Recursive seed amplification detects distinct α-synuclein strains in cerebrospinal fluid of patients with Parkinson's disease

Abstract

Parkinson's disease (PD) is a heterogeneous neurodegenerative disorder with a wide range of clinical phenotypes. Pathologically, it is characterized by neuronal inclusions containing misfolded, fibrillar alpha-synuclein (aSyn). Prion-like properties of aSyn contribute to the spread of aSyn pathology throughout the nervous system as the disease progresses. Utilizing these properties, seed amplification assays (SAA) enable the detection of aSyn pathology in living patients. We hypothesized that structurally distinct aSyn aggregates, or strains, may underlie the clinical heterogeneity of PD. To test this hypothesis, we recursively amplified aSyn fibrils from the cerebrospinal fluid (CSF) of 54 patients (34 people with PD and 20 controls). These fibrils were then characterized regarding SAA kinetic properties and detergent resistance. In addition, cultured cells were transfected with SAA products, and the extent of seeded aSyn pathology was quantified by staining for phosphorylated aSyn followed by automated high-throughput microscopy and image analysis. We found that fibrils, amplified from CSF by recursive SAA, exhibit two types of distinct biophysical properties and have different seeding capacities in cells. These properties are associated with clinical parameters and may therefore help explain the clinical heterogeneity in PD. Measuring aSyn strains may be relevant for prognosis and for therapies targeting aSyn pathology.

Keywords: Alpha-synuclein; Parkinson’s disease; RT-QuIC; Seed amplification assay; Strains.

Fig. 1

Recursive seed amplification uncovers distinct seeding properties. a Schematic of experimental design: Cerebrospinal fluid (CSF) samples from n = 34 patients with Parkinson’s disease (PD) were used to seed a seed amplification assay (SAA). The SAA product (Amp1) was diluted and used as seed for another round of amplification (Amp2). Following the second amplification, fibrils were purified, adjusted in concentration and used as seeds for a third SAA (Amp3). b SAA aggregation curves of the initial amplification (Amp1) from CSF samples. Each sample was run in four technical replicates. The replicate with the second fastest lag phase is shown. Curves were color-coded based on the kinetic of the last round of amplification (Amp3). c SAA aggregation curves in the second amplification (Amp2) using the diluted SAA products from Amp1. Each curve represents the average fluorescence of four independent replicates. d SAA aggregation curves in the third amplification (Amp3) using the diluted SAA products from Amp2. Each curve represents the average fluorescence of four independent replicates. Patient samples were classified as “fast kinetic” (orange) or “slow kinetic” (blue) for (b-d) based on this third round of amplification (Amp3). e Summary of the kinetic results in Amp3 for each patient as obtained by different experiments. Each panel represents the Amp3 result of the respective patient sample (n = 34). For each panel, the seeds were amplified in independent experiments in a new round of CSF-> Amp1-> Amp2-> Amp3 amplification (= rSAA). For four patients, the CSF from a second, independent lumbar puncture (Re-LP) was available. For samples with white panels, not enough CSF was available to perform five independent experiments

Fig. 2

Kinetic types are stable in additional rounds of amplification. a Schematic overview of the experiments. b Adding additional rounds of amplification (Amp4, Amp5), the kinetic type was unchanged for the analysed samples (P01, P07, P10, P18, P20, P22)

Fig. 3

Kinetic types are stable in multiple amplifications from the same Amp1 product. a Schematic overview of the experiments. The products of the initial seed amplification from CSF (Amp1) were amplified over several rounds of amplifications (Amp2-> Amp3). b Each panel represents the results of one round of amplification (Amp2-> Amp3 (orange = fast kinetic; blue = slow kinetic; white = not done). Each row represents amplifications from the same Amp1 product. c Curves depict Amp3 results of several amplifications of the Amp1(1) product from sample P8 (orange) and sample P15 (blue)

Fig. 4

Kinetic types are unaffected by donor CSF components other than seeds. a Schematic overview of the experiments. In addition, fibrils from Amp1 were separately amplified as described in Fig. 1 to test their kinetic type. CSF (slow) = CSF from patients defined as slow kinetic sample; CSF (fast) = CSF from patients defined as fast kinetic sample; AmpX (slow) = slow kinetic fibrils; AmpX (fast) = fast kinetic fibrils. b Figure shows the third amplification (Amp3) of aggregates from three fast kinetic samples (P01, P07, P10) dissolved in CSF from slow kinetic samples (P18, P20, P22) (light orange, orange and dark orange curves) and vice versa (light blue, blue and dark blue curves). For each sample, 3 different dilutions of aggregates were tested (15 fg/µl = dotted line; 150 fg/µl = dashed line; 15,000 fg/µl = solid line). The kinetic type was not influenced by the CSF under these conditions. c Figure shows the fourth amplification (Amp4), here without CSF, with the same dilution of Amp3 products (1:100,000). The fibrils preserved their initial kinetic type regardless of the CSF or the amount of seeds added in Amp3

