Differential Regulation of Amyloid-β Endocytic Trafficking and Lysosomal Degradation by Apolipoprotein E Isoforms^,

Background: Apolipoprotein E (apoE) regulates amyloid-β (Aβ) clearance in an isoform-dependent manner.

Results: Internalized Aβ traffics to lysosomal and recycling pathways. ApoE3 more efficiently promotes Aβ lysosomal trafficking and degradation than apoE4.

Conclusion: ApoE isoforms differentially affect Aβ lysosomal trafficking and degradation.

Significance: Differential effects of apoE isoforms on Aβ cellular degradation may explain why apoE4 is a risk factor for Alzheimer disease.

Abstract

Aggregation of amyloid-β (Aβ) peptides leads to synaptic disruption and neurodegeneration in Alzheimer disease (AD). A major Aβ clearance pathway in the brain is cellular uptake and degradation. However, how Aβ traffics through the endocytic pathway and how AD risk factors regulate this event is unclear. Here we show that the majority of endocytosed Aβ in neurons traffics through early and late endosomes to the lysosomes for degradation. Overexpression of Rab5 or Rab7, small GTPases that function in vesicle fusion for early and late endosomes, respectively, significantly accelerates Aβ endocytic trafficking to the lysosomes. We also found that a portion of endocytosed Aβ traffics through Rab11-positive recycling vesicles. A blockage of this Aβ recycling pathway with a constitutively active Rab11 mutant significantly accelerates cellular Aβ accumulation. Inhibition of lysosomal enzymes results in Aβ accumulation and aggregation. Importantly, apolipoprotein E (apoE) accelerates neuronal Aβ uptake, lysosomal trafficking, and degradation in an isoform-dependent manner with apoE3 more efficiently facilitating Aβ trafficking and degradation than apoE4, a risk factor for AD. Taken together, our results demonstrate that Aβ endocytic trafficking to lysosomes for degradation is a major Aβ clearance pathway that is differentially regulated by apoE isoforms. A disturbance of this pathway can lead to accumulation and aggregation of cellular Aβ capable of causing neurotoxicity and seeding amyloid.

Introduction

Accumulation and aggregation of amyloid-β (Aβ)² peptides cleaved from amyloid precursor protein (APP) are likely initiating events in the pathogenesis of Alzheimer disease (AD) (1, 2). Overproduction or impaired clearance can both lead to Aβ accumulation. Recent evidence indicates that late-onset AD cases, referring to patients who develop AD after the age 65, are likely caused by an overall impairment in Aβ clearance (3). Genetically, among the three polymorphic alleles (ϵ2, ϵ3, and ϵ4), the ϵ4 allele of the apolipoprotein E (APOE) gene is the strongest genetic risk factor for late-onset AD (4, 5). Although the primary function of apoE in the brain is to mediate cholesterol transport through apoE receptors, apoE is readily associated with Aβ in AD brains (4). In vivo experiments have shown that apoE4 significantly decreases Aβ clearance when compared with apoE3 in amyloid model mice without affecting Aβ production (6). However, how apoE isoforms differentially regulate Aβ clearance is not clear.

A major Aβ clearance pathway is cellular uptake by different cell types in brain parenchyma and in cerebral vasculature (4). Recent genome-wide association studies have also identified several endocytosis-related genes (BIN1, CD2AP, PICALM, CD33) that are closely related to the risk of AD in addition to APOE (7–9). Endocytosis, including receptor-mediated endocytosis and fluid phase pinocytosis, is an efficient pathway by which extracellular proteins gain entry into intracellular compartments. The majority of endocytosed proteins traffic through several endocytic compartments before being delivered to lysosomes for degradation. Selected ligands can also be delivered to specialized organelles or recycled back to extracellular space. Rab GTPase family members play critical roles in mediating vesicular transport to different compartments. In particular, endocytic vesicles acquire Rab5 for their fusion with early endosomes, from where specific cargo proteins are either delivered to recycling endosomes through the function of Rab11 or transported to late endosomes/multivesicular bodies in a manner that depends on the function of Rab7 (10).

