Reversible Conformational Conversion of α-Synuclein into Toxic Assemblies by Glucosylceramide

Abstract

α-Synuclein (α-syn) aggregation is a key event in Parkinson's disease (PD). Mutations in glycosphingolipid (GSL)-degrading glucocerebrosidase are risk factors for PD, indicating that disrupted GSL clearance plays a key role in α-syn aggregation. However, the mechanisms of GSL-induced aggregation are not completely understood. We document the presence of physiological α-syn conformers in human midbrain dopamine neurons and tested their contribution to the aggregation process. Pathological α-syn assembly mainly occurred through the conversion of high molecular weight (HMW) physiological α-syn conformers into compact, assembly-state intermediates by glucosylceramide (GluCer), without apparent disassembly into free monomers. This process was reversible in vitro through GluCer depletion. Reducing GSLs in PD patient neurons with and without GBA1 mutations diminished pathology and restored physiological α-syn conformers that associated with synapses. Our work indicates that GSLs control the toxic conversion of physiological α-syn conformers in a reversible manner that is amenable to therapeutic intervention by GSL reducing agents.

Keywords: GBA1; Gaucher disease; Lewy body; Parkinson’s disease; glucocerebrosidase; iPS-derived dopaminergic neurons; lysosomal storage disease; protein aggregation; α-synuclein.

Copyright © 2017 Elsevier Inc. All rights reserved.

Figures

Figure 1. GSL Accumulation Induces Pathological α-Syn Aggregates in Human Midbrain iPSns

(A) iPSC-derived midbrain dopamine neurons from healthy control lines were treated for 2 or 7 days with GCase inhibitors CBE or IFG (50 μM) and phosphate buffered saline (PBS) as a control. Neurons were subjected to sequential extraction followed by western blot using α-Syn antibodies C-20 and syn303. β-iii-tubulin (tub) and GAPDH were used as loading controls. Right: quantification of the blots (n = 4). (B) Immunofluorescence analysis of α-Syn- (red) or thioflavin S- (Thio S, green) positive neurons as diffuse (i) or punctated inclusions (ii) after a 7 day treatment with 50 μM CBE. Nuclei were detected with DAPI (blue). The arrow indicates a representative amyloidogenic inclusion. Scale bars, 5 μm. (C) Left: electron microscopy (EM) analysis of WT control iPSns showing normal morphology of vesicles in the cytoplasm (i and ii) and a close-up of membrane delimited electron lucent vesicles (iii). Scale bars, 1 (i), 0.5 (ii), and 0.25 μm (iii). N, nucleus. Right: EM analysis of 7 day CBE-treated WT iPSns showing accumulation of vesicles and multilamellar membrane inclusions (i, white arrow) or fibrillogranular material accumulating in enlarged vacuoles. (ii) shows a close-up of the fibrillar material indicated with the black arrow in (i). Scale bars, 200 (i) and 25 nm (ii). (D) Tubular (i) or fibrillar (ii and iii) accumulations indicated by black arrows within enlarged vacuoles of 7 day CBE-treated iPSns. Scale bars, 100 (i and iii) and 200 nm (ii). (E) Soluble lysates were analyzed by native SEC (inputs shown in A). HMW (#1–4) and monomeric fractions (#5–10) were analyzed by western blot. Molecular radius is shown horizontally in angstroms (Å). (F) Quantification of western blot in (E) (n = 5). (G) Measurement of α-Syn from CBE-treated iPSns by SEC/ELISA (soluble monomers or HMW species) or sequential extraction (total soluble or insoluble) (n = 4).For all quantifications, values are the mean ± SEM, *p

Figure 2. HMW Physiological Conformers Are Structurally Distinct from GSL-Induced Conformers

(A) Isolated SEC fractions from PBS control (HMW-ctrl) or CBE-treated (HMW-GSL) iPSns were analyzed by immuno-FRET. α-Syn KO lysate was used as a negative control, whereas pre-formed fibrils (PFFs) were used as a positive control. HMW GD species were obtained from L444P/L444P patient neurons and compared to a distinct control line. Direct excitation of the acceptor antibody (647) was used to quantify total α-Syn (n = 5). (B) H4 cells were treated with CBE for 2 days, fixed, and analyzed by immuno-FRET (ex = 488 nm, em = 660 nm; white). Direct excitation of donor (green) or acceptor (red) and dapi (blue) are shown as controls. Scale bar, 10 μm. (C) Quantification of (B) (n = 3). (D) iPSns were treated with CBE for indicated times and FRET efficiency was measured as in (B) and (C) (n = 4). (E) HMW conformers were isolated from either PBS- or CBE-treated H4 cells by SEC, digested with PK, and analyzed by western blot. Quantification is shown below (C-20). α-Tubulin (α-Tub) and Coomassie brilliant blue (CBB) were used as loading controls (n = 6). M, marker lane. (F) HMW conformers isolated from soluble brain lysates of WT or Gaucher mice (4L/PS-NA) were analyzed as in(E). Quantification is shown below (n=3 mice). For all quantifications, values are the mean ± SEM, *p

