PrP is a central player in toxicity mediated by soluble aggregates of neurodegeneration-causing proteins

Abstract

Neurodegenerative diseases are an enormous public health problem, affecting tens of millions of people worldwide. Nearly all of these diseases are characterized by oligomerization and fibrillization of neuronal proteins, and there is great interest in therapeutic targeting of these aggregates. Here, we show that soluble aggregates of α-synuclein and tau bind to plate-immobilized PrP in vitro and on mouse cortical neurons, and that this binding requires at least one of the same N-terminal sites at which soluble Aβ aggregates bind. Moreover, soluble aggregates of tau, α-synuclein and Aβ cause both functional (impairment of LTP) and structural (neuritic dystrophy) compromise and these deficits are absent when PrP is ablated, knocked-down, or when neurons are pre-treated with anti-PrP blocking antibodies. Using an all-human experimental paradigm involving: (1) isogenic iPSC-derived neurons expressing or lacking PRNP, and (2) aqueous extracts from brains of individuals who died with Alzheimer's disease, dementia with Lewy bodies, and Pick's disease, we demonstrate that Aβ, α-synuclein and tau are toxic to neurons in a manner that requires PrPC. These results indicate that PrP is likely to play an important role in a variety of late-life neurodegenerative diseases and that therapeutic targeting of PrP, rather than individual disease proteins, may have more benefit for conditions which involve the aggregation of more than one protein.

Keywords: Alzheimer’s disease; Aβ; Dementia with Lewy bodies; Prion protein; Tau; α-Synuclein.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1

Preparation and characterization of soluble aggregates of Aβ, α-synuclein and tau. a–c Representative chromatograms depicting isolation of monomeric (a) Aβ1–42, (b) α-synuclein (αSyn), and (c) tau. Downward arrows indicate elution of globular molecular weight standards, the shaded region indicates the fraction retained for aggregation experiments and the inset SDS-PAGE/coomassie (CBB) depicts migration of the purified monomers. d–f Freshly SEC-isolated monomer was diluted to 20 µM (Aβ1-42; d) and 35 µM (αSyn; e) and 50 µM (tau; f), combined with Thioflavin-T (ThT) and aggregation monitored until ThT signals plateaued (Tmax). Aβ and αSyn were aggregated without additives while tau was aggregated in the presence of 50 and 100 µM heparin and DTT, respectively. Binding of ThT to Aβ1–42 and αSyn was monitored continuously, whereas aliquots of tau were removed, mixed with ThT and assessed at 3-day intervals. In each case, identical reactions without protein (buffer) served as negative control. Fibrils were harvested at time points indicated by downward arrows, mounted on grids and representative negative-stain electron microscope (EM) micrographs of Aβ, αSyn and tau Tmax fibrils are presented to the right of each graph. g–i Representative negative-stain EM micrographs of immersion sonicated soluble aggregates (Aβ SBAs, g; αSyn SAAs, h; tau STAs, i). Size distribution of SBAs, SAAs and STAs as determined by negative-stain EM are presented to the right of each micrograph. j–l Neat Tmax fibrils (N), the supernatant of centrifuged fibrils (S), resuspended washed fibril pellets (P), and SBAs (j), SAAs (k), and STAs (l) were used for ThT binding. Buffer alone (B) and freshly isolated monomer (M) were included as negative controls. Scale bar in micrographs from d–i = 100 nm. Data in d–f and j–l are the mean ± SD, indicate three technical replicates and are representative of at least three independent experiments. Molecular weight markers (in kDa) are indicated to the right of the inset CBB gels in a, b and c

Fig. 2

Soluble aggregates of Aβ, α-Synuclein and Tau bind to the N-terminus of PrP. a–c Binding of Aβ (a), αSyn (b) and tau (c) monomers, soluble aggregates and fibrils to immobilized PrP23–231 was assessed using an ELISA-like microtiter plate assay. Data shown are the mean ± SD from a single experiment, whereas inset EC50s are from at least four independent experiments. d–f Binding of soluble protein aggregates to immobilized PrP23–231 can be inhibited by anti-PrP monoclonal antibodies (mAbs) to Sites I (MI-0131) and II (ICSM35), but not a nonspecific mAb (46–4). Data shown are the mean ± SD from a single experiment, whereas inset IC50s are from at least four independent experiments. g–i To determine the regions of PrP involved in binding soluble protein aggregates, PrP23–231, PrP91–231, and PrP119–231 were immobilized and binding of SBAs (g), SAAs (h) and STAs (i) measured. Data shown are the mean ± SD from a single experiment, whereas inset EC50s are from at least 4 independent experiments