Several studies have shown that endocytosed Aβ can be delivered to lysosomes (11–14); however, how Aβ traffics through various endocytic compartments to reach lysosomes and what is the fate of Aβ once it is delivered there are not clear. Thus, we systematically dissected the itinerary of Aβ endocytic trafficking in neuronal cells using various organelle markers. We also assessed the effects of altered endocytic trafficking and function on cellular Aβ distribution using both wild-type and mutant forms of Rab proteins (Rab5, Rab11, and Rab7). The effects of lysosomal enzyme inhibitors on Aβ degradation and aggregation were also assessed. Finally, we compared the differential roles of apoE3 and apoE4 on Aβ uptake and degradation. Our studies established a primary trafficking pathway of Aβ to lysosomes for degradation, revealed a previously undefined recycling pathway through Rab11-positive compartments, and defined differential functions of apoE isoforms in regulating Aβ cellular uptake and endocytic trafficking.

EXPERIMENTAL PROCEDURES

Reagents

Aβ42, FAM-Aβ42, and tetramethylrhodamine (TMR)-Aβ42 were purchased from AnaSpec (Fremont, CA). Aβ peptides were pretreated with trifluoroacetic acid, distilled under nitrogen, washed with 1,1,1,3,3,3-hexafluoro-2-propanol, distilled under nitrogen, stored at −20 °C, and dissolved in dimethyl sulfoxide (DMSO) to 200 μm before use. Aβ42 solution was centrifuged at 17,000 × g for 30 s to remove aggregated Aβ. Leupeptin, pepstatin A, and E-64d were purchased from Sigma. Rabbit-anti-Rab5A antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit anti-Rab7 and rabbit anti-Rab11 antibodies were purchased from Cell Signaling (Danvers, MA). Rat anti-LAMP1 antibody was purchased from BD Biosciences. Mouse anti-apoE antibody was purchased from Abcam (Cambridge, MA). Alexa Fluor 568-labeled human transferrin (Tf), DiI-low-density lipoprotein (LDL), and LysoTracker Red were purchased from Invitrogen. Heparin was purchased from Baxter Healthcare Corp. (Round Lake, IL). Recombinant human apoE3 and apoE4 proteins were generously provided by Dr. Yadong Huang (Gladstone Institutes, San Francisco, CA). GFP-Rab5-WT, Rab5-DN (S34N), and Rab5-CA (Q79L) plasmids were kindly provided by Dr. Steve Caplan (University of Nebraska, Lincoln, NE), GFP-Rab11-CA (Q70L) plasmid was kindly provided by Dr. Wim Annaert (VIB Center for the Biology of Disease, Leuven, Belgium), and GFP-Rab7-WT (Addgene plasmid 12605), Rab7-DN (T22N) (Addgene plasmid 12660), Rab11-WT (Addgene plasmid 12674), and Rab11-DN (S25N) (Addgene plasmid 12678) plasmids (15) were kindly provided by Dr. Richard Pagano (Mayo Clinic, Rochester, MN) and obtained from Addgene.

Cell Culture and Confocal Microscopy

Mouse neuroblastoma N2a cells were cultured in DMEM/Opti-MEM I (1:1) (Gibco) medium supplemented with 5% fetal bovine serum (Gibco) and maintained at 37 °C in humidified air containing 5% CO₂. Primary cortical neurons were obtained from 17-day-old embryos of wild-type C57BL/6 mice and grown in Neurobasal medium (Gibco) supplemented with 0.5 mm GlutaMAX (Gibco), 2% B27 (Gibco), and 1% penicillin-streptomycin (Invitrogen), in cell culture dishes precoated with poly-d-lysine solution (100 μg/ml). At 5 days in vitro, neurons were treated with 10 μm cytosine arabinofuranoside (Sigma) for 2 days to remove glial cells, and media were changed to fresh Neurobasal medium containing B27 and penicillin-streptomycin at 7 days in vitro. In some experiments, cells were cultured on 8-well chambered cover glasses (Nalge Nunc International, Rochester, NY) and observed by a confocal laser-scanning fluorescence microscope (model LSM 510 invert; Carl Zeiss, Jena, Germany). The pinhole size was set to 1 airy unit. Co-localization in the images was quantified by ImageJ software (National Institutes of Health), and the Manders' coefficient was calculated with JACoP (rsbweb.nih.gov/ij/plugins/track/jacop.html).

Transfection of Rab Plasmids

N2a cells were transfected with plasmids using Lipofectamine 2000 by adding prepared complexes directly to a suspension of cells in the plate, according to the manufacturer's instructions. Medium was changed on the following day. Cells were used 36 h after the start of transfection.

