Crystal Violet Selectively Detects Aβ Oligomers but Not Fibrils In Vitro and in Alzheimer's Disease Brain Tissue

Abstract

Deposition of extracellular Amyloid Beta (Aβ) and intracellular tau fibrils in post-mortem brains remains the only way to conclusively confirm cases of Alzheimer's Disease (AD). Substantial evidence, though, implicates small globular oligomers instead of fibrils as relevant biomarkers of, and critical contributors to, the clinical symptoms of AD. Efforts to verify and utilize amyloid oligomers as AD biomarkers in vivo have been limited by the near-exclusive dependence on conformation-selective antibodies for oligomer detection. While antibodies have yielded critical evidence for the role of both Aβ and tau oligomers in AD, they are not suitable for imaging amyloid oligomers in vivo. Therefore, it would be desirable to identify a set of oligomer-selective small molecules for subsequent development into Positron Emission Tomography (PET) probes. Using a kinetics-based screening assay, we confirm that the triarylmethane dye Crystal Violet (CV) is oligomer-selective for Aβ42 oligomers (AβOs) grown under near-physiological solution conditions in vitro. In postmortem brains of an AD mouse model and human AD patients, we demonstrate that A11 antibody-positive oligomers but not Thioflavin S (ThioS)-positive fibrils colocalize with CV staining, confirming in vitro results. Therefore, our kinetic screen represents a robust approach for identifying new classes of small molecules as candidates for oligomer-selective dyes (OSDs). Such OSDs, in turn, provide promising starting points for the development of PET probes for pre-mortem imaging of oligomer deposits in humans.

Keywords: Alzheimer’s Disease; Crystal Violet; Thioflavin T; amyloid beta; amyloid oligomers; fluorescence; tissue staining.

Figure 1

Schematic of kinetic dye assay to identify potential oligomer-selective dyes (A,B). Schematic of the transition in ThT kinetics from (A) fibril-dominated sigmoidal kinetics with a well-defined lag phase to (B) biphasic ThT kinetics with a lag-free, oligomer-dominated phase. The purple curve indicates that an Oligomer-Selective Dye (OSD) should display (A) a weak response during fibril-dominated sigmoidal growth but (B) a robust response to the oligomer-dominated phase of biphasic growth with only a weak or no response to the fibril-related secondary upswing. (C) Chemical structures of ThT and CV dye.

Figure 2

Thioflavin T vs. Crystal Violet kinetics during Aβ42 aggregation at pH 7.4. (A) Thioflavin T (ThT) vs. (B) Crystal Violet (CV) fluorescence responses during growth of Aβ42 at the indicated concentrations (in μM) at pH 7.4, 150 mM NaCl at 27 °C. Shown are three dye traces collected for each protein concentration. Thioflavin T displays the transition from fibril-dominated sigmoidal kinetics (orange traces) at 2 μM to progressively more oligomer-induced biphasic kinetics (blue traces), highlighted by the use of semi-log plots. Crystal Violet responds to the oligomer-dominated phase of biphasic ThT kinetics but lacks responses to the delayed secondary upswing due to accelerating fibril kinetics. (C,D) Overlay of the fractional changes in ThT vs. CV fluorescence during (C) fibril-dominated sigmoidal growth at 2 μM and (D) prominent biphasic growth at 30 μM. CV shows no discernible response to fibril growth (2 μM) but perfectly tracks the initial oligomer phase of biphasic ThT kinetics, which persists deep into the fibril-dominated secondary phase. We have previously shown that biphasic ThT kinetics is due to the superposition of lag-free and rapidly increasing formation of off-pathway oligomers with sigmoidal fibril kinetics [35,36] The persistent CV responses deep into the fibril growth phase match with the continued presence of AβOs observed with TEM and oligomer-selective antibodies (see Figure 3).

