted to do a bin-to-bin comparison between LB and noLB signatures. (B) Signatures were aligned with RRHO and show low similarity in highly enriched variables (light orange background) and higher similarity for lower rank variables (pale blue background). (C and D) First 20 enriched variables for both LB-injected animals (C) and noLB-injected animals (D). Color code is similar to Fig. 2A. Detailed of variable names can be found in table S1. Bars are means ± 99% confidence interval estimated by bootstrap. Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 6 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE dysfunction in prodromal phases of PD pathology (19). In addition, and marmoset monkeys (27) using the same methodology (fig. S8, increased levels of phosphorylated Ser129 -syn in sensorimotor C and D). Here, overexpression of -syn did not trigger accumula- (h.syn.ctx.sma.ant) and cingulate cortices (h.syn.ctx.cg.ant), shared tion of Zn2+ in the SNpc [despite inducing dopaminergic neuro- by both LB and noLB signatures, have already been reported by our degeneration (27)], suggesting that this phenomenon is specific to group in an independent cohort of nonhuman primates (9). seeding experiment paradigms. This assertion could be confirmed LB and, especially, noLB signatures showed that variables related in other seeding-based models such as striatal preformed fibril in- to -syn aggregation status were among the best predictors (LB, one jection (8, 10) or more refined models using specific types of fibrils in top 10 best predictors; noLB, three in top 10 best predictors): or oligomers (28). aggregated -syn in the putamen measured by filter retardation Last, we analyzed a publicly available cortical proteomic data- assay (db.synAGG.putc.ultra.p2 and db.syn.putc.ultra.s2) and high–­ base of healthy individual and patients with PD. We observed that molecular weight -syn in the SN (wb.synHMW.sn). This was several Zn2+ transporters were elevated in the brains of patients with highly expected from the literature as -syn aggregation has been PD, thus suggesting a zinc dyshomeostasis in patients (fig. S8F). associated with PD pathology (20). Plasma membrane transporters such as the zinc transporter 1 (ZnT1), Variables related to the proteostasis network (levels of the lysosomal the Zrt-/Irt-like protein 6 (ZIP6), and the ZIP10 showed increased receptor LAMP2—wb.lamp2.sn—sixth or amount of ubiquinated levels (fig. S8, G to I), while the synaptic vesicle membrane trans- proteins—wb.ub.sn—ninth) were more specifically associated with porter ZnT3 remained constant (fig. S8J). the noLB signature. This is of high interest as proteostasis defect is more and more considered as a key step in pathogenicity (21, 22). Association metric shows independence of strong predictors Levels of the microglia marker, ionized calcium-binding adapter As we used combinations of three variables and because of the molecule 1 (Iba1), was ranked as the third best predictor of neuro- structure of MLPs, one could expect that some combinations would degeneration in the LB signature. Microglial inflammatory response complement each other to allow finer prediction of neurodegeneration was shown to be implicated in neurodegeneration in many animal levels. To address this question, we used a classic measurement of models, including -syn overexpressing and toxin-­based animal model association in the field of data mining: lift (29) and plotted the results of PD (23). as network plots showing association (edge size) and enrichment in Last, postmortem analysis of Zn2+ concentration in the brains of the best learners (node size). Lift calculation was corrected for error patients with PD has shown elevated levels in the striatum and prediction to avoid detrimental association between variables. The SNpc (24), and a recent meta-analysis showed a decrease in circu- first observation was that the most enriched variables (top 3 to 5) lating Zn2+ levels in patients with PD (25). In experimental models appeared to be self-sufficient to predict the neurodegeneration levels of PD, Zn2+ accumulation has been associated with dopaminergic with minimal error (Fig. 6). Some variables, with modest enrich- degeneration in rodent exposed to mitochondrial toxins (26). ment showed strong positive associations that were specific to each experimental group. Associated variables in LB-injected monkeys Experimental confirmation of MLPs’ prediction were (i) -syn–related parameters along the SNpc-striatum-cortex We aimed to confirm the relevance of the top first MLP selected axis, an impairment of locomotion and the ethologically defined variables. Because the LB signature was associated with changes in orientation of the animals toward their environment (Fig. 6, inset -syn phosphorylation in cortical areas, we analyzed side by side #1); and (ii) oligomeric -syn species measured in the midbrain and the levels of -syn and phosphorylated Ser129 -syn in 18 brain striatum equally associated, but to lesser extent, with -syn levels in regions (Fig. 5A). In agreement with the LB signature obtained from cortex and plasma (Fig. 6, inset #2). the MLP, LB-injected monkeys displayed a stronger accumulation In noLB-injected animals, the analysis shed light upon the relative of phosphorylated Ser129 -syn compared to noLB-injected animals abundance of two members of the macroautophagy pathway (Fig. 6B, (Fig. 5, A and B). In addition, the two most enriched variables of the inset #3) and the balance between monomeric and high–molecular LB signature [i.e., phosphorylated -syn levels in parahippocampal weight species of -syn in the putamen (Fig. 6B, inset #4). This and entorhinal cortices (Fig. 4C)] showed significant negative disruption of the nigrostriatal pathway reverberated upon the basal correlations with degrees of degeneration (Fig. 5, C and D), thus ganglia physiology as -aminobutyric acid (GABA) levels in the confirming their ability to predict neurodegeneration. internal globus pallidus, i.e., the basal ganglia output structure, were Then, we decided to confirm the relevance of one of the strongest associated with a decreased social behavior (Fig. 6B, inset #5). predictors, the levels of Zn2+ in the SNpc in independent experiments. First, we observed a significant increase in Zn2+ in noLB-injected mice compared to sham- or LB-injected mice (fig. S8A). Second, we DISCUSSION analyzed the levels of Zn2+ in LB-injected macaque monkeys from a In the present study, we report that injection of distinct -syn previous study of our laboratory (9). Despite the fact that these assemblies derived from patients with PD in nonhuman primates experiments were performed in a different nonhuman primate sub- leads to dopaminergic degeneration through discrete mechanisms. species, injection of LB in the putamen (similar to the present study) Applying a machine learning method, we gained insight into unique or above the SNpc (different from the present study) induced elevation signatures of degeneration induced by injection of two distinct of Zn2+ levels in the SNpc, as measured by synchrotron radiation -syn pathogenic assemblies (i.e., those contained in the LB and x-ray fluorescence (SR-XRF) (fig. S8B). Notably, the dimension of noLB fractions derived from idiopathic PD patients’ brains). To do so, the effect was similar across studies (fig. S8E). Then, to understand we built a large dataset with 180 variables obtained from behavioral, whether that modulation Zn2+ levels was specific to our experimental histological, biochemical, transcriptional, and biophysical approaches paradigm, we measured Zn2+ levels in the context of adeno-associated applied to several brain areas for each individual. By using a distributed virus–mediated overexpression of mutant human -syn in both rats MLP algorithm that we developed for the purpose of this study, we Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 7 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE Fig. 5. Levels of -syn and phosphorylated -syn in different brain regions. (A) Heat map representing the surface of -syn and S129 phosphorylated -syn immuno­ staining intensity in the brain of noninoculated and LB- and noLB-inoculated baboon monkeys. The heat maps show all brain regions measured and are organized according in three main groups: cortical, basal ganglia, and subcortical areas. From top to bottom: cingulate cortex (ctx.cg), sensorimotor cortex (ctx.sm), retro-insular cortex (ctx.retins), parahippocampal cortex (ctx.phipp), entorhinal cortex (ctx.ent), hippocampus (hipp), caudate nucleus (cd), putamen (put), substantia nigra (sn), ventral tegmental area (vta), red nucleus (rn), subthalamic nucleus (stn), lateral dorsal nucleus (ldn), lateral geniculate nucleus (cgen), claustrum (cltm), fornix (frx), white matter (wm), and corpus callosum (corcal). The color bars represent the log2 value of the ratio of each brain regions. (B) Representative pictures of -syn and phosphorylated -syn (p-Syn S129) staining in the entorhinal and parahippocampal cortices. (C and D) Correlation between levels of phosphorylated -syn in the parahippocampal cortex (C) and the entorhinal cortex (D) with levels of TH staining in the substantia nigra (SN). Dashed lines indicate the linear regression. Gray area indicates the 95% confidence interval around of the linear regression. identified characteristics that give insight into the strongest predictors our laboratory showed that noLB injection did not have any observ- of neurodegeneration for each experimental group. We have there- able impact on dopaminergic degeneration, -syn accumulation, or fore described that distinct -syn assemblies leading to similar phosphorylation (9). In agreement, other groups also showed the degeneration in monkeys are associated with different mechanisms, absence of toxicity of soluble recombinant -syn in mice (10). One hence experimentally confirming the true multifactorial nature of possible explanation is that primate dopaminergic neurons