Small soluble α-synuclein aggregates are the toxic species in Parkinson's disease

Parkinson’s disease (PD) is a prevalent neurodegenerative disorder characterized by dopaminergic neuronal loss and the presence of Lewy bodies enriched in aggregated α-synuclein. α-Synuclein aggregates exist as a heterogeneous mixture of sizes and morphologies, complicating the identification of the specific toxic species in vivo. Prior in vitro studies have implicated small soluble prefibrillar aggregates (oligomers) as key cytotoxic species, though some work suggests mature fibrils can be more toxic under certain conditions. Many prior experiments used acute, nonphysiological doses, stabilizers, or mutants, raising uncertainty about relevance to human brain species. The central question addressed here is how toxicity relates to aggregate size and morphology and whether the toxic in vitro species correspond to soluble aggregates present in human PD brain. The study aims to fractionate α-synuclein aggregates by size without perturbation, characterize their structures, assess toxicity via membrane permeabilization and microglial inflammatory responses, and compare in vitro fractions to soluble aggregates extracted from post-mortem PD and control brains.

Multiple studies report that small, soluble α-synuclein aggregates (often termed oligomers) are more cytotoxic than mature fibrils in vitro, disrupting membranes, impairing synapses, and inducing organelle dysfunction and neuroinflammation. Oligomers exhibit seeding capacity and have been linked to progressive neurodegeneration; however, other studies argue fibrils can be more toxic and sustain neuronal loss. Prior oligomer generation often relied on trapping intermediates, cross-linkers/stabilizers, high concentrations, or mutant α-synuclein, which may not mirror native human brain aggregates. Single-molecule measurements in human substantia nigra reported higher proportions of oligomeric α-synuclein in PD versus controls. Despite these insights, the field lacks consensus on oligomer definitions and a robust comparison of size-dependent toxicity at physiologically relevant conditions and direct comparison to human brain-derived soluble aggregates.

Aggregation and size-fractionation: Wild-type monomeric α-synuclein was expressed and purified from E. coli, diluted to 1 mg/mL in PBS (pH 7.4), ultracentrifuged (350,000 × g, 1 h, 4 °C), and aggregated at 37 °C under shaking (200 rpm). Aliquots collected at 0–48 h were combined and layered onto a discontinuous sucrose density gradient (10, 20, 30, 40, 50% v/v; 400 µL per layer) and centrifuged (113,000 × g, 4 h, 4 °C; SW 60 Ti). Fractions were collected top-to-bottom, filtered (0.2–0.45 µm), dialyzed (3.5 kDa MWCO), aliquoted, and stored at −80 °C.

Structural characterization: Fractions were imaged by TEM (uranyl acetate negative stain) and by TIRF microscopy using Thioflavin T (ThT) to quantify single-aggregate fluorescence intensities. Super-resolution imaging used Thioflavin X (ThX) and aptamer-based AD-PAINT to measure aggregate length and eccentricity in solution. High-resolution atomic force microscopy (AFM), including phase-controlled imaging on functionalized mica, measured aggregate height and cross-sectional diameter with angstrom precision, favoring detection of smaller species. Antibody-based single-molecule pulldown (SiMpull) employed a conformation-specific MJF antibody (aggregated/filamentous α-synuclein) and a sequence-specific SC antibody (C-terminus aa121–125) to probe epitope accessibility across fractions and brain-derived samples; isotype controls assessed specificity.

Toxicity assays: Fractions were normalized to a monomer-equivalent concentration of 500 nM and incubated with BV2 microglial-like cells (24 h). TNF-α release in supernatants was quantified to assess inflammatory potential, normalized to buffer (negative) and LPS (10 ng/mL, positive) controls. Membrane permeabilization was measured via Ca2+ influx into dye-loaded POPC liposomes under TIRF; signals were normalized to ionomycin controls.

Human brain aggregates: Soluble aggregates were extracted from amygdala tissue of three PD cases and three age/sex-matched controls (Cambridge Brain Bank). Tissue (300 mg) was soaked in aCSF with protease inhibitors (1.5 mL, 30 min, 4 °C), centrifuged (2,000 × g, 10 min), supernatant re-centrifuged (14,000 × g, 2 h), and the upper 90% dialyzed for 72 h (2 kDa MWCO) with buffer exchanges to remove cytokines. Extracts were aliquoted and stored at −80 °C. Immunohistochemistry on FFPE amygdala sections confirmed Lewy body pathology in PD but not controls. Brain-derived aggregates were analyzed by AD-PAINT, AFM, and SiMpull (MJF/SC) and tested for BV2 TNF-α responses over 96 h (normalized to total protein by BCA). Statistical analyses included KS tests, t-tests, Mann–Whitney, and matched one-way ANOVA with Dunnett’s correction.

