The grey matter is a central target of pathological processes in neurodegenerative disorders such as Parkinson’s and Alzheimer’s diseases. The grey matter is often also affected in multiple sclerosis, an autoimmune disease of the central nervous system. The mechanisms that underlie grey matter inflammation and degeneration in multiple sclerosis are not well understood. Here we show that, in Lewis rats, T cells directed against the neuronal protein β-synuclein specifically invade the grey matter and that this is accompanied by the presentation of multifaceted clinical disease. The expression pattern of β-synuclein induces the local activation of these T cells and, therefore, determined inflammatory priming of the tissue and targeted recruitment of immune cells. The resulting inflammation led to significant changes in the grey matter, which ranged from gliosis and neuronal destruction to brain atrophy. In humans, β-synuclein-specific T cells were enriched in patients with chronic-progressive multiple sclerosis. These findings reveal a previously unrecognized role of β-synuclein in provoking T-cell-mediated pathology of the central nervous system.
The complex neuronal networks of the grey matter in the central nervous system (CNS) constitute the site at which the signals for body control are received, processed and distributed. Consequently, any damage to this neuronal fulcrum has a most severe and enduring impact on the integrity and function of the CNS, as is seen in chronic neurodegeneration of the brain in disorders such as Parkinson’s and Alzheimer’s diseases. In multiple sclerosis (MS), an autoimmune disease of the CNS, it is not only the myelinated white matter that is affected by the autoimmune attack, as previously thought1,2,3, but also the grey matter. In fact, it is the destruction of the grey matter that is primarily responsible for the brain atrophy and disease progression observed in MS4. An important gap in our understanding of the pathogenesis of MS results from the fact that the currently available experimental models—in particular, experimental autoimmune encephalomyelitis (EAE)—do not reproduce this critical hallmark of the human disease. The target autoantigens of classical EAE models are generally myelin proteins, such as myelin basic protein (MBP), myelin oligodendrocyte glycoprotein or proteolipid protein5. These proteins are predominantly expressed in CNS white matter, in which the EAE lesions are also located.
β-Synuclein-specific T cells induce grey matter disease
Here we studied T cells specific for β-synuclein, which is abundantly present within neuronal synapses6 and has previously been proposed to be a potential CNS autoimmune target antigen7. T cell lines specific for β-synuclein and for MBP were established and labelled with fluorescent proteins by retroviral gene transfer (Tβ-syn and TMBP cells, respectively) (Extended Data Fig. 1a–c). After intravenous transfer into healthy recipient rats, TMBP cells infiltrated the white matter of the spinal cord and the brain, as expected, and evoked the typical monophasic EAE characterized by ascending paresis and weight loss8,9. Tβ-syn cells were also highly pathogenic. However, in notable contrast to TMBP cells, the Tβ-syn cells almost exclusively invaded the grey matter, and in particular the cerebral cortex, an area that is almost devoid of TMBP cells in EAE (Fig. 1a and Extended Data Fig. 1d). Clinically, the rats developed a very heterogeneous disease phenotype that included atypical signs such as paresis of individual limbs, head tilting and ataxia (Extended Data Fig. 1e, f).
Our T cell lines consist of polyclonal populations with potentially distinct T cell receptor (TCR) reactivity (Extended Data Fig. 1g). To ascertain whether β-synuclein acted as the target autoantigen for the observed grey matter disease, we generated a TCR-transgenic rat, which had an immune repertoire that mainly (over 95%) contained T cells that were reactive against β-synuclein (Fig. 1b, c and Extended Data Fig. 1h, i). T cell lines from these TCR-transgenic rats (Extended Data Fig. 1j–o) induced massive grey matter lesions similar to wild-type Tβ-syn cells, and the majority of rats (more than 80%) displayed an atypical disease characterized by paroxysmal bouts of monotonous scratching movements, jerks, pareses and ataxic and/or rotational movements (Fig. 1d, Extended Data Fig. 1p and Supplementary Video 1).
Migratory paths and cues of Tβ-syn cells
Disease signs induced by Tβ-syn cells started, similarly to EAE induced by TMBP cells, after a preclinical phase of two to three days after T cell transfer (Extended Data Fig. 1q). During this prodromal phase, Tβ-syn cells—as do TMBP cells8,10—travelled along a defined route through the lung, mediastinal lymph nodes, spleen and blood before they arrived at the CNS. On their way, the Tβ-syn cells assumed a migratory mode; that is, they showed downregulation of activation markers and concurrent upregulation of chemokine receptors and adhesion molecules that are required for crossing the blood–brain barrier (BBB)10 (Extended Data Fig. 2a, b). Intravital two-photon laser-scanning microscopy recordings revealed that the Tβ-syn cells accumulated at the leptomeninges before disease onset (Fig. 2a and Extended Data Fig. 2a). At the same time as the start of clinical signs, Tβ-syn cells invaded deep into the adjacent cortical grey matter (Fig. 1d, Extended Data Fig. 1p and Supplementary Video 2). The peak of Tβ-syn cell infiltration was reached shortly after the manifestation of clinical signs. Thereafter, Tβ-syn cell numbers rapidly declined (Extended Data Fig. 2a). TMBP cells also used the CNS-encasing leptomeninges as the point of entry; however, they were substantially confined to the spinal cord (Fig. 2a, Extended Data Fig. 2a and Supplementary Video 2).
The intravascular motility pattern did not explain the different migratory targets of Tβ-syn and TMBP cells: both types of T cells rolled along or interacted more tightly with the vascular walls of brain and spinal cord vessels—that is, after fixed adherence, the cells crawled with comparable velocity along the endothelium11 (Extended Data Fig. 2c, d and Supplementary Video 3). Even CNS-ignorant ovalbumin-specific T cells (TOVA cells) had similar intravascular locomotion behaviour (Supplementary Video 4).
Previous reports have suggested that different cytokine and integrin profiles could direct T cells into different CNS regions. T helper 17 (TH17) cells invaded the brain using LFA-1, whereas TH1 cells entered the spinal cord through VLA-4 integrin12,13,14,15. Transcriptome analyses revealed that Tβ-syn and TMBP cells had almost identical expression patterns of master transcription factors, cytokines, cytokine and chemokine receptors, and adhesion molecules (Fig. 2b and Extended Data Fig. 2e). The cells expressed IFNγ and IL-17 (Extended Data Fig. 1o), LFA-1 and VLA-4, and CCR6 and CXCR3, indicating a mixed TH1/TH17 cell type. Their chemotactic responsiveness was almost identical (Extended Data Fig. 2f). As in TMBP cells11, VLA-4 but not LFA-1 was crucial for Tβ-syn cell invasion of the CNS. VLA-4 monoclonal antibodies almost completely abolished Tβ-syn cell rolling and crawling along brain vessels, as well as subsequent grey matter invasion and disease progression (Fig. 2c, Extended Data Fig. 2g–m and Supplementary Video 5). Interference with IFNγ and, to a lesser extent, IL-17 signalling reduced the entry of Tβ-syn cells into CNS tissue and ameliorated the disease (Fig. 2c and Extended Data Fig. 2n, o).
