Article Text

PDF
Research paper
Association between naturally occurring anti-amyloid β autoantibodies and medial temporal lobe atrophy in Alzheimer's disease
  1. Akio Kimura1,
  2. Masao Takemura2,3,
  3. Kuniaki Saito4,
  4. Nobuaki Yoshikura1,
  5. Yuichi Hayashi1,
  6. Takashi Inuzuka1
  1. 1Department of Neurology and Geriatrics, Gifu University Graduate School of Medicine, Gifu, Japan
  2. 2Department of Informative Clinical Medicine, Gifu University Graduate School of Medicine, Gifu, Japan
  3. 3Advanced Diagnostic System Research Laboratory, Fujita Health University Graduate School of Health Sciences, Toyoake, Aichi, Japan
  4. 4Department of Disease Control and Prevention, Fujita Health University Graduate School of Health Sciences, Toyoake, Aichi, Japan
  1. Correspondence to Dr Akio Kimura, Department of Neurology and Geriatrics, Gifu University Graduate School of Medicine, Gifu, 1-1 Yanagido, Gifu 501-1194, Japan; kimura1{at}gifu-u.ac.jp

Abstract

Background Naturally occurring autoantibodies against amyloid β (Aβ) peptide exist in the serum and cerebrospinal fluid (CSF) of healthy individuals. Recently, it was reported that administration of intravenous immunoglobulin at the mild cognitive impairment (MCI) stage of Alzheimer's disease (AD) reduces brain atrophy.

Objective To examine the association between naturally occurring anti-Aβ autoantibodies and brain atrophy in patients with cognitive impairment.

Methods Serum and CSF levels of anti-Aβ autoantibodies and CSF biomarkers were evaluated in 68 patients with cognitive impairment, comprising 44 patients with AD, 19 patients with amnestic MCI and five patients with non-Alzheimer's dementia. The degree of brain atrophy was assessed using the voxel-based specific regional analysis system for AD, which targets the volume of interest (VOI) in medial temporal structures, including the whole hippocampus, entorhinal cortex and amygdala.

Results CSF levels of anti-Aβ autoantibodies were inversely correlated with the extent and severity of VOI atrophy, and the ratio of VOI/grey matter atrophy in patients with AD, but not in MCI or non-AD patients. Serum levels of anti-Aβ autoantibodies were not associated with these parameters in any of the patient groups.

Conclusions These results indicate that CSF levels of naturally occurring anti-Aβ autoantibodies are inversely associated with the degree of the VOI atrophy in patients with AD. Although the mechanism is unclear, CSF levels of naturally occurring anti-Aβ autoantibodies may be implicated in the progression of atrophy of the whole hippocampus, entorhinal cortex and amygdala, in AD.

Statistics from Altmetric.com

Introduction

Alzheimer's disease (AD) represents the most common neurodegenerative disorder and the leading cause of cognitive impairment.1 The pathophysiology of AD is thought to involve the accumulation and deposition of amyloid β (Aβ) peptide in the form of plaques,2–4 and clearance of Aβ from the brain has been extensively researched as a therapeutic strategy for AD.5 It has been reported that naturally occurring autoantibodies against Aβ peptide are present in healthy individuals.7 In an AD mouse model, administration of monoclonal anti-Aβ antibodies was effective in reducing plaques.6 In open-label clinical trials in patients with AD, administration of intravenous immunoglobulin (IVIg) was associated with a modest improvement in cognitive function as well as amyloid clearance.8 ,9 Short-term treatment with IVIg at the mild cognitive impairment (MCI) stage of AD has been demonstrated to reduce brain atrophy, prevent cognitive decline and delay the progression to dementia.10 These data suggest that naturally occurring anti-Aβ autoantibodies in IVIg preparations may have a therapeutic effect. However, the association between levels of naturally occurring anti-Aβ autoantibodies and the degree of brain atrophy in patients with AD has not been investigated.

Several reports have described the levels of naturally occurring anti-Aβ autoantibodies in patients with AD.11–13 Some results indicate that CSF anti-Aβ autoantibody levels are reduced in patients with AD,11 whereas others have found no difference between patients with AD and healthy individuals.12 ,13 Different methods of measurement may be one explanation for the conflicting data.