could be synucleinopathies. highly susceptible to -syn toxicity. This could be, in part, due to Our results illustrate that both small oligomeric and larger -syn their unique cellular architecture (30), a feature already known to assemblies induce dopaminergic degeneration in nonhuman primates. contribute to the selective vulnerability of these neurons in PD (31). This finding was unexpected, since previous mouse studies from The large and complex axonal arbor of dopamine neurons makes Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 8 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE Fig. 6. Association metric shows independence of strong predictors and beneficial association of weaker predictors. Both network plots were built using number of counts in the top 1% as node size and color and lift (association measure) as edges. To allow better visualization, only 10% of the strongest edges are shown. (A) Net- work plot for LB-injected animals showing independence of strong predictors: S129 phosphorylated -syn (psyn) in the entorhinal (h.psyn.ctx.er) and the parahippocam- pal cortex (h.psyn.ctx.phipp), microglia activation in the putamen (h.iba1.put), -syn in the cingulate cortex (h.syn.ctx.cg), and the supplementary motor area (h.syn.ctx. sma) and -aminobutyric acid (GABA) levels in the internal part of the globus pallidus (hlpc.gaba.gpi). Inset #1 highlights the association between actimetry measure (actim) and a scan-sampling measure of body direction toward a closed environment (ss.enf) with -syn levels in the caudate nucleus (h.syn.cd), the red nucleus (h.syn.rn), and psyn in the sensorimotor cortex (h.psyn.ctx.sm). Inset #2 highlights the association between pathological -syn in the putamen (wb.syn.put and db.syn.put) and the SNpc (db.syn.sn) as well as psyn in the ventral tegmental area (h.psyn.vta) and peripheral levels of -syn in the plasma (bm.plasma). (B) Network plot for noLB-injected animals showing independence of strong predictors: levels of Zn in the SNpc (s.zn.sn), pathological -syn in the putamen (db.syn.put), -syn in the supplementary motor area (h.syn.ctx.sma), and aggregated -syn in the SNpc (wb.synHMW.sn). Inset #3 highlights the association between autophagosomes (wb.lc3.put) and lysosomes (wb.lamp2.put) levels in the putamen and -syn in the SNpc (wb.syn.sn). Inset #4 highlights the association between GABA levels in the internal part of the globus pallidus (hlpc.gaba.gpi), -syn in the caudate nucleus (wb.syn.cd), and microglia activation in the entorhinal cortex (h.iba1.ctx.er). Inset #5 highlights the association between soluble (wb.syn.putc) and aggregated (wb.synHMW.putc) levels of -syn in the putamen. them particularly vulnerable to factors that contribute to cell death, were mostly enriched in some LB fractions used in our study (fig. S2A), and in primates, this axonal arbor is 10-fold the size of that in a fraction in which we already know that -syn is essential to the rodents (30). In addition, primate dopamine neurons display unique toxicity mechanism (9). It is, however, still possible that the presence molecular characteristics (e.g., the presence of neuromelanin, the of other pathogenic agents could act as facilitators or aggravators of intracellular levels of which have been shown to be important in the the disease, as proposed by Johnson and colleagues (33). Additional threshold for the initiation of PD) (32). These unique features of studies are now needed to fully address the question of host-seed primate dopaminergic neurons might be important in explaining the interactions, but our results highlight the relevance and the need of toxic mechanisms of the relatively low content of -syn aggregates the nonhuman primate model for the study of synucleinopathies. in the noLB fractions. One could hypothesize that other pathogenic We also confirmed that the toxicity mechanisms associated with agents such as tau or A also contribute to the toxicity. Although patient-derived -syn aggregates are shared features among patients follow-up studies would be necessary to confirm this view, we and therefore common to the disease. LB and noLB fractions used observed, in the fractions used in this study, that both tau and A in this study were isolated from a pool of five patients who were Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 9 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE different from the pool of three patients used in our previous study than, humans. Instead of the classification algorithms (the algorithm in mice (9). In mice experiment performed in this study (fig. S3B), learns to identify in which category a sample belongs) that were mostly we observed the same level of dopaminergic degeneration (~40% at used in recent applications of machine learning in biology (43), we 4 months after injection). chose, in this study, to predict continuous and biologically relevant The unexpected observation that, in nonhuman primates, the variables using MLPs. Our choice was justified by the limited sam- noLB fraction is toxic to the same extent as the LB fraction suggests ple size that is often a constraint of experimental biology. Although the existence of previously unrecognized forms of -syn toxicity. it might have been possible to use other feature selection methods, Several studies have suggested that prefibrillar oligomeric species the use of MLPs with a distributed architecture allowed us to avoid are the toxic -syn species (6, 34). Our biochemical studies showed overfitting issues and to develop a method particularly well suited that noLB and LB fractions had different amyloid properties (Fig. 1), for low–sample size datasets (15). As both LB- and noLB-injected contents (figs. S1 and S2A), and structures (fig. S2, B to E). LB monkeys displayed similar levels of degeneration, they were in- fractions contained mostly large aggregated -syn fibrils and some distinguishable using that end point. Instead of using a clustering smaller aggregates, while noLB fractions contained a much lower analysis or a classification method, hence making the a priori as- amount (10-fold) of smaller aggregates and soluble -syn. The sumption that these groups were different, we preferred to submit smaller aggregates in LB and noLB fractions were different in nature, the two experimental groups to the MLP independently. as shown by micro-infrared spectroscopy (fig. S2, B to E). One could The combination of this constrained, distributed architecture hypothesize that the observed effect is due to a species in common and the holistic approach allowed us to rank input variables accord- between LB and noLB. However, because of the extent of degeneration, ing to the number of times they appeared in the group of best pre- which was similar between the two experimental groups, and the dictors (defined as top 1% of best networks). A major issue in the -syn content dissimilarities, both in terms of amount and nature, use of machine learning in experimental biology in the “black box” this appears very unlikely. is the fact that it is usually impossible to “understand” how an algo- We cannot rule out, however, that other proteins contained in rithm predicts an output (44). By using a reverse engineering method, these fractions play a role. The question of factors contributing to we aimed to tackle that issue. Because we explored all possible the pathology when using patient-derived material is of utmost im- combinations of our variables, we could rank the input variables, portance. We performed a filter retardation–based study to charac- assuming that the more they appeared in the top 1%, the more they terize if these proteins are enriched in our fractions. As reported in contained information allowing precise prediction of the neuro- fig. S2A, tau and A are enriched in some LB fractions, which is degeneration levels. Our two experimental groups showed that some expected as LBs are mature aggregates that have been shown to contain of the best predictors were similar (about 30%) but the majority was many proteins. In our previous paper (9), we used two approaches different. One could hypothesize that the variables similar between to ensuring that -syn was essential to the observed toxicity. First, the two signatures probably embedded information related to we injected immunodepleted fractions and observed an absence of disease progression (for example, common paths of -syn spread- dopaminergic degeneration. Second, we injected fractions to -syn– ing once pathology is initiated), while the different ones probably knockout mice and also observed an absence of dopaminergic contain information regarding the process of disease initiation. Further degeneration. Together, these results allowed us to conclude that experimental studies are now needed to confirm the relevance of both the exogenous -syn coming from the injected material and these variables. the endogenous -syn were essential to the toxicity. Building on In addition, as these two kinds of -syn assemblies were associated these previous results, we are confident that the observed toxicity with different signatures identified by our MLP approach, we propose engages -syn both in the fraction and in the host. that our results illustrate the multifactorial nature of the disease as We believe that our results support the notion of the existence of different mechanisms (i.e., signatures) initiated by different triggers a range of -syn pathogenic structures with distinct toxic properties (i.e., -syn assemblies) led to similar consequences (i.e., degeneration within the PD brain, in agreement with recent reports (35, 36). A levels). More important than the enrichment of a given signature recent study, using recombinant seeds, showed that among full- for a single variable, we favor an overall interpretation, taking into length, long, and short fibrils and oligomers, the short fibrils displayed consideration the multifactorial aspect of the disease. the highest toxicity (28). Further work is necessary to