T cell activation drives grey matter inflammation
Because the differences in the migratory properties or adhesion molecule and cytokine profiles of the T cells could not explain the preferential targeting of the grey or white matter, we next investigated whether the distribution of CNS antigens regulates invasion. Whereas MBP is predominantly expressed within highly myelinated white matter, β-synuclein is mainly expressed in grey matter with its high density of neuronal processes and synapses16. Tβ-syn cells accumulated mainly in areas of high β-synuclein expression (Extended Data Fig. 3a–c). Thus, antigen expression could lead to preferential T cell activation within distinct brain regions. Indeed, Tβ-syn cells isolated from the leptomeninges and, in particular, from the brain parenchyma displayed an activatory gene profile. Vice versa, TMBP cells were activated mainly within the spinal cord white matter and significantly less so within the grey matter (Extended Data Fig. 4a–g). Using nuclear translocation of the fluorescently labelled nuclear factor of activated T cells (NFAT) as an activation-dependent biosensor17,18, we could confirm Tβ-syn cell activation within the meninges and deep within the grey matter (Fig. 3a, b, Extended Data Fig. 4h–k and Supplementary Video 6).
The dominant pattern of activation of Tβ-syn cells within the grey matter indicates that β-synuclein is preferentially presented in grey matter tissue. In fact, grey-matter-derived microglia induced a stronger activation response in Tβ-syn cells than did microglia from whole-brain tissue, the latter of which also contained microglial cells from the white matter of the brain. In the presence of the appropriate neuronal or myelin antigen, microglia from the grey matter or whole-brain tissue equally stimulated both Tβ-syn and TMBP cells. As expected, TOVA cells were not activated by the CNS antigen-presenting cells (APCs). Blockade of the interaction between TCRs and MHC class II by monoclonal antibodies prevented the activation of Tβ-syn cells (Fig. 3c and Extended Data Fig. 5a).
The relevance of in situ activation of Tβ-syn cells was supported by the following findings. Intrathecal application of anti-MHC class II monoclonal antibodies decreased Tβ-syn cell activation in the CNS, Tβ-syn cell recruitment and clinical disease (Fig. 3d and Extended Data Fig. 5b–d). Interference with the Tβ-syn cell-signalling cascade through treatment with FK506, which prevents nuclear NFAT translocation, resulted in very similar effects (Fig. 3d and Extended Data Fig. 5e–g). By contrast, promotion of T cell activation had the opposite effect. Even brain antigen-irrelevant T cells could be guided into the brain tissue. To achieve this, we used TOVA cells, which carried a chimeric antigen receptor (Extended Data Fig. 6a, b), that were stimulated by local application of CAR crosslinking monoclonal antibodies—thus avoiding the need for the local antigen presentation machinery. Involving CNS APCs in the T cell activation process by local ovalbumin application uniformly triggered increased invasion of TOVA cells (Extended Data Fig. 6c). Furthermore, local application of MBP redirected white matter-affine TMBP cells into the brain cortex (Extended Data Fig. 6d). Additionally, invasion of Tβ-syn cells could be enhanced by increasing the availability of the autoantigen by neuronal overexpression of β-synuclein using viral gene transfer (Extended Data Fig. 6e–g).
The activation of Tβ-syn cells triggered an inflammatory reaction of endothelial cells and the grey matter parenchyma with significant upregulation of adhesion molecules and chemokines (Fig. 4a, b). As a consequence, strong recruitment of host T cells and myeloid cells was observed (Extended Data Fig. 7a). The correlations between effector T cell activation, local tissue response and subsequent cell recruitment also became evident in studies in which Tβ-syn and TOVA cells were transferred together. When transferred alone, very few TOVA cells entered the leptomeninges. After transfer with Tβ-syn cells, however, numerous TOVA cells were recruited to the cortical leptomeninges and parenchyma. Their location, migratory properties and kinetics in the grey matter resembled those of Tβ-syn cells, although their activation state remained low (Extended Data Fig. 7b–d and Supplementary Video 7). Interfering with Tβ-syn cell activation efficiently reduced tissue inflammation and TOVA cell recruitment (Extended Data Figs. 5d, 7e, f).
Tβ-syn cells induce permanent neurodegeneration
We next analysed the consequences of Tβ-syn-cell-driven autoimmune attacks for the grey matter, which in its healthy state is almost devoid of any peripheral immune cells. Gadolinium (Gd)-enhanced magnetic resonance imaging (MRI) revealed marked signal increases in the meninges and within the cortex during disease onset, which indicates a substantial leakiness of the BBB. Intravital two-photon laser-scanning microscopy analyses after MRI confirmed these findings. Furthermore, subsequent histological analysis revealed a strong inflammatory response in the regions with Gd leakage (Fig. 5a, Extended Data Fig. 8a–d and Supplementary Video 8).
Tβ-syn cells within the grey matter crawled at high speed (around 10 μm min−1) through the dense parenchyma (Extended Data Fig. 8e–h and Supplementary Video 2) and were regularly found to be in close association with neuronal structures. Some cells were even located within neurons (emperipolesis), which suggests neuronal damage19,20 (Fig. 5b, c and Supplementary Video 9). In fact, in addition to significant glial activation—upregulation of IBA1 associated with significant morphological changes in microglia—we found signs of neuronal damage, such as a significant reduction in synaptic spines and the emergence of apoptotic neurons (Fig. 5c and Extended Data Fig. 8i–m). Most of the inflammatory changes disappeared after recovery from the acute autoimmune attack (Fig. 5a and Extended Data Fig. 8l, m).
Patients with MS generally develop multiple disease bouts with clinical remissions between relapses. This could faithfully be reproduced in our rat model by performing consecutive transfers of effector T cells in the TCR-transgenic hosts (Extended Data Fig. 9a–d). Morphological analyses three weeks after the peak from a fourth Tβ-syn-cell-induced disease bout showed that—similarly to the recovery state after an acute lesion—the strong inflammation of the grey matter had almost completely vanished. Only single groups of T cells remained around vessels or within the leptomeninges and the adjacent parenchyma (Extended Data Fig. 9e). However, in contrast to the recovery state after a single inflammatory bout, we found significant persisting gliosis with activated microglia and astrocytes after repeated bouts (Fig. 5d, e and Extended Data Fig. 9f). Moreover, clear signs of structural neuronal damage, such as destruction of the regular architecture of neurons and their processes, could be observed (Extended Data Fig. 9g). In fact, quantitative analyses revealed significant atrophy of the cortical tissue (Fig. 5f and Extended Data Fig. 9h). This permanent damage of the grey matter was confirmed using longitudinal MRI studies that demonstrated significant thinning of the brain cortex and enlargement of the ventricles. MRI analyses also revealed the dynamics of the atrophic process that had already started after the first bout, but which were significantly increased after consecutive attacks (Fig. 5g and Extended Data Fig. 9i–k).
