Apolipoprotein E isoforms and apolipoprotein AI protect from amyloid ...

Journal of Neurochemistry, 2004, 91, 1312–1321

doi:10.1111/j.1471-4159.2004.02818.x

Apolipoprotein E isoforms and apolipoprotein AI protect from amyloid precursor protein carboxy terminal fragment-associated cytotoxicity Izumi Maezawa,* Lee-Way Jin,* Randall L. Woltjer,* Nobuyo Maeda, George M. Martin,* Thomas J. Montine* and Kathleen S. Montine* *Department of Pathology, University of Washington, Seattle, Washington, USA Department of Pathology, University of North Carolina, Chapel Hill, North Carolina, USA

Abstract Inheritance of the apolipoprotein (APO) E gene e4 or e2 allele alters the risk of developing Alzheimer disease (AD), while increased alpha-tocopherol (AT) intake appears to lower the risk of AD. As APOE is a major apolipoprotein in the CNS and AT in vivo is transported in lipoproteins, we tested the hypothesis that CNS lipoproteins, as modeled by relevant concentrations of high density lipoprotein (HDL), and AT would interact to suppress neurotoxicity in a cell culture model of amyloid beta (Ab)- related toxicity. These cells conditionally express C99-derived peptides, proposed to be a key step in AD pathogenesis; this expression is closely associated with subsequent cell death. We found that physiologic

concentrations of lipoproteins present in the CNS protected from C99-associated toxicity and provided evidence for two mechanisms of protection. The first was AT-independent, APOE isoform-dependent, and most potent for the APOE2 isoform. The second was a synergistic protection afforded by a combination of APOAI, or less so APOE, and AT. These data provide a novel explanation for the apparent AD-protective effect of inheriting an e2 APOE allele, and suggest that optimizing AT enrichment of CNS lipoproteins or devising APOAI mimetics may augment AT efficacy in treating AD. Keywords: alpha-tocopherol, apolipoproteins, APO, C99, HDL. J. Neurochem. (2004) 91, 1312–1321.

Alzheimer’s disease (AD) is an adult–onset neurodegenerative syndrome that can be inherited as a rare autosomal dominant trait or more commonly appears sporadically, probably as the confluence of genetic susceptibilities and environmental factors. Substantial evidence indicates that the etiology of AD is related to proteolytic products from the C-terminus of the amyloid precursor protein (APP) (Selkoe 2002; Walsh et al. 2002). Cleavage of APP by b-secretase generates C99, a peptide consisting of the 99 C-terminal amino acids of APP. Further proteolytic cleavage produces a number of peptides, several of which, as with C99 itself, have been shown to be neurotoxic in vitro and in vivo (OsterGranite et al. 1996; Fraser et al. 1997; McPhie et al. 1997; McPhie et al. 2001). Which species is the primary neurotoxin in vivo remains to be clarified, although greatest attention has been focused on one particular family of C99 fragments, the amyloid beta (Ab) peptides, especially Ab42. Much of the data examining the neurotoxicity of Ab peptides has been performed in cell culture with extracellular application of micromolar concentrations of artificially

aggregated synthetic peptides, likely yielding an array of assembly forms that may or may not reflect what occurs

1312

Received May 15, 2004; revised manuscript received August 13, 2004; accepted August 13, 2004. Address correspondence and reprint requests to Kathleen S. Montine, Department of Pathology, University of Washington, Box 359645, Harborview Medical Center, 325 9th Ave, Seattle, WA 98104, USA. E-mail: [email protected] Abbreviations used: AD, Alzheimer’s disease; APO, apolipoprotein; APP, amyloid precursor protein; AT, alpha-tocopherol; Ab, amyloid beta; CM, conditioned medium; DMPC, dimyristoylphosphatidylcholine; HDL, high density lipoprotein; HDL-ATsat, HDL enriched with 0.28 nmoles AT/lg HDL; HDLfixed-AT, HDL at 0.75 lg/mL with unspecified AT enrichment; HDLfixed-AT1/10sat, HDLfixed enriched with 0.028 nmoles AT/lg HDL; HDLfixed-AT1/2sat, HDLfixed enriched with 0.14 nmoles AT/lg HDL; HDLfixed-ATsat, HDLfixed enriched with 0.28 nmoles AT/lg HDL; HMWC, high molecular weight complexes; LDL, low density lipoprotein; PBS, phosphate-buffered saline; Tet, tetracycline; Tet–, cell culture medium without Tet; Tet+, cell culture medium containing 1 lg/mL Tet; tHDL, trypsinized HDL; TR, targeted replacement; 6E10, monoclonal antibody directed against Ab1)17; 4G8, monoclonal antibody directed against Ab17)24.

2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1312–1321

Apolipoprotein protection from C99 toxicity 1313

in vivo at low nanomolar concentrations (Selkoe 2002). While high concentrations of exogenously applied aggregated synthetic Ab peptides may test the hypothesis that C99-derived peptides are neurotoxic through the formation of large insoluble extracellular aggregates, it does not address the more recent proposal that small soluble aggregates that accumulate intraneuronally may be the ultimate neurotoxic species (Selkoe 2002; Walsh et al. 2002). MC65 cells are one means to address this limitation; these human neuroblastoma cells are stably transfected with a C99 expression vector under the control of a tetracycline (Tet) sensitive repressor (Fukuchi et al. 1992a; Fukuchi et al. 1992b). Previous reports from Dr Martin’s laboratory have demonstrated an association between C99 expression and cytotoxicity in these cells (Fukuchi et al. 1992a; Fukuchi et al. 1992b; Fukuchi et al. 1993; Sopher et al. 1994; Sopher et al. 1996; Jin et al. 2002). Specifically, removal of tetracycline results in initial expression of C99, a 10 kDa protein that is reactive with antibodies directed against Ab peptide. Subsequent processing yields a smaller protein that comigrates with Ab1)40, but not Ab1)42, on western blots. This 4 kDa protein, although not completely characterized, is reactive with the 6E10 antibody that recognizes the amino-terminal 17 residues of Ab and undergoes progressive aggregation to detergent-soluble 6E10-immunoreactive high molecular weight complexes (HMWCs); formation of these HMWCs is strongly associated with cell death in this model (Fukuchi et al. 1992a, Fukuchi et al. 1992b, Fukuchi et al. 1993; Sopher et al. 1994; Sopher et al. 1996; Jin et al. 2002; Woltjer et al. 2003). In addition to C99-derived peptides, apolipoprotein (APO) E isoforms also appear to play a critical role in AD pathogenesis. Humans possess three common alleles of the APOE gene (APOE), e2, e3, and e4, which are expressed as APOE2, APOE3, and APOE4 isoforms, respectively (Rall et al. 1982). Numerous genetic association studies have observed that inheritance of the e4 APOE allele imparts a gene dosage-dependent increased risk and an earlier onset for developing AD (Saunders et al. 1993). CNS APOE is largely produced in astrocytes, secreted on nascent lipoproteins, and remains distinctly separated from peripheral APOE (Linton et al. 1991; LaDu et al. 1998). These nascent APOEcontaining particles circulate within the CNS and ultimately form one of the two major CNS lipoproteins, an APOEbearing high density lipoprotein (HDL)-like particle; the other major CNS lipoprotein is an APOAI-bearing HDL-like particle (Roheim et al. 1979; Pitas et al. 1987; Borghini et al. 1995). Elegant experiments with double transgenic mice have shown that the increased risk for AD associated with APOE4 is related, at least in part, to interactions between this protein and Ab peptides in the formation of neuritic plaques (Bales et al. 2002). Interestingly, genetic association studies have also observed that inheritance of the e2 APOE allele is associated with a decreased risk and later