provide a Using this methodology, we confirmed the interest of highly complete structural characterization of those species in patient-­ expected variables but, more importantly, also unexpected variables derived fractions, including the contribution of oligomeric species. that appear to be excellent predictors of -syn–associated dopami- As yet, very few studies have reported the high-resolution structures nergic degeneration. The first hit for LB-injected animals was phos- of -syn aggregates, which are, on the one hand, only derived from phorylated -syn in the entorhinal cortex (as we have previously shown), studies using recombinant -syn and, on the other hand, limited to near followed by phosphorylated -syn in the parahippocampal cortex atomic resolution (37–39). Encouragingly, much effort is currently (unexpected), striatal microglial activation, and GABA dysregulation being devoted to this field of research, and three recent studies in the internal part of the globus pallidus (expected) (fig. S6). Con- reported the atomic structure of -syn fibrils determined by cryo– versely, Zn homeostasis was a strong predictive variable (unexpected), electron microscopy (40–42). Unfortunately, these studies were followed by -syn aggregation–related terms (expected) in noLB-­ limited to recombinant-generated -syn and not to -syn isolated injected animals (fig. S7). from human brain tissue. To confirm the prediction made by the MLP approach, we first To perform a characterization of the effects of the two fractions, performed a retrospective literature analysis. This analysis showed we developed a machine learning method to identify their biological that a significant part of the best predictors has been shown in the characteristics. It is now well accepted that machine learning algo- literature to be correlated with disease progression. Then, we attempted rithms can be trained to detect patterns as well as, or even better to confirm the interest of one of the top hits, the accumulation of Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 10 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE Zn2+ in the SNpc, in independent experimental cohorts. We here The Institutional Animal Care and Ethical Committee of the Bordeaux describe an elevation of Zn2+ levels in the SNpc of both mice injected University (CE50, France) approved experiments under the license with noLB and macaque monkeys (a different nonhuman primate number 5012099-A (mice). The Institutional Animal Care and Ethical subspecies than the baboons used in this study) injected with LB in Committee of the Murcia University (Spain) approved experiments striatum or SNpc. However, in mice, Zn2+ dyshomeostasis was not under the license number REGA ES300305440012 (monkeys). associated with neurodegeneration in the noLB group (at odds with what was observed in monkeys), suggesting a species difference in Animals and stereotactic injections the relationship between zinc levels and dopaminergic tract integrity. Mice Unexpectedly, that result was not observed in rats and marmoset Wild-type C57BL/6 mice (4 months old) received 2 l of either LB monkeys overexpressing human mutant -syn. This observation fractions or noLB fractions by stereotactic delivery to the region might suggest that Zn2+ dyshomeostasis is a feature of disease not immediately above the right SN (coordinates from Bregma: ante- triggered in the context of human mutant -syn overexpression rior/posterior, −2.9; lateral, −1.3; dorsal/ventral, −4.5) at a flow rate that is associated with fast progressing pathology (27). Then, to of 0.4 l/min, and the pipette was left in place for 5 min after injec- expand our results to human pathology, we analyzed a publicly tion to avoid leakage. Mice were euthanized 4 months after injec- available proteomic dataset of human samples. According to that tion. Ten to 15 mice were used in each group. analysis, patients with PD displayed increased levels of plasma Monkeys membrane Zn2+ transporters, hence suggesting a Zn dyshomeostasis Animals, which were from the research animal facility of the Uni- in patients. In the context of PD, Zn2+ dyshomeostasis has been versity of Murcia (Murcia, Spain) and housed in two multi-male associated with autophagy/lysosomal dysfunction in the context of multi-female exterior pens, were studied in a breeding farm over PARK9 mutations (45). Further studies are now needed to fully un- 2 years (Murcia, Spain). Animals were fed fruits, vegetables, and ravel this connection. monkey pellets twice a day before 9:00 a.m. and after 5:00 p.m. Water Together, our findings show that primate dopaminergic neurons was available ad libitum. Seventeen healthy adult olive baboons are sensitive to both small, mostly soluble, -syn extracts and larger, (Papio papio) were used in this study. Group sizes were chosen, aggregated, -syn extracts derived from patients with PD. These assuming a one-tailed  of 0.05, with sample size of at least three per findings involve two immediate outcomes. First, soluble -syn toxicity group, which provided >80% power to detect a difference between has not been reported so far, suggesting the existence of species dif- the treatment groups and the control group, using a Fisher’s exact ferences that would need to be thoroughly investigated (46) and test. Animals were randomized into treatment or control groups. calling for a systematic appraisal of proteinopathies in primates, Six baboons were used for LB injections, four were used for noLB particularly for validating therapeutic strategies before clinical testing injections, and seven were untreated control animals. Intrastriatal (47). Second, the present study highlights the complex structure-­ injections of either LB or noLB fractions were performed at two toxicity relationship of -syn assemblies and corroborates the multi- rostrocaudal levels of the motor striatum (anterior commissure, factorial origin of synucleinopathies. Distinct assemblies can induce −1 and −5 mm) under stereotactic guidance as previously described similar degeneration (that would probably lead to similar clinical (48). The total injected volume per hemisphere was 100 l (two in- manifestation in patients) through different mechanisms, nigrostria- jection sites with 50 l each at 3 l/min at each location site). After tal or extranigral brain pathways, calling for molecular diagnosis to each injection, the syringe was left in place for 10 min to prevent identify patient subpopulations before launching large-scale, hetero- leakage along the needle track. A number of parameters were mon- geneous in nature, clinical trials. Last, we developed a machine learn- itored during the course of the 2-year study, including survival and ing approach allowing quantitative assessment of the explanatory clinical observations. At the end of the experiment (24 months after power of a given set of variables compatible with the constrained injection), all monkeys were euthanized with pentobarbital overdose sample size of experimental biology. (150 mg/kg, intravenously), followed by perfusion with room tem- perature 0.9% saline solution (containing 1% heparin) in accordance with accepted European Veterinary Medical Association guidelines. MATERIALS AND METHODS Brains were removed quickly after death. Each brain was then Access to data and machine learning code for replicability dissected along the midline, and each hemisphere was divided into and further use by the community three parts. The left hemisphere was immediately frozen by immersion The entire raw dataset is made available to the readers (table S2). Authors in isopentane at −50°C for at least 5 min and stored at −80°C. The chose not to provide representative examples of each procedure for right hemisphere was fixed for 1 week in 10 volume per tissue of 4% the sake of space and because the entire dataset is fully disclosed. Further paraformaldehyde at 4°C, cryoprotected in two successive gradients information and requests, for examples, should be directed to and of 20 and then 30% sucrose in phosphate-buffered saline (PBS) will be fulfilled by the corresponding contacts. Hyperlink to the machine before being frozen by immersion in isopentane (−50°C) for at least learning code (http://doi.org/10.5281/zenodo.1240558) is provided 5 min, and stored at −80°C until sectioning. Cerebrospinal fluid (https://zenodo.org/record/1240558#.XC8pqy17Su4). (CSF) and blood samples (plasma, serum, and whole blood) in the 17 animals were carefully collected before euthanasia. No samples Ethics statement were excluded from analysis in these studies. Experiments were performed in accordance with the European Union directive of 22 September 2010 (2010/63/EU) on the protection of Purification of LBs from human PD brains animals used for scientific purposes. The Animal Experimentation The samples were obtained from brains collected in a Brain Donation Ethical Committee (CEEA) of the Vall d’Hebron Institute of Research Program of the Brain Bank “GIE NeuroCEB” run by a consortium approved experiments under the license number CEEA 81/13 (rats). of Patients Associations: ARSEP (association for research on multiple Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 11 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE sclerosis), CSC (association for research on cerebellar ataxias), and syn211, Thermo Fisher Scientific; 1:1000) for 30 min, washed three times France Alzheimer and France Parkinson. The consents were signed with PBS, and incubated with a goat anti-mouse tetramethylrhodamine by the patients themselves or their next of kin in their name, in (TRITC) (the Jackson laboratory; 1:500) before being coverslipped accordance with the French Bioethical Laws. The Brain Bank GIE for microscopic visualization using fluorescence mounting medium. NeuroCEB (Bioresource Research Impact Factor number BB-0033- Dot blotting analysis 00011) has been declared at the Ministry of Higher Education and To evaluate proteinase K–resistant -syn contained in