The medial temporal lobe (MTL) is a system of anatomically related structures that are essential for declarative memory (conscious memory of facts and events).14 The core pathological changes in AD take place in the MTL, and start to occur several years before the clinical manifestation of dementia.15–17 Thus, the MTL is a region of interest for early diagnosis of AD. MTL atrophy, assessed by structural MRI, predicts the development of AD in individuals with MCI.18–20 There are several ways to determine the degree of MTL atrophy. Computer-aided voxel-based morphometry (VBM) has been applied to detect early atrophic changes in AD. The voxel-based specific regional analysis system for AD (VSRAD), which targets the MTL, was developed as a sensitive diagnostic tool for the early stages of AD.21 VSRAD enables the degree of MTL atrophy to be assessed by comparing a given subject's grey matter (GM) volume with that of a healthy control database template. In this study, the association between naturally occurring anti-Aβ autoantibodies and MTL atrophy was examined using VSRAD in patients with different types of cognitive impairment. The MTL was defined as the region comprising the medial temporal structures, specifically, the whole hippocampus, entorhinal cortex and amygdala.

Patients and methods

Patients

The study enrolled 68 consecutive patients who were admitted to the Department of Neurology and Geriatrics (Gifu University Graduate School of Medicine), because of memory loss, from February 2013 to July 2015. Among these 68 patients, 44 had AD, 19 had amnestic MCI and 5 had non-Alzheimer's dementia (non-AD). All patients underwent physical and neurological examinations, neuropsychological testing, including the mini-mental state examination (MMSE) and frontal assessment battery (FAB), laboratory studies and brain MRI. All patients with AD met the NINCDS-ADRDA criteria for probable AD.22 Among the five non-AD patients, two had diffuse Lewy body disease, two had corticobasal syndrome and one had frontotemporal dementia. CSF samples were collected from all 68 patients. Serum samples were collected from 38 of the 44 patients with AD and all 19 patients with MCI. Written informed consent was obtained from all patients. This study was approved by the Institutional Review Board of Gifu University Graduate School of Medicine, Gifu, Japan.

Evaluation

CSF samples for routine diagnostic investigations were obtained by lumbar puncture without trauma and stored at −80°C until analysis. Anti-Aβ autoantibody levels, Aβ40, Aβ42 and phosphorylated 181 tau (P-Tau) protein were measured in the CSF samples. CSF Aβ40 levels were determined using commercially available ELISA kits (Wako, Osaka, Japan), according to the manufacturer's instructions. Levels of Aβ42 and P-Tau were analysed by a commercial laboratory (SRL, Tokyo, Japan).

Determination of anti-Aβ autoantibody levels

ELISA was used to measure anti-Aβ autoantibody levels. An indirect test was performed using 96-well plates (Maxisorp Loose Nunc-Immuno Module; Nunc, Kamstrup, Denmark) coated with 1 µg/100 μL/well of human recombinant Aβ40 protein (hydrochloride) (Wako, Osaka, Japan) in carbonate-bicarbonate buffer for 1–2 hours at room temperature, and then overnight at 4°C. After washing with 1× phosphate-buffered saline (PBS) containing 0.05% Tween 20 (PBST), non-specific binding sites were blocked using 1% skimmed milk in 1× PBS (100 μL/well) for 2 hours at room temperature. After two washes with 1×PBST, patients' sera (1:100) and CSF (1:2) diluted with 1× PBS were applied (100 μL/well) for 1.5 h at room temperature. The washing procedure was repeated three times, and horseradish peroxidase-conjugated goat antihuman IgG antibodies (Vector, Burlingame, California, USA) diluted (1:3000) with 1× PBS were added (100 μL/well), and incubated for 1 h at room temperature. After three washes, substrate solution (100 μL/well) (TMB Substrate Reagent Set; BD Biosciences, Franklin Lakes, New Jersey, USA) was applied. After 30 min, the reaction was stopped using 1 M H2SO4 (100 μL/well) and the absorbance was read at 450 nm using an ELISA processor (ELISA Processor II; Dade Behring, Marburg, Germany). For standard curve generation, pooled serum samples from five patients with AD who showed relatively high optical absorbance in the preliminary test were used at different concentrations by serial dilution. Quantitative values (in U/100 μL) for anti-Aβ autoantibody levels were assigned using a standard curve consisting of 1.562, 6.25, 25 and 100 U, where 100 U=1:10 dilution of a standard sample that was used for all ELISA runs.