Detection of Cell-associated Aβ and ApoE by ELISA

After incubation with Aβ42 (500 nm) and/or apoE (500 nm) in serum-free medium for 24 h at 37 °C, cells were harvested by incubating with trypsin for 5 min at 37 °C. Cell pellets were collected by centrifugation at 1,000 × g for 5 min. After washing two times with PBS, cells were dissolved in 5 m guanidine in 50 mm Tris-HCl (pH 8.0). To measure human Aβ42, samples were captured with mAb 2.1.3 antibody followed by detection with HRP-conjugated Ab5 antibody (16). To measure human apoE, samples were captured using AB947 antibody (Millipore) and detected with biotin-conjugated goat anti-apoE antibody (Meridian Life Science) and poly-HRP-conjugated streptavidin (Fitzgerald). In some experiments, lysosomal inhibitors (pepstatin A, 10 μm; leupeptin, 100 μm; E-64d, 50 μm) were added during incubation with Aβ42 and/or apoE.

In Vitro Amyloid Seeding Assay

After incubation with Aβ42 (500 nm) in the presence of lysosomal inhibitors in serum-free medium for 24 h at 37 °C, cells were harvested by trypsin, washed twice with PBS, and homogenized in 2% SDS in PBS. In vitro seeding assay was performed as described previously (17). Briefly, freshly prepared TMR-Aβ42 (100 nm) was incubated with cell homogenates on 8-well chambered cover glasses for 48 h, and amyloid aggregates were observed using a confocal microscope. For each chamber, the fluorescence intensity of Aβ aggregates in nine randomly selected fields (140 × 140 μm) was quantified by ImageJ software.

Fluorescence-activated Cell Sorter (FACS)-based Binding Assay

Cells were plated onto 12-well plates, allowed to grow to 90% confluency and then detached from the plates using cell dissociation solution (Sigma), and incubated with FAM-Aβ42 (2 μm), together with vehicle control and recombinant apoE3 or apoE4 protein (500 nm), in the presence or absence of heparin (30 units/ml) at 4 °C for 2 h in PBS with 1.5% heat-inactivated FBS and 0.1% sodium azide. Then cells were washed twice with PBS/FBS. Samples (1 × 10⁴ cells/sample) were analyzed for fluorescence on a BD FACSCalibur (BD Biosciences). Control cells without any exposure to fluorescence were used to assess background fluorescence.

Statistical Analysis

All quantified data represent an average of triplicate samples. Statistical significance was determined by Student's t test, and p < 0.05 was considered significant.

RESULTS

Aβ Traffics through Rab5- and Rab7-positive Endosomal Compartments

To assess Aβ endocytic trafficking, we examined the cellular localization of TMR-Aβ42 in N2a cells transiently transfected with GFP vector control, GFP-Rab5-WT, a GFP-Rab5-dominant negative mutant (Rab5-DN, S34N), or a GFP-Rab5-constitutively active mutant (Rab5-CA, Q79L). After a 24-h incubation of cells with TMR-Aβ42, 6.3 ± 1.1% of Aβ was co-localized with Rab5-WT-positive early endosomes (Fig. 1, A and B). Aβ42 was more diffusely distributed and less co-localized with Rab5-DN mutant (0.4 ± 0.2%) (Fig. 1, A and B). In the case of the Rab5-CA mutant, Aβ42 was accumulated in Rab5-CA-positive compartments, in particular in enlarged endosomes. The co-localization of Aβ42 was higher in Rab5-CA-positive compartments (15.1 ± 1.8%) (Fig. 1, A and B) when compared with Rab5-WT. When the amounts of cell-associated Aβ42 were quantified by ELISA after a 24-h incubation with native, unlabeled Aβ42, expression of Rab5-DN and Rab5-CA, but not Rab5-WT, significantly increased cell-associated Aβ42 when compared with control (Fig. 1C). To further define Aβ42 endocytic trafficking, we used similar approaches to assess the trafficking of Alexa Fluor 568-conjugated Tf, which traffics through early and recycling endosomes. When compared with Aβ42, more Tf was detected in Rab5-WT-positive compartments (48.2 ± 6.8%) (Fig. 1, D and E). Tf was also captured in Rab5-CA-positive compartments (59.2 ± 6.5%) (Fig. 1, D and E). Similar to Aβ42, expression of Rab5-DN mutant decreased Tf co-localization in these compartments (6.4 ± 1.6%) (Fig. 1, D and E). These results indicate that cell-internalized Aβ42 traffics through Rab5-positive early endosomes. A disturbance of endosomal functions, as observed in AD brains (18, 19), will likely lead to Aβ42 accumulation in these intracellular compartments.

