Background
Amyloid-β (Aβ) is a major factor in the pathogenesis of Alzheimer’s disease (AD) [
1‐
3]. Aβ may be involved in the cognitive impairment of memory that leads to AD, particularly in the form of aggregated 42-amino acid isoform of the Aβ peptide (Aβ
42), which is a major neurotoxic species among Aβ isomers, including Aβ
40, Aβ
42, and other truncated forms of Aβ [
4‐
10]. Since the initial identification of these Aβ isomers in bodily fluids, the measurement of Aβ levels in the blood and cerebrospinal fluid (CSF) has been a research priority [
11,
12]. Three biomarkers in CSF, specifically Aβ
42, total tau, and hyperphosphorylated tau, are widely accepted as AD determinants on the basis of their close correlations with AD pathology [
13,
14]. Diagnostic imaging using positron emission tomography (PET) is the preferred method of investigating pathological and functional changes in the brain of patient with AD.
11C- or
18F-labeled Pittsburgh compound B-positron emission tomography (PiB-PET), which involves binding to amyloid depositions, is particularly beneficial for understanding the underlying processes of AD. Therefore, this type of imaging is used both in research and in the clinic to differentiate patients with AD from control subjects and individuals with other types of dementia [
15‐
17].
Several noninvasive diagnostics for AD, based on diverse biomarkers in the saliva, urine, and blood, have been reported and are still in the research phase of development [
18‐
21]. Mapstone et al. [
22] identified a panel of ten plasma phospholipids as potential diagnostic biomarkers of AD, which included lysophosphatidylcholine, phosphatidylcholine metabolites, and acylcarnitine metabolites. This plasma lipid panel predicted AD conversion, suggesting 90% sensitivity and 85% specificity in differentiating an at-risk group from the cognitively intact group. Despite encouraging results, low positive predictive values limited the clinical usefulness of this panel as a screening tool in subjects aged 70–80 years or younger. In another study, significant differences in soluble CD40 (sCD40) and sCD40 ligand (sCD40L) levels in plasma were observed between AD cases and control subjects. sCD40 was approximately three times higher in patients with AD than in control subjects with sensitivity and specificity of 68% and 84%, respectively. Similarly, concentrations of sCD40L were 2.27 times higher in AD cases than in control subjects with sensitivity and specificity of 51% and 76%, respectively [
23]. A biomarker panel of cortisol, pancreatic polypeptide, insulin-like growth factor binding protein 2, β
2-microglobulin, vascular cell adhesion molecule 1, carcinoembryonic antigen, matrix metalloproteinase 2, CD40, macrophage inflammatory protein 1α, superoxide dismutase, and homocysteine was shown to significantly increase in plasma from patients with AD. In addition, apolipoprotein E (ApoE), epidermal growth factor receptor, hemoglobin, calcium, zinc, interleukin (IL)-17, and albumin were revealed to be decreased in patients with AD. Cross-validated accuracy measures from the Australian Imaging, Biomarker & Lifestyle Flagship Study of Ageing (AIBL) cohort reached a mean (SD) of 85% (3.0%) for sensitivity and specificity and 93% (3.0) for the AUROC. A second validation using the Alzheimer’s Disease Neuroimaging Initiative cohort showed accuracy measures of 80% (3.0%) for sensitivity and specificity and 85% (3.0) for the AUROC [
24]. Eighteen signaling blood proteins in hematopoiesis, immune responses, apoptosis, and neuronal supports were suggested to differentiate patients with AD from control subjects with close to 90% accuracy and also to identify patients who would convert from mild cognitive impairment to AD 2–6 years later [
25].
Among these, methods employing blood-based biomarkers have been focused on the detection of amyloid-β oligomers (AβOs) and other surrogate biomarkers of AD [
26‐
31]. Experimental cross-sectional analyses undertaken to detect AβOs in the plasma have demonstrated limited and inconclusive results [
26‐
28]. Other blood-based surrogate biomarkers, including ApoE, inflammatory markers (IL-8, IL-1a), Aβ autoantibodies, total serum cholesterol, and microRNAs (specifically miR-9, miR-29a, miR-29b, miR101, miR-125b, miR-132, miR-134, and miR-181c), have also demonstrated variability as disease correlates [
29‐
31]. Although there have been difficulties in developing methods for AD diagnosis using blood-based biomarkers, a reliable and reproducible blood-based assay is still needed for clinical use [
32].