Voxel-based specific regional analysis system for Alzheimer's disease

All 68 patients underwent 1.5T brain MRI scans and all patients, except for 1 non-AD patient, underwent VSRAD analysis, using free software (VSRAD advance, Eisai Co, Ltd, Tokyo, Japan). Detailed information on VSRAD analysis has been described in previous reports.23 ,24 Briefly, MRI scans were segmented into GM, white matter (WM) and CSF images, using a unified tissue-segmentation procedure after image intensity non-uniformity correction. Each processed segmented image was compared with the mean and SD of GM or WM images from 80 healthy volunteers using voxel-by-voxel Z-score analysis, with and without voxel normalisation to global mean intensities (global normalisation), Z-score=[(control mean)-(individual value))/(control SD)]. The Z-score maps were superimposed on tomographic sections and surface rendered to the standardised brain. A representative Z-score map of GM volume reductions in a patient with AD using VSRAD is shown in figure 1. VSRAD recorded the target volume of interest (VOI) in medial temporal structures, including the entire region of the entorhinal cortex, hippocampus and amygdala. A Z-score >2 was defined as significant atrophy, consistent with previous studies.23 ,25 In addition, five indicators of atrophy in the target VOI and the whole brain were defined, as follows: (1) the severity of atrophy obtained from the averaged positive Z-score in the target VOI (the severity of VOI atrophy); (2) the extent of a region showing significant atrophy in the target VOI, that is, the percentage rate of the coordinates with a Z-value exceeding the threshold value of 2 in the target VOI (the extent of VOI atrophy); (3) the extent of a region showing significant atrophy in the GM, that is, the percentage rate of the coordinates with a Z-value exceeding the threshold value of 2 in the GM (the extent of GM atrophy); (4) the extent of a region showing significant atrophy in the WM, that is, the percentage rate of the coordinates with a Z-value exceeding the threshold value of 2 in the WM (the extent of WM atrophy); and (5) the extent of a region showing significant atrophy in the target VOI relative to the extent of a region showing significant atrophy in the whole brain (the ratio of VOI/GM atrophy).

Figure 1

Z-score map of grey matter volume reductions in a patient with Alzheimer's disease, using the voxel-based specific regional analysis system for Alzheimer's disease (VSRAD). A representative Z-score map of grey matter volume reductions is shown for a 63-year-old woman with Alzheimer's disease. Coloured areas with Z-score of >1 were overlaid as atrophied regions on axial sections (A) and the cortical surface (B) of the standardised MRI template.

Statistical analysis

Data analysis was performed using statistical software (Ekuseru-Toukei 2012, Social Survey Research Information, Tokyo, Japan). Statistically significant differences among the three disease groups were calculated using the Steel–Dwass test, and differences between two disease groups were calculated using the Mann-Whitney U test. Spearman's rank correlation was used to assess the correlations between data. A p value of ≤0.05 was considered as denoting statistical significance.

Results

Comparison of the levels of anti-Aβ autoantibodies and CSF biomarkers, and the five indicators on VSRAD

A comparison of levels of anti-Aβ autoantibodies and CSF biomarkers, and the five indicators of VSRAD analysis in the three patient groups, is presented in table 1. The extent of GM atrophy was significantly higher in patients with AD than in patients with MCI (p=0.001). The extent of VOI atrophy was significantly higher in patients with AD than in patients with MCI (p=0.031). There were no significant differences between the three groups in the levels of CSF anti-Aβ autoantibody and CSF biomarkers, and no significant difference between AD and patients with MCI in serum anti-Aβ autoantibody levels.