• Sucrose gradient fractionation separated α-synuclein aggregates by size and structure: 20–30% fractions were enriched in smaller, predominantly non-fibrillar species; 40–50% fractions in elongated fibrils. ThT single-aggregate intensity distributions increased with sucrose density (KS p < 0.001 between adjacent fractions), consistent with larger aggregate size.
• Super-resolution length distributions showed increasing lengths with higher sucrose fractions: 20% (~170–263 nm), 30% (~226–241 nm), 40% (~275–334 nm), 50% (~361–410 nm). Statistical testing across replicates indicated differences (e.g., 20% vs 30% p = 0.014; 30% vs 40% p = 0.496; 40% vs 50% p = 0.923).
• AFM revealed increasing height and cross-sectional diameter with sucrose density; 10–20% fractions contained many monomers/small oligomers (≤0.5–0.8 nm height), while 40–50% contained larger mature aggregates (up to ~6 nm height).
• Antibody accessibility differed by size: SC (C-terminus) detected more aggregates in low-density fractions, indicating C-terminus accessibility in smaller species; MJF (filament-specific) detected most in high-density (50%) fractions enriched in fibrils. Isotype controls reduced detected spots ~20-fold, confirming specificity.
• Toxicity correlated inversely with size: BV2 TNF-α release was greatest for the 20% fraction (average length ~190 nm) and low for 40–50% fibrillar fractions; 30% was intermediate. Matched one-way ANOVA with Dunnett’s test: 10% vs 20% p = 0.288; 20% vs 30% p = 0.006; 20% vs 40% p = 0.020; 20% vs 50% p = 0.035.
• Membrane permeabilization (Ca2+ influx) was higher for 10% and 20% fractions than for 30–50%: 10% vs 20% p = 0.899; 20% vs 30% p = 0.001; 20% vs 40% p = 0.142; 20% vs 50% p = 0.008.
• PD vs control brain-derived soluble aggregates: AD-PAINT showed most aggregates <100 nm, with PD having a larger proportion of smaller aggregates (KS p < 0.001). AFM showed significantly smaller median height and diameter in PD vs control (height: PD 4.2 ± 0.05 nm vs control 4.4 ± 0.05 nm, p < 0.001; diameter: PD 20 ± 1 nm vs control 22.1 ± 1 nm, p < 0.001). SiMpull indicated fewer MJF-positive (fibrillar) species in PD relative to controls, consistent with a shift toward smaller, less fibrillar aggregates.
• Brain-derived PD aggregates elicited greater BV2 TNF-α responses over 96 h than control aggregates when normalized to total protein, aligning toxicity with smaller aggregate size and higher particle numbers rather than total protein load.

The study addresses the long-standing question of which α-synuclein aggregate species drive toxicity in PD by separating aggregates by size without cross-linkers or mutants and directly testing their bioactivity. Combining orthogonal biophysical readouts (ThT/ThX single-molecule imaging, TEM, AFM) and epitope accessibility (MJF vs SC antibodies) established that smaller, non-fibrillar aggregates predominate in lower sucrose fractions with accessible C-termini, whereas higher fractions are enriched in β-sheet-rich fibrils. Functional assays demonstrated that smaller aggregates are more potent in membrane permeabilization and inducing microglial TNF-α release than larger fibrils. Importantly, soluble aggregates extracted from PD brains are smaller and less fibrillar than those from controls and are more inflammatory, closely matching the properties of the in vitro small fractions. These findings support a mechanistic model where small, non-fibrillar α-synuclein aggregates are key drivers of neuroinflammation and toxicity in PD. The results reconcile conflicting literature by indicating that while fibrils are abundant and seed aggregation, the smaller soluble species may be the principal effectors of acute membrane disruption and inflammatory signaling, akin to observations in Aβ systems. The non-perturbative fractionation also provides a framework to infer toxicity from size distributions when sample amounts preclude direct bioassays.

Small, soluble, non-fibrillar α-synuclein aggregates (<~200 nm) are the most toxic species in terms of membrane permeabilization and microglial inflammatory activation, while larger fibrillar aggregates are less toxic in these assays. Soluble aggregates extracted from PD amygdala are smaller and less fibrillar than controls and elicit stronger inflammatory responses, linking aggregate size to disease-relevant toxicity in human tissue. Methodologically, a reproducible sucrose density gradient enables non-destructive size fractionation and multi-modal structural/functional characterization. Future work should determine cellular mechanisms that govern aggregate size and structure in vivo, assess region- and stage-specific aggregate distributions, extend analyses to neuronal toxicity and synaptic function at physiological concentrations, and evaluate therapeutic strategies that target formation or clearance of small soluble aggregates.

• Sample size of human tissue was limited (n = 3 PD and n = 3 controls), potentially affecting generalizability.
• Imaging modalities have complementary biases: AFM favors detection of smaller species and under-samples long aggregates on mica; ThT/ThX detect β-sheet-rich species and may miss non-β-sheet oligomers; thus, absolute size values differ across methods.
• Membrane permeabilization assay lacked sensitivity for brain-derived samples, likely due to lower aggregate abundance compared to recombinant preparations.
• Complexity of brain extracts precluded resolving monomers by AFM; aggregate heterogeneity remains high.
• BV2 microglial cell line may not fully recapitulate primary human microglial responses; assays were performed at set concentrations (e.g., 500 nM monomer-equivalents), which may differ from in vivo levels.
• Some author-affiliation details and certain methodological parameters for all antibodies/strains are not fully detailed in the excerpt; preparation steps, filtration, and dialysis may alter aggregate distributions despite efforts to be non-perturbative.