Alzheimer Amyloid-β Peptide Forms Denaturant-Resistant Complex with Type ε3 but Not Type ε4 Isoform of Native Apolipoprotein E

Abstract

Background

The apolipoprotein E (apoE) type ε4 isoform specifies increased cerebral and cerebrovascular accumulation of amyloid-β protein (Aβ) and contributes to the genetic susceptibility underlying a large proportion (~60%) of typical, sporadic Alzheimer disease. Unfortunately, in vitro biochemical studies of direct apoE isoform-specific interactions with Aβ have been inconsistent, perhaps due to the use by different research groups of apoE isoform preparations in different conformational states (purified denatured versus native).

Materials and Methods

In the current study, we have investigated the possibility that synthetic Aβ^1–40 preferentially associates with native apoE of either the type ε3 or the type ε4 isoform.

Results

Here, we demonstrate the preferential association of synthetic Aβ^1–40 with native apoE ε3. The complex between apoE ε3 and Aβ^1–40 could not be disrupted by sodium dodecyl sulfate. In a parallel assay, no denaturant-resistant association of Aβ^1–40 with apoE ε4 was detectable.

Conclusions

These results support the notion that the apoE ε4 isoform may foster β-amyloidogenesis because apoE ε4 is inefficient in forming complexes with Aβ.

Introduction

The apolipoprotein E type ε4 allele contributes to the genetic susceptibility underlying a large proportion (~60%) of typical, sporadic Alzheimer disease (1–5). Correlations between genetic and neuropathological results indicate that the apolipoprotein E (apoE) type ε4 isoform specifies increased cerebral (6–10) and cerebrovascular (11–13) accumulation of amyloid-β protein (Aβ). Conversely, the apoE ε2 isoform can apparently prevent (14,15) the expression of clinical Alzheimer-type dementia, which is otherwise typically associated with amyloidogenic mutations in the amyloid-β protein precursor (AβPP) (16). The importance of apoE in human amyloid-β biology has been emphasized by studies indicating that apoE and Aβ are present as nondissociable complexes within the cerebral amyloid plaques characteristic of this disease (17,18). Therefore, it is important to search for relevant interaction(s) between Aβ and various isoforms of apoE.

Currently proposed mechanisms for the proamyloidogenic activity of apoE ε4 include the possibility that apoE ε4 might have a relatively greater propensity (compared with apoE ε3) for modulating Aβ misfolding and fibrillogenesis (19–23), although we recently observed equipotent profibrillogenic activities for native apoE ε3 and ε4 isoforms (Sweeney et al., submitted for publication). Alternatively, apoE ε4 might be inferior to apoE ε3 in some bioactivity related to modulation of Aβ clearance (7,24). Since both Aβ and apoE are secreted proteins, either of these mechanisms might be attributable to direct apoE/Aβ interactions (2,24) in the extracellular space. To date, however, the results of in vitro biochemical studies of direct interactions between Aβ and different apoE isoforms have been inconsistent, perhaps due to the use by different research groups of either purified, denatured (2) or native (24) apoE isoforms. In view of the great importance of clarifying the molecular mechanisms underlying the role of apoE in human amyloid-β biology, we have in the current study compared the ability of native apoE ε3 and ε4 isoforms to associate with synthetic Aβ^1–40.

Materials and Methods

Aβ^1–40 was prepared at the Keck Foundation Protein Facility at Yale University (New Haven, CT, U.S.A.). RAW264 mouse macrophage cells were obtained from American Type Culture Collection (Rockville, MD, U.S.A.). ApoE serum standards were kindly provided by Dr. Petar Alaupovic of the Oklahoma Medical Research Foundation (Oklahoma City, OK, U.S.A.).

Transfected Cells

RAW264 cells were stably transfected with apoE isoforms using the SV40 early promoter to drive expression of genomic apoE DNA fragments which extended from the AatII site in exon 1 to the EcoRI site 626 bp 3′ to the end of the apoE gene (25). The initial clone contained the apoE ε4 allele, and the apoE ε3 clone was created by exchange of two 2-kb genomic EcoRI cassettes. Stably expressing lines were prepared by cotransfection of apoE with pUCneo. Geneticin (Life Technologies, Inc. Gaithersburg, MD, U.S.A.) resistant colonies were expanded and screened for apoE production by in situ immunocytochemistry of samples of colonies, using a commercial apoE antibody (INCSTAR Stillwater, MN, U.S.A.).