Figure 3. Characterization of the Structural Differences of HMW Physiological and GSL-Induced Species

(A) Soluble HMW fractions (#1–3) containing α-Syn were isolated from iPSn cultures treated with either PBS control (HMW-ctrl) or CBE (HMW-GSL). Equivalent amounts of HMW species from each condition were incubated with 0.1% SDS in vitro, and breakdown of the HMW species was determined by SEC/ELISA (n = 3). (B) HMW-ctrl or HMW-GSL α-Syn were dot-blotted under native conditions (N) followed by western blot. The same blots were denatured (D) by boiling for 10 min in SDS, then re-probed with the same antibodies to determine exposed or buried regions. α-Syn KO lysate was used to determine antibody specificity. Fully denatured recombinant monomers were used as a control, showing no difference between N and D conditions. Neuron-specific enolase (NSE) was used as a loading control. Right: quantification of changes in antibody reactivity after denaturation (n = 3). (C) Native dot blot/western analysis of HMW species using conformational specific antibodies (n = 3). (D) Equivalent amounts of HMW-ctrl and HMW-GSL isolated from H4 cells were analyzed by western blot. C-20 immunoreactivity was used to measure total α-Syn found in each HMW fraction. Recombinant (recomb) purified α-Syn was used as a standard, and α-Syn KO lysate was used to determine antibody specificity. Asterisks indicate non-specific bands. Below: quantification of western blots. (E) Native dot blot/western analysis of HMW α-Syn derived from GD patient neurons (L444P/L444P) compared to controls using oligomer-specific antibody A11 (n = 3). Syn211 and α-Tub were used as loading controls. (F) Diagram summarizing the buried and exposed regions of HMW-ctrl and HMW-GSL. N.T., not tested. (G) HMW and LMW (monomer [mon.]) species were isolated by SEC and mixed with purified recombinant α-Syn monomers in vitro at pH 7.4 under quiescent conditions, and amyloid formation was determined by thioflavin T (ThT) fluorescence. RFU, relative fluorescence units. PBS, bovine serum albumin (BSA), α-SynKO lysate, or HMW GSL-induced species immunodepleted (I.D.) of α-Syn were used as negative controls (n = 4). Pre-formed seed fibrils were used as a positive control (*p < 0.05 seeds and HMW-GSL versus all control conditions, **p < 0.01 HMW-GSL versus BSA and GSL mon., ***p < 0.001 HMW-GSL versus PBS control). (H) Purified recombinant monomers were mixed with BSA as a control, GSL monomers, or HMW-GSL and aggregation was assessed at 68 hr by sedimentation/ western blot. Supe, supernatant (n = 4, *p

Figure 4. GluCer Converts Physiological α-Syn Conformers into Pathogenic Species

(A) iPSn-derived monomers isolated from control neurons were incubated with phosphatidylcholine (PC) as a control, and GluCer under quiescent conditions at pH 5.5, followed by SEC/ELISA to monitor conversion into HMW species (n = 4). (B) Physiological conformers (HMW-ctrl or mon.-ctrl) were isolated from iPSns and mixed with PC or GluCer in vitro as in (A), followed by immuno-FRET analysis (n = 4, *p < 0.05 compared to time = 0). (C) HMW-ctrl was isolated and mixed with PC or GluCer in vitro at pH 5.0, 6.0, or 7.4 for 30 min at 37°C followed by SEC/ELISA. SDS (0.1%) was used as a control to disrupt HMW conformers (n = 4). (D) HMW-ctrl fractions from iPSns were treated with PC or GluCer as in (B) for 30 min, followed by PK digest/western blot (C-20)(n = 3). CBB was used as a loading control. (E) HMW-ctrl from H4 cells was mixed with PC or GluCer in vitro and analyzed as in (D) (n = 3). (F) HMW-ctrl from iPSns was incubated with BSA, PC, or GluCer in vitro as in (D), then mixed with purified recombinant monomers. Amyloid formation was monitored over time under quiescent conditions at pH 7.4 by ThT (n = 4). For all quantifications, values are the mean ± SEM (*p < 0.05, ***p < 0.0002,****p < 0.0001).