Fig. 5

Unseeded aggregation curves in negative controls does not resemble fast or slow kinetic. a Schematic overview of the experiments. b SAA aggregation curves of the initial amplification (Amp1) from CSF. 24 samples were measured on the same plate, each sample was run in four technical replicates. 20 samples were previously measured negative in the SAA (black curves), 2 PD-samples were previously categorized as fast kinetic (orange curves) and 2 PD-samples were previously categorized as slow kinetic (blue curves). The replicate with the second fastest lag phase is shown. c SAA aggregation curves in the second amplification (Amp2) using the diluted SAA products from Amp1. Each curve represents the average fluorescence of four independent replicates. The kinetic type of the 4 PD samples is already distinguishable. Two negative controls show a small increase in the fluorescence signal. d SAA aggregation curves in the third amplification (Amp3) using the diluted SAA products from Amp2. Each curve represents the average fluorescence of four independent replicates. The kinetic type of the 4 PD samples is distinguishable. Three of the negative samples show an increase in the fluorescence signal, two of them already did in Amp2

Fig. 6

Distinguishability of fibril types is preserved despite differing fibril concentrations. a Schematic overview of the experiments. In addition, fibrils from Amp1 were separately amplified as described in Fig. 1 to test their kinetic type. b The third round of amplification (Amp3) was seeded with different amounts of fast (light orange, orange and dark orange curves) and slow kinetic fibrils (light blue, blue and dark blue curves). Each curve represents the average of four independent replicates

Fig. 7

Curve kinetics in fibril mixtures are dominated by fast kinetic fibrils. a Schematic overview of the experiments. In addition, fibrils from Amp1 were separately amplified as described in Fig. 1 to test their kinetic type. b The third round of amplification (Amp3) was seeded with 75 mg of a fibril mixture, containing different proportions of fast and slow kinetic fibrils. Each curve represents the average of four independent replicates. c Figure shows the fourth amplification (Amp4), diluted Amp3 products (1:10,000) were used as seeds. Each curve represents the average of four independent replicates

Fig. 8

Unseeded aSyn pre-formed fibrils show a different kinetic type in rSAA compared to fast and slow kinetic fibrils. a Schematic overview of the experiments. In addition, fibrils from Amp1 were separately amplified as described in Fig. 1 to test their kinetic type. b The third round of amplification (Amp3) was seeded with 150 pg of either fast kinetic fibrils (orange line), slow kinetic fibrils (blue line), pre-formed fibrils (PFF) (black line) or a mixture of either fast or slow kinetic fibrils with PFF (dashed and dotted lines). Each curve represents the average of four independent replicates. PFF show a distinct kinetic curve (slope; maximum fluorescence) compared to fast and slow kinetic fibrils. c Figure shows the fourth amplification (Amp4), diluted Amp3 products (1:5,000) were used as seeds. Each curve represents the average of four independent replicates. The kinetic differences between PFF, fast kinetic and slow kinetic fibrils are still present

Fig. 9

Fast and slow kinetic fibrils differ in their conformational stability. Comparison of fibrils with fast versus slow recursive seed amplification assay (rSAA) aggregation kinetics regarding their resistance to increasing concentrations of the chemical denaturant guanidinium chloride (GdnHCl). a Raw fluorescence values (minus blank control) after incubation with thioflavin T show a lower baseline (0 M GdnHCl) fluorescence of slow kinetic fibrils (P = 0.0014). b Non-linear curve fit of fluorescence values normalized to baseline shows that slow kinetic fibrils have a higher GdnHCl₅₀ value, i.e., the concentration of GdnHCl required to denature 50% of the fibrils (2.31 M vs. 1.54 M, P = 0.0001). Values are mean ± standard deviation (n = 3). P values from two-tailed unpaired t test. RFU = Relative fluorescence units

Fig. 10

Patient-derived fibrils differ in their capacity to induce cellular α-synuclein pathology. a Schematic of experimental design: α-synuclein (aSyn) seeds present in cerebrospinal fluid (CSF) samples from patients with Parkinson’s disease (PD, n = 20) were amplified twice using a seed amplification assay (SAA). Following amplification, fibrils were purified, adjusted in concentration, and sonicated. Successful and comparable fragmentation was verified by dynamic light scattering (DLS). The sonicated fibrils were subsequently transfected into three HEK cell clones stably overexpressing both aSyn and GFP separately (3 wells per patient sample, clone, and experiment). After incubating for 48 h, the cells underwent fixation, staining, and automated imaging. b Representative immunofluorescence images show staining for phosphorylated aSyn (pS129), GFP, and Hoechst of cells treated with aSyn monomer (control) and patient-derived fibrils (fast kinetic = Patient 01, slow kinetic = Patient 21). c Thresholding results of images from (b) after automated analysis using CellProfiler. Regions of interest were obtained for areas positive for pS129 (white) and GFP (green). d pS129 area fraction (percentage of GFP-positive cytosol covered by pS129 signal) after transfection with patient-derived fibrils with fast (n = 8) versus slow (n = 12) rSAA kinetics. Each dot represents the mean of all experiments (n = 3–8, see Supplementary Table 2) of one patient in all three HEK cell clone. Lines represent mean ± standard deviation, P value from two-tailed unpaired t test