β-Synuclein reactivity in patients with MS
Myelin-reactive T cells are known to be part of the immune repertoire of healthy people and are increased in patients with MS21,22,23,24. Studying the blood of healthy subjects, we found both MBP- and β-synuclein-reactive T cells. In patients with MS, the frequency of both cell specificities was higher (Fig. 6 and Extended Data Fig. 10a). As described previously for potentially auto-aggressive T cells25,26,27, MBP- and β-synuclein-reactive T cells upregulated ICOS and CD25, expressed the CXCR3 and CCR6 chemokine receptors and produced proinflammatory cytokines, such as IFNγ and, to a lesser extent, IL-17 (Extended Data Fig. 10b–e). Notably, the frequency of β-synuclein-specific T cells was especially increased in patients with chronic-progressive MS, whereas the frequency of MBP-specific T cells was mainly increased in patients with relapsing-remitting MS (Fig. 6). The frequency of β-synuclein-specific T cells correlated with disease severity and duration (Extended Data Fig. 10f, g). Recently, α-synuclein-reactive T cells have been reported to be increased in patients with Parkinson’s disease28. We confirmed the increased frequency of α-synuclein-reactive T cells in patients with Parkinson’s disease, whereas β-synuclein-reactive T cells were not overrepresented in the blood of these patients. In patients with MS, the numbers of not only β-synuclein-specific T cells but also α-synuclein-specific T cells were increased (Fig. 6).
Our β-synuclein model makes it possible to study autoimmune processes that are directed against the complex and vulnerable grey matter, which distinguishes it from classic EAE models that affect only the ‘wiring’ and therefore the transmission of signals. The inability of EAE models to reflect grey matter pathology has generated speculation that alternative pathological mechanisms such as hypoxia or currently undefined soluble toxic factors initiate the degenerative grey matter process in MS29. Our data show that inflammation not only in white matter but also in grey matter can be induced by a targeted T-cell-driven attack. Antigen expression within the CNS and the specificity of T cells were crucial factors that determined the CNS target structures—not, as previously hypothesized, a special set of cytokines or adhesion molecules, or gravity12,13,14,15,30. Activation-induced release of cytokines by the T cells triggered a glial and vascular reaction that resulted in the formation of inflammatory lesions in the grey matter. It remains to be further clarified why it is that immunity against β-synuclein but not any of the other previously tested neuronal antigens, such as neurofilament or contactin-231,32, induces this overt, severe grey matter disease. Apart from a genetically encoded susceptibility—for example, the presence of certain MHC haplotypes that regulate the presentation capacity of antigenic peptides—the high expression levels of β-synuclein might have a conducing role. Furthermore, the association of β-synuclein with synapses and released vesicles—structures that have a high turnover within the CNS and as a consequence are in the reach of brain resident APCs—should render this neuronal autoantigen particularly recognizable by autoreactive T cells33. The tendency of synucleins to form aggregates34 may be another contributing factor. Notably, β-synuclein-specific T cells not only cause inflammation but also induce significant chronic glial activation and neurodegeneration, particularly after repeated inflammatory bouts. The extent to which glial activation and/or inflammatory factors, such as cytokines, or the disruption of the BBB participate in the neurodegenerative process remains to be elucidated1,35,36,37. The neurodestructive component of our model reflects well the situation in MS, in which destruction of the grey matter is a hallmark of the disease—in particular, in the case of longstanding disabling MS4. Indeed, we found that β-synuclein-specific T cells were mainly increased in patients with chronic-progressive MS. Myelin-reactive T cells were predominant in patients with relapsing-remitting MS. This distinct chronological preponderance of the cell specificities could indicate that the β-synuclein-reactive T cells arise secondarily to the myelin-reactive T cells (for example, in the process of epitope spreading)38. However, β-synuclein-specific T cells were also detectable early in the disease process of cases with relapsing-remitting MS. Furthermore, inflammatory grey matter lesions were also observed early in MS2. A grey matter attack by neuronal antigen-reactive T cells might therefore be present from the very beginning of the autoimmune process and the accumulating irreversible neurological deficits of patients with chronic MS might be due to an exhaustion of the functional reserve, not necessarily because of epitope spreading. Clearly, future studies are needed to provide more insights into when β-synuclein-specific T cells expand, the functional role of these cells and how different treatments influence their functionality. In summary, our autoimmune grey matter inflammation model lends itself to the study of autoimmunity-mediated neurodegeneration and may open new avenues for the development of individual prognostic profiles and therapeutic measures for patients with MS.
Lewis rats on a LEW/Crl background (Rattus norvegicus) were bred at the animal facility of the University Medical Center Göttingen (Germany). The rats were kept under standardized conditions on a 12-h light–dark cycle. They were provided with food and drink ad libitum. All experiments were performed according to the local animal welfare regulations. Both male and female rats were used in the EAE experiments. No differences were noted between the sexes.
Generation of transgenic rats
Wild-type Lewis rats were immunized with rat β-synuclein protein (Creative BioMart). After 10 days, lymph node cells were isolated and amplified by exposing them to β-synuclein (10 μg ml−1). Seven days later, lymphocytes were rechallenged using a β-synuclein peptide library covering the entire protein sequence. The peptides β-syn79–98 and β-syn93–111 gave the most pronounced proliferative response (Extended Data Fig. 1a). β-Syn93–111 was then selected for further testing and for the generation of transgenic rats. The generation of β-synuclein-specific TCR-transgenic Lewis rats and their detailed characterization will be described elsewhere (L.D., manuscript in preparation). In brief, clonal effector T cell lines established from immunized Lewis rats were used to determine the sequence of rearranged α and β TCR chains that confer specificity to rat β-syn93–111. These sequences were fused as a tricistronic eGFP-T2A-TCRα-P2A-TCRβ cassette, which was cloned into the self-inactivating lentiviral vector FUGW by replacing the GFP insert. Gene transfer was performed by microinjection of concentrated viral particles into the perivitelline space of fertilized Lewis oocytes (donors obtained from Charles River) as previously described39. Founder rats were identified by genomic PCR of skin biopsies and flow cytometry analysis of blood samples. Rats were bred to homozygosity and maintained for several generations under standard conditions.