onset for AD (Corder et al. 1994); however, the mechanisms that underlie this apparent neuroprotective effect of APOE2 are not clear. Current therapies commonly used to treat AD include centrally acting cholinesterase inhibitors, which temporarily improve cognitive function, and alpha-tocopherol (AT), which presumably acts to suppress the now well-documented increased lipid peroxidation to brain that occurs at all phases of AD (reviewed in Montine et al. 2002). Evidence for AT efficacy in treating AD includes a clinical trial where dietary supplementation with AT provided modest benefit to patients with established AD, and some, but not all, observational studies that associated dietary supplementation with vitamins E and C or diets high in these antioxidants with reduced risk of developing AD (Sano et al. 1997; Zandi et al. 2004). In support of a neuroprotective role for AT in AD pathogenesis, cell culture experiments have shown that free AT added to medium efficiently suppresses C99-associated toxicity in Tet– MC65 cells (Sopher et al. 1996), as well as in other models of Ab peptide-associated neurotoxicity that use exogenously applied aggregated synthetic Ab (Butterfield et al. 1999; Yatin et al. 2000); these results have been interpreted as supporting the proposal that lipid peroxidation is a major effector of neurotoxicity from C99 and its proteolytic products. A potentially important caveat is that all these cell culture experiments involved the addition of free AT directly to medium, while in vivo virtually all of circulating AT is bound to lipoproteins. From these data we decided to test the hypothesis that CNS lipoproteins modulate the protective effect of AT in AD, perhaps through APOE-dependent mechanisms in the MC65 model of intracellular C99-associated cell death.

Materials and methods Materials Human HDL and low density lipoprotein (LDL), APOE purified from human plasma, and antihuman APOAI and APOE monoclonal antibodies were purchased from Chemicon International (Temecula, CA, USA). Human recombinant APOE isoforms were purchased from Panvera (Madison, WI, USA). Sequencing-grade trypsin was purchased from Promega (Madison, WI, USA). (+/–) AlphaTocopherol and dimyristoylphosphatidylcholine (DMPC) were purchased from Sigma (St Louis, MO, USA). Monoclonal antibody 6E10 (directed against Ab1)17) was purchased from Senetek (Napa, CA, USA). Monoclonal antibody 4G8 (directed against Ab17)24) was purchased from Signet Laboratories (Dedham, MA, USA). Preparation of HDL-AT particles AT was dissolved in ethanol and mixed with phosphate-buffered saline (PBS) to make a 1 mM stock. HDL from the manufacturer was mixed with PBS to make a 200 lg/mL suspension. One milliliter of the HDL suspension, 50 lL of 1 mM AT, and 50 lL PBS were incubated for 1 h at 37C with gentle shaking. To confirm


1314 I. Maezawa et al.

that AT was incorporated into HDL, a small portion of the mixture was centrifuged using a 50 kDa MWCO Microcon filter (Millipore) at 14 000 r.p.m., 4oC in a 5145C Eppendorf centrifuge (Brinkmann Instrument Inc) for 2 min. Both flow-through and liquid remaining on the column were subjected to AT measurement by HPLC as described below. HDL particles were incubated with increasing concentrations of free AT until maximal enrichment was achieved at 0.28 nmol AT/lg HDL; at higher ratios AT became detectable in the flow-through. Trypsinized HDL (tHDL) was prepared according to the method of Mendez et al. (Mendez and Oram 1997). Briefly, HDL dialyzed against 0.1 M NH4HCO3 at 4C for 24 h was incubated with trypsin (30 min at 37C, HDL–trypsin protein ratio of 40 : 1), with trypsinization terminated by dialysis against PBS containing 0.1 mM phenylmethylsulfonyl fluoride for 24 h at 4C. tHDL was subjected to 4–15%Tris/HCl SDS gel electrophoresis and analyzed by silver stain and western blotting with anti-APOAI and anti-APOE antibodies. Cholesterol concentrations were determined with total cholesterol reagent (Boehringer Ingelheim) following the manufacturer’s protocol. MC65 cell culture and cell viability MC65 cells were cultured as described previously (Sopher et al. 1994; Woltjer et al. 2003). For cell viability experiments, cells were plated onto 96 well plates at 2 · 104 cells/well and the indicated treatments applied in 200 lL to each well. For western blot analysis and intracellular AT measurement, cells were plated onto 60-mm dishes at 2.5 · 106 cells/dish. To perform parallel experiments using both size plates, equivalent treatments were based on moles or micrograms per 105 cells to control for differences in culture medium volume and cell number in different size wells. n ‡ 3 for all assays and experiments were repeated on separate days. Before incubation with lipoproteins, cells were washed with PBS and OptiMEM (Gibco/BRL) once each and then incubated with Opti-MEM without Tet and the indicated compound. Cell viability was determined using MTT (tetrazolium salt) as previously described (Woltjer et al. 2003) with the average viability of parallel Tet+ cells set as 100% for each experiment. Confocal immunofluorescence microscopy Immunofluorescence labeling and confocal microscopy for Ab-immunoreactive aggregates were performed according to the method of Vuletic et al. (2003). Briefly, MC65 cells were plated onto Laboratory-Tec Chamber slides eight-well (Nunc, Naperville, IL, USA) at 2 · 104 cells/well in Opti-MEM with or without Tet. Indicated amounts of HDL or HDL-ATsat were added, and 24 h later cells were fixed in 4% paraformaldehyde for 10 min. The cells were permeabilized with 0.03% saponin, followed by incubation with primary antibody (biotinylated 4G8 1 : 200) for 2 h. After incubation with AlexaFluro 488-conjugated streptavidin for 30 min the cells were mounted in Vectorshield mounting media (Vector Laboratory, Burlington, CA, USA) containing DAPI for nuclear label. Images were examined with a Leica DM IRBE confocal microscope and captured using a Leica TCSSP camera and Leica confocal software. Primary astrocyte culture Homozygous APOE2, APOE3 and APOE4 targeted replacement (TR) mice ‘humanized’ at apoE were developed by Dr Maeda