A multimer detection system (MDS) was originally developed to detect prion oligomers in the blood of scrapie-infected animals. MDS is a sandwich enzyme-linked immunosorbent assay (ELISA) that preferentially detects oligomers over monomers by creating steric hindrance between capturing and detection antibodies that are specific to a unique/overlapping epitope [
33].
In the present study, MDS for AD was developed to detect AβOs using two different antibodies against the N-terminus of Aβ. Initially, MDS was unable to differentiate AβOs in the blood of patients with AD from those of normal control subjects. Synthetic Aβ was then spiked into the plasma of patients with AD and control subjects. Using MDS, the dynamic changes of AβO formation were detected in the spiked plasma of patients with AD but not in the spiked plasma of control subjects. Therefore, in this study, we evaluated the dynamic changes of AβO levels in the plasma of patients with AD compared with those of normal age-matched control subjects.
Methods
Clinical data
This study was approved by the institutional review board of Seoul National University Bundang and Chung-Ang University Hospital [B-0905-075-003, B-1202-145-003, C2012048(743), C2013142(1102)]. Pooled plasma samples were collected from 11 patients with AD and 9 elderly normal control subjects, and individual plasma samples of 24 patients with AD and 29 healthy elderly normal control subjects were collected from either Seoul National University Bundang Hospital or Chung-Ang University Hospital (Table
1). Written informed consent was obtained from all patients who participated in this study or from their caregivers. AD cases were each diagnosed with a probable AD amnestic type on the basis of clinical criteria of the National Institute on Aging-Alzheimer’s Association workgroups within a clinical setting with clinical data and follow-up longer than 6 month before inclusion into PiB-PET or CSF studies. Hence, the recruited patients were clinically well-characterized patients with AD, and only they were included in the study. They were diagnosed with AD after initial workup and had not shown any possibility of other neurodegenerative disorders except AD or secondary dementia disorders on the basis of more than 6 months of follow-up. The Mini Mental State Examination, identification of the ApoE phenotype, PET imaging with PiB and
18F-fluorodeoxyglucose, and CSF analysis were performed. The characteristics of all participants are described in Table
1. Fifty-one additional plasma samples from elderly normal control subjects were included to avoid false positivity (Table
2).
Table 1
Characterization of patients with Alzheimer’s disease and healthy normal control subjects
Total sample number | 24 | 29 |
Sex |
Female | 13 (54.2%) | 16 (55.2%) |
Male | 11 (45.8%) | 13 (44.8%) |
Age, years (SD) | 67.6 (±7.4) | 62.4 (±5.7) |
Education, years (SD) | 13.1 (±3.9) | 13.2 (±3.5) |
CDR-SOB, mean | 6.35 | 0.03 |
MMSE score, mean | 17.7 | 29.03 |
ApoE ε4, % | 47.8 | 21.7 |
Note test | 1 | 0 |
Number of plasma samples | 24 | 29 |
CSF markers | 23 | 28 |
Aβ42, pg/ml, mean (SD) | 258.6 (±70.8) | 464.8 (±114.4) |
p-Tau, pg/ml, mean (SD) | 58.6 (±18.6) | 28.0 (±14.3) |
t-Tau, pg/ml, mean (SD) | 132.1 (±61.8) | 62.1 (±20.3) |
PiB-PET number | 23 | 28 |
Mean SUVR | 1.57 | 1.14 |
FDG-PET number | 18 | 28 |
Mean SUVR | 0.9 | 1.06 |
Table 2
Supplementary information on healthy normal control subjects
Total sample number | 51 |
Sex |
Female | 21 (43.1%) |
Male | 29 (56.9%) |
Age, years (SD) | 62.25 (±7.89) |
Education, years (SD) | 10.35 (±3.29) |
CDR-SOB | 0.03 |
MMSE score | 28.27 |
Sample preparation
Blood samples were collected in heparin-containing tubes and centrifuged at 850 × g for 30 minutes. The plasma (supernatant) was divided into aliquots and stored at −80 °C until analysis. Plasma samples of patients with AD (n = 11) and elderly normal control subjects (n = 9) were separately pooled for initial method optimization, whereas the remaining samples were assessed individually.