MBP was extracted from guinea pig brain as described40, OVA was purchased from Sigma, recombinant rat β-synuclein was purchased from Creative BioMart or purified in-house from bacterial E. coli lysates expressing recombinant protein with C-terminal 6×His tag using a Ni-agarose column. All peptides were synthetized by the peptide facility of the Charité (Berlin).
Generation and culturing of T cells
CD4+ T cells retrovirally engineered to express fluorescent proteins (eGFP, mCherry, TdTomato, Turquoise2 or LifeAct-Turquoise2) or NFAT reporter (NFAT–YFP or NFAT–YFP/Cherry-H2B) and reactive against β-syn93–111, MBP or OVA (Tβ-syn, TMBP or TOVA cells) were established as reported17,41. For the generation of chimeric antigen receptor-expressing T cell lines (CAR T cells), the sequence encoding SCA431scFv-mIgG1-CD4tm-CD3ζ CAR specific for carcinoembryonic antigen42 (courtesy of H. Abken, University of Cologne) was excised from the pBullet vector and cloned into a pMSCV-Puro-P2A vector, which was then used to establish a stable retroviral packaging cell line. The mouse single-chain antibody domain of CAR was used to activate CAR T cells in situ by application of crosslinking anti-IgG1 monoclonal antibodies.
In brief, 6–8-week-old female Lewis rats were immunized subcutaneously with 150 μl of emulsion prepared from β-syn93–111, MBP or OVA (stock concentration: 1 mg ml−1) and an equal volume of complete Freund’s adjuvant containing Mycobacterium tuberculosis extract (Difco) at a concentration of 2 mg ml−1. The cell suspension isolated from the draining lymph nodes 9–10 days after immunization was co-cultured with GP+E86 packaging cell lines producing replication-deficient retroviruses in the presence of 10 μg ml−1 MBP or OVA or 8 μg ml−1 of β-syn93–111. After 48 h, medium containing horse serum (10%, Biochrom AG) and mouse IL-2 was added to the culture for T cell expansion. Starting from day 4 after antigen stimulation, transduced T cells were selected by the addition of G418 (400 mg ml−1, Thermo Fisher) for two weeks or puromycin (1 μg ml−1, Carl Roth) for one week. Seven days after primary stimulation, T cells were rechallenged with the cognate antigen in the presence of 30-Gy-irradiated thymic APCs. The primary cell lines underwent at least three cycles of stimulation in culture before being used for transfer experiments. For the derivation of transgenic T cell lines, lymph node cells were explanted six days after immunization and antigen restimulation was performed every six days. All established T cell lines were CD4+CD8−αβTCR+. They displayed an effector memory phenotype (L-selectin−CD45RClowCD44high) and upon stimulation produced IFNγ and IL-17. Phenotype, cytokine profile, antigen specificity, pathogenicity and the absence of mycoplasma contamination were verified in each cell line.
In vitro activation assays were performed in 96-well plates by co-culturing 5 × 104 T cells and 106 30-Gy-irradiated thymocytes per well. For unspecific stimulation, T cells were added onto plates coated with anti-CD3 (1 μg ml−1, clone G4.18, BD Biosciences) and anti-CD28 (2 μg ml−1, clone JJ316, BD Biosciences) monoclonal antibodies or treated with 2 μg ml−1 concanavalin A (Sigma). For blocking MHC class II-mediated antigen presentation, APCs were pretreated with 1 μg ml−1 anti-RT1b monoclonal antibodies (clone OX-6, Biolegend) before being added to T cells.
Adoptive transfer EAE was induced by intravenous injection of T cell blasts (day 2 after stimulation). Unless stated differently, the following number of T cells were injected: 7–8 × 106 wild-type Tβ-syn cells, 2–3 × 106 transgenic Tβ-syn cells or 5 × 106 TMBP cells. For TOVA-cell transfer, 5 × 106 cells were injected. In cotransfer experiments, equal numbers (3–4×106) of TOVA–Tomato and Tβ-syn–Turquoise cells were injected intravenously into naive recipients.
Relapsing-remitting EAE was induced by repeated transfer of 2.5 × 106 transgenic Tβ-syn or TMBP cell blasts into TCR-transgenic rats. After each transfer, the rats developed a severe monophasic disease with comparable severity of disease symptoms. β-Synuclein transgenic animals repeatedly injected with TOVA cells (5 × 106 cells per rat) were used as control. As expected, upon repeated transfer of transgenic Tβ-syn cell blasts, wild-type rats became completely resistant to disease induction33.
Weight and clinical scores were recorded daily. Classical signs of EAE were scored as follows: 0, no disease; 1, flaccid tail; 2, gait disturbance; 3, complete hindlimb paralysis; 4, tetraparesis; 5, death. For atypical symptoms, the following classification was used: 0, no disease; 1, occasional twitches and scratching with or without flaccid tail; 2, frequent twitches and scratching, ataxia; 3, severe tonic and myoclonic movements, severe gait impairment; 4, tetraparesis; 5, death.
No statistical method was used to predetermine sample size. Rats from different cages but within the same experimental group were selected to assure randomization. Experimenters were blinded to the identity of experimental group in the experiments as follows: clinical effect of cytokine blockers; morphological analysis; acquisition and analysis of the human samples.
Cell isolation, flow cytometry and fluorescence-activated cell sorting
Fluorescently labelled T cells were isolated from the tissues as previously described8,10. In brief, mononuclear cells were isolated from EDTA-treated blood by density gradient. Single-cell suspensions were obtained from parenchymal organs, grinding the tissues through a mesh. The isolated cells underwent different procedures: spleen-derived single-cell suspensions were treated with ACK buffer for erythrocyte lysis. Immune cells were extracted from the different CNS compartments by a two-phase Percoll-density gradient. Isolation of CD11b+ microglial cells was performed as described43. In brief, rats were perfused with saline, meninges were removed and grey matter was isolated under a stereotactic microscope. CD11b+ microglial cells were then isolated with a three-phase Percoll-density gradient. Of note, no enzymatic digestion of the tissue was performed.