noLB and LB Research and has received approval to distribute samples (agreement fractions derived from PD brains, we subjected each fraction to AC-2013-1887). Human SNpc was dissected from freshly frozen post- digestion with proteinase K (1 g/ml) for 0, 15, 30, 45, and 60 min. mortem midbrain samples from five patients with sporadic PD The reaction was stopped by boiling for 5 min before dot blotting exhibiting conspicuous nigral LB pathology on neuropathological with syn211 antibody. To analyze their stability, we treated noLB examination (mean age at death, 75 ± 2.75 years; frozen postmortem and LB fractions with increasing concentrations of urea (7 and 8 M) interval, 31.8 ± 7.45 hours; GIE NeuroCEB BB-0033-00011). Tissue or SDS (0.5, 1, and 2%) for 6 hours at room temperature. -Syn was was homogenized in 9 volumes (w/v) of ice-cold MSE buffer [10 mM visualized as described above. Mops/KOH (pH 7.4), 1 M sucrose, 1 mM EGTA, and 1 mM EDTA] Filter retardation assay of noLB and LB fractions was probed with protease inhibitor cocktail (cOmplete Mini, Boehringer with antibodies against phosphorylated -syn (Abcam, EP1536Y; Mannheim) with 12 strokes of a motor-driven glass/Teflon dounce 1:1000), ubiquitin (Sigma-Aldrich, U5379; 1:1000), p62 (Progen, homogenizer. For LB purification, a sucrose step gradient was pre- GR62-C; 1:1000), hyperphosphorylated tau (AT8, Thermo Fisher pared by overlaying 2.2 M with 1.4 M and lastly with 1.2 M sucrose Scientific, MN1020), or A (clone 6F/3D, DAKO; 1:1000). in volume ratios of 3.5:8:8 (v/v). The homogenate was layered on Human -syn aggregation TR-FRET immunoassay the gradient and centrifuged at 160,000g for 3 hours using a SW TR-FRET–based immunoassays were validated for total and oligomeric 32.1 rotor (Beckman Coulter). Twenty-six fractions of 1500 l were -syn (49). Ten microliters of noLB and LB samples were analyzed collected from each gradient from top (fraction 1) to bottom (frac- for total -syn quantification with the TR-FRET immunoassays kit tion 26) and analyzed for the presence of -syn aggregates by filter against human -syn aggregation kit (Cisbio, number 6FASYPEG) retardation assay, as previously described (9). Further characterization according to the manufacturer’s instructions. of LB fractions was performed by immunofluorescence, -syn ELISA Velocity sedimentation and density floatation -syn profiles quantification, and electron microscopy as previously described in noLB and LB fractions (9). For stereotactic injections, LB-containing fractions from pa- Frozen noLB and LB fraction aliquots (100 l) were thawed and tients with PD were mixed together in the same proportion (PD #1, solubilized in solubilization buffer (SB) to reach 10 mM tris (pH 7.5), fractions 19 and 20; PD #2, fractions 19 and 20; PD #3, fraction 150 mM NaCl, 0.5 mM EDTA, 1 mM dithiothreitol (DTT), cOmplete 22; PD #4, fractions 17 to 19; PD #5, fractions 20, 21, and 23). noLB-­ EDTA-free protease inhibitors (Roche), PhosSTOP phosphatase in- containing fractions (i.e., fraction 3, at the beginning of the 1.2 M hibitors (Roche), benzonase (1 U/l; Novagen), 2 mM MgCl2, and interface) derived from the same patients with PD (which contain 2% (w/v) N-lauroyl-sarcosine (sarkosyl, Sigma-Aldrich) final con- soluble or finely granular -syn) but lack large LB-linked -syn centrations by incubating at 37°C under constant shaking at 600 rpm aggregates were obtained from the same sucrose gradient purifica- (ThermoMixer, Eppendorf) for 45 min. For velocity sedimentations, tion. Using ELISA kit against human -syn (Invitrogen, #KHB0061; a volume of 400 l of solubilized noLB/LB fraction was loaded on following the manufacturer’s recommendations), -syn concentra- top of an 11-ml continuous 5 to 20% iodixanol gradient (OptiPrep, tion was measured, and both LB and noLB fractions were adjusted Sigma-Aldrich) in SB containing 0.5% (w/v) final sarkosyl concen- to ~24 pg of -syn per microliter. In all cases, samples were tration and linearized directly in ultracentrifuge 11-ml tubes (Seton) bath-sonicated for 5 min before in vitro and in vivo injections. with a Gradient Master (Biocomp). For density floatation gradients, a volume of 400 l of solubilized noLB/LB fraction was mixed to Characterization of noLB and LB fractions reach 40% iodixanol in SB with 0.5% (w/v) final sarkosyl concentra- Electron microscopy tion and loaded within an 11-ml 10 to 60% discontinuous iodixanol Briefly, carbon-coated nickel grids were covered for 1 min with cor- gradient in SB with 0.5% (w/v) final sarkosyl concentration. The responding fractions of interest and then washed three times with gradients were centrifuged at 180,000g for 3 hours (velocity) or distilled water. They were then washed again in distilled water and 17 hours (density) in a swinging-bucket SW 40 Ti rotor using an stained for 5 min with 2% uranyl acetate before being air-dried. Optima L-90K ultracentrifuge (Beckman Coulter). Gradients were Digital images were obtained with a computer linked directly to a then segregated into 16 equal fractions from the top using a piston charge-coupled device camera (Gatan) on a Hitachi H-7650 electron fractionator (Biocomp) and a fraction collector (Gilson). Fractions microscope. In all cases, samples were bath-sonicated for 5 min were aliquoted for further analysis of their content by dot blot. before the in vitro applications. Gradient linearity was verified by refractometry. Immunofluorescence analysis of noLB and LB fractions For dot blotting, aliquots of the collected native fractions were Indicated fractions from the sucrose gradient were spread over slides spotted onto Hybond polyvinylidene difluoride 0.2-m membranes coated with poly-d-lysine and fixed with 4% paraformaldehyde in (GE Healthcare) using a dot-blot vacuum device (Whatman). For PBS for 30 min. Fixed slides were stained with 0.05% thioflavin S for total (MJFR1) and phosphorylated pS129 (EP1536Y) -syn immuno- 8 min and then washed three times with 80% EtOH for 5 min, labeling, a step of fixation in PBS–0.1% glutaraldehyde was performed followed by two washes in PBS for 5 min. Last, all samples were at this point, followed by three washes in PBS. Membranes were then washed three times with PBS and blocked with 2% casein and 2% blocked with 5% (w/v) skimmed milk powder in PBS–0.1% (v/v) normal goat serum for 30 min. For immunofluorescence analyses, Tween and probed with anti-human -syn (rabbit, Abcam, MJFR1; samples were incubated with human -syn–specific antibody (clone 1:10,000), anti-phosphorylated pS129 -syn (rabbit, Abcam, EP1536Y; Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 12 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE 1:5000), or anti–-syn aggregate-specific FILA-1 (rabbit, Abcam, uridine and 68 M uridine to control glial outgrowth and in glial MJFR14-6-4-2; 1:10,000) primary antibodies in phosphate buf- cell–derived neurotrophic factor (10 ng/ml). They were incubated fer saline with tween20 PBS-T–4% (w/v) bovine serum albumin for a further 7 to 8 days until the start of experiments. All TH neu- (BSA) and secondary goat anti-rabbit immunoglobulin G horse- rons were counted on each plate following the addition of noLB and LB radish peroxidase (HRP)–conjugated antibodies (the Jackson lab- fractions after 1, 2, 5, and 7 days of treatment. oratory; 1:10,000) in phosphate buffer saline with tween20 PBS-T 1% (w/v) milk. Immunoreactivity was visualized by chemiluminescence Nonhuman primate behavioral assessment (GE Healthcare). The amount of the respective protein in each frac- Following a 4-hour minimum habituation phase performed 1 day tion was determined by the Image Studio Lite software after acquisition before the beginning of the observations, baboon behavior was ob- of chemiluminescent signals with a ChemiDoc imager (Bio-Rad). Pro- served outside the feeding and cleaning times, in a random order at files obtained by immunoblot were normalized and plotted with SEM two time points (morning and afternoon), over 4 to 9 days (eight using the Prism software. sessions per group). On the first observational time point (i.e., 1 month Fourier transform infrared microspectroscopy after surgery), the habituation phase was performed over 3 days, One to 2 l of each suspension was deposited on a CaF2 window and allowing the observer to recognize the animals individually. We dried at room pressure and temperature. The protein aggregates were used a scan sampling method, appropriate for time budgeting (51), then measured in transmission at 50 m by 50 m spatial resolution in which behavioral parameters were assessed every 5 min during with an infrared microscope (50). Depending on its size, it was possible 2-hour sessions, resulting in 192 scans per individual. Extra obser- to collect 1 to 20 spectra inside each aggregate. The infrared micro- vational sessions were performed to avoid missing data. A unique scope was a Thermo Scientific Continuum equipped with a mercury trained observer [SC (Sandrine Camus); intraobserver reliabili- cadmium telluride (MCT) detector and a 32× 0.65–numerical aperture ty, Spearman rank-­order correlation R = 0.987] collected the data Reflachromat objective and matching condenser, coupled to a Thermo live on the two time points of the study: at 1 and 24 months after Scientific Nicolet 8700 spectrometer with a globar source and KBr surgery. The observer was standing 1 m away from the outdoor cag- beamsplitter. The microscope was operated in dual-path single-aper- es. We focused on behavioral profiles rather than single items and ture mode. Spectra were recorded between 650 and 4000 cm−1 at used two repertoires: One reports the interaction with the environ- 2 cm−1 resolution, with Happ-Genzel apodization and Mertz phase ment, and one describes the position within the environment, accord- correction. Spectra were processed in Omnic 9.2 for automatic at- ing to published protocols (52–54). We investigated the percentages mospheric correction to remove water vapor contribution. of occurrence