Co-aggregation with Apolipoprotein E modulates the function of Amyloid-β in Alzheimer's disease

Abstract

Which isoforms of apolipoprotein E (apoE) we inherit determine our risk of developing late-onset Alzheimer's Disease (AD), but the mechanism underlying this link is poorly understood. In particular, the relevance of direct interactions between apoE and amyloid-β (Aβ) remains controversial. Here, single-molecule imaging shows that all isoforms of apoE associate with Aβ in the early stages of aggregation and then fall away as fibrillation happens. ApoE-Aβ co-aggregates account for ~50% of the mass of diffusible Aβ aggregates detected in the frontal cortices of homozygotes with the higher-risk APOE4 gene. We show how dynamic interactions between apoE and Aβ tune disease-related functions of Aβ aggregates throughout the course of aggregation. Our results connect inherited APOE genotype with the risk of developing AD by demonstrating how, in an isoform- and lipidation-specific way, apoE modulates the aggregation, clearance and toxicity of Aβ. Selectively removing non-lipidated apoE4-Aβ co-aggregates enhances clearance of toxic Aβ by glial cells, and reduces secretion of inflammatory markers and membrane damage, demonstrating a clear path to AD therapeutics.

Fig. 1. ApoE and Aβ transiently co-aggregate en route to fibrils.

A Aβ42 aggregation (4 µM) in the presence of different non-lipidated isoforms of apoE (0 or 80 nM), monitored by ThT fluorescence (n = 3 independent replicates). B Time points at which samples were taken for further analysis. C SiMPull assay for Aβ42 aggregates and apoE-Aβ co-aggregates (1 µM Aβ monomer equivalents), using biotinylated 6E10 antibody for capture, and Alexa-Fluor-647-labeled 6E10 (500 pM) and Alexa-Fluor-488-labeled EPR19392 (1 nM) antibodies for detection. D–F Representative two-color TIRF images of aggregates were captured at t₁ (D), t₂ (E), and t₃ (F). G, H Colocalization between Aβ and apoE at different time points, quantified by aggregate counting (G) and 6E10 fluorescence intensity (H). Data are plotted as the mean and standard deviation of three independent replicates. I Average sizes of colocalized and non-colocalized aggregates (N.B. in diffraction-limited imaging, the minimum apparent aggregate size is ~0.2 µm²). Source data are provided as a Source Data file.

Fig. 2. ApoE interacts with Aβ in early-stage aggregates but not fibrils.

Aβ42 aggregates and apoE-Aβ co-aggregates were co-immunoprecipitated using an Aβ-specific antibody (6E10) and then analyzed by western blotting using both Aβ-specific 6E10 and apoE-specific EPR19392 antibodies. This image represent one of three independent experiments.

Fig. 3. Aβ-apoE co-aggregates form in human brain tissue, but their concentration is isoform-dependent.

A Representative two-color TIRF images of aggregates from frontal-cortex extracts of homozygous APOE4 and APOE3 AD patients. B Number of Aβ-containing species captured from APOE4 and APOE3 homozygotes. C, D Colocalization between Aβ and apoE in extracts from APOE4 and APOE3 homozygotes, quantified by aggregate counting (C) and 6E10 (Aβ) fluorescence intensity (D). E Sizes of apoE-colocalized and non-colocalized Aβ aggregates in APOE4 and APOE3 homozygotes. Data points in panels (B–E) are plotted as the mean +/− standard deviation of three biological replicates. Statistical significance was calculated using a unpaired two-sample t-test. *P < 0.05, **P < 0.01. Source data are provided as a Source Data file.

Fig. 4. Clearance of apoE-Aβ co-aggregates by glial cells depends on aggregate maturity, apoE isoform and lipidation.

A Assay for uptake of aggregates by glial cells. B, C Representative two-color epifluorescence images showing uptake of Aβ42 (1 µM monomer equivalents) and non-lipidated apoE (0 or 80 nM) from aggregation mixtures at t₁ (end of lag phase) by iMGLs (B) and iAstrocytes (C). D, E Quantified uptake of Aβ by iMGLs (D) and iAstrocytes (E). Units of uptake = integrated fluorescence of sample divided by the integrated fluorescence of internalized early-stage Aβ42 aggregates at t₁ for each replicate. Data points are plotted as the mean +/− standard deviation of three biological replicates. Statistical significance was calculated using a unpaired two-sample t-test. *P < 0.05, **P < 0.01, ***P < 0.001, ns, non-significant (P  ≥  0.05). Source data are provided as a Source Data file.