and colleagues (Sullivan et al. 1997; Sullivan et al. 1998). These mice were backcrossed six generations to C67BL/6 genetic background. Primary astroctye cultures were prepared from the cerebral cortices of postnatal day 1 APOE TR pups according to the method of Furukawa et al. (1986) and as approved by the University of Washington IACUC. Briefly, cortices were removed, placed into Hank’s Balance Salt Solution, the meninges removed under a dissecting microscope, and tissues minced into small pieces (1 mm3). After washing with DMEM, tissues were incubated with DMEM digestion buffer containing 0.5 mM EDTA, 0.2 mg/mL L-cysteine, 30 U/mL papain, and 200 lg/mL Dnase I for 30 min at 37C with gentle shaking. After centrifugation at 235 g for 10 min, supernatant was carefully aspirated and tissues were dissociated by gentle trituration in DMEM containing 10% fetal bovine serum and 100 lg/mL penicillin-streptomycin. The dissociated cells were plated into T75 flasks coated with 0.001% poly-ornithine (one half brain per flask) and medium was changed the next day. For experiments, cells were plated onto 24 well plates at 5 · 105 cells/well in Opti-MEM and 24 h later conditioned medium (CM) was collected and centrifuged at 14 000 r.p.m. for 2 min at 4C. AT concentration was determined as described in the next section. AT measurement by HPLC To determine the change in intracellular AT concentration following various treatments, 2.5 · 106 cells were plated onto a 60-mm dish in Opti-MEM without Tet. Twenty-four hours after Tet removal, medium was changed and cultures incubated with free AT, HDL or HDL-AT for 1 h at 37C with gentle shaking. Cells then were washed twice with ice-cold PBS, resuspended in PBS, and sonicated. AT levels were quantified by reverse-phase HPLC according to the method of Bieri et al. (1979). Five microliters of internal standard (1 mM AT acetate) was added to a mixture of sample and methanol (1 : 1 v/v) and then vortexed for 30 s. Three volumes of HPLC grade hexane was added, and the mixture vortexed vigorously for 2 min and separated into phases. The organic phase was transferred to another tube and evaporated under a stream of nitrogen in a 37C water bath. The residue was reconstituted in ethanol, and 30 lL of sample injected onto the HPLC. The HPLC system consisted of a Shimadzu LC-10ADvp pump, a C18 reverse phase column and a Shimadzu RF-10Axl fluorescence detector. Samples were eluted with 100% methanol at a flow rate of 1 mL/min. The fluorescence detector was set to monitor wavelengths of Em ¼ 292 nm and Ex ¼ 335 nm. Western blotting Twenty-four hours after treatment, MC65 cells were harvested and washed with ice-cold PBS twice and homogenized in 50 mM TrisHCl, pH 7.4, 150 mM NaCl, 0.1 mM EDTA, 0.32 M sucrose in the presence of protease-inhibitor cocktail, followed by dilution and boiling in 2 · Laemmli buffer. Protein concentration was determined by Bio-Rad protein dye reagent. 5 lg of cell protein was subjected to bicine gel electrophoresis (Jin et al. 2002; Woltjer et al. 2003). Proteins were transferred to PVDF membranes and probed with 6E10 (1: 2000), a monoclonal antibody to Ab1)17. Western blotting of APOE and APOAI was performed as previously described by us (Montine et al. 1998). Recombinant APOE was used to construct a standard curve for estimation of APOE



concentration in conditioned medium (CM) using methods similar to those previously described by us (Montine et al. 1999). Statistical methods Data were analyzed by the tests and approaches specified using GraphPad Prism software (San Diego, CA, USA).

Results

Cell viability was determined in these experiments using an MTT conversion assay; we have shown previously that the LIVE/DEAD assay gives similar results (Woltjer et al. 2003). As shown previously, MC65 cells cultured in the presence of 1 lg/mL Tet (Tet+) did not show any significant loss of viability over 72 h, whereas following the removal of Tet (Tet–) there was a delayed time-dependent loss of viability that began about 48 h after Tet removal and culminated in 90–95% reduction in viability after 72 h (vida infra) (Sopher et al. 1996; Woltjer et al. 2003). In excellent agreement with previous work, free AT showed significant concentrationdependent cytoprotection in Tet– MC65 cultures from 0.4 to 3.4 lM (R2 ¼ 0.99, slope ¼ 21.4 ± 0.6% viability/lM AT, y-intercept ¼ 9.9 ± 0.8% viability), resulting in approximately 80% viability at the highest concentration of AT (Sopher et al. 1996); AT less than 0.4 lM was not detectably cytoprotective in Tet– cells. HDL was enriched with saturating amounts of AT (HDLATsat) as described in Methods with a final ratio of 0.28 nmol AT/lg HDL; others have obtained comparable levels of AT-enrichment in HDL (Laureaux et al. 1997). When HDLATsat was added at the time of Tet removal (t ¼ 0), significant concentration-dependent cytoprotection was observed at 72 h (Fig. 1). HDL-ATsat had an EC50 of approximately 0.6 lg/mL HDL and gave complete protection at 3 lg/mL HDL; expressed in terms of AT concentration in these particles, this corresponds to an EC50 of 0.2 lM AT and complete protection at 0.8 lM AT. In contrast, the same concentrations of free AT were approximately 10-fold less protective (Fig. 1) with a calculated EC50 of 1.9 lM (p < 0.01 at all points). These data demonstrated that ATenriched HDL is much more effective at protecting Tet– MC65 cells than the same concentrations of free AT. Interestingly, this protective effect appeared to be partially due to HDL, as HDL alone also showed significant cytoprotection with an EC50 of approximately 2.6 lg/mL HDL and complete protection at 6 lg/mL HDL (Fig. 1), corresponding to an apolipoprotein concentration of approximately 3 lg apolipoprotein/mL. For comparison, the protein concentration of HDL-like particles in human CSF is approximately 10 lg/mL (Roheim et al. 1979) (p < 0.01 for all points except 0 and 6 lg/mL HDL). This protective effect was specific to HDL as neither LDL nor DMPC offered any significant cytoprotection to Tet– cells at up to 5 lg/mL (p > 0.05). Indeed, incubation for 72 h with LDL