Fig. 3

SBAs, SAAs and STAs bind to primary neurons in a dose-dependent and saturable manner. a, c, e Soluble protein aggregates were added to primary mouse neurons (MPNs) and binding assessed using serial-permeabilized immunocytochemistry. Representative images SBAs (a), SAAs (c) and STAs (e) are shown. Staining for MAP2, PrP and bound proteins are shown in blue, red and yellow, respectively. b, d, f Relative dose–response binding of SBAs (a), SAAs (c) and STAs (e) to MPNs was quantified using a custom FIJI macro that identified soluble protein aggregate puncta that were at least 50% PrP colocalized. Relative values were determined by normalizing binding signals to those obtained with the highest protein concentration analyzed (3 µM). Inset images depict enlarged triple-colocalization images for neurons treated with 1 μM SPAs (from a, c and e). Scale bar in a, c and e = 50 µm, and data in b, d and f represent three independent experiments with 60 images analyzed per experiment per dose

Fig. 4

SBAs, SAAs and STAs bind to primary neurons in a PrP-dependent manner. a, c and e Soluble protein aggregates were added to wild-type (WT) and PrP-null (Prnp−/−) MPNs and binding assessed using serial-permeabilized immunocytochemistry. Representative images for SBAs (a), SAAs (c) and STAs (e) are shown. Staining for MAP2, PrP and bound proteins are shown in blue, red and yellow, respectively. Enlarged triple-colocalization images, indicated by boxed regions, are presented to the right of each panel. b, d and f, Relative binding of SBAs (b), SAAs (d) and STAs (f) to WT and Prnp−/− MPNs was determined as in Fig. 3. Binding of soluble protein aggregates to Prnp−/− MPNs is expressed as % binding to WT neurons. Scale bar in a, c and e = 50 µm, and data in b, d and f are the mean ± SD and represent three independent experiments with 45–300 images analyzed per experiment per genotype

Fig. 5

SBAs, SAAs and STAs are toxic to primary neurons in a manner requiring PrP. a, c and e, WT MPNs were incubated without (Media), with monomers (1 μM), or with SBAs (a), SAAs (c) and STAs (e) and neurite length measured using live-cell imaging. Each well was imaged for 6 h prior to the addition of samples, and NeuroTrack-identified neurite lengths were used to normalize values obtained across 96 h after sample addition. Each point represents the mean ± SEM of four images taken from three independent wells. b, d, f Plots of normalized neurite length for WT PMNs incubated without (Media), with monomers (1 μM), or with SBAs (b), SAAs (d) and STAs (f). Data are derived from the last 6 h of the traces shown in a, c and e, respectively. Each point represents the mean ±  SEM of 4 images from three wells for each 2 h bin. g, h, i Plots of normalized neurite length for WT and PrP-null (Prnp−/−; gray shading) MPNs incubated without (Media) or with SBAs (g), SAAs (h) and STAs (i). Data are derived from the last 6 h of treatment. Each point represents the mean ± SEM of four images from three wells for each 2 h bin from three independent experiments