Fig. 5. Early-stage, non-lipidated apoE-Aβ co-aggregates inflame glial cells in an isoform-dependent way.

A Aggregate-induced cytokine and chemokine secretion by iMGLs and iAstrocytes. B–I MCP-1, IL-1β, IL-6 and TNF-α release by iMGLs (B–E) and iAstrocytes (F–I) induced by early-stage (t₁) and fibrillar (t₃) aggregates. Each data point represents one of three biological replicates; error bars represent mean values +/− standard deviation. Statistical significance was calculated using a unpaired two-sample t-test. *P < 0.05, **P < 0.01, ***P < 0.001, ns, non-significant (P ≥ 0.05). Source data are provided as a Source Data file.

Fig. 6. ApoE modulates the toxicity of Aβ aggregates isoform-dependently.

A, B Permeabilization of lipid bilayers by early-stage (t₁) and fibrillar (t₃) Aβ aggregates ([Aβ42] = 4 µM in monomer equivalents; [apoE] = 0 or 80 nM). Ca²⁺ influx is referenced to the influx caused by the ionophore, ionomycin. C, D Neurotoxicity of t₁ and t₃ aggregates to human neuroblastoma SH-SY5Y cells, assayed by lactate dehydrogenase (LDH) release ([Aβ42] = 4 µM in monomer equivalents; [apoE] = 0 or 80 nM). Data points represent one of three independent experiments (B) or one of three or four biological replicates (D); error bars represent mean values +/− standard deviation of three independent (C) or four biological (D) replicates. Statistical significance was calculated using a unpaired two-sample t-test. *P < 0.05, **P < 0.01, ***P < 0.001, ns, non-significant (P ≥ 0.05). Source data are provided as a Source Data file.

Fig. 7. Removing non-lipidated apoE4-Aβ co-aggregates enhances Aβ clearance, reducing secretion of inflammatory markers and cytotoxicity.

A Immunoprecipitation of non-lipidated early-stage co-aggregates from a mixture of non-lipidated and lipidated apoE4-Aβ co-aggregates using the HAE-4 antibody. B Representative two-color TIRF images and sizes of individual apoE4-Aβ co-aggregates before (number of aggregates: 10,277, average size: 0.47 ± 1.95 µm²) and after (number of aggregates: 4354, average size: 0.23 ± 0.39 µm²) immunoprecipitation (data are plotted in log₁₀ scale in inset). C–L Effect of immunoprecipitation on Aβ uptake and cytokine/chemokine release by iMGLs (C–G) and iAstrocytes (H–L). M, N Effect of immunoprecipitation on neurotoxicity of apoE4-Aβ co-aggregates, measured by permeabilization of lipid membranes (M) and LDH release by human neuroblastoma SH-SY5Y cells (N). Data points represent one of three biological replicates (C–L, N) or one of three independent experiments (M); Units of uptake = integrated fluorescence of sample divided by the integrated fluorescence of internalized early-stage Aβ42 aggregates at t₁ for each replicate, as in Fig. 3 (C, H); error bars represent mean values +/− standard deviation of three biological replicates. Statistical significance was calculated using one-way ANOVA with post-hoc Tukey. *P < 0.05, **P < 0.01, ***P < 0.001, ns, non-significant (P ≥ 0.05). Source data are provided as a Source Data file.

Fig. 8. Early-stage apoE4-Aβ coaggregates from the brains of APOE4/4 AD patients are poorly lipidated.

A Representative two-color TIRF images of early-stage aggregates extracted from frontal cortices of APOE4 homozygotes, before and after immunoprecipitation (IP) with the non-lipidated-apoE4-specific antibody, HAE-4. B Sizes of individual Aβ-containing species before (number of aggregates: 3814, average size: 0.28 ± 0.69 µm) and after (number of aggregates: 3576, average size: 0.19 ± 0.26 µm) immunoprecipitation (data are plotted in log₁₀ scale in inset). C, D Effect of immunoprecipitation on colocalization between Aβ and apoE4 quantified by aggregate counting (C), and fluorescence intensity (D). Data points in panels (C, D) represent one of three biological replicates; error bars represent mean values +/− standard deviation of theree independent experiments. Statistical significance was calculated using a paired two-sample t-test. *P < 0.05, **P < 0.01, ***P < 0.001, ns, non-significant (P ≥ 0.05). Source data are provided as a Source Data file.

Fig. 9

Proposed mechanism for how apoE influences Aβ aggregation and modulates Aβ clearance and toxicity in  an isoform- and lipidation-specific manner.