Preparation of synthetic Aβ42
Lyophilized AggreSure β-Amyloid (1–42) peptide (AnaSpec, Fremont, CA, USA) and double-mutant F19S/L34P Aβ42 (mutAβ42; AnyGen Co., Ltd., Gwangju, South Korea) were each dissolved in 50 mM Tris/150 mM NaCl (pH 7.2) at a concentration of 1 mg/ml and then sonicated for 5 minutes to obtain a homogeneous solution. The peptide solution was further diluted with phosphate-buffered saline containing Tween 20 (PBST; Sigma-Aldrich, St. Louis, MO, USA) to a desired concentration of 10 μg/ml. Solutions of diluted peptides were divided into aliquots and kept at −80 °C until further use.
Thioflavin T assay
Aβ aggregation was monitored using a thioflavin T (ThT) assay kit following the suggested protocol of the manufacturer (AnaSpec). Ninety microliters of test sample and 10 μl of 2 mM ThT solution were added to each well of a 96-well plate (Thermo Fisher Scientific, Waltham, MA, USA), and the plates were incubated for different lengths of time. Then, changes in ThT fluorescence intensity were detected by measuring excitation and emission wavelengths of 440 nm and 484 nm, respectively, using a multispectrophotometer (Victor 3TM; PerkinElmer, Waltham, MA, USA) with 15 seconds of shaking before reading and analysis.
TEM
AβO, protofibrils, and fibrils were characterized by TEM at various incubation times (0, 1, 3, 6, 24, and 48 h). Five microliters of each sample was applied to carbon-coated TEM grids that had previously been glow-discharged for 3 minutes in the air and immediately negatively stained (~5 seconds) with 2% uranyl acetate. Excess solution was removed with blotting paper. Image acquisition was carried out using a Philips CM10 transmission electron microscope (Philips Research, Eindhoven, The Netherlands) with an accelerating voltage of 80 kV.
Sodium dodecyl sulfate-PAGE and immunoblotting
The aggregation state of Aβ was also analyzed by sodium dodecyl sulfate-PAGE followed by Western blotting. Synthetic peptide samples were electrophoresed on a 10–20% Tris-Tricine precast gel (Bio-Rad Laboratories, Hercules, CA, USA) and visualized by Coomassie blue staining (Bio-Rad Laboratories). After electrophoresis, the proteins were transferred to a polyvinylidene fluoride membrane (Bio-Rad Laboratories), which was blocked with 2% Block Ace (Bio-Rad Laboratories) in Tris-buffered saline containing Tween 20 (TBST; Sigma-Aldrich) for 1 h at room temperature (RT) under conditions to reduce nonspecific binding. The membrane was incubated for 1 h at RT with a horseradish peroxidase (HRP)-conjugated FF51 antibody (FF51-HRP antibody; PeopleBio Inc., Seoul, South Korea) diluted in 0.4% Block Ace in TBST. Proteins bound to the antibody were visualized with 3,3′,5,5′-tetramethylbenzadine reagent (Sigma-Aldrich).
MDS for Alzheimer’s disease
A modified MDS was used to measure AβOs. With this method, epitope-overlapping antibodies specific for the N-terminus of Aβ were used to capture and detect the Aβ antigen in its multimeric or oligomeric form. Because MDS was initially developed to detect prion oligomers using prion antibodies, over 100 sets of antibodies against Aβ were screened (data not shown). In addition, in-house Aβ antibodies were developed. The mouse monoclonal antibody 6E10 (BioLegend, San Diego, CA, USA) and an in-house FF51-HRP antibody were chosen to detect AβOs in our modified MDS, owing to their sensitivity and specificity. The epitopes for these antibodies overlap at the N-terminus of Aβ. The FF51 antibody specifically recognizes amino acid residues 1–4 of Aβ.