For the isolation of endothelial cells, euthanized rats were transcardially perfused with ice-cold PBS. Brain tissue was dissected and meninges were removed. The parenchyma was homogenized in a Dounce tissue grinder before the tissue was digested for 1 h in digestion buffer (HEPES-buffered DMEM medium supplemented with 0.4 U ml−1 liberase (Roche), 120 U ml−1 DNase I (Roche)) at 37 °C. Digestion was stopped by adding ice-cold buffer with 2 mM EDTA and samples were passed through a 20G syringe and filtered through a 70-μm cell strainer. Myelin debris was removed by 30% Percoll gradient centrifugation for 20 min at 700g. Cells were stained in staining buffer (PBS with 2 mM EDTA and 0.5% BSA) with anti-CD31–PE antibodies (Bio-Rad, clone TLD-3A12, dilution 1:100), anti-CD45–AF647 (Biolegend, clone OX-1, dilution 1:100) and anti-CD11b–AF647 (Biolegend, clone OX-42, dilution 1:100) for 30 min at 4 °C, washed twice with staining buffer then resuspended in HEPES-buffered DMEM with 2 mM EDTA for fluorescence-activated cell sorting (FACS). Labelled cell populations were sorted by a FACSAria 4L SORP cell sorter (Becton Dickinson). Sorted cells were pelleted for 5 min at 300g and taken up in QIAzol Lysis Reagent (Qiagen) for RNA isolation.
Flow cytometry analysis was performed with a FACSCalibur operated by Cell Quest software (Becton Dickinson) or by a CytoFLEX flow cytometer operated by CytExpert software (v.2.3, Beckmann Coulter). All procedures were performed at 4 °C. For surface staining, the following mouse anti-rat monoclonal antibodies were used: CD134–biotin (clone OX-40, Serotec), CD25–PE (clone OX-39, Biolegend), CD11b/c–AF647 (clone OX-42, Biolegend), αβTCR–AF647 or αβTCR–PE (clone R73, Biolegend), TCR Vβ8.2 (clone R78, Santa Cruz Biotechnology), TCR Vβ8.3 (clone B73, Serotec, described to recognize Vβ8.544, but it actually recognized Vβ8.3 as described45), TCR Vβ10 (clone G101, Serotec), TCR Vβ16.1 (clone His42, Serotec), γδTCR–PE (clone V65, Biolegend), CD45–PE (clone OX-1, Biolegend), CD45RA–PE (clone OX-33, Biolegend), CD8α–PE (OX-8, Biolegend), CD8β-PE (clone eBio341, eBioscience), CD3–AlexaFluor647 (clone 1F4, Biolegend), CD4–PECy5 (clone W3/25, BD Biosciences), CD4–PE–Cy7 (clone OX-35, BD Biosciences), CD28–APC (clone JJ319, Biolegend), CD62L–PE (clone OX-85, Biolegend), CD45RC–PerCP–Cy5.5 (clone OX-22, Abcam). Mouse IgG1κ (MOPC 31C, Sigma-Aldrich) served as isotype control; APC-labelled anti-mouse antibody (Dianova) was used as secondary antibody. For intracellular staining, mouse anti-rat IFNγ (DB-1, conjugated to AF647, Biolegend), rat anti-mouse IL-17–BV421 (clone TC11-8H4, Biolegend) and anti-Ki-67–AF647 (clone B56, BD Biosciences) antibodies were used. Before staining, cells were fixed and permeabilized using a BD Cytofix/Cytoperm kit (BD Biosciences) as recommended by the manufacturer. Mouse IgG1κ–AF647 (clone MOPC21, Biolegend) and rat IgG1κ–BV421 (clone RTK2071, Biolegend) antibodies were used as control. For the analysis of human peripheral blood mononuclear cell (PBMC) cultures the following antibodies were used: CD3–APC or CD3–APCFire750 (clone UCHT1, Biolegend), CD4–PE (clone RPA-T4, Biolegend), CD8–PECy5 (clone RPA-T8, Biolegend), CD25–PE–Cy5 (clone BC96, Biolegend), ICOS–PacificBlue (clone C398.4A, Biolegend). Cytofluorometric quantification of T cells was performed by relating the number of cells to a known absolute amount of fluorescent beads (BD Biosciences). Data analysis was performed using FlowJo software (FlowJo).
Cell proliferation analysis using BrdU incorporation assay
BrdU incorporation was measured as previously reported11. In brief, for ex vivo analysis BrdU (120 mg kg−1 body weight, Sigma-Aldrich) was injected intraperitoneally 2 h before organ collection. This time interval between BrdU injection and sample collection was chosen to minimize cell trafficking and to allow BrdU to enter the CNS parenchyma. For in vitro labelling, T cells were incubated with 10 μM BrdU for 2 h. Intracellular staining for BrdU was then performed using the BrdU APC Flow kit (BD Biosciences) according to the manufacturer’s instructions. The T cells were then analysed by flow cytometry.
Intravital two-photon laser-scanning microscopy
Two-photon laser-scanning microscopy (TPLSM) was used for imaging fluorescently labelled T cells in the brain of living rats. Imaging was performed in the preclinical phase and at the onset and peak of the disease. For this purpose, rats were pre-anaesthetized by intramuscular injection of 10 mg kg−1 xylazine combined with 50 mg kg−1 ketamine and then intubated via a small incision of the trachea and immediately ventilated with 1.5–2% isoflurane. During imaging, rats were stabilized in a custom-made microscope stage and their body temperature regulated and maintained (36–37 °C) via a heated pad connected to a custom-built thermocontroller. Fluid supply during imaging sessions was maintained using a perfusing device (Ismatec) set to a 0.6 ml h−1 flow rate. To access the parietal cortex, an open-skull window preparation was performed. In brief, a midline scalp incision was performed to expose the parietal, interparietal and frontal bones. The skull was then stereotactically fixed with dental paste to a head holder. Subsequently, a microdrill was used to cut an ellipsoid area of around 0.5 cm2 in the parietal bone (coordinates from bregma: x = +1.5 mm, y = −1 mm). The dura mater was carefully removed. For all preparations, tissue was immediately covered with sodium chloride solution to prevent dehydration.