FIGURE 1.

Aβ42 traffics through Rab5-positive early endosomal compartments. A, N2a cells were transiently transfected with GFP vector (Control), GFP-Rab5-WT, GFP-Rab5-DN, or GFP-Rab5-CA and incubated with TMR-Aβ42 (500 nm) for 24 h. Cells were observed by fluorescence confocal microscopy, and representative cells are shown. The inset shows a high magnification view of Aβ/Rab5 co-localization. Scale bar: 10 μm. B, the percentages of Aβ that co-localized with Rab5 in comparison with total Aβ were quantified by ImageJ. Data are plotted as mean ± S.E. (n = 9). C, N2a cells transiently transfected with the same set of constructs as in A were incubated with unlabeled Aβ42 (500 nm) for 24 h, and the amounts of cell-associated Aβ were quantified by ELISA. Data are plotted as mean ± S.E. (n = 3). N.S., not significant; *, p < 0.05. Con, control. D, N2a cells transiently transfected with the same set of constructs as in A were incubated with Alexa Fluor 568-conjugated Tf (20 μg/ml) for 24 h. Cells were observed by fluorescence confocal microscopy with representative images shown. Scale bar: 10 μm. E, the percentages of Tf that co-localized with Rab5 in comparison with total Tf were quantified by ImageJ. Data are plotted as mean ± S.E. (n = 9).

Next, we analyzed the effects of overexpressing GFP-fused WT or a DN mutant (T22N) of Rab7 on cellular localization of internalized Aβ42. When DiI-labeled LDL, which is transferred from early to late endosomes/lysosomes, was used as a control, Rab7-DN expression resulted in a disturbed trafficking of LDL (data not shown). In the case of Aβ, 45.8 ± 3.9% of TMR-Aβ42 was co-localized with Rab7-WT-positive late endosomal compartments, indicating that Aβ traffics mainly to Rab7-positive late endosomes, whereas only 10.4 ± 1.6% of Aβ was transferred to Rab7-DN-positive compartments. ELISA showed that expression of neither Rab7-WT (1.09 ± 0.06-fold) nor Rab7-DN (1.10 ± 0.17-fold) affected the amounts of cell-associated Aβ42 when compared with control, suggesting that Aβ might traffic to other endocytic pathways when the late endocytic pathway is perturbed.

A Small Portion of Aβ Traffics through Rab11-positive Recycling Endosomes

To investigate whether recycling endosomes participate in the endocytic trafficking of Aβ, we examined the effects of expressing GFP-fused Rab11-WT, Rab11-DN (S25N) or Rab11-CA (Q70L) mutant on Aβ42 trafficking in N2a cells (Fig. 2, A–C). Approximately half of internalized Tf was transported into Rab11-WT-positive endosomes, whereas Rab11-CA facilitated and Rab11-DN suppressed Tf accumulation in Rab11-positive compartments (Fig. 2, D and E). A small portion of internalized TMR-Aβ42 was co-localized with Rab11-WT (6.7 ± 0.9%), and the expression of Rab11-CA increased this co-localization up to 17.4 ± 2.5% (Fig. 2, A and B). A negligible amount of Aβ42 was co-localized with Rab11-DN (Fig. 2, A and B). ELISA showed that cell-associated Aβ42 level was significantly increased in cells expressing Rab11-CA when compared with control (Fig. 2C), confirming that some Aβ42 indeed traffics through Rab11-positive recycling endosomes. These results indicate that despite a primary endocytic trafficking pathway through early and late endosomes, a portion of cell-internalized Aβ42 can traffic through the recycling endosomes.

FIGURE 2.

Aβ42 traffics through Rab11-positive recycling endosomal compartments. A, N2a cells were transiently transfected with GFP vector (Control), GFP-Rab11-WT, GFP-Rab11-DN, or GFP-Rab11-CA and incubated with TMR-Aβ42 (500 nm) for 24 h. Cells were observed by fluorescence confocal microscopy, and representative cells are shown. Insets show high magnification views of Aβ/Rab11 co-localization. Scale bar: 10 μm. B, the percentages of Aβ that co-localized with Rab11 in comparison with total Aβ were quantified by ImageJ. Data are plotted as mean ± S.E. (n = 9). C, N2a cells transiently transfected with the same set of constructs as in A were incubated with native, unlabeled Aβ42 (500 nm) for 24 h, and the amounts of cell-associated Aβ were quantified by ELISA. Data are plotted as mean ± S.E. (n = 3). N.S., not significant; *, p < 0.05. Con, control. D, N2a cells transiently transfected with the same set of constructs as in A were incubated with Alexa Fluor 568-conjugated Tf (20 μg/ml) for 24 h. Cells were observed by fluorescence confocal microscopy with representative images shown. Scale bar: 10 μm. E, the percentages of Tf that co-localized with Rab11 in comparison with total Tf were quantified by ImageJ. Data are plotted as mean ± S.E. (n = 9).

Internalized Aβ42 Is Co-localized with Rab5, Rab7, Rab11, and LAMP1 in Primary Neurons

To confirm that Aβ traffics through endosomal/lysosomal pathways in neurons, we performed immunostaining with antibodies against Rab5A, Rab11, Rab7, or LAMP1 in mouse primary cultured cortical neurons after incubations with FAM-Aβ42 (500 nm) (Fig. 3). We found that Aβ42 partially co-localized with Rab5A in neurites and cell body after a 1-h incubation (Fig. 3A). After a 3-h incubation, Aβ42 was detected in Rab11- and Rab7-positive compartments (Fig. 3, B and C). Finally, Aβ42 was co-localized with lysosomal marker LAMP1 in the cell body, but not in the neurites, after a 24-h incubation (Fig. 3D). These results confirm that internalized Aβ42 is transported to either the endosome/lysosome degradation pathway or the recycling pathway in neurons.

FIGURE 3.

Co-localization of internalized Aβ42 with Rab5, Rab11, Rab7, and LAMP1 in primary neurons. Mouse primary cortical neurons were incubated with FAM-Aβ42 (500 nm) for 1 h (Rab5), 3 h (Rab11 and Rab7), or 24 h (LAMP1). Following immunostaining with antibodies to Rab5A (A), Rab11 (B), Rab7 (C), or LAMP1 (D), neurons were observed by fluorescence confocal microscopy with representative neuronal cell bodies and neurites shown. The panels to the right are high magnification views of neurites from the white rectangle areas in the original images. The insets show high magnification views of cell bodies from the white square areas in the original images. Scale bar: 10 μm.

Inhibition of Lysosomal Degradation Increases Cell-associated Aβ and Accelerates the Formation of Aβ Aggregates

To assess lysosomal trafficking and degradation of Aβ, we tested the effects of lysosomal inhibitors (leupeptin, pepstatin A, and E-64d), which block lysosomal enzyme activities (20). N2a cells were transfected with Rab5-WT, Rab7-WT, Rab11-WT, or vector (control) and then incubated with Aβ42 in the presence or absence of lysosomal inhibitors for 24 h. ELISA showed that the presence of lysosomal inhibitors significantly increased cell-associated Aβ42 under all conditions (Fig. 4A), indicating that lysosomal degradation of Aβ42 is a significant event following cellular Aβ42 uptake. Expression of Rab5-WT, Rab7-WT, or Rab11-WT did not affect cell-associated Aβ42 levels in the absence of lysosomal inhibitors. However, in the presence of lysosomal inhibitors, expression of Rab5-WT and Rab7-WT, but not Rab11-WT, further increased cell-associated Aβ42 when compared with control (Fig. 4A). These results indicate that up-regulation of Rab5 or Rab7 expression enhances Aβ transport to lysosomes where Aβ is efficiently degraded. They also suggest that a disturbance of lysosomal enzyme functions could lead to intracellular Aβ accumulation.

FIGURE 4.

Aβ degradation or aggregation in the lysosomes. A, N2a cells were transiently transfected with GFP vector (Con), GFP-Rab5-WT, GFP-Rab7-WT, or GFP-Rab11-WT and incubated with Aβ42 (500 nm) in the presence or absence of lysosomal inhibitors (LI) for 24 h. The amounts of cell-associated Aβ were quantified by ELISA. Data are plotted as mean ± S.E. (n = 3). N.S., not significant; *, p < 0.05. B, N2a cells transfected and treated with Aβ42 as in A were homogenized and incubated with TMR-Aβ42 (100 nm) for 48 h. Aβ aggregates were observed by fluorescence confocal microscopy with representative fields shown. Cells without Aβ treatment were used as a negative control (panel a), and preformed Aβ fibrils were used as a positive control (panel f). Scale bar: 20 μm. C, the amounts of Aβ aggregates were quantified by ImageJ software as shown as combined fluorescence intensity from nine different fields of one chamber. Data are plotted as mean ± S.E. (n = 3). N.S., not significant, *, p < 0.05.

To examine whether cell accumulated Aβ42 possesses “seeding effect” to further facilitate Aβ42 aggregation, we tested the ability of cell-associated Aβ42 to seed aggregates in vitro. N2a cells transfected with vector (control), Rab5-WT, Rab7-WT, or Rab11-WT were incubated with Aβ42 (500 nm) in the presence of lysosomal inhibitors for 24 h, homogenized, and incubated with freshly prepared TMR-Aβ42 (100 nm) for 48 h. When fluorescent amyloid aggregates were observed and quantified by confocal microscopy, cell lysates from Rab5-WT- and Rab7-WT-expressing, but not Rab11-WT-expressing, cells induced more Aβ aggregates when compared with controls (Fig. 4, B and C), suggesting that lysosomal accumulated Aβ42 has sufficient ability to seed amyloid aggregates.

ApoE Facilitates Aβ Lysosomal Trafficking and Degradation in an Isoform-dependent Manner

ApoE4 is a strong risk factor for AD (4). To investigate the role of apoE isoforms in neuronal Aβ endocytic trafficking, we incubated N2a cells with Aβ42, together with vehicle control, apoE3, or apoE4 for 24 h, in the presence or absence of lysosomal inhibitors. Cell-associated Aβ42 levels were significantly increased in the presence of apoE, with apoE3 showing a greater effect than apoE4 (Fig. 5A). In the presence of lysosomal inhibitors, cell-associated Aβ42 levels were further increased by apoE3 and apoE4 (Fig. 5A), again with apoE3 having a greater effect than apoE4. When lysosomal degradation of Aβ42 was calculated by subtracting the cell-associated Aβ in the absence of inhibitors from that in the presence of inhibitors, apoE3 enhanced Aβ42 lysosomal degradation by 1.75-fold when compared with apoE4 (Fig. 5B). In addition, confocal microscopy also revealed that internalized FAM-Aβ42, which was largely co-localized with LysoTracker, was greatly increased in the presence of apoE (Fig. 5C). Upon quantification, we confirmed that the amounts of Aβ42 in lysosomes were higher in apoE3-treated cells than apoE4-treated cells (Fig. 5D).

FIGURE 5.

ApoE facilitates cellular Aβ uptake, lysosomal trafficking, and degradation in an isoform-dependent manner. A, N2a cells were incubated with Aβ42 (500 nm), together with vehicle control (Con), apoE3 (E3), or apoE4 (E4) (500 nm), in the presence or absence of lysosomal inhibitors for 24 h, and the amounts of cell-associated Aβ were quantified by ELISA. Data are plotted as mean ± S.E. (n = 3). *, p < 0.05; **, p < 0.01. B, the amounts of degraded Aβ were calculated by subtracting the mean values of cell-associated Aβ in the absence of inhibitors from the values in the presence of inhibitors. Data are plotted as mean ± S.E. (n = 3). *, p < 0.05; **, p < 0.01. C, cells were incubated with FAM-Aβ42 (500 nm), together with vehicle control, apoE3, or apoE4 (500 nm), in the presence of lysosomal inhibitors for 24 h. LysoTracker Red (75 nm) was included in the final hour of incubation prior to processing for fluorescence microscopy. Scale bar: 10 μm. D, the amounts of Aβ that co-localized with lysosomes were quantified by ImageJ. Data are plotted as mean ± S.E. (n = 9). *, p < 0.05; **, p < 0.01.

To address whether apoE isoforms differentially affect Aβ42 binding to the cell surface, we carried out a FACS-based binding assay in the presence or absence of apoE3 or apoE4. Both apoE isoforms enhanced Aβ42 binding to the cell surface; however, apoE3 increased cell surface-bound Aβ42 more than apoE4 (1.7-fold versus 1.4-fold when compared with that of control, respectively) (Fig. 6). Consistent with our previous finding that Aβ42 binds to cell surface heparan sulfate proteoglycans (HSPG) (13), the differential effects of apoE isoforms on Aβ42 binding were eliminated in the presence of heparin (Fig. 6). These results demonstrate that apoE facilitates Aβ42 binding to cell surface HSPG in an isoform-dependent manner.

