Elevated Aβ aggregates in feces from Alzheimer's disease patients: a proof-of-concept study

Abstract

Background: Misfolding and aggregation of amyloid β (Aβ), along with neurofibrillary tangles consisting of aggregated Tau species, are pathological hallmarks of Alzheimer's disease (AD) onset and progression. In this study, we hypothesized the clearance of Aβ aggregates from the brain and body into the gut.

Methods: To investigate this, we used surface-based fluorescence intensity distribution analysis (sFIDA) to determine the Aβ aggregate concentrations in feces from 26 AD patients and 31 healthy controls (HC).

Results: Aβ aggregates were detectable in human feces and their concentrations were elevated in AD patients compared to HC (specificity 90.3%, sensitivity 53.8%).

Conclusion: Thus, fecal Aβ aggregates constitute a non-invasive biomarker candidate for diagnosing AD. Whether digestion-resistant Aβ aggregates in feces are secreted via the liver and bile or directly from the enteric neuronal system remains to be elucidated.

Keywords: Amyloidosis; Aβ oligomer quantitation; Brain-gut-microbiota axis; Clearance; Fecal/stool samples; Leaky gut; sFIDA.

Fig. 1

Scheme of sFIDA principle. In sFIDA, capture antibodies directed against a linear epitope on Aβ (Nab228, directed against epitope amino acids 1 − 11) are immobilized on a glass surface, and unoccupied surface area is blocked with bovine serum albumin to reduce unspecific binding events. During sample incubation, monomeric and aggregated Aβ species are bound to the capture antibody. (A) Because sFIDA uses the same or overlapping epitopes for capture and detection, only Aβ aggregates are subsequently detected with fluorescence-labeled antibodies IC16-CF633, which is overlapping with the epitope of the capture antibody (directed against epitope amino acids 2 − 8). (B) For monomeric Aβ, this epitope is already masked by the capture antibody and cannot be bound by the detection antibody. Afterward, the assay surface is imaged by fluorescence microscopy, and pixels above a defined cutoff threshold are counted by image-data analysis (called pixel count). Finally, pixel-based readouts are calibrated into molar particle concentrations using silica nanoparticles (SiNaPs) coated with Aβ as calibration standards. Created with BioRender.com

Fig. 2

Schematic illustration of the sample collection process. Study participants were divided into two groups (normal cognition, impaired cognition) based in their clinical symptoms. CSF analyses or imaging were performed for the participants with impaired cognition, to confirm AD related pathology. In total, 31 HC and 26 participants with impaired cognition due to amyloid pathology were included into this study. Created with BioRender.com

Fig. 3

Percent signal reduction of different assay controls for the assessment of assay selectivity. A) Aβ-coated SiNaPs, IQC, and Aβ aggregates in three fecal samples were analyzed by sFIDA. Based on the observed pixel counts, the percent signal reduction of each assay control (no capture antibody: capture control, dark red; no detection probe: autofluorescence control, light red; detection probe against α-synuclein: cross-reactivity control, rose) in comparison to reference (standard assay setup) were calculated. (B) All samples were subjected to immunodepletion using magnetic beads linked to Aβ-specific antibody (Nab228) and control beads linked to α-synuclein-specific antibody (211). Based on the observed pixel counts of non-depleted (dark gray), 211-depleted (light gray), and Nab228-depleted (red) samples, percent signal reduction was calculated

Fig. 4

Independent measurements of aggregated Aβ yield high comparability. For Aβ-coated SiNaPs (dark grey), IQC sample (light gray), and 13 fecal samples (red), pixel counts of the second measurement were plotted against pixel counts of the first measurement. Because the second measurement was carried out months later, there were minor changes in used reagent lots, e.g., manufacturing date of washing buffers, and the used homogenized fecal samples were subjected to an additional freeze-thaw cycle. Please note the logarithmic scaling

Fig. 5

Representative TIRFM images, molar particle concentration of fecal Aβ aggregates and receiver operating characteristic. (A) Representative TIRFM images for the red channel (IC16-CF633, excitation 635 nm, emission 705 nm, exposure time 1000 ms, gain 1000) of 1 pM Aβ-coated SiNaPs, 100 pM synthetic Aβ1 − 42 oligomers (based on total Aβ concentration), fecal sample (AD patient) and sample buffer control. For better illustration of 14-bit images, color and contrast were adjusted using ImageJ software (colormap: red hot, contrast: maximum grayscale value 8000). Scale bar: 20 μm. (B) Concentrations of fecal Aβ aggregates of AD patients were significantly elevated with a p-value of 0.009 compared to HC. Significant differences between both cohorts were calculated with a Mann − Whitney U test (**p: ≤ 0.01). Please note the logarithmic scaling (line = median, square = mean). (C) In receiver operating characteristic (ROC) analysis, discrimination of AD patients versus HC showed a specificity of 90.3% and a sensitivity of 53.8% with an AUC of 0.703

Fig. 6

Association of fecal Aβ aggregates with additional biomarkers affecting AD pathology. The combination of fecal Aβ aggregates with further biomarkers may provide new insights into mechanism of brain-gut-microbiota axis and AD pathogenesis. Therefore, in addition to CSF biomarkers (Aβ40, Aβ42, phosphorylated and total Tau) and Bristol scale, 16 S rRNA profiles, fecal calprotectin, short-chain fatty acids, secondary bile acids, liver biomarkers, and lipopolysaccharides should be determined in the future. Because amyloids produced by gut microbiome share similarities in tertiary structure with CNS amyloids, they may act in a prion-like manner and induce misfolding, aggregation, and deposition of Aβ and may cross-seed with neuronal amyloids once they have entered the brain due to increased permeability of the blood-brain barrier. Created with BioRender.com