Fig. 1 Tetracycline (Tet) was removed from MC65 cells and the indicated concentrations of HDL or HDL enriched with saturating amounts of AT (HDL-ATsat, 0.28 nmoles/lg HDL) added at time 0. Free AT was added to parallel cultures at the concentrations indicated on the upper axis. After 72 h, viability was assessed by MTT assay. Data are expressed as mean percentage viability ± SEM (n ¼ 6) with parallel Tet+ cultures set at 100% viability. Absence of a visible error bar is due to an SEM smaller than the data symbol. EC50s for HDLATsat (as AT concentration) and free AT were 0.2 lM and 1.9 lM AT, respectively. Two-way ANOVA for percentage viability had p < 0.0001 for free AT versus HDL-ATsat and AT concentration; corrected repeated pair comparisons showed that HDL-ATsat was significantly greater than free AT (p < 0.01) at all points (0.05–1.7 lM AT). EC50 values for HDL-ATsat (as HDL concentration) and HDL were 0.6 and 2.6 lg/mL HDL, respectively. Two-way ANOVA for percentage viability for cells protected with HDL or HDL-ATsat had p < 0.0001 for protectant and concentration with Bonferroni-corrected repeated pair comparisons having p < 0.01 for all points except 0 and 6 lg/mL HDL. There was no effect of AT, HDL or HDL-ATsat on viability in Tet+ cultures (not shown).

was even toxic to Tet+ cells; while 1 lg/mL LDL did not reduce viability, 2.5 lg/mL resulted in viability of only 6.2 ± 1.6% in Tet+ cells. In contrast, HDL showed no toxicity to Tet+ cells; viability after 72 h with 5 lg/mL HDL was 108.2 ± 2.8%. Numerous reports have shown by western blotting that accumulation of intracellular Ab-immunoreactive HMWCs precedes and is closely associated with cell death in Tet– MC65 cells (Fukuchi et al. 1992a; Fukuchi et al. 1992b; Fukuchi et al. 1993; Sopher et al. 1994; Sopher et al. 1996; Jin et al. 2002; Woltjer et al. 2003); although the composition of these HMWCs remains to be fully elucidated, the cells accumulate Ab1)40 but not Ab1)42 upon Tet removal (Jin et al. 2002; data not shown). We confirmed this association as part of the present study (as illustrated in Fig. 5, lane 2 below) and expanded this characterization using immunofluorescence microscopy to visualize cytoplasmic Ab-immunoreactive deposits (Fig. 2). As expected, Tet+ cells had no visible Ab-immunoreactivity (Fig. 2a), while 24 h after Tet removal cells showed abundant cytoplasmic Ab-immunoreactive aggregates (Fig. 2b). The addition of 2 lg/mL HDL-ATsat at the time of Tet removal, resulting in complete cytoprotection (see Fig. 1), also resulted in complete suppression of Ab-immunoreactive aggregate formation (Fig. 2c), while the same concentration of HDL,



Fig. 3 Equal amounts of HDL or tHDL (0.5 lg cholesterol/mL, corresponding to approximately 4 lg/mL HDL), 0.5 lg/mL DMPC, 2.5 lg/mL of purified apolipoprotein (APO) AI or E, or 2.5 lg/mL of recombinant APOE2, APOE3, or APOE4 was added to MC65 cells upon removal of Tet and viability measured 72 h later by MTT assay. Data are expressed as mean percentage viability ± SEM (n ‡ 3) with parallel Tet+ cultures set at 100% viability. ANOVA had p < 0.0001 with Bonferroni-corrected repeated pair comparisons having *p < 0.001 versus untreated control; all corrected paired comparisons with HDL had p < 0.001. Fig. 2 Twenty-four hours after Tet removal and treatment with the indicated compounds, MC65 cells were fixed, permeabilized, and labeled with 4G8, a monoclonal antibody to Ab17)24. Antibody binding was visualized with avidin-Alexa488 conjugate using DAPI counter stain followed by confocal microscopy. (a) Tet+, (b) Tet–, (c) 2 lg/mL HDL-ATsat, (d) 2 lg/mL HDL, (e) 6 lg/mL HDL-ATsat, and (f) 6 lg/mL HDL.

corresponding to approximately 25% viability (see Fig. 1), only partially prevented Ab-immunoreactive aggregate formation (Fig. 2d). HDL-ATsat and HDL at 6 lg/mL, both of which afforded complete cytoprotection, also completely ablated Ab-immunoreactive aggregate formation (Fig. 2e,f). These data suggested that the processes by which HDL and HDL-AT protected Tet– MC65 cell viability also suppressed Ab-immunoreactive aggregate formation. In order to determine whether the effect of HDL alone was due to endogenous AT in unmodified HDL particles, we measured the AT concentration of unmodified HDL particles, as well as the intracellular AT concentration of untreated cells and cells incubated with up to 6 lg/mL HDL for 1 h. The AT concentration in unmodified HDL was 6 pmol/lg, or approximately 1/50 that of HDL-ATsat particles. In addition, there was no significant difference in the intracellular AT concentration between untreated MC65 cells and those treated with 6 lg/mL HDL for 1 h (0.07 ± 0.04 and 0.12 ± 0.04 pmoles/105 cells, respectively, p > 0.05). As HDL-mediated protection was independent of endogenous AT delivery to cells, we examined which components of HDL provided the observed cytoprotective effect (Fig. 3). HDL incubated with trypsin sufficient to ablate immunoreactivity for APOE and APOAI (not shown) largely reduced cytoprotection (Fig. 3), suggesting that apolipoproteins were

critically important to the protection of Tet– MC65 cells. Similar results were obtained by denaturation of HDL by boiling (not shown). To determine whether cytoprotection was provided by the apolipoproteins themselves, we measured viability in cells treated with 2.5 lg/mL purified APOE or APOAI, concentrations close to those of CSF (Fig. 3). APOE provided modest but statistically significant cytoprotection (p < 0.001) while APOAI did not. As the purified APOE used was derived from pooled human plasma and thus represented a mixture of APOE isoforms, we next tested recombinant APOE2, APOE3, and APOE4 at the same concentration. Recombinant APOE2 duplicated the effect of purified APOE; however, neither APOE3 nor APOE4 significantly protected Tet– MC65 cells (Fig. 3). Although both purified APOE and recombinant APOE2 provided significant cytoprotection in MC65 cells following Tet removal, these effects were relatively small compared with the effects of HDL alone. An important limitation to using purified APOE or recombinant APOE isoforms is that the protein is lipid-poor and does not resemble CNS APOE produced in astrocytes and secreted in nascent lipid particles (LaDu et al. 1998). As one way to overcome this limitation, we prepared primary cultures of astrocytes from apoEtargeted replacement (TR) mice ‘humanized’ at the apoE locus to APOE2, APOE3, or APOE4 and collected serumfree conditioned medium (CM) as a source of nascent APOEbearing particles; albeit as part of an ensemble of secreted factors. Western blot analysis showed no significant difference in APOE band intensity among CM collected from the three cultures; in each case a twofold dilution of CM yielded APOE band densities similar to that seen with 2.5 lg/mL recombinant APOE (not shown). Parallel western blots