Figure 2

Thioflavin T vs. Crystal Violet kinetics during Aβ42 aggregation at pH 7.4. (A) Thioflavin T (ThT) vs. (B) Crystal Violet (CV) fluorescence responses during growth of Aβ42 at the indicated concentrations (in μM) at pH 7.4, 150 mM NaCl at 27 °C. Shown are three dye traces collected for each protein concentration. Thioflavin T displays the transition from fibril-dominated sigmoidal kinetics (orange traces) at 2 μM to progressively more oligomer-induced biphasic kinetics (blue traces), highlighted by the use of semi-log plots. Crystal Violet responds to the oligomer-dominated phase of biphasic ThT kinetics but lacks responses to the delayed secondary upswing due to accelerating fibril kinetics. (C,D) Overlay of the fractional changes in ThT vs. CV fluorescence during (C) fibril-dominated sigmoidal growth at 2 μM and (D) prominent biphasic growth at 30 μM. CV shows no discernible response to fibril growth (2 μM) but perfectly tracks the initial oligomer phase of biphasic ThT kinetics, which persists deep into the fibril-dominated secondary phase. We have previously shown that biphasic ThT kinetics is due to the superposition of lag-free and rapidly increasing formation of off-pathway oligomers with sigmoidal fibril kinetics [35,36] The persistent CV responses deep into the fibril growth phase match with the continued presence of AβOs observed with TEM and oligomer-selective antibodies (see Figure 3).

Figure 3

Aggregate morphologies in the sigmoidal vs. biphasic pathway. (A) Representative sigmoidal (orange, 2 μM) and biphasic (blue, 30 μM) ThT kinetics of Aβ42, with arrows indicating typical early stage (~3 h) and endstage (36+ h) time points for aliquots collected for AFM and/or TEM imaging. (B,C) Sigmoidal pathway (2 μM): AFM images during the lag phase (B;~3 h) yielded no discernible protein aggregates (inset: typical height profile at line indicated in image). End-stage TEM images (C) displayed mixtures of isolated and bundled fibrils. (D–F) Biphasic pathway (5–50 μM): AFM (D) and TEM (E) images of oligomeric/protofibrillar aggregates observed during the initial phase (≤3 h) of biphasic kinetics (inset: height profile at the scan line indicated in the AFM image). The images show the correspondence of the beads-on-a-string morphology for oligomers/protofibrils in AFM and TEM imaging modalities. Late-stage TEM images (>36 h) (F) yielded mixtures of darkly stained oligomeric/protofibrillar aggregates (dark arrow) together with lightly stained and increasingly clumped fibrils (white arrow).

Figure 4

CV kinetics correlates with immunostaining against Aβ42 oligomers. The persistence and amplitude of the CV fluorescence (orange: no oligomers; blue: oligomer response; data from Figure 2B) correlates with immunostaining with oligomer-selective antibody A11 at the end of the incubation period (right side). Note that Aβ42 aliquots were all diluted to 2 μM concentration prior to blotting. In addition, aliquots from a 60 μM sample, sampled at 0 h, 3 h, and 20 h and stained with A11 (bottom panel), match with the build-up of oligomer populations well past the lag phase observed with ThT under those conditions (Figure 2A).

Figure 5

CV selectively detects AβOs over AβFs in APP/PS1 mouse and human AD brains. (A) Representative images of 8-month-old APP/PS1 and WT littermate mouse brains stained with Aβ antibody (D54D2, Cell Signaling) (green) and CV (red). White boxes magnified to the right. White arrows: examples of Aβ antibody-stained plaques with major CV staining; red arrows: examples of Aβ antibody-stained plaques with little CV staining; yellow arrows: examples of CV-stained plaques with little Aβ antibody immunoreactivity. (B) Representative images of 8-month-old APP/PS1 mouse brains stained with Thioflavin S (ThioS, green) and CV (red). (C) A representative image of an APP/PS1 mouse brain section showing clear juxtaposition of CV staining to ThioS+ fibrils but without colocalization (white arrows). (D) Z-stacked image of a ThioS+ plaque lacking CV staining. (E) Representative images of 8-month-old APP/PS1 mouse brain sections stained with oligomer-selective A11 antibody, CV, and ThioS. (F) Quantification of ThioS and A11 colocalization with CV (left graph) and CV colocalization with ThioS and A11 (right graph) (t-test, **** p < 0.0001, n = 25 cortical images/group from 2 APP/PS1 mice). (G) Representative images of ThioS/CV and A11/CV staining from the frontal cortex of five AD patient brains. (H) Quantification of ThioS and A11 colocalization with CV in AD brains (t-test, **** p < 0.0001, n = 26–29 images/group from five AD patient brains).