β-Amyloid promotes accumulation of lipid peroxides by inhibiting CD36-mediated clearance of oxidized lipoproteins

Aβ_1-42 and reverse Aβ_42-1 (rev Aβ) peptides were obtained from Biosource International (Camarillo, California). To induce fibril formation, Aβ_1-42 was resuspended in H₂O at 1 mg/ml and incubated for 1 week (37°C) and fibril formation was confirmed by thioflavine S (Sigma-Aldrich Co., St. Louis, Missouri) fluorescent staining as we previously described [17, 18].

The Cd36^-/- mice were generated in our laboratory as previously described [17] and SraI/II null (Sra^-/-) mice were generously provided from Dr. T. Kodama (University of Tokyo, Japan) [24]. Both mouse lines were backcrossed to C57BL/6 mice for 7 generations (98.6% C57BL/6) prior to intercrossing to generate mice lacking both Sra and Cd36. Double knockout mice (Sra^-/-/Cd36^-/-) were generated from heterozygote intercrosses at the expected ration of 1:16. Wild type age-matched C57BL/6 mice (The Jackson Laboratory, Bar Harbor, Maine) were used as controls for these three lines. Lyn^-/- and Fyn^-/- mice were obtained from The Jackson Laboratory and Lyn^-/-, Fyn^-/- and wild type littermate control mice were generated from heterozygote intercrosses. All mice were maintained in a pathogen-free facility with free access to rodent chow and water. All experimental procedures were carried out in accordance with Massachusetts General Hospital's institutional guidelines for use of laboratory animals.

Macrophages were collected from 6–8 week old mice by peritoneal lavage 4 days after i.p. injection with 3% thioglycollate as we previously described [17, 25]. Cells were washed in PBS, cultured for 2 h in DMEM with 5% FCS, and washed again to remove non-adherent cells. Adherent cells were incubated in DMEM with 1% FCS overnight prior to use and were routinely >95% CD11b⁺ and F4/80⁺ as determined by flow cytometric analysis. Primary microglia were prepared from mixed brain cultures of neonatal mice as we previously described [17]. Briefly, whole brains were incubated in 0.25% trypsin and 1 mM EDTA (10 min, 25°C) and dissociated to obtain a single cell-suspension. Cells were washed in HBSS (4x, 10 min) and cultured in DMEM containing 10% FCS, 1% Fungizone for 10–12 days. Microglia accumulating above astrocyte monolayers were collected after gentle agitation, washed and incubated in DMEM with 1% FCS overnight prior to use. Microglia prepared in this manner were routinely >95% CR3⁺ and express SR-A and CD36 [14, 17, 18].

Human ¹²⁵I-LDL and LDL (d = 1.019 - 1.063) were purchased from Biomedical Technologies (Stoughton, Massachusetts) and oxidized as we previously described [12, 26]. LDL was diluted to 250 μg/ml, dialyzed against PBS at 4°C to remove EDTA, and then dialyzed against 5 μM CuSO₄ in PBS at 37°C for 6 or 10 h. Oxidation was terminated by the addition of 50 μM butylated hydroxytoluene and 200 μM EDTA and oxLDL was used within 2 days of preparation. Moderately oxidized LDL (6 h oxidation) had a relative electrophoretic mobility of approximately 2.5–3 times that of native, unmodified LDL, whereas extensively oxidized LDL (10 h oxidation) had a relative mobility four times that of native LDL.