Figure 5. GSLs Induce Toxicity through α-Syn

(A) iPSns were infected with lentiviral particles that express either scrambled (scr) or shRNA to knock down α-Syn (7 days, MOI = 5). Western blot shows knockdown efficiency. β-iii-tubulin and GAPDH were used as loading controls. Quantification is shown to the right (n = 3). (B) Neuron viability was determined by neurofilament quantification in iPSns treated with PBS or CBE (n = 4). (C) Western blot of α-Syn demonstrating an efficient KO in three replicates. β-iii-tubulin (βiii tub) and GAPDH were used as loading controls. (D) Neuron viability was determined as in (B) from ctrl of α-Syn KO iPSns treated with 50 μM CBE for 7 days (n = 4). (E) HMW-ctrl or HMW-GSL was isolated and quantified against recombinant α-Syn standards. Equal amounts were applied into the culture media of control iPSns and neurotoxicity was determined by neurofilament and cell volume quantification (n = 4). (F) Recombinant α-Syn was incubated with GluCer to generate HMW-GSL species in vitro, then applied to H4 cultures. Cultures were treated with α-Syn monomers alone, α-Syn + PC, or PC/GluCer lipids without α-Syn as controls, and viability was assessed as in (E). Dynasore, endocytosis inhibitor. (G) H4 cells were incubated using the same conditions described in (F), and cellular toxicity was assessed through release of lactate dehydrogenase (LDH) into the culture media (n = 4). Values are expressed as fold change compared to PC + α-Syn for (F) and (G). For all quantifications, values are the mean ± SEM (*p

Figure 6. Reduction of Intracellular GSLs Diminishes Pathological α-Syn

(A–C) Measurement of soluble and insoluble α-Syn by sequential extraction/western blot of GD iPSns (N370S/c.84dupG) (A), H4 α-Syn cells (B), or SNCA triplication (trp) iPSns (C), with pre-existing pathology, treated with a glucosylceramide synthesis inhibitor (GCSi,50 nM). DMSO was used as a control. Quantification is shown below (n = 4). β-iii-tubulin, GAPDH, or CBB was used as loading control. (D) Analysis of amyloidogenic α-Syn aggregates in cell bodies (top) or neurites (bottom) of fixed patient iPSns treated with GCSi by immunostaining (LB509, red) and thioflavin S (ThioS, green). Nuclei were detected with DAPI (blue). Scale bars, 10 μm. Right: quantification of cells containing ThioS-positive α-Syn in cell bodies (n = 4). (E–G) SEC/western blot of GD patient iPSns (N370S/c.84dupG) (E), H4 α-Syn cells (F), and SNCA trp patient iPS midbrain neurons (G). Quantification is shown below (n = 3). For all quantifications, values are the mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 7. Restoration of Physiological α-Syn Conformers by GSL Reduction in Patient-Derived iPSns

(A) HMW-ctrl was converted into pathogenic species by GluCer, followed by the removal of GluCer with anti-GluCer antibodies (time = 0). Samples were then incubated for 30 min under physiological conditions and structural conversion was analyzed by FRET. Total α-Syn was measured by direct excitation of the acceptor antibody (n = 5, **p SNCA trp patient iPSns and analyzed by PK digestion/western blot. Quantification is shown below. CBB was used as a loading control (n = 3, *p < 0.05 DMSO compared to GCSi). (D) Immunofluorescence analysis of GD and SNCA trp iPSns showing localization of α-Syn (red) and the synaptic marker synapsin (green) within neurites. Scale bar, 5 μm. (E) Quantification of α-Syn/synapsin co-localization (Pearson's coefficient) (n = 3, *p

Figure 8. Schematic Diagram Summarizing the Effects of GSLs on α-Syn Aggregation

In healthy human midbrain neurons, α-Syn exists as both physiological monomers and HMW conformers. GSLs such as GluCer preferentially act on HMW conformers, and to a minor extent on monomers, to convert them into assembly-state oligomers (AOs) with molecular radius that is identical to HMW physiological conformers. HMW physiological species are non-toxic, loosely self-associated species and can be directly converted into AOs. AOs are cytotoxic, compact, and PK resistant, and can accelerate the formation of assembled amyloid fibrils from free monomers.