Preparation of Conditioned Media

ApoE isoform-transfected RAW264 cells were used to generate conditioned media containing native apoE isoforms. Conditioned media containing apoE were prepared using Dulbecco’s minimal essential medium, which in some cases was supplemented with 0.2% (w/v) bovine serum albumin. Media were harvested after a 24-hr conditioning period, and apoE isoform levels were determined by quantitative immunoblotting of conditioned media and serum standards containing apoE (Fig. 1). ApoE concentrations were adjusted to 20 µg/ml using medium conditioned by untransfected RAW264 cells as diluent. These apoE concentrations are within the range reported for human cerebrospinal fluid (26). Aliquots were stored at — 120°C until use.

Fig. 1

Quantitative immunoblotting for apoE content of conditioned media

Anti-apoE antibody was used to probe electroblotted conditioned medium proteins following SDS-PAGE in a 10% gel. Lanes 1–6, human plasma apoE standards of 10 ng (Lane 1), 20 ng (Lane 2), 50 ng (Lane 3), 100 ng (Lane 4), 200 ng (Lane 5), and 400 ng (Lane 6); Lane 7, conditioned medium from untransfected RAW264 cells; Lane 8, conditioned medium from apoE ε3-transfected RAW264 cells; Lane 9, conditioned medium from apoE ε4-transfected RAW264 cells.

Detection of ApoE/Aβ^1–40 Complexes by Immunoblotting

ApoE/Aβ^1–40 complexes were studied using a modification of the procedure of LaDu et al. (24). Synthetic Aβ^1–40 was freshly prepared as a 5.0 mg/ml stock solution in distilled water. Aliquots of this stock solution were added to conditioned medium samples containing various amounts of either apoE ε3 or apoE ε4, yielding a final Aβ^1–4 concentration of 0.5 mg/ml (~100 µM). The reaction mixtures were incubated at room temperature for 2 hr, and then aliquots were removed and analyzed by SDS-PAGE in a 10–20% tricine gel under nonreducing conditions. ApoE/Aβ^1–40complexes were detected by immunoblotting with either antibody 6E10 against Aβ^1–17(Senetek, Maryland Heights, MO, U.S.A.) or anti-apoE antibody (INCSTAR).

Results

Preferential association of Aβ^1–40 with native apoE ε3 was demonstrated using an experimental protocol based on immunoblotting of denaturant-resistant apoE/Aβ complexes, as previously described by LaDu et al. (24). In the current study, conditioned media containing native apoE isoforms were derived from stably transfected RAW264 macrophage cells. Human kidney 293 cells were used in the study by LaDu et al. (24).

Figure 2 illustrates typical results obtained from one of three independent experiments. Aβ^1–40 was incubated with either conditioned medium from control untransfected RAW264 cells (Fig. 2, Lanes 1 and 4), apoE ε3-transfected RAW264 cells (Fig. 2, Lanes 2 and 5), or apoE ε4-transfected RAW264 cells (Fig. 2, Lanes 3 and 6). Following incubation, aliquots were removed, mixed with an equal volume of nonreducing Laemmli buffer, and boiled for 3 min prior to loading and separation by SDS-PAGE in a 10–20% tricine gel. Separated proteins were then electroblotted to nitrocellulose and probed with either antibody 6E10 (Fig. 2, left panel) against Aβ^1–17 or anti-apoE antibody (Fig. 2, right panel).

Fig. 2

Comparison of ability of native apoE ε 3 and ε 4 to form a denaturant-resistant complex with synthetic A β ^1–40

Aβ^1–40 (0.5 mg/ml) was added to each of the three types of conditioned media. For media containing apoE isoforms, apoE concentration was adjusted to 20 µg apoE/ml as described in Materials and Methods. Incubations were carried out for 2 hr at room temperature. Aliquots of the reaction mixtures were separated by SDS-PAGE in a 10–20% tricine gel, electroblotted to nitrocellulose and probed with either anti-Aβ^1–17 antibody 6E10 (left panel) or anti-apoE antibody (right panel). Lanes 1 and 4, conditioned medium from untransfected RAW264 cells; Lanes 2 and 5, conditioned medium from apoE ε3-transfected RAW264 cells; Lanes 3 and 6, conditioned medium from apoE ε4-transfected RAW264 cells. Arrowhead indicates the location of apoE ε3/Aβ complex which migrates as a discrete band in Lane 2 and as the slowly migrating component of the apoE ε3 band in Lane 5. Bracket indicates position of free Aβ^1–40, visible in Lanes 1–3. Markings in center of figure represent positions of molecular weight standards of 200, 97.4, 69, 46, 30, 21.5, and 14.3 kD, respectively, from top to bottom.