Measurement of ¹²⁵I-oxLDL binding, degradation and uptake was performed on confluent monolayers of peritoneal macrophages (7 × 10⁵) and microglia (5 × 10⁵) in 24 well plates as we previously described [12, 26]. Briefly, 10 μg/ml of ¹²⁵I-oxLDL was added to cells in the presence or absence of 30-fold excess unlabeled oxLDL, native LDL, Aβ_1-42, or rev Aβ peptide for 5 h at 37°C. To measure ¹²⁵I-oxLDL degradation, media were removed and assayed for TCA-soluble non-iodide degradation products. To measure ¹²⁵I-oxLDL binding in the presence Aβ_1-42 or rev Aβ, cells were washed 3x with 50 mM Tris pH 7.4, 0.15 N NaCl and 2 mg/ml BSA, 1x with 50 mM Tris pH 7.4 and 0.15 N NaCl and treated with 0.4% dextran sulfate to release surface bound ¹²⁵I-oxLDL [27]. To measure ¹²⁵I-oxLDL uptake, cells were washed 3x in 50 mM Tris pH 7.4 and 0.15 N NaCl, lysed in 0.1 N NaOH and assayed for ¹²⁵I and cellular protein content. In some experiments, cell-association of oxLDL (cell-surface bound and endocytosed oxLDL) was measured by omitting the dextran sulfate treatment. Cellular protein content was measured by BCA assay (Pierce, Rockford, IL) and degradation, binding and uptake activity are expressed as ng ¹²⁵I-oxLDL/mg protein. Specific degradation was calculated as the difference of total cellular degradation of ¹²⁵I-oxLDL in the presence and absence of 30-fold excess unlabelled oxLDL competitor. All measurements were performed in triplicate and are representative of at least 3 experiments.

Macrophages and microglia were cultured with 40 μg/ml of oxLDL for 48 h in the presence or absence of Aβ_1-42 or revAβ. Cholesterol ester accumulation was assessed by gas chromatography-mass spectrometry (GC-MS) and oil red O staining as we previously described [12, 26]. For GC-MS analysis, lipids were extracted with hexane:isopropanol (3:2) and stigmasterol (Sigma, St. Louis, Missouri) was added as an internal standard. Lipid extracts were washed once with water and divided equally. One lipid aliquot was saponified for determination of total cholesterol and the second aliquot analyzed for free cholesterol using gas chromatography-mass spectrometry. The samples were injected (splitless) into an Agilent 6890 GC-MS-(G2613A system, Agilent Technologies, Palo Alto, CA) equipped with a J&W DB17 fused silica capillary column (15 m × 0.25 mm inner diameter × 0.5 μm; J&W Scientific, Folsom, CA). The GC temperature program was as follows: the initial temperature was 260°C for 5 min, then increased to 280°C (5°C/min) and held 280°C for 11 min. A model 5973N mass-selective detector (Agilent Technologies) was used in scan modes to identify the samples. Cholesterol measurements were made in triplicate and normalized to cellular protein content. Cholesterol ester content was calculated by subtracting free cholesterol from total cholesterol measured after saponification. To assess neutral lipid accumulation, cells were fixed in 4% paraformaldehyde and stained with oil red O for 30 min. Staining was recorded on an Olympus X10 microscope equipped with a digital camera.

Total RNA was extracted using Trizol B reagent and real-time quantitative RT-PCR (QRT-PCR) was performed using the QuantiTect SYBR Green PCR kit (Qiagen Inc, Valencia, CA) as we previously described [17, 18]. Each reaction contained 0.3 μM of CD36, SRA or GAPDH primers, 3 μl of cDNA, SYBR Green, and HotStarTaq polymerase. PCR was performed using a BioRad i Cycler under the following conditions: 15 min at 95°C, followed by 30 cycles of 30 sec at 95°C, 30 sec at 55°C and 30 sec at 72°C. Each sample was analyzed in triplicate and the amount of CD36, SRA and GAPDH mRNA in each sample was calculated from a standard curve of known template. Data are expressed as the mean number of CD36 and SRA molecules normalized to GAPDH.

Cells were washed in ice-cold PBS and lysed in radioimmune precipitation buffer containing protease and phosphatase inhibitors. For detection of CD36, 30 μg of protein was run on an 8% denaturing SDS-polyacrylamide gel, transferred to nitrocellulose and blocked overnight in 5% nonfat dry milk and 3% BSA in Tris-buffered saline containing 0.1% Tween 20 (TBS-T) as we previously described [17, 26]. Membranes were incubated with a rabbit anti-CD36 antiserum (1:500 dilution) generated in our laboratory [17] for 2 hours, washed three times in TBS-T, and incubated with horseradish peroxidase-conjugated anti-rabbit IgG (1:10,000 dilution) for 1 hour. Blots were washed 3x in TBS-T, exposed to ECL reagent (Amersham Biosciences, Piscataway, NJ), and signal was recorded and quantified using an Alpha Innotech Fluorchem 8800 image analysis system. Blots were stripped and probed with an anti-actin rabbit polyclonal antibody (Santa Cruz Biotechnology) as described above as an internal standard for equivalent loading.

Treatment of peritoneal macrophages with Aβ_1-42, but not rev Aβ, dose-dependently inhibited lysosomal degradation of ¹²⁵I-oxLDL (Fig. 1a). Half-maximal inhibition of macrophage ¹²⁵I-oxLDL degradation was achieved with 10 μM Aβ_1-42. This was equivalent to the inhibitory effect of 15-fold excess of unlabelled oxLDL competitor (Fig. 1b). At 20 μM, Aβ_1-42 reduced macrophage degradation of ¹²⁵I-oxLDL by up to 90%, while treatment with the same concentration of non-fibrillar rev Aβ peptide reduced degradation by only 10%, and this concentration was selected for all further experiments. Because engulfment of Aβ_1-42 has previously been reported to disrupt endosomal/lysosomal integrity in a neuronal cell line [28], we investigated whether the observed reduction in oxLDL degradation could be attributed to lysosomal accumulation of Aβ_1-42 which occurs within 1 h of treatment. After exposure to Aβ_1-42 for 3 hours, macrophages were washed extensively to remove extracellular Aβ_1-42 and exposed to ¹²⁵I-oxLDL or ¹²⁵I-oxLDL + Aβ_1-42 for 5 h. While cells continuously exposed to Aβ_1-42 showed a profound impairment of oxLDL degradation, cells pre-treated with Aβ_1-42 were similar to untreated and revAβ-treated cells, indicating that intracellular accumulation of Aβ_1-42 does not block subsequent lysosomal degradation of oxLDL (Fig. 1c).

The inhibition of ¹²⁵I-oxLDL degradation by Aβ_1-42 would be predicted to reduce cellular cholesterol ester accumulation. Excess unesterified "free" cholesterol is cytotoxic and is thus rapidly converted by the microsomal enzyme acyl-coenzyme A:cholesterol acyltransferase (ACAT) to cholesterol ester for storage. This neutral lipid is retained in cytoplasmic lipid droplets for storage and/or efflux from the cell. Using gas chromatograpy-mass spectrometry, we quantified the cholesterol ester content of macrophages treated with oxLDL in the presence and absence of Aβ_1-42. As expected, untreated cells did not contain measurable cholesterol ester, while macrophages treated with 40 μg/ml oxLDL for 48 h accumulated approximately 80 μg cholesterol ester/mg cellular protein (Fig. 1d). By contrast, macrophages treated with both oxLDL and Aβ_1-42 showed no measurable cholesterol ester accumulation after 48 h, similar to untreated cells.

As seen in peripheral macrophages, Aβ_1-42 substantially inhibited ¹²⁵I-oxLDL binding, uptake, and degradation by primary microglia indicating that it has a similar effect on lipoprotein metabolism in these two myeloid cell types (Fig. 2a,2b,2c). In the presence of 20 μM Aβ_1-42, microglia demonstrated a 55% reduction in ¹²⁵I-oxLDL binding, an 80% reduction in ¹²⁵I-oxLDL uptake and a 95% reduction of ¹²⁵I-oxLDL degradation. The absence of cholesterol ester in oxLDL treated microglia exposed to Aβ_1-42 was confirmed by staining cells with the neutral lipid stain oil red O. Microglia treated with oxLDL alone demonstrate oil red O positive lipid droplets in their cytoplasm characteristic of cholesterol ester storage (Fig. 2d). However, in the presence of Aβ_1-42, oxLDL treated microglia show a dramatic reduction in lipid droplets that is not seen with treatment with the same concentration of rev Aβ. As expected, cells treated with Aβ_1-42 or rev Aβ alone do not accumulate cholesterol ester in the absence of exogenously added oxLDL (Fig. 2d). Similar results were observed in macrophages (data not shown). Together, these data demonstrate that Aβ blocks cholesterol ester accumulation in macrophages and microglia by inhibiting oxLDL clearance.

To address the mechanism by which Aβ_1-42 inhibits oxLDL metabolism, we first evaluated cellular expression of the scavenger receptors SRA and CD36. Fibrillar Aβ_1-42 reduced expression of CD36 mRNA by 40 and 60% after 6 and 24 h, respectively (Fig. 3a), but showed no effect on macrophage expression of SRA. Western blotting confirmed a 40% decrease in CD36 protein in Aβ_1-42 treated macrophages (Figure 3b), which would be expected to reduce the ability of these cells to bind oxLDL.

β-Amyloid has previously been reported to bind to the class A scavenger receptors SRA I & II and to block uptake of LDL modified by acetylation [14, 21]. We employed Sra and Cd36 single null mice to investigate the role of these receptors in the inhibition of oxLDL clearance by Aβ_1-42. In addition, we used Sra/Cd36 double null mice to evaluate the role of SRA/CD36-independent mechanisms, including those of additional scavenger receptor family members. Because of the difficulty of culturing sufficient numbers of primary microglia for binding and degradation experiments, studies involving knock-out mice were performed with peritoneal macrophages. In Sra^-/- and wild type macrophages Aβ_1-42 blocked cell association (binding and uptake) of ¹²⁵I-oxLDL by greater than 50%, indicating that this scavenger receptor is not essential for the inhibitory action of Aβ (Fig. 4a). By contrast, in the absence of Cd36, impairment of ¹²⁵I-oxLDL cell association by Aβ_1-42 was reduced to 8%, indicating that this receptor was the primary target of Aβ_1-42 inhibition (Fig. 4a).

The finding that CD36 is required for Aβ_1-42 inhibition of oxLDL suggests two possible mechanisms of action: (1) direct competition for CD36 binding, or (2) inhibition of oxLDL metabolism as a result of Aβ/CD36 signal transduction. To address whether CD36 signaling inhibits cellular oxLDL degradation, we used macrophages with targeted deletions in two kinases in this pathway, Lyn and Fyn, which have previously been shown to be required for CD36-mediated p44/42 activation, MCP-1 secretion and ROS production [17]. However, as in wild type macrophages, Aβ_1-42 effectively inhibited ¹²⁵I-oxLDL degradation in Lyn^-/- and Fyn^-/- macrophages, suggesting that this signaling pathway does not inhibit oxLDL metabolism (Fig. 4b). Furthermore, treatment of macrophages with the general phosphotyrosine kinase inhibitor genistein did not reverse Aβ_1-42 inhibition of ¹²⁵I-oxLDL degradation, confirming that phosphotyrosine kinase signaling does not mediate this effect of Aβ_1-42 (data not shown). Interestingly, in untreated Fyn^-/- macrophages ¹²⁵I-oxLDL degradation was increased 2-fold (Fig. 4b) indicating that this kinase may play a role in regulating oxLDL uptake. However, despite a doubling of oxLDL degradation in Fyn^-/- macrophages, this process was still inhibited by Aβ_1-42 by up to 90%. Together, these experiments suggest that Aβ_1-42 inhibition of oxLDL metabolism is not the result of CD36-Lyn/Fyn signal transduction and support the hypothesis that Aβ_1-42 competes for oxLDL binding to CD36. Analysis of ¹²⁵I-oxLDL cell-surface binding showed that Aβ inhibited ¹²⁵I-oxLDL binding by approximately 60% in wild type macrophages (Fig. 5). This inhibitory effect was lost in Cd36^-/- macrophages, confirming that Aβ inhibited oxLDL binding to this receptor. Of note, wild type macrophages bound 60% more oxLDL than macrophages lacking Cd36 as has previously been reported, and this correlated with the percentage reduction of oxLDL binding by Aβ in wild type macrophages (57%), suggesting that the CD36-dependent contribution to oxLDL binding was totally inhibited. To confirm that other myeloid scavenger receptors were not inhibited by Aβ, assesed ¹²⁵I-oxLDL binding in Sra^-/-Cd36^-/- macrophages. No effect of Aβ was observed in these cells, demonstrating the specificity of Aβ inhibition of oxLDL binding to CD36.