confirmed the absence of APOAI, consistent with others’ finding that brain parenchymal cells do not express APOAI (Elshourbagy et al. 1985a; Elshourbagy et al. 1985b; Zannis et al. 1985). AT in CM from astrocytes expressing each of the three APOE isoforms was below the limit of detection of our assay (0.15 pmol AT). CM from APOE TR astrocytes added to MC65 cells at the time of Tet removal provided a concentration-dependent cytoprotection that was greatest for APOE2, less for APOE3, and least for APOE4 (Fig. 4). In striking contrast to similar concentrations of purified APOE and recombinant APOE2 (see Fig. 3), astrocyte-CM from APOE2 and APOE3 TR cultures completely protected MC65 cells from loss of viability following Tet removal. While these results do not rule out other factors in CM contributing to the response, they do confirm that the isoform-specific difference seen with recombinant APOE extended to other sources of APOE, either as a direct or indirect effect. While the experiments described above bear on the mechanisms by which HDL and astrocyte-derived nascent APOE-bearing lipoproteins protect Tet– MC65 cells, they do not address the mechanisms of enhanced cytoprotection by AT-enriched HDL. Interpretation of data presented in Fig. 1 from HDL-ATsat is limited by the fact that both HDL and AT were increasing simultaneously at a constant ratio of 0.28 nmoles AT/lg HDL. Therefore, we refined our experiments to use a fixed concentration of HDL particles (0.75 lg/mL, HDLfixed), which by itself offered little or no cytoprotection (see Fig. 1), with increasing amounts of AT enrichment. AT was added to achieve 1/10 saturation (HDLfixed-AT1/10 sat), ½ saturation (HDLfixed-AT1/2 sat) or saturation (HDLfixed-ATsat); when these particles were added to cultures the resulting AT concentrations were 0.02, 0.1, or

Fig. 4 Conditioned serum-free medium (CM) was harvested from targeted replacement transgenic mice expressing ‘humanized’ APOE2, APOE3, or APOE4, and the indicated volume diluted to 200 lL and added to Tet– MC65 cells at the time of Tet removal (t ¼0). Viability was determined 72 h later by MTT assay. Data are expressed as mean percentage viability ± SEM (n ¼ 3) with parallel Tet+ cultures set at 100% viability. Absence of a visible error bar is due to an SEM smaller than the data symbol. Two-way ANOVA had p < 0.0001 for both APOE isoform and concentration.

0.2 lM, respectively, all subcytoprotective concentrations of free AT (see Fig. 1). There was significant AT concentrationdependent protection of Tet– cells by HDLfixed enriched with increasing amounts of AT (Fig. 5a). (p < 0.01 for each HDLfixed-AT preparation compared with untreated Tet– cultures). These data show that AT combined with HDL had a synergistic protective effect in Tet– MC65 cells. This synergistic cytoprotective effect of HDLfixed-AT was accompanied by suppressed accumulation of Ab-immunoreactive aggregates, similar to that seen in Fig. 2. As shown in Fig. 5(b), semiquantitative assessment of accumulated C99-derived Ab-immunoreactive peptides and HMWCs by western blot 24 h after Tet removal showed endogenous APP in all cultures and multiple Ab-immunoreactive HMWCs in Tet– cells at this time, similar to previous reports (Sopher et al. 1994; Sopher et al. 1996; Woltjer et al. 2003).

(a)

* * *

(b)

Fig. 5 (a) AT was incorporated into 0.75 lg/mL HDL (HDLfixed) to yield particles that were 1/10 saturated with AT (HDLfixed-AT1/10 sat), ½ saturated with AT (HDLfixed-AT1/2 sat), or fully saturated with AT (HDLfixed-ATsat). Tet was removed from MC65 cells and HDL alone or the AT-enriched HDL particles added at a constant concentration of 0.75 lg/mL HDL at time 0; the corresponding concentrations of AT added were 0, 0.02, 0.1, and 0.2 lM AT, respectively. Data in graph are expressed as percentage viability ± SEM (n ¼ 3) 72 h after Tet removal. Absence of a visible error bar is due to an SEM smaller than the data symbol. ANOVA had p < 0.0001 with Bonferroni-corrected repeated pair comparisons having *p < 0.01 for each HDLfixed-AT preparation, but not HDL alone, compared with untreated Tet– cultures. (b) Western blot of cells treated with the same exposures (normalized to lg/105 cells) as in (a) showing accumulation of high molecular weight complexes (HMWCs) derived from C99 and its proteolytic products that are immunoreactive with 6E10, a monoclonal antibody to Ab1)17. Lane 1, Tet+ lane 2, no treatment (Tet–); lane 3, HDL alone; lane 4, HDLfixed-AT1/10 sat; lane 5, HDLfixed-AT1/2sat; and lane 6, HDLfixed-ATsat.



Interestingly, HDLfixed alone was associated with partial reduction in HMWCs in the absence of a significant effect on cell viability. This is consistent with our observation that while APOAI also partially reduced HMWCs (55% reduction) in the absence of significant cytoprotection, the larger effect seen with APOE (78% reduction in HMWCs) was associated with modest but significant cytoprotection. As seen in Fig. 5(b), increasing levels of AT-enrichment of HDLfixed resulted in progressive further reduction in the cellular levels of Ab-immunoreactive HMWCs. These data indicate that in parallel with cytoprotection, AT-enriched HDLfixed showed an AT concentration-dependent suppression of Ab-immunoreactive HMWCs in Tet– MC65 cells. One obvious potential explanation for the increased cytoprotective effect of HDL-AT is enhanced delivery of AT to cells. We measured intracellular AT concentration and found that all three HDLfixed-AT particles significantly increased intracellular AT concentration in Tet– cells after 1 h in a concentration-dependent manner (slope ¼ 6.6 ± 1.2 pmol AT increase/nmol AT added, R2 ¼ 0.94), while over the same concentration range free AT had a smaller effect (slope ¼ 2.4 ± 0.1 pmol AT increase/nmol AT added, R2 ¼ 0.99); these data show that enrichment of AT in HDL particles almost tripled the efficiency of delivery to MC65 cells over free AT. In order to determine whether this increased AT delivery could completely accounted for the increased viability seen with HDLfixed-AT over HDL alone, intracellular AT concentration after one hour of treatment with HDLfixed-AT was plotted against viability (as shown in Fig. 5) for each of the three particles and compared with the same plot for free AT (with viability values from Fig. 1). As seen in Fig. 6, the slope of the best fit line for free AT was 4.3 ± 0.2% viability per pmol AT/105 cells (R2 ¼ 0.99, p < 0.0001), while the slope of the best fit line for AT-enriched HDLfixed particles was 34.5 ± 6.4% viability per pmol AT/105 cells (R2 ¼ 0.94, p < 0.0001), demonstrating that even after adjusting for increased AT uptake, AT delivered by HDL was, mole for mole, approximately sevenfold more effective than free AT as a cytoprotectant. Note that the concentration of HDL used in these experiments was minimally cytoprotective on its own, while the AT in HDLfixed-AT was at subcytoprotective concentrations (see Fig. 1). Thus the synergy between HDL and AT in protecting Tet– MC65 cells cannot be explained by increased AT uptake alone. As the protective effect of HDL partially resided in apolipoproteins, we determined whether the synergism between HDL and AT might also be apolipoprotein-dependent, thus reflecting a more physiological setting than apolipoproteins in isolation as in Fig. 3. In these experiments, purified APOE or APOAI (0–5 lg/mL) was mixed with AT (0.1 or 0.5 lM) for 30 min at 37oC, and then applied to cells upon Tet removal, just as with HDL. As seen in Fig. 7, purified APOE had a small, but significant, concen-