Table 1

Clinical characteristics, laboratory findings and indicators on the VSRAD analysis in patients with cognitive impairment

Correlation between anti-Aβ autoantibody levels and the five indicators on VSRAD

Correlations between anti-Aβ autoantibody levels and each of the five indicators on VSRAD were analysed (table 2, figure 2). In all patients, CSF anti-Aβ autoantibody levels were inversely correlated with the extent of VOI atrophy (r=−0.258, p=0.035), the severity of VOI atrophy (r=−0.271, p=0.026) and the ratio of VOI/GM atrophy (r=−0.255, p=0.037) (table 2, figure 2A–C). In patients with AD, CSF anti-Aβ autoantibody levels were also inversely correlated with the extent of VOI atrophy (r=−0.319, p=0.035), the severity of VOI atrophy (r=−0.305, p=0.044) and the ratio of VOI/GM atrophy (r=−0.314, p=0.038) (table 2, figure 2D–F). In contrast, neither patients with MCI nor non-AD patients had significant correlations on VSRAD analysis between CSF anti-Aβ autoantibody levels and each indicator. No significant correlations were detected between CSF anti-Aβ autoantibody levels and the extent of GM or WM atrophy.

Table 2

Correlations between anti-Aβ autoantibody levels and each indicator on the VSRAD, and CSF biomarkers

Figure 2

Correlation between anti-amyloid β autoantibody levels in CSF and each indicator on voxel-based specific regional analysis system for Alzheimer's disease (VSRAD) analysis. Correlations between CSF anti-Aβ autoantibody levels and each indicator on VSRAD in all patients (A–C) and patients with AD (D–F). Correlations between CSF anti-Aβ autoantibody levels and the severity of VOI atrophy (A, D), extent of VOI atrophy (B, E) and ratio of VOI/GM atrophy (C, F). Aβ, Amyloid β; AD, Alzheimer's disease; CSF, cerebrospinal fluid; GM, grey matter; VOI, volume of interest; VSRAD, voxel-based specific regional analysis system for Alzheimer's disease.

In addition, no significant correlations were detected between serum anti-Aβ autoantibody levels and each indicator on VSRAD analysis, except for the extent of WM atrophy in patients with MCI (r=0.483, p=0.036).

Correlation between anti-Aβ autoantibody levels and cognitive function

Neither serum nor CSF anti-Aβ autoantibody levels—and cognitive function—had significant correlations between them, as assessed by either MMSE or FAB scores.

Correlation between anti-Aβ autoantibody levels and CSF biomarkers

The relationships between CSF levels of anti-Aβ autoantibodies and CSF levels of Aβ40, Aβ42 and P-Tau, were analysed. CSF anti-Aβ autoantibody levels were inversely correlated with CSF Aβ42 levels in all patients (r=−0.405, p=0.001), patients with AD (r=−0.395, p=0.008) and patients with MCI (r=−0.617, p=0.005) (table 2, figure 3). There were no significant correlations between CSF anti-Aβ autoantibody levels and CSF levels of Aβ40 or P-Tau. Serum anti-Aβ autoantibody levels were not significantly correlated with any of the CSF biomarkers.

Figure 3

Correlation between anti-amyloid β autoantibody levels and amyloid β42 levels in CSF, Correlations between CSF anti-Aβ autoantibody levels and CSF Aβ42 levels in all patients (A), patients with AD (B) and patients with MCI (C). Aβ, Amyloid β; AD, Alzheimer's disease; CSF, cerebrospinal fluid; MCI, mild cognitive impairment.