To use MDS, the 6E10 antibody was coated overnight at 4 °C in the wells of a 96-well black plate (Thermo Fisher Scientific) at a dilution of 3 μg/ml in carbonate-bicarbonate buffer (Sigma-Aldrich). The plates were blocked for 2 h with 0.4% Block Ace (100 μl) at RT. After washing three times with PBS (Sigma-Aldrich), the plate was stored at 4 °C until use. Prior to the assay, aliquots of plasma samples were thawed at 37 °C for 15 minutes. Ten microliters of plasma, 4.04 μl of HBR-1, a HAMA blocker (Scantibodies Laboratory, Santee, CA, USA), and PBST were mixed. We spiked the synthetic Aβ42 into plasma mixture and incubated it at 37 °C for the indicated durations.
The plasma sample mixture and serially diluted standards were added to each well of the plate in a total volume of 100 μl. The plates were incubated at RT for 1 h. After washing three times with TBST, the FF51-HRP antibody in TBST containing 0.4% Block Ace was added to the wells, and the plate was incubated for 1 h at RT. To increase the sensitivity of detection, 100 μl/well of enhanced chemiluminescence substrate solution (Rockland Immunochemicals Inc., Limerick, PA, USA) was used, and the luminescent signal was detected and quantified using a Victor 3TM multispectrophotometer.
Measurement of Aβ monomers (Aβ40 and Aβ42)
Sandwich ELISAs were performed to measure Aβ40 and Aβ42 monomer levels. Aβ40 monomers were captured with the 11A50 antibody (specific for the C-terminus of Aβ40) and detected with the 1E11 antibody conjugated to biotin. Aβ42 monomers were captured with the 12F4 antibody (specific for the C-terminus of Aβ42) and detected with the 1E11 antibody conjugated to biotin.
Statistics
Statistical evaluations were performed using the Mann-Whitney U test followed by the calculation of two-tailed p values to determine the significance between groups.
Discussion
MDS was initially developed to detect prion disease. This method consists of a sandwich ELISA that exclusively detects oligomeric forms of antigens and relies on two different epitope-overlapping antibodies to capture and detect antigens by creating steric hindrance over a specific epitope [
33,
39]. MDS was modified to detect AβOs for AD diagnosis using two epitope-overlapping Aβ antibodies specific for the N-terminus amino acids 1–18 of Aβ
42.
We monitored the formation of AβO, protofibrils, and fibrils from synthetic Aβ
42 in PBST buffer employing two methods: a ThT assay and MDS (Fig.
1). Aβ oligomerization was closely correlated between these two assays. When the time-course changes of synthetic Aβ in PBST buffer were monitored by MDS, ThT assay, and TEM, the MDS detected a continuous increase of Aβ
42 oligomer levels from 0 to 24 h from the start of incubation, followed by a decline until 120 h. Conversely, an increase of fluorescent signal in the ThT assay was correlated with an increase of Aβ
42 fibril formation, which was confirmed by TEM imaging (Fig.
1c and d). The MDS signal decreased at that time of fibrillization; thus, it would be reasonable to assume that MDS detected specifically oligomeric and protofibril forms of Aβ. During the early phase of incubation, MDS revealed decreasing AβO levels accompanied by the formation of protofibrils and fibrils, but the levels detected by the ThT assay were stable. Thus, MDS is able to detect AβOs during the early stage of Aβ oligomerization. Decreasing AβO levels, as measured by MDS, in conjunction with the formation of protofibrils and fibrils is potentially explained by two hypotheses. First, MDS has a higher sensitivity for oligomers than for protofibrils or fibrils. Second, the total number of oligomers decreases over time. After full oligomerization, the spiked Aβ converts to protofibrils, resulting in decreased MDS signals.