Fig. 6

PrP is required for SPA-mediated inhibition of LTP. Field excitatory postsynaptic potential (fEPSP) were recorded simultaneously from sets of four slices using a MED64 Quad II system. a–c Time-course traces show that SBAs (a), SAAs (b) and STAs (c) dose-dependently inhibit hippocampal LTP in wild-type (WT) slices. In a and b, aCSF control (Ctr), black; 50 nM SPAs, grey; 100 nM SPAs, blue; 250 nM SPAs, green; and 500 nM SPAs, red. In c, aCSF control (Ctr), black; 0.5 nM SPAs, gray; 1 nM SPAs, blue; 2 nM SPAs, green; and 10 nM SPAs, red. d–f Time-course traces indicate that SBAs (d; 500 nM) and SAAs (e; 500 nM) and STAs (f; 10 nM) potently inhibit LTP in WT, but not PrP-null (Prnp−/−), slices. aCSF control on WT slices (WT Ctr), black; aCSF control on Prnp−/− (KO Ctr) slices, gray; SPAs on WT slices, red; SPAs on Prnp−/− slices, blue. In a–f, the gray horizontal bar represents the duration of SPA treatment and data represent the mean ± SD of 5–6 animals per group. Statistical analysis of results are provided in Supplementary Figure 5

Fig. 7

SBAs, SAAs and STAs are toxic to induced neurons in a manner requiring PrP. a–c iPSC-derived human neurons (iNs) expressing (CR-C) or lacking PrP (CR-PRNP; gray shading) were incubated without (Media) or with SBAs (a), SAAs (b) and STAs (c) and neurite length measured using live-cell imaging. The CR-PRNP line was generated using CRISPR editing (Supplementary Figure 6). Data were collected and analyzed as in Fig. 5 and represent the mean ± SEM of four images from three wells for each 2 h bin from three independent experiments. d–f To determine whether immunotargeting PrP could protect against neurotoxicity induced by soluble protein aggregates, iNs were treated with or without anti-PrP mAbs to Site I (MI-0131) or II (ICSM35) before the addition of SBAs (d), SAAs (e) and STAs (f) and neurite length measured using live-cell imaging. Data were collected and analyzed as in Fig. 5 and represent the mean ± SEM of four images from three wells for each 2 h bin from three independent experiments

Fig. 8

Soluble extracts from AD, DLB and PiD brains are toxic to neurons in a manner requiring PrP. a Aβx−42 immunoassay quantification of Aβ present in homogenates from AD1 brain after immunodepletion with AW7 (AW7 ID) or preimmune serum (Mock). b iNs were incubated without (Media) or with extracts from brain AD1 ID’ed with AW7 (AW7 ID) or preimmune serum (Mock) and neurite length measured using live-cell imaging. c Aβx−42 immunoassay quantification of Aβ present in homogenates from AD2 brain after immunodepletion with AW7 (AW7 ID) or preimmune serum (Mock). d iNs expressing (CR-Control; circles) or lacking PrP (CR-PRNP; diamonds, grey shading) were incubated without (Media) or with extracts from brain AD2 ID’ed with AW7 or preimmune serum (Mock) and neurite length measured using live-cell imaging. e ELISA quantification of αSyn present in homogenates from DLB1 brain after ID with 2F12 (2F12 ID) or isotype control (Mock). f iNs were incubated without (Media) or with extracts from brain DLB1 ID’ed with 2F12 (2F12 ID) or isotype control (Mock) and neurite length measured using live-cell imaging. g ELISA quantification of αSyn present in homogenates from DLB2 brain after ID with 2F12 (2F12 ID) or isotype control (Mock). h CR-Control and CR-PRNP iNs were incubated without (Media) or with extracts from brain DLB2 ID’ed with 2F12 (2F12 ID) or isotype control (Mock) and neurite length measured using live-cell imaging. i ELISA quantification of tau present in homogenates from PiD1 brain after ID with Tau5 (Tau5 ID) or isotype control (Mock). j iNs were incubated without (Media) or with extracts from brain PiD1 ID’ed with Tau5 (Tau5 ID) or isotype control (Mock) and neurite length measured using live-cell imaging. k ELISA quantification of tau present in homogenates from PiD2 brain after ID with Tau5 (Tau5 ID) or isotype control (Mock). l, CR-Control and CR-PRNP iNs were incubated without (Media) or with extracts from brain PiD2 ID’ed with Tau5 (Tau5 ID) or isotype control (Mock) and neurite length measured using live-cell imaging. In a, c, e, g, i and k, data represent the mean ± SD of three technical replicates and are representative of at least two independent experiments. In b, d, f, h, j and l, data were collected and analyzed as before and represent the mean ± SEM of four images from three wells for the last 6 h of three independent experiments