Fig. 6 Cells were incubated with AT or HDLfixed-AT for 1 h at identical exposures as for viability experiments (normalized to lg/105 cells), rinsed and harvested. Intracellular AT concentration was determined by HPLC (expressed as pmoles/105 cells) and plotted against viability at the same exposure (from Fig. 1 for free AT and Fig. 5 A for HDLfixedAT). Absence of a visible error bar is due to an SEM smaller than the data symbol. Linear regression analysis for free AT gave a slope ¼ 4.39% viability per picamole intracellular AT/105 cells and intercept ¼ 10.56% viability (r2 ¼ 0.99) and for HDLfixed-AT gave a slope ¼ 34.51% viability per picamole intracellular AT/105 cells and intercept ¼ 11.44% viability (r2 ¼ 0.94).

Fig. 7 Cells were prepared as described for Fig. 1. At the time of Tet removal, apolipoprotein (E or AI), AT, both or neither was added to cells and viability determined at 72 h. Apolipoproteins were added at the concentrations shown on the x-axis, while AT was added at a fixed concentration of either 0.1 lM or 0.5 lM. Data are expressed as mean percentage viability ± SEM (n ‡ 3) with parallel Tet+ cultures set at 100% viability. Absence of a visible error bar is due to an SEM smaller than the data symbol. For reference, the dashed line shows the cytoprotective effect of 0.5 lM AT alone. Two-way ANOVA for experiments using APOAI had p < 0.0001 for both APOAI and AT concentrations. Two-way ANOVA for experiments using APOE had p < 0.0001 for both APOE and AT concentrations.

tration-dependent cytoprotective effect on Tet– MC65 cells, while APOAI showed a trend toward protection at the highest concentration that was not statistically significant. The combination of APOAI plus AT had a significantly greater cytoprotective effect than either alone and was dependent on AT concentration (p < 0.0001 for APOAI concentration and AT concentration). Indeed, the combination of APOAI with 0.5 lM AT completely protected Tet– MC65 cells, largely reproducing the effect with HDL-ATsat



observed in Fig. 1. Interestingly, although APOE alone had a greater cytoprotective effect than APOAI, combination of purified APOE with AT was significantly less protective than the combination of APOAI with AT. Recombinant APOE isoforms (2.5 lg/mL) were also investigated using 0.5 lM AT. Similar to the relative isoform effects in Fig. 3, recombinant APOE4 plus AT was 39 ± 2% less cytoprotective than purified APOE plus AT (p £ 0.001). Moreover, neither recombinant APOE2 nor APOE3 plus AT was significantly different from purified APOE plus AT while both were significantly more cytoprotective than APOE4 plus AT (p £ 0.01 for both). These data suggest that the synergistic cytoprotective effect of HDL-AT in Tet– MC65 cells derives primarily from an interaction between APOAI and AT and that the effect can be completely replicated by APOAI and AT in the absence of lipoprotein particles. Discussion

Genetic linkage studies have consistently associated an increased risk of AD with inheritance of the e4 APOE allele and a decreased risk of AD with the e2 APOE allele (Corder et al. 1993; Corder et al. 1994). Moreover, epidemiologic studies have repeatedly associated AT intake with reduced risk of AD, and one clinical trial has shown that large doses of AT may slow the progression of dementia in patients with AD (Sano 2003; Zandi et al. 2004). As APOE is one of the major CNS apolipoproteins and as AT in vivo is transported in lipoproteins, we tested the hypothesis that CNS lipoproteins, as modeled by relevant concentrations of HDL, and AT would interact to suppress the neurotoxicity associated with C99-derived peptides, a key component in the pathogenesis of AD. Our results provided evidence for two mechanisms of protection from C99-associated toxicity; the first was APOE isoform-dependent, greatest for APOE2, likely indirect, and independent of AT; the second was synergistic between apolipoproteins and AT, greater for APOAI than APOE, and not explained by enhanced delivery of AT to cells. HDL alone protected Tet– MC65 cells, a novel observation. The basis by which HDL could suppress C99-associated cytotoxicity was shown to be dependent on HDL protein and partially replicated by purified APOE or recombinant APOE2 alone, but not DMPC or APOAI. As mentioned, there are several potential limitations to the biological activity of purified or recombinant APOE so we pursued this mechanism further by harvesting conditioned medium from cultures of TR mouse primary astrocytes; others have shown that murine astrocyte-conditioned medium contains APOE-bearing nascent lipoprotein particles (LaDu et al. 1998). Here we were able to completely replicate the protective effect of HDL using these CNS cell-derived particles with APOE2-derived CM being the most potent protectant, followed by APOE3, and finally APOE4. While this could be interpreted as evidence that the packaging into

lipid of APOE vastly improves its cytoprotective effect, it seems much more likely that other products in the conditioned medium are contributing to this effect; however, important for interpretation of our data, we did rule out both APOAI and AT as factors. In fact, secretion of several molecules has been demonstrated to be APOE isoformdependent (Vitek et al. 1994) Thus, the partial cytoprotection seen with recombinant APOE2 and the complete cytoprotection seen with APOE2 CM, both of which were isoformspecific, were consistent with both a direct and indirect facet of APOE2 cytoprotection. As far as we are aware, this APOE isoform-dependent difference in protection from C99-associated toxicity represents a new pathway by which APOE isoforms may influence the pathogenesis of AD and may help explain the reduced risk of AD reported in individuals who inherit an e2 APOE allele. AT alone also protected MC65 cells, as has been reported previously (Sopher et al. 1996). Free AT-mediated protection of Tet– MC65 cells was linearly related to the amount of AT added and had an EC50 of approximately 1.9 lM, while AT-enriched HDL protected at much lower concentrations of AT (EC50 ¼ 0.2 lM). One criticism of AT cytoprotection in cell culture models is that its effective concentrations in vitro are high compared with levels attained in vivo, calling into question the relative importance of AT as an antioxidant for AD pathogenesis in vivo. Our results showing synergism between AT and HDL, APOAI, or APOE address this criticism and suggest that AT activity in models of AD pathogenesis can be more accurately assessed using AT enriched in lipoproteins or combined with apolipoproteins. Synergistic protection from C99-associated cytotoxicity was observed for HDL enriched with AT. Indeed, enrichment of minimally (< 5%) protective HDL with subprotective concentrations of AT led to markedly increased cytoprotection of up to approximately 60%; attempts at further protection by HDL-AT were precluded by the saturation of this small amount of HDL with AT. Although AT enriched in HDL was delivered to cells with greater efficiency that free AT, this still did not account for the synergistic protection observed. Indeed, Fig. 6 shows that even when adjusted for increased delivery by HDL, HDL-delivered AT was approximately sevenfold more protective against C99-associated toxicity than free AT. These data demonstrated that it was not simply the amount of AT delivered, but also the context in which it was delivered that appeared to underlie the synergistic protective effect of AT-enriched HDL. We tested this hypothesis by combining the components present in HDL-AT in an attempt to replicate its synergistic protection. Even the simplest model of combining purified APOAI directly with AT was able to almost exactly replicate the protective action of HDL-AT; purified APOE combined with AT also was synergistically protective but less potent than APOAI and AT. The precise mechanism by which these apolipoproteins act synergistically with AT to protect from