Technical equipment and labelling procedures
TPLSM imaging was performed as previously described46, using two different systems: (1) a Zeiss Laser Scanning Microscope 710 (Carl Zeiss) combined with a Coherent 10 W Ti:Sapphire chameleon laser (Coherent), controlled by Zeiss ZEN 2012 software; and (2) an Olympus FVMPE-RS TPLSM equipped with a Spectra-Physics Mai Tai Ti:Sapphire oscillator and a Mai Tai DeepSee Ti:Sapphire oscillator. The excitation wavelength was tuned to 880 nm or 1,010 nm and routed through a 20× water 1.0 NA immersion objective W Plan Apochromat (Carl Zeiss) and a 25× water 1.05 NA immersion objective Olympus Scaleview in the Zeiss and in the Olympus TPLSM, respectively. Farred fluorescent proteins (for example, TdTomato) were excited at 1,100 nm either using a Ti:Sapphire laser pumped OPO in the Zeiss TPLSM or a Mai Tai DeepSee Ti:Sapphire oscillator in the Olympus TPLSM. Emitted fluorescence was detected using non-descanned detectors equipped with 442/46 nm, 550/49 nm and 624/40 nm band-pass filters. Typically, areas of 424.27 μm × 424.27 μm (512 × 512 pixels) width were scanned and 50–100-μm z-stacks were acquired. The acquisition rate during bidirectional scanning was approximately 1.3 s per z-plane including 2 times line-averaging. For reproducible motility analyses, the interval time was kept to 30 s. Blood vessels were labelled by intravenous injection of 2,000,000 Da dextran tetramethylrhodamine (Fischer Scientific) during the imaging session. Meningeal phagocytes were labelled by intrathecal injection of 3,000 Da dextran Texas red (Fischer Scientific) 24 h before imaging. To label myeloid cells, a monoclonal antibody against CD11b (OX-42) labelled in-house with SeTau647 (SETA BioMedicals) was injected intrathecally immediately before imaging. For evaluating BBB permeability, 70,000 Da dextran Texas red (Fischer Scientific) was injected before imaging.
Analysis of time-lapse videos
Acquired 3D time-lapse videos were analysed using Imaris 8.0.1 software (Bitplane AG). Cells were tracked using the automated Imaris Track module with subsequent manual revision. Motility parameters including T cell velocity (instantaneous speed average), crawling duration and meandering index (total T cell path length/single-cell displacement sum) were calculated as described46 within a 30-min recording interval. Rolling T cells were defined as cells appearing as single or several round-shaped dots with >50 μm min−1 instantaneous velocity. Owing to limitations in temporal resolution of the TPLSM, fast rolling events may have escaped our analysis.
In vivo integrin signal blockade
Integrin-mediated binding was blocked as previously described11 using mouse anti-rat monoclonal antibodies directed either against VLA-4 (anti-CD49d, clone TA-2; courtesy of T. Issekutz, Dalhousie University, Canada) or against CD11a (integrin αL, anti-LFA-1, clone WT.1, Serotec). For intravital TPLSM recordings in the intravascular phase, monoclonal antibodies were injected intravenously at a single dose of 1 mg kg−1 during the imaging session. After recording, ex vivo isolated T cells were tested by flow cytometry for antibody saturation. For assessing the clinical effects monoclonal antibodies were administered intrathecally twice (40 μg per rat per injection), namely on day 2.5 and 3.5 after T cell transfer.
In vivo inflammatory cytokine blockade
IFNγ and IL-17 were blocked by anti-rat IFNγ (provided by W. van der Meide) and anti-rat IL-17A (provided by Novartis) monoclonal antibodies, respectively. Antibodies were injected intrathecally (80 μg per rat per injection) on day 1.5, 2.5 and 3.5 after T cell transfer. In a preliminary experiment, mouse IgG1 or PBS were used as control. Given that results were identical between the IgG1 and PBS groups, in all subsequent experiments, PBS was used as control.
In vitro antigen presentation
Microglial cells (5 × 104 per 100 μl) isolated as described above were cocultured with Tβ-syn, TMBP or TOVA cells (3 × 104 per 100 μl for each cell line) in 96-well plates with or without the cognate antigen (10 μg ml−1) and in the presence or absence of MHC-II blocking monoclonal antibodies (10 μg ml−1). After 48 h, the cell culture supernatant was collected. Enzyme-linked immunosorbent assay (ELISA) assays were used to detect the concentrations of IFNγ and were carried out according to the manufacturer’s protocol (PeproTech).
A trans-well migration assay was used to test the chemotactic capacity of blood-isolated antigen-specific cells towards CCL5 (0.1 ng ml−1), CCL19 (0.1 ng ml−1), CCL20 (0.1 ng ml−1), CXCL11 (0.1 ng ml−1), CXCL12 (0.1 ng ml−1) as described47. All chemokine ligands were purchased from Peprotech. The number of effector T cells migrating in response to the chemokine ligands was counted by flow cytometry and quantified in relation to the basic migration in the absence of chemokine ligands.
In vivo MHC class II blockade
MHC class II-mediated antigen presentation was blocked by intrathecal injection of anti-rat MHC class II monoclonal antibodies (80 μg per rat per injection) performed on days 1.5, 2.5 and 3.5 after T cell transfer.
FK506 (Sigma-Aldrich) was administered as previously described17. The drug was dissolved in DMSO (Carl Roth) to a concentration of 2 mg ml−1 and injected intraperitoneally at a dose of 2 mg kg−1 once per day. Control rats received an intraperitoneal injection of an equal volume of DMSO.
Activation of TOVA–CAR cells
In vitro activation of resting TOVA–CAR cells (day 6 after antigen challenge) was achieved by two different methods: (1) stimulation with increasing doses (0.5, 5, 50 μg ml−1) of biotinylated anti-mouse IgG1 monoclonal antibodies (clone RMG1-1, BioLegend) with or without crosslinking streptavidin (0.5 μg ml−1, Biolegend); and (2) stimulation with anti-rat CD3 (0.1 mg ml−1, clone G4.18, BD Biosciences) and CD28 (0.5 mg ml−1, clone JJ316, BD Biosciences) monoclonal antibodies, which had been coated on the culture plate before adding the T cells.
For in vivo activation, rats were injected with 10 μl anti-mouse IgG1 monoclonal antibody (50 μg ml−1) 3 days after TOVA–CAR cell transfer. The monoclonal antibody was injected with a stereotactic devise (Narishige) either in the cisterna magna (intrathecally) or in the right ventricle (intracerebroventricular (i.c.v.); coordinates from bregma: x = +1.5 mm; y = −1.0 mm, z = −4.3 mm). The solution was injected at a rate of 10 μl in 5 min to limit increase in intracranial pressure. After 5 min, the needle was slowly retracted (1 mm every 5 min) to avoid reflow through the injection route. The two injection modes gave equivalent results and therefore the results were pooled.
In vivo delivery of soluble antigen
To boost antigen presentation, 3 days after transfer of TMBP cells or TOVA cells, rats were injected with MBP or OVA (10 μg per rat), respectively. The antigens were injected either intrathecally or i.c.v. as described above. The two injection modes gave equivalent results and therefore the results were pooled. TPLSM or FACS quantification was performed 12 h later.
Neuronal overexpression of β-synuclein
AAV vectors were constructed to express β-syn–eGFP, γ-syn–eGFP, or eGFP under the human synapsin promoter as previously described48. An AAV2 serotype was selected, because of its higher selectivity for neuronal cells. AAV vectors were injected intracranially into neonatal rats 24–48 h after birth. Each rat received 2 μl of solution containing 1.5 × 108 transducing units. Six weeks later, the rats were intravenously transferred with 2.5 × 106 Tβ-syn cells.
RNA extraction, cDNA library preparation and RNA sequencing
RNA extraction, cDNA library preparation and RNA sequencing (RNA-seq) was performed as described47. For T cell sequencing, total RNA from wild-type or transgenic Tβ-syn cells was isolated either from in vitro Tβ-syn cell blasts (20 h after antigen encounter) or ex vivo from blood, brain leptomeninges and brain parenchyma 3 days post transfer (p.t.). Between 200,000–400,000 Tβ-syn cells were sorted from each sample achieving more than 98% purity. Three different biological replicates were prepared for each sample. Four to five rats were pooled for each replicate. Library preparation for RNA-seq was performed using the TruSeq RNA Sample Preparation Kit (Illumina, RS–122–2002) starting from 500 ng of total RNA. Single-read (45-bp) sequencing was conducted using a HiSeq 2000 (Illumina). Sequences were aligned to the genome reference sequence of R. norvegicus allowing for 2 mismatches within 45 bases. Gene annotation was performed using R. norvegicus entries from Ensembl v.78. Data of gene expression profiles of TMBP cells were obtained from a previously published dataset47.
Quantitative PCR was performed on a StepOnePlus Real-Time PCR System (Applied Biosystems) using FAM-TAMRA-labelled probes as described10,47. ACTB was used as house-keeping gene. For detecting cytokine and chemokine expression on human samples the following primers/probe combination was used with RPL13A as house-keeping gene: IFNG: forward, TCTCTCGGAAACGATGAAATATACAA; reverse, TTTTACATATGGGTCC TGGCAGTA; probe, ATCTTGGCTTTTCAGCTCTGCATCGTTTTG. IL17: forward, TGTGATCTGGGAGGCAAAGT; reverse, CCCACGGACACCAGTA TCTT; probe, AGCCTCCCTGCGCAGGACCA. IL4: forward, GCCTGG CGGGCTTGA; reverse, AGCCTTTCCAAGAAGTTTTCCA; probe, CCTGTCCTGTGAAGGAAGCCAACCA. CXCR3: forward, CTATGACTA TGGAGAAAACGAGAGTGA; reverse, GGCCCGGTCGAAGTTCA; probe, TCGTGCTGTACCTCCCCGCCC. CCR6: forward, TTGTACAGGCGACTA AGTCATTCC; reverse, CCCACACAACAAGGCAGATG; probe, CTCCGA TCCAGAACACTACCGCGC. RPL13A: forward, CTGTGAAGGCATCAACA TTTCTG; reverse, TCGGGAAGGGTTGGTGTTC; probe, TTGAAGTACCTGG CTTTCCTCCGCAA
Magnetic resonance imaging
MRI was performed at 9.4 T. Radiofrequency excitation and signal reception were accomplished by a birdcage resonator and a four-channel phased-array surface coil, respectively (both from Bruker Biospin MRI). Multislice fast spin-echo MRI was used to perform T2-weighted axial MRI with repetition time (TR) of 9,286 ms, echo time (TE) of 11 ms, RARE factor of 12, 60 slices, in-plane resolution of 120 μm × 120 μm, slice thickness of 480 μm and a total acquisition time (TA) of 195 s. In addition sagittal MRI (TR/TE = 4,333/11 ms, RARE factor = 12, 28 slices, in-plane resolution = 120 μm × 60 μm, slice thickness of 480 μm and TA = 182 s) was performed. T1-weighted fat-suppressed gradient-echo MRI (3D FLASH, TR/TE = 14.8/4.2 ms, flip angle = 25° and TA = 16 min) was performed at an isotropic resolution of 120 μm before and after intravenous injection of 30 μl gadobutrol, a Gd-based contrast agent (stock concentration: 1 mM, Gadovist, Bayer). The analysis followed a strategy previously developed for intra-individual comparisons of MR images49. For quantifying Gd enhancement, MRI signal intensities were evaluated. For this purpose, regions of interest (ROIs) were selected in a standardized manner: the signal intensity in the meninges was calculated using mid-sagittal T1-weighted planes; for the analysis of signal intensities in parietal cortex, corpus callosum and thalamus, T1-weighted images 0.5 mm lateral from the mid-sagittal plane were used. The mean signal intensity was measured for all ROIs and normalized to the signal intensity of the thalamus. Subsequently, the ratio between normalized signal intensity before and after Gd injection was calculated. For measuring its thickness, the corpus callosum was marked in T2-weighted planes 0.5 mm lateral from the mid-sagittal section by two parallel lines. The distance between those lines was measured at both endpoints. For measuring cortical thickness, the distance between corpus callosum and meninges was measured at three points. The measurements were taken in both hemispheres and subsequently averaged. Ventricular size was assessed by 3D reconstruction of the third and lateral ventricles from T2-weighted images. All analyses were performed with Fiji software.
Analysis of neuronal synaptic spine density
Rats were transcardially perfused with ice-cold 1× PBS (2 min) and 4% PFA (10 min) at the peak of T cell infiltration. Hemispheres were then separated and post-fixed in 4% PFA for 30 min. One hemisphere was set aside for histology, the other embedded in 2–3% low-melt agarose (Roth) and cut into 300-μm thick slices on a vibratome (Leica VT1200S). Slices were washed three times 15 min in 1× PBS. Cortical neurons were analysed using DiOlistic labelling on acute slices as described50. In brief, DiI-coated tungsten particles (0.7 and 1.7 μm in diameter; Bio-Rad) were delivered to the slices using a hand-held gene gun (Helios Gene Gun System, Bio-Rad). After 48 h, images were acquired by confocal microscopy. Synaptic spines in apical dendrites from DiI-stained L2/3 pyramidal neurons were counted on deconvoluted images using Fiji software. The number of spines was normalized per micrometre of dendritic length.
Confocal microscopy was performed using a Zeiss Laser Scanning Microscope 710 (Carl Zeiss) controlled by Zeiss ZEN 2012 software (Carl Zeiss) and a custom-made microscope that combines two-photon, confocal and stimulated emission depletion microscopy (Abberior instruments). Fluorophores were excited using 405-nm UV-diode for DAPI, 488-nm Argon laser for GFP, a 561-nm DPSS laser for mCherry or DiI and a 633-nm HeNe laser for Cy5 dyes. Accordingly, filters were set from 415 to 470 nm for DAPI acquisition, 498 to 546 nm for GFP and 571 to 660 nm for mCherry/DiI/Cy5. A 40× oil/1.3 NA immersion objective Plan Apochromat objective (Carl Zeiss) was used. Images were acquired using a pinhole size of 50 μm. A z-step size of 0.5 μm was chosen to guarantee optimal z-resolution. T cell quantification was performed on single z-stack using the cell-counter plug-in in ImageJ with manual corrections.