Free Aβ^1–40 was visible as a diffusely migrating, 6E10-immunoreactive species at the bottom of the gel, near the bromphenol blue tracking dye indicator (Fig. 2, left panel). Using antibody 6E10, an apoE/Aβ^1–40 complex (Fig. 2, arrowhead) was detectable following co-incubation of Aβ^1–40 with apoE ε3 (Fig. 2, Lane 2), but not following co-incubation of Aβ^1–40 with apoE ε4 (Fig. 2, Lane 3). As expected, apoE ε3 and ε4 were detected only in media from transfected RAW264 cells (Fig. 2, Lanes 5 and 6). An apoE ε3/Aβ^1–40 complex was also visible as the slowly migrating component of the apoE ε3 band, as detected with the anti-apoE antibody (Fig. 2, Lane 5). ApoE ε4/Aβ^1–40 complexes were not detectable in any of three identical independent experiments.

Discussion

The present results confirm the observations of LaDu et al. (24), who also used native apoE isoforms. The results obtained by LaDu et al. (24), using kidney 293 cells, and in the present study using RAW264 macrophage cells, differ from those obtained by Strittmatter et al. (2), who described preferential association of apoE ε4 with Aβ. The latter results are most likely attributable to the use of purified, delipidated, denatured apoE isoforms (27). Our RAW264-produced apoE isoforms are lipid-poor (6.0 ± 2.6 µgphospholipid/100 µg apoE), suggesting that denaturation of apoE during purification (27) is the most likely explanation for the results obtained by Strittmatter et al. (2).

Since apolipoprotein E ε4-related Alzheimer disease is the most prevalent molecular subtype of the disease identified to date (4,5), an elucidation of the molecular pathology resulting from the apolipoprotein E ε4 genotype is particularly crucial. Along this line, the ability of apoE ε4 to promote cerebral and cerebrovascular accumulation of Aβ (6–13) would appear to be an important clue, especially considering that genetically heterogeneous forms of Alzheimer disease foster Aβ accumulation via at least three independent amyloidogenic pathways: excess Aβ precursor expression (28), excess Aβ generation (29), or generation of aberrant, excessively aggregable Aβ (16,30).

Studies of Aβ levels and plaque accumulation in affected individuals support the notion that the role of apoE ε4 in β-amyloidogenesis might involve a deficiency of Aβ clearance. Soluble Aβ levels in human cerebrospinal fluid are apparently unaffected by the apoE ε4 allele (31). Parenchymal amyloid plaque density (plaques/mm²) is increased by apoE ε4, while plaque size is increased in Down syndrome (32). These data are consistent with a proposed model (32) in which elevations in soluble Aβ concentration (as would be predicted to occur in Down syndrome) lead to enlarged plaque size whereas the apoE ε4 “effect” leads to an increased frequency of plaque initiation. The current data demonstrating differences in the association of specific apoE isoforms with Aβ indicate the importance of assessing the role of the apoE ε3/Aβ complex in the fate of Aβ. It will be important to determine the fate of both extracellular and cell-associated Aβ. The use of primary brain cultures as well as in vivo approaches should be of value in determining the importance of apoE in Aβ clearance.

Alternatively, with regard to the molecular basis of cognitive decline in Alzheimer disease, apoE ε4 may well be less efficient than apoE ε3 in exerting important neurobiologic effects via an as yet unclarified role in neuronal or synaptic plasticity (33). This possibility is supported by recent evidence that apoE-deficient mice exhibit learning and memory deficits (34,35), even in the absence of consistent obvious structural neuropathology (35). It is encouraging to note that many of the proposed actions for apoE isoforms are experimentally testable. Data from such experiments may provide valuable opportunities for slowing or preventing the clinical course of Alzheimer disease.