C99-associated cytotoxicity is not clear, but may include altered trafficking of AT to different compartments in cells or activation of cell-signaling pathways that substantially augment the efficacy of AT. In summary, we observed that CNS lipoproteins, as modeled by HDL, could completely protect MC65 cells from C99-associated toxicity. We provided evidence for two mechanisms by which CNS lipoproteins may protect neurons from C99 production. The first was AT-independent, APOE isoform-dependent, and most potent for APOE2 versus APOE3 or APOE4. The second was a synergistic protection afforded by APOAI and AT, and less so by APOE and AT. Interestingly, this supra-additive effect was not explained by enhanced delivery of AT to cells by AT-enriched HDL. These data provide evidence that the two major CSF apolipoproteins at physiologic concentrations can protect neurons from C99-associated cytotoxicity through at least two different mechanisms. They also provide a novel explanation for the apparent AD-protective effect of inheriting an e2 APOE allele and suggest that therapeutic interventions in AD that include AT may be enhanced by optimizing AT enrichment of CNS lipoproteins or by devising APOAI mimetics to augment AT efficacy. Acknowledgements The authors wish to thank Steve Bogh for assistance with manuscript preparation. This work was supported by the Nancy and Ellsworth Alvord Endowment, grants from the NIH (AG16835, AG05144, AG022040, AG019033), and a grant from the Alzheimer Disease Public Awareness Fund from the University of Washington.

References Bales K. R., Dodart J. C., DeMattos R. B., Holtzman D. M. and Paul S. M. (2002) Apolipoprotein E, amyloid, and Alzheimer disease. Mol. Interv. 2, 363–375. Bieri J. G., Tolliver T. J. and Catignani G. L. (1979) Simultaneous determination of a-tocopherol and retinol in plasma or red cells by high pressure liquid chromatography. Am. J. Clin. Nutr. 32, 2143– 2149. Borghini I., Barja F., Pometta D. and James R. W. (1995) Characterization of subpopulations of lipoprotein particles isolated from human cerebrospinal fluid. Biochim. Biophys. Acta 1225, 192–200. Butterfield D. A., Koppal T., Subramaniam R. and Yatin S. (1999) Vitamin E as an antioxidant/free radical scavenger against amyloid b-peptide-induced oxidative stress in neocortical synaptosomal membranes and hippocampal neurons in culture: insights into Alzheimer’s disease. Rev. Neurosci. 10, 141–149. Corder E. H., Saunders A. M., Strittmatter W. J., Schmechel D. E., Gaskell P. C., Small G. W., Roses A. D., Haines J. L. and PericakVance M. A. (1993) Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families. Science 261, 921–923. Corder E. H., Saunders A. M., Risch N. J. et al. (1994) Protective effect of apolipoprotein E type 2 allele for late onset alzheimer disease. Nat. Genet. 7, 180–184.

Elshourbagy N. A., Liao W. S. L., Mahley R. W. and Taylor J. M. (1985a) Apolipoprotein E mRNA is abundant in the brain and adrenals, as well as in the liver, and is present in other peripheral tissues of rats and marmosets. Proc. Natl Acad. Sci. USA 82, 203– 207. Elshourbagy N. A., Boguski M. S., Liao W. S. L., Jefferson L. S., Gordon J. I. and Taylor J. M. (1985b) Expression of rat apolipoprotein A-IV and A-I genes: mRNA induction during development and in response to glucocorticoids and insulin. Proc. Natl Acad. Sci. USA 82, 8242–8246. Fraser S. P., Suh Y. H. and Djamgoz M. B. (1997) Ionic effects of the Alzheimer’s disease b-amyloid precursor protein and its metabolic fragments. Trends Neurosci. 20, 67–72. Fukuchi K., Kamino K., Deeb S. S., Smith A. C., Dang T. and Martin G. M. (1992a) Overexpression of amyloid precursor protein alters its normal processing and is associated with neurotoxicity. Biochem. Biophys. Res. Commun. 182, 165–173. Fukuchi K., Kamino K., Deeb S. S., Furlong C. E., Sundstrom J. A., Smith A. C. and Martin G. M. (1992b) Expression of a carboxy-terminal region of the b-amyloid precursor protein in a heterogeneous culture of neuroblastoma cells: evidence for altered processing and selective neurotoxicity. Brain Res. Mol. Brain Res. 16, 37–46. Fukuchi K., Sopher B., Furlong C. E., Smith A. C., Dang N. and Martin G. M. (1993) Selective neurotoxicity of COOH-terminal fragments of the b-amyloid precursor protein. Neurosci. Lett. 154, 145–148. Furukawa S., Furukawa Y., Satoyoshi E. and Hayashi K. (1986) Synthesis and secretion of nerve growth factor by mouse astroglial cells in culture. Biochem. Biophys. Res. Commun. 136, 57–63. Jin L. W., Hua D. H., Shie F. S., Maezawa I., Sopher B. and Martin G. M. (2002) Novel tricyclic pyrone compounds prevent intracellular APP C99-induced cell death. J. Mol. Neurosci. 19, 57–61. LaDu M. J., Gilligan S. M., Lukens J. R., Cabana V. G., Reardon C. A., VanEldik L. J. and Holtzman D. M. (1998) Nascent astrocyte particle differ from lipoproteins in CSF. J. Neuorchem. 70, 2070– 2081. Laureaux C., Therond P., Bonnefont-Rousselot D., Troupel S. E., Legrand A. and Delattre J. (1997) a-Tocopherol enrichment of high-density lipoproteins: stabilization of hydroperoxides produced during copper oxidation. Free Rad. Biol. Med. 22, 185–194. Linton M. F., Gish R., Hubl S. T., Butler E., Esquivel C., Bry W. I., Boyles J. K., Wardell M. R. and Young S. G. (1991) Phenotypes of apolipoprotein B an apolipoprotein E after liver transplantation. J. Clin. Invest. 88, 270–281. McPhie D. L., Lee R. K., Eckman C. B., Olstein D. H., Durham S. P., Yager D., Younkin S. G., Wurtman R. J. and Neve R. L. (1997) Neuronal expression of b-amyloid precursor protein Alzheimer mutations causes intracellular accumulation of a C-terminal fragment containing both the amyloid b and cytoplasmic domains. J. Biol. Chem. 272, 24 743–24 746. McPhie D. L., Golde T., Eckman C. B., Yager D., Brant J. B. and Neve R. L. (2001) b-Secretase cleavage of the amyloid precursor protein mediates neuronal apoptosis caused by familial Alzheimer’s disease mutations. Brain Res. Mol. Brain Res. 97, 103–113. Mendez A. J. and Oram J. F. (1997) Limited proteolysis of high density lipoprotein abolishes its interaction with cell-surface binding sites that promote cholesterol efflux. Biochim. Biophys. Acta 1346, 285– 299. Montine K. S., Bassett C. N., Ou J. J., Markesbery W. R., Swift L. L. and Montine T. J. (1998) Apolipoprotein E allelic influence on human ventricular fluid apolipoproteins. J. Lipid Res. 38, 2443–2451. Montine T. J., Montine K. S., Olson S. J., Graham D. G., Roberts L. J., Morrow J. D., Linton M., Fazio S. and Swift L. L. (1999) Increased cerebral cortical lipid peroxidation and abnormal phospholipids in



