Autocatalytic amplification of Alzheimer-associated Aβ42 peptide aggregation in human cerebrospinal fluid

Abstract

Alzheimer's disease is linked to amyloid β (Aβ) peptide aggregation in the brain, and a detailed understanding of the molecular mechanism of Aβ aggregation may lead to improved diagnostics and therapeutics. While previous studies have been performed in pure buffer, we approach the mechanism in vivo using cerebrospinal fluid (CSF). We investigated the aggregation mechanism of Aβ42 in human CSF through kinetic experiments at several Aβ42 monomer concentrations (0.8-10 µM). The data were subjected to global kinetic analysis and found consistent with an aggregation mechanism involving secondary nucleation of monomers on the fibril surface. A mechanism only including primary nucleation was ruled out. We find that the aggregation process is composed of the same microscopic steps in CSF as in pure buffer, but the rate constant of secondary nucleation is decreased. Most importantly, the autocatalytic amplification of aggregate number through catalysis on the fibril surface is prevalent also in CSF.

Keywords: Intrinsically disordered proteins; Peptides.

Fig. 1

Aggregation kinetics in buffer. a The ThT fluorescence intensity as a function of time (h) for 0.8–10 µM Aβ42 in buffer. b The calculated t_1/2 (h) as a function of Aβ42 concentration, with a fitted power function as in Eq. (1). The error bars are the standard deviations from the n = 4 repeats per Aβ42 concentration. c Shows the double log representation of (b). The fluorescence data underlying each time trace as well as the t_1/2 values and the associated standard deviations can be found in Supplementary Data 1

Fig. 2

Aggregation kinetics in constant Aβ42 concentration. a The ThT fluorescence as a function of time for 3 µM Aβ42 in buffer with various concentrations of CSF (0–80%). The hexagonal figure represents the CSF pool used. b The t_1/2 as estimated from the data in panel (a) as a function of CSF concentration. The error bars are the standard deviations from the n = 3 repeats per CSF concentration. The fluorescence data underlying each time trace as well as the t_1/2 values and the associated standard deviations can be found in Supplementary Data 2

Fig. 3

Cryo-EM images. Cryo-EM images of the final aggregates formed in solutions with a total monomer concentration of 10 µM and a 0%, b 15%, or c 66% CSF. The hexagonal figure represents the CSF pool used. Scale bar = 200 nm

Fig. 4

Aggregation kinetics in CSF. The ThT fluorescence as a function of time (h) for 0.8–10 µM Aβ42 in buffer with a 15% CSF; b 32% CSF; c 66% CSF. The hexagonal figures represent the CSF pool used. d is the calculated t_1/2 (h) as a function of Aβ42 concentrations for all CSF concentrations; it is also plotted with a fitted power function, Eq. (1). e is the double log representation of (d). The error bars are the standard deviations from the n = 4 repeats per Aβ42 concentration. The fluorescence data underlying each time trace as well as the t_1/2 values and the associated standard deviations can be found in Supplementary Data 3

Fig. 5

The double nucleation mechanism. The double nucleation mechanism discovered for Aβ42 in a phosphate buffer. Primary nucleation is a reaction starting from monomers in solution and in the kinetic description used here leads directly to growth competent species. The rate at which this process yields new fibrils, [P], is given by k_n[m]^nc, where [m] is the monomer concentration, [P] is the fibril concentration in fibril units and n_c is the reaction order, which constitutes a lower bound for the number of monomers in the nucleus. Secondary nucleation is a reaction whereby monomers from solution react on the fibril surface and in the kinetic description used here leads directly to growth competent species in solution. The rate at which this process yields new fibrils, [P], is given by k₂[M][m]ⁿ², where [m] is the monomer concentration, [M] is the fibril concentration in monomer units, and n₂ is the reaction order, which constitutes a lower bound for the number of monomers in the nucleus. Elongation is a reaction whereby monomers add to the fibril ends to extend the fibril by one monomer unit. The rate at which this process yields new fibril mass, [M], is given by 2k₊[P][m], where [m] is the monomer concentration

Fig. 6

Normalized kinetics with fitted models. Normalized ThT fluorescence (relative aggregate concentration) as a function of time (h) for 0.8–10 µM Aβ42 in buffer with a, b 0% CSF, c, d 15% CSF, e, f 32% CSF, and g, h 66% CSF, together with fits obtained using AmyloFit. Graphs (b, d, f, h) show the fit using a multi-step secondary nucleation dominated mechanism, while (a, c, e, g) show the same data fitted with a model including primary nucleation and elongation only. The legend at the bottom shows the color coding of the concentrations of Aβ42 in µM. The normalized fluorescence data as well as the two fits evaluated at each time point can be found in Supplementary Data 4

Fig. 7

Seeded data. Normalized fluorescence intensity from seeded aggregation experiments with 3 µM Aβ42 monomer concentration in a buffer and c 66% CSF and addition of 0–30% seeds. The hexagonal figure represents the CSF pool used. b and d are the estimated values of t_1/2 (h) as a function of concentration of the added seed in %. The error bars are the standard deviations from the n = 3 repeats per seed concentration. The fluorescence data underlying each time trace as well as the t_1/2 values and the associated standard deviations can be found in Supplementary Data 5