of each item with regard to the total number of scans to obtain mean behavioral and postural time budgets, body orienta- Rat ventral midbrain primary cultures tion, and location profiles. Postnatally derived ventral midbrain cultures were prepared two steps. In the first step, rat astrocyte monolayers were generated as Histopathological analysis follows. The entire cerebral cortex from a single rat pup (postnatal Extent of lesion days 1 and 2) was removed, diced, and then mechanically dissociated To assess the integrity of the nigrostriatal pathway, TH immunohisto- by gentle trituration. The isolated cells were plated at 80,000 cells chemistry was performed on SNpc and striatal sections. Briefly, 50-m per well under which a laminin-coated coverslip was affixed. The free-floating sections from one representative level of the striatum cells were housed at 37°C in an incubator in 5% CO2 and were fed (anterior, medial, and posterior) and serial sections (1 every 12) corre- on glial media [89% minimum essential medium (MEM), 9.9% fetal sponding to the whole SNpc were incubated with a mouse monoclonal calf serum, 0.33% glucose, 0.5 mM glutamine, and insulin (5 g/ml)]. antibody raised against human TH (Millipore, MAB318; 1:5000) for one Once confluence had been attained (about 1 week in vitro), fluoro- night at room temperature and revealed by an anti-mouse peroxidase deoxyuridine (6.7 mg/ml) and uridine (16.5 mg/ml) were added to EnVision System (DAKO, K400311), followed by 3,3′-diaminobenzidine prevent additional proliferation. In the second stage, which occurred (DAB) visualization. Free-floating SNpc sections were mounted on 1 week later, rat pups aged between 1 and 2 days were anesthetized, gelatinized slides, counterstained with 0.1% cresyl violet solution, and 1-mm3 blocks containing ventral midbrain neurons were dis- dehydrated, and coverslipped, while striatal sections were mounted on sected from 1-mm-thick sagittal sections taken along the midline of gelatinized slides and coverslipped. The extent of the lesion in the the brain. Tissues were collected immediately into cold phosphate striatum was quantified by optical density. Sections were scanned in buffer and were treated enzymatically using papain (20 U/ml) with an Epson Expression 10000XL high-resolution scanner, and images kynurenate (500 M) at 37°C under continuous oxygenation with were used in ImageJ open-source software to compare the gray level gentle agitation for 2 hours. A dissociated cell suspension was in each region of interest, i.e., caudate nucleus and putamen. TH-positive achieved by gentle trituration and was then plated onto the pre- SNpc cells were counted by stereology blind with regard to the ex- established glia wells at a density of 0.5 to 1.7 million neurons per perimental condition using a Leica DM6000 B motorized microscope well. Cultures were maintained in specially designed neuronal coupled with the Mercator software (ExploraNova, France). The media [47% DMEM (Dulbecco’s modified Eagle’s medium), 40% SN was delineated for each slide, and probes for stereological Dulbecco’s MEM, 10% Ham’s F-12 nutrient medium, 1% calf serum, counting were applied to the map obtained (size of probes was 0.25% albumin, glucose (2 mg/ml), 0.4 mM glutamine, catalase (10 100 × 80 m spaced by 600 × 400 m). Each TH-positive cell with its g/ml), 50 M kynurenic acid, 10 M CNQX (6-cyano-7-nitroqui- nucleus included in the probe was counted. The optical fractionator noxaline-2,3-dione), insulin (25 g/ml), transferrin (100 g/ml), method was lastly used to estimate the total number of TH-positive superoxide dismutase (5 g/ml), putrescine (2.4 g/ml), Na2SeO3 cells in the SNpc of each monkey hemisphere. In addition, we measured (5.2 ng/ml), triiodothyronine (0.02 g/ml), progesterone (62.5 Nissl cell count, the volume of SN, and the surface of TH-occupied in ng/ml), and cortisol (40 ng/ml)] containing 27 M fluorodeoxy- SN to fully characterize the pattern of dopaminergic cell loss in the SN. Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 13 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE -Syn pathology GAGGGCAGGTGTACT (forward) and CCTCCAATGGCAAAG- Synucleinopathy was assessed with a mouse monoclonal antibody GACTG (reverse); EIF4a2 (NM_001967), TGACATGGACCAGAAG- raised against human -syn (syn211) and phosphorylated -syn GAGAGA (forward) and TGATCAGAACACGACTTGACCCT (clone 11A5, Elan; 1:5000) immunostaining, as we previously reported (reverse); SNCA (CR457058), GGGCAAGAATGAA GAAGGAGC (9, 27). Briefly, selected sections at two rostrocaudal levels were in- (forward) and GCCTCATTGTCAGGATCCACA (reverse). cubated in a same well to allow direct comparison of immunostaining intensity. Sections were incubated overnight at room temperature Biochemical analysis with the aforementioned antibodies. The following day, revelation Total protein extraction and quantification was performed with antispecies peroxidase EnVision System (DAKO), Immunoblot analyses were performed on SN, putamen, and caudate followed by DAB incubation. Sections were then mounted on gelatin­ nucleus. Five tissue patches were extracted on ice using 100 l of ized slides, dehydrated, counterstained if necessary, and coverslipped radioimmunoprecipitation assay (RIPA) buffer [50 mM tris-HCl until further analysis. Gray-level quantification or immunostaining-­ (pH 7.4), 150 mM NaCl, 1.0% Triton X-100, 0.5% Na-­deoxycholate, positive surface quantification in 40 brain regions (Fig. 2B) was per- and 0.1% SDS] with a protease inhibitor cocktail tablet (cOmplete formed, as previously described (27). Mini, Roche Diagnostics). The lysate was placed on ice for 20 min Inflammation and then centrifuged at 14,000 rpm for 15 min at 4°C. The super- Inflammatory process in the striatum, the entorhinal cortex, and natant was collected, and the bicinchoninic acid (BCA) assay was the white matter of noLB- and LB-injected monkeys was measured used to determine the total amount of protein in the lysates and through glial fibrillary acidic protein (GFAP)/S-100 (DAKO, Z0334/ then stored at −80°C. Abnova, PAP11341) and Iba1 (Abcam, ab5076) immunohisto- On the basis of total protein concentrations calculated from the chemistry. Striatal sections of all animals were incubated together BCA assays, aliquots of tissue lysates corresponding to known amounts overnight with a mix of rabbit antibodies raised against GFAP and of total protein per lane were prepared for each animal in Laemmli S-100 for the astroglial staining (respective dilutions, 1:2000 and buffer [25 mM tris-HCl (pH 6.8), 7.5% glycerol, 1% SDS, 250 mM 1:1000) and with a goat anti-Iba1 antibody for the microglial stain- DTT, and 0.05% bromophenol blue) for immunoblotting experiment. ing (dilution, 1:1000). These signals were revealed with antispecies Biochemical fractionation peroxidase EnVision system (DAKO), followed by DAB incubation. Tissue patches (n = 10) were homogenized in 200 l of high-salt (HS) Sections were mounted on slides, counterstained in 0.1% cresyl buffer (50 mM tris, 750 mM NaCl, 5 mM EDTA, and a cocktail of violet solution, dehydrated, and coverslipped. Sections stained by protease inhibitors and phosphatase inhibitors). Samples were sed- GFAP/S-100 were numerized at ×20 magnification with a NanoZoomer imented at 100,000g for 20 min, and supernatants were removed for (Hamamatsu, France), and the quantification of GFAP-positive astro- analysis. Pellets were rehomogenized in successive buffers, after cytic reaction was estimated by immunostaining-positive surface quan- which each was sedimented, and supernatant was removed: HS con- tification at regional levels with the Mercator software (ExploraNova, taining 1% Triton X-100 (HS/Triton) (variable names terminated as France). Sections stained by Iba1 were used for the microglial mor- ultra.s1), RIPA (50 mM tris, 150 mM NaCl, 5 mM EDTA, 1% NP-40, phology analysis through fractal dimension quantification based on 0.5% Na-deoxycholate, and 0.1% SDS) (variable names terminated microscopic acquisitions, as previously described (55). All analyses as ultra.s12), and SDS/urea [8 M urea, 2% SDS, and 10 mM tris were performed blinded to the researcher. (pH 7.5)] (variable names terminated as ultra.p2). SDS sample buffer was added, and samples were heated to 100°C for 5 min before mRNA extraction and quantitative reverse transcription immunoblot analysis. polymerase chain reaction Western blot analysis SN samples were homogenized in TRI reagent (Euromedex, Western blots were run under all conditions from 20 g of protein France), and RNA was isolated using a standard chloroform/ separated by SDS–polyacrylamide gel electrophoresis and transferred isopropanol protocol. RNA was processed and analyzed following to nitrocellulose. Incubation of the primary antibodies was per- an adaptation of published methods (56). Complementary DNA formed overnight at 4°C with rabbit anti-LC3 (Novus Biologicals; (cDNA) was synthesized from 2 g of total RNA using RevertAid 1:1000), rabbit anti–LAMP-2 (Santa Cruz Biotechnology; 1:1000), Premium Reverse Transcriptase (Fermentas) and primed with mouse anti-TH (Millipore; 1:1000), goat p62 (Progen; 1:1000), and oligo-dT primers (Fermentas) and random primers (Fermentas). mouse anti-human -syn (Thermo Fisher Scientific; 1:1000). For Quantitative polymerase chain reaction (qPCR) was performed using detection of ubiquitinated proteins, proteins were transferred to a LightCycler 480 Real-Time PCR System (Roche, Meylan, France). polyvinylidene fluoride membranes (Millipore) and subjected to qPCRs were performed in duplicate for each sample using transcript-­ Western blot analysis using a rabbit anti-ubiquitin (Sigma-Aldrich, specific primers, cDNA (4 ng), and LightCycler 480 SYBR Green I U5379; 1:1000). Anti-actin (Sigma-Aldrich; 1:5000) was used to con- Master (Roche) in a final volume of 10 l. The PCR data were ex- trol equal loading. Appropriate secondary antibodies coupled to ported and analyzed in an informatics tool (Gene Expression Analysis peroxidase were revealed using a SupeSignal West Pico Chemilumi- Software Environment) developed at the NeuroCentre Magendie. nescent kit (Immobilon Western, Chemiluminescent HRP Substrate, For the determination of the reference gene, the geNorm method was Millipore). Chemiluminescence images were acquired using the used. Relative expression analysis was corrected for PCR efficiency ChemiDoc XRS+ system measurement (Bio-Rad). Signals per and normalized against two reference genes. The proteasome sub- lane were quantified using ImageJ, and a ratio of signal on load- unit beta type 6 (Psmb6) and eukaryotic translation initiation factor ing per animal was performed and used in statistical analyses. 4a2 (EIF4A2) genes were used as reference genes. The relative level Dot blotting analysis of -syn of expression was calculated using the comparative (2−∆∆Ct) method. This technique was performed as we previously described (9, 34). Primer sequences are as follows: Psmb6 (NM_002798), CAAGAAG- After heating at 100°C for 5 min, 20 g of protein extract was diluted Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 14 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE in buffer (25 mM tris-HCl, 200 mM glycine, and 1% SDS) and 6.1 pg/ml in serial fourfold dilutions. One percent of BSA/PBS-T filtered through either a nitrocellulose membrane or an acetate served as blank. For the different specimen, the following dilutions cellulose membrane (Bio-Rad; 0.2-m pore size). Membranes were were applied: 1 in 10,000 for whole blood and 1 in 8 for serum, plasma, then saturated in 5% dry-skimmed milk in PBS and probed with and CSF. All dilutions were prepared in 1% BSA/PBS-T. After antibodies against -syn (syn211; 1:1000), both -syn fibrils and washing the plates, 25-l of the calibrator solution and diluted sam- -syn oligomers [Syn-O1; 1:10,000; (57)] (provided by O.M.A.E.-A.). ples were applied to the wells and incubated as indicated above. Revelation was performed as described in the Western blot analysis Plates were washed again, and 25 l of Sulfo-TAG–labeled Syn1 section. antibody (BD Biosciences, Heidelberg, Germany) diluted to 1 g/ml in 1% PBS-T was applied to the wells as detection antibody. Sulfo- SR-XRF microscopy elemental mapping of brain tissue TAG labeling was performed according to the manufacturer’s in- cryosections struction using MSD Sulfo-TAG NHS-Ester (MSD). Incubation was The synchrotron experiments were carried out at the Diamond Light done for 1 hour at 700 rpm. Plates were washed, 150 l of 2× Read Buffer Source, Harwell Science and Innovation Campus (Didcot, UK) with (MSD) was applied, and the plates were read on a MSD Sector Imager a 3-GeV energy of the storage ring and 300-mA currents with top-up 2400. Data analysis was performed using Workbench software (MSD). injection mode. All SR-XRF microscopy investigations reported herein were carried out on the microfocus spectroscopy beamline Neurotransmitter analysis (I18). The micro-XRF elemental mapping was acquired at room Brain patches were dissected out on ice-cold plate, weighed, and temperature with an incident x-ray energy set to 12 keV using an put into 1.5-ml Eppendorf tubes. Samples were homogenized in Si(111) monochromator and resulting in a x-ray photon flux of 2 × methanol/water (50:50%, v/v) and then centrifuged at 14,000 rpm 1011 photons/s. The SN of each animal was collected from free-floating for 15 min at 4°C. The supernatant was aliquoted and stored at sections and mounted onto an x-ray transparent metal-­free 4-m-thick −80°C until amino acid derivatization. Glutamate and GABA content Ultralene foil (SPEX CertiPrep, Metuchen, NJ, USA) secured to a in the samples was measured by high-performance liquid chroma- customized polyetheretherketone holder ensuring contamination-free tography (HPLC) coupled with fluorometric detection (FP-2020 samples and reduced x-ray scattering contribution. The samples Plus fluorimeter, Jasco, Tokyo, Japan) after precolumn derivatization were affixed to a magnetic plate that connects to the sample stage. The with o-phthaldialdehyde/mercaptoethanol (OPA) reagent. Thirty four-element Si drift Vortex ME4 energy-dispersive detector (Hitachi microliters of OPA reagent was automatically added to 28 l of High-Technologies Science America), with Xspress 3 processing sample by a refrigerated autosampler kept at 4°C (Triathlon, Spark electronics, was operated in the 90° geometry; hence, it minimizes Holland, Emmen, The Netherlands). Fifty microliters of the mixture the background signal. The sample-detector distance was fixed was injected onto a 5-C18 Hypersil ODS column (3 × 100 mm; (75 mm). The sample was held at 45° to the incident x-ray beam Thermo Fisher Scientific, USA) perfused at 0.48 ml/min (Jasco and rastered in front of the beam, while the XRF spectra were PU-2089 Plus Quaternary Pump; Jasco, Tokyo, Japan) with a mobile collected. An area of 500 m by 500 m within the SNpc was mapped phase containing 0.1 M sodium acetate, 10% methanol, and 2.2% for each sample with a step size that matches the beam size (5 m) tetrahydrofuran (pH 6.5). Chromatograms were acquired and ana- and a dwell time of 1 s per pixel due to low concentration of the lyzed using a ChromNAV software (Jasco, Tokyo, Japan). Under element. A thin (100 m) pellet of the National Institute of Standards these conditions, the limits of detection for glutamate and GABA and Technology (NIST) standard reference materials SRM1577c were ~1 and ~0.5 nM, and their retention times were ~3.5 and (bovine liver material, NIST, Gaithersburg, MD, USA) was measured ~18.0 min, respectively. to calibrate experimental parameters and a thin-film XRF reference material (AXO DRESDEN GmbH). This was followed by elemental Multiple-layer perceptrons quantification through the open-source software PyMCA in which Each MLP had the same architecture rule: three neurons as input, both the reference material and the sample are modeled in terms of three neurons in the hidden layer, and three neurons as output. Ac- main composition, density, and thickness. The fluorescence spec- tivation function of neurons was the hyperbolic tangent. Each net- trum obtained from each pixel was fitted, the elemental concentra- work was trained over 1000 presentations of a subset of the dataset. tion (micrograms per gram of dry weight or parts per million) maps We used as error measure the mean square of differences between were generated, and an average elemental concentration of the SNpc the expected output and the actual output. Our implementation regions was obtained. comprises two parameters: a learning rate set at 0.05 (regulating the learning speed) and a momentum set at 0.05 (introducing purpose- Measurement of -syn in monkey biological fluid samples fully a conservatism bias). Before learning, inputs were first scaled Multi-array 96-well plates [Meso Scale Discovery (MSD), Gaithersburg, and centered (z scoring) to avoid dimensionality issues and then MD, USA] were coated with 30 l of MJFR1 (3 l/ml; Abcam, normalized between −0.5 and 0.5. For every combination of three Cambridge, UK) as capture antibody and incubated overnight at variables used as inputs, 50 instances of MLP were trained with dif- 4°C without shaking. The next day, plates were washed three times ferent subsets of the dataset. Eighty percent of available data have with 150 l of PBS-T [PBS (AppliChem, Darmstadt, Germany) sup- been used for learning, and the remaining 20% have been used for plemented with 0.05% Tween 20 (Roth, Karlsruhe, Germany)] per testing the performance of the network [elements of each subset well. Unspecific binding of proteins was prevented by incubation were randomly (and uniformly) drawn for each network]. The per- with 150 l of 1% BSA (SeraCare Life Sciences, Milford, MA, USA)/ formance from a given set of input variables was the mean of the PBS-T per well for 1 hour and shaking at 700 rpm. Calibrators (pro- error of the 50 instances of MLP that had data for these variables vided by O. El-Agnaf) were prepared from single-use aliquots of as inputs. Code was written using Python (v. 3.7.3) and the Python -syn (1 g/ml stored at −80°C until use) and ranged from 25,000 to scientific stack: numpy (v. 1.17.2), scipy (v. 1.3.1), and Cython Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 15 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE (v. 0.29.13) (58). The code is fully available here (DOI: 10.5281/ 5. B. Dehay, M. Vila, E. Bezard, P. Brundin, J. H. Kordower, Alpha-synuclein propagation: New insights from animal models. Mov. Disord. 31, 161–168 (2016). zenodo.1240558). Computation has been done using the Avakas 6. B. Winner, R. Jappelli, S. K. Maji, P. A. Desplats, L. Boyer, S. Aigner, C. Hetzer, T. Loher, cluster of the Mesocentre de Calcul Intensif Aquitain (MCIA). M. Vilar, S. Campioni, C. Tzitzilonis, A. Soragni, S. Jessberger, H. Mira, A. Consiglio, RRHO test was performed as previously described (59) using RRHO E. Pham, E. Masliah, F. H. Gage, R. Riek, In vivo demonstration that -synuclein oligomers package (1.14.0) in R (60) on variable list after ranking between ex- are toxic. Proc. Natl. Acad. Sci. U.S.A. 108, 4194–4199 (2011). perimental groups. Plotting was made using matplotlib (v. 3.1.1) in 7. N. Bengoa-Vergniory, R. F. Roberts, R. Wade-Martins, J. Alegre-Abarrategui, Alpha-synuclein oligomers: A new hope. Acta Neuropathol. 134, 819–838 (2017). Python (v. 3.7.3) environment. 8. Y. Chu, S. Muller, A. Tavares, O. Barret, D. Alagille, J. Seibyl, G. Tamagnan, K. Marek, The association metric was based on lift calculation. Let a and b K. C. Luk, J. Q. Trojanowski, V. M. Y. Lee, J. H. Kordower, Intrastriatal alpha-synuclein fibrils be the two variables and na the number of combinations including in monkeys: Spreading, imaging and neuropathological changes. Brain 142, 3565–3579 variable a and n the total number of combinations considered in the (2019). analysis. Lift calculation was then 9. A. Recasens, B. Dehay, J. Bové, I. Carballo-Carbajal, S. Dovero, A. Pérez-Villalba, P.-O. Fernagut, J. Blesa, A. Parent, C. Perier, I. Fariñas, J. A. Obeso, E. Bezard, M. Vila, ​n​  ab​​ ​n​  b​​ Lewy body extracts from Parkinson disease brains trigger -synuclein pathology ​​Lift​  ab​​  = ​─ n ​​ ​n​ a​​ ​ / ​─ and neurodegeneration in mice and monkeys. Ann. Neurol. 75, 351–362 (2014). 10. K. C. Luk, V. Kehm, J. Carroll, B. Zhang, P. O’Brien, J. Q. Trojanowski, V. M. Y. Lee, The lift calculation was then corrected for performance to avoid Pathological -synuclein transmission initiates Parkinson-like neurodegeneration selection of detrimental association by being divided by the mean in nontransgenic mice. Science 338, 949–953 (2012). 11. W. Peelaerts, L. Bousset, A. Van der Perren, A. Moskalyuk, R. Pulizzi, M. Giugliano, prediction error of the duo. C. Van den Haute, R. Melki, V. Baekelandt, -Synuclein strains cause distinct synucleinopathies after local and systemic administration. Nature 522, 340–344 Quantification and statistical analysis (2015). Regarding the data analysis for Fourier transform infrared microspec- 12. E. Bezard, S. Dovero, C. Prunier, P. Ravenscroft, S. Chalon, D. Guilloteau, A. R. Crossman, troscopy, spectra were analyzed by PCA. PCA is a multivariate B. Bioulac, J. M. Brotchie, C. E. Gross, Relationship between the appearance of symptoms and the level of nigrostriatal degeneration in a progressive 1-methyl-4-phenyl-1,2,3,6- statistical analysis technique that captures independent sources of tetrahydropyridine-lesioned macaque model of Parkinson's disease. J. Neurosci. 21, variance in the data and represents them in principal components 6853–6861 (2001). (eigenvectors) that carry the underlying spectral information and in 13. M. Bourdenx, A. Nioche. (2018). a score plot that shows the relation between spectra and can be used 14. K. Hornik, M. Stinchcombe, H. White, Multilayer feedforward networks are universal to cluster the data based on the spectral information. PCA was per- approximators. Neural Netw. 2, 359–366 (1989). formed in The Unscrambler X 10.3 (CAMO Software) using the 15. L. I. Kuncheva, J. J. Rodríguez, On feature selection protocols for very low-sample-size data. Pattern Recognit. 81, 660–673 (2018). singular value decomposition algorithm with leverage correction. 16. R. L. Albin, A. B. Young, J. B. Penney, The functional anatomy of basal ganglia disorders. Two series of preprocessing were applied before PCA and compared. Trends Neurosci. 12, 366–375 (1989). Spectra were either baseline-corrected in the amide I region between 17. M. Neumann, V. Müller, H. A. Kretzschmar, C. Haass, P. J. Kahle, Regional distribution 1590 and 1700 cm−1 and vector-normalized or their second deriva- of proteinase K-resistant -synuclein correlates with Lewy body disease stage. tives were computed and vector-normalized. J. Neuropathol. Exp. Neurol. 63, 1225–1235 (2004). Statistical analyses were performed with GraphPad Prism 6.0 18. E. Kövari, G. Gold, F. R. Herrmann, A. Canuto, P. R. Hof, C. Bouras, P. Giannakopoulos, Lewy body densities in the entorhinal and anterior cingulate cortex predict cognitive deficits (GraphPad Software Inc., San Diego, CA). For all experiments, in Parkinson’s disease. Acta Neuropathol. 106, 83–88 (2003). comparisons among means were performed using one-way analysis 19. L. Silveira-Moriyama, J. L. Holton, A. Kingsbury, H. Ayling, A. Petrie, W. Sterlacci, W. Poewe, of variance (ANOVA), followed, if appropriate, by a pairwise H. Maier, A. J. Lees, T. Revesz, Regional differences in the severity of Lewy body pathology comparison between means by Tukey post hoc analysis. All values across the olfactory cortex. Neurosci. Lett. 453, 77–80 (2009). are expressed as the means ± SEM. Size effect was assessed with 20. Y. C. Wong, D. Krainc, -synuclein toxicity in neurodegeneration: Mechanism and therapeutic strategies. Nat Med 23, 1–13 (2017). Cohen’s d analysis. In all analyses, statistical significance was set 21. B. Dehay, J. Bové, N. Rodríguez-Muela, C. Perier, A. Recasens, P. Boya, M. Vila, at P < 0.05. Pathogenic lysosomal depletion in Parkinson’s disease. J. Neurosci. 30, 12535–12544 (2010). SUPPLEMENTARY MATERIALS 22. S. Kaushik, A. M. Cuervo, Proteostasis and aging. Nat. Med. 21, 1406–1415 (2015). Supplementary material for this article is available at http://advances.sciencemag.org/cgi/ 23. E. C. Hirsch, S. Vyas, S. Hunot, Neuroinflammation in Parkinson’s disease. Parkinsonism content/full/6/20/eaaz9165/DC1 Relat. Disord. 18, S210–S212 (2012). 24. D. T. Dexter, A. Carayon, F. Javoy-Agid, Y. Agid, F. R. Wells, S. E. Daniel, A. J. Lees, P. Jenner, View/request a protocol for this paper from Bio-protocol. C. D. Marsden, Alterations in the levels of iron, ferritin and other trace metals in Parkinson’s disease and other neurodegenerative diseases affecting the basal ganglia. REFERENCES AND NOTES Brain 114, 1953–1975 (1991). 1. H. Braak, K. Del Tredici, U. Rüb, R. A. I. de Vos, E. N. H. Jansen Steur, E. Braak, Staging 25. K. Du, M.-Y. Liu, X. Zhong, M.-J. Wei, Decreased circulating zinc levels in Parkinson’s of brain pathology related to sporadic Parkinson’s disease. Neurobiol. Aging 24, 197–211 disease: A meta-analysis study. Sci. Rep. 7, 3902 (2017). (2003). 26. J.-Y. Lee, H. J. Son, J. H. Choi, E. Cho, J. Kim, S. J. Chung, O. Hwang, J.-Y. Koh, Cytosolic 2. J. H. Kordower, Y. Chu, R. A. Hauser, T. B. Freeman, C. W. Olanow, Lewy body-like labile zinc accumulation in degenerating dopaminergic neurons of mouse brain after pathology in long-term embryonic nigral transplants in Parkinson’s disease. Nat. Med. 14, MPTP treatment. Brain Res. 1286, 208–214 (2009). 504–506 (2008). 27. M. Bourdenx, S. Dovero, M. Engeln, S. Bido, M. F. Bastide, N. Dutheil, I. Vollenweider, 3. J. Y. Li, E. Englund, J. L. Holton, D. Soulet, P. Hagell, A. J. Lees, T. Lashley, N. P. Quinn, L. Baud, C. Piron, V. Grouthier, T. Boraud, G. Porras, Q. Li, V. Baekelandt, D. Scheller, S. Rehncrona, A. Björklund, H. Widner, T. Revesz, O. Lindvall, P. Brundin, Lewy bodies A. Michel, P.-O. Fernagut, F. Georges, G. Courtine, E. Bezard, B. Dehay, Lack of additive in grafted neurons in subjects with Parkinson’s disease suggest host-to-graft disease role of ageing in nigrostriatal neurodegeneration triggered by -synuclein propagation. Nat. Med. 14, 501–503 (2008). overexpression. Acta Neuropathol. Commun. 3, 46 (2015). 4. I. Mendez, A. Viñuela, A. Astradsson, K. Mukhida, P. Hallett, H. Robertson, T. Tierney, 28. J. M. Froula, M. Castellana-Cruz, N. M. Anabtawi, J. D. Camino, S. W. Chen, D. R. Thrasher, R. Holness, A. Dagher, J. Q. Trojanowski, O. Isacson, Dopamine neurons implanted into J. Freire, A. A. Yazdi, S. Fleming, C. M. Dobson, J. R. Kumita, N. Cremades, L. A. Volpicelli-Daley, people with Parkinson’s disease survive without pathology for 14 years. Nat. Med. 14, Defining -synuclein species responsible for Parkinson's disease phenotypes in mice. 507–509 (2008). J. Biol. Chem. 294, 10392–10406 (2019). Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 16 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE 29. G. Piatetsky-Shapiro, in Knowledge Discovery in Databases, G. Piatetsky-Shapiro, W. J. Frawley, 52. S. M. J. Camus, C. Blois-Heulin, Q. Li, M. Hausberger, E. Bezard, Behavioural profiles Eds. (AAAI/MIT Press, 1991). in captive-bred cynomolgus macaques: Towards monkey models of mental disorders? 30. J. P. Bolam, E. K. Pissadaki, Living on the edge with too many mouths to feed: Why PLOS ONE 8, e62141 (2013). dopamine neurons die. Mov. Disord. 27, 1478–1483 (2012). 53. S. M. Camus, C. Rochais, C. Blois-Heulin, Q. Li, M. Hausberger, E. Bezard, Birth origin 31. D. J. Surmeier, J. A. Obeso, G. M. Halliday, Selective neuronal vulnerability in Parkinson differentially affects depressive-like behaviours: Are captive-born cynomolgus monkeys disease. Nat. Rev. Neurosci. 18, 101–113 (2017). more vulnerable to depression than their wild-born counterparts? PLOS ONE 8, e67711 32. I. Carballo-Carbajal, A. Laguna, J. Romero-Giménez, T. Cuadros, J. Bové, M. Martinez-Vicente, (2013). A. Parent, M. Gonzalez-Sepulveda, N. Peñuelas, A. Torra, B. Rodríguez-Galván, A. Ballabio, 54. S. M. J. Camus, C. Rochais, C. Blois-Heulin, Q. Li, M. Hausberger, E. Bezard, T. Hasegawa, A. Bortolozzi, E. Gelpi, M. Vila, Brain tyrosinase overexpression implicates Depressive-like behavioral profiles in captive-bred single- and socially-housed rhesus age-dependent neuromelanin production in Parkinson’s disease pathogenesis. and cynomolgus macaques: A species comparison. Front. Behav. Neurosci. 8, 47 (2014). Nat. Commun. 10, 973 (2019). 55. F. N. Soria, M. Engeln, M. Martinez-Vicente, C. Glangetas, M. J. López-González, S. Dovero, 33. M. E. Johnson, B. Stecher, V. Labrie, L. Brundin, P. Brundin, Triggers, facilitators, and B. Dehay, E. Normand, M. Vila, A. Favereaux, F. Georges, C. L. Bianco, E. Bezard, aggravators: Redefining Parkinson’s disease pathogenesis. Trends Neurosci. 42, 4–13 P.-O. Fernagut, Glucocerebrosidase deficiency in dopaminergic neurons induces (2019). microglial activation without neurodegeneration. Hum. Mol. Genet. 26, 2603–2615 34. M. Bourdenx, N. S. Koulakiotis, D. Sanoudou, E. Bezard, B. Dehay, A. Tsarbopoulos, Protein (2017). aggregation and neurodegeneration in prototypical neurodegenerative diseases: 56. S. A. Bustin, V. Benes, J. A. Garson, J. Hellemans, J. Huggett, M. Kubista, R. Mueller, Examples of amyloidopathies, tauopathies and synucleinopathies. Prog. Neurobiol. 155, T. Nolan, M. W. Pfaffl, G. L. Shipley, J. Vandesompele, C. T. Wittwer, The MIQE guidelines: 171–193 (2017). Minimum information for publication of quantitative real-time PCR experiments. Clin. Chem. 35. A. Lau, R. W. L. So, H. H. C. Lau, J. C. Sang, A. Ruiz-Riquelme, S. C. Fleck, E. Stuart, 55, 611–622 (2009). S. Menon, N. P. Visanji, G. Meisl, R. Faidi, M. M. Marano, C. Schmitt-Ulms, Z. Wang, 57. N. N. Vaikath, N. K. Majbour, K. E. Paleologou, M. T. Ardah, E. van Dam, W. D. J. van de Berg, P. E. Fraser, A. Tandon, B. T. Hyman, H. Wille, M. Ingelsson, D. Klenerman, J. C. Watts, S. L. Forrest, L. Parkkinen, W.