aged homozygous apoE-deficient C57BL/6J mice. Exp. Neurol. 158, 234–241. Montine T. J., Neely M. D., Quinn J. F., Beal M. F., Markesbery W. R., Roberts L. J. and Morrow J. D. (2002) Lipid peroxidation in aging brain and Alzheimer’s disease. Free Rad. Biol. Med. 33, 620. Oster-Granite M. L., McPhie D. L., Greenan J. and Neve R. L. (1996) Age-dependent neuronal and synaptic degeneration in mice transgenic for the C terminus of the amyloid precursor protein. J. Neurosci. 16, 6732–6741. Pitas R. E., Boyles J. K., Lee S. H., Hui D. and Weisgraber K. H. (1987) Lipoproteins and their receptors in the central nervous system. J. Biol. Chem. 262, 14 352–14 360. Rall S. C. Jr, Weisgraber K. H. and Mahley R. W. (1982) Human apolipoprotein E: the complete amino acid sequence. J. Biol. Chem. 257, 4171–4178. Roheim P. S., Carey M., Forte T. and Vega G. L. (1979) Apolipoproteins in human cerebrospinal fluid. Proc. Natl Acad. Sci. USA 76, 4646– 4649. Sano M. (2003) Noncholinergic treatment options for Alzheimer’s disease. J. Clin. Psychiatry 64, 23–28. Sano M., Ernesto C., Thomas R. G. et al. (1997) A controlled trial of selegiline, a-tocopherol, or both as a treatment for Alzheimer’s disease. N. Engl. J. Med. 336, 1216–1222. Saunders A. M., Strittmatter W. J., Schmechel D. et al. (1993) Association of apolipoprotein E allele e4 with late-onset familial and sporadic Alzheimer’s disease. Neurology 43, 1467–1472. Selkoe D. J. (2002) Alzheimer’s disease is a synaptic failure. Science 298, 789–791. Sopher B. L., Fukuchi K., Smith A. C., Leppig K. A., Furlong C. E. and Martin G. M. (1994) Cytotoxicity mediated by conditional expression of a carboxyl-terminal derivative of the b-amyloid precursor protein. Brain Res. Mol. Brain Res. 26, 207–217. Sopher B. L., Fukuchi K., Kavanagh T. J., Furlong C. E. and Martin G. M. (1996) Neurodegenerative mechanisms in Alzheimer disease: a role for oxidative damage in amyloid b protein precursor-mediated cell death. Mol. Chem. Neuropathol. 29, 153–168.

Sullivan P. M., Mezdour H., Aratani Y., Knouff C., Najib J., Reddick R. L., Quarfordt S. H. and Maeda N. (1997) Targeted replacement of the mouse apolipoprotein E gene with the common human APOE3 allele enhances diet-induced hypercholesterolemia and atherosclerosis. J. Biol. Chem. 272, 17 972–17 980. Sullivan P. M., Mezdour H., Quarfordt S. H. and Maeda N. (1998) Type III hyperlipoproteinemia and spontaneous atherosclerosis in mice resulting from gene replacement of mouse Apoe with human APOE*2. J. Clin. Invest. 102, 130–135. Vitek M. P., Bhattacharya K., Glendening J. M., Stopa E., Vlassara H., Bucala R., Manogue K. and Cerami A. (1994) Advanced glycation end products contribute to amyloidosis in Alzheimer disease. Proc. Natl Acad. Sci. USA 91, 4766–4770. Vuletic S., Jin L. W., Marcovina S. M., Peskind E. R., Moller T. and Albers J. J. (2003) Widespread distribution of PLTP in human CNS: evidence for PLTP synthesis by glia and neurons, and increased levels in Alzheimer’s disease. J. Lipid Res. 44, 1113–1123. Walsh D. M., Klyubin I., Fadeeva J. V., Rowan M. J. and Selkoe D. J. (2002) Amyloid-b oligomers: their production, toxicity and therapeutic inhibition. Biochem. Soc. Trans. 30, 552–557. Woltjer R. L., Maezawa I., Ou J. J., Montine K. M. and Montine T. J. (2003) Advanced glycation endproduct precursor alters intracellular amyloid-b/AbPP carboxy-terminal fragment aggregation and cytotoxicity. J. Alzheimer’s Dis. 5, 467–476. Yatin S. M., Varadarajan S. and Butterfield D. A. (2000) Vitamin E prevents Alzheimer’s amyloid b-peptide (1–42)-induced neuronal protein oxidation and reactive oxygen species production. J. Alzheimer’s Dis. 2, 123–131. Zandi P. C., Anthony J. C., Khachaturian A. S., Stone S. V., Gustafson D., Tschanz J. T., Norton M. C., Welsh-Bohmer K. A. and Breitner J. C. (2004) Reduced risk of Alzheimer disease in users of antioxidant vitamin supplements: the Cache County Study. Arch. Neurol. 61, 82–88. Zannis V. I., Cole F. S., Jackson C. L., Kurnit D. M. and Karathanasis S. K. (1985) Distribution of apolipoprotein A-I, C-II, C-III and E mRNA in fetal human tissues. Biochemistry 24, 4450–4455.