When synthetic Aβ
42 is spiked into plasma, the plasma composition may dictate and interfere with Aβ oligomerization. Difficulties in detecting spiked synthetic Aβ
42 via routine ELISA may be attributable to the presence of many different interfering factors in the plasma, which bind to the spiked synthetic Aβ
42 and thus reduce its detection. These factors include naturally occurring Aβ autoantibodies, albumin, fibrinogen, immunoglobulin, ApoJ, ApoE, transthyretin, α
2-macroglobulin, serum amyloid P component, plasminogen, and amylin [
35‐
38,
40]. In addition, these molecules in bodily fluids could also inhibit Aβ fibrillization. Our hypothesis was that the composition of those components in blood from patients with AD would be different from that in healthy control subjects. If the same amount of Aβ were spiked into AD and control plasma samples, a different phenomenon would be observed between the two groups.
In this study, when equal amounts of synthetic Aβ
42 were spiked into plasma, MDS signals for AβOs declined from 0 to 48 h in both groups (Fig.
2b), potentially due to AβO binding to interfering factors. Although binding affinities and Aβ epitopes likely vary among different binding factors, high concentrations in the plasma would likely result in the scavenging of spiked synthetic Aβ
42. Binding of these factors to Aβ may naturally influence the normal functions and sequestration of Aβ, leading to clearance and reduced oligomerization potential in elderly normal persons [
35,
37].
Forty-eight hours after spiking with Aβ
42, AβO levels measured by MDS began to increase in the plasma from patients with AD but not in that of normal control subjects (Fig.
2b). Patients with AD may exhibit different binding profiles based on their plasma composition in the context of spiked synthetic Aβ
42, which in turn may increase oligomerization potential and decrease sequestration capacity. Alternatively, the characteristics of endogenous plasma Aβ may differ between the two groups, permitting the dynamic changes of Aβ oligomerization to be detected.
It is challenging work to detect crude AβOs in plasma because the concentrations of Aβ in blood are very low. Furthermore, the concentrations of AβOs would be a subset of total Aβ in blood. The size of 4.5 kDa could be another reason why the MDS failed to measure the endogenous plasma Aβ from patients with AD. As shown in Fig.
3, the MDS did not discriminate between patients with AD and healthy normal control subjects without incubation conditions. Even though incubation of plasma samples without spiking external Aβ made slight differences between AD and normal control samples, the difference was not significant.
However, even at ultralow concentrations, the formation of AβO in the blood of patients with AD may be initiated via incubation with spiked synthetic Aβ
42. The first 48 h of incubation represent a slow nucleation-dependent oligomerization phase during which ultralow concentrations of AβO nuclei are required to bind to spiked synthetic Aβ
42. The period after 48 h and up to 144 h represents a rapid-growth phase for the formation of oligomers, protofibrils and fibrils, surpassing the critical detection limit [
34,
41‐
43]. As previously mentioned, the interfering factors may be saturated with spiked Aβ; hence, they may not affect to MDS signal even after 48 h of incubation. In addition, the binding interactions between Aβ and another Aβ may be stronger than other interfering factors for oligomerization. In previous studies, by using protein misfolding cyclic amplification technology, researchers were able to differentiate AβO levels by catalyzing the misfolding and amplification of Aβ aggregates by spiking Aβ
42 into the CSF of patients with AD and control individuals [
44]. However, spiking Aβ into the plasma to differentiate AβO has not previously been published.
We observed the phenomenon that MDS signals of control subjects were still stable while we spiked the same large amount of synthetic peptide into both AD and normal plasma samples, even after the identical incubation step. Currently, we do not know the exact cause of the phenomenon. Oligomerization of Aβ could be influenced by potential factors in plasma of patients with AD but not plasma samples from normal control. The concentrations of these potential factors could be different in the disease state, but they may not be present in the normal state.
We detected differential AβO dynamic changes in the blood of patients with AD and normal control subjects, but a direct correlation between blood and brain pathology remains uncharacterized. The properties of Aβ plaques in the brain may differ from those in the blood because Aβ in the blood also originates from amyloid precursor protein metabolism in skeletal muscle, organs, skin, and peripheral cells [
45,
46]. However, on the basis of previous reports, Aβ peptides cross the blood-brain barrier, resulting in elevated Aβ levels in the CSF and plasma during intracerebroventricular injection of synthetic Aβ
42 monomers into normal imprinting control region mice [
47,
48]. It will be interesting to identify the correlation between AβO concentrations in the plasma and amyloid plaque deposition in the brains of patients with AD.