Discussion

The findings of this study revealed that CSF anti-Aβ autoantibody levels were inversely correlated with the extent and severity of VOI atrophy, and the ratio of VOI/GM atrophy in patients with AD, but not in MCI and non-AD patients. However, there were no significant correlations between CSF anti-Aβ autoantibody levels and the extent of GM or WM atrophy. These data suggest that CSF levels of naturally occurring anti-Aβ autoantibodies are inversely associated with the degree of MTL atrophy in patients with AD. The MTL is implicated as a key brain region in the pathogenesis of AD and cognitive impairment.15–17 Tau tangle aggregation in this region may develop concurrently with cortical Aβ deposition in preclinical AD; however, the pathological relationship between tau and Aβ remains unclear.26–28 The amyloid hypothesis of AD, which has been the dominant theory of disease causation for decades, proposes that the accumulation of tau tangles is a downstream event in brain Aβ deposition.2–4

The mechanism by which naturally occurring anti-Aβ autoantibodies in the CSF influence MTL atrophy in AD is unclear. One possibility is that naturally occurring anti-Aβ autoantibodies in the CSF delay the progression of MTL atrophy in AD. Passive administration of monoclonal antibodies specific for synthetic peptide fragments of Aβ was effective in clearing Aβ and improving memory deficits in a mouse model of AD.6 ,29 It was reported that naturally occurring anti-Aβ autoantibody can inhibit Aβ aggregation and block Aβ neurotoxicity.29 ,30 Naturally occurring anti-Aβ autoantibodies may be a physiological mechanism to prevent AD development.31

Our study demonstrated that CSF anti-Aβ autoantibody levels are inversely associated with CSF Aβ42 levels in AD and MCI. CSF anti-Aβ autoantibodies may bind to soluble Aβ42, reducing CSF Aβ42 levels and promoting amyloid clearance from the AD brain. Other proposals include the peripheral sink hypothesis, which postulates that intravenous administration of Aβ-specific antibodies results in a net efflux of Aβ from the brain to the plasma.32 ,33 Intravenous anti-Aβ antibody administration is thought to shift the CNS-to-plasma Aβ equilibrium due to protein binding in the periphery. We consider that naturally occurring anti-Aβ autoantibodies in the CSF may shift the parenchyma-to-CSF Aβ equilibrium by binding to soluble Aβ42 in the CSF, resulting in a net efflux of Aβ from the parenchyma of the AD brain. Thus, naturally occurring anti-Aβ autoantibodies may interfere with the pathogenesis of AD by various mechanisms. Further studies are needed to examine whether CSF levels of naturally occurring anti-Aβ autoantibodies play a role in amyloid clearance in the AD brain.

Kile et al10 reported that administration of IVIg every 2 weeks for a total of five infusions in patients with amnestic MCI reduces brain atrophy, prevents cognitive decline and delays conversion to dementia for at least 1 year; however, the beneficial effects of IVIg were diminished at 2 years. Although the mechanism of action of IVIg has not been fully elucidated, IVIg containing naturally occurring anti-Aβ autoantibodies may induce amyloid clearance. Importantly, in this study, CSF levels of naturally occurring anti-Aβ autoantibodies, but not serum levels, were inversely associated with the degree of MTL atrophy in patients with AD. To demonstrate the effects of IVIg on the progression of brain atrophy and cognitive impairment, it may be necessary to maintain increased CSF levels of naturally occurring anti-Aβ autoantibodies for a longer time period during the preclinical stage of AD.

In this study, there were no significant correlations between CSF anti-Aβ autoantibody levels and MMSE or FAB scores. One explanation for this result is the limited sample size of the present study; also, the cognitive impairment of the patients with AD was relatively mild. Prospective studies that include a higher number of patients are required.

Acknowledgments

The authors would like to thank Dr A Takekoshi, Dr H Segawa, Dr M Yasunishi and Dr A Koumura (Departments of Neurology and Geriatrics, Gifu University Graduate School of Medicine), for providing patients' clinical information. This research was partially supported by a Grant-in-Aid for Scientific Research (C) (26461290) from the Japan Society for the Promotion of Science.

References

View Abstract

Footnotes

  • Contributors AK was involved in the study design, data collection, data analysis, statistical analysis and drafting of the manuscript. MT and KS was involved in the data analysis and drafting of the manuscript. NY and YH was involved in the data collection and drafting of the manuscript. TI was involved in the study design and drafting of the manuscript.

  • Funding Grant-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science (grant number 26461290).

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Institutional Review Board of Gifu University Graduate School of Medicine.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles