Double immunostaining of a senile plaque in Alzheimer disease brain tissue with end-specific monoclonal antibodies against amyloid β (Aβ) peptide isoforms Aβ42 (left) and Aβ40 (right). The Aβ42 antibody was directly conjugated to Cy3 for fluorescent detection. Aβ42 was the predominant species in Alzheimer disease brain, and Aβ40 was generally localized to the core of a subset of plaques.
The amyloid burden for amyloid β (Aβ) peptide isoforms Aβ40 and Aβ42 vs duration of illness. The weak correlation coefficients (R2=0.1 for Aβ42 and R2=0.14 for Aβ40) suggest that total amounts of both Aβ40 and Aβ42 do not increase with longer duration of illness.
The amounts of amyloid β (Aβ) peptide isoforms Aβ40 and Aβ42 in Alzheimer disease brain tissue with varying degrees of severity of congophilic amyloid angiopathy (CAA). We found no relationship between the severity of CAA and the relative amounts of Aβ40 and Aβ42 in Alzheimer disease brain tissue.
McNamara MJ, Gomez-Isla T, Hyman BT. Apolipoprotein E Genotype and Deposits of Aβ40 and Aβ42 in Alzheimer Disease. Arch Neurol. 1998;55(7):1001-1004. doi:10.1001/archneur.55.7.1001
To examine the differential deposition of amyloid β (Aβ) peptide isoforms Aβ40 and Aβ42 in the Alzheimer disease (AD) brain in relation to the apolipoprotein E (APOE) genotype.
The APOE ϵ4 genotype is an inherited risk factor for AD and is associated with increased deposition of Aβ protein in the cerebral cortex. Previous data from familial AD due to mutations in presenilin 1 and presenilin 2 genes and the amyloid precursor protein suggest that the long form of Aβ peptide, Aβ42, is selectively increased in these circumstances. Herein, we examine whether APOE genotype influenced the species of Aβ peptide deposited.
Design and Methods
The amount of Aβ40, Aβ42, and total Aβ deposited in immunostained temporal lobe tissue of 28 cases of AD of known APOE genotype was determined.
Individuals with the APOE ϵ4 genotype (APOE ϵ4/4) were associated with both increased Aβ40 (P<.05) and Aβ42 (P<.05) compared with individuals without the APOE ϵ4/4 genotype.
Our results differ from the data from AD due to mutations in presenilin 1 and presenilin 2 genes and the amyloid precursor protein and suggest that the APOE ϵ4 genotype mediates increased Aβ deposition by a mechanism that differs from that found in other genetic causes of AD.
ALZHEIMER DISEASE (AD) is a progressive neurodegenerative disorder characterized in part by deposition of the amyloid β peptide (Aβ) at amino acids 39 through 43 in the brains of affected individuals. An early-onset familial form of AD has been found to cosegregate with mutations in 3 different genes: the presenilin 1 (PS1) gene on chromosome 14,1 the presenilin 2 (PS2) gene on chromosome 1,2 and the amyloid precursor protein (APP) gene on chromosome 21.3 Although the mechanisms by which these genetic defects exert their pathogenic effects are unknown, evidence from in vitro experiments suggests that the APP717 and PS1 and PS2 gene mutations alter APP processing such that an increased amount of long Aβ (or Aβ42) is produced.4- 7 Further support for the Aβ42 overproduction hypothesis comes from quantitative immunohistochemical studies of cerebrospinal fluid, and plasma that demonstrate a specific elevation of Aβ42 deposition in brain and plasma samples of individuals with APP and PS1 gene mutations.4,8- 11 Moreover, Aβ42 is the species deposited earliest in the disease process12,13 and has been shown to be more fibrillogenic in vitro than Aβ40.14 All together, there is compelling evidence that increased production of Aβ42 in the brain is critical for the initiation of early-onset familial AD.
In addition to the autosomal-dominant inherited gene defects associated with early-onset familial AD, inheritance of 1 allele of a common polymorphism of the APOE gene is associated with an increased risk of developing AD in the general population. Apolipoprotein E is present in 3 common alleles: ϵ2, ϵ3, and ϵ4. Inheritance of APOE ϵ4 is associated with increased risk of AD, whereas inheritance of APOE ϵ2 is associated with decreased risk of AD compared with the most common APOE ϵ3/3 genotype.15- 19 Like AD associated with APP, PS1, and PS2 mutations, APOE ϵ4 leads to a marked increase in Aβ deposition.15,19,20 However, in contrast to the specific elevation of Aβ42 deposits in the brains of individuals expressing mutations in PS1, PS2, or APP, it has been reported21,22 that the APOE ϵ4 allele is associated with increased amounts of Aβ40 and an increased Aβ40/Aβ42 ratio in the AD brain. This result suggests that the increased senile plaque frequency observed with an APOE ϵ4 allele is due to an increase in Aβ40-positive rather than Aβ42-positive senile plaques and suggests a pathogenic mechanism different from that of the AD-associated gene mutations in APP, PS1, and PS2.
To gain insight into the pathogenic mechanisms of AD associated with APOE ϵ4 and to test the hypothesis that AD associated with APOE ϵ4 follows a pathogenic route involving Aβ40 rather than Aβ42, we quantitated Aβ40 and Aβ42 immunostaining in relation to APOE genotype and duration of illness in 28 cases of sporadic AD.
Tissue samples from 28 cases of AD were selected on the basis of the APOE genotype status. Temporal lobe blocks were first fixed in paraformaldehyde lysine metaperiodate (Sigma Chemical Co, St Louis, Mo) for 24 to 48 hours and then placed in a cryoprotecting solution of 15% glycerin in 0.1-mol/L phosphate-buffered saline (pH, 7.4) overnight. The tissue was sectioned at 50 µm on a freezing sledge microtome and stored in sterile tubes containing cryoprotecting solution at −20°C until use.
Adjacent sections were immunostained using the Aβ antibody 10D5 (1:350 dilution) (Athena Neuroscience, South San Francisco, Calif)23 and C-terminal specific monoclonal antibodies to identify Aβ40 (1:50 dilution) (14C2; Athena Neuroscience) and Aβ42 (1:50 dilution) (21F12; Athena Neuroscience).9 Free-floating sections were pretreated with 70% formic acid (25°C, 10 minutes) and then with 0.01-mol/L citrate buffer (100°C, 10 minutes) to enhance staining. Three percent hydrogen peroxide containing 0.5% (vol/vol) of a mild detergent (alkylaryl polyether alcohol, Triton X-100, VWR Scientific, Boston, Mass) was applied to the sections for 20 minutes followed by 1 hour of protein blocking in 3% milk in Tris-buffered saline solution. Sections were then incubated overnight at 4°C in primary antibody and developed using horseradish peroxidase–linked secondary antibodies (Jackson Immunoresearch, West Grove, Pa).
The superior temporal sulcus region was chosen for morphometric analysis for several reasons. The superior temporal sulcus is anatomically unique because it is 1 of only 3 areas of association cortex identified in the monkey that receives afferent input from all sensory modalities.24 Thus, it is a higher order association cortex and is known from previous neuropathological studies to be severely and consistently affected in AD.25,26 In addition, the superior temporal sulcus has clearly defined boundaries and its structure is remarkably consistent across brains, decreasing potential anatomical variability among brains.
Amyloid deposition was quantified using an image analysis system (Bioquant, R and M Biometrics, Nashville, Tenn). Video images were captured and a threshold optical density was obtained to discriminate the staining from the background. Manual editing of each field eliminated artifacts, separated contiguous structures, and deleted vessel-associated staining. A strip of cortex approximately 1 cm from the crown of the gyrus on the inferior bank of the superior temporal sulcus measuring 700 mm wide by the depth of gray matter was chosen for analysis. The same area was analyzed on each of the 3 slides per case. The total percentage of cortical surface area covered by senile plaques positive for Aβ40, Aβ42, and total Aβ deposits was calculated for each case. The individual (M.J.M.) performing the quantitative analysis was unaware of the genotypes at the time of the analysis.
To determine any association between the amount of congophilic amyloid angiopathy (CAA) and Aβ40 or Aβ42 deposition in the neuropil, each case was scored for CAA by counting the number of times the crosshairs of a 10×10 block (525 µm2) grid overlapped a vessel with amyloid deposits. The grid was moved through the depth of cortex (the same area previously quantitated for amyloid burden) and the actual area (in micrometers) as well as the number of grids necessary to cover the area were recorded. A CAA score was obtained by dividing the number of "hit" crosshairs by the number of grids. A score higher than 1.0 was identified as severe CAA. Scores ranging from 0.5 to 1.0 were identified as moderate, and those less than 0.5 were mild CAA.
Amyloid deposits consisted primarily of Aβ42 species, with Aβ40 generally localized to the central core of amyloid plaques (Figure 1). Many plaques contained only Aβ42 deposits, whereas few plaques with only Aβ40 deposits were recognized. Image analysis results are presented for amyloid burden or percentage of cortical surface area covered by immunoreactivity.
We initially examined whether Aβ40 or Aβ42 deposits varied with the duration of AD. As shown in Figure 2, the Aβ42 amyloid burden is consistently higher than the Aβ40 amyloid burden at all points in the illness. The total amount of Aβ42 and Aβ40 deposits measured in this way did not increase with longer duration of illness in a consistent fashion. The weak correlation coefficients for these 2 measures compared with the duration of illness (R2=0.1 for Aβ42 and R2 =0.14 for Aβ40) suggest that the duration of illness is not a major confounder in determining either Aβ40 or Aβ42 deposition.
We selected the cases studied herein because they had specific APOE genotype (ϵ3/3, ϵ3/4, and ϵ4/4) and sufficient clinical records to determine the age of onset and the duration of illness. These 3 groups were matched in their age at death, although the APOE ϵ4/4 group had a younger age of onset and therefore a longer duration of illness (Table 1). Since duration of illness does not appear to affect the Aβ40/Aβ42 ratio, we directly compared the various APOE genotypes with one another. Analysis of variance showed a significant difference (P<.05) between the genotype groups in the Aβ40 measures. Post hoc tests (Fisher protected least significant difference) showed that Aβ40 deposition was significantly increased in the APOE ϵ4/4 group (P<.05). Similarly, analysis of variance suggested that the 3 groups also differed in the amount of Aβ42 deposits (P=.05), and the area occupied by Aβ42 deposits was increased in the APOE ϵ4/4 group (P=.02). The Aβ40/Aβ42 ratio was not statistically different among the 3 groups (range, P=.43-.56) but there was individual variation in this measure.
It is well established that Aβ40 is the predominant form of amyloid deposited in amyloid angiopathy.12 We then examined whether the amount of CAA assessed by the CAA score altered the relative deposition of Aβ40 and Aβ42 in the neuropil. As shown in Figure 3, there was no relationship between the CAA score and the Aβ40/Aβ42 ratio.
Deposits of Aβ42, with additional hydrophobic residues at the C-terminus, are more fibrillogenic than those of Aβ40.14 An increase in Aβ42 levels or the ratio of the 2 species is possibly a pathogenic event in AD mediated by PS1, PS2,5- 10 and APP mutations.4,11 Our current data test whether this is also the case for another genetic risk factor for AD, APOE ϵ4.
Our data do not support the hypothesis that levels of Aβ42 are specifically elevated in persons with APOE ϵ4. In fact, in this series, Aβ40 levels were also elevated in the APOE ϵ4/4 cases, and no statistically significant difference (P=.43) in the Aβ40/Aβ42 ratio was observed. This finding is in general agreement with previous reports showing elevated Aβ40 levels associated with APOE ϵ4 but contradicts these reports by demonstrating an elevation in Aβ42 levels as well.21,22,27
Certain caveats to this study should be noted. We directly measured the amount of immunodetectable Aβ, Aβ40, and Aβ42 in the neocortex; the exact relationship between these measures and amounts of soluble or formic acid–soluble Aβ species measured by enzyme-linked immunosorbent assay remains to be determined, although Ishii et al27 suggest that the results with immunostaining and enzyme-linked immunosorbent assay are comparable. In addition, only a limited anatomical area was examined, and immunohistochemical approaches must always be interpreted conservatively. Nonetheless, the data strongly suggest that a specific elevation in Aβ42 levels does not occur in association with APOE ϵ4. We postulate that the mechanism of enhanced deposition of total Aβ in APOE ϵ4 is different from the mechanism in APP, PS1, or PS2 mutations. In the latter circumstances, it has been suggested that the primary effect is an alteration of APP metabolism and increased synthesis of Aβ42. The current data are consistent with the possibility that the effect of APOE ϵ4 is to diminish clearance of Aβ,15 thus similarly affecting Aβ40 and Aβ42.
Accepted for publication January 22, 1998.
Supported by grants AG12406 and AG05134 from the National Institutes of Health, Bethesda, Md.
We thank the Alzheimer Disease Research Center Brain Bank (E. Tessa Hedley-Whyte, MD) for access to brain tissue and diagnostic information, Suzanne Sampson, BS, for assistance with the database, and Peter Seubert, PhD, and Dale Schenk, PhD (Athena Neurosciences, South San Francisco, Calif) for anti-Aβ monoclonal antibodies.
Corresponding author: Bradley T. Hyman, MD, PhD, Massachusetts General Hospital, Department of Neurology/Alzheimer's Unit, 149 13th St (CNY 6405), Charlestown, MA 02129 (e-mail: firstname.lastname@example.org).