FIGURE 6.

ApoE increases Aβ cell surface binding to HSPG in an isoform-dependent manner. N2a cells were incubated with FAM-Aβ42 (2 μm), together with vehicle control, apoE3 (E3), or apoE4 (E4) (500 nm), in the presence or absence of heparin (30 units/ml) at 4 °C for 2 h. Cell-bound Aβ was quantified by FACS. Data are plotted as mean ± S.E. (n = 3). N.S., not significant; *, p < 0.05; **, p < 0.01.

To further dissect the mechanisms underlying the effects of apoE isoforms on Aβ42, we analyzed apoE lysosomal trafficking and degradation. We found that apoE3 was transported into lysosomes and degraded more efficiently than apoE4 (Fig. 7, A and B). The more efficient trafficking of apoE3 than apoE4 was confirmed by immunostaining and confocal microscopy (Fig. 7, C and D). These results suggest that apoE isoforms likely facilitate Aβ42 lysosomal trafficking and degradation by serving as Aβ42 trafficking chaperones.

FIGURE 7.

ApoE3 lysosomal trafficking and degradation is more efficient than apoE4. A, N2a cells were prepared as in Fig. 5A, and the amounts of cell-associated apoE were quantified by ELISA. E3, apoE3; E4, apoE4. Data are plotted as mean ± S.E. (n = 3). **, p < 0.01. B, the amounts of degraded apoE were calculated by subtracting the mean values of cell-associated apoE in the absence of inhibitors from the values in the presence of inhibitors. Data are plotted as mean ± S.E. (n = 3). **, p < 0.01. C, cells were incubated with Aβ42 (500 nm), together with vehicle control, apoE3, or apoE4 (500 nm), in the presence of lysosomal inhibitors for 24 h, immunostained with antibodies against apoE and LAMP1, and observed by fluorescence confocal microscopy. Scale bar: 10 μm. D, the amounts of apoE that co-localized with lysosomes were quantified by ImageJ. Data are plotted as mean ± S.E. (n = 9). *, p < 0.05.

DISCUSSION

In the brains of AD patients, Aβ deposition is more abundantly detected in APOE ϵ4 carriers than noncarriers (21–23). Although several Aβ-independent mechanisms exist, the primary pathway by which apoE4 increases the risk of AD is likely to facilitate Aβ accumulation and aggregation in the brain (24). In this study, we defined the cellular itinerary of Aβ endocytic trafficking and lysosomal degradation. More importantly, we demonstrated an apoE isoform-dependent effect on Aβ lysosomal trafficking and degradation in neuronal cells. Although both apoE3 and apoE4 facilitate cellular Aβ binding and subsequent lysosomal trafficking and degradation, apoE3 consistently exhibits greater effects than apoE4. The differences of cell-associated Aβ between apoE3 and apoE4 were further exacerbated in the presence of lysosomal inhibitors, indicating that apoE3 possesses greater ability to accelerate lysosomal degradation of Aβ than apoE4. Recently, Castellano et al. (6) have used an in vivo microdialysis technique to demonstrate an apoE isoform-dependent Aβ clearance in PDAPP/TRE mice, with higher concentration of interstitial fluid Aβ and slower Aβ clearance rate in the hippocampus of PDAPP/E4 mice than in PDAPP/E3 mice, suggesting a lesser role of apoE4 in supporting Aβ clearance than apoE3. Our in vitro findings demonstrating a greater ability of apoE3 to facilitate Aβ clearance than apoE4 are consistent with these in vivo results.

A possible mechanism by which apoE isoforms differentially accelerate Aβ cellular uptake and lysosomal trafficking is to enhance Aβ binding to cell surface receptors. Our recent work has shown that cell surface HSPG constitutes a major Aβ binding site on the neuronal cell surface (13). As apoE is also a high affinity heparin-binding protein (25), it is possible that apoE isoforms differentially affect Aβ binding to cell surface HSPG. Indeed, our results clearly show that apoE3 increases Aβ binding to neuronal cell surface more than apoE4 in a manner that depends on the availability of HSPG. Whether such effects depend on a direct interaction between apoE and Aβ is not clear. Previous in vitro studies have demonstrated that apoE3 binds to Aβ with higher affinity than apoE4 (26, 27), and native apoE3 forms a larger amount of SDS-stable complex with Aβ than apoE4 (28). Thus, it is possible that apoE3 more readily forms apoE/Aβ complexes than apoE4 and that these apoE/Aβ complexes have a greater affinity to cell surface HSPG than Aβ alone. Alternatively, binding of apoE isoforms to cell surface HSPG generates higher affinity binding sites for Aβ. Finally, binding of apoE isoforms to cell surface HSPG might alter its conformation and/or charge properties such that it is more prone to Aβ binding. Further studies are needed to address these possibilities.