-P. Gai, N. Hattori, M. Takanashi, S.-J. Lee, D. M. A. Mann, -Synuclein strains target distinct brain regions and cell types. Nat. Neurosci. 23, Y. Imai, G. M. Halliday, J.-Y. Li, O. M. A. el-Agnaf, Generation and characterization of novel 21–31 (2019). conformation-specific monoclonal antibodies for -synuclein pathology. Neurobiol. Dis. 36. T. Strohäker, B. C. Jung, S.-H. Liou, C. O. Fernandez, D. Riedel, S. Becker, G. M. Halliday, 79, 81–99 (2015). M. Bennati, W. S. Kim, S.-J. Lee, M. Zweckstetter, Structural heterogeneity of -synuclein 58. P. Virtanen, R. Gommers, T. E. Oliphant, M. Haberland, T. Reddy, D. Cournapeau, fibrils amplified from patient brain extracts. Nat. Commun. 10, 5535 (2019). E. Burovski, P. Peterson, W. Weckesser, J. Bright, S. J. van der Walt, M. Brett, J. Wilson, 37. J. A. Rodriguez, M. I. Ivanova, M. R. Sawaya, D. Cascio, F. E. Reyes, D. Shi, S. Sangwan, K. Jarrod Millman, N. Mayorov, A. R. J. Nelson, E. Jones, R. Kern, E. Larson, C. J. Carey, E. L. Guenther, L. M. Johnson, M. Zhang, L. Jiang, M. A. Arbing, B. L. Nannenga, J. Hattne, İ. Polat, Y. Feng, E. W. Moore, J. V. Plas, D. Laxalde, J. Perktold, R. Cimrman, I. Henriksen, J. Whitelegge, A. S. Brewster, M. Messerschmidt, S. Boutet, N. K. Sauter, T. Gonen, E. A. Quintero, C. R. Harris, A. M. Archibald, A. H. Ribeiro, F. Pedregosa, P. van Mulbregt, D. S. Eisenberg, Structure of the toxic core of -synuclein from invisible crystals. SciPy 1.0 Contributors, SciPy 1.0--Fundamental algorithms for scientific computing in Nature 525, 486–490 (2015). Python. arXiv:1907.10121 [cs.MS] (23 July 2019). 38. M. D. Tuttle, G. Comellas, A. J. Nieuwkoop, D. J. Covell, D. A. Berthold, K. D. Kloepper, 59. S. B. Plaisier, R. Taschereau, J. A. Wong, T. G. Graeber, Rank-rank hypergeometric overlap: J. M. Courtney, J. K. Kim, A. M. Barclay, A. Kendall, W. Wan, G. Stubbs, C. D. Schwieters, Identification of statistically significant overlap between gene-expression signatures. V. M. Y. Lee, J. M. George, C. M. Rienstra, Solid-state NMR structure of a pathogenic fibril Nucleic Acids Res. 38, e169 (2010). of full-length human -synuclein. Nat. Struct. Mol. Biol. 23, 409–415 (2016). 60. R. C. Team, R: A Language and Environment for Statistical Computing (R Foundation for 39. L. Bousset, L. Pieri, G. Ruiz-Arlandis, J. Gath, P. H. Jensen, B. Habenstein, K. Madiona, Statistical Computing, 2016). V. Olieric, A. Böckmann, B. H. Meier, R. Melki, Structural and functional characterization of two alpha-synuclein strains. Nat. Commun. 4, 2575 (2013). Acknowledgments: We wish to express our gratitude to A. R. Crossman (University of 40. Y. Li, C. Zhao, F. Luo, Z. Liu, X. Gui, Z. Luo, X. Zhang, D. Li, C. Liu, X. Li, Amyloid fibril Manchester, UK) for comments for language supervision. We also thank M. Bosc (Cold structure of -synuclein determined by cryo-electron microscopy. Cell Res. 28, 897–903 Spring Harbor, USA) for valuable comments on the manuscript. We thank C. L. Martínez (2018). (Veterinary Service, University of Murcia) for administrative assistance; M. F. Ros Romero 41. R. Guerrero-Ferreira, N. M. Taylor, D. Mona, P. Ringler, M. E. Lauer, R. Riek, M. Britschgi, and J. Martínez Rabadán (University of Murcia) for veterinary and husbandry support; H. Stahlberg, Cryo-EM structure of alpha-synuclein fibrils. eLife 7, e36402 (2018). A. L. Gil, L. Cuenca, and I. Mascarell from the Clinical and Experimental Neuroscience 42. D. R. Boyer, B. Li, C. Sun, W. Fan, M. R. Sawaya, L. Jiang, D. S. Eisenberg, Structures of fibrils group (University of Murcia) for technical help with various parts of the in vivo part of formed by -synuclein hereditary disease mutant H50Q reveal new polymorphs. these complex experiments. We would like to thank P. Hantraye (MIRCen) for providing Nat. Struct. Mol. Biol. 26, 1044–1052 (2019). baboon stereotactic frame. The University of Bordeaux and the CNRS provided 43. A. Esteva, B. Kuprel, R. A. Novoa, J. Ko, S. M. Swetter, H. M. Blau, S. Thrun, infrastructural support. Funding: This work was supported by a grant from the Michael Dermatologist-level classification of skin cancer with deep neural networks. J. Fox Foundation (project grant number 2013-8499), Fundacion de Investigacion HM Nature 542, 115–118 (2017). Hospitales (Madrid, Spain), the Fundación Séneca (project grant number FS19540/PI/14), 44. D. Castelvecchi, Can we open the black box of AI? Nature 538, 20–23 (2016). the TARGET PD ANR grant, and The Simone and Cino Del Duca Prize from the French 45. J.-S. Park, B. Koentjoro, D. Veivers, A. Mackay-Sim, C. M. Sue, Parkinson's disease- Academy of Sciences. J.-A.O. and I.T.D. were funded by MINECO/AEI/FEDER-UE associated human ATP13A2 (PARK9) deficiency causes zinc dyshomeostasis (SAF2015-67239-P). M.B. and M.-L.A. were supported by a Ministère de l’Enseignement and mitochondrial dysfunction. Hum. Mol. Genet. 23, 2802–2815 (2014). Supérieur et de la Recherche fellowship and the France Parkinson Foundation (M.B.). The 46. E. Diguet, C. E. Gross, F. Tison, E. Bezard, Rise and fall of minocycline in neuroprotection: help of the Bordeaux Imaging Center, part of the national infrastructure France Need to promote publication of negative results. Exp. Neurol. 189, 1–4 (2004). BioImaging, granted by ANR-10INBS-04-0, is acknowledged. The human -Syn 47. R. Aron Badin, M. Vadori, E. Cozzi, P. Hantraye, Translational research for Parkinson’s aggregation TR-FRET immunoassay was performed in the Biochemistry and Biophysics disease: The value of pre-clinical primate models. Eur. J. Pharmacol. 759, 118–126 Platform of the Bordeaux Neurocampus at the Bordeaux University funded by the LABEX (2015). BRAIN (ANR-10-LABX-43) with the help of Y. Rufin. Computing time for this study was 48. M. Ahmed, A. Berthet, E. Bychkov, G. Porras, Q. Li, B. H. Bioulac, Y. T. Carl, B. Bloch, S. Kook, provided by the MCIA, the public research HPC center in Aquitaine, France. The samples I. Aubert, S. Dovero, E. Doudnikoff, V. V. Gurevich, E. V. Gurevich, E. Bezard, Lentiviral were obtained from the Brain Bank GIE NeuroCEB (BRIF number 0033-00011), funded by overexpression of GRK6 alleviates l-dopa-induced dyskinesia in experimental Parkinson’s the patients’ associations France Alzheimer, France Parkinson, ARSEP, and “Connaître les disease. Sci. Transl. Med. 2, 28ra28 (2010). Syndromes Cérébelleux” to which we express our gratitude. The synchrotron Diamond is 49. M. Bidinosti, D. R. Shimshek, B. Mollenhauer, D. Marcellin, T. Schweizer, G. P. Lotz, acknowledged for provision of beam time (exp. SP13009). Author contributions: M.B., M. G. Schlossmacher, A. Weiss, Novel one-step immunoassays to quantify -Synuclein: M.V., J.A.O., P.D., B.D., and E.B. conceived and designed the study. M.B., G.P., I.T.D., C.E., Applications for biomarker development and high-throughput screening. J. Biol. Chem. N.G.-C., M.T.H., B.D., and E.B. performed the surgeries. S.C. and C.E. performed the 287, 33691–33705 (2012). behavioral analyses. M.G. set up the actimetry behavioral platform. S.D., M.-L.A., A.P., and 50. G. Zandomeneghi, M. R. H. Krebs, M. G. McCammon, M. Fändrich, FTIR reveals structural P.A. performed the histologic and immunohistochemical analysis of the data. S.D., A.P., differences between native -sheet proteins and amyloid fibrils. Protein Sci. 13, and M.-L.A. performed the imaging experiments. E.D. performed the electron microscopy 3314–3321 (2004). analysis. F.L., M.-L.A., and M.-L.T. performed the biochemistry experiments. C.P. 51. J. Altmann, Observational study of behavior: Sampling methods. Behaviour 49, 227–267 performed and analyzed the primary culture experiment. M.-L.A., S.B., and B.D. performed (1974). the synchrotron analysis. C.S. performed the infrared microscopy. N.K. and B.M. Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 17 of 18 SCIENCE ADVANCES | RESEARCH ARTICLE performed biological fluid analysis. S.N. and M.M. performed the HPLC analysis. T.L.-L. (http://doi.org/10.5281/zenodo.1240558) is provided (https://zenodo.org/ performed the mRNA extraction and qPCR analysis. M.B., A.N., S.D., M.-L.A., S.C., N.P.R., record/1240558#.XC8pqy17Su4). S.B., C.S., F.L., N.K., B.M., S.N., M.M., C.P., A.R., N.N.V., O.M.A.E.-A., M.T.H., P.D., M.V., J.A.O., B.D., and E.B. analyzed the data. M.B., A.N., and N.P.R. developed the MLP approach. M.B., Submitted 21 October 2019 M.V., J.A.O., B.D., and E.B. wrote the paper. M.T.H. supervised the in vivo phase of the Accepted 6 March 2020 experiments, while B.D. and E.B. supervised the whole project. All authors discussed the Published 13 May 2020 results, assisted in the preparation, and contributed to the manuscript. All authors 10.1126/sciadv.aaz9165 approved the final version of the manuscript. Competing interests: E.B. is a director and a shareholder of the Motac Neuroscience Ltd. The other authors declare that they have Citation: M. Bourdenx, A. Nioche, S. Dovero, M.-L. Arotcarena, S. Camus, G. Porras, M.-L. Thiolat, no competing interests. Data and materials availability: The entire raw dataset is made N. P. Rougier, A. Prigent, P. Aubert, S. Bohic, C. Sandt, F. Laferrière, E. Doudnikoff, N. Kruse, available to the readers (table S2). The authors chose not to provide representative B. Mollenhauer, S. Novello, M. Morari, T. Leste-Lasserre, I. T. Damas, M. Goillandeau, C. Perier, examples of each procedure for the sake of space and because the entire dataset is fully C. Estrada, N. Garcia-Carrillo, A. Recasens, N. N. Vaikath, O. M. A. El-Agnaf, M. T. Herrero, P. Derkinderen, disclosed. Further information and requests for examples should be directed to and will M. Vila, J. A. Obeso, B. Dehay, E. Bezard, Identification of distinct pathological signatures induced be fulfilled by the corresponding contacts. Hyperlink to the machine learning code by patient-derived -synuclein structures in nonhuman primates. Sci. Adv. 6, eaaz9165 (2020). Bourdenx et al., Sci. Adv. 2020; 6 : eaaz9165 13 May 2020 18 of 18