Histology and immunohistochemistry
Rats were intracardially perfused with saline (10 min) followed by a fixative containing 4% PFA (15 min). Samples were post-fixed in the same fixative for 24 h and then transferred in 30% sucrose. Histological analysis was performed according to established procedures on paraffin and frozen sections. The following antibodies were used: mouse anti-rat RT1B-MHC class II antigen (clone OX-6, Serotec), rabbit anti-IBA1 (Wako), mouse anti-rat CD68 (clone ED1, Serotec), mouse anti-neurofilament marker (Biolegend), mouse anti-GFAP (Thermo Fisher), mouse anti-rat CD43 (clone W3/13, Biolegend), rabbit anti-β-synuclein (AB 5086, Millipore), rabbit anti-tyrosine hydroxylase (Millipore) and goat anti-mouse or goat anti-rabbit IgG antibodies labelled with Alexa Fluor 488, Alexa Fluor 555 or Alexa Fluor 647 (Invitrogen) as secondary antibodies. Images were acquired using either a VS120 Virtual Slide Microscope (Olympus) equipped with a 10× objective or with a Zeiss LSM700 confocal microscope equipped with a 40× Zeiss objective. For morphological quantification, images of brain sections cut coronally in series were acquired with a BZ-9000 Keyance microscope. The total length of the cortex ranging from the corpus callosum to the meningeal surface was measured along the orientation of the neuronal processes (dendrites) in the motor region M1 using the BZ-9000 Keyance software. In addition, the corpus callosum was measured at its thickest site close to the midline. Morphological analysis and 3D reconstruction of microglia stained with anti-rat IBA1 and astrocytes stained with anti-rat GFAP were done using Imaris 8.0.1 software (Bitplane AG) as previously described51. Confocal z-stacks with 0.5-μm step size were opened in Imaris and smoothed with both median and Gaussian filters. Single cells were rendered by first using the filament tool to reconstruct the processes and subsequently the surface tool to render the cell body. Process length and number of segments (segments between branch points or branch and endpoints) were analysed.
Rats were intracardially perfused with saline followed by a fixative containing 4% PFA and 1% glutaraldehyde. After postfixation in Ito’s solution and washing in cacodylate buffer, brain tissue was cut into 2 mm × 2 mm specimens. After postfixation in 1% osmium tetroxide, the specimens were dehydrated with graded alcohols and embedded in Epon (Roth). Then, 1-μm semi-thin sections were cut by a microtome (Reichert-Jung) and stained with Richardson’s staining. Ultrathin sections were cut with an Ultracut E microtome (Reichert-Jung), contrasted with lead citrate and uranyl acetate and imaged with a Zeiss EM 109.
Patients and control donors
T cell assays were performed using PBMCs prepared from 29 healthy controls, 40 patients with MS and 20 patients with Parkinson’s disease. Detailed clinical information is included in Extended Data Fig. 10a. This study was approved by the local ethical committee, and all patients gave their informed consent for the study.
Isolation of PBMCs, proliferation assay and CD3 stimulation
PBMCs isolated from human peripheral blood by Ficoll gradient were labelled with 5-chloromethylfluorescein diacetate (CFSE, Invitrogen) at a density of 1 × 106 cells per ml using 0.5 μM CFSE and cultured for 7 days in the presence of the following antigens: human β-synuclein (50 μg ml−1, Abcam), human α-synuclein (50 μg ml−1, provided by M. Zweckstetter, MPIBC, Germany), MBP (50 μg ml−1), OVA (50 μg ml−1, Sigma-Aldrich) or staphylococcal enterotoxin B (10 μg ml−1, Sigma-Aldrich). Before measurements, cells were stained with a mix of anti-CD3–AF647 (clone UCTH1, Biolegend), anti-CD4–PE (clone RPA-T4, Biolegend) and anti-CD8–PECy5 (clone RPA-T8, Biolegend) antibodies. Dilution of fluorescent labels was measured in CD3+CD4+ T cells on FACSCalibur or FACSAria 4L SORP cell sorter flow cytometers (BD Biosciences). The percentage of proliferating CD3+CD4+ T cells was used as a read-out.
For unspecific stimulation of cytokine production, CFSE-labelled PBMC cultures after 7 days of incubation with specific antigens were added onto plates coated with anti-human CD3 monoclonal antibody (5 μg ml−1 in PBS, clone OKT3, Biolegend). After 12 h, brefeldin A (Sigma-Aldrich) was added to a concentration of 5 μg ml−1 for 3 h incubation at 37 °C, and then cells were stained with anti-CD3 and anti-CD4 antibodies (described above) for 30 min on ice. Intracellular staining was performed using a BD Cytofix/Cytoperm staining kit (BD Biosciences) according to the manufacturer’s protocol. The following antibodies were used: anti-human IL-17A monoclonal antibody (clone BL168, labelled with AlexaFluor647, Biolegend, dilution 1:50), anti-human IFNγ monoclonal antibody (clone B27; labelled PacificBlue, dilution 1: 50) and isotype control antibodies (mouse IgG1k–PB and IgG1k–AF647, both from Biolegend).
ELISA for IFNγ and IL-17
PBMCs were cultured as described above the in presence of either human β-synuclein or MBP. After seven days in culture, supernatants were collected. An ELISA assay for detecting IFNγ and IL-17 concentrations was performed according to the manufacturer’s protocol (R&D).
Statistical analysis was performed using GraphPad Prism 7.03 software (GraphPad Software). The tests performed and confidence intervals are indicated in the figure legends.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
The datasets generated during and/or analysed during the current study are available from the corresponding authors on reasonable request.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank A. Stas, S. Mole, A. Mönnich, S. Hamann, M. Weig and H. Nguyen for technical assistance, C. Ludwig for text editing, G. Salinas-Riester for her support in performing the transcriptome analyses, K. Raithatha for her help in analysing the transcriptome data, H. Abken, S. Brioschi, L. Flügel, T. Issekutz, P. and W. von der Meide, M. Simons, T. Michaelis and M. Korte for providing reagents and/or technical, scientific or clinical advice and M. Gößwein for help with the artwork. This work was supported by the German Research Foundation (RK-Grant FL 377/3-1, FL 377/2-2; SFB 1328/1 A01, OD 87/1-1), the Federal Ministry for Education and Research (Competence Network Multiple Sclerosis, KKNMS), ERANET consortium MELTRA-BBB and the Ministry of Science and Culture of Lower Saxony (Niedersachsen-Research Network on Neuroinfectiology, N-RENNT).
Extended data figures and tables
About this article