Receptor-mediated endocytosis often leads to trafficking through early/late endosomes en route to lysosomes for degradation as for the case of LDL and a variety of other extracellular ligands. In this work, we defined the endocytic trafficking pathway by which extracellular Aβ is eventually delivered to lysosomes. We found that the majority of internalized Aβ traffics through Rab5- and Rab7-positive early and late endosomes, respectively. Most internalized Aβ is delivered to the lysosomal pathway for degradation. Because both overexpression of Rab5 and overexpression of Rab7 accelerate Aβ trafficking to lysosomes and its degradation, the endosome/lysosome route likely represents a major endocytic trafficking pathway for Aβ. Interestingly, we also identified a co-localization of Aβ with recycling endosome marker Rab11, albeit to a smaller extent when compared with those for Rab5 and Rab7, suggesting that at least a portion of endocytosed Aβ can be recycled. To our knowledge, this is the first study demonstrating a potential pathway for Aβ recycling. The physiological and pathophysiological significance of this potential Aβ recycling pathway is not clear.

Autophagy also plays an important role in AD pathogenesis where Aβ is subsequently degraded by autophagolysosomes/lysosomes (29, 30). Autophagy, receptor-mediated endocytosis, and pinocytosis (fluid phase endocytosis) converge in lysosomes where a variety of acid hydrolases, including cathepsin B and cathepsin D, serve as Aβ-degrading enzymes (31). These findings suggest that lysosomes have a high capacity to degrade Aβ and that a disturbance of this Aβ degradation pathway could critically affect Aβ metabolism. As high concentrations of Aβ (32) and the acidic environment in the endosomal/lysosomal compartments (33) promote Aβ aggregation, a disturbance of lysosomal function can lead to harmful Aβ aggregation. Such a possibility is highlighted in our current study where the presence of lysosomal inhibitors leads to Aβ lysosomal accumulation and aggregation. More importantly, these Aβ aggregates are capable of seeding additional Aβ aggregates, presenting a possibility that Aβ aggregation and deposition in AD brains might be initiated within the lysosomes. Pathological conditions such as increased Aβ production and/or impaired endosomal/lysosomal functions can lead to similar Aβ aggregates in the lysosomes. In addition to seeding amyloid, intracellular Aβ aggregates might also be exocytosed to the extracellular environments where they injure synapses and trigger neurodegeneration. In this regard, mounting recent work has demonstrated a greater toxicity associated with soluble Aβ oligomers (34–38). Whether Aβ oligomers are generated intracellularly, extracellularly, or both requires further investigation. Aβ aggregates also destabilize the lysosomal membrane (39), which further compromises lysosomal function and induces leakage of hydrolases into cytoplasmic compartment, contributing to ultimate neuronal death (40, 41). Several studies have shown that endosome/lysosome abnormality, including their enlargement, is often induced in early stages of AD (18, 19, 42, 43). Furthermore, presenilin 1 (PSEN1) mutations, causing early-onset AD, induce a disturbed lysosomal/autophagy phenotype by affecting lysosome acidification and proteolysis (44). Together, these evidences suggest that a disturbance of lysosome functions could be the first step to induce Aβ aggregation in lysosomes, which results in further damage to lysosomes, generating a vicious cycle in the pathogenesis of AD.

In summary, we have established an Aβ endocytic trafficking itinerary that includes a major pathway to the lysosomes for degradation and a minor pathway to Rab11-positive compartments for recycling. We have also demonstrated that apoE facilitates Aβ binding and lysosomal trafficking in an isoform-dependent manner, with apoE3 more efficiently enhancing Aβ trafficking and degradation than apoE4. Our work suggests a possibility that early endosome/lysosome dysfunctions likely lead to Aβ intracellular accumulation and aggregation, initiating a cascade of events that ultimately cause injuries to synapses and neurons. Better understanding of apoE isoform-dependent effects on Aβ metabolism might allow us to address why apoE4 is a strong risk factor for AD and how we can identify new